Seeds of twelve T. aestivum cultivars, classified as drought-tolerant or drought-sensitive, were obtained from the General Directorate of Agricultural Research and Policies (TAGEM), Turkey ( Supplementary Table S1 ). The seeds were surface sterilized (5 min with 70% EtOH and 5 min with 5% hypochlorite) and pre-germinated in Petri dishes for 10 days on moist filter paper at 4°C in the dark. Uniform seedlings were transplanted into 1.5 L plastic pots filled with a turf: soil: sand mixture (3:3:1) and grown in a controlled environment chamber at 18–20°C, 60–70% relative humidity. For each cultivar, three pots were assigned to the control and three to the drought-stress treatment.

Lipid peroxidation was quantified by measuring thiobarbituric acid reactive substances (TBARS), a widely used indicator of oxidative membrane damage under stress conditions. TBARS levels were determined following the method of Madhava Rao and Sresty (2000), using an extinction coefficient of 155 mM^-¹ cm^-¹ for calculation.

To further understand the antioxidant defense strategies, the activities of key ROS-scavenging enzymes were assessed. All enzyme extractions were carried out at 4°C. Fresh tissue samples (0.5 g) were ground in liquid nitrogen and homogenized in 1.5 ml of 50 mM Tris-HCl buffer (pH 7.8) containing 0.1 mM EDTA, 0.2% (w/v) Triton X-100, 1 mM phenylmethylsulfonyl fluoride (PMSF), and 2% (w/v) polyvinylpyrrolidone (PVP). For ascorbate peroxidase (APX) extraction, 5 mM ascorbate was added to the homogenization buffer. The homogenates were centrifuged at 14,000 × g for 30 min, and the resulting supernatants were used for protein quantification and enzyme assays. Total soluble protein content was determined using the Bradford method with bovine serum albumin as a standard. All spectrophotometric measurements were performed using a Shimadzu UV-1600 spectrophotometer.

Total RNA isolation was performed from leaf and root tissues of 4- and 8-hours droughts stressed and control groups using the RNeasy Plant Mini kit (Qiagen, Hilden, Germany) according to the manufacturers’ instructions. Approximately 100 mg of tissue was ground into a fine powder using a pre-chilled mortar and pestle with liquid nitrogen. The powdered tissue was transferred to lysis buffer containing guanidine thiocyanate to inactivate RNases, followed by homogenization through the kit-provided shredder column to remove cellular debris. RNase-free DNaseI (Roche Applied Science GmbH, Germany) digestion and purification were carried out for the elimination of the genomic DNA from total RNA as described Cevher-Keskin et al. (2019). RNA concentration and purity were initially determined using a NanoDrop ND-1000 spectrophotometer (Thermo Fisher Scientific, USA), ensuring A260/A280 ratios between 1.9 and 2.1 and A260/A230 ratios above 2.0. RNA integrity was further assessed with an Agilent 2100 Bioanalyzer (Agilent Technologies, USA), and only samples with an RNA Integrity Number (RIN) ≥ 8.0 were selected for downstream RNA-seq analysis. All samples were stored at –80 °C until library preparation.

Leaf and root tissues from three biological replicates of each cultivar were analyzed under four conditions (4-hours drought, 4-hours control; 8-hours drought, and 8-hours control), resulting in a total of 72 samples (3 genotypes × 4 conditions × 3 replicates × 2 tissues). The RNAseq library for each sample was prepared with a 1250 ng of total RNA using the TruSeq RNA Sample Preparation kit (Illumina) according to the manufacturer’s instructions. Paired-end sequencing was performed with a current next generation sequencing instrument, HiSeq2000 (Illumina, user guide; Part# 15011190 Rev. H) using TruSeq SBS Kit v3 (cBot-HS) (Illumina, user guide; Part#15023333 Rev. B). The prepared libraries were enriched using 15 cycles of PCR and purified by the QIAquick PCR purification kit (Qiagen). The Agilent 2100 Bioanalyzer was used to control the size and purity of the samples using the Agilent High Sensitivity DNA Kit. A total of 12 indexes were prepared for 72 samples and run-on Illumina HiSeq 2000 for 6 lanes (Pingault et al., 2015). The enriched libraries were diluted with the elution buffer to a final concentration of 10 nM. Sequencing was performed on each library to generate 100-bp PE reads for transcriptome sequencing on an Illumina High-Seq 2000 platform.

The quality control was performed for the Illumina paired-end sequencing files of each sample. FastQC Software” was used for the detection of faulty sequences (Andrews, 2022). RNA-seq data were trimmed using the Fastx Toolkit (http://hannonlab.cshl.edu/fastx_toolkit, Hannon Lab, 2022). After quality control, de novo assembly was carried out from a total of 311 GB of transcript data. The assembly was performed as recommended by Duan et al. (2012). The resultant data were evaluated using the software “Trinity Assembly”, which combines three independent software modules (Inchworm, Chrysalis and Butterfly) and 323 Mbs of FASTA files were obtained. To remove the expected redundancy in this assembly file, “the cd-hit-est tool” to place the contigs into clusters was applied, so that a sequence is not represented more than once in our reference assembly. Subsequently, the RNA-seq data were mapped to our de novo reference genome using Bowtie (https://bowtie-bio.sourceforge.net/index.shtml, 2022; Langmead et al., 2022). The resulting mapped reads were evaluated by using the RSEM tool to obtain Fragments per Kilobase of transcript per Million mapped reads (FPKM) data. FPKM files belonging to each sample were subjected to pairwise comparison using the edgeR differential expression tool, which is included in the R-Bioconductor package (Robinson et al., 2010). Through differential expression analysis, we pooled replicates belonging to each condition into a single file by averaging the counting information corresponding to each gene. As a result, comparisons between different conditions were carried out and differentially expressed transcripts were obtained. However, some transcripts were not informative, as they were not annotated due to a lack of well-annotated reference genome. In this case, the Trinotate annotation tool (https://rnabio.org/module-07-trinotate/0007/02/01/Trinotate) was used which uses various well referenced methods for functional annotation including homology search for known sequence data (NCBI-BLAST), protein domain identification (HMMER/PFAM), protein signal prediction (singalP/tmHMM), and comparison to currently curated annotation databases (EMBL Uniprot eggNOG/GO Pathways databases) have been applied. To account for multiple hypothesis testing, we first calculated adjusted p-values using the Benjamini-Hochberg procedure, which controls the expected proportion of false positives (Benjamini and Hochberg, 1995; Benjamini and Yekutieli, 2001). These adjusted p-values were recommended as the primary statistic for interpreting significance. Additionally, multiple testing corrections were performed using the False Discovery Rate (FDR) method to further control type I error, applying a stringent threshold of FDR ≤ 0.05 for the identification of DEGs and enriched functional terms. In addition to this statistical threshold, we assessed the biological relevance of the expression change using the log₂ Fold Change (log₂FC), which quantified both the direction (up- or down-regulation) and the magnitude of the transcriptional difference. A log₂ fold change threshold was applied to filter genes, ensuring that only those with biologically meaningful shifts were retained for downstream analysis. Genes exceeding the defined cut-off, either positively or negatively, were considered significant and included in the final dataset. The colour intensity, based on the adjusted logarithmic scale of FC values, demonstrates the level of significance of each term. If there was no log2FC score for the corresponding enriched term, this was depicted as white in the heatmap. Pathway enrichment analysis is performed by using WebGestaltR. The terms belong to KEGG and WikiPathways. Adjusted values indicate adjusted p-values. Negative logarithm base 10 was applied to the p-values of pathway terminologies.

Primers were designed for the selected genes using FastPCR and Primer 3 programs. The quality of the primers was validated by BLASTN queries against the entire wheat EST unigene set. The primers, wherever possible, were designed spanning an intron or intron-intron junctions to detect any genomic DNA contamination. All the primers were adjusted to 100–140 bp amplicon size and 55°C annealing temperature and controlled by conventional PCR by housekeeping genes (β actin, EF-1 and EF2 primers).

First-strand cDNA was synthesized from 1 μg of total RNA in a 20 μl reaction using MMLV reverse transcriptase (Roche High Fidelity cDNA Synthesis Kit) following the manufacturer’s instructions. cDNA quality was verified by conventional PCR using housekeeping gene primers (β-actin, EF1, EF2). Differential expression analysis was performed with SYBR Green Mix (Roche FastStart Universal SYBR Green Master) and gene-specific primers ( Supplementary Table S5 ) on an iQ5 System (Bio-Rad, Hercules, USA) as described by Cevher-Keskin et al. (2011). Each reaction was run in triplicate to ensure accuracy. Relative transcript abundance was normalized to housekeeping genes (EF-α1 and EF-α2), and fold changes were calculated using the comparative CT (ΔΔCq) method (Schmittgen et al., 2000). Error bars represent standard deviation across three technical replicates.

Data from three independent biological replicates were analyzed using one-way analysis of variance (ANOVA), or Student’s t-test, as appropriate, to evaluate the effects of treatments on the measured parameters. When significant differences were observed, Tukey’s Honest Significant Difference (HSD) post hoc test was applied for pairwise comparisons. Statistical significance was set at p < 0.05 (*, significant) and p < 0.01 (**, highly significant); results with p ≥ 0.05 were considered not significant (ns). All statistical analyses were performed using GraphPad Prism (v. 10.5.0 for Mac; GraphPad Software). Bar graphs for enzyme analysis were generated using Microsoft Excel (Microsoft Corporation, Redmond, WA, USA).

In the present study, 12 bread wheat (Triticum aestivum L.) cultivars with diverse genetic backgrounds were initially evaluated to determine the most promising drought stress tolerant and sensitive cultivars under progressive drought stress ( Supplementary Figure S2 ).

Initially, we conducted a physiological screen. Soil water content (SWC) was monitored throughout the drought stress period. In control plants, SWC ranged from 35% to 45%, depending on the cultivar. Under drought conditions, SWC generally declined to 8–22%. Some cultivars, such as Sultan, Kırgız, and Gerek, showed a larger SWC difference between control and drought treatments, suggesting faster water depletion or higher water loss. In contrast, Müfitbey and Harmankaya-99 exhibited smaller differences, indicating slower water uptake or more efficient water conservation. Relative water content (RWC) also decreased during the 10-day drought treatment, with Atay 85 showing the most pronounced reduction, dropping below 70% ( Supplementary Figure S2 ). Conversely, Gerek 79, Müfitbey, Altay, and Harmankaya-99 maintained higher RWC levels under stress, reflecting better water retention and drought tolerance.

To contextualize these findings, we reviewed agronomic and historical performance data from the Ministry of Agriculture (Geçitkuşağı Agricultural Research Institute, Eskişehir, Turkey). Müfitbey is a drought-tolerant winter cultivar, maintaining photosynthetic function, canopy cooling, and stable PSII efficiency (Fv/Fm) under stress (Dirik and Sakin, 2024; Tarım Orman, 2025a). Gerek 79, widely used as a tolerant check, shows broad adaptation and resilience to combined drought and heat stresses (Tarım Orman, 2025b). In contrast, Atay 85 is drought-sensitive, with sharp declines in RWC, SPAD, and Fv/Fm under water deficit (NBC Agriculture). Atay 85 performs best under irrigation, whereas Gerek 79 thrives in drought-prone, cold regions, and Müfitbey combines winter hardiness with drought tolerance. Based on these contrasting traits, we selected Gerek 79 (tolerant), Müfitbey (highly tolerant), and Atay 85 (sensitive) for further analysis.

To capture robust transcriptional responses, we applied a shock dehydration treatment, as acute stress elicits stronger gene expression changes than progressive drought. Seedlings at a similar developmental stage were grown in aerated ½ Hoagland’s solution (renewed every three days) under controlled conditions (16 h photoperiod, 22/18 °C, 60% RH) ( Figure 1A ). They were then removed and subjected to 4 and 8 hours of shock dehydration ( Figure 1B, C ). Leaf and root tissues were immediately frozen in liquid nitrogen, stored at −80 °C, and used for RNA-seq.

Lipid peroxidation (TBARS assay) used as an indicator of oxidative membrane damage under drought stress. To elucidate antioxidant defense strategies, we analyzed the activities of key ROS-scavenging enzymes: superoxide dismutase (SOD), catalase (CAT), peroxidase (POX), and ascorbate peroxidase (APX). SOD catalyzes the dismutation of superoxide radicals, protecting photosynthetic pigments and stabilizing PSII under stress (Rao et al., 2025). CAT complements this by decomposing hydrogen peroxide into water and oxygen, maintaining redox balance (Calderón et al., 2018). Together with POX and APX, these enzymes form a coordinated antioxidant network mitigating drought-induced oxidative stress.

TBARS levels varied significantly among genotypes and time points ( Figure 2 ). Atay 85 showed a sharp increase at 8 hours (AD8), indicating severe oxidative damage. Gerek 79 exhibited the highest TBARS accumulation at GD8, while Müfitbey maintained consistently low levels, suggesting superior oxidative stress tolerance. SOD activity displayed tissue- and genotype-specific patterns ( Figure 3A ). Atay 85 roots showed strong induction at 4 hours and remained high at 8 hours, with leaves following a similar trend. Gerek 79 roots increased at 4 hours, but leaves showed marked suppression under drought. Müfitbey maintained stable SOD activity in both tissues, with only minor fluctuations. CAT responses were highly genotype-dependent ( Figure 3B ). Atay 85 roots exhibited the strongest induction, peaking at AD8 (>9 units mg^-¹ protein), whereas Gerek 79 roots declined steadily under stress. Müfitbey showed moderate root responses. In leaves, Atay 85 activity decreased under stress, while Gerek 79 and Müfitbey increased, with Müfitbey reaching the highest levels at 8 hours. POX activity highlighted contrasting strategies ( Figure 3C ). Atay 85 roots showed delayed induction, while Gerek 79 roots peaked at 4 h before declining. In leaves, Müfitbey exhibited a rapid and strong POX increase (~135% at 4 h), whereas Atay 85 decreased, suggesting reliance on alternative antioxidant systems. APX responses were also genotype-specific ( Figure 3D ). Atay 85 strongly upregulated APX in both roots and leaves, particularly at 8 h (+84.8%), indicating a robust ascorbate–glutathione cycle. Conversely, Gerek 79 and Müfitbey roots showed significant reductions, and Müfitbey leaves declined at 8 hours, suggesting limited APX contribution under prolonged stress.

RNA-seq was performed on root and leaf tissues of the selected cultivars, Atay 85, Gerek 79, and Müfitbey, subjected to 4- or 8-hour shock dehydration or maintained under control conditions to assess transcript-level differences ( Figure 4A, B ). To further explore cultivar- and tissue-specific responses, Gene Ontology (GO) enrichment analysis was conducted on differentially expressed genes (DEGs). Significant DEGs were categorized into biological processes, molecular functions, and cellular components ( Figures 5 – 8 ; Supplementary Figures S3 - S8 ), enabling comparisons of conserved and unique drought-response mechanisms among cultivars.

The distribution of gene expression levels was more variable in leaf tissues than in roots. In Atay 85 roots under 8-hour drought stress (p < 0.01), upregulated genes were enriched in categories related to ion transport, including GO:0015078~hydrogen ion transmembrane transporter activity, GO:0015077~monovalent inorganic cation transmembrane transporter activity, and GO:0022890~inorganic cation transmembrane transporter activity. In contrast, the tolerant cultivar Müfitbey showed enrichment in biosynthetic and energy-related processes, such as GO:0034404~nucleobase, nucleoside, and nucleotide biosynthetic process, GO:0034654~nucleic acid biosynthetic process, GO:0016469~proton-transporting two-sector ATPase complex, GO:0045259~proton-transporting ATP synthase complex, and GO:0044271~nitrogen compound biosynthetic process. In Gerek 79, leaves under 8-hour drought stress, upregulated genes were associated with GO:0005506~iron ion binding, GO:0046906~tetrapyrrole binding, and GO:0009767~photosynthetic electron transport chain ( Figures 4 , 6 - 8 ).

At 4 hours of drought stress ( Figure 5A ), 37 genes were commonly regulated across all three cultivars. Müfitbey had the highest number of unique genes (8), followed by Atay 85 (7) and Gerek 79 (3). Pairwise overlaps included 5 shared genes between Atay 85 and Gerek 79, 3 between Atay 85 and Müfitbey, and 3 between Gerek 79 and Müfitbey. At 8 hours of drought stress ( Figure 5B ), the number of shared genes across all cultivars decreased to 15, indicating more divergent responses over time. Atay 85 had 9 unique genes, Gerek 79 had 10, and Müfitbey had 2. Pairwise overlaps included 3 genes between Atay 85 and Gerek 79, 4 between Gerek 79 and Müfitbey, while Atay 85 and Müfitbey shared none. Unique DEGs for sensitive and tolerant cultivars are listed in Supplementary Tables S1 and S2 .

Pathway enrichment analysis of significantly differentially expressed genes revealed tissue-specific drought responses ( Figure 9 ).

In roots, enrichment of secondary metabolite biosynthesis and phenylpropanoid pathways suggests enhanced production of protective compounds (e.g., flavonoids, lignin) that reinforce root structure and improve drought tolerance. Glutathione metabolism plays a central role in detoxifying drought-induced reactive oxygen species (ROS), while nucleotide and purine metabolism supports DNA/RNA synthesis and energy balance, ensuring root growth and repair under stress. Interestingly, pathways linked to seed development were also upregulated, possibly reflecting a survival mechanism to secure future reproduction. In leaves, enrichment of porphyrin metabolism indicates an effort to sustain chlorophyll synthesis and maintain photosynthetic activity despite water deficit. Activation of ubiquitin-mediated proteolysis facilitates protein turnover by removing damaged proteins and recycling amino acids—an essential adaptive mechanism during stress. Similar to roots, glutathione and phenylpropanoid pathways in leaves contribute to ROS scavenging and structural reinforcement, protecting cellular integrity. Together, roots and leaves exhibit activation of secondary metabolite biosynthesis, highlighting a broad stress-adaptive strategy. However, roots prioritize metabolic adjustments for growth maintenance, oxidative stress reduction, and developmental adaptation, whereas leaves focus on preserving photosynthesis, regulating protein turnover, and maintaining cellular homeostasis under drought.

To validate the RNA-seq results, twelve drought-responsive genes identified from the DEG analysis were selected for expression profiling using qRT-PCR. Although fold-change values differed between RNA-seq and qRT-PCR, the overall expression trends were consistent, confirming the reliability of the RNA-seq data.

The selected genes represent key stress-related pathways, including cell wall remodeling, oxidative stress response, hormone signalling, and metabolic regulation. These include:

TaPME42 (pectinesterase/pectinesterase inhibitor 42), TaExLP (extensin-like protein), TaGLP9-1 (germin-like protein 9-1), TaZFP36 (CCCH-type zinc finger protein 36), TaMC5 (metacaspase-5), TaPGM (phosphoglycerate/bisphosphoglycerate mutase), TaPP2CA (protein phosphatase 2C), TaGI (GIGANTEA), TaRBP45B (RNA-binding protein), TaFER (ferritin), TaADT (arogenate dehydratase 5), and TaFBW2 (F-box protein). Expression was assessed in root and leaf tissues of drought-stressed and control plants. qRT-PCR validation confirmed differential expression of genes under drought stress ( Figures 10 – 12 ).

TaZFP36 (zinc finger protein) was strongly induced in roots and leaves of the tolerant cultivar Müfitbey, but not in Atay 85. TaFER (ferritin) increased in leaves of both cultivars, peaking at 8 hours ( Supplementary Figure S9 ). TaPME42 (pectinesterase) was consistently upregulated in roots and leaves of both cultivars. TaExLP (extensin-like protein) showed strong induction in roots of both cultivars but cultivar-specific patterns in leaves, higher in Atay 85 at 4 hours, reduced in Müfitbey.

TaGLP9-1 (germin-like protein) was upregulated in roots of both cultivars, but only Müfitbey showed increased expression in leaves. In defense responses, TaMC5 (metacaspase-5) was induced in roots of Müfitbey but unchanged in Atay 85. TaADT-5 (arogenate dehydratase) increased in leaves of both cultivars at 4 hours, with an eight-fold rise in Müfitbey at 8 hours, while Atay 85 declined ( Supplementary Figure S10 ). TaPGM (Phosphoglycerate mutase) increased in roots of Atay 85 at both time points and in leaves of Müfitbey at 4 hours. TaPP2CA was significantly upregulated in roots and leaves of both cultivars, peaking at 4 hours in most tissues. Finally, TaGI (GIGANTEA) decreased in Müfitbey leaves at 8 hours but increased in Atay 85 at 4 hours. TaRBP45B (RNA-binding protein) was induced in roots of both cultivars and transiently in leaves at 4 hours. TaFBW2 (F-box protein) showed cultivar-specific induction, suggesting a role in ubiquitin-mediated proteolysis and stress signalling.

Our study revealed drought-induced upregulation of metal ion-binding genes in both leaf and root tissues of T. aestivum, including those involved in heme, 2Fe-2S cluster, and zinc, iron, and copper binding. Proteins such as AtTZF1–3 are known to regulate plant growth and stress responses (Wang et al., 2008).

In Arabidopsis, 11 CCCH-type TZFs with plant-specific motifs were identified (Wang et al., 2008; Pomeranz et al., 2010). AtTZF1–6 and AtTZF9 are involved in ABA signaling, seed germination, and PAMP-triggered immunity, and localize to stress granules and processing bodies, influencing post-transcriptional and epigenetic regulation (Anderson and Kedersha, 2009). Gain-of-function AtTZF1 lines show enhanced drought and cold tolerance via ABA/GA modulation (Lin et al., 2011).

In wheat, 269 TaZFPs exhibit stress-responsive cis-elements and tissue-specific expression, suggesting roles in growth and abiotic stress adaptation (Wu et al., 2022). Notably, TaZFP36 was significantly upregulated in both tissues of drought-tolerant cultivars, but not in the sensitive cultivar Atay 85, indicating its potential role in drought tolerance, consistent with AtTZF1 studies.

We also observed regulation of TaFER, a ferritin gene involved in iron storage and oxidative stress response (Wu et al., 2022). Ferritin accumulation in chloroplasts is triggered by various stressors including ozone, ethylene, and iron overload (Fobis-Loisy et al., 1995; Van Wuytswinkel and Briat, 1995; Murgia et al., 2002). Our qRT-PCR results showed differential TaFER expression in leaf tissues of tolerant and sensitive cultivars, highlighting its role in iron regulation and drought adaptation.

Differential gene expression analysis under drought stress revealed several cell wall-related genes, including Beta-galactosidase 1, Glucose-6-phosphate/phosphate-translocator, TaExLP4, TaExLP6, TaGLP9-1, and lignin biosynthesis genes. Three were selected for further analysis due to elevated expression: TaPME49, TaExLP, and TaGLP9-1 ( Supplementary Table S3 ).

TaPME49, involved in pectin demethylesterification, affecting cell wall plasticity (Kohli et al., 2015) showed increased in drought-stressed leaf tissues of both tolerant and sensitive cultivars, but decreased in roots, indicating tissue-specific regulation (Al-Qsous et al., 2004; Micheli, 2001). TaExLP, a hydroxyproline-rich glycoprotein (HRGP) involved in development and stress responses (Kurepa and Smalle, 2023) was strongly induced in roots at 4 hours, with higher early expression in sensitive cultivars. Its suppression in tolerant ones suggests a role in early stress signaling rather than long-term adaptation. TaGLP9-1, a Germin-like proteins contribute to ROS-mediated defence and are associated with abiotic stress responses (Christensen et al., 2004; Cevher-Keskin et al., 2019), was upregulated in drought-tolerant leaf tissues, supporting its role in resistance, consistent with proteomic data (Faghani et al., 2015).

ABA regulates drought responses and interacts with salicylic acid (SA) pathways to modulate defence against pathogens (Gupta et al., 2020; Cao et al., 2011) with defence-related genes, including TaADT5 and TaMC5, were upregulated in drought-stressed leaf tissues. TaADT5, a key enzyme in lignin biosynthesis and possibly anthocyanin production (Corea et al., 2012a; Corea et al., 2012b; Muhammad et al., 2023) showed increased expression in tolerant cultivars suggests a role in drought resilience. Related genes ADT1 and ADT3 also contribute to anthocyanin synthesis under stress (Chen et al., 2016). TaMC5, a metacaspase involved in programmed cell death (PCD) and defence (Uren et al., 2000; Valandro et al., 2020) was elevated in 8-hour drought-stressed root and leaf tissues of Müfitbey, indicating its role in drought tolerance via PCD regulation.

Phosphoglycerate/Bisphosphoglycerate Mutase (PGM) is a key enzyme in the glycolysis pathway, catalyzing the transfer of phosphate groups among the three carbon atoms of phosphoglycerate. PGM also dephosphorylates and activates Actin-Depolymerizing Factor 1 (ADF1), a protein that governs the re-modelling of the actin cytoskeleton which is essential for maintaining cell structure and intracellular transport under stress conditions (Oslund et al., 2017).

In our study, TaPGM expression significantly increased in the roots of the drought-sensitive cultivar Atay 85 after 8 hours of stress ( Figure 9E, F ). In contrast, the tolerant cultivar showed early induction in leaf tissue at 4 hours, with no notable change at 8 hours ( Figure 7E, F ). This suggests a dual role for TaPGM in drought response: late root-specific induction in sensitive plants may reflect a delayed compensatory mechanism, while early leaf expression in tolerant plants indicates a proactive adaptation strategy. Overall, TaPGM expression timing and tissue specificity appear to be key factors in drought tolerance in T. aestivum.

Protein Phosphatase 2A (PP2A), a serine/threonine phosphatase, plays diverse roles in biotic and abiotic stress responses (Pais et al., 2009). It negatively regulates ABA signaling and influences ABA-dependent gene expression and light-mediated nitrate reductase activation (Chen et al., 2015; Creighton et al., 2017).

In rice (Oryza sativa), all five catalytic subunit genes (OsPP2A-1–5) are upregulated under salinity stress (Yu et al., 2003). Similarly, salt stress elevates StPP2Ac1–3 transcripts in potato leaves, and okadaic acid inhibition confirms PP2A’s positive role in stress regulation (Pais et al., 2009). In wheat, TaPP2Ac-1 accumulates under water deficit, and its overexpression in tobacco enhances drought tolerance (Xu et al., 2007). In Arabidopsis, PP2A-C5 overexpression activates chloride channels (AtCLCa, AtCLCc), improving ion sequestration and tolerance to salt and drought (Hu et al., 2017; Balasubramanian et al., 2022).

In our study, TaPP2CA expression was significantly upregulated in both leaf and root tissues of tolerant and sensitive cultivars after 4 and 8 hours of drought stress ( Figure 12A, B ). This early and sustained induction suggests a key role in drought perception and ABA-mediated signalling.

Prolonged drought (8 hours) led to downregulation of genes involved in photosynthesis, light harvesting, photosystem I stabilization, and photorespiration in both tolerant and sensitive cultivars, indicating reduced photosynthetic activity due to stress-induced damage and energy conservation.

One key regulator is GIGANTEA (GI), a multifunctional protein involved in circadian rhythm, photoperiodism, phytochrome B signaling, and flowering (Krahmer et al., 2018). GI is modulated by environmental cues such as cold, hydrogen peroxide, blue light, and karrikin (Waters et al., 2014; Krahmer et al., 2018), and stabilizes ADO3 and ADO1/ZTL, regulating CONSTANS (CO) in the long-day flowering pathway. GI also enhances salinity tolerance via SOS2 interaction and induces EARLY FLOWERING (ELF) under drought (Kim et al., 2013; Riboni et al., 2013). Mutations in GI improve oxidative and freezing stress tolerance through CDF upregulation (Fornara et al., 2015). GI also promotes the “Drought Escape” (DE) response, accelerating flowering under drought (Bader et al., 2023).

In our study, TaGI expression was reduced in 8-hour drought-stressed leaf tissues of the tolerant cultivar Müfitbey ( Figure 12C, D ), supporting its negative role in drought tolerance, consistent with GI knockout studies.

RNA-Binding Protein RBP45B, part of the hnRNP family, binds poly(A) tails and is involved in mRNA maturation and translation initiation. RBPs are widely upregulated under abiotic stress (salt, drought, heat, cold, ozone, hypoxia, flooding), highlighting their role in stress tolerance (Yan et al., 2022). TaRBP45B was induced in drought-stressed root tissues at both time points, with cultivar-specific expression in leaves ( Figure 12E, F ), suggesting a positive role in drought adaptation.

The three wheat cultivars examined in this study exhibit distinct physiological and molecular responses to drought stress. Atay 85, the drought-sensitive cultivar, shows pronounced metabolic suppression and a decline in ATP production. This is accompanied by poor coordination between root and leaf responses, resulting in oxidative damage and premature senescence. In contrast, Gerek 79, which displays moderate tolerance, manages to conserve energy while activating osmoprotective mechanisms and reinforcing structural components. Müfitbey, the highly tolerant cultivar, maintains metabolic stability and initiates robust hormonal and antioxidant responses. It appears to adopt a dormancy-like strategy, allowing it to endure prolonged drought conditions more effectively.

Across all cultivars, photosynthesis-related genes are generally downregulated under drought stress, while genes associated with stress responses and structural integrity are upregulated. Mitochondrial activity varies among cultivars, reflecting species-specific metabolic adjustments. These findings highlight the diverse drought adaptation strategies employed by wheat and point to potential genetic targets for breeding more resilient cultivars.

This study identifies candidate drought-responsive genes in wheat (e.g., TaZFP36, TaFER, TaPP2CA, TaGI, TaRBP45B) and highlights cultivar-specific strategies of tolerance. These genes provide targets for functional validation and genome editing to improve drought resilience. They also represent potential molecular markers that could be incorporated into breeding programs for selecting drought-tolerant genotypes. Integrating transcriptomic insights with breeding tools could support sustainable wheat production under water-limited conditions.