Genetically engineered Fusarium as a tool to evaluate the effects of environmental factors on initiation of trichothecene biosynthesis

Authors


  • Editor: Claire Remacle

Correspondence: Makoto Kimura, Plant & Microbial Metabolic Engineering Research Unit, Discovery Research Institute (DRI), RIKEN, 2-1 Hirosawa, Wako, Saitama 351-0198, Japan. Tel.: +81 48 467 9796; fax: +81 48 462 4394; e-mail: mkimura@riken.jp

Abstract

Fusarium graminearum was engineered for expression of enhanced green fluorescent protein gene (egfp) as a reporter regulated in a manner similar to Tri5, a key pathway gene in trichothecene biosynthesis. Using the transgenic fungus, it was found that the reporter gene was induced to express in aerial hyphae developed on trichothecene noninducing medium YG solidified by agar. Unexpectedly, the transcriptional activation of egfp was markedly suppressed by adding NaCl that does not significantly affect fungal growth. As suggested by these findings, wild-type F. graminearum that formed aerial hyphae on YG agar plates produced trichothecenes and the production was effectively suppressed by adding 1% NaCl to the agar. To evaluate the effects of abiotic stress on the expression of trichothecene biosynthesis (Tri) genes, a sensitive plate assay was established using GYEP medium (which very weakly induces trichothecene production) solidified with gellan gum. Using this assay, triazole fungicides were shown to cause transcriptional activation of egfp at sublethal concentrations. Indeed, trichothecene production significantly increased when F. graminearum was grown in rice medium (which moderately induces trichothecene) amended with low doses of tebuconazole. The real-time monitoring system described here may help predict the risks of trichothecene contamination by the fungus under various environmental conditions.

Introduction

Trichothecenes are a large family of sesquiterpene mycotoxins characterized by a 9,10-double bond and a 12,13-epoxide (Ueno, 1984). Numerous kinds of trichothecenes are produced by taxonomically unrelated fungal genera, including Fusarium, Trichothecium, Myrothecium, Stachybotrys, and other saprophytic fungi. Among the trichothecene producers, Fusarium culmorum and Fusarium graminearum are major hemibiotrophic pathogens that cause a devastating disease in cereal crops, collectively known as Fusarium head blight (FHB) (Goswami & Kistler, 2004). In addition to significant losses of yield, these Fusarium species cause mycotoxin contamination in infected grains. Therefore, most of the concerns about trichothecenes are related to Fusarium trichothecenes such as deoxynivalenol, nivalenol, and acetylated derivatives thereof, which actually pose serious health threats to humans and animals (Pestka & Smolinski, 2005).

The biosynthesis of Fusarium trichothecenes has been studied in detail and most of the genes involved (Tri genes) have been identified (Desjardins et al., 1993; Kimura et al., 2001). The first pathway gene to be isolated from a trichothecene-producing fungus was Tri5 (formerly Tox5), which encodes a key sesquiterpene cyclase for the committed step in the biosynthesis (Hohn & Beremand, 1989). Tri5 is located nearly at the center of the trichothecene gene cluster together with other Tri genes essential for toxin biosynthesis (Hohn et al., 1993; Kimura et al., 1998, 2003; Brown et al., 2001; Lee et al., 2002). One of these, Tri6, encodes a zinc finger transcription factor involved in the regulation of Tri5 and other Tri genes (Proctor et al., 1995). Tri6 itself is regulated by another regulatory protein encoded by Tri10 (Tag et al., 2001). The expression of Tri genes is also affected by changes in the G-protein signal transduction pathway (Tag et al., 2000). However, little is known about the effects of environmental factors (e.g., pesticides, ionic strength, temperature, oxidative stress, etc.) on the regulation of trichothecene biosynthesis.

Several fungicides are used under various field conditions for the control of FHB. However, the efficacy with which a fungicide reduces the disease's severity is not necessarily correlated with the toxin levels in grains. In some cases, levels of trichothecenes actually increase when a fungicide is applied (D'Mello et al., 1998). Considering the number and combination of individual factors, it would be useful to develop a simple protocol to rapidly assess the risks of increased toxin production by Fusarium under various conditions.

To monitor the expression of foreign genes, green fluorescent protein genes are often used as a nondestructive reporter in many organisms, including fungi. To evaluate the effects of various environmental factors on the initiation of trichothecene biosynthesis, transgenic F. graminearum was generated, which contains this highly sensitive reporter gene in place of the coding region of Tri5. In this paper, the effects of stress factors (salt, fungicides, and active oxygen) that resulted in the altered expression of the reporter gene in the fungus was reported.

Materials and methods

Strain, reagents and media

Fusarium graminearum MAFF 111233 (Fusarium asiaticum) used in this study shows coproduction of 4-acetylnivalenol (4-ANIV) and 4,15-diacetylnivalenol (4,15-diANIV). Standard samples of 4-ANIV and 4,15-diANIV were purified from this fungal strain grown in RF medium (Tokai et al., 2005). All the fungicides used in this study (Table 1) were purchased from Wako Chemical Industries, Ltd. (Osaka, Japan). The following media were used for fungal cultures; YG medium (2% glucose and 0.5% yeast extract), YG agar (YG medium solidified with 1.5% agar), GYEP gellan gum (5% glucose, 0.1% yeast extract, 0.1% peptone, and 1% gellan gum), rice medium [4% rice (30 min boiled and glass filtered), 3% sucrose, and 0.1% yeast extract], and rice agar (rice medium solidified with 1.5% agar). To examine the effects of tebuconazole and trifloxystrobin on production of trichothecenes, the fungal strain was cultured in rice medium supplemented with the fungicides with gyratory shaking (120 rpm at 25°C).

Table 1.   Activation of egfp expression by sublethal concentrations of various fungicides in the transgenic Fusarium graminearum strain*
FungicidesConcentration
(μg mL−1)
EGFP
intensity
Colony size
(% diameter)
  • *

    Incubated at 25°C for 4 days.

  • No (−), little (±), or strong (+<++) EGFP fluorescence.

Control 100
Azoxystrobin0.167
0.05±72
0.0193
Kresoxim-methyl0.165
0.05±69
0.0185
Trifloxystrobin0.154
0.0552
0.0161
Difenoconazole0.0133
0.005±45
0.001++69
Ipconazole0.01++46
0.005++57
0.001+83
Metconazole0.01++60
0.005++63
0.001++73
Tebuconazole0.1++50
0.05++68
0.01++89
Tridemorph5047
25±64
10+67
Thiophanatemethyl536
182
0.1100
Captan584
1102
0.1103
Iprodione538
192
0.1104

Construction of targeted gene replacement vector

The neo gene was amplified from pBF-Srf-Neo (Tokai et al., 2005) by PCR with primers NEO-F (5′-AGCATCGATTCGCATGATTGAACAAGATG-3′) and NEO-R (5′-AGGCCTTCAGAAGAACTCGTCAAGAA -3′), digested with ClaI and StuI (restriction sites underlined), and the 0.8-kb fragment was recovered. The plasmid pNF312 was constructed by replacing the 0.5 kb ClaI–SmaI fragment of pBF312 (Banno et al., 2003), a vector that contains a Clontech's synthetic gene (egfp) encoding a red-shifted mutant version (S65T) of EGFP, with the 0.8 kb ClaI–StuI fragment of neo (Beck et al., 1982). The gene replacement vector, pProm5GFPN, was constructed by replacing the complete coding region of Tri5 with pNF312 using the inverse-PCR (IPCR) method (Akiyama et al., 2000) as follows (see Fig. 1a): (1) the region containing Tri5 was amplified by long PCR with inward primers 1F (5′-CTCGGGATCCGACGTAACGTCGCAATTGAG-3′) and 2R (5′-AGGAGGATCCCGTAATCTTCAAATGGTGCT-3) containing a BamHI recognition site (underlined), which does not exist in the corresponding PCR products, (2) the amplified products were self-ligated after digestion with BamHI, (3) the flanking regions were amplified by IPCR with primers 4R (5′-ACCGGTGGTGTATTGGTAACAGTTATTC-3′) and 3F (5′-GCGGCCGCCCGAATGCGAGTTTAGAAGT-3′) containing AgeI and NotI recognition sites, respectively (underlined), and (4) the IPCR products were cloned between the AgeI and NotI recognition sites upstream of egfp in pNF312.

Figure 1.

 Construction of a genetically engineered Fusarium graminearum strain for real-time monitoring of Tri gene expression. (a) Targeted replacement of Tri5 with egfp in the core trichothecene gene cluster. Homologous regions between the genomic DNA and replacement vector pProm5GFPN are shown by thick lines. (b) Southern blot analysis of DNA from wild-type and transgenic F. graminearum. The locations of probes A and B are illustrated in (a). (c) RT-PCR of RNA from the transgenic F. graminearum grown in YG medium and rice medium. Control reactions (RT; –) did not yield the PCR products. (d) Northern blot analysis of RNA from the transgenic F. graminearum grown in YG medium and rice medium. The formaldehyde gel stained with ethidium bromide is shown below each lane to demonstrate equal loading of RNA samples. Although the digital image is not clear, an extremely faint band of the egfp transcript could be seen on the original blot in the lane of rice medium. (e) Epifluorescence micrographs (BL, bright field; FL, fluorescence) of wild-type and transgenic F. graminearum grown on a cellophane sheet placed on YG agar or rice agar plates. Scale bar; 100 μm.

Generation of a transgenic F. graminearum strain for real-time monitoring of Tri gene expression

Fusarium graminearum was transformed with BamHI-linearlized pProm5GFPN according to the method as described previously (Tokai et al., 2005). For selection of the Fusarium transformants, 20 μg mL−1 of G418 (Sigma) was added to YG agar (the bottom layer) and 1% seaplaque (the upper layer; Takara Bio Inc., Otsu, Japan), both containing 1 M sorbitol. To confirm the expression pattern of egfp under an epifluorescence microscope (DM RA, Leica Ltd, Cambridge), mycelial plugs of the transgenic fungus were cultured on cellophane sheets placed on YG agar or rice agar. After appropriate incubation periods, the sheets with the mycelia were stripped off from the agar and used for microscopic observation.

Effects of stress factors on expression of egfp

The suppressive effect of salt stress on the expression of egfp in aerial hyphae was examined by adding NaCl (0%, 1%, and 2%) to YG agar. To examine the stimulative effects of fungicides (Table 1) and H2O2 (0, 1, 2.5, and 5 mM) on activation of the reporter gene, GYEP gellan gum was used as the culture media. In either case, the plates were incubated at 25°C and the mycelia were observed under an epifluorescence stereomicroscope (MZIII; Leica).

Quantification of type B trichothecenes

For the quantification of type B trichothecenes produced by F. graminearum grown on YG agar plates (14 mL), half of the culture with agar was extracted with 7 mL of ethyl acetate saturated with NaCl. For the toxin quantification of the fungus grown in rice medium, 5 mL of the culture supernatant was extracted with an equal volume of ethyl acetate saturated with NaCl. In either case, the ethyl acetate extract was evaporated under a gentle stream of nitrogen and the dried material was reconstituted in 8 mL of acetonitrile/water (84 : 16). It was then applied to a MycoSep 227 column (Romer Labs®, Union, MO), and 4 mL of the eluate was recovered. After drying, the sample was dissolved in 0.1 mL of water/acetonitrile/methanol (80 : 15 : 5) and an aliquot (40 μl) was analyzed using a PEGASIL ODS column (diameter, 4.6 mm; length, 250 mm; Senshu Scientific Co., Tokyo, Japan) connected to an HPLC system (SCL-10A; Shimadzu, Kyoto, Japan). The column was run by isocratic elution with water/acetonitrile/methanol (80 : 15 : 5) at a flow rate of 1.0 mL min−1, and the eluates were monitored at 220 nm. The total amount of type B trichothecenes was calculated from the peak areas of 4-ANIV and 4,15-diANIV, which gave a nearly equal value of molar absorbance.

Quantification of ergosterol contents

To evaluate the fungal biomass on YG agar plates, ergosterol content was analyzed (Bluhm & Woloshuk, 2005). Half of the culture with agar was extracted with 10 mL of chloroform/methanol (2 : 1), the solvent layer recovered and evaporated under nitrogen, and the sample dissolved in 3 mL of methanol. Ergosterol was analyzed using a PEGASIL ODS column at a flow rate of 1 mL min−1 with 100% methanol. The eluates were monitored at 282 nm and the peak area of samples was normalized to an ergosterol standard (Wako) of known concentration.

Hybridization of nucleic acids and reverse transcription-polymerase chain reaction (RT-PCR)

For Southern and Northern blot analyses, a digoxigenin-labeled probe was prepared using a PCR DIG Probe Synthesis kit (Roche Diagnostics GmbH, Mannheim, Germany) with the following primers; probe A, primers 5F (5′-ATGGAGAACTTTCCCACCGA-3′) and 6R (5′-TCACTCCACTAGCTCAATCG-3′); probe B, primers 1F and 4R; probe C, EGFP-U/ATG (5′-ATGGTGAGCAAGGGCGAGG-3′) and EGFP-D/TAA (5′-TTACTTGTACAGCTCGTCC-3′). Nucleic acids were transferred to a Nytran® SuperCharge membrane (Schleicher & Schell GmbH, Dassel, Germany) using a Turboblotter (Schleicher & Schell) blotting apparatus. Hybridizations were carried out using the DIG Easy Hyb (Roche Diagnostics) hybridization solution at 68°C (Southern) or 50°C (Northern). Standard hybridization and washing techniques recommended by the manufacturer were used. For RT-PCR of egfp, the cDNA was synthesized from total RNA (treated with the RNase-free DNase I) with the Superscript First-strand Synthesis System (Invitrogen, Carlsbad, CA) and used as a template for the PCR with primers EGFP-U/ATG and EGFP-D/TAA.

Results

Characterization of a transgenic Fusarium strain carrying egfp in place of Tri5

For the real-time monitoring of Tri gene expression, a transgenic F. graminearum strain carrying egfp under the control of the Tri5 promoter was produced. As shown in Fig. 1a, targeted gene replacement was conducted by transforming the fungus with a BamHI-linearized gene replacement vector, pProm5GFPN. Genomic DNA was isolated from five G418 resistant colonies and used for PCR with primers 1F and 2R to screen for a transformant with the targeted gene replacement. A single PCR product of 9.8 kb (i.e., without the wild-type 4.8 kb product), expected from targeted gene replacement, was obtained from the DNA of two transformants (data not shown). One of the transformants, strain T1, was used for further analyses. In the DNA blot analysis, strain T1 lost the wild-type SphI band (3.3 kb) when hybridized with probe A (Tri5 coding region) and showed a reasonably sized shifted XhoI band (from 2.3 kb of the wild-type to 8.6 kb of T1) when hybridized with probe B (upstream region of Tri5) (Fig. 1b). The egfp mRNA was detected as a cDNA of expected size by RT-PCR (Fig. 1c; 720 bp) when strain T1 was cultured in rice medium that moderately induces trichothecene production, but not in YG medium (trichothecene noninducing medium). Northern blot analysis revealed that the expression level of the gene was as low as the detection limit of the technique; an extremely faint band of the probe-RNA hybrid was barley observed on the original blot (although it is hardly visible on the digital image of Fig. 1d). Mycelial plugs of the transgenic strain were cultured on a cellophane sheet placed on YG agar or rice agar plates. As shown in Fig. 1e, the transgenic fungus grown on rice agar plates, but not on YG agar plates, revealed strong fluorescence during the short incubation period (4 days of incubation at 25°C) under the epifluorescence microscope. Because toxin accumulates marginally if at all when YG medium is used for liquid culture of the wild-type strain, this result reasonably explains the features of F. graminearum regarding the transcriptional activation of Tri genes by nutritional factors.

Tri gene expression is induced in aerial hyphae, which is suppressed by hypertonic NaCl that least affects fungal growth

To establish a system for investigating the effects of abiotic stress on egfp expression, the EGFP fluorescence of the fungus was next tried to be detected directly under the epifluorescence stereomicroscope. Because the transcriptional activation of egfp does not occur when the fungus is grown in YG medium (Fig. 1c and d) or on YG agar (Fig. 1e), this system was expected to be used to identify chemicals that have stimulative effects on trichothecene biosynthesis. Indeed, no fluorescence was observed in the fresh mycelia actively extending outward in contact with the surface of the agar (Fig. 2a; see area ‘c’ of strain T1 grown on YG agar). However, with longer incubation periods, the transgenic fungus formed aerial hyphae and showed very bright EGFP fluorescence (Fig. 2a; see area ‘a’ and ‘b’ of strain T1 grown on YG agar) when observed under the epifluorescence stereomicroscope. Unexpectedly, addition of NaCl to the medium significantly suppressed the development of the EGFP fluorescence in aerial hyphae; at a concentration of 2% (0.34 M), the EGFP fluorescence was no longer observed [Fig. 2a; see areas ‘a’, ‘b’, and ‘c’ of strain T1 grown on YG agar (2% NaCl)]. At this concentration, NaCl did not markedly affect mycelial growth or the formation of aerial hyphae.

Figure 2.

 Hypertonic NaCl suppresses trichothecene production in the aerial hyphae of Fusarium graminearum developed on YG agar plates. (a) Use of the transgenic F. graminearum strain T1 for real-time monitoring of the Tri gene expression. The transgenic fungus was observed under the epifluorescence stereomicroscope. Enclosed boxes (a–c) on the photograph of YG agar plates indicate the areas observed under the epifluorescence stereomicroscope (BF, bright field; FL, fluorescence). (b) Production of trichothecenes by F. graminearum grown on YG agar plates after 5 and 10 days of incubation at 25°C (three independent experiments). Colored bars indicate the relative molar amounts of 4-ANIV/4,15-diANIV in proportion to the 0% NaCl control at day 5 in each experiment. Ergosterol content (μg g−1 medium) is shown at the bottom of the graph. Blue, 0% NaCl; Green, 1% NaCl; Red, 2% NaCl.

Wild-type F. graminearum was grown on YG agar plates to examine whether (1) trichothecenes are produced in aerial hyphae, and if so, (2) NaCl suppresses trichothecene production as predicted by the disappearance of the EGFP fluorescence. Three independent experiments, each using a fresh mycelial plug from the same plate as an inoculum, were carried out. The trichothecene content was expressed as a ratio to the 0% control (11.8, 2.2, and 6.2 μg of trichothecene per gram culture 5 days after inoculation) in each experiment. After 5 and 10 days of incubation on YG agar plates with 0%, 1%, or 2% NaCl, total amounts of ergosterols and trichothecenes were quantified by HPLC analysis. In the presence of 1% and 2% NaCl, the fungal biomass represented by the ergosterol content decreased only slightly compared with that obtained in the absence of NaCl (Fig. 2b; values at the bottom of the graph). In contrast to the small effects of NaCl on fungal growth, 1% NaCl markedly suppressed the accumulation of trichothecenes and 2% NaCl completely inhibited the toxin production (Fig. 2b; colored bars). The inhibitory effects of the chemicals on toxin production were also observed with 1% (0.13 M) KCl, but not with 0.34 M sorbitol (data not shown). The results suggest that salt stress inhibits trichothecene biosynthesis in aerial hyphae.

Establishment of a sensitive plate assay (with GYEP gellan gum plates) to evaluate the effects of various fungicides on toxin production

Because YG agar plates were not suitable for easy monitoring of the stimulative effects of certain fungicides on trichothecene production, a sensitive plate assay system was next established, which enables prediction of such effects. The assay that was developed uses GYEP medium (which very weakly induces trichothecene production by the F. graminearum strain) in place of YG medium for stimulating trichothecene production and the polysaccharide gelling agent gellan gum as a substitute for agar to suppress the formation of aerial hyphae. GYEP gellan gum plates have almost perfect visual clarity to enable easy and more-sensitive stereomicroscopic observation of the EGFP fluorescence on the surface of the medium.

Mycelial plugs of the engineered F. graminearum were placed at the center of GYEP gellan gum plates and incubated in the presence of various concentrations of fungicides at 25°C (Fig. 3a; see also Supplementary MultiMedia S1 for an overview of the monitoring system). By monitoring the outermost fresh part of the transgenic mycelia on GYEP gellan gum plates (Fig. 3a; boxed area), the transcriptional activation of egfp by certain fungicides at sublethal concentrations could be observed (Table 1). The stimulative effects were observed with a group of fungicides collectively known as triazoles: fungicides with other modes of action (e.g., strobilurins, benzimidazole, and dicarboximide) did not cause transcriptional activation of egfp. Typical examples of stereomicrographs are shown in Fig. 3b; i.e., while the reporter gene expression was not detected in the mycelia treated with various concentrations of trifloxystrobin, it was observed in the mycelia treated with 0.05 μg mL−1 of tebuconazole. The range of triazole concentrations effective for Tri gene activation was rather narrow and the stimulative effect was observed only in the range of c. 0.01–0.10 μg mL−1 of tebuconazole. At concentrations higher than 0.10 μg mL−1 of tebuconazole, the growth of the fungus was significantly impaired and morphologically abnormal mycelia clustered without the EGFP fluorescence. A typical example is shown in Fig. 3b (strain T1 at 0.25 μg mL−1 of tebuconazole).

Figure 3.

 Evaluation of the effects of various fungicides on expression of Tri genes. (a) Overview of the monitoring system. The stimulative effect of fungicides on Tri gene expression is most sensitively detected by comparing the EGFP fluoresence of the boxed area to that of the control (i.e., without fungicides). See also Supplementary MultiMedia S1 for overview. (b) Epifluorescence stereomicrographs of the wild-type and strain T1 of F. graminearum grown on GYEP gellan gum medium amended with tebuconazole and trifloxystrobin (4 days at 25°C). White bars represent relative growth rates of the fungus (i.e., diameter of the colony) compared with the control. (c) Production of trichothecenes by F. graminearum grown in rice media (three independent experiments). Compared with the growth on GYEP gellan gum plates, the fungal growth in rice media was less affected by the fungicides. Colored bars indicate the relative molar amounts of 4-ANIV/4,15-diANIV in proportion to the control at day 3 (without fungicides) in each experiment. Fungal biomass (dry weight mg mL−1 medium) is shown at the bottom of the graph. C, control (blue); Te, tebuconazole (green); Tr, trifloxystrobin (red).

The amount of trichothecenes affected by the application of the fungicides, tebuconazole and trifloxystrobin, was determined using liquid culture to exclude the effect of aerial hyphae formation activating trichothecene production. Rice medium (i.e., moderate level of trichothecene induction) was used instead of GYEP medium (i.e., very low level of trichothecene induction) because the former is more appropriate to demonstrate clearly both the suppressive and stimulative effects of fungicides. Three independent experiments were carried out and the trichothecene content was expressed as a ratio to the control (16.6, 15.2, and 1.4 μg of trichothecene per mL culture 3 days after inoculation). As shown in Fig. 3c, the addition of tebuconazole (0.05 μg mL−1) increased the amount of trichothecenes produced by wild-type F. graminearum by c. two to threefold without significant changes in fungal biomass (dry weight). In contrast, trifloxystrobin (0.05 μg mL−1) markedly decreased the amount of the toxin. The result suggests that the plate assay system that was developed is useful for predicting the stimulative effects of various fungicides on enhancement of trichothecene production by wild-type F. graminearum in rice medium.

Use of GYEP gellan gum plates for investigating the effects of oxidative stress on Tri gene expression

To examine whether the Tri gene-monitoring system described here can be used to evaluate of the effects of active oxygen associated with hypersensitive reactions in plant defense, H2O2 was added to the GYEP gellan gum plates. Because oxidative stress is known to stimulate production of trichothecenes in toxigenic Fusarium species (Ponts et al., 2006), the EGFP fluorescence was expected to be detected in the engineered F. graminearum as in the case of triazole fungicides. When the GYEP gellan gum plates were amended with H2O2 at concentrations that do not inhibit fungal growth, the EGFP fluorescence was observed (Fig. 4). The result demonstrates the utility of this assay system for predicting the effects of oxidative stress on the transcriptional activation of Tri genes.

Figure 4.

 Evaluation of the effects of H2O2 (0, 1, 2.5, and 5 mM) on expression of Tri genes (4 days at 25°C). The mycelia were observed under the epifluorescence stereomicroscope (BF, bright field; FL, fluorescence). White bars represent relative growth rates of the fungus (i.e., diameter of the colony) compared with the control (0 mM).

Discussion

Fusarium graminearum encounters various stress factors during the infection to host cereal plants. To evaluate the effects of such factors on trichothecene biosynthesis, a transgenic Fusarium strain was generated, which carries egfp in place of the coding region of Tri5 in the trichothecene gene cluster. RT-PCR analysis revealed that egfp connected to the Tri5 promoter was tightly regulated but it was difficult to detect the transcript under the trichothecene-producing conditions on Northern blots. A low transcript level of a foreign gene targeted to the Tri5 region was observed not only with egfp but also with other foreign genes, such as those for plant sesquiterpene biosynthesis (unpublished result). Regardless of the very low expression level of the transgene under the trichothecene-producing conditions, EGFP fluorescence could be detected, owing to the high sensitivity of this reporter gene. A similar case was previously reported in Saccharomyces cerevisiae, in which the fluorescence could be observed even if the transcript level of the reporter gene is under the detection limit of the Northern blot analysis (Chambers et al., 2004).

Although the toxin level is also influenced by the amount of farnesyl pyrophosphate (FPP), a branching point intermediate of the primary metabolic pathway (e.g., ergosterol) and the trichothecene (i.e., secondary metabolic) pathway, EGFP-based monitoring of Tri gene expression alone gave useful information for predicting enhancement of toxin production. This suggests that activation of egfp in the transgenic strain can reasonably be used to predict enhancement of trichothecene production in the wild-type strain. Indeed, genes necessary for the biosynthesis of FPP (e.g., mevalonate kinase genes, FPP synthase gene) are known to be transcriptionally activated when Tri gene are induced to express in Fusarium sporotrichioides (Tag et al., 2001; Peplow et al., 2003).

With the transgenic Fusarium strain for real-time monitoring of Tri gene expression, two new findings could be made: (1) trichothecene production is transcriptionally activated in aerial hyphae developed on YG agar (although YG medium does not induce trichothecene production), and (2) this transcriptional activation is markedly suppressed by adding NaCl (as low as 1%) to the agar. To predict the effect of applying the fungicide on the basis of the fluorescence, a highly sensitive plate assay system needed to be developed due to the low activity of the fungicides to cause expression of egfp (which is interfered with by the appearance of the fluorescence in aerial hyphae developed on YG agar). Using GYEP gellan gum that weakly induces toxin production and significantly suppresses aerial hyphae formation, the stimulative effects of triazole fungicides could be directly monitored on Tri gene expression under the epifluorescence steromicroscope. Interestingly, triazoles are inhibitors of cytochrome P450 lanosterol 14α-demethylase in the biosynthesis of ergosterol (Ghannoum & Rice, 1999). The inhibition of the ergosterol pathway may cause accumulation of the earlier intermediate FPP and may eventually trigger the transcriptional activation of Tri genes in F. graminearum.

Several authors have reported that sublethal concentrations of triazole fungicides (e.g., tebuconazole) directly stimulate trichothecene biosynthesis in vitro (D'Mello et al., 1998; Magan et al., 2002), but others have reported contradictory results (Matthies & Buchenauer, 1996; Covarelli et al., 2004). These controversial findings may be attributed to differences in the timing and quantity of fungicide application, growth stage of the fungus, and/or combined effects of various environmental factors. Indeed, use of the EGFP monitoring system suggested that triazoles activate the transcription of Tri genes only over a certain range of concentrations, with which the fungal growth is slightly inhibited (see Fig. 3b; strain T1 treated with 0.05 μg mL−1 of tebuconazole).

Although it was not possible to predict the suppressive effect of strobilurin fungicides (e.g., trifloxystrobin) on trichothecene production, the real-time monitoring system has significant advantages over the conventional analytical methods in speed, costs, and operations. Thus, the transgenic Fusarium established in this study will be useful for assessing the possible undesirable effects of abiotic stress activating toxin production during infection of the pathogen to host cereal grains. In addition, the transgenic strain may also be used as a parental strain for screening of mutants that are blocked in the signal transduction pathway of Tri gene activation. This is the reason why the most efficient fungal transformation marker gene hph (that confers hygromycin B resistance) was saved in this study.

Acknowledgements

Dr Isamu Yamaguchi, President of the Food and Agricultural Materials Inspection Center (FAMIC), is thanked for information about fungicides. Thanks are also due to Dr Paul Nicholson for detailed information about trichothecene production media of F. culmorum. This work was supported by the Program for Promotion of Basic Research Activities for Innovative Biosciences (PROBRAIN).

Ancillary