www.elsevier.com/locate/freeradbiomed

Free Radical Biology &

Original Contribution

Mitochondrial redox state regulates transcription of the nuclear-encoded

mitochondrial protein manganese superoxide dismutase: a proposed

adaptive response to mitochondrial redox imbalance

Aekyong Kima, Michael P. Murphyb, Terry D. Oberleyc,d,*

aMolecular and Environmental Toxicology Center, University of Wisconsin, Madison, WI, USAbMedical Research Council Dunn Human Nutrition Unit, Hills Road, Cambridge, UK

cDepartment of Pathology and Laboratory Medicine, University of Wisconsin Medical School, Madison, WI, USAdPathology and Laboratory Medicine Service, William S. Middleton Veterans Memorial Hospital, Madison, WI, USA

Received 21 July 2004; revised 6 October 2004; accepted 22 October 2004

Available online 20 November 2004

Abstract

Overexpression of human manganese superoxide dismutase (MnSOD) in mouse NIH/3T3 cells using an inducible retroviral system led to

alterations in the mitochondrial redox state since levels of reactive oxygen species rapidly increased after induction of human MnSOD

(Antioxid. Redox Signal. 6:489–500; 2004). Alterations in exogenous human MnSOD led to large increases in levels of endogenous mouse

MnSOD (sod2) and thioredoxin 2 (txn2) mRNAs, but smaller increases in MnSOD and thioredoxin 2 protein expression. Tight regulation of

mitochondrial protein levels seems to be necessary for optimal cellular function, since mitochondrial antioxidant protein levels did not

increase to the same extent as antioxidant protein mRNA levels. We hypothesize that these changes in antioxidant proteins are adaptations to

the altered mitochondrial redox state elicited by MnSOD overexpression. The mitochondrial-specific antioxidant MitoQ reversed cell growth

inhibition, and greatly decreased levels of endogenous sod2 and txn2 transcripts following induction of exogenous MnSOD. Elevated levels

of mouse sod2 transcripts resulted from transcriptional activation of the endogenous sod2 gene since actinomycin D prevented transcription

of this gene. Therefore, the mitochondrial redox state appears to modulate a nuclear-driven biochemical event, i.e., transcriptional activation

of a nuclear gene encoding a protein targeted to mitochondria.

D 2004 Elsevier Inc. All rights reserved.

Keywords: MnSOD; Thioredoxin 2; Transcription; mRNA; Mitochondrial redox; H2O2; Mitochondrial antioxidant; MitoQ; Free radicals

0891-5849/$ - see front matter D 2004 Elsevier Inc. All rights reserved.

doi:10.1016/j.freeradbiomed.2004.10.030

Abbreviations: AOE, antioxidant enzyme; DCF,2V,7V - dichlorofluor-

escein diacetate; GAPDH, glyceraldehyde-3-phosphate dehydrogenase;

H2O2, hydrogen peroxide; IPTG, isopropyl h-thiogalactoside; MitoQ, a

mitochondrial-specific antioxidant which is a mixture of mitoquinol and

mitoquinone; MnSOD, manganese containing superoxide dismutase; O2S�,

superoxide anion; Prxns, peroxiredoxins; RNS, reactive nitrogen species;

ROS, reactive oxygen species; SOD, superoxide dismutase; sod2, MnSOD

gene; TNFa, tumor necrosis factor-a; txn2, thioredoxin 2 gene.

* Corresponding author. William S. Middleton Veterans Memorial

Hospital, Room A35, 2500 Overlook Terrace, Madison, WI 53705, USA.

Fax: (608) 280 7087.

E-mail address: toberley@wisc.edu (T.D. Oberley).

Introduction

In eukaryotic cells, three isoforms of superoxide dis-

mutase (SOD) are present: extracellular copper/zinc-con-

taining SOD (sod3:ECSOD), mitochondrial manganese-

containing SOD (sod2:MnSOD), and cytoplasmic/nuclear

copper/zinc-containing SOD (sod1:CuZnSOD), although

the latter also localizes to the mitochondrial intermembrane

space [1]. While the SOD isoenzymes catalyze the identical

dismutation reaction involving the conversion of superoxide

anion (O2S�) to oxygen (O2) and hydrogen peroxide (H2O2),

the function of each SOD isoform in cellular physiology

appears to be very different, and often one SOD cannot

compensate for another. For example, the lethal phenotype

Medicine 38 (2005) 644–654

A. Kim et al. / Free Radical Biology & Medicine 38 (2005) 644–654 645

exhibited by sod2�/� knockout mice was not corrected or

delayed by sod1 overexpression, suggesting that the

subcellular location of SOD is important in physiological

functions of these enzymes [2].

In a resting cell, superoxide anion is produced at 1–2% of

total daily oxygen consumption during electron transfer and

oxidative phosphorylation for ATP generation by mitochon-

dria [3]. Consequently, mitochondria are the major source of

reactive oxygen species (ROS) in a resting cell. Due to their

ubiquitous reactivities, ROS can be converted to other

radicals such as reactive nitrogen species (RNS) and

reactive carbon species in subcellular location- and local

redox state-dependent manners [4]. Mitochondrial ROS are

now appreciated as regulators of mitochondrial functions

including electron transfer chain enzymes and mitochondrial

membrane potential [5,6]. In addition, they are involved in

regulation of other subcellular organelles [4].

The inverse correlation between MnSOD activity and

cell growth is a paradoxical phenomenon when one

postulates that MnSOD functions only as an antioxidant

enzyme (AOE) to protect a cell from oxidative stress caused

by O2S� [7]. Proposed hypotheses regarding mechanisms by

which MnSOD exerts growth inhibition often emphasize

increased H2O2 production secondary to elevated MnSOD

activity resulting in oxidative environments first in mito-

chondria and subsequently in the cytoplasm [8–10].

Previous studies from our laboratory demonstrated that

overexpression of human MnSOD in NIH/3T3 fibroblasts

utilizing an inducible retroviral system resulted in inhibition

of cell growth via prolongation of G1 and S phases of the cell

cycle (see Fig. 5 for a summary of data from the previous

study) [11]. We postulated that this was a physiological

mechanism to regulate cell cycle progression since prolon-

gation of G1 and S phases was completely reversible

following removal of the inducer and hence return of

MnSOD expression to the original level. When cellular

ROS levels were measured using DCF fluorescent dye and

flow cytometry, a burst of ROS with concomitantly

decreased mitochondrial membrane potential preceded tran-

sient and reversible cell cycle modulation following MnSOD

induction. Sustained ROS increase was thus not necessary

for MnSOD-mediated growth inhibition (see Fig. 5) [11].

As a result of these previous observations, we hypothe-

sized that changes in the mitochondrial redox state by

elevated MnSOD activity were utilized to coordinate

physiological and biochemical events in two subcellular

compartments. Also, MnSOD may serve as a regulatory

enzyme modified by the mitochondrial redox environment

[11].

In addition to a transient ROS burst after MnSOD

induction, newly established steady-state levels of ROS

after the initial ROS burst were approximately 50% lower

than control levels (see Fig. 5) [11]. We postulated that there

must be alterations in mitochondrial antioxidant proteins

and/or antioxidant enzymes responsible for the newly

established low steady-state levels of ROS. Guo et al. [12]

have previously proposed that overexpression of MnSOD

results in redox alterations with subsequent expression of

stress-responsive nuclear genes. We speculated that adaptive

changes in mitochondrial redox capacity following MnSOD

overexpression could also occur at the transcriptional level

of nuclear genes encoding mitochondrial antioxidant pro-

teins and/or AOEs.

To test the hypothesis of mitochondrial redox state-

mediated transcriptional regulation of nuclear genes encod-

ing mitochondrial antioxidant proteins and/or AOEs,

MitoQ, a mitochondrial specific antioxidant, was utilized

in this study. MitoQ refers to the mixture of the mitochond-

rially targeted-quinol (mitoquinol) and –quinone (mitoqui-

none); these ubiquinone derivatives preferentially

accumulate in the mitochondria matrix at levels over several

hundred-fold compared to cytoplasm [13]. The antioxidant

properties of MitoQ are mediated by oxidation of mitoqui-

nol to mitoqinone, which is subsequently regenerated

through reduction to mitoquinol by respiratory complexes.

The efficacy of MitoQ as a mitochondrial antioxidant was

previously demonstrated: (1) prevention of lipid peroxida-

tion by H2O2 and ferrous iron, (2) scavenging of peroxyni-

trite, and (3) prevention of apoptosis by H2O2, but not by

staurosporine or tumor necrosis factor-a (TNFa) [14].

In this paper, we utilized MnSOD as a molecular tool to

manipulate the mitochondrial redox state through its

antioxidant activity of removing O2S� and/or its pro-oxidant

activity of generating H2O2. mRNA levels of endogenous

MnSOD (sod2) and thioredoxin 2 (txn2) genes were

modulated by inducible expression of exogenous MnSOD

cDNA. MnSOD and thioredoxin 2 are mitochondrial

proteins. In studies utilizing a mitochondrial-specific anti-

oxidant (MitoQ), transient redox imbalance in mitochondria

following MnSOD induction was shown to be responsible

for transcriptional activation of the nuclear-encoded endo-

genous sod2 gene. These observations provide another

example, besides the previously described modulation of

cell cycle kinetics [11], in which the mitochondrial redox

state coordinates cellular events directed by other subcel-

lular organelles, that is, the transcriptional regulation of a

nuclear gene encoding a protein targeted to mitochondria.

Materials and methods

Chemicals and reagents

All chemicals were purchased from Sigma Chemical Co.

(Saint Louis, MO), unless otherwise specified. Tissue

culture supplies were from Falcon (Becton Dickinson

Labware, Franklin Lakes, NJ). WI-38 VA 13 (SV-40-

transformed human lung fibroblast) and NIH/3T3 (mouse

embryo fibroblast) cell lines were obtained from the

American Type Culture Collection (Manassas, VA). DMEM

with high glucose (4.5 g/liter), and IPTG (isopropyl h-thiogalactoside) were obtained from Life Technology

A. Kim et al. / Free Radical Biology & Medicine 38 (2005) 644–654646

(Gaithersburg, MD). Fetal bovine serum was purchased

from HyClone (Logan, UT). Anti-thioredoxin 2 antibody

was purchased from Lab Frontier (Seoul, Korea). Primary

antibody against human MnSOD was generously provided

by Dr. Larry W. Oberley of the University of Iowa (Iowa

City, Iowa). MitoQ refers to a mixture of the mitoquinol

(10-(6V-ubiquinolyl)decyltriphenylphosphonium bromide)

and the mitoquinone (10-(6V-ubiquinonyl)decyltriphenyl-phosphonium bromide). The synthesis and characteristics of

MitoQ were described previously [13,14]. In this study, 10

mM solution of MitoQ in ethanol was stored at �208C in

the dark with desiccant under nitrogen. TPMP (methyltri-

phenylphosphonium bromide), used as a control lipophilic

cation without significant antioxidant capacity [13], was

prepared as a 50 AM solution in H2O and stored under the

same conditions as MitoQ.

Establishment and characterization of the hms1 cell line

Detailed information concerning characteristics of hms1

cells was published elsewhere [11]. Briefly, the hms1 cell

line was developed from NIH/3T3 mouse embryo fibro-

blasts utilizing an inducible LNXCO3 retroviral vector

containing human MnSOD cDNA. The human cDNA was

sequenced and was identical to the sequence in GenBank

Accession Number Y00985 with the following changes: a

site-specific mutation of cytidine339 to thymidine339 in the

coding region, and a nucleotide deletion of thymidine931 in

the 3V UTR region. The former was introduced since human

MnSOD with isoleucine at amino acid position 58 was

shown to have higher enzyme activity (twofold) compared

to MnSOD with threonine at the same position [15],

whereas the latter was an unexpected outcome of the

cloning process. When expressed, mature sod2 mRNAs

contain the 5V UTR region of 94 bases and the 3V UTR

region of 212 bases. When cells were treated with inducer

(IPTG), for example, with 25 Ag/ml IPTG for 24 h,

immunoreactive MnSOD protein was increased approxi-

mately twofold over control; exclusive subcellular locali-

zation of MnSOD protein to mitochondria was confirmed

using immunogold electron microscopy and elevation of

MnSOD biochemical activity following induction has been

documented elsewhere [11].

Cell culture

The hms1 cells utilized in this study were routinely

maintained in T75 flasks at 25–90% confluency in DMEM

media with 10% serum at 378C in a humidified atmosphere

of 95% air and 5%CO2.Media were renewed every 3–4 days.

A solution of 0.05% trypsin and 0.53 mM EDTA was used

for routine subculture. Cells with a passage number of less

than 30 were used in all experiments. Mycoplasma contam-

ination was regularly tested using the Immun-Mark Mycotest

kit (ICN Biomedicals Inc., Aurora, OH), and was negative in

the cultures used for these studies.

DNA fluorometric assay

Detailed procedures are described elsewhere [11]. Cells

were seeded at 3000 per well in 96 well plates. Twenty-four

hours later, MnSOD was induced with 25 Ag/ml IPTG in the

presence or absence of 50 nM MitoQ treatment. Forty-eight

hours later, cells were washed twice with Krebs-Ringer

buffer (KR buffer, pH 7.4) (Sigma) and frozen at �708C in

100 Al KR buffer per well overnight. On the day of assay,

plates were thawed at room temperature for at least 2 h. A

200 Al TNE high salt solution (10 mM Tris base, 1 mM

EDTAd 2Na, 2 M NaCl, pH 7.4) containing 10 Ag/ml

Hoechst 33258 (Sigma) was added to each well. Plates were

incubated for at least 2 h at room temperature before

measuring Hoechst fluorescence in each well at wave-

lengths Ex360 nm/Em515 nm. KR buffer was used as a blank,

and Hoechst fluorescence from media was subtracted as

background.

Quantitative real-time PCR (QPCR)

hms1 cells were plated at a density of 100 cells/mm2

surface area and cultured for 24 h to reach early log phase of

the growth curve. Initiation of MnSOD induction was

achieved by adding fresh media containing 25 Ag/ml IPTG.

For MitoQ treatment, 50 nM of MitoQ or TPMP, the latter

as a control for nonspecific effects of lipophilic cations, was

added 1 h prior to the addition of IPTG. Cells were

harvested at the indicated times after initiation of induction

using cell dissociation solution (Sigma), and 20,000 cells

were collected and stored at �708C. Pellets were thawed on

ice and lysed using Cell Lysis II Buffer, and cDNA was

synthesized with 10 Al of lysed sample (Cells-to-cDNA II

kit, Ambion, Austin, TX) according to the manufacturer’s

instructions.

cDNAs equivalent to 200 cells were mixed with 120 nM

of primers and SYBR Master Mix (Applied Biosystems,

Foster City, CA) in a total volume of 25 Al. QPCR was

conducted using the following parameters: 958C for 15 min,

and 50 cycles at 958C for 10 s and 608C for 20 s. Also,

melting curves were verified between 55 and 958C with

0.58C of temperature increments. QPCR was performed in

96-well optical reaction plates using iCycler (Bio-Rad

Laboratories, Hercules, CA). Threshold cycle number (Ct

value) was analyzed using iCycler iQ optical system

software (Bio-Rad, version 3.0a), and amplification effi-

ciency of each QPCR was corrected for analysis as

previously described [16]. QPCR was normalized to the

Ct value of 18S ribosomal RNA from the same sample.

Primer sequences were as follows: human sod2 transcript

forward primer, 5V-GCTGACGGCTGCATCTGTT-3V;reverse primer, 5V-CCTGATTTGGACAAGCAGCAA-3V;mouse sod2 transcript forward primer, 5V-ATTAACGCG-CAGATCATGCA-3V; reverse primer, 5V-TGTCCCCCAC-CATTGAACTT-3V [17]; mouse thioredoxin 2 transcript

forward primer, 5V-GGACCGCGGCTAGAGAAGAT-3V;

A. Kim et al. / Free Radical Biology & Medicine 38 (2005) 644–654 647

reverse primer, 5V-GCTGGTCCTCGTCCTTGATC-3V[17]; 18S ribosomal RNA forward primer, 5V-CGCCGCTAGAGGTGAAATTCT-3V; reverse primer, 5V-CGAACCTCCGACTTTCGTTCT-3V.

Actinomycin D treatment and QPCR

To test transcriptional activation of the mouse sod2 gene,

hms1 cells were plated, and MnSOD expression was

induced as described above. Ten hours after the initiation

of MnSOD induction, a time point at which mouse sod2

mRNA level was rapidly rising, actinomycin D (Sigma,

A5156) was added to media at a final concentration of 5 AM.

Cells were collected every hour for 2 h, and QPCR was

performed as described above.

Western blotting

Cells were washed and scrape-harvested in PBS. Whole

cell lysates were prepared with M-PER mammalian protein

extraction reagent (Pierce, Rockford, IL). Samples were

centrifuged at 18,845g for 15 min at 48C, and total protein

concentration of supernatant was measured using the

Bradford assay (Bio-Rad). The amount of 10–50 Ag of

total proteins was separated and transferred to a PVDF

membrane using standard SDS-PAGE procedure. Mem-

branes were blocked in 5% nonfat milk in TBST (153 mM

Tris HCl, 150 mM NaCl, and 0.05% Tween 20, pH 7.8).

After anti-MnSOD antibody (1:5000 dilution) or anti-

thioredoxin 2 antibody (1:1000 dilution) and horseradish

peroxidase-conjugated secondary antibody (1:50,000 dilu-

tion) (Bio-Rad) incubations, signals were detected using

ECL plus and Storm 860, and quantitated using ImageQuant

software version 5.0 (Amersham Bioscience Corp., Piscat-

away, NJ). Anti-glyceraldehyde-3-phosphate dehydrogenase

(GAPDH) (1:10,000 dilution) (Advanced Immuno Chemi-

cal Inc., Long Beach, CA) antibody was used to normalize

MnSOD or thioredoxin 2 bands. Anti-MnSOD antibody

used in the current studies was shown to detect both human

and murine MnSOD proteins [8,9].

Statistics

Statistical analysis was performed using either Student’s

t test with two-tailed distribution or one-way ANOVA for

comparison of different groups.

Results

Overexpression of human MnSOD cDNA led to transient

increases in mRNA levels of endogenous sod2 and txn2 genes

Since the hms1 cell line was developed to express human

MnSOD protein in murine cells, it was possible to

discriminate the origin of sod2 mRNAs using a primer pair

specific for either human or murine MnSOD mRNA. Each

sod2 primer pair was cross-tested in human WI-38 VA 13

and mouse NIH/3T3 cells for its species specificity, with

results demonstrating the generation of sod2 amplification-

products in a species specific manner (data not shown). As

shown in Fig. 1A, human sod2 mRNA levels were

increased approximately 3-fold compared to control as early

as 3 h after initiation of induction. Although there was a

significant decrease at 6 h after induction, the levels of

human sod2 mRNA remained elevated for 15 h after

initiation of induction at an average level of 2.4-fold over

control. In our previous study [11], it was shown that there

was increased immunoreactive MnSOD protein expression

as early as 3 h after initiation of induction using quantitative

immunogold electron microscopy. The kinetics of human

sod2 mRNA expression were in excellent agreement with

the expression profile of immunoreactive MnSOD protein.

mRNA levels of endogenous AOE were also screened as

a function of MnSOD induction time. Sequence information

for endogenous AOE primer pairs screened in this study is

available elsewhere [17]. Of 18 primers of endogenous

antioxidant and/or AOEs screened (for the complete list,

refer to Table 1), only sod2 and txn2 mRNA levels were

significantly elevated. Levels of mRNA of endogenous

sod2 and txn2 genes were altered as a function of

exogenous MnSOD expression time; mouse sod2 mRNAs

increased approximately 15-fold over control at 12 h (Fig.

1C), whereas mouse txn2 transcripts reached about 6-fold

over control at 9 h (Fig. 1E) after initiation of exogenous

MnSOD induction.

When changes in mRNA levels were expressed as a

ratio compared to the control level at a given time point

(Figs. 1A, C, and E), distinct alterations of exogenous sod2

and endogenous sod2 and txn2 transcripts were present

during different time periods. The first 6-h period after

initiation of human MnSOD induction showed increases in

both human and mouse sod2 mRNA amounts at 3 h (Figs.

1A and C). It was expected to have increased human sod2

mRNA level following IPTG treatment in hms1 cells. On

the other hand, simultaneously increased mouse sod2

mRNA level was unexpected. We postulate that this could

be an adaptive response secondary to the ROS burst at 3 h

after initiation of MnSOD induction reported in the

previous study [11].

The second 6-h period showed peak increases in

endogenous sod2 at 12 h (Fig. 1C) and txn2 mRNA at 9

h (Fig. 1E) after human MnSOD induction. These peaks of

endogenous sod2 and txn2 transcripts gradually decreased

and returned to control levels (Figs. 1C and E). Elevated

sod2 and txn2 mRNA levels suggest that the initial ROS

increase observed at 3 h after MnSOD induction established

an oxidative mitochondrial redox environment, however

transiently. The temporal profiles and magnitudes of

alterations in human sod2, mouse sod2 and txn2 transcripts

strongly suggest transcriptional regulation of sod2 and txn2

genes.

Fig. 1. Modulation of endogenous sod2 and txn2 transcripts by exogenous sod2 mRNA expression. (A) Expression profile of human sod2 mRNA after IPTG

treatment. (B) Altered expression profile of human sod2 mRNA by simultaneous MitoQ treatment. (C) Alterations in mouse sod2 transcript as a function of

time after human sod2 mRNA expression. (D) Decreased mouse sod2 transcripts by simultaneous MitoQ treatment. (E) Changes in mouse txn2 transcript after

initiation of human MnSOD induction. (F) Decreased mouse txn2 transcripts by simultaneous MitoQ treatment. hms1 cells were treated without (A, C, and D)

or with 50 nMMitoQ for 1 h (B, D, and F). Then, induction of human MnSOD cDNAwas initiated by the addition of 25 Ag/ml IPTG at 0 h. Cells were collected

every 3 h, and mRNAs were analyzed by quantitative PCR as described under Materials and methods. mRNA level of interest is expressed as a fraction of 18S

RNA from the same sample in control condition (open circle with dotted line) or induced condition (open square with dotted line). The ratio of induced

condition over control condition at a given time point is shown as a closed circle with bold line. Values are meanF SD, n = 3; a is P b 0.05 compared with 0 h

control; b is P b 0.05 compared with control at a given time point; c is P b 0.05 compared with 50 nM MitoQ treated control condition at a given time point.

A. Kim et al. / Free Radical Biology & Medicine 38 (2005) 644–654648

MitoQ treatment prevented transient increases in mRNA

levels of endogenous sod2 and txn2 genes by exogenous

sod2 mRNA

From the previous study [11], we hypothesized that upon

MnSOD induction, mitochondrial redox status was changed

due to altered ROS levels, i.e., decreased O2S�, increased

H2O2, and/or chemical derivatives of these species. It was

Table 1

Mouse genes screened using real-time PCR in the current study

Gene Gene name Gene Gene name

Prxn1 Peroxiredoxin 1 Gpx3 Glutathione peroxidase 3

Prxn2 Peroxiredoxin 2 Gpx4 Glutathione peroxidase 4

Prxn3 Peroxiredoxin 3 Sod1 Superoxide dismutase 1

Prxn4 Peroxiredoxin 4 Sod2 Superoxide dismutase 2

Prxn5 Peroxiredoxin 5 Sod3 Superoxide dismutase 3

Prxn6 Peroxiredoxin 6 Txn1 Thioredoxin 1

Cat Catalase Txn2 Thioredoxin 2

Gpx1 Glutathione peroxidase 1 Glrx1 Glutaredoxin 1

Gpx2 Glutathione peroxidase 2 Glrx2 Glutaredoxin 2

Endogenous antioxidant and antioxidant enzyme genes screened using real-

time PCR in this study are listed. The primer sequence of each gene is

provided elsewhere [17].

not possible to distinguish which radical species was

ultimately responsible for changes in mitochondrial redox

status because of the ability of DCF to react with several

free radical species [18].

We hypothesized that changes in mitochondrial redox

status may be the cause for subsequent transiently increased

endogenous sod2 and txn2 transcripts. We expected to

attenuate/eliminate alterations in mitochondrial radical

species and/or mitochondrial redox state due to increased

MnSOD activity through strengthening mitochondrial anti-

oxidant buffer capacity with the mitochondrial-specific

antioxidant, MitoQ. The efficacy of MitoQ as a mitochon-

drial antioxidant was demonstrated previously [13,19,20].

To test the role of alterations in mitochondrial radicals

and/or mitochondrial redox status on the MnSOD-mediated

elevation of endogenous sod2 and txn2 transcripts, cells

were treated with 50 nM MitoQ prior to MnSOD induction;

then, changes in mRNA levels of exogenous sod2,

endogenous sod2, and txn2 genes were measured as

described above over a 24-h time period.

As shown in Fig. 1B, the kinetics of human sod2

mRNA changes was slightly altered with 50 nM MitoQ

treatment. The initial 3-fold peak at 3 h after MnSOD

Fig. 2. Effect of MitoQ treatment on growth inhibition by MnSOD

overexpression. MitoQ treatment prevented growth inhibition caused by

increased MnSOD expression as previously reported [11]. hms1 cells were

treated with (filled bar) or without 50 nM MitoQ (hatched bar) for 1 h prior

to the addition of 25 Ag/ml IPTG. Cellular proliferation at 48 h after

initiation of MnSOD induction was measured using Hoechst fluorescence

as described under Materials and methods. Hoechst fluorescence was

expressed as a percentage of control without MitoQ treatment. Values are

mean F SD, n = 3; a is P b 0.05 compared with control without 50 nM

MitoQ treatment at 48 h; b is P b 0.05 compared with induced condition

without 50 nM MitoQ treatment at 48 h.

Fig. 3. Transcriptional activation of endogenous sod2 gene by exogenous

sod2 mRNA expression. hms1 cells were treated with 25 Ag/ml IPTG at 0 h

to induce human MnSOD expression. At 10 h after initiation of MnSOD

induction, actinomycin D (Act D) was added at a final concentration of

5 AM. Cells were collected every hour for 2 h and levels of mouse sod2

mRNA were analyzed by quantitative PCR as described under Materials

and methods and expressed as a fraction of 18S RNA from the same

sample. Values are mean F SD, n = 3; a is P b 0.05 compared with control

at a given time point; b is P b 0.05 when comparing induced condition

without 5 AM Act D treatment to induced condition with 5 AM Act D

treatment at a given time point.

A. Kim et al. / Free Radical Biology & Medicine 38 (2005) 644–654 649

induction appeared to be delayed to 6 h with a 4-fold

increase, but overall levels of increase at an average of 2.2-

fold over control were similar to an average level of 1.9-

fold without MitoQ treatment. Significant changes were

observed in the kinetic profiles of mouse sod2 and txn2

transcripts. A 50 nM MitoQ treatment almost completely

blocked elevation of mouse sod2 and txn2 transcripts at 12

and 9 h after MnSOD induction, respectively (compare

Figs. 1D and F to Figs. 1C and E). Preliminary studies with

TPMP did not show an effect on mRNA levels of sod2 and

txn2 (data not shown).

Thus, enhanced mitochondrial redox buffer capacity by

MitoQ treatment successfully prevented or delayed altera-

tions in endogenous sod2 and txn2 transcripts following

increases in exogenous sod2 mRNA levels. These results

strongly suggest a causative role of mitochondrial radicals

and/or mitochondrial redox status in the elevation of

endogenous sod2 and txn2 transcripts by increased exoge-

nous sod2 mRNA expression.

MitoQ reversed cell growth inhibition caused by

overexpression of MnSOD

An inverse correlation between cell growth and MnSOD

activity was previously shown in hms1 cells [11]. As shown

in Fig. 2, 50 nM MitoQ treatment reversed cellular growth

inhibition due to increased MnSOD expression in hms1

cells, while only minimally affecting cellular proliferation in

control noninduced cells over a 48-h time period. Thus, the

previously observed inhibition of cell growth by MnSOD

overexpression must be mediated at least partially by

alterations in mitochondrial radicals and/or mitochondrial

redox state.

Expression of exogenous sod2 mRNA led to transcriptional

activation of the endogenous sod2 gene

Considering the time and magnitude of mRNA accumu-

lation, transcriptional modulation of nuclear genes appeared

to be a feasible mechanism to explain accumulation of sod2

and txn2 transcripts after human MnSOD induction. To test

whether transcriptional activation of the mouse sod2 gene

occurred as a response to human MnSOD expression, cells

were treated with 5 AM actinomycin D at 10 h after

initiation of human MnSOD induction, since this was a time

point at which endogenous sod2 mRNA was increasing

rapidly; cells were collected every hour for 2 h. As shown in

Fig. 3, induced cells treated with actinomycin D failed to

accumulate mouse sod2 transcripts, whereas levels in

untreated cells were increased approximately 14-fold over

control at 12 h after initiation of MnSOD induction. Since

mRNA profiles of both control and induced cells treated

with actinomycin D were indistinguishable from control

levels, alterations in mRNA stability seemed to play a minor

role, if any, in the accumulation of endogenous sod2

transcripts by exogenous sod2 mRNA expression. Consid-

ering the fact that MitoQ treatment appeared to almost

completely block increases in sod2 and txn2 mRNAs,

transcriptional activity of the nuclear sod2 gene encoding a

A. Kim et al. / Free Radical Biology & Medicine 38 (2005) 644–654650

mitochondrial MnSOD protein seemed to be modulated by

mitochondrial radicals and/or mitochondrial redox status.

Altered mRNA levels of mouse sod2 and txn2 genes by

MitoQ treatment led to changes in immunoreactive protein

levels of MnSOD and thioredoxin 2

To confirm that sod2 and txn2 mRNA increases led to

changes in protein amounts, immunoreactive protein levels

of MnSOD and thioredoxin 2 were measured following

MitoQ treatment. Since the anti-MnSOD antibody used in

the current study reacts with both human and mouse

MnSOD proteins, it was not possible to discriminate the

origin of immunoreactive MnSOD protein. In this study,

GAPDH was used as a loading control in the Western

analysis, and a previous study showed similar results when

normalized to h-actin [11].

As shown in Fig. 4A, maximal MnSOD protein

expression was obtained at 15 h after initiation of MnSOD

induction at a level of approximately 2-fold over control,

and expression levels were higher than control up to 24 h

after MnSOD induction. MnSOD activity gel, which

recognizes human but not mouse MnSOD activity (for a

Fig. 4. Effects of MitoQ treatment on expression of MnSOD and thioredoxin 2 p

profile of immunoreactive MnSOD protein in the absence (A) or presence (B) of 50

2 protein in the absence (C) or presence (D) of 50 nM MitoQ treatment. (E) W

representative quantified values of indicated protein/GAPDH below Western imag

the addition of 25 Ag/ml IPTG at 0 h and collected every 3 h. Immunoreactive prot

as fold over 0 h control. Values are mean F SD, n = 3; a is P b 0.05 compared wit

P b 0.05 compared with 50 nM MitoQ treated control conditions at a given time

detailed method, refer to [21]), clearly showed an increase in

human MnSOD activity as a function of time after IPTG

treatment (data not shown), demonstrating the functioning

of the retroviral inducible gene expression system.

When MnSOD was induced and simultaneously treated

with 50 nM MitoQ, there was a significant decrease in

MnSOD protein amount at 18 h after initiation of MnSOD

induction compared to the induced condition without MitoQ

treatment (Figs. 4B and E). When 50 nM TPMP was used as

a control for nonspecific effects of a lipophilic cation

inherent in the chemical structure of MitoQ [13], the

expression of MnSOD protein at 18 h after initiation of

induction was similar to control levels (Fig. 4E). We

postulate that the modulation in MnSOD protein amounts

observed at 18 h after initiation of MnSOD induction in the

absence or presence of MitoQ treatment resulted from the

modulation of endogenous sod2 mRNA that was observed

at 12 h after MnSOD induction (Fig. 1C).

The levels of immunoreactive thioredoxin 2 protein were

also modulated as a function of MnSOD induction time and

dependent on treatment with MitoQ. As shown in Figs. 4C

and F, the amount of thioredoxin 2 protein was lower in

MnSOD-induced cells than control cells up to 9 h but

rotein following human sod2 mRNA expression. Alterations in expression

nM MitoQ treatment. Modulations in levels of immunoreactive thioredoxin

estern blot images from Figs. A and C or (F) from Figs. B and D with

e. hms1 cells were pretreated with or without 50 nM MitoQ for 1 h prior to

eins were detected as described under Materials and methods and expressed

h 0 h control; b is P b 0.05 compared with control at a given time point; c is

point.

A. Kim et al. / Free Radical Biology & Medicine 38 (2005) 644–654 651

gradually recovered up to 15 h after initiation of MnSOD

induction.When cells were treated withMitoQ uponMnSOD

induction, the overall amounts of thioredoxin 2 protein were

similar to control levels (Figs. 4D and F). The transient

recovery of thioredoxin 2 protein level in the MnSOD-

induced condition at 15 h after initiation ofMnSOD induction

(Figs. 4C and F) was absent with MitoQ treatment (Figs. 4D

and F). These observations suggest the increased endogenous

txn2 mRNA level at 9 h after MnSOD induction (Fig. 1E)

resulted in transient recovery of thioredoxin 2 protein at 15 h

after MnSOD induction (Figs. 4C and F). The 50 nM TPMP

treatments resulted in levels of thioredoxin 2 protein similar

to those under control conditions (Fig. 4F). Taken together,

elevated levels of endogenous sod2 and txn2 transcripts

resulted in the increased expression of immunoreactive

MnSOD and thioredoxin 2 proteins.

Discussion

In the present study, we have shown that increased

MnSOD expression resulted in alterations in endogenous

mitochondrial antioxidant protein expression, i.e., MnSOD

and thioredoxin 2. These observations were not totally

unexpected since it has been also observed from studies

utilizing a constitutive gene expression system to over-

express MnSOD that MnSOD-overexpressing clones often

had alterations in their AOE profiles. These changes were

considered as adaptive phenotypes of MnSOD-overexpress-

Fig. 5. Temporal profile of characteristic alterations in hms1 cells after inducible ov

study are summarized. Time scale represents time points after initiation of MnS

previous study are shown with arrow (above the time scale); main results from th

represent kinetics of altered mitochondrial characteristics with density gradient re

mitochondrial radicals and/or mitochondrial redox status where the most significa

induction. The middle bar depicts alterations in mitochondria directly due to elevate

bottom bar represents resultant adaptive changes in mitochondria such as modified

profiles. DCm, mitochondrial membrane potential; EM, quantitative immunogold

ing cells to accommodate elevated H2O2 production due to

increased MnSOD activity [8,9,22,23]. In addition,

increased expression of the endogenous sod2 gene follow-

ing constitutive overexpression of exogenous sod2 cDNA

was previously reported [24].

As illustrated in Fig. 5, we documented from our

previous study using hms1 cells [11] that steady-state levels

of ROS were increased as early as 3 h after initiation of

MnSOD induction. However, they decreased promptly at

6 h, reaching new steady-state levels of ROS at approx-

imately 50% lower levels than control. Alteration(s) in the

mitochondrial redox environment must have been present

after increased MnSOD expression since the mitochondrial

membrane potential was transiently impaired at 3 h [11].

In the current study, modulation of the mitochondrial

redox environment by MnSOD was further demonstrated by

kinetic profiles of thioredoxin 2 mRNA and protein. Dec-

reased immunoreactive protein levels of thioredoxin 2 were

detected as early as 3 h after initiation of MnSOD induction

(Fig. 4C). Western analysis measures the total amount of

thioredoxin 2 immunoreactive protein (both reduced and

oxidized forms). Therefore, the decreased amount of thio-

redoxin 2 immunoreactive protein following increased

MnSOD expression may reflect the accelerated consumption

rates of mitochondrial redox equivalents due to elevated

H2O2 production. Possible mechanism(s) for decreased

thioredoxin 2 immunoreactive protein includes faster protein

turnover and/or enhanced association with other proteins,

resulting in sequestration from the free thiol pool.

erexpression of MnSOD. Observations from the previous study [11] and this

OD induction. Alterations in hms1 cells after MnSOD induction from the

is study are described with dotted arrow (below the time scale). Three bars

presenting the severity of alteration. Top bar illustrates putative changes in

nt change coincides with the burst of ROS at 3 h after initiation of MnSOD

d MnSOD activity such as impaired mitochondrial membrane potential. The

mitochondrial antioxidant protein and/or mitochondrial antioxidant enzyme

electron microscopy.

A. Kim et al. / Free Radical Biology & Medicine 38 (2005) 644–654652

The mitochondrial-specific antioxidant, MitoQ, success-

fully blocked the decrease in the amount of thioredoxin 2

immunoreactive protein at 3 h after initiation of MnSOD

induction (Fig. 4D), further suggesting that the initial ROS

elevation promoted an oxidative redox environment in

mitochondria.

An inverse correlation between MnSOD activity and cell

proliferation has been demonstrated by several studies [8–

10,22,23]. It has been proposed that H2O2 production

resulting from the enzymatic activity of MnSOD is

responsible for inhibition of growth [10,25]. The prevention

of MnSOD-mediated growth inhibition by MitoQ treatment

demonstrated in this study (Fig. 2) also supports this

hypothesis. Initial changes in ROS resulting from elevated

MnSOD expression elicited adaptive cellular responses

from other cellular compartments: (1) cell cycle regulation

in cytoplasm [11], and (2) transcriptional regulation of a

nuclear gene encoding a mitochondrial protein (this study).

Up-regulation of sod2 and txn2 transcripts shown in this

study would ensure elimination of O2S� and peroxides in the

mitochondria in the presence of an oxidative redox environ-

ment. Both MnSOD and thioredoxin 2 are essential for cell

viability since sod2�/� or txn2�/� knockout mice showed

neonatal or embryonic lethality, respectively [2,26]. Also,

conditional knockout of thioredoxin 2 resulted in potenti-

ated mitochondria-dependent apoptosis [27]. An important

concept developed from the current study, however, is, the

presence of coordinated regulation and expression kinetics

of sod2 and txn2 genes. Mitochondria seem to have evolved

to coregulate MnSOD expression and mitochondrial H2O2-

removing capacity to prevent excessive changes in redox

status due to modulation of ROS levels as a result of

enzymatic activity of MnSOD. The level of tolerance of

mitochondria for redox disturbance (namely, mitochondrial

redox threshold) appears to be low as demonstrated by rapid

ROS fluctuations after MnSOD induction [11] and consid-

ering how promptly sod2 and txn2 transcripts increased

following MnSOD induction as shown in the current study.

Temporal patterns of MnSOD expression, changes in

ROS levels, and alterations in sod2 and txn2 transcripts

suggested that only a minimal alteration in MnSOD

expression, if not accompanied by H2O2–removing capacity,

is required to elicit significant biochemical changes. Up to

9 h after initiation of MnSOD induction, western analysis

failed to detect statistically significant increases in immu-

noreactive MnSOD protein (Fig. 4A), although a more

sensitive ultrastructural quantitative approach using immu-

nogold electron microscopy did document a small increase

in immunoreactive MnSOD protein in mitochondria as early

as 3 h after initiation of MnSOD induction [11]. However,

as illustrated in Fig. 5, within the first 9 h after MnSOD

induction, fluctuations in ROS and mitochondrial membrane

potential levels were manifested, and presumably as an

adaptive response, txn2 mRNAs accumulated (Fig. 1E).

Therefore, mitochondria appear to attempt to maintain

narrow optimal redox ranges. We hypothesize that MnSOD

is capable of manipulating mitochondrial redox status

utilizing not only antioxidant but also pro-oxidant catalytic

reactions, which may explain multiple levels of regulation

of MnSOD expression.

The necessity of maintaining optimal levels of MnSOD

expression is also suggested by the discrepancy between the

magnitudes of increases in mouse MnSOD mRNA and

immunoreactive protein. Elevated endogenous sod2 tran-

scripts at 12 h after induction of human MnSOD expression

were approximately 15-fold over control (Fig. 1C); how-

ever, immunoreactive MnSOD protein at 18 h after

induction, presumably reflecting elevated mouse sod2

mRNA levels at 12 h after human MnSOD induction, was

approximately 70% higher than control level (Fig. 4A).

These observations imply that expression of MnSOD

protein must be tightly controlled despite greatly increased

sod2 mRNA availability for translation.

A rapid decrease in the levels of mouse sod2 mRNA at

15 h after initiation of human MnSOD expression suggests

that posttranscriptional regulation of sod2 mRNA must be

critical (Fig. 1B). In addition, the presence of a distinctive

decrease in the levels of human sod2 mRNA at 6 h after

initiation of MnSOD induction may also suggest the

importance of posttranscriptional regulation of sod2 tran-

scripts to accommodate rapidly changing mitochondrial

redox environments (Fig. 1A). We propose that mitochon-

dria regulate the amount of MnSOD protein, therefore,

presumably MnSOD activity, at both transcriptional and

posttranscriptional levels, and that the mitochondrial redox

status drives regulation at both levels. The identification of

mechanisms utilized by mitochondria to achieve these

regulatory functions is of great importance for the develop-

ment of our understanding of the physiological roles of

MnSOD.

Another characteristic of MnSOD which allows tight

regulation of its activity in mitochondria is the product

inhibition property of this enzyme. It has been suggested

that an evolutionarily conserved hydrogen bond network in

the active site of MnSOD allows controlled H2O2 release,

protecting mitochondria from overproduction of H2O2 [28].

The importance of cellular protection against H2O2 gene-

rated in physiological cellular environments has been further

emphasized by the reversible overoxidation of peroxiredox-

ins [29,30] and identification of sestrins as mammalian

proteins involved in the regeneration of overoxidized

cytosolic peroxiredoxins [31]. Therefore, levels of MnSOD

activity in the mitochondria must be regulated within

optimal ranges; the lower limit of MnSOD activity must

be sufficient to remove mitochondrial O2S�, whereas the

upper limit should be within mitochondrial H2O2-removing

capacity.

In this study, we have demonstrated that increased

MnSOD expression, when not accompanied by simulta-

neous appropriate levels of H2O2-removing capacity, led to

changes in endogenous sod2 and txn2 transcripts;

increased sod2 mRNA levels were demonstrated to result

A. Kim et al. / Free Radical Biology & Medicine 38 (2005) 644–654 653

from the transcriptional activation of the mouse sod2 gene.

The necessity of tight regulation of MnSOD expression in

mitochondria to maintain an optimal mitochondrial redox

environment was suggested by the discrepancy in the

levels of mouse sod2 mRNA and MnSOD protein

expression. Our observations further suggest that the

mitochondrial redox state serves as an initiator of cellular

events occurring outside mitochondria, i.e., the nucleus,

and that MnSOD plays a central role in the modulation of

mitochondrial redox states via both antioxidant and pro-

oxidant properties.

Acknowledgments

The authors thank Dr. Larry W. Oberley (The University

of Iowa, Iowa City, IA) for excellent discussion and

suggestions during the course of this study. This work was

supported by a Veterans Administration Merit Review grant

(to T.D.O.).

References

[1] Okado-Matsumoto, A.; Fridovich, I. Subcellular distribution of

superoxide dismutases (SOD) in rat liver. Cu,Zn-SOD in mitochon-

dria. J. Biol. Chem. 276:38388–38393; 2001.

[2] Copin, J. -C.; Gasche, Y.; Chan, P. H. Overexpression of copper/zinc

superoxide dismutase does not prevent neonatal lethality in mutant

mice that lack manganese superoxide dismutase. Free Radic. Biol.

Med. 28:1571–1576; 2000.

[3] Cadenas, E.; Davies, K. J. A. Mitochondrial free radical generation,

oxidative stress, and aging. Free Radic. Biol. Med. 29:222–230;

2000.

[4] Mikkelsen, R. B.; Wardman, P. Biological chemistry of reactive

oxygen and nitrogen and radiation-induced signal transduction

mechanisms. Oncogene 22:5734–5754; 2003.

[5] Nulton-Persson, A. C.; Szweda, L. I. Modulation of mitochondrial

function by hydrogen peroxide. J. Biol. Chem. 276:23357–23361;

2001.

[6] Zorov, D. B.; Filburn, C. R.; Klotz, L. -O.; Zweier, J. L.; Sollott, S. J.

Reactive oxygen species (ROS)-induced ROS release: A new

phenomenon accompanying induction of the mitochondrial perme-

ability transition in cardiac myocytes. J. Exp. Med. 192:1001–1014;

2000.

[7] Oberley, L. W. Anticancer therapy by overexpression of superoxide

dismutase. Antioxid. Redox Signal. 3:461–472; 2001.

[8] Li, N.; Oberley, T. D.; Oberley, L. W.; Zhong, W. Overexpression of

manganese superoxide dismutase in DU145 human prostate carcinoma

cells has multiple effects on cell phenotype. Prostate 35:221–233;

1998.

[9] Li, N.; Oberley, T. D.; Oberley, L. W.; Zhong, W. Inhibition of cell

growth in NIH/3T3 fibroblasts by overexpression of manganese

superoxide dismutase: Mechanistic studies. J. Cell. Physiol. 175:

359–369; 1998.

[10] Rodriguez, A. M.; Carrico, P. M.; Mazurkiewicz, J. E.; Melendez, J.

A. Mitochondrial or cytosolic catalase reverses the MnSOD-depend-

ent inhibition of proliferation by enhancing respiratory chain activity,

net ATP production, and decreasing the steady state levels of H2O2.

Free Radic. Biol. Med. 29:801–813; 2000.

[11] Kim, A.; Zhong, W.; Oberley, T. D. Reversible modulation of cell

cycle kinetics in NIH/3T3 mouse fibroblasts by inducible over-

expression of mitochondrial manganese superoxide dismutase. Anti-

oxid. Redox Signal. 6:489–500; 2004.

[12] Guo, G.; Yan-Sanders, Y.; Lyn-Cook, B. D.; Wang, T.; Tamae, D.;

Ogi, J.; Khaletskiy, A.; Li, Z.; Weydert, C.; Longmate, J. A.; Huang,

T. -T.; Spitz, D. R.; Oberley, L. W.; Li, J. J. Manganese superoxide

dismutase-mediated gene expression in radiation-induced adaptive

responses. Mol. Cell. Biol. 23:2362–2378; 2003.

[13] Smith, R. A.; Kelso, G. F.; James, A. M.; Murphy, M. P. Targeting

coenzyme Q derivatives to mitochondria. Methods Enzymol. 382:

45–67; 2004.

[14] Kelso, G. F.; Porteous, C. M.; Coulter, C. V.; Hughes, G.; Porteous,

W. K.; Ledgerwood, E. C.; Smith, R. A. J.; Murphy, M. P. SelecQ

tive targeting of a redox-active ubiquinone to mitochondria within

cells. Antioxidant and antiapoptotic properties. J. Biol. Chem. 276:

4588–4596; 2001.

[15] Zhang, H. J.; Yan, T.; Oberley, T. D.; Oberley, L. W. Comparison of

effects of two polymorphic variants of manganese superoxide

dismutase on human breast MCF-7 cancer cell phenotype. Cancer

Res. 59:6276–6283; 1999.

[16] Liu, W.; Saint, D. A. A new quantitative method of real time reverse

transcription polymerase chain reaction assay based on simulation of

polymerase chain reaction kinetics. Anal. Biochem. 302:52–59; 2002.

[17] Wang, X.; Phelan, S. A.; Forsman-Semb, K.; Taylor, E. F.; Petros, C.;

Brown, A.; Lerner, C. P.; Paigen, B. Mice with targeted mutation of

peroxiredoxin 6 develop normally but are susceptible to oxidative

stress. J. Biol. Chem. 278:25179–25190; 2003.

[18] Halliwell, B.; Whiteman, M. Measuring reactive species and oxidative

damage in vivo and in cell culture: How should you do it and what do

the results mean? Br. J. Pharmacol. 142:231–255; 2004.

[19] Sch7fer, M.; Sch7fer, C.; Ewald, N.; Piper, H. M.; Noll, T. Role of

redox signaling in the autonomous proliferative response of endothe-

lial cells to hypoxia. Circ. Res. 92:1010–1015; 2003.

[20] Jauslin, M. L.; Meier, T.; Smith, R. A.; Murphy, M. P. Mitochondria-

targeted antioxidants protect friedreich ataxia fibroblasts from

endogenous oxidative stress more effectively than untargeted anti-

oxidants. FASEB J. 17:1972–1974; 2003.

[21] Beauchamp, C.; Fridovich, I. Superoxide dismutase: Improved

assays and an assay applicable to acrylamide gels. Anal. Biochem.

44:276–287; 1971.

[22] Li, N.; Oberley, T. D. Modulation of antioxidant enzymes, reactive

oxygen species, and glutathione levels in manganese superoxide

dismutase-overexpressing NIH/3T3 fibroblasts during the cell cycle.

J. Cell. Physiol. 177:148–160; 1998.

[23] Li, N.; Zhai, Y.; Oberley, T. D. Two distinct mechanisms for inhibition

of cell growth in human prostate carcinoma cells with antioxidant

enzyme imbalance. Free Radic. Biol. Med. 26:1554–1568; 1999.

[24] Yan, T.; Oberley, L. W.; Zhong, W.; St Clair, D. Manganese-

containing superoxide dismutase overexpression causes phenotypic

reversion in SV40-transformed human lung fibroblasts. Cancer Res.

56:2864–2871; 1996.

[25] Li, S.; Yan, T.; Yang, J. Q.; Oberley, T. D.; Oberley, L. W. The role of

cellular glutathione peroxidase redox regulation in the suppression of

tumor cell growth by manganese superoxide dismutase. Cancer Res.

60:3927–3939; 2000.

[26] Nonn, L.; Williams, R. R.; Erickson, R. P.; Powis, G. The absence of

mitochondrial thioredoxin 2 causes massive apoptosis, exencephaly,

and early embryonic lethality in homozygous mice. Mol. Cell. Biol.

23:916–922; 2003.

[27] Tanaka, T.; Hosoi, F.; Yamaguchi-Iwai, Y.; Nakamura, H.; Masutani,

H.; Ueda, S.; Nishiyama, A.; Takeda, S.; Wada, H.; Spyrou, G.; Yodoi,

J. Thioredoxin-2 (TRX-2) is an essential gene regulating mitochondria-

dependent apoptosis. EMBO J. 21:1695–1703; 2002.

[28] Davis, C. A.; Hearn, A. S.; Fletcher, B.; Bickford, J.; Garcia, J. E.;

Leveque, V.; Melendez, J. A.; Silverman, D. N.; Zucali, J.; Agarwal,

A.; Nick, H. S. Potent anti-tumor effects of an active site mutant of

human manganese-superoxide dismutase: Evolutionary conservation

of product inhibition. J. Biol. Chem. 279:12769–12776; 2004.

A. Kim et al. / Free Radical Biology & Medicine 38 (2005) 644–654654

[29] Woo, H. A.; Chae, H. Z.; Hwang, S. C.; Yang, K. -S.; Kang, S. W.;

Kim, K.; Rhee, S. G. Reversing the inactivation of peroxiredoxins

caused by cysteine sulfinic acid formation. Science 300:653–656;

2003.

[30] Woo, H. A.; Won Kang, S.; Kim, H. K.; Yang, K. -S.; Chae, H. Z.;

Rhee, S. G. Reversible oxidation of the active site cysteine of

peroxiredoxins to cysteine sulfinic acid: Immunoblot detection with

antibodies specific for the hyperoxidized cysteine-containing

sequence. J. Biol. Chem. 278:47361–47364; 2003.

[31] Budanov, A. V.; Sablina, A. A.; Feinstein, E.; Koonin, E. V.;

Chumakov, P. M. Regeneration of peroxiredoxins by p53-regulated

sestrins, homologs of bacterial AhpD. Science 304:596–600; 2004.