Identification of b-amyrin and sophoradiol 24-hydroxylase by expressed sequence tag mining and functional expression assay Masaaki Shibuya 1 , Masaki Hoshino 1 , Yuji Katsube 1 , Hiroaki Hayashi 2 , Tetsuo Kushiro 1 , and Yutaka Ebizuka 1 1 Graduate School of Pharmaceutical Sciences, The University of Tokyo, Japan 2 Gifu Pharmaceutical University, Japan Triterpene saponins are glycosides of cyclic C30 terpe- nes and include a number of active constituents of medicinal plants, as exemplified by glycyrrhizin in Glycyrrhiza glabra, ginsenosides in Panax ginseng, sai- kosaponins in Bupleurum falcatum, etc. [1]. Extensive pharmacological studies on triterpene saponins from medicinal plants revealed their important biological activities. For example, ginsenosides and ⁄ or their agly- cones show various activities including central nervous system-stimulating (or -suppressing) activity, and anti- cancer activity, etc. [2]. Their distribution is not limited to medicinal plants. They are rather ubiquitously distri- buted in the plant kingdom. Legumes such as Glycine max, Pisum sativum, and Medicago sativa are known as rich sources of triterpene saponins [1]. Recently, avicins, saponins isolated from the Australian desert tree Acacia victoriae (Leguminosae), have been repor- ted to induce apoptosis in tumor cells (Jurkat human T cell line) by affecting mitochondrial function and are promising anticancer agents [3]. Despite these promis- ing activities for medicinal use, great difficulties in obtaining sufficient quantities of these triterpene Keywords b-amyrin 24-hydroxylase; CYP93E1; Glycine max; P450; sophoradiol 24-hydroxylase Correspondence Y. Ebizuka, Graduate School of Pharmaceutical Sciences, The University of Tokyo, Hongo Bunkyo-ku, Tokyo 113–0033, Japan Fax: +81 3 5841 4744 Tel: +81 3 5841 4740 E-mail: yebiz@mol.f.u-tokyo.ac.jp (Received 28 October 2005, revised 12 December 2005, accepted 23 December 2005) doi:10.1111/j.1742-4658.2006.05120.x Triterpenes exhibit a wide range of structural diversity produced by a sequence of biosynthetic reactions. Cyclization of oxidosqualene is the ini- tial origin of structural diversity of skeletons in their biosynthesis, and sub- sequent regio- and stereospecific hydroxylation of the triterpene skeleton produces further structural diversity. The enzymes responsible for this hydroxylation were thought to be cytochrome P450-dependent mono- oxygenase, although their cloning has not been reported. To mine these hy- droxylases from cytochrome P450 genes, five genes (CYP71D8, CYP82A2, CYP82A3, CYP82A4 and CYP93E1) reported to be elicitor-inducible genes in Glycine max expressed sequence tags (EST), were amplified by PCR, and screened for their ability to hydroxylate triterpenes (b-amyrin or sophora- diol) by heterologous expression in the yeast Saccharomyces cerevisiae. Among them, CYP93E1 transformant showed hydroxylating activity on both substrates. The products were identified as olean-12-ene-3 b,24-diol and soyasapogenol B, respectively, by GC-MS. Co-expression of CYP93E1 and b-amyrin synthase in S. cerevisiae yielded olean-12-ene-3b,24-diol. This is the first identification of triterpene hydroxylase cDNA from any plant species. Successful identification of a b-amyrin and sophoradiol 24-hydroxy- lase from the inducible family of cytochrome P450 genes suggests that other triterpene hydroxylases belong to this family. In addition, substrate specific- ity with the obtained P450 hydroxylase indicates the two possible biosyn- thetic routes from triterpene-monool to triterpene-triol. Abbreviations EST, expressed sequence tags. 948 FEBS Journal 273 (2006) 948–959 ª 2006 The Authors Journal compilation ª 2006 FEBS saponins from natural sources and ⁄ or by chemical syn- thesis prevent them from being used in clinical trials. If triterpene saponins are to be developed as therapeutic agents, the problem of supply must be resolved. As the practical supply of triterpene saponins by chemical syn- thesis is difficult both in terms of quantity and cost, biological production has been considered to be an alternative method to obtain them in sufficient quanti- ties. Production by plant cell or hairy root cultures as a source of triterpene saponins has been attempted for decades, but without practical success so far [4–6]. In order to improve the biological production method, a detailed understanding of the biosynthesis of triterpene saponins is required, including the enzymes catalyzing the sequence of reactions and the genes encoding these enzymes. The biosynthesis of triterpene saponins involves the initial cyclization of 2,3-oxidosqualene, a common pre- cursor of all the sterol and triterpene biosyntheses, into various cyclic triterpenes, followed by oxidative modifi- cation of these carbon skeletons and transfer of the sugar moiety (Fig. 1). More than 80 different types of skeleton are generated at the cyclization step [7]. Successful cloning of oxidosqualene cyclases in recent years has disclosed the molecular origin of the skeletal diversity of triterpenes [8–24]. It is notable that multi- functional triterpene synthases yielding more than two products exist in plants in addition to single-product- specific triterpene synthase and contribute to the skeletal diversity of triterpenes. Subsequent regio- and stereospecific hydroxylations of the skeleton produce further structural diversity. In contrast to the rapid progress in the research on skeletal formation, little is known about the subsequent oxidation and sugar transfer reactions. Enzymological studies indicated that oxidation of inert methylene and methyl groups of tri- terpene skeletons is mediated by cytochrome P450 monooxygenase (P450) [25,26]. However, no gene encoding the triterpene-hydroxylating P450 has yet been reported. In general, purification of microsomal P450 enzymes from higher plants for amino-acid sequencing is diffi- cult because a number of P450 exist even in a single plant species. For example, 272 P450 genes were found in the Arabidopsis thaliana genome [27,28], whose products may be very similar in physical properties and therefore be difficult to separate from each other. Therefore, the reverse genetic method is not practical for cloning P450 involved in triterpene biosynthesis. An alternative approach by functional analysis of heterologously expressed P450 based on genomic sequences or expressed sequence tags (EST) appeared HO β-Amyrin HO Sophoradiol OH HO Olean-12-ene-3β,24-diol HO HO Soyasapogenol B OH HO RO R= -GlcA-Gal, Soyasaponin III etc. OH HO O 2,3-Oxidosqualene 3 24 22 Fig. 1. Biosynthesis of soyasaponin. M. Shibuya et al. CYP93E1 FEBS Journal 273 (2006) 948–959 ª 2006 The Authors Journal compilation ª 2006 FEBS 949 promising, which requires information on the reaction catalyzed (what is substrate, and what is product, etc.). However, phytochemical information on A. thaliana metabolites is still lacking [29], and the details of tri- terpene metabolism need further study. In a recent review, some 50 terpenes were listed as A. thaliana metabolites [29]. The 15 triterpenes in the list, however, were not isolated from the plant itself, but they were identified as the products of heterologously expressed triterpene synthases [10,14,17,20,24]. Although the presence of some triterpenes (lupeol, b-amyrin, etc.) was confirmed in the whole plant by GC-MS analysis (data not shown), none of hydroxylated triterpenes and their glycosides has been identified. The lack of understanding of triterpene metabolism in A. thaliana makes it difficult to identify the function of each P450 from the A. thaliana genome, even though there are not more than 272. In this study, an alternative approach based on EST information was taken to clone triterpene-hydroxylat- ing P450 from soybean (Glycine max). G. max is one of the most important crops in the world, and the accumulated EST information revealed the existence of more than 200 types of P450 genes (TIGR Soy- bean Gene Index, http://www.tigr.org/tigr-scripts/tgi/ T_index.cgi?species ¼ soybean). Fourteen of them have been obtained in full length, including cinnamic acid 4-hydroxylase [30], isoflavone synthase [31], di- hydroxypterocarpan 6a-hydroxylase [32], and flavonoid 6-hydroxylase [33]. G. max produces triterpene saponins known as soyasaponins. More than 10 types of soyasaponin have been isolated, all of which are glycosides of oleanene triterpene [34]. Their aglycone structures are restricted to two, soyasapogenols A and B (Fig. 2). Soyasapoge- nol A has four hydroxyl groups at C-3, C-21, C-22, and C-24, whereas soyasapogenol B has three at C-3, C-22, and C-24 [34]. In addition to these two agly- cones, soyasapogenols C and D (dehydrated or oxi- dized soyasapogenol B at the C-22 hydroxyl group, respectively) were reported [35]. However, saponins with soyasapogenols C and D as aglycone have not been isolated. Therefore, they are considered to be artifacts during the isolation procedure [35]. This evi- dence reduces the potential number of triterpene hydroxylases responsible for soyasaponin biosynthesis. Two possible routes from b-amyrin to soyasapogenol B shown in Fig. 1 indicate the presence of four types of hydroxylase. Biosynthetic route for soyasapogenol A is not as simple as that for soyasapogenol B, and the presence of additional hydroxylases must be considered. Fortunately, the aglycone of the major soyasaponins is soyasapogenol B, and glycosides of soyasapogenol A are minor saponins in G. max. This abundance ratio strongly suggests high level of tran- scription of 22- and 24-hydroxylase genes. Soyasaponin biosynthesis in cell suspension cultures of G. glabra is reported to be induced by methyl jasmonate [36]. In this study, functional analysis of elicitor-inducible P450s already isolated as EST from G. max, was carried out by heterologous expression in yeast. Results and Discussion P450 genes induced by elicitors in the cell culture of G. max Accumulation isoflavones phytoalexin glyceollins, in the seedling infected with Phytophythora sojae or in the cell cultures treated with a glucan elcicitor from this oomycete have been reported [37]. Their accumu- lation was caused by transcriptional activation of their biosynthetic genes [37]. Such activation has also been reported in other legumes, Phaseolus vulgaris [38] and Glycyrrhiza echinata [39]. Not only isoflavo- noids, but also soyasaponins are induced by methyl jasmonate in the cell cultures of G. glabra. The triter- pene biosynthetic genes, namely squalene synthase and b-amyrin synthase, are transcriptionally activated [36], suggesting that triterpene hydroxylase genes might also be elicitor inducible. Nine full-length cyto- chrome P450 genes were isolated from G. max as the genes inducible by the yeast extract elicitor [30], three of which were shown to encode cinnamic acid 4-hy- droxylase (CYP73A11) [30], flavonoid 6-hydroxylase HO Soyasapogenol B OH HO HO Soyasapogenol A OH HO OH HO Soyasapogenol C HO HO Soyasapogenol E HO O 3 24 21 22 Fig. 2. Sapogenols in Glycine max. CYP93E1 M. Shibuya et al. 950 FEBS Journal 273 (2006) 948–959 ª 2006 The Authors Journal compilation ª 2006 FEBS (CYP71D9) [33], and dihydroxypterocarpan 6a-hy- droxylase (CYP93A1) [32], leaving the remaining clones (CYP71A9, CYP71D8, CYP82A2, -3, and -4, CYP93A3) unidentified. As none of CYP82A sub- family member has been identified for their enzyme function, they are chosen in this study with the expec- tation of detecting triterpene hydroxylase activity. Four genes (CYP82A2, -3, and -4 together with CYP71D8) were cloned using RT-PCR based on the reported sequences [30] and RNA prepared from yeast extract-treated cell cultures of G. max. (data not shown). The genes obtained were ligated into expres- sion vector pYES2 (Invitrogen) and expressed in S. cerevisiae strain INVSC2 (Invitrogen). However, in vitro assay of the cell-free extracts of these trans- formants with 14 C-labeled b-amyrin as a substrate did not show any hydroxylase activity, and in vivo assay by feeding the same substrate to the culture of each transformant did not yield any detectable hydroxylat- ed product (data not shown). Function of genes belonging to the CYP93 family The CYP93 family consists of five subfamilies inclu- ding several members with identified functions. Elici- tor-inducible CYP93A1, CYP93B1, and CYP93C2 encode dihydroxypterocarpan 6a-hydroxylase [32] (2S)- flavanone-2-hydroxylase [40], and isoflavone synthase [31], respectively. As the majority of CYP93 family members are flavonoid biosynthesis-related genes, other members of this family, including CYP93D1 (found in the A. thaliana genome sequence, EMBL and GenBank accession number AB010697; protein id BAB11147.1) and CYP93E1 (inducible by infection with P. sojae and obtained together with CYP93C2 from G. max), were considered to be flavonoid bio- synthesis-related genes. Despite extensive trials, failure to detect isoflavone synthase activity in CYP93E1 suggests its monooxygenase activity towards other substrates [31]. As the production of not only flavonoid but also of triterpene saponins is induced by elicitation as mentioned above, triterpene hydroxylase is one possible function of CYP93E1. Feeding of b-amyrin and sophoradiol to S. cerevisiae transformed with pESC-CYP93E1 As reported for brassinosteroid-6-oxidase (CYP85A1) [41,42] and taxane 10b-hydroxylase [44], enzyme activ- ities of heterologously expressed P450s were demon- strated by feeding the substrate to the transformed yeast. To examine hydroxylating activity toward b-amyrin and sophoradiol, possible intermediates in soyasapogenol B biosynthesis (Fig. 1), they were administered to the transformant (INVSC2 ⁄ pESC- CYP93E1) after induction of the GAL1 promoter. Cells were harvested, disrupted by boiling with 20% KOH ⁄ 50% aqueous methanol solution, and extracted with hexane. After acetylation, products were analyzed with GC-MS. Expecting the formation of olean-3b,24- hydroxy-12-ene from b-amyrin, and soyasapogenol B from sophoradiol, GC was monitored by the intensity of the respective base peaks (m ⁄ z ¼ 218 or m ⁄ z ¼ 216), retro-Diels–Alder fragments at the C-ring, as shown in Fig. 3. The b-amyrin feeding experiments generate a peak with the same retention time (15.4 min) (entry B) as that of authentic sample (entry A). The MS fragmentation pattern of this peak (B in Fig. 4) was completely identical to that of the authen- tic olean-3b,24-diacetoxy-12-ene (A in Fig. 4). This peak was not observed in the negative controls (C: without substrate, D: no induction of GAL1 pro- moter, E: transformant with void vector). These results confirm that CYP93E1 encodes b-amyrin 24-hydroxy- lase. 24-Hydroxylase activity was also observed with sophoradiol, as shown in Fig. 5, indicating that R 1 R 1 = OAc, R 2 = R 3 = H : O -Ac-β-amyrin, m/z = 468 R 1 = R 2 = OAc, R 3 = H : 3β,24-diacetoxyolean-12-ene, m/z = 526 R 1 = R 3 = OAc, R 2 = H : di - O -Ac-sophoradiol, m/z = 526 R 1 = R 2 = R 3 = OAc : tri - O -Ac-soyasapogenol B, m/z = 584 R 3 R 2 AB C D E R 3 D E C D E C R 3 = H, m/z = 218 R 3 = OAc, m/z = 276 m/z = 216 R 3 = OAc - AcOH Fig. 3. Base peaks of oleanene-type triterpenes due to retro-Diels-Alder fragmentation in GC-MS analysis. M. Shibuya et al. CYP93E1 FEBS Journal 273 (2006) 948–959 ª 2006 The Authors Journal compilation ª 2006 FEBS 951 CYP93E1 has 24-hydroxylase activities for both b-am- yrin and sophoradiol substrates. b-amyrin and sophoradiol hydroxylase activities in the cell-free extract of S. cerevisiae harboring pESC-CYP93E1 To demonstrate in vitro activity, a cell-free extract was prepared from the transformed yeast. b-Amyrin or sophoradiol was incubated with the extract. After extraction with hexane and acetylation, the products were analyzed with GC-MS. As shown in Fig. 6, when b-amyrin was incubated, a peak at 15.4 min corres- ponding to olean-3b,24-diacetoxy-12-ene was found in the complete assay mixture (entry B), but not in the negative controls (C: without substrate, D: boiled cell-free extract, E: no induction of GAL1 promoter, E: void vector). The MS fragmentation pattern was also identical to that of the authentic olean-3b,24-di- acetoxy-12-ene, except for the presence of several back- ground peaks (the amount of authentic sample was adjusted to equalize the height of both peaks in GC). When the extract was incubated with sophoradiol, a peak was found at 19.5 min in the complete assay mix- ture (entry B), as shown in Fig. 7. The major peaks (m ⁄ z ¼ 201 and 216) in MS fragmentation were identi- cal to those of the authentic tri-O-acetyl-soyasapogenol B. To the best of our knowledge, this is the first demonstration of in vitro hydroxylase activity for a triterpene substrate in a heterologous expression sys- tem, although activities of several diterpene hydroxy- lases were demonstrated in vitro [45–48]. Fig. 4. GC-MS analysis of the extract from transformant fed with b-amyrin. GC was monitored based on intensity of the base peak (m ⁄ z 218), which was a fragment of the D,E-ring moiety due to retro-Diels–Alder fragmentation at the C-ring in olean-3b,24-diacetoxy-12-ene. Entry A in the upper panel: 20 pmol of authentic olean-3b,24-diacetoxy-12-ene; B: complete conditions as described in Experimental proce- dures; C: without feeding with b-amyrin; D: without induction of the GAL1 promoter; E: transformant with void vector. MS fragmentations of entries A and B are shown in the lower panel. CYP93E1 M. Shibuya et al. 952 FEBS Journal 273 (2006) 948–959 ª 2006 The Authors Journal compilation ª 2006 FEBS Co-expression of PSY (b-amyrin synthase) and CYP93E1 in lanosterol synthase-deficient S. cerevisiae strain GIL77 As shown in Fig. 3, there is no doubt that one hydro- xyl group is introduced into the A- or B-ring of b-amyrin, but the present results do not exclude the possibility that the product has a hydroxyl group at positions other than C-24, as there is no information on the retention time and MS fragmentation of such compounds. Feeding of b-amyrin (50 nmol) to the cul- tures (20 mL) of the yeast transformed with pESC- CYP93E1 yielded about 0.2 nmol of hydroxylated b-amyrin (Fig. 4). Based on this conversion ratio, the yield of the product after b-amyrin (2.5 mmol, 1 mg) feeding in 1-L cultures was estimated to be 10 nmol (4.5 lg). If the efficiency of uptake of b-amyrin from media is one of the reasons for the low conversion, it would be improved by in situ supply of the substrate through coexpression with b-amyrin synthase. In our previous studies, more than 10 mg of b-amyrin was produced by 1-L culture of yeast transformant with the plasmid harboring P. sativum b-amyrin synthase gene (PSY) [13]. CYP93E1 and PSY genes were sub- cloned into the S. cerevisiae expression vector pESC harboring two expression cassettes. Cells were harves- ted from 1-L of induced culture, lysed by boiling with 20% KOH ⁄ 50% aqueous methanol solution, and extracted with hexane. The product was purified on sil- ica gel column to yield 1.0 mg of product as crystals. Fig. 5. GC-MS analysis of the extract from the transformant fed with sophoradiol. GC was monitored based on the intensity of the base peak (m ⁄ z 216), which was a fragment of the D,E-ring moiety due to retro-Diels–Alder fragmentation at the C-ring in tri-O-acetyl-soyasapoge- nol B. Entry A in the upper panel: 20 pmol of authentic tri-O-acetyl-soyasapogenol B; B: complete conditions as described in Experimental procedures; C: without feeding with b-amyrin; D: without induction of the GAL1 promoter; E: transformant with void vector. MS fragmenta- tions of entries A and B are shown in the lower panel. M. Shibuya et al. CYP93E1 FEBS Journal 273 (2006) 948–959 ª 2006 The Authors Journal compilation ª 2006 FEBS 953 1D-NMR ( 1 H- and 13 C-NMR) spectra were completely identical to those reported for olean-12-ene-3b,24-diol [49], and correlations observed in 2D-NMR (HMQC, HMBC, and NOESY) further confirmed its identity (data not shown). The agylcone of the major soyasaponins in G. max is soyasapogenol B, which is biosynthesized via two hy- droxlyations at C-22 and C-24 of b-amyrin. In this study, CYP93E1 was demonstrated to hydroxylate the methyl group (C-24) of both b-amyrin and sophoradiol. This result indicates that CYP93E1 has substrate specif- icity for the 3-hydroxyolean-12-ene structure, and a hydroxyl group at C-22 was not recognized. Hydroxyla- tion only at C-24 methyl group points to very strict regiospecificity for hydroxylation. To further investigate substrate specificity, CYP93E1 was coexpressed with YUP8H12R.43, a multifunctional triterpene synthase (protein ID; AAC17070.1, BAC clone, EMBL and Gen- Bank accession number; AC002986) from A. thaliana producing lupeol, butyrospermol, tirucalla-7,21-dien- 3b-ol, taraxasterol, b-amyrin, w-taraxasterol, bauerenol, a-amyrin, and multiflorenol [14], in the same manner as above. No hydroxylated products other than olean- 3b,24-diacetoxy-12-ene were detected in GC-MS ana- lysis (data not shown). Among the nine triterpenes with different skeletons produced by A. thaliana YUP8H12R.43, only b-amyrin was hydroxylated, sug- gesting the strict substrate specificity of CYP93E1 for the 3-hydroxyolean-12-ene structure. The 24-hydroxylase activity of CYP93E1 for b-amy- rin and sophoradiol suggests that the biosynthesis of soyasaponin might form a metabolic grid, via Fig. 6. GC-MS analysis of in vitro reaction products with b-amyrin as a substrate. GC was monitored based on the intensity of the base peak (m ⁄ z 218) as described in the legend to Fig. 4. Entry A in the upper panel: authentic olean-3b,24-diacetoxy-12-ene (injected amount was not determined); B: complete conditions as described in Experimental procedures; C: removal of b-amyrin from complete conditions; D: using boiled cell-free extract; E: using cell-free extract prepared from the transformant with no induction of the GAL1 promoter; F: using cell-free extract prepared from the transformant with void vector. MS fragmentations of entries A and B are shown in the lower panel. CYP93E1 M. Shibuya et al. 954 FEBS Journal 273 (2006) 948–959 ª 2006 The Authors Journal compilation ª 2006 FEBS olean-12ene-3b,24-diol or sophoradiol from b-amyrin to soyasapogenol B (Fig. 1), and the hydroxylation of b-amyrin or olean-12-ene-3b,24-diol at the C-22 posi- tion is catalyzed by other P450, probably an enzyme similar to CYP93E1. A similar metabolic grid branch- ing at the hydroxylation of campestanol by two inde- pendent P450 (6-hydroxylase and 22-hydroxylase) and joining at the formation of castasterone by the same enzymes was proposed in brassinosteroid biosynthesis [42]. Not only glycosides but also triterpene aglycones show interesting biological activities. For example, soyasapogenol B has hepatoprotective activity [50], and oleanolic acid and ursolic acid show anti-inflam- matory and antitumor-promoting activities [51], etc. As the supply of triterpenes including oxygenated derivatives through organic synthesis is not practical, these compounds must be isolated from natural sources. Successful production of olean-12-ene-3b,24- diol by coexpression of b-amyrin synthase and triter- pene hydroxylase in S. cerevisiae in this study opened a way for the production of useful oxygenated triterpe- nes through fermentation. This methodology will be useful for the production of triterpene saponins after cloning of the sugar transferases. As all CYP93 family members thus far identified are flavonoid biosynthesis-related enzymes, it was unex- pected that CYP93E1 would encode b-amyrin and sophoradiol 24-hydroxylase. The identification of CYP93E1 as triterpene hydroxylase implies that the function of other members of the CYP93E subfamily are not necessarily related to flavonoid biosynthesis. Fig. 7. GC-MS analysis of in vitro reaction products with sophoradiol as a substrate. GC was monitored based on the intensity of the base peak (m ⁄ z 216) as described in the legend to Fig. 5. Entry A in the upper panel: authentic tri-O-acetyl-soyasapogenol B (injected amount was not determined); B: complete conditions as described in Experimental procedures; C: removal of b-amyrin from complete conditions; D: using boiled cell-free extract; E: using cell-free extract prepared from the transformant with no induction of the GAL1 promoter; F: using cell- free extract prepared from the transformant with void vector. MS fragmentations of entries A and B are shown in the lower panel. M. Shibuya et al. CYP93E1 FEBS Journal 273 (2006) 948–959 ª 2006 The Authors Journal compilation ª 2006 FEBS 955 Cloning of the full-length cDNA and functional analy- sis of several other EST clones belonging to this sub- family is in progress. On the other hand, sterol oxygenases thus far identified do not show high sequence homology and are assigned to different CYP families (obtusifoliol 14a-demethylase; CYP51 [52], brassinosteroid 6-oxidase; CYP85A1 [41,42], brassino- steroid 22a-hydroxylase; CYP90B1 [43]). Therefore, mining of triterpene hydroxylase from other members of the CYP family must be continued. Experimental procedures RNA extraction and amplification of CYP93E1 cDNA Soybean seeds (cultivar Wase–Edamame purchased from Atariya Nouen, Tokyo, Japan) were germinated in a growth cabinet under 16-h daylight and 8-h dark conditions at 25 °C. After 2 weeks, the leaves (4.7 g) were detached, immediately frozen in liquid nitrogen, and homogenized with a mortar and pestle. RNA was extracted with the phe- nol-chloroform method as reported previously [8], to give 98 lg of total RNA. A single-strand cDNA pool was pre- pared with reverse transcriptase (Superscript II, Invitrogen, Carlsbad, CA, USA) and 0.5 lg of oligo dT primer, RACE32 (5¢-GACTCGAGTCGACATCGATTTTTTTTT TTTTT-3¢) with dNTP (0.2 mm) in a volume of 20 lL fol- lowing the manufacturer’s recommended protocol. The resulting cDNA mixture was used as the template in subse- quent PCR. The open reading frame of CYP93E1 was amplified by PCR (initial denaturing for 5 min at 94 °C, 30 cycle of 94 °C for 1 min, 65 °C for 2 min, and 72 °C for 3 min, and a final extension reaction for 10 min at 72 °C) using Ex. Taq DNA polymerase (Takara Bio Inc, Shiga, Japan) with template (the cDNA pool described above) and primers (5¢-AAACACTAGTATGCTAGACATCAAAGG CTAC-3¢, and 5¢-TTCAATCGATTCAGGCAGCGAACG GAGTGAA-3¢), which were synthesized based on the reported sequences. The obtained clone was sequenced in both strands. This sequence has been submitted to the DDBJ sequence database and is available under accession number AB231332. Construction of S. cerevisiae expression vector pESC-CYP93E1 and S. cerevisiae transformant The amplified cDNA fragment was ligated into the restric- tion enzyme sites (SpeI and ClaI) of pESC-URA (Invitro- gen) after digestion with these enzymes. The plasmid obtained was designated pESC-CYP93E1. S. cerevisiae strain INVSC2 (Invitrogen) was transformed with pESC- CYP93E1 using a Frozen-EZ Yeast Transformation II kit (Zymo Research, California, USA). Feeding of b-amyrin and sophoradiol to the transformant with pESC-CYP93E1 The transformant with pESC-CYP93E1 was inoculated in 20 mL of synthetic complete medium [53] without uracil, containing hemin chloride (13 lgÆmL )1 ), and raffinose (2%) in place of glucose (SCR-U), and incubated at 30 °C for 1 day. Then, 1 mL of 40% galactose solution (final concen- tration 2%), 0.2 mL of hemin chloride solution (final concentration 13 lgÆmL )1 ), and the substrate (50 nmol of b-amyrin or sophoradiol, in methanol solution) were added. Cells were incubated under the same conditions for 1 day and then harvested by centrifugation at 500 g for 5 min. After the addition of 0.25 mL of 40% potassium hydroxide solution and 0.25 mL of methanol, the cell suspension was boiled for 5 min. Products were extracted three times with 0.2 mL of hexane and evaporated. Pyridine and acetic anhydride 0.01 mL each were added to the residue, and then the mixture was left to stand at room temperature overnight. The reaction was terminated by the addition of 0.1 mL each of methanol and water. The products were extracted three times with hexane (0.2 mL). After evapor- ation, the residue was dissolved in 0.02 mL of hexane, and 0.001 mL of hexane solution was used for GC-MS analysis using a Shimadzu (Kyoto, Japan) GCMS-QP2010 with a Restec Rtx-5MS glass capillary column (30 m in length, 0.25 mm in diameter, 0.25-lm film thickness) and He as a carrier gas (45 cmÆmin )1 ) on the program (held at 230 °C for 2 min, then temperature increased at the rate of 10 °CÆ min )1 until 330 °C). The temperature of the ioniza- tion chamber was 250 °C, and ionization was by electron impact at 70 eV. In vitro assay of b-amyrin and sophoradiol hydroxylase using cell-free transformant extracts The transformant with pESC-CYP93E1 was inoculated in 20 mL of SCR-U medium (described above), and incubated at 30 °C for 1 day. Then, 1 mL of 40% galactose solution was added (final concentration 2%). The cells were incuba- ted under the same conditions for 1 day, harvested by cen- trifugation at 500 g for 5 min, resuspended in 0.1 mL of 50 mm potassium phosphate buffer (pH 7.5, containing 10% sucrose, 5 mm EDTA, and 14 mm 2-mercaptoetha- nol), and broken using a Beat-beater (Biospec Products, Oklahoma, USA) with glass beads (0.4–0.6 mm diameter, one-third of total volume) at 4 °C. Then an additional 0.4 mL of the same buffer was added to suspend broken cells. Cell homogenates were centrifuged at 3000 g for 5 min. The supernatant (0.4 mL) was used as the enzyme solution. With a solution (0.1 mL) of the NADPH-re- generating system (10 mm NADPH, 38 mm glucose-6-phos- phate, 2.5 UÆmL )1 glucose-6-phosphate dehydrogenase), the enzyme solution and the substrate (b-amyrin or sophoradiol 50 nmol) were incubated for 6 h at 30 °C. The reaction was CYP93E1 M. Shibuya et al. 956 FEBS Journal 273 (2006) 948–959 ª 2006 The Authors Journal compilation ª 2006 FEBS terminated by the addition of 0.5 mL of 40% potassium hydroxide solution. After boiling for 5 min, products were extracted twice with 1 mL of hexane. The hexane extract was evaporated, acetylated, and analyzed using GC-MS fol- lowing the same procedure as described above. Production of olean-12-ene-3b,24-diol by coexpression of CYP93E and b-amyrin synthase in lanosterol synthase-deficient S. cerevisiae strain GIL77 Two oligo DNAs (5¢-CTTCGTCGACAAGATGTGGAG GTTGAAGATA-3¢ and 5¢-GTCCGCTAGCTCAAGGCA AAGGAACTCTTCT-3¢), corresponding to the N- and C-terminal sequences of b-amyrin synthase from P. sativum, were synthesized. The open reading frame was amplified by PCR with these primers and the plasmid pYES-PSY [13] as a template under the same conditions as described above, except that the annealing temperature was 58 °C. The amplified DNA fragment was ligated into the restriction enzyme sites (SalI and NheI) of pESC after digestion with these enzymes. The plasmid obtained was designated pESC- PSY. S. cerevisiae strain GIL77 [8] was transformed with pESC-PSY. Production of b-amyrin by the resulting trans- formant was confirmed following the reported method [13]. pESC-CYP93E1 and pESC-PSY were digested with restric- tion enzymes (SalI and ClaI). The fragment containing CYP93E1 and promoters (GAL1 ⁄ GAL10) from pESC- CYP93E1 was separated upon agarose gel electrophoresis and ligated into SalI and ClaI-digested and dephosphory- rated pESC-PSY. The resulting plasmid was designated pESC-PSY-CYP93E1. S. cerevisiae strain GIL77 was trans- formed with this plasmid using the same method as described above. The transformant with pESC-PSY- CYP93E1 was inoculated in 1 L of SCR-U medium (des- cribed above), and incubated at 30 °C for 1 day. Then, 50 mL of 40% galactose solution was added (final concen- tration 2%). The cells were incubated under the same con- ditions for 1 day, harvested by centrifugation at 500 g for 5 min, resuspended in 100 mL of 0.1 m potassium phos- phate buffer (pH 7.0) supplemented with 2% glucose and hemin chloride (13 lgÆmL )1 ), and further incubated at 30 °C for 24 h. Then, the cells were collected and suspen- ded in 25 mL of 40% potassium hydroxide and 25 mL of methanol and refluxed for 2 h. Products were extracted with 50 mL of hexane. The hexane layer was washed with 25 mL of saturated sodium bicarbonate. Extraction was repeated three times, and then the hexane layers were com- bined, dehydrated with sodium sulfate, and evaporated. The residue was applied on a silica gel column (4 g of Wako FC-40, Wako Pure Chemical Industries, Osaka, Japan) with the eluent (benzene:acetone ¼ 4 : 1). The frac- tions containing the products were combined, evaporated (1.4 mg), and again applied on a silica gel column (2 g of Wako FC-40) with the eluent (hexane:ethyl acetate ¼ 4:1) to isolate the product (1.0 mg) of GIL77 ⁄ pESC-PSY- CYP93E1. 1 H- and 13 C-NMR spectra were measured in CDCl 3 (ECA, JEOL, Tokyo, Japan, 1 H: 500 MHz, 13 C: 125 MHz) with the signal of CHCl 3 (7.26 p.p.m. in 1 H, 77.00 in 13 C) as an internal standard. 1 H-NMR (500 MHz, CDCl 3 ): d 0.82 (3H, s), 0.87 (6H, s), 0.88 (3H, s), 0.93 (3H, s), 1.13 (3H, s), 1.25 (3H, s), 3.34 (1H, d, J ¼ 11.5 Hz, C-24), 3.45 (1H, dd, J ¼ 5.5 Hz, 11.5 Hz, C-3), 4.21 (1H, d, J ¼ 11.5 Hz, C-24), 5.17(1H, t, J ¼ 3.5 Hz, C-12). 13 C-NMR (125 MHz, CDCl 3 ): d 16.0, 16.7, 18.4, 22.3, 23.7, 23.8, 25.9, 26.1, 26.9, 27.7. 28.4, 31.1, 32.5, 32.8, 33.3, 34.7, 36.6, 37.1, 38.3, 39.8, 41.7, 42.8, 46.8, 47.2, 47.7, 55.8, 64.5, 80.9, 121.5, 145.2. Acknowledgements The authors are grateful to Dr S. Nishimaya (Meiji Seika Kaisha, Ltd) for the gift of authentic olean-12- ene-3b,24-diol, sophoradiol, and soyasapogenol B. A part of this research was financially supported by a Grant-in-Aid for Scientific Research (S) (No. 15101007) to Y.E. from the Ministry of Education, Culture, Sports, Science and Technology, Japan. References 1 Waller GR & Yamasaki K, eds. (1996) Saponins Used in Food and Agriculture: Advances in Experimental Medi- cine and Biology. Vol. 405. Plenum Press, NY & London. 2 Shibata S (2001) Chemistry and cancer preventing activ- ities of ginseng saponins and some related triterpenoid compounds. J Korean Med Sci 16, Suppl. 28–37. 3 Haridas V, Higuchi M, Jayatilake GS, Bailey D, Mujoo K, Blake ME, Arntzen CJ & Gutterman JU (2001) Avi- cins: triterpenoid saponins from Acacia victoriae (Ben- tham) induce apoptosis by mitochondrial perturbation. Proc Natl Acad Sci USA 98, 5821–5826. 4 Ko K-S, Noguchi H, Ebizuka Y & Sankawa U (1989) Oligoside production by hairy root cultures transformed by Ri plasmids. Chem Pharm Bull 37, 245–248. 5 Hayashi H, Sakai T, Fukui H & Tabata M (1990) Formation of soyasaponins in licorice cell-suspension cultures. Phytochemistry 29, 3127–3129. 6 Hayashi H, Fukui H & Tabata M (1988) Examination of triterpenoids produced by callus and cell-suspension cultures of Glycyrrhiza glabra. Plant Cell Rep 7, 508– 511. 7 Xu R, Fazio GC & Matsuda SPT (2004) On the origins of triterpenoid skeletal diversity. Phytochemistry 65, 261–291. 8 Kushiro T, Shibuya M & Ebizuka Y (1998) b-Amyrin synthase: Cloning of oxidosqualene cylcase that catalyzes M. Shibuya et al. CYP93E1 FEBS Journal 273 (2006) 948–959 ª 2006 The Authors Journal compilation ª 2006 FEBS 957 [...]... (2004) Comparative genomics of rice and Arabidopsis Analysis of 727 cytochrome P450 genes and pseudogenes from a monocot and a dicot Plant Physiol 135, 756–772 29 D’Auria JC & Gershenzon J (2005) The secondary metabolism of Arabidopsis thaliana: Growing like a weed Curr Opin Plant Biol 8, 308–316 30 Schopfer CR & Ebel J (1998) Identification of elicitorinduced cytochrome p450s of soybean (Glycine max L.)... L.) using differential display of mRNA Mol Gen Genet 258, 315–322 31 Steele CL, Gijzen M, Qutob D & Dixon RA (1999) Molecular characterization of the enzyme catalyzing the aryl migration reaction of isoflavonoid biosynthesis in soybean Arch Biochem Biophys 367, 146–150 32 Schopfer CR, Kochs G, Lottspeich F & Ebel J (1998) Molecular characterization and functional expression of dihydroxypterocarpan 6a-hydroxylase,... compilation ª 2006 FEBS M Shibuya et al 34 Kitagawa I, Yoshikawa M, Wang HK, Saito M, Tosirisuk V, Fujiwara T & Tomita K (1982) Revised structures of soyasapogenol A, soyasapogenol B, and soyasapogenol E, oleanene sapogenols from soybean, structures of soyasaponin I, soyasaponin II, and soyasaponin III Chem Pharm Bull 30, 2294–2297 35 Ireland PA & Dziedzic SZ (1986) Effect of hydrolysis on sapogenin release... Up-regulation of soyasaponin biosynthesis by methyl jasmonate in cultured cells of Glycyrrhiza glabra Plant Cell Physiol 44, 404–411 37 Ebel J & Grisebach H (1988) Defense strategies of soybean against the fungus Phytophthora megasperma f sp glycinea: a molecular analysis Trends Biochem Sci 13, 23–27 38 Lawton MA & Lamb CJ (1987) Transcriptional activation of plant defense genes by fungal elicitor, wounding, and. .. Mol Cell Biol 7, 335–341 39 Aoki T, Akashi T & Ayabe S (2000) Flavonoids of leguminous plants: structure, biological activity, and biosynthesis J Plant Res 113, 475–488 40 Akashi T, Aoki T & Ayabe S (1999) Cloning and functional expression of a cytochrome P450 cDNA encoding 2-hydroxyisoflavanone synthase involved in biosynthesis of the isoflavonoid skeleton in licorice Plant Physiol 121, 821–828 41 Bishop... Croteau RB (2001) Molecular cloning of a cytochrome P450 CYP93E1 45 46 47 48 49 50 51 52 53 taxane 10b-hydroxylase cDNA from Taxus and functional expression in yeast Proc Natl Acad Sci USA 98, 1501–1506 Wheeler AL, Long RM, Ketchum REB, Rithner CD, Williams RM & Croteau RB (2001) Taxol biosynthesis: differential transformations of taxadien-5a-ol and its acetate ester by cytochrome P450 hydroxylases from... Hasegawa J & Nishino H (2004) Triterpene acids from the leaves of Perilla frutescens and their anti-inflammatory and antitumor-promoting effects Biosci Biotech Biochem 68, 85–90 Cabello-Hurtado F, Zimmerlin A, Rahier A, Taton M, DeRose R, Nedelkina S, Batard Y, Durst F, Pallett KE & Werck-Reichhart D (1997) Cloning and functional expression in yeast of a cDNA coding for an obtusifoliol 14a-demethylase (CYP51)... evolution of oxidosqualene cyclases in higher plants Biol Pharm Bull 26, 642–650 958 20 Ebizuka Y, Katsube Y, Tsutsumi T, Kushiro T & Shibuya M (2003) Functional genomics approach to the study of triterpene biosynthesis Pure Appl Chem 75, 361–366 21 Iturbe-Ormaetxe I, Haralampidis K, Papadopoulou K & Osbourn AE (2003) Molecular cloning and characterization of triterpene synthases from Medicago truncatula and. .. 12923–12928 Morita M, Shibuya M, Kushiro T, Masuda K & Ebizuka Y (2000) Molecular cloning and functional expression of triterpene synthases from pea (Pisum sativum): New a-amyrin producing enzyme is a multifunctional triterpene synthase Eur J Biochem 267, 3453–3460 Kushiro T, Shibuya M, Masuda K & Ebizuka Y (2000) A novel multifunctional triterpene synthases from Arabidopsis thaliana Tetrahedron Lett 41, 7705–7710... Y, Kushiro T, Shibuya M & Ebizuka Y (2001) Cloning and characterization of a cDNA encoding b-amyrin synthase involved in glycyrrhizin and soyasaponin biosynthesis in licorice Biol Pharm Bull 24, 912–916 Hayashi H, Huang P, Inoue K, Hiraoka N, Ikeshiro Y, Yazaki K, Tanaka S, Kushiro T, Shibuya M & Ebizuka Y (2001) Molecular cloning and characterization of isomultiflorenol synthase, a new triterpene synthase . Identification of b-amyrin and sophoradiol 24-hydroxylase by expressed sequence tag mining and functional expression assay Masaaki Shibuya 1 ,. approach by functional analysis of heterologously expressed P450 based on genomic sequences or expressed sequence tags (EST) appeared HO β-Amyrin HO Sophoradiol OH HO Olean-12-ene-3β,24-diol HO HO Soyasapogenol