Genotype-specific transcriptomic response to drought stress in potato cultivars modulated by microbial biostimulants Authors: Rachele Tamburino, Lorenza Sannino, Francesca Palomba, Emanuela Russo, Maria Consiglia Esposito, Adriana Sacco, Michelina Ruocco, Nunzia Scotti Published: 2025 ABSTRACT AbstractBackgroundDrought is a major abiotic stress that significantly limits potato productivity and tuber quality. Microbial biostimulants have emerged as promising tools to improve crop resilience and promote sustainable agriculture. This study aimed to investigate the transcriptomic response to severe drought stress in three potato cultivars (Camelia, Cicero, and Agata) characterized by differing polyphenol content. In addition, the potential mitigating effects of a microorganism-based biostimulant mixture composed by two Trichoderma species, were evaluated.ResultsSevere drought stress led to a significant reduction in tuber yield in Camelia and Agata, whereas Cicero exhibited greater tolerance. Application of the microbial biostimulant slightly alleviated yield loss in Camelia but was associated with an increased proportion of non-commercial tubers. Polyphenol content varied according to genotype and treatment, Cicero displayed elevated polyphenol levels in the skin under drought, while Camelia showed increased levels in both skin and pulp when drought was combined with biostimulant application. Transcriptomic analysis revealed genotype-specific drought response strategies. Camelia exhibited enhanced proteostasis and osmoprotection via anthocyanin accumulation, while Cicero maintained photosynthetic activity, redox homeostasis, and genome stability. Trichoderma-based biostimulant treatment modulated these responses by enhancing endoplasmic reticulum protein quality control in Camelia, and promoting photosynthesis and sugar metabolism in Cicero.ConclusionsPotato responses to drought are highly genotype-dependent and involve complex trade-offs among growth, stress defense, and metabolic reprogramming. Microbial biostimulants offer a promising approach to enhance drought resilience and tuber nutraceutical quality. However, their efficacy is genotype-specific and requires targeted application strategies for optimal results.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12870-025-07810-1.Keywords: Solanumtuberosum, Drought stress, Trichoderma spp., NGS, Differential gene expression Background Climate change is exerting an increasing impact on global food production systems. Rising temperatures, coupled with the growing frequency and severity of drought events, are significantly impairing crop development and yield, thereby reducing food availability and accessibility. These adverse conditions could continue to escalate, posing a critical challenge to the sustainability of agriculture. In this context, enhancing crop resilience has become a strategic priority for global food security, especially for nutritionally and economically important crops. To develop climate-resilient crops, it is essential a comprehensive understanding of plant responses to environmental stress and the identification of genes conferring tolerance [1–4]. Potato (Solanum tuberosum L.) is a staple food crop ranked fourth in global consumption. Its tubers are rich in carbohydrates produced via leaf photosynthesis, and contain significant levels of vitamins, minerals, and phenolic compounds [5]. The accumulation of these health-promoting compounds is genotype-dependent and influenced by environmental conditions, particularly temperature. Generally, low temperatures promote phenolic biosynthesis [6]. Despite its global importance, potato is highly susceptible to drought stress due to its shallow root system and physiological sensitivity [7]. Drought is the second major cause of potato yield loss after pathogen infection and adversely affects multiple physiological and biochemical processes, including photosynthesis, nutrient assimilation, and tuber development [8, 9]. The extent of yield reduction depends on both environmental factors (e.g., severity and duration of drought) and plant characteristics (e.g., genotype and developmental stage). Among these processes, photosynthesis is particularly critical. Drought impairs carbon partitioning by triggering a cascade of molecular and physiological responses that determine plant survival [3, 10, 11]. Furthermore, drought stress negatively regulates genes involved in stomatal function, hormonal signaling, particularly abscisic acid, and antioxidant defense, ultimately compromising plant growth and productivity [12]. Facing climate-induced stress, plant scientists are adopting integrated strategies that combine traditional breeding approaches with modern genomic tools and the use of biostimulants. Among biostimulants, soil microorganisms have shown significant promise in improving plant growth and stress resilience while maintaining ecological sustainability. In particular, Trichoderma spp. are widely recognized for their ability to stimulate root development, improve water and nutrient uptake, enhance photosynthetic efficiency, and mitigate oxidative damage by modulating reactive oxygen species (ROS) levels [13–18]. These beneficial effects have been documented in a variety of crops under drought conditions, including rice, tomato, and durum wheat [19–21]. Although several studies have characterized the molecular response of potato to drought [1, 4, 6, 9, 11, 22, 23], the molecular mechanisms underlying the interaction between Trichoderma-based biostimulants and drought tolerance remains largely unexplored. Hence, in-depth analyses of these interactions are crucial to optimize the use of these beneficial microorganisms in sustainable crop management. This study aimed to select potato genotypes with high polyphenol content because polyphenols are key molecules involved in the detoxification of reactive oxygen species (ROS) generated under drought stress. These compounds act as powerful antioxidants, protecting cellular structures and maintaining redox homeostasis. Moreover, elevated polyphenol concentrations in potato tubers enhance their nutritional value due to the well-documented health-promoting properties of these compounds in the human diet [24, 25]. Previous research has demonstrated that water deficit can modulate the accumulation of phenolic compounds and the antioxidant capacity of potato tubers [26, 27]. Therefore, genotypes with higher tuber polyphenol content may possess an enhanced constitutive antioxidant potential, supporting greater plant tolerance to stress conditions. Identifying such genotypes is valuable not only for understanding stress adaptation mechanisms but also for improving the nutritional profile of potatoes. Subsequently, we evaluated the performance of the three most promising potato genotypes under drought, with and without a Trichoderma-based biostimulant application. The two genotypes showing the best performance in terms of yield-related parameters (i.e., tuber number, weight, and size) and polyphenol content, were selected for transcriptomic analysis using RNA sequencing (RNA-seq) to investigate the molecular basis of their drought response. Our results provide novel insights into the gene networks modulated by drought and microbial treatment, laying the groundwork for breeding and biostimulant-based strategies to enhance potato resilience under climate change scenario. Methods Plant, materials and treatment conditions A collection of nine potato cultivars (kindly provided by Coppola Patate srl and Dr. M. Mazzei), were included for preliminary screening from those available on the market (Table 1). Table 1Potato genotypes used in this workGenotypeTuber number/plantTuber weight/plant (g)Tuber shapeSkin colourPulp colourAgata9–11350–400OvalYellowYellowBlue Star12–14400–450Long ovalPurplePurpleCamelia9–11400–420OvalDark yellowDark yellowCayman12–14420–450Round/ovalYellowLight yellowCicero11–12400–420OvalYellowYellowDouble Fun12–14400–450OvalPurpleYellow- PurpleEmanuelle12–14420–460Oval/long ovalDark yellowDark yellowSunita9–11380–400Round/ovalYellowYellowViolet Queen15–17300–360Long ovalPurplePurple Potato genotypes used in this work To ensure homogeneity of the plant material for drought experiments, apical buds of plants obtained from the tubers were first sterilized by immersion in 70% ethanol for 30 s, then in 1.5% NaClO containing 0.1% (v/v) Tween20 for 15 min with gentle stirring. The buds were then washed five times, each for 5 min, in Milli-Q water. Subsequently, the sterilized buds were sown in vitro under controlled conditions (16 h light 40 µmol photons m− 2 s− 1 and 8 h dark at 24 °C) on Murashige & Skoog (MS) medium with B5 vitamins (Duchefa, The Netherlands), solidified with 0.8% (w/v) agar, with 30 g/l sucrose and 250 mg/l of cefotaxime. In vitro rooted seedlings (3 weeks old) of the potato (Solanum tuberosum L.) cultivars Agata, Camelia, and Cicero were transplanted into polystyrene trays containing S-type soil and maintained for three weeks in an environmentally controlled walk-in chamber under the following conditions: 14 h light (200 µmol photons m⁻² s⁻¹) at 25 °C and 10 h dark at 20 °C. Subsequently, forty uniformly developed plants per cultivar were transplanted into pots (18 cm height × 18 cm diameter) filled with a mixture of S-type soil, virgin sandy soil, and Baltic blond peat (5:4:1) (Fig. 1). Plants were irrigated daily with tap water at 80 ± 5% soil relative water content (SRWC) for two weeks without the addition of fertilizers. The experiments were conducted in a greenhouse under the following environmental conditions: temperature 12–23 °C and relative humidity 60–80%. For each cultivar, forty plants were divided into two subsets differing in the application of a biostimulant formulation, which was administered twice via irrigation, at transplanting and two days prior to water withholding. Each subset was further divided into five experimental treatments (four plants per treatment): Control (C), Drought 1 (D1), Rewatering 1 (R1), Drought 2 (D2), and Rewatering 2 (R2) (Fig. 1). Schematic representation of the experimental design for drought stress and biostimulant application in potato plants. Three-weeks-old seedlings were transplanted into cell trays and grown under controlled conditions for an additional three weeks. Forty uniformly developed plants were transplanted into individual pots and divided into two groups with or without biostimulant treatment, applied twice at transplanting into pots (t = 0) and after 2 weeks (t = 14), prior to the first drought cycle. Plants underwent two consecutive drought-rewatering cycles. The first drought (D1, 14 days) began 16 days after transplanting, followed by two days of rewatering (R1, 2 days). The second drought cycle (D2, 10 days) was initiated after R1 sampling and followed by an additional rewatering cycle (R2, 2 days). Leaves were sampled at each stage, and tubers were collected at physiological maturity for yield and biochemical analyses A novel bioformulation containing two Trichoderma strains from the collection of CNR-IPSP was used for the treatment. The Trichoderma strains were characterised at the morphological and molecular levels as reported by [28], then inoculated into flasks containing Potato Dextrose Broth (PDB) and left to grow separately in a rotating incubator at 25 °C for 7 days at 150 rpm. Once the minimum concentration of the culture was 105 CFU the two strains were mixed and let it grow for 7 more days in a bioreactor. The resulting fermented product, diluted at 106 CFU, was applied to the potato plants. Drought stress was imposed during the stolon induction phase by withholding irrigation until the SRWC decreased to 10 ± 2%. This water deficit was maintained for 14–16 days (14 days for Camelia and Agata, and 16 days for Cicero) through daily monitoring of soil moisture to sustain 10 ± 2% SRWC; this phase was designated as D1. Control pots were maintained throughout the experiment at 80 ± 5% SRWC. Subsequently, plants in the R1, D2, and R2 treatments were rewatered to 80% SRWC for 2 days, after which a second drought cycle (D2 and R2) was imposed for 10 days. Thereafter, R2 plants were rewatered and maintained under well-watered conditions until crop maturity. Soil water status was estimated gravimetrically before water application to pots [29]. Leaves from all experimental conditions were harvested from second and third nodes at the end of each treatment stage and stored at − 80 °C. At the end of each stage, plants were subsequently maintained under well-watered conditions until maturity for tuber production. After tuberization, tubers were collected, weighted, and measured to assess yield under all applied treatments. Polyphenols content determination Total polyphenol content (TPC) was assessed both in pulp and skin of tubers by Folin-Ciocalteu assay as previously described [30]. Tubers were washed with distilled water, sliced separating skin and pulp and lyophilized using a freeze-dryer alpha 1–2 LD plus (Martin Christ Gefriertrocknungsanlagen GmbH, Germany). Briefly, 100 mg of lyophilized pulp and skin of tubers were extracted in 1 ml 80% (v/v) methanol and sonicated for 30 min at 10 °C and then centrifuged at 12,000 xg for 20 min at 4 °C. Supernatant was recovered and kept on ice at dark till the usage. In a 96-well plate, 20 µL of each sample extract was added to 150 µL of Folin–Ciocalteu reagent diluted with distilled water (1:30, v/v) and 30 µL of 20% (w/v) Na2CO3, incubated at room temperature in the dark for 45 min and then absorbance at 765 nm was measured in a Multiskan™ Sky microplate spectrophotometer (Thermo Scientific, Waltham, MA, USA). Gallic acid (GA) was used as reference standard and TPC was estimated from the GA calibration curve (range 31.25–750 µg/mL, 6 levels; r2 = 0.998). The results were expressed as GA equivalents per g of dry leaf (mg GAE/g DW, means ± SD of three biological replicates). Statistical differences between treatments and controls samples were assessed using Student’s t test (*P ≤ 0.05, **P ≤ 0.01); whilst a threeway ANOVA was performed to evaluate the effects of genotype, treatment, biostimulant and their interactions. Total RNA isolation and sequencing Four-five leaves per plant from second and third nodes were harvested for RNA extraction and analysis. Each plant represented a biological replicate. Total RNA was extracted from leaves of three biological replicates of drought stressed Camelia and Cicero plants (i.e., D1 and D1 + mix conditions) and their respective controls using the RNeasy Plant Mini Kit (Qiagen, Germania) following manufacturer’s instructions. The integrity and quality of RNA was evaluated by denaturing agarose gel electrophoresis and nanodrop measurements (Additional file 1) and 2 µg of RNA per sample were used for sequencing. High-throughput sequencing service and preliminary data processing of 24 RNA-seq libraries were carried out by Sequentia Biotech SL (https://www.sequentiabiotech.com/). The raw reads were quality-checked using Trimmomatic v0.39, applying a minimum read length of 35 bp and a minimum sequencing quality of 20. Filtered reads were mapped to the potato reference genome v3 (Ensembl Genomes 57) using STAR aligner v2.7.9a with default parameters. Count matrix was build using featureCounts v2.0.3, considering only reads with a minimum alignment quality score of 40. The resulting count matrix was normalized using the TMM method, and low expressed genes were filtered out using HTSfilter v1.28.0. Subsequent analyses, including principal component analysis (PCA) were performed using FactoMineR and pheatmap R packages (R software 4.3.0), respectively. After both analyses, libraries that did not group with their respective biological replicates were excluded from downstream analyses. Differential expression and gene ontology (GO) analyses Differential expression analysis was carried out with edgeR R package (R software 4.3.0), only considering differential expressed genes (DEGs) with a log fold-change (LogFC) higher than 1 or lower than − 1 and False Discovery Rate (FDR) lower than 0.05. Four different contrasts were analysed per genotype: Drought1 vs. Control (D1 vs. C). Drought1 vs. Control (D1 vs. C). Drought1 + mix vs. Control + mix (D1 + mix vs. C + mix). Drought1 + mix vs. Control + mix (D1 + mix vs. C + mix). Control + mix vs. Control (C + mix vs. C). Control + mix vs. Control (C + mix vs. C). Drought1 + mix vs. Drought1 (D1 + mix vs. D1). Drought1 + mix vs. Drought1 (D1 + mix vs. D1). Taking these comparisons, a GO enrichment analysis of up-regulated and downregulated genes was performed using clusterProfiler R package (R software 4.3.0). Quantitative real-time PCR validation To validate the RNA-seq data, five genes were selected among either the most positively or negatively affected categories for qRT-PCR analysis. An aliquot of each sample was retrotranscribed to cDNA by digesting trace amount of DNA using DNase kit (Invitrogen) and using RevertAid 1 st strand cDNA synthesis kit (ThermoFisher Scientific, USA). cDNA quality was further evaluated by RT-PCR amplification of 18 S rRNA gene and amplification products were visualized by agarose gel electrophoresis (Additional file 1). Diluted cDNA (1:20) was used as template for semi-quantitative real time using the Platinum™ SYBR™ Green qPCR SuperMix-UDG (Applied Biosystems, USA). All reactions were set using 6.25 µl of 2x SYBR green dye, 0.6 µM of forward and reverse primer mix, 4.5 µl template cDNA and run on the 7900HT Fast Real Time PCR (Applied Biosystems, USA). Primers used and program cycle parameters are listed in Additional file 2. The relative expression levels were calculated using the 2−∆∆CT method [31]. EF-1α was used as internal control. Data representing three biological replicates and three technical replicates are expressed as mean ± standard deviation. Statistical analysis The statistical analysis was performed in Microsoft Excel (Windows 11 Enterprise) and R software. One- and three-way ANOVA and RNA-seq data analysis were conducted in R software 4.3.0. Significancy of polyphenol contents and gene expression levels was analysed using Student’s t-test. The differences between different treatments were considered significant when the P-value was lower than 0.05. Results Screening of potato genotypes for drought stress tolerance In order to select the cultivar for drought stress experiment, the total polyphenol content (TPC) was measured in both tubers’ skin and pulp of collected cultivars (Fig. 2). The Folin-Ciocalteau assay revealed that polyphenol content was higher in the skin than in the pulp for all the tested genotypes, among them, Camelia exhibited the highest TPC in both tissues (i.e., 132.9 ± 5.5 and 36.7 ± 4.9 mg GAE/g DW in skin and pulp, respectively), followed by Cicero, whose pulp showed a polyphenol content comparable to that of Camelia’s pulp (i.e., 33.3 ± 5.2 mg GAE/g DW). Total polyphenol content (TPC) in skin and pulp tubers of a potato cultivars’ collection. Data represent the mean (± SD) of three biological replicates (n = 3; each replicate is a set of tubers from independent plants). Different letters indicate significant differences between TPC within the same tissue using the Tukey test (P < 0.05) Based on their high TPC, suitability for multiple purposes (i.e., processed, table consumption and industrial use) and agronomic value, Camelia, Cicero and Agata were selected for experiment under severe water deficit conditions, with or without treatment using the microorganism-based biostimulant mixture according to the scheme reported in Fig. 1. Plant performance was assessed by measuring tuber yield (Table 2) and polyphenol content in both skin and pulp following water deficit treatments (Fig. 3). The three cultivars exhibited distinct responses to drought, which were also influenced by biostimulant treatment. Table 2Yield of tubers expressed as mean weight and number of tubers in all experimental conditionsTreatmentCameliaCiceroAgataWeight ± sd (g)N. tubers (average)SizeWeight ± sd (g)N. tubers (average)SizeWeight ± sd (g)N. tubers (average)Size C 85.7 ± 11.94.75M/L64.9 ± 6.32.8S/M129.4 ± 6.111.5M/L D1 53.4 ± 19.8*3.75M/L71.2 ± 5.42M85.6 ± 20.5**4M/L R1 65.3 ± 17.24.5M/L83.8 ± 4.6**3.5S/M72.1 ± 19.4**5.5M/L D2 73.9 ± 10.63.5M/L78.1 ± 6.92.5S/M77.3 ± 7.0**3.8M/L R2 57.3 ± 19.8*2.75M/L78.4 ± 3.4*2S/M99.0 ± 11.3**4.3M/L C + mix 74.3 ± 21.55M/L96.9 ± 17.82.3M99.5 ± 28.14.8M/L D1 + mix 76.5 ± 9.24.5S/M83.4 ± 3.9*1.5S/M71.4 ± 12.04.3M/L R1 + mix 71.9 ± 6.93.25S/M78.3 ± 3.5*3S/M77.6 ± 21.25.5M/L D2 + mix 66.3 ± 14.93.5S/M69.8 ± 6.2**2.2S/M100.4 ± 9.76M/L R2 + mix 74.0 ± 13.15S/M77.4 ± 6.92S/M93.1 ± 7.04M/L Three way ANOVA_weight P value Three way ANOVA_number P valueGenotype***Genotype***Treatment***Treatment**biostimulantnsBiostimulantNsgenotype: treatmentnsgenotype: treatmentNsgenotype: biostimulantnsgenotype: biostimulantNstreatment: biostimulantnstreatment: biostimulant*genotype: treatment: biostimulantnsgenotype: treatment: biostimulant*S Small, size ≤ 3 cm, M Medium, 3 cm < size < 5 cm, L Large, size ≥ 5 cm, ‘***’, ‘**’, ‘*’ and ‘ns’ indicate significant at P < 0.001, P < 0.01, P < 0.05 P > 0.05 and not significant, respectively Yield of tubers expressed as mean weight and number of tubers in all experimental conditions S Small, size ≤ 3 cm, M Medium, 3 cm < size < 5 cm, L Large, size ≥ 5 cm, ‘***’, ‘**’, ‘*’ and ‘ns’ indicate significant at P < 0.001, P < 0.01, P < 0.05 P > 0.05 and not significant, respectively Total polyphenol content (TPC) in tubers of drought-stressed potato cultivars grown without and with biostimulant. TPC in skin and pulp tubers of Camelia (A, D), Cicero (B, E) and Agata (C, F). Data represent the mean (± SD) of three biological replicates. Asterisks indicate significant differences: P < 0.05 (*), P < 0.01 (**), P < 0.001 (***); ns = not significant Tuber weight decreased in Camelia and Agata tubers under all applied conditions (up to 37.7% and 44.3%, respectively) compared to their control, although this reduction was mitigated (up to 10.7% and 28.3%, respectively) when biostimulant mix was applied (Table 2). Interestingly, Camelia tubers from D1 + mix and Agata from D2 + mix even showed a slight increase in weight (3 and 1%, respectively) compared to the control. Conversely, Cicero tubers significantly increased in weight under drought (up to 29.1% in R1) compared to the control, but this gain was reversed upon biostimulant application, with a weight reduction of up to 27.9%. Regarding size, biostimulant-treated Camelia tubers under all applied conditions were smaller than controls, whereas Cicero and Agata showed no significant size changes. Tuber nutritional quality was evaluated by measuring the polyphenol content in both pulp and skin (Fig. 3). In Camelia, pulp polyphenols significantly increased (up to 1.7-fold) under all experimental conditions with biostimulants, while no significant changes were observed in the absence of biostimulants. TPC from skin also increased significantly up to 1.2-fold in both R1 and R1 + mix conditions. In Cicero, skin polyphenols significantly increased up to 1.6-fold under drought stress but slightly decreased with biostimulant application (i.e., D2 + mix). Pulp TPC in Cicero remained largely unchanged, except for a 59% reduction in R2. In Agata, TPC increased significantly in D1 pulp (1.4-fold) and R1 and R2 skin (up to 2.4-fold) without biostimulants, while no significant variations were observed with biostimulant mix. Threeway ANOVA indicated that both genotypes and treatment significantly affected all measured parameters (P < 0.01). However, genotype-treatment interactions were only significant for polyphenols in tuber pulp. The biostimulant application also had a significant effect on polyphenol content in both tissues, and its interaction with genotype was statistically significant. Principal Component Analysis (PCA) of tuber weight, tuber number and polyphenol content in skin and pulp revealed that the first two principal components together explained the entire variance (PC1: 77.9%. PC2: 22.1%; Additional file 3). Tuber weight and TPC in skin and pulp were the main contributors to PC1. Notably, polyphenol content in skin and pulp was positively correlated, with particularly high values in Camelia genotype, while TPC negatively correlated with tuber weight. Transcriptional changes induced by drought and drought + biostimulant treatment in leaves. Based on yield and TPC results, Camelia and Cicero cultivars were selected for RNA-seq analysis considering four conditions: control (C) drought (D1), control with biostimulant (C + mix), and drought with biostimulant (D1 + mix), with D1 and D1 + mix corresponding to the first imposed water stress (Fig. 1). Indeed, due to poor RNA quality obtained from leaves of some conditions (i.e., R1, D2 and R2), RNA-seq analysis was restricted to D1. PCA of all the libraries revealed a clear genotype-specific clustering. Notably, the Camelia samples grouped differently depending on the biostimulant treatment (Additional file 4). PC1 explaining 40.6% of the total variance, primarily separated the genotypes, suggesting a genotype-specific response to drought. Biostimulant treatment induced substantial transcriptional reprogramming in Camelia. A summary of differentially expressed genes (DEGs) is showed in Fig. 4, in general, Camelia showed a higher number of differentially expressed genes. Specifically, in the D1 vs. C comparison, Camelia had 4,305 DEGs (1,749 up-regulated, 2,556 down-regulated), while Cicero had only 426 DEGs (229 up-regulated and 197 down-regulated) (Fig. 4). The two genotypes shared just 4.3% of total DEGs (Additional file 5 A). In D1 + mix vs. C + mix, Camelia had 5,215 DEGs (2,150 up-regulated and 3,065 down-regulated) and Cicero 934 (360 up-regulated and 574 down-regulated) (Fig. 4). Number of identified Differentially Expressed Genes (DEGs) in Camelia and Cicero genotypes. Number of DEGs identified in the comparisons D1 vs. C, D1 + mix vs. C + mix, D1 + mix vs. D1, and C + mix vs. C. DEGs were selected using a cut-off of P-adjust < 0.05 and |log2FC| ≥ 1 The 7.7% of total DEGs were shared (Additional file 5 A). For D1 + mix vs. D1 comparison, Camelia had 855 DEGs (413 up-regulated and 442 down-regulated), and Cicero 757 (160 up-regulated and 597 down-regulated), with 3.4% overlap (Fig. 4, Additional file 5 A). In C + mix vs. C, Camelia had 164 DEGs (68 up-regulated and 94 down-regulated), whereas Cicero had just one, with no overlap (Fig. 4, Additional file 5 A). In all comparisons the number of down-regulated genes was higher than up-regulated which means that plants stop regular active metabolism as main mechanism of response to the drought stress. These results indicate that Camelia and Cicero exhibited genotype-specific response to water deficit and biostimulant application. Further analysis of shared DEGs (Additional file 5B) showed that in the D1 vs. C, 156 genes (59 up-regulated and 97 down-regulated) had the same trend in both cultivars, while 39 genes showed opposite trend. In D1 + mix vs. C + mix, 267 genes (60 up-regulated and 207 down-regulated) shared the same trend, whilst 170 genes were oppositely regulated. In D1 + mix vs. D1, 24 shared genes (2 up-regulated and 22 down-regulated) were similarly regulated, while 28 genes displayed opposite trend. Camelia also had more DEGs consistently present across all conditions (Additional file 6 A), particularly in the D1 vs. C and D1 + mix vs. C + mix comparisons (2994 overlapping DEGs). In Cicero, the largest overlap (313 DEGs) was between the D1 + mix vs. C + mix and D1 + mix vs. D1 (Additional file 6B). qRT-PCR validation of five randomly selected DEGs confirmed the RNA-seq findings (Additional file 7). For each gene, qRT-PCR validation was performed only the treatment combinations that showed significant expression changes in the RNA-Seq data. Gene ontology (GO) enrichment analysis In Camelia, the higher number of DEGs identified compared to Cicero resulted in a greater overlap across the analysed comparisons, reflecting a broader range of affected Gene Ontology terms (Additional file 8 C and 8 F). The most impacted biological processes (BPs) were consistently enriched across the D1 vs. C, D1 + mix vs. C + mix and D1 + mix vs. D1 comparisons. These included up-regulated categories associated to transmembrane transport, protein folding and xyloglucan metabolic processes, as well as down-regulated categories related to translation and protein modification (Fig. 5A and Additional file 9 A). Interestingly, the photosynthesis-related categories exhibited divergent expression patterns depending on the condition, being predominantly up-regulated in the D1 vs. C comparison. Secondary metabolism and senescence-related processes were exclusively up-regulated in drought-stressed plants while GO categories linked to calcium-mediated signalling were strongly induced by biostimulant application (Additional file 9 A). GO enrichment analysis of potato genotypes Camelia and Cicero. Gene Ontology (GO) enrichment analysis of the four pairwise comparisons in potato genotypes Camelia (A) and Cicero (B), categorized into Biological Processes (BP). Each circle represents a GO term; circle size and color indicate the number of counts and the p-value, respectively At the molecular functions (MF) level, in D1 vs. C and D1 + mix vs. C + mix comparisons, the most affected categories included those related to energy and signal transduction, such as ATP and nucleotide binding, along with transferase and protein kinase activities (Additional file 8 A). These molecular functions were primarily associated with membrane-localized compartments Additional file 8B and 9 C). In Cicero, the most significantly enriched BPs in both the D1 vs. C and D1 + mix vs. C + mix comparisons were primarily related to photosynthesis (Fig. 5B and Additional file 10 A). Furthermore, exposure to drought (D1 vs. C) enhanced the expression of genes involved in chromatin remodelling consistent with a possible inhibition of cell division to conserve energy. Conversely, pathways associated with abscisic acid (ABA) and ethylene signalling, as well as defense responses, protein folding and glutathione metabolism, were down-regulated under drought stress (Additional file 10 A). Notably, biostimulant application under drought conditions (i.e., D1 + mix vs. C + mix), triggered the activation of primary metabolic pathways including gluconeogenesis, fructose metabolism and carbon fixation (Fig. 5B and Additional file 10 A). These pathways are known to contribute to stress adaptation via the accumulation of soluble sugars that act as osmoprotectors [32]. However, other related processes such as carbohydrate metabolism were suppressed. Additionally, genes related to circadian rhythm were up-regulated, potentially enhancing photosynthetic and metabolic efficiency under stress [33, 34]. Consistently, increased expression of chlorophyll binding activity at the MF level suggested a possible role for the biostimulant in preserving photosynthetic capacity (Additional file 8D and 10B). However, in the D1 + mix vs. D1 comparison, photosynthesis-related genes were down-regulated. This may seem contradictory with the D1 + mix vs. C + mix result, but it can be explained by a mild induction of these genes mediated by Trichoderma treatment. When compared with the C + mix condition, this induction appears as up-regulation, whereas when compared with the D1 conditions, it appears as down-regulation due to the stronger effect of drought on photosynthesis. Such an effect suggests a possible adaptive modulation mediated by the biostimulant. Similarly, transcriptomic profile in tomato treated with Ascophyllum nodosum seaweed showed mild up-regulation of photosynthesis-related genes, exceeding the baseline in irrigated controls [35]. Overall, these findings suggest that Cicero adopts a distinct, biostimulant-supported strategy, to cope to water deficit, largely involving metabolic reprogramming. The altered biological processes were predominantly associated with subcellular compartments such as the chloroplast, thylakoid and other plastid structures (Additional file 8E and 10 C). A clear genotype-dependent transcriptional response emerged between Camelia and Cicero under stress condition. To further explore their intrinsic transcriptional differences, their transcriptomic profiles under non-stress conditions, specifically in the control (C) and control with biostimulant (C + mix) were compared. Under control conditions (i.e., C Camelia vs. C Cicero) a total of 5,929 DEGs were identified with 2,325 up-regulated in Camelia and 3,604 up-regulated in Cicero. Gene Ontology (GO) enrichment analysis revealed that Camelia exhibited a significant up-regulation of BPs related to translation, ribosome biogenesis, and photosynthesis including key functions such as light harvesting and photosystem I activity (Additional file 11 A). Additional enrichment was observed in genes associated with protein folding, phosphorylation, response to hydrogen peroxide, and phosphate starvation response, suggesting a transcriptional state primed for stress anticipation or mitigation. In contrast, BPs down-regulated in Camelia, and thus up-regulated in Cicero in control conditions, included primary metabolic and structural processes, such as carbohydrate metabolic process, cell wall organization, cell wall biogenesis, glucan metabolic process, and proteolysis (Additional file 11 A). Further reductions in expression were seen in genes involved in lipid catabolism, phosphate ion transport, and defense response to other organisms, indicating a lower investment in baseline metabolic and defense functions in Camelia compared to Cicero (Additional file 11 A). Consistently, one of the most enriched MFs in Camelia was the structural constituent of the ribosome (Additional file 11B), aligned with the enrichment of ribosome-associated cellular compartments (Additional file 11 C). Under biostimulant-treated control conditions (i.e., C + mix Camelia vs. C + mix Cicero), the transcriptomic comparison between Camelia and Cicero revealed 5,260 DEGs, with 2,106 genes up-regulated in Camelia, and 3,154 in Cicero. GO analysis showed that Camelia maintained a transcriptional profile strongly oriented toward protein biosynthesis, with significant up-regulation of BPs such as translation, ribosome biogenesis, and rRNA processing. This pattern was accompanied by enhanced activation of photosynthesis-related pathways, including light harvesting (Additional file 12 A). These trends were further supported by enriched MFs related to ribosomal activity and protein folding, such as structural ribosome constituents, ATP-dependent chaperone activity, unfolded protein binding, protein self-association and heat shock protein binding (Additional file 12B and 12 C). Moreover, Camelia showed increased expression of genes involved in stress responses, including those induced by heat, hydrogen peroxide, and salt stress, suggesting a transcriptional state primed for environmental resilience. In contrast, Cicero exhibited a transcriptional profile more oriented toward defense-related GO categories, including responses to biotic stimuli (e.g., fungi and other organisms), as well as pathways associated with lipid and carbohydrate metabolism, proteolysis, and transmembrane transport (Additional file 12 A). Discussion Drought stress is one of the most significant factors that limit plant growth and yield. Potato is particularly susceptible to water withholding thus in the last years there has been an increasing attention to research in this field [7, 36–40]. Beneficial soil microorganisms support crops by promoting growth and enhancing tolerance to both abiotic and biotic stresses [14, 41]. In this study, we investigated the performances of three potato genotypes (i.e., Camelia, Cicero, and Agata), selected for their polyphenol content and agronomic value, under severe drought stress also evaluating the effect of a Trichoderma-based biostimulant mixture. Trichoderma spp., a well-known genus of plant growth-promoting fungi, are common in soil and root ecosystems and can establish robust, long-lasting colonizations of root surfaces, positively influencing crop productivity and tolerance to environmental stresses [14, 42]. Trichoderma-based biostimulant mixture used in this study was tested in open-field experiments, where its effectiveness and persistence were verified [28]. Previous studies demonstrated that short periods of water deficit imposed during the vegetative or tuberization phases markedly reduce both tuber number and yield, while increasing the proportion of small tubers (< 35 mm in diameter) [43]. Consistent with these findings, in the present study, drought stress significantly reduced tuber yield in the cultivars Camelia and Agata, confirming their sensitivity to water limitation (Table 2). Notably, in Camelia, biostimulant treatment partially mitigated yield loss, suggesting enhanced drought resilience. However, biostimulant treatment also led to a higher proportion of small-sized tubers, with a consequent reduction of the commercial quality. This aligns with the results of [29], who observed that PGPR-based biostimulants could mitigate yield losses under drought. Fungal biostimulants may prioritize stress tolerance over sink development by modulating specific hormonal and metabolic pathways. These results underscore the importance of considering both “genotype” and “environmental context” when evaluating biostimulant efficacy. Drought triggers alterations, both stimulatory and inhibitory effects, in non-enzymatic antioxidants such as phenolic compounds that clearly reflect the complex metabolic response to environmental stressors [27, 44–46]. The increase of phenolic compounds plays an essential role in neutralizing ROS and maintaining redox balance [47, 48]. The impact of drought on phenolic content can be genotype- and treatment-specific. In our study, the treatment with Trichoderma induced in Camelia a consistent increase in polyphenols in both skin and pulp (Fig. 3), suggesting enhanced secondary metabolism and nutritional quality of tubers. Similar effects were observed in tomato, where a simulated drought stress, by PEG treatment, either with or without Trichoderma application, induced an increase in total phenolic content [49]. In contrast, Cicero and Agata displayed more variable and less pronounced changes. For example, Cicero showed increased skin polyphenols under drought alone, which was reversed by biostimulant treatment, while Agata exhibited increased pulp polyphenols only in the absence of biostimulant. These results highlight the role of genotype in adapting secondary metabolite responses to environmental stresses. To investigate the molecular responses of the most drought-performing potato cultivars (i.e., Camelia and Cicero), high-throughput RNA sequencing was employed to profile transcriptome changes in the leaves of plants exposed to water deficit either alone (D1) or in combination with a Trichoderma-based biostimulant mixture (D1 + mix). Each genotype exhibited distinct molecular responses. In Camelia, drought (D1) triggered the up-regulation of genes related to protein folding, protein oligomerization, and various amino acid transport and carbohydrate metabolic processes, suggesting an attempt to preserve metabolic homeostasis under stress. At the same time, GO terms related to protein synthesis, including translation, ribosome biogenesis, and rRNA processing, were significantly suppressed. Interestingly, photosynthesis-related genes were up-regulated, which may reflect an adaptative strategy to maintain energy production despite reduced stomatal conductance and CO2 availability [50]. Additionally, activation of oxidative and osmotic stress-related genes, alongside suppression of biotic defense genes, suggested a trade-off between biotic and abiotic stress pathways [51]. Genes involved in the anthocyanin biosynthesis were also up-regulated, supporting their role in both antioxidant defense and osmotic adjustment [52]. Carbohydrate metabolism was also elicited, consistent with a drought tolerance strategy, as these molecules are essential for energy production, osmoregulation and osmoprotection, thus linking growth, development, and carbon status as observed by [53]. Suppression of DNA replication and repair and hormonal signalling pathways further indicated cell division and growth inhibition [54]. Down-regulation of water transport and glutathione-related genes reflected impaired oxidative stress mitigation, although subcellular redox dynamics are known to affect glutathione metabolism [55, 56]. Enhanced expression of leaf senescence genes may support nutrient remobilization from aging tissues to other vital areas of the plant [57]. Altogether, these responses reflect a shift toward cellular adaptation and metabolic conservation under drought. When Camelia was treated with the biostimulant under drought (D1 + mix), the transcriptomic response in both D1 + mix vs. C + mix and D1 + mix vs. D1 comparisons was broader and more integrative. Similarly to what observed in the D1 vs. C, growth-related processes (e.g., translation, ribosome biogenesis, and DNA replication) were suppressed. However, in contrast to the untreated drought conditions, photosynthesis-related genes, including light harvesting and photosystem I, were also repressed suggesting a shift from maintaining photosynthetic activity toward energy conservation. This contrast with previous reports in rice where Trichoderma treatment enhanced photosynthetic gene expression (i.e., ribulose-bisphosphate carboxylase small chain A, ferredoxin-NADP reductase, photosystem I subunit O, photosystem II core complex proteins psbY) under drought [58]. Water transport genes remained down-regulated, consistent with a conserved drought avoidance strategy. Notably, biostimulant treatment strongly induced genes involved in protein quality control (e.g., protein folding, chaperones, endoplasmic reticulum (ER) unfolded protein response), suggesting an enhanced capacity for proteostasis and ER stress response [59]. Ion transport and signalling pathways, especially calcium signalling, were activated, in agreement with their role in abiotic stress perception [60, 61]. Up-regulation of genes involved in membrane remodelling and cell wall modification further supports structural adaptation under stress [62]. These data suggest that the biostimulant modulates Camelia’s stress response from a defense-maintenance to a damage-mitigation mode, promoting cellular integrity and survival. In Cicero, drought (D1) induced strong transcriptional down-regulation of defense-related genes (e.g., glutathione metabolism-related genes), hormonal signalling (i.e., ethylene and ABA) and protein processing pathways. Conversely, photosynthesis-related genes (e.g., light-harvesting complex and chlorophyll biosynthesis) were up-regulated, possibly to maintain redox balance and prevent ROS accumulation as previously observed in Arabidopsis and Malus domestica [63, 64]. Additional up-regulated genes included those involved in transmembrane transport, nitric oxide biosynthesis (a redox signaling molecule), and chromatin remodelling, suggesting mechanisms for genome stability and stress adaptation [65–69]. Moderate induction of amino acid biosynthesis pathways (e.g., lysine and threonine) supports limited but targeted metabolic reprogramming for osmotic protection as previously observed in sandbur and wheat [70, 71]. Overall, Cicero’s response reflected a minimalist survival strategy, characterized by selective activation of key stress-related processes and suppression of non-essential metabolic activity. Under combined drought and biostimulant treatment (D1 + mix), Cicero’s transcriptomic response shifted toward a more adaptive reprogramming. When transcriptomic response was analysed in comparison with Trichoderma-treated plant in control conditions (i.e., D1 + mix vs. C + mix), photosynthesis, photorespiration, and sugar metabolism pathways (e.g., sucrose and fructose biosynthesis) were activated, indicating improved energy production and osmoprotection. This is consistent with previous studies showing sugar signaling’s role in stress responses [70, 72]. Similar results were reported in Pinus massoniana seedlings inoculated with Trichoderma longibrachiatum, which induced the accumulation of osmotic substances (e.g., sugars, proline) to regulate osmotic pressure under drought stress [73]. Notably, photosynthesis was up-regulated in D1 + mix vs. C + mix, but down-regulated in D1 + mix vs. D1, suggesting that the biostimulant’s effect on photosynthesis is context-dependent. Circadian rhythm-related genes were also activated, potentially restoring physiological homeostasis under stress, as previously observed in the model plant Arabidopsis [74, 75]. Furthermore, hormonal signalling (e.g., cytokinin) and phosphate starvation responses were enhanced. Although cytokinins can negatively interact with phosphate signalling [76], they are known to improve drought tolerance in several crops [77–80]. Meanwhile, key genes related to primary metabolism (carbohydrates, lipids, amines), cell wall biogenesis, and general transmembrane transport were strongly down-regulated, indicating a strategic reduction in energy-consuming processes in favour of sustaining core metabolic and photosynthetic functions. The analysis of transcriptomic profile in the D1 + mix vs. D1 comparison only the GO category related to the assembly of transcription preinitiation complex resulted up-regulated. Among the down-regulated BPs, in addition to those described in the D1 + mix vs. C + mix comparison, were detected water and sulfate transport, proteolysis, DNA replication and cell wall biogenesis suggesting a stronger inhibition of cell growth-related processes in this condition. Conclusions Overall, the results of the present work reveal significant genotype-dependent variability in potato responses to drought particularly at the transcriptomic level. Camelia and Cicero, the two cultivars exhibiting the strongest drought performance, employed distinct molecular strategies to cope with water deficit. Camelia demonstrated a pronounced suppression of growth alongside strong activation of proteostasis mechanisms and anthocyanin accumulation, suggesting a stress tolerance mechanism centered on cellular protection (Fig. 6A). In contrast, Cicero maintained photosynthetic efficiency and redox homeostasis, while down-regulated growth and defense pathways (Fig. 6C). Schematic representation of the most regulated biological processes in Camelia and Cicero. Panels A and C show the most up-regulated biological processes under D1 vs. C conditions, while panels B and D refer to D1 + mix vs. C + mix conditions for Camelia and Cicero, respectively. For processes up-regulated in both conditions, arrows indicate the treatment with the higher number of expressed genes. Down-regulated processes are highlighted in blue. Abbreviations: Cp, chloroplast; Vc, vacuole; Nc, nucleus; Mt, mitochondrion; Px, peroxisome The application of Trichoderma-based biostimulant mixture modulated these responses in a genotype-specific manner. In Camelia, it enhanced protein folding, stability, and ion transport regulation (Fig. 6B), while in Cicero, it stimulated photosynthesis and soluble sugar accumulation (Fig. 6D). From an agronomic perspective, our results demonstrated that the biostimulant mix partially mitigated drought-induced yield loss in Camelia but also increased the proportion of small, non-commercial tubers. While this represents a drawback from a market perspective, it can be considered a beneficial stress-adaptative strategy by increasing reproductive output. Moreover, polyphenol accumulation exhibited distinct genotype-specific patterns, with Camelia emerging as the most responsive genotype in terms of improving nutritional value under combined drought and biostimulant treatment. These findings highlight the importance of a genotype-tailored approach when assessing the effectiveness and agronomic relevance of microbial biostimulants, taking into account the complex interplay between plant genetics, environmental stress, and microbial composition. Supplementary Information Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Supplementary Material 4 Supplementary Material 5 Supplementary Material 6 Supplementary Material 7 Supplementary Material 8 Supplementary Material 9 Supplementary Material 10 Supplementary Material 11 Supplementary Material 12 Acknowledgements We thank Mr. R. Nocerino and Dr. Marco Porcelli (CNR-IBBR, Portici, Italy) for assistance in plant growth. Authors’ contributions N.S. and M.R. conceived and designed the study. R.T. and L.S. performed drought experiments and biochemical analysis. R.T., L.S., E.R., F.P., M.C.E. performed RNA isolation and qRT-PCR analysis experiments. A.S. carried out the microbial biostimulant growth and mixture preparation. R.T., A.S., M.R., N.S. analysed the data. R.T. and N.S. wrote the manuscript. R.T., A.S., M.R., N.S. revised the manuscript. All authors have read and agreed to the final version of the manuscript. Funding This research was partially funded by grants Agritech National Center (European Union Next-GenerationEU, PIANO NAZIONALE DI RIPRESA E RESILIENZA (PNRR) – MISSIONE 4 COMPONENTE 2, INVESTIMENTO 1.4 – D.D. 1032 17/06/2022, CN00000022) and “Ministero dello Sviluppo Economico” FCS-Horizon 2020 PON I&C 2014–2020, Project No. F/200088/01–03/X45 “Sviluppo di nuove tecnologie nell’agricoltura di precisione per la produzione sostenibile di genotipi di patata con elevate qualità nutrizionali (SOS-TATA)”. Data availability All data generated or analysed during this study are included in this published article and its supplementary information files. The datasets generated and/or analysed during this study are available in the Zenodo repository, [10.5281/zenodo.17084816] (10.5281/zenodo.17084816). 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. References Associated Data Supplementary Materials Supplementary Material 1 Supplementary Material 2 Supplementary Material 3 Supplementary Material 4 Supplementary Material 5 Supplementary Material 6 Supplementary Material 7 Supplementary Material 8 Supplementary Material 9 Supplementary Material 10 Supplementary Material 11 Supplementary Material 12 Data Availability Statement All data generated or analysed during this study are included in this published article and its supplementary information files. The datasets generated and/or analysed during this study are available in the Zenodo repository, [10.5281/zenodo.17084816] (10.5281/zenodo.17084816).