Antagonistic and Detoxification Potentials of Trichoderma Isolates for Control of Zearalenone (ZEN) Producing Fusarium graminearum

Fungi belonging to Fusarium genus can infect crops in the field and cause subsequent mycotoxin contamination, which leads to yield and quality losses of agricultural commodities. The mycotoxin zearalenone (ZEN) produced by several Fusarium species (such as F. graminearum and F. culmorum) is a commonly-detected contaminant in foodstuffs, posing a tremendous risk to food safety. Thus, different strategies have been studied to manage toxigenic pathogens and mycotoxin contamination. In recent years, biological control of toxigenic fungi is emerging as an environment-friendly strategy, while Trichoderma is a fungal genus with great antagonistic potentials for controlling mycotoxin producing pathogens. The primary objective of this study was to explore the potentials of selected Trichoderma isolates on ZEN-producing F. graminearum, and the second aim was to investigate the metabolic activity of different Trichoderma isolates on ZEN. Three tested Trichoderma isolates were proved to be potential candidates for control of ZEN producers. In addition, we reported the capacity of Trichoderma to convert ZEN into its reduced and sulfated forms for the first time, and provided evidences that the tested Trichoderma could not detoxify ZEN via glycosylation. This provides more insight in the interaction between ZEN-producing fungi and Trichoderma isolates.

Both ZEN and its reduced forms are stable compounds, which exhibit hepatotoxic, hematotoxic, immunotoxic, and genotoxic effects. In addition, ZEN-related mycotoxins can competitively bind the estrogen receptors, causing alterations in genitals and reproduction disorders, posing a threat to human and animals health (Zinedine et al., 2007). To protect consumers, prevention before harvest seems to be an effective strategy for mycotoxin management (Jard et al., 2011;Tian et al., 2016a). Application of antagonistic biological control agents for controlling the toxigenic Fusarium spp., is a promising biological control based approach to reduce ZEN contamination. As potential antagonistic microbes, the genus Trichoderma has been widely studied for their capabilities against plant pathogenic fungi, and its biological control mechanisms mainly include faster growth speed and antibiotic production to compete nutrients and living space with pathogens, mycoparasitism mediated by producing cell wall degrading enzymes, and the ability to induce plant's defense system (Benítez et al., 2004;Sellamani et al., 2016;Tian et al., 2016a). Two Trichoderma isolates could effectively decrease the amount of mycotoxin ZEN produced by Fusarium spp. by a dual-culture assay (Gromadzka et al., 2009). In addition, other studies showed that some Trichoderma isolates also could inhibit mycotoxin deoxynivalenol (DON) production of Fusarium spp. (Busko et al., 2008;Matarese et al., 2012;Tian et al., 2016b). DON, a common type B trichothecene mycotoxin (Cuomo et al., 2007), usually co-occurs with ZEN in the foods and feeds (Molto et al., 1997;Castillo et al., 2002;Döll and Dänicke, 2011;Pietsch et al., 2013;Kovalsky Paris et al., 2014). Both DON and ZEN are frequently detected mycotoxins with high contamination levels (Stepien and Chełkowski, 2010). Recent work showed that DON could be bio-transformed into its modified form deoxynivalenol-3-glucoside (D3G) by Trichoderma isolates. D3G was generated in the defense of plants after infected by DON-producing pathogens, and D3G was regarded as a detoxification product of DON catalyzed by UDP-glucosyltransferases (Poppenberger et al., 2003;Schweiger et al., 2010;Li et al., 2015;Pasquet et al., 2016). Our recent study reported the occurrence of D3G in metabolism of selected Trichoderma isolates against DON producers (Tian et al., 2016b). However, little is known about the metabolism of ZEN in Trichoderma isolates. Thus, our particular interest was that whether Trichoderma spp. also possess the capacity to glycosylate ZEN into glycosylated forms for self-protection. Herein, the antagonistic potentials of Trichoderma isolates against ZENproducing F. graminearum and the metabolism of ZEN in Trichoderma isolates was investigated in this work. A targeted LC-MS/MS method for simultaneous determination of ZEN and its reduced forms (α-ZOL, β-ZOL, α-ZAL, β-ZAL, and ZAN) and glycosylated forms (Z14G, α-ZOL14G, and β-ZOL14G) was applied to explore the anti-toxigenic activity of antagonists and ZEN metabolization in Trichoderma isolates. It was observed that three Trichoderma isolates could effectively suppress the mycelia spread and mycotoxin production of ZEN-producing F. graminearum. In addition, results of ZEN-treated experiment showed that the tested Trichoderma isolates could not detoxify ZEN via glycosylation, but could convert ZEN to its reduced (α-ZOL and β-ZOL) and sulfated metabolites (Z14S and ZOL14S). As far as we know, this is the first report of the metabolic activity of Trichoderma isolates on ZEN, which would provide more insights in the interaction between mycotoxin ZEN producing fungi and antagonistic Trichoderma isolates.

Antagonistic Potentials of Trichoderma
Isolates on Growth and Mycotoxin Production of F. graminearum F1 The dual-culture test was performed to examine the antagonistic potentials of Trichoderma isolates for controlling ZENproducing F. graminearum as described before (Matarese et al., 2012;Tian et al., 2016b). The mycelial disks (F. graminearum and Trichoderma spp. combinations) from actively-growing colonies were placed on a 9-cm diameter dish. In addition, a disk of F. graminearum was placed without disks of Trichoderma isolates (control). The Fusarium-Trichoderma combinations, as well as the controls were incubated at 25 • C for 2 weeks. To evaluate the inhibition efficacy of Trichoderma isolates on mycelia growth of F. graminearum F1, the radius of each F. graminearum colony was measured to create growth curve as described in Matarese et al. (2012), and then the data were subjected to analysis of variance of regression to compare the slope of growth curves of the pathogen in the presence/absence of tested Trichoderma isolates.

Treatment of Trichoderma Isolates with ZEN
The tested Trichoderma isolates were activated on PDA medium at 25 • C. Then, a mycelial disk of each activated Trichoderma strain was moved from the edge of the colony, and inoculated in a new dish containing of 10 ml PDA medium mixed with ZEN at different concentrations (0, 0.5, 1, 2, and 4 µg/ml). Pure ZEN was added into the PDA medium as described before (Utermark and Karlovsky, 2007). The mycelial disk of Trichoderma isolates placed on PDA medium without mycotoxin was as control. The dishes were incubated at 25 • C, and growth radius of the tested Trichoderma isolates were measured two times a day until the mycelia of tested strains spread over the whole dish. Regression analysis of the growth data was performed to evaluate the inhibitory effects of ZEN on mycelial growth of Trichoderma isolates.

Mycotoxin Extraction
After incubation, the PDA medium and mycelia in the dish were dried and ground into powder for preparation, followed by adding 10 ml of ACN/water/formic acid (84/15.9/0.1, v/v/v) solution. The mixture was then shaken for 10 min, and subjected to ultrasonication for 30 min. Next the mixture was centrifuged at 4,000 rpm for 30 min. 1 ml of the supernatant was passed through a Cleanert MC column for clean-up by following the manufacturer instructions. Thereafter, the purified mixture was moved into a new tube, and evaporated to dryness by nitrogen gas at 45 • C. Finally, the residue was re-dissolved with 1 mL of methanol/water (1/1, v/v) and filtered through a 0.22-µm filter for LC-MS/MS analysis.
For MS/MS analysis, the parameters were set as follows: interface voltage of 2.5 kV (ESI − ); desolvation temperature of 270 • C; nebulizing gas (N 2 ) pressure of 50 psi and drying gas (N 2 ) pressure of 25 psi; heat block temperature of 300 • C. The quantitation and identification of target mycotoxins were performed in selected reaction monitoring (SRM) mode. The optimized MS/MS parameters for each analyte are listed in Table 1. Xcalibur TM software (Thermo Fisher Scientific, San Jose, CA, USA, 2011) was used for data processing.

Statistical Analysis
All the experiments were set up in triplicates. Data were presented as the mean ± standard error of the mean (SEM), and data were subjected to two-tailed student's t-test analysis or regression analysis with Graphpad Prism 5.01 (GraphPad Software, San Diego, CA, USA, 2007).

Effect of Trichoderma Isolates on Growth of ZEN-Producing F. graminearum F1 in Co-culture Assay
Trichoderma isolates and F. graminearum F1 were co-cultured on PDA medium. These antagonists rapidly occupied the living space, and inhibited the mycelial spread of F. graminearum F1 due to their antagonistic activities (Figure 2). Growth inhibition is the main pattern for antagonists to manage the pathogen, and we found that seven of the tested Trichoderma isolates were able to significantly suppress the mycelial growth of F. graminearum F1 (Table 2). Furthermore, T. harzianum Q710613, T. atroviride Q710251 and T. asperellum Q710682 displayed more effective inhibitory effects, as we observed that these three Trichoderma isolates overgrew the colony of F. graminearum F1, and the mycelium of Fusarium in these pathogen-antagonist combinations were restricted to extend vertically (Figure 2). These results showed that T. harzianum Q710613, T. atroviride Q710251, and T. asperellum Q710682 were more effective suppressors for controlling the mycelia growth of F. graminearum F1.

Effect of Trichoderma Isolates on Mycotoxin Production of ZEN-Producing F. graminearum F1 in Co-culture Assay
To investigate the effect of Trichoderma isolates on ZEN-related mycotoxins production of F. graminearum, and verify whether Trichoderma could glycosylate ZEN into glycosylated forms. ZEN and its reduced derivatives (α-ZOL, β-ZOL, α-ZAL, β-ZAL, and ZAN), as well as three glycosylated mycotoxins (Z14G, α-ZOL14G, and β-ZOL14G), were monitored in this work.
F. graminearum F1 used in this dual-culture assay could produce 1562 µg/g ZEN, 27 µg/g ZAN, 2.6 µg/g α-ZOL, and 15 µg/g β-ZOL on PDA medium (Figure 3). ZEN was the major mycotoxin produced by the tested F. graminearum. When F. graminearum F1 grew against antagonistic Trichoderma isolates, the amount of mycotoxins produced by F. graminearum F1 was inhibited because of the antagonistic activity of Trichoderma. The inhibition rate of ZEN production ranged from 9 to 97%, for ZAN ranged from 22 to 98%, for α-ZOL ranged from 31 to 87%, and for β-ZOL ranged from 34 to 89%. Among the tested isolates, T. koningii GIM3.137 exhibited weaker inhibitory effects on the mycotoxin production of F. graminearum F1 (Figure 3). While T. harzianum Q710613, T. atroviride Q710251, and T. asperellum Q710682 exhibited a better efficiency in inhibiting mycotoxin production of Fusarium. When co-cultured with these three isolates, the amount of ZEN and ZAN produced by F. graminearum F1 was inhibited by over 93%, and the amount of α-ZOL and β-ZOL produced by F. graminearum F1 was inhibited by over 80%. Unexpectedly, no glycosylated forms of ZEN and ZOL were observed when Trichoderma isolates were co-cultured with ZEN-producing Fusarium. The experiment, as described below, pinpointed the treatment of Trichoderma isolates with ZEN, and analyzed the metabolites to confirm the obtained result.

Analysis of the Metabolites When
Trichoderma Grew on PDA Medium Amended with Pure Mycotoxin ZEN

The Inhibition of ZEN on Growth of Trichoderma
It has been reported that ZEN could inhibit the growth of some filamentous fungi, which help ZEN-producing Fusarium species compete with other microbes, so ZEN is regarded as a contributive factor for ZEN-producers (Utermark and Karlovsky, 2007). Firstly, we evaluated the inhibitory effects of ZEN on growth of various Trichoderma isolates, as the toxic effects of ZEN have not yet been elucidated in the genus Trichoderma. These effects were evaluated by comparing the mycelia growth rate of Trichoderma spp. when exposed to different ZEN concentrations (0, 0.5, 1, 2, and 4 µg/ml) on PDA medium. Results demonstrated that ZEN exhibited significant fungal toxic effects on Trichoderma isolates: the mycelia growth of five isolates were significantly inhibited by 1 µg/ml of ZEN. While for T. longibranchiatum GIM3.534, T. atroviride Q710251 and T. asperellum Q710682, the inhibitory effects were observed when treated with 2 µg/ml of ZEN (Figure 4).
Subsequently, all the Trichoderma isolates treated with 2 µg/ml ZEN were selected for further study of the metabolic activity of Trichoderma isolates on ZEN. The whole medium and mycelia were collected and prepared for analysis.

Analysis of Metabolites by LC-HRMS
Besides glycosylation, sulfation is another detoxification process for different mycotoxins in plants and fungi. Zearalenone-14sulfate (Z14S) was found to be a metabolite in Arabidopsis thaliana, Rhizopus spp. and Aspergillus spp. when exposed to ZEN, and zearalenol-14-sulfate (ZOL14S) was observed as a fungal metabolite in ZEN-treated trial (Berthiller et al., 2007;Brodehl et al., 2014). However, there were no reference standards (Z14S and ZOL14S) available for quantitative analysis by LC-MS/MS. For detection of sulfated forms of ZEN in prepared    (Brodehl et al., 2014;Binder et al., 2017). In conclusion, these results revealed the presence of Z14S and ZOL14S in Trichoderma metabolism with ZEN treatment. For the first time, we reported that antagonistic Trichoderma isolates possess the detoxification capability to sulfate ZEN, and these sulfated forms would be quantified when reference standards are available in future.

DISCUSSION
The effective methods to manage mycotoxin contamination include application of antagonistic microbes to prevent mycotoxin production before harvest and using detoxification agents to treat contaminated foodstuffs (Atanasova-Penichon et al., 2016;Perczak et al., 2016;De Saeger and Logrieco, 2017). Due to its potentials to control plant pathogens, the non-toxigenic Trichoderma genus has been intensively investigated (Benítez et al., 2004). In the present study, we co-cultured Trichoderma isolates with ZEN-producing F. graminearum F1 to assess the inhibition and detoxification capacities of tested Trichoderma isolates T. harzianum Q710613, T. atroviride Q710251, T. asperellum Q710682 displayed promising antagonistic potentials to control the growth and mycotoxin production of ZEN-producing F. graminearum F1. In order to exhaustively access their antagonistic potentials, these Trichoderma isolates were dual cultured with other ZEN-producing Fusarium species in the later experiment. These antagonists exhibited prominent inhibitory actions on both mycelia spread ( Figure S1) and mycotoxin production ( Figure S2) of the ZEN-producers. Taken together, our recent progress indicates that the three candidates are potential biological control antagonists to combat toxigenic fungi, which deserve attention and further analysis of their ability to control disease development in field experiments.  Plants possess the capacity to detoxify phytotoxic compounds into low-toxic products after infected by toxigenic fungi. Mycotoxins are toxic xenobiotics for plants, which can be conjugated to polar metabolites in detoxification reactions of plants, generating low-toxic metabolites with structure changed (Berthiller et al., 2016). The detoxification mechanisms of plants against mycotoxins mainly include three phases: transformation phase, conjugation phase and compartmentation phase (Berthiller et al., 2007). Both glycosylation and sulfation are common processes in detoxification reactions of different plants against mycotoxins (Lemmens et al., 2016). It has been showed that DON and ZEN can be bio-transformed into glycosylated and sulfated forms in the detoxification process of plants (Berthiller et al., 2007). The UDP-glucosyltransferase (UGT) capable of converting DON into D3G was firstly identified in Arabidopsis thaliana (Poppenberger et al., 2003), and then the first UGT capable of converting ZEN into Z14G was also identified in Arabidopsis thaliana (Poppenberger et al., 2006). With regard to DON, our previous work proved that Trichoderma spp. possess the ability to metabolize DON into its glycosylated form (Tian et al., 2016b). Consequently, we explored whether Trichoderma isolates possess the ability to modify mycotoxin ZEN in this work. Not similar to plants, the tested Trichoderma isolates could not bio-transform ZEN into its glycosylated forms, but could convert ZEN into its reduced and sulfated form(s) (Figure 7). Evidence was provided that Trichoderma isolates were able to detoxify ZEN via sulfation when competing with ZEN-producing F. graminearum.
LC-MS/MS is a useful tool for simultaneous determination of different co-existing mycotoxins when standards are available (Righetti et al., 2016). However, it is still challenging to identify and quantify modified mycotoxins by using LC-MS/MS due to the limited commercial availabilities of modified mycotoxin standards. The HRMS has the advantage of providing accurate ion mass-to-charge that can be used for structure elucidation of compounds in a targeted or untargeted strategy (Righetti et al., 2016), so it has become a promising tool for analyzing the predicted metabolites without standards (De Boevre et al., 2016;Righetti et al., 2016). In our current work, the HRMS was used to obtain accurate mass and fragmentation patterns of analytes, and the sulfated metabolites (Z14S and ZOL14S) produced by Trichoderma isolates were discovered. This contributes to further investigations of the defense mechanism of biological control agents against toxigenic fungi.

AUTHOR CONTRIBUTIONS
AW and YT conceived and designed the experiments; YT, YLT, and ZY performed the experiments and analyzed the data; AW and YT wrote the paper; YL, JC, SD, and MD contributed materials and amended the manuscript.