New assay method based on Raman spectroscopy for enzymes reacting with gaseous substrates

Article abstract of DOI:10.1002/pro.3569

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/H₂O 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 (k_out) 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.

Full text of DOI:10.1002/pro.3569

Protein Science
Methods and Applications
DOI 10.1002/pro.3569
New Assay Method Based on Raman Spectroscopy for Enzymes Reacting with
Gaseous Substrates
Yuka Kawahara-Nakagawa1†, Koji Nishikawa1,2†, Satoru Nakashima3 Shota Inoue1, Takehiro
Ohta4, Takashi Ogura1‡, Yasuteru Shigeta5, Katsuyuki Fukutani6, Tatsuhiko Yagi7, Yoshiki
Higuchi1,2*
†These authors contributed equally to this work.
‡Deceased July 23, 2017
1 Graduate School of Life Science, University of Hyogo, 3-2-1 Koto, Kamigori-cho, Ako-gun,
Hyogo 678-1297, Japan
2
Core Research for Evolutional Science and Technology (CREST), Japan Science and
Technology Agency (JST), Kawaguchi Center Building, 4-1-8 Honcho, Kawaguchi-shi,
Saitama 332-0012, Japan
3
Graduate School of Material Science, Nara Institute of Science and Technology, 8916-5
Takayama, Ikoma, Nara 630-0192, Japan
4
Department of Applied Chemistry, Faculty of Engineering, Sanyo-Onoda City University,
1-1-1 Daigakudori, Sanyo-Onoda, Yamaguchi 756-0884, Japan
This article has been accepted for publication and undergone full peer review but has not been
through the copyediting, typesetting, pagination and proofreading process which may lead to
differences between this version and the Version of Record. Please cite this article as
doi: 10.1002/pro.3569
© 2019 The Protein Society
Received: Nov 08, 2018; Revised: Dec 27, 2018; Accepted: Dec 31, 2018
This article is protected by copyright. All rights reserved.

5 Center for Computational Science, University of Tsukuba, 1-1-1 Tennodai, Tsukuba,
Ibaraki 305-8571, Japan
6Institute of Industrial Science, The University of Tokyo, 4-6-1 Komaba Meguro-ku, Tokyo
153-8505, Japan
7 Shizuoka University, 836 Oya, Suruga-ku, Shizuoka 422-8529, Japan
Abbreviations: DvMF, Desulfovibrio vulgaris Miyazaki F
*Correspondence to: Yoshiki Higuchi, Graduate School of Life Science, University of Hyogo,
3-2-1 Koto, Kamigori-cho, Ako-gun, Hyogo, 678-1297, Japan
Email address: hig@sci.u-hyogo.ac.jp
Running title: H/D exchange reaction of H2ase by Raman method (46/50 characters)
Manuscript pages: 30
supplementary material pages: 2
1 Table; 6 Figures
Supplementary materials (filename: Supplementary materials_revised.docx) have 2 Tables
2
This article is protected by copyright. All rights reserved.

Abstract: 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 non-physiological 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/H2O system. Based on a
kinetic model, the ratio of the rate constants of the H/D exchange reaction (k) at the active
3
This article is protected by copyright. All rights reserved.

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. (244/250 words)
Keywords: Raman spectroscopy, H/D exchange reaction, [NiFe] hydrogenase, kinetic
experiment, quantitative analysis of gaseous substrates, non-invasive measurement
50-75-word statement, written for a broader audience
A new assay system for measuring the time course of enzymatic reactions involving gaseous
substrates by Raman spectroscopy was established. By this system, we can estimate the gas
composition in the reaction cuvette continuously in a non-invasive manner over a prolonged
period. We investigated the H/D exchange reaction catalyzed by [NiFe] hydrogenase and
determined that the ratio of the rate constants for the H/D exchange reaction (k) and product
release (kout) was 1.9±0.2. (73/75 words)
4
This article is protected by copyright. All rights reserved.

Introduction
Understanding the detailed mechanisms of catalytic reactions mediated by enzymes
is fundamental for developing new industrial biocatalysts and applying them to biomimetic
processes. Enzymatic reaction mechanisms have been studied using biochemical and
spectroscopic techniques, and structural information of the active site has been fundamentally
important. The experimental data are typically analyzed based on simple kinetic models
focusing on the chemical reaction occurring at the active site.1,2 However, the overall
catalytic performance of an enzyme depends on not only the reaction occurring at the active
site but also the protein system, which regulates the transfer of the substrates and products as
well as that of electrons. Consequently, the protein system contributes to the suppression of
side reactions. For enzymatic reactions involving gaseous substrates, the gas channel through
which the substrate molecules pass from the enzyme surface to the active site and the release
route of the products to the reaction medium are important for high catalytic performance.
Hydrogenase, which catalyzes the reversible conversion of molecular hydrogen (H2) to
protons (H+) and electrons (e−), is one example of enzymes that react with gaseous
substrates.3 Since hydrogenase is a potential candidate for a biofuel cell catalyst,
understanding its catalytic mechanism is desirable for developing more active and
5
This article is protected by copyright. All rights reserved.

inexpensive catalysts. In addition to the physiological reaction, namely, the reversible
oxidation of H2, hydrogenases catalyze two reactions in vitro without electron transfer to the
mediators. These are the H/D exchange reaction between H+(H2O)/D+(D2O) and D2/H2 and
the nuclear spin isomer conversion reaction between para-H2 and ortho-H2.4–8 The former
reaction consists of the transfer of the gaseous substrate via the gas channel followed by
exchange between H+ and D+ (which requires the transfer of H+/D+ via the proton pathway)
after the cleavage of the covalent bond of H2, and subsequent re-formation of the bond at the
active site. The hydrogenase-mediated H/D exchange reaction is conventionally measured by
gas chromatography, mass spectrometry, or a combination thereof (GC/MS).8,9 Although
these methods are well established, they require the use of cumbersome procedures to
determine the composition of the gas phase in the reaction cuvette. Moreover, sampling the
gas from the cuvette at regular intervals, which is necessary to monitor the reaction, changes
the pressure of the gas phase after each operation. Alternatively, the gas composition can be
measured in situ using Raman spectroscopy without the need to remove sample material from
the reaction cuvette. This means that Raman spectroscopy enables us to trace the time course
of the enzymatic activity without requiring any data corrections owing to the change in gas
pressure inside the cuvette. Raman spectroscopy has the additional advantage of permitting
6
This article is protected by copyright. All rights reserved.

the continuous and simultaneous measurement of the changes in the ratios of isotopes and
nuclear spin isomers. Despite the great potential of Raman spectroscopy, there have been a
few reports of the application of this technique for monitoring enzymatic reactions10 and no
reports involving enzymatic reactions of gaseous substrates to the best of our knowledge. In
this study, we developed a new Raman-spectroscopy-based assay system for measuring the
enzymatic activity toward gaseous substrates and successfully measured the H/D exchange
reaction catalyzed by the [NiFe] hydrogenase from Desulfovibrio vulgaris Miyazaki F
(DvMF).
7
This article is protected by copyright. All rights reserved.

Results
Continuous monitoring of the catalytic reaction by Raman spectroscopy
In the system shown in Figure 1A, only the gas phase in the reaction cuvette is
illuminated by the excitation light, such that the enzyme in the solution phase is not damaged
or otherwise affected. Since there is no need to withdraw aliquots of gas from the reaction
cuvette to measure the Raman spectra, the reaction conditions such as pressure and
temperature can be kept constant. In fact, Raman intensity of D2/H2 (1 atm) in the reaction
cuvette did not change at all for over one day. It is common practice to add a small amount of
dithionite to remove residual O2 from the assay system of hydrogenases.8,11,12 In this study,
however, the reaction mixture was anaerobically prepared in the custom-built reaction cuvette
shown in Fig. 1B inside a glovebox and no reductants are required (see Materials and
Methods). The highest peaks in the rotational–vibrational Raman spectrum of D2, HD, and H2
were located at 2987, 3628, and 4155 cm−1, respectively, as shown in Figure 2.13 All of the
peaks originating from the sample gases were observed simultaneously in a single spectrum.
The resolution of the system was sufficient for discrimination of the rotational states (spin
isomers); five peaks corresponding to the possible isotopic/isomeric species (para-D2,
ortho-D2, HD, para-H2, and ortho-H2) could be clearly distinguished as labeled in Figure 2.
8
This article is protected by copyright. All rights reserved.

H/D exchange reaction in the D2/H2O system
Figure 3(A) shows the time course of the Raman peak intensities of D2, HD, and H2
during the H/D exchange reaction in the D2/H2O system catalyzed by the [NiFe] hydrogenase
from DvMF (10 M) at 30°C. During the course of the reaction, the Raman peak intensities
of D2 and H2 exhibited a simple decay and rise, respectively. In contrast, the Raman peak
intensity of HD increased over the period from 0 to 200 min, began to decay after 200 min,
and then finally disappeared at 1000 min. The amounts of D2, HD, and H2 were estimated by
calculating the area intensities of the corresponding Raman bands, which were normalized
using standard gases (see Materials and Methods). Since the total number of gas molecules in
the reaction cuvette during the H/D exchange reaction was kept unchanged, each of the
gaseous components is hereafter represented as a molar fraction. As shown in Figure 3(B),
the time course data obtained for the H/D exchange reaction were sufficiently accurate to
permit precise evaluation of the kinetics of our system. The time course of the D2 fraction
revealed that this component underwent exponential decay, where the time constant (pseudo–
first-order rate constant) (D) at the enzyme concentration of 10 M was 0.0050±0.0004
min−1. The dependency of enzyme concentration was observed in D at the range of 2–20 M
in Figure 4. Since no accelerated features were observed on the H/D exchange reaction at the
9
This article is protected by copyright. All rights reserved.

initial stage of the reactions [Fig. 3(B)], the enzyme samples used in this study have a full
activity from the beginning of the measurements.
In our developed system, the reaction mixture could not be stirred vigorously (see
Materials and Methods). In addition, this measurement setup resulted in the large volume of
the gas phase compared to the liquid phase. For these reasons, we had to use the concentrated
enzyme solutions (2, 5, 10 and 20 µM) to measure the time course of the H/D exchange
reaction in an appropriate period of time. The H/D exchange activities (2–6 mol (D2) s-1 (mol
enzyme)-1) calculated from the initial rates of D2 consumption for the concentrated enzyme
solutions (Table S1) were considerably lower than those estimated from the previous
reports8,14 (about 120 mol s-1 (mol enzyme)-1 ). Since the dependency of the H/D exchange
activity on the enzyme concentration was reported before,8 we carried out experiments with
the diluted enzyme solutions. The activity estimated from 20 nM enzyme showed comparable
value (47 mol s-1 (mol enzyme)-1) as shown in Table S1, if the large differences of the
experimental conditions between two methods are taken into account (see Materials and
Methods). In order to investigate the characteristics of the reaction system, D at various
volumes of enzyme solutions were calculated (Table S2).
The initial production rate of HD (v1, 0.0018±0.0002 min−1 at the enzyme
10
This article is protected by copyright. All rights reserved.

concentration of 10 M) was almost equal to that of H2 (v2, 0.0022±0.0003 min−1). These
results could not be adequately analyzed using the simple stepwise reaction mechanism
shown in equations 1 and 2, indicating that not only the single exchange reactions (e.g., D2 +
H+ → HD + D+) but also the double exchange reaction (i.e., D2 + 2H+ → H2 + 2D+)
occurred macroscopically in the reaction cuvette.
D2 + H+→ HD + D+
HD + H+→ H2 + D+
(1)
(2)
Effect of the product release rate from the hydrogenase in the H/D exchange
reaction using a kinetic model
To analyze the time course of the H/D exchange reaction in the D2/H2O system
quantitatively, we adopted the reaction model proposed by Leroux et al., as depicted in
Figure 5.15 In this model, k is the rate constant of the H/D exchange reaction at the active site
and kin and kout are the rate constants for the trapping and release, respectively, of molecular
hydrogen species by the hydrogenase molecule (including the binding or dissociation of them
at the Ni–Fe active site and the transfer through the gas channel). To obtain the ratios of k to
kout, the time courses of D2, HD and H2 at several enzyme concentrations were fitted against
the analytical solutions derived from the model equations presented in Figure 5. As indicated
11
This article is protected by copyright. All rights reserved.

by the green lines in Figure 4, the fitted curves were in good agreement with the experimental
data. The values of k/kout (1.8–2.1) were approximately constant for enzyme concentrations in
the range of 2–20 M (Table I), and similar to those calculated by the method where the
maximum points of HD are used.15
12
This article is protected by copyright. All rights reserved.

Discussion
Until the 1990s, H/D exchange reactions by hydrogenases were studied and
discussed with respect to the initial reaction rates of HD and H2 production, where the main
focus was the mechanism of the double exchange reaction.4,5,8 Recently, the technique of
online mass spectrometry, which allows monitoring of the time course of the variation of the
gas composition of the reaction, has been used to obtain more detailed information regarding
the mechanism of the H/D exchange reaction.15,16 However, it is difficult to survey the entire
time course of the reaction by mass spectrometry owing to the continuous consumption of the
gas in the reaction cuvette during sampling. Using the newly developed system based on
Raman spectroscopy, we confirmed that H2 (i.e., the fully exchanged product) is
concomitantly produced alongside HD during the H/D exchange reaction in the D2/H2O
system. As shown in Figure 3(B), the initial production rates of HD and H2 were almost equal.
It is generally accepted that D2 is heterolytically cleaved to produce D− and D+ following
binding to the Ni–Fe active site. The D− bridges the Ni and Fe and the D+ is proposed to bind
to the cysteine (Cys546) sulfur atom that coordinates the Ni (the Ni-R state in Fig. 6).3,7,14,17
In the D2/H2O system, the formation of HD and H2 at the Ni–Fe active site is supposed to
proceed according to equations 3–6:
13
This article is protected by copyright. All rights reserved.

A + D2 → A-D− + D+
A-D− + H+→ A + HD
A + HD → A-H− + D+
A-H− + H+→ A + H2
(3)
(4)
(5)
(6)
where “A” and “A-H− + H+” represent the Ni–Fe active site and the Ni-R state, respectively,
as shown in Figure 6. According to equations 3–6, H2 should appear after HD production. In
the D2/H2O system, the fully exchanged product (H2) could be produced by two successive
single H/D exchange reactions (1) in different active sites or (2) at the same active site
(macroscopically double exchange reaction).6,8,18 Since the ratio of v2/v1 depended little on
the enzyme concentrations (Table I), the former is denied. Besides, D depended on the
enzyme concentration and did not depend on the volume of the enzyme solution. The results
indicate that the reaction probably occurred near the gas-liquid interface and not controlled
by the diffusion of the substrate. Hence, the entire time courses of individual components in
the reaction were quantitatively analyzed and well fitted to the equations derived from the
previously described reaction model (Fig. 5).15 From the fitting results in Fig. 4, k/kout was
estimated to be 1.9±0.2 (Table I). This obtained value indicates that the rates of the H/D
exchange reaction at the active site and the release of the products from the enzyme were
14
This article is protected by copyright. All rights reserved.

comparable. Hamdan et al., reported that k and kout can be separately calculated under the
condition of largely excess concentration of substrate over Michaelis constant, Km.19
Although k/kout was successfully estimated in this study, it is difficult to apply this
approximation because there is insufficient substrate concentration in the solution (760 M)
against Km of DvMF (130 M for D2 in the H/D exchange reaction)8 and to discuss the
relationship between kin and kout based on the present reaction model. In addition, the
assumption that all molecular hydrogen species share the same kinetics in this model is not
valid, because the masses of D2 and HD are twofold and 1.5-fold that of H2, respectively,
while the kinetic isotope effects may be as high as 17-fold and fourfold that of H2,
respectively. To clarify these issues, the following reactions must be investigated: (1) the H/D
exchange reaction in the H2/D2O system; and (2) the nuclear spin isomer conversion reaction,
that is, the conversion between para-H2 and ortho-H2 (or between ortho-D2 and para-D2) in
H2O or D2O. These measurements can be performed using exactly the same assay system
proposed in this study.
In the experimental setup of our Raman system using concentrated enzyme solution,
the enzymatic reaction mainly takes place at the interface between the gas and liquid phases.
This is one of the reasons that the activities of the H/D exchange reaction showed the
15
This article is protected by copyright. All rights reserved.

dependency of the enzyme concentrations (Table S1). Nevertheless, D of D2 decay with the
concentrated enzyme in this study was dependent on the enzyme concentrations (probably on
the enzyme density at the interface) and k/kout was successfully estimated from the entire time
course of the H/D exchange reaction. These results indicate that the data obtained in this
study were not controlled by the diffusion limit of the substrate at the liquid-gas interface, but
demonstrated the typical feature of the enzymatic reaction, even though the improvement of
the measurement system is required for further assay of the H/D exchange activity of the
enzyme more accurately. We believe that our newly developed Raman method is potentially
useful and is applicable to assay the enzymes reacting with gaseous substrates.
16
This article is protected by copyright. All rights reserved.

Materials and Methods
Materials and sample preparation
[NiFe] hydrogenase from DvMF was purified as described in the previous report.17 Enzyme
purification was performed under strict anoxic conditions. The concentration of the
hydrogenase was determined using the absorption coefficient at 400 nm (47 mM−1 cm−1).
H2-uptake activity of the anaerobically purified enzyme assayed by the conventional
spectroscopic method with methyl viologen was 5620 mol s-1 (mol enzyme)-1, which was a
sufficiently high value compared to that [840 mol s-1 (mol enzyme)-1] prepared aerobically.14
The H/D exchange activities of [NiFe] hydrogenases were ranging from 45 to 450,20,21
whereas that obtained using our system was 47 mol s-1 (mol enzyme)-1 at 20 nM. Although it
is difficult to compare them directly due to the large differences of the experimental
conditions, including the reaction cuvettes, bubbling conditions, additives (electron transfer
media), the obtained value in this study was reasonable. The D2/H+ exchange assay was
performed in 25 mM Tris-HCl buffer (pH = 7.4). Dissolved oxygen in the buffer solution was
removed by bubbling with argon gas prior to use in a glovebox. The glovebox (Coy
Laboratory Products Inc.) was maintained under anoxic conditions with H2 (1.5±0.5%) and
N2 (98.5±0.5%). H2 gas (99.99%) was purchased from Taiyo Nippon Sanso Corporation. D2
17
This article is protected by copyright. All rights reserved.

(99.8%) and HD (D: 97%) were purchased from Cambridge Isotope Laboratories Inc.
Raman spectroscopy
Raman scattering was measured using a 532.0 nm Nd:YVO4 laser (Spectra-Physics,
Millennia X) as the excitation source with a power of 800 mW at the sample point. The
detector was a liquid-nitrogen-cooled CCD (Roper Scientific Inc., Spec-10:400B) attached to
a 30-cm spectrometer (Princeton Instruments Inc., Acton SpectraPro 300i) with a 500-nm
blazed grating with 1200 grooves/mm. This setup enabled us to carry out the simultaneous
measurement of a wide Raman spectral range including the D2, HD, and H2 signals. The
exposure time was 60 s and five scans were accumulated for each spectrum to provide a
sufficiently high signal-to-noise ratio to allow quantitative discussion. The measurements
were performed in a custom-built reaction cuvette as shown in Figure 1(B). All of the optical
components were installed on an optical table and arranged as depicted in Figure 1(A).
Assay of H/D exchange reaction
The reaction vessel consisted of a glass reaction cuvette and a two-way cock unit attached to
the top of the cuvette via a ground glass joint. The hydrogenase solution was gently placed in
18
This article is protected by copyright. All rights reserved.

the bottom of the reaction cuvette. After connecting the two units at the ground glass joint
with high-vacuum sealant, the gas inside the cuvette was replaced with D2 gas at 100 kPa via
a vacuum line connected to the inlet of the two-way cock unit. The two-way cock was closed
prior to removal of the vacuum line. Finally, the reaction cuvette was placed in the cuvette
holder in front of the spectrometer to initiate the Raman measurements. The reaction mixture
was continuously stirred (1000 rpm) using a magnetic stirrer and maintained at 30°C in a
thermostated chamber. Unless otherwise noted, the volumes of the reaction mixture and gas
phase were 1 mL and 10 mL, respectively. Since the droplets attached on the wall of the
inside of the cuvette by the splashed solution of the reaction mixture disturb the transmission
of the laser light, resulting in the inaccurate measurement of the Raman peak intensity, the
enzyme solution was not able to be stirred vigorously in the present measurement system.
This could be solved by improving the reaction system in future work. All of the procedures
except the Raman measurements were performed in the glovebox.
Calculation of the gas composition
We measured the area intensities of the Raman bands for each pure isotope individually. The
D2:HD:H2 intensity ratio was 0.71:0.88:1. The observed area intensities of the Raman bands
19
This article is protected by copyright. All rights reserved.

of the individual isotopic species were corrected by dividing by these values to obtain the
molar fraction of each isotopic species.
Data analysis
The fractions of the isotopic species in the period from the start of the measurement to half of
the half-life of D2 were used to calculate the initial rates of HD and H2 production. The data
were least-squares fitted with a straight line to obtain the slopes corresponding to v1 and v2
for HD and H2, respectively.
The time courses of the D2/H+ exchange reactions were fitted using the analytical
solution of the reaction model depicted in Figure 5. The calculations were described in detail
in the previous report.15 The initial conditions were [D2]0 = 1, [HD]0 = 0, and [H2]0 = 0. The
equations used for the fitting were as follows:
[D2] = [D2]0 e−[ka / (1+kb)] t (7)
[HD] = 2[D2]0 {e−[ka /(2+kb)] t − e−[ka /(1+kb)] t} (8)
[H2] = [D2]0 (1− 2e−[ka /(2+kb)] t + e−[ka /(1+kb)] t) (9)
20
This article is protected by copyright. All rights reserved.

where ka = (k × kin ×[E]0)/(kout + kin×[D2]0), kb = k/kout, and [E]0 is the enzyme concentration.
The other terms are explained in the Results section and Figure 5.
Acknowledgements
This work was supported by the Japan Science and Technology Agency CREST grant
JPMJCR12M4 (Y.H.), a Japan Society for Promotion of Science Grants-in Aid for Scientific
Research (B), Grant No. 25291038 (Y.H.), for Challenging Exploratory Research, No.
24657077 (Y.H.) and for JSPS Fellows JP18J11420 (Y.K.), and a Ministry of Education,
Culture, Sports, Science and Technology Grant-in-Aid for Scientific Research on Innovative
Area (Hydrogenomics, 18H05518 (K.F.) and 18H05516 (Y.H.)). We thank Dr. N. Shibata, Dr.
T. Hiromoto, Dr. S. Yanagisawa, and Dr. M. Kubo for helpful discussions, Ms. S. Inomata
and Ms. A. Kakimoto for experimental assistance, and Ms. K. Hataguchi and Ms K.
Matsumoto for bacterial cultures. There are no conflicts of interest to declare.
21
This article is protected by copyright. All rights reserved.

References
1. Lippard SJ, Berg JM (1994) Principles of bioinorganic chemistry. University Science
Books.
2. Rehder D (2014) Bioinorganic chemistry. Oxford University Press.
3. u t , t , er , e erse 2 1 r en ses he ev 11 1–
4148.
4. Krasna AI, Rittenberg D (1954) The mechanism of action of the enzyme hydrogenase1. J
Am Chem Soc 76:3015–3020.
5. Berlier Y, Fauque GD, LeGall J, Choi ES, Peck HD, Lespinat PA (1987) Inhibition studies
of three classes of Desulfovibrio hydrogenase: application to the further characterization of
the multiple hydrogenases found in Desulfovibrio vulgaris hildenborough. Biochem Biophys
Res Commun 146:147–153.
6. Vignais PM (2005) H/D exchange reactions and mechanistic aspects of the hydrogenases.
Coord Chem Rev 249:1677–1690.
7. Yagi T, Higuchi Y (2013) Studies on hydrogenase. Proc Japan Acad Ser B 89:16–33.
8. Yagi T, Tsuda M, Inokuchi H (1973) Kinetic studies on hydrogenase:
parahydrogen-orthohydrogen conversion and hydrogen-deuterium exchange reactions. J
22
This article is protected by copyright. All rights reserved.

Biochem 73:1069–1081.
9. Berlier Y, Lespinat PA, Dimon B (1990) A gas chromatographic-mass spectrometric
technique for studying simultaneous hydrogen-deuteron exchange and para-orthohydrogen
conversion in hydrogenases of Desulfovibrio vulgaris Hildenborough. Anal Biochem
188:427–431.
10. Westley C, Fisk H, Xu Y, Hollywood KA, Carnell AJ, Micklefield J, Turner NJ,
Goodacre R (2017) Real-time monitoring of enzyme-catalysed reactions using deep UV
resonance Raman spectroscopy. Chemistry 23:6983–6987.
11. Vignais PM, Henry M-F, Berlier Y, Lespinat PA (1982) Effect of pH on H-2H exchange,
H2 production and H2 uptake, catalysed by the membrane-bound hydrogenase of Paracoccus
denitrificans. Biochim Biophys Acta Bioenergetics 681:519–529.
12. McTavish H, Sayavedra-Soto LA, Arp DJ (1996) Comparison of isotope exchange, H2
evolution, and H2 oxidation activities of Azotobacter vinelandii hydrogenase. Biochim
Biophys Acta Protein Struct Mol Enzymol 1294:183–190.
13. Stoicheff BP (1957) High resolution Raman spectroscopy of gases: IX. spectra of H2, HD,
and D2. Can J Phys 35:730–741.
14. Yagi T, Ogo S, Higuchi Y (2014) Catalytic cycle of cytochrome-c3 hydrogenase, a
23
This article is protected by copyright. All rights reserved.

[NiFe]-enzyme, deduced from the structures of the enzyme and the enzyme mimic. Int J
Hydrogen Energy 39:18543–18550.
15. Leroux F, Dementin S, Burlat B, Cournac L, Volbeda A, Champ S, Martin L, Guigliarelli
B, Bertrand P, Fontecilla-Camps J, Marc R, Christophe L (2008) Experimental approaches to
kinetics of gas diffusion in hydrogenase. Proc Natl Acad Sci USA 105:11188–11193.
16. Cournac L, Guedeney G, Peltier G, Vignais PM (2004) Sustained photoevolution of
molecular hydrogen in a mutant of Synechocystis sp. strain PCC 6803 deficient in the type I
NADPH-dehydrogenase complex. J Bacteriol 186:1737–1746.
17. Ogata H, Nishikawa K, Lubitz W (2015) Hydrogens detected by subatomic resolution
protein crystallography in a [NiFe] hydrogenase. Nature 520:571.
18. Krasna AI (1979) Hydrogenase: properties and applications. Enzyme Microb Technol
1:165–172.
19. Abou Hamdan A, Dementin S, Liebgott PP, Gutierrez-Sanz O, Richaud P, De Lacey AL,
Roussett M, Bertrand P, Cournac L, Léger C (2012) Understanding and tuning the catalytic
bias of hydrogenase. J Am Chem Soc 134:8368–8371.
20. Bertrand P, Dole F, Asso M, Guigliarelli B (2000) Is there a rate-limiting step in the
catalytic cycle of [Ni-Fe] hydrogenases? J Biol Inorg Chem 5:682–691.
24
This article is protected by copyright. All rights reserved.

21. Fauque GD, Berlier YM, Czechowski MH, Dimon B, Lespinat PA, LeGall J (1987) A
proton-deuterium exchange study of three types of Desulfovibrio hydrogenases. J Ind
Microbiol 2:15–23.
25
This article is protected by copyright. All rights reserved.

Table I. D, v2/v1, ka and kb for several enzyme concentrations
D /min-1
v2/v1
ka
kb
Enzyme / M
0.0022
1.34
0.007
2.14
2
± 0.0003
± 0.08
 0.001
0.18
0.0038
1.20
0.011
1.86
5
± 0.0003
± 0.05
 0.001
0.02
0.0050
1.22
0.015
1.83
10
20
± 0.0004
± 0.07
 0.001
0.03
0.0074
1.18
0.016
1.83
± 0.0010
± 0.06
 0.003
0.04
All data are the average results obtained from three independently purified enzyme solutions.
ka = (k × kin ×[E]0)/(kout + kin×[D2]0)
kb = k/kout
26
This article is protected by copyright. All rights reserved.

Figure legends
Figure 1. Experimental setup. (A) Schematic illustration of the Raman-spectroscopy-based
assay system. (B) Custom-built reaction cuvette for anaerobic measurements of the reaction.
The glass cuvette possesses four transparent surfaces for Raman measurements.
Figure 2. Typical rotational–vibrational spectrum of the gaseous molecular hydrogen
species (H2, HD, and D2) measured using the Raman-spectroscopy-based assay system.
The indicated wavenumbers are for the highest peaks of the individual isotopic hydrogen
molecules. The characters p and o indicate the para and ortho states, respectively. The upper
two boxes show the enlarged views of D2 and H2.
Figure 3. H/D exchange reaction in the D2/H2O system catalyzed by [NiFe] hydrogenase
from Desulfovibrio vulgaris Miyazaki F. (A) Waterfall plot of the Raman spectra obtained at
25-min intervals during the H/D exchange reaction in the D2/H2O system. (B) Time course of
the isotopic composition of molecular hydrogen. The black, blue, and red markers represent
the fractions of D2, HD, and H2, respectively. The gray line over the black markers represents
the fitting curve for the exponential decay of D2. The pale blue and pale red lines in the inset
27
This article is protected by copyright. All rights reserved.

represent the fitting curves for the initial reaction rates of HD and H2 production,
respectively.
Figure 4. Kinetic analysis of the time course of the H/D exchange reaction catalyzed by
[NiFe] hydrogenase. The time course of the reaction was measured at enzyme
concentrations of 2, 5, 10, and 20 M. The black, blue, and red markers represent the molar
fractions of D2, HD, and H2, respectively. The green lines over the markers represent the
fitting curves calculated using the equations based on the model depicted in Figure 5. The
interval between each measurement was 5 min.
Figure 5. Scheme of the H/D exchange reaction used for the kinetic analysis.15
E
represents the free enzyme, and E(D2), E(HD), and E(H2) represent the enzyme bound to D2,
HD, and H2, respectively.
Figure 6. Proposed catalytic mechanism for H2 at the active site of [NiFe] hydrogenase.
The catalytic cycle consists of Ni-Sia, Ni-C and Ni-R states. Ni-R state harbors a hydride
bridge
between
the
Ni
and
Fe,
and
S
of
Cy
28
This article is protected by copyright. All rights reserved.

s546
is
protonated.
29
This article is protected by copyright. All rights reserved.

30
This article is protected by copyright. All rights reserved.