pnas201318547 273..278


Evaluation of intramitochondrial ATP levels identifies
G0/G1 switch gene 2 as a positive regulator of
oxidative phosphorylation
Hidetaka Kiokaa,b,1, Hisakazu Katoa,1, Makoto Fujikawac, Osamu Tsukamotoa, Toshiharu Suzukid,e, Hiromi Imamuraf,
Atsushi Nakanoa,g, Shuichiro Higoa,b, Satoru Yamazakih, Takashi Matsuzakib, Kazuaki Takafujii, Hiroshi Asanumaj,
Masanori Asakurag, Tetsuo Minaminob, Yasunori Shintania, Masasuke Yoshidae, Hiroyuki Nojik, Masafumi Kitakazeg,
Issei Komurob,l, Yoshihiro Asanoa,b,2, and Seiji Takashimaa,2

Departments of aMedical Biochemistry and bCardiovascular Medicine and iCenter for Research Education, Osaka University Graduate School of Medicine,
Osaka 565-0871, Japan; cDepartment of Biochemistry, Faculty of Pharmaceutical Science, Tokyo University of Science, Chiba 278-8510, Japan; dChemical
Resources Laboratory, Tokyo Institute of Technology, Yokohama 226-8503, Japan; eDepartment of Molecular Bioscience, Kyoto Sangyo University, Kyoto
603-8555, Japan; fThe Hakubi Center for Advanced Research and Graduate School of Biostudies, Kyoto University, Kyoto 606-8501, Japan; Departments of
gClinical Research and Development and hCell Biology, National Cerebral and Cardiovascular Center Research Institute, Osaka 565-8565, Japan; jDepartment
of Cardiovascular Science and Technology, Kyoto Prefectural University School of Medicine, Kyoto 602-8566, Japan; and kDepartment of Applied Chemistry,
School of Engineering and lDepartment of Cardiovascular Medicine, Graduate School of Medicine, University of Tokyo, Tokyo 113-8656, Japan

Edited by Gottfried Schatz, University of Basel, Reinach, Switzerland, and approved November 19, 2013 (received for review October 7, 2013)

The oxidative phosphorylation (OXPHOS) system generates most
of the ATP in respiring cells. ATP-depleting conditions, such as
hypoxia, trigger responses that promote ATP production. How-
ever, how OXPHOS is regulated during hypoxia has yet to be
elucidated. In this study, selective measurement of intramitochon-
drial ATP levels identified the hypoxia-inducible protein G0/G1
switch gene 2 (G0s2) as a positive regulator of OXPHOS. A
mitochondria-targeted, FRET-based ATP biosensor enabled us to
assess OXPHOS activity in living cells. Mitochondria-targeted,
FRET-based ATP biosensor and ATP production assay in a semi-
intact cell system revealed that G0s2 increases mitochondrial ATP
production. The expression of G0s2 was rapidly and transiently
induced by hypoxic stimuli, and G0s2 interacts with OXPHOS
complex V (FoF1-ATP synthase). Furthermore, physiological en-
hancement of G0s2 expression prevented cells from ATP depletion
and induced a cellular tolerance for hypoxic stress. These results
show that G0s2 positively regulates OXPHOS activity by interact-
ing with FoF1-ATP synthase, which causes an increase in ATP pro-
duction in response to hypoxic stress and protects cells from a
critical energy crisis. These findings contribute to the understand-
ing of a unique stress response to energy depletion. Additionally,
this study shows the importance of assessing intramitochondrial
ATP levels to evaluate OXPHOS activity in living cells.

energy metabolism | live-cell imaging

Maintaining cellular homeostasis and activities requires astable energy supply. Most eukaryotic cells generate ATP
as their energy currency mainly through the mitochondrial oxi-
dative phosphorylation (OXPHOS) system. The OXPHOS sys-
tem consists of five large protein complex units (i.e., complexes
I–V), comprising more than 100 proteins. In this system, oxygen
(O2) is essential as the terminal electron acceptor for complex IV
to finally produce the proton-motive force that drives the ATP-
generating molecular motor complex V (FoF1-ATP synthase).
Hypoxia causes the depletion of intracellular ATP and triggers

adaptive cellular responses to help maintain intracellular ATP
levels and minimize any deleterious effects of energy depletion.
Although the metabolic switch from mitochondrial respiration to
anaerobic glycolysis is widely recognized (1–4), several recent
reports have shown that hypoxic stimuli unexpectedly increase
OXPHOS efficiency as well (5–7). In other words, cells have
adaptive mechanisms to maintain intracellular ATP levels by
enhancing OXPHOS, particularly in the early phase of hypoxia,
in which the O2 supply is limited but still remains. However, the
mechanism by which OXPHOS is regulated during this early
hypoxic phase is still not fully understood.

Revealing the mechanism of this fine-tuned regulation of
OXPHOS requires accurate and noninvasive measurements of
OXPHOS activity. Although researchers have established methods
to measure OXPHOS activity, precise measurement, especially in
living cells, is still difficult. Measuring the intracellular ATP
concentration is one of the most commonly used methods for
evaluating OXPHOS activity. However, there are two major
problems with this method. First, the intracellular ATP concen-
tration does not always accurately reflect OXPHOS activity, be-
cause it can also be affected by glycolytic ATP production, cytosolic
ATPases, and ATP buffering enzymes, such as creatine kinase
and adenylate kinase (8). Second, because measurements of the
ATP concentration by chromatography (9), MS (10), NMR (11),
or luciferase assays (12) are based on cell extract analysis, these
methods cannot be used to measure the serial ATP concentration
changes in living cells in real time.
In this study, we overcame these problems by the selective

measurement of the intramitochondrial matrix ATP concentra-
tion ([ATP]mito) in living cells. In the final step of OXPHOS,
ATP is produced not in the cytosol but in the mitochondrial
matrix. Therefore, we hypothesized that a selectively measuring
[ATP]mito is suitable for the highly sensitive evaluation of cellular
ATP production by OXPHOS. In fact, real-time evaluation of
both [ATP]mito and the cytosolic ATP concentration ([ATP]cyto)
in living cells revealed that [ATP]mito reflected OXPHOS activity
with far more sensitivity than [ATP]cyto. Using this fine method,
we found that G0/G1 switch gene 2 (G0s2), a hypoxia-induced

Significance

We developed a sensitive method to assess the activity of ox-
idative phosphorylation in living cells using a FRET-based ATP
biosensor. We then revealed that G0/G1 switch gene 2, a pro-
tein rapidly induced by hypoxia, increases mitochondrial ATP
production by interacting with FoF1-ATP synthase and protects
cells from a critical energy crisis.

Author contributions: Y.A. and S.T. designed research; H. Kioka, H. Kato, O.T., and A.N.
performed research; M.F., T.S., H.I., S.H., S.Y., T. Matsuzaki, K.T., H.A., M.A., T. Minamino,
Y.S., M.Y., H.N., M.K., and I.K. contributed new reagents/analytic tools; H. Kioka and
H. Kato analyzed data; and Y.A. and S.T. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.
1H. Kioka and H. Kato contributed equally to this work.
2To whom correspondence may be addressed. E-mail: asano@cardiology.med.osaka-u.ac.jp
or takasima@cardiology.med.osaka-u.ac.jp.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1318547111/-/DCSupplemental.

www.pnas.org/cgi/doi/10.1073/pnas.1318547111 PNAS | January 7, 2014 | vol. 111 | no. 1 | 273–278

C
EL
L
B
IO
LO

G
Y

D
o
w

n
lo

a
d
e
d
 a

t 
C

a
rn

e
g
ie

 M
e
llo

n
 U

n
iv

e
rs

ity
 o

n
 A

p
ri
l 5

, 
2
0
2
1
 

mailto:asano@cardiology.med.osaka-u.ac.jp
mailto:takasima@cardiology.med.osaka-u.ac.jp
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental
www.pnas.org/cgi/doi/10.1073/pnas.1318547111


protein in cardiomyocytes, increases OXPHOS activity. G0s2
interacted with FoF1-ATP synthase and increased the ATP
production rate. Our results suggest that hypoxia-induced pro-
tein G0s2 is a positive regulator of OXPHOS and protects cells
by preserving ATP production, even under hypoxic conditions.

Results
Establishment of a Sensitive Method to Assess OXPHOS Activity in
Living Cells. To elucidate the mechanism by which OXPHOS is
regulated under hypoxia, it is essential to establish a sensitive
method for assessing OXPHOS activity in living cells. For this
purpose, we used an ATP indicator based on e-subunit for an-
alytical measurements (ATeam), which is an ATP-sensing
FRET-based indicator (13). We introduced this ATP biosensor
into cardiomyocytes that possess an abundance of mitochondria
and produce the highest levels of ATP among all primary cells
(14, 15). The ATeam assay can measure both [ATP]cyto (i.e., the
Cyto-ATeam assay) and [ATP]mito when a duplex of the mito-
chondrial targeting signal of cytochrome c oxidase subunit VIII is
attached to the indicator (i.e., the Mit-ATeam assay). In this
case, the YFP/CFP emission ratio of the ATeam fluorescence
represents the ATP concentration in each compartment. In-
terestingly, the Mit-ATeam assay was a far more sensitive
method than the Cyto-ATeam assay in determining OXPHOS
activity in living cells. For example, a very low dose of oligomycin
A (0.01 μg/mL), a specific OXPHOS complex V (FoF1-ATP
synthase) inhibitor, greatly reduced the YFP/CFP emission ratio
of the Mit-ATeam fluorescence that represents [ATP]mito within
10 min (Fig. 1 A, Upper and B and Movie S1). In contrast, the
same dose of oligomycin A resulted in a slight and slow decline
of the YFP/CFP emission ratio of Cyto-ATeam fluorescence
(Fig. 1 A, Lower and B and Movie S1). The same phenomenon
was observed when the cells were exposed to hypoxia, which
suppresses the activity of OXPHOS complex IV (cytochrome c
oxidase). Again, [ATP]mito decreased more markedly than [ATP]cyto
during 2.5 h of hypoxia (Fig. 1 C and D and Movie S2). These
results indicate that the Mit-ATeam assay is far more sensitive
for measuring the activity of OXPHOS than the Cyto-ATeam

assay. In addition, OXPHOS inhibition decreased the YFP/CFP
emission ratio of the Mit-ATeam fluorescence of HeLa cells as
well as cardiomyocytes (Fig. S1), suggesting the broad applica-
bility of this assay. Therefore, we used Mit-ATeam for the as-
sessment of the OXPHOS activity in living cells.

Hypoxia-Induced Gene G0s2 Affects the Intramitochondrial ATP
Concentration. The expression of genes involved in OXPHOS
regulation is considered to be up-regulated in the early phase of
hypoxia. Thus, to find unique OXPHOS regulators, we focused
on the rapidly induced genes in response to hypoxic stimulation.
We compared the gene expression profiles of cultured rat car-
diomyocytes at three different time points during hypoxic con-
ditions (0, 2, and 12 h) (Fig. S2A). The expression of well-known
hypoxia-induced genes, such as VEGF-α and hexokinase 2 mRNA
(16, 17), was slightly up-regulated at 2 h and further enhanced at
12 h of hypoxia. In contrast, three other genes (Adamts1, Cdkn3,
and G0s2) underwent rapid increases in expression at 2 h but
declined at 12 h of sustained hypoxia (Fig. S2 B and C). This
rapid and transient time course of expression implies that these
three genes may play distinct regulatory roles, especially in the
early hypoxic phase, in which oxygen is limited but still available.
To examine whether these genes are involved in the regulation of
OXPHOS activity, we knocked down these genes by shRNA (see
Fig. S7A) and examined [ATP]mito using the Mit-ATeam assay. In
this experiment, [ATP]mito in cardiomyocytes treated with shRNA
for G0s2 clearly declined within 24 h compared with the control
cardiomyocytes (Fig. 2A and Movie S3). In addition, the time
course of ATP decline was in agreement with the time course of
G0s2 depletion (Fig. 2A and Fig. S3A). Importantly, the over-
expression of G0s2 restored normal ATP levels (Fig. 2 B and C),
and again, the Cyto-ATeam assay could not detect a significant
effect of G0s2 knockdown within this time frame (Fig. S3B and
Movie S4). These findings imply that mitochondrial ATP pro-
duction through OXPHOS was inhibited by G0s2 ablation.
We confirmed that the mRNA and protein levels of G0s2 both
increased after 2–6 h of hypoxia and then declined after 12 h of
hypoxia (Fig. 2 D and E). G0s2 was first reported as a gene with

Fig. 1. Establishment of a sensitive method to assess OXPHOS activity in living cells. (A) YFP/CFP emission ratio plots of (Upper) Mit-ATeam and (Lower) Cyto-
ATeam fluorescence in cardiomyocytes. Various concentrations (0.001, 0.01, 0.1, 1, and 10 μg/mL) of oligomycin A or DMSO (control) were added at 5 min
(arrowhead; n = 3). (B) Representative sequential YFP/CFP ratiometric pseudocolored images of (Upper) Mit-ATeam and (Lower) Cyto-ATeam in car-
diomyocytes. Oligomycin A (0.01 μg/mL) was added at 5 min. (Scale bars: 20 μm.) (C) YFP/CFP emission ratio plots of Mit-ATeam and Cyto-ATeam fluorescence
in cardiomyocytes (n = 10). (D) Representative sequential YFP/CFP ratiometric pseudocolored images of (Upper) Mit-ATeam and (Lower) Cyto-ATeam in
cardiomyocytes. Cells were exposed to 1% hypoxia from the time point 30 min. All of the measurements were normalized to the YFP/CFP emission ratio at
0 min. Data are represented as the means ± SEMs. (Scale bars: 20 μm.)

274 | www.pnas.org/cgi/doi/10.1073/pnas.1318547111 Kioka et al.

D
o
w

n
lo

a
d
e
d
 a

t 
C

a
rn

e
g
ie

 M
e
llo

n
 U

n
iv

e
rs

ity
 o

n
 A

p
ri
l 5

, 
2
0
2
1
 

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/sm01.mov
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/sm01.mov
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/sm02.mov
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF1
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF2
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF2
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF7
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/sm03.mov
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF3
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF3
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/sm04.mov
www.pnas.org/cgi/doi/10.1073/pnas.1318547111


expression that was induced during the cell cycle switch from G0
to G1 phase (18). G0s2 is expressed in many tissues and espe-
cially abundant in heart, skeletal muscle, liver, kidney, brain, and
adipose tissue (19). Although G0s2 may play a role in cell cycle
progression (20), the function of G0s2 in the hypoxic response
remains unknown.

G0s2 Rescues the Decline of ATP Production During Hypoxia. We next
tested whether the overexpression of the G0s2 before hypoxic
stress could prevent hypoxia-induced ATP depletion. We prepared
cardiomyocytes overexpressing G0s2 and control cardiomyocytes.
During sustained hypoxia, [ATP]mito gradually declined in control
cardiomyocytes as measured by the Mit-ATeam assay. Notably,
the overexpression of G0s2 before the onset of hypoxia reduced
this decline in [ATP]mito, which allowed the cardiomyocytes to
promptly recover to baseline levels of [ATP]mito after reoxyge-
nation (Fig. 3 A and B and Movie S5). In addition, the prehypoxia
overexpression of G0s2 preserved cell viability during sustained
hypoxia (Fig. 3C). These results suggest that G0s2 can preserve

mitochondrial ATP production even under hypoxia and protect
cells from the energy crisis under hypoxia.

G0s2 Binds to FoF1-ATP Synthase but Not Other OXPHOS Protein
Complexes. To reveal the mechanism by which G0s2 affects
[ATP]mito, we sought to identify the biochemical targets of G0s2.
We screened for G0s2 binding proteins by immunoaffinity
purification of cell lysates from cardiomyocytes expressing
C-terminally Flag-tagged G0s2 (G0s2-Flag). G0s2-Flag is expressed
in cardiomyocytes localized to the mitochondria (Fig. S4A). MS
analysis revealed that multiple FoF1-ATP synthase subunits, but
no other mitochondrial respiratory chain complex subunits, were
coimmunoprecipitated with G0s2-Flag (Fig. S4B and Table S1).
FoF1-ATP synthase is a well-known ATP-producing enzyme
composed of a protein complex that contains an extramembra-
nous F1 and an intramembranous Fo domain linked by a periph-
eral and a central stalk (21–24). The binding of FoF1-ATP synthase
to G0s2-Flag was confirmed by immunoblotting with anti-
bodies against several subunits of FoF1-ATP synthase (Fig. 4A).

Fig. 2. G0s2, a hypoxia-inducible protein, affects
intramitochondrial ATP concentration in cardiomyo-
cytes. (A) Sequential YFP/CFP ratiometric pseu-
docolored images of Mit-ATeam fluorescence in
cardiomyocytes expressing (Upper) shRNAs for LacZ
(shLacZ) or (Lower) G0s2 (shG0s2). Oligomycin A
(1 μg/mL) was added at the end of the time-lapse
imaging to completely inhibit ATP synthesis. The
indicated time represents the period after adenovi-
rus infection. (B) Representative YFP/CFP ratiometric
pseudocolored images of Mit-ATeam fluorescence in
cardiomyocytes expressing the indicated adenovirus
for 24 h. (Scale bar: A and B, 20 μm.) (C) The bar
graph shows the mean YFP/CFP emission ratio of
Mit-ATeam fluorescence in cardiomyocytes express-
ing shLacZ (n = 30), shG0s2 #1 (n = 30), shG0s2 #2
(n = 29), and shG0s2 #2 + G0s2 WT (n = 32) for 24 h.
All of the measurements were normalized to the
average of the control cells (shLacZ). ***P < 0.001.
(D) Gene expression value plots of G0s2 (red line)
and VEGF-α (Vegfa; black line) levels in cardiomyo-
cytes under hypoxic conditions (1% O2). Each value
was compared with the level of Actb expression (n =
3). Values represent the means ± SEMs. (E) Immu-
noblotting of the G0s2 expression in cardiomyocytes
under hypoxic conditions (1% O2).

Fig. 3. Overexpression of G0s2 before hypoxia rescues the decline of mitochondrial ATP production during hypoxia. (A) Sequential YFP/CFP ratiometric
pseudocolored images of Mit-ATeam fluorescence in cardiomyocytes expressing (Upper) G0s2 WT or (Lower) LacZ during hypoxia and reoxygenation. (Scale
bar: 20 μm.) (B) YFP/CFP emission ratio plots of Mit-ATeam fluorescence in cardiomyocytes expressing G0s2 WT (n = 20) or LacZ (n = 19) during hypoxia and
reoxygenation. All of the measurements were normalized to the ratio at time 0 and compared between cardiomyocytes with G0s2 WT and LacZ at each time
point. (C) The bar graph shows the cell viability of cardiomyocytes overexpressing G0s2 under hypoxic conditions. Cardiomyocytes expressing either LacZ or
G0s2 WT were cultured under normoxic or hypoxic conditions for 18 h (n = 8). The asterisks denote statistical significance comparing G0s2 with LacZ. Data are
represented as the means ± SEMs. n.s., not significant. *P < 0.05; **P < 0.01; ***P < 0.001.

Kioka et al. PNAS | January 7, 2014 | vol. 111 | no. 1 | 275

C
EL
L
B
IO
LO

G
Y

D
o
w

n
lo

a
d
e
d
 a

t 
C

a
rn

e
g
ie

 M
e
llo

n
 U

n
iv

e
rs

ity
 o

n
 A

p
ri
l 5

, 
2
0
2
1
 

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/sm05.mov
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF4
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF4
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=ST1


Conversely, G0s2-Flag was coimmunoprecipitated with FoF1-
ATP synthase (Fig. S4C). G0s2-Flag was also found to be asso-
ciated with the FoF1-ATP synthase in 293T and HeLa cells (Fig.
S4C). Both coimmunoprecipitation using an anti-G0s2 antibody
and a reciprocal immunoprecipitation revealed that endogenous
G0s2 interacts with FoF1-ATP synthase, whereas none of the
proteins in complexes I–IV or adenine nucleotide translocase 1
(ANT1; also referred to as ADP/ATP carrier) were coimmuno-
precipitated with G0s2 (Fig. 4 B and C).
Given that the G0s2 protein contains an evolutionarily conserved

amino terminus and one hydrophobic domain (HD) (19), we cre-
ated three G0s2 partial deletion mutants to identify the domain in
G0s2 that is important for binding to FoF1-ATP synthase (Fig.
S4D). Among these mutants, G0s2 ΔC and G0s2 ΔN but not G0s2
ΔHD bound to the FoF1-ATP synthase complex (Fig. 4D and Fig.
S4 E and F). Furthermore, we confirmed that G0s2 directly inter-
acts with FoF1-ATP synthase in an in vitro pull-down assay using
a recombinant maltose-binding protein–fused G0s2 protein and
purified FoF1-ATP synthase from bovine heart mitochondria (Fig.

S5). Immunocytochemical analysis revealed that endogenous G0s2
colocalized with the β-subunit of FoF1-ATP synthase (Fig. 4E). The
knockdown of G0s2 expression by shRNA abolished G0s2 staining
(Figs. S6 and S7A), indicating that both antibodies used for
immunostaining specifically recognize G0s2. These data suggest
that G0s2 interacts with the FoF1-ATP synthase complex through
its HD in mitochondria and regulates OXPHOS activity.

G0s2 Increases Mitochondrial ATP Production Rate. [ATP]mito is
mainly determined by the rate of ATP synthesis by FoF1-ATP
synthase and ATP/ADP exchange by the ATP/ADP translocase
ANT1. This theory means that the increased [ATP]mito observed
in the G0s2-overexpressing cells may result from the increased
ATP synthesis and/or decreased ATP/ADP exchange, although
G0s2 did not interact with ANT1 (Fig. 4B). To resolve this issue
and directly measure the rate of ATP production in mitochon-
dria, we used a semiintact cell system called the mitochondrial
activity of streptolysin O permeabilized cells (MASC) assay (25).
In this assay, we permeabilized the plasma membrane to wash
out any cytosolic components, such as creatine and glycolytic
substrates, but left the mitochondria intact. Furthermore, we
treated the cells with P1, P5-di(adenosine-5′) pentaphosphate to
inhibit the activity of adenylate kinase. These steps allowed us to
measure the ATP production rate mostly from OXPHOS, with
a minimal contribution of ATP buffering systems in the cytosol.
The MASC assay was suitable for accurate measurement of mi-
tochondrial ATP production rate, because mitochondria in this
semiintact cell system suffered much smaller damage than the
isolated mitochondria in the conventional method. Surprisingly, in
the MASC assay, the ATP production rate markedly increased
when G0s2 was expressed in HeLa cells that lacked endogenous
G0s2 (Fig. 5A). In cardiomyocytes, shRNA-mediated G0s2
knockdown decreased the ATP production rate in mitochondria,
and the expression of G0s2 WT but not G0s2 ΔHD could restore
the ATP production rate (Fig. 5B and Fig. S7A). In both cells,
complete inhibition of ATP production by oligomycin A indicated
that the observed ATP synthesis was catalyzed by OXPHOS but
not other metabolism (Fig. 5 A and B).
Next, to evaluate the physiological role of G0s2, we examined

whether endogenous G0s2 induced by hypoxia could enhance the
ATP production rate. Cardiomyocytes were pretreated with hyp-
oxia for 4 h, during which G0s2 expression was largely induced.
We then evaluated the ATP production rate of both hypoxia-
pretreated and nontreated cardiomyocytes under room air con-
ditions. Even under these equivalent normoxic conditions, hyp-
oxia-pretreated cardiomyocytes produced ATP faster than
nontreated control cardiomyocytes (Fig. 5C and Fig. S7B). G0s2
knockdown attenuated this increase in the rate of ATP produc-
tion, indicating that the enhanced ATP production rate resulting
from hypoxia pretreatment primarily depends on endogenous
G0s2 induction. This increased G0s2 expression was essential for
cell survival, because G0s2-depleted cells died earlier than control
cells under conditions of hypoxic stress (Fig. 5D)
Furthermore, to assess the effect of G0s2 on cellular respiration,

we continuously measured the oxygen consumption rate (OCR)
using an XF96 Extracellular Flux Analyzer. G0s2 knockdown de-
creased the basal OCR of cardiomyocytes, most likely because of
the decreased activity of ATP synthesis (Fig. 5 E and F). In con-
trast, the proton leakage of the mitochondrial inner membrane
and the maximum respiratory capacity of OXPHOS complexes
I–IV were unaffected by G0s2 ablation (Fig. 5 E and F). These
data show that G0s2 knockdown reduced respiration caused by
ATP synthesis without affecting respiration caused by proton
leakage, nonmitochondrial respiration, or the maximal respira-
tion capacity.
All these findings indicate that G0s2 enhances the mitochondrial

ATP production rate by increasing the activity of FoF1-ATP synthase.

Discussion
In this study, we showed that G0s2 kinetically increased OXPHOS
activity through direct binding to FoF1-ATP synthase. Our previous

Fig. 4. G0s2 interacts with the FoF1-ATP synthase in mitochondria. (A) Im-
munoprecipitation (IP) of G0s2-Flag in cardiomyocytes. Cell lysates from
cardiomyocytes expressing G0s2-Flag or LacZ were immunoprecipitated with
an anti-Flag antibody. (B) IP of endogenous G0s2 in cardiomyocytes. En-
dogenous G0s2 was induced by hypoxia and immunoprecipitated using an
anti-G0s2 antibody. C, OXPHOS complex; FoF1, FoF1-ATP synthase. *IgG light
chain. (C) IP of FoF1-ATP synthase in cardiomyocytes under normoxic or
hypoxic conditions. Cell lysates from cardiomyocytes cultured under nor-
moxia or hypoxia for 4 h were immunoprecipitated with an antibody against
the whole FoF1-ATP synthase complex or a control IgG. *Nonspecific band.
(D) IP of G0s2 mutants expressed in cardiomyocytes. Cell lysates were immu-
noprecipitated with an anti-Flag antibody. (E) Immunostained images of
hypoxia-stimulated (4 h) cardiomyocytes with anti-G0s2 (green) and anti–FoF1-
ATP synthase β-subunit (red) antibodies. (Scale bars: 20 μm.)

276 | www.pnas.org/cgi/doi/10.1073/pnas.1318547111 Kioka et al.

D
o
w

n
lo

a
d
e
d
 a

t 
C

a
rn

e
g
ie

 M
e
llo

n
 U

n
iv

e
rs

ity
 o

n
 A

p
ri
l 5

, 
2
0
2
1
 

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF4
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF4
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF4
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF4
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF4
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF4
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF4
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF5
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF5
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF6
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF7
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF7
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF7
www.pnas.org/cgi/doi/10.1073/pnas.1318547111


studies of FoF1-ATP synthase have revealed that this enzyme has
a specific structure that connects two molecular nanomotors that
synchronize with each other to produce ATP (26–30). These
physically distinct structures suggest that a specific activating fac-
tor for FoF1-ATP synthase must exist. Combined with the find-
ings from this study, we hypothesize that G0s2 may lower the
activation barrier of the FoF1-ATP synthase nanomotor and
enhance the ATP production rate with the equivalent proton
motive driving force (PMF; i.e., the sum of the membrane po-
tential and the pH gradient). Activation barriers might be gen-
erated by various factors, such as friction between the stator and
rotor of FoF1-ATP synthase, physical and electrical resistance to
proton transport through the channel, and the existence of ro-
tary blockers such as the bacterial e-subunit and cyclophilin D
(31). The increased ATP production rate caused by G0s2 over-
expression observed in the MASC assay supports this hypothesis,
because the PMF in the initial phase of this assay should be the
same. If this hypothesis is true, even with reduced PMF, cells that
express G0s2 should produce ATP faster than cells that express

little or no G0s2. In fact, G0s2 overexpression attenuated the de-
cline of [ATP]mito under hypoxic conditions that reduced the PMF.
Precise real-time measurement of the PMF is currently difficult, but
these hypotheses might be proven in future studies. Kinetically
faster ATP production should accompany greater consumption of
both O2 and PMF; however, our results suggest that preserving
ATP production is more beneficial than preserving PMF for cell
viability, particularly when the O2 supply is restricted but still exists.
The transience of endogenous G0s2 expression induced by hypoxia
might serve to protect tissues in the early phase of energy crisis.
There may be specific mechanisms to decrease G0s2 expression
under prolonged ischemia that have yet to be identified. Another
possible mechanism by which G0s2 could increase the ATP pro-
duction rate is that G0s2 increases the Fo-F1 coupling efficiency of
FoF1-ATP synthase. However, this hypothesis is less likely, because
G0s2 altered the oxygen consumption rate to increase the ATP
production rate. Although this uncoupling phenomenon has rarely
been reported for mammalian mitochondrial FoF1-ATP synthase,
we cannot completely eliminate the possibility that intrinsically

Fig. 5. G0s2 enhances the mitochondrial ATP production rate. (A and B) MASC assay of (A) permeabilized HeLa cells expressing the indicated plasmids or (B) car-
diomyocytes expressing the indicated adenovirus in the presence (dotted lines) or absence (solid lines) of 1 μg/mL oligomycin A (Oli A). Upper shows the ATP pro-
duction plots, and Lower shows the mean ATP production rates between 0 and 10 min. (A) n = 12. (B) Solid lines, n = 12; dotted lines, n = 8. (C) MASC assay of
permeabilized cardiomyocytes pretreated with hypoxia. Cells expressing the indicated adenovirus were pretreated with or without hypoxia for 4 h. After the pre-
treatment, the cells were permeabilized under room air conditions followed by MASC assay in the presence (dotted lines; n = 8) or absence (solid lines; n = 12) of 1 μg/
mL Oli A. Upper shows the ATP production plot, and Lower shows the mean ATP production rate between 0 and 10 min. (D) The bar graph represents the cell viability
of G0s2-depleted cardiomyocytes under hypoxic conditions. Cardiomyocytes expressing shLacZ or shG0s2 (#2) were cultured under normoxic or hypoxic conditions for
18 h. (E and F) The OCR in cardiomyocytes expressing shLacZ and shG0s2 (#2) under basal conditions and in response to the indicated mitochondrial inhibitors (n = 8).
FCCP, carbonyl cyanide-4-(trifluoromethoxy)-phenylhydrazone. Data are represented as the means ± SEMs. n.s., not significant. **P < 0.01; ***P < 0.001.

Kioka et al. PNAS | January 7, 2014 | vol. 111 | no. 1 | 277

C
EL
L
B
IO
LO

G
Y

D
o
w

n
lo

a
d
e
d
 a

t 
C

a
rn

e
g
ie

 M
e
llo

n
 U

n
iv

e
rs

ity
 o

n
 A

p
ri
l 5

, 
2
0
2
1
 



uncoupled FoF1-ATP synthase exists, because we could not ac-
curately measure the amount of uncoupled FoF1-ATP synthase in
intact cells.
G0s2 was first identified in cultured monocytes during the

drug-induced cell cycle transition from G0 to G1 phase (18, 32).
A limited number of studies have implied that G0s2 is involved
in cell proliferation (33), differentiation (19), apoptosis (34),
inflammation (35), and lipid metabolism (36) in various cellular
settings. Moreover, G0s2 was reported to localize to the cytosol
(33), endoplasmic reticulum (19), mitochondria (34), or the
surface of lipid droplets (36). How G0s2 distinguishes these
multiple functions is still not clear. In our hands, G0s2 is always
localized to mitochondria, which was shown by immunostaining
with two antibodies against different epitopes of G0s2 (Fig.
S6). Complete depletion of mitochondrial staining by G0s2
knockdown strongly suggests the specific localization of G0s2 to
mitochondria. We also showed that G0s2 specifically bound to
mitochondrial FoF1-ATP synthase but not other OXPHOS
protein complexes and functionally regulated OXPHOS activity.
Together, these data suggest that G0s2 acts in the mitochondria.
However, different cellular conditions may change the localiza-
tion and role of G0s2. Additionally, G0s2-mediated changes in
ATP metabolism may possibly affect the lipid metabolism or
cellular proliferation. Additional studies will reveal the
functional mechanisms by which G0s2 exerts these multiple
functions in different cellular conditions.
In this study, we evaluated [ATP]mito and [ATP]cyto separately

using FRET-based ATP biosensors in living cells. This dual
evaluation revealed that [ATP]mito reflected mitochondrial ATP
production with much greater sensitivity than [ATP]cyto (Fig. 1
and Movies S1 and S2). Because [ATP]cyto is strongly influenced
by the activity of various cytosolic ATP hydrolytic enzymes and

ATP buffering enzymes, [ATP]cyto does not always reflect the
ATP availability that determines cellular function.
Taken together, our results indicate that G0s2 is a positive

regulator of OXPHOS that works to increase the mitochondrial
ATP production rate even under hypoxic conditions. Therefore,
enhancing the level and function of G0s2 could be beneficial for
hypoxia- and mitochondria-related disorders, such as ischemic
diseases, metabolic diseases, and cancer.

Materials and Methods
Cells were infected with adenovirus encoding FRET-based ATP indicators AT1.03
or mit-AT1.03 to measure changes in cytosolic or mitochondrial ATP concen-
trations, respectively. Image acquisitions and FRET analyses were performed as
described previously with some modifications (13). For the control of oxygen
concentration during time-lapse imaging, digital gas mixer for stage-top in-
cubator GM8000 (Tokai Hit) was used to create hypoxic (1% O2) or normoxic
(20% O2) condition. Additional methods are found in SI Materials and Methods.

ACKNOWLEDGMENTS. We thank M. Murata for helpful discussions and
advice, H. Miyagi (Olympus Co. Ltd.) for technical advice regarding mi-
croscopy, T. Miyazaki (Cyclex Co. Ltd.) for making antibodies, S. Ikezawa and
A. Ogai for technical assistance, K. Tanaka for help with the purification of
bovine FoF1-ATP synthase, and Y. Okada and H. Fujii for secretarial support.
This research was supported by the Japan Society for the Promotion of Science
through the Funding Program for Next Generation World-Leading Researchers
(NEXT Program) initiated by the Council for Science and Technology Policy;
grants-in-aid from the Ministry of Health, Labor, and Welfare–Japan; and
grants-in-aid from the Ministry of Education, Culture, Sports, Science, and
Technology–Japan. This research was also supported by grants from Takeda
Science Foundation, Japan Heart Foundation, Japan Cardiovascular Research
Foundation, Japan Intractable Diseases Research Foundation, Japan Founda-
tion of Applied Enzymology, Japan Medical Association, Uehara Memorial
Foundation, Mochida Memorial Foundation, Banyu Foundation, Naito Foun-
dation, Inoue Foundation for Science, Osaka Medical Research foundation for
Intractable Diseases, Ichiro Kanehara Foundation, and Showa Houkoukai.

1. Kim JW, Tchernyshyov I, Semenza GL, Dang CV (2006) HIF-1-mediated expression of
pyruvate dehydrogenase kinase: A metabolic switch required for cellular adaptation
to hypoxia. Cell Metab 3(3):177–185.

2. Papandreou I, Cairns RA, Fontana L, Lim AL, Denko NC (2006) HIF-1 mediates adaptation to
hypoxia by actively downregulating mitochondrial oxygen consumption. Cell Metab 3(3):
187–197.

3. Semenza GL (2012) Hypoxia-inducible factors in physiology and medicine. Cell 148(3):
399–408.

4. Semenza GL, et al. (1996) Hypoxia response elements in the aldolase A, enolase 1, and
lactate dehydrogenase A gene promoters contain essential binding sites for hypoxia-
inducible factor 1. J Biol Chem 271(51):32529–32537.

5. Chen YC, et al. (2012) Identification of a protein mediating respiratory supercomplex
stability. Cell Metab 15(3):348–360.

6. Fukuda R, et al. (2007) HIF-1 regulates cytochrome oxidase subunits to optimize ef-
ficiency of respiration in hypoxic cells. Cell 129(1):111–122.

7. Strogolova V, Furness A, Robb-McGrath M, Garlich J, Stuart RA (2012) Rcf1 and Rcf2,
members of the hypoxia-induced gene 1 protein family, are critical components of
the mitochondrial cytochrome bc1-cytochrome c oxidase supercomplex. Mol Cell Biol
32(8):1363–1373.

8. Saks V, et al. (2006) Cardiac system bioenergetics: Metabolic basis of the Frank-Star-
ling law. J Physiol 571(Pt 2):253–273.

9. Smolenski RT, Lachno DR, Ledingham SJ, Yacoub MH (1990) Determination of sixteen
nucleotides, nucleosides and bases using high-performance liquid chromatography
and its application to the study of purine metabolism in hearts for transplantation.
J Chromatogr A 527(2):414–420.

10. Shimura D, et al. (2013) Metabolomic profiling analysis reveals chamber-dependent me-
tabolite patterns in the mouse heart. Am J Physiol Heart Circ Physiol 305(4):H494–H505.

11. Kemp GJ, Meyerspeer M, Moser E (2007) Absolute quantification of phosphorus
metabolite concentrations in human muscle in vivo by 31P MRS: A quantitative re-
view. NMR Biomed 20(6):555–565.

12. Ford SR, et al. (1996) Use of firefly luciferase for ATP measurement: Other nucleotides
enhance turnover. J Biolumin Chemilumin 11(3):149–167.

13. Imamura H, et al. (2009) Visualization of ATP levels inside single living cells with
fluorescence resonance energy transfer-based genetically encoded indicators. Proc
Natl Acad Sci USA 106(37):15651–15656.

14. Lopaschuk GD, Kelly DP (2008) Signalling in cardiac metabolism. Cardiovasc Res 79(2):
205–207.

15. Hattori F, et al. (2010) Nongenetic method for purifying stem cell-derived car-
diomyocytes. Nat Methods 7(1):61–66.

16. Forsythe JA, et al. (1996) Activation of vascular endothelial growth factor gene
transcription by hypoxia-inducible factor 1. Mol Cell Biol 16(9):4604–4613.

17. Wolf A, et al. (2011) Hexokinase 2 is a key mediator of aerobic glycolysis and pro-
motes tumor growth in human glioblastoma multiforme. J Exp Med 208(2):313–326.

18. Russell L, Forsdyke DR (1991) A human putative lymphocyte G0/G1 switch gene
containing a CpG-rich island encodes a small basic protein with the potential to be
phosphorylated. DNA Cell Biol 10(8):581–591.

19. Zandbergen F, et al. (2005) The G0/G1 switch gene 2 is a novel PPAR target gene.
Biochem J 392(Pt 2):313–324.

20. Heckmann BL, Zhang X, Xie X, Liu J (2013) The G0/G1 switch gene 2 (G0S2): Regu-
lating metabolism and beyond. Biochim Biophys Acta 1831(2):276–281.

21. Dimroth P, von Ballmoos C, Meier T (2006) Catalytic and mechanical cycles in F-ATP
synthases. Fourth in the Cycles Review Series. EMBO Rep 7(3):276–282.

22. Senior AE (2007) ATP synthase: Motoring to the finish line. Cell 130(2):220–221.
23. Walker JE (1998) ATP synthesis by rotary catalysis (Nobel Lecture). Angew Chem Int Ed

37:5000–5011.
24. Yoshida M, Muneyuki E, Hisabori T (2001) ATP synthase—a marvellous rotary engine

of the cell. Nat Rev Mol Cell Biol 2(9):669–677.
25. Fujikawa M, Yoshida M (2010) A sensitive, simple assay of mitochondrial ATP synthesis

of cultured mammalian cells suitable for high-throughput analysis. Biochem Biophys
Res Commun 401(4):538–543.

26. Adachi K, et al. (2007) Coupling of rotation and catalysis in F(1)-ATPase revealed by
single-molecule imaging and manipulation. Cell 130(2):309–321.

27. Itoh H, et al. (2004) Mechanically driven ATP synthesis by F1-ATPase. Nature
427(6973):465–468.

28. Noji H, Yasuda R, Yoshida M, Kinosita K, Jr. (1997) Direct observation of the rotation
of F1-ATPase. Nature 386(6622):299–302.

29. Rondelez Y, et al. (2005) Highly coupled ATP synthesis by F1-ATPase single molecules.
Nature 433(7027):773–777.

30. Uchihashi T, Iino R, Ando T, Noji H (2011) High-speed atomic force microscopy reveals
rotary catalysis of rotorless F₁-ATPase. Science 333(6043):755–758.

31. Giorgio V, et al. (2009) Cyclophilin D modulates mitochondrial F0F1-ATP synthase by
interacting with the lateral stalk of the complex. J Biol Chem 284(49):33982–33988.

32. Siderovski DP, Blum S, Forsdyke RE, Forsdyke DR (1990) A set of human putative
lymphocyte G0/G1 switch genes includes genes homologous to rodent cytokine and
zinc finger protein-encoding genes. DNA Cell Biol 9(8):579–587.

33. Yamada T, Park CS, Burns A, Nakada D, Lacorazza HD (2012) The cytosolic protein
G0S2 maintains quiescence in hematopoietic stem cells. PLoS ONE 7(5):e38280.

34. Welch C, et al. (2009) Identification of a protein, G0S2, that lacks Bcl-2 homology
domains and interacts with and antagonizes Bcl-2. Cancer Res 69(17):6782–6789.

35. Kobayashi S, et al. (2008) Expression profiling of PBMC-based diagnostic gene
markers isolated from vasculitis patients. DNA Res 15(4):253–265.

36. Yang X, et al. (2010) The G(0)/G(1) switch gene 2 regulates adipose lipolysis through
association with adipose triglyceride lipase. Cell Metab 11(3):194–205.

278 | www.pnas.org/cgi/doi/10.1073/pnas.1318547111 Kioka et al.

D
o
w

n
lo

a
d
e
d
 a

t 
C

a
rn

e
g
ie

 M
e
llo

n
 U

n
iv

e
rs

ity
 o

n
 A

p
ri
l 5

, 
2
0
2
1
 

http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF6
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=SF6
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/sm01.mov
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/sm02.mov
http://www.pnas.org/lookup/suppl/doi:10.1073/pnas.1318547111/-/DCSupplemental/pnas.201318547SI.pdf?targetid=nameddest=STXT
www.pnas.org/cgi/doi/10.1073/pnas.1318547111