Annual tea plant cuttings (Camellia sinensis L. cv. ‘Longjing43’) were pre-cultured hydroponically in a growth chamber (16 h light/8 h dark cycle, 25°C/23°C, and 50% humidity). The nutrient solution containing 15% PEG-6000 was used to simulate drought stress. Plants exposed to PEG were treated with 0, 5, 10 and 30 μM GR24 via foliar spray every 48 hours, ensuring complete coverage but not dripping. One bud and one leaf of tea plant were harvested after 1 day (24 h)with the first treatment for subsequent experiment. After 8 days of treatment, the shoots of tea plants were collected for physiological indicator determination.

Two-year-old seedlings of Camellia sinensis L.cv. Shancha NO.1 (SC), Huangjinya (HJY), Longjin43 (LJ43) and Zijuan (ZJ) were grown in the greenhouse of Tea Research Institute of the Chinese Academy of Agricultural Sciences. Tea seedlings with consistent growth were selected and randomly divided into three groups: (1) Control group (CK), no stress and no GR24 treatment; (2) Drought stress group (DS), no watering was applied during the treatment period.; (3) GR24 treatment group (SL), drought treatment with 10 μM GR24 spray application. The CK and DS groups were sprayed with a 3% acetone solution, while the SL group was sprayed with a 10 μM GR24 solution. All treatments were uniformly watered and sprayed with the corresponding solutions, designated as Day 0. Subsequently, the solutions were sprayed every 2 days (48 h) for a total of 10 days. Following 10 days of water deprivation, the shoots were harvested, immediately flash-frozen in liquid nitrogen, and stored at −80°C for subsequent analyses.

Wild-type Arabidopsis thaliana (Columbia-0 ecotype, Col-0) and three independent CsD27-overexpressing T3 generation homozygous lines (OE5-9, OE9-2, OE12-1) were selected and used for further experiments. For the germination assays, the seeds were sown on 1/2MS medium with and without 200 mM mannitol (control). For the drought stress treatment, the 3-weeks-old seedlings of transgenic lines and WT were grown on soil for one week with or without watering (CK treatment and DS treatment, respectively).

The OJIP curves of tea plant were measured at 1, 2, 4, 6 and 8 d after exposure to drought stress and foliar application of GR24 with various concentrations. Chlorophyll α fluorescence was measured after a 30-minute dark treatment using FluorPen FP110 portable fluorometer (Photon Systems Instruments, Czech Republic). Formulae and explanation of the fluorescence parameter used in this study are provided in Supplementary Table 1 .

Fresh leaves were rinsed with deionized water to remove surface contaminants and gently blotted dry with tissues. For each treatment group, six leaves were punched by hole puncher (0.5 cm diameter). The leaf discs were transferred into 50 mL conical tubes containing 15 mL of deionized water. The tubes were shaken (200 rpm) for 1 h to measure the conductivity of the solution, recorded as EC1. Subsequently, samples were boiled for 30 min at 100 °C. Followed by cooling to room temperature, the solution was measured and the results were recorded as EC2. The fresh samples were pulverized in liquid nitrogen, and 0.1 g of each samples was accurately weighed to determined the production rate of MDA and H₂O₂ by using the detection kits (Suzhou Comin Biotechnology Co., Ltd, Suzhou, China).

Total RNA was extracted using the RNA prep Pure Plant kit (TIANGEN biotech CO., Ltd, Beijing, China) and 2 μg of total RNA was used for first-strand cDNA synthesis by the PrimeScript™ RT Reagent Kit (TaKaRa, Dalian). Subsequently, quantitative PCR was performed using SYBR^® Premix Ex Taq™ (TaKaRa, Dalian). Transcript levels were normalized to GAPDH, and the calculation of relative transcript abundance using the 2^−ΔΔCt method. For each treatment, three biological replicates were used to analyze the expression of each gene.

The CDS of CsD27 (excluding the termination codon) were PCR-amplified and cloned into the pBWA(V)HS-GFP fusion vector. The recombinant constructs were introduced into A. tumefaciens through electroporation, followed by transient transformation of N. benthamiana leaf epidermis using syringe infiltration. An empty vector control (35S:GFP) was co-infiltrated under identical conditions. At 2 days post-infiltration, transformed leaf sections were excised and mounted on glass slides to detect the subcellular localization of the CsD27 protein using a confocal laser-scanning microscope (Nikon C2-ER).

The samples were lyophilized with FreeZone freeze dryer (LABCONCO, America) for 4 days and ground with a RETSCH MM400 grinder (Duesseldorf, Germany). To analyze catechin concentrations, 0.1000g of each freeze-dried samples was used to extract following a high-performance liquid chromatography (HPLC) –based method as previously described in detail (Tong et al., 2023).

To further investigate the relationship between SL and drought tolerance in tea plants, we examined the phenotypic characteristics of four cultivars and measured the MDA content of four cultivars after 10 days drought stress and GR24 treatments ( Figure 3 , Supplementary Figure 1 ), revealing different levels of stress response. SC and LJ43 exhibited milder wilting and lower MDA content, while ZJ showed moderate wilting and MDA levels. In contrast, HJY suffered the most severe damage from drought stress, characterized by burning symptoms appearing on wilted shoots and the highest MDA content.

To confirm the impact of SL on tea catechins, the main flavonoid antioxidants, we measured the catechin components in the shoots of drought-insensitive cultivar LJ43 and drought-sensitive cultivar HJY under drought stress ( Figure 4 ). Generally, epigallocatechin gallate (EGCG) was the most abundant catechin, followed by epigallocatechin (EGC). In the drought-insensitive cultivar which showed no visible symptoms of drought stress, there were no significant differences in catechin components across the three treatments. However, HJY showed a significantly decrease in both total catechin and individual catechin components. Consistent with the phenotypic observations and MDA content, the application of GR24 alleviated drought-induced damage and promoted catechin accumulation, indicating that SL positively influenced tea quality under drought stress.

D27 has been confirmed to be involved in the plant response to abiotic stress across multiple species. Previous studies have shown that the drought resistance of D27 knockout mutants is significantly reduced, while overexpressing lines exhibit the opposite result (Haider et al., 2018). Given the central role of SL in drought response based on the above observation, we identified 8 CsD27, the key SL biosynthesis genes, from tea plant genome of Camellia sinensis L.cv. Shuchazao ( Supplementary Figure 2 ). Specifically, the member (CSS0025398.1) was regulated in Camellia sinensis L.cv. Tieguanyin by long-term drought stress as revealed in previous database (unpublished). Additionally, we measured the expression levels of CsD27 across four tea plant cultivars. While CsD27 exhibited higher expression levels in SC and LJ43 (drought-tolerant cultivars), its expression was significantly reduced in both HJY and ZJ (drought-sensitive cultivars) ( Figure 5B ). Correlation analysis showed a significant and negative correlation between expression of CsD27 and MDA content, indicating that cultivars with higher expression levels exhibited higher drought tolerance ( Figure 5C ). Therefore, we cloned the CsD27 gene (CSS0025398.1) and performed a multiple sequence alignment, followed by the construction of a phylogenetic tree ( Figures 5A, D ). The results demonstrated that CsD27 shared conserved regions with D27 proteins from other species, and exhibited the highest similarity to SlD27 and GmD27.

In order to identify the subcellular localization of CsD27 protein, the 35S:CsD27-GFP plasmid was introduced into tobacco leaves. The results showed that strong GFP fluorescence of the 35S:CsD27-GFP fusion protein was predominantly observed in the chloroplast, indicating that CsD27 is mainly localized in the chloroplast, consistent with the predicted localization ( Figure 5E ).

To investigate the tissue-specific expression pattern of CsD27, total RNA from various tea plant tissues including root, stem, mature leaves, the first leaf of shoot and bud were extracted and used for qPCR ( Figure 5F ). The results showed that the expression level of CsD27 was higher in mature leaves than other tissues, while exhibited extremely low expression levels in both buds and stems.

Meanwhile, CsD27 was differentially expressed under stress and hormone treatment. The expression of CsD27 after 24 hours of DS treatment slightly decreased compared to CK in LJ43. Under GR24 treatment, the expression of CsD27 transcript was effectively upregulated, with a 2.2-fold increase at 5 μM, a maximum induction of 2.4-fold at 10 μM, and an increase of approximately 1.7-fold at 30 μM ( Figure 5G ). These results implied that CsD27 might be a crucial gene in response to drought, and the regulation networks were probably related to GR24 concentration.

To further investigate its function in drought stress tolerance, we developed three transgenic Arabidopsis lines with CsD27 overexpression (OE5-9, OE9-2, OE12-1). The result indicated that the expression levels of CsD27 were significantly higher in the transgenic lines than in the wild-type (WT) ( Figure 6A , Supplementary Figure 3 ).

Seeds of each genotype were planted on 1/2 Murashige and Skoog (MS) media containing 0 or 200 mM mannitol to assess the seed germination rates under osmotic stress. Although germination and cotyledon greening rates were comparable between WT and transgenic lines under 200 mM mannitol treatment, the transgenic lines exhibited larger cotyledon areas and enhanced growth performance compared to WT seedlings under osmotic stress conditions ( Figures 6C, D ). Additionally, we selected three-week-old seedlings of each genotype to expose to one week of drought stress to investigate the role of CsD27 in drought response of Arabidopsis. Compared to WT, the majority of overexpression line seedlings retained green coloration and exhibited better growth under drought stress, as evidenced by less severe wilting phenotype and significantly higher fresh weight ( Figures 6E, F ). Similarly, the MDA result was demonstrated that OE5-9, OE9–2 and OE12–1 exhibited significant reduction of drought-induced damage, which was consistent with the phenotypic observation ( Figure 6B ). After 3 days of re-watering, the OE5-9, OE9–2 and OE12–1 seedlings exhibited superior recovery, whereas WT seedlings showed only 25% survival rates ( Figure 6C ). These results indicated that CsD27 enhanced the drought tolerance of Arabidopsis.

Water deficit ​induces multilevel photosynthetic dysfunction, ​manifested through stomatal closure, ​structural disorganization of thylakoid membranes, disruption of photosynthetic pigments, and degradation of photosynthetic proteins (Vanlerberghe et al., 2016; Razi and Muneer, 2021). The ΔV_t in tea plant revealed that drought stress induced the emergence of a distinct phase in the OJIP kinetic curve, the J peak ( Figure 1D ) caused by the accumulation of Q_A ^- in PSII reaction centers (RCs), indicating that drought stress disrupted the electron flow beyond Q_A (Guo et al., 2020; Tsimilli-Michael, 2020; Jiang et al., 2025). A significant increase in M_o and V_j under drought stress further corroborated this result ( Figure 1B ). Whereas, the GR24-treatments efficiently improved the probability of PSII photosynthetic electron transport, thereby conferring protection against drought in the acceptor-side of PSII. Moreover, this effect became more pronounced with increasing GR24 concentrations. The L-band analysis revealed a positive L-band under drought stress, indicating reduced energetic connectivity and system stability (Antunović Dunić et al., 2023). Conversely, GR24 treatment enhanced PSII unit connectivity ( Figures 1G, H ), ​implying ​optimized excitation energy distribution (Yusuf et al., 2010).

Significantly, ​we identified ​a drought-induced K-band ​appearance at 150 μs ( Figures 1I, J ), ​indicative of OEC ​dysfunction (Strasser et al., 2004). Tea plants with 10 μM and 30 μM GR24 had negative K-band, indeed improving the fraction of the active OEC centers. However, the K-band of 5 μM treatment exceeded that of DS treatment. This phenomenon may be caused by SL-induced stomatal closure via ABA-dependent/independent ways, which exacerbated CO₂ limitation and thereby amplified oxidative stress on the PSII donor side (Cui, 2017; Korek and Marzec, 2023; Zhang et al., 2024). In addition, the results demonstrated that drought stress caused a decline in F_v/F_m, while exogenous application of SL mitigated this reduction, indicating enhanced PSII functionality and alleviation of photoinhibition. This findings align with previous reports in wheat (Wei, 2021), grapevine (Min et al., 2019) and purpureum (Li et al., 2022), ​suggesting ​conserved SL functions ​across plant taxa.

The plant cell membrane plays a vital role in maintaining normal metabolic activities. Our results demonstrated that drought-induced redox imbalance triggered substantial membrane damage in both hydroponic and pot experiments ( Figures 2 , 3 , Supplementary Figure 1 ), consistent with the findings in grapevine (Wang et al., 2021) and soybean (Xie et al., 2024). GR24 treatment ​attenuated damage to cell integrity and oxidative stress with significant reduction in REC and MDA content, which might be attributed to ROS-scavenging capacity ​through ​both expression and activities of antioxidant enzymes (Wang et al., 2021; Xie et al., 2024). Intriguingly, we observed a concentration-dependent dichotomy in H₂O₂ regulation ( Figure 2C ), whereby low-dose GR24 suppressed H₂O₂ accumulation, while high-dose treatment potentiated ROS signaling. This dual functionality mirrors the hormetic effect of phytohormone signaling, where low-dose GR24 attenuates acute oxidative damage by limiting ROS accumulation, thereby maintaining membrane integrity. In contrast, elevated H₂O₂ levels under high-dose GR24 may reflect an adaptive signaling phase, potentially involving ABA-mediated activation of respiratory burst oxidase homologs (RBOHs) (Rodrigues and Shan, 2022). Moreover, the possibility that SLs directly promote ROS production through ABA-independent pathways cannot be excluded (Lv et al., 2018; Wei, 2021). These dual roles of GR24 highlight its complex involvement in redox regulation and drought adaptation, necessitating complementary analyses of antioxidant system activity and ABA biosynthesis profiles.

D27 is involved in abiotic stress resistance. For example, overexpression of D27 could decrease drought tolerance in transgenic rice (Haider et al., 2018). A correlation analysis was made to investigate the relationship between CsD27 and drought tolerance. We found that the cultivars with higher expression levels exhibited better drought tolerance compared to those with lower expression. In Arabidopsis, SL-deficient mutants which have a 70–75% reduction in SL content exhibited a hypersensitivity to drought, and they could be rescued when sprayed with SL to almost the same level of WT plants (Ha et al., 2014). Both of these results implicated the involvement of SL as a positive regulator in abiotic stress responses.

Phylogenetic analysis showed that CsD27 had the closest relationship with SlD27 and GmD27 ( Figure 5A ). It was reported that GmD27 could reduce the accumulation of ROS in an ABA–dependent manner (Han et al., 2025), providing phylogenetic evidence for CsD27’s putative role in drought adaptation. To verify this hypothesis, we generated CsD27-overexpressing Arabidopsis lines (OE5-9, OE9–2 and OE12-1). It was found that the transgenic plants with overexpression of CsD27 had enhanced osmotic resistance for the better growth, lower MDA content and higher survival rate under stress ( Figure 6 ). This might contribute to the crucial role for D27 in determining ABA and SL content (Haider et al., 2018). Future investigations should prioritize expanding genetic diversity across tea cultivars to systematically characterize the functional association between CsD27 allelic variation and drought resistance phenotypes.

Catechins are the main flavonoid antioxidants in tea plants, functioning primarily by donating hydrogen atoms and electrons to scavenge ROS (Lv et al., 2021). In the drought-sensitive cultivar HJY, both the total catechins and individual catechin levels declined under drought stress, whereas ​the drought-insensitive cultivar LJ43 ​maintained catechin content, particularly the ​key antioxidants epigallocatechin gallate (EGCG) and epigallocatechin (EGC). Such cultivar-specific response might ​align with ​PAL-mediated ​phenylpropanoid flux redirection under mild drought stress (Chakraborty et al., 2002). Additionally, GR24 application enhanced both total and individual catechin accumulation under drought conditions, suggesting that SL contribute to improved non-enzymatic antioxidant capacity, consistent with the results in crab apple (Xu et al., 2023). Notably, we observed that the expression of CsD27 was significantly negatively correlated with epicatechin (EC). This correlation may reflect complex SL-mediated effects on isomerization processes and branch-point regulation within the catechin biosynthesis pathway. Future studies are warranted to elucidate the precise role of SL in catechin metabolism and to explore strategies for optimizing catechin synthesis and accumulation through modulation of SL signaling. Moreover, the results also indicated that SLs application under drought stress contributed to tea quality. By preserving catechin levels, SLs not only directly enhanced the catechin accumulation, but also facilitated the production of downstream quality-related metabolites. Since oxidative catechin are inherently unstable, they might be further oxidized to facilitate ​theaflavins (TFs) and thearubigins (TRs), which are critical for taste, liquor color, and aroma quality of black tea (Lv et al., 2021). However, more evidence and experiments are needed to delineate SL-regulated ​biosynthetic nodes ​in ​catechin metabolism, ​potentially ​enabling ​targeted metabolic engineering.