Freshly harvested cut roses (‘Carola’) were obtained from a rose plantation in Huadu District, Guangzhou, China (longitude 113°28′12″E, latitude 23°26′24″N), with a commercial maturity level (cultivated for 2 years). The flower stems were placed in plastic buckets filled with tap water and transported to the postharvest experiment at the Biotechnology Research Institute of Zhongkai University of Agriculture and Engineering within 1 h. Cut roses of uniform developmental stage and stem thickness, free from pests and diseases were selected for the experiments. The stems, with a length of 35 cm and retaining the first and second leaves, were trimmed. The stems were initially placed in conical flasks containing deionized water and subjected to three hours of dark treatment. Subsequently, the cut roses were individually transferred to separate conical flasks containing different concentrations of KNO₃ treatments: A (control, CK), B (25 mmol/L), C (50 mmol/L), D (75 mmol/L), and E (100 mmol/L), under a light intensity of 100 μmol m^–2 s^–1, temperature of 20°C ± 2°C, and humidity of 60% ± 5% (Fig. 1).

The water loss rate of the cut roses within a 9-h light period was continuously monitored using an automated device for precise water loss detection. Data were collected every 5 min, with three biological replicates selected for each treatment.

Four biological replicates, exhibiting consistent growth conditions, were chosen for each treatment. Samples were collected at 0, 1, 3, 6, and 9 h of light exposure. For each stem, a 0.25 cm² leaf sample was excised using a surgical blade. Two random fields of view from the lower epidermis of rose petals were captured using an optical microscope, resulting in a total of 8 fields of view for statistical analysis. Subsequently, 40 random stomata were selected, and their apertures were quantified using Image-win64 software, enabling the calculation of the mean aperture value (Fig. 1).

To induce stomatal closure, the cut roses were placed in darkness for 3 h. Following the dark treatment, the cut roses were individually placed in conical flasks containing deionized water, and 25 mmol/L KNO₃ solutions, respectively. Sampling time points of 0, 1, and 3 h were selected under a light intensity of 100 μmol m^–2 s^–1 (0 h denoted the time point after 3 h of darkness, coinciding with the initiation of white light exposure). Leaf samples were taken from the second leaf of the roses using a surgical blade, immediately flash-frozen in liquid nitrogen, and subsequently stored in a −80°C ultra-low temperature freezer. Each treatment consisted of three biological replicates, with each biological replicate comprising one cut rose. The experimental conditions were maintained at a temperature of 20°C ± 2°C and relative humidity of 60% ± 10%.

RNA extraction, library construction, quality control, and transcriptome sequencing were outsourced to Guangzhou Genedenovo Biotechnology Co., Ltd., Principal component analysis was performed using R based on gene expression levels. Low-quality data filtering was conducted following the approach described by Vidal et al. [43]. The functional annotation of differentially expressed genes (DEGs) was performed by aligning the NCBI rose genome. Alignment analysis based on the reference genome was conducted using HISAT2 software [44]. Differential expression analysis and weighted gene co-expression network analysis (WGCNA) analysis were performed using the OmicSmart transcriptome analysis platform (Fig. 1).

Two differentially expressed genes were selected for analysis, and primer sequences were designed via NCBI (https://www.ncbi.nlm.nih.gov/tools/primer-blast), with the housekeeping gene GAPDH serving as the reference gene. The reaction mixture was prepared using 2 × Real Star SYBR Mixture (with ROX II) reagent (YEASEN, Shanghai, China) as follows: 10 μL of SYBR Mixture (with ROX II), 1 μL of cDNA template, 1 μL of each primer (10 μmol/μL), and 8 μL of nuclease-free water. Real-time PCR reactions were conducted using the CFX Connect Real-Time PCR Detection System (Bio-Rad, USA) under the following conditions: 95°C for 5 min; 95°C for 10 s, 60°C for 30 s, for 40 cycles. Relative quantitative analysis was performed using the 2^−ΔΔCT method.

All data were presented as mean ± standard error (SE). Significance testing was carried out using the least significant difference method (LSD) (p < 0.05). Data analysis was performed using SPSS 26.0 (International Business Machines Corporation, USA) and Sigmaplot 14.0 (Systat Software Inc., USA).

The water loss rate in cut roses ‘Carola’ exhibited distinct responses to various concentrations of KNO₃ treatments. During the initial 60 min, the 25 and 75 mmol/L KNO₃ treatments exhibited the highest rates of water loss, surpassing the control (CK) treatment, while the 50 and 100 mmol/L KNO₃ treatments resulted in the lowest rates. Subsequently, the water loss rates gradually increased in the KNO₃ treatments, whereas the control showed a gradual decline. After 150 min, all KNO₃ treatments showed higher water loss rates compared to the control, with the 25 and 75 mmol/L KNO₃ treatments displaying the highest rates, significantly exceeding the CK treatment (Fig. 2A). Furthermore, the cumulative water loss after 300 min was notably higher in the 25 and 75 mmol/L KNO₃ treatments than in the other treatments, including the control (Fig. 2B).

As depicted in Figs. 3A and 3B, the stomatal apertures of cut roses leaves exhibited distinctive characteristics under the control and various concentration of KNO₃ treatments. Specifically, the 75 mmol/L KNO₃ treatment consistently demonstrated the widest stomatal aperture under light conditions, significantly exceeding the control. Furtherore, the stomatal apertures of the 25, 50, and 100 mmol/L KNO₃ treatments exhibited a rapid increase within the 0 to 3-h time frame, with values noticeably surpassing those of the control by the 3-h mark (Fig. 3).

Total RNA from ‘Carola’, a cultivar of cut roses, was extracted and subjected to agarose gel electrophoresis. The ensuing electrophoretic pattern exhibited distinct bands, underscoring the RNA’s suitability for transcriptome sequencing (Fig. S1). Each sample yielded a surplus of 5.4 billion effective reads, with GC content spanning 45.78% to 46.65%. Notably, the Q30 percentage surpassed 91% across all samples, affirming the robustness of the sequencing data underpinning this investigation (Table S1).

Utilizing the expression levels of known genes in each sample, we conducted Principal Component Analysis (PCA), calculated Pearson correlation coefficients, performed sample relationship clustering, and conducted inter-sample correlation analysis. The findings demonstrated high reproducibility among the samples in this transcriptional sequencing study, validating their suitability for subsequent analysis (Fig. 4A).

A total of 5456 DEGs were up-regulated, while 6607 DEGs were down-regulated (Fig. 4B). Gene ontology (GO) enrichment analysis indicated that the top 20 enriched terms in the cellular component category were chromosome-related (GO:0000785, GO:0044427, GO:0005694), as well as photosynthetic membrane (GO:0034357). In terms of biological process, the major enriched terms were related to biological regulation in response to chemical and light stimuli (GO:0042221, GO:0009416, GO:0050896, etc.). The molecular function category showed enrichment for protein dimerization activity (GO:0046983), suggesting that cut roses exhibit stress responses and initiate gene expression and translation under potassium nitrate treatment in light conditions (Fig. 4C).

Correspondingly, the Kyoto encyclopedia of genes and genomes (KEGG) enrichment analysis revealed that the top 20 enriched pathways included biosynthesis of secondary metabolites, photosynthesis, starch and sucrose metabolism, plant-pathogen interaction, and plant hormone signal transduction (Fig. 4D).

Differential analysis identified 918 genes with significant expression differences between the 25 mmol/L KNO₃ treatment and the control after 1 h, along with 662 genes after 3 h of treatment (Fig. 5A). Employing the R programming language, Venn diagrams and UpSetR plots were generated, revealing 246 genes that exhibited differential expression at both 1 and 3 h, underscoring their close association with the response to KNO₃ treatment (Fig. 5A).

Based on the expression levels of these differentially expressed genes across various samples, GO functional annotation was performed. The top 20 enriched processes encompassed nitrate metabolic processes, terpenoid metabolic processes, vacuolar components (e.g., vacuolar membranes), ion transport, and response to stimuli. These results indicate transcriptional changes in these processes in response to KNO₃ treatment, potentially influencing associated physiological processes (Fig. 5B).

Furthermore, the differentially expressed genes were mapped to the KEGG database, revealing highly enriched pathways such as nitrogen metabolism, the pentose phosphate pathway, secondary metabolite biosynthesis, metabolic pathways, and the MAPK (mitogen-activated protein kinase) signaling pathway (Fig. 5C). This suggests the involvement of certain genes in these pathways in the response of cut rose leaves to KNO₃ treatment.

Heatmap analysis of expression levels for genes related to NO₃⁻ transport and metabolism revealed three classes of differentially expressed genes (Fig. 6A). Class I genes, including nitrate regulatory genes NRG2, APSR, and nitrate transporter NAR2.1, exhibited relatively high expression levels after 1 h of KNO₃ treatment, indicating their early response to KNO₃ treatment. Class II genes, such as nitrate transporters NRT3.1, NRT2.5, and nitrate reductase NIA1, showed higher expression levels after 1 h of KNO₃ treatment, suggesting nitrate transport into cells and associated redox reactions. Class III genes, including nitrate regulatory gene REL2 and nitrate transporter NRT2.7, exhibited decreased expression levels after KNO₃ treatment, indicating their suppression during KNO₃ treatment (Fig. 6A).

Differential gene heatmap analysis identified five classes of K⁺ channel and transporter genes (Fig. 6B). Class I genes displayed relatively low expression levels at 0 and 1 h in the control group but exhibited increased expression levels at 3 h, as well as after 1 and 3 h of KNO₃ treatment. Class II genes exhibited relatively low expression levels at 0 and 1 h, with increased expression at 3 h. Among them, POT7 (potassium transporter 7) and POT6 displayed higher expression levels after 3 h of KNO₃ treatment compared to the control, while HAK13 (high-affinity K⁺ transporter 13), POT1, and SKOR (stellar K⁺ outward rectifier) exhibited lower expression levels after 3 h of KNO₃ treatment compared to the control. Class III genes, represented by POT10 and HAK17, showed higher expression levels after 1 h of KNO₃ treatment compared to other groups. Class IV genes exhibited higher expression levels without light exposure and KNO₃ treatment, but their expression decreased under light exposure. Class V genes displayed lower expression levels during KNO₃ treatment compared to the control (Fig. 6B).

Five classes of differentially expressed genes related to vacuoles were identified. Class I genes exhibited relatively high expression levels after 1 h of KNO₃ treatment, including vacuolar sorting-associated proteins (VPS28-1, VPS32.1, VPS36), vacuolar cation/proton exchangers (CAX3, CAX2), and an unidentified vacuolar membrane protein YML018C. Class II genes showed relatively high expression levels after 3 h of KNO₃ treatment, including vacuolar sorting receptors (VSR6), vacuolar sorting-associated proteins (VPS26A and VPS13b), and an amino acid transporter (YPQ1). Class III genes exhibited higher expression levels after KNO₃ treatment compared to the control, including vacuolar sorting receptors (BP80) and vacuolar sorting-associated proteins (VPS24-1, VPS20.1, VPS29). Class IV genes displayed higher expression levels without light exposure and KNO₃ treatment, with decreased expression under light exposure. Class V genes showed lower expression levels during KNO₃ treatment compared to the control (Fig. 6C).

Four classes of aquaporin-related genes were identified, and their expression patterns varied among different treatments. Class IV genes, NIP3-1 (nodulin 26-like intrinsic protein 3–1) and PIP2-1 (plasma membrane intrinsic protein 2–1), displayed higher expression levels after 3 h of KNO₃ treatment, indicating their close association with the response to potassium nitrate treatment (Fig. 6D).

By employing WGCNA, we explored the correlation between gene expression levels and their response to potassium nitrate treatment in cut roses. Related genes and their expression patterns were identified, which were then classified into modules based on their similarity in expression. To construct the gene clustering tree, the expression correlation coefficients among genes and determined a power value of 7 was computed (Fig. 7A). The resulting clustering relationships facilitated the partitioning the genes into different modules, and the modules with similar characteristic expression patterns were subsequently merged based on the similarity of their module eigengenes (Fig. 7B). Our network comprised a total of 4860 genes, divided into 18 modules. The turquoise module contained the highest number of genes (1050), while the darkturquoise module had the fewest (58) (Figs. 7A and 7B). Notebly, the turquoise module exhibited decreased gene expression under light exposure, suggesting a potential association between these genes and the response of cut roses to light. Additionally, the lightgreen and greenyellow modules exhibited decreased expression upon KNO₃ treatment, while the midnightblue module showed increased expression, indicating their involvement in the response to potassium nitrate treatment. Moreover, the darkturquoise, darkgreen, darked, and yellow modules exhibited relatively higher expression levels after 1 h of potassium nitrate treatment compared to other groups. Similarly, the black module showed relatively higher expression levels after 3 h of potassium nitrate treatment (Fig. 7C). These modules provide valuable insights into subsequent analysis of the gene expression dynamics associated with KNO₃ treatment in cut roses.

GO analysis revealed that the turquoise module was significantly enriched in processes such as nucleic acid binding transcription factor activity, phosphatidylethanolamine metabolic process, second-messenger-mediated signaling, and chemical homeostasis (Fig. 8A). Moreover, the turquoise module showed a high enrichment level in several KEGG pathways, including Plant hormone signal transduction, Photosynthesis-antenna proteins, and Plant-pathogen interaction (Fig. 8C).

GO analysis of the black module demonstrated a significant enrichment in processes such as anion channel activity, chloride transmembrane transporter activity, cellular response to inorganic substance, glutamate metabolic process, and nitrate metabolic process (Fig. 8B). Furthermore, the black module exhibited a strong enrichment in several KEGG pathways, including the Pentose phosphate pathway, Carbon metabolism, Ascorbate and aldarate metabolism, and Nitrogen metabolism (Fig. 8D).

To identify hub genes within each module, we selected the top 100 gene pairs with the highest correlation coefficients (weight values) and constructed a network. Based on our screening, the turquoise module contained 45 genes, the black module contained 83 genes, and the darkgreen module contained 40 genes. According to the NR database annotation, TSJT1 was identified as the hub gene in the turquoise module, known as a stem-specific protein (Fig. 9A); LAX2 (ncbi_112193093) was identified as the hub gene in the black module, characterized as an auxin transporter-like protein involved in signal transduction (Fig. 9B); SCPL34 (ncbi_112197638) was identified as the hub gene in the darkgreen module, classified as a serine carboxypeptidase-like gene (Fig. 9C). Furthermore, several transcription factors such as BEH2, NAC021, CDF3, ethylene-responsive transcription factor (ERF053), ethylene receptor 2 (ETR2), and auxin response factor (ARF6) displayed high connectivity in the turquoise, black, and darkgreen modules, indicating their involvement in the regulatory processes associated with the response of cut roses to potassium nitrate treatment under light conditions (Fig. 9).

To verify the precision and dependability of the RNA-Seq dataset, we employed qRT-PCR for the evaluation of the expression profiles of two selected DEGs, NIA1 and NIR1 (Table S2). The expression patterns of DEGs assessed through q-PCR exhibited a noteworthy degree of concordance, underscoring the robust reliability of the RNA-Seq data produced within the scope of this investigation (Fig. S2).

Previous studies conducted by Huang et al. have provided evidence that cut roses possess the highest stomatal density on their lower epidermis, while petals, sepals, and stems have only scattered stomata [49]. This finding suggests that, like most cut flowers, water loss through stomata is the primary pathway for leaf desiccation in cut roses. In our study, we observed the changes in water loss and stomatal opening on the lower leaf surface of ‘Karola’ cut roses under light conditions after treatment with different concentrations of KNO₃. Our results clearly indicate a close relationship between leaf stomata and water loss in the cut roses. We observed varying degrees of differences in the effects of different concentrations of KNO₃ treatment on stomatal opening on the hypodermis (Fig. 3), as well as the rate (Fig. 2A) and cumulative amount (Fig. 2B) of water loss in cut roses. Notably, the treatment with 75 mmol/L KNO₃ exhibited the most prominent promotion of stomatal opening. In conclusion, our findings demonstrate that different concentrations of KNO₃ treatment significantly regulate stomatal opening and water loss in cut roses leaves. Specifically, varying concentrations of KNO₃ treatment can promote stomatal opening on the hypodermis (Fig. 3) to different extents and accelerate the water loss rate (Fig. 2) in the flower stems.

Numerous key ion channel families are widely present in terrestrial plants; however, moss plants lack ion channels specifically expressed in guard cells, highlighting the essential role of ion osmotic pressure in driving stomatal movement during the evolution of land plants [50]. The activation of nitrate reductase through illumination explained the inverse relationship between light intensity and nitrate content, as it reduces nitrate accumulation [51]. Recent studies have identified the involvement of ion channels or transporters associated with potassium, calcium, and NO₃⁻ transport on the vacuolar membrane of guard cells, regulating the opening and closing of plant stomata [52–55]. Aforementioned research demonstrated that the utilization of potassium fertilizer not only impacts rice yield formation but also has consequences for the absorption and utilization of nitrogen within the rice [56]. Upon absorption, NO₃⁻ is reduced to NO₂⁻, NH₄⁺, and glutamine by NR, NiR, and glutamine synthetase (GS), respectively, and subsequently metabolized into various amino acids and nitrogenous compounds through glutamate synthase (GOGAT) [34]. Hence, the activity of nitrogen metabolism enzymes is crucial for NO₃⁻ absorption and assimilation. In this study, transcriptome GO and KEGG analysis demonstrated the highest enrichment in nitrate metabolism and nitrogen metabolism in cut roses treated with 25 mmol/L KNO₃ (Fig. 5). Differentially expressed genes related to nitrate, such as ferredoxin--nitrite reductase (NIR1), nitrate reductase (NIA1), and glutamine synthetase leaf isozyme (GLN2), exhibited expression levels 4 to 46 times higher after treatment with 25 mmol/L KNO₃ compared to the control (Fig. 6A). The expression levels of nitrate transporter genes, including NRT2.5 and NRT3.1, increased by 14 and 18.24 times, respectively, after KNO₃ treatment compared to the control (Fig. 6A). These findings suggest that following NO₃^- absorption and assimilation, these genes may significantly contribute to the increased stomatal aperture and water loss rate in cut roses treated with KNO₃. Under potassium deficiency, the expression levels of nitrate transporters in Arabidopsis roots were significantly downregulated but rapidly recovered upon potassium supplementation. Potassium activation is essential for nitrogen assimilation-related enzymes and facilitates the transport of NO₃⁻ from roots to shoots. Nitrogen levels impact potassium uptake, and excessive nitrogen supply reduces potassium utilization efficiency. Potassium plays a significant role in enhancing NO₃⁻ uptake, serving as a crucial factor for nitrogen uptake and assimilation. Notably, the assimilation rate of inorganic nitrogen is higher under light conditions compared to dark conditions. The distribution of nitrogen in plants influences nitrogen uptake and assimilation, and increased potassium supply under high nitrogen conditions enhances the rate of nitrogen distribution to the leaves. In this study, the 25 mmol/L KNO₃ treatment resulted in a higher K⁺ concentration, which partially promoted NO₃⁻ absorption.

Stomatal movement is tightly regulated by the turgor pressure of guard cells, with the intracellular concentration of K⁺ playing a pivotal role in modulating turgor pressure. Unlike other cell types, mature guard cells lack plasmodesmata, necessitating the passage of all inorganic ions across the cell membrane. In this context, potassium ion channels assume a critical role [52,57]. Remarkably, even minor fluctuations in K⁺ concentration can exert profound effects on membrane charge, triggering a cascade of signaling events associated with K⁺ acquisition, salinity response, plant immunity, and programmed cell death [58–60]. Within the scope of this study, gene ontology analysis revealed significant enrichment in the category of ion transport (Fig. 5B). Differentially expressed genes implicated in potassium transport exhibited varying degrees of expression variation across different treatments and time points (Fig. 6B). The impact of varying potassium supply levels on potassium ion channels and transporters was evident, highlighting the regulatory role of potassium absorption governed by feedback mechanisms based on the plant’s nutritional status [61]. Accumulation of potassium in the roots leads to a reduced rate of potassium uptake, while under conditions of low potassium levels, the activation and enhancement of potassium ion channels and transporters occur [62]. High salt concentrations in water negatively impact fruit growth, physiology, production, and post-harvest quality by reducing water availability, causing toxicity through Cl⁻ and Na⁺, and creating nutrient imbalances due to ion competition, including deficiencies of essential nutrients (Ca²⁺, Mg²⁺, K⁺, and NO₃⁻) [63]. K⁺ and NO₃⁻ represent the most abundant cations and anions in plant tissues, with their absorption rates exhibiting a positive correlation and mutual promotion to enhance charge balance. This phenomenon can be attributed to the improvements in charge balance during nutrient uptake and long-distance transport processes, as well as the activation of enzymes involved in nitrate assimilation induced by potassium ions [64]. Excessive nitrogen adversely affects potassium uptake and utilization efficiency, as evidenced by the downregulation of gene expression related to MdAKT1, MdHAK1, and MdPT5 under conditions of high nitrogen stress [38]. Nitrogen transporters, including NRT1.1 and NRT1.5, participate in K⁺ transport, and intricate physiological and molecular mechanisms interconnect nitrogen and potassium uptake [47,65]. A recent investigation has unveiled that the NO₃⁻ channel SLAH3 collaborates synergistically with the K⁺ channels GORK and SKOR, exerting an influence over nitrogen-potassium homeostasis and the modulation of membrane potential within the realm of plant biology [48]. Within the context of this study, treatment with 25 mmol/L KNO₃ resulted in elevated potassium ion concentration, thereby facilitating K⁺ uptake. However, this process is subject to regulation by the internal potassium status of the plant, and the high concentration of NO₃⁻ negatively modulates potassium uptake, ultimately leading to a complex expression pattern of different potassium ion channels and transporters.

Plants can respond to various signals, such as blue light, osmotic stress, sugars, and plant hormones, by activating H⁺-ATPase [66–68]. The activated H⁺-ATPase drives K⁺ uptake by activating inward-rectifying K⁺ (K⁺ _in) channels, while anions like NO₃⁻ and malate enter guard cells, resulting in K⁺ accumulation and subsequent stomatal opening [1,69,70]. The turquoise module genes exhibited the highest expression levels in darkness (Fig. 7C), which decreased under light conditions. GO analysis revealed enrichment in nucleic acid binding transcription factor activity, phosphatidylethanolamine metabolic process, second-messenger-mediated signaling, and chemical homeostasis. This suggests that processes such as H⁺-ATPase undergo phosphorylation, signal transmission via second messengers, enhanced transcription factor activity, and regulation of ion and chemical substance homeostasis, ultimately controlling physiological processes like stomatal opening (Fig. 8A). Furthermore, the highly enriched KEGG pathways included plant hormones and photosynthetic antenna proteins, indicating the involvement of photosynthetic antenna proteins in light signal perception and the participation of plant hormones in subsequent signal transduction and physiological responses (Fig. 8C).

Previous studies have provided evidence for the activation of the isopentenyltransferase (IPT) gene by NO₃⁻, facilitating the essential steps of cytokinin synthesis crucial for cell division [71]. Increased NO₃⁻ supply induces the biosynthesis of ethylene by activating the 1-aminocyclopropane-1-carboxylate (ACC) synthase (ACS) and ACC oxidase (ACO) genes [31]. Nitrogen influences primary root growth by modulating auxin concentrations, while it affects lateral root growth through the regulation of the AFB3 (an auxin receptor gene) and NPF6.3 [27]. Under NO₃⁻ treatment, the transcription levels of the auxin signaling F-box3 (AFB3) receptor gene are enhanced, consequently inducing NAC gene expression. The aminotransferase gene DNR1 (Dull Nitrogen Response1) in rice was identified as a regulator of nitrate uptake and assimilation through modulation of auxin homeostasis. This regulation is achieved by a 520-bp deletion in the DNR1 promoter, leading to decreased transcript levels, elevated auxin content, and enhanced expression of OsARFs, which play a crucial role in activating genes associated with nitrate metabolism [72]. Simultaneously, microRNA 393 (miRNA393) negatively modulates this signaling pathway [34,73]. At 3 h of KNO₃ treatment, the black module exhibits relatively higher expression levels (Fig. 6C), with the hub gene identified as an auxin transporter-like protein. GO analysis reveals enrichment in anion channel activity, glutamate metabolic process, nitrate metabolic process, and the KEGG pathway nitrogen metabolism. These findings suggested that after 3 h of KNO₃ treatment, the influx of NO₃⁻ activates nitrogen metabolism processes, with the involvement of plant hormones such as auxin, promoting stomatal opening and water loss in cut roses (Figs. 8B, 8D).

During stomatal opening, vacuoles occupy approximately 90% of the guard cell volume, and a significant portion of ions entering the guard cells through the plasma membrane is stored in vacuoles. Conversely, stomatal closure in Arabidopsis guard cells is accompanied by a reduction in cell volume and approximately 20% decrease in vacuole volume [74]. The relationship between vacuole morphology and ion transport represents a dynamic equilibrium. The vacuolar K⁺ pool is dynamic and serves as a storage reservoir that replenishes ions during abundance or waste, maintaining cytoplasmic concentrations under starvation conditions [75,76]. In the GO analysis of this study, the vacuole-related category and vacuolar membrane showed significant enrichment (Fig. 5B). Differential expression of genes associated with vacuolar protein sorting (VPS) and cationic amino acid transporter (CAX) exhibited notable expression differences in response to the treatment with 25 mmol/L KNO₃, suggesting the involvement of vacuole-related genes in regulating the response of cut roses to KNO₃ treatment.