A novel maize microRNA negatively regulates resistance to Fusarium verticillioides

Abstract Although microRNAs (miRNAs) regulate the defence response against multiple pathogenic fungi in diverse plant species, few efforts have been devoted to deciphering the involvement of miRNA in resistance to Fusarium verticillioides, a major pathogenic fungus affecting maize production. In this study, we discovered a novel F. verticillioides‐responsive miRNA designated zma‐unmiR4 in maize kernels. The expression of zma‐unmiR4 was significantly repressed in the resistant maize line but induced in the susceptible lines upon exposure to F. verticillioides exposure, whereas its target gene ZmGA2ox4 exhibited the opposite pattern of expression. Heterologous overexpression of zma‐unmiR4 in Arabidopsis resulted in enhanced growth and compromised resistance to F. verticillioides. By contrast, transgenic plants overexpressing ZmGA2ox4 or the homologue AtGA2ox7 showed impaired growth and enhanced resistance to F. verticillioides. Moreover, zma‐unmiR4‐mediated suppression of AtGA2ox7 disturbed the accumulation of bioactive gibberellin (GA) in transgenic plants and perturbed the expression of a set of defence‐related genes in response to F. verticillioides. Exogenous application of GA or a GA biosynthesis inhibitor modulated F. verticillioides resistance in different plants. Taken together, our results suggest that the zma‐unmiR4–ZmGA2ox4 module might act as a major player in balancing growth and resistance to F. verticillioides in maize.

insects (Feng et al., 2021;Li et al., 2018). The expression of miR393, the first miRNA identified to be involved in plant immunity, is induced by bacterial flagellin-derived peptide and restricts the growth of Pseudomonas syringae by repressing auxin signalling (Navarro et al., 2006). miR159a plays a positive role in rice resistance to Magnaporthe oryzae (Chen et al., 2021), whereas miR156 negatively regulates rice resistance to bacterial blight by Xanthomonas oryzae (Liu et al., 2019). In addition, the important roles of miR160a (Li et al., 2014), miR166 , miR528 (Wu et al., 2017), miR398b (Li et al., 2014), miR164 (Hu et al., 2020), and miR168 (Wu et al., 2015) in disease resistance by regulating specific target genes in various crops have been well characterized. For instance, miR528 negatively regulates rice resistance to rice stripe virus by cleaving the transcripts of the l-ascorbate oxidase (AO) gene (Wu et al., 2017). Loss of function of the Osa-miR159a target genes, including OsGAMYB, OsGAMYBL, and OsZF, results in enhanced resistance to M. oryzae, consistent with the related phenotypes of Osa-miR159a overaccumulation plants (Chen et al., 2021). miR156 negatively regulates rice resistance against bacterial blight through decreasing the expression levels of its targets IPA1 and OsSPL7 (Liu et al., 2019). Considering the extensive regulation of miRNAs during plant immunity, further characterization of pathogen-responsive miRNAs and resultant miRNA-mediated disease defence processes will have a profound impact on the development of new strategies for controlling disease damage in crop production.
Fusarium verticillioides is one of the most commonly occurring pathogenic fungi and causes various prevalent diseases in crops, especially maize, posing a great challenge to food and feed safety (Gai et al., 2018;Ju et al., 2017;Liu et al., 2020;Mu et al., 2018;Septiani et al., 2019). F. verticillioides infection occurs throughout the whole growth period of maize and results in seedling blight, stalk rot, ear rot, and seed rot (Machado et al., 2013;Septiani et al., 2019;Stagnati et al., 2019). Most importantly, F. verticillioides-infected plants or seeds may accumulate fumonisins, a family of mycotoxins associated with several diseases in livestock and humans and classified as probable carcinogens (Rosa Junior et al., 2019). Thus, it is of great significance to dissect the molecular mechanism of resistance to F. verticillioides. Although many genetic and omics studies have identified a series of quantitative trait loci/genes associated with F. verticillioides resistance (Butrón et al., 2019;Chen et al., 2016;Lanubile et al., 2017;Maschietto et al., 2017;Schiwek et al., 2020;Yao et al., 2020), the molecular mechanisms underlying the response of plants to F. verticillioides remain largely elusive, especially the role of miRNAs in this process. In our previous study using highthroughput sequencing (Zhou et al., 2020), a number of miRNAs, including known and new predicted miRNAs, were identified to be potentially associated with resistance to F. verticillioides ear rot.
Further functional analysis of these miRNAs is important to dissect the molecular mechanisms underlying the plant-F. verticillioides interaction and ultimately improve disease resistance.
In the current study, we focused on a novel F. verticillioidesresponsive miRNA designated zma-unmiR4 and aimed to reveal its function in the response of plants against F. verticillioides. We found that the expression levels of zma-unmiR4 were significantly down-regulated in the resistant maize line but up-regulated in the susceptible lines after F. verticillioides infection, whereas the target gene ZmGA2ox4 displayed the opposite profiles of expression.
Heterologous accumulation of zma-unmiR4 resulted in impaired resistance to F. verticillioides infection and enhanced growth in Arabidopsis; however, transgenic plants overexpressing ZmGA2ox4 or the homologue AtGA2ox7 showed high resistance to F. verticillioides as well as retarded growth. Further analyses indicated that zma-unmiR4 was able to regulate F. verticillioides resistance through gibberellin (GA) signalling by suppressing AtGA2ox7 expression in Arabidopsis. These results provide direct evidence for the crucial role of zma-unmiR4 in regulating plant growth and disease resistance to F. verticillioides.

| zma-unmiR4 is a novel maize miRNA responsive to F. verticillioides infection
Deep sequencing of small RNA libraries from maize kernels untreated or treated with F. verticillioides previously revealed a number of F. verticillioides-responsive miRNAs (Zhou et al., 2020), including 92 potentially novel miRNAs. These predicted miRNAs displayed various expression profiles in response to F. verticillioides ( Figure 1a). One novel miRNA candidate, designated zma-unmiR4, was characterized in more detail for its differential expression in the F. verticillioides-susceptible maize line N6 and the resistant line BT-1 ( Figure 1b). Amplification of its precursor sequence indicated that zma-unmiR4 is transcribed as an individual transcriptional unit from the maize genome ( Figure 1c). In addition, zma-unmiR4 transcription was confirmed through RNA blotting in maize kernels ( Figure 1d).
A high degree of complementarity for the precursor structure was observed using the RNAfold web server (Figure 1e). These observations support the notion that zma-unmiR4 represents a novel miRNA potentially regulating resistance to F. verticillioides in maize.
Moreover, zma-unmiR4 was found to be expressed in various maize tissues ( Figure S1), implying its potential functions during various developmental stages.

| ZmGA2ox4 and its homologue AtGA2ox7 are the targets of zma-unmiR4
Based on target gene prediction (http://rna.infor matik.uni-freib urg.de), zma-unmiR4 showed extensive sequence complementarity with the gene Zm00001d017294 encoding gibberellin 2-oxidase 4 (ZmGA2ox4; Figure 2a). Notably, the accumulation of ZmGA2ox4 transcripts was drastically increased in BT-1 but decreased in N6 Reverse transcription-quantitative PCR (RT-qPCR) also showed that ZmGA2ox4 transcript levels were significantly decreased when both transgenes were coexpressed (Figure 2f). To confirm a direct interaction between zma-unmiR4 and ZmGA2ox4, we also constructed a reporting system containing mutated vectors of ZmGA2ox4 (ZmGA2ox4M) and zma-unmiR4 (zma-unmiR4M) (Figure 2a,d). As shown in Figure 2e,f, neither cotransformation of zma-unmiR4M and normal ZmGA2ox4 nor cotransformation of ZmGA2ox4M and normal zma-unmiR4 effectively reduced the GUS signals and ZmGA2ox4 expression. Together, these results demonstrate that ZmGA2ox4 is a target of zma-unmiR4.
Arabidopsis AtGA2ox7 and AtGA2ox8, encoding homologous proteins of ZmGA2ox4, were predicted to be the putative heterologous targets of zma-unmiR4 (Figures 3a and S2). We compared the expression changes of AtGA2ox7 or AtGA2ox8 between wild-type (WT) and zma-unmiR4-overexpressing (zma-unmiR4 OE) plants ( Figure 3b). As shown in Figure 3c, AtGA2ox7 was significantly down-regulated while AtGA2ox8 displayed no obvious changes in both zma-unmiR4 overexpressors, suggesting that AtGA2ox7 may be targeted by zma-unmiR4. To verify this regulation in planta, AtGA2ox7 was fused with the gene encoding GUS, and this fusion gene was transiently coexpressed with 35S:pre-unmiR4 in tobacco. GUS activity and transcript levels were dramatically decreased compared with the vector control ( Figure 3d,e), and AtGA2ox7 transcript levels were greatly reduced ( Figure 3f). These data demonstrate that zma-unmiR4 negatively regulates AtGA2ox7 in Arabidopsis.

| Overexpression of zma-unmiR4 confers Arabidopsis growth and F. verticillioides susceptibility
To investigate the biological functions of zma-unmiR4, we developed homozygous transgenic Arabidopsis lines overexpressing F I G U R E 1 Identification and validation of zma-unmiR4. (a) Expression heatmap of the 92 predicted novel miRNA candidates identified by small RNA sequencing in BT-1 and N6 kernels after Fusarium verticillioides inoculation (Zhou et al., 2020). Maize kernels of BT-1 and N6 at 0 or 3 days postinoculation were sampled for construction of a small RNA library. (b) Reverse transcription (RT)-quantitative PCR analysis of differential expression of zma-unmiR4 in BT-1 and N6 maize lines after F. verticillioides inoculation. Maize EF1a was used as the internal control. Data are means ± standard deviation from three biological replicates. **p < 0.01 by Student's t test. (c and d) Verification of zma-unmiR4 production by RT-PCR amplifying its precursor (c) and RNA blotting (d). (e) Hairpin structure of zma-unmiR4 predicted by RNAfold software.
BT (e and f) Tobacco leaves were cotransfected with different reporter plasmids and different effector plasmids. After 2 days, the transfected leaves were used for β-glucuronidase (GUS) staining (e) and isolation of total RNA for reverse transcription-quantitative PCR analysis to determine the expression levels of GUS and ZmGA2ox4 (f). C, E, ME, R, and MR indicate various constructs of control, effector, mutated effector, reporter, and mutated reporter in panel (d), respectively. Data are means ± standard deviation from three biological replicates. **p < 0.01 by Student's t test.    Figure S6). The same results as in Figure 4b were obtained, suggesting that the differences of F. verticillioides resistance were due to the genotypic variation. In addition, the rosette leaves of zma-unmiR4 OE and atga2ox7 plants displayed more severe blight or death phenotypes after spraying with F. verticillioides spore suspension; however, the transgenic plants of ZmGA2ox4 OE or AtGA2ox7 OE were almost unaffected ( Figure 4e).
We further tested whether there existed differences in F. verticillioides seed rot among WT, atga2ox7 mutant, and the transgenic plants indicated above. The seeds from various genotypes were incubated with a F. verticillioides spore suspension, and the phenotypes of fungal mycelia growth on the seed surface were recorded after 6 days. Compared to the water treatment control, the growth and invasion areas of fungal mycelia showed remarkable differences among various genotypes after F. verticillioides inoculation. The seeds from zma-unmiR4 OE or atga2ox7 mutant plants were more sensitive to F. verticillioides but seeds from ZmGA2ox4 OE or AtGA2ox7 OE transgenic plants were more resistant ( Figure S7). In detail, more than half of the seeds from zma-unmiR4 OE and atga2ox7 mutant plants exhibited disease grades II and III; however, most seeds from ZmGA2ox4 OE and AtGA2ox7 OE plants belonged to grade I according to the three grades of disease resistance (Figure 4f,g). These data suggested that zma-unmiR4 could positively regulate plant growth and negatively regulate F. verticillioides resistance by manipulating AtGA2ox7 or ZmGA2ox4 expression.

| Altered resistance to F. verticillioides by zma-unmiR4 is associated with the production of H 2 O 2
As a necrotrophic fungal pathogen, F. verticillioides might ultimately kill and benefit from the infected host cells (Rivas-San Vicente

| Development of F. verticillioides resistance by zma-unmiR4 is correlated with the expression of defence-related genes
We The relative lesion area (lesion area/total area of each leaf) was measured by ImageJ software. Ten leaves from three biological replicates were analysed for each genotype. (d) Investigation of F. verticillioides content in the inoculated leaves of indicated genotypes. F. verticillioides tubulin level determined by quantitative PCR was used as an indicator of F. verticillioides content. Arabidopsis Actin 2 was used as internal control. (e) Disease symptoms of different Arabidopsis genotypes after F. verticillioides spraying. Four-week-old plants were sprayed with a F. verticillioides spore suspension or sterile water. Bar = 3 cm. (f) Disease grades of F. verticillioides-caused seed rot. Healthy dry seeds were sterilized, immersed in F. verticillioides spore suspension for 48 h, and placed on sterile filter paper for 6 days for disease grade investigation. Three disease grades were classified according to the infected area (IA, the area covered by mycelia on one seed/the total area of the same seed): grade I, 0 < IA ≤ 25%; grade II, 25% < IA ≤ 50%; grade III, 50% < IA ≤ 100%. Bar = 250 μm. (g) Disease grades of seed rot among indicated genotypes. Data are from a total of 150 seeds for each genotype. For panels (c) and (d), data are means ± standard deviation from three biological replicates. Letters above the bars indicate significant differences (p < 0.05).    These data suggested that zma-unmiR4-mediated suppression of AtGA2ox7 might disturb the induction of defence-related genes by F. verticillioides, thus resulting in resistance variations.

| GA accumulation is associated with F. verticillioides resistance
AtGA2ox7, a member of the gibberellin 2-oxidase family, is a 2-oxoglutarate-dependent dioxygenase that regulates the deactivation of bioactive GAs (Li et al., 2019). We analysed the endogenous content of bioactive GA1, GA3, GA4, and GA7 in the rosette leaves of 4-week-old plants. The levels of GA3 and GA4 were too low to detect in WT, zma-unmiR4 OE, and AtGA2ox7 OE samples tested, but GA1 accumulated to higher levels in zma-unmiR4 OE plants than in WT and AtGA2ox7 OE plants (Figure 6a). In addition, compared with WT, zma-unmiR4 OE transgenic plants accumulated higher levels of GA7 while the contents of GA7 were significantly decreased in AtGA2ox7 OE (Figure 6b). These results suggested that the zma-unmiR4-AtGA2ox7 module mediates plant growth and F. verticillioides resistance probably through regulating endogenous bioactive GA accumulation.
To further investigate the effects of GA on plant disease resistance and growth, 17-day-old seedlings of WT, zma-unmiR4 OE, and AtGA2ox7 OE transgenes were sprayed with GA or the GA synthesis inhibitor uniconazole. As expected, the growth of WT and AtGA2ox7 OE seedlings was enhanced by GA treatment, but growth inhibition was observed for both zma-unmiR4 OE and WT plants when treated with uniconazole ( Figure 7a). We then inoculated the leaves with a F. verticillioides spore suspension. Compared to the water control treatment, the leaves of WT and AtGA2ox7 OE plants treated with GA displayed larger yellow necrotic lesions, significantly increased F. verticillioides content, and higher H 2 O 2 levels as well as cell death (Figure 7b-e). In contrast, both WT and zma-unmiR4 OE plants treated with uniconazole exhibited significantly smaller necrotic lesions and less F. verticillioides content, and the H 2 O 2 and cell death levels were much lower than in the water-treated control ( Figure 7b-e).
Moreover, we applied GA or uniconazole on the susceptible maize line N6 to test the changes in resistance to F. verticillioides.
Compared with the application of water, the susceptibility of N6 seedlings to F. verticillioides was greatly promoted by GA, consistent with increased F. verticillioides content, necrotic lesions, H 2 O 2 accumulation, and cell death (Figure 7f-k). In contrast, maize seedlings treated with uniconazole displayed obviously smaller necrotic lesions ( Figure 7g,h), significantly less F. verticillioides content (Figure 7i), decreased H 2 O 2 levels (Figure 7j,k), and mild cell death (Figure 7j).
Furthermore, similar results were observed in rice seedlings treated with GA and uniconazole ( Figure S9). When the rice seedlings were sprayed directly with a F. verticillioides spore suspension, the disease symptoms of the seedlings treated with GA were obviously enhanced, while the opposite was observed in seedlings treated with uniconazole ( Figure S9). Collectively, these results demonstrate that GA plays a negative role in plants resistance to F. verticillioides.

| DISCUSS ION
F. verticillioides is one of the most common pathogenic fungi and can cause many prevalent diseases in crops, especially in maize, such as seedling blight, root rot, stalk rot, ear rot, and seed rot, Given that miRNAs provide quantitative regulation of target gene expression rather than switching regulation, the dynamic accumulation of pathogen-responsive miRNAs can provide finetuning of target gene expression during pathogen infection, thus in turn enhancing the plant's disease resistance (Campo et al., 2013).
High-throughput sequencing of small RNAs is an effective method F I G U R E 6 Quantification of endogenous bioactive gibberellins in wildtype (WT), zma-unmiR4 overexpression (OE), and AtGA2ox7 OE plants. Shoots of 4-week-old Arabidopsis plants of various genotypes were sampled for quantifying gibberellins GA1 and GA7 content. Data are means ± standard deviation of three biological replicates. **p < 0.01 by Student's t test. For panels (f) to (k), 7-day-old seedlings were sprayed with water, GA (50 μM), or uniconazole (20 μM) once a day for 7 days, inoculated with a F. verticillioides spore suspension, and photographed, stained, or sampled at 4 days after inoculation. For panels (c) and (h), relative lesion area was measured by ImageJ software, and more than 10 leaves from three biological replicates were analysed for each group. For panels (d)   to discover pathogen-responsive miRNAs, including conserved and novel miRNAs. Although false-positive prediction of novel miRNAs cannot be ruled out during sequencing and data processing, the function of these miRNAs in pathogen resistance should be fully considered. For instance, Md-miRln20 , osa-miR7695 (Campo et al., 2013), and Md-miRLn11 (Ma et al., 2014) were characterized by small RNA sequencing and their function in disease resistance was experimentally validated. In a previous study, multiple F. verticillioides-responsive miRNAs were identified using small RNA deep sequencing (Zhou et al., 2020), GAs are phytohormones that play multiple roles in plant development and stress responses (Rizza & Jones, 2019;Schomburg et al., 2003). Endogenous levels of bioactive GAs are maintained through a balance of biosynthesis and inactivation. AtGA2ox7 is a 2-oxoglutarate-dependent dioxygenase that regulates the deactivation of bioactive GAs (Li et al., 2019). Consistently, transgenic plants overexpressing AtGA2ox7 showed a significant reduction of bioactive GAs compared to the WT, thus exhibiting GA-deficient phenotypes such as dwarfism, delayed flowering, and small dark green leaves (Porri et al., 2012;Schomburg et al., 2003;Shu et al., 2016) ( Figures 4a,b, 6 and S4). By contrast, dysfunction of AtGA2ox7 resulted in GA-induced phenotypes, including enhanced growth, large leaf size, and early flowering (Magome et al., 2008;Rieu et al., 2008;Shu et al., 2016), which was consistent with the phenotypes of zma-unmiR4 OE plants and higher GA contents (Figures 4a,b, 6 and S4).
Therefore, we have reason to believe that the high level of bioactive GAs by zma-unmiR4-mediated repression of AtGA2ox7 is responsible for the phenotypic changes of zma-unmiR4 OE plants.
Although the function of bioactive GA in plant growth and development is well known, the role of GA in plant resistance to F. verticillioides remains unclear. In fact, GA was first identified from Gibberella fujikuroi (Fusarium moniliforme), a necrotrophic fungus that causes rice bakanae disease (Yabuta & Sumiki, 1938). Overexpression of the GA-deactivating enzyme Eui can increase resistance to bacterial blight and rice blast caused by X. oryzae and M. oryzae, respectively; however, transgenic rice overexpressing OsGA20ox3 (encoding a GA biosynthesis enzyme) was more susceptible to both diseases (Qin et al., 2013;Yang et al., 2008). Similarly, our current results through genetic and physiological analysis in Arabidopsis or maize plants demonstrated that GAs also exhibit a negative effect on the resistance to F. verticillioides (Figures 4 and 7). Modification of the expression levels of AtGA2ox7 or ZmGA2ox4, which encode GA-deactivating enzymes, could change F. verticillioides resistance ( Figure 4). Despite the enhanced resistance to F. verticillioides upon overaccumulation of AtGA2ox7 in Arabidopsis, many adverse effects on development were observed, such as dwarfism and delayed flowering (Figure 4a), which would also be expected to occur in maize. Application of the GA biosynthesis inhibitor uniconazole significantly reduces the lodging rate and enhances yield in maize (Ahmad et al., 2021). Our data also showed the positive effects of uniconazole on F. verticillioides resistance and dwarf traits in maize and rice (Figure 7f-k and S9). Therefore, fine-tuning of the zma-unmiR4-ZmGA2ox4 regulatory module could theoretically be an alternative way to generate desirable resistance to F. verticillioides and to lodging without growth or yield penalty in maize breeding.
In rice, overexpression or dysfunction of the GA-deactivating enzyme Eui results in disturbed homeostasis of SA and JA, thus leading to altered disease susceptibility (Yang et al., 2008). In addition, the SA receptor-encoding gene NPR1 is a master regulator of systemic acquired resistance in plants, and overaccumulation of NPR1 leads to enhanced disease resistance to diverse pathogens (Ding et al., 2018). In support of this notion, we found that NPR1 In conclusion, we discovered a novel F. verticillioides-responsive miRNA, zma-unmiR4, in maize kernels and revealed that ZmGA2ox4 and its homologue AtGA2ox7 are the targets of zma-unmiR4. We showed that zma-unmiR4-mediated suppression of AtGA2ox7 disturbed the accumulation of bioactive GA, and zma-unmiR4-ZmGA2ox4/AtGA2ox7-mediated bioactive GA dynamics act as a crucial regulator in F. verticillioides resistance and plant growth ( Figure 8). However, as a master regulator of plant growth and development, GA confers susceptibility to multiple plant diseases (Qin et al., 2013;Yang et al., 2008;Zhang et al., 2020). Our study provides a new strategy for a better balance between F. verticillioides disease resistance and growth in maize breeding by engineering the zma-unmiR4-ZmGA2ox4 module.

| Plant materials and growth conditions
The F. verticillioides-susceptible maize inbred line N6 is a Tangsipingtou line, whereas the F. verticillioides-resistant maize line BT-1 is improved by tropical Asia material .
Seedlings were grown at 25 ± 2°C with a 14/10-h light/dark photo- pPROTO driven by a 35S promoter. The primers used for vector construction are listed in Table S1.

| Transfection of maize protoplasts
Maize protoplasts were isolated as previously described .
35S::ZmGA2ox4-YFP was cotransfected into the protoplasts with 35S::pre-zma-unmiR4 or empty vector. After incubation in the dark for 16 h, YFP and mCherry signals were observed using a laser scanning confocal microscope (A1HD25; Nikon). The relative fluorescence intensity (YFP/mCherry) was calculated by ImageJ software.

| F. verticillioides inoculation and phenotype investigation
The F. verticillioides strain was isolated from naturally infected maize kernels in Zhengzhou. A single spore of F. verticillioides was isolated and propagated on sterilized maize kernels at 28°C for 7 days. The spores were then collected and diluted to the indicated concentration using sterile distilled water with 0.2 μl/ml Tween 80. The ear inoculation was performed as previously described (Wu et al., 2020;Zhou et al., 2020). The middle of the ears was injected with 2 ml F.
For Arabidopsis leaf inoculation, the healthy rosette leaves at indicated times were inoculated with 20 μl F. verticillioides spore suspension (1 × 10 7 spores/ml). After 4-6 days of culture at 22°C, leaves were photographed or sampled for histological staining, F. verticillioides quantification, and H 2 O 2 content determination. For spore suspension spraying, 5-week-old plants were sprayed with a F. verticillioides spore suspension (2 × 10 7 spores/ml) or sterile water once a day for 10 days. For seed inoculation, sterilized seeds were soaked in a F. verticillioides spore suspension (1 × 10 7 spores/ml) at 28°C darkness for 48 h. Then the seeds were evenly placed on wet filter paper in Petri dishes at 28°C in darkness for 6 days.
For maize leaf inoculation, healthy leaves were lacerated with a needle and injected with 10 μl of a F. verticillioides spore suspension (1 × 10 7 spores/ml). After incubation at 25°C for 2-5 days, leaves were photographed or sampled for histological staining, F. verticillioides quantification, and H 2 O 2 content determination.
For rice leaf inoculation, healthy leaves were scratched with a needle, immersed in 3 ml of a F. verticillioides spore suspension (2 × 10 7 spores/ml), and incubated at 25°C for 5 days. For seedling inoculation, healthy seedlings were sprayed with a F. verticillioides spore suspension (2 × 10 7 spores/ml) once a day for 6 days.

| RNA analyses
About 1 μg RNA was treated with DNase I (Promega) and reverse transcribed using the Transcriptor First Strand cDNA Synthesis Kit (TOYOBO). The RT-qPCR assay was performed using SYBR Green I Master reagent and a STEP ONE PLUS system (ThermoFisher).
The expression levels of target genes were normalized to the internal control genes using the 2 −ΔΔCt method. RNA gel blot analyses of miRNA were performed as described previously (Zhang & Li, 2013).
Primer and probe sequences are listed in Table S1.

| GUS, DAB, and TB staining and H 2 O 2 quantification
The transient expression assay for GUS analysis was performed as previously described . For DAB staining, leaves were immersed in 0.1% DAB solution, infiltrated under vacuum conditions for 15 min, and then incubated at room temperature for 12 h in the dark. Chlorophyll was removed by immersion in 95% ethanol.
H 2 O 2 quantification was performed as previously described (Garg et al., 2012). TB staining was performed by submerging the leaves in TB solution (10 ml lactic acid, 10 ml phenol, 10 ml glycerol, 10 ml sterile water, and 10 mg trypan blue) for 30-60 min. Chlorophyll was removed by immersion in 95% ethanol.

| Hormone treatments
Seven-day-old maize seedlings were sprayed with 20 μM uniconazole, 50 μM GA3, or water once a day for 7 days and photographed, and then leaves were inoculated. Fourteen-day-old rice seedlings were sprayed with 20 μM uniconazole, 50 μM GA3, or water once a day for 4 days and then sprayed with a F. verticillioides spore suspension. Seventeen-day-old Arabidopsis seedlings were sprayed with 20 μM uniconazole, 50 μM GA3, or water once a day for 5 days and photographed, and then leaves were inoculated as described above.

| Determination of chlorophyll concentration
The measurement of chlorophyll concentration was performed as previously described (Arnon, 1949). Leaves from 5-week-old plants were sampled. The absorbance was measured at 663 nm, 645 nm, and 652 nm using an ELISA instrument.

| Gibberellin measurement
Healthy rosette leaves of 4-week-old plants were harvested, immediately frozen in liquid nitrogen, and ground into powder. Next, 50 mg of plant sample was dissolved in 500 μl HPLC-grade acetonitrile/water (90:10, vol/vol). As internal standards for the quantification, 10 μl internal standard solution (100 ng/ml) was added into the extract. GA contents were calculated by MetWare (http:// www.metwa re.cn/) based on the AB Sciex QTRAP 6500 liquid chromatography-tandem mass spectrometry platform. Three biological replicates were performed.
analysed the data. Y.X. and H.Z. prepared the figures and wrote the article. All authors read and approved this manuscript.

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.

S U PP O RTI N G I N FO R M ATI O N
Additional supporting information may be found in the online version of the article at the publisher's website.