Purification and characterization of norcoclaurine synthase. The first committed enzyme in benzylisoquinoline alkaloid biosynthesis in plants.

Norcoclaurine synthase (NCS; EC ) catalyzes the condensation of dopamine and 4-hydroxyphenylacetaldehyde (4-HPAA) as the first committed step in benzylisoquinoline alkaloid biosynthesis in plants. NCS was purified 1590-fold to homogeneity from cell suspension cultures of meadow rue (Thalictrum flavum ssp. glaucum). The purification procedure, which resulted in a 4.2% yield, involved hydrophobic interaction, anion exchange, hydroxyapatite, and gel filtration chromatography. Purified NCS displayed native and denatured molecular masses of approximately 28 and 15 kDa, respectively, suggesting that the enzyme is composed of two subunits. Two-dimensional polyacrylamide gel electrophoresis revealed two major and two minor isoforms with pI values between 5.5 and 6.2. NCS activity was maximal at pH 6.5 to 7.0 and temperatures between 42 and 55 degrees C and was not affected by divalent cations. The enzyme showed hyperbolic saturation kinetics for 4-HPAA (K(m) = 335 microm) but sigmoidal saturation kinetics for dopamine (Hill coefficient = 1.8) suggesting cooperativity between the dopamine binding sites on each subunit; thus, NCS might play a regulatory, or rate-limiting, role in controlling the rate of pathway flux in benzylisoquinoline alkaloid biosynthesis. Product inhibition kinetics performed at saturating levels of one substrate and with norlaudanosoline as the inhibitor showed that NCS follows an iso-ordered bi-uni mechanism with 4-HPAA binding before dopamine. NCS activity was highest in soluble protein extracts from roots followed by stems, leaves, and flower buds.

Benzylisoquinoline alkaloids are a large and diverse group of secondary metabolites found mainly in five related plant families, including the Papaveraceae and Ranunculaceae. Many benzylisoquinoline alkaloids are pharmacologically active including the analgesic and antitussive drugs morphine and codeine, the antibiotic sanguinarine, and the muscle relaxants papaverine and tubocurarine. The structural complexity of these pharmaceuticals generally precludes chemical synthesis as an alternative to cultivated plants for their commercial production. All benzylisoquinoline alkaloids share a common biosynthetic origin beginning with a lattice of decarboxylations, ortho-hydroxylations, and deaminations that convert L-tyrosine into both dopamine and 4-hydroxyphenylacetaldehyde (4-HPAA) 1 (1). The first committed step in benzylisoquinoline alkaloid biosynthesis is catalyzed by norcoclaurine synthase (NCS; EC 4.2.1.78), which condenses dopamine and 4-HPAA to form the trihydroxylated alkaloid (S)-norcoclaurine (see Fig. 1).
(S)-Norcoclaurine is now accepted as the central precursor to all benzylisoquinoline alkaloids produced in plants (2,3). However, NCS was first isolated based on its ability to convert dopamine and 3,4-dihydroxyphenylacetaldehyde (3, to the tetrahydroxylated alkaloid (S)-norlaudanosoline (4). The ability of NCS to accept either 4-HPAA or 3,4-DHPAA as a substrate contributed to the incorrect conclusion that (S)-norlaudanosoline is a common pathway intermediate (4,5). However, only (S)-norcoclaurine has been found to occur in plants (6). Sequential 6-O-methylation of (S)-norcoclaurine, followed by N-methylation, a P450-dependent 3Ј hydroxylation, and subsequent 4Ј-O-methylation lead to the formation of (S)-reticuline, an important branch-point intermediate in benzylisoquinoline alkaloid biosynthesis. Although cDNA clones have been isolated for the enzymes involved in the conversion of (S)-norcoclaurine to (S)-reticuline (7-9), NCS has not been extensively characterized.
Previously, we developed a sensitive assay to isolate NCS activity in protein extracts of opium poppy (Papaver somniferum) and related species (10). We showed that NCS activity occurs at different levels in all organs of the plant and is induced in opium poppy cell suspension cultures treated with a fungal elicitor. Opium poppy NCS was shown to exhibit hyperbolic saturation kinetics for 4-HPAA but sigmoidal saturation kinetics for dopamine suggesting that it plays a regulatory role in the regulation of benzylisoquinoline alkaloid pathways. In this paper, we report the purification to homogeneity and characterization of NCS from cell suspension cultures of meadow rue (Thalictrum flavum). T. flavum is a medicinal member of the Ranunculaceae that accumulates the benzylisoquinoline alkaloids berberine and magnoflorine.

EXPERIMENTAL PROCEDURES
Plants and Cell Cultures-Meadow rue (T. flavum ssp. glaucum) plants were grown under a photoperiod of 16 h of light and 8 h of darkness at 23°C. Cell suspension cultures were maintained at 23°C in diffuse light on a medium consisting of B5 salts and vitamins (11), 100 mg liter Ϫ1 myo-inositol, 1 g liter Ϫ1 hydrolyzed casein, 20 g liter Ϫ1 sucrose, and 1 mg liter Ϫ1 2,4 dichlorophenoxyacetic acid. Cells were collected by vacuum filtration and frozen at Ϫ80°C.
Enzyme Extraction and Assay-Protein fractions were incubated in buffer A with 311 pmol (5.0 nCi) [8-14 C]dopamine and 10 nmol 4-HPAA in a total volume of 30 l unless stated otherwise. After an incubation period of 1.5 h at 37°C, the entire reaction volume was applied to a silica gel 60 F 254 TLC plate (EM Science). The TLC plates were developed in a solvent system consisting of n-butanol:acetic acid:water (4:1:5) and, subsequently, autoradiographed for 12-24 h using X-OMAT film (Kodak). Radiolabeled spots with an R f of 0.60 were scrapped from the plate, and the radioactivity quantified by liquid scintillation counting. Controls were performed using protein samples incubated in boiling water for 15 min. NCS activity was calculated after subtracting the background radioactivity produced in control reactions.
Protein Determination and Preparation of the Crude Protein Extract-Protein concentration was determined according to the method of Bradford (12) using bovine serum albumin as the standard. Plant cells (1000 g) were ground under liquid nitrogen to a fine powder with a mortar and pestle and extracted in buffer A. The slurry was centrifuged at 20,000 ϫ g for 20 min, and the supernatant was desalted on a Sephadex G-25 column (50 ϫ 230 mm). Solid ammonium sulfate was added to 20% saturation, the extract was centrifuged at 10,000 ϫ g for 20 min, and the pellet was discarded. Additional ammonium sulfate was added to 60% saturation, and the sample was centrifuged again. The pellet was dissolved in buffer B containing 15% ammonium sulfate.
Phenyl-Sepharose Chromatography-The protein extract was loaded onto a Phenyl-Sepharose column (12 mm ϫ 100 mm) equilibrated in buffer B containing 15% ammonium sulfate. The loaded column was thoroughly washed with buffer B before elution of the bound proteins using a 15-0% linear ammonium sulfate gradient over 2 h at a flow rate of 1 ml min Ϫ1 . After the collection of fraction 78, buffer B was replaced with deionized water at a flow rate of 2 ml/min. Fractions (5 ml) were collected and assayed for NCS activity.
Q-Sepharose Chromatography-Active fractions from phenyl-Sepharose chromatography were pooled and loaded onto a Q-Sepharose column (12 mm ϫ 100 mm) equilibrated in buffer C. The column was thoroughly washed with buffer C before applying a linear gradient of 0 -1 M KCl over 2 h at a flow rate of 0.5 ml min Ϫ1 . Fractions (2.5 ml) were collected and assayed for NCS activity.
Hydroxyapatite Chromatography-Active protein from Q-Sepharose chromatography was pooled and the buffer changed to 10 mM potassium phosphate, pH 7.2, using Centricon Plus-20 ultrafiltration units (Millipore). The concentrated protein was loaded onto a hydroxyapatite column (12 mm ϫ 100 mm) equilibrated with buffer D. A linear 10 -500 mM potassium phosphate gradient was applied over 2 h at a flow rate of 0.5 ml min Ϫ1 . Fractions (2.5 ml) were collected and assayed for NCS activity.
Superose-12 Gel Filtration Chromatography-Active fractions from hydroxyapatite chromatography were pooled, the protein was concentrated, and the buffer was changed to buffer A by ultrafiltration. The 200-l protein sample was applied to two Superose-12 columns (5 mm ϫ 300 mm each) connected in tandem and equilibrated with buffer A. The proteins were eluted with buffer A at a flow rate of 0.2 ml min Ϫ1 . Fractions (200 l) were collected and assayed for NCS activity Gel Electrophoresis-Protein preparations at different stages of purification were visualized by gel electrophoresis under denaturing conditions according to Laemmli (13) using a 17% (w v Ϫ1 ) polyacrylamide gel (SDS-PAGE). The molecular weight of purified NCS was determined from a graph of the log molecular weight versus migration distance of protein standards. Two-dimensional electrophoresis was performed according to the method of O'Farrell (14) except that immobilized protein gradient gel strips were used for the first dimension and 17% (w v Ϫ1 ) denaturing polyacrylamide gels for the second dimension. Proteins were visualized by staining gels with Coomassie Brilliant Blue.
Estimation of Native Molecular Mass, Chemical Cross-linking, and Mass Spectrometry-The molecular weight of native NCS was estimated by gel filtration chromatography on tandem Superose-12 columns (5 mm ϫ 300 mm each) at a flow rate of 0.2 ml/min. The columns were calibrated using bovine serum albumin (67 kDa), ovalbumin (45 kDa), chymotrypsinogen A (25 kDa), and ribonuclease A (12 kDa). The void volume was calculated by the elution of blue dextran 2000. The native protein was treated with dimethyl suberimidate as a chemical cross-linking agent as described previously (15). Protein spots from two-dimensional gel electrophoresis gels were analyzed using a PerSeptive Biosystems Voyager DE-STR matrix-assisted laser desorption ionization-time of flight mass spectrometer after in-gel trypsin cleavage.

RESULTS
Distribution of NCS Activity in T. flavum-NCS activity was detected in roots, stems, leaves, and flower buds of T. flavum plants (Fig. 2). When normalized against total soluble protein, NCS activity was highest in roots and was 1.4-, 1.8-, and 4.5-fold lower in stems, leaves, and flower buds, respectively. NCS activity was also detected in T. flavum cell cultures at levels similar to those found in stems and leaves (Fig. 2).
Purification of NCS-NCS was purified from suspensioncultured T. flavum cells as summarized in Table I. The homogenate from 1000 g of cells was initially centrifuged at 13,000 ϫ g to remove insoluble debris. NCS activity was never associated with the pellet. The desalted supernatant was fractionated by ammonium sulfate precipitation and subjected to hydrophobic interaction chromatography. NCS eluted from the phenyl-Sepharose column in two peaks of activity (Fig. 3A). An initial activity peak eluted between fractions 51 and 77 after the ammonium sulfate concentration in buffer B was reduced to 0%. Continued elution of the column with deionized water beginning at fraction 79 resulted in a second peak of NCS activity between fractions 81 and 88. Protein eluted and stored in deionized water retained NCS activity for several days at 4°C. Ammonium sulfate precipitation and hydrophobic interaction chromatography enriched NCS activity 13-fold. Subsequent anion exchange chromatography on Q-Sepharose followed by hydroxyapatite chromatography resulted in an additional purification of 4.5-fold. NCS activity eluted as single peaks from the Q-Sepharose column between 500 and 1000 mM potassium chloride (fractions 13 through 23; Fig. 3B) and from the hydroxyapatite column between 200 and 350 mM potassium phosphate (fractions 11 through 17; Fig. 3C). The final purification step was gel filtration chromatography using tandem Superose-12 columns, which resulted in an overall purification of 1590-fold. NCS activity correlated with the elution of a single protein peak between fractions 54 and 65. However, only 18% of the activity loaded onto the gel filtration column was recovered (Table I). As the purification progressed, a 15-kDa protein became prominent on Coomassie Blue-stained SDS-polyacrylamide gels and was the only detectable protein recovered from the Superose-12 chromatography step (Fig. 4).  General Properties of NCS-The molecular mass of native NCS was determined by chromatography on tandem Superose-12 columns calibrated with molecular mass standards. Purified NCS eluted with an apparent molecular mass of ϳ28 kDa. Chemical cross-linking of native NCS with dimethyl suberimidate also resulted in a 28-kDa polypeptide detectable by SDS-PAGE. The purified protein was subjected to twodimensional gel electrophoresis using an isoelectric focusing gel with an immobilized linear pH gradient (Fig. 5). Two major and two minor charge isoforms were visualized after SDS-PAGE on a Coomassie Blue-stained gel. Calibration of the gel using isoelectric point standards showed pI values of 5.7 and 6.0 for the two major isoforms, and 5.5 and 6.2 for the two minor isoforms. All charge isoforms had the same molecular mass of ϳ15 kDa. Matrix-assisted laser desorption ionizationtime of flight mass spectrometry on the two major isoforms identified two independent tryptic peptides, representing a total coverage of ϳ16%, with identical molecular masses in each protein based on a stringent 0.001% maximum deviation of mass accuracy.
Purified NCS showed optimal activity between pH 6.5 and 7.0 and half-maximal activity at pH 6.2, indicating a rapid decline in enzyme function with increasing acidity. However, the enzyme was relatively insensitive to pH increases in the basic range. NCS showed high levels of activity at temperatures between 45 and 55°C, although half-maximal activities were found at 37 and 57°C. The purified enzyme was unaffected by the inclusion of either 40 mM CaCl 2 or MgCl 2 in the standard assay. No detectable loss in NCS activity occurred when the purified protein was stored at Ϫ80°C for 7 days. However, a 75% loss in activity occurred when the protein was stored for 7 days at 4°C. Product Inhibition Patterns-The order of substrate binding was determined by an analysis of product inhibition patterns using norlaudanosoline as the inhibitor (Fig. 6). A Hanes plot showing the activity of purified NCS at different concentrations of one substrate and saturating levels of the second substrate using various fixed concentrations of the inhibitor showed that the binding of norlaudanosoline is uncompetitive with respect to 4-HPAA (Fig. 6A), but noncompetitive with respect to dopamine (Fig. 6B). The K i value for norlaudanosoline was 4.7 mM with respect to dopamine. All experiments were performed in duplicate with reproducible results. DISCUSSION Norcoclaurine synthase was purified to homogeneity from cell suspension cultures of T. flavum using a 5-step purification procedure. The protocol resulted in an overall purification of 1590-fold with a recovery of 4.2%. Based on the initial NCS activity of the crude extract, the enzyme constituted ϳ0.06% of the soluble protein. Despite the relatively low abundance of NCS activity in cultured T. flavum cells, the combination of ammonium sulfate precipitation and hydrophobic interaction, anion exchange, hydroxyapatite, and gel filtration chromatography allowed the enzyme to be purified due to its relatively high hydrophobicity and small molecular mass (Fig. 3). It should be noted that desalting the crude protein extract was necessary to detect NCS activity suggesting the presence of inhibitory compounds. Phenyl-Sepharose chromatography was performed using an ammonium sulfate gradient of 15-0% saturation to prevent precipitation of the protein fraction obtained by an ammonium sulfate fractionation between 20 and 60% saturation.
The rapid growth of cultured plant cells has facilitated the purification of several enzymes involved in alkaloid biosynthesis (16). Even though T. flavum roots showed 2-fold higher NCS activity than cultured cells (Fig. 2), large quantities of suspension-cultured cells were easier to produce due to the relatively slow growth rate of the intact plants. T. flavum cell cultures were selected over those of other benzylisoquinoline alkaloid- producing species, such as P. somniferum and Eschscholzia californica, for the purification of NCS due to the relative instability of the enzyme in some crude protein extracts (10). For example, NCS activity could not be recovered after ammonium sulfate precipitation of crude protein extracts from P. somniferum or E. californica cell cultures.
The purified 15-kDa band (Fig. 4) was identified as NCS by subjecting consecutive fractions across the peak of enzyme activity obtained by Superose-12 gel filtration chromatography (Fig. 3) to SDS-PAGE analysis (data not shown). The intensity of the 15-kDa, Coomassie Blue-stained band correlated with the relative level of NCS activity and the relative abundance of protein in these fractions. The native molecular mass of 28 kDa, determined by gel filtration chromatography, suggests that the native enzyme is a dimer comprised of two 15-kDa subunits. The dimeric structure of NCS was corroborated by the formation of a 28-kDa polypeptide by chemical cross-linking of the native protein complex (15). These results are at variance with the native molecular mass of 15.5 kDa previously reported for NCS (5). However, it should be noted that a partially purified enzyme extract and a lower performance Sephadex G-100 column were used in this earlier study.
The detection of four NCS isoforms, two of relatively high abundance and two of lower abundance (Fig. 5), is in agreement with a previous characterization of the enzyme (5). The existence of four charge isoforms was indicated by the migration of NCS activity in partially purified protein extracts using disc gel electrophoresis. The presence of four isoforms suggests that native NCS is either a homodimer consisting of only one isoform or a heterodimer comprised of two different isoforms. Matrix-assisted laser desorption ionization-time of flight mass spectrometry showed that the two major NCS charge isoforms are related; thus, the native protein complex consists of two similar, if not identical, subunits. The characterization of NCS as a dimer is also consistent with the sigmoidal binding kinetics of dopamine, which demonstrates the cooperative binding of this substrate. The Hill coefficient for dopamine (n H ϭ 1.8) suggests that NCS displays positive cooperativity between the dopamine binding sites on each subunit. The sigmoidal binding kinetics of dopamine and hyperbolic binding kinetics of 4-HPAA for NCS from T. flavum are consistent with the substrate binding kinetics of the enzyme from P. somniferum and E. californica (10).
Invariably, enzymes exhibiting sigmoidal substrate saturation kinetics play a regulatory role in metabolism. As such, NCS responds to a relatively modest increase in dopamine concentration with a substantial increase in activity because the binding of dopamine to one subunit increases the affinity for dopamine of the other subunit. Dopamine was detected at a level of 16% (w dry w Ϫ1 ) in cultured Papaver bracteatum cells (17) and found at concentrations of 1 mg ml Ϫ1 in the latex of P. bracteatum and P. somniferum (18). However, the dopamine pool was shown to be localized within a vacuolar compartment in cultured P. bracteatum cells (17) suggesting that the subcellular trafficking of dopamine, and its availability to NCS, represent additional levels of regulation.
NCS from T. flavum exhibited optimum catalytic activity between pH 6.5 and 7.0, did not require Mg 2ϩ or Ca 2ϩ as reported for NCS from Eschscholzia tenuifolia (5), and displayed a broad temperature optimum between 45 and 55°C suggesting that the enzyme is relatively thermostable. Overall, the properties of NCS from T. flavum are consistent with those of the enzyme from P. somniferum and E. californica (10). The kinetic mechanism for NCS was also investigated. NCS catalyzes a condensation reaction between dopamine and 4-HPAA resulting in the formation of a single product, (S)-norcoclaurine ( Fig. 1). Hanes plots were used to determine the binding order of 4-HPAA and dopamine (Fig. 6). Norlaudanosoline was used as the inhibitor due to the unavailability of (S)-norcoclaurine. NCS has been shown to catalyze the condensation of dopamine with 4-HPAA or 3,4-DHPAA with equal efficiency, leading to the formation of (S)-norcoclaurine or (S)-norlaudanosoline, respectively (4). Thus, norlaudanosoline is expected to inhibit NCS activity in the same manner as norcoclaurine. It is notable that (S)-norcoclaurine and (S)-norlaudanosoline, or their 6-Omethylated derivatives, are accepted as substrates by the 6-Omethyltransferase and N-methyltransferase, respectively, involved in subsequent steps of the benzylisoquinoline alkaloid pathway (19,20).
The Hanes plot showing the activity of purified NCS at different concentrations of dopamine, saturating levels of 4-HPAA, and various fixed concentrations of norlaudanosoline reflects a decrease in V max , but constant K m , as the amount of inhibitor was increased (Fig. 6A). In contrast, the Hanes plot showing the activity of purified NCS at different concentrations of 4-HPAA, saturating levels of dopamine, and various fixed concentrations of norlaudanosoline reflects a decrease in both V max and K m , with a constant V max K m Ϫ1 ratio as the amount of inhibitor was increased (Fig. 6B). Thus, the product inhibition pattern is consistent with uncompetitive binding with respect to 4-HPAA and noncompetitive binding with respect to dopamine, indicating that 4-HPAA binds to the enzyme before dopamine. The Hill coefficient for dopamine (n H ϭ 1.8) did not change in response to increasing inhibitor concentrations.
The product inhibition patterns also imply that 4-HPAA combines with a different form of the enzyme than the alkaloid because the inhibitor and first substrate do not bind competitively (21). After the product is released, NCS appears to undergo a conformational change reverting back to a form to which 4-HPAA can bind before another reaction sequence can begin (22). Overall, the product inhibition patterns are consistent with an iso-ordered, bi-uni mechanism for NCS.
NCS shows several features that are similar to strictosidine synthase (STR), which couples tryptamine and secologanin as the first committed step in monoterpenoid indole alkaloid biosynthesis. Both enzymes catalyze condensation reactions between amine and aldehyde substrates. Four STR isoforms were isolated from cell cultures of Catharanthus roseus, all of which exhibited cooperative binding kinetics for secologanin (23). Cooperative substrate binding might have evolved as a parallel mechanism to provide NCS and STR with a regulatory role in their respective pathways. STR and NCS also display similar native molecular masses of 31 and 28 kDa, respectively (24). However, STR is composed of a single polypeptide, whereas NCS appears to be a dimer, suggesting that the enzymes are not structurally related. It is also interesting to note that deacetylipecoside synthase, which catalyzes the condensation of dopamine and secologanin, is a monomer with a native molecular mass of 30 kDa (25).
The purification of NCS represents the first step in the isolation of the corresponding gene, the availability of which will reveal its evolutionary relationship with STR. Moreover, the key function of NCS in the formation of valuable benzylisoquinoline alkaloids makes the enzyme an intriguing target for molecular characterization. The unique biochemical characteristics of NCS, including its low molecular weight and sigmoidal substrate binding kinetics, also warrant crystallographic studies to better understand its catalytic mechanism and determine its role in the regulation of benzylisoquinoline alkaloid biosynthesis.