Microbiology (2006), 152, 1119–1128 DOI 10.1099/mic.0.

28612-0

Glucose-6-phosphate dehydrogenase and

ferredoxin-NADP(H) reductase contribute to
damage repair during the soxRS response of
Escherichia coli
Mariana Giró, Néstor Carrillo and Adriana R. Krapp
Correspondence Instituto de Biologı́a Molecular y Celular de Rosario (IBR, CONICET), División Biologı́a
Adriana R. Krapp Molecular, Facultad de Ciencias Bioquı́micas y Farmacéuticas, Universidad Nacional de Rosario,
krapp@ibr.gov.ar Suipacha 531, 2000 Rosario, Argentina

The NADP(H)-dependent enzymes glucose-6-phosphate dehydrogenase (G6PDH) and

ferredoxin(flavodoxin)-NADP(H) reductase (FPR), encoded by the zwf and fpr genes, respectively,
are committed members of the soxRS regulatory system involved in superoxide resistance in
Escherichia coli. Exposure of E. coli cells to the superoxide propagator methyl viologen (MV) led to
rapid accumulation of G6PDH, while FPR was induced after a lag period of several minutes.
Bacteria expressing G6PDH from a multicopy plasmid accumulated higher NADPH levels and
displayed a protracted soxRS response, whereas FPR build-up had the opposite effects.
Inactivation of either of the two genes resulted in enhanced sensitivity to MV killing, while further
increases in the cellular content of FPR led to higher survival rates under oxidative conditions.
In contrast, G6PDH accumulation over wild-type levels of expression failed to increase MV
tolerance. G6PDH and FPR could act concertedly to deliver reducing equivalents from
carbohydrates, via NADP+, to the FPR acceptors ferredoxin and/or flavodoxin. To evaluate
whether this electron-transport system could mediate reductive repair reactions, the pathway
was reconstituted in vitro from purified components; the reconstituted system was found to be
functional in reactivation of oxidatively damaged iron–sulfur clusters of hydro-lyases such as
aconitase and 6-phosphogluconate dehydratase. Recovery of these activities after oxidative
challenge was faster and more extensive in transformed bacteria overexpressing FPR than in
wild-type cells, indicating that the reductase could sustain hydro-lyase repair in vivo. However,
Received 20 October 2005 FPR-deficient mutants were still able to fix iron–sulfur clusters at significant rates, suggesting that
Revised 29 December 2005 back-up routes for ferredoxin and/or flavodoxin reduction might be called into action to rescue
Accepted 4 January 2006 inactivated enzymes when FPR is absent.

INTRODUCTION this regulon is the SoxR protein, a dimeric transcription

factor that contains [2Fe–2S] centres (Nunoshiba et al.,
Reactive oxygen species (ROS), such as the superoxide anion
1992; Wu & Weiss, 1992). When E. coli cells are exposed
radical and its derivatives hydrogen peroxide and the
to superoxide-propagating compounds such as the redox-
hydroxyl radical, represent a serious threat to aerobic life.
cycling herbicide methyl viologen (MV), the iron–sulfur
Along the path of evolution, organisms thriving in air have
cluster of SoxR undergoes univalent oxidation to yield the
developed a number of adaptive devices to cope with this
oxidized, active form of the protein, which subsequently
menace. In Escherichia coli and other enterobacteria, the stimulates the synthesis of SoxS, a transcriptional activator
induction of a suite of unlinked genes by the soxRS regula- of the AraC/XylS family, by productive interaction with its
tory system confers tolerance to superoxide and nitric oxide promoter (Ding & Demple, 1997; Gaudu et al., 1997;
(Gaudu et al., 1997; Greenberg et al., 1990; Nunoshiba et al., Pomposiello & Demple, 2001). Increased SoxS levels, in
1992; Pomposiello & Demple, 2001; Pomposiello et al., 2001; turn, induce expression of the various genes of the regulon
Tsaneva & Weiss, 1990; Wu & Weiss, 1992). The sensor of via s70 RNA polymerase (Gaudu et al., 1997; Pomposiello &
Demple, 2001). Components of the soxRS system are recruited
Abbreviations: Fd, ferredoxin; Fld, flavodoxin; FPR, ferredoxin(flavodoxin)-
to combat the toxic effects of oxidants at various levels,
NADP(H) reductase; G6PDH, glucose-6-phosphate dehydrogenase; MV, including ROS scavenging, replacement of sensitive targets
methyl viologen; 6PGD, 6-phosphogluconate dehydratase; ROS, reactive by resistant counterparts and damage repair (Gaudu et al.,
oxygen species. 1997; Pomposiello & Demple, 2001). Most soxRS members
Downloaded from www.microbiologyresearch.org by
0002-8612 G 2006 SGM Printed in Great Britain 1119
IP: 45.225.232.158
On: Mon, 04 Dec 2017 15:11:38
M. Giró, N. Carrillo and A. R. Krapp

are also subject to transcriptional modulation by the SoxS acid to bind the leaving hydroxyl group during substrate
homologues MarA and Rob, which activate the closely dehydration (Djaman et al., 2004; Imlay, 2003). Their sus-
related regulons marRAB and rob, involved in multiple anti- ceptibility to superoxide stems from the ability of this ROS
biotic resistance and tolerance to organic solvents, respec- to oxidize the catalytic iron to generate [4Fe–4S]3+, an unstable
tively (Alekshun & Levy, 1997). intermediate that rapidly decomposes into [3Fe–4S]1+ and
Fe2+ (Imlay, 2003), thereby inactivating the corresponding
Among the SoxS targets present in E. coli, there are a number enzymes and disabling the metabolic pathways to which
of genes encoding enzymes and proteins engaged in oxido- they belong (Flint et al., 1993; Gardner & Fridovich, 1991a,
reductive processes. They include glucose-6-phosphate b; Kennedy et al., 1983; Varghese et al., 2003). Conversion
dehydrogenase (G6PDH), ferredoxin(flavodoxin)-NADP(H) of the [3Fe–4S]1+ cluster back to the active form requires
reductase (FPR) and its electron acceptor substrates flavo- reduction and metallation, taking place within a few minutes
doxin (Fld) I and II (Griffith & Wolf, 2001; Pomposiello & after the oxidative condition has subsided (Djaman et al.,
Demple, 2001; Zheng et al., 1999). G6PDH, encoded by the 2004). E. coli mutants lacking Fd are still able to repair iron–
zwf gene, catalyses the first step in the oxidative branch of the sulfur centres, albeit at lower rates (Djaman et al., 2004),
pentose phosphate pathway, which generates ribose for suggesting that this carrier may be a physiological electron
nucleoside synthesis and NADPH for reductive pathways donor for the process. A comparable contribution of iso-
and repair reactions (Csonka & Fraenkel, 1977; Fraenkel, functional Fld has not been evaluated so far. G6PDH and
1987). Besides its induction by the soxRS/marRAB/rob sys- FPR could act concertedly in the provision of reducing
tems, G6PDH expression undergoes growth-rate-dependent equivalents for dehydratase reactivation by establishing a
regulation on different carbon sources (Rowley et al., 1991; short electron-transport chain in which G6PDH supplies
Wolf et al., 1979). E. coli and Salmonella strains devoid of NADPH, which could be later used by FPR as a substrate in
G6PDH activity are still able to grow on glucose (Fraenkel, the reduction of Fd and/or Fld.
1968), but display increased susceptibility to oxidants and
killing by murine macrophages (Greenberg et al., 1990; Build-up of NADPH levels during episodes of oxidative
Lundberg et al., 1999; Nunoshiba et al., 1995). Yeast G6PDH stress may also have unwanted consequences. Accumulation
mutants are also abnormally sensitive to oxidative stress of the reduced nucleotide is expected to downregulate the
(Nogae & Johnston, 1990), and in humans, the G6PDH soxRS system by keeping the SoxR sensor in a reduced state
deficiency responsible for haemolytic anaemia is character- (Gardner & Fridovich, 1993; Koo et al., 2003), thus slowing
ized by enhanced oxidant sensitivity and decline of NADPH down or even switching off the entire response. In addition,
levels in erythrocytes (Scott et al., 1991). NADPH may favour the propagation of deadly hydroxyl
radicals through the Fenton reaction by redox-cycling the
The physiological role played by FPR during normal growth free iron leached from iron–sulfur clusters, either directly
is poorly understood. In nonphotosynthetic organisms and (Brumaghim et al., 2003) or as a substrate of flavin reductase
tissues, this FAD-containing enzyme mediates electron (Woodmansee & Imlay, 2002). This enzyme generates
transfer from NADPH to ferredoxin (Fd) or Fld, providing reduced flavins, which are the preferred physiological Fen-
low-potential electron carriers required for a plethora of ton reductants during hydroxyl radical formation in vivo
oxido-reductive pathways (reviewed by Carrillo & Ceccarelli, (Woodmansee & Imlay, 2002). The coordinated induction
2003; Ceccarelli et al., 2004). Fld is employed in E. coli for of G6PDH and FPR during the soxRS response could exert
the reductive activation of several anaerobic enzymes counteracting effects on the NADPH pool that might be
(Blaschkowski et al., 1982; Wan & Jarrett, 2002), and Fd important for the maintenance of redox homeostasis in the
for the assembly of iron–sulfur clusters (Djaman et al., stressed cell (Krapp et al., 2002).
2004). Insertional mutants lacking FPR display no obvious
growth penalty, the only phenotype being, once again, To gain insights into the roles of G6PDH and FPR in the
reduced tolerance to oxidative damage (Bianchi et al., 1995; protection against oxidative stress, we investigated several
Krapp et al., 1997, 2002). aspects of their function as members of the soxRS regulon.
We report herein that G6PDH and FPR could establish a
While most components of the soxRS regulon play distinct minimal electron-transport system that provides reducing
and well-recognized protective roles, little is known of the equivalents for scavenging and repair reactions, including
actual contributions of G6PDH and FPR to the concerted reductive reactivation of oxidized hydro-lyases. FPR partici-
cell response against oxidative stress. The beneficial effects of pates in this pathway by supplying reduced Fd and/or Fld,
G6PDH have been attributed to the provision of NADPH but its contribution did not determine survival in vivo,
for scavenging and repair reactions (Lundberg et al., 1999), presumably due to the existence of alternative routes for
but the identities of the pathways that benefit from increased reduction of these low-potential electron carriers.
provision of reducing equivalents remain conjectural.
Among the early targets of superoxide toxicity there is a
family of metal-dependent hydro-lyases that includes METHODS
fumarase A, aconitase B, 6-phosphogluconate dehydratase Bacterial strains and culture media. E. coli cells used in this
(6PGD) and hydroxyacid dehydratase. These enzymes work (Table 1) were routinely grown at 37 uC in Luria–Bertani (LB)
employ a solvent-exposed [4Fe–4S]2+ cluster as a Lewis or M9 minimal media (Sambrook et al., 1989), supplemented with
Downloaded from www.microbiologyresearch.org by
1120 Microbiology 152
IP: 45.225.232.158
On: Mon, 04 Dec 2017 15:11:38
 G6PDH and FPR contribution to the soxRS response

Table 1. E. coli strains and plasmids

Strain or plasmid Genotype or relevant characteristics* Source or reference

Strains
HB301 W3110 D(argF–lac)U169 Rowley et al. (1991)
HB351 W3110 D(argF–lac)U169 zeb-1 : : Tn10 D(edd-zwf)22 Rowley et al. (1991)
DF100 HfrC tonA22 garB10 ompF627 relA1 pit-10 spoT1T2r Rowley et al. (1992)
DF102 DF100 zwfL2 Rowley et al. (1992)
B247 MC4100 (lJW2), Ampr Kanr [lJW2:W(soxS9 : : lacZ9)] Wu & Weiss (1992)
r
BL21(DE3)pLysS F2 ompT hsdSB (r{ {
B mB ) dcm gal (DE3), pLysS, Cam Novagen
2
GC4468 F Dlac U169 rpsL Greenberg et al. (1990)
Plasmids
pSU18 pACYC184 carrying the multiple cloning site of pUC18, Camr Martinez et al. (1988)
pDR17 pBR322 carrying the E. coli zwf gene, Ampr Rowley et al. (1991)
pSUG6PDH pSU18 carrying the E. coli zwf gene, Camr This study
pETG6PDH pET-28b(+) carrying E. coli zwf gene, Kanr This study
pSUFPR pSU18 carrying the E. coli fpr gene, Camr Krapp et al. (2002)

*Ampr, ampicillin resistance; Camr, chloramphenicol resistance; Kanr, kanamycin resistance.

0?4 % (w/v) glucose and the corresponding antibiotics. IPTG was added Preparation of recombinant G6PDH and antisera. The zwf
at a final concentration of 0?5 mM when expression of plasmid-borne gene present in plasmid pDR17 (Rowley et al., 1991) was amplified
genes (Table 1) was desired. Cells used for determination of aconitase by PCR using two primers with specific restriction sites incorporated
and 6PGD activities were cultured in a gluconate medium containing (BamHI and HindIII). The forward and reverse oligonucleotides
60 mM K2HPO4, 33 mM KH2PO4, 7?6 mM (NH4)2SO4, 1?7 mM were designed to hybridize with the +68 to +81 and the +1542 to
sodium citrate, 1 mM MgSO4, 10 mg thiamine hydrochloride ml21, +1559 regions of the zwf gene, respectively, counting from the tran-
0?25 % (w/v) Casamino acids and 0?4 % (w/v) potassium gluconate, scription initiation site. The PCR product was cloned in pGEM-T-
adjusted to pH 7. easy (Promega), digested with BamHI and HindIII, and finally
ligated to compatible sites of pET-28b(+) (Novagen). The resulting
Bacterial viability. Cells were grown in LB broth to OD600 0?3, plasmid, pETG6PDH, contained the entire zwf coding region fused
serially diluted and spotted on M9 agar plates containing 100 mg in-frame to an N-terminal hexahistidine tag. After expression in E.
chloramphenicol ml21 and 0?5 mM IPTG, when required, and the coli BL21(DE3)pLysS, the fusion protein was isolated by passage
indicated concentrations of MV. In the disk diffusion method, through a Ni-NTA agarose column (Qiagen). Purified G6PDH dis-
100 ml of a bacterial suspension containing ~109 cells ml21 was played a specific activity of about 100 units mg21 and migrated as a
mixed with 3 ml 0?7 % (w/v) molten agar at 42 uC and poured onto single 55 kDa polypeptide in SDS-PAGE. Antisera directed against
M9 agar plates. After hardening, MV solutions of various concentra- both G6PDH and FPR were prepared by rabbit immunization.
tions were added in 5 ml aliquots to paper disks (5 mm diameter)
placed on the agar surface. The zones of growth inhibition were MV-dependent induction of G6PDH and FPR. Exposure of E.
measured after incubation for 30 h at 37 uC. Statistical analysis was coli GC4468 cells to MV was carried out by diluting 10 ml of an
conducted using a two-sided t-test. overnight culture in 1 litre of fresh LB medium. The resulting sus-
pension was incubated for 30 min at 37 uC with vigorous shaking
Enzymic assays. For the preparation of E. coli extracts, cells were and MV was added to a final concentration of 100 mM. Fractions
disrupted by sonic oscillation, and the resulting lysates cleared by were transferred at various times to prechilled tubes, centrifuged and
centrifugation at 15 000 g for 15 min. Protein concentration was resuspended in 1 ml of a solution containing 50 mM phosphate
estimated in the supernatants by a dye-binding assay (Sedmak & buffer, pH 7?6, 0?1 mM EDTA, 0?1 mM phenylmethylsulfonyl fluor-
Grossberg, 1977), using bovine serum albumin as standard. G6PDH, ide and 200 mg chloramphenicol ml21 to prevent de novo protein
FPR and b-galactosidase activities were determined according to translation. Cells were lysed and the contents of G6PDH and FPR
Kao & Hassan (1985), Krapp et al. (2002) and Miller (1992), respec- analysed by SDS-PAGE and immunoblotting with specific antisera.
tively. Aconitase was measured at 25 uC in 90 mM Tris/HCl, pH 8, Secondary antibodies conjugated to alkaline phosphatase were
20 mM sodium isocitrate, by following the formation of cis- employed for detection. To estimate the half-life of the enzymes
aconitate at 240 nm (e240=3?6 mM21 cm21). 6PGD activity was upon removal of the inducer, cultures were exposed to 100 mM MV
determined in 50 mM Tris/HCl, pH 7?6, 10 mM MgCl2, 8 mM for 4 h at 37 uC, harvested, washed in LB broth to eliminate MV,
6-phosphogluconate. After 5 min at 25 uC, the reaction was stopped and processed as indicated above. Immunoreactive bands were inte-
by dilution with 1 ml of 50 mM Tris/HCl, pH 7?6, and heated in a grated using the Multi-Analyst Package 1.1 from Bio-Rad, and the
boiling water bath for 2 min. Samples were centrifuged and the amounts of G6PDH and FPR were estimated by comparison with
amounts of pyruvate were determined in the supernatants by reac- blotted pure enzymes of known concentration.
tion with NADH and lactate dehydrogenase (Gardner & Fridovich,
1991a). Time-course of soxRS induction by MV. Overnight cultures of
B247 cells carrying a chromosomal soxS9 : : lacZ fusion (Wu & Weiss,
In all cases, one activity unit is defined as the amount of enzyme that 1992) and transformed with either pSU18, pSUFPR (Krapp et al.,
catalyses the transformation of 1 mmol substrate per minute under the 2002) or pSUG6PDH (containing the zwf gene cloned in the
conditions of the assay. BamHI/HindIII sites of pSU18) were diluted 1/100 in fresh LB broth
Downloaded from www.microbiologyresearch.org by
http://mic.sgmjournals.org 1121
IP: 45.225.232.158
On: Mon, 04 Dec 2017 15:11:38
M. Giró, N. Carrillo and A. R. Krapp

supplemented with 25 mg ml21 chloramphenicol. Cells were cultured We compared the time-courses of G6PDH and FPR
at 37 uC, and 0?5 mM IPTG and 100 mM MV were successively accumulation in cells exposed to 100 mM MV, as well as
added at 30 min and 150 min of incubation, respectively. Samples
the kinetics of disappearance after removal of the reagent
were removed at various times after MV challenge to assay for
b-galactosidase and G6PDH activities. (Fig. 1). This concentration of MV provided for almost
maximal induction of the soxRS response (see below), with
The NADP(H) levels were estimated by a redox cycling assay, after negligible effects on growth rate. The levels of the two
alkaline extraction of the pyridine nucleotides (Krapp et al., 2002). enzymes were estimated by immunoblotting with specific
antisera, instead of using promoter fusions as is common
Oxidative damage and repair of dehydratases containing practice, to incorporate post-transcriptional and transla-
[4Fe–4S]2+ centres. To measure the MV-dependent inactivation
tional regulation, if any, and to measure actual accumula-
and in vivo recovery of aconitase and 6PGD, GC4468 cells trans-
formed with either pSU18 or pSUFPR were cultured at 37 uC in tion rates. The results confirmed that FPR is a minor protein
500 ml gluconate medium to OD600 0?6–0?8. MV was then added to in unstressed E. coli cells (~0?01 % of the total soluble
a final concentration of 100 mM, and incubation was continued for
30 min in the presence of 200 mg kanamycin ml21 to block new
protein synthesis. Cells were collected by centrifugation, rinsed to
remove MV, and resuspended in 60 ml of the same medium con-
taining kanamycin. The bacterial suspension was incubated at 37 uC
without agitation. Aliquots were removed at intervals and centri-
fuged at 4 uC. The collected cells were lysed by sonic oscillation in
1 ml ice-cold 50 mM Tris/HCl, pH 7?6, containing 0?6 mM MnCl2
and 20 mM barium DL-fluorocitrate to stabilize aconitase (Gardner
& Fridovich, 1991b). Lysates were clarified by centrifugation at
15 000 g for 15 min, and the supernatants were immediately assayed
for the corresponding enzymic activities. The same procedure was
used to estimate inactivation and repair of hydro-lyases in FPR-
deficient mutants, except that chloramphenicol (200 mg ml21) was
employed instead of kanamycin to inhibit translation.

To evaluate dehydratase reactivation in vitro, GC4468 cells transformed

with pSU18 were grown in gluconate medium to OD600 0?8, harvested
and ruptured as described above. Lysates were stirred for 30 min
at 30 uC to inactivate oxygen-sensitive dehydratases. Reactivation
experiments were carried out in a reconstituted system made up of
50 mM Tris/HCl, pH 7?6, 0?3 mM NADP+, 3 mM glucose 6-
phosphate, 1 unit G6PDH, 0?3 mM FPR, 5 mM of either Fd or Fld,
1 mM L-cysteine and bacterial extracts corresponding to 150 mg Fig. 1. Induction and decay of FPR and G6PDH during the
soluble protein (complete system). Fractions were taken at various soxRS response of E. coli. The upper parts of the figure show
times and assayed for aconitase and 6PGD activities. Recombinant
typical immunoblots obtained during induction (a) and decay (b)
FPR, Fd and Fld were prepared by published procedures (Wan &
of the soxRS response, and the lower parts illustrate the
Jarrett, 2002). Statistical analysis was conducted using a two-sided
t-test. G6PDH (#) and FPR ($) time-courses. E. coli ribosomal
protein S1 was used as a control to detect any lane-to-lane
differences in total protein loading. (a) Time-courses of FPR
and G6PDH build-up during MV challenge. Aliquots were with-
RESULTS AND DISCUSSION drawn at the times indicated, cells were lysed and supernatants
corresponding to 7 (G6PDH) or 20 (FPR) mg soluble protein
Time-course of G6PDH and FPR induction were separated by SDS-PAGE and immunoblotting. G6PDH
during the soxRS response of E. coli and FPR amounted to 3?21±0?55 (0?32 %) and 0?12±0?02
(0?01 %) mg per mg of total soluble protein, respectively, in
All members of the soxRS/marRAB/rob regulons contain
uninduced E. coli cells. Further increases in both enzymes with
recognition sequences for the corresponding transcription
respect to total protein were normalized to these concentra-
factors, termed the ‘soxbox’ or ‘marbox’ sites (Griffith &
tions. (b) Half-lives of FPR and G6PDH after MV removal. Cells
Wolf, 2001; Martin et al., 2000; Wood et al., 1999). Each
that had been incubated for 4 h at 37 6C with 100 mM MV
regulatory protein differs in the extent to which it activates were washed with fresh LB broth and processed as indicated
particular genes, with the consequence that regulon mem- above. The relative amounts of each enzyme accumulated after
bers are expressed to different levels during, for instance, the 4 h of MV induction (~3?9 % and ~0?3 % of total soluble
soxRS response. In a number of experiments using tran- protein for G6PDH and FPR, respectively) were taken as 1,
scriptional fusions between the promoters and lacZ, and subsequent declines expressed as a fraction of these
maximal induction of fpr (and fldA) by both MV and values. The inset in (b) shows a semilogarithmic representation
SoxS was more than twofold higher than that of zwf (Martin of the decline in G6PDH and FPR. Data points represent the
et al., 2000; Wood et al., 1999). However, these assays only means of three independent experiments with SE¡15 % of the
addressed the transcriptional component of expression and mean. Curves were fit to the experimental data using Sigma
did not investigate the induction kinetics. Plot 8.02.
Downloaded from www.microbiologyresearch.org by
1122 Microbiology 152
IP: 45.225.232.158
On: Mon, 04 Dec 2017 15:11:38
 G6PDH and FPR contribution to the soxRS response

protein), whereas basal G6PDH levels were comparatively normally in the overexpressing bacteria, although they
high (~0?3 % of the total soluble protein), in accordance accumulated 40 % less b-galactosidase than the correspond-
with its involvement in the pentose phosphate pathway. ing controls at all times assayed (Fig. 2). G6PDH activities
Expression of this dehydrogenase was rapidly induced by increased further on exposure to MV (data not shown),
MV, while FPR levels showed little change during the first indicating that SoxS-dependent induction was still func-
30 min of exposure. After that lag period, the reductase tional to a significant extent. As already reported (Gaudu
contents increased steeply and overtook the maximal induc- et al., 2000), zwf mutants displayed a stronger response to
tion reached by G6PDH (Fig. 1a), in agreement with the MV with a maximum induction of the soxS promoter that
data obtained with fused promoters (Martin et al., 2000; was approximately twofold over that of wild-type cells (data
Wood et al., 1999). G6PDH and FPR accumulation corre- not shown), but kinetic characterization of the time-course
lated with concomitant increases in specific activity (data proved difficult due to the extreme MV sensitivity of the
not shown). When the cells were returned to normal growth deficient bacteria.
conditions, the levels of the two enzymes declined to a stable
value within 2 h of MV removal (Fig. 1b). The behaviour of the zwf strains contrasted with that of cells
in which the FPR contents were modified. Overexpression of
G6PDH is considered the main source of NADPH in E. coli the flavoprotein resulted in higher soxRS induction in the
and related bacteria (Csonka & Fraenkel, 1977). Hence, presence of MV, about threefold above the level reached by
the cellular nucleotide pool may be reduced immediately cells transformed with pSU18 (Fig. 2, inset). Time-courses
after superoxide challenge due to rapid induction of this of soxRS induction in the transformants expressing FPR
dehydrogenase, while NADPH-consuming FPR only began were similar to those previously reported (data not shown,
to accumulate at later stages of the soxRS response. but see Krapp et al., 2002). The inducing effect of FPR is
similar to that of other NADPH consumers that stimulate
Effect of G6PDH and FPR contents on the soxRS response, such as Fld I (Zheng et al., 1999) or
MV-driven induction of the soxRS regulon desulfoferrodoxin (Gaudu et al., 2000).
The size and redox state of the NADP(H) pool are expected
to influence the progress of the soxRS response because
NADPH is the source of reducing equivalents for the SoxR
reductase, which maintains the sensor protein reduced and
inactive (Koo et al., 2003). Indeed, activation of the soxRS
regulon by redox-cycling agents has been directly attributed
to faulty SoxR reductase function due to NADPH depletion
(Gaudu et al., 2000; Koo et al., 2003; Liochev & Fridovich,
1992). As already indicated, changes in G6PDH expression
might affect the onset of the soxRS response by modifying
the redox status of the pyridine nucleotide pool. To test
this contention, we monitored the MV-dependent induc-
tion of a reporter soxS9 : : lacZ gene fusion in E. coli strains
expressing various levels of the dehydrogenase. A survey of
the existing literature indicates that induction of the soxRS
response has been studied under a wide range of MV
concentrations, from less than 10 mM (Gaudu et al., 2000)
up to 500 mM (Griffith et al., 2004). In a detailed study of
dose dependency, Gort & Imlay (1998) showed that SoxS
induction saturated above 150 mM MV in a 45 min assay. Fig. 2. Induction of single-copy soxS9 : : lacZ by MV in E. coli
We therefore used 100 mM MV, a concentration employed cells expressing G6PDH. Bacteria containing a chromosomal
by several authors (Chander et al., 2003; Liochev et al., soxS9 : : lacZ fusion (strain B247), and transformed with either
1994). pSU18 ($) or pSUG6PDH (#), were incubated at 37 6C with
100 mM MV in the presence of 0?5 mM IPTG. Samples were
Cell-free extracts obtained from transformants overexpres- taken at intervals to assay for b-galactosidase activity. Each
sing a plasmid-borne G6PDH displayed specific activities data point represents the mean of determinations from four
of 1?62 units (mg total soluble protein)21, as compared to independent experiments with SE¡10 % of the mean. Inset,
0?09 units mg21 in siblings carrying the supporting vector maximal levels of soxS9 : : lacZ induction in bacteria transformed
pSU18. Higher G6PDH activities correlated with an with pSU18 (black bars), pSUG6PDH or pSUFPR (grey bars).
approximately twofold increase in the NADPH/NADP+ Experimental conditions were as above, and b-galactosidase
ratio (from ~2?3 to ~4?3). Both ratios declined to about activities were determined at 90 min after MV challenge.
50 % after a 90 min challenge with 100 mM MV, but the Induction levels of the overexpressing strains were normalized
difference between cells transformed with pSUG6PDH or to those attained by cells transformed with the supporting plas-
pSU18 was maintained. The soxRS response proceeded mid, which were taken as 1.
Downloaded from www.microbiologyresearch.org by
http://mic.sgmjournals.org 1123
IP: 45.225.232.158
On: Mon, 04 Dec 2017 15:11:38
M. Giró, N. Carrillo and A. R. Krapp

Thus, G6PDH and FPR did exert opposite effects on soxRS Gardner & Fridovich, 1991a, b; Kennedy et al., 1983), and
induction. It is conceivable that rapid accumulation of the since they are involved in many important metabolic path-
zwf product after exposure to superoxide could eventually ways, prompt repair of the iron–sulfur clusters is essential if
slow down the protective response to suboptimal levels, the cells are to survive. Indeed, their importance can be
while the subsequent expression of FPR (and Fld) will gauged by considering that oxidant-resistant fumarase and
counterbalance this effect by reoxidizing the pyridine aconitase are induced as part of the soxRS regulon to com-
nucleotide pool and preventing further increases in pensate for the loss of the corresponding activities during
NADPH level. the oxidative stress situation (Liochev & Fridovich, 1992;
Pomposiello & Demple, 2001; Varghese et al., 2003).
Overexpression of G6PDH does not increase
the MV tolerance of E. coli Assembly of iron–sulfur centres shows a strict requirement
for Fd (a conspicuous FPR substrate), but the role of this
E. coli cells deficient in either G6PDH or FPR displayed protein in cluster repair is less clear (Djaman et al., 2004).
enhanced sensitivity to various sources of oxidative stress, The possible contribution of isofunctional Fld to this
when compared to isogenic strains harbouring functional process has not been evaluated. To determine if FPR could
versions of the corresponding genes (Bianchi et al., 1995; be the source of reducing power during cluster repair, we
Greenberg et al., 1990; Krapp et al., 1997, 2002; Lundberg et al., designed a reconstituted electron-transport system in which
1999). On the other hand, FPR build-up in an otherwise the oxidation of glucose 6-phosphate was coupled via
wild-type background led to increased cell survival, indicat- G6PDH and FPR to reduction of Fd or Fld (Fig. 3a).
ing that the protective effects of this reductase were dose- Exposure of E. coli extracts to air caused rapid inactivation
dependent even beyond endogenous levels of expression and of typical hydro-lyases such as aconitase and 6GPD. Sub-
induction (Bianchi et al., 1995; Krapp et al., 1997, 2002). sequent incubation of the extract with the reconstituted
In contrast, E. coli cells transformed with pSUG6PDH were system in the presence of L-cysteine led to recovery of the
as resistant as their wild-type siblings when spotted on either two activities within 30 min (Fig. 3b, c). Controls lacking
minimal or rich media supplemented with MV, as described
in Methods (data not shown). To rule out the possibility that
the lack of G6PDH effect could be caused by deficiencies in
the expression system, the use of a plasmid-borne gene or
other shortcomings of the experimental set-up, we also
evaluated tolerance by a disk diffusion method, and included
an E. coli strain that overproduces G6PDH from a chromo-
somal zwf gene with activating mutations in the promoter
region (Fraenkel & Banerjee, 1971; Fraenkel & Parola, 1972).
The G6PDH-deficient mutant was still abnormally suscep-
tible to MV toxicity in this system, but the overproducers
failed to display increased tolerance, irrespective of whether
the enzyme was expressed from the plasmid or the chromo-
some (data not shown). These results suggest that the
maximum protective effects of G6PDH (or its relevant pro-
ducts, i.e. NADPH) are attained during the soxRS response
in stressed E. coli cells, and that any further increase is
inconsequential in terms of tolerance.

G6PDH and FPR can activate oxidant-sensitive

dehydratases in vitro
The delayed appearance of FPR and the distinctive features
of its contribution to the defensive system against oxidative
Fig. 3. In vitro repair of iron–sulfur centres in oxidized hydro-
stress argue against involvement of this reductase in ROS lyases. Dehydratases present in E. coli extracts were inactivated
avoidance and detoxification, and suggest that its protective by exposure to air and subsequently reactivated by a reconstituted
role might be related to repair activities required at later electron-transport chain containing G6PDH and FPR (a). G6P,
stages of the adaptive response, once the oxidative damage glucose 6-phosphate; 6PG, 6-phosphogluconate. Aconitase (b)
has occurred. Among the oxido-reductive pathways that and 6PGD (c) activities were determined after 30 min incuba-
could benefit from the surplus of low-potential electron tion in the complete system containing either Fd or Fld, or in
carriers generated during the soxRS response, reactivation mixtures lacking Fd/Fld, FPR, G6PDH or cysteine (Cys).
of oxidatively damaged hydro-lyases containing [4Fe–4S]2+ Controls (‘None’) were incubated in 50 mM Tris/HCl, pH 7?6,
clusters is a likely candidate. These enzymes are rapidly for the same time. Values reported are the mean±SE of three
inactivated by superoxide and other ROS (Flint et al., 1993; independent experiments.
Downloaded from www.microbiologyresearch.org by
1124 Microbiology 152
IP: 45.225.232.158
On: Mon, 04 Dec 2017 15:11:38
 G6PDH and FPR contribution to the soxRS response

Fd, G6PDH, FPR or cysteine failed to significantly reacti-

vate the enzymes in the time-frame of the experiment, while
addition of Fld instead of Fd resulted in similar levels of
recovery (Fig. 3b, c). Fld is not associated with the group of
genes involved in assembly of iron–sulfur centres to which
Fd belongs (Djaman et al., 2004). Conversely, Fld (but
not Fd) is induced during episodes of oxidative stress
(Pomposiello & Demple, 2001; Zheng et al., 1999).

Cysteine could be replaced in the reconstituted system by

other thiols, including glutathione and dithiothreitol, but
with lower efficiency (data not shown). The strict require-
ment of thiol groups is intriguing, since modified clusters
that are still amenable to rapid reactivation are not expected
to disintegrate beyond the [3Fe–4S]1+ state (Djaman et al.,
2004). However, the dependence of hydro-lyase reactivation
on the addition of thiols is well documented, and both
oxidized and nitrosylated clusters are quickly repaired
in vitro when incubated with iron and SH-containing com-
pounds (Kennedy et al., 1983; Schwarz et al., 2000; Varghese
et al., 2003; Yang et al., 2002). In our hands, however,
incorporation of ferrous ions into the reactivation assay had
little effect on activity recovery (data not shown), indicating
that the iron present in the bacterial extracts is probably
recycled for reconstitution of the functional clusters. The
preference for cysteine also suggests that cysteine desulfurase
encoded by the iscS gene (and eventually other members of
the isc operon) could be involved in repair, but further
research will be necessary to properly address this question.

Involvement of FPR in reactivation of hydro-

lyases in vivo Fig. 4. Oxidative damage and repair of hydro-lyases in bacteria
overexpressing FPR. E. coli GC4468 cells, transformed with
To investigate if the iron–sulfur cluster repair activities pSU18 ($) or pSUFPR (#), were exposed to 100 mM MV for
displayed by FPR in the reconstituted system have physio- 30 min, rinsed and further incubated in gluconate medium sup-
logical consequences, E. coli cells expressing various levels of plemented with 200 mg kanamycin ml”1. Aliquots were removed
the reductase were exposed to MV, and then shifted to an at the indicated times and assayed for aconitase (a) and 6PGD
MV-free medium to monitor dehydratase reactivation. (b) activities as described in Methods. Data points represent
Kanamycin or chloramphenicol was used to prevent new the mean of determinations from three independent experiments
protein synthesis during the periods of inactivation and with SE¡15 % of the mean. (c) FPR accumulation in cells
recovery. When wild-type cells were subjected to a 30 min transformed with pSU18 or pSUFPR after MV-dependent
challenge with the superoxide propagator, aconitase activ- induction. Samples were collected at the times indicated as
ities declined to ~10 % of their initial values, while 6PGD ”30 or 0 min in (a) and (b), lysed and processed essentially as
dropped to less than 5 % of its pre-treatment activity described in Fig. 1. Cleared extracts corresponding to 15 mg
(Fig. 4a, b). Removal of the reagent resulted in time- total soluble protein were loaded for immunoblot detection of
dependent recovery of both activities to about 50 % of their FPR.
original levels (Fig. 4). Exposure of fpr mutants to the same
treatment led to a similar inactivation, whereas recovery
was slightly delayed in this strain, relative to the wild-type small fraction of 6PGD were inactive under aerobic growth
cells (data not shown). These results essentially reproduced conditions.
those obtained by Krapp et al. (2002) and Djaman et al.
(2004), who considered the observed differences between
Concluding remarks
the strains of little significance, and suggested that the
contribution of FPR to iron–sulfur cluster repair was thus G6PDH is rapidly induced at the onset of the soxRS response
marginal. However, overexpression of a plasmid-encoded (Fig. 1), leading to accumulation of NADPH that could be
reductase dramatically accelerated activity recovery to levels used by scavenging enzymes such as glutathione reduc-
beyond those present in unstressed bacteria (Fig. 4), tase and thioredoxin reductase. As an electron donor for
indicating that a significant part of the aconitase and a SoxR reductase (Koo et al., 2003) and flavin reductase
Downloaded from www.microbiologyresearch.org by
http://mic.sgmjournals.org 1125
IP: 45.225.232.158
On: Mon, 04 Dec 2017 15:11:38
M. Giró, N. Carrillo and A. R. Krapp

(Woodmansee & Imlay, 2002), NADPH may have a pro- G6PDH and FPR is not a dedicated pathway, since both
oxidant effect by limiting the soxRS response and fostering reduced Fld and NADPH can be supplied by alternative
Fenton-type reactions. Delayed expression of FPR could in sources. Although pyruvate-ferredoxin(flavodoxin) reduc-
principle counterbalance these effects through its NADPH- tase is apparently not regulated by oxidants, its involvement
consuming activity (Krapp et al., 1997, 2002). It is not clear in the protection against oxidative stress deserves further
whether the differential time-courses displayed by the two investigation.
enzymes are important for the correct deployment of the
soxRS response, or merely reflect different sensitivities of the
corresponding promoters to SoxS build-up. Rapid FPR ACKNOWLEDGEMENTS
accumulation could lead to overoxidation of the NADP(H)
The authors wish to thank Dr Alejandro Viale (IBR, Argentina) for
pool (Krapp et al., 2002), removing NADPH necessary critical reading of the manuscript and many helpful suggestions. This
for scavenging enzymes required at an early stage of the work was supported by grant PICT-07853 from the National Science
response, whereas the flavoprotein is needed later during the Agency (ANPCyT, Argentina). N. C. and A. R. K. are staff members of
stress situation, once the damage has been done and repair the National Research Council (CONICET, Argentina).
mechanisms are invoked. Our results therefore indicate that
NADPH increases rapidly during the soxRS response despite
the activities of FPR and other reductases until the regulon is REFERENCES
self-restrained by negative feed-back regulation, presumably
Alekshun, M. N. & Levy, S. B. (1997). Regulation of chromosomally
through the activity of SoxR reductase (Fig. 2). G6PDH
mediated multiple antibiotic resistance: the mar regulon. Antimicrob
plays a dominant protective role during this stage, but Agents Chemother 41, 2067–2075.
further increases have little effect, probably because optimal
Bianchi, V., Haggard-Ljungquist, E., Pontis, E. & Reichard, P.
NADPH levels have already been reached. The role of FPR (1995). Interruption of the ferredoxin(flavodoxin) NADP+ oxido-
appears to be more related to the reduction of low-potential reductase gene of Escherichia coli does not affect anaerobic growth
electron carriers (Fd and Fld), rather than NADPH oxida- but increases sensitivity to paraquat. J Bacteriol 177, 4528–4531.
tion. The flavoenzyme was able to act as electron donor for Blaschkowski, H. P., Neuer, G., Ludwig-Festl, M. & Knappe, J.
Fd- or Fld-dependent reactivation of oxidatively damaged (1982). Routes of flavodoxin and ferredoxin reduction in Escherichia
hydro-lyases in vitro (Fig. 3) and in vivo (Fig. 4). However, coli. CoA-acylating pyruvate : flavodoxin and NADPH : flavodoxin
the contribution of FPR to this repair mechanism only oxidoreductases participating in the activation of pyruvate formate-
lyase. Eur J Biochem 123, 563–569.
becomes evident when the reductase accumulates over
wild-type levels of expression and induction, whereas its Brumaghim, J. L., Li, Y., Henle, E. & Linn, S. (2003). Effects of
hydrogen peroxide upon nicotinamide nucleotide metabolism in
participation during the normal soxRS response elicited by Escherichia coli. Changes in enzyme levels and nicotinamide nucleo-
sublethal concentrations of MV appears to be relatively tide pools and studies on the oxidation of NAD(P)H by Fe(III). J Biol
minor. The reasons for this behaviour are not entirely clear, Chem 278, 42495–42504.
but they may be related to the presence of alternative routes Carrillo, N. & Ceccarelli, E. A. (2003). Open questions in ferredoxin-
for Fd and Fld reduction in E. coli, which might partially or NADP+ reductase catalytic mechanism. Eur J Biochem 270, 1900–
totally compensate the FPR deficiency. One such pathway 1915.
has been already identified; it involves pyruvate-ferredoxin- Ceccarelli, E. A., Arakaki, A. K., Cortez, N. & Carrillo, N. (2004).
(flavodoxin) reductase (Blaschkowski et al., 1982). The role Functional plasticity and catalytic efficiency in plant and bacterial
of FPR in the provision of low-potential electron carriers for ferredoxin-NADP(H) reductases. Biochim Biophys Acta 1698, 155–165.
damage repair during the soxRS response could therefore be Chander, M., Raducha-Grace, L. & Demple, B. (2003). Transcription-
masked in experiments involving FPR-deficient mutants by defective soxRS mutants of Escherichia coli: isolation and in vivo
characterization. J Bacteriol 185, 2441–2454.
the partly overlapping activities of the pyruvate-dependent
enzyme. Similarly, anaerobic pathways that require reduced Csonka, L. & Fraenkel, D. G. (1977). Pathways of NADPH formation
in Escherichia coli. J Biol Chem 152, 3382–3391.
Fld for the activity of pyruvate-formate lyase, ribonucleotide
reductase and other key enzymes can proceed in fpr null Ding, H. & Demple, B. (1997). In vivo kinetics of a redox-regulated
transcriptional switch. Proc Natl Acad Sci U S A 94, 8445–8449.
mutants (Bianchi et al., 1995; Krapp et al., 1997).
Djaman, O., Outten, F. & Imlay, J. A. (2004). Repair of oxidized iron–
Blaschkowski et al. (1982) determined the kinetic properties sulfur clusters in Escherichia coli. J Biol Chem 279, 44590–44599.
of the two reductases, and showed that they display similar Flint, D. H., Tuminello, J. F. & Emptage, M. H. (1993). The inactiva-
catalytic efficiencies (kcat/Km values of 0?01–0?02 mM21 s21, tion of Fe–S cluster containing hydro-lyases by superoxide. J Biol
Chem 268, 22369–22376.
based on Fld affinity). However, the Km for pyruvate is
1?6 mM (Blaschkowski et al., 1982), compared to less than Fraenkel, D. G. (1968). Selection of Escherichia coli mutants lacking
glucose 6-phosphate dehydrogenase or gluconate 6-phosphate dehydro-
10 mM for NADPH (Carrillo & Ceccarelli, 2003). The actual
genase. J Bacteriol 95, 1267–1271.
performance of the enzymes as providers of reduced Fld
Fraenkel, D. G. (1987). Glycolysis, pentose phosphate pathway, and
and/or Fd in vivo will therefore depend on the concentra- Entner-Doudoroff pathway. In Escherichia coli and Salmonella
tions of electron donors and the enzyme levels, especially typhimurium: Cellular and Molecular Biology, pp. 142–150. Edited
after FPR induction under oxidative stress. The results thus by F. C. Neidhart and others. Washington, DC: American Society for
suggest that the electron-transport system established by Microbiology.
Downloaded from www.microbiologyresearch.org by
1126 Microbiology 152
IP: 45.225.232.158
On: Mon, 04 Dec 2017 15:11:38
 G6PDH and FPR contribution to the soxRS response

Fraenkel, D. G. & Banerjee, S. (1971). A mutation increasing the typhimurium virulence and resistance to reactive oxygen and nitrogen
amount of a constitutive enzyme in Escherichia coli, glucose 6-phosphate intermediates. Infect Immun 65, 5371–5375.
dehydrogenase. J Mol Biol 56, 183–194. Martin, R. G., Gillette, W. K. & Rosner, J. L. (2000). Promoter
Fraenkel, D. G. & Parola, A. (1972). ‘‘Up-promoter’’ mutations of discrimination by the related transcriptional activators MarA and
glucose 6-phosphate dehydrogenase in Escherichia coli. J Mol Biol 71, SoxS: differential regulation by differential binding. Mol Microbiol
107–111. 35, 623–634.
Gardner, P. R. & Fridovich, L. (1991a). Superoxide sensitivity of the Martinez, E., Bartolomé, B. & de la Cruz, F. (1988). pACYC184-
Escherichia coli 6-phosphogluconate dehydratase. J Biol Chem 266, derived cloning vectors containing the multiple cloning site and
1478–1483. lacZa reporter gene of pUC8/9 and pUC18/19 plasmids. Gene 68,
Gardner, P. R. & Fridovich, L. (1991b). Superoxide sensitivity of the 159–162.
Escherichia coli aconitase. J Biol Chem 266, 19328–19333. Miller, J. H. (1992). A Short Course in Bacterial Genetics: a Laboratory
Gardner, P. R. & Fridovich, L. (1993). NADPH inhibits transcription Manual for E. coli and Related Bacteria. Cold Spring Harbor, NY:
of the Escherichia coli manganese superoxide dismutase gene (sodA) Cold Spring Harbor Laboratory.
in vitro. J Biol Chem 268, 12958–12963. Nogae, I. & Johnston, M. (1990). Isolation and characterization of
Gaudu, P., Moon, N. & Weiss, B. (1997). Regulation of the soxRS the ZWF1 gene encoding glucose 6-phosphate dehydrogenase of
oxidative stress regulon. J Biol Chem 272, 5082–5086. Saccharomyces cerevisiae. Gene 96, 161–169.
Gaudu, P., Dubrac, S. & Touati, D. (2000). Activation of SoxR by Nunoshiba, T., Hidalgo, E., Amabile Cuevas, C. F. & Demple, B.
overproduction of desulfoferrodoxin: multiple ways to induce the (1992). Two-stage control of an oxidative stress regulon: the
soxRS regulon. J Bacteriol 182, 1761–1763. Escherichia coli SoxR protein triggers redox-inducible expression of
the soxS regulatory gene. J Bacteriol 174, 6054–6060.
Gort, A. S. & Imlay, J. A. (1998). Balance between endogenous super-
oxide stress and antioxidant defenses. J Bacteriol 180, 1402–1410. Nunoshiba, T., de Rojas-Walker, T., Tannenbaum, S. R. & Demple,
B. (1995). Roles of nitric oxide in inducible resistance of Escherichia
Greenberg, J., Monach, P., Chou, J., Josepphy, D. & Demple, B.
coli to activated murine macrophages. Infect Immun 63, 794–798.
(1990). Positive control of a global antioxidant defense regulon
activated by superoxide-generating agents in Escherichia coli. Proc Pomposiello, P. J. & Demple, B. (2001). Redox-operated genetic
Natl Acad Sci U S A 87, 6181–6185. switches: the SoxR and OxyR transcription factors. Trends Biotechnol
19, 109–114.
Griffith, K. L. & Wolf, R. E. (2001). Systematic mutagenesis of the
DNA binding sites for SoxS in the Escherichia coli zwf and fpr Pomposiello, P. J., Bennik, M. H. & Demple, B. (2001). Genome-
promoters: identifying nucleotides required for DNA binding and wide transcriptional profiling of the Escherichia coli responses to
transcription activation. Mol Microbiol 40, 1141–1154. superoxide stress and sodium salicylate. J Bacteriol 183, 3890–3902.
Griffith, K. L., Shah, I. M. & Wolf, R. E. (2004). Proteolytic degradation Rowley, D. L., Pease, A. & Wolf, R. E. (1991). Genetic and physical
of Escherichia coli transcription activators SoxS and MarA as the analyses of the growth rate-dependent regulation of Escherichia coli
mechanism for reversing the induction of the superoxide (soxRS) zwf expression. J Bacteriol 173, 4660–4667.
and multiple antibiotic resistance (mar) regulons. Mol Microbiol 51, Rowley, D. L., Fawcet, W. P. & Wolf, R. E. (1992). Molecular
1801–1816. characterization of mutations affecting expression level and growth
Imlay, J. A. (2003). Pathways of oxidative damage. Annu Rev rate-dependent regulation of the Escherichia coli zwf gene. J Bacteriol
Microbiol 57, 395–418. 174, 623–626.
Kao, S. M. & Hassan, H. M. (1985). Biochemical characterization of Sambrook, J., Fritsch, E. & Maniatis, T. (1989). Molecular Cloning: a
a paraquat-tolerant mutant of Escherichia coli. J Biol Chem 260, Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring
10478–10481. Harbor Laboratory.
Kennedy, M. C., Emptage, M. H., Dreyer, J. L. & Beinert, H. (1983). Schwarz, C. J., Djaman, O., Imlay, J. A. & Kiley, P. J. (2000). The
The role of iron in the activation-inactivation of aconitase. J Biol cysteine desulfurase, IscS, has a major role in in vivo Fe–S cluster for-
Chem 258, 11098–11105. mation in Escherichia coli. Proc Natl Acad Sci U S A 97, 9009–9014.
Koo, M. S., Lee, J. H., Rah, S. Y., Yeo, W. S., Lee, J. W., Lee, K. L., Scott, M., Zuo, L., Lubin, B. & Chiu, D. (1991). NADPH, not glutathione,
Koh, Y. S., Kang, S. O. & Roe, J. H. (2003). A reducing system of the status modulates oxidant sensitivity in normal and glucose 6-phosphate
superoxide sensor SoxR in Escherichia coli. EMBO J 22, 2614–2622. dehydrogenase-deficient erythrocytes. Blood 77, 2059–2064.
Krapp, A. R., Tognetti, V. B., Carrillo, N. & Acevedo, A. (1997). The Sedmak, J. J. & Grossberg, S. E. (1977). A rapid, sensitive, and
role of ferredoxin-NADP+ reductase in the concerted cell defense versatile assay for protein using Coomassie brilliant blue G250. Anal
against oxidative damage. Studies using Escherichia coli mutants and Biochem 79, 544–552.
cloned plant genes. Eur J Biochem 249, 556–563. Tsaneva, I. R. & Weiss, B. (1990). soxR, a locus governing a
Krapp, A. R., Rodriguez, R. E., Poli, H. O., Paladini, D. H., Palatnik, superoxide response regulon in Escherichia coli K-12. J Bacteriol 172,
J. F. & Carrillo, N. (2002). The flavoenzyme ferredoxin(flavodoxin)- 4197–4205.
NADP(H) reductase modulates NADP(H) homeostasis during the Varghese, S., Tang, Y. & Imlay, J. A. (2003). Contrasting sensitivities
soxRS response of Escherichia coli. J Bacteriol 184, 1474–1480. of Escherichia coli aconitases A and B to oxidation and iron
Liochev, S. I. & Fridovich, L. (1992). Fumarase C, the stable fumarase depletion. J Bacteriol 185, 221–230.
of Escherichia coli, is controlled by the soxRS regulon. Proc Natl Acad Wan, J. T. & Jarrett, J. T. (2002). Electron acceptor specifity of
Sci U S A 89, 5892–5896. ferredoxin (flavodoxin) : NADP+ oxidoreductase from Escherichia
Liochev, S. I., Hausladen, A., Beyer, W. F., Jr & Fridovich, L. (1994). coli. Arch Biochem Biophys 406, 116–126.
NADPH-ferredoxin oxidoreductase acts as a paraquat diaphorase and is Wolf, R. E., Prather, D. M. & Shea, F. M. (1979). Growth rate-
a member of the soxRS regulon. Proc Natl Acad Sci U S A 91, 1328–1331. dependent alteration of 6-phosphogluconate dehydrogenase and
Lundberg, B., Wolf, R. E., Dinauer, M., Xu, Y. & Fang, F. (1999). glucose 6-phosphate dehydrogenase levels in Escherichia coli K-12.
Glucose 6-phosphate dehydrogenase is required for Salmonella J Bacteriol 139, 1093–1096.
Downloaded from www.microbiologyresearch.org by
http://mic.sgmjournals.org 1127
IP: 45.225.232.158
On: Mon, 04 Dec 2017 15:11:38
M. Giró, N. Carrillo and A. R. Krapp

Wood, T. I., Griffith, K. L., Fawcett, W. P., Jair, K. W., Schneider, T. D. Wu, J. & Weiss, B. (1992). Two-stage induction of the soxRS
& Wolf, R. E. (1999). Interdependence of the position and orientation (superoxide response) regulon of Escherichia coli. J Bacteriol 174,
of SoxS binding sites in the transcriptional activation of the class I 3915–3920.
subset of Escherichia coli superoxide-inducible promoters. Mol Yang, W., Rogers, P. & Ding, H. (2002). Repair of nitric oxide-
Microbiol 34, 414–430. modified ferredoxin [2Fe–2S] cluster by cysteine desulfurase (IscS).
Woodmansee, A. N. & Imlay, J. A. (2002). Reduced flavins promote J Biol Chem 277, 12868–12873.
oxidative DNA damage in non-respiring Escherichia coli by delivering Zheng, M., Doan, B., Schneider, T. D. & Storz, G. (1999). OxyR and
electrons to intracellular free iron. J Biol Chem 277, 34055–34066. SoxRS regulation of fur. J Bacteriol 181, 4639–4643.

Downloaded from www.microbiologyresearch.org by

1128 Microbiology 152
IP: 45.225.232.158
On: Mon, 04 Dec 2017 15:11:38