Fungal and Bacterial Communities Associated with Northern Corn Leaf Blight in Resistant and Susceptible Sweet Corn
Article information
Abstract
Northern corn leaf blight (NCLB), caused by Exserohilum turcicum (Setosphaeria turcica), is a major disease that negatively impacts the yield and quality of sweet corn. Plant-associated microbes hold great potential for enhancing crop productivity and sustainability. This study investigated the fungal and bacterial communities associated with NCLB in resistant and susceptible sweet corn cultivars using amplicon metagenomic sequencing. The structural composition and diversity of the fungal community in symptomatic NCLB-susceptible cultivars differed significantly from those in asymptomatic NCLB-resistant cultivars. In contrast, the bacterial communities showed no significant differences between resistant and susceptible cultivars in both the phyllosphere and rhizosphere. Exserohilum and Alternaria were significantly more abundant in the phyllosphere of symptomatic NCLB-susceptible plants, while fungal genera such as Sporobolomyces and Aureobasidium, along with the order Dothideales and the bacteria Bacillus, were significantly more abundant in the phyllosphere of asymptomatic NCLB-resistant cultivars. Microbial metabolic functions related to sugar metabolism—including sucrose biosynthesis and the degradation of glucose and xylose, compounds abundant in plant cell walls—were enriched in the phyllosphere of symptomatic NCLB-susceptible plants. In contrast, functions associated with detoxification and defense responses to plant phenolic compounds were enriched in microbes from asymptomatic NCLB-resistant cultivars. Additionally, Bacillus, identified as part of the core microbiome, and the epiphytic yeast Sporobolomyces, identified as a hub in the microbial network, exhibited antimicrobial activity that may suppress E. turcicum. These findings offer valuable insights into the role of microbial communities in plant health and disease resistance, with promising implications for developing microbiome-based strategies to manage NCLB.
Sweet corn (Zea mays L. var. rugosa Bonaf.) is a significant crop, both in terms of consumption and economic value. Based on the gene composition in the endosperm, sweet corn is categorized into five types: normal, sugary enhanced, super sweet, synergistic, and augmented shrunken (Singh et al., 2014). In Thailand, sweet corn varieties containing the shrunken-2 gene, which promotes high sucrose levels, dominate the market, accounting for around 95% of total consumption (Jompuk et al., 2020; Nwanosike et al., 2015). In Southeast Asia, sweet corn competes with traditional waxy maize, which serves as both a staple food and a snack (Sinkangam et al., 2011). Climate and soil conditions in Thailand make it an ideal location for sweet corn production. Despite its popularity, sweet corn is vulnerable to several fungal diseases, which can significantly impact yield and quality. Common foliar diseases include northern corn leaf blight (NCLB) (Exserohilum turcicum), southern corn leaf blight (Bipolaris maydis), gray leaf spot (Cercospora zea-maydis), common rust (Puccinia sorghi), anthracnose leaf blight (Colletotrichum graminicola), and downy mildews (Peronosclerospora spp. and Sclerospora spp.) (Nsibo et al., 2024). These diseases can cause significant yield losses and affect the quality and safety of sweet corn for human consumption.
NCLB caused by the fungus Exserohilum turcicum (Teleomorph: Setosphaeria turcica), is one of the important diseases that affect the yield and quality of sweet corn. This fungal pathogen is an ascomycete fungus that also caused Turcicum leaf blight of sorghum (Leonard et al., 1989). This disease is prevalent in maize growing areas worldwide and is associated with moderate-to-severe yield losses (Raymundo and Hooker, 1981). It can cause significant damage to sweet corn crops, that affects the leaves, husks, and stalks of the corn plant. The symptom of this disease characterized by long, cigar-shaped lesions on the leaves, which can turn brown and cause the leaves to die prematurely. E. turcicum is a hemibiotroph, being biotrophic pathogen before switching to be necrotrophic pathogen (Human et al., 2020). During infection, this pathogen penetrate host through the epidermis and formation of the haustorium, invades new tissues, causing necrosis, and then systemically spreads throughout the plant and blocks the vascular tissues (Kotze et al., 2019). This causes plant lodging and a reduction in photosynthetic leaf area, in severe cases, the disease can cause the husks to become infected, which can lead to premature kernel abortion and reduced yield (Raymundo and Hooker, 1981). A severe NCLB infection prior to flowering may cause >50% losses in maize final yields (Raymundo and Hooker, 1981).
Sweet corn foliar diseases are best managed by an integrated approach consisting of timing of planting, hybrid selection, crop debris management, disease scouting, biological control, and foliar fungicides (Shah and Dillard, 2010). In addition, genetic resistance is one of the most economical and effective strategies for managing NCLB and can be incorporated into an IPM program for preventive disease management. Quantitative resistance in maize to NCLB is associated with multiple candidate genes related to plant defense, including receptor-like kinase genes similar to those involved in basal defense (Poland et al., 2011). In Thailand, certain sweet corn cultivars, such as Insee 2, have demonstrated resistance to NCLB and experienced reduced yield loss when grown under conditions with high disease prevalence (Truong and Samphantharak, 2006).
Microbial communities play a crucial role in plant disease resistance by interacting with both the plant and potential pathogens (Amoo et al., 2023). These microbial populations, which include bacteria, fungi, yeasts, and other microorganisms, can interact with plants in ways that enhance their resistance to diseases (Xiong et al., 2023). These interactions protect plants by outcompeting pathogens, producing antimicrobial compounds, inducing plant immune responses, and adapting to environmental stresses (Lemanceau et al., 2017; Lindow and Brandl, 2003). Managing and promoting these beneficial microbial populations is a promising approach for improving plant health and disease resistance, particularly in sustainable agricultural systems.
Recent advances in next-generation sequencing technologies have led to a growing number of studies investigating the corn-associated microbiome and its dynamics with the plant (Wallace, 2023). These studies reveal that microbial communities are highly diverse and complex, with composition varying based on factors such as geographic location, climate, soil type, and agricultural practices (Brisson et al., 2019; Cavaglieri et al., 2009). The composition and differentiation of microbial communities are influenced not only by environmental factors and agricultural management but also by the host plant phenotype, genotype, and disease status (Chiarini et al., 1998). For example, the fungal communities in the phyllosphere of highly resistant maize varieties exhibit greater complexity, integrity, and stability compared to those in susceptible varieties (Luo et al., 2023). Additionally, the presence of specific bacterial species on corn leaves has been shown to enhance resistance to B. maydis infection (Balint-Kurti et al., 2010). In the rhizosphere, the production and release of primary and secondary root metabolites shape the root-associated microbiome. For instance, benzoxazinoids, tryptophan-derived heteroaromatic metabolites produced by maize, regulate global root metabolism while concurrently influencing the bacterial and fungal communities in the rhizosphere (Cotton et al., 2019). Understanding the functional roles of plant-microbe interactions and the factors involved in the assembly of plant-associated microbial communities can deepen our knowledge of plants as meta-organisms and how they utilize their microbiota to adapt to challenging environments.
In this study, we aimed to investigate the fungal and bacterial communities associated with NCLB in Thailand, focusing on a resistant commercial hybrid (Insee 2) and a susceptible cultivar (Hi-Brix 3) under field conditions, using amplicon metagenomic sequencing of fungal internal transcribed spacer (ITS) and bacterial 16S rRNA regions. Additionally, Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt2) analysis was employed to study shifts in bacterial community structure and functional potential in the phyllosphere and rhizosphere of the resistant (Insee 2) and susceptible (Hi-Brix 3) cultivars. By understanding the interactions between these fungal and bacterial communities, we aim to contribute to the development of effective disease management strategies and promote sustainable crop production practices.
Materials and Methods
Experimental site and sample collection
Two sweet corn cultivars were used in this study: the resistant cultivar Insee 2, representing asymptomatic NCLB plants, and the susceptible cultivar Hi-Brix 3, which is susceptible to NCLB caused by Exserohilum turcicum, representing symptomatic plants. The resistance levels of both cultivars were confirmed in greenhouse experiments following inoculation with three isolates of E. turcicum. Hi-Brix 3 exhibited typical NCLB symptoms, with infected leaf areas of 15%, 16%, and 30%, respectively, while Insee 2 showed no visible symptoms of NCLB (Supplementary Fig. 1B and C). Both cultivars were subsequently cultivated under field conditions at the National Corn and Sorghum Research Center (Suwan Farm) in Pak Chong, Nakhon Ratchasima, Thailand (Supplementary Fig. 1A). Suwan Farm has cultivated various maize cultivars, including Insee 2 and Hi-Brix 3, and has reported NCLB occurrence in its fields for over 20 years. The two cultivars were grown in four replicated blocks under field conditions. All Hi-Brix 3 plants displayed typical NCLB symptoms (Supplementary Fig. 1A), with infected leaf areas ranging from 15% to 30%, while all Insee 2 plants remained asymptomatic (Supplementary Fig. 1A). At the R3 growth stage (75 days after planting), symptomatic Hi-Brix 3 plants and asymptomatic Insee 2 plants were randomly selected for sampling. From each replicated block, four symptomatic (Hi-Brix 3) and four asymptomatic (Insee 2) plants were collected at 3-meter intervals in a zigzag pattern across the rows, totaling 16 plants per cultivar group.
Leaf samples were collected from the upper to middle sections of each sweet corn plant and placed into sterile plastic bags. Excess soil was gently removed from the roots, and representative root samples were similarly collected and bagged under sterile conditions. During sampling, plants were handled using gloved hands, and all tools and equipment were disinfected with 70% ethanol between each plant to prevent cross-contamination. All leaf and root samples were kept on ice and transported to the laboratory within four hours. In addition, soil samples were collected from the field, and their physical and chemical properties were analyzed (Supplementary Table 1).
Extraction of phyllosphere and rhizosphere microorganisms
Leaf samples from 16 NCLB symptomatic plants and 16 asymptomatic resistant plants (4 plants per block) were used to collect phyllosphere microorganisms. The collection procedure was based on the method described by Dong et al. (2019), with modifications. Leaves from each plant were washed in sterile phosphate-buffered saline containing 0.01% Triton X-100 (PBST: Na2HPO4 1.42 g/L, KH2PO4 0.24 g/L, NaCl 8 g/L, KCl 0.2 g/L; pH 7.4) while shaking at 250 rpm for 90 min at 25°C. The wash solution was then centrifuged at 12,000 rpm for 15 min at 4°C to collect microbial pellets. These pellets were stored at −80°C as phyllosphere samples until DNA extraction.
For rhizosphere microorganism extraction, root samples were first washed with sterilized PBST buffer for 1 min to remove bulk soil. The roots were then shaken in sterilized PBST buffer at 250 rpm for 90 min at 25°C. After centrifugation at 12,000 rpm for 15 min at 4°C, the pellets were collected as rhizosphere samples and stored at −80°C until DNA extraction.
DNA extraction and sample pooling
Microbial DNA from each pellet was extracted using the DNeasy PowerSoil Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. DNA quantification was performed using a NanoDrop ND-1000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA), and DNA integrity was verified using 1% agarose gel electrophoresis. For amplicon metagenomic sequencing, DNA from four samples collected from either the phyllosphere or rhizosphere within the same block was pooled to create a single composite sample per block. This method resulted in a single DNA sample representing the four asymptomatic resistant plants from one block and a separate DNA sample representing the four NCLB symptomatic plants from the same block. The samples included the asymptomatic resistant phyllosphere (CPH, n = 4), asymptomatic resistant rhizosphere (CRH, n = 4), NCLB - symptomatic phyllosphere (CPD, n = 4), and NCLB symptomatic rhizosphere (CRD, n = 4). The extracted DNA was stored at −20°C for future analysis.
Library preparation and amplicon-based metagenomic sequencing
To characterize fungal communities, the ITS1-1F region of the nuclear ribosomal ITS was amplified using primers ITS1-1F-F (5′-CTTGGTCATTTAGAGGAAGTAA-3′) and ITS1-1F-R (5′-GCTGCGTTCTTCATCGATGC-3′). For bacterial communities, the V3–V4 regions of the 16S rRNA gene were amplified using universal primer sets 341F (5′-CCTAYGGGRBGCASCAG-3′) and 806R (5′-GGACTACNNGGGTATCTAAT-3′), as described by Mayer et al. (2021). Sequencing libraries were prepared using the TruSeq DNA PCR-Free Sample Preparation Kit (Illumina, San Diego, CA, USA), following the manufacturer’s instructions. Index codes were added to enable sample identification. The quality of the prepared 16S rRNA and ITS libraries was assessed using a Qubit 2.0 Fluorometer (Thermo Scientific) and an Agilent Bioanalyzer 2100 system. Libraries were sequenced on an Illumina NovaSeq 6000 platform, generating 250 bp paired-end reads at NovogeneAIT Genomics Singapore Pte. Ltd. (Singapore).
Sequence data processing
The raw paired-end sequences underwent quality control, including trimming of adapters and primers, using the FASTP program (Chen et al., 2018). Further sequence quality checks were performed with FASTQC and MultiQC tools (Ewels et al., 2016). Clean paired-end sequences were merged using FLASH version 1.2.7 (Magoč and Salzberg, 2011). Data analysis of the 16S rRNA and ITS gene fragments was carried out using QIIME 2 (version 2020.06). The merged and cleaned sequences were imported into QIIME using the “qiime tools import” plugin, and sequence quality was visualized through “qiime demux summarize”.
An amplicon sequence variant (ASV) table was created using DADA2 (version 1.14) within QIIME 2, providing high-resolution taxonomic classification (Callahan et al., 2016). Taxonomic assignments were made using the UNITE database (version 9.0, 2023-07-18) for fungi (Abarenkov et al., 2024) and the SILVA 16S database (release 132; feature classifier classify-sklearn) for bacteria (Quast et al., 2013). Any sequences annotated as chloroplast or mitochondria were filtered out using the “qiime taxa filter-table” and “qiime taxa filter-seq” functions.
Microbial communities analysis
Microbial diversity analyses were conducted using R packages, including phyloseq 1.46.0 (McMurdie and Holmes, 2013), microbiome 1.24.0, eulerr 7.0.2 (Larsson, 2024), EasyStat 0.1.0, and vegan 2.6.4. (Oksanen et al., 2022). Alpha diversity, which measures the uniformity of relative abundance distributions of ASVs, was assessed using the Chao1, Pielou’s Evenness, and Shannon indices (Haegeman et al., 2013). Statistical differences in alpha diversity indices were evaluated through the Kruskal-Wallis test via the “alpha-group-significance” command. The α level of significance was set to 0.05 for all adjusted P-values. The box plots representing alpha diversity indices were created using ggplot2 (Wickham, 2016), developed by the R Core Team (2020).
Beta diversity, reflecting differences in community structure between sample groups, was quantified using the Bray-Curtis index (Bray and Curtis, 1957). Principal Coordinate Analysis (PCoA) was applied to Bray-Curtis distance matrices to visualize variations in microbial community composition across the sample groups. To test for significant differences in community composition due to factors such as plant compartment and NCLB disease, permutational multivariate analysis of variance (PERMANOVA) was conducted using 999 permutations (Anderson, 2017). Statistical significance was determined at P < 0.05.
Differences in microbial relative abundance at the genus level were evaluated with Welch’s t-test, followed by Bonferroni correction, using STAMP (structural time series analyzer, modeler, and predictor) software (Parks et al., 2014). To predict functional abundances and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway involvement, Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt2) (Douglas et al., 2020) was utilized within the QIIME2 plugin. Statistical analyses of the predicted functional profiles were further conducted using STAMP to visualize the abundance profiles generated by PICRUSt2. Statistical significance was determined at P < 0.05.
Network analysis of microbial communities in the phyllosphere was performed using the ASV table agglomerated at the genus level. The top 50 most abundant fungal and bacterial ASVs, based on total abundance across all samples, were selected to construct fungal community networks, bacterial community networks, and cross-kingdom interaction networks. Co-occurrence networks were inferred and analyzed using the SParse InversE Covariance Estimation for Ecological Association Inference (SPIEC-EASI) v1.1.3 R package (Kurtz et al., 2015). After applying centered log-ratio transformation, the Meinshausen-Bühlmann neighborhood selection method was employed to estimate the conditional dependence between each pair of ASVs, with the Stability Approach to Regularization Selection (StARS) used to determine the optimal sparsity parameter. The resulting sparse association matrices were passed to the NetCoMi R package (Peschel et al., 2021) for visualization and further analysis of network properties.
The efficacy of phyllospheric microorganism to inhibit E. turcicum
The phyllosphere microorganisms from leaf surface of sweet corn were isolated on nutrient agar (peptone, 5 g/L; beef extract, 3 g/L with 15% agar; HiMedia Laboratories, LLC., Mumbai, India). Individual colonies displaying distinct shapes and/or colors were selected and identified.
For bacteria, the universal primers 16SF (5′-GGAGAGTTAGATCTTGGCTCAG-3′) and 16SR (5′-GTGCTGCAGGGTTACCTTGTTACGACT-3′) were used to amplify the 16S rRNA gene (Newman et al., 2008). For yeasts, the primers ITS1 (5′-TCCGTAGGTGAACCTGCGG-3′) and ITS4 (5′-TCCTCCGCTTATTGATATGC-3′) (White et al., 1990), as well as NL1 (5′-GCATATCAATAAGCGGAGGAAAAG-3′) and NL4 (5′-GGTCCGTGTTTCAAGACGG-3′) (O’Donnell and Gray, 1995), were used to amplify the ITS gene and the 26S rRNA gene (D1/D2 domain of the large-subunit [LSU] rRNA gene), respectively. The amplified products were bidirectionally sequenced by Macrogen Inc. (Seoul, Korea) and analyzed through BLAST alignment in the NCBI database (http://blast.ncbi.nlm.nih.gov/Blast.cgi).
The biocontrol activity of phyllospheric microorganisms was assessed using the confrontation culture method, following the procedure established by (Li et al., 2021). Briefly, a 5-mm mycelium plug of E. turcicum was placed in the center of a potato dextrose agar plate. Drops of 10 μL of the broth culture from each phyllospheric microorganism were applied at four opposite points, each 25 mm from the center, and incubated at room temperature (~28°C). The control group consisted of plates inoculated with E. turcicum mycelia without the addition of phyllospheric microorganisms. Each treatment was repeated three times. The diameter of the fungal colony, with or without confrontation from bacterial and yeast strains, was measured and used to calculate the inhibition rate using the formula: Mycelial growth inhibition (%) = [(Control colony diameter – Treatment colony diameter)/(Control colony diameter – Inoculation plug diameter)] × 100.
Availability of data and materials
The raw sequence data for ITS and 16S rRNA generated during the study are available in the NCBI Short Read Archive at https://www.ncbi.nlm.nih.gov/sra, BioProject: PRJNA987847. The codes used for the analyses and plotting are available from the corresponding author upon reasonable request.
Results
Shared taxa and core fungal and bacterial communities
The sequencing statistics and ASVs were provided in the Supplementary Table 2. Analysis of shared ASVs across all samples revealed a total of 73 fungal taxa and 367 bacterial taxa common to all conditions (Fig. 1A and B). Among the unique taxa, 8 fungal ASVs were unique to the phyllosphere of the susceptible cultivar exhibiting NCLB symptoms, while 11 fungal ASVs were found exclusively in the phyllosphere of the resistant cultivar. For bacterial ASVs, 80 bacterial ASVs were unique to the phyllosphere of the susceptible cultivar exhibiting NCLB symptoms, while 26 bacterial ASVs were exclusive to the phyllosphere of the resistant cultivar (Fig. 1A and B).
The number of amplicon sequence variants (ASVs) and the core fungal and bacterial microbiota identified in this study. Venn diagrams of shared and unique ASVs for fungal ASVs (A) and bacterial ASVs (B), associated with the phyllosphere and rhizosphere of asymptomatic resistant plants (Insee 2) and symptomatic northern corn leaf blight (NCLB)-infected plants (Hi-Brix 3). The samples were categorized as follows: the asymptomatic resistant phyllosphere (CPH), asymptomatic resistant rhizosphere (CRH), NCLB - symptomatic phyllosphere (CPD), and NCLB symptomatic rhizosphere (CRD). Heatmaps of the fungal core microbiome (C) and bacterial core microbiome (D) show ASVs present in at least 90% of the sweet corn samples in this study. The data represent ASV data aggregated at the genus level.
Fungal ASVs present in at least 90% of the samples identified as core in this study included Periconia, Curvularia, Fusarium, Bipolaris, and Alternaria (Fig. 1C). These taxa are known for their roles in both symbiosis and plant pathogens. For bacterial communities, core microbiome genera included Bacillus, Pantoea, Sphingomonas, Pseudomonas, Enterobacter, and Microbacterium, which are often associated with plant health and important plant-microbe interactions (Fig. 1D).
Diversity of fungal and bacterial communities
Our preliminary data in greenhouse experiment found that without the presence of E. turcicum, the fungal and bacterial communities of uninfected plants of these two cultivar Insee 2 and Hi-Brix 3 were not significantly different in either the phyllosphere (fungi: R2 = 0.072, P = 0.667; bacteria: R2 = 0.358, P = 0.667) or the rhizosphere (fungi: R2 = 0.203, P = 1; bacteria: R2 = 0.088, P = 0.667) (Supplementary Fig. 2B).
Under natural field condition, the results showed that fungal communities in the phyllosphere of symptomatic NCLB leaves were clustered differently from those in asymptomatic resistant leaves (Fig. 2A), suggesting that the presence of NCLB symptoms may influence shifts in the fungal communities in the phyllosphere. These results were further confirmed by PERMANOVA on Bray-Curtis dissimilarity values (Table 1). The presence of NCLB symptoms significantly affected fungal communities in the phyllosphere (R2 = 0.542, P = 0.028), with a lesser impact observed in the rhizosphere (R2 = 0.312, P = 0.035) (Table 1, Fig. 2A).
Community diversity and structure of fungal and bacterial communities in the phyllosphere and rhizosphere of asymptomatic resistant plants (Insee 2) and symptomatic northern corn leaf blight (NCLB)-infected plants (Hi-Brix 3). β-Diversity is shown as principal coordinates analysis (PCoA) of Bray-Curtis dissimilarity for amplicon sequence variants. Fungal community composition differed significantly between asymptomatic resistant and susceptible NCLB-infected plants (A), while bacterial communities did not show significant differences in either the phyllosphere or rhizosphere (B). α-Diversity indices for fungal communities (C) and bacterial communities (D) in the phyllosphere and rhizosphere of asymptomatic resistant and symptomatic NCLB-infected plants are presented as box-and-whisker plots, visualizing the Chao1 index, Pielou’s evenness, and Shannon index based on disease factors. The Kruskal-Wallis test was used to assess the effects of disease on α-diversity indices. The significance level (α) was set at 0.05 for all adjusted P-values. The samples were categorized as follows: the asymptomatic resistant phyllosphere (CPH), asymptomatic resistant rhizosphere (CRH), NCLB - symptomatic phyllosphere (CPD), and NCLB symptomatic rhizosphere (CRD).
PERMANOVA results for the analysis of differences in fungal (ITS) and bacterial communities (16S) across the plant compartment and disease factors
However, bacterial communities in symptomatic NCLB samples, whether from the phyllosphere (R2 = 0.242, P = 0.055) or rhizosphere (R2 = 0.209, P = 0.093), did not significantly differ from those in asymptomatic resistant samples (Fig. 2B). These findings suggest that NCLB infection has a more substantial impact on fungal community composition, while bacterial communities exhibit greater resilience to these factors.
However, there were no significant differences in fungal alpha diversity indices between symptomatic NCLB samples and asymptomatic resistant samples in both the phyllosphere and rhizosphere (Fig. 2C). For bacterial communities, the Chao1 and Richness indices in symptomatic NCLB samples were significantly lower (P < 0.05) than in asymptomatic resistant samples (Fig. 2D). This suggests that NCLB infection reduces bacterial richness in the phyllosphere.
Microbial relative abundance
The fungal community in all sweet corn samples was predominantly composed of Cladosporium in the phyllosphere (79.29–89.34%) and the rhizosphere (55.04–74.16%) (Supplementary Fig. 3). After excluding Cladosporium from the communities, the phylum Ascomycota was the most abundant taxa in both the phyllosphere and rhizosphere (Fig. 3A). At the genus level, the fungal genera Exserohilum (34.81%), the pathogen responsible for northern corn leaf blight, and Papiliotrema (11.22%) showed increased abundance in the NCLB susceptible leaves whereas Sporobolomyces (20.43%) and Aureobasidium (11.07%) were the most common genera in the asymptomatic resistant phyllosphere samples (Fig. 3A). In the rhizosphere of susceptible plants, Aspergillus (19.54%), Fusarium (11.92%), and Acremonium (11.94%) were the most abundant genera. Fusarium (12.99%) and Acremonium (9.36%) were also abundant in the asymptomatic resistant rhizosphere (Fig. 3A).
The composition of fungal communities (A) and bacterial communities (B) at genus levels in the phyllosphere and rhizosphere of asymptomatic resistant plants (Insee 2) and symptomatic northern corn leaf blight (NCLB)-infected plants (Hi-Brix 3). Taxa occurring with less than 0.1% relative abundance are shown as “Others”. The samples were categorized as follows: the asymptomatic resistant phyllosphere (CPH), asymptomatic resistant rhizosphere (CRH), NCLB - symptomatic phyllosphere (CPD), and NCLB symptomatic rhizosphere (CRD).
For bacterial communities, Pantoea was the most abundant genus in the phyllosphere of both asymptomatic resistant and NCLB diseased plants (Fig. 3B). In NCLB symptomatic susceptible plant, the phyllosphere showed a higher relative abundance of Lactobacillus (11.26%) and Sphingomonas (7.95%), while Bacillus and Enterobacter were more abundant in the asymptomatic resistant samples (Fig. 3B). In the rhizosphere, Bacillus and Pseudomonas were the predominant genera in asymptomatic resistant plants, whereas Bacillus, Paenarthrobacter, and Pantoea were more abundant in the rhizosphere of NCLB symptomatic susceptible plants (Fig. 3B).
Differential abundance of fungal and bacterial genera in healthy resistant and NCLB-infected plants
STAMP analysis showed that Exserohilum and Alternaria were significantly more abundant in the NCLB symptomatic susceptible plants, while the fungal genera Sporobolomyces, Aureobasidium, and members of the order Dothideales were significantly more abundant in the phyllosphere of asymptomatic resistant plants (Fig. 4A). In the rhizosphere, Talaromyces and Aspergillus were more abundant in NCLB symptomatic susceptible plants, whereas members of the family Nectriaceae and the fungal genus Curvularia were enriched in asymptomatic resistant plants.
Differentially abundant fungal genera (A) and bacterial genera (B) in the phyllosphere and rhizosphere of symptomatic northern corn leaf blight (NCLB)-infected plants (Hi-Brix 3) and asymptomatic resistant plants (Insee 2). The mean proportion of sequences (percentage) of each genus, confidence intervals, and P-values are provided for each comparison. Only P-value < 0.05 (Welch’s test) are shown. The samples were categorized as follows: the asymptomatic resistant phyllosphere (CPH), asymptomatic resistant rhizosphere (CRH), NCLB - symptomatic phyllosphere (CPD), and NCLB symptomatic rhizosphere (CRD).
For bacterial communities, the genera Lactobacillus, Methylobacterium, Salinicoccus, and Jeotgalicoccus were significantly more abundant in the phyllosphere samples of NCLB symptomatic susceptible plants. In contrast, several uncultured bacteria, along with Bacillus, Nocardiodes, Paenarthrobacter, Streptomyces, and Tumebacillus, were more abundant in the phyllosphere of asymptomatic resistant plants, with Bacillus showing the most significant difference. In the rhizosphere, uncultured bacteria, Solirubrobacter, and Sinomonas were more abundant in NCLB symptomatic susceptible samples (Fig. 4B).
Differential abundance of microbial metabolic functions
PICRUSt2 was used to predict microbial functional pathways and assess potential functional changes in comparison between NCLB symptomatic susceptible plant and asymptomatic resistant plants (Fig. 5). Heatmap and PCoA revealed that the distribution of KEGG pathways differed between the phyllosphere and rhizosphere of both asymptomatic resistant plants and NCLB symptomatic susceptible plant (Fig. 5A and 5B). Three pathways were significantly enriched in the phyllosphere of NCLB symptomatic susceptible plant (Welch’s t-test; P < 0.05) (Fig. 5C). These included pathways related to sucrose biosynthesis (from photosynthesis) and the superpathway of glucose and xylose degradation, which crucial for breaking down glucose and xylose, fundamental sugars in plant tissues. In addition, N10-formyl-tetrahydrofolate biosynthesis, which important for folate metabolism, was also enriched in the phyllosphere of NCLB symptomatic susceptible plant. These results suggest that microbial communities associated with NCLB infection are more engaged in sugar metabolism, likely indicating that they encounter these sugars in the infected tissues.
Differential abundance of microbial metabolic functions from the phyllosphere and rhizosphere of symptomatic northern corn leaf blight (NCLB)-infected plants (Hi-Brix 3) and asymptomatic resistant plants (Insee 2). (A) Heatmap analysis of the abundance of predicted Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways across all samples. (B) Principal coordinate analysis of predicted KEGG pathways compositions across all samples. Extended error bar plot of significantly different KEGG pathways in the phyllosphere (C) and rhizosphere (D). Bar plots on the left display the mean proportion of each KEGG pathway, while dot plots on the right show differences in mean proportions between the two groups based on P-values. Only pathways with P-values < 0.05 (Welch’s test) are shown. The samples were categorized as follows: the asymptomatic resistant phyllosphere (CPH), asymptomatic resistant rhizosphere (CRH), NCLB - symptomatic phyllosphere (CPD), and NCLB symptomatic rhizosphere (CRD).
In addition, 27 pathways were significantly more abundant in the phyllosphere of resistant plants (Welch’s t-test; P < 0.05) (Fig. 5C). Many of these pathways are linked to functions that help microbes cope with plant defense responses, including coumarins biosynthesis (engineered), 4-coumarate degradation (anaerobic), catechol degradation to 2-oxopent-4-enoate II, mandelate degradation I, 2-aminophenol degradation, 2-nitrobenzoate degradation I, and ergothioneine (EGT) biosynthesis I. These findings suggest that microbial communities in resistant sweet corn plants focus on detoxification and defense-related functions, enabling them to be more compatible with the unique environment found in resistant plants.
In the rhizosphere, 23 metabolic pathways were significantly different (Welch’s t-test; P < 0.05) (Fig. 5D). Metabolic pathways related to various functions, particularly those involved in carbohydrate metabolism, energy metabolism (Calvin-Benson-Bassham cycle) and amino acid metabolism (the superpathway of L-methionine biosynthesis via transsulfuration) were enriched in the rhizosphere of NCLB symptomatic susceptible plant. Moreover, pathways involved in the metabolism of cofactors and vitamins such as superpathway of S-adenosyl-L-methionine biosynthesis and terpenoid and polyketide biosynthesis (taxadiene biosynthesis) were notably enriched in infected plant rhizospheres.
In contrast, pathways that were more prominent in the rhizosphere of resistant plants included those related to fatty acid biosynthesis (palmitate biosynthesis II), amino acid degradation (L-leucine degradation, L-tyrosine degradation, and L-tryptophan degradation XII from Geobacillus), as well as vitamin B12 metabolism (adenosylcobalamin salvage from cobinamide II and adenosylcobalamin biosynthesis from cobyrinate a,c-diamide I). These pathways suggest a heightened microbial response to metabolic stress and adaptation in resistant plants.
Microbial network structures in the phyllosphere
Network analysis revealed distinct differences in both intrakingdom and interkingdom interactions within the phyllosphere communities of asymptomatic resistant and NCLB-symptomatic susceptible plants (Fig. 6, Supplementary Table 3).
Microbial network analysis revealed notable differences between the phyllosphere communities of healthy resistant plants and northern corn leaf blight (NCLB)-infected sweet corn leaves. The interaction networks were constructed using the top 50 amplicon sequence variants at the genus level for the fungal–fungal intra-kingdom network (A, B), bacterial–bacterial intra-kingdom network (C, D), and the fungal–bacterial interkingdom network (E, F) in the healthy (asymptomatic) resistant phyllosphere (CPH) and symptomatic NCLB-infected phyllosphere (CPD), respectively. Green color indicates positive associations, red indicates negative associations, and edge width represents the strength of the association. The modularity score was calculated to measure the strength of the division of each network into clusters, with cluster numbers identified using fast greedy modularity optimization. Network hubs were defined by the highest eigenvector centrality values, determined as those above the empirical 95% quantile of all eigenvector centralities. The network plots were visualized using the Fruchterman-Reingold layout algorithm.
In the fungal–fungal networks, Didymosphaeriaceae, Exserohilum, and Rhodotorula emerged as key hubs in asymptomatic resistant plants (Fig. 6A). In contrast, Bipolaris, Curvularia, and Stachylidium were identified as central hubs in the networks of NCLB-symptomatic susceptible plants. Notably, Exserohilum clustered with Striaticonidium, Phaeosphaeriaceae, and Fusarium, showing a direct positive interaction with Striaticonidium and a direct negative interaction with Phaeosphaeriaceae (Fig. 6B). Additionally, fungal interactions in symptomatic plants were characterized by a higher number of clusters a slightly lower clustering coefficient (0.049 vs. 0.06) and modularity (0.62 vs. 0.68), along with a marginally greater proportion of positive edges (50.85% vs. 50.79%) compared to resistant plants (Fig. 6A and B, Supplementary Table 3).
The bacterial–bacterial networks for both plant types consisted of 7 clusters (Fig. 6C and D). In the resistant plants, Intrasporangiaceae (F), Jeotgalicoccus, and Tumebacillus were identified as central hubs (Fig. 6C), whereas in the susceptible plants, Bacillus, Staphylococcus, and Streptomyces played a dominant role (Fig. 6D).
In the bacterial–fungal interkingdom networks, NCLB-symptomatic susceptible plants exhibited higher complexity, comprising 9 clusters with a clustering coefficient of 0.181 and modularity of 0.613. In comparison, the asymptomatic resistant plants had 8 clusters and slightly lower values for both the clustering coefficient (0.165) and modularity (0.609) (Fig. 6E and F, Supplementary Table 3). Key interkingdom hubs in symptomatic plants included fungal taxa such as Pleosporales, Saccharomycetales, and Alfaria, as well as the bacterial genus Nocardioides and unidentified bacterial taxa. In contrast, the interkingdom network of resistant plants featured fungal hubs like Aureobasidium, Mucor, and unidentified fungi, along with bacterial genera such as Bifidobacterium and Streptomyces.
Efficacy of phyllospheric bacteria and yeast in inhibiting Exserohilum turcicum
To evaluate the biocontrol potential of phyllospheric microorganisms in suppressing the growth of E. turcicum, the causal agent of northern corn leaf blight, various bacteria and yeasts were isolated and identified. These included Bacillus sp., Brevibacillus sp., Curtobacterium citreum, Sphingomonas yabuuchiae, Sporobolomyces patagonicus, and Rhodotorula taiwanensis.
The bacterial strains Bacillus sp. (CD2-2), Bacillus sp. (DC5), and Brevibacillus sp. (CD4), along with the yeasts S. patagonicus (CD2) significantly (P < 0.05) inhibited the mycelial growth of E. turcicum, with inhibition percentages of 70.13%, 67.89%, 62.66%, and 35.79%, respectively (Fig. 7A and B). However, C. citreum, S. yabuuchiae, and R. taiwanensis did not demonstrate the ability to inhibit the growth of E. turcicum. Consistent results were observed across independent experiments. This analysis highlights the potential of certain phyllosphere microorganisms to act as biocontrol agents against E. turcicum, offering promising avenues for the development of biological control strategies against NCLB.
In vitro inhibition of Exserohilum turcicum by phyllospheric bacteria and yeast isolated from sweet corn leaves. (A) Box plot showing the inhibition percentages of E. turcicum mycelial growth by bacteria and yeast (n = 5). Statistically significant differences are indicated by different letters based on Tukey’s honest significant difference test (P < 0.05). Similar results were observed in repeated independent experiments. (B) The interactions between Exserohilum turcicum (control) and phyllospheric microorganisms belonging to the genera Sphingomonas yabuuchiae CD1, Sporobolomyces patagonicus CD2, Bacillus sp. CD2-2, Curtobacterium citreum CD3, Brevibacillus sp. CD4, Bacillus sp. CD5, Hannaella siamensis CH1, Rhodotorula taiwanensis CH2, and Curtobacterium albidum CH4.
Discussion
This study provided valuable insights into the fungal and bacterial communities associated with the phyllosphere and rhizosphere of a resistant cultivar (Insee 2), representing asymptomatic NCLB-resistant plants, and a susceptible cultivar (Hi-Brix 3), representing NCLB-symptomatic susceptible plants. These sweet corn cultivars were studied in relation to E. turcicum, the causal agent of NCLB. Additionally, we identified certain phyllosphere bacteria and yeasts as core microbiota or key hubs in the microbial network, demonstrating antifungal activity that inhibits the NCLB pathogen.
The identification of 421 fungal and 2,503 bacterial ASVs across different compartments and cultivars highlights the complexity and diversity of the sweet corn microbiome. The identification of core fungal taxa, such as Periconia, Curvularia, Fusarium, Bipolaris, and Alternaria, is particularly noteworthy, as these genera include several species known for their roles as plant pathogens or opportunistic fungi. Periconia species are often found as fungal endophytes, and some species can cause disease in plants (Bovio et al., 2023; Tian et al., 2022). Curvularia such as Curvularia lunata, is a well-known pathogen that causes maize leaf diseases in the world. Bipolaris maydis, is also the leaf pathogen, a causal agent of Southern corn leaf blight, one of the major leaf diseases in maize (Tatum, 1971). It has been reported that Fusarium oxysporum, Alternaria alternata, and Periconia are highly abundant in plant leaves exposed to drought conditions (Garcia et al., 2018). The presence of these fungi in both resistant and susceptible sweet corn cultivars in Thailand suggests that they may be endemic to the sweet corn microbiome, potentially playing dual roles as commensal organisms and opportunistic pathogens.
The bacterial communities were more diverse than the fungal communities, with 2,503 ASVs identified, including core bacterial taxa Bacillus, Pantoea, Sphingomonas, Pseudomonas, Enterobacter, and Microbacterium. Our results are consistent with other corn microbiomes which also found that these genera are the most common bacterial species in the maize phyllosphere and rhizosphere (Singh and Goodwin, 2022).
Our study showed that both plant compartment and NCLB infection significantly influenced the composition of fungal communities. Our results are consistent with several studies on plant microbiomes, where microbial assemblages are shown to be associated with specific plant organs (Trivedi et al., 2020; Vorholt, 2012). It has been reported that the fungal communities in the phyllosphere of the four varieties of maize with various levels of resistances to E. turcicum were significantly different and clustered into the respective maize variety they inhabited (Luo et al., 2023). This is expected as the pathogen thrives and proliferates in infected tissues. In contrast, bacterial community compositions in both phyllosphere and rhizosphere were not significantly affected by NCLB infection but did affect the richness of bacterial community. This suggests that bacterial communities in sweet corn are more resilient to changes induced by Exserohilum infection.
However, both disease presence and host genetic resistance can influence shifts in the plant microbiome (Chiarini et al., 1998; Luo et al., 2023). In this study, samples from the resistant cultivar (Insee 2) and the susceptible cultivar (Hi-Brix 3) were collected under natural field conditions, where all Insee 2 plants were asymptomatic, while all Hi-Brix 3 plants exhibited symptomatic leaves. However, our preliminary data indicated that the fungal and bacterial communities in the phyllosphere of healthy leaves from both cultivars were not significantly different (Supplementary Fig. 2B). These findings suggest that the observed microbial shifts are primarily driven by the presence of NCLB rather than inherent host resistance. Nevertheless, the individual contributions of genetic resistance and disease presence warrant further investigation.
In the fungal community, the significant increase in Exserohilum and Alternaria in NCLB-symptomatic susceptible plants. Members of the genus Alternaria range from saprophytes to endophytes, and also can be pathogens of plants (Wallace, 2023). Alternaria species including A. alternata, A. tenuissima, A. burnsii, and an unclassified Alternaria species has been globally reported as a prevalent causal agent of maize leaf blight (Gonçalves et al., 2013; Xu et al., 2022). This is consistent with their role as pathogens or opportunistic fungi that proliferate under weakened host conditions. For the bacterial community, Lactobacillus, Methylobacterium, and Salinicoccus, were found to be more abundant in NCLB-symptomatic susceptible plants. Notably, bacterial populations such as Methylobacterium spp. and Lactobacillus spp. have also been shown to be enriched in trees infected with the pathogen (Ginnan et al., 2020). Conversely, beneficial fungi from the genera Sporobolomyces and Aureobasidium were found to be more abundant in the phyllosphere of the resistant cultivar. This study is the first to demonstrate that resistant plants harbor a higher abundance of these taxa. Several yeasts, such as those in the genera Sporobolomyces, Aureobasidium, Rhodotorula, and Papiliotrema, have been associated with plant leaves (Kemler et al., 2017; Labancová et al., 2023) and have been reported to exhibit biological activity against various plant pathogens. Some strain of epiphytic yeast such as A. pullulans, Pseudozyma aphidis has been shown to induce plant resistance (Buxdorf et al., 2013; Zeng et al., 2023). In our study, we found that Sporobolomyces isolated from sweet corn leaves was able to inhibit E. turcicum, suggesting that these taxa are potentially outcompeted or inhibited in NCLB diseased plants.
E. turcicum is a hemibiotrophic fungus, which initially lives as a biotroph, feeding on living host tissue, before switching to a necrotrophic lifestyle where it kills infected host cells (Levy and Cohen, 1983; Ohm et al., 2012). During pathogen infection, an increase in the number of sugar transporters of several plants results in sugar accumulation in the apoplast, which is used by pathogens as a source of carbon and energy (Lemoine et al., 2013). In this study, we observed that several pathways related to carbohydrate metabolism, such as sucrose biosynthesis and glucose and xylose degradation, were significantly enriched in the phyllosphere of susceptible plants infected with NCLB. These pathways are crucial for breaking down fundamental sugars, suggesting heightened microbial activity in response to infection (Lievens et al., 2015). This increased carbohydrate metabolism could indicate that microbes are utilizing plant-derived sugars released during E. turcicum infection, potentially contributing to disease progression. Folate, or vitamin B9, is a generic term for tetrahydrofolate and its C1-substituted derivatives, which are synthesized de novo in bacteria, fungi, and plants, including sweet corn (Xiao et al., 2022). Moreover, the enrichment of the N10-formyl-tetrahydrofolate biosynthesis pathway, essential for folate metabolism, was increased in the microbial community of NCLB-infected leaves, suggesting that folate may play a role in the growth and establishment of phyllosphere bacterial communities on the plant leaves.
In contrast, resistant sweet corn plants exhibited elevated activity in several defense-related metabolic pathways within the phyllosphere microbial community. Pathways such as catechol degradation, mandelate degradation, and EGT biosynthesis were significantly increased. In the interaction between pathogens and plants, plant secondary metabolites such as phenolics, terpenes, and nitrogen- and sulfur-containing compounds play important roles (Kliebenstein et al., 2005). Plant phenolic compounds are ubiquitous in plant species, and include simple phenols, phenolic acids, coumarins, flavonoids, stilbenes, hydrolyzable and condensed tannins, lignans, and lignins derived from phenylalanine and tyrosine (Naczk and Shahidi, 2004). Plant-derived phenols containing a catechol group and mandelic acid (MA) are plant-derived compounds commonly involved in plant defense mechanisms (Pagare et al., 2015), suggesting that these microbes possess traits allowing them to survive and thrive under the plant defense conditions. For instance, catechol has been identified in resistant cultivars of Poplar (Populus spp.), where it demonstrated significant antifungal activity against Botryosphaeria dothidea (Li et al., 2020). MA, an 8-carbon α-hydroxy acid and a typical aromatic compound, has been reported as a phytochemical with antimicrobial potential (Abed et al., 2022). The MA degradation pathway has been discovered in a variety of prokaryotic and eukaryotic microorganisms that can convert MA to benzoic acid (Wang et al., 2022) or directly use MA as a carbon source. Examples include Acinetobacter, Azotobacter, Bacillus, Pseudomonas, Aspergillus, Neurospora, and Rhodotorula (Wang et al., 2022). This pathway involves enzymes such as mandelate racemase, S-mandelate dehydrogenase, benzoylformate decarboxylase, and benzaldehyde dehydrogenase, which feed into downstream benzoic acid degradation pathways (Wang et al., 2022). Our study reveals that Bacillus was also significantly increased in the resistant sweet corn leaves. The increased abundance of microorganisms capable of degrading MA in resistant sweet corn plants suggests that microbes may contribute to breaking down harmful plant defense chemicals or play a role in balancing the plant defense response with microbial symbiosis. EGT is a sulfur-containing antioxidant derived from histidine, synthesized by actinomycetes, cyanobacteria, methylobacteria, and some fungi, but not by animals or plants (Cumming et al., 2018; Fujitani et al., 2018; Genghof, 1970). Our results showed that several Gram-positive bacteria, such as Bacillus and Streptomyces, as well as yeasts Aureobasidium, were more enriched in the phyllosphere of resistant plants and Rhodotorula was found as a hub in the network. Aureobasidium pullulans and Rhodotorula mucilaginosa have previously been reported as EGT-producing eukaryotic strains (Fujitani et al., 2018). This suggests that the capacity of these microbes to synthesize EGT, an antioxidant, may enable them to mitigate oxidative stress caused by the plant immune response, benefiting both the plant and its associated microbial community. Furthermore, the superpathway of GDP-mannose-derived O-antigen building block biosynthesis and dTDP-L-rhamnose biosynthesis, which are components of bacterial outer membranes and play roles in evading plant immune responses. The enrichment of these pathways in resistant plants may indicate an adaptive response by bacteria to survive in the altered environment of the phyllosphere, allowing them to persist despite the plant defense mechanisms.
The fungal–fungal intrakingdom network in the NCLB-symptomatic susceptible phyllosphere showed more clusters and a lower clustering coefficient, suggesting a more fragmented and less cohesive community structure compared to the asymptomatic resistant plant network. Exserohilum clustered with Striaticonidium, Phaeosphaeriaceae, and Fusarium, showing a direct positive interaction with Striaticonidium and a negative interaction with Phaeosphaeriaceae. This pattern may reflect specific ecological interactions, where Exserohilum and Striaticonidium potentially co-occur or cooperate under certain environmental or host conditions, while Phaeosphaeriaceae may be competitively excluded. The interkingdom network analysis revealed more clusters in the NCLB-symptomatic susceptible plant, with slightly higher positive edge connectivity, suggesting more frequent co-occurrence of bacterial and fungal species in response to NCLB infection. The bacterial-bacterial network analysis found that Bacillus and Streptomyces were present as hub in the NCLB-symptomatic susceptible plant network points to the critical roles these genera may play in stabilizing bacterial communities under disease stress. Our study showed that phyllospheric bacteria Bacillus sp. and Brevibacillus sp. and phyllospheric yeast S. patagonicus showed strong biocontrol potential against E. turcicum. It has been reported that Bacillus are the core microbiota of resistant maize cultivars and exhibited the ability to induce plant defense against corn stalk rot pathogen, Fusarium graminearum (Xia et al., 2024). These antagonistic yeasts have also been widely used as biological control agents and plant growth-promoting rhizobacteria (Palmieri et al., 2022). The effectiveness of Bacillus and Sporobolomyces in controlling NCLB in planta, both in greenhouse and field experiments, warrants further investigation.
Although the fungal and bacterial communities identified in this study align with findings from other corn microbiome research, the relatively small number of pooled samples (n = 4 per group) may limit statistical power. Network analysis was performed using SPIEC-EASI, a statistically robust method for inferring microbial associations from compositional data. However, it assumes network sparsity and linear relationships, is sensitive to small sample sizes, and does not capture temporal dynamics or infer direct biological causality. Therefore, future studies with larger sample sizes are needed to validate and expand upon these findings.
This study demonstrates that the fungal community is highly susceptible to disruption by NCLB infection, with notable shifts in community structure and network interactions observed in the diseased phyllosphere. In contrast, bacterial communities show greater resilience to the effects of NCLB, although some shifts in abundance and network structure were still observed. Furthermore, the results highlight distinct metabolic shifts in microbial communities between resistant and NCLB-susceptible sweet corn plants. In resistant plants, the metabolic functions of the microbial community are more aligned with the degradation of defense compounds produced by the plant, whereas in infected plants, the metabolic functions of the microbial community demonstrate increased carbohydrate metabolism and bacterial survival strategies, potentially contributing to disease progression. The phyllospheric bacteria Bacillus sp. and Brevibacillus sp., along with the phyllospheric yeast S. patagonicus, identified as core microbiome members or hubs in the network, were able to inhibit the growth of E. turcicum. The efficacy of these isolates in providing disease protection under greenhouse and field conditions should be further investigated. Understanding these interactions between the plant, its microbiome, and pathogen infection could be key to developing strategies to enhance plant resistance and manage NCLB effectively.
Notes
Conflicts of Interest
No potential conflict of interest relevant to this article was reported.
Acknowledgments
We thank the personnel at National Corn and Sorghum Research Center, Kasetsart University for their assistance during sample collection. This research is funded by the Fundamental Fund under project FF(KU)2.66. from Kasetsart University Research and Development (KURDI), and Kasetsart University through the Graduate School Fellowship Program.
Electronic Supplementary Material
Supplementary materials are available at The Plant Pathology Journal website (http://www.ppjonline.org/).