Báo cáo khoa học: A new sulfurtransferase from the hyperthermophilic bacterium Aquifex aeolicus Being single is not so simple when temperature gets high potx

A new sulfurtransferase from the hyperthermophilic
bacterium Aquifex aeolicus
Being single is not so simple when temperature gets high
Marie-Ce
´
cile Giuliani
1
, Pascale Tron
1
, Gise
`
le Leroy
1
, Corinne Aubert
1
, Patrick Tauc
2
and
Marie-The
´
re
`
se Giudici-Orticoni
1
1 Laboratoire de Bioe
´
nerge
´
tique et Inge
´
nierie des Prote

´
ines (BIP), IBSM-CNRS, Marseille, France
2 Laboratoire de Biotechnologie et de Pharmacologie Ge
´
ne
´
tique Applique
´
e – ENS-CACHAN, Cachan, France
Sulfur adds considerable functionality to a wide variety
of biomolecules because of its unique properties: its
chemical bonds are both easily made and easily bro-
ken, and sulfur serves as both an electrophile (e.g.
in disulfides) and a nucleophile (e.g. as thiol) [1–3].
For incorporation into biomolecules, sulfur must be
reduced and ⁄ or activated, and sulfate or polysulfides
are substrates for reductases that are widespread in
nature. The activated form of sulfur, the ‘sulfane sul-
fur’ (R-S-SH) was suggested as the biologically rele-
vant active sulfur species in the early 1980s. Sulfane
sulfur is produced enzymatically with the IscS protein,
the SufS protein and rhodanese being the most promi-
nent biocatalysts [3].
Rhodaneses (thiosulfate : cyanide sulfurtransferase
or TSTs) are widespread enzymes that catalyse the
transfer of a sulfane sulfur. Thiosulfate is generally
used as a substrate for rhodaneses in vitro assays, and
cyanide is used as a sulfur acceptor to regenerate the
covalent catalytic cysteinyl residue (Eqn 1a,b):
SSO

2À
3
þ Rho-SH ! SO
2À
3
þ Rho-S-SH ð1aÞ
Rho-S-SH þ CN
À
! Rho-SH þ SCN
À
ð1bÞ
In addition to cyanide, other thiophilic acceptor com-
pounds are also acceptable [4,5]. Despite numerous
studies, the physiological role of rhodaneses remains
unclear and is still widely debated because the in vivo
substrate has not been identified [6–13]. The difficulties
in establishing the in vivo functions of rhodaneses lie in
the multiplicity of rhodanese modules and rhodanese
Keywords
Aquifex aeolicus; hyperthermophile;
oligomerization; sulfurtranferase;
thermostability
Correspondence
M T. Giudici-Orticoni, Laboratoire de
Bioe
´
nerge
´
tique et Inge
´

nierie des Prote
´
ines,
IBSM-CNRS, 31 chemin Joseph Aiguier,
13402 Marseille cedex 20, France
Fax: +33 4 91 16 45 78
Tel: +33 4 91 16 45 50
E-mail:
(Received 20 April 2007, revised 6 July
2007, accepted 12 July 2007)
doi:10.1111/j.1742-4658.2007.05985.x
Sulfur is a functionally important element of living matter. Rhodanese is
involved in the enzymatic production of the sulfane sulfur which has been
suggested as the biological relevant active sulfur species. Rhodanese
domains are ubiquitous structural modules occurring in the three major
evolutionary phyla. We characterized a new single-domain rhodanese with
a thiosulfate : cyanide transferase activity, Aq-477. Aq-477 can also use
tetrathionate and polysulfide. Thermoactivity and thermostability studies
show that in solution Aquifex sulfurtranferase exists in equilibrium
between monomers, dimers and tetramers, shifting to the tetrameric state
in the presence of substrate. We show that oligomerization is important
for thermostability and thermoactivity. This is the first characterization of
a sulfurtransferase from a hyperthermophilic bacterium, which moreover
presents a tetrameric organization. Oligomeric Aq-477 may have been
selected in hyperthermophiles because subunit association provides extra
stabilization.
Abbreviations
BN, Blu native gel; MST, mercaptopyruvate sulfurtransferase; rec, recombinant; Rho, rhodanese; SR, sulfur reductase; ST, sulfurtransferase;
Sud, sulfide dehydrogenase; TST, thiosulfate sulfurtransferase.
4572 FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS

activities. However, the role of persulfide cysteine
at the catalytic site has been demonstrated [14–17]. A
typical feature of the rhodanese superfamily is the
modular structure of its various members [18]. Rhoda-
neses or sulfurtransferases are ubiquitous enzymes
found in many organisms from all three domains of
life. Even though the discovery of the hyperthermo-
philes has important ramifications, not only in micro-
bial physiology and evolution, but also in
biotechnology, no rhodanese has been characterized
from these extremophilic organisms.
The prototypic enzyme, bovine liver rhodanese, con-
sists of an N-terminal inactive rhodanese module (the
catalytic cysteinyl residue is replaced) and a C-terminal
catalytic module, each encompassing about 120 amino
acids [19,20]. This domain organization is also typical
for many rhodanese sequences distributed in all king-
doms. These two domains show weak sequence simi-
larity [18]. In addition to the two-domain rhodaneses,
single-domain versions are known [5,21–23], with
Escherichia coli GlpE protein as the prototype [5,16].
Characterization of single-domain rhodanese indicates
that the N-terminal domain in two-domain rhodaneses
is not essential for catalysis. Rhodanese modules may
also be involved in processes beside sulfur transfer.
An interesting example is the rhodanese-homologous
domain of the E. coli YbbB protein, which is responsi-
ble for the exchange of sulfur for selenium in 2-thio-
uridine in vivo [12].
In addition, characterization of sulfide dehydroge-

nase (Sud) from a mesophilic bacterium, Wolinella
succinogenes revealed for the first time the direct inter-
vention of rhodanese in energetic sulfur metabolism,
because this protein is the sulfur donor for the termi-
nal acceptor of respiratory chain sulfur reductase in
W. succinogenes.
Microorganisms with the remarkable property of
growing at temperatures near and above 100 °C have
been isolated from shallow submarine and deep-sea
volcanic environments over the last 20 years. The
majority of these hyperthermophilic microorganisms
are archaea and they are considered to represent the
most slowly evolving forms of life [24–26]. Numerous
hyperthermophilic archaea are known, but very few
hyperthermophilic bacteria have been discovered to
date. Most known hyperthermophilic bacteria are
members of the genus Aquifex and have an optimal
growth temperature of 85 °C [24,27]. Aquifex is a
hyperthermophilic, hydrogen-oxidizing, microaero-
philic, obligate chemolithoautotrophic bacterium. It
obtains energy for growth from hydrogen, oxygen and
sulfur (thiosulfate or elemental sulfur), and uses the
reductive tricarboxylic acid cycle to fix CO
2
. Stimulated
by the exceptional phylogenetic and physiological prop-
erties of A. aeolicus, as well as by its potential as a
source of extremely stable enzymes, we undertook sev-
eral studies on the energetic metabolism of this organ-
ism, whose genome has been completely sequenced [28].

In particular, we studied its hydrogen ⁄ sulfur metabo-
lism [29,30]. Two rhodanese-coding genes, rhdA1 and
rhdA2 are annotated in the A. aeolicus genome. Both
belong to the two-domain family of rhodaneses.
Here, we describe the identification, purification and
biochemical and biophysical characterization of a new
single-domain rhodanese in A. aeolicus, which was not
identified by annotation of the genome. The results shed
light on some particularities of the protein which may
be linked to the need for extremophiles and their macro-
molecules to develop molecular mechanisms adapted to
extreme physicochemical conditions. Its possible meta-
bolic roles in A. aeolicus are also discussed.
Results
Evidence for sulfurtransferase (ST) activity in
A. aeolicus
ST activity was routinely measured under an argon
atmosphere in an assay mixture containing thiosulfate
and cyanide. The amount of SCN
–
produced was rep-
resentative of the catalysis. After centrifugation of the
cell extract, ST activity was found to be associated
with the soluble fraction, whereas the membrane
fraction was inactive. After cellular separation, 90% of
the activity was present in the cytoplasmic extract and
around 10% in the periplasmic extract. As no lactate
dehydrogenase activity could be detected in the peri-
plamic extract, we propose that the enzymes required
for sulfur transfer are present in A. aeolicus periplas-

mic and cytoplasmic spaces.
We obtained evidence for the presence of sulfur-
transferase activity in A. aeolicus grown on elemental
sulfur or thiosulfate. However, extracts of cells grown
with elemental sulfur showed a specific sulfurtrans-
ferase activity 5.5· the specific activity measured with
cells grown with thiosulfate. We thus decided to use
cells cultured on H
2
⁄ S° medium to work on the ST
enzymes.
Purification of a new ST
A protein with ST activity was purified by Q-Sepha-
rose, HTP and Superdex 200 gel-filtration chromato-
graphy from the cytoplasmic fraction of A. aeolicus
grown on H
2
⁄ S° medium (Table 1). At the final step,
ST activity was detected in different fractions. The
M C. Giuliani et al. Oligomeric sulfurtransferase from Aquifex aeolicus
FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS 4573
calibration curve shows that these fractions corres-
pond roughly to a molecular mass of 50 to 15 kDa.
We divided this activity peak into three fractions
corresponding to approximately 15, 25 and 50 kDa
(Fig. 1). N-Terminal sequence determination of these
fractions led to the identification of these peaks as
containing the same protein. A search for this
sequence in the Aquifex proteins database reveals one
protein, Aq-477. This protein, encoded by aq477, was

annotated as a protein of unknown function with a
molecular mass of 12 804 Da as deduced from the
amino acid sequence. Mass spectra of each fraction
demonstrated one major protein with molecular mass
of 12 810 Da. This suggests a possible oligomerization
state of the protein in the dimer and tetramer, in line
with the masses corresponding to the different activity
fractions detected. As the protein contains only one
cysteine residue, Cys69, a disulfide bridge may (or
may not) be formed between two monomers. MS data
demonstrate the absence, in the samples, of one or
more possible interdisulfide bridges possibly involved
in the oligomerization.
Aq-477 is a single-domain ST
The product of aq477 purified from A. aeolicus,is
able to transfer sulfur from thiosulfate to cyanide.
According to the sequence deduced from the gene,
no signal peptide is detected, which is in agreement
with the purification of the protein from the cyto-
plasm of A. aeolicus. aq477 is flanked by genes panD
and aq478. Because a single base pair separates the
termination and initiation codons for panD and
aq477, they appear to be organized as an operon.
panD encodes an aspartate decarboxylase which cata-
lyses the decarboxilation of aspartate to produce
b-alanine, a precursor of coenzyme A. aq478 encodes
a protein similar to proteins involved in signal trans-
duction mechanisms. The product of aq477 was
annotated as a hypothetical protein related to a
member of COG0607P, which regroup 168 rhoda-

nese-related sulfurtransferases. All these homologues
belong to a a ⁄ b-fold protein domain found dupli-
cated in the rhodanese proteins. Each protein from
this family contains at least one cysteine residue,
which has been found to be essential for the protein’s
function [18]. Unlike classical two-domain rhodane-
ses, Aq-477 is composed of a single-domain rhoda-
nese fold, the catalytic domain as it contains the
characteristic catalytic cysteine. Few single-domain
rhodaneses have been characterized in detail. The pri-
mary sequence of Aq-477 shows only slight similarity
with other single-domain proteins of known 3D
structure, i.e. 21% homology with GlpE from E. coli,
25% with Sud from W. succinogenes and 27% with
a TTHA0613 ORF from T. thermophilus [21,31,32].
However, structural alignment (Fig. 2) shows the
same global fold for all single-domain STs with a
typical a ⁄ b topology and the extension and location
of the regular secondary structure elements approxi-
mately coincide in all these proteins. However, some
differences are found. The major difference is the
presence of an extra a helix in Sud protein. More-
over, the mesophilic ST (Sud and GlpE) have an
insertion between b3 and a4. In addition, in Aq-477,
the prediction proposes a shift of b5 through the
C-terminus. The a6 helix is absent in thermo ⁄ hyper-
thermophiles enzymes.
Despite the low overall sequence identity, a few
residues are conserved in b strands b2 and b4, which
Table 1. Yields and enrichments of the ST activity. One unit of ST

activity corresponds to the uptake of 1 lmol of SCNÆmin
)1
.
Preparation
Sulfurtransferase
activity
Yield (%)
Specific
activity
(UÆmg
)1
)
Total
activity
(U)
Soluble proteins 134 35,160
Q Sepharose (300 m
M NaCl) 306 7800 22
HTP (300 m
M phosphate) 550 3000 8.5
Superdex 200 850 1000 2.8
Ve/Vo
1, 1, 2, 2,
log (PM)
4,
4,
4,
4,
4,
5,

5,
5,
Elution volume (ml)
14 15.5 16.9
123
0
50
100
150
200
mAU
( )
Activity

( ) u/ml
100
200
300
Fig. 1. Size-exclusion chromatography of wild-type Aq-477. An
S200 column (1 · 30 cm) was equilibrated in 100 m
M Tris ⁄ HCl,
50 m
M NaCl, pH 7.6 at 20 °C. The protein (200 lLat4mgÆmL
)1
)
was detected by its absorbance (d) and its activity (m). (Inset) Cali-
bration curve in 100 m
M Tris ⁄ HCl, 50 mM NaCl, pH 7.6, flow rate
0.3 mLÆmin
)1

, sample volume 200 lLat2mgÆmL
)1
of rusticyanine
(17 kDa); dihemic cytochrome c (21 kDa); ovalbumin (43 kDa); alco-
hol dehydrogenase (150 kDa).
Oligomeric sulfurtransferase from Aquifex aeolicus M C. Giuliani et al.
4574 FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS
constitute the structural core of the protein. In b2, the
sequence XDXR (X being hydrophobic residues) is
conserved. In b4, a set of four hydrophobic residues
preceding the position occupied by cysteine is
conserved. These residues are also conserved in two-
domain rhodaneses. We can conclude that these
residues play an important role in the global folding
processes of the mesophilic, themophilic and hyper-
thermophilic proteins. The potential active site is
located between a central b strand and a a helix. In
Aq-477, Cys69 is the first residue of loop 69–74. More-
over, the positive charges (R30, R70, R74) found in
the substrate-binding pocket for the negative polysul-
fide chain are conserved. In Sud protein R46 and E50
are involved in substrate binding [31]. These residues
are conserved in Aq-477 (R30, E34).
Various proteins can catalyse sulfur transfer. In
addition to the two classical two-domain rhodanese
proteins, annotated as rhdA1 and rhdA2 in the A. aeo-
licus genome, a fourth gene exhibiting sequence simi-
larity to the rhodaneses family [18] is detected in the
genome using the catalytic motif for ST as query for
a BLAST search. This other protein, Aq-1599, is

annotated as a protein of unknown function in the
A. aeolicus genome. We identified it as single domain
ST. Aq-1599 is predicted to be periplasmic.
Cloning, heterologous production and purifica-
tion of recombinant Aq-477 from A. aeolicus
The aq477 gene encoding A. aeolicus ST, was amplified
by PCR, inserted into a pET22 expression vector and
expressed in E. coli BL21-CodonPlus (DE3)-RIPL
strain. After induction, a high level of protein was
detected in the soluble extract. Ten milligrams of solu-
ble protein were obtained from 1 L of culture after
two purification steps, as described in the Experimental
procedures. The N-terminal sequence was identical to
the native protein purified from A. aeolicus. The recon-
structed mass spectra of recombinant Aq-477 gave
only one peak corresponding to a molecular mass of
12 813 Da. This is consistent with the molecular mass
calculated from the sequence and supports this protein
corresponding to the mature enzyme. Moreover, it
demonstrates the absence of interdisulfide bridges in
the enzyme.
Physicochemical and catalytic properties
UV–Vis spectroscopy performed in the oxidized and
reduced states showed no signal except at 280 nm, cor-
responding to the protein absorption. Aq-477 does not
contain prosthetic groups or heavy metal ions.
Fig. 2. Structural alignment of Aq-477 from A. aeolicus, GlpE from E. coli, Sud from W. succinogenes and TTHA0613 from T. termophilus
sequences Secondary structures are indicated, based on the 3D prediction ( for A. aeolicus
and from the 3D structure for Glpe, Sud and TTHA0613. Residues involved in the a helix are boxed and those in b sheet are underlined. Con-
served residues involved in substrate binding are in bold. The arrow indicates the active-site cysteine.

M C. Giuliani et al. Oligomeric sulfurtransferase from Aquifex aeolicus
FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS 4575
As the active site of sulfurtransferases involves a cys-
teine residue, Aq-477 was incubated with a cysteine-
modifying reagent, iodoacetamide to verify that Cys69
is required for ST activity. All the activity disappeared
at a 1 : 1 molar ratio (iodoacetamide ⁄ Aq-477). This
demonstrates that Cys69 is: (a) involved in the cataly-
sis, and (b) not involved in disulfide bond formation.
Several previously characterized rhodaneses are spe-
cifically inhibited by anions [33]. As seen for GlpE, the
only single-domain enzyme tested [5], slight inhibition
by anions was observed for Aq-477. Addition of
ammonium sulfate or ammonium acetate, at an ionic
strength of 300 mm, resulted in 25% inhibition of rho-
danese activity.
Various compounds were tested as sulfur donors.
Cysteine, dithiothreitol and b-mercaptopyruvate were
unable to replace thiosulfate as the sulfur donor. These
results show that Aq-477 is not a mercaptopyruvate
sulfurtransferase. Kinetics analysis was carried out
with thiosulfate, tetrathionate and polysulfide as sulfur
donors. In contrast to Sud from W. succinogenes or
GlpE from E. coli, Aq-477 was active with the three
substrates. All the kinetics show a Michaelis–Menten
behaviour and the parameters are summarized in
Table 2. Clearly, it appears that polysulfide sulfur was
a very efficient sulfur donor, indicating that this sulfur
compound is probably the real substrate. H
2

S produc-
tion was also tested. Because of the high unspecific
reaction with tetrathionate and polysulfide in the pre-
sence of dithiothreitol, the test was performed in pre-
sence of NaBH
4
instead of KCN as described by
Klimmeck et al. [34]. To detect H
2
S production, 100-
fold more enzymes were needed compared with the
kinetics of thiocyanate production, suggesting that this
reaction was not physiological.
It has been shown that reduced E. coli thioredoxin 1
serves as a sulfur-acceptor substrate for GlpE from
E. coli [5]. To test thioredoxin as a sulfur acceptor sub-
strate for Aq-477, we adapted the assay method
described for GlpE at 50 °C (see Experimental proce-
dures for the basis for this assay). Aq-477 may also be
able to utilize dithiol proteins such as thioredoxin as a
sulfur acceptor. However, owing to the amount and
stability of the proteins (thioredoxin and thioredoxin
reductase) needed in this test at 50 °C, determination
of the kinetic parameters was not possible.
The purified Aq-477 was active, with thiosulfate or
polysulfide as a sulfur donor and cyanide as sulfur
acceptor, over a large range of temperatures (25–
80 °C) (Table 3). Aq-477 showed a temperature opti-
mum at 80 °C and only 8% of the activity was
observed at 25 °C. Aq-477 showed a pH optimum at 9

and no sulfurtransferase activity was observed at pH
values < 7.
Oligomerization state of Aq-477
Four single-domains ST have been described to date.
For two of them, i.e. GlpE from E. coli and Sud from
W. succinogenes, biochemical characterization has
shown a homodimeric organization [5,34]. Gel-filtra-
tion chromatography was used to determine the
apparent molecular mass of recombinant Aq-477. One
peak is obtained, corresponding to a mix of the tetra-
meric, dimeric and monomeric forms as for the wild-
type protein (Fig. 3, trace A). When each fraction was
concentrated and run through the column, the same
elution profile was obtained. Use of Superose 12
instead of an S200 column did not enable us to obtain
a more homogeneous form. This indicates that Aq-477
forms soluble oligomers that are in equilibrium
Table 2. Steady-state kinetic parameters of Aq-477 from A. aeoli-
cus. Values were obtained from direct experimental measurements
fitted to the Michaelis–Menten equation. Apparent K
m
values for
polysulfide refer to polysulfide sulfur concentration. nd, no data.
Activity V
m
(s
)1
) Apparent K
m
(mM)

Thiosulfate rhodanese
sulfurtransferase
7865 ± 200 5.7 ± 0.9 (S
2
O
3
2–
);
2.1 ± 0.37 (CN
–
)
Thiosulfate
sulfurtransferase
no activity nd
Tetrathionate rhodanese
sulfurtransferase
8802 ± 500 6.9 ± 2 (S
4
O
6
)
Polysulfide sulfur
rhodanese
sulfurtransferase
165,000 < 0.05 (S
n
2–
)
Polysulfide sulfur
sulfurtransferase

36 nd
Thioredoxin
sulfurtransferase
72 nd
3-Mercaptopyruvate
rhodanese
sulfurtransferase
no activity nd
Table 3. Temperature dependence of thiosulfate rhodanese
sulfurtransferase and polysulfide sulfur rhodanese sulfurtransferase
activities of Aq-477 from A. aeolicus. nd, no data.
Temperature
(°C)
Thiosulfate
rhodanese
sulfurtransferase
activity (s
)1
)
Polysulfide
sulfur rhodanese
sulfurtransferase
activity (s
)1
)
85 8000 nd
60 5000 165000
37 2500 85000
25 1000 32000
Oligomeric sulfurtransferase from Aquifex aeolicus M C. Giuliani et al.

4576 FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS
with the monomer and the energy barriers for the
interconversion are not very high, as judged by the
ease of interconversion.
Electrophoretic migration of rec-Aq-477, under dena-
turing conditions, shows two bands, at 25 and 12 kDa
(Fig. 4A, lane 2). After transfer on a poly(vinylidene
difluoride) membrane, automated Edman degradation
yielded the same N-terminal sequence up to the 10th
cycles for the two bands. In a similar way, western blot-
ting after SDS ⁄ PAGE of each fraction from the S200
column (1.5 mgÆmL
)1
) showed two bands at 12 and
25 kDa for the recombinant enzyme (Fig. 4A, lane 3).
The same pattern was observed in the presence or
absence of dithiothreitol (data not shown) confirming
the absence of a disulfide bridge. These results show
that, even under denaturing conditions, the dimeric
form was detected. This suggests a tight interaction
between the two subunits and a possible physiological
role for the oligomeric form of Aq-477.
With the native enzyme from A. aeolicus, two bands
are detected by antibodies at 12 and 50 kDa, suggesting
a tetrameric organization of the enzyme (Fig. 4B,
lane 3). When the samples are boiled in the presence of
higher SDS concentrations (2%) the higher molecular
mass bands tend to disappear (data not shown). Only
one band around 50 kDa was detected by western blot-
ting performed after denaturing electrophoresis of solu-

ble crude extract (Fig. 4B, lane 2). The same experiment
performed with crude extract from A. aeolicus cultivated
on H
2
⁄ Na
2
SO
3
shows that Aq-477 was less abundant
under these growth conditions than with S° as the sulfur
electron acceptor (Fig. 4B, lane 4). These results demon-
strate that: (a) the oligomeric state of the enzyme con-
tains at least four subunits in vivo, and (b) the high
stability of the oligomeric state under these experimental
conditions as a high SDS concentration was necessary
to destabilize the subunit interactions.
The oligomeric state of Aq-477 influenced
by substrate, protein concentration and salt
The previous experiments suggest the existence of an
equilibrium reaction between monomer, dimer and at
least tetramer. First, we tested the effect of the sulfur
donor on the Aq-477 oligomeric state. Gel-filtration
chromatography was realized in presence of 10 mm
Na
2
S
2
O
3
or 1 mm polysulfide (Fig. 3, trace B). Com-

pared with the same experiment without substrate
(Fig. 3, trace A) the peak was more homogenous and
the molecular mass was 50 kDa. This result demon-
strates the tetrameric organization of Aq-477 in the
presence of substrate.
For a more defined correlation between the concen-
tration and the different states of oligomerization, a
stock solution of 4 mgÆmL
)1
wild-type or recombinant
Aq-477 was diluted to 0.4 and 0.04 mgÆmL
)1
at 25 °C
in 50 mm Tris ⁄ HCl, 100 mm NaCl, pH 7.6 and same
amount of protein was loaded onto 10–20% Blue
native (BN) gel. The gel patterns were different because
the diluted enzyme (1 ⁄ 100) presents a lower molecular
mass than the enzyme without dilution or if diluted at
1 ⁄ 10 (Fig. 5, lanes A and B). When the same experi-
ment was carried out in the absence of salt in the pro-
tein sample, the monomeric form was observed (Fig. 5,
lane C). The presence of 100 mm NaCl in the sample
induced oligomerization of the enzyme, suggesting
hydrophobic interactions between the subunits. In the
same way, the presence of thiosulfate at 10 mm in the
sample stabilizes the oligomeric form (Fig. 5, lane D).
Time-resolved fluorescence anisotropy experiments
confirmed most of the results obtained by size-exclusion
chromatography and BN gel. Fluorescence anisotropy
measurements are based on the depolarization of light

that occurs during the rotational diffusion of macro-
molecules or biological complexes. The extent of light
Elution volume (ml)
15
A
B
Fig. 3. Size-exclusion chromatography of recombinant Aq-477. An
S200 (1 · 30 cm) column was equilibrated in (A) 100 m
M Tris ⁄ HCl,
100 m
M NaCl, pH 7.6 or (B) 100 mM Tris ⁄ HCl, 100 mM NaCl,
10 m
M Na
2
S
2
O
3
pH 7.6 at 20 °C. Profile A: injection of 50 lLof
rec-Aq-477 at 2.6 mgÆmL
)1
; profile B: injection of 50 lL of rec-
Aq-477 at 2.6 mgÆmL
)1
in the presence of 10 mM Na
2
S
2
O
3.

The
protein was detected by its absorbance at 280 nm.
M C. Giuliani et al. Oligomeric sulfurtransferase from Aquifex aeolicus
FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS 4577
depolarization between excitation and emission times is
then related to the molecular size of the macromolecule.
Analysis of the time-resolved fluorescence anisotropy
data displays the distribution of rotational correlation
times (h), which are related to the hydrodynamics vol-
umes. Aq-477 contains one tryptophan residue and it
was studied using intrinsic tryptophan fluorescence.
Excitation at 295 nm resulted in an emission spectrum
with one maximum at 330 nm. In the presence of a satu-
rating concentration of thiosulfate, an increase of 30%
was observed without a significant shift of the maximum
(data not shown). This behaviour has previously been
seen with rhodanese from A. vinelandi [35]. Between 40
and 0.4 lm, Aq-477 displayed different rotational corre-
lation time (Table 4), confirming that its oligomeric
state is strongly dependent upon the protein concentra-
tion. At low concentrations and 25 °C, the rotational
correlation time was 7 ns, which corresponds to a globu-
lar protein of around 14 kDa. This demonstrates that at
low concentrations the major part of the enzyme was
monomeric. The equilibrium was shifted to the dimeric
form (14 ns) then to a trimeric or tetrameric form
(20 ns) when the protein concentration was increased
(Table 3). The same measurements were done with
Aq-477 sample freshly prepared in the presence of a sat-
urating concentration of thiosulfate. At low protein con-

centrations, instead of the monomer, the dimeric form
was detected with a longer correlation time of 11 ns. In
a similar way, at higher concentrations, the trimer and
tetrameric forms were detected in the presence of sub-
strate. Over the concentration range used, there is equi-
librium between monomer, dimer, trimer and tetramer.
Our results show that the initial step in the oligo-
merization process is the formation of dimers. This is
in agreement with the result obtained by electrophore-
sis under denaturing conditions. The dimeric form is
an active intermediate structural conformation that
evolves to at least a tetrameric form. Our results
2
3
A
1
2341
14,4
43
30
20,1
67
B
20
80
50
40
30
Fig. 4. SDS polyacrylamide gel of the purified Aq-477 from A. aeolicus. (A) SDS polyacrylamide gel of the purified recombinant Aq-477.
Lane 1, molecular mass markers (in kDa); lane 2, 1 lg of recombinant Aq-477; lane 3, immunoblotting experiments of recombinant Aq-477,

1 lg of recombinant Aq-477 was loaded onto the gel before detection by immunoblotting using anti-(Aq-477) sera. (B) SDS polyacrylamide
gel of Aq-477 from A. aeolicus. Lane 1, molecular mass markers (in kDa); lane 2, immunoblotting experiments of soluble crude extract from
A. aeolicus cultivated on H
2
⁄ S
o
medium. Protein (1 lg) was loaded onto the gel before detection by immunoblotting using anti-(Aq-477) sera.
Lane 3, immunoblotting experiments of soluble crude extract from A. aeolicus cultivated on H
2
⁄ NaS
2
O
3
medium. Protein (1 lg) was
loaded onto the gel before detection by immunoblotting using anti-(Aq-477) sera. Lane 4, immunoblotting experiments of enriched fraction
of Aq-477 from A. aeolicus. Protein (100 ng) was loaded onto the gel before detection by immunoblotting using anti-(Aq-477) sera.
AB CD
Fig. 5. BN gel of purified recombinant Aq-477. Lane A, 20 l gof
Aq-477 at 0.4 mgÆmL
)1
; lane B, 20 lg of Aq-477 at 0.04 mgÆmL
)1
;
lane C, 20 lg of Aq-477 at 0.4 mgÆmL
)1
without salt; lane D, 20 lg
of Aq-477 at 0.04 mgÆmL
)1
+S
2

O
3
10 mM.
Table 4. Comparison of lifetime and long rotational correlation
times of Aq-477 for different protein concentrations, in the absence
or presence of substrate at 25 °C. Correlations times were mea-
sured by monitoring tryptophan fluorescence (k
ex
¼ 295 nm;
k
e
m ¼ 330 nm) as indicated in the Experimental procedures. The
normalized values of the correlation times (reference temperature
20 °C) were obtained using the Perrin equation.
Aq-477 concentration (l
M)
0.4 4 40
Thiosulfate 10 m
M –+ –+ –+
Lifetime (ns) 1.66 1.41 1.8 1.57 1.11 1.16
Long rotational
correlation
time (ns)
71114212025
Oligomeric sulfurtransferase from Aquifex aeolicus M C. Giuliani et al.
4578 FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS
confirm that: (a) a dilution step generates the mono-
meric form, and (b) the substrate induces the oligomer-
ization state and a more homogenous state.
Stability and activity of Aq-477

We generated monomeric and oligomeric forms of
Aq-477 using a dilution step. We first verified the
global fold of the protein by CD experiments in the
far-UV. Independent of the degree of polymerization,
the enzyme was correctly folded. Deconvolution of the
spectra gives 28% a helix and 23% b sheet. This was in
the same range as obtained using various secondary
prediction software (around 30% a helix and 23%
b sheet). These experiments show the absence of dena-
turation of the enzyme directly after the dilution step
at room temperature. As the enzyme presents activity
at 80 °C and no activity at 25 °C, it was interesting to
study the effect of temperature on the structural stabil-
ity by CD ellipticity at 222 nm [36]. Changes in helical
content of the various forms of Aq-477 after thermal
treatment are shown in Fig. 6. Decrease in ellipticity at
222 nm was observed for the monomeric form of
Aq-477 indicating a loss of protein secondary structure
at high temperature (Fig. 6A, black circles). Compared
with the oligomeric form (Fig. 6A, black triangles), the
monomeric form is less stable with a transition around
60 °C. Denaturation of the monomer of Aq-477
appears to be an irreversible process as the initial sig-
nal is not recovered at 25 °C after thermal treatment.
The heat tolerance of the oligomeric Aq-477 indicates
a considerable contribution of the oligomerization
process to the thermal stability. The same experiments
were carried out with a monomeric enzyme sample
freshly prepared in the presence of 500 lm polysulfide
or 3 mm Na

2
S
2
O
3
. The spectrum obtained (Fig. 6B)
was similar to that of the oligomeric form, indicating a
role for the substrate in the stabilization or induction
of the more stable form. To better understand the sta-
bility of the various forms of Aq-477, its activity was
measured for the oligomeric and monomeric forms
preincubated at different temperatures from 4 to
80 °C. Independent of the incubation temperature
(4, 25 or 80 °C), the activity of the monomeric form of
Aq-477 decreased to 10% with a similar time-course of
inactivation over the whole temperature range
(Fig. 7A). Addition of substrate after 40 or 60 min of
incubation did not induce any modification in the
traces. This indicates the irreversibility of the inactiva-
tion process. The same experiments were carried out
with the oligomeric enzyme (Fig. 7B). Independent of
the temperature incubation used, the activity at 80 °C
was stable and 95% activity was still present after
150 min. The corresponding half-life of irreversible
inactivation increased from 19 min for the monomeric
form to 320 min for the oligomeric form. As shown
previously by fluorescence and CD experiments, deacti-
vation of the Aq-477 monomer is prevented by the
presence of a sulfur donor. As shown in Fig. 7C, addi-
tion of 5 or 10 mm S

2
O
3
or 500 lm polysulfide sulfur
in a freshly prepared monomeric form, results in
$ 80% stabilization of the activity at 80 °C.
These results demonstrate that: (a) the mono-
meric form is unstable, (b) monomer inactivation is
irreversible, (c) the substrate prevents inactivation, and
(d) temperature alone does not induce the active form.
Tem
p
erature (°C)
30 40 50 60 70 80 90
Relative CD Intensity at 222 nm (%)
20
40
60
80
100
30 40 50 60 70 80 90
20
40
60
80
100
AB
Fig. 6. Changes in relative CD intensity at 222 nm for monomeric (d) or oligomeric (m) Aq-477. (A) Relative CD intensity at 222 nm at vari-
ous temperatures for oligomeric or freshly prepared monomeric form of Aq-477. (B) Relative CD intensity at 222 nm for freshly prepared
monomeric form of Aq-477 prepared in the presence of 3 m

M Na
2
S
2
O
3
(d) or 500 lM polysulfide (.). The degree of polymerization was con-
trolled by BN gel.
M C. Giuliani et al. Oligomeric sulfurtransferase from Aquifex aeolicus
FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS 4579
The ability of proteins to adopt different quaternary
structures is essential for many biological processes
such as signal transduction, cell-cycle regulation and
enzyme catalysis. We tested the impact of the oligo-
merization state on the kinetic of ST to determine
whether oligomerization can promote regulation of the
kinetic behaviour.
Steady-state kinetics were measured with monomeric
enzyme and compared with values obtained with oligo-
meric enzyme. Lineweaver–Burk plots indicated that
whatever the enzyme forms there was no cooperativity
between the different active sites of the oligomeric
enzyme. However, the apparent V
m
was fourfold smal-
ler for the monomeric enzyme than for oligomeric
enzyme. The K
app
m
values were similar (5.48 ± 1 mm).

This is typically observed in the case of an irreversible
inactivation of the enzyme as the amount of active
enzyme in the test became smaller.
In conclusion, these results demonstrate: (a) the
absence of kinetic regulation, such as cooperativity, in
the oligomeric enzyme; and (b) that the oligomeric
form is the active form of the enzyme.
Discussion
Aq-477 is a single-domain rhodanese
We purified and characterized a protein from A. aeo-
licus annotated as a hypothetical protein. In vitro,it
catalyses the transfer of sulfane sulfur from thiosul-
fate to cyanide to form thiocyanate. According to
this activity and its amino acid sequence, Aq-477
belongs to the rhodanese (or sulfurtransferase) fam-
ily. It is the first single-domain sulfurtransferase to
be characterized from hyperthermophilic bacteria.
Moreover, it is the only one-domain STS to present
at least a tetrameric thermoactive and thermostable
organization which is controlled and induced by the
substrate.
In recent years, a considerable number of proteins
with a rhodanese homology fold have been detected.
The rhodanese fold was first observed in the crystal
structure of bovine mitochondrial rhodanese [19] and
later in the crystal structures of TTHA0613 from Ther-
mus thermophilus HB8 and At5g66040.1 from Arabid-
opsis thaliana [32,37]. This domain was found in the
Cdc25 class of protein phosphatases and in a variety
of proteins such as sulfite dehydrogenase, in certain

stress proteins and in cyanide and arsenate resistance
proteins [13].
Genome sequencing has shown that ORFs coding
for rhodanese or the mercaptopyruvate sulfurtransfer-
ase (MST) homologue are present in most eubacteria,
archaea and eukaryota [38,39]. Often, several genes
encoding for distinct ‘rhodanese-like’ proteins are
found in the same genome, suggesting that the encoded
proteins may have distinct biological functions. In
A. aeolicus, two genes encode two multidomain rhoda-
neses rhdA1 and rhdA2. Besides Aq-477, we have
detected only one other ORF that potentially encodes
a protein with a rhodanese fold Aq-1599. Characte-
rization of this protein is in progress.
Aq-477 is a thermostable and thermoactive
tetramer ST
Few single-domain rhodaneses have been characterized
in detail. Resolution of the 3D structure of Sud, sul-
furtransferase from W. succinogenes, shows a dimeric
organization [28] and this enzyme probably functions
as a dimer in solution. GlpE was also described as a
dimeric enzyme but the 3D structure did not confirm
this [7]. Dimerization in Sud occurs via the a1 helix
which is absent in Aq-477 and GlpE. The two last
120100806040200
Residual activity (%)
0
20
40
60

80
100
120
0
20
40
60
80
100
120
0
20
40
60
80
100
120
A
C
Time (min)
500 100 150 500 100 150200
B
Fig. 7. Thermal inactivation of freshly prepared monomer or oligomer of Aq-477. (A) Time course of irreversible inactivation of a freshly pre-
pared monomeric form of Aq-477 at 4 °C(d), 25 °C(.), and 80 °C(j). (B) Time course of irreversible inactivation of a freshly prepared
monomeric form of Aq-477 (.) or oligomeric form (d)at25°C. (C) Time course of irreversible inactivation of a freshly prepared monomeric
form of Aq-477 at 25 °C without (d) and in the presence of 500 l
M (r) polysulfide sulfur in the buffer.The degree of polymerization was
controlled by BN gel.
Oligomeric sulfurtransferase from Aquifex aeolicus M C. Giuliani et al.
4580 FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS

structures solved were those of T. thermophilus and
Ar. thaliana rhodaneses [29,33]. These two enzymes
were monomeric. However, no data are available on
their organization in solution, active form or physio-
logical role. Dimerization of RhdA from A. vinelandii
has also been shown, but only when mutations were
introduced into the catalytic loop inducing an interdis-
ulfide bridge [16] or when the enzyme was considerably
overexpressed [17]. However, in all cases, the active
enzyme was the monomeric form [16,17]. Our results
on Aq-477 from A. aeolicus show that this enzyme
exists at least as a monomer, dimer and tetramer at
25 °C. Western blotting on crude extracts revealed one
major band around 50 kDa, suggesting that the active
form in vivo is the oligomeric form. Aq-477 is the only
single-domain rhodanese characterized to date as a
thermoactive and thermostable tetramer. The crystal
structures of many proteins from hyperthermophiles
have been solved, and several factors responsible for
their extreme thermostability have been proposed,
including an increase in the number of ion pairs and
hydrogen bonds, core hydrophobicity and packing
density, as well as the oligomerization of several
subunits and an entropic effect due to the relatively
shorter surface loops and peptide chains [40]. Protein
stability arises from a combination of many factors,
which each contribute to various extents in different
proteins. It seems that there is no single dominating
factor, even in hyperthermophilic proteins [41].
Comparative examination of the primary structure of

ST did not point to any obvious features that could
explain the high thermostability of Aq-477 from A. aeo-
licus except for a decrease in the number of asparagine
residues (4 versus 12 in Sud), a diminution in glycine
residues and an increase in the number of hydropho-
bic residues (53 versus 48%). The same features were
observed in TTHA0613 from T. thermophilus HB8.
We have shown that: (a) the monomeric form was
less stable than the oligomers, and (b) the concentra-
tion and ⁄ or substrate induce the dimerization and the
tetramerization.
Few enzymes from hyperthermophilic organisms are
higher-order oligomers than their counterparts in me-
sophilic organisms and potential stabilizing role of
increased subunit interactions via oligomerization has
been suggested [41–44].
Moreover, the oligomeric organization of proteins,
and especially of enzymes, provides an additional level
of complexity and plays an important role in numerous
biological processes. In the simplest case of homodi-
mers, the intersubunit interface can provide an
additional shared binding site for noncompetitive
ligands, and ⁄ or mediate conformational changes [45].
The kinetic behaviour of Aq-477 does not present
any cooperativity processes. As a consequence, oligo-
merization of the enzyme was not in line with allosteric
regulation of the activity but more probably with ther-
mal stability. Thus, protein stability and not efficiency
has been selected for in the evolution of this oligomer
and assembly of identical subunits to noncovalently

associated oligomers is thought to ensure their survival
in hyperthermophiles. This is also the case for other
hyperthermophily enzymes such as phosphoribosyl-
anthranilate isomerase from Thermotoga maritima [46],
and formyltransferase from the hyperthermophile
Methanopyrus kaudleris [47].
One of the major driving forces for protein
oligomerization originates from shape complementarity
between the associating molecules, brought about by a
combination of hydrophobic and polar interactions
(e.g. hydrogen bonds and salts bridges) [48]. Our
results show the probable role of hydrophobic interac-
tions between subunits because dissociation occurs in
the absence of salt.
Functional role of Aq-477
Like all enzymes belonging to the rhodanese family, the
function of the single-domain enzymes in vivo is seri-
ously debated. When mercaptopyruvate was used as the
sulfur donor, no activity was detected with Aq-477, sug-
gesting that this protein was not a MST. This is in agree-
ment with the amino acid composition of the active site
loop which is different from the characteristic motif of
MST i.e. CG[S ⁄ T]GVT with no charged residues in the
loop [18]. In the same way, the Cd25 phosphatase
domain and arsenate resistance role were excluded as in
these enzymes an elongated seven amino acid active-site
loop was present. The Aq-477 amino acid loop presents
the motif of the catalytic domain of thiosulfate cyanide
sulfurtransferase (TST) which is distributed among bac-
teria, archaea and eukaryotes. Aq-477 catalyses sulfur

transfer from thiosulfate, tetrathionate and polysulfide.
To date, this is the only enzyme in which use of these
different sulfur donors has been demonstrated in vitro,
because Sud is inactive with thiosulfate [34] and the
polysulfide sulfurtranferase activity of GlpE has not
been demonstrated [5]. We propose that aq477 encodes
a monomeric rhodanese with polysulfide sulfurtran-
ferase activity and, therefore to rename this gene rhdB1.
Members of the genus Aquifex were obtained from
marine hydrothermal systems [27] where sulfur is the
predominant compound. Therefore, it is not surprising
to find numerous genes that encode putative proteins
involved in sulfur metabolism in the A. aeolicus gen-
ome. A supercomplex from A. aeolicus involved in
M C. Giuliani et al. Oligomeric sulfurtransferase from Aquifex aeolicus
FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS 4581
sulfur reduction in the cytoplasmic space has been
characterized by our group [30]. In Acidithiobacillus
ferrooxidans, proteomic experiments have shown the
induction of ST when sulfur was used as the electron
donor, and the role of this enzyme as a sulfur carrier
for various energetic complexes has been suggested
[49]. The well-studied Wolinella system proposes also a
direct interaction between Sud and the polysulfur
reductase. Experiments are currently being developed
by our group to determine if Aq-477 is the physiologi-
cal partner of the sulfur reductase.
Experimental procedures
All restriction enzymes were obtained from Promega (Mad-
ison WI). PCR was carried out using PWO DNA polyme-

rase from Roche (Mannheim, Germany). DNA ligase was
obtained from Roche. Cloning oligonucleotides were pur-
chased from MWG and DNA sequencing was performed
by Genome Express (Marseilles, France).
Bacterial strains and plasmids
The bacterial strains and plasmids used are listed in
Table 5. BL21 (DE3)-RIPL (Stratagene, La Jolla, CA) har-
bouring peT22aq477 was used to overproduce Aq-477 for
purification. To construct peT22aq477, the aq477 gene was
amplified by PCR with the primers 5¢-GG
CATATGTTTA
TGAACGTTCCGG-3¢ and 5¢-GG
GTCGACTTAAGATT
TAGCAGGT-3¢, where the underlined letters indicate the
restriction sites for, respectively, NdeI and SalI. After cleav-
age with NdeI and SalI, the amplified product was cloned
into peT22 to create peT22aq477.
Growth conditions for overproduction and
purification of recombinant Aq-477
E. coli BL21-CodonPlus (DE3)-RIPL was transformed with
the peT22aq477 plasmid. A culture (0.5 L) of recombinant
E. coli was grown at 37 °CtoD
600
¼ 0.6 and then induced
with 1 mm isopropyl thio-b-d-thiogalactoside for 3 h at
37 °C. Cells were harvested by centrifugation and disrupted
by French press in 50 mm Tris ⁄ HCl, pH 7.6 containing pro-
teases inhibitors cocktail from Roche. Crude extract was cen-
trifuged at 14 000 g for 10 min and immediately heated to
80 °C for 40 min to precipitate heat-labile E. coli proteins.

Following centrifugation at 4000 g for 15 min, the superna-
tant containing the overproduced protein was concentrated
using Centriprep concentrators (Amicon France, Epernon,
France) with YM-10 membranes and was loaded onto a
Superdex S200 high-resolution column (1 · 30 cm, FPLC
apparatus, Amersham Pharmacia Biotech, Piscataway, NJ)
equilibrated in buffer 100 mm Tris ⁄ HCl, 50 mm NaCl
pH 7.6 and eluted in the same buffer (0.3 mLÆmin
)1
). Active
fractions were concentrated, frozen in liquid nitrogen and
stored at )20 °C. All steps were performed at room tempera-
ture.
Growth conditions and purification of Aq-477
wild-type
A. aeolicus was cultivated in 1.4 Nm
)2
bottles under
1.4 bar H
2
⁄ CO
2
(80:20) in SME medium modified accord-
ing to Guiral et al. [30] at pH 6.8 in the presence of thiosul-
fate (1 gÆL
)1
)orS° (7.5 gÆL
)1
) and harvested in the late
exponential growth phase. Typical yields were $ 400 mg of

cell material per L of culture. After centrifugation (30 min
3700 g,4°C), the pellet was frozen and stored at )80 °C.
Periplasmic extraction was performed as described by Bru-
gna et al. [50]. Lactate dehydrogenase activity was mea-
sured to show whether the extract was contaminated by
cytoplasmic proteins. After periplasmic extraction, cell
material (50 g) was resuspended in 50 mm Tris ⁄ HCl, 5 mm
EDTA, 10 lgÆmL
)1
DNase, 5% glycerol and proteases
inhibitors (pH 7.6) under argon, and broken in French
press cell. Debris and unbroken cells were removed by
centrifugation (10 000 g, 15 min). Membrane and soluble
proteins were separated by ultracentrifugation (45 min,
15 000 g, rotor 45 Ti, Beckman). After dialysis the superna-
tant containing the soluble proteins was loaded onto a
Q-Sepharose (1.6 · 10 cm, FLPC) equilibrated in 50 mm
Tris ⁄ HCl (pH 7.6) buffer. The absorbed proteins were
eluted by a linear gradient of NaCl (0–1 m) in the same
buffer. Sulfurtransferase activity was found in the 300 mm
NaCl fraction. The fraction was loaded onto a HTP col-
umn (1 · 10 cm biogel; Bio-Rad, Hercules, CA) equili-
brated in 50 mm Tris ⁄ HCl, 300 mm NaCl (pH 7.6). The
column was washed with the same buffer and proteins were
eluted by a linear gradient of potassium phosphate (0–1 m),
pH 7.6. The 300 mm phosphate fraction contained the
major part of sulfurtransferase activity. After concentration
using Centriprep concentrators (Amicon) with YM-10
membranes, the last step of purification was done with a
Superdex 200 (1 · 30 cm) high-resolution column (FPLC

apparatus, Amersham Pharmacia Biotech) equilibrated in
100 mm Tris ⁄ HCl, 50 mm NaCl pH 7.6 and eluted in the
Table 5. Bacterial strains and DNA vectors.
Strain or vector
Genotype, comments
and ⁄ or reference
E. coli BL21-CodonPlus
(DE3)-RIPL
E. coli BF
–
ompT hsdS(r
B
–
m
B
–
)
dcm
+
Tet
r
gal k(DE3)
endA Hte [argU proL Cam
r
]
[argU ileY leuW Strep ⁄ Spec
r
]
Aquifex aeolicus VF5 (Deckert) 63744–3 Novagen
pET-22b(+) Novagen

pET22Aq477 aq477 gene from A. aeolicus in
NdeI-SalI fragment of pET-22b(+)
Oligomeric sulfurtransferase from Aquifex aeolicus M C. Giuliani et al.
4582 FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS
same buffer (0.3 mL min
)1
). Active fractions were concen-
trated, frozen in liquid nitrogen and stored at )20 °C. All
steps were performed at room temperature.
Sulfurtransferase activity assays
All buffers used for activity assays were preincubated under
argon, all assays were done at 80 °C. One unit of ST acti-
vity corresponds to the production of 1 lmol of thiocya-
nate or H
2
S per minute.
Thiosulfate rhodanese sulfurtransferase
ðS
2
O
2À
3
þ CN
À
: SCN
À
þ SO
2À
3
Þ

Thiosulfate rhodanese sulfurtransferase activity was deter-
mined by measuring SCN formation as the red Fe(SCN)
3
complex from cyanide and thiosulfate [34]. The reaction
mixture contained 100 mm Tris ⁄ HCl, pH 9.0, 10 mm KCN,
5 mm mercaptoethanol, and enzyme extract and was initi-
ated by addition of 10 mm Na
2
S
2
O
3
. This was realized in
presence and in absence of 2.5 mm dithiothreitol. After
incubation at 85 °C for 10 min the reaction was stopped by
addition of 200 lL acidic iron reagent (FeCl
3
,50gÆl
1
; 65%
HNO
3
, 200 mLÆL
1
). After centrifugation at 13 000 g for
3 min the absorption was read at 460 nm. Spontaneous
rates of thiocyanate formation were determined by omitting
the crude extract from the reaction mixture. Amounts of
product formation were quantified using a standard curve
done with NaSCN. Test without enzyme was done as con-

trol. S
2
O
3
stability at 80 °C and linearity of SCN
–
produc-
tion up to 20 min has been verified. Steady state kinetics
were done as described in [5]. The final concentration of
Aq-477 was 30 nm.
3-Mercaptopyruvate rhodanese sulfurtransferase
(HSCH
2
COCOO
)
+CN
)
:CH
3
COCOO
)
+SCN
)
)
The assay mixture consisted of 100 mm Tris ⁄ HCl, pH 9.0,
10 mm KCN, 2.5 mm dithiothreitol and enzyme extract and
was initiated by 5 mm 3-mercaptopyruvate. Assays were
incubated and treated as described above. The final concen-
tration of Aq-477 was 30 nm.
Thiosulfate sulfurtransferase

ðS
2
O
2À
3
þ BH
À
4
: HS
À
þ SO
2À
3
þ BH
3
Þ
Assay mixtures (1 mL) contained 100 mm Tris ⁄ HCl,
pH 9.0, 2.5 mm dithiothreitol and protein extracts as stated
above, and were started by adding 200 lm sodium thiosul-
fate. Reactions were incubated for 20 min at 37 °C. The
amount of H
2
S developed during the reaction was fixed by
adding 100 lL30mm FeCl
3
dissolved in 1.2 m HCl and
100 lL20mm N,N¢-dimethyl-p-phenylene-diamine dis-
solved in 7.2 m HCl. Samples were kept in the dark for
20 min, centrifuged and the absorption of methylene blue
formed was measured at 670 nm. For quantification,

standard curves were prepared or the molar extinction
coefficient of 15 · 10
6
Æcm
)1
Æm
)1
was used. The final concen-
tration of Aq-477 was 3 lm.
3-Mercaptopyruvate s ulfurtransferase ðHSCH
2
COCOO
À
þ BH
À
4
: HS
À
þ BH
3
þ CH
3
COCOO
À
Þ
Enzyme assays were performed as described for thiosulfate
sulfurtransferase with the exception of adding 3-mercapto-
pyruvate, rather than thiosulfate, as a substrate at a final
concentration of 1 mm.
Tetrathionate rhodanese sulfurtransferase

ðS
4
O
2À
6
þ CN
À
: SCN
À
þ S
3
O
2À
6
Þ
Enzyme assays were carried out as described for thiosulfate
sulfurtransferase with the exception of adding tetrathionate,
rather than thiosulfate, as a substrate at a final concentration
of 10 mm. The final concentration of Aq-477 was 30 nm.
Tetrathionate sulfurtransferase
ðS
4
O
2À
6
þ BH
þ
4
: HS
À

þ BH
3
þ S
3
O
2À
6
Þ
In this case NaBH
4
(5 mm) replaced KCN in the mixture
assay as described by Klimmek et al. [34]. H
2
S production
was measured as described for thiosulfate sulfurtransferase.
The final concentration of Aq-477 was 3 lm.
Polysulfide rhodanese sulfurtransferase
ðS
2À
n
þ CN
À
: SCN
À
þ S
2À
nÀ1
Þ
Polysulfide sulfur was generated as described by Klimmeck
et al. [34] and the polysulfide assay was carried out

following polysulfide consumption directly by measuring
A
360
(a
˚
¼ 0.38 mm
)1
.cm
)1
polysulfide sulfur) at 60 °Cas
described previously [34]. No dithiothreitol was added to
the medium. The final concentration of Aq-477 was 8 nm.
Polysulfide sulfurtransferase
ðS
2À
n
þ BH
4À
:HS
À
þ S
nÀ1
2À
þ BH
3
Þ
In this case, NaBH
4
(5 mm) replaced KCN in the mixture
assay as described by Klimmek et al. [34]. H

2
S production
was measured following polysulfide consumption directly
by measuring A
360
(a
˚
¼ 0.38 mm
)1
.cm
)1
)at60°C. The
final concentration of Aq-477 was 1 lm.
Thioredoxin sulfurtransferase
We used an NADPH-coupled assay with thioredoxin reduc-
tase to show whether reduced thioredoxin is an effective
M C. Giuliani et al. Oligomeric sulfurtransferase from Aquifex aeolicus
FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS 4583
sulfur acceptor substrate for Aq-477. NADPH, thioredox-
in f and thioredoxin reductase f stability has been verified
at various temperatures. The system was not stable at
85 °C, so the activity was measured at 50 °C. Each assay
(final volume, 0.5 mL) contained 50 mm Tris ⁄ HCl (pH 7.6),
0.25 UÆmL
)1
thioredoxin reductase, 100 lm NADPH, and
16 lm thioredoxin. Cuvettes containing all reagents except
NADPH were used as blanks. NADPH was added, and the
reaction mixtures were allowed to equilibrate. Measure-
ments of the absorbance at 340 nm were used to ensure

that the mixtures had reached equilibrium. After equilib-
rium had been reached, purified Aq-477 (160 nm), thiosulf-
onate (30 mm), or both were added.
For steady-state kinetics studies, thiosulfate, tetrathionate
and polysulfide were used as sulfur donor and cyanide as sul-
fur acceptor. The final concentration of Aq-477 was 35 nm.
Steady-state kinetics experimental measurements were fitted
to the Michaelis–Menten equation using sigma-plot.
The temperature dependence of the thiosulfate sulfur-
transferase activity as well as polysulfide rhodanese sulfur-
transferase was measured between 20 and 80 °C in 100 mm
Tris ⁄ HCl, pH 9.0. pH dependence of the thiosulfate sulfur-
transferase activity was carried out in 100 mm Mes (pH 6
and 6.5), 100 mm Mops (pH 7.1), 100 mm Hepes (pH 7.1,
7.6 and 8), 100 mm Tris ⁄ HCl (pH 8 and 9) or 100 m m gly-
cine (pH 9.2 and 10).
Inhibition and inactivation studies
Anion inhibition studies were carried out using ammonium
acetate (50–400 mm) and ammonium sulfate (50–400 mm)
in thiosulfate rhodanese sulfurtransferase assay.
Inactivation of cysteine was performed on purified
Aq-477 by incubation in 100 mm Tris ⁄ HCl pH 7.6 with
iodoacetamide reagent in 1 : 0.5, 1 : 1 and 1 : 2 molecular
ratios, for 1 h at room temperature. The remaining ST
activity was determined and compared with that of Aq-477
incubated without iodoacetamide.
Stability studies measurements were performed incuba-
ting monomeric or oligomeric form of Aq-477 at 0 °C,
25 °C and 80 °C. Activity was measured at various times.
The same experiments were carried out with enzyme pre-

pared with various substrate concentrations (1–20 mm).
The enzymatic assay was carried out with Na
2
S
2
O
3
10 mm
final or polysulfide sulfur at 500 lm.
Stability studies experimental measurements were fitted
to a decreasing exponential equation using sigma-plot.
CD studies
CD spectra were recorded on a Jasco J-715 spectropolarime-
ter equipped with a Peltier-type temperature control system
(model PTC-348WI) in the far-UV (195–250 nm). Experi-
mental conditions were 0.2 nmÆmin
)1
; temperature 20 or
80 °C. Spectra were averaged from five acquisitions. The
protein concentration was 0.05 mgÆmL
)1
in 20 mm Tris ⁄ HCl
buffer, pH 7.6. All CD spectra were baseline corrected by
buffer and analysed using CD spectroscopy deconvolution
cdnn 2.1 software for determining the secondary-structure
classes and K2d algorithm [51]. CD spectra were also moni-
tored at 222 nm at various temperatures (20–90 °Cat
1 °CÆs
)1
) for the monomeric and the oligomeric form. Sam-

ples were prepared in the absence of substrate in 20 mm
Tris ⁄ HCl buffer, pH 7.6 or in 20 mm Tris ⁄ HCl buffer,
pH 7.6 with 3 mm thiosulfate or 500 lm polysulfide.
N-Terminal sequence determination
N-Terminal amino acid sequences were determined from
soluble protein or after SDS ⁄ PAGE. After electrophoresis
on 12% polyacrylamide gel under denaturing conditions,
proteins were transferred onto a poly(vinylidene difluoride)
membrane for 45 min at a current intensity of 0.8 mAÆcm
)2
using a semidry electrophoretic transfer unit. Sequence
determinations were carried out with an Procise 494 micros-
equencer (Applied Biosystems, Foster City, CA). Quantita-
tive determination of phenylthiohydantoin derivates was
done by high-pressure liquid chromatography (Waters,
Manchester, UK) monitored by a data and chromatogra-
phy control station (Waters).
Molecular mass determination
MALDI-MS was performed on a reflectron time-of-flight
mass spectrometer equipped with delayed extraction (Voy-
ager DE-RP, Perspective Biosystem Inc., Paris, France).
The sample (0.7 lL) was mixed directly on the support with
an equal volume of matrix (saturated solution of sinapinic
acid in 40% acetonitrile and 60% water made 0.1% in tri-
fluoroacetic acid).
Analytical procedures
Native gel electrophoresis
Purified enzyme was loaded onto a native 4% polyacrylamide
stacking ⁄ 10% running gel (Mini-Protean II, Bio-Rad) or on
12.5% polyacrylamide Phast Gels with Phast Gel native buf-

fer strips (Phast System; Pharmacia, Uppsala, Sweden).
BN gels (10–20%) were used as described by Schagger &
von Jagow [52]. Oligomeric Aq-477 (20 lg) was loaded onto
the gel, and dilution shock was carried out or not on the pro-
tein (in the same buffer). To study the effect of salt or sub-
strate, Aq-477 was incubated for 10 min at 80 °C with
100 mm NaCl or 5 and 10 mm thiosulfate, before migration.
Denaturing gel electrophoresis
Purified enzyme (1 lg) was incubated for 3 min at 90 °C
with a sample loading buffer containing SDS 2% and
Oligomeric sulfurtransferase from Aquifex aeolicus M C. Giuliani et al.
4584 FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS
20 mm dithiothreitol, and loaded onto a 4% polyacryl-
amide stacking ⁄ 12% running SDS gel (Mini-Protean II,
Bio-Rad) or on 12.5% polyacrylamide Phast Gels with
SDS buffer strips (Phast System; Pharmacia). After migra-
tion, the gel was stained as described previously [53].
Immunoblotting
After migration, western blotting was performed using stan-
dard procedures. Anti-(Aq-477) sera against A. aeolicus
recombinant Aq-477 were used and the detection reaction
was performed using goat peroxidase-conjugated anti-(rab-
bit IgG) serum (Sigma, St. Louis, MO) and SuperSignal
West Pico chemiluminescent substrate reagents (Pierce,
Rockford, IL).
Time-resolved fluorescence experiments
Time-resolved fluorescence parameters (lifetimes and
correlation times) were obtained from the two polarized
fluorescence decays I
^

(t) and I
//
(t), using time-correlated
single-photon counting technique. The instrumentation
setup was essentially similar to those previously described
[54,55] with modifications. Briefly, a time-correlated single-
photon counting SPC-430 card (Becker-Hickl GmbH, Paris,
France) was used for the acquisition; the time scaling was
19.63 ps per channel and 4096 channels were used. The exci-
tation light pulse source was a Ti-sapphire femtosecond laser
(Millenia-pumped Tsunami, Spectra Physics), the repetition
rate of the laser was set down at 8 MHz, associated with a
third harmonic generator tuned to 300 nm. Fluorescence
emission was detected through a monochromator (ARC
SpectraPro-150) set at 345 nm (Dk ¼ 15 nm) by a micro-
channel plate photomultiplier (Hamamatsu R1564U-06)
connected to an amplifier Phillips Scientific 6954 at 25 °C.
The two polarized components of the fluorescence decay
were collected alternately over a period of 30 s until the total
count of the I
//
component reached $ 15 · 10
6
. The micro-
cell was thermostated with a Haake type-F3 circulating
nbath.
Analysis of the decay curves was performed using the
quantified maximum entropy method [56]. The anisotropy
decay is described by the following equation:
rðtÞ¼

I
II
ðtÞÀGI ?ðtÞ
I
II
ðtÞþ2GI ?ðtÞ
¼
X
n
i¼1
q
i
 e
Àt=h
i
ð2Þ
with
P
n
i¼1
q
i
¼ q
0
where G is the G-factor, the h
i
are the
individual correlation times and the q
i
are the associated

fractional amplitudes. Normalization of h for a given tem-
perature was performed using:
h ¼
g
ðTÞ
V
kT
ð3Þ
where g is the viscosity, V is the volume of the rotating
unit, k the Bolzmann constant, and T the temperature (K).
Sequence analysis
Multiple sequence alignments were performed using clus-
tal w [57]. Sequences were retrieved via the NCBI server
().
Secondary structure topology was predicted after multi-
ple sequence alignment at .
Acknowledgements
We gratefully acknowledge Marielle Bauzan (Fermen-
tation Plant Unit IBSM., Marseilles, France) for
growing the bacteria, Re
´
gine Lebrun (Proteomic
Analysis Center, Marseilles, France) for N-terminal
sequencing and mass determination, and Marianne
Guiral, Wolfgang Nitschke, Mari Lu` z Cardenas,
Athel Cornish-Bowden and Mireille Bruschi for help-
ful discussions.
References
1 Beinert H (2000) A tribute to sulfur. Eur J Biochem
267, 5657–5664.

2 Beinert H (2000) Iron–sulfur proteins: ancient struc-
tures, still full of surprises. J Biol Inorg Chem 5, 2–15.
3 Kessler D (2006) Enzymatic activation of sulfur for
incorporation into biomolecules in prokaryotes. FEMS
Microbiol 30, 825–840.
4 Cianci M, Gliubich F, Zanotti G & Berni R (2000) Spe-
cific interaction of lipoate at the active site of rhoda-
nese. Biochim Biophys Acta 1481, 103–108.
5 Ray WK, Zeng G, Potters MB, Mansuri AM & Larson
TJ (2000) Characterization of a 12-kilodalton rhodanese
encoded by glpE of Escherichia coli and its interaction
with thioredoxin. J Bacteriol 182, 2277–2284.
6 Cerletti P (1986) Seeking a better job for an under-
employed enzyme: rhodanese. Trends Biochem Sci 11 ,
369–372.
7 Leimku
¨
hler S & Rajagopalan KV (2001) A sulfurtrans-
ferase is required in the transfer of cysteine sulfur in the
in vitro synthesis of molybdopterin from precursor Z in
Escherichia coli. J Biol Chem 276, 22024–22031.
8 Lauhon CT, Skovran E, Urbina HD, Downs DM &
Vickery LE (2004) Substitutions in an active site loop of
Escherichia coli IscS result in specific defects in Fe–S
cluster and thionucleoside biosynthesis in vivo. J Biol
Chem 279, 19551–19558.
9 Lauhon C & Kambampati R (2000) The iscS gene in
Escherichia coli is required for the biosynthesis of 4-
thiouridine, thiamin, and NAD. J Biol Chem 275,
20096–20103.

10 Lauhon CT (2002) Requirement for IscS in biosynthesis
of all thionucleosides in Escherichia coli. J Bacteriol
184, 6820–6829.
M C. Giuliani et al. Oligomeric sulfurtransferase from Aquifex aeolicus
FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS 4585
11 Bui BT, Escalettes F, Chottard G, Florentin D & Mar-
quet A (2000) Enzyme-mediated sulfide production for
the reconstitution of [2Fe)2S] clusters into apo-biotin
synthase of Escherichia coli. Sulfide transfer from cyste-
ine to biotin. Eur J Biochem 267, 2688–2694.
12 Wolfe MD, Ahmed F, Lacourciere GM, Lauhon CT,
Stadtman TC & Larson TJ (2004) Functional diversity
of the rhodanese homology domain: the Escherichia
coli ybbB gene encodes a selenophosphate-dependent
tRNA 2-selenouridine synthase. J Biol Chem 279,
1801–1809.
13 Cereda A, Carpen A, Picariello G, Iriti M, Faoro F,
Ferranti P & Pagani S (2007) Effects of the deficiency
of the rhodanes-like RhdA in Azotobacter vinelandii.
FEBS Lett 581, 1625–1630.
14 Bordo D, Deriu D, Colnaghi R, Carpen A, Pagani S &
Bolognesi M (2000) The crystal structure of a sulfur-
transferase from Azotobacter vinelandii highlights the
evolutionary relationship between the rhodanes and
phosphatase enzyme families. J Mol Biol 298, 691–704.
15 Bordo D, Forlani F, Spallarossa A, Colnaghi R, Carpen
A, Bolognesi M & Pagani S (2000) A persulfate cysteine
promotes active site reactivity in Azotobacter vinelandii
rhodanese. Biol Chem 382 , 1245–1252.
16 Forlani F, Carpen A & Pagani S (2003) Evidence that

elongation of the catalytic loop of the Azotobacter vine-
landii rhodanese changed selectivity from sulphur- to
phosphate-containing substrates. Protein Eng 16,
515–519.
17 Cereda A, Forlani F, Lametti S, Bernhardt R, Ferranti
P, Picariello G, Pagani S & Bonomi F (2003) Molecular
recognition between Azotobacter vinelandii rhodanese
and sulphur acceptor protein. Biol Chem 384,
1473–1481.
18 Bordo D & Bork P (2002) The rhodanese ⁄ Cdc25 phos-
phatase superfamily. Sequence–structure–function rela-
tions. EMBO Reports 3, 729–746.
19 Ploegman JH, Drent G, Kalk KH, Hol WG, Heinrik-
son RL, Keim P, Weng L & Russell J (1978) Structure
of bovine liver rhodanese. I. Structure determination at
2.5 A
˚
resolution and a comparison of the conforma-
tion and sequence of its two domains. Nature 273,
124–129.
20 Ploegman JH, Drent G, Kalk KH, Hol WG, Heinrikson
RL, Keim P, Weng L & Russell J (1979) The structure
of bovine liver rhodanese. II. The active site in the sul-
fur-substituted and the sulfur-free enzyme. J Mol Biol
127 (2), 149–162.
21 Spallarossa A, Donahue JL, Larson TJ, Bolognesi M &
Bordo D (2001) Escherichia coli GlpE is a prototype
sulfurtransferase for the single-domain rhodanese
homology superfamily. Structure 9, 1117–1122.
22 Adams H, Teertstra W, Koster M & Tommassen J

(2002) PspE (phage-shock protein E) of Escherichia coli
is a rhodanese. FEBS Lett 518
, 173–176.
23 Kleerebezem M, Crielaard W & Tommassen J (1996)
Involvement of stress protein PspA (phage shock pro-
tein A) of Escherichia coli in maintenance of the proton-
motive force under stress conditions. EMBO J 15,
162–171.
24 Kelly RM & Adams MWW (1994) Metabolism in
hyperthermophilic microorganisms. Antonie V Leeuwen-
hoek 66, 247–270.
25 Adams MW (1994) Biochemical diversity among sulfur-
dependent, hyperthermophilic microorganisms. FEMS
Microbiol Rev 15, 261–277.
26 Stetter KO (1999) Extremophiles and their adaptation
to hot environments. FEBS Lett 452 (1–2), 22–25.
27 Olsen GJ, Woese CR & Overbeek R (1994) The winds
of (evolutionary) change: breathing new life into micro-
biology. J Bacteriol 176, 1–6.
28 Deckert G, Warren PV, Gaasterland T, Young WG,
Lenox AL, Graham DE, Overbeek R, Snead MA,
Keller M, Aujay M et al. (1998) The complete genome
of the hyperthermophilic bacterium Aquifex aeolicus.
Nature 392, 353–358.
29 Brugna-Guiral M, Tron P, Nitschke W, Stetter KO,
Burlat B, Guigliarelli B, Bruschi M & Giudici-Orticoni
MT (2003) [NiFe] hydrogenases from the hyperthermo-
philic bacterium Aquifex aeolicus: properties, function,
and phylogenetics. Extremophiles 7, 145–157.
30 Guiral M, Tron P, Aubert C, Gloter A, Iobbi-Nivol C

& Giudici-Orticoni MT (2005) A membrane-bound mul-
tienzyme, hydrogen-oxidizing, and sulfur-reducing com-
plex from the hyperthermophilic bacterium Aquifex
aeolicus. J Biol Chem 280, 42004–42015.
31 Lin YJ, Dancea F, Lohr F, Klimmek O, Pfeiffer-
Marek S, Nilges M, Wienk H, Kroger A & Ruterjans
H (2004) Solution structure of the 30 kDa
polysulfide-sulfurtransferase homodimer from Wolinella
succinogenes. Biochemistry 43, 1418–1424.
32 Hattori M, Mizohata E, Tatsuguchi A, Shibata R,
Kishishita S, Murayama K, Terada T, Kuramitsu S,
Shirouzu M & Yokoyama S (2006) Crystal structure
of the single-domain rhodanese homologue TTHA0613
from Thermus thermophilus HB8. Proteins 64,
284–287.
33 Alexander K & Volini M (1987) Properties of an
Escherichia coli rhodanese. J Biol Chem 262, 6595–6604.
34 Klimmek O, Kreis V, Klein C, Simon J, Wittershagen
A & Kroger A (1998) The single cysteine residue of the
Sud protein is required for its function as a polysulfide-
sulfurtransferase in Wolinella succinogenes. Eur J
Biochem 253, 263–269.
35 Finazzi Agro A, Federici G, Giovagnoli C, Cannella C
& Cavallini D (1972) Effect of sulfur binding on rhoda-
nese fluorescence. Eur J Biochem 28, 89–93.
36 Greenfield NJ (1996) Methods to estimate the confor-
mation of proteins and polypeptides from circular
dichroism data. Anal Biochem
235, 1–10.
Oligomeric sulfurtransferase from Aquifex aeolicus M C. Giuliani et al.

4586 FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS
37 Cornilescu G, Vinarov DA, Tyler EM, Markley JL &
Cornilescu CC (2006) Solution structure of a single-
domain thiosulfate sulfurtransferase from Arabidopsis
thaliana. Protein Sci 15, 2836–2841.
38 Schultz J, Milpetz F, Bork P & Ponting CP (1998)
MART, a simple modular architecture research tool:
identification of signaling domains. Proc Natl Acad Sci
USA 95, 5857–5864.
39 Mueller EG (2006) Trafficking in persulfides: delivering
sulfur in biosynthetic pathways. Nat Chem Biol 1,
185–194.
40 Petsko GA (2001) Structural basis of thermostability in
hyperthermophilic proteins, or ‘there’s more than one
way to skin a cat’. Methods Enzymol 334, 469–478.
41 Opitz U, Rudolph R, Jaenicke R, Ericsson L & Neurath
H (1987) Proteolytic dimers of porcine muscle lactate
dehydrogenase. Characterization, folding, and reconsti-
tution of the truncated and nicked polypeptide chain.
Biochemistry 26, 1399–1406.
42 Walden H, Taylor GL, Lorentzen E, Pohl E, Lilie H,
Schramm A, Knura T, Stubbe K, Tjaden B & Hensel R
(2004) Structure and function of a regulated archaeal
triosephosphate isomerase adapted to high temperature.
J Mol Biol 342, 861–875.
43 Vieille C & Zeikus GJ (2001) Hyperthermophilic
enzymes: sources, uses, and molecular mechanisms for
thermostability. Microbiol Mol Biol Rev 65, 1–43.
44 Tanaka Y, Tsumoto K, Yasutake Y, Umetsu M, Yao
M, Fukada H, Tanaka I & Kumagai I (2004) How

oligomerization contributes to the thermostability of an
archaeon protein. Protein 1-isoaspartyl-O-methyltrans-
ferase from Sulfolobus tokodaii. J Biol Chem 279,
32957–33267.
45 Traut TW (1994) Dissociation of enzyme oligomers: a
mechanism for allosteric regulation. CRC Crit Rev Bio-
chem 29, 125–163.
46 Thoma R, Hennig M, Sterner R & Krischner K
(2000) Structure and function of mutationally gener-
ated monomers of dimeric phosphoribosylanthranilate
isomerase from Thermotoga maritima. Structure 8,
265–276.
47 Shima S, Thauer RK, Ermler U, Durchschlag H, Tzi-
atzios C & Schubert D (2000) A mutation affecting the
association equilibrium of formyltransferase from the
hyperthermophilic Methanopyrus kandleri and its influ-
ence on the enzyme’s activity and thermostability. Eur
J Biochem 267, 6619–6623.
48 Jones S & Thornton JM (1996) Principles of protein–
protein interactions. Proc Natl Acad Sci USA 93, 13–20.
49 Acosta M, Beard S, Ponce J, Vera M, Mobarec JC &
Jerez CA (2005) Identification of putative sulfurtransfer-
ase genes in the extremophilic Acidithiobacillus ferrooxi-
dans ATCC 23270 genome: structural and functional
characterization of the proteins. OMICS 9, 13–29.
50 Brugna M, Giudici-Orticoni MT, Spinelli S, Brown K,
Tegoni M & Bruschi M (1998) Kinetics and interaction
studies between cytochrome c
3
and Fe-only hydrogenase

from Desulfovibrio vulgaris Hildenborough. Proteins 33,
590–600.
51 Andarde MA, Chacon P, Merelo JJ & Moran F (1993)
Evaluation of secondary structure of proteins from UV
circular dichroism spectra using an unsupervised learn-
ing neural network. Protein Eng 6, 383–390.
52 Schagger H & von Jagow G (1991) Blue native electro-
phoresis for isolation of membrane protein complexes in
enzymatically active form. Anal Biochem 199, 223–231.
53 Mortz E, Krogh TN, Vorum H & Gorg A (2001)
Improved silver staining protocols for high sensitivity
protein identification using matrix-assisted laser desorp-
tion ⁄ ionization-time of flight analysis. Proteomics 1,
1359–1363.
54 Deprez E, Tauc P, Leh H, Mouscadet JF, Auclair C,
Hawkins ME & Brochon JC (2001) DNA binding
induces dissociation of the multimeric form of HIV-1
integrase: a time-resolved fluorescence anisotropy study.
Proc Natl Acad Sci USA 98, 10090–10095.
55 Deprez E, Tauc P, Leh H, Mouscadet JF, Auclair C &
Brochon JC (2000) Oligomeric states of the HIV-1
integrase as measured by time-resolved fluorescence
anisotropy. Biochemistry 39, 9275–9284.
56 Brochon JC (1994) Maximum entropy method of data
analysis in time-resolved spectroscopy. Methods Enzy-
mol 240, 262–311.
57 Thompson JD, Higgins DG & Gibson TJ (1994) CLUS-
TAL W: improving the sensitivity of progressive multi-
ple sequence alignment through sequence weighting,
position-specific gap penalties and weight matrix choice.

Nucleic Acids Res 22, 4673–4680.
M C. Giuliani et al. Oligomeric sulfurtransferase from Aquifex aeolicus
FEBS Journal 274 (2007) 4572–4587 ª 2007 The Authors Journal compilation ª 2007 FEBS 4587