Mature seeds of Hyoscyamus niger were harvested from the medicinal plant garden of the Xizang Agricultural and Husbandry College of Nyingchi (Tibet, China) in August 2016, with their taxonomic identity confirmed by Professor Xiaozhong Lan. These seeds were germinated into plantlets in substrate composed of vermiculite:pindstrap moss:perlite (6:3:1) and grown at 25 ± 1°C under an 16 h-light/8 h-dark conditions. Once the plantlets reached 10 cm in height, their roots and leaves were respectively harvested for the tissue profile analysis of HnODC, HnADC1, HnADC2, and TA-biosynthesis genes, including HnPMT, HnTRI, and HnH6H. To determine whether MeJA influenced the expression levels of these genes and the metabolism of polyamines and TAs, the 10-cm-tall H. niger plants were treated with 100 μM of MeJA for 0, 1, 6, 12, and 24 h. Each duration had three replicate plants, from which the roots were harvested for RNA isolation and metabolite analysis. The leaves collected from same plants were used for metabolite detection. Plant material treated with a solution lacking MeJA for 24 h served as the control. Three or more independent plants per treatment were used in all analyses.

Total RNA was extracted from the H. niger roots with RNAsimple Total RNA Kit, according to the manufacturer’s protocols (Tiangen Biotech, Beijing, China). 50–100 mg of material from each plant part was used to extract total RNA. The first-strand cDNA chain was synthesized by using a FastKing RT kit (Tiangen Biotech, Beijing, China). The prepared reaction mixture, with a total volume of 10 μl, contained 2 μl of buffer (DNase solution provided by the FastKing RT kit), 2 μg of total RNA, and ddH₂O was incubated at 42°C for 3 min to remove any potential genomic DNA. Next, 2 μl of King RT buffer, 1 μl of FastKing RT Enzyme Mix, and 2 μl of FQ-RT Primer Mix were added into the reaction mixture; ddH₂O was then also added to obtain the final volume of 20 μl. Then, this 20 μl of the RT reaction mixture was incubated at 42°C for 15 min and at 95°C for 3 min. All the cDNA samples were diluted 50 times with RNase-free water, after which 8 μl of cDNA solution served as templates for the RT-PCR.

A pair of gene-specific primers, HnODC-F and HnODC-R (Supplementary Table S1), was used to isolate the coding sequence of HnODC based on sequenced H. niger transcriptomes (data not published). Amplification reactions were performed in a final volume of 50-μl buffer containing 5 μl of TransTaq HiFi Buffer I (10×) with 20 mM MgSO₄ (TransGen Biotech, Beijing, China), 4 μl of 2.5-mM dNTPs, 1 μl of each primer (10 mM), 2.5 U of TransTaq HiFi DNA polymerase, and 50 ng of template DNA. PCR conditions were set as follows: the templates were denatured at 94°C for 5 min, followed by 28 cycles (94°C for 30 s, 56°C for 30 s, and 72°C for 90 s), and finally incubated at 72°C for 8 min. PCR products were purified from 1.0% (w/v) agarose gel with a DNA purification kit (BioFlux, Hangzhou, China), then subcloned into pMD19-T for sequencing by using the following protocol: a reaction mixture that contained 1 μl of PMD19-T vector, 1 μl of DNA fragment, 3 μl of ddH₂O, and 5 μl of Solution I was first prepared, then incubated at 16°C for 30 min and transformed into E. coli. The recombinant plasmid harboring HnODC was extracted from E. coli for sequencing. The sequence of HnODC was confirmed by sequencing it on a 3730 DNA Analyzer (Thermo Fisher Scientific, Waltham, MA, United States) using M13 forward and M13 reverse universal primers, via the Sanger sequencing approach. The BLAST analysis was performed online at the website¹ (Johnson et al., 2008).Then multiple alignments were performed using the ClustalX bioinformatics program (Larkin et al., 2007). A phylogenetic tree was built by the neighbor-joining method in MEGA software v.5 (Tamura et al., 2011). Its bootstrapped values were generated from n = 1000 replicates to evaluate the accuracy of the phylogenetic construction.

To analyze the tissue profile of TA-biosynthesis genes HnODC, HnADC1, HnADC2, HnPMT, HnTRI, and HnH6H, their total RNAs were respectively extracted from the leaves and roots of H. niger plants (three biological replicates) according to the methods described above. Likewise, to analyze the expression patterns of these genes under the MeJA treatment, their RNAs were respectively extracted from the roots of plants treated with MeJA and control plants. After reverse-transcription into cDNAs, expression levels of the genes were analyzed by real-time quantitative PCR (qPCR), using the phosphoglycerate kinase gene (PGK) as an internal reference, by following the method of Li et al. (2014). The qPCR kits were purchased from BIO-RAD and the qPCR system was an IQ5 thermocycler (BIO-RAD, Hercules, CA, United States). The 2^-ΔΔCT method was used to calculate the relative gene expression levels (Livak and Schmittgen, 2001). At least three independent plants were used in this gene expression analysis. All primers were designed with the software tool Beacon Designer (Premier Biosoft International, Palo Alto, CA, United States), based on sequences publicly available from the NCBI GenBank database. Primer specificity was validated by melting profiles and consisted of a single product-specific melting temperature. The associated GenBank accession numbers are HnODC (MK169378), HnPMT (AB018572), HnTRI (D88156), and HnH6H (DQ812529). All the primers used are listed in Supplementary Table S1.

N-methylputrescine and polyamines (putrescine, spermidine, and spermine) were extracted using the method described by Do et al. (2013). Root and leaf samples (1.00 g fresh weight, FW) from 24-h-MeJA-treated plants and corresponding control plants were homogenized respectively in liquid nitrogen, and then extracted in 4 ml of 0.2 N perchloric acid (PCA) at 4°C for 1 h. After centrifugation at 16,000 g at 4°C for 30 min, the supernatant was used to determine the plant content of polyamines. To do this, 1 ml of the supernatant was added with 10 μl of benzoyl chloride. After incubation at 37°C for 25 min in the dark, the benzoylzed polyamines were extracted with 2 ml of chloroform, dried with nitrogen flow, and then dissolved in 1 ml of methanol. From each ensuing sample, 20 μl were injected into HPLC for metabolite analysis (Flores and Galston, 1982). The detection methods used here were similar to those applied in the enzymatic assays described below. Tropane alkaloid content was quantified by adhering to previously described methodologies (Qiang et al., 2016; Zhao et al., 2017; Geng et al., 2018). For this, 200 mg of dry powder from each plant part was accurately weighed for alkaloid extraction and detection, with at least three independent plants used for metabolite detections.

The HnODC coding region was amplified by using two primers (HnODC-PF/HnODC-PR) containing a restriction site for BamHI and SacI (Supplementary Table S1). First, the PCR products of HnODC were purified and digested using BamHI and SacI. Then the purified HnODC coding region harboring the restriction sites of BamHI and SacI was inserted into a pET-28a+ vector, to generate the prokaryotic expression vector. The ligation mixture consisted of 1 μl of T4 DNA ligase, 1 μl of reaction buffer, 2 μl of linearized pET-28a+ vector, and 6 μl of purified HnODC. Next, the constructs were introduced to E. coli Rosetta for protein expression. Bacteria were cultured in an LB liquid medium with 50 mg/L of kanamycin and 34 mg/L of chloramphenicol at 37°C. When the OD value of bacterial cultures had reached 0.6, HnODC expression was induced by adding isopropyl β-D-1-thiogalactopyranoside (IPTG) at a final concentration of 0.5 mM. Bacteria were further cultured at 25°C for 7 h, then harvested for protein purifications. The recombinant His-tagged HnODC was purified via Ni²⁺-chelating resin columns using the same methods we reported in a previous study (Qiang et al., 2016).

To perform the enzymatic assays, we followed Docimo et al. (2012). To test the catalytic activity of HnODC, its purified form was tested in a 10-mM Hepes buffer (pH 8.0) containing 1 mM dithiothreitol and 1 mM pyridoxal phosphate (PLP) for 60 min, by using different concentrations of L-ornithine as the substrate. The products (putrescine) from enzymatic assays and authentic putrescine were benzoylated into benzoyl putrescine (as carried out by Geng et al., 2018). The benzoylated samples were used for the HPLC analysis or stored at -80°C for future use. All benzoylated samples were detected at 234 nm at an oven temperature of 30°C on a Shimadzu LC-20 HPLC system (Shimadzu Corp., Kyoto, Japan). A YMC-Pack ODS-A column was used (150 × 4.6 mmI. D.S-5 μm, 12 nm) and the flowing phage consisted of methanol:water (41:59) at a flow rate of 1 ml/min throughout the analysis. 20°μl of sample was injected for analysis. The Michaelis–Menten curve and Lineweaver-Burk plot of the HnODC enzyme were drawn to determine its K_m and V_max values, on which calculations of turnover rate (K_cat) and catalytic efficiency (K_cat/K_m) were based.

The 1293-bp coding sequence of HnODC encoded a 430-amino-acid polypeptide (Supplementary Figure S1). The BLASTP analysis indicated HnODC belonged to the superfamily of type III PLP-dependent enzymes, and that it resembled the ODCs in GenBank. As the major binding site of α-DFMO (Coleman et al., 2003), the GPTCD motif was found at the C-regions in all of these ODC and LDC proteins (Figure 2). For HnODC, this motif was located at positions 372–376 (Figure 2), while the YAVKCN motif was at positions 90–95, and present in all the ODC and LDC sequences (Figure 2). Lysine in this motif was postulated to bind to the cofactor pyridoxal-5′-phosphate (Lee and Cho, 2001). But compared with the ODC enzymes of mammals, all plant ODC enzymes we found lacked the PEST regions required for constitutive and conditional degradation of ODC by the 26S proteasome system (Lee and Cho, 2001). Our phylogenetic analysis showed a distant evolutionary relationship between mammalian ODC enzymes and those from plants (Figure 3). All plant L/ODCs and ODCs occupied the same branch, which had two subgroups, and the L/ODCs are mainly from Leguminosae. Furthermore, HnODC showed closer evolutionary relationships with ODC proteins of Solanaceae plants and EcODC (Figure 4). HnODC has been deposited in GenBank (accession number MK169378).

Expression levels of HnODC, HnADC1, HnADC2 and three TA-biosynthesis genes (HnPMT, HnTRI, and HnH6H) were detected in the roots and leaves of H. niger (Figure 4). However, all three genes—HnPMT (Figure 4B), HnTRI (Figure 4C), and HnH6H (Figure 4D)—were specifically expressed in the roots of H. niger, consistent with earlier reported findings (Hashimoto and Yamada, 1987; Hashimoto et al., 1992; Geng et al., 2018). Unlike the root-specific expression of those three TA biosynthesis genes, HnODC was expressed in both roots and leaves, but its expression level was still much higher in roots than leaves (Figure 4A). HnADC1 was expressed in roots and leaves at similar level, and HnADC2 was expressed in roots and leaves with no significant difference (Supplementary Figure S2). However, these six genes responded differently to the MeJA treatment of H. niger plants (Figure 4): the transcript levels of HnODC (Figure 4E) and HnPMT (Figure 4F) were significantly increased, while the expression of HnTRI (Figure 4G), HnH6H (Figure 4H) and HnADC1/2 (Supplementary Figure S2) was unchanged. Across 1–24 h of the MeJA treatment, the expression of HnODC increased up to 6-fold that of the control, while that of HnPMT increased by 6–30 folds. These results indicated HnODC is mainly expressed in roots and induced by MeJA.

Since the above gene expression analysis revealed HnODC was upregulated by MeJA at the transcriptional level, it was reasonable to investigate ornithine-derived metabolite production in H. niger plants treated with MeJA (Figure 5). Specifically, we examined polyamines, N-methylputrescine, hyoscyamine, and scopolamine content in their roots and leaves. Putrescine production was significantly promoted in roots (Figure 5A) and leaves (Figure 5B), where its concentrations were respectively increased by 34.98 and 90.32% over the corresponding control plants. By contrast, the MeJA treatment did not affect spermidine and spermine production in the roots (Figure 5A), but it did elevate their concentrations significantly in leaves (Figure 5B). Roots of MeJA-treated plants had a 6.14-fold higher N-methylputrescine content (30.14 ± 6.52 nmol/g FW) relative to control roots (4.91 ± 2.54 nmol/g FW) (Figure 5A). In leaves of either MeJA-treated or control plants, N-methylputrescine was not detectable (Figure 5B). The production of hyoscyamine and scopolamine was also detected in H. niger roots and leaves, for which more scopolamine than hyoscyamine was found, consistent with former’s greater abundance than the latter in H. niger (Han Woo et al., 1995). However, hyoscyamine and scopolamine concentrations were not significantly changed in roots and leaves (Figure 5A,B) by MeJA, suggesting it did not significantly affect their production.

To determine its enzymatic activity, HnODC was expressed in E. coli to produce its recombinant proteins (Figure 6A). The HnODC enzymes could be readily obtained in the supernatants of lysed E. coli. Then, the His-tagged HnODC was purified using Ni²⁺-chelating resin column through elution with 50 mM of imidazole (Figure 6B). The molecular weight of recombinant HnODC was approximately 50 kDa (Figure 6A,B), consistent with its calculated molecular weight and similar to the molecular weight of other reported plant ODC enzymes. When the substrate, ornithine, was fed to HnODC, the products were successfully detected by HPLC with a retention time of 14.8 min (Figure 6C), which agreed with that of the standard (Figure 6C). In our negative controls, no product was detected when HnODC was boiled (Figure 6C). Together, these results show HnODC did catalyze ornithine to produce putrescine.

Enzymes catalyzing the same reaction in different organisms usually differ from each other in their enzymatic kinetics, such as affinity to substrate and catalytic efficiency. To obtain kinetics information on HnODC behavior, we derived its K_m, V_max, and K_cat values for ornithine: respectively, 2.62 ± 0.11 mM, 1.87 ± 0.023 nmol min^-1 μg^-1 and 1.57 ± 0.015 at pH 8.0 and at 30°C based on the Michaelis–Menten curve (Figure 7A). Notably, the K_m value of HnODC was higher than that of either EcODC (0.395 mM) or NgODC (0.56 mM) (Table 1), suggesting that HnODC had lower affinity to ornithine than EcODC and NgODC (Docimo et al., 2012). Since the V_max value of HnODC also exceeded that of EcODC and NgODC, its resulting K_cat value likewise greater. The K_cat/K_m value, which expresses catalytic efficiency, was 599 M^-1 s^-1 for HnODC and thus higher corresponding values reported for EcODC (465 M^-1 s^-1) and NgODC (16.53 M^-1 s^-1) (Table 1). Lineweaver-Burk plot for evaluation of K_m and V_max for HnODC was shown in the Figure 7B. In sum, the kinetic analysis demonstrated HnODC had a lower affinity to ornithine but a much higher catalytic efficiency than displayed by EcODC and NgODC.

Generally, all ODCs contain the conserved PLP-binding motifs composed of PFYAVKCN, and the GPTCD sequences, both of which are necessary for ODC activity (Coleman et al., 2003). Both of the motifs present in HnODC strongly suggest that it is a functional enzyme which catalyzes the decarboxylation of ornithine. Although the ODC sequences are similar to those of LDC, the phylogenetic analyses are able to distinguish between them. Since ODC and LDC each had its own clade in the phylogenetic tree, it has been suggested that they may have had common ancestors and evolved to different types of enzymes with slight modification (Bunsupa et al., 2012). Furthermore, the relatively high sequence similarity between the HnODC and other plant ODC proteins also suggest that HnODC should have similar functions to the other plant ODC enzymes.

Our biochemical assays confirmed that HnODC catalyzed the decarboxylation of ornithine to produce putrescine. HnODC had a lower affinity to ornithine than do the ODC enzymes of tobacco and coca tree, but its catalytic efficiency was found to be much greater that of ODCs of tobacco and coca tree. Particularly, HnODC showed about 36-fold increase in catalytic efficiency over NgODC. The low catalytic efficiency of tobacco ODC led it to be a limiting enzyme in nicotine biosynthesis, and consequently overexpression of yeast ODC enhanced the production of nicotine in transgenic tobacco (Hamill et al., 1990). In TA-producing plant species, the production of TAs was greatly reduced when ODC was inhibited by DFMO, suggesting that ODC might play a crucial role in TA biosynthesis (Nyman, 1994). High catalytic efficiency of HnODC facilitated the production of putrescine that entered the biosynthetic pathway of TAs. Due to higher catalytic efficiency of HnODC, it might be a better candidate for engineering the biosynthesis of putrescine-derived metabolites than the reported plant ODC enzymes.

Biosynthesis genes involved in the same pathway usually have similar tissue expression patterns. We found that in Hyoscyamus niger plants the TA biosynthesis genes of HnPMT, HnTRI, and HnH6H were expressed almost exclusively in secondary roots. Unlike them, HnODC was expressed in both roots and leaves yet its expression level was much higher in roots than in leaves. The different expression levels of HnODC likely reflect the changing metabolic demands for putrescine by root and leaf organs. Both polyamines and TAs were synthesized in roots and this required more production of its key precursor (putrescine); therefore, a high expression of HnODC in roots matched this requirement. The phytohormone, MeJA, could up-regulate the TA biosynthesis genes in a species-dependent way. HnPMT expression was dramatically elevated by MeJA in hairy root cultures of H. niger (Geng et al., 2018), while it was not affected by MeJA in Atropa belladonna (Li et al., 2014). HnODC was found strongly induced by MeJA in roots and leaves of the H. niger plants; however, neither HnTRI nor HnH6H was changed by MeJA at the transcriptional level. Hence, we conclude that MeJA positively regulated the TA-biosynthesis genes, including HnODC and HnPMT, though it did not regulate the TA-biosynthesis genes, such as TRI and H6H. Considering that ADC enzymes contribute to the putrescine biosynthesis, their expression was also detected. Unlike HnODC and TA biosynthesis genes with high or specific expression in roots, HnADC1/HnADC2 was expressed in roots and leaves, with no significant difference. The two ADC genes were not responsive to MeJA treatment. Gene expression analysis suggested that the increased production of putrescine was mainly caused by the up-regulation of HnODC.

Due to the MeJA-elevated expression of HnODC, putrescine—the product given by HnODC—production was significantly increased in roots and leaves, consistent with previously reported results that overexpression of ODC enhanced the production of putrescine in transgenic tobacco and rice (Lepri et al., 2001; Kumria and Rajam, 2002). In particular, we found N-methylputrescine production markedly promoted in roots when H. niger plants were treated with MeJA. This increased production could have been caused by both MeJA-induced expression of HnPMT and a greater supply of putrescine provided by MeJA-induced expression of HnODC. In the leaves with or without MeJA treatment, N-methylputrescine production was undetectable, obviously due to the lack of HnPMT expression in leaves; this results also indicates that N-methylputrescine synthesized in roots was hardly translocated to aboveground to leaf parts. Yet MeJA clearly promoted the production of spermidine and spermine in leaves, whereas their respective production in roots was not altered. Biosynthesis genes involved in TA pathway, such as HnPMT, HnTRI, and HnH6H, were not expressed in leaf, suggesting that putrescine was not metabolized into TA biosynthesis in leaf. However, putrescine was able to go into biosynthesis of spermidine and spermine in leaf. The increased putrescine production induced by MeJA resulted in more putrescine available for biosynthesis of spermidine and spermine in leaf. In roots, putrescine entered the biosynthesis of N-methylputrescine, spermidine, and spermine. MeJA-elevated HnPMT expression led to enhanced conversion of putrescine into N-methylputrescine in roots, and consequently the production of spermidine and spermine was at relatively stable levels. For hyoscyamine and scopolamine, the MeJA treatment did not affect their levels in roots and leaves, suggesting that the elevated expression of HnODC and HnPMT were not enough to promote the production of the two pharmaceutical TAs. Previously, overexpression tobacco PMT gene markedly promoted the N-methylputrescine production but did not enhanced the TA accumulation in root cultures of H. niger (Zhang et al., 2004). To conclude, the up-regulation of HnODC at transcriptional level was able to provide the precursors (putrescine and N-methylputrescine) at higher levels for TA biosynthesis and thereby promote polyamine production in H. niger.