Blast Resistance Gene Pish Derived Reprogramming of Noncoding RNAs in Rice

Abstract

Rice and Magnaporthe oryzae have co-evolved sophisticated molecular interaction mechanisms. While protein-coding genes (PCGs) involve in the response to M. oryzae have been intensively studied, yet the role of long noncoding RNAs (lncRNAs) remain poorly unclear. In this study, we performed whole transcriptome strand-specific RNA sequencing of rice seedlings from the susceptible recurrent parent cultivar LTH (compatible interaction) and the resistant rice monogenic line IRBLsh-S (incompatible interaction, harboring the resistance gene Pish) following the inoculation with the blast fungus strain Y92-66b. The results revealed that 16.8% of the identified lncRNAs were responsive to blast-infection in the Pish-line at 24-hour post inoculation (hpi). Notably, 45.1% of these blast responsive lncRNAs were novel. Functional analysis indicated that, 49 differentially expressed lncRNAs were co-expressed with genes enriched for the response to oxidative stress and diterpene phytoalexin biosynthetic process (particularly the synthesis of momilactone A). Furthermore, the expression pattern of four lncRNAs correlated with those of genes related to the cell wall macromolecule metabolic process. Here, two lncRNAs (XLOC_046130 and XLOC_040277) were predicted to act as endogenous target mimics (eTMs) for miRNAs and were co-expressed with transcription factors to induce the expression of genes involved in the synthesis of momilactone A. This demonstrates that lncRNAs regulate innate immunity through complex networks involving PCGs, transcription factors, and miRNAs. This study provides new insight into the regulatory mechanisms governed by R genes and highlight the potential role of lncRNAs in breeding and agricultural practices for improved disease management.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12284-026-00883-y.

Background

Rice (Oryza sativa L.) contributes 23% of total calories to global food consumption, and it is urgent to improve rice production to satisfy the increasing population demands (Wilson and Talbot 2009). Rice blast caused by Magnaporthe oryzae is one of the most destructive and damaging disease, responsible for huge losses annually in rice production worldwide (Neupane and Bhusal 2021; Xiao et al. 2023). There is a need for new strategies to control blast disease in rice. During these years, many cultivars harboring significant resistance (R) genes were bred, which improved the tolerance and resistance in fields against rice blast disease (Khan et al. 2014; Ngernmuen et al. 2020). Moreover, various defence-regulated genes were cloned successfully through two fast-developing technologies, including genome editing and genome-wide sequencing, which accelerated to deepen the genetic resources with durable and broad resistance (Yano et al. 2016). The arms races between rice and M. oryzae are considered the classic model for researching plants and pathogens. Gene-gene theory helps better understand the evolution and battles between rice and M. oryzae (Devanna et al. 2022). Furthermore, noncoding RNAs, including miRNAs, have emerged as significant regulators of rice blast resistance (Li et al. 2023). Their regulatory functions suggest that a substantial layer of immune regulation remain to be elucidated.

Long noncoding RNAs (lncRNAs) are transcripts with over 200 bp in length that possess no apparent coding sequences or open reading frame (ORF) (Chen and Kim 2024; Marsico et al.,20222). Based on their genomic location, lncRNAs are mainly classified as intergenic lncRNA (lincRNA), intronic lncRNA, or antisense lncRNA (NATs) (Xia et al. 2025). In the past, the lncRNAs were regarded as useless constitutions in organisms. Nevertheless, recently, it has been essential for signal transduction and transcriptomic reprogram for various creatures like miRNA. Previous studies suggested that plants lncRNAs play an significant role in different biological processes, such as regulation of flowering time, root organogenesis, photomorphogenesis, fertility, phosphate (Pi) homeostasis, and many other developmental pathways (Ghorbani et al. 2021; Liu et al. 2022; Wang et al. 2023). Previously, it was reported that lncRNAs regulate immune signaling by directly targeting and modulating the expression of host NBS-LRR genes (Liao et al. 2022). Furthermore, species-specific and differentially expressed key lncRNAs were also identified, that regulate host resistance to pathogens attack in various plants, such as pepper, tomoato, grapevine and walnut (Bhatia et al. 2021; Feng et al. 2021; Yin et al. 2021; Delarue et al. 2025), which highlighted the role of lncRNAs in the plant immunity response. However, many researches about lncRNA was more concentrated on rice physiology. For example, Li et al. 2022 identified 445 Gibberellin-responsive (GARRs) lncRNAs were differentially expressed in Zea mays. Among these, four shared GARRs, and from which GARR2 has been shown to interact with ZmUPL1, which affecting GA-related gene expression and primary auxin responsive in maize (Li et al. 2022; Yadav et al. 2023). Another study identified 1549 lncRNAs, among them, 229 differentially expressed lncRNAs directly targeted stress-responsive genes, including AG, and various TFs, such as bHLH, LBD, MYB, NAC, and WRKY. Functional analysis revealed that some lncRNAs act as important regulators by competing with mRNAs for miRNAs binding. This mechanism modulate a critical network that control plant growth, development, and stress resistance (Sahraei et al. 2024). Moreover, a genome-wide study of M. oryzae at six infection stages identified 2601 novel lncRNAs, including 560 infection-specific expressed lncRNAs (ISELs). These ISELs were found to regulate fungal PR genes, such as those encoding xylanases and effectors. This suggests that lncRNAs mediate a dynamic molecular crosstalk from both sides of the host-pathogen interaction (Choi et al. 2022). Additionally, a study found 21 DELs in both susceptible Lijiangxintuanheigu (LTH) and resistance (IR25) rice varieties at different time hours, which are responsible for the regulation of ethylene (ET) and jasmonic acid (JA) pathways. The correlation network analysis revealed a key lncRNA, such as LncRNA-9497.1 particularly enhance the expression of R genes, RLKs, and phytohormone signalling against M. oryzae in resistant line (Shan et al. 2025). While, Liu et al. 2024 demonstrated that numerous lncRNAs, such as MOIRA1, MOIRA2, and MOIRA3 were key candidate against rice blast pathogen. The overexpression of MOIRA1, and MOIRA2 enhance the susceptibility to M. oryzae, and increasing the tiller number and single rice plant yield (Liu et al. 2024), suggesting the dual role of these lncRNAs against rice blast disease. Therefore, Yu et al. (2020) found that 73 lncRNAs regulate the expression of JA-related genes associated with bacterial blight resistance in rice. Moreover, the pathogen-responsive lncRNAs in rice were identified, which provide more valuable references for immunity-associated lncRNAs (Wang et al. 2020).

This research aims to elucidate the role of immunity-related lncRNA mediated by a major resistance gene in rice. We used the susceptible rice line Lijiangxintuanheigu (LTH) and its monogenic line IRBLsh-S, which harbors a broad-spectrum resistance gene Pish to construct whole transcriptome strand-specific RNA sequencing (ssRNA-Seq) library and identified regulating lncRNAs by R gene Pish. Our results show that many lncRNAs were coexpressed with defense-related pathways, especially momilactone A synthesis genes in IRBLsh-S, implying that lncRNAs were induced during the interaction between rice and blast fungus, and the lncRNAs involving defense networks mediated by R genes might be indispensable for plant immunity.

Materials and Methods

Plant Materials and Growth Conditions

Lijiangxintuanheigu (LTH), a Japonica rice variety that originated from Yunnan province in China, is highly compatible with almost all the rice blast isolates, and it is commonly used as a susceptible parental line for the identification and cloning of the resistance (R) genes (Yang et al. 2022). IRBLsh-S was derived from LTH and Shin 2 bred by the International Rice Research Institute (IRRI), harboring Pish R gene (Tsunematsu et al. 2000). In this study, the compatible rice line LTH and its monogenic line IRBLsh-S were planted under similar conditions in a greenhouse, as described in previous reports (Wei et al. 2013).

Fungus Strain and Inoculation

The M. oryzae strain Y92-66b was collected from the infected rice field and used for inoculation. This strain cannot cause the infection in the IRBLsh-S. The conidial suspension 1 × 10⁵ conidia mL^− 1 of pathogen was sprayed on IRBLsh-S and LTH with 0.02% tween 20. While, the distilled sterilize water (ddH₂O) with 0.02% tween 20 solution was sprayed on control (mock) plants. The inoculated and un-inoculated plants were placed in the dark chamber with 100% humidity at 28 °C for 24 h, and then transferred to the greenhouse. Disease reactions were checked at 7 days post inoculation (dpi) based on lesion grade (Li et al. 2014).

Whole Transcriptome ssRNA-Seq Library Preparation and Sequencing

The fully expanded twice and third leaves of 15 rice seedlings from each group were harvested and immediately frozen in liquid nitrogen for ssRNA-Seq at 24 hpi. There were three biological replicates in each treatment. RNA was extracted from rice leaves using the OMEGA plant RNA kit and quality was checked on a 1.0% agarose gel, and purity was assessed with a Nano Photometer^® spectrophotometer. RNA concentration was measured using a Qubit^® 2.0 Fluorimeter and the Qubit^® RNA Assay Kit, while integrity was evaluated with the RNA Nano 6000 Assay Kit on a Bioanalyzer 2100 system. Each sample used 3.0 µg of RNA for preparation. The preparation of libraries for ssRNA-Seq and deep sequencing was performed by the Novogene Corporation (Beijing, PR China). Ribosomal RNA was eliminated using the Epicentre Ribo-zero™ rRNA Removal Kit. Sequencing libraries were created from the rRNA-depleted RNA with the NEBNext^® Ultra™ Directional RNA Library Prep Kit for Illumina^®, following the manufacturer’s guidelines. The libraries were verified with the Bioanalyzer 2100 system before being sequenced on an Illumina Hiseq 2000 platform, producing 125 bp paired-end reads. The sequencing data have been submitted in the NCBI Sequence Read Archive (SRA) under the BioProject number (PRJNA716694), with the accession no (SRX10430967-SRX10430978).

LncRNAs Identification Pipeline

The rice transcriptome was assembled by independently aligning each RNA-seq dataset to the rice genome using TopHat 2.0, and then assembling the transcriptomes with Cufflinks 2.0. These individual transcriptomes were merged with Cuffmerge to create a final transcriptome, and RSEM was used to estimate transcript abundance. Firstly, transcripts of tRNA or rRNA without strand information or transcripts shorter than 200 bp were discarded. Secondly, the remaining transcripts with blast e-value < 1e-5 or with a PFAM domain were filtered out. Thirdly, transcripts with amino acid length ≥ 100 were removed. Lastly, the Coding Potential Calculator (CPC) was used to predict transcripts with coding genes (Kong et al. 2007). All transcripts with a CPC scores > 0 were discarded as potentially protein coding. The rest transcripts were considered dependable and putative expressed lncRNAs. To further validate the retrieving non-coding transcripts, all transcripts were re-analyzed using the CPC2 online tool available at (https://cpc2.gao-lab.org/run_cpc2_result.php?userid=251104913812484). This analysis confirms the accuracy and reliability of our final lncRNAs datasets, which initially identified by CPC (Kang et al. 2017).

Analysis of Differential Expression Genes and LncRNAs

To identify the DEGs and DELs, the EdgeR software package was used (Robinson et al. 2010). The significant DEGs and DELs were identified from the three biological replicates by comparing fungus- and mock-treatment in both the LTH and IRBLsh-S. The threshold for significant transcripts were a P-value < 0.05 and |log2 FC| ≥1, respectively.

Comparison of lncRNAs with Reference Ensembl Database

To further explore the novelty of our identified lncRNAs, we compared all lncRNAs with the previous annotated lncRNAs in the Ensemble Plants database release 62. The complete O. sativa Japonica group, IRGSP-1.0 genome assembly (Oryza_sativa.IRSP.1.0.62.gtf) was downloaded from Ensembl Plants, and all featured transcripts of lncRNAs (transcript_biotype_lncRNAs) were extracted as reference transcript IDs comparison. After that, we extracted the transcripts IDs generated by Cufflinks for all our putative lncRNA. These transcripts IDs of lncRNAs were directly compared against the reference (Ensembl Plants) transcripts IDs by Venny 2.1 online tool. The matching transcripts IDs were subjected as “known lncRNAs” and non-matching were classified as “novel lncRNAs”.

GO Enrichment and Pathway Analysis

The AgriGo gene ontology (GO) tool was used for the enrichment analysis of DEGs from incompatible and compatible rice lines (Tian et al. 2017). The biological process (BP), cellular component (CC) and molecular functions (MF) GO terms were selected based on the two rice lines’ FDR values of < 0.05. Moreover, two analysis tools, DAVID (Version 6.8) and KEGG, were evaluated to find out the related pathways and functions.

Co-expression Network Analysis

To identify resistance gene clusters, DEGs and DELs from the treated IRBLsh-S and LTH lines were selected for further functional analysis. In IRBLsh-S, a matrix with 12 columns and 2024 rows was produced, with each row and column corresponding to one transcript and sample, respectively. Consequently, the WGCNA analysis was performed from R package (Langfelder and Horvath 2008). The matrix served as the input for WGCNA, where data filtering was conducted to eliminate transcripts with excessive missing values, following WGCNA recommended cutoff thresholds. After that, a soft-thresholding power of 9 was selected from 1 to 20 calculation. Transcripts of lncRNAs, TFs, and other coding RNAs clustered into different co-expression modules, and then GO enrichment and pathways analysis were conducted in every module. The interesting GO terms or pathways genes were subject to co-expressed analysis with lncRNAs and TFs using an R script in different modules. The relationship between lncRNA and mRNA and TF and mRNA with R² > 0.8 and FDR < 0.05 was considered a strong correlation. To explore the association among diverse transcripts in the same module, Cytoscape 3.3.0 was used for visualizing the network of each module (Demchak et al. 2014). The potential role of a separate lncRNA was determined according to the protein-coding transcripts associated with it in the network module (Morandin et al. 2016).

Quantitative Real-Time PCR Validation

To validate the expression levels of lncRNAs and mRNAs selected from the ssRNA-seq analysis, total RNA was isolated from the sample using an Eastepr^®Super total RNA extraction kit for qRT-PCR (Promega, Madison, USA). First-strand cDNA was reverse transcribed by GoScript™ Reverse Transcription System (Promega, Madison, USA). The qRT-PCR was performed using SYBR^® Premix Ex Taq™ II (TaKaRa, Dalian, China) with a final volume of 20 µL on a Bio-Rad Thermal Cycler. All primers used in this study are listed in Additional file 2: Table S1. All experiments contained three biological replicates.

Prediction of Endogenous Target Mimic for miRNAs

LncRNA as an endogenous target mimic (eTM) of miRNA was predicted using RegRNA 2.0 (http://regrna2.mbc.nctu.edu.tw/detection.html).

Results

Symptoms of Different Rice Lines Inoculated with Blast Fungus

Lijiangxintuanheigu (LTH) is a very susceptible variety against most strains of rice blast pathogen and is used as a recurrent parent to develop monogenic lines such as IRBLsh-S. Thus, the genetic relationship between LTH and IRBLsh-S is very close and can be considered ideal materials for researching R gene Pish regulating defense pathways. The disease reactions of IRBLsh-S and LTH to the strain Y92-66b were recorded at 7 dpi, revealed typical symptoms on LTH ranging from 3 to 5 (Fig. 1A), while IRBLsh-S were incompatible with Y92-66b (Fig. 1B). This phenotypic contrast was used to explore resistance-specific lncRNAs and mRNAs.

Fig. 1.

Phenotypic difference between IRBLsh-S and LTH after 7 days of inoculation by M. oryzae strain Y92-66b. A Compatible reaction of LTH(n = 70); B Incompatible reaction of IRBLsh-S (n = 61)

DEGs Identified Based on Whole Transcriptome ssRNA-Seq and Enrichment Analysis

High-throughput ssRNA-Seq was carried out to obtain the transcripts from the leaves of IRBLsh-S and LTH after inoculated with Y92-66b and mock (Additional file 1: Fig. S1). More than 724 million raw read sequences were generated from 12 libraries. After removing low quality reads and adapter sequences, 98.5% of the raw reads were clean. These data’s average Q20 and Q30 were 95.49 and 91.30, respectively. All clean reads were associated with the rice genome; about 91% could be mapped to the rice reference genome, and 84% were uniquely mapped (Table 1). Additionally, the principal component analysis (PCA) was performed based on the normalized FPKM values of all transcripts, revealed that PCA1 explained 32.94% of the total variance, while PCA2 accounted for 12.30% of the variance, respectively (Additional fie 1: Fig. S2). Hierarchical clustering analysis of the FPKM normalized gene count matrix demonstrated that the biological repeats of all samples clustered together, and the transcriptomic characteristics significantly differed between LTH and IRBLsh-S after inoculation with strain Y92-66b (Additional file 1: Fig. S3). The results indicated that our data was highly credible, and expression levels of transcripts obtained could be used for further analysis.

Table 1.

Statistical data of the RNA-Seq reads for samples

Samples	Raw reads	Clean reads	Raw bases(G)	Clean bases(G)	Q20%	Q30%	Mapped Reads%	Concordant Pairs%
IRBLsh-S-mock-1	67 525 358	66 449 891	16.88	16.61	95.19	90.81	92.80	85.40
IRBLsh-S-mock-2	61 713 799	60 826 761	15.43	15.21	95.64	91.66	93.80	86.20
IRBLsh-S-mock-3	67 865 145	67 138 375	16.97	16.78	95.82	91.94	93.60	86.50
IRBLsh-S-Y92-66b-1	58 987 779	58 376 088	14.75	14.59	95.94	92.15	88.30	82.10
IRBLsh-S-Y92-66b-2	59 657 433	58 969 513	14.91	14.74	95.60	91.52	89.80	82.60
IRBLsh-S-Y92-66b-3	59 728 333	59 001 296	14.93	14.75	95.56	91.47	89.10	82.80
LTH-mock-1	53 196 075	52 779 742	13.30	13.19	95.66	91.48	94.40	86.80
LTH-mock-2	55 439 514	54 885 568	13.86	13.72	95.75	91.72	94.10	86.10
LTH-mock-3	58 355 031	57 738 635	14.59	14.43	94.66	89.80	92.40	83.50
LTH-Y92-66b-1	66 622 072	65 378 575	16.66	16.34	95.06	90.51	90.50	81.90
LTH-Y92-66b-2	56 633 803	56 006 823	14.16	14.00	95.45	91.19	89.70	82.80
LTH-Y92-66b-3	58 289 663	57 632 744	14.57	14.41	95.55	91.34	91.70	85.00
AVERAGE	60 334 500	59 598 668	15.08	14.90	95.49	91.30	91.68	84.31
TOTAL	724 014 005	715 184 011	181.01	178.77

Among all expressed transcripts, 83.8% (28709) of the transcripts were common in both IRBLsh-S and LTH with Y92-66b and mock, demonstrating genetic resemblance based on their expression profiles (Fig. 2A). Moreover, a total of 1955 DEGs (610 specific up-regulated and 593 specific down-regulated) in IRBLsh-S, and 2377 DEGs (1037 specific up-regulated and 588 specific down-regulated) in LTH were identified (Fig. 2B; Additional file 2: Table S2 and Table S3). The number of DEGs showed that the transcriptional reprogramming emerged from IRBLsh-S and LTH after inoculated with Y92-66b. The identified DEGs in IRBLsh-S and LTH were used for functional enrichment analysis, respectively.

Fig. 2.

Expressed transcripts and DEGs in IRBLsh-S and LTH. A Venn diagram of expressed transcripts in two rice lines. B Significant differentially expressed mRNAs (log2 FC ≥ 1 or ≤-1 and p-value < 0.05)

There were 63 significantly enriched GO terms found among 1955 DEGs in IRBLsh-S, while 47 GO terms were enriched based on 2377 DEGs in LTH. Among these, 15 BP, 22 MF, and 6 CC were similar in the rice lines. Notably, 14 BP and 6 MF were specifically enriched in IRBLsh-S, including BP terms related to the “diterpene phytoalexin biosynthetic process” (GO:0051502), “phytoalexin biosynthetic process” (GO:0052315), “response to stimulus” (GO:0050896), “response to stress” (GO:0006950), “response to chemical stimulus” (GO:0042221), “response to oxidative stress” (GO:0006979), “polysaccharide metabolic process” (GO:0005976), “isoprenoid metabolic process” (GO:0006720), “terpenoid metabolic process” (GO:0006721), “cell wall macromolecule metabolic process” (GO:0044036), “diterpenoid metabolic process” (GO:0016101), “diterpene phytoalexin metabolic process” (GO:0051501), “phytoalexin metabolic process” (GO:0051501) (Fig. 3). Furthermore, 6 MF such as; “transition metal ion binding” (GO:0046914), “oxidoreductase activity, acting on peroxide as acceptor” (GO:0016684), “peroxidase activity” (GO:0004601), “antioxidant activity” (GO:0016209), “carbon-nitrogen lyase activity” (GO:0016840), and "manganese ion binding" (GO:0030145). In contrast, only 4 CC terms were mainly enriched in LTH (Fig. 3).

Fig. 3.

Enriched GO terms in IRBLsh-S and LTH. The number in the different color circles represents significantly enriched BP and MF pathways in IRBLsh-S and LTH

For KEGG pathways analysis, we found that the DEGs were significantly involved in osa01110 “biosynthesis of secondary metabolites”, osa00904 “diterpenoid biosynthesis” and osa00940 “phenylpropanoid biosynthesis” in IRBLsh-S (Table 2), which were consistent with GO enrichment. Moreover, the expressions of three genes were induced related to a synthetic pathway of a defense-related compound, momilactone A, belonging to the diterpenoid biosynthesis pathway (osa00904) in IRBLsh-S (Additional file 1: Fig. S3; Additional file 2: Table S4). Diterpenoid-related phytoalexin is often regarded as the typical defense response under ETI. The much higher regulation of genes related to momilactone A might be characterized as R gene Pish to trigger immunity in rice.

Table 2.

Different KEGG pathways present in IRBLsh-S based on DEGs in specific-enriched GO terms

Pathway ID	Pathway Description	Count in Gene Set	p-value	FDR
osa01110	Biosynthesis of secondary metabolites	33	4.60E-09	1.90E-07
osa00904	Diterpenoid biosynthesis	7	1.40E-06	2.90E-05
osa00940	Phenylpropanoid biosynthesis	11	7.20E-06	9.50E-05
osa01100	Metabolic pathways	30	3.80E-02	1.20E-01
osa00360	Phenylalanine metabolism	4	1.30E-02	1.00E-01
osa04075	Plant hormone signal transduction	8	1.40E-02	1.20E-01
osa00591	Linoleic acid metabolism	3	1.80E-02	2.00E-01
osa00944	Flavone and flavanol biosynthesis	2	4.20E-02	1.90E-01

Identified LncRNAs in Rice Leaves and Different LncRNAs Expression Profiling Between IRBLsh-S and LTH

The pipeline of lncRNA identification strategy can be seen in Fig. 4. A total of 1884 transcripts from 1730 gene loci was deemed the putative lncRNAs in rice among 111,608 transcripts according to the filtering criteria (Additional file 2: Table S5). To robustly validate the non-coding potential of these transcripts, all identified lncRNAs were analyzed with the CPC2 online tool, revealed that large number of these putative lncRNAs were confidently classified as non-coding by this independent tool, demonstrating the reliability of the final lncRNAs set (Additional file 2: Table S6). To further explore the novelty of our candidate, we compared all identified lncRNAs with Ensembl Plants database release 62 lncRNAs by using all extracted transcript IDs from reference and Cufflinks annotated file. The result revealed that 74.73% (1408) lncRNAs were novel and rest 25.26% (476) lncRNAs correspond to previously annotated lncRNAs. From which, 1147 lncRNAs overlapped with annotated genes (designated as OLR: ORF-overlapped lncRNAs), while the other 737 were Non-ORF-overlapped lncRNAs (NOL, implying new functional loci). These rice lncRNAs consist of three types, including 959 lincRNAs (392 NOL and 567 OLR), 368 antisense lncRNAs (300 NOL and 68 OLR), and 557 natural antisense transcripts (NATs) (45 NOL and 512 OLR). The antisense lncRNAs type showed the highest NOL ratio more than 81.5%, respectively (Fig. 5A). Moreover, the identified lncRNAs were distributed on all 12 chromosomes without a remarkable location bias (Fig. 5B). Therefore, many lncRNAs (79.3%) consistent of a single exon, while (18.84%) smaller portion contained two to four exons, whereas, a few lincRNAs (1.86%) were found to cover five to nine exons (Fig. 5C). The length of lncRNAs ranged from 200 to 9319 bp, and more than half (61%) of all lncRNAs were 200–1000 bp in length (Fig. 5D).

Fig. 4.

An integrative computational pipeline for the systematic identification of lncRNAs in rice. Coding Potential Calculator (CPC)

Fig. 5.

The types of lncRNAs in rice seedlings leaves and their characters. A Percentages of different types of lncRNAs; B Distribution of different types of lncRNAs on different chromosomes; C Number of exons in different types of lncRNAs; D Distribution of lncRNAs with different lengths

To Screen the LncRNAs Correlated with Rice Responses to Fungus, Analyzed All LncRNAs by T-Test

There were 317 differently expressed lncRNAs (DELs) identified in rice responding to blast fungus, accounting for 16.8% of all lncRNAs in our data (Fig. 6A), and 45.1% of these DELs belonging to NOL. We found 59 DELs (36 up-regulated and 23 down-regulated) and 92 DELs (25 up-regulated and 67 down-regulated) specifically expressed in IRBLsh-S and LTH, respectively. However, only 10 DELs were shared in the rice lines with similar expression patterns (Fig. 6B, Additional file 2: Table S2 and Table S3). Therefore, the induction of specific lncRNAs between IRBLsh-S and LTH in response to blast fungus implied that the roles of lncRNAs in immunity regulated by the R gene could not be neglected.

Fig. 6.

DELs in IRBLsh-S and LTH. A Differentially expressed lncRNAs (DELs) in IRBLsh-S and LTH (p-value < 0.05); B Significant DELs were selected on the base of (log₂ FC ≥ 1 or ≤ -1 and p-value < 0.05)

Co-expression Modules of DELs and DEGs Related to Rice Resistance

The functions of lncRNAs can be deduced by coding-non-coding co-expression networks (Liao et al. 2011; Hao et al. 2015). To capture the potential function of lncRNAs involved in rice resistance to blast fungus, we classified the whole transcripts, including DELs and DEGs, based on their expression patterns in all samples by WGCNA. We obtained five modules in IRBLsh-S (Fig. 7: Additional file 2: Table S2) and five modules in LTH (Additional file 2: Table S3), which suggests the DELs and DEGs with similar expression patterns in each module (Additional file 2: Table S7). Moreover, we performed GO enrichment and KEGG pathway using DEGs and provided references to explore the functions of lncRNAs in each module.

Fig. 7.

Modules of significant differential transcripts in IRBLsh-S identified by WGCNA. The color of each cell represents the correlation coefficient between the modules and traits

In IRBLsh-S line, three modules (M1, M2, and M3) significantly enriched the GO terms (Additional file 2: Table S8). Especially, the genes belonging to immunity-related GO terms, such as cell wall macromolecule catabolic process and diterpene phytoalexin biosynthetic process were found in M2 and M3, respectively (Additional file 2: Table S8), indicated that the transcripts in M2 and M3 were involved in Pish regulating defense and considered to deepen the functions of lncRNAs in these two modules. Compared with DEGs belonging to M2 and M3 in LTH inoculated by blast fungus, the regulated directions of DEGs were identical in IRBLsh-S after inoculation. However, greater changes in expression levels of DEGs were found in IRBLsh-S. There were 4 DELs and 189 DEGs clustered in M2 and 2 GO terms connected with resistance in IRBLsh-S about the “cell wall macromolecule metabolic process” and “polysaccharide metabolic process” significantly enhanced in this module.

Moreover, many transcripts (49 DELs, 1407 DEGs) were enriched in M3. These transcripts were significantly enriched in 65 GO terms (Additional file 2: Table S8). Interestingly, 13 GO terms were consistent with resistance related to biological process (BP) and molecular function (MF) in IRBLsh-S based on all DEGs, including genes involved in “diterpene phytoalexin biosynthetic process”, “phytoalexin biosynthetic process”, “response to stimulus”, “response to stress” and “transition metal ion binding”. To comprehend the function of DELs and DEGs in M3, we performed pathway analysis of genes in M3 and found that mainly genes were involved in biosynthesis of secondary metabolites, diterpenoid biosynthesis, and phenylpropanoid biosynthesis pathways (Additional file 2: Table S9). Therefore, the genes involving diterpenoid biosynthesis were mainly associated with the metabolic pathway of momilactone A. Moreover, the results showed that lncRNAs in rice could regulate innate immunity response to blast fungus, such as momilactone A and cell wall.

The Networks of DELs, DEGs and TFs in Rice Innate Immunity to Fungus

To further elucidate the lncRNA regulating networks in IRBLsh-S responding to M. oryzae, we analyzed the relationships among DELs, DEGs, and TFs of different GO terms in M2 and M3 by co-expression analysis and network was visualized by using Cytoscape software. We found that momilactone A synthesis genes were regulated by 12 lncRNAs, and 29 TFs belong to the MYB family, the WRKY family, and the bHLH family (Fig. 8A). The DEGs related to the cell wall in M2 demonstrated that two induced genes and one suppressed gene specifically were highly correlated with lncRNAs (XLOC_033508; XLOC_039551; XLOC_011573) and TFs (XLOC_040281; XLOC_026879) in IRBLsh-S (Fig. 8B). The expression level of these lncRNAs and TFs were validated by qRT-PCR (Additional file 2: Table S4). The number of transcripts with consistent expression levels exceeded 85%. The results demonstrated that lncRNAs participated in rice innate immunity to fungus via inducing the expressions of momilactone A and cell wall-related DEGs and TFs in IRBLsh-S.

Fig. 8.

Transcripts network visualization and heapmaps in M2 and M3. A Network visualization of 3 genes about momilactone A and connected lncRNAs and transcriptional factors (TFs). B Heatmap and network visualization of differentially expressed genes (DEGs) in special GO terms, lncRNAs, and TFs of module M2

Prediction of LncRNAs as Endogenous Target Mimic for miRNAs

LncRNA could act as an Endogenous Target Mimic (eTMs) to compete with endogenous RNAs targeted by miRNA and avoid inaccurate expression of transcripts regulated by miRNAs (Wu and Wang 2013). Therefore, we analyzed the 53 lncRNAs in M2 and M3 as eTMs using RegRNA 2.0. In total, 10 lncRNAs (9 lncRNAs in M3 and 1 lncRNA from M2) were predicted to be the potential eTMs for 12 miRNAs (Additional file 2: Table S10). To further explore the interaction between the lncRNAs, miRNAs, mRNAs, we constructed a competing endogenous RNA (ceRNA) network (Additional file 1: Fig. S5). These lncRNAs were co-expressed for 81 DEGs about regulating transcription, transport, and response to stress (Additional file 2: Table S11). Moreover, our results demonstrated that there were two lncRNAs (XLOC_040277 and XLOC_046130) co-expressing with the genes of momilactone A binding miRNAs (osa-miR2927, osa-miR1433, osa-miR169q), implying these two lncRNAs might be as eTMs of 3 miRNAs to ensure the correct expressions of momilactone A related genes in IRBLsh-S.

Discussion

The effective deployment of disease resistance genes necessitates systematic elucidation of their underlying resistance mechanisms, with particular emphasis on transcriptional regulation. Remarkable progress has been made in rice blast resistance research, including the identification and cloned of many resistance (R) genes. The protein-coding genes (PCGs) involved in upstream and downstream regulatory networks of these R genes have been extensively characterized (Li et al. 2023). While recent studies have revealed exciting findings regarding the regulatory roles of non-coding RNAs (particularly miRNAs) in defense pathways, the involvement of long non-coding RNAs (lncRNAs) in rice blast resistance regulation remains poorly understood. Using monogenetic lines harbored different R genes favors deepening the resistance mechanism characterized mediated by various R genes. This study evaluated LTH and its monogenic line, IRBLsh-S, to represent the compatible and incompatible interaction with M. oryzae. We constructed an ssRNA-Seq library to identify DELs and DEGs related to rice defense, revealed that 83.8% of all expressed transcripts were common in both IRBLsh-S and LTH, which was about 10% higher than the 74.0% and 74.2% reported in previous studies (Bagnaresi et al. 2012; Jain et al. 2017), due to their highly conserved genetic background. Pish was localized on chromosome 1 and isolated by extensive characterization of retrotransposon-tagged suppressive mutants (Takahashi et al. 2010). Therefore, further research for understanding Pish regulation and functional roles and downstream events is critical for utilizing the loci in agricultural practice. Previously, it was reported that different varieties carrying Pish gene showed high resistance against multiple strains of M. oryzae (Khan et al. 2014).

In contrast, Ngernmuen et al. (2020) reported that Pish and Pik rice line revealed unique DEGs, which were related to defense-related pathways and phytohormones, including chitinase substilin, brassinosteroids, jasmonic acid (JA) and salicylic acid (SA), showed high resistance against blast pathogen. In another study, the findings suggest that integrating several R genes, including Pish and Pi64, can exhibit durable blast resistance against multiple isolates of M. oryzae (Ma et al. 2015). Expression studies on rice lines carrying different blast resistance genes, including Pish revealed the role of a plethora of transcriptional regulons acting through multiple signalling kinases and transcription factors (TFs) upon infection by different M. oryzae strains (Sureshkumar et al. 2019).

The BL1 variety carrying Pib and Pish genes exhibited different expressions and broad-spectrum resistance against various Philippine isolates of M. oryzae (Kato et al. 2004). Another study also found that the varieties harboring Pish and Pi54 genes conferred broad-spectrum resistance against blast pathogen, while only Pish carrying variety displayed moderate resistance to isolates (Xiao et al. 2018). These results implying Pish possibly contributing superficially while co-expression with other resistance genes. The near-isogenic line (Pish) highly consistent with the genetic background of the susceptible parent may be more conducive to screening the associated genetic elements contributing to disease resistance, so the mechanism of Pish is needed for further research.

lncRNAs increasingly play an important role in plants challenges with biotic and abiotic stresses (Zhu et al. 2014). Our results showed that 1884 lncRNAs in rice seedlings and 317 in Pish-line were identified. The basic features were consistent with those identified in other studies, including length and covering exons (Zhang et al. 2014; Li et al. 2017). Among these lncRNAs, only 304 (16.14%) and 610 (32.4%) were not found in previous results which focus on the roles of lncRNA on sexual reproduction and agronomic traits (Zhang et al. 2014; Wang et al. 2015). Our study suggested that lncRNAs were differently and specifically reprogrammed by infection of pathogens. The responsive lncRNAs to blast fungus infection in rice were identified recently (Wang et al. 2020). Compared with their results, only a few lncRNAs were the same. Our results, combined with the previous reports, suggest that lncRNAs are like other regulators of disease resistance, can be specifically derived by rice genotype, pathogen isolate and environment conditions, and therefore need to be identified systemically.

Resistant DEGs were enriched in IRBLsh-S at 24 hpi in R gene Pish-derived manner and via improving the expressions of phytoalexin synthesis and cell wall-related genes. Moreover, our results show that lncRNAs are massively responsive to blast fungus infection at 24 hpi, implying unneglected roles of lncRNAs in immunity. The results of the present study are consistent with another study who reported lncRNAs are pivotal regulators and played important defense regulating roles across different plant-microbe interaction (Zhu et al. 2014; Cui et al. 2017). For instance, it has been reported that many lncRNAs have been involved in Fusarium wilt defense response by overlapping and regulation of FOC-responsive defense-related genes expression (Cheng et al. 2021). In cotton, numerous Fov-regulated lncRNAs positively correlate with the highest susceptibility to disease resistance, and associated with many defense-related pathways, indicating the crucial role of lncRNAs in the response to pathogen infection and the establishment of disease resistance (Yao et al. 2019). While the silencing of lncRNACXE20 decrease the expression pattern of carboxylesterase-related genes, thereby increasing the disease resistance against Botrytis cinerea in tomato plants (Chen et al. 2024). Moreover, a comprehensive analysis of RNA-Seq data revealed that the lncRNAs were correlate with TFs, and each lncRNAs regulated by its corresponding TFs can activate the mRNA targeted by the TF, indicating the involvement of TFs in mRNA expression by the associated lncRNAs, which enhance resistance against different viruses in rice (Cao et al. 2022), highlighting the sophisticated layered control of immunity in plants, of which our lncRNAs are likely a part. However, the functions of lncRNAs in rice resistant response to blast fungus remained unknown largely (Wang et al. 2020). Thus, we applied a WGCNA approach to classifying different clusters based on their expression in IRBLsh-S and LTH, which favored conducting the function annotation of specific lncRNAs (Li et al. 2017). Comparing the Pish-gene harboring line with recurrent susceptible line LTH, 53 lncRNAs were involved in Pish-mediated resistance via regulated diterpene phytoalexins and cell walls. This finding aligns with the growing perspectives that lncRNAs serves as key regulators of particular defense mechanisms. For instance, Yu et al. (2020) identified 567 disease-related lncRNAs, among them 73 were correlated with the 39 JA-related PCG, underscoring the significant role of lncRNAs in the modulation of JA pathway. Therefore, the functional analysis revealed that the overexpression of ALEX1 is participate in the JA pathway activation in rice against Xanthomonas oryzae (Yu et al. 2020). While the analysis of ceRNAs network in rice revealed that lncRNAs can regulate the expression of genes associated with plant-pathogen interaction, hormone signaling pathways, and TFs by acting as miRNAs decoys (Li et al. 2023). This indicated that resistance gene Pish could recruit momilactone A to cope with the blast fungus. Phytoalexin is a basal defense component, and priming phytoalexin biosynthesis is a critical response in resistance lines to blast fungus (Han et al. 2015). Diterpene phytoalexins are a major phytoalexin in rice and were accumulated after blast fungus inoculation more rapidly and highly in resistant lines than in susceptible lines (Hasegawa et al. 2010). Momilactone A was reported to display anti-M. oryzae activity (Cartwright et al. 1977), no momilactones were detected in healthy or dry rice leaves but could be detected in the blast fungus-infected leaves (Umemura et al. 2003). Several types of research indicated that momilactone A played roles in rice defence against different fungal and bacterial pathogens (Tamogami and Kodama 2000; Okada et al. 2007; Hasegawa et al. 2010). Moreover, the identification of these above 53 lncRNAs specifically associated with the Pish R gene, represent a significant finding of this study, indicating a high degree of regulatory specificity. This suggests that the immune responses in plants are not solely activated by R gene itself, but rather is modulated by a specialized set of lncRNAs that coordinate downstream defense responses and mechanisms, particularly those associated with the diterpene phytoalexin and cell wall reinforcement pathways. In addition, this Pish-specific regulatory defense and immunity-related layer, unique to the Pish R gene, may facilitate a more precise and efficient defense method, by directing the plant energy and metabolic resources towards the most effective pathways, rather than initiating a large and non-specific transcriptional reprogramming in rice.

Furthermore, the Os-CPS4 momilactone A-coding gene was also found in the current study, plays a significant role in fungal non-host disease resistance (Lu et al. 2018) and could be elicited by brown planthopper honeydew-associated symbiotic microbes (Wari et al. 2019). In addition, momilactone A has a wide range of functions, showing a broad-spectrum resistance. Momilactone A has been patented as an herbicide (Chung et al. 2006) based on its allelochemicals properties, and it showed highly antibacterial activity against different bacteria such as Bacillus cereus, B. pumilus, Escherichia coli and Pseudomonas ovalis (Fukuta et al. 2007). Momilactone-like compounds could play an important role in crop-friendly herbicides and anti-fungal and antibacterial agents (Minh and Xuan 2024). These results suggested that lncRNAs found in this study potentially play the roles for basal defence to blast fungus and have multiple regulation function to response to other stresses.

The results of the present study demonstrated that regulating the function of momilactone A via combining lncRNAs and other pathways is promising in rice resistance breeding. In addition, previous research indicated that the knocked-down expression of Os-cps4 played a significant role in rice defense through OSCPS4-dependent labdane-related diterpenoids against M. oryzae (Toyomasu et al. 2014). Therefore, another study found that the resistance variety produces high momilactones and phytocassanes compared to susceptibility at the starting point (Hasegawa et al. 2010). Consequently, such decreased and late production (i.e., even in the parental/wild-type line) may cause the detected absence of effect with Os-cps4 plants (Lu et al. 2018). Four lncRNAs significantly related to DEGs involved polysaccharide and cell wall macromolecule in M2 of IRBLsh-S. Studies showed that the cell wall-associated defense plays a crucial role in plants basal resistance (Hückelhoven 2007), and genes of this term were enriched in resistant plants when responding to stress, including fungi and bacteria (Du et al. 2015; Xu et al. 2016). Molina et al. (2021) reported that cell wall composition regulates plant resistance against diseases. Moreover, a study showed that trisaccharide 3¹-β-D-Cellobiosyl-glucose and tetrasaccharide 3¹-β-D-Cellotriosyl-glucose are specific DAMPs released from the hemicellulose of the rice cell wall, which is perceived by an OsCERK1 and OsCEBiP immune complex during M. oryzae infection in rice (Yang et al. 2021). The present study results identified the cell wall genes related to the resistance in rice harboring Pish to M. oryzae. They showed that lncRNAs may regulate the basal immunity of IRBLsh-S via the genes involved with cell walls.

In plant-microbe interaction, lncRNAs are associated with many defense responses in plants such as; systemic acquired resistance, salicylic acid-mediated defense and innate immunity response (Li et al. 2017; Wang et al. 2017), cellulose biosynthesis (Huang et al. 2016), ROS scavenging (Cui et al. 2017). Our results indicated that lncRNAs regulated multiple aspects of rice resistance to M. oryzae. It has been reported that transcription factors (TFs) are involved in the plant immune responses between the interaction of M. oryzae and rice, especially WRKYs, MYB, and bZIP (Inoue et al. 2013; Wei et al. 2013; Jain et al. 2017). We identified that the networks with TFs, lncRNAs and mRNAs are associated with rice resistance in present study. Genes were up-regulated significantly by 29 TFs to regulate momilactone A synthesis, and 3 cell wall genes were also significantly related to 2 TFs. These TFs belong to WRKY and MYB families mainly. It has been reported that lncRNA could affect miRNA by competing for the binding to their mRNA-binding sites (Yamamura et al. 2018). A lncRNA (IPS1) was demonstrated to be associated with miRNA (miR399) regulation as target mimicry in Arabidopsis thaliana (Wu and Wang 2013). Some lncRNAs as eTMs for miRNA in several plant species were identified by genome-wide analyses to affect the development and response to stress (Shumayla et al. 2017; Huang et al. 2018). Subsequently, the current study investigated that lncRNAs in rice seedlings may regulate innate immunity to blast fungus via miRNAs, especially by networks of lncRNAs, genes, TFs and miRNAs. Some lncRNAs may played resistant roles probably by other ways, such as interaction with protein and nucleic acids (Kopp and Mendell 2018).

Conclusions

We identified 1884 lncRNAs in rice seedlings using ssRNA-Seq, 317 lncRNAs involved in the interaction between rice and blast fungus strain Y92-66b, and many of these loci are newly detected. We found fifty-three DELs in IRBLsh-S after inoculation, that are specific lncRNAs associated with Pish-mediated resistance. While, their exact resistance mechanism remains elusive, however, the specific induction of these lncRNAs during the defense response suggests an important role in modulating the rice phytoalexin momilactone A, which is crucial and recognized component of basic defense in rice (Fig. 9). Future work will focus on determining whether these lncRNAs function by directly co-regulating biosynthesis genes or by appropriating miRNAs that target this pathway. Further, our data suggest that some lncRNAs might be the critical regulation factors linking to networks of lncRNAs, genes, TFs and miRNAs. The current study provided insights for understanding regulation by R genes, the context of lncRNAs in breeding, and agricultural practice for disease management. Based on transcriptome analysis of Pish harboring near-isogenic line IRBLsh-S and its parental LTH at 24-hour post-inoculation with Magnaporthe oryzae, we identified lncRNAs participating in basal defense processes during fungal infection, including oxidative stress modulation, diterpene phytoalexin biosynthesis (particularly momilactone A production), and cell wall reinforcement coordination. These lncRNAs were further found to interact with networks of protein-coding genes (PCGs), transcription factors (TFs), and miRNAs, demonstrating their essential regulatory roles in Pish-mediated blast resistance. Our findings establish lncRNAs as pivotal regulators in the Pish-conferred resistance pathway. Future investigations distinguishing: (i) Pish-specific regulators, (ii) shared regulators among different resistance genes, and (iii) sustained regulators during prolonged defense responses will provide crucial scientific foundations for their application in resistance breeding and pest management practice of rice.

Fig. 9.

The putative network about lncRNAs to mediate rice resistance to M. oryzae derived by R-gene Pish. Different color showed that lncRNAs are co-expressed closely with genes enriched in biological process and molecular functions especially response to oxidative stress and diterpene phytoalexin biosynthetic process. Two lncRNAs (XLOC_046130 and XLOC_040277) are involved in the regulation of momilactone A genes, which are recognized as a component of basic defense in rice immunity

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

CPC

Coding potential calculator

DEGs

Differentially expressed genes

DELs

Differently expressed lncRNAs

ETI

Effector triggered immunity

IncRNA

Intronic lncRNA

JA

Jasmonic acid

lncRNAs

Long noncoding RNAs

NAT

Antisense lncRNA

NOL

Non ORF-overlapped lncRNAs

OLR

Overlapped lncRNAs

ORF

Open reading frame

RSEM

RNA-seq by expectation-maximization

Author Contributions

LL and OI: Investigation, Methodology, Data curation and Writing-original draft of manuscript. BY: Formal Analysis, Writing– review & editing. YX: Writing– review & editing. PY: Software, Writing– review & editing. MHRH: Formal Analysis, Writing– review & editing. WY: Formal analysis, Supervision, Writing– review & editing. LC: Conceptualization, Supervision, Formal Analysis, Funding acquisition, Project administration, Writing– review & editing. All authors have read and approved the final manuscript.

Funding

This work was supported by the National Key R&D Program of China (2023YFD1400800). The Talents Training Program in Yunnan (YNYC-2022-0217). Major Science and Technology Projects in Yunnan (202402AE090026). The National Natural Science Foundation of China (32560630, 32202254). The Yunnan Provincial Talent Program for Wang Yi (XDYC.QNRC 2023 − 0421).

Data Availability

The datasets presented in this study can be found in online repositories with sequence archive number (SRX10430967-SRX10430978) under the BioProject number PRJNA716694. Further inquiries can be directed to the corresponding author.

Declarations

Ethics Approval and Consent to Participate

Not applicable.

Consent for Publication

Not applicable.

Competing Interests

The authors declare no competing interests.