Apocynum venetum seeds collected from Dongying City in China’s Shandong Province and stored in the laboratory were used in this study. In transgenic experiments, Arabidopsis thaliana Columbia-0 (Col-0) was used as the wild type. Our previous study reported on these transgenic plants (AvF3’H-OE, AvF3H-OE, and AvFLS-OE) (Xu et al., 2021).

A. venetum seeds were grown in vermiculite pots in a growth chamber at 24 °C with a 14-hour photoperiod and watered with 10% liquid MS. To investigate the expression patterns of AvF3’H, AvF3H, and AvFLS, six-week-old seedlings were watered with a 100 mM NaCl solution and the second completely spread (full grown) leaf from the top was collected at 0 min, 5 min, 15 min, 30 min, 1 h, 2 h, 4 h, 8 h, 12 h, 24 h, 48 h, and 72 h after watering with NaCl and immediately frozen in liquid nitrogen before storing them at -80 °C.

The sterile transgenic (AvF3’H-OE, AvF3H-OE, and AvFLS-OE) and wild type Arabidopsis thaliana seeds were grown in the growth chamber described above on 1/2 MS growth medium. After 10 days, some of these seedlings were transplanted into plastic trays. Three-week-old seedlings on plastic trays were evenly divided into two groups; one group was watered with water, while the other was watered with a 100 mM NaCl solution. After a week of treatment, the aerial phenotype was photographed with a digital camera (Canon 5D Mark III, Japan) and data was collected with Image J software. The seedlings were immediately frozen in liquid nitrogen and stored at -80 °C for further analysis, which included total flavonoid and compound composition content analysis, Na⁺ and K⁺ content detection, antioxidant enzymes activity test, ROS staining, and transcriptome analysis. The remainder of the original lot’s seedlings were transferred to 1/2 MS solid medium with or without 100 mM NaCl. Photographs of the seedlings were taken two weeks later, and data was collected using Image J software. Transgenic and WT seedlings were also transferred to 1/2 MS medium containing 100 mM NaCl with or without 100 µM rutin and eriodictyol for the flavone component complementation experiment. ROS staining was performed on both treated and untreated plants after a week. The flow chart of the experimental design is shown in Supplementary File 1 .

The A. venetum genome database (https://www.ncbi.nlm.nih.gov/genome/?term=txid377125[orgn]) was used to extract the 2000-bp sequences of flavonoid synthetase enzyme gene promoters from the NCBI database. The promoter Cis-acting elements were detected and identified using the PLACE database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/).

The total flavonoids extraction method was based on a previous method (Zhang et al., 2015) with minor modifications. In brief, 50 mg dry sample powder was added to 3 mL 80% (w/v) methyl alcohol extracting in a water bath at 95 °C overnight. In a total of 110 µL supernatant, 440 µL NaNO₂ (0.066 M, room temperature for 5 minutes), 60 µL AlCl₃ (0.75 M, 6 minutes), and 400 µL NaOH (0.5 M) were added in turns. The absorbance of the mixture was measured at 510 nm. The total flavonoids were presented as mg rutin equivalents, i.e. mg/g DW. HPLC analysis method was used to detect the main flavonoid compounds content in transgenic Arabidopsis plants (Xu et al., 2020).

The Na⁺ and K⁺ concentration measurement method was based on a previous method (Xu et al., 2020) with minor modifications. In brief, 0.25 g dried sample powder was digested in 5 mL of HNO₃ at 110°C for about 6 hours until a colorless liquid was obtained. The cooled liquid mixture was diluted to 10 mL with deionized water. The Na⁺ and K⁺ content was then determined using an optical emission spectrometer (ICP, Optima 8000, PerkinElmer, USA)

The SOD activity detection method was based on a previous method (Stewart and Bewley, 1980). SOD activity was defined as the ability of inhibiting the photochemical reduction of nitro blue tetrazolium. CAT activity was determined according to the method of (Patra et al., 1978), in which the ability of CAT was related to the extinction of H₂O₂. POD activity was measured by referring to the methodology of (Nickel and Cunningham, 1969).

The histochemical staining methods using 3′3′-diaminobenzidine (DAB) and nitro blue tetrazolium (NBT) to detect H₂O₂ and O₂•^-, respectively, was referred (Zhao et al., 2016). Likewise, the H₂O₂ content detection method was based on a previous method reported by (Libik et al., 2005). The O₂•^- content was detected with a specific kit (SA-2-G; Cominbio, Suzhou, China) following the manufacturer’s instructions.

IAA and JA detection was done according to (Du et al., 2013) with miner modification. In brief, twelve seedlings of transgenic plants or WT samples under salt stress were ground into powder. 100 mg of powder were extracted twice with 900 µL of extraction buffer [methanol:H₂O:acetonitrile = 90:9:1 (v/v)]. Quantification was performed in an ABI Qtrap6500 LC-MS system (Applied Biosystems, USA) with stable-isotope-labeled ABA and auxin as standards (Sigma).

Total RNA was extracted from samples with the EasyPure Plant RNA Kit (TransGen, ER301-01, Beijing, China) and genomic DNA contamination was removed using RNase-free DNase I (TransGen, K21109). First-strand cDNA synthesis and qRT-PCR amplification were performed as previously described by (Wang et al., 2018). Supplementary File 2 lists the gene-specific primers that were used for qRT-PCR. The relative expression levels of target genes were calculated using the 2^-ΔΔCt method (Livak and Schmittgen, 2001) and normalized to those of the AtTUBULIN reference gene (Xu et al., 2017). For transcriptome analysis, the RNA purification, reverse transcription, library construction and sequencing were performed at Shanghai Majorbio Bio-pharm Biotechnology Co., Ltd. (Shanghai, China). Transcriptome sequencing data were analyzed on their online platform (www.majorbio.com) (Ren et al., 2022). The transcripts per million reads (TPM) was used to calculate the expression levels of each gene using RSEM (http://deweylab.biostat.wisc.edu/rsem/). Differential expression genes (DEGs) analysis was performed using DESeq2 (Love et al., 2014). The Q value ≤ 0.05 and |log2(Fold Change)| > 1 were used to identify DEGs. The GO and KEGG functional enrichment pathway analyses were performed using GO (https://github.com/tanghaibao/Goatools) and KOBAS (http://kobas.cbi.pku.edu.cn/home.do) tools.

All charts (bar charts, boxplots, bubble charts, Venn diagrams and heatmap, so on) were drawn using Adobe Illustrator cs5 software and Majorbio Cloud Platform (www.majorbio.com). Significant differences among all the experimental treatments were determined using a one-way analysis of variance (ANOVA) followed by Tukey’s test at the probabilities of p < 0.05 level with SPSS version 17.0 software (SPSS, Chicago, IL, USA).

The main biosynthesis pathways of flavonoid metabolism are integrated in Figure 1A . The red marked genes (AvF3’H, AvF3H, and AvFLS) were the main focus in this study. The qRT-PCR results revealed that the expression patterns of these three genes in response to salt stress were similar. Under salt stress, they were induced to express at a high level for four hours. The expression level peaked at 5 minutes and 2 hours, with AvF3’H showing the highest level of expression compared to the other two. Salt stress inhibited the expression of three genes after four hours of exposure ( Figure 1B ). The expression patterns of other important genes in the flavonoid metabolic pathway were also studied, and the expression profiles were divided into two distinct groups. The expression patterns of AvCHI, Av4CL3, AvF3’5’H, AvRT, AvCHS, AvPAL1, and AvC4H2 were similar to that of AvFLS, AvF3H, and AvF3’H, which showed a higher expression level in the first 4 hours, and reached the peak within 2 to 4 hours ( Supplementary File 3 ). The expression patterns of others genes such as Av4CL1, Av4CL2, AvC4H1, AvC4H3, AvPAL2, Av4CL4, and Av4CL were clustered together. They showed rapid response to salt stress. The expression of these genes was rapidly induced in the first 30 minutes and then started to decline. It reached a low point at 4 hours, after which it gradually began to rise again ( Supplementary File 3 ). The Cis-acting elements in the gene promoter, which usually respond to the types of binding transcription factors, were identified in the PLACE database to help elucidate the potential regulatory mechanism of these genes’ expression under salt stress. From the 2 kb promoters of AvF3’H, AvF3H, and AvFLS, a total of 133 different non-repetitive Cis-acting elements were identified ( Supplementary File 4 ). These Cis-acting elements were found to be primarily involved in abiotic stress (32%), miscellaneous function (17%), hormones (16%), biotic stress (7%), and seed development (6%). ( Figure 1C ). The top ten Cis-acting elements in each gene were listed. “DOFCOREZM,” “CACTFTPPCA1,” and “ROOTMOTIFTAPOX1” were found in the highest frequency in all three genes, indicating that they may be involved in basic functions. The frequency of a salt-stress response element “GT1GMSCAM4” in AvF3’H, on the other hand, was significantly higher than in AvF3H and AvFLS ( Figure 1D ), suggesting that AvF3’H may play an important role in response to salt stress.

OE-transgenic Arabidopsis plants were developed to investigate the potential roles of AvF3’H, AvF3H, and AvFLS in the response to salt stress. Semi-quantitative PCR was used to confirm the expression of AvF3’H, AvF3H, and AvFLS in transgenic plants ( Supplementary File 5 ). Under normal growth conditions, the aerial parts of all three transgenic lines had longer petioles and larger leaves than WT, however, the statistical analysis results were not significant. Under salt stress, these parameters were significantly larger in transgenic plants than WT ( Figures 2A, D–F ), indicating that these three genes have a salt tolerance function. Phenotypic and statistical analysis revealed that there was no significant difference in the main root length of transgenic plants and WT under normal conditions ( Figures 2B, G ). Under 100 mM NaCl stress, however, transgenic plant roots were significantly longer than WT roots ( Figures 2C, G ). Furtherly, the endogenous flavonoid content results of these plants revealed that transgenic plants owed the higher flavonoid than WT both under normal and salt stress conditions ( Figure 2H ), suggesting a potential link between flavonoid content and salt tolerance. HPLC analysis was used to detect the flavonoids ingredients content. The content of rutin, eriodictyol, and naringerin in transgenic plants and WT were somewhat complex under normal conditions. However, all the transgenic plants showed significantly higher content compared to WT under salt stress ( Figures 2I–K ), suggesting the potential positive roles of them in response to salt stress.

Transgenic plants had significantly higher Na⁺ content under normal and salt stress than WT plants. Furthermore, the Na⁺ content of WT and transgenic plants was 3.6 to 7.9 times higher in saline than in normal conditions ( Figure 2L ). Interestingly, the K⁺ content in WT was 2.7 times higher under salt stress than under normal conditions, whereas the K⁺ content in transgenic plants decreased by 39.86%, 18.54%, and 33.46% in the three transgenic lines under salinity compared to normal conditions ( Figure 2M ), resulting in a higher Na/K ratio in transgenic plants under salt stress ( Figure 2N ). These findings suggested that overexpression of the flavonoid synthetase genes could increase the Na⁺ content and reduce the K⁺ content in transgenic plants, resulting in the ion homeostasis changes.

Salt stress is one of the detrimental environmental stresses inducing factors that lead to ROS accumulation. Thus, using 3,3’-diaminobenzidine (DAB) and nitrotetrazolium blue chloride (NBT) staining, histochemical analysis was used to investigate the effect of two ROS species, hydrogen peroxide (H₂O₂) and superoxide anion (O₂•^-). Under salt stress, the accumulation of brown and blue precipitates (showing DAB and NBT staining, respectively) in transgenic lines was much lower than in WT plants ( Figures 3A, B ). These findings suggest that transgenic plants’ improved salt stress tolerance may be due to reduced ROS production, which correlates with higher endogenous flavonoid levels in the three transgenic plants ( Figure 2H ). As expected, control plants with exogenous rutin and naringerin were significantly less damaged by ROS than those without under salt stress ( Figures 3A, B ). According to the spectrophotometry analysis results, the content of H₂O₂ and O₂•^- that accumulated in the leaves of these transgenic plants showed no significant difference with that of WT under normal condition, however, significantly reduced compared with the WT plants under salt stress ( Figures 3C, D ). To evaluate the plants abiotic stress damage caused by salt stress, the MDA content was measured. As shown in Figure 3E , there was no significant difference in MDA content between the transgenic and WT plants under normal conditions. However, the MDA content in transgenic plants was significantly reduced (28.39% to 32.32%) compared to WT under salt stress. The antioxidant scavenging system, on the other hand, was activated in the transgenic plants. Under salt stress, the activity of CAT in both WT and transgenic plants was significantly higher than that under normal conditions, with the transgenic plants showing significantly higher activities than the control. Although the activities of SOD and POD were inhibited in WT and transgenic plants under salt stress (except SOD in AvF3’H-OE) compared to normal conditions, the activities of SOD and POD in transgenic plants were significantly higher than those in WT plants under salt stress ( Figures 3F, H ).

In three independent experiments, we performed transcriptome deep sequencing (RNA-seq) analysis on transgenic and WT plants treated with 100 mM NaCl to further investigate the molecular mechanism of endogenous flavonoid accumulation in reducing the effect of salt stress. Three biological replicates of each treatment were compactly gathered together in the PCA diagram, indicating that the RNA-seq results were stable and reliable. Furthermore, while the four treatment samples were clearly separated, AvF3H-OE was separated from the other three samples, indicating that AvF3H overexpressing plants had a relatively distinct transcriptome profile from the other three samples ( Supplementary File 6 ). We then identified differentially expressed genes (DEGs). A total of 1299 non-repeat genes were identified in transgenic plants when compared with WT ( Supplementary File 7 ), of which 289 were identified in AvF3’H vs WT (148 up-regulated and 141 down-regulated) ( Figure 4A ), 1028 were identified in AvF3H vs WT (844 up-regulated and 184 down-regulated) ( Figure 4B ), 235 were identified in AvFLS vs WT (122 up-regulated and 113 down-regulated) ( Figure 4C ), suggesting that AvF3H overexpressing had a significant effect on gene regulatory networks. A venn map was used to identify shared DEGs among different comparison groups in order to investigate the general rule of flavonoid metabolism pathway related genes in response to salt stress. A total of 43 shared DEGs were identified, with 30 up-regulated and 13 down-regulated ( Figure 4D , Supplementary File 7 ). GO enrichment analysis revealed that some JA related terms such as “Response to jasmonic acid”, “Jasmonic acid hydrolase”, and “Regulation of jasmonic acid mediated signaling pathway” were significantly enriched in these shared DEGs ( Figure 4E ). The most enriched KEGG pathway was “Nitrogen metabolism” pathway ( Figure 4F ). These findings suggested that JA and nitrogen metabolism play important roles in the response to salt stress. Furthermore, the top 10 enrichment GO and KEGG terms from each comparison group were integrated. Finally, a total of 23 non-repeated GO terms ( Supplementary File 8A ) and 18 KEGG pathways ( Supplementary File 8B ) were listed. We noticed that GO and KEGG terms of the AvF3H vs WT comparison group possessed more DEGs compared to the other two, which is consistent with the highest DEGs identified in this comparison group. GO terms “cell wall” and KEGG pathways “MAPK signaling pathway-plant”, “plant-pathogen interaction”, and “cyanoamino acid metabolism” were shared in all three comparison groups indicating the importance of these genes’ involvement in these terms. Furthermore, hormone related GO terms and KEGG pathways, such as “regulation of jasmonic acid mediated signaling pathway”, “jasmonic acid hydrolase”, “response to jasmonic acid”, “Plant hormone signal transduction”, and “alpha-Linolenic acid metabolism”, cell wall related terms such as “cell wall” and “plant-type cell wall”, flavonoid metabolism related terms such as “flavonoid biosynthesis”, “isoflavonoid biosynthesis”, and “phenylpropanoid biosynthesis” were significantly enriched in the transgenic vs WT groups, which suggests their involvement in salt tolerance. A total of 465 DEGs were involved in this enrichment GO or KEGG terms, and most of them were up-regulated in transgenic plants under salt stress compared to WT plants ( Supplementary File 8C ).

Flavonoids, as an important member of the non-enzymatic antioxidant system, could directly eliminate the excess ROS produced by salt stress, working synergistically with the enzymatic antioxidant system. Importantly, flavonoids can also induce the expression of jasmonic acid and auxin synthesis genes via unknown pathways, resulting in the IAA and JA content increase, and activating the JA and IAA signal transduction pathways to promote the expression of cell wall biosynthesis or modification genes that promote cell wall expansion ( Figure 6 ).