Abstract
Gummy stem blight (GSB), a severe and widespread disease causing great losses to cucurbit production, is a major threat to melon production. However, the melon-GSB interaction remains largely unknown, which significantly impedes the genetic improvement of melon for GSB resistance. Here, full-length transcriptome and widely targeted metabolome were used to reveal the early defense responses of resistant (PI511890) and susceptible (Payzawat) melon to GSB. Differentially expressed genes were specifically enriched in the secondary metabolite biosynthesis and MAPK signaling pathway in PI511890, while in carbohydrate metabolism and amino acid metabolism in Payzawat. More than 1000 novel genes were identified in PI511890, which were enriched in the MAPK signaling pathway. There were 11,793 alternative splicing events identified and involved the defense response to GSB. A total of 910 metabolites were identified, with flavonoids as the dominant metabolites. Integrated full-length transcriptome and metabolome analysis showed that eriodictyol and oxalic acid may be used as marker metabolites for GSB resistance in melon. Moreover, post-transcription regulation was widely involved in the defense response of melon to GSB. These results improve our understanding of the interaction between melon and GSB and may facilitate the genetic improvement of GSB resistance of melon.
Key words: Melon; Gummy stem blight; Full-length transcriptome; Metabolome
Introduction
Melon (Cucumis melo L.), an important horticultural crop with great economic significance, is widely grown for fresh consumption. According to the statistics of Food and Agriculture Organization, the global melon production reached 28.6 million tons in 2021 (FAOSTAT, www.fao.org/faostat). China is the leading producing country of melon, accounting for approximately half of the total production (14.1 million tons), followed by Turkey, India, and Kazakhstan, whose annual production was 1.4–1.6 million tons in 2021. However, the yield and quality of melon are faced with serious threats of diseases caused by pathogen attack.
Gummy stem blight (GSB) caused by Stagonosporopsis cucurbitacearum (syn. Didymella bryoniae ) is a prevalent and devastating fungal disease of melon throughout the world (Li et al., 2017; Stewart et al., 2015). It has been reported that GSB pathogens can attack 37 species of the Cucurbitaceae family (Rennberger & Keinath, 2018). Under favorable environmental conditions, the pathogen can infect all aboveground parts of susceptible plants throughout the whole growth period, causing the formation of necrotic spots and serious reduction of yield and quality. The incidence of GSB can reach up to 80% for melon cultivated in greenhouse, and the yield loss can reach 100% once infected (Rennberger & Keinath, 2018; Virtuoso et al., 2022). Currently, chemical control, particularly fungicides, is the most widely used method to control GSB. However, excessive application of fungicides inevitably causes negative impacts on the environment and food safety. In addition, the effect is declining due to increasing resistance of certain pathogenic isolates to chemicals (Keinath & Zitter, 1998; Hassan et al., 2018).
Breeding of resistant cultivars is the most efficient approach for disease control. In recent years, some research efforts have been devoted to screening GSB-resistant melon germplasm (Wolukau et al., 2007; Zhang et al., 1997). A review has summarized the currently identified melon GSB-resistant resources (Luo et al., 2022). Another study investigated the inheritance of GSB-resistant traits, resulting in the identification of five independent monogenic resistance loci from PI accessions of PI140471, PI157082, PI511890, PI482398, and PI482399, which were designated as Gsb-1 , Gsb-2 , Gsb-3 ,Gsb-4 , and gsb-5 , respectively (Frantz & Jahn, 2004). Only a limited number of molecular markers associated with GSB resistance have been developed for maker-assisted selection of melon (Hassan, Rahim et al., 2018; Hassan, Robin et al., 2018; Wolukau et al., 2009). By using an ultra-dense genetic map, a previous study mapped GSB resistance QTLs from an inbred line of Cucumis melo spp.conomon into a 108-kb interval on chromosome 4 and proposed a candidate gene (Hu et al., 2018). Recently, Gsb-7(t) conferring GSB resistance was mapped on chromosome 7 and MELO3C010403-T2 was proposed as the candidate gene (Ma et al., 2023). However, functions of these candidate genes have not been validated yet (Seblani et al., 2023).
Clarifying the defense response of host to pathogen infection is important for understanding the disease resistance mechanism. High-throughput omics technologies have become powerful tools for studying plant defense response to biotic stresses, among which transcriptome is widely employed to identify the genes, signal transduction pathways, and regulatory networks involved in plant-pathogen interaction. For example, a transcriptomic analysis in a recent study demonstrated that an apyrase-like gene plays an important role in the defense response of pumpkin to GSB (Zhao et al., 2022). Alternative splicing (AS), which is usually identified by full-length transcriptome, is an important post-transcriptional regulatory mechanism that increases the diversity of transcripts and proteins (Ule & Blencowe, 2019). Several studies have shown that many genes undergo AS in response to biotic stresses in plants (Zhang et al., 2019). Functional analysis of alternative transcripts has become a powerful tool to develop new strategies for improvement of plant tolerance to environmental stress (Kufel et al., 2022). Metabolome can act as a bridge between genotypes and phenotypes, and is also a powerful tool for decoding plant-pathogen interaction (Serag et al., 2023). Disease infection can cause great perturbation on plant metabolism. The widely targeted metabolome allows comprehensive metabolic profiling of plants during pathogen attack. It is known that plant secondary metabolites, including phenolic compounds, alkaloids, glycosides, and terpenoids, play pivotal roles in plant-pathogen interaction (Anjali et al., 2023). However, gene expression profiles, AS landscape, and metabolites involved in the defense response of melon to GSB remain largely unknown.
In this study, we selected two melon accessions with contrasting resistance to GSB, and determined their defense responses to GSB based on full-length transcriptome and widely targeted metabolome. We also characterized the novel genes, AS events, differentially expressed genes (DEGs), and differentially accumulated metabolites (DAMs) involved in the defense response of melon to GSB. The results are expected to provide a comprehensive understanding on the defense response of melon to GSB at transcriptomic and metabolic levels, as well as valuable information for elucidating the mechanism for the resistance of melon to GSB.
Materials and methods
Plant materials and artificial inoculation
PI511890 (C. melo var. chito ) from Mexico and Payzawat from China (C. melo var. inodorus ) were used as the plant materials in this study. PI511890 is a wild melon accession and exhibits resistance to GSB (Frantz & Jahn, 2004). Payzawat is a widely cultivated landrace and susceptible to GSB. The seeds were firstly sterilized with 1.5% sodium hypochlorite, and then sown in pots containing sterilized peat-perlite substrate (2: 1, v/v). The pots were placed in a greenhouse and seedling management followed the commercial production practices. At the third true leaf stage, uniform and healthy seedlings were selected for subsequent experiments.
Pathogenic fungi were isolated from melon stem with typical GSB symptoms and identified as Stagonosporopsis cucurbitacearum . Purified fungi were inoculated on potato dextrose agar (PDA) culture medium, cultured at 25℃ under darkness for three days, then treated with 12 h photoperiod under ultra-violet light (280–360 nm) for five days, and maintained at darkness for two days to obtain the spores. The spores on the medium were washed off and filtered to obtain the spore suspension. The inoculum suspension was adjusted to 5 × 105spores/mL by adding ddH2O. For inoculation, the spore solution was sprayed on the seedlings until reaching the point of runoff. After inoculation, the seedlings were covered with a plastic tunnel and the relative humidity was kept over 90% with the temperature varying from 25℃ to 30℃. At the same time, spraying of distilled water was performed for the other seedlings to be used as the controls. Completely randomized block experimental design with three biological replicates was adopted for the treatments and controls, with each replicate consisting of 20 seedlings. Leaves were sampled at 0, 12, 24, 36, 48, 60, 72 h after inoculation (hpi) for subsequent analyses. For transcriptome and metabolome analyses, the leaves were immediately frozen in liquid nitrogen and then stored in a refrigerator (–80°C).
Histochemical staining
Trypan Blue staining was performed for the leaves to determine the growth of spores and hyphae according to the previous reports (Bhadauria et al., 2010; van Wees, 2008). Briefly, the leaves were punched to discs with a diameter of 10 mm and soaked in the Trypan Blue staining solution, then immediately heated in 90℃ water for 10 min. After the solution was allowed to cool down to room temperature, the staining solution was discarded and the leaf discs were decolorized using 2.5 mg/mL chloral hydrate solution, which was replaced after every 24 h until the leaf discs were completely decolorized.
Accumulation of H2O2 and O2- in leaves was measured using 3,3’-diaminobenzidine (DAB) and nitro blue tetrazolium chloride (NBT) staining methods, respectively (Daudi & O’Brien, 2012). In brief, the leaves were punched into several discs with a diameter of 10 mm. For each biological replicate, 10 discs were selected and immersed in 1 mg/mL DAB solution under 25 ℃/light for 5 h and 0.5 mg/mL NBT solution under 25 ℃/dark for 5 h, respectively. Then, the leaf discs were decolorized with 95% ethanol under 95℃ for 20 min and immersed in anhydrous ethanol for store and photo. The staining results were observed under a light optical microscope (OLYMPUS C × 41) with an ocular magnification lens at 40 × (400 um scale). The staining areas were calculated using Image J with the formula of (stained leaf area/leaf disc area) × 100% (Schneider et al., 2012). The larger staining area means higher accumulation of H2O2 or O2-. ANOVA was used to test the differences in staining areas among treatments and the least significant difference was used for multiple comparisons of the means. The agricolae package of R was used for statistical analysis.
Full-length transcriptome analysis
Samples at 24 hpi were selected for full-length transcriptome analysis, which included GSB-inoculated samples of PI511890 (TRT) and Payzawat (TST), and controls of PI511890 (TRC) and Payzawat (TSC). Extraction of RNA and construction of sequencing libraries were performed according to the protocols providing by the Oxford Nanopore Technologies (ONT). The libraries were sequenced on a PromethION platform to obtain the full-length transcriptome according to the standard protocol of ONT.
The pipeline for full-length transcriptome analysis is shown in Supplementary Figure 1. The short fragments and low-quality reads (length < 100 bp, Qscore < 7) were filtered out using NanoFilt (v2.8.0; Coster et al., 2018). The clean data were then processed with Pychopper (v2.4.0) to identify and orient full-length sequences. The melon genome of DHL92 (v4.0) was used as the reference (http://cucurbitgenomics.org/v2/organism/23). The full-length sequences were mapped to the reference genome using minimap2 (v2.17-r941; Li, 2018). Samtools (v1.7) was used to extract the uniquely mapped reads with a minimum quality score of 10. After polishing and clustering the full-length sequences, the consensus sequence was obtained using Pinfish pipeline (v0.1.0; Chen et al., 2021). The resulting consensus transcripts were then mapped to the reference genome using minimap2.
Transcript isoforms were identified for the full-length transcriptome. All consensus transcripts were merged and assembled to obtain a non-redundant transcript set using StringTie (Pertea et al., 2015). The assembled transcripts were compared with the reference genome using gffcompare (v0.12.1; Pertea & Pertea, 2020). After filtering transcripts with single exon, transcripts with class codes (”u”, ”x”, ”i”, ”j”, ”o”) and length longer than 200 bp were defined as novel transcripts. The novel transcripts were further classified into isoforms of novel genes (”u”, ”x”, ”i”) and novel isoforms of known genes (”j”, ”o”).
Ballgown was used to estimate transcript abundance (v2.26.0; Pertea et al., 2016). DEGs were identified using DEseq2 with |log2FoldChange| > 1 and adjusted p < 0.05 (Liu et al., 2021). Enrichment analyses of Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) for DEGs were performed using clusterProfiler with the cutoff of p < 0.05 (Yu et al., 2012).
AS events and fusion genes were identified. SUPPA2 (v2.3) was used to generate seven main types of the local AS events, including retained intron (RI), alternative 5’ splice-site (A5), alternative 3’ splice-site (A3), skipping exon (SE), alternative first exon (AF), alternative last exon (AL), and mutually exclusive exons (MX) (Trincado et al., 2018). Salmon (v0.13.1) was used to calculate the transcript abundance (TPM), which was then used to calculate the value of percentage spliced-in (PSI) by SUPPA2 (Patro et al., 2017). Furthermore, diffSplice was applied to identify differentially alternative splicing events with the cutoff of p < 0.05 (Hu et al., 2013). Candidate fusion genes were identified using the ToFU (fusion_finder.py) in cDNA_Cupcake program (v29.0.0, https://github.com/Magdoll/cDNA_Cupcake).
Widely targeted metabolome analysis
The samples at 24 hpi were further selected for widely targeted metabolome analysis, which included GSB-inoculated samples of PI511890 (MRT) and Payzawat (MST), and controls of PI511890 (MRC) and Payzawat (MSC). Extraction, detection, identification, and quantification of metabolites were performed according to the reported methods (Chen et al. 2013). Briefly, the freeze-dried sample was crushed using a mixer mill (MM 400, Retsch) with a zirconia bead for 1.5 min at 30 Hz, and approximately 100 mg of powder was extracted with 70% aqueous methanol. The sample extracts were analyzed using an ultra-performance liquid chromatography-electrospray ionization-mass spectrometry (UPLC-ESI-MS/MS) system (UPLC, Shim-pack UFLC SHIMADZU CBM30A system; MS, Applied Biosystems 4500 Q TRAP). The analytical conditions were as follows, UPLC: column, Agilent SB-C18 (1.8 µm, 2.1 mm × 100 mm); column temperature, 40°C; flow rate, 0.35 mL/min; injection volume, 4 µL. LIT and triple quadrupole (QQQ) scans were acquired on a triple quadrupole-linear ion trap mass spectrometer (Q TRAP). Instrument tuning and mass calibration were performed with 10 and 100 µmol/L polypropylene glycol solutions in QQQ and LIT modes, respectively. A specific set of MRM transitions were monitored for each period according to the metabolites eluted within this period.
Based on the detected metabolites, principal component analysis (PCA) was performed to reveal the relationships among the samples using FactoMineR and factoextra packages in R. The orthogonal partial least squares-discriminant analysis (OPLS-DA) was performed to determine the DAMs with |log2FoldChange| > 1 and variable importance in project (VIP) ≥1 (Eriksson et al., 2006). Enrichment analysis for DAMs was conducted using the Metabolites Biological Role (MBROLE) (v2.0; López-Ibáñez et al., 2016).
3. Results
3.1 Growth of GSBpathogenic fungi on melon leaves
The growth process of S. cucurbitacearum on the leaves of PI511890 and Payzawat was observed at 0, 12, 24, 36, 48, 60, and 72 hpi using trypan blue staining method. The results showed that the growth process of S. cucurbitacearum consisted of germination of conidia, formation and elongation of germ tube, production of appresorium, as well as growth and spread of hyphae (Figure 1A). On the leaves of Payzawat (GSB-susceptible), germ tubes and hyphae were observed at 12 hpi, followed by appressoria at 24 hpi. The hyphae were apparently elongated at 60 hpi and even began to invade into epidermis of Payzawat leaves at 72 hpi. However, on the leaves of PI511890 (GSB-resistant), germ tubes and appressoria appeared until 24 hpi and 36 hpi, respectively. Hyphae were observed at 60 hpi, which grew slowly thereafter. These results indicated that the spore germination and hyphal growth of S. cucurbitacearum on PI511890 leaves were inhibited compared with those on Payzawat leaves. Moreover, 24 hpi was the key time point to determine the different responses of PI511890 and Payzawat to S. cucurbitacearum infection.