Integrative transcriptome and proteome analysis reveals maize responses to Fusarium verticillioides infection inside the stalks

Abstract

Fusarium stalk rot caused by Fusarium verticillioides is one of the most devastating diseases of maize that causes significant yield losses and poses potential security concerns for foods worldwide. The underlying mechanisms of maize plants regulating defence against the disease remain poorly understood. Here, integrative proteomic and transcriptomic analyses were employed to identify pathogenesis‐related protein genes by comparing differentially expressed proteins (DEPs) and differentially expressed genes (DEGs) in maize stalks after inoculation with F. verticillioides. Functional enrichment analysis showed that DEGs and DEPs were mainly enriched in glutathione metabolism, starch and sucrose metabolism, amino sugar and nucleotide sugar metabolism, linoleic acid metabolism, and phenylpropanoid biosynthesis. Fourteen DEGs and DEGs that were highly elevated after inoculation with F. verticillioides were confirmed with parallel reaction monitoring and reverse transcription‐quantitative PCR, demonstrating the accountability and reliability of proteomic and transcriptomic data. We also assessed the potential roles of defence‐related genes ZmCTA1, ZmWIP1, and ZmLOX2, identified from the multi‐omics analysis, during the process of F. verticillioides infection through virus‐induced gene silencing. The elevation of stalk rot symptomatic characteristics in the silenced plants revealed their contribution to resistance. We further functionally characterized the roles of ZmLOX2 expression in the defence response of maize plants conditioning fungal invasion via the salicylic acid‐dependent pathway. Collectively, this study provides a comprehensive analysis of transcriptome and proteome of maize stalks following F. verticillioides inoculation, and defence‐related genes that could inform selection of new genes as targets in breeding strategies.

This study analysed the transcriptome and proteome of maize stalks following Fusarium verticillioides infection and characterizes the roles of ZmLOX2 in the maize defence response via the salicylic acid‐dependent pathway.

1. INTRODUCTION

Stalk rot is one of the most destructive diseases of maize (Zea mays) worldwide (Mueller & Wise, 2015). The ascomycete fungal genus Fusarium is considered to be the most frequently reported causative agent of maize stalk rot diseases (Jackson et al., 2009). Fusarium verticillioides, an important maize pathogen that produces fumonisins and causes stalk rot and ear rot, has become one of the most aggressive causal agents of maize diseases in China in recent years (Liu et al., 2019).

The fungal infection and consequent contamination with mycotoxins dramatically affect yield and quality, thereby causing multimillion dollar losses annually and posing serious health hazards to both humans and animals (Duan et al., 2016; Mueller et al., 2016; Pechanova & Pechan, 2015; Silva et al., 2017). Agronomic and chemical practices have had a little success in preventing infection of F. verticillioides (Wilke et al., 2007). Development of inbred and hybrid resistance to F. verticillioides is therefore the most economically practical strategy for environmentally friendly and sustainable long‐term prevention. However, totally immune genotypes are not available and commercial hybrids have less resistance than desired (Ju et al., 2017; Maschietto et al., 2017). No specific resistance genes that confer immunity to Fusarium stalk rot have been documented to date. A few causal genes for resistance to Gibberella (F. graminearum) stalk rot and the underlying mechanism of resistance have been examined (Wang et al., 2017; Ye et al., 2019), whereas for the F. verticillioides stalk rot there are few studies focused on the molecular interactions between the pathogen and the host maize plants, and the underlying molecular mechanisms of resistance are not clear, which limits progress toward effectively controlling the disease.

Plants have developed many sophisticated mechanisms to adapt or survive under the invasion of different types of pathogens. The response to pathogen attack in plants is a complex network of expression of genes including not only those that participate directly in defence but also those that moderate key major metabolic pathways (Pechanova & Pechan, 2015; Taylor et al., 2004) such as hormone regulation and signalling (Pieterse et al., 2012; Wang et al., 2016; Ye et al., 2019), generation of antioxidants, and induction of stress proteins (Lanubile et al., 2017). Previous studies demonstrated signal transduction to the nucleus to trigger a defence response or hypersensitive response (HR) through a cellular signalling network involving secondary messengers such as reactive oxygen species (ROS) and calcium, calcium‐associated proteins, and kinase cascades such as mitogen‐activated protein (MAP) kinase cascades (Qi et al., 2018; Sewelam et al., 2016; Wang & Balint‐Kurti, 2016) after perception of pathogen invasion. Wang and Balint‐Kurti (2016) documented significant associations of genes encoding two key enzymes in lignin biosynthesis, hydroxycinnamoyl transferase, and caffeoyl CoA O‐methyltransferase with variation in the severity of the rust resistance gene Rp1‐D21‐induced HR (Wang & Balint‐Kurti, 2016). Pathogen‐associated molecular patterns trigger the small heat shock protein (sHSP) family, several secondary metabolites, and the signalling pathways of abscisic acid, jasmonic acid (JA), or salicylic acid (SA) to defend against F. verticillioides infection (Wang et al., 2016); four genes known to participate in secondary metabolism, five genes encoding sHSPs, and 10 genes involved in signal transduction processes were identified as overlapping or very near to two quantitative trait loci (QTLs) for resistance to F. verticillioides in maize. The pathogenesis‐related genes participating in JA and ethylene (ET) signalling pathways and the shikimate biosynthesis pathway are also related to resistance to F. verticillioides (Campos‐Bermudez et al., 2013; Lanubile et al., 2014). Shu et al. (2017) characterized a set of genes including PR‐5, PR‐10, PR‐10.1, Protein P21, and four chitinase genes responsible for resistance to F. verticillioides.

Protein‐based molecular characterization of resistance to Fusarium infection with high‐throughput proteomics (Pechanova & Pechan, 2015) has provided a more accurate reflection of functional heterogeneity and stronger predictors of response to fungal invasion. However, very few studies have examined the proteomic profiling of maize resistance to F. verticillioides. Mohammadi et al. (2011) documented the proteomic profiling of two maize inbred lines, B73 and CO441, in responce to the early stages of infection by F. graminearum. The defence‐responsive proteins, including pathogenesis‐related 10 (PR‐10), chitinases, xylanase inhibitors, proteinase inhibitors, and a class III peroxidase, were more abundant in the tolerant inbred line CO441. Pathogenesis‐related proteins β‐1,3‐glucanase (PR‐2) and chitinase (PR‐3) were observed in greater abundance in wheat line Arina, resistant to F. graminearum infection, in comparison to the susceptible line Agent (Kang & Buchenauer, 2002), inhibiting the hydrolysis of plant cell walls by pathogenic fungi (Nicholson et al., 2007). Reid and Altosaar (2019) observed up‐regulation of defence proteins in response to F. graminearum infection in maize lines CO441 and CL30. Therefore, proteomic analysis provides a promising tool to explore the pathogenesis of F. verticillioides at the protein level.

Proteomics technology could be used to explore proteins related to plant disease resistance, such as plant oxylipins. Mounting evidence has shown that plant oxylipins play crucial roles in the defence response to pathogens (Deboever et al., 2020). Oxylipins are mainly generated from linoleic acid (C18:2) or α‐linolenic acid (C18:3), the catalysis of which is initiated by lipoxygenases (LOXs) and α‐dioxygenases (DOX) (Andreou et al., 2009; Feussner & Wasternack, 2002;). LOXs are divided into 9‐ and 13‐LOXs according to the oxidized position of carbon (C) (Feussner & Wasternack, 2002). It has been well‐documented that maize 13‐LOX, ZmLOX10, is a highly induced on pathogen infection (Nemchenko et al., 2006). Reports have also showed that 9‐oxylipins play a role in plant–pathogen interactions. In maize, ZmLOX1, 2, 3, 4, 5, and 12 are 9‐LOXs (Christensen et al., 2014). Among these, ZmLOX3, ZmLOX4, ZmLOX5, and ZmLOX12 are required for fungal pathogenesis by F. verticillioides and other pathogens, or defence against pathogens in maize root, stem, and ear (Battilani et al., 2018; Christensen et al., 2014; Gao et al., 2007, 2008, 2009; Gorman et al., 2020; Lanubile et al., 2021; Park et al., 2010; Wang et al., 2020).

In the present study, we analysed the abundance of transcripts and proteins with transcriptomic analysis and isobaric tags for relative and absolute quantitation (iTRAQ) labelling (Ross et al., 2004) coupled with liquid chromatography‐tandem mass spectrometry (LC–MS/MS) to probe candidate genes and metabolic pathways participating in defence responses to F. verticillioides infection in maize inbred line Chang7‐2 (CH7‐2). Parallel reaction monitoring (PRM) and reverse transcription‐quantitative PCR (RT‐qPCR) were employed to quantitatively validate the expression levels of candidate genes associated with tolerance to F. verticillioides. Based on the multi‐omics analysis, we further analysed the effect of ZmLOX2 knockdown during pathogen infection using brome mosaic virus (BMV)‐based virus‐induced gene silencing (VIGS) of ZmLOX2. Silencing of ZmLOX2 facilitated the infection of F. verticillioides in the maize plants. ZmLOX2 regulated the maize response to F. verticillioides infection through the SA‐triggered defence pathway. These findings provide a foundation for exploring the protein expression profile of resistance‐related genes in maize, and the role of ZmLOX2 expression in defence responses of maize plants following fungal invasion.

2. RESULTS

2.1. The process of F. verticillioides infection inside maize stalks

To gain insight into the process of how F. verticillioides survives and proliferates inside the maize stalk, we artificially wounded the stalk of plants of maize inbred line CH7‐2 approximately at the appearance of the ninth leaf, and inoculated plants with F. verticillioides mycelial plug inocula (Figure 1a). From 24 h to 2 weeks after inoculation, the challenged plants were harvested and the inoculated internodes were split to observe fungal spread and the appearance of the stalk symptoms in detail. To the naked eye, the split internodes were almost symptomless for up to 36 h postinoculation (hpi); stalk tissues at the inoculated site turned pale brown by 48 hpi, and by 3 days postinoculation (dpi), the brown colour became darker and the brown area elongated above and below the site of the wound, whereas at roughly 5 dpi, more than half of the inoculated internodes exhibited typical dark‐brown symptoms, and even the diseased tissues near the wound site were totally collapsed (Figure 1b).

FIGURE 1.

The process of Fusarium verticillioides infection inside maize stalks. (a) The stem at the second or third internode above the soil line of maize inbred line Chang7‐2 (CH7‐2) seedlings at the ninth‐leaf stage were artificially inoculated with F. verticillioides; plants inoculated with a potato dextrose agar (PDA) plug served as a mock treatment. (b) The dark‐brown symptoms were observed in the split internodes at 5 days postinoculation (dpi). (c) Measurement of physiological and biochemical characters of CH7‐2 seedlings after F. verticillioides inoculation. The samples were harvested at 1, 3, and 5 dpi. Water‐soluble protein content (WSPC), water‐soluble saccharide concentration (WSSC), catalase (CAT) activity, dehydrogenase (DHA) activity, lipid peroxidation MDA (malondialdehyde) level, polyphenol oxidase (PPO) activity, peroxidase (POD) activity, superoxide dismutase (SOD) activity, phenylalanine ammonia‐lyase (PAL) activity, and lignin content were measured. Bars show mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001 (paired Student's t test). Scale bars in (a) and (b), 5 cm. DW, dry weight; MK, mock; Fv, F. verticillioides.

In addition to the symptomatic aspects, the different physiological and biochemical responses of CH7‐2 during infection were also characterized at 1, 3 and 5 dpi. In brief, water‐soluble saccharide content (WSSC), the concentrations of malondialdehyde (MDA) and lignin, and the activities of catalase (CAT), dehydrogenase (DHA), polyphenol oxidase (PPO), peroxidase (POD), superoxide dismutase (SOD), and phenylalanine ammonia‐lyase (PAL) were found to be significantly increased after F. verticillioides fungal infection (Figure 1c). In contrast, the contents of water‐soluble protein concentration (WSPC) and of chlorophyll in the F. verticillioides‐inoculated plants were remarkably decreased compared to those observed in the mock‐inoculated CH7‐2 plants (Figures 1c and S1).

Considering the proliferation process of F. verticillioides inside the maize stalks and the physiological and biochemical responses of maize plants to the fungal inoculation, the upper and lower segments immediately adjacent to the diseased stalk segments were harvested at 3 dpi and subjected to analysis of differentially expressed genes (DEGs) and proteins after fungal infection.

2.2. Transcriptomic analysis of DEGs and function analysis

To better understand the differences between the F. verticillioides‐inoculated plants and the mock‐inoculated (MK) plants, we performed high‐throughput transcriptome sequencing of stem tissue samples from near the inoculation site. Six cDNA libraries were constructed, and a total of 291.88 million and 284.93 million raw sequence reads were generated from the F. verticillioides and MK libraries, respectively. After removing low‐quality reads and adaptor sequences, 290.25 million and 283.15 million clean reads were obtained, respectively, with 93.67%–94.25% Q30 bases and 51.30%–54.04% GC content (Table S1). The clean reads were mapped onto the maize B73 RefGen_v4 reference genome. The average mapping rate was 89.99% for the six samples (Table S2). In the F. verticillioides and MK groups, 32,017 and 31,961 genes, respectively, were identified from all three replicates. Moreover, a total of 31,073 genes were found to be expressed in both groups, accounting for more than 97% of the total number (Figure 2a). Principal component analysis (PCA) revealed that the three replicates of each group were located nearest to each other (Figure 2b). We used fragments per kilobase of transcript per million mapped reads (FPKM) values to represent gene expression levels and drew boxplots and expression density plots for the expression of all genes in the two groups. The results were consistent with the above analysis (Figure S2a,b). These analyses demonstrated that the raw RNA‐sequencing (RNA‐Seq) data exhibited high levels of reliability and reproducibility.

FIGURE 2.

Overall transcriptome analysis results in the Fusarium verticillioides‐inoculated and mock‐inoculated (MK) groups. (a) Venn diagram of the common genes expression between F. verticillioides and MK groups. The left and right sides of the Venn diagram represent the genes specifically expressed in the F. verticillioides and MK groups, respectively, and the overlapping part in the middle represents the number of genes co‐expressed in both groups. (b) Plots of principal component analysis in the MK and F. verticillioides groups. The blue triangles represent the MK group and the orange dots represent the F. verticillioides group. (c) Histogram of differential gene expression. (d) Volcano plot of differential gene expression. Up‐regulated and down‐regulated genes are labelled with red and blue dots, respectively. GO (e) and KEGG (f) enrichment of differentially expressed genes (DEGs). The bubble diagram shows the degree of enrichment of GO and KEGG terms and the top 20 GO terms with the lowest p _adj values are used in the diagram. The abscissa represents the enrichment ratio and the ordinate denotes the GO term or KEGG pathway. The size of bubbles indicates the number of genes annotated to a certain GO term or KEGG pathway, and the colour represents the p _adj value. GO terms and pathways relevant to our study are marked yellow in the figure. (g) Cluster heatmap of differential gene expression in plant disease resistance‐related pathways. Red and blue indicate high and low expression levels (log₂ fragments per kilobase of transcript per million mapped reads, FPKM), respectively.

To investigate transcriptome changes resulting from F. verticillioides infection, we compared the gene expression between the F. verticillioides and MK groups. According to DESeq2 analysis, |log₂(fold change)| ≥ 1 and p _adj < 0.05 were used as screening conditions to identify the DEGs. A total of 2814 DEGs were identified, among which 1538 were up‐regulated and 1276 were down‐regulated. The volcano plot demonstrates DEGs between the F. verticillioides and MK groups (Figure 2c,d). We plotted the cluster heatmap for all DEGs and found that the DEGs within each group had a high correlation (Figure S2c).

The Gene Ontology (GO) annotation of up‐ and down‐regulated DEGs indicated that most down‐regulated DEGs in the GO terms of the biological process (BP) were involved in the metabolic process (GO:0008152), the cellular process (GO:0009987), the single‐organism process (GO:0044699), and the response to stimulus (GO:0050896), while up‐regulated DEGs were more involved in regulation of the biological process (GO:0050789), and biological regulation (GO:0065007), suggesting that up‐ and down‐regulated DEGs may have different functions. Similar results were also observed in molecular function (MF) and cell components (CCs) (Figure S3). Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis indicated that most DEGs were annotated on metabolic pathways (ko01100), amino sugar and nucleotide sugar metabolism (ko00520), biosynthesis of secondary metabolites (ko01110), plant–pathogen interaction (ko04626), phenylpropanoid biosynthesis (ko00940), plant hormone signal transduction (ko04075), and MAPK signalling pathway ‐ plant (ko04016) (Figure S4).

We further performed GO and KEGG enrichment analysis of DEGs. The pathway enrichment charts show the top 20 GO terms and KEGG pathways with the most reliable enrichment significance (the lowest p _adj value) (Figure 2e,f). GO enrichment analysis was conducted for all the DEGs and some interesting GO terms were significantly enriched, for instance the lignin catabolic process (GO:0046274), the cell wall macromolecule catabolic process (GO:0016998), the chitin catabolic process (GO:0006032), chitinase activity (GO:0004568), hydrolase activity, acting on ester bonds (GO:0016788), response to oxidative stress (GO:0006979), peroxidase activity (GO:0004601), and the hydrogen peroxide catabolic process (GO:0042744). The enriched GO terms belong to the biological process and molecular function categories. KEGG enrichment analysis was also performed and some important pathways related to plant disease resistance were significantly enriched, such as MAPK signalling pathway – plant (ko04016), tyrosine metabolism (ko00350), amino sugar and nucleotide sugar metabolism (ko00520), plant–pathogen interaction (ko04626), cutin, suberin and wax biosynthesis (ko00073), and phenylpropanoid biosynthesis (ko00940). Some pathways crucial for plant disease resistance, such as linoleic acid metabolism (ko00591), flavonoid biosynthesis (ko00941), and plant hormone signal transduction (ko04075), were not in the top 20 pathways with the lowest p _adj values, having values of 0.009984316, 0.206875, and 0.024095385, respectively. The DEGs involved in these pathways were identified as candidate genes (Figure 2g and Table S3).

2.3. Proteomic analysis of DEPs and function analysis

Proteomics provides a complementary approach to genomic technologies for understanding the adaptation mechanisms of organisms under stresses. In this study, proteomic analyses were also conducted for F. verticillioides‐inoculated and MK plants. A total of 296,599 spectra were obtained between the F. verticillioides and MK groups, of which 92,028 were matched to known spectra (Table S4). Overall, 18,056 peptides corresponding to 3969 unique proteins were identified at the 95% confidence level. These results indicate that iTRAQ has high sensitivity, with a spectral identification rate of 31.03% (Table S2a).

Before analysis of the proteins identified from the high‐throughput assay, we first examined the quality of the data obtained from the mass spectrometry. The distribution of the peptide number defining each protein is illustrated in Figure S5a. More than 73.87% (2932/3963) of the unique proteins harboured at least two peptides. The average length of the polypeptide identified in the assay was 15.45 amino acid residues, which is within a reasonable range of the peptide length. The results also showed that the length of the peptides was mainly concentrated between 8 and 22 amino acid residues, with 12 being the commonest length (Figure S5b). Subsequently, repeatability analysis was conducted to confirm the reproducibility of the biological replicates by coefficient of variance (CV) and PCA. The PCA result indicated that the three biological replicates of the F. verticillioides and MK groups showed a high confidence level in proteome (Figure 3a). CV analysis between the F. verticillioides and MK groups showed that their CV cumulative percentages were 54.29% and 49.95%, respectively, with the CV threshold value set at 20% (Figure S5c,e). Comparatively, the F. verticillioides group had better repeatability, which is consistent with the PCA results. For a given protein, the greater the number of peptides that support the protein, the higher the confidence of the protein identification. Therefore, the coverage of the protein indirectly reflects the overall accuracy of the identification results. Proteins with a coverage of ≥20% accounted for 37.34% of the total protein, and the average protein with an identified coverage was 19.20%, which suggested high confidence (Figure S5d). Among the 3969 proteins quantified by iTRAQ coupled with LC–MS/MS analysis, 3703, 2490, and 2104 proteins were annotated by GO, COG, and KEGG, respectively (Figure 3b). In the relative quantitative results, 387 among 3963 proteins were identified as differentially expressed proteins (DEPs) (p < 0.05, fold change > 2.0 for up‐regulated proteins or fold change < 0.50 for down‐regulated proteins), including 274 up‐regulated and 113 down‐regulated proteins, which are presented in a histogram and a volcano plot (Figure 3c,d). Furthermore, we performed a cluster analysis to identify similarities and differences of DEPs, with the colour denoting the change in the protein abundance (Figure S5f).

FIGURE 3.

Result overview of differential proteomics based on liquid chromatography–tandem mass spectrometry analysis. (a) Principal component analysis based on protein expression. (b) The statistical plot of proteins annotated by different databases. The abscissa represents the different annotation databases and the ordinate represents the number of proteins. (c) Number of up‐ or down‐accumulated protein species between the Fusarium verticillioides‐inoculated and mock‐inoculated (MK) groups. (d) Volcano plots of protein abundance between the F. verticillioides and MK groups. (e) The top 20 significant biological process terms in GO enrichment analysis at p < 0.05. (f) The top 20 significant pathways in KEGG enrichment analysis at p < 0.05. Pathways relevant to our study are marked yellow.

Moreover, we performed functional annotation with GO and KEGG to analyse the function of the DEPs. We found that up‐ and down‐regulated DEPs were significantly different in GO functional classification and functional concentration. In the biological process classification, more than 20% of the DEPs were annotated in metabolic (GO:0008152) and cellular (GO:0009987) processes. Correspondingly, in the cellular component classification, DEPs were annotated as cell (GO:0005623), cell part (GO:0044464), and organelle (GO:0043226). Catalytic activity (GO:0003824) and binding (GO:0005488) accounted for the highest proportions of DEPs in the molecular function (MF) category (Figure S5g). The top 10 pathways of F. verticillioides:MK up‐ and down‐regulated DEPs are shown in pie charts in Figure S6a. Six of the top 10 pathways were identical: metabolic pathways (ko01100), biosynthesis of secondary metabolites (ko01110), phenylpropanoid biosynthesis (ko00940), glutathione metabolism (ko00480), microbial metabolism in diverse environments (ko01120), and starch and sucrose metabolism (ko00500) (Figure S6a).

We performed enrichment analysis separately on these GO and KEGG entries to ascertain which entries were significantly enriched. The enrichment charts showed the top 20 GO terms and KEGG pathways with the most reliable enrichment significance (the lowest p value). The GO enrichment analysis presented three categories: biological process, cellular component, and molecular function. The entries with the largest number of DEPs in the biological process category were metabolic process (315 DEPs), cytoplasm (287 DEPs), and chitinase activity (274 DEPs) (Figures 3e and S6b,c). KEGG enrichment analysis showed that some important pathways of interest were significantly enriched, such as phenylpropanoid biosynthesis (40 DEPs), amino sugar and nucleotide sugar metabolism (14 DEPs), benzoxazinoid biosynthesis (13 DEPs), phenylalanine, tyrosine, and tryptophan biosynthesis (nine DEPs), flavone and flavonol biosynthesis (seven DEPs), and linoleic acid metabolism (six DEPs) (Figure 3f).

2.4. Comparative analysis of transcriptomic and proteomic data to identify DEPs and DEGs associated with candidate pathways

To gain deeper insights into the disease resistance mechanism of maize, a combined transcriptome and proteome analysis was used to identify the key genes and proteins of maize during F. verticillioides infection. In this study, the reference libraries of the transcriptome and the proteome were compared by BLAST alignment. A total of 3652 genes or proteins were matched in the combined transcriptome and proteome analysis, accounting for 11.09% and 92.01%, respectively. Moreover, integration of the proteome and transcriptome data showed that 74 DEPs were matched with their DEGs. In addition, 57 (49 up‐ and eight down‐regulated) DEPs showed the same tendency as DEGs (Figure 4a). The fold‐changes in DEPs indicated differentially positive correlations with their corresponding DEGs based on Pearson's correlation tests. A limited correlation (r = 0.078) was detected between the proteome and transcriptome, and a relatively higher positive correlation was identified with the same trend (r = 0.766) for DEGs and DEPs (Figure 4b,c).

FIGURE 4.

Correlations between mRNA and protein expression. (a) Venn diagram of genes quantified in the transcriptome (blue) and proteome (orange), differentially expressed genes (DEGs, green) and differentially expressed proteins (DEPs, pink) in Fv_MK. (b) Scatterplot of the relationship between genes identified in both the transcriptome and proteome in Fv_MK. (c) Scatterplot and correlation coefficients between DEGs and DEPs (the same trend) in Fv_MK. (d) Boxplots of KEGG enrichment associations. The x axis represents the log₂(fold change) of the gene difference at the protein/mRNA level and the y axis represents the pathway name. (e) Combined enrichment of proteome and transcriptome pathways in Fv_MK. The blue bars represent the number of DEGs and the orange bars represent the number of DEPs. (f) In the pathways identified by the combined transcriptome and proteome analysis, both genes and proteins were significantly differentially clustered as heatmaps. The colour in the figure indicates the level of gene and protein log₂(fold change): red, up‐regulation; blue, down‐regulation. The genes/proteins marked in red and marked * in the figure are studied further in this paper.

Combined analysis of the proteome and the transcriptome in F. verticillioides versus MK identified enrichment in linoleic acid metabolism, amino sugar and nucleotide sugar metabolism, phenylpropanoid biosynthesis, starch and sucrose metabolism, and glutathione metabolism (Figure 4d,e). The enrichment analysis result indicated that these pathways play an important role in plant resistance to fungal pathogen infestation. Based on literature reports and previous studies, we selected linoleic acid metabolism and amino sugar and nucleotide sugar metabolism as key pathways for further research. The trend of DEGs and DEPs in these two pathways was consistent, and both were significantly up‐regulated (Figure 4f and Table S5).

2.5. Validation of data reliability through RT‐qPCR and PRM

The RT‐qPCR experiments were performed for 14 selected DEGs associated with the response to F. verticillioides (Table S5). In the F. verticillioides and MK groups, these DEGs include LOX genes (Zm00001d042541, ZmLOX1; Zm00001d042540, ZmLOX2; Zm00001d033623, ZmLOX3; Zm00001d053675, ZmLOX10) related to steroid receptor modulators, disease resistance index genes, and pathogenesis‐related genes such as PRP6 and PRP9 (Zm00001d028816 and Zm00001d023811), and other resistance‐related genes such as Zm00001d035038 (ZmFRK2, metabolism‐related gene), Zm00001d048021 (ZmAOS1, allene oxide synthase), Zm00001d042143 (ZmGEB1, glucan endo‐1,3‐β‐glucosidase homologue 1), Zm00001d045431 (ZmENO1, enolase 1), Zm00001d008548 (ZmWIP1, wound‐ or infection‐induced gene), Zm00001d007718 (ZmBX13, benzoxazinone synthesis 13), Zm00001d003190 (ZmCTA1, resistance to chitin‐containing fungi), and Zm00001d038703 (ZmZRP4, O‐methyltransferase). The expression of 11 of these 14 candidate genes, namely ZmLOX1, ZmLOX2, ZmLOX3, ZmPRP6, ZmPRP9, ZmFRK2, ZmAOS1, ZmENO1, ZmWIP1, ZmCTA, and ZmZRP4, showed a common upward trend; three of the 14 genes (ZmLOX10, ZmGEB1, and ZmBX13) showed a common decreasing trend with the corresponding RNA‐Seq data (Figure 5a).

FIGURE 5.

Validation of RNA‐Seq by reverse transcription‐quantitative PCR (RT‐qPCR) and validation of iTRAQ data by parallel reaction monitoring (PRM). (a) Concordance analysis of RNA‐Seq and RT‐qPCR results. The blue bars represent the RNA‐Seq results and the pink bars indicate the RT‐qPCR results. Error bars indicate the standard deviation of three biological replicates. (b) Concordance analysis of proteomic and PRM results. The pink bars represent the iTRAQ results and the blue bars indicate the PRM results. The horizontal axis shows the gene/protein names and the vertical axis represents log₂ (Fusarium verticillioides/mock‐inoculated).

In addition, the DEPs corresponding to the above 14 genes were verified by PRM. The PRM quantitative results (Figure 5b) showed that the expression trends of the proteins corresponding to these 14 genes were consistent in the iTRAQ data and PRM verification results, indicating that the data captured by the iTRAQ technique coupled with LC–MS/MS were accountable and reliable. The RNA‐Seq and iTRAQ data were verified by RT‐qPCR and PRM, respectively, and the results showed that they were highly consistent, indicating that the obtained results were accurate and reliable.

2.6. Transient silencing of ZmLOX2 facilitates F. verticillioides infection in maize plants

To understand the potential roles of ZmLOX2 in regulating the maize defence response, the development of typical maize stalk rot symptoms in the roots and the stalks of ZmLOX2 transiently silenced plants was examined. An improved BMV‐based silencing vector (Ding et al., 2018) was used to knock down ZmLOX2 expression through VIGS. The ZmLOX2 VIGS vector BMV‐LOX2 and control plasmid BMV‐GFP (green fluorescent protein) were introduced into Nicotiana benthamiana leaves to propagate the BMV. Maize cv. Va35 seedlings were inoculated with partially purified BMV particles to transiently silence ZmLOX2 as done previously (Cao et al., 2012; Zhu et al., 2014). The RT‐qPCR data showed that the expression of ZmLOX2 in the BMV‐ZmLOX2 inoculated plants was about 20.5%–27% that of the control plants (Figure 6a). In the ZmLOX2 knocked‐down plants, the expression of ZmLOX1, ZmLOX4, and ZmLOX5, which share 82.4%, 64.52% and 63.52% homology with ZmLOX2 (Figure S9), was not affected (Figure 6a). These results demonstrated that silencing of ZmLOX2 was successfully achieved.

FIGURE 6.

Transient silencing of ZmLOX2 facilitates Fusarium verticillioides infection in maize plants. (a) Silencing efficiency and specificity of ZmLOX2 in the BMV‐LOX2 inoculated seedlings were evaluated by measuring the transcript levels of ZmLOX2 and its homologous genes ZmLOX1, ZmLOX4, and ZmLOX5 at 7 days postinoculation (dpi) and 10 dpi. (b) At 7 dpi, the BMV‐GFP and BMV‐LOX2 pre‐inoculated cv. Va35 seedlings were challenged with F. verticillioides and the inoculated seedling roots were examined for symptoms at 48 h postinoculation (hpi). Inoculation zone indicated by red lines, and magnified in the white dashed box. (c) The measurement of stalk rot score on average (SRS_A), stalk rot disease severity index (DSI) of F. verticillioides‐inoculated maize seedling roots. (d) At 10 dpi, the stalks of BMV‐GFP and BMV‐LOX2 pre‐inoculated Va35 plants were challenged with F. verticillioides. The symptoms in the stalks were recorded in the split internodes at 72 h after inoculation with F. verticillioides. (e) Measurement of SRS_A and DSI for F. verticillioides‐inoculated maize stalks. Scale bars: (b) 5 cm, (d) 2 cm. Values shown are mean ± SD. *p < 0.05, **p < 0.01, ***p < 0.001 (according to a paired Student's t test). NS, not significant; Fv, F. verticillioides.

At 7 dpi, the root tips of BMV‐GFP‐ and BMV‐LOX2‐preinoculated maize Va35 seedlings were challenged with F. verticillioides. The fungal infection in the roots of these seedlings was analysed by measuring the development of symptoms, the stalk rot score on average (SRS_A) and disease severity index (DSI). The infection of F. verticillioides caused more severe symptoms in the ZmLOX2‐silenced roots than those seen in the control plants at 48 hpi (Figure 6b). The SRS_A and DSI of the roots were consistent with the symptom development: silencing of ZmLOX2 in Va35 seedlings through VIGS caused about a 1.5‐fold change of SRS_A, and a 25.30% increase of DSI compared with that seen in the controls (Figure 6c). A parallel test was also conducted to analyse the effect of ZmLOX2 on F. verticillioides infection inside maize stalks. The stalks of the ZmLOX2‐silenced plants were artificially challenged with F. verticillioides at 10 dpi, and the internodes of the inoculated internodes were split to observe the stalk symptoms at 48 hpi, similar to the test of progress of the disease in the seedling roots. Transient silencing of ZmLOX2 also facilitated the infection of F. verticillioides inside maize stalks (Figure 6d,e).

2.7. ZmLOX2 regulates the maize defence response to F. verticillioides infection through the SA‐mediated pathway

Plant hormones such as JA and SA are known to act in the defence against fungal infection. RNA‐Seq and iTRAQ‐based proteomic analysis showed that multiple genes in SA pathways were activated (Figure 4e). 9‐LOXs mediate a defence response through the JA‐mediated pathway (Battilani et al., 2018; Christensen et al., 2015). To explore whether SA or JA triggered ZmLOX2‐mediated defence against F. verticillioides infection, maize seedlings at the three‐leaf stage were used to determine the expression profiles of ZmLOX2 under SA or JA stress. Exogenous treatment with SA drastically induced ZmLOX2 in maize leaves, whereas ZmLOX2 transcripts were weakly induced in leaves treated with methyl jasmonate (MeJA) (Figure 7a). The enhanced susceptibility of ZmLOX2‐silenced maize Va35 seedlings to fungal infection led us to evaluate whether ZmLOX2 silencing affects the expression of maize defence‐related genes. The transcriptional level of PR genes ZmPR3, ZmPR4, and ZmPR5, which are also SA‐responsive marker genes, was analysed by RT‐qPCR. The data indicated that the expression levels of ZmPR3, ZmPR4, and ZmPR5 were reduced by approximately 76%, 90%, and 67%, respectively, in the ZmLOX2‐silenced seedlings compared with the BMV‐GFP controls (Figure 7b). We also measured the H₂O₂ content in the ZmLOX2‐silenced leaves. As expected, lower levels of H₂O₂ were detected during F. verticillioides infection in ZmLOX2‐silenced leaves compared with BMV‐GFP control leaves (Figure 7c). SA analysis using high‐performance liquid chromatography (HPLC) revealed that lower free and total SA (free SA plus glucosyl‐conjugated SA) levels accumulated in ZmLOX2‐silenced seedlings inoculated with F. verticillioides (Figure 7d). In addition, the induction of the SA pathway and related physiological and biochemical characteristics, such as the activity of CAT, POD, PAL, and lignin content (Chen et al., 1993; Dempsey et al., 2011; Kimura & Kawano, 2015), were tested. The data indicated that silencing of ZmLOX2 impaired the increase in these characteristics during fungal infection (Figure S8). Although JA levels were obviously induced during the F. verticillioides infection process, no significant differences in JA levels were observed between ZmLOX2‐silenced seedlings and the controls inoculated with BMV‐GFP vector during infection by F. verticillioides (Figure 7e). These data indicate that silencing of ZmLOX2 attenuates the basal defence response of maize plants to F. verticillioides infection.

FIGURE 7.

ZmLOX2 regulates the maize defence response to Fusarium verticillioides infection through the salicylic acid (SA)‐mediated pathway. (a) Relative expression of ZmLOX2 after exogenous treatment with methyl jasmonate (MeJA) and SA. Ctl., control. (b) The expression profiles of the pathogenesis‐related (PR) genes ZmPR3, ZmPR4, and ZmPR5 in the ZmLOX2‐silenced plants. (c) Quantification of the H₂O₂ level in maize Va35 leaves of ZmLOX2‐GFP control plants and ZmLOX2‐silenced plants at 24 and 48 h postinoculation (hpi) with F. verticillioides. Levels of SA (d) and jasmonic acid (JA) (e) in the controls and ZmLOX2‐silenced leaves infected by F. verticillioides. (f) Lipoxygenase (LOX) activity in maize leaves of unsilenced and ZmLOX2‐silenced plants inoculated with F. verticillioides at 24 and 48 hpi. (g) Quantification of lipid peroxidation (malondialdehyde [MDA] concentration) in leaves of empty vector control and silenced maize plants after F. verticillioides infection. (h) Pretreatment of the ZmLOX2‐silenced plants with SA‐enhanced maize resistance to F. verticillioides fungal infection. Values shown are mean ± SD. Different letters on the columns indicate a significant difference (p < 0.05) by the Tukey–Kramer test. *p < 0.05, **p < 0.01, ***p < 0.001 (according to a paired Student's t test). ns, not significant; MK, mock‐inoculated; Fv, F. verticillioides.

To test whether ZmLOX2 activity is induced by fungal infection, LOX activity in protein extracts of ZmLOX2‐silenced seedlings infected with F. verticillioides were assayed. The unsilenced plants exhibited higher levels of LOX activity during F. verticillioides infection compared with the mock‐inoculated controls, but LOX activity was not significantly enhanced in ZmLOX2‐silenced plants by F. verticillioides infection (Figure 7f), indicating that transient silencing of ZmLOX2 compromises LOX activity. In addition, lipid peroxidation was also evaluated by determining the levels of malondialdehyde (MDA), which is a decomposition product formed by peroxidation of polyunsaturated fatty acids in membranes during infection. It was found that MDA concentration was significantly enhanced in maize seedlings during the infection establishment, and lower levels of MDA were detected in ZmLOX2‐silenced seedlings than in the controls (Figure 7g).

The involvement of SA in the process of ZmLOX2‐mediated defence response was further verified by assessing whether exogenous treatment of ZmLOX2‐silenced seedlings with SA could enhance resistance to F. verticillioides infection. The maize Va35 seedlings were sprayed with SA or MeJA for 12 h, then inoculated with F. verticillioides mycelial inocula. In the nonsilenced seedlings, exogenous application with SA and MeJA enhanced maize resistance to F. verticillioides infection, whereas for the ZmLOX2 transiently silenced plants, those pretreated with SA but not with MeJA displayed enhanced disease resistance (Figure 7f). Collectively, these results indicate that ZmLOX2 regulates the maize response to F. verticillioides infection through the SA‐triggered defence pathway.

3. DISCUSSION

In this study, quantitative transcriptomic and iTRAQ‐based proteomic analyses were employed to determine the different responses of maize plants to F. verticillioides infection. Based on the transcriptomes, we found that a large number of disease resistance‐related genes were differentially expressed, although they often showed low level expression (Figure 4b). Due to the low efficiency of proteomics technology in detecting low‐abundance proteins (Liu et al., 2016; Vogel & Marcotte, 2012), the number of DEPs identified in the proteome was relatively small, only 387. On the plus side, most DEPs were enriched in the pathways of interest, such as phenylpropanoid biosynthesis, benzoxazinoid biosynthesis, flavone and flavonol biosynthesis, and linoleic acid metabolism (Figure 3f). It is undeniable that high‐quality data to quantify protein expression are important to understand the stress response (Liu et al., 2016). DEGs and DEPs were mainly assigned to glutathione metabolism, starch and sucrose metabolism, amino sugar and nucleotide sugar metabolism, linoleic acid metabolism, and phenylpropanoid biosynthesis (Figure 4d). Fourteen pathogenesis‐related proteins and genes from KEGG and GO annotation were validated with RT‐qPCR and PRM coupled with LC‐MS/MS analysis. These genes and proteins showed comparable trends to RNA‐Seq and iTRAQ (Figure 5a,b). The findings provide a solid foundation for identification of candidate resistance genes by analysing the underlying proteomic regulation in host–pathogen interactions.

Oxylipins have been implicated as a newly emerging group of signals that have defence roles or promote virulence in plants (Battilani et al., 2018). Among these, JA is a well‐known lipoxygenase (LOX)‐derived oxylipin (Andreou & Feussner, 2009; Borrego & Kolomiets, 2016). JA signalling regulates diverse physiological processes and is a key regulator of the defence response to pathogen infection via induction of defensive metabolites and proteolytic enzymes (Robert‐Seilaniantz et al., 2011). ZmLOX3 has been proved to be a major susceptibility factor, negatively regulating JA biosynthesis in diverse maize organs to facilitate fungal infection (Gao et al., 2008, 2009). ZmLOX4, ZmLOX5, and ZmLOX12 are required for normal JA biosynthesis during maize defence against F. verticillioides seed infection (Battilani et al., 2018; Christensen et al., 2014; Lanubile et al., 2021). In contrast, ZmLOX1 and ZmLOX2 are not involved in seed defence against F. verticillioides infection (Battilani et al., 2018). JA usually shows antagonism with another important defence phytohormone, SA, which typically induces resistance to hemibiotrophic pathogens (Glazebrook, 2005). By using the maize mutant lox10 and the JA‐deficient double mutant opr7‐5 opr8‐2, previous research indicated that green leaf volatiles increase susceptibility of maize to Colletotrichum graminicola through JA‐dependent suppression of SA shortly after inoculation (Gorman et al., 2020). Here, we report that transient silencing of ZmLOX2 facilitated F. verticillioides infection inside maize stalks and roots (Figure 6). The activity was reduced in ZmLOX2‐silenced plants (Figure 7f) and was further validated by lower levels of lipid peroxidation increase in the ZmLOX2‐silencing seedlings compared with those observed in the nonsilenced controls after fungal infection (Figure 7g). This suggests that maize seedlings defend themselves against F. verticillioides infection by activating the LOX pathway.

Previous studies documented that LOX expression is regulated by JA during infection with fungal pathogens (Melan et al., 1993). ZmLOX2 transcript was weakly induced in leaves pretreated with MeJA (Figure 7a), and although the JA levels were drastically induced after fungal infection, no differences in JA levels were observed between the ZmLOX2‐silenced and nonsilenced seedlings (Figure 7e). It could be speculated that ZmLOX2 is not involved in the JA‐mediated maize disease resistance response to F. verticillioides infection.

LOXs catalyse the conversion of linoleic acid into hydroperoxides, which are in turn converted to oxylipins. These primary products may cause oxidative damage to plant membranes during the HR (Slusarenko, 1996). In addition to the lower levels of peroxidation, silencing of ZmLOX2 was also accompanied by reduced expression of defence‐related genes such as the SA‐responsive ZmPRs (Figure 7b), as well as lowered ROS (Figure 7c) and decreased SA accumulation (Figure 7d). These results suggest that lipid peroxidation and ROS and SA accumulation in maize seedlings are crucial for the execution of defence responses induced by ZmLOX2. The exogenous pretreatment with SA but not MeJA restored disease resistance in the ZmLOX2‐silenced plants (Figure 7h). All these data demonstrate that ZmLOX2 shows a remarkable defence response to F. verticillioides infection via the SA‐dependent signalling pathway.

Maize chitinase genes have been proposed to be involved in maize resistance to F. verticillioides infection in maize kernels (Shu et al., 2017). The maize WIP1 gene encodes a wound‐induced Bowman–Birk inhibitor (BBI) protein, which is a type of serine protease inhibitor, and its expression is induced by wounding or infection, conferring resistance against pathogens (Zhang et al., 2013). In this study, maize chitinase 1 gene (ZmCTA1) and ZmWIP1 were found to be regulated in maize plants after F. verticillioides infection through bioinformatics analysis and experimental validation (Figure 5). Their potential roles in defence against F. verticillioides infection were supported by the enhanced susceptibility to F. verticillioides in the ZmCTA1‐ and ZmWIP1‐silenced plants (Figure S7).

In this work, we speculated that ZmLOX2 regulates maize defence to F. verticillioides infection through SA‐triggered signalling pathway. The mechanism by which this occurs remains obscure. Thus, the identification of the partners that directly interact with ZmLOX2 will help to elucidate the underlying mechanism of its function.

4. EXPERIMENTAL PROCEDURES

4.1. Plant materials and fungal infection

F. verticillioides strain HNYY‐03‐1 was preserved and propagated in our laboratory and cultured on fresh potato dextrose agar (PDA) plates (c.15 ml per plate) at 25°C in darkness for 5–7 days. The inoculum was obtained by the homogenization of five plates of hyphal mats (approximately 125 ml) with a kitchen blender, adjusting to a final volume of 200 mL with sterile double‐deionized water (ddH₂O).

Maize inbred line Chang7‐2 (CH7‐2) with moderate resistance to F. verticillioides was grown in a nursery pot in the greenhouse. The greenhouse was operated on a 16 h:8 h light:dark photoperiod with daytime temperature 28°C and night‐time temperature 22°C. The light intensity was 230 μmol m⁻² s⁻¹ at a height of 107 cm. The soil was irrigated with water or Peters Excel Cal‐Mag Special 15–5‐15 water‐soluble fertilizer supplemented with calcium and magnesium (Hummert cat. 07–5660‐1) to keep it reasonably wet. The maize plant inoculation was based on previously described protocols (Zhang et al., 2016) with minor modifications. Seedlings at the nine‐leaf stage were selected for uniform performance. Maize plants were inoculated by punching a hole in the stem at the second or third internode above the soil line using a sterile micropipette tip (10 mm hole depth), followed by injection of 50 μL of freshly prepared F. verticillioides inoculum. A similar number of plants were inoculated with PDA as a mock treatment. The wounds were sealed with Vaseline after inoculation. The upper and lower stem segments immediately adjacent to the inoculation segments (Figure 1) were sampled at 3 dpi and each sample was replicated three times. All individual samples were frozen in liquid nitrogen and stored at −80°C for further proteome and transcriptome analysis. For the inoculation of seedling roots, the root tips were challenged with 20 μL of the mycelial plug inoculum. Mock‐inoculated maize roots treated with a PDA plug to serve as the control.

4.2. Disease symptoms evaluation

The roots of the inoculated seedlings were scored at the disease region at 24 or 48 hpi with a rating grade of 1–5 as previously described (Ye et al., 2019). For the evaluation of stalk rot symptoms, from 24 h to 2 weeks after inoculation, the inoculated internodes of the individual maize plants were split and symptoms observed with scores of 0, 1, 3, 5, 7, and 9 according to the previously described classification standard (Duan et al., 2022).

$${\mathrm{SRS}}_A={\sum }_1^n\mathrm{SRS}/n$$

$$\begin{array}{cc}\mathrm{DSI}\left(\%\right)=\hfill & \sum \left(\text{grade}\mathbf{\times }\text{number of plants in grade}\right)\\ & \mathbf{\times }100/\left(5\mathbf{\times }\text{total number of plants}\right).\hfill \end{array}$$

4.3. Physiological and biochemical characteristics measurement

To assay the physiological and biochemical characters related to the plant defence reaction after fungal infection, the roots and leaves of the seedlings were harvested at different time points after inoculation. The physiological and biochemical characteristics, including dehydrogenase (DHA) activity (μg/g), lipid peroxidation malondialdehyde (MDA) level (mmol/g), phenylalanine ammonia‐lyase (PAL) activity (U/g), peroxidase (POD) activity (U/g), polyphenol oxidase (PPO) activity (U/g), catalase (CAT) activity (U/g), superoxide dismutase (SOD) activity (U/g), soluble saccharide concentration (mg/g), total protein concentration with BCA method (mg/g), lignin content (mg/g fresh weight), and chlorophyll content (mg/g) were measured using assay kits form Comin Biotechnology Co., Ltd following the manufacture's protocols.

4.4. RNA isolation, library construction, and sequencing

Total RNA was isolated and purified using TRIzol reagent (Invitrogen) following the manufacturer's procedure. The RNA amount and purity of each sample was quantified using a ND‐1000 spectrometer (NanoDrop). The RNA integrity was assessed by Bioanalyzer 2100 (Agilent) with RIN >7, and confirmed by electrophoresis with a denaturing agarose gel. Three micrograms of RNA were used as input material for the RNA sample preparations. Sequencing libraries were generated according to the following steps. First, mRNA was purified from total RNA using poly‐T oligo‐attached magnetic beads. Fragmentation was carried out using divalent cations under elevated temperature in fragmentation buffer. The first cDNA strand was synthesized by random hexamers, and the second cDNA strand was synthesized by adding buffer, dNTPs, RNase H and DNA polymerase I. Double‐stranded cDNA was repaired and A was added to the 3′ end. Hieff NGS DNA selection beads were used for purification and fragment selection. After purification and fragment selection, the products were amplified and enriched by PCR. The target region library of the duplex was denatured, cycled, and digested to yield a single‐stranded circular DNA. Single‐stranded circular DNA was amplified by rolling circle amplification (RCA), known as DNA nano balls (DNB). After library construction, the library was quantified using the Qubit method, then the library was sequenced on a DNBSEQ‐T7 genetic sequencer using PE150 sequencing.

4.5. Identification of DEGs and enrichment analysis

Samples were sequenced on the platform to obtain image files, which were transformed by sequencing platform software, and the original data in FASTQ format (raw data) were generated. The sequencing data contained a number of connectors, low‐quality reads, so we used fastp (v. 0.21.0) (Chen et al., 2018) software to filter the sequencing data to get a high‐quality sequence (clean data) for further analysis. The reference genome and gene annotation files were downloaded from the genome website. The filtered reads were mapping to the reference genome using HISAT2 v. 2.1.0 (Pertea et al., 2016).

We used StringTie v. 2.1.5 (Pertea et al., 2015) statistics to compare the read count values on each gene as the original expression of the gene, and then used FPKM (Mortazavi et al., 2008) to standardize the expression. The differential expression of genes was analysed by DESeq2 v. 1.30.1 (Love et al., 2014) with screened conditions as follows: expression difference multiple log₂ fold change ≥ |1|, significant p _adj < 0.05. We visualized the results as a heatmap using the pheatmap v. 1.0.8 R package. We mapped all the genes to terms in the GO database (Ashburner et al., 2000) and calculated the numbers of differentially enriched genes in each term. Using cluster Profiler to perform GO enrichment analysis on the differential genes, the p value was calculated by the hypergeometric distribution method (the standard of significant enrichment was p _adj < 0.05). The main biological functions performed by significantly enriched genes were determined using the GO terms. Cluster Profiler v. 3.18.1 (Yu et al., 2012) software was used to carry out the enrichment analysis of the KEGG pathway of differential genes, focusing on the significantly enriched pathways with p _adj < 0.05.

4.6. Protein extraction and trypsin digestion

Samples were first ground to powder in liquid nitrogen and incubated in lysis buffer (7 M urea, 2 M thiourea, 4% SDS, 40 mM Tris–HCl, pH 8.5) containing 1 mM phenylmethylsulfonyl fluoride and 2 mM EDTA (final concentration) for 5 min, then 10 mM dithiothreitol (DTT, final concentration) was added to the sample. The suspension was sonicated for 5–15 min on ice and then centrifuged at 4°C, 20,817 × g for 20 min. The supernatant was mixed with 4 volumes of precooled acetone at −20°C for 2 h. After centrifugation, the protein pellets were air‐dried and resuspended in 8 M urea/100 mM triethylammonium hydrogen carbonate buffer (TEAB, pH 8.0). Protein samples were reduced with 10 mM DTT at 56°C for 30 min and alkylated with 50 mM iodoacetamide at room temperature for 30 min in the dark, then 4 volumes of precooled acetone at −20°C were added over 2 h. After centrifugation, the protein pellets were air‐dried and resuspended in 8 M urea/100 mM TEAB (pH 8.0). The total protein concentration was measured using the Bradford method. Equal amounts of protein from each sample (c.100 μg) were used for tryptic digestion. Trypsin was added at an enzyme:protein ratio of 1:50 (wt/wt), and the digestion was performed at 37°C for 12–16 h. After digestion, peptides were desalted using C18 columns and the desalted peptides were dried with a vacuum concentrator. The dried peptide power was redissolved to 20 μL with 0.5 M TEAB for peptide labelling.

4.7. iTRAQ labelling and fractionation

The peptide sample was dissolved in 20 μL of dissolution buffer (0.5 M TEAB), 70 μL of isopropanol was added then the mixture was centrifuged with shaking. Samples were labelled with iTRAQ Reagent‐8 plex Multiplex Kit (AB Sciex) according to the manufacturer's instructions. Samples were iTRAQ labelled as following: MK1, 113; MK2, 114; MK3, 115; Fv1, 116; Fv2, 117; Fv3, 118. All of the labelled samples were mixed in equal amounts. Next, the labelled samples were fractionated using an HPLC system (Thermo DINOEX Ultimate 3000 BioRS) using a Welch C18 column (5 μm, 100 A, 4.6 × 250 mm) at high pH conditions. Finally, the collected fractions were combined into 15 fractions. The combined components were desalted on a Strata‐X column and dried in vacuo.

4.8. LC–MS/MS and bioinformatics analysis

The peptide samples were dissolved in 2% acetonitrile/0.1% formic acid and analysed using a TripleTOF 5600plus mass spectrometer coupled with the Eksigent microLC system (AB Sciex). Peptides were loaded onto a C18 trap column (5 μm, 100 μm × 20 mm) and eluted at 5 μL min⁻¹ onto a C18 analytical column (3 μm, 300 μm × 150 mm) over a 60 min gradient. The 60 min solvent gradient was 5% B, 0 min; 5%–27% B, 45 min; 27%–50% B, 3 min; 50%–80% B, 2 min; 80% B, 6 min; 80%–5%, 0.5 min; 5%, 3.5 min. The two mobile phases were buffer A (2% acetonitrile/0.1% formic acid/98% water) and buffer B (98% acetonitrile/0.1% formic acid/2% water). For information‐dependent acquisition (IDA), survey scans were acquired in 250 ms and 35 product ion scans were collected at 50 ms per scan. MS1 spectra were collected in the range 350–1500 m/z, and MS2 spectra were collected in the range 100–1500 m/z. Precursor ions were excluded from reselection for 10 s. In the IDA advanced tab, the option “Adjust CE when using iTRAQ reagent” was selected for iTRAQ samples. The resulting MS/MS data were processed using the Proteinpilot v. 4.5 (AB Sciex) search engine with default parameters. Functional annotation of the significantly different proteins was completed by accessing the GO and KEGG databases (Ashburner et al., 2000; Kanehisa et al., 2016).

4.9. Integrated analysis of transcriptome and proteome

This study associated the transcriptome with the proteome reference library by BLAST alignment. The correlations between mRNA and protein expression levels in the F. verticillioides and MK groups were estimated by Pearson correlation analysis. The transcriptomic and proteomic data were converted with log₂ to the average fold change and displayed as scatter plots using the R program, with the selection criteria: log₂(fold change) ≥ |1| and p < 0.05 for mRNA and proteins.

4.10. RT‐qPCR

Total RNA was purified from different samples using TRIzol reagent (Invitrogen), then treated with RNase‐free DNase I (TaKaRa). The first‐strand cDNA was synthesized using 2 μg of total RNA per 20 μL of reaction mixture and an oligo(dT) primer. Ten‐fold diluted cDNA, a set of gene‐specific primers (Table S6) and TB Green Premix Ex Taq II (TaKaRa) were mixed for real‐time qPCR to determine the accumulation levels of the maize genes on the CFX96 real‐time PCR detection system (Bio‐Rad). The expression level of GAPDH mRNA was determined and used as an internal control. The relative expression level of each gene was calculated using the 2^−ΔΔCt method (Livak & Schmittgen, 2001). Differences between the treatments were then analysed using Student's t tests. All experiments were carried out at least three times.

4.11. PRM validation

To verify the protein expression levels obtained by iTRAQ analysis, the expression levels of selected proteins were further quantified by PRM analysis. Peptide fragmentation and targeted PRM mass spectrometry were conducted with a TripleTOF 5600plus LC–MS equipped with the Eksigent nanoLC system (Sciex). The two mobile phases were aqueous phase buffer A (2% acetonitrile/0.1% formic acid/98% water) and organic phase buffer B (98% acetonitrile/0.1% formic acid/2% water). For IDA detection, a first‐order mass spectrum was scanned with an ion accumulation time of 250 ms and a second‐order mass spectrum of 30 precursor ions was collected with an ion accumulation time of 100 ms. MS1 acquisition was performed in the range 350–1500 m/z and MS2 acquisition was performed in the range 100–1500 m/z. The precursor ion dynamic exclusion time was set to 15 s. The list of 15 target peptides was added to the inclusion list for PRM data collection. The peptides were selected according to size and fragmented for secondary ion spectra collection. MS raw data captured from the target proteomics analysis were examined against the Proteinpilot database and a spectra library created using Skyline v. 3.1.0 (MacLean et al., 2010) with a cut‐off score of 0.90. The precursor m/z of the target peptides was sequentially determined in agreement with the inclusion list for secondary mass spectrometry analysis (MS2), in which MS2 was settled. Each sample was dissolved in loading buffer (0.1% formic acid/3% acetonitrile) for centrifugation, and the supernatant was used for loading to capture six PRM spectrum files for protein quantification.

4.12. VIGS

An improved BMV‐derived VIGS vector (Ding et al., 2018) was used to knockdown the target genes. Sequence‐specific DNA fragments representing the target genes were amplified using specific primer pairs (Table S6) to construct the relevant VIGS vector. To evaluate the genetic stability of foreign inserts in the VIGS vector during RNA silencing in plants, we predicted the folding structure formed by the full‐length BMV RNA3 sequence in pF3‐13 m of pBMV‐CP5 vector or with gene fragment inserts using a deep learning‐based RNA secondary structure prediction tool MXFold2 (Sato et al., 2021; http://www.dna.bio.keio.ac.jp/mxfold2/). Only those vectors harbouring foreign gene fragment inserts without changing the stability of BMV RNA3 were chosen for subsequent analysis (Figure S10). The resulting constructs BMV‐LOX2, BMV‐CTA1, BMV‐WIP1, and the control BMV‐GFP were then transformed into Agrobacterium tumefaciens C58C1. N. benthamiana leaves were infiltrated with A. tumefaciens cultures and collected for virion preparation, as described previously (Zhu et al., 2014).

The third leaves of three‐leaf stage Va35 plants were rub‐inoculated with approximately 20 μg of partially purified virions. More than 20 seedlings were used for each treatment, and the inoculated plants were grown inside a growth chamber at 20/18°C (day/night) for 7–10 days before being challenged with F. verticillioides. Systemically infected leaves (or equivalent leaves from mock‐inoculated plants) from BMV‐LOX2, BMV‐CTA1, BMV‐WIP1 or BMV‐GFP inoculated plants were harvested from individual plants at 7 and 10 dpi, and subjected to RT‐qPCR to evaluate the efficiency of the gene silencing.

4.13. Quantification of H₂O₂ by xylenol orange assay

H₂O₂ production in plants was spectrophotometrically measured using a xylenol orange assay (Bindschedler et al., 2001) following the modified protocol described by Hwang and Hwang (2010).

4.14. Procaryotic expression of ZmLOX2 in Escherichia coli

The ZmLOX2 gene was cloned into the in vitro expression vector pET28a (EMD Bioscience) with gene‐specific primers (Table S6). The generated plasmid was transformed into E. coli BL21. The transformant bacteria were grown in Luria–Bertani medium supplemented with kanamycin (50 mg/mL) and cultured at 37°C until OD₆₀₀ = 0.5. The ZmLOX2 protein expression was induced in the bacterial culture by IPTG. pET28a:His‐tagged ZmLOX2 was purified with 1.5 mL of nickel‐nitrilotriacetate agarose resin (Invitrogen) in a 10‐mL purification column from which the target protein was eluted with an imidazole solution (250 mM). Purified ZmLOX2 was subjected to 10% SDS‐PAGE and stained with Coomassie brilliant blue to confirm purity.

4.15. LOX activity assay

Protein extracts were prepared from the maize leaves following the previously described method (Cao et al., 2012). Protein concentrations were determined with a dye‐binding protein assay kit (Bio‐Rad) following the manufacturer's instructions. LOX enzyme activity was analysed according to the previously described protocol (Hwang & Hwang, 2010). The assay mixture contained 0.1 M sodium phosphate buffer (pH 6.5), 0.1 mM linoleic acid (Sigma), which was used as a substrate, and 5–10 μL of resuspended ZmLOX2 enzyme purified from E. coli or protein extracts from BMV‐LOX2 or BMV‐GFP control seedlings. The reaction was adjusted to a final volume of 1.2 mL. The change in the absorbance was spectrophotometrically measured by monitoring the increase in the conjugated diene hydroperoxide at A₂₃₄ at 25°C for 10 min. For each test, three different biological samples with three independent experiments were analysed. All experiments were performed three times with similar results.

4.16. Oxylipin profiling

Oxylipin profiling in the roots and leaves of the silenced and wild‐type maize plants was measured according to the previously described method (Gao et al., 2008).

4.17. SA measurements

SA and SA glycoside were extracted according to the method described by Verberne et al. (2002) and measured by ultrahigh performance liquid chromatography‐triple quadrupole mass spectrometry following the method of Pan et al. (2010) at the mass spectrometry centre of the College of Biological Sciences, China Agricultural University. There were three biological repeats for each treatment investigated.

4.18. Measurement of JA

Endogenous levels of JA and its derivative compounds in maize plants were quantified by comparing the levels of endogenous metabolites to isotopically labelled standards from Sigma‐Aldrich and Larodan AB using LC–MS/MS according to the description by Wang et al. (2020).

4.19. Chemical treatments

For treatment of maize plants with SA, three‐leaf‐stage maize plants were sprayed with 2 mM SA (Sigma‐Aldrich). SA was dissolved in 100% ethanol diluted in double‐distilled water containing 0.2% (vol/vol) Triton X‐100 as a wetting agent. Mock‐inoculated maize plants treated with 0.2% (vol/vol) Triton X‐100 were used as a control. For the treatment with MeJA, the seedlings were sprayed with 100 μM MeJA dissolved in 0.1% acetone. Plants treated with MeJA were tightly sealed in a plastic bag for 12 h. The experiment was carried out at least three times and 18 plants were used for each treatment.

4.20. Statistical analysis

All data collected from phenotypic and physiological parameters, RT‐qPCR, and PRM analyses were subjected to one‐way variance analysis (ANOVA) and Student's t test using software SPSS v. 22.0 (IBM). p < 0.05 indicated a statistically significant difference; p < 0.01 indicated a very significant difference.

CONFLICT OF INTEREST STATEMENT

The authors declare that the research was conducted without any commercial or financial relationships that could be construed as a potential conflict of interest.

Supporting information

Figure S1 Measurement of chlorophyll contents of CH7‐2 seedlings after Fusarium verticillioides inoculation.

Figure S2 Transcriptome overall expression and differential expression analysis. (a) The boxplot shows the overall range and distribution of the gene expression level with the log₁₀(FPKM +1) value in Fusarium verticillioides and the mock‐inoculated (MK) group. (b) FPKM density distribution of F. verticillioides and the MK group. The y axis corresponds to gene density and the x axis displays the log₁₀(FPKM) of groups. (c) Heatmap from hierarchical clustering of all differentially expressed genes (DEGs) in F. verticillioides and the MK group.

Figure S3 GO annotation of the differentially expressed genes (DEGs) between Fusarium verticillioides and the mock‐inoculated (MK) group. Functional annotation of DEGs for Gene Ontology (GO) categories in molecular function, biological process, and cellular component. The blue, orange, and cyan bar graphs indicate the annotation of total differential genes, up‐regulated differential genes, and down‐regulated differential genes in different GO categories, respectively. The abscissa represents the name of the GO term and the ordinate represents the number of differential genes annotated on this term.

Figure S4 KEGG annotation of the differentially expressed genes (DEGs) between Fusarium verticillioides and the mock‐inoculated (MK) group. Functional annotation of up‐ and down‐regulated DEGs for Kyoto Encyclopedia of Genes and Genomes (KEGG). The result format is similar to the Gene Ontology annotation in Figure S3.

Figure S5 Overview of proteomic results and differential expressed protein (DEP) analysis. (a) Distribution of unique peptides. The abscissa presents the unique peptide numbers of the identified proteins. The vertical ordinate at the left denotes the protein number corresponding to the abscissa value. The vertical ordinate at the right denotes the cumulative protein ratio. (b) Distribution of peptide lengths. Most peptides had a length of 12. (c) The distribution of coefficient of variance (CV) values for different samples. The abscissa is the CV value and the ordinate corresponds to the cumulative percentage of the two groups of samples at the 20% CV threshold value. (d) Coverage of identified proteins. Different colours indicate different identification coverage ranges of proteins. (e) Box plots of the CV values for the different samples. The black horizontal lines on the box represent the location of the median. (f) Heatmap of hierarchical clustering of DEPs. Red indicates high expression protein and blue indicates low expression protein. (g) Functional annotation of DEPs for the Gene Ontology database. The orange column presents the functional classification ratios of up‐regulated proteins under the categories biological progress, cellular component, and molecular function. The green column presents the functional classification ratios of down‐regulated proteins under the categories biological progress, cellular component, and molecular function.

Figure S6 Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis of differentially expressed proteins (DEPs) between Fusarium verticillioides and the mock‐inoculated (MK) group. (a) KEGG annotation results for DEPs. The percentages of the top 10 pathways related to the DEPs are shown. The number of proteins in each pathway is shown in parentheses. The top 20 significant and cell components (CCs) (b) and molecular function (MF) (c) terms in GO enrichment analysis at p < 0.05.

Figure S7 Silencing of ZmCTA1, ZmWIP1 enhanced maize susceptibility to Fusarium verticillioides. (a), (c) Silencing efficiency of ZmCTA1, ZmWIP1 in maize cv. Va35 plants was measured by reverse transcription‐quantitative PCR. (b), (d) Measurement of F. verticillioides infection in the ZmCTA1‐, ZmWIP1‐silenced plants through stalk rot disease severity index (DSI).

Figure S8 Measurement of salicylic acid (SA) pathway‐related physiological and biochemical characters in the ZmLOX2‐silenced plants after Fusarium verticillioides infection. The activities of (a) catalase (CAT), (b) peroxidase (POD), (c) phenylalanine ammonia‐lyase (PAL), and (d) lignin contents were measured 24 h and 48 h after inoculation with F. verticillioides.

Figure S9 The alignment of different ZmLOX2 orthologues. The alignment was performed using BioEdit (v. 7.2.5). The targeting sites of gene‐specific primers used for amplification of fragment inserts are indicated with red lines.

Figure S10 Deep learning‐predicted BMVCP5 (+) strand RNA3 secondary structure without and with gene fragment inserts. Full‐length (+) stands of BMV RNA3, without (a) or with gene fragment inserts, representing ZmLOX2 (b), ZmCTA1 (c), and ZmWIP1 (d) were folded with the MXFold2 Web server (http://www.dna.bio.keio.ac.jp/mxfold2/). Locations of the two cloning sites are indicated. Gene‐specific fragment inserts are shown with magenta dashed lines.

Table S1 Overview of RNA sequencing data

Table S2 RNA sequencing data and reference genome alignment results

Table S3 Screening of differential genes of plant disease resistance key pathways

Table S4 Statistical results of protein identification information

Table S5 Targeted pathways and key genes identified by combined transcriptome and protein analysis

Table S6 Primers used in this study

Zhang, L. , Hou, M. , Zhang, X. , Cao, Y. , Sun, S. , Zhu, Z. et al. (2023) Integrative transcriptome and proteome analysis reveals maize responses to Fusarium verticillioides infection inside the stalks. Molecular Plant Pathology, 24, 693–710. Available from: 10.1111/mpp.13317

DATA AVAILABILITY STATEMENT

The raw RNA‐Seq data are publicly available at the SRA database of NCBI at www.ncbi.nlm.nih.gov/sra/, accession no. PRJNA897094. The MS data have been deposited at the ProteomeXchange Consortium via the iProX repository at www.iprox.cn with the data set identifier IPX0005342001.