Overexpression of Nepenthesin HvNEP-1 in Barley Endosperm Reduces Fusarium Head Blight and Mycotoxin Accumulation

Fusarium head blight (FHB) causes substantial losses of yield and quality in grains, both in the field and in post-harvest storage. To date, adequate natural genetic resistance is not available for the control of FHB. This study reports the cloning and overexpression of a barley (Hordeum vulgare L.) antifungal gene, nepenthesin 1 (HvNEP-1), in the endosperm of barley grains. Transgenic barley lines overexpressing HvNEP-1 substantially reduced FHB severity and disease progression after inoculation with Fusarium graminearum or Fusarium culmorum. The transgenic barley also showed reduced accumulation of the mycotoxin deoxynivalenol (DON) in grain, far below the minimum value allowable for food. Semi-field evaluation of four HvNEP-1 transgenic lines revealed substantial reduction of FHB severity and progression as compared with the control H. vulgare cultivar Golden promise (GP) plants. Our study demonstrated the utility of HvNEP-1 for the control of FHB in barley, and possibly other grains such as wheat and maize.


Introduction
Fungal pathogens cause considerable loss of yield and quality in economically important crops. Fusarium head blight (FHB) or scab is one of the major fungal diseases of the Triticeae family in temperate and warm, humid regions of the world [1]. FHB is caused by several Fusarium species, but Fusarium graminearum and Fusarium culmorum are seen as two of the most economically relevant species [1][2][3]. In the 1990s, an FHB epidemic caused an estimated loss of 2.7 billion USD in the US alone [4]. In 2010, a report on the occurrence of FHB in the US indicated that the disease was on the decline in most states. However, in parts of Ohio, the FHB incidence reached up to 60% [5]. In the same year, the overall impact of the disease in the state of Kansas was valued at 13 million USD. More recently, the state of Kansas experienced an estimated 2.1% yield loss, due to FHB in 2019 [6]. Overall, FHB could lead to >1 billion USD economic losses per year [7].

Plasmid Construction for Plant Transformation
For plant transformation, the sequence for the signal peptide lacking HvNEP-1 was PCR amplified from gDNA using primer pairs P1 (see primer information in Table S1). PCR was carried out in a 50 µL reaction mixture containing 100 ng gDNA template (2 µL), 10 µl 5× Herculase II Buffer, 10 µL 2 mM dNTP mixture, 4 µL primer mixture P1 (10 pmol/µL), 0.5 µL Herculase II DNA polymerase (2 U/µL), 3 µL DMSO, and 20.5 µL MQ. The PCR conditions were a denaturation step at 96 • C for 2 min, then, 40 cycles of 96 • C for 1 min, 60 • C for 20 sec, 72 • C for 2.30 min, and a final extension step at 72 • C for 2.30 min. Amplicons were detected by electrophoresis in 1% agarose gels stained with ethidium bromide. An endoplasmic reticulum (ER) sorting sequence (SE) KDEL was introduced at the sequence for the C-terminal end of HvNEP-1. The barley HordD signal peptide (HordDsp) was used in place of the HvNEP-1 native signal peptide. The HordD promoter (HordDp) and signal peptide were PCR amplified with primer pairs P2 under the following conditions: initial denaturation at 96 • C for 2 min, then 40 cycles of 96 • C for 1 min, 60 • C for 20 sec, 72 • C for 1 min, and a final extension step at 72 • C for 2.30 min, and a 4 • C hold. The amplicon was fused upstream of the ∆HvNEP-1 and inserted into the barley transformation vector pWBVec8 by In-fusion cloning [38]. In the resulting construct, pWBVec8: HordDp: HordDsp: ∆HvNEP-1: NOS, HvNEP-1 was fused downstream of the HordD promoter and upstream of the Agrobacterium tumefaciens-derived NOS terminator.

Generation of Transgenic Barley Plants
The construct was used to transform competent Agrobacterium strain AGL0 using the method already described by An et al. [39]. The bacteria were grown on LB plates containing 100 µg/mL spectinomycin and 25 µg/mL rifampicin for 72 h at 28 • C. Positive colonies were identified by PCR using gene-specific primers. A positive clone was inoculated into 5 ml MG/L medium (5 g/L mannitol, 1 g/L L-glutamic acid, 0.25 g/L KH 2 PO 4 , 0.1 g/L NaCl, 0.1 g/L MgSO 4 ·7H 2 O, 1ng/L biotin, 5 g/L tryptone, and 2.5 g/L yeast extract) [40], containing 100 µg/mL spectinomycin and 25 µg/mL rifampicin and grown by shaking (150 rpm, 28 • C, 48 h). Sterile 15% glycerol stocks were prepared and stored at −80 • C. For barley transformation, 200 µL of the glycerol stock was used to inoculate 5 ml MG/L medium and grown overnight by shaking (180 rpm, 28 • C). The culture was used for transformation of immature barley embryos isolated from barley cultivar Golden promise (GP) following the procedure described by Holme et al. [41] and Bartlett et al. [42].

PCR Analysis of Transformed Barley Plants
After transformation, selection, and regeneration of T0 plants, the gDNA was isolated from young leaves using the procedure described in [43]. PCR was performed to confirm the presence of the hygromycin resistance gene hpt using primer pair P3. Positive transformants were further confirmed with PCR using primer pair P4 designed inside the HordDp and HvNEP-1. The resulting PCR fragment size was 759 bp. Untransformed GP plants used as the negative control were analyzed similarly using both primer pairs. Eight weeks after germination (Zadoks growth stage 60 [45] ), spikes of 10 T 0 plants were spray inoculated with a spore suspension (1 × 10 5 spores per ml in water containing 0.04% tween 20) ( Figure S1a). Briefly, tillers of the T 0 plants were subdivided into three sections. Two of the sections were inoculated with the two Fusarium isolates and the remaining section was sprayed with Milli-Q water (MQ water) (Milli-Q®Direct Water Purification System, Merck, Germany) as a control. Two untransformed GP plants at the same stage of development were treated similarly with Fusarium spores and MQ water. Subsequently, the plants were covered with plastic bags for 3 days and transferred into the greenhouse (18 to 21 • C and relative humidity 70% to 75%).

Disease Scoring and Mycotoxin Analysis
Spikes of the individual transgenic lines and GP control plants were scored for disease symptoms (FHB severity) at 1, 2, and 3 weeks after inoculation. Percent disease severity was scored as the percentage of infected seeds in a spike. Since the transgene segregated in the T 0 plants, the first 3 matured spikes from each transgenic and GP control plants were selected and scored for disease severity. Grains from the selected spikes were used for analysis of mycotoxins and HvNEP-1 expression. Mycotoxin accumulation (DON, NIV, ZEA, T-2, and HT-2) in the seeds of the transgenic lines and control GP plants was quantified in an LC-MS/MS system consisting of an Agilent 1260 infinity chromatographic system (Santa Clara, CA, USA) coupled to an AB Sciex 3200 triple-quadrupole trap mass spectrometer (QTRAP/MS) (AB Sciex, Framingham, USA) according to Etzerodt et al. [46]. Analyst 1.6.3 software (AB Sciex, Framingham, USA) was used to control the LC-MS/MS system. The same numbers of seeds per treatment were used for the analysis of mycotoxins and HvNEP-1 expression. In the Fusarium-treated transgenic and GP control plants, seeds were collected and analyzed separately for F. graminearum (FG) and F. culmorum (FC). For each treatment, 3 seeds per spike (9 seeds per treatment) were collected in liquid nitrogen and dissected longitudinally into equal parts. The seeds were pooled by treatment and one half of each pooled sample was freeze dried, ground, and used for mycotoxin analysis; the other pooled half was used for analysis of HvNEP-1 expression. Similarly, seeds from the non-inoculated control and the GP control plants were collected and analyzed for mycotoxins.

HvNEP-1 Expression in the Transgenic Barley Lines
Total RNA was extracted from the pooled seeds using the method described by Chomczynski and Sacchi [47]. RNA samples were treated with DNase according to the manufacturer's recommendation (Roche). Reverse transcription of mRNA was performed using Superscript III-RT (Invitrogen) and oligo (dT) 21T-anchor-containing primers. The qPCR was performed in a final volume of 12 µL containing 6 µL Power SYBR Green master mix (Applied Biosystems), 1 µL diluted cDNA, 2.4 µL of 1.5 µM primer mix, and 2.6 µL sterile Milli Q water. The samples were set up in 384-well plates and the qPCR was run and detected in an AB7900HT sequence detection system (Applied Biosystems, USA). Gene-specific qPCR primers were designed according to the available HvNEP-1 sequence. Expression analysis of the seeds from Fusarium (FG, FC) and water (MQ) treated spikes were conducted separately for the transgenic and control plants. The level of Splicing factor 2 (SP2) gene expression was used to normalize HvNEP-1 expression. The qPCR primers are described in the Table S1.

Field Soil Plot Analysis of Transgenic Barley Lines for Fusarium Resistance
The segregating T 1 generation from 4 selected lines (NEP02-01, NEP16, NEP18, and NEP20) and the GP-1 control plants were grown on 5 plots (0.6 m × 2.2 m) in a controlled semi-field facility at Aarhus University (Flakkebjerg, 55.321547 • N) ( Figure S1b). Transgene-positive plants from the segregating T 1 generation were selected by PCR using hpt primers. The transgenic and non-transgenic control plants (Zadoks stage 60) were spray-inoculated with F. graminearum 7775 spores and covered with nylon mesh for 3 days. Ten transgene-positive plants from the transgenic lines and 53 GP-1 control plants were used to study disease infection rate and progression. The plants were assessed for disease severity on a scale of 1 to 9. In addition, the disease progression (AUDPC) was also examined at 1, 2, and 3 weeks after inoculation.

HvNEP-1-Overexpressing Barley Lines and Assessment of FHB Resistance
The overexpression construct was designed to express HvNEP-1 specifically in the endosperm of barley grains (Figure 1a). Immature barley embryos were transformed with the construct using the Agrobacterium strain AGL0. A total of 525 immature embryos were infected with Agrobacterium and 82 T 0 plants were regenerated. Among the 82 T 0 plants, 24 plants were selected and tested for the insertion of HvNEP-1 by PCR. All tested plants were positive for the transgene. Plants regenerated from 20 out of the 24 lines were transferred into soil and grown in the greenhouse. Ten of the transgenic lines were randomly selected for downstream analysis (Figure 1b,c). The lines analyzed by qPCR for seed HvNEP-1 expression showed varying expression of the transgene between the lines. The highest relative expression was seen in line NEP20 (0.4166 ∆Ct) and the lowest was detected in line NEP20-02 (0.0114 ∆Ct) as compared with the reference gene SP2 (Figure 1d).  The transgenic lines were assessed for FHB resistance and mycotoxin accumulation at the 85 to 87 growth stage (in the Zadoks scale, [45]). Disease severity of the 10 T 0 lines was compared to two untransformed control GP plants. FHB severity was scored as the percentage of infected seeds in a spike after one, two, and three weeks after inoculation. The transgenic lines showed a substantial reduction in FHB severity as compared with the control plants. The mean percent of infection in the transgenic plants varied from 3.41% to 23.08%, whereas the mean percent of infection in the control GP plants ranged from 31.88% to 50% for both F. graminearum and F. culmorum strains (Figure 2a). The progression of FHB in the spikes of the transgenic lines and control GP plants was assessed for the first three weeks after inoculation, and the area under the disease progress curve (AUDPC) was calculated (Figure 2b). The mean AUDPC of FHB progress was higher in the control GP plants than that in the transgenic lines.

Semi-Field Analysis of the HvNEP-1 Transgenic Lines
To test the reliability of HvNEP-1-mediated resistance to FHB, four Fusarium-infected transgenic lines and control GP plants were grown in a field plot covered with an insect-proof net ( Figure S1b). Ten transgene-positive plants from each transgenic line were compared to 53 untransformed GP-1 plants. Line NEP20 showed the lowest disease infection rate and progression as compared with the

Mycotoxin Analysis of the HvNEP-1-Overexpressing Lines
The mycotoxins levels were analyzed in the grains of Fusarium-infected and water-treated transgenic (T 0 ) and non-transformed GP plants three weeks after inoculation (Table 1). Fusarium-infected seeds were sorted into seeds with FHB symptoms (FC/FG, FHB + ) and seeds with no FHB symptoms (FC/FG, FHB -). In FHB-seeds from transgenic plants, mycotoxin levels were below detection level in all samples, except from some NIV accumulation in NEP02-01 (FC, 77 µg/kg DM and FG, 124 µg/kg DM) and NEP06 (FG, 171 µg/kg DM). Water-treated control plants showed no sign of infections and mycotoxin accumulation. Similarly, water-treated transgenic plants did not develop infections, although a minor NIV level (163 µg/kg DM) was seen in line NEP16. In the FHB + non-transgenic plants, both F. culmorum and F. graminearum caused a strong accumulation of DON, and also a strong accumulation of NIV (except for GP-2) and ZEA (except for GP-1). The highest levels of DON were seen in F. culmorum-infected (FC, FHB + ) GP-1 control seeds. In the transgenic plants infected with F. culmorum (FC, FHB + ), DON levels were substantially reduced as compared with the F. culmorum-infected (FC, FHB + ) non-transgenic GP plants. For nine out of the 10 transgenic plants, DON levels were below the detection level. For the NEP20 line, in which DON levels were detected, we measured 152 µg DON/kg DW, which was still far below the DON levels in the controls. For the F. graminearum-infected (FG, FHB + ) seeds of transgenic plants, DON levels were below the detection level in six out of the 10 lines. The NEP04 line had a DON content just above the detection level, but NEP19 and NEP20 had high levels of DON (8170 and 40,073 µg/kg DW, respectively). ZEA levels were also strongly reduced in the transgenic lines. Only line NEP20 had a detectable ZEA level (256 µg/kg DW). In contrast, high levels of NIV were still seen in the transgenic lines. For some lines (NEP02-01, NEP06, and NEP20 for F. culmorum-infected (FC, FHB + ) seeds and NEP02-01, NEP05, and NEP06 for F. graminearum-infected (FG, FHB + ) seeds), NIV levels were higher than in the untransformed GP plants. In the non-transgenic FHBplants, higher levels of DON were detected in both F. culmorum and F. graminearum infected plants as compared with transgenic FHBplants.

Semi-Field Analysis of the HvNEP-1 Transgenic Lines
To test the reliability of HvNEP-1-mediated resistance to FHB, four Fusarium-infected transgenic lines and control GP plants were grown in a field plot covered with an insect-proof net ( Figure S1b). Ten transgene-positive plants from each transgenic line were compared to 53 untransformed GP-1 plants. Line NEP20 showed the lowest disease infection rate and progression as compared with the other transgenic lines (NEP02-01, NEP16 and NEP18) and the untransformed control plant GP-1 (Figure 3a). Three weeks after infection, line NEP20 had an average infection score of 1.9, whereas the control GP plants were scored at 4.15 (on a one to nine scale) (Figure 3b). AUDPC was calculated at one, two, and three weeks after inoculation. After three weeks, the AUDPC was significantly higher in the GP control plants (35%) than in the transgenic NEP20 plants (24.5%) (Figure 3c).

Discussion
Fusarium head blight is a devastating disease of barley and other crops worldwide. Resistant cultivars could reduce damage from FHB. However, so far, only limited resources of resistance genes are available. In the current study, the antifungal potential of barley HvNEP-1 was investigated by transgenic overexpression in the developing barley endosperm. The ten investigated T0 barley plants had variable levels of HvNEP-1 expression, but no noticeable phenotypic differences were observed between the transgenic and control GP plants, even in the highest expressing lines (Figure 1b). Studies of Fusarium-infected plants showed that overexpression of HvNEP-1 in transgenic barley improved resistance to FHB both by limiting primary infection (Type-I resistance) and the spread of the disease (AUDPC) (Type-II resistance) as compared with the untransformed control plants [48]. Limiting infection and spread of disease is important, particularly, from a yield perspective. However, accumulation of mycotoxins is another unfortunate effect of Fusarium infections. The current study demonstrated that, endosperm-specific overexpression of HvNEP-1 reduced the accumulation of mycotoxins in the infected grains significantly (Type-IV resistance) [49]. In contrast to the nontransgenic control plants infected with F. culmorum, in which the level of DON was high, the accumulation of DON in the infected transgenic grains was very low (below the detection level in nine out of the ten lines). This was in agreement with our previous in vitro experiments, where the accumulation of 15-ADON was reduced in cultures with rHvNEP-1 [36]. For the F. graminearuminfected plants, seven transgenic lines had DON levels below the detection level. The lower mycotoxin formation in the transgenic seeds was accompanied by decreased disease progression in

Discussion
Fusarium head blight is a devastating disease of barley and other crops worldwide. Resistant cultivars could reduce damage from FHB. However, so far, only limited resources of resistance genes are available. In the current study, the antifungal potential of barley HvNEP-1 was investigated by transgenic overexpression in the developing barley endosperm. The ten investigated T 0 barley plants had variable levels of HvNEP-1 expression, but no noticeable phenotypic differences were observed between the transgenic and control GP plants, even in the highest expressing lines (Figure 1b). Studies of Fusarium-infected plants showed that overexpression of HvNEP-1 in transgenic barley improved resistance to FHB both by limiting primary infection (Type-I resistance) and the spread of the disease (AUDPC) (Type-II resistance) as compared with the untransformed control plants [48]. Limiting infection and spread of disease is important, particularly, from a yield perspective. However, accumulation of mycotoxins is another unfortunate effect of Fusarium infections. The current study demonstrated that, endosperm-specific overexpression of HvNEP-1 reduced the accumulation of mycotoxins in the infected grains significantly (Type-IV resistance) [49]. In contrast to the non-transgenic control plants infected with F. culmorum, in which the level of DON was high, the accumulation of DON in the infected transgenic grains was very low (below the detection level in nine out of the ten lines). This was in agreement with our previous in vitro experiments, where the accumulation of 15-ADON was reduced in cultures with rHvNEP-1 [36]. For the F. graminearum-infected plants, seven transgenic lines had DON levels below the detection level. The lower mycotoxin formation in the transgenic seeds was accompanied by decreased disease progression in the transgenic plants as compared with the non-transgenic controls. Contamination of food and feed with mycotoxins is a worldwide problem. Although acute mycotoxicosis caused by high doses is rare in animals and humans, exposure to Fusarium mycotoxins can alter the human and animal susceptibility to infectious diseases by affecting the intestinal health and the innate and adaptive immune system [50]. The current study demonstrates that Fusarium resistance in barley induced by HvNEP-1 can facilitate a protection against mycotoxins.
Furthermore, the reliability of HvNEP-1 as an antifungal gene was examined under semi-field conditions where plants were grown in soil plots covered by an insect net (Figure 3a, Figure S1b). The result from the semi-field experiment was consistent with the greenhouse study where the overexpression of HvNEP-1 strongly suppressed FHB disease severity and progression (Figure 3b,c). On the basis of the semi-field analysis, the expression level of HvNEP-1 in line NEP20 was sufficient to reduce the infection rate and disease progression.
The current study describes a new resistance mechanism against FHB in barley. HvNEP-1 is demonstrated to have a strong potential as an FHB resistance gene and for reducing mycotoxins in infected seed. The transgenic approach to manage FHB used here can potentially minimize fungicide use, post-harvest sorting of infected grains, and combat mycotoxin-associated health risks. The use of azole fungicides for Fusarium control has aroused growing concern over the last years. Fungicides such as tebuconazole induces triazole-resistance in Aspergillus fumigatus [51]. At the same time, broad-spectrum triazoles are the primary drugs in the treatment of patients with prophylaxis of aspergillosis [52]. Fusarium resistant plants require less fungicide, and thereby alleviate the formation of triazole-resistant microorganisms in the environment.
Since the HvNEP-1 gene was isolated from barley, it can reduce safety concerns and could potentially also be used in cisand intra-genesis approaches [41]. Similarly, NEP-based strategies can be developed in other species where NEP candidates are available. Further studies are essential to understand the exact mechanism of HvNEP-1-based resistance, as well as to show if variation in HvNEP-1 levels is available in the natural germplasm and the extent to which these variants have the potential to protect the crop against FHB.

Conclusions
The results of this study provide insights into the novel HvNEP-1 based Fusarium resistance mechanism in barley that limits primary infection, the spread of the disease, and synthesis of mycotoxins in the grains. Utilization of HvNEP-1 based Fusarium resistance in our crops could help to reduce the use of azole fungicides in agriculture. Future studies would show to what extent HvNEP-1 based Fusarium resistance can be utilized in other crops.