All plant materials were cultivated in a greenhouse under a controlled environment of 22°C with a 16-h light/8-h dark photoperiod. For JA response analysis, 1-month-old S. baicalensis seedlings were transferred to an aqueous solution containing 200 μM MeJA. The roots of the seedlings were collected at four time points, including 0, 1, 3, and 7 h of MeJA treatment.

Roots of S. baicalensis seedlings subjected to MeJA treatment for 0, 1, 3, and 7 h were used for transcriptome analysis. Total RNA was extracted using TRIzol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. LC-Bio Technology Co., Ltd. (Hangzhou, China) generated the sequencing libraries and conducted sequencing. Reference transcriptome data are deposited in the Sequence Read Archive (SRA) under accession number SRA: PRJNA961700.

WGCNA was conducted using the OmicStudio tools available at https://www.omicstudio.cn/tool/WGCNA. A pairwise Pearson’s correlation matrix was constructed and transformed into a weighted matrix. To create a topological overlap matrix, a soft threshold of 0.85 was applied. Module-trait associations were estimated to identify the relationship between modules and various treatment stages, by calculating the correlation between the module eigengene and the flavone contents at each infection stage. This approach facilitated the identification of expression modules that were highly correlated with the flavone contents. The networks were visualized using Cytoscape software.

S. baicalensis transcription factor genes were identified through a comparative analysis of transcripts assembled from RNA-seq data against the PlantTFDB database using the BlastX algorithm. Candidate SbMYC2 genes were further selected by BLAST using the AtMYC2 sequence as query. Additionally, the hidden Markov model (HMM) profiles PF06200 (TIFY domain) and PF09425 (Jas motif) of the JAZ family were retrieved from the Pfam database, and the transcriptome sequence was subjected to hmmsearch analysis to identify candidate JAZ genes. Finally, a phylogenetic tree was constructed with all candidate TFs using MEGA7 software to facilitate visualization and interpretation of the results.

The coding region of SbWRKY75 was cloned into the pHB vector to generate the pHB : SbWRKY75 overexpression vector. A fragment of the SbWRKY75 coding sequence (bp 9–293) was selected for RNA interference (RNAi) and cloned into the intermediate vector pENTR. The target fragment was then recombined into the pHELLSGATE vector via LR Clonase to obtain the RNAi interference construct pHELLSGATE : SbWRKY75 ( Figure S2 ). The pHB : SbWRKY75 and pHELLSGATE : SbWRKY75 clones were subsequently transformed into Agrobacterium (Agrobacterium tumefaciens) strain GV3101.

The seeds were washed with running water for a duration of 30 minutes, followed by a cold storage period of 72 hours at 4°C. Subsequently, the seeds were immersed in 70% ethanol for 30 seconds and rinsed thrice with sterile water to ensure surface sterilization. A 1% sodium hypochlorite treatment was then administered for 15 minutes, followed by three additional rinses with sterile water. The surface-sterilized seeds were then inoculated onto Murashige and Skoog (MS) medium to facilitate the cultivation of sterile seedlings. Leaf and stem segments were excised from S. baicalensis plantlets and used as explants to induce somatic embryogenesis. The explants were immersed in Agrobacterium suspension (OD=0.6) with uniform shaking in the dark for 10 min. After being blotted dry onto sterile tissue paper, the explants were transferred to MS medium with 10.0 mg/L 1-naphthaleneacetic acid. The inoculated explants were selected on selective medium containing 200 mg/L timentin and cultivated at 25°C in the dark to induce somatic embryogenesis. After the formation of somatic embryos, they were subjected to an alternating light/dark cycle of 16 hours of light and 8 hours of darkness under a light intensity of 2,000 lux. The seedlings were then transferred to 1/2 MS medium to induce root formation. The rooted seedlings were transferred to flower discs containing substrate for further cultivation. Genomic DNA was extracted from the young leaves of transgenic S. baicalensis seedlings for confirmation of transformation.

The promoter fragments of baicalin biosynthesis pathway genes were cloned into the pLacZ vector as bait. The coding sequences of SbWRKY75, SbWRKY41, and SbRAP2.1 were recombined into the pB42AD vector as prey. The primer sequences are listed in Table S1 . Different combinations of vectors were co-transformed into yeast strain EGY48. Transformants were selected on synthetic defined (SD) medium lacking tryptophan and leucine. Positive yeast colonies were grown for 3 days at 30°C, collected, and resuspended in sterile water, before being spotted onto an X-gal medium plate, which was then incubated at 30°C for 24 h. Negative controls were established using the empty pLacZ and pB42AD plasmids.

The coding sequences of SbWRKY75, SbWRKY41, and SbRAP2.1 were cloned into the pET32a expression vector to produce His-tagged fusion proteins. The sequences of the primers used for cloning are provided in Table S1 . The resulting plasmids were transformed into E. coli BL21 cells, and the production of recombinant proteins was induced by adding 0.1 mM IPTG at 18°C for 12 h. The bacterial cells pelleted by centrifugation and lysed using a low-temperature ultra-high-pressure continuous flow cell disrupter (JNBIO). The supernatant was purified using Ni-NTA resin (Qiagen, Cat. No. 30210), and the bound proteins were eluted with 200 mM imidazole and subsequently concentrated using a Millipore (Merck, Cat. No. UFC903024) ultrafiltration device.

To investigate whether SbWRKY75, SbWRKY41, and SbRAP2.1 can bind to the promoters of their target genes, electrophoretic mobility shift assays (EMSAs) were conducted. The probes used in EMSA were labeled with digoxigenin; their sequences are listed in Table S1 . The EMSA reactions were performed using a DIG Gel Shift Kit, 2nd Generation, following the manufacturer’s instructions (Roche). The reaction products were loaded onto a 5% polyacrylamide gel and electrophoresed at 80 V for 90 min. The DNA-protein complexes were transferred to a nylon membrane for 60 min at 300 mA and crosslinked at 120 MJ. Probe detection was carried out according to the protocol provided by Roche.

The coding sequences of SbWRKY75 and SbWRKY41 without a stop codon were individually cloned into the fusion YFP vector pHB through one-step cloning. The resulting constructs were transformed into Agrobacterium strain GV3101 and infiltrated into Nicotiana benthamiana leaves using needleless syringes. YFP fluorescence in lower epidermal cells was observed 48 h to 72 h after Agrobacterium infiltration using a confocal microscope (Leica TCS SP8). Primers used are listed in Table S1 .

To perform the dual-luciferase reporter gene assay (dual-LUC), the promoter sequences of SbCLL-7, SbF6H, and SbUGT were cloned individually into pGreen II 0800 by one-step cloning (Lv et al., 2019). The resulting constructs were transformed into Agrobacterium strain GV3101. N. benthamiana leaves were infiltrated with the Agrobacterium cultures using needleless syringes. The leaves were collected 48 h after infiltration for dual-LUC assays, which were performed using a Dual-Luciferase Reporter Assay System according to the manufacturer’s instructions (Promega). Primers used are listed in Table S1 .

The quantification of baicalin, baicalein, wogonoside, and wogonin was carried out using an Agilent 6410 Triple Quad LC/MS system. Root samples of S. baicalensis were collected and homogenized to obtain a fine powder, which was then subjected to ultrasonic extraction with 70% (v/v) ethanol at 60°C for 1 h. The obtained extracts were centrifuged at 10,000 g at 4°C for 15 min and subsequently filtered through a Millipore Millex-HV filter (Merck Millipore, Billerica, MA, USA). To determine the flavone content, an absolute quantification method with external standards was employed. Calibration curves were constructed using standard solutions of baicalin, baicalein, wogonoside, and wogonin at known concentrations. The analytical parameters of the Agilent 6410 Triple Quad LC/MS were as follows: column, Waters T3 column (2.1 mm × 100 mm, 3 μm); oven temperature, 350°C. The mobile phase contained D: water (containing 0.1% [v/v] formic acid); C: acetonitrile; flow rate, 0.3 mL/min; column temperature, 50°C; and injection volume, 5 μL.

Total RNA was extracted from the roots of control S. baicalensis seedlings and those of seedlings treated with 200 µM MeJA for 0, 1, 3, or 7 h using TRIzol Reagent. Total RNA samples was then reverse-transcribed to first-strand cDNA using PrimeScript II RTase (Takara cat. no. 6110A). Primer pairs were designed using Primer Premier 5 software. Quantitative PCR was performed on a QuantStudio 3 System following the manufacturer’s protocol. The relative expression levels of target genes were calculated using the 2^–ΔΔCt method with SbACT7 as the reference gene. The primer sequences are available in Table S1 .

All boxplots, bar graphs, and connecting lines were generated using Origin 2018 Software. The error bars in the figures represent the standard deviation of the means obtained from three independent biological replicates.

To investigate the effects of MeJA treatment on baicalin biosynthesis in S. baicalensis roots, we treated plants with an aqueous MeJA solution for 0, 1, 3, or 7 h ( Figure 1A ). We determined that the levels of baicalein, baicalin, wogonin, and wogonoside change rapidly over the course of the treatment ( Figure 1B ). Specifically, the levels of baicalein, baicalin, wogonin, and wogonoside significantly decreased by 54%, 96%, 31% and 97%, respectively, after 1 h of MeJA treatment compared to the 0-h samples. However, after 7 h of MeJA exposure, the levels of these compounds had increased to reach 95% (baicalein), 19% (baicalin), 113% (wogonin), and 13% (wogonoside) of their levels in the 0-h samples. These results suggest that the treatment resulted in a substantial decrease in flavone glycoside content, followed by a slow recovery. Conversely, the decrease in glycoside content was modest and this essentially returned to pre-treatment levels after 7 h of MeJA treatment. Therefore, MeJA treatment has a significant influence on the flavone contents of S. baicalensis roots.

We examined the expression of various genes known to encode enzymes involved in baicalin biosynthesis in S. baicalensis ( Figure 1C ). These genes consisted of SbPAL-1 (Phenylalanine ammonia-lyase 1), SbPAL-2, SbPAL-3 (Park et al., 2012), SbCLL-7 (Cinnamate-CoA ligase 7), SbCHS-2 (Chalcone synthase 2), SbCHI (Chalcone isomerase), SbFNSII-2 (Flavone synthase II-2), SbF6H (Flavonoid 6-hydroxylase) (Zhao et al., 2016b), SbUGT (UDP-Glucosyl transferase) (Nagashima et al., 2000) and SbGUS (BETA-D-Glucuronidase) (Ikegami et al., 1995). We observed significant changes in the expression levels of these genes in seedling roots upon treatment with exogenous MeJA, in agreement with the observed alterations in flavone contents described above. These findings suggest that the MeJA-induced regulation of flavone biosynthesis in S. baicalensis is mediated through transcriptional modulation of various key biosynthetic genes in this pathway.

We employed the WGCNA R package to identify modules of co-expressed genes in response to MeJA treatment. With a soft minimal threshold for Pearson’s correlation coefficient of 0.85, we identified 15 co-expression modules, as defined on differential expression of their constituent genes, with each module being indicated by a different color ( Figure 2A , Table S2 ). We clustered the consensus module eigengenes and visualized them to analyze module-phenotype correlations. Among the different gene expression modules, we found that the ‘MEturquoise’ module exhibited a strong and positive correlation with both baicalin and wogonoside contents, suggesting that the genes within this module play important roles in regulating the biosynthesis of these two compounds. Additionally, the ‘MEbrown’ module showed a strong and positive correlation specifically with wogonin content, indicating that the genes within this module may be involved in the specific biosynthesis of wogonin ( Figure 2B ). KEGG pathway enrichment analysis revealed that the ‘MEturquoise’ module does not contain genes directly involved in the phenylpropanoid pathway. Therefore, we focused on the ‘MEbrown’ module for subsequent analysis ( Figure S1 ). We selected genes encoding baicalin biosynthetic enzymes and TFs within the top 30% genes from the ‘MEbrown’ module as candidate genes and identified key transcription factor genes regulating baicalin biosynthesis through correlation analysis ( Figure 2C ).

To investigate the TFs involved in regulating baicalin biosynthesis during MeJA processing, we looked at the number of differentially expressed TF genes in different TF gene families at different time points. We observed that the bHLH, ERF, MYB, NAC, and WRKY families are represented by the highest numbers of differentially expressed genes, with the proportion of differentially expressed TF genes in the ERF and WRKY families being particularly high ( Figure 2D ). Notably, TFs in the WRKY and ERF families have been shown to play important roles in flavonoid biosynthesis (Lloyd et al., 2017; Wu et al., 2022). Therefore, we chose five WRKY TF genes and four ERF TF genes with a p<0.01, as determined by Pearson’s correlation analysis with baicalin biosynthesis pathway genes, from the top 30% central TF genes in the ‘MEbrown’ module defined by WGCNA analysis. We investigated the function of these selected TF genes below.

We cloned the promoter sequence of several baicalin biosynthesis pathway genes into the pLacZ vector. To identify TFs potentially binding to these promoter sequences, we conducted a yeast one-hybrid (Y1H) assay using the above five WRKY TF genes and four ERF TF genes. As indicated by the blue color of the corresponding yeast colonies harboring the relevant constructs, we determined that SbWRKY75 (Sb01t34690) bound to the SbCLL-7 promoter, while SbWRKY41 (Sb06t15420) bound to the SbF6H and SbUGT promoters, and SbRAP2.1 (Sb08t10780) bound to the SbCLL-7 promoter ( Figure 3A ).

We validated the above results by conducting an electrophoretic mobility shift assay (EMSA) and a dual-luciferase (LUC) assay. To this end, we produced recombinant SbWRKY75 and SbRAP2.1 fused to a thioredoxin tag and incubated the purified proteins with digoxigenin-labeled probes for the SbCLL-7 promoter. We detected specific binding of both recombinant proteins to labeled probes, as evidenced by the lower mobility of the probes from the corresponding target genes ( Figure 3B ). Importantly, this lower mobility was gradually competed by incubation with unlabeled probe ( Figure 3B ). In the dual-LUC assay, using the reporter constructs proSbCLL-7:LUC, proSbF6H:LUC, and proSbUGT : LUC, we observed that co-expression of SbWRKY75 with the proSbCLL-7:LUC reporter construct resulted in a 103% increase in the LUC/REN ratio (firefly to Renilla luciferase activity). Additionally, co-expression of SbWRKY41 with the proSbF6H:LUC and proSbUGT : LUC reporter constructs resulted in a 113% increase and a 60% increase in the LUC/REN ratio, respectively ( Figure 3C ).

We also generated constructs consisting of the SbWRKY75 or SbWRKY41 coding sequence cloned in-frame with the sequence of yellow fluorescent protein (YFP) and driven by the cauliflower mosaic virus (CaMV) 35S promoter to establish the subcellular localization of these two TFs. We detected YFP fluorescence for SbWRKY75:YFP and SbWRKY41:YFP mainly in the nucleus of N. benthamiana leaf cells, indicating that SbWRKY75 and SbWRKY41 localize to the nucleus ( Figure 3D ).

To investigate the effect of SbWRKY75 on the baicalin biosynthesis pathway in S. baicalensis, we transiently overexpressed SbWRKY75 via Agrobacterium-mediated transient transformation. We detected a significant 73% increase in SbWRKY75 transcript levels in the roots of these transiently overexpressing plants ( Figure 4A ). Conversely, we measured about a 43% decrease in SbWRKY75 transcript levels in S. baicalensis plants transiently expressing an RNAi construct specific for SbWRKY75 ( Figure 4B ). We then analyzed the relative expression levels of ten baicalin biosynthesis genes in these SbWRKY75 overexpression and RNAi plants. The expression levels of SbCLL-7, SbPAL-1, SbPAL-2, SbCHS-2, SbFNSII-2, SbF6H, and SbUGT were significantly upregulated in the SbWRKY75 overexpression seedlings, showing approximately 0.9-, 1.1-, 1.77-, 0.45-, 7.70-, 1.64-, and 0.17-fold increases, respectively. Conversely, the expression levels of SbCLL-7, SbPAL-1, SbCHS-2, SbCHI, SbUGT, and SbGUS were approximately 84%, 34%, 21%, 59%, 48%, and 84% lower in the SbWRKY75 RNAi seedlings compared to wild-type S. baicalensis seedlings.

We also measured baicalin contents in the roots of SbWRKY75 overexpression and RNAi seedlings ( Figure 4C ). Compared to those in wild-type S. baicalensis seedlings, we detected baicalin and wogonoside levels that were higher by 14% and 13%, respectively, in SbWRKY75 overexpression seedlings, whereas the levels of these compounds decreased by 22% and 19% in the SbWRKY75-RNAi seedlings. These results further support a role for SbWRKY75 in regulating baicalin accumulation.

The altered expression of baicalin biosynthesis genes in MeJA-treated roots suggested that transcriptional regulation contributes to JA-induced baicalin biosynthesis. MYC TFs and JAZ proteins are critical regulators of the JA signaling pathway, making them potential targets for investigation in S. baicalensis (Li et al., 2017). Analysis of SbMYC2.1 and SbMYC2.2 revealed the presence of conserved domains shared with their homologs in Arabidopsis ( Figure 5A ). In addition, we identified three and one W-boxes in their promoters, respectively, raising the possibility that their transcription may be regulated via direct binding by WRKY TFs. Moreover, using HMMsearch, we identified 11 members of the JAZ family in S. baicalensis and reconstructed a phylogenetic tree together with JAZ proteins from Arabidopsis ( Figure 5B ). This analysis revealed that SbJAZ3 is similar to SbJAZ7, AtJAZ1, and AtJAZ2, while SbJAZ4 was more similar to AtJAZ5 and AtJAZ6. Additionally, SbJAZ5 clustered together with AtJAZ9, suggesting that these TFs may exert similar functions in plants. However, SbJAZ8, SbJAZ9, and SbJAZ10 were grouped together in a distinct branch, indicating that they may have undergone unique changes in their evolution and have different biological functions.

Turning to our RNA-seq analysis of gene expression in the roots of S. baicalensis following MeJA treatment, we observed similar upregulation of gene expression patterns after MeJA treatment for SbWRKY75, SbJAZ9, SbJAZ7, SbWRKY41, SbMYC2.1, and SbJAZ3 ( Figure 5C ), indicating that SbWRKY75 and SbWRKY41 are necessary for JA-induced baicalin biosynthesis and suggesting that they may regulate the transcription of both SbMYC2s and SbJAZs.

We investigated the binding of SbWRKY75 to the promoters of SbJAZ7 and SbJAZ9, and of SbWRKY41 to the promoters of SbMYC2.1 and SbJAZ3 using a Y1H assay. We tested various promoter fragments, finding that SbWRKY41 binds to the W2-box motif in the SbMYC2.1 promoter and to the W1-box motif in the SbJAZ3 promoter ( Figure 6A ). To confirm this interaction, we purified full-length recombinant SbWRKY41 and performed an EMSA using digoxigenin-labeled fragments containing the W2-box motif from the SbMYC2.1 promoter and the W1-box from the SbJAZ3 promoter motif as probes. We determined that SbWRKY41 binds to the W-box motifs of the SbMYC2.1 and SbJAZ3 promoters; importantly, binding decreased when unlabeled probes were added as competitors ( Figure 6B ). This finding indicates that SbWRKY41 specifically binds to the W-box motifs in the SbMYC2.1 and SbJAZ3 promoters.

The enrichment of W-box motifs in WRKY promoters suggests that WRKY TFs can regulate their own expression or cross-regulate with other WRKY TFs (Maleck et al., 2000). To investigate the potential regulatory relationship between SbWRKY75 and SbWRKY41, we conducted Y1H and EMSA experiments. We established that SbWRKY41 can modulate the expression of SbWRKY75 and may function upstream of SbWRKY75 ( Figure 6B ).

The construction of gene regulatory networks is a powerful approach for identifying TF-gene interactions in diverse biological contexts, and WGCNA has demonstrated superior performance in detecting functionally associated gene pairs through examination of pairwise correlations in gene expression patterns (Kuang et al., 2021). In this study, we performed a transcriptome analysis of control S. baicalensis roots and of roots exposed to MeJA treatment for 1, 3, or 7 h to establish a co-regulated gene network associated with baicalin biosynthesis. Through this analysis, we identified nine TF genes and their corresponding regulatory pathways responsible for baicalin biosynthesis. We confirmed the validity of the network through Y1H assays, EMSAs, and dual-LUC assays, which demonstrated that SbWRKY75, SbWRKY41, and SbRAP2.1 bound to the promoters of their respective target genes.

We showed that SbWRKY75 activated transcription from the SbCLL-7 promoter, while SbWRKY41 activated that from two baicalin biosynthesis pathway genes. Our investigation indicated that the overexpression and RNA interference of SbWRKY75 in transiently transformed seedlings significantly influenced the expression of baicalin biosynthetic genes and the accumulation of baicalin. Similarly, TcWRKY1 is a TF encoded by a gene whose expression is induced by MeJA and enhances the expression of DBAT when overexpressed in Taxus chinensis suspension cells. Moreover, RNAi-mediated knockdown of TcWRKY1 transcript levels decreases DBAT transcript levels, thereby decreasing paclitaxel biosynthesis (Li et al., 2013). We established here that SbWRKY75 acts as a positive regulator of baicalin biosynthesis by modulating the expression of genes involved in its biosynthetic pathway. SbWRKY75 overexpression may affect the expression of SbCLL-7 or modulate other signaling pathways, thereby either directly or indirectly influencing the expression of baicalin biosynthesis pathway genes. However, as plants seek to maintain a balance between growth, development, and secondary metabolite biosynthesis, negative feedback regulation can inhibit gene expression when the concentration of secondary metabolites is too high. Thus, some baicalin pathway genes in S. baicalensis may experience a more modest decrease or change in their expression levels. Previous studies have demonstrated that WRKY TFs are closely associated with anthocyanin accumulation in S. baicalensis (Zhang et al., 2022). Our findings suggest that both SbWRKY75 and SbWRKY41 transcriptionally activate the expression of baicalin biosynthesis genes, thereby positively regulating baicalin accumulation.

In addition, we demonstrated that SbWRKY41 also regulates the expression of SbMYC2.1 and SbJAZ3. MYC2 and JAZ proteins are key regulators of the JA signaling pathway, interacting with TFs that control plant secondary metabolism (Li et al., 2022). JA induces the expression of WRKY TF genes and regulates the activity and stability of their encoded proteins through various modifications, including phosphorylation and acetylation (Mao et al., 2007; Liao et al., 2016). Furthermore, several WRKY TFs have been implicated in the JA signaling pathway, such as WRKY40a in peppers (Capsicum annuum), which directly regulates the expression of CaJAZ8 to modulate JA signaling (Raffeiner et al., 2022), and GhMYC2, which serves as a link between the WRKY–MAPK cascade and flavonoid biosynthesis in upland cotton (Gossypium hirsutum) (Wang et al., 2022a). Given the interaction between SbWRKY41, SbMYC2.1, and SbJAZ3, we speculate that JA-responsive SbWRKY41 also regulates baicalin biosynthesis via the JA signaling pathway in S. baicalensis.

Both the Y1H assay and EMSA revealed potential interactions between SbWRKY41 and SbWRKY75, suggesting that these two TFs may form a regulatory network to optimize the signal transduction and cooperatively regulate baicalin biosynthesis. In plants, WRKY TFs interact and cross-regulate one another, and many possess self-regulatory feedback (Turck et al., 2004). For instance, WRKY6 can inhibit its own activity by binding to the W-box in the promoter of its encoding gene and that of WRKY42 in Arabidopsis (Robatzek and Somssich, 2002). Similar observations have been reported for CaWRKY6 and CaWRKY40 in pepper, with CaWRKY6 showing functional and expression similarities with CaWRKY40 (Cai et al., 2015). Based on the regulation of SbWRKY75 by SbWRKY41, it is reasonable to speculate that SbWRKY41 and SbWRKY75 may functionally interact and form a transcriptional network to achieve optimal coordination and fine-tuning of baicalin biosynthesis.

Based on the above findings, we propose a mechanistic model outlining the possible mode of action of SbWRKY75 and SbWRKY41 in regulating the biosynthesis of baicalin in S. baicalensis. In the presence of bioactive JAs, SbWRKY75 and SbWRKY41 are activated, with SbWRKY41 in turn regulating the expression of SbWRKY75. Additionally, SbWRKY41 participates in the JA signaling pathway by modulating the expression of SbMYC2.1 and SbJAZ3. By binding to the W-box cis-elements present in the promoters of baicalin biosynthetic genes, SbWRKY75 and SbWRKY41 activate the transcription of these genes. Our findings provide evidence for a positive regulatory role for SbWRKY75 and SbWRKY41 in baicalin biosynthesis in S. baicalensis, thereby offering a promising strategy for enhancing the production of baicalin through genetic engineering of specific TF genes.