Báo cáo khoa học: Probing suggested catalytic domains of glycosyltransferases by site-directed mutagenesis Tobias Hefner and Joachim Stockigt ¨ ppt

Probing suggested catalytic domains of glycosyltransferases by site-directed mutagenesis Tobias Hefner and Joachim Sto¨ ckigt Lehrstuhl fu ¨ r Pharmazeutische Biologie, Johannes Gutenberg-Universita ¨ t Mainz, Germany The plant enzyme arbutin synthase isolated from cell sus- pension cultures of Rauvolfia serpentina and heterologously expressed in Escherichia coli is a member of the NRD1b family of glycosyltransferases. This enzyme was used to prove, by site-directed mutagenesis, suggested catalytic domains and reaction mechanisms proposed for enzyme- catalyzed glycosylation. Replacement of amino acids far from the NRD domain do not significantly affect arbutin synthase activity. Exchange of amino acids at the NRD site leads to a decrease of enzymatic activity, e.g. substitution of Glu368 by Asp. Glu368, which is a conserved amino acid in glycosyltransferases located at position 2 and is important for enzyme activity, does not serve as the nucleophile in the catalytic centre as proposed. When it is replaced by Ala, the resulting mutant enzyme E368A exhibits comparable acti- vity as found for E368D in respect to vanillin. Enzyme activities of wild-type and E368A towards several substrates were not affected at the same level. His360 at position 1 of NRD1b glycosyltransferases occupies a more crucial role as expected. When it is exchanged against other basic amino acids such as Lys or Arg the enzyme activity decreases  1000-fold. Replacement of His360 by Glu leads to a mutant enzyme (H360E) with an  4000-fold lower activity compared with the wild-type. This mutein still produces a b-glucoside, not an a-glucoside and therefore indicates that generation of the typical E–E motif of NRD1a glycosyl- transferases does not convert a NRD1b enzyme into a NRD1a enzyme. The presented data do not support several suggestions made in the literature about catalytic amino acids involved in the glycosyltransfer reaction. Keywords: arbutin synthase; catalytic domains; NRD glycosyltransferases; reaction mechanism; site-directed mutagenesis. The transfer of a monosaccharide moiety from an activated sugar donor to monomeric and polymeric acceptor mole- cules is a common reaction in nature. The glycosylation of an enormous variety of natural compounds and also of a broad range of xenobiotics is catalyzed by more than 300 known glycosyltransferases identified from human, animal, microbial and plant sources [1,2]. Although some of these transferases have already been exhaustively investigated for many decades the molecular mechanism of their action including details of their catalytic domains remains mostly unexplored. A rapidly growing amount of sequence data of these enzymes and numerous sequence alignment studies provide first insights into the process of glycosylation. They also deliver working hypo- theses, which might help to establish a better understanding of these processes although the end conclusions based on sequence alignments are still highly speculative. Application of additional approaches such as site-direc- ted mutagenesis and X-ray analyzes must have priority in order to solve the catalytic mechanism of glucosyl transfer in the near future. Heterologous expression of glucosyl- transferases will be, however, a prerequisite to succeed in this research. We have recently isolated a novel glucosyltransferase catalyzing the glucosylation of hydroquinone from cell suspension cultures of the Indian medicinal plant Rauvolfia serpentina Benth. ex Kurz [3]. We named this enzyme arbutin synthase. Functional heterologous expression of this synthase in Escherichia coli by the approach of Ôreverse geneticsÕ allowedustousetheenzymeasoneofthemost promising candidates to prove general suggestions made recently on the reaction mechanism of glycosyltransferases. In this paper we report on appropriate site-directed mutagenesis experiments performed on arbutin synthase, which are applied to evaluate the validity of general mechanistic models of glucose transfer. Materials and methods Site directed mutagenesis Mutagenesis of arbutin synthase (AS) was achieved using the QuickChange TM Site-Directed Mutagenesis Kit (Strat- agene, La Jolla, USA). As template for PCR AS-pQE60 [¼ AS(His) 6 ] construct and the following primer pairs were used (substituted amino acids are underlined): L(86)Ifor: Correspondence to J. Sto ¨ ckigt, Department of Pharmaceutical Biology, Institute of Pharmacy, Johannes Gutenberg-University Mainz, Staudinger Weg 5, 55099 Mainz, Germany. Fax: + 49 6131 3923752, Tel.: + 49 6131 3925751, E-mail: stoeckig@mail.uni-mainz.de Abbreviations: AS, arbutin synthase; NRD, nucleotide recognition domain. Enzymes: arbutin synthase (EC 2.4.1.218). Note: The cDNA sequence of arbutin synthase from Rauvolfia serpentina wassubmittedtoGenBankwithaccessionnumber AJ310148. Note: Dedicated to Professor Zenk on his seventieth birthday. (Received 15 April 2002, revised 27 September 2002, accepted 2 December 2002) Eur. J. Biochem. 270, 533–538 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03409.x 5¢-GACCCGTATTTGTATCACCATCACTCGCTCTCT CCCG-3¢, L(86)Irev: 5¢-CGGGAGAGAGCGAGTGAT GGT GATACAAATACGGGTC-3¢; A(204)Vfor: 5¢-GGC CAAGAGATACCGGTTA GTTGAGGGTATCATGG- 3¢, A(204)Vrev: 5¢-CCATGATACCCTC AACTAACCGG TATCTCTTGGCC-3¢; E(368)Dfor: 5¢-GGAACTCT ATTCTT GACAGTGTAGTTAATGGGGTGCCG-3¢; E(368)Drev: 5¢-CGGCACCCCATTAACTACACT GTC AAGAATAGAGTTCC-3¢; E(368)A: 5¢-GGAACTCT ATTCTT GCGAGTGTAGTTAATGGGGTGCCG-3¢; E(368)Arev: 5¢-CGGCACCCCATTAACTACACT CGCA AGAATAGAGTTCC-3¢; H(360)Rfor: 5¢-GGGTGGATT TCTAACC CGATGCGGGTGGAAC-3¢; H(360)Rrev: 5¢- GTTCCACCCGCA TCGGGTTAGAAATCCACCC-3¢; H(360)Kfor: 5¢-CGGGTGGATTTCTAACC AAGTGCG GGTGGAAC-3¢; H(360)Krev: 5¢-GTTCCACCCGCA CTTGGTTAGAAATCCACCC-3¢; H(360)Efor: 5¢-CG GGTGGATTTCTAACC GAGTGCGGGTGGAAC-3¢; H(360)Erev: 5¢-GTTCCACCCGCA CTCGGTTAGAAA TCCACCC-3¢. The resulting plasmids were transformed into E. coli TOP10 and sequenced after purification using the Nucleo Spin Plasmid Kit (Macherey-Nagel, Du ¨ ren, Germany). Protein expression For expression of the mutant enzymes the plasmids were transformed into E. coli M15 cells. Cultures in 100 mL LB (Luria–Bertani) medium, containing 100 mgÆL )1 ampicillin and 25 mgÆL )1 kanamycin, were grown overnight at 200 r.p.m and 37 °C. These cultures which were used to inoculate 2 L LB medium (antibiotics as before plus 0.3 m M IPTG), were cultivated at 100 r.p.m and 25 °C. After 24 h the cells were harvested by centrifugation at 4000 g for 10 min. The resulting pellets were resuspended in 50 mL buffer (K 2 HPO 4 50 m M ,pH8.0,300m M NaCl, 10 m M imidazole and 20 m M b-mercaptoethanol) and 1 mgÆmL )1 lysozyme was added. After incubation on ice for 30 min the cells were lyzed by sonification (70 W, 6 · 10 s) and centrifuged at 12 000 g for 30 min. The resulting superna- tants were pumped through Ni-nitrilotriacetic acid (Qiagen, Hilden, Germany) columns (each 1 mL volume) at a flow rate of 0.5 mLÆmin )1 . After washing the column with buffer containing 20 m M imidazole the enzyme was eluted by a linear gradient (20–250 m M imidazole) over 20 column volumes. The purity of the eluted enzymes was checked by Coomassie-blue stained SDS/PAGE [4]. Protein concentration and activity Protein concentrations were measured using the method of Bradford [5] and a standard curve derived from bovine serum albumin. For testing the activity and determining the kinetic parameters of arbutin synthase wild-type and mutant enzymes the following assay was used. A solution of 1 m M hydroquinone or the substrates tested (Fig. 5), 2m M UDP-Glu, 100 m M Tris/HCl, pH 7.5 in a total volume of 127.6 lL was prepared. Enzyme in different amounts was added to this solution and the mixture was incubated at 50 °C for various times. After terminating the enzymatic reaction with 300 lL MeOH and centrifugation at 18 000 g for 5 min the supernatant was analyzed by HPLC. A 250 · 4 mm LiChrospher 60 RP-select B column (5 lm) (Merck, Darmstadt, Germany) was used and a solvent system consisting of 2% acetonitrile and 98% water, pH 2.3 (H 3 PO 4 ). For verifying whether a-glucosides or b-glucosides were formed with the mutein H360E, an assay with the following conditions was applied: 1 mg hydroqui- none, 10 mg UDPG, 50 lg enzyme in 550 lLwater, containing 100 m M Tris, pH 7.5 and 20 m M mercaptoeth- anol was prepared. After incubation at 37 °Cfor17h,the reaction was terminated with 300 lL MeOH, centrifuged and freeze-dried. The residue was dissolved in 500 lL MeOH/H 2 O (7 : 3) and applied to a TLC plate (Silica gel 60 F 254 , solvent system EtOAc/MeOH/H 2 O (7:2:1)). The bands identified as the glucoconjugates were scratched out and eluted with 1.5 mL CH 2 Cl 2 /MeOH (7 : 3). Ten micro- litres of this fraction were mixed with 190 lLMeOHand analyzed by the above described HPLC method. A peak at 3.6 min clearly showed glucosylated hydroquinone. The samples were freeze-dried and dissolved in 100 lLH 2 O. To 25 lL of this solution 175 lL citrate buffer (100 m M , pH 5.0) containing 20 nkat almond-derived b-glucosidase (Sigma, Deisenhofen, Germany) or a-glucosidase (20 nkat) from brewers yeast (Sigma, Deisenhofen, Germany) in 100 m M K 2 HPO 4 (pH 6.0) were added. After incubation at 37 °C for 1 h the reaction was terminated with 300 lL MeOH followed by centrifugation (5 min, 18 000 g). The supernatant was analyzed by the HPLC and TLC methods described above. Results and discussion In our previous studies we have described the isolation from plant cell suspension cultures of R. serpentina a UDP- glucose dependent enzyme which glucosylates hydroqui- none with formation of the O-b- D -glucoside arbutin (Fig. 1). Arbutin synthase has also been heterologously expressedinanactiveforminE. coli [6] followed by a detailed sequence analysis and investigation of the enzyme properties, especially of its substrate specificity [7]. Based on these substrate studies, arbutin synthase is not only a glycosyltransferase with an exceptionally broad substrate acceptance but also in this respect exceeds all the so far known proteins of this particular enzyme family. Indeed it is an unique enzyme which at the present time exhibits the most multifunctional character in the metabolism of natural compounds by converting members of many different groups of natural products, e.g. phenyl-propanoids, cou- marins, anthraquinones, flavonoids and protoberberines. In addition this glucosyltransferase also glucosylates a large number of phenolic xenobiotics. In general, all the glycosyltransferases belong to only two types of enzymes, transferases retaining the stereochemistry at the anomeric carbon or those inverting the configuration Fig. 1. Catalyzed reaction of arbutin synthase (AS) isolated from cell suspension cultures of Rauvolfia s erpentina or heterologously expressed in E. coli. 534 T. Hefner and J. Sto ¨ ckigt (Eur. J. Biochem. 270) Ó FEBS 2003 at that centre during sugar transfer [8–10]. Extensive computer alignments of the amino acid sequences of these enzymes were used in the past not only for further classification but also to propose catalytic domains of these proteins and the nature of involved reaction mechanisms. When applying the approach of Campbell et al.[1,2],who divided the glycosyltransferases into 26 families, we could place arbutin synthase clearly in family 1, which consists of transferases from viruses, bacteria, fungi, higher plants and animals. The same result was obtained when we followed the classification of the Cazy-Server (http://afmb.cnrs-mrs. fr/cazy/CAZY/index.html), subdividing the glycosyltrans- ferases into 56 different families based on the system of Campbell et al. [1,2]. Other authors classify glycosyltrans- ferases after the appearance of the so-called DxD-motif [11–15], which is believed to be involved in binding the UDP moiety. In the sequence of arbutin synthase this motif could not be unambiguously identified. Although there are several sequences, which could be potential candidates (Fig. 2), the surrounding amino acids at these sites do not fit an extended DxD-motif taking into account the properties of neigh- bouring amino acids. Using hydrophobic cluster analysis (http://smi.snv.jussieu.fr/hca/hca-form.html) appropriate clusters, which seem to be important for a typical DxD- motif [11,15], could not be detected in arbutin synthase (data not shown). Based on an overwhelming amount of data of their primary structures, which derive from cloning of the appropriate cDNAs, the glycosyltransferases are grouped into NRD1 and NRD2 proteins because of their nucleotide-recognition-domain. The NRD1 family is fur- ther classified into the NRD1a and NRD1b subgroups, depending on the stereochemical course of glycosylation, which can proceed with retention or inversion of the stereochemistry at the anomeric centre of the glucose moiety. The inverting enzymes (NRD1b and NRD2) were further subdivided into a class showing the motif of a His (or Arg/Asp) representing position 1 which in general is located eight amino acids upstream from a Glu residue (position 2) [8]. As previously classified, most of the NRD1b H(R/K)-E domain-containing transferases form one specific family. This family is well separated from another, exhibiting instead the E-E motif [1,2]. The resulting O-b)glucosides of arbutin synthase but also conserved sequences such as the small NRD1bSandlarge NRD1bL domains (Fig. 2) clearly place the enzyme into the NRD1b family. In agreement with this classification, arbutin synthase exhibits the His-Glu site (Fig. 2), proposed as a catalytic domain of general importance for the mechanism of sugar transfer. The significance of this His- Glu site we also explored by an alignment based on amino acid sequences of 24 plant-derived glucosyltransferases including arbutin synthase. As illustrated in Fig. 3 this motif is, in fact, completely conserved within the NRD domain which is a part of the so called plant secondary product glucosyltransferase (PSPG)-box [9,10]. Therefore arbutin synthase became an interesting example for the evaluation of earlier suggestions concerning catalytic amino acids and proposed reaction mechanisms of glycosyltrans- ferases [8]. Because site-directed mutagenesis studies have not so far been performed for eucaryotic glycosyltrans- ferases in order to prove their catalytic amino acids, we have generated for this study wild-type arbutin synthase-(His) 6 and seven mutant enzymes of it. These enzymes were expressed in E. coli M15 strain using the expression vector pQE-60. Purification of the muteins was facilitated by introducing a His-tag onto the C-terminus and linear- gradient elution with imidazole from Ni 2+ -nitrilotriacetic acid columns. Based on Coomassie-blue staining all these mutant enzymes showed high purity (Fig. 4). Because the K m value of the natural substrate of the synthase, hydro- quinone, is extremely small (< 1 l M ) and very difficult to measure, we used the substrate vanillin for the determin- ation of kinetic parameters and enzyme activity (K m , V max , k cat , k cat /K m ) of the muteins. But the specific enzyme activity could still be determined with hydroquinone as substrate. As control mutations we replaced some amino acids which were not discussed in the literature as important for enzyme activity and were also far from the NRDs and the His-Glu domains; Lys86 was changed to Ile and Ala204 to Val (Fig. 2, Table 1). The results indicated that enzyme activity was not drastically influenced by these replacements, probably because the mutations are far from the nucleotide recogni- tion sites. Exchange of Lys86 against the neutral Ile resulted in only a slight increase of the K m -value and approximately 18% decrease of V max . Replacement of Ala204 to Val caused a greater decrease of the catalytic efficiency (k cat /K m ), i.e. approximately threefold. The specific activity of this Fig. 2. Arbutin synthase amino acid sequence showing the NRD1bS (boxed) and NRD1bL (dot boxed) domains, putative DxD motifs (written white on black), and the H–E site in position 1 and position 2 (marked with arrows). Ó FEBS 2003 Probing catalytic domains of NRD glycosyltransferases (Eur. J. Biochem. 270) 535 mutant enzyme compared with the wild-type was still between 55 and 60% for the substrates, hydroquinone and vanillin. The following mutations did, however, give more intrigu- ing results. As recently suggested, conserved glutamic acids of the NRD region located at position 1 and 2 may occupy the role of catalytic residues for those glycosyltransferase reactions which proceed with retention of the sugar donor configuration (NRD1a family). Transferases catalyzing the inverting reaction (NRD1b family) are, however, characte- rized by His instead of Glu in position 1. In arbutin synthase this histidine is identified as His360 and the glutamic acid at position 2 as Glu368. Provided that the model of transferase mechanisms proposed in the literature is correct, substitution of Glu in position 2 for Asp in any member of the NRD1a or NRD1b family should result in a dramatic reduction in the reaction rate of such a mutein [8]. This suggestion is due to the assumption that in both reactions the Glu residue acts as Fig. 3. Sequence alignment, showing the PSPG-box [9,10], prepared with 24 plant-derived glucosyltransferase sequences. The alignment was created using the CLUSTALW program at the server of the EBI. *, homo- logous amino acids; Ô:Õ, conserved substitutions have been observed. RS: arbutin synthase (R. serpentina, AJ310148), AT I: putative gluco- syltransferase (GENE:AT2G23260) (A. thaliana, O22182), AT II: putative glucosyltransferase (GENE:AT2G23250) (A. thaliana, O22183), AT III: putative glucosyltransferase (GENE:AT2G23210) (A. thaliana, O22186), DB: betanidin 6-O-glucosyltransferase (D. bel- lidiformis, Q8W237), FI: flavonoid 3-O-glucosyltransferase (GENE: UFGT) (F. intermedia, Q9XF16), GM: putative glucosyltransferase (G. max, Q8S3B7), GT: flavonol 3-O-glucosyltransferase (G. triflora, Q96493), HV: flavonol 3-O-glucosyltransferase (H. vulgare), LE: putative glucosyltransferase (L. esculentum, Q8RXA4), ME: flavonol 3-O-glucosyltransferase 1 (M. esculenta, Q40284), NT I: glucosyl- transferase NTGT2 (N. tabacum, Q8RU71), NT II: UDP-glucose: salicylic acid glucosyltransferase (N. tabacum, Q9M6E7), PA: Gluco- syltransferase-14 (GENE:ADGT-14) (P. angularis, Q8S995), PF I: flavonoid 3-O-glucosyltransferase (P. frutescens, O04114), PF II: UDP-glucose:anthocyanin 5-O-glucosyltransferase (P. frutescens, Q9ZR27), PH: anthocyanin 5-O-glucosyltransferase (P. hybrida, Q9SBQ2), PL: putative glucosyltransferase (P. lunatus, Q8S3B5), SB: UDP-glucose: flavonoid 7-O-glucosyltransferase (S. baicalensis, Q9SXF2), SolB: UDPG glucosyltransferase (S. berthaultii, O24341), SOB: UDP-glucose glucosyltransferase (S. bicolour, Q9SBL1), ST: UDP-glucose glucosyltransferase (S. tuberosum, P93789), VV: UDP flavonoid 3-O-glucosyltransferase (V. vinifera, O22304), ZM: flavonol 3-O-glucosyltransferase (Z. mays, P16166). Fig. 4. Purity of arbutin synthase wild-type and mutant enzymes after Ni 2 -nitrilotriacetic acid chromatography SDS/PAGE and staining by Coomassie-blue. (I, marker proteins; II, AS-WT; III, AS-L86I; IV, AS-A204V; V, AS-E368D; VI, AS-E368A; VII, AS-H360K; VIII, AS-H360R; IX, H360E; w, arbutin synthase and its muteins). Table 1. Comparison of kinetic parameters of wild-type and muteins of arbutin synthase-(His) 6 expressed in E. c oli. Values of K m and k cat were calculated from Lineweaver–Burk plots using vanillin as substrate (n.d. ¼ not detectable, detection limit < 10 pkatÆmg )1 ). Enzyme arbutin synthase K m vanillin [lmolÆL )1 ] V max vanillin [pkat] Specific activity hydroquinone [nkatÆmg )1 ] Specific activity vanillin [nkatÆmg )1 ] k cat vanillin [s )1 ] k cat /K m vanillin [LÆmol )1 Æs )1 ] Wild type 440 46.5 202.9 24.1 2.59 5886.4 L86I 452 38.3 180.4 20.4 1.84 4070.8 A204V 706 12.3 139.1 13.1 1.30 1841.4 E368D 411 1.9 17.8 0.9 0.12 292.0 E368A 398 6.2 24.0 1.1 0.08 201.0 H360R – – 0.22 n.d. – – H360K – – 0.17 n.d. – – H360E – – 0.05 n.d. – – 536 T. Hefner and J. Sto ¨ ckigt (Eur. J. Biochem. 270) Ó FEBS 2003 the nucleophile. We were able to prove for the first time the suggested model by site-directed mutagenesis of arbutin synthase. First the E368D mutein was generated. Deter- mination of the kinetic properties of this mutein indeed indicated a clear decrease of activity, e.g. of the specific enzyme activity more than 10- and 25-fold for hydroqui- none and vanillin, respectively. Also the k cat /K m value for the substrate vanillin decreased dramatically in comparison withthewild-type(Table1).InthecasethattheGlu(orthe Asp) residue at this position is really crucial for the sugar transfer as a nucleophile, its exchange against another amino acid, e.g. by a neutral one, must lead to a total loss of enzyme activity as it has been discussed [8]. As a consequence we created and tested the mutant arbutin synthase E368A, in which the putative Glu was exchanged by Ala. But this mutein still exhibited remarkable enzyme activity (Table 1). Each of the measured kinetic parameters of this mutant enzyme with regard to vanillin were in the same range as those obtained for the former mutant enzyme E368D. The specific activities of mutein E368A were even slightly higher with both substrates hydroquinone and vanillin than those of E368D. This mutagenesis experiment obviously excludes Glu368 as the nucleophile. A nucleophilic residue is, however, a prerequisite for the S N 2 reaction, leading from the a-con- figured UDP-glucose to the b-configured glucosylated product. As shown by our alignment study Glu384 might remain as an appropriate candidate (Fig. 3). But Glu384 (in AS) is far from Glu368 and appears not to be strictly conserved. Whereas it is detected in five of the 24 sequences it is replaced by its homologue Asp in the remaining 19 enzymes. Future experiments must show, whether the both residues at position 384 provide the nucleophile for the reaction instead of Glu368 or whether an acidic amino acid outside the PSPG-box may occupy the nucleophilic role. For deeper investigation of the properties of the mutein AS-E368A, an additional eight substrates were tested (Fig. 5). By determination of the specific activities towards these substrates, it was possible to compare these with the activities obtained with the AS wild-type enzyme (Table 2). Surprisingly changes of the relative enzyme activities were not at the same level. Eugenol, for instance, was not glucosylated at all. If substitution of Glu by Ala causes an affect at the acidic catalytic centre only, activities towards different substrates should not change. By the obtained results we may conclude, that replacing Glu by Ala does not only affect the NRD, but also has an influence on recognition of the substrates. It may be that the observed effect is due to alteration of the stereochemical and electronic situation at the substrate binding pocket, but at the present time these domains of glucosyltransferases are also unknown. Therefore any conclusions drawn on correlations between enzyme acti- vity towards different substrates and mutations must be considered tentative until confirmed by X-ray crystallo- graphic analyzes. Mutations at the second typical and conserved residue in position 1, which is His360 in arbutin synthase, did not support at all the suggested reaction mechanism model. For this model it has been assumed that mutation of His by other basic amino acids such as Arg or Lys would probably be tolerated by enzymes of the NRD1 family. In contrast to this theory, the appropriate mutant enzymes of arbutin synthase, H360R and H360K, showed such small conversion rates that determinations of K m , V max or k cat values were not attainable and only measurement of the Table 2. Specific activities of AS-WT and E368A. For better comparison, the relative activities of the wild-type enzyme were divided through the activity values obtained by the mutein E368A. Substrate AS-WT specific activity [nkatÆmg )1 ] relative activity [%] AS-E368A specific activity [nkatÆmg )1 ] relative activity [%] relative activity WT/ relative activity E368A Hydroquinone 208.9 100 28.9 100 1.0 b-Naphthol 38.4 18.4 2.1 7.3 2.5 8-Hydroxyquinoline 24.9 11.9 1.0 3.5 3.4 3-Methoxyphenol 24.7 11.8 0.6 2.1 5.6 Vanillin 22.2 10.6 1.8 6.2 1.7 Resorcinol 16.3 7.8 0.8 2.8 2.8 4-Hydroxybenzaldehyde 15.3 7.3 0.8 2.8 2.6 Umbelliferone 10.8 5.1 0.9 3.1 1.6 Eugenol 8.5 4.1 < 0.03 < 0.1 > 40 Scopoletin 7.4 3.5 0.6 2.1 1.7 Fig. 5. Structures of substrates that were tested with arbutin synthase wild-type and E368A mutein. Ó FEBS 2003 Probing catalytic domains of NRD glycosyltransferases (Eur. J. Biochem. 270) 537 specific activities of both mutants with the natural highly accepted substrate hydroquinone (K m <1l M ) was pos- sible. However, the measured enzyme activities were approximately 1000-fold diminished compared to the wild-type or even 100-fold smaller as determined for the above discussed mutein E368A. We therefore believe, that the functional role of His in position 1 of NRD1b family members is apparently more crucial than has previously been accepted. This observation is additionally supported by a further mutation experiment. If glycosyl- transferases, which catalyze sugar transfer with retention of configuration, also depend on the presence of a glutamic acid residue in the same position as histidine occupies in inverting transferases (position 360), it would be an exciting challenge to convert an inverting to a retaining enzyme just by such a point mutation. For that reason we created the H360E mutant enzyme of arbutin synthase which however, exhibited an  4000-fold decreased enzyme activity. This enzyme did in fact, reveal the lowest specific activity of all the mutant enzymes described here. Nevertheless, arbutin could be enzymatically synthesized with high amounts of this mutein (approximately 50-fold compared to the standard assay) and much longer incubation times (approximately 200-fold) because of excellent enzyme expression and simple purification. The isolated and purified glucosidic product was clearly identified as the O-b-glucoside of hydroquinone, because it resisted incuba- tion in the presence of a-glucosidase. In contrast, it was completely hydrolyzed in the presence of b-glucosidase as shown by TLC and HPLC analysis. This experiment is again not in agreement with the suggested mechanism for the glycosyltransferase reaction which, would lead to the hydroquinone-O-a)glucoside and not to the O-b-glucoside arbutin as observed. Conclusions As arbutin synthase fulfils all the requirements of a member of the NRD1b enzyme family, the recent suggestion on catalytic important amino acids of glycosyltransferases, which is based on sequence alignment studies, is not satisfactory due to the results of the site-directed mutagen- esis experiments presented here. The question concerning the mechanism of one of the basic reactions in cells, the transfer of a sugar moiety during the formation of an a-or b-glucoside, still awaits an answer. Especially when consid- ering the results obtained by the mutein AS-E368A, where obviously the substrate recognition site was affected by this point-mutation. Crystallization and cocrystallization of such a glucosyltransferase with nucleotide sugars and substrates followed by X-ray analysis might be the best strategy for future success in elucidating the molecular nature of the glucosylation process. Acknowledgements The financial support provided by Deutsche Forschungsgemeinschaft (Bonn, Germany) and by the Fonds der Chemischen Industrie (Frankfurt/Main, Germany) is highly appreciated. We also thank J. Arend (Mainz) for advice in enzyme purification and W. E. Court (Mold) for linguistic help. References 1. Campbell, J.A., Davies, G.J., Bulone, V. & Henrissat, B. (1997) A classification of nucleotide-diphospho-sugar glycosyltransferases based on amino acid sequence similarities. Biochem. J. 326, 929– 939. 2. Campbell, J.A., Davies, G.J., Bulone, V. & Henrissat, B. (1998) A classification of nucleotide-diphospho-sugar glycosyltransferases based on amino acid sequence similarities. Biochem. J. 329, 719. 3. Arend, J., Warzecha, H. & Sto ¨ ckigt, J. (2000) Hydroquinone: O-glucosyl-transferase from cultivated Rauwolfia cells – Enrich- ment and partial amino acid sequences. Phytochemistry 53, 187–193. 4. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680–685. 5. Bradford, M.M. (1976) A rapid and sensitive method for quan- titation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 248–254. 6. Arend, J., Warzecha, H., Hefner, T. & Sto ¨ ckigt, J. (2001) Utilizing genetically engineered Escherichia coli to produce plant specific glucosides. Biotechnol. Bioeng. 76, 126–131. 7. Hefner, T., Arend, J., Warzecha, H., Siems, K. & Sto ¨ ckigt, J. (2002) Arbutin synthase, a novel member of the NRD1b glyco- syltransferase family, is an unique multifunctional enzyme con- verting various natural products and xenobiotics. Bioorgan. Med. Chem. 10, 1731–1741. 8. Kapitonov, D. & Yu, R.K. (1999) Conserved domains of glyco- syltransferases. Glycobiology 9, 961–978. 9. Vogt, T. & Jones, P. (2000) Glycosyltransferases in plant natural product synthesis: characterization of a supergene family. Trends. Plant Sci. 5 (9), 380–386. 10. Hughes, J. & Hughes, M.A. (1994) Multiple secondary plant product UDP-glucose glucosyltransferase genes expressed in cas- sava. DNA Seq. 5, 41–49. 11. Breton, C. & Imberty, A. (1999) Structure/function studies of glycosyltransferases. Curr.Opin.Struct.Biol.9, 563–571. 12. Breton, C., Bettler, E., Joziasse, D.H., Geremia, R. & Imberty, A. (1998) Sequence function relationships of prokaryotic and eukaryotic galactosyltransferases. J. Biochem. 123, 1000–1009. 13. Busch, C., Hofmann, F., Selzer, J., Munro, S., Jeckel, D. & Aktories, K. (1998) A common motif of eukaryotic glycosyl- transferases is essential for the enzyme activity of large clostridial cytotoxins. J. Biol. Chem. 273, 19566–19572. 14. U ¨ nligil, U.M. & Rini, J.M. (2000) Glycosyltransferase structure and mechanism. Curr.Opin.Struct.Biol.10, 510–517. 15. Wiggins, C.A.R. & Munro, S. (1998) Activity of the yeast MNN1 a-1,3-mannosyltransferase requires a motif conserved in many other families of glycosyltransferases. Proc. Natl Acad. Sci. 95, 7945–7950. 538 T. Hefner and J. Sto ¨ ckigt (Eur. J. Biochem. 270) Ó FEBS 2003 . Probing suggested catalytic domains of glycosyltransferases by site-directed mutagenesis Tobias Hefner and Joachim Sto¨ ckigt Lehrstuhl fu ¨ r Pharmazeutische. a member of the NRD1b family of glycosyltransferases. This enzyme was used to prove, by site-directed mutagenesis, suggested catalytic domains and reaction