Engineering tropane biosynthetic pathway in Hyoscyamus niger hairy root cultures

Abstract

Scopolamine is a pharmaceutically important tropane alkaloid extensively used as an anticholinergic agent. Here, we report the simultaneous introduction and overexpression of genes encoding the rate-limiting upstream enzyme putrescine N-methyltransferase (PMT) and the downstream enzyme hyoscyamine 6 β-hydroxylase (H6H) of scopolamine biosynthesis in transgenic henbane (Hyoscyamus niger) hairy root cultures. Transgenic hairy root lines expressing both pmt and h6h produced significantly higher (P < 0.05) levels of scopolamine compared with the wild-type and transgenic lines harboring a single gene (pmt or h6h). The best line (T₃) produced 411 mg/liter scopolamine, which was over nine times more than that in the wild type (43 mg/liter) and more than twice the amount in the highest scopolamine-producing h6h single-gene transgenic line H₁₁ (184 mg/liter). To our knowledge, this is the highest scopolamine content achieved through genetic engineering of a plant. We conclude that transgenic plants harboring both pmt and h6h possessed an increased flux in the tropane alkaloid biosynthetic pathway that enhanced scopolamine yield, which was more efficient than plants harboring only one of the two genes. It seems that the pulling force of the downstream enzyme (the faucet enzyme) H6H plays a more important role in stimulating scopolamine accumulation in H. niger whereas the functioning of the upstream enzyme PMT is increased proportionally. This study provides an effective approach for large-scale commercial production of scopolamine by using hairy root culture systems as bioreactors.

Secondary metabolites are low-molecular-weight compounds produced widely throughout the plant kingdom. Plant alkaloids constitute the largest groups of natural products, providing many pharmacologically active compounds. The in-depth understanding of biosynthetic pathways, along with the increasing number of cloned genes involved in biosynthesis, enable the exploration of metabolic engineering as a potential effective approach to increase the yield of specific metabolites by enhancing rate-limiting steps or by blocking competitive pathways. A few genera of the plant family Solanaceae, including Hyoscyamus, Duboisia, Atropa, and Scopolia, are able to produce biologically active nicotine and tropane alkaloids simultaneously (1-3). Both tropane and pyridine alkaloid biosynthetic pathways share a common polyamine metabolism in their early steps. Putrescine is a common precursor of both polyamines, such as spermidine and spermine, and tropane/pyridine alkaloids (4-6). Putrescine N-methyltransferase (PMT; EC 2.1.1.53) is the enzyme involved in the removal of putrescine from the polyamine pool because it catalyses the N-methylation of this diamine to form N-methylputrescine (mP). Because both the tropane ring moiety of the tropane alkaloids and the pyrrolidine ring of nicotine are derived from putrescine by way of mP synthesis, the N-methylation of putrescine catalyzed by PMT is the first committed step in the biosynthesis of these alkaloids (7). Scopolamine, which is the 6,7-β-epoxide of hyoscyamine, is formed from hyoscyamine by means of 6β-hydroxyhyoscyamine. Hyoscyamine 6β-hydroxylase (H6H, EC 1.14.11.11), a 2-oxo-glutarate-dependent dioxygenase, catalyzes the hydroxylation of hyoscyamine to 6β-hydroxyhyoscyamine, as well as the epoxidation of 6β-hydroxyhyoscyamine to scopolamine (1, 8, 9) (Fig. 1). The tropane alkaloids hyoscyamine (its racemic form being atropine) and scopolamine are structurally related and are derived from a common intermediate, the N-methylpyrrolinium cation. They are used medicinally as anticholinergic agents that act on the parasympathetic nerve system. Because they differ in their actions on the central nervous system, there is currently a 10-fold higher commercial demand for scopolamine, in the N-butylbromide form, than for hyoscyamine and atropine combined (9). Hyoscyamine and scopolamine are mostly synthesized in young root cells and translocated to the aerial parts of the plant (10). Hence, cultured roots are capable of accumulating high concentrations of these metabolites. Small-scale jar fermenters for the hairy roots of several Solanaceous species have been developed as prospective in vitro systems for commercial production of tropane alkaloids, but the scopolamine levels in these systems are often much lower than those of hyoscyamine (11). Researchers have also carried out genetic engineering of pharmaceutically important tropane alkaloids (12), in which the conversion of hyoscyamine to the much more valuable scopolamine is the major goal. Releasing the expression of key enzymatic activities from the strict regulation to which they are normally subjected is expected to increase product formation. A rough correlation has been found between H6H activity and the ratio of scopolamine to hyoscyamine in scopolamine-producing cultured roots (9). H6H, therefore, is a promising target enzyme that, if overexpressed in hyoscyamine-accumulating tissues, would result in increased scopolamine levels in the transformants. Several hyoscyamine-rich but scopolamine-poor plants that had been considered unattractive for commercial exploitation may now become promising candidates for large-scale scopolamine production by means of cultured roots. The hydroxylase gene from Hyoscyamus niger has been introduced into Atropa belladonna, a typical hyoscyamine-rich tropane alkaloid-producing plant species (13). Several transgenic root clones showed 5-fold higher concentrations of scopolamine than the wild-type hairy roots. By overexpressing h6h in Hyoscyamus muticus hairy root cultures, the best transgenic clone had a 100-fold increase of scopolamine. Furthermore, this clone produced hyoscyamine as the major alkaloid in similar amounts compared with controls (14). Recent efforts have been aimed at increasing the flux through the biosynthetic pathways. Overexpression of PMT increased the nicotine content in Nicotiana sylvestris, implying increased flux through tropane and pyridine pathways whereas suppression of endogenous PMT activity severely decreased the nicotine content and induced abnormal morphologies (15). When overexpressing the pmt gene in A. belladonna (15) and Duboisia hybrid (16), there was no significant increase in either tropane- or pyridine-type alkaloid concentrations.

Fig. 1.

Biosynthetic pathway of nicotine and tropane alkaloids (1). ArgDC, arginine decarboxylase; OrnDC, ornithine decarboxylase; PMT, putrescine N-methyltransferase; DAO, diamine oxidase; TR, tropinone reductase; H6H, hyoscyamine 6β-hydroxylase. [Adapted with permission from ref. 1 (Copyright 1994, Annual Reviews, Inc.).]

Although a substantial increase in productivity is feasible when a rate-limiting enzyme is targeted (17), in most biosynthetic pathways for secondary metabolites there often exists more than one rate-limiting step. Overexpression of one enzyme often renders subsequent reactions more rate-limiting, and thus the effect of single-enzyme overexpression may be dampened. Strategies should include fortification of multiple steps by overexpressing multiple biosynthetic genes, manipulating regulatory genes that control the expression of multiple pathway enzyme genes, or both. A good example is the bioengineering of the β-carotene biosynthetic pathway in the major staple food crop rice (Oryza sativa). By overexpressing the combination of genes encoding for three key enzymes (phytoene synthase, phytoene desaturase, and lycopene β-cyclase), a final β-carotene (previtamin A) concentration of 2 mg/kg was detected in dry rice endosperm (18).

So far, explants of H. niger have not been used for scopolamine bioengineering. Commercial cultivars of H. niger contain the natural amounts of scopolamine, which are not very high. Hence, there is a very strong need to increase scopolamine production rates for commercial production. Here, we report the introduction of gene constructs containing cDNA clones of pmt and h6h, driven by the constitutive cauliflower mosaic virus 35S promoter, into H. niger, either one gene at a time, or both genes together. The morphology, growth rate, activities of alkaloid pathway enzymes, and alkaloid production capacities of these engineered hairy root lines were investigated.

Materials and Methods

Construction of pmt and h6h Binary Expression Vector. The monovalent pmt expression plasmid pBMI was previously constructed (16). Disarmed Agrobacterium tumefaciens strain C58C1 harboring both pBMI and Agrobacterium rhizogenes Ri plasmid pRiA4, containing a single pmt gene, were used for plant transformation (16, 19).

The monovalent h6h expression plasmid pLAL21, constructed by Jouhikainen et al. (14), was used in the study. Plasmid LAL21 was isolated from Escherichia coli strain DH5α and transformed into A. rhizogenes LBA9402 by electroporation (20). A positive clone, after confirmation by PCR and enzymatic digestion analysis for the presence of the h6h gene, was used for plant transformation.

The pBMI and pLAL21 were used to construct a bivalent expression plasmid containing both pmt and h6h genes. The pBMI was linearized by HindIII digestion and blunted with Klenow fragment of E. coli DNA polymerase I. After double digestion of pLAL21 with EcoRI and EcoRV, the 1,250-bp fragment, containing the h6h coding sequence and 87-bp 3′-end sequence of the cauliflower mosaic virus (CaMV) 35S promoter, was recovered. It was then ligated with an EcoRI/EcoRV fragment from the double digestion of pGFP to form the intermediate plasmid GFP-h6h. After digesting pGFP-h6h with SphI, the 2,100-bp fragment, containing the complete h6h expression cassette, was recovered, blunted with E. coli T4 DNA polymerase, and finally ligated with the blunted, linearized pBMI to form the recombinant expression vector pXI (Fig. 2A). The pXI contained two separate expression cassettes for pmt and h6h, both driven by the CaMV 35S promoter and the nptII cassette for conferring kanamycin resistance.

Fig. 2.

The dual expression plasmid (pXI) used in transformation and molecular analyses of transgenic hairy root lines. (A) Schematic representation of the pXI. (B) Representative PCR analyses for the presence of rolB and rolC genes in transgenic hairy root lines. M, DL-2000 Marker (100-2,000 bp); P, pRiA4 (positive control); N, the wild-type H. niger root (negative control); T, transgenic hairy root lines containing both pmt and h6h genes induced by A. tumefaciens C58C1 strain (pRiA4, pXI). (C) Representative PCR analyses for the presence of pmt and h6h genes in transgenic hairy root lines. M, DL-2000 Marker (100-2,000bp); PC, pXI (positive control); NC, pRiA4 (negative control). (D) Northern blot analyses for the expression of pmt and h6h in transgenic hairy root lines. A4, transformed hairy root lines generated through blank transformation with A. rhizogenes strain A4; H, single h6h transgenic hairy root lines; P, single pmt transgenic hairy root lines; WT, wild type; T1 to T10, different transgenic lines.

The three plasmids, pBMI, pLAL21 and pXI, were separately joined to an expression vector within the left and right T-DNA [portion of the Ti (tumor-inducing) plasmid that is transferred to plant cells] borders. The pXI was isolated from E. coli strain DH5α and transformed into disarmed A. tumefaciens strain C58C1 containing A. rhizogenes Ri plasmid pRiA4 (20). A positive clone, after confirmation by PCR and enzymatic digestion analysis for the presence of both pmt and h6h genes, was used to transform plant tissues for simultaneous expression of pmt and h6h.

Plant Transformation and Root Cultivation. H. niger seeds were obtained from the Second Military Medical University (Shanghai, China) and germinated into plants. Transformation of leaf explants from these H. niger plants was carried out basically following the previously described method for tobacco leaf disk transformation (21), with the simultaneous transformation with strains A4 and LBA9402 of A. rhizogenes as controls. Wild-type plants were grown in the same growth chamber. Roots developed at cut edges 2-3 weeks after cocultivation were excised and cultured on solid, hormone-free, half-strength B5 medium (22), supplemented with 30 g/liter sucrose as the carbon source. All of the culture media contained 100 μg/ml kanamycin and 500 μg/ml cefotaxime. Root culture clones were maintained at 26°C in the dark and routinely subcultured every 25-30 days. The rapidly growing kanamycin-resistant lines with no bacterial contamination were used to establish hairy root lines. About 100 mg of fresh roots (3 cm in length) were inoculated into 150-ml conical flasks containing 40 ml of liquid and half-strength B5 medium and cultured on an orbital shaker (100 rpm) at 25°C in the dark. After 28 days of culture, the roots were filtered and washed with 10 ml of sterile distilled water and lyophilized (23). Root tissues from three flasks of cultures were collected individually at days 3, 7, 14, 21, 28, and 35 after inoculation. They were then treated as described above.

PCR Analysis. Genomic DNA was isolated from hairy root samples by using the acetyl trimethyl ammonium bromide (CTAB) method (24). The DNA was then used in PCR analysis for detecting the presence of Agrobacterium rol (B, C), pmt, and h6h genes in transgenic hairy root tissues, following the previously described method (25, 26). PCR primers for pmt detection were FPMT (5′-GCCATTCCCATGAACGGCC-3′) and RPMT (5′-CCTCCGCCGATGATCAAAACC-3′) whereas the primers FH6H (5′-CCGGA AT TCGGATCCCA ACGTATAGATTCTTC-3′) and RH6H (5′-CGGGAATTCGGATCCCAAACCATCACTGCAAT-3′) were used for h6h detection. PCR was carried out in total volumes of 50 μl reaction mixtures, containing 1 μl of each primer (10 μmol/liter), 1 μl of 10 mmol/liter dNTPs, 5 μl of 10× PCR buffer (Mg²⁺ plus) and 2.5 units of Taq^r DNA polymerase (TaKaRa) with 200 ng of genomic DNA as template. For detection of the pmt gene, the template was denatured at 94°C for 5 min followed by 35 cycles of amplification (1 min at 94°C, 1 min at 60°C, 45 s at 72 °C) and by 5 min at 72°C. For the detection of the h6h gene, the template was denatured at 94°C for 5 min followed by 35 cycles of amplification (1 min at 94°C, 1 min at 54°C, 1.5 min at 72°C) and by 5 min at 72°C. The mixture of the above four primers was used for simultaneous detection of pmt and h6h from hairy root samples derived from transformation (T lines) by using the same PCR conditions for the detection of the h6h gene.

Northern Blot Analysis. Total RNA was isolated from separately generated 4-week-old root lines of culture by using TRIzol Reagent (GIBCO/BRL) and subjected to Northern blot analysis for the expression of pmt and h6h. Aliquots of total RNA (10 μg per sample) were denatured and separated on a 1.1% formaldehyde-denatured (wt/vol) agarose gel (27). After electrophoresis, the RNA was transferred onto a positively charged Hy-bond-N⁺ nylon membrane (Amersham Pharmacia) through capillary transfer (27). The probe was generated by PCR (PCR DIG Probe Synthesis Kit, Roche) with pXI as template using primers FPMT/RPMT for pmt and FH6H/RH6H for h6h. PCR labeling of the probes with digoxigenin (DIG)-dUTP and hybridization (30-min prehybridization at 50°C followed by 16-h hybridization at 50°C) were performed according to the manufacturer's instructions (Roche). Hybridizing bands were detected by using the DIG Luminescent Detection Kit (Roche), and signals were visualized by exposure to Fuji x-ray film at 37°C for 10 min.

Enzyme Activity Assay. PMT enzymatic activity was evaluated by using the method of Feth et al. (28) with some modifications. Tissues (0.5-1.0 g of FW) were extracted on ice with 3 vol of 100 mmol/liter potassium phosphate buffer (pH 7.5, buffer A), containing 5 mmol/liter EDTA, 10 mmol/liter mercaptoethanol, 0.5% sodium ascorbate, and 2% polyethyleneglycol 400, followed by centrifugation at 27,000 × g for 30 min. The supernatant was loaded onto a Sephadex G25 prepacked PD-10 column (Amersham Pharmacia Biotech) equilibrated and eluted with 50 mmol/liter potassium phosphate buffer (pH 8, buffer B) containing 1 mmol/liter EDTA and 5 mmol/liter mercaptoethanol. The reactions were performed by incubating 100 μl of the purified supernatant with 20 μl of 25 mmol/liter putrescine (final concentration: 3.6 mmol/liter), 8 μl of 10 mmol/liter S-adenosylmethionine (final concentration: 0.6 mmol/liter) and 12 μlofbufferBat37°C for 30 min. After stopping the reactions by heating in boiling water, 65 mmol/liter borate-KOH buffer and a solution of dansylchloride (5.4 mg/ml acetonitrile) were added to the incubation mixture. After heating at 60°C for 15 min, the dansylated amines were extracted by adding 0.5 ml of toluene followed by vortex mixing for 30 s. After an aliquot (400 μl) of the toluene was removed, the residue was dried and resuspended in a fixed volume of acetonitrile, which was then injected into the HPLC. The chromatographic conditions for the separation of dansylated N-methylputrescine were as described before (29). The retention time of N-methylputrescine was 24 min.

H6H enzymatic activity was assayed by using GLC by measuring the formation of 6β-hydroxyhyoscyamine or scopolamine (8). Subsequent extraction of the reaction products and the conditions for GLC analysis were as described by Hashimoto and Yamada (9).

Alkaloid Extraction and Analysis. Extraction of tropane alkaloids (hyoscyamine and scopolamine) was based essentially on the method described by Hashimoto et al. (13) and analyzed by HPLC as described (30). Pyridine alkaloid (nicotine) was extracted from hairy root cultures with 25 mmol/liter sodium phosphate buffer (pH 7.8) at 30°C for 24 h with constant agitation and analyzed by HPLC by using a mobile phase of 40% methanol containing 0.2% phosphoric acid buffered to pH 7.25 with triethylamine. Standards for hyoscyamine, scopolamine, and nicotine (Sigma) were prepared in methanol at a final concentration of 1 mg/ml.

Results and Discussion

Transformation of H. niger with Plasmids Containing pmt and h6h. Three plasmids, pBMI, pLAL21, and pXI, containing the expression cassettes of the cDNAs encoding pmt and h6h, separately or together (Fig. 2 A), were separately introduced into H. niger leaf explants by using disarmed A. tumefaciens C58C1 strain and A. rhizogenes LBA9402. The abbreviations P, H, and T refer to the transgenic hairy root lines generated from pmt single gene-, h6h single gene-, and pmt/h6h double gene-transformations, respectively. Hairy root lines generated from transformations with strains A4 and LBA9402 containing no pmt or h6h are denoted as A4 and LBA, respectively. In total, 41 P, 50 H, and 25 T hairy root lines were generated, of which 28 P, 7 H, and 16 T lines survived the successive subculture process (Table 1). Some transgenic hairy root lines turned brown and aged considerably faster than wild-type hairy root lines. These lines were discarded, and the remaining hairy root lines were subcultured for 4-5 weeks in hormone-free, half-strength B5 liquid medium (Fig. 3A).

Table 1. Gene constructs and derived root cultures.

		Number of established root lines
T-DNA construct	Strain	Total	Antibiotic-resistant	PCR-positive	Established root cultures
<nptll<pmt<h6h	C58C1	25	16	11	T1, T2, T3, T6, T8, T10
<nptll<pmt>	C58C1	41	28	28	P4, P5, P18
<nptll<h6h>	LBA9402	50	7	3	H4, H11, H20

The antibiotic resistance gene (nptll) is always placed near the left border of the T-DNA; < and > indicate the direction of transcription. In addition, two lines transformed with C58C1 (pRiA4) were also established. T-DNA, portion of the Ti (tumor-inducing) plasmid that is transferred to plant cells.

Fig. 3.

Analyses for the morphology, growth rate, enzyme activities, and alkaloid contents in transgenic H. niger hairy root lines. (A) Phenotype of developed root lines. I, pmt and h6h dual plasmid generated transgenic hairy root lines (T) on solid 1/2 B5 medium; II, transformed hairy root lines generated through blank transformation with A. rhizogenes strain A4 on solid 1/2 B5 medium; III, the covalent pmt and h6h dual plasmid generated transgenic hairy root lines (T) in liquid 1/2 B5 medium; IV, wild-type H. niger root culture in liquid 1/2 B5 medium. (B) Time courses of growth of five randomly selected hairy root lines. Each value represents the means of three to five determinations ±SD; g, grams of tissues. (C) PMT activity in the T₃ line (n = 32, ± SD) and WT line (n = 16, ± SD). (D) H6H activity in the T₃ line (n = 57, ±SD) and WT line (n = 22, ±SD). (E) Alkaloid contents in root lines. A4, transformed hairy root lines generated through blank transformation with A. rhizogenes strain A4; H, h6h transgenic hairy root lines; P, pmt transgenic hairy root lines; T, transgenic hairy root lines containing both pmt and h6h genes.

Significant differences of growth characteristics were detected among independent transformed root lines. All of the P and T lines grew fast and vigorously with thick and fewer branches whereas H lines grew slowly with slender, dense and white branches. Significant differences (P < 0.01) were found between T/P lines and H lines with respect to growth rate, but no significant difference (P > 0.05) was detected among H lines and hairy roots generated by blank transformation (A4, LBA9402) (Fig. 3B). Generally, a fairly high content of secondary metabolites in the tissues is associated with poor growth, and the actual total productivity of secondary metabolites therefore remains low (14). We found in this study, however, that root morphology had a considerable influence on secondary metabolite production whereas the growth rate in the high scopolamine-producing lines was not reduced as compared with those with low scopolamine production. Line T₃, for example, which produced the highest level of scopolamine, grew very rapidly. Aging was observed along with the generation of brown pigments in the roots, most likely due to the accumulation of phenolic compounds. All types of hairy root lines reached the highest growth rate at the third week and achieved maximum fresh weight at the fifth week.

Molecular Analysis of Genetically Engineered Hairy Roots. All of the hairy roots contained the rol genes (rol B, rol C), revealed by PCR analysis (Fig. 2B). The integration of pmt and h6h in transformed hairy roots was also confirmed by PCR (Fig. 2C). In total, the PCR-positive, kanamycin-screened hairy root lines amounted to 82.35% (42/51), with 100% (28/28) for P, 42.86% (3/7) for H, and 68.75% (11/16) for T lines, respectively. The percent of PCR-positive hairy roots in this study is much higher than that previously reported in Nicotiana tabacum (13%) (31). Because both FH6H and RH6H primers were designed to cover the h6h sequence and the vector sequence, none of the checked DNA band was amplified from the control wild-type root samples.

Total RNA was isolated from separately generated root lines at the fourth week of cultivation for Northern blot analysis of pmt and h6h expression. Our results showed that the pmt and h6h transcripts accumulated at quite variable levels among independent transgenic hairy root lines (Fig. 2D). When compared with the wild-type plants, the h6h transcript in all of the transgenic H and T lines (including H₄,H₁₁,T₁,T₂,T₃, and T₁₀), except for line H₂₀, expressed at higher or comparable levels. However, the transcript level of h6h in line H₂₀ was lower than that of the wild type, implying the occurrence of cosuppression, which subsequently resulted in the low scopolamine accumulation in this line (see Fig. 2D). The cosuppression phenomenon of pmt was previously observed in transgenic N. sylvestris plant (15). Because PMT is a common key enzyme catalyzing polyamine metabolism existing in most Solanaceae species (1, 7), all of the transgenic lines showed pmt transcripts at various levels. The H lines (H₄, H₁₁, and H₂₀) expressed lower levels of pmt whereas T lines showed higher levels compared with the wild type.

Analysis for Enzyme Activity and Alkaloid Concentration. The PMT and H6H enzymatic activities in the transgenic hairy roots were determined by HPLC (28) and GLC (8), respectively. A good correlation was observed between mRNA levels and enzymatic activities in transgenic hairy root lines. PMT activity (50.7 pkat/mg ± 4.3, n = 32, ±SD) and H6H activity (10.2 pkat/mg ± 0.45, n = 57, ±SD) in line T₃ were significantly higher (P < 0.01) than those in wild-type lines (2.53 pkat/mg ± 0.25, n = 16, ±SD for PMT and 4.03 pkat/mg ± 0.85, n = 22, ±SD for H6H) (Fig. 3 C and D). All of the transgenic lines achieved maximum PMT activity at the end of culture period (35 days) but achieved peak H6H activity already after 3 wk of culture.

The profiles of alkaloids (nicotine, hyoscyamine, and scopolamine) of transgenic hairy root lines were determined by HPLC. The capacities of transgenic root lines to biosynthesize hyoscyamine and scopolamine are shown in Fig. 3E. Except for line H₄, which showed a low level of nicotine content (<26 mg/liter), no pyridine alkaloid nicotine was detected in any of the hairy root lines, implying that overexpression of pmt did not promote the accumulation of nicotine in H. niger. All lines expressing h6h transcripts produced tropane alkaloids in which scopolamine was the staple compound. However, scopolamine levels in H lines were highly variable, ranging from 25.1 to 184.4 mg/liter. The expression levels of the h6h transgene also varied from line to line but showed positive correlation to their corresponding scopolamine concentrations. Elevated h6h activity was widely detected in the highly productive H and T lines. The inability of line H₂₀, the P lines, A4, and wild-type control lines to accumulate scopolamine may partly result from their low h6h activities. These findings suggest that high H6H activity has a positive influence on the flow of metabolites through the pathway. This positive correlation between H6H activity and scopolamine production has also been found in tissues of A. belladoma (13), H. muticus (14), and N. tabacum (31). These data demonstrate the desirability of overexpressing h6h for elevated scopolamine production.

In the present study, although having higher pmt transcript levels than the control lines (Fig. 2D), the pmt single-gene transgenic hairy root lines (P lines) produced alkaloids (hyoscyamine, scopolamine, and nicotine) at similar levels to those of the control lines (Fig. 3E). This result indicates that enhancement of pmt transcript alone was not sufficient to boost scopolamine biosynthesis, suggesting that this pathway is mainly downstream-limited, rather than limited at the step where putrescine is converted to methylputrescine in H. niger.

Our observation is consistent with the results of pmt single-gene overexpression in A. belladonna (13), N. sylvestris (15) and Duboisia (16), but contrary to those in Hyoscyamus albus (7) and Datura hybrid (32). It is conjectured that the response of Solanaceous species to pmt overexpression is species-related, which may be partly due to a different and specific posttranslational regulation of the endogenous enzyme in respect to the exogenous one (16).

Two of the T lines (T₃ and T₁₀) presented the highest scopolamine production (411.2 and 224.8 mg/liter, respectively). Are the high scopolamine levels associated only with h6h single-gene overexpression, or with a coordinative effect of both pmt and h6h transgenes? Northern blot analysis showed that line H₁₁ had a relatively higher h6h transcript level than T₃ and T₁₀, but its scopolamine content was not the highest (184.4 mg/liter). Whereas H₄ and T₃ showed similar h6h transcript levels, the scopolamine concentration in T₃ was significantly higher than that in H₄. The best line (T₃) produced 411.2 mg/liter scopolamine, more than nine times that in the wild type (43.7 mg/liter) and over two times that in the highest scopolamine-producing h6h single-gene transgenic line H₁₁ (184.4 mg/liter). To our knowledge, this is the highest scopolamine content achieved in planta developed through genetic engineering to date.

It is concluded that, as in the case of transgenics that include both pmt and h6h in H. niger, the pulling force from the metabolically downstream enzyme (the faucet enzyme) H6H plays a more important role in stimulating scopolamine accumulation whereas the proportional functioning of the upstream enzyme PMT is well coordinated. The expression of pmt and h6h in H. niger is similarly regulated at the transcriptional level. The first specific precursor of the tropane alkaloid pathway is N-methylputrescine, whose formation from putrescine is catalyzed by PMT, which seems to be flux-limiting and hence represents the first committed step in tropane alkaloid biosynthesis (1, 7). Because h6h encodes an enzyme catalyzing the final two steps from hyoscyamine to scopolamine that is not inhibited by end product (8), it is suggested that it is completely desirable to fundamentally enhance the scopolamine production by simultaneously promoting the hydroxylation activity of H6H and directing the flux from the polyamine pool to the tropane alkaloid pathway. This result is in accordance with our anticipation that fortification of multiple steps by overexpressing multiple biosynthetic genes is crucial for improved production of useful plant secondary metabolites.

This report on engineering pmt and h6h simultaneously into scopolamine-producing plant species results in significant enhancement of scopolamine accumulation in cultured hairy root lines. Overexpression of multiple biosynthetic genes or transcription factors that control the expression of genes in the target bioengineering pathway is a promising strategy to alter the accumulation of certain secondary metabolic products. When the steps of the biosynthetic pathway are elucidated and the respective genes have been cloned, exact regulation toward enhanced productivity of medicinal natural products will be possible. Metabolic engineering, either alone or in combination with traditional breeding, provides a practical means to stimulate valuable secondary metabolites production.

The current study provides an effective approach for commercially large-scale production of scopolamine by using the hairy root systems as bioreactors. This work also sheds light on how to effectively increase the end products of secondary metabolic pathways by appropriate genetic engineering strategies.