We selected the drought-tolerant spring wheat pure line BM14 (Bamai14) as the experimental material, and a widely cultivated spring wheat variety NC4 (NingChun4) as the control. Seeds of both genotypes were kindly provided by Dr. Lei Yang (Bayannur Academy of Agricultural Sciences, Bayannur, China). After surface sterilization and germination in Petri dishes, uniformly germinated seeds with coleoptile approximately 1 cm long were transferred to hydroponic boxes containing 1/2×Hoagland solution. The seedlings were grown in a controlled growth chamber (24 °C/17 °C, 16 h light/8 h dark), with the nutrient solution replaced daily. After four days of cultivation, the seedlings were treated with 1/2×Hoagland solution supplemented with 15% (w/v) PEG-6000 for an additional 10 days, and control plants were cultivated with 1/2×Hoagland solution without PEG. PEG-6000 was widely applied to mimic osmotic stress (Chen et al., 2025; Wang et al., 2025; Yan et al., 2025), and a concentration of 15% was used because preliminary experiments revealed markedly different responses of the two varieties to PEG-6000 of this concentration without causing irreversible damage. After 10 days of treatments, plant phenotypic traits were documented via photographing, and then various physiological and biochemical parameters were measured. The second fully expanded leaf from each biological replicate was collected and stored at −80 °C for metabolomic analysis. All experiments were conducted with three biological replicates, each replicate consisting of a pooled sample from 8–12 plants from a box.

Relative water content (RWC) was determined following the method of Li et al. (2023), with minor modifications. Fresh weight (FW), saturated weight (SW, after immersion in distilled water), and dry weight (DW, after oven-drying) were recorded. RWC (%) = (FW - DW)/(SW - DW) × 100%. The rate of water loss (RWL) was measured according to Jiang et al. (2021). Fresh weight (FW₁), weight after air-drying at room temperature in the dark for 5 h (FW₂), and dry weight (DW) were recorded. RWL (g·g^-¹·h^-¹) = (FW₁ - FW₂)/(DW × t). Gas exchange and chlorophyll fluorescence parameters were measured on fully expanded mature leaves using a portable photosynthesis and fluorescence system (LI-6800; LI-COR Biosciences, Lincoln, NE, USA). Measurement conditions were maintained at a light intensity of 1200 μmol m⁻²·s⁻¹, a CO₂ concentration of 400 μmol·mol⁻¹, and a relative air humidity of 55%. Chloroplast pigment contents were determined following the method of Ni et al. (2009). Dry samples were digested with HNO₃ at 120°C for measurements of mineral elements. Potassium (K⁺) content was determined using an atomic absorption spectrophotometer (TAS-990; Beijing Puxi Instrument Co., China). The contents of calcium (Ca), and magnesium (Mg), iron (Fe), manganese (Mn) mineral elements were analyzed using inductively coupled plasma atomic emission spectrometry (ICP-AES). Electrolyte leakage rate was assessed using a conductivity meter (DDS-307A; LeiMagnet, China). Superoxide anion (O₂^-) content was measured according to the method of Wan et al. (2014). The activities of superoxide dismutase (SOD), peroxidase (POD) and catalase (CAT) were determined following the methods of Zhou et al. (2021) and Chen and Zhang (2016). One unit (U) of enzyme activity was defined as a change in absorbance of 0.01 per minute under the assay conditions, and enzyme activities were expressed as U·g^-¹ fresh weight (FW). Malondialdehyde (MDA) content was measured according to Tang (1999) and Fu et al. (2022). Soluble protein content was determined using the Coomassie Brilliant Blue method (Bradford, 1976). The activities of sucrose phosphate synthase (SPS) and sucrose synthase (SuSy) were determined spectrophotometrically (Verma et al., 2011) and expressed as nmol·g^-¹·min^-¹ FW.

Data are presented as means ± standard deviation (SD). Graphical visualizations were performed using Origin 2025 (OriginLab Corporation, Northampton, MA, USA). Physiological traits were analyzed using one-way analysis of variance (ANOVA) under a completely randomized design. Least significant difference (LSD) test (P < 0.05) was used for mean comparisons with SPSS software (IBM SPSS Statistics, Armonk, NY, USA). Principal component analysis (PCA) for metabolic data was performed with Origin 2025. Differentially accumulated metabolites (DAMs) were initially screened using the criteria VIP > 1, |log2 (fold change) | > 1, and P < 0.05 (Student’s t-test). Stress-upregulated metabolites unique to BM14 (SUMs) were further identified based on Benjamini-Hochberg false discovery rate (FDR)-adjusted P < 0.05, VIP > 1, and |log2 (fold change) | > 1.

After 10 days of osmotic stress treatment, no significant differences were observed between the drought-tolerant spring wheat pure line BM14 and the widely cultivated spring wheat variety NC4 under control conditions. However, under osmotic stress, BM14 exhibited better growth performance compared to NC4 (Figure 1A). Osmotic stress significantly lowered plant height and biomass, as well as the relative growth rates in both varieties, with less reduction in BM14 than in NC4 (Figures 1B-E). After the stress treatment, BM14 exhibited a smaller reduction in RWC than NC4 (Figure 1F). Under osmotic stress, RWL was much lower in BM14 than in NC4 (Figure 1G). BM14 exhibited a stronger leaf water retention capacity and a lower water loss rate, which contributed to maintaining a more stable intracellular water status. Osmotic stress displayed much stronger inhibitory effects on photosynthesis in NC4 than in BM14 (Figure 2). Chlorophyll fluorescence parameters of BM14 were unaffected by osmotic stress, and its photosynthetic pigment content showed a slight decline under the stress condition. In contrast, NC4 exhibited a significant reduction in both chlorophyll fluorescence parameters and pigment content. Ion homeostasis was also disrupted under osmotic stress, leaf K⁺ content in NC4 decreased to 31% of the control, while in BM14 it decreased to 89% of the control (Figure 3A). Although the K⁺ content in the roots of both varieties decreased under stress condition, there was no significant difference between the two varieties under osmotic stress. Other mineral elements in BM14 exhibited slight changes after the stress treatment, whereas those mineral elements in NC4 displayed more pronounced upregulation or downregulation (Figures 3B-E). Osmotic stress elevated O₂^- levels in both genotypes, with lower enhancement in BM14 than in NC4 (Figure 4A). Consistently, changes in MDA content (Figure 4B) and electrolyte leakage rate (Figure 4C) indicated that BM14 experienced less oxidative damage. Under osmotic stress, in the leaves, the CAT and SOD activity of BM14 was 0.34 times and 0.18 higher than that of NC4, respectively. In the roots, CAT activity in BM14 was 1.28 times higher than that in NC4 under osmotic stress (Figures 4D–F). Under osmotic stress, soluble protein content significantly decreased in the leaves of NC4, while remaining unchanged in BM14 (Figure 4G). Additionally, SPS and SuSy activity were significantly increased in the roots of BM14 after osmotic stress treatment, whereas no significant changes were detected in NC4 (Figures 4H-I).

A widely targeted metabolomics approach was employed to characterize metabolic responses to osmotic stress in the leaves and roots of BM14 and NC4. A total of 1,716 metabolites were collectively detected in the leaves and roots of BM14 and NC4. These metabolites encompassed a wide range of compound classes, including 221 alkaloids, 220 amino acids and derivatives, 373 flavonoids, 81 lignans and coumarins, 69 nucleotides and derivatives, 59 organic acids, 142 lipids, 193 phenolic acids, 19 quinones, 8 steroids, 9 tannins, 104 terpenoids, 58 carbohydrates, 17 vitamins, and 143 compounds classified as others. Based on metabolite identification confidence criteria in the MWDB database, the detected metabolites were classified into three confidence levels. Among them, 545 metabolites were assigned to level 1, 314 to level 2, and 857 to level 3 (Supplementary Table 1), indicating a high overall reliability of metabolite annotation. The QC plots were presented in Supplementary Figure 1. Principal Component Analysis (PCA) revealed distinct separations between treatment groups, indicating significant metabolite changes in response to the stress (Supplementary Figure 2). To assist group discrimination, OPLS-DA was performed for each tissue and genotype. For a representative comparison, the OPLS-DA model explained a substantial proportion of the variance, with R²X = 0.447 and R²Y = 0.991, and showed acceptable predictive performance, as indicated by a Q² value of 0.888 (Supplementary Figure 3). Differentially accumulated metabolites (DAMs) were identified for each comparison based on the criteria of VIP > 1, |log2 (fold change) | > 1, and P < 0.05. The number and proportion of DAMs in each comparison group were shown in Table 1 and Figure 5, and the detailed information was listed in the Supplementary Table 1. In the comparison between BM14 and NC4 leaves under control conditions, 279 DAMs were identified, whereas osmotic stress induced 645 DAMs between BM14 and NC4 leaves. Analysis of DAM revealed distinct metabolic response patterns between the two genotypes. In both organs of BM14, osmotic stress induced a notable accumulation of carbohydrates and lipids, but it downregulated the levels of many alkaloids and amino acid or amino acid derivatives. In contrast, in both organs of NC4, osmotic stress upregulated the accumulation levels of many alkaloids and amino acid or amino acid derivatives, suggesting distinct metabolic response strategies between the two genotypes.

To elucidate the genotype-specific metabolic regulation mechanisms of the drought-tolerant spring wheat under osmotic stress, we focused on stress-upregulated metabolites unique to BM14 (SUM). According to the statistical test criterion FDR < 0.05, VIP > 1, and |log2 (fold change) | > 1, we identified SUMs if the levels of the metabolites were elevated in BM14 not in NC4 as well as the levels of the metabolites were upregulated in both genotypes but BM14 displayed higher levels than NC4 under the stress condition. Detailed information of SUMs was displayed in Table 2; Supplementary Table 1. Carbohydrates were the primary SUMs, with 14 ones in leaves and 15 ones in roots. Besides carbohydrates, a number of flavonoids and phenolic acids were identified as SUM in leaves, while many of phenolic acids, lignans and coumarins were identified as SUM in roots.

We discovered 14 and 15 carbohydrates in list of leaf and root SUM, respectively (Figures 6, 7). The 14 carbohydrates in leaves mainly included D-arabinose, D-fructose, D-galactose, D-glucose, DL-xylose, D-mannose, Sorbose. Notably, 4 carbohydrates (D-arabinose, D-galactose, DL-xylose, L-xylose) were identified as SUMs in both leaves and roots, and all of them were upregulated in BM14 not in NC4. In the list of leaf SUMs, we discovered 3 organic acids, including citric acid glucoside, D-malic acid, and α-ketoglutaric acid (Supplementary Figure 4). In list of root SUM, we found 6 organic acids, including citric acid, citric acid diglucoside, citric acid glucoside, isocitric acid, malic acid-1-O-diglucoside and tartronate semialdehyde (Supplementary Figure 4). Among them, citric acid glucoside was identified as SUM in both organs. We discovered 6 flavonoids with SUM pattern in leaves, including 3 flavonols, 2 flavones, and 1 flavanone (Supplementary Figure 5). In SUM list of roots, we found 4 flavonoids, levels of which all were significantly upregulated in BM14 under osmotic stress but remained unchanged in NC4. Furthermore, 3 phenolic acids were upregulated in the leaves of BM14 but not in the leaves of NC4 (Supplementary Figure 6), and 13 phenolic acids were upregulated in the roots of BM14 but not in the roots of NC4. A few lipids were identified as SUM (Supplementary Figure 7). Additionally, levels of 3 B vitamins (pyridoxine, 4-pyridoxic acid, D-panthenol) were enhanced in the leaves of BM14 but not in the leaves of NC4 (Supplementary Figure 8). Only one B vitamin was identified as SUM in roots. To obtain an integrated view of metabolic adjustments associated with osmotic stress tolerance in the drought-tolerant genotype BM14, important SUMs (carbohydrates, organic acids and phenylpropanoids) were mapped in metabolic pathway network according to literature and metabolic pathway databases (Figure 8).

Global spring wheat production is increasingly challenged by drought stress (Ceccarelli et al., 2010; Dong et al., 2018; Fan et al., 2018; Khan, 2019), highlighting the importance of elucidating physiological and metabolic mechanisms underlying drought tolerance. In the present study, we found that under osmotic stress, the drought-tolerant pure line BM14 maintained better growth, higher leaf water content, lower rate of water loss, and higher photosynthetic activity. Osmoregulation and energy metabolism play critical roles in plant drought tolerance (Gupta et al., 2020; Haghpanah et al., 2024; Xu and Fu, 2022). Under drought or high osmotic stress, plants typically accumulate inorganic ions and compatible solutes to lower intracellular water potential, thereby maintaining cellular water balance and structural stability. Previous studies have demonstrated that energy cost of accumulating inorganic ions is much lower than de novo synthesis of organic osmolytes to lower intracellular water potential (Hou et al., 2021; Shabala and Shabala, 2011; Zeng et al., 2015). In response to osmotic stress, plants primarily reduce cellular water potential through the uptake and accumulation of inorganic ions such as K⁺, Na⁺, NO₃^-, and Cl^- (Hou et al., 2021). In our study, the K⁺ content in the leaves of both genotypes significantly decreased under osmotic stress; however, the reduction was markedly smaller in BM14 than in NC4, with its K⁺ content being 2.67-fold higher than that in NC4. We also observed that many potassium channel genes and potassium transporter genes were upregulated in roots of BM14 but not in those of NC4 (unpublished manuscript), indicating that the elevated K⁺ accumulation is specific metabolic response strategy of BM14 to osmotic stress. Enhanced K⁺ accumulation has been widely associated with drought tolerance across diverse plant species (Feng et al., 2020; Wang et al., 2023; Sun et al., 2024; Li et al., 2024), revealing the role of K⁺ metabolism in drought tolerance is conserved among plants. In addition to accumulated higher concentration of K⁺ under osmotic stress, BM14 also exhibited genotype-specific accumulation for multiple types of metabolites. We focused on SUMs, and a total of 42 SUMs were identified in the leaves, including 14 carbohydrates, 6 flavonoids, and 3 phenolic acids. In the roots, 77 SUMs were identified, including 15 carbohydrates, 13 phenolic acids. These findings suggest that the enhanced accumulation of carbohydrates and flavonoids in leaves and the enhanced accumulation of carbohydrates and phenolic acids in roots, is a distinctive metabolic feature of BM14, which contributes to strong tolerance of this wheat genotype to osmotic stress. Free carbohydrates are well-recognized as primary organic osmolytes under drought stress and simultaneously function as key substrates for energy metabolism (Ozturk et al., 2021; Amoah and Adu-Gyamfi, 2025; Boriboonkaset et al., 2013). In our study, BM14 showed marked accumulation of multiple hub carbohydrates under osmotic stress, including fructose, galactose, glucose, sorbose and inositol in leaves, as well as trehalose, sucrose, panose and maltose in roots. Notably, under osmotic stress, sucrose accumulation was unaffected in NC4, but it was markedly upregulated in the roots of BM14. Sucrose has been widely recognized as a major organic osmolyte in response to salinity, osmotic, and drought stresses (Amoah and Adu-Gyamfi, 2025; Mathan et al., 2021; Suprasanna et al., 2016). To further explore the regulation mechanism by which osmotic stress induces sucrose synthesis in BM14, we assessed the activities of the key enzymes involved in sucrose synthesis. Consistent with enhanced sucrose accumulation, the activities of key sucrose synthesis enzymes SPS and SuSy were significantly elevated in the roots of BM14 under the stress condition, whereas their activities were unaffected in NC4. This coordinated upregulation of SPS and SuSy activities likely facilitates rapid sucrose accumulation, thereby supporting both osmotic adjustment and energy supply. Additionally, several important glycolysis-related carbohydrates (glucose, fructose, galactose, and sucrose) and intermediates of TCA cycle (malate, α-ketoglutarate, citrate, and isocitrate) were significantly upregulated in BM14 but remained unchanged in NC4. Similar enhancements in central carbon metabolism have been reported in other drought-tolerant crop genotypes, suggesting that strengthened energy metabolism represents a conserved adaptive trait that sustains stress responses under water deficit conditions (Li et al., 2024; Yin et al., 2024). These findings highlight energy metabolism-related pathways as promising targets for genetic improvement of drought resistance.

Osmotic stress induces excessive accumulation of reactive oxygen species (ROS) in plants, leading to membrane lipid peroxidation, protein denaturation, and DNA damage, ultimately disrupting cellular structure and function (Bao et al., 2024; Hasanuzzaman et al., 2021). To mitigate ROS-induced damage, plants typically enhance antioxidant enzyme activities and accumulate non-enzymatic ROS scavengers (Hasanuzzaman et al., 2021; Irato and Santovito, 2021; Rao et al., 2025). In this study, we observed that under osmotic stress, the levels of O₂^- in both leaves and roots of BM14 were significantly lower than those in NC4. Conversely, the activities of antioxidant enzymes, including SOD and CAT, were significantly higher in BM14 than in NC4 under osmotic stress. These findings indicated that enhanced antioxidant enzyme activity may contribute to the superior ROS-scavenging capacity of BM14, consistent with previous studies demonstrating the close association between antioxidant enzyme activity and drought tolerance in wheat (Abid et al., 2018; Ge et al., 2024; Hasanuzzaman et al., 2021). Regarding non-enzymatic antioxidants, BM14 exhibited pronounced organ-specific accumulation patterns in response to osmotic stress. These non-enzymatic antioxidants have been revealed to play critical roles in drought tolerance of other crops. For example, Wang et al. (2024b) reported that drought-tolerant soybean (Glycine max (L.) Merr.) varieties significantly increased the accumulation of flavonoids and phenolic acids to maintain ROS homeostasis under drought stress conditions. The inhibitive effect of abiotic stress on plant growth and metabolism has been associated with the degradation of B vitamins (Hanson et al., 2016). Although vitamins and plant hormones were discovered around the same time, vitamins have received considerably less attention than plant hormones in studies of abiotic stress (Hanson et al., 2016; Kohler, 1975). However, increasing evidence has revealed their vital antioxidant roles in plant stress responses (Asensi-Fabado and Munne-Bosch, 2010; Cao et al., 2024; Fitzpatrick, 2024). For instance, application of thiamine and pyridoxine has been shown to enhance antioxidant capacity and salt stress tolerance in faba bean (Ahmed and Sattar, 2024), and application of vitamin B2 (riboflavin) can improve drought tolerance in Agaricus bisporus (Guhr et al., 2017). Additionally, it had been reported that abiotic stress can upregulate B vitamin levels in tobacco plants (Huang et al., 2013), Arabidopsis thaliana (Rapala-Kozik et al., 2012), and maize (Rapala-Kozik et al., 2008). Based on these observations, we propose that B vitamins, together with phenolic acids and flavonoids, may mediate ROS scavenging and contribute to strong osmotic stress tolerance of BM14. Overall, our findings suggest that drought-tolerant spring wheat genotype employs organ-specific ROS scavenging strategies (Figure 9), relying on coordinated regulation of enzymatic and non-enzymatic antioxidants to maintain ROS homeostasis under osmotic stress.

Previous studies have reported that linoleic acid, α-tocotrienol, and L-leucine are closely associated with tolerance to PEG-induced osmotic stress in winter wheat (Li et al., 2022), while allantoin, galactaric acid, gluconic acid, and glucose have been linked to drought tolerance in rice (Ghorbanzadeh et al., 2023). In line with these findings, our results demonstrate that the drought-tolerant spring wheat genotype responds to osmotic stress through enhanced accumulation of key glycolysis-related carbohydrates, tricarboxylic acid (TCA) cycle intermediates, B vitamins, phenolic acids, and flavonoids, indicating a distinct metabolic response strategy in spring wheat. Although PEG-6000 treatment does not fully reproduce the complexity of soil drought, particularly with respect to root–soil and root–microorganism interactions, it effectively induces osmotic stress, a major component of soil drought. Consequently, PEG-based treatments may bias the identification of differentially accumulated metabolites (DAMs), especially in roots, compared with soil-based drought conditions. Nevertheless, both our results and soil drought–based studies consistently highlighted the importance of K⁺ homeostasis, carbohydrate metabolism, and phenolic acids and flavonoids in osmotic stress or drought tolerance, supporting robust reliability of the metabolic traits based on PEG treatment in the present study (Chen et al., 2025; Wang et al., 2025; Yan et al., 2025; Cui et al., 2019; Xu et al., 2021; de Souza Mateus et al., 2022).