Na+/Ca2+ exchanger mediates cold Ca2+ signaling conserved for temperature-compensated circadian rhythms

Abstract

Circadian rhythms are based on biochemical oscillations generated by clock genes/proteins, which independently evolved in animals, fungi, plants, and cyanobacteria. Temperature compensation of the oscillation speed is a common feature of the circadian clocks, but the evolutionary-conserved mechanism has been unclear. Here, we show that Na+/Ca2+ exchanger (NCX) mediates cold-responsive Ca2+ signaling important for the temperature-compensated oscillation in mammalian cells. In response to temperature decrease, NCX elevates intracellular Ca2+, which activates Ca2+/calmodulin-dependent protein kinase II and accelerates transcriptional oscillations of clock genes. The cold-responsive Ca2+ signaling is conserved among mice, Drosophila, and Arabidopsis The mammalian cellular rhythms and Drosophila behavioral rhythms were severely attenuated by NCX inhibition, indicating essential roles of NCX in both temperature compensation and autonomous oscillation. NCX also contributes to the temperature-compensated transcriptional rhythms in cyanobacterial clock. Our results suggest that NCX-mediated Ca2+ signaling is a common mechanism underlying temperature-compensated circadian rhythms both in eukaryotes and prokaryotes.

Full text

Kon et al., Sci. Adv. 2021; 7 : eabe8132 30 April 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
1 of 12
NEUROSCIENCE
Na+/Ca2+ exchanger mediates cold Ca2+ signaling
conserved for temperature-compensated
circadian rhythms
Naohiro Kon1, Hsin-tzu Wang1, Yoshiaki S. Kato2, Kyouhei Uemoto3,4, Naohiro Kawamoto5,
Koji Kawasaki5, Ryosuke Enoki6,7, Gen Kurosawa8, Tatsuto Nakane9, Yasunori Sugiyama9,
Hideaki Tagashira10, Motomu Endo3, Hideo Iwasaki5, Takahiro Iwamoto10*,
Kazuhiko Kume2, Yoshitaka Fukada1*
Circadian rhythms are based on biochemical oscillations generated by clock genes/proteins, which independently
evolved in animals, fungi, plants, and cyanobacteria. Temperature compensation of the oscillation speed is a com-
mon feature of the circadian clocks, but the evolutionary-conserved mechanism has been unclear. Here, we
show that Na+/Ca2+ exchanger (NCX) mediates cold-responsive Ca2+ signaling important for the temperature-
compensated oscillation in mammalian cells. In response to temperature decrease, NCX elevates intracellular
Ca2+, which activates Ca2+/calmodulin-dependent protein kinase II and accelerates transcriptional oscillations of
clock genes. The cold-responsive Ca2+ signaling is conserved among mice, Drosophila, and Arabidopsis. The mam-
malian cellular rhythms and Drosophila behavioral rhythms were severely attenuated by NCX inhibition, indicating
essential roles of NCX in both temperature compensation and autonomous oscillation. NCX also contributes to
the temperature-compensated transcriptional rhythms in cyanobacterial clock. Our results suggest that NCX-
mediated Ca2+ signaling is a common mechanism underlying temperature-compensated circadian rhythms both
in eukaryotes and prokaryotes.
INTRODUCTION
Among a wide variety of biological functions, the circadian clock is
of particular interest because of its unique property, i.e., temperature-
compensated oscillation with a period of approximately 24 hours
(1). Generally, an increase in temperature by 10°C accelerates rates
of biochemical reactions by two- to threefold (Q10=2 to 3), whereas
Q10 of the oscillation speed of the clock is 0.8 to 1.2. The property
was originally termed temperature independence, but later termed
temperature compensation on the basis of the finding of overcompen-
sation for the effect of temperature on the period length in photo-
synthetic dinoflagellates (Lingulodinium polyedra) (1). The temperature
compensation is a common property of the circadian clocks, impli-
cating that a mechanism underlying the compensation is tightly
associated with machinery for cell-autonomous oscillation.
Most of the overt circadian rhythms are based on biochemical
oscillations generated by clock genes and their encoded proteins
(2–5). Homologies of the clock genes are limited among animals,
fungi, plants, and cyanobacteria, suggesting that the clock genes in-
dependently evolved after divergence of the lineages. In cyano bacteria,
KaiC phosphorylation rhythms constitute a core circadian oscillator
termed posttranslational oscillator (PTO) (5). The phosphorylation
rhythms of KaiC in the KaiA-KaiB-KaiC protein complex are tem-
perature compensated invitro. In eukaryotes, clock genes and their
encoded proteins constitute transcriptional/translational feedback
loops (TTFLs) (2–4). Because a HES (Hairy and Enhancer of Split)–
based TTFL in segmentation clock is temperature sensitive (Q10=2
to 3) (6), temperature compensation is not a general property in-
trinsic to TTFLs. This suggests the existence of an important mech-
anism regulating the circadian oscillation of the TTFLs.
Historically, before the discovery of the clock genes, a feedback
system involving ions and ion regulators in plasma membranes was
proposed as the oscillation mechanism of the circadian clock (7).
This “membrane model” is based on the observation that the circadian
rhythms are notably affected by manipulating ion concentra-
tions or ion regulator activities in various eukaryotes (7). To date,
several ions, especially Ca2+, have been shown to play an essential
role for oscillation of the TTFLs in mammals (8), insects (9), and
plants (10). In mice and Drosophila, intracellular Ca2+ levels were
shown to exhibit robust circadian oscillations (11–13), which elicit
rhythmic activation of Ca2+/calmodulin-dependent protein kinase
II (CaMKII) (14–16). CaMKII phosphorylates CLOCK to activate
CLOCK-BMAL1 heterodimer, a key transcriptional activator in the
animal TTFLs (3). The upstream regulator of the Ca2+-dependent
phosphorylation signaling has been a missing link between the TTFL
and the membrane model.
RESULTS
Ca2+ signaling is a key for
temperature-compensated oscillation
To uncover key regulators involved in temperature-compensated os-
cillation in mammals, we investigated effects of various small-molecule
1Department of Biological Sciences, School of Science, The University of Tokyo, Hongo
7-3-1, Bunkyo-ku, Tokyo 113-0033, Japan. 2Department of Neuropharmacology,
Graduate School of Pharmaceutical Sciences, Nagoya City University, Nagoya 467-8603 ,
Japan. 3Graduate School of Biological Science, Nara Institute of Science and Tech-
nology, Ikoma 630-0192, Japan. 4Graduate School of Biostudies, Kyoto University,
Kyoto 606-8501, Japan. 5Department of Electrical Engineering and Bioscience, Waseda
University, Tokyo 162-8480, Japan. 6Biophotonics Research Group, Exploratory Re-
search Center on Life and Living Systems (ExCELLS), National Institutes of Natural
Sciences, Higashiyama 5-1, Myodaiji, Okazaki, Aichi 444-8787, Japan. 7Division of
Biophotonics, National Institute for Physiological Sciences, National Institutes of Nat-
ural Sciences, Higashiyama 5-1, Myodaiji, Okazaki, Aichi 444-8787, Japan. 8iTHEMS,
RIKEN, Wako 351-0198, Japan. 9Department of Life Sciences, Faculty of Agriculture,
Kagawa University, Kagawa 761-0795, Japan. 10Department of Pharmacology, Fac-
ulty of Medicine, Fukuoka University, Fukuoka 814-0180, Japan.
*Corresponding author. Email: tiwamoto@fukuoka-u.ac.jp (T.I.); sfukada@mail.
ecc.u-tokyo.ac.jp (Y.F.)
Copyright © 2021
The Authors, some
rights reserved;
exclusive licensee
American Association
for the Advancement
of Science. No claim to
original U.S. Government
Works. Distributed
under a Creative
Commons Attribution
License 4.0 (CC BY).

Kon et al., Sci. Adv. 2021; 7 : eabe8132 30 April 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
2 of 12
compounds targeting protein kinases or ion regulators (table S1)
on cellular rhythms of Rat-1 fibroblasts stably expressing Bmal1-
luciferase reporter (14,17). A Q10 value was calculated from the
period lengths of the bioluminescence rhythms recorded at 32° and
37°C (figs. S1, A and B, and S2A). All the screening results were
evaluated by using a Q10 value, which was defined as a difference
of the Q10 values between drug-treated cells and control [0.1%
dimethyl sulfoxide (DMSO)–treated] cells (Fig.1A). We found a
remarkable increase in Q10 by CaMKII inhibitor KN-93 (Fig.1A) in
a dose-dependent manner (Fig.1,BandC). The treatment with
10 M KN-93 shortened the period at 37°C, whereas it lengthened
the period at 32°C (Fig.1D). Such a temperature-dependent bidirectional
effect of KN-93 was unique in that many compounds showed a
unidirectional period-modifying effect at 32° and 37°C (fig. S1C). KN-92,
an inactive analog of KN-93, had no significant effect on Q10 value
(Fig.1E and fig. S2B), supporting the specific effect of KN-93 on CaMKII.
In detailed analysis of the effects of the compounds, we noticed
that the treatment with KB-R7943, an inhibitor of Na+/Ca2+ exchanger
(NCX) (18), increased the Q10 value in a dose-dependent manner
(Fig.1,BandC). Similar to the CaMKII inhibitor, KB-R7943 exhib-
ited the temperature-dependent bidirectional effect on the circadian
period (Fig.1D). Another NCX inhibitor, SEA0400 (18), also showed
the bidirectional period-modifying effect (Fig.1D) and the Q10-
increasing effect (Fig.1E). On the other hand, none of these effects
were observed after treatment of Rat-1 cells with nifedipine and
verapamil, blockers of L-type Ca2+ channel, or with IC261, a period-
lengthening inhibitor of casein kinase I (Fig.1,C andD, and fig.
S3) (15).
The period-modifying effects of KN-93 and KB-R7943 were further
analyzed at various temperatures between 32° and 37°C (fig. S4). As
a control, Rat-1 cells treated with DMSO showed shorter periods at
lower temperatures (Fig.2A), a phenomenon termed overcompen-
sation observed in a wide range of species (1–5). In contrast, the
oscillation speed was slowed down by decreasing the temperature in
the presence of KN-93 or KB-R7943 (Fig. 2A), and this period-
lengthening effect was particularly obvious below 35°C (Fig.2B).
The Q10 value calculated from the circadian periods at 32° and 35°C
was 0.89 (vehicle), 1.49 (20 M KB-R7943), or 2.01 (10 M KN-93).
It is evident that the overcompensated oscillation becomes tempera-
ture sensitive by inhibiting CaMKII or NCX activity. These results
together demonstrate that CaMKII and NCX are key players for
temperature compensation in the mammalian cellular clock.
NCX-Ca2+-CaMKII signaling is important for cellular
circadian oscillation
Note that the KB-R7943 treatment of Rat-1 fibroblasts decreased the
amplitude of the cellular rhythms (Fig.1B). Among the chemicals
targeting ion channels and transporters, only KB-R7943 suppressed
the relative amplitude of the rhythms (Fig.3A), suggesting an im-
portant role of NCX in the cell-autonomous oscillation mechanism,
in addition to the temperature compensation.
NCX exchanges 3 Na+ for 1 Ca2+ across the plasma membrane.
NCX is a unique bidirectional regulator of cytosolic Ca2+ concen-
tration because it can mediate both Ca2+ influx and efflux, depending
on not only the membrane potential but also local concentrations of
Na+ and Ca2+ (18). In response to an increase in cytoplasmic Ca2+
levels, NCX mediates Ca2+ efflux, while NCX can maintain steady-
state levels of intracellular Ca2+ by promoting Ca2+ influx in several
types of cells (18). We examined roles of NCX in regulation of intracel lular
Ca2+ levels in NIH3T3 fibroblasts. Fluo-4 acetoxymethyl ester
(Fluo- 4 AM)–based Ca2+ imaging revealed that the basal fluorescence
level in the cultured cells was remarkably reduced by the addition of
5 to 20 M NCX inhibitor KB-R7943 to the culture medium
(Fig.3B), indicating that NCX contributes to net Ca2+ influx in the
quiescent state. Then, we evaluated the effect of KB-R7943 on cellular
CaMKII activity, which reflects intracellular Ca2+ level (14). One-day
treatment of NIH3T3 cells with 20 M KB-R7943 significantly
decreased the CaMKII activity toward syntide-2, a model substrate
specific to CaMKII (Fig.3C) (14). These results reveal an important
role of NCX in the maintenance of the activity level of Ca2+-CaM-
KII signaling in the fibroblasts.
To address the role of NCX-Ca2+-CaMKII signaling in cell-
autonomous oscillation, we investigated effects of chronic inhibition
of the signaling on circadian rhythms in Rat-1 reporter cells (14,17).
The relative amplitude of the cellular bioluminescence rhythm
detected by Bmal1-luc was markedly reduced by chronic treatment
with KN-93, trifluoperazine (calmodulin antagonist), 1,2-bis(2-
aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA)–AM
(intracellular Ca2+ chelator), or KB-R7943 (Fig.3D). The amplitude-
reducing effect by chronic treatment with KB-R7943 or SEA0400
was also observed in Per2-luc reporter cells (Fig.3E). The severe
damping of the transcriptional rhythms was reversed by washing
out the drug- containing medium (Fig.3E). On the other hand, a
pulse treatment (for 1 to 2 hours) of Rat-1–Bmal1-luc cells with
KN-93, KB-R7943, or SEA0400 caused a phase-dependent phase
shift of the bioluminescence rhythms with their maximal responses
at circadian time (CT)21 (Fig.3F). The amplitude-reducing effects
(Fig.3,DandE) and the overt phase-resetting actions of the inhibi-
tors (Fig.3F) together suggest that NCX-dependent Ca2+-CaMKII
signaling functions as a state variable of the circadian oscillator in a
limit cycle interpretation (fig. S5) (14,19).
Lowering temperature activates NCX-Ca2+-CaMKII signaling
In the experiments examining the relationship between temperature
and the cellular rhythms, we found that the amplitude of the rhythm
was decreased by lowering temperature particularly below 34°C
(Fig.4A, DMSO, and fig. S6). In the same range of temperatures,
KN-93 treatment (2 to 10 M) caused a much larger decrease in the
amplitude in a dose-dependent manner (Fig.4,A andB, and fig.
S6). The results indicate that CaMKII activity compensates for am-
plitude decrease in the TTFL at temperatures below 34°C.
Note that hypothermia is clinically defined as a drop in core body
temperature below 35°C (20). We hypothesized that Ca2+ signaling
may be activated for cold response in mammalian cells, as reported
for cold tolerance mechanism of insects and plant cells (21). This
idea was tested by investigating intracellular Ca2+ levels in cultured
fibroblasts. In Fluo-4 AM–based Ca2+ imaging, lowering of the tem-
perature from 37° to 25°C significantly increased free Ca2+ levels in
NIH3T3 cells (Q10=0.73) (Fig.4C). The hypothermic response was
blocked by treatment with SEA0400 or KB-R7943 (Fig.4D). We found
that the CaMKII activity of lysates prepared from the cells cultured
at 27°C was higher than that at 37°C (Q10=0.78) (Fig.4E, DMSO).
In addition, the hypothermic activation of CaMKII was inhibited in
the cells cultured with 20 M KB-R7943 (Fig.4E). These results in-
dicate that NCX enhances Ca2+ influx and activates CaMKII signaling
in response to the temperature decrease in the mammalian cells.
We then examined how intracellular Ca2+ levels affect the clock
gene expression rhythms. In Rat-1–Bmal1-luc cells, 1-hour treatment

Kon et al., Sci. Adv. 2021; 7 : eabe8132 30 April 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
3 of 12
E
1086420
20
22
24
26
28
Period (hours)
37ºC
32ºC
20
21
22
23
24
25
10020
37ºC
32ºC
515
C
D
Relative amplitude
1.0
0.5
0
–1.0
–0.5
1.0
0.5
0
–1.0
–0.5
21345
21345
37ºC
32ºC
10
8
6
4
2
0
KN-93 (µM)
1.0
0.5
0
–1.0
–0.5
1.0
0.5
0
–1.0
–0.5
21345
21 345
20
10
0
KB-R7943 (µM)
37ºC
32ºC
Time (days)Time (days)
B
∆Q10 of period
0.6
0.5
0.4
0.3
0.2
0.1
0.0
–0.1
–0.2
A
Roscovitine
SB203580
SB202190
DRB
SP600125
TG003
D4476
LY294002
Rapamycin
PI-103
Lidocaine
Amiloride
Trifluoperazine
Mepivacaine
Flecainide
Calmidazolium
KNK437
Wortmannin
Nifedipine
Ru360
DS16570511
Verapamil
Nilotinib
Benzamil
IC261
TTFA
DS44170716
Diltiazem
Go6983
TBBt
Furosemide
DMAT
SB431542
HMA
Veratridine
Bay-K-8644
NPPB
Mibefradil
Phloretin
SKF-96365
KN-93
KB-R7943
SB216763
19
20
1002030
21
22
23
37ºC
32ºC
KN-93 KB-R7943 SEA0400
29
(µM)
10 µM KN-92
DMSO
∆Q10 of period
20 10 30
SEA0400
(µM)
0.4
0.3
0.2
0.1
0
10 µM KN-93
100 20
0.8
1.0
1.2
1.4
1.6
1.8
Q10 of period
5 15
KN-93
KB-R7943
200
0.8
1.0
1.2
1.4
1.6
1.8
10 30 200
0.8
1.0
1.2
1.4
1.6
1.8
10 30 200
0.8
1.0
1.2
1.4
1.6
1.8
10 30
Nifedipine Verapamil IC261
100203010020301002030
20
22
24
25
Period (hours)
37ºC
32ºC
Nifedipine
23
21
Verapamil IC261
37ºC
32ºC
37ºC
32ºC
20
22
24
25
23
21
20
22
24
25
23
21
(µM)
(µM)
(µM)
(µM)
(
µ
M)
(µM) (µM) (µM) (µM)
Fig. 1. CaMKII and NCX activities are essential for temperature compensation. (A) Effects of chemical inhibitors on Q10 of bioluminescence rhythms in Rat-1–Bmal1-
luc cells. To normalize experiment-to-experiment variations, Q10 values compared to the vehicle (DMSO) control were used for comparison of the all screening data. The
waveforms and the other parameters are shown in figs. S1 to S3. (B) Representative bioluminescence rhythms of Rat-1–Bmal1-luc cells in the presence of KN-93 (left) or
KB-R7943 (right) at 32° or 37°C. (C) Dose-dependent effect of KN-93 or KB-R7943 on Q10. ★P < 0.05 compared to DMSO (Dunnett’s test). (D) Dose- and temperature-dependen t
effect of KN-93, KB-R7943, or SEA0400 on period length at 32° or 37°C. ★P < 0.05 compared to DMSO (Dunnett’s test). (E) Effect of KN-92, KN-93, or SEA0400 on Q10 value.
We used a concentration of 10 M consistently in the first screening (A), and two compounds, KN-93 and KB-R7943, met our criteria. Then, we performed reproducibility
test and dose dependency test for the two compounds with several control compounds (B to E). Representative data (B) or the means with SEM from three independent
samples (A and C to E) are shown.

Kon et al., Sci. Adv. 2021; 7 : eabe8132 30 April 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
4 of 12
with Ca2+ ionophore, ionomycin, or A23187 up-regulated transcripts
of Per1, Per2, and Dec1 (Fig.4F), which are regulated by CaMKII
(14,16) through E-box and/or CRE, a DNA cis-element responsive
to Ca2+/cyclic adenosine monophosphate signaling (3). Consistently,
a decrease in the temperature from 37°C down to 29°C elevated the
expression levels of many E-box–regulated genes, such as Per1, Per2,
Per3, Dec1, Dec2, Rev-erb, Rev-erb, and Cry1 (Fig.4G). In addi-
tion, E4bp4, which is regulated by Ca2+-NFAT signaling (22), is also
up-regulated by the temperature decrease (Fig.4G). Consistent with
the decrease in relative amplitude of the bioluminescence rhythms
at lower temperatures (Fig.4A), a peak-trough ratio of Bmal1 ex-
pression rhythm was reduced by lowering the temperature (Fig.4G).
We found that the hypothermic up-regulation of Per1 and Per2
transcripts was significantly attenuated in the presence of NCX in-
hibitor KB-R7943 or CaMKII inhibitor KN-93 (Fig.4H). These results
together indicate that the temperature changes have a marked influence
on the clock gene expression levels through NCX-Ca2+-CaMKII
signaling.
Cold-responsive Ca2+ signaling compensates for slowdown
of TTFL at lower temperature
In 1957, Hastings and Sweeney (1) hypothesized that temperature
compensation of the circadian clock is based on a combination of
temperature-sensitive period-shortening and period-lengthening
processes. Most biochemical reactions in the TTFL are slowed
down by decreasing the temperature (table S2). In an invitro assay,
kinase activity of purified CaMKII toward a CLOCK peptide (Ser/
Pro-rich region of CLOCK) (15) was reduced by lowering the tem-
perature (Q10=2.9) (Fig.5A). In contrast, as described above (Fig.4E),
CaMKII activity in the cultured cells was enhanced by lowering the
temperature (Q10=0.78), indicating that Ca2+ influx is a key factor
for accelerating CaMKII-mediated processes in the circadian clock
at lower temperatures. Overexpression of CaMKII-T286D, a con-
stitutive active form of CaMKII (14), accelerated the oscillation speed
(shortened the period) and increased the amplitude of Bmal1-luc
rhythms in cultured NIH3T3 cells (Fig.5,BandC). To understand
the experimental results theoretically, we simulated the effect of
phosphorylation-dependent activation of CLOCK-BMAL1 on the
gene expression rhythm by using a previously published mathematical
model (23). In the horizontal axis of this simulation (Fig. 5D), a
standard phosphorylation rate of CLOCK-BMAL1 estimated from
previous experimental results was set to 1 (23). We found that an
increase in the phosphorylation rate of CLOCK-BMAL1 accelerated
the oscillation speed (shortened the period) and increased the am-
plitude of Bmal1 mRNA rhythm (Fig.5D). These theoretical analysis
and experimental data collectively indicate that the cold-responsive
Ca2+ signaling compensates for the period lengthening and ampli-
tude reduction of the TTFL caused by lowering the temperatures.
Considering the roles of intracellular Ca2+ in the circadian oscilla-
tion of the TTFL (Fig.3) and in its temperature compensation
(Figs.1,2,4, and 5,AtoD), we propose an oscillation model in which
the TTFL couples with a Ca2+ oscillator for temperature-compensated
circadian rhythms (Fig. 5E). We then examined responses of the
Ca2+ oscillator to temperature changes. Circadian rhythms of intra-
cellular Ca2+ levels in cultured slices of the mouse suprachiasmatic
nucleus (SCN) were monitored by using adeno-associated virus-
mediated gene transfer of GCaMP6s (11). Lowering the temperature
from 35° to 28°C caused upward shifts of both the peak and trough
levels of the intracellular Ca2+ (Fig.5,FandG) with no significant
change in the period length of the Ca2+ oscillation (Q10 = 1.02).
These results together suggest that the circadian Ca2+ oscillator is
highly responsive to temperature changes to maintain constant pe-
riod lengths of cellular circadian rhythms.
Cold-responsive phosphorylation signaling is conserved
among animals and plants
The cold-responsive Ca2+ signaling was investigated invivo in several
organisms. Cold exposure of mice to 4°C for 10min remarkably
decreased the temperatures of the body surface (Fig.6A) without a
large change in the core body temperature (fig. S7A). Infrared ther-
mography revealed that the temperatures of the ear and tail dropped
by 12.0° and 16.5°C, respectively (Fig.6A). We found that CaMKII
activities (toward syntide-2) in the tissue lysates were enhanced by
1.34-fold (ear) and 2.25-fold (tail) after 90-min exposure at 4°C (Fig.6B).
The cold response of CaMKII was also analyzed in Drosophila
melanogaster. CaMKII activities in the fly heads were enhanced by
1.57-fold (Fig.6B), when the flies (maintained at 25°C) were ex-
posed to 4°C for 90min. In plants, Ca2+-dependent protein kinases
(CDPKs) are the major transducers of Ca2+ signaling (24). Catalytic
domains of CDPKs are highly homologous to animal CaMKII, and
syntide-2 is a model substrate of CDPKs (24). The enzymatic activities
phosphorylating syntide-2 in the shoot (leaf and stem) lysates of
Arabidopsis thaliana (kept at 22°C) were remarkably enhanced by
90-min exposure at 4°C (Fig.6B). These results suggest that activation
10 µM KN-93
20 µM KB-R7943
DMSO
18
Period (hours)
28
26
24
22
20
32 33 34 35 36 37
Temperature (ºC)
A
6
5
4
3
2
1
0
2
1
0
–1
–2
–3
Period change (hours)
–1
–2
–3
Period change (hours)
32 33 34 35 36 37 32 33 34 35 36 37
10 µM KN-93 20 µM KB-R7943
Temperature (ºC) Temperature (ºC)
B
Fig. 2. Temperature compensation is compromised by CaMKII or NCX inhibitor. (A) Period length of Rat-1–Bmal1-luc cells in the presence of KN-93 or KB-R7943 at
32°, 33°, 34°, 35°, 36°, or 37°C. ★P < 0.05 compared to DMSO (Student’s t test). (B) Temperature-dependent effect of KN-93 or KB-R7943 on the period length. The means
with SEM from three independent samples (A and B) are shown.

Kon et al., Sci. Adv. 2021; 7 : eabe8132 30 April 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
5 of 12
KB-R7943 (µM)
100
120
80
60
40
20
0
DMSO
KB-R7943
SKF-96365
DS44170716
Flecainide
TTFA
Lidocaine
KCN
HMA
Benzamil
Phloretin
Veratridine
Mibefradil
Mepivacaine
DS16570511
Amirolide
Bafilomycin
Ru360
NPPB
Bay-K-8644
Lanthanum Cl
Diltiazem
Furosemide
Nifedipine
Verapamil
100
80
60
40
20
0
120
Relative amplitude (%)
0 13 16 20 10 40 80 100 10 20
KN-93 TFP BAPTA KBR
Relative amplitude (%)
100
80
60
40
20
0
120
(µM)
AB
C
KB-R7943 (µM)
100
80
60
40
20
0
120
Cellular CaMKII activity (%)
0 20
D
F
Circadian time of treatment
20 µM KN-93 20 µM KB-R7943 20 µM SEA0400
0
6
–6
–12
1539 21
0
6
Phase shift (hours)
–6
–12
1539 21 0
6
–6
–12
1539 21
nd
02
0105
Relative fluorescence
intensity (%)
12
012
3
Bioluminescence levels (×103 counts/min)
30 µM KB-R7943 30 µM SEA0400
Bmal1-luc
Time (days)
Bmal1-luc
Per2-luc Per2-luc
30 µM KB-R7943 30 µM SEA0400
6
5
8
4
0
0
0
0
2
1
4
2
4
3
2
1
34567 01234567
Time (days)
012 3 4567 01234567
E
Fig. 3. NCX-dependent Ca2+-CaMKII signaling is a key determinant of the state of circadian oscillator. (A) Effects of various ion channel modulators on the ampli-
tude of Rat-1–Bmal1-luc cells. Bioluminescence rhythm data are shown in fig. S2. (B) Effects of NCX inhibitors on intracellular Ca2+ levels. The Ca2+ level changes by the
drug were measured by using Fluo-4 in NIH3T3 cells. ★P < 0.05 compared to DMSO (Dunnett’s test). (C) Effects of the NCX inhibitors on intracellular CaMKII levels in NI-
H3T3 cells. After 1-day treatment with the inhibitor or DMSO, phosphorylation activity of the cell lysate was measured with syntide-2. ★P < 0.05 compared to DMSO
(Student’s t test). (D) Effects of Ca2+-CaMKII signaling inhibitors on amplitude of the rhythms in Rat-1–Bmal1-luc cells. ★P < 0.05 compared to DMSO (Dunnett’s test). The
level of DMSO control was set to 100% (A to D). TFP, trifluoperazine. (E) Reversible effects of NCX inhibitors on bioluminescence rhythm of Rat-1–Bmal1-luc cells or Rat-
1–Per2-luc cells. (F) Effects of pulse inhibition of CaMKII or NCX on phase of the oscillator. Two-hour treatment of KN-93 (left) or KB-R7943 (middle) or 1-hour treatment of
SEA0400 (right) was applied to Rat-1–Bmal1-luc cells at various circadian time (CT). CT12 was defined as trough level of Bmal1-luc rhythm. Because the treatment of KN-93
at CT24 resulted in the disappearance of the cellular rhythm, the extent of the phase shift was not determined. Data shown are means with SEM from three (A, B, D, and
F) or eight (C) independent samples. All experimental data in this figure were obtained from cells cultured at 37°C. nd, not determined.

Kon et al., Sci. Adv. 2021; 7 : eabe8132 30 April 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
6 of 12
Per1
Per2
Per3
Cry2
Dbp
Rev-erb
Rev-erb E4bp4
Bmal1
Clock
Npas2
Dec1
Dec2
Ror
Ror
Cry1
1
0
2
3
0
2
0
2
4
0
2
6
4
0
2
6
4
1
0
2
3
0
2
6
4
0
1
2
0
2
24 30 36 42
6
4
1
0
2
4
0
12
8
1
0
2
3
1
0
2
3
24 30 36 42 0
2
24 30 36 42
4
1
0
2
3
24 30 36 42
1
0
2
3
4
Time after Dex shock (hours)
Relative mRNA levels
37ºC 33ºC 29ºC
4
6
37 36 35 34 33 32
Temperature (ºC)
100
80
60
40
20
0
120
Relative amplitude (%)
0 2 4
6 8 10 KN-93 (µM)
37 34 31 28 25
100
120
Temperature (ºC)
Relative fluorescence
intensity (%)
37ºC 25ºC
AB
CD
37 34 31 28 25
Temperature (ºC)
100
80
60
120
37 34 31 28 25
Temperature (ºC)
100
80
60
120
DMSO
10 µM
20 µM
DMSO
10 µM
40 µM
SEA0400KB-R7943
E
37ºC
27ºC
100
80
60
40
120
140
CaMKII activity (%)
KB-R7943
DMSO
39 38 37 36 35 34 33 32 31 30 29 28 27
Temperature (ºC)
100
80
60
40
20
0
120
Relative amplitude (%)
DMSO
10 µM KN-93
#1
#2
#3
F
G
Pretreat
DMSO
A23187
Ionomycin
Pretreat
DMSO
A23187
Ionomycin
Pretreat
DMSO
A23187
Ionomycin
120
80
40
0
140
160 Per1 Per2 Dec1
Relative mRNA levels (%)
ns
Relative fluorescence
intensity (%)
Per1 Per2
Relative mRNA levels (%)
100
0
200
300
100
0
200
37ºC 27ºC 37ºC 27ºC
DMSO 10 µM KB-7943 10 µM KN-93
H
Fig. 4. Hypothermic activation of NCX-dependent Ca2+-CaMKII signaling. (A) Effects of temperature on amplitude of cellular rhythm in Rat-1–Bmal1-luc cells.
(B) Temperature- and dose-dependent effects of KN-93 on amplitude of cellular rhythm in Rat-1–Bmal1-luc cells. (C) Hypothermic Ca2+ response in NIH3T3 cells. The mean
value at 37°C is set to 100. ★P < 0.05 compared to 37°C (Dunnett’s test). Right panels are representative images of intracellular Ca2+ levels in NIH3T3 cells at 37° or 25°C.
(D) KB-R7943 or SEA0400 blocks hypothermic Ca2+ response in NIH3T3 cells. Initial value of each cell at 37°C was set to 100%. ★P < 0.5 × 10−7 compared to DMSO (Stu-
dent’s t test). The Ca2+ imaging analysis was started from 37°C down to 25°C (C and D). (E) NCX mediates hypothermic CaMKII activation in NIH3T3 cells. The mean value
of DMSO at 37°C is set to 100%. ★P < 0.05 (Student’s t test). ns, not significant. (F) Ca2+ ionophore up-regulates clock gene Per1, Per2, and Dec1. Thirty-six hours after
the rhyth m induction with dexamethasone, 10 M (final concentration) ionomycin, A23187, or the same volume of DMSO was applied to Rat-1–Bmal1-luc cells. One hour
after the treatment, the cells were harvested to detect clock gene mRNA levels. The mean value of pretreatment is set to 100%. ★P < 0.05 compared to DMSO (Student’s
t test). (G) Hypothermic response of clock genes in Rat-1–Bmal1-luc cells. (H) NCX and CaMKII mediate hypothermic up-regulation of Per1 and Per2 in Rat-1–Bmal1-luc cells.
The mean value at 37°C is set to 100%. ★P < 0.05 and ★★P < 0.005 compared to DMSO-treated cells at 27°C (Student’s t test). The cells were harvested to detect clock gene
mRNA levels at indicated time points (G) or 5 days (H) after rhythm induction by dexamethasone. Representative data [panels of (C)] or means with SEM from 3 (A, B, F,
and G), 8 (E), 9 (H), or 20 (C and D) independent samples are shown.

Kon et al., Sci. Adv. 2021; 7 : eabe8132 30 April 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
7 of 12
CaMKII
BMAL1CLOCK
P
Ca2+
Tr
anscriptional loop
Ca2+ oscillator
E-box
PER CRY
NCX
Temperature signal
CaM
100
80
40
0
5101520
Temperature (ºC)
Phosphorylated CLOCK (%)
60
20
120
Q10 = 2.9
Period
Period length (hours)
Bmal1 mRNA level (%)
120
80
40
0
22
23
24
25
Peak
Trough
01234
Phosphorylation rate of CLOCK-BMAL1
01234
AB C
E
D
Bioluminescence levels
(×103 counts/min)
Time (days)
Empty
LacZ
CaMKII-T286D
3
2
1
01234
0
5
21.0
21.5
22.0
Empty
LacZ
CaMKII
-T286D
Period length (hours)
120
100
80
Amplitude (%)
Empty
LacZ
CaMKII
-T286D
Amplitude
Mathematical simulation
Phosphorylation rate of CLOCK-BMAL1
1023456 789101112
Relative Ca2+ levels
350
550
510
470
430
390
Time (days)
35ºC 28ºC 35ºC
Peak
Trough
1st 2nd 3rd
35ºC 28ºC
Normalized Ca2+ (%)
90
100
110
120
130
Peak/trough
1st 2nd 3rd
FG
GCaMP6s
SCN
Q10 = 1.02
Fig. 5. Mechanism of temperature compensation. (A) Effect of temperature on phosphorylation activity of purified CaMKII. Phosphorylation activity of rat CaMKII
against CLOCK peptide was measured by autoradiography. The level of phosphorylated CLOCK at 20°C was set to 100%. (B) Effect of CaMKII overexpression on bio-
luminescence rhythm by Bmal1-luc in NIH3T3 cells. (C) Effect of CaMKII overexpression on period length and amplitude of cellular rhythm. ★P < 1.0 × 10−6 (Student’s t test
with Bonferroni correction). (D) Mathematical simulation of effect of CLOCK-BMAL1 activation on period length and amplitude of Bmal1 expression rhythms. (E) Circadian
Ca2+ oscillator regulates TTFL to generate temperature-compensated overt rhythms in mammalian circadian clock. (F) Effect of temperature on Ca2+ oscillation in SCN.
(G) Hypothermia increases trough and peak levels of Ca2+ oscillation in SCN. The fluorescence levels of GCaMP6s were divided by those of mRubby for normalization of
effect of temperature on the fluorescence indicator. Representative data [top panels of (A), (B) and (F)] or means with SEM (A, C, and G) from three (A), four (B and C), or
seven (F and G) independent samples are shown.
A
at 23ºC
at 4ºC
20
4
36
0
100
50
150
200
250
Phosphorylation activity (%)
Ear Tail
B
Mus musculus
Head
Drosophila
melanogaster
0
Arabidopsis
thaliana
423
100
50
150
423 425
Shoot
0
100
422
200
300
400
500
CoreEar Tail
4ºC23ºC
Ambient temperature
Body temperature (ºC)
20
30
40
10
04ºC
23ºC
Max
Min
(ºC
)
Fig. 6. Cold-responsive phosphorylation signaling conserved in animals and plants. (A) Body temperature of mice at normal (23°C) or cold (4°C) temperature. Surface
body temperature was measured by infrared thermography. Representative images are shown in the left panel. Core body temperature of mice was measured by im-
plantable device in peritoneal cavity (fig. S7). (B) Hypothermic activation of cellular phosphorylation activity against syntide-2 in the ear or tail of Mus musculus, the head
of D. melanogaster, or the shoot of A. thaliana. The phosphorylation activity at normal temperature was set to 100%. ★P < 0.01 and ★★P < 1.0 × 10−5 compared to non-
treated samples (Student’s t test). Data are means with SEM from three independent samples (A and B).

Kon et al., Sci. Adv. 2021; 7 : eabe8132 30 April 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
8 of 12
23.2
23.3
23.4
23.5
23.6
23.7
WT NCX2+/–
Period length (hours)
WT calxAcalxB
0.03
0.02
0.01
0
FFT power
WT calxAcalxB
A B
D E
10
012012
WT NCX2+/–
012012
30
20
40
1
LD
DD
Mus musculus
Drosophila melanogaster
Days
(hours)
120120120120120120
4
Bioluminescence (×104 cpm)
12
3
2
1
8
4
12340
Arabidopsis thaliana CCA1::LUC
Normal
Ca2+ free
22ºC 17ºC
00
Time (days)
G
25
26
27
28
29
30
Period length (hours)
31
ns
F
12340
25ºC30ºC
Q10
1.6
1.1
1.2
1.3
1.4
1.5
Period length (hours)
23
24
25
26
27
28
29
30
25ºC 30ºC
WT
∆yrbG
Bioluminescence (×104 cpm)
20
10
0
10
5
0
012301234
Time (days)
WT
∆
yrbG
IJK
WT
∆yrbG
Synechococcus elongatus
PkaiBC::luxAB
17ºC 22ºC
Normal
Ca2+ free
Q10
0.7
0.8
0.9
1.0
Normal
Ca2+ free
H
NCX2+/–
NCX3–/–
23.8
NCX2+/– NCX3–/– #1
012012
LD
DD
(hours)
NCX2+/– NCX3–/– #2
012012
Mus musculus
C
10
30
20
40
1
Days
50
70
60
80
Fig. 7. Conserved roles of Ca2+ signaling in circadian clockwork. (A) Wheel-running rhythm of NCX mutant mice. LD, light-dark; DD, constant dark.(B) Period length of
wheel-running rhythm of the NCX mutant mice. Animal number of wild type (WT), NCX2+/−, or NCX2+/− NCX3−/− is 6, 10, or 7, respectively. ★P < 0.05 compared to WT
(Student’s t test). (C) Aberrant pattern of morning and evening activity rhythms in NCX2+/− NCX3−/− mice. (D) Locomotor activity rhythm of calx mutants of D. melanogaster.
(E) Relative FFT power of locomotor activity rhythm of calx mutants of D. melanogaster. Animal number of WT, calxA, or calxB is 15, 16, or 14, respectively. ★P < 0.5 × 10−4
compared to WT (Student’s t test). (F) Effect of Ca2+ depletion on gene expression rhythm by CCA1::LUC reporter in A. thaliana. (G) Effect of Ca2+ depletion on period
length of gene expression rhythm by CCA1::LUC reporter in A. thaliana. (H) Effect of Ca2+ depletion on Q10 of gene expression rhythm by CCA1::LUC reporter in A. thaliana.
The sample number of each group is 20. ★P < 1.0 × 10−7 (Student’s t test). (I) Effect of knockout of yrbG on gene expression rhythm by PkaiBC::luxAB in Synechococcus
elongatus PCC 7942 at 25° or 30°C. (J) Effect of knockout of yrbG on period length of gene expression rhythm at 25° or 30°C. ★P < 0.05 and ★★P < 0.0005. (K) Effect of
knockout of yrbG on Q10 of gene expression rhythm at 25° or 30°C. *P < 0.005. The sample number of each group is 4 (J and K). Data shown are representative (A, C, D, F,
and I) or means with SEM (B, E, G, H, J, and K).

Kon et al., Sci. Adv. 2021; 7 : eabe8132 30 April 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
9 of 12
of Ca2+-dependent phosphorylation signaling is a conserved mech-
anism underlying the cold responses.
Roles of NCX in circadian clockworks are conserved
in eukaryotes and prokaryotes
Mammalian NCXs form a multigene family composed of three mem-
bers: NCX1, NCX2, and NCX3. NCX1 is ubiquitously expressed in
a variety of tissues, while NCX2 and NCX3 are expressed in the brain
and muscle (18,25,26). A previous study demonstrated that homo-
zygous knockout of NCX1 or NCX2 results in lethality (18,25). We
examined wheel-running activity rhythms of NCX2+/− mice and
NCX2+/− NCX3−/− double-mutant mice. In constant dark condition,
NCX2+/− and NCX2+/− NCX3−/− mice showed free-running rhythms
with circadian periods significantly longer than that of wild-type
mice (Fig.7,AandB). One double-mutant mouse exhibited unstable
coordination between onset and offset of the wheel-running activity
bouts under constant darkness (Fig. 7C), a phenotype similar to
that observed for CaMKIIK42R kinase-dead knock-in mice (14).
These results indicate that NCX2 and NCX3 play important roles in
maintaining normal behavioral rhythms in mice.
In D. melanogaster, NCX is encoded by a single gene, calx. We
analyzed behavioral rhythms of two different lines of calx mutants,
calxA deficient for Na+/Ca2+ exchange currents and calxB deficient
for CALX protein expression (27). Both calxA and calxB homozy-
gous mutants showed severely weakened rhythmicity inlocomotor
activities at 25°C under constant darkness (Fig.7D). Fast Fourier
transform (FFT) analysis revealed a significant reduction in the be-
havioral rhythmicity in the calx mutants (Fig.7,DandE), indicat-
ing an essential role of CALX in the Drosophila clock governing the
behavioral rhythms.
Roles of Ca2+ in plant clocks were investigated in A. thaliana
expressing CCA1::LUC reporter. Because Arabidopsis has 13 NCX
genes (28), it is difficult to evaluate roles of NCXs genetically. In-
stead, we investigated effects of Ca2+ depletion in a growth medium
on the bioluminescence rhythms. The Ca2+ depletion resulted in
significant shortening of the free-running period in constant light
condition at 22°C (Fig.7,FandG), whereas the period-shortening
effect was undetectable at 17°C. The Ca2+ depletion caused an in-
crease in the Q10 value from 0.80 (in the normal medium) to 0.95
(Fig.7H), indicating that Ca2+ signaling is required for accelerating
the oscillation speed at lower temperatures in the plant as well.
Roles of NCX in prokaryotic circadian clocks were investigated
by generating a cyanobacterial strain lacking yrbG, a bacterial homo-
log of NCX (28). The circadian rhythms in the yrbG strain were
monitored with PKaiBC::luxAB reporter under constant light condition
(Fig.7I). We found that yrbG deficiency caused significant shorten-
ing of the period length at 30°C, whereas the period was lengthened
at 25°C when compared with the wild-type strain (Fig.7,IandJ).
Hence, the Q10 value of the bioluminescence rhythms was increased
from 1.19 to 1.49 by the depletion of yrbG (Fig.7K). These results
demonstrate that NCX-dependent Ca2+ signaling plays a conserved
role in both the TTFL-based eukaryotic clock and the PTO-based
prokaryotic clock systems.
DISCUSSION
Circadian TTFLs are an elaborate system that drives a wide range of
overt rhythms with various phase angles and amplitudes. The oscil-
lation speed of the TTFLs is temperature compensated, although
many of the biochemical reactions in TTFLs are slowed down by
decreasing temperature (table S2). The present study demonstrates
that the temperature compensation of the TTFL in mammalian cells
was compromised when Ca2+-dependent phosphorylation signaling
was inhibited (Fig.2A). We found an important role of NCX-CaMKII
activity as the state variable of the circadian oscillator (Fig.3,DtoF,
and fig. S5). The present study and a series of preceding works
demonstrate that the Ca2+ oscillator plays essential roles in the cir-
cadian oscillation mechanism (Fig.5E) (8–16). Functional studies
clearly demonstrated essential roles of NCX-dependent Ca2+ signaling
in the three important properties of the circadian clock, i.e., cell-
autonomous oscillation (Figs.3,AtoE, and 7, A to C), temperature
compensation (Figs.1,2,4,and5), and entrainment (Fig.3F). The
circadian Ca2+ oscillation is observed in mice lacking Bmal1 or Cry1/
Cry2 (11,12), implicating that the Ca2+ oscillator is an upstream
regulator of the TTFL in mammals.
The effects of NCX2 and NCX3 deficiencies on the regulation of
mouse behavioral rhythms (Fig.7,AtoC) suggest involvement of
Na+/Ca2+ exchanging activity in the Ca2+ dynamics of the SCN. Previous
studies showed that L-type Ca2+ channel (LTCC) and voltage-gated
Na+ channel (VGSC) are required for high-amplitude Ca2+ rhythms
in the SCN (11,12). Because NCX activities are regulated by local
concentrations of Na+/Ca2+ and the membrane potential (18), co-
operative actions of LTCC, VGSC, and NCX seem to play impor-
tant roles in generation mechanism of the robust Ca2+ oscillations
in the SCN.
It should be emphasized that the role of Ca2+/calmodulin-dependent
protein kinases is conserved among clockworks in insects (9,13,16),
fungi (29), and plants (10,24), suggesting that the Ca2+ oscillator might
be a core timekeeping mechanism in their common ancestor (Fig.8,
Eukaryota). After divergence of each lineage, a subset of clock genes
should have independently evolved in association with the Ca2+ os-
cillator. It is noteworthy that NCX is also required for temperature
compensation of PTO-based cyanobacterial clock (Fig.7,I toK).
CLOCK BMAL
PER CRY
TIM
WC
FRQ
LHY
TOC1
CCA
SasA
RpaA
KaiA
KaiB
CKI
CKII
GSK3
NCX
Transcriptional
rhythm
Posttranslational
timekeepers
Bacteria
Archaea
Eukaryota
Mammals Drosophila Neurospora Arabidopsis Synechococcus
CaMK
Ca2+ homeostasis
KaiC
Fig. 8. Involvement of ancient Ca2+ signaling for temperature-compensated
circadian rhythms. Clock genes involved in the TTFLs evolved independently after
divergence of each lineage. In animals, fungi, and plants, common multifunctional
kinases, such as casein kinase I (CKI), CKII, glycogen synthase kinase 3 (GSK3), or Ca2+-
dependent kinase (CaMK), are involved in posttranslational regulation of clock gene
products. In cyanobacteria, posttranslational oscillator by KaiA/KaiB/KaiC drives
the TTFL. NCX, a highly conserved molecule among three domains of life, is a com-
mon circadian timekeeping element in the eukaryotes and prokaryotes, and its
original function is regulation of Ca2+ homeostasis.

Kon et al., Sci. Adv. 2021; 7 : eabe8132 30 April 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
10 of 12
Because intracellular Ca2+ in cyanobacteria is elevated in response
to temperature decrease (30), YrbG-mediated Ca2+ signaling may
regulate the PTO invivo. Conservation of NCX among eukaryotes,
eubacteria, and archaea (Fig.8) (31) suggests that NCX-dependent
temperature signaling is essential for adaptation of a wide variety of
organisms to environment. Further studies on NCX-regulated Ca2+
flux will provide evolutionary insights into the origin of the circa-
dian clocks.
MATERIALS AND METHODS
Real-time monitoring of gene expression rhythms
in mammalian cells
Real-time monitoring of gene expression rhythms in mammalian
cells was performed by using Rat-1 fibroblasts stably expressing
Bmal1-luciferase reporter (14,15,17). The fibroblasts were plated on
35-mm dishes (1.0×106 cells per dish) and cultured at 37°C under
5% CO2 in a culture medium of Dulbecco’s modified Eagle’s medi-
um (DMEM) (Sigma-Aldrich, catalog no. 5796) supplemented with
10% fetal bovine serum (FBS; Equitech-Bio Inc.), penicillin (50 U/ml),
and streptomycin (50 g/ml). One day after the plating, the cells
were treated with 0.1 M dexamethasone for 2hours, and the medium
was replaced with a recording medium of DMEM (Sigma-Aldrich,
catalog no. D2902) supplemented with 10% FBS, glucose (3.5 mg/ml),
penicillin (25 U/ml), streptomycin (25 g/ml), 0.1 mM luciferin, and
10 mM Hepes-NaOH (pH 7.0). The bioluminescence signals were
continually recorded from the cells cultured under air in a dish-type
bioluminescence detector, Kronos (ATTO, AB-2500), or LumiCycle
(Actimetrics). For overexpression of constitutive active CaMKII,
pcDNA3.1-rat CaMKII-T286D was transfected to NIH3T3 cells
expressing the Bmal1 reporter 1 day after plating of the cells (0.5 ×
106 cells per 35-mm dish) (14). The bioluminescence rhythms from
the cells were monitored (as described above) from 1 day after the
transfection.
For normalization of dish-to-dish variation of the bioluminescence
levels, the raw data were divided by the mean bioluminescence sig-
nals recorded for 7 days. The normalized rhythms were detrended
by subtracting 24-hour centered moving averages, and the areas un-
der the curves (arbitrary units) were used for calculating the relative
amplitudes of the rhythms (14). Period lengths were calculated us-
ing the average value of peak-to-peak periods and trough-to-trough
periods 1 day after the dexamethasone treatment of cultured cells.
Q10 value was calculated by the following equation
Q 10 = (1 / 2) 10/(T2‐T1)
where 1 and 2 are the periods at temperature T1 and T2, respectively.
Real-time monitoring of gene expression rhythms in plants
Monitoring of bioluminescence rhythms of A. thaliana (ecotype
Columbia-0) expressing CCA1::LUC was performed as described pre-
viously (32). The plants were grown on a growth medium containing
10 mM KCl, 0.6 mM NH4NO3, 0.5 mM H3BO3, 0.75 mM MgSO4,
0.015 mM ZnSO4, 0.05 mM MnSO4, 0.05 mM FeSO4, 1.5 mM CaCl2,
0.05 mM Na2-EDTA, 10 mM NH4NO3, and 0.8% agar (pH 6.3) at
22°C under 12-hour light (approximately 80 mol m−2 s−1)/12-hour
dark cycles for 2 weeks. Then, the plants were transferred to the
growth medium without CaCl2. Two days after the transfer, lucifer-
in (final concentration of 0.125 mM) was added to the medium, and
bioluminescence signals were measured with photomultiplier tubes
under continuous light conditions.
Real-time monitoring of gene expression rhythms
in cyanobacteria
A strain that harbored a PkaiBC::luxAB reporter cassette with a chlor-
amphenicol resistance gene at the targeting site (neutral site I) on
the genome (ILC 976) was used as a wild-type strain. To disrupt the
yrbG gene, a plasmid (pIL 1000) was constructed to harbor up-
stream and downstream regions of yrbG (Synpcc7942_0242) with a
gentamicin resistance gene in the pGEM-T Easy backbone (Promega).
The ∆yrbG strain (ILC 1383) was generated by transformation of
ILC 976 with pIL 1000. Cells were grown in BG-11 media in the
absence of calcium source (250 M CaCl2). The bioluminescence
profiles were measured with photomultiplier tubes under continu-
ous light (LL, 40 mol/m2 s) conditions after 2 days of 12-hour
light/12-hour dark cycles (33).
Reverse transcription polymerase chain reaction analysis
Total RNA was prepared from cultured cells using TRIzol reagent
(Invitrogen) according to the manufacturer’s protocol. Reverse
transcription polymerase chain reaction analysis was performed as
described previously (14,17).
Intracellular Ca2+ imaging
For Ca2+ imaging in cultured cells, NIH3T3 cells were plated on 35-mm
dishes (1.0×106 cells per dish) and cultured at 37°C under 5% CO2
in the culture medium. One day after the plating, the medium was
replaced by an imaging buffer of Hanks’ balanced salt solution (Sigma-
Aldrich, catalog no. H8264) containing 0.04% Pluronic F-127 and
1.25 mM probenecid. One hour after loading of 2 M Fluo-4 AM at
37°C, the fluorescence intensity of the cells was monitored by a flu-
orescence microscope (Olympus, BX51W1) equipped with an elec-
tron multiplying charge-coupled device digital camera (Hamamatsu
Photonics, C9100-13 ImagEM) in the imaging buffer. The buffer
was perfused by using a peristatic pump (Gilson, MINIPULS 3) for
control of the buffer temperature, which is continuously monitored
by thermoelectric couple and controlled by a dual automatic tem-
perature controller (Warner, TC-344B).
Circadian Ca2+ imaging in the SCN was performed as described
previously (11). Briefly, the SCN slices were prepared from neonate
mice (C57BL/6, 5 days old, both male and female). Ca2+ indicator
protein GCaMP6s and control fluorescence protein mRuby were
expressed under the control of the human synapsin-1 promoter by
using adeno-associated virus (Addgene, 50942-AAV1).
Measurement of CaMKII activity
For analysis of CaMKII activities in cultured cells, NIH3T3 cells were
plated on 100-mm dishes (1.0×107 cells per dish) and cultured at
37°C under 5% CO2 in the culture medium. One day after the plat-
ing, the medium was replaced by the recording medium containing
the NCX inhibitor or 0.1% DMSO (vehicle), and the cells were cul-
tured at 27° or 37°C. One day after the culture, cells were harvested
by a cell scraper with 2ml of a sampling buffer [20 mM tris-HCl, 5 mM
EDTA, 1 mM EGTA, 10 mM sodium pyrophosphate, 50 mM NaF,
1 mM Na3VO4, 1 mM dithiothreitol (DTT), 0.1 mM phenylmethylsulfonyl
fluoride, leupeptin (0.04 mg/ml), and aprotinin (0.04 mg/ml), pH 7.5].
For analysis of tissues, the tails or ears of C57BL/6 mice (7 weeks
old, male), the heads of D. melanogaster (W1118, male), or the shoots

Kon et al., Sci. Adv. 2021; 7 : eabe8132 30 April 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
11 of 12
of A. thaliana (ecotype Columbia-0, 14 days old) were prepared at
ZT5, and 1mg of the tissue was homogenized in 1ml of the sam-
pling buffer. The cells or tissues were homogenized by using a glass/
Teflon homogenizer (20 strokes). CaMKII activity levels of the lysates
phosphorylating syntide-2 were measured by using CaM-kinase II
Assay kit (CycLex, catalog no. CY-1173) according to the manufac-
turer’s protocol.
For analysis of purified CaMKII activity phosphorylating a CLOCK
peptide (GST-SP), CaM and rat brain CaMKII were prepared as de-
scribed previously (15,34). The assay was carried out at 5°, 10°, 15°,
or 20°C in a reaction mixture (10 l) composed of 40 mM tris-HCl
(pH 8.0), 2 mM DTT, 5 mM MgCl2, 0.5 mM CaCl2, 1 mM [-32P]
ATP, 1 M CaM, 100ng of rat brain CaMKII, and 500ng of GST-SP
peptide. After incubation for 30 min, the reaction was stopped by
the addition of 10 l of 2× SDS sample buffer. Phosphorylated pro-
teins or peptides were resolved by SDS–polyacrylamide gel electro-
phoresis and detected by autoradiography. We found that CaMKII
activity purified from the rat brain was inactivated by incubation
above 30°C, as reported by the previous study (34). Thus, the activ-
ity levels of the purified CaMKII were analyzed in the range of 5°
to 20°C.
Animal experiments
The animal experiments were conducted in accordance with the guide-
lines of the University of Tokyo. NCX2 heterozygous knockout
mice (NCX2+/−) were produced as described previously (25). NCX3
homozygous knockout mice (NCX3−/−) were generated as follows:
The targeting vector was constructed by replacing the 1.9 kilo–base
pairs Eco RI–Mun I fragment containing exon 2 of the NCX3 gene
with a PGK (Phosphoglycerate kinase promoter)–neo cassette. The
targeted ES (embryonic stem cell) clones were confirmed by Southern
blot analysis and used for the generation of germline chimeras. Chimeric
male mice were crossed with female C57BL/6 mice to establish the
germline transmission and backcrossed to C57BL/6 mice for more
than 10 generations. The mutant mice (C57BL/6 background, male,
6 to 8 weeks old) were housed individually at 23°C in cages (13 × 23 ×
15 cm) equipped with a running wheel (diameter, 10 cm) with food
and water available ad libitum. Wheel- running rhythms were mon-
itored under constant dark condition after housing under 12-hour
light/12-hour dark cycles for at least 2 weeks. The numbers of wheel
revolution were collected every minute into a computer system. All the
behavioral data were analyzed by using ClockLab software (Actimetrics).
For measurement of the internal body temperature, the activity- and
temperature-measuring device, nano tag (KISSEI COMTEC Co. Ltd.),
was implanted into the peritoneal cavity or subcutaneous site in mice
(C57BL/6 background, male, 8 weeks old). For measurement of the
surface body temperature, an infrared camera (FLIR, E6) was used,
and the image data were analyzed by FLIR Tools software (FLIR).
Locomotor activity rhythms of D. melanogaster were monitored
as described previously (35). Male flies (2 to 5 days old) were individ-
ually housed in glass tubes (length, 65 mm; inside diameter, 3 mm)
containing sucrose-agar (1% agar supplemented with 5% sucrose)
food at one end and a cotton plug on the other end. The glass tubes
were placed in the Drosophila activity monitor system (TriKinetics),
and the locomotor activity of each fly was recorded as the numbers
of infrared beam crossing in 1-min bin. Free-running rhythms were
recorded under constant dark condition after housing under 12-hour
light/12-hour dark cycles for at least 3 days. calxA or calxB mutant
flies were obtained from the Bloomington Drosophila Stock Center.
Mathematical analysis
By using a previously published mathematical model (23), we inves-
tigated an effect of CaMKII activation on the TTFL of the mammalian
circadian clock. Because CaMKII phosphorylates CLOCK to acti-
vate transcriptional activity of the CLOCK-BMAL1 complex (14–17),
we varied the corresponding parameter, which was represented as
“phos” in the original model (23). Ordinary differential equations were
solved numerically by using the Euler method with delta t=0.001.
SUPPLEMENTARY MATERIALS
Supplementary material for this article is available at http://advances.sciencemag.org/cgi/
content/full/7/18/eabe8132/DC1
View/request a protocol for this paper from Bio-protocol.
REFERENCES AND NOTES
1. J. W. Hastings, B. M. Sweeney, On the mechanism of temperature independence
in a biological clock. Proc. Natl. Acad. Sci. U.S.A. 43, 804–811 (1957).
2. J. C. Dunlap, J. J. Loros, Making time: Conservation of biological clocks from fungi
to animals. Microbiol. Spectr. 5, 3 (2017).
3. J. S. Takahashi, Transcriptional architecture of the mammalian circadian clock. Nat. Rev.
Genet. 18, 164–179 (2017).
4. M. A. Nohales, S. A. Kay, Molecular mechanisms at the core of the plant circadian
oscillator. Nat. Struct. Mol. Biol. 23, 1061–1069 (2016).
5. S. E. Cohen, S. S. Golden, Circadian rhythms in Cyanobacteria. Microbiol. Mol. Biol. Rev. 79,
373–385 (2015).
6. A. B. Webb, A. C. Oates, Timing by rhythms: Daily clocks and developmental rulers. Dev.
Growth Differ. 58, 43–58 (2016).
7. D. Njus, F. M. Sulzman, J. W. Hastings, Membrane model for the circadian clock. Nature
248, 116–120 (1974).
8. G. B. Lundkvist, Y. Kwak, E. K. Davis, H. Tei, G. D. Block, A calcium flux is required
for circadian rhythm generation in mammalian pacemaker neurons. J. Neurosci. 25,
7682–7686 (2005).
9. M. C. Harrisingh, Y. Wu, G. A. Lnenicka, M. N. Nitabach, Intracellular Ca2+ regulates
free-running circadian clock oscillation in vivo. J. Neurosci. 27, 12489–12499 (2007).
10. M. C. Martí Ruiz, K. E. Hubbard, M. J. Gardner, H. J. Jung, S. Aubry, C. T. Hotta, N. I. Mohd-Noh,
F. C. Robertson, T. J. Hearn, Y. C. Tsai, A. N. Dodd, M. Hannah, I. A. Carré, J. M. Davies, J. Br aam,
A. A. R. Webb, Circadian oscillations of cytosolic free calcium regulate the Arabidopsis
circadian clock. Nat. Plants 4, 690–698 (2018).
11. R. Enoki, D. Ono, S. Kuroda, S. Honma, K. I. Honma, Dual origins of the intracellular
circadian calcium rhythm in the suprachiasmatic nucleus. Sci. Rep. 7, 41733 (2017).
12. T. Noguchi, T. L. Leise, N. J. Kingsbury, T. Diemer, L. L. Wang, M. A. Henson, D. K. Welsh,
Calcium circadian rhythmicity in the suprachiasmatic nucleus: Cell autonomy
and network modulation. eNeuro 4, ENEURO.0160-17.2017 (2017).
13. F. Guo, J. Yu, H. J. Jung, K. C. Abruzzi, W. Luo, L. C. Griffith, M. Rosbash, Circadian neuron
feedback controls the Drosophila sleep-activity profile. Nature 536, 292–297 (2016).
14. N. Kon, T. Yoshikawa, S. Honma, Y. Yamagata, H. Yoshitane, K. Shimizu, Y. Sugiyama,
C. Hara, I. Kameshita, K. I. Honma, Y. Fukada, CaMKII is essential for the cellular clock
and coupling between morning and evening behavioral rhythms. Genes Dev. 28,
1101–1110 (2014).
15. N. Kon, Y. Sugiyama, H. Yoshitane, I. Kameshita, Y. Fukada, Cell-based inhibitor screening
identifies multiple protein kinases important for circadian clock oscillations. Commun.
Integr. Biol. 8, e982405 (2015).
16. T. S. Kaiser, B. Poehn, D. Szkiba, M. Preussner, F. J. Sedlazeck, A. Zrim, T. Neumann,
L. T. Nguyen, A. J. Betancourt, T. Hummel, H. Vogel, S. Dorner, F. Heyd, A. von Haeseler,
K. Tessmar-Raible, The genomic basis of circadian and circalunar timing adaptations
in a midge. Nature 540, 69–73 (2016).
17. N. Kon, T. Hirota, T. Kawamoto, Y. Kato, T. Tsubota, Y. Fukada, Activation of TGF-/activin
signalling resets the circadian clock through rapid induction of Dec1 transcripts. Nat. Cell
Biol. 10, 1463–1469 (2008).
18. T. Iwamoto, Y. Watanabe, S. Kita, M. P. Blaustein, Na+/Ca2+ exchange inhibitors: A new
class of calcium regulators. Cardiovasc. Hematol. Disord. Drug Targets 7, 188–198 (2007).
19. C. H. Johnson, J. A. Elliott, R. Foster, Entrainment of circadian programs. Chronobiol. Int.
20, 741–774 (2003).
20. D. J. Brown, H. Brugger, J. Boyd, P. Paal, Accidental hypothermia. N. Engl. J. Med. 367,
1930–1938 (2012).
21. N. M. Teets, S.-X. Yi, R. E. Lee Jr., D. L. Denlinger, Calcium signaling mediates cold sensing
in insect tissues. Proc. Natl. Acad. Sci. U.S.A. 110, 9154–9159 (2013).
22. Y. Nishimura, T. Tanaka, Calcium-dependent activation of nuclear factor regulated by
interleukin 3/adenovirus E4 promoter-binding protein gene expression by calcineurin/nucl ear

Kon et al., Sci. Adv. 2021; 7 : eabe8132 30 April 2021
SCIENCE ADVANCES | RESEARCH ARTICLE
12 of 12
factor of activated T cells and calcium/calmodulin-dependent protein kinase signaling.
J. Biol. Chem. 276, 19921–19928 (2001).
23. J. K. Kim, D. B. Forger, A mechanism for robust circadian timekeeping via stoichiometric
balance. Mol. Syst. Biol. 8, 630 (2012).
24. A. Liese, T. Romeis, Biochemical regulation of in vivo function of plant calcium-dependent
protein kinases (CDPK). Biochim. Biophys. Acta 1833, 1582–1589 (2013).
25. N. Morimoto, S. Kita, M. Shimazawa, H. Namimatsu, K. Tsuruma, K. Hayakawa, K. Mishima,
N. Egashira, T. Iyoda, I. Horie, Y. Gotoh, K. Iwasaki, M. Fujiwara, T. Matsuda, A. Baba,
I. Komuro, K. Horie, J. Takeda, T. Iwamoto, H. Hara, Preferential involvement of Na+/Ca2+
exchanger type-1 in the brain damage caused by transient focal cerebral ischemia
in mice. Biochem. Biophys. Res. Commun. 429, 186–190 (2012).
26. Y. C. Wang, Y. S. Chen, R. C. Cheng, R. C. Huang, Role of Na+/Ca2+ exchanger in Ca2+
homeostasis in rat suprachiasmatic nucleus neurons. J. Neurophysiol. 113, 2114–2126
(2015).
27. T. Wang, H. Xu, J. Oberwinkler, Y. Gu, R. C. Hardie, C. Montell, Light activation, adaptation,
and cell survival functions of the Na+/Ca2+ exchanger CalX. Neuron 45, 367–378 (2005).
28. L. Emery, S. Whelan, K. D. Hirschi, J. K. Pittman, Protein phylogenetic analysis of Ca2+/
cation antiporters and insights into their evolution in plants. Front. Plant Sci. 3, 1 (2012).
29. Y. Yang, P. Cheng, G. Zhi, Y. Liu, Identification of a calcium/calmodulin-dependent
protein kinase that phosphorylates the Neurospora circadian clock protein frequency.
J. Biol. Chem. 276, 41064–41072 (2001).
30. I. Torrecilla, F. Leganés, I. Bonilla, F. Fernández-Piñas, Use of recombinant aequorin
to study calcium homeostasis and monitor calcium transients in response to heat
and cold shock in cyanobacteria. Plant Physiol. 123, 161–176 (2000).
31. M. C. Weiss, F. L. Sousa, N. Mrnjavac, S. Neukirchen, M. Roettger, S. Nelson-Sathi,
W. F. Martin, The physiology and habitat of the last universal common ancestor. Nat.
Microbiol. 1, 16116 (2016).
32. K. Negishi, M. Endo, M. Abe, T. Araki, SODIUM POTASSIUM ROOT DEFECTIVE1 regulates
FLOWERING LOCUS T expression via the microRNA156-SQUAMOSA PROMOTER BINDING
PROTEIN-LIKE3 module in response to potassium conditions. Plant Cell Physiol. 59,
404–413 (2018).
33. S. Kutsuna, Y. Nakahira, M. Katayama, M. Ishiura, T. Kondo, Transcriptional regulation
of the circadian clock operon kaiBC by upstream regions in cyanobacteria. Mol. Microbiol.
57, 1474–1484 (2005).
34. A. Ishida, H. Fujisawa, Stabilization of calmodulin-dependent protein kinase II through
the autoinhibitory domain. J. Biol. Chem. 270, 2163–2170 (1995).
35. H. Funato, C. Miyoshi, T. Fujiyama, T. Kanda, M. Sato, Z. Wang, J. Ma, S. Nakane, J. Tomita,
A. Ikkyu, M. Kakizaki, N. Hotta-Hirashima, S. Kanno, H. Komiya, F. Asano, T. Honda,
S. J. Kim, K. Harano, H. Muramoto, T. Yonezawa, S. Mizuno, S. Miyazaki, L. Connor,
V. Kumar, I. Miura, T. Suzuki, A. Watanabe, M. Abe, F. Sugiyama, S. Takahashi, K. Sakimura,
Y. Hayashi, Q. Liu, K. Kume, S. Wakana, J. S. Takahashi, M. Yanagisawa, Forward-genetics
analysis of sleep in randomly mutagenized mice. Nature 539, 378–383 (2016).
36. K. Sidaway-Lee, M. J. Costa, D. A. Rand, B. Finkenstadt, S. Penfield, Direct measurement
of transcription rates reveals multiple mechanisms for configuration of the Arabidopsis
ambient temperature response. Genome Biol. 15, R45 (2014).
37. J. F. Sayegh, A. Lajtha, In vivo rates of protein synthesis in brain, muscle, and liver of five
vertebrate species. Neurochem. Res. 14, 1165–1168 (1989).
38. N. Shulga, P. Roberts, Z. Gu, L. Spitz, M. M. Tabb, M. Nomura, D. S. Goldfarb, In vivo nuclear
transport kinetics in Saccharomyces cerevisiae: A role for heat shock protein 70 during
targeting and translocation. J. Cell Biol. 135, 329–339 (1996).
39. J. M. Fagan, A. L. Goldberg, The rate of protein degradation in isolated skeletal muscle
does not correlate with reduction-oxidation status. Biochem. J. 227, 689–694 (1985).
40. Y. Isojima, M. Nakajima, H. Ukai, H. Fujishima, R. G. Yamada, K. H. Masumoto, R. Kiuchi,
M. Ishida, M. Ukai-Tadenuma, Y. Minami, R. Kito, K. Nakao, W. Kishimoto, S. H. Yoo,
K. Shimomura, T. Takao, A. Takano, T. Kojima, K. Nagai, Y. Sakaki, J. S. Takahashi,
H. R. Ueda, CKI/-dependent phosphorylation is a temperature-insensitive, period-
determining process in the mammalian circadian clock. Proc. Natl. Acad. Sci. U.S.A. 106,
15744–15749 (2009).
Acknowledgments: We are grateful to members of the Fukada laboratory, S. Gibo at RIKEN,
J. Tomita at Nagoya City University, I. Daichi at Waseda University, H. Ito and M. Seki at Kyushu
University, K.-I. Homma and S. Honma at Hokkaido University, and members of the
Transformative Research Area “Hibernation Biology” for the helpful discussions. Funding: This
work was supported, in part, by the Japanese Society for the Promotion of Science (JSPS)
Grants-in-Aid for Scientific Research (KAKENHI) to Y.F. (17H06096), N.Ko. (18H06066,
20H03292, and 20H05769), and T.I. (17K08610). N.Ko. was supported by the Tomizawa Jun-ichi
and Keiko Fund of Molecular Biology Society of Japan for Young Scientist and the Cooperative
Study Program (9273) of the National Institute for Physiological Sciences. H.-t.W. is supported
by JSPS Research Fellowship for Young Scientists. Author contributions: N.Ko. and Y.F.
planned the research project. N.Ko. and H.-t.W. performed the analysis using mammalian cells
and mice. Y.S.K. and K.Ku. performed the analysis using Drosophila. K.U. and M.E. performed
the analysis using Arabidopsis. N.Ka., K.Ka., and H.I. performed the analysis using cyanobacteria.
R.E. performed the Ca2+ imaging experiments using the mouse SCN. G.K. performed the
mathematical simulation. T.N. and Y.S. performed the in vitro CaMKII kinase assay. H.T. and T.I.
generated NCX2 and NCX3 knockout mice, and T.I. provided useful advice on the NCX
knockout mouse experiments and reviewed the manuscript. N.Ko. and Y.F. wrote the
manuscript with support from all authors. Competing interests: The authors declare that they
have no competing interests. Data and materials availability: All data needed to evaluate
the conclusions in the paper are present in the paper and/or the Supplementary Materials.
Additional data related to this paper may be requested from the authors.
Submitted 17 September 2020
Accepted 11 March 2021
Published 30 April 2021
10.1126/sciadv.abe8132
Citation: N. Kon, H.-t. Wang, Y. S. Kato, K. Uemoto, N. Kawamoto, K. Kawasaki, R. Enoki, G. Kuros awa,
T. Nakane, Y. Sugiyama, H. Tagashira, M. Endo, H. Iwasaki, T. Iwamoto, K. Kume, Y. Fukada, Na+/
Ca2+ exchanger mediates cold Ca2+ signaling conserved for temperature-compensated circadian
rhythms. Sci. Adv. 7, eabe8132 (2021).