Enlarging the gas access channel to the active site renders
the regulatory hydrogenase HupUV of Rhodobacter
capsulatus O
2
sensitive without affecting its transductory
activity
Ophe
´
lie Duche
´
1
, Sylvie Elsen
1
, Laurent Cournac
2
and Annette Colbeau
1
1 Laboratoire de Biochimie et Biophysique des Syste
`
mes Inte
´
gre
´
s (UMR 5092 CNRS-CEA-UJF), De
´
partement Re
´
ponse et Dynamique
Cellulaires, Grenoble, France
2 CEA Cadarache, De
´

are quickly and irreversibly inactivated in the presence
of O
2
[3]. In contrast, most [NiFe] hydrogenases are
only reversibly inhibited by O
2
.
The structure of the bimetallic active site and the
mechanisms of hydrogen oxidation in [NiFe] hydro-
genases have been thoroughly studied by various bio-
physical methods (reviewed in [4,5]). The information
obtained has given clues to the inactivation of the
enzyme by O
2
.InDesulfovibrio hydrogenases, it has
been shown that the Fe atom is linked to three non-
protein ligands: 1 CO and 2 CN
–
[6]. The Ni and Fe
ions are asymmetrically bridged by two cysteine sulfur
atoms and one oxygenic species (O
2
–
or OH
–
), which
appears in the oxidized enzyme [7–9]. The catalytic
Keywords
gas access channel; hydrogenases; oxygen
sensitivity; Rhodobacter capsulatus

2
inaccessibility to the active site, we introduced two mutations in order to
enlarge the gas access channel in the HupUV protein. We showed that such
mutations (Ile65 fi Val and Phe113 fi Leu in HupV) rendered HupUV
sensitive to O
2
inactivation. Also, in contrast with the wild-type protein,
the mutated protein exhibited an increase in hydrogenase activity after
reductive activation in the presence of reduced methyl viologen (up to 30%
of the activity of the wild-type). The H
2
-sensing HupUV protein is the first
component of the H
2
-transduction cascade, which, together with the two-
component system HupT ⁄ HupR, regulates HupSL synthesis in response to
H
2
availability. In vitro, the purified mutant HupUV protein was able to
interact with the histidine kinase HupT. In vivo, the mutant protein exhib-
ited the same hydrogenase activity as the wild-type enzyme and was equally
able to repress HupSL synthesis in the absence of H
2
.
Abbreviations
MG medium, malate ⁄ glutamate medium; MN medium, malate ⁄ ammonia medium; RH, regulatory hydrogenase; SH, soluble NAD-linked
hydrogenase.
FEBS Journal 272 (2005) 3899–3908 ª 2005 FEBS 3899
activity of [NiFe] hydrogenases, i.e. binding and
oxidation of H

sensor protein, this system comprises a histidine
kinase and a response regulator (HupT and HupR
respectively in R. capsulatus), which form a two-
component regulatory system functioning by phosphate
transfer [20]. We have demonstrated that the H
2
sen-
sor, HupUV, interacts directly with the histidine kinase
HupT [13], thus promoting its autophosphorylation in
the absence of H
2
. The phosphate is then transfered to
the response regulator HupR, which, in contrast with
most response regulators, is active in the unphosphory-
lated state [20]. Consequently, this phosphorylation
leads to the inactivation of the transcriptional factor
HupR and to the decrease in the synthesis of HupSL
hydrogenase in the absence of H
2
. A homologous
system has been found in R. eutropha, namely the
HoxBC ⁄ HoxJ ⁄ HoxA system [21].
Compared with standard hydrogenases, RHs from
R. capsulatus and R. eutropha exhibit unusual bio-
chemical features. The most interesting feature is that
they are O
2
insensitive [14,16,18,22], and thus could
offer an attractive option for applications in a future
hydrogen economy. However, the hydrogenase activity

decreased, suggesting that a partial blocking of the gas
channel by the presence of bulky residues may indeed
explain the O
2
insensitivity of the sensor enzymes [25].
In this study, we replaced Ile65 and Phe113 (corres-
ponding to amino acids 74 and 122 in the large sub-
unit of D. fructosovorans hydrogenase) of the large
subunit (HupV) of HupUV with Val and Leu, respect-
ively, and showed that these amino acids are indeed
involved in the O
2
insensitivity of the isolated protein.
We have also shown that the mutated HupUV protein
is as active in vivo as wild-type HupUV and is func-
tional in the H
2
-transduction system.
Results
Overproduction of mutated HupUV proteins
in R. capsulatus
We used site-directed mutagenesis to modify two bulky
residues lining the putative gas access channel in the
large subunit HupV (Ile65 and Phe113 replaced by
Val and Leu, respectively) After mutagenesis, the
hupUV genes were cloned into the expression vector
pSE102. In pSE103 and pOD7, the wild-type and
mutated hupUV genes, respectively, are expressed from
the strong nif promoter.
To assess H

HupU subunit was produced as a fusion protein with
an N-terminal His
6
tag, we were able to purify the
complex His
6
HupUHupV by affinity chromatography
on a Ni
2+
-charged column. Figure 1A shows the last
O
2
sensitivity of the regulatory hydrogenase HupUV O. Duche
´
et al.
3900 FEBS Journal 272 (2005) 3899–3908 ª 2005 FEBS
step of the purification of wild-type and mutated pro-
teins, with the two subunits in a stoichiometric ratio.
Under native conditions (Fig. 1B), the two proteins
displayed the same pattern. The molecular masses of
the two bands ( 80 and 170 kDa) were estimated by
runs in native gels with different acrylamide concentra-
tions [26], as previously described [13]. They corres-
pond, respectively, to the dimeric form, HupUV, and
the tetrameric form, Hup(UV)
2
, both of which exhibit
hydrogenase activity (see below). These results suggest
that the quaternary structure of the mutated protein
was well conserved.

tially protected when stored under N
2
; it exhibited
 50% activity during the same time (as compared
with 20% under air) (not shown). Thus in the mutant
protein, there was specific inactivation of the catalytic
activity by O
2
, but the mutation could also modify the
conformation of the protein, rendering it unstable.
H–D exchange activity catalysed by wild-type and
mutated HupUV proteins
The effect of O
2
on the activity of aerobically purified
HupUV proteins was then assessed directly by a MS
method monitoring continuously the H–D exchange in
either the absence or presence of O
2
. The results are
given in Table 1.
In the wild-type HupUV protein, the activity and
the rate of HD and H
2
formation were similar under
aerobic and anaerobic conditions. These results are in
agreement with a previous study reporting that the H–D
exchange reaction catalyzed by the HupUV protein was
high in the presence of O
2

did not further activate the HupUV protein,
whereas, interestingly, the activity of the OD7 protein
AB
Fig. 1. SDS ⁄ polyacrylamide gel (A) and native gel (B) of wild-type
and mutated HupUV proteins. (A) Cell extracts from 5 L were puri-
fied on two successive Ni
2+
-charged columns. Then 10 lL of the
pools purified on the second HiTrap column and eluted with
250 m
M imidazole were loaded on to an SDS ⁄ 12% polyacrylamide
gel. Lane 1, wild-type; lane 2, mutant. (B) An 8-lg sample of each
protein was run on a native polyacrylamide gel and stained with
Coomassie Brilliant Blue. Lane 1, wild-type; lane 2, mutant.
Fig. 2. Inactivation of wild-type and mutated HupUV proteins in air.
Soluble extracts obtained after centrifugation of sonicated cells at
50 000 r.p.m. for 1 h (A) and purified proteins (B) were kept at 4 °C
under air, and H
2
-uptake hydrogenase activity was assayed every
day during 1 week. Wild-type, diamonds; mutant, circles. Data rep-
resent the mean results from two or three independent assays.
O. Duche
´
et al. O
2
sensitivity of the regulatory hydrogenase HupUV
FEBS Journal 272 (2005) 3899–3908 ª 2005 FEBS 3901
was fourfold higher after reduction by MV
+

complex, as previously determined for wild-type
HupUV [13], and the amount of free HupUV decreas-
ed. There was no difference in migration between the
two HupUV–HupT complexes, suggesting that the
mutations did not substantially modify the interaction.
Table 1. H–D exchange activity and rate of H
2
and HD formation by wild-type and mutated HupUV proteins of R. capsulatus. The values are
initial rates corrected for gas consumption by the mass spectrometer. Activity and H
2
or HD rate of formation are expressed as lmol formedÆ
min
)1
Æ(mg protein)
)1
as described [44]. Assays under aerobiosis and anaerobiosis were performed separately. When noted, reduced methyl
viologen (MV
+
) was present at 0.16 mm. Data are means from two or three independent experiments, with variation of less than 15%.
Proteins
Aerobiosis Anaerobiosis Anaerobiosis + MV
+
Activity
HD
formation
H
2
formation Activity
HD
formation

synthesis of HupSL hydrogenase
The next question we addressed was to check whether
the O
2
-sensitive mutated protein was able to function
in vivo, i.e. to transduce the H
2
signal, and in the
absence of H
2
, to repress hydrogenase synthesis, even
in presence of O
2
. The plasmids pSE103 and pOD7,
which expressed hupUV genes from the nif promoter,
were not suitable for in vivo experiments, because this
promoter is not active under aerobiosis [28]. For this
reason, the plasmids pSE60 and pOD15, in which the
hupUV genes were cloned under control of the fruc-
tose-induced fru promoter, were constructed and used
to complement the hupUV mutant strain BSE16. The
complemented cells were grown under aerobiosis or
anaerobiosis in the presence of 3 mm fructose and in
either the presence (derepressing conditions) or absence
(repressing conditions) of H
2
.H
2
was produced endo-
genously as a by-product of nitrogenase activity during

2
resistance by blocking O
2
access to the active site
[25]. To check this hypothesis, we replaced, by site-
directed mutagenesis, two amino acids that line the gas
access channel, Ile65 and Phe113, with Val and Leu,
respectively. Interestingly, these replacements rendered
the protein O
2
sensitive, demonstrating that these resi-
dues are involved in O
2
sensitivity of the RH. This was
corroborated by experiments showing that the H–D
exchange activity of the mutant protein increased
greatly in the presence of reduced MV, at variance
with that of the wild-type protein. However, even after
reductive activation, the hydrogenase activity of puri-
fied mutated HupUV protein remained twice as low as
that of the wild-type, suggesting that O
2
may also irre-
versibly inactivate the active site. Another explanation
is that the mutations could also modify the structure
around the active site and⁄ or the binding of ligands,
thus decreasing the catalytic efficiency of the enzyme.
Our results suggest that, in vivo, the mutated HupUV
protein is protected from O
2

ive NiFe site, thus inactivating the reaction with H
2
[30].
It should be noted that the occurrence of direct binding
of O
2
to the active NiFe site is under debate and was not
observed for hydrogenase from Desulfovibrio gigas [8].
Some hydrogenases, however, are able to consume
H
2
in the presence of O
2
, and exhibit noticeable resist-
ance to this gas. The best-known enzyme is the soluble
Table 2. Hydrogenase activities of the wild-type B10 and hupUV
BSE16 strains from R. capsulatus, complemented with wild-type
and mutated hupUV genes. Cells were grown overnight at 30 °C
anaerobically in the light (MN or MG medium) or aerobically in the
dark (MN or MN medium + 10% H
2
)toanA
660
of  1.5. In MG
medium, H
2
was evolved from nitrogenase activity. Fructose
(3 m
M) was added at the beginning of growth at an A
660

groups, tenta-
tively assigned to the Fe atom and the Ni atom [6,32].
It has been hypothesized that these two CN
–
groups
may shield the active site from O
2
attack by steric hin-
drance [32,33]. The Ni-bound CN
–
seems to be respon-
sible for the O
2
insensitivity of the enzyme, and is
linked to the presence of the hypX gene [34,35], found
also in other aerobic bacteria such as Rhizobium [36].
Indeed, in SH purified from an HypX
–
strain, the cata-
lytic turnover (the hydrogenase activity) was shown to
be independent of the presence of O
2
, but the enzyme
was irreversibly inactivated if O
2
was present during
the autocatalytic activation [35], probably because of
formation of some peroxide or superoxide. In a recent
study, a mutant of HoxH, the active-site-containing
subunit of the SH, was constructed by replacement of

atom liganded by 1 CO and 2 CN
–
[16], and the binding
of an hydride to Ni and Fe after H
2
reduction has
recently been demonstrated [38]. However, in contrast
with standard hydrogenases, the RH exists only as two
redox forms, i.e. ready oxidized and reduced. The O
2
and MV
+
responses observed in the mutant HupUV
protein suggest that it has reached unready states, and
further studies will be needed to determine which ones.
In a recent study using X-ray absorption spectroscopy,
Haumann et al. [39] suggested that the specific Ni
co-ordination may also be crucial to the O
2
insensitivity
of the R. eutropha RH. In particular, the number of S
ligands was decreased by one upon formation of the
active state, but binding of O
2
to the active site was pre-
vented because an O ⁄ N ligand from an amino acid was
already bound at the free position at the Ni site.
In any case, it appears that in the O
2
-resistant

O
2
-sensitive HupUV proteins by biophysical methods
may lead to the improved understanding of the mecha-
nisms of O
2
resistance ⁄ sensitivity in [NiFe] hydrogen-
ases in general. RHs that are insensitive to O
2
and, as
isolated, ready to function are potentially of great bio-
technological interest, but their activity is low. When
the basis of their O
2
resistance is understood, it will be
possible to design a hydrogenase that exhibits high
activity together with O
2
insensitivity.
Experimental procedures
Bacterial strains and plasmids
The strains and plasmids used in this study are listed in
Table 3. R. capsulatus strains were grown heterotrophically
at 30 °C under anaerobiosis in the light or under aerobiosis
in the dark with shaking, in MG medium (7 mm glutamate,
30 mmdl-malate) or MN medium (7 mm ammonium sul-
fate, 30 mmdl-malate) [19]. Escherichia coli strains were
grown at 37 °C in Luria–Bertani medium. Antibiotics were
used at the following concentrations: 100 (ampicillin) and
10 (tetracycline) mgÆL

Phe113 from HupV with Leu. The 3.2-kb fragment corres-
ponding to the mutated hupUV genes was excised from plas-
mid pOD585 with NdeI–BamHI and cloned into pSE50
digested with the same enzymes in place of the wild-type
hupUV genes, leading to plasmid pSE504. Plasmid pSE102
was cleaved with NcoI–BamHI, to clone 3.2-kb fragments
from pSE50 and pSE504 digested with the same enzymes,
leading to plasmids pSE103 and pOD7, respectively, which
were introduced into R. capsulatus hupUV mutant BSE16 or
hupSL mutant JP91 by conjugation. From these plasmids,
the HupU subunit will carry an N-terminal His
6
tag for easy
purification of the HupUV complex.
Purification of the His
6
-HupUV proteins
In the plasmids pSE103 and pOD7, wild-type and
mutated hupUV genes were expressed from the nifHDK
promoter. For this reason, cells (from 5 L culture) were
grown under conditions allowing strong expression of the
nif promoter (MG medium, anaerobiosis, under light).
Proteins were purified on a HiTrap chelating column
(Amersham Pharmacia Biotech, Piscataway, NJ, USA) as
described previously [13]. Elution of the 5-mL column
with buffer containing 100 mm imidazole gave an active
pool, which was concentrated on a 1-mL column by elu-
tion with 250 mm imidazole in the buffer. The pools were
dialyzed three times in 25 mm Tris ⁄ HCl (pH 8) contain-
ing 10% (v ⁄ v) glycerol and 150 mm NaCl, at 4 °C. The

introduced into the vessel, and the changes in D
2
,HD
and H
2
were monitored by scanning masses 4, 3 and 2,
Table 3. Bacterial strains and plasmids used in this study.
Strain or plasmid Relevant characteristics Source or reference
Strains
R. capsulatus
B10 Wild-type [45]
BSE16 hupUV Hup
c
[12]
JP91 hupSL Hup
–
[46]
Plasmids
pUC18 Ap
r
[47]
pFRK-I Ap
r
Ble
r
Gm
r
Km
r
; fruP fusion vector [48]

r
; pSE102 with 3.2-kb NcoI-BamHI from pSE504 This work
pOD12 Ap
r
Gm
r
; pFRK-I with a 3.2-kb NdeI-BamHI from pOD585 This work
pSE60 Tc
r
; pPHU234 with 6.2-kb HindIII containing hupUV [12]
pOD15 Tc
r
; pPHU231 with 6-kb HindIII-BamHI from pOD12 This work
O. Duche
´
et al. O
2
sensitivity of the regulatory hydrogenase HupUV
FEBS Journal 272 (2005) 3899–3908 ª 2005 FEBS 3905
respectively. When required, the medium was made aero-
bic by the addition of H
2
O
2
(5 lL 0.3% H
2
O
2
) decom-
posed by the addition of catalase (500 U) thus liberating

Km
r
cartridge)
of the plasmid pFRK-I, leading to plasmid pOD12.
pFRK-I contains a fructose-activated promoter, pfru, from
R. capsulatus. From pOD12, the HindIII–BamHI fragment
containing mutated hupUV genes downstream of pfru was
cloned into the broad host range plasmid pPHU231 diges-
ted with the same enzymes. The resulting plasmid, pOD15,
was introduced into the R. capsulatus hupUV mutant
BSE16 by triparental conjugation [13].
Acknowledgements
We thank P. M. Vignais and M. Satre for critical read-
ing of the manuscript. We also thank P. Carrier for
excellent technical assistance. O.D. was supported by a
two-year postdoctoral grant from the Commissariat a
`
l’Energie Atomique (CEA). The work was supported
by research grants from the CEA, the Centre National
de la Recherche Scientifique (CNRS: ACI ‘Energie Con-
ception Durable’) and the Universite
´
Joseph Fourier
(UJF) de Grenoble.
References
1 Vignais PM, Billoud B & Meyer J (2001) Classification
and phylogeny of hydrogenases. FEMS Microbiol Rev
25, 455–501.
2 Vignais PM & Colbeau A (2004) Molecular biology
of microbial hydrogenases. Curr Issues Mol Biol 6,

9 Volbeda A, Garcin E, Piras C, De Lacey AL, Fernan-
dez VM, Hatchikian EC, Frey M & Fontecilla-Camps
JC (1996) Structure of the [NiFe]hydrogenase active site:
evidence for uncommon Fe-ligands. J Am Chem Soc
118, 12989–12996.
10 Garcin E, Vernede X, Hatchikian EC, Volbeda A, Frey
M & Fontecilla-Camps JC (1999) The crystal structure
of a reduced [NiFeSe]hydrogenase provides an image of
the activated catalytic center. Structure Fold Des 7,
557–566.
11 Higuchi Y, Ogata H, Miki K, Yasuoka N & Yagi T
(1999) Removal of the bridging ligand atom at the
Ni-Fe active site of [NiFe]hydrogenase upon reduction
with H
2
, as revealed by X-ray structure analysis at 1.4
A
˚
resolution. Structure Fold Des 7, 549–556.
12 Elsen S, Colbeau A, Chabert J & Vignais PM (1996)
The hupTUV operon is involved in negative control of
hydrogenase synthesis in Rhodobacter capsulatus.
J Bacteriol 178, 5174–5181.
13 Elsen S, Duche
´
O & Colbeau A (2003) Interaction
between the H
2
sensor HupUV and the histidine kinase
HupT controls HupSL hydrogenase synthesis in Rhodo-

ber of the subclass of regulatory hydrogenases. J Bacter-
iol 182, 2716–2724.
18 Vignais PM, Cournac L, Hatchikian EC, Elsen S,
Serebryakova N, Zorin NA & Dimon B (2002) Con-
tinuous monitoring of the activation and activity of
[NiFe]-hydrogenases by membrane-inlet mass spectro-
metry. Int J Hydrogen Energy 27, 1441–1448.
19 Colbeau A & Vignais PM (1992) Use of hupS, lacZ
gene fusion to study regulation of hydrogenase expres-
sion in Rhodobacter capsulatus : stimulation by H
2
.
J Bacteriol 174, 4258–4264.
20 Dischert W, Vignais PM & Colbeau A (1999) The
synthesis of Rhodobacter capsulatus HupSL hydrogenase
is regulated by the two-component HupT ⁄ HupR sys-
tem. Mol Microbiol 34 , 995–1006.
21 Buhrke T, Lenz O, Porthun A & Friedrich B (2004)
The H
2
-sensing complex of Ralstonia eutropha: interac-
tion between a regulatory [NiFe]hydrogenase and a his-
tidine protein kinase. Mol Microbiol 51, 1677–1689.
22 Vignais PM, Dimon B, Zorin NA, Tomiyama M &
Colbeau A (2000) Characterization of the hydrogen-
deuterium exchange activities of the energy-transducing
HupSL hydrogenase and H
2
-signaling HupUV hydro-
genase in Rhodobacter capsulatus. J Bacteriol 182, 5997–

by stopped-flow infrared spectroscopy. Biochemistry
43, 6820–6831.
30 George SJ, Kurkin S, Thorneley RNF & Albracht SPJ
(2004) Reaction of H
2
, CO, and O
2
with active [NiFe]-
hydrogenase from Allochromatium vinosum. A stopped-
flow infrared study. Biochemistry 43, 6808–6819.
31 Schneider K, Schlegel HJ, Cammack R & Hall DO
(1979) The iron-sulphur centres of soluble hydrogenase
from Alcaligenes eutrophus. Biochim Biophys Acta 578,
445–461.
32 Van der Linden E, Burgdorf T, Bernhard B, Bleijlevens
B, Friedrich B & Albracht SPJ (2004) The soluble
[NiFe]-hydrogenase from Ralstonia eutropha contains
four cyanides in its active site, one of which is respon-
sible for the insensitivity towards oxygen. J Biol Inorg
Chem 9, 616–626.
33 Happe RP, Roseboom W, Egert G, Friedrich CG, Mas-
sanz C, Friedrich B & Albracht SPJ (2000) Unusual
FTIR and EPR properties of the H
2
-activating site of
the cytoplasmic NAD-reducing hydrogenase from
Ralstonia eutropha. FEBS Lett 466, 259–263.
34 Buhrke T & Friedrich B (1998) hoxX (hypX) is a func-
tional member of the Alcaligenes eutrophus hyp gene
cluster. Arch Microbiol 170, 460–463.

FEBS Journal 272 (2005) 3899–3908 ª 2005 FEBS 3907
HYSCORE and ENDOR spectroscopy. J Am Chem
Soc 125, 13075–13083.
39 Haumann M, Porthun A, Buhrke T, Liebisch P, Meyer-
Klaucke W, Friedrich B & Dau H (2003) Hydrogen-
induced structural changes at the nickel site of the
regulatory [NiFe]hydrogenase from Ralstonia eutropha
detected by X-ray absorption spectroscopy. Biochemistry
42, 11004–11015.
40 Buhrke T, Lenz O, Krauss N & Friedrich B (2005) Oxy-
gen tolerance of the H
2
-sensing [NiFe] hydrogenase
from Ralstoniae utropho H16 is based on limited access
of oxygen to the active site. J Biol Chem in press.
41 Sambrook J, Fritsch EF & Maniatis T (1989) Molecular
Cloning: a Laboratory Manual, 2nd edn. Cold Spring
Harbor Laboratory Press, Cold Spring, N.Y.
42 Colbeau A, Godfroy A & Vignais PM (1986) Cloning
of DNA fragments carrying hydrogenase genes of Rho-
dopseudomonas capsulata. Biochimie 68, 147–155.
43 Colbeau A, Kelley BC & Vignais PM (1980) Hydro-
genase activity in Rhodopseudomonas capsulata: relation-
ship with nitrogenase activity. J Bacteriol 144, 141–148.
44 Cournac L, Guedeney G, Peltier G & Vignais PM
(2004) Sustained photoevolution of molecular hydrogen
in a mutant of Synechocystis sp. strain PCC 6803 defi-
cient in the type I NADPH-dehydrogenase complex.
J Bacteriol 186, 1737–1746.
45 Marrs B (1974) Genetic recombination in Rhodopseudo-