The FASEB Journal • Research Communication
Crucial yet divergent roles of mitochondrial redox state
in skeletal muscle vs. brown adipose tissue energetics
Ryan J. Mailloux, Cyril Nii-Klu Adjeitey, Jian Ying Xuan, and Mary-Ellen Harper1
Department of Biochemistry, Microbiology, and Immunology, Faculty of Medicine, University of
Ottawa, Ottawa, Ontario, Canada
ABSTRACT Reduced glutathione (GSH) is the major
determinant of redox balance in mitochondria and as
such is fundamental in the control of cellular bioener-
getics. GSH is also the most important nonprotein
antioxidant molecule in cells. Surprisingly, the effect of
redox environment has never been examined in skele-
tal muscle and brown adipose tissue (BAT), two tissues
that have exceptional dynamic range and that are
relevant to the development of obesity and related
diseases. Here, we show that the redox environment
plays crucial, yet divergent, roles in modulating mito-
chondrial bioenergetics in skeletal muscle and BAT.
Skeletal muscle mitochondria were found to naturally
have a highly reduced environment (GSH/GSSG⬇46),
and this was associated with fairly high (⬃40%) rates of
state 4 (nonphosphorylating) respiration and decreased
reactive oxygen species (ROS) emission. The degluta-
thionylation of uncoupling protein 3 (UCP3) following
an increase in the reductive potential of mitochondria
results in a further increase in nonphosphorylating
respiration (⬃20% in situ). BAT mitochondria were
found to have a much more oxidized status (GSH/
GSSG⬇13) and had basal reactive oxygen species emis-
sion that was higher (⬃250% increase in ROS genera-
tion) than that in skeletal muscle mitochondria. When
redox status was subsequently increased (i.e., more
reduced), UCP1-mediated uncoupling was more sensi-
tive to GDP inhibition. Surprisingly, BAT was found to
be devoid of glutaredoxin-2 (Grx2) expression, while
there was abundant expression in skeletal muscle.
Taken together, these findings reveal the importance of
redox environment in controlling bioenergetic func-
tions in both tissues, and the highly unique character-
istics of BAT in this regard.—Mailloux, R. J., Adjeitey,
C. N.-K., Xuan, J. Y., and Harper, M.-E. Crucial yet
divergent roles of mitochondrial redox state in skeletal
muscle vs. brown adipose tissue energetics. FASEB J. 26,
363–375 (2012). www.fasebj.org
Key Words: glutathione 䡠 thermogenesis 䡠 uncoupling proteins
By virtue of its high concentration (⬃10 mM) and
powerful reductive potential (E0⫽⫺240 mV), reduced
glutathione (GSH) is the major redox buffer in mito-
chondria (1, 2). The redox buffering capacity of GSH
allows this molecule to function as the principal non-
0892-6638/12/0026-0363 © FASEB
protein antioxidant species in cells (3). Reactive oxygen
species (ROS) are a normal byproduct of fuel metabo-
lism in cells and are generated predominantly in mito-
chondria. As the singlet electrons of ROS can damage
cellular constituents and functions, there are many
enzymatic and nonenzymatic mechanisms to keep cel-
lular ROS in check. The GSH-mediated scavenging of
ROS generates inactive glutathione disulfide (GSSG),
which, in turn, is rapidly converted back to GSH by
NADPH and glutathione reductase (3). This antioxida-
tive system allows mitochondria to maintain a high
GSH/GSSG ratio and buffer locally generated ROS
during metabolic processes. Recent work has also
shown that GSH can regulate the redox environment of
mitochondria by forming disulfide bridges with ex-
posed thiols on proteins, a process referred to as
glutathionylation (4 –7). Very recent studies reveal that
this type of reversible reaction is a key post-translational
control mechanism required to modulate protein func-
tion in response to changes in redox status (8, 9).
Fluctuations in ROS induce the glutathionylation and
deglutathionylation of proteins, and the processes are
highly specific (5, 10). For example, certain proteins,
such as actin, SDH, uncoupling protein 2 (UCP2), and
UCP3, are deglutathionylated by ROS (5, 11, 12). In
contrast, other proteins, including some electron trans-
port chain respiratory complexes and tricarboxylic acid
cycle enzymes, are controlled by glutathionylation in
response to ROS (13–16). More overt oxidative stres-
sors that induce a great decrease in GSH/GSSG can
prompt the spontaneous hyperglutathionylation of pro-
teins, resulting in metabolic dysfunction (17).
Two tissues that have remarkable degrees of meta-
bolic activity and flexibility are skeletal muscle and
brown adipose tissue (BAT). Both have the fairly
unique ability to oxidize a range of carbon sources to
satisfy energy demands. Regardless of energy source,
the graded transfer of electrons from NADH and
FADH2 to diatomic oxygen (O2) is used to form the
1
Correspondence: University of Ottawa, Faculty of Medi-
cine, Department of Biochemistry, Microbiology, and Immu-
nology, 451 Smyth Road, Ottawa, ON, Canada, K1H 8M5.
E-mail: maryellen.harper@uottawa.ca
doi: 10.1096/fj.11-189639
This article includes supplemental data. Please visit http://
www.fasebj.org to obtain this information.
363
protonmotive force across the mitochondrial inner
membrane (MIM) (18). Protonmotive force can then
be used to drive ATP production (coupled respiration)
or can be dissipated by proton-leak reactions (nonphos-
phorylating, uncoupled respiration) (18). Proton leak-
age is mediated through uncoupling protein (UCP)
and non-UCP mechanisms (19). Recent work has indi-
cated that brown fat and skeletal muscle cells can share
the same ancestral cellular lineage (20). In fact, Sca1⫹
progenitor cells found in white fat and skeletal muscle
can easily be induced to differentiate into brown adi-
pocytes, while those in brown fat are constitutively
committed as brown adipocyte precursors (21). Despite
their similarity in fuel oxidation and common cell
lineages, substrate oxidation in skeletal muscle and
BAT serves very different purposes. In skeletal muscle,
substrate oxidation is mostly used to support coupled
respiration to generate ATP for contractile functions
(22). However, it must be kept in mind that uncoupled
respiration in skeletal muscle can account for ⬃20 –
40% of energy expenditure in this tissue, and its
quantitative importance is greatest when ATP demands
of muscle are low (e.g., when muscle is at rest; refs. 22,
23). In BAT, substrate oxidation rates are very high
when UCP1 is activated to fulfill this tissue’s essential
role in thermoregulation (24, 25). Under conditions of
thermoneutrality, BAT is inactive, and only basal cellu-
lar reactions occur (26). However, when an animal is
exposed to the cold, BAT metabolic activity increases
substantially to support thermogenesis (27, 28). While
UCP3 is also expressed in BAT, UCP3 is the only UCP
protein expressed in muscle. UCP3-mediated proton
leakage is thought to be activated by ROS to provide a
negative feedback loop mitigating further ROS produc-
tion by mitochondria (29, 30). We recently demon-
strated that UCP3, unlike UCP1, is controlled by revers-
ible glutathionylation/deglutathionylation, a process
induced by ROS (11).
Although mitochondrial metabolism in both tissues
has been investigated extensively, very little informa-
tion exists on the influence of redox state. This is
particularly true for BAT, a tissue that has recently
attracted a great deal of attention since it has been
unequivocally demonstrated that BAT exists in adult
humans (31–33). Here, we examined the effect of the
redox environment on bioenergetics and ROS in skel-
etal muscle and BAT mitochondria. We report differ-
ences in redox environment and ROS production and
their control in these two tissues. We further demon-
strate that these differences were matched by variations
in GSH and GSSG levels.
MATERIALS AND METHODS
Mice and tissue extraction
Male C57BL6J mice (9 –12 wk of age) fed a standard ad libitum
diet (44.2% carbohydrate, 6.2% fat, 18.6% crude protein; diet
T.2018; Harlan, Indianapolis, IN, USA) were used for all
364 Vol. 26 January 2012 The FASEB Journal 䡠 www.fasebj.org
determinations. Mice were housed either at room tempera-
ture (RT; 23°C) or 4°C for 24 h prior to experimentation. All
experiments on skeletal muscle mitochondria were per-
formed on mice housed at RT. During cold exposure, mice
were given free access to food and water but no housing
materials. On the day of experiments, mice were quickly
euthanized (decapitation), and the skeletal muscle and BAT
were immediately extracted and placed in ice-cold buffer with
or without 2 mM DTT. All experiments were performed
according to the principles and guidelines of the Canadian
Council of Animal Care and the study was approved by the
Animal Care Committee of the University of Ottawa.
Mitochondrial isolation
Buffers were ice-cold, and procedures were performed on ice
or at 4°C to maintain mitochondrial redox status and func-
tional characteristics. To test the effect of redox state on
skeletal muscle and BAT mitochondria, isolations were per-
formed in the presence or absence of 2 mM DTT. DTT is a
strong reducing agent that maintains a more reduced GSH
pool in mitochondria (7). Notably, however, DTT was not
included in any of the mitochondrial resuspension and assay
buffers. Muscle mitochondria were isolated as described by
Mailloux et al. (11). Skeletal muscle from forelimbs and
hindlimbs was dissected, cleaned of connective tissue and fat,
and placed in buffer A (140 mM KCl, 20 mM HEPES, 5 mM
MgCl2, and 1 mM EGTA; pH 7.0). Muscle was weighed and
minced on a Teflon board and then placed in 30 ml of
homogenizing medium (buffer A supplemented with 1 mM
ATP and 1% BSA, w/v) containing 1 U of protease (subtilisin
A). Tissue was homogenized using a glass/Teflon Potter-
Elvehjem tissue grinder and then centrifuged at 800 g (9 min,
4°C) to remove undisrupted tissue and whole cells. The
supernatant was centrifuged at 12,000 g (9 min, 4°C), and the
resulting pellet was resuspended in 1 ml of buffer A and
incubated on ice for 5 min to allow myofibrillar repolymer-
ization. Samples were then diluted to 30 ml with buffer A and
centrifuged at 800 g (9 min, 4°C) to remove myofibrils.
Finally, the supernatant was centrifuged at 12,000 g (9 min,
4°C) to pellet the mitochondria. The final pellet was resus-
pended in buffer A devoid of DTT.
BAT mitochondria were isolated as described by Mailloux
et al. (11). Briefly, interscapular BAT was removed, cleaned of
white adipose tissue, and placed in 30 ml of 250 mM sucrose
solution. The tissue was minced and homogenized using a
glass/Teflon Potter-Elvehjem tissue grinder in 30 ml of 250
mM sucrose solution. The homogenate was centrifuged at
8500 g (9 min, 4°C), and the pellet was resuspended in 250
mM sucrose plus 0.2% BSA and spun at 800 g (9 min, 4°C).
The supernatant was centrifuged at 8500 g (9 min, 4°C), and
the mitochondrial pellet was resuspended in buffer B (125
mM sucrose, 25 mM Trizma base, 5 mM MgCl2, and 1 mM
EGTA, pH 7.4) devoid of DTT. Protein concentration was
determined by Bradford assay using BSA as the standard.
Mitochondrial preparations were kept on ice, and all exper-
iments were conducted within 2 h following isolation. Exper-
iments using skeletal muscle mitochondria were conducted in
buffer A. All BAT mitochondria experiments were performed
using buffer B and performed under state 4 respiratory
conditions. Aliquots of mitochondria were also used for
HPLC analyses or stored at ⫺80°C for later immunoblot
determinations.
Mitochondrial membrane potential (⌬⌿m) determinations
Kinetic determinations were conducted in isolated mitochon-
dria and the fluorophore, safranin O (5 ␮M; excitation 485
MAILLOUX ET AL.
nm, emission 580 nm) using a BioTEK FLx800 reader and
Gen5 software (BioTek Instruments, Inc., Winooski, VT,
USA). Mitochondria (0.5 mg/ml) were preincubated in ei-
ther buffer A or buffer B at 37°C prior to initiating experi-
ments. For skeletal muscle, mitochondria were assayed under
state 3 and state 4 conditions. Mitochondria were energized
for 1.5 min with 5 mM succinate and then deenergized with
0.5 mM ADP for 4 min. Mitochondria were then treated with
1.3 ␮g/ml of oligomycin to assess the effect of redox environ-
ment on state 4 respiration. Rates of change in fluorescence
(AU/min/mg protein) were then calculated for state 3 and
state 4 conditions, respectively. For BAT mitochondria, the
effect of redox environment on UCP1 function was tested by
GDP and oleic acid titrations. Immediately prior to the assay,
mitochondria were preincubated in 1.3 ␮g/ml oligomycin to
induce state 4 (nonphosphorylating) respiratory conditions;
subsequent changes in membrane potential were probed
using UCP1 inhibitors (GDP) and activators (oleate). Mito-
chondria were treated with 5 mM pyruvate and 3 mM malate
for 1 min, and then GDP was added incrementally over time.
Oleic acid was then titrated to test UCP1 reactivation (incre-
ments of 0.1 mM). Fresh oleate was used in every experiment.
The time of addition for each chemical is indicated in the
figures. As a control, safranin O fluorescence in the presence
of mitochondria and in the absence of substrate was per-
formed immediately before each experiment.
Oxygen consumption measurements
BAT mitochondria (0.5 mg/ml) isolated in the presence or
absence of DTT at 37°C were studied in a Clark-type oxygen
electrode (Hansatech, Norfolk, UK). Prior to determinations,
mitochondria were warmed for 10 min in the electrode
chamber in buffer B at 37°C. Buffer B was assumed to contain
406 nmol O/ml. State 4 respiratory conditions were then
induced by adding oligomycin (1.3 ␮g/ml) to the chamber.
State 4 respiration was then tested following the addition of 5
mM pyruvate and 3 mM malate. Experiments were performed
under state 4 conditions to assess UCP1-mediated respiration
only. Mitochondria were then treated acutely with GDP (final
concentration of 1 mM), followed by a treatment with oleate
(final concentration of 0.1 mM). Measurements of oxygen
consumption rate (OCR) were completed within an 8- to
12-min period.
Mitochondrial ROS emission determinations
ROS emission experiments were conducted using the BioTEK
SynergyMX microplate reader and Gen5 software. Mitochon-
dria isolated in the presence or absence of DTT were tested
kinetically for ROS emission in the presence of different
substrates and allosteric regulators. ROS production was
assessed fluorometrically using 20 ␮M DCF diacetate (DCFH-
DA). For all experiments, mitochondria were prewarmed in
the requisite buffer for 10 min at 37°C. Baseline measure-
ments of DCFH-DA fluorescence were performed in the
presence of mitochondria prior to addition of substrates. For
skeletal muscle mitochondria (0.5 mg/ml), ROS production
was tested in the presence of succinate (5 mM) or pyruvate/
malate (5 mM/3 mM). ROS production was then tested
under oligomycin-induced state 4 conditions (1.3 ␮g/ml).
ROS emission in BAT mitochondria was tested in buffer B
under oligomycin-induced state 4 conditions. BAT mitochon-
dria (0.5 mg/ml) were treated with 5 mM pyruvate and 3 mM
malate, and then ROS emission was ascertained following the
addition of GDP (1 mM) and oleate (0.1 mM). Experiments
were conducted for 20 –30 min. Rates of change in ROS
production (AU/min/mg protein) were calculated after the
REDOX MODULATION OF MITOCHONDRIAL METABOLISM
addition of substrates or allosteric regulators. Data were
normalized to background fluorescence and mitochondrial
concentration.
Western blot analysis
Levels of glutaredoxin-1 (Grx1), glutaredoxin-2 (Grx2), and
thioredoxin-2 (Trx2) were tested by immunoblot. Immedi-
ately following isolation, mitochondria were diluted to 1
mg/ml in Laemmli buffer and then electrophoresed on a
10% isocratic SDS gel. On completion, gel slabs were equili-
brated in transfer buffer and then electroblotted onto nitro-
cellulose membranes at RT for 1 h at 100 V with a maximum
amperage of 400 mA. The membranes were blocked with
Tris-buffered saline containing 1% (v/v) Tween-20 (TBS-T).
Membranes were then probed for 24 h at 4°C with primary
antibodies directed against Grx1 (1:500; Abcam, Cambridge,
MA, USA), Grx2 (1:500; Abcam), Trx2 (1:500; Abcam), UCP1
(1:10,000; Sigma, St. Louis, MO, USA), and UCP3 (Abcam;
1:1000). For UCP1, membranes were probed for only 1 h at
RT. SDH served as the loading standard (1:2000, SDHA
antibody; Santa Cruz Biotechnology, Santa Cruz, CA, USA).
Membranes were then incubated for 1 h at RT with the
requisite horseradish peroxidase-conjugated secondary anti-
bodies (anti-rabbit, 1:2000; Santa Cruz Biotechnology). Blots
were incubated in enhanced chemiluminescent substrate for
visualization (ECL kit; Thermo Scientific, Waltham, MA,
USA).
Protein-glutathione (PSSG) adduct detection
Following mitochondrial isolation and determination of pro-
tein content, mitochondria were washed once with either
buffer A (skeletal muscle) or buffer B (BAT) and then treated
with Laemmli buffer containing 25 mM N-ethylmaleimide
(NEM). The NEM was included to prevent spontaneous
glutathionylation events and to quench any residual DTT.
Protein (30 ␮g) was electrophoresed in a 10% linear gradient
SDS gel under standard conditions. On completion, gel slabs
were incubated in transfer buffer and then used for electro-
blotting, as described above. Membranes were blocked for 1 h
at RT in TBS-T containing 5% (w/v) nonfat skim milk. PSSG
adducts were detected using the anti-PSSG (1:500; Virogen,
Boston, MA, USA). Following a 24-h incubation in primary
antibody at 4°C, membranes were probed for 1 h at RT with
anti-mouse horseradish peroxidase-conjugated secondary an-
tibody (1:2000, Santa Cruz Biotechnology). Immunoreactive
bands were detected as described above. Blots were quanti-
fied using ImageJ software (U.S. National Institutes of Health,
Bethesda, MD, USA).
GSH/GSSG determinations
Mitochondrial GSH and GSSG levels were assessed by HPLC.
Following one wash in either buffer A or buffer B, mitochon-
dria isolated from skeletal muscle and BAT were diluted
to 0.5 mg/ml in ice-cold 1% (w/v) trichloroacetic acid
(TCA)/1% (w/v) meta-phosphoric acid solution and incu-
bated on ice for 10 min. Precipitate was removed by centrif-
ugation for 10 min at 12,000 g at 4°C. The supernatant was
then collected and stored at ⫺80°C. On the day of experi-
ments, samples were injected into an Agilent 1100 Series
HPLC equipped with a Pursuit C18 column (150⫻4.6 mm, 5
␮m; Agilent Technologies, Santa Clara, CA, USA). For proper
separation of GSH and GSSG, a flow rate of 1 ml/min was
used, with a mobile phase consisting of 0.1% (w/v) TCA and
HPLC-grade methanol in a 90:10 ratio. GSH and GSSG were
detected using an Agilent UV-Vis variable-wavelength detec-
365
tor operating at 215 nm. Retention times were confirmed by
injecting GSH and GSSG standards. GSH and GSSG were
quantified using Agilent Chemstation software.
In situ measurement of bioenergetics
The XF24 Extracellular Flux Analyzer (Seahorse Bioscience,
North Billerica, MA, USA) was employed to determine the
mitochondrial bioenergetic characteristics in primary myo-
tubes incubated in the presence or absence of 2 mM DTT.
Primary myoblasts were isolated and purified from wild-type
C57BL6J mice, as described by Rando et al. (34). Myoblasts
were subcultured and maintained in DMEM containing 10%
(v/v) FBS, 5 mM glucose, 1 mM pyruvate, 2% (v/v) antimy-
cotics-antibiotics (AA), and 2.5 ng/ml fibroblast growth fac-
tor (FGF). For experiments, primary myoblasts were seeded at
50,000 cells/ml and grown in Matrigel-coated 24-well Sea-
horse XF24 culture plates. On reaching ⬃80% confluency,
cells were differentiated for up to 5 d in DMEM containing
2% (v/v) FBS, 5 mm glucose, 1 mM pyruvate, and 2% AA. On
the day of experiments, differentiated cells were incubated
for 45 min at 37°C in HCO3-free DMEM containing 5 mM
glucose, 4 mM l-glutamine, and 1 mM pyruvate (Seahorse
medium) in the absence or presence of 2 mM DTT at
ambient CO2. The medium was then aspirated, cell monolay-
ers were washed once with warmed PBS, and then fresh
prewarmed DTT-free Seahorse medium was added. Fluori-
metric sensors enabled the sensitive in situ measurement of
OCR and extracellular acidification rate (ECAR) in the
microenvironment surrounding the cells. Measurements of
OCR and ECAR were performed for 3 measurement intervals
to assess basal rates (1 measurement interval consists of a
2-min mixing, 2-min incubation, and 2-min measurement
step). The effect of DTT on state 4 respiration was ascertained
by injecting oligomycin (1.3 ␮g/ml). State 4 respiration was
tested over 3 measurement intervals. OCR and ECAR were
normalized to total cellular protein/well using the Bradford
assay.
To assess whether DTT activates UCP3 in situ, C2C12 cells
were grown [DMEM, 5 mM glucose, 1 mM pyruvate, 10%
(v/v) FBS, and 2% (v/v) AA] and differentiated [DMEM, 5
mM glucose, 1 mM pyruvate 2% (v/v) FBS, and 2% (v/v) AA]
on Matrigel-coated Seahorse plates (35). On full differentia-
tion, cells were transiently transfected with pCMV-UCP3
plasmid (Origene, Rockville, MD, USA) using Lipofectamine
2000 (Invitrogen, Carlsbad, CA, USA), according to the
manufacturer’s instructions. Transfection protocols were op-
timized to avoid UCP3 overexpression and resultant uncon-
trolled proton leakage. Following a 36-h transfection, cells
were washed twice with PBS and incubated in Seahorse
medium containing 2 mM DTT, as described above. Cells
incubated in the absence of DTT served as the control.
Immediately prior to loading the plate in the Seahorse XF-24,
the medium was removed and replaced with fresh prewarmed
DTT-free Seahorse medium. Basal and state 4 respiration
TABLE 1. Absolute levels of GSH and GSSG in skeletal muscle and BAT mitochondria isolated in the presence or absence of DTT
Skeletal muscle
Parameter ⫺DTT ⫹DTT
GSH (nmol/mg protein) 31.50 ⫾ 0.021 34.50 ⫾ 0.028*
GSSG (nmol/mg protein) 0.709 ⫾ 0.014 0.438 ⫾ 0.004*
BAT mitochondria were collected from mice exposed to RT and 4°C conditions. Absolute GSH and GSSG levels were determined by HPLC,
as described in Materials and Methods. Statistical significance was determined by comparing ⫺DTT and ⫹DTT; n ⫽ 4. *P ⬍ 0.05, **P ⬍ 0.005;
Student’s t test.
366 Vol. 26 January 2012 The FASEB Journal 䡠 www.fasebj.org
rates were tested as described above. Data were normalized to
protein content/well. State 4 respiration was expressed as a
percent of the basal respiration.
For diamide titrations, C2C12 cells were grown, differenti-
ated, transfected, and then treated with or without 2 mM
DTT, as described above. Cells were transfected with either
pCMV-UCP3 or empty pCMV vector. Cells were tested for the
presence or absence of UCP3 by immunoblot (Supplemental
Fig. S1). Immediately prior to loading the cell plate into the
Seahorse XF-24, medium was removed and replaced with
fresh prewarmed DTT-free Seahorse medium containing 1.3
␮g/ml of oligomycin to induce state 4 respiratory conditions.
Following assessment of state 4 respiration, diamide was
injected sequential to final concentrations of 50, 100, 200,
and 300 ␮M, respectively. OCR was tested for two intervals
following each injection, each including 2 min of mixing, 2
min of incubation, and 2 min of measurement. Data were
normalized to protein content/well.
Statistical analysis
Student’s t tests were performed using Excel software
(Microsoft, Redmond, WA, USA). Results are expressed as
means ⫾ se unless indicated otherwise in the figure
captions.
RESULTS
Glutathione levels are drastically different between
skeletal muscle and BAT mitochondria and are
differentially affected by the reducing agent, DTT
The absolute GSH and GSSG levels in skeletal muscle
and BAT mitochondria (from RT mice) treated or not
treated with DTT are presented in Table 1. Comparing
the GSH/GSSG of skeletal muscle and BAT revealed
that the GSH pool in BAT mitochondria was far more
oxidized than skeletal muscle. This would indicate that,
in contrast to skeletal muscle, the BAT mitochondrial
environment is far more oxidizing. Indeed, isolation of
skeletal muscle mitochondria in DTT increased the
GSH/GSSG ratio substantially (Fig. 1A). BAT mito-
chondria isolated from mice exposed to RT conditions
and treated with DTT also contained a more reduced
GSH pool in comparison to mitochondria isolated in
the absence of DTT (Fig. 1A).
Changes in the GSH/GSSG ratio are often accompa-
nied by alterations in the levels of PSSG adducts.
Skeletal muscle mitochondria isolated in the presence
of DTT displayed a significant 36% decrease in PSSG
BAT, RT BAT, 4°C
⫺DTT ⫹DTT ⫺ DTT ⫹DTT
30.73 ⫾ 0.002 34.56 ⫾ 0.003** 9.43 ⫾ 0.40 9.56 ⫾ 0.79
2.417 ⫾ 0.001 1.242 ⫾ 0.006** 1.66 ⫾ 0.36 1.41 ⫾ 0.34
MAILLOUX ET AL.
Figure 1. Effects of DTT pretreatment on redox environment in skeletal muscle and BAT mitochondria. BAT mitochondria were
isolated from mice exposed to RT conditions. Mitochondria from skeletal muscle and BAT were isolated in the presence of 2
mM DTT, as described in Materials and Methods. Mitochondria isolated in the absence of DTT served as the control. A) Effect
of DTT pretreatment on the mitochondrial GSH/GSSG ratio. Immediately following isolation, mitochondria were diluted to 0.5
mg/ml in a solution containing 1% (w/v) TCA/1% (w/v) meta-phosphoric acid and then clarified for HPLC analysis of GSH
and GSSG levels. GSH and GSSG retention times were confirmed, and levels were quantified by injecting varying amounts of
standard solutions. Following quantification, the GSH/GSSG ratio for skeletal muscle and BAT pretreated with DTT was
determined. Absolute amounts of mitochondrial GSH and GSSG are shown in Table 1; n ⫽ 5. B) Assessment of PSSG levels in
skeletal muscle and BAT mitochondria. Following isolation, 30 ␮g of mitochondrial protein was electrophoresed and tested for
PSSG levels using an anti-PSSG antibody (virogen). Blots were subsequently quantified using ImageJ software; n ⫽ 4.
C) Immunoblot analysis for enzymes involved in reversible glutathionylation. Mitochondrial protein (60 ␮g) was electropho-
resed and tested for Grx1, Grx2, and Trx2 levels, respectively. SDH served as the loading control. Values are expressed as
means ⫾ sd, using Student’s t test.

adducts (P⬍0.01; Fig. 1B). Immunoreactive bands in
the low-molecular-weight range (⬃34 kDa) displayed
the most dramatic decreases in intensity when skeletal
muscle mitochondria were isolated in DTT. In contrast,
DTT decreased PSSG levels by 27% in BAT mitochon-
dria, but this did not reach statistical significance (Fig.
1B). There were apparently a number of proteins that
were glutathionylated in BAT mitochondria; however,
DTT pretreatment had less of an effect on decreasing
the amount of PSSG adducts when compared to effects
in skeletal muscle. We next tested the levels of putative
glutathionylating enzymes in mitochondria from both
tissues. In Fig. 1C, skeletal muscle mitochondria con-
tained Grx1, Grx2, and Trx2 proteins. In contrast, only
Grx1 and Trx2 were detected in BAT mitochondria.
There also appeared to be a slight shift in the mobility
of Grx1 between the skeletal muscle and BAT mito-
chondria. BAT mitochondria contained more Trx2,
which is thought to play a greater role in limiting
oxidative damage than in glutathionylation (36). Over-
all, these findings reveal that the mitochondrial redox
state and key redox enzymes are substantially different
between skeletal muscle and BAT.
A more reductive mitochondrial environment
increases UCP3-mediated proton leakage in skeletal
muscle
Protein glutathionylation is often associated with
changes in redox state, and we recently demonstrated
that UCP3-mediated proton leakage is activated by the
deglutathionylation of UCP3 (9, 11). Hence, we exam-
ined the effect of the presence and absence of DTT on
mitochondrial bioenergetics. First, we tested the effect
of DTT on skeletal muscle mitochondrial membrane
potential using safranin O. Mitochondria were initially
energized for ⬃1.5 min with 5 mM succinate, and then
ADP and oligomycin were added sequentially to induce
state 3 and 4 respiratory conditions, respectively. While
there was a trend for an increase, the pretreatment
of mitochondria with DTT had no significant effect
on rate of mitochondrial membrane depolarization
under state 3 respiratory conditions (ADP⫹succinate;
Fig. 2A). Examination of membrane potentials under
oligomycin-induced state 4 respiratory conditions did,
however, reveal that DTT pretreatment diminished the
rate of membrane potential repolarization, consistent
with the induction of leak-dependent processes (Fig.
2B). The increase in state 4 respiration was confirmed
in situ using primary myotubes from wild-type mice.
Indeed, preincubation in DTT led to a significant
increase in state 4 respiration in cells (Fig. 2C). Assess-
ment of the cellular ECAR revealed that DTT treatment
diminished glycolytic metabolism (Fig. 2C). We then
determined whether DTT pretreatment diminished ROS
production from the skeletal muscle mitochondria (Fig.
2D). No significant changes in the rate of ROS production
were observed following succinate addition (slope⫺DTT⫽
36.98⫾14.05 vs. slope⫹DTT⫽25.83⫾14.79, P⫽0.1, Stu-
dent’s t test). However, DTT pretreatment significantly

REDOX MODULATION OF MITOCHONDRIAL METABOLISM 367

 Figure 2. A more reducing redox environment increases state 4 respiration and
decreases ROS production in skeletal muscle mitochondria. Mitochondria from
skeletal muscle were isolated in the presence of 2 mM DTT, as described in
Materials and Methods. Mitochondria isolated in the absence of DTT served as
the control. A) Effect of alterations in redox environment on mitochondrial
membrane potential under state 3 conditions. Mitochondrial membrane poten-
tial was determined using safranin O. Mitochondria (0.5 mg/ml) from skeletal
muscle were energized with 5 mM succinate and then treated with ADP (0.5
mM). Rate of change in fluorescence following ADP addition was then calcu-
lated; n ⫽ 5. B) Effect of alterations in redox environment on mitochondrial
membrane potential under state 4 conditions. Mitochondria (0.5 mg/ml) from
skeletal muscle that were energized with succinate (5 mM) and treated with ADP
(0.5 mM) were then treated with oligomycin (1.3 ␮g/ml). Rate of change in
fluorescence following oligomycin addition was then calculated; n ⫽ 5. C) Pretreatment of primary myotubes isolated
from wild-type (WT) mice with DTT increases oligomycin-induced state 4 respiration. Primary myotubes grown and
differentiated in Seahorse tissue culture plates were treated with 2 mM DTT for 45 min, and then OCR and ECAR were
tested in situ using the Seahorse Extracellular Flux Analyzer. Following the assessment of OCR and ECAR under basal
conditions, OCR and ECAR were tested under oligomycin-induced state 4 conditions. Data are expressed as a
percentage of basal conditions. OCR was normalized to protein content/well; n ⫽ 4. D) A more reductive environment
diminishes state 4 ROS generation. Mitochondria (0.5 mg/ml) were tested for ROS emission using DCFH-DA (20 ␮M).
ROS emission was tested kinetically with succinate (5 mM) and oligomycin (1.3 ␮g/ml). Rate of ROS production
following the addition of substrate or oligomycin was then calculated; n ⫽ 5. Values are expressed as means ⫾ se, using
Student’s t test.

diminished the rate of ROS production during oligomy-
cin-induced state 4 conditions (after oligomycin treat-
ment, slope⫺DTT⫽71.41⫾18.22 vs. slope⫹DTT⫽38.84⫾
18.19, Pⱕ0.01, Student’s t test). Rotenone treatment
diminished ROS production in mitochondria energized
by succinate and treated with oligomycin, confirming that
complex I is a major source for succinate-mediated ROS
production (Supplemental Fig. S2). However, mitochon-
dria not exposed to DTT produced more ROS following
the oligomycin and rotenone treatments. We also tested
the effect of complex I substrates (pyruvate/malate)
on the rate of ROS production from mitochondria iso-
lated in the presence or absence of DTT (Supplemental
Fig. S2). Again, ROS production following induction of
state 4 respiration with oligomycin was found to be lower
in mitochondria pretreated with DTT.
Degree of glutathionylation controls UCP3-mediated
proton leakage
The above results indicate that increased leak-depen-
dent respiration accounts, in part, for the decrease in
mitochondrial ROS generation. Hence, we decided to
test whether DTT-induced increase in leak-dependent
respiration was due to UCP3 deglutathionylation, and
whether this DTT-mediated activation of UCP3 could
be reversed by acute treatment with diamide, a gluta-
thionylation catalyst. We first determined whether DTT
could enhance leakage through UCP3 using C2C12
cells transiently transfected with pCMV-UCP3. It is
important to note that C2C12 cells do not express
endogenous UCP3 (29). C2C12 cells preincubated in
the presence or absence of DTT did not display any
changes in the basal respiration and oligomycin-in-
duced state 4 respiration (Fig. 3A). However, DTT
pretreatment significantly increased the contribution
of nonphosphorylating respiration to the total respira-
tion in the UCP3-overexpressing cells (Fig. 3B). Di-
amide titration experiments revealed that this DTT-
mediated activation of UCP3 could be reversed by
glutathionylation. For this set of experiments, C2C12
cells transiently transfected with either pCMV-UCP3 or
empty pCMV vector were preincubated in the presence
or absence of DTT, and then state 4 respiratory condi-
tions were induced with oligomycin. Titration of di-
amide up to a concentration of 200 ␮M resulted in a
progressive decrease in state 4 respiration (up to 40%)
in UCP3-overexpressing cells treated with DTT (Fig.
3C). Raising the diamide concentration to 300 ␮M had
no further effect on state 4 respiration. Diamide treat-
ment also diminished state 4 respiration in the cells
transfected with empty pCMV vector, but not to the
same extent as the UCP3-expressing cells, indicating
that other proteins involved in catalyzing proton

368 Vol. 26 January 2012 The FASEB Journal 䡠 www.fasebj.org MAILLOUX ET AL.
Figure 3. UCP3 is required for the DTT-mediated increases in state 4 respiration. For all experiments, respiratory rates were
tested in situ with the Seahorse Extracellular Flux Analyzer. A) Absolute basal and oligomycin-induced respiratory rates in C2C12
cells transiently transfected with pCMV-UCP3 treated with or without 2 mM DTT. Cells were treated with or without 2 mM DTT
45 min prior to the assay. Following assessment of basal OCR, oligomycin-induced state 4 was tested; n ⫽ 4. B) Contribution of
leak-dependent respiration to total respiration. Oligomycin-induced state 4 respiration is expressed as a percentage of basal
OCR; n ⫽ 4. C, D) Deglutathionylation of UCP3 is responsible for the DTT-mediated increases in state 4 respiration. C2C12
myotubes transiently transfected with empty pCMV vector (⫺UCP3) or pCMV-UCP3 (⫹UCP3) were treated with (C) or without
(D) 2 mM DTT for 45 min, and then state 4 conditions were induced immediately with oligomycin (1.3 ␮g/ml). Effect of
glutathionylation on state 4 respiration was then tested by diamide titration. Following assessment of state 4 respiration in
UCP3⫺/⫹ cells treated with or without DTT, cells were acutely treated with increasing amounts of diamide (50, 100, 200, and
300 ␮M, respectively). Data are expressed as a percentage of the state 4 OCR induced with oligomycin prior to diamide
treatment (0 ␮M). State 4 OCR was normalized to protein content/well; n ⫽ 6. Values are expressed as means ⫾ sd, using
Student’s t test.

leakage are also inhibited by glutathionylation (Fig.
3C). Cells not treated with DTT were also tested for
UCP3 inhibition by diamide titration. Small, but
significant, decreases in state 4 respiration in UCP3-
expressing cells were observed when treated with 100
␮M and 200 ␮M diamide (Fig. 3D). Intriguingly, at
200 ␮M diamide, only a 20% decrease in state 4
respiration was observed, indicating that pretreat-
ment with DTT may liberate cysteine residues from
glutathione in situ, making these residues available
for reglutathionylation by diamide. Moreover, bring-
ing the concentration of diamide up to 300 ␮M
actually restored oligomycin-induced state 4 respira-
tion to control levels (Fig. 3D). Hence, our results
indicate that, at least in C2C12 cells, UCP3 is partially
glutathionylated under normal conditions and that
full deglutathionylation is required for the full acti-
vation of UCP3-mediated proton leakage.
Effects of a reductive environment on BAT
mitochondrial bioenergetics
Although DTT treatment increased the GSH/GSSG
ratio in BAT mitochondria, it did not significantly
decrease the amount of PSSG adducts. We previously
demonstrated that in BAT mitochondria, unlike skel-
etal muscle mitochondria, glutathionylation appears
to play a much less important role in modulating
UCP1 (11). Hence, we tested the effect of redox
environment on the bioenergetics of BAT mitochon-
dria. Membrane potential determinations revealed
that DTT pretreatment sensitized UCP1 to inhibition
by GDP (Fig. 4A). In the mitochondria treated with
DTT, GDP induced a 2.5-fold increase in safranin O
fluorescence (indicative of large increases in mem-
brane potential), whereas GDP only increased the
fluorescence by 1.6-fold in mitochondria not treated

REDOX MODULATION OF MITOCHONDRIAL METABOLISM 369

Figure 4. Influence of a changing redox environment on UCP1 in BAT mitochondria from RT mice. Mitochondria from BAT
were isolated in the presence of 2 mM DTT as described in Materials and Methods. Mitochondria isolated in the absence of DTT
served as the control. Mitochondria were treated with oligomycin (1.3 ␮g/ml) to induce state 4 respiratory conditions. A) Effect
of alterations in redox environment on mitochondrial membrane potential. Mitochondrial energization was tested using
safranin O. Mitochondria (0.5 mg/ml) from BAT were energized at time 0 with 5 mM pyruvate and 3 mM malate, and then GDP
(G) was titrated into the reaction at 0.5 mM increments to test the effect of DTT on UCP1 inhibition. Effect of DTT on UCP1
reactivation was then tested by adding increasing amounts of oleate (OA) at 0.1 mM increments. The change in fluorescence
following the addition of GDP and oleate was calculated from the background fluorescence. *P ⱕ 0.05. B) DTT increases
sensitivity of UCP1 to GDP but not oleate. O2 consumption determinations were performed using a Clark-type electrode under
oligomycin-induced state 4 conditions. Respiration was initially tested with 5 mM pyruvate and 3 mM malate, followed by GDP
(1 mM) and oleate (0.1 mM) treatment. C) DTT pretreatment increases ROS emission in BAT mitochondria isolated from RT
mice. ROS emission was tested under oligomycin-induced state 4 respiratory conditions. ROS production was detected with
DCFH-DA (20 ␮M). Following the addition of 5 mM pyruvate and 3 mM malate, ROS production was determined following GDP
(1 mM) and oleate (0.1 mM) treatment. Data were normalized to background fluorescence and amount of protein. Values are
expressed as means ⫾ se, using Student’s t test; n ⫽ 5.

with DTT (fold change calculated from the total GDP
response). It is important to note that GDP was
titrated in 0.5 mM increments, reaching a final
concentration of 2 mM (Fig. 4A). Since UCP1 has
been reported to be fully inhibited by 1 mM GDP,
and GDP is also known to inhibit ANT, some of the
GDP effects may be due to ANT inhibition (37). In
contrast to GDP, no difference in oleate sensitivity
was observed between mitochondria isolated in the
absence or presence of DTT (Fig. 4A). BAT mito-
chondrial oxygen consumption determinations con-
firmed these observations. A more pronounced
decrease in oxygen consumption was observed in
DTT-exposed mitochondria acutely treated with 1
mM GDP (Fig. 4B). Although a trend for increased
respiration was observed following oleate addition to
the DTT-treated mitochondria, this did not reach
significance. We also tested ROS emission in BAT
mitochondria isolated in the presence or absence of
DTT. The addition of GDP did not increase the rate
of ROS production (Fig. 4C). However, the addition
of oleate increased ROS emission in mitochondria pre-
treated with DTT. A comparison of the rates of ROS forma-
tion following oleate addition indicates that DTT pretreat-
ment enhances mitochondrial ROS formation in BAT
(slope⫺DTT⫽236.94⫾63.63 vs. slope⫹DTT⫽396.71⫾67.52,
P⫽0.05, Student’s t test). Oleate did not increase in the rate
of ROS emission from mitochondria not pretreated with
DTT. Hence, unlike skeletal muscle, BAT mitochondria
treated with DTT display a metabolic shift that favors in-
creased sensitivity to UCP1 inhibition and higher ROS
generation.
BAT mitochondrial redox environment in
mitochondria from cold-exposed mice
Given that UCP1 expression and uncoupling increase
substantially in BAT when animals are exposed to a
cold environment and that the activation of mitochon-
drial uncoupling can drastically alter mitochondrial
redox poise, we measured GSH and GSSG pools in BAT
mitochondria isolated from cold-exposed mice. The
absolute amounts of GSH and GSSG are presented in
Table 1. The GSH/GSSG ratio was even more oxidized
in the BAT mitochondria from cold-exposed mice
(Fig. 5A). As expected, cold exposure altered the
bioenegetics of BAT mitochondria. Specifically, mito-
chondrial membrane potential was more depolarized
following cold exposure, and UCP1 was more sensitive
to the allosteric regulators GDP and oleate (Fig. 5B). In
addition, OCRs were double what was observed with
mitochondria from RT mitochondria (Fig. 5C). In-
creasing mitochondrial redox state by DTT pretreat-
ment still elevated the sensitivity of UCP1 to GDP
inhibition (Fig. 5B). No differences in membrane po-
larity were detected prior to GDP addition. In addition,
there was no statistically significant difference in the
effect of oleate on UCP1 in mitochondria isolated in
the presence or absence of DTT. Measurements of
OCR confirmed these results (Fig. 5C). Exposure of
DTT-treated mitochondria to GDP resulted in a more
significant decrease in O2 consumption. No differences
in oxygen consumption were recorded when GDP-
inhibited UCP1 was reactivated with oleate. We also
tested the effect of the presence or absence of DTT on

370 Vol. 26 January 2012 The FASEB Journal 䡠 www.fasebj.org MAILLOUX ET AL.
 Figure 5. Influence of a changing redox environment on UCP1 in BAT
mitochondria from cold-exposed mice. Mice were exposed to cold (4°C) for
24 h, and then mitochondria from BAT were isolated in the presence of 2 mM
DTT, as described in Materials and Methods. Mitochondria isolated in the
absence of DTT served as the control. Mitochondria were treated with
oligomycin (1.3 ␮g/ml) to induce state 4 respiratory conditions. A) Effect of
DTT treatment on GSH/GSSG ratio. Immediately following isolation, mito-
chondria were diluted to 0.5 mg/ml in a solution containing 1% (w/v) TCA
and 1% (w/v) meta-phosphoric acid, and then clarified for HPLC analysis of
GSH and GSSG levels. GSH and GSSG retention times were confirmed, and
levels were quantified by injecting varying amounts of standard solutions.
Following quantification, GSH/GSSG ratio was determined. Absolute amounts
of mitochondrial GSH and GSSG are shown in Table 1. Values are expressed
as means ⫾ sd, using Student’s t test; n ⫽ 4. B) Effect of alterations in redox environment on mitochondrial membrane
potential. Mitochondrial energization was tested using safranin O. Mitochondria (0.5 mg/ml) were energized at time 0 with
5 mM pyruvate and 3 mM malate under oligomycin-induced state 4 conditions, and then GDP (G) was titrated into the
reaction at 0.5 mM increments to test the effect of DTT on UCP1 inhibition. Effect of DTT on UCP1 reactivation was then
tested by adding oleate (OA) at 0.1 mM increments. Change in fluorescence following the addition of GDP and oleate was
calculated from the background fluorescence. Values are expressed as means ⫾ se, using Student’s t test; n ⫽ 4. C) DTT
increases sensitivity of UCP1 to GDP but not oleate. O2 consumption determinations were performed using a Clark-type
electrode under oligomycin-induced state 4 conditions. Respiration was initially tested with 5 mM pyruvate and 3 mM
malate, followed by GDP (1 mM) and oleate (0.1 mM) treatment. Values are expressed as means ⫾ se, using Student’s t
test; n ⫽ 4. D) Assessment of ROS production rates. ROS emission was tested under oligomycin-induced state 4 respiratory
conditions. ROS production was detected with DCFH-DA (20 ␮M). Following the addition of 5 mM pyruvate and 3 mM
malate, ROS production was determined following GDP (1 mM) and oleate (0.1 mM) treatment. Data were normalized to
background fluorescence and amount of protein. Values are expressed as means ⫾ se; n ⫽ 4.

ROS emission from BAT mitochondria isolated from
cold-exposed mice. Specifically, GDP did not increase
the rate of ROS production in mitochondria pretreated
with or without DTT (Fig. 5D). However, the addition
of oleate increased the rate of mitochondrial ROS
production in mitochondria pretreated with DTT (Fig.
5D). Indeed, the slope of the line following oleate
addition for the DTT-treated mitochondria was much
steeper, indicating faster kinetics of ROS production
(slope⫺DTT⫽979⫾420 vs. slope⫹DTT⫽1868⫾608, P⫽
0.05, Student’s t test).
DISCUSSION
In the present study, we provide the first analysis of the
effect of redox environment on bioenergetics in skele-
tal muscle and BAT mitochondria. Although a more
reductive environment was beneficial to skeletal muscle
mitochondria, these conditions seemed to have an
opposite effect on BAT mitochondria. These differ-
ences may be related to the vastly different physiologi-
cal roles of skeletal muscle and BAT. However, both
tissues are characterized by a high degree of metabolic
flexibility. BAT relies largely on glycolysis for ATP
production, and when fully activated, uses fatty acids to
support uncoupled respiration, while glycolytic ATP
production continues. In the fed state, skeletal muscle
is highly glycolytic, whereas in the fasting state and
during aerobic exercise, its reliance on fatty acid oxi-
dation is high. Such metabolic distinctions may account
for the observed differences in mitochondrial redox
state between these two tissues. We have demonstrated
that the differences in mitochondrial redox environ-
ment are required to maintain the differences between
mitochondrial bioenergetics in skeletal muscle and
BAT mitochondria. To manipulate the redox environ-
ment, mitochondria were isolated in the absence or
presence of DTT, a strong reducing agent. DTT is
commonly used to maintain a reductive environment
during mitochondrial isolation since its highly negative
redox potential maintains glutathione in a reduced
state (38). The process of mitochondrial isolation, due
to the exposure of mitochondria to ambient oxygen
levels can severely diminish mitochondrial function
(39). The recent work of Garcia et al. (7) reveals that
mitochondria isolated from other tissues using a Per-

REDOX MODULATION OF MITOCHONDRIAL METABOLISM 371

coll gradient protocol diminishes redox state and mi-
tochondrial function. Hence, the isolation of mito-
chondria in the presence of reducing agents, such as
DTT, protects enzymes from oxidative deactivation. In
the presence of DTT, skeletal muscle mitochondria
displayed an increase in UCP3-dependent state 4 respi-
ration, which coincided with a decrease in ROS emis-
sion. Although this is thought to be due to the deglu-
tathionylation and activation of UCP3, the increased
availability of GSH would also quench ROS.
Early studies of liver mitochondria established that the
redox environment influences mitochondrial metabolism
(40). Indeed, these initial observations showed that sup-
plementing mitochondria with DTT increases mitochon-
drial oxygen consumption. However, one caveat to these
early studies was that DTT was present during the actual
measurements of oxygen consumption, and reducing
agents (carrying very negative redox potentials) decrease
oxygen concentrations in solution. In the recent study by
Garcia et al. (7), the importance of isolating mitochondria
from tissues other than BAT or skeletal muscle in a
reductive environment was demonstrated. The research-
ers showed that DTT treatment maintains the metabolic
efficiency of neuronal mitochondria by preserving the
GSH pool, and thus the reductive potential of the matrix
environment (7). Hill et al. (41) made similar observations
in situ in vascular smooth muscle cells. These seminal
findings illustrate that redox plays a central role in mod-
ulating mitochondrial metabolism. However, it should be
pointed out that these more current studies also included
DTT in their reaction mixtures, which again can alter
oxygen levels in solution, and thus the measured rates of
oxygen consumption. In our study, we pretreated mito-
chondria or cells with DTT and immediately prior to
determinations replaced the DTT-containing solutions
with buffers devoid of this compound. Hence, the ob-
served changes were due to alterations in mitochondrial
redox environment unencumbered by any DTT-induced
decreases in oxygen tension.
Earlier studies in BAT mitochondria from mice ex-
posed to different temperatures showed that the GSH
pools are highly oxidized (42). This initial observation
is intriguing, since we know now that a large fraction of
the mitochondrial GSH pool in other tissues is main-
tained in a reduced state (2). A high GSH/GSSG ratio
ensures that the matrix environment is highly reducing,
a physical property required for the proper functioning
of mitochondria (e.g., catalytic thiols are maintained in
a reduced and active state, and a reductive environ-
ment aids in quenching ROS). However, in the case of
BAT mitochondria, the matrix environment is highly
oxidizing, especially in mitochondria isolated from
cold-exposed mice. On BAT activation, O2 consump-
tion increases severalfold and appears to be accompa-
nied by substantial increases in ROS production, which
further oxidizes the GSH pool. However, mitochondria
from mice housed at RT conditions also displayed high
rates of ROS production, especially after oleate addi-
tion. In this study, RT was ⬃23°C, which presents a mild
thermal stress to mice. Thermoneutrality for mice is
372 Vol. 26 January 2012 The FASEB Journal 䡠 www.fasebj.org
⬃30°C (43). Hence, the BAT from the RT mice used in
our study is slightly activated for thermogenesis, but
clearly not to the same extent as that from mice
exposed to 4°C. It would be interesting to perform
additional studies on BAT from mice housed at ther-
moneutrality. Cold exposure does increase SOD in BAT
mitochondria (44). Immunoblots on mitochondria
from the BAT of mice exposed to RT conditions
revealed higher amounts of Trx2 and Grx1 following
DTT treatment, which can aid in quenching ROS. But
Grx2, the matrix-associated Grx isoform, was too low in
amount to detect. Hence, increased ROS generation
from BAT mitochondria is most likely due to the highly
oxidative nature of this tissue, a unique characteristic
required to drive its thermogenic function. In the
present study, we not only confirmed that the GSH pool
is far more oxidized in comparison to skeletal muscle,
but that an oxidized GSH pool may be required for the
proper functioning of BAT mitochondria. In mitochon-
dria from both RT and cold-acclimated mice, UCP1
displayed increased sensitivity to GDP inhibition when
the redox environment was altered with DTT. We also
tested the effect of DTT on ROS emission from BAT
mitochondria. DTT did not have any effect on the rate
of mitochondrial ROS production when malate and
pyruvate were used as substrates or when GDP was
added to inhibit UCP1. However, the addition of oleate
substantially increased the rate of ROS production in
mitochondria pretreated with DTT. It should be noted
that the increases in ROS production are likely to be
associated with oleate oxidation pathways. Indeed,
acute increases in fatty acid oxidation can be a signifi-
cant source of ROS in mitochondria (45). Cold expo-
sure increased mitochondrial ROS generation even
further in comparison to RT mitochondria. In addi-
tion, BAT mitochondria generated far more ROS than
skeletal muscle mitochondria. Although we do not have
an explanation for this high level of ROS generation by
BAT mitochondria, we can only hypothesize that this is
directly or indirectly due to the presence of UCP1
and/or the absence of Grx2 (46). Indeed, oleate, a
well-known UCP1 activator, increased ROS generation.
Moreover, cold exposure, which increases mitochondrial
UCP1 content and activity increased ROS generation
even more. The observed increases in ROS production
following oleate treatment and cold acclimation indicate
that UCP1 does not participate in mitigating mitochon-
drial ROS emission. However, more in-depth investiga-
tions are warranted. Also, the effects of redox environ-
ment on UCP1 function should be investigated further in
UCP1-null mice. Hence, BAT mitochondria generate
large amounts of ROS, especially after stimulation of
UCP1. To our knowledge, there is only one previously
published study in which BAT mitochondrial ROS pro-
duction was measured. Mracek et al. (47) measured the
effect of different substrates on ROS emission in BAT
mitochondria; however, the researchers did not assess the
effects of GDP and oleate on ROS emission.
Recently, we corroborated and extended the original
findings of Echtay et al. (30) by showing that proton
MAILLOUX ET AL.
leakage through UCP3 is activated by a ROS-induced
deglutathionylation (11). Only small amounts of ROS
activated UCP3, since higher amounts actually deacti-
vated the protein. Moreover, we demonstrated that the
ROS-mediated activation of UCP3 required the deglu-
tathionylation of Cys25 or Cys259 and that the subse-
quent reglutathionylation of these residues inhibits
UCP3-mediated proton leakage (11). In the present
study, we not only confirmed our previous observations,
but also show that the degree of UCP3 glutathionyla-
tion dictates the contribution of this protein to proton
leakage. C2C12 cells overexpressing UCP3 displayed a
small, but significant, increase in state 4 respiration
when pretreated with DTT. However, DTT pretreat-
ment increased the sensitivity of UCP3 to subsequent
diamide-mediated inhibition. Indeed, in UCP3-express-
ing C2C12 cells pretreated with DTT, there was a more
pronounced decline in state 4 respiration with increas-
ing doses of diamide. In the DTT-pretreated cells, a
concentration of up to 200 ␮M diamide was needed to
fully inhibit UCP3. In contrast, UCP3-expressing cells
not exposed to DTT were less sensitive to diamide
treatment; UCP3 was fully inhibited by 100 ␮M di-
amide. This would indicate that UCP3 is partially
glutathionylated (and thus inhibited) in cells not pre-
treated with DTT. Furthermore, the inhibition of UCP3
by diamide seemed to dissipate over time, since state 4
respiration gradually returned to normal, indicating
that glutathionylation/deglutathionylation of UCP3 is a
transient event. Hence, these findings indicate that
DTT pretreatment (inducing deglutathionylation) is
required to assess the full leakage capacity of UCP3.
As shown in Fig. 6, our findings indicate that redox
state affects the control of UCP3 by modulating the
extent of its glutathionylation. Of note, we also ob-
served that C2C12 cells transiently transfected with
empty pCMV vector were sensitive to diamide-mediated
inhibition of state 4 respiration, although not to the
same degree as cells expressing UCP3. DTT pretreat-
REDOX MODULATION OF MITOCHONDRIAL METABOLISM
ment increased the sensitivity of these cells to diamide
treatment. Since the only protein expressed in C2C12
cells that is known to be associated with proton leakage
is ANT, and since ANT may be modulated by glutathio-
nylation (16), our current findings indicate that further
research into the control of ANT-mediated proton
leakage by glutathionylation in muscle is warranted.
Like UCP3, UCP1 has been suggested to play a role
in curtailing mitochondrial ROS emission through an
inducible proton-leak mechanism (48). However, the
role of UCP1 in controlling ROS production from
mitochondria is highly contentious (49). We have also
recently showed that, unlike UCP3, UCP1 is not mod-
ulated by diamide-induced glutathionylation in BAT
mitochondria (11). However, in the same publication,
we did observe that UCP3 was controlled by glutathio-
nylation in BAT mitochondria (11). Indeed, BAT mi-
tochondria do express UCP3, and on the basis of our
observations, it would seem that UCP3 and UCP1 are
controlled differently, which is surprising since they
share high homology in cysteine residues. The differ-
ence in control of these two closely related proteins
may be due to divergences in their physiological func-
tions. While UCP1 is used for thermogenesis, UCP3 is
required to guard cells from ROS and/or to control
ROS-mediated signaling processes in cells (19). In the
present study, we investigated the glutathionylated pro-
teome and the redox poise of BAT mitochondria. How-
ever, we found no significant DTT-induced changes in the
glutathionylated proteome; however, a trend for de-
creased protein glutathionylation following DTT pretreat-
ment was observed. Intriguingly, though, we found that
BAT mitochondria, unlike skeletal muscle, do not contain
Grx2 (or the levels are too low for detection by immuno-
blot). The main enzymes thought to be involved in
driving reversible glutathionylation in the mitochondria
are Grx1, Grx2, and, to a limited extent, Trx2. While Grx1
is positioned in the intermembrane space/cytosol, Grx2 is
located in the matrix, where most ROS are produced
Figure 6. Model for the control of
UCP3-mediated proton leakage by re-
versible glutathionylation. Partial gluta-
thionylation of UCP3 inhibits proton
leakage. Complete deglutathionylation
of UCP3 results in the activation of the
protein and increases proton leakage.
Full glutathionylation of available cys-
teines inhibits UCP3-mediated proton
leakage.
373
(50). The absence of Grx2 in the matrix of BAT mito-
chondria would indicate that this tissue lacks at least one
enzyme involved in GSH conjugation reactions in the
matrix. BAT mitochondria did, nevertheless, contain
abundant Trx2, which can drive glutathionylation reac-
tions. However, Trx2 is not as efficient as Grx in mediat-
ing glutathionylation reactions (51). BAT mitochondria
were also found to have a far more oxidized GSH pool
than skeletal muscle mitochondria, consistent with the
propensity of BAT mitochondria to generate much
higher amounts of ROS. These conditions in BAT mito-
chondria could prompt the nonenzymatic glutathionyla-
tion of proteins. Indeed, ROS are well known to drive
spontaneous addition of GSH to exposed cysteinyl groups
by oxidizing sulfur residues to more reactive sulfinic acids
(50). GSSG could also glutathionylate the proteome
through a simple thiol disulfide exchange reaction with
exposed cysteines (52, 53). Although we do not provide
any conclusive evidence regarding UCP1 glutathionyla-
tion, we can conclude that the mechanisms governing
glutathionylation in BAT mitochondria differ greatly
from those in skeletal muscle.
In summary, this is the first study to profile the effect
of redox environment on skeletal muscle and BAT
bioenergetics. Surprisingly, we found that BAT mito-
chondria have a far more oxidizing redox environment,
and our results indicate that this is required for UCP1
function. These fundamental observations indicate the
very unique nature of BAT, as our results herein and
those of others (7, 41) demonstrate that all other
tissues are reliant on a more reductive environment
for their function. Indeed, the major physiological
function of BAT is nonshivering thermogenesis, and
this is unequivocally dependent on UCP1. Because of
their significant contributions to whole body energy
homeostasis, skeletal muscle and BAT are studied ex-
tensively in order to better understand the pathogene-
sis of obesity and type 2 diabetes mellitus (54, 55). The
observations, herein, provide further insights into the
unique importance of mitochondrial redox balance on
metabolic processes in these two tissues.
The authors thank Mahmoud Salkhordeh for the isolation
and purification of the primary myoblasts. R.J.M. was funded
by a postdoctoral fellowship provided by the Canadian Insti-
tutes of Health Research (CIHR). Operating funding for this
research was provided by CIHR (Institute of Nutrition, Me-
tabolism, and Diabetes; to M.E.H.).
REFERENCES
1. Schafer, F. Q., and Buettner, G. R. (2001) Redox environment
of the cell as viewed through the redox state of the glutathione
disulfide/glutathione couple. Free Radic. Biol. Med. 30, 1191–
1212
2. Hurd, T. R., Costa, N. J., Dahm, C. C., Beer, S. M., Brown,
S. E., Filipovska, A., and Murphy, M. P. (2005) Glutathiony-
lation of mitochondrial proteins. Antioxid. Redox Signal. 7,
999 –1010
3. Mari, M., Morales, A., Colell, A., Garcia-Ruiz, C., and Fernandez-
Checa, J. C. (2009) Mitochondrial glutathione, a key survival
antioxidant. Antioxid. Redox Signal. 11, 2685–2700
374 Vol. 26 January 2012 The FASEB Journal 䡠 www.fasebj.org
4. Taylor, E. R., Hurrell, F., Shannon, R. J., Lin, T. K., Hirst, J., and
Murphy, M. P. (2003) Reversible glutathionylation of complex I
increases mitochondrial superoxide formation. J. Biol. Chem.
278, 19603–19610
5. Wang, J., Boja, E. S., Tan, W., Tekle, E., Fales, H. M., English, S.,
Mieyal, J. J., and Chock, P. B. (2001) Reversible glutathionyla-
tion regulates actin polymerization in A431 cells. J. Biol. Chem.
276, 47763– 47766
6. Naoi, M., Maruyama, W., Yi, H., Inaba, K., Akao, Y., and
Shamoto-Nagai, M. (2009) Mitochondria in neurodegenerative
disorders: regulation of the redox state and death signaling
leading to neuronal death and survival. J. Neural Transm. 116,
1371–1381
7. Garcia, J., Han, D., Sancheti, H., Yap, L. P., Kaplowitz, N., and
Cadenas, E. (2010) Regulation of mitochondrial glutathione
redox status and protein glutathionylation by respiratory sub-
strates. J. Biol. Chem. 285, 39646 –39654
8. Shelton, M. D., Chock, P. B., and Mieyal, J. J. (2005) Glutare-
doxin: role in reversible protein S-glutathionylation and regula-
tion of redox signal transduction and protein translocation.
Antioxid. Redox Signal. 7, 348 –366
9. Mieyal, J. J., Gallogly, M. M., Qanungo, S., Sabens, E. A., and
Shelton, M. D. (2008) Molecular mechanisms and clinical
implications of reversible protein S-glutathionylation. Antioxid.
Redox Signal. 10, 1941–1988
10. Fratelli, M., Demol, H., Puype, M., Casagrande, S., Eberini, I.,
Salmona, M., Bonetto, V., Mengozzi, M., Duffieux, F., Miclet, E.,
Bachi, A., Vandekerckhove, J., Gianazza, E., and Ghezzi, P.
(2002) Identification by redox proteomics of glutathionylated
proteins in oxidatively stressed human T lymphocytes. Proc. Natl.
Acad. Sci. U. S. A. 99, 3505–3510
11. Mailloux, R. J., Seifert, E. L., Bouillaud, F., Aguer, C., Collins, S.,
and Harper, M. E. (2011) Glutathionylation acts as a control
switch for uncoupling proteins UCP2 and UCP3. J. Biol. Chem.
286, 21865–21875
12. Chen, Y. R., Chen, C. L., Pfeiffer, D. R., and Zweier, J. L. (2007)
Mitochondrial complex II in the post-ischemic heart: oxidative
injury and the role of protein S-glutathionylation. J. Biol. Chem.
282, 32640 –32654
13. Hurd, T. R., Requejo, R., Filipovska, A., Brown, S., Prime, T. A.,
Robinson, A. J., Fearnley, I. M., and Murphy, M. P. (2008)
Complex I within oxidatively stressed bovine heart mitochon-
dria is glutathionylated on Cys-531 and Cys-704 of the 75-kDa
subunit: potential role of CYS residues in decreasing oxidative
damage. J. Biol. Chem. 283, 24801–24815
14. Applegate, M. A., Humphries, K. M., and Szweda, L. I. (2008)
Reversible inhibition of alpha-ketoglutarate dehydrogenase by
hydrogen peroxide: glutathionylation and protection of lipoic
acid. Biochemistry 47, 473– 478
15. Kil, I. S., and Park, J. W. (2005) Regulation of mitochondrial
NADP⫹-dependent isocitrate dehydrogenase activity by gluta-
thionylation. J. Biol. Chem. 280, 10846 –10854
16. Queiroga, C. S., Almeida, A. S., Martel, C., Brenner, C., Alves,
P. M., and Vieira, H. L. (2010) Glutathionylation of adenine
nucleotide translocase induced by carbon monoxide prevents
mitochondrial membrane permeabilization and apoptosis.
J. Biol. Chem. 285, 17077–17088
17. Brennan, J. P., Wait, R., Begum, S., Bell, J. R., Dunn, M. J., and
Eaton, P. (2004) Detection and mapping of widespread inter-
molecular protein disulfide formation during cardiac oxidative
stress using proteomics with diagonal electrophoresis. J. Biol.
Chem. 279, 41352– 41360
18. Harper, M. E., Green, K., and Brand, M. D. (2008) The
efficiency of cellular energy transduction and its implications
for obesity. Annu. Rev. Nutr. 28, 13–33
19. Azzu, V., and Brand, M. D. (2010) The on-off switches of the
mitochondrial uncoupling proteins. Trends Biochem. Sci. 35,
298 –307
20. Seale, P., Bjork, B., Yang, W., Kajimura, S., Chin, S., Kuang, S.,
Scime, A., Devarakonda, S., Conroe, H. M., Erdjument-Brom-
age, H., Tempst, P., Rudnicki, M. A., Beier, D. R., and Spiegel-
man, B. M. (2008) PRDM16 controls a brown fat/skeletal
muscle switch. Nature 454, 961–967
21. Schulz, T. J., Huang, T. L., Tran, T. T., Zhang, H., Townsend,
K. L., Shadrach, J. L., Cerletti, M., McDougall, L. E., Giorgadze,
N., Tchkonia, T., Schrier, D., Falb, D., Kirkland, J. L., Wagers,
A. J., and Tseng, Y. H. (2011) Identification of inducible brown
MAILLOUX ET AL.
 adipocyte progenitors residing in skeletal muscle and white fat.
Proc. Natl. Acad. Sci. U. S. A. 108, 143–148
22. Rolfe, D. F., and Brand, M. D. (1996) Contribution of mitochon-
drial proton leak to skeletal muscle respiration and to standard
metabolic rate. Am. J. Physiol. 271, C1380 –C1389
23. Rolfe, D. F., and Brand, M. D. (1996) Proton leak and control of
oxidative phosphorylation in perfused, resting rat skeletal mus-
cle. Biochim. Biophys. Acta 1276, 45–50
24. Cannon, B., and Nedergaard, J. (2004) Brown adipose tissue:
function and physiological significance. Physiol. Rev. 84, 277–359
25. Enerback, S., Jacobsson, A., Simpson, E. M., Guerra, C., Ya-
mashita, H., Harper, M. E., and Kozak, L. P. (1997) Mice lacking
mitochondrial uncoupling protein are cold-sensitive but not
obese. Nature 387, 90 –94
26. Goubern, M., Yazbeck, J., Chapey, M. F., Diolez, P., and Moreau,
F. (1990) Variations in energization parameters and proton
conductance induced by cold adaptation and essential fatty acid
deficiency in mitochondria of brown adipose tissue in the rat.
Biochim. Biophys. Acta 1015, 334 –340
27. Nicholls, D. G., and Rial, E. (1999) A history of the first
uncoupling protein, UCP1. J. Bioenerg. Biomembr. 31, 399 – 406
28. Himms-Hagen, J. (1990) Brown adipose tissue thermogenesis:
interdisciplinary studies. FASEB J. 4, 2890 –2898
29. Mailloux, R. J., Dumouchel, T., Aguer, C., Dekemp, R., Beanlands, R.,
and Harper, M. E. (2011) Hexokinase II acts through UCP3 to
suppress mitochondrial reactive oxygen species production and main-
tain aerobic respiration. Biochem. J. 437, 21865–21875
30. Echtay, K. S., Roussel, D., St-Pierre, J., Jekabsons, M. B., Cadenas, S.,
Stuart, J. A., Harper, J. A., Roebuck, S. J., Morrison, A., Pickering, S.,
Clapham, J. C., and Brand, M. D. (2002) Superoxide activates mito-
chondrial uncoupling proteins. Nature 415, 96–99
31. Cypess, A. M., Lehman, S., Williams, G., Tal, I., Rodman, D.,
Goldfine, A. B., Kuo, F. C., Palmer, E. L., Tseng, Y. H., Doria, A.,
Kolodny, G. M., and Kahn, C. R. (2009) Identification and
importance of brown adipose tissue in adult humans. N. Engl.
J. Med. 360, 1509 –1517
32. Virtanen, K. A., Lidell, M. E., Orava, J., Heglind, M., Westergren,
R., Niemi, T., Taittonen, M., Laine, J., Savisto, N. J., Enerback,
S., and Nuutila, P. (2009) Functional brown adipose tissue in
healthy adults. N. Engl. J. Med. 360, 1518 –1525
33. Nedergaard, J., Bengtsson, T., and Cannon, B. (2010) Three
years with adult human brown adipose tissue. Ann. N. Y. Acad.
Sci. 1212, E20 –E36
34. Rando, T. A., and Blau, H. M. (1994) Primary mouse myoblast
purification, characterization, and transplantation for cell-me-
diated gene therapy. J. Cell Biol. 125, 1275–1287
35. Mailloux, R. J., and Harper, M. E. (2010) Glucose regulates
enzymatic sources of mitochondrial NADPH in skeletal muscle
cells; a novel role for glucose-6-phosphate dehydrogenase.
FASEB J. 24, 2495–2506
36. Caprioli, J., Munemasa, Y., Kwong, J. M., and Piri, N. (2009)
Overexpression of thioredoxins 1 and 2 increases retinal gan-
glion cell survival after pharmacologically induced oxidative
stress, optic nerve transection, and in experimental glaucoma.
Trans. Am. Ophthalmol. Soc. 107, 161–165
37. Parker, N., Crichton, P. G., Vidal-Puig, A. J., and Brand, M. D.
(2009) Uncoupling protein-1 (UCP1) contributes to the basal
proton conductance of brown adipose tissue mitochondria.
J. Bioenerg. Biomembr. 41, 335–342
38. Cleland, W. W. (1964) Dithiothreitol, a new protective reagent
for Sh groups. Biochemistry 3, 480 – 482
39. Nulton-Persson, A. C., and Szweda, L. I. (2001) Modulation of
mitochondrial function by hydrogen peroxide. J. Biol. Chem.
276, 23357–23361
REDOX MODULATION OF MITOCHONDRIAL METABOLISM
40. Haugaard, N., Lee, N. H., Kostrzewa, R., and Haugaard, E. S.
(1969) Effects of a disulfide (Ellman’s reagent) and thiols on
oxidative phosphorylation and ion transport by rat liver mito-
chondria. Biochem. Pharmacol. 18, 2385–2391
41. Hill, B. G., Higdon, A. N., Dranka, B. P., and Darley-Usmar,
V. M. (2010) Regulation of vascular smooth muscle cell bioen-
ergetic function by protein glutathiolation. Biochim. Biophys. Acta
1797, 285–295
42. Barja de Quiroga, G., Lopez-Torres, M., Perez-Campo, R.,
Abelenda, M., Paz Nava, M., and Puerta, M. L. (1991) Effect
of cold acclimation on GSH, antioxidant enzymes and lipid
peroxidation in brown adipose tissue. Biochem. J. 277, 289 –
292
43. Feldmann, H. M., Golozoubova, V., Cannon, B., and Neder-
gaard, J. (2009) UCP1 ablation induces obesity and abolishes
diet-induced thermogenesis in mice exempt from thermal stress
by living at thermoneutrality. Cell Metab. 9, 203–209
44. Petrovic, V., Buzadzic, B., Korac, A., and Korac, B. (2009)
Antioxidative defense and mitochondrial thermogenic response
in brown adipose tissue. Genes. Nutr. 5, 225–235
45. Seifert, E. L., Estey, C., Xuan, J. Y., and Harper, M. E. (2009)
Electron transport chain-dependent and -independent mecha-
nisms of mitochondrial H2O2 emission during long-chain fatty
acid oxidation. J. Biol. Chem. 285, 5748 –5758
46. Wu, H., Xing, K., and Lou, M. F. (2010) Glutaredoxin 2 prevents
H2O2-induced cell apoptosis by protecting complex I activity in
the mitochondria. Biochim. Biophys. Acta 1797, 1705–1715
47. Mracek, T., Pecinova, A., Vrbacky, M., Drahota, Z., and Houstek,
J. (2009) High efficiency of ROS production by glycerophos-
phate dehydrogenase in mammalian mitochondria. Arch.
Biochem. Biophys. 481, 30 –36
48. Oelkrug, R., Kutschke, M., Meyer, C. W., Heldmaier, G., and
Jastroch, M. (2010) Uncoupling protein 1 decreases superoxide
production in brown adipose tissue mitochondria. J. Biol. Chem.
285, 21961–21968
49. Silva, J. P., Shabalina, I. G., Dufour, E., Petrovic, N., Backlund,
E. C., Hultenby, K., Wibom, R., Nedergaard, J., Cannon, B., and
Larsson, N. G. (2005) SOD2 overexpression: enhanced mito-
chondrial tolerance but absence of effect on UCP activity.
EMBO J. 24, 4061– 4070
50. Gallogly, M. M., and Mieyal, J. J. (2007) Mechanisms of revers-
ible protein glutathionylation in redox signaling and oxidative
stress. Curr. Opin. Pharmacol. 7, 381–391
51. Chrestensen, C. A., Starke, D. W., and Mieyal, J. J. (2000) Acute
cadmium exposure inactivates thioltransferase (glutaredoxin),
inhibits intracellular reduction of protein-glutathionyl-mixed
disulfides, and initiates apoptosis. J. Biol. Chem. 275, 26556 –
26565
52. Gilbert, H. F. (1995) Thiol/disulfide exchange equilibria and
disulfide bond stability. Methods Enzymol. 251, 8 –28
53. Shaked, Z., Szajewski, R. P., and Whitesides, G. M. (1980) Rates
of thiol-disulfide interchange reactions involving proteins and
kinetic measurements of thiol pKa values. Biochemistry 19, 4156 –
4166
54. Sastre, J., Pallardo, F. V., Llopis, J., Furukawa, T., Vina, J. R., and
Vina, J. (1989) Glutathione depletion by hyperphagia-induced
obesity. Life Sci. 45, 183–187
55. Robertson, R. P., Harmon, J., Tran, P. O., and Poitout, V.
(2004) Beta-cell glucose toxicity, lipotoxicity, and chronic
oxidative stress in type 2 diabetes. Diabetes 53(Suppl. 1),
S119 –S124
Received for publication June 9, 2011.
Accepted for publication September 1, 2011.
375