Introduction

Bananas including plantains are one of the most important fruits crops all over the world. From the production context, after rice, wheat and maize, bananas are the fourth most important crop and are grown in all the tropical and subtropical regions including coastal, deltaic, plain inland to hills, up to an altitude of 1800 m (Mustaffa and Sathiamoorthy 2004). According to the Food and Agriculture Organization of the United Nations,  the world’s banana production was 116 million tons during 2017–2019. It is a highly nutritious fruit crop and is a major source of carbohydrates and has high contents of potassium, vitamin B and C. Most of the modern cultivated bananas that produce edible fruits are hybrids of intra- or interspecies crosses between two diploid wild species, Musa acuminata (A genome) and M. balbisiana (B genome) both belonging to Eumusa section of the Musaceae family (Simmonds and Sheppard 1955; Heslop-Harrison and Schwarzacher 2007). M. paradisiaca (ABB) is a hybrid between M. acuminata and M. balbisiana, and the vernacular name of the cultivar in Assam is ‘Kachkal’. It is a culinary banana found in the entire North-Eastern region of India and has been reported to be relatively resistant to common banana pests and pathogens (Mishra 2000).

The global production of bananas is threatened by several biotic stresses caused by pests and pathogens. Among the fungal diseases, the Sigatoka leaf spot disease complex is one of the most devastating foliar diseases of bananas. The causal agent of Sigatoka disease comprises of three species of Mycosphaerella. Black Sigatoka is caused by Mycosphaerella fijiensis (anamorph Pseudocercospora fijinensis), Yellow Sigtoka is caused by M. musicola (anamorph P. musae) and Eumusae leaf spot is caused by M. eumusae (anamorph P. eumusae) (Carlier et al. 2000; Jones 2000; Arzanlou et al. 2007). Black Sigatoka disease is the most virulent among the three leaf spot diseases and is distributed worldwide affecting a wider range of banana genotypes (Carlier et al. 2000). But in India, Yellow Sigatoka and Eumusae leaf spot are more prevalent and severe. In the North-Eastern regions, Sigatoka disease results in significant yield loss predominantly in the cultivars of the Cavendish subgroup. Although studies are being carried out worldwide concerning host resistance in banana against Black Sigatoka disease, molecular studies for an incompatible interaction between Musa spp. and the Yellow Sigatoka pathogen is still lagging behind, particularly in India. Management of the Sigatoka disease in commercial bananas is highly dependent on the use of fungicides and cultural practices. But the use of chemicals brings collateral problems such as environmental pollution, human/animal health hazards and pathogen resistance.

In recent years, high-throughput next-generation sequencing (NGS) platforms have evolved as revolutionary tools for gaining complete insight into genomes through large scale genome and transcriptome sequencing. RNA sequencing (RNA-seq) is an affordable, accurate and high-resolution deep sequencing technology that sequences the entire transcriptome of an organism within a short period. It is being widely used for transcriptome sequencing in many plant species to understand the mechanisms of plant defense responses to biotic stresses, and to identify defense or resistance-related genes implicated in host–pathogen interactions. Such defense-related or resistance-related candidate genes could further be utilized for crop improvement through plant genetic modification and marker-assisted breeding technologies. A number of studies have been carried out in several genotypes of Musa spp. to unravel host responses to various biotic and abiotic stresses including responses to the Sigatoka pathogen, by investigating transcriptomic changes and identifying differentially expressed genes (DEGs). In a study by Passos et al. (2013) where M.acuminata was challeged by M. musicola, several defense-related genes and transcription factors such WRKY transcription factors, MAP kinase, pathogenesis-related protein1, putative chitinase, endochitinase, serine/threonine-protein kinase, NBS-LRR protein, etc. were found to be upregulated. In another study by Sun et al. (2019), where banana was infected with Fusarium oxysporum f. sp. cubense (Foc) TR4, 9612 DEGs were identified, of which 72.5% were assigned in biological process, 42.2% assigned in cellular component and 82.2% in molecular functioning. Similarly, a comparative transcriptome analysis was carried out in resistant banana ‘Pahang’ and susceptible cultivar ‘Brazilian’ infected by Foc TR4 (Zhang et al. 2019) in which the resistant genotype was found to exhibit constitutive defense responses  even before pathogen inoculation and inducible defense responses ahead of the susceptible genotype during initial stages of Foc infection. In the present study, several transcription factors particularly the NAC, MAPK and WRKY transcription factors have also been identified to be associated with stress responses in banana. Recently, Yan et al. (2021) have reported that the transcriptional activator NAC42 regulates fruit ripening in banana under oxidative stress. Previous studies have reported MaNAC42 and MusaNAC68 to be stress-related transcription factors in banana, as their overexpression in transgenic lines improved tolerance to salinity and drought stresses (Negi et al. 2016; Tak et al. 2017). Similarly, the MpSNAC67 transcription factor was also demonstrated to be a positive regulator of senescence in banana that regulates the expression of chlorophyll catabolic genes (Tak et al. 2018). Furthermore, MusaSNAC1, which modulates stomatal closure and H2O2 content in response to drought tolerance in banana, was found to have a potential role in cross talk between shoot multiplication and drought responses (Negi et al. 2018, 2021). Mitogen-activated protein kinases (MAPKs) are proteins involved in many functions such as signaling, stress response and plant development (Hamel et al. 2006). The MAPK gene MusaMPK5 has been reported to be involved in the regulation of cold tolerance in banana and is found to be associated with phosphorylation of NAC42 and SNAC67 (Tak et al. 2020). The WRKY transcription factors have several established biological roles in banana plants such as those associated with stress responses, fruit ripening, magnesium deficiency, etc. (Yang et al. 2021; Jia et al. 2022). The MusaWRKY18 was found to be strongly induced in transgenic tobacco lines in response to abiotic stresses (cold, salinity and drought) and applications of several plant growth regulators, indicating its role under multiple stress conditions (Tak et al. 2021). To devise a complete and effective management strategy against diseases, understanding the host–pathogen interactions at the molecular level is a prerequisite. Investigating such interactions can unravel the host defense mechanisms and provide information on genes involved in resistance responses. Further, elucidation of the expression patterns of such genes would help in the identification of candidate genes involved in conferring resistance. With this background, the present study was designed to undertake a comprehensive study on the banana defense transcriptome during the ‘M. paradisiaca–P. musae’ interaction, to identify specific molecular events as well as candidate genes associated with defense or resistance against the pathogen.

Materials and methods

Plant preparation, pathogen inoculation and sample collection

The banana cultivar Kachkal (M. paradisiaca, Genome ABB), which is resistant to Yellow Sigatoka disease, was chosen for the present study. We collected young and healthy banana suckers with 1–2 opened leaves from the Horticulture Orchard of Assam Agricultural University, Jorhat, and planted in pots (size 40 cm height × 35 cm width) containing mixture of compost and soil (in the ratio 1:2). The plants were grown inside a green house till complete emergence of 4–5 leaves (Supplementary file 1). Leaf samples showing typical Yellow Sigatoka leaf spot symptoms were collected from a susceptible banana cultivar M. accuminata var. Jahaji (AAA), cut into 2 cm pieces and surface sterilized using 0.1% mercuric chloride. The pathogen P. musae was isolated from infected leaves of banana identified by their typical symptoms following methodology described by Selvarajan et al. (2001). Leaves with typical lesions of Sigatoka were collected, cut into small sections of 2 cm2 sizes and dipped in sterile distilled water for 15 min. The sections were then stapled to sterile filter paper and placed on the lid of a petriplate facing 3% water agar. The plates were incubated at room temperature and the release of spores was monitored regularly under a microscope. The spores identified based on their morphology were then transferred to V8 juice agar medium and incubated at 28 °C. For microscopic detection of the fungus, a small amount of the culture was stained with Lactophenol blue and observed under a light microscope (Supplementary file 1). For molecular detection, genomic DNA of the fungus was isolated by CTAB method followed by PCR amplification using specific primers MMus137/R635 described by Johanson and Jeger (1993) (Supplementary file 1). Young plants of cv. Kachkal with 3–4 opened leaves were artificially inoculated using P. musae spore suspension prepared at a concentration of 106 spores mL−1. The suspension was inoculated into the plants by spraying on both the upper and lower leaf surfaces. The plants meant to be used as control were mock inoculated by spraying with sterile distilled water. Leaf samples were collected from the pathogen challenged and mock inoculated plants in triplicates at three time points, i.e., 24 h post inoculation (hpi), 48 hpi and 72 hpi. Samples were collected in liquid nitrogen and immediately stored at − 80 °C. The plants were subsequently monitored for development of symptoms of infection.

Total RNA isolation and RNA-seq

Total RNA was isolated from the leaf samples collected at different time points using the RNeasy Plant Mini Kit (QIAGEN). After checking the quality and quantity of the RNA samples using a Nanodrop 1000 spectrophotometer (Thermo Scientific), we pooled the total RNA extracted from three biological replicates for each treatment (control and P. musae-infected tissues at 24, 48 and 72 hpi) for RNA-seq. The samples were also checked by electrophoresing in a 1% denaturing agarose gel (Supplementary file 1). The samples collected at 48 hpi were then precipitated using 0.1 volume of 3 M Sodium Acetate (NaOAc, pH 5.5) and 100% ethanol and shipped in dry ice to Bencos Pvt. Ltd., Mumbai, for sequencing works. RNA-seq was carried out using the Illumina HiSeq 2500 platform following 2 × 101 bp chemistry. Paired-end sequencing libraries were prepared from the total RNA using the Illumina TruSeq RNA sample preparation kit (Cat. No. RS-122-2101). The sequenced raw data was processed to obtain high-quality clean reads using Sequencing Control Software HCS v2.2. The BCL (base calls) binary was converted into FASTQ utilizing the Illumina package bcl2fastq (v1.8.4). The low-quality bases and adapters were trimmed away and the high-quality filtered reads were aligned to the reference M. acuminata genome using the software Hisat2. The draft genome and its GTF files were downloaded from the NCBI database to assist the reference-guided assembly. The aligned reads were further processed using the Cufflink software package. The assembled reads were merged using Cuffmerge (a component of Cufflink). The GTF reference file was used to merge new isoforms and known isoforms and maximize overall assembly quality. The Cuffdiff program further estimated the significant changes in the expression of the transcripts. The transcripts were then classified as upregulated and downregulated based on their Generalized Fold change (GFold) value. We considered the transcripts with a difference of at least two-fold change and p-value ≤ 0.01.

Differential gene expression analysis and functional annotation of transcripts regulated during P. musae infection in M. paradisiaca

The differentially expressed transcripts of both control and treated samples were with the BLASTx program (NCBI) for homology searches against publicly available protein database Swissprot (http://www.ebi.ac.uk/uniprot/), annotated and classified on Gene Ontology (http://www.geneontology.org/) and the KEGG pathway (http://www.genome.jp/kegg/) with a threshold E value of 10–5 for functional annotation and classification. The GO terms were further classified into biological process, molecular function, and cellular components using the WEGO software. Pathview was used to generate KEGG pathways and Mapman was used to view the DEGs in specific biological processes. Venn diagrams were constructed using the online tool Venny (version 2.1.0). The biological pathways for DEGs were assigned by mapping with KEGG database by inbuilt tool of Blast2GO and KAAS annotator.

Validation of RNA-seq data by quantitative real-time PCR (qRT-PCR)

To validate the gene expression patterns obtained through RNA-seq experiments, we carried out qRT-PCR based expression analysis for all the three time points (24, 48, and 72 hpi). We synthesized first-strand cDNA using the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems, USA) and diluted (1:10) with nuclease-free water. The gene-specific primers were designed using the NCBI PrimerBlast tool (http://www.ncbi.nlm.nih.gov/tools/primer-blast/) for amplification of selected DEGs based on the functional annotation analysis, considering FC value and p-value. A SYBR green assay was performed using a QuantStudio 5 Real-Time PCR system (Applied Biosystems, USA) using three biological replicates in three technical replicates. β-Tubulin gene was taken as the endogenous control gene. A reaction volume of 10 μL was set up for each primer set with 5 µM of forward and reverse primers, 1X SYBR Green PCR Master Mix and nuclease-free water. Cycling conditions consisted of initial template denaturation at 95 °C for 5 min followed by 30 cycles of denaturation at 94 °C for 20 s, annealing at 50–60 °C (primer specific) for 20 s, extension at 72 °C for 25 s and, final extension at 72 °C for 7 min. Finally, the expression level of each gene was relatively quantified based on calculated normalized ratio using advanced relative quantification using QuantStudio5 software version v1.5.x. The mean delta delta Ct value (2−ΔΔCt value) or RQ (relative quantification) value of technical replicates was taken and presented as a histogram to illustrate the expression patterns. One-way ANOVA was tested between the RQ values obtained from qRT-PCR. Dunnett's multiple comparisons test was done and significance was tested at p-value < 0.005 to test the statistical significance of the expression parameters at different time points for individual genes.

Results

RNA sequencing and transcriptome assembly of DEGs 

Total RNA extraction was done from P. musae infected M. paradisiaca var. Kachkal samples and water-treated controls (mock treated) at different time points 24, 48 and 72 hpi. Paired-end RNA-seq of leaf samples of Kachkal at 48 hpi by P. musae (KIc) along with water-treated control (KC) were done in Illumina HiSeq 2500. The transcriptome sequencing generated a total of 43,506,340 reads from mock-treated control sample and 45,459,572 reads from the P. musae-infected sample. The ratio of reads having Phred quality score of over 20, i.e., Q20 (%) of control and P. musae-infected Kachkal samples were found to be 97.79% and 97.866%, respectively (base call accuracy 99%). The ratio of reads that have Phred quality score of over 30, i.e., Q30 (%) of control and P. musae- infected Kachkal samples were found to be 96.207% and 96.353%, respectively (base call accuracy 99.9%). The GC content of the mock-treated control and P. musae-infected samples were found to be 49.502% and 50.339%, respectively. After reference-based transcriptome assembly and differential gene expression analysis, it was observed that out of a total of 45,857 genes, 9851 genes were upregulated and 1303 genes were downregulated in the cultivar Kachkal infected with P. musae (Supplementary file 2). After applying a threshold value of fold change ± 2 and p-value > 0.01, 429 upregulated genes and 156 downregulated genes were filtered out. The results show that 17 DEGs were exclusively expressed in the control and 52 DEGs were exclusively expressed in P. musae-infected Kachkal plant. The heat map of the DEGs in M. paradisiaca upon P. musae infection is shown in Fig. 1 which depicts the vast array of genes modulated in the resistant banana cultivar upon infection at 48 hpi. The sequences were submitted to nucleotide sequence database of NCBI under Sequence Read Archive (SRA) (Accession: PRJNA493890 ID: 493890).

Fig. 1
figure 1

Heat map of differentially expressed genes (DEGs) in M. paradisiaca upon P. musae infection

Full size image

Functional annotation through gene ontology and KEGG analysis

To understand the molecular mechanism of the interaction between M. paradisiaca var. Kachkal and the Sigatoka pathogen P. musae, GO (gene ontology) enrichment analyses of the DEGs in Kachkal compared with mock inoculated control sample was carried out using the WEGO software. The upregulated genes were classified into 49 GO classes of biological process (BP), 20 classes of molecular function (MF) and 9 classes of cellular component (CC) (Supplementary file 3). Similarly, the downregulated genes were classified into 35 GO classes of BP, 28 classes of MF and 6 classes of CC (Supplementary file 4). The graphical representations of the top 20 GO categories under BP, MF and CC in which the upregulated and downregulated DEGs associated with the M. paradisiaca–P. musae interaction were enriched, are depicted in Figs. 2 and 3, respectively. The Venn diagram (Fig. 4) of the enrichment data depicts that among the upregulated genes, 51 were common between BP and MF categories, 32 were common among BP and CC, 36 are common among MF and CC, while 15 genes were common in all the 3 GO categories. In case of downregulated genes, BP and MF share 29 common genes, BP and CC share 80 common genes, CC and MF share 45 common genes, while 16 genes are commonly represented in all the categories (Fig. 4).

Fig. 2
figure 2

Graphical representation of the top 20 GO classes, under which the genes upregulated during the M. paradisiaca–P. musae interaction were enriched

Full size image
Fig. 3
figure 3

Graphical representation of the top 20 GO classes, under which the genes downregulated during the M. paradisiaca–P. musae interaction were enriched

Full size image
Fig. 4
figure 4

Venn diagram depicting GO distribution showing numbers of overlapping genes (upregulated and downregulated genes) among the classes of biological process, molecular function and cellular component

Full size image

Some important defense-related genes significantly upregulated in M. paradisiaca var. Kachkal in response to P. musae infection at 48 hpi, include LRR receptor-like serine/threonine-protein kinase, non-specific lipid-transfer protein, ethylene-responsive factor, cytochrome P450, peroxidase, phenylalanine ammonia-lyase, plant intracellular Ras-group-related LRR, PR1, germin-like protein, GDSL esterase/lipase, UDP-glucose 6-dehydrogenase, PR4, chitinase, mannose-specific lectin, polyphenol oxidase, WRKY, MAPK, Bowman–Birk-type proteinase inhibitor, zinc finger protein, O-methyltransferase and F-box/LRR-repeat protein. The G-fold change in expression at 48 hpi and specific annotations of the genes as obtained from the RNA-seq analysis are presented in Table 1. Apart from genes, several transcription factors known to be involved in biotic stress responses have been identified to be upregulated as a result of infection. The study shows that 34 transcription factors including WRKY transcription factors, ethylene-responsive transcription factors, bHLH transcription factors, MYB family transcription factors, heat stress transcription factors, GATA transcription factors, MYC transcription factors, etc. are involved in the interaction, of which 27 were overexpressed while 7 were underexpressed (Table 2).

Table 1 Defense-response genes highly upregulated during M. paradisiaca–P. musae interaction at 48 hpi
Full size table
Table 2 Transcription factors differentially expressed in the M. paradisiaca–P. musae interaction
Full size table

To elucidate the functional involvement of transcripts in metabolic pathways, the DEGs were annotated against the KEGG database. The results of the KEGG analysis depict that, the DEGs which were upregulated in response to infection were significantly enriched in 20 KEGG pathways (Supplementary file 3). The pathways in which a large number of genes were represented include metabolic pathways (286 genes), biosynthesis of secondary metabolites (173 genes), plant hormone signal transduction (47 genes), and phenylpropanoid biosynthesis (27 genes). The downregulated genes were annotated into 17 pathways (Supplementary file 4) and the highly enriched ones include metabolic pathways (286 genes), biosynthesis of secondary metabolites (173 genes), plant hormone signal transduction (62 genes), amino sugar and nucleotide sugar metabolism (46 genes). Furthermore, to visualize the functional involvement of DEGs in biological pathways administrating plant defense in M. paradisiaca in response to Sigatoka infection at 48 hpi, GO results assigned were mapped using Mapman for plant–pathogen interaction and plant hormone signal transduction pathways. The plant–pathogen interaction pathway generated from the analysis indicates several upregulated DEGs (Fig. 5) such as CDPK, Rboh, Cam/CML, Pti1, Pti6, PR1, etc. which are important factors involved in PAMP-triggered immunity. Other factors such as CERK1, PBS1, and WRKY associated with effector-triggered immunity were also found to be upregulated. All these genes are important components of the plant–pathogen interaction pathway that are associated with hypersensitive response (HR), programmed cell death (PCD) and defense-related gene induction. A map of the plant hormone signal transduction pathway depicts several genes/factors associated with signal transduction mechanisms leading to cell enlargement (AUX/IAA, GH3, SAUR), cell division (AHP, ARR), stem growth (GID, TF), stomatal closure (PYR/PYL, PP2C, SnRK2), senescence/stress response (JAR1, JAZ, MYC2), disease resistance (NPR1, TGA, PR1) (Fig. 6).

Fig. 5
figure 5

Visualization of DEGs involved in the plant–pathogen interaction pathway in M. paradisiaca upon P. musae infection at 48 hpi. White box or no color fill means no DEG was assigned to that ontology term and red represents upregulation

Full size image
Fig. 6
figure 6

Visualization of DEGs involved in the plant hormone signal transduction pathway in M. paradisiaca upon P. musae infection at 48 hpi White box or no color fill means no DEG was assigned to that ontology term and red represents upregulation

Full size image

Validation of RNA-seq data through expression profiling of defense response genes

To confirm and validate the differential expression profiles obtained from the transcriptome analysis data, we randomly selected a total of 10 defense-related DEGs for qRT-PCR analysis. The genes exhibiting defense response-related traits (based on their functions related to defense as established by previous studies) which were significantly upregulated in the resistant host upon infection were chosen for qRT-PCR validation viz. pathogenesis-related protein PR1-like (PR1), PR4, pathogen-related protein-like (PR), ethylene-responsive transcription factor (ERF), plant intracellular Ras-group-related LRR protein 1-like (Ras-LRR), WRKY transcription factor (WRKY), LRR receptor-like serine/threonine-protein kinase (LRR-STK), putative O-methyltransferase (OMT), mitogen-activated protein kinase (MAPK) and germin-like protein (GLP) (Table 3). The qRT-PCR-based expression profiling revealed the differential expression of the selected genes at three time points viz. 24, 48 and 72 hpi (Fig. 7) and also validated the transcriptome expression data. The RNA-seq study was carried out using the samples (mock inoculated control and infected) from 48 hpi and the fold change in expression was found to be analogous to the expression patterns of the genes in qRT-PCR analysis.

Table 3 Defense-related genes associated with M. paradisiaca defense against P. musae infection, selected for qRT-PCR showing fold changes in expression (at 48 hpi) as observed from the RNA-seq and qRT-PCR analyses
Full size table
Fig. 7
figure 7

qRT-PCR-based differential expression analysis of selected defense-related genes upregulated in M. paradisiaca in response to infection by P. musae: (i) PR1 (ii) PR4, (iii) PR-like, (iv) GLP (v) OMT, (vi) Ras-LRR, (vii) LRR-STK, (viii) UGP, (ix) ERF, (x) WRKY, across 24, 48 and 72 hpi using (xi) Tubulin (as internal control) (Y axis –RQ obtained from RT-PCR analysis of the genes; X axis samples at different conditions viz. control and infected leaf tissue at 24 hpi, 48 hpi and 72 hpi)

Full size image

The qRT-PCR study revealed that the PR1 gene was significantly upregulated by 2.064-fold within 24 hpi and the expression further increased at 48 hpi (2.867-fold). However, its expression rapidly declined at 72 hpi. In case of PR4, there was no significant change in expression at 24 hpi but the gene was significantly highly upregulated at 48 hpi by 3.949-fold. Interestingly, its expression also declined quite rapidly at 72 hpi. Similarly, a pathogen-related protein showed a slight induction after 48 hpi by 2.326-fold but was rapidly and highly upregulated by 6.314-fold at 72 hpi as compared to control (Fig. 7).

In case of ERF (Ethylene-responsive transcription factor ERF037-like), there was an increase in expression by 1.903-fold at 24 hpi, and 1.799-fold at 48 hpi, but the gene was highly upregulated by 2.911-fold at 72 hpi. On the other hand, the Ras-group-related LRR protein 1-like gene (Ras-LRR) did not show significant upregulation at 24 hpi but its expression increased at 48 hpi significantly by 2.152-fold. Later at 72 hpi, the gene was highly upregulated by 5.018-fold. A similar expression pattern was also seen in case of WRKY. The WRKY transcription factor showed no significant change in expression at the initial time point but later, at 48 hpi, it was highly upregulated by 2.986-fold. At the last time point (72 hpi), WRKY showed a sudden spike in expression by 7.533-fold (Fig. 7).

Further, in case of LRR-STK (LRR receptor-like serine/threonine-protein kinase) significantly high upregulation by 21.95-fold was observed at the initial time point. In the later time points, at 48 and 72 hpi, although its expression was observed to be gradually declining, it remained significantly higher as compared to control (12.324- and 10.39-fold during 48 hpi and 72 hpi, respectively). In contrast, although no significant change in expression was observed in OMT (O-methyltransferase ZRP4) at 24 hpi, the expression gradually increased at the later time points. OMT showed significantly high upregulation at 48 hpi (4.592-fold) and 72 hpi (7.917-fold). Similarly, MAPK showed no upregulation at 24 hpi but at 48 hpi, the gene was significantly upregulated by 3.398-fold and was highly upregulated by 9.024-fold at 72 hpi. Finally, in case of Germin-like protein (GLP), although no change in expression was seen initially, the gene was significantly upregulated by 2.445- and 4.384-fold at 48 hpi and 72 hpi, respectively (Fig. 7).

The investigation, thus, established that a resistant host can quickly deploy defense strategies within 24–48 h of infection. Majority of the defense-related genes selected for the validation study were found to be highly and significantly upregulated within these time points. Moreover, for most of the genes, the expression remained consistently high till the later time point of 72 hpi. These results, thus, indicate that the early defense-related genes identified from the study are important not only for pathogen recognition and perception but also for activation of defense-related genes and persistence of defense in M. paradisiaca against  P. musae invasion.

Discussion

High throughput next generation sequencing technologies such as RNA-seq can help understand mechanisms underlying resistance responses in plant–pathogen interactions. Several studies have been carried out to study the transcriptomic changes in Musa germplasms against Sigatoka leaf spot disease (Portal et al. 2011; Passos et al. 2013; Rodriguez et al. 2016), however, the defense transcriptome of M. paradisiaca in response to Sigatoka disease has not been investigated so far. In the current investigation, the transcriptome sequencing of M. paradisiaca var. Kachkal, during Sigatoka infection caused by P. musae was carried out using the Illumina platform. The differential gene expression analysis revealed that a vast set of genes were upregulated while many genes were downregulated in Kachkal at 48 hpi as a result of P. musae infection. The RNA-seq analysis also depicted that a wide range of important defense-related genes are upregulated in ‘Kachkal’ in response to P. musae infection with functions related to oxidoreductase activity, response to biotic stimulus, protein kinase activity, hydrolase activity, metabolic pathways, catalytic activity, mannose binding, sequence-specific DNA binding, MAPK cascade, etc., all of which are responsible for either mediating or activating resistance responses in the host. The KEGG enrichment analysis shows that the upregulated genes were most highly represented in ‘metabolic’ and ‘biosynthesis of secondary metabolites’ pathways. Additionally, ‘plant hormone signal transduction’, ‘plant–pathogen interaction’ and ‘phenylpropanoid biosynthesis’ pathways were also highly significantly enriched indicating their involvement in imparting resistance to the pathogen. The analysis also indicated the involvement of several classes of transcription factors in the interaction such as WRKY, ERF, bHLH, MYB, most of which are known to be involved in regulating downstream defense gene activation.

The qRT-PCR study evaluated the expression patterns of the selected defense response genes in the host banana plant Kachkal in response to infection by P. musae. The study also corroborated the RNA-seq results, as the fold change in expression patterns of the selected genes at 48 hpi, obtained from qRT-PCR analysis, was found to be comparable to those obtained from the transcriptome data. The genes selected for the validation study have defense associated functions. The PR proteins, induced during pathogen attack as well as other stresses, belong to multi-gene families. Several PR proteins have been identified to exhibit plant defenses against a range of plant pathogens including fungi (van Loon and van Strien 1999), and are a vital part of plant innate immunity. PR1 is one of the most abundant groups of PR proteins mostly induced by pathogen attack or salicylic acid and is used as a marker for pathogen-induced systemic acquired resistance (SAR) (van Loon et al. 2006). Another PR protein, PR4, is reported to have antifungal activity and is induced by pathogen infection or SA, JA and ethylene (Van Loon and Van Strien 1999; Elvira et al. 2008). PR4 can disrupt cell polarity by binding to chitin present in the fungal cell wall (Portal et al. 2011). In the present study, a rapid but slight induction in expression of PR1 was observed in the resistant cultivar immediately after infection at 24 hpi which further enhanced at 48 hpi as revealed by qRT-PCR. However, its expression declined rapidly at 72 hpi. In case of PR4, although no significant change in expression was seen at 24 hpi but the gene was highly upregulated at 48 hpi. Interestingly, just like PR1, the expression of PR4 also rapidly declined at 72 hpi. The pathogen-related protein also did not show any significant upregulation at 24 hpi but was upregulated at 48 hpi. At 72 hpi, it showed a quite rapid and high upregulation as compared to control. These results are in agreement with Rodriguez et al. (2016) who reported that a resistant variety exhibits an earlier and stronger gene expression as compared to a susceptible one. They had conducted microarray analysis in resistant M. acuminata ssp. Burmannicoides ‘Calcutta 4’ after inoculation with M. fijiensis and detected high levels of relative gene expression in PR1 and PR4 at 24 hours after inoculation. In contrast, these genes were downregulated in the susceptible genotype ‘Williams’ (Rodriguez et al. 2016). Zhang et al. (2019) carried out comparative transcriptome analysis in corms of resistant banana ‘Pahang’ and susceptible Cavendish cultivar ‘Brazilian’ infected by Foc TR4. They observed that the resistant genotype exhibited constitutive defense responses before pathogen inoculation and inducible defense responses before the susceptible genotype at the initial stages of Foc infection. They further demonstrated that several defense response DEGs including PR1 were upregulated rapidly within one day post inoculation. Earlier, Bai et al. (2013) had also reported that the resistant cultivar ‘Yueyoukang 1’ could trigger a much faster defense response against Foc TR4 infection compared to ‘Brazilian’.

The Ethylene Response Factor (ERF) superfamily has been reported to play important roles in response to several stresses including pathogen infection (Mizoi et al. 2012). The ethylene regulated expression of secondary transcription factors like ERF can eventually regulate the expression of downstream defense-related genes in abiotic as well as biotic stresses (Berrocal-Lobo et al. 2002; Zhang et al. 2004; Zhang et al. 2007). In the present investigation, it was found that the expression of ERF increased immediately at 24 hpi as shown by qRT-PCR analysis. At 48 hpi, its expression was still significantly high as compared to the control and was further upregulated at the late time point. This gradual increase in the expression of ERF protein over time in the resistant host upon infection by P. musae indicates its possible role in conferring defense. McGrath et al. (2005) have observed that ERFs show altered transcript levels following infection with the incompatible pathogen Alternaria brassicicola as early as at 6 hpi. They have further identified that 10 members of the APETALA2/ethylene response factor family were induced by F. oxysporum as well as methyl jasmonate treatments. In tomato, the AP2/ERF transcription factors were found to induce defense-related genes, which when overexpressed in transgenic Arabidopsis, led to resistance against Erysiphe orontii and P. syringae (Gu et al. 2002). Again in transgenic wheat, overexpression of ERF has been found to confer enhanced resistance to fungal pathogen Rhizoctonia cerealis (Zhu et al. 2014).

The WRKY transcription factors, along with other signaling proteins, regulate the defense-related signal transduction pathways as well as the overall expression of defense-response genes (Mao et al. 2011). Several studies have reported and established the functions of WRKY in plant resistance responses particularly in response to fungal pathogens (AbuQamar et al. 2006; Xu et al. 2006; Zheng et al. 2006). Our study indicates strong involvement of WRKY in mediating defense-related signaling in response to pathogen infection. Although no significant change in expression of WRKY was noticed in Kachkal at 24 hpi upon infection by P. musae, it upregulated at 48 hpi, and its expression increased further at 72 hpi. This is in equivalence with the study by Aamir et al. (2018) which reported that F. oxysporum f. sp. lycopersici challenged tomato shows increased expression of two WRKY genes (SolyWRKY33 and SolyWRKY37) at 48 h. Moreover, the expression of these genes was lesser at the earlier phase of infection (0–24 h) as compared to that in 48 h, which also corroborates our results. In Brassica rapa, the relative expression of WRKY genes was recorded to be highest on the 6th day post inoculation by fungus F. oxysporum sp. conglutinans and on the 7th day post inoculation by bacterium Pectobacterium carotovorum sub. sp. carotovorum (Kayum et al. 2014). In banana, Niu et al. (2018) have reported the upregulation of MaWRKY50 against F. oxysporum infection in a resistant cultivar, both before and after pathogen infection. In contrast to these reports, Tripathi et al. (2019) reported that the expression of nine WRKYs transcription factors was suppressed in response to infection by the bacterial pathogen Xanthomonas campestris cv. musacearum in the Banana Xanthomonas Wilt-resistant genotype indicating their involvement as negative regulators of defense.

Plant innate immunity is composed of pathogen-associated molecular pattern (PAMP)-triggered immunity (PTI) and effector-triggered immunity (ETI). PTI is initiated by the recognition of the molecular structure of pathogens and it often activates the downstream mitogen-activated protein kinase (MAPK) signaling cascades leading to expression of a range of defense genes. MAPKs can activate and regulate the expression of many defense-related or resistance-related genes through the phosphorylation of proteins and transcription factors (Zhang and Klessig 2001). Wang et al. (2017) have demonstrated that the MAPK signaling cascade participates in multiple stress signaling in bananas. They identified 10 MAPKK and 77 MAPKKK genes in the banana genome involved in tissue development, fruit development and ripening, and response to abiotic stress. In the present study, it was observed that MAPK was significantly upregulated in Kachkal in response to P. musae at 48 hpi as revealed by qRT-PCR. At 72 hpi, MAPK expression increased to a great extent. These results indicate the importance of the MAPK cascade in mediating resistance to the pathogen. A transcriptome analysis involving two contrasting barley genotypes in response to Blumeria graminis f. sp. hordei infection have reported that several DEGs involved in the MAPK signaling pathway are upregulated at 24 and 48 hpi in the resistant genotype Feng 7 (Li et al. 2019). Similarly, Sun et al. (2019) have also reported the upregulation of a MAPK pathway gene in response to Fusarium wilt infection in the resistant banana cultivar ‘Guijiao 9’. Studies in sugarcane and soybean have also reported a peak in expression of MAPK genes in response to infection by fungal pathogens at 72 hpi, particularly in resistant cultivars (Wu et al. 2013; Lanubile et al. 2015).

Plants have pattern recognition receptors (PRRs), belonging to classes of receptor-like kinases (RLKs) which can recognize conserved molecular signatures of invading pathogens and initiate PTI. The leucine-rich repeat (LRR)-RLKs are the largest group of plasma membrane proteins that are known to be involved in pathogen perception and recognition and mediating the downstream responses. The Ras-group are a class of intracellular LRR proteins that interact directly with Ras GTPases, and are involved in defense signal transduction in animals, lower eukaryotes and plants (Claudianos and Campbell 1995). In the present investigation, the Ras-group-related LRR protein 1-like gene did not show significant upregulation at 24 hpi but its expression was significantly induced at 48 hpi and later, was highly upregulated at 72 hpi. A receptor-like serine/threonine-protein kinase (LRR-STK) was found to be one of the most highly upregulated gene in the study, thus indicating its possible involvement in inducing defense responses against the pathogen. The gene was highly upregulated at all the time points, however, the expression showed a decreasing trend. Several studies in bananas have reported differential upregulation of LRR receptor-like serine/threonine-protein kinases with higher expression levels in the resistant cultivar compared to the susceptible in response to infection by Foc (Niu et al. 2018), M. musicola (Passos et al. 2013) and X. campestris cv. musacearum (Tripathi et al. 2019) establishing the role of these LRR kinases in mediating pathogen resistance.

The plant O-methyltransferases (OMTs) enzymes, involved in the biosynthesis of several compounds essential for plant defense and growth, methylate the oxygen atom (O-methylation reaction) of plant defense-related secondary metabolites such as alkaloids, flavonoids, phenylpropanoids etc., thus participating in lignification, stress tolerance and disease resistance (Lam et al. 2007). In pea, alfalfa, medicago and chickpea, the O-methylation of isoflavonoids leads to the biosynthesis of defense chemicals called ‘phytoalexins’, which are essential for disease resistance (Liu et al. 2006). Yang et al. (2017) have reported that a gene coding for a caffeoyl-CoA O-methyltransferase (ZmCCoAOMT2) is associated with the phenylpropanoid pathway and lignin production confers quantitative resistance to southern leaf blight and gray leaf spot disease in maize. The present investigation also establishes a possible role of an OMT in conferring resistance to P. musae infection in the cultivar Kachkal, since it was found to be highly and significantly upregulated at 48 hpi as well as 72 hpi. Our findings are consistent with several other studies that have demonstrated that OMTs are differentially expressed in response to infection by fungal pathogens such as Foc TR4 (in banana) (Zhang et al. 2019) and Sclerotinia sclerotiorum (in oilseed rape) (Jian et al. 2018) within 48–72 hpi.

Germin-like proteins (GLPs) have a crucial role in basal resistance and are induced due to pathogen infection or application of disease resistance-associated chemicals such as H2O2, SA, and ethylene (Zimmermann et al. 2006). These proteins function as cofactors of plant cell wall reinforcement through facilitating cross-linking of the cell wall. This leads to the generation of H2O2 which in turn activates the SA/JA signaling cascades leading to the expression of defense genes and induction of SAR. Several studies have established the involvement of GLPs in enhancing defense responses and broad spectrum disease resistance to many pathogens (Breen and Bellgard 2010; Liu et al. 2016). The expression study revealed that in Kachkal, a GLP has a potential role in conferring resistance to P. musae as it was significantly upregulated during 48–72 hpi. Similar expression patterns of induction of GLP at 72 hpi were also reported by Lanubile et al. (2015) in a soybean cultivar resistant to F. oxysporum. However, a study by Tripathi et al. (2019) on disease resistance of M. balbisiana against banana Xanthomonas wilt disease, found that GLP was activated in both BXW-susceptible and BXW-resistant genotypes at 12 hpi and 48 hpi, respectively.

The present study, thus, reveals that a number of genes and transcription factors are induced in M. paradisiaca immediately after infection by P. musae which are important for pathogen perception and activation of defense responses in the host. From the qRT-PCR results, the genes such as PR-like, GLP, OMT and LRR-STK and the transcription factors ERF and WRKY were observed to be highly induced within 24-48 hpi. Moreover, their expression was found to remain consistently high till the last time point under study, thus indicating their roles in defending the fungus attack. Therefore, these genes can be further studied to be used as potential candidates for engineering P. musae resistant banana varieties using biotechnological approaches.

Conclusion

The study aimed to investigate the defense-related transcriptomic responses triggered in the resistant cultivar M. paradisiaca var. Kachkal in response to infection by the Sigatoka pathogen P. musae, for the first time. We further envisaged identifying a set of defense-related genes modulated as a result of the infection with the hypothesis that their expression should be relatively high during the initial stages of infection (24 and 48 hpi) to provide the host plant a long-lasting and broad spectrum of resistance. The transcriptomic as well qRT-PCR analysis established the fact that a resistant host tries to initiate and establish defense response by deploying an array of genes and transcription factors involved in a wide range of biological processes and pathways. The results also indicate that the early defense-related genes identified from the study are important not only for pest/pathogen recognition and perception but also for activation and persistence of defense in M. paradisiaca. These genes would serve as important potential candidates for future banana genetic improvement programs for the development of disease-resistant varieties.