Archives of Biochemistry and Biophysics 547 (2014) 18–26

Contents lists available at ScienceDirect

Archives of Biochemistry and Biophysics journal homepage: www.elsevier.com/locate/yabbi

Proximal FAD histidine residue influences interflavin electron transfer in cytochrome P450 reductase and methionine synthase reductase Carla E. Meints, Sarah M. Parke, Kirsten R. Wolthers ⇑ Department of Chemistry, University of British Columbia Okanagan, 3333 University Way, Kelowna, BC V1V 1V7, Canada

a r t i c l e

i n f o

Article history: Received 15 January 2014 and in revised form 17 February 2014 Available online 28 February 2014 Keywords: Methionine synthase reductase Cytochrome P450 reductase Electron transfer Flavin Stopped-flow spectroscopy

a b s t r a c t Cytochrome P450 reductase (CPR) and methionine synthase reductase (MSR) transfer reducing equivalents from NADPH to FAD to FMN. In CPR, hydride transfer and interflavin electron transfer are kinetically coupled steps, but in MSR the two catalytic steps are represented by two distinct kinetic phases leading to transient formation of the FAD hydroquinone. In human CPR, His322 forms a hydrogen-bond with the highly conserved Asp677, a member of the catalytic triad. The catalytic triad is present in MSR, but Ala312 replaces the histidine residue. To examine if this structural variation accounts for differences in their kinetic behavior, reciprocal substitutions were created. Substitution of His322 for Ala in CPR does not affect the rate of NADPH hydride transfer or the FAD redox potentials, but does impede interflavin electron transfer. For MSR, swapping Ala312 for a histidine residue resulted in the kinetic coupling of hydride and interflavin electron transfer, and eliminated the formation of the FAD hydroquinone intermediate. For both enzymes, placement of the His residue in the active site weakens coenzyme binding affinity. The data suggest that the proximal FAD histidine residue accelerates proton-coupled electron transfer from FADH2 to the higher potential FMN; a mechanism for this catalytic role is discussed. Ó 2014 Elsevier Inc. All rights reserved.

Introduction Diflavin reductases are multi-domain enzymes that shuttle electrons derived from NADPH via two non-covalently bound flavin cofactors, FAD and FMN, to their cognate redox partner proteins. The prototypic member of this small enzyme family, cytochrome P450 reductase, is the redox partner for microsomal cytochrome P450 monooxygenases, which function in steroid, fatty acid and prostagladin metabolism as well as the detoxification of xenobiotic compounds [1]. A second eukaryotic family member, methionine synthase reductase, transfers an electron to the cob(II)alamin prosthetic group of methionine synthase, a process that leads to reductive remethylation of the vitamin B12 cofactor and restoration of methionine synthase activity [2,3]. Methionine synthase is vital for cell homeostasis as it generates tetrahydrofolate and methionine by transferring a methyl group from methyltetrahydrofolate to homocysteine [4]. Other members of the diflavin reductase family include novel oxidoreductase 1, and the reductase domains of nitric oxide synthase and P450 BM3, and the alpha subunit of bacterial sulfite reductase [5].

⇑ Corresponding author. Fax: +1 250 807 8009. E-mail address: [email protected] (K.R. Wolthers). http://dx.doi.org/10.1016/j.abb.2014.02.009 0003-9861/Ó 2014 Elsevier Inc. All rights reserved.

The diflavin reductase protein scaffold comprises three domains: an NADP(H) and FAD binding domain that is homologous to ferredoxin NADP+-oxidoreductase (FNR),1 an FMN-containing flavodoxin (Fld)-like domain, and a connecting domain that tethers the flavin domains and positions the cofactors such that their dimethylbenzyl moieties are 520 nm in the stopped-flow multi-wavelength data (Fig. 3A and B) also reveals that NADP+-FADH2 charge-transfer (CT) complex does not accumulate following hydride transfer. This is noteworthy, as the CT band has been observed in analogous stopped-flow studies with the Trp677 variants, which are also defective in interflavin electron transfer [27]. Thus, the perpetual stacking of the nicotinamide and FAD rings does not appear to be the cause for the attenuated FAD to FMN electron transfer in H322A. The absorbance changes and associated rate constants are also observed in single wavelength stopped-flow traces at 454 and 600 nm in Fig. 4A and B. Fits of the absorbance traces at 454 nm to a single (H322A) or double (H322Q) exponential equation gen-

erated rate constants that were comparable to that obtained from global analysis of the PDA data. Fig. 4B clearly shows that the 600 nm signal which typically forms in the first kinetic phase, is diminished in the H322Q variant and is barely detectable in H322A. Over a longer time course (i.e. 150 s), there is a gradual build-up of the disemiquinone (noted by an absorbance increase at 600 nm). The disproportionation of electrons between the two flavins in this final slow kinetic phase is the non-physiological thermodynamic equilibration of the enzyme/coenzyme mixture, as previously noted for the native enzyme [14]. The combined stopped-flow data suggest that His322 promotes the flux of electrons from FADH2 to the higher potential FMN cofactor. Pre-steady state analysis of the isolated FAD domain We compared the pre-steady state rate of NADPH reduction of the isolated NADPH/FAD-domain with the H322A substitution (FADH322A) with the native form of the truncated protein (FADwt) to determine if His322 affects the rate of hydride transfer. Fig. 5 shows single wavelength traces at 454 nm following the rapid mixing of 10-fold excess of NADPH against FADwt and FADH322A. The traces were fit to a double exponential equation generating a kobs1 = 30 s1 and kobs2 = 2 s1 for both enzyme forms. The identical rate constants suggest that removal of the imidazole side chain does not influence the rate of NADPH-hydride transfer. Redox potentiometry Anaerobic spectroelectrochemical titrations of the FADH322A variant were measured to determine if removal of the imidazole side chain influenced the midpoint potentials of the FAD cofactor. Previous studies with native CPR showed that the redox potentials

C.E. Meints et al. / Archives of Biochemistry and Biophysics 547 (2014) 18–26

23

A

B Fig. 5. Single-wavelength stopped-flow traces following NADPH reduction of isolated FAD domain CPR: FADWT (black) and FADH322A (grey). The reactions were performed in 50 mM Tris–HCl, pH 7.5 at 25 °C with 20 lM enzyme and 200 lM NADPH. The data were fit to a double exponential giving a kobs1 of 31.9 ± 0.1 s1 and 31.6 ± 0.3 s1 and kobs2 of 1.0 ± 0.1 s1 and 1.2 ± 0.1 s1 for FADWT and FADH322A, respectively.

Stopped-flow analysis of MSR flavin reduction

C

Fig. 4. Single wavelength stopped-flow traces following NADPH reduction of CPR (a) and the H322Q (b) and H322A (c) variants. The reactions were performed in 50 mM Tris–HCl, pH 7.5 at 25 °C under pseudo-first order conditions with the enzyme concentration at 20 lM and the coenzyme concentration at 200 lM. Panel A shows an average of 4 absorbance traces at the flavin absorbance maxima (454 nm) following rapid mixing of the enzyme with NADPH. Soluble CPR and the H322Q variant were fit to a double exponential equation from 0.001 to 1 s, giving rate constants of 15 and 3 s1 (sCPR) and 17 and 6 s1 (H322Q). The H322A variant was fit to a single exponential equation, giving an observed rate constant of 8 s1. Panels B and C show the relative absorbance changes at 600 nm over short (B) and extended (C) time frames during the course of the reductive half reaction.

of the FAD couples were similar in the full-length enzyme and the isolated NADPH/FAD domain [30]. Therefore, to simplify the fitting process, the redox potentials of FADH322A were measured. Fig. 6A reveals that reduction of FADH322A by sodium dithionite under anaerobic conditions does lead to the formation of the FAD semiquinone. The absorbance values at 600 nm plotted against the potential values normalized to the standard hydrogen electrode clearly show the appearance and disappearance of the FAD semiquinone during the redox titration. The data in Fig. 6B were fit to Eq. (1) which gave midpoint potentials of 273 ± 6 mV for FADox/sq and 398 ± 20 mV for FADsq/hq. These values are similar to that obtained for the native FAD domain indicating that the His322 imidazole side chain does not affect the midpoint potentials of the FAD cofactor [30].

Like wild-type MSR, the photodiode array derived spectra of A312Q were fitted to a three-step kinetic model (Fig. 7A and C) [17]. The first kinetic phase showed bleaching of the flavin absorption maxima at 454 nm without the formation of a broad absorbance band at 600 nm. This step represents the formation of the two-electron reduced enzyme species (E-FADH2-FMN) upon oxidation of the first equivalent of NADPH. A312Q exhibited a modest 1.9-fold reduction in kobs1, with a similar amplitude change as that of the native enzyme. The second slower kinetic phase, with a rate constant of 2 s1 is associated with FAD to FMN electron transfer, denoted by a build-up of the disemiquinone signal at 600 nm and loss of absorbance at 454 nm. The final kinetic phase follows the disappearance of absorbance peaks at 454 and 600 nm. This step represents the conversion of two-electron reduced species to the full four-electron reduced species by a second equivalent of NADPH. In contrast, NADPH reduction of A312H resembles that of native CPR, in that it occurs in two discrete kinetic phases, without the transient formation of the E-FADH2-FMN intermediate (Fig. 7B and D). Like CPR, the initial kinetic event is associated with a large amplitude change at the flavin absorbance maxima that is accompanied by a smaller absorbance increase at 600 nm. These simultaneous absorbance changes indicate that hydride and interflavin electron transfer are coupled events in the A312H variant. However, the rate constant for this step (0.77 s1) is 30-fold slower than that of wild-type CPR. With the A312H mutant, we measured the rate of flavin reduction with (R)-[4-2H]-NADPH to determine if the slower rate of flavin reduction is attributed to a slower rate of hydride transfer. The initial rate of flavin reduction (kobs1) was 0.54 s1 with (R)-[4-2H]-NADPH giving a KIE of 1.3, slightly smaller than the KIE of 1.7 obtained for the native enzyme [17]. Thus, it appears from these data that hydride transfer is less rate-determining in H322A compared to the native enzyme. Single-wavelength absorbance studies were conducted at 454 and 600 nm to measure the rate constants associated with absorbance changes for A312Q and A312H (Fig. 8). At 454 nm, flavin bleaching of A312Q and A312H were best fitted to a single and double exponential equation, respectively, which yielded a kobs1 and kobs2 that were comparable to the rate constants determined from multi-wavelength analyses. Both variants displayed an ‘‘updown’’ optical signal at 600 nm that reached a maxima at 5–10 s post mixing and then decayed over 100 s. The traces, which likely correspond to transient formation of the disemiquinone, could not

24

C.E. Meints et al. / Archives of Biochemistry and Biophysics 547 (2014) 18–26

Fig. 6. Redox titration of FADH322A. Panel A: Spectral features of FADH322A over the redox titration. Spectra were recorded after each addition of dithionite to FADH322A once the potential readings equilibrated. Panel B: Absorbance at 600 nm versus the normalized redox potential. These data were fitted to Eq. (1) giving midpoint potential values of 273 ± 6 mV for FADox/sq and 398 ± 20 mV for FADsq/hq.

A

B

C

D

Fig. 7. Anaerobic multi-wavelength stopped-flow analysis of the MSR A312Q and A312H variants. The photodiode array spectra of 20 lM of MSR A312Q (A) and A312H (B) after rapid mixing with 200 lM NADPH in 50 mM Tris–HCl, pH 7.5 at 25 °C. The flavin absorbance spectra, acquired for 150 s after the mixing event, were analyzed by singular value decomposition and global analysis. The A312Q and A312H spectral data were fit to a three-step and a two-step model, respectively, generating spectral profiles for the oxidized (solid line), partially reduced forms of the enzyme (dashed, dotted, and dashed/dotted lines). The observed rate constants are summarized in Table 2.

be reliably fit to a double exponential equation. Nevertheless, it is worth noting that the disemiquinone species in A312H and A312Q occurred at an earlier time point compared to the native enzyme, suggesting that these enzymes are more efficient at FADH2 to FMN interflavin electron transfer.

Discussion In this paper, we sought to explain why interflavin electron transfer is efficient in CPR and not in MSR. We focused on the structural differences surrounding the FAD isoalloxazine ring, and in particular the catalytic triad. While the arrangement of the conserved Asp/Glu, Ser and Cys is known to be critical for catalysis, the precise roles of these residues in mediating electron flux through the difla-

vin reductases have yet to be fully established. The hydroxyl group of Ser is within van der Waals contact of the FAD N5 (3.26 Å), but structural analysis of reduced spinach FNR reveals that the hydroxyl group moves 0.2 Å closer to the reduced FADH2, bringing it within hydrogen bonding distance of the N5 [33]. The N5 atom accepts a proton from the C4 of the nicotinamide ring during the hydride transfer process forming the FAD hydroquinone. Establishment of a hydrogen-bond to the Ser hydroxyl presumably favors the formation of the FAD hydroquinone. The next step of catalysis is single electron transfer from FADH2 to the FMN cofactor. For CPR, this step is kinetically coupled to hydride transfer, such that after the first fast phase of flavin reduction, 70% of the FAD is re-oxidized to the quinone, while the remaining 30% is in the semiquinone state. Reformation of the FAD quinone in this initial kinetic event requires rapid deprotonation of the N5 of the FAD,

C.E. Meints et al. / Archives of Biochemistry and Biophysics 547 (2014) 18–26

Fig. 8. Single wavelength absorbance traces for NADPH reduction of wild type MSR and the A312Q and A312H variants. The reactions were performed in 50 mM Tris– HCl, pH 7.5 at 25 °C under pseudo-first order conditions with the enzyme concentration at 20 lM and the coenzyme concentration at 200 lM. Absorbance traces at the flavin absorbance maxima after rapid mixing with NADPH (454 nm) are shown in Panel A. Spectral traces at 454 nm were fitted to a double exponential equation for A312Q yielding kobs1 of 12.8 ± 0.1 s1 and kobs2 of 1.11 ± 0.01 s1 and to a single exponential for A312H yielding a kobs1 of 0.77 ± 0.01 s1. Panel B represent the relative absorbance changes at 600 nm.

which in CPR is shielded from the solvent. Given its proximity to the Ser hydroxyl, the Glu/Asp carboxylate is a likely a proton acceptor to the N5 atom of the FAD via the Ser hydroxyl. This hypothesis was examined through mutagenesis of the glutamate side chain in FNR from spinach and Anabaena; however, no definitive conclusion could be reached [34,35]. Nevertheless, structural studies show that the Glu re-positions itself upon flavodoxin binding to form a hydrogen-bond with the conserved Ser [36], thereby increasing its proton donating capacity to the FAD N5. Computational studies also implicate the glutamate as a general acid in the proton-coupled electron transfer process to the N5 FAD [24]. The distance between the Asp/Glu carboxylate and Ser hydroxyl in CPR and FNR is 2.93 Å and 2.65 Å respectively, indicating that a moderate to strong hydrogen bond forms between these two residues [33,37]. In contrast, the electrostatic interaction between the corresponding residues in MSR (Asp695 and Ser454) is weaker as the distance between the donor and acceptor atoms is 3.40 Å [23]. Instead, a water molecule (water930) bridges the two residues forming hydrogen bonds with both side chains. Asp695 also forms a hydrogen bond with a second water molecule (water912), and the side chain is also within van der Waals contact (3.26 Å) of the indole –NH group of Trp697, which is in a different orientation than that observed in CPR. In general, Asp695 is more solvent exposed in MSR, compared to Asp677 of CPR, which forms hydrogen-bond contact with Ser460, His322, Cys632, and water692.

25

The results from the steady-state kinetic data indicate that the presence of a histidine in the active site weakens the binding affinity for NADPH. The crystal structure of the W677X mutant of CPR provides structural insight for how the imidazole side chain may weaken NADP+ binding. Removal of the last two C-terminal residues (W677X) enabled the nicotinamide ring to align with a tilted geometry with respect to the FAD isoalloxazine ring. In this catalytically productive complex, the carboxylate side chain of Asp forms a strong hydrogen bond (2.5 Å) to the carboxamide group of the NADP(H) nicotinamide ring. Polar contact with the carboxamide of the coenzyme pulls the Asp677 carboxylate chain away from Ser460 and Cys632, severing the hydrogen bonds with the side chains. Perhaps, the imidazole side chain competes with the carboxamide group for the polar interaction with the Asp side chain. In doing so, this may potentially weaken NADP(H) binding, increasing the rate constant associated with the release of the cofactor. Hubbard et al. 2001 proposed that displacement of the oxidized nicotinamide promotes interflavin electron transfer and FAD semiquinone formation as it allows the hydrogen-bond network between the catalytic triad to be re-established [38]. As mentioned above, the Trp677 variants of CPR were also defective in interflavin electron transfer [29]. However, reduction of these mutants led to formation of a CT complex between FADH2 and the oxidized nicotinamide ring. Presumably, disruption of the hydrogen-bond network within the catalytic triad, namely between Asp677 and Ser460 enforced by the binding of the nicotinamide ring prevents the deprotonation of the N5 of the FAD hydroquinone. Pre-steady state analysis of CPR H322A showed that the imidazole side chain is also vital for interflavin electron transfer and reduction of the enzyme by a second equivalent of NADPH. However, unlike the Trp677 variants, there is no optical signal that is attributed to the CT complex. Thus, despite the fact that the coenzyme binds to H322A with a tighter binding affinity, the mutation does not lead to persistent stacking of the nicotinamide against the FAD. Therefore, formation of the CT complex does not account for the attenuated rate of interflavin electron transfer. The rate of NADPH reduction of the isolated and H322A variants of the NADPH/FAD domain are similar indicating that the imidazole side chain does not influence hydride transfer. The 2.5-fold reduction in the initial kinetic phase of flavin reduction of H322A in the fulllength enzyme is likely linked to disrupted electron transfer to the more electropositive FMN cofactor. Impeded interflavin electron transfer is further supported by a more pronounced affect on cytochrome c3+ reduction, which is dependent on electron delivery from FMN compared to FeCN, which can receive reducing equivalents from FAD or FMN. We observed under aerobic conditions, that the flavin semiquinone does form following a five-minute incubation in the presence of NADPH. Under these conditions, the flavin semiquinone likely forms by single electron transfer to O2 to form the superoxide radical. The mechanism by which His322 promotes interflavin electron transfer in CPR is unclear, but given that it is part of the hydrogen-bond network of the catalytic triad, we suggest that it enables the correct positioning of the Asp side chain such that it can form good hydrogen bond contacts with the Ser. It is also possible that non-covalent interaction between His322 and Asp677 affects the pKa of the latter residue and its ability to act as a general base. There is the possibility that His322 is also acting as a second general base in concert with the Asp during the deprotonation of FAD, but given that the glutamine side chain can partially substitute for the imidazole in mediating electron flux, this possibility seems unlikely. Attempting to mimic hydrogen-bond network surrounding the conserved Asp in MSR by substituting the Ala312 for a His or a Gln, led to differential effects on catalysis. In general, the A312Q variant was similar to native MSR, with the exception of slightly

26

C.E. Meints et al. / Archives of Biochemistry and Biophysics 547 (2014) 18–26

weaker coenzyme binding and the earlier appearance of the flavin disemiquinone intermediate following NADPH-dependent reduction of the fully oxidized enzyme. Faster disemiquinone formation was also observed for the A312H, but in the case of this mutant, the rate of flavin reduction decreased so dramatically that hydride and interflavin electron transfer occur within the same kinetic phase, such that there is no longer the accumulation of the E-FADH2FMN intermediate. The earlier appearance of the disemiquinone in both MSR variants suggest that the more solvent exposed Asp695 the native enzyme is less efficient at deprotonation of the N5 of the FAD. In summary, we have shown that His322 in CPR is key for rapid intramolecular electron transfer from FAD to FMN. We propose based on its proximity to the catalytic triad and to FAD, that the histidine optimally positions the Asp side chain so that is can form a strong hydrogen-bonding network to promote deprotonation of the FAD hydroquinone. Through these studies, the role of a previously overlooked amino acid residue neighbouring the FAD cofactor has provided more mechanistic insight into the function of key active site residues in mediating electron flux in diflavin reductases. Acknowledgment This work is supported by a Grant from the Natural Sciences and Engineering Research Council of Canada. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.abb.2014.02.009. References [1] F.P. Guengerich, Human Cytochrome P450 Enzymes, Plenum, New York, 2005. [2] D. Leclerc, A. Wilson, R. Dumas, C. Gafuik, D. Song, D. Watkins, H.H. Heng, J.M. Rommens, S.W. Scherer, D.S. Rosenblatt, R.A. Gravel, Proc. Natl. Acad. Sci. U.S.A. 95 (1998) 3059–3064. [3] H. Olteanu, R. Banerjee, J. Biol. Chem. 276 (2001) 35558–35563. [4] R.G. Matthews, Acc. Chem. Res. 34 (2001) 681–689. [5] M.B. Murataliev, R. Feyereisen, F.A. Walker, Biochim. Biophys. Acta 1698 (2004) 1–26. [6] M. Wang, D.L. Roberts, R. Paschke, T.M. Shea, B.S. Masters, J.J. Kim, Proc. Natl. Acad. Sci. U.S.A. 94 (1997) 8411–8416. [7] T. Iyanagi, N. Makino, H.S. Mason, Biochemistry 13 (1974) 1701–1710.

[8] J. Ellis, A. Gutierrez, I.L. Barsukov, W.C. Huang, J.G. Grossmann, G.C. Roberts, J. Biol. Chem. 284 (2009) 36628–36637. [9] W.C. Huang, J. Ellis, P.C. Moody, E.L. Raven, G.C Roberts, Structure 21 (2013) 1581–1589. [10] M. Jenner, J. Ellis, W.C. Huang, E. Lloyd Raven, G.C. Roberts, N.J. Oldham, Angew. Chem. Int. Ed. Engl. 50 (2011) 8291–8294. [11] C.R. Pudney, B. Khara, L.O. Johannissen, N.S. Scrutton, PLoS Biol. 9 (2011) e1001222. [12] S. Hay, S. Brenner, B. Khara, A.M. Quinn, S.E. Rigby, N.S. Scrutton, J. Am. Chem. Soc. 132 (2010) 9738–9745. [13] D. Hamdane, C. Xia, S.C. Im, H. Zhang, J.J. Kim, L. Waskell, J. Biol. Chem. 284 (2009) 11374–11384. [14] D.D. Oprian, M.J. Coon, J. Biol. Chem. 257 (1982) 8935–8944. [15] A. Gutierrez, L.Y. Lian, C.R. Wolf, N.S. Scrutton, G.C. Roberts, Biochemistry 40 (2001) 1964–1975. [16] S. Brenner, S. Hay, A.W. Munro, N.S. Scrutton, FEBS J. 275 (2008) 4540–4557. [17] K.R. Wolthers, N.S. Scrutton, Biochemistry 43 (2004) 490–500. [18] A.L. Shen, C.B. Kasper, Biochemistry 35 (1996) 9451–9459. [19] A.L. Shen, M.J. Christensen, C.B. Kasper, J. Biol. Chem. 266 (1991) 19976– 19980. [20] A.L. Shen, D.S. Sem, C.B. Kasper, J. Biol. Chem. 274 (1999) 5391–5398. [21] Z. Deng, A. Aliverti, G. Zanetti, A.K. Arakaki, J. Ottado, E.G. Orellano, N.B. Calcaterra, E.A. Ceccarelli, N. Carrillo, P.A. Karplus, Nat. Struct. Biol. 6 (1999) 847–853. [22] C. Xia, D. Hamdane, A.L. Shen, V. Choi, C.B. Kasper, N.M. Pearl, H. Zhang, S.C. Im, L. Waskell, J.J. Kim, J. Biol. Chem. 286 (2011) 16246–16260. [23] K.R. Wolthers, X.D. Lou, H.S. Toogood, D. Leys, N.S. Scrutton, Biochemistry 46 (2007) (1844) 11833–11844. [24] V.I. Dumit, T. Essigke, N. Cortez, G.M. Ullmann, J. Mol. Biol. 397 (2010) 814– 825. [25] R.E. Viola, P.F. Cook, W.W. Cleland, Anal. Biochem. 96 (1979) 334–340. [26] K.R. Wolthers, J. Basran, A.W. Munro, N.S. Scrutton, Biochemistry 42 (2003) 3911–3920. [27] C.E. Meints, S. Simtchouk, K.R. Wolthers, FEBS J. 280 (2013) 1460–1474. [28] A. Grunau, M.J. Paine, J.E. Ladbury, A. Gutierrez, Biochemistry 45 (2006) 1421– 1434. [29] C.E. Meints, F.S. Gustafsson, N.S. Scrutton, K.R. Wolthers, Biochemistry 50 (2011) 11131–11142. [30] A.W. Munro, M.A. Noble, L. Robledo, S.N. Daff, S.K. Chapman, Biochemistry 40 (2001) 1956–1963. [31] P.L. Dutton, Methods Enzymol. 54 (1978) 411–435. [32] S.N. Daff, S.K. Chapman, K.L. Turner, R.A. Holt, S. Govindaraj, T.L. Poulos, A.W. Munro, Biochemistry 36 (1997) 13816–13823. [33] C.M. Bruns, P.A. Karplus, J. Mol. Biol. 247 (1995) 125–145. [34] A. Aliverti, Z. Deng, D. Ravasi, L. Piubelli, P.A. Karplus, G. Zanetti, J. Biol. Chem. 273 (1998) 34008–34015. [35] M. Medina, M. Martinez-Julvez, J.K. Hurley, G. Tollin, C. Gomez-Moreno, Biochemistry 37 (1998) 2715–2728. [36] G. Kurisu, M. Kusunoki, E. Katoh, T. Yamazaki, K. Teshima, Y. Onda, Y. KimataAriga, T. Hase, Nat. Struct. Biol. 8 (2001) 117–121. [37] C. Xia, S.P. Panda, C.C. Marohnic, P. Martasek, B.S. Masters, J.J. Kim, Proc. Natl. Acad. Sci. U.S.A. 108 (2011) 13486–13491. [38] P.A. Hubbard, A.L. Shen, R. Paschke, C.B. Kasper, J.J. Kim, J. Biol. Chem. 276 (2001) 29163–29170.