The Reductive Half-reaction of Xanthine Dehydrogenase from Rhodobacter capsulatus

Background: Kinetic characterization of wild-type xanthine dehydrogenase and variants. Results: Comparison of the pH dependence of both kred and kred/Kd, as well as kcat and kcat/Km. Conclusion: Ionized Glu232 of wild-type enzyme plays an important role in catalysis by discriminating against the monoanionic form of xanthine. Significance: Examining the contributions of Glu232 to catalysis is essential for understanding the mechanism of xanthine dehydrogenase. The kinetic properties of an E232Q variant of the xanthine dehydrogenase from Rhodobacter capsulatus have been examined to ascertain whether Glu232 in wild-type enzyme is protonated or unprotonated in the course of catalysis at neutral pH. We find that kred, the limiting rate constant for reduction at high [xanthine], is significantly compromised in the variant, a result that is inconsistent with Glu232 being neutral in the active site of the wild-type enzyme. A comparison of the pH dependence of both kred and kred/Kd from reductive half-reaction experiments between wild-type and enzyme and the E232Q variant suggests that the ionized Glu232 of wild-type enzyme plays an important role in catalysis by discriminating against the monoanionic form of substrate, effectively increasing the pKa of substrate by two pH units and ensuring that at physiological pH the neutral form of substrate predominates in the Michaelis complex. A kinetic isotope study of the wild-type R. capsulatus enzyme indicates that, as previously determined for the bovine and chicken enzymes, product release is principally rate-limiting in catalysis. The disparity in rate constants for the chemical step of the reaction and product release, however, is not as great in the bacterial enzyme as compared with the vertebrate forms. The results indicate that the bacterial and bovine enzymes catalyze the chemical step of the reaction to the same degree and that the faster turnover observed with the bacterial enzyme is due to a faster rate constant for product release than is seen with the vertebrate enzyme.

thine and xanthine to uric acid in the overwhelming majority of extant organisms. Most of these enzymes are dehydrogenases utilizing NAD ϩ as oxidizing substrate, although in mammals the enzyme can be post-translationally modified (by either cysteine oxidation or limited proteolysis) to an oxidase that utilizes O 2 instead (1,2). The x-ray crystal structures of several of these enzymes are known, including the bovine enzyme in both dehydrogenase and oxidase forms (3) and the dehydrogenase from Rhodobacter capsulatus (4). As shown in Fig. 1, the architecture of the active site is highly conserved despite the fact that the vertebrate enzyme is a homodimer with all four redox-active centers (FAD and a pair of [2Fe-2S], in addition to the active site molybdenum center) in a single subunit, whereas the R. capsulatus enzyme is an (␣␤) 2 tetramer with the FAD and [2Fe-2S] centers in one pair of subunits, and the molybdenum center is a second pair of subunits. Substrate binds between two active site phenylalanine residues (Phe 914 and Phe 1009 in the bovine enzyme, Phe 344 and Phe 459 in the R. capsulatus enzyme), which serve to orient substrate for catalysis. A universally conserved glutamate (Glu 1261 in the bovine enzyme and Glu 730 in the bacterial enzyme) is located immediately beneath the molybdenum center and is essential for catalytic activity (5,6). Other conserved residues in the active site include Glu 802 and Arg 880 (Glu 232 and Arg 310 in the bacterial enzyme). Given the virtually identical active site structures, the readily expressed recombinant bacterial enzyme constitutes a convenient system to conduct mutational studies (7).
The mechanism of xanthine hydroxylation is now generally understood to proceed as shown in Fig. 2 (8), beginning with proton abstraction from the equatorial Mo-OH by the highly conserved active site glutamate, followed by nucleophilic attack on the sp 2 -hybridized carbon of substrate that becomes hydroxylated (C2 in the case of hypoxanthine, C8 in the case of xanthine). Concomitant hydride transfer to the MoϭS group results in reduction of the molybdenum center from Mo VI to Mo IV . The intermediate thus formed has product coordinated to the now reduced molybdenum center via the catalytically introduced hydroxyl oxygen and breaks down by electron transfer out of the molybdenum to the other redox-active cen-ters of the enzyme and displacement of product from the molybdenum coordination sphere by hydroxide from solvent. Depending on reaction conditions and substrate, the latter sequence of events can give rise to a Mo(V) species containing product coordinated to the molybdenum center that gives rise to the long-studied "very rapid" EPR signal (9,10), although with most substrates and under most reaction conditions, this is a minor pathway that contributes no more than ϳ10% of the total catalytic throughput. It is notweworthy, however, that formation of the very rapid Mo(V) species is thus an oxidative event, arising from the loss of an electron from a precursor Mo(IV) species that is EPR-silent.
Previous mutagenesis studies of the recombinant R. capsulatus enzyme have involved substitution of Glu 730 and Glu 232 to Ala (6), and Arg 310 to Met (11) (equivalent to Glu 1261 , Glu 802 , and Arg 880 , respectively, in the bovine enzyme). Mutation of Glu 730 to Ala abolishes all activity toward xanthine in rapid reaction kinetic studies at neutral pH, with a minimum reduction in k red (and, correspondingly, the steady-state k cat ) of 10 7 , consistent with the role of this residue as a general base, abstracting the Mo-OH proton to initiate catalysis (6). Mutation of Arg 310 to Met also results in a large reduction in k red , by a factor of 10 4 , an effect attributed to Arg 310/880 stabilizing negative charge on the heterocycle in the course of nucleophilic attack (11). The E232A variant exhibited a more modest 12-fold decrease in k red and a 12-fold increase in K d in the reductive half-reaction, relative to wild-type enzyme, comparable with the effect seen on the steady-state kinetic behavior of an E803V variant of the human enzyme (13). Given the effect of the mutation on both K d and k red , it is evident that Glu 232 is involved with both substrate binding and transition state stabilization, although its specific role has been somewhat controversial. Computational studies have suggested that Glu 232/802 is ionized and plays a role in facilitating proton tautomerization between N3 and N9 of xanthine in the course of catalysis (6, 14 -16). Such a role is consistent with the orientation of xanthine observed in complexes of the inactive, desulfo forms of both the bovine and R. capsulatus enzymes (Fig. 3A), where Glu 802/232 appears to be deprotonated and serving as a hydrogen-bonding acceptor to the protonated N3 of substrate (17,18). On the other hand, in the structure of xanthine oxidase with the slow substrate FYX-051 (19) Glu 802 appears to be protonated, serving as a hydrogen-bonding donor to an aromatic nitrogen of the substrate. Glu 802 also appears to be protonated in crystal structures of the rat enzyme in complex with urate (20), again appearing to serve as a hydrogen-bonding donor, this time to N7 of urate, which is seen in an inverted orientation relative to that observed with xanthine (Fig. 3B). Quantum mechanics/ molecular mechanics studies have also been performed on xanthine oxidase, in which it is presumed that Glu 802 is protonated at the outset of catalysis (21).
The above mutational effects are predicated on the assumption that the chemical step of the reaction is rate-limiting in the variants. It is known, however, that at least with the wild-type bovine enzyme, product release is in fact rate-limiting. To further evaluate the role of Glu 232 in catalysis, kinetic studies on the reductive half-reaction of a nonionizing E232Q variant of the recombinant R. capsulatus enzyme have been conducted. We find that the E232Q variant exhibits the same 12-fold decrease in k red as the E232A variant, a result that is inconsistent with the argument that a neutral (protonated) Glu 802/232 acts as a hydrogen-bonding donor to the substrate during catal-

The Reductive Half-reaction of Xanthine Dehydrogenase
ysis. The pH dependence of the reductive half-reaction has also been determined that, when compared with the behavior of the wild-type enzyme, suggests a role for Glu 232 in discriminating against the monoanionic form of substrate. We have also undertaken a kinetic isotope effect study to evaluate the extent to which the chemical step of the reaction is rate-limiting in the case of the wild-type R. capsulatus enzyme and found that its considerably faster turnover relative to the bovine enzyme is due to significantly faster product release, the chemical step of the reaction for the two enzymes occurring at comparable rates.

EXPERIMENTAL PROCEDURES
Enzyme Purification, Metal, and Cofactor Content Quantification-Recombinant R. capsulatus xanthine dehydrogenase was heterologously expressed in Escherichia coli TP 1000 cells and purified as described previously (7), by nickel-nitrilotriacetic acid chromatography, Q-Sepharose, and size exclusion chromatography. E232Q was mutated to glutamine in the enzyme by PCR mutagenesis, as described previously (18). Metal content analysis was performed using an Optima 2100DV inductively coupled plasma-optical emission (ICP-OES) 3 spectrometer (PerkinElmer Life Sciences). Protein samples were incubated overnight in a 1:1 mixture with 65% nitric acid (Suprapur; Merck) at 100°C. Protein samples were diluted with a 10-fold volume of ultrapure water (Millipore) prior to ICP-OES analysis. The multielement standard solution XVI (Merck) was used as a reference (22).
Enzyme Assay Methods-UV-visible absorption spectra were obtained, and steady-state experiments were performed using a Hewlett-Packard 8452 diode array spectrophotometer. To follow the fractional conversion of xanthine to uric acid by xanthine oxidase, substrate and enzyme were added to 1 ml of air equilibrated buffer (0.1 M sodium pyrophosphate, 0.3 mM EDTA, pH 8.5). The formation of uric acid was monitored at 295 nm (⌬⑀ 295 nm ϭ 9600 M Ϫ1 cm Ϫ1 ). For bacterial XDH1, xanthine, NAD ϩ , and enzyme were subsequently added to 1 ml of air-equilibrated buffer (50 mM potassium phosphate, 0.1 mM EDTA, pH 7.9). The formation of NADH (⌬⑀ 340 nm ϭ 6200 M Ϫ1 cm Ϫ1 ) was monitored in addition to the formation of uric acid.
Rapid Reaction Kinetics-The reductive half-reaction kinetics were performed using an Applied Photophysics, Inc. SX-18MV stopped flow spectrophotometer. Enzyme was made anaerobic in glass tonometers by evacuating and flushing with oxygen scrubbed argon gas (obtained from Airgas) for between 30 and 60 min. Substrate solutions were made anaerobic by bubbling solutions in glass syringes for 15-30 min. Standard  NOVEMBER 14, 2014 • VOLUME 289 • NUMBER 46 buffer conditions for xanthine oxidase were 100 mM sodium pyrophosphate, 0.3 mM EDTA, pH 8.5, at 25°C. Reactions were observed at 450 nm, monitoring the disappearance of absorbance because of reduction. For bacterial xanthine dehydrogenase, standard buffer conditions were 50 mM potassium phosphate, 0.1 mM EDTA, pH 7.8. Reactions involving the bacterial xanthine dehydrogenase were monitored at 465 nm. The reaction mixtures after mixing typically had 5-10 M enzyme and varying concentrations of substrate. The reaction with each substrate concentration used was repeated in triplicate. The kinetic transients were fit to single exponentials to obtain rate constants, which were averaged for each substrate concentration. Each k obs was then plotted against substrate concentration, and the hyperbolic curve fit using SigmaPlot to obtain values for k red and the dissociation constant, K d . pH dependence studies on the reductive half-reaction were conducted between pH units of 6 -11 using conditions previously reported by Kim et al. (23). Buffers at the various pH units included 0. For both the steady-state and rapid reaction data, pH profiles for a given kinetic parameter L were fit to the following equation for a double-ionization mechanism, with the maximum theoretical value for L, L max , and the two pK a values pK 1 and pK 2 as floating variables.

The Reductive Half-reaction of Xanthine Dehydrogenase
Kinetic Isotope Effect Studies-The deuterium isotope effect on the steady-state kinetics of the R. capsulatus xanthine dehydrogenase was analyzed in the conventional manner in solutions. Typical experiments involved addition of catalytic amounts of enzyme to 2.5 ml of buffer solution at pH/D 7.8 containing 50 mM potassium phosphate, 0.1 mM EDTA, and 500 M NAD ϩ as the terminal electron acceptor. The solutions were made anaerobic by bubbling argon gas through it for 20 min in a sealed quartz cuvette. The buffer solution, including NAD ϩ , was preincubated at 25°C followed by the addition of first xanthine and then enzyme. Changes in absorbance change at 295 nm were monitored. The initial reaction mix contained a fixed concentration of functional R. capsulatus xanthine dehydrogenase (ϳ2 nM functional active sites) and varying concentrations of substrates (5-200 M). The initial rates were calculated from linear fits of the first 20 -120 s of each transient. Xanthine oxidase assays used 5 nM functionally active sites and were measured in air equilibrated 0.1 M sodium pyrophosphate, pH 8.5, that did not contain NAD ϩ .
Competition experiments were used to determine the tritium isotope effect on (V/K) were performed following a slight modification of the method described by D'Ardenne and Edmondson (24). Catalytic amounts of enzyme (5 nM functional active sites for xanthine oxidase and 2 nM functional active sites for bacterial XDH) were incubated at room temperature with xanthine concentrations of 40 M for bovine xanthine oxidase and 200 M xanthine for bacterial XDH. Air-saturated buffers were used for xanthine oxidase, whereas anaerobic solutions of 500 M NAD ϩ were used for the bacterial XDH samples. Levels of radiolabeled [8-3 H]xanthine in each sample were ϳ0.1 Ci/ ml. Fractional conversion of xanthine to uric acid was determined via UV-visible spectroscopy, monitoring the reaction to completion and determining the point at which 10 -20% of the total absorbance change at 295 nm was reached. Once 10 -20% fractional conversion was reached, the reaction mixture was quenched by the addition of 10% (w/v) trichloroacetic acid. Substrate ([8-3 H]xanthine) and product ( 3 H-labeled water) were separated by HPLC (Supelco C18 column, isocratic conditions, 1 ml/min, 50 mM sodium phosphate, pH 6.0) with tritiated water eluting at ϳ3 min and [8-3 H]xanthine eluting at ϳ10 min.
Competitive (V/K) isotope effects were calculated from quantitation of the radioactivity in the tritiated water and the [8-3 H]xanthine using the following equation (24,25), where f is the fractional conversion of substrate to products, R p t is the 3 H ratio of products at time t, and R s 0 is the 3 H ratio of the substrate at t ϭ 0. The ratio at 0% conversion was monitored to correct for the naturally occurring 3 H exchange with water in the absence of enzyme.
Knowing H/D (V/K) and H/T (V/K), the intrinsic isotope effect was calculated using Equation 3 (26), with both the primary deuterium isotope effect ( H/D V) and the intrinsic isotope effect ( D k) thus known, the f v or fractional velocity of the chemical step in relation to the overall rate of the reaction was calculated using Equation 4 (24), where f v represents the fractional extent to which the chemical step of the reaction (k C-H ) contributes to the overall k cat ; 1/f v thus gives how much faster the chemical step of the reaction is than k cat , thus multiplying the inverse of f v by the value for k cat gives the approximate rate of the chemical step. The estimated values for the rate of product dissociation (k off ) was calculated using Equation 5.

RESULTS AND DISCUSSION
Steady-state Kinetics of the E232Q Variant of R. capsulatus Xanthine Dehydrogenase-Structural studies with both the bovine and R. capsulatus enzymes indicate that Glu 802 (equivalent to Glu 232 in the R. capsulatus enzyme) is within hydrogenbonding distance of substrate and could act as either a hydrogen bonding acceptor (in its ionized state) to substrate N3 or hydrogen-bonding donor (in the neutral form, although this would require a pK a above 8, well outside the normal range for glutamates) to the C 6 ϭO carbonyl (Fig. 3) (17, 18). Past studies by Leimkühler et al. have demonstrated a 12-fold decrease of E232A variant in k red compared with the wild-type enzyme (6).
Although the alanine variant assesses the effect of removing the carboxylate side chain of glutamate, it does not provide specific insight its ionization state. We have thus prepared an E232Q variant of the R. capsulatus xanthine dehydrogenase variant and examined its kinetic behavior in detail. If, as has been suggested (19), Glu 232 is neutral and acts as a hydrogen-bonding donor during catalysis, then the E232Q variant should still be able to hydrogen bond with the substrate, and there should be little change in activity. If, on the other hand, Glu 232 is ionized, a decrease in activity with the E232Q variant is expected, at least as much as the 12-fold decrease in k red seen previously with the E232A mutant (6).
The pH dependence the steady-state values k cat and k cat /K m for wild-type xanthine dehydrogenase and the E232Q variant are shown in Fig. 4. For both enzyme forms, each kinetic parameter exhibits bell-shaped pH dependence. For wild-type enzyme, the pK a values that define the pH profile for k cat are 7.4 and 9.6; for the E232Q variant, these values decrease to 6.5 and 9.2. The maximum of k cat for the wild-type enzyme is ϳ150 s Ϫ1 , which decreases with the E232Q variant to 3.3 s Ϫ1 (Table 1). For k cat /K m , the pK a values for wild-type enzyme are 6.8 and 8.8, which are lowered only very modestly in the E232Q variant to 6.6 and 8.5. The pK a values for k cat /K m with wild-type enzyme determined here are in fair agreement with the values of 6.3 and 7.7 reported previously by Aguey-Zinsou et al. (27), who monitored catalysis electrochemically (the pH dependence of k cat was not reported in this work). However, here the autocatalytic reaction with uric acid was measured directly at the electrode, which might explain the difference in the pH profile, because the latter is rather related to the ionization of uric acid (28). The maximal theoretical value for k cat /K m , obtained from the fit to the pH profile as described under "Experimental Procedures," decreases from 2.0 ϫ 10 6 M Ϫ1 s Ϫ1 with the wild-type enzyme to 6 ϫ 10 4 M Ϫ1 s Ϫ1 for the E232Q variant.

The Reductive Half-reaction of the E232Q Variant with
Xanthine-We next examined the reductive half-reaction kinetics of the E232Q variant with xanthine, with the results summarized in Tables 2 and 3. As shown in Fig. 5A, at pH 7.8, 25°C, the kinetics of reduction of both wild-type xanthine dehydrogenase and the E232Q variant by xanthine exhibit hyperbolic dependence on xanthine concentration. The limiting rate of reduction, k red , for the wild-type enzyme with xanthine was found to be 217 Ϯ 13 s Ϫ1 , and the apparent dissociation constant, K d , was found to be 42.1 Ϯ 10.0 M, in good agreement with previously published results (k red ϭ 67.3 s Ϫ1 , K d ϭ 33.6 M) (7) after taking into account the lower temperature in the earlier work, 4°C. The E232Q variant exhibits a 12-fold decrease in k red and a ϳ3-fold increase in K d compared with wild type. These values can be compared with the published results seen with the E232A variant, which exhibited a 12-fold decrease in k red and 12-fold increase in K d compared with wild type. It is evident that both variants compromise k red to the same degree, although the effect on K d is somewhat more modest in the E232Q variant than in the E232A variant. Nevertheless, the comparable effects of the two mutations on k red are consistent with the interpretation that Glu 802/232 is in fact deprotonated in the course of catalysis (6, 14 -16). Both the magnitude of k red seen with the E232Q variant and its pH dependence is consistent with the behavior of the steady-state k cat , as expected given that the overall rate-limiting step in catalysis is known to reside in the reductive half-reaction (24).
The pH Dependence of Wild-type R. capsulatus Xanthine Dehydrogenase and the E232Q Variant-Previous pH dependence studies of bovine xanthine oxidase with xanthine as substrate (23,29) have yielded a sigmoidal curve (ascending at higher pH) for the rapid reaction kinetic parameter k red (which tracks the breakdown of the E⅐S complex) with a pK a of 7.5. This is the sole ionization that is clearly defined between pH 6 and 10, indicating that no other active site residue ionizes in a catalytically critical way in this pH range. (Although there is evidence of decrease at very high pH, it is not sufficiently well defined as to assign a specific pK a .) For k red /K d (which tracks the reaction of free enzyme with free substrate in the low [substrate] regime and which will respond to ionizations of free enzyme as well as free substrate), on the other hand, a bellshaped curve is clearly observed. Fits using Equation 6, yield pK a values of 6.6 and 7.4 for the ascending and descending limbs of the plot, respectively. The latter ionization agrees well with the known pK a for xanthine of 7.6 (31), and the former has been assigned to Glu 1261 acting as an active site base (32). The implication is that enzyme binds to and acts on the neutral rather than monoanionic form of the substrate (consistent with a reaction mechanism predicated on nucleophilic attack on substrate).

The Reductive Half-reaction of Xanthine Dehydrogenase
In examining the pH dependence of the reductive half-reaction of the bacterial enzyme (summarized in Table 2), we find that k red /K d for the wild-type enzyme is also bell-shaped, consistent with the rate-limiting step being in the reductive halfreaction of the overall catalytic sequence (as is the case with the bovine enzyme). The pH dependence on k red /K d also exhibits bell-shaped curve with pK a values of ϳ7 and 8.5 (Fig. 5B). By contrast to the bovine enzyme, however, we find that the pH dependence of k red with the wild-type bacterial enzyme is also clearly bell-shaped curve, yielding pK a values of 7.3 and 10 for the ascending and descending limbs, respectively, and a theoretical maximum value for k red of ϳ270 s Ϫ1 (Fig. 5C). The somewhat different pH dependence of k red seen with the bovine and bacterial enzymes, with a clear descending limb in the pH profile at high pH only observed with the bacterial enzyme, suggests that although both enzymes preferentially bind the neutral form of substrate-that is, shift its pK a to a higher value-the bovine enzyme does so to a somewhat greater degree than the bacterial form. Still, the shift in pK a for xanthine on binding the bacterial enzyme is increased 2.5 pH units over that in free solution (as compared to the bovine enzyme, where the shift is at least 3.5 pH units).
With the E232Q variant, the pH dependence on k red /K d also exhibits bell-shaped curve with pK a values of 6.1 and 8.6 (Fig.  5C), generally consistent with the results seen with wild-type enzyme taking into account the significantly lower maximal value seen with the variant (2.0 ϫ 10 5 M Ϫ1 s Ϫ1 as compared with ϳ4 ϫ 10 6 M Ϫ1 s Ϫ1 ). As shown in Fig. 5C, the pH profile for k red is also bell-shaped (as seen with the wild-type bacterial enzyme), but the maximal value is 21 s Ϫ1 , significantly lower than the value of 270 s Ϫ1 seen with the wild-type enzyme. The pK a values for the pH dependence of k red with the E232Q variant are 6.3 and 8.6, comparable with the values seen in the pH dependence of k red /K d with the E232Q variant (Fig. 5C). The significant reduction in the maximal value for k red upon substitution of glutamine for glutamate, comparable with that seen previously with the E232A variant (6), clearly indicates that negative charge and not hydrogen bond donating capacity is the catalytically essential feature of this residue. Our interpretation of the difference in pK a values for k red with wild-type and E232Q variant is that the variant does not shift the pK a of substrate to nearly the same extent as does wild-type enzyme, with the result that the monoanion of xanthine is able to bind in the active site. Deprotonation of xanthine typically occurs at N3 (15,31), so based upon the orientation of xanthine in the crystal structures of the wild-type enzymes shown in Fig. 3 (17), the negative charge of Glu 802/232 would repel xanthine exhibiting a negative charge delocalized over N3 and the C2 oxygen (Fig. 6,  A and B). Substitution of Glu 232 to Gln abolishes this electrostatic repulsion (Fig. 6, C and D), and the neutral but polar glutamine would also be able to interact favorably with both the neutral and monoanionic forms of substrate (although only the former would be catalytically effective, because of the significantly lower susceptibility of the negatively charged monoanion to nucleophilic attack).
We note that substrate orientation in the active site is likely to depend critically on its ionization state. The negative charge at N3 of the monoanion could well be stabilized by the posi- for the anaerobic reductive half-reaction of wild-type (blue) and E232Q (red) enzyme. All reactions were carried out under anaerobic conditions in 50 mM potassium phosphate, 0.1 mM EDTA, pH 7.8, at 25°C, following reduction of enzyme by excess xanthine was followed at 465 nm. Hyperbolic fits to the data (solid lines) yielded k red ϭ 217 s Ϫ1 and K d 432 M for wild-type enzyme and 17 s Ϫ1 and 124 M for the E232Q variant. B, the pH dependence of k red /K d for wild-type xanthine dehydroganase (black circles) and the E232Q variant (white circles) with xanthine. The fit for the pH dependence of k red /K d for the wild-type enzyme gave pK a values of ϳ7 and ϳ8.5 and a theoretical maximum value of ϳ4 ϫ 10 6 M Ϫ1 s Ϫ1 . The fit for the pH dependence of k red /K d for XDH E232Q variant gave pK a values of 6.1 and 8.5 and a theoretical maximum value of 0.20 ϫ 10 6 M Ϫ1 s Ϫ1 . It is to be noted that because of the two pK a values for k red /K d for wild-type bacterial XDH being close to each other, the theoretical maximum is significantly underestimated. C, the pH dependence of k red for wild-type xanthine dehydrogenase (black circles) and the E232Q variant (white circles) with xanthine. The fit for the pH dependence of k red for wild-type bacterial XDH gave pK a values of ϳ7.3 and ϳ10 and a theoretical maximum value of ϳ270 s Ϫ1 . The fit for the pH dependence of k red for XDH E232Q variant gave pK a values of 6.18 and 8.59 and a theoretical maximum value of 21.2 s Ϫ1 .
tively charged Arg 310 , resulting in inversion of substrate in the active site that would allow Gln 232 in the variant bacterial enzyme could hydrogen bond to the C6 carbonyl group in either of two ways (Fig. 6, E and F). In the wild-type enzyme, the negative charge on Glu 232 driving substrate protonation, this binding mode would be unlikely under physiological conditions. We reiterate, however, that an inverted orientation has been observed crystallographically for reduced bovine xanthine oxidoreductase in complex with urate (20), as shown in Fig. 3B and corresponding to the orientation shown in Fig. 6E. At neutral pH, urate (which has a first pK a of 5.8 (33)) is predominantly monoanionic, with the negative charge largely localized on the C 2 oxygen. We suggest that the negative charge on urate at neutral pH gives rise to an inverted orientation in the active site relative to the catalytically productive one for xanthine. This being the case, urate must be considered a poor mimic of substrate in attempting to understand the orientation of the latter in the active site at the outset of catalysis.
In summary, replacing Glu 232 of R. capsulatus xanthine dehydrogenase with glutamine shifts the observed pH profile of k red , consistent with the negatively charged glutamate increasing the pK a of enzyme-bound substrate by at least two pH units relative to its value in free solution. Further, given the significant (12-fold) decrease in k red associated with the mutation, it is clear that Glu 232 is in fact ionized and negatively charged in wild-type enzyme. Our results suggest that the ionized Glu 232 plays an essential role in substrate specificity, preventing binding of the negatively charged monoanion by electrostatic repulsion. The decrease in k red above pH 7.4 seen here with the E232Q variant observed is consistent the xanthine monoanion now being able to bind in the active site, although it is not acted upon by the enzyme.
Kinetic Isotope Effect Studies of R. capsulatus Xanthine Dehydrogenase-The reductive half-reaction of xanthine consists of two major steps: 1) the initial chemical step of the reaction (nucleophilic attack and hydride transfer to the molybdenum center) that yields product coordianted to the molybdenum center via the catalytically introduced hydroxyl group and 2) product release by displacement of product from the molybdenum center by hydroxide, accompanied by electron transfer out of the molybdenum center. In the wild-type bovine and chicken enzymes, the rate-limiting step has been shown to be product release (12), which is some 75-fold slower than the chemical step of the reaction (24). In light of the significantly larger value k cat and k red seen with wild-type R. capsulatus xanthine dehydrogenase relative to the bovine enzyme (30), we have determined the extent to which the chemical step of the reaction is rate-limiting in the bacterial enzyme using a kinetic isotope effect protocol.
Using xanthine deuterated at position 8 as substrate, the apparent primary kinetic isotopic effect on k cat and k cat /K m ( D k cat and D (V/K), respectively) are found to be very small for the bacterial enzyme: 1.39 and 2.12, respectively (Fig. 7). These values are comparable with those seen previously with the bovine and chicken enzymes (24,29). To determine the intrinsic isotope effect for the chemical step of the reaction, the tritium isotope effect on (V/K) must also be determined, which allows for the subsequent calculation of the extent that the chemical step is rate-limiting. Using a modification of the method of D'Ardenne and Edmondson (24) as detailed under "Experimental Procedures," tritium isotope effect values were determined for the bacterial xanthine dehydrogenase and, as a control, the bovine oxidase. The results are shown in Table 4, where it can be seen that the tritium isotope effect on (V/K) for the bovine and bacterial enzymes were comparable, at 3.62 and 3.64, respectively. The intrinsic isotope effects calculated as described under "Experimental Procedures" for the bovine and bacterial enzymes are 7.01 and 5.44, respectively, the former value being in acceptable agreement with the previously published value of 7.4 (24). After calculating the fractional velocity (f v ) of the chemical step for each of the enzymes (see "Experimental Procedures"), k C-H is found to be some ϳ62-fold faster than k cat for the bovine enzyme and ϳ12-fold faster for the bacterial enzyme. With the k cat values for each enzyme (Table  3), the rate constant for the chemical step of the reaction for each of the enzymes could be estimated, yielding values of 930 and 766 s Ϫ1 for the bovine and bacterial enzymes, respectively, which are remarkably similar to one another. Knowing k cat and k C-H allows calculation of the rate constant for product dissociation, k off , using the following equation.
The values thus obtained are 15.2 and 76.9 s Ϫ1 for the bovine and bacterial enzymes, respectively. These results demonstrate that the increased rates seen with the bacterial enzyme are due to significantly faster product dissociation rather than acceleration of the chemical step of the reaction. Unfortunately, the tritium kinetic isotope effects for the E232Q variant (or others, including E730A, E232A, and R310M) could not be reliably determined because of the slow rate of reaction. Even when rates were fast enough to be observed experimentally, consumption of product NADH began to occur (resulting in loss in absorbance at 340 nm) along with a decrease in uric acid (a decrease in absorbance at 295 nm). This does not occur to any significant degree with the wild-type enzyme and could be attributed to reversal of the reaction of bacterial XDH with uric acid and NADH. Given the decrease in the limiting rate of reduction of the variants as compared the wild-type enzyme, the E232A and R310M XDH variants would be expected to exhibit a smaller difference between the rates of the forward and backward reactions, allowing for the reverse reaction to occur much more quickly than the forward reaction after the initial accumulation of uric acid and NADH. The Reductive Half-reaction of Xanthine Dehydrogenase