Benzalacetone synthase

A novel polyketide synthase that plays a crucial role in the biosynthesis of phenylbutanones in Rheum palmatum

Authors


I. Abe, University of Shizuoka, School of Pharmaceutical Sciences, 52-1 Yada, Shizuoka 422-8526, Japan. Fax/Tel.: + 81 54 264 5662, E-mail: abei@ys7.u-shizuoka-ken.ac.jp

Abstract

Benzalacetone synthase (BSA) is a novel plant-specific polyketide synthase that catalyzes a one step decarboxylative condensation of 4-coumaroyl-CoA with malonyl-CoA to produce the C6–C4 skeleton of phenylbutanoids in higher plants. A cDNA encoding BAS was for the first time cloned and sequenced from rhubarb (Rheum palmatum), a medicinal plant rich in phenylbutanoids including pharmaceutically important phenylbutanone glucoside, lindleyin. The cDNA encoded a 42-kDa protein that shares 60–75% amino-acid sequence identity with other members of the CHS-superfamily enzymes. Interestingly, R. palmatum BAS lacks the active-site Phe215 residue (numbering in CHS) which has been proposed to help orient substrates and intermediates during the sequential condensation of 4-coumaroyl-CoA with malonyl-CoA in CHS. On the other hand, the catalytic cysteine-histidine dyad (Cys164–His303) in CHS is well conserved in BAS. A recombinant enzyme expressed in Escherichia coli efficiently afforded benzalacetone as a single product from 4-coumaroyl-CoA and malonyl-CoA. Further, in contrast with CHS that showed broad substrate specificity toward aliphatic CoA esters, BAS did not accept hexanoyl-CoA, isobutyryl-CoA, isovaleryl-CoA, and acetyl-CoA as a substrate. Finally, besides the phenylbutanones in rhubarb, BAS has been proposed to play a crucial role for the construction of the C6–C4 moiety of a variety of natural products such as medicinally important gingerols in ginger plant.

Abbreviations
BAS

benzalacetone synthase

CHS

chalcone synthase

STS

stilbene synthase

CTAS

4-coumaroyltriacetic acid synthase

2PS

2-pyrone synthase

ACS

acridone synthase

BBS

bibenzyl synthase

VS

valerophenone synthase

RppA

red-brown pigment producing enzyme

CoA

coenzyme A

LC-ESIMS

liquid chromatography electron spray ionization mass spectrometry.

Chalcone synthase (CHS) is a plant-specific polyketide synthase that catalyze a condensation of 4-coumaroyl-CoA (1) with three C2 units from malonyl-CoA (2) to produce naringenin chalcone (4), a tetraketide with a new aromatic ring system (Scheme 1)[1]. CHSs have been cloned and sequenced from more than 40 plant species, and many of them have been functionally expressed in Escherichia coli. They are homodimers of a 40–45-kDa polypeptide, and share 60–75% amino-acid sequence identity with other members of the CHS-superfamily enzymes [2] including stilbene synthase (STS) [3] and 2-pyrone synthase (2PS) [4]. The CHS-superfamily enzymes assemble structurally diverse natural products using a common mechanistic strategy that relies on a cysteine residue to anchor the polyketide during a series of condensation reactions. The structure-function relationship of the enzymes has thus elicited intense chemical and biological interest. The recently reported crystal structure of CHS from alfalfa (Medicago sativa, Leguminosae) has provided detailed mechanistic and stereochemical insights into the enzyme reactions including the formation of the polyketide intermediate and the structural features governing the stereospecific cyclization reaction leading to chalcone in CHS [5,6].

Figure Scheme 1. .

Proposed mechanism for the conversion of 4-coumaroyl-CoA (1) and malonyl-CoA (2) to benzalacetone (3) by benzalacetone synthase (BAS), and to naringenin chalcone (4) by chalcone synthase (CHS).

Rhubarb (Pheum palmutum, Polygonaceae) is a medicinal plant that contains considerable amount of aromatic polyketides such as anthraquinones [7], stilbenes [8], naphtalenes [9], and phenylbutanones [10,11]. Therefore, in addition to regular CHSs involved in the biosynthesis of flavonoids, presence of functionally different polyketide synthases that catalyze the initial key reactions in the biosynthesis of these metabolites were expected. Here in this paper, we describe for the first time cloning and characterization of benzalacetone synthase (BAS), a novel plant-specific CHS-related polyketide synthase that catalyzes a one-step decarboxylative condensation of 4-coumaroyl-CoA (1) with malonyl-CoA (2) to produce benzalacetone; 4-(4-hydroxyphenyl)but-3-en-2-one (3) (Scheme 1). The enzyme plays a crucial role in the biosynthesis of pharmaceutically important phenylbutanone glucoside, lindleyin (6), the active principle of the anti-inflammatory action of the medicinal plant [10,11].

Materials and methods

Chemicals

4-Coumaroyl-CoA and 3-(4-hydroxyphenyl)propionyl-CoA were chemically synthesized as described previously [12,13]. Malonyl-CoA, hexanoyl-CoA, isobutyryl-CoA, isovaleryl-CoA, and acetyl-CoA were purchased from Sigma. 4-(4-hydroxyphenyl)but-3-en-2-one (3) was from Lancaster Synthesis Ltd., Lancaster, UK.

PCR and DNA sequencing

PCR was carried out with Ex-Taq DNA polymerase (Takara) according to the manufacturer's instruction using Robocycler Gradient 40 (Stratagene). The nucleotide sequences were determined from both strands with a DNA sequencer ABI 373A using Dye Terminator Cycle Sequencing FS Ready Reaction Kit (PE Applied Biosystems).

Extraction of plant RNA and cDNA preparation

Young leaves of R. palmatum were harvested at the medicinal plant garden of University of Shizuoka, Japan in July 2000, and immediately frozen with liquid nitrogen. Total RNA was extracted by the acid guanidium thiocyanate/phenol/chloroform method, and reverse-transcribed using Reverscript (Wako) and oligo dT primer (RACE32 = 5′-GGCCACGCGTCGACTAGTACTTTTTTTTTTTTTTTTT-3′) according to the manufacturer's protocol. The obtained cDNA mixture was diluted with Tris/EDTA (10 mm Tris, 1 mm EDTA, pH 8.0) and used as a template for the following PCR reactions.

PCR amplification of core cDNA fragment

As described previously [14], inosine-containing degenerate oligonucleotide primers (112S, 174S, 368A and 380A, the number of primer indicates the amino-acid number of corresponding CHS) based on the highly conserved sequences of known CHSs were used for amplification of a core fragment of the cDNA. The sequences of the primers are as follows: 112S = 5′-(A/G)A(A/G)GCIITI(A/C)A(A/G)GA(A/G)TGGGGICA-3′, 174S = 5′-GCIAA(A/G)GA(T/C)ITIGCIGA(A/G)AA(T/C)AA-3′, 368A = 5′-CCC(C/A)(A/T)ITCIA(A/G)ICCITCICCIGT IGT-3′, and 380A = 5′-TCIA(T/C)IGTIA(A/G)ICCIGG ICC(A/G)AA-3′. Following the conditions described in the paper [14], nested PCR was carried out with the primer sets of 112S and 380 A, and then with 174S and 368 A, to amplify a 571-bp DNA fragment. For the PCR, 30 cycles of reactions (94 °C for 0.5 min, 42 °C for 0.5 min and 72 °C for 1 min) were performed each time with a 10-min final extension. The gel-purified PCR product was ligated into pT7Blue T-Vector (Novagen) and sequenced.

3′ and 5′ RACE

For the 3′-end amplification, two specific primers; 217S = 5′-TTGGCGCAGATCCAGACCTA-3′ and 225S = 5′-GTTGAGAGACCCATATTCGA-3′, were designed based on the obtained core sequence. First RT-PCR was carried out with the primer set of 217S and RACE32, and the second PCR with 225S and RACE32, to amplify a 638-bp DNA fragment. For the PCR, 30 cycles of reactions (94 °C for 1 min, 55 °C for 1 min and 72 °C for 1 min) were performed each time. The nucleotide sequence of the PCR product was determined as described above.

The 5′ end amplification was carried out using the 5′ RACE system kit (BRL) and two specific primers; 188A = 5′-TCGTCATCTCGGAGCAAACG-3′ and 180A = 5′-ATGAGAACACGAGCTCCCTT-3′, based on the obtained core sequence. Thus, template cDNA was prepared by reverse transcription of the R. palmatum RNA using the antisense 188A primer according to manufacturer's instruction, and poly(A) tail added by terminal deoxynucleotidyl transferase. First PCR was carried out with the primer set of 188A and RACE32, and the second PCR with 180A and RACE32, to amplify a 666-bp DNA fragment. The PCR conditions were the same as described above for the 3′ RACE.

Cloning and expression of full-length cDNA

A full-length cDNA was obtained using N-terminal and C-terminal PCR primers; 5′-TCTATCGGATCCATGGC AACTGAGGAGATG-3′ (sense, the BamHI site is underlined) and 5′-TGACCGGTCGACGCTAATTACGGGCATACT-3′ (antisense, the SalI site is underlined). The amplified DNA was digested with BamHI/SalI, and cloned into the BamHI/SalI site of pET-22b(+) (Novagen). Thus, the recombinant enzyme contains an additional hexahistidine tag at the C-terminal. After confirmation of the sequence, the plasmid was transformed into E. coli BL21(DE3)pLysS. The cells harboring the plasmid were cultured to a D600 of 0.6 in Luria–Bertani medium containing 100 µg·mL−1 of ampicillin at 30 °C. Then, 0.4 mm isopropyl thio-β-d-galactoside was added to induce protein expression, and the culture was incubated further at 30 °C for 16 h.

Enzyme purification

The E. coli cells were harvested by centrifugation and resuspended in 50 mm potassium phosphate buffer, pH 8.0, containing 0.1 m NaCl. Cell lysis was carried out by the freeze-thaw method, and centrifuged at 15 000 g for 60 min. The supernatant was passed through a column of Pro-Bond™ resin (Invitrogen) which contained Ni2+ as an affinity ligand. After washing with 50 mm potassium phosphate buffer, pH 7.9, containing 0.5 m NaCl and 40 mm imidazole, the recombinant BAS was finally eluted with 15 mm potassium phosphate buffer, pH 7.5, containing 10% glycerol and 500 mm imidazole. Protein concentration was determined by the Bradford method (Protein Assay, Bio-Rad) with bovine serum albumin as standard.

Finally, in order to determine subunit composition, the purified enzyme was applied to HPLC gel filtration column (TSK-gel G3000SW, 7.5 × 600 mm, TOSOH), which was eluted with 0.1 m potassium phosphate buffer, pH 6.8, containing 10% glycerol and 0.2 m KCl at a flow rate of 1.0 mL·min−1. The following standard molecular mass markers were used: β-amylase (200 kDa), alcohol dehydrogenase (150 kDa), bovine albumin (66 kDa), carbonic anhydrase (29 kDa), and cytochrome c (12.4 kDa).

Enzyme reaction

The standard reaction mixture contained 27 nmol of 4-coumaroyl-CoA (or other CoA ester), 54 nmol of malonyl-CoA, and 105 pmol of the purified recombinant BAS in a final volume of 500 µL of 100 mm potassium phosphate buffer, pH 8.0, containing 1 mm EDTA. Incubations were carried out at 30 °C for 1 h, and stopped by adding 50 µL of 20% HCl. The products were then extracted with 1000 µL of ethyl acetate, and concentrated by N2 flow. The residue was dissolved in MeOH containing 0.1% trifluoroacetic acid, and analyzed by reverse-phase HPLC and LC-ESIMS as described below. For large-scale enzyme reaction, 15 mg of purified recombinant enzyme was incubated with 4-coumaroyl-CoA (10.0 mg, 11.0 µmol) and malonyl-CoA (10.0 mg, 11.7 µmol) in 500 mL of 100 mm phosphate buffer, pH 8.0, containing 1 mm EDTA, at 30 °C for 3 h. The reaction was quenched by addition of 20% HCl (50 mL), and extracted with ethyl acetate (500 mL × 2). After HPLC separation, pure benzalacetone; 4-(4-hydroxyphenyl)but-3-en-2-one (3) (1.5 mg, 84% yield from 10.0 mg of 1) was obtained. Spectroscopic data for the enzyme reaction product were as follows: LC-ESIMS: Rt = 18.2 min, m/z 163 [M + H]+. UV: λmax 329 nm 1H NMR (400 MHz, CD3OD): δ 7.59 (1H, d, J = 16.0 Hz), 7.50 (2H, d, J = 8.8 Hz), 6.81 (2H, d, J = 8.8 Hz), 6.62 (1H, d, J = 16.0 Hz), 2.34 (3H, s).

HPLC and LC-ESIMS

The enzyme reaction products were separated by reverse-phase HPLC (JASCO 880) on a TSK-gel ODS-80Ts column (4.6 × 150 mm, TOSOH) with a flow rate of 0.8 mL·min−1[12,13]. Gradient elution was performed with H2O and MeOH, both containing 0.1% trifluoroacetic acid: 0–5 min, 30% MeOH; 5–17 min, linear gradient from 30 to 60% MeOH; 17–25 min, 60% MeOH; 25–27 min, linear gradient from 60 to 70% MeOH. Elutions were monitored by a multichannel UV detector (MULTI 340, JASCO) at 290 nm, 330 nm and 360 nm. UV spectra (198–400 nm) were recorded every 0.4 s.

On-line LC-ESIMS spectra were measured with a Hewlett-Packard HPLC 1100 series (Wilmington, DE) coupled to a Finnigan MAT LCQ ion trap mass spectrometer (San Jose, CA, USA) fitted with an ESI source. HPLC separations were carried out under the same conditions as described above. The ESI capillary temperature and capillary voltage were 275 °C and 3.0 V, respectively. The tube lens offset was set at 20.0 V. All spectra were obtained in both negative and positive mode; over a mass range of m/z 120–350, at a range of one scan every 2 s. The collision gas was helium, and the relative collision energy scale was set at 30.0% (1.5 eV).

Phylogenetic tree

A total of 32 amino-acid sequences of CHS-superfamily enzymes were aligned and phylogenetic tree was developed with the clustal w (v. 1.8) program (DNA Data Bank of Japan, http://www.ddbj.nig.ac.jp) [15]. Bootstrap resampling (1000 bootstrap, 111 seed) of the original data was used as a pseudo-empirical test of the reliability of the tree topology [16,17]. The tree was constructed by use of the majority rule and strict consensus algorithm implanted in phylip[18]. The accession numbers of GenBank sequence data bank used in the analysis are as follows; Arabidopsis thaliana CHS (AF112086), Arachis hypogaea STS (L00952), Camellia sinensis CHS (D26593), Gerbera hybrida CHS (Z38096), G. hybrida 2PS (Z38097), Glycine max CHS (X53958), Humulus lupulus CHS (AJ304877), H. lupulus VS (AB047593), Hydrangea macrophylla CHS (AB011467), H. macrophylla CTAS (AB011468), Ipomoea purpurea CHS-A (U15946), I. purpurea CHS-B (U15947), I. purpurea CHS-D (AB004905), I. purpurea CHS-E (AB027534), Medicago sativa CHS (L02902), Oryza sativa CHS (AB000801), Petunia hybrida CHS (X04080), Phalaenopsis sp. bibenzyl synthase (BBS) (X79903), Phaseolus vulgaris CHS (X06411), Pinus strobus CHS (AJ004800), P. strobus STS (Z46915), Pinus sylvestris CHS (X60754), P. sylvestris STS (S50350), Pisum sativum CHS (X63333), Pueraria lobata CHS (D10223), Ruta gravenolens acridone synthase (ACS) (Z34088), R. graveolens CHS (AJ297789), Vitis vinifera CHS (X75969), V. vinifera STS (S63221), Zea mays CHS (X60205), Streptomyces griseus red-brown pigment producing enzyme (RppA) (AB018074). The β-ketoacyl carrier protein synthase III (FABH) of E. coli (M96793) [19] was employed as an outgroup as described previously [20].

Results and discussion

A cDNA encoding BAS was cloned and sequenced from young leaves of Rheum palmatum by RT-PCR amplification using inosine-containing degenerate oligonucleotide primers based on the highly conserved sequences of known CHS enzymes [14]. The terminal sequences of cDNA were obtained by 3′ and 5′ RACE method. A 1437-bp full-length cDNA contained a 124-bp 5′ noncoding region, a 1152-bp ORF encoding a 42 223 kDa protein of 384 amino acids, and 161-bp of 3′ noncoding region. The deduced amino-acid sequence of R. palmatum BAS showed 60–75% identity to those of CHS-related enzymes from other plant species (Fig. 1). One of the most characteristic features of R. palmatum BAS is that it lacks the active-site Phe215 (numbering in CHS) residue conserved in all the known CHS-related enzymes [5]. On the other hand, the catalytic cysteine-histidine dyad (Cys164–His303) in CHS [21,22] is well conserved in R. palmatum BAS.

Figure 1.

Comparison of deduced amino-acid sequences of CHS-related enzymes. M.s CHS, Medicago sativa chalcone synthase; A.h STS, Arachis hypogaea stilbene synthase; H.m CTAS, Hydrangea macrophylla 4-coumaroyltriacetic acid synthase; G.h 2PS, Gerbera hybrida 2-pyrone synthase; R.g ACS, Ruta gravenolens acridone synthase; R.p BAS, Rheum palmatum benzalacetone synthase. The active-site amino-acid residues conserved in all the known CHS-related enzymes (Cys164, Phe215, His303, and Asn336, numbering in CHS) were marked with ◆.

The phylogenetic tree (Fig. 2) showed that the members of the CHS-superfamily enzymes did not form a species-specific cluster, but instead grouped into subfamilies according to their enzymatic function. Interestingly, the cloned BAS from R. palmatum (Polygonaceae) forms a separate cluster with other nonchalcone forming enzymes including valerophenone synthase (VS) from Humulus lupulus (Moraceae) [23,24], 4-coumaroyltriacetic acid lactone synthase (CTAS) from Hydrangea macrophylla (Saxifragaceae) [25], 2-pyrone synthase (2PS) from Gerbera hybrida (Asteraceae) [4], acridone synthase (ACS) from Ruta gravenolens (Rutaceae) [26], and two other CHS-like enzymes with unknown function from Ipomoea purpurea (Convolvulaceae) [27]. These nonchalcone forming polyketide synthases involved in the biosynthesis of species-specific secondary metabolites appear to be evolutionary more primitive than CHS for the flavonoids. On the other hand, stilbene synthase (STS), grouping with the CHS from the same or related plants, has been proposed to have evolved independently several times from CHS [20].

Figure 2.

Phylogenetic tree of the CHS-superfamily enzymes. Multiple sequence alignment performed by clustal w (v. 1.8). The β-ketoacyl carrier protein synthase III (FABH) of E. coli served as outgroup. The indicated scale represents 0.1 amino-acid substitutions per site. Numbers at the forks indicate how often the group to the right appeared among 1000 bootstrap replicates. Abbreviations: CHS, chalcone synthase; STS, stilbene synthase; 2PS, 2-pyrone synthase; ACS, acridone synthase; BBS, bibenzyl synthase; CTAS, 4-coumaroyltriacetic acid synthase; BAS, benzalacetone synthase; VS, valerophenone synthase; RppA, red-brown pigment producing enzyme.

As in the case of other CHS-related enzymes, recombinant BAS was functionally expressed in E. coli with an additional hexahistidine tag at the C-terminal. Purification by Ni-chelate affinity column chromatography afforded ≈ 3 mg of homogeneous recombinant BAS from 1 g of E. coli cell pellet. The purified enzyme gave a single band with molecular mass of 42 kDa on SDS/PAGE (Fig. 3A), while the native BAS appeared to be a homodimer as it had apparent molecular mass of ≈ 90 kDa as determined by HPLC gel-filtration (Fig. 3B). In the presence of 10% glycerol, the enzyme could be stored at −80 °C without significant loss of activities for at least 1 month.

Figure 3.

SDS/PAGE and HPLC gels. (A) SDS/PAGE (10%) gel stained with Coomassie Brilliant Blue. Lane 1, E. coli cell-free extract; Lane 2, BAS purified by Ni-chelate affinity column, giving a single band with molecular mass of 42 kDa. (B) Calibration curve obtained with HPLC gel filtration analysis (TSK-gel G3000SW, TOSOH). The native enzyme appeared to be a homodimer as it had apparent molecular mass of ≈ 90 kDa.

When incubated with 4-coumaroyl-CoA and malonyl-CoA as substrates under the standard assay condition [12,13], the recombinant BAS efficiently afforded benzalacetone; 4-(4-hydroxyphenyl)but-3-en-2-one (3) as a single product. The enzyme showed the Km = 10.0 µm and kcat = 1.79 min−1 for 4-coumaroyl-CoA, with a pH optimum within a range of 8.0–8.8. On the other hand, formation of pyrone derivatives; 4-coumaroyltriacetic acid lactone and bis-noryangonin, the common byproducts of CHS enzyme reactions in vitro[25], were not detected in the enzyme reaction mixture. Further, in contrast with CHS that showed broad substrate specificity toward aliphatic CoA esters [12,13], BAS did not accept hexanoyl-CoA, isobutyryl-CoA, isovaleryl-CoA, and acetyl-CoA as a substrate. Moreover, enzymatic conversion of 3-(4-hydroxyphenyl)propionyl-CoA to 4-(4-hydroxyphenyl)butan-2-one (raspberry ketone) (5) was not detected.

Confirmation of the structure of the enzyme reaction product was obtained as follows. First, the LC-ESIMS spectrum of the product gave a parent ion peak [M + H]+ at m/z 163, and a UV spectrum (λmax 329 nm). Second, the 1H NMR spectrum of the product obtained from a large scale enzyme reaction (84% yield from 10 mg of 4-coumaroyl-CoA) showed A2B2-type (δ 7.50, 2H, d, J = 8.8 Hz, and 6.81, 2H, d, J = 8.8 Hz) aromatic signals together with a pair of trans-coupled α,β-unsaturated olefinic protons (δ 7.59 and 6.62, each 1H, d, J = 16.0 Hz) and α-methyl protons (2.34, 3H, s), supporting the structure. Finally, the spectroscopic data (1H NMR, UV, MS, and LC-MS) of the reaction product was identical with those of the commercially available authentic compound.

This is the first report of cloning and functional expression of BAS that catalyzes a one-step decarboxylative condensation of 4-coumaroyl-CoA with malonyl-CoA to produce the C6–C4 skeleton of phenylbutenone (Scheme 1). As benzalacetone was previously identified as an early released derailed byproduct of CHS that catalyzes a sequential condensation with three molecules of malonyl-CoA [2], the decarboxylation of the chemically unstable diketide intermediate may not require such an elaborate enzymatic assistance.

As described above, R. palmatum BAS lacks the active-site Phe215 (numbering in CHS) residue which has been proposed to help orient substrates and intermediates during the sequential condensation with malonyl-CoA in CHS [6]. Therefore, the replacement of the Phe residue with Leu in R. palmatum BAS may thus block further chain elongation of the diketide intermediate. Indeed, it has been reported that formation of increased proportion of truncated reaction products were observed when the Phe215 of M. sativa CHS was substituted with other amino acid [6]. In order to test this hypothesis, site-directed mutagenesis experiments of BAS enzyme are now in progress in our laboratory.

BAS catalyzes the initial key reaction step in the biosynthesis of phenylbutanoids in rhubarb, which contains considerable amount of phenylbutanone glucosides such as pharmaceutically important lindleyin (6) (Fig. 4), the active principle for the anti-inflammatory action of the medicinal plant [10,11]. Besides phenylbutanones in rhubarb, benzalacetone is known to be an immediate precursor of raspberry ketone; 4-(4-hydroxyphenyl)butan-2-one (5), the characteristic aroma of raspberry fruits (Rubus ideaeus, Rosaceae) [28,29]. As described above, as R. palmatum BAS did not accept 3-(4-hydroxyphenyl)propionyl-CoA as a substrate, it is likely that the double bond of benzalacetone (3) is later reduced by a reductase in the biosynthesis of lindleyin as in the case of raspberry ketone. Furthermore, possible involvement of the phenylbutenone intermediate has been also proposed for the biosynthesis of curcumin (7) and [6]-gingerol (8), the pungent principles of turmeric (Curcuma longa, Zingiberaceae) and ginger (Zingiber officinale, Zingiberaceae), respectively [2]. In fact [6]-gingerol has been shown to be biosynthesized from ferulic acid, malonate, and hexanoyl moieties by precursor feeding experiments in Z. officinale[30]. Similar biosynthetic pathway would be also likely for the construction of the C6-C4 moiety of curcumin (Fig. 4).

Figure 4.

Structures of the diketide (phenylbutenone) derived natural products. The C6–C4 moieties are shown by bold lines.

In summary, we have for the first time succeeded in the cloning and characterization of BAS, a novel plant-specific polyketide synthase that plays a crucial role in the biosynthesis of phenylbutanoids in higher plants. Further characterization of the enzyme as well as crystallization trials will be reported in due course.

Acknowledgements

We thank Dr T. Shibata (NIHS, Experimental Station for Medicinal Plants, Holekaido), Dr T. Miyahara and Mr H. Miyamoto (University of Shizuoka) for Rheum palmatum plant used in this study. We are also indebded to Ms. A. Nomura (University of Shizuoka) for technical assistance. This work was in part supported by a Grant-in-Aid for Scientific Research [(B) 13480188] from the Ministry of Education, Sciences, Sports and Culture Japan.

Footnotes

  1. Enzymes: benzalacetone synthase (EC 2.3.1.-); chalcone synthase (EC 2.3.1.74); stilbene synthase (EC 2.3.1.95).

  2. Note: the novel nucleotide sequence data published here have been deposited in the EMBL/DDBJ/GenBank databases under accession number AF326911.

Ancillary

Advertisement