Integrated transcriptomic and biochemical analyses reveal mechanisms of drought adaptation in strawberry (Fragaria × ananassa L.) Authors: Hayat Topçu, Umutcan Ünaldı, Mesut Ada, Harun Karcı, Duygu Ayvaz Sönmez, Azam Akbari, Luca Mazzoni, Nesibe Ebru Kafkas, Salih Kafkas Published: 2026 ABSTRACT AbstractDrought is a major abiotic stress that limits strawberry productivity worldwide, yet the underlying physiological and molecular mechanisms of cultivar-specific tolerance remains poorly understood. This study integrated physiological, biochemical, and transcriptomic analyses to dissect drought responses of two contrasting cultivars of Fragaria × ananassa: drought-tolerant ‘Kara’ and drought-sensitive ‘E22’. Plants were exposed to both progressive water-deficit (Def) and PEG-induced osmotic stress under controlled greenhouse conditions. Physiologically, the two cultivars diverged substantially: E22 exhibited progressive chlorophyll decline and severe reduction in PSII efficiency (Fv/Fm dropping to 0.64 under PEG), whereas Kara maintained and even increased its chlorophyll content throughout the experiment while sustaining higher Fv/Fm values (> 0.72). Proline accumulation — a key osmoprotective response — was significantly higher in Kara (1.50 vs 0.78 µmol g⁻1 FW under PEG at peak stress). Transcriptomic profiling (18 RNA-Seq libraries, ~ 872 million reads) revealed fundamentally different stress strategies. E22 showed extensive transcriptome reprogramming, with 22,558 differentially expressed transcripts (DETs) under PEG, including down-regulation of photosynthesis-related pathways and up-regulation of ABA signaling, MAPK cascades, and heat shock response. In contrast, Kara exhibited a restrained response (< 900 DETs per stress condition), with differential expression focused on homeostasis-related functions such as transcription regulation, protein folding, and RNA processing. PEG-induced osmotic stress consistently elicited a stronger transcriptomic response than progressive water deficit in both cultivars, suggesting that osmotic sensing is a primary driver of large-scale gene regulation under drought. Enrichment analyses (GO/KEGG via ORA and GSEA) confirmed that E22 activated canonical drought and heat stress pathways, while Kara selectively engaged transcriptional regulators and RNA-processing genes. Transcription factor profiling showed extensive activation of AP2/ERF, NAC, WRKY, and MYB families in E22, with far fewer TFs altered in Kara. qRT-PCR validation of selected DEGs corroborated the RNA-Seq findings. These results provide new insights into cultivar-specific drought adaptation and identify candidate genes for breeding stress-resilient strawberry varieties.Supplementary InformationThe online version contains supplementary material available at https://doi.org/10.1186/s12870-026-09061-0.Keywords: Fragaria × ananassa, Physiological analysis, RNA-Seq, Transcription factor Introduction Drought is a major abiotic stress that increasingly threatens global agriculture due to climate change, leading to more frequent and severe water deficits in crop-growing regions [1]. As one of the most damaging environmental stressors, drought impairs photosynthesis, nutrient uptake, and cellular homeostasis, while promoting oxidative damage—often resulting in drastic yield losses [2, 3]. A comprehensive understanding of plant responses to drought is essential for developing resilient cultivars in the face of mounting climatic pressures. Strawberry (Fragaria × ananassa Duch.), an octoploid member of the Rosaceae family (2n = 8x = 56), is one of the most economically valuable berry crops worldwide. According to the Food and Agricultural Organization of the United Nations, more than nine million tonnes of strawberries were produced globally in 2021 [4, 5]. In addition to its commercial relevance, the strawberry is renowned for its rich nutritional content, including vitamins, minerals, and bioactive compounds [6]. However, the species remains particularly vulnerable to drought due to its shallow root system, expansive leaf area, and high water content in vegetative tissues, which collectively exacerbate reductions in biomass, fruit quality, and yield under water-limiting conditions [7]. The release of the first chromosome-scale reference genome for F. × ananassa cv. 'Camarosa' [8] has greatly facilitated genome-wide transcriptomic analyses in this complex allopolyploid. Although some genotypic variation in drought tolerance has been observed among strawberry cultivars—particularly in traits related to carbohydrate metabolism and physiological stress responses—the molecular basis of these differences remains poorly characterized [9]. Drought stress is known to elicit extensive transcriptomic reprogramming and metabolic remodeling [10], often accompanied by the accumulation of reactive oxygen species (ROS), which damage cellular macromolecules if not tightly regulated [11]. Plants are known to combat oxidative stress through enzymatic (e.g., SOD, POD, CAT) and non-enzymatic antioxidant systems, while osmolytes such as proline serve primarily as osmoprotectants with additional ROS-scavenging properties [12, 13]. Proline accumulation under drought has been documented in strawberry, where tolerant genotypes typically maintain higher proline levels than sensitive ones [14, 15]. Abscisic acid (ABA) is widely recognized as a key mediator of drought signaling in plants. Under stress, ABA biosynthesis is upregulated through 9-cis-epoxycarotenoid dioxygenase (NCED) genes, while its catabolism is epigenetically regulated [16–18]. ABA perception and signal transduction via PYR/PYL/RCAR receptors, PP2Cs, and SnRK2 kinases control key adaptive responses, including stomatal closure, osmolyte accumulation, and transcriptional reprogramming [18, 19]. Hormonal crosstalk among auxin, jasmonic acid, cytokinins, and salicylic acid further complicates drought adaptation [20]. Transcription factors (TFs) serve as key regulators of drought responses by orchestrating gene expression networks involved in osmoprotection, detoxification, and metabolism. Families such as WRKY, MYB, NAC, bZIP, and AP2/ERF are frequently implicated in stress signaling and tolerance mechanisms [21, 22]. For instance, overexpression of TaWRKY31 in wheat has been shown to improve drought resilience [23]. Additionally, non-coding RNAs are increasingly recognized as modulators of TF activity in stress responses [24, 25]. RNA sequencing (RNA-seq) has enabled high-throughput discovery of drought-responsive genes in various crops, revealing enrichment in hormone signaling, MAPK pathways, and secondary metabolism [26, 27]. Comparative transcriptomic profiling of contrasting genotypes—tolerant versus sensitive—has proven particularly effective for dissecting cultivar-specific stress adaptation strategies in crops such as rice, wheat, and soybean [28]. In strawberries, integrated omics approaches have begun to uncover transcriptomic, epigenomic, and metabolomic shifts during drought stress. Combined transcriptome–methylome analyses in Fragaria nilgerrensis revealed epigenetic regulation of key stress genes including NCED1 and CYP707A2 [16]. In cultivated strawberry, Jiang et al. [29] used transcriptomic and metabolomic profiling to demonstrate coordinated changes in ABA-related processes and carbohydrate metabolism under water deficit. Furthermore, Merlaen et al. [31] characterized contrasting water use behaviors in two F. × ananassa cultivars under drought and PEG-induced osmotic stress, linking cultivar-specific stomatal regulation to aquaporin expression patterns. PEG treatment induces osmotic stress by lowering the medium's water potential without causing ionic toxicity, thereby mimicking drought conditions at the physiological and molecular levels. PEG 6000 consists of high-molecular-weight polymers that are not readily taken up by plant cells, thereby reducing the water potential of the growth medium while avoiding direct cytotoxic effects [30]. PEG-induced osmotic stress is widely used as a complementary approach to soil-based water deficit for studying plant dehydration responses under controlled conditions. High-molecular-weight PEG (e.g., PEG-6000) reduces the water potential of the growth medium without being absorbed by plant roots, thereby simulating the osmotic component of drought stress [30]. Combining PEG treatment with progressive soil water deficit allows dissection of both osmotic and complex soil-drying responses [31]. Despite these insights, integrated transcriptomic analyses in cultivated strawberry (F. × ananassa) remain scarce. Most RNA-seq studies in Fragaria have been conducted on the diploid F. vesca or have focused on a limited number of candidate genes in the octoploid. Comparative transcriptomic studies between tolerant and sensitive cultivars of F. × ananassa are particularly lacking, hampering the identification of cultivar-specific drought adaptation strategies and molecular markers for breeding. Notably, no study to date has combined genome-wide transcriptomic profiling with physiological and biochemical characterization across multiple drought stress modalities in contrasting cultivars of cultivated strawberry. To address this gap, the objectives of this study were: (1) to characterize genome-wide transcriptional responses in two commercial strawberry cultivars exposed to water-deficit and PEG-induced osmotic stress; 2) to identify differentially expressed genes, TFs, and enriched pathways associated with drought adaptation; (3) to integrate transcriptomic findings with biochemical traits, such as proline accumulation and photosynthetic parameters; and (4) to elucidate cultivar-specific mechanisms underlying drought tolerance. This integrated approach aims to advance the molecular understanding of strawberry’s drought responses and provide molecular targets for the development of more resilient cultivars. Materials and methods Plant materials and growth conditions Two commercial strawberry (Fragaria × ananassa Duch.) cultivars with contrasting drought tolerance were used in this study: the drought-tolerant cultivar 'Kara' and the drought-sensitive cultivar 'E22'. 'Kara' is a native Turkish landrace predominantly cultivated in the Black Sea region of Türkiye, well-adapted to local environmental conditions and known for its resilience under water-limited conditions. 'E22' (syn. 'Florida Beauty') is a drought-sensitive cultivar widely used in strawberry production. The drought tolerance classification of these cultivars was established based on physiological and biochemical screening under PEG-induced osmotic stress [9]. Both cultivars were propagated from runner plants and established in 20 × 20 cm pots prior to stress treatments. Plants were cultivated in a substrate composed of peat, perlite, and vermiculite at a 2:1:1 (v/v/v) ratio under controlled greenhouse conditions maintained at approximately 25 °C, ~ 60% relative humidity, and a 16/8 h (light/dark) photoperiod. The experiment was arranged in a randomized complete block design (RCBD) with three biological replicates per treatment, each consisting of 10 pots per cultivar. Drought stress treatments Three treatments were applied to both cultivars: Control (Ctrl), Water Deficit Stress (Def), and Osmotic Stress (PEG). Ctrl plants were irrigated regularly every two days. For the Def treatment, irrigation was progressively reduced to impose 25%, 50%, 75%, and finally 100% water deficit relative to field capacity, with adjustments applied every two days following the methodology outlined by Doorenbos and Pruitt [32]. For the PEG treatment, regular irrigation was replaced by application of a 30% (w/v) polyethylene glycol 6000 (PEG-6000) solution every two days, applied until the substrate was fully saturated, to simulate osmotic stress. At this concentration, the estimated osmotic potential is approximately − 1.2 MPa [30], selected to induce severe but sub-lethal osmotic stress. Although PEG-induced osmotic stress is more commonly applied in hydroponic systems, its use in solid substrates provides a complementary approach that allows genotypic comparisons under controlled and reproducible osmotic conditions. The primary rationale for using PEG 6000 is that it mimics water-deficit conditions by limiting water availability to the roots, thereby inducing physiological responses similar to those of soil drought. Leaf tissues were sampled at two critical time points: (1) the onset of visible drought symptoms (day 14) and (2) peak stress severity (day 20). These two stress points were determined according to the results of SPAD (chlorophyll content) and IR (leaf temperature) measurements (Fig. 1). Samples were flash frozen in liquid nitrogen and stored at − 80 °C for transcriptomic and biochemical analyses. Physiological and biochemical measurements were performed at both time points, while RNA for transcriptomic analysis was extracted from leaf samples collected at day 20 (peak stress). RNA was isolated from the samples collected from stressed and control plants; the resulting RNA was sequenced, with the sequencing outsourced to BGI. All computational analyses were performed on a Linux-based high-performance computing (HPC) server running Ubuntu 20.04.6 LTS with the Linux kernel 5.15.0–139-generic. All scripts were implemented in Python (v3.11) [33] and R (v4.5.1) [34]. Control, restricted irrigation, and PEG-treated samples of E22 and Kara varieties Physiological and biochemical analyses Leaf chlorophyll content Chlorophyll content was measured on the upper 2–3 leaves of plants in each pot using a ClorofiLOG chlorophyll meter (Falker, Porto Alegre, Brazil). PSII maximum quantum yield (Fv/Fm) The maximum quantum yield of Photosystem II (Fv/Fm) was measured in dark-adapted, fully expanded leaves of each plant using a FluorPen FP 100 fluorometer (Photon Systems Instruments, Drásov, Czech Republic). Fv/Fm was calculated as (Fm − Fo)/Fm, where Fo is the minimal fluorescence and Fm is the maximal fluorescence yield of a dark-adapted leaf. Leaf temperature Leaf temperature (°C) was measured on intermediate-aged leaves near the growing tips using an infrared thermometer (IR). Measurements were taken between 10:00 and 11:00 on three plants per replicate. Proline content Free proline concentration was quantified according to the protocol of Bates et al., [35]. Leaf tissues from each treatment group were homogenized in 3% (w/v) sulfosalicylic acid and filtered through Whatman filter paper. After adding acid ninhydrin and ice-cold acetic acid, the reaction mixture was incubated in a 100 °C water bath for 1 h for color development. The reaction was stopped in an ice bath, extracted with toluene, and the absorbance of the toluene layer was measured at 520 nm. Proline concentration was calculated against a standard calibration curve and expressed as µmol of proline per g fresh weight (µmol g⁻1 FW). Statistical analysis Experimental data were analyzed using JMP® software (Version 13.0; SAS Institute Inc., Cary, NC, USA), following a randomized complete block design (RCBD). Treatment means were compared using the Least Significant Difference (LSD) test at the 5% significance level (p < 0.05). For each of the two strawberry cultivars (E22 and Kara), three replicates per treatment (Ctrl, Def and PEG) were used, with three plants per replicate. Molecular and bioinformatic analyses RNA extraction and sequencing Total RNA was extracted following Carvalho et al., [36]. RNA quality and concentration were assessed using a Qubit 3 RNA BR Analysis Kit (Invitrogen), and integrity was checked via 0.8% denaturing agarose gel electrophoresis. cDNA synthesis for qRT-PCR validation used the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Sequencing was outsourced to BGI (Beijing Genomics Institute, China). Adapter trimming and quality filtering were performed using SOAPnuke v2.2.1 [37]. Reads shorter than 15 nt, with > 5% ambiguous bases, or average quality below 0.5 were discarded. High-quality reads were aligned to the chromosome-scale octoploid Fragaria × ananassa ‘Camarosa’ reference genome assembly v1.0.a1 [8], obtained from the Genome Database for Rosaceae (GDR), using HISAT2 (v2.2.1) [38]. The assembly contained separate homeologous chromosome copies (e.g., Fvb1-1 to Fvb1-4 through Fvb7-1 to Fvb7-4), rather than a collapsed chromosome representation. Gene and transcript abundances were then estimated using RSEM (v1.3.3) (https://pubmed.ncbi.nlm.nih.gov/21816040/), which probabilistically assigns multi-mapped reads through an expectation–maximization framework. Transcript assembly was conducted with StringTie (v2.2.1) [39]. Novel transcripts were identified, and coding sequences were predicted using TransDecoder (v5.7.1) [40]. Non-coding RNA candidates were filtered using the LGC classifier and Rfam database using INFERNAL (v1.1.4) [41, 42]. Differential gene expression and network analyses Gene-level quantification was performed with tximport (v1.30.0), followed by normalization and differential expression analysis using DESeq2 (v1.44.0) [43]. K-means clustering and PCA were used to explore global expression patterns. Functional annotation and enrichment analysis Functional annotation was conducted against NR, NT, UniProt (Swiss-Pro/TrEMBL) [44], KEGG [45], COG/KOG [46], and PlantTFDB (v5.0) [47] databases using BLAST + (v2.16.0) [48] and DIAMOND (v2.0.15) [49]. GO [50] and KEGG enrichment analyses were performed using goseq (v1.52.0) and clusterProfiler (v4.10.0), respectively, and corrected for transcript length bias. Enriched terms were visualized using ggplot2 [51]. Gene expression heatmaps Clustered heatmaps were created to display expression patterns of key genes across samples. Genes with strong expression were selected via the filterByExpr function from edgeR [52], grouped by cultivar and treatment. After filtering out low-expression genes, the count matrix was variance-stabilized (VST). Heatmaps were generated using pheatmap (v1.0.12) [53], with both row and column clustering (Euclidean distance, complete linkage), and annotated by cultivar (E22 or Kara) and treatment (Ctrl, Def, PEG) per column. This allowed visualization of gene clusters sharing similar expression profiles across treatments. qRT-PCR validation Eight DEGs were selected for validation via qRT-PCR. Primer design was conducted using Primer3Plus, with exon-exon spanning primers to avoid genomic DNA amplification. cDNA was synthesized using High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Thermo Fisher Scientific), following the manufacturer’s protocol. qRT-PCR reactions were performed using Roche LightCycler® 96 with ACTIN as the reference gene. The relative expression was calculated using the 2⁻ΔΔCT method [54]. All assays included three biological replicates per treatment. The qRT-PCR primers used are listed in Supplementary Table 1. Results Chlorophyll content Chlorophyll content (SPAD values) was measured at two time points (day 14 and day 20) in both cultivars under all treatments. In the sensitive cultivar E22, chlorophyll content decreased significantly over time across all conditions: from 53.46 to 42.13 in control, from 44.63 to 38.16 under Def, and from 41.60 to 36.63 under PEG (Table 1). In contrast, the tolerant cultivar Kara showed a markedly different pattern: chlorophyll content increased over time in all treatments, rising from 42.56 to 53.36 in control, from 40.93 to 51.76 under Def, and from 40.56 to 51.30 under PEG (p < 0.0001; Table 1). These results indicate that E22 suffered progressive chlorophyll degradation under stress, whereas Kara maintained and even enhanced its chlorophyll content throughout the experimental period. At the end-point measurement, Kara retained significantly higher chlorophyll levels than E22 under both stress treatments, suggesting superior maintenance of photosynthetic pigments as a component of drought tolerance. As a result of comparing the results of the 14th and 20th days of the sensitive cultivar with the chlorophyll content of the tolerant cultivar Kara from the 14th day to the 20th day, all measurements of Ctrl plants, Def plants, and PEG-irrigated plants, all the results we obtained suggested that there was already a pre-existing structure related to tolerance in the structure of the tolerant plant, as it continued to develop under water stress. It was determined that the sensitive cultivar did not have this structure (Fig. 2). Table 1Physiological and biochemical parameters of E22 and Kara strawberry cultivars under control (Ctrl), progressive water deficit (Def), and PEG-induced osmotic stress (PEG) conditions. Values represent mean ± three biological replicates. Different lowercase letters within columns indicate significant differences (Duncan’s multiple range test, P < 0.05). NS, not significantSampleChlorophyll (SPAD)Leaf Temperature (°C)PSII (Fv/Fm)Proline (µmol g⁻1 FW)FirstEndFirstEndFirstEndFirstEndE22_Ctrl53.46 a42.13 c20.2322.500.8300 a0.8258 a0.31 d0.38 dE22_Def44.63 b38.16 d20.7623.760.7611 b0.7388 b0.52 c0.55 cdE22_PEG41.60 cd36.63 e22.3024.860.7376 b0.6436 c0.70 b0.78 bcKara_Ctrl42.56 c53.36 a23.7624.630.7909 ab0.7587 b0.55 bc0.55 cdKara_Def40.93 de51.76 b24.0626.260.7742 a0.7329 b0.63 bc0.86 bKara_PEG40.56 e51.30 b25.8027.300.7548 ab0.7188 b1.23 a1.50 aProb > F < 0.0001* < 0.0001*0.88330.51100.0415*0.0032*0.0069*0.0146*LSD1.0020.556NSNS0.062430.056890.1710.247SPAD Soil Plant Analysis Development, PSII (Fv/Fm) Maximum quantum yield of Photosystem II, FW Fresh weight*indicates statistical significance at P < 0.05 Physiological and biochemical parameters of E22 and Kara strawberry cultivars under control (Ctrl), progressive water deficit (Def), and PEG-induced osmotic stress (PEG) conditions. Values represent mean ± three biological replicates. Different lowercase letters within columns indicate significant differences (Duncan’s multiple range test, P < 0.05). NS, not significant SPAD Soil Plant Analysis Development, PSII (Fv/Fm) Maximum quantum yield of Photosystem II, FW Fresh weight *indicates statistical significance at P < 0.05 Chlorophyll content (SPAD) for both E22 and Kara genotypes under drought-stress treatments (○ Circle markers = first measurement, ▲ Triangle markers = end measurement) Leaf temperature Leaf surface temperature was measured with an infrared thermometer between 10:00–11:00 a.m. at both time points. Numerically, leaf temperature tended to increase under drought in both cultivars, with PEG-treated plants showing the highest values (Table 1). However, the overall treatment effect was not statistically significant for either time point (first measurement: p = 0.88; end measurement: p = 0.51; LSD test, Table 1). Although individual treatment means were not significantly different, the numerical trend of higher leaf temperatures under stress is consistent with the expected physiological response of reduced transpiration and partial stomatal closure under drought (Fig. 3). Leaf temperature measured by IR for both E22 and Kara varieties under different drought stress treatments (P < 0.001) PSII maximum quantum yield (Fv/Fm) Photosystem II (PSII) maximum quantum yield (Fv/Fm) decreased under drought in both cultivars. In Kara, Fv/Fm declined modestly — from 0.7909 (control) to 0.7548 under PEG at the first measurement; at the end measurement, Fv/Fm dropped further to 0.7188. In contrast, E22 showed a more severe decrease: from 0.8300 (control) to 0.7376 under PEG, reaching as low as 0.6436 at the end of measurement (Fig. 4). The treatment effect was statistically significant at both time points (first measurement: p = 0.042; end measurement: p = 0.003; Table 1). These findings indicate that PSII photochemical efficiency is more severely compromised under drought in the sensitive cultivar, especially under PEG-induced osmotic stress. PS II activity for both E22 and Kara varieties under different drought stress treatments (P < 0.001) Proline accumulation Proline content, an established osmoprotectant marker, increased under both drought treatments in both cultivars (p = 0.007 at first measurement; p = 0.015 at end measurement; Table 1). However, the magnitude of accumulation was significantly higher in Kara plants compared to E22 plants. Under PEG treatment, Kara accumulated the highest proline levels among all treatment groups (1.23 µmol g⁻1 FW at the first measurement, reaching 1.50 µmol g⁻1 FW at end measurement), whereas E22 under PEG reached only 0.70 and 0.78 µmol g⁻1 FW, respectively. Even though E22 plants also showed proline increase under both stress conditions, its accumulation remained substantially lower compared to Kara plants (Fig. 5). This supports the hypothesis that enhanced osmotic adjustment via proline contributes to drought tolerance in strawberry (Table 2). Changes in proline content of strawberry cultivars subjected to control plant, restricted irrigation, and PEG application (P < 0.001) Table 2Primer sequences used for qRT-PCR validation of differentially expressed transcripts (DETs) identified from RNA-Seq analysisNoPrimer NameSequence (5' → 3')Length (bp)1E22K_pd_116.3FAGTGACCATTCCAGGGCCAG202E22K_pd_116.3RTCGTCGTCAGAGTCGTCGTC203E22K_pd_83.56_1-2FATGGAGACTCGGGTCCGTTT204E22K_pd_83.56_1-2RACGTGCCTGGTGAGTACGTT205E22K_pd_83.56_2FGCTGCGTAACACTCGTCTGC206E22K_pd_83.56_2RCCGGTGGACACAGCACTACT207E22K_pd_70.17_1FTCCGTAGCCAGTTCCGGTTC208E22K_pd_70.17_1RGGAAACCCCGTGGGACGTAA209E22K_pd_70.17_2FTGGGCATGACACCCTTTGATCT2210E22K_pd_70.17_2RGCAGACGAGTGTTACGCAGC2011E22K_pd_61.56_1FTTCCTGTTCCGGTTCCGGTG2012E22K_pd_61.56_1RGGAAACCCCGTGGGACGTAA2013E22K_pd_61.56_2FGTAGTGGTGTGTCCACCGGG2014E22K_pd_61.56_2RTCCCCACAACACCACTCACA2015E22K_pd_72.47FGACCTGCTTCCTCACGACCA2016E22K_pd_72.47RAACCCTCAGCCCATGAAGCA20Eight gene targets were selected based on differential expression between cultivars and stress treatmentsF Forward primer, R Reverse primer Primer sequences used for qRT-PCR validation of differentially expressed transcripts (DETs) identified from RNA-Seq analysis Eight gene targets were selected based on differential expression between cultivars and stress treatments F Forward primer, R Reverse primer RNA-seq quality, read mapping, and global transcriptome characteristics Sequencing of 18 libraries generated a total of 872.15 million paired-end reads (150 bp). After adapter and quality trimming (removing reads < 15 nt, reads with > 5% ambiguous nucleotides, or average quality below Q20), 768.96 million clean read pairs were retained (~ 88.2% of raw reads), corresponding to ~ 115.33 Gb of data (~ 6.4 Gb per library on average). Base quality was high: 98.37% of bases had Q20 or above, and 94.15% had Q30 or above. Alignment to the Fragaria × ananassa ‘Camarosa’ reference genome (v1.0) yielded an average mapping rate of 97.27%, with 81.05% uniquely mapped reads. Strand-specific orientation was confirmed, with ~ 88.19% of reads aligning to the expected strand (Table S1). Variance-stabilized count data (via DESeq2) showed high reproducibility among biological replicates: Pearson correlation coefficients ranged from 0.957 to 0.992 among all 18 libraries (mean 0.973). Within treatment groups, correlations ranged between ~ 0.960 and ~ 0.985. Alternative normalization using log₂(TPM + 1) produced comparable consistency (mean r = 0.944), confirming robustness across normalization methods (Figures S1–S2; Tables S2–S3). Transcript discovery and global expression patterns Per-sample reference-guided transcriptome assemblies yielded an average of 110,243 ± 5,303 transcript models per library (range 97,600–118,400). On average, 31.3% of transcripts matched annotated reference genes, while 4.2% were classified as novel intergenic transcripts; the remainder represented novel isoforms of known genes (Table 3). Across all libraries, a total of 1,984,366 transcript instances (620,748 known + 83,561 novel intergenic) were reconstructed, encompassing 10,699,563 exons. The average exon count per transcript was ~ 5.39 (range 4.50–6.18). Notably, Kara libraries consistently showed slightly higher total transcript counts than E22, especially under PEG stress. Table 3Summary of transcript discovery per library. Total, known, and novel intergenic transcripts, and the number of exons identified in each sample library using StringTie assemblySampleTotal TranscriptsKnown TranscriptsNovel Intergenic TranscriptsNo. of ExonsE22_Ctrl1113,02832,0804,652698,991E22_Ctrl2108,46233,9294,587620,464E22_Ctrl3108,70835,1804,590591,112E22_Def1102,52134,5354,342540,215E22_Def2105,37430,7474,378604,742E22_Def3111,22733,3104,641634,969E22_PEG1109,52632,2914,507586,912E22_PEG2106,33630,1764,360596,761E22_PEG3114,21031,3464,687667,348Kara_Ctrl197,60936,3244,233439,374Kara_Ctrl2113,67536,4764,859591,682Kara_Ctrl3112,87636,5004,732591,227Kara_Def1107,01235,8664,561569,888Kara_Def2110,39434,4664,712603,604Kara_Def3110,95736,6414,747585,886Kara_PEG1116,25136,7454,921578,231Kara_PEG2117,79036,7384,985601,232Kara_PEG3118,41037,3985,067596,925Total1,984,366620,74883,56110,699,563 Summary of transcript discovery per library. Total, known, and novel intergenic transcripts, and the number of exons identified in each sample library using StringTie assembly Global expression analysis of the 2,000 most variable genes via PCA revealed clear clustering by cultivar and treatment (Fig. 6). Biological replicates grouped tightly, indicating no major batch effects. Hierarchical clustering further separated the two cultivars across all conditions: E22 samples (Ctrl, Def and PEG) formed one cluster, and all Kara samples formed another. Within E22, stress-treated samples (Def and PEG) clustered closely, while in Kara, samples showed a more graded separation: PEG stress diverged the most, followed by Def, with control remaining closer to stress conditions (Fig. 7, Figure S3–S4). These patterns suggest extensive transcriptome reprogramming in E22 under stress, but a more moderated response in Kara. PCoA analysis showing high consistency within each group and between each group Heat map of all samples connected to each other Differentially expressed genes (DEGs) Using thresholds of |log₂FC|≥ 1 and FDR-adjusted p ≤ 0.01, a cumulative total of 301,279 differentially expressed transcript (DET) instances were detected across all 15 pairwise comparisons (note: individual transcripts appearing in multiple comparisons are counted each time). In polyploid species, gene expression patterns are further complicated by the presence of homeologous gene copies derived from different subgenomes. These homeologs may exhibit differential expression, subgenome dominance, or condition-specific regulation, particularly under stress conditions. Therefore, the high number of differentially expressed transcripts observed in the sensitive genotype may reflect not only transcriptional reprogramming but also differential regulation among homeologous gene copies. In contrast, the more limited response in the tolerant genotype may indicate a more coordinated and stable regulation of homeolog expression. In E22, PEG treatment (E22_PEG vs E22_Ctrl) yielded 22,558 DE transcripts (11,197 upregulated, 11,361 downregulated), while water-deficit stress (E22_Def vs E22_Ctrl) resulted in 11,081 DE transcripts (5,376 up, 5,705 down). The direct comparison between PEG and water-deficit treatments in E22 (E22_PEG vs E22_Def) yielded only 1,144 DE transcripts, indicating substantial overlap in the stress responses. In contrast, the tolerant cultivar Kara showed far fewer DETs: Kara_Def vs Kara_Ctrl produced 442 DETs, and Kara_PEG vs Kara_Ctrl produced 858 DETs (602 upregulated, 256 downregulated). However, comparison between PEG and water-deficit conditions in Kara (Kara_PEG vs Kara_Def) revealed 4,005 DE transcripts, suggesting PEG induces a distinct transcriptional adjustment even in the tolerant genotype. The markedly higher number of differentially expressed transcripts in E22 under PEG treatment relative to its control (22,558 DETs) compared with Kara (858 DETs) suggests that the sensitive genotype undergoes a much broader transcriptional reprogramming in response to osmotic stress. This extensive response likely reflects greater disruption of cellular homeostasis and a more reactive stress response. In contrast, the limited number of DETs in the tolerant genotype Kara suggests a more restrained, potentially more efficient transcriptional response to stress, although this could also partly reflect reduced statistical power due to higher within-group variability. Notably, the large number of DETs already detected between E22_Ctrl and Kara_Ctrl (15,352 DETs) demonstrates that the two genotypes differ substantially at the basal transcriptomic level even before stress exposure. This finding suggests that drought tolerance in Kara may be associated not only with stress-induced responses but also with constitutive gene expression patterns that provide a pre-adapted molecular background for stress resilience (Table 4). Table 4Number of up- and down-regulated differentially expressed transcripts (DETs) for key pairwise comparisons between E22 and Kara cultivars under control, water deficit, and PEG stress conditions (|log₂FC|> 1, adjusted P-value < 0.01; DESeq2 v1.44.0)ControlCaseUp-regulatedDown-regulatedTotal DETsE22_CtrlE22_Def5,3765,70511,081E22_CtrlE22_PEG11,19711,36122,558E22_DefE22_PEG8872571,144E22_DefKara_Def15,18414,62829,812E22_PEGKara_PEG19,52717,25936,786Kara_CtrlKara_Def207235442Kara_CtrlKara_PEG602256858Kara_DefKara_PEG2,2651,7404,005Total55,24551,441106,686Comparisons shown represent the focal contrasts used in the main text analysisDETs Differentially expressed transcripts Number of up- and down-regulated differentially expressed transcripts (DETs) for key pairwise comparisons between E22 and Kara cultivars under control, water deficit, and PEG stress conditions (|log₂FC|> 1, adjusted P-value < 0.01; DESeq2 v1.44.0) Comparisons shown represent the focal contrasts used in the main text analysis DETs Differentially expressed transcripts Cross-cultivar contrasts revealed large baseline differences: under control conditions, E22_Ctrl vs Kara_Ctrl yielded 15,352 DE transcripts. Under stress, cross-genotype comparisons (e.g., E22_PEG vs Kara_Ctrl) resulted in tens of thousands of DE transcripts (e.g., ~ 40,673), highlighting fundamentally different transcriptomic landscapes between the two cultivars (Table S4; Fig. 8). Volcano plot of stress responses of two genotypes at the transcriptome level Functional enrichment and transcription factor regulation under drought Both Over-Representation Analysis (ORA) and Gene Set Enrichment Analysis (GSEA) were applied to DEG sets, controlling for transcript length bias and multiple testing (FDR ≤ 0.01). In E22, upregulated genes under stress were strongly enriched in classical drought-responsive GO terms (Figure S5), including “response to water deprivation,” “osmotic stress,” “cellular water homeostasis,” “ABA-activated signaling,” and “stomatal movement”. Regulatory pathways such as MAPK cascade regulation, ubiquitin-mediated proteolysis, and hormone signal transduction were also enriched, reflecting complex hormonal and stress signaling crosstalk. Downregulated GO categories included photosynthesis, chloroplast and plastid organization, porphyrin/chlorophyll metabolism, ribosome biogenesis, and DNA replication — indicating resource reallocation away from growth and photosynthesis. PEG stress uniquely induced enrichment in protein processing in the endoplasmic reticulum, suggesting increased demand for proteostasis under osmotic stress. In Kara, upregulated genes under stress were modestly enriched for “transcription regulator activity,” “protein folding,” and “RNA processing/splicing,” but classical drought-response terms (e.g., “water deprivation,” “ABA signaling,” “osmotic stress”) were absent or weakly enriched (Figure S6). Downregulated genes showed minimal enrichment in photosynthetic or metabolic pathways, indicating that core metabolism remained largely intact under stress in the tolerant cultivar. Given that osmotic stress is known to activate heat stress-responsive pathways through shared proteotoxic signaling mechanisms, we specifically examined heat- and temperature-stress responses. In E22 (under both Def and PEG), there was strong enrichment of heat shock factors (HSFs), molecular chaperones, and stress-related proteins; GO terms such as “response to heat,” “response to high temperature,” and “heat acclimation” were significantly enriched (Figure S5). This activation likely reflects an osmotic stress-induced cross-response rather than a direct consequence of leaf warming, as leaf temperature differences were not statistically significant (p > 0.05). Transcripts associated with dehydration and cold stress (e.g., dehydrins, cold-shock proteins) were also induced, reflecting the well-documented overlap between osmotic and temperature stress signaling. In contrast, Kara showed only modest induction of a few chaperones under PEG, and very slight enrichment under water deficit, suggesting more efficient proteostatic regulation under stress (Figure S6). In the matched cultivar comparisons under stress, E22 also showed broader functional disruption than Kara (Figure S7). Under water deficit, enriched GO terms were dominated by photosynthesis- and plastid-related categories, including photosynthesis, light reaction, photosynthetic electron transport chain, photosynthetic electron transport in photosystem I, and plastid structural components Molecular function terms were likewise centered on redox and pigment-associated activities, including oxidoreductase activity, pigment binding, chlorophyll binding, and tetrapyrrole binding. In the PEG comparison, this same plastid/photosynthesis signature remained prominent, but was accompanied more clearly by stress-responsive biological processes such as response to water, response to water deprivation, response to abscisic acid, response to osmotic stress, and response to salt stress. The corresponding molecular functions also expanded to include regulatory activities such as hormone receptor binding and transcription regulator activity. Taken together, these matched-stress comparisons indicate that, relative to the tolerant Kara cultivar, E22 undergoes a stronger reorganization of chloroplast-associated functions, photosynthetic metabolism, and stress-signaling pathways under both water deficit and PEG treatment.The enrichment network maps emphasized clear differences in functional organization between the two cultivars (Fig. 9). In the control comparison (E22_Ctrl vs Kara_Ctrl), enriched processes already separated into distinct modules, including photosynthesis and energy metabolism on one side and stimulus response, cell cycle, and chromosome organization on the other, indicating substantial constitutive divergence between E22 and Kara. Under PEG treatment, E22 (E22_Ctrl vs E22_PEG) showed a large stress-centered network dominated by response to water deprivation, osmotic stress, hormone-related signaling, and abiotic stimulus, together with a connected photosynthesis/light-reaction branch. In contrast, Kara under PEG (Kara_Ctrl vs Kara_PEG) displayed a smaller and more compact network enriched mainly for oxidative stress, reactive oxygen species, temperature-related responses, hydrogen peroxide response, and protein folding. In the direct PEG comparison (E22_PEG vs Kara_PEG), stress-signaling terms such as abscisic acid, osmotic stress, water deprivation, salt stress, and chemical response remained prominent, while photosynthesis-related terms formed a separate module, further distinguishing the two cultivars under the same treatment. Enrichment network map (ORA) of the most enriched GO terms for E22_Ctrl vs Kara_Ctrl, E22_Ctrl vs E22_PEG, Kara_Ctrl vs Kara_PEG and E22_PEG vs Kara_PEG in biological processes Gene Set Enrichment Analysis (GSEA) revealed pronounced genotype-dependent differences in pathway regulation under both deficit and PEG-induced stress conditions. In the tolerant genotype Kara, pathways related to photosynthesis, carbon fixation, and core metabolic processes were consistently enriched, indicating the preservation of energy metabolism and physiological stability under stress. In contrast, the sensitive genotype E22 exhibited significant enrichment of stress-responsive and regulatory pathways, including MAPK signaling (Fig. 10), RNA processing, and osmotic and abiotic stress response pathways, reflecting a broader and more reactive transcriptional reprogramming. These findings suggest that drought tolerance in Kara is associated with the maintenance of fundamental metabolic functions, whereas E22 responds through extensive activation of stress-related signaling networks. Summary bar charts (GSEA) of the most enriched terms E22_Def vs Kara_Def, E22_PEG vs Kara_PEG (KEGG) To explore transcriptional regulation under drought, all differentially expressed (DE) transcripts were mapped to known plant TF families using PlantTFDB [47]. In the drought-sensitive cultivar E22, drought stress led to strong upregulation of TFs, particularly from the AP2/ERF (notably DREB), NAC, MYB, WRKY, and bZIP families—key regulators of abiotic stress. Additional activation in bHLH, HD-ZIP, and HSF families suggested integration of hormonal, osmotic, and heat stress signaling. Few TFs were downregulated, indicating a primarily activatory response. PEG stress induced slightly more TF changes than water deficit, consistent with stronger osmotic signaling. In contrast, the tolerant cultivar Kara showed far fewer TF changes, with a balanced mix of up- and down-regulation. Upregulated TFs belonged to the same core families but were limited in number, suggesting reliance on a smaller set of master regulators rather than broad transcriptional shifts. PEG also triggered more TF changes than water deficit in Kara. Cross-cultivar comparisons (e.g., E22_Def vs. Kara_Def) confirmed these trends: many TFs induced in E22 were unchanged in Kara. This highlights E22's reactive regulatory reprogramming versus Kara's restrained, potentially constitutive tolerance. Detailed TF family-level changes are in Table S5, with key patterns in Fig. 11. Summary visualization of TF family contributions to each response qRT-PCR validation of selected DEGs To validate RNA-seq findings, eight differentially expressed genes (DEGs) were randomly selected for quantitative real-time PCR. cDNA was synthesized from leaf samples of both cultivars under control, Def, and PEG treatments were collected on day 14 (onset of drought symptoms) and day 20 (peak stress). Relative expression analysis using the 2⁻ΔΔCT method showed that seven out of eight genes were significantly upregulated under drought, with the highest expression typically observed on day 20 in PEG-treated plants. In all cases, expression levels were higher in Kara than in E22 under stress conditions (Fig. 12). One gene (E22K_8356_2) remained undetectable in all control and stress samples for both cultivars, indicating very low basal expression or cultivar-/condition-specific regulation. The strong concordance between RNA-Seq and qRT-PCR data supports the reliability of the transcriptomic results and the potential role of these genes in drought response. qRT-PCR in samples taken from tolerant Kara and susceptible E22 varieties during two stress periods. (First (1) the period when the first symptoms of drought appeared (14th day) and second (2) the period when the drought intensity reached the maximum (20th day)) (Relative expression of primer; a: E22K_116_3, b: E22K_8356_2, c: E22K_,7017_1, d: E22K_7017_2, e: E22K_,6156_1, f: E22K_,6156_2, g: E22K_7247) Discussion This study presents a comparative analysis of drought responses in two commercial strawberry cultivars with contrasting tolerance — ‘Kara’ (tolerant) and ‘E22’ (sensitive) — by integrating physiological measurements, biochemical profiling, and transcriptomic data. The results collectively indicate that Kara’s drought tolerance is driven by preemptive homeostatic control and restrained gene expression, while E22 responds with broad stress reprogramming that compromises core metabolic processes. Our findings are largely consistent with prior studies on strawberry and other crops: drought reduces chlorophyll, disrupts photosynthesis, increases proline and other osmolytes, and causes broad transcriptomic shifts. For example, a glasshouse study on strawberry cultivars under field capacity–based water deficit reported decreased total chlorophyll, reduced leaf gas exchange, increased proline content, and overall yield decline under severe drought [15]. Similarly, Attar et al. [9] screened 16 F. × ananassa cultivars under PEG-induced stress and found that drought significantly reduced relative water content, chlorophyll, and Fv/Fm, while cultivar-specific responses in sugar accumulation and antioxidant capacity distinguished tolerant from sensitive genotypes. Another study showed that stress tolerant plants often maintain higher proline levels and better physiological performance than sensitive ones under drought [14]. However, our study adds to the literature by combining physiological, biochemical, and genome-wide transcriptomic data — enabling a comprehensive view of cultivar-specific drought responses. Physiologically, the two cultivars diverged substantially. In E22, chlorophyll content declined progressively under both Def and PEG treatments, consistent with drought-induced chloroplast degradation widely documented in strawberry [9, 15]. In contrast, Kara not only maintained but increased its chlorophyll content over the experimental period across all treatments — including control — suggesting active chloroplast biogenesis or delayed leaf senescence rather than simple preservation of existing pigments. This finding is noteworthy, as most drought studies in strawberry report chlorophyll decline [9, 15]; the increase observed in Kara may reflect a genotype-specific developmental trajectory or stress-induced stay-green phenotype that warrants further investigation. The PSII maximum quantum yield (Fv/Fm) decreased under stress in both cultivars, but E22 suffered a more severe decline (reaching 0.64 under PEG), indicating substantial photoinhibition, whereas Kara remained above 0.72 in all treatments. It should be noted that leaf surface temperature showed a numerical increase under drought in both cultivars; however, this trend did not reach statistical significance (p > 0.05), possibly due to the controlled greenhouse environment that limited ambient thermal variation. The higher levels of proline in Kara under stress, on the other hand, suggest that osmotic adjustment and solute accumulation are essential for its ability to handle drought. Proline is a well-known osmoprotectant, stabilizing proteins and membranes, scavenging free radicals, and helping maintain cell turgor under water deficit [55]. Kara showing significantly greater proline accumulation than E22 (1.50 vs 0.78 µmol g⁻1 FW under PEG at peak stress) suggests that osmoregulation is part of its protective strategy. The increased proline accumulation in Kara substantiates its function as an osmoprotectant and ROS scavenger, aligning with prior studies in strawberry and other crops [13]. Together, these physiological and biochemical data support the idea that Kara adopts a “conservative” but effective drought-tolerance strategy—preserving photosynthetic capacity, maintaining osmotic balance, and limiting stress-induced damage—whereas E22 responds with aggressive stress signals but suffers more substantial photosynthetic and likely growth impairment. Transcriptomically, the contrast between cultivars was striking: E22 exhibited over 20,000 DEGs under PEG treatment, with significant enrichment in ABA signaling, MAPK cascades, ubiquitin-mediated proteolysis, and heat response pathways — a classic high-stress reprogramming profile. The much larger number of differentially expressed transcripts (DETs) in E22 (tens of thousands) compared with Kara (only a few hundred under each stress) indicates fundamentally different transcriptional strategies. Such reprogramming — diverting resources from growth/photosynthesis to stress mitigation — is a classic drought response strategy observed in many plants [56, 57]. Kara, in contrast, displayed < 500 DEGs per stress condition and selectively upregulated genes related to transcription regulation, protein folding, and RNA processing — a restrained but targeted response suggesting efficient regulatory control. However, we acknowledge that the substantially lower number of DETs in Kara could also partly reflect reduced statistical power arising from higher within-group transcriptomic variability or overall lower expression levels in the tolerant genotype, rather than solely a genuinely restrained transcriptional response. This pattern is consistent with the concept of "constitutive" or "primed" tolerance, where tolerant genotypes maintain a baseline protective state that requires only minor transcriptional adjustment upon stress, as opposed to the extensive emergency-response reprogramming seen in sensitive genotypes. Indeed, Merlaen et al. [31] described a similar dichotomy in strawberry at the physiological level: the tolerant cultivar 'Figaro' showed gradual, conservative stomatal regulation under drought, while the sensitive 'Flair' exhibited sudden, drastic stomatal closure — paralleling the restrained vs reactive transcriptomic patterns observed in Kara and E22, respectively [58]. Notably, our experimental design allowed direct comparison of two distinct drought modalities. PEG-induced osmotic stress consistently triggered a stronger transcriptomic response than progressive water deficit in both cultivars: E22 produced 22,558 DETs under PEG vs 11,081 under Def; Kara produced 4,005 DETs when comparing PEG to Def conditions. This is consistent with the nature of PEG treatment, which imposes a more acute osmotic challenge at a defined water potential, whereas progressive soil drying is a gradual process that allows time for physiological acclimation [30]. The substantially stronger response to PEG underscores that the rate and mode of water potential decline shape transcriptional outcomes, and suggests that osmotic sensing — rather than dehydration per se — may be a primary trigger for large-scale gene regulation under drought. However, one limitation of the transcriptomic analysis should be noted. Although reads were aligned to a non-collapsed octoploid reference genome and multi-mapped reads were handled probabilistically during quantification, ambiguity among highly similar homeologous loci remains possible because explicit subgenome-specific assignment was not performed. The network-level patterns in enrichment network maps also support the broader conclusion that drought response in E22 is extensive and highly interconnected, whereas Kara relies on a more restricted and targeted adjustment. In E22, stress-responsive processes expanded into large modules that linked osmotic, hormonal, and photosynthetic functions, consistent with widespread transcriptional reprogramming. By contrast, Kara was characterized by a tighter network centered on oxidative protection, temperature-related responses, and protein homeostasis, suggesting better maintenance of cellular balance under stress. Together, these patterns reinforce the idea that Kara tolerates drought through controlled adjustment, whereas E22 responds through a broader and potentially more costly stress program. A striking finding was the large baseline transcriptomic divergence between the two cultivars: 15,352 DETs were detected between E22_Ctrl and Kara_Ctrl alone, before any stress was applied. This indicates that the two cultivars possess fundamentally different constitutive gene expression landscapes, which likely reflect the distinct genetic backgrounds of a locally adapted Turkish cultivar (Kara) and a breeding line (E22). Such pre-existing transcriptomic differences may precondition the stress response, with Kara's baseline expression profile potentially already including protective gene networks that E22 must actively induce upon stress.The divergent activation of TF families further supports this interpretation: E22 massively upregulated TFs from stress-responsive families (AP2/ERF, NAC, MYB, WRKY, bZIP, HSF, etc.), whereas Kara induced only a small subset of these families. Broad activation of stress TF networks in sensitive genotypes has been documented previously as part of reactive transcriptional reprogramming [21, 28]. Kara induced only a small, focused subset of these TFs, potentially representing master regulators. This difference implies that the tolerant cultivar may possess an inherently stable gene regulatory network, requiring minimal perturbation to adapt. Importantly, our findings resonate with recent insights from transcriptomic studies in Fragaria vesca and other strawberry systems, where functional analyses of stress genes suggest that targeted regulatory changes can confer drought resilience [16, 29, 59]. A recent integrated transcriptomic and metabolomic comparison of 'Kaorino' (tolerant) and 'Benihoppe' (sensitive) F. × ananassa cultivars under drought also identified cultivar-specific transcriptional responses and highlighted key signaling pathways including MAPK and glycerophospholipid metabolism as determinants of differential tolerance [59]. However, unlike many previous studies that focus solely on transcriptomics or physiology, our integrative approach allows for direct correlation between expression dynamics and physiological performance, strengthening causal interpretations. The consistent results between RNA Seq and qRT PCR validation for selected DEGs reinforce the reliability of our transcriptome data and support that the observed gene expression changes likely underpin, at least partially, the physiological differences between cultivars. Several limitations of this study should be acknowledged. First, biochemical analyses were limited to proline as an osmoprotective marker; activities of core antioxidant enzymes (SOD, POD, CAT) and levels of non-enzymatic antioxidants (ascorbate, glutathione) were not measured. Consequently, the role of enzymatic ROS scavenging in the differential drought tolerance of the two cultivars cannot be directly assessed from the present data, although the transcriptomic evidence provides indirect support for differential regulation of antioxidant-related pathways. Second, despite the emphasis on hormonal signaling in the Introduction, direct measurements of ABA, JA, and ethylene were not performed; the observed changes in hormone-related gene expression therefore remain indirect evidence for hormonal pathway activation. Third, the exclusive focus on leaf tissue precludes conclusions about root-specific adaptations, which are central to drought perception and water uptake. Fourth, the absence of day 0 baseline chlorophyll measurements limits our ability to fully distinguish developmental effects from stress-induced changes. Finally, although RNA-seq at day 14 could have revealed early transcriptional dynamics, the present study prioritized peak stress (day 20) to maximize genotypic differences. In conclusion, our results suggest that minimal, precise transcriptomic modulation — rather than widespread gene activation — may be a hallmark of effective drought tolerance in strawberry. The tolerant cultivar Kara combined physiological resilience (maintained chlorophyll, stable Fv/Fm, high proline accumulation) with a restrained transcriptomic profile, whereas the sensitive E22 mounted an extensive but metabolically costly emergency response. The candidate genes validated by qRT-PCR, particularly those upregulated in Kara under stress, represent promising targets for marker-assisted selection in breeding programs. Future studies should prioritize field-level validation under long-term water-deficit scenarios, multi-omics integration (e.g., proteomics, metabolomics), and functional characterization of key regulatory genes to translate these findings into practical breeding strategies for climate-resilient strawberry production. Conclusion The central finding of this study is that effective drought tolerance in strawberry may rely on minimal, targeted transcriptomic modulation rather than widespread gene activation. The tolerant cultivar Kara combined physiological resilience with a highly restrained regulatory response, suggesting that its tolerance is at least partly constitutive—rooted in pre-existing gene expression patterns that require only fine-tuned adjustment upon stress. In contrast, the extensive transcriptional reprogramming observed in E22 reflects a reactive, metabolically costly emergency strategy that ultimately fails to prevent photosynthetic damage. This dichotomy between constitutive preparedness and reactive reprogramming may represent a general framework for understanding drought tolerance variation not only in strawberry but also in other polyploid crop species with complex genomes. From a translational perspective, the candidate genes validated via qRT-PCR provide promising molecular markers for marker-assisted selection in breeding programs. However, several limitations of the present study should be considered when interpreting these results. The exclusive focus on leaf tissue, the absence of direct measurements of antioxidant enzyme activities and hormone levels, and the lack of day 0 baseline data for chlorophyll all constrain the scope of our conclusions. Future research should prioritize: (i) root-level transcriptomic and physiological analyses to capture the full spectrum of tissue-specific drought responses; (ii) comprehensive antioxidant and hormone profiling to link transcriptomic changes to biochemical outputs; (iii) multi-omics integration incorporating proteomics and metabolomics; and (iv) field-based validation under long-term, realistic drought scenarios to assess the agronomic relevance of the identified candidate genes. Supplementary Information Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Acknowledgements Not applicable. Authors’ contributions HT: Conceptualization, Formal analysis, Investigation, Data curation, Bioinformatics analysis, Writing – original draft, Writing – review & editing, Visualization. UÜ: Investigation, Data curation, Methodology, Sample preparation, RNA extraction. AA: Writing – review & editing. DAS: Investigation, Plant material maintenance, Physiological and biochemical measurements. HK: Bioinformatics analysis, Software, Data curation, Formal analysis, Validation. MA: Investigation, Methodology, Physiological measurements, Data collection. LM: Writing – review & editing, Methodology, Biochemical analysis. NEK: Supervision, Resources, Writing – review & editing, Funding acquisition, Project administration. SK: Conceptualization, Supervision, Funding acquisition, Project administration, Writing – review & editing. All authors have read and approved the final version of the manuscript. Funding This work was supported by the Scientific and Technological Research Council of Türkiye (TÜBİTAK) under the TOVAG Project No. 221O354. Data availability The RNA-sequencing datasets generated and analysed during the current study will be deposited in the NCBI Sequence Read Archive (SRA) and made publicly available upon acceptance of the manuscript. The accession numbers will be provided before publication. Declarations Ethics approval and consent to participate Not applicable. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Azam Akbari, Email: AAkbari7345@gmail.com. Nesibe Ebru Kafkas, Email: ebruyasakafkas@gmail.com. Salih Kafkas, Email: skafkas@cu.edu.tr. References Associated Data Supplementary Materials Supplementary Material 1. Supplementary Material 2. Supplementary Material 3. Data Availability Statement The RNA-sequencing datasets generated and analysed during the current study will be deposited in the NCBI Sequence Read Archive (SRA) and made publicly available upon acceptance of the manuscript. The accession numbers will be provided before publication.