# 2020, Vol.49, No.2

164-173

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.

Hydrogenases catalyze the reversible oxidation of molecular hydrogen.17 According to the metal composition of the active site, they are classified into three major groups; 1) [NiFe]-hydrogenases, including the subgroup [NiFeSe]-hydrogenases,814 2) [FeFe]-hydrogenases, and 3) [Fe]-hydrogenases.15 [NiFe]-hydrogenases catalyze mainly the heterolytic cleavage of molecular hydrogen into protons and electrons. [FeFe]-hydrogenases are more active in H2 production.5 [Fe] hydrogenases transfer hydride to the acceptor methenyl-H4MPT+.16 Given their high activity under ambient conditions and the terrestrial abundancy of the transition metals incorporated into the metal centers, hydrogenases are of great interest as environmentally friendly energy conversion catalysts. Numerous model complexes have been synthesized and characterized.1720 For example, functional heterodinuclear (NiFe or NiRu) compounds21 that mimic [NiFe]-hydrogenases have been synthesized by Ogo et al., and mononuclear Ni complexes provide interesting insights into the hydrogen conversion mechanism.14,2230 Furthermore, enzymatic and bio-inspired fuel cells have been developed,3137 with biofuel cells combining photosystems, and hydrogenases being particularly intensively studied. Another application of hydrogenases is the use of bacterial cells as biomolecular catalysts for hydrogen production.38,39

[NiFe]-hydrogenases are the most divergently evolved enzymes among [NiFe]-, [FeFe]-, and [Fe]-hydrogenases in terms of higher specificity and unique or complex function. They are categorized into groups I–IV, and further divided into subgroups according to their function in cells.40 The heterodimeric unit (the large subunit which harbors the Ni-Fe active site) is believed to be essential for the enzyme activity of all [NiFe]-hydrogenases. Small subunits containing several [4Fe4S] or [3Fe4S] clusters involved in electron transfer are also critical (Figure 1A).41,42 Many [NiFe]-hydrogenases have more than three subunits, such as a membrane-bound cytochrome b-type subunit or a flavin-binding subunit that interacts with NADH/NADPH. Recently, a new type of [NiFe]-hydrogenase that is catalytically active at ambient O2 concentrations has been discovered in several aerobic microorganisms. [NiFe]-hydrogenases can thus be distinguished by their stability to O2 as O2-sensitive or O2-tolerant, with the former being inactivated quickly upon exposure to O2 but regaining activity following the removal of O2, and the latter maintaining activity. O2-sensitive two-subunit [NiFe]-hydrogenases are generally called ‘standard’ enzymes.

Decades of intensive study have elucidated the mechanism underlying H2 conversion in [NiFe]-hydrogenases. The active site is composed of a Ni atom and an Fe atom, with the Fe ligated to two CN and one CO. The Fe is in the low-spin Fe2+ state during catalysis. The Ni has four cysteine thiolate ligands, two of which bridge the two metal atoms. The third bridging position is occupied or unoccupied depending on the oxidation state (marked as X in Figure 1B). The small subunit typically contains three Fe-S clusters. Standard O2-sensitive [NiFe]-hydrogenases contain two [4Fe4S] clusters and one [3Fe4S] cluster, with the proximal [4Fe4S] cluster close to the Ni-Fe active site being ligated by four cysteine residues. The medial [3Fe4S] cluster is located in the middle of the [4Fe4S] clusters. The distal [4Fe4S] cluster located beneath the molecule’s surface is ligated by three cysteine residues and a histidine residue. Recent studies revealed that O2-tolerant [NiFe]-hydrogenases, such as membrane-bound [NiFe]-hydrogenases (MBH-type), possess an unprecedented proximal [4Fe3S] cluster ligated by six cysteine residues.4345 The remaining Fe-S clusters (i.e., one distal [4Fe4S] and one medial [3Fe4S] cluster) are coordinated similarly to those in O2-sensitive standard [NiFe]-hydrogenases.46 These Fe-S clusters are related to electron transfer, and these enzymes contain a sophisticated proton transfer pathway and a gas-access channel.47,48

Recent structural analysis of a NAD+-reducing soluble [NiFe]-hydrogenase in the oxidized state related to NADH:quinone oxidoreductase (complex I) revealed an unprecedented active site geometry.49 Furthermore, there have been many structural studies on different types of [NiFe]-hydrogenases (e.g., an electron-bifurcating [NiFe]-hydrogenase complexed with heterodisulfide reductase (HdrABC-MvhAGD)50), a Hyb-type [NiFe]-hydrogenase from Citrobacter sp. S-77,51 Hyd-2 from Escherichia coli,52 a 14-subunit MBH from the hyperthermophile Pyrococcus furiosus,53 and an actinobacterial-type [NiFe]-hydrogenase from Ralstonia eutropha.54

This short review presents an overview of our current understanding of the mechanism of [NiFe]-hydrogenases based on high-resolution X-ray structure analyses and spectroscopic studies, as well as theoretical investigations of proton and electron transfer pathways in the protein matrix. In the last section, we introduce a new assay method using Raman spectroscopy, which is a useful tool for both probing the metal centers5557 and analyzing the kinetic properties of [NiFe]-hydrogenases, especially H/D exchange and nuclear spin conversion reactions.58

The Ni-Fe active site passes through several oxidation states during the catalytic cycle. Figure 2 shows the four states of the catalytic cycle: Ni-SIa, Ni-R, Ni-C, and Ni-L. Hydrogen binding likely occurs in the Ni-SIa state, and theoretical calculations suggest that molecular hydrogen is attached to the Ni ion in the singlet state (Ni2+) in the Ni-SIa state.59 A complementary experiment showed that the competitive inhibitor carbon monoxide can bind to the Ni atom in O2-sensitive [NiFe] hydrogenases.6062 However, in heteronuclear model complexes the Fe ion is preferred as the initial hydrogen binding site.17,63 This discrepancy is likely due to differences in ligand coordination between the enzyme active site and the model complexes. The distance between the metals of the model complexes is much larger than that of the enzyme. In the model complexes the geometry of Ni is always square planar, while the enzymes vary during the catalytic cycle. The first experimentally determined structure was very recently reported for the Ni-SIa state and shows that the coordination position attacked by hydrogen molecules is vacant.62 H2 binds to the Ni-Fe active site and is split into a hydride and a proton, resulting in the Ni-R state. Between the Ni-SIa state and the Ni-R state, density functional theory (DFT) calculations predicted two intermediates: a H2-coordinated Ni2+ (the Ni-H2 state)59,64 and a bridging hydride with a second (additional) terminal hydride on Ni4+.65,66 Neutron diffraction studies are required to confirm the coordination of H2 bound in the Ni-Fe active site.

In its most reduced state (Ni-R) the hydride is attached between the Ni and Fe atoms and the proton is bound to the terminal cysteine thiolate. It is usually difficult to observe the electron density of light atoms such as hydrogen by X-ray diffraction methods. However, analysis of the subatomic (0.89 Å) resolution crystal structure revealed the precise geometry of the protein structure, and more than 90% of the amino acid residue hydrogen atoms near the active site were assigned in the electron density map.67 Using least-square full-matrix inversion, the standard deviations for the bond distances were estimated to be 0.002–0.007 Å for the Ni-Fe active site and 0.006 Å for the amino acid residues. Furthermore, the hydride bridged between the Ni and Fe atoms was located slightly closer to the Ni atom (Ni-H, 1.58 Å; Fe-H, 1.78 Å). The axial ligation geometry of the Ni2+ coordination in the Ni-R state indicated that the nickel-ion ligand sphere is essentially square pyramidal. Nuclear resonance vibrational spectroscopy (NRVS) together with DFT calculations confirmed that the Ni ion is indeed in the low spin state in the Ni-R state.68

The Ni-C state can be detectable by EPR spectroscopy due to the Ni3+ ion (S = 1/2). In this state, the hydride remains at the bridging position and the proton attached to the cysteine is released.69 In the next step, the bridging hydride is transferred to the terminal cysteine thiolate as a proton, resulting in the Ni-L state, where the Ni ion is reduced (Ni3+ to Ni+).70,71 The Ni-L state has been first identified as a light induced state at cryogenic temperatures.72 Several Ni-L states (Ni-L2, and L3) have been identified at different pH conditions.70

It has been suggested that the Ni-L state converts to the Ni-SIa state by one electron oxidation and H+ transfer.73 A recent spectroscopic study suggested that the Ni-L state can be observed at ambient temperature and could be involved in the catalytic cycle.74,75 However, theoretical calculations indicated several intermediate states during the catalytic cycle not involving the Ni-L state.65 Dyer et al. recently used nanosecond transient infrared spectroscopy, enabling rapid detection of proton-coupled electron transfer dynamics, and reported that the Ni-SIa to Ni-C transition involves three new intermediate states (electron-proton transfer, hydrogen bond cleavage of glutamate, and hydride formation).76,77

Most [NiFe]-hydrogenases are inactivated upon exposure to O2. There are two highly oxidized states (Ni-A and Ni-B) and each has a one-electron reduced intermediate state (Ni-SU and Ni-SIr, respectively).13 Ni-A requires long activation time (up to hours), while Ni-B is activated within a few minutes under the H2 conditions. Single-crystal ENDOR spectroscopic studies suggested that a hydroxide ligand is harbored between the two metals at the active site in the Ni-B state,78,79 whereas the identity of the bridging ligand in the Ni-A state remains controversial. The modified first coordination sphere on the Ni atom in the electron density map is typically regarded as characteristic of the Ni-A state,8083 but a recent DFT analysis suggests that the spectroscopic difference between Ni-A and Ni-B is due to the slightly distorted ligand orientation of the cysteine residues.78

The diamagnetic Ni-SU state resulting from the one-electron reduction of Ni-A was recently proposed to have a terminal coordinated water molecule to the Fe atom and to exhibit modified coordination of Ni.66 In contrast, a one-electron reduction of Ni-B results in the diamagnetic Ni-SIr state.84 The OH ligand is bridged between the Ni and Fe atoms in the Ni-SIr state and further reduction induces release of the ligand as a water molecule, resulting in transition to the Ni-SIa state. These results suggest that the OH ligand in the Ni-B state originates in a nearby water molecule to protect the active site from O2 damage.

The distal [4Fe4S] cluster is ligated by three well-conserved cysteines and one histidine residue located at the protein surface, and interacts with an electron acceptor. Stein et al. calculated that the electron transfer rate changes when this histidine is replaced by cysteine, indicating that the protein environment is tuned for optimum electron transfer relay.85 Another amino acid variant near the distal [4Fe4S] cluster showed that the redox potential can be tuned by the second coordination sphere of the electron transfer site.86 The electron transfer route from the active site to the molecular surface is complex, with multiple routes via the amino acid residues between the Fe-S clusters co-existing simultaneously during catalysis.87 O2-tolerant hydrogenases contain a dimeric structure in which the distal [4Fe4S] clusters in each monomer are located approximately 12 Å apart, suggesting that intra-dimer electron transfer may be important for O2-tolerance.88,89

A recent unprecedented result demonstrated that the isolated large subunit from [NiFe]-hydrogenase from Ralstonia eutropha, which lacks the small subunit, shows H2 activation (H/D exchange reaction) under O2 conditions.90 No H2 oxidation/H+ reduction activity was observed due to the lack of the electron transfer relay. The isolated large subunit formed a mixture of the dimeric and the monomeric forms. The presence of the C-terminal extension might prevent access of an electron acceptor to the active site.

Substrate H2 diffuses from the molecular surface to the active site buried deeply in the protein matrix. Several models predict H2 diffusion channels: most of the models suggest that the channels are connected to the Ni atom at the active site through a “bottle-neck” path located at two hydrophobic amino acid residues, typically valine and leucine. Crystal structures have shown that molecular O2 uses similar pathways.47,48

The protons, produced by heterolytic cleavage of H2 at the active site, are transferred to the molecular surface.91 Several proton transfer pathways have been proposed to date, including the glutamate pathway and the arginine pathway. High-resolution crystallography and nuclear resonance vibrational spectroscopy studies show that in the Ni-R state the proton is attached to the terminal cysteine residue at the first coordination sphere of the Ni atom in the active site.67,68 In the next step, the glutamate residue located near the cysteine residue accepts the protons. The proton is further transferred through various pathways (Figure 3).92 The threonine residue close to the glutamate residue can accept the proton and transfer it to a water molecule through the hydrogen bond network at the molecular surface.67 The role of the glutamate pathway is further supported by recent FTIR spectroscopic studies71 showing that the terminal Ni-coordinated cysteine sulfur atom is protonated in the Ni-L state and deprotonated in the Ni-C and Ni-SIa states. The glutamate COOH stretching band was detected in the Ni-L state with a double hydrogen bond and in the Ni-C and Ni-SIa states with single hydrogen bonds.

O2-tolerant MBH-type [NiFe] hydrogenases have a complicated proton and electron transfer mechanism because they must rapidly reactivate following O2 attack. This requires that the electrons and protons be supplied to the active site for quick O2 reduction. A possible mechanism for the proton and electron transfer pathways between the active site and the unique proximal [4Fe3S] cluster has been investigated for the O2-tolerant [NiFe]-hydrogenases from Hydrogenovibrio marinus (HmMBH), Escherichia coli (EcMBH), and Ralstonia eutropha (ReMBH). When O2 invades the H2-uptake catalytic cycle, the proximal [4Fe3S]3+ cluster donates two electrons back to the active site (or to the still-unknown location for O2 reduction) to prevent formation of the inactive state, resulting in the super-oxidized state [4Fe3S]5+.85,87,9396 In response to this opposite electron flow, Fe4 in the super-oxidized proximal cluster prefers the deprotonated amide nitrogen of the polypeptide backbone (cysteine residue) as a ligand relative to the original µ3-S (S3), and is stabilized by this coordination (Figure 4).44,45,97 In addition, a glutamate carboxylate near Fe4 (HmMBH and EcMBH)44,45 or an OH (ReMBH)43 at Fe1 also help stabilize the super-oxidized [4Fe3S].

A different inactive state was proposed for O2-tolerant NAD+-reducing soluble [NiFe]-hydrogenases compared with standard hydrogenases. Recent crystallographic studies of the enzyme from Hydrogenophilus thermoluteolus TH-1 in the oxidized state revealed a unique six-coordinated Ni with three bridging cysteine thiolate ligands, one terminal cysteine sulfur, and bidentate ligation of the glutamine side chain.49,98,99 Three bridging thiol ligands between the metals (both octahedral coordination) and a bidentate glutamate ligand found in oxidized NAD+-reducing [NiFe]-hydrogenase are believed to be an alternative system for protecting the active site from direct attack by O2 due to no coordination space for O2. Upon H2 activation, the active site undergoes a conformational change, resulting in structural features similar to the reduced state of standard [NiFe]-hydrogenases.

The crystal structure and EPR spectroscopic data of an EcMBH variant in which cysteine is mutated to glycine (C19G) showed that the modified proximal Fe-S cluster can transfer only one electron and loses O2-tolerance.100

The proton liberated from the amide nitrogen of the cysteine residue may be transferred towards the active site via a glutamate/histidine network to reduce O2 (Figure 5), after which the deformed [4Fe3S] is restored to the original form with concomitant conformational changes of the nearby residues. Consequently, the [4Fe3S] cluster is considered essential for O2-tolerance. Note that replacement of the valine residue at the bottle-neck of the gas-channel by cysteine in EcMBH results in increased O2-tolerance.101

An electrochemical study of the E28Q variant of EcMBH showed approximately 1% activity at pH 7 and suggested that the glutamate near the active site is a key residue for the exit route for protons from the active site.102 Another site-directed mutagenesis study near the proximal cluster showed that the E73Q variant does not affect O2-tolerance but increases H2 production.103 This glutamate (E73) is highly conserved in O2-tolerant hydrogenases whereas glutamine is conserved in O2-sensitive hydrogenases.

A proposed alternative proton transfer mechanism employing a frustrated Lewis pair (FLP) suggests that hydrogen is polarized between the acid metal(s) and the deprotonated guanidine moiety (the base) of the arginine residue (Figure 1B) that is highly conserved in [NiFe]-hydrogenases.104106

Recently, another type of O2-stable Hyb-type heterotetrameric [NiFe]-hydrogenase was found, in Citrobacter sp. S-77 (S77HYB).51 The hydrogenase unit of this enzyme is stable to O2 exposure and its catalytic activity is recovered when immersed in an H2-containing solution with about 1.0% O2,21 indicating moderate O2-reducing ability. S77HYB does not contain [4Fe3S] but has a [4Fe4S] cluster proximal to the Ni-Fe active site, similar to O2-sensitive ‘standard’ enzymes. The [4Fe4S] in S77HYB likely donates two electrons to the active site, as proposed for the proximal [4Fe3S] in MBH-type enzymes, but the molecular mechanism is slightly different. When S77HYB is exposed to O2, the proximal [4Fe4S] donates two electrons and likely super-oxidizes and deforms with a shift in Fe4, which is stabilized by the coordination of the nearby COO and OH between Fe2 and Fe4. Upon reduction, the deformed [4Fe4S] is immediately restored to its original cubane configuration upon the concomitant relocation of water molecules (Figure 6).

Standard and other O2-sensitive Hyb-type enzymes have no or one O2-protecting system, that is, an electron supply system (flexible Fe-S cluster and nearby residue) or a water relocation system that restores the deformed cluster. The medial cluster [3Fe-4S] in S77HYB has a high potential and may play a crucial role in the back-donation of electrons to the Ni-Fe active site, given the lower redox potential of [4Fe4S] compared with those of [4Fe3S] in MBH-type enzymes.107 The molecular mechanism in S77HYB for protecting the active site from O2 damage is intermediate between those of O2-sensitive (standard) and O2-tolerant (MBH-type) [NiFe]-hydrogenases (Figure 6).

Hydrogenases can catalyze non-physiological reactions such as proton/deuterium (H/D) exchange and nuclear para-H2/ortho-H2 spin conversion reactions in the absence of an electron mediator.108 These features are not directly linked to the catalytic reaction in cells but are useful for understanding the kinetics of the enzymatic reactions. At ambient temperature, molecular hydrogen is an equilibrium mixture of para- and ortho-H2 (called normal-H2), that is, H2; ortho-H2:para-H2 = 3:1, and D2; ortho-D2:para-D2 = 1:2. The ground states (J = 0), para-H2 and ortho-D2, are enriched at cryogenic temperatures and are stable for a prolonged time period. Little spontaneous spin conversion (i.e., conversion to normal-H2/D2) occurs in the absence of catalyst even at ambient temperature.

During the H/D exchange reaction of [NiFe]-hydrogenases in a D2/H2O system, D2 is transferred via the gas channel to the active site, then is split into D+ and D heterolytically. H/D exchange then occurs without electron transfer (eq 1), where E and E:D represent the active site and the coordination of D, respectively.

$$\text{E} + \text{D}_{2} + \text{H}^{+} \leftrightarrow \text{E}{:}\text{D}^{-} + \text{D}^{+} + \text{H}^{+} \leftrightarrow \text{E} + \text{HD} + \text{D}^{+}$$
(1)
The H/D exchange reaction is conventionally investigated using mass spectroscopy. The spin conversion reaction between para-H2 and ortho-H2 (eq 2) occurs at the active site without electron or proton transfer and is generally analyzed by gas chromatography and thermal conductivity.
$$\text{E} + \textit{para}\text{-H}_{2} \leftrightarrow \text{E}{:}\text{H}^{-} + \text{H}^{+} \leftrightarrow \text{E} + \text{normal-H}_{2}$$
(2)

The kinetics for hydrogenase from Clostridium pasteurianum and Proteus vulgaris were first reported by measuring the production of HD and D2 in an H2/D2O system.109 The experimental results suggested the heterolytic cleavage of H2 into H+ and H. Yagi et al. reported a detailed kinetic analysis of the purified standard [NiFe]-hydrogenase from Desulfovibrio vulgaris Miyazaki F (DvMSTD).110 They concluded that the two unequal H atoms cleaved from H2 were exchangeable, but with different exchange rate constants. The H/D exchange activities of several [NiFe]-hydrogenases are summarized in Table 1.111,112 Prior to the 1990s, H/D exchange reactions were discussed in terms of the initial production rates of HD and H2 and focused mainly on the mechanism of the double exchange reaction (i.e., D2 + 2H+ → H2 + 2D+).109,110,112118 Jouanneau and Vignais reported an online mass spectrometry technique that allows continuous monitoring of the variation of the gas composition in a cell extract system.117,119121 Leroux et al. concluded that gas diffusion rates are defined not only by the narrowness of the gas pathway, but also by the dynamics of amino acid side chains composing the gas pathway.117

 Table 1. H/D exchange activities of several [NiFe]-hydrogenases
Table 1. H/D exchange activities of several [NiFe]-hydrogenases
 species Type isotope exchange activityµmol min−1 mg−1 H2 evolutionµmol min−1 mg−1 H2 uptakeµmol min−1 mg−1 Additives ref Thiocapsa roseopersicina [NiFe] 220(pH = 5.5) 65(pH = 4.0) 55(pH = 9.5) MV 112 Azobacter vinelandii [NiFe] 34(pH = 5.0) 32(pH = 5.0) — dithionite 112 Desulfovibrio gigas [NiFe] 120(pH = 4.0–4.5) 440 (pH = 4.0–4.5) 1500 (pH = 8.0) dithionite 122,123 [NiFe] 267(pH = 7.5–8.0) — — dithionite 122 Desulfovibrio fructosovorans [NiFe] 223(pH = 5.5) 65 (pH = 7.6) 330 (pH = 8.5) — 112 Desulfovibrio vulgaris Miyazaki F [NiFe] 28(pH = 7.4) — — — 58 [NiFe] 282(pH = 7.4) 75600(pH = 7) 50400(pH = 7) Cytc3 110 Desulfomicrobium baclatum [NiFeSe] 350(pH = 4.0) 466 (pH = 4.0) 120 (pH = 7.5) — 112,122,123 Desulfovibrio salexigens British Guiana [NiFeSe] 900(pH = 3.5–4.0) 1830 (pH = 4.5–5.0) 1300 (pH = 7.5–8.0) — 122 Desulfovibrio vulgaris strain Hildenborough [FeFe] 2700(pH = 5.0) 4800 (pH = 5.6–6.0) 50000 (pH = 8.4) — 122 Desulfovibrio desulfuricans [FeFe] 3360(pH = 5.5) 8200(pH = 8.0) 62200(pH = 8.0) dithionite 124

A novel assay method for enzymes reacting with gaseous substrates using Raman spectroscopy has recently been published.58 In this system, the gas phase in the reaction mixture is illuminated by excitation light (an 800 mW, 532.0 nm Nd:YVO4 laser) that does not damage the enzyme in the solution phase, allowing non-invasive monitoring of the catalytic reaction over a prolonged time period. Furthermore, the setup with a 30-cm spectrometer and a 500-nm blazed grating (1200 grooves/mm) allows simultaneous measurements of a wide Raman spectral range corresponding to the rotational and vibrational signals of ortho/para-D2, HD, and ortho/para-H2 (2900 to 4200 cm−1) (Figure 7).

Applying this Raman method to the H/D exchange reaction by DvMSTD in a D2/H2O system allowed estimation of the initial production rates of HD (v1) and H2 (v2). The obtained ratio v2/v1 was not affected by enzyme concentration, suggesting that the double-exchanged product (H2) could be produced by two successive single H/D exchange reactions at the same active site. The time courses of individual components of the D2/H2O exchange reactions were quantitatively analyzed according to the reaction model (Figure 8).117 The data were fitted to the kinetic model and showed that the H/D exchange rate at the active site and the product release rate from the enzyme were comparable.

[NiFe]-hydrogenases are found in microorganisms such as archaea, bacteria and also in some eukaryotes. Most of them live in anoxic environments such as fresh water ponds, marine sediments, and hot springs. These [NiFe]-hydrogenases are O2-sensitive, but in organisms living in aerobic environments some O2-tolerant enzymes have been described such as Hydrogenovibrio marinus, Ralstonia eutropha or Escherichia coli. The catalytic cycle of [NiFe]-hydrogenases have been determined by crystallography, spectroscopy and theoretical calculations. In the Ni-SIa state the third bridging ligand is vacant. After heterolytic splitting of H2, the hydride binds to the bridging position between the metals and the proton attached to the terminal cysteine thiolate in the Ni-R state. In the Ni-C state the hydride remains and the proton is released, subsequently the hydride is removed as the proton and the Ni atom is reduced in Ni+ in the Ni-L state. Several proton transfer pathways have been proposed by theoretical calculations. The neutron diffraction studies will help to identify the proton transfer pathways.

X-ray crystal structures of some O2-tolerant [NiFe]-hydrogenases have been determined. In the case of MBH, the unique proximal [4Fe3S] cluster provides the extra electrons to reduce O2. The O2-tolerant NAD+-reducing soluble [NiFe]-hydrogenase from Hydrogenophilus thermoluteolus TH-1 shows another O2 resistance mechanism that the six-coordinated Ni atom by four cysteine thiolate ligands and bidentate ligation of glutamate protect the active site from the direct O2 attack. In the case of the Hyb-type hydrogenase from Citrobacter sp. S-77, an extra electron can be provided by the proximal [4Fe4S] cluster which is distorted in the oxidized state.

Recent development of a new Raman spectroscopy assay method can provide useful insight into the kinetics with small molecules, such as hydrogen or methane.

We thank all co-workers and collaboration partners named in the cited references.

The work was partly supported by a MEXT KAKENHI Grant-in-Aid for Scientific Research on Innovative Areas to Y.H. (Hydrogenomics, No. 18H05516), JSPS KAKENHI Grants-in-Aid for Scientific Research (A) to Y.H. (No. 19H00984), JST CREST grant JPMJCR12M4 to Y.H., JSPS KAKENHI Returning Researcher Development Research to H.O. (No. 16K21748), and Kinoshita Kinen Jigyo to K.N.

Yoshiki Higuchi is currently a full professor in the Graduate School of Life Science, University of Hyogo. He obtained his Ph.D. from Osaka University (1984). After a postdoctoral position at the Institute for Protein Research in Osaka University, he was appointed as an assistant professor at the Himeji Institute of Technology (1985). He moved to Kyoto University in 1995 as an associate professor of the Graduate School of Science. In 2002, he moved to the Himeji Institute of Technology (presently University of Hyogo) to become a full professor in the Graduate School of Science. He received the award of The Crystallographic Society of Japan in 1999, 38th Iue Culture Award for Science and Technology in 2014, and Hyogo Science Award in 2017.

Hideaki Ogata is a specially appointed associate professor, Institute of Low Temperature Science, Hokkaido University, Sapporo, Japan since 2017. He earned a Ph.D. from Kyoto University (2003). He worked at Himeji Institute of Technology as a postdoc for a half year. He joined the Max Planck Institute for Chemical Energy Conversion, Mülheim an der Ruhr, Germany as a postdoc, later on he worked as a group leader (2003–2017). His research interest lies in understanding the structure and function of metalloenzymes.

Koji Nishikawa is an assistant professor in the Graduate School of Life Science, University of Hyogo since 2012. He obtained his Ph.D. from University of Hyogo (2010). He joined the Institute for Protein Research, Osaka University (2010) and Max Planck Institute for Chemical Energy Conversion, Mülheim an der Ruhr, Germany (2010–2012) as a postdoc. His research interest lies in structural biology.