Furanocoumarin biosynthesis in Ammi majus L. Cloning of bergaptol O -methyltransferase Marc Hehmann 1, *, Richard Lukac ˇ in 1, *, Halina Ekiert 2 and Ulrich Matern 1 1 Institut fu ¨ r Pharmazeutische Biologie, Philipps-Universita ¨ t Marburg, Germany; 2 Department of Pharmaceutical Botany, Collegium Medicum, Jagiellonian University, Krako ´ w, Poland Plants belonging to the Apiaceae or Rutaceae accumulate methoxylated psoralens, such as bergapten or xanthotoxin, as the final products of their furanocoumarin biosynthesis, and the rate of accumulation depends on environmental and other cues. Distinct O-methyltransferase activities had been reported to methylate bergaptol to bergapten and xantho- toxol to xanthotoxin, from induced cell cultures of Ruta graveolens, Petroselinum crispum and Ammi majus. Bergap- tol 5-O-methyltransferase (BMT) cDNA was cloned from dark-grown Ammi majus L. cells treated with a crude fungal elicitor. The translated polypeptide of 38.7 kDa, composed of 354 amino acids, revealed considerable sequence similar- ity to heterologous caffeic acid 3-O-methyltransferases (COMTs). For homologous comparison, COMT was cloned from A. majus plants and shown to share 64% identity and about 79% similarity with the BMT sequence at the polypeptide level. Functional expression of both enzymes in Escherichia coli revealed that the BMT activity in the bacterial extracts was labile and rapidly lost on purification, whereas the COMT activity remained stable. Furthermore, the recombinant AmBMT, which was most active in potassium phosphate buffer of pH 8 at 42 °C, showed narrow substrate specificity for bergaptol (K mSAM 6.5 l M ; K m Bergaptol 2.8 l M ) when assayed with a variety of sub- strates, including xanthotoxol, while the AmCOMT accep- ted 5-hydroxyferulic acid, esculetin and other substrates. Dark-grown A. majus cells expressed significant BMT activity which nevertheless increased sevenfold within 8 h upon the addition of elicitor and reached a transient maxi- mum at 8–11 h, whereas the COMT activity was rather low and did not respond to the elicitation. Complementary Northern blotting revealed that the BMT transcript abun- dance increased to a maximum at 7 h, while only a weak constitutive signal was observed for the COMT transcript. The AmBMT sequence thus represents a novel database accession specific for the biosynthesis of psoralens. Keywords: Ammi majus L.; Apiaceae; furanocoumarin biosynthesis; bergaptol O-methyltransferase; caffeate O-methyltransferase. Cell suspension cultures of the Apiaceae, in particular Ammi majus L. [1,2] and Petroselinum crispum [3–6], have served in numerous model studies on the induced plant disease resistance response. Upon treatment with fungal elicitor, these cells produce linear furanocoumarins (psoralens) besides lignin-like compounds for reinforcement of their cell walls [7–9]. Various crude cell wall elicitors, particularly Pmg (from Phythophthora sojae,formerlyPhythophthora mega- sperma f. sp. glycinea), have been used in the past, and at least in case of the induction of parsley cells the eliciting principle has been identified as a peptide [10]. Furanocoumarins are potentially toxic compounds which probably function as phytoalexins in the response to fungal infection [3], but their accumulation can also be triggered by other means, e.g. wounding of plants or exposure to acid fog [11,12]. A considerable proportion of psoralens may be recovered from the surface of plants [13], and most of the psoralens elicited in cell cultures also accumulated in the culture fluid. Psoralens are capable of intercalating with DNA, and the methoxyl- ated psoralens bergapten and xanthotoxin are the most relevant natural furanocoumarins in terms of their thera- peutic potential. These psoralens exhibit photosensitizing and antiproliferative activities [14] and were evaluated as photosensitizing drugs in oral psoralen plus UVA irradiation (PUVA) therapy of psoriasis and vitiligo [15,16]. A. majus cells also uniquely produce derivatives of 7-O- prenylumbelliferone under conditions of elicitation (Fig. 1) [17], and the induction of coumarin biosynthesis in these cell cultures provided the basis for extensive in vitro investiga- tions. Both the umbelliferone 7-O-prenyltransferase activity and a 6-C-prenyltransferase activity forming demethylsube- rosin en route to the psoralens (Fig. 1) were found associated with the microsomal fraction [18]. Such a 6-C- prenyltransferase activity had been reported initially as a Mn 2+ -dependent enzyme from Ruta graveolens and assigned to the plasitidic membranes [19]. Individual Correspondence to U. Matern, Institut fu ¨ r Pharmazeutische Biologie, Philipps-Universita ¨ t Marburg, Deutschhausstrasse 17A, D-35037 Marburg, Germany. Fax: + 49 6421 282 6678, Tel.: + 49 6421 282 2461, E-mail: matern@staff.uni-marburg.de Abbreviations:SAM,S-adenosyl- L -methionine; BMT, bergaptol 5-O-methyltransferase; XMT, xanthotoxol 8-O-methyltransferase; OMT, O-methyltransferase; COMT, caffeic acid 3-O-methyltrans- ferase; PUVA, psoralen plus UVA irradiation; RACE, rapid amplification of cDNA ends; RLM-RACE, RNA ligase-mediated rapid amplification of cDNA ends. *These authors contributed equally to the work described. (Received 3 November 2003, revised 17 December 2003, accepted 15 January 2004) Eur. J. Biochem. 271, 932–940 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.03995.x cytochrome P450 monooxygenases sequentially convert demethylsuberosin to (+)-marmesin and psoralen, and these activities were also demonstrated in the microsomal fraction of elicited A. majus cells [1,2]. Moreover, the conversion of (+)-marmesin to psoralen (Fig. 1) was proven to proceed by syn-elimination releasing acetone [20]. The subsequent hydroxylation of psoralen in the 5- or 8-position yields bergaptol and xanthotoxol, respectively (Fig. 1). Both of these hydroxylations to yield 5,8-dihyd- roxypsoralen are required for the formation of isopimpin- ellin, which commonly accumulates as a minor byproduct upon induction, but the order of hydroxylations and O-methylations remains unresolved [19]. However, psoralen 5-monooxygenase activity forming bergaptol was demon- strated in vitro with microsomes from elicited A. majus cells [2]. The biosynthesis of bergaptol from umbelliferone is thus entirely catalyzed through membrane-bound enzymes, involving one prenyltransferase and three P450 monoxy- genases, and is preceded by the formation of umbelliferone from 4-coumaric acid, which was also proposed to depend on a P450 monooxygenase. The 5- or 8-hydroxylated furanocoumarins (bergaptol or xanthotoxol) are further processed by O-methylation to bergapten and xanthotoxin, and the corresponding O-methyltransferases (OMTs) were identified as distinct entities and purified by affinity chromatography from Ruta graveolens [21] and later also from Petroselinum crispum [22]. S-Adenosyl- L -methionine– bergaptol 5-O-methyltransferase (SAM–BMT) and S-aden- osyl- L -methionine–xanthotoxol 8-O-methyltransferase (SAM–XMT) (Fig. 1) are also expressed in dark-grown A. majus cells, and in all instances these methyltransferases are soluble and inducible enzymes. We report here the cloning and functional characteriza- tion of BMT from elicitor-treated Ammi majus L. cells as a major step towards a molecular understanding of psoralen biosynthesis. For comparison, the closely related S-adeno- syl- L -methionine–caffeate 3-O-methyltransferase (SAM– COMT) was also cloned from A. majus plants, and the differential regulation of these methyltransferases was examined upon elicitation of the cell cultures. Materials and methods Ammi majus cell cultures and induction Cell suspension cultures of A. majus L. (40 mL B5 + - medium in 250 mL flasks) were initiated and grown continuously in the dark as described elsewhere [1,2]. Pmg Fig. 1. Schematic outline of linear furanocou- marin biosynthesis. The sequence of hydroxy- lations and O-methylations of psoralen leading to isopimpinellin has not been established. Ó FEBS 2004 Bergaptol O-methyltransferase in Ammi majus 1 (Eur. J. Biochem. 271) 933 elicitor was suspended in distilled water (5 mgÆmL )1 ), the suspension was heated to boiling point and added to 6-day- old cell cultures (1 mL per 40 mL culture). The cells were harvested 4 h later and immediately frozen in liquid nitrogen and stored at )70 °C until use. Chemicals Biochemicals were purchased from Roth (Karlsruhe, Germany), vectors and Escherichia coli host strains from Invitrogen (Karlsruhe, Germany) or Qiagen (Hilden, Germany). Restriction enzymes and DNA modifying enzymes were from MBI-Fermentas (St. Leon-Rot, Germany), Promega (Mannheim, Germany) or Stratagene (Heidelberg, Germany). Bergaptol was bought from Extrasynthese (Genay, France), caffeic acid from Roth (Karlsruhe, Germany), and [methyl- 14 C]S-adenosyl- L -methionine was purchased from Hartmann Analytic (Braunschweig, Germany). RNA isolation, PCR cloning and heterologous expression Total RNA was isolated from Pmg elicitor-induced cells following the protocol of Giuliano et al. [23]. The time of elicitor-induction was chosen from previous induction experiments in which the time course of furanocoumarin- specific enzyme activities in A. majus had been monitored [1,2]. Alternatively, the RNA was isolated from the stems and leaves of 4–6 week-old A. majus plants. cDNA frag- ments were generated by RT-PCR amplification using degenerate oligonucleotide primers [24] which had been designed according to conserved amino acid sequences of plant OMTs [25,26]. The cDNA fragments were cloned, sequenced, and full length clones were generated by the rapid ampification of cDNA ends (RACE) and RNA ligase-mediated rapid amplification of cDNA ends (RLM- RACE) techniques, respectively, using gene-specific pri- mers. Cloning of the PCR products was performed by TOPO TA Cloning (Invitrogen, Karlsruhe, Germany). Briefly, the protein coding regions of the putative BMT and COMT were amplified with 5¢-primers providing an NcoI site directly before the start codon and 3¢-primers inserting a BamHI site after the stop codon before they were cloned into the pCR2.1-TOPO vector. An internal NcoIsite contained in the ORF of the COMT was deleted by using QuikChangeÒ Multi Site-Directed Mutagenesis Kit as described by the manufacturer (Stratagene, Heidelberg, Germany) without altering the amino acid sequence. The mutation was verified by DNA sequencing [27], and the BMT- and COMT-coding DNA clones were subsequently isolated by digestion with NcoIandBamHI. The cDNAs were subcloned into pQE60 vector (Qiagen, Hilden, Germany) for functional expression in E. coli strain M15 (Qiagen) harboring the plasmid pRep4 and employed for BMT and COMT activity assays, respectively. The expres- sion was induced by the addition of 1.0 m M isopropyl thio- b- D -galactoside [28]. Following the induction, the cells were harvested by centrifugation [29], disrupted by ultrasonica- tion, the crude extract was cleared by centrifugation (30 000 g, 4 °C, 10 min), and enzyme assays were carried out with the supernatants. Sequence analysis The cDNAs amplified by RT-PCR were sequenced by the dideoxy nucleotide chain termination technique [27]. The cDNA sequences were subjected to BLAST searches (advanced WU-Blast2; EMBL) and alignments with CLUSTALW algorithm (EMBL). Purification procedure The crude extract was fractionated by ammonium sulfate precipitation from 0 to 45%, 45–60% and 60–80% saturation. The 60–80% fraction (BMT) and the 45–60% fraction (COMT) were dissolved in 200 m M potassium phosphate buffer of pH 8.0 and 200 m M Tris/ HCl buffer of pH 7.5, respectively. The extracts were desalted by size exclusion chromatography 2 through PD-10 columns (Amersham, Freiburg, Germany) and on Frac- togel EMD BioSEC (S) (Merck, Darmstadt, Germany). The purification was monitored by SDS/PAGE [30] and enzyme activity assays of individual fractions. Enzyme assays and other analytical methods BMT activity was routinely measured at 42 °C in 200 m M potassium phosphate buffer pH 8.0 in the presence of sodium ascorbate (20 m M ), magnesium chloride (1.5 m M ), bergaptol (250 l M )andtherecombinantbacterial enzyme extract. The reaction was started by the addition of S-adenosyl- L -[methyl- 14 C]methionine (40 l M ), and the product was identified as bergapten by silica thin-layer chromatography employing trichloromethane/ethylacetate (2 : 1, v/v; R F bergaptol 0.52, R F bergapten 0.77), trichlorometh- ane/methanol (95 : 5, v/v; R Fbergaptol 0.26, R F bergapten 0.78), n-hexane/ethylacetate/methanol (5 : 5 : 1, v/v/v; R Fbergaptol 0.67, R F bergapten 0.78) or toluene/ethylacetate (3 : 2, v/v; R Fbergaptol 0.34, R F bergapten 0.68)asthesolventsystems.The COMT activity assay was carried out at 32 °C in 200 m M potassium phosphate buffer pH 7.5 containing sodium ascorbate (20 m M ), magnesium chloride (1.5 m M ), caffeic acid (250 l M )andS-adenosyl- L -[methyl- 14 C]methionine (40 l M ) in addition to 3 the crude enzyme protein. The reactions were stopped by the addition of 1 M 4 HCL (30 lL) and extracted with 400 lL ethyl acetate. Aliquots of the organic phase (200 lL) were mixed with 5 mL scintillation cocktail (Roth, Karlsruhe, Germany) and measured in a liquid scintillation counter (1214 Rackbeta; PerkinElmer, Wellesley, MA, USA). Incubations with boiled enzyme (5 min at 100 °C), or mixtures lacking bergaptol and caffeic acid, were run for control and served for background corrections. Linear conditions for kinetic assays were established by adjusting the amount of enzyme protein (BMT: 0.15–3.0 lg per assay; COMT: 0.5–10.25 lg per assay). The BMT assays were usually conducted for 20 min, using 1.5 lgdesalted protein and 4.0 nmol S-adenosyl- L -[methyl- 14 C]methionine and 25.0 nmol bergaptol per 100 lL incubation, which secured linear conversion rates for about 60 min. The COMT assays were carried out accordingly using 5 lg desalted protein. Protein was determined according Lowry et al. [31] and the data were extrapolated from Lineweaver–Burk plots. 934 M. Hehmann et al.(Eur. J. Biochem. 271) Ó FEBS 2004 Northern blotting Following the addition of Pmg elicitor to the A. majus cell suspensions, total RNA was isolated from the cells every 0.5 h up to 8 h and used for Northern blot analysis (RNA dot blot). The RNA (4 lg) was denaturated in 0.5· Mops buffer pH 6.0, containing 50% (v/v) formamide and 2.2 M formaldehyde, and transferred to a Hybond-N + nylon membrane (Amersham Biosciences, Freiburg) [32] in an I-SRc 96-Dot Blot Minifold (Schleicher and Schu ¨ ll, Dassel, Germany). The full size A. majus BMT cDNA (1062 bp) and COMT cDNA (1095 bp) were 32 P-labeled using a Rediprime TM II-random prime labeling system (Amersham Biosciences, Freiburg), and used as probes. The blots were blocked and then hybridized with 25 ng of one of these two labeled probes. Hybridization was carried out overnight at 68 °Cin2· Denhardt’s solution in the presence of 0.5% (w/v) SDS and 100 lgÆmL )1 salmon sperm DNA (Sigma, Deisenhofen, Germany). After stringent washings for 20 min at room temperature in 2· NaCl/Cit followed by 15 min at 68 °Cin2· NaCl/Cit the membranes were exposed to a Bio Imaging Analyzer FLA-2000 (Fujifilm). Results Induction of O -methyltransferase activities Elicitor–induction studies had been carried out previously with A. majus cell cultures, and the coumarin-specific enzyme activities (dimethylallyldiphosphate: umbelliferone 6-C-and7-O-dimethylallyltransferases) commonly reached a first maximum at 12 h of induction [18]. This time course suggested maximal transcript abundances within the first 6 h of elicitation and corresponded to the patterns reported for the elicitor induction of phenylalanine ammonia lyase and 4-coumarate:CoA ligase activities in Petroselinum crispum cell cultures [33]. However, a different induction profile was reported by Hauffe et al. [22] for the BMT activity in P. crispum cells with a maximum beyond 25 h. Therefore, preliminary assays of BMT and XMT activities were conducted with crude extracts of dark-grown A. majus cells and under the conditions described for Ruta graveolens [34]. These assays revealed that both O-methyltransferase activities were expressed already in the controls (0.4 lkatÆ kg )1 BMT and 66 nkatÆkg )1 XMT, on average), but increased considerably in response to treatment of the cell suspensions with Pmg elicitor. The activities were measured every three hours over the time period from 2 to 23 h following the addition of elicitor, and shown to increase sevenfold within 8 h to reach a transient maximum at about 8–11 h. (H. Ekiert and R. Lukac ˇ in, unpublished data). Thus, the cells at 4 h after the addition of elicitor were considered to contain the highest BMT and XMT transcript abundances. The activity of COMT in A. majus was examined for comparison, and only low levels (2.5 lkatÆkg )1 on average) were observed in crude extracts of suspension-cultured cells, which hardly changed upon elicitation and might be due to related OMTs with specificity towards catechols [26,35]. However, higher activity (4.2 lkatÆkg )1 ) was determined in extracts of leaf and stem tissues. The moderate rate of COMT expression in the suspension cells is reminiscent of the correspondingly low COMT activity in cultured Petro- selinum crispum cells [36]. In Petroselinum as well as in Ruta cells, the COMT activity had been clearly distinguished from the BMT and XMT activities [22,34]. cDNA cloning and functional expression The total RNA from A. majus cells that had been elicited for 4 h was used as a template for RT-PCR amplifications, with degenerate oligonucleotide primers designed for the cloning of COMT-related enzymes from other plant sources [24–26]. These experiments generated two different frag- ments of 215 bp which were cloned into the pCR2.1-TOPO vector and extended to the full size cDNAs of 1062 and 1074 bp, respectively, by RACE and RLM-RACE (Gen- Bank accession nos. AY443006 and AY443008). Prelimin- ary sequence alignments had already revealed a close similarity of the cDNAs with those of other plant OMTs. Therefore, the inserts were subcloned into an expression vector for the expression in E. coli, and the recombinant polypeptides were extracted from the induced transformants Fig. 2. Thin-layer cochromatography of the labeled product from BMT assays with authentic bergaptol and bergapten on silica F 254 plates developed with n-hexane/ethylacetate/methanol (5 : 5 : 1, v/v/v). Ref- erence furanocoumarins separated in the absence (lane 1) and in the presence of the enzymatic product (lane 3) were spotted by their quenching under irradiation at 254 nm, and the enzymatic product (lanes 2 and 3) was detected by autoradiography using a Bioimager (inverse presentation). S, start line; F, solvent front. Ó FEBS 2004 Bergaptol O-methyltransferase in Ammi majus 1 (Eur. J. Biochem. 271) 935 in 70 m M Tris/HCl buffer pH 7.5, containing 10 m M EDTA. The enzyme activity of the crude supernatants was determined with a variety of potential substrates, and the 1062 bp transformant was found to methylate bergaptol to bergapten with narrow substrate specificity. The identity of the enzymatic product was firmly established by thin- layer cochromatography with authentic bergapten in four solvent systems (Fig. 2), and hence the transformant encoded a BMT, designated AmBMT. The functionality of the 1074 bp clone encoding a COMT-like protein, however, has so far not been assigned. Because the RT-PCR from cell culture RNA failed to amplify a full-size COMT sequence, we turned back to A. majus plants and used the RNA from leaf and stem tissues as a template. These experiments yielded a cDNA of 1095 bp (GenBank accession no. AY443007), which was expressed in E. coli andshowntoencodeaCOMTthat converts caffeate to ferulate. Both labeled cDNAs were used as probes for Northern blotting experiments employing the total RNA of A. majus cells at various time points following the addition of the Pmg elicitor. The abundance of BMT transcripts, which were hardly detectable in control cells, increased signifi- cantly to a transient maximum at 7 h (Fig. 3). However, a very weak hybridization signal was recorded for the COMT transcripts that hardly changes in intensity over the time of the experiment and corresponded with the low constitutive level of enzyme activity in the cells. Characterization of enzymes The BMT activity ( 2.5 lkatÆkg )1 ) in the desalted bacterial extracts was very labile and could not be purified exten- sively, whereas the COMT activity (1.6 lkatÆkg )1 )remained stable upon storage. The enzyme extracts were therefore subjected only to ammonium sulfate fractionation (45–60% saturation for COMT; 60–80% saturation for BMT) and subsequent desalting through PD-10 columns, enhancing the apparent specific activities to 20.2 and 10.3 lkatÆkg )1 , respectively, for BMT and COMT. The rates of enzyme activity were compared in various buffers in pH range 2.0–10.0 and at temperatures ranging from 20 to 50 °C. Significant activity of the recombinant BMT was observed between 38 °Cand44°C and from pH 6.5–9.0, and the optimum was recorded at 42 °C in potassium phosphate buffer pH 8.0, whereas the optimal COMT activity was observed at 32 °C in potassium phosphate buffer pH 7.0. These conditions were routinely chosen for all further assays. This activity profile of the recombinant BMT was fully compatible with the data measured for BMT extracted from A. majus cells (H. Ekiert and R. Lukac ˇ in, unpublished data), and the pH dependency corresponded to that of the BMTs from P. crispum [22] or R. graveolens [34]. The effect of a number of metal ions (Co 2+ ,Cu 2+ ,Fe 2+ ,Fe 3+ , Mg 2+ ,Mn 2+ ,Ni 2+ ,Zn 2+ )at1.5m M or 0.1 m M concen- tration was also examined. Significant inhibition of the BMT activity was observed in the presence of Cu 2+ (100% Fig. 3. Induction of BMT transcript abundance in Ammi majus cell cultures. Total RNA (4 lg per dot) isolated from the cells at different time intervals following the addition of elicitor (lanes 1 +, 2 +) or from controls (lanes 3 –, 4 –) treated with water (1 mLÆ40 mL )1 culture) was employed for Northern dot blot hybridization using labeled AmBMT cDNA as a probe. Table 1. Substrate specificities of A. majus OMTs. Substrates were used at 10 m M concentration in the assays. Neither of the OMTs accepted umbelliferone, psoralen, xanthotoxol, sinapate, 4-coumarate, 2-coumarate, 3-coumarate, catechol, kaempferol, quercetin, dihydrokaempferol, apigenin or naringenin to a significant extent (< 1%) as a substrate. The relative activity values relate to caffeate as the standard substrate. Substrate COMT BMT Rel. activity (%) K m Rel. activity (%) K m Bergaptol < 1 100 2.8 l M b Caffeate 100 122.0 l M a <1 5-Hydroxyferulate 177 29.0 l M <1 Caffeic acid methyl ester 207 42.0 l M <1 Caffeoyl coenzyme A 13.5 219.0 l M <1 3-(3,4-dihydroxyphenyl)propionate 10.4 2.2 m M <1 Esculetin 27 133.0 l M <1 Daphnetin 8 103.0 l M <1 a K m SAM ¼ 2 l M . b K m SAM ¼ 6.5 l M . 936 M. Hehmann et al.(Eur. J. Biochem. 271) Ó FEBS 2004 and 51%) and Ni 2+ (47% and 16%) as well as Co 2+ (21% and 10%). COMT activity was completely inhibited at either of the Cu 2+ concentrations but less by Ni 2+ (91.5% and 47%), Mn 2+ (83% and 7.5%), or Co 2+ ,Fe 3+ and Zn 2+ (23–30% and 0–5%). Fe 2+ and Mg 2+ did not affect the turnover rates. Similar results were reported for heterologous OMTs [37,38]. A variety of potential substrates, including xanthotoxol (Table 1), was employed to determine the substrate specificities of the recombinantly expressed A. majus BMT and COMT. However, only bergaptol was accepted as a substrate by the BMT. Kinetic assays revealed the affinities to S-adenosyl- L -methionine and bergaptol at K m ¼ 6.5 and 2.8 l M , respectively. The COMT was much less selective and showed the highest affinity to 5-hydroxyferulate (K m ¼ 29 l M ) followed by caffeic acid methyl ester, caffeate, esculetin, caffeoyl-CoA, 3-(3,4-dihydroxyphenyl)propionate (dihydrocaffeate) or daphnetin (Table 1). Relationship of sequences S-Adenosyl- L -methionine-dependent O-methyltransferases are characterized by a common signature of five highly Fig. 4. Alignment of AmBMT and AmCOMT polypeptides from Ammi majus. Hyphens were inserted for maximal alignment. The consensus sequence (COMTcons) derived from the COMT polypeptides of Ocimum basilicum, Catharanthus roseus, Capsicum annuum, Capsicum chinense, Prunus dulcis and Rosa chinensis is given in the bottom line. Identical amino acid residues are denoted by asterisks, and dots mark conservative exchanges. The amino acids are numbered in the right margin. Highly conserved regions I–V proposed as a signature of S-adenosyl- L -methionine- dependent O-methyltransferases [26,39,40] are underlined, and the motifs 1 and 2 considered to govern the substrate specificity [44,47] are in bold. Ó FEBS 2004 Bergaptol O-methyltransferase in Ammi majus 1 (Eur. J. Biochem. 271) 937 conserved regions [39–41], and a corresponding consensus sequence was assigned from plant OMTs [26]. These elements were also recognized in both the BMT and COMT sequences from A. majus (Fig. 4; regions I–V) showing 94.5% and 97% identity with the consensus sequence. In case of rat liver catechol OMT regions I and IV were shown by X-ray diffraction to be involved in S-adenosyl- L -methionine and metal binding [42], and the other three regions are likely to serve the same purpose. Generally, five different structural folds have been reported to bind SAM during catalysis [43], subclassifying the OMTs, but most plant OMTs, including COMT and the homologous BMT, belong to class I. In contrast to BMTs, COMTs occur ubiquituously in plants and have been cloned from many different sources. Based on the COMTs of Ocimum basilicum, Catharanthus roseus, Capsicum annuum, Capsi- cum chinense, Prunus dulcis and Rosa chinensis,whichshare a sequence identity of about 50% and 72.3% similarity, a consensus sequence was derived also for the full size polypeptides (Fig. 4). The AmCOMT sequence was fully compatible with this consensus sequence showing 57% identity at 80% similarity. However, in the light of such a relationship it is particularly notable that the alignment of the AmBMT polypeptide with the AmCOMT sequence revealed 64% identity and 78.4% similarity. Discussion The activities of XMT and BMT were described previously from Ruta graveolens [21] and Petroselinum crispum cells [22]. In case of dark-grown Petroselinum cells both activities were induced upon elicitor treatment to transient maxima at 30–35 h with PcXMT reaching a three- to fourfold higher value than PcBMT [22], whereas the constitutive BMT activity of irradiated Ruta cells far exceeded that of XMT activity [34]. The native enzymes purified from induced parsley cells were reported as stable homodimers of 67 kDa (XMT) and 73 kDa (BMT) being most active in potassium phosphate buffer of pH 7.5–8.0 (XMT) or of 8.0–8.5 (BMT).ThenativeenzymesfromRuta appeared to be larger (85 kDa for BMT and 110 kDa for XMT [34]), and their stability differed greatly in desalted crude extracts. XMT activity was lost rapidly while the BMT activity decreased at only a moderate rate [34]. In contrast, BMT from A. majus was found to be a rather labile enzyme in crude extract from plant cells or after recombinant expres- sion, which did not enable the extensive purification. Furthermore, XMT activity was induced to a negligible extent in elicited A. majus cells as compared to BMT. Nevertheless, the expression of AmBMT in E. coli yielded highly active extracts that revealed a molecular mass corresponding to that of bovine serum albumin (67 ± 5 kDa) on size exclusion chromatography calibrated with alcohol dehydrogenase, bovine serum albumin, oval- bumin, chymotrypsinogen A and ribonuclease A. Apparent K m values were determined at 2.8 l M (bergaptol) and 6.5 l M (SAM) which is in accordance with the values reported for native PcBMT (K mbergaptol 4.0 l M ; K mSAM 3.1 l M ) [22]. The cloning of AmBMT revealed a molecular mass of 38.7 kDa for the translated polypeptide, strongly suggesting a homodimer composition for the native Ammi enzyme as was shown previously for other OMTs [44–46]. The high degree of homology with heterologous COMTs prompted us to clone the AmCOMT also, which was achieved using seedlings. The alignment for these two OMT polypeptides revealed a surprisingly high degree of homology (Fig. 4). The similarity to annotated COMTs had been proposed also for the PcBMT [26], but, unfortunately, the relevant sequence data have not been published and cannot be compared. AmBMT and AmCOMT are regulated differ- ently upon elicitation of A. majus cells, because the low level of COMT transcript abundance and activity of the cells hardly changed over the time of the experiments, whereas the AmBMT transcript was transiently induced to a maximum at 7 h. More importantly, despite the homology (Fig. 4) the substrate specificities of the recombinant OMTs differed greatly. The AmBMT exclusively methylated ber- gaptol to bergapten (Fig. 1), whereas the AmCOMT accepted several substrates apart from caffeate, with a preference for 5-hydroxyferulate (Table 1). It is thus obvious that small sequence elements, in addition to the five highly conserved regions required for SAM binding [39,41], strongly affect the substrate specificity of OMTs. AmBMT is a typical member of the COMT family of enzymes which differ from the recently crystallized small-molecule methyl ester OMTs [45]. In the case of two crystallized OMTs (chalcone OMT, daidzein 7-OMT) from Medicago sativa, two such regions of 14 (motif I) and 11 amino acids (motif II) were identified and proposed to control the specificity [44]. From these and corresponding sequence elements in COMTs and flavonoid OMTs a consensus sequence was established (Fig. 4) and used to predict the substrate specificity of a novel OMT [47]. In summary, the studies suggested that substitutions of two to three amino acids in motifs I or II may provide a basis for the OMT classification, although a reliable prediction was not possible. This is reminiscent of the few amino acid substitutions reported for Clarkia OMTs to discriminate caffeate and (iso)eugenol substrates [48]. On comparison of AmCOMT and AmBMT only subtle differences in the motifs I and II were noticed due to five substitutions each (Fig. 4), most of which were conservative exchanges. However, the mutation of a single residue at other locations might considerably shift the specificity of OMTs for related substrates as has been shown for phenylpropene OMTs from sweet basil [46]. Future point mutations will reveal the relevance of these substitutions. Acknowledgements Financial support by the Deutsche Forschungsgemeinschaft and Fonds der Chemischen Industrie is gratefully acknowledged. The authors are indebted to S. Specker and S. Schreiner for fruitful discussions, to B. Rohde for chromatography and to A. Batschauer and O. Panajotow (Pflanzenphysiologie und Photobiologie, Fachb- ereich Biologie, Universita ¨ t Marburg) for their part in the subcloning experiments. References 1. Hamerski, D. & Matern, U. (1988) Elicitor-induced biosynthesis of psoralens in Ammi majus L. suspensions cultures. Microsomal conversion of demethylsuberosin into (+)-marmesin and psora- len. Eur. J. Biochem. 171, 369–375. 938 M. Hehmann et al.(Eur. J. Biochem. 271) Ó FEBS 2004 2. Hamerski, D. & Matern, U. (1988) Biosynthesis of psoralens. Psoralen 5-monooxygenase activity from elicitor-treated Ammi majus cells. FEBS Lett. 239, 263–265. 3. Tietjen, K.G., Hunkler, D. & Matern, U. (1983) Differential response of cultured parsley cells to elicitors from two non- pathogenic strains of fungi. 1. Identification of induced products as coumarin derivatives. Eur. J. Biochem. 131, 401–407. 4. Wendorff, H. & Matern, U. (1986) Differential response of cul- tured parsley cells to elicitors from two non-pathogenic strains of fungi. Eur. J. Biochem. 161, 391–398. 5. Hahlbrock, K. & Scheel, D. (1989) Physiology and molecular biology of phenylpropanoid metabolism. Annu. Rev. Plant Phy- siol.Mol.Biol.40, 347–369. 6. Hahlbrock, K., Scheel, D., Logemann, E., Nu ¨ rnberger, T., Parniske, M., Reinold, S., Sacks, W.R. & Schmelzer, E. (1995) Oligopeptide elicitor-mediated defense gene activation in cultured parsley cells. Proc. Natl Acad. Sci. USA 92, 4150–4157. 7. Matern, U. (1991) Coumarins and other phenylpropanoid compounds in the defense response of plant cells. Planta Med. 57, 15–20. 8. Matern 5 , U. & Grimmig, B. (1994) Natural phenols as stress metabolites. Natural phenols in plant resistance. Acta Hortic. 381, 448–462. 9. Matern, U., Grimmig, B. & Kneusel, R.E. (1995) Plant cell wall reinforcement in the disease resistance response: molecular com- position and regulation. Can.J.Bot.73, 511–517. 10. Nu ¨ rnberger, T., Nennstiel, D., Jabs, T., Sacks, W.R., Hahlbrock, K. & Scheel, D. (1994) High affinity binding of a fungal oligo- peptide elicitor to parsley plasma membranes triggers multiple defense responses. Cell 78, 449–460. 11. Zangerl, A.R. & Berenbaum, M.R. (1990) Furanocoumarin induction in wild parsnip: evidence for an induced defense against herbivores. Ecology 71, 1933–1940. 12. Dercks, W., Trumble, J. & Winter, C. (1990) Impact of atmo- spheric pollution alters linear furanocoumarin content in celery. J. Chem. Ecol. 16, 443–454. 13. Zobel, A.M. & Brown, S.A. (1990) Seasonal changes of furanocoumarin concentrations in leaves of Heracleum lanatum. J. Chem. Ecol. 16, 1623–1634. 14. Pathak, M.A., Parrish, J.A. & Fitzpatrick, T.B. (1981) Psoralens in photochemotherapy of skin diseases. Farmaco 36, 479–491. 15. Honigsmann, H., Jaschke, E., Gschnait, F., Brenner, W., Fritsch, P. & Wolff, K. (1979) 5-Methoxypsoralen (bergapten) in photo- chemotherapy of psoriasis. Br.J.Dermatol.101, 369–378. 16. Hann, S.K., Cho, M.Y., Im, S. & Park, Y.K. (1991) Treatment of vitiligo with oral 5-methoxypsoralen. J. Dermatol. 18, 324–329. 17. Hamerski, D., Beier, R.C., Kneusel, R.E., Matern, U. & Him- melspach, K. (1990) Accumulation of coumarins in elicitor-treated cell suspension cultures of Ammi majus. Phytochemistry 29, 1137– 1142. 18. Hamerski, D., Schmitt, D. & Matern, U. (1990) Induction of two prenyltransferases for the accumulation of coumarin phytoalexins in elicitor-treated Ammi majus cell suspension cultures. Phytochemistry 29, 1131–1135. 19. Murray, R.D.H., Me ´ ndez,J.&Brown,S.A.(1982)The Natural Coumarins: Occurrence, Chemistry and Biochemistry. Wiley, New York. 20. Stanjek, V., Miksch, M., Lu ¨ er,P.,Matern,U.&Boland,W. (1999) Biosynthesis of psoralen: mechanism of a cytochrome P450 catalyzed oxidative bond cleavage. Angew. Chem. Int. Ed. 38, 400–402. 21. Sharma, S.K., Garrett, J.M. & Brown, S.A. (1979) Separation of the S-adenosylmethionine: 5- and 8-hydroxyfuranocoumarin O-methyltransferases of Ruta graveolens L. by general ligand affinity chromatography. Z. Naturforsch. 34c, 387–391. 22. Hauffe, K.D., Hahlbrock, K. & Scheel, D. (1986) Elicitor- stimulated furanocoumarin biosynthesis in cultured parsley cells: S-adenosyl- L -methionine: bergaptol and S-adenosyl- L -methio- nine: xanthotoxol O-methyltransferases. Z. Naturforsch. 41c, 228–239. 23. Giuliano, G., Bartley, G.E. & Scolino, P.A. (1993) Regulation of carotenoid biosynthesis during tomato development. Plant Cell 5, 379–387. 24. Frick, S. & Kutchan, T.M. (1999) Molecular cloning and func- tional expression of O-methyltransferase common to isoquinoline alkaloid and phenylpropanoid biosynthesis. Plant J. 17, 329–339. 25. Dumas, B., van Doorsselaere, J., Gielen, J., Legrand, M., Fritig, B., van Montagu, M. & Inze ´ , D. (1992) Nucleotide sequence of a complementary DNA encoding O-methyltransferase from poplar. Plant Physiol. 98, 796–797. 26. Ibrahim, R.K., Bruneau, A. & Bantignies, B. (1998) Plant O- methyltransferases: molecular analysis, common signature and classification. Plant. Mol. Biol. 36, 1–10. 27. Sanger, F., Nicklen, S. & Coulsen, A.R. (1977) DNA sequencing with chain-terminating inhibitors. Proc. Natl Acad. Sci. USA 74, 5463–5467. 28. Junghanns, K.T., Kneusel, R.E., Baumert, A., Maier, W., Gro ¨ ger, D. & Matern, U. (1995) Molecular cloning and heterologous expression of acridone synthase from elicited Ruta graveolens L. cell suspension cultures. Plant Mol. Biol. 27, 681–692. 29. Luka  cin, R. & Britsch, L. (1997) Identification of strictly con- served histidine and argenine residues as part of the active site in Petunia hybrida flavanone 3b-hydroxylase. Eur. J. Biochem. 249, 748–757. 30. Laemmli, U.K. (1970) Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature 227, 680–685. 31. Lowry, O.H., Roesbrough, N.J., Farr, A.L. & Randall, R.J. (1951) Protein measurements with the Folin phenol reagent. J. Biol. Chem. 193, 265–275. 32.Sambrock,J.&Russel,D.W.(2001)Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 33.Hahlbrock,K.,Lamb,C.J.,Purwin,C.,Ebel,J.,Fautz,E.& Scha ¨ fer, E. (1981) Rapid response of suspension-cultured parsley cells to the elicitor from Phytophthora megasperma var. Sojae. Plant Physiol. 67, 768–773. 34. Thompson, H.J., Sharma, S.K. & Brown, S.A. (1978) O-Methyl- transferases of furanocoumarin biosynthesis. Arch. Biochem. Biophys. 188, 272–281. 35. Maury, S., Geoffroy, P. & Legrand, M. (1999) Tobacco O- methyltransferases involved in phenylpropanoid metabolism. The different caffeoyl-coenzyme A/5-hydroxyferuloyl-Coenzyme A 3/ 5-O-methyltransferase and caffeic acid/5-hydroxyferulic acid 3/5- O-methyltransferase classes have distinct substrate specificities and expression patterns. Plant Physiol. 121, 215–223. 36. Ebel, J., Hahlbrock, K. & Grisebach, H. (1972) Purification and properties of an o-dihydricphenol meta-O-methyltransferase from cell suspension culture of parsley and its relation to flavonoid biosynthesis. Biochim. Biophys. Acta 268, 313–326. 37. Sato, F., Tsujita, T., Katagiri, Y., Yoshida, S. & Yamada, Y. (1994) Purification and characterization of S-adenosyl- L -methio- nine: norcoclaurine 6-O-methyltransferase from cultured Coptis japonica cells. Eur. J. Biochem. 225, 125–131. 38. Morishige, T., Tsujita, T., Yamada, Y. & Sato, F. (2000) Mole- cular characterization of the S-adenosyl- L -methionine: 3¢-hyd- roxy-N-methylcoclaurine 4¢-O-methyltransferase involved in isoquinoline alkaloid biosynthesis in Coptis japonica. J. Biol. Chem. 275, 23398–23405. 39. Joshi, C.P. & Chiang, V.L. (1998) Conserved sequence motifs in plant S-adenosyl- L -methionine-dependent methyltransferases. Plant Mol. Biol. 37, 663–674. Ó FEBS 2004 Bergaptol O-methyltransferase in Ammi majus 1 (Eur. J. Biochem. 271) 939 40. Schluckebier, G., O’Gara, M., Saenger, W. & Cheng, X. (1995) Universal catalytic domain structure of AdoMet-dependent methyltransferases. J. Mol. Biol. 247, 16–20. 41. Ibrahim, R.K. (1997) Plant O-methyltransferase signatures. Trends Plant Sci. 2, 249–250. 42. Vidgren, J., Svensson, L.A. & Liljas, A. (1994) Crystal structure of catechol O-methyltransferase. Nature 368, 354–358. 43. Schubert, H.L., Blumenthal, R.M. & Cheng, X. (2003) Many paths to methyltransfer: a chronicle of convergence. Trends Bio- chem. Sci. 28, 329–335. 44. Zubieta, C., He, X Z., Dixon, R.A. & Noel, J.P. (2001) Structures of two natural product methyltransferases reveal the basis for substrate specificity in plant O-methyltransferases. Nat. Struct. Biol. 8, 271–279. 45. Zubieta, C., Ross, J.R., Koscheski, P., Yang, Y., Pichersky, E. & Noel, J.P. (2003) Structural basis for substrate recognition in the salicylic acid carboxyl methyltransferase family. Plant Cell 15, 1704–1716. 46.Gang,D.R.,Lavid,N.,Zubieta,C.,Chen,F.,Beuerle,T., Lewinsohn, E., Noel, J.P. & Pichersky, E. (2002) Characterization of phenylpropene O-methyltransferases from sweet basil: facile change of substrate specificity and convergent evolution within a plant O-methyltransferase family. Plant Cell 14, 505–519. 47. Schro ¨ der, G., Wehinger, E. & Schro ¨ der, J. (2002) Predicting the substrates of cloned plant O-methyltransferases. Phytochemistry 59,1–8. 48. Wang, J. & Pichersky, E. (1999) Identification of specific residues involved in substrate discrimination in two plant O-methyl- transferases. Arch. Biochem. Biophys. 368, 172–180. 940 M. Hehmann et al.(Eur. J. Biochem. 271) Ó FEBS 2004 . Mole- cular characterization of the S-adenosyl- L -methionine: 3¢-hyd- roxy-N-methylcoclaurine 4¢ -O-methyltransferase involved in isoquinoline alkaloid biosynthesis. Furanocoumarin biosynthesis in Ammi majus L. Cloning of bergaptol O -methyltransferase Marc Hehmann 1, *, Richard Lukac ˇ in 1, *, Halina Ekiert 2 and