Uniform, healthy, three-year-old potted plants of C. oleifera ‘Xianglin 210’ were supplied by the National Engineering Research Center for Oil-tea Camellia (Changsha, China). The seedlings were cultivated in plastic pots (top diameter: 24 cm; height: 20 cm) filled with a homogenized substrate of lateritic red soil (collected from Changsha, Hunan; 112°58′ E, 28°12′ N) and peat soil (Pindstrup Mosebrug A/S, Ryomgaard, Denmark) at a 3:1 (v/v) ratio. At the beginning of the experiment, seedlings had an average height of 75.9 cm and a basal stem diameter of 7.9 mm. The experiment was conducted in the Center’s greenhouse (113°01′ E, 28°06′ N), where daytime and nighttime temperatures were maintained at 30°C and 25°C, respectively, with a relative humidity of 80–85%. Water-soluble iron dihydroporphyrin powder was purchased from Anqing baite Biology Engineering Co., Ltd. (Anqing, China), dissolved to prepare solutions of 20, 40, and 80 µg/L, and applied as foliar sprays. Plants were sprayed with a handheld atomizer until leaf surfaces were uniformly wetted to the point of slight runoff, ensuring consistent coverage across treatments. Treatments commenced on day 1 of the experiment and were repeated at 15-day intervals for a total of three applications. SPAD values were measured on intact leaves using a SPAD-502 Chlorophyll Meter (Konica Minolta, Osaka, Japan), which estimates relative chlorophyll content based on leaf transmittance at 650 nm and 940 nm.

To investigate the physiological and biochemical responses of C. oleifera saplings under normal conditions following ICE6 treatment, four treatment groups with different ICE6 concentrations were established: 0 µg/L (T0, control), 20 µg/L (T1), 40 µg/L (T2), and 80 µg/L (T3). Throughout the normal condition experiment, all plants were maintained under normal growth conditions with adequate watering to avoid any water stress. ICE6 treatments were applied on days 1, 15, and 30 ( Figure 1 ). Each treatment was performed with three biological replicates, and each replicate consisted of five uniformly sized plants. On day 31, photosynthetic parameters (n = 15) and SPAD values (n = 15) were measured on the fully expanded middle leaves of each plant between 9:00 and 11:00 in the morning. Subsequently, the same leaves were randomly collected, immediately frozen in liquid nitrogen, and transported back to the laboratory for further analysis. The indices measured included soluble protein (SP) content, soluble sugar (SS) content, indole-3-acetic acid (IAA) content, malondialdehyde (MDA) content, and proline (Pro) content (n = 3).

Building upon the experiment described in Section 2.2, a drought stress treatment was further applied to the potted plants ( Figure 1 ). The soil water content was measured using a ProCheck handheld multifunction reader/data logger. Drought stress was induced by withholding water on day 31, allowing the soil to dry naturally. Sampling was performed on the seventh day (On day 38), when the relative soil water content had decreased to approximately 30%. Once the target soil moisture level was reached, photosynthetic parameters (n = 15) and SPAD values (n = 15) were measured on the fully expanded leaves in the middle each plant between 9:00 and 11:00 in the morning. Same leaves were then collected, immediately frozen in liquid nitrogen, and transported back to the laboratory for subsequent physiological and biochemical analyses. The indices measured included superoxide dismutase (SOD) activity, peroxidase (POD) activity, catalase (CAT) activity, malondialdehyde (MDA) content, proline (Pro) content, soluble sugar (SS) content, abscisic acid (ABA) content, and soluble protein (SP) content (n = 3).

Photosynthetic parameters, including net photosynthetic rate (Pn), intercellular CO₂ concentration (Ci), stomatal conductance (Gs), and transpiration rate (Tr), were measured using a 6400XT Portable Photosynthesis System (LI-COR, Lincoln, NE, USA) on functional leaves of randomly selected C. oleifera saplings. The system was configured with an airflow rate of 500µmol/s, a photosynthetically active radiation (PAR) intensity of 1000µmol/m²/s, a CO₂ concentration of 400 µmol/mol. Water use efficiency (WUE) was calculated as the ratio of Pn to Tr, to assess the efficiency of water utilization under specific conditions.

The activities of peroxidase (POD), superoxide dismutase (SOD) and catalase (CAT), together with the contents of malondialdehyde (MDA), proline (Pro), soluble sugars(SS) and soluble proteins (SP), were determined using commercial 96-well microplate assay kits (Quanzhou Ruixin Biotechnology Co., Ltd., Quanzhou, China). The following kits were used: Superoxide Dismutase (SOD) Activity Assay Kit (WST-8 method; Cat. No. G0101W), Peroxidase (POD) Activity Assay Kit (Cat. No. G0107W), Catalase (CAT) Activity Assay Kit (Cat. No. G0105W), Malondialdehyde (MDA) Content Assay Kit (Cat. No. G0109W), Proline (Pro) Content Assay Kit (Cat. No. G0111W), Soluble Sugar Content Assay Kit (Cat. No. G0501W) and Soluble Protein Content Assay Kit (Coomassie Brilliant Blue method; Cat. No. G0417W).

Endogenous indole-3-acetic acid (IAA) and abscisic acid (ABA) were quantified by ultra-performance liquid chromatography–tandem mass spectrometry (UPLC–MS/MS). For each treatment, three biological replicates were prepared. Leaves from five uniformly sized plants were pooled per replicate. Each pooled sample was ground in liquid nitrogen, thoroughly mix, and accurately weigh approximately 0.3 g (to 0.0001 g) into a 15 mL centrifuge tube. Add 5 mL of extraction solution (isopropanol:water:formic acid, 80:19:1, v/v/v), homogenize for 2 min, and perform ultrasonic extraction at 4°C for 1 h. Centrifuge the extract at 10,000 rpm for 10 min to obtain the supernatant. Subsequently, add 1 mL of dichloromethane, perform low-temperature ultrasonic extraction for 30 min, and centrifuge at 10,000 rpm for 10 min to collect the supernatant. Repeat the ultrasonic extraction once more, and combine the three supernatants. Evaporate the combined supernatant under reduced pressure with nitrogen at room temperature to the aqueous phase, adjust the volume to 2 mL, vortex to mix, dilute 2-fold with methanol, and filter through a 0.22 µm membrane for analysis. All operations are conducted on wet ice and protected from light.

Chromatographic separation was performed on a Waters FTN UPLC system equipped with a BEH C₁₈ column (2.1 × 50 mm, 1.7 µm). The column temperature was maintained at 40°C and the autosampler at 10°C. The mobile phase consisted of 0.1% formic acid in water (A) and acetonitrile (B), delivered at 0.3 mLmin⁻¹ using the following gradient: 0.0 min, 80/20 (A/B); 1.5 min, 65/35; 3.0 min, 10/90; 4.5 min, 10/90; 4.7 min, 65/35; and 5.0 min, 80/20. The total runtime was 5 min, followed by a 5 min post time at the initial conditions (80/20) for column re-equilibration. Injection volume was 10 µL.

Mass spectrometric analysis was performed on an AB SCIEX 4000 triple quadrupole mass spectrometer equipped with an electrospray ionization (ESI) source operated in both positive and negative modes (ESI±). The MS/MS parameters were set as follows: curtain gas (CUR) 35 psi; collision gas (CAD) 7 psi; IonSpray voltage +5500 V in positive mode and –4500 V in negative mode; source temperature 450°C; ion source gas 1 (GS1) 40 psi; and ion source gas 2 (GS2) 50 psi.

The original chromatograms, calibration curves, and quantification data of IAA and ABA from UPLC–MS/MS analyses are included in the supplementary file ( Supplementary Data S1 ).

Total RNA was extracted from C. oleifera leaves (n = 3) using the RNAprep Pure Plant Plus Kit (TIANGEN, Beijing, China). First-strand cDNA was synthesized using the HiScript III 1st Strand cDNA Synthesis Kit (+gDNA wiper) (Vazyme, Nanjing, China), followed by five-fold dilution of the cDNA for quantitative real-time PCR (RT-qPCR) analysis. The ChamQ Universal SYBR qPCR Master Mix (Vazyme, Nanjing, China) was used for the RT-qPCR assays. Primers for the RT-qPCR were designed using Primer Premier 5 software (Premier Biosoft, Palo Alto, CA, USA), and the sequences of the primers are shown in Table 1 . The 2{sp}−ΔΔ{it}Ct {/sp}{/it}method was used to calculate the relative expression levels of the genes. Tubulin was used as the reference gene, as it has been validated as a stable housekeeping gene in C. oleifera (Zhang et al., 2025). Descriptions of gene functions are provided in the supplementary file ( Supplementary Data S2 ).

To investigate the relationships among physiological and biochemical traits under drought stress, we performed correlation analysis, principal component analysis (PCA), hierarchical clustering, and heatmap visualization. Pearson correlation coefficients were calculated based on trait values, and the correlation matrix was visualized using the corrplot package in R. Heatmaps of treatment means (standardized as Z-scores) were generated with the pheatmap package to illustrate clustering patterns of treatments and traits. PCA was conducted to examine the contribution of traits to the principal components. In addition, traits were clustered using the unweighted pair-group method with arithmetic mean (UPGMA), based on a distance matrix derived from Pearson’s correlation (1 – r), to identify groups of traits with similar response patterns.

Statistical analyses were conducted using IBM SPSS Statistics 27 (IBM Corporation, Armonk, NY, USA). For physiological and biochemical indices, including SPAD value, photosynthetic rate, antioxidant enzyme activities (SOD, POD,CAT), MDA content, Proline, soluble protein, soluble sugars, and hormone levels (ABA, IAA), means and standard deviations were calculated for descriptive statistics. All data were first subjected to two-way analysis of variance (ANOVA), followed by Duncan’s new multiple range test at a significance level of P < 0.05. In each column, means sharing the same letter (a, b, c, d) are not significantly different (P ≥0.05), whereas means with different letters differ significantly (P < 0.05).

As illustrated in Figure 2A , the SPAD values of C. oleifera seedlings significantly decreased under drought conditions compared to normal conditions (T0). Furthermore, the SPAD values of ICE6-treated Camellia oleifera seedlings (T1, T2, T3) were higher than those of the control group (T0) under both normal and drought conditions. Under normal conditions, the SPAD values of ICE6-treated seedlings increased significantly compared to the control, with rises of 7.99% for T1 (20 µg/L), 11.36% for T2 (40 µg/L), and 11.57% for T3 (80 µg/L), corresponding to increasing ICE6 concentrations. Under drought stress, compared to T0, the SPAD values in the ICE6-treated groups remained relatively high, with T2 showing the most notable increase.

Under normal conditions, the Pn of the ICE-treated groups was significantly higher than that of the control group (T0), and Pn increased with higher ICE treatment concentrations. The highest rate was observed in the T3 group, showing a 113.73% increase compared to T0. The patterns of Gs ( Figure 2C ) and Tr ( Figure 2E ) under normal condition were consistent with the Pn. Moreover, the Ci in ICE-treated groups was higher than in T0, with similar Ci values across the different ICE concentrations. Interestingly, WUE was significantly lower in the ICE-treated groups compared to the control group T0 ( Figure 2F ). These observations suggest that treating C. oleifera saplings with ICE can significantly enhance their photosynthetic rates.

Under drought stress, the Pn in the T1 and T3 groups was significantly higher than that of the control group (T0), while T2 showed no significant difference compared to the control ( Figure 2B ). Gs and Tr in ICE-treated groups were also significantly higher than in T0, with the lowest Gs and Tr values observed in T2. ( Figures 2C , E ). The Ci among ICE-treated groups showed complex variations: Ci in T2 was significantly higher than in T0, while T1 had significantly lower values compared to T0, and T3 exhibited no significant difference from the control ( Figure 2D ). Furthermore, WUE in T1 was significantly higher than in T0, whereas T2 and T3 showed no significant differences from the control ( Figure 2F ). These results suggest that different ICE concentrations elicit varied photosynthetic responses under drought stress. Specifically, T1 was able to sustain higher photosynthetic activity during drought, whereas T2 and T3 appeared to actively downregulate its photosynthetic activity.

Under normal conditions, the IAA content in the ICE-treated groups was higher than that of T0. However, it gradually decreased as the ICE concentration increased ( Figure 3 ). Under drought stress, ABA content in ICE-treated groups was significantly lower than in T0, with T2 exhibiting the highest ABA levels among all treatments, followed by T1 and T3 ( Figure 3 ).

Antioxidant enzymes such as SOD, POD, and CAT play crucial roles in maintaining redox balance. They effectively eliminate excess reactive oxygen species (ROS), such as superoxide anions and hydrogen peroxide, which are frequently induced by drought stress (Laxa et al., 2019). These ROS attack unsaturated fatty acids in the cell membranes, triggering lipid peroxidation processes, with malondialdehyde (MDA) being a direct byproduct (Das and Roychoudhury, 2014; Huang et al., 2024). Consequently, an increase in MDA levels typically reflects the level of oxidative stress induced by ROS within the cells (Hasanuzzaman et al., 2020).

Under drought stress, SOD activity in ICE-treated groups was significantly higher than that in T0, with the highest activity observed in T2, followed by T1 and T3 ( Figure 4A ). Although the POD ( Figure 4B ) and CAT ( Figure 4C ) activities in ICE-treated C. oleifera seedlings were lower than those in the control group (T0) under drought stress, these activities were overall elevated compared to their respective levels under normal conditions. Furthermore, MDA levels in all ICE-treated groups were significantly lower than in T0, with reductions of 18.32% for T1 (20 µg/L), 14.14% for T2 (40 µg/L), and 30.84% for T3 (80 µg/L), indicating enhanced drought tolerance in C. oleifera saplings treated with ICE ( Figure 4D ). These results suggest that ICE treatment can effectively activate the antioxidant system in C. oleifera seedlings under drought stress. In particular, the T3 treatment showed the lowest MDA level under drought conditions, with no significant difference compared to its respective level under normal conditions.

Under normal conditions, both proline ( Figure 5A ) and soluble sugar levels ( Figure 5B ) in ICE-treated seedlings were not significantly different from those in the control group (T0). However, under drought stress, the proline and soluble sugar levels in C. oleifera seedlings were higher than their respective levels under normal conditions.

Regarding soluble protein (SP) content, under normal conditions, ICE-treated groups exhibited higher SP levels than T0, with the highest content observed in the T2 group, followed by T3 and T1 ( Figure 5C ). Under drought stress, SP content in ICE-treated groups was significantly higher than in T0, indicating that ICE promotes protein accumulation.

To investigate how ICE6 influences gene expression in C. oleifera saplings, we focused on eight key genes previously identified in our transcriptomic analyses (He et al., 2022). These genes, which show significant differential expression under drought stress, are crucial for understanding hormonal signaling pathways, particularly those involving auxin and ABA.

Under normal conditions, the expression levels of RbcL in T2 and T3 were significantly higher than the control, while T1 showed slightly lower levels. Meanwhile, RbcS expression in T1 and T2 were significantly higher than in the control, whereas T3 was significantly lower ( Figure 6 ). Under drought stress, RbcL expression in T1 and T2 did not significantly differ from the control, while T3 was significantly higher. RbcS expression in T2 and T3 were significantly higher than in the control, with no significant difference observed in T1.

Under normal conditions, IAA9, a negative regulator of auxin, exhibited significantly lower expression levels in T2 and T3 compared to the control, while T1 showed no significant difference. However, under drought stress, the highest expression levels of IAA9 were observed in T2, followed by T1, with T3 showing no significant difference from the control. These results suggest that, under normal conditions, ICE6 treatment in T2 and T3 may activate downstream genes in auxin pathways by inhibiting the negative regulator IAA9. Conversely, under drought stress, ICE6 treatment appears to upregulate IAA9 expression in T1 and T2, which could lead to the suppression of auxin pathway activities. Additionally, expression levels of TCP4 were significantly reduced across all ICE6-treated groups under normal conditions. During drought stress, TCP4 expression in T1 was significantly higher than in the control, while T2 showed no significant difference, and T3 was significantly lower. The varied expression of TCP4 suggests that drought tolerance mechanisms differ across the different ICE6 concentrations.

Under drought stress, the expression levels of PP2C16, PP2C24, PP2C51, and SnRK2.8 in T1 were higher than in the control, indicating enhanced ABA signaling. In T3, expression levels of PP2C24, PP2C51, and SnRK2.8 were also significantly higher than in the control. However, in T2, the expression of these genes showed no significant differences compared to the control.

Principal component analysis (PCA) revealed that the first two principal components (PC1 and PC2) explained 54.98% and 16.95% of the total variance, respectively, accounting for a cumulative variance of 71.93% ( Figure 7A ). The enzymatic antioxidant traits such as POD and CAT were strongly associated with the negative axis of PC1, whereas SOD was mainly associated with PC2. Lipid peroxidation marker MDA, osmolyte Pro and SS showed moderate contributions to both PC1 and PC2. Photosynthetic traits such as SPAD, Pn, Gs, and Tr were positioned opposite to POD and CAT vectors, indicating negative associations with these enzymatic antioxidant traits along PC1. This pattern was further supported by the Pearson correlation heatmap ( Figure 7B ), in which POD and CAT were positively correlated with each other but negatively correlated with SPAD, Pn, Gs, and Tr. Photosynthetic traits showed strong positive intercorrelations, while MDA and Pro exhibited moderate positive correlations with POD and CAT and negative correlations with most photosynthetic parameter.

The UPGMA dendrogram grouped the physiological and biochemical traits into two major clusters ( Figure 8A ). Photosynthetic-related parameters (SPAD, Pn, Gs, Tr) were closely associated, while oxidative stress and antioxidant indicators (MDA, SOD, POD, CAT, Pro, etc.) formed another cluster. This classification highlights the coordinated responses of photosynthetic traits and stress-related traits under drought conditions. Consistently, the heatmap of standardized traits revealed distinct patterns among treatments ( Figure 8B ). Under well-watered conditions, T3 exhibited the most favorable profile, with the highest SPAD, Pn, Gs, and Tr, indicating optimal photosynthetic performance under non-stress conditions. Under drought stress, T3 also showed the lowest MDA while largely maintaining photosynthetic traits, suggesting effective mitigation of oxidative damage. Overall, T3 demonstrated superior performance under both normal and drought stress conditions.