Detoxification and Excretion of Trichothecenes in Transgenic Arabidopsis thaliana Expressing Fusarium graminearum Trichothecene 3-O-acetyltransferase

Fusarium graminearum, the causal agent of Fusarium head blight (FHB), produces trichothecenes including deoxynivalenol (DON), nivalenol (NIV), and 3,7,15-trihydroxy-12,13-epoxytrichothec-9-ene (NX-3). These toxins contaminate grains and cause profound health problems in humans and animals. To explore exploiting a fungal self-protection mechanism in plants, we examined the ability of F. graminearum trichothecene 3-O-acetyltransferase (FgTri101) to detoxify several key trichothecenes produced by F. graminearum: DON, 15-ADON, NX-3, and NIV. FgTri101 was cloned from F. graminearum and expressed in Arabidopsis plants. We compared the phytotoxic effects of purified DON, NIV, and NX-3 on the root growth of transgenic Arabidopsis expressing FgTri101. Compared to wild type and GUS controls, FgTri101 transgenic Arabidopsis plants displayed significantly longer root length on media containing DON and NX-3. Furthermore, we confirmed that the FgTri101 transgenic plants acetylated DON to 3-ADON, 15-ADON to 3,15-diADON, and NX-3 to NX-2, but did not acetylate NIV. Approximately 90% of the converted toxins were excreted into the media. Our study indicates that transgenic Arabidopsis expressing FgTri101 can provide plant protection by detoxifying trichothecenes and excreting the acetylated toxins out of plant cells. Characterization of plant transporters involved in trichothecene efflux will provide novel targets to reduce FHB and mycotoxin contamination in economically important plant crops.


Introduction
Mycotoxins are secondary metabolites produced by fungi and pose serious threats to human and animal health by contaminating food and feed supplies. Fusarium species produce a variety of trichothecenes, including T-2 toxin, 4,15-diacetoxyscirpenol (DAS), nivalenol (NIV), deoxynivalenol (DON), and its acylated derivatives 3-acetyldeoxynivalenol (3-ADON) and 15-acetyldeoxynivalenol . Trichothecenes produced by the fungal pathogen F. graminearum, a major causal agent for Fusarium head blight (FHB), contaminate wheat, barley, and other small grains. Trichothecenes have multiple inhibitory effects on eukaryotic cells, including DNA and RNA synthesis, protein translation and elongation, mitochondria translation, and cell division [1,2]. Consumption of trichothecenecontaminated products causes immunotoxicity and cytotoxicity in humans and animals. In the United States, the Food and Drug Administration has issued advisory levels for DON in cereals and associated products for human consumption (1 ppm) and animal feed (5-10 ppm). Trichothecenes play an important role during F. graminearum and host interactions. DON is a critical factor for FHB spread in wheat. Mutants deficient in DON production are restricted to inoculated florets and cannot pass the rachis node to spread to adjacent spikes [3,4]. On the other hand, transgenic plants expressing genes that detoxify DON increase DON and FHB resistance. For example, transgenic Arabidopsis and wheat plants expressing a barley UDP-glucosyltransferase, which converts DON to the less toxic DON-3-O-glucoside (D3G), exhibit resistance to DON [5,6]. Transgenic wheat plants expressing Fhb7, encoding a glutathione S-transferase, both detoxify trichothecenes and enhance FHB resistance [7]. Thus, the utilization of trichothecene detoxification is a promising approach to reduce FHB and mycotoxin contamination.
The gene cluster and biosynthetic pathways of trichothecenes are well characterized in Fusarium species. A core TRI gene cluster contains 12 TRI genes located on chromosome 2. TRI5, the first gene for trichothecene biosynthesis, and TRI6, the regulatory gene for trichothecene biosynthesis, reside in the TRI gene cluster [8]. For self-protection, Fusarium species employ a transporter (TRI12) to pump toxins out of the fungal cells, and a trichothecene 3-O-acetyltransferase (TRI101) to reduce toxicity of isotrichodermol and other intermediates during the biosynthesis of trichothecenes [9]. Although TRI101 is located outside of the TRI gene cluster, the expression of TRI101 is regulated by TRI6. In addition, Tri101 catalyzes the conversion of toxic Fusarium trichothecenes, such as DON, T-2 toxin, and HT-2 toxin, into less toxic C3-acetylated products [9].
The detoxification ability of TRI101 has been exploited to reduce trichothecenes and FHB, and studies have shown that transgenic plants expressing TRI101 convert DON to 3-ADON and have increased tolerance to trichothecenes such as DAS or DON [10][11][12][13]. However, most of these studies expressed F. sporotrichioides TRI101 (FsTRI101), and transgenic plants expressing FsTRI101 did not significantly increase wheat and barley resistance toward F. graminearum [11,12]. Further studies demonstrated that the proteins encoded by FgTri101 and its ortholog FsTri101 share about 80% identity. FgTri101 is much more effective in converting DON to 3-ADON in comparison to FsTri101 [14]. Partly because of this reason, transgenic wheat plants expressing FsTRI101 only displayed partial resistance to FHB caused by F. graminearum [11]. Therefore, it is necessary to reexamine the potential to reduce FHB and trichothecenes contamination in transgenic plants expressing FgTri101. In addition, three genetically distinct populations of F. graminearum have been characterized as primary causal agents of FHB in North America [15]: NA1, which predominantly produces 15-ADON; NA2, a more aggressive population that produces 3-ADON; and a recently identified population, referred to as NA3, which produces the trichothecene analog, 3α-acetoxy, 7,15-dihydroxy-12,13-epoxytrichothec-9-ene (NX-2). Although the three populations produce different trichothecene phenotypes (chemotypes) in liquid culture, in planta the trichothecenes are primarily found in their deacetylated form. Both 3-ADON and 15-ADON are converted to DON, and NX-2 to 3,7,15-trihydroxy-12,13-epoxytrichothec-9-ene (NX-3). The toxic effects of the NX-3 toxin in inducing cytotoxicity and intestinal inflammatory are comparable to DON [16]. Control of FHB and mycotoxin contamination remains a challenge due to a lack of completely resistant varieties, a complicated fungal infection process, and emerging fungicide-resistant strains. In order to identify a successful strategy to reduce FHB and mycotoxin contamination, it is important to determine whether FgTri101 can effectively detoxify the novel trichothecene NX-3 as well as DON.
Since wheat transformation and regeneration are time-consuming processes, in this study, we generated transgenic Arabidopsis plants expressing FgTri101. The goal of this study was to determine whether the transgenic Arabidopsis plants expressing the FgTri101 gene can increase resistance to several key trichothecenes. We also examined whether FgTri101 transgenic plants can convert DON, 15-ADON, NX-3, and NIV to their acetylated forms.

Generation of Transgenic Arabidopsis Expressing FgTri101
Arabidopsis Col-0 was transformed with the plasmid pBinARS/plus-FgTri101, driven by a double CaMV35S promoter ( Figure S1). A total of 20 T 0 plants were obtained on kanamycin-containing selection media. The presence of the FgTri101 gene was confirmed in nine independent T 0 transgenic Arabidopsis plants by PCR amplification ( Figure S1). The expression of FgTri101 in these transgenic plants was determined by RT-qPCR. The transgenic plant Tri101-12 displayed the highest expression level of FgTri101 (765-fold), followed by Tri101-17 (21-fold) and Tri101-8 (14-fold), relative to the lowest expression plant Tri101-10 that had C t values of 22.8 ± 0.3 ( Figure 1). Four transgenic lines, Tri101-8, 10, 12, and 17, were selected for further study.
gene can increase resistance to several key trichothecenes. We also examined whether FgTri101 transgenic plants can convert DON, 15-ADON, NX-3, and NIV to their acetylated forms.

Generation of Transgenic Arabidopsis Expressing FgTri101
Arabidopsis Col-0 was transformed with the plasmid pBinARS/plus-FgTri101, driven by a double CaMV35S promoter ( Figure S1). A total of 20 T0 plants were obtained on kanamycin-containing selection media. The presence of the FgTri101 gene was confirmed in nine independent T0 transgenic Arabidopsis plants by PCR amplification ( Figure S1). The expression of FgTri101 in these transgenic plants was determined by RT-qPCR. The transgenic plant Tri101-12 displayed the highest expression level of FgTri101 (765-fold), followed by Tri101-17 (21-fold) and Tri101-8 (14-fold), relative to the lowest expression plant Tri101-10 that had Ct values of 22.8 ± 0.3 ( Figure 1). Four transgenic lines, Tri101-8, 10, 12, and 17, were selected for further study. The T2 seeds collected from the T1 transgenic plants (Tri101-8, 10, 12, and 17) were examined for segregation on a kanamycin-containing selection medium. These transgenic lines displayed no segregation for kanamycin resistance and therefore were considered homozygous. Furthermore, we determined the copy number of FgTri101 in three transgenic homozygous lines (TRI101-8-T3, 12-T3, and 17-T3) and nine hemizygous T0 TRI101 plants by quantitative real-time PCR. The Arabidopsis gene 4-hydroxyphenypyruvate dioxygenase (4HPPD) was used as an endogenous reference. The amplification efficiency of the target gene FgTri101 and the Arabidopsis 4HPPD was 91% and 101.9% respectively. Using the ratio formula calculation, one transgenic plant, TRI101-6, appeared to have two insertions using DNA from hemizygous T0 plants (Table S1). We estimated a single FgTri101 insertion was present in the rest of the Arabidopsis transgenic lines (Table S1).

Enhanced Root Growth of Transgenic Arabidopsis Plants Expressing FgTri101 on Media Containing Trichothecenes
Since in planta trichothecenes are primarily found in their deacetylated forms, the germination and growth of transgenic Arabidopsis seeds were assessed on MS media containing DON, NX-3, and NIV, respectively, to evaluate the protective effect provided by The T 2 seeds collected from the T 1 transgenic plants (Tri101-8, 10, 12, and 17) were examined for segregation on a kanamycin-containing selection medium. These transgenic lines displayed no segregation for kanamycin resistance and therefore were considered homozygous. Furthermore, we determined the copy number of FgTri101 in three transgenic homozygous lines (TRI101-8-T 3 , 12-T 3 , and 17-T 3 ) and nine hemizygous T 0 TRI101 plants by quantitative real-time PCR. The Arabidopsis gene 4-hydroxyphenypyruvate dioxygenase (4HPPD) was used as an endogenous reference. The amplification efficiency of the target gene FgTri101 and the Arabidopsis 4HPPD was 91% and 101.9% respectively. Using the ratio formula calculation, one transgenic plant, TRI101-6, appeared to have two insertions using DNA from hemizygous T 0 plants (Table S1). We estimated a single FgTri101 insertion was present in the rest of the Arabidopsis transgenic lines (Table S1).

Enhanced Root Growth of Transgenic Arabidopsis Plants Expressing FgTri101 on Media Containing Trichothecenes
Since in planta trichothecenes are primarily found in their deacetylated forms, the germination and growth of transgenic Arabidopsis seeds were assessed on MS media containing DON, NX-3, and NIV, respectively, to evaluate the protective effect provided by FgTri101. On the media containing 10 mg/L DON or NX-3, the control Arabidopsis Col-0 and GUS transgenic seeds germinated, but displayed small yellow cotyledons, necrotic lesions on leaves, and very short roots. In contrast, the plants expressing FgTri101 exhibited green leaves and long shoots and roots (Figures 2 and S2). Statistical analyses showed that all four FgTri101 transgenic lines displayed a significantly longer root length compared to the Arabidopsis Col-0 and GUS controls (p < 0.001) on DON and NX-3 plates (Figures 3 and S3). However, on the media containing 10 mg/L NIV, FgTri101, Arabidopsis Col-0, and GUS transgenic lines displayed similar phenotypes: green leaves and long roots ( Figure 2 and Figure S2). The roots of Arabidopsis Col-0 and GUS transgenic plants on media containing NIV were significantly longer compared with those on the media containing DON or NX-3 ( Figure S3). These results suggest that expression of FgTri101 did not provide plant protection toward NIV, and that NIV is less phytotoxic. Taken together, our observations indicate that the expression of FgTri101 in the transgenic Arabidopsis plants showed enhanced resistance to DON and NX-3, but not to NIV.
FgTri101. On the media containing 10 mg/L DON or NX-3, the control Arabidopsis Col-0 and GUS transgenic seeds germinated, but displayed small yellow cotyledons, necrotic lesions on leaves, and very short roots. In contrast, the plants expressing FgTri101 exhibited green leaves and long shoots and roots (Figures 2 and S2). Statistical analyses showed that all four FgTri101 transgenic lines displayed a significantly longer root length compared to the Arabidopsis Col-0 and GUS controls (p < 0.001) on DON and NX-3 plates (Figures 3 and S3). However, on the media containing 10 mg/L NIV, FgTri101, Arabidopsis Col-0, and GUS transgenic lines displayed similar phenotypes: green leaves and long roots (Figures 2 and S2). The roots of Arabidopsis Col-0 and GUS transgenic plants on media containing NIV were significantly longer compared with those on the media containing DON or NX-3 ( Figure S3). These results suggest that expression of FgTri101 did not provide plant protection toward NIV, and that NIV is less phytotoxic. Taken together, our observations indicate that the expression of FgTri101 in the transgenic Arabidopsis plants showed enhanced resistance to DON and NX-3, but not to NIV.

Acetylation and Excretion of Trichothecenes by Transgenic Arabidopsis Seedlings Expressing FgTri101
To confirm the acetylation function of FgTri101 in transgenic Arabidopsis, we treated Arabidopsis seedlings with 50 mg/L DON, 15-ADON, NX-3, or NIV in liquid half-strength MS media. After seedlings were treated for 2 days with individual toxins, the

Acetylation and Excretion of Trichothecenes by Transgenic Arabidopsis Seedlings Expressing FgTri101
To confirm the acetylation function of FgTri101 in transgenic Arabidopsis, we treated Arabidopsis seedlings with 50 mg/L DON, 15-ADON, NX-3, or NIV in liquid half-strength MS media. After seedlings were treated for 2 days with individual toxins, the concentrations of trichothecenes were measured in Arabidopsis seedlings and in the media. Our results showed that DON was efficiently converted to 3-ADON in the transgenic seedlings expressing FgTri101, whereas no 3-ADON was detected in wild-type Col-0 and GUS transgenic seedlings (Figure 4a). This observation confirmed that the expression of FgTri101 in transgenic Arabidopsis can convert DON to 3-ADON. Interestingly, we detected 3-ADON in the media growing FgTri101 transgenic seedlings when DON was added to the media ( Figure 4b). Surprisingly, over 96% of 3-ADON was detected in the media out of the total converted 3-ADON ( Figure 4c). These results suggest that Arabidopsis transporter/transporters can efficiently pump out 3-ADON following the FgTri101 detoxification of DON. .

Acetylation and Excretion of Trichothecenes by Transgenic Arabidopsis Seedlings Expressing FgTri101
To confirm the acetylation function of FgTri101 in transgenic Arabidopsis, we treated Arabidopsis seedlings with 50 mg/L DON, 15-ADON, NX-3, or NIV in liquid half-strength MS media. After seedlings were treated for 2 days with individual toxins, the concentrations of trichothecenes were measured in Arabidopsis seedlings and in the media. Our results showed that DON was efficiently converted to 3-ADON in the transgenic seedlings expressing FgTri101, whereas no 3-ADON was detected in wild-type Col-0 and GUS transgenic seedlings (Figure 4a). This observation confirmed that the expression of FgTri101 in transgenic Arabidopsis can convert DON to 3-ADON. Interestingly, we detected 3-ADON in the media growing FgTri101 transgenic seedlings when DON was added to the media (Figure 4b). Surprisingly, over 96% of 3-ADON was detected in the media out of the total converted 3-ADON (Figure 4c). These results suggest that Arabidopsis transporter/transporters can efficiently pump out 3-ADON following the FgTri101 detoxification of DON.   Figures 5a and S4). In contrast, a large amount of 3,15-diADON (98%) was detected in the media when 15-ADON was added to FgTri101 transgenic seedling cultures (Figure 5b). 3,15-diADON was not detected in wild-type Col-0 and GUS transgenic plants or in their growth media. These results indicate that the transgenic Arabidopsis seedlings expressing FgTri101 convert 15-ADON to 3,15-ADON and pump most of it to the media. Similarly, NX-3 was converted to NX-2 in transgenic Arabidopsis seedling expressing FgTri101, but there was no conversion in wild-type Col-0 and GUS seedlings (Figure 6a). NX-2 was excreted to the media (Figure 6b), and about 90% of NX-2 was exported to the media (Figure 6c). Similarly, NX-3 was converted to NX-2 in transgenic Arabidopsis seedling expressing FgTri101, but there was no conversion in wild-type Col-0 and GUS seedlings (Figure 6a). NX-2 was excreted to the media (Figure 6b), and about 90% of NX-2 was exported to the media (Figure 6c).
However, 3-ANIV was not detected in FgTri101 transgenic seedlings. The added NIV was observed in Arabidopsis seedlings expressing GUS or Tri101 ( Figure S5), suggesting that Arabidopsis seedlings can uptake NIV, but have no ability to convert it to 3-ANIV. This explained the growth results that transgenic Arabidopsis expressing FgTri101, wild-type Col-0, or GUS had similar phenotypes on NIV-supplemented media.
Taken together, our results suggest that Arabidopsis expressing FgTri101 can detoxify DON, 15-ADON, and NX-3, and efficiently pump out acetylated trichothecenes to protect the transgenic plants. However, Arabidopsis expressing FgTri101 provides no protection against NIV.

Transgenic Arabidopsis Expressing FgTri101 Accumulated Higher Weight with Addition of Trichothecenes
To examine the protective effect of transgenic Arabidopsis expressing FgTri101, the total weights of Arabidopsis seedlings were determined after being treated with 50 mg/L DON, 15-ADON, NX-3, or NIV in liquid MS media. After treatments of 48 h with DON, 15-ADON, or NX-3, the total weights of the FgTri101 transgenic lines were significantly higher compared to the GUS transgenic controls (Figure 7a-c). However, no significant difference was observed between the NIV-treated GUS and FgTri101 seedlings (Figure 7d). These results suggest that FgTri101 can provide plant protection from trichothecenes DON and NX-3, but not from NIV produced by F. graminearum. However, 3-ANIV was not detected in FgTri101 transgenic seedlings. The added NIV was observed in Arabidopsis seedlings expressing GUS or Tri101 ( Figure S5), suggesting that Arabidopsis seedlings can uptake NIV, but have no ability to convert it to 3-ANIV. This explained the growth results that transgenic Arabidopsis expressing FgTri101, wild-type Col-0, or GUS had similar phenotypes on NIV-supplemented media.
Taken together, our results suggest that Arabidopsis expressing FgTri101 can detoxify DON, 15-ADON, and NX-3, and efficiently pump out acetylated trichothecenes to protect the transgenic plants. However, Arabidopsis expressing FgTri101 provides no protection against NIV.

Transgenic Arabidopsis Expressing FgTri101 Accumulated Higher Weight with Addition of Trichothecenes
To examine the protective effect of transgenic Arabidopsis expressing FgTri101, the total weights of Arabidopsis seedlings were determined after being treated with 50 mg/L DON, 15-ADON, NX-3, or NIV in liquid MS media. After treatments of 48 h with DON, 15-ADON, or NX-3, the total weights of the FgTri101 transgenic lines were significantly higher compared to the GUS transgenic controls (Figure 7a-c). However, no significant difference was observed between the NIV-treated GUS and FgTri101 seedlings ( Figure  7d). These results suggest that FgTri101 can provide plant protection from trichothecenes DON and NX-3, but not from NIV produced by F. graminearum.

Morphology of Transgenic Arabidopsis Plants Expressing FgTri101
To assess whether the transgenic expression of FgTri101 affects Arabidopsis growth and morphology, FgTri101 transgenic lines (Tri-8, 12, and 17) and the GUS transgenic control were germinated on MS media containing kanamycin and transferred to soil. The flowering time, rosette leaf number, plant height, and number of shoots were evaluated. All three FgTri101 expression lines did not show a significantly different flowering time, plant height, and number of shoots compared with the GUS control (Table 1). However,

Morphology of Transgenic Arabidopsis Plants Expressing FgTri101
To assess whether the transgenic expression of FgTri101 affects Arabidopsis growth and morphology, FgTri101 transgenic lines (Tri-8, 12, and 17) and the GUS transgenic control were germinated on MS media containing kanamycin and transferred to soil. The flowering time, rosette leaf number, plant height, and number of shoots were evaluated. All three FgTri101 expression lines did not show a significantly different flowering time, plant height, and number of shoots compared with the GUS control (Table 1). However, the FgTri101 line 12, which had the highest expression level of FgTri101, displayed significantly fewer rosette leaf numbers compared with the GUS control. Nevertheless, the expression of the FgTri101 gene without extremely high levels did not significantly affect the overall Arabidopsis morphology.

Discussion
Fusarium species employ TRI101 encoding trichothecene 3-O-acetyltransferase for self-protection from trichothecene mycotoxins [9]. In this study, we generated transgenic Arabidopsis plants expressing FgTri101 and demonstrated that these transgenic plants had enhanced resistance to several trichothecenes compared to Arabidopsis Col-0 and GUS transgenic plants. We showed that FgTri101 transgenic plants converted DON, 15-ADON, and NX-2 to their acetylated forms. Interestingly, we observed that 3-ADON, 3,15-diADON, and NX-3 were efficiently excreted into the media from FgTri101 transgenic plants. This study demonstrated that the overexpression of FgTri101 provides plant protection from important mycotoxins by detoxification and excretion. Further characterization of the transporters involved in toxin efflux will provide novel targets for generating transgenic crops to enhance FHB resistance and reduce mycotoxin contamination.
Gene expression analysis revealed varying levels of FgTri101 expression in the different transgenic Arabidopsis plants. Among four selected transgenic lines, the gene expression of FgTri101 was 700-fold higher in the line Tri101-12 than Tri101-10. Although the phenotypes of the four FgTri101 lines in root growth assays using 10 mg/L individual toxins were not significantly different, Tri101-10, which had the lowest FgTri101 expression, showed relatively low toxin conversion rates compared to Tri101-8, 12, and 17. A good correlation was demonstrated between the levels of lactoferrin protein produced in transgenic wheat lines and the level of resistance against F. graminearum [17]. Further investigations are needed to generate FgTri101 antibodies and use Western blots to quantify the levels of the FgTri101 protein.
The transgenic Arabidopsis plants expressing FgTri101 displayed enhanced resistance to trichothecenes, such as DON and NX-3, with longer roots and higher growth weights. Trichothecenes are phytotoxic and affect plant morphology, such as causing inhibition of root elongation and dwarfism [18]. Plant root growth inhibition assays are commonly used to evaluate the direct phytotoxic effects of trichothecenes. DON and other structural analogs have been shown to inhibit plant growth [19,20]. Some studies showed that 3-ADON is highly toxic to wheat [21]. However, transgenic rice expressing FgTri101 displayed enhanced DON tolerance [13]. We showed that expression of FgTri101 not only increased Arabidopsis tolerance to DON, but also enhanced plant resistance toward NX-3 that is produced by a newly identified F. graminearum population (NA3). More importantly, we discovered that the acetylated trichothecenes were pumped out of plant cells, which may greatly reduce their phytotoxic effects. It will be interesting to determine whether the high toxicity caused by 3-ADON in wheat is due to a lack of transporters for 3-ADON excretion.
Our study demonstrated that approximately 89-98% of converted 3-ADON, 3,15-diADON, or NX-2 by Tri101 transgenic Arabidopsis was excreted into the media. Since 3-ADON and 3,15-di ADON have been shown to be phytotoxic to plants [20], the enhanced root growth of FgTri101 transgenic plants may be due to most of the acylated trichothecenes being exported out of plant cells. Plant ATP-binding cassette (ABC) transporters are involved in toxin transportation and detoxification. For self-protection, many toxic metabolites are conjugated and transported to the plant vacuole for detoxification and sequestration [22]. Studies demonstrated that wheat plants contain two vacuolar membrane-localized transporters, TaABCC3.1 and TaABCC3.2, which are induced by DON in wheat and associated with DON tolerance [23]. Although lacking direct evidence of TaABCC3.1 in DON transport and detoxification, the TaABCC3.1 transporter has been demonstrated to contribute to FHB resistance in wheat [23]. Arabidopsis AtABCC3, the closest homologue of TaABCC3 proteins, has been shown to localize within the vacuolar membrane (tonoplast) and transport glutathione conjugates and chlorophyll catabolites [24,25]. It is possible that AtABCC3 functions in DON sequestration. Most of the ABC transporters have been characterized to localize in the vacuolar membranes and sequester toxins. For example, PDR5 from Saccharomyces cerevisiae is a multidrug transporter [26] and TRI12 from F. sporotrichioides is a trichothecenes efflux pump for fungus self-protection [27]. In the Arabidopsis genome, from a total of 130 annotated ABC transporters, about 20 of them have been functionally characterized. In addition to sequestering toxins in vacuoles, there are many ABC transporters involved in detoxification and transportation of various compounds [28]. A plant plasma membrane ABC transporter, NpABC1, has been identified and is involved in the secretion of an antifungal terpenoid that plays a role in plant defense [29]. To the best of our knowledge, this is the first report showing trichothecenes excreted out of plants. Further investigations are needed to determine whether wheat and barley plants expressing FgTri101 have the ability to excrete acetylated toxins. We plan to identify the transporter/transporters that are involved in trichothecene excretion from plant cells, which may provide a novel target to reduce FHB and mycotoxin contamination.
FgTri101 and GUS transgenic Arabidopsis displayed similar healthy root length and growth weight when NIV was added, whereas GUS transgenic plants were significantly inhibited when DON or NX-3 was added. These observations suggest NIV is less phytotoxic to plants. Similarly, our recent study showed that DON, 3-ADON, and 15-ADON, but not NIV, inhibit wheat root elongation [20]. We did not observe NIV acetylation in transgenic Arabidopsis plants expressing FgTri101. F. graminearum strains produce NIV and acetylated NIV (typically 4,15-diANIV and 4-ANIV) in cultures [30]. To determine if Tri101 proteins are conserved between DON-producing strains and NIV-producing strains, we compared FgTri101 from PH-1 (used in this study for generation of Arabidopsis transgenic plants) and Tri101 from selected NIV strains. These Tri101 proteins share high identity and the catalytic histidine (FgH156 and 160D). We did observe that two amino acids were substituted in selected NIV strains, including A215P and L392F ( Figure S6). The motif DFGWG(FgD389-G393) is conserved in Fusarium species. However, the fourth position (Fg392) was observed as a Phe (F) in FsTRi101, while it is a Leu (L) in FgTri101 [14]. In comparison, we observed that 392F was present in all NIV-procuring Fusarium species ( Figure S6). This suggests that the DFGWG motif may be important for substrate specificity. Further study is needed to determine if 392F is important for FgTri101 to bind NIV.
Regarding the effects of 3-ADON on the growth of cereal plants, there are conflicting results. Several studies show that 3-ADON is highly toxic to wheat at low concentrations [21]. On the other hand, studies find that 3-ADON does not inhibit rice root growth [13]. Ohsato and associates suggested that the higher toxicity of 3-ADON observed in cereal plants may be due to the instability of the C3 acetyl form in plant cells. In our study, we observed most of the 3-ADON was excreted out from plant cells, which greatly reduced 3-ADON phytotoxicity. More interestingly, while we only detected a trace amount of 3,15-diADON in a few transgenic Arabidopsis samples, we observed significant amounts of 3,15-diADON in the media. This suggests that either 3,15-diADON is unstable inside of plant cells, or it is quickly transported out of plant cells. We speculate that transgenic plants expressing FgTri101 may constantly export C-3 acetylated trichothecenes out of plant cells and protect crops from the phytotoxic effect of trichothecenes. Transgenic wheat and barley plants expressing FsTRI101 converted DON to an acetylated form, 3-ADON, but did not display effective FHB resistance [11,12]. This may be because FsTRI101 is less efficient in the conversion of DON to 3-ADON than FgTri101 [14]. It is also possible that wheat and barley lack transporters that excrete acetylated trichothecenes. Further investigations are needed to determine whether gene stacking of FgTri101 and the transporters for acylated toxin excretion provide effective protection for wheat and barley toward FHB and mycotoxin contamination.
Acetylation at the C-3 position has been shown to significantly reduce the phytotoxicity of a number of trichothecenes [31,32]. Therefore, expressing fungal Tri101 transacetylase could improve plant resistance to the toxins. This approach may work with other mycotoxins, but it would require different detoxification genes.
In conclusion, we demonstrated that transgenic Arabidopsis plants expressing FgTri101 displayed resistance to trichothecenes and converted several important trichothecenes to their less toxic acylated forms. More importantly, we showed that the acylated trichothecenes were excreted into the media from transgenic Arabidopsis plants expressing FgTri101. Further identification of the transporters for trichothecene excretion will provide novel breeding targets to reduce FHB and mycotoxin contamination.

Construction of FgTri101 Vector and Arabidopsis Transformation
Arabidopsis thaliana (ecotype Columbia, Col-0) were grown in a growth chamber at 23/20 • C with a 16/8 h light/dark cycle. The pBinARS/plus vector was used for FgTri101 vector construction [33]. FgTri101 contains no intron and was amplified using F. graminearum genomic DNA with primers FgTri101-ORF5 and ORF3 , including the two enzyme cut sites Spe I and Xba I (Table S2). The PCR products of FgTri101 were purified, digested, and ligated into pBinARS/plus treated with the same enzymes. The ligation mixture was used to transform TOP10 competent cells (Invitrogen, Carlsbad, CA, USA). The positive clone containing FgTri101 was obtained and sequenced. The purified plasmids of pBinARS/plus-FgTri101 driven by a double 35S (D35S) promoter were introduced into Agrobacterium tumefaciens strain GV2260 for Arabidopsis transformation.
Arabidopsis transformation was performed with pBinARS/plus-FgTri101 using floral dip methods [34]. Similarly, the pBinARS/plus vector carrying β-glucuronidase (GUS) was used to generate transgenic Arabidopsis plants as controls. Transgenic plants were selected on Murashige and Skoog (MS) media with the addition of 50 mg/L kanamycin. After about two weeks, the kanamycin-resistant plants were transferred to soil and grown in the growth chamber as described above.

Molecular Analysis of Transgenic Arabidopsis
Plant genomic DNA was isolated from leaves of kanamycin-resistant plants using ZR Fungal/Bacterial DNA miniprep kit (Zymo Research, Boston, MA, USA). DNA concentration and quality were evaluated using a spectrophotometer (NanoDrop 2000, Thermofisher Scientific, Waltham, MA, USA). PCR was used to amplify the target gene using the primers D35S-F and Nos-R (Table S2).
Total RNA was extracted from leaves of transgenic Arabidopsis plant containing FgTri101 using a Trizol reagent (Sigma-Aldrich, St. Louis, MO, USA) combined with the Ambion RNA isolation kit [35]. RNA was quantified with a spectrophotometer (Nanodrop; Thermo Fisher Scientific, Waltham, MA, USA) and treated with RQ1 RNase-free DNase (Promega Corp. Madison, WI, USA). The absence of genomic DNA contamination was verified by quantitative PCR (qPCR). The first-strand cDNA was synthesized, and reverse transcriptase qPCR (RT-qPCR) was performed on a Bio-Rad CFX96 real-time system (Bio-Rad Laboratories, Hercules, CA, USA) using primers FgTri101-RT-5' and FgTri101-RT-3' (Table S2). The Arabidopsis gene elongation factor 1-alpha (EF1α) was used to normalize the values as an internal control. The gene expression level was calculated with the 2 −∆∆Ct method using CFX manager software (Bio-Rad), relative to the transgenic plant with the lowest expression level (Tri101-10), which was set as 1. The qPCR reactions were set up in triplicate and repeated three times with similar results.
To obtain homozygotes for further analysis, T2 seeds from transgenic lines were sown on MS media containing 50 mg/L kanamycin for a progeny test. The transgenic lines for which T2 seedlings showed no segregation were considered as homozygotes. The FgTri101 copy number was estimated by qPCR in homozygous T2 lines Tri101-8, 12, 17 and all T 0 transgenic lines as described [35,36]. The Arabidopsis gene 4-hydroxyphenypyruvate dioxygenase (4HPPD), which is a single-copy gene, was used as the endogenous reference gene [36]. Primer efficiency for FgTri101 and 4HPPD was determined using a standard curve consisting of a dilution series for each primer pair. The qPCR amplification efficiency was calculated according to the following equation: Efficiency = 10 (−1/slope) − 1. The ratio of the copy number of TRI101 was calculated using the following equation: Ratio = (1 + Efficiency (Ct FgTri101 ))/(1 + Efficiency (Ct 4HPPD )). The qPCR reactions were set up in three technical triplicates and repeated three times. Data are shown as means ± SD of three replicates.

Root Growth Assays
Although F. graminearum produces the acetylated trichothecenes in culture, e.g., 3-ADON, 15-ADON, and NX-2, most of these toxins are converted to their deacetylated forms, DON, NIV, and NX-3 in planta. Therefore, root growth assays were conducted on a MS medium with the addition of DON, NIV, and NX-3 in square vertical plates (Greiner bio-one North America Inc. Monroe, NC). T 2 seeds of the transgenic plants expressing GUS or FgTri101 (lines 8, 10, 12, and 17) were surface sterilized using 70% ethanol for 2 min, then 50% Clorox for 10 min, and rinsed three times with sterile water. The seeds were sown on MS medium containing 10 mg/L DON, NIV, or NX-3. A total of nine surface-sterilized seeds were placed on each agar plate. Four plates from each transgenic line were treated with each toxin. The plates were placed vertically in a growth chamber as described above. The root length was measured every 2 days for 12 days after germination. Root length (n = 36-40) from each transgenic line was analyzed and compared with GUS controls using JMP.

Conversion of Trichothecenes by Transgenic Arabidopsis Seedlings
T 2 seeds of the transgenic plants expressing GUS and FgTri101 (lines 8, 10, 12, and 17) were surface sterilized as described above. The sterilized seeds were placed on a MS medium containing 50 mg/L kanamycin. The plates were kept at 4 • C for 2 days and moved to a growth chamber at 23/20 • C in a 16/8 h light/dark cycle. After 7-8 days, about 16 plants were placed into 50 mL tubes to grow in 5 mL of half-strength MS liquid medium containing 1% sucrose. After 3-4 days, the media were removed and substituted with 5 mL of fresh medium plus 50 mg/L DON, 15-ADON, NIV, or NX-3, respectively. The samples were collected two days after toxin treatments. The seedlings were rinsed three times with 5 mL of water to remove external toxins. The rinsed water was combined with the media for toxin measurement. The fresh weigh of seedlings was measured. The seedling samples and culture media were used for toxin analysis. Each transgenic line was treated with three biological replicates for each toxin. Media containing toxins without plants served as controls.

Mycotoxin Extraction and Quantification
Arabidopsis seedlings were extracted with 10 mL of acetonitrile:water (86:14) by shaking for 1 h at 250 rpm on an orbital shaker. The samples were centrifuged for 5 min at 4000 rpm, and 9 mL of extract was dried under a stream of nitrogen [37]. For liquid media, the toxin was extracted from 1.4 mL aliquots from 20 mL that were combined with 8.6 mL acetonitrile, and the mixture was dried under a stream of nitrogen.
The total excreted toxin was calculated by the following formula: 3-acetylated toxin × 20/1.4, and divided by the total acetylated toxin in the plant and media.

Statistical Analyses
All statistical analyses were performed using JMP software. Individual analyses are discussed in each Section 4.

Supplementary Materials:
The following are available online at https://www.mdpi.com/article/ 10.3390/toxins13050320/s1, Figure S1: (a) Diagram of FgTri101 construct in the binary vector pBi-nARS/PLUS. (b) PCR amplification of FgTri101 gene from transgenic Arabidopsis plants. Figure S2 Arabidopsis seedling growth on MS media containing different trichothecenes (10 mg/L). Figure  S3 Root length comparison after a two-week incubation on MS media containing toxins. Figure S4: Chromatographs show 15-ADON converted to 3,15-ADON by FgTri101 transgenic seedlings, but not in GUS controls. Figure S5: Chromatographs show no detectable acetylated NIV by FgTri101 transgenic Arabidopsis. Figure S6: Comparison of deduced Tri101 amino acid sequences from DON producing strain PH-1 and NIV producing strains from Fusarium family. Table S1: The copy number of FgTri101 in transgenic Arabidopsis. Table S2: Primers used in the study.