Vol. 54 No. 2/2007, 379–385 Regular paper

In order to clarify the poorly understood mechanisms of two-electron reduction of quinones by flavoenzymes, we examined the quinone reductase reactions of a member of a structurally distinct old yellow enzyme family, Enterobacter cloacae PB2 pentaerythritol tetranitrate reductase (PETNR). PETNR catalyzes two-electron reduction of quinones according to a 'ping-pong' scheme. A multiparameter analysis shows that the reactivity of quinones increases with an increase in their single-electron reduction potential and pK(a) of their semiquinones (a three-step (e(-),H(+),e(-)) hydride transfer scheme), or with an increase in their hydride-transfer potential (E(7)(H(-))) (a single-step (H(-)) hydride transfer scheme), and decreases with a decrease in their van der Waals volume. However, the pH-dependence of PETNR reactivity is more consistent with a single-step hydride transfer. A comparison of X-ray data of PETNR, mammalian NAD(P)H : quinone oxidoreductase (NQO1), and Enterobacter cloacae nitroreductase, which reduce quinones in a two-electron way, and their reactivity revealed that PETNR is much less reactive, and much less sensitive to the quinone substrate steric effects than NQO1. This may be attributed to the lack of pi-pi stacking between quinone and the displaced aromatic amino acid in the active center, e.g., with Phe-178' in NQO1.


INTRODUCTION
The therapeutic action and/or toxicity of quinoidal drugs and environmental pollutants is frequently attributed to their flavoenzyme-catalyzed reduction.Single-electron reduction of quinones to their anion-radicals initiates their redox cycling under aerobic conditions, and leads to the oxidative stress (O'Brien, 1991).Typically, these reactions are performed by flavoenzymes dehydrogenases-electrontransferases, e.g., NADPH : cytochrome P-450 reductase (EC 1.6.4.2) and ferredoxin : NADP + reductase (EC 1.18.1.2),which contain stabilized neutral flavin semiquinones, and possess natural single-electron accepting oxidants (Iyanagi, 1990, and references therein).These reactions follow an 'outer-sphere' electron transfer with a weak electronic interaction between the reactants (Marcus & Sutin, 1985), and are characterized by a linear or parabolic dependence of log (rate constant) on the single-electron reduction potential of quinones, E 1 7 , with an absence of pronounced effects of quinone structure (Čėnas et al., 1994;Anusevičius et al., 1997).

2007
L. Misevičienė and others Two-electron reduction of quinones frequently contributes to their detoxification, because the relatively stable hydroquinones undergo conjugation, and are readily excreted from organism (O'Brien, 1991, and references therein).However, due to very few mechanistic studies, the mechanism(s) of two-electron (hydride) reduction of quinones by flavoenzymes are unclear, except for the requirement of the instability of anionic flavin semiquinone in the enzyme active center (Tedeschi et al., 1995;Koder et al., 2002, and references therein).In the most thoroughly studied reactions of NAD(P)H : quinone oxidoreductase (NQO1, DT-diaphorase, EC 1.6.99.2) and Enterobacter cloacae nitroreductase (NR, EC 1.6.99.7), the reactivity of quinones in general increases upon an increase in their electron (hydride)-accepting potency, but it is also strongly influenced by their structural peculiarities (Anusevičius et al., 2002;Nivinskas et al., 2002).Thus, it is unclear whether the two-electron reduction of quinones requires their specific interaction(s) with the flavoenzyme active center as well, e.g., the π-π stacking of quinone with the displaced aromatic amino-acid residue, like Phe-178' in NQO1 (Li et al., 1995).For this reason, it is necessary to extend mechanistic studies of two-electron reduction using flavoenzymes of different groups and families.
In this study, we have obtained a quantitative structure-activity relationship in the reduction of quinones by PETNR, and distinguished the effects of the electron/hydride-accepting potency of quinones and their steric parameters.
All the kinetic measurements were carried out in 0.05 M K-phosphate (pH 7.0), containing 1 mM EDTA at 25 o C, using a Hitachi-557 spectrophotometer.The initial rates of PETNR-catalyzed oxidation of NADPH by quinones were determined by monitoring NADPH oxidation (∆ε 340 = 6.2 mM -1 cm -1 ).The kinetic parameters of reactions, i.e., the catalytic constant (k cat ) and the bimolecular rate constant (k cat /K m ) of quinone reduction correspond to the reciprocal intercepts and slopes of the Lineweaver-Burk plots, where v is the reaction rate, [E] is the enzyme concentration, and [Q] is quinone concentration.The k cat corresponds to the number of molecules of NADPH oxidized by the enzyme active center per second.The quinone substrate inhibition constants (K iq ) at fixed NADPH concentration were obtained from the intercepts of plots [E]/v vs. [Q] with the x axis.1,4-Benzoquinone-mediated reduction of cytochrome c by PETNR was monitored according to Δε 550 = 20 mM -1 cm -1 (concentrations of 1,4-benzoquinone and cytochrome c, 50 μM, concentration of NADPH, 150 μM).

RESULTS
Monitoring 1,4-benzoquinone-mediated reduction of cytochrome c by PETNR at pH 7.0, we found that the rate of cytochrome c reduction was equal to 5% of the NADPH enzymatic oxidation rate by benzoquinone, determined in the absence of cytochrome c.This shows that in the reduction of 1,4benzoquinone single-electron flux constitutes about 2.5% (the ratio of double cytochrome c reduction rate and NADPH oxidation rate), because 1,4-benzosemiquinone reduces cytochrome c rapidly (k ~ 10 6 M -1 s -1 ), whereas benzohydroquinone reduces it at a negligible rate at pH < 7.2 (Iyanagi, 1990, and refer-Two-electron reduction of quinones by Enterobacter cloacae PB2 ences therein).Thus, PETNR reduces quinones in a two-electron way to hydroquinones.
Like in our previous studies of two-electron transferring NQO1 and E. cloacae NR (Anusevičius et al., 2002;Nivinskas et al., 2002), one should consider two possible hydride transfer mechanisms from reduced PETNR (E-FMNH -) to quinones.The first is a three-step (e -,H + ,e -) mechanism: In this case, quinone reactivity should increase with an increase in their single-electron reduction potential (E 1 7 ) and pK a of semiquinone (pK a (QH .)) (Fukuzumi et al., 1989).Alternatively, in a singlestep transfer, quinone reactivity should increase with an increase in their hydride-transfer potential ((E 7 (H -)), i.e., the potential of the quinone/anionic hydroquinone (Q/ QH -) redox couple (Carlson & Miller, 1985).The latter is expressed as where E m,7 is the midpoint potential of the quinone/ hydroquinone couple at pH 7.0, and pK a (QH 2 ) is pK a of hydroquinone.If pK a (QH 2 ) ≤ 7.0, then E 7 (H -) ≅ E m,7 .We examined the reactivity of PETNR with structurally diverse quinoidal compounds possessing a broad range of E 1 7 , pK a (QH .), E 7 (H -), and van der Waals volumes (VdWvol) (Table 1).
In the reactions of PETNR with 5-hydroxy-1,4napthoquinone (Fig. 1A), 5,8-dihydroxy-1,4-naphthoquinone, 2,6-dimethyl-, trimethyl-and tetramethyl-1,4-benzoquinones (not shown), we obtained a series of parallel plots in Lineweaver-Burk coordinates at varied concentrations of quinones and several fixed NADPH concentrations, with an upward curvature at high quinone concentrations.This points to a 'ping-pong' scheme with quinone substrate inhibition.The quinone substrate inhibition constants (K iq ) increased with an increase in NADPH concentration (Fig. 1B).This shows that quinones bind to the oxidized enzyme form, thus acting as competitive inhibitor towards NADPH.The kinetic parameters of reaction, k cat and k cat /K m for quinone reduction, and K iq of quinones are given in Table 1.We were unable to determine k cat /K m of 1,4-benzoquinone, because its K m was below 1.0 μM (k cat ~ 25 s -1 at [NADPH] = 150 μM).The k cat /K m values for NADPH oxidation, determined using several quinone oxidants, did not depend on their nature, being in the range of 4.2 × 10 5 -2.9 × 10 5 M -1 s -1 .The k cat values for the most active quinones were close to 34 s -1 (Table 1), i.e., close to the maximal rate of the reduction of PETNR by NADPH, determined in the presteadystate kinetics (Basran et al., 2003).Thus, in these cases, the overall reaction rate becomes limited by the reductive half-reaction, and not by the oxidation of reduced PETNR by quinones.

The values of E 1
7 , E 7 (H -), and pK a (QH .) are taken from Wardman (1989) and Čėnas et al. (2004).a Kinetic parameters and K iq were determined at fixed NADPH concentration (150 μM); b k cat determined at saturating NADPH concentration, K is determined by extrapolation to [NADPH] = 0; c Assumed to be the same as for 1,4-naphthoquinone.

Two-electron reduction of quinones by Enterobacter cloacae PB2
giving an uncertain dependence of quinone reactivity on the proton-accepting potency of their radicals (Eqn.5).We failed to improve the regressions by the introduction of (VdWvol) 2 as an additional variable.
single-electron reduction potential of tetramethyl-1,4benzoquinone at pH 5.5-8.0 (pK a (QH .) = 4.6, Table 1), and the negative dependence of E 1 of riboflavin on pH with ∆E 1 /∆pH = -0.059V (pK a (QH .) = 8.8, Table 1).In the case of a three-step hydride transfer with the partly rate-limiting first electron transfer (Eqn.1), an increase in pH should relatively decrease the reactivity of riboflavin with respect to the reactivity of tetramethyl-1,4-benzoquinone.However, the log k cat / K m values of both compounds were pH-independent (Fig. 3B).This is consistent with a single-step hydride transfer (Eqn.2), because the pH-dependence of hydride-transfer potentials of both compounds is the same, ∆E(H -)/∆pH = -0.030V.

DISCUSSION
Our data show that the reactivity of quinones towards two-electron transferring PETNR is relatively insensitive to their structural peculiarities, and depends mainly on the energetics of charge transfer (Figs. 2A,3A,.In this aspect, PETNR differs from NQO1, whose reactivity is much more sensitive to the steric effects of quinones with the coefficients ∆log (k cat /K m )/∆VdWvol = -1.46-1.69× 10 -2 (Anusevičius et al., 2002;Čėnas et al., 2004), which distort the reactivity dependence on the quinone reduction potential.PETNR also differs from E. cloacae NR, which is relatively insensitive to the VdWvol of quinones, but shows a pronounced specificity towards 2-hydroxy-1,4-naphthoquinone structure (Nivinskas et al., 2002).A similar parabolic log k cat / K m vs. E 1 7 dependence for a limited number of quinones has been demonstrated in their two-electron reduction by Clostridium kluyveri diaphorase (Čėnas et al., 1987).Thus, our data provide further evidence that in certain cases the two-electron (hydride) reduction of quinones by flavoenzymes may not be significantly influenced by the specific interactions, similar to their single-electron reduction by flavoenzymes dehydrogenases-electrontransferases.Thus, the flavin semiquinone instability may be a sufficient factor for the two-electron (hydride) reduction of quinones.Although the reduction of quinones by PETNR is interesting mainly as a model reaction, it has been recently shown that a member of the OYE family is the main enzyme in Trypanosoma cruzi responsible for the NADPH-dependent bioreductive activation of trypanocidal quinones and aromatic nitrocompounds (Kubata et al., 2002).
It is of certain interest to compare the 'intrinsic' reactivities of two-electron transferring flavoenzymes, i.e., their k cat /K m of quinone reduction at ΔE 7 (H -) about 0. Because E. cloacae NR contains anionic FMNH -at pH 7.0, its E 7 (H -) = E m,7 = -0.190V (Koder et al., 2002).X-Ray studies also suggest the presence of anionic reduced flavins in NQO1 and PETNR (Li et al., 1995;Barna et al., 2001).Thus, their E 7 (H -) may be equal to or below their E m,7 (Eqn. 3, i.e., -0.159 V (NQO1, Tedeschi et al., 1995), and -0.267 V (PETNR, Khan et al., 2004).At ΔE 7 (H -) ~ 0, the quinone reduction rate constant by NQO1 may reach 10 7 M -1 s -1 (Anusevičius et al., 2002).In E. cloacae NR the rate constant varies between about 10 7 M -1 s -1 (2-hydroxy-1,4-naphthoquinone) and about 10 5 M - 1 s -1 (other compounds) (Nivinskas et al., 2002).In contrast, the 'intrinsic' reactivity of PETNR is lower, being close to 10 4 M -1 s -1 (Table 1, Fig. 3A).This may point to a less efficient electronic interaction between quinones and FMNH -isoalloxazine ring in PETNR.Indeed, the binding of quinones in the active center of NQO1 involves their π-π stacking with the flavin isoalloxazine ring, with the displacement of the aromatic residues of Tyr-128' and Phe-178', the latter stacking behind the quinone ring (Li et al., 1995).The binding of benzoate to E. cloacae NR involves some displacement of Phe-124 and Tyr-123, however, these residues do not stack with benzoate (Haynes et al., 2002).In contrast, the binding of steroid or nitroaromatic ligands to PETNR involves only small perturbations of Tyr-351, Tyr-186 and/or Trp-102 (Barna et al., 2001;Khan et al., 2002;2004), without any stacking interactions.These data enabled us to conclude that the active center of PETNR is flexible, and is able to accommodate large ligands without significant aminoacid displacements.This property is most probably responsible for the less efficient electronic interaction between quinones and FMNH - in PETNR, and, in parallel, for the absence of pronounced discrimination of bulky oxidants.

Figure 2 .
Figure 2. Dependence of the reactivity of PETNR on the electron-accepting potency and steric parameters of quinones.A. Dependence of k cat /K m of quinones on their single-electron reduction potentials (E 1 7 ).B. Dependence of k cat /K m of quinones on their Van der Waals volume (VdWvol).The numbers of compounds are taken from Table1.

Figure 3 .
Figure 3. Dependence of the reactivity of PETNR on the hydride-accepting potency of quinones and on pH. A. Dependence of k cat /K m of quinones on their hydride-transfer potential (E 7 (H -)).The numbers of compounds are taken from Table 1.B. pH dependence of k cat /K m of tetramethyl-1,4-benzoquinone (1) and riboflavin (2).