Planta (2003) 217: 138–146 DOI 10.1007/s00425-002-0969-0
O R I GI N A L A R T IC L E
Burkhard Meßner Æ Oliver Thulke Æ Anton R. Scha¨ffner
Arabidopsis glucosyltransferases with activities toward both endogenous and xenobiotic substrates
Received: 15 May 2002 / Accepted: 23 November 2002 / Published online: 30 January 2003 Springer-Verlag 2003
Abstract Arabidopsis thaliana Heynh. harbors UDPglucose-dependent glucosyltransferase (UGT; EC 2.4.1.-) activities that are able to glucosylate xenobiotic substrates as a crucial step in their detoxification, similar to other plants. However, it has remained elusive whether side-activities of UGTs acting on endogenous substrates could account for that property. Therefore, seven recombinantly expressed A. thaliana enzymes were tested using the phytotoxic xenobiotic model compound 2,4,5-trichlorophenol (TCP) as a substrate. The enzymes were selected from the large Arabidopsis UGT gene family because their previously identified putative endogenous substrates comprised both carboxylic acid, and phenolic and aliphatic hydroxyl moieties as biochemical targets. In addition, UGT75D1, which was shown to accept the endogenous flavonoid kaempferol as a substrate, was included. All enzymes tested, except the sterol-conjugating UGT80A2, glucosylated TCP as a parallel activity. The Km values for TCP ranged from 0.059 to 1.25 mM. When tested at saturating concentrations of the native substrates the glucosylation of TCP by the glucose-ester-forming UGT84A1 and UGT84A2 was suppressed by p-coumaric acid and sinapic acid, respectively. In contrast, the activities of UGT72E2 and UGT75D1 toward their phenolic native substrates and the xenobiotic TCP were mutually inhibited. TCP was a competitive inhibitor of sinapyl alcohol glucosylation by UGT72E2. These overlapping in vitro activities suggest cross-talk between the detoxification of xenobiotics and endogenous metabolism at the biochemical level, depending on the presence of competing substrates and enzymes.
B. Meßner Æ O. Thulke Æ A.R. Scha¨ffner (&) Institute of Biochemical Plant Pathology, GSF – National Research Center for Environment and Health, 85764, Neuherberg, Germany E-mail: schaeff
[email protected] Fax: +49-89-31872930
Keywords Arabidopsis Æ Competition assays Æ Glucosyltransferase Æ Substrate specificity 2,4,5-Trichlorophenol Xenobiotics Abbreviations GST: glutathione-S-transferase Æ TCP: 2,4,5-trichlorophenol Æ UDPG: UDP-glucose Æ UGT: UDP-glucose-dependent glucosyltransferase
Introduction Plants need to detoxify or regulate the bioactivity of a diverse set of low-molecular-weight compounds. These chemicals are either produced as endogenous defense or signaling molecules or they are imposed on plants from exogenous sources. A plethora of secondary and several primary plant metabolites have been identified, that are reported to occur in their glucosylated forms (Jones and Vogt 2001). Similar to these endogenous compounds, agrochemicals and xenobiotics are commonly detoxified via glucosylation (Cole and Edwards 2000), when functional substituents such as hydroxyl, amino, sulfhydryl, or carboxylic acid groups are present or are introduced by enzymatic reactions. Corresponding to this high number of low-molecular-weight compounds being conjugated a huge number of UDP-glucose dependent glucosyltransferases (UGTs; EC 2.4.1.-) have been revealed by genome projects. They are defined by a conserved amino acid region in the C-terminal parts of the deduced protein sequences. The Arabidopsis thaliana genome harbors 118 putative UGT members (Jones and Vogt 2001; Li et al. 2001; Ross et al. 2001; and our own unpublished data). Research during the last decade on the metabolism of xenobiotic phenols and carboxylic acids revealed that UGT enzymes acting on these substrates are widespread among lower and higher plants, although individual species showed particularly high activities (Sandermann 1994; Pflugmacher and Sandermann 1998). Since glucosylation is an efficient step in the detoxification of xenobiotics these studies proved that the ability of plant
139
species to participate in the removal of xenobiotics from the environment is widely distributed. Several studies were successfully undertaken to biochemically purify plant UGTs that glucosylate xenobiotic compounds such as 2,4,5-trichlorophenol (TCP), pentachlorophenol, chlorinated anilines or herbicide metabolites like 6-hydroxybentazone (Sandermann et al. 1991; Leah et al. 1992; Wetzel and Sandermann 1994; Gallandt and Balke 1995; Schmidt et al. 1995; Brazier et al. 2002). Although high enrichments of these activities had been achieved it could not be excluded that similar enzymes had been co-purified in view of the enormous number of isoenzymes revealed by genome projects. Nevertheless, the lack of activities toward a few selected potential endogenous substrates suggested that these enzymes might be specific for the xenobiotic substrates (Sandermann et al. 1991; Wetzel and Sandermann 1994). In contrast, Leah et al. (1992) described the biochemical purification of a soybean glucosyltransferase acting on both kaempferol and 6-hydroxybentazone. This raises the question of whether this activity was the property of a single enzyme or due to co-purifying activities. Thus, without examining individual UGT gene products it remained elusive whether the glucosylation of xenobiotics was attributable to specific UGT isoenzymes exclusively conjugating xenobiotics or to UGT members displaying side-activities toward xenobiotics. In recent years, a few glucosyltransferases from various species have been cloned, and recombinantly expressed proteins have been assayed to determine their substrates (Fraissinet-Tachet et al 1998; Moehs et al. 1997; Warnecke et al. 1997; Ford et al. 1998; Jones et al. 1999; Lee and Raskin 1999; Martin et al. 1999; Yamazaki et al. 1999; Kita et al. 2000; Milkowski et al. 2000; Taguchi et al. 2001; von Rad et al. 2001; for review: Jones and Vogt 2001). Only recently, however, have systematic analyses of numerous UGTs identified in the A. thaliana genome been initiated. Based on a phylogenetic alignment, 14 A. thaliana UGT groups A to N have been compiled (Li et al. 2001; Ross et al. 2001). Groups A, D, E, H and L are comprised of more than 10 members each, whereas a few groups contain only up to 3 members. A UGT that was active toward sterols (Warnecke et al. 1997) had not been included in these alignments and is now named UGT80A2, defining an additional branch. Analyses of recombinantly expressed A. thaliana UGTs identified a few enzymes that are active toward known endogenous substrates possessing different functional moieties of the aglycones, such as carboxylic acids and phenolic or aliphatic hydroxyl groups (Ross et al. 2001; Warnecke et al. 1997). Most of the enzymes for which a putative endogenous substrate has been identified belong to groups E and L. In addition to related enzymes from other species, these studies identified Arabidopsis enzymes glucosylating indole-3acetic acid (UGT84B1, group L; Jackson et al. 2001), phenylpropanoid derivatives such as coumaric acid, sinapic acid, and sinapyl alcohol (UGT84A1, UGT84A2, both group L; UGT72E2, group E;
Milkowski et al. 2000; Li et al. 2001; Lim et al. 2001; Ross et al. 2001). Group L enzymes were mostly, but not exclusively involved in glucose ester formation, whereas group E enzymes produced O-glucosides (Ross et al. 2001). The ability to accept hydroxybenzoates as substrates was found for members of UGT groups B, D, E, F and L (Lim et al. 2002). We were interested in identifying Arabidopsis UGTs acting on xenobiotics. Since several recent studies reported a rather broad specificity toward putative endogenous substrates, summarized as regioselectivity (Vogt and Jones 2000; Taguchi et al 2001; Vogt 2002), the possibility that this property of plant glucosyltransferases toward potential endogenous substrates could be extended to xenobiotic compounds was intriguing. Thus, several known UGTs from Arabidopsis were selected, including mostly the best-studied group E and group L representatives which accept various known endogenous substrates (Warnecke et al. 1997; Milkowski et al. 2000; Jackson et al. 2001; Lim et al. 2001, 2002). These recombinantly expressed enzymes were tested toward TCP as a model xenobiotic compound (Pflugmacher and Sandermann 1998; Brazier et al. 2002). This biochemical study shows for the first time that a side-activity toward a phytotoxic xenobiotic substrate is shared by several UGTs and may mutually interfere with their activity toward the respective putative endogenous substrates.
Materials and methods Plant material Plants of Arabidopsis thaliana Heynh. ecotype Columbia (Col-0; obtained from G. Redei, University of Missouri) were either grown on soil or hydroponically according to Gibeaut et al. (1997) with addition of 0.5 g/l 2-[N-morpholino]ethanesulfonic acid and adjusting the pH to 5.4 with phosphoric acid. Chemicals All chemicals used in this study were of research-grade purity and purchased from Sigma. Sinapyl alcohol was obtained from Aldrich (Steinheim, Germany). Enzyme preparations from leaves and roots A partially purified UGT fraction was prepared essentially as described by Pflugmacher and Sandermann (1998). Enzymes were extracted at 4 C. Leaves and roots from 3- to 4-week-old Arabidopsis plants (6 g fresh weight) were ground in a mortar to a fine powder with liquid nitrogen. This powder was immediately suspended without clotting in 30 ml buffer, containing 0.1 M Tris–borate (pH 8), 14 mM mercaptoethanol, 2 mM MgCl2, 0.5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride (PMSF), and 600 mg insoluble polyvinylpyrrolidone (Polyclar AT; Serva, Heidelberg, Germany). After sonification for 30 s and passing the extract through Miracloth, cell debris was removed by centrifugation at 15,000 g for 20 min. In order to remove inhibitory compounds, proteins were precipitated from the initial supernatant by ammonium sulfate (40% saturation) in the presence of 1% (w/v) insoluble polyvinylpyrrolidone. The precipitate was removed by
140 centrifugation (15 min, 15,000 g) and the supernatant adjusted to 80% saturation with ammonium sulfate. The precipitated protein was collected by centrifugation and dissolved in 2.5 ml 0.1 M Tris– HCl (pH 7.5) containing 14 mM mercaptoethanol, followed by desalting on Sephadex G-25 (PD-10; Amersham–Pharmacia Biotech, Freiburg, Germany). The desalted enzyme fraction was supplemented with 20% glycerol and stored at 20 C.
Construction of glutathione-S-transferase (GST)–UGT expression plasmids DNA fragments encompassing the coding sequences were amplified from A. thaliana DNA by PCR or, in the case of UGT80A2, RT–PCR. Gene-specific oligomers including tion sites (underlined) for cloning were used:
sequences of UGT Columbia genomic from total RNA by appropriate restric-
– GUGT72E2 – – – – – –
(At5g66690), 5¢-GAATAAGAAGAATTCACGATGGATATCACAAAACC, 5¢-GAGCTAGAAAGCGGCCGCAATTTCTAAGCAC; UGT75B1 (At1g05560), 5¢-GGAGCTAAGACAAGTAGAATTCAAAATGGCGC, 5¢-CTCTTTTGTCCTCGAGTTACTTTACTTTTACCTCC; UGT75D1 (At4g15550), 5¢-CACCAACCAAGAATTCAACAAATGG, 5¢-CAAAATACTGTCGACTGTATTGTG; UGT80A2 (At3g07020) 5¢-TCTGCTGGTCGACATAGAGAGAGAATCACGATG, 5¢-GGAGGGAGAGCGGCCGCTTACGAACAGCCAAAG; UGT84A1 (At4g15480), 5¢-TCCATGTCGACAATGGTGTTCGAAACTTGTCCA, 5¢-AAAGGCGGCCGCCTAGTATCCATTATCTTTAGT; UGT84A2 (At3g21560), 5¢-GAGAAACCAAGTTGAATTCAAAATGGAGCTAG, 5¢-CTTAAATCTCGAGTTAAAAGCTTTTGATTGATCCAG; UGT84B1 (At2g23260), 5¢-ATAAGAATTCATGGGCAGTAGTGAGGG, 5¢-GTTTGAGTCGACAGTTAGGCGATTGTG. Restricted PCR fragments were inserted into the GST gene fusion vector series pGEX-6P (Amersham–Pharmacia Biotech). Escherichia coli strain DH5a was used for transformation.
Purification of recombinant UGTs For recombinant expression of the GST–UGT fusion proteins the corresponding plasmids were transformed into E. coli XL-1 Blue or BL 21. Recombinant enzymes were released and purified essentially according to Jackson et al. (2001). However, E. coli spheroblasts were osmotically shocked by suspending in 0.2· phosphate-buffered saline (PBS: 28 mM NaCl, 0.54 mM KCl, 2 mM Na2HPO4, 0.36 mM KH2PO4, pH 7.3) containing 0.2 mM PMSF, and an additional short sonification was employed to efficiently but gently disrupt the cells and release the protein content. After centrifugation and re-adjusting the supernatant to 1·PBS, the recombinant protein was affinity-purified using a glutathione-coupled Sepharose gel according to the manufacturer’s instructions (Amersham– Pharmacia Biotech). The eluted fusion proteins were concentrated by membrane filtration (Microsep 30 K; Pall, Dreieich, Germany) to about 1 mg/ml and supplemented with 20% glycerol for storage at )20 C. The protein assays were carried out with Bio-Rad Protein Assay Dye (Bio-Rad, Mu¨nchen, Germany) using bovine serum albumin as a reference. The purified recombinant proteins were analyzed by SDS–PAGE (Laemmli et al. 1970). As known for GST-fusion peptides some polypeptides partially released the GST moiety (Jackson et al. 2001). This portion of the intact fusion protein was determined by photometric scanning of the Coomassie-stained SDS–polyacrylamide gels and used to correct the protein amount for the calculation of the specific enzyme activities.
Glucosyltransferase activities in enzyme preparations from leaves and roots Similar to Pflugmacher and Sandermann (1998) the assay mixtures (100 ll) contained 0.1 M Tris–HCl (pH 7.5), 0.1 mM UDP-[14C]glucose (1 kBq/10 nmol), 20–40 lg protein, 0.2 mM aglycone (2 ll each of 10 mM stock solution in ethylene glycol monomethyl ether). 2.5 mM 2-hydroxymethylphenyl-b-D-glucopyranoside and 2.5 mM 4-nitrophenyl-b-D-glucopyranoside were routinely included in order to protect the products from degradation by endogenous b-glucosidases. After incubation (30–60 min, 30 C) the reaction was stopped by addition of 50 ll of 0.1 M H3PO4 and 160 ll ethyl acetate, followed by vortexing (3·60 s) and centrifugation (15,000 g, 2 min) for phase separation. A 100-ll sample of the organic phase was used to determine the formed 14C-glucosides by liquid scintillation counting. Incubation without an aglycone was used as a blank. Specific activity was expressed as lkat/kg protein.
Glucosyltransferase activity assay with recombinant UGTs The enzyme activity was determined with intact fusion proteins since removal of the GST moiety did not alter the activity in the case of UGT75D1 (data not shown) and other UGTs (Jackson et al. 2001). Assay mixtures contained 0.1 M Tris–HCl (pH 7.5), 5 mM UDP-glucose, 2–5 lg fusion protein, 0.1, 0.2, or 2.0 mM aglycone (2 ll each of a stock solution in ethylene glycol monomethyl ether). The final test volume was 50 ll. After incubation for 30–60 min at 30 C the reaction was stopped by addition of 15 ll of 0.5 M H3PO4 and cleared by centrifugation (15,000 g, 2 min). Reverse-phase HPLC was performed using an Agilent 1100 HPLC system (Agilent Technologies, Waldbronn, Germany) and a Prodigy 5 l ODS (3) column (250 mm long, 4.60 mm i.d.; Phenomenex, Aschaffenburg, Germany). A linear gradient (8–100%) with increasing acetonitrile against 0,1% H3PO4 at a flow rate of 1 ml/min over 16 min was used to separate the glucose conjugates from their aglycones. The following wavelengths and retention times (Rt) were used for detection: TCP, 205 nm, Rt=15.2 min; indole-3-acetic acid, 230 nm, Rt=10.3 min; p-coumaric acid, 311 nm, Rt=8.8 min; sinapic acid, 306 nm, Rt=8.9 min; sinapyl alcohol, 285 nm, Rt=8.4 min; kaempferol, 370 nm, Rt=11.9 min. To quantify the glucose conjugates, changes in the extinction coefficients due to glucosylation were taken into account by determining the ratio (r) of the integrated peaks of the completely glucosylated compounds vs. the free aglycone. The detection of the glucose conjugates was as follows: TCP-glucoside, 205 nm, Rt=9.8 min, r=1.8; IAA-glucose, 230 nm, Rt=8.0 min, r=1.0; p-coumaroylglucose, 311 nm, Rt=6.9 min, r=0.88; sinapoylglucose, 306 nm, Rt=7.0 min, r=0.94; syringin, 285 nm, Rt=6.3 min, r=1.6; kaempferolglucose, 370 nm, Rt=8.7 min, r=2.0.
Reverse Northern hybridization DNA fragments to be used as probes for reverse Northern hybridization were amplified from genomic DNA by PCR using the following primers and cloned into vector pGEM-Teasy (Promega, Mannheim, Germany):
– UGT72E2 – – – –
(At5g66690), 5¢-CGCACGAGTCGCTTTGCAGA, 5¢-CTTGCACTCGAACACCGGTTA; UGT75B1 (At1g05560), 5¢-GTGGTGGATATTTGTGGAG, 5¢-CGCTTGCTTGGAGAGTCTC; UGT75D1 (At4g15550), 5¢-GGTGTAAGAGTGATGGAG, 5¢-ACTAACCAATCAACATCC; UGT80A2 (At3g07020) 5¢-AGACACTAGCAAAGGCGATGAAG, 5¢-GAAGTTGCAGCTCAAGAAACTGAA; UGT84A1 (At4g15480), 5¢-GGTGGCTCCAGGAGGTTCG, 5¢-GTCCAAACTTATGGTGAGTC;
141
– UGT84A2 –
(At3g21560), 5¢-GGGTGCCAAACCTGTGGG, 5¢-GGCAAGATATAGTATTAACC; UGT84B1 (At2g23260), 5¢-ATCACAATCGCCTAACTCT, 5¢-GGTCAATATCGACACAAAG.
Their specificity was verified by FASTA analyses against the A. thaliana genome sequence; none of the probes showed a cross-hybridization above a threshold of 70% homology over a stretch of 70 nucleotides. Similarly, probes were prepared for AtPIP1;2 and AtTIP3;1 (data not shown). T7 and Sp6 oligonucleotides binding to flanking vector sequences were used for amplification of all probes before transfer to a nylon membrane (Amersham–Pharmacia Biotech). RNA was isolated from different A. thaliana tissues according to Chang et al. (1993). After reverse transcription and 33 P-radiolabeling the cDNA was hybridized to the membrane. Normalization with respect to the spotted amount of PCR fragments was achieved by an independent hybridization with a radiolabeled T7 oligonucleotide targeting all probes. Signals from three independent experiments were recorded by scanning image plates (FLA-3000; Fuji, Du¨sseldorf, Germany) and quantified using ArrayVision software (Interfocus, Suffolk, UK).
Results Glucosylation of xenobiotics by glucosyltransferases from Arabidopsis leaves and roots To investigate the potential of A. thaliana for glucosylating and detoxifying xenobiotics, partially purified glucosyltransferase preparations from leaves and roots were examined. Leaves and roots constitute the major target organs to soil- and air-borne xenobiotics in ecosystems. Both preparations showed similar activities toward several xenobiotic model compounds containing nucleophilic hydroxyl or carboxylic acid groups, such as 2,4,5-trichlorophenol (TCP), pentachlorophenol, 4-nitrophenol, 3,5-dibromo-4-hydroxybenzoic acid, and 2,2-bis-(4-chlorophenyl)-acetic acid (Fig. 1). However, the root and leaf enzyme extracts exhibited differential specific activities toward the aglycones tested. Whereas the lower activities toward 2,2-bis-(4-chlorophenyl)-acetic acid and 4-nitrophenol were nearly indistinguishable, the root enzyme preparation was substantially more active toward both chlorophenols, pentachlorophenol and TCP. In contrast, the herbicide metabolite 3,5-dibromo4-hydroxybenzoic acid was efficiently glucosylated by the glucosyltransferase preparation from A. thaliana leaves only (Fig. 1). This survey demonstrated the potential of A. thaliana UGTs to glucosylate different xenobiotic compounds and indicated a differential expression in leaves and roots. Similar differences were found when analyzing enzyme preparations from different plant species with xenobiotic phenols (Pflugmacher and Sandermann 1998; Harvey et al. 2002) and may reflect differences in the presence, in the number or in the activity of isoenzymes capable of glucosylating these xenobiotics. TCP, as the most potent substrate in crude enzyme preparations, was selected as a xenobiotic compound for further studies using recombinant enzymes.
Fig. 1 Glucosyltransferase activities toward xenobiotics in Arabidopsis thaliana leaves and roots. Enzyme activities in partially purified glucosyltransferase preparations from Arabidopsis leaves (black columns) and roots (stippled columns) toward xenobiotic phenols and carboxylic acids are shown: 2,4,5-trichlorophenol (a), pentachlorophenol (b), 2,2-bis-(4-chlorophenyl)-acetic acid (c), 3,5dibromo-4-hydroxybenzoic acid (d), and 4-nitrophenol (e). All assays were carried out with 0.1 mM UDP-[14C]-glucose and 0.2 mM aglycone. Conjugation reactions were measured by scintillation counting of the formed 14C-conjugates after ethyl acetate extraction. Results represent the average of three independent replicates ± SD
Activities of recombinant A. thaliana UGTs toward TCP as a xenobiotic model substrate TCP contains an acidic, phenolic hydroxyl group (pKa 7.4) attached to a single aromatic ring as a target for glucosylation. When analyzing potential endogenous substrates, a regioselectivity with respect to the accepted substrates rather than a strict substrate specificity had frequently been observed (Jones and Vogt 2001). In order to examine whether this property can be extended to xenobiotic compounds, enzymes accepting known natural substrates that contain acidic alcohol or carboxyl groups attached to aromatic ring systems as targets for glucosylation were analyzed for side-activities toward TCP. Thus, we selected the A. thaliana glucosyltransferases UGT72E2, UGT75B1, UGT84A1, UGT84A2, and UGT84B1 that had been shown to glucosylate the natural substrates sinapyl alcohol, 4-hydroxybenzoic acid, p-coumaric acid, sinapic acid, and indole-3-acetic acid, respectively (Milkowski et al. 2000; Jackson et al. 2001; Lim et al. 2001, 2002). In addition, UGT75D1, a distinct member of the same UGT group L, like UGT75B1, UGT84A1, UGT84A2, and UGT84B1 (Ross et al. 2001), was included; so far, no endogenous substrate had been identified for UGT75D1 (see below). The expression of all selected UGT genes was verified and determined in different A. thaliana organs. Genespecific fragments designed to distinguish the highly homologous genes were cloned and served as probes for reverse Northern hybridizations. With the exception of
142
the ubiquitously detected UGT75D1 and UGT80A2 (see below) all other transcripts had a differential, organspecific expression pattern showing expression in roots (UGT72E2, UGT84A1, UGT84A2), leaves (UGT75B1), stems (UGT75B1), flowers (UGT84A1, UGT84A2, UGT75B1), and siliques (UGT84A1, UGT84A2, UGT84B1, UGT75B1) (Table 1). However, we cannot rule out the possibility that minor transcription, not detected here, occurred in other organs, such as the weak root expression of UGT84B1 detected by RT–PCR (Jackson et al. 2001). The expression levels of the UGTs examined were generally low; for comparison, expression data for the highly and ubiquitously found aquaporin AtPIP1;2 and the seed-specific AtTIP3;1 are included (Weig et al. 1997; Table 1). All selected enzymes were expressed as fusion proteins with GST and affinity-purified. Enzyme assays were HPLC-based in order to allow quantification as well as monitoring of additional substrates and products. Surprisingly, TCP was a substrate of all six recombinant enzymes, but to different extents (Table 2). The Km values ranged from 0.059 to 1.25 mM, indicating a broad and differential affinity of the examined Table 1 Expression pattern of UGTs in tissues of Arabidopsis thaliana. Gene-specific probes were used for reverse Northern hybridization UGT gene
Roota
Leafa
Stema
Flowera
Siliquea
72E2 75D1 84A1 84A2 84B1 75B1 80A2 AtPIP1;2b AtTIP3;1b
+++ ++ ++ + ) ) +++ ++++ )
) ++ ) ) ) + ++ +++ )
) ++ ) ) ) + ++ ++++ )
) ++ ++ ++ ) ++ +++ ++++ )
) ++ ++ ++ ++ ++ +++ +++ +++
a
Expression levels were classified as – (not detectable), + (very low level but significantly above background; gene expression values 0.1–0.5), ++ (low level; 0.6–6.0),+++ (medium level; 6.1–25), ++++ (high level; >25). Arbitrary numbers for gene expression values are the mean of three independent experiments b Gene-specific 3¢-UTR probes for the highly expressed aquaporin AtPIP1;2 andseed-expressed aquaporin AtTIP3;1 were used as references for gene expression levels and organ specificity (Weig et al. 1997)
UGTs toward TCP. UGT84B1, UGT72E2, and UGT75D1 exhibited the highest affinities. In addition, UGT72E2 showed the highest Kcat value among all enzymes tested. When comparing the enzyme affinities toward TCP and their putative endogenous substrates, both higher (UGT75D1, UGT84A1, UGT84A2, UGT75B1) and lower (UGT72E2, UGT84B1) Km values were observed for the xenobiotic TCP (Lim et al. 2001, 2002; Jackson et al. 2001; for UGT75D1, see below). As indicated by the highest Kcat/Km value, the recombinant UGT72E2 and UGT75D1 were the strongest enzymes glucosylating TCP. To identify an endogenous substrate of UGT75D1, several aglycones were screened. UGT75D1 showed a very low rate of indole-3-acetic acid, 4-hydroxybenzoic acid and sinapic acid glucosylation; no activity could be identified for sinapyl alcohol (data not shown). In contrast, a substantial activity toward the flavonoid kaempferol was detected. The Km of kaempferol was determined to be 0.012 mM and Kcat was 0.060 s)1, suggesting kaempferol as a potential endogenous substrate of UGT75D1 parallel to its activity toward TCP. On the other hand, it was interesting to examine whether any plant UGT would be active toward TCP. UGT80A2 was chosen as a more distantly related member of the glucosyltransferase family that efficiently conjugated non-phenolic hydroxyl groups of several sterols like cholesterol (Warnecke et al. 1997). However, no activity toward TCP could be detected for this enzyme (Table 2). Competition and mutual interference of native substrates and TCP Several recombinant enzymes showed activity toward the xenobiotic model compound TCP. In planta, however, these enzymes may not independently encounter a xenobiotic substrate but rather interact simultaneously with endogenous substrates. Therefore, competition experiments were performed for the most actively TCPconjugating recombinant enzymes UGT72E2 and UGT75D1, offering TCP along with the respective native substrate. Since these enzymes act on phenolic hydroxyl moieties of their putative endogenous
Table 2 Kinetic data for TCP as a substrate of UGTs.Mean values from three independent experiments ± SD are shown. The first fourenzymes were used for competition studies (Figs. 2, 3). n.s. Not a substrate UGT
Km (mM)
Kcat (s)1)
72E2 75D1 84A1 84A2 84B1 75B1 80A2
0.11±0.016 0.076±0.009 1.25±0.17 0.41±0.035 0.059±0.016 0.54±0.035 n.s.
0.23±0.043 0.058±0.012 0.13±0.044 0.079±0.007 0.027±0.010 0.077±0.008 n.s.
a
Lim et al. 2001 Jackson et al. 2001
b
Kcat/Km (mM)1 s)1) 2.1 0.77 0.10 0.19 0.47 0.14 – c
Lim et al. 2002 Warnecke et al. 1997
d
Putative endogenous substrate Sinapy lalcohola Kaempferol p-Coumaric acida Sinapic acida Indole-3-acetic acidb 4-Hydroxybenzoic acidc Sterolsd
143
substrates, two additional TCP-conjugating enzymes, UGT84A1 and UGT84A2, forming glucose esters of their natural substrates were included in this study. The enzyme activities toward TCP at 0.2 mM were compared with the respective native substrates at 0.2 and 2.0 mM, except for kaempferol, which was used at 0.1 mM, close to its maximum solubility in the aqueous buffer. Both UGT84A1 and UGT84A2 showed preferential activity toward p-coumaric acid and sinapic acid vs. TCP when individually tested at 0.2 mM (Fig. 2) in accordance with their relative Km values (Table 2; Lim et al. 2001). Addition of the endogenous substrates at the same or a 10-fold higher concentration together with TCP suppressed and almost eliminated the activity of both enzymes toward TCP. In contrast, the presence of TCP did not appreciably interfere with the glucosylation of the native substrate. A different scenario was found for UGT72E2 and UGT75D1. At 0.2 mM of either substrate the activity of UGT72E2 toward TCP was about one-sixth of that toward sinapyl alcohol (Fig. 3a). However, in the presence of both substrates a mutual suppression of both activities was found. A 10-fold excess of the natural substrate strongly reduced TCP-conjugation. Nevertheless, there was still a parallel repression of the formation of the sinapyl alcohol glucoside, in agreement with the somewhat lower Km of TCP. The experimental data for the competition assay of UGT75D1 were different from those of the other enzymes because of the limitation of the concentration of kaempferol to 0.1 mM. Nevertheless, a mutual interference and repression was found in the presence of kaempferol (0.1 mM) and TCP
Fig. 2 Competition assays: activities of UGT84A1 (a) and UGT84A2 (b) toward TCP and native carboxylic acids. Conjugation reactions were measured in the presence of one or two substrates at concentrations as indicated (0.2, 2.0 mM) and 5 mM UDPG. Cac p-Coumaric acid, Sac sinapic acid. Activities for TCP (black columns) and natural substrates (stippled columns) were determined using HPLC. Results represent the average of three independent replicates ± SD
(0.2 mM) similar to that observed for UGT72E2 (Fig. 3). The mutual inhibition by the natural and xenobiotic substrate appeared to be different for the two glucose-ester-forming UGTs in comparison to the two glucoside-forming UGTs (Figs. 2, 3). Therefore, the nature of this interaction was further examined for UGT84A2 acting on sinapic acid (forming the glucose ester) and TCP in comparison to UGT72E2 acting on the structurally related sinapyl alcohol (forming the phenolic O-glucoside) and TCP. The influence of TCP on the glucosylation of the native substrates was analyzed according to Dixon (1953). UGT84A2 did not show any inhibition of the formation of sinapoyl glucose ester by TCP (Fig. 4a), corroborating the data presented above (Fig. 2b). Thus, sinapic acid is the clearly preferred substrate and TCP-glucosylation could be only found as a side-activity of UGT84A2 if it was present as a single substrate. Obviously, no true competition between the two substrates could be detected under these conditions and it remained unclear whether TCP and sinapic acid use the same substrate binding sites. In contrast, the analogous Dixon analysis of UGT72E2 revealed a competitive inhibition of sinapyl alcohol glucosylation by TCP (Fig. 4b). The inhibition constant Ki of TCP was 0.06±0.03 mM, which is in the range of the Km constant for TCP (Table 2). An additional plot according to Cornish-Bowden (1974) showed parallel graphs, revealing that the inhibition was purely competitive (Fig. 4c).
Fig. 3 Competition assays: activities of UGT72E2 (a) and UGT75D1 (b) toward TCP and native phenolic compounds. Conjugation reactions were measured in the presence of one or two substrates with concentrations as indicated (0.1, 0.2, 2.0 mM) and 5 mM UDPG. Sal Sinapyl alcohol, Kae kaempferol. Kaempferol was added at 0.1 mM, close to its maximum solubility. Activities for TCP (black columns) and natural substrates (stippled columns) were determined using HPLC. Results represent the average of three independent replicates ± SD
144
Fig. 4 Influence of TCP on glucosylation of sinapic acid by UGT84A2 (a) and sinapyl alcohol by UGT72E2 (b, c). Different concentrations of TCP were added to reactions with fixed concentration of sinapic acid (Sac; a) or sinapyl alcohol (Sal; b, c) as indicated. Data for glucoside formation were plotted according to Dixon (1953) (a, b) and Cornish-Bowden (1974) (c). The experiments were repeated three times; results from representative plots are shown
Discussion Enzyme activities toward endogenous and xenobiotic aglycones Several recombinantly expressed plant UGTs possess a side-activity toward the phytotoxic, xenobiotic TCP in addition to their putative endogenous substrates such as p-coumaric acid, sinapic acid, sinapyl alcohol or kaempferol. The latter had been identified as a potential native substrate of UGT75D1. Previous biochemical purification of plant UGTs glucosylating chlorophenols or pesticide metabolites had raised the question of whether these activities indicated specific UGTs acting on xenobiotics or whether they were side-activities of
UGTs acting on native substrates (see Introduction). The glucosylation of TCP by several recombinantly expressed UGT enzymes identified in this study clearly proved that there are plant enzymes with overlapping, parallel activity toward their putative endogenous substrates and the xenobiotic TCP. Although these unequivocal results may render unlikely the existence of UGTs that specifically act on xenobiotics, they cannot exclude such a possibility. So far, a broad activity toward both endogenous and exogenous aglycones had been only confirmed for mammalian UGT1.1, which was able to glucuronidate a large number of various compounds in the liver as a step in the detoxification of endogenous catabolites or xenobiotic drugs (King et al. 1996). The TCP-conjugating UGTs identified here belong to subfamilies E and L of A. thaliana UGTs that had been selected on the basis of both chemical and structural considerations regarding their putative endogenous aglycones in comparison with TCP. Interestingly, TCP was found to be a substrate of both glucose-ester- and O-glucoside-forming UGTs. The surprising side-activity of the ester-forming enzymes may be caused by the rather high acidity of the phenolic hydroxyl group of TCP having a pKa value of 7.4 leading to a substantial negative charge and ease of deprotonation under the physiological pH conditions (7.5) used in the enzymatic assays. In addition, all active enzymes target functional groups in electronic conjugation with an aromatic system like the phenolic group of TCP. Indeed, UGT80A2, which exhibits a broad substrate specificity toward steroids with non-aromatic hydroxyl groups (Warnecke et al. 1997), did not glucosylate TCP. Considering these aspects of substrate properties, we suggest that ‘chemiselectivity’, emphasizing the chemical reactivity of the target group besides the known influence of regioselectivity (Vogt et al. 1999; Jones and Vogt 2001; Taguchi et al. 2001), constitutes an additional feature governing the activity of plant UGTs. Although focusing mainly on the regioselectivity of tobacco UGTs recognizing a 2-naphthol substructure, Taguchi et al. (2001) had already found additional activities toward structurally non-related substrates like hydroxycinnamic acid that also contained phenolic or acidic target moieties. It will be interesting to investigate whether xenobiotics other than TCP are glucosylated by the enzymes studied in this work and whether this can be explained by chemi- and regioselectivity. In parallel, it will be important to extend the study to UGT enzymes belonging to other subfamilies defined by Ross et al. (2001), especially once native substrates are discovered. Despite the basic side-activity toward TCP the mutual interference between TCP and the native substrates appeared to be different for the glucose-ester-forming vs. the O-glucoside-forming enzymes studied. Although UGT84A1 and UGT84A2 exhibited a clear side-activity toward TCP, TCP-glucosylation was strongly sup-
145
pressed by the putative endogenous hydroxycinnamic acid substrates. On the other hand, TCP did not inhibit the formation of glucose-esters of p-coumaric acid and sinapic acid, respectively (Figs. 2, 4a). Therefore, no competitive inhibition occurred under these conditions. In contrast, a mutual inhibition was demonstrated for the O-glucoside-forming UGT72E2 and UGT75D1 (Fig. 3). In the case of UGT72E2 the xenobiotic TCP competitively inhibited the glucosylation of sinapyl alcohol, indicating overlapping binding sites for their substrates. In terms of evolution, the contribution of at least several UGTs to the detoxification of a single xenobiotic such as TCP may be a consequence of the ‘‘snowball’’ effect, i.e. the more genes already present in a family the faster the generation of even more new genes (Pichersky and Gang 2000). Thus, by extrapolating the findings for TCP as a model xenobiotic substrate, we speculate that the likelihood of evolving enzymes that show sideactivities toward xenobiotic substrates may increase with the number of genes present, leading to a broad-spectrum anti-xenobiotic enzymatic machinery based on several individual enzymes. Possible implications for the function of UGT enzymes in planta The biochemical data presented suggest that several UGT enzymes may contribute to the glucosylation of TCP in vivo. Thus, the capacity for detoxification will depend on their developmental and spatio-temporal expression patterns. In addition, detoxification may be influenced by a direct induction of these isoenzymes in response to xenobiotics or to other environmental factors. However, these activities have been identified as overlapping with the glucosylation of the putative endogenous substrates. In vitro competition assays showed both a strong suppression of TCP conjugation by the native compounds (Figs. 2, 4a) and a mutual inhibition of the formation of the respective glucosides (Figs. 3, 4b, c). In vivo, the substrate concentrations will be considerably lower than those used in this and other in vitro studies. Thus, the reactions occurring in vivo will be mainly governed by the ability to bind different substrates efficiently, as indicated by low Km values. They will also depend on the presence of the respective enzymes and the concentrations of the competing substrates at a specific site. Consequently, differences in the composition of metabolites and the related UGTs may influence the ability and efficacy to detoxify TCP, or other xenobiotics if these findings could be extended to other compounds. On the other hand, apparently specific UGT activities toward xenobiotics such as those induced after safener treatments of wheat (Brazier et al. 2002) may also influence the formation of yet unidentified endogenous glucosides and thus affect properties of the plant other than its detoxification capability, such as development, stress response, or nutritional value.
Indeed, these mutual interactions may indicate crosstalk between detoxification pathways and endogenous metabolism at the biochemical level. Specifically, the TCP-conjugating enzymes identified here are probably involved in lignin biosynthesis, e.g. the putative sinapyl alcohol glucosyltransferase UGT72E2, or in the biosynthesis of protective secondary metabolites, e.g. the putative sinapic acid glucosyltransferase UGT84A2. In conclusion, our findings strongly support the notion that UGTs acting on xenobiotics may well constitute ‘‘old enzymes for a new job’’ (Kreuz et al. 1996). However, since they may be still engaged in their old jobs, any kind of overlap of the activities toward endogenous and exogenous substrates may lead to interferences between different branches of (secondary) metabolism and detoxification. A future challenge will be to reveal whether these overlapping and interfering biochemical features can be verified in planta and how they influence the properties of such plants. This will require the expression level of single as well as multiple UGT enzymes in transgenic plants to be altered by overexpression and by suppression. Acknowledgements The authors thank Birgit Geist, Elke Gerstner, Elisabeth Schindler, and Susanna Holzinger for technical assistance. Matthias Affenzeller from our laboratory provided RNA from different A. thaliana tissues. We are indebted to Drs. Werner Heller, Christian Langebartels, Dieter Ernst and Heinrich Sandermann (all our Institute) for critical reading of the manuscript. Two anonymous referees provided valuable comments.
References Brazier M, Cole DJ, Edwards R (2002) O-glucosyltransferase activities toward phenolic natural products and xenobiotics in wheat and herbicide-resistant and herbicide-susceptible blackgrass (Alopecurus myosuroides). Phytochemistry 59:149–156 Chang S, Puryear J, Cairney J (1993) A simple and efficient method for isolating RNA from pine trees. Plant Mol Biol Rep 11:113– 116 Cole DJ, Edwards R (2000) Secondary metabolism of agrochemicals in plants. In: Roberts T (ed) Metabolism of agrochemicals in plants. Wiley, New York, pp 107–154 Cornish-Bowden A (1974) A simple graphical method for determining the inhibition constants of mixed, uncompetitive and non-competitive inhibitors. Biochem J 137:143–144 Dixon M (1953) The determination of enzyme inhibitor constants. Biochem J 55:170–171 Ford CM, Boss PK, Hoj BP (1998) Cloning and characterization of Vitis vinifera UDP-glucose:flavonoid 3-O-glucosyltansferase, a homologue of the enzyme encoded by the maize Bronze-1 locus that may primarily serve to glucosylate anthocyanidins in vivo. J Biol Chem 273:9224–9233 Fraissinet-Tachet L, Baltz R, Chong J, Kauffmann S, Fritig B, Saindrenan P (1998) Two tobacco genes induced by infection, elicitor and salicylic acid encode glucosyltransferases acting on phenylpropanoids and benzoic acid derivatives, including salicylic acid. FEBS Lett 437:319–323 Gallandt ER, Balke NE (1995) Xenobiotic glucosyltransferase activity from suspension-cultured Glycine max cell. Pestic Sci 43:31–40 Gibeaut DM, Hulett J, Cramer GR, Seemann JR (1997) Maximal biomass of Arabidopsis thaliana using a simple, low-maintenance hydroponic method and favorable environmental conditions. Plant Physiol 115:317–319
146 Harvey PJ, Campanella BF, Castro PML, Harms H, Lichtfouse E, Scha¨ffner AR, Smrcek S, Werck-Reichhart D (2002) Phytoremediation of polyaromatic hydrocarbons, anilines and phenols. Environ Sci Pollut Res 9:29–47 Jackson RG, Lim E, Li Y, Kowalczyk M, Sandberg G, Hoggett J, Ashford DA, Bowles DJ (2001) Identification and biochemical characterization of an Arabidopsis indole-3-acetic acid glucosyltransferase. J Biol Chem 276:4350–4356 Jones P, Vogt T (2001) Glycosyltransferases in secondary plant metabolism: tranquilizers and stimulant controllers. Planta 213:164–174 Jones PR, Moller BL, Hoj PB (1999) The UDP-glucose: p-hydroxymandelonitrile-O-glucosyltransferase that catalyzes the last step in synthesis of the cyanogenic glucoside dhurrin in Sorghum bicolor. J Biol Chem 274:35483–35491 King CD, Green MD, Rios GR, Coffman BL, Owens IS, Bishop WP, Tephly TR (1996) The glucuronidation of exogenous and endogenous compounds by stably expressed rat and human UDP-glucuronosyltransferase 1.1. Arch Biochem Biophys 332:92–100 Kita M, Hirata Y, Moriguchi T, Endo-Inagaki T, Matsumoto R, Hasegawa S, Suhayda CG, Omura M (2000) Molecular cloning and characterization of a novel gene encoding limonoid UDPglucosyltransferase in Citrus. FEBS Lett 469:173–178 Kreuz K, Tommasini R, Martinoia E (1996) Old enzymes for a new job. Plant Physiol 111:349–353 Laemmli UK (1970) Cleavage of structural proteins during the assembly of the head pf bacteriophage T4. Nature 227:680–685 Leah JM, Worrall TL, Cobb AH (1992) Isolation and characterization of two glucosyltransferases from Glycine max associated with bentazone metabolism. Pestic Sci 34:81–87 Lee H-I, Raskin I (1999) Purification, cloning, and expression of a pathogen inducible UDP-glucose:salicylic acid glucosyltransferase from tobacco. J Biol Chem 274:36637–36642 Li Y, Baldauf S, Lim E, Bowles DJ (2001) Phylogenetic analysis of the UDP-glycosyltransferase multigene family of Arabidopsis thaliana. J Biol Chem 276:4338–4343 Lim E, Li Y, Parr A, Jackson R, Ashford DA, Bowles DJ (2001) Identification of glucosyltransferase genes involved in sinapate metabolism and lignin synthesis in Arabidopsis. J Biol Chem 276:4344–4349 Lim E, Doucet CJ, Li Y, Elias L, Worrall D, Spencer SP, Ross J, Bowles DJ (2002) The activity of Arabidopsis glycosyltransferases toward salicylic acid, 4-hydroxybenzoic acid, and other benzoates. J Biol Chem 277:586–592 Martin RC, Mok MC, Mok DWS (1999) Isolation of a cytokinin gene, ZOG1, encoding zeatin O-glucosyltransferase from Phaseolus lunatus. Proc Natl Acad Sci USA 96:284–289 Milkowski C, Baumert A, Strack D (2000) Identification of four Arabidopsis genes encoding hydroxycinnamate glucosyltransferases. FEBS Lett 486:183–184 Moehs CP, Allen PV, Friedman M, Belknap WR (1997) Cloning and expression of solanidine UDP-glucose glucosyltransferase from potato. Plant J 11:227–236
Pflugmacher S, Sandermann H (1998) Taxonomic distribution of plant glucosyltransferases acting on xenobiotics. Phytochemistry 49:507–511 Pichersky E, Gang DR (2000) Genetics and biochemistry of secondary metabolites in plants: an evolutionary perspective. Trends Plant Sci 5:439–445 Ross J, Li Y, Lim E, Bowles DJ (2001) Protein family review. Higher plant glycosyltransferases. Genome Biol 2:3004.13004.6 Sandermann H (1994) Higher plant metabolism of xenobiotics: the "green liver" concept. Pharmacogenetics 4:225–241 Sandermann H, Schmitt R, Eckey H, Bauknecht T (1991) Plant biochemistry of xenobiotics: isolation and properties of soybean O- and N-glucosyl and O- and N-malonyltransferases for chlorinated phenols and anilines. Arch Biochem Biophys 287:341–350 Schmidt B, Rivero C, Thiede B (1995) 3,4-Dichloroaniline N-glucosyl- and N-malonyltransferase activities in cell cultures and plants of soybean and wheat. Phytochemistry 39:81–84 Taguchi G, Yatawa T, Hayashida N, Okazaki M (2001) Molecular cloning and heterologous expression of novel glucosyltransferases from tobacco cultured cells that have broad substrate specificity and are induced by salicylic acid and auxin. Eur J Biochem 268:4086–4094 Vogt T (2002) Substrate specificity and sequence analysis define a polyphyletic origin of betanidin 5- and 6-O-glucosyltransferase from Dorotheanthus bellidiformis. Planta 214:492–495 Vogt T, Jones P (2000) Glycosyltransferases in plant natural product synthesis: characterization of a supergene family. Trends Plant Sci 5:380–386 Vogt T, Grimm R, Strack D (1999) Cloning and expression of a cDNA encoding betanidin 5-O-glucosyltransferase, a betanidin- and flavonoid-specific enzyme with high homology to inducible glucosyltransferases from Solanaceae. Plant J 19:509–519 von Rad U, Hu¨ttl R, Lottspeich F, Gierl A, Frey M (2001) Two glucosyltransferases are involved in detoxification of benzoxazinoids in maize. Plant J 28:633–642 Warnecke DC, Baltrusch M, Buck F, Wolter FP, Heinz E (1997) UDP-glucose:sterol glucosyltransferase: cloning and functional expression in Escherichia coli. Plant Mol Biol 35:597– 603 Weig A, Deswarte C, Chrispeels MJ (1997) The major intrinsic protein family of Arabidopsis has 23 members that form three distinct groups with functional aquaporins in each group. Plant Physiol 114:1347–1357 Wetzel A, Sandermann H (1994) Plant biochemistry of xenobiotics: isolation and characterization of a soybean O-glucosyltransferase of DDT metabolism. Arch Biochem Biophys 314:323– 328 Yamazaki M, Gong Z, Fukuchi-Mizutani M, Fukui Y, Tanaka Y, Kusumi T, Saito K (1999) Molecular cloning and biochemical characterization of a novel anthocyanin 5-O-glucosyltransferase by mRNA differential display for plant forms regarding anthocyanin. J Biol Chem 274:7405–7411
