PMC:11492142 JSONTXT 2 Projects

Salicylic acid improves chilling tolerance via CsNPR1–CsICE1 interaction in grafted cucumbers Abstract Abstract Salicylic acid (SA) plays a role in the regulation of grafting-induced cold tolerance. However, the molecular mechanism behind it is still unknown. Here, we established that the phenylalanine ammonia-lyase (PAL) pathway-dependent elevate in SA content in grafted cucumber leaves was not only synthesized in the leaves but also transported from the roots under chilling stress. RNAi-CsPAL with low SA content as rootstock reduced SA accumulation in grafted seedling leaves while decreasing rootstock-induced cold tolerance, as evidenced by higher electrolyte leakage (EL), hydrogen peroxide (H2O2), and superoxide anion (O2·−) contents and lower expression of cold-responsive genes (CsICE1, CsDREB1A, CsDREB1B, and CsCOR47), whereas OE-CsPAL with high SA content as rootstock improved the cold tolerance of grafted plants in comparison with the wild type (WT). In addition, CsNPR1 was significantly upregulated in grafted cucumber under chilling stress, with exogenous and endogenous overexpressed SA inducing its transcriptional expression and protein stability, which exhibited higher expression in grafted plants than in self-root plants. While CsNPR1-overexpression (OE-CsNPR1) seedlings as scions were more tolerant to chilling stress than WT seedlings, CsNPR1-suppression (Anti-CsNPR1) seedlings as scions were more vulnerable to chilling stress. Notably, CsNPR1–CsICE1 interactions alleviated ROS accumulation and activated the expression of CsDREB1A, CsDREB1B, CsCOR47, CsCOR15, CsCOR413, and CsKIN1 to enhance SA-mediated chilling tolerance in grafted cucumber. Overall, our findings reveal that SA enhances chilling tolerance in grafted cucumbers via the model of the CsNPR1–CsICE1 transcriptional regulatory cascade. Introduction Cucumber (Cucumis sativus L.), a globally cultivated crop species, exhibits a significant susceptibility to cold stress. In the northern region of China, cucumbers are primarily grown in solar greenhouses during the winter season, but these plants are frequently subjected to chilling stress, which has a significant impact on their growth, productivity, and quality during this period. Grafting is the simplest and most effective method for improving plant resistance to both biotic and abiotic stresses [1] and has been widely used in horticulture crops. Grafting using strong-resistant rootstocks can improve plant tolerance to abiotic stress; for example, heat, cold, drought, and salt [2–5]. Additionally, these rootstocks can enhance the development of above-ground organs by transporting water, mineral salts, and hormones to scions, which then transmit photosynthetic products to rootstocks to promote root growth [6], indicating that a signal interaction exists between rootstock and scion [7]. As the major regulatory component of cold acclimation, the inducer of CBF expression 1 (ICE1)–c-repeat binding factor (CBF)–cold-responsive gene (COR) significantly contributes to the cold stress signaling response pathway [8–11]. It is thought that CBFs are essential for plants in response to cold signal, and it has been found that CBF1, CBF2, and CBF3 [12], which can greatly improve the resistance to cold of overexpression-Arabidopsis seedlings [13–15]. In the ICE1–CBF–COR cascade, ICE1 is an upstream transcription factor (TF) that can trigger the expression of CBF3 and enhance the cold tolerance [16] and stimulates the expression of CBFs and COR genes by identifying specifically the MYC elements in the CBF promoter [17]. Many components can regulate ICE1 activity. For instance, open stomata 1 (OST1) can phosphorylate ICE1 directly, and it can also bind competitively with high expression of osmotically responsive gene 1 (HOS1) to inhibit the ubiquitination degradation of ICE1, thereby maintaining ICE1 stability and regulating downstream genes expression [18, 19]. Agarwal et al. [20] reported that the MYB15 protein cooperates with ICE1 and binds to the MYB recognition site located at the CBF gene promoter to control CBF expression in response to cold stress. An et al. [21] demonstrated that MdABI4–MdICE1 interactions improved the transcriptional regulatory function of MdICE1 on MdCBF1, hence promoting the abscisic acid-induced cold tolerance of apples. Moreover, ICE1 transcriptional activity is inhibited through interaction with the inhibitors of jasmonic acid (JA) signaling, that include JAZ1 and JAZ4 [22]. Salicylic acid (SA), as a vital signal molecule, mediates complex biological functions in plants. Plant defense response, particularly the systemic acquired resistance (SAR) as well as hypersensitive response (HR), have been extensively researched in relation to SA [23–25]. Phenylalanine ammonlyase (PAL) as well as isochorismate synthetase (ICS) pathways are two synthetic pathways of SA in higher plants [26] and the PAL pathway is considered to be the main pathway of SA synthesis in cucumber under chilling stress [27]. Phenylalanine ammonlyase is a key rate-limiting enzyme in the PAL pathway and plays a crucial role in the synthesis of SA in response to low temperature [28, 29]. Many studies have established that SA can improve plant tolerance in response to abiotic stresses; for example salt, drought, and high or low temperature [30–33]. While SA plays a role in the communication between rootstock and scion, long-distance transport of SA from rootstock to scion can improve cucumber leaf cold tolerance [34]. Nevertheless, the potential mechanism of SA involvement in rootstock–scion interaction as well as the pathway for SA signal transport remain unclarified. The SA receptor known as the non-expressor of pathogenesis-related gene 1 (NPR1) includes the BTB/POZ (broad-complex, tramtrack, and bric-a-brac/poxvirus and zinc finger) domain, ankyrin repeat domain (ANK), and NPR1-like structure [35]. In normal circumstances, NPR1 usually localizes as a polymer in the plant cytoplasm, but stress conditions can cause rapid SA accumulation, resulting in polymeric NPR1 being reduced to monomer and transported to the plant nucleus. NPR1 can regulate pathogens-plant response through a transcriptional cascade that is promoted by SA and TGA TFs [36]. Furthermore, SA enhanced the interactions among NPR1, CDK8 (cyclin-dependent kinase 8), and WRKY18 (WRKY DNA-binding protein 18) in Arabidopsis. NPR1 improved its own and target gene levels to enhance plant immunity by recruiting WRKY18, CDK8, and TGA TFs, as well as RNA polymerase II [67, 71]. Notably, Olate et al. [37] revealed that NPR1 improved the Arabidopsis cold tolerance via the independent upregulation of cold-induced genes, not mediated by SA and TGA factors. NPR1 interaction with heat shock transcription factor 1 (HSFA1) was also found to stimulate the HSFA1-regulated genes expression and, consequently, plant cold tolerance. NPR1 has been proven to serve as a centralized point that synergistically regulates SA and cold signals [38]. Despite the importance of NPR1 in plant immunity and cold response having been well established, whether its participation in the SA involvement in grafting-induced cold tolerance as well as the underlying regulatory mechanism of it remain understudied. In this study, we present molecular results demonstrating the involvement of SA in the communication between rootstock and scion, and that CsNPR1 expression in cucumber scion can be encouraged by pumpkin rootstock. While establishing that CsNPR1 overexpression improves the chilling tolerance of grafted cucumber, we also demonstrate that the CsNPR1-CsICE1 interaction positively regulates grafted cucumber chilling tolerance. This study uncovers a new SA signal regulatory pathway for grafting-induced cold tolerance in plants, which is mediated by NPR1 and cold stress regulatory protein CsICE1. Results SA improved the grafting-induced cold tolerance of cucumbers In our earlier studies, it was shown that using pumpkin as a rootstock enhanced the cold tolerance of cucumber and the application of 1 mM exogenous SA enhanced the resistance of grafted cucumber plants to cold stress [34, 39]. Here, we found that Cs/Cm plants exhibited a cold-damage mitigation effect, significantly lower chilling injury index (CI), electrolyte leakage (EL), and malondialdehyde (MDA) accumulation, and strikingly higher mRNA levels of CsDREB1A and CsICE1 in comparison to Cs/Cs plants under cold stress (Fig. S1, see online supplementary material), but at optimum growth temperature (25°C/18°C), most of these indices did not differ. Exogenous SA mitigated the chilling damage and Pn reduction caused by cold stress while upregulating the chilling-induced mRNA levels of the carbon assimilation key genes (CsRCA, CsrbcL) and cold-responsive genes (CsICE1, CsDREB1A, CsCOR47) in cucumbers, especially in the Cs/Cm plants (P < 0.05). However, the SA synthesis inhibitor of L-a-aminooxy-phenylpropionic acid (AOPP) treatment resulted in cold-damage as well as reduced net photosynthetic rate (Pn) and expression of carbon assimilation-related genes and cold-related genes compared to H2O treatment under cold stress (Fig. S2, see online supplementary material). Phenylalanine ammonialyase (PAL) is the key enzyme of salicylic acid synthesis, and PAL is the main pathway of salicylic acid synthesis at low temperature [27, 40]. To further verify SA involvement in grafted cucumber chilling tolerance, we obtained overexpression and RNAi transgenic cucumber plants of CsPAL (CsaV3_6G039720) using Agrobacterium-mediated genetic transformation (Fig. S3, see online supplementary material), and the transgenic lines of OE-CsPAL-3 and RNAi-CsPAL-92 were chosen for subsequent experiments. At normal temperature (25°C/18°C), the seven grafting combinations of WT/WT, OE-PAL/OE-PAL, and RNAi-PAL/RNAi-PAL self-grafted plants, WT grafted onto rootstocks of RNAi-PAL and OE-PAL transgenic plants (WT/RNAi-PAL and WT/OE-PAL), as well as RNAi-PAL and OE-PAL as scion grafted onto rootstocks of WT plants (RNAi-PAL/WT and OE-PAL/WT) showed no differences (Fig. S4, see online supplementary material), while the RNAi-PAL/RNAi-PAL, WT/RNAi-PAL, and RNAi-PAL/WT plants showed severe wilting symptoms and significantly higher CI, EL, and MDA content than the WT/WT plants under chilling stress, the OE-PAL/OE-PAL, WT/OE-PAL, and OE-PAL/WT plants markedly alleviated the cold damage and showed lower values (Fig. 1a–d). Compared with the WT/WT plants, the RNAi-PAL/RNAi-PAL, WT/RNAi-PAL, and RNAi-PAL/WT plants accumulated more hydrogen peroxide (H2O2) and superoxide anion (O2.−), whereas the OE-PAL/OE-PAL, WT/OE-PAL, and OE-PAL/WT plants accumulated less under chilling stress. The diaminobenzidine (DAB) and nitroblue tetrazolium (NBT) staining results were consistent with the quantitative results (Fig. 1e–h). Cold stress upregulated the expressions of CsICE1, CsCOR47, CsDREB1A, and CsDREB1B in all grafting combinations. Compared with the WT/WT self-grafted plants, the increase of the relative expression of cold responsive genes in RNAi-PAL/RNAi-PAL, WT/RNAi-PAL, and RNAi-PAL/WT plants was significantly lower, which was higher in OE-PAL/OE-PAL, WT/OE-PAL, and OE-PAL/WT plants after treatment at 8°C/5°C for 72 h (Fig. 1i–l). Under chilling stress, a similar result was observed at the protein level of DREB1A, which was lower in the RNAi-PAL/RNAi-PAL, WT/RNAi-PAL, and RNAi-PAL/WT plants but higher in the OE-PAL/OE-PAL, WT/OE-PAL, and OE-PAL/WT plants compared to the WT self-grafted plants (Fig. 1m). Figure 1 Effect of RNAi and overexpress CsPAL on the cold tolerance of self-grafted and grafted cucumbers under cold stress. (a) Chilling injury index; (b) phenotype of seedlings; (c–f) EL, MDA, H2O2, and O2·- content, respectively; (g, h) DAB and NBT staining of cucumber leaves. (i–l), Expression of CsICE1, CsDREB1A, CsDREB1B, and CsCOR47, respectively; (m) DREB1A protein accumulation. The self-grafted and grafted cucumbers plants were treated at 8/5°C for 72 h. Data represent the mean values of four biological replicates, with standard deviations (± SDs) included. With a significance level of P < 0.05, different letters differ significantly between samples. SA participated in rootstock–scion communication of grafted cucumbers Compared with the WT/WT plants, SA content and PAL activity in the leaves and roots of RNAi-PAL/RNAi-PAL transgenic plants were significantly decreased at both room temperature and low temperature, but the OE-PAL/OE-PAL plants exhibited obviously higher PAL activity and SA content. At normal temperature (25°C), the WT/RNAi-PAL and WT/OE-PAL leaves as well as the OE-PAL/WT and RNAi-PAL/WT roots showed similar CsPAL mRNA abundance, PAL activity, and SA content to the WT/WT in leaves and roots. However, the CsPAL mRNA abundance, PAL activity, and SA content in WT/OE-PAL roots and OE-PAL/WT leaves significantly exceeded those of WT/WT roots and leaves, but the three parameters in the WT/RNAi-PAL roots and RNAi-PAL/WT leaves were significantly less than those of WT/WT roots and leaves, respectively. After 12-h exposure at 5°C, the CsPAL mRNA abundance, PAL activity, and SA content of all grafting combinations were accumulated significantly, which in the WT/RNAi-PAL leaves were 19.9%, 14.4%, and 16.5% lower than those of the WT/WT leaves, respectively. However, these three indices in the WT/OE-PAL leaves were 67.5%, 19.8%, and 25.9% higher than those in the WT/WT leaves, respectively (Fig. 2a, d, f). Notably, these three indices markedly decreased in roots of WT/RNAi-PAL at both 25°C and 5°C, they strikingly increased in roots of WT/OE-PAL (Fig. 2b, e, g), while there is no significant difference among the OE-PAL/WT, RNAi-PAL/WT, and WT/WT roots. In order to investigate the presence of long-distance transmission of SA between rootstock and scion, we analysed the variation in SA content in xylem sap. As shown in Fig. 2c, the rise of SA content in xylem sap was significantly greater in the OE-PAL/OE-PAL, WT/OE-PAL, and OE-PAL/WT plants, while it was noticeably lower in the RNAi-PAL/RNAi-PAL, WT/RNAi-PAL, and RNAi-PAL/WT plants than in WT/WT plants. According to these findings, SA signal participates in the rootstock-scion communication of grafted cucumber, which is dependent on the PAL pathway. Figure 2 SA is involved in rootstock-scion communication under cold stress. (a–c) SA content in leaves, roots and xylem sap, respectively; (d, e)  CsPAL expression in leaves and roots, respectively; (f, g) PAL activity in leaves and roots, respectively. Samples were taken from the third leaf of plants at the three-leaves stage following chilling treatments at 0 h and 12 h. Data represent the mean values of four biological replicates, with standard deviations (± SDs) included. With a significance level of P < 0.05, different letters differ significantly between samples. CsNPR1 participated in Cm rootstock-induced cold tolerance To explore the possible molecular regulated mechanism of SA associated with enhanced cold resistance, we performed transcriptome analysis of Cs/Cs and Cs/Cm leaves before and after 12 h of being subjected to cold stress (5°C). Among the 912 differentially expressed genes (DEGs) between Cs/Cs and Cs/Cm plants at 25°C, 459 genes exhibited an upregulation in expression, whereas 453 genes showed a downregulation in expression in Cs/Cm leaves. At 5°C, the DEGs add to 6110, include 2774 upregulated genes and 3336 downregulated genes in Cs/Cm leaves (Fig. S5a and b, see online supplementary material). Our previous research has analysed the changing level of SA synthesis related genes, such as PAL, ICS, and SABP2 in grafted cucumber under cold stress [34], and here we mainly analysed the SA signaling pathway associated genes and found that 9 out of 10 genes were significantly up-regulated (Fig. S6a, see online supplementary material), as determined by the qRT-PCR assays of some key genes of SA-related signal transduction, the relative expression levels of NPR, TGA, and PR in Cs/Cm plants exhibited a considerable increase compared to those in Cs/Cs plants at 5°C, corroborating the transcriptome data results. The expression of NPR (CsaV3_3G040450; CsaV3_4G007550) in Cs/Cm plants response to low temperature was upregulated by 3.15-fold and 2.50-fold, and the expression of TGA (CsaV3_3G033620; CsaV3_3G040130) was upregulated by 1.21-fold and 1.37-fold, indicating that NPR was more responsive to low temperature in Cs/Cm plants (Fig. S5c–e, see online supplementary material). NPR1 is established as a central element in the signal transduction pathway for SA [41], which serves as the primary factor in this pathway [42] and requires interaction with other TFs to control the downstream genes expression. Previous studies demonstrated that NPR1 could collaborate with TGA to trigger PRs expression [43]. More importantly, we also found other up-regulated TFs genes, such as ICE1, which was a crucial element of chilling signal transduction pathway, implying that a new NPR1-mediated possible signal transduction pathway of SA exists in grafted cucumber under chilling stress. Thus, NPR1 (CsaV3_4G007550) was chosen as the target gene to be further studied in the next experiment, and we found CsNPR1 protein sequences have conserved BTB/POZ domain, ankyrin repeats, and NPR1/NIM1-like defense protein C-terminal domain, which is consistent with Arabidopsis thaliana (Fig. S6b, see online supplementary material). To determine whether CsNPR1 participates in cold tolerance induced by rootstock, we measured the change of CsNPR1 mRNA abundances in Cs/Cs and Cs/Cm leaves exposed to 5°C for 0–12 h. Cold stress significantly induced the mRNA expression levels of CsNPR1, with upregulation being most pronounced after 6 h of treatment. The Cs/Cm leaves exhibited higher CsNPR1 mRNA abundance than the Cs/Cs leaves (Fig. 3a). In normal conditions, the leaves of cucumber pretreated with SA and OE-PAL/OE-PAL, WT/OE-PAL, and OE-PAL/WT transgenic plants displayed an increase in CsNPR1 expression, while Cs/Cs and Cs/Cm plants pretreated with 2-aminoindan-2-phosphonic acid (AIP) and AOPP as well as the RNAi-PAL/RNAi-PAL, WT/RNAi-PAL, and RNAi-PAL/WT plants had no difference in CsNPR1 mRNA abundance, compared with H2O treatment and WT/WT plants, respectively (Fig. S7, see online supplementary material). Under chilling stress, SA significantly increase the CsNPR1 mRNA and protein expression levels in both Cs/Cs and Cs/Cm plants, while the treatment of AIP or AOPP down-regulated the CsNPR1 mRNA and protein expressions, compared with the H2O treatment, but the CsNPR1 mRNA and protein expressions in each treatment of Cs/Cm leaves was always significantly higher than that of Cs/Cs leaves (Fig. 3b and d). Moreover, the CsNPR1 mRNA abundance in OE-PAL/OE-PAL, WT/OE-PAL, and OE-PAL/WT leaves was higher compared to WT/WT leaves, the RNAi-PAL/RNAi-PAL, WT/RNAi-PAL, and RNAi-PAL/WT leaves was lower compared to WT/WT leaves (Fig. 3c). Likewise, a similar result was observed at the protein level of CsNPR1 (Fig. 3e). As determined by the cell-free protein degradation assay, proteasome inhibitor (MG132) markedly inhibited CsNPR1 protein degradation, with SA treatment repressing and AIP and AOPP treatment promoting this process. CsNPR1 degradation was greatly slower in Cs/Cm leaves than in Cs/Cs leaves in all treatments (Fig. 3f). Compared to WT/WT plants, the OE-PAL/OE-PAL, WT/OE-PAL, and OE-PAL/WT improved the stability of CsNPR1 protein in vitro, while the RNAi-PAL/RNAi-PAL, WT/RNAi-PAL, and RNAi-PAL/WT accelerated the CsNPR1 degradation (Fig. 3g). These findings revealed that CsNPR1 plays a role in grafting-induced cold tolerance and exogenous or endogenous overexpressed SA-induced CsNPR1 protein expression in Cs/Cm plants under cold stress. Figure 3 CsNPR1 participates in cold tolerance induced by SA and grafting. (a) Response of CsNPR1 mRNA abundance to cold stress (5°C) in Cs/Cs and Cs/Cm leaves. (b)  CsNPR1 expression in Cs/Cs and Cs/Cm leaves subjected to cold stress as affected by SA, AIP, and AOPP. The three-leaves seedlings were sprayed with 1.0 mM SA, 30 μM AIP, 0.1 mM AOPP, or distilled water (H2O, control). Twenty-four hours later, it was exposed to 5°C for 24 h. (c)  CsNPR1 expression in CsPAL transgenic self-grafted and grafted cucumbers at 5°C. Data represent the mean values of four biological replicates, with standard deviations (± SDs) included. With a significance level of P < 0.05, different letters differ significantly between samples. (d) CsNPR1 protein expression in Cs/Cs and Cs/Cm leaves at 5°C. The seedlings were treated in the same way as in (b). (e) CsNPR1 protein in CsPAL transgenic self-grafted and grafted cucumbers at 5°C. (f) Cell-free degradation experiment of recombinant CsNPR1-HIS protein incubated. The seedlings were pretreated with 1.0 mM SA, 30 μM AIP, 0.1 mM AOPP, or distilled water (H2O, control), respectively 24 h before exposure to cold of Cs/Cs and Cs/Cm plants. (g) Cell-free degradation experiment of recombinant CsNPR1-HIS protein in CsPAL transgenic self-grafted and grafted cucumbers at 5°C. Then, the proteins were incubated with CsNPR1-HIS for 0, 0.5, 1.0, and 1.5 h. The proteasome inhibitor MG132 was used as control. Anti-His antibody immunoblotting was used to detect the level of proteins in the sample. The protein level at 0 hour was set to 1, and RBC was utilized as actin. CsNPR1 positively regulated grafted cucumber cold tolerance To further explore the role of CsNPR1 response to cold stress in grafted cucumber, the Anti-CsNPR1 transgenic plants were obtained (Fig. S8, see online supplementary material) and overexpressed CsNPR1 transgenic leaves as scion of Cs/Cs and Cs/Cm using Agrobacterium-mediated genetic transformation and Agrobacterium-mediated transient transformation, respectively. The relative CsNPR1 mRNA expression was significantly higher in Anti-CsNPR1/Cm and OE-CsNPR1/Cm leaves than in Anti-CsNPR1/Cs and OE-CsNPR1/Cs leaves under cold stress (Fig. S9a and b, see online supplementary material). We selected the transgenic line of Anti-CsNPR1–2 and analysed the effect of suppressed expression of NPR1 (Anti-CsNPR1) on the cold tolerance of Anti-CsNPR1/Cs and Anti-CsNPR1/Cm plants. As shown in Fig. 4a–d, the plant phenotype, H2O2 and O2·− accumulation or EL between Anti-CsNPR1 and WT plants have no significant differences at optimum growth temperature (25/18°C). Under cold stress, all plants exhibited wilting and a significantly increased H2O2 and O2·− accumulation and EL, but the Anti-CsNPR1 plants exhibited severe water loss and considerably higher H2O2 and O2·− accumulation and EL than the WT plants. The H2O2, O2·− and EL in Anti-CsNPR1 leaves were elevated by 27.1%, 562.8%, and 122.2%, respectively, and those in WT leaves were elevated by 11.1%, 361.7%, and 80.7%, respectively. The Anti-CsNPR1/Cm plants exhibited significantly lighter cold damage as well as the H2O2 and O2·− accumulation and EL in Anti-CsNPR1/Cm plants elevated by 7.38%, 140.7%, and 68.1%, but that in Anti-CsNPR1/Cs plants declined by 10.1%, 151.2%, and 62.9%, respectively, under chilling stress. H2O2 and O2·− accumulation was also analysed via DAB and NBT staining, and the results corroborated the biochemical measurement results (Fig. 4h). In addition, the expression of CsICE1, CsDREB1A, and CsCOR47 significantly upregulated at 5°C. While the Anti-CsNPR1 leaves displayed a strikingly lower cold-responsive genes expression than the WT leaves, the Anti-CsNPR1/Cm plants had markedly higher values than the Anti-CsNPR1/Cs plants (Fig. 4e–g). The DREB1A protein level corroborated its mRNA expression result (Fig. 4i). Figure 4 Suppression expression of CsNPR1 reduces the cold tolerance. (a) Phenotype of seedlings; (b–d) accumulation of H2O2 and O2·− and EL; (e–g) expression of CsICE1, CsDREB1A, and CsCOR47; (h) NBT and DAB staining of H2O2 and O2·−; (i) DREB1A protein. The leaves were sampled after chilling treatment at 0 h and 24 h. Data represent the mean values of four biological replicates, with standard deviations (± SDs) included. With a significance level of P < 0.05, different letters differ significantly between samples.Then, we examined the impact of overexpression of NPR1 (OE-CsNPR1) on the cold tolerance of cucumber plants. The H2O2, O2·−, and MDA content in the empty vector control (WT) leaves increased by 201.9%, 258.3%, and 105.8% at 5°C, respectively, but that in OE-CsNPR1 leaves increased by 195.7%, 166.7%, and 66.6%, respectively. OE-CsNPR1 was found to markedly decrease ROS accumulation and MDA content in comparison to the WT (Fig. 5a–d). Moreover, the mRNA abundances of cold-responsive genes in the OE-CsNPR1/Cs and OE-CsNPR1/Cm were detected. CsICE1, CsDREB1A, and CsCOR47 expression levels in OE-CsNPR1/Cm leaves were upregulated by 16.8-, 7.4-, and 11.3-fold, respectively, and in OE-CsNPR1/Cs leaves were upregulated by 12.4-, 4.1-, and 7.3-fold, respectively, following at 5°C for 12 h (Fig. 5e–g). Overall, these findings demonstrate that CsNPR1, as a positive regulator of cold response, participates in grafting-induced cold tolerance. Figure 5 Overexpression of CsNPR1 improves the cold tolerance of Cs/Cs and Cs/Cm cucumbers. (a) DAB and NBT staining of H2O2 and O2·−, respectively at 5°C; (b–d) Accumulations of H2O2, O2·− and MDA at 25°C and 5°C, respectively; (e–g)  ICE1, DREB1A, and COR47 expression at 25°C and 5°C, respectively. Empty vector (WT)/Cs, OE-CsNPR1/Cs, and WT/Cm, OE-CsNPR1/Cm transient transgenic cucumbers were treated at 5°C for 12 h. Data represent the mean values of four biological replicates, with standard deviations (± SDs) included. With a significance level of P < 0.05, different letters differ significantly between samples. CsICE1 overexpression improved the cold stress response of grafted cucumbers ICE, a key TF of the cold signaling pathway, is crucial for regulating plant response to cold signals. Among the upregulated genes, an ICE gene named CsICE1 (CsaV3_3G027730) was induced by Cm (Figs S1e and S5b, f, see online supplementary material). Because SA treatment and OE-CsNPR1 upregulated CsICE1 expression (Fig. 5e; Fig. S2d, see online supplementary material) in Cs/Cm, we then investigated whether CsICE1 participated in the cold tolerance of grafted cucumber. The results reveled that cold stress improved the CsICE1 mRNA level in Cs/Cs seedlings, with pronounced improvement at 6-h treatment (Fig. 6a). In order to further investigate the role of CsICE1 in the cold stress response in grafted cucumber, the Cs/Cs and Cs/Cm plants of OE-CsICE1 transient transgenic leaves as scion were obtained (Fig. S9c, see online supplementary material). After subjecting to 5°C for 12 h, the H2O2, O2·− and MDA content in the WT leaves increased by 57.1%, 97.3%, and 80.6%, respectively, but that in OE-CsICE1 leaves only increased by 36.9%, 77.8%, and 64.5%, respectively (Fig. 6b–d). Compared with WT, the mRNA abundance of CsDREB1A and CsCOR47 in OE-CsICE1/Cm leaves was upregulated by 11.2- and 3.60-fold, but that in OE-CsICE1/Cs leaves only increased by 7.58- and 2.41-fold, respectively, under chilling stress (Fig. 6e and f). Overall, these findings indicate that CsICE1 positively modulates the cold tolerance of grafted cucumber. Figure 6 Overexpression of CsICE1 improves the cold tolerance of Cs/Cs and Cs/Cm cucumbers. (a)  CsICE1 expression in grafted cucumbers under cold stress. (b–d) MDA content, H2O2 and O2·− accumulation at 25 and 5°C, respectively; (e, f)  DREB1A and COR47 expression at 25 and 5°C, respectively. Empty vector (WT)/Cs, OE-CsICE1/Cs, and WT/Cm, OE-CsICE1/Cm transient transgenic cucumber plants were treated at 5°C for 12 h. Data represent the mean values of four biological replicates, with standard deviations (± SDs) included. With a significance level of P < 0.05, different letters differ significantly between samples. CsNPR1 directly interacted with CsICE1 Considering that CsNPR1 and CsICE1 exhibit similar expression patterns under cold stress and that ICE1 is crucial for plants to respond to cold signal, we hypothesize that CsNPR1 may interact with CsICE1 to contribute to grafting-induced cold tolerance. An in vitro detection experiment was conducted based on yeast two-hybrid (Y2H) analysis. To avoid CsNPR1 auto-toxication, we deleted the 3′ region of CsNPR1 (561aa–585aa) to construct the CsNPR1-pGBKT7* vector (the trimmed CsNPR1 sequence is shown in Table S1, see online supplementary material) and used 150 ng/ml Aureobasidin A (AbA) to inhibit CsNPR1–pGBKT7* auto-activation (Fig. S10, see online supplementary material). As determined by the Y2H assay results, only the co-transformed combination with CsICE1 and CsNPR1 yeast cells grew on SD (−Trp/−Leu/-His/−Ade/+X-α-gal) media, whereas other combinations all failed to induce growth on the same media (Fig. 7a), indicating that CsNPR1 can directly interact with CsICE1. To verify this result, in vitro pull-down assay was performed. As shown in Fig. 7b, the CsNPR1-GST fusion protein was bound by CsICE1-HIS, while the GST control did not bind. In vivo dual-luciferase assays conducted to validate CsNPR1–CsICE1 interaction revealed a fluorescence signal in the co-transformation region of CsNPR1 and CsICE1 (Fig. 7c). The interaction between CsNPR1 and CsICE1 was further confirmed using a bimolecular fluorescence complementation (BiFC) assay (Fig. 7d). These findings establish that CsNPR1 can directly interact with CsICE1 in vitro and in vivo to modulate the cold stress response in grafted cucumber. Figure 7 CsNPR1 directly interacts with CsICE1. (a) The Y2H experiment is used to determine the interactions between CsNPR1 and CsICE1., the co-transformation yeast cells were plated on SD/−Trp-Leu (SD/−T-L) and SD/−Trp-Leu-His-Ade (SD/−T-L-H-A) media with 150 ng·ml−1 AbA for 2 d, and 10−0 and 10−1 represent the dilutions of yeast solution. The positive interactions were further confirmed using X-α-gal. (b) Pull-down assay to identify the in-vitro interaction of CsNPR1 and CsICE1. The CsICE1-HIS proteins were respectively incubated with CsNPR1-GST or GST proteins, and then pulled down by GST purification kit. Immunoblotting using anti-GST and anti-His antibodies allowed for the detection of the eluted proteins. (c) Luciferase complementarity assay to identify the in-vivo interactions of CsNPR1 with CsICE1 in tobacco. (d) BiFC assay showing the in-vivo interactions of CsNPR1 with CsICE1 in tobacco. Bars = 50 μm. Similar results were seen in all three of the experiments. CsNPR1–CsICE1 interaction promoted the rootstock-induced cold tolerance of cucumber To verify the role of the CsNPR1–CsICE1 interaction in regulating cold tolerance, we obtained the Cs/Cs and Cs/Cm plants of CsNPR1 and CsICE1 co-expressed (OE-CsNPR1–OE-CsICE1) transient transgenic leaves as scion (Fig. S11, see online supplementary material). OE-CsNPR1–OE-CsICE1 significantly decreased the accumulation of H2O2 and O2.− in detached leaves induced by cold stress, as determined by NBT and DAB staining, in comparison to OE-CsICE1. The decrease in H2O2 and O2.− accumulation in OE-CsNPR1–OE-CsICE1/Cm was markedly greater than that in OE-CsNPR1–OE-CsICE1/Cs under cold stress (Fig. 8a and b). OE-CsNPR1–OE-CsICE1 co-expressed leaves showed much higher mRNA abundances of the key downstream genes of ICE1, including CsDREB1A, CsDREB1B, CsCOR47, CsCOR15, CsWCOR413, and CsKIN1 than OE-CsICE1. Compared with OE-CsNPR1–OE-CsICE1/Cs leaves, a greater improvement in the ICE1 downstream gene mRNA level was observed in OE-CsNPR1–OE-CsICE1/Cm leaves at 5°C (Fig. 8c–h). Overall, these results suggest that the CsNPR1–CsICE1 interaction positively regulates the rootstock-induced cold tolerance of cucumber. Figure 8 Co-overexpression of CsNPR1 and CsICE1 improves the cold tolerance of Cs/Cs and Cs/Cm cucumbers. (a, b) DAB and NBT staining of H2O2 and O2·−, respectively; (c–h) the expression of DREB1A, DREB1B, COR15, COR47, COR413, and KIN1 of empty vector (WT), OE-CsICE1 and OE-CsICE1-CsNPR1 transient transgenic cucumber leaves, respectively, at 25 and 5°C for 9 h. Data represent the mean values of four biological replicates, with standard deviations (± SDs) included. With a significance level of P < 0.05, different letters differ significantly between samples. Discussion Plants’, especially solanaceous fruit and melon vegetables, response to abiotic stresses including heat, cold, and salinity can be enhanced by grafting with resistant rootstocks. In this study, grafted seedlings showed lower CI, EL, and MDA content, as well as higher expression of chilling response genes, such as DREB1A and ICE1 (Fig. S1, see online supplementary material), corroborating previous findings [7, 34]. Despite the importance of long-distance signaling molecules like phytohormones, water and mineral ion transport, osmotic substances, miRNAs, and proteins in rootstock–scion interaction, how these signaling molecules, especially phytohormones signals, participate in regulating abiotic stress resistance in grafting plants remains unclarified. SA is a crucial phytohormone that plays a significant role in the plant resistance to chilling stress [28]. Here, we also discovered that SA could alleviate damage to the cucumber photosystem and stimulate the expression of cold responsive genes (ICE1, DREB1A, COR47) in both Cs/Cs and Cs/Cm plants when exposed to cold stress (Fig. S2, see online supplementary material). The application of SA synthesis inhibitors AIP and AOPP dramatically reduced the chilling tolerance of cucumbers enhanced by grafting, highlighting SA involvement in grafting-induced cold tolerance. Although SA, as a long-distance transduction signaling molecule, could participate in rootstock–scion information exchange [44, 45, 46], the molecular verification of SA function during this process was limited. Previous research has informed that the SA synthesis is approximately 90% dependent on the ICS pathway in Arabidopsis, with the PAL pathway producing the isomer of SA, 4-hydroxybenzoic acid, instead of SA [46]. However, different SA synthesis pathways differ greatly in different species. Under low temperature, the SA content and the AtICS expression had no significant changes in Arabidopsis for 7 days [47, 71], whereas the SA content was significantly accumulated at 1.4- and 2.2-fold in cucumber and watermelon within 1 day, respectively [27, 48]. The PAL pathway has been evidenced as the main pathway for SA synthesis in cucumbers responding to chilling stress [27, 28, 34]. Thus, we obtained CsPAL transgenic cucumber plants and found that CsPAL overexpression markedly increased SA content, whereas CsPAL inhibition declined SA content (Fig. S3, see online supplementary material). Notably, OE-PAL/OE-PAL with CsPAL overexpression not only alleviated chilling damage of ungrafting seedlings, as indicated by lower EL, MDA content, ROS accumulation, CI, and relative mRNA expression of CsICE1, CsDREB1A, CsDREB1B, and CsCOR47 in comparison to WT, but also promoted the cold tolerance of WT/OE-PAL and OE-PAL/WT grafting seedlings. RNAi-PAL with CsPAL inhibition showed the opposite results (Fig. 1). To further explore SA participation in rootstock–scion interaction, we measured the change in CsPAL mRNA abundance, PAL activity, and SA content in both the leaves and roots of grafted and ungrafted transgenic plants. It was demonstrated that at both room temperature and low temperature, the OE-CsPAL plants shown significantly higher CsPAL mRNA abundance, PAL activity, and SA content in the leaves and roots. In accordance with this, these three indices for WT/OE-PAL in both leaves and roots exceeded those for the WT/WT leaves and roots, respectively, after 12-h chilling stress. However, RNAi-PAL decreased PAL activity and SA content in both leaves and roots of WT/RNAi-PAL, compared with WT/WT seedlings. It means the SA decline in rootstock caused by CsPAL inhibition is the main cause for the weakening of chilling tolerance in WT/RNAi-CsPAL seedlings. Notably, the SA content of xylem sap in WT/OE-PAL was significantly higher than that in WT/WT plants and WT/RNAi-PAL was lower than WT/WT plants (Fig. 2). Furthermore, these three indices of roots in RNAi-PAL/WT and OE-PAL/WT had no difference, demonstrating that the SA dependence on the PAL pathway in rootstock-scion communication played the vital role in grafting-induced chilling tolerance of cucumber. SA signal transduction is primarily initiated by the specific binding between SA and its receptor NPR1, after which NPR1 interacts with TGA, WRKY, NIMIN, or other TFs to activate the downstream genes’ expression during plant immunity [49]. Transcriptome analysis showed significant difference in SA signal transduction-related genes expression between Cs/Cs and Cs/Cm plants at low temperatures. This indicates that the SA signaling pathway takes part in the cold tolerance of cucumber induced by grafting (Fig. S5, see online supplementary material). Considering the significance of NPR1 in the SA signaling pathway, the change in CsNPR1 attracted our attention. Chilling stress was previously shown to upregulate AtNPR1 expression in Arabidopsis [50]. In our study, the data of real-time PCR verification and further KEl pathway analysis of the SA signal transduction-related DGEs both proved the upregulation of CsNPR1 in Cs/Cm plants, and the nucleotide sequences of CsNPR1 were highly conserved (Figs S5 and 6, see online supplementary material). Moreover, exogenous and endogenous over-expressed SA increased CsNPR1 mRNA and protein expression and reduced CsNPR-HIS degradation in both Cs/Cs and Cs/Cm plants, whereas endogenous SA inhibition blocked this induction effect. However, pumpkin-grafted plants exhibited a higher CsNPR1 level following AIP and AOPP treatments than the Cs/Cs plants under cold stress (Fig. 3), indicating that CsNPR1 expression was mediated by the raising of SA content in scion or grafting under low temperature stress. Cao et al. [71] and Glazebrook et al. [51] revealed that npr1 displayed increased susceptibility to pathogens and decreased pathogenesis-related (PR) gene expression; however, NPR1 overexpression in npr1 could restore the resistance of disease and PR expression in Arabidopsis [72]. The same results were obtained under cold stress. As an example, the cold resistance of npr1 is much weaker than that of WT, and OE-AtNPR1 enhances the cold adaptation of Arabidopsis [37]. In our study, Anti-CsNPR1 as the scion notably reduced the cold tolerance of Cs/Cm seedlings, as proved by the higher ROS content and EL and the lower mRNA abundance of CsICE1, CsDREB1A, and CsCOR47 (Fig. 4), whereas OE-CsNPR1 improved cold tolerance with decreased MDA and ROS accumulation and upregulated cold responsive genes expression in comparison to the WT seedlings (Fig. 5). However, the cold tolerance of cucumber induced by grafting is a very complex process, there are many other signals involved in this process, such as SA, ABA, MeJA, MT, H2O2, mRNA, etc. [3, 7, 52, 53]. These signals not only have function singly, but also interact with each other to regulate graft-induced cold tolerance in cucumber and thus the cold tolerance of Anti-CsNPR1/Cm was still better than that of Anti-CsNPR1/Cs under chilling stress. According to the above results, we speculated that SA-induced CsNPR1 expression could positively participate in the regulation of cold tolerance in grafted cucumber. Due to NPR1 lacking the DNA-binding domain and only acting as a coactivator [54, 55], it needs to interact with other proteins. For example, NPR1 mediated the WRKY18–WRKY60 module to activate ABI4 and ABI5 expression and then positively regulated the ABA signal [56]. Meanwhile, NPR1 induced plant chilling tolerance by interacting with HSFA1 [37]. ICE, a bHLH-type TF, is believed to be a CBF inducer. Under chilling stress, ICE could bind with a CBF promoter, hence upregulating the downstream genes expression level that respond to cold, such as COR15A, KIN1, COR47, and RD29A, and promoting the cold tolerance of plants [57–59]. Here, we found that CsICE1 overexpression declined MDA and ROS accumulation and improved CsDREB1A and CsCOR47 expression under cold stress, and OE-CsICE1/Cm plants showed lower MDA and ROS accumulation and higher CsDREB1A and CsCOR47 expression than OE-CsICE1/Cs plants (Fig. 6). TF complexes often regulate cold signals more effectively than single TF, and ICE–CBF can be regulated by multifarious TFs or proteins and then take part in the hormone signaling pathway to regulate chilling stress [21, 22]. For instance, SnRK2.6/OST1, which participate in the ABA signal pathway, interacts with ICE1 to yield phosphorylated ICE1, which then activates CBF–COR gene expression, thereby conferring low-temperature resistance to this gene [18]. Because CsICE1 and CsNPR1 exhibited comparable variations under chilling stress (Figs 3a and 6a; Fig. S5, see online supplementary material), we speculated whether the CsICE1–CsNPR1 interaction existed under chilling stress. Fortunately, Y2H, pull-down, luciferase complementation and BiFC experiments proved that CsNPR1 actually interacted with CsICE1 (Fig. 7), and the co-transformation of CsNPR1 and CsICE1 to cucumbers further demonstrated that the CsNPR1–CsICE1 module participated in regulating cold tolerance in grafted cucumber seedlings. Compared to WT, grafting seedlings with single CsICE1 overexpression showed lower MDA and ROS accumulation and higher mRNA expression of CsDREB1A, CsDREB1B, CsCOR47, CsCOR15a, CsWCOR413, and CsKIN1. Notably, co-transforming CsNPR1 with CsICE1 notably alleviated ROS accumulation and enhanced the downstream chilling-sensitive genes expression (Fig. 8). Conclusion In conclusion, SA was involved in regulating the cold tolerance of grafted cucumber which is induced by using pumpkin as the rootstock. The SA content increase in Cs/Cm seedling leaves, which was largely dependent on the PAL pathway, was not only synthesized in leaves but also transported upward from the rootstock under chilling stress. Cold stress stimulated the CsNPR1 and CsICE1 expression and enhanced cold tolerance in grafted cucumber. More importantly, the SA signal in scion also upregulated the CsNPR1–CsICE1 module to positively mitigate lipid peroxidation damage and promote the cold responsive genes expression, resulting in improving the cold tolerance of grafted cucumber (Fig. 9). Materials and methods Plant materials and growth conditions Cucumber (C. sativus L., cv. ‘Jinyou 35’, Cs), pumpkin (Cucurbita moschata D., cv. ‘Jinmama 519’, Cm), phenylalanine ammonia-lyase (PAL) gene overexpression, and RNAi-silenced transgenic cucumber lines were used as the experimental materials. The transgenic plant lines were obtained using the Agrobacterium tumefaciens genetic transformation system, and the cucumber cultivar ‘Xintaimici’ was utilized as the wild type (WT). The grafting procedure was conducted using the top insertion method once the cotyledons of both the rootstock and scion seedlings had fully expanded. The resulting self-root and hetero-root grafted seedlings, as well as the transgenic plants, were cultivated in an artificial climate chamber with the following conditions: temperature of 26/18°C (day/night), photon flux density (PFD) of 600 μmol m−2 s−1, photoperiod of 11/13 h (light /dark), and relative humidity of 80%. Chilling and SA treatments All experiments were conducted in an artificial climate chamber to investigate the impact of grafting on cucumber cold tolerance. Two types of cucumber plants were used: the self-rooted (cucumber as rootstock and scion, Cs/Cs) and grafted cucumber plants (with pumpkin as rootstock, Cs/Cm). These plants were treated at 8°C/5°C for 24, 72, and 120 h, with those treated at 25°C/18°C serving as controls. The chilling injury index (CI), reactive oxygen (ROS) accumulation, and cold-responsive genes expression were measured at 72 h. Figure 9 A possible model of SA involved in rootstock-scion communication to improving the cold tolerance of grafted cucumber. The yellow solid circle means SA synthesized from leaf and the red solid circle means SA synthesized from root.To determine the role of SA in grafting-induced cold tolerance, Cs/Cs and Cs/Cm plants were sprayed with SA (1.0 mM), SA synthesis inhibitors 2-aminoindan-2-phosphonic acid (AIP, 0.03 mM), L-2-aminooxy-3-phenylpropionic acid (AOPP, 0.1 mM), or distilled water (control) for 24 h and then the plants were treated at 5°C for 48 h to determine the photosynthetic rate (Pn) and RCA, rbcL, and cold-responsive genes expression. Subsequently, WT/WT, overexpressed CsPAL (OE-PAL/OE-PAL), and RNAi (RNAi-PAL/RNAi-PAL) self-grafted transgenic plants and WT/OE-PAL, WT/RNAi-PAL, and OE-PAL/WT, RNAi-PAL /WT grafted plants were exposed to cold conditions for 12 or 72 h in order to assess the electrolyte leakage (EL), malondialdehyde (MDA) content, hydrogen peroxide (H2O2) and superoxide anion (O2.−) contents, cold-responsive genes expression, as well as SA content, PAL activity, and CsPAL expression. At the two-leaf stage, the transgenic plants of Anti-CsNPR1/Cs and Anti-CsNPR1/Cm were subjected to 8°C/5°C for 24 h, with those plants subjected to 25°C/18°C serving as controls. The EL, MDA content, ROS accumulation, and the cold-responsive genes expression were determined. Meanwhile, the pBI121-GFP empty vectors, as well as overexpressed CsNPR1 and CsICE1 vectors were injected into the cotyledon leaves of Cs/Cs and Cs/Cm seedlings. After a 12-h dark incubation, half of the WT and overexpressed leaves of CsNPR1 and CsICE1 were treated at 5°C and the other half were treated at 25°C. The cold-responsive genes expression, MDA content, and ROS accumulation were determined after chilling for 9 or 12 h. Measurement of CI, EL, MDA, and ROS accumulation The CI was estimated using the method in [60]. The calculation formula is as follows: CI = Σ (plants of different grade × grade)/[total plants × 5 (the maximum grade)]. Grades in the formula represent the cold damage degree of plants. The EL was determined using the approach [61]. The measurement of electrical conductivity (EC) was conducted with a DDB-303A conductivity meter (Lei—ci, Shanghai, China). The initial EC and final EC (A/B) were used to calculate EL, where A is the EC detected after incubating leaves in distilled water at 25°C for 3 h, and B is the EC detected after boiling for 30 min and cooling to room temperature. The MDA content was determined following the described method by Zhao and Cang [62] using the thiobarbituric acid (TBA) colorimetric approach. For H2O2 determination, the H2O2 kit (Njjcbio, Nanjing, China) was utilized by the guideline description provided. The levels of O2.− content were assessed following the method outlined by Li and Gong [63]. Qualitative analysis of H2O2 and O2.− were conducted using a 3, 3-diaminobenzidine (DAB) kit (Solaribio, Beijing, China) and nitroblue tetrazolium (NBT) (Biotopped, Beijing, China), as outlined in the studies by Thordal-Christensen et al. [64] and Jabs et al. [65]. SA content and PAL activity assay SA content quantification was carried out following the previously described method [34]. In brief, freeze-dried cucumber leaves and roots were extracted with 80% methanol for 16 h, and the resulting supernatants were separated by centrifugation and then evaporated to dryness using a rotary evaporator (EYELA N-1210B, Shanghai, China). Following the elimination of pigments and phenolic impurities using chloroform and polyvinylpolypyrrolidone (PVPP), SA was extracted into ethyl acetate, rotary evaporated until dry, and then dissolved in 50% methanol. The xylem sap was collected beneath the scion cotyledon, and the EP tube was filled with absorbent cotton and closely contacted with the section on the stem. After absorption for 10 h, the absorbent cotton was weighed and leached in 80% methanol for 16 h, then concentrated and extracted with 50% methanol. The extract was filtered via a 0.22 μm filter and analysed using TSQ Quantum Access (Thermo Fisher Scientific, Waltham, MA, USA) for measurement. The PAL activity measurement followed the protocol described by Shang et al. [29]. Transcriptome analysis Cs/Cs and Cs/Cm plants were subjected to cold stress for 12 h, while plants grown at 25°C served as controls. We extracted total RNA from leaves of Cs/Cs and Cs/Cm plants, both before and after chilling stress, and performed transcriptome analyses using the method described by Fu et al. [34]. Each sample was replicated three times. To determine the differential genes, we applied a significant threshold of P-value <0.05 and fold change >1.5. The data of this study have been deposited in the Gene Expression Omnibus of NCBI and can be accessed through SRA accession: PRJNA701131230 (http://www.ncbi.nlm.nih.go v/bioproject/701131). Vector construction and genetic transformation The coding sequence (CDS) of CsPAL, CsNPR1, and CsICE1 was inserted into the pBI121-GFP vector, which contains the CaMV35S promoter, to create overexpression vectors. The pBI121-GFP vector was used to insert a 432-bp fragment sequence of CsNPR1 (Table S2, see online supplementary material) to generate antisense vectors. Similarly, the pBWA(V)KS vector was used to insert a 201-bp fragment sequence of CsPAL (Table S3, see online supplementary material) to generate RNAi-generated vectors. Table S4 (see online supplementary material) provides a comprehensive list of the primers utilized in this work. Using the freeze–thaw procedure, the recombinant vector plasmids were transformed into Agrobacterium LBA4404 (Weidi, Shanghai, China). Peeling cotyledons were used as explants for the Agrobacterium-mediated genetic transformation of cucumbers [66]. To perform transient transformation of cucumber leaves, we selected cucumber leaves that exhibited consistent growth and then injected Agrobacterium containing overexpression vectors of CsPAL, CsNPR1, or CsICE1 into the cucumber leaves using a medical syringe. The transgenic plants were detected by PCR analysis. RNA extraction and qRT–PCR analysis Using an RNA extraction kit (Trizol, TRANs, Beijing, China), total RNA was extracted from plant materials and reverse transcription to cDNA using HiScript® III RT SuperMix (Vazyme, Nanjing, China). The mRNA abundance of cold responsive genes, rubisco large-subunit, transcriptome-related genes, and CsNPR1 in cucumber leaves were assessed by quantitative real-time PCR (qRT–PCR) using the LightCycler® 480II equipment (Roche, Penzberg, Germany). β-actin served as the control. Table S5 (see online supplementary material) includes the primers designed for the genes. Protein degradation and western blot analysis A protein degradation buffer containing 25 mM Tris–HCl (pH = 7.5), 10 mM MgCl2, 10 mM NaCl, 10 mM ATP, 5 mM dithiothreitol (DTT), and 4 mM phenylmethanesulfonyl fluoride (PMSF) was used to extract the total protein from the cucumber leaves. Equal amounts of extracted proteins and Escherichia coli BL21(DE3) (Vazyme, Nanjing, China) induced CsNPR1-HIS recombinant proteins were subjected to incubation at 22°C for the specified sampling times. For protease inhibitor MG132 treatment (as control), cucumber leaf extracts were pretreated with DMSO (dissolve agent of MG132) or 50 μM MG132 for 30 min and then mixed with the CsNPR1–MG132 protein and incubated at the specified sampling times. The relative protein level of CsNPR1 was detected by western blotting (Tanon, Shanghai, China) with anti-His-tag mouse monoclonal antibody (CWBio, Beijing, China). Western blotting was conducted as previously mentioned [34]. Yeast two-hybrid (Y2H) assay The CDS of CsNPR1 (remove self-toxic sequence) was cloned into the pGBKT7 bait vector, the CsICE1 CDS was cloned into the pGADT7 prey vector. The recombinant plasmid combinations were co-transferred into Y2H-sensitive state cells. The transformed yeast strains were grown on SD/−Trp/−Leu medium, and the resulting colonies were transferred to SD/−Trp/−Leu/-His/−Ade medium for the screening of interacting proteins. Table S4 (see online supplementary material) contains the primers used in this experiment. Pull-down assay The insertion of CsNPR1 CDS into the pGEX-4 T-1 vector yielded the glutathione S-transferase (GST)-tag fusion protein, whereas the insertion of CsICE1 CDS into pET-32a (+) yielded the HIS-tag fusion protein. The fusion expression proteins of GST-CsNPR1 and HIS-CsICE1 were produced by 1 M isopropyl-β-d-thiogalactoside (IPTG) induction after these recombinant plasmids were transformed into the E. coli strain BL21 (DE3). The pull-down assays were conducted with the BeaverBeads™ GSH (Suzhou, China) according to the instructions. Table S4 (see online supplementary material) contains the primers used in this experiment. Luciferase complementation imaging assays CsNPR1 and CsICE1 coding regions were cloned into the pCAMBIA1300-cLUC and pCAMBIA1300-nLUC vectors, respectively. Tobacco leaves were injected with GV3101(pSoup) Agrobacterium strain solutions carrying recombinant plasmids using syringes. An imaging system (Tanon, Shanghai, China) was used to create the image. The primers used in this experiment are listed in Table S4 (see online supplementary material). Bimolecular fluorescence complementation assays The CDS of CsNPR1 and CsICE1 were inserted into pSm35s-nYFP-ccdb and PpSm35s-ccdb-cYFP vectors, respectively. The solutions of Agrobacterium strain containing recombinant plasmids were mixed and transmitted into tobacco leaves using syringes. The YFP fluorescence was detected by a confocal laser-scanning microscope (Zeiss LSM 510 Meta, Jena, Germany). Table S4 (see online supplementary material) includes the specific primers used in this experiment. Statistical analysis The data were displayed as means ± standard deviation (SD), and the complete experiment was carried out at least four times. DPS software was used to statistically evaluate all of the data. Duncan’s multiple range tests were used to evaluate the statistical analysis of the data at P < 0.05. Supplementary Material Web_Material_uhae231 Acknowledgements This work is funded by The National Science Foundation of China (31872155, 32372685), The National Key Research and Development Program of China (2019YFD1000300), The Special Fund of Vegetable Industrial Technology System of Shandong Province in China (SDAIT-05–10), and The Funds of Shandong ‘Double Tops’ Program (SYL2017YSTD06). Author contributions X.F conducted the majority of the experiment, analysed the data, and finished the initial draft. X. A. and H.B. were responsible for the research design and paper editing. Y.F. and Y.Z. collaborated with X.F. to carry out the experiment. Data availability All relevant data can be found within the manuscript and its supporting materials.The transcriptome datasets have been uploaded to the Gene Expression Omnibus of NCBI and can be accessed through SRA accession: PRJNA701131230. Conflict of interest statement The authors declare that they have no conflicts of interest. Supplementary data Supplementary data is available at Horticulture Research online.