Evidence for proton tunneling and a transient covalent flavin-substrate adduct in choline oxidase S101A

BBA - Proteins and Proteomics 1865 (2017) 1470–1478

Contents lists available at ScienceDirect

BBA - Proteins and Proteomics journal homepage: www.elsevier.com/locate/bbapap

Evidence for proton tunneling and a transient covalent flavin-substrate adduct in choline oxidase S101A

MARK

Rizvan Uluisika, Elvira Romeroa,1, Giovanni Gaddaa,b,c,d,⁎ a

Department of Chemistry, Georgia State University, Atlanta, GA 30302-3965, USA Department of Biology, Georgia State University, Atlanta, GA 30302-3965, USA c Department of Center for Biotechnology and Drug Design, Georgia State University, Atlanta, GA 30302-3965, USA d Department of Center for Diagnostics and Therapeutics, Georgia State University, Atlanta, GA 30302-3965, USA b

A R T I C L E I N F O

A B S T R A C T

Keywords: Choline oxidase Flavin-substrate adduct Kinetic isotope effect Proton transfer Quantum mechanical tunneling Stopped-flow

The effect of temperature on the reaction of alcohol oxidation catalyzed by choline oxidase was investigated with the S101A variant of choline oxidase. Anaerobic enzyme reduction in a stopped-flow spectrophotometer was biphasic using either choline or 1,2-[2H4]-choline as a substrate. The limiting rate constants klim1 and klim2 at saturating substrate were well separated (klim1/klim2 > 9), and were >15-fold slower than for wild-type choline oxidase. Solvent deuterium kinetic isotope effects (KIEs) ~4 established that klim1 probes the proton transfer from the substrate hydroxyl to a catalytic base. Primary substrate deuterium KIEs ≥7 demonstrated that klim2 reports on hydride transfer from the choline alkoxide to the flavin. Between 15 °C and 39 °C the klim1 and klim2 values increased with increasing temperature, allowing for the analyses of H+ and H– transfers using Eyring and Arrhenius formalisms. Temperature-independent KIE on the klim1 value (H2Oklim1/D2Oklim1) suggests that proton transfer occurs within a highly reorganized tunneling-ready-state with a narrow distribution of donor-acceptor distances. Eyring analysis of the klim2 value gave lines with the slope(choline) > slope(D-choline), suggesting kinetic complexity. Spectral evidence for the transient occurrence of a covalent flavin-substrate adduct during the first phase of the anaerobic reaction of S101A CHO with choline is presented, supporting the notion that an important role of amino acid residues in the active site of flavin-dependent enzymes is to eliminate alternative reactions of the versatile enzyme-bound flavin for the reaction that needs to be catalyzed.

1. Introduction Understanding how enzymes acquire the catalytic power to accelerate biochemical reactions by factors as high as 1020 while maintaining exquisite selectivity and unprecedented accuracy is one of the current frontiers of science [1,2]. The transfers of protons (H+) and hydride ions (H−) are integral parts of a vast number of enzyme-catalyzed reactions and have been investigated to address the contribution of quantum mechanical (QM) tunneling to enzymatic catalysis [3–7]. Examples include soybean lipoxygenase-1 [8,9], alcohol dehydrogenase [10–12], dihydrofolate reductase [13,14], morphinone reductase [15,16], glucose oxidase [17–19], choline oxidase (CHO) [20–23], and glycolate oxidase [24]. As a diagnostic tool to investigate tunneling and conformational sampling in enzymes catalyzing H+ or H– transfers the effect of temperature on the reaction rate constants and their associated

deuterium kinetic isotope effects (KIEs) are determined [11,23,25–33]. The resulting data can be interpreted using an extension of the Transition State Theory (TST) as previously described [34]. In the framework of this model the enzyme-substrate complex samples different conformations to bring the donor and acceptor closer to each other. Temperature-dependent rates and temperature-independent KIEs are observed when conformational sampling results in a narrow distribution of donor-acceptor distances (DADs) at the tunneling-ready-state (TRS). In contrast, temperature-dependent KIEs are observed in enzymes presenting a poorly reorganized TRS with a wide range of DADs at thermal equilibrium, because the lighter isotope can tunnel over longer distances than the heavier isotope and thermally activated DAD fluctuations populate shorter DADs at high temperatures. In the present work, we used this model for investigating H+ and H– transfers in the S101A mutant of CHO (E.C. 1.1.3.17, choline‑oxygen 1-

Abbreviations: CHO, choline oxidase; DAD, donor-acceptor distance; D-choline, 1,2-[2H4]-choline; FAD, flavin adenine dinucleotide; KIE, kinetic isotope effect; QM, quantum mechanical; TRS, tunneling-ready-state; TST, transition state theory ⁎ Corresponding author at: Department of Chemistry, Georgia State University, P.O. Box 3965, Atlanta, GA 30302-3965, USA. E-mail address: [email protected] (G. Gadda). 1 Present Address: Molecular Enzymology Group, University of Groningen, Nijenborgh 4, 9747AG Groningen, The Netherlands. http://dx.doi.org/10.1016/j.bbapap.2017.08.004 Received 22 March 2017; Received in revised form 8 August 2017; Accepted 10 August 2017 Available online 24 August 2017 1570-9639/ © 2017 Elsevier B.V. All rights reserved.

BBA - Proteins and Proteomics 1865 (2017) 1470–1478

R. Uluisik et al.

Scheme 1. Four-electron oxidation of choline catalyzed by CHO.

oxidoreductase). CHO has been extensively investigated mechanistically, biochemically and structurally [20,21,35–46]. The enzyme catalyzes the fourelectron oxidation of choline to glycine betaine with dioxygen as electron acceptor [47–49]. The reaction occurs through two subsequent FAD-mediated oxidations of the substrate [48,50,51]. The alcohol substrate is oxidized to betaine aldehyde in the first oxidation reaction; then, the hydrated form of betaine aldehyde, i.e., gem-diol choline, is oxidized to glycine betaine (Scheme 1). The aldehyde intermediate stays bound in the active site of CHO from bacteria, as established by a colorimetric assay and kinetic studies [50,51]. In CHO from fungi, instead, it is released to the bulk solvent [52]. Despite the potential complexity of the steady-state kinetic mechanism of bacterial CHO, the two oxidation reactions are the sole limiters of the overall turnover of the enzyme [50]. CHO from Arthrobacter globiformis has thus emerged as a model system to investigate the reaction of alcohol oxidation catalyzed by flavoprotein oxidases. This enzyme is also of high interest in biotechnology because glycine betaine is a powerful and widespread osmoprotectant in many plants [53]. The abstraction of the hydroxyl H+ of choline to form choline alkoxide in CHO triggers alcohol oxidation (Scheme 2), as demonstrated by pH and solvent deuterium KIEs [48,50,54]. Recent mutagenesis, kinetic and spectroscopic studies showed that His466 acts as a catalytic base in the deprotonation of choline [55]. The reaction of H+ transfer is fast (≥1900 s− 1) in WT CHO, occurring in the mixing time of the stopped-flow spectrophotometer [50]. In contrast, the subsequent H– transfer from the α-C of the choline alkoxide to the N(5) atom of the enzyme-bound FAD is the slowest kinetic step in catalysis (Scheme 2) [50]. The effect of temperature on the kinetic parameters and their KIEs for WT CHO revealed that H– transfer occurs within a highly reorganized TRS with a narrow distribution of DADs [23]. Hydrogen bonding and electrostatic interactions with several active site residues, including Ser101, Glu312, His351 and His466, contribute to the stabilization and the optimal positioning of the alkoxide species in the enzyme-alkoxide complex for the subsequent H– transfer reaction [37,39,44,45,56]. No conclusions on the mode of transfer of the H+ could be drawn for WT CHO because the step is too fast and it is not observable in a stopped-flow spectrophotometer [50]. The crystal structure of WT CHO in a complex with glycine betaine

Fig. 1. Superimposition of the active site of WT CHO in complex with glycine betaine (gray carbons; PDB ID: 4MJW) and S101A CHO mutant (green carbons; PDB ID: 3NNE). Distances in Å between residues and either the ligand or FAD are indicated close to the broken lines.

shows that the hydroxyl group of Ser101 is ~3.8 Å away from the carboxylate of the ligand and the flavin C(4a) and N(5) atoms (Fig. 1) [57]. Other flavoproteins also have a residue acting as an hydrogenbond donor located within hydrogen-bond distance to the flavin N(5) atom [58]. Examples include pyranose 2-oxidase (T169) [59], alditol oxidase (S106) [60], the electron-transfer flavoprotein from the methylotrophic bacterium W3A1 (S254) [61], and the oxygenase component of p-hydroxyphenylacetate-3-hydroxylase (S171). Our group has previously shown that mutagenesis of Ser101 in CHO yields variant enzymes displaying at least 15-fold decreases in the rate constant for the cleavage of the substrate OH bond as compared to the WT CHO, with values ≤125 s− 1 at 25 °C [37]. We also observed that the hydrophilic character of Ser101 has a positive effect on the rate constant for hydride transfer from the substrate to the flavin N(5) atom [37]. In Scheme 2. Proton and hydride transfers in the reaction of WT CHO with choline. For FAD, R is ribitol adenosine diphosphate.

1471

BBA - Proteins and Proteomics 1865 (2017) 1470–1478

R. Uluisik et al.

respectively, and C is the absorption at infinite time that accounts for the absorbance of the reduced enzyme. The kinetic parameters associated with the fast and slow phases of flavin reduction were determined by using Eq. (2) [37]. In this equation, S is the concentration of substrate, klim is the limiting rate constant at saturated substrate concentration and appKd is the apparent dissociation constants for substrate and enzyme binding to yield competent enzyme-substrate complexes evolving to enzyme-product complexes. The temperature dependence of klim was determined according to the Eyring formalism (Eq. (3)), where kB and h are the Boltzmann and Planck constants, respectively. The enthalpy of activation (ΔH‡) is calculated from the slope of the plot, whereas the entropy of activation (ΔS‡) is calculated from the y-intercept of the plot. Both ΔS‡ and ΔH‡ are assumed to be independent of temperature, which generally holds well for the relatively small temperature range investigated [63]. The temperature dependence of the KIEs on klim were determined by fitting the data with the logarithmic form of the Arrhenius equation (Eq. (4)), in which the isotope effect on the preexponential factor (AH/AD) is calculated from the y-intercept of the plot and the isotope effect on the energy of activation [ΔEa = Ea(D) − Ea(H)] is calculated from the slope of the plot. R is the universal gas constant (8.314 J K− 1 mol− 1), and T is the absolute temperature. The temperature dependence of the klim values and their KIE were interpreted using Eq. (5), which is an extension of the TST equation including the tunneling probability. The terms in this equation are described in [34]. Global fitting of the diode array data to an A → B → C model was carried out using SPECFIT/32 software (Spectrum Software Associates) to generate calculated spectra, rate constants, and concentration profiles for the spectral species.

all Ser101 variants investigated, the decrease in absorbance at 450 nm that is associated with the cleavage of the substrate OH bond was too large to be due to the presence of the alkoxide intermediate close to the flavin [37]. These findings prompted us to perform the mechanistic and spectroscopic studies presented here. We selected the S101A CHO among all S101 variants for the present study because previous crystallographic observations showed that both the 3D structure of the enzyme and the topology of the active site are not affected by replacing Ser101 with alanine [40]. In addition, the rates of the S101A variant for the cleavage of the substrate OH and CH bonds are adequate for stopped-flow measurements under the required conditions. Specifically, we have investigated in this study the mode of H+ and H– transfers in the S101A enzyme by using temperature KIEs and rapid kinetics techniques, allowing us to make conclusions on the mode of H+ in the reaction catalyzed by CHO. Also, this study revealed that the anaerobic reaction of S101A CHO with choline involves the formation of a transient covalent flavin-substrate adduct previously not detected in the WT enzyme and several of the other mutant variants of CHO previously investigated. 2. Materials and methods 2.1. Materials Recombinant CHO from A. globiformis strain ATCC 8010 with Ser101 replaced with alanine was expressed from the plasmid pET/ codAmg-S101A using Escherichia coli Rosetta(DE3)pLysS cells and purified to high levels as judged by SDS-PAGE as previously described [40]. Choline chloride was from ICN (Aurora, OH). 1,2-[2H4]-Choline bromide (D-choline) was from Isotech INC (Miamisburg, OH). Aspergillus niger glucose oxidase and glucose were from Sigma-Aldrich (St. Louis, MO). All other reagents used were of the highest purity commercially available.

A 450 = B1 e−k obs1 t + B2 e−k obs2 t + C

(1)

k obs = (klim S ) (K dapp + S )

(2)

ln(klim T ) = ln(kB h) + (ΔS ‡ R) − (ΔH ‡ RT )

(3)

ln KIE = ln(AH AD ) − (ΔEa RT )

(4)

2.2. Rapid kinetics Stopped-flow experiments were conducted using a Hi-Tech SF-61 double-mixing stopped-flow spectrophotometer (TgK Scientific, Bath, UK) thermostatted at 15–39 °C. A xenon lamp and either photomultiplier tube detection (450 nm or 400 nm) or diode array detection were used. The enzyme, substrate, and instrument were made anaerobic as previously described [24]. Equal volumes of S101A CHO and choline or D-choline were mixed in the presence of glucose (5 mM) and glucose oxidase (0.5 μM). The final concentration of S101A CHO was 10–30 μM, based on its extinction coefficient at 452 nm (9.5 mM− 1 cm− 1) [37]. Substrate concentrations were ≥ 10-times higher than the enzyme concentration, ensuring pseudo-first-order kinetic conditions. All solutions were prepared in 50 mM sodium pyrophosphate, pL 10.0. For the determination of solvent deuterium KIEs, all solutions were prepared using 99.9% deuterium oxide (D2O). The pD values were determined by addition of 0.4 to the pH electrode readings [62]. NaOD was used to adjust the pD value. The CHO solutions in buffer prepared with H2O were loaded into a PD-10 column equilibrated with buffer made with D2O just prior starting the procedure to make anaerobic the enzyme solution inside a stopped-flow tonometer.

k = C(T) e−ΔG

‡ (RT )

∫0

∞

P(m,DAD) e−(E(DAD)

(kB T )) dDAD

(5)

3. Results 3.1. Kinetic isotope effects at 15 °C Solvent deuterium KIEs were determined for S101A CHO to investigate H+ transfer in the reaction of choline oxidation catalyzed by the enzyme. The kobs values for anaerobic flavin reduction were determined in a stopped-flow spectrophotometer by mixing the enzyme with choline at pL 10.0 and 15 °C and monitoring absorbance changes over time at 450 nm. pH 10 has previously been shown to be in the pH independent region for the WT CHO [48,50,54] and a number of enzyme variants with mutations in the active site, including E312D [45], H351A [44], H466A [56], H99N [21], and the S101A enzyme (data not shown). Anaerobic mixing of the enzyme with choline in buffer prepared with H2O resulted in the two-electron reduction of the enzymebound flavin, following a biphasic process (Fig. 2A). Accordingly, the stopped-flow traces were best fit with a two-exponential process, defining fast (kobs1) and slow (kobs2) observed rate constants that were well separated (i.e., with kobs1/kobs2 ≥ 9). The klim values at saturating choline were determined for both the fast and slow kinetic phases by fitting the hyperbolic patterns of kobs1 and kobs2 as a function of choline concentration (Fig. 2B). In contrast, the appKD values could not be determined accurately since they were smaller than the lowest concentration of substrate that could be used in the reaction to maintain pseudo first-order conditions. For this reason the appKD values are not considered in this study.

2.3. Data analysis The KaleidaGraph software package (Synergy Software, Reading, PA) and the Kinetic Studio Software Suite (Hi-TgK Scientific, Bath, UK) were used to fit the kinetic data. Stopped-flow traces were best fit with the Eq. (1), which describes a double exponential process where kobs1 and kobs2 are the observed first-order rate constants associated with the absorption changes of the fast and slow phases, respectively, t is the time, A is the absorption at 450 nm at any given time, B1 and B2 are the amplitudes of the absorption changes for the fast and slow phases, 1472

BBA - Proteins and Proteomics 1865 (2017) 1470–1478

R. Uluisik et al.

Fig. 2. Solvent and substrate KIEs on the anaerobic reaction of S101A CHO. S101A CHO was reacted with either choline or D-choline at 15 °C to obtain stoppedflow traces at 450 nm. Buffer prepared with either H2O or D2O at pL 10.0 was used. Panels A and C show representative reaction traces using saturating choline or D-choline (6 mM). For the sake of clarity one experimental point of every five is shown (vertical lines). Each reaction was assayed in triplicate.

suggested that the cleavage of the substrate OH bond in S101A CHO may be associated with the transient formation of an enzyme-substrate intermediate. To evaluate this hypothesis, additional stopped-flow studies using either photomultiplier tube detection (at 400 nm or 450 nm) or diode array detection were carried out. Low temperature (15 °C), D-choline, and solutions prepared with D2O were used for the experiments to slow down the enzyme reactions. The first spectrum recorded in the stopped-flow spectrophotometer after mixing S101A CHO with D-choline showed the characteristic spectrum of oxidized S101A CHO with maxima at 373 nm and 452 nm [37]. Reaction of the oxidized S101A CHO with the substrate resulted in the accumulation of an intermediate with decreased absorbance at 450 nm and increased absorbance at 400 nm (Fig. 3). This becomes evident by observing the difference spectra obtained when the first spectrum in Fig. 4A (0.001 s) is subtracted from all spectra in Fig. 4A (Fig. 4B). This analysis also shows that the maximum amount of intermediate is achieved after 0.3 s. The intermediate decayed yielding the two-electron reduced enzyme-bound flavin, as evidenced by the complete bleaching of the absorption band at 452 nm (Fig. 4A). The diode array data were analyzed globally by fitting observed spectra to an A → B → C model using the SpecFit software. An excellent agreement was found between rates of enzyme reaction estimated by global analysis and values obtained by single-wavelength analysis at 450 nm (Fig. 4D and Tables S1–S2). In this model, A, B and C are spectral species reflecting the fully oxidized enzyme, a mix of oxidized enzyme and a reaction intermediate, and the fully reduced enzyme, respectively. Fig. 4C shows the computed absorption spectrum of the reaction intermediate, which is the difference spectrum obtained when the spectrum corresponding to half of the initial enzyme concentration is subtracted from the spectrum B. This approach was carried out under the assumption that the reaction intermediate does not absorb at 450 nm. The extracted spectrum of the reaction intermediate calculated as described above has an absorbance maximum at 389 nm (Fig. 4C). This assumption seems to be possible, based on previous studies showing that flavin-substrate adducts exhibit an absorbance maximum in the 300–400 nm region and low absorbance in the 450 nm region [71,72]. Therefore, we believe that in the spectral species B the remaining absorbance at 450 nm may due to the presence of a fraction of oxidized enzyme and thus the subtraction is justified to study this hypothesis.

Replacement of H2O with D2O had a large effect on the limiting value of kobs1 at saturating choline (klim1), with values of 81 s− 1 in H2O and 18 s− 1 in D2O (Fig. 2B). This corresponds to a solvent KIE of 4.5 (± 0.3) on the fast kinetic phase seen in the stopped-flow spectrophotometer at 15 °C. In contrast, much smaller differences were seen on the klim2 value, for which the solvent KIE was 1.7 (± 0.1). These results are in line with previous studies at 25 °C showing solvent KIEs of 3.8 on the klim1 value and 1.7 on the klim2 value [37]. In that study it was also demonstrated that the klim1 and klim2 values were indifferent to nonprotogenic solvents of increased viscosity [37]. The lack of solvent viscosity effects establishes the solvent KIE as a probe for H+ transfers involving solvent exchangeable sites rather than solvent-sensitive equilibria involving enzyme-substrate complexes. Thus, with the S101A enzyme the fast phase of flavin reduction is assigned to the cleavage of the substrate OH bond, i.e., the H+ transfer reaction catalyzed by CHO. Substrate KIEs were determined with choline and 1,2-[2H4]-choline in the same conditions described above to probe H– transfer in the S101A enzyme. While there were small differences in the klim1 value upon substituting choline with deuterated choline, with rate constants of 81 s− 1 and 71 s− 1, large differences where seen in the klim2 value, with rate constants of 3.1 s− 1 and 0.4 s− 1 (Fig. 2C–D). Accordingly, the substrate KIE was 1.1 (±0.1) on the klim1 value and 8.2 (± 0.3) on the klim2 value at 15 °C. These data agree well with a previous study at 25 °C showing substrate KIEs of 1 on the klim1 value and 6.5 on the klim2 value, in which it was concluded that the slow phase of flavin reduction is associated with the H– transfer reaction catalyzed by the S101A enzyme variant [37]. 3.2. Spectral evidence for a transient flavin-substrate adduct The anaerobic reduction of S101A CHO with choline is a biphasic process, based on the reaction traces at 450 nm obtained with a stopped-flow instrument configured with photomultiplier tube detection. KIE studies revealed that the first phase corresponds to the cleavage of the substrate OH bond while the second phase is associated with the H– transfer reaction (Section 3.1). Fig. 2 shows that the first phase of the reaction accounts for ~50% of the total absorbance change at 450 nm. Deprotonation of ligands in the flavin binding pocket is usually accompanied by small changes in the absorbance of the enzyme-bound flavin. The large absorption change seen after the first phase of reaction 1473

BBA - Proteins and Proteomics 1865 (2017) 1470–1478

R. Uluisik et al.

39 °C. As expected, the klim1 value increased with increasing temperature. The fit of the data acquired in H2O and D2O to the Eyring formalism (Eq. (3)) yielded lines with similar slopes and different y-intercepts (Fig. 5A and Table 1). Accordingly, an Arrhenius analysis of the D2O klim1 value yielded a line with a slope very close to zero (Fig. 5B). By comparing the D2Oklim1 values at for example 15 °C and 39 °C (Table 1 and Table S1) it was evident that the solvent KIE for H+ transfer is temperature-independent, and thus we concluded that the calculated Δ Ea value can be considered negligible and it is not zero due to the experimental error often associated to this type of experiments. The Arrhenius analysis of the temperature dependence of the klim1 value returned an isotope effect on the pre-exponential factor (AH2O/AD2O) of ~110. The effect of temperature on the klim2 and Dklim2 values was investigated to establish if H– transfer in the S101A CHO occurs within a highly reorganized TRS, as for the case of WT CHO [23]. Anaerobic enzyme reactions with either choline or D-choline were carried out in aqueous buffer at pH 10.0 in the temperature range from 15 °C to 39 °C. When the klim2 values were analyzed using an Eyring plot (Eq. (3), Fig. 5C), the protiated substrate yielded a ΔH‡H value significantly larger than the ΔH‡D value obtained for the deuterated substrate (Table 1). Accordingly, a negative Δ Ea value was calculated using the Arrhenius formalism (Eq. (4), Fig. 5D), suggesting that kinetic complexity may be present in the reaction of H– transfer in S101A CHO [64]. The ΔS‡D could not be accurately determined, as evidenced by the high standard deviation value calculated for this parameter. As a result, the standard deviation for the ΔG‡D was very high and thus it was not possible to carry out a fair comparison of the ΔG‡H and ΔG‡D values. Nevertheless, it is clear that kinetic complexity plays a role in the H– transfer reaction catalyzed by S101A CHO, based on the ΔH‡H and ΔH‡D values. This fact precludes us from drawing further mechanistic conclusions about the hydride transfer reaction in this mutant.

Fig. 3. Stopped-flow traces for the reaction of S101A CHO with D-choline in solutions prepared with D2O at 15 °C. PMT detection was used. For the sake of clarity one experimental point out of every five is shown (vertical lines).

3.3. Temperature effects on H+ and H– transfers The effect of temperature on the klim1 and D2Oklim1 values was determined to assess the contribution of QM tunneling to the H+ transfer catalyzed by S101A CHO. Anaerobic enzyme reductions were carried out as described above with choline as substrate and either H2O or D2O as buffered solvent at pL 10.0 in the temperature range from 15 °C to

4. Discussion The oxidation of choline catalyzed by CHO involves the cleavage of the substrate OH and CH bonds, with H+ and H– transfers taking place on separate kinetic steps (Scheme 2) [46,50]. In the WT enzyme H+ Fig. 4. Spectral course of the reaction of S101A CHO with D-choline in solutions prepared with D2O at 15 °C. (A) Spectra recorded using diode array detection. (B) Difference spectra obtained when the first spectrum in panel A (0.001 s) is subtracted from all spectra in panel A. (C) Calculated absorption spectra for the fully oxidized enzyme (black), mix of oxidized enzyme and covalent flavin-substrate adduct (red), fully reduced enzyme (blue), and covalent flavin-substrate adduct (green). (D) Concentration profiles of the spectral species A, B, and C shown in panel C.

1474

BBA - Proteins and Proteomics 1865 (2017) 1470–1478

R. Uluisik et al.

Fig. 5. Temperature dependence of the klim1 (A), D2O klim1 (B), klim2 (C), and Dklim2 (D) values for S101A CHO. S101A CHO was anaerobically reacted with increasing concentrations of either choline or D-choline at 13–39 °C to obtain stopped-flow traces at 450 nm and the corresponding klim values. Buffer prepared with either H2O or D2O at pL 10.0 was used. Each reaction was assayed in triplicate. The error bars represent the standard error. When no error bar is visible, it is smaller than the symbol.

temperature-independent KIEs observed for H+ transfer in S101A CHO indicate that conformational sampling results in a narrow distribution of DADs at the TRS. Contrarily, temperature-dependent KIEs with a positive Δ Ea value (i.e., activation energy for the H– transfer using nondeuterated substrate lower than that using deuterated substrate) is expected in enzymes presenting a poorly reorganized TRS with a wide range of DADs at thermal equilibrium. This is due to the fact that the lighter isotope can tunnel over longer distances than the heavier isotope and thermally activated DAD fluctuations populate shorter DADs at high temperatures. The slow phase of flavin reduction seen in the stopped-flow spectrophotometer is associated with the H– transfer reaction from the alkoxide choline to the flavin N(5) atom, based on the substrate KIE studies. The activation energy for the H– transfer using choline was higher than that using D-choline, suggesting that kinetic complexity from kinetic steps other than H– transfer affects the Dklim2 values. Consequently, the observed KIEs are deflated from the intrinsic values and, unless kinetic complexity is accounted for, cannot be used for mechanistic conclusions [64,65]. For these reasons, we could not carry out an in-depth analysis of the contribution of the serine hydroxyl to the H– tunneling reaction catalyzed by CHO. Kinetic complexity arises from the presence of kinetic steps, other than the chemical step under study, which contribute to the observed rate. For S101A CHO, we hypothesize that the source of kinetic complexity may be a slow isomerization of the enzyme-substrate complex taking place before fast flavin reduction. Indeed, it was previously proposed that deprotonation of the alcohol substrate may trigger a protein isomerization prior the hydride transfer step for WT CHO and E312D CHO [22,45]. In the case of WT CHO, support for this conclusion comes from the comparison of the effects of pH and temperature on the primary kinetic isotope effect on the steadystate kinetic parameters determined in reversible and irreversible catalytic regimes [22]. Glu312 is an active site residue important for binding and positioning of the substrate in CHO. In E312D CHO, the presence of an isomerization of the enzyme-substrate complex occurring prior to the cleavage of the substrate CeH bond was revealed by studying the effect of pH and viscosity on the primary kinetic isotope effect on the steady-state kinetic parameters [45]. A similar approach was used with other enzyme catalyzing the oxidation of hydroxyl groups yielding carbonyl moieties, flavocytochrome b2, to discover a solvent-sensitive isomerization of the enzyme-substrate complex taking

Table 1 Thermodynamic parameters for CHO S101A variant.a H– transfer

H+ transfer ΔH‡H2O, kcal/mol ΔH‡D2O, kcal/mol ΔS‡H2O, kcal/mol K ΔS‡D2O, kcal/mol K ΔG‡H2Ob, kcal/mol ΔG‡D2Ob, kcal/mol AH2O/AD2O ΔEa, kcal/mol D2O klim1 15 °C D2O klim1 39 °C

8.8 ± 0.4 6.9 ± 0.8 −0.019 ± 0.001 −0.028 ± 0.004 14.5 ± 0.5 15.3 ± 1.5 103 ± 29 −2 ± 1 4.5 ± 0.3 5.2 ± 0.4

ΔH‡H, kcal/mol ΔH‡D, kcal/mol ΔS‡H, kcal/mol K ΔS‡D, kcal/mol K ΔG‡Hb, kcal/mol ΔG‡Db, kcal/mol AH/AD ΔEa, kcal/mol D klim2 15 °C D klim2 39 °C

12.4 ± 1.2 6.0 ± 1.2 − 0.013 ± 0.001 − 0.039 ± 0.020 16.3 ± 1.2 17.7 ± 6.2 (5.4 ± 1.0) × 105 −6 ± 1 8.2 ± 0.3 20.3 ± 1.1

a Conditions: anaerobic reactions at 15–39 °C and pL 10.0. Choline and D-choline were used as substrates. b Gibbs free energy of activation (ΔG‡ = ΔH‡-TΔS‡) calculated for 25 °C.

transfer is fast occurring within the mixing time of the stopped-flow spectrophotometer [50]. Replacement of Ser101 with alanine in the enzyme active site significantly decreases the rate of reaction [37], allowing for a mechanistic investigation of the H+ transfer reaction that was not previously possible in CHO. The reaction of H+ transfer in S101A CHO occurs within a highly reorganized TRS with a narrow distribution of DADs. The evidence for this conclusion comes from the determination of solvent KIEs at various temperatures for the fast phase observed in the reaction of S101A CHO with choline. The results were interpreted using the model described in [34], which is an extension of the TST for adiabatic reactions like the H+ transfer catalyzed by S101A CHO. The klim1 values for S101A CHO increased with increasing temperature as shown in Fig. 5A. By fitting the D2Oklim1 values to the Arrhenius equation, the activation energies for the H+ transfer using solutions prepared with H2O or D2O were similar, yielding a temperature-independent D2Oklim1 value and a deuterium isotope effect on the Arrhenius pre-exponential factor significantly larger than one (Table 1). To date, the temperature dependence of the KIE is the most robust tool for investigating tunneling and conformational sampling in enzymes catalyzing H+ or H– transfers [11,23,25–33]. During the enzyme reaction, the enzyme-substrate complex samples different conformations to bring the donor and acceptor closer to each other [34]. The temperature-dependent rates and 1475

BBA - Proteins and Proteomics 1865 (2017) 1470–1478

R. Uluisik et al.

Scheme 3. Proposed mechanism for the reaction of S101A CHO with choline. For FAD, R is ribitol adenosine diphosphate.

place prior to CeH bond cleavage [66]. It is improbable that the removal of a functional group from the active site of an enzyme, as for the case of the serine hydroxyl of CHO that decreases ≥ 15-fold the rate constant for H+ transfer, may increase the reorganization of the TRS for H+ transfer with respect to the WT enzyme. We therefore suggest that a reorganized TRS is likely already present in the WT enzyme, probably due to the extended network of Hbonding and electrostatic interactions of the alcohol substrate with active site residues other than Ser101, e.g., with Glu312 [45], His351 [44], and His466 [56]. Thus, the decrease in the rate constant for H+ transfer associated with the removal of the serine hydroxyl from the enzyme active site is probably not due to functional changes that alter the conformational sampling to achieve a narrow distribution of DADs at the TRS. Previous crystallographic studies showed that the only difference between the S101A mutant and WT CHO is the lack of a hydroxyl group on residue 101 [40,57]. Therefore, the decrease in the rate constant for H+ transfer in the S101A CHO compared to WT CHO can not be attributed to structural changes in the topology of the enzyme active site either. Instead, we suggest that the abstraction of the substrate hydroxyl proton in WT CHO is faster than in S101A CHO because, in WT CHO, the transition state for the OeH bond cleavage is stabilized by a hydrogen bond between the side chain of Ser101 and the oxygen atom of the alkoxide species. In alcohol oxidation reactions catalyzed by enzymes a substrate OH bond cleavage occurs concomitantly or it precedes CH bond cleavage either in separate steps or a concerted fashion [68,69]. The simplest such mechanism involves direct H– transfer from the alkoxide intermediate to the flavin N(5) atom (Scheme 2) [67,68]. However, alternative mechanisms have been previously considered due to the great versatility of the isoalloxazine ring [69,70], among which, a nucleophilic attack of the alkoxide on the flavin C(4a) atom was proposed [69,70]. The resulting flavin-substrate adduct would exhibit an absorbance maximum in the 300–400 nm region, based on previous studies [71,72]. These spectral properties agree with the spectrum that we calculated for the S101A CHO using the diode array data and an A → B → C model, which exhibits a maximum at 389 nm. An increase in absorbance in the long-wavelength region (λ > 500 nm) was not observed during the first reaction phase for S101A CHO, ruling out that the intermediate is a charge transfer complex between the oxidized S101A CHO and D-choline. The intermediate in the S101A CHO reaction was not assigned either to the anionic flavin semiquinone state. The formation of this semiquinone state in CHO results in an increment in absorbance at both 372 nm and 530 nm, in contrast to the formation of the transient intermediate in spectral species B (Fig. 4C) [48]. Previous studies suggested that the anionic flavin semiquinone of CHO does not participate in the catalysis [48]. We also ruled out a model in which the fast and slow kinetic phases are due to the presence of an enzyme mix containing fast- and slow-reacting species with a different rate-limiting step controlling the flavin reduction rate (H+ and H–

transfers, respectively), since this model does not explain the increase in absorbance at 389 nm associated to the first observed phase. Therefore, we propose that the first reaction phase in S101A CHO involves the transient occurrence of a covalent flavin-substrate adduct, probably involving a nucleophilic attack of the alkoxide on the flavin C(4a) atom (Scheme 3). In this regard, there is no evidence so far suggesting that a similar enzyme-substrate adduct is also formed in the reaction catalyzed by WT CHO or other mutant variants of CHO. 5. Conclusions The mechanistic investigation presented here established that conformational sampling in the S101A variant of CHO results in a narrow distribution of DADs at the TRS for H+ transfer. H+ tunneling in WT CHO, which is too fast to be investigated by rapid kinetics, may also occur within a highly-reorganized enzyme-substrate complex since the alternate scenario that the removal of the serine hydroxyl from the enzyme active site confers increased reorganization is unlikely. Regarding H– tunneling, a previous study demonstrated a highly reorganized TRS for WT CHO [22,23]. However, mechanistic conclusions could not be drawn here for S101A CHO due to the presence of kinetic complexity. This observation further demonstrates that kinetic complexity, which was initially considered to explain why measured and intrinsic KIEs sometimes differ when using steady-state kinetics [65], may also affect rate constants determined using rapid kinetic approaches. The study presented here has also revealed that a transient covalent enzyme-substrate adduct is observed in the reductive half-reaction of the enzyme lacking a hydroxyl group on residue 101. In contrast, such a covalent enzyme-substrate adduct is not formed in WT CHO, which shows the flavin oxidized after deprotonation of the alcohol substrate in catalysis. Thus, the removal of the serine hydroxyl side chain proximal to the flavin C4a atom of the enzyme results in a drastic change in the chemical mechanism of catalysis of CHO. This conclusion strongly supports the notion that, besides modulating the thermodynamic properties of the flavin, participating in substrate binding or catalysis, an important role of amino acid residues in the active site of flavin-dependent enzymes is to eliminate unproductive reactions of the versatile enzyme-bound flavin for the reaction that needs to be catalyzed. Conflict of interest The authors declare no conflict of interest. Transparency document The http://dx.doi.org/10.1016/j.bbapap.2017.08.004 associated with this article can be found in the online version. 1476

BBA - Proteins and Proteomics 1865 (2017) 1470–1478

R. Uluisik et al.

[25] C.R. Pudney, A. Guerriero, N.J. Baxter, L.O. Johannissen, J.P. Waltho, S. Hay, N.S. Scrutton, Fast protein motions are coupled to enzyme H-transfer reactions, J. Am. Chem. Soc. 135 (2013) 2512–2517. [26] Z. Wang, T. Abeysinghe, J.S. Finer-Moore, R.M. Stroud, A. Kohen, A remote mutation affects the hydride transfer by disrupting concerted protein motions in thymidylate synthase, J. Am. Chem. Soc. 134 (2012) 17722–17730. [27] V. Stojković, L.L. Perissinotti, D. Willmer, S.J. Benkovic, A. Kohen, Effects of the donor–acceptor distance and dynamics on hydride tunneling in the dihydrofolate reductase catalyzed reaction, J. Am. Chem. Soc. 134 (2012) 1738–1745. [28] D. Roston, C.M. Cheatum, A. Kohen, Hydrogen donor–acceptor fluctuations from kinetic isotope effects: a phenomenological model, Biochemistry 51 (2012) 6860–6870. [29] S. Hay, L.O. Johannissen, P. Hothi, M.J. Sutcliffe, N.S. Scrutton, Pressure effects on enzyme-catalyzed quantum tunneling events arise from protein-specific structural and dynamic changes, J. Am. Chem. Soc. 134 (2012) 9749–9754. [30] M. Yoon, H. Song, K. Hakansson, E.N.G. Marsh, Hydrogen tunneling in adenosylcobalamin-dependent glutamate mutase: evidence from intrinsic kinetic isotope effects measured by intramolecular competition, Biochemistry 49 (2010) 3168–3173. [31] O.A. Oyeyemi, K.M. Sours, T. Lee, K.A. Resing, N.G. Ahn, J.P. Klinman, Temperature dependence of protein motions in a thermophilic dihydrofolate reductase and its relationship to catalytic efficiency, Proc. Natl. Acad. Sci. 107 (2010) 10074–10079. [32] E.J. Loveridge, L.-H. Tey, R.K. Allemann, Solvent effects on catalysis by Escherichia coli dihydrofolate reductase, J. Am. Chem. Soc. 132 (2010) 1137–1143. [33] J.P. Klinman, An integrated model for enzyme catalysis emerges from studies of hydrogen tunneling, Chem. Phys. Lett. 471 (2009) 179–193. [34] A. Kohen, Role of dynamics in enzyme catalysis: Substantial versus semantic controversies, Acc. Chem. Res. 48 (2015) 466–473. [35] S. Gannavaram, G. Gadda, Relative timing of hydrogen and proton transfers in the reaction of flavin oxidation catalyzed by choline oxidase, Biochemistry 52 (2013) 1221–1226. [36] G. Gadda, Oxygen activation in flavoprotein oxidases: the importance of being positive, Biochemistry 51 (2012) 2662–2669. [37] H. Yuan, G. Gadda, Importance of a serine proximal to the C(4a) and N(5) flavin atoms for hydride transfer in choline oxidase, Biochemistry 50 (2011) 770–779. [38] K. Rungsrisuriyachai, G. Gadda, Role of asparagine 510 in the relative timing of substrate bond cleavages in the reaction catalyzed by choline oxidase, Biochemistry 49 (2010) 2483–2490. [39] O. Quaye, T. Nguyen, S. Gannavaram, A. Pennati, G. Gadda, Rescuing of the hydride transfer reaction in the Glu312Asp variant of choline oxidase by a substrate analogue, Arch. Biochem. Biophys. 499 (2010) 1–5. [40] S. Finnegan, H. Yuan, Y.-F. Wang, A.M. Orville, I.T. Weber, G. Gadda, Structural and kinetic studies on the Ser101Ala variant of choline oxidase: catalysis by compromise, Arch. Biochem. Biophys. 501 (2010) 207–213. [41] S. Finnegan, J. Agniswamy, I.T. Weber, G. Gadda, Role of valine 464 in the flavin oxidation reaction catalyzed by choline oxidase, Biochemistry 49 (2010) 2952–2961. [42] Y. Xin, G. Gadda, D. Hamelberg, The cluster of hydrophobic residues controls the entrance to the active site of choline oxidase, Biochemistry 48 (2009) 9599–9605. [43] A.M. Orville, G.T. Lountos, S. Finnegan, G. Gadda, R. Prabhakar, Crystallographic, spectroscopic, and computational analysis of a flavin C4a-oxygen adduct in choline oxidase, Biochemistry 48 (2009) 720–728. [44] K. Rungsrisuriyachai, G. Gadda, On the role of histidine 351 in the reaction of alcohol oxidation catalyzed by choline oxidase, Biochemistry 47 (2008) 6762–6769. [45] O. Quaye, G.T. Lountos, F. Fan, A.M. Orville, G. Gadda, Role of Glu312 in binding and positioning of the substrate for the hydride transfer reaction in choline oxidase, Biochemistry 47 (2008) 243–256. [46] G. Gadda, Hydride transfer made easy in the reaction of alcohol oxidation catalyzed by flavin-dependent oxidases, Biochemistry 47 (2008) 13745–13753. [47] F. Fan, M. Ghanem, G. Gadda, Cloning, sequence analysis, and purification of choline oxidase from Arthrobacter globiformis: a bacterial enzyme involved in osmotic stress tolerance, Arch. Biochem. Biophys. 421 (2004) 149–158. [48] M. Ghanem, F. Fan, K. Francis, G. Gadda, Spectroscopic and kinetic properties of recombinant choline oxidase from Arthrobacter globiformis, Biochemistry 42 (2003) 15179–15188. [49] S. Ikuta, S. Imamura, H. Misaki, Y. Horiuti, Purification and characterization of choline oxidase from Arthrobacter globiformis, J. Biochem. 82 (1977) 1741–1749. [50] F. Fan, G. Gadda, On the catalytic mechanism of choline oxidase, J. Am. Chem. Soc. 127 (2005) 2067–2074. [51] G. Gadda, Kinetic mechanism of choline oxidase from Arthrobacter globiformis, Biochim. Biophys. Acta 1646 (2003) 112–118. [52] K. Lambou, A. Pennati, I. Valsecchi, R. Tada, S. Sherman, H. Sato, R. Beau, G. Gadda, J.-P. Latgé, Pathway of glycine betaine biosynthesis in Aspergillus fumigatus, Eukaryot. Cell 12 (2013) 853–863. [53] T.H. Chen, N. Murata, Glycinebetaine protects plants against abiotic stress: mechanisms and biotechnological applications, Plant Cell Environ. 34 (2011) 1–20. [54] G. Gadda, pH and deuterium kinetic isotope effects studies on the oxidation of choline to betaine-aldehyde catalyzed by choline oxidase, Biochim. Biophys. Acta, Proteins Proteomics 1650 (2003) 4–9. [55] C.L. Smitherman, K. Rungsrisuriyachai, M.W. Germann, G. Gadda, Identification of the catalytic base for alcohol activation in choline oxidase, Biochemistry 54 (2015) 413–421. [56] M. Ghanem, G. Gadda, On the catalytic role of the conserved active site residue His466 of choline oxidase, Biochemistry 44 (2005) 893–904.

Acknowledgments This work was supported in part by the National Science Foundation grant CHE-1506518 (G.G.). Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.bbapap.2017.08.004. References [1] D. Herschlag, A. Natarajan, Fundamental challenges in mechanistic enzymology: progress toward understanding the rate enhancements of enzymes, Biochemistry 52 (2013) 2050–2067. [2] B.G. Miller, R. Wolfenden, Catalytic proficiency: the unusual case of OMP decarboxylase, Annu. Rev. Biochem. 71 (2002) 847–885. [3] J.P. Klinman, A. Kohen, Hydrogen tunneling links protein dynamics to enzyme catalysis, Annu. Rev. Biochem. 82 (2013) 471–496. [4] J.P. Klinman, Importance of protein dynamics during enzymatic CeH bond cleavage catalysis, Biochemistry 52 (2013) 2068–2077. [5] S. Hammes-Schiffer, Catalytic efficiency of enzymes: a theoretical analysis, Biochemistry 52 (2012) 2012–2020. [6] J. Pu, J. Gao, D.G. Truhlar, Multidimensional tunneling, recrossing, and the transmission coefficient for enzymatic reactions, Chem. Rev. 106 (2006) 3140–3169. [7] D.G. Truhlar, Tunneling in enzymatic and nonenzymatic hydrogen transfer reactions, J. Phys. Org. Chem. 23 (2010) 660–676. [8] M.P. Meyer, J.P. Klinman, Investigating inner-sphere reorganization via secondary kinetic isotope effects in the CeH cleavage reaction catalyzed by soybean lipoxygenase: tunneling in the substrate backbone as well as the transferred hydrogen, J. Am. Chem. Soc. 133 (2010) 430–439. [9] S.C. Sharma, J.P. Klinman, Experimental evidence for hydrogen tunneling when the isotopic Arrhenius prefactor (AH/AD) is unity, J. Am. Chem. Soc. 130 (2008) 17632–17633. [10] Z.D. Nagel, S. Cun, J.P. Klinman, Identification of a long-range protein network that modulates active site dynamics in extremophilic alcohol dehydrogenases, J. Biol. Chem. 288 (2013). [11] Z.D. Nagel, C.W. Meadows, M. Dong, B.J. Bahnson, J.P. Klinman, Active site hydrophobic residues impact hydrogen tunneling differently in a thermophilic alcohol dehydrogenase at optimal versus nonoptimal temperatures, Biochemistry 51 (2012) 4147–4156. [12] T. Shiou-chuan, J.P. Klinman, Probes of hydrogen tunneling with horse liver alcohol dehydrogenase at subzero temperatures, Biochemistry 40 (2001) 2303–2311. [13] A. Yahashiri, G. Nimrod, N. Ben-Tal, E.E. Howell, A. Kohen, The effect of electrostatic shielding on H tunneling in R67 dihydrofolate reductase, Chembiochem 10 (2009) 2620–2623. [14] L. Wang, N.M. Goodey, S.J. Benkovic, A. Kohen, The role of enzyme dynamics and tunnelling in catalysing hydride transfer: studies of distal mutants of dihydrofolate reductase, Philos. Trans. R. Soc. B 361 (2006) 1307–1315. [15] C.R. Pudney, S. Hay, J. Pang, C. Costello, D. Leys, M.J. Sutcliffe, N.S. Scrutton, Mutagenesis of morphinone reductase induces multiple reactive configurations and identifies potential ambiguity in kinetic analysis of enzyme tunneling mechanisms, J. Am. Chem. Soc. 129 (2007) 13949–13956. [16] J. Pang, S. Hay, N.S. Scrutton, M.J. Sutcliffe, Deep tunneling dominates the biologically important hydride transfer reaction from NADH to FMN in morphinone reductase, J. Am. Chem. Soc. 130 (2008) 7092–7097. [17] S.L. Seymour, J.P. Klinman, Comparison of rates and kinetic isotope effects using PEG-modified variants and glycoforms of glucose oxidase: the relationship of modification of the protein envelope to CeH activation and tunneling, Biochemistry 41 (2002) 8747–8758. [18] Q. Su, J.P. Klinman, Nature of oxygen activation in glucose oxidase from Aspergillus niger: the importance of electrostatic stabilization in superoxide formation, Biochemistry 38 (1999) 8572–8581. [19] A. Kohen, T. Jonsson, J.P. Klinman, Effects of protein glycosylation on catalysis: changes in hydrogen tunneling and enthalpy of activation in the glucose oxidase reaction, Biochemistry 36 (1997) 2603–2611. [20] Q. Osbourne, G. Gadda, Effect of a conservative mutation of an active site residue involved in substrate binding on the hydride tunneling reaction catalyzed by choline oxidase, Arch. Biochem. Biophys. 489 (2009) 10–14. [21] Q. Osbourne, S. Cowins, G. Gadda, Contribution of flavin covalent linkage with histidine 99 to the reaction catalyzed by choline oxidase, J. Biol. Chem. 284 (2009) 16990–16997. [22] F. Fan, G. Gadda, An internal equilibrium preorganizes the enzyme-substrate complex for hydride tunneling in choline oxidase, Biochemistry 46 (2007) 6402–6408. [23] F. Fan, G. Gadda, Oxygen- and temperature-dependent kinetic isotope effects in choline oxidase: correlating reversible hydride transfer with environmentally enhanced tunneling, J. Am. Chem. Soc. 127 (2005) 17954–17961. [24] E. Romero, S.T. Ladani, D. Hamelberg, G. Gadda, Solvent-slaved motions in the hydride tunneling reaction catalyzed by human glycolate oxidase, ACS Catal. 6 (2016) 2113–2120.

1477

BBA - Proteins and Proteomics 1865 (2017) 1470–1478

R. Uluisik et al.

catalysis, J. Am. Chem. Soc. 126 (2004) 4778–4779. [65] W.W. Cleland, The use of isotope effects to determine transition-state structure for enzymic reactions, Methods Enzymol. 87 (1982) 625–641. [66] P. Sobrado, S.C. Daubner, P.F. Fitzpatrick, Probing the relative timing of hydrogen abstraction steps in the flavocytochrome b2 reaction with primary and solvent deuterium isotope effects and mutant enzymes, Biochemistry 40 (2001) 994–1001. [67] E. Romero, G. Gadda, Alcohol oxidation by flavoenzymes, Biomol. Concepts 5 (2014) 299–318. [68] T. Wongnate, P. Chaiyen, The substrate oxidation mechanism of pyranose 2-oxidase and other related enzymes in the glucose-methanol-choline superfamily, FEBS J. 280 (2013) 3009–3027. [69] P.F. Fitzpatrick, Substrate dehydrogenation by flavoproteins, Acc. Chem. Res. 34 (2001) 299–307. [70] P.F. Fitzpatrick, Oxidation of amines by flavoproteins, Arch. Biochem. Biophys. 493 (2010) 13–25. [71] S. Ghisla, B. Entsch, V. Massey, M. Husein, On the structure of flavin-oxygen intermediates involved in enzymatic reactions, Eur. J. Biochem. 76 (1977) 139–148. [72] S.M. Miller, V. Massey, D. Ballou, C.H. Williams Jr., M.D. Distefano, M.J. Moore, C.T. Walsh, Use of a site-directed triple mutant to trap intermediates: demonstration that the flavin C(4a)-thiol adduct and reduced flavin are kinetically competent intermediates in mercuric ion reductase, Biochemistry 29 (1990) 2831–2841.

[57] F. Salvi, Y.F. Wang, I.T. Weber, G. Gadda, Structure of choline oxidase in complex with the reaction product glycine betaine, Acta Crystallogr. Sect. D: Biol. Crystallogr. 70 (2014) 405–413. [58] M.W. Fraaije, A. Mattevi, Flavoenzymes: diverse catalysts with recurrent features, TIBS 25 (2000) 126–132. [59] W. Pitsawong, J. Sucharitakul, M. Prongjit, T.-C. Tan, O. Spadiut, D. Haltrich, C. Divne, P. Chaiyen, A conserved active-site threonine is important for both sugar and flavin oxidations of pyranose 2-oxidase, J. Biol. Chem. 285 (2010) 9697–9705. [60] F. Forneris, D.P.H.M. Heuts, M. Delvecchio, S. Rovida, M.W. Fraaije, A. Mattevi, Structural analysis of the catalytic mechanism and stereoselectivity in Streptomyces coelicolor alditol oxidase, Biochemistry 47 (2008) 978–985. [61] K.-Y. Yang, R.P. Swenson, Modulation of the redox properties of the flavin cofactor through hydrogen-bonding interactions with the N (5) atom: role of αSer254 in the electron-transfer flavoprotein from the methylotrophic bacterium W3A1, Biochemistry 46 (2007) 2289–2297. [62] K.B. Schowen, R.L. Schowen, Solvent isotope effects on enzyme systems, Methods Enzymol. 87 (1982) 551–606. [63] E.V. Anslyn, D.A. Dougherty, Modern Physical Organic Chemistry, University Science Books, 2006, pp. 355–420 (Place Published). [64] R.S. Sikorski, L. Wang, K.A. Markham, P.R. Rajagopalan, S.J. Benkovic, A. Kohen, Tunneling and coupled motion in the Escherichia coli dihydrofolate reductase

1478