Three New O-Methyltransferases Are Sufficient for All O-Methylation Reactions of Ipecac Alkaloid Biosynthesis in Root Culture of Psychotria ipecacuanha

The medicinal plant Psychotria ipecacuanha produces ipecac alkaloids, a series of monoterpenoid-isoquinoline alkaloids such as emetine and cephaeline, whose biosynthesis derives from condensation of dopamine and secologanin. Here, we identified three cDNAs, IpeOMT1–IpeOMT3, encoding ipecac alkaloid O-methyltransferases (OMTs) from P. ipecacuanha. They were coordinately transcribed with the recently identified ipecac alkaloid β-glucosidase Ipeglu1. Their amino acid sequences were closely related to each other and rather to the flavonoid OMTs than to the OMTs involved in benzylisoquinoline alkaloid biosynthesis. Characterization of the recombinant IpeOMT enzymes with integration of the enzymatic properties of the IpeGlu1 revealed that emetine biosynthesis branches off from N-deacetylisoipecoside through its 6-O-methylation by IpeOMT1, with a minor contribution by IpeOMT2, followed by deglucosylation by IpeGlu1. The 7-hydroxy group of the isoquinoline skeleton of the aglycon is methylated by IpeOMT3 prior to the formation of protoemetine that is condensed with a second dopamine molecule, followed by sequential O-methylations by IpeOMT2 and IpeOMT1 to form cephaeline and emetine, respectively. In addition to this central pathway of ipecac alkaloid biosynthesis, formation of all methyl derivatives of ipecac alkaloids in P. ipecacuanha could be explained by the enzymatic activities of IpeOMT1–IpeOMT3, indicating that they are sufficient for all O-methylation reactions of ipecac alkaloid biosynthesis.

The medicinal plant Psychotria ipecacuanha produces ipecac alkaloids, a series of monoterpenoid-isoquinoline alkaloids such as emetine and cephaeline, whose biosynthesis derives from condensation of dopamine and secologanin. Here, we identified three cDNAs, IpeOMT1-IpeOMT3, encoding ipecac alkaloid O-methyltransferases (OMTs) from P. ipecacuanha. They were coordinately transcribed with the recently identified ipecac alkaloid ␤-glucosidase Ipeglu1. Their amino acid sequences were closely related to each other and rather to the flavonoid OMTs than to the OMTs involved in benzylisoquinoline alkaloid biosynthesis. Characterization of the recombinant IpeOMT enzymes with integration of the enzymatic properties of the IpeGlu1 revealed that emetine biosynthesis branches off from N-deacetylisoipecoside through its 6-O-methylation by IpeOMT1, with a minor contribution by IpeOMT2, followed by deglucosylation by IpeGlu1. The 7-hydroxy group of the isoquinoline skeleton of the aglycon is methylated by IpeOMT3 prior to the formation of protoemetine that is condensed with a second dopamine molecule, followed by sequential O-methylations by IpeOMT2 and IpeOMT1 to form cephaeline and emetine, respectively. In addition to this central pathway of ipecac alkaloid biosynthesis, formation of all methyl derivatives of ipecac alkaloids in P. ipecacuanha could be explained by the enzymatic activities of IpeOMT1-IpeOMT3, indicating that they are sufficient for all O-methylation reactions of ipecac alkaloid biosynthesis.
Plants produce a diverse array of secondary metabolites, whose number has been estimated to be over 200,000. In plants, a large portion of those natural products are considered to be synthesized for chemical defense against microbial attack and herbivore predation because of their biological activities. Because some of them exhibit a wide range of pharmacological activities, they have also been studied from the aspect of phar-maceutical sources. Alkaloids are one of the better studied classes of plant natural products due to their pharmacological activities. Psychotria ipecacuanha Stokes (Rubiaceae), which is native to South and Central America, has been used in therapy since the early 17th century as an emetic and expectorant and later as a medication for amebic dysentery (1). Such medicinal effects of the root extracts derive from the principal alkaloid emetine. P. ipecacuanha also produces many kinds of emetinerelated alkaloids such as cephaeline and ipecoside, known as ipecac alkaloids ( Fig. 1) (2). Alangium lamarckii Thwaites (Alangiaceae), a medicinal plant indigenous to India (2), also produces ipecac alkaloids.
Structurally, ipecac alkaloids possess a monoterpenoid-isoquinoline skeleton that originates from secologanin, a glucosidal monoterpenoid, and dopamine. Secologanin and dopamine are condensed in a Pictet-Spengler manner to form the (1␣S)-diastereomer N-deacetylisoipecoside and the (1␤R)-diastereomer N-deacetylipecoside (Fig. 1). The reaction proceeds nonenzymatically under weakly acidic conditions, but it was demonstrated that two enzymes, each of which stereospecifically catalyzes the formation of each diastereomer, are involved in the reaction in Alangium (3,4). The (1S)-diastereomer N-deacetylisoipecoside is deglucosylated and converted to protoemetine, which is then condensed with a second molecule of dopamine and further converted to cephaeline and emetine (Fig. 1). The (1R)-diastereomer N-deacetylipecoside is converted to alkaloidal glucosides, such as ipecoside and alangiside, through N-acetylation and spontaneous lactamization ( Fig. 1) (5,6). They are accumulated as main alkaloidal glucosides in Psychotria and Alangium.
Ipecac alkaloids possess one isoquinoline moiety in the carbon skeleton prior to condensation with the second molecule of dopamine to form cephaeline and emetine, which contain two isoquinoline groups (Fig. 1). At completion of emetine biosyn- thesis, four hydroxy groups are methylated. Two hydroxy groups of the first isoquinoline skeleton are methylated during protoemetine formation. IpeGlu1 is involved in the deglucosylation step, but it is unknown whether O-methylations occur before or after deglucosylation (Fig. 1). In addition, O-methylation reactions of the second isoquinoline skeleton to form the final product emetine have not been determined. Moreover, a large part of the chemical diversity of ipecac alkaloids is attributable to the distinct O-methylation patterns with respect to the four hydroxy moieties. Thus, OMTs are essential targets to be identified for the elucidation of the order of transformations occurring along the ipecac alkaloid biosynthetic pathway. In this study, we report the identification of three cDNAs encoding ipecac alkaloid OMTs in P. ipecacuanha through EST analysis coupled with the characterization of the substrate specificity and kinetic parameters of recombinant enzymes. The substrate specificities of the three recombinant OMTs are sufficient to explain all O-methylation reactions of ipecac alkaloid biosynthesis. Based on their functional characterization, we propose the detailed biosynthetic network of ipecac alkaloids in P. ipecacuanha.

EXPERIMENTAL PROCEDURES
Plant Materials-P. ipecacuanha root cultures were maintained as described by Ref. 15.
Cloning of O-Methyltransferase cDNAs-1050 ESTs from a -ZAP cDNA library constructed from mRNA of P. ipecacuanha in vitro roots were recently sequenced (15). Homology searches of the ESTs using the BLASTX algorithm revealed that one of the contigs (contig-213) was similar to known plant OMTs. The contig consisted of four EST members (ph1903-35, ph1903-62, ph1803-II-61, and ph3011-41), of which ph1903-62 and ph1803-II-61 were identical in the overlapped region, but they were different from ph1903-35 and ph3011-41. Oligonucleotide primers were designed from common sequences among them for the isolation of full-length coding sequences by RT-PCR, 5Ј-AACTTGGCAAATGGAAACTG-3Ј (forward) and 5Ј-CTTGAATTCAAGGAGAAAGCTC-3Ј (reverse). Root mRNA was purified as described in Ref. 15, and the cDNA synthesized using a SMART rapid amplification of cDNA ends cDNA amplification kit (Clontech) was used for the PCR as follows: 30 s at 98°C, followed by 35 cycles of amplification (10 s at 98°C, 30 s at 55°C, and 1 min at 72°C) in a 50-l reaction mixture containing 3 ng of cDNA, 0.5 M primers, 0.2 mM dNTPs, 1ϫ reaction buffer, and 1 unit of Phusion HS-HF polymerase (Finnzymes). The PCR products were inserted into pCR-BluntII-TOPO vector (Invitrogen) and sequenced to obtain IpeOMT1 and IpeOMT2. IpeOMT1 corresponded to the EST ph3011-41, and IpeOMT2 to the ph1903-62 and ph1803-II-61. Because the full-length cDNA clone corresponding to the EST ph1903-35 was not obtained, 3Ј-rapid amplification of cDNA ends was performed using ph1903-35-specific forward primer (5Ј-GTCTAGATTAATAGCATCATAAGTTAGG-3Ј) and the adapter primer under the same conditions as described above except for the annealing temperature at 60°C. The PCR products were inserted into pCR-BluntII-TOPO and sequenced to obtain IpeOMT3, which corresponded to the EST ph1903-35.
Heterologous Expression and Purification of IpeOMT Enzymes-The entire coding region of each of the three IpeOMTs was amplified by PCR and ligated between the NdeI and EcoRI restriction sites of a pET28a vector (Novagen). The resulting vector was introduced into the Escherichia coli strain BL21 CodonPlus(DE3) RIL (Stratagene) for protein expression. The recombinant E. coli was grown in 500 ml of LB medium containing 30 g/ml kanamycin and 35 g/ml chloramphenicol at 37°C until the A 600 reached 0.6 -0.8. After cooling, isopropyl 1-thio-␤-D-galactopyranoside was added at 1 mM, and the culture was incubated for 18 h at 18°C on an orbital shaker at 200 rpm. The cells were harvested (6000 ϫ g, 10 min) and resuspended in 20 ml of 50 mM Tris-HCl buffer (pH 7.5) containing 20% (v/v) glycerol, 10 mM 2-mercaptoethanol, and 0.5% (v/v) protease inhibitor mixture (Sigma). After sonication (20 s at 50 watts, six times) and centrifugation (20,000 ϫ g, 20 min), the supernatant containing soluble protein was collected for the purification of the N-terminal His-tagged IpeOMT proteins.
The recombinant protein was purified with TALON His tag purification resin (Clontech). All operations were done at 4°C. The resin (1 ml of bed volume per 20 ml of the supernatant) equilibrated in 50 mM Tris-HCl buffer (pH 7.5) containing 0.5 M KCl, 20% glycerol, and 10 mM 2-mercaptoethanol was mixed with the supernatant for 1 h. After the resin was washed with equilibration buffer, the recombinant IpeOMT protein was eluted in a stepwise manner by increasing imidazole concentrations (5 and 200 mM) in equilibration buffer (5 ml for each elution). The 200 mM imidazole fraction was subjected to a PD10 column (GE Healthcare) equilibrated with storage buffer (50 mM Tris-HCl (pH 7.5) containing 0.1 M KCl, 20% glycerol, and 5 mM dithiothreitol) and was concentrated appropriately by ultrafiltration using a Centriprep YM-10 (Millipore). Purified protein was stored at Ϫ80°C. The protein profile was analyzed by 12% SDS-PAGE. Protein concentration was measured by the method of Ref. 25 using bovine serum albumin as a standard.
3-O-Methylredipecamine was prepared from N-deacetylisoipecoside via 6-O-methyl-N-deacetylisoipecoside. The reaction mixture (30 ml) containing 0.1 M citrate, 0.2 M phosphate buffer (pH 6.5), 1 mM N-deacetylisoipecoside, 10 mM AdoMet, and 4.6 mg of IpeOMT1 enzyme was incubated at 40°C. After a 7-h incubation, AdoMet (65 mg) and 0.9 mg of IpeOMT1 enzyme were added and further incubated for 2 h. The reaction was stopped by adding 6.2 ml of 1 N HCl, and the precipitation formed was removed by filtration (0.2 m, Whatman), and the pH was adjusted to 5.0 with 0.2 M Na 2 HPO 4 . To the solution (55 ml) was added 1.8 ml of 10 mM NaBH 3 CN dissolved in 0.1 M citrate, 0.2 M phosphate buffer (pH 5.0), and 360 g of IpeGlu9 enzyme. After a 1-h incubation at 55°C, the reaction was stopped by adding 12 ml of 1 N HCl. The pH was adjusted to 9.0 with 10% Na 2 CO 3 , and ethyl acetate extraction was performed (100 ml three times). The organic layer was washed with saturated NaCl solution and evaporated in vacuo to dryness. The residue was dissolved in 1 ml of acetonitrile containing 0.1% trifluoroacetic acid and subjected to preparative HPLC as described above, and the collected fraction was concentrated in vacuo and freeze-dried. 2-O-Methylredipecamine was prepared from 7-O-methyl-N-deacetylisoipecoside using the same procedure.
Enzyme Assay-In a series of screening of substrates for the IpeOMT1, IpeOMT2, and IpeOMT3 enzymes, standard enzyme reaction was performed in 0.1 M KP i buffer (pH 7.5) containing 10 g of enzyme, 1 mM AdoMet, and 1 mM substrate (varied between 0.05 and 1 mM with solubility and availability) in a total volume of 100 l. After incubation at 30°C for 2 h, the reaction was terminated by the addition of 20 l of 1 N HCl. After centrifugation (16,000 ϫ g, 2 min), the supernatant was subjected to HPLC analysis to detect the reaction products (column, Nova-Pak C18, 3.9 ϫ 300 mm, 4 m, Waters; eluent, A ϭ 0.1% trifluoroacetic acid, B ϭ acetonitrile, gradient ϭ 12-80% B/(A ϩ B) in 30 min; flow rate, 0.8 ml/min; detection, 230 nm). For the analysis of dopamine, 3-hydroxy-4-methoxyphenethylamine, 4-hydroxy-3-methoxyphenethylamine, guaiacol, and catechol, conditions for UV detection and solvent gradient, were changed appropriately.
For the determination of pH optimum, the same buffer systems were used as described in Ref. 15. Temperature optimum was determined by the reactions at 25-55°C. N-Deacetylisoipecoside was used as substrate for IpeOMT1 and IpeOMT2 and (S)-reticuline for IpeOMT3.
Enzyme reactions for the determination of kinetic parameters were performed in 0.1 M KP i buffer at the optimum pH for each IpeOMT at 30°C. Kinetic parameters were calculated from three replicates by nonlinear fitting of the data to the Michaelis-Menten equation using SigmaPlot 2001.
Identification of Enzymatic Products-The enzymatic products that are dependent on the presence of AdoMet were identified by comparing the retention time and the fragmentation pattern in liquid chromatography (LC)-tandem mass spectrometry (MS/MS) analysis with authentic compounds. For LC-MS/MS analysis, HPLC separation was conducted with a Shimadzu LC20A system, and mass spectra were obtained using a 4000 Q-TRAP triple stage quadruple mass spectrometer equipped with a TurboIonSpray ionization source (Applied Biosystems) in a positive ion mode. The HPLC separation was performed under the following conditions: column, Nova-Pak C18, 3.9 ϫ 300 mm, 4 m, Waters; eluent, solvent A ϭ 0.2% formic acid, solvent B ϭ acetonitrile; flow rate, 0.8 ml/min. Solvent gradient conditions basically followed that described above but were changed appropriately when the 6-O-methylate and 7-O-methylate of isoquinolines needed to be separated. For the analyses of IpeOMT1 and IpeOMT2 reaction mixtures with cephaeline and 7Ј-O-demethylcephaeline to detect emetine and cephaeline, respectively, the column was changed to Lichrospher 60 RP-select B (4 ϫ 250 mm, 5 m, Merck) for better peak shape of those substrates and products. To identify the molecular ion of the reaction products, MS detection was done by enhanced mass scan (collision energy, 10 V; declustering potential, 60 V), and then the fragmentation pattern of the molecular ion detected was analyzed by enhanced product ion scan, for which the declustering potential was kept at 80 V while the collision energy was adjusted appropriately.
To identify the enzymatic reaction products of N-deacetylisoipecoside by IpeOMT1 and IpeOMT2, the products were derivatized with di-tert-butyldicarbonate before LC-MS/MS analysis, because 6-O-methyl-N-deacetylisoipecoside and 7-Omethyl-N-deacetylisoipecoside could not be separated by HPLC under any conditions tested. For the analysis of IpeOMT1 reaction product, the enzyme reaction was performed in 0.1 M KP i (pH 7.5) containing 1.7 mM N-deacetylisoipecoside, 2 mM AdoMet, and 200 g of IpeOMT1 enzyme in a total volume of 1 ml. After a 4-h incubation at 30°C, the reaction was terminated by addition of 0.2 ml of 1 N HCl, followed by centrifugation (16,000 ϫ g, 2 min). The supernatant was subjected to preparative HPLC under the same conditions as described above except for the gradient condition (10 -21% B/(A ϩ B) in 60 min). The collected fraction was evaporated in vacuo and freeze-dried. The residue was treated with 1 mol of di-tert-butyldicarbonate in methanol containing 1 mol of triethylamine for 1 h at room temperature. The reaction mixture was dried under N 2 flow and dissolved in 500 l of 50% methanol. The chromatographic behavior of the N-Boc derivative was compared with those of standard 6-O-methyl-N-Bocdeacetylisoipecoside and 7-O-methyl-N-Boc-deacetylisoipecoside synthesized according to Ref. 15. To identify the enzymatic reaction products of N-deacetylisoipecoside by IpeOMT2, the enzyme reaction was performed in 0.1 M citrate, 0.2 M phosphate buffer (pH 6.5) containing 0.5 mM N-deacetylisoipecoside, 1 mM AdoMet, 1 mM EDTA, 1 mg/ml bovine serum albumin and 400 g of IpeOMT2. After the overnight incubation at 40°C, the reaction was terminated by 0.2 ml of 1 N HCl. Purification and N-Boc derivatization of the products were conducted in the same procedure as described above and subjected to LC-MS/MS analysis.
Subcellular Localization of IpeOMTs and IpeGlu1-The coding sequences of IpeOMT1, IpeOMT2, IpeOMT3, and Ipeglu1 (GenBank TM accession number AB455576) were amplified by PCR and inserted into pBA103 vector (gift from Brian Kelly, Donald Danforth Plant Science Center), which possesses an enhanced cauliflower mosaic virus 35S promoter, an enhanced green fluorescence protein gene, and a nopaline synthase terminator, to express C-terminal GFP-fused proteins, IpeOMT1-GFP, IpeOMT2-GFP, IpeOMT3-GFP, and Ipe-Glu1-GFP. DNA-coated gold particles (0.6 mm, Bio-Rad) were prepared according to the manufacturer's instructions and bombarded into a small piece (ϳ2 ϫ 2 cm) of onion with a PDS1000 He Biolistic Particle Delivery System (Bio-Rad) according to manufacturer's instructions. The bombarded onion was wrapped with a wet paper towel and incubated at room temperature overnight in a Petri dish. Cytosol-and vacuole-targeted enhanced yellow fluorescence protein constructs (gift from Dr. Isabel Ordiz, Donald Danforth Plant Science Center), endoplasmic reticulum-targeted enhanced GFP construct (gift from Dr. Nigel Taylor, Donald Danforth Plant Science Center), and pBA103 without insert were also bombarded as controls. After the incubation, the epidermis was peeled off and analyzed using an LSM 510 confocal/multiphoton microscope (Zeiss). Fluorescence of GFP and YFP was visualized by excitation at 488 nm and a bandpass emission filter (505-550 nm). Confocal images were analyzed using the Imaris software (Bitplane).
Quantitative RT-PCR Analysis-P. ipecacuanha in vitro roots precultured for 8 weeks were transferred to new LS liquid medium (26), and the roots were harvested in duplicate 1 and 3 days and 1, 2, 4, 6, and 8 weeks after transfer. Total RNA was purified using an RNeasy plant mini kit (Qiagen), followed by DNase treatment and re-purification. The synthesis of cDNA was performed with a SuperScript III reverse transcriptase (Invitrogen) (0.5 g of RNA in 20 l of reaction mixture). Duplicates were further pooled, and 5 ng of cDNA was used for PCR. PCR was performed in triplicate on a StepOnePlus real time PCR system (Applied Biosystems) under the following conditions: 5 min at 95°C, followed by 40 cycles of amplification (15 s at 95°C and 1 min at 60°C) in a 20-l reaction mixture containing template cDNA, 0.5 M primers, and 10 l of PerfeCTa SYBR Green FastMix Reaction Mixes (Quanta Biosciences). Primers specific to each of IpeOMT1, IpeOMT2, IpeOMT3, and Ipeglu1 were used (supplemental Table S1). A standard curve was made for every experiment with the dilution series of the known quantities of a plasmid having each of the target cDNAs as template. Specificity of the amplification was verified by a melt-curve analysis at the end of each PCR.

Isolation of O-Methyltransferase cDNAs from P. ipecacuanha-
RT-PCR from mRNA of P. ipecacuanha root culture resulted in the isolation of two full-length cDNAs IpeOMT1 (ipecac alkaloid O-methyltransferase 1, GenBank TM accession number AB527082) and IpeOMT2 (GenBank TM accession number AB527083). IpeOMT1 corresponded to the EST ph3011-41 and IpeOMT2 to the ph1903-62 and ph1803-II-61. The 3Ј-rapid amplification of cDNA ends PCR-specific to the EST ph1903-35 resulted in the isolation of IpeOMT3 (GenBank TM accession number AB527084).
Enzymatic Characterization of Recombinant IpeOMT Enzymes-N-terminal His-tagged IpeOMT enzymes were purified to apparent homogeneity from the soluble protein fraction by metal chelation chromatography (Fig. 3), yielding 6 mg (IpeOMT1), 16 mg (IpeOMT2), and 4 mg (IpeOMT3) of recombinant enzymes from 1 liter of E. coli culture. Enzymatic activities were first tested with seven glucosidal ipecac alkaloids ( Table 1, substrates 1-7, and supplemental Fig. S1), including N-deacetylisoipecoside, N-deacetylipecoside, and their 6-and 7-O-monomethylates. In the reaction of IpeOMT1 using N-deacetylisoipecoside as substrate, product dependent upon the presence of IpeOMT1 and the methyl donor AdoMet was detected by HPLC analysis. Because the possible products 6-Omethyl-N-deacetylisoipecoside and 7-O-methyl-N-deacetylisoipecoside could not be separated by reverse-phase HPLC under conditions tested (Fig. 4A), the product was purified and chemically converted to the N-Boc derivative. LC-MS/MS profiles of the derivative were the same as those of authentic 6-O-methyl-N-Boc-deacetylisoipecoside (Fig. 4B), showing that IpeOMT1 catalyzed 6-O-methylation of N-deacetylisoipecoside regiospecifically to form 6-O-methyl-N-deacetylisoipecoside. IpeOMT2 also reacted with N-deacetylisoipecoside, and AdoMet-dependent products were detected (Fig. 4A). Product identification was performed after N-Boc derivatization. Two peaks were detected, and they showed the same LC-MS/MS profiles as authentic 6-O-methyl-N-Boc-deacetylisoipecoside and 7-O-methyl-N-Boc-deacetylisoipecoside (Fig. 4B). Thus, it was found that IpeOMT2 catalyzed both a 6-and 7-Omonomethylation of N-deacetylisoipecoside, where 6-OMT activity was approximately three times higher than 7-OMT activity, as determined from the chromatogram of the N-Boc derivatives (Fig. 4B). In contrast to IpeOMT1 enzyme that showed no activity toward (1R)-diastereomer N-deacetylipecoside, IpeOMT2 reacted with N-deacetylipecoside as well as (1S)-diastereomer N-deacetylisoipecoside. By LC-MS/MS analysis, the products were identified to be 6-O-methyl-Ndeacetylipecoside and 7-O-methyl-N-deacetylipecoside (Fig.  4C). In the reaction, 7-O-methylation occurred approximately eight times more efficiently than 6-O-methylation. In addition to the substrate and the products, their lactams formed by nonenzymatic intramolecular reaction were also detected (Fig. 4C). The enzyme assay with O-monomethylates of N-deacetylisoipecoside and N-deacetylipecoside revealed that IpeOMT1 also catalyzed 6-O-methylation of 7-O-methyl-N-deacetylisoipecoside to form 6,7-O,O-dimethyl-N-deacetylisoipecoside (Fig. 4D). Enzymatic activities for lactam substrates (Table 1, substrates 8 -13, and supplemental Fig. S1) were also tested. IpeOMT1 did not show any activity for the lactam substrates. IpeOMT2 catalyzed O-methylation of the 2-hydroxy group of demethylalangiside, which corresponds to the 7-hydroxy group with respect to the isoquinoline skeleton, to form 3-O-demethyl-2-O-methylalangiside (Fig. 4E). Its activity was regio-and stereoselective, in contrast to the activities for non-lactam glucosidal ipecac alkaloids as described above. IpeOMT3 accepted none of the 13 glucosidal ipecac alkaloids ( Table 1, substrates 1-13) tested as substrate.
To examine the involvement of IpeOMT enzymes in 6Ј-and 7Ј-O-methylation of the second isoquinoline moiety in emetine biosynthesis (Fig. 1), their activities toward 7Ј-O-demethylcephaeline, the penultimate intermediate for emetine biosynthesis, and toward cephaeline ( Table 1,

Ipecac alkaloids 1 N-Deacetylisoipecoside
Although an endogenous substrate for IpeOMT3 was not known at this stage, because (S)-reticuline, one of the benzylisoquinoline alkaloids, was found to serve as its substrate as described below, pH and temperature optima were determined using N-deacetylisoipecoside as substrate for IpeOMT1 and IpeOMT2 and (S)-reticuline for IpeOMT3. The pH optimum of IpeOMT1 was 8.0, whereas the pH optima of IpeOMT2 and IpeOMT3 were 7.0. Temperature optima of IpeOMT1, IpeOMT2, and IpeOMT3 were 40, 45, and 40°C, respectively.
IpeOMT2 catalyzed 6-O-methylation of (R,S)-isococlaurine, (R,S)-norcoclaurine, and (R,S)-isoorientaline (Table 1), showing that IpeOMT2 catalyzes 6-O-methylation of 6,7-dihydroxyand 6-hydroxy-7-methoxyisoquinolines. This is different from the IpeOMT2 activities for ipecac alkaloids described above, i.e. 6-O-monomethylation and 7-O-monomethylation activities toward the 6,7-dihydroxyisoquinoline moiety. Considering that 6-and 7-O-methylation of N-deacetylipecoside was catalyzed only by IpeOMT2 and that 7-O-methyl-N-deacetylisoipecoside was synthesized only by IpeOMT2 (Figs. 7 and 8), this relaxed regiospecificity appears to be an essential feature in ipecac alkaloid biosynthesis. It is notable, however, that IpeOMT2 regiospecifically catalyzed the methylation of the 7Ј-hydroxy group, but not the 6Ј-hydroxy group, of 7Ј-O-demethylcephaeline to form cephaeline, which is considered to be the primary reaction of IpeOMT2 judging from the catalytic parameters (Table 2). Although 3-OMT activity toward oripavine was detected, because the 3-hydroxy group of oripavine is not bound to an isoquinoline skeleton, this activity seems to be fortuitous.
IpeOMT3  (Table 1). IpeOMT3 accepted only 7-hydroxy-6-methoxyisoquinolines. Although the methylated position of the enzymatic product of coreximine could not be assigned due to the lack of authentic 2-O-methylcoreximine and 11-O-methylcoreximine, and possibly due their identical fragmentation pattern in MS/MS analysis, methylation of the 2-hydroxy group that corresponds to 7-hydroxy group of isoquinoline skeleton is most likely, judging from the regiospecificity of IpeOMT3 enzyme for other benzylisoquinoline alkaloids.
Although IpeOMT3 was found to catalyze 7-O-methylation of 7-hydroxy-6-methoxybenzylisoquinoline alkaloids, they are not the endogenous substrates, because none of these benzylisoquinoline alkaloids have been found in P. ipecacuanha. However, the results encouraged us to examine the ipecac alkaloid aglycon as a possible endogenous substrate. As shown in Ref. 15, deglucosylation of N-deacetylisoipecoside gives multiple products formed by nonenzymatic intramolecular reactions of the highly reactive aglycon. Redipecamine (supplemental Fig.  S1) that is formed by the deglucosylation reaction under reducing conditions could, however, be stably prepared (15). Therefore, we used redipecamine, 2-O-methylredipecamine, and 3-O-methylredipecamine (Table 1, substrates 16 -18, and supplemental Fig. S1) as mimics of endogenous ipecac alkaloid aglycons. IpeOMT1 reacted with none of them. IpeOMT2 methylated the 2-hydroxy group of redipecamine (Fig. 4H) that corresponds to the 7-hydroxy group of the isoquinoline skeleton, whereas IpeOMT3 catalyzed 2-O-methylation of 3-Omethylredipecamine (Fig. 4H). These results strongly suggested that IpeOMT2 and IpeOMT3 accept endogenous ipecac alkaloid aglycons that are formed by deglucosylation of N-deacetylisoipecoside and 6-O-methyl-N-deacetylisoipecoside, respectively.
Because IpeOMT sequences exhibited the highest identities to the flavonoid OMT of C. roseus, and IpeOMT3 also exhibited comparably high identity to the resveratrol OMT of V. vinifera, IpeOMT activities toward kaempferol, quercetin, myricetin, and resveratrol (Table 1, substrates 49 -52, and supplemental Fig. S1) were examined, but none were accepted by IpeOMTs as substrate.
Kinetic parameters of the IpeOMT enzymes for ipecac alkaloids, synthetic mimics of the ipecac alkaloid aglycons, and the methyl donor AdoMet were determined (Table 2). Apparent K m and apparent k cat values of IpeOMT1 for N-deacetylisoipecoside were 107 M and 11.3 ϫ 10 Ϫ3 s Ϫ1 , respectively. The apparent K m value for 7-O-methyl-N-deacetylisoipecoside (105 M) was equal to that for N-deacetylisoipecoside, but the apparent k cat value for 7-O-methyl-N-deacetylisoipecoside (20.4 ϫ 10 Ϫ3 s Ϫ1 ) was 2-fold higher than that for N-deacetylisoipecoside. The catalytic efficiency of IpeOMT1 for cephaeline (3.5 s Ϫ1 ⅐M Ϫ1 ) was remarkably lower than those for N-deacetylisoipecoside (106 s Ϫ1 ⅐M Ϫ1 ) and 7-O-methyl-Ndeacetylisoipecoside (194 s Ϫ1 ⅐M Ϫ1 ) mainly due to the lower k cat value (0.28 ϫ 10 Ϫ3 s Ϫ1 ) despite the comparable K m value (80 M).
Kinetic parameters of IpeOMT2 for N-deacetylisoipecoside were evaluated based on the total amount of the two products, 6-O-methyl-N-deacetylisoipecoside and 7-O-methyl-N-deacetylisoipecoside, because they could not be separated by HPLC analysis. The apparent reaction efficiency (0.81 s Ϫ1 ⅐M Ϫ1 ) was ϳ130-fold lower than that of IpeOMT1 (106 s Ϫ1 ⅐M Ϫ1 ) due to larger K m value (420 M) and lower k cat value (0.34 ϫ 10 Ϫ3 s Ϫ1 ). Kinetic parameters of IpeOMT2 for N-deacetylipecoside were not determinable, because considerable amounts of sub-strate and the enzymatic products (6-O-methyl-N-deacetylipecoside and 7-O-methyl-N-deacetylipecoside) were nonenzymatically converted to their lactams under the reaction conditions used (pH 7.0) (see Fig. 4C). The apparent reaction efficiency for demethylalangiside (8.4 s Ϫ1 ⅐M Ϫ1 ) was 10-fold higher than that for N-deacetylisoipecoside due to the 10-fold increase of the k cat value (3.8 ϫ 10 Ϫ3 s Ϫ1 ) with a similar K m value (450 M). IpeOMT2 catalyzed the 2-O-methylation of the synthetic N-deacetylisoipecoside aglycon mimic, redipecamine, with greater reaction efficiency (66 s Ϫ1 ⅐M Ϫ1 ) than those for N-deacetylisoipecoside and demethylalangiside. The apparent reaction efficiency of IpeOMT2 for 7Ј-O-demethylcephaeline (32,600 s Ϫ1 ⅐M Ϫ1 ) was the highest among the substrates tested, which is attributable to notably small K m (1.0 M) and high k cat (32.6 ϫ 10 Ϫ3 s Ϫ1 ) values.
Subcellular Localization of IpeOMT and IpeGlu1 Enzymes-To examine the subcellular localization of IpeOMT1, IpeOMT2, and IpeOMT3 enzymes, C-terminal GFP-fused enzymes (IpeOMT1-GFP, IpeOMT2-GFP, and IpeOMT3-GFP) were transiently expressed in onion epidermal cells. In addition, subcellular localization of the ipecac alkaloid ␤-glucosidase (IpeGlu1) (15) was also examined by expressing the IpeGlu1-GFP fusion protein to see whether or not the IpeGlu1 enzyme is compartmentalized separately from the IpeOMT enzymes. Three-dimensional images were reconstructed from the confocal images, and the localization of GFP/YFP signals was analyzed. Fig. 5 shows the representative sections obtained from the transformed onion cells. Fluorescence profiles of onion cells expressing each of the IpeOMT-GFPs and IpeGlu1-GFP were the same as those expressing nontargeted GFP and cytosol-targeted YFP but were totally different from those expressing endoplasmic reticulum-targeted GFP and vacuoletargeted YFP. These results showed the cytosolic localization of IpeOMT and IpeGlu1 enzymes. This was supported by the absence of signal peptides in IpeOMT and IpeGlu1 sequences as predicted by the computer programs iPSORT, SignalP, and Target P.  Transcript Profiles of IpeOMT and IpeGlu1 Genes in P. ipecacuanha in Vitro Roots-Transcript profiles of the IpeOMT1, IpeOMT2, and IpeOMT3 genes, as well as the Ipeglu1 gene, were analyzed in P. ipecacuanha root culture to see whether the three IpeOMT genes are coordinately transcribed with Ipeglu1. Time course changes in the gene transcription in root cultures were measured by quantitative RT-PCR analysis (Fig. 6). Throughout the time course, transcript patterns of the three IpeOMTs were well correlated with that of Ipeglu1, where transcript levels started to increase after subculture, reached maxima in 1-2 weeks, and then decreased to lower levels. The transcript level of IpeOMT3 was lower than the other two IpeOMTs and Ipeglu1.

Involvement of IpeOMTs in the Biosynthesis of Ipecac Alkaloids Derived from (1S)-N-Deacetylisoipecoside-Identification
of the three ipecac alkaloid O-methyltransferases, IpeOMT1, IpeOMT2, and IpeOMT3, in this study, as well as the ipecac alkaloid ␤-glucosidase, IpeGlu1 (15), enabled us to elucidate a major portion of the biosynthetic pathways to ipecac alkaloids. Although the EST data base of Psychotria root culture that was constructed consisted of a relatively small number of ESTs (1050 members) (15), the results indicate that it is enriched in ipecac alkaloid biosynthetic genes.
The elucidated pathway and the contribution of each of those enzymes are shown in Figs. 7 and 8 for S-and R-forms of ipecac alkaloids, respectively. Both IpeOMT1 and IpeOMT2 enzymes catalyzed the 6-O-methylation of N-deacetylisoipecoside, an intermediate in emetine biosynthesis. Kinetic analysis showed that the reaction catalyzed by IpeOMT1 is 130 times as efficient as that catalyzed by IpeOMT2 due to smaller K m and higher k cat values, indicating that IpeOMT1 plays a major role in the formation of 6-O-methyl-N-deacetylisoipecoside. 7-O-Methylation of N-deacetylisoipecoside was catalyzed by IpeOMT2, although the reaction efficiency was much lower than the 6-Omethylation reaction catalyzed by IpeOMT1 and IpeOMT2. 6,7-O,O-Dimethyl-N-deacetylisoipecoside was found to be synthesized via 7-O-methyl-N-deacetylisoipecoside, because none of the IpeOMTs catalyzed 7-O-methylation of 6-O-methyl-N-deacetylisoipecoside but IpeOMT1 catalyzed the 6-Omethylation of 7-O-methyl-N-deacetylisoipecoside. These enzymatic properties seem to be the mechanism to effectively prepare 6-O-methyl-N-deacetylisoipecoside to be deglucosylated by IpeGlu1 for emetine biosynthesis. In fact, 6-O-methyl-N-deacetylisoipecoside is one of the best substrates for IpeGlu1 among the glucosidal ipecac alkaloids having an S-configuration, although its activity was extremely poor toward 7-O-methyl-N-deacetylisoipecoside and 6,7-O,O-dimethyl-N-deacetylisoipecoside (15).
Deglucosylation of 6-O-methyl-N-deacetylisoipecoside by IpeGlu1 gives multiple products due to nonenzymatic intramolecular reactions, of which the iminium cation formed by dehydration of the immediate IpeGlu1 product was supposed to serve as intermediate for emetine biosynthesis (15). To examine the involvement of IpeOMTs in the other O-methylation reaction (i.e. methylation of the 7-hydroxy group with respect to isoquinoline skeleton) prior to protoemetine formation, we tested synthetic 3-O-methylredipecamine (see supplemental Fig. S1) as a mimic of endogenous 6-O-methyl-N-deacetylisoipecoside aglycon. That resulted in the detection of 2-Omethylation activity of IpeOMT3 toward 3-O-methylredipecamine, which strongly supports the hypothesis that the methylation of the 7-hydroxy group of the isoquinoline skeleton occurs after deglucosylation of 6-O-methyl-N-deacetylisoipecoside at a certain biosynthetic step leading to protoemetine (Fig. 7).
These results suggest that methylation of the 6-hydroxy group before the 7-hydroxy group of the isoquinoline skeleton is a common feature in the biosynthesis of isoquinoline alkaloids. In the biosynthesis of the benzylisoquinoline alkaloids papaverine, laudanine, and palmatine, 6-hydroxy group of isoquinoline skeleton is methylated in the early step by norcoclaurine 6-OMT (16,18,22), and then the 7-hydroxy group is methylated in a later step by norreticuline 7-OMT (19), reticuline 7-OMT (16), and columbamine 2-OMT (21), respectively. In P. ipecacuanha, however, two sequential O-methylation steps leading from 7Ј-O-demethylcephaeline, the condensation product of protoemetine and the second dopamine, to emetine were performed in the opposite manner, where the 7Ј-hydroxy group of 7Ј-O-demethylcephaeline was methylated first by the highly specific (K m ϭ 1.0 M) IpeOMT2 to form cephaeline, followed by 6Ј-O-methylation of cephaeline to form the end product emetine by IpeOMT1 (Fig. 7). The catalytic efficiency of IpeOMT1 toward cephaeline (3.5 s Ϫ1 ⅐M Ϫ1 ) was much lower than that of IpeOMT2 toward 7Ј-O-demethylcephaeline (32,600 s Ϫ1 ⅐M Ϫ1 ). These enzymatic properties may be the reason for accumulation of both cephaeline and emetine as major ipecac alkaloids in P. ipecacuanha (1) despite the expression of IpeOMT1 and IpeOMT2 at the comparable levels (Fig. 6).
Involvement of IpeOMTs in the Biosynthesis of Ipecac Alkaloids Derived from (1R)-N-Deacetylipecoside-Ipecac alkaloids derived from (1R)-N-deacetylipecoside is known to be accumulated as alkaloidal glucosides such as ipecoside and alangiside (5, 6). Their O-methyl derivatives (Fig. 8) have also been identified (2). It was found herein that O-methylation reactions of the alkaloidal glucosides having an R-configuration are catalyzed only by IpeOMT2 (Table 1), the catalytic property of which revealed the biosynthetic pathway to the R-form ipecac alkaloids (Fig. 8) ipecacuanha. This would be justified by the fact that none of the IpeOMTs catalyzed their formation (Fig. 8). However, we cannot exclude the possibility that other OMTs that were not discovered in this study are involved in their formation, because methylalangiside has been isolated in Alangium (31,32). Deep transcriptome sequencing would enable the complete identification of the remaining ipecac alkaloid biosynthetic genes, including such OMT(s), as well as those involved in the unidentified pathway leading to protoemetine from 6-O-methyl-N-deacetylisoipecoside aglycon.
Subcellular Network of Ipecac Alkaloid Biosynthetic Enzymes-It has been demonstrated that the terpenoid-indole alkaloid, strictosidine, which is formed by condensation of tryptamine and secologanin, is synthesized and stored in the vacuole (33,34). Likewise, P. ipecacuanha N-deacetylisoipecoside and N-deacetylipecoside that are formed by the condensation of dopamine and secologanin should be synthesized in the vacuole. We demonstrated that IpeOMTs and IpeGlu1, which work toward N-deacetylisoipecoside and N-deacetylipecoside, are localized in cytosol (Fig. 5). This means that N-deacetylisoipecoside and N-deacetylipecoside need to be transported outside the vacuole to react with the subsequent enzymes. For emetine biosynthesis, N-deacetylisoipecoside transported into the cytosol immediately has to be captured by IpeOMT1 for 6-Omethylation prior to deglucosylation by IpeGlu1. Despite colocalization of IpeOMT1 and IpeGlu1 in the cytosol, this may be possible because IpeOMT1 showed a considerably lower K m value (107 M) for N-deacetylisoipecoside than did IpeGlu1 (5.8 mM) (15).
The biosynthetic process of glucosidal ipecac alkaloids having an R-configuration appears to be more complicated. Considering that ipecoside is accumulated in P. ipecacunaha regardless of the presence of IpeGlu1 that exhibits the highest catalytic activity toward ipecoside (15), ipecoside and IpeGlu1 must be localized in distinct subcellular compartments. Thus, we hypothesized that the R-diastereomer N-deacetylipecoside is synthesized and acetylated in the vacuole. This is supported by the fact that nonglucosidal ipecac alkaloid with an R-configuration has not been found so far in P. ipecacuanha. In P. ipecacuanha, however, 6-O-methylipecoside and 7-O-methylipecoside are also accumulated (2). Because IpeOMT2 accepted N-deacetylipecoside but not ipecoside as substrate, 6-O-methylipecoside and 7-O-methylipecoside are presumed to be formed via N-acetylation of 6-O-methyl-N-deacetylipecoside and 7-O-methyl-N-deacetylipecoside, respectively. For this to occur, N-deacetylipecoside has to be transported into the cytosol to react with IpeOMT2. N-Deacetylipecoside, however, can then react with the cytosolic IpeGlu1, which exhibits a high catalytic activity toward it. Moreover, the 6-O-methyl-Ndeacetylipecoside and 7-O-methyl-N-deacetylipecoside produced need to be transported into the vacuole again based on the above-mentioned hypothesis that N-acetylation occurs in the vacuole. Discrepancies in subcellular localization between substrate and the corresponding enzyme in natural product biosynthesis have also been reported for 2-O-␤-D-glucopyranosyl-4-hydroxy-1,4-benzoxazin-3-one (DIBOA-Glc)-metabolizing enzymes in maize, where the substrate DIBOA-Glc was considered to be sequestered in the vacuole, but two enzymes (2-oxoglutarate-dependent dioxygenase and O-methyltransferase) metabolizing DIBOA-Glc were localized in cytosol (35). As exemplified by anthocyanin transport into the vacuole by a multidrug resistance-associated protein-type ATP-binding cassette transporter in maize (36,37), the mechanism of vacuolar transport of plant secondary metabolites is becoming unveiled (38). To reveal the subcellular network of ipecac alkaloid biosynthetic enzymes, it is essential to identify such transporters, as well as to determine subcellular localization of all biosynthetic enzymes.
Moreover, the cell-specific expression of the biosynthetic genes must be taken into consideration. Immunofluorescence analysis of the morphine biosynthetic enzymes in opium poppy P. somniferum demonstrated that in capsule and stem, 3Ј-hydroxy-N-methylcoclaurine 4Ј-OMT, reticuline 7-OMT, and salutaridinol 7-O-acetyltransferase were found predominantly in parenchyma cells within the vascular bundle, whereas codeinone reductase is localized to laticifers (39). Accordingly, we can hypothesize that ipecac alkaloids having the S-and R-configuration would be synthesized in distinct types of cells; the former cells need to co-express IpeOMTs and IpeGlu1 in an identical cell to complete the emetine biosynthesis, but the latter might express IpeOMT2 and an enzyme for N-acetylation but not IpeGlu1 to avoid deglucosylation of N-deacetylipecoside and its IpeOMT2 products, as well as their acetylated products.
Functional Differentiation of IpeOMTs in the Evolution of P. ipecacuanha-Because IpeOMTs catalyzed O-methylation reactions of 6-and/or 7-hydroxy groups of the isoquinoline skeleton of ipecac alkaloids, we expected that each of the three IpeOMTs shows the closest relationship according to its regiospecificity to the known 6-and 7-OMTs for benzylisoquinoline alkaloid biosynthesis. Phylogenetic analysis, however, demonstrated that three IpeOMTs are closely related to each other, and the clade is closer to the flavonoid OMTs and 16-hydroxytabersonine OMT than to those benzylisoquinoline alkaloid OMTs (Fig. 2). Similarly, norreticuline 7-OMT showed a closer relationship to norcoclaurine 6-OMT rather than to reticuline 7-OMT in P. somniferum (19). Pienkny et al. (19) hypothesized that it would be because 7-OMT had been generated by duplication and the following speciation of 6-OMT locus in the evolutionary process. This hypothesis is supported by the fact that IpeOMTs having distinct regiospecificity showed the closest relationship to each other. It is most likely that they have evolved in P. ipecacuanha after the differentiation from its ancestor through duplications and speciation of an ancestral form of OMT. Because ipecac alkaloids are produced not only in the genus Psychotria, but also in the genus Alangium despite their different plant families (Rubiaceae and Alangiaceae, respectively), molecular and enzymatic characterizations of ipecac alkaloid OMTs in Alangium should give an important clue to the evolution not only of ipecac alkaloid OMTs but also of plant OMTs involved in natural product biosynthesis in general.