西川 幸志

We report the first neutron structure of [NiFe]-hydrogenase in its oxidized state. This study leads to new insights into the oxidized active site and visualization of the protons characteristic of the oxidized enzyme.

Hydrogenases catalyze the reversible oxidation of H2. Carbon monoxide (CO) is known to be a competitive inhibitor of O2-sensitive [NiFe]-hydrogenases. Although the activities of some O2-tolerant [NiFe]-hydrogenases are unaffected by CO, the partially O2-tolerant [NiFe]-hydrogenase from Citrobacter sp. S-77 (S77-HYB) is inhibited by CO. In this work, the CO-bound state of S77-HYB was characterized by activity assays, spectroscopic techniques and X-ray crystallography. Electron paramagnetic resonance spectroscopy showed a diamagnetic Ni2+ state, and Fourier-transform infrared spectroscopy revealed the stretching vibration of the exogenous CO ligand. The crystal structure determined at 1.77 Å resolution revealed that CO binds weakly to the nickel ion in the Ni-Fe active site of S77-HYB. These results suggest a positive correlation between O2 and CO tolerance in [NiFe]-hydrogenases.

X-ray crystallography is the most powerful tool for obtaining structural information about protein molecules, affording accurate and precise positions for all of the atoms in the protein except for hydrogen. However, hydrogen species play crucial roles in the physiological functions of enzymes, including molecular recognition through hydrogen bonding and catalytic reactions involving proton transfer. Neutron crystallography enables direct identification of the positions of hydrogen species. [NiFe]-hydrogenase from Desulfovibrio vulgaris Miyazaki F is an enzyme that catalyzes the reversible oxidation of molecular hydrogen. It contains a bimetallic Ni-Fe active site for the catalytic reaction and three Fe-S clusters for electron transfer. Previous X-ray structure analyses of the enzyme under various oxidation conditions have revealed that the active site changes its coordination structure depending on the redox state. In the inactive air-oxidized form, an oxygen species was identified between the Ni and Fe atoms, whereas in the active H-2-reduced form, subatomic-resolution X-ray structure analysis and single-crystal EPR analyses indicated a hydride ligand between the two metal atoms. However, the assignment of the hydride moiety by X-ray crystallography remains controversial, and the proton transfer pathways in the molecule are still ambiguous. To allow neutron diffraction experiments, large crystals of [NiFe]-hydrogenase were prepared by the vapor diffusion method with the macroseeding technique according to the two-dimensional phase diagram (protein concentration vs. precipitant concentration). Neutron diffraction data were collected at approximately 2.0 angstrom resolution at cryogenic temperature using a gas-stream cooling system to trap short-lived intermediates in the catalytic reaction.

A membrane-bound hydrogenase from Desulfovibrio vulgaris Miyazaki F is a metalloenzyme that contains a binuclear Ni-Fe complex in its active site and mainly catalyzes the oxidation of molecular hydrogen to generate a proton gradient in the bacterium. The active-site Ni-Fe complex of the aerobically purified enzyme shows its inactive oxidized form, which can be reactivated through reduction by hydrogen. Here, in order to understand how the oxidized form is reactivated by hydrogen and further to directly evaluate the bridging of a hydride ligand in the reduced form of the Ni-Fe complex, a neutron structure determination was undertaken on single crystals grown in a hydrogen atmosphere. Cryogenic crystallography is being introduced into the neutron diffraction research field as it enables the trapping of short-lived intermediates and the collection of diffraction data to higher resolution. To optimize the cooling of large crystals under anaerobic conditions, the effects on crystal quality were evaluated by X-rays using two typical methods, the use of a cold nitrogen-gas stream and plunge-cooling into liquid nitrogen, and the former was found to be more effective in cooling the crystals uniformly than the latter. Neutron diffraction data for the reactivated enzyme were collected at the Japan Photon Accelerator Research Complex under cryogenic conditions, where the crystal diffracted to a resolution of 2.0 Å. A neutron diffraction experiment on the reduced form was carried out at Oak Ridge National Laboratory under cryogenic conditions and showed diffraction peaks to a resolution of 2.4 Å.

Hydrogenases control the proton concentration in cells, which is an essential function for hydrogen metabolism in several microorganisms. Some [NiFe]-hydrogenases are catalytically active under air and are thus of great interest for developing bio-inspired synthetic models and new devices for clean energy conversion. Here, we provide an overview of the structural basis of the reaction mechanism of [NiFe]-hydrogenases, and the recent development of a new assay method which may uncover hidden properties of hydrogenases.

A [NiFe] hydrogenase (H2 ase) is a proton-coupled electron transfer enzyme that catalyses reversible H2 oxidation; however, its fundamental proton transfer pathway remains unknown. Herein, we observed the protonation of Cys546-SH and Glu34-COOH near the Ni-Fe site with high-sensitivity infrared difference spectra by utilizing Ni-C-to-Ni-L and Ni-C-to-Ni-SIa photoconversions. Protonated Cys546-SH in the Ni-L state was verified by the observed SH stretching frequency (2505 cm-1 ), whereas Cys546 was deprotonated in the Ni-C and Ni-SIa states. Glu34-COOH was double H-bonded in the Ni-L state, as determined by the COOH stretching frequency (1700 cm-1 ), and single H-bonded in the Ni-C and Ni-SIa states. Additionally, a stretching mode of an ordered water molecule was observed in the Ni-L and Ni-C states. These results elucidate the organized proton transfer pathway during the catalytic reaction of a [NiFe] H2 ase, which is regulated by the H-bond network of Cys546, Glu34, and an ordered water molecule.

Enzyme activity is typically assayed by quantitatively measuring the initial and final concentrations of the substrates and/or products over a defined time period. For enzymatic reactions involving gaseous substrates, the substrate concentrations can be estimated either directly by gas chromatography or mass spectrometry, or indirectly by absorption spectroscopy, if the catalytic reactions involve electron transfer with electron mediators that exhibit redox-dependent spectral changes. We have developed a new assay system for measuring the time course of enzymatic reactions involving gaseous substrates based on Raman spectroscopy. This system permits continuous monitoring of the gas composition in the reaction cuvette in a non-invasive manner over a prolonged time period. We have applied this system to the kinetic study of the [NiFe] hydrogenase from Desulfovibrio vulgaris Miyazaki F. This enzyme physiologically catalyzes the reversible oxidation of H2 and also possesses the nonphysiological functions of H/D exchange and nuclear spin isomer conversion reactions. The proposed system has the additional advantage of enabling us to measure all of the hydrogenase-mediated reactions simultaneously. Using the proposed system, we confirmed that H2 (the fully exchanged product) is concomitantly produced alongside HD by the H/D exchange reaction in the D2 /H2 O system. Based on a kinetic model, the ratio of the rate constants of the H/D exchange reaction (k) at the active site and product release rate (kout ) was estimated to be 1.9 ± 0.2. The proposed assay method based on Raman spectroscopy can be applied to the investigation of other enzymes involving gaseous substrates.

Bilirubin oxidase (BOD) belongs to the family of blue multicopper oxidases, and catalyzes the concomitant oxidation of bilirubin to biliverdin and the reduction of molecular oxygen to water via a four-electron reduction system. The active sites of BOD comprise four copper atoms; type I copper (T1Cu) forms a mononuclear site, and a cluster of three copper atoms forms a trinuclear center. In the present study, we determined the high-resolution crystal structures of BOD from the fungus Myrothecium verrucaria. We investigated wild-type (WT) BOD and a BOD mutant called Met467Gln, which is inactive against bilirubin. The structures revealed that a novel post-translational crosslink between Trp396 and His398 is formed in the vicinity of the T1Cu site in WT BOD, whereas it is absent in the Met467Gln mutant. Our structural and computational studies suggest that the His-Trp crosslink adjusts the redox potential of T1Cu to that of bilirubin to efficiently abstract electrons from the substrate.

Citrobacter sp. S-77 [NiFe]-hydrogenase harbors a standard [4Fe-4S] cluster proximal to the Ni-Fe active site. The presence of relocatable water molecules and a flexible aspartate enables the [4Fe-4S] to display redox-dependent conformational changes. These structural features are proposed to be the key aspects that protect the active site from O2 attack.

Standard [NiFe]-hydrogenase from Desulfovibrio vulgaris Miyazaki F (DvMF-H2ase) catalyzes the uptake and production of hydrogen (H2) and is a promising biocatalyst for future energy devices. However, DvMF-H2ase experiences oxidative inactivation under oxidative stress to generate Ni-A and Ni-B states. It takes a long time to reactivate the Ni-A state by chemical reduction, whereas the Ni-B state is quickly reactivated under reducing conditions. Oxidative inhibition limits the application of DvMF-H2ase in practical devices. In this research, we constructed a mediated-electron-transfer system by co-immobilizing DvMF-H2ase and a viologen redox polymer (VP) on electrodes. The system can avoid oxidative inactivation into the Ni-B state at high electrode potentials and rapidly reactivate the Ni-A state by electrochemical reduction of VP. H2 oxidation and H+ reduction were realized by adjusting the pH from a thermodynamic viewpoint. Using carbon felt as a working-electrode material, high current densities-up to (200 ± 70) and -(100 ± 9) mA cm-3 for the H2-oxidation and H+-reduction reactions, respectively-were attained.

Hydrogenase is a key enzyme for a coming hydrogen energy society, because it has strong catalytic activities on both uptake and production of dihydrogen. We, however, have to overcome the sensitivity against O2 of the enzyme, because hydrogenase is, generally, easily inactivated in the presence of O2. In this study, we have revisited the crystal structures of [NiFe]‑hydrogenase from sulfate-reducing bacterium in the several oxidized and reduced conditions. Our results revealed that the Ni-Fe active site of the enzyme exposed into O2 showed two forms, Form-1 and Form-2. The Ni-Fe active site in Form-1 showed the typical Ni-B (inactive ready) structure, whereas those in Form-2 lost Ni with no relation to an exposure time to O2, and two cysteinyl sulfur ligands made a disulfide bond. On the other hand, the formation of sulfenylation of the cysteinyl ligand to Ni, which is often observed in the oxidized form, did not correlate with the Ni-elimination, but with exposure time to O2.

Previously, the Ni-SIr state of [NiFe] hydrogenase was found to convert to the Ni-SIa state by light irradiation. Herein, large activation energies and a large kinetic isotope effect were obtained for the reconversion of the Ni-SIa state to the Ni-SIr state after the Ni-SIr-to-Ni-SIa photoactivation, suggesting that the Ni-SIa state reacts with H2O and leaves a bridging hydroxo ligand for the Ni-SIr state.

Electrostatic interactions between proteins are key factors that govern the association and reaction rate. We spectroscopically determine the second-order reaction rate constant (k) of electron transfer from [NiFe] hydrogenase (H2ase) to cytochrome (cyt) c3 at various ionic strengths (I). The k value decreases with I. To analyze the results, we develop a semi-analytical formula for I dependence of k based on the assumptions that molecules are spherical and the reaction proceeds via a transition state. Fitting of the formula to the experimental data reveals that the interaction occurs in limited regions with opposite charges and with radii much smaller than those estimated from crystal structures. This suggests that local charges in H2ase and cyt c3 play important roles in the reaction. Although the crystallographic data indicate a positive electrostatic potential over almost the entire surface of the proteins, there exists a small region with negative potential on H2ase at which the electron transfer from H2ase to cyt c3 may occur. This local negative potential region is identical to the hypothetical interaction sphere predicted by the analysis. Furthermore, I dependence of k is predicted by the Adaptive Poisson-Boltzmann Solver considering all charges of the amino acids in the proteins and the configuration of H2ase/cyt c3 complex. The calculation reproduces the experimental results except at extremely low I. These results indicate that the stabilization derived from the local electrostatic interaction in the H2ase/cyt c3 complex overcomes the destabilization derived from the electrostatic repulsion of the overall positive charge of both proteins.

X-ray structure analysis method has been widely utilized to solve the structure of biological macromolecules. It is, however, also well known that the hydrogen atoms and protons, which are important to understand the structure-function relationship of them, are hardly visible in the electron density map due to the small atomic scattering factor of hydrogen. To visualize hydrogen species, neutron structure analysis is the most powerful technique, because the nuclear scattering length of hydrogen is comparable to those of other atoms comprising the protein molecules. Though many x-ray crystal structures of [NiFe] hydrogenases were reported so far, the real proton pathway and reaction mechanism at the Ni-Fe active site are still unclear. In this paper, we report how to prepare the high-quality large single crystals by using an improved crystallization phase diagram for [NiFe] hydrogenase.

A membraneless direct electron transfer (DET)-type dihydrogen (H-2)/air-breathing biofuel cell without any mediator was constructed wherein bilirubin oxidase from Myrothecium verrucaria (BOD) and membrane-bound [NiFe] hydrogenase from Desulfovibrio vulgaris Miyazaki F (MBH) were used as biocatalysts for the cathode and the anode, respectively, and Ketjen black-modified water proof carbon paper (KB/WPCC) was used as an electrode material. The KB/WPCC surface was modified with 2-aminobenzoic acid and p-phenylenediamine, respectively, to face the positively charged electron accepting site of BOD and the negatively charged electron-donating site of MBH to the electrode surface. A gas-diffusion system was employed for the electrodes to realize high-speed substrate supply. As result, great improvement in the current density of O2 reduction with BOD and H-2 reduction with MBH were realized at negatively and postively charged surfaces, respectively. Gas diffusion system also suppressed the oxidative inactivation of MBH at high electrode potentials. Finally, based on the improved bioanode and biocathode, a dual gas-diffusion membrane- and mediatorless H-2/air-breathing biofuel cell was constructed. The maximum power density reached 6.1 mW cm(-2) (at 0.72 V), and the open circuit voltage was 1.12 V using 1 atm of H-2 gas as a fuel at room temperature and under passive and quiescent conditions. (C) 2016 Elsevier B.V. All rights reserved.

The Ni-SIr state of [NiFe] hydrogenase from Desulfovibrio vulgaris Miyazaki F was photoactivated to its Ni-SIa state by Ar(+) laser irradiation at 514.5 nm, whereas the Ni-SL state was light induced from a newly identified state, which was less active than any other identified state and existed in the "as-isolated" enzyme.

H-2/O-2 biofuel cells utilizing hydrogenases and multicopper oxidases as bioelectrocatalysts are clean, sustainable, and environmentally friendly power devices. In this study, we constructed a novel gas diffusion bioelectrode with a sheet of waterproof carbon cloth as the electrode base and optimized the hydrophilicity/hydrophobicity of the electrode for both high gas permeability and high direct electron transfer bioelectrocatalytic activity. The electrode exhibited a large current density of about 10 mA cm(-2) in the steady-state for both H-2 oxidation and O-2 reduction. The biocathode and the bioanode were coupled to construct a gas diffusion H-2/O-2 biofuel cell. The dual gas diffusion system allowed the separate supply of gaseous substrates (H-2 and O-2) to the bioanode and biocathode, with consequent suppression of the oxidative inhibition of the hydrogenases. The cell exhibited a maximum power density of 8.4 mW cm(-2) at a cell voltage of 0.7 V under quiescent conditions.

The purification procedure of Hyd-2-type [NiFe]-hydrogenase from Citrobacter sp. S-77 was improved by applying treatment with trypsin before chromatography. Purified protein samples both with and without trypsin treatment were successfully crystallized using the sitting-drop vapour-diffusion method with polyethylene glycol as a precipitant. Both crystals belonged to space group P21, with unit-cell parameters a = 63.90, b = 118.89, c = 96.70 Å, β = 100.61° for the protein subjected to trypsin treatment and a = 65.38, b = 121.45, c = 98.63 Å, β = 102.29° for the sample not treated with trypsin. The crystal obtained from the trypsin-treated protein diffracted to 1.60 Å resolution, which is considerably better than the 2.00 Å resolution obtained without trypsin treatment. The [NiFe]-hydrogenase from Citrobacter sp. S-77 retained catalytic activity with some amount of O2, indicating that it has clear O2 tolerance.

Different light-induced Ni-L states of [NiFe] hydrogenase from its Ni-C state have previously been observed by EPR spectroscopy. Herein, we succeeded in detecting simultaneously two Ni-L states of [NiFe] hydrogenase from Desulfovibrio vulgaris Miyazaki F by FT-IR spectroscopy. A new light-induced νCO band at 1890 cm(-1) and νCN bands at 2034 and 2047 cm(-1) were detected in the FT-IR spectra of the H2-activated enzyme under N2 atmosphere at basic conditions, in addition to the 1910 cm(-1) νCO band and 2047 and 2061 cm(-1) νCN bands of the Ni-L2 state. The new bands were attributed to the Ni-L3 state by comparison of the FT-IR and EPR spectra. The νCO and νCN frequencies of the Ni-L3 state are the lowest frequencies observed among the corresponding frequencies of standard-type [NiFe] hydrogenases in various redox states. These results indicate that a residue, presumably Ni-coordinating Cys546, is protonated and deprotonated in the Ni-L2 and Ni-L3 states, respectively. Relatively small ΔH (6.4 ± 0.8 kJ mol(-1)) and ΔS (25.5 ± 10.3 J mol(-1) K(-1)) values were obtained for the conversion from the Ni-L2 to Ni-L3 state, which was in agreement with the previous proposals that deprotonation of Cys546 is important for the catalytic reaction of the enzyme.

The enzyme hydrogenase reversibly converts dihydrogen to protons and electrons at a metal catalyst. The location of the abundant hydrogens is of key importance for understanding structure and function of the protein. However, in protein X-ray crystallography the detection of hydrogen atoms is one of the major problems, since they display only weak contributions to diffraction and the quality of the single crystals is often insufficient to obtain sub-ångström resolution. Here we report the crystal structure of a standard [NiFe] hydrogenase (∼91.3 kDa molecular mass) at 0.89 Å resolution. The strictly anoxically isolated hydrogenase has been obtained in a specific spectroscopic state, the active reduced Ni-R (subform Ni-R1) state. The high resolution, proper refinement strategy and careful modelling allow the positioning of a large part of the hydrogen atoms in the structure. This has led to the direct detection of the products of the heterolytic splitting of dihydrogen into a hydride (H(-)) bridging the Ni and Fe and a proton (H(+)) attached to the sulphur of a cysteine ligand. The Ni-H(-) and Fe-H(-) bond lengths are 1.58 Å and 1.78Å, respectively. Furthermore, we can assign the Fe-CO and Fe-CN(-) ligands at the active site, and can obtain the hydrogen-bond networks and the preferred proton transfer pathway in the hydrogenase. Our results demonstrate the precise comprehensive information available from ultra-high-resolution structures of proteins as an alternative to neutron diffraction and other methods such as NMR structural analysis.

[NiFe] hydrogenase catalyzes the reversible cleavage of H2. The electrons produced by the H2 cleavage pass through three Fe-S clusters in [NiFe] hydrogenase to its redox partner. It has been reported that the Ni-SI(a), Ni-C, and Ni-R states of [NiFe] hydrogenase are involved in the catalytic cycle, although the mechanism and regulation of the transition between the Ni-C and Ni-SI(a) states remain unrevealed. In this study, the FT-IR spectra under light irradiation at 138-198 K show that the Ni-L state of [NiFe] hydrogenase is an intermediate between the transition of the Ni-C and Ni-SI(a) states. The transition of the Ni-C state to the Ni-SI(a) state occurred when the proximal [Fe4S4]p(2+/+) cluster was oxidized, but not when it was reduced. These results show that the catalytic cycle of [NiFe] hydrogenase is controlled by the redox state of its [Fe4S4]p(2+/+) cluster, which may function as a gate for the electron flow from the NiFe active site to the redox partner.

Nitrophorins are proteins occurring in the saliva of the blood-sucking insect Rhodnius prolixus to carry NO as a vasodilator and blood-coagulation inhibitor into the victim's tissue. It was suggested that the rate of NO release can be enhanced by the blood-plasma component L-cysteine [J.M.C.Ribeiro, Insect Biochem. Mol. Biol. 26 (1996) 899-905]. However, the mechanism of the reaction is not clear. In the attempt to exploit the reaction in detail, complexes of nitrophorin 4 (NP4) with the thiols 2-mercaptoethanol, L-cysteine, and L-homocysteine and with HS(-) were formed and characterized under anaerobic conditions using absorption spectroscopy, X-ray crystallography, and EPR spectroscopy. In contrast to met-myoglobin, which is reduced by L-cysteine, all four compounds form low-spin Fe(III) complexes with NP4. The weak equilibration constants (167-5200 M(-1)) neither support significant complexation nor the simple displacement of NO in vivo. Both amino acid based thiols form additional H-bonds with side chains of the heme pocket entry. Glutathione and L-methionine did not form a complex, indicating the specificity of the complexes with L-cysteine and L-homocysteine. Continuous wave EPR spectroscopy reveals the simultaneous existence of three low-spin systems in each case that are attributed to various protonation and/or conformational stages in the heme pocket. Electron nuclear double resonance (ENDOR) spectroscopy demonstrates that the thiol sulfurs are, at least in part, protonated. Overall, the results not only demonstrate the good accessibility of the NP4 heme center by biologically relevant thiols, but also represent the first structural characterization of a ferriheme protein in complex with L-cysteine L-homocysteine.

NADH:rubredoxin oxidoreductase (NROR), an O(2)-inducible protein, is a versatile electron donor for scavengers of O(2) and reactive oxygen species (ROS) in Clostridium acetobutylicum. Recombinant NROR was overexpressed in Escherichia coli and purified to homogeneity; it was subsequently crystallized using the sitting-drop vapour-diffusion method at 293 K. Preliminary crystallographic analysis revealed that the crystals belonged to space group P4(1)22 or P4(3)22, with unit-cell parameters a = b = 98.6, c = 88.3 A, and diffracted to 2.1 A resolution. Assuming that the crystals contained one molecule per asymmetric unit, the Matthews coefficient was calculated to be 2.7 A(3) Da(-1) and the solvent content to be 54.1%.