Archives of Biochemistry and Biophysics 440 (2005) 54–64 www.elsevier.com/locate/yabbi
Unusual pseudosubstrate speciWcity of a novel 3,5-dimethoxyphenol O-methyltransferase cloned from Ruta graveolens L. Laura Burga a,1, Frank Wellmann a,1, Richard Lukabin a,2, Simone Witte a, Wilfried Schwab b, Joachim Schröder c, Ulrich Matern a,¤ b
a Institut für Pharmazeutische Biologie, Philipps-Universität Marburg, Deutschhausstraße 17A, D-35037 Marburg, Germany Fachgebiet Biomolekulare Lebensmitteltechnologie der Technischen Universität München, Lise-Meitner-Straße 34, D-85354 Freising, Germany c Institut für Biologie II, Biochemie der PXanzen, Universität Freiburg, Schänzlestraße 1, D-79104 Freiburg, Germany
Received 19 April 2005, and in revised form 25 May 2005 Available online 20 June 2005
Abstract A cDNA was cloned from Ruta graveolens cells encoding a novel O-methyltransferase (OMT) with high similarity to orcinol or chavicol/eugenol OMTs, but containing a serine-rich N-terminus and a 13 amino acid insertion between motifs IV and V. Expression in Escherichia coli revealed S-adenosyl-L-methionine-dependent OMT activity with methoxylated phenols only with an apparent Km of 20.4 for the prime substrate 3,5-dimethoxyphenol. The enzyme forms a homodimer of 84 kDa, and the activity was insigniWcantly aVected by 2.0 mM Ca2+ or Mg2+, whereas Fe2+, Co2+, Zn2+, Cu2+ or Hg2+ were inhibitory (78–100%). Dithiothreitol (DTT) suppressed the OMT activity. This eVect was examined further, and, in the presence of Zn2+ as a potential thiol methyltransferase (TMT) cofactor, the recombinant OMT methylated DTT to DTT-monomethylthioether. Sets of kinetic OMT experiments with 3,5-dimethoxyphenol at various Zn2+/DTT concentrations revealed the competitive binding of DTT with an apparent Ki of 52.0 M. Thus, the OMT exhibited TMT activity with almost equivalent aYnity to the thiol pseudosubstrate which is structurally unrelated to methoxyphenols. 2005 Elsevier Inc. All rights reserved. Keywords: Ruta graveolens L.; Trimethoxyphenol; 3,5-Dimethoxyphenol O-methyltransferase; Dithiothreitol S-methyltransferase; O-Methyltransferase pseudosubstrate
Numerous S-adenosyl-L-methionine-dependent Omethyltransferases (OMTs)3 have been isolated from plants methylating nucleophilic hydroxyl or carboxyl *
Corresponding author. Fax: +49 6421 282 6678. E-mail address:
[email protected] (U. Matern). 1 These authors contributed equally to the work presented. 2 Present address: Chromsystems Instruments and Chemicals GmbH, Heinburgstraße 3, D-81243 München, Germany. 3 Abbreviations used: SAM, S-adenosyl-L-methionine; DTT, 1,4-dithiothreitol; LC–MS, liquid chromatography-mass spectrometry; LSC, liquid scintillation counting; OMT, O-methyltransferase; SEC, size exclusion chromatography; TLC, thin-layer chromatography; TMT, thiol methyltransferase.; RLM-RACE, RNA-ligase mediated rapid ampliWcation of cDNA ends; Pmg, elicitor from Phytophthora sojae (formerly P. megasperma f. sp. glycinea). 0003-9861/$ - see front matter 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.abb.2005.05.026
groups of a variety of substrates [1–5]. Most of these enzymes show preferences for aromatic substrates with narrow speciWcity concerning the pattern of ring substitution [6–9], but multifunctional OMTs have also been reported [10,11]. Noteworthy in the latter instances, the OMTs accepted structurally related substrates only, i.e., various phenolics. Extensive studies were conducted on “lignin-speciWc” OMTs from various plants, and the assignment of two classes was proposed [12]: low molecular weight (23–27 kDa subunits), Mg2+-dependent OMTs which do not accept caVeic acid as substrate were grouped to class I, and OMTs (38–43 kDa subunits) methylating caVeic acid or caVeyl aldehyde and caVeyl alcohol (COMTs) independently of Mg2+ were designated
L. Burga et al. / Archives of Biochemistry and Biophysics 440 (2005) 54–64
class II. Class I enzymes were initially identiWed as caVeoyl-CoA OMTs [1] methylating the guaiacyl moiety in lignin biosynthesis [13], whereas class II COMTs were considered to methylate in situ primarily 5-hydroxyconiferaldehyde residues at a later stage in the lignin pathway, a conclusion that was supported recently by phenolic proWling of COMT-deWcient transformants [14]. An N-terminal domain was proposed to determine the speciWcity of class I OMT [15]. However, the subclassiWcation of “lignin-speciWc” OMTs on the basis of substrate speciWcities is not absolute, because an OMT accepting both caVeoyl-CoA and caVeic acid was reported from loblolly pine [16] and, more recently, a class I OMT with broader substrate speciWcity was reported from Mesembryanthemum crystallinum [17]. Furthermore, a class I OMT from Ammi majus was shown to switch substrate speciWcity towards caVeic and 5-hydroxyferulic acid when Mg2+ was replaced by Mn2+ or Co2+ in the assays [18]. Class II OMTs which are related to catechol OMT [19] appear to be even less discriminate in their choice of substrates. A number of the respective cDNAs was functionally expressed in Escherichia coli, and several of these recombinant OMTs were shown to methylate both phenylpropanoids and Xavonoids [20,21] or benzaldehydes [22]. The numerous COMT polypeptide sequences deposited in data libraries Wrmly revealed the conserved motifs required for SAM-binding [2,12,23]. However, the molecular basis of substrate speciWcity still remains to be established, because residues determining the individual substrate aYnities were assigned in only a few instances [4,24,25], and the predictive value of the alignments of class II OMT sequences is still under debate [21]. The search has recently been extended to related OMTs involved in the formation of volatile phenolics as constituents of Xoral fragrance, i.e., phenylpropenes [26– 28] and methoxybenzenes [29,30]. While in case of the phenylpropene OMTs from basil the exchange of a single amino acid residue was shown to modulate the relative speciWcities of chavicol vs. eugenol OMTs [26], the residues controlling speciWcity in methoxybenzene OMTs (orcinol OMTs and phloroglucinol OMT) remain to be elucidated. However, phylogenetic analysis revealed that class II OMTs fall into two distinct categories: the Wrst includes mainly COMTs, preferring caVeic acid as substrate, while the second category designates enzymes with a more diverse range of speciWcities [26,29]. Both chavicol and eugenol OMTs as well as orcinol and phloroglucinol OMTs belong to the second distinct category of OMTs [29]. The concerted action of the latter two enzymes is required to develop the full bouquet of rose fragrance, because 1,3,5-trimethoxybenzene as the major scent component besides 3,5dimethoxytoluene is generated from phloroglucinol via 3,5-dihydroxyanisole and 3,5-dimethoxyphenol [29]. Orcinol OMT was shown in vitro to catalyze both the
55
methylations converting 3,5-dihydroxyanisole to 1,3,5trimethoxybenzene. This study describes a novel methoxybenzene OMT from Ruta graveolens L., the common rue, which appears to be also involved in the formation of volatile 1,3,5-trimethoxybenzene but clearly diVers from orcinol OMT. Moreover, the enzyme shows a puzzling thioltransferase activity towards the pseudosubstrate dithiothreitol (DTT) which has not been reported before.
Results cDNA cloning and sequence alignments Suspension cultures of R. graveolens were previously established for the investigation of acridone alkaloid biosynthesis [31,32], and the RNA isolated from cells that had been induced for 2–4 h with crude fungal elicitor were used to clone acridone synthase cDNA [33]. The biosynthesis of acridone alkaloids requires several methylations, i.e., the N-methylation of anthranilate, and the RNA was therefore employed also for RT-PCR at rather low stringency with degenerate oligonucleotide primers designed for the ampliWcation of class II OMTs [6]. Amplicons of about 200 bp length were subcloned and sequenced for provisional veriWcation of OMT-relationship. A PCR fragment of 214 bp was Wnally chosen for the design of gene speciWc primers, and the full-size cDNA of 1590 nucleotides was generated by RACE and RLM-RACE. DNA sequencing revealed an open reading frame of 1122 bp encoding a 374 residue polypeptide of Mr 41,559. Comparison of the translated polypeptide with eight class II OMTs from data base accessions (Fig. 1) using multiple CLUSTAL-W [34] alignments revealed 61–62% sequence similarity with chavicol and orcinol OMTs at a 46–50% level of identity. A little less identity was noticed with a Xavonoid OMT from Catharanthus roseus (41%) and reticuline OMT from Papaver somniferum (36%), whereas the identity with caVeic acid OMTs was in the range of 30% only. The Wve sequence elements conserved in SAM-dependent OMTs and proposed to be involved in SAM and metal binding, mainly inferred from X-ray diVraction analysis of crystallyzed rat liver catechol or alfalfa chalcone and isoXavone OMTs [19,25], were also recognized in the Ruta polypeptide (Fig. 1, regions I–V). These elements are ubiquitously present in plant class II OMTs [2,21], and the sequence relationship of the Ruta enzyme was further corroborated by comparison with 50 methyltransferases annotated in databases using the program WU-Blast2 (EMBL, Heidelberg), which did not reveal any signiWcant sequence similarity to plant class I OMTs or S- and N-methyltransferases. Thus, the cDNA isolated from R. graveolens was provisionally assigned as RgOMT (GenBank Accession No. AY894417). Nevertheless, two peculiar features
56
L. Burga et al. / Archives of Biochemistry and Biophysics 440 (2005) 54–64
although a particular function could not be assigned. Furthermore, the R. graveolens polypeptide carries an insertion of thirteen residues (Glu308 to Thr321) separating motifs IV and V (Fig. 1) which had been proposed as a metal binding and a catalytic region in catechol OMT [19]. Finally, the phylogenetic analysis (not shown) grouped the R. graveolens polypeptide closer to orcinol OMTs from roses [29,27] than to class I COMT or COMT-related enzymes. The full-size cDNA was also used for Northern blotting to monitor the speciWc transcript abundance in R. graveolens cell cultures for 8 h following the addition of fungal elicitor, which revealed the constitutive expression of the OMT transcript only (data not shown). Functional expression and assessment of OMT activity
Fig. 1. Alignment of the polypeptide translated from R. graveolens cDNA (RgOMT) with orcinol OMT from Rosa hybrida (RhOOMT; Accession No. AAM23005), orcinol OMT from Rosa chinensis (RcOOMT; CAD29458), eugenol OMT from Ocimum basilicum (ObEOMT; AAL30424), chavicol OMT from Ocimum basilicum (ObCOMT; AAL30423), (R,S)-reticuline 7-OMT from Papaver somniferum (PsROMT; AAQ01668), caVeic acid OMT from Medicago sativa (MsCOMT; P28002), pinosylvin OMT from Pinus sylvestris (PsPOMT; [10]), and caVeic acid OMT from Chrysosplenium americanum (CaCOMT; AAA86982). The level of identity (sequence similarity) ranged from 50% (62%) with RhOOMT and RcOOMT over 47% (59%) with ObEOMT, 46% (61%) with ObCOMT, 38% (54%) with PsROMT, 30% (47%) with MsCOMT, 24% with PsPOMT and 29% with CaCOMT. The regions designated I–V represent conserved motifs [2,21,57].
distinguish this polypeptide from the plant class II OMTs reported so far: eight serine residues are present in the N-terminal stretch of 17 amino acids (Fig. 1), and the evaluation by the Signal P program [35,36] suggested a signal domain in the N-terminal 24 amino acids,
The open reading frame of the RgOMT cDNA including the stop codon was ligated into vector pQE-60, and functional expression of the enzyme was accomplished in IPTG-induced E. coli M15. Crude enzyme was obtained by extraction of the cells in 0.1 M Tris–HCl, pH 7.5, and used for the initial OMT activity assays with Sadenosyl-L-[methyl-14C]methionine. A variety of low molecular weight benzenoid phenolics was tested as potential substrates, because of the sequence relationship of the translated polypeptide to orcinol OMTs (Fig. 2), and additional assays were conducted with a large number of other phenolic compounds, e.g., cinnamic acids, phenylpropanoid esters, coumarins, stilbenes, Xavonoids, lignans, monoterpenes, and acridone alkaloids (Fig. 2). Formation of labeled product was observed only with methoxylated phenols, i.e., 3,5dimethoxyphenol, guaiacol or 3-methoxyphenol, while extracts from E. coli transformed with the empty vector and fractionated with ammonium sulfate (60–80% saturation) did not yield any product. The recombinant OMT was puriWed about 30-fold with 4.4% recovery from the crude extracts by ammonium sulfate precipitation and successive fractionation by size exclusion, anion exchange, and aYnity chromatographies (Table 1). The OMT activity of the puriWed enzyme with 3,5-dimethoxyphenol reached 12.2 kat/kg (Table 1), and the subsequent SDS–PAGE documented the apparent homogeneity revealing one protein band of about 42 kDa (Fig. 3). Additional size exclusion chromatography of the recombinant enzyme on a calibrated superose column indicated a molecular mass of roughly 84 § 5 kDa for the active methoxyphenol OMT suggesting a homodimeric composition. Upon shock freezing of the enzyme in liquid nitrogen and storage at ¡80 °C the activity remained stable for 3 weeks at least. The rates of methylation of 3,5-dimethoxyphenol were examined in various buVers ranging from pH 2.0 to 10.0 and at temperatures from 20 to 47 °C, which revealed the maximal rate at pH 7.0–7.5 with a broad temperature optimum
L. Burga et al. / Archives of Biochemistry and Biophysics 440 (2005) 54–64
57
The activity of the recombinant OMT was apparently independent of divalent cations, and the presence of Mg2+ or Ca2+ at 2.0 mM concentration in the assay only insigniWcantly aVected the activity, while Co2+ or Fe2+ suppressed the activity by 80% (Table 3) and Cu2+, Hg2+ or Zn2+ completely abolished the activity (Table 3). The eVect of Zn2+ was examined further, and the inhibition of OMT activity reached over 90, 72, and 33%, respectively, at 10.0, 5.0, or 2.0 M ZnCl2 in the assays in comparison to control incubations lacking zinc. However, the addition of 5.0 or 10.0 mM EDTA fully reversed the inhibitory eVect of 2.0 mM ZnCl2 on the OMT activity. The activity was also completely inhibited by preincubation with thiol blocking reagents, e.g., phenylmethylsulfonyl Xuoride, p-hydroxymercuribenzoate or N-ethylmaleimide, and this eVect could be partly reversed by the addition of 5 mM of 1,4-dithiothreitol (DTT) (Table 3). Characterization of TMT activity
Fig. 2. Substrate preferences of the recombinant OMT from R. graveolens.
between 33 and 42 °C (Fig. 4). Kinetic analysis of the puriWed recombinant enzyme documented the highest aYnity to 3,5-dimethoxyphenol (apparent Km 20.4 M; Km SAM 2.0 M) followed by 3-methoxyphenol and guaiacol, whereas 3,5-dihydroxyanisole was a poor substrate (Table 2). The product from semipreparative incubations of 3,5-dimethoxyphenol with the puriWed enzyme was collected for LC–MS analysis, and the retention time as well as the mass signal at m/z 213 ([M+HCOO]¡) identiWed the product as 1,3,5-trimethoxybenzene in comparison with an authentic sample. Therefore, the enzyme clearly belongs to the methoxyphenol OMTs and classiWes as 3,5-dimethoxyphenol OMT.
Assays conducted with the recombinant 3,5-dimethoxyphenol OMT in the presence of 5 mM DTT in the course of stability and inhibition studies surprisingly suggested a 24% inhibition of the OMT activity (Table 3), which is in sharp contrast to reports of the eVect of DTT on other O-methyltransferases. Therefore, DTT was assumed a potential substrate and used in sets of activity assays in the absence and in the presence of Zn2+ or other metal ions, because metal cofactors had been proposed for various methyltransferases including thiol methyltransferases [37]. For these assays, the enzyme had to be partially puriWed at least, because non-transformed E. coli expressed some minor TMT activity that was completely removed on ammonium sulfate precipitation (60–80% saturation) of the crude extracts. The Smethylating activity of the enzyme eluted from SEC (Table 1) and incubated with 5.0 mM DTT was negligible (below 0.02 kat/kg). However, the addition of 1.0 or 2.0 mM ZnCl2 to the assays greatly stimulated the TMT activity (1.9 and 2.6 kat/kg, respectively), and the new product could be separated by thin-layer chromatography. The proper choice of Zn2+ concentrations initially presented a problem, because the rate of turnover of DTT as a function of DTT concentration in the assay depended strongly on the Zn2+ concentration (Fig. 5): while at 10 M Zn2+ maximal conversion rates were achieved at about 0.1 mM DTT, the stoichiometry approached 1:1 at concentrations exceeding 0.5 mM Zn2+, suggesting an inhibition at elevated concentration as was observed previously for the 3,5-dimethoxyphenol OMT activity (Table 3). The data furthermore suggested that excess DTT completely titrated the available Zn2+ and restored the methylating activity. Semipreparative incubations were then conducted at 0.1 mM DTT with the pure recombinant R. graveolens enzyme in the
58
L. Burga et al. / Archives of Biochemistry and Biophysics 440 (2005) 54–64
Table 1 PuriWcation of recombinant R. graveolens OMT PuriWcation step
Protein (mg)
SpeciWc activity (kat/kg)
Apparent puriWcation (-fold)
Recovery (%)
Crude extract Ammonium sulfate (0.6–0.8) Fractogel EMD BioSEC (S) Fractogel EMD DEAE 650 (S) Adenosine–agarose
206.6 75.0 15.4 3.9 0.3
0.4 0.9 1.9 4.3 12.2
1 2.3 4.8 10.8 30.5
100 81.7 35.4 20.3 4.4
The enzyme was expressed in IPTG-induced E. coli M15, and the cells (14.0 g wet pellet) were extracted in 0.1 M Tris–HCl buVer, pH 7.5. The OMT activity was measured with 3,5-dimethoxyphenol under standard assay conditions.
Fig. 3. SDS–PAGE separation of the R. graveolens OMT expressed in E. coli. The enzyme was puriWed from crude bacterial extract to apparent homogeneity by ammonium sulfate precipitation (60–80% saturation) and successive size exclusion, anion exchange and aYnity chromatographies (Table 1). The crude extract (19.5 g protein, lane 1) or the protein eluted from Fractogel EMD BioSEC (S) column (6.5 g, lane 2) , Fractogel DEAE column (1.4 g, lane 3), and adenosine–agarose (1.0 g, lane 4) was separated on a 5% stacking and 12.5% separation gel, and staining of the gels was accomplished with Coomassie brilliant blue R250. Molecular mass markers (lane M) are indicated in the left margin.
presence of 10.0 M ZnCl2, and the product was collected for LC–MS analysis. The retention time, in comparison to an authentic sample, and the mass signal at m/ z 167 ([M¡H]¡) identiWed the product as dithiothreitol monomethylthioether. The reference sample was synthesized and unequivocally characterized by 1H and 13C NMR and mass spectrometry (data not shown). Thus, the pure R. graveolens OMT is capable of methylating also the pseudosubstrate DTT proving TMT activity (27.1 kat/kg). Furthermore, the pH (7.0–7.5 max.) and temperature dependencies of the turnover rates fully matched the pattern determined for the OMT activity (Fig. 4). From a few additional thiol compounds tested only 2-mercaptoethanol served as substrate with low aYnity, while methionine, homocysteine, cysteine, 6mercaptopurine or the fungicide zinc-ethylene-bisdithiocarbamate was not accepted.
Fig. 4. The relative conversion rates of 3,5-dimethoxyphenol to 1,3,5trimethoxybenzene were determined at 36 °C in various 0.1 M buVers over a pH range from 2.0 to 10.0 (A), and the O-methyltransferase activity was measured in 0.1 Tris–HCl buVer, pH 7.5, using a range of temperatures from 20 to 46 °C (B). Table 2 Substrate aYnities of recombinant R. graveolens OMT Substrates
Apparent Km (M)
Apparent Vmax (nmol s¡1)
3,5-Dimethoxyphenol 3-Methoxyphenol Guaiacol 3,4-Dimethoxyphenol 2,3-Dimethoxyphenol 3,5-Dihydroxyanisole SAM
20.4 44.6 66.6 85.3 275.8 346.2 2.0
7.1 1.5 0.9 3.4 2.3 1.4 0.9
Competition of OMT and TMT activities The bifunctionality of the recombinant methyltransferase cloned from R. graveolens and the fact that the
OMT activity was apparently inhibited in the presence of Zn2+/DTT suggested equivalent substrate binding sites for the enzyme. Therefore, the inhibition kinetics of
L. Burga et al. / Archives of Biochemistry and Biophysics 440 (2005) 54–64
59
Table 3 Inhibition of 3,5-dimethoxyphenol OMT activity Inhibitor in the standard assay
Concentration (mM)
Relative activity (%)
None 1,4-Dithiothreitol (DTT) 2-Mercaptoethanol Zn2+, Cu2+ or Hg2+ Co2+ or Fe2+ Mn2+, Ca2+ or Fe3+ Mg2+ Phenymethylsulfonyl Xuoride (PMSF) PMSF + DTTa p-Hydroxymercuribenzoate (pHMB) pHMB + DTTa N-Ethylmaleimide (NEM) NEM + DTTa
— 5 50 2 2 2 2 5
100 76 80 ND 22 74 85 ND
5/5 0.1
1.3 ND
0.1/5 0.5 0.5/5
38 ND 60
The partially puriWed recombinant R. graveolens enzyme, eluted from SEC, was preincubated for 15 min with the potential inhibitor and Sadenosyl-L-[methyl-14C]methionine in the absence of substrate. ND, not detected. a In case of pHMB, PMSF or NEM this was followed by the addition of DTT and another preincubation for 15 min. The OMT incubations were started subsequently by the addition of 3,5-dimethoxyphenol, and activities were related to the methylating activity for 3,5-dimethoxyphenol in the absence of inhibitor (1.9 kat/kg, 100%).
Fig. 6. Kinetic analysis of the eVect of various Wxed ZnCl2/DTT concentrations on the turnover rate of the recombinant R. graveolens OMT with 3,5-dimethoxyphenol as the variable substrate. The standard OMT assay was conducted in the absence of ZnCl2/DTT (symbol solid square) or in the presence of ZnCl2/DTT Wxed at concentrations of 2.0/20.0 M (solid triangle), 5.0/50.0 M (cross) or 10.0/100.0 M (open square). The apparent Km values inferred from the Lineweaver– Burk diagram plotted vs. the concentration of DTT (inset) revealed an apparent Ki of 52.0 M.
partially puriWed recombinant enzyme (eluted from size exclusion chromatography). The concentration of substrate (3,5-dimethoxyphenol) was varied from 14.0 to 100.0 M. Apparent Km values from 20.4 (in the absence of Zn2+/DTT) up to 57.8 M (in the presence of the highest Zn2+/DTT concentration) were recorded, and double-reciprocal replotting indicated an apparent Ki of 52.0 M for DTT (Fig. 6). The absolute value was most likely lower because probably only the Zn-chelated DTT is used as substrate. In any case, the shift in apparent Km and the constant Vmax suggested that 3,5-dimethoxyphenol and DTT compete for the same substrate binding site of the R. graveolens enzyme. Tissue distribution of OMT and TMT activities
Fig. 5. Relative TMT activity of the recombinant R. graveolens OMT determined with 1,4-dithiothreitol as substrate and S-adenosyl-L[methyl-14C]methionine (7.9 Ci mol¡1) in the presence of various concentrations of zinc chloride. The maximal turnover rate (100%) corresponds to 3.7 kat/kg. The product was extracted from the incubation with ethylacetate and quantiWed by liquid scintillation counting. Control thin-layer chromatography on silica gel revealed only one product corresponding to 1,4-dithiothreitol monomethylthioether.
3,5-dimethoxyphenol OMT activity was investigated in sets of experiments in the absence or in the presence of ZnCl2/DTT-combinations of 2.0/20.0, 5.0/50.0, and 10.0/ 100.0 M, respectively (Fig. 6). In these combinations, the DTT chelation of Zn2+ prevails avoiding the metalinhibition of OMT activity. The assays were conducted at a saturating concentration of S-adenosyl-L-[methyl14 C]methionine (11.54 M; 52 Ci mol¡1) in 0.1 M Tris– HCl buVer, pH 7.5 (0.1 ml total), and using 50.0 g of the
The physiological function of 3,5-dimethoxyphenol OMT is likely connected with the scent production, and the relative tissue distribution of the methyltransferase activities was examined in outdoor grown adult R. graveolens plants and compared to the expression in suspension culture (Fig. 7). Obviously, the 3,5-dimethoxyphenol OMT activity was predominantly present in the stem extracts. However, the patterns of methylating activities for 2,3-dimethoxyphenol (primarily expressed in leaf extract), guaiacol and 3,5-dihydroxyanisole (leaf extracts) or 3,4-dimethoxyphenol (leaf and stem extracts), as well as the speciWc activities of the crude extracts related to guaiacol OMT (100%), suggested that more than one OMT must be responsible for these methylations (Fig. 7). This is also to be inferred for the TMT activities, because the speciWc activity of Xower extracts for 2-mercaptoethanol exceeded that for DTT,
60
L. Burga et al. / Archives of Biochemistry and Biophysics 440 (2005) 54–64
Fig. 7. Relative OMT and TMT activities in crude extracts of diVerent tissues of R. graveolens plants and cell cultures determined in standard assays (0.1 M Tris–HCl buVer, pH 7.5 at 36 and 38 °C, respectively, in the presence of 25.0 M labeled SAM) using various substrates. The highest conversion rate observed (8.4 kat/kg OMT activity with guaiacol in leaf extract) was arbitrarily set at 100%.
and cell cultures showed almost equivalent activities with both substrates (Fig. 7), whereas the OMT cloned from R. graveolens exhibited relatively low activity towards 2-mercaptoethanol (Table 2). For comparison, the relative abundance of the 3,5-dimethoxyphenol OMT transcript in the outdoor grown plants was examined by RT-PCR (Fig. 8) and revealed the expression in all the Ruta tissues with high template activity in stem and developing Xowers. R. graveolens plants are known to develop an intensive scent, and it is conceivable that more than one OMT contribute volatile methoxybenzenes, in addition to trimethoxybenzene, as documented by the preferred methylation of guaiacol or 2,3-dimethoxyphenol in the crude plant extracts.
Discussion Methoxylated benzenes, particularly 1,3,5-trimethoxybenzene, and toluenes, besides mono- and sesquiterpenes, are the predominant components of the scent of roses, and the fragrance biochemistry of this important Xoricultural crop has been the subject of genomic approaches [28]. At least two types of OMTs are essential for the synthesis of trimethoxybenzene from phloroglucinol [38], which have been studied from roses only [27,29,30]. Orcinol (3,5-dihydroxytoluene or 5-methyl-
Fig. 8. RT-PCR ampliWcation of dimethoxyphenol OMT from template of various R. graveolens tissues. The tissues (young leaf from the plant tip and adult leaf from a more basipetal region; top quarter stem segment; emerging Xowers) were harvested from outdoor grown adult plants (about 1.4 m tall). Equivalent aliquots of cDNA generated from total RNA (about 5 g per tissue) were heated for 2 min at 95 °C and ampliWed subsequently with end-to-end oligonucleotide primers using (lane 1) 15, (lane 2) 30 or (lane 3) 35 cycles of 0.5 min at 95 °C followed by 1.0 min at 65 °C and 2 min at 72 °C. All samples were Wnally heated to 72 °C for 2 min prior to separation on 1.5% agarose (OMT at 1122 bp). The total RNA contents of the various tissue extracts (triple the amount as used for cDNA synthesis) are shown for comparison (bottom).
1,3-benzenediol) OMTs catalyze the successive methylations of orcinol to 3,5-dimethoxytoluene as well as those from 3,5-dihydroxyanisole (5-methoxy-1,3-benzenediol) via 3,5-dimethoxyphenol to 1,3,5-trimethoxybenzene, although the relative activity for the latter reaction is rather low [29]. A separate phloroglucinol OMT is required for the Wrst methylation step [30]. It is unlikely that such volatiles are restricted to roses, recalling also the multifunctionality of some OMTs [6,26]. Plants of the Rutaceae family are known to produce aromatic volatiles in nearly all tissues, and the data presented in this report clearly demonstrate the expression of an enzyme in R. graveolens plants (Figs. 7 and 8) and cell cultures highly homologous to orcinol OMTs. This enzyme, however, most eYciently methylated 3,5-dimethoxyphenol to 1,3,5-trimethoxybenzene, whereas 3,5-dihydroxyanisole was a rather poor substrate (Table 2). Therefore, the R. graveolens OMT clearly diVers in substrate speciWcity from rose orcinol OMTs, showing a rather narrow speciWcity (Fig. 2), and represents a novel methoxybenzene OMT. Irrespective of these diVerences, the polypeptides of R. graveolens OMT (Fig. 3) and orcinol OMTs revealed the highest homology (about 62%) among the annotated sequences, and the most obvious diVerence is a 13 amino acid insertion in the C-terminal portion separating motifs IV and V (Fig. 1). This insertion contains four glutamic and two aspartic acid residues besides three lysine residues which conceivably contribute to the surface charge and modify folding of the enzyme polypeptide. The precise eVect of the insertion on the substrate speciWcity remains to be established.
L. Burga et al. / Archives of Biochemistry and Biophysics 440 (2005) 54–64
While the inhibition of enzymes by thiol reagents (Table 3) is a common phenomenon, thiols such as 2mercaptoethanol or DTT are often included in preparative procedures to preserve enzyme activities on puriWcation and storage [39]. The apparent suppression of the R. graveolens OMT activity by low DTT concentrations was therefore rather puzzling and indicated the possibility of substrate competition. Alkylthiol S-methyltransferases involved in biotransformation and detoxiWcation processes have been reported from bacterial and mammalian sources, and these studies attributed a role to Zn2+ as cofactor [39–41]. Much less is known about Smethyltransferases from plants, primarily concerning the methylation of methionine or homocysteine [42–45]. A novel class of thiol methyltransferases (TMTs) was reported only recently from Brassica oleracea L. producing sulfur volatiles and thought to be involved in glucosinolate metabolism [46,47]. These monomeric enzymes of about 25 kDa were capable of methylating HS¡, thiocyanate, and some organic thiols including thiophenol and thiosalicylic acid. However, the Brassica TMTs do not methylate methionine or homocysteine, lack N- or O-methyltransferase activity and do not share any signiWcant sequence similarity with such proteins [47]. In the presence of Zn2+, the recombinant R. graveolens OMT eYciently catalyzed the methylation of DTT, and the product was unequivocally identiWed as the monomethylthioether. The enzyme also accepted 2mercaptoethanol with lower aYnity, but did not methylate methionine or homocysteine, which diVers from all OMTs and TMTs reported so far. Such a surrogate activity has never been described for an OMT and was therefore investigated further. Dual functions can be ascribed to Zn2+ in these assays, which are likely the reason for the observed anomalous Zn2+/DTT stoichiometry: the inhibitory eVect of Zn2+ on the OMT activity suggests that the metal interacts with the enzyme polypeptide, presumably via Asp, Glu, His, and Cys residues as the most common zinc ligands [41], which may even cause the dissociation of the homodimeric enzyme. It must be emphasized that the translation of the recombinant R. graveolens OMT terminated at the endogenous stop codon lacking a His tag, thus excluding that additional metal binding was introduced. The enzyme belongs to class II OMTs which act independently of divalent cations but contain the conserved metal binding motifs known from class I OMTs [19,2]. In the TMT assays, the Lewis acid zinc coordinates also to the substrate DTT shifting the thiol–thiolate equilibrium at near neutral pH to the thiolate and thus activating the thiol for nucleophilic attack [37,48,49]. Model reactions for Zn-dependent TMTs clearly documented that the transfer becomes eYcient, when the alkyl thiol is able to chelate Zn2+, and that the dissociated thiolate anion is the reacting species [37,48,49]. Furthermore, the thioether product did not remain coordinated to zinc [37].
61
One would expect a 1:1 reaction of zinc and DTT, although this may change with protein binding [50]. Our TMT activity assays with DTT in the presence of increasing zinc-concentrations (Fig. 5) suggested that DTT essentially reversed the inhibitory eVect of Zn2+ on enzyme activity by chelating titration before the thiolate substrate was accepted; in support of this assumption, suppression of the TMT activity was observed upon the addition of Zn2+ in excess at any DTT concentration. The fact that competitive kinetics were observed for Zn– DTT inhibition with respect to 3,5-dimethoxyphenol in the OMT assay (Fig. 6) indicated that the TMT active site is close or at the OMT catalytic centre. This Wnding is surprising, although competitive zinc–DTT inhibition was reported also with ribonucleotides in rho transcription termination assays determining the poly(dC)ribo(C)10-dependent ATPase activity [50].
Conclusion A novel methoxyphenol OMT was cloned from R. graveolens and shown to be expressed in leaf, stem, and Xower tissues. The enzyme is related to plant class II OMTs with highest homology to orcinol OMTs from roses. The recombinant enzyme was capable of methylating the pseudosubstrate DTT to the monomethylthioether and thus possessing TMT activity. The eYcient methylation of DTT required the presence of Zn2+, and competitive kinetics of inhibition suggested that the OMT and TMT activities rely on the same active site. These Wndings may indicate broader substrate speciWcities and functions of plant class II OMTs than anticipated from standard in vitro assays, although at present no physiological relevance can be assigned to the TMT activity.
Materials and methods Chemicals and enzymes Substrates and Wne biochemicals of analytic grade were purchased from Roth (Karlsruhe, Germany) and Sigma (Deisenhofen, Germany) or were from the lab collection. Restriction enzymes, vectors, and E. coli host strains were obtained from Qiagen (Hilden, Germany), MBI-Fermentas (St. Leon-Rot, Germany), Promega (Mannheim, Germany), and Stratagene (Heidelberg, Germany). S-Adenosyl-L-[methyl-14C]methionine (52 Ci mol¡1) was purchased from Hartmann Analytic (Braunschweig, Germany). Acridone alkaloids were extracted from suspension-cultured R. graveolens cells (80 g) induced for 16 h with Pmg elicitor; the cells were disrupted with quartz sand in 0.1 M Tris–HCl buVer, pH 7.5 (100 ml), and the alkaloids were extracted in
62
L. Burga et al. / Archives of Biochemistry and Biophysics 440 (2005) 54–64
trichloromethane [32] from the clear supernatant after centrifugation (30,000g for 10 min at 4 °C). Lignans were kindly provided by Prof. M. Petersen, Institut für Pharmazeutische Biologie, Philipps-Universität Marburg. Plant material Ruta graveolens L. plants were harvested from the local Botanical Gardens. The stems, fully developed leaves as well as emerging young leaves and Xowers were collected separately in May 2004, immediately frozen, ground to Wne powder in liquid nitrogen, and stored at ¡80 °C until use. Cell suspension cultures of R. graveolens, cell line R-20 [31], were propagated in M20 medium [51,52]. The cells were propagated and harvested as described elsewhere [53]. For induction, Pmg elicitor was added to 7-day-old cell cultures, as described previously [8]. The cells were incubated for 3 h, subsequently harvested, frozen in liquid nitrogen and kept at ¡80 °C until use. RNA isolation, PCR ampliWcation, and cloning Total RNA was extracted from ground frozen Pmg elicitor-induced cells (200 mg) according to Giuliano et al. [54]. Degenerate oligonucleotide primers described for the cloning of OMTs involved in isoquinoline alkaloid biosynthesis [6] were employed for RT-PCR using the R. graveolens RNA as template. The cDNA fragments were cloned and sequenced, and full-size cDNAs were generated by 3⬘-/5⬘-RACE and RLM-RACE using gene speciWc primers. The OMT related sequences were ampliWed and ligated into the pCR2.1-TOPO vector. AmpliWcation of the OMT transcript from various plant tissues was accomplished with end-to-end oligonucleotide primers, 5⬘-CACCATGGCGATGGCTAATGGA TC-3⬘ (forward) and 5⬘-GGGATAAGCTTCAATGA GTGACC-3⬘ (reverse), using total RNA as template. Heterologous expression and puriWcation The R. graveolens cDNA was modiWed by the introduction of NcoI and BamHI sites, respectively, adjacently before the start codon and after the stop codon, allowing the correct termination, using adequately designed oligonucleotide primers with PCR ampliWcation. Two internal NcoI restriction sites were removed using QuikChange Multi Site-Directed mutagenesis Kit as described by the manufacturer (Stratagene, Heidelberg, Germany), without altering the amino acid sequence. The NcoI/BamHI fragment was cloned in pQE-60 vector (Qiagen, Hilden, Germany) and expressed in E. coli strain M15 (Qiagen, Hilden, Germany) harbouring the plasmid pRep4. The incubation was carried out at 37 °C with shaking at 220 rpm for 3 h, and the expression was induced by adding IPTG to a
Wnal concentration of 1 mM. The cells were harvested, disrupted by ultrasonication (0.1 M Tris–HCl buVer, pH 7.5), and the debris was removed by centrifugation (30,000g for 10 min). All extraction and puriWcation steps were carried out at 4 °C. The crude bacterial extract containing the OMT activity was fractionated by ammonium sulfate precipitation, and the protein precipitated at 60–80% saturation was separated further by size exclusion chromatography (SEC) on Fractogel EMD BioSEC (S) (Merck, Darmstadt, Germany) in 0.1 M Tris–HCl buVer, pH 7.5. Active fractions were combined and loaded on a Fractogel EMD DEAE 650 (S) anion exchange column (Merck, Darmstadt, Germany), and bound protein was eluted in 0.1 M Tris–HCl buVer, pH 7.5, using a linear NaCl gradient from 0 to 1.0 M at a Xow rate of 1 ml/min. Active fractions were pooled, desalted through a PD10 column in 50 mM Tris–HCl buVer, pH 8.5, and applied to a 1 ml adenosine–agarose column prepared from adenosine 5⬘monophosphate–agarose (Sigma, Deisenhofen, Germany) by dephosphorylation with bovine alkaline phosphatase [42,55] equilibrated with the same buVer. After loading the sample, the column was washed successively with equilibration buVer (2.0 ml) and 50 mM Tris–HCl buVer, pH 8.5, containing 0.2 M NaCl (2 ml). The bound protein was eluted with 50 mM Tris–HCl buVer, pH 8.5, containing 0.2 mM NaCl and 2.0 mM SAM (10.0 ml) at a Xow rate of 0.5 ml/min. All procedures were performed at 4 °C, and protein amounts were quantiWed according to Lowry et al. [56]. Methyltransferase assays The OMT assays were routinely carried out at 36 °C in 0.1 M Tris–HCl buVer, pH 7.5 (200 l total) containing 0.25 mM substrate. The incubation was started by the addition of 25.0 M S-adenosyl-L-[methyl14 C]methionine (4.72 Ci mol¡1; 55,500 dpm) and continued for 30 min. The reaction was stopped by addition of 0.2 N HCl (20 l), and the products were extracted in 400 l ethyl acetate. Aliquots of the organic phase (200 l) were measured by liquid scintillation counting (LSC) in a 1214 Rackbeta counter (PerkinElmer, Wellesley, MA, USA). The product from OMT incubations with 3,5-dimethoxyphenol was identiWed by chromatography on silica thin-layer plates (Merck, Darmstadt, Germany) in toluene:acetic acid (4:1, v/v) and co-chromatography with authentic 1,3,5trimethoxybenzene (Rf 3,5-dimethoxyphenol 0.46; Rf 1,3,5-trimethoxybenzene 0.58). The compounds were spotted under 280 nm irradiation or by a BioImager Analyzer FLA2000 (FujiWlm, Japan). Kinetic measurements were performed under pseudo-Wrst-order conditions using partially puriWed enzyme eluted from SEC chromatography (0.1 mg protein). The apparent Km values were recorded in the presence of 0.25 mM of 3,5-
L. Burga et al. / Archives of Biochemistry and Biophysics 440 (2005) 54–64
dimethoxyphenol or at a saturating concentration of SAM (25.0 M) and examined by Lineweaver–Burk plots. Control incubations were carried out for background corrections omitting the substrate or the enzyme or using heat-denatured protein (95 °C for 5 min), and all incubations were performed in duplicate at least. The S-methyltransferase assays were conducted with the homogeneous or partially puriWed recombinant R. graveolens OMT, eluted from SEC, at 38 °C in 0.1 M Tris–HCl buVer, pH 7.5 (200 l total) containing 5.0 mM (i.e., DTT) or 50.0 mM (2-mercaptoethanol) substrate in the presence of 2.0 mM ZnCl2. The incubation was started by the addition of 25.0 M S-adenosyl-L-[methyl14 C]methionine (4.72 Ci mol¡1; 55,500 dpm) and continued for 30 min. The reaction was stopped, extracted, and aliquots were measured by LSC as described for OMT assays. Alternatively, 100 l of the extract was applied to silica gel plates, and the plates were developed in toluene:ethylacetate (3:2, v/v) for DTT or 2-mercaptoethanol as substrate (Rf DTT-monomethylthioether 0.27). In case of L-methionine or homocysteine, aliquots (30 l) of the aqueous incubation were applied to cellulose thinlayer plates and separated in n-butanol:acetic acid:water (3:1:1, v/v/v) or n-propanol:formic acid:water (20:1:5, v/ v/v) [44] (Rf S-methylmethionine 0.93; Rf SAM 0.21). The radioactivity was located by a BioImager Analyzer FLA-2000 (FujiWlm, Japan). Competition assays of the OMT activity were conducted with 3,5-dimethoxyphenol (from 14.0 to 100.0 M) at 36 °C in 0.1 M Tris–HCl buVer, pH 7.5 (100 l total) containing 11.54 M S-adenosyl-L-[methyl14 C]methionine (52 Ci mol¡1) and in the presence of mixtures of ZnCl2/DTT Wxed at concentrations of 2.0/ 20.0, 5.0/50.0 or 10.0/100.0 M. The reactions were started by the addition of the partially puriWed OMT eluted from SEC (0.05 mg) and stopped after 30 min by adding 0.2 N HCl (50 l). The incubation mixture was extracted twice with 200 l ethyl acetate each, the combined organic phase was concentrated under vacuum and applied to silica thin-layer plates, and the plates were developed in toluene:acetic acid (4:1, v/v) and examined subsequently by Bioimaging. Mass spectrometry Products were collected from 200 standard OMT or TMT assays employing 3,5-dimethoxyphenol (0.25 mM) and Zn2+/DTT (0.01:0.1 mM), respectively, and subjected to LC–MS analysis. The system used for LC–MS analysis consisted of a Bruker esquire 3000 plus mass spectrometer, equipped with an Agilent 1100 HPLC system, composed of an Agilent 1100 quaternary pump and an Agilent 1100 variable wavelength detector. The column was a Eurosphere C18 column, particle size 5 m, 10 £ 2 mm (Grom Analytik
63
& HPLC GmbH, Rottenburg, Germany). The ionisation parameters were as follows. The voltage of the capillary was 3079 V and the end plate was set to ¡500 V. The capillary exit was ¡102.5 V and the Octopole RF amplitude 119.1 Vpp. The temperature of the dry gas (N2) was 300 °C at a Xow of 5.1 min¡1. The full scan mass spectra were measured from m/z 50 to 500 until the ICC target reached 20,000 or 200 ms, whichever was reached Wrst. Tandem mass spectrometry was performed using helium as the collision gas, and the collision energy was set at 1.00 V. All mass spectra were acquired in the negative ionisation mode. The LC parameters were from 0% acetonitrile and 100% water (acidiWed with 0.05% formic acid) to 50% acetonitrile and 50% acidic water in 35 min, then in 2.5 min to 100% acetonitrile, and kept for 2.5 min at these conditions, Wnally back to 100% water and 0% acetonitrile in 5 min at a Xow rate of 0.2 ml. The detection wavelength was 220 nm. Authentic DTT-monomethylthioether was synthesized from equimolar amounts of DTT and methyliodide at pH 10.0, and the structure was conWrmed by 1H, 13C NMR and LC–MS data (manuscript in preparation).
Northern blotting Crude Pmg elicitor was added to cell suspension cultures of R. graveolens [8], the cells were harvested every 30 min for 8 h following the addition, and the total RNA was isolated. The RNA (5 g per time point) was denaturated in 0.5£ Mops buVer, pH 7.0, containing 50% formamide and 2.2 M formaldehyde for dot-blot analysis as described previously [8] on Hybond-N+ nylon membrane (Amersham Biosciences, Freiburg). Noninduced cell cultures served as controls. Labeled OMT cDNA (1300 bp, released from pQE-60 vector by NcoI/ BamHI digestion) was generated as described by Hehmann et al. [8] and used as probe. The hybridized Wlter was washed twice with 2£ SSC, 0.1% SDS buVer for 20 min at room temperature and 20 min at 68 °C, and exposed to BioImaging.
Acknowledgments We thank our colleague, Prof. M. Petersen, for providing the various authentic lignans. This work was supported by grants from the Deutsche Forschungsgemeinschaft and Fonds der Chemischen Industrie.
References [1] D. Schmitt, A.E. Pakusch, U. Matern, J. Biol. Chem. 266 (1991) 17416–17423.
64
L. Burga et al. / Archives of Biochemistry and Biophysics 440 (2005) 54–64
[2] R.K. Ibrahim, A. Bruneau, B. Bantignies, Plant Mol. Biol. 36 (1998) 1–10. [3] N. Lavid, W. Schwab, E. Kafkas, M. Koch-Dean, E. Bar, O. Larkov, U. Ravid, E. Lewinsohn, J. Agric. Food Chem. 50 (2002) 4025–4030. [4] C. Zubieta, J.R. Ross, P. Koscheski, Y. Yang, E. Pichersky, J.P. Noel, Plant Cell 15 (2003) 1704–1716. [5] M.B. Pott, F. Hippauf, S. Saschenbrecker, F. Chen, J. Ross, I. Kiefer, A. Slusarenko, J.P. Noel, E. Pichersky, U. EVmert., B. Piechulla, Plant Physiol. 135 (2004) 1946–1955. [6] S. Frick, T. Kutchan, Plant J. 17 (1999) 329–339. [7] A. Ounaroon, G. Decker, J. Schmidt, F. Lottspeich, T.M. Kutchan, Plant J. 36 (2003) 808–819. [8] M. Hehmann, R. Lukaçin, H. Ekiert, U. Matern, Eur. J. Biochem. 271 (2004) 932–940. [9] G. Schröder, E. Wehinger, R. Lukaçin, F. Wellmann, W. Seefelder, W. Schwab, J. Schröder, Phytochemistry 65 (2004) 1085–1094. [10] H. Chiron, A. Drouet, A.C. Claudot, C. Eckerskorn, M. Trost, W. Heller, D. Ernst, H. Sandermann, Plant Mol. Biol. 44 (2000) 733–745. [11] M. Wein, N. Lavid, S. Lunkenbein, E. Lewinsohn, W. Schwab, R. KaldenhoV, Plant J. 31 (2002) 755–765. [12] C.P. Joshi, V.L. Chiang, Plant Mol. Biol. 37 (1998) 663–674. [13] Z.H. Ye, R.E. Kneusel, U. Matern, J.E. Varner, Plant Cell 6 (1994) 1427–1439. [14] K. Morreel, J. Ralph, F. Lu, G. Goeminne, R. Busson, P. Herdewijn, J.L. Goeman, J. Van der Eycken, W. Boerjan, E. Messens, Plant Physiol. 136 (2004) 4023–4036. [15] T. Vogt, FEBS Lett. 561 (2004) 159–162. [16] L. Li, J.L. Popko, X.H. Zhang, K. Osakabe, C.S. Tsai, C.P. Joshi, V.L. Chiang, Proc. Natl. Acad. Sci. USA 94 (1997) 5461–5466. [17] M. Ibdah, X.H. Zhang, J. Schmidt, T. Vogt, J. Biol. Chem. 278 (2003) 43961–43972. [18] R. Lukacin, U. Matern, S. Specker, T. Vogt, FEBS Lett. 577 (2004) 367–370. [19] J. Vidgren, L.A. Svensson, A. Liljas, Nature 368 (1994) 3354–3358. [20] A. Gauthier, P.J. Gulick, R.K. Ibrahim, Arch. Biochem. Biophys. 351 (1998) 243–249. [21] G. Schröder, E. Wehinger, J. Schröder, Phytochemistry 59 (2002) 1–8. [22] P. Kota, D. Guo, C. Zubieta, J. Noel, R.A. Dixon, Phytochemistry 65 (2004) 837–846. [23] G. Schluckebier, M. O’Gara, W. Saenger, X. Cheng, J. Mol. Biol. 247 (1995) 16–20. [24] J. Wang, E. Pichersky, Biochem. Biophys. 368 (1999) 172–180. [25] C. Zubieta, X.Z. He, R.A. Dixon, J.P. Noel, Nat. Struct. Biol. 8 (2001) 271–279. [26] D.R. Gang, N. Lavid, C. Zubieta, F. Chen, T. Beuerle, E. Lewinsohn, J.P. Noel, E. Pichersky, Plant Cell 14 (2002) 505–519. [27] N. Lavid, J. Wang, M. Shalit, I. Gutermann, E. Bar, T. Beuerle, N. Menda, S. ShaWr, D. Zamir, Z. Adam, A. Vainstein, D. Weiss, E. Pichersky, E. Lewinsohn, Plant Physiol. 129 (2002) 1899–1907. [28] I. Guterman, M. Shalit, N. Menda, D. Piestun, M. Dafny-Yelin, G. Shalev, E. Bar, O. Davydov, M. Ovadis, M. Emanuel, J. Wang, Z. Adam, E. Pichersky, E. Lewinsohn, D. Zamir, A. Vainstein, D. Weiss, Plant Cell 14 (2002) 2325–2338.
[29] G. Scalliet, N. Journot, F. Jullien, S. Baudino, J.L. Magnard, S. Channeliere, P. Vergne, C. Dumas, M. Bendahmane, J.M. Cock, P. Hugueney, FEBS Lett. 523 (2002) 113–118. [30] S. Wu, N. Watanabe, S. Mita, H. Dohra, Y. Ueda, M. Shibuya, Y. Ebizuka, Plant Physiol. 135 (2004) 95–102. [31] A. Baumert, I.N. Kuzovkina, G. Krauss, M. Hieke, D. Gröger, Cell Rep. 1 (1982) 168–171. [32] A. Baumert, W. Maier, B. Schuhmann, D. Gröger, J. Plant, Physiology 139 (1991) 224–228. [33] K.T. Junghanns, R.E. Kneusel, D. Gröger, U. Matern, Phytochemistry 49 (1998) 403–411. [34] J.D. Thompson, D.G. Higgins, T.J. Gibson, Nucleic Acids Res. 22 (1994) 4673–4680. [35] H. Nielsen, J. Engelbrecht, S. Brunak, G. von Heijne, Protein Eng. 10 (1997) 1–6. [36] J.D. Bendtsen, H. Nielsen, G. von Heijne, S. Brunak, J. Mol. Biol. 340 (2004) 783–795. [37] R.G. Matthews, C.W. Goulding, Curr. Opin. Chem. Biol. 1 (1997) 332–339. [38] S. Wu, N. Watanabe, S. Mita, Y. Ueda, M. Shibuya, Y. Ebizuka, J. Biosci. Bioeng. 96 (2003) 119–128. [39] M. Fujioka, Int. J. Biochem. 24 (1992) 1917–1924. [40] M. Krüer, M. Haumann, W. Meyer-Klaucke, R.K. Thauer, H. Dau, Eur. J. Biochem. 269 (2002) 2117–2123. [41] B. González, M.A. Pajares, M. Martinez-Ripoll, T.L. Blundell, J. Sanz-Aparicio, J. Mol. Biol. 338 (2004) 771–782. [42] F. James, K.D. Nolte, A.D. Hanson, J. Biol. Chem. 270 (1995) 22344–22350. [43] B. Neuhierl, M. Thanbichler, F. Lottspeich, A. Bock, J. Biol. Chem. 274 (1999) 5407–5414. [44] F. Bourgis, S. Roje, M.L. Nuccio, D.B. Fisher, M.C. Tarczynski, C. Li, C. Herschbach, H. Rennenberg, M.J. Pimenta, T.L. Shen, D.A. Gage, A.D. Hanson, Plant Cell 11 (1999) 1485–1497. [45] P. Ranocha, S.D. McNeill, M.J. Ziemak, C. Li, M.C. Tarczinski, A.D. Hanson, Plant J. 25 (2001) 575–584. [46] J. Attieh, R. Djiana, P. Koonjul, C. Étienne, S.A. Sparace, H.S. Saini, Plant Mol. Biol. 50 (2002) 511–521. [47] J. Attieh, S.A. Sparace, H.S. Saini, Arch. Biochem. Biophys. 380 (2000) 257–266. [48] M. Machuqueiro, M. Darbre, J. Inorg. Biochem. 94 (2003) 193–196. [49] A. Krezel, W. Lesniak, J. Jezowska-Bojczuk, P. Mlynarz, J. Brasun, H. Kozlowski, W. Bal, J. Inorg. Biochem. 84 (2001) 77–88. [50] T.P. Weber, W.R. Widger, H. Kohn, Biochemistry 42 (2003) 9121– 9126. [51] H.P. Schmauder, D. Gröger, J. Biotech. 8 (1988) 239–242. [52] E.M. Linsmaier, F. Skoog, Physiol. Plant 18 (1965) 100–127. [53] W. Maier, A. Baumert, D. Gröger, J. Plant Physiol. 145 (1995) 1–6. [54] G. Giuliano, G.E. Bartley, P.A. Scolino, Plant Cell 5 (1993) 379–387. [55] M. Kato, K. Mizuno, T. Fujimora, M. Iwama, M. Irie, A. Crozier, H. Ashihara, Plant Physiol. 120 (1999) 579–586. [56] O.H. Lowry, N.J. Roesbrough, A.L. Farr, R.J. Randall, Biol. Chem. 193 (1951) 265–275. [57] X. Ni, L.P. Hager, Proc. Natl. Acad. Sci. USA 95 (1998) 12866– 12871.
