Molecular cloning and characterization of norcoclaurine synthase, an enzyme catalyzing the first committed step in benzylisoquinoline alkaloid biosynthesis

Authors


(fax +403 289 9311; e-mail pfacchin@ucalgary.ca).

Summary

(S)-Norcoclaurine synthase (NCS) (EC 4.2.1.78) catalyzes the condensation of 3,4-dihydroxyphenylethylamine (dopamine) and 4-hydroxyphenylacetaldehyde (4-HPAA) as the first committed step in the biosynthesis of benzylisoquinoline alkaloids such as morphine, sanguinarine, and berberine, in plants. A molecular clone encoding NCS was isolated from a meadow rue (Thalictrum flavum ssp. glaucum) cell suspension culture cDNA library. Heterologous expression of the NCS cDNA, truncated to remove a putative signal peptide, produced a recombinant protein with NCS activity. Recombinant NCS showed sigmoidal saturation kinetics for dopamine (Hill coefficient=1.98), hyperbolic saturation kinetics for 4-HPAA (Km of 700 μm), and pH and temperature optima of 7.0 and 40°C, respectively, all similar to the purified, plant-derived enzyme. NCS exhibits 28–38% identity, and putative structural homology, with the Bet v 1 allergen and pathogenesis-related (PR)10 protein families. NCS also displays 35% identity with the enzyme (HYP1) responsible for hypericin biosynthesis in St John's wort (Hypericum perforatum). The novel catalytic functions of NCS and HYP1 define a new class of plant secondary metabolic enzymes within the Bet v 1 and PR10 protein families. Weaker homology was also detected between NCS and proteins identified in the latex of Papaver somniferum (opium poppy), and in Arabidopsis thaliana. A family of three to five NCS genes is abundantly expressed in the rhizome, followed by petioles and roots of T. flavum. NCS transcripts were localized to the immature endodermis and pericycle in roots, and the protoderm of leaf primordia in rhizomes; thus, the sites of NCS gene expression and berberine accumulation are temporally and spatially separated in roots and rhizomes respectively.

Introduction

Benzylisoquinoline alkaloids are a large and diverse group of secondary metabolites found in several related plant families including the Papaveraceae, Ranunculaceae, Berberidaceae, Fumariaceae, and Menispermaceae (Facchini, 2001). Many benzylisoquinoline alkaloids are pharmacologically active including the analgesics morphine and codeine, the antimicrobials sanguinarine and berberine, 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 to both 3,4-dihydroxyphenylethylamine (dopamine) and 4-hydroxyphenylacetaldehyde (4-HPAA) (Rueffer and Zenk, 1987). 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 (Figure 1).

Figure 1.

(S)-Norcoclaurine synthase (NCS) catalyzes the condensation of dopamine and 4-hydroxyphenylacetaldehyde to form (S)-norcoclaurine, the common precursor to all benzylisoquinoline alkaloids produced in plants including berberine, sanguinarine, and morphine.

Norcoclaurine synthase was first isolated based on its ability to convert dopamine and 3,4-dihydroxyphenylacetaldehyde (3,4-DHPAA) to the tetrahydroxylated alkaloid (S)-norlaudanosoline (Rueffer et al., 1981). 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 (Rueffer et al., 1981; Schumacher et al., 1983). However, only (S)-norcoclaurine has been found to occur in nature (Stadler and Zenk, 1990) and is now accepted as the central precursor to all benzylisoquinoline alkaloids produced in plants (Stadler et al., 1987, 1989). The 6-O-methylation of (S)-norcoclaurine, followed by N-methylation, P450-dependent 3′-hydroxylation, and subsequent 4′-O-methylation, leads to the formation of (S)-reticuline, a key branch-point intermediate in benzylisoquinoline alkaloid biosynthesis. Although cDNAs have been isolated for the enzymes involved in the conversion of (S)-norcoclaurine to (S)-reticuline (Choi et al., 2002; Morishige et al., 2000; Pauli and Kutchan, 1998), the molecular cloning of NCS has not been reported.

Previously, we isolated NCS activity in protein extracts from various benzylisoquinoline alkaloid-producing plants (Samanani and Facchini, 2001), and purified the enzyme to homogeneity from Thalictrum flavum cell cultures (Samanani and Facchini, 2002). We also showed that berberine accumulates in a cell type-specific manner in T. flavum plants – the endodermis and pericycle in roots, and the pith and cortex in rhizomes (Samanani et al., 2002). NCS is expected to play a regulatory role in the biosynthesis of berberine and other benzylisoquinoline alkaloids due to its entry-point location in the pathway and cooperative substrate-binding kinetics (Samanani and Facchini, 2001, 2002). In this paper, we report the molecular cloning and characterization of NCS from T. flavum. Recombinant NCS displays similar catalytic properties compared with the plant-derived enzyme, and is ancestrally related to an enzyme (HYP1) from St John's wort (Hypericum perforatum) responsible for the biosynthesis of the bioactive naphthodianthrone hypericin (Bais et al., 2003). The novel catalytic functions of NCS and HYP1 define a new class of plant secondary metabolic enzymes within the pathogenesis-related (PR)10 and Bet v 1 protein families.

Results

Heterologous expression of NCS

A full-length cDNA was isolated from a T. flavum cell culture library using a polymerase chain reaction (PCR) product amplified with degenerate primers based on peptide sequences obtained from the purified, native NCS protein. The cDNA contained a 630-bp open-reading frame (ORF) flanked by a 50-bp 5′-untranslated region (UTR), and a 142-bp 3′ UTR followed by a poly(A) tract (Figure 2). The ORF encoded a predicted translation product of 210 amino acids with a molecular mass of 23.3 kDa, and a pI of 5.5 (Figure 2). The microsequenced peptides used to design degenerate PCR primers were found in the deduced amino acid sequence between residues 159 and 171 (fraction 96) and residues 114 and 122 (fraction 105). Expression of the full-length NCS ORF did not result in detectable levels of recombinant protein or NCS activity in the bacterial protein extracts. However, constructs encoding truncations of the first 10 (pNCSΔ10) or the first 19 (pNCSΔ19) N-terminal amino acids from recombinant NCS directed the production of polypeptides with C-terminal His tags in protein extracts from bacterial cell cultures induced with isopropyl-β-d-thiogalactopyranoside (IPTG) (Figure 3a). Recombinant proteins were not produced in extracts isolated prior to the addition of IPTG to cultures harboring these constructs, or in extracts of cells containing the empty pET29b vector in the absence or presence of IPTG. The recombinant proteins displayed apparent molecular masses of 32 kDa by SDS-PAGE, which is similar to the predicted molecular masses of 27 and 28 kDa for pNCSΔ10 and pNCSΔ19 respectively. MALDI-TOF mass spectrometry analysis of NCSΔ10 revealed the expected masses for seven independent tryptic peptides, representing a total coverage of 48%, based on a stringent 0.001% maximum deviation of mass accuracy (data not shown). NCS activity was detected in protein extracts containing His-tagged, recombinant proteins, but was absent from control extracts (Figure 3b).

Figure 2.

Nucleotide and predicted amino acid sequences of a norcoclaurine synthase (NCS) cDNA from Thalictrum flavum ssp. glaucum cell suspension cultures. The underlined amino acids represent a putative signal peptide, whereas the boxed regions show residues identified through direct Edman sequencing of peptides obtained from the purified NCS protein.

Figure 3.

Heterologous expression of NCS in Escherichia coli. Western blot analysis performed using a His-tag monoclonal antibody shows the occurrence of recombinant proteins possessing a histidine tag (a). Numbers on the left refer to the location of standard protein molecular weight markers in kDa. NCS activity in total soluble protein extracts from non-induced (−) and induced (+) E. coli cells harboring the NCS expression constructs: pET29b, empty vector control; pNCSΔ10, pET29b harboring NCS truncated by the first 30 nucleotides of the open-reading frame (ORF); pNCSΔ19, pET29b harboring NCS truncated by the first 57 nucleotides of the ORF (b). The arrowhead shows the migration distance of authentic (S)-norcoclaurine on the thin layer chromatograph (TLC). IPTG, isopropyl-β-d-thiogalactopyranoside; M, molecular weight markers; O, TLC origin; F, TLC front.

General enzymatic properties of recombinant NCS

Purified, recombinant NCS showed optimal activity at pH 7.0, and half-maximal activities at pH 6.0 and 7.5, indicating a rapid decline in enzyme function under mildly acidic and basic conditions (Figure 4a). Recombinant NCS showed maximum activity at 40°C and half-maximal activities at 25 and 53°C (Figure 4b). Varying the concentration of dopamine from 55 to 280 μm at a saturating level (2.0 mm) of 4-HPAA produced sigmoidal saturation kinetics, suggesting positive cooperativity in the binding of dopamine to recombinant NCS (Figure 4c). Hill plot analysis (log[V (Vmax − V)−1] versus log[substrate concentration]) of the effect of dopamine concentration on recombinant NCS activity produced a linear function with a Hill coefficient (nH) of 1.98. Varying the concentration of 4-HPAA from 0.2 to 1.3 mm at a saturating level (280 μm) of dopamine produced hyperbolic saturation kinetics, indicating the lack of cooperativity on the binding of 4-HPAA to recombinant NCS (Figure 4d). Hill plot analysis of the effect of 4-HPAA concentration on recombinant NCS activity produced a linear function with nH = 0.98. Hanes plots ([substrate concentration]nHV−1 versus [substrate concentration]nH) confirmed the positive cooperativity of dopamine binding at a saturating level of 4-HPAA and Michaelis–Menten kinetics for 4-HPAA binding at a saturating level of dopamine by producing linear functions where nH = 1.98 and nH = 0.98 for dopamine and 4-HPAA respectively. Recombinant NCS showed a K′ constant of 0.0025 mmnH for dopamine and a Km of 700 μm for 4-HPAA (Figure 4e,f).

Figure 4.

General enzymatic properties of recombinant norcoclaurine synthase (NCS). The effect of pH (a) and temperature (b) on recombinant NCS activity. Hill plots showing the effect of substrate concentration on the reaction velocity (V) of NCS for dopamine (c) and 4-hydroxyphenylacetaldehyde (4-HPAA) (d). Hill coefficients were used to calculate Hanes plots for dopamine (e) and 4-HPAA (f). The negative x-intercept in (e) (arrowhead) represents the K′ of NCS for dopamine. The apparent Km for 4-HPAA was calculated as the negative x-intercept in (f) (arrowhead). R2 values in (c–f) refer to the best-fit linear functions.

Phylogenetic analysis

A blast search performed with the predicted amino acid sequence of NCS identified substantial homology with the Bet v 1 allergen and PR10 protein families from a variety of plant species (Figure 5). Most notably, NCS displayed 35% identity with the enzyme (HYP1) responsible for the biosynthesis of the bioactive naphthodianthrone hypericin in St John's wort (H. perforatum) (Bais et al., 2003). Amino acid sequence alignment also showed 28–38% identity and 50–60% homology between NCS and major allergen proteins, e.g. from birch (Betula pendula) (Breiteneder et al., 1989) and celery (Apium graveolens) (Breiteneder et al., 1995), and PR10 proteins, e.g. from rice (Oryza sativa), mung bean (Vigna radiata) (Fujimoto et al., 1998), pine (Pinus monticola) (Liu and Ekramoddoullah, 2003), beech (Fagus sylvatica) (Calvo et al., 1999), and potato (Solanum tuberosum) (Matton et al., 1990). NCS is unique among known Bet v 1 and PR10 proteins because of oligopeptide extensions at the N- and C-termini of the predicted translation product (Figure 5). Analysis of the NCS amino acid sequence using the SignalP Server (http://www.cbs.dtu.dk/services/SignalP) revealed the presence of a putative 19-amino acid N-terminal signal peptide (Neilsen et al., 1997). Generally, weaker homology was also detected between NCS and proteins identified in the latex of opium poppy (Nessler et al., 1990), and in Arabidopsis thaliana (Osmark et al., 1998). An unrooted, neighbor-joining phylogram based on the amino acid sequences of NCS, HYP1, selected Bet v 1, PR10 and major latex protein (MLP) homologs is shown in Figure 6. The phylogenetic tree topology supports an ancient ancestral origin for the Bet v 1/PR10 and MLP families.

Figure 5.

Alignment of the deduced amino acid sequence of Thalictrum flavum norcoclaurine synthase (NCS) with Bet v 1 and pathogenesis-related (PR)10 proteins from a variety of plant species: Hypericum perforatum HYP1 (Genbank accession number AAN65449); Betula pendula Bet v 1 major allergen (CAA07327); Oryza sativa PR10 (CAE04140); Apium graveolens Api g 1 (P49372); Vigna radiata PR10 (BAA74451); Pinus monticola PR10 (AAL49994); Fagus sylvatica PR10 (CAA10235); Solanum tuberosum PR10 (P17641). Sequences were aligned using clustal_x (Thompson et al., 1997). Shaded boxes indicate residues that are identical in at least 33% of the aligned proteins. Dots represent gaps introduced into sequences to maximize alignments.

Figure 6.

An unrooted, neighbor-joining phylogenetic tree based on the sequences for Thalictrum flavum norcoclaurine synthase (NCS), Hypericum perforatum HYP1, Betula pendula Bet v 1, Oryza sativa pathogenesis-related (PR)10, Apium graveolens Api g 1, Pinus monticola PR10, Fagus sylvatica PR10, Solanum tuberosum PR10, and two major latex proteins (MLPs) from Papaver somniferum, MLP1 (S28427) and MLP2 (X54305) and four Arabidopsis thaliana MLP homologs, MLP1 (NM102160), MLP2 (NM105754), MLP3 (NM105756), and MLP4 (NM117481). Sequence alignments and phylogeny were performed using clustal_x (43) and the phylogenetic data were displayed using treeview (Page, 1996).

Gene copy number and transcript localization

Gel-blot hybridization analysis of EcoRI- or HindIII-digested T. flavum genomic DNA at high stringency revealed strong (i.e. at least 80%) nucleotide sequence identity between NCS and three to five restriction fragments in each digest (Figure 7). RNA gel-blot hybridization analysis showed that NCS transcripts accumulated most abundantly in the rhizome and to a lesser extent in petioles, roots, leaves, and flower buds of mature T. flavum plants (Figure 8a). The highest level of NCS activity was also detected in the rhizome, followed by roots, leaves, and flower buds (Figure 8b). NCS activity could not be determined in protein extracts from mature petioles as [8-14C]dopamine was converted to an unknown reaction product, which interfered with the measurement of (S)-norcoclaurine levels (data not shown). In situ hybridization analysis using a digoxigenin (DIG)-labeled antisense probe demonstrated that NCS transcripts were localized to the immature endodermis (i.e. lacking a Casparian strip) and the pericycle in roots (Figure 9a–c), and the protoderm of leaf primordia in rhizomes (Figure 9d–f). NSC transcripts were also detected in localized regions inside the immature stele (Figure 9a). It is notable that NCS transcripts were not detected in mature endodermal or pericycle cells of root sections undergoing secondary growth or in cortical cells of the rhizome. No signals were detected in any tissues exposed to a DIG-labeled sense probe for NCS (Figure 9c,f).

Figure 7.

Southern blot analysis of Thalictrum flavum genomic DNA hybridized to the full-length norcoclaurine synthase (NCS) cDNA. Genomic DNA (20 μg) was digested with EcoRI or HindIII, fractionated, transferred to a nylon membrane, and hybridized at high stringency to the 32P-labeled NCS cDNA. Numbers on the left refer to the location of standard DNA size markers in kilobases (kb).

Figure 8.

Relative abundance of norcoclaurine synthase (NCS) transcripts (a) and NCS activity (b) in various organs of Thalictrum flavum. RNA gel-blot hybridization analysis (a) was performed using total RNA (15 μg), which was fractionated, transferred to a nylon membrane, and hybridized at high stringency to the 32P-labeled NCS cDNA. Gels were stained with ethidium bromide prior to blotting to ensure equal loading. Error bars (b) represent the standard deviation of three independent measurements. n.d., not determined.

Figure 9.

In situ RNA hybridization using a digoxigenin (DIG)-labeled antisense probe for NCS performed on root cross (a) and longitudinal sections (b), and rhizome apical region cross (d) and longitudinal (e) sections. In situ hybridization using a DIG-labeled sense probes for NCS performed on root (c) and rhizome apical region (f) cross sections. Abbreviations: ap, apical meristem; co, cortex; en, endodermis; lp, leaf primordial; pr, protoderm; rc, root cap; st, stele. Bars = 100 μm.

Discussion

Plants are able to produce a remarkable array of complex natural products based on modifications to relatively simple precursors. More than 2500 known benzylisoquinoline alkaloids produced in plants are all derived via the elaboration of (S)-norcoclaurine, a common intermediate resulting from a condensation reaction involving the amine moiety of dopamine and the aldehyde group of 4-HPAA. The central biosynthetic role of (S)-norcoclaurine suggests the ancient recruitment of NCS from a Bet v 1/PR10 ancestor as the key evolutionary event allowing certain plant taxa to produce numerous, structurally diverse benzylisoquinoline alkaloids with a range of physiologic and pharmacologic functions (Figures 5 and 6). Bet v 1, a major allergen from birch trees, and related proteins of the Bet v 1 family are the main cause of pollen-related allergies (Ipsen and Lowenstein, 1993; Vieths et al., 2002). Bet v 1 polypeptides show 35–60% homology with PR10 proteins, which are ubiquitous in plants (Wen et al., 1997) and are generally expressed in response to microbial challenge (Swoboda et al., 1995; Warner et al., 1992). PR10 and Bet v 1 proteins have been reported to show kinase (Saraste et al., 1990) and RNase activities (Bufe et al., 1996; Swoboda et al., 1996), but definitive catalytic functions for members of this protein family have not been identified. Recently, Bet v 1 proteins have been suggested to possess a steroid-binding domain (Neudecker et al., 2001) and appear to bind a wide range of ligands such as fatty acids, flavonoids, and cytokinins (Morgensen et al., 2002). Although the physiologic role of this binding capacity is not known, the ability of Bet v 1 and PR10 proteins to interact with natural product ligands might have resulted in the evolution of unique catalytic functions, such as NCS activity, by some ancestors of this family.

A role for certain Bet v 1/PR10 proteins in plant secondary metabolism is supported by the recent cloning of the gene encoding HYP1, which is responsible for the biosynthesis of hypericin in H. perforatum (Bais et al., 2003). Hypericin biosynthesis involves the initial condensation of emodine and emodine anthrone, followed by dehydration to form emodine dianthrone, which appears to undergo successive phenolic oxidations to yield the bioactive naphthodianthrone (Bais et al., 2003). The sequence similarity between NCS and HYP1 suggests the occurrence of other Bet v 1/PR10-type enzymes involved in similar condensation reactions situated early in other plant natural product pathways. It is notable that other, evolutionarily unrelated, pollen allergen-type proteins have been implicated in plant secondary metabolism. Phenylcoumarin benzylic ester reductase (Karamloo et al., 2001) and isoflavone reductase (Gang et al., 1999) are involved in isoflavonoid and lignan biosynthesis and share homology with the Bet v 6-class allergen.

The homology between NCS and MLP (Figure 6) is noteworthy as the benzylisoquinoline alkaloids of opium poppy accumulate in the latex of specialized laticifers, where MLPs are the most abundant proteins (Facchini, 2001; Nessler et al., 1990). It is perhaps an intriguing coincidence that the enzyme catalyzing the first step in the biosynthesis of morphine and related alkaloids might be derived from the same ancestor as the major protein in the latex, where these natural products are stored. The physiologic functions of MLP are not known. An analysis of the homology between Bet v 1/PR10 proteins and MLP homologs in opium poppy, Arabidopsis, and other plant species, held together with the three-dimensional structure of Bet v 1 (Gajhede et al., 1996), supports a strong structural similarity among these protein families (Osmark et al., 1998). The tertiary structure of Bet v 1 and PR10 proteins consists of a seven-stranded β sheet bending around a long α helix (Biesiadka et al., 2002; Gajhede et al., 1996). The high sequence identity between NCS and Bet v 1 predicts a conserved tertiary structure. However, Bet v 1 was described as a monomer, whereas native NCS appears to assemble as a dimer (Samanani and Facchini, 2002).

Recombinant and plant-derived NCS displayed similar catalytic properties in terms of pH and temperature optima, relative thermal stability, sigmoidal saturation kinetics for dopamine, and Michaelis–Menten kinetics for 4-HPAA (Figure 4) (Samanani and Facchini, 2002). The apparent Km for 4-HPAA of 700 μm for the recombinant NCSΔ10 is approximately double that of the native enzyme (Samanani and Facchini, 2002). Minor differences might have resulted from a 4-kDa peptide encoded by the pET29b expression vector and fused to the N-terminus of the recombinant enzyme. Overall, the recombinant protein will facilitate the reliable determination of additional enzymatic properties, including the three-dimensional structure and reaction mechanism of NCS.

Norcoclaurine synthase responds to a relatively modest increase in dopamine concentration with a substantial increase in activity, as the binding of dopamine to one subunit increases the affinity for dopamine of the other subunit. Enzymes exhibiting such sigmoidal substrate saturation kinetics invariably play a regulatory role in metabolism. Dopamine was detected at a level of 16% (w dry w−1) in cultured Papaver bracteatum cells (Kutchan et al., 1986), and found at concentrations of 1 mg ml−1 in the latex of P. somniferum (opium poppy) and P. bracteatum (Roberts et al., 1983). The cellular abundance of dopamine is surprising if the availability of this substrate is expected to play a role in the regulation of metabolic flux into benzylisoquinoline alkaloid pathways. However, the dopamine pool was localized within a vacuolar compartment in cultured P. bracteatum cells (Kutchan et al., 1986) suggesting that the subcellular trafficking of dopamine, and its availability to NCS, represent additional levels of regulation. The presence of a putative signal peptide suggests that NCS is associated with a subcellular compartment other than the cytosol (Figure 2), which could mediate accessibility to the cellular dopamine pool. The cytosol is the expected subcellular location for all related Bet v 1/PR10 proteins. The duplication and subsequent insertion of an NCS ancestor into another gene encoding an N-terminal signal peptide could explain the unique occurrence of this sequence in NCS. Subcellular compartmentation of NCS is consistent with the association of other benzylisoquinoline alkaloid biosynthetic enzymes with the endoplasmic reticulum (Facchini, 2001). The TargetP Server (http://www.cbs.dtu.dk/services/TargetP) also supports the targeting of NCS to the secretory pathway (Emanuelsson et al., 2000).

DNA gel-blot analysis suggests the presence of a small family of three to five NCS genes in T. flavum (Figure 7), which is consistent with the detection of four charge isoforms by two-dimensional PAGE (Samanani and Facchini, 2002). The pI of 5.5 for the predicted NCS translation product is also in agreement with the pI values of the native protein (Samanani and Facchini, 2002). The abundance of NCS transcripts was consistent with the occurrence of NCS activity (Figure 8a,b), the abundance of berberine (Samanani et al., 2002), and various organs of mature T. flavum plants. Previously, we reported high levels of NCS activity in the roots of young plantlets (Samanani and Facchini, 2002), but NCS activity was considerably lower in the roots of mature plants (Figure 8b).

The localization of NCS transcripts in immature endodermis and pericycle of roots (Figure 9a–c) shows that expression of a key biosynthetic gene involved in the benzylisoquinoline alkaloid pathway is temporally distinct from the sites of berberine accumulation, which are restricted to mature endodermis and pericycle tissues after the initiation of secondary growth in roots (Samanani et al., 2002). Our data suggest that alkaloids are synthesized and accumulate in the same cell types at different stages of root development in a temporal-specific manner. In contrast, the detection of abundant NCS transcripts in the protoderm of leaf primordia (Figure 9d–f) suggests that intercellular translocation of benzylisoquinoline alkaloids also occurs in rhizomes as berberine accumulates throughout the cortex and pith (Samanani et al., 2002). In the rhizome, our data suggest that alkaloid biosynthesis and accumulation are spatially separated. However, the relationship between the cellular sites of alkaloid synthesis and accumulation in both the roots and shoots of T. flavum is dramatically distinct from the localization of these processes in P. somniferum (Bird et al., 2003). In opium poppy, benzylisoquinoline alkaloids such as morphine accumulate in specialized laticifers, which are adjacent to the sieve elements of the phloem that contain biosynthetic enzymes (Bird et al., 2003). Gene transcripts encoding these enzymes were localized to the companion cell associated with each sieve element (Bird et al., 2003). The species-specific cell types involved in the biosynthesis and accumulation of structurally related alkaloids in T. flavum and P. somniferum show that even alkaloid pathways with a common phylogenetic origin do not necessarily share a conserved developmental program. Endodermis and pericycle have also been implicated in the biosynthesis of tropane and pyrrolizidine alkaloids (De Luca and St Pierre, 2000; Facchini, 2001; Moll et al., 2002), which are biosynthetically unrelated to benzylisoquinoline alkaloids. The tissue-specific localization of other steps in berberine biosynthesis in T. flavum will show whether the pathway is restricted to the endodermis, pericycle, and leaf primordia epidermis, or if other cell types are also involved.

A molecular clone for NCS almost completes the collection of cDNAs encoding enzymes involved in berberine biosynthesis. Previously, cDNAs for the berberine biosynthetic enzymes (S)-norcoclaurine 6-O-methyltransferase (Morishige et al., 2000), (S)-coclaurine N-methyltransferase (Choi et al., 2002), (S)-3′-hydroxy-N-methylcoclaurine 4′-O-methyltransferase (Morishige et al., 2000), berberine bridge enzyme (Chou and Kutchan, 1998), (S)-scoulerine 9-O-methyltransferase (Takeshita et al., 1995), and (S)-canadine synthase (Ikezawa et al., 2003) have been reported. Together with the cDNA for (S)-N-methylcoclaurine 3′-hydroxylase from a related benzylisoquinoline alkaloid-producing plant (Pauli and Kutchan, 1998), molecular clones for all pathway enzymes from l-dopa to berberine have now been isolated with the exception of the final oxidase (Facchini, 2001). The availability of a complete set of genes encoding the berberine biosynthetic enzymes will create unprecedented opportunities to understand the regulation of alkaloid biosynthesis in plants. The molecular cloning and characterization of NCS also heightens our appreciation for the evolutionary ability to recruit unique catalytic functions from a wide variety of protein families and incorporate new enzymes into novel natural product pathways.

Experimental procedures

Plants and cell cultures

Thalictrum flavum ssp. glaucum plants were grown under greenhouse conditions at a day/night temperature of 20/18°C.

Chemicals

[8-14C]Dopamine hydrochloride (57.72 GBq mol−1) was purchased from Sigma-Aldrich (St Louis, MO, USA). 4-HPAA was synthesized as described previously (Samanani and Facchini, 2001). All reagents were passed through a 0.45-μm filter.

Nucleic acid isolation and analysis

Total RNA was isolated using the method of Logemann et al. (1987) and poly(A)+ RNA was selected by oligo(dT) cellulose chromatography. For gel-blot analysis, 15 μg of RNA was fractionated on a 1.0% (w/v) agarose gel, containing 7.5% (w/v) formaldehyde, before transfer to a nylon membrane. Thalictrum flavum genomic DNA (20 μg) was isolated, digested with restriction endonucleases, electrophoresed on a 1.0% (w/v) agarose gel, and transferred to a nylon membrane. Blots were hybridized with a random-primer 32P-labeled, full-length NCS probe. Hybridizations were performed at 65°C in 0.25 mm sodium phosphate buffer, pH 8.0, 7% (w/v) SDS, 1% (w/v) BSA, and 1 mm EDTA. Blots were washed at 65°C, twice with 2x SSC and 0.1% (w/v) SDS and twice with 0.2x SSC and 0.1% (w/v) SDS (1x SSC = 0.15 m NaCl and 0.015 m sodium citrate, pH 7.0), and autoradiographed with an intensifier at −80°C.

cDNA library construction

A unidirectional oligo(dT)-primed cDNA library was constructed in λUni-ZAPII XR, according to the manufacturer's instructions (Stratagene, La Jolla, CA, USA) using poly(A)+ RNA isolated from T. flavum cell suspension cultures. The primary library containing 1 × 107 phage was amplified prior to screening.

Peptide microsequencing

The purified NCS protein (Samanani and Facchini, 2002) was concentrated by ultrafiltration and subjected to SDS-PAGE on a 17% (w/v) acrylamide gel according to the method of Laemmli (1970). Gel fragments containing Coomassie blue-stained NCS protein bands were excised, rinsed twice with high-performance liquid chromatography (HPLC)-grade acetonitrile and water (1:1) for 3 min, and frozen at −80°C. Sequence analysis was performed after trypsin digestion of NCS at the Harvard Microchemistry Facility (Harvard University, Cambridge, MA, USA) by microcapillary reverse-phase HPLC nano-electrospray tandem mass spectrometry (μLC/MS/MS) on a Finnigan LCQ DECA quadrupole ion trap mass spectrometer (Thermo Electron Corporation, Waltham, MA, USA). Four distinct peptides were detected by HPLC that eluted at 74, 96, 105, and 137 min. Matrix-assisted laser desorption-ionization-time-of-flight (MALDI-TOF) mass spectrometry was performed on a Perspective Biosystems Voyager-DE STR mass spectrometer (Framingham, MA, USA). Peptides eluting at 96 and 105 min were large enough for Edman sequencing. The primary sequences for fraction 96 (SSTEYHVKPEFVK) and fraction 105 (FILVDNEHR) were supported by MALDI and partial MS/MS analysis. Underlined residues were regarded with the highest confidence.

Molecular cloning of NCS

Pairs of degenerate sense and antisense primers were synthesized based on high-confidence peptide sequences, where the letter I represents inosine: 96-1s (5′-ACNGARTAYCAYGTNAARCC-3′); 96-1a (5′-GGYTTNACRTGRTAYTCNGT-3′); 96-2s (5′-TAYCAYGTNAARCCNGARTT-3′); 96-2a (5′-AAYTCNGGYTTNACRTGRTA-3′); 105-1s (5′-GTNGAYAAYGARCAYMG-3′); 105-1a (5′-CKRTGYTCRTTRTCNAC-3′); 105-2s (5′-ATHYTIGTNGAYAAYGARCA-3′); 105-2a (5′-TGYTCRTTRTCIACNARDAT-3′). All combinations of sense and antisense primers failed to amplify DNA fragments displaying an ORF using the T. flavum cDNA library as the template. However, at an annealing temperature of 50°C, both the 96-1a and 96-2a primers yielded single PCR amplicons of approximately 550 bp when paired with the T3 primer, which anneals to a region in pBluescript SK flanking the multiple cloning site. The amplicons were ligated into pGEM-T (Promega, Madison, WI, USA), sequenced, and used to screen the cDNA library at high stringency. The longest of several cDNA clones was sequenced.

Heterologous expression of NCS

The PCR primers containing BamHI restriction sites were synthesized to amplify the ORF of NCS, minus 30 nucleotides (5′-TTGGATCCATTAATGTTGTTAGGA-3′) or 57 nucleotides (5′-TTGGATCC-ACAGAAACTGATTCTG-3′) from the 5′-end. The antisense PCR primer contained an XhoI restriction site and was specific to the 3′-end of the ORF (5′-TTCTCGAGGACTGTTATTATTGCG-3′). The amplicons were ligated into pET29b (Novagen, Madison, WI, USA) using the introduced BamHI and XhoI sites, and the vectors were used to transform Escherichia coli ER2566 pLys S cells (New England Biolabs, Beverly, MA, USA). Bacteria were grown at 37°C in Lauria-Bertani medium to an A600 of 0.6. The cultures were then induced with 0.3 mm IPTG for 3 h and collected by centrifugation.

NCS enzyme assay

Bacterial cells were resuspended in 100 mm Tris-HCl, pH 7.0, containing 12 mm 2-mercaptoethanol (2-ME), and lysed using a French press at 15 000 psi. The lysate was centrifuged at 5000 g for 15 min and the supernatant concentrated by ultrafiltration. Unless otherwise indicated, NCS activity was determined by incubating 15 μg of soluble protein in 100 mm Tris-HCl, pH 7.0, containing 311 pmol (187 Bq) [8-14C]dopamine and 10 nmol 4-HPAA in a total volume of 30 μl for 1.5 h at 37°C. Buffers used to determine optimum pH were 200 mm potassium phosphate, pH 5–7, and 200 mm Tris-HCl, pH 7–9, containing 12 mm 2-ME. The reaction was applied to a silica gel 60 F254 TLC plate (EM Science, Gibbstown, NJ, USA), which was developed in n-butanol:acetic acid:water (4:1:5, by vol.) and autoradiographed. Radiolabeled spots with an Rf of 0.60, corresponding to (S)-norcoclaurine (Samanani and Facchini, 2001), were scraped off the plates and the radioactivity quantified by liquid scintillation counting. Protein concentration was determined according to Bradford (1976).

In situ RNA hybridization analysis

The NCS cDNA in pBluescript SK was used to synthesize sense and antisense DIG-labeled RNA probes using T3 and T7 RNA polymerases respectively. DIG-labeled probes were hydrolyzed at 60°C in 40 mm sodium carbonate buffer, pH 10, to produce 200–400 nucleotide fragments. The pH was neutralized using 10% (v/v) acetic acid and the precipitated RNA resuspended in 50 μl of deionized water. Sections were deparaffinized with xylene and rehydrated using an ethanol series (1:0, 1:0, 9.5:0.5, 7:3, and 1:1 ethanol:water), incubated for 30 min in prehybridization buffer (100 mm Tris-HCl pH 8.0, 50 mm EDTA) containing 2–10 μg ml−1 proteinase K (Roche Diagnostics, Laval, Quebec, Canada), and blocked in TBS (10 mm Tris-HCl pH 7.5, 150 mm NaCl) containing 2 mg ml−1 glycine. Sections were postfixed in 3.7% (v/v) formaldehyde in PBS (100 mm sodium phosphate buffer, pH 7.2, 140 mm NaCl), incubated in 100 mm triethanolamine buffer, pH 8.0, containing 0.5% (v/v) acetic anhydride, and rinsed in TBS. Slides were inverted onto 100 μl of hybridization buffer [10 mm Tris-HCl, pH 6.8, 10 mm sodium phosphate buffer, pH 6.8, 40% (v/v) deionized formamide, 10% (w/v) dextran sulfate, 300 mm NaCl, 5 mm EDTA, 1 mg ml−1 yeast tRNA, 0.8 U ml−1 RNase inhibitor (Invitrogen, Burlington, Ontario, Canada), and 200–1000 ng ml−1 DIG–RNA] spread over a coverslip. Slides were sealed in a Petri dish lined with filter paper soaked in 50% (v/v) formamide, incubated overnight at 50°C, then immersed in 2x SSC (1x SSC = 300 mm NaCl, 30 mm sodium citrate, pH 7.0) at 37°C. Sections were incubated in 50 μg ml−1 RNase A (Roche Diagnostics) in 500 mm NaCl, 10 mm Tris-HCl, pH 7.5, 1 mm EDTA for 30 min at 37°C, and washed in 2 l of the following solutions for 1 h: 2x SSC and 1x SSC at room temperature, and 0.1x SSC at 65°C. Slides were rinsed in TBST [10 mm Tris-HCl pH 7.5, 150 mm NaCl, 0.3% (v/v) Triton-X100], blocked for 1 h in TBST containing 2% (w/v) BSA, inverted onto coverslips carrying 100 μl of goat anti-DIG- alkaline phosphatase (AP) conjugate (Roche Diagnostics) diluted 1:200 in TBST containing 1% (w/v) BSA, incubated for 2 h in sealed Petri dishes lined with TBST-soaked filter paper, and rinsed in TBST buffer. Color development was performed in AP buffer containing 400 μm 5-bromo-4-chloro-3-indolyl phosphate and 428 μm nitroblue tetrazolium for 24 h.

Acknowledgements

PJF is the Canada Research Chair of Plant Biotechnology. DKL is the recipient of a Natural Sciences and Engineering Research Council of Canada Graduate Scholarship. This work was funded by a grant from the Natural Sciences and Engineering Research Council of Canada to PJF.

Genbank accession number: AY376412.

Ancillary