The transcriptomic data utilized for lncRNA identification were derived from Liu et al. (2015) (DOI: 10.1186/s12870-015-0511-8). This dataset was originally generated to study the transcriptional response of wheat (Triticum aestivum cv. TAM 107) to drought stress, heat stress (Hirose et al.), and combined heat-drought stress (Sahraei et al.). The data are publicly available under NCBI SRA accession number SRP045409 (Liu et al., 2015). TopHat and Cufflinks software were used for transcriptome mapping and assembly (Trapnell et al., 2012). The sequences were screened for protein-coding transcripts and other non-coding RNA families using Blastx, Rfam and the UTR databases (http://utrdb.ba.itb.cnr.it/). The assembled transcripts from this method were used to identify lncRNAs based on these criteria: the transcript should be longer than 200 bp and have an open reading frame < 100 bp; the coding potential calculator value should be < −1. Gene expression levels were determined using FPKM value. Heat maps were created with the log2-transformed FPKM values using Sangerbox software.

Potential cis-target genes were identified as those within 100 kb upstream and downstream of the lncRNA (Zhang et al., 2020). Potential trans-target genes were predicted based on sequence complementarity using the LncTar program (http://www.cuilab.cn/lnctar). The wheat genome database was consulted for functional annotations of target genes, which were then analyzed using the Blast2GO and WEGO programs.

PsRobot was used to predict 323 lncRNAs which significantly responsive to drought stress targeted by miRNAs (Wu et al., 2012). Pearson’s correlation coefficients were used to analyze the expression of lncRNAs and mRNAs, and ceRNAs were identified using the ceRNA-score principle (Shu et al., 2019). lncRNAs and mRNAs were selected based on having more than three shared miRNAs, with p-value < 0.05, and FDR value < 0.1. The ceRNA network was constructed by selecting pairs exhibiting negative correlations between miRNAs-mRNAs and positive correlations between lncRNA-mRNAs. The ceRNA network was analyzed and visualized using Cytoscape.

Wheat ‘Chinese Spring’ plants were used in the study. The plants were cultivated in soil within a greenhouse kept at 25 ± 2 °C with a 16-hour day and 8-hour night light cycle. We treated 10-day-old seedlings with 20% (w/v) PEG6000 in 6h to simulate drought stress (Hu et al., 2013). Seedlings cultivated in a normal (non-stress) environment served as the control group. Leaves of the seedlings were collected for RNA isolation and gene expression analysis.

Gene-specific primers were designed from the nucleotide sequence (Traes_5DL_7D14C1A38.1) using Primer6 software and subsequently used to amplify the full TalncR9 sequence from wheat cDNA. Following sequencing, the suitable fragments were attached to the tobacco rattle virus RNA2 (TRV2) vector. The recombinant plasmids were then transferred into Agrobacterium tumefaciens strain GV3101. The TRV2 empty vector was used as a negative control and the phytoene desaturase (TaPDS) gene was used as a positive control, which produced a photobleached leaf phenotype. The experiment was performed using the whole-plant silencing method with minor modifications (Zhang et al., 2017). Three experimental replicates were inoculated, and each replicate consisted of more than 20 seedlings. Approximately 2 weeks after virus inoculation, when the PDS-silenced seedlings had become white, total RNA was isolated from leaves of both the silenced and control plants for gene expression analyses. The list of primer sequences utilized is provided in Supplementary Table S1 .

In accordance with previously described methods, two subsets of plants were maintained for drought-stress treatment. The control group was cultivated under a normal (non-stress) environment and the other seedlings were planted at 50% field capacity by withholding water during the entire duration of the study (Manmathan et al., 2013). Leaves from seedlings grown under the non-stress and drought treatments were sampled for physiological indicator measurements.

The complete sequence of TalncR9 was cloned into a pCAMBIA1304 vector to create overexpression construction. The pCAMBIA1304-TalncR9 vector and empty vector were introduced into the A.tumefaciens EHA105 strain for transformation of Arabidopsis (ecotype Columbia 0) via the floral dip technique. The identification of transgenic lines was achieved by DNA PCR. Depending on the transgene expression level, two independent homozygous lines of TalncR9-Overexpressed plants (L-5 and L-9) of the T₃ generation were selected for subsequent drought-stress tolerance assays.

To evaluate drought stress tolerance, sterilized seeds from wild type (WT), vector control (VC), and transgenic lines L-5 and L-9 were germinated in a greenhouse on 1/2 MS medium, with and without mannitol concentrations of 150 mM or 300 mM (Ge et al., 2024). Seed germination rates were recorded for a week. Following 10 days of vertical growth, the length of primary roots was measured. Subsequently, seedlings were collected for RNA extraction and physiological indicators measurements.

Physiological indicators were measured using a malondialdehyde (MDA), proline assay, soluble sugar content test, and peroxidase (POD) assay kit (Jiancheng, China) in accordance with the manufacturer’s instructions. Each sample was analyzed in three biological replicates.

Total RNA was extracted from the samples and reverse-transcribed with the RNAprep Pure Plant Kit and FastKing cDNA Kit (Tiangen, Beijing, China). Gene expression levels were estimated by qRT-PCR assays, which were performed using gene-specific primers ( Supplementary Table S1 ) on a CFX real time PCR machine (Bio-Rad, Hercules, CA, USA). The qRT-PCR primers’ specificity and efficiency were confirmed through melting curve analysis and agarose gel electrophoresis, with each primer pair showing a PCR efficiency of 90%–110%. β-Actin were used as the internal reference gene for wheat and Arabidopsis, respectively. Relative gene expression levels were determined using the 2^−ΔΔCt method. All samples were analyzed with three biological replicates and the data were represented as mean ± standard deviation (SD).

RNA-seq data of wheat under heat and drought stress conditions, obtained from the NCBI sequence read archive database, were used to identify lncRNAs (https://www.ncbi.nlm.nih.gov/sra, accession number SRP045409).

After de novo data assembly and screening, 2830 lncRNA sequences were identified from a total of 101,245 transcripts within the heat- and drought-treatment SRA data. In addition, 2485, 2830, and 2350 lncRNAs were responsive to heat, drought, and heat plus drought stress, respectively ( Figure 1A ). And the 2830 drought-responsive lncRNAs also included 2485 heat-responsive lncRNAs and 2350 drought-heat-responsive lncRNAs. The expression profiles of lncRNAs were analyzed using RNA-seq data. ( Figure 1B ; Supplementary Table S2 ). As the expression levels of a higher number of lncRNAs were significantly changed after 6h of heat or drought treatment than after 1h of the same treatments, the subsequent analyses were based on 6h of stress treatments. The expression levels of 8.0%, 10.1%, and 9.6% of the lncRNAs were significantly up-regulated by heat, drought, and drought plus heat stresses, respectively (value > 1). In contrast, the expression levels of 3.2%, 1.3%, and 3.2% of the lncRNAs were significantly down-regulated by the same stresses, respectively (value < −1) ( Figures 1C–E ). Collectively, 323 significant drought-responsive lncRNAs (value > 1 or < −1) were identified in wheat for further study ( Supplementary Table S3 ).

LncRNA can regulate the expression of target genes by binding to DNA, RNA and protein, and its regulatory modes can be divided into cis-regulation and trans-regulation (Font-Cunill et al., 2018). A total of 6970 target genes were predicted for 323 significant drought-responsive lncRNAs, including 2532 cis-target and 4438 trans-target genes. To identify biological functions, the predicted target genes were organized into GO term classifications. ( Figure 2A ). Consequently, 4284 target genes were classified into the three primary GO categories: biological process (BP), cellular component (CC), and molecular function (MF). In the BP category, the dominant categories are cellular process, metabolic process, and response to stimulus. The most highly represented GO terms in the CC category were cell part, cell and organelle. For the MF category, binding and catalytic activity were the most abundant GO terms. The target genes were also identified to be enriched in additional stress-related terms, including response to jasmonic acid, response to karrikin, response to high light intensity and divalent metal ion transport. These annotations implied that the predicted target genes of drought-responsive lncRNAs are involved in extensive metabolic activities, especially in stress responses.

A metabolic pathway enrichment analysis was performed using the KEGG database to explore the target genes of drought-responsive lncRNAs. We identified 20 significantly enriched KEGG pathways ( Figures 2B, C ). The top three pathways included carbon metabolism (115 target genes), biosynthesis of amino acids (91 target genes), and plant hormone signal transduction (64 target genes). Among these, the pathways most enriched were peroxisome and the metabolism of cysteine and methionine, which are vital for drought response. Peroxisomes help to detoxify reactive oxygen species (ROS), which accumulate during drought stress and can cause oxidative damage (del Río et al., 2002). Cysteine and methionine are involved in the synthesis of antioxidants and redox-active compounds, which are involved in drought stress by regulation of redox state (Reest et al., 2018). These KEGG annotations served as a valuable tool for exploring pathways linked to drought resistance in wheat.

While some lncRNAs may be degraded by miRNAs (Borah et al., 2018), in the context of target mimicry, lncRNAs competitively bind miRNAs to modulate their activity toward target mRNAs. The psRobot software was utilized to predict possible lncRNA-miRNA target mimics. The results revealed that 56 lncRNAs exhibiting differential expression could act as target mimics for 38 unique miRNAs (see Supplementary Table S4 ). The majority of lncRNAs were associated with a single miRNA, whereas certain miRNAs influenced multiple lncRNAs. For example, tae-miR399, tae-miR9778 and tae-miR9780 targeted four lncRNAs, respectively. Consequently, lncRNAs could substantially influence miRNA activity.

A network involving lncRNA, miRNA, and mRNA was constructed in wheat by analysis of 323 lncRNAs which significantly responsive to drought stress. Using the identified mRNAs associated with drought response, we created a network comprising 19 lncRNA-miRNA-mRNA, including 9 lncRNAs, 6 miRNAs, and 14 mRNAs. ( Figure 3 ; Supplementary Table S5 ). In this network, 2 target genes (TaRPM1 and TaNBS-LRR1) belong to the NBS-LRR protein family of plants, which are usually responsible for recognizing pathogens and activating downstream immune signaling pathways, could enhance plant disease resistance. These two target genes with 2 lncRNAs (TalncR296 and TalncR302) were targeted by tae-miR9778, involved in ceRNA interactions. Additionally, indole-3-acetic acid-amido synthetase (TaGH3.8) was identified had function in the metabolism of the plant auxin (IAA), and potassium transporter 25 protein (TaKT25) had function in potassium ion transmembrane transporter activity. These 2 target genes along with 3 lncRNAs (TalncR27, TalncR242 and TalncR256), were targeted by tae-miR5384-3p. TalncR9 could targeted 4 genes by tae-miR167a, in which 2 target genes (TaE3–1 and TaE3-2) were identified as RING-type E3 ubiquitin transferase, which played key roles in plant development, stress response, and signal transduction through ubiquitination modification. The other 2 target genes (TaARF12–1 and TaARF12-2) were identified as auxin response factor, which participated in plant development and hormone signal transduction by responding to auxin. TalncR124, TalncR99 and TalncR181 also involved in ceRNA networks by targeting 1 to 3 target genes, respectively. The ceRNA interactions analysis indicated that lncRNA may play significant roles within ceRNA networks by modulating miRNAs and their downstream target genes in response to drought stress.

To confirm the RNA-seq data, the 9 lncRNAs involved in ceRNA networks and another 10 most significant up-regulated and down-regulated lncRNAs in response to drought stress were selected for qRT-PCR analysis. After normalization, the expression patterns of the chosen lncRNAs exhibited a similar trend, in which 12 lncRNAs were both up-regulated, and 7 lncRNAs were both down-regulated in RNA-seq data and qRT-PCR analyses. These outcomes suggest that RNA-seq data serve as a reliable source for revealing the expression profiles of lncRNAs in wheat ( Figure 4 ).

As the qRT-PCR indicated that the expression level of TalncR9 was increased under drought treatment, and it was involved in ceRNA networks, TalncR9 may be a potentially significant functional gene implicated in the drought-stress response of wheat. We cloned TalncR9 for further functional study. The gene-silenced plants of TalncR9 were constructed by Virus-induced gene silencing (VIGS) technology. The VIGS: TalncR9 vector was used for inoculation of the wheat seedlings’ two-leaf stage leaves. VIGS: TaPDS and VIGS: TRV2 (CK-TRV2) were served as positive and negative controls, respectively. The photobleached phenotype was observed on the leaves of VIGS: TaPDS plants, which indicated that the VIGS system was effective. No differences in the morphology of the VIGS seedlings and the CK (Control Check) seedlings were observed. The qRT-PCR analysis indicated that TalncR9’s relative expression level was lower in the majority of VIGS seedlings than in the CK and CK-TRV2 seedlings. ( Figure 5A ). Due to insufficient biomass in individual silenced lines, a pooled sample comprising 20 independent TalncR9-silenced lines was used for analysis. After drought treatment, all plants exhibited increased wilting severity. However, the symptoms of VIGS: TalncR9 plants exhibited more severity compared to the CK and CK-TRV2 plants ( Figure 5B ). Most VIGS: TalncR9 plants had severely wilted or desiccated leaves, whereas few leaves of the CK and CK-TRV2 seedlings showed yellowing or curling. Physiological indicators—including soluble sugar content, proline content, malondialdehyde (MDA) contents, and peroxidase (POD) activity were measured under drought stress ( Figures 5C–F ). Prior to drought stress, no significant differences were observed in any indicator across the control (CK), empty vector control (CK-TRV2), and TalncR9-silenced plants. However, following drought stress: In control groups (CK and CK-TRV2), soluble sugar, proline, and POD activity significantly increased (p<0.05), while MDA content rose moderately. Conversely, in TalncR9-silenced plants: Soluble sugar, proline, and POD activity showed either attenuated increases or no significant change, and were significantly lower than controls (p<0.05). Moreover, MDA content exhibited a greater increase, resulting in higher levels than both controls (p<0.01). Collectively, these results demonstrate that silencing TalncR9 compromises drought tolerance in wheat by impairing osmotic adjustment (via sugar/proline) and antioxidant capacity (via POD), while simultaneously exacerbating membrane lipid peroxidation (via MDA).

To investigate the role of TalncR9, we generated transgenic Arabidopsis lines overexpressing TalncR9. The germination rate and root elongation were examined to investigate the function of TalncR9 in transgenic plants under drought conditions ( Figure 6 ). For the germination rate tests, there were no differences in seed germination rates between transgenic lines and controls (WT, VC) on MS medium under normal conditions. However, the transgenic lines exhibited a greater germination rate compared to the controls on 1/2 MS medium containing 150 mM and 300 mM mannitol. ( Figures 6A, C ). For the root elongation experiment, both control and TalncR9 transgenic lines grew similarly without treatment, while the transgenic seedlings exhibited greater root lengths compared to the control groups (WT and VC) under mannitol treatments ( Figures 6B, D ).

To examine the physiological processes related to the role of TalncR9 in drought stress response, the soluble sugar, MDA, proline contents and POD activity were assessed in the transgenic plants and controls. Before drought stress, no significant differences were detected for all analyzed indicators among the WT, VC, and TalncR9 transgenic lines. After drought stress, the TalncR9 transgenic plants exhibited lower levels of MDA contents, higher levels of soluble sugar, proline contents, and POD activity compared with the controls (WT and VC) under drought condition ( Figures 6E–H ). The results suggested that overexpression of TalncR9 improved the drought resistance in transgenic Arabidopsis.

The transcript levels of genes involved in drought stress were detected in TalncR9 transgenic lines and controls ( Figure 7 ). Before drought stress, the expression levels showed no significant differences for all analyzed genes among all the lines. Under drought stress, the AtLEA30, AtERD1A, AtDREB2 and AtSPS1 genes were up-regulated, whereas the gene expression levels in the TalncR9 transgenic plants were higher than in the controls (WT and VC). These findings indicate that TalncR9 enhances drought tolerance through up-regulation of drought-related genes in transgenic plants.