B. cusia plants were grown in perennial dominant Shufeng Farm in Fujian province, China (25°25 N 118°39 C). I. indigotica plants were cultivated in the Naval Medical University’s (NMU) experimental field in Shanghai, China. Professor Hanming Zhang (School of Pharmacy) conducted the species verification. Six-month-old B. cusia plants were subjected to hormone treatments to study the influence of abiotic stress on the expression of BcTSA. The aerial parts were sprayed with a solution containing either 100 μM MeJA, 100 μM ABA, or 100 μM SA, respectively. The leaves were collected at 0, 2, 4, 6, 8, 12, and 24 hours following treatment and rapidly frozen in liquid nitrogen before being kept at -80°C. Water was used to mock-treat control plants. Three to five leaves from different plants were combined together to make a biological repeat, and three biological repetitions were performed.

Following the manufacturer’s instructions, 100 mg of B. cusia samples were used to extract total RNA using the Column Plant Total RNA Kit (TransGen, Beijing, China). RNA samples were tested for quality and concentration using ethidium bromide-stained agarose gel electrophoresis and spectrophotometer measurement on a NanoDrop 2000C Spectrophotometer (Thermo Scientific, Waltham, MA, USA). For cDNA synthesis, samples having an optical density absorption ratio at OD260/280 between 1.9 and 2.2 and an OD260/230 more than 2.0 were employed. TransScript One-Step gDNA Removal and cDNA Synthesis SuperMix (TransGen, Beijing, China) was used to create first-strand cDNA from 1 μg of total RNA, according to the manufacturer’s instructions.

The full length of the BcTSA ORF sequence was derived from the B. cusia transcriptome database. The coding sequences of the BcTSA gene was amplified by PCR from cDNA with pfu DNA polymerase (New England Biolabs, Ipswich, MA, USA) using the gene-specific primers BcTSA-F and BcTSA-R (Table S1). The targeted fragments were purified with the EasyPure Quick Gel Extraction Kit (TransGen, Beijing, China) and cloned into the T-vector with the pEASY™-Blunt Zero Cloning Kit (TransGen, Beijing, China). The generated plasmids were transformed into E. coli DH5α cells and selected by means of antibiotic resistance. The sequences were validated by DNA sequencing via Sangon (Shanghai, China).

To identify the whole ORF of TSA sequences, NCBI ORF Finder (http://www.ncbi.nlm.nih.gov/orffinder/) and Vector NTI Advance™ 11.0 were used. The deduced amino acid sequences were examined using Vector NTI Advance software to calculate the theoretical isoelectric point (pI) and molecular weight (MW) of BcTSA protein. PSORT (http://psort.hgc.jp/) was used to predict the subcellular localization of the BcTSA protein. To get the amino acid sequences of TSAs different from plant species, a keyword search of the National Center for Biotechnology Information database-NCBI (https://www.ncbi.nlm.nih.gov) was undertaken. The deduced polypeptide sequence of BcTSA was aligned with those of SsTSA (TEY16427), SiTSA (XP_011082598), HiTSA (PIN15896), PbTSA (XP_009354864), MdTSA (XP_008382958), PmTSA (XP_008235332), PyTSA (PQQ06723), FvTSA (XP_004289899), CsTSA (XP_028076212), LgIGL (ACJ02772) and AsIGL (ACJ02768) by Clustal X v2.1. The phylogenetic tree was constructed using the Neighbor-Joining method with MEGA 5.0 software (The Biodesign Institute, Tempe, AZ, USA). The SOPMA tool (https://npsa-pbil.ibcp.fr/cgi-bin/npsaautomat.pl?page=npsasopma.html) was utilized to predict the secondary structure and functional domains of the BcTSA protein. The Simple Modular Architecture Research Tool (SMART, http://smart.embl-heidelberg.de/) was used to identify the conserved sequence domains of BcTSA. The three-dimensional structure of BcTSA was modeled using Swiss-Model (http://swissmodel.expasy.org/) and PyMOL (https://www.pymol.org/2/).

The real-time quantitative PCR (RT-qPCR) mixture contained 1.0 μL cDNA template, 10 μL 2 × master mix, 1.0 μL forward primer (10 μM), 1.0 μL reverse primer (10 μM), and 7 μL RNase-free ddH₂O and was performed on a Thermal Cycler Dice Real Time System TP800 (Takara, Tokyo, Japan) according to the instructions of SYBR PreMix Ex Taq (Takara Bio, Dalian, China). The expression levels were normalized with the 18S control gene using the 2^-ΔΔCt method (Table S1) (Huang et al., 2017). All RT-qPCR experiments were performed in triplicate.

To create BcTSA-GFP, the ORF of BcTSA was fused to the pCAMBIA1301-GFP vector using Bgl II and Spe I restriction sites. The primers used for subcloning are listed in Table S1. DNA sequencing validated the sequence and fusion of GFP under the control of the cauliflower mosaic virus (CaMV) 35S promoter. Using the PEG-mediated transfection technique, the BcTSA-GFP fusion protein was transiently expressed in freshly produced rice protoplasts. Transformed protoplasts were incubated overnight at room temperature (23-25°C), and transient expressions were seen using a confocal laser scanning microscope (Nikon, Tokyo, Japan). The GFP fluorescence was recorded at excitation wavelengths of 488 nm and emission wavelengths of 505-530 nm. A wavelength longer than 650 nm was used to capture the red autofluorescence caused by chlorophylls as previously described (Tan et al., 2015).

The complementation assays were performed essentially as described with some modifications (Zhang et al., 2008). The E. coli ΔtrpA mutant strain CGSC 9129, which contains an interruption/deletion of the trpA gene, was purchased from Yale University’s E. coli Genetic Stock Center. The ORF of BcTSA was cloned in pBAD-TOPO (Invitrogen, Carlsbad, CA, USA) expression vector for functional expression in E. coli using specific primers specified in Table S1 according to the manufacturer’s instructions. The untransformed E. coli mutant strain was cultured in LB medium with kanamycin at a concentration of 50 mg/mL. All strains were cultivated on M9 basic medium, supplemented with 1M MgSO4, 1M CaCl2, 20% glucose, glycerol, 100 mg/mL of ampicillin, and 20% arabinoses for inducible expression of BcTSA, then cultured at 37°C in the dark for 2 d. The untransformed strains or transformed strains containing the pBAD-TOPO vector serve as a negative control. The medium supplemented with 100 mg/mL tryptophan was used as a positive control.

The whole coding sequence of BcTSA was amplified from leaf cDNA using primers PHB-TSA-F and PHB-TSA-R (Table S1). The fragment was subsequently inserted into the Bam HI and Spe I restriction sites of the modified binary vector PHB-flag, which was driven by two 35S promoters. Agrobacterium tumefaciens C58C1 transmitted the final construct to I. indigotica leaf explants. The hairy root cultures were created by transforming A. rhizogenes C58C1 as previously reported (Ma et al., 2017). After 45 days of culturing, the hairy roots were harvested for DNA and RNA extraction, as well as metabolite analyses. In this research, at least three separate lines were evaluated for gene expression and metabolism.

The CTAB technique was used to extract genomic DNA from transgenic hairy roots after 45 days of cultivation. To determine the presence of inserted BcTSA fragments in BcTSA overexpressed lines, the primers PHB-TSA-F and rbcsr based on the sequence of vector PHB-flag were utilized. In A. tumefaciens strain C58C1, rolB and rolC represent DNA fragments of T-DNA in Ri, and hpt represents the hygromycin resistance gene of PHB-flag, indicating that PHB-flag: BcTSA has been effectively transformed into hairy roots. The expression levels of BcTSA genes in all positive lines were examined using RT-qPCR. All primers are listed in Supplementary Table S1. At least three separate control lines were evaluated in these studies, with the mean value serving as the control.

Hairy roots of I. tinctoria were dried at 50°C and pestled into fine powder. 0.2 g of samples were extracted with 5 mL of methanol: chloroform (1:1 vol). The suspension was supersonic for 30 min and then centrifuged at 10 000 rpm for 15 min to remove the suspended particles. The supernatant was transferred and the powder was extracted with 5 mL of reagents again (Yu et al., 2019). The final supernatant was filtered through a 0.22 μm organic membrane. Five millilitres of the extracting solution was evaporated to dryness and redissolved in 200 μL methanol.

LC-MS/MS analysis was run on an Agilent 1200 infinity LC column coupled with an Agilent 6410 Triple Quadrupole LC/MS System (Agilent, USA). The LC operating conditions were as follows: LC column, Poroshell 120 EC-C18, 3.0 μm×150 mm, 2.7 μm; mobile phase, 5 mM ammonium acetate and 0.2% solution of formic acid (A phase) and B phase consisted of a mixture of acetonitrile and methanol (50:50 v/v) with 5 mM ammonium acetate; total flow rate of mobile phase, 0.3 mL/min; total run time including equilibration, 26 min. The injection volume was 5 μL. The MS/MS was operated with an electrospray ionization source in positive ion mode with multiple reaction monitoring. The nebulizer gas pressure was set at 40 psi with a source temperature of 350°C and gas flow at 10 L/min. The capillary voltage was 4 000 V (positive mode). High-purity nitrogen gas was used as collision cell gas. The raw chromatograph and mass spectrogram data were processed with the MassHunter Workstation software (Agilent, USA).

Experiments were performed in triplicate, and are presented as mean ± standard error of the mean (SEM) calculated by GraphPad Prism 8.0 software. A Student’s t- test was used to compare two groups. One-way ANOVA followed by the Dunnett post hoc test was used for multiple comparisons versus the control group. p < 0.05 and p < 0.01 were set as the criterion for statistical significance.

RT-PCR was used to isolate the full-length genomic and cDNA sequences of BcTSA from B. cusia. The full-length genomic BcTSA gene was 4381 bp in size and consisted of 8 introns and 9 exons (Figure 2A). The cDNA of BcTSA is composed of 948 bp and encodes a putative alpha-tryptophan protein of 315 amino acids. The predicted protein had a calculated MW of 33.545 kDa with a pI of 5.61. According to amino acid analysis, the BcTSA putative protein featured a conserved Trp_syntA domain at the 57th-314th amino acids, which is similar to TSAs found in other plants. The secondary structure estimation revealed that the BcTSA peptide consisted of 43.49% alpha helices, 17.14% extended strands, 6.67% beta twists, and 32.7% random coils. The most frequent structural features were alpha helices and random coils, which penetrated much of the BcTSA secondary structure, whereas extended strands and beta turns were scattered throughout the protein. The 3D structure of BcTSA was simulated by the SWISS-MODEL server using the crystal structure of Tryptophan synthase alpha chain homolog BX1 (PDB id: 1rd5.1) as a template with the consistency of predicted results at 59.46%, implying that it contains the Tryptophan synthase alpha-subunit functional domain.

To discover more regarding the phylogenetic relationship between BcTSA and its close homologs, a rooted phylogenetic tree was built using the amino acid sequences of BcTSA from B.cusia and 11 other closet homologs (including IGL from Lamium galeobdolon and Aphelandra squarrosa, as well as contain Trp_syntA domain) from 11 angiosperms (Figure 2B). The phylogenetic tree showed BcTSA was structurally closely related to AsIGL in Aphelandra squarrosa (Figure 2B), the two plant species of which belonged to the Acanthaceae family, suggesting a close phylogenetic relationship. BcTSA also shared 68-78.15% amino acid similarity with other IGL or TSA proteins from other species, according to multiple sequence alignment (Figure 2C).

To characterize the transcription pattern of BcTSA in B. cusia, the transcription level of BcTSA was analyzed by RT-qPCR. As shown in Figure 3A, the BcTSA transcript was detected in all analyzed tissues of B. cusia, with the highest transcript levels found in the stems, followed by the roots, and the lowest levels found in the leaves. The RT-qPCR showed the transcription level of BcTSA rose dramatically under the stress of ABA, SA, and MeJA with significant variations depending on the time and/or the phytohormones (Figures 3B–D). During the ABA treatment, the transcript level of BcTSA decreased gradually for the first two hours, then rapidly rose to a peak at 4 h (5.66-fold), and then downregulated (Figure 3B). After treatment with MeJA, BcTSA expression decreased modestly for 2 hours before rapidly increasing. The highest level of BcTSA expression was observed after 12 h of 0.1 mM MeJA treatment, with an expression of approximately 18.29 times higher than control levels (Figure 3C). The transcript of BcTSA rapidly increased 1.77-fold after SA treatment for 1 h and then declined after 2 h of treatment, but again increased up to 2-fold at 4 h of treatment. Along with the extension of treatment time, the transcript level of BcTSA reached its highest point at 16 h, with an increase of ~8.75-fold, and subsequently dropped (Figure 3D).

Given the possibility that the indole alkaloids route exists as a metabolic channel connected with chloroplasts, it was important to identify BcTSA’s subcellular location. As shown in Figure 4A-E, when the GFP fusion proteins of BcTSA were constitutively expressed in rice protoplast, their fluorescent signals merged nicely with the chlorophyll autofluorescence from chloroplasts, demonstrating that the BcTSA protein could be localized to the chloroplasts. In contrast, free GFP was discovered uniformly throughout the cytoplasm (Figure 4F–J).

To determine whether BcTSA has the enzyme activity that synthesizes the production of indole from IGP, the recombinant expression plasmid pBAD-TOPO for BcTSA was used to transform into the E. coli (9129) mutant strain ΔtrpA, which is a Trp-auxotrophic mutant with a deleted TSA gene. In contrast to the control transformant TOPO, which only contains the pBAD-TOPO vector, E. coli (TSA) colonies transformed with the BcTSA cDNA were able to survive on M9 basic minimal medium containing 0.02% arabinose without tryptophan. The presence of BcTSA cDNA fragments in plasmids isolated from transformant colonies growing without tryptophan were determined by PCR using BcTSA specific primers (Table S1). The observation suggests that BcTSA can restore the E. coli ΔtrpA mutant (Figure 5). As a consequence, we conclude that BcTSA catalyzes indole synthesis in B. cusia.

The PHB-BcTSA construct was created by successfully subcloning the full-length BcTSA cDNA without a stop codon into the plasmid PHB-flag utilizing the BamHI and SpeI restriction sites (Figure 6A). The transcript levels of endogenous BcTSA were examined by RT-qPCR in positive lines (named T). The expression level of BcTSA was massively increased in the positive transgenic lines, with 16.89-, 19.95-, and 14.43-fold enhancements in T-2, T-15, and T-20 compared to WT plants, according to RT-qPCR data (Figure 6B). Four separate lines (2, 4, 15, 20) with higher BcTSA transcript levels were chosen from all of the independent hygromycin-resistant transgenic I. indigotica hairy roots harboring the PHB-BcTSA construct. LC-MS/MS analysis revealed that the concentration of indole alkaloids in four different BcTSA-overexpressed I. indigotica hairy root lines was considerably enhanced compared to the WT group (Figure 6C and Figure S1). The T-4 line with the maximum isatin concentration (105.58 μg/g DW), followed by the T-20 and T-2 lines, which were 5.1-, 3.4-, and 2.9-fold higher than that in its WT (20.79 μg/g DW) counterparts, respectively. Line T-20 showed the greatest level of indigo and indirubin, at 55.73 μg/g DW and 568.47 μg/g DW, respectively, which increased 10.82-fold and 22.45-fold when compared to the WT. This observation demonstrated BcTSA’s beneficial properties in the metabolic engineering of indole alkaloids synthesis.