The water dropwort variety ‘Fuqin No.1’ was used as plant material in this study, which was stored in the aquatic vegetable experimental base of Yangzhou University (32°39′N, 119°42′E). The water dropwort was grown in the pots (38 cm × 28 cm) containing fertile soil under natural condition. The water dropwort was irrigated with water every 5 days and the 30-day-old plants were treated with different concentrations of GA₃. The treatments of T1, T2, T3, T4 represented 40 mg/L, 80 mg/L, 120 mg/L, 160 mg/L of GA₃, respectively. T5 indicated the treatment of uniconazole (a gibberellin inhibitor, 50 mg/L). The water dropwort treated with distilled water was set as control group (CK). Each treatment was performed for three biological replicates. The 65-day-old plants of water dropwort were measured and the samples were harvested for further analysis.

The petioles of water dropwort under different treatments were sampled for anatomical structure analysis. The middle part of petioles was cut into 3 mm and the tissue structure was immobilized in the phosphate buffer solution containing 2.5% of glutaraldehyde (pH = 7.2). The slices were dewaxed and dehydrated with xylene and ethanol, respectively. The samples were then stained by the safranin-O (1%) for 2 h and cleaned by ethanol (Que et al., 2018). The anatomical structure of water dropwort was observed by light microscope. The Image J software was used to detect the cell length of water dropwort under different treatments (Schindelin et al., 2015).

To further investigate the changes of GAs contents in water dropwort, the petioles of CK, T2, and T5 groups were collected for determination of GAs contents. The samples of water dropwort were frozen in liquid nitrogen and stored in -80 °C refrigerator. The samples were then sent to MetWare (http://www.metware.cn/) for the determination of GAs contents. The measurement was performed using AB Sciex QTRAP 6500 LC-MS/MS platform (He et al., 2020).

The broadly targeted metabolomics of water dropwort were conducted in MetWare (http://www.metware.cn/). Briefly, the plant samples were first freeze-dried and crushed. Fifty mg of lyophilized powder were dissolved in methanol solution (70%). The samples were then centrifugated and filtrated for further UPLC-MS/MS analysis. Four μL of the above extraction were injected into the UPLC-ESI-MS/MS system. The effluent was further analyzed by an ESI-triple quadrupole-linear ion trap (QTRAP)-MS (Li et al., 2021). The differentially expressed metabolites (DEMs) among different groups were identified based on the variable importance in projection (VIP) and fold change (FC) values. The VIP values were obtained from the OPLS-DA model. The metabolites with a VIP value ≥ 1 and FC value ≥2 or ≤0.5 were determined as DEMs (Cao et al., 2022). The identified metabolites were annotated based on the KEGG database (http://www.kegg.jp/kegg/compound/).

The petioles of water dropwort from three biological replicates under CK, T2, and T5 treatments were sampled and frozen in liquid nitrogen for transcriptome analysis (Zhang et al., 2017). The purity and integrity of extracted total RNA was detected by NanoDrop 2000 and Agient2100/LabChip GX, respectively. The RNA-seq libraries were constructed by Biomarker Technologies Co., Ltd. (Beijing, China). The constructed libraries were sequenced by the Illumina NovaSeq6000 platform. Based on the high-quality genome of water dropwort (Liu et al., 2023), the clean reads of transcriptome were compared to the reference genome by the HISAT2 software (Kim et al., 2015). The expression levels of the gene were calculated based on the read counts and normalized by the fragments per kilobase of transcript per million fragments mapped (FPKM) method (Trapnell et al., 2010). The transcriptome analysis of water dropwort was conducted on the BMKCloud (www.biocloud.net). The raw data has been deposited in SRA database with the ID of PRJNA977200.

In order to identify the members of the cellulose synthase (CES) family in water dropwort, the HMM model (PF03552) was used to search the CES family member from water dropwort by TBtools (Chen et al., 2020). The obtained candidate CES family sequences were further determined by analysis of conserved domains in Pfam databases (Finn et al., 2014). The sequences of Arabidopsis CES family were downloaded from The Arabidopsis Information Resource (TAIR) (Rhee et al., 2003). The protein sequences of CES family from water dropwort and Arabidopsis were used to construct the phylogenetic tree by MEGA 7.0 using neighbor-joining method (Kumar et al., 2016). The promoter sequence (upstream 2000 bp) of different CESA family genes were extracted from the genome of water dropwort (Liu et al., 2023). The cis-acting elements analysis of promoter was performed on the PlantCARE (Lescot et al., 2002) and the different cis-acting elements were visualized by TBtools (Chen et al., 2020).

To investigate the effects of GA treatment on the growth of water dropwort, the GA₃ with five concentrations (0, 40, 80, 120, 160 mg/L) and uniconazole (GA biosynthesis inhibitor) were used to treat water dropwort plants. The water dropwort was measured and sampled after 30 days of treatment. The growth of water dropwort was influenced by GA treatment ( Figure 1A ). Plant height of water dropwort was significantly increased after GA₃ treatment, and the GA₃ with concentration of 80 mg/L (T2) showed the most significant effect on the growth of water dropwort ( Figure 1B ). The effects of GA₃ and uniconazole on the petiole numbers were not obvious ( Figure 1C ).

The anatomical structure of petioles under GA and uniconazole treatments was examined. The results indicated that the cell length of petiole was significantly increased after GA treatment ( Figures 2A–F ). The cell length of water dropwort petiole without any treatment was approximately 253.6 μm. By contrast, the cell length of petiole with 80 mg/L of GA₃ treatment was 343.5 μm, which was significantly longer than CK and uniconazole treatments ( Figure 2G ). These results suggested that the GA-mediated increase of plant height was mainly due to cell elongation of water dropwort. Compared with other concentrations of GA treatments, the T2 (80 mg/L) treatment had the most significant promoting effect on the cell length of water dropwort.

The metabolomic analysis of water dropwort was performed to understand the metabolites changes under GA treatment. A total of 870 metabolites were identified from all samples of water dropwort ( Supplementary Table S1 ). The detected metabolites can be classified into different categories. The flavonoids accounted for the highest proportion of 17.59%, followed by phenolic acids, amino acids and derivatives, and organic acids, which accounted for 16.21%, 10% and 8.16% respectively ( Figure 3 ). Cluster analysis indicated that different metabolites accumulated in CK, T2, and T5 treatments.

The differentially expressed metabolites (DEMs) among CK, T2, and T5 groups of water dropwort were analyzed ( Supplementary Table S2 ). The VIP based on the OPLS-DA model and FC values were combined to identify the DEMs in different groups of water dropwort ( Figures 4A–C ). The results indicated that 59 DEMs were identified in CK and T2 group, including 37 down-regulated and 22 up-regulated DEMs. A total of 114 DEMs were identified in CK and T5 group, including 66 down-regulated and 48 up-regulated DEMs. A total of 142 DEMs were identified in CK and T5 group, including 84 down-regulated and 58 up-regulated DEMs. The KEGG enrichment analysis of DEMs in different groups of water dropwort was also conducted. The results indicated that the metabolic pathways and biosynthesis of secondary metabolites were the most abundant pathways ( Figures 4D–F ).

To identify the gene expression profiles under exogenous GA treatment, the transcriptome analysis of water dropwort was conducted. A total of 63.98 Gb clean data were obtained and the percentage of Q30 bases of each sample was no less than 92.52% ( Supplementary Table S3 ). The mapped efficiency of different samples with reference genome ranged from 91.20% to 94.46%. The differential expressed analysis of water dropwort under different treatments was performed ( Figure 5A ). The volcano plot indicated that a total of 5,367 DEGs were identified between CK and T2 treatments, including 2,694 up-regulated DEGs and 2,673 down-regulated DEGs ( Figure 5B ).

The annotation and enrichment analysis of DEGs was conducted according to the GO and KEGG databases. The GO enrichment analysis mainly includes the biological process, cellular component, and molecular function. As for biological process branch, the ‘carbohydrate metabolic process’ had the most abundant DEGs after exogenous GA treatment. The components involved in cell wall were also identified in the biological process branch, such as ‘cell wall organization’ and ‘cell wall biogenesis’ ( Supplementary Figure S1 ). In cellular branch, the plasma membrane had the largest number of DEGs in all components. In addition, the cell wall-related components were also detected from the branch. The exogenous GA₃ may promote plant elongation by regulating cell-related life processes in water dropwort. KEGG annotation and enrichment analysis were conducted to identify the metabolic pathways of water dropwort ( Figure 5C ). The DEGs were most enriched in the ‘plant hormone signal transduction’ pathway after exogenous GA treatment. This suggested that the exogenous GA treatment could regulate plant growth by affecting various hormone signaling pathways. In addition, the DEGs were also abundantly enriched in many metabolisms, including ‘starch and sucrose metabolism’, ‘phenylpropanoid biosynthesis’, and ‘carbon metabolism’.

To investigate the effect of exogenous GA₃ treatment on different kinds of GA, the GAs contents of CK, T2, and T5 were determined. Compared with CK and T5 treatments, the GA₃ content under T2 treatment was significantly increased after exogenous GA₃ treatment. In addition, the contents of GA₁, GA₄, and GA₈ were also increased after GA₃ treatment. However, some GAs were significantly decreased after GA₃ treatment, such as GA₁₅, GA₁₉, GA₂₄, GA₂₉, and GA₅₃ ( Figure 6A ). The transcriptomic analysis indicated that the GA biosynthesis and signaling transduction was significantly influenced by the exogenous GA treatment. GAs is synthesized from the geranylgeranyl-PP (GGPP). The expression trends of GA biosynthesis genes (KAO, GA20ox, GA3ox) in different groups varied. The transcripts of GA metabolism-related gene (GA2ox) were up-regulated after exogenous GA treatment. The active GAs can bind with GID proteins to form a complex in plants. The GID genes were up-regulated and DELLA genes were down-regulated after exogenous GA treatment ( Figure 6B ).

The biosynthesis of cellulose was related to the cell wall formation in plants. A total of 59 cellulose synthase (CES) family members were identified from the genome of water dropwort. Based on the phylogenetic relationships with Arabidopsis, the CES proteins of water dropwort can be divided into different cellulose subfamilies, including synthase (CESA) and cellulose synthase-like (CSLA/B/C/D/E/G) subfamilies ( Figure 7 ).

A total of 22 members were detected from CESA subfamily, and the cis-acting elements analysis were conducted by PlantCARE ( Figure 8A ). The gibberellin-responsive elements were discovered in the promoters of CESA family genes, suggesting these CESA genes may be regulated by the gibberellin signal in water dropwort. In addition, the CESA genes also contained abundant other hormone-related elements, such as auxin responsiveness elements, MeJA responsiveness elements, and abscisic acid responsiveness elements. The biosynthesis of cellulose was catalyzed by the CESA proteins, the transcriptions of CESA genes were analyzed in water dropwort ( Figure 8B ). The results indicated that most CESA genes were up-regulated under GA₃ treatment. The changes of gene expression promoted the biosynthesis of cellulose, which further provided the materials for the cell wall in the cell elongation of water dropwort under GA₃ treatment.