Circadian control by the reduction/oxidation pathway: Catalase represses light-dependent clock gene expression in the zebrafish Jun Hirayama*, Sehyung Cho†, and Paolo Sassone-Corsi*‡ *Department of Pharmacology, School of Medicine, University of California, Irvine, CA 92697; and †Kyung Hee Institute of Age-Related and Brain Diseases, Kyung University, Seoul 130-701, Korea
Light is the key entraining stimulus for the circadian clock, but several features of the signaling pathways that convert the photic signal to clock entrainment remain to be deciphered. Here, we show that light induces the production of hydrogen peroxide (H2O2) that acts as the second messenger coupling photoreception to the zebrafish circadian clock. Treatment of light-responsive Z3 cells with H2O2 triggers the induction of zCry1a and zPer2 genes and the subsequent circadian oscillation of zPer1. Remarkably, the induction kinetics and oscillation profile in response to H2O2 are identical to those initiated by light. Catalase (Cat), an antioxidant enzyme degrading H2O2, shows an oscillating pattern of expression and activity, antiphasic to zCry1a and zPer2. Interestingly, overexpression of zCAT results in a reduced light-dependent zCry1a and zPer2 gene induction. In contrast, inhibition of zCAT function enhances light-mediated inducibility of these clock genes. These findings implicate the enzymatic function of zCAT enzyme in the negative regulation of light-dependent clock gene transcriptional activation. Our findings provide an attractive link between the regulation of the cellular reduction/oxidation (redox) state and the photic signaling pathways implicated in circadian control. light signaling 兩 hydrogen peroxide 兩 transcription 兩 circadian rhythms
O
rganisms ranging from bacteria to humans have daily rhythms driven by endogenous oscillators called circadian clocks that regulate various biochemical, physiological, and behavioral processes with a periodicity of ⬇24 h (1–3). Under natural conditions, rhythms are entrained to a 24-h cycle by environmental time cues, light being the most important. These circadian clock mechanisms have been investigated and specific ‘‘clock genes’’ that affect the daily rhythm have been identified. The core of the clock mechanism in Drosophila, Neurospora, mammals, and cyanobacteria is commonly represented by a transcription/translation-based negative-feedback loop that relies on positive and negative oscillator elements. Although the organization of the negative feedback loop in Drosophila, Neurospora, mammals, and cyanobacteria is conceptually similar, its components differ among species (1). In mammals, two basic helix–loop– helix PAS(PER-ARNT-SIM) domain-containing transcription factors, CLOCK and BMAL1, constitute the positive elements (4, 5). Upon heterodimerization, CLOCK:BMAL1 drives the transcription of the negative components of the clock machinery, three Period genes (Per1, Per2, and Per3) and two Cryptochrome genes (Cry1 and Cry2). The PERs and CRYs negatively regulate their own expression, setting up the rhythmic oscillations of gene expression that drive the circadian clock (6 – 8). The zebrafish constitutes an attractive alternative to the mammalian system to study the complexity of the circadian clock machinery and the influence that light has on it (9, 10). Characterization of the molecular components of the zebrafish circadian oscillator has revealed that the zebrafish negativefeedback loop consists of components similar to those of mamwww.pnas.org兾cgi兾doi兾10.1073兾pnas.0705614104
mals (10). Organ and tissue culture explant experiments have demonstrated that peripheral circadian oscillators are present throughout the tissues and organs of the zebrafish and that they display the remarkable feature of being light-responsive (11–13). We have established a cell line derived from zebrafish embryos, designated Z3, in which circadian clock components display distinct and differential light-dependent activation and expression profiles under various light conditions (14). Z3 cells, therefore, nicely recapitulate the zebrafish clock system and constitute an invaluable tool for investigating the link between light-dependent gene activation and the signaling pathways responsible for the generation of vertebrate circadian rhythms. Hydrogen peroxide (H2O2) is one of the reactive oxygen species (ROS) which is generated from various sources, including the mitochondria and the peroxisome, as a by-product of normal metabolism (15, 16). H2O2 can be generated by specific enzymes, such as NADPH oxidase, Acyl-Coa oxidase (AOX), and amine oxidase (17–19). A variety of extracellular stimuli, like visible light, low and high temperature, and wounding, enhance cellular H2O2 generation through the activation of these enzymes (20, 21). Excess production of H2O2 has deleterious effects because H2O2 can react with various cellular targets such as DNA, proteins, and lipids, causing cellular damage (22). As a protective system against increased intracellular levels of H2O2, cells possess several antioxidant enzymes that degrade H2O2. One prominent antioxidant enzyme in the cell is catalase, which converts H2O2 to O2 and H2O (23). Thus, a critical balance between the production and the metabolism of H2O2 is required to determine the cell fate. Although H2O2 can act as a toxic metabolite, it is ideally suited to be a signaling molecule because of its small size, so it can diffuse short distances within the cell, and because several mechanisms for its production and its rapid removal are present in the cell. Indeed, various studies have established that H2O2 has an intracellular signaling role in many biological systems. For example, H2O2 has been implicated in a wide range of plant responses, which include the closure of stomata and the regulation of cell expansion and development (24, 25). In addition, it has been demonstrated that H2O2 controls the expression of many genes, such as c-fos, c-jun, and c-myc (26), and that it activates cell signaling cascades, such as those involving mitogenactivated protein kinases (MAPKs), to mediate responses to a variety of extracellular stimuli (27). It is thereby established that H2O2 can act as a signaling molecule to induce a range of molecular, biochemical, and physiological responses in cells, and Author contributions: J.H. and P.S.-C. designed research; J.H. and S.C. performed research; J.H., S.C., and P.S.-C. analyzed data; and J.H. and P.S.-C. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. ‡To
whom correspondence should be addressed. E-mail:
[email protected].
© 2007 by The National Academy of Sciences of the USA
PNAS 兩 October 2, 2007 兩 vol. 104 兩 no. 40 兩 15747–15752
CELL BIOLOGY
Edited by Solomon H. Snyder, Johns Hopkins University School of Medicine, Baltimore, MD, and approved August 7, 2007 (received for review June 15, 2007)
Relative DCM-H2DCFDA Fluorescence
A
3
B
Time after H2O2 treatment (h) t 0 0.5 1 2 3 4 zCry1a
H2O2 concentration (uM) t 0 10-310-210-1 1 10 50102
2 zCry1a
zPer2 zPer1
1
zPer2
zCat zAox
zAox
0
RNA
D L NI
D L zCAT
D L LUC
Fig. 1. Light produces H2O2 in Z3 cells. Quantitative analysis of DCF fluorescence. Z3 cells were loaded with CM-H2DCFDA and then were exposed to light. CM-H2DCFDA fluorescence was detected by spectrofluorimetry. Values are means ⫾ SE of three independent experiments. In each experiment, the CM-H2DCFDA fluorescence of the nontransduced cells kept in the dark was set as 1. Where indicated, the cells were transduced with zCat or luciferase genes by retroviral vector (shown by zCAT and LUC, respectively).
C
Results Light Induces H2O2 Production in Z3 Cells. H2O2 is recognized to
function as a ROS signaling molecule that mediates responses to various stimuli in plants and animals (15, 16). Because light has been reported to induce production of H2O2 in cultured cells (19, 28), we hypothesized that it could be the signaling molecule that would convey the photic signal to the circadian transcriptional machinery in Z3 cells. We therefore evaluated H2O2 production in Z3 cells after light exposure. Z3 cells were loaded with a nonf luorescent, membrane-permeable derivative of CMH2DCFDA that fluoresces when oxidized by ROS and then were exposed to light. Light treatment leads to a robust increase in CM-H2DCFDA fluorescence in Z3 cells (Fig. 1). To assess the selectivity of the fluorescent response for H2O2, we infected Z3 cells with retroviral vectors expressing either the catalase or the luciferase genes. Catalase converts H2O2 to O2 and H2O (23). Expression of transduced genes was confirmed by catalase or luciferase activity assay (data not shown). The response of CM-H2DCFDA intensity was shown to be coupled to H2O2, because catalase-mediated transduction markedly decreased the fluorescent signal, whereas luciferase had no effect (Fig. 1). Taken together, these results demonstrate that light induces H2O2 production in Z3 cells. H2O2 Induces zCry1a and zPer2 Expression in Z3 Cells. We have
previously established that light induces zPer2 gene expression, an event that initiates the light-dependent entrainment process of Z3 cells permanently cultured in the dark (14, 29). To obtain evidence that H2O2 functions as signaling mediator for light-dependent circadian transcription, we examined the effect of H2O2 on circadian clock gene expression. Z3 cells were kept in constant darkness for 3 days and then stimulated with 100 M H2O2. zCry1a and zPer2 15748 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0705614104
t
D
L H
U N
zCry1a zPer2
D
Ctr t D L H
U0126 D L H
SB D L H
zCry1a
zAOX zPer2
E it is therefore conceivable that it could mediate signaling pathways for additional biological events. Here, we show that H2O2 is a key signaling mediator for light-dependent circadian gene expression in the zebrafish. In addition, we provide several lines of evidence that the antioxidant enzyme catalase is involved in circadian gene control by governing intracellular levels of H2O2. Our data strongly suggest that the alteration of the reduction– oxidation (redox) status within a cell is used as a signaling mechanism to transduce photic signals to circadian transcriptional regulation.
RNA
zCAT zAOX
Ctr U0126 SB t D L D L D L zAOX
RNA
Fig. 2. H2O2 induces zCry1a and zPer2 expression in Z3 cells and this induction is mediated by MAPK pathway. (A) A time course of circadian gene induction upon H2O2 treatment of confluent Z3 cells was examined by RNA protection assay. Z3 cells permanently cultured in the dark were treated with H2O2 (100 M). At indicated time points after the treatment, cells were harvested for RNA preparation. tRNA serves as our negative control reaction (t). Relative amounts of total RNA used for each sample are displayed (RNA). (B) Circadian gene induction by different concentrations of H2O2. Z3 cells kept in the dark and stimulated with H2O2 at the concentrations of 10⫺3, 10⫺2, 10⫺1, 1, 10, 50, and 100 M. After 100 min, cells were harvested for RNA preparation. As the negative control (0 M H2O2 concentration), we added water to the culture media. (C) Z3 cells kept in dark and stimulated with light (L), H2O2 (H) (100 M), UV (U) (10mJ/m2, 30 seconds), or MNU (M) (2 mM). After 100 min, cells were harvested for RNA preparation. (D) Confluent Z3 cells were treated with the indicated inhibitors or with the vehicle (DMSO, 0 M). The concentrations of U0126 and SB203580 inhibitors used were 40 M and 10 M, respectively. After 1 h, cells were stimulated with light (L) or H2O2 (100 M) (H), or were left in the dark (D). After 100 min, cells were harvested, and RNA was prepared. (E) Confluent Z3 cells were treated with the indicated inhibitors or with the vehicle (DMSO, 0 M). The concentrations of U0126 and SB203580 inhibitors used were 40 M and 10 M, respectively. After 1 h, cells were stimulated with light (L) or were left in the dark (D). After 7 h, cells were harvested, and RNA was prepared.
expression was strongly induced by H2O2, reaching maximal levels 2 h after treatment (Fig. 2A). Importantly, the profile of H2O2induced gene expression is analogous to that of light-dependent activation of zCry1a and zPer2, peaking 2 h after light onset (compare Figs. 2 A and 6A). On the other hand, H2O2 induced a slight reduction in zPer1 expression and had no apparent effect on zCat and zAox (Fig. 2 A). To better characterize the response of cells to H2O2, Z3 cells were stimulated for 100 min with H2O2 at various concentrations. Interestingly, a concentration as low as 1 nM caused a significant increase in zCry1a and zPer2 expression (Fig. 2B). Because H2O2 induces DNA damage in cells, we reasoned that DNA damage could represent a stimulus for clock gene expression. To test this possibility we treated Z3 cells with other genotoxic reagents, such as UV light or N-methyl-N-nitrosourea (MNU). Importantly, none of these DNA damage-inducing Hirayama et al.
t
0
2
6 12 18 24 26 30 36 42 48
A
Day1
zCry1a
Day2
Day3
Day4
t d 2 6 121824 2 6 121824 2 6 121824 2 6 121824
zPer2
zCry1a
zPer1 zPer2 zCat zCat RNA zPer1 Fig. 3. H2O2 induces circadian gene oscillation in Z3 cells. Z3 cells maintained in constant darkness were stimulated with H2O2 (100 M), and RNA was harvested at each time point indicated after the treatment. zCry1a, zPer1, zPer2, and zCat gene expression was examined by RPA, as shown in Fig. 2 A.
zAox
RNA 10
zCry1a
5 0 0
12
24
36
48
60
72
84
96 zPer2
10 5 0 0
12
24
36
48
60
72
84
96
3 zCat 2 1 0 0
12
24
36
48
60
72
84
96
4
H2O2 Induces Circadian Gene Oscillation in Z3 Cells. We have previously demonstrated that circadian gene expression in Z3 cells can be entrained by light (14). If H2O2 is the signaling molecule that conveys the photic signal to the circadian machinery, it should induce circadian gene expression. We therefore examined the expression profile of circadian genes during 48 h after H2O2 treatment (Fig. 3). Interestingly, zPer1 oscillated in H2O2-treated Z3 cells. The expression of zPer1 gradually diminished to a minimal value at ⬇12 h after H2O2 treatment to then rise again and fluctuate in a circadian fashion. Importantly, the expression profile of zPer1 in response to H2O2 treatment is analogous to the one obtained by exposing Z3 cells to a 12-h light (L) and 12-h dark (D) cycle (compare Figs. 3 and 4). The expression of zCry1a and zPer2 was different from that of zPer1, being only transiently induced after H2O2 treatment. Indeed, zCry1a and zPer2 mRNAs did not fluctuate significantly at times later than 2 h. Finally, H2O2 treatment had no effect on the expression of the zCat gene.
zCat Oscillates Antiphase to zCry1a and zPer2, and Its Oscillation Is Strictly Light-Dependent. For H2O2 to act as a second messenger,
there must be mechanisms for its removal subsequent to a signaling event. One of the well established enzymes regulating the cellular levels of H2O2 is catalase (Cat). Catalase is present in all aerobic organisms and protects cells from the damaging effects of H2O2 by converting H2O2 to O2 and H2O (23). Importantly, rhythmic oscillations in Cat transcripts have been reported in plants (31). We therefore examined the lightdependent regulation of zCat in Z3 cells (Fig. 4). Cells permanently cultured in the dark were exposed to light-dark (LD) cycles for 2 days and then placed in constant darkness (DD) for Hirayama et al.
zPer1 2 0 0
12
24
36
48
60
72
84 96 Time (h)
Fig. 4. Oscillation of zCat and circadian clock genes under LD and DD cycles in Z3 cells. Z3 cells maintained in constant darkness were exposed to 12:12 LD cycles for 2 days and then transferred to constant dark condition for another 2 days. RNA was harvested at each time point indicated after light onset. zCry1a, zPer1, zPer2, zCat, and zAox gene expression was examined by RPA (A) or RT-PCR analysis (B), as shown in Fig. 2. In RT-PCR analysis, the value from the cells kept in dark was set as 1 for each gene, and zebrafish actin gene was used for normalization. The bar above the blots indicates light (white) and dark (black) periods.
another 2 days. As described previously (14), the zPer1 gene oscillated under both LD and DD conditions, evidence of the strict circadian regulation of zPer1 transcription. On the other hand, expression of zPer2 and zCry1a genes is strictly lightdependent (14). Their transcriptional regulation was uncoupled from the circadian clock machinery, because zPer2 and zCry1a oscillation is present only under LD conditions but not under DD conditions. Interestingly, expression of zCat oscillates under LD conditions, but this oscillation ceased after transferring the cells to the DD condition. This light-dependent expression pattern is reminiscent of zPer2 and zCry1a expression, although zCat mRNA oscillates antiphase to zPer2 and zCry1a (Figs. 4 and 5A). Finally, expression of the zAox gene did not fluctuate under any PNAS 兩 October 2, 2007 兩 vol. 104 兩 no. 40 兩 15749
CELL BIOLOGY
MAPKs Mediate the Effect of H2O2. We have previously demonstrated that light-induced zPer2 expression in Z3 cells is mediated by the activation of the MAPK pathway (30). To examine whether H2O2 also relies on the same intracellular signaling pathway, we treated Z3 cells with a panel of kinase inhibitors (Fig. 2D). Importantly, the specific mitogen-activated protein (MAP)/extracellular signal-regulated kinase(ERK) kinase (MEK) inhibitor, U0126, blocked both H2O2- and light-induced zCry1a and zPer2 expression. We next tested the SB203580 inhibitor, specific for the stress-regulated p38 MAP kinases. This inhibitor did not show the drastic effect on both H2O2- and light-dependent zCry1a and zPer2 induction as U0126 did. These results demonstrate that MAPKs are essential for the signaling cascade that couples H2O2 to clock gene expression and that H2O2- and light-dependent effects operate within the same signaling cascade.
B
Relative mRNA expression level
treatments activated expression of zCry1a and zPer2 (Fig. 2C), demonstrating that H2O2 induction of both genes is not related to the DNA damage response.
zCry1a 5 4
Relative zCry1a/zPer2 expression
3
0
3
6
9
12
zPer2 11 9 7 5 3 1
2.6 2.2 1.8 1.4 1.0 0 3 6 9 12 Duration of light exposure (h)
2.0
0
6
12
18
24 Time (h) LD
1.6 DD 1.2
0.8
zCry1a 5 2.0 zCry1a /zPer2 1.5
4 3 2 1 0 zPer2
11 9 7 5 3 1
3
6
9
1.0 Catalase Activity 12 2.0
Relative Catalase Activity
Relative zCry1a/zPer2 expression
C
1
Relative Catalase Activity
B
2
2.6 2.2 zCry1a 1.8 /zPer2 1.4 1.0 zCat
Relative zCat expression
A
1.5 1.0 0 3 6 9 12 Duration of light exposure (h)
Fig. 5. zCat gene expression and catalase activity in Z3 cells cycle in antiphase to zPer2 and zCry1a mRNAs. (A) Z3 cells maintained in constant darkness were exposed to light, and RNA was harvested at each time point indicated after light onset. zCry1a, zPer2, and zCat gene expression was examined by RT-PCR analysis as shown in Fig. 4B. (B) Z3 cells maintained in constant darkness were exposed to a 12:12 LD cycle. Cells were harvested at each time point indicated after light onset for catalase activity assay. The catalase activity of cells at time 0 was taken as 1. Values are means ⫾ SE of three independent experiments. The bar above the graph indicates light (white) and dark (black) periods. (C) Catalase activity was assayed at each time point indicated, as shown in B.
condition. It is notable that H2O2 is not able to induce zCat gene expression (Figs. 2 A and 3). In addition, the U0126 MEK inhibitor, which blocks light induction of the zCry1a and zPer2 genes, has no effect on light induction of the zCat gene (Fig. 2E). These findings, together with the different kinetics of light inducibility for zCat with respect to zPer2 or zCry1a (Figs. 4 and 5A), strongly suggest that the light signaling pathway that leads to zCat expression is distinct from that for zPer2 or zCry1a. Catalase Enzymatic Activity Oscillates Antiphase to zPer2 and zCry1a Expression. We next decided to examine the temporal changes in
catalase enzymatic activity in Z3 cells. Z3 cells were cultured in constant darkness for 3 days and then entrained to a LD cycle. Catalase activity in Z3 cells showed a robust oscillation under LD conditions, reaching its maximal activity 12 h after light onset (Fig. 5B), with a precise antiphasic timing with respect to zPer2 and zCry1a expression (Fig. 5C). In contrast, no change in 15750 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0705614104
Fig. 6. zCAT is a key negative regulator for light-dependent activation of circadian genes. (A and B) Light induction of zCry1a and zPer2 expression in non- (Ctr, black line), zCat- (zCAT, red line), and luciferase- (LUC, green line) infected Z3 cells was examined by RPA (A) or RT-PCR analysis (B), as shown in Fig. 2 A and 4B, respectively. (C) Confluent Z3 cells were treated with catalase inhibitor sodium azide (350 M) or with the vehicle (DMSO, 0 M). After 2 h, cells were stimulated with light, and RNA was harvested at each time point indicated after light onset. zCry1a, zPer2, and zAox gene expression was examined by RPA, as shown in Fig. 2D. (D) A model depicting the molecular mechanism underlying light-dependent transcription in zebrafish circadian clock.
catalase activity was observed under DD conditions (Fig. 5B). These observations perfectly match the zCat transcriptional profile, which cycles only under LD condition, with a phase opposite to zCry1a and zPer2 expression (Figs. 4 and 5A). zCAT Regulates the Timing of Light-Dependent Clock Gene Induction.
Our observations indicate that H2O2 functions as the signaling molecule for zCry1a and zPer2 induction by light. We reasoned that zCAT could then operate as a key negative regulator of light-input pathway to the circadian machinery. The antiphasic expression and enzymatic activity of zCat with respect to the cyclic transcription of zCry1a and zPer2 supports this idea (Figs. 4 and 5). To further confirm this, we infected Z3 cells with retroviral vectors inducing ectopic expression of either zCat (pCLNCX-zCat) or luciferase (pCLNCX-Luc) genes. Expression levels of the transduced genes were confirmed by measuring their enzymatic activities (data not shown). In cells infected with pCLNCX-zCat, light induction of zCry1a and zPer2 genes was significantly reduced compared with noninfected cells or to cells infected with the pCLNCX-Luc retrovirus (Fig. 6 A and B). In noninfected cells or in cells infected with pCLNCX-Luc, zPer2 expression reached maximal levels 2 h after light stimulation, whereas it is maximal 3 h after the treatment in cells infected with pCLNCX-zCat. The regulatory effect of catalase was confirmed by the use of sodium azide, a catalase inhibitor. Treatment of Z3 cells with sodium azide results in enhanced light-mediated inducibility of zCry1a and zPer2 genes (Fig. 6C). Taken together, these results provide evidence that zCAT is an essential regulator of light-dependent circadian transcription, inhibiting zCry1a and zPer2 expression by reducing cellular H2O2 levels. Hirayama et al.
Hirayama et al.
the circadian clock. Importantly, light also increases the intracellular catalase activity by eliciting an increase of zCat transcription, after the maximum expression of zCry1a and zPer2 genes. The increased catalase activity diminishes the lightinduced cellular H2O2 levels leading to a lower zCry1a and zPer2 gene expression, thus creating a negative-feedback loop. An important issue raised by our model concerns the identity of the phototransducing molecules responsible for light-dependent H2O2 production in the zebrafish system. Because H2O2 can transduce a photic signal to the circadian machinery, those molecules should function as the circadian photoreceptor. Both the two prominent candidates for the circadian photopigments, melanopsin (42) and cryptochromes (43), were never shown to produce H 2 O 2 . One plausible hypothesis is that f lavincontaining oxidases function as the phototransducing enzymes responsible for the light-dependent H2O2 production. These oxidases are versatile flavoenzymes that catalyze oxygenation in a large number of metabolic pathways, generating H2O2 as a by-product (44). Importantly, it has been shown that flavincontaining oxidases can absorb light in the near violet–blue region (45), and that violet–blue light activates these enzymes and induces the photoreduction of flavin adenin dinucleotide (FAD), leading to H2O2 production (19). These results are in remarkable agreement with our previous data showing the involvement of a blue-light photoreceptor in the light induction of zPer2 gene (30). In addition, AOX, one of the flavincontaining oxidases, is constantly expressed in Z3 cells in all of the different lighting regimes (naı¨ve, LD, and DD conditions) (Figs. 2 and 4), which further supports the role of flavincontaining oxidases as putative circadian photoreceptors. Materials and Methods Cell Culture. The Z3 light-responsive cell line has been described
(14, 46). Cultured flasks were kept in a water-jacketed, thermostatically controlled, light-sealed incubator. Illumination was achieved by using a halogen light source fed into the incubator through a fiber optic line. A programmable timer connected to the light source controlled the light cycles. Light-dark cycles consisted of 12 h of light and 12 h of dark, unless otherwise noted. RNA Analyses. RNA-Solv (Omega Biotek, Doraville, GA) was used for total RNA extraction. A miniaturized RNase protection assay (RPA) was performed as described (46). Zebrafish Per1, Per2, and Cry1a probes were described (14, 30). The sequences for the zebrafish Cat and Aox are available on request. For RT-PCR analysis, total RNA was reverse-transcribed into cDNA by using M-MLV Reverse Transcriptase (GIBCO–BRL, Carlsbad, CA) with oligo random hexamers. Prepared cDNA was purified and subjected to quantitative PCR analysis by using Light Cycler (Roche Diagnostics, Indianapolis, IN) with SYBR green PCR Kit (Qiagen, Valencia, CA). Retroviral Infection. The RetroMax expression system (IM-
GENEX, San Diego, CA) was used to produce retrovirus according to the manufacturer’s instructions. zCat or luciferase genes were cloned into pCLNCX retroviral vector, in which the cloned genes are under the control of CMV promoter. We used pMD.G/vsv-g as enveloping vector. We confirmed the high infection efficiency (95–100%) by pCLNCX retroviral vector expressing GFP in Z3 cell line. Assay of Intracellular H2O2. Intracellular H2O2 levels were measured by f luorescence of 5-(and-6)-chloromethyl-2⬘,7⬘dichlorodihydrof luorescein diacetate acetyl ester (CMH2DCFDA; Molecular Probes, Eugene, OR). Z3 cells were incubated with CM-DCF-DA and then exposed to light. The fluorescence of CM-DCF-DA was measured by using a fluorescence image scanner. PNAS 兩 October 2, 2007 兩 vol. 104 兩 no. 40 兩 15751
CELL BIOLOGY
Discussion This study provides several lines of evidence supporting the role of H2O2 as the second messenger transducing the photo signal to clock gene transcription in the light-responsive zebrafish system. Indeed, light enhances H2O2 production in Z3 cells (Fig. 1). Treatment of Z3 cells with H2O2 triggers the induction (zCry1a and zPer2) and the circadian oscillation (zPer1) of clock genes with kinetics analogous to those elicited by light (Figs. 2 and 3). Interestingly, in contrast to zebrafish cells, H2O2 is not able to trigger the induction or the oscillation of circadian genes (Per1 and Per2) in mammalian cultured cells (data not shown), presumably because peripheral tissues are not light-responsive in mammals (32). Taken together, these findings strongly suggest that the capacity of zebrafish peripheral clocks to sense H2O2 underlies their direct light responsiveness. In further support of the role of H2O2 as light-signaling mediator, we have shown that the antioxidant enzyme zCAT is involved in the light-dependent regulation of zCry1a and zPer2 genes. Interestingly, in plants, Cat gene expression is regulated by light, circadian clock, or both (31, 33). This regulation is essential for a wide range of plant responses, because H2O2 mediates a variety of plant signaling pathways (22). In the present study, we found that light, but not the circadian clock, transactivates zCat gene as well as zCry1a and zPer2 genes (Fig. 4). Light induction of the zCat gene, however, is delayed compared with that of other genes, so that intracellular catalase activity starts to increase significantly with remarkable timing, that is, after the maximum expression of zCry1a and zPer2 genes (Figs. 4 and 5). We propose that this increased catalase activity reduces light-induced intracellular H2O2 levels, thus leading to down-regulation of zCry1a and zPer2 gene expression. Consistent with this scenario, zCry1a and zPer2 mRNAs oscillate antiphase to zCat gene expression and activity (Figs. 4 and 5), and overexpression of zCAT suppresses light induction of zPer2 and zCry1a, whereas inhibition of cellular catalase activity enhances it (Fig. 6). Although the suppression we observe is drastic, it is, however, not complete, suggesting the involvement of other antioxidant systems. In fact, cells possess many other enzymatic and nonenzymatic mechanisms to reduce intracellular H2O2 levels, such as glutathione peroxidase, vitamin C, and vitamin E (15). Whether these other systems contribute to light signaling pathway and/or circadian clock function still needs to be addressed. zCry1a and zPer2 genes are strictly light-dependent, and they both show the same profile of gene expression after a light pulse (Figs. 4 and 5). We therefore propose that both zCRY1a and zPER2 proteins are critical for the light-dependent entrainment of the zebrafish circadian clock (Fig. 6D). Studies in a variety of species have reported that each single cell contains a cellautonomous and self-sustained oscillator whose phases are not synchronized to each other in the absence of external time cues (34–39). zCRY1a and zPER2 interact with each other, forming a heterodimer that functions as a repressor of CLOCK:BMALmediated transcription (40, 41). Therefore, it is conceivable that the light-induced zCRY1a/zPER2 complex inhibits the transcription of clock genes by bringing them to their minimal activity levels, thus resetting all of the cell’s clocks to the same phase. This idea is somehow confirmed by the expression profile of the clock-controlled gene, zPer1, which starts to decrease after the induction of zCry1a and zPer2 genes (Figs. 2–4). In summary, we propose a working model diagrammed in Fig. 6D, in which the critical balance between the production and metabolism of H2O2 is required for the light-dependent regulation of zCry1a and zPer2 transcription. Light facilitates cellular H2O2 production, probably by activating flavin-containing oxidases (see below). This, in turn, stimulates the MAPK signaling pathway and then induces transactivation of zCry1a and zPer2 genes, an event that we believe is critical for light-entrainment of
Catalase Activity. Z3 cells were washed twice with PBS, homogenized in medium consisting of 50 mM phosphate buffer (pH 7.0), and then clarified by centrifugation for 10 min at 10,000 ⫻
g. The supernatant was used for the assay. Total catalase activity was measured by monitoring the decomposition of 10 mM H2O2 (47).
Dunlap JC (1999) Cell 96:271–290. King DP, Takahashi JS (2000) Annu Rev Neurosci 23:713–742. Cermakian N, Sassone-Corsi P (2000) Nat Rev Mol Cell Biol 1:59–67. Hogenesch JB, Gu YZ, Moran SM, Shimomura K, Radcliffe LA, Takahashi JS, Bradfield CA (2000) J Neurosci 20:1–5. Bunger MK, Wilsbacher LD, Moran SM, Clendenin C, Radcliffe LA, Hogenesch JB, Simon MC, Takahashi JS, Bradfield CA (2000) Cell 103:1009–1017. Kume K, Zylka MJ, Sriram S, Shearman LP, Weaver DR, Jin X, Maywood ES, Hastings MH, Reppert SM (1999) Cell 98:193–205. Shearman LP, Sriram S, Weaver DR, Maywood ES, Chaves I, Zheng B, Kume K, Lee CC, van der Horst GT, Hastings MH, Reppert SM (2000) Science 288:1013–1019. Hirayama J, Sassone-Corsi P (2005) Curr Opin Genet Dev 15:548–556. Pando MP, Sassone-Corsi P (2002) BioEssays 24:419–426. Cahill GM (2002) Cell Tissue Res 309:27–34. Whitmore D, Foulkes NS, Strahle U, Sassone-Corsi P (1998) Nat Neurosci 1:701–707. Whitmore D, Foulkes NS, Sassone-Corsi P (2000) Nature 404:87–91. Cermakian N, Whitmore D, Foulkes NS, Sassone-Corsi P (2000) Proc Natl Acad Sci USA 97:4339–4344. Pando MP, Pinchak AB, Cermakian N, Sassone-Corsi P (2001) Proc Natl Acad Sci USA 98:10178–10183. Neill S, Desikan R, Hancock J (2002) Curr Opin Plant Biol 5:388–395. Hancock JT, Desikan R, Neill SJ (2001) Biochem Soc Trans 29:345–350. Kawano T (2003) Plant Cell Rep 21:829–837. Allan AC, Fluhr R (1997) Plant Cell 9:1559–1572. Hockberger PE, Skimina TA, Centonze VE, Lavin C, Chu S, Dadras S, Reddy JK, White JG (1999) Proc Natl Acad Sci USA 96:6255–6260. Bolwell GP (1999) Curr Opin Plant Biol 2:287–294. Orozco-Cardenas ML, Narvaez-Vasquez J, Ryan CA (2001) Plant Cell 13:179– 191. Neill SJ, Desikan R, Clarke A, Hurst RD, Hancock JT (2002) J Exp Bot 53:1237–1247. Noctor G, Foyer CH (1998) Annu Rev Plant Physiol Plant Mol Biol 49:249–279. Kwak JM, Mori IC, Pei ZM, Leonhardt N, Torres MA, Dangl JL, Bloom RE, Bodde S, Jones JD, Schroeder JI (2003) EMBO J 22:2623–2633.
25. Foreman J, Demidchik V, Bothwell JH, Mylona P, Miedema H, Torres MA, Linstead P, Costa S, Brownlee C, Jones JD, et al. (2003) Nature 422:442–446. 26. Fornace AJ, Jr (1992) Annu Rev Genet 26:507–526. 27. Kamata H, Hirata H (1999) Cell Signal 11:1–14. 28. Squirrell JM, Wokosin DL, White JG, Bavister BD (1999) Nat Biotechnol 17:763–767. 29. Hirayama J, Cardone L, Doi M, Sassone-Corsi P (2005) Proc Natl Acad Sci USA 102:10194–10199. 30. Cermakian N, Pando MP, Thompson CL, Pinchak AB, Selby CP, Gutierrez L, Wells DE, Cahill GM, Sancar A, Sassone-Corsi P (2002) Curr Biol 12:844–848. 31. Zhong HH, McClung CR (1996) Mol Gen Genet 251:196–203. 32. Schibler U, Sassone-Corsi P (2002) Cell 111:919–922. 33. Zhong HH, Young JC, Pease EA, Hangarter RP, McClung CR (1994) Plant Physiol 104:889–898. 34. Hardin PE (1994) Mol Cell Biol 14:7211–7218. 35. Plautz JD, Kaneko M, Hall JC, Kay SA (1997) Science 278:1632–1635. 36. Mihalcescu I, Hsing W, Leibler S (2004) Nature 430:81–85. 37. Welsh DK, Yoo SH, Liu AC, Takahashi JS, Kay SA (2004) Curr Biol 14:2289–2295. 38. Nagoshi E, Saini C, Bauer C, Laroche T, Naef F, Schibler U (2004) Cell 119:693–705. 39. Carr AJ, Whitmore D (2005) Nat Cell Biol 7:319–321. 40. Hirayama J, Fukuda I, Ishikawa T, Kobayashi Y, Todo T (2003) Nucleic Acids Res 31:935–943. 41. Hirayama J, Nakamura H, Ishikawa T, Kobayashi Y, Todo T (2003) J Biol Chem 278:35620–35628. 42. Provencio I, Jiang G, De Grip WJ, Hayes WP, Rollag MD (1998) Proc Natl Acad Sci USA 95:340–345. 43. Sancar A (2000) Annu Rev Biochem 69:31–67. 44. Kim JJ, Miura R (2004) Eur J Biochem 271:483–493. 45. Setoyama C, Tamaoki H, Nishina Y, Shiga K, Miura R (1995) Biochem Biophys Res Commun 217:482–487. 46. Hirayama J, Kaneko M, Cardone L, Cahill G, Sassone-Corsi P (2005) Methods Enzymol 393:186–204. 47. Aebi H (1984) Methods Enzymol 105:121–126.
1. 2. 3. 4. 5. 6. 7.
8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.
15752 兩 www.pnas.org兾cgi兾doi兾10.1073兾pnas.0705614104
Hirayama et al.