Báo cáo khoa học: Redox-sensitive loops D and E regulate NADP(H) binding in domain III and domain I–domain III interactions in proton-translocating Escherichia coli transhydrogenase potx

Redox-sensitive loops D and E regulate NADP(H) binding in domain III
and domain I–domain III interactions in proton-translocating
Escherichia coli
transhydrogenase
Carina Johansson
1
, Anders Pedersen
1
,B.Go¨ ran Karlsson
2
and Jan Rydstro¨m
1
1
Department of Biochemistry, Go
¨
teborg University, Sweden;
2
Department of Molecular Biotechnology, Chalmers University of
Technology, Go
¨
teborg, Sweden
Membrane-bound transhydrogenases are conformationally
driven proton-pumps which couple an inward proton
translocation to the reversible reduction of NADP
+
by
NADH (forward reaction). This reaction is stimulated by an
electrochemical proton gradient, Dp, presumably through an
increased release of NADPH. The enzymes have three
domains: domain II spans the membrane, while domain I
and III are hydrophilic and contain the binding sites for

NAD(H) and NADP(H), respectively. Separately expressed
domain I and III together catalyze a very slow forward
reaction due to tightly bound NADP(H) in domain III.
With the aim of examining the mechanistic role(s) of loop
D and E in domain III and intact cysteine-free Escherichia
coli transhydrogenase by cysteine mutagenesis, the con-
served residues bA398, bS404, bI406, bG408, bM409 and
bV411 in loop D, and residue bY431 in loop E were selected.
In addition, the previously made mutants bD392C and
bT393C in loop D, and bG430C and bA432C in loop E,
were included. All loop D and E mutants, especially bI406C
and bG430C, showed increased ratios between the rates of
the forward and reverse reactions, thus approaching that of
the wild-type enzyme. Determination of k
NADPH
d
values
indicated that the former increase was due to a strongly
increased dissociation of NADPH caused by an altered
conformation of loops D and E. In contrast, the cysteine-free
G430C mutant of the intact enzyme showed the same inhi-
bition of both forward and reverse rates. Most domain III
mutants also showed a decreased affinity for domain I. The
results support an important and regulatory role of loops D
and E in the binding of NADP(H) as well as in the inter-
action between domain I and domain III.
Keywords: transhydrogenase; NADP; proton pump; mem-
brane protein.
Transhydrogenase is a membrane protein, which is found in
the inner membrane of mitochondria and in the cytoplasmic

membrane of bacteria. It couples the reduction of NADP
+
by NADH to the electrochemical proton gradient (Dp)
according to the reaction
NADH þ NADP
þ
þ H
þ
ðoutÞ
! NAD
þ
þ NADPH þ H
þ
ðinÞ
ð1Þ
‘Out’ and ‘in’ denote the cytosol and matrix, respectively,
in mitochondria and periplasmic space and cytosol, respect-
ively, in bacteria. Key features of this reaction is that?p
stimulates the rate of reduction of NADP
+
by NADH
some 10-fold and causes a shift in the apparent equilibrium
constant from 1 to approximately 500 [1].
Transhydrogenase from Escherichia coli is composed of
an a subunit of about 54 kDa and a b subunit of about
48 kDa. The active form of the enzyme is a
2
b
2
.Likeall

other membrane-bound transhydrogenases, the E. coli
enzyme is composed of three domains. Domain I (dI)
and domain III (dIII) are hydrophilic and contain the
binding sites for NAD(H) and NADP(H), respectively,
whereas domain II (dII) spans the membrane. The genes
encoding the hydrophilic and nucleotide-binding domains
of transhydrogenase from several species have been
overexpressed and the proteins have been purified and
characterized. In all cases, dI is purified as a dimer and
lack bound substrates. Separately expressed dIII exists as a
monomer and contains tightly bound NADP(H), reflecting
a dramatically increased affinity for NADP(H) as com-
pared to the intact enzyme. Even in the absence of the
membrane bound dII, dI and dIII from the same or
different species form a catalytically active complex
capable of catalyzing the various transhydrogenation
reactions. However, the tight binding of NADP(H), results
in low reactions rates for the reverse and forward reactions
catalyzed by the dI + dIII complex, as these reactions are
limited by a slow release of NADP
+
and NADPH,
respectively (reviewed in [2–4]).
The 3D structures of dI from Rhodospirillum rubrum [5]
and dIII from bovine [6], human [7], E. coli [8–10] and
R. rubrum [11,12] transhydrogenases have been studied
by both X-ray crystallography [5–7] and NMR [8–12]. The
global fold of dIII is a six-stranded parallel b sheet
Correspondence to J. Rydstro
¨

m, Department of Biochemistry,
Go
¨
teborg University, Box462, 405 30 Go
¨
teborg, Sweden.
Fax: + 46 31 7733910, Tel.: + 46 31 7733921,
E-mail:
Abbreviations: dI, transhydrogenase domain I; dIII, transhydrogenase
domain III; ecI, E. coli dI; ecIII, E. coli dIII; rrI, R. rubrum dI;
rrIII, R. rubrum dIII; cfTH, cysteine-free transhydrogenase;
AcPyAD
+
, oxidized 3-acetylpyridine-NAD
+
MIANS, 2-(4¢-malei-
midylanilino)naphthalene-6-sulfonic acid (sodium salt).
(Received 30 May 2002, revised 12 July 2002,
accepted 26 July 2002)
Eur. J. Biochem. 269, 4505–4515 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03144.x
surrounded by helices and irregular loops. All dIII
structures were solved using dIII with NADP
+
bound in
a nonclassical binding mode of the substrate, i.e. as
compared to other NADP(H)-dependent enzymes,
NADP
+
in dIII is bound in a reversed orientation. The
3D structure of dIII with bound NADPH is still unknown,

even though NMR experiments in which NADPH was
added provided some information regarding regions that
were affected by a change in the redox-state of the substrate
[9–11]. The crystal structure of dI from the R. rubrum
transhydrogenase [5] showed that it was a dimer, with the
monomer comprised of the two subdomains 1 A and 1B [5].
The 1B subdomain has a Rossman fold responsible for
binding NAD(H) in a novel mode as compared to other
NAD(H)-binding enzymes [5].
By using NMR in combination with mutagenesis, an
extensive characterization of the dynamic interface between
E. coli transhydrogenase dI (ecI) and dIII (ecIII) was
recently carried out [10]. In addition to information
regarding important residues involved in the ecI–ecIII
interface, the results revealed unexpected redox-dependent
changes of the ecI–ecIII interface suggested to be relevant
for the overall reaction mechanism of the intact enzyme.
The regions at the C-terminal end of the bsheet comprised
of residues bG389-bI406 (part of loop D, linking b-strands 4
and 5) and bG430–bV434 (part of loop E, linking b strands
5 and 6) were identified as redox-sensitive regions, regulated
by the redox-state of NADP(H), that strongly influenced
the ecI–ecIII interface [10]. A structure of the dI
2
-dIII
complex from R. rubrum has recently been resolved by
X-ray crystallography [13], which revealed dI–dIII inter-
faces at an atomic level similar to those derived from the
NMR studies [10]. Based on the crystal structures of
dIII from the human heart [4,7] and the dI

2
-dIII com-
plex from R. rubrum [13], loops D and E have previously
been suggested to be important in NADP(H)-binding
and possibly also coupling to proton translocation
[4,13]. The corresponding regions in ecIII are indicated in
an NMR-derived structural model (Fig. 1) and in the
amino-acid sequence with secondary structure elements
(Fig. 2).
The functional importance of loops D and E have so far
not been functionally established. In the present work, the
roles of these regions were investigated in greater detail
using site-directed mutagenesis. The results suggest that
loop D is important in communicating affinity changes in
the NADP(H)-binding site to domain I, and that both loop
D and E regulate the release of NADP(H). In support of the
previous suggestion [4,13] both loops are suggested to play a
key role in the regulation of the enzyme by an electrochem-
ical proton gradient.
MATERIALS AND METHODS
Site-directed mutagenesis
The Quikchange mutagenesis kit (Stratagene) was used to
introduce single cysteine mutations in isolated ecIII and
cfTH. A modified pET8c plasmid used for the prepar-
ation of histidine-tagged ecIII [14] served as a DNA
template in the construction of the seven cysteine
mutants ecIIIA398C, ecIIIS404C, ecIIII406C, ecIIIG408C,
ecIIIM409C, ecIIIV411C and ecIIIY431C. In addition,
four previously produced mutants, i.e. ecIIID392C [15],
and ecIIIT393C, ecIIIG430C and ecIIIA432C [9], were

further characterized. The cfbG430C mutant was based
on the pCLNH plasmid to which an N-terminal histi-
dine tag has been added to the cfTH gene [14]. The
correctness of the mutant products was checked by DNA
sequencing.
Fig. 1. Two views of the partial 3-dimensional
structure of ecIII. The NADP(H)-binding site
of ecIII. This illustration was based on the
NMR structure of E. coli dIII [10]. The
beginning and end of loop D (excluding a5)
and E are P403-V411 and M427-G433,
respectively. The illustration was prepared
using the software
MOLMOL
[26].
Fig. 2. The amino-acid sequence of ecIII and secondary structure ele-
ments based on NMR data. Secondary elements were based on results
from NMR experiments [10]. Black shaded residues are conserved, and
green-shaded residues are similar among different species of transhy-
drogenase.
4506 C. Johansson et al.(Eur. J. Biochem. 269) Ó FEBS 2002
Expression and purification
The gene encoding ecIII was based on the 177 C-terminal
residues of E. coli transhydrogenase b-subunit (bQ286-
bL462). EcIII mutants [14] and E. coli cysteine-free tran-
shydrogenase enzymes [15] were expressed and purified as
described. The plasmid pCD1 encoding R. rubrum domain I
(rrI) was expressed in E. coli TG1 cells [16] and purified
according to the method described by Bizouarn et al. [17]
with modifications. After sonication and centrifugation of

1 L culture, the supernatant was loaded onto a 15-mL
Q-Sepharose HP column (Pharmacia) equilibrated with
20 m
M
Tris/HCl and 10 m
M
(NH
4
)
2
SO
4
(pH 8.0). Protein
was eluted with about 60 mL of the same buffer, after which
(NH
4
)
2
SO
4
was added to a final concentration of 1.6
M
.
After 10 h of incubation the sample was centrifuged for 1 h
at 18 000 r.p.m. in a Beckman JA20 rotor, and the
supernatant loaded onto a 15-mL Butyl Toyopearl column
(Tosohas). Protein was eluted with a gradient (300 mL) of
1.6 to 0
M
(NH

4
)
2
SO
4
in 20 m
M
Tris/HCl (pH 8.0), dialyzed
against 20 m
M
Tris/HCl, 10 m
M
(NH
4
)
2
SO
4
pH 8.0 and
stored at )20 °C in 20% glycerol.
All domains displayed a purity greater than 90% as
judged by SDS-polyacrylamide gel electrophoresis using
8–25% gradient gels in the Phast system (Pharmacia) for
transhydrogenase domains and 10–20% gradient gels
(Novex) for the cfTH enzymes.
Determination of protein concentration and substrate
content
Protein concentrations were determined using the bicinch-
oninic acid assay with bovine serum albumin as standard
[18]. The content of bound NADPH in the ecIII mutants

was determined by absorbance spectroscopy at 339 nm,
using e
NADPH
¼ 6100
M
)1
Æcm
)1
, whereas the content of
NADP
+
was determined by fluorescence using a modified
Klingenberg procedure as described previously [14].
Activity assays
Unless stated otherwise, transhydrogenation reactions cata-
lyzed by mutant and wild-type ecIII were assayed as
described [19] in buffer A [20 m
M
each of Mes, Mops, Ches,
and Tris, 50 m
M
NaCl (pH 7.0)], using rrI. Protein-protein
titrations were performed in which the ecIII concentration
was kept constant and the rrI concentration varied until a
maximal rate was reached.
The forward and reverse reactions catalyzed by the cfTH
and cfTH mutant enzymes were measured as described [20]
in buffer B [20 m
M
each of Mes, Ches, Tris and Hepes,

50 m
M
NaCl and 0.01% Brij (pH 7.0)] whereas the cyclic
reaction [19,20] was normally assayed in buffer C [20 m
M
each of Mes, Ches, Tris and Hepes, 50 m
M
NaCl, 2 m
M
MgCl
2
,1m
M
EDTA and 0.01% Brij (pH 6.0)]. The reverse
and cyclic reactions were followed optically at 375 nm as
reduction of AcPyAD
+
as described [19]. The forward
reaction was measured spectroscopically at 398 nm as
reduction of thio-NADP
+
[21].
For comparative reasons, the transhydrogenation reac-
tions catalyzed by rrI + wild-type and mutant ecIII
mixtures were also assayed in the cfTH assay buffers. All
measurements were performed at 25 °C.
Fluorescence measurements
Fluorescence measurements were carried out on a SPEX
Model FL1T1 s2 and Shimadzu RF5001PC spectrofluo-
rometers at 25 °C. A cuvette with a 10 · 10 mm cross

section was used and the excitation and emission slits were
both 2.5 nm.
Determination of the NADPH release rate from ecIII
by fluorescence
The release rate of NADPH from ecIII was determined
from the exponential decrease in fluorescence as bound
NADPH was released from ecIII and oxidized by glutathi-
one and glutathione reductase using excitation and emission
wavelengths of 340 and 460 nm, respectively. Oxidized
glutathione (2 m
M
) was added to the cuvette containing
1.5–2 l
M
of mutant ecIII enzyme and 1–4 U of glutathione
reductase; the fluorescence was monitored for up to 20 min.
In the case of the ecIIIV431C mutant, it was preincubated
with an equimolar concentration (about 1 l
M
)ofNADPH
for 5 min prior to the assay. The measurements were carried
out in a buffer composed of 20 m
M
Mops and 5 m
M
MgCl
2
(pH 7.0).
RESULTS
Characterization of single-cysteine mutations in ecIII

In order to elucidate the mechanistic roles of loop D and
E in greater detail, the single ecIII cysteine mutants,
ecIIIA398C, ecIIIS404C, ecIIII406C, ecIIIG408C,
ecIIIM409C, ecIIIV411C (loop D), and ecIIIY431C (loop
E), were expressed in the cysteine-free background and
purified as described in Materials and methods. In addition,
the previously made mutants ecIIID392C and ecIIIT393C
in loop D, and ecIIIG430C and ecIIIA432C in loop E, were
included. All mutants were characterized with respect to
substrate binding, catalytic activities of the different trans-
hydrogenation reactions and affinity for rrI. In these assays
and under the conditions used dimer formation by all
mutants amounted to less than 10% as tested by SDS/
PAGE (data not shown).
Figure 3 A and B show the detailed positions of some of
the mutated as well as other important residues in ecIII, and
bound NADP
+
, in ecIII, viewed from two different angles.
Note that loops D and E are much more defined in the
crystal structure of bovine dIII shown in Fig. 3 than in
Fig. 1. A recently produced high resolution NMR structure
of ecIII with bound NADP
+
is essentially identical to that
of bovine dIII with bound NADP
+
(C. Johansson, J. T.
Johansson, A. Pedersen, J. Rydstro
¨

m and B. G. Karlsson,
unpublished results).
Content of bound NADP(H) in dIII
The increased affinity of dIII for NADP(H) is reflected in
the content of tightly bound NADP(H) in almost 100% of
the molecules of separately expressed ecIII [10,19]. The
presence of NADP(H) changes the absorbance maximum,
k
max
, from 278 nm for the apo-protein to 268 nm for the
NADP(H)-containing ecIII. Consequently, k
max
is an
indication of the fraction of apo-protein. An additional
Ó FEBS 2002 Redox-sensitive loops D and E in transhydrogenase (Eur. J. Biochem. 269) 4507
typical property of isolated ecIII is the percentage of bound
NADP(H), which is highly reproducible for different
preparations. In order to assess the effects on substrate-
binding site occupancy, the proportions of bound NADP
+
and NADPH were determined for the mutants generated in
this investigation i.e. ecIIIA398C, ecIIIS404C, ecIIII406C,
ecIIIG408C, ecIIIM409C, ecIIIV411C and ecIIIY431C
(Table 1).
Fig. 3. NADP(H) binding region of dIII viewed
perpendicular to (A) and parallell with the
b sheet (B). The structure was modelled using
MOLMOL
[26] based on the crystal structure of
the bovine dIII with bound NADP

+
(PDB
entry 1D40). Important residues and loops D
and E, are indicated. Numbering of mutated
residues is according to ecIII.
Table 1. Content of bound NADP(H) in wild-type and mutant ecIII. The concentrations of NADPH in the ecIII enzymes were calculated from
UV-Vis spectra and the content of NADP
+
was determined by a modified Klingenberg procedure (see Materials and methods). The k
max
value
corresponds to the wavelength at which maximal absorption was observed for the interval 200–400 nm in UV-Vis spectra.
Enzyme Loop affected k
max
NADP
+
(%) NADPH (%) Apo-form (%)
ecIII 267 87 5 8
ecIIIA398C D 267 79 16 5
ecIIIS404C D 267 66 33 1
ecIIII406C D 267 72 19 9
ecIIIG408C D 267 65 24 11
ecIIIM409C D 267 70 27 3
ecIIIV411C D 267 40 6 54
ecIIIY431C E 267 43 28 29
4508 C. Johansson et al.(Eur. J. Biochem. 269) Ó FEBS 2002
All ecIII mutants contained bound substrate, where
the percentages of NADPH and NADP
+
varied, often

with an increased content of NADPH. This effect was
most pronounced in the case of the ecIIIS404C,
ecIIIM409C and ecIIIY431C mutants, with 33%, 27%
and 28% bound NADPH, respectively, and a relatively
constant amount of apoprotein. EcIIIV411C and
ecIIIY431C did not follow the same pattern, but
contained approximately 54% and 29% apoprotein,
respectively (Table 1). Thus, it is clear that the mutations
in the redox-sensitive loops D and E strongly affected the
binding site.
In all mutants in loop D, except ecIIIV411C, the
nucleotide content was at least 89%, with a 3–5 fold
increase in the percentage of bound NADPH. (Table 1).
Clearly, the small fraction of apo-protein in these mutants
reflected the fact that this region of the protein is not directly
involved in substrate binding.
Forward reaction catalyzed by rrI + ecIII mutant mixtures
The forward reaction (reduction of thio-NADP
+
by
NADH) catalyzed by rrI and ecIII mutant mixtures was
examined by protein-protein titrations in which the ecIII
concentration was kept constant and the rrI concentration
varied until a maximal rate was reached. In this and other
dI-dependent assays in this investigation, rrI rather than ecI
was used due to the generally higher activities obtained with
this dI preparation (cf. Fig. 2). The [rrI]/[ecIII] ratio at half-
saturation is a measure of the affinity of the complex formed
for rrI and ecIII [10] and may be used to gain information
about the role of a particular amino-acid residue in the

interactions with dI. However, because the rrI concentration
required for half maximal rate is dependent on both the K
d
for the rrI + ecIII complex and the release rate of thio-
NADPH from domain III, it can not be considered as a true
measure of the affinity between the two domains. Due to the
tight binding of nucleotides to wild-type ecIII, the forward
reaction catalyzed by ecIII with saturating concentrations of
rrI is limited at pH 7.0 by the slow release of thio-NADPH
from domain III [16]. An increase in the forward reaction
rate may thus be regarded as an increase in the release rate
of NADPH.
Table 2 summarizes the results obtained for measure-
ments of the forward reaction catalyzed by rrI and ecIII
mutants. The maximal rates for the ecIIIG430C and
ecIIIY431C mutants were 275 and 100 times that of the
wild-type ecIII, respectively. This pronounced increase
demonstrates the dramatic effect on the rate of dissociation
of NADPH caused by mutations in loop E. Depending on
the position in which the cysteine residue was introduced,
mutations in loop D had different consequences on the
forward reaction rate. The ecIIII406C mutant showed a
35-fold higher rate than wild-type ecIII, whereas the
ecIIIS404C mutation had a minor effect (Table 2).
Likewise, the ecIIIA398C, ecIIIG408C, ecIIIM409C,
ecIIIV411C and ecIIIA432 mutants showed relatively minor
changes. As the ecIIII406 residue does not participate
directly in substrate binding [6,7], the increased forward
reaction rate demonstrated by the ecIIII406C mutant was
possibly a result of perturbation of the surroundings of this

position caused by the mutation. An obvious candidate
responsible for this effect is the conserved bD392 residue,
which is located within 6 A
˚
from bI406 in the NADP
+
-
complexed dIII crystal structure [6,7].
The [rrI]/[ecIII] at half-saturation was increased for all
cysteine mutants, but appeared to be correlated to the
maximal rate displayed by the mutants (Table 2).
Reverse reaction catalyzed by rrI + ecIII mutant mixtures
Analyses of the reverse reaction catalyzed by rrI + ecIII
mutant mixtures were performed by protein-protein titra-
tions in the same way as for the forward reaction. Like the
forward reaction, the reverse reaction is limited at pH 7.0 by
the slow release of the product bound to dIII [16], but is
several-fold faster. The maximal rate of the reverse reaction
is thus an excellent tool for examining if a mutation has
altered the rate of dissociation of NADP
+
. The [rrI]/[ecIII]
necessary for half V
max
is an indication of the dissociation
constant for the rrI + ecIII complex but, like the forward
reaction, this ratio is also dependent on how fast the product
is released from ecIII. An elevated release rate needs more dI
to saturate the reaction.
As shown in Table 3 the maximal rates obtained for both

the ecIIII406C and ecIIIY431C mutants were 4.5-fold
higher than that of wild-type ecIII, indicating an increased
dissociation of NADP
+
. Again, the result obtained with the
ecIIII406C mutant was probably an indirect effect caused
by perturbations in its environment. The ecIIIS404C,
ecIIIG408C, ecIIIM409C and ecIIIV411C mutations did
not affect the reverse reaction rate significantly, which is
consistent with the fact that these positions are not in the
substrate binding-site [6,7]. The [rrI]/[ecIII] ratios at half-
saturation of the reverse reaction correlated well with the
maximal rate of the ecIII mutants.
Cyclic reaction catalyzed by rrI + ecIII mutant mixtures
The cyclic reaction, i.e. the reduction of AcPyAD
+
by
NADH via NADP(H) bound to dIII, was analyzed by
protein-protein titrations in which the ecIII concentration
Table 2. Properties of the forward reaction catalyzed by wild-type and
mutant ecIII in the presence of rrI. The values are estimations from
protein-protein titration curves in which the ecIII concentration was
fixed and the rrI concentration varied (not shown). The assays were
performed as described in Materials and methods. The [ecIII] values
refer to the fixed enzyme concentration used in the titrations. The [rrI]
values correspond to the concentration of rrI at half V
max
.
Enzyme
[ecIII]

(n
M
)
[rrI]
(n
M
)
[rrI]/
[ecIII]
V
max
(mol thio-NADPH)Æ
(mol ecIII)
)1
Æmin
)1
%
ecIII 5000 5 0.001 0.04 100
ecIIIA398C 2500 3 0.001 0.12 300
ecIIIS404C 5000 4 0.0008 0.08 200
ecIIII406C 2500 20 0.08 1.4 3500
ecIIIG408C 2500 5 0.002 0.13 325
ecIIIM409C 5000 15 0.003 0.13 325
ecIIIV411C 5000 10 0.002 0.22 550
ecIIIG430C 2500 400 0.16 11 27500
ecIIIY431C 2500 140 0.06 4 10000
ecIIIA432C 2500 10 0.004 0.13 325
Ó FEBS 2002 Redox-sensitive loops D and E in transhydrogenase (Eur. J. Biochem. 269) 4509
was kept constant and the rrI concentration varied until the
reaction rate reached a maximum. NADP(H) remains

bound to dIII during the entire catalytic cycle and the rate of
the cyclic reaction at pH 7.0 has been shown to be limited
by the hydride transfer steps [16]. Measurements of the
cyclic reaction thus offers an opportunity to examine the
affinity of ecIII mutants for dI. The [rrI][/ecIII] ratio at half-
saturation is only dependent on the dissociation constant for
the rrI + ecIII complex. Using the actual rrI and ecIII
concentrations at half V
max
the K
d
for the complex can be
calculated according to K
d
¼ [rrI]-[ecIII]/2.
In Table 4 the data from the protein-protein titrations for
the ecIII mutants are listed. Except for the ecIIIS404C
mutant, the rrI concentration required for half V
max
was
considerably higher in all of the mutants made in the
redox-sensitive loops. The most affected mutants were
the ecIIIA398C (loop D), ecIIIM409C (loop D) and
ecIIIY431C (loop E) mutants which exhibited 6–15 times
higher K
d
for the rrI + ecIII complex as compared to wild-
type ecIII (Table 4). In agreement with the crystal structure
of the dI
2

–dIII complex [13], these results show that both
loops D and E directly or indirectly make crucial contacts
with dI. Mutations in the I406–V411 region affected the
hydride transfer efficiency and only 24–31% of the
maximal rate of the cyclic reaction could be reached,
despite saturating concentrations of rrI. Interestingly, the
ecIIIS404C and ecIIIA398C mutants were still able to
catalyze the hydride transfer at a wild-type rate, even though
the affinity for rrI had been substantially lowered.
Release rate of NADPH measured by fluorescence
The fact that NADPH, but not NADP
+
, fluoresces at
460 nm when using excitation at 340 nm, was utilized in
order to determine the rate of release of NADPH from
ecIII. By this method bound NADPH was oxidized by
glutathione reductase and glutathione. The reaction is
limited by the rate of release of NADPH and the resulting
decrease in fluorescence could consequently be used to
calculate the K
offNADPH
[22]. Fig. 4 shows the oxidation by
glutathione and glutathione reductase of NADPH bound to
ecIIIG432C and ecIIIG430C. The rate obtained with
ecIIIA432C is representative of the wild-type rate. In
contrast, the ecIIIG430C mutant showed a dramatic 110-
fold increase in oxidation rate. Based on similar oxidation
traces, K
offNADPH
values for several ecIII mutants were

calculated (Table 5). The release rate of NADPH was only
significantly increased for the ecIIII406C, ecIIIG430C and
ecIIIY431C mutants. In contrast to other mutants, the
NADP(H) bound to ecIIIY431C was rapidly lost upon
storage. The NADPH released was subsequently oxidized,
requiring reloading with equimolar NADPH prior to assay.
The ecIIII406C and ecIIIY431C mutants displayed a
fourfold to fivefold faster release of NADPH. These results
also indicate that there is no obvious relationship between
the release rate of NADPH and the content of bound
NADP(H) (cf. Table 1).
Characterization of the G430C mutant in intact
cysteine-free
E. coli
transhydrogenase
In order to examine the effects of a mutation in the redox-
sensitive region of loop E in the intact E. coli transhydro-
genase, the cfTHG430C mutant was constructed and the
Table 3. Properties of the reverse reaction catalyzed by wild-type and
mutant ecIII in the presence of rrI. The values are estimations from
protein-protein titration curves in which the ecIII concentration was
fixed and the rrI concentration varied (not shown). The assays were
performed as described in Materials and Methods. The [ecIII] values
refer to the fixed enzyme concentration used in the titrations. The [rrI]
values correspond to the concentration of rrI at half V
max
.
Enzyme
[ecIII]
(n

M
)
[rrI]
(n
M
)
[rrI]/
[ecIII]
V
max
(mol AcPyADH)Æ
(mol ecIII)
)1
Æmin
)1
%
ecIII 4900 20 0.004 4 100
ecIIIA398C 2500 30 0.012 4 100
ecIIIS404C 4000 30 0.008 3 75
ecIIII406C 2500 250 0.100 18 450
ecIIIG408C 2500 40 0.016 4 100
ecIIIM409C 5000 150 0.030 3 75
ecIIIV411C 2500 110 0.044 5 125
ecIIIY431C 2500 240 0.096 18 450
Table 4. Properties of the cyclic reaction catalyzed by wild-type and mutant ecIII in the presence of rrI. The values are estimations from protein-
protein titration curves in which the ecIII concentration was fixed and the rrI concentration varied (not shown). The assays were performed as
described in Materials and Methods. The [ecIII] values refer to the fixed enzyme concentration used in the titrations. The [rrI] values correspond to
the concentration of rrI at
1
/

2
V
max
.K
d
is the estimated dissociation constant derived according to K
d
¼ [rrI]-[ecIII]/2.
Enzyme
[ecIII]
(n
M
)
[rrI]
(n
M
)
K
d
(n
M
)
V
max
(mol AcPyADH)Æ
(mol ecIII)
)1
Æmin
)1
%

ecIII 40 70 50 4900 100
ecIIIA398C 12.5 300 294 5000 102
ecIIIS404C 12.5 95 89 4900 100
ecIIII406C 40 250 230 1400 29
ecIIIG408C 40 200 180 1400 29
ecIIIM409C 40 450 430 1200 24
ecIIIV411C 40 150 130 1500 31
ecIIIY431C 40 470 450 800 16
4510 C. Johansson et al.(Eur. J. Biochem. 269) Ó FEBS 2002
resulting mutant protein was characterized with respect to
catalytic activities. The kinetic properties of the various
transhydrogenation activities catalyzed by the cfTHG430C
mutant are summarized in Table 6. The severe effect of
mutating this conserved glycine into a cysteine was clearly
reflected in the resulting maximal rates of the reverse,
forward and cyclic reactions which were all between 7 and
10% of the corresponding wild-type cfTH activities. The K
m
for NADPH in the reverse reaction was 40 times higher
than that for wild-type cfTH whereas the K
m
for thio-
NADP
+
in the forward reaction, remained essentially
unchanged. Consequently, the cfTHG430C mutation resul-
ted in a substantial loss of affinity for NADPH, while the
affinity for NADP
+
was unaffected (Table 6).

For comparative reasons, the transhydrogenation reac-
tions catalyzed by rrI + wild-type ecIII and ecIIIG430C
enzymes mixtures and by cfTH and cfTHG430C enzymes,
Fig. 4. Release of NADPH from ecIII mutants studied by fluorescence.
K
offNADPH
for ecIII mutants were estimated from curves obtained
when monitoring the decrease in fluorescence intensity as ecIII
enzymes were treated with glutathione and glutathione reductase (see
Materials and methods). Upper trace denotes ecIIIA432C and lower
trace ecIIIG430C.
Table 5. Release rates of NADPH from wild-type and mutant ecIII
determined by fluorescence. The release rates of NADPH from ecIII
enzymes were determined by fluorescence as described in Materials
and Methods. The K
offNADPH
was derived from curves obtained by
monitoring the decrease in fluorescence as NADPH was oxidized.
Enzyme
K
offNADPH
(s
)1
)
Relative
rate
ecIII 0.005 1
ecIIIT393C 0.010 2
ecIIIS404C 0.002 0.4
ecIIII406C 0.020 4

ecIIIM409C 0.004 0.8
ecIIIG430C 0.560 110
ecIIIY431C 0.025 5
ecIIIA432C 0.011 2
Table 6. Kinetic parameters of purified cfTH and cfTHG430C enzymes. The K
thioÀNADPþ
m
, K
NADPH
m
and V
max
values were derived from Eadie-Hofstee plots. The fixed concentration of AcPyAD
+
used in the
reverse reaction was 400 l
M
for cfTH and 1500 l
M
for cfTHG430C. The fixed concentration of NADH used in the forward reaction was 400 l
M
for cfTH and 800 l
M
for cfTHG430C. For the cyclic reaction
the following concentrations were used for cfTH; 200 l
M
NADP
+
,200l
M

AcPyaD
+
and 100 l
M
NADH and for cfTHG430C; 1500 l
M
NADP
+
, 1500 l
M
AcPyAD
+
and 750 l
M
NADH.
Forward reaction Reverse reaction Cyclic reaction
V
max
V
max
V
max
Enzyme (lmol thio-NADPH)Æ(mg cfTH)
)1
Æmin
)1
%
K
thioÀNADPþ
m

(l
M
)(lmol AcPyADH)Æ(mg cfTH)
)1
Æmin
)1
%
K
NADPH
m
(l
M
)(lmol AcPyADH)Æ(mg cfTH)
)1
Æmin
)1
%
cfTH 0.5 100 67 3.6 100 4 13.4 100
cfTHG430C 0.05 10 47 0.3 8 175 0.9 7
Ó FEBS 2002 Redox-sensitive loops D and E in transhydrogenase (Eur. J. Biochem. 269) 4511
were all analyzed in buffer C. The maximal rates listed in
Table 7 were obtained from the pH optima of the
respective mixtures of domains or enzymes. For the
reconstituted system, rrI + ecIII, there was a pronounced
difference between the rates of the forward and reverse
reaction, the reverse rate being 150 times higher. For the
reconstituted rrI + ecIIIG430C mutant, this difference in
rates had largely disappeared, the reverse reaction being
only 4.5 times faster (Table 7). Thus, the ratio between the
forward and the reverse reaction rates was approximately

the same for wild-type cfTH and cfTHG430C enzymes,
but the activities displayed by the cfTHG430C mutant
were only 8–10% of those catalyzed by wild-type cfTH
(Table 7). However, the maximal rates of the forward and
reverse reactions exhibited by the rrI + ecIIIG430C
complex and the cfTHG430C enzyme were almost the
same (Table 7).
DISCUSSION
Conformational changes involved in the proton pumping
mechanism of transhydrogenase have been suggested to be
dependent on binding/release of NADP(H) and the redox-
state of this substrate [2,4]. The exact location of NADP
+
and the bonds stabilizing its association with dIII have been
established both in isolated dIII [6,7] and in the dI
2
-dIII
complex [13]; the two forms of dIII do not reveal any major
differences with regard to NADP
+
-binding [13]. Structur-
ally, loop E is involved in the binding of the pyrophosphate
group as well as the ribose of 2¢-5¢-ADP through the
conserved bG430. The preceding conserved K424-R425-
S426 residues bind the 2¢-phosphate [6,7,13]. The structural
role of loop D is less obvious, but it appears to interact with
loop E as well as dI [13], and the semiconserved I406 points
into an apparent crevice formed by loop D and E towards
the nicotinamide ring of NADP
+

. The distance between
I406 and the nicotinamide ring is about 6 A
˚
[13]. Based on
this structural information, it was proposed that loop D and
E(theÔlidÕ)wereinvolvedintheinteractionwithdIand
binding of NADP
+
, respectively, especially the change from
the ÔoccludedÕ state to the open state and vice versa [2,4,13].
These conclusions were supported by NMR studies where
NADP
+
hadbeenreplacedbyNADPHintheR. rubrum
dIII [11]. Studies of ecIII by NMR [10], especially chemical
shifts caused by the presence of ecI and/or NADP(H),
showed that the ecIII itself and the ecI–ecIII interface were
altered upon a change in the redox-state of the bound
NADP(H).
Despite the large amount of structural information
available for loops D and E, their functional roles have
not been systematically examined by site-directed mutage-
genesis. Residues subjected to mutagenesis were therefore
chosen based on the magnitude of their chemical shift
perturbations in the NMR experiments [10], i.e. especially
G389-I406 and G430-V434, and surrounding residues, and
on their conservation among 62 transhydrogenase gene
sequences. Some of these mutants, i.e. ecIIIK424C,
ecIIIH345C, ecIIIA348C, ecIIIR350C [15], and ecIIIT393C,
ecIIIR425C, ecIIIG430C and ecIIIA432C [19], were partly

characterized previously. In order to be able to react these
with thiol-specific reagents, all selected residues were
mutated to cysteines in the cysteine-free E. coli transhydro-
genase.
Table 7. Comparison of the maximal rates of transhydrogenation reactions catalyzed by wild-type and G430C mutants. The V
max
values listed in the table are the maximal rates obtained in the pH range 5.0–9.0
and are given with reference to mol ecIII. The concentrations of substrates and enzymes were as described in the respective section.
Enzyme
Forward reaction Reverse reaction Cyclic reaction
V
max
(mol thio-NADPH) (mol ecIII)
)1
Æmin
)1
pH
V
max
(mol AcPyADH)Æ(mol ecIII)
)1
Æmin
)1
pH
V
max
(molAcPyADH)Æ(mol ecIII)
)1
Æmin
)1

pH
rrI + ecIII 0.04 7.0 6 7.0 2800 7.0
rrI + ecIIIG430C 7 7.0 32 7.0 850 7.0
cfTH 48 7.0 370 7.0 1370 6.0
cfTHG430C 5 7.0 31 5.0 87 6.0
4512 C. Johansson et al.(Eur. J. Biochem. 269) Ó FEBS 2002
The role of loop D
Mutations were introduced in loop D in ecIII, a region that
was suggested by NMR experiments to be involved in
redox-regulation of the interactions of ecIII with ecI [10]. In
addition to ecIIIA398C, ecIIIS404C and ecIIII406C, muta-
tions were also made in the adjacent G408-M409-P410-V/
I411 region, i.e. ecIIIG408C, ecIIIM409C and ecIIIV411C.
Earler made mutants in this region include ecIIID392C [15]
and ecIIIT393 [10]. All of these mutants show a varying
content of bound NADP(H), the most conspicuous being
ecIIIR425C [10] and ecIIID392C [15] which are isolated as
100% apo-form, and ecIIIV411C which has 54% apo-form
(Table 1).
Introduction of a cysteine in the S404 position of ecIII
had little or no effect on the substrate-binding properties
and affinity for dI. This ecIII mutant behaved wild-type like
in all experiments. It should be noted, however, that the
replacement of serine with cysteine is a rather mild
substitution. When the ecIIIS404C mutant was reacted
with MIANS and NEM (A. Pedersen, C. Johansson and
J. Rydstro
¨
m, unpublished results), the reverse reaction was
stimulated by a factor of 1.6 and 2.0, respectively, indicating

that the side-chain of S404 is pointing towards the substrate.
The ecIIII406C mutation led to higher release rates for
NADP
+
and especially NADPH, as indicated by elevated
reverse and forward reaction rates, as well as a faster release
of NADPH in fluorescence measurements. These observa-
tions might seem peculiar as the I406 residue does not make
any specific interactions with NADP(H) in the crystal
structure [6,7,13]. However, as a working hypothesis, it is
possible that I406 creates a suitable environment for the
essential substrate-binding D392. This aspartic acid residue
is located within 6 A
˚
from I406 in the crystal structure and
has been proposed on the basis of mutagenesis to be a key
residue in catalysis/binding as well as the proton pumping,
possibly constituting one end of the proton wire
[2,10,15,20,23–25,27]. Structural evidence [2,4,6,7] supports
these proposals. I406 may take part in the regulation of the
accessibility of the D392 side-chain and thereby control
protonation events. The rrI affinity was also affected by the
ecIIII406C mutation, as shown by the five-fold increase in
the K
d
for the rrI–ecIII complex.
The local environment of I406C was recently demonstra-
ted by MIANS labeling experiments to depend on the
redox-state of the added substrate (A. Pedersen, C.
Johansson, B.G. Karlsson & J. Rydstro

¨
m, unpublished,
results). This difference might reflect a movement of the
I406C side-chain, and probably the entire loop D, that is
coupled to events in the NADP(H)-binding site.
Mutations in the G408-M409-P410-V/I411 region did
not influence the substrate-binding characteristics of ecIII,
but caused a substantial decrease in the affinity for domain
I. The K
d
for the rrI + ecIIIM409C complex was ninefold
higher than that for wild-type ecIII and the maximal rate of
the cyclic reaction was only about 24%, indicating that the
complex was distorted by this mutation.
The role of loop E
Mutations in loop E in isolated ecIII had dramatic
consequences on its interactions with both dI and
NADP(H). The most remarkable property of these mutants
was the high dissociation rates of NADP(H), suggested by
both high forward (Table 2) and reverse (Table 3) reaction
rates, particularly the fast release of NADPH in fluores-
cence measurements. The ecIIIG430C and ecIIIA432C
mutants were earlier shown to be catalyze 850- and 150-fold
increased reverse rates, respectively [10]. In the presence of
rrI, the forward reaction catalyzed by the ecIIIG430C and
ecIIIY431C mutants was 275 and 100 times faster, respect-
ively, than that of wild-type ecIII. As this reaction catalyzed
by the rrI + ecIII complex normally is limited by the slow
release of the NADPH, these high rates indicate high
dissociation rates of NADPH. Indeed, measured directly, a

110-fold increase in the K
offNADPH
was demonstrated for the
ecIIIG430C mutant.
A careful analysis of the structure of ecIII revealed that a
plausible explanation for the above observation is that the
side-chains of G430-Y431-A432 appear to be involved in
specific interactions with NADP(H) and thus could con-
tribute to an increased affinity for this substrate (see
[6,7,13]). The side-chain of a cysteine residue most likely
adopts a different angle than those of the original GYA
residues. This is particularly apparent from the content of
bound NADP(H) and percentage apo-form of the ecIII
mutants. In contrast to wild-type ecIII, both ecIIIG430C
[10] and ecIIIA432C [10] have a reversed content of
NADP(H), i.e. predominantly NADPH with no or little
NADP
+
cIIIY431C also has an altered NADP(H) content,
and all three mutants have high apo-form content.
In addition to being important for the regulation of
NADPH release, loop E also plays a crucial role in the
interactions with dI. This was demonstrated by the high
concentration of rrI needed for half-saturation of the
cyclic reaction catalyzed by the ecIIIY431C mutant
(Table 4), as well as the ecIIIG430C and ecIIIA432C
mutants [10]. The K
d
for the rrI + ecIIIG430C and
rrI + ecIIIG430C complex was about 10 and 15 times,

respectively, higher than that for the rrI + ecIII complex.
As expected, the rate of the cyclic reaction was inversely
proportionate to the K
d
value (Table 4 and [10]), whereas
the rate of the reverse reaction catalyzed by the
ecIIIG430C [10], ecIII/431C (Table 3) and ecIIIA432C
[10] mutants was proportionate to the [rrI]/[ecIII] ratio at
1/2 V
max
, respectively.
The inherent tight binding of NADP(H) in isolated
domain III has previously been explained by the hypo-
thesis that separately expressed dIII mimics the ‘occluded’
conformation of domain III in the intact enzyme [2,4,16].
This occluded conformation is assumed to correspond to a
state in which the hydride transfer step takes place [2].
Consequently, the activities of the reverse and forward
reactions catalyzed by rrI + ecIII, in which release of
NADP(H) is limiting, were very low as compared to that
of intact cfTH (Table 7). The comparison of the rates of
the various transhydrogenation reactions catalyzed by
wild-type and mutants of rrI + ecIII complexes and
intact transhydrogenases allowed an important conclusion
regarding the differences between isolated ecIII and dIII
as it functions in intact transhydrogenase. The ratio of the
rates of the reverse and forward reactions catalyzed by
rrI + ecIIIG430C was similar to that for cfTHG430C, i.e
approximately 7 (Table 6). Normally, this ratio is the
same for cfTH, but 150 for rrI + ecIII (Table 6).

Introduction of a cysteine residue in the G430 position
Ó FEBS 2002 Redox-sensitive loops D and E in transhydrogenase (Eur. J. Biochem. 269) 4513
of ecIII obviously perturbed the conformation of loop E
in such a way that allowed it to function essentially as in
the wild-type intact enzyme/cfTH, i.e. with a strongly
increased rate of the forward reaction and a less strong
increase in the reverse reaction. Indeed, this change is
consistent with the 40-fold elevated K
mNADPH
for
cfTHG430C while K
m
thio–NADP+
was unchanged. The
elevated K
m
for NADPH is most likely a consequence of a
higher dissociation rate of this substrate, assuming that the
K
on
of NADPH is unchanged. Thus, even though the
evidence is rather indirect, it is conceivable that the tight
binding of NADP(H) in the occluded state of domain III
directly or indirectly involves G430 and/or the conforma-
tion of loop E. An interesting possibility is therefore that
G430 and loop E in the resting state of the intact enzyme
(and in the ecIIIG430C mutant) are much less associated
with NADP(H), i.e. the ÔlidÕ is open.
As proton translocation in transhydrogenase is very likely
associated with conformational changes that affect binding

and release of NADP(H) [2,4], loop E may play a major role
in the coupling mechanism of transhydrogenase. In addi-
tion, the changes in affinity for NADP(H) could be
communicated to domain I as loop E forms part of the
region that confers a redox regulation of the ecI + ecIII
complex interface.
In conclusion, the present results suggest that loop D is
involved in the interactions with domain I and that the I406
residue is a potential candidate for the regulation of the
accessibility of the side-chain of the D392 residue that is
essential for proton-pumping. Moreover, the results support
the notion that loop E functions as a mobile lid [4,7,13],
regulating the release of NADPH, a step that probably is of
central importance in the coupling mechanism of transhy-
drogenase. It is proposed that movements of these two loops
work in concert to regulate the affinity of NADP(H),
protonation events in their surroundings and to communi-
cate these changes to domain I.
ACKNOWLEDGEMENTS
This work was supported by the Swedish Natural Science Research
Council. AP acknowledges a grant from the Sven and Lilly Lawski
Foundation.
REFERENCES
1. Olausson, T., Fjellstro
¨
m, O., Meuller, J. & Rydstro
¨
m, J. (1995)
Molecular biology of nicotinamide nucleotide transhydrogenase –
a unique proton pump. Biochim. Biophys. Acta 1231, 1–19.

2. Bizouarn, T., Fjellstro
¨
m, O., Meuller, J., Axelsson, M., Bergkvist,
A., Johansson, C., Karlsson, B.G. & Rydstro
¨
m, J. (2000) Proton
translocating nicotinamide nucleotide transhydrogenase from
E. coli. Mechanism of action deduced from its structural and
catalytic properties. Biochim. Biophys. Acta 1457, 211–228.
3. Bragg, P.D. (1998) Site-directed mutagenesis of the proton-
pumping pyridine nucleotide transhydrogenase of Escherichia coli.
Biochim. Biophys. Acta 1365, 98–104.
4. Jackson,J.B.,Scott,S.A.,Quirk,P.G.&Venning,J.D.(2002)The
alternating site, binding change mechanism for proton transloca-
tion by transhydrogenase. Biochemistry 41, 4173–4185.
5. Buckley, P.A., Jackson, J.B., Schneider, T., White, S.A., Rice,
D.W. & Baker, P.J. (2000) Protein-protein recognition, hydride
transfer and proton pumping in the transhydrogenase complex.
Structure 8, 809–815.
6. Prasad, G.S., Sridhar, V., Yamaguchi, M., Hatefi, Y. & Stout,
C.D. (1999) Crystal structure of transhydrogenase domain III at
1.2 angstrom resolution. Nat. Struct. Biol. 6, 1126–1131.
7.White,S.A.,Peake,S.,McSweeney,S.,Leonard,G.,Cotton,
N.P.J. & Jackson, J.B. (1999) The high-resolution structure of the
NADP(H)-binding component (dIII) of proton-translocating
transhydrogenase from human heart mitochondria. Structure
8, 1–12.
8. Johansson, C., Bergkvist, A., Fjellstro
¨
m, O., Rydstro

¨
m, J. &
Karlsson, G. (1999) Sequential assignment and secondary struc-
ture analysis of the NADP(H)-binding domain of Escherichia coli
transhydrogenase. J. Biomol. NMR. 14, 295–296.
9. Johansson, C., Bergkvist, A., Fjellstro
¨
m, O., Rydstro
¨
m, J. &
Karlsson, B.G. (1999) NMR characterization of the NADP(H)-
binding domain of Escherichia coli transhydrogenase: sequential
assignment and global fold. FEBS Lett. 458, 180–184.
10. Bergkvist, A., Johansson, C., Johansson, T., Rydstro
¨
m, J. &
Karlsson, B.G. (2000) Interactions of the NADP(H)-binding
domain III of proton-translocating transhydrogenase from
Escherichia coli with NADP (H) and the NAD (H) -binding
domain I studied by NMR and site-directed mutagenesis.
Biochemistry 39, 12595–12605.
11. Quirk,P.G.,Jeeves,M.,Cotton,N.P.,Smith,J.K.&Jackson,J.B.
(1999) Structural changes in the recombinant, NADP (H) -binding
component of proton translocating transhydrogenase revealed by
NMR spectroscopy. FEBS Lett. 446, 127–132.
12. Jeeves, M., Smith, J., Quirk, P.G., Cotton, N.P.J. & Jackson, J.B.
(2000) Solution structure of the NADP(H)-binding component
(dIII) of proton-translocating transhydrogenase from Rhodospir-
illum rubrum. Biochim. Biophys. Acta. 1459, 248–257.
13. Cotton, N.P.J., White, S.A., Peake, S.J., McSweeney, S. & Jack-

son, J.B. (2001) The crystal structure of an asymmetric complex of
the two nucleotide binding components of proton-translocating
transhydrogenase. Structure 9, 165–176.
14. Fjellstro
¨
m, O., Bizouarn, T., Zhang, J.W., Rydstro
¨
m, J., Venning,
J.D. & Jackson, J.B. (1999) Catalytic properties of hybrid com-
plexes of the NAD(H)-binding and NADP(H)-binding domains
of the proton-translocating transhydrogenases from Escherichia
coli and Rhodospirillum rubrum. Biochemistry 38, 415–422.
15. Fjellstro
¨
m,O.,Axelsson,M.,Bizouarn,T.,Hu,X.,Johansson,C.,
Meuller, J. & Rydstro
¨
m, J. (1999) Mapping of residues in the
NADP(H)-binding site of proton-translocating nicotinamide
nucleotide transhydrogenase from Escherichia coli.Astudyof
structure and function. J. Biol. Chem. 274, 6350–6359.
16. Diggle, C., Bizouarn, T., Cotton, N.P. & Jackson, J.B. (1996)
Properties of the purified, recombinant, NADP(H)-binding
domain III of the proton-translocating nicotinamide nucleotide
transhydrogenase from Rhodospirillum rubrum. Eur. J. Biochem.
241, 162–170.
17. Bizouarn, T., Grimley, R., Diggle, C., Thomas, C.M. & Jackson,
J.B. (1997) Mutations at tyrosine-235 in the mobile loop region of
domain I protein of transhydrogenase from Rhodospirillum
rubrum strongly inhibit hydride transfer. Biochim. Biophys. Acta

1320, 265–274.
18. Smith, P.K., Krohn, R.I., Hermanson, G.T., Mallia, A.K.,
Gartner, F.H., Provenzano, M.D., Fujimoto, E.K., Goeke, N.M.,
Olson, B.J. & Klenk, D.C. (1985) Measurement of protein using
bicinchoninic acid. Anal Biochem. 150, 76–85.
19. Fjellstro
¨
m, O., Johansson, C. & Rydstro
¨
m, J. (1997) Structural
and catalytic properties of the expressed and purified NAD(H)-
and NADP(H)-binding domains of proton-pumping transhy-
drogenase from Escherichia coli. Biochemistry 36, 11331–11341.
20. Meuller,J.,Zhang,J.,Hou,C.,Bragg,P.D.&Rydstro
¨
m, J. (1997)
Properties of a cysteine-free proton-pumping nicotinamide
nucleotide transhydrogenase. Biochem. J. 324, 681–687.
21. Zhang, J., Hu, X., Osman, A.M. & Rydstro
¨
m, J. (1997)
Effects of metal ions on the substrate-specificity and activity of
4514 C. Johansson et al.(Eur. J. Biochem. 269) Ó FEBS 2002
proton-pumping nicotinamide nucleotide transhydrogenase from
Escherichia coli. Biochim. Biophys. Acta 1319, 331–339.
22. Hutton, M., Day, J.M., Bizouarn, T. & Jackson, J.B. (1994)
Kinetic resolution of the reaction catalysed by proton-translo-
cating transhydrogenase from Escherichia coli as revealed by
experiments with analogues of the nucleotide substrates. Eur. J.
Biochem. 219, 1041–1051.

23. Meuller,J.,Hu,X.,Bunthof,C.,Olausson,T.&Rydstro
¨
m, J.
(1996) Identification of an aspartic acid residue in the beta subunit
which is essential for catalysis and proton pumping by transhy-
drogenase from Escherichia coli. Biochim. Biophys. Acta 1273,
191–194.
24. Hu, X., Zhang, J., Fjellstro
¨
m, O., Bizouarn, T. & Rydstro
¨
m, J.
(1999) Site-directed mutagenesis of charged and potentially pro-
ton-carrying residues in the beta subunit of the proton-translo-
cating nicotinamide nucleotide transhydrogenase from Escherichia
coli. Characterization of the beta H91, beta D392, and beta K424
mutants. Biochemistry 38, 1652–1658.
25. Bragg, P.D. & Hou, C. (2001) Characterization of mutants of b
aspartate213, and b aspartate222, possibly components of the
energy transduction pathway of the proton-translocating pyridine
nucleotide transhydrogenase of Escherichia coli. Arch. Biochem.
Biophys. 388, 299–307.
26. Koradi, R., Billeter, M. & Wu
¨
thrich, K. (1996) MOLMOL:
a program for display and analysis of macromolecular structures.
J. Mol. Graphics 14 (51–5), 29–32.
27. Rodrigues, D.J., Venning, J.D., Quirk, P.G. & Jackson, J.B.
(2001) A change in ionization of the NADP(H)-binding compo-
nent (dIII) of proton-translocating transhydrogenase regulates

both hydride transfer and nucleotide release. Eur. J. Biochem. 268,
1430–1438.
Ó FEBS 2002 Redox-sensitive loops D and E in transhydrogenase (Eur. J. Biochem. 269) 4515