Fresh leaves of Boju (0.5 g) were homogenized in 2 mL of 0.05 M phosphate buffer (pH 7.8). The resulting mixture was transferred to a centrifuge tube, where an additional 6 mL of buffer was added. The mixture was then centrifuged at 10,000 g for 20 min at 4°C. The supernatant was collected and used as the enzyme solution for subsequent assays.

To measure the malondialdehyde (MDA) content, 2 mL of the enzyme solution was combined with 2 mL of a 0.6% thiobarbituric acid solution, along with 4 mL of 5% trichloroacetic acid (TCA). This mixture was incubated in a boiling water bath for 10 min, subsequently removed, and left to cool. Following the cooling process, the solution underwent centrifugation at a speed of 3,000 g for a 15-min duration. The optical density (OD) of the supernatant was then assessed at wavelengths of 532 nm, 600 nm, and 450 nm. Subsequently, the MDA content was calculated utilizing the formula:

Each prepared solution was aliquoted into four clear EP(Eppendorf) tubes ( Supplementary Table S1 ). Two tubes were designated for the experimental group, while the remaining two acted as controls, with the enzyme solution in the control tubes replaced with distilled water. After thorough mixing, one control tube was shielded to prevent light exposure, while the other tubes, including those in the experimental group, were exposed to 4,000× fluorescent lamps for 10 min, maintaining a reaction temperature between 25–35°C. At the end of the incubation period, the reaction was halted by covering all tubes with a black cloth to obstruct further light exposure. The control tube exposed to light served as the blank, and the OD at 560 nm was measured for each tube. Superoxide dismutase (SOD) activity was then calculated using the subsequent equation:

In the formula, SOD₁ represents total enzyme activity, while SOD₂ denotes specific enzyme activity. A_cx stands for the absorbance of the control tube, and A_g signifies the change in absorbance of the sample tube. V_T refers to the total volume of the sample solution, V₁ indicates the volume of the sample used in the assay, and W represents the fresh weight of the sample.

The reaction mixture was prepared by combining phosphate buffer, guaiacol, and hydrogen peroxide in appropriate proportions. For the control tube, 3 mL of this reaction mixture was mixed with 1 mL of phosphate buffer, while the experimental tube contained 3 mL of the reaction mixture and 1 mL of an enzyme solution. The resulting reaction mixture was promptly moved to a cuvette, kick-starting the stopwatch for time tracking. The absorbance at the 470 nm wavelength was gauged at 1-min intervals using a spectrophotometer. Finally, the peroxidase activity was computed using the given formula:

The above equation represents the change in absorbance throughout the reaction. Here, ΔA470 indicates the change in absorbance during the reaction time, W denotes the fresh weight of the sample, V_T refers to the total volume of the sample solution, V_r indicates the volume of the enzyme solution used in the assay, and “t” represents the reaction time.

Freshly harvested leaves were placed in a mortar. Then, a small quantity of quartz sand and calcium carbonate powder, accompanied by 2–3 mL of 95% ethanol, were added to the leaves. This concoction was ground into a pulp. Subsequently, an extra 10 mL of ethanol was added, and grinding was continued until the tissue turned white. The mixture was left to stand for 3–5 min. Next, the solution was filtered through a funnel into a 25 mL brown volumetric flask. The mortar, pestle, residue, and filter paper were rinsed with a minimal amount of ethanol, and the rinse solution was filtered into the volumetric flask. Finally, the solution was diluted to 25 mL using ethanol. The resulting solution, transferred from the volumetric flask to a cuvette, had its absorbance measured at wavelengths of 665 nm, 649 nm, and 470 nm, using 95% ethanol as the blank. The computed concentrations of chlorophyll a, chlorophyll b, and carotenoids were calculated using the given equations.

Grind 0.5 g of fresh Boju leaves into a homogenate with 5 mL of distilled water, then centrifuge at 10,000 g for 10 min. Collect 1.0 mL of the supernatant and add 5 mL of Caulmers Brilliant Blue G-250 solution into a test tube; mix thoroughly. After 2 min, measure the absorbance at 595 nm and determine the protein content using the standard curve. The soluble protein content can be calculated using the given equation. In this formula, C represents the calculated protein content value obtained from the standard curve, VT is the total volume of the extracted solution, WF is the fresh weight of the sample, and VS is the volume of the sample spiked during the determination.

Fresh Boju samples were thoroughly ground in liquid nitrogen. These ground samples were accurately measured and placed into a test tube, followed by the addition of 10 mL of acetonitrile and 8 μL of the internal standard solution. The mixture was then extracted overnight at 4°C, before being centrifuged at 12,000 g for 5 min at a similar temperature. The supernatant was gathered, and 5 mL of acetonitrile was added to the remaining residue. This extraction process was performed twice more. Supplied supernatants were combined, and a suitable quantity of C18 and GCB(graphitized carbon black) was added to purify the extract. The mixture was subsequently centrifuged at 12,000 g for 5 min at 4°C, and then the supernatant was collected. The solvent underwent evaporation under nitrogen, and the leftover residue was re-dissolved in 400 μL of methanol. Before the hormone content analysis using HPLC-MS/MS(High-performance liquid chromatography-tandem mass spectrometry), the solution was filtered through a 0.22 μm organic-phase filter membrane.

The chromatographic conditions were as follows: A Poroshell 120 SB-C18 reversed-phase column was used (2.1 × 150 mm, 2.7 μm) with a column temperature of 30°C; the mobile phase was a ratio of A to B = (methanol/0.1% formic acid): (water/0.1% formic acid). The elution gradient was set as: 0–1 min, 20% A; 1–3 min, 20–50% A; 3–9 min, 50–80% A; 9–10.5 min, 80% A; 10.5–10.6 min, 80–20% A; and 10.6–13.5 min, 20% A. The changes in the endogenous levels of GA₃, GA₄, IAA, and ABA in Boju (Boju) after exogenous IAA treatment were measured.

The Boju samples, given similar growth conditions, were separated into two groups: control and treatment, each containing six biological replicates. The chrysanthemums in the treatment group were sprayed twice daily with a 10 mg/L IAA solution, while the control group was left untreated. Nine days post-treatment, samples were taken from both groups for metabolomics analysis following the standards stipulated by BGI Genomics Co., Ltd, Shenzhen. Each sample, weighing 50 μg, was soaked in 1.5 mL test tubes containing 800 μL of a pre-cooled extraction solution (methanol: H₂O ratio of 7: 3, v/v) and 20 μL of internal standard 1 (IS1). Homogenization utilized a braided grinder at 50 Hz for 10 min, which was then followed by sonication in a 4°C water bath for 30 min. After resting for 1 h at −20°C, the extracts underwent centrifugation at 14,000 g for 15 min at 4°C. The supernatant (600 μL) was then filtered through a 0.22 μm membrane, and 20 μL of this filtered solution from each sample was combined with mixed QC(quality control) samples to evaluate the reproducibility and stability of the LC/MS (Liquid chromatography-mass spectrometry)analysis. The filtered solutions and mixed QC samples were transferred to 1.5 mL sample vials for instrumental runs. Metabolites were subsequently separated and detected using the UPLC-MS(Ultra-performance liquid chromatography-mass spectrometry) method.

Frozen samples were sent to BGI Genomics Co., Ltd, Shenzhen for transcriptome sequencing. The total RNA(ribonucleic acid) enrichment method was used for treatment, and the desired RNA was obtained after purification. The purified RNA was then fragmented. Random N6 primers were used for reverse transcription, followed by cDNA(complementary DNA) duplex synthesis to form double-stranded DNA. The synthesized double-stranded DNA was end-repaired, phosphorylated at the 5′ end, and a sticky ‘A’ overhang was added at the 3′ end. A bulged adapter with a protruding ‘T’ at the 3′ end was then ligated. The PCR(polymerase chain reaction) product was heat-denatured to obtain single-stranded DNA, which was then circularized using primers to generate a single-stranded circular DNA library for sequencing.

The experimental data was analyzed using a one-way analysis of variance and independent samples t-test via SPSS Statistics 26.0 software and illustrated using GraphPad Prism 9. A p-value less than 0.05 indicates a significant difference whereas p greater than 0.05 signifies a non-significant discrepancy. Each treatment included three replicates.

It has been demonstrated that stress treatment triggers the accumulation of reactive oxygen species (ROS), accelerates membrane lipid peroxidation, and increases malondialdehyde (MDA) content. This study revealed that treatment with IAA significantly raised the MDA content in Boju, which was 4.9, 1.87, and 1.79 times higher after 3, 6, and 9 days of treatment, respectively, compared to day 0. The MDA content stabilized by the ninth day, showing a notable increase on day 3 compared to day 0 ( Figure 1A ). These results suggest that IAA treatment at this concentration leads to the accumulation of ROS and an increase in MDA content in Boju, resulting in cellular damage similar to that caused by abiotic stress.

ROS are toxic compounds that, if not promptly removed, can considerably affect plant growth and development. One key way that plants mitigate ROS effects is by boosting the activity of antioxidant enzymes. The most common ROS scavengers in plants include peroxidase (POD) and superoxide dismutase (SOD). This study revealed that SOD activity in chrysanthemums gradually increased with the application of the hormone IAA, reaching its peak on the 9th day – a 45% increase compared to day 0 ( Figure 1B ). Similarly, POD activities also significantly increased after IAA treatment compared to day 0, showing considerable differences compared to day 0. After 3, 6, and 9 days of IAA treatment, POD activity saw increases of 92.9%, 91.3%, and 93.8%, respectively ( Figure 1C ). These findings suggest that IAA treatment, at this concentration, amplified the activity of antioxidant enzymes and the resilience of Boju, hence preventing lipid membrane peroxidation damage happened due to IAA over-treatment. It ensured a response similar to adversity stress at this time due to IAA over-treatment.

Chlorophyll plays an integral role in plant photosynthesis; higher chlorophyll content indicates stronger photosynthetic activity. However, excessive chlorophyll can produce ROS, leading to cell death. Following IAA treatments, we noted an increase in chlorophyll a, chlorophyll b, and carotenoids in Boju over time. Specifically, the content of chlorophyll a rose by 34%, 32%, and 37% at days 3, 6, and 9, respectively, compared to day 0. Similarly, chlorophyll b content increased by 37%, 40%, and 46% on days 3, 6, and 9, respectively, compared to day 0; carotenoid content also augmented by 18%, 20%, and 25% at days 3, 6, and 9, respectively, compared to day 0 ( Figures 1D-F ). These findings suggest that IAA treatment boosts chlorophyll and carotenoid accumulation in Boju, promoting photosynthesis to help the plant manage the stress induced by excessive IAA. However, the rise in chlorophyll content also resulted in increased ROS production, signaling that Boju was under stress and sustained some damage at this time.

Soluble protein is a crucial substance in osmoregulation. Studies have indicated that soluble protein content typically increases dramatically under stress in plants during normal conditions, yet gradually decreases as stress intensifies. This is because most protein synthesis is inhibited during heightened stress periods, while specific proteins such as aquaporins and late embryogenesis abundant proteins are produced to augment the plant’s resistance to stress. In our research, we observed a decrease in the protein content of Boju leaves following IAA hormone treatment compared to day 0, with all observed changes being significantly different from day 0; decreasing by 15%, 16%, and 13% respectively ( Figure 1G ). This suggests that as the IAA hormone treatment duration extends and stress increases, the synthesis of specific proteins is triggered thereby enhancing the plant’s resilience to the damage induced by excessive IAA treatment at that point in time.

The current study is designed to investigate the impact of the over-application of exogenous IAA on the growth of Boju plants. Hormones serve as crucial signaling factors enabling plants to respond to stress from adverse conditions. Some research has indicated that the levels of growth hormone (IAA) and gibberellin (GA₃) in tea tree leaves decrease significantly under drought and high-temperature stress. This suggests that the tea tree slows its growth rate to adapt to these stress conditions by limiting the production of its growth-promoting hormones. Hence, we aim to explore whether the application of exogenous IAA impacts the fluctuation of endogenous hormones in the plant. This could offer a means to gauge the impact of IAA on the plant during such times.

In this study, we predominantly examined the fluctuations in four hormones – endogenous IAA, GA₄, GA₃, and ABA – after administering an exogenous IAA hormone. IAA plays a fundamental role as a growth hormone within the plant body, performing a range of functions such as maintaining apical dominance, enabling transport, and facilitating tissue and organ differentiation. It is particularly worth noting that IAA demonstrates dual effects: it promotes growth at low concentrations while inhibiting it at high concentrations Analysis of endogenous IAA in Boju plants before and after the treatment showed a significant 99% decrease in IAA levels by day 9, relative to day 0 ( Figure 2A ).

Gibberellins (GAs) are diterpenoid hormones that regulate various physiological processes, including plant growth, cell elongation, seed germination, and development. GAs also play a role in regulating plant responses to environmental stress. Although 129 gibberellin molecules have been isolated, only a few – including GA₃ and GA₄ – are biologically active. Previous studies have demonstrated that IAA positively regulates the expression of GA biosynthesis genes, thereby influencing GA formation. In this study, we observed significant decreases in GA₃ and GA₄ levels following IAA treatment. Specifically, GA₃ decreased by 93% compared to day 0, while GA₄ decreased by 45% ( Figures 2B, C ).

Thimann et al. proposed that increased levels of abscisic acid (ABA) in plants are a significant cause of leaf senescence, as ABA inhibits protein synthesis and accelerates the degradation of proteins and RNAs in the leaves. Concurrently, ABA enhances plant resistance against abiotic stresses by promoting processes including protein transport, carbon metabolism, and resistance gene expression patterns. In our study, after a 9-day IAA treatment, endogenous ABA levels in Boju decreased by 99% compared to day 0, demonstrating a significant difference from the control ( Figure 2D ). This indicates that while IAA treatment stresses the plants, it does not expedite senescence.

In conclusion, our study determined that all four endogenous hormones – IAA, GA₃, GA₄, and ABA – experienced significant reductions in Boju plants following exogenous IAA treatment. This finding suggests that overextended exposure to exogenous IAA does not foster plant growth as expected, but instead incites stress. This stress, in turn, provokes the plant to slow its growth as a response to the hormonal imbalance. Such an adaptive response might encompass a reduction in hormones to alleviate the stress effects induced by the excessive IAA treatment.

The reproducibility of the QC sample assay was used to assess data quality. The primary components included the intensity of the highest ion acquired at each time point, employed to generate the base-peak chromatogram ( Supplementary Figures S1A, B ). Furthermore, the raw mass spectrometry data were processed using MetaX to produce a summary data table. From this table, a detection of 9121 metabolites was confirmed. QC samples were subsequently analyzed, with reproducibility evaluated by the correlation coefficient (R). An R-value closer to 1 indicates fewer systematic errors and superior experimental reproducibility, thus reflecting a higher data quality ( Supplementary Figure S2 ). The coefficient of variation (CV) was calculated based on the quantification values of each metabolite in the QC samples, where a lower CV value indicates improved reproducibility of the quantification results ( Supplementary Table S2 ).

The analysis of overall metabolites primarily involved qualitative methods, which included demonstrating the detection effects of each sample group through spectrograms and classifying all identified metabolites, followed by counting the number of metabolites in each category. For quantitative analysis, principal component analysis (PCA) plots are utilized to observe contrasting group similarities as a whole and to cluster expression patterns among groups. Furthermore, the number of metabolites detected corresponded with the number of isolated peaks in the base-peak chromatogram (BPC) ( Supplementary Figure S3 ). From the data summary table, a total of 9,121 metabolites were counted, out of which 2,308 were identified. Metabolites with classification information in the Kyoto Encyclopedia of Genes and Genomes (KEGG) and HMDB(Human Metabolome Database) databases were categorized into 33 groups ( Supplementary Figure S4 ), and the identified metabolites were counted based on the broad categories of the 22 KEGG metabolic pathways they were involved in. Metabolites were predominantly found in lipids, terpenoids, alkaloids, and flavonoids, as well as in the amino acid and carbohydrate metabolic pathways ( Supplementary Figure S5 ).

Univariate and multivariate analyses were conducted to detect differential metabolites between the two groups. Initially, we investigated the overall differences using PCA ( Supplementary Figure S6 ) and Partial Least Squares Discriminant Analysis (PLS-DA) ( Supplementary Figure S7 ). We then validated the OPLS-DA(Orthogonal Partial Least Squares Discriminant Analysis) model ( Supplementary Figure S8 ) using a 200-response permutation test. The results suggested a well-fitted model when the red dashed line tilted upwards and the intercept between Q2 and the vertical axis was less than 0 ( Supplementary Figure S9 ). Subsequently, we screened differential metabolites based on the following criteria: 1) VIP(variable importance in projection) ≥ 1 in the OPLS-DA model; 2) Fold change ≥ 1.2 or ≤ 0.83; 3) p-value< 0.05 in univariate analysis. We identified a total of 263 differential metabolites, with 69 being up-regulated and 194 down-regulated ( Figure 3A ). A volcano plot of these differential metabolites is illustrated in Figure 3B , wherein down-regulated significant metabolites appear in green, up-regulated significant ones in red, metabolites with VIP ≥ 1 as circles, those with VIP< 1 as triangles, and non-significant metabolites in gray.

After screening for differential metabolites, examining the expression patterns and biological functions of these metabolites is crucial. In this study, we analyzed expression patterns through clustering of differential metabolite expression, correlation clustering analysis, and network analysis. The biological functions of these metabolites were scrutinized using pathway annotation and enrichment analysis. We also conducted a comparison analysis of the differential metabolites in individual groups. In addition, we assessed the potential of differential metabolites as markers for growth.

Due to marked differences in physiological indicators between the CK and IAA treatments, a differential transcriptome analysis was conducted on these two groups, producing a total of 38.11 Gb of data. Following de novo assembly and redundancy removal of clean reads with Trinity, we obtained a total of 334,744 unigenes with a cumulative length of 231,850,720 bp, an average length of 692 bp, an N50 of 1,109 bp, and a GC content of 39.55%. These unigenes were then annotated by comparison with seven functional databases, yielding the following results: 188,725 unigenes in the Non-Redundant Protein Sequence Database (NR, 56.38%), 115,568 in the Nucleotide Sequence Database (NT, 34.52%), 112,100 in Swiss-Prot (33.49%), 94,686 in the Eukaryotic Clusters of Orthologous Groups (KOG, 28.29%), 116,277 in KEGG (34.74%), 149,638 in GO (44.70%), and 86,485 in the Protein Families Database (Pfam, 25.84%) were functionally annotated.Using TransDecoder, 103,247 coding sequences (CDSs) were detected. Additionally, 22,799 simple sequence repeats (SSRs) were identified in 20,082 unigenes, and 3,601 unigenes encoding transcription factors were predicted. PCA was executed on control and treated samples ( Supplementary Figure S12 ) to pinpoint differentially expressed genes. The dispersion of gene expression levels across samples was evaluated using box plots ( Supplementary Figure S13 ), and Venn diagrams were created for genes in treatment and control groups ( Supplementary Figure S14 ). Interestingly, a total of 206,098 genes were expressed in the treatment and control groups.

Intergroup difference analysis was conducted using Poisson Distribution, with the thresholds being a Fold Change ≥ 2 and an FDR(False Discovery Rate) ≤ 0.001. A total of 144 differential genes were identified, with 49 up-regulated and 95 down-regulated ( Figure 6A ). Subsequently, based on GO and KEGG annotation results, these differential genes underwent further functional classification and enrichment analysis.

In the GO analysis, results were categorized into three functional groups: molecular function, cellular component, and biological process ( Figure 6B ). Functional categories that displayed a Q value ≤ 0.05 were deemed significantly enriched. This analysis exposed that the differential genes were significantly enriched in 183 GO terms. The top 20 GO terms with the smallest Q values were chosen and visualized as an enriched GO bubble plot. The plot reveals significant enrichment in glutamate decarboxylase activity, glutamate metabolic process, protein O-linked glycosylation, gibberellic acid-mediated signaling pathway, and protein O-GlcNAc transferase activity ( Figure 6C ).

Moreover, KEGG pathway analysis identified that differential genes partook in 51 metabolic pathways. These pathways were sorted into seven branches, with the differential genes in this study mostly involved in five major branches: Cellular Processes, Environmental Information Processing, Genetic Information Processing, Metabolism, and Organismal Systems ( Figure 6D ).

Based on the KEGG pathway annotation classification, p-values were calculated, and pathways with a Q value ≤ 0.05 were deemed significantly enriched. We identified 43 KEGG pathways significantly enriched in the differentially expressed genes found in this study. The top 20 pathways with the smallest Q values were selected for the KEGG enrichment bubble plot. The analysis indicated that the differentially expressed genes were mainly enriched in pathways associated with photosynthesis, taurine, and hypotaurine metabolism, other types of O-glycan biosynthesis, and terpenoid backbone biosynthesis. Other enriched pathways included butanoate metabolism, beta-alanine metabolism, N-glycan biosynthesis, alanine, aspartate, and glutamate metabolism, among others ( Figure 6E ).