Carbonyl sulfide hydrolase from thiobacillus thioparus strain thi115 is one of the β-carbonic anhydrase family enzymes

Article abstract of DOI:10.1021/ja307735e

Carbonyl sulfide (COS) is an atmospheric trace gas leading to sulfate aerosol formation, thereby participating in the global radiation balance and ozone chemistry, but its biological sinks are not well understood. Thiobacillus thioparus strain THI115 can grow on thiocyanate (SCN^-) as its sole energy source. Previously, we showed that SCN^- is first converted to COS by thiocyanate hydrolase in T. thioparus strain THI115. In the present work, we purified, characterized, and determined the crystal structure of carbonyl sulfide hydrolase (COSase), which is responsible for the degradation of COS to H₂S and CO₂, the second step of SCN^- assimilation. COSase is a homotetramer composed of a 23.4 kDa subunit containing a zinc ion in its catalytic site. The amino acid sequence of COSase is homologous to the β-class carbonic anhydrases (β-CAs). Although the crystal structure including the catalytic site resembles those of the β-CAs, CO₂ hydration activity of COSase is negligible compared to those of the β-CAs. The α5 helix and the extra loop (Gly150-Pro158) near the N-terminus of the α6 helix narrow the substrate pathway, which could be responsible for the substrate specificity. The k _cat/K_m value, 9.6 × 10⁵ s^-1 M^-1, is comparable to those of the β-CAs. COSase hydrolyzes COS over a wide concentration range, including the ambient level, in vitro and in vivo. COSase and its structurally related enzymes are distributed in the clade D in the phylogenetic tree of β-CAs, suggesting that COSase and its related enzymes are one of the catalysts responsible for the global sink of COS.

Full text of DOI:10.1021/ja307735e

Article
pubs.acs.org/JACS
Carbonyl Sulfide Hydrolase from Thiobacillus thioparus Strain THI115
Is One of the β‑Carbonic Anhydrase Family Enzymes
Takahiro Ogawa,†,⊗ Keiichi Noguchi,‡ Masahiko Saito,†,▽ Yoshiko Nagahata,§,○ Hiromi Kato,†,▲
Akashi Ohtaki,∥,¶ Hiroshi Nakayama,⊥ Naoshi Dohmae,⊥ Yasuhiko Matsushita,# Masafumi Odaka,∥
Masafumi Yohda,∥ Hiroshi Nyunoya,# and Yoko Katayama*,†,§
§
†Department of Science of Resources and Environment, United Graduate School of Agricultural Science, Department of
Environmental Science on Biosphere, Graduate School of Agriculture, and #Gene Research Center, Tokyo University of Agriculture
and Technology, Fuchu, Tokyo 183-8509, Japan
∥
‡Instrumentation Analysis Center and Department of Biotechnology and Life Science, Graduate School of Technology, Tokyo
University of Agriculture and Technology, Koganei, Tokyo 184-8588, Japan
⊥Biomolecular Characterization Team, Advanced Technology Support Division, Advanced Science Institute, RIKEN, Wako, Saitama
351-0198, Japan
S
* Supporting Information
ABSTRACT: Carbonyl sulfide (COS) is an atmospheric trace gas
leading to sulfate aerosol formation, thereby participating in the global
radiation balance and ozone chemistry, but its biological sinks are not
well understood. Thiobacillus thioparus strain THI115 can grow on
thiocyanate (SCN−) as its sole energy source. Previously, we showed
that SCN− is first converted to COS by thiocyanate hydrolase in T.
thioparus strain THI115. In the present work, we purified, characterized,
and determined the crystal structure of carbonyl sulfide hydrolase
(COSase), which is responsible for the degradation of COS to H2S and CO2, the second step of SCN− assimilation. COSase is a
homotetramer composed of a 23.4 kDa subunit containing a zinc ion in its catalytic site. The amino acid sequence of COSase is
homologous to the β-class carbonic anhydrases (β-CAs). Although the crystal structure including the catalytic site resembles
those of the β-CAs, CO2 hydration activity of COSase is negligible compared to those of the β-CAs. The α5 helix and the extra
loop (Gly150−Pro158) near the N-terminus of the α6 helix narrow the substrate pathway, which could be responsible for the
substrate specificity. The kcat/Km value, 9.6 × 105 s−1 M−1, is comparable to those of the β-CAs. COSase hydrolyzes COS over a
wide concentration range, including the ambient level, in vitro and in vivo. COSase and its structurally related enzymes are
distributed in the clade D in the phylogenetic tree of β-CAs, suggesting that COSase and its related enzymes are one of the
catalysts responsible for the global sink of COS.
catalyze the transformation of COS (Table 1).12−22 These
enzymes are diversely distributed, but their activities on COS
are much less efficient than their activities on their natural
substrates. Rather, COS has been used as an inhibitor of these
INTRODUCTION
■
Carbonyl sulfide (COS; OCS) is the most abundant sulfur
compound in the troposphere,1,2 with the average mixing ratio
of ∼500 parts per trillion by volume (pptv)1 and is an
enzymes. Recently, carbon disulfide (CS2) hydrolase from an
important source of stratospheric sulfate aerosol influencing the
Earth’s radiation balance3−5 and ozone chemistry.6 COS is also
acidophilic and thermophilic archaeon, Acidianus sp. strain A1-
3, has been shown to efficiently catalyze the hydrolysis of COS
regarded as a greenhouse gas due to its high global-warming
as well as CS2.23 However, its habitat is limited in the extreme
environments such as solfataric springs.24 There have been no
reports on an enzyme that is found commonly and that only
catalyzes the degradation of COS. Furthermore, the under-
standing of COS degradation at the enzyme level is critical to
elucidate soil microorganisms as a major sink of COS.9
potential.7 Although there is still uncertainty in the global
budget of atmospheric COS, both anthropogenic and natural
activities have strong influences on the processes of source and
sink of COS.8,9 Biological COS degradation in a soil
environment is considered an important COS sink, as the
uptake of COS was inhibited by the addition of ethoxzolamide,
an inhibitor for carbonic anhydrase (CA; EC 4.2.1.1),10 and by
autoclaving the soil.11 Therefore, studying COS degrading
enzymes in the soil is important for solving global environ-
mental problems.
Thiobacillus thioparus is an obligately chemolithoautotrophic
and mesophilic sulfur-oxidizing bacterium. It is widely
distributed in soil and freshwater.25 T. thioparus strain
CA, a ubiquitous enzyme involving numerous biological
processes, and three other enzymes have been demonstrated to
Received: August 4, 2012
Published: February 13, 2013
© 2013 American Chemical Society
3818
dx.doi.org/10.1021/ja307735e | J. Am. Chem. Soc. 2013, 135, 3818−3825

Journal of the American Chemical Society
Article
Table 1. Comparison of Enzyme Kinetics of COS Degradation between COSase and Enzymes Harboring COS Degrading
Activity
enzyme
COSase
organism
kcat (s−1
)
Km (μM)
kcat/Km (s−1 M−1
)
reference
a
a
T. thioparus strain THI115
Acidianus sp. strain A1−3
Bos taurus
58
60
9.6 × 105
this study
b
CS2 hydrolase
CA
1800
41
22
8.2 × 107
23
17
20
21
22
22
1.9 × 103
3.1 × 103
2.2
2.2 × 104
b
nitrogenase
CO dehydrogenase
RuBisCO
Azotobacter vinelandii
Rhodospirillum rubrum ATCC11170T
Rhodospirillum rubrum
Spinacia oleracea
0.16
0.52
6.3
52
b
2.4 × 105
5.6 × 103
1.8 × 103
1.1 × 103
2.2 × 103
RuBisCO
3.8
a
b
Calculated by dimer. Calculated based on reference.
●
○
Figure 1. Degradation of COS by whole cells and COSase. (A) Time course of COS ( ) degradation and H2S ( ) production by COSase. COSase
(0.066 μg) was injected into a vial (4.9 mL) filled with 320 nmol of COS and 20 nmol of H2S. Because addition of the enzyme solution into the vial
caused disturbance of COS and H2S concentrations, the sampling of the headspace gas was started 5 min after the addition of the enzyme. (B)
●
○
Degradation of ambient level COS by whole cells. Symbols: , with cells; , without cell. Cells in a 10 mL culture containing fully grown T.
thioparus strain THI115 were put in a Petri dish and then placed in a gas-sampling bag (∼3 L). (C) Degradation of ambient level of COS by
recombinant COSase. Symbols: , with COSase; , without COSase. The reaction was started by adding the reaction mixture containing 10 μg of
COSase to a gas-sampling bag as described in the text. The volume of the introduced nitrogen gas that contained 510 pptv (55 pmol) of COS was
2.7 L. The amounts of COS hydrolyzed by COSase was calculated as the difference between the concentrations with and without COSase (indicated
by arrows).
●
○
THI115 was isolated as a thiocyanate (SCN−)-degrading
RESULTS
■
microorganism from activated sludge used for the wastewater
treatment of a coke-oven factory.26 SCN− is common in nature
and is generated by the hydrolysis of glucosinolates in plants
and the detoxication of cyanide in mammals.27,28 In addition to
its natural origins, various industrial processes including coal
gasification have been considered part of its anthropogenic
origin.28 Enzymatic transformation of SCN− is initiated by
thiocyanate hydrolase (SCNase; EC 3.5.5.8), which was first
identified in T. thioparus strain THI115.26 SCNase is a non-
corrin cobalt enzyme that is highly similar to nitrile
hydratases29−31 and hydrolyzes SCN− to ammonia and
COS.26 The resultant COS is metabolized to hydrogen sulfide
(H2S), which is further oxidized to sulfate as the end product to
produce energy.32 Thus, COS hydrolysis in T. thioparus strain
THI115 must be linked to a metabolic pathway of chemo-
lithotrophic energy formation. However, the enzyme respon-
sible for the conversion of COS remains unknown.
Purification and Characterization of COSase. A crude
extract of T. thioparus strain THI115 initially degraded 1200
ppmv of COS at a rate of 0.085 μmol mg−1 min−1, indicating
that the bacterium possesses an enzyme catalyzing the
degradation of COS. This enzyme was purified by sequential
column chromatography and ran as a single band on sodium
dodecyl sulfate−polyacrylamide gel electrophoresis with a
subunit molecular mass of 27 kDa (Figure S1A). Inductively
coupled plasma mass spectrometric analysis revealed that the
enzyme contained one zinc ion per subunit. The N-terminal 35
amino acid sequence was determined to be MEKSNTDAL-
LENNRLYAGGQATHRPGHPGMQPIQP. The decrease of
COS by enzymatic degradation, balanced with the production
of H2S throughout the incubation (Figure 1A), indicated that
the overall reaction by the enzyme was as follows:
COS + H2O → H2S + CO2
We named the enzyme carbonyl sulfide hydrolase (COSase).
Gene Cloning and Expression of COSase in Escher-
ichia coli. Using an oligo-DNA probe designed for the N-
terminal amino acid sequence, 1864 bp, including 660 bp of
open reading frame encoding the COSase gene, was identified
(Figure S2A). Southern blotting analysis using the COSase
gene as a probe detected one band for the T. thioparus strain
THI115 genomic DNA digested with each restriction enzyme.
This result confirmed the presence of only one copy of the
In this study, we describe the purification, gene cloning,
characterization, and crystallography of the novel enzyme
carbonyl sulfide hydrolase (COSase) from T. thioparus strain
THI115. COSase acts over a wide concentration range,
including the ambient level, highlighting the possibility that
COSase is one of the mediators that act as a COS sink.
3819
dx.doi.org/10.1021/ja307735e | J. Am. Chem. Soc. 2013, 135, 3818−3825

Journal of the American Chemical Society
Article
COSase gene in the genome (Figure S3). COSase shares a high
amino acid sequence homology with β-CAs and retains most of
their highly conserved residues, including the Zn-binding site
(Figure S2B). COSase was expressed in E. coli, and the
molecular mass of the purified COSase was determined to be
∼94 kDa by size-exclusion chromatography with multi-angle
light scattering (SEC-MALS), suggesting it has a homotetra-
meric structure (Figure S1B).
Activity of Recombinant COSase on COS, CO2, and
CS2. The COSase expressed in E. coli exhibited a hydrolysis
activity of 18 μmol mg−1 min−1 for COS at a COS
concentration of 26 μM (1400 ppmv), approximately
equivalent to that of the enzyme purified from T. thioparus
strain THI115. COSase activity followed Michaelis−Menten-
type kinetics with Km = 60 μM for COS and kcat = 58 s−1.
COSase exhibited the CO2 hydration activity with a rate of 24
μmol mg−1 min−1, which corresponds to a turnover number
(TON) of 19 s−1 by dimer at a CO2 concentration of 17 mM.
Although COSase shares a high similarity with β-CAs, its CO2
hydration activity was ∼3−4 orders of magnitude smaller than
those of β-CAs.33 Furthermore, we examined COSase’s CS2
hydrolysis activity and found that at a CS2 concentration of 460
μM, H2S was produced at a rate of 3.9 μmol mg−1 min−1 with a
trace amount of COS (<0.1 μmol mg−1 min−1). The CS2
hydrolysis activity of COSase is much smaller than that of the
Acidianus sp. strain A1-3 CS2 hydrolase (∼30 μmol mg−1
min−1).23
Crystal Structure of COSase. To understand the substrate
specificity, the crystal structures of COSase in its native and
SCN−-complexed forms were determined at 1.20 and 1.33 Å
resolution, respectively. SCN− was used as a COS analogue
because of its similar size and shape. The structure of the native
COSase can be superimposed to the structure of the SCN−
complex with a root-mean-squared deviation (RMSD) of 0.16
Å for 213 Cα pairs and 0.38 Å for all atoms. Hereafter, we
describe the structure of COSase complexed with SCN−. A
COSase subunit consists of a five-stranded β-sheet core
(β1−β5) and four flanking α-helices (α1, α2, α3, and α6),
with two additional helices (α4 and α5) protruding from its
core. Pairs of the subunits assemble into tightly packed
homodimers to form an extended ten-stranded β-sheet across
the dimer (Figure 2A). Similar dimeric structures were
commonly found in the crystal structures of β-CAs.34 Two
symmetrically related dimers are weakly associated to form
homotetramers. The tetramerization interface of ∼960 Å2/
subunit is smaller than the dimerization interface (∼3420 Å2/
subunit). Consistent with the SEC-MALS results, the
homotetramer found in the crystal is likely to be the
biologically active structure in solution.
Figure 2. Crystal structure of COSase. (A) The ribbon diagram of the
dimer structure. One subunit is colored blue to red from the N to the
C terminus, and the other one is shown in gray. The zinc ions in the
catalytic site are indicated as dark gray spheres. (B) The catalytic site
of COSase complex with SCN−. The residues originating from one
subunit are shown in cyan and those from the other in pink. The
OMIT electron density map (5σ) for the SCN− and surrounding water
molecules are colored orange. The zinc ion (dark gray sphere) is
coordinated by two cysteine and one histidine residues and a water
molecule (red sphere). Interactions between the zinc ion and its
ligands are shown as red broken lines. Blue broken lines indicate the
hydrogen bonds. Nitrogen, oxygen, and sulfur atoms are shown in
blue, red, and yellow, respectively.
SCN− molecule is located in the catalytic-site pocket and
interacts with a zinc-bound water molecule (O(W1)···N(SCN)
= 3.00 Å), a water molecule hydrogen-bonded to His63′
(meaning His63 of the other subunit in the homodimer,
O(W2)···N(SCN) = 3.01 Å). In addition, the SCN− molecule
is surrounded by the Cys44, Met45, Asp46, Ala68, Met102,
Ile33′, His63′, Leu67′, and Ile82′ residues (Table S1). Since the
catalytic-site pocket of COSase is small and hydrophobic, its
shape is perfectly matched with that of a SCN− molecule.
Structural Comparison with β-CAs. Crystal structures of
β-CAs can be categorized into two subclasses, the plant-type
and Cab-type classes, depending on the conserved residues near
their active-site pockets.35 The residues Gln, Phe, and Tyr are
conserved in the plant-type β-CAs, but variable in the Cab type.
The corresponding residues in COSase are Ile33, His63, and
Ile82, respectively, suggesting that COSase resides in the Cab
type. Structural searches with the program DALI36 revealed that
COSase has a high similarity to the Mycobacterium tuberculosis
β-CA Rv1284 (Protein Data Bank (PDB) code 1YLK)37 and
the Methanobacterium thermoautotrophicum β-CA (PDB code
1G5C)38 with RMSD values of 1.2 and 2.4 Å for 146 Cα pairs,
Structure around the Catalytic Site in SCN−-Com-
plexed COSase. The catalytic site is located at the interface
between the two subunits of the homodimer and largely
sequestered from solvent. As predicted by sequence con-
servation, a zinc ion coordinates residues Cys44, His97, and
Cys100 from the same subunit with coordination distances of
2.28 Å (Zn−Sγ(Cys44)), 2.04 Å (Zn−Nε(His97)), and 2.31 Å
(Zn−Sγ(Cys100)), respectively (Figure 2B). The fourth site is
occupied by a water molecule with a coordination distance of
2.08 Å. The water molecule is stabilized by hydrogen bonding
to the carboxylate group of neighboring Asp46 (2.59 Å), which
formed the salt bridge with Arg48, and the amide nitrogen of
Gly101 (2.98 Å). These interactions in the catalytic site of
COSase are essentially identical to those found in β-CAs.34 An
3820
dx.doi.org/10.1021/ja307735e | J. Am. Chem. Soc. 2013, 135, 3818−3825

Journal of the American Chemical Society
Article
Figure 3. Overlay of the catalytic sites of COSase (pink), the Myc. tuberculosis β-CA Rv1284 (blue), the Myc. tuberculosis β-CA Rv3588c (purple), the
Meth. thermoautotrophicum β-CA (orange) and the Acidianus sp. strain A1−3 CS2 hydrolase (green). The zinc ion and water molecules in COSase
are represented as gray and red spheres, respectively. Interactions between the zinc ion and its ligands are shown as red broken lines. Blue broken
lines indicate the hydrogen bonds. Nitrogen, oxygen, and sulfur atoms are shown in blue, red, and yellow, respectively.
respectively. Since both β-CAs can be classified into Cab type,
COSase appears more closely related to the Cab type than to
the plant type. The notable structural differences from the Cab-
type β-CAs are observed in the β3−β4 connection, especially in
the α5 helix and the following loop (Gly129−Ala135). This
arrangement of the α5 helix in the overall architecture has never
been observed in the structures of β-CAs and may be unique to
COSase. In addition, COSase has an extra loop (Gly150−
Pro158) near the N-terminus of the α6 helix. All of these
additional structures found in COSase are located near its
active-site pocket. As shown in the structure-based alignment
(Figure S2B), the COSase sequence has an insertion in this
region relative to the β-CAs.
mM NaSCN were 70% and 20%, respectively, showing that
SCN− acts as a very weak inhibitor against COSase.
COS Hydrolysis Activity at Ambient Level. Considering
the effect of COSase on the atmosphere, we examined the
degradation of ambient level COS by whole cells of T. thioparus
strain THI115 (Figure 1B). Initially, 600 pptv COS decreased
to 350 pptv in 6 h and then gradually reached less than the
detection limit (280 pptv) in 25 h. A small increase of COS in
the control appeared to originate from the mTC medium
because no COS production was detected in the absence of the
medium. The purified COSase retained its COS hydrolysis
activity of 22 pmol mg−1 min−1, even under the ambient mixing
ratio (Figure 1C). These results suggest that COSase from T.
thioparus strain THI115 functions as one of the microbial sinks
for ambient COS.
The catalytic site of COSase is analogous to those of the Myc.
tuberculosis β-CA Rv128437 and the Meth. thermoautotrophicum
β-CA38 (Figure 3). An SCN− molecule from the buffer solution
is located in the catalytic-site pocket of Myc. tuberculosis β-CA
Rv1284, and its position is almost the same as that in COSase.
On the other hand, an SCN− molecule directly coordinates the
zinc ion in the tetrameric form of Myc. tuberculosis β-CA
Rv3588c which can be categorized into plant type.39 The recent
structural study on the CS2 hydrolase from the Acidianus sp.
strain A1-3 has revealed that the enzyme has the β-CA-type fold
and catalytic-site structure.23 The catalytic sites of COSase and
the CS2 hydrolase are similar to each other, and the notable
differences are observed in the Ala47, Ile33′, and Leu87′, which
are substituted with Glu38, Arg20′, and Phe78′ in the CS2
hydrolase, respectively (Figure 3). These results suggest that
COSase, Cab-type β-CA, and CS2 hydrolase have similar
catalytic mechanisms in spite of their different substrate
specificity.
DISCUSSION
■
The conservation in amino acid sequence and crystal structure
between COSase and β-CAs indicates that COSase belongs to
the β-CA super family. However, COSase is unique in its
substrate specificity. It shows high COS degradation activity but
has little CO2 hydration activity and poor CS2 hydrolysis
activity compared to β-CAs and CS2 hydrolase, respec-
tively.23,33 Unfortunately, we could not determine the kinetic
parameters of COSase for CO2 hydration because of its low
reactivity. But, the fact that the COS hydrolysis activity at a
COS concentration of 26 μM is comparable to the CO2
hydration activity at a CO2 concentration of 17 mM indicates
that COSase is at least moderately specific for COS compared
to CO2.33 Additionally, COSase is very weakly inhibited by
SCN−, one of the inhibitors known against CA.18,19,40 T.
thioparus strain THI115 can use SCN− as its sole energy source.
Previously, we demonstrated SCNase catalyzes the first step of
SCN− degradation to generate energy for the growth. The COS
hydrolyzing activity in the crude T. thioparus strain THI115
extract reaches ∼0.4% of that of the purified COSase. The kcat/
Km value of COSase for COS is much larger than that of
Inhibition of COSase Activity by SCN−. The structure of
SCN−-complexed COSase implied that this molecule may
inhibit COS hydration activity. Because CA is inhibited by
various anion including SCN−,40 we examined the effect by
adding 10 and 100 mM NaSCN to reaction mixture at 18 μM
of COS. The residual activities in the presence of 10 and 100
3821
dx.doi.org/10.1021/ja307735e | J. Am. Chem. Soc. 2013, 135, 3818−3825

Journal of the American Chemical Society
Article
Figure 4. Possible substrate tunnels (yellow) from the active site to the surrounding solvent calculated using CAVER: (A) COSase, (B) Myc.
tuberculosis β-CA Rv1284, (C) Myc. tuberculosis β-CA Rv3588c (tetrameric form), (D) Meth. thermoautotrophicum β-CA, and (E) CS2 hydrolase from
the Acidianus sp. strain A1-3. One subunit is shown in cyan and the other in pink. Dark orange and dark green parts in (E) correspond to the N- and
C-terminals of the adjacent subunit, and light orange and light green parts represent those of the opposite subunit. Dark gray spheres represent the
zinc ions in the active sites.
SCNase for SCN−, and the large kcat/Km value of COSase is
appropriate for the hydrolysis of the COS produced by SCNase
in T. thioparus strain THI115.26 Thus, COSase is competent for
the second step of SCN− degradation in T. thioparus strain
THI115. Additionally, COSase is a novel β-CA family enzyme
that uses COS as a natural substrate.
CAVER.43 These β-CAs have widely opened tunnels toward
their surfaces, suggesting a diffusion-controlled catalytic
mechanism. In contrast, the substrate pathway of COSase is
highly hydrophobic and extremely narrow because of the α5
helix and the extra loop (Gly150−Pro158) near the N-terminus
of the α6 helix. The size of the substrate pathway of COSase is
wide enough for the uptake and release of COS and its
products, H2S and CO2, but would be too narrow for the
The conservation of the environment around the catalytic
zinc center and the hydroxide ligand suggests that the catalytic
mechanism of COSase for COS is conserved with that
proposed for β-CAs and CS2 hydrolase.23,34 Briefly, the
hydroxide ion coordinated to the zinc ion nucleophilically
attacks the carbon atom of the COS molecule to generate the
intermediate having the zinc ion bound both to the hydroxide
oxygen and by the sulfur atom of the COS molecule. The
subsequent release of the oxygen from the zinc ion results in
the formation of the product, CO2. Then, a solvent water
molecule reacted with the sulfur-coordinated zinc ion, resulting
in the regeneration of the active site and release of H2S (Figure
S4). The reason for the substrate specificity of COSase is not
well understood because the structure around the catalytic site
of COSase is highly conserved with those of the Cab-type β-
CAs from Myc. tuberculosis37 and Meth. thermoautotrophicum38
(Figure 3) and because both β-CAs hydrate CO2 efficiently.41,42
We hypothesize that the difference in the CO2 specificity
between COSase and Cab-type β-CAs may be attributed to the
accessibility of the CO2 to the active site. Figure 4 shows the
proposed substrate tunnels of COSase, the β-CAs from Myc.
tuberculosis and Meth. thermoautotrophicum, and the CS2
hydrolase from the Acidianus sp. strain A1-3 calculated using
−
release of the CO2-hydration product HCO3 . The Acidianus
sp. strain A1-3 CS2 hydrolase shows no CA activity despite its
highly conserved catalytic site with β-CA structures.23 Although
the oligomeric states of COSase (tetramer) and the CS2
hydrolase (hexadecamer) are different, their dimeric structures
(Figure 4) and catalytic sites (Figure 3) are similar to each
other. In the case of the CS2 hydrolase, the N-terminal α-helix
of the neighboring subunit in the dimer and the C-terminal
loop from adjacent dimers in the hexadecamer were located
near its catalytic-site pocket. Smeulders et al. have suggested
that its substrate pathway is also hydrophobic and narrowed by
its subunit interfaces among its hexadecameric structure.23
Although the substrate pathways are restricted by different
means, the structure around the substrate pathways could be
responsible for the substrate specificity of both enzymes.
However, we cannot rule out the possibilities of flexibility of the
substrate pathway and other mechanisms controlling the
substrate specificity because the pathway of some plant-type
CAs are narrow but the activity is efficient.34,35 CS2 hydrolysis
activity of COSase is 7.7 times less than that of the Acidianus sp.
strain A1-3 CS2 hydrolase at a CS2 concentration of 460 μM.
3822
dx.doi.org/10.1021/ja307735e | J. Am. Chem. Soc. 2013, 135, 3818−3825

Journal of the American Chemical Society
Article
Figure 5. Phylogenetic tree of COSase and β-CAs. A phylogenetic tree was constructed with MEGA560 using sequences of COSase and β-CAs. β-
CAs were selected from the PDB and used the sequences from UniProt Knowledgebase (UniProtKB) corresponding to β-CAs. β-CAs from the
NCBI conserved domain database (CDD) were used to classify β-CAs into four clades. For each clade, single sequence data from eukarya and
archaea, and those of bacteria that represent clades A, B, and C at the phylum level and clade D at the genus level were used to construct the tree
when the sequence data of β-CAs was unavailable in PDB. Sequence data of Halothiobacillus neapolitanus deposited in PDB were not included due to
the low homology with COSase. The transit peptide sequence of β-CA from Pisum sativum was also not included. The accession numbers of the β-
CAs used here are indicated by its NCBI GI number or UniProtKB number and appear in parentheses. The COSase, CS2 hydrolase, and β-CAs of
clade D that were deposited to PDB are shown in bold.
The reason for this difference is not known. We note that Ile33
and Ala47 of COSase are substituted by Arg20 and Glu38 in
CS2 hydrolase (Figure 3). These residues increase the
hydrophilicity of the corresponding area in the substrate pocket
of CS2 hydrolase and may affect its substrate specificity.
Detailed studies on the substrate recognition mechanism of
COSase are the subject for further investigations.
widely distributed in clade D. β-CA is suggested to be an
ancient enzyme.45 Sulfur isotope analysis indicates that the
mixing ratio of COS in the present troposphere is 10,000-times
lower than that of the Archean eon more than 3 billion years
ago.47 Thus, it is important for COSase and/or COSase-related
enzymes to harbor a higher specificity and catalytic efficiency to
be a global sink of tropospheric COS in the present day. In
addition, COSase contributes to biological energy production
by a SCN− degradation pathway and is one of sulfur’s
biogeochemical processes.
Both the Km and kcat of COSase for COS hydrolysis are ∼2
orders of magnitude smaller than those of β-CAs for CO2
hydration.41,42 As a result, the overall catalytic efficiency of
COSase, kcat/Km = 9.6 × 105 s−1 M−1, is comparable to those of
β-CAs. COSase shows the highest kcat/Km value among widely
distributed enzymes with COS-degrading activity (Table 1).
Previously, we showed that phylogenetically diverse soil
bacteria belonging to Actinomycete, such as the genus
Mycobacterium and the genus Williamsia, and to Proteobacteria,
such as the genus Cupriavidus, exhibited COS-degrading
activity.44 Prokaryotic β-CAs are classified into four subgroups
of clade: A, B, C, and D (Figure 5).33,45 Both the T. thioparus
strain THI115 COSase and the Acidianus sp. strain A1-3 CS2
hydrolase as well as two structurally similar β-CAs from Myc.
tuberculosis and Meth. thermoautotrophicum belong to the clade
D. We showed that soil bacteria in the genus Mycobacterium
and filamentous fungus in the genus Fusarium degraded COS at
an ambient mixing ratio.44,46 The enzymes with higher
specificities for COS than CO2, such as COSase, might be
CONCLUSIONS
■
In summary, COSase, a novel enzyme hydrolyzing COS to H2S
and CO2, has been purified from T. thioparus strain THI115, a
chemolithoautotrophic bacterium degrading SCN− to COS in
its energy producing metabolism, and characterized structurally
as well as biochemically. COSase exists as a homotetramer in
which two pairs of subunits formed tightly associated dimers.
Each subunit possesses a zinc ion in the active site. Although
COSase shares high amino acid sequence homology and
conserves structure including the zinc catalytic center as found
in the β-CAs, CA activity is negligible compared to those of β-
CAs. The difference in the substrate specificity could be due to
the accessibility and hydrophobicity of the substrate pathway
from the surface of molecule to the active site. COSase had Km
3823
dx.doi.org/10.1021/ja307735e | J. Am. Chem. Soc. 2013, 135, 3818−3825

Journal of the American Chemical Society
Article
mixing 1.2 μL of protein solution (6 mg·mL−1) with 1.2 μL of
reservoir solution (1.2 M (NH4)2SO4, 0.2 M NaCl, 30% glycerol, 0.1
M Tris-HCl, pH 8.5). Crystals of COSase with SCN− were obtained
by addition of 0.01 M NaSCN to the crystallization drop containing
native crystals.
X-ray diffraction data were collected at 95 K using an ADSC/CCD
detector system (ADSC Quantum 315r and Quantum 4R) on the BL-
5A and BL-6A beamlines at the Photon Factory (Tsukuba, Japan).
Diffraction data were processed using the program HKL2000.54 Data
collection statistics are summarized in Table S2.
Structure Determination and Refinement. The structures of
COSase and the COSase/thiocyanate complex were solved by a
molecular replacement method with the program Molrep55 using the
partial structures of β-CA from Myc. tuberculosis (PDB code 1YLK,
residues 23−99) and the native COSase as a search model,
respectively. After a rigid-body refinement using REFMAC,56 the
model was improved by rebuilding in COOT57 and refining in
SHELXL.58 To avoid over fitting of the diffraction data, a free R-factor
was monitored by setting aside 5% of the reflections as a test set.59
Finally, the structures were refined to Rwork/Rfree = 0.135/0.170 and
0.156/0.204 for COSase and COSase/thiocyanate, respectively (Table
S2). The atomic coordinates and structure factors for COSase and the
COSase/thiocyanate complex were deposited in the PDB under the
accession codes 3VQJ and 3VRK, respectively.
= 60 μM for COS, Vmax = 74 μmol mg−1 min−1, and kcat/Km =
9.6 × 105 s−1 M−1, showing that COSase is one of the most
efficient COS-degrading enzymes and suggesting it as a
potential mediator to degrade low concentrations of COS,
such as that in the atmosphere.
MATERIALS AND METHODS
More details are given in the Supporting Information.
Purification of COSase. T. thioparus strain THI115 (NBRC
105750)26 was harvested by centrifugation, and the cell pellets were
disrupted by sonication. COSase was purified from the cell-free extract
by column chromatography.
■
Measurement of COSase Activity. The reaction was started by
the addition of 200 μL of the enzyme into a vial (4.9 mL) containing
COS and incubated at 30 °C at a static condition. COS and H2S in the
headspace gas was measured by a gas chromatograph equipped with a
flame photometric detector (FPD-GC, GC-14B, Shimadzu, Kyoto,
Japan). H2S in the reaction mixture was measured by a colorimetric
method.48 The amounts of COS in the solution and H2S in the
headspace gas or the solution were estimated based on the solubility of
these compounds.49−51 COSase activity is expressed as micromoles of
H2S produced per milligram of protein per minute. Also, the TONs
are calculated by dimer.
Degradation of Ambient Level COS by Whole Cells. Cells of
T. thioparus strain THI115 (0.15 g as wet weight) were washed three
times and resuspended in a basal salt mTC medium. The resultant cell
suspension corresponding to 10 mL-equivalent of culture was filtered
with glass fiber filter (GF-75, Advantec Toyo Kaisha, Tokyo, Japan)
and then placed on three pieces of wet glass microfiber filters (GF/C
circle 90 mm o.d., Whatman, Kent, UK) in 4 mL of mTC medium on
the bottom of a Petri dish (9.3 cm i.d.). The dish was placed in a 5 L
gas-sampling bag (aluminized polyethylene bag, GL Sciences, Tokyo,
Japan) and then exposed to ∼3 L of air. A 300 mL air sample was
taken from the bag, loaded onto a pre-cryo column under liquid
oxygen, and then introduced to a FPD-GC, as described previously.44
Measurement of COSase Activity at Ambient Mixing Ratio
of COS. The gas-sampling bag was used for this experiment as
described above except N2 was used instead of air, and pure COS was
introduced to make up the final mixing ratio to ∼500 pptv. Ten
micrograms of COSase in 10 mL of 50 mM Tris-HCl, pH 8.5, was
injected onto a glass dish (8.5 cm i.d.) placed in the bag with a syringe
through a butyl rubber stopper fitted to the bag. COS was measured
by FPD-GC after cryophilic concentration as described above.
Measurement of CO2 Hydration Activity. The CA activity was
measured using stopped-flow spectrophotometry (SX-20, Applied
Photophysics, Surrey, UK).52 Enzyme solution contained 50 mM
Hepes, pH 7.5 as buffer and 0.2 mM phenol red as indicator, and 200
mM Na2SO4 to maintain the ionic strength. The enzyme solution and
CO2-saturated water were mixed 1:1 at 25 °C, and the absorbance at
557 nm was measured. CO2 hydration activity is expressed as μmol of
CO2 hydrated per mg of protein per minute. Also, the TONs are
calculated by dimer.
Measurement of CS2 Hydrolysis Activity. The activity was
estimated according to the procedure employed to measurement of
COSase activity, except that CS2 gas (10,200 ppmv in N2 as a balanced
gas, Nissan Tanaka, Saitama, Japan) was used instead of COS. The
initial CS2 concentration in solution was estimated to be 460 μM
based on the solubility of 0.047 M atm−1 at 30 °C.53 After a 20-min
incubation, the amount of COS and H2S in a headspace gas was
measured by FPD-GC.
Expression and Purification of Recombinant COSase. The
COSase gene was screened from the genomic library and subcloned
into a GST fusion protein expression vector. COSase was expressed in
E. coli and purified by affinity chromatography. The nucleotide
sequence data for COSase gene has been submitted to the DDBJ
database and will appear in the DDBJ, EMBL, and GenBank
nucleotide sequence databases with the accession no. AB583934.
Crystallization and Data Collection. Crystals of COSase were
grown at 20 °C using the hanging-drop vapor diffusion method by
ASSOCIATED CONTENT
* Supporting Information
Details of materials and methods, Tables S1−S3, and Figures
S1−S4. This material is available free of charge via the Internet
at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
katayama@cc.tuat.ac.jp
■
Present Addresses
⊗T.O.: Research Center for Science and Technology, Tokyo
University of Agriculture and Technology, Fuchu, Tokyo 183-
8509, Japan
▽M.S.: Department of Planning, Kamitsuga Agriculture
Promotion Office, Kanuma, Tochigi 322-0023, Japan
○Y.N.: JST, ERATO, Suematsu Gas Biology Project, Shinjuku,
Tokyo 160-8582, Japan
▲H.K.: Department of Life Sciences, Graduate School of Life
Sciences, Tohoku University, Sendai, Miyagi 980-8577, Japan
¶A.O.: Taitec Co., Ltd, Koshigaya, Saitama 343-0822, Japan
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported in part by a Grant-in Aid for Scientific
Research from the Ministry of Education, Science, Sports and
Culture of Japan (KAKENHI B 18310020 (to Y.K.) and
KAKENHI C 22550147 (to K.N.)). We are grateful to the
beamline assistants at the Photon Factory for data collection at
BL-5A and BL-6A.
REFERENCES
■
(1) Chin, M.; Davis, D. D. J. Geophys. Res.: Atmos. 1995, 100, 8993−
9005.
(2) Andreae, M. O.; Crutzen, P. J. Science 1997, 276, 1052−1058.
(3) Crutzen, P. J. Geophys. Res. Lett. 1976, 3, 73−76.
(4) Baldwin, B.; Pollack, J. B.; Summers, A.; Toon, O. B.; Sagan, C.;
Van Camp, W. Nature 1976, 263, 551−555.
(5) Turco, R. P.; Whitten, R. C.; Toon, O. B.; Pollack, J. B.; Hamill,
P. Nature 1980, 283, 283−286.
3824
dx.doi.org/10.1021/ja307735e | J. Am. Chem. Soc. 2013, 135, 3818−3825

Journal of the American Chemical Society
Article
(6) Solomon, S.; Sanders, R. W.; Garcia, R. R.; Keys, J. G. Nature
1993, 363, 245−248.
(39) Covarrubias, A. S.; Bergfors, T.; Jones, T. A.; Hogbom, M. J.
̈
Biol. Chem. 2006, 281, 4993−4999.
(40) De Simone, G.; Supuran, C. T. J. Inorg. Biochem. 2012, 111,
117−129.
(7) Bruhl, C.; Lelieveld, J.; Crutzen, P. J.; Tost, H. Atmos. Chem. Phys.
̈
2012, 12, 1239−1253.
(41) Minakuchi, T.; Nishimori, I.; Vullo, D.; Scozzafava, A.; Supuran,
C. T. J. Med. Chem. 2009, 52, 2226−2232.
(8) Aydin, M.; Williams, M. B.; Tatum, C.; Saltzman, E. S. Atmos.
Chem. Phys. 2008, 8, 7533−7542.
(42) Smith, K. S.; Ferry, J. G. J. Bacteriol. 1999, 181, 6247−6253.
(9) Watts, S. F. Atmos. Environ. 2000, 34, 761−779.
(10) Kesselmeier, J.; Teusch, N.; Kuhn, U. J. Geophys. Res.: Atmos.
1999, 104, 11577−11584.
̌ ́ ̌ ̌ ́ ̌
(43) Petrek, M.; Otyepka, M.; Banas, P.; Kosinova, P.; Koca, J.;
Damborsky, J. BMC Bioinf. 2006, 7, 316.
́
(44) Kato, H.; Saito, M.; Nagahata, Y.; Katayama, Y. Microbiology
2008, 154, 249−255.
(11) Saito, M.; Honna, T.; Kanagawa, T.; Katayama, Y. Microbes
Environ. 2002, 17, 32−38.
(45) Smith, K. S.; Jakubzick, C.; Whittam, T. S.; Ferry, J. G. Proc.
Natl. Acad. Sci. U.S.A. 1999, 96, 15184−15189.
(46) Li, X. S.; Sato, T.; Ooiwa, Y.; Kusumi, A.; Gu, J.-D.; Katayama,
Y. Microb. Ecol. 2010, 60, 96−104.
(12) Chengelis, C. P.; Neal, R. A. Biochem. Biophys. Res. Commun.
1979, 90, 993−999.
(13) Miller, A. G.; Espie, G. S.; Canvin, D. T. Plant Physiol. 1989, 90,
1221−1231.
(47) Ueno, Y.; Johnson, M. S.; Danielache, S. O.; Eskebjerg, C.;
Pandey, A.; Yoshida, N. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 14784−
14789.
(48) Chae, Y. M.; Tabatabai, M. A. Anal. Lett. 1983, 16, 1197−1206.
(49) Wilhelm, E.; Battino, R.; Wilcock, R. J. Chem. Rev. 1977, 77,
219−262.
(50) De Bruyn, W. J.; Swartz, E.; Hu, J. H.; Shorter, J. A.; Davidovits,
P.; Worsnop, D. R.; Zahniser, M. S.; Kolb, C. E. J. Geophys. Res.: Atmos.
1995, 100, 7245−7251.
(51) Dean, J. A. Electrolytes, Electromotive Force, and Chemical
Equilibrium: Lange’s Handbook of Chemistry, 15th ed.; McGraw-Hill:
New York, 1999; pp 8.1−8.169.
(52) Khalifah, R. G. J. Biol. Chem. 1971, 246, 2561−2573.
(53) Staudinger, J.; Roberts, P. V. Chemosphere 2001, 44, 561−576.
(54) Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307−
326.
(55) Vagin, A.; Teplyakov, A. J. Appl. Crystallogr. 1997, 30, 1022−
1025.
(14) Kesselmeier, J.; Merk, L. Biogeochemistry 1993, 23, 47−59.
(15) Protoschill-Krebs, G.; Wilhelm, C.; Kesselmeier, J. Bot. Acta
1995, 108, 445−448.
(16) Protoschill-Krebs, G.; Wilhelm, C.; Kesselmeier, J. Atmos.
Environ. 1996, 30, 3151−3156.
(17) Haritos, V. S.; Dojchinov, G. Comp. Biochem. Physiol., Part C:
Toxicol. Pharmacol. 2005, 140, 139−147.
(18) Supuran, C. T. Nat. Rev. Drug Discovery 2008, 7, 168−181.
(19) Alterio, V.; Di Fiore, A.; D’Ambrosio, K.; Supuran, C. T.; De
Simone, G. Chem. Rev. 2012, 112, 4421−4468.
(20) Seefeldt, L. C.; Rasche, M. E.; Ensign, S. A. Biochemistry 1995,
34, 5382−5389.
(21) Ensign, S. A. Biochemistry 1995, 34, 5372−5381.
(22) Lorimer, G. H.; Pierce, J. J. Biol. Chem. 1989, 264, 2764−2772.
(23) Smeulders, M. J.; Barends, T. R. M.; Pol, A.; Scherer, A.;
Zandvoort, M. H.; Udvarhelyi, A.; Khadem, A. F.; Menzel, A.;
Hermans, J.; Shoeman, R. L.; Wessels, H. J. C. T.; van den Heuvel, L.
P.; Russ, L.; Schlichting, I.; Jetten, M. S. M.; Op den Camp, H. J. M.
Nature 2011, 478, 412−416.
(56) Murshudov, G. N.; Vagin, A. A.; Dodson, E. J. Acta Crystallogr.,
Sect. D: Biol. Crystallogr. 1997, 53, 240−255.
(24) Huber, H.; Stetter, K. O. Genus II. Acidianus Segerer, Neuner,
Kristjansson and Stetter 1986, 561VP. In Bergey’s Manual of Systematic
Bacteriology, 2nd ed.; Garrity, G. M., Boone, D. R., Castenholz, R. W.,
Eds.; Springer: New York, 2001; Vol.1, pp 202−204.
(25) Kelly, D. P.; Wood, A. P.; Stackebrandt, E. Genus II. Thiobacillus
Beijerinck 1904b, 597AL. In Bergey’s Manual of Systematic Bacteriology,
2nd ed.; Garrity, G. M., Brenner, D. J., Krieg, N. R., Staley, J. T., Eds.;
Springer: New York, 2005; Vol. 2, Part C, pp 764−769.
(26) Katayama, Y.; Narahara, Y.; Inoue, Y.; Amano, F.; Kanagawa, T.;
Kuraishi, H. J. Biol. Chem. 1992, 267, 9170−9175.
(27) Fahey, J. W.; Zalcmann, A. T.; Talalay, P. Phytochemistry 2001,
56, 5−51.
(57) Emsley, P.; Cowtan, K. Acta Crystallogr., Sect. D: Biol. Crystallogr.
2004, 60, 2126−2132.
(58) Sheldrick, G. M. Acta Crystallogr., Sect. A: Found. Crystallogr.
2008, 64, 112−122.
(59) Brunger, A. T. Nature 1992, 355, 472−475.
̈
(60) Tamura, K.; Peterson, D.; Peterson, N.; Stecher, G.; Nei, M.;
Kumar, S. Mol. Biol. Evol. 2011, 28, 2731−2739.
(28) Valdes
34, 9−23.
́
, M. G.; Díaz-García, M. E. Crit. Rev. Anal. Chem. 2004,
(29) Katayama, Y.; Hashimoto, K.; Nakayama, H.; Mino, H.; Nojiri,
M.; Ono, T.; Nyunoya, H.; Yohda, M.; Takio, K.; Odaka, M. J. Am.
Chem. Soc. 2006, 128, 728−729.
(30) Katayama, Y.; Matsushita, Y.; Kaneko, M.; Kondo, M.; Mizuno,
T.; Nyunoya, H. J. Bacteriol. 1998, 180, 2583−2589.
(31) Arakawa, T.; Kawano, Y.; Kataoka, S.; Katayama, Y.; Kamiya, N.;
Yohda, M.; Odaka, M. J. Mol. Biol. 2007, 366, 1497−1509.
(32) Kim, S.-J.; Katayama, Y. Water Res. 2000, 34, 2887−2894.
(33) Smith, K. S.; Ferry, J. G. FEMS Microbiol. Rev. 2000, 24, 335−
366.
(34) Rowlett, R. S. Biochim. Biophys. Acta, Proteins Proteomics 2010,
1804, 362−373.
(35) Kimber, M. S.; Pai, E. F. EMBO J. 2000, 19, 1407−1418.
(36) Holm, L.; Kaariainen, S.; Rosenstrom, P.; Schenkel, A.
̈
̈
̈
̈
Bioinformatics 2008, 24, 2780−2781.
(37) Covarrubias, A. S.; Larsson, A. M.; Hogbom, M.; Lindberg, J.;
̈
Bergfors, T.; Bjorkelid, C.; Mowbray, S. L.; Unge, T.; Jones, T. A. J.
̈
Biol. Chem. 2005, 280, 18782−18789.
(38) Strop, P.; Smith, K. S.; Iverson, T. M.; Ferry, J. G.; Rees, D. C. J.
Biol. Chem. 2001, 276, 10299−10305.
3825
dx.doi.org/10.1021/ja307735e | J. Am. Chem. Soc. 2013, 135, 3818−3825