Comparative metabolites analysis of resistant, susceptible and wild rice species in response to bacterial blight disease

Abstract

Globally, rice bacterial blight disease causes significant yield losses. Metabolomics is a vital tool for understanding this disease by analyzing metabolite levels and pathways involved in resistance and susceptibility. It enables the development of disease-resistant rice varieties and sustainable disease management strategies. This study has focused on the metabolic response to bacterial blight disease in three rice varieties: the near isogenic rice line IRBB27, wild rice (Oryza minuta-CG154:IRGC No. 93259, accession No. EC861737), and the susceptible control IR24. However, detailed metabolomics studies in wild rice remain largely unexplored. So, metabolic analysis with untargeted liquid chromatography mass spectrometry analysis (LC–MS/MS) was performed at various time points, including pre infection and post infection at 12 h and 24 h with Xanthomonas oryzae pv. oryzae (Xoo). In this study, a total of 6067 metabolites were identified. Pre-infection stage of the susceptible, resistant, and wild rice had 675, 660, and 702 identified metabolites, respectively, but these numbers were altered at post-infection stages. Various defense-related metabolites, including amino acids, flavonoids, alkaloids, terpenoids, nucleotide derivatives, organic acids, inorganic compounds, fatty acid and lipid derivatives have been identified. PCA and PLS-DA plots revealed differences in the metabolome among susceptible, resistant, and wild genotypes, suggesting distinct metabolic profiles for each. In this study, we found 149 metabolites were upregulated and 162 downregulated in the wild type (CG154) compared to the susceptible cultivar (IR24). Similarly, 85 metabolites were upregulated and 92 downregulated in the resistant near isogenic line (IRBB27) compared to IR24, while 156 were upregulated and 149 downregulated in CG154 compared to IRBB27. Key metabolites, including flavonoids, terpenoids, and phenolic compounds, showed significantly higher levels (P ≤ 0.01) in resistant varieties. These identified defense metabolites could serve as potential biomarkers for bacterial blight resistance in rice. The findings from this study have important implications for the development of new rice cultivars with tolerance to bacterial blight disease.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-025-06154-0.

Introduction

Rice (Oryza sativa L.) is one of the very important crops consumed worldwide [1]. It is a primary staple food of Asian and African countries and a crucial calorie source of half of the global population. While demand from poverty-stricken populations is rising, yield and total production have stagnated in major rice-producing countries. The global rice yield is plagued by several microbial diseases caused by bacteria, fungus, nematodes etc. Bacterial blight disease caused by Xanthomonas oryzae pv. oryzae (Xoo) is the most devastating disease which can reduce the production upto 60–70% while during severe infection the yield loss can increase up to 81% [2]. Catering to the needs of a rapidly growing global population presents a significant challenge to ensure sufficient food supply. To achieve this, agriculture must adopt compatible strategies to promote sustainable rice production [3] understanding the host–pathogen interactions are pivotal to design better rice production strategies.

Generally, plant develops several multiple resistance strategies to combat themselves from the pathogen. Plants have evolved a sophisticated innate ability to defend against the attacks [4]. Their immune response involves a two-layered system: the first layer relies on cell surface-localized pattern recognition receptors (PRRs) that detect pathogen-associated molecular patterns (PAMPs) triggering a relatively weak immunity known as PTI (Pattern-Triggered Immunity). PTI encompasses various responses, including reactive oxygen species (ROS) production, increased calcium concentration, callose deposition, phytoalexin release, and mitogen-activated protein kinase (MAPK) activation [3]. Moreover, Xoo can suppress plant immune responses using it’s effectors secreted through the type III secretion system (T3SS) to rice cell. These effectors target plant proteins involved in immunity [5]. The TALE effectors secreted from Xoo interact with susceptible genes of the rice cell to suppress rice immunity. Rice plants counter by recognizing effectors with resistance genes (R genes), leading to effector-triggered immunity (ETI) and a strong defense response called the hypersensitive response (HR), characterized by rapid cell death and local necrosis [6]. Because of the adaptability or pathogen-avoiding response, the metabolic response varies in the rice plants. Thus, understanding the complex host–pathogen interplay requires a holistic approach wherein host responses at each functional level can be captured. Currently omics approaches has emerged as a powerful tool to decipher the systemic responses of host–pathogen interactions. Omics captures responses at different functional levels of a system such as genomics, metagenomics, transcriptomics, proteomics, volatilomics, spectranomics, ionomics and phenomics have been employed successfully to study plant-pathogen interaction system [7].

Omics approaches such as genomics, transcriptomics, metabolomics and proteomics were successfully implemented in rice crop against the bacterial blight disease [8, 9]. Despite the potential benefits of other omics approaches in comprehending plant defense responses, metabolomics holds special importance for a deeper understanding because of its inherent characteristics. As it includes the fact that metabolites are the byproducts of gene expression and that they give an instantaneous picture of metabolic processes that directly impact phenotype. This study focuses on the metabolomics approach, a well-established and rapidly growing technology. It excels in identifying responsive metabolites and their changes in disease response through high-resolution liquid chromatography linked to mass spectrometry (LC–MS), valued for its sensitivity and extensive metabolic coverage [10].

Metabolomics is the study of all small molecules (≤ 1000 Da) in an organism during environmental and genetic changes, reflecting gene regulation and providing a biochemical description of the phenotypic status [8]. Metabolomics is a versatile approach widely used in diverse research fields, such as abiotic and biotic stress studies, functional genomics, biomarker discovery, biotechnology, and integrative systems biology [9]. It uncovers the specific defense metabolites indicative of pathogen, pest, or environmental disturbances, leading to a comprehensive metabolic changes at various levels. So, use of such altered metabolites in the respective resistant rice cultivar will be the most effective method to control and manage the Xoo attack. Plant synthesizes specialized metabolites to adopt the several disease caused by different pathogens. Plant metabolites such as proline, flavonoids, saponins, phenylpropanoids, glucosinolates, sorbitol, mannitol, alkaloid compounds, terpenoid compounds, amino acid derivatives etc. have been involved in biotic stress mechanism [11]. Potential of identified metabolites has been investigated by using new advanced analytical strategies and tools to recognize defense-responsive metabolic compounds. Metabolomics analysis was done in several crops with resistant and susceptible cultivar in biotic stress condition such as capsicum (Colletotrichum sp.) [12], Soybean (Phytophthora sojae) [13], tomato (Ralstonia solanacearum) [14], rice (Nilaparvata lugens) [15], cotton (Verticillium dahliae) [16], sorghum (Colletotrichum sublineolum) [17], sugar beet (Cercospora beticola) [18], groundnut (Alternaria alternata) [19, 20] etc. Currently, there is limited available information regarding the metabolic responses triggered by Xoo infections in rice. It has been reported that changes of metabolomics such as fatty acids, glutathione, lipid, xanthophylls, acetophenone etc. were greatly affected in transgenic resistant rice variety expressing Xa21 receptor TP309_Xa21 than susceptible variety TP309 after the infection with PXO99 under a single time point [21]. Collectively, integrating metabolomics and transcriptomic analysis provides an insight on the comprehensive biochemical and genetic responses of a Xa23-resistant rice variety CBB23 to bacterial leaf blight pathogens, deciphering resistance mechanisms for the improvement of more durable rice cultivars [22]. A study on black rice cultivars ‘Melik’ and ‘Pari Ireng’ revealed that it has exhibited higher concentrations of primary metabolites and enhanced total phenolic and flavonoid concentrations compared to white rice cultivars after infection with Xoo, indicating distinct chemical defense strategies in pigmented rice against bacterial leaf blight (BLB) [23]. To date, no studies have reported metabolomics analysis comparing susceptible, resistant, and wild rice relatives, which are reservoirs of numerous R genes conferring resistance to various Xoo strains. The present study addresses this by analyzing the metabolomics profiles of these rice genotypes to uncover key compounds linked to BLB resistance.

Resistant rice varieties provide more economic benefits and are effective ways to avoid the bacterial blight disease. Development of resistant varieties with major resistance (R) genes have proven to be very effective in controlling bacterial blight disease. The choice of R gene depends on the prevalent Xoo pathotype stain. Rice R-gene Xa27 plays a major role in providing disease resistance response against Xoo. Avrxa27, a specific TALE (Transcription Activator Like Effector), released by Xoo enters the cytoplasm of rice cells through the type III secretion system. The R-gene Xa27, also referred to as the executor R gene, functions by deceiving this TALE, which in turn activates the host resistance [24]. In this study, we have used three rice lines near isogenic line-IRBB27 (resistant), CG154 (wild rice O. minuta) and IR24 (susceptible). It has been reported that Xa27 R gene was derived from wild rice O. minuta from chromosome 6. In a previous study, a group of researchers transferred a resistance (R) gene from wild rice (Oryza minuta) to cultivated rice (Oryza sativa L.) and this mapped gene was named as Xa27 [25]. Then Xa27 R gene was isolated through map-based from the near isogenic rice variety IRBB27 [26]. IR24 is the susceptible genotype having recessive xa27 allele whereas IRBB27 is the resistant genotype having dominant Xa27 allele. Thus the metabolic differences between resistant, susceptible and wild rice will provide insights for creating stable resistant rice. Metabolites are the sensitive indicators of the physiological status of plants. The presence or abundance of specific metabolites in a particular condition of infection can reflect a particular traits or characteristics in plants. Metabolites can be used as markers for quality traits and disease resistance, enabling breeders to select and advance plants with desirable traits, resulting in crop varieties that are better suited to agriculture and food production [27]. In order, to investigate the regulation of metabolites before and after infection, we examined changes in metabolites at 12 h and 24 h after Xoo infection in resistant, wildtype and susceptible rice. The study aimed at deciphering the metabolic responses of all three types of rice plants against the Xoo infection that may provide the metabolic basis of resistance.

Materials and methods

Rice plant collection

The seeds of two rice lines IRBB27 (resistant; near isogenic line), IR24 (susceptible) and tiller of wild rice (O. minuta; CG154) were collected from ICAR-Indian Institute of Rice Research, Hyderabad, Telangana, India. All the seeds were germinated on wet cotton petri plates for initial 10–15 days. The seedlings were then transferred into larger pots of 15 cm diameter which persisted the paddy mixture of black and red soil. Tillers of wild rice were well maintained in large pots with 7–8 plants in each pot. Seedlings and tillers were grown in a greenhouse under controlled conditions with 25–28 °C temperature and 80% humidity, 16 h light and 8 h dark day-night cycle. Triplicate sets of plants were used for the experiment.

Xoo culture preparation

IXO20, a highly localized virulent strain of Xoo in Hyderabad, Telangana, was collected from the Indian Institute of Rice Research, Hyderabad, Telangana, India. The bacterial cultures were cultured for 72 h at 28ºC on modified Wakimoto's medium. Following that, the bacterial cultures were suspended in sterile, double-distilled water with concentrations of 1 × 10⁸–1 × 10⁹ CFU/ml with an OD600 value of 0.1–0.2.

Plant inoculation and sample collection

The infection was carried out during the booting stage of the plant life cycle in the resistant, susceptible, and wild plants. The control sample leaves were collected prior to the bacterial culture infection. Four leaves from each plant were inoculated by the syringe infiltration method on the adaxial surface of the leaves at six in the morning. Further samples of infected leaves were collected at 12 and 24 h following the infection. Two biological replicates of each leaf samples were used in this study. The collected leaves were promptly frozen in liquid nitrogen to stop the current metabolic process, and they were then kept in a freezer at -80ºC until for further extraction of metabolites for upcoming experiment.

Extraction of leaf metabolites

Sample extractions of susceptible (IR24), resistant (IRBB27) and wild (CG154) were performed. Initially, 600 mg of leaf samples both infected and control were weighed and grounded using liquid nitrogen. Subsequently, 1 millilitre of 80% methanol was added to the powdered samples. The samples containing the methanolic mixture solution were incubated for 30 min at 30ºC in a shaker with ideal shaking conditions. After centrifuging the slurry solution for five minutes at 14,000 rpm, the supernatants were collected and passed through a 0.2 μm filter column by following the protocol [28]. This step was repeated twice to accumulate more metabolites. The filter solution of the above process has the final metabolites stored at -80℃ till analysis.

Analysis of liquid chromatography-high resolution tandem mass spectrometry (LC–MS)

The filtered solution with 10 μl volume was transferred into micro-LC vials and injected into a LCMS-8045 mass spectrometer (Shimadzu Corp., Kyoto, Japan) that interfaced to a Shimadzu triple quadrupole LCMS. The electrospray ionization (ESI)–MS analysis was performed in both positive and negative ion modes. Full-scan mass spectra were acquired over a mass range of m/z 50–2000. During the LC–MS analysis, specific chromatography conditions were employed. Separation of metabolites was achieved by utilizing a mobile phase system of 0.04% acetic acid in water (A), and 100% acetonitrile (B). Chromatographic separation was performed through a CLC0181 column (Shimpack Gist C18 75*4.0 mm, 3 μm). The diluent used was methanol which facilitates the optimal elution and the separation was conducted at a flow rate of 0.8 ml/min with a column oven temperature of 30 °C. Additionally, the sample cooler was set at 15 °C to maintain stability. The LC program spanned a run time of 45 min for a refined separation. In tandem with these chromatographic conditions, the MS parameters played a crucial role. The acquisition time was synchronized at 45 min, interfacing with ESI and operating in both positive and negative scan modes. For effective nebulization, a gas flow of 3 L/min was employed, supported by heating gas flow 10 L/min. The interface temperature was maintained at 300 °C, with a drying gas flow of 10 L/min, while the dynamic equilibrium was upheld by a DL temperature of 250 °C, collectively facilitating efficient ionization and seamless transfer of analytes from liquid to gas phase for accurate molecular characterization. These precise parameters collectively facilitated the comprehensive analysis of the sample, enabling accurate compound identification and quantification.

Metabolomics-based data detection, interpretation, processing and multivariate data analysis (MVDA) in statistical methods

The preprocessing of raw LC–MS data was conducted using MZmine 2.53 (http://mzmine.github.io/), an open-source software designed for LC–MS data processing. MZmine 2.53, developed in Java for platform independence, facilitated mass feature detection. The initial steps included importing and cropping MZML files to retain a retention time range of 0–45 min. Mass detection was performed, chromatograms were built and peaks were aligned using the respective algorithms provided by MZmine 2.53. For metabolite identification, KEGG/Plantcyc compound databases were employed. Resulting data was exported in CSV format to facilitate further analysis, including the generation of analysis plots. Data processing encompassed normalization, where raw data was transformed for compatibility with MetaboAnalyst 5.0. Peak intensity tables were uploaded in CSV format, followed by a data integrity check that ensured accuracy and non-negativity of values. Missing value imputations were applied, replacing zeros and missing values with a small value. Data filtering was optionally performed to improve results by removing variables unlikely to contribute significantly. Data normalization was undertaken via quantile normalization, log transformation (base 10), and pareto scaling. Statistical data analysis incorporated univariate and multivariate techniques. Univariate methods included fold change analysis, t-tests and volcano plots for identifying significant features. Correlation analysis visualized feature correlations and patterns. Multivariate techniques including Principal Component Analysis (PCA) and Partial Least Squares—Discriminant Analysis (PLS-DA), revealing sample relationships and discriminating variables. Hierarchical clustering, employing diverse linkage methods, yielded heatmaps for data visualization, volcano plot by performing Fold Change (FC) analysis, t-tests, and generating for the two-group data. Pathway impact analysis employed KEGG Mapper, querying KEGG objects against pathway maps. The resultant KEGG IDs were utilized for samples' analysis and outcomes were exported.

Results

Metabolic investigation of Xanthomonas oryzae pv. oryzae induced alterations in resistant, susceptible, and wild rice

To unravel the metabolic responses of the susceptible rice accession (IR24) against Xanthomonas oryzae pv. oryzae (Xoo), with its resistant counterpart (IRBB27) and a wild rice relative (CG154; Oryza minuta) at a metabolic level, we analyzed methanolic extracts of Xoo-untreated (pre infection stage) and treated (post infection stage at 12 h and 24 h) IR24, IRBB27 and CG154 rice lines utilizing the tandem LC–MS/MS system. The metabolites identities were also verified by comparing the obtained MS/MS spectra and retention times (RTs). The chromatographic separation facilitated the discrimination of diverse classes of molecules constituting the plant metabolomes, yielding base peak intensity MS chromatograms. Visual examination of the MS chromatograms uncovered quantitative accumulation of metabolites or peak intensities and qualitative (presence/absence) variations in the metabolic profiles of treated and untreated samples. These differences indicate high-level overview of metabolic changes in the cellular metabolism and overall metabolomes of the plants under investigation (Fig S1).

Identification of metabolites from the rice varieties

The analysis revealed variations in the number of detected metabolites among different rice varieties under Xoo infection at various time points. In the susceptible variety IR24, 675 metabolites were identified in the control, while the counts decreased to 630 after 12 h and slightly increased to 672 after 24 h of infection. For the resistant accession IRBB27, the control had 660 metabolites, which increased to 669 at 12 h and further to 679 at 24 h. Notably, the wild relative CG154 exhibited 702 metabolites in the control, 687 at 12 h, and 693 at 24 h. These findings suggest dynamic metabolic responses in different rice varieties under Xoo infection, highlighting potential distinctions in the defense mechanisms and metabolic pathways associated with susceptibility, resistance and wild-type responses (Fig S2). Several defense-related metabolites were classified into different classes such as amino acids, flavonoids, alkaloids, nucleotide derivatives, organic acid derivatives etc.

Multivariate statistical analysis reveals distinct metabolic responses in rice varieties to Xoo infection

To better understand the metabolic responses of different rice varieties to Xoo infection, multivariate statistical analysis was conducted using unsupervised learning methods like PCA and HCA. PCA modeling of the metabolic data displayed distinct clustering patterns, with samples of susceptible, resistant and wild rice varieties each forming separate groups. These findings resulted the differential metabolic responses in different rice varieties when responding to Xoo infection. Data from the LC–MS/MS analysis was used to compute PLS-DA models, which helped visualize the time-dependent metabolic reconfiguration of infected plants compared to their uninfected samples. The resulting PLS-DA models highlighted time-dependent variations in the plants' metabolic profiles. This was further applied by the BPI chromatograms, which provided evidence of altered metabolism. Quantitative and qualitative variations in the number of peaks of IRBB27, IR24 and CG154 were responsible for the observed differences in the separations from the PLS-DA models, resulting the significant metabolic shifts induced by Xoo infection. To identify and validate significant biomarkers, variable importance in the projection (VIP) scores were calculated, with only scores above 1 considered as significant. Subsequent heatmap analysis of the resistant samples revealed differential accumulation of metabolites in control versus Xoo-infected plants, highlighting metabolic reprogramming induced by Xoo infection across various metabolite classes, including flavonoids, fatty acids, phenolics, amino acid derivatives, and organic acids etc.

Comprehensive metabolite profiling of susceptible (IR24), resistant (IRBB27) and wild (CG154) after infection of Xoo

Three rice varieties namely IRBB27 (resistant), IR24 (susceptible), CG154 (O. minuta) were examined at different time intervals of 12 h and 24 h following infection. From the ESI ( +) and (-) mode, a set of 6067 metabolite features were acquired and subsequently analyzed using an unsupervised Principal Component Analysis (PCA) to visualize the variation in metabolite profiles across sample groups at different timepoints. QC samples in the PCA plot serve as reference points, indicating the reproducibility and reliability of the analytical method across different batches or experimental conditions. There was 13.9% of variance has explained in the PCA score plot PC1, whereas 10.1% variance has explained in PC2. Principal component analysis (PCA) was conducted to identify the plausible metabolic variations across the samples. The PC1 kept together all the IRBB27, IR24 as well as CG154 samples at 12 h and 24 h respectively. The PCA plot showed CG154 was deviating from IRBB27 and IR24. Each stage of time period from pre infection to 12 h and 24 h post infection creates an environment for differential accumulation of metabolites and a substantial effect on different metabolic compositions. The PCA plots represented clearly that wild rice was distinct from the resistant and susceptible samples and visualizing or assuming major changes in metabolites (Fig S3a).

A comparison was conducted between the metabolites of IRBB27 and IR24, resulting in a PCA score plot where PC1 accounted for 23.7% which is the higher proportion of variance and PC2 for 10.4% which is the lower proportion of variance. This analysis revealed that the metabolite profiles of IRBB27 and IR24 were closely grouped together at pre-infection and up to 12 h post-infection, indicating a high degree of similarity. However, at the 24 h post-infection, a remarkable discrimination in metabolite profiles between IRBB27 and IR24 emerged. This difference led to a significant deviation in the metabolite composition of IRBB27 at 24 h post-infection compared to its pre-infection state, highlighting a distinct metabolic response in IRBB27 at this time point (Fig S3b). Similarly, in order to identify the key factors influencing metabolite composition, separate PCA plots were generated for CG154, the wild rice relative, along with IR24 and IRBB27. It was observed that the second component (PC2, 15.9%) distinctly separated IR24 samples at control, 12 h and 24 h from those of CG154 at similar time points. Moreover, the first component (PC1, 17.2%) exhibited less separation among the metabolites compared to PC2 (Fig S3c). Further investigation was conducted to explore the relationship between the metabolite profiles in the resistant variety IRBB27 and wild resistant relative CG154. The PCA plot for CG154 and IRBB27 revealed a variance of 16.6% for PC1 and 15.1% for PC2. In this plot, the separation of metabolites were not notably distinct, as both PC1 and PC2 placed the metabolites relatively close together (Fig S3d). Additionally, the grouping of samples by the computed PCA models suggests differential metabolic changes in various rice varieties, such as CG154, IRBB27, and IR24, both pre and post infection.

Exploring rice varietal metabolic profiles with PLS-DA in IR24, IRBB27 and CG154

PLS-DA (Partial Least Squares Discriminant Analysis) is a statistical method widely utilized in metabolomics to classify and analyze data. PLS-DA aims to maximize separation between samples in this space while minimizing regression residuals [29]. Overall, PLS-DA plots offer a robust means to classify samples based on their metabolic characteristics thus it was utilized to analyze the susceptible, resistant, and wild rice groups. In the PLS-DA graph each rice varieties were distinguished (Fig S4a) and the analysis revealed several metabolites highlighted as discriminants.

The metabolites of wild rice CG154 were clearly separated from IRBB27 and IR24 by PLS-DA (component 1 with 11.2%, component 2 with 8.1% variance), while wild vs susceptible, analysis it was observed that CG154 metabolites were distinct from IR24 by (14.9% component 1 and 10.5% component 2 variance) (Fig S4b) showing a distinguished separation of metabolites between wild and susceptible. However, when resistant and susceptible were compared, it was found that the IRBB27 variety deviated from IR24 with PLS-DA component 1 accounting for 9% of the variation and component 2 accounting for 13.5% (Fig S4c). The metabolite signal intensities identified significant metabolic differences between resistant IRBB27 and wild-type CG154 at the same post-infection stage. Most of the metabolite variability in the resistant and wild-type samples was captured by component 1 (14.1%) and component 2 (10.4%), with IRBB27 samples showing positive loadings (Fig S4d). The PLS-DA models demonstrated time-dependent shifts in plant metabolite profiles, confirmed by BPI (Base Peak Intensity) chromatograms. This altered metabolism, with distinct quantitative and qualitative variations observed in IR24, IRBB27, and CG154 control versus infected samples over time. These findings offer valuable insights into the biochemical response to infection and the dynamic metabolic rearrangements during plant growth.

Identification of discriminant metabolites in genotypic variance using Volcano Plot analysis

The PLS-DA analysis revealed that there were significant metabolites identified which were discriminated by Volcano plot and those metabolites were significantly responsible for discrimination separation of wild (CG154) and susceptible (IR24) genotype, similarly for separation between resistant (IRBB27) and susceptible (IR24). Volcano plot is one of the variate analysis methods are commonly used in exploratory data analysis. This analysis identified significant features that discriminated between conditions. For paired fold change analysis, variables were deemed significant if more than 75% of pairs per variable exceeded the FC threshold. The results, plotted on a log2 scale, facilitated comparison of absolute value changes between group means. The volcano plot highlighted significant metabolites, showing statistical significance against fold change. Metabolites with large fold changes and high statistical significance were easily identified, with the X and Y axis representing the fold change threshold and the t-test threshold respectively. Metabolites above these thresholds were considered significant candidates for further investigation, while those located at the origin (0,0) were identified as the most significant metabolites. This analysis provided clear insights into the key metabolic changes distinguishing the conditions under study (Fig. 1a-c).

Fig. 1.

Volcano plots highlighting important metabolites in different sample comparisons; (a) Volcano plot showing significant metabolites in the CG154_vs_IR24 comparison; (b) Volcano plot of CG154_vs_IRBB27 comparison; (c) Volcano plot comparing IRBB27_vs_IR24. Metabolites with a fold change greater than 2 (x-axis) and a t-test p-value less than 0.05 (y-axis) are marked with red circles. Both fold changes and p-values are log-transformed. The distance from the origin (0,0) corresponds to the metabolite’s significance, with greater distances indicating higher significance

Some significant metabolites such as oryzalexin E, ( +)-secoisolariciresinol, ITP, 2-succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexene-1-carboxylate, 3-methoxy-4',5-dihydroxy-trans-stilbene, 8-[(1R,2R)-3-oxo-2-{(Z)-pent-2-enyl} cyclopentyl]octanoate, UDP-2-acetamido-4-dehydro-2,6-dideoxyglucose etc. were upregulated in CG154 as compared to IR24. Additionally, hydrogen selenide, gentiodelphin, ancymidol, spermidine, oleamide, biliverdin, pterostilbene etc. were significantly downregulated in CG154 as compared to IR24 (Fig. 1a). Trans-Zeatin riboside triphosphate, N-(L-arginino) succinate, (2S)-Flavan-4-ol, beta-tocotrienol, myristoleic acid, (S)-dihydroorotate, sinapoyl aldehyde, norbixin, alpha-ribazole were upregulated and 3-methyl-2-butenal, L-arogenate, propane-1-ol, 2-Iminopropanoate, bergapten, (S)-hydroxydecanoyl-CoA, catechol, cysteamine, selenite, silver etc. were downregulated in CG154 than IRBB27 (Fig. 1b). CoA, CDP-choline, D-glyceraldehyde, 3-hexenol, raffinose, isoorientin, syringetin, tryptophan, 3-octaprenyl-4-hydroxybenzoate, paclitaxel, nicotianamine, L-carnitine, mirtillin etc. were upregulated in IRBB27 and shikimate-3-phosphate, 2-(2'-methylthio) ethylmalic acid, 3-oxopalmitoyl-CoA, 1,4-dihydroxy-2-naphthoate, methyl (indol-3-yl) acetate, myristoleic acid, trans-tetradec-2-enoyl-CoA, orotate, hexadecanoic acid, 1-naphthaldehyde, lead, 2-oxoglutaramate, O-phospho-L-homoserine etc. were down regulated in IRBB27 compared to IR24 (Fig. 1c).

Identification of metabolic dynamics through Venn diagram analysis

The analysis of Differentially Accumulated Metabolites (DAMs) was visualized through a Venn diagram, revealing unique and shared DAMs among and between pairs of rice varieties under treatment and control conditions. Utilizing a selection of DAMs identified by their m/z values via MS/MS analysis, we generated a total of six-way of the Venn diagrams. These diagrams vividly represented the number of DAMs shared across all time points, as well as the presence of unique DAMs specific to particular rice varieties in the context of rice bacterial blight disease. This comprehensive visualization indicates dynamic metabolic responses associated with different time points and rice varieties under bacterial blight biotic stress.

In the control condition, a total of 419 DAMs were identified as common among IR24, IRBB27, and CG154. Additionally, 80 DAMs were unique to IR24, 71 to IRBB27 and 103 to CG154, showcasing distinct metabolic profiles within each rice variety under normal conditions (Fig. 2a). Transitioning from the pre-infection stage to 12 h post-Xoo infection, the shared DAMs decreased to 404 across all three rice lines and were also identified by a reduction in the number of unique metabolites in IR24 to 71. Also, DAMs in IRBB27 increased by 4, reaching a total of 75 compared to the normal condition and DAMs were increased up to 118 in CG154 (Fig. 2b). Interestingly, the common metabolites or DAMs at 24 h post-infection increased to 432, surpassing both control and 12 h post-infection conditions. Additionally, the number of unique DAMs decreased to 62 in IR24, while in IRBB27 it was 73, and in CG154, it drastically dropped to 93 (Fig. 2c).

Fig. 2.

Differentially Accumulated Metabolites (DAMs) across rice varieties under normal and Xoo-infected conditions; (a) Venn diagrams showing DAMs across the susceptible (IR24), resistant (IRBB27), and wild rice (CG154) varieties under control conditions. A total of 419 DAMs were common among the three varieties, with 80, 71, and 103 unique DAMs in IR24, IRBB27, and CG154, respectively; (b) Venn diagrams at 12 h post infection. Shared DAMs across the three rice varieties decreased to 404. Unique metabolites in IR24 were reduced to 71, while DAMs in IRBB27 and CG154 increased to 75 and 118, respectively; (c) Venn diagrams at 24 h post infection. The shared DAMs increased to 432, surpassing the number observed under control and 12 h post-infection conditions. Unique DAMs decreased to 62 in IR24, while they increased to 73 in IRBB27 and 93 in CG154; (d) DAMs identified in IR24 at different time points. A total of 97 DAMs were detected under control conditions, which decreased to 73 after 12 h of infection and increased slightly to 83 at 24 h, but did not exceed the initial number observed under control conditions; (e) DAMs identified in IRBB27. A total of 83 DAMs were detected during the control stage, with minor changes post-infection, detecting 85 DAMs at 12 h and 88 DAMs at 24 h; (f) DAMs identified in CG154. A total of 97 DAMs were detected under control conditions, which decreased to 89 after 12 h of infection and further decreased to 77 at 24 h post-infection

When observing the impact or variation of DAMs in individual rice lines, it was identified that initially, 97 DAMs were present during the normal stage of IR24. After 12 h post-infection, the number decreased to 73, and then further increased to 83 at 24 h, although it did not surpass the initial count observed in the normal stage (Fig. 2d). Similarly, 83 DAMs were detected during the normal stage of IRBB27, with minor changes observed at 12 h and 24 h post-infection, where 85 and 88 DAMs were detected respectively (Fig. 2e). In CG154, 97 DAMs were identified in the control condition, decreasing chronologically after infection to 89 at 12 h and 77 at 24 h (Fig. 2f).

The Venn diagram provides an overview of the total significant metabolites, highlighting both common and unique responses among three rice varieties. In the susceptible variety (IR24), there is a remarkable decrease in the number of unique metabolites post-infection with Xoo. However, in the resistant variety (IRBB27), the number of unique metabolites remains relatively consistent compared to the control (pre-infection stage), with only minor fluctuations involving 4 to 5 metabolites. This suggests that there weren’t significant alterations in unique metabolites according to the Venn diagram. On the other hand, in the wild relative (CG154), being inherently wild, there was an increase in the number of metabolites observed at 12 h post-infection, followed by a subsequent decrease at 24 h post-infection. These observations indicated distinct metabolic responses across the three varieties following Xoo infection. The rapid decline in secondary metabolism in IR24 makes the plant vulnerable to Xoo, exhibiting a slower defense response compared to the resistant IRBB27. Moreover, the wild CG154 showed a number of metabolite changes over time post-infection, suggesting its potential for improved resistance through induced secondary metabolism. Furthermore, it was evident that disease resistance efficiency was influenced by these metabolic changes. The varying responses in metabolite profiles among the rice varieties showed their role in modulating resistance mechanisms against Xoo infection. These findings suggest that metabolic responses exhibit greater time dependency and genotype specificity compared to transcriptional responses.

Exploring metabolic variability by Hierarchical Clustering Analysis (HCA)

To explore differential metabolite accumulation among the samples multivariate statistical analysis has been done. The HCA (Hierarchical Clustering Analysis) revealed that the abundance of metabolites observed in substantial variation among different samples with different times of infection. PCA-extracted data trends were further explored using hierarchical clustering analyses on low-dimensional data from the PC analyses. The resulting dendrogram visually represented the hierarchical clustering, assessing whether distinct natural groupings of data points could be discerned based on their similarities or dissimilarities. Numerous metabolite compounds were identified and subsequently employed for unsupervised hierarchical clustering analysis. Each genotype exhibited a unique metabolomic profile, as demonstrated by the consistent clustering of all replications of the same genotype within distinct sub-clusters. It demonstrated the variance in Xoo infection response among susceptible, resistant and wild rice lines. Metabolic responses in Xoo-susceptible IR24 and Xoo-resistant IRBB27 lines under different treatments revealed distinct patterns, elucidating the metabolic pathways involved in the defense mechanisms against Xoo infection. In HCA, each set of replicate samples forms distinct clusters initially, which were then systematically merged until all samples unite into a single cluster, utilizing similarity measures such as euclidean distance and clustering algorithms such as average linkage. Heatmaps, with dendrograms, enhanced visualization clarity. Employing the top 25 metabolites, HCA was conducted and the resulting clustering patterns were represented through a heatmap (Fig. 3a). Moreover, hierarchical clustering analysis (HCA) has visualized the alteration of metabolites in different time periods in resistant vs susceptible as well as wild vs susceptible and wild vs resistant varieties (Fig. 3b-d).

Fig. 3.

Comparative metabolomic profiling and Hierarchical Clustering Analysis (HCA) of Xoo-susceptible, resistant, and wild rice lines. Hierarchical Clustering Analysis (HCA) was performed on the top 25 individual metabolites to illustrate variation in metabolic responses at different time points. Metabolites that are up-regulated are marked in dark red, while those down-regulated are marked in dark blue. The data were normalized using the median, log-transformed, and Pareto-scaled in MetaboAnalyst; (a) Metabolomic analysis of leaf samples from Xoo-susceptible (IR24), resistant (IRBB27), and wild (CG154) rice relative enabled the identification and annotation of key metabolites; (b) HCA comparing metabolomic responses between the susceptible variety IR24 and resistant variety IRBB27. In IR24, several metabolites show significant downregulation at 24 h post-infection, with minor upregulation at 12 h. In contrast, these metabolites are consistently upregulated in IRBB27 at both the control and 24 h post-infection stages, with variable levels at 12 h, demonstrating differential metabolic response patterns in resistant and susceptible plants upon infection; (c) HCA comparing IR24 and the wild rice CG154. The analysis reveals a stronger defense response in CG154, with key secondary metabolites significantly upregulated in response to Xoo infection compared to IR24, suggesting superior metabolic defense mechanisms in the wild relative; (d) HCA between IRBB27 and CG154. The wild rice relative CG154 exhibits a more robust defense response to Xoo infection than the resistant IRBB27, as evidenced by the substantial upregulation of key biomarker metabolites at 24 h post infection, highlighting its enhanced metabolic resilience against the pathogen

Metabolite levels in susceptible, resistant and wild rice

The metabolite glycerophosphoglycerol was initially upregulated in the control condition of IR24 but displayed significant downregulation at 12 h post-infection. Subsequently, it showed an upregulation at 24 h post-infection in IR24. However, this metabolite did not exhibit upregulation in IRBB27 at either the pre- or post-infection stages. Farnesyl diphosphate exhibited a remarkable upregulation in IRBB27 at the 12 h post-infection stage, contrasting with its downregulation in all other treatment conditions of IR24 and CG154. Additionally, other metabolites such as coproporphyrinogen III and arsenic acid were notably upregulated in the control stage of IRBB27 but experienced downregulation in post-infection. Similarly, Methyl (indol-3-yl) acetate displayed upregulation in both the control and 24 h post-infection stages in IRBB27, yet showed significant downregulation at the 12 h post-infection stage. This observation suggests a dynamic response of this metabolite in the resistant variety following pathogen challenge, with subsequent recovery at 24 h, aligning once again with the control condition, indicative of its resilience post-infection. The metabolite (2R,3S)-2,3-Dimethylmalate exhibited pronounced upregulation at 12 h post infection in CG154 wild rice compared to the susceptible and resistant varieties both pre and post infection. Similarly, phaseic acid showed higher expression levels in CG154 at the 12 h post-infection stage compared to its pre-infection stage and the 12 h post-infection stage of IRBB27, but it was downregulated in other samples. Additionally, 1-napthoic acid was significantly upregulated in CG154 wild rice at the 24 h post-infection stage, while being downregulated in the control and 12 h post-infection conditions, as well as in other rice varieties such as IR24 and IRBB27. Also, metabolites like chlordecone and kaempferol 3-O-beta-D-glucosyl-(1- > 2)-beta-D-glucoside exhibited significant downregulation in IR24 control and CG154 at 12 h post-infection, respectively. Digalacturonate, on the other hand, was downregulated only in IRBB27 at the 24 h post-infection stage and in CG154 control, while being upregulated in other treatment and control samples. These findings suggest the presence of unique metabolites expressing predominantly at specific stages in wild rice species (Fig. 3a). In the analysis and discussion of individual metabolites within each genotype, certain distinct metabolites were identified, showcasing clear differences and deviations of one genotype from others (Fig S2). However, no evident separation of genotypes based on the dendrogram was observed.

The hierarchical clustering models revealed three major clusters: one comprising samples from the highly susceptible IR24 cultivar, distinct from the other two rice (IRBB27 and CG154) (Fig. 3b-c). Both PCA and HCA modeling provided descriptive insights into the overall data structure, revealing underlying patterns and sub-structures such as variety-related clustering, treatment-dependent groupings (pre and post-infection stages) and time-related variation (12 h, 24 h) (Fig. 3b-d). These findings suggest a biological phenomenon in the extracted metabolite space, where differential metabolite profiles delineate temporal cellular responses of rice plants to Xoo infection.

Differentiation comparative metabolites: IR24 vs IRBB27

Comparing susceptible and resistant rice plants revealed significant differences in metabolite expression profiles. In IR24, oxalyl-CoA, UDP-N-acetyl-D-glucosamine, tetrahydrofolate, kaempferol 3-O-beta-D-glucosyl-(1- > 2)-beta-D-glucoside, and molybdate were significantly downregulated at the 24 h post-infection stage, while showing slight upregulation at the 12 h post-infection stage compared to the control condition. These metabolites were upregulated at both the 24 h post-infection and control stages in IRBB27, while with variations at the 12 h post-infection stage. Distinct changes were observed in IR24 and IRBB27 for metabolites such as 5-phospho-alpha-D-ribose 1-diphosphate, deoxyribose and cysteamine, with downregulation in IR24 at 24 h post-infection, contrasting with upregulation in both 12 h and 24 h post-infection stages of IR24 and IRBB27 respectively. Asparagine and N6-(L-1,3-Dicarboxypropyl)-L-lysine exhibited higher upregulation in IRBB27 at 24 h post-infection but were downregulated in other samples. Farnesyl diphosphate showed prominent upregulation at the 12 h stage of IRBB27 compared to IR24, suggesting its potential as a marker metabolite in IRBB27 at this stage. Trans-Dec-2-enoyl-CoA, D-glucono-1,5-lactone 6-phosphate exhibited minimal accumulation in IR24 control but showed higher accumulation post-infection in both IR24 and IRBB27, including their control conditions. Metabolic changes were observed in response to stress from Xoo infection, such as the fluctuation of methyl (indol-3-yl) acetate expression in IRBB27. Metabolic compounds like S-Inosyl-L-homocysteine, sedoheptulose 7-phosphate, 5-O-(1-Carboxyvinyl)-3-phosphoshikimate, delphinidin-3,5,3'-triglucoside, glycerophosphoglycerol, methanophenazine, betanin, L-cysteate, and NADH were highly upregulated in IR24 control and 24 h infection stages but significantly downregulated at the 12 h post-infection stage, indicating a maintenance of susceptibility-related metabolic levels in IR24 despite alterations in response at 12 h. Similarly, in IRBB27, variation was observed in the expression patterns of these metabolites between control and post-infection stages (Fig. 3b). Based on the observed metabolic responses, IRBB27 appears to be the better genotype against Xoo infection, as it exhibits upregulation of defense metabolites such as oxalyl-CoA, UDP-N-acetyl-D-glucosamine, tetrahydrofolate, and kaempferol 3-O-beta-D-glucosyl-(1- > 2)-beta-D-glucoside at both post-infection and control stages, along with significant upregulation of farnesyl diphosphate and other key metabolites, indicating a robust defense mechanism.

Differentiation comparative metabolites: IR24 vs CG154

Phaseic acid, (2R,3S)-2,3-Dimethylmalate, and (S)-2,3,4,5-tetrahydropyridine-2-carboxylate secondary metabolites exhibited significant upregulation in CG154 at 12 h post-infection, contrasting with downregulation observed in all other samples except for phaseic acid and (S)-2,3,4,5-tetrahydropyridine-2-carboxylate which showed slight upregulation in CG154 control condition. However, in IR24, all three metabolites were downregulated. Oleamide and nicotinate D-ribonucleotide were found at higher levels in CG154 at both 12 h and 24 h post-infection, contrasting with lower accumulation or downregulation observed in control and all stages of IR24. Similarly, metabolites such as sirohydrochlorin, solanine, FADH2, delphinidin-3-glucoside-5-caffoyl-glucoside and delphinidin-3,5,3'-triglucoside displayed downregulation at the 24 h post-infection stage in CG154 but were upregulated at control and 12 h infection stages; most of these metabolites exhibited similar upregulation in IR24 at both control and infection stages. Additionally, L-Cysteate, oryzalexin E, acetoacetyl-CoA, pelargonidin 3-(6-p-coumaroyl) glucoside, glycerophosphoglycerol and sedoheptulose 7-phosphate were identified as downregulated in both 12 h and 24 h post-infection stages of CG154, while they were upregulated in IR24 pre and 24 h post-infection. So, based on the observed metabolic responses, CG154 demonstrates a potentially better defense mechanism against Xoo compared to IR24. CG154 exhibits a superior response against Xoo infection, as evidenced by its significant upregulation of key secondary metabolites at 12 h post-infection and higher levels of metabolites like oleamide, (2R,3S)-2,3-Dimethylmalate, and (S)-2,3,4,5-tetrahydropyridine-2-carboxylate compared to IR24, indicating a more robust defense mechanism (Fig. 3c).

Differentiation comparative metabolites between IRBB27 vs CG154

HCA between IRBB27 and CG154 revealed notable differences. As kempferol 3-O-beta-D-glucosyl-(1- > 2)-beta-D-glucoside was consistently upregulated in both lines except at the 12 h post-infection stage in CG154 where it was remarkably downregulated. Key biomarkers such as desacetoxyvindoline and 1-napthholic acid were highly upregulated in CG154 at 24 h post-infection, unlike in IRBB27. The compound 3-methyl-2-butenal, specifically expressed in CG154, decreased post-infection stages in IRBB27. Certain metabolites like L-2,3-Dihydrodipicolinate, maltohexaose, trans-zeatin ribose, and N-(L-arginino) succinate were upregulated in both IRBB27 varieties at 12 h and 24 h infection condition but downregulated in the same stages of CG154, except for maltohexose which was upregulated in CG154 at 12 h post-infection. Defense-related metabolites like CMP-N-acetylneuraminate, farensyl diphosphate, and 2-Dehydro-3-deoxy-D-octonate-8-phosphate were highly upregulated in IRBB27 at 24 h post-infection, potentially contributing to its survival, whereas these were downregulated in CG154 samples except for CMP-N-acetylneuraminate. Certain normal stage metabolites like arsenic acid and coproporphyrinogen III were highly accumulated only in IRBB27 control samples. Metabolites like (2R,3S)-2,3-Dimethylmalate and 2-Oxo-3-hydroxy-4-phosphobutanoate were highly upregulated at 12 h post-infection of CG154 but downregulated in all remaining samples of both CG154 and IRBB27. Based on the analysis, CG154 demonstrates a better defense response compared to IRBB27. This conclusion is drawn from the significant upregulation of key biomarker metabolites such as desacetoxyvindoline, 1-napthholic acid at 12 h, (2R,3S)-2,3-Dimethylmalate and 2-Oxo-3-hydroxy-4-phosphobutanoate at 24 h post-infection in CG154, in contrast to IRBB27 samples. Additionally, certain defense-related metabolites like CMP-N-acetylneuraminate, farensyl diphosphate, and 2-Dehydro-3-deoxy-D-octonate-8-phosphate were highly upregulated in IRBB27 at 24 h post-infection, suggesting a potential survival mechanism, while these metabolites were downregulated in CG154 samples. These findings indicate that CG154 exhibits a more robust and effective defense response against Xoo infection compared to IRBB27 (Fig. 3d).

Exploring metabolic relationships by correlation heatmap analysis

A correlation heatmap visually represents the correlation between different variables in a dataset, often used in metabolomic studies to explore relationships between metabolites. Heatmaps uncovered relationships, outliers, and linear/nonlinear connections. Color-coded cells ease interpretation, identifying both linear and nonlinear associations. In Fig S5, a correlation heatmap illustrates Pearson Correlation Coefficients between samples of IR24, IRBB27, and CG154 under various control and treated conditions, focusing to discern linear relationships within the metabolic dataset across susceptible, resistant and wild rice relative. Higher coefficients near 1 signify greater sample similarity. Each coefficient is depicted as a square, with color denoting the degree of correlation on a blue-to-red scale, where blue indicates 0 and red indicates 1.

Identification of biomarker metabolites through Variable Importance Projection (VIP) score analysis in rice genotypes

Variable Importance in the Projection (VIP) scores were calculated to identify, characterize, and statistically confirm the selected biomarkers for validation. VIP score plots from PLS-DA showed differentially expressed metabolites pre-infection stage and after Xoo-infected. Metabolites with a VIP score ≥ 1 was chosen as significantly affected by the treatments. The VIP scores from two components of PLS-DA were employed to discern metabolites capable of discriminating between control and infected rice varieties (Fig. 4a). The VIP score plots displayed distinct separation of metabolites between each pair of comparative samples, such as IR24 vs IRBB27 (Fig. 4b), IR24 vs CG154 (Fig. 4c), and IRBB27 vs CG154 (Fig. 4d). The identified metabolites played a crucial role in genotype discrimination, with some metabolites significantly upregulated in one genotype while downregulated in another. These key metabolites across the three rice lines, highlighted in the VIP score plot, were essential for identifying biomarker metabolites.

Fig. 4.

Metabolite profiling and genotype differentiation in response to Xoo infection; (a) Discrimination of metabolites among leaves of IR24, IRBB27, and CG154 under control and treated conditions from VIP (Variable importance in the projection) scores. Metabolites with VIP scores ≥ 1 were deemed affected by Xoo. The figure showing varying accumulation patterns of significant metabolites in infected plants compared to controls across pre and post-infection stages, with some metabolites increasing and others decreasing, showing genotype-dependent differences. Additionally, Xoo infection prompted both up- and down-regulation of metabolites compared to controls; (b) The VIP plot revealed distinct metabolite signatures between IR24 and IRBB27; (c) The VIP plot revealed distinct metabolite signatures between IR24 and CG154; (d) The VIP plot revealed distinct metabolite signatures between IR24 and CG154

The VIP plot illustrates the discriminatory components between IR24 and IRBB27 samples. For instance, in IR24, metabolites like CDP-choline and CoA were exclusively upregulated in the control, while others such as 2-Succinyl-5-enolpyruvyl-6-hydroxy-3-cyclohexene-1-carboxylate, carbamate, MurNAc 6-phosphate, ATP and sirohydrochlorin were specifically upregulated at the 12 h post-infection stage. Also, oxaloacetate and tryptophan showed heightened expression at the 24 h infection stage in IR24. Moreover, unique metabolites like UDP, arsenic acid, quing hau sau, and lead exhibited elevated expression solely in IRBB27 under control conditions. Similarly, CO₂ (total), 1,4-Dihydroxy-2-naphthoate, 2-(2'-Methylthio)ethylmalic acid, and 3-Oxopalmitoyl-CoA were highly expressed only in 12 h IRBB27 samples, while methyl (indol-3-yl)acetate, biochanin A 7-O-(6-O-malonyl-beta-D-glucoside) and shikimate-3-phosphate were specific to IRBB27 at the 24 h post-infection stage. Tryptophan, CoA, CDP-choline, secoisolariciresinol, sirohydrochlorin were upregulated in IR24 but downregulated in IRBB27. Like that, 2-(2'-Methylthio) ethylmalic acid, 3-Oxopalmitoyl-CoA and hexadecanoic acid were highly upregulated in IRBB27 but downregulated in IR24. These findings suggest that these specific metabolites serve as biomarkers, uniquely upregulated at particular stages in specific genotypes. Metabolites with higher enrichment values serve as reliable candidates for differentiating genotypes from susceptible to resistant, thus identified as key biomarker metabolites from this VIP plot (Fig. 4b). Unique metabolites in the wild relative CG154 differ from the susceptible genotype IR24, exhibiting elevated expression levels at the 12 h post-infection stage in CG154, including oleamide, nicotinamide-beta-riboside and chlorogenate, which are downregulated in IR24 at the same stage. However, 2-amino-4-hydroxy-6-(D-erythro-1,2,3-trihydroxypropyl)-7,8-dihydropteridine and 1-linoleoylglycerophosphocholine are highly expressed in CG154 at the 24 h infection stage, while nicotianamine shows heightened expression in IR24. Furthermore, metabolites such as 3-Methoxy-4',5-dihydroxy-trans-stilbene and acetoacetyl-CoA, which are highly expressed in IR24 at the 12 h infection stage, are indicative of susceptibility, as they are significantly downregulated or not expressed in CG154 at the same stage. Similarly, in the pre-infection stage, metabolites like cysteamine and (S)-hydroxyhexanoyl-CoA are highly expressed in CG154 but downregulated in IR24, whereas some metabolites are uniquely expressed in IR24 at the pre-infection stage and downregulated in CG154 (Fig. 4c). When comparing the metabolite expression between wild and resistant genotypes, certain metabolites were identified with higher expression in the resistant genotype but very low expression in the wild type. These metabolites include norbixin, chlorophyllide, trans-zeatin riboside, 3-methoxy-4',5-dihydroxy-trans-stilbene, pelargonidin 3-(6-p-coumaroyl) glucoside, phloretin, beta-tocotrienol, N6-(delta2-Isopentenyl)-adenosine 5'-monophosphate and methylarsonate among which norbixin, chlorophyllide, trans-zeatin riboside, 3-Methoxy-4',5-dihydroxy-trans-stilbene, pelargonidin 3-(6-p-coumaroyl) glucoside, and pelargonidin 3-(6-p-coumaroyl) glucoside exhibited high enrichment values, primarily responsible for separating the resistant genotype from the wild type. Additionally, a larger number of unique metabolites such as norbixin, 3-Methoxy-4',5-dihydroxy-trans-stilbene, pelargonidin 3-(6-p-coumaroyl) glucoside, arsenic acid, phloretin, thiamin diphosphate, N6-(delta2-Isopentenyl)-adenosine 5'-monophosphate and DIBOA were highly expressed in the normal or pre-infection stage of the resistant line IRBB27 compared to the wild relative species CG154, which showed metabolites like cysteamine, murNAc 6-phosphate, and (S)-hydroxydecanoyl-CoA. At the 12 h post-infection stage in IRBB27, beta-tocotrienol and methylarsonate were highly upregulated, whereas in CG154, they were significantly downregulated, confirming their importance as key metabolites expressed at the 12 h infection stage for plant survival. In contrast, at the 12 h infection stage in CG154, other metabolites such as GDP-4-dehydro-6-deoxy-D-mannose, vitexin, selenite, and chlorogenate were upregulated. Similarly, at the 24 h post-infection stage, the IRBB27 genotype showed higher expression of metabolites like trans-zeatin riboside, chlorophyllide, methyl (indol-3-yl) acetate, pyropheophorbide a, and 8-[(1R,2R)-3-Oxo-2-{(Z)-pent-2-enyl}cyclopentyl]octanoate while in CG154, it was L-arogenate, 3-hexenol, and canavanine (Fig. 4d). Thus, the VIP score plot serves as a tool for identifying candidate metabolites at specific infection stages, indicative of susceptibility or resistance in plant growth and regulation. In the resistant rice variety IRBB27, the metabolite biochanin A 7-O-(6-O-malonyl-beta-D-glucoside) (ID: C12625; molecular formula: C₂₅H₂₄O₁₃) showed the largest change with the highest upregulation compared to the susceptible variety IR24, with a fold change of 16.697 and a log₂(FC) value of 4.0615. Similarly, in the wild rice line CG154, the metabolite gibberellin A8-catabolite (ID: C11870; molecular formula: C₁₉H₂₂O₇) exhibited the largest change compared to IR24, with a fold change of 26.841 and a log₂(FC) value of 4.7464.

Metabolic pathways impacted to Xoo infection

To assess the impact of Xoo infection on rice plant metabolomes, metabolic pathway analysis (MetPA) was conducted utilizing putatively annotated metabolite IDs with MetaboAnalyst bioinformatics tool suite (version 5.0 https://www.metaboanalyst.ca/). MetPA integrates univariate and multivariate analyses with pathway topology analysis, presenting significant pathways linked to the metabolomic data [27, 28]. This analysis was crucial for enhanced data visualization, prioritizes significant pathways based on lower log10(p) scale p-values and pathway impact factors. Analyzing the metabolic pathways' influence coefficients revealed potential target pathways with impact values exceeding 0.1 consequently, emerged as closely linked pathways in this investigation.

Upon detailed pathway analysis, significant pathway activations were observed primarily in the susceptible cultivar IR24 post-infection, including pyruvate metabolism, glycolysis/gluconeogenesis and cyanoamino acid metabolism at 12 h post-infection (hpi) and also butanoate metabolism, tryptophan metabolism, and isoquinoline alkaloid biosynthesis at 24 hpi. Interestingly, pyruvate metabolism, isoquinoline alkaloid biosynthesis and glycolysis/gluconeogenesis were already active in the resistant variety IRBB27 at pre-infection stages. Additional pathways, such as butanoate metabolism, pyruvate metabolism, isoquinoline alkaloid biosynthesis, beta-alanine metabolism and porphyrin metabolism, exhibited differential activity in the control stage of CG154. Post-infection stages of resistant and wild lines showed varying pathway impacts depending on the infection, with significant impacts observed in pathways such as alanine, aspartate, and glutamate metabolism, citrate cycle (TCA cycle), pyrimidine metabolism, glycine, serine, and threonine metabolism, purine metabolism, glyoxylate and dicarboxylate metabolism, vitamin B6 metabolism, arginine biosynthesis, zeatin biosynthesis, carbon fixation in photosynthetic organisms, cysteine and methionine metabolism, anthocyanin biosynthesis, nicotinate and nicotinamide metabolism, riboflavin metabolism, thiamine metabolism, phenylalanine, tyrosine, and tryptophan biosynthesis, fatty acid degradation, sulfur metabolism, cyanoamino acid metabolism, and isoquinoline alkaloid biosynthesis (Fig. 5 a-i). We focused on the significantly different metabolites primarily involved in the fatty acid degradation, the TCA cycle, and alanine, aspartate, and glutamate metabolism, as well as purine and pyrimidine metabolism pathways and have used heatmaps to illustrate the changes in metabolite accumulation across the genotypes (Fig. 6). These enriched pathways were represented using normalized metabolite values, selected based on the highest match status in the different rice varieties and their respective treatment conditions.

Fig. 5.

Pathway impact analysis of rice metabolomes reveals differential responses across rice lines and infection stages. MetaboAnalyst pathway analysis reveals significant pathways identified from the metabolomes of IR24, IRBB27 and CG154 rice. Pathways are ranked by p-value (y-axis) for enrichment analysis and pathway impact values (x-axis) for topology analysis. Node color indicates significance level (red = lowest p-value), while node size represents pathway impact. The size of each circle corresponds to the pathway's impact score, highlighting the most impacted pathways with high statistical significance scores. The figures above illustrate the impact pathways exhibiting significant variations among genotypes and across pre- and post-infection stages; (a) IR24 control; (b) IR24 12 hpi; (c) IR24 24 hpi; (d) IRBB27 control; (e) IRBB27 12 hpi; (f) IRBB27 24 hpi; (g) CG154 control; (h) CG154 12 hpi; (i) CG154 24 hpi

Fig. 6.

Metabolomic differences in Key KEGG pathways. Variations in the metabolome compounds of rice genotypes (susceptible, resistant, wild) treated with BLB post infection 12hpi, 24hpi, and pre infection (0hpi) that are involved in important KEGG pathways of fatty acid degradation, the TCA cycle, alanine, aspartate, and glutamate metabolism, as well as purine and pyrimidine metabolism. The linked KEGG pathways are shown by the green typeface metabolism, and the direct and indirect regulation of the downstream compound's synthesis is indicated by the black dotted arrows

Discussion

Rice is a vital food source for many people worldwide, but it's been facing substantial yield loss from bacterial leaf blight (BLB) caused by a harmful gram negative proteobacterium Xanthomonas oryzae pv. oryzae (Xoo) [30]. BLB affects most types of rice and which is a major challenge for rice farmers everywhere [31]. Previous studies have reported that the progression of disease Xoo is due to various metabolic and molecular strategies it employes to successfully colonies host and evade host defense mechanisms [3]. One key factor is its type III secretion system (T3SS), which delivers proteins into host cells to counteract the defense mechanisms of host plant. Among these proteins are transcription activator-like (TAL) effectors, which target specific genes like SWEET, while non-TAL effectors like XopR also disrupt host defenses [32]. Additionally, the bacterium's lipopolysaccharide (LPS) triggers callose deposition [33], but Xanthomonas can break down this barrier using xyloglucan depolymerization machinery [34]. Another important factor is putative phytase A (phyA), which helps the bacterium obtain phosphate from plant tissues for nutrition [35]. Lastly, certain amino acids like leucine, isoleucine, and valine are crucial for the bacterium's growth and virulence in rice plants. To end this, understanding of rice-Xoo interplay especially at the level of metabolism is pivotal to gain insights into metabolic basis of diseases and resistance. To understand host's metabolic responses and metabolic basis of defense mechanisms to Xoo infection we investigated these dynamics using LC–MS analysis. By examining pre and post-infection stages of rice leaf samples from genotypes IR24, IRBB27 and CG154, we aimed to find out metabolic changes and their distinctions between susceptible, resistant and wild relative. Infection of susceptible, resistant and wild rice plants resulted in metabolic changes impacting both primary and specialized secondary metabolism, with specific metabolites being up or down-regulated. In the resistant samples, both upregulation and downregulation were observed in phytochemical compounds, phenylpropanoids and related compounds, flavonoids, alkaloids, terpenoids, organic acids, amino acid derivatives, inorganic compounds, fatty acids and lipid derivatives compared to IR24, as indicated by log₂ fold change values. Similarly, in CG154 a mixture of upregulation and downregulation was noted in phenylpropanoid compounds, flavonoids, terpenoids, alkaloids, glucosinolates, amino acid derivative compounds and flavones compared to IR24. Metabolites showing alterations in post-pathogen infection are implicated in energy metabolism, suggesting that these changes could be because of the Xoo infection. Also, distinct metabolic responses in resistant and susceptible tomato plants to TYLCV (Tomato Yellow Leaf Curl Virus) infection, has highlighted the importance of salicylic acid (SA) biosynthesis in defense for resistant plants. The findings suggest the intricate coordination of primary and secondary metabolites as a crucial defense mechanism against TYLCV [36]. Temporal differences in metabolite and protein responses, particularly within the phenylpropanoid pathway, contribute to contrasting outcomes between rust-resistant and rust-susceptible Eucalyptus grandis genotypes, shedding light on functional modulation mechanisms against Eucalyptus rust [37]. The metabolomic responses of late blight-resistant (Ziyun No.1) and susceptible (Favorita) potato cultivars to Phytophthora infestans were investigated. Significant metabolic differences emerged after 48 h post-inoculation, with 198 and 115 differentially expressed metabolites (DEMs) identified in compatible and incompatible interactions, respectively. The resistant cultivar exhibited elevated levels of SA and key metabolites involved in phenylpropanoid biosynthesis, suggesting their role in defense mechanisms against late blight [38].

In the present investigation, comparison between the wild type and susceptible cultivar, 149 metabolites were upregulated and 162 were downregulated in the wild type (CG154) compared to the susceptible (IR24). When comparing the resistant line and susceptible cultivar, 85 metabolites were upregulated in the resistant (IRBB27) and 92 were downregulated compared to the susceptible. Similarly, 156 metabolites were upregulated and 149 were downregulated in the wild type (CG154) compared to the resistant (IRBB27). The results suggest that there are significant differences in metabolite expression patterns between the wild-type (CG154) and susceptible cultivar (IR24), with a higher number of metabolites being upregulated in CG154. Similarly, there are also significant differences between the resistant line (IRBB27) and the susceptible cultivar (IR24), with fewer metabolites being upregulated in IRBB27. Additionally, the comparison between CG154 and IRBB27 reveals distinct metabolic profiles, with a higher number of metabolites being upregulated in CG154. These findings indicate potential metabolic mechanisms underlying susceptibility and resistance in the investigated rice lines. Employing liquid chromatography-mass spectrometry-based untargeted metabolomics, it has revealed the defense mechanisms of the Pst-resistant wheat plant with Koonap cultivar, characterized by an increase in phenolic compounds in infected plants contrasting with their decrease in control plants, while these metabolites were down-regulated in the susceptible wheat Gariep cultivar, providing insights into dynamic metabolic reprogramming and potential biomarkers for resistance against Puccinia striiformis [39]. Similarly, in the present study, the Xoo-resistant IRBB27 and CG154 rice plants quickly built up defense metabolites when infected by the pathogen, unlike the susceptible IR24 cultivar, revealing valuable information about plant–microbe interactions and potential strategies for enhancing plant resistance to bacterial blight disease through metabolic engineering. A study using metabolomics and LC/MS found clear metabolic shifts in wheat grains infected with Tilletia controversa, suggesting changes in fungal toxin-related compounds and defense mechanisms. Also, the study observed increased levels of cucurbic acid and octadecatrienoic acid post-infection, and identified eight metabolic pathways potentially affected by the interaction [40]. Moreover, the study conducted on sorghum revealed the presence of 23 potential metabolic pathways, of which nine were notably significant. Among these pathways are phenylalanine metabolism, stilbenoid and gingerol biosynthesis, flavonoid biosynthesis, tryptophan metabolism, riboflavin and tyrosine metabolism, and phenylpropanoid biosynthesis. Together, these pathways contribute to trace the metabolic state associated with the induced resistance of sorghum against Colletotrichum sublineolum infection [17]. We looked at certain molecules and pathways in rice plants which changes during a Xoo attack. These could be important markers and regulators of how susceptible, resistant and wild plants react to the bacterial blight infection. In this current study we have identified different metabolic pathways activated during Xoo infection which include alanine, aspartate, and glutamate metabolism, citrate cycle (TCA cycle), pyrimidine metabolism, glycine, serine, and threonine metabolism, purine metabolism, glyoxylate and dicarboxylate metabolism, vitamin B6 metabolism, arginine biosynthesis, zeatin biosynthesis, carbon fixation in photosynthetic organisms, cysteine and methionine metabolism, anthocyanin biosynthesis, nicotinate and nicotinamide metabolism, riboflavin metabolism, thiamine metabolism, phenylalanine, tyrosine, and tryptophan biosynthesis, fatty acid degradation, sulfur metabolism, cyanoamino acid metabolism and isoquinoline alkaloid biosynthesis. Heatmaps showed the changes in metabolite accumulation in susceptible, resistant and wild rice lines in some selected enriched pathways (Fig. 6). Investigating metabolome profiles of common bean genotypes Teebus-RR-1 (resistant) and Golden Gate Wax (susceptible) revealed differential responses to Uromyces appendiculatus races (1 and 3) at 14- and 21-days post-infection (dpi). Statistically significant metabolites, including flavonoids, terpenoids, alkaloids, and lipids were induced by rust infections in both genotypes. The resistant genotype exhibited differential enrichment of key metabolites such as aconifine, D-sucrose and galangin indicating a defense mechanism against the rust pathogen, suggesting a potential strategy for understanding plant defense against rust [41].

During a pathogen infection, metabolic profiling indicates a significant change in the plant's primary and secondary metabolites [42]. Oryzalexins E which was upregulated in wild rice line CG154 as compared to susceptible IR24 is an important diterpene phytoalexins in rice that enhance the plant's defense against Xoo. Specifically, oryzalexin E and oryzalexin S have increased in response to blast infections, indicating their role in blast disease resistance [43]. Hydrogen selenide (H₂Se), upregulated in CG154 rice line as compared to susceptible IR24, is a selenium-derived compound that enhances plant disease resistance by preventing fungal invasion, damaging pathogen structures, and altering soil microbial communities [44]. Spermidine was found upregulated in CG154 as compared to susceptible IR24. In addition to this it also plays a crucial role in eggplant's defense against bacterial wilt by being upregulated through the SmSPDS gene, which is activated by the SmMYB44 transcription factor. Higher spermidine levels in resistant lines boost the plant's resistance to Ralstonia solanacearum, providing a key mechanism for disease resistance [45]. Pterostilbene was found upregulated in CG154 as compared to susceptible IR24 which was also significantly accumulates in DM-immune grape cultivars after Plasmopara viticola inoculation, enhancing their resistance to downy mildew [46] (Fig. 1a). Norbixin was found downregulated in CG154 as compared to IRBB27. Norbixin biosynthesis is up-regulated in response to wheat streak mosaic virus (WSMV) infection at both normal and elevated temperatures at 48 hpi/20 °C and 72 hpi/32 °C, contributing to the plant's defense mechanisms in resistant wheat cultivars [47]. Catechol was found downregulated in CG154 as compared to IRBB27. It plays a crucial role in plant defense by accumulating in infected tissues, such as onion scales infected with Colletotrichum circinans, and by being present in high concentrations in strawberry leaves, which increases resistance to the two-spotted spider mite by suppressing mite development [48]. Cysteamine was found downregulated in CG154 than IRBB27. Cysteamine is identified as a promising molecular marker of resistance to late blight disease, showing potential in breeding programs to enhance plant resistance mechanisms [49] (Fig. 1b). In our study, raffinose was upregulated in the IRBB27 rice line compared to the susceptible IR24 line, suggesting its role in pathogen recognition. Raffinose levels were about 10 times higher in infected samples but decreased significantly near sporogenesis and whip development, indicating its involvement in the initial defense response, similar to its role in sugarcane smut [50] (Fig. 1c). Glycerophosphoglycerol was upregulated after 12 h post infection in susceptible IR24 and downregulated in resistant IRBB27 as compared to IR24. The study found lipidome alterations in grapevine genotypes infected with Erysiphe necator. Resistant genotypes, such as BC4 and F26P92, showed significant changes in their lipid profiles, including down-accumulation of certain lipid classes like glycerophosphoglycerol, potentially contributing to their defense mechanisms against the pathogen [51]. Farnesyl diphosphate was identified as a remarkable upregulation after the infection in IRBB27 which is crucial in soybean defense against Phakopsora pachyrhizi, acting as a key precursor in terpenoid biosynthesis, which is upregulated in resistant genotypes [52]. It was upregulated in pre infection stage of IRBB27 but downregulated after the infection. Coproporphyrinogen, including both coproporphyrinogen I and III, plays a crucial role in chlorophyll biosynthesis, and its downregulation in Candidatus Phytoplasma infected sesame plants suggests a significant impairment in the chlorophyll biosynthesis pathway, contributing to chlorophyll breakdown and impaired chloroplast physiology [53]. Phaseic acid was found higher expression in CG154 after the infection as compared to pre infection or control stage. Phaseic acid plays a role in the differential plant immune responses to co-colonization by Metarhizium robertsii and Fusarium solani, with its levels significantly increasing in bean plants mutual colonized by both fungi compared to un-inoculated plants, indicating its involvement in the complex regulation of plant defense mechanisms under biotic stress [54]. Digalacturonate which was downregulated at 24 h of post infection the infection of IRBB27, a product of pectin degradation, is influenced by the down-regulation of genes like PME53 (pectinesterase), PL8 and PL18 (pectate lyase), and At1g48100 (polygalacturonase) in pear fruit treated with Meyerozyma guilliermondii combined with alginate oligosaccharide [55]. CDP-choline accumulates in Trichoderma-treated tomato plants infested with Macrosiphum euphorbiae, potentially enhancing defense responses against aphid infestation [56] (Fig. 4b). Shikimate 3-phosphate was found uniquely expressed in IRBB27 24 h post infection stage which is a key intermediate in the shikimate pathway, where the enzyme 5-enolpyruvylshikimate-3-phosphate synthase (EPSPS) converts it into 5-enolpyruvylshikimate-3-phosphate. It is a crucial step targeted by the herbicide glyphosate; thus, genes coding for glyphosate-tolerant EPSPS can enhance glyphosate tolerance in crops by reducing shikimate accumulation and allowing normal growth even under herbicide treatment [57]. Hexadecanoic acid which was found upregulation in IRBB27 but downregulation in IR24 in our study it was also upregulated in other study reported in rice plant in response to the Rhizoctonia solani toxin, which is involved in the biosynthesis of saturated and unsaturated fatty acids, cutin, suberin, and wax, contributing to necrosis in infected rice plants [58] (Fig. 4b). Oleamide that was incredibly found in CG154 after 12 h of infection was also detected in mulberry root exudates, inhibits Ralstonia pseudosolanacearum growth by altering cell morphology, reducing extracellular polysaccharide content, inducing oxidative stress, and downregulating key virulence-related genes, thus contributing to the delay or slowdown of bacterial wilt in mulberry [59]. Similarly Acetyl-CoA was identified downregulated in CG154 after the infection as compared to susceptible IR24, serving as the primary energy source for the TCA cycle, which has also been downregulated in Cytospora mali under biocontrol by Bacillus veleznesis L-1, leading to energy shortage and reduced virulence in the fungus [60] (Fig. 4c). Chlorophyllide found upregulation and very unique to IRBB27 which can be identified as one of the defense metabolite, formed by chlorophyllase (CLH) upon cell disruption, plays a role in plant defense by exhibiting toxicity to insect herbivores, such as Spodoptera litura larvae, thereby inhibiting their growth [61]. Chlorogenate which was detected highly expressed in CG154 at 12 h post infection stage also accumulates in polyphenol oxidases (PPO)-downregulated potato tubers, enhancing defense against Phytophthora infestans by redirecting phenylpropanoid metabolism toward the production of defensive phenolic compounds [62] (Fig. 4d).

In our study, it was identified that the concentration of metabolites, which includes flavonoids, terpenoids, organic acids, inorganic compounds, amino acids and fatty acids derivatives and other related phytochemical compounds, showed a significantly higher (P ≤ 0.01) in IRBB27 compared to IR24. Similarly, in CG154 relative to IR24, phenylpropanoids and related compounds, diterpenoids, flavonoids, and terpenoids displayed higher levels (P ≤ 0.01). Analysis of PLSDA VIP data revealed that 344 metabolites exhibited upregulation, with a VIP score ≥ 1.0, in IRBB27 compared to IR24. The results suggest that there are significant differences in metabolite concentrations between the resistant genotype IRBB27 and the susceptible genotype IR24, as well as between the wild type CG154 and susceptible IR24. Additionally, the upregulation of a large number of metabolites in both IRBB27 and CG154 compared to IR24, along with the differential upregulation of metabolites between CG154 and IRBB27, highlights distinct metabolic responses associated with resistance and wild-type genotypes. By utilizing untargeted metabolomics to elucidate the distinctive metabolic signatures of four tomato cultivars with varying resistance levels to Ralstonia solanacearum, revealed potential biomarkers associated with resistance traits. Two highly resistant cultivars showed prominent levels of flavonoids and steroidal glycoalkaloids, while the less resistant cultivars exhibited distinctive fatty acid derivatives in root extracts. The intermediate resistant cultivar displayed lower levels of cinnamic acid derivatives and flavonoids compared to highly resistant cultivars, suggesting a potential correlation between metabolic profiles and resistance levels to Ralstonia solanacearum [63].

Conclusion

In our current study, we found that the resistant IRBB27 genotype exhibited the presence of amino acid derivatives, alkaloid phytochemical compounds, flavonoids, phenylpropanoids, terpenoids, anthocyanidins and anthocyanins. Also, in the wild CG154 genotype, phenylpropanoids (( +)-secoisolariciresinol, 3-methoxy-4',5-dihydroxy-trans-stilbene, pterostilbene) and related compounds, flavonoids (gentiodelphin, delphinidin-3-glucoside-5-caffoyl-glucoside, delphinidin-3,5,3'-triglucoside, pelargonidin 3-(6-p-coumaroyl) glucoside), terpenoids (oryzalexin E, acetoacetyl-CoA), alkaloids ((S)-2,3,4,5-Tetrahydropyridine-2-carboxylate, solanine), abscisic acid derivatives (phaseic acid), nucleotides and derivatives (nicotinate D-ribonucleotide, FADH₂, glucosinolates, anthocyanins, flavones), fatty acid derivative (oleamide) and amino acid derivatives ((2R,3S)-2,3-Dimeth, L-cysteate) were identified which are deviating from the susceptible genotype IR24. These findings suggest potential differences in metabolic responses between the susceptible genotype IR24 and the resistant genotype IRBB27 like oxalyl-CoA (CoA derivative), UDP-N-acetyl-D-glucosamine (nucleotide sugar), tetrahydrofolate (vitamin and derivative), kaempferol, isoorientin, syringetin, mirtillin 3-O-beta-D-glucosyl-(1- > 2)-beta-D-glucoside (flavonoid), Molybdate (metal/inorganic compound), 5-Phospho-alpha-D-ribose 1-diphosphate, D-Glucono-1,5-lactone 6-phosphate and 5-O-(1-Carboxyvinyl)-3-phosphoshikimate (phosphate compound), asparagine (amino acid), cysteamine and L-cysteate (amino acid derivative), farnesyl diphosphate and trans-Dec-2-enoyl-CoA (lipid derivative), sedoheptulose 7-phosphate (sugar/oligosaccharide), delphinidin-3,5,3'-triglucoside, betanin, methyl(indol-3-yl) acetate, glycerophosphoglycerol, methanophenazine and Inosyl-L-homocysteine (secondary metabolite) as well as between the resistant genotype IRBB27 and the wild type CG154 such as kaempferol 3-O-beta-D-glucosyl-(1- > 2)-beta-D-glucoside (flavonoid), desacetoxyvindoline (alkaloid), 1-Naphtholic acid (phenolic compound), 3-Methyl-2-butenal (aldehyde), L-2,3-Dihydrodipicolinate and n-(L-arginino)succinate (amino acid derivative), maltohexaose (carbohydrate), trans-zeatin ribose (cytokinin nucleotide), CMP-N-acetylneuraminate (nucleotide sugar), farnesyl diphosphate (terpenoid), 2-Dehydro-3-deoxy-D-octonate-8-phosphate (phosphate compound), arsenic acid (inorganic compound), coproporphyrinogen, (2R,3S)-2,3-Dimethylmalate (organic acid) and 2-Oxo-3-hydroxy-4-phosphobutanoate (organic acid) were identified. The differential presence of metabolites among these genotype highlights distinct metabolic responses associated with resistance in crop plants. Further exploration of these metabolic differences could aid in the development of targeted breeding strategies, leveraging metabolite-QTLs and pathway modifications, to enhance crop resistance against pathogens and improve agricultural productivity. Additionally, investigating the functional roles of specific metabolites could uncover novel mechanisms underlying plant defense mechanisms and inform future biotechnological interventions for crop improvement.

Abbreviations

Xoo

Xanthomonas oryzae Pv. oryzae

PCA

Principal component analysis

PLS-DA

Partial least squares-discriminant analysis

PRR

Pattern recognition receptors

PAMPs

Pathogen-associated molecular patterns

PTI

Pattern-Triggered Immunity

ROS

Reactive oxygen species

MAPK

Mitogen-activated protein kinase

R genes

Resistance genes

ETI

Effector-triggered immunity

HR

Hypersensitive response

LC-MS

Liquid chromatography mass spectrometry

TALE

Transcription Activator Like Effector

CFU/ml

Colony-forming unit per millilitre

OD

Optical density

Rpm

Revolutions Per Minute

ESI

Electrospray ionization

KEGG

Kyoto Encyclopedia of Genes and Genomes

RTs

Retention times

HCA

Hierarchical cluster analysis

DAM

Differentially Accumulated Metabolites

BPI

Base peak intensity

VIP

Variable importance in the projection

QC

Quality control

FC

Fold Change

ITP

Inosine triphosphate

UDP

Uridine diphosphate

CoA

Coenzyme A

CDP

Cytidine diphosphate

NADH

Nicotinamide adenine dinucleotide (NAD) + hydrogen (H)

FADH2

Flavin adenine dinucleotide

CMP

Cytidine monophosphate

MurNAc

N-acetyl-D-muramic acid

GDP

Guanosine diphosphate

HPI

Hour post infection

TCA

Tricarboxylic acid

BLB

Bacterial leaf blight

T3SS

Type III secretion system

XOP

Xanthomonas outer protein

LPS

Lipopolysaccharide

PhyA

Phytase A

DEMs

Differentially expressed metabolites

H₂Se

Hydrogen selenide

WSMV

Wheat streak mosaic virus

EPSPS

Enolpyruvylshikimate-3-phosphate synthase

CLH

Chlorophyllase

PPO

Polyphenol oxidases

SA

Salicylic acid

Authors’ contributions

PPD conceived and designed the research, conducted experiments, analyzed the data and written the complete manuscript. AK conducted experiments. MM reviewed and supervised the work. KB, KRB and KPB reviewed and edited and collected literatures. RMS and IAG edited, reviewed and approved the manuscript.

Funding

This work is funded by CSIR, New Delhi [Grant No. 38(1522)/21/EMR-II], and the authors IAG and PPD extend their appreciation to CSIR for the funding.

Data availability

The data and materials are provided in the article and supplementary material. Further inquiries for raw data can be made available upon request from the corresponding author.

Declarations

Ethics approval and consent to participate

Ethical approval was applied for this research work but the Institutional Biosafety Committee (IBSC) members reiterated that the scope of work as described in this proposal does not require IBSC approval.

Consent for publication

All authors have consented for its publication.

Competing interests

The authors declare no competing interests.