Oxidation state governs structural transitions in peroxiredoxin II that correlate with cell cycle arrest and recovery
Inactivation of eukaryotic 2-Cys peroxiredoxins (Prxs) by hyperoxidation has been proposed to promote accumulation of hydrogen peroxide (H2O2) for redox-dependent signaling events. We examined the oxidation and oligomeric states of PrxI and -II in epithelial cells during mitogenic signaling and in response to fluxes of H2O2. During normal mitogenic signaling, hyperoxidation of PrxI and -II was not detected. In contrast, H2O2-dependent cell cycle arrest was correlated with hyperoxidation of PrxII, which resulted in quantitative recruitment of approximately 66- and approximately 140-kD PrxII complexes into large filamentous oligomers. Expression of cyclin D1 and cell proliferation did not resume until PrxII-SO2H was reduced and native PrxII complexes were regenerated. Ectopic expression of PrxI or -II increased Prx-SO2H levels in response to oxidant exposure and failed to protect cells from arrest. We propose a model in which Prxs function as peroxide dosimeters in subcellular processes that involve redox cycling, with hyperoxidation controlling structural transitions that alert cells of perturbations in peroxide homeostasis.
THE JOURNAL OF CELL BIOLOGY
© The Rockefeller University Press $8.00
The Journal of Cell Biology, Vol. 175, No. 5, December 4, 2006 779–789
Reentry into the cell cycle requires integration of signals from
several redox-dependent processes (Burch and Heintz, 2005).
For example, production of hydrogen peroxide (H
) is required
for mitogenic signaling in response to EGF, bFGF, PDGF, and
thrombospondin2 (Gabbita et al., 2000; Karin and Shaulian,
2001). One mechanism by which H
acts in mitogenic signal-
ing is through the transient oxidation of cysteine residues pres-
ent in signaling targets such as the phosphatases protein tyrosine
phosphatase 1B and PTEN (phosphatase and tensin homologue
on chromosome 10), which regulate signaling through the extra-
cellular signal–related kinase (ERK) 1/2 and PI3-kinase–Akt
pathways, respectively (for review see Forman et al., 2004).
Given the prominent role of oxidants in cell cycle reentry,
the G0–G1 transition can be considered an oxidative phase of
the cell cycle, as suggested by a recent study on metabolic cy-
cles in yeast (Tu et al., 2005). However, although production of
in response to growth factors is required for cell cycle
reentry (Finkel, 2003), high levels of H
during the G0–G1
transition cause cell cycle arrest. In serum-stimulated mouse
lung epithelial cells, as in many other cell types (for review see
Schwartz and Assoian, 2001), signals from the ERK1/2 and
PI3-kinase–Akt pathways are integrated temporally at the level
of expression of cyclin D1 (Yuan et al., 2003, 2004; Burch et al.,
2004). Recently, we showed that pathways regulating expres-
sion of cyclin D1 are targeted by reactive oxygen species (ROS)
and reactive nitrogen species, resulting in cell cycle arrest (Yuan
et al., 2003, 2004; Burch et al., 2004). Arrest can be bypassed
by loading cells with catalase (Yuan et al., 2003), supporting the
notion that intracellular levels of H
represent one mecha-
nism for redox-dependent control of cell cycle progression.
Peroxiredoxins (Prxs) are a highly abundant family of
widely expressed antioxidant enzymes (for reviews see Wood
et al., 2003b; Immenschuh and Baumgart-Vogt, 2005; Rhee et al.,
2005). Because PrxI interacts with c-Abl (Wen and Van Etten,
1997) and c-Myc (Mu et al., 2002; Egler et al., 2005) and PrxII
modulates signaling through the PDGF receptor (Choi et al.,
2005), Prxs have emerged as important factors that link ROS
metabolism to redox-dependent signaling events. All Prxs use a
redox-active peroxidatic cysteine to attack peroxide substrates,
resulting in the formation of a cysteine sulfenic acid (Cys-SOH).
As is typical for 2-Cys Prxs, PrxI and -II are obligate
Oxidation state governs structural transitions
in peroxiredoxin II that correlate with cell cycle
arrest and recovery
Timothy J. Phalen,
Paula B. Deming,
Alan K. Howe,
Albert van der Vliet,
Thomas J. Jönsson,
Leslie B. Poole,
and Nicholas H. Heintz
Department of Pathology and
Department of Pharmacology, University of Vermont College of Medicine, Burlington, VT 05405
Center for Structural Biology, Wake Forest University School of Medicine, Winston-Salem, NC 27157
nactivation of eukaryotic 2-Cys peroxiredoxins (Prxs)
by hyperoxidation has been proposed to promote ac-
cumulation of hydrogen peroxide (H
) for redox-de-
pendent signaling events. We examined the oxidation
and oligomeric states of PrxI and -II in epithelial cells dur-
ing mitogenic signaling and in response to ﬂ uxes of H
During normal mitogenic signaling, hyperoxidation of
PrxI and -II was not detected. In contrast, H
cell cycle arrest was correlated with hyperoxidation
of PrxII, which resulted in quantitative recruitment of
66- and 140-kD PrxII complexes into large ﬁ lamen-
tous oligomers. Expression of cyclin D1 and cell prolifera-
tion did not resume until PrxII-SO
H was reduced and
native PrxII complexes were regenerated. Ectopic expres-
sion of PrxI or -II increased Prx-SO
H levels in response to
oxidant exposure and failed to protect cells from arrest.
We propose a model in which Prxs function as peroxide
dosimeters in subcellular processes that involve redox cy-
cling, with hyperoxidation controlling structural transitions
that alert cells of perturbations in peroxide homeostasis.
Correspondence to Nicholas H. Heintz: email@example.com
Abbreviations used in this paper: DNCB, 1-chloro-2,4-dinitrobenzene; ERK, ex-
tracellular signal–related kinase; GOx, glucose oxidase; GSH, glutathione;
HMC, high molecular mass complex; Prx, peroxiredoxin; ROS, reactive oxygen
species; Trx, thioredoxin; TrxR, thioredoxin reductase.
JCB • VOLUME 175 • NUMBER 5 • 2006 780
homodimers, and in these enzymes the Cys-SOH of the peroxi-
datic cysteine in one subunit is attacked by a resolving cysteine
in the neighboring subunit, resulting in an intersubunit disul de
bond. In mammalian cells, the intersubunit disul de is reduced
by thioredoxin (Trx), which is then regenerated by Trx reduc-
tase (TrxR) using reducing equivalents from NAD(P)H (Fig. 1).
Calcium concentration, pH, and oxidation state in uence the
assembly of 2-Cys Prx dimers into decamers, and decamers into
high molecular mass oligomers (for reviews see Wood et al.,
2003b; Immenschuh and Baumgart-Vogt, 2005; Rhee et al.,
2005). Recent work also provides evidence for a link between
structural transitions in the oligomeric state of Prxs and their
peroxidase and protein chaperone activities (Wood et al., 2003a;
Parsonage et al., 2005; Jang et al., 2006).
In contrast to prokaryotic homologues, eukaryotic 2-Cys
Prxs have a particularly interesting biochemical characteristic
in that they are readily inactivated by their own substrate, H
Because of a C-terminal domain that induces a kinetic pause in
the catalytic cycle, the peroxidatic cysteine of PrxI and -II is
susceptible to hyperoxidation, leading to the formation of
sul nic acid (Cys-SO
H), which cannot participate in disul de
bond formation with the resolving cysteine (Wood et al., 2003a).
Inactivation through hyperoxidation has been proposed to allow
to accumulate to substantial levels, thereby facilitating
redox-dependent signaling, a concept known as the “ oodgate”
hypothesis (Wood et al., 2003a). The fact that the sul nic acid
form of 2-Cys Prxs is not a terminal end product but can be
reduced in an ATP-dependent manner by sul nyl reductases,
such as sul redoxins (Biteau et al., 2003; Chang et al., 2004)
and p53-inducible sestrins (Budanov et al., 2004), suggests that
H may participate in regulatory signaling loops.
We tested the relevance of the oodgate hypothesis during
mitogenesis by investigating the connection between the oxida-
tive state of Prx isoforms and cell cycle entry and arrest. Our
studies indicate that widespread inactivation of PrxI and -II by
hyperoxidation is not a facet of normal mitogenic signaling.
Rather, examination of dose-dependent responses to uxes of
demonstrate that cell cycle arrest in response to oxidative
stress correlates with recruitment of PrxII-SO
H into cytoplas-
mic oligomers and that recovery of cell proliferation occurs af-
H is reduced. Unexpectedly, transient overexpression
of PrxI and -II led to increased levels of hyperoxidized Prxs in
response to oxidative stress and failed to protect cells from ar-
rest. We propose that Prx-SO
H functions in stress response
pathways that warn cells of perturbations in oxidant metabolism
and thereby contribute to oxidant-induced cell cycle arrest.
Effects of H
on mitogenic signaling
To examine the oxidation state of Prxs during mitogenic signal-
ing, mouse C10 lung epithelial cells were collected in G0 by
serum deprivation, and the formation of Prx-SO
H in response to
serum stimulation was assessed using an antibody speci c for
H was not detected above background levels
in cells stimulated for 15 min with medium containing serum
concentrations from 2 to 20%, a range that induces dose-depen-
dent induction of tyrosine phosphorylation (Fig. 2 A, lanes 2–5),
activation of the ERK1/2 and PI-3 kinase–Akt mitogenic sig-
naling pathways, and expression of cyclin D1 (Ranjan et al.,
2006). These results indicate that normal mitogenic signaling
does not require inactivation of Prxs by hyperoxidation, in
agreement with a recent report on the role of PrxII in PDGF
signaling (Choi et al., 2005).
To further explore Prx oxidation in cell cycle control, we
adopted an experimental paradigm that utilizes a dose-depen-
generating system to evoke transient cell cycle arrest
(Burch et al., 2004). C10 cells were synchronized in G0 by se-
rum deprivation and induced to reenter the cell cycle by adding
medium containing 10% FBS with or without glucose oxidase
(GOx). In complete medium with glucose and 10% FBS, GOx
caused the dose-dependent production of H
in a linear fash-
ion for at least 8 h (Fig. 2 B). For example, in complete medium,
5.0 mU/ml GOx generated 10 μM H
During the rst 6 h of serum stimulation, 1.0 or 2.5 mU/ml
GOx had little effect on the expression of cyclin D1, whereas
doses of 5.0 mU/ml or greater blocked expression of cyclin D1
(Fig. 2 C, lanes 7–9). In response to continuous exposure to 1.0
mU/ml GOx, the levels of activated ERK1/2 were similar to the
serum control, cyclin D1 was expressed, and hyperoxidized 2-
Cys Prxs were not observed (Fig. 2 C, lane 5), suggesting that
C10 cells are able to metabolize considerable amounts of exog-
during the G0–G1 transition without accumulating
hyperoxidized 2-Cys Prxs. At 2.5 mU/ml, levels of phospho-
ERK1/2 were unaffected, Prx-SO
H was barely detectable after
6 h of exposure, and cyclin D1 was expressed at nearly normal
levels. In contrast, at 5.0 mU/ml, hyperoxidized Prx-SO
cumulated to substantial levels and cyclin D1 was not expressed
(Fig. 2 C, lane 7). Concentrations of GOx ≥10.0 mU/ml in-
duced accumulation of hyperoxidized Prx-SO
H, caused hyper-
activation of ERK1/2, and blocked expression of cyclin D1
(Fig. 2 C, lanes 8 and 9).
Figure 1. The catalytic cycle of C-Cys eukaryotic Prxs. (A) When exposed
, the peroxidatic cysteine (S
H) of 2-Cys Prxs is oxidized to sulfenic
acid (Prx-SOH). Upon reaction with the resolving cysteine (S
H), a Prx
dimer with an intermolecular disulﬁ de bond is formed, which is then re-
duced by Trx to regenerate active enzyme. Because of a pause in the cata-
lytic cycle, the S
H of eukaryotic 2-Cys Prxs is susceptible to hyperoxidation,
resulting in the formation of a sulﬁ nic acid form (Prx-SO
H) that is catalyti-
cally inactive. Sulﬁ redoxins and sestrins are ATP-dependent sulﬁ nyl reduct-
ases that participate in retroreduction of Prx-SO
H, regenerating active
enzyme. 2-Cys Prxs are obligate homodimers that can assemble into
decamers and higher molecular mass oligomers, depending on oxidation
state, pH, calcium concentrations, and posttranslational modiﬁ cations such
PEROXIREDOXINS AND REDOX SIGNALING • PHALEN ET AL. 781
We previously showed that termination of ERK1/2 signal-
ing after 3 h of exposure to the highest dose of GOx (15 mU/ml)
restores expression of cyclin D1 but not cell proliferation
(Burch et al., 2004). Hence, prolonged activation of ERK1/2 is
a useful marker of oxidant-induced arrest at the G0–G1 transi-
tion of the cell cycle. Although GOx in uenced the levels of
phospho-ERK1/2 in a dose-dependent manner as before, it did
not induce phosphorylation of JNK in synchronized cells at any
dose (Fig. 2 C, lanes 5–9). In asynchronous cells, activation of
JNK in C10 cells by H
is associated with cell death (Pantano
et al., 2003).
To determine if retroreduction of Prx-SO
H prevented the
accumulation of Prx-SO
H, serum-stimulated cells were treated
with 1-chloro-2,4-dinitrobenzene (DNCB), with or with out
GOx. DNCB depletes cells of reduced glutathione (GSH) and
blocks reduction of Trx by inhibiting TrxR (Arner et al., 1995),
thereby impairing the ability of Trx and GSH to participate in
the retroreduction of Prx-SO
H to catalytically active forms.
Within 10 min, 5 μM DNCB caused a 90% reduction in GSH
levels that persisted for at least 3 h (unpublished data).
In the absence of GOx, DNCB blocked the ability of
serum to induce expression of cyclin D1 but did not prevent
phosphorylation of ERK1/2 (Fig. 2 C, lane 4) or cause the accu-
mulation of hyperoxidized Prxs. In contrast, DNCB markedly
sensitized 2-Cys Prxs to hyperoxidation by GOx (Fig. 2 C,
compare lanes 6–9 with lanes 10–13), suggesting that Prx retro-
reduction pathways are active during cell cycle reentry. Al-
though phospho-ERK1/2 levels were increased in cells treated
with GOx and enhanced in cells treated with DNCB and GOx,
only with DNCB were high concentrations of GOx able to
induce phosphorylation of JNK (Fig. 2 C, lanes 10–13).
Cell proliferation was then examined in serum-stimulated
cells treated with DNCB and/or GOx (Fig. 2 D). GOx and/or
DNCB were added to serum-stimulated cells, and proliferation
was examined over a 3-d period without changing the culture
media. C10 cells exposed to 1.0 or 2.5 mU/ml GOx proliferated
as well as untreated controls, whereas those exposed to doses of
GOx ≥5.0 mU/ml failed to proliferate by 3 d (Fig. 2 C). Greater
than 70% of cells arrested in response to all but the highest dose
of GOx (15.0 mU/ml) remained viable for at least 3 d (Fig. 2 D
and not depicted). Caspase 3 was not activated in serum-stimu-
lated cells at any dose of GOx, although it was readily activated
after exposure to GOx by staurosporin (unpublished data), indi-
cating that proapoptotic pathways were functional in arrested
C10 cells. Cells treated with DNCB alone recovered slowly
(Fig. 2 D), whereas cells treated with DNCB and any dose of
GOx did not proliferate (not depicted).
Although DNCB sensitized Prxs to hyperoxidation by
GOx, it did not sensitize Prxs to hyperoxidation in response to
serum at any time point. Together, these studies indicate that
formation of Prx-SO
H may not be required for mitogenic sig-
naling during the G0–G1 transition of the cell cycle. In contrast,
dose-response experiments with GOx revealed a sharp transi-
tion from unimpeded cell proliferation to cell cycle arrest that
occurred between concentrations of 2.5 and 5.0 mU/ml, and
that arrest was re ected in failure to express cyclin D1.
Oxidation of PrxI and -II and cell
Transitions between dimers, decamers, and high molecular
mass oligomers of Prxs are governed by oxidation state (Wood
et al., 2002; Moon et al., 2005), phosphorylation during G2/M
Figure 2. Hyperoxidation of Prxs in serum-
stimulated cells correlates with inhibition of cell
proliferation. (A) C10 cells synchronized by se-
rum deprivation were stimulated with medium
containing the indicated concentration of FBS
for 15 min, and levels of phosphotyrosine and
H were assessed by immunoblotting.
Actin was used as a loading control. (B) For
generating ﬂ uxes of H
, GOx was added to
DME with 10% FBS at the indicated concentra-
tion, and the amount of H
in medium was
measured as a function of time. (C) Serum-
starved mouse lung epithelial (C10) cells (time
0) were stimulated with DME containing 10%
FBS and the indicated concentrations of GOx
(mU/ml), with or without 5 μM DNCB, for 6 h.
Cell extracts were examined for expression of
the indicated proteins by immunoblotting. PrxI
and -II comigrate under reducing conditions on
SDS gels; based on molecular mass, the band
at 26 kD is mitochondrial PrxIII. (D) The indi-
cated concentrations of GOx were added to
culture medium during serum stimulation, and
replicate C10 cultures were counted over a 3-d
period to assess cell proliferation. Control cul-
tures (time 0) were maintained in DME with
0.5% FBS for the duration of the experiment or
stimulated with 10% FBS alone. Error bars indi-
cate mean ± SD.
JCB • VOLUME 175 • NUMBER 5 • 2006 782
(Chang et al., 2002; Jang et al., 2006), and other parameters (for
review see Wood et al., 2003b). To study the oxidation state of
2-Cys Prxs under various conditions, an immunoblotting
method was devised to detect the relative amounts of reduced or
oxidized Prx (Prx-SH, Prx-SOH, or Prx-S-S-Prx) versus hyper-
oxidized Prx (Prx-SO
H). With this method, it was possible to
estimate the fraction of catalytically active PrxI and -II despite
the limitation that the Prx-SO
H antibody recognizes hyperoxi-
dized PrxI and -II with equivalent ef ciency.
When extracts were resolved by standard SDS-PAGE, to-
tal PrxI and -II levels detected by immunoblotting and quanti-
ed by densitometry varied less than ±8% during the rst 6 h
after serum stimulation, with or without GOx (Fig. 3). When
probed rst for Prx-SO
H and then for either PrxI or -II after
stripping the membrane, immunoblotting produced reciprocal
signals that re ected the fraction of PrxI or -II that was not cata-
lytically inactivated versus the fraction that was inactivated by
hyperoxidation. Using densitometry, the levels of reduced/oxi-
dized PrxI (Fig. 4 A), reduced/oxidized PrxII (Fig. 4 B), and
H (Fig. 4 C) were estimated as a function of GOx con-
centration after 3 h of exposure and after 3 h of recovery in fresh
medium (Fig. 3). At 2.5 mU GOx/ml, >85% of PrxI was hyper-
oxidized after a 3-h exposure (Fig. 3, lane 6). After recovery,
<50% of PrxI was hyperoxidized, and the reduction in Prx-
H levels (Fig. 4 C) was accompanied by recovery of the sig-
nal for reduced PrxI (Fig. 3, lane 15; and Fig. 4 A), con rming
the activity of retroreduction pathways in C10 cells. PrxII ap-
peared to be less sensitive to hyperoxidation than PrxI; at 2.5
mU/ml GOx (Fig. 3, lane 6), only 25% of PrxII had been in-
activated by 3 h (Fig. 4 B). At 10 or 15 mU/ml, both PrxI and -II
were quantitatively hyperoxidized (Fig. 3 A, compare lanes 8
and 9 with lanes 17 and 18), and little signal for reduced PrxI
and -II was regained after a 3-h recovery period (Fig. 4, A and B).
In cells treated with GOx, expression of cyclin D1 was inversely
correlated with the levels of Prx-SO
H (Fig. 3).
To assess the relationship between Prx hyperoxidation
and cellular redox status, GSH levels were measured as a func-
tion of GOx concentration after exposure and recovery (Fig. 4 D).
A considerable drop in GSH levels was not observed at 3 h until
concentrations of GOx exceeded 5.0 mU/ml, and at all concen-
trations of GOx, GSH levels increased after recovery in fresh
medium (Fig. 4 D). These results agree well with a report that
shows PrxII is hyperoxidized in response to levels of H
do not inhibit the TrxR–Trx system or deplete cells of GSH
(Baty et al., 2005). Hence, cells treated with 5.0 mU/ml GOx
for 3 h that retained near normal levels of GSH underwent tran-
sient cell cycle arrest, whereas those treated with either 10 or
15 mU/ml GOx that accumulated hyperoxidized PrxI and -II that
could not be reduced after 3 h of recovery (Fig. 3), perhaps be-
cause of low GSH levels (Fig. 4 D), were not able to proliferate.
Serum stimulation engages PrxI and -II
in peroxide metabolism
When assessed under standard conditions, the total levels of
PrxI and -II did not change during the rst 6 h of serum stimula-
tion (Fig. 3). When samples were denatured in the presence of
SDS, but without reducing agents to preserve disul de bonds,
gel electrophoresis showed that both PrxI (Fig. 5, lane 1) and
PrxII (lane 7) from serum-starved cells were partitioned be-
tween 23-kD Prx-SH/Prx-SOH monomers and 38-kD Prx-S-S-
Prx homodimers. Upon addition of serum, the levels of PrxI
(Fig. 5, lanes 2–6) and Prx II (lanes 8–12) monomers decreased,
Figure 3. Dose-dependent hyperoxidation of PrxI and -II by ﬂ uxes of
hydrogen peroxide. Serum-starved C10 cells were stimulated with DME
with 10% FBS containing the indicated concentration of GOx for 3 h. After
exposure, cells were washed and allowed to recover for 3 h in fresh me-
dium. Asynchronous cells plated at the same density in DME with 10% FBS
were used as controls. Cell extracts were examined for the expression of
cyclin D1, total PrxI, total PrxII, and Prx-SO
H by standard SDS-PAGE and
immunoblotting. Note that cyclin D1 is degraded in response to GOx in
asynchronous cells, whereas GOx did not affect the levels of total PrxI and
-II under any condition. To assess levels of PrxI and -II that were not hyper-
oxidized in serum-stimulated cells, immunoblots ﬁ rst probed for Prx-SO
were stripped and reprobed for PrxI or -II (see Materials and methods).
Figure 4. Prx oxidation occurs before depletion of cellular GSH. Using the
strip–reprobe immunoblotting method, densitometry was used to quantify
the signals for the fraction of PrxI (A) or PrxII (B) that was not hyperoxidized
as a function of GOx concentration after 3 h of exposure to the indicated
concentration of GOx (squares) or after a 3-h recovery period (diamonds).
(C) Total Prx-SO
H levels are expressed as a percentage of the maximal
signal obtained with 15 mU/ml GOx, which caused quantitative hyperoxi-
dation of PrxI and -II. (D) Reduced GSH levels were measured in extracts
from C10 cells treated with GOx and allowed to recover in the same fashion.
Error bars indicate mean ± SD.
PEROXIREDOXINS AND REDOX SIGNALING • PHALEN ET AL. 783
and PrxI and -II homodimers with intersubunit disul de bonds
increased (Fig. 5, lanes 2–6 and 8–12, respectively). After expo-
sure to 15 mU/ml GOx, all dimers with intersubunit disul de
bonds were lost by 30 min, and only hyperoxidized PrxI and -II
monomers were detected for the duration of the experiment
(Fig. 5, lanes 13–17; and not depicted). Because homodimers
with intersubunit disul de bonds are produced only during per-
oxide catalysis (Fig. 1), these results indicate that PrxI and -II
produced in response to serum stimulation.
Upon hyperoxidation, a condition in which intersubunit disul de
bonds cannot form, only Prx-SO
H monomers were observed,
Recruitment of hyperoxidized PrxII
into high molecular mass oligomers
At 2.5 mU/ml GOx, 85% of PrxI was hyperoxidized, and yet cells
expressed cyclin D1 and proliferated normally. In contrast, cells
treated with 5.0 mU/ml GOx did not express cyclin D1 or prolif-
erate. To better understand this difference, native gel electropho-
resis was used to examine the effect of GOx on the oligomerization
state of PrxI and -II. When cell extracts were resolved by electro-
phoresis in the absence of reducing agents and SDS, immunoblot-
ting indicated that PrxI was organized exclusively in complexes
>660 kD (unpublished data). In contrast, PrxII was detected in
two sets of bands that we refer to as A–A’ and B–B’ (Fig. 6).
Compared with the mobility of native molecular mass markers,
A–A’ migrated with an apparent molecular mass of 66 kD and
B–B’ with a mass of 140 kD. Although similar PrxII complexes
have been observed in other cell types (Moon et al., 2005), the
precise constituents of these complexes are not known.
In extracts of serum-starved cells, bands A and B were the
predominant form of PrxII (Fig. 6 A, lane 1). Addition of DNCB
or FBS alone for 3 h did not change the mobility of PrxII on na-
tive gels (Fig. 6 A, lanes 2 and 3), but together DNCB and FBS
increased the signal of band B’ (lane 4). Because FBS and
DNCB do not induce Prx hyperoxidation (Fig. 3), changes in
band B may re ect structural transitions during formation of
PrxII-S-S-PrxII dimers during peroxide metabolism (Fig. 5), in
Figure 5. Serum stimulation increases the levels of Prx-S-S-Prx dimers.
At the indicated times, extracts were prepared from serum-stimulated C10
cells in the absence of reducing agents and resolved by gel electrophoresis
in the presence of SDS. After transfer, blots were probed for PrxI or -II as in-
dicated. Extracts from serum-stimulated cultures treated with15 mU/ml
GOx were prepared in the same fashion and probed for Prx-SO
starved cells (time 0) were used as controls.
Figure 6. Hyperoxidation of PrxII induces
structural transitions that correlate with cell
cycle arrest. (A and B) Serum-starved C10
cells (time 0) were stimulated with medium
containing 10% FBS and the indicated con-
centration of GOx with or without 5 μM
DNCB. Cell extracts prepared in the ab-
sence of reducing agents were resolved by
native gel electrophoresis to preserve pro-
tein complexes. After transfer, immunoblots
ﬁ rst were probed for PrxII (A), and after
stripping, for Prx-SO
H (B). Complexes A–A’
and B–B’ migrated with apparent molecular
masses of 66 and 140 kD, respectively.
HMCs detected by the anti–Prx-SO
body in B migrated with apparent molecu-
lar masses >500 kD. To examine the
dynamics of these complexes in response to
oxidative stress, extracts were prepared at
the indicated times from cells stimulated
with FBS alone (C) or from cells stimulated
with FBS containing 2.5 mU/ml (D) or 5.0
mU/ml (E) GOx for the ﬁ rst 3 h of serum
stimulation. Samples were resolved under
reducing conditions for assessing total lev-
els of Prx-SO
H and cyclin D1 and native
conditions for visualizing HMCs and PrxII
complexes A–A’ and B–B’.
JCB • VOLUME 175 • NUMBER 5 • 2006 784
agreement with studies that show PrxII metabolizes H
produced in response to growth factors (Choi et al., 2005) and
-activated signaling by phospholipase D1 (Xiao
et al., 2005).
In response to exposure to 1.0 or 2.5 mU/ml GOx, band B’
increased in abundance relative to band B, perhaps re ecting
increased engagement of the PrxII 140-kD complex in peroxide
metabolism (Fig. 6 A, lanes 5 and 6). At concentrations of GOx
of 5.0 mU/ml or higher, bands B and B’ disappeared, band A
decreased, and band A’ appeared (Fig. 6 A, lanes 7–9). As ob-
served in Fig. 3, DNCB shifted the dose response for the A–A’
and B–B’ complexes to lower concentrations of GOx (Fig. 6 A,
When reprobed for Prx-SO
H, little hyperoxidized PrxII
was observed for cells treated with 1.0 mU/ml GOx (Fig. 6 B,
lane 5), whereas hyperoxidized Prx-SO
H was observed to comi-
grate with band B’ in extracts from cells treated with 2.5 mU/ml
GOx (Fig. 6 B, lane 6). At concentrations of GOx ≥5.0 mU/ml,
H was incorporated into several discrete high molecular
mass complexes (HMCs) with apparent molecular masses >500
kD and considerable levels of A’ accumulated (Fig. 6 B, lanes 7–9).
Recruitment of Prx-SO
H into HMCs correlated with loss of
signal from the PrxII B–B’complex (Fig. 6 A, lanes 7–9).
PrxII complexes accumulate
during cell proliferation
In time course experiments, the A–A’ and B–B’ complexes re-
sponded to serum stimulation and cell proliferation and, during
recovery from exposure, to 5.0 mU/ml GOx. The levels of the
140-kD B–B’ complex uctuated during the rst 12 h of serum
stimulation (Fig. 6 C, lanes 1–6) and increased markedly in
abundance as cells reached con uence 48–96 h later (lanes 8–10).
As cells reached con uence, increases in the A–A’ also
were observed (Fig. 6 C, lanes 8–10). Serum stimulation and
cell proliferation for >3 d caused no change in the signal for to-
H detected under reducing and denaturing conditions
H in HMCs detected by native gel electrophoresis
(Fig. 6 C, lanes 2–10). The PrxII complexes were largely unaf-
fected by exposing cells to 2.5 mU/ml GOx for the rst 3 h of
serum stimulation (Fig. 6 D, lanes 1–9), even though substantial
levels of Prx-SO
H were observed under these conditions (Fig.
6 D, lanes 1–4) and the cultures took slightly longer to reach
con uence. Note that 2.5 mU/ml GOx did not increase HMCs
At 5.0 mU/ml GOx, the B–B’ complex was not observed
during the 3-h exposure, HMCs containing Prx-SO
in abundance, and cyclin D1 was not expressed (Fig. 6 E, lanes
1 and 2). After GOx was removed at 3 h, total Prx-SO
were reduced over time, and Prx-SO
H in HMCs returned to
background levels (Fig. 6 E, lanes 3–9). As signal for Prx-SO
diminished in HMCs, A’ was lost, the B–B’ complex reappeared,
and cyclin D1 was expressed (Fig. 6 E, lanes 5–9). By 96 h, the
HMCs and PrxII A–A’ and B–B’ complexes observed by native
gel electrophoresis were identical in extracts from cells exposed
to all three conditions, even though proliferation to con uence
was delayed in cells treated with 5.0 mU/ml GOx (e.g., total
cellular protein at 72 h was 50% of the 10% FBS control).
Localization of hyperoxidized 2-Cys Prxs
Immuno uorescence confocal microscopy was used to localize
H within C10 cells treated with various doses of GOx.
In all cells, the Prx-SO
H antibody reacted with the cell nucleus,
but this signal did not correlate with the level of Prx hyperoxi-
dation detected by immunoblotting. In cells treated with 1.0
mU/ml GOx for 3 h, immunostaining was occasionally ob-
served in small patches at the cell periphery (Fig. 7 D), and this
pattern was more obvious in cells treated with 2.5 mU/ml GOx
(Fig. 7 E). At 5.0 mU/ml, GOx staining was observed in a la-
mentous pattern in the cell cytoplasm (Fig. 7 F). Prx-SO
cytoplasmic laments was particularly evident in cells treated
with 10.0 mU/ml GOx, and at 15 mU/ml GOx, staining was
prominent around the cell periphery (Fig. 7, G and H). At higher
doses of GOx, the peripheral Prx-SO
H staining pattern corre-
lated with changes in morphology that included a considerable
increase in cell diameter. A lamentous cytoplasmic staining
pattern for Prx-SO
H was not observed in asynchronous cells at
any dose of GOx (Fig. 7 I and not depicted).
Ectopic expression of HA-PrxI and -PrxII
and oxidant-induced arrest
Up-regulation of PrxI is thought to counteract the effects of en-
hanced oxidant production in tumor cells and thereby promote
cell survival and proliferation (Chang et al., 2005; Park et al., 2006).
Figure 7. Organization of cytoplasmic Prx-SO
H in response to oxidative
stress. C10 cells were plated on coverslips and synchronized in G0 by
serum deprivation. After serum stimulation for 3 h, with or without exposure
to the indicated concentration of GOx (mU/ml), cells were stained for Prx-
H and examined by confocal microscopy. Asynchronous C10 cells
treated with GOx were used as controls.
PEROXIREDOXINS AND REDOX SIGNALING • PHALEN ET AL. 785
To test the effects of Prx expression on responses to GOx, we
generated expression vectors for HA-tagged PrxI, PrxII, and
PrxII-∆C, a robust mutant of PrxII that is 100-fold less sensitive
to inactivation by H
(Koo et al., 2002; Wood et al., 2003a).
HA-PrxI interacts with endogenous PrxI in coimmunoprecipita-
tion experiments, and HA-PrxI and -PrxII are hyperoxidized in
response to GOx and reduced during recovery (unpublished
data), indicating that HA-tagged Prxs function in peroxide me-
tabolism in a manner similar to their endogenous counterparts.
C10 cells were rst transfected with expression constructs, and
24 h later the cultures were trypsinized and cells were plated at
identical cell densities and synchronized by serum deprivation
for 72 h as before. The transfected and serum-starved cell cul-
tures were then treated with 5.0 mU/ml GOx as before.
In synchronized cells, immunoblotting showed HA-PrxI
(Fig. 8 A, lanes 10–12) and HA-PrxII (lanes 13–15) were ex-
pressed at levels about fourfold that of their endogenous
counterparts. Because of addition of the HA epitope tag and de-
letion of the PrxII C-terminal domain, HA-PrxII-∆C comigrated
with endogenous PrxII. As compared with untransfected cells
(Fig. 8 A, lane 3) or vector controls (lane 6), expression of cata-
lase (lane 9) and the robust PrxII-∆C mutant (lane 18) reduced
but did not eliminate Prx-SO
H levels generated in response to
GOx during a 3-h exposure, with 3 h of recovery period as before.
HA-PrxI (Fig. 8 A, lane 12) and HA-PrxII (Fig. 8 A, lane 15)
were hyperoxidized under these conditions and thereby in-
creased the total cellular levels of Prx-SO
H as measured by
densitometry (Fig. 8 B). After recovery, expression of HA-PrxI
or -PrxII did not reduce the levels of phospho-ERK1/2 or pro-
mote expression of cyclin D1 (Fig. 8 A). Although cells express-
ing catalase (Fig. 8 A, lanes 7–9) or PrxII-∆C (Fig. 8 A, lanes
16–18) showed lower levels of total Prx-SO
H and pERK1/2
after recovery, cells had not expressed cyclin D1 or resumed
proliferation by this time. Expression of HA-PrxI or -PrxII did
not affect expression of cyclin D1 in response to serum alone
(Fig. 8 A, lanes 11 and 14). When cells treated with 5.0 mU/ml
GOx were examined after 72 h of recovery, cells expressing
HA-PrxI and -PrxII proliferated in a manner similar to vector
controls, whereas cells expressing PrxII-∆C resumed prolifera-
tion earlier during recovery (Fig. 8 C). Thus, as in serum-stimu-
lated cells, the accumulation of Prx-SO
H in cells overexpressing
PrxI or -II was correlated with delays in cell cycle progression
To con rm that HA-PrxII-∆C was cytoprotective, stable
cell lines were generated and treated with 5.0 mU GOx/ml con-
tinuously for 16 h. Flow cytometry showed that after 16 h 30%
of control cells exhibited a sub-G1 DNA content, whereas in
comparison, 10% of the cell population expressing HA-PrxII
was detected in the sub-G1 fraction. In contrast, <0.5% of cells
expressing PrxII-∆C were detected in the dead cell fraction
The susceptibility of 2-Cys Prxs to inactivation by hyperoxida-
tion is highly conserved in eukaryotes, inspiring the hypothesis
that the Prx inactivation loop evolved to support peroxide-
dependent signaling (Wood et al., 2003a). Here, we have exam-
ined the relationship between the oxidation state of PrxI and -II
and transition from G0 into G1, a portion of the cell cycle known
to respond to peroxide-dependent signaling (Finkel, 2003).
Based on the presence of homodimers containing intersubunit
Figure 8. Elevated expression of PrxI and -II
does not prevent cell cycle arrest in response to
oxidative stress. (A) C10 cells were transfected
with the indicated expression vectors or vector
control, synchronized by serum deprivation,
and treated with 5.0 mU/ml GOx for 3 h as in
Fig. 3. After 3 h of recovery in fresh medium,
cell lysates were assayed for Prx-SO
cyclin D1, pERK/12, and ERK1/2 levels using
standard immunoblotting or for active PrxI and
-II using the strip–reprobe blotting method.
Nontransfected (NT) cells were included as
controls for transfection. (B) Densitometry was
used to quantify the total Prx-SO
H signal for
each sample presented in A. (C) After 72 h,
cell numbers were determined to assess cell
proliferation in cultures transfected with the in-
dicated expression vectors. Error bars indicate
mean ± SD.
JCB • VOLUME 175 • NUMBER 5 • 2006 786
disul de bonds that are generated only during the Prx catalytic
cycle, both PrxI and -II appear to metabolize H
response to serum stimulation (Fig. 5), although the source of
, rate of catalysis, and sites of metabolism are unknown. In
cells treated with DNCB (Fig. 3), which depletes cells of GSH
and disrupts both Trx and GSH-dependent steps in the retrore-
duction cycle (Jeong et al., 2006), Prxs were sensitized to hy-
peroxidation by GOx. Nonetheless, in the presence of DNCB,
hyperoxidation of Prxs was not observed in serum-stimulated
cells at any time point, indicating that hyperoxidation of PrxI or -II
may not be required during mitogenic signaling, a result that is
in agreement with studies on the role of PrxII in PDGF signaling
(Choi et al., 2005). Rather, our studies suggest that oligomers of
hyperoxidized PrxII play a role in cell cycle arrest.
Hyperoxidized PrxI accumulated more rapidly in response
to exogenous uxes of H
than did hyperoxidized PrxII (Fig. 3),
but levels of PrxI-SO
H did not correlate with arrest. In contrast
to PrxI, cell cycle progression, arrest, and recovery were corre-
lated with changes in the oligomeric state of PrxII. As assessed
by native gel electrophoresis, PrxII existed in two complexes of
66 kD (A–A’) and 140 kD (B–B’). As cells proliferated to
con uence, both A–A’ and B–B’ increased in abundance and
B–B’ increased in complexity (Fig. 6 C). In con uent cells, the
B–B’ complex encompassed three distinct bands, suggesting
recruitment of additional factors as cells exited the cell cycle, a
matter presently under investigation.
In GOx dose-response experiments, C10 cells were able
to accumulate substantial levels of hyperoxidized PrxI or -II
during mitogenic signaling without marked effects on cell cycle
progression. For example, exposure to 2.5 mU/ml GOx resulted
in nearly complete hyperoxidation of PrxI and considerable lev-
els of hyperoxidized PrxII, yet C10 cells were able to express
cyclin D1 and proliferate. At these levels of exposure, GSH lev-
els were unaffected, and hyperoxidized PrxI and -II were read-
ily reduced once GOx was removed (Fig. 3). At levels of GOx
that induced transient cell cycle arrest upstream of cyclin D1,
but did not alter GSH levels, the B–B’ PrxII complexes disap-
peared and hyperoxidized PrxII appeared to be incorporated
into HMCs. When oxidative stress was terminated, Prx-SO
HMCs was readily reduced, the B–B’ complex reappeared, and
cells resumed expression of cyclin D1 and cell proliferation. In
contrast to the rate of hyperoxidation of PrxII seen in response
to GOx (Fig. 4 B), the dose-dependent structural transitions in
PrxII were abrupt (Fig. 6 B), suggesting a threshold effect for
delimiting choices between cell cycle progression and arrest.
Prx-dependent thresholds that regulate responses to increasing
doses of H
have been observed in yeast (Bozonet et al.,
2005; Vivancos et al., 2005).
Electron microscopy shows that in vitro PrxII decamers
are able to stack up on one another in an oblique fashion, form-
ing short laments (Harris et al., 2001). Immunostaining showed
H becomes organized in lamentous structures in
the cytoplasm of serum-stimulated C10 cells (Fig. 7). Because
this was not observed in asynchronous cells, recruitment of Prx-
H oligomers into cytoplasmic laments may be linked to a
process active in serum-stimulated cells, such as actin stress
ber formation, or re ect acquisition of chaperone function by
hyperoxidized PrxII (Moon et al., 2005). Linking the organiza-
tion of PrxII to actin stress ber formation is an attractive possi-
bility, for actin stress ber formation is a redox-dependent
process that regulates signaling through ERK1/2 and expression
of cyclin D1 (Roovers and Assoian, 2003).
Elevated expression of PrxI, PrxII, and robust mutants of
these enzymes has been shown to protect cells against oxidative
stress (Mu et al., 2002; Moon et al., 2004), but these studies
were not conducted in synchronized cells. In our previous stud-
ies, we have observed very different responses to oxidative
stress that depend on cell cycle position and cell density (Persinger
et al., 2001; Yuan et al., 2003; Burch et al., 2004; Ranjan and
Heintz, 2006). Differential sensitivity may be related to wiring
of MAPK pathways, for JNK is not activated by H
synchronized C10 cells (Fig. 2), whereas it is readily activated
in asynchronous C10 cells at levels that result in
hyperoxidation of <20% of PrxI (Pantano et al., 2003; unpub-
In synchronized cells, a fourfold increase in expression of
HA-PrxI and -PrxII relative to endogenous PrxI and -II did not
reduce the level of hyperoxidized endogenous PrxI or -II in re-
sponse to GOx but, rather, resulted in increased levels of total
H. Expression of HA-PrxI and -PrxII also did
not promote cell proliferation during recovery (Fig. 8 C).
Together with the GOx dose-response studies, these results in-
dicate that oligomers of hyperoxidized Prx-SO
H may be sensed
as an anti-mitogenic signal.
Although the propensity of eukaryotic 2-Cys Prxs to be
inactivated by H
may provide a “ oodgate” for permitting
to accumulate for redox-dependent signaling, our data
provide evidence for an additional hypothesis for the conserva-
tion of the inactivation shunt in mammals. Rather than simply
buffering intracellular peroxide, Prx enzymes may continuously
interpret and report peroxide levels, using their redox and oligo-
meric states as posttranslational modi cations to interface with
and modulate redox-sensitive cellular events (Fig. 9). Thus,
Prxs may serve as highly sensitive peroxide dosimeters that link
oxidant metabolism to a variety of redox-dependent processes
required for cell cycle reentry. Upon serum stimulation, these
enzymes become engaged in metabolizing H
response to activation of growth factor receptors, actin stress
ber formation, cell migration, and other processes. If oxidant
metabolism goes awry or the cell is exposed to threshold levels
of exogenous ROS, structural transitions regulated by hyperoxi-
dation would terminate the Prx catalytic cycle, thereby inter-
rupting interactions with regulatory factors or disrupting redox
cycling of other factors. Alternatively, PrxII-SO
may be sensed directly as an anti-mitogenic signal. Linking Prx
hyperoxidation to cell cycle progression would allow cells to
respond to perturbations in peroxide homeostasis well before
depletion of GSH or disruption of the TrxR–Trx system.
It is intriguing that p53 is activated by oxidative stress and
that downstream targets of p53 include factors that in uence
cellular redox state, including sestrins that regenerate Prx activity
(Budanov et al., 2004). Retroreduction of Prxs by sul nyl
reductases is a reasonable facet of stress responses only if resto-
ration of Prx activity contributes to recovery of activity in cell
PEROXIREDOXINS AND REDOX SIGNALING • PHALEN ET AL. 787
signaling pathways, for degradation by the proteasome would
be equally effective in ridding cells of hyperoxidized Prxs.
Hence, 2-Cys Prxs may provide an exquisitely sensitive, widely
distributed, and dose-dependent “smoke alarm” for alerting
cells to oxidative stress.
Materials and methods
Cell culture, cell cycle synchronization, and oxidant exposure
C10 mouse lung epithelial cells (Malkinson et al., 1997) were cultured,
synchronized, and stimulated with serum as described previously (Burch
et al., 2004). For oxidant exposures, recombinant GOx (Roche) in 10 mM
phosphate buffer, pH 7.4, was diluted in medium immediately before use.
Levels of H
generated by GOx in complete medium were determined
as described previously (van der Vliet et al., 1997). DNCB (Sigma-Aldrich)
was dissolved in DMSO and used at a ﬁ nal concentration of 5 μM.
For growth curves, cells were plated in duplicate in 6-well dishes and
treated as described (see Results) before trypsinization and counting
with a hemocytometer.
Cell extracts were prepared with NP-40 lysis buffer (150 mM NaCl, 1.0%
NP-40, 50 mM Tris, pH 8.0, 1 μg/ml leupeptin, 1 μg/ml aprotinin, 1 mM
NaF, 1 mM NaVO
, and 1 mM PMSF) or passive lysis buffer (Promega) as
noted. At harvest, cells in 60-mM dishes were washed once with cold PBS,
pH 7.4, 100 μl of lysis buffer were added, and lysates were collected by
scraping with a rubber policeman. The insoluble fraction was pelleted
by centrifugation in a microfuge for 5 min, and the protein concentra-
tion of the soluble fraction was determined using a protein assay
Electrophoresis and immunoblotting
For reducing SDS-PAGE, lysates were diluted 1:5 with 5× sample buffer
(10% SDS, 500 mM DTT, 300 mM Tris, pH 6.8, 0.05% bromophenol blue,
and 50% glycerol), heated at 95°C for 5 min before resolution on
12% SDS-PAGE gels. Nonreducing SDS-PAGE was performed in the same
manner, except that DTT was omitted from sample buffer. For native-PAGE,
NP-40 lysates were diluted in 5× sample buffer without DTT or SDS and
resolved on 8% polyacrylamide gels without SDS. Under all conditions,
proteins were transferred onto Immobilon-P PVDF (Millipore). Membranes
were blocked with 5% nonfat milk in TBS/T (25 mM Tris, pH 8, 150 mM
NaCl, and 0.1% Tween-20) and incubated with primary antibodies diluted
in 5% milk in TBS/T. Reactive proteins were visualized by HRP-conjugated
secondary antibodies (GE Healthcare) and chemiluminescence using West-
ern Lightning ECL (PerkinElmer).
Total levels of PrxI and -II were determined using SDS-PAGE and im-
munoblotting conditions as described previously (Burch et al., 2004). For
assessing relative levels of PrxI and -II that were not hyperoxidized, blots
were ﬁ rst probed for Prx-SO
H using anti-PrxSO
antibody (Lab Frontier)
and then stripped at 50°C for 15 min in 62.5 mM Tris, pH 6.8, 2% SDS,
and 100 mM β-mercaptoethanol. Stripped blots were washed several
times with TBS/T, blocked in 5% nonfat milk in TBS/T for 30 min, and
reprobed with anti-PrxI or anti-PrxII antibody. In contrast to probing for
H ﬁ rst, probing for PrxI or -II before stripping and reprobing with
H antibody did not inﬂ uence detection of Prx-SO
Antibodies for PrxI (LF-PA0001), PrxII (LF-PA0007), and Prx-SO
PA0004) were obtained from Lab Frontier. Antibodies to ERK1/2 (9102),
phospho-ERK1/2 (9101), and phospho-JNK (9251) were obtained from
Cell Signaling Technologies. Anti-phosphotyrosine mouse monoclonal
4G10 was purchased from Upstate Cell Signaling Solutions, anti–cyclin
D1 (sc-450) was purchased from Santa Cruz Biotechnology, Inc., and anti-
actin from Sigma-Aldrich. Mouse monoclonal anti-HA 12CA5 was a gift
from E. Harlow (Harvard University, Cambridge, MA).
Plasmid construction and transfection
Full-length coding sequences for human PrxI and -II were recovered with
BamHI ends from pET-17 (Novagen) vectors (Kang et al., 1998) using PCR
and the following primer sets: PrxI forward, 5′-cgcggatccatgtcttcaggaaatg-3′;
PrxI reverse, 5′-cgcggatcctcacttctgcttgg-3′; PrxII forward, 5′-cgcggatc-
catggcctccggtaacg-3′; PrxII reverse, 5′-gcgggatccctaattgtgtttggag-3′.
PrxII-∆C was generated from a previously described pET-19 (Novagen)
PrxII construct (Jönsson et al., 2005) by introducing a stop codon at D188
using the QuikChange site-mutagenesis kit (Stratagene) with the following
primers: forward, 5′-G A C A C G A T T A A G C C C A A C G T G T A G G A C A G C A A-
G G A A T A T T T C -3′; reverse, 5′-G A A A T A T T C C T T G C T G T C C T A C A C G T T
G G G-
C T T A A T C G T G T C -3′. PrxII-∆C was subcloned from the PrxII pET-19 vector
using the PrxII primers listed. PCR products were cloned ﬁ rst using topo-TA
cloning vector pCR2.1 (Invitrogen). Positive clones were digested with
BamHI, and fragments were subcloned into pCMV-HA to introduce an
N-terminal HA epitope tag. The pZeoSV-catalase expression vector (Arnold
et al., 2001) was a gift from D. Lambeth (Emory University, Atlanta, GA).
Expression constructs were propagated in DH5α cells and prepared for
transfection by alkaline lysis and sedimentation to equilibrium in CsCl.
Asynchronous C10 cells at 70% conﬂ uence in 60-mm plates were cotrans-
fected with Prx expression plasmids and an EGFP expression vector
(pEGFP-N2; CLONTECH Laboratories, Inc.) using Lipofectamine 2000 (In-
vitrogen) according to manufacturer’s protocols. Based on EGFP expres-
sion, transfection efﬁ ciency was routinely >70%.
C10 cells were lysed in 1% Triton, 50 mM Hepes, 250 mM NaCl, 10%
glycerol, 1.5 mM MgCl
, 1 mM PMSF, 1 mM EGTA, 2 mM Na
10 μg/ml aprotinin, and 10 μg/ml leupeptin, pH 7.4. GSH was measured
as previously described with some modiﬁ cations (van der Vliet et al.,
1998). In brief, samples were mixed 1:1 with 2 mM monobromobimane
(Thiolyte; Calbiochem) in 50 mM N-ethylmorpholine, pH 8.0, and incu-
bated at RT for 5 min in the dark. Trichloroacetic acid was added to the re-
action mixture to a ﬁ nal concentration of 5%. Samples were centrifuged at
3,000 g for 5 min, and supernatants were injected onto a Waters Symme-
tree C-18 column (150 × 4.5 mm). The GSH-monobromobimane adduct
was eluted with 10% CH
CN/0.25% glacial acetic acid and detected by
ﬂ uorescence emission of 480 nm after excitation at 395 nm.
C10 cells were plated on glass coverslips in 100-mm tissue culture dishes,
synchronized or allowed to grow asynchronously to 70% conﬂ uence, and
treated as indicated. Coverslips were rinsed with PBS, ﬁ xed with 3% para-
formaldehyde for 15 min at RT, and washed several times with PBS, and
cells were permeabilized with 0.1% Triton X-100 in PBS for 15 min at RT.
After gentle washing, coverslips were blocked for 1 h at RT with 10% nor-
mal goat serum in PBS and incubated with 1 μg/ml Prx-SO
H antibody in
PBS with 1% BSA overnight at 4°C. Alexa Fluor 594 (Invitrogen) conju-
gated goat anti–rabbit secondary antibody at 1 μg/ml in PBS was added
for 25 min at RT in the dark. Coverslips were mounted on slides, and
Figure 9. A model for Prx hyperoxidation in cell signaling. Prxs may
serve as dosimeters for redox-dependent signaling events, with structural
transitions induced by hyperoxidation disrupting interactions with regula-
tory factors or modulating other redox-dependent processes, such as actin
stress ﬁ ber formation. In addition, Prx-SO
H oligomers may be directly rec-
ognized as a signal that warns cells of perturbations in oxidant metabolism
and thereby contribute to stress responses that mediate oxidant-induced
cell cycle arrest.
JCB • VOLUME 175 • NUMBER 5 • 2006 788
images were generated at RT using a confocal scanning laser microscope
(MRC 1024 ES; Bio-Rad Laboratories) on a stand (BX50; Olympus), using
a 40× Plan-Apo lens (Olympus) with a 0.95 NA and a correction collar.
Digital images were collected with Laser Sharp Capture Software (Bio-Rad
Laboratories) and processed as black-and-white images. Contrast was ad-
justed using Photoshop (Adobe).
We thank Peter Burch and Anne Loonen for measuring the generation of H
in medium by GOx and Yvonne Janssen-Heininger and Todd Lowther for useful
This work was supported by grants from the National Heart, Lung, and
Blood Institute (P01 HL67004) and General Medical Sciences (R01
GM074204). T.J. Phalen was supported by an National Institute of Environ-
mental Health Sciences environmental pathology training grant (T32
Submitted: 2 June 2006
Accepted: 30 October 2006
Arner, E.S., M. Bjornstedt, and A. Holmgren. 1995. 1-Chloro-2,4-dinitroben-
zene is an irreversible inhibitor of human thioredoxin reductase. Loss
of thioredoxin disul de reductase activity is accompanied by a large
increase in NADPH oxidase activity. J. Biol. Chem. 270:3479–3482.
Arnold, R.S., J. Shi, E. Murad, A.M. Whalen, C.Q. Sun, R. Polavarapu,
S. Parthasarathy, J.A. Petros, and J.D. Lambeth. 2001. Hydrogen perox-
ide mediates the cell growth and transformation caused by the mitogenic
oxidase Nox1. Proc. Natl. Acad. Sci. USA. 98:5550–5555.
Baty, J.W., M.B. Hampton, and C.C. Winterbourn. 2005. Proteomic detection
of hydrogen peroxide-sensitive thiol proteins in Jurkat cells. Biochem.
Biteau, B., J. Labarre, and M.B. Toledano. 2003. ATP-dependent reduction of cys-
teine-sulphinic acid by S. cerevisiae sulphiredoxin. Nature. 425:980–984.
Bozonet, S.M., V.J. Findlay, A.M. Day, J. Cameron, E.A. Veal, and B.A. Morgan.
2005. Oxidation of a eukaryotic 2-Cys peroxiredoxin is a molecular
switch controlling the transcriptional response to increasing levels of hy-
drogen peroxide. J. Biol. Chem. 280:23319–23327.
Budanov, A.V., A.A. Sablina, E. Feinstein, E.V. Koonin, and P.M. Chumakov.
2004. Regeneration of peroxiredoxins by p53-regulated sestrins, homo-
logs of bacterial AhpD. Science. 304:596–600.
Burch, P.M., and N.H. Heintz. 2005. Redox regulation of cell-cycle re-entry:
cyclin D1 as a primary target for the mitogenic effects of reactive oxygen
and nitrogen species. Antioxid. Redox Signal. 7:741–751.
Burch, P.M., Z. Yuan, A. Loonen, and N.H. Heintz. 2004. An extracellular signal-
regulated kinase 1- and 2-dependent program of chromatin traf cking of
c-Fos and Fra-1 is required for cyclin D1 expression during cell cycle
reentry. Mol. Cell. Biol. 24:4696–4709.
Chang, J.W., S.H. Lee, J.Y. Jeong, H.Z. Chae, Y.C. Kim, Z.Y. Park, and Y.J. Yoo.
2005. Peroxiredoxin-I is an autoimmunogenic tumor antigen in non-small
cell lung cancer. FEBS Lett. 579:2873–2877.
Chang, T.S., W. Jeong, S.Y. Choi, S. Yu, S.W. Kang, and S.G. Rhee. 2002.
Regulation of peroxiredoxin I activity by Cdc2-mediated phosphorylation.
J. Biol. Chem. 277:25370–25376.
Chang, T.S., W. Jeong, H.A. Woo, S.M. Lee, S. Park, and S.G. Rhee. 2004.
Characterization of mammalian sul redoxin and its reactivation of hy-
peroxidized peroxiredoxin through reduction of cysteine sul nic acid in
the active site to cysteine. J. Biol. Chem. 279:50994–51001.
Choi, M.H., I.K. Lee, G.W. Kim, B.U. Kim, Y.H. Han, D.Y. Yu, H.S. Park, K.Y.
Kim, J.S. Lee, C. Choi, et al. 2005. Regulation of PDGF signalling and
vascular remodelling by peroxiredoxin II. Nature. 435:347–353.
Egler, R.A., E. Fernandes, K. Rothermund, S. Sereika, N. de Souza-Pinto,
P. Jaruga, M. Dizdaroglu, and E.V. Prochownik. 2005. Regulation of reac-
tive oxygen species, DNA damage, and c-Myc function by peroxiredoxin 1.
Finkel, T. 2003. Oxidant signals and oxidative stress. Curr. Opin. Cell Biol.
Forman, H.J., J.M. Fukuto, and M. Torres. 2004. Redox signaling: thiol chemis-
try de nes which reactive oxygen and nitrogen species can act as second
messengers. Am. J. Physiol. Cell Physiol. 287:C246–C256.
Gabbita, S.P., K.A. Robinson, C.A. Stewart, R.A. Floyd, and K. Hensley. 2000.
Redox regulatory mechanisms of cellular signal transduction. Arch.
Biochem. Biophys. 376:1–13.
Harris, J.R., E. Schroder, M.N. Isupov, D. Schef er, P. Kristensen, J.A. Littlechild,
A.A. Vagin, and U. Meissner. 2001. Comparison of the decameric structure
of peroxiredoxin-II by transmission electron microscopy and X-ray
crystallography. Biochim. Biophys. Acta. 1547:221–234.
Immenschuh, S., and E. Baumgart-Vogt. 2005. Peroxiredoxins, oxidative stress,
and cell proliferation. Antioxid. Redox Signal. 7:768–777.
Jang, H.H., S.Y. Kim, S.K. Park, H.S. Jeon, Y.M. Lee, J.H. Jung, S.Y. Lee, H.B.
Chae, Y.J. Jung, K.O. Lee, et al. 2006. Phosphorylation and concomitant
structural changes in human 2-Cys peroxiredoxin isotype I differentially
regulate its peroxidase and molecular chaperone functions. FEBS Lett.
Jeong, W., S.J. Park, T.S. Chang, D.Y. Lee, and S.G. Rhee. 2006. Molecular
mechanism of the reduction of cysteine sul nic acid of peroxiredoxin to
cysteine by mammalian sul redoxin. J. Biol. Chem. 281:14400–14407.
Jönsson, T.J., M.S. Murray, L.C. Johnson, L.B. Poole, and W.T. Lowther. 2005.
Structural basis for the retroreduction of inactivated peroxiredoxins by
human sul redoxin. Biochemistry. 44:8634–8642.
Kang, S.W., H.Z. Chae, M.S. Seo, K. Kim, I.C. Baines, and S.G. Rhee. 1998.
Mammalian peroxiredoxin isoforms can reduce hydrogen peroxide gen-
erated in response to growth factors and tumor necrosis factor-alpha.
J. Biol. Chem. 273:6297–6302.
Karin, M., and E. Shaulian. 2001. AP-1: linking hydrogen peroxide and oxi-
dative stress to the control of cell proliferation and death. IUBMB Life.
Koo, K.H., S. Lee, S.Y. Jeong, E.T. Kim, H.J. Kim, K. Kim, K. Song, and
H.Z. Chae. 2002. Regulation of thioredoxin peroxidase activity by C-
terminal truncation. Arch. Biochem. Biophys. 397:312–318.
Malkinson, A.M., L.D. Dwyer-Nield, P.L. Rice, and D. Dinsdale. 1997. Mouse
lung epithelial cell lines—tools for the study of differentiation and the
neoplastic phenotype. Toxicology. 123:53–100.
Moon, E.Y., Y.H. Han, D.S. Lee, Y.M. Han, and D.Y. Yu. 2004. Reactive oxygen
species induced by the deletion of peroxiredoxin II (PrxII) increases the
number of thymocytes resulting in the enlargement of PrxII-null thymus.
Eur. J. Immunol. 34:2119–2128.
Moon, J.C., Y.S. Hah, W.Y. Kim, B.G. Jung, H.H. Jang, J.R. Lee, S.Y. Kim, Y.M.
Lee, M.G. Jeon, C.W. Kim, et al. 2005. Oxidative stress-dependent struc-
tural and functional switching of a human 2-Cys peroxiredoxin isotype II
that enhances HeLa cell resistance to H2O2-induced cell death. J. Biol.
Mu, Z.M., X.Y. Yin, and E.V. Prochownik. 2002. Pag, a putative tumor sup-
pressor, interacts with the Myc Box II domain of c-Myc and selectively
alters its biological function and target gene expression. J. Biol. Chem.
Pantano, C., P. Shrivastava, B. McElhinney, and Y. Janssen-Heininger. 2003.
Hydrogen peroxide signaling through tumor necrosis factor receptor 1
leads to selective activation of c-Jun N-terminal kinase. J. Biol. Chem.
Park, J.H., Y.S. Kim, H.L. Lee, J.Y. Shim, K.S. Lee, Y.J. Oh, S.S. Shin, Y.H.
Choi, K.J. Park, R.W. Park, and S.C. Hwang. 2006. Expression of perox-
iredoxin and thioredoxin in human lung cancer and paired normal lung.
Parsonage, D., D.S. Youngblood, G.N. Sarma, Z.A. Wood, P.A. Karplus, and
L.B. Poole. 2005. Analysis of the link between enzymatic activity and
oligomeric state in AhpC, a bacterial peroxiredoxin. Biochemistry.
Persinger, R.L., W.M. Blay, N.H. Heintz, D.R. Hemenway, and Y.M. Janssen-
Heininger. 2001. Nitrogen dioxide induces death in lunch epithelial
cells in a density-dependent manner. Am. J. Respir. Cell Mol. Biol.
Ranjan, P., and N. Heintz. 2006. S-phase arrest by reactive nitrogen species is
bypassed by okadaic acid, an inhibitor of protein phosphatases PP1/PP2.
Free Radic. Biol. Med. 15: 247–249.
Ranjan, P., V. Anathy, P.M. Burch, K. Weirather, J.D. Lambeth, and N.H.
Heintz. 2006. Redox-dependent expression of cyclin D1 and cell prolif-
eration by Nox1 in mouse lung epithelial cells. Antioxid. Redox Signal.
Rhee, S.G., S.W. Kang, W. Jeong, T.S. Chang, K.S. Yang, and H.A. Woo. 2005.
Intracellular messenger function of hydrogen peroxide and its regulation
by peroxiredoxins. Curr. Opin. Cell Biol. 17:183–189.
Roovers, K., and R.K. Assoian. 2003. Effects of rho kinase and actin stress bers
on sustained extracellular signal-regulated kinase activity and activation
of G(1) phase cyclin-dependent kinases. Mol. Cell. Biol. 23:4283–4294.
Schwartz, M.A., and R.K. Assoian. 2001. Integrins and cell proliferation: regu-
lation of cyclin-dependent kinases via cytoplasmic signaling pathways.
J. Cell Sci. 114:2553–2560.
Tu, B.P., A. Kudlicki, M. Rowicka, and S.L. McKnight. 2005. Logic of the yeast
metabolic cycle: temporal compartmentalization of cellular processes.
PEROXIREDOXINS AND REDOX SIGNALING • PHALEN ET AL. 789
van der Vliet, A., J.P. Eiserich, B. Halliwell, and C.E. Cross. 1997. Formation of
reactive nitrogen species during peroxidase-catalyzed oxidation of nitrite.
A potential additional mechanism of nitric oxide-dependent toxicity.
J. Biol. Chem. 272:7617–7625.
van der Vliet, A., P.A. Hoen, P.S. Wong, A. Bast, and C.E. Cross. 1998.
Formation of S-nitrosothiols via direct nucleophilic nitrosation of thiols
by peroxynitrite with elimination of hydrogen peroxide. J. Biol. Chem.
Vivancos, A.P., E.A. Castillo, B. Biteau, C. Nicot, J. Ayte, M.B. Toledano, and E.
Hidalgo. 2005. A cysteine-sul nic acid in peroxiredoxin regulates H
sensing by the antioxidant Pap1 pathway. Proc. Natl. Acad. Sci. USA.
Wen, S.T., and R.A. Van Etten. 1997. The PAG gene product, a stress-induced
protein with antioxidant properties, is an Abl SH3-binding protein and
a physiological inhibitor of c-Abl tyrosine kinase activity. Genes Dev.
Wood, Z.A., L.B. Poole, R.R. Hantgan, and P.A. Karplus. 2002. Dimers to
doughnuts: redox-sensitive oligomerization of 2-cysteine peroxiredoxins.
Wood, Z.A., L.B. Poole, and P.A. Karplus. 2003a. Peroxiredoxin evolution and
the regulation of hydrogen peroxide signaling. Science. 300:650–653.
Wood, Z.A., E. Schroder, J. Robin Harris, and L.B. Poole. 2003b. Structure,
mechanism and regulation of peroxiredoxins. Trends Biochem. Sci.
Xiao, N., G. Du, and M.A. Frohman. 2005. Peroxiredoxin II functions as
a signal terminator for H
-activated phospholipase D1. FEBS J.
Yuan, Z., H. Schellekens, L. Warner, Y. Janssen-Heininger, P. Burch, and N.H.
Heintz. 2003. Reactive nitrogen species block cell cycle re-entry through
sustained production of hydrogen peroxide. Am. J. Respir. Cell Mol. Biol.
Yuan, Z., D.J. Taatjes, B.T. Mossman, and N.H. Heintz. 2004. The duration of
nuclear extracellular signal-regulated kinase 1 and 2 signaling during cell
cycle reentry distinguishes proliferation from apoptosis in response to
asbestos. Cancer Res. 64:6530–6536.