Tài liệu Báo cáo khoa học: Intrinsic GTPase activity of a bacterial twin-arginine translocation proofreading chaperone induced by domain swapping ppt

Intrinsic GTPase activity of a bacterial twin-arginine
translocation proofreading chaperone induced by domain
swapping
David Guymer
1
, Julien Maillard
2
, Mark F. Agacan
1
, Charles A. Brearley
3
and Frank Sargent
1
1 College of Life Sciences, University of Dundee, Dundee, UK
2 ENAC-ISTE ⁄ Laboratoire de Biotechnologie Environnementale (LBE), EPF Lausanne, Switzerland
3 School of Biological Sciences, University of East Anglia, Norwich, UK
Keywords
Escherichia coli; molecular chaperones;
noncanonical GTPase; TorD protein;
twin-arginine transport pathway
Correspondence
F. Sargent, Division of Molecular
Microbiology, College of Life Sciences,
University of Dundee, Dundee DD1
5EH, UK
Fax: +44 1382 388 216
Tel: +44 1382 386 463
E-mail:
(Received 28 July 2009, revised 12
November 2009, accepted 18 November
2009)

doi:10.1111/j.1742-4658.2009.07507.x
The bacterial twin-arginine translocation (Tat) system is a protein targeting
pathway dedicated to the transport of folded proteins across the cytoplas-
mic membrane. Proteins transported on the Tat pathway are synthesised as
precursors with N-terminal signal peptides containing a conserved amino
acid motif. In Escherichia coli, many Tat substrates contain prosthetic
groups and undergo cytoplasmic assembly processes prior to the transloca-
tion event. A pre-export ‘Tat proofreading’ process, mediated by signal
peptide-binding chaperones, is considered to prevent premature export of
some Tat-targeted proteins until all other assembly processes are complete.
TorD is a paradigm Tat proofreading chaperone and co-ordinates the mat-
uration and export of the periplasmic respiratory enzyme trimethylamine
N-oxide reductase (TorA). Although it is well established that TorD binds
directly to the TorA signal peptide, the mechanism of regulation or control
of binding is not understood. Previous structural analyses of TorD homo-
logues showed that these proteins can exist as monomeric and domain-
swapped dimeric forms. In the present study, we demonstrate that isolated
recombinant TorD exhibits a magnesium-dependent GTP hydrolytic activ-
ity, despite the absence of classical nucleotide-binding motifs in the protein.
TorD GTPase activity is shown to be present only in the domain-swapped
homodimeric form of the protein, thus defining a biochemical role for the
oligomerisation. Site-directed mutagenesis identified one TorD side-chain
(D68) that was important in substrate selectivity. A D68W variant TorD
protein was found to exhibit an ATPase activity not observed for native
TorD, and an in vivo assay established that this variant was defective in
the Tat proofreading process.
Structured digital abstract
l
MINT-7302371, MINT-7302377: TorD (uniprotkb:P36662) and TorD (uniprotkb:P36662)
bind (

MI:0407)bymolecular sieving (MI:0071)
l
MINT-7302402: TorD (uniprotkb:P36662) and TorD (uniprotkb:P36662 ) bind (MI:0407)by
comigration in non denaturing gel electrophoresis (
MI:0404)
l
MINT-7302387: TorD (uniprotkb:P36662) and TorD (uniprotkb:P36662 ) bind (MI:0407)by
cosedimentation in solution (
MI:0028)
Abbreviations
GAP, GTPase activating protein; IMAC, immobilised metal affinity chromatography; MGD, molybdopterin guanine dinucleotide;
Tat, twin-arginine translocation; TMAO, trimethylamine N-oxide; TorA, trimethylamine N-oxide reductase.
FEBS Journal 277 (2010) 511–525 ª 2009 The Authors Journal compilation ª 2009 FEBS 511
Introduction
The twin-arginine translocation (Tat) system is a pro-
tein-targeting pathway present in the cytoplasmic
membrane of many prokaryotes [1]. Tat-targeted pro-
teins are synthesised as precursors with cleavable
N-terminal signal peptides incorporating the distinctive
SRRxFLK ‘twin-arginine’ amino acid motif [2]. A key
feature of Tat translocation is the requirement for
physiological substrates to be fully folded before suc-
cessful translocation can occur [3]. Escherichia coli pro-
duces 27 Tat substrates [4], the majority of which bind
complex prosthetic groups, fold, activate, and often
oligomerise, in the cytoplasm prior to membrane trans-
location [1,3,5]. It is considered that the Tat translo-
case itself may be able to accept or reject pre-proteins
on the basis of their folded state in a ‘Tat quality con-
trol’ process [3]. In addition, some proteins are sub-

jected to a chaperone-mediated ‘Tat proofreading’
process in the cytoplasm prior to export. Tat proof-
reading involves the direct binding of particular Tat
signal peptides by dedicated chaperones aiming to pre-
vent premature targeting of immature proteins [6,7].
The Escherichia coli trimethylamine N-oxide (TMAO)
reductase (TorA) is a Tat-dependent periplasmic redox
enzyme encoded by the torCAD operon [8]. TorA is the
archetypal model Tat substrate synthesised with a 39
residue signal peptide and contains molybdopterin
guanine dinucleotide (MGD) as a single prosthetic
group [9]. Acquisition of the MGD cofactor by TorA in
the cell cytoplasm is an essential pre-requisite to translo-
cation of this protein [10]. The torD gene product is
involved in cofactor loading into TorA [11–18], and
additionally operates as a Tat proofreading chaperone,
binding directly to the TorA twin-arginine signal peptide
[5,19,20].
TorD belongs to a family of peptide-binding
proteins specifically dedicated to molybdoprotein
assembly [6,7,16,21]. Phylogenetic analysis of the TorD
family allows separation of the members into three
broad ‘clades’: the TorD clade, the DmsD clade and
the NarJ clade [16,21]. A number of 3D structures
now exist for members of the TorD family and these
all share a unique all a-helical architecture [22–25].
The helices are arranged into two domains (N-terminal
and C-terminal), which are connected by a hinge
region [23]. Crystal structures of monomeric forms
have been described [22,24,25]; however, higher-order

oligomers of TorD-like proteins have been observed
[26,27] and the crystal structure of a homodimer has
been solved [23]. Dimerisation of the Shewanella massi-
lia TorD protein is driven by ‘domain-swapping’ in
which the N-terminal domain of one protomer packs
onto the C-terminal domain of a second protomer
(and vice versa); however, the physiological role of this
domain-swapping was not clear [23].
The isolated, recombinant E. coli TorD monomer
has been shown to bind to the TorA signal peptide
in vitro with an apparent dissociation constant (K
d
)of
59 nm [19]. Such relatively tight binding led to some
speculation as to how binding and release cycles could
be regulated. The crystal structure of the Sh. massilia
TorD homodimer was observed to contain tightly-
bound oxidised (and therefore cyclic) dithiothreitol
[23]. As a result, Hatzixanthis et al. [20] hypothesised
that this observation could suggest that a common cyc-
lic regulatory molecule, perhaps a nucleotide, could
normally be bound by TorD. Indeed, the monomeric
form of the E. coli TorD protein was subsequently
shown to bind guanine nucleotides with low affinity
(apparent K
d
 370 lm for GTP) [20], and recent inde-
pendent computational analysis predicted a potential
GTP-binding site on the DmsD protein from Salmo-
nella enterica serovar Typhimurium [24].

In the present study, the relationship between E. coli
TorD and GTP was investigated. We demonstrate that
recombinant TorD possesses an intrinsic magnesium-
dependent GTP hydrolysis activity in vitro.Itis
revealed that this activity is a property only of the
dimeric form of the protein, suggesting that domain-
swapping is required to generate the active site. In
addition, site-directed mutagenesis identifies one resi-
due, D-68, that is involved in the substrate selectivity
of this protein.
Results
TorD has magnesium-dependent GTPase activity
Early ligand-binding experiments [20] and recent bioin-
formatic analysis [24] suggested that TorD-like proteins
may bind GTP. In addition, overproduction of TorD
family proteins has been reported to result in the isola-
tion of a range of stable homo-oligomeric forms
[26,27]. In the initial development of our purification
strategies, we established that an ion exchange chro-
matographic step, immediately following metal affinity
chromatography, resulted in isolation of stable mono-
meric TorD
his
[20,28]. This monomeric form was ideal
for biophysical studies [19,20] but, despite its ability to
bind guanine nucleotides, no nucleotide hydrolysis
activity could be detected [20]. We therefore decided to
explore the possibility that different folding forms of
TorD
his

may harbour different biological activities.
The E. coli TorD homodimer has GTPase activity D. Guymer et al.
512 FEBS Journal 277 (2010) 511–525 ª 2009 The Authors Journal compilation ª 2009 FEBS
First, E. coli TorD
his
was overproduced and isolated
by nickel-affinity chromatography. Eluate from the
metal affinity column was then assayed directly for GTP
and ATP hydrolytic activity using a malachite green
method for the quantification of free inorganic phos-
phate (P
i
). This assay measures free P
i
in solution by
spectrophotometric determination of the complex
formed between malachite green, molybdate and free P
i
.
The initial assay chosen already included magnesium
chloride in the reaction mixture because magnesium ions
are essential for the GTPase activity of the majority of
characterised GTPases [29,30]. The assay demonstrated
that TorD exhibits hydrolytic activity towards GTP
(Fig. 1A). No P
i
release was detected in the negative con-
trols, which included a sample of the elution buffer used
in the chromatographic experiment and a sample of a
maltose-binding protein isolated in the identical buffers,

and on the same immobilised metal affinity chromatog-
raphy (IMAC) column, as the TorD
his
described here
(Fig. 1A). In addition, our recombinant TorD
his
shows
no hydrolytic activity towards ATP (Fig. 1A).
The binding of a magnesium cofactor has long been
established as essential for the activity of canonical
GTP-binding proteins [29,31]. The precise role of
Mg
2+
may vary between enzymes but is clearly essen-
tial for GTP hydrolysis and is almost always required
to allow GTP binding [29,30]. To determine the
requirement for magnesium of the GTPase activity of
TorD, the reaction was performed in varying amounts
of MgCl
2
(Fig. 1B). The GTPase activity of TorD
without added MgCl
2
is negligible, establishing that
magnesium is required for TorD GTPase activity
(Fig. 1B). Titration of increasing amounts of MgCl
2
gradually enhances GTPase activity, with a peak at
approximately 1 mm (Fig. 1B). The malachite green
assay was also performed with 1.2 mm MgCl

2
in the
presence of 10 mm EDTA. In this experiment, the
presence of EDTA completely inhibited the reaction
(data not shown). Furthermore, no other divalent
cations, including manganese, could replace magne-
sium in this assay (not shown).
Finally, the product of the GTP hydrolysis reaction
catalysed by TorD was shown to be GDP by HPLC
analysis (Fig. S1). Taken altogether, these data demon-
strate the initial identification of a strictly magnesium-
dependent GTPase activity associated with the IMAC
pool of recombinant TorD protein.
GTP hydrolytic activity is a feature of the TorD
homodimer
Having established that GTPase activity was associ-
ated with TorD collected immediately following metal
affinity chromatography, the next step was to further
purify the hydrolytic activity by alternative chromato-
graphic techniques. Size exclusion chromatography
using a SuperdexÔ 75 column identified a range of
molecular mass species present in the nickel-affinity
purified TorD sample (Fig. 2A). The major peak corre-
sponded to an approximate molecular mass of
25.5 kDa, in agreement with the predicted molecular
mass of TorD
his
of 24.2 kDa, and so likely represents
monomeric TorD. Lesser protein peaks representing
TorD species with approximate molecular mass of

Fig. 1. TorD has magnesium-dependent GTPase activity. (A) The
malachite green assay for in vitro P
i
release from nucleotides.
Reaction mixtures [50 lL aliquots containing 5 m
M GTP or ATP,
1.2 m
M MgCl
2
and 10 mM Tris–HCl (pH 7.5)] were incubated at
22 °C for 24 h containing 0.1 m
M of metal affinity chromatography-
purified TorD (‘TorD + GTP’ and ‘TorD + ATP’). As controls, an
equal amount of a His-tagged maltose binding protein-TorA signal
peptide fusion (‘MalE:TorA-SP + GTP’), or protein-free column
buffer (‘no protein’), were also assayed in the presence of 5 m
M
GTP. (B) The GTP hydrolysis reaction is magnesium dependent.
Reactions containing 0.1 m
M of metal affinity chromatography-puri-
fied TorD with 5 m
M GTP buffered in 10 mM Tris–HCl (pH 7.5)
were incubated at 22 °C for 24 h in the presence of increasing
amounts of MgCl
2
. Released P
i
was assayed by the malachite
green method. In both (A) and (B), the total phosphate released in
each reaction is shown and error bars represent the SEM (n = 3).

D. Guymer et al. The E. coli TorD homodimer has GTPase activity
FEBS Journal 277 (2010) 511–525 ª 2009 The Authors Journal compilation ª 2009 FEBS 513
49.0 kDa, similar to the predicted 48.5 kDa of a dimer
of TorD
his
, and higher-order oligomers at approxi-
mately 85.7 kDa (beyond the linear range of the
SuperdexÔ 75 column of 3–70 kDa) were also
observed (Fig. 2A). The presence of TorD oligomers is
supported by gel electrophoresis (Fig. 2B, C). Denatur-
ing SDS–PAGE showed the presence of TorD polypep-
tide in all fractions tested, whereas PAGE performed in
the absence of SDS allowed a ready visualisation of the
different oligomeric forms of TorD present in the higher
molecular mass fractions (Fig. 2B, C). Magnesium-
dependent GTPase activity was restricted to the higher
molecular mass fractions, and was completely absent
from the monomer form (data not shown).
CibacronÔ Blue F3G-A is a dye molecule that can
be immobilised to a Sepharose matrix (Blue Sepha-
roseÔ HP), and which is able to bind specifically to
some nucleotide-binding proteins as a result of its
structural similarity to nucleotide cofactors. Specifi-
cally-bound proteins are then normally eluted by the
application of either an amount of cofactor or by
increasing the ionic strength. Initial experiments with
TorD under ‘standard’ conditions for analysing classi-
cal nucleotide-binding proteins [32] suggested that the
protein did not bind tightly to Cibacron Blue under
conditions of low ionic strength, and that nothing was

therefore eluted upon application of a high ionic
strength solution (data not shown). Surprisingly, how-
ever, upon washing the Blue Sepharose column in
water after each experiment, a small peak of protein
was observed in the eluate. To explore this further, the
buffer conditions were changed and the metal affinity
pool of TorD was applied to a Blue Sepharose column
equilibrated in buffer containing 0.5 m NaCl. The elu-
tion profile revealed that the majority of the sample
did not bind to the CibacronÔ Blue dye (Fig. 3A).
However, upon switching from the relatively high salt
A
B
C
Fig. 2. Oligomeric forms of E. coli TorD can be readily identified.
(A) Elution profile of a pool of TorD derived from metal affinity chro-
matography applied to a SuperdexÔ 75 (10 ⁄ 30) size exclusion col-
umn at 0.5 mLÆmin
)1
in 50 mM Tris–HCl (pH 7.5) and 200 mM NaCl.
Values for the apparent molecular mass of peak proteins were cal-
culated using control proteins of known molecular mass (not
shown). (B) Equivalent volumes (5 lL) of the protein fractions indi-
cated were diluted 1 : 1 in either Laemmli or native sample buffer
and subjected directly (unboiled) to SDS–PAGE (top panel) and non-
SDS–PAGE (bottom panel) on 12.5% (w ⁄ v) polyacrylamide gels.
Protein bands were visualised with Coomassie Brilliant Blue R-250
stain. (C) Equivalent amounts (3.6 lg) of protein in each indicated
fraction were separated directly (unboiled) by SDS–PAGE (top
panel) and non-SDS–PAGE (bottom panel) on 12.5% (w ⁄ v) poly-

acrylamide gels and stained with Coomassie R-250.
The E. coli TorD homodimer has GTPase activity D. Guymer et al.
514 FEBS Journal 277 (2010) 511–525 ª 2009 The Authors Journal compilation ª 2009 FEBS
concentration buffer (0.5 m) to a very low ionic
strength solution (pure water in this case), a small
peak of protein was eluted from the column (Fig. 3A).
SDS–PAGE and western analysis revealed that this
fraction contained TorD protein (Fig. 3B, C), and
TorD isolated in this way is referred to in the present
study as ‘TorD
Blue
’.
The TorD
Blue
protein peak isolated by the ‘reverse’
Blue Sepharose chromatography protocol was analysed
for GTPase activity using the malachite green assay
(Fig. 3D). Very interestingly, TorD
Blue
demonstrated
GTPase activity, whereas the ‘flow-through’ fraction of
TorD showed negligible activity (Fig. 3D). Control
experiments using a column fraction containing no
protein also showed no activity (Fig. 3D). Very unusu-
ally, therefore, all of the GTPase activity was bound
to the Cibacron Blue column in the presence of 0.5 m
salt (something that a ‘canonical’ nucleotide-binding
protein would not do) and, subsequently, all of the
GTPase activity could be eluted in a solution of very
low ionic strength (again not the typical behaviour of

a classical nucleotide-binding protein).
The relative purity of the TorD
Blue
preparation was
analysed further by SDS–PAGE and MS. Both the
TorD
his
fraction (from the initial metal affinity chro-
matography pool) and the TorD
Blue
preparation were
separated by SDS–PAGE and the proteins were visual-
ised using the most sensitive silver staining method
available (Fig. 3E). This method revealed a single
A
B
C
D
E
1400
1200
1000
800
600
Absorbance units (280 nm)
V
e
(ml)
Conductance units (mS·cm
–1

)
400
200
0
12 17 22 32 37
0
10
20
30
40
50
60
27
Fig. 3. TorD GTPase activity can be isolated by Cibacron Blue affin-
ity chromatography. (A) Unusual behaviour of TorD on CibacronÔ
Blue affinity media. A sample of 0.5 m
M metal affinity chromatogra-
phy-purified TorD was applied to a 1 mL HiTrapÔ Blue column,
attached to an FPLC system, in 0.5
M NaCl-containing buffer at
1mLÆmin
)1
. Bound proteins were eluted in a single step to pure
water. (B) Protein fractions from the unbound flow-through (‘FT’) or
single 1 mL fractions eluted in water (numbered 35–38) were
diluted 1 : 1 in either Laemmli or ‘native’ sample buffer and sub-
jected (unboiled) to SDS–PAGE (top panel) and non-SDS–PAGE
(bottom panel). Three microlitres of the flow-through sample was
used to give an equivalent amount of protein loaded compared to
that of 36 mL fraction ( 1.1 lg). Gels were stained with Coomas-

sie R-250. (C) A western immunoblot was carried out on 6 ng sam-
ples of the original metal affinity-purified material (‘IMAC’), the
unbound flow-through (‘FT’), and pooled fractions 36–38. Proteins
were mixed with Laemmli disaggregation buffer and heat-treated at
100 °C for 2 min before being separated by SDS–PAGE, blotted
onto nitrocellulose, and challenged with an anti-TorD serum
(1 : 10 000 dilution). (D) Analysis of the ‘TorD
Blue
’ fractions for
GTPase activity. Protein samples (1.2 lg) of fractions eluted in
water from the Cibacron Blue Sepharose column were subjected to
the malachite green GTPase assay. The fraction at 33 mL contained
no detectable protein and was assayed as an internal control to
establish that incubation with the column buffers alone did not facil-
itate P
i
release. Error bars indicate the SEM (n = 3). (E) Pooled pro-
tein samples (4 lg) from the original metal affinity column (‘IMAC’)
and the combined eluate from the CibacronÔ Blue affinity column
(‘Blue’) were mixed with Laemmli disaggregation buffer and heat-
treated at 100 °C for 2 min before being separated by SDS–PAGE
using Bio-Rad (Bio-Rad, Hercules, CA, USA) pre-cast 15% (w ⁄ v)
polyacrylamide gels. Proteins were visualised using the Silver-
quest
Ò
(Invitrogen) silver-staining kit and, where indicated, the reac-
tion was stopped after 2 or 4 min.
D. Guymer et al. The E. coli TorD homodimer has GTPase activity
FEBS Journal 277 (2010) 511–525 ª 2009 The Authors Journal compilation ª 2009 FEBS 515
strong band in each fraction corresponding to the

TorD protein (Fig. 3E). In addition, the TorD
Blue
preparation was subjected to MS analysis (Fig. S2).
Recombinant TorD was found to the dominant species
in this preparation and was intact, except for partial
modifications to the initiator methionine (Met-1),
which are common in bacterial systems (Fig. S2).
Thus, from a total of 45 mg of recombinant TorD
his
isolated by metal affinity chromatography and applied
to the Cibacron Blue column, 2.3 mg of the TorD
Blue
protein harbouring all of the GTPase activity was
recovered.
Further analysis of the two different TorD pools by
SDS–PAGE and native PAGE suggested that the key
difference lay in the oligomeric state of the proteins.
The TorD population that failed to bind to the Ciba-
cron
TM
Blue migrated at a low apparent molecular
mass under native conditions (Fig. 3B), displaying sim-
ilar behaviour to the monomer form of TorD charac-
terised by molecular exclusion chromatography
(Fig. 2). However, the GTPase-active TorD
Blue
clearly
comprised oligomeric TorD (Fig. 3B). The precise olig-
omeric state of the TorD
Blue

species was therefore
investigated by analytical ultracentrifugation.
Sedimentation velocity analysis was performed on a
sample taken from the initial nickel-affinity chroma-
tography and this material was shown to contain
predominantly monomeric TorD with detectable quan-
tities of dimer and higher oligomeric species (Fig. 4A).
By contrast, however, analytical ultracentrifugation
reveals that the GTPase-active TorD
Blue
species iso-
lated here contains no monomeric TorD at all
(Fig. 4B). Instead, it was unequivocally established
that TorD
Blue
is dominated by the homodimeric form
of the protein (Fig. 4B). Trace amounts of higher-
order oligomeric species are also present in this sample
(Fig. 4B).
Taken together, these data strongly indicate that
gross overproduction of recombinant TorD results in a
mixed population comprising numerous different oligo-
meric forms. A small sub-population of proteins
(approximately 5% by mass of the total) was found to
adopt a stable homodimeric conformation that har-
bours a magnesium-dependent GTPase activity.
Kinetics of the GTPase reaction
The Cibacron Blue affinity chromatography approach
presented the opportunity to study the kinetics of GTP
hydrolysis catalysed by the active TorD fraction. For

these experiments, P
i
release from GTP hydrolysis by
TorD
Blue
was assayed continuously using the Enz-
Chek
Ò
Phosphate Assay Kit (Invitrogen, Carlsbad,
CA, USA). This coupled assay is based on the proto-
col described by Webb and Hunter [33] whereby, in
the presence of P
i
, 2-amino-6-mercapto-7-methylpurine
riboside is converted by purine nucleoside phosphory-
lase to ribose-1-phosphate and 2-amino-6-mercapto-
7-methylpurine. The product is detected by an increase
in A
360
.
The GTPase activity of TorD
Blue
in a range of GTP
concentrations was assayed in 96-well plates, the
P
i
-dependent product measured continuously at A
360
and a curve plotted of the initial velocities (V
0

) against
the substrate concentration (Fig. 5A). The best fit to
Fig. 4. The homodimer form of TorD dominates the GTPase-active
fraction. Sedimentation velocity analytical ultracentrifugation profiles
of (A) the TorD pool immediately following metal affinity chroma-
tography at 0.5 mgÆmL
)1
, rmsd = 0.024682, f ⁄ f
0
= 1.199770, and
(B) the TorD
Blue
pool immediately following Cibacron Blue aff-
inity chromatography at 0.5 mgÆmL
)1
, rmsd = 0.024537, f ⁄ f
0
=
1.201629. Samples at 0.25 and 0.75 mgÆmL
)1
were also analysed
(not shown) and gave similar profiles.
The E. coli TorD homodimer has GTPase activity D. Guymer et al.
516 FEBS Journal 277 (2010) 511–525 ª 2009 The Authors Journal compilation ª 2009 FEBS
these data resulted in a sigmoidal curve (Fig. 5A),
indicative of a co-operative substrate binding and
hydrolysis model with non Michaelis–Menten kinetics.
To extract the kinetic parameters from these data, a
Hill plot (Fig. 5B) was drawn of V
0

)(V
max
⁄ V
0
) against
log
10
[GTP], using an estimate for V
max
of 39 lmÆmin
)1
obtained from the linear plot (Fig. 5A). A line of best
fit revealed the values for K
m
and the Hill coefficient,
h. A value of 2.13 for h was obtained, which indicates
positive co-operativity of GTP binding and hydrolysis
by the TorD homodimer. TorD was calculated to
hydrolyse GTP with a K
m
of 1.42 mm and a K
cat
of
3.9 min
)1
, both contributing to a very low specificity
constant (K
cat
⁄ K
m

) of 45.77 M
)1
Æs
)1
.
TorD residue D-68 confers substrate specificity
Analysis of the primary amino acid sequence data
suggests that TorD differs substantially from any
canonical GTPase enzymes. For example, regulatory
GTP hydrolases (‘G-proteins’) all possess a conserved
guanine nucleotide-binding domain, with the five
polypeptide loops that comprise the guanine nucleo-
tide-binding site being the most highly conserved
elements that define the GTPase superfamily [31].
These five loops are designated G-1 to G-5, with G-1,
G-3 and G-4 being present in all canonical GTPases
[34,35]. G-1 corresponds to the ‘Walker A motif’ [36]
and includes the consensus sequence Gx
4
GK[S ⁄ T]. G-2
(Dx
n
T) is a structurally mobile element and, in many
GTPases, GTP-binding alters the conformation of this
loop, bringing a conserved threonine essential for
Mg
2+
co-ordination into a position to facilitate GTP
hydrolysis [35]. G-3 corresponds to the ‘Walker B
motif’ [36] and includes the consensus sequence Dx

2
G
[35]. Although the G-1 and G-3 consensus motifs are
found in many nucleotide-binding proteins, it is the
G-4 motif that provides the specificity for guanine nu-
cleotides. The characteristic sequence motif of G-4 is
[N ⁄ T][K ⁄ Q]xD [34,37,38] and it is often preceded by a
stretch of four hydrophobic or nonpolar amino acids
[35]. Finally, G-5 ([T ⁄ G][C ⁄ S]A) is less well conserved
than the other motifs and cannot always be unambigu-
ously identified from primary sequence data alone [35].
TorD lacks each of the canonical G-1 and G-3
motifs that are considered essential for GTP recogni-
tion and hydrolysis. However, examination of the
amino acid sequence reveals a potential candidate for
a G-4 guanine specificity motif in TorD. Praefcke et al.
[39] described the human guanylate-binding protein
that has guanine specificity conferred by a G-4 motif
with the sequence TLRD. In addition, the homologous
GBP in chicken contained a TVRD motif at this posi-
tion. This is of particular interest because the E. coli
TorD possesses a TVRD tetrapeptide at positions 65–
68, which is predicted to be located on an exposed
loop between helix 4 and helix 5.
The residues of this putative G-4 motif were sub-
jected to site-directed mutagenesis. The focus was resi-
due D-68, the final residue in the TVRD motif,
because examples have been described where substitu-
tions at this position have altered the substrate speci-
ficity of canonical GTPases [34,37,39–42]. In classical

GTPases, the aspartate is considered to form a hydro-
A
B
Fig. 5. Kinetics of GTP hydrolysis by the TorD homodimer. (A) A linear
plot of initial reaction rate, V
0
(released P
i
; lMÆmin
)1
) against GTP con-
centration (m
M) plotted using ORIGIN software (OriginLab Corporation,
Northampton, MA, USA). Aliquots (10 l
M) of a pooled fraction of
TorD
Blue
isolated by Cibacron Blue affinity chromotagraphy were
asayed in a 96-well plate format at 37 °C with shaking in the presence
of increasing amounts of substrate and the formed product measured
continuously at A
360
using the EnzCheck assay. (B) A Hill plot of the
same data shown in (A) converted into log (V
0
– V
max
⁄ V
0
) against

log[GTP]. V
max
was estimated as 39 lMÆmin
)1
. The straight line was
calculated by Microsoft Excel (Microsoft Corp., Redmond, WA, USA)
using liner regression: y = 2.1302x – 0.3237 (R
2
= 0.9478). From
these data, the Hill coefficient (equal to the gradient of the line),
h = 2.13, K
m
[10
(y ⁄ )h)
] = 1.42 mM, K
cat
= 3.9 min
)1
, and K
cat
⁄ K
m
=
45.77
M
)1
Æs
)1
.
D. Guymer et al. The E. coli TorD homodimer has GTPase activity

FEBS Journal 277 (2010) 511–525 ª 2009 The Authors Journal compilation ª 2009 FEBS 517
gen bond with the amino group at position 2 of the
guanine base, which is absent from the xanthine base
[39,40]. Thus, D-68 was substituted by asparagine and
the hydrolytic activity of the TorD
D68N
variant towards
GTP and XTP was measured using the malachite green
assay (Fig. 6A, B). TorD
D68N
was unaltered for
GTPase activity, although this variant demonstrated
enhanced XTPase activity (Fig. 6A, B). To test for an
increased affinity for xanthine nucleotides by a different
means, the standard malachite green assay was modi-
fied to perform an assay measuring the effect of compe-
tition of a ten-fold excess of XMP on the GTPase
activities of the native protein and the TorD
D68N
vari-
ant. The GTPase activity of TorD
D68N
was found to be
inhibited to a greater degree than the native TorD by
the presence of excess XMP (Fig. 6B).
Examples exist of naturally-occurring GTPases
where the G-4 consensus aspartate is substituted by
tryptophan, and these are able to hydrolyse ITP in
addition to GTP [34,37,41]. A TorD D68W variant
was therefore tested for its ITPase and ATPase activity

(Fig. 6C). The TorD
D68W
protein showed no obvious
increase in ITPase activity compared to the native pro-
tein, which could clearly hydrolyse ITP already
(Fig. 6C). More interestingly, however, TorD
D68W
was
observed to possess a hydrolytic activity towards ATP
(Fig. 6C), a substrate that native TorD was unable to
recognise (Figs 1A and 6C). Taken together, these data
implicate the TorD ‘G-4’ motif in playing a key role in
substrate selectivity for this enzyme, especially with
respect to the ability of the enzyme to distinguish
between GTP and ATP.
A TorD D68W variant is defective in the Tat
proofreading process
The physiological role of TorD residue D-68 was
tested in vivo. TorD has two physiological functions
that can be independently measured and assays have
been developed to study the overall biosynthesis of the
TorA enzyme, as well as the isolated Tat proofreading
activity. First, the ability of the torD gene to rescue
TMAO reductase activity in a DtorD mutant when
expressed in trans was explored. A chromosomal dele-
tion strain (FTD100) was transformed with pUNI-
PROM derivatives encoding native TorD and
TorD
D68W
. The strains were grown anaerobically in

LB supplemented with glycerol and TMAO and the
benzyl viologen-linked TMAO reductase activity of
periplasmic fractions was assayed (Fig. 7A). TorD
D68W
was observed to support assembly of the periplasmic
TMAO reductase activity to a level equivalent to
native TorD (Fig. 7A).
120%
A
B
C
100%
80%
60%
40%
NTPase activity (% GTPase)
NTPase activity (% GTPase)
20%
0%
120%
100%
80%
60%
40%
20%
0%
4.0
3.5
3.0
2.5

2.0
1.5
1.0
0.5
0.0
02
wtTorD + GTP
wtTorD + GTP
wtTorD + ATP
wtTorD + ITP
D68W + GTP
D68W + ATP
D68W + ITP
wtTorD + GTP + XMP
D68N + GTP D68N + GTP + XMP
46
Elapsed time (h)
[P
i
] mM
81012
wtTorD – GTP wtTorD – XTP D68N – GTP D68N – XTP
Fig. 6. TorD residue D68 controls substrate specificity. (A) The
GTPase and XTPase activities of 0.1 m
M ( 0.122 mg) native TorD
and the D68N variant purified by immobilised metal affinity chroma-
tography were assayed by the malachite green method in 50 lL reac-
tions containing either 5 m
M GTP or XTP, 1.2 mM MgCl
2,

and 10 mM
Tris–HCl (pH 7.5). Reactions were incubated at 22 °C for 24 h. The
results are shown as a percentage of the native TorD GTPase activ-
ity. (B) A timecourse competition ⁄ inhibition assay of GTPase activity
using 0.1 m
M metal affinity chromatography-purified TorD and
TorD
D68N
in 100 lL reactions containing 10 mM Tris–HCl (pH 7.5),
1.2 m
M MgCl
2
and 5 mM GTP ± 50 mM XMP. Phosphate release
was quantified at 2 h intervals by withdrawing 10 lL aliquots and
subjecting those to the malachite green assay. (C) Nucleotide hydro-
lysis activities of 0.1 m
M samples of metal affinity chromatography-
purified TorD and TorD
D68W
assayed by the malachite green method.
Each 50 lL reaction contained 5 m
M of either GTP, ATP, or ITP as
well as 1.2 m
M MgCl
2
and 10 mM Tris–HCl (pH 7.5). Reactions were
incubated at 22 °C for 24 h and the results are shown as a percent-
age of GTPase activity exhibited by native TorD. In all cases, the error
bars represent the SEM (n = 3).
The E. coli TorD homodimer has GTPase activity D. Guymer et al.

518 FEBS Journal 277 (2010) 511–525 ª 2009 The Authors Journal compilation ª 2009 FEBS
Next, a specific assay for Tat proofreading was
employed. Jack et al. [5] developed an assay based on
a strain (RJ607) producing a TorA-signal-peptide-
HybO fusion protein. Cells producing the TorA-HybO
fusion have impaired hydrogenase-2 activity because
assembly of the enzyme is disrupted; however, co-
expression of active torD restores the Tat proofreading
of this enzyme and so rescues hydrogenase-2 activity
in the mutant strain. RJ607 was transformed with a
pUNIPROM vector encoding TorD and TorD
D68W
,
grown anaerobically in LB supplemented with glycerol
and fumarate, and benzyl viologen-linked hydroge-
nase-2 activity was assayed in whole cells (Fig. 7B).
Interestingly, TorD
D68W
was observed to have a
reduced Tat proofreading activity in vivo.
Discussion
The crystal structure of Sh. massilia TorD is a homod-
imer formed through 3D domain swapping [23]. It has
been established in the present study, in common with
other TorD homologues [26,27], that E. coli TorD can
also be purified in a range of stable oligomeric forms,
suggesting that oligomerisation may be a characteristic
feature of the TorD family. Most significantly, this
work has now established the biochemical relevance of
the dimerisation exhibited by TorD. The E. coli TorD

homodimer displays an intrinsic specific GTPase activ-
ity, whereas the monomer form remains completely
inactive. In terms of kinetics, the sigmoidal curve of
V
0
versus [GTP] is indicative of a co-operative binding
and hydrolysis model, and the Hill coefficient (h)of
2.13 implies a strongly positive co-operative binding
model whereby binding and hydrolysis of GTP at one
site enhances the affinity (or activity) towards GTP at
other sites. The Hill coefficient also enables an estima-
tion of the number of hydrolytic sites present in the
enzyme because current dogma states that h cannot be
greater than the number of ligand-binding sites present
within the molecule. This suggests either that the active
form of TorD could contain at least three active sites,
or that the active oligomeric form could be larger than
dimeric. Indeed, the TorD protein from Sh. massilia
has been previously observed as stable trimers [27].
Note that h provides only a minimum estimate for the
number of binding sites. Haemoglobin, for example,
shows a Hill coefficient for oxygen binding varying in
the range 1.7–3.0, but clearly possesses four binding
sites [43,44].
The discovery of an enzymatic activity associated
with the domain swapped TorD dimmer, which is
absent from the monomer, brings E. coli TorD sharply
into line with other biological systems that utilise
domain swapping to regulate activity. For example,
glyoxylase I of Pseudumonas putida is a metastable

domain-swapped dimer with two Zn
2+
cofactors that
also exists as a monomer with a significantly reduced
activity and only a single Zn
2+
cofactor [45]. An addi-
tional example is the bleomycin resistance protein,
Fig. 7. Residue D68 is involved in the Tat proofreading process.
Physiological activity of the TorD
D68W
variant. (A) E. coli strain
FTD100 (DtorD) was transformed with a pUNIPROM vector
expressing either torD (‘TorD’), or the D68W mutant (‘D68W’), and
grown anaerobically in LB containing 0.5% (v ⁄ v) glycerol and 0.4%
(w ⁄ v) TMAO before intact cells were assayed for TMAO:BV oxido-
reductase activity. (B) E. coli strain RJ607 (/torA::hybO, DhybA)
was transformed with a pUNIPROM vector expressing either torD
(‘TorD’), or the D68W mutant (‘D68W’), and grown anaerobically in
LB containing 0.5% (v ⁄ v) glycerol and 0.4% (w ⁄ v) fumarate before
whole cells were assayed for hydrogen:BV oxidoreductase activity.
In all cases, the error bars represent the SEM (n = 3).
D. Guymer et al. The E. coli TorD homodimer has GTPase activity
FEBS Journal 277 (2010) 511–525 ª 2009 The Authors Journal compilation ª 2009 FEBS 519
which sequesters two molecules of the bleomycin anti-
biotic within two crevices formed at the interface of
the domain swapped dimer [46]. Equally intriguing is
the T7 endonuclease I, representing one of the most
striking examples of a domain-swapped dimer, which
comprises a composite catalytic site containing ele-

ments from both polypeptide chains [47,48]. It is
tempting to speculate that TorD dimerisation generates
an analogous composite active site, which would
enable the dimer to acquire a completely new biologi-
cal function through domain swapping. Indeed, such
acquisition of a novel function would be consistent
with the theory that domain swapping provides a
model for the evolution of oligomeric enzymes [49,50].
Although the mechanism of GTP hydrolysis by the
TorD dimer remains to be elucidated, there are aspects
of this activity that are mirrored in canonical nucleo-
tide hydrolases. For example, the GTPase activity of
TorD is dependent upon the presence of magnesium
ions and this is clearly in line with the requirements of
the classical GTPases [29,30]. Magnesium often func-
tions as a phosphate ligand in canonical GTPases in
addition to a variety of proposed activities involved in
the actual hydrolysis reaction [30,31,51]. Although the
exact nature of the GTP-binding site remains to be
determined, it is clear that TorD represents a novel
family in this regard. This is reinforced by the observa-
tion that active E. coli TorD behaves very unusually
on Cibacron Blue affinity media, eluting only under
conditions of low ionic strength, perhaps suggesting
hydrophobic interactions between the cofactor and
protein dominate. Interestingly, a more typical G-4
‘guanine specificity’ loop was identified in TorD, and
the D-68 side chain unequivocally was demonstrated
to regulate substrate access to the active site(s). The
E. coli TorD sequence in question, 65-TVRD-68, is a

close match to the G-4 consensus [N ⁄ T][K ⁄ Q]xD
[34,37,38]. This motif is conserved in the TorD and
DmsD clades of the wider TorD family and is always
located in an unstructured loop between helices 4 and
5 (Fig. 8). The final aspartate in the TVRD tetrapep-
tide is occasionally naturally replaced by glutamine or
glutamate (Fig. 8G), although these side chains should
still able to confer guanine specificity in classical GTP-
ases [41]. Note also that the TorD G-4 motif is often
immediately followed by another conserved aspartate
(Fig. 8G), which could also have a role in determining
nucleotide specificity.
This TorD
D68W
variant was unaffected in its GTPase
activity but, in addition, showed a new ATPase activity.
The fact that this protein was defective in the Tat proof-
reading process is intriguing. Why would ATP hydroly-
sis not be able to substitute for GTP hydrolysis in vivo?
Because the ATP hydrolysis reaction catalysed by the
variant TorD form is obviously slower, and the concen-
tration of ATP in the cell is higher than GTP, it is possi-
ble that ATP is inhibiting the Tat proofreading function
of TorD
D68W
in vivo
.
In an attempt to identify catalytic residues essential
for GTP hydrolysis, extensive mutagenesis was con-
ducted on E. coli torD (Fig. S3) based on recent bio-

chemical and in silico experiments [5,19,20,24].
However, from the 17 different TorD variants tested,
none were found to be completely devoid of in vitro
GTPase activity (Fig. S3). TorD variants Q7L, as orig-
inally identified by Buchanan et al. [19], and F41W
[20] showed the lowest hydrolytic activity towards
GTP (Fig. S3), and these side-chains are in close prox-
imity to the G-4 loop (Fig. 8).
What is the purpose of the GTPase activity of
TorD? GTP hydrolysis is often used to regulate pro-
cesses where multiple proteins must function co-ordi-
nately [52] and the maturation of TorA is certainly a
complex process involving the co-ordination of the two
distinct functions of TorD with the folding of TorA
and possible interactions of both proteins with the
MGD biosynthetic apparatus [12]. The concentration
of GTP in the cytoplasm of an exponentially growing
E. coli cell has been estimated at approximately
0.9 mm [53]. This is below the K
m
of GTP hydrolysis
by TorD and perhaps suggests that TorD has only a
low level of this activity in vivo. Most GTPases bind
very stably to GTP and have a very low intrinsic level
of GTPase activity, and also require the action of
GTPase activating proteins (GAPs), or the interaction
with a specific effecter, to catalyse the hydrolysis of
GTP. The transition from GTP- to GDP-bound forms
can therefore be limited by the intrinsic rate of GTP
hydrolysis, by the action of a GAP, or by interaction

with a cognate target. GTPases governed by these fac-
tors are often referred to as ‘clocks’, ‘switches ⁄ adap-
tors’ and ‘sensors’, respectively [31]. The turnover
number of TorD for GTP of 3.9 min
)1
is within the
range expected for classical GTPases [31] but is very
low in general terms, and it is possible that TorD may
function as a ‘clock’ with the low rate of GTP hydro-
lysis coinciding with the maturation rate of TorA.
Another possibility is that arginines of the signal pep-
tide, or perhaps residue R22 of TorD [20], may be
contributing an ‘arginine finger’, thus activating the
GTPase activity in response to signal binding. Arginine
fingers have been observed within the Ffh-FtsY com-
posite active site and provided by the Ras and Rho
GTPase GAPs [54–56]. However, the experiments
conducted in the present study were unable to demon-
strate an increase in the rate of GTP hydrolysis in
The E. coli TorD homodimer has GTPase activity D. Guymer et al.
520 FEBS Journal 277 (2010) 511–525 ª 2009 The Authors Journal compilation ª 2009 FEBS
response to addition of excess TorA signal peptide
(data not shown). There is currently no evidence for
the requirement of a specific GAP for TorD GTPase
in vitro, although it remains a possibility that such
may exist in vivo and further work to identify potential
interacting partners will be required to investigate this.
The recent finding that TorD can bind the molybde-
num cofactor, and may also interact with MobA in
addition to two sites on apoTorA [12], suggests that

TorD is intimately involved with all stages of TorA
maturation. The possibility that the GTP-binding
pocket could also represent the binding site for the
GTP-derived molybdenum cofactor should not be dis-
counted; however, in line with other biological systems
discussed above, it appears more likely that the magne-
sium-dependent GTP hydrolysis activity is regulating
the intra- and inter-molecular interactions exhibited by
the TorD protein.
ABC
D
G
EF
Fig. 8. Putative GTP binding regions of TorD family proteins. (A, B) Showing surface and cartoon representations, respectively, of the E. coli
TorD monomer model [19]. The TVRD G-4 motif, T65-D68, is displayed in yellow together with other residues implicated in nucleotide bind-
ing (Q7, F41 and F180) [19,20,24]. (C) The Sh. massilia TorD homodimer [23] is shown as a surface representation with the two intertwined
protomers coloured red and blue and the equivalent potential GTP-binding residues as highlighted in (A) and (B) (H8, W40, F182 and the G-4
loop K66-E69) displayed in yellow. (D) A cartoon representation of a single functional unit of the Sh. massilia TorD dimer. (E, F) Showing car-
toon and surface representations, respectively of S. enterica serovar Typhimurium DmsD [24], again with predicted GTP-binding residues
D8, L37, F175 and the G-4 loop T62-E66 displayed in yellow. All images were generated using
PYMOL (). (G) Multiple
sequence alignment of partial TorD sequences from different bacteria (from top to bottom: E. coli, Shigella dysenteriae, S. enterica serovar
Typhimurium, Vibrio fischeri, Vibrio cholerae, Sh. masillia, and Pastuerella multicoda) showing overall conservation of the G-4 loop tetrapep-
tide together with the adjacent C-terminal residue (boxed). Alignments were generated using
CLUSTALW2 ( />clustalw2) and
BOXSHADE ( />D. Guymer et al. The E. coli TorD homodimer has GTPase activity
FEBS Journal 277 (2010) 511–525 ª 2009 The Authors Journal compilation ª 2009 FEBS 521
Experimental procedures
Plasmids
Overproduction of TorD

his
was achieved from plasmid
pQI-TorD [19], which incorporates a C-terminal hexahisti-
dine tag onto the protein. For constitutive expression of
torD, plasmids pSU-TorD [5] and pUP-TorD [5], deriva-
tives of pSUPROM and pUNIPROM [5], respectively, were
used. Point mutations were introduced into the torD
sequence by site directed mutagenesis using the Quikchange
protocol (Stratagene, La Jolla, CA, USA).
Protein production and purification methods
The host strain for TorD
his
overproduction was E. coli C43
(DE3) [57] harbouring pREP4 (lacI
+
, Kan
R
; Roche Diag-
nostics, Basel, Switzerland). Protein production was per-
formed in LB medium supplemented with appropriate
antibiotics at 37 °C. Cells were grown with shaking until D
600
of  0.5 was reached followed by a further 3 h under
‘inducing’ conditions in the presence of 1 mm (final) isopro-
pyl b-d-thiogalactoside. Following overproduction, cells were
harvested by centrifugation, resuspended in 20 mm Tris–HCl
(pH 7.6) and 1 mm dithithreitol, and broken by two passages
through a French pressure cell at 8000 psi. Debris and unbro-
ken cells were removed by and centrifugation and the super-
natant, which contains soluble proteins and membrane

vesicles, was applied to a 5 mL of a HisTrapÔ IMAC col-
umn (GE Healthcare, Milwaukee, WI, USA) equilibrated
with 20 mm Tris–HCl (pH 7.6), 150 mm NaCl, 25 mm imid-
azole (Fluka, Buchs, Switzerland) and 1 mm dithiothreitol.
Bound proteins were eluted using a 30 mL linear gradient in
the range 20–500 mm imidazole in the same buffer. Samples
containing TorD
his
as judged by SDS–PAGE were pooled
and concentrated in a Vivaspin (Millipore, Billerica, MA,
USA) device with a 10 kDa molecular mass cut-off.
Overproduction of /MalE-TorA
his
[58] was performed in
the same manner as for TorD
his
, except the host strain was
BL21 (DE3) [59] containing pREP4, and purification was
performed as previously described [58].
Molecular exclusion chromatography was performed
using a SuperdexÔ 75 10 ⁄ 30 column (GE Healthcare)
equilibrated in 20 mm Tris–HCl, 150 mm NaCl and 1 mm
dithiothreitol.
Cibacron blue affinity chromatography was performed
using a 5 mL HiTrapÔ Blue HP column (GE Healthcare)
in a protocol adapted specifically for TorD in the present
study. The column was first equilibrated with 20 mm Tris–
HCl (pH 7.5), 500 mm NaCl, 1 mm dithiothreitol and
IMAC-purified TorD
his

was applied in this buffer. Follow-
ing re-equilibration, bound protein was eluted in a single
step from loading buffer to deionised water.
Protein concentrations of purified preparations were
determined by measuring A
280
and applying molar extinc-
tion coefficients of 30 940 MÆcm
)1
for TorD
his
and
66 350 MÆcm
)1
for FMalE-TorA
his
.
Protein analysis
The SDS–PAGE method was as described by La
¨
mmli
[60] and non-SDS-PAGE followed the same protocol but
omitted SDS at every stage. Western immunoblotting was
performed as described previously [61].
Analytical ultracentrifugation (sedimentation
velocity)
Proteins were prepared as 1.0 mgÆmL
)1
stock solutions and
diluted to give three samples at 0.25, 0.50 and 0.75

mgÆmL
)1
in 50 mm Tris–HCl (pH 7.5). The samples were
loaded into assembled cells with quartz windows and two-
sector 12 mm path length flow-through charcoal-filled epon
centrepieces. Each cell was manually fitted and aligned in
an An50-Ti rotor (Beckman Coulter, Fullerton, CA,
USA). The rotor was allowed to reach 16 °C and then
equilibrated at that temperature for 2 h before accelerating
to 150 000 g. Data were collected using the absorbance
detection system of a XL-I analytical ultracentrifuge
(Beckman Coulter) and the cells were scanned at 280 nm
every 120 s. Two hundred scans were collected for each
concentration of protein. The data were analysed using the
sedfit [62,63]. Solvent density and solvent viscosity were
calculated using the public domain software sednterp
(). The c(s) continuous distribu-
tion was generated in sedfit and then optimised using the
meniscus and f ⁄ f
0
as floating parameters. The relative con-
centration and weight-average sedimentation coefficient
were determined for each species using the integration tool
in sedfit.
GTPase actvity assays
Two GTPase assays were employed in the present study,
a discontinuous assay based on detecting released phos-
phate using malachite green, and a continuous coupled
reaction available commercially as a kit. For the discontin-
uous assay, 0.1 mm samples of purified protein were incu-

bated (typically overnight) in a reaction mixture containing
10 mm Tris–HCl (pH 7.5), 5 mm GTP and 1.2 mm MgCl
2
at 37 °C. On completion, samples were diluted 250-fold in
water and mixed with malachite green solution [0.7 m HCl,
0.3 mm malachite green oxalate, 8.3 mm Na
2
MoO
4
, 0.05%
(v ⁄ v) Triton X-100] in a 1 : 1 ratio. Following a 15 min
incubation at 22 °C, A
630
was measured and compared with
a phosphate standard curve prepared using identical
solutions.
For the continuous assay, the activity of 10 lm TorD
Blue
was determined using the EnzChek
Ò
Phosphate Assay Kit
The E. coli TorD homodimer has GTPase activity D. Guymer et al.
522 FEBS Journal 277 (2010) 511–525 ª 2009 The Authors Journal compilation ª 2009 FEBS
(Invitrogen). Reactions were started by the addition of
protein and phosphate release was measured continuously
using a Synergy 2 Plate reader (BioTek Inc., Winooski, VT,
USA) set to measure A
360
. Absorbance data were converted
into released P

i
(lmÆmin
)1
) by comparison with a standard
curve of known phosphate concentration.
All nucleotides were supplied by Sigma-Aldrich
(St Louis, MO, USA), with the exception of XTP, which
was supplied by Jena Bioscience GmbH (Jena, Germany).
GTP, ITP, GMP and XMP were stored as 50 mm stocks in
10 mm Tris–HCl (pH 8.0). ATP was buffered in 75 mm
Tris–HCl (pH 8.0). XTP was supplied as a 10 mm solution
buffered to pH 7.5.
In vivo Tat proofreading and TMAO reductase
activity assays
The signal peptide binding-dependent Tat proofreading
activity of TorD can be isolated and assayed in vivo,as
described previously [5]. Strain RJ607, which carries a chro-
mosomal torA::hybO fusion, was grown anaerobically in
LB medium supplemented with 0.5% (v ⁄ v) glycerol and
0.4% (w ⁄ v) fumarate and washed whole cells assayed for
hydrogen:benzyl viologen oxidoreductase activity as
described previously [5].
Periplasmic TMAO reducatase activity was assayed in
strain FTD100 (DtorD) [5]. The strain was grown anaerobi-
cally in LB medium supplemented with 0.5% (v ⁄ v) glycerol
and 0.4% (w ⁄ v) TMAO, harvested by centrifugation, and
separated into periplasmic fraction and sphaeroplasts using
a lysozyme ⁄ EDTA method [64]. TMAO-dependent oxida-
tion of reduced benzyl viologen was then assayed as
described previously [65].

Acknowledgements
We thank Tracy Palmer (Dundee) for useful discus-
sions and Martin Zoltner (Dundee) for help with anal-
ysing the kinetic data. Mass spectrometry was carried
out by Kenneth Beattie and Samantha Kosto of the
Fingerprints Proteomics Facility at the University of
Dundee. This work was funded in the UK by the
BBSRC through a Doctoral Training Grant awarded
to the University of Dundee, and J.M. was supported
by the Swiss National Science Foundation.
References
1 Berks BC, Palmer T & Sargent F (2003) The Tat
protein translocation pathway and its role in microbial
physiology. Adv Microb Physiol 47, 187–254.
2 Berks BC (1996) A common export pathway for pro-
teins binding complex redox cofactors? Mol Microbiol
22, 393–404.
3 DeLisa MP, Tullman D & Georgiou G (2003) Folding
quality control in the export of proteins by the bacterial
twin-arginine translocation pathway. Proc Natl Acad
Sci USA 100, 6115–6120.
4 Tullman-Ercek D, DeLisa MP, Kawarasaki Y, Iran-
pour P, Ribnicky B, Palmer T & Georgiou G (2007)
Export pathway selectivity of Escherichia coli twin argi-
nine translocation signal peptides. J Biol Chem 282,
8309–8316.
5 Jack RL, Buchanan G, Dubini A, Hatzixanthis K, Pal-
mer T & Sargent F (2004) Coordinating assembly and
export of complex bacterial proteins. EMBO J 23,
3962–3972.

6 Sargent F (2007) Constructing the wonders of the bacte-
rial world: biosynthesis of complex enzymes. Microbiol-
ogy 153, 633–651.
7 Sargent F (2007) The twin-arginine transport system:
moving folded proteins across membranes. Biochem Soc
Trans 35, 835–847.
8 Mejean V, Iobbi-Nivol C, Lepelletier M, Giordano G,
Chippaux M & Pascal MC (1994) TMAO anaerobic
respiration in Escherichia coli: involvement of the tor
operon. Mol Microbiol 11, 1169–1179.
9 Kisker C, Schindelin H & Rees DC (1997) Molybde-
num-cofactor-containing enzymes: structure and mecha-
nism. Annu Rev Biochem 66, 233–267.
10 Santini CL, Ize B, Chanal A, Muller M, Giordano G &
Wu LF (1998) A novel sec-independent periplasmic pro-
tein translocation pathway in Escherichia coli. EMBO J
17, 101–112.
11 Genest O, Ilbert M, Mejean V & Iobbi-Nivol C (2005)
TorD, an essential chaperone for TorA molybdoenzyme
maturation at high temperature. J Biol Chem 280,
15644–15648.
12 Genest O, Neumann M, Seduk F, Stocklein W, Mejean
V, Leimkuhler S & Iobbi-Nivol C (2008) Dedicated
metallochaperone connects apoenzyme and molybde-
num cofactor biosynthesis components. J Biol Chem
283, 21433–21440.
13 Genest O, Seduk F, Ilbert M, Mejean V & Iobbi-
Nivol C (2006) Signal peptide protection by specific
chaperone. Biochem Biophys Res Commun 339,
991–995.

14 Genest O, Seduk F, Theraulaz L, Mejean V &
Iobbi-Nivol C (2006) Chaperone protection of immature
molybdoenzyme during molybdenum cofactor limitation.
FEMS Microbiol Lett 265, 51–55.
15 Ilbert M, Mejean V, Giudici-Orticoni MT, Samama JP
& Iobbi-Nivol C (2003) Involvement of a mate chaper-
one (TorD) in the maturation pathway of molybdoen-
zyme TorA. J Biol Chem 278, 28787–28792.
16 Ilbert M, Mejean V & Iobbi-Nivol C (2004) Functional
and structural analysis of members of the TorD family,
a large chaperone family dedicated to molybdoproteins.
Microbiology 150, 935–943.
D. Guymer et al. The E. coli TorD homodimer has GTPase activity
FEBS Journal 277 (2010) 511–525 ª 2009 The Authors Journal compilation ª 2009 FEBS 523
17 Pommier J, Mejean V, Giordano G & Iobbi-Nivol C
(1998) TorD, a cytoplasmic chaperone that interacts
with the unfolded trimethylamine N-oxide reductase
enzyme (TorA) in Escherichia coli. J Biol Chem 273,
16615–16620.
18 Genest O, Mejean V & Iobbi-Nivol C (2009) Multiple
roles of TorD-like chaperones in the biogenesis of
molybdoenzymes. FEMS Microbiol Lett 297, 1–9.
19 Buchanan G, Maillard J, Nabuurs SB, Richardson DJ,
Palmer T & Sargent F (2008) Features of a twin-
arginine signal peptide required for recognition by
a Tat proofreading chaperone. FEBS Lett 582,
3979–3984.
20 Hatzixanthis K, Clarke TA, Oubrie A, Richardson DJ,
Turner RJ & Sargent F (2005) Signal peptide-chaperone
interactions on the twin-arginine protein transport path-

way. Proc Natl Acad Sci USA 102, 8460–8465.
21 Turner RJ, Papish AL & Sargent F (2004) Sequence
analysis of bacterial redox enzyme maturation proteins
(REMPs). Can J Microbiol 50, 225–238.
22 Kirillova O, Chruszcz M, Shumilin IA, Skarina T,
Gorodichtchenskaia E, Cymborowski M, Savchenko A,
Edwards A & Minor W (2007) An extremely SAD case:
structure of a putative redox-enzyme maturation protein
from Archaeoglobus fulgidus at 3.4 A
˚
resolution. Acta
Crystallogr D Biol Crystallogr 63 , 348–354.
23 Tranier S, Iobbi-Nivol C, Birck C, Ilbert M, Mortier-
Barriere I, Mejean V & Samama JP (2003) A novel pro-
tein fold and extreme domain swapping in the dimeric
TorD chaperone from Shewanella massilia. Structure 11,
165–174.
24 Qiu Y, Zhang R, Binkowski TA, Tereshko V,
Joachimiak A & Kossiakoff A (2008) The 1.38 A
˚
crystal
structure of DmsD protein from Salmonella typhimurium,
a proofreading chaperone on the Tat pathway. Proteins
71, 525–533.
25 Stevens CM, Winstone TM, Turner RJ & Paetzel M
(2009) Structural analysis of a monomeric form of
the twin-arginine leader peptide binding chaperone
Escherichia coli DmsD. J Mol Biol 389, 124–133.
26 Sarfo KJ, Winstone TL, Papish AL, Howell JM, Kadir
H, Vogel HJ & Turner RJ (2004) Folding forms of

Escherichia coli DmsD, a twin-arginine leader binding
protein. Biochem Biophys Res Commun 315, 397–403.
27 Tranier S, Mortier-Barriere I, Ilbert M, Birck C,
Iobbi-Nivol C, Mejean V & Samama JP (2002) Charac-
terization and multiple molecular forms of TorD from
Shewanella massilia, the putative chaperone of the mo-
lybdoenzyme TorA. Protein Sci 11, 2148–2157.
28 Hatzixanthis K, Richardson DJ & Sargent F (2005)
Chaperones involved in assembly and export of N-oxide
reductases. Biochem Soc Trans 33, 124–126.
29 Zhang B, Zhang Y, Wang Z & Zheng Y (2000) The
role of Mg
2+
cofactor in the guanine nucleotide
exchange and GTP hydrolysis reactions of Rho family
GTP-binding proteins. J Biol Chem 275, 25299–25307.
30 Pai EF, Krengel U, Petsko GA, Goody RS, Kabsch W
& Wittinghofer A (1990) Refined crystal structure of
the triphosphate conformation of H-ras p21 at 1.35 A
˚
resolution: implications for the mechanism of GTP
hydrolysis. EMBO J 9, 2351–2359.
31 Sprang SR (1997) G protein mechanisms: insights from
structural analysis. Annu Rev Biochem 66, 639–678.
32 Palmer T, Vasishta A, Whitty PW & Boxer DH (1994)
Isolation of protein FA, a product of the mob locus
required for molybdenum cofactor biosynthesis in
Escherichia coli. Eur J Biochem 222, 687–692.
33 Webb MR & Hunter JL (1992) Interaction of GTPase-
activating protein with p21ras, measured using a contin-

uous assay for inorganic phosphate release. Biochem J
287, 555–559.
34 Dever TE, Glynias MJ & Merrick WC (1987)
GTP-binding domain: three consensus sequence
elements with distinct spacing. Proc Natl Acad Sci USA
84, 1814–1818.
35 Bourne HR, Sanders DA & McCormick F (1991) The
GTPase superfamily: conserved structure and molecular
mechanism. Nature 349, 117–127.
36 Walker JE, Saraste M, Runswick MJ & Gay NJ (1982)
Distantly related sequences in the alpha- and beta-
subunits of ATP synthase, myosin, kinases and other
ATP-requiring enzymes and a common nucleotide
binding fold. EMBO J 1, 945–951.
37 Hwang YW & Miller DL (1987) A mutation that alters
the nucleotide specificity of elongation factor Tu, a GTP
regulatory protein. J Biol Chem 262, 13081–13085.
38 la Cour TF, Nyborg J, Thirup S & Clark BF (1985)
Structural details of the binding of guanosine diphos-
phate to elongation factor Tu from E. coli as studied by
X-ray crystallography. EMBO J 4, 2385–2388.
39 Praefcke GJ, Geyer M, Schwemmle M, Robert
Kalbitzer H & Herrmann C (1999) Nucleotide-binding
characteristics of human guanylate-binding protein 1
(hGBP1) and identification of the third GTP-binding
motif. J Mol Biol 292, 321–332.
40 Weijland A, Sarfati R, Barzu O & Parmeggiani A
(1993) Asparagine-135 of elongation factor Tu is a
crucial residue for the folding of the guanine nucleotide
binding pocket. FEBS Lett 330, 334–338.

41 Cavallius J & Merrick WC (1998) Site-directed
mutagenesis of yeast eEF1A. Viable mutants with
altered nucleotide specificity. J Biol Chem 273, 28752–
28758.
42 Carr-Schmid A, Durko N, Cavallius J, Merrick WC &
Kinzy TG (1999) Mutations in a GTP-binding motif of
eukaryotic elongation factor 1A reduce both transla-
tional fidelity and the requirement for nucleotide
exchange. J Biol Chem 274, 30297–30302.
The E. coli TorD homodimer has GTPase activity D. Guymer et al.
524 FEBS Journal 277 (2010) 511–525 ª 2009 The Authors Journal compilation ª 2009 FEBS
43 Hill AV (1910) The possible effects of the aggregation
of the molecules of haemoglobin on its dissociation
curves. J Physiol 40, iv–vii.
44 Weiss JN (1997) The Hill equation revisited: uses and
misuses. FASEB J 11, 835–841.
45 Saint-Jean AP, Phillips KR, Creighton DJ & Stone MJ
(1998) Active monomeric and dimeric forms of Pseudo-
monas putida glyoxalase I: evidence for 3D domain
swapping. Biochemistry 37, 10345–10353.
46 Dumas P, Bergdoll M, Cagnon C & Masson JM (1994)
Crystal structure and site-directed mutagenesis of a
bleomycin resistance protein and their significance for
drug sequestering. EMBO J 13, 2483–2492.
47 Hadden JM, Convery MA, Declais AC, Lilley DM &
Phillips SE (2001) Crystal structure of the Holliday
junction resolving enzyme T7 endonuclease I. Nat
Struct Biol 8 , 62–67.
48 Newcomer ME (2002) Protein folding and three-dimen-
sional domain swapping: a strained relationship? Curr

Opin Struct Biol 12, 48–53.
49 Bennett MJ, Schlunegger MP & Eisenberg D (1995) 3D
domain swapping: a mechanism for oligomer assembly.
Protein Sci 4 , 2455–2468.
50 Green SM, Gittis AG, Meeker AK & Lattman EE
(1995) One-step evolution of a dimer from a monomeric
protein. Nat Struct Biol 2, 746–751.
51 Schweins T, Scheffzek K, Assheuer R & Wittinghofer A
(1997) The role of the metal ion in the p21ras catalysed
GTP-hydrolysis: Mn
2+
versus Mg
2+
. J Mol Biol 266,
847–856.
52 Bernstein HD, Poritz MA, Strub K, Hoben PJ,
Brenner S & Walter P (1989) Model for signal
sequence recognition from amino-acid sequence of 54K
subunit of signal recognition particle. Nature 340,
482–486.
53 Jewett MC, Miller ML, Chen Y & Swartz JR (2009)
Continued protein synthesis at low [ATP] and [GTP]
enables cell adaptation during energy limitation.
J Bacteriol 191, 1083–1091.
54 Bourne HR (1997) G proteins. The arginine finger
strikes again. Nature 389, 673–674.
55 Focia PJ, Shepotinovskaya IV, Seidler JA & Freymann
DM (2004) Heterodimeric GTPase core of the SRP tar-
geting complex. Science 303, 373–377.
56 Doudna JA & Batey RT (2004) Structural insights into

the signal recognition particle. Annu Rev Biochem 73,
539–557.
57 Miroux B & Walker JE (1996) Over-production of pro-
teins in Escherichia coli: mutant hosts that allow synthe-
sis of some membrane proteins and globular proteins at
high levels. J Mol Biol 260, 289–298.
58 Maillard J, Spronk CA, Buchanan G, Lyall V,
Richardson DJ, Palmer T, Vuister GW & Sargent F
(2007) Structural diversity in twin-arginine signal pep-
tide-binding proteins. Proc Natl Acad Sci USA 104,
15641–15646.
59 Studier FW & Moffatt BA (1986) Use of bacteriophage
T7 RNA polymerase to direct selective high-level
expression of cloned genes. J Mol Biol 189, 113–130.
60 Laemmli UK (1970) Cleavage of structural proteins
during the assembly of the head of bacteriophage T4.
Nature 227, 680–685.
61 Towbin H, Staehelin T & Gordon J (1979) Electropho-
retic transfer of proteins from polyacrylamide gels to
nitrocellulose sheets: procedure and some applications.
Proc Natl Acad Sci USA 76, 4350–4354.
62 Dam J & Schuck P (2004) Calculating sedimentation
coefficient distributions by direct modeling of sedimen-
tation velocity concentration profiles. Methods Enzymol
384, 185–212.
63 Lebowitz J, Lewis MS & Schuck P (2002) Modern ana-
lytical ultracentrifugation in protein science: a tutorial
review. Protein Sci 11, 2067–2079.
64 Hatzixanthis K, Palmer T & Sargent F (2003) A subset
of bacterial inner membrane proteins integrated by the

twin-arginine translocase. Mol Microbiol 49, 1377–1390.
65 Silvestro A, Pommier J & Giordano G (1988) The
inducible trimethylamine-N-oxide reductase of Escheri-
chia coli K12: biochemical and immunological studies.
Biochim Biophys Acta 954, 1–13.
Supporting information
The following supplementary material is available:
Fig. S1. The product of GTP hydrolysis by TorD is
GDP.
Fig. S2. Analysis of TorD
Blue
by MS.
Fig. S3. GTPase activity of TorD variants.
This supplementary material can be found in the
online version of this article.
Please note: As a service to our authors and readers,
this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and
may be re-organized for online delivery, but are not
copy-edited or typeset. Technical support issues arising
from supporting information (other than missing files)
should be addressed to the authors.
D. Guymer et al. The E. coli TorD homodimer has GTPase activity
FEBS Journal 277 (2010) 511–525 ª 2009 The Authors Journal compilation ª 2009 FEBS 525