Scion-Rootstock interactions shape early transcriptional responses to controlled water deficit in newly selected table grape genotypes

Abstract

Water limitation represents a major constraint in viticulture, affecting both physiological and molecular responses in grapevines. This study evaluated the physiological responses and scion transcriptomic profiles of two newly selected table grape genotypes, Genusia and Maula, grafted onto two contrasting rootstocks, 1103P (V. berlandieri × V. rupestris) and SO4 (V. berlandieri × V. riparia), under controlled water deficit conditions. Physiological measurements, including stomatal conductance and stem water potential, were integrated with RNA sequencing analyses to evaluate differential gene expression in actively growing shoots. Results revealed that rootstock identity significantly modulated scion responses to water limitation. SO4-grafted plants displayed a higher number of differentially expressed genes and a broader transcriptional reprogramming compared to 1103P, indicating a stronger molecular response to reduced water availability. Functional enrichment analyses indicated that SO4 combinations preferentially activated stress-related and defense-associated pathways, whereas 1103P combinations maintained a more stable expression of photosynthetic and primary metabolic processes. Overall, these findings highlight the central role of rootstock–scion interaction in shaping early physiological and transcriptional adjustments to water deficit and provide valuable insights for optimizing grafting combinations in table grape breeding programs to cope with water deficit.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12870-026-08616-5.

Introduction

Grapevines (Vitis vinifera L.) are a primary global fruit crop, producing approximately 78 million tons annually (OIV Report, 2024 https://www.oiv.int/index.php/). However, productivity is increasingly threatened by water scarcity, particularly in Mediterranean regions where climate change intensifies drought [12, 24]. Understanding the complex physiological, molecular, and structural responses to reduced water availability is essential for maintaining the yield and quality of high-value crops like table grapes. Key indicators like stomatal conductance and water potential are essential for monitoring grapevine responses to water stress, as they directly influence photosynthesis and transpiration [11, 24, 50]. Grapevines respond to reduced water availability through a combination of physiological adjustments (e.g. regulation of stomatal conductance and plant water status), molecular responses, and genetically determined traits, which may also include anatomical modifications affecting water transport and retention during water deficit conditions, such as reduced leaf area, altered stomatal density and distribution, increased cuticle thickness, modifications of xylem vessel diameter and lignification, and changes in root system architecture affecting water uptake and transport [12, 29, 63]. These strategies depend on the coordinated interaction between scion genotype and rootstock identity, which integrates hydraulic traits and hormonal signaling [17, 40, 49], highlighting the importance of grafting in grapevine cultivation. Grapevine responses to water stress are influenced by environmental conditions, affecting stomatal activity, leaf water status, and resource partitioning [32, 59]. While genotypes are often categorized by their water-use strategies—ranging from conservative (isohydric) to risk-taking (anisohydric)—experimental evidence suggests these patterns are rarely fixed. Instead, stomatal regulation is modulated by the environment and the scion–rootstock interaction, which shapes water absorption and hormonal signaling [9, 10, 12, 22, 25, 28]. Abscisic acid (ABA) is a primary physiological responder to water deficit, regulating stomatal closure to protect xylem functionality and water-use efficiency [33, 46, 63]. However, prolonged ABA signaling can limit carbon fixation, vegetative growth, and fruit maturation [57]. Water deficit also reshapes metabolism, specifically the phenylpropanoid pathways and antioxidant responses [60]. Transcriptomic advances have identified thousands of differentially expressed genes (DEGs) related to ABA signaling, osmotic adjustment, and ROS detoxification [9, 22]. Key molecular pathways, such as phenylpropanoid biosynthesis involving enzymes like PAL, C4H, and 4CL, are upregulated in tolerant genotypes to reinforce cell walls and scavenge ROS [18]. Additionally, ABA-related regulators (e.g., SnRK2, PP2C) and osmolytes such as proline and raffinose help maintain turgor and protect cellular structures [56, 68]. Genes involved in sugar metabolism and osmotic adjustment are often differentially regulated during water deficit, contributing to the maintenance of cellular homeostasis in tolerant genotypes [8, 14, 16, 20, 21]. In contrast to extreme drought conditions often investigated in experimental studies, Mediterranean table grape systems more frequently experience moderate and transient water deficits. Understanding responses under these agronomically relevant conditions is essential for sustainable management. Research shows that rootstocks significantly influence scion stomatal regulation, hydraulics, and gene expression [17, 19, 45, 49, 71]. However, most studies focus on established cultivars, leaving a gap in knowledge regarding newly selected table grape genotypes [48, 61]. Moreover, evaluating scion–rootstock combinations at early selection stages is critical, as fruit quality is highly sensitive to stress.

In this study, we evaluated physiological traits and scion transcriptomic profiles under controlled water limitation in two newly selected and contrasting table grape genotypes, Maula and Genusia, grafted onto the widely used rootstocks 1103 Paulsen and SO4, which differ in vigor and water-use-related traits [19, 45] enabling a comparative analysis of scion–rootstock combinations.

This experimental framework allowed us to investigate how different scion–rootstock combinations modulate early physiological and transcriptional responses to reduced irrigation. By focusing on scion responses across rootstock–scion combinations, this work aims to identify grafting combinations potentially suitable for water-limited conditions and relevant for selection-oriented breeding programs.

Material and methods

Plant materials

Grapevine plants used in this study consisted of newly selected table grapevine genotypes, Maula and Genusia, obtained from breeding programs at CREA-Viticulture and Enology research center. The two scion genotypes selected for this study differ markedly in their phenological behavior and irrigation requirements. ‘Maula’ is an early-ripening cultivar harvested at the end of June and typically requires modest irrigation inputs until maturation, whereas ‘Genusia’ is a medium-late cultivar harvested between late September and October and generally requires sustained water supply during berry development. The inclusion of these contrasting phenological types allowed the evaluation of rootstock–scion interactions under reduced irrigation across distinct developmental and water-demand scenarios. All plants used in this study were heterografted, onto 1103P and SO4 rootstocks. Autografted or own-rooted plants were not included, as the experimental design was specifically conceived to evaluate rootstock–scion interactions relevant for applied breeding and genotype selection programs.

Experimental design to evaluate water deficit responses

In table grape production systems, unlike wine grapes, severe water stress is generally avoided because it may compromise berry size, quality, and marketability [35]. For this reason, the study was designed to assess plant responses under moderate and controlled water limitation, reflecting irrigation regimes realistically applicable under field conditions. The aim was not to induce severe drought, but to evaluate how different scion–rootstock combinations respond to reduced water supply within agronomically acceptable thresholds. The water‑stress treatment therefore consisted of a controlled mild deficit, imposed through reduced irrigation to trigger early physiological responses without reaching severe or canonical drought levels. This approach is consistent with previous studies examining rootstock‑mediated modulation of scion physiological and molecular responses under water stress [9]. The trial was conducted from 1 to 14 September 2023 in Turi (Southern Italy; 40°55′05″ N). Plants were grown individually in 5‑L pots inside a screenhouse, using a homogeneous peat–perlite–sand substrate that ensured adequate root development while allowing controlled manipulation of soil water availability. Water deficit was imposed by providing 1 L of water per week to stressed plants, whereas controls were irrigated to maintain soil moisture near field capacity. Rather than applying a fixed duration or predetermined stem water potential thresholds, stress progression was monitored by comparing reduced‑irrigation plants with their well‑watered counterparts. The trial was stopped once at least one grafting combination showed a statistically significant reduction in stomatal conductance and/or stem water potential with respect to the corresponding control plants. All physiological measurements and subsequent sampling were performed simultaneously across combinations, ensuring exposure to same environmental and irrigation conditions. This design allowed assessment of scion–rootstock responses under comparable stress progression without driving plants into severe drought. The water‑stress phase lasted approximately two weeks, during which physiological parameters were repeatedly monitored. Because the objective was to characterize early responses under moderate, agronomically relevant water limitation rather than post‑stress recovery, no rehydration cycle was included. During the experiment, mean air temperature was 24.0 °C (range 18.9–29.1 °C), and mean relative humidity averaged 60% (48–71%) (complete data available at https://protezionecivile.regione.puglia.it/bollettini-meteorologici-regionali-mensili). No rainfall occurred. Solar irradiance data from the ENEA SolarItaly database indicated a mean daily Global Horizontal Irradiance (GHI) of 4.784 kWh m⁻² day⁻¹ for September.

Physiological measurements

Stomatal conductance was measured using a portable gas exchange system Decagon SC-1 Leaf Porometer (METER, München,Germany) on three leaves per plant, selected from the same biological replicates used for transcriptomic analyses. Measurements were performed on five biological replicates per grafting combination and treatment. Stem water potential was measured on the same plants used for stomatal conductance assessments using a Scholander-type pressure chamber, following the methodology described by Turner et al. [65]. All physiological measurements were conducted during the central hours of the day (between 11am and 1 pm), when diurnal fluctuations in gas exchange and plant water status are minimized and physiological parameters are representative of stress conditions. This measurement window is consistent with established protocols for table grapevine water stress studies [9]. Whole-plant water loss or cumulative transpiration was not directly quantified. However, the combination of controlled irrigation regimes, repeated measurements of stomatal conductance, and stem water potential provided a consistent physiological framework to define and compare plant water status across treatments. Finally, to confirm that our irrigation strategy did not cause permanent damage to the plants, fresh root weight was measured during the following winter season to verify that root development was not affected by the water treatment. The weight of the roots was measured after washing the root system to remove any soil residue.

Sampling for transcriptomic analyses

Sampling for transcriptomic analyses was performed at the end of the water deficit imposition, once plants had reached a moderate and physiologically significant stress level, as defined accordingly with previous reported strategy. For transcriptomic assays all material collections were performed within the same time window to minimize variability related to diurnal fluctuations. Actively growing shoots were collected from three out of five of the same plants used for physiological measurements. This sampling strategy was intentionally adopted to investigate early molecular responses to water stress in developing tissues, accordingly to [9].

Statistical analysis

Data normality and homogeneity of variances were preliminarily assessed using Shapiro–Wilk’s and Levene’s tests, respectively. Assumptions were met and no data transformation was required. A three-way ANOVA was performed to evaluate the effects of Scion, Rootstock, and Water Condition, as well as all possible interactions, on physiological parameters. Post-hoc comparisons of estimated marginal means were conducted using the emmeans package in R (v. 4.2.2), and simple-effect contrasts were performed when significant interactions were detected.

This factorial design allowed the separation of genetic (scion, rootstock) and environmental (condition) effects and enabled the detection of context-dependent responses emerging from their interaction.

RNA extraction and sequencing

RNA extraction was performed as reported by Prencipe et al. [55]. Briefly, the sorbitol pre-wash step was performed with a buffer containing Tris–HCl, sorbitol, EDTA, PVP-40, and β-mercaptoethanol to remove polysaccharides and polyphenols from macerated plant material. Samples were vortexed, centrifuged at 2500 × g, and the supernatant was discarded. RNA was then extracted using the Norgen Plant/Fungi Total RNA Purification Kit. RNA purity and yield were calculated using a Qubit 4 Fluorometer (Thermo Fisher Scientific, Waltham, MA, USA), while the 260/280 ratio was calculated using a NanoDrop 2000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). RNA integrity was assessed using the Agilent 2100 Bioanalyzer Instrument (Agilent, Santa Clara, CA, USA) and RIN values were calculated. RNA libraries were prepared using the Illumina TruSeq Stranded mRNA Library Prep Kit and sequenced by Novogene (Cambridge, UK) on the Illumina NovaSeq X PE150 platform. Given the sequencing depth, data quality, and use of a standardized analytical pipeline, transcriptomic profiling was considered robust for comparative analyses. Root tissues were not analyzed, as the aim of the study was to characterize scion-level physiological and transcriptional responses relevant for genotype evaluation rather than to dissect the molecular origin of rootstock-derived signals.

Differential gene expression analysis and function enrichment analysis

V. vinifera PN40024_v4 was used as a reference for read mapping and gene annotation, directly downloaded from the ENSEMBL repository (en-sembl_56_vitis_vinifera_pn40024_v4_toplevel). Further, the gene annotation files were downloaded from the Novogene genome website directly (https://www.novogene.com/us-en/Reference/ReferenceForSpecies/index.html). The index of the reference genome was built using Hisat2 v2.0.5 and paired-end clean reads were aligned to the reference genome using Hisat2 v2.0.5. We selected Hisat2 [47] as the mapping tool because Hisat2 can generate a database of splice junctions based on the gene model annotation file and, thus, a better mapping result than other non-splice mapping tools. featureCounts v1.5.0-p3 [41] was used to count the read numbers mapped to each gene. And then FPKM of each gene was calculated based on the length of the gene and reads count mapped to this gene. Differential expression analysis was performed using DeSeq2 [2]. DESeq2 provides statistical routines for determining differential expression in digital gene expression data using a model based on the negative binomial distribution. The resulting p-values were adjusted using Benjamini and Hochberg’s approach for controlling the false discovery rate. Genes with an adjusted p-value < = 0.05 found by DESeq2 were assigned as differentially expressed.

Gene Ontology (GO) enrichment analysis and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were performed to identify functional and pathway level associations for differentially expressed genes derived from RNA sequencing data. Differential expressions were determined by comparing transcriptomic profiles of samples from the two genotypes grafted on 1103P, or SO4 rootstock at water deficit or control (Ctr) condition using a false discovery rate (FDR) threshold of < 0.05 and a log2 fold-change cutoff of ± 1. For GO enrichment analysis, DEGs were categorized into biological processes (BP), molecular functions (MF), and cellular components (CC) using the Gene Ontology database. Enrichment analysis was conducted using Fisher’s exact test with FDR correction, and significantly enriched GO terms were identified based on a corrected p-value < 0.05, focusing particularly on BP.

For KEGG pathway enrichment analysis, DEGs were mapped to pathways in the KEGG database using the KEGG Automatic Annotation Server (KAAS). Enriched pathways were identified by comparing observed gene counts in each pathway to an expected distribution using a hypergeometric test. Pathways with an FDR-adjusted p-value < 0.05 were considered significantly enriched. Visualization of enriched GO terms and KEGG pathways was performed using R packages (v. 4.2.2) (“R Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2016; Available online: http://www.R-project.org/,” [58]), including clusterProfiler [70] and ggplot2 [69], to provide an integrated view of genotype- and time-dependent transcriptional responses.

Results

Physiological responses to water stress

Physiological measurements were performed to monitor plant water status during the controlled water deficit (hereafter referred to as WS) treatment and to define comparable stress conditions across genotypes and grafting combinations. The irrigation regime was intentionally designed to induce an early and controlled deviation from well-watered conditions rather than to reach severe drought thresholds commonly reported for potted grapevines [23, 29, 43, 44]. After approximately two weeks, and based on the comparative assessment of plant water status relative to controls, the treatment was stopped once significant differences were detected, maintaining a moderate and agronomically relevant level of WS. Three-way ANOVA (Table 1) revealed that physiological responses were modulated in a genotype- and context-dependent manner. For stomatal conductance (gs), Rootstock showed a significant main effect (F = 9.976, p = 0.003), while Scion and Condition alone were not independently significant. However, a significant Rootstock × Condition interaction (F = 9.872, p = 0.004) and a significant three-way interaction Scion × Rootstock × Condition (F = 4.751, p = 0.037) were detected, indicating that stomatal response to reduced irrigation depended on the specific scion–rootstock combination rather than on water limitation alone.

Table 1.

Statistical analysis of main effect of scion, rootstock and condition, and their interaction on physiological parameters

Effect	DF	gs		Ψ		Plant height		Root weight
		F-value	p < 0.05	F-value	p < 0.05	F-value	p < 0.05	F-value	p < 0.05
Scion	1	1.589	0.217	16.643	0.0003*	2.337	0.136	3.015	0.092
Rootstock	1	9.976	0.003*	1.473	0.234	2.337	0.136	0.297	0.589
Condition	1	3.700	0.063	16.417	0.0003*	16.440	0.0003*	0.623	0.436
Scion*Rootstock	1	2.460	0.127	0.096	0.759	0.059	0.8090	5.060	0.031*
Scion*Condition	1	1.569	0.219	1.225	0.277	2.475	0.1260	0.068	0.797
Rootstock*Condition	1	9.872	0.004*	8.429	0.007*	2.475	0.1260	0.033	0.858
Scion*Rootstock*Condition	1	4.751	0.037*	0.662	0.422	2.475	0.1260	0.190	0.666

* next to p<0.05 indicates significant values

Simple-effect comparisons within each Scion × Rootstock combination confirmed that Maula and Genusia grafted onto SO4 exhibited significantly lower gs under WS (198.33 and 167.50 mmol m⁻² s⁻¹, respectively) compared with their controls (Fig. 1A), together with reduced stem water potential (− 0.63 and − 0.93 MPa, respectively) (Fig. 1B). In contrast, Genusia grafted onto 1103P maintained comparatively higher gs values under WS (410.80 mmol m⁻² s⁻¹), consistent with the moderating effect of this rootstock on stomatal regulation. Absolute stem water potential (Ψ) values remained above − 1.0 MPa across combinations, confirming an early physiological response to reduced irrigation relative to well-watered controls rather than severe water deficit conditions. For Ψ, significant main effects of Scion (F = 16.643, p = 0.0003) and Condition (F = 1.417, p = 0.0003) were observed, together with a significant Rootstock × Condition interaction (F = 8.429, p = 0.007). These results indicate that plant water status was influenced both by genotype and by its interaction with irrigation regime, further supporting differential sensitivity among grafting combinations. Plant height was measured as a supplementary indicator of short-term growth response, acknowledging the limited duration of the experiment. Only the main effect of Condition was significant (F = 16.440, p = 0.0003), with a modest reduction observed exclusively in Genusia/SO4 (Fig. 1C). Considering the two-week timeframe, this result should be interpreted as an early trend rather than as evidence of pronounced growth inhibition. Finally, root weight was significantly affected only by the Scion × Rootstock interaction (F = 5.060, p = 0.031), while Condition showed no significant effect (Fig. 1D). This suggests that root growth differences primarily reflected intrinsic genetic background rather than short-term water limitation, consistent with the moderate intensity and limited duration of the imposed water reduction. Overall, physiological data indicate that the applied treatment elicited measurable physiological responses to reduced irrigation compared with well-watered controls. Importantly, significant interaction terms for gs and Ψ indicate that stress perception and regulation emerged from the combined effect of scion genotype, rootstock identity, and irrigation regime. This genotype-dependent modulation of early physiological responses provides a coherent framework for interpreting the differential transcriptomic reprogramming observed among grafting combinations.

Fig. 1.

Effects of water stress on physiological parameters of Genusia and Maula cultivars grafted on 1103P and SO4 rootstocks. A Stomatal conductance; B Stem water potential; C Plant height; D Root weight. Estimated marginal means were computed using the emmeans test and pairwise comparisons between condition levels were carried out within each Scion-Rootstock combination. * = p < 0.05. ** = p < 0.01; *** = p < 0.001

Transcriptomic analysis of grapevine shoots under water stress

To investigate the molecular responses to WS in relation to scion genotype and rootstock genotype, RNA-seq was performed. A total of 24 cDNA libraries were generated and sequenced, yielding an average of 56.7 million raw reads per library. Rigorous quality control was applied to the sequencing data, resulting in 95.7% of the reads classified as clean, with a base error rate of only 0.01%. Key sequencing metrics, including total raw reads, clean reads, GC content, and read quality scores (Q20 and Q30), were assessed across all 24 samples. Q20 scores ranged from 96.66% to 97.24%, while Q30 scores ranged from 90.40% to 93.52%, confirming the high accuracy and reliability of the sequencing data (Supplementary Table 1). Reads were aligned to the Vitis vinifera reference genome (PN40024_v4) and the expression level for each gene, in each sample, was determined (Supplementary Table 2). Based on the gene expression profile in each sample, Principal Component Analysis (PCA) was further performed, revealing a good grouping of replicates confirming data reliability, although higher variability was observed in some combinations with SO4 rootstock. The distribution of samples suggests both genotype and treatment-influence. Dim2 mainly separated the two scion genotypes, with Genusia samples positioned in the upper quadrants and Maula samples in the lower quadrants. Dim1 contributed to the separation of several WS-treated samples toward positive values. In particular, Maula WS samples clustered in the lower right quadrant irrespective of rootstock, whereas Genusia/SO4 WS samples were located in the upper right quadrant, appearing more distant from the remaining combinations (Supplementary Fig. 1).

Differential gene expression analysis and comparison across scion/rootstock combinations

DEGs across scion-rootstock combinations under WS in comparison with Ctr were determined. Genusia grafted onto SO4 exhibited 7334 DEGs (3584 up-regulated, 3750 down-regulated), significantly more than its 1103P counterpart (2750 DEGs: 1305 up, 1445 down), suggesting a stronger transcriptional reprogramming when SO4 was used as rootstock (Fig. 2A and Supplementary Table 3). Similarly, Maula on SO4 had 7594 DEGs (3589 up, 4015 down) compared to 3203 DEGs on 1103P (1649 up, 1554 down), reinforcing SO4’s influence on gene expression changes under WS (Fig. 2A and Supplementary Table 3). To evaluate the effect of rootstock identity on scion transcriptome under WS, we compared gene expression in scions grafted onto 1103P versus those grafted onto SO4. Under WS, Genusia showed 5692 DEGs (3024 up, 2668 down) between 1103P- and SO4-grafted plants, whereas only 522 DEGs (192 up, 330 down) were detected in Maula, indicating that the Genusia scion transcriptome was more strongly modulated by rootstock identity (Fig. 2B and Supplementary Table 3). When comparing scion genotypes under the same rootstock, 5592 DEGs were detected in Maula vs Genusia on 1103P, while 6833 DEGs were identified in Maula vs Genusia on SO4. (Fig. 2C and Supplementary Table 3).

Fig. 2.

UpSet plot of DEGs across scion–rootstock combinations in WS and Ctr conditions. Total number of DEGs for each comparison is shown by the horizontal bars on the left; intersection sizes among comparisons are shown by the vertical bars on the right. A Comparison between rootstocks, DEGs in 1103P with respect to SO4 at Ctr and under WS across scions are reported; B Comparison between conditions: DEGs in WS with respect to Ctr condition in each scion/rootstock combination are reported; C Comparison between scions: DEGs in Maula with respect to Genusia at Ctr and under WS across rootstocks are reported. Ma = Maula; Ge = Genusia; WS = Water stress; Ctr = Control

To identify unique and shared DEGs across different scion-rootstock combinations under WS and Ctr conditions, detailed comparisons were performed to go insight into scion-specific, rootstock-dependent, and conserved transcriptional responses.

Comparing genes that are DE in Maula with respect to Genusia under WS across rootstocks, 3465 were unique to SO4 and 2078 to 1103P, showing rootstock-driven transcriptional shifts specific for each scion. Meanwhile, 1355 genes were commonly regulated between scions, across the two rootstocks, highlighting genes DE in response to drought influenced only by scion genotype. By contrast, under control conditions, only 207 DEGs were conserved between scion genotypes across both rootstocks, suggesting that WS induces stronger scion-specific transcriptional reprogramming compared to non-stress conditions (Fig. 2C and Supplementary Table 3). Further comparison (Fig. 2B and Supplementary Table 3) showed that the number of DEGs in response to WS was higher in SO4-grafted combinations than in 1103P-grafted ones for both scions. Specifically, 2908 and 2768 DEGs were identified in Maula/SO4 and Genusia/SO4, respectively, compared with 650 and 987 DEGs in Maula/1103P and Genusia/1103P.

A core set of 367 DEGs was shared across all WS comparisons, representing a common transcriptional response to water deficit. Additionally, 65 DEGs were shared between Maula and Genusia under 1103P, whereas 1952 DEGs were common to both scions under SO4, indicating a stronger overlap of WS-responsive genes in SO4-supported scions.

Moreover, 497 and 727 DEGs were unique to Genusia and Maula, respectively, across rootstocks, reflecting scion-specific transcriptional components (Fig. 2B and Supplementary Table 3). To assess the effect of rootstock identity on scion transcriptome, gene expression profiles of scions grafted onto 1103P and SO4 were compared under both WS and control conditions (Fig. 2A and Supplementary Table 3).

Under WS, 5106 DEGs were detected between 1103P- and SO4-grafted Genusia plants, whereas only 267 DEGs were identified in Maula. Under control conditions, 423 and 403 DEGs were detected in Genusia and Maula, respectively, indicating baseline rootstock-dependent transcriptional differences.

A total of 212 DEGs were shared between Genusia and Maula under WS, representing common rootstock-modulated transcriptional components, while only 25 DEGs were conserved across scions under control conditions.

Gene ontology enrichment analysis

GO enrichment analysis was conducted to identify significantly enriched Biological Processes (BP) under WS in both Genusia and Maula cultivars grafted onto 1103P and SO4 rootstocks (Supplementary Data 1; Supplementary Fig. 2 A and B). If 1103P was considered, for down-regulated genes Genusia were significantly enriched in BP categories related to cell cycle regulation, DNA replication, chromosome segregation, and cytoskeletal organization (Supplementary Fig. 2 A). In Maula grafted onto 1103P, enriched down-regulated processes additionally included phenylpropanoid metabolism, lignin biosynthesis, and ribosome-related processes (Supplementary Fig. 2B). For up-regulated genes, Genusia on 1103P exhibited significant enrichment in processes linked to stress adaptation, including oxidative stress responses, osmotic balance, and hormone signaling (Supplementary Fig. 2 A). In contrast, Maula on 1103P displayed enrichment in processes related to protein stabilization, ubiquitination, and cell redox homeostasis (Supplementary Fig. 2B).

On SO4 rootstock on the contrary, both cultivars exhibited a broader spectrum of enriched down-regulated BP. Genusia showed a more pronounced enrichment in ribosomal biogenesis and cytoskeletal components compared to 1103P (Supplementary Fig. 3 A), whereas Maula displayed enrichment in DNA repair- and stress-response-related processes suggesting a stress avoidance strategy rather than active resilience (Supplementary Fig. 3B). Similarly, both cultivars grafted onto SO4 showed enrichment of up-regulated BP related to ABA-responsive processes, protein turnover, and stress-related signaling pathways (Supplementary Fig. 3 A and B).

Further comparisons between different scion-rootsock combinations under WS highlighted scion-dependent enrichment patterns. Genusia predominantly showed modulation of genes belonging to BP associated with secondary metabolism and cell wall biosynthesis (Supplementary Fig. 4 A), whereas Maula displayed stronger enrichment in processes linked to redox balance and cellular transport pathways (Supplementary Fig. 4B).

KEGG pathway enrichment analysis

KEGG enrichment analysis was performed to identify significantly over-represented functional pathways among differentially expressed genes (DEGs) in each scion–rootstock combination under WS in the Genusia and Maula cultivars grafted onto 1103P and SO4 rootstocks (Supplementary Data 2). KEGG enrichment analysis identified biological pathways that are significantly over-represented among differentially expressed genes, providing a pathway-level view of transcriptional modulation under water stress. Overall, the transcriptional patterns observed at the pathway level were found consistent with the early and moderate physiological response described above, suggesting that gene expression reprogramming occurs even in the absence of severe water deficit. For down-regulated genes (Fig. 3A, B, C and D), Genusia on 1103P showed significant enrichment of pathways related to DNA replication (vvi03030), mismatch repair (vvi03430), and ribosome function (vvi03010) (Fig. 3A). This transcriptional profile indicates a reduction in genes associated with cell cycle progression and protein synthesis, consistent with a transient slowing of growth-related processes under reduced irrigation. Maula on 1103P displayed a similar enrichment in DNA replication and repair pathways, together with oxidative phosphorylation (vvi00190) and lysine biosynthesis (vvi00300) (Fig. 3B). Rather than indicating confirmed metabolic reprogramming, this pattern reflects coordinated transcriptional modulation of genes associated with energy metabolism and amino acid biosynthesis during early stress perception. On SO4 rootstock, a broader enrichment of down-regulated pathways was observed. Genusia displayed downregulation in nucleotide metabolism (vvi01250), biotin metabolism (vvi00780), and phagosome-related pathways (vvi04145) (Fig. 3C). Maula on SO4 exhibited additional enrichment in citrate cycle (vvi00020) and galactose metabolism (vvi00052) (Fig. 3D). Given that physiological measurements indicated mild-to-moderate stress levels, these results likely represent early transcriptional adjustments rather than severe metabolic impairment. The down-regulation of pathways associated with DNA replication and ribosome function is coherent with a transient adjustment of growth-related processes under early water limitation, rather than with severe metabolic disruption.

Fig. 3.

KEGG dot plots showing metabolic pathway analysis of down-expressed genes. A Genusia cultivar on the 1103P rootstock in WS condition compared to Ctr condition; B Maula cultivar on the 1103P rootstock in WS condition compared with Ctr condition; C Genusia cultivar on the SO4 rootstock in WS condition compared to Ctr condition; D Maula cultivar on the SO4 rootstock, in WS condition compared to Ctr condition

For up-regulated genes (Fig. 4A, B, C and D), Genusia grafted onto 1103P showed enrichment of carbon fixation (vvi00710) and photosynthesis-antenna proteins (vvi00196) (Fig. 4A). Of note, in combinations where stomatal conductance was comparatively maintained under reduced irrigation, enrichment of photosynthesis-related pathways reflects coordinated transcriptional regulation aligned with the observed physiological behavior. Maula on 1103P displayed enrichment of MAPK signaling (vvi04016) and plant hormone signal transduction (vvi04075) pathways (Fig. 4B), indicating activation of stress-related signaling components. Importantly, enrichment of hormone-related pathways reflects transcriptional activation of signaling modules, not direct measurement of hormone concentrations.

Fig. 4.

KEGGdot plots showing metabolic pathway analysis of up-expressed genes. A Genusia cultivar on the 1103P rootstock in WS condition compared to Ctr condition; B Maula cultivar on the 1103P rootstock in WS condition compared with Ctr condition; C Genusia cultivar on the SO4 rootstock in WS condition compared to Ctr; D Maula cultivar on the SO4 rootstock, in WS condition compared to Ctr condition

On SO4, both cultivars both cultivars showed enrichment of ABA-related and osmotic adjustment-associated pathways. Genusia exhibited enrichment in galactose metabolism (vvi00052) (Fig. 4C), while Maula displayed enrichment in circadian rhythm regulation (vvi04712) (Fig. 4D). These patterns suggest scion-specific transcriptional responsiveness under reduced irrigation. Moreover, the broader pathway enrichment observed in SO4-grafted plants parallels the stronger physiological modulation detected in this rootstock background, indicating a more pronounced transcriptional responsiveness under the same irrigation regime.

Finally, direct comparison between cultivars under WS (Fig. 5A–D) revealed that Maula exhibited stronger down-regulation of secondary metabolism-related pathways (vvi00940) on 1103P (Fig. 5A), whereas on SO4 it showed greater enrichment in down-regulated carbohydrate metabolism pathways (vvi00052) (Fig. 5B). In contrast, Genusia maintained enrichment of photosynthesis- and signaling-related pathways across both rootstocks (Fig. 5C and D). Overall, KEGG analysis indicates that transcriptional reprogramming under moderate water limitation is strongly dependent on the scion–rootstock combination. Importantly, given the mild intensity of the imposed stress, these pathway-level differences likely reflect early regulatory adjustments rather than late-stage drought acclimation.

Fig. 5.

KEGG dot plots showing metabolic pathway analysis of down- and up-expressed genes in the comparison between the two cultivars. A Pathways of down-regulated genes in Maula with respect Genusia under WS condition, both on the 1103P rootstock; B Pathways of down-regulated genes in Maula with respect Genusia under WS condition, both on the SO4 rootstock; C Pathways of up-regulated genes in Maula with respect Genusia under WS condition, both on the 1103P rootstock; D Pathways of up-regulated genes in Maula with respect Genusia under WS condition, both on the SO4 rootstock

Overall, the pathway-level reorganization observed across graft combinations mirrors the genotype-dependent physiological modulation detected under reduced irrigation, supporting the interpretation that scion–rootstock interactions shape early adaptive responses to water limitation.

Integrated analysis of candidate DEGs in scion- and rootstock-dependent responses to water stress

KEGG and GO enrichment analyses identified six key pathways: Hormone Signaling (vvi04075) and MAPK mediated signaling (vvi04016), central to hormonal control of WS adaptation; Phenylpropanoid Biosynthesis (vvi00940) related to cell wall-associated and redox processes; Galactose Metabolism (vvi00052) including genes involved in sugar metabolism potentially linked to osmotic regulation; amino acid degradation (vvi00280) and cytoskeletal remodeling (vvi04814). We therefore focused on DEGs in these pathways to identify key players in the transcriptional reprogramming associated with water stress–induced transcriptional adjustments across the four scion/rootstock combinations (Maula/SO4, Maula/1103P, Genusia/SO4, and Genusia/1103P) (Supplementary Table 4).

In plant hormone signaling (vvi04075)

Key genes were found related to ABA signaling. ABA-related genes (PP2C: Vitvi06g00533, Vitvi09g00156, Vitvi11g01392) were upregulated in SO4-grafted plants and Maula/1103P, while Genusia/1103P showed weaker induction. ABF (Vitvi03g01574) was upregulated in all combinations, strongest in Genusia/SO4, except Vitvi04g01505, which was downregulated in Genusia (both rootstocks). Ethylene-related genes (ETR, EIN3) were upregulated in SO4-grafted plants, while jasmonic acid signaling components (JAZ, COI1) showed rootstock- and genotype-specific regulation. Auxin-responsive genes (AUX/IAA) exhibited contrasting expression patterns across combinations (Supplementary Fig. 5 and Supplementary Fig. 6).

Other key DEGs were found in MAPK Signaling (vvi04016)

RbohD (Vitvi19g00233) was consistently upregulated, strongest in Genusia/SO4. MAPK cascade genes (MPK3/6, MAPK17/18) were upregulated in Genusia/SO4 and Maula (both rootstocks). WRKY22/29 (Vitvi02g00039) was downregulated only in Maula/1103P, suggesting genotype-specific suppression of defense-related transcription factors. Several aquaporin genes (PIP, TIP families) were downregulated, except SIP2 (Vitvi08g00904), which was upregulated in Genusia/SO4.

Genes involved in lignin biosynthesis (Phenylpropanoid Biosynthesis vvi00940)

Showed differential regulation (Supplementary Table 4). Cinnamyl-alcohol dehydrogenases (Vitvi03g00564) were upregulated in Maula (both rootstocks) and Genusia/SO4, while Vitvi07g01673 and Vitvi03g00561 were specific to Genusia/SO4 and Maula/1103P, respectively. Vitvi14g01589 (4-coumarate-CoA ligase) was downregulated in Genusia/1103P but upregulated in Maula (both rootstocks). Several peroxidases, including Vitvi13g00143 and Vitvi05g00235, were upregulated in SO4-grafted plants, strongest in Genusia.

Different regulated genes related to sugar accumulation were also found (Galactose Metabolism, vvi00052)

3-alpha-galactosyltransferase (Vitvi01g00714) was upregulated across all combinations, particularly in SO4 and Genusia. Conversely, raffinose synthase (Vitvi19g00768) was consistently downregulated. Beta-fructofuranosidase (Vitvi00g04570) was upregulated in SO4 plants, while other isoforms (Vitvi08g01434, Vitvi09g00193) showed contrasting regulation in Genusia/1103P and Maula/SO4. Invertase (Vitvi16g00713) was downregulated only in Genusia/SO4.

Amino acid and fatty acid metabolism (vvi00280)

Genes involved in valine, leucine, and isoleucine degradation were mostly upregulated in Genusia/SO4 and Maula (both rootstocks), highlighting a transcriptional modulation of alternative respiratory pathways (Supplementary Table 4). Dihydrolipoyl dehydrogenase (Vitvi14g00104) was upregulated in Maula/SO4 but downregulated in Genusia/1103P.

Motor proteins (vvi04814)

Tubulin (TUBB) isoforms (Vitvi06g00569, Vitvi14g01929) were downregulated in Genusia/SO4, except Vitvi01g04308, which was upregulated. Kinesins (KIF15, KIF2/3, KIF10) were strongly downregulated in SO4, with some isoforms showing rootstock-specific regulation (Vitvi17g00321, Vitvi18g01055). Dyneins (DYNLL) were mostly upregulated, except for Vitvi09g00355, which was downregulated in Genusia/1103P.

Discussion

Deciphering the molecular and physiological bases of grapevine responses to water deficit is fundamental for the identification of tolerant genotypes. Both scion genotype and rootstock identity are known to influence plant physiological behavior under limited water availability [17, 49]. The present study investigated transcriptional responses of two newly selected table grape genotypes grafted onto two contrasting rootstocks under a controlled reduction in water supply. The treatment was designed to induce a mild and controlled water deficit rather than severe drought, reflecting agronomically relevant conditions for table grape production where excessive stress may compromise fruit quality. Previous studies have shown that genotype-dependent differences in grapevine responses often emerge during early stress perception, when transcriptional adjustments precede major physiological changes [9]. Accordingly, transcriptomic analyses were conducted on actively growing shoot apices, a highly responsive tissue allowing detection of early regulatory differences among scion genotypes and rootstock–scion combinations under controlled water limitation. Physiological measurements revealed significant differences in stomatal conductance and stem water potential between control and water-deficit plants (Fig. 1,Table 1). Although absolute values did not reach the typical thresholds for moderate or severe drought in potted grapevines [23, 43, 44], the consistent reductions in gs and Ψ in specific graft combinations indicate that plants clearly perceived and responded to reduced water availability. Assimilation rates and intrinsic water use efficiency (iWUE were not measured; therefore, carbon assimilation dynamics were not directly assessed. Our analysis instead focused on transcriptional differences among scion–rootstock combinations under controlled water limitation. The occasional overlap between control and water-deficit plants likely reflects intrinsic genotype-dependent variability, underscoring the relevance of scion × rootstock interactions [18, 29]. The significant scion × rootstock interaction identified by ANOVA indicates that physiological responses to reduced irrigation are combination-dependent, reinforcing the concept of grafted plants as integrated systems rather than independent genotypes [17, 49]. This has practical implications for vineyard management under water-limited conditions [3, 4, 12]. Maula grafted onto SO4 exhibited a stronger reduction in stomatal conductance under reduced irrigation, indicating a more pronounced stomatal response compared with the other combinations. Although carbon assimilation was not measured, this pattern supports differential regulation of stomatal behavior in this combination, whereas the more gradual modulation observed in Genusia may reflect an alternative regulatory balance.

Overall, water deficit effects were more pronounced in vines grafted onto SO4, whereas 1103P was associated with comparatively moderated physiological adjustments under the same irrigation regime, consistent with its widely reported drought-resilient behavior [5, 14, 28, 32, 38, 52, 53]. These genotype- and rootstock-dependent differences highlight the importance of evaluating scion–rootstock combinations when selecting cultivars for specific irrigation regimes and environmental conditions [40, 53, 64].

Our data indicate that SO4-mediated responses involve stronger stomatal regulation and enhanced activation of ABA-related signaling, whereas 1103P supports a comparatively moderated transcriptional adjustment under controlled water limitation. Rather than reflecting contrasting drought strategies, these differences likely represent rootstock-dependent modulation of early stress perception and signaling sensitivity. The divergence in stomatal behavior may be linked to previously described differences in hydraulic and hormonal regulation [28, 64], although these parameters were not directly measured. Overall, these findings support distinct regulatory modes of water-deficit responsiveness among rootstocks. Transcriptomic analysis revealed clear differences among combinations in the magnitude of DEGs (Fig. 2A–C; Supplementary Table 3). Across both scion genotypes, SO4 triggered substantially broader transcriptional reprogramming than 1103P under the same irrigation regime, indicating a strong rootstock influence on scion molecular responses to water limitation [9, 18, 71]. Notably, Genusia displayed a higher number of DEGs in response to rootstock variation than Maula, suggesting genotype-specific regulatory plasticity (Fig. 2A and Supplementary Table 3).

GO and KEGG enrichment allowed us to identify key pathways modulated by WS across different scion-rootstock combinations responsible of the mentioned differences (Fig. 6).

Fig. 6.

Circos diagram of DEGs in scion- and rootstock-dependent responses to water stress. The figure shows key metabolic pathways and biological processes enriched in DEGs under water stress across four grapevine scion/rootstock combinations: Maula/SO4, Maula/1103P, Genusia/SO4, and Genusia/1103P. Upregulated pathways are shown in red, while downregulated are colored in green, with lines connecting those common to different combinations

In SO4-grafted plants, particularly in Maula, we observed a pronounced downregulation of genes associated with cell cycle progression and oxidative phosphorylation (Figs. 3 and 4; Supplementary Data 2). Rather than indicating increased metabolic costs, this pattern likely reflects a transcriptional reallocation away from growth-associated processes under moderate water limitation. A similar reduction in cell cycle-related gene expression was observed in Genusia/SO4, supporting the idea that SO4 may induce a stronger early adjustment of growth-related pathways compared to 1103P. Comparable transcriptional modulation of cell cycle and energy-related genes under water limitation has been reported in grapevine [18, 42]. At the same time, SO4 consistently showed stronger activation of hormone- and MAPK-related signaling pathways (Figs. 3, 4 and 5; Supplementary Figs. 5 and 6), suggesting enhanced sensitivity in stress signal perception and amplification. This is in line with previous studies describing genotype-dependent modulation of stress-responsive cascades in grafted grapevine systems [9, 10, 13, 17, 18, 49, 63]. In contrast, 1103P-supported scions displayed a comparatively moderated transcriptional reprogramming, consistent with its widely reported drought-resilient physiological behavior [31, 31]; with several photosynthesis- and oxidative stress–related genes remained stable or were moderately upregulated (Figs. 3 and 4), reflecting a comparatively moderated transcriptional adjustment under controlled water limitation. Although assimilation rates were not measured, the transcriptional profile of 1103P-supported scions is consistent with previous reports describing distinct drought-responsive behaviors for this rootstock under water deficit [1, 15, 27, 30, 33, 34, 37, 39, 67]. Overall enrichment analysis highlighted that the interaction between scion and rootstock significantly influences grapevine molecular responsiveness through differential modulation of hormone signaling, metabolic regulation, and structural pathways, as previously demonstrated in grafted systems [9, 27, 37, 39, 68].

Among hormone signaling pathways (Supplementary Figs. 5 and 6), ABA signaling emerged as a central regulatory hub across combinations. Core components including PP2C (Vitvi06g00533), PYR/PYL (Vitvi16g01226), ABF (Vitvi03g01574), and JAZ (Vitvi01g02293; Genusia/SO4) were differentially regulated, particularly in SO4-rooted plants, supporting enhanced activation of ABA-dependent cascades [13, 28]. Coordinated regulation of downstream targets such as raffinose-related genes (Vitvi08g01890; vvi00052) and phenylpropanoid pathway genes (Vitvi02g00263; vvi00940) further indicates hormone-driven early responses to water limitation, consistent with the stronger stomatal responsiveness observed in SO4 combinations (Fig. 7A).

Fig. 7.

Expression levels of significant differentially expressed genes, along with corresponding stomatal behavior and water potential (Ψ) across different scion/rootstock combinations (Maula/SO4, Maula/1103P, Genusia/SO4, and Genusia/1103P) under water stress condition. Upregulated genes are shown on the left, while downregulated genes are shown on the right. A Genes involved in the plant hormone signaling pathway. ABA-related genes are highlighted in yellow, AUX/IAA-related genes in blue, and GH3-related genes in red. B Genes involved in the mitogen-activated protein kinase (MAPK) pathway. Aquaporin-related genes are highlighted in blue

Auxin-related pathways also showed rootstock-dependent modulation. In Genusia/SO4, GH3 (Vitvi01g01791), involved in auxin conjugation, was strongly downregulated (− 5.6292), whereas it was upregulated in Genusia/1103P (3.9623) (Supplementary Fig. 5). Similarly, selected AUX/IAA genes were downregulated in Genusia/SO4 and Genusia/1103P (Vitvi07g00521, Vitvi11g00301) but upregulated in Maula/SO4 (Vitvi05g00838, Vitvi14g00483, Vitvi11g00301), highlighting genotype- and rootstock-dependent divergence in auxin regulatory networks. Although auxin levels were not quantified, these patterns suggest modulation of hormonal balance rather than direct evidence of altered growth. Induction of ethylene-related genes (ETR, EIN3, ERF1/2) further supports activation of multiple hormone signaling cascades under water deficit [26, 67, 68], indicating coordinated hormonal integration in scion-specific responses. Beyond hormone signaling, combination-specific metabolic adjustments were observed. Genusia/SO4 showed enrichment of phenylpropanoid-related genes, whereas Maula/SO4 exhibited stronger modulation of carbohydrate-associated pathways (Figs. 3, 4 and 5,Supplementary Data 2), consistent with stress-related metabolic reorganization [18, 31]. Upregulation of respiratory burst oxidase homolog D (RboHD) in SO4-grafted plants suggests enhanced ROS-related signaling rather than direct oxidative damage [31, 36, 62]. In Genusia/1103P, aldehyde dehydrogenase genes (ALDH; Vitvi09g00078) were strongly upregulated, whereas reduced ALDH expression was observed in Maula/1103P (Vitvi01g00785), indicating combination-specific modulation of detoxification pathways. Finally, cytoskeletal organization (vvi04814) emerged as a potential component of early water-deficit responsiveness. In Genusia/1103P, downregulation of motor proteins (KIF15, KIF22) and aquaporins including PIP2-7-C3H (Vitvi03g00155) and TIP1-1-C3H (Vitvi08g01602) suggests coordinated transcriptional modulation of intracellular trafficking and membrane water transport (Fig. 7B). Conversely, Maula/SO4 exhibited upregulation of selected dynein genes alongside PIP1-2-ERF (Vitvi02g00310). These patterns align with previous reports linking cytoskeletal dynamics and aquaporin regulation in plant stress responses [6, 7, 39, 51, 54, 66].

This study focused on early transcriptional responses under controlled water deficit rather than severe drought tolerance. Hydraulic traits, whole-plant transpiration, root architecture, and metabolite profiles were not assessed and warrant future integrative investigation. Nonetheless, the experimental conditions allowed consistent comparison of physiological and molecular responses across genotypes. Overall, the results reveal clear differences in molecular responsiveness among scion–rootstock combinations, distinguishing combinations with stronger signaling amplification (e.g., SO4-mediated responses) from those showing more moderated transcriptional adjustment (e.g., 1103P-supported scions). These differences are particularly relevant for table grape systems, where optimizing performance under moderate irrigation constraints is a primary goal.

By identifying key regulatory pathways and candidate genes associated with early water-deficit sensitivity (Table 2; Supplementary Figs. 5 and 6), this study provides a molecular framework for optimizing scion–rootstock pairing strategies in breeding programs targeting sustainable irrigation management. Unlike most previous studies focusing on established cultivars, the present work evaluates newly selected table grape genotypes during early selection stages, providing information directly relevant for breeding programs where the choice of the most suitable scion–rootstock combination represents a critical step. The RNA-seq dataset generated here represents a valuable comparative resource for future integrative analyses combining physiological, hormonal, and metabolomic approaches to clarify the functional relevance of genotype-specific transcriptional plasticity. By highlighting the complex scion–rootstock interactions underlying water-deficit responses, including coordinated modulation of hormonal signaling, metabolic pathways, and cytoskeletal dynamics, our data reinforce the key role of rootstock selection in shaping grapevine performance under water-limited conditions [22, 28, 29] and support the development of molecular markers for breeding and rootstock selection under increasingly variable climatic scenarios.

Table 2.

Candidate genes selected from overall data analysis

KEGG pathway	Functional module	Gene ID	Gene name	Genusia/1103P	Genusia/SO4	Maula/1103P	Maula/SO4
vvi00052	Osmoprotectant/sugar metabolism	Vitvi08g01890	raffinose synthase		Up (2.3184)	Up (1.5349)	Up (2.1058)
vvi00052	Osmoprotectant/sugar metabolism	Vitvi05g00170	inositol 3-alpha-galactosyltransferase	Up (3.1927)	Up (13.9218)		Up (5.8764)
vvi00052	Osmoprotectant/sugar metabolism	Vitvi00g04570	beta-fructofuranosidase		Up (2.9081)		Up (1.8658)
vvi00280	Stress-associated primary metabolism	Vitvi01g00785	aldehyde dehydrogenase (NAD +)			Down (−2.9199)
vvi00280	Stress-associated primary metabolism	Vitvi01g00787	aldehyde dehydrogenase (NAD +)		Up (3.9774)		Up (1.4578)
vvi00280	Stress-associated primary metabolism	Vitvi04g01402	aldehyde dehydrogenase (NAD +)		Up (1.4632)	Up (1.2581)
vvi00280	Stress-associated primary metabolism	Vitvi09g00078	aldehyde dehydrogenase (NAD +)	Up (4.6066)
vvi00280	Stress-associated primary metabolism	Vitvi10g02269	acetyl‑CoA C‑acetyltransferase		Up (1.3247)	Up (0.7195)	Up (1.1675)
vvi00280	Stress-associated primary metabolism	Vitvi18g02625	amino acid aminotransferase		Up (2.6671)	Up (1.4179)	Up (2.5344)
vvi00280	Stress-associated primary metabolism	Vitvi06g01161	hydroxymethylglutaryl‑CoA lyase	Up (0.91096)	Up (1.1835)		Up (0.5623)
vvi00940	Phenylpropanoid/cell wall metabolism	Vitvi03g00564	cinnamyl‑alcohol dehydrogenase		Up (1.9576)	Up (1.4949)	Up (2.3501)
vvi00940	Phenylpropanoid/cell wall metabolism	Vitvi05g00235	peroxidase		Up (4.8749)	Up (1.9664)	Up (2.6512)
vvi00940	Phenylpropanoid/cell wall metabolism	Vitvi02g00263	caffeic acid O‑methyltransferase		Up (4.6960)	Up (1.5389)	Up (2.6314)
vvi04075	ABA core signaling	Vitvi03g01574	ABF	Up (0.9319)	Up (1.6046)	Up (1.3748)	Up (1.5623)
vvi04075	Auxin homeostasis/signaling	Vitvi05g00838	AUX/IAA		Down (−1.3613)		Up (1.2699)
vvi04075	Auxin homeostasis/signaling	Vitvi07g00521	AUX/IAA		Down (−1.5574)	Down (−0.8929)
vvi04075	Auxin homeostasis/signaling	Vitvi11g00301	AUX/IAA	Down (−0.8549)	Down (−1.4685)		Up (0.8821)
vvi04075	Auxin homeostasis/signaling	Vitvi14g00483	AUX/IAA		Down (−1.0655)		Up (1.1547)
vvi04075	Jasmonate signaling	Vitvi13g00512	COI1		Up (0.6057)	Up (0.6323)	Up (0.6646)
vvi04075	Ethylene signaling	Vitvi07g00359	ETR		Up (0.8013)		Up (0.7367)
vvi04075	Auxin homeostasis/signaling	Vitvi01g01791	GH3	Up (3.9623)	Down (−5.6292)		Up (3.4641)
vvi04075	Auxin homeostasis/signaling	Vitvi07g00293	GH3			Down (−0.8474)	Up (1.7563)
vvi04075	Jasmonate signaling	Vitvi01g02293	JAZ		Up (3.1665)
vvi04075	Jasmonate signaling	Vitvi12g02718	JAZ	Up (5.7340)
vvi04075	ABA core signaling	Vitvi06g00533	PP2C		Up (2.6199)	Up (2.2155)	Up (2.3546)
vvi04075	ABA core signaling	Vitvi16g01226	PYR/PIL	Up (0.6131)	Up (0.8354)	Up (0.4476)	Up (0.6695)
vvi04075/vvi4016	Ethylene signaling	Vitvi13g01126	EIN3		Up (0.8598)		Up (0.8039)
vvi04075/vvi4016	Ethylene signaling	Vitvi05g00715	ERF1/2		Up (3.2021)	Up (1.9759)	Up (3.7921)
vvi04075/vvi4016	Ethylene signaling	Vitvi06g00518	ETR	Up (1.5036)	Up (2.1683)
vvi04075/vvi4016	Ethylene signaling	Vitvi19g01359	ETR			Up (0.6295)	Up (0.7797)
vvi04814	Cytoskeleton/intracellular transport	Vitvi19g00302	DYNLL	Up (4.9448)			Up (2.3011)
vvi04814	Cytoskeleton/intracellular transport	Vitvi09g00355	DYNLL	Down (−0.8394)	Up (1.9422)	Up (1.3237)	Up (1.3237)
vvi04814	Cytoskeleton/intracellular transport	Vitvi06g00964	KIF15	Down (−1.5027)	Down (−4.0374)		Down (−3.0122)
vvi04814	Cytoskeleton/intracellular transport	Vitvi08g01197	KIF22	Down (−1.7732)	Down (−4.5211)	Down (−1.4375)	Down (−2.8902)
vvi4016	Aquaporins/water transport	Vitvi13g00012	PIP1‑2‑C3H				Down (−1.1676)
vvi4016	Aquaporins/water transport	Vitvi02g00310	PIP1‑2‑ERF		Up (0.7764)	Up (0.9538)	Up (0.7594)
vvi4016	Aquaporins/water transport	Vitvi06g00281	PIP2‑7‑C3H		Down (−1.8206)
vvi4016	Aquaporins/water transport	Vitvi03g00155	PIP2‑7‑C3H	Down (−1.0826)	Down (−2.1705)
vvi4016	Aquaporins/water transport	Vitvi06g01346	TIP1‑1‑C3H		Down (−4.7853)		Down (−2.6727)
vvi4016	Aquaporins/water transport	Vitvi08g01602	TIP1‑1‑C3H	Down (−1.4169)	Down (−1.8461)
vvi4016	Aquaporins/water transport	Vitvi04g00307	TIP4‑3‑BES		Down (−1.9668)		Down (−1.3868)
vvi4016	Stress transcription factors	Vitvi06g00741	WRKY33		Up (1.3051)
vvi4016	ROS detoxification	Vitvi18g00095	CAT1	Up (2.0949)	Up (2.1431)
vvi4016	ROS detoxification	Vitvi19g00233	RboHD		Up (2.1978)	Up (1.0433)	Up (2.4092)

Conclusions and future perspectives

This study highlights the role of rootstock–scion interactions in shaping grapevine responses to controlled water deficit, revealing genotype-dependent differences in early molecular responsiveness across grafting combinations. Differences in transcriptional responses suggest that combinations vary in their molecular sensitivity to water limitation. Coordinated modulation of hormonal signaling, aquaporins, phenylpropanoid metabolism, and cytoskeletal components indicates the complexity of regulatory networks underlying early water-deficit responses. Notably, the combined modulation of cytoskeletal genes and aquaporins represents a relatively underexplored aspect of scion–rootstock interactions under reduced water availability.

Unlike most previous studies focusing on established cultivars, this work evaluates newly selected table grape genotypes across different rootstock combinations during early selection stages. The candidate genes identified (Table 2; Supplementary Figs. 5 and 6) represent potential molecular markers for discriminating scion–rootstock combinations based on their sensitivity to reduced irrigation, providing a useful framework for breeding programs aimed at optimizing table grape performance under moderate water limitation [5]. Notably, the two scion genotypes displayed distinct transcriptional responsiveness patterns: Maula generally showed stronger transcriptional reprogramming under reduced irrigation, whereas Genusia exhibited comparatively moderated responses. These results suggest that scion–rootstock combinations may differ in their molecular sensitivity thresholds to irrigation reduction. More broadly, this study shows how transcriptomic profiling can complement physiological screening to support early evaluation of scion–rootstock combinations in table grape breeding. Future studies integrating transcriptomic data with physiological analyses and field validation will be essential to support the development of climate-resilient grapevine cultivars under variable water availability.

Authors’ contributions

M.F.C., R.P., and M.D. conceptualized the study. F.A.M.M., M.S., A.D.M., A.P., T.B., C.B., L.R.F. performed the experiments and analyzed the data. M.F.C., M.D. and F.A.M.M. wrote and revised the manuscript. All the authors have read and approved the manuscript.

Funding

This research was funded by the Italian Ministry of University and Research (MUR), project ‘PRIMA’ ‘Regenerative agricultural approaches to improve ecosystem services in Mediterranean vineyards—REVINE’.

Data availability

Raw sequencing data from Illumina sequencing experiments from this study have been submitted to the Sequence Read Archive (SRA; https://www.ncbi.nlm.nih.gov/sra/, accessed on 18 December 2024) under the BioProject ID PRJNA1253408.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.