Báo cáo khóa học: Active site residues and mechanism of UDP-glucose dehydrogenase ppt

Active site residues and mechanism of UDP-glucose dehydrogenase Xue Ge 1 , Lisa C. Penney 1 , Ivo van de Rijn 2 and Martin E. Tanner 1 1 Department of Chemistry, University of British Columbia, Vancouver, Canada; 2 Wake Forest University School of Medicine, Winston-Salem, NC, USA UDP-glucose dehydrogenase catalyzes the NAD + -depend- ent twofold oxidation of UDP-glucose to give UDP-glucu- ronic acid. A sequestered aldehyde intermediate is produced in the first oxidation step and a covalently bound thioester is produced in the second oxidation step. This work demon- strates that the Streptococcus pyogenes enzyme incorporates a single solvent-derived oxygen atom during catalysis and probably does not generate an imine intermediate. The reaction of UDP-[6¢¢,6¢¢-di- 2 H]- D -glucose is not accompan- ied by a primary kinetic isotope effect, indicating that hydride transfer is not rate determining in this reaction. Studies with a mutant of the key active site nucleophile, Cys260Ala, show that it is capable of both reducing the aldehyde intermediate, and oxidizing the hydrated form of the aldehyde intermediate but is incapable of oxidizing UDP- glucose to UDP-glucuronic acid. In the latter case, a ternary Cys260Ala/aldehyde intermediate/NADH complex is pre- sumably formed, but it does not proceed to product as both release and hydration of the bound aldehyde occur slowly. A washout experiment demonstrates that the NADH in this ternary complex is not exchangeable with external NADH, indicating that dissociation only occurs after the addition of a nucleophile to the aldehyde carbonyl. Studies on Thr118Ala show that the value of k cat is reduced 160-fold by this mutation, and that the reaction of UDP-D-[6¢¢,6¢¢-di- 2 H]- glucose is now accompanied by a primary kinetic isotope effect. This indicates that the barriers for the hydride transfer steps have been selectively increased and supports a mech- anism in which an ordered water molecule (H-bonded to Thr118) serves as the catalytic base in these steps. UDP-glucose dehydrogenase catalyzes an NAD + -depend- ent twofold oxidation of UDP-glucose to generate UDP- glucuronic acid (Fig. 1) [1]. In mammals, UDP-glucuronic acid is used in the biosynthesis of hyaluronan and various glycosaminoglycans such as heparin sulfate and chondroitin sulfate [2]. In addition, it is used in the liver where the act of glucuronidation targets molecules for excretion [3]. UDP- glucuronic acid also serves as a precursor to UDP-xylose which provides a major component of the cell wall polysac- charides in plants [4]. In many strains of pathogenic bacteria, such as group A streptoccoci and Streptococcus pneumoniae type 3, UDP-glucuronic acid is used in the construction of the antiphagocytic capsular polysaccharide [5,6]. This cap- sule protects the bacteria from the immune system of the host and thus serves as a major virulence factor. UDP-glucose dehydrogenase is of mechanistic interest because it belongs to a family of sugar nucleotide-modifying enzymes that catalyze a net four-electron oxidation and thus effectively serve as both alcohol dehydrogenases and aldehyde dehydrogenases [7]. Other members of this family include UDP-ManNAc dehydrogenase [8] and GDP-mannose dehydrogenase [9,10]. Extensive studies on both the bovine and S. pyogenes UDP-glucose dehydro- genases have led to the mechanistic pathway outlined in Fig. 1. The enzyme is thought to follow a Bi-Uni-Uni-Bi ping-pong mechanism in which UDP-glucose is bound first and UDP-glucuronic acid is released last [11,12]. The first oxidation involves the transfer of the C-6¢ pro-R hydride of UDP-glucose to the si face (B face) of NAD + to form NADH and an aldehyde intermediate [13,14]. This inter- mediate is bound tightly to the enzyme and is not accessible to external aldehyde-trapping reagents [15,16]. Nevertheless, it has been demonstrated that synthetic samples of the aldehyde intermediate will serve as a kinetically competent substrate for the second step of the reaction [17]. It is quite likely that the bound form of the aldehyde exists largely as a covalent hemithioacetal adduct as this species has been implicated as an intermediate in the second step of catalysis, although an imine linkage via an active site lysine has also been proposed [18]. The second oxidation step involves the addition of a cysteine thiol to the aldehyde to generate the thiohemiacetal intermediate with subsequent hydride trans- fer to the second NAD + molecule [19,20]. The resulting thioester is hydrolyzed in a final step to generate the product, UDP-glucuronic acid. Strong evidence in support of this covalent catalysis mechanism was obtained when a Cys260Ser mutant of the S. pyogenes enzyme was incubated with UDP-glucose and NAD + [20]. The mutant enzyme was essentially inactive; however, mass spectral analysis indicated that a covalent adduct had accumulated in which UDP-glucuronic acid was attached to Ser260 via an ester linkage. The mutant enzyme was therefore capable of catalyzing both oxidation steps of the reaction but was incapable of hydrolyzing the unnatural ester bond at any significant rate. Interestingly, when the Correspondence to M. E. Tanner, Department of Chemistry, University of British Columbia, 2036 Main Mall, Vancouver, British Columbia, Canada V6T 1Z1. Tel.: +1 604 822 9453, Fax: +1 604 822 2847, E-mail: mtanner@chem.ubc.ca Abbreviations: GAPDH, glyceraldehyde phosphate dehydrogenase. Enzyme: UDP-glucose dehydrogenase (EC 1.1.1.22). (Received 29 August 2003, revised 9 October 2003, accepted 14 October 2003) Eur. J. Biochem. 271, 14–22 (2004) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03876.x Cys260Ala mutant was examined, the enzyme was essen- tially inactive towards UDP-glucose oxidation but readily oxidized the intermediate aldehyde at a rate within an order of magnitude of that seen with the wild-type enzyme. Apparently, the normal reaction proceeds via covalent catalysis; however, with the alanine mutant the second oxidation step can proceed directly from the hydrated aldehyde (Fig. 2). In subsequent work, X-ray structures were solved of the native and Cys260Ser dehydrogenases from S. pyogenes in complex with UDP-xylose/NAD + and UDP-glucuronic acid/NAD(H), respectively [21]. In each case, Cys/Ser260 was positioned appropriately to participate in covalent catalysis, as expected. Other active site residues that could potentially play roles in the hydride transfer and/or hydro- lysis steps were also identified. These include Thr118, Glu141, Glu145, Lys204, Asn208, and Asp264. All but Glu145 are strictly conserved among all family members. Inspection of the hydrogen bonding networks at the site that would be occupied by the C-6¢ hydroxyl of UDP-glucose in the Michaelis complex led to the proposal of two scenarios for the hydride transfer steps (presumably both hydride transfer steps will employ the same catalytic residues). In the first scenario, Lys204 acts as the catalytic base that deprotonates the C-6¢ hydroxyl, whereas Asn208 and an ordered water molecule serve as hydrogen bond donors to the hydroxyl oxygen (Fig. 3A). In the second scenario, the ordered water molecule serves as the catalytic base with the assistance of Asp264 as the ultimate proton acceptor (Fig. 3B). In this case, Asn208 and Lys204 (presumably in the ammonium form) serve as hydrogen bond donors to the hydroxyl oxygen, and Thr118 and a ribose hydroxyl of NAD + form hydrogen bonds with the ordered water. The identity of the catalytic residues involved in hydrolysis of the thioester intermediate are harder to predict from these structures; however, Glu141 and Glu145 are potential candidates. In this work, the possibility of imine formation involving Lys204 is examined with the use of H 2 18 O, and the nature of the rate determining steps of catalysis is probed using deuterated UDP-glucose. Studies on the Cys260Ala mutant and other active site mutants are described which help to further delineate the roles these residues play in catalysis. Experimental procedures General procedures UDP-glucose dehydrogenase from S. pyogenes was expressed in Escherichia coli using the plasmid pGAC147 as described elsewhere [5]. All proteins were purified as described previously [12] and analyzed by ESI-MS to ensure they had the expected molecular masses. All enzymatic assays for dehydrogenase activity used the method des- cribed previously [12], unless otherwise indicated. The aldehyde intermediate, uridine diphospho-a- D -gluco- hexodialdose, was synthesized as described previously [17], and stock concentrations were calculated from measure- ments of A 262 using a value of e 262 ¼ 8700 M )1 for the uridine chromophore. H 2 18 O (95% enriched) was from Cambridge Isotope Laboratories. [6,6-di- 2 H]- D -Glucose (98% enriched) was from Aldrich. Commercial enzymes and sugar nucleotides were from Sigma or Boehringer Mannheim Biochemicals, unless stated otherwise. Site-directed mutagenesis and mutant protein expression The mutagenesis protocol to produce the active site mutants followed an adaptation of enzymatic PCR [22] using plasmid pGAC147 [5] which contains a copy of hasBas template. The primers for the construction of the mutant plasmids were as follows: T118A, 5¢-CGCGGTACCAAC TCTTATCATCAAATCAGCAATTC-3¢ and 5¢-ATAG GTACCATGGCTATTAACGCTTAGTACC-3¢; E141Q, 5¢-CGCGGTACCTTTATATGACAACTTATATC-3¢ and 5¢-ATAGGTACCTTTAGATTCTCTTAAAAACTG AGGGC-3¢; E145Q, 5¢-CGCGGTACCTTTATATGAC AACTTATATC-3¢ and 5¢-ATAGGTACCTTTAGACTG TCTTAAAAATTCAG-3¢;K204A,5¢-CGCGGTACCC AATACTTATTTAGCG-3¢ and 5¢-ATAGGTACCAAA TAGTGCTACTGCTTCAGC-3¢; N208A, 5¢-CGCGGTA CCGTTAAGGGTAGC-3¢ and 5¢-ATAGGTACCTAAA Fig. 1. The mechanism of the reaction cata- lyzed by UDP-glucose dehydrogenase. Ó FEBS 2003 Active site residues of UDP-glucose dehydrogenase (Eur. J. Biochem. 271)15 TAAGTAGCGGCAAATAG-3¢; D264N, 5¢-CGCGGTA CCCAATTATTGGCAAATTAC-3¢ and 5¢-ATAGGTAC CTTCGTGTTTTTAGGTAGACA-ATAAC-3¢.Nucleo- tides designated in bold led to the desired mutations. Cys260Ala was expressed using the plasmid pGAC400 as described previously [20]. All mutations and constructs were confirmed by sequencing the entire gene. Expression experiments of the mutant constructs were initiated by transforming each plasmid into E. coli JM109(DE3) and colonies were grown overnight in TYPG (3 mL). The TYPG media contained 8 g of tryptone, 8 g of yeast extract, 2.5 g of NaCl, 1.25 g of K 2 HPO 4 and 2.5 g of glucose per 500 mL of distilled water. For inductions, either the JM109(DE3) culture harboring pGAC147 (3 mL) or the JM109(DE3) culture harboring the desired mutant construct (0.5 mL) was inoculated into TYPG (250 mL) and grown at 37 °C with vigorous shaking until the cultures reached D 600 ¼ 0.8–1.2. An aliquot of the culture (10 mL) was removed just prior to induction by isopropyl thio-b- D -galactoside (0.4 m M final concentration) and the incubation was resumed. Following induction aliquots of culture (10 mL) were removed at 1-h intervals, rapidly chilled on ice, and a portion (1.5 mL) was prepared for SDS/PAGE analysis. Bacteria were sedimented at 13 000 g (10 s), the pellets were resuspended in TE (150 lL; 10 m M Tris pH 7.5, 1 m M EDTA), and finally the cells were solubilized by boiling for 5 min in the presence of 6 · loading buffer (30 lL [23]). SDS/PAGE was performed by the method of Laemmli [24] using 10% gels and loading 20 lL lysate per lane to confirm the overexpression of the mutant constructs. Solvent isotope incorporation study A solution of sodium phosphate buffer (50 m M , 500 lL), pH 8.0, containing 50% H 2 18 O (v/v), UDP-glucose (50 m M ), NAD + (10 m M ), FMN (10 m M ), dithiothreitol (2 m M ), and UDP-glucose dehydrogenase (0.35 mg) was incubated at 37 °C for 20 h. At periodic intervals the reaction was gently mixed and exposed to atmospheric oxygen. After the incubation, EDTA (16 m M ) was added and the mixture was lyophilized to dryness. The residue was redissolved in D 2 Oand 13 C NMR spectra were recorded. A control sample was run containing 100% H 2 16 O. For further purification, the samples were applied to a column of Bio-Gel P-2 resin (40 mL) and eluted with water. Fractions containing the UDP-glucuronic acid product were lyophilized and submitted for MALDI MS analysis. Kinetic isotope effect studies UDP-[6¢¢,6¢¢-di- 2 H]- D -glucose was prepared using a slight modification of a procedure previously described for the preparation of tritiated UDP-glucose [25]. A solution of Tris/HCl buffer (70 m M , 40 mL) pH 7.8, containing [6,6-di- 2 H]- D -glucose (8 mg, 44 lmol, 1.1 m M ), ATP (73 mg, 132 lmol), UTP (64 mg, 132 lmol), glucose 1,6-diphosphate (0.165 lmol), MgSO 4 (36 mg, 146 lmol), hexokinase (66 U), phosphoglucomutase (109 U), UDP- glucose pyrophosphorylase (12.5 U), and inorganic pyro- phosphatase (33 U) was incubated at 30 °Cfor20h.The reaction mixture was applied to a column of DE-52 anion exchange resin (65 mL) and eluted with a linear gradient of 0–400 m M triethylammonium bicarbonate (800 mL total). Fractions containing UDP-glucose (as assayed by UDP- glucose dehydrogenase and NAD + ) were pooled, lyophi- lized, redissolved in water, and lyophilized a second time. The product was redissolved in water and passed through a column of Amberlite IR-120 (plus) resin (10 mL, Na + form, eluted with water). The resulting solution was applied to a column of Bio-Gel P-2 resin (2.5 · 45 cm) and eluted with water. UV active fractions were pooled and lyophilized to dryness to give the disodium salt of UDP-[6¢¢,6¢¢-di- 2 H]- D -glucose as a white solid (29 lmol, 66% yield assayed enzymatically, see below). 1 H NMR and mass spectra were consistent with the assignment of the product as UDP- [6¢¢,6¢¢-di- 2 H]- D -glucose, and indicated the extent of deuter- ium incorporation to be > 95%; LSI(–) MS (thioglycerol) m/z 567 (M(di- 2 H)–H + , 100%). The rates of the dehydrogenase reaction were determined under saturating conditions for both substrates using a previously described kinetic assay [12]. The values of [UDP- glucose] (deuterated or undeuterated) and [NAD + ]were as follows: Wild-type enzyme (0.5 m M ,1.2m M ), T118A (0.5 m M ,4m M ), E141Q (0.5 m M ,2m M ), and E145Q (0.8 m M ,2m M ). The error reported for the kinetic isotope effects are the SD of the data points from the average determined ratios (five independent measurements). The concentrations of the stock UDP-glucose solu- tions (both deuterated and undeuterated) were determined Fig. 3. Two scenarios outlining the residues involved in catalyzing the hydride transfer steps. (A) Lys204 serving as the key acid/base residue. (B) An ordered water molecule and Asp264 serving as the key acid/ base residue. X ¼ H for the first oxidation step, X ¼ SR (from Cys260) for the second oxidation step. Fig. 2. The proposed mechanism for the oxidation of the hydrated aldehyde intermediate by the Cys260Ala mutant. 16 X. Ge et al. (Eur. J. Biochem. 271) Ó FEBS 2003 enzymatically by running the UDP-glucose dehydrogenase reaction to completion. The substrate concentrations were calculated from the changes in A 340 assuming a stoichio- metry of two equivalents of NADH produced for every molecule of UDP-glucose consumed, and an extinction coefficient for NADH of e ¼ 6220 M )1 . Cys260Ala studies Initial velocity kinetics for the reduction of the aldehyde intermediate. Assays were performed at 30 °CinTrien/ HCl buffer (50 m M , 0.50 mL total volume) pH 8.7, con- taining NADH (0.15 m M ) and dithiothreitol (2 m M )with varying amounts of aldehyde. Initial velocities were meas- ured during the first 60 s after initiation with the Cys260Ala mutant and calculated from the decrease in A 340 using an extinction coefficient for NADH of e ¼ 6220 M )1 . Full time course analysis of incubations with the aldehyde intermediate and NADH. Thealdehydeintermediate (75 l M with wild-type enzyme, 92 l M with Cys260Ala) was incubated at 30 °C with NADH (0.15 m M ) and either wild-type dehydrogenase (0.22 l M )ortheCys260Ala mutant (0.22 l M ) in Trien/HCl buffer (50 m M ,0.50mL total volume), pH 8.7, containing dithiothreitol (2 m M ). The changes in NADH concentrations were monitored using A 340 assuming an extinction coefficient for NADH of e ¼ 6220 M )1 . To confirm the identity of the products at the end-point of these reactions, aliquots were analyzed by ion- paired reversed phase HPLC as described previously and compared with authentic standards [14]. Deuterium washout experiment. A solution of Trien/HCl buffer (50 m M , 2 mL total volume) pH 8.7, containing UDP- D -[6¢,6¢-di- 2 H]-glucose (2 m M ), NADH (20 m M ), NAD + (4 m M ), dithiothreitol (2 m M ), and the Cys260Ala mutant (0.1 mg) was incubated at 30 °C for 23 h. The resulting solution was applied to a column of DE-52 anion exchange resin (60 mL) and eluted with a linear gradient of 0–300 m M ammonium bicarbonate (800 mL total). Frac- tions containing UDP-glucose (as assayed by UDP-glucose dehydrogenase and NAD + ) were pooled and lyophilized. The resulting solids were dissolved in water and passed through a column of Amberlite IR-120 (plus) resin (10 mL, Na + form, eluted with water). The eluent was lyophilized to dryness, redissolved in water, applied to a column of Bio- Gel P-2 resin (2.5 · 45cm),andelutedwithwater.UV active fractions were pooled and lyophilized to dryness. The UDP- D -[6¢¢,6¢¢-di- 2 H]-glucose obtained in this fashion was analyzed for deuterium content by 1 H NMR spectroscopy and LSI-MS, and showed identical spectral characteristics to the original labeled material. Results Solvent 18 O-isotope incorporation study The fact that the aldehyde intermediate in the UDP-glucose dehydrogenase reaction is never released into solution and is inaccessible to external trapping reagents led to the sugges- tion that this species is covalently bound to an active site lysine residue via an imine linkage [18]. This suggestion was supported by experiments on the bovine liver enzyme in which the active site cysteine thiol had been chemically modified to a thiocyanide moiety. When this inactivated form of the enzyme was incubated with UDP-glucose and NAD + (or with the aldehyde intermediate alone) and then treatedwithNaBH 4 , a covalent enzyme-substrate adduct was generated. The properties of the adduct were consistent with that expected for a reduction product of a Schiff’s base formed between the aldehyde intermediate and an active site lysine, and the authors concluded that an imine was formed on the normal reaction pathway. The suggestion was somewhat at odds with earlier 18 O-isotopic labeling studies showing that the reaction proceeds with the incorporation of only one solvent-derived oxygen atom into the carboxy- late product [26]. The proposal of an imine intermediate could only be consistent with this observation if the original C6¢¢ oxygen atom was sequestered in the active site during the second oxidation step and then re-delivered during hydrolysis of the thioester. More recent work has focused on the bacterial enzyme from S. pyogenes, and the X-ray crystal structure has shown that Lys204 is in the active site and is in close proximity to C6¢ of UDP-glucose [21]. In order to re-examine the possibility of Schiff base formation with this enzyme, the reaction was carried out in H 2 18 O water and the UDP- glucuronic acid produced was examined for 18 O-isotope content. Samples of UDP-glucose were incubated with substoichiometric amounts of NAD + in the presence of FMN and oxygen [27,28]. The FMN/O 2 acts as an NAD + regenerating system and facilitates purification of the UDP- glucuronic acid from the minor dinucleotide contaminants. Reactions were carried out in both H 2 16 Oand50% H 2 16 O/50% H 2 18 O, and the resulting UDP-glucuronic acid was analyzed by MALDI TOF MS. In the latter case, signals corresponding to both unlabeled (m/z 648.3, M-2H + +3Na + ) and singly 18 O-labeled product (m/z 650.3) were observed in approximately equal amounts, confirming that the reaction was accompanied by incor- poration of a single solvent-derived oxygen atom. In order to determine the position of the incorporated 18 O-atom, both samples were analyzed by 13 C NMR spectroscopy. In the control sample, a single carboxylate signal is observed at 176.671 p.p.m., whereas, in the 50% H 2 16 O/50% H 2 18 O sample two signals of approximately equal intensity were observed at 176.673 and 176.647 p.p.m. (Fig. 4). The 0.026 p.p.m. separation of the signals in the latter sample is consistent with the magnitude expected for an 18 O-isotope induced shift in a mono-labeled carboxylate, and confirms that the label was introduced into the carboxylate of UDP-glucuronic acid [28]. Deuterium kinetic isotope effect study In order to determine if either hydride transfer step in the reaction catalyzed by UDP-glucose dehydrogenase was rate determining, samples of 6¢¢,6¢¢-dideuterated- UDP-glucose were prepared and analyzed for the pres- ence of a primary kinetic isotope effect on catalysis. Since both C6¢¢ positions were labeled, a single experi- ment could be used to simultaneously probe both hydride transfer steps. UDP-[6¢¢,6¢¢-di- 2 H]- D -glucose was prepared enzymatically from [6,6-di- 2 H]- D -glucose using Ó FEBS 2003 Active site residues of UDP-glucose dehydrogenase (Eur. J. Biochem. 271)17 hexokinase, phosphoglucomutase, UDP-glucose pyro- phosphorylase, and inorganic pyrophosphatase. The resulting UDP-[6¢¢,6¢¢-di- 2 H]- D -glucose was found to be > 95% enriched with two deuterium labels when ana- lyzed by MS. The rates of the UDP-glucose dehydro- genase reaction for both the labeled and unlabeled substrates were determined under saturating conditions and the value of k H /k D was found to be 1.1 ± 0.1. This demonstrates that there is no primary kinetic isotope effect on the reaction of this substrate and therefore the hydride transfer steps are not rate determining in this reaction. Site-directed mutagenesis In an effort to further understand the roles of the active site residues in catalysis several mutant proteins were targeted for study. Plasmids encoding Thr118Ala, Glu141Gln, Glu145Gln, Lys204Ala, Asn208Ala, Cys260Ala, and Asp264Asn were generated and shown to result in high levels of protein expression. Unfortunately, with several of the key mutants (K204A, N208A, and D264N), all of the expressed protein was produced in insoluble inclusion bodies and attempts to solubilize these proteins were unsuccessful. With Cys260Ala, much of the protein was present in inclusion bodies, however, enough remained in solution to allow for the purification and study of this mutant. Thr118Ala, Glu141Gln, and Glu145Gln were soluble and could be isolated in good yield. Studies on Cys260Ala Cys260 provides the key active site thiol involved in covalent catalysis with this enzyme and thus mutants of this residue warrant further investigation. In previous studies, the Cys260Ala mutant was reported to be essentially inactive towards the oxidation of UDP-glucose (< 0.01% activity of wild-type enzyme) [20]. However, the mutant was quite capable of catalyzing the oxidation of the aldehyde inter- mediate at rates within an order of magnitude of that of thewild-typeenzyme(k cat ¼ 0.19 s )1 , K M ¼ 0.26 m M vs. k cat ¼ 1.2 s )1 , K M ¼ 0.014 m M , respectively). Since this mutant is presumably incapable of participating in covalent catalysis, it would appear that it simply binds the hydrated form of the aldehyde from solution and oxidizes it directly to UDP-glucuronic acid (Fig. 2). The question remains, however, as to why UDP-glucose is not a substrate for the Cys260Ala mutant. One possibility is that the Cys260 thiol plays a key role in the first oxidation step of the reaction, in addition to its role in covalent catalysis during the second oxidation step. An alternate possibility is that Cys260 is not required for the first oxidation step and the aldehyde intermediate is readily formed by the mutant enzyme; however, it is tightly bound and there is no mechanism by which it can be hydrated and proceed forward to the second oxidation step. In order to address the previous question, the ability of the Cys260Ala mutant to catalyze the reverse of the first step in catalysis, namely reduction of the aldehyde intermediate, was examined. The aldehyde intermediate was generated via chemical synthesis and was found to be an excellent substrate for reduction by NADH using the Cys260Ala mutant. The values of k cat and K M for the aldehyde were determined (at 0.15 m M NADH) to be 1.9 ± 0.1 s )1 and 58 ± 7 l M , respectively (it was not possible to measure these values with the wild-type enzyme due to a dismutation process, vide infra). It should be noted that the value of K M essentially represents an Ôapparent K M Õ since the majority of the aldehyde in solution exists as a hydrate [17] and the reduction process requires that the enzyme bind the unhydrated form of the aldehyde. The observation that Cys260Ala catalyzes a reasonably rapid reduction of the aldehyde clearly demonstrates that Cys260 does not play a key role in catalysis of the first oxidation step. It also means that Cys260Ala must be capable of catalyzing the reverse reaction, namely the oxidation of UDP-glucose to give the free aldehyde. The Haldane equation dictates that the extremely slow rate of this oxidation process must ulti- mately be attributed to an unfavorable equilibrium constant reflecting the higher energies of the free aldehyde and NADH. These observations are further supported by monitoring the full time course of the reaction of the aldehyde intermediate with excess NADH (Fig. 5). When the wild- type enzyme is the catalyst, an initial decrease in A 340 is observed due to the reduction of the aldehyde that generates NAD + and UDP-glucose. As NAD + accumulates, how- ever, it is possible for the enzyme to oxidize either the aldehyde intermediate or UDP-glucose to UDP-glucuronic acid and regenerate NADH. Ultimately, the A 340 value returns to its initial position, indicating there is no net consumption of NADH. This phenomenon can be under- stood by considering a dismutation process in which the aldehyde disproportionates between the alcohol and the acid without consuming NADH (Fig. 6). The thermodynamic stability of UDP-glucuronic acid ultimately drives the Fig. 4. 13 C NMR spectra of the carboxyl group of UDP-glucuronic acid (A) generated in 50% H 18 2 O/50% H 16 2 O and (B) generated in 100% H 16 2 O. 18 X. Ge et al. (Eur. J. Biochem. 271) Ó FEBS 2003 dismutation to completion. With the Cys260Ala mutant, a rapid decrease in A 340 is observed due to reduction of the aldehyde intermediate and the consumption of NADH (Fig. 5). In this case, however, the NADH is not regenerated by the dismutation process and the value of A 340 does not return to its original position. This occurs because the UDP- glucose that is generated by reduction of the aldehyde cannot be oxidized by this mutant and is therefore kinetically trapped in that form. By calculating the amount of NAD + consumed, it was found that 80% of the aldehyde was converted to UDP-glucose in this fashion. The remaining 20% of the aldehyde did undergo dismutation since a fraction of it was oxidized directly to the acid by the enzyme and the accumulating NAD + .Inbothexperiments,HPLC analysis with authentic standards of the UDP-sugars was used to confirm the expected product distributions. This kinetic behavior is entirely consistent with the curious observation that Cys260Ala can both oxidize the aldehyde to the acid and reduce it to the alcohol, but that it cannot oxidize the alcohol to the acid at any appreciable rate. Since the Cys260Ala mutant is capable of catalyzing the first hydride transfer step, the ternary complex of bound NADH and aldehyde intermediate must be formed upon incubation of the alcohol and NAD + . In order to proceed with the second oxidation step, however, a nucleophile must add to the aldehyde, and in the absence of the active site thiol a water molecule must assume this role. If neither release, nor hydration, of the bound aldehyde can proceed at a reasonable rate, no turnover would be observed and the only fate of the bound species would be to return to starting materials via the reverse reaction. In order to probe further the nature of this ternary complex an experiment was devised to determine whether the bound NADH could exchange into bulk solution at any measurable rate. A sample of UDP- D -[6¢¢,6¢¢-di- 2 H]-glucose was incubated with the Cys260Ala mutant in the presence of 20 m M NADH and 4 m M NAD + . An initial oxidation event should produce a ternary complex of the monodeuterated aldehyde and NAD 2 H. If the bound NAD 2 H could exchange with unlabeled NADH, then back reaction would lead to the formation of monodeuterated UDP-glucose, and a net ÔwashoutÕ of the pro-R deuterium should be observed. Even upon extensive incubations, however, no significant loss of deuterium label could be detected in the recovered UDP- D - [6¢¢,6¢¢-di- 2 H]-glucose as analyzed by both MS and 1 HNMR spectroscopy. This indicates that the NADH is not released into solution from the mutant ternary complex and suggests that in the wild-type reaction, the first-formed NADH is not released until the Cys260 thiolate adds to bound aldehyde and generates the thiohemiacetal intermediate. Studies on Thr118Ala, Glu141Gln, and Glu145Gln The three remaining mutants that could be obtained in a soluble form were analyzed for their ability to catalyze the UDP-glucose dehydrogenase reaction (Table 1). E141Q and E145Q showed similar behavior in that the values for Fig. 6. The dismutation of the aldehyde intermediate during an incuba- tion with NADH and wild-type UDP-glucose dehydrogenase. Boxed structures represent species present at the completion of the reaction. Fig. 5. Kinetic trace following the full time-course of the incubation of the aldehyde intermediate and NADH with either wild-type UDP- glucose dehydrogenase (solid line) or the Cys260Ala mutant (dashed line). Table 1. Kinetic constants and kinetic isotope effects for the reactions catalyzed by the wild-type and mutant UDP-glucose dehydrogenases. Enzyme UDP-Glc K M (l M ) NAD + K M (l M ) k cat (s )1 ) k H /k D a Wild-type 20 ± 4 65 ± 6 1.8 ± 0.1 1.1 ± 0.1 T118A 59 ± 9 400 ± 100 0.011 ± 0.003 1.9 ± 0.1 E141Q 60 ± 9 135 ± 7 0.14 ± 0.02 1.4 ± 0.1 E145Q 125 ± 24 187 ± 20 0.20 ± 0.04 1.4 ± 0.1 a Rate of UDP-glucose oxidation vs. the rate of UDP-[6¢¢,6¢¢- di- 2 H]glucose oxidation measured under saturating conditions (see Experimental procedures). Ó FEBS 2003 Active site residues of UDP-glucose dehydrogenase (Eur. J. Biochem. 271)19 k cat were  10-fold lower than those obtained with the wild- type dehydrogenase, and the values for K M were two- to six- fold higher. The modest changes in the catalytic constants suggest these active site residues do not play key roles in either catalysis or binding. In the case of Thr118Ala, the value of k cat dropped by 160-fold, whereas the K M values increased slightly. This indicates that Thr118 is reasonably important for catalysis, although it is probably not serving as a key acid/base catalyst or nucleophile. In order to probe the nature of the steps that were affected by the mutations, the mutants were examined for the possible presence of a primary kinetic isotope effect on the oxdiation of UDP- [6¢¢,6¢¢-di- 2 H]- D -glucose. In the case of both E141Q and E145Q, the k cat isotope effects were found to be 1.4 ± 0.1. In the case of Thr118Ala, however, k H /k D was found to be 1.9 ± 0.1, consistent with the presence of a primary kinetic isotope effect. This indicates that with the Thr118Ala mutant, one or both of the hydride transfer steps has become the rate-limiting step of catalysis. Discussion The isotope incorporation studies described in this manu- script confirm that the S. pyogenes UDP-glucose dehydro- genase reaction proceeds with the incorporation of a single solvent derived oxygen atom into the product carboxylate. This is consistent with previous results obtained using the bovine liver enzyme [26]. This observation does not disprove the existence of an imine intermediate during catalysis, since the original C6¢¢ oxygen atom from UDP-glucose could conceivably be sequestered within the active site and redelivered into the product carboxylate during the final hydrolysis step. Nevertheless, the absence of any di-labeled product, combined with a lack of any chemical rationale as to a how imine formation could facilitate catalysis, argues against the formation of such an intermediate. Instead, it is likely that once the aldehyde is formed, the thiol of Cys260 readily adds to generate the covalently bound thiohemi- acetal intermediate (Fig. 1). In the previous studies that used the thiocyanide-modified bovine enzyme and NaBH 4 trapping, the isolated adduct was probably formed as a result of the unnatural modification to the active site thiol that prevents the aldehyde from proceeding forward in catalysis [18]. Thus, the imine formation observed with the modified enzyme probably does not reflect a step that occurs in the normal reaction pathway. The absence of any primary kinetic isotope effect during the oxidation of UDP- D -[6¢¢,6¢¢-di- 2 H]-glucose indicates that neither of the hydride transfer steps are rate determining in the reaction of the wild-type enzyme. Instead, another chemical step such as the hydrolysis of the thioester intermediate may be rate determining. This was clearly the case with the Cys260Ser mutant in which the ester intermediate accumulated and could be isolated [20]. Kinetic studies have also led to the suggestion that an irreversible thioesterhydrolysisstepisratedetermininginthecaseofthe beef liver enzyme [11]. Studies with the Cys260Ala mutant showed that this mutant is still capable of oxidizing the aldehyde intermedi- ate at a reasonable rate [20]. It is likely that the mutant is binding the hydrated aldehyde from solution and oxidizing it directly to the acid without using covalent catalysis (Fig. 2). Similar observations have been made with the phosphorylating glyceraldehyde 3-phosphate dehydro- genase (GAPDH) from E. coli [29]. This enzyme normally oxidizes an aldehyde using a thiol-based covalent catalysis strategy similar to the second step of the UDP-glucose dehydrogenase reaction. The resulting thioester intermedi- ate is attacked by phosphate to generate the acyl phosphate product, 1,3-diphosphoglycerate. When the active site cysteine was converted to an alanine, however, the resulting mutant catalyzed the formation of 3-phosphoglycerate as the sole product. It was suggested that the mutant had oxidized the hydrated form of the aldehyde directly to the acid, and was thereby converted into a nonphosphorylating GAPDH. Other enzymes, such as histidinol dehydrogenase [30,31] and alcohol dehydrogenase [32,33], are also known to be able to oxidize hydrated aldehydes directly to acids without using covalent catalysis. The observation that the Cys260Ala mutant of UDP- glucose dehydrogenase can efficiently catalyze the reverse of the first oxidation step, namely reduction of the aldehyde intermediate, indicates that the Cys260 is not required for this step. Instead, the inability of this mutant to catalyze the overall oxidation of the alcohol to the acid must be due to the reasonably high energy of the aldehyde intermediate. In the reaction of the wild-type enzyme, this intermediate is stabilized by binding interactions with active site residues and is readily converted to the thiohemiacetal by attack of Cys260. In the case of Cys260Ala, however, the only way for catalysis to proceed is for the intermediate to be released into solution and then rebound as the hydrate. Apparently release of the aldehyde intermediate is very slow, and hydration of the bound aldehyde does not readily occur, hence the overall rate of catalysis is also very slow. This phenomenon is readily apparent when the full time course of the aldehyde reduction is observed. With the wild-type enzyme, an initial reduction of the aldehyde is observed, followed by oxidation of the aldehyde/alcohol as NAD + accumulates. The net result is a dismutation in which the aldehyde is converted into equimolar amounts of alcohol and acid, with no net consumption of NADH. Similar dismutation processes have been observed with alcohol dehydrogenases that show aldehyde dehydrogenase activit- ies [32,33]. In the case of Cys260Ala, however, the initial reduction phase generates a great deal of UDP-glucose that is kinetically trapped and cannot be reoxidized at any appreciable rate. Only 20% of the aldehyde underwent dismutation in this case. The final experiment with Cys260Ala was a test for NADH exchange in the ternary dehydrogenase/aldehyde/NADH complex. The absence of deuterium washout from UDP-[6¢¢,6¢¢-di- 2 H]- D -glucose dur- ing an incubation with the mutant, NAD + , and excess NADH shows that the NAD 2 H formed in the initial oxidation step does not exchange with free NADH. Its only fate, therefore, is a back reaction to produce the initial starting materials. A similar ternary complex is also formed in the case of the wild-type reaction; however, the thiol group of Cys260 can readily add to the aldehyde carbonyl group. The first formed NADH may then exchange with NAD + and the second oxidation step may proceed. The results obtained with the mutant enzyme are consistent with the idea that cofactor exchange only takes place after the nucleophilic thiol adds to the carbonyl. This helps to explain 20 X. Ge et al. (Eur. J. Biochem. 271) Ó FEBS 2003 how the enzyme efficiently sequesters the aldehyde inter- mediate during catalysis. Our attempts at investigating the roles of other key active site residues via mutagenesis studies were hampered by an inability to isolate several of the mutants (Lys204Ala, Asn208Ala, and Asp264Asn) in a soluble form. Of the three mutants that were amenable to purification (Thr118Ala, Glu141Gln, and Glu145Gln), only Thr118Ala showed a substantial reduction in catalytic efficiency. This mutant also showed a primary kinetic isotope effect upon the oxidation of UDP-[6¢¢,6¢¢-di- 2 H]- D -glucose, indicating that a hydride transfer step was rate limiting for catalysis. Con- sidering the two scenarios for residues involved in promo- ting the hydride transfer steps [21], these observations are most consistent with the one in which the ordered water molecule serves as the catalytic base with the assistance of Asp264 as the ultimate proton acceptor (Fig. 3B). The water molecule is within hydrogen bonding distance of Thr118 and removing such an interaction would certainly perturb its environment. If the water molecule does play a key role in the hydride transfer steps, then mutations to Thr118 may increase the barrier to these steps and render them rate determining. Further investigations will be necessary to fully outline the roles of the active site residues in both the hydride transfer and thioester hydrolysis steps of this interesting enzymatic transformation. Acknowledgements This research was supported by NSERC (operating grant to M.E.T) and the NIH (Public Health Service Grant AI37320 to I.v.d.R). References 1. Oppenheimer, N.J. & Handlon, A.L. (1992) Mechanism of NAD- dependent enzymes. Enzymes 20, 453–504. 2. Rode ´ n, L. (1980) In The Biochemistry of Glycoproteins and Pro- teoglycans (Lennarz, W.J. & , eds), pp. 267–371. Plenum Pub- lishing Co, New York, NY. 3. Dutton, G.J. (1980) Glucuronidation of Drugs and Other Com- pounds. CRC Press Inc., Boca Raton, FL. 4. Dalessandro, G. & Northcote, D.H. (1977) Changes in enzymic activities of nucleoside diphosphate sugar interconversions during differentiation of cambium to xylem in sycamore and poplar. Biochem. J. 162, 267–279. 5. Dougherty, B.A. & van de Rijn, I. (1993) Molecular character- ization of hasB from an operon required for hyaluronic acid synthesis in group A streptococci. J. Biol. Chem. 268, 7118– 7124. 6. Arrecubieta, C., Lo ´ pez, R. & Garcı ´ a, E. (1994) Molecular char- acterization of cap3A, a gene from the operon required for the synthesis of the capsule of Streptococcus pneumoniae type 3: Sequencing of mutations responsible for the unencapsulated phenotype and localization of the capsular cluster on the pneu- mococcal chromosome. J. Bacteriol. 176, 6375–6383. 7. Feingold, D.S. & Franzen, J.S. (1981) Pyridine nucleotide-linked four electron- transfer dehydrogenases. Trends Biochem. Sci. 6, 103–105. 8. Kawamura, T., Ishimoto, N. & Ito, E. (1979) Enzymatic synthesis of uridine diphosphate N-acetyl- D -mannosaminuronic acid. J. Biol. Chem. 254, 8457–8465. 9. Snook, C.F., Tipton, P.A. & Beamer, L.J. (2003) Crystal structure of GDP-mannose dehydrogenase: a key enzyme of alginate bio- synthesis in P. aeruginosa. Biochemistry 42, 4658–4668. 10. Naught,L.E.,Gilbert,S.,Imhoff,R.,Snook,C.,Beamer,L.& Tipton, P. (2002) Allosterism and cooperativity in Pseudomonas aeruginosa GDP-mannose dehydrogenase. Biochemistry 41, 9637– 9645. 11. Ordman, A.B. & Kirkwood, S. (1977) UDP-glucose dehydro- genase. Kinetics and their mechanistic implications. Biochim. Biophys. Acta 481, 25–32. 12.Campbell,R.E.,Sala,R.F.,vandeRijn,I.&Tanner,M.E. (1997) Properties and kinetic analysis of UDP-glucose dehydro- genase from group A streptococci. J. Biol. Chem. 272, 3416– 3422. 13. Ridley, W.P. & Kirkwood, S. (1973) The stereospecificity of hydrogen abstraction by uridine diphosphoglucose dehydro- genase. Biochem. Biophys. Res. Commun. 54, 955–960. 14. Campbell, R.E. & Tanner, M.E. (1999) UDP-Glucose analogues as inhibitors and mechanistic probes of UDP-glucose dehydro- genase. J. Org. Chem. 64, 9487–9492. 15. Strominger, J.L., Maxwell, E.S., Axelrod, J. & Kalckar, H.M. (1957) Enzymatic formation of uridine diphosphoglucuronic acid. J. Biol. Chem. 224, 79–90. 16. Nelsestuen, G.L. & Kirkwood, S. (1971) The mechanism of action of uridine diphosphoglucose dehydrogenase. J. Biol. Chem. 246, 3828–3834. 17. Campbell, R.E. & Tanner, M.E. (1997) Uridine diphospho-a- D -gluco-hexodialdose: Synthesis and kinetic competence in the reaction catalyzed by UDP-glucose dehydrogenase. Angew. Chem. Int., English edition 36, 1520–1522. 18. Ordman, A.B. & Kirkwood, S. (1977) Mechanism of action of uridine diphosphoglucose dehydrogenase. J. Biol. Chem. 252, 1320–1326. 19. Ridley, W.P., Houchins, J.P. & Kirkwood, S. (1975) Mechanism of action of uridine diphosphoglucose dehydrogenase. J. Biol. Chem. 250, 8761–8767. 20. Ge, X., Campbell, R.E., van de Rijn, I. & Tanner, M.E. (1998) Covalent adduct formation with a mutated enzyme: Evidence for a thioester intermediate in the reaction catalyzed by UDP-glucose dehydrogenase. J. Am. Chem. Soc. 120, 6613–6614. 21. Campbell, R.E., Mosimann, S.C., van de Rijn, I., Tanner, M.E. & Strynadka, N.C.J. (2000) The first structure of UDP-glucose dehydrogenase reveals the catalytic residues necessary for the two- fold oxidation. Biochemistry 39, 7012–7023. 22. Hughes, M.J. & A.D. (1996) Creation of deletion, insertion and substitution mutations using a single pair of primers and PCR. Biotechniques 20, 192–196. 23. Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seid- man, J.G., Smith, J.A. & Struhl, K. (1987) Current Protocols in Molecular Biology. Wiley, New York. 24. Laemmli, U.K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680– 685. 25. Snetkova,E.V.,Akulov,G.P.,Gordeeva,L.S.&Kaminskii,Y.L. (1987) Synthesis of UDP-[1- 3 H]glucose and UDP-[6- 3 H]glucose. Khimiya Prirodnykh Soedinenii 1, 125–128. 26. Schiller, J.G., Bowser, A.M. & Feingold, D.S. (1972) Studies on the mechanism of action of UDP- D -glucose dehydrogenase from beef liver. Carbohydr. Res. 25, 403–410. 27. Jones, J.B. & Taylor, K.E. (1976) Nicotinamide coenzyme regeneration. Flavin mononucleotide (riboflavin phosphate) as an efficient, economical, and enzyme compatible recycling agent. Can. J. Chem. 54, 2969–2973. 28. Grubmeyer, C.T. & Insinga, S. (1990) Histidinol dehydrogenase: 18 O isotope shift in 13 C NMR reveals the origin of histidine oxy- gens. J. Am. Chem. Soc. 112, 5906–5908. 29. Corbier, C., Della Seta, F. & Branlant, G. (1992) A new chemical mechanism catalyzed by a mutated aldehyde dehydrogenase. Biochemistry 31, 12532–12535. Ó FEBS 2003 Active site residues of UDP-glucose dehydrogenase (Eur. J. Biochem. 271)21 30. Nagai, A., Kheirolomoom, A. & Ohta, D. (1993) Site-directed mutagenesis shows that the conserved cysteine residues of histi- dinol dehydrogenase are not essential for catalysis. J. Biochem. 114, 856–861. 31. Teng, H., Segura, E. & Grubmeyer, C. (1993) Conserved cysteine residues of histidinol dehydrogenase are not involved in catalysis. Novel chemistry required for enzymatic aldehyde oxidation. J. Biol. Chem. 268, 14182–14188. 32. Henehan, G.T.M., Chang, S.H. & Oppenheimer, N.J. (1995) Aldehyde dehydrogenase activity of Drosophila melanogaster alcohol dehydrogenase: Burst kinetics at high pH and aldehyde dismutase activity at physiological pH. Biochemistry 34, 12294– 12301. 33. Henehan, G.T.M. & Oppenheimer, N.J. (1993) Horse liver alcohol dehydrogenase-catalyzed oxidation of aldehydes: Dismutation precedes net production of reduced nicotiamide adenine dinu- cleotide. Biochemistry 32, 735–738. 22 X. Ge et al. (Eur. J. Biochem. 271) Ó FEBS 2003 . Active site residues and mechanism of UDP-glucose dehydrogenase Xue Ge 1 , Lisa C. Penney 1 , Ivo van de Rijn 2 and Martin E. Tanner 1 1 Department of. mechanism of the reaction cata- lyzed by UDP-glucose dehydrogenase. Ó FEBS 2003 Active site residues of UDP-glucose dehydrogenase (Eur. J. Biochem. 271)15 TAAGTAGCGGCAAATAG-3¢;