Role of p53 in the entrainment of mammalian circadian behavior rhythms


  • Toshiyuki Hamada,

    1. Applied Molecular-Imaging Physics, Hokkaido University Graduate School of Medicine, Sapporo, Hokkaido 060-8638, Japan
    2. Ishida Group of Clock Gene, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki 305-8566, Japan
    Search for more papers by this author
  • Tomoko Niki,

    1. Ishida Group of Clock Gene, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki 305-8566, Japan
    Search for more papers by this author
  • Norio Ishida

    Corresponding author
    1. Ishida Group of Clock Gene, National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Ibaraki 305-8566, Japan
    2. Graduate School of Life and Environmental Sciences, University of Tsukuba, Tsukuba, Ibaraki 305-8566, Japan
    Search for more papers by this author

  • Communicated by:Shuh Narumiya


p53 protein plays a role for control of cell proliferation and the induction of apoptosis in normal cells. However, its role in the circadian rhythms that control many physiological functions including locomotor behavior remains unknown. The present study examined the locomotors activity rhythms of mice which have homozygous mutations of p53 gene. The period of drinking activity rhythms in p53 knockout (p53 KO) mice became unstable under constant dark. Light pulse causes a big phase shifts at CT15.5–17, when p53 mRNA expression peaks in the suprachiasmatic nucleus (SCN). Furthermore, photic entrainment of p53 KO mice is unusual under light–dark conditions. These findings suggest that p53 is involved in entrainment of the circadian behavioral rhythm.


Many biochemical, physiological and behavioral processes in life-forms ranging from microorganisms to vertebrates exhibit circadian rhythms (Edmunds 1988). The period (per) gene is required for the circadian rhythms of locomotor activity and eclosion behavior in Drosophila, and Per homologues in mammals have been identified (Sun et al. 1997; Tei et al. 1997; Sakamoto et al. 1998). Several recent studies of the molecular clock mechanism have suggested that circadian pacemakers consist of transcriptional/translational auto regulatory feedback loops of many clock gene products (Kriegsfeld et al. 2002). The casein kinase 1ɛ (CK1 ɛ) gene modifies PERIOD protein phosphorylation and affects the circadian rhythms of animal species ranging from Drosophila to humans (Ralph & Menaker 1988; Kloss et al. 1998; Lowrey et al. 2000; Toh et al. 2001; Xu et al. 2005). These facts indicate that autoregulatory feedback loops exist for several clock gene products among many species.

The cell division cycle in the dinoflagellate alga Gonyaulax as well as in the algal flagellates Euglena and Chlamydomonas is gated by a circadian rhythm (Edmunds 1988). Edmunds et al. (1979) also demonstrated that when the yeast S. cerevisiae is cultured at low temperatures, the circadian rhythm of division activity elongates the generation time. However, whether the cell division cycle is gated by the circadian clock in yeasts remains unknown. Nonetheless, the cell cycles of Gonyaulax, Euglena, Chlamydomonas and S. cerevisiae might also be closely associated with the circadian clock. Period 1 (Per1) interacts with the checkpoint proteins ATM and Chk2 in human cancer cells and plays an important role in regulating growth and the control of DNA damage (Gery et al. 2006).

The p53 tumor suppressor gene plays a critical role in the cell cycle, and it is frequently mutated in a range of human cancers (Kastan et al. 1991, 1992; Harris & Hollstein 1993). Many reports suggest that the gene is required for G1 growth arrest after DNA damage or the induction of apoptosis (Shaw et al. 1992; Clarke et al. 1993; Lowe et al. 1993). Although mice lacking p53 are viable, they develop various tumors from approximately four to 6 months of age (Donehower et al. 1992; Jacks et al. 1994; Tsukada et al. 1993). The loss of Cryptochrome (Cry), which is one of clock genes that forms a negative feedback loop in central circadian clock system, reduce can-cer risk in p53 mutant mice (Ozturk et al. 2009). Another circadian clock component, BMAL1 (also known ARNTL; Aryl hydrocarbon receptor nuclear translocator-like protein 1), works as a regulator of the p53 tumor suppressor pathway (Mullenders et al. 2009). Fu et al. (2002) have shown that the Per2 mutation induces disruption of the circadian oscillator and decreases p53 protein content, which cause mice to radiation-induced cancers. These reports suggest that p53 may be linked with molecular function of circadian clock.

Here, we examined the role of p53 in circadian rhythms using p53 KO mice. We found that the period of drinking activity rhythms became unstable and underwent large phase shifts in p53 KO mice after exposure to light pulses at CT15.5–17 when p53 mRNA expression peaks in the SCN.


Drinking activity rhythm of p53 KO mice under LD conditions

To determine whether p53 is involved in the circadian clock mechanism, we recorded the drinking activity rhythms of homozygotic p53 KO mice generated using TT2-ES-cells. TT2 cells originated from an F1 embryo derived from mating C57 BL/6 and CBA mice (Yagi et al. 1993). Wild-type mice were generated from TT2 -ES-cells.

All wild-type mice were entrained to the LD cycle and showed robust circadian activity rhythms (Fig. 1A), whereas entrainment of the p53 KO mice was quite unusual. The activity onset of p53 KO mice tended to be quite disordered compared with wild-type mice (+0.60 ± 0.62 h, n = 15 vs. +0.06 ± 0.46 h, n = 8, P < 0.05 Student's t-test; Fig. 1B,C,D). Some p53KO mice tended to split the drinking activity rhythm under LD (Fig. 1B) into D and early L phases. Three mice that did not become entrained to the LD cycle (Fig. 1E). One of the mice died during this study period.

Figure 1.

Drinking activity rhythm of p53 KO mice under LD. Double plots show 48 hours for each horizontal trace, and daily records are included right and below that of each preceding day. Solid bars, periods of intense drinking and activity; dotted lines, no drinking activity. Vertical units represent 5 days. (A) Wild-type mice can adequately entrain activity onset to LD cycle (Lights on, 07:00; lights-off, 19:00). (B–E) Activities of p53 KO mice. (B) Drinking activity rhythm is split; (C, D) Activity onset time (ZT 12) is abolished or E, not entrained to LD cycle. The dead day was indicated by arrow.

Characteristics of drinking activity rhythm of p53 KO mice under DD

We investigated drinking activity rhythms for approximately 3 months under DD using wild-type mice derived from TT2-ES-cells, C57BL/6 (Schwartz & Zimmerman 1990) and ICR (Oishi et al. 2002) mice as normal controls. Wild-type, C57 BL/6 and ICR mice showed free-running with a stable period (<24 h) of drinking activity rhythms after transfer to DD (Fig. 2A and data not shown). All three groups of control mice had periods that averaged between 23.7 and 23.85 h and stable circadian periods of <24 h. C57 BL/6 mice, ICR and CBA mice have stable circadian periods of <24 h (Schwartz & Zimmerman 1990; Yoshimura & Ebihara 1996, 1998; Oishi et al. 2002). Drinking activity rhythm of seven p53 KO mice were examined. But, unfortunately, four mice (4/7 mice) of p53 KO mice died during this study (the mice in Fig. 2B,D). We named each p53 KO mice in Fig 2B,C,D,E 8-p53 KO, 9-p53 KO, 10-p53 KO, 11-p53 KO, respectively. The period of drinking activity rhythms of p53 KO mice was unstable under DD and changed after exposing to light pulses (Fig 2B,C,D,E). Figure 2B shows that exposing 8-p53 KO mice to light at CT 20 induces a large phase shift and lengthened the period of drinking activity rhythms (>24 h). In contrast, exposing 8-p53 KO mice to light at CT 3.5 induced a phase shift and shortened the period of drinking activity rhythm (<24 h; Fig. 2B). The periods of 8-p53 KO mice appeared to be very unstable (Fig. 2B). Two p53 KO mice died after showing arrhythmic drinking activity for several days (Fig. 2B,D). Figure 2C (9-p53 KO) shows that light exposure at CT 16 induced a large phase shift and lengthened the period of drin-king activity rhythm (>24 h). The period of drinking activity rhythms changed in 10-p53 KO mice exposed to a light pulse at CT 16, but not CT 0.5 (Fig. 2D). The period of the 10-p53 KO mice appeared to be longer than 24 h; activity onset was gradually delayed (Fig. 2D). Interestingly, 11-p53 KO mice showed normal free-running with a stable period (<24 h) when exposed to a light pulse at CT 19.5 or at CT 15 (Fig. 2E). However, the period in 11-p53 lengthened to a light pulse at CT 12 (Fig. 2E). Taken together, p53 KO mice showed unstable periodicity under constant dark.

Figure 2.

Unstable periodicity of p53 KO mice under DD. Drinking activity rhythms in control and p53 KO mice exposed to light pulses (approximately 300 lux) for 1 h. Vertical black bar represents 5 days. (A) Wild mice have stable circadian periods of <24 h. (B–E) Activities of p53 KO mice (homozygotes). The dead day was indicated by arrow in Fig. 2B,D: Period of drinking activity rhythms these mice tended to become destabilized under DD with frequent large changes and phase shifts especially after exposure to light.

Under LL condition, the period of drinking activity rhythm in p53 KO mice (25.42 ± 0.28 h, n = 5) had no difference from that of wild mice (25.40 ± 0.38 h, n = 4; P = 0.46, Student's t-test; Fig. S1 in Supporting Information), suggesting that photic input is normal in p53 KO mice.

Phase response curve to light pulse

Figure 3a shows the phase response curve (PRC) of C57BL/6 and ICR mice exposed to light. The phase shifts in C57BL/6 and ICR mice during subjective day (CT0–11) were −0.02 ± 0.16 h (n = 5) and 0.17 ± 0.16 h (n = 3), respectively, and during early subjective night (CT12.5–17), −1.76 ± 0.29 h (n = 8) and −1.77 ± 0.06 h (n = 3), respectively. During late subjective night (CT19–23), these values were +0.66 ± 0.59 h (n = 6) and +0.76 ± 0.11 (n = 3) h, respectively. The PRC of wild-type mice exposed to light was +0.13 ± 0.44 h (n = 6), −1.74 ± 0.34 h (n = 7) and +0.65 ± 0.41 h (n = 5) during subjective day, and early and late subjective night, respectively (Fig. 3B). Figure 3C showed that the phase delay was very large in p53 KO mice compared with control mice (Fig. 3A,B). Large phase delays were generated in p53 KO mice after exposure to light at early subjective night (−2.52 ± 1.08 h, n = 8, P < 0.05 vs. wild mice, Student's t-test). The phase delay from CT 15.5 to 17 was particularly large (−3.23 ± 0.45 h, n = 5), being about twice that of control mice. The phases shifts during late subjective night and subjective day were −0.41 ± 1.35 h (n = 7; P = 0.06 vs. wild mice, Student's t-test) and 0.67 ± 0.89 h (n = 4; P = 0.08 vs. wild mice, Student's t-test). The data suggest that a role of p53 gene might be involved in light entrainment.

Figure 3.

Phase response curves of control and p53 KO mice to light. Each point in phase response curves represents phase shift in drinking activity rhythm, induced by light pulses (approximately 300 lux) at circadian time corresponding to position on abscissa. Responses to light pulse of A, C57BL/6 and ICR; (B) wild type; (C) p53 KO mice.

Expression of p53 mRNA in the SCN, eye and pineal

We examined the morphology of the SCN of p53 KO mice by cresyl violet staining. The hypothalamic areas including the SCN and the magnified SCN (Fig. 4A) of p53 KO mice did not apparently differ from that of wild-type mice (Silver et al. 1999). Therefore, the disrupted behavior of circadian rhythms in p53 KO mice cannot attribute to gross developmental defects in the SCN.

Figure 4.

Expression of p53 mRNA in the SCN, eye and pineal. (A) Morphological staining (cresyl violet) of SCN in p53 KO mice. Third ventricle (3V); optic chasm (OC); suprachiasmatic nucleus (SCN). (B) Rhythmic expression of p53 mRNA in mouse SCN determined by RPA (RNase protection assay). Signals in RPA were quantified using BAS 2000 image analyzer and normalized by β-actin. One-way ANOVA followed Fisher's test. *< 0.05 vs. CT6. (C) Expression of p53 mRNA in mice eye by RPA. (D) Expression of p53 mRNA in rat pineal by Northern blotting analysis.

Then, we quantified p53 mRNA in the SCN, eyes and pineal (Guilding & Piggins 2007) of wild-type mice under constant darkness using RNase Protection Assay (RPA) or Northern blotting assay. We also examined the effects of light pulses on beh-avior rhythms (Fig. 3) by comparing the amounts of mRNA obtained at the middle of subjective day, and at early and late subjective night. The amount of p53 mRNA in the SCN had changed with circadian manner with a peak CT 15 (Fig. 4B), whereas that in the eye did not changed (Fig. 4C). In addition, p53 mRNA expression in the eye of wild mice had not affected by light pulse (300 lux, 1 h exposure; Fig. 4C). But Pineal showed circadian expression of p53 mRNA by Northern blotting analysis (Fig. 4D). These results suggest that circadian expression of p53 mRNA in the SCN or pineal might have a role of circadian behavior.


The period of circadian rhythm is normally stable (Schwartz & Zimmerman 1990; Yoshimura & Ebihara 1996, 1998; Oishi et al. 2002), and the circadian clock is robust and constant. However, the present study showed that the period of p53 KO mice showed very unstable. Furthermore, light pulses caused a large phase shift at early subjective night and changed the time of the response to light pulses in p53 KO mice compared with controls. It is considered that the large phase shifts and period change induced by light pulses were time dependent. These phenomena mainly occurred at CT 15.5–17, when the p53 mRNA content in the SCN peaked, suggesting that p53 is involved in a component of the circadian clock mechanism. Light pulses did not significantly alter the content of p53 mRNA in the SCN (data not shown) and eyes at CT 15.5–17 (Fig. 4C). Therefore, the present findings suggest that the activation of p53 protein rather that of p53 mRNA plays an important role in light-induced phase shifts.

In mouse liver, it is reported that p53 mRNA has no rhythmic expression (Fu et al. 2002; Gréchez-Cassiau et al. 2008). But p 21Waf1/CIP1 mRNA shows rhythmic expression which is a major target of p53 activated after DNA damage and plays an important role during epidermis differentiation (Gréchez-Cassiau et al. 2008). MDM2 mRNA involved in post-transcriptional regulation of p53 also shows rhythmic expression in mouse liver (Fu et al. 2002). At post-transcriptional levels, p53 in oral mucosa is controlled by circadian clock (Bjarnason et al. 1999). Therefore, p53 protein might have an important role in circadian system in both central nervous system and peripheral tissue.

The relationship between clock genes and cancer remains unclear. Disruption of the circadian oscillator by the Per2 mutation caused mice to radiation-induced cancers (Fu et al. 2002). However, Cry 1/Cry2 and clock/clock mutant and wild-type mice are indistinguishable with respect to the incidence of tumors induced spontaneously and by ionizing radiation (IR; Gauger & Sancar 2005; Antoch et al. 2008). A recent study has shown that p53 KO mice live 50% longer when Cry is deleted (Ozturk et al. 2009). Direct interactions between p53 and clock genes in cancer remain to be investigated.

Very recently, CC Lee and colleges reported p53 binding to a response element in the Per2 promoter and induced circadian expression of Per2 (Miki et al. 2013). They also found out unstable behavior of p53 KO mice as we reported in this article. Behavioral analysis of p53 KO mice showed a widely distributed period that exhibited sudden changes during free-running conditions. These data suggest that abnormal behavior of p53 KO mice might be due to abnormal circadian expression of Per2 gene. Furthermore, phase shifts response to a 15-min light pulse enhanced phase delay in p53 KO mice. In the present study, our results also suggest that p53 have a role in the entrainment of mammalian circadian behavior rhythms.

In conclusion, our findings indicate that p53 might be involved in photic entrainment of the circadian clock mechanism in mice.

Experimental procedures

Knockout mice

To introduce a mutation into the p53 gene by homologous recombination in embryonic stem (ES) cells, the neor (neomycin phosphotransferase) gene was inserted into the NcoI site in the second exon in the opposite orientation to the p53 gene in the targeting vector (Tsukada et al. 1993). This insertion disrupted the translation initiation codon so that homologous recombination with this vector would generate a null mutation of the p53 gene. We confirmed the genotype of the homozygotic p53 KO mice using the polymerase chain reaction (PCR; data not shown). All animal experiments were carried out in a humane manner after receiving approval from the Institutional Animal Experiment Committee of AIST (Ibaragi, Japan).

Recording drinking activity rhythms

The mice were individually housed in cages equipped with an infrared sensor, and food and water were provided ad libitum. Drinking activity was monitored using an on-line PC (Chronobiology Kit; Stanford Software System Stanford, CA, USA), and daily activity onset was visually estimated from standard double-plot actograms of drinking activity frequency. Behavioral activity rhythms were analyzed using a Chronobiology kit. Circadian periods were analyzed using chi-square periodgrams.

Mice maintained for at least 1 week on a light–dark (12 h light: 12 h dark; LD) cycle were transferred to constant darkness (12 h dark: 12 h dark; DD). After establishing a stable free-running rhythm, the mice were exposed to light pulses (approximately 300 lux) for 1 h at a specified circadian time (CT) when the activity onset was designated as CT12. Drinking activity rhythms of the mice were recorded from 6 to 8 weeks of age and continued for approximately 3 months. Some mice were transferred to constant light (12 h light: 12 h light; LL) after housing under LD cycle and then transferred to DD condition.

The phase angle of entrainment was measured as the time of activity onset relative to the time of lights-off and determined from successive lines of activity onset times from at least 1 week of activity onset time in LD to initial free-running in DD. A line was drawn manually on plots through the times of activity onset.

mRNA expression analysis

After 2 days in DD, wild mice were decapitated at a specified CT and then SCN were dissected. Sampling for eyes, after 2 days in DD, wild-type mice were exposed to a 1 h light pulse at CT6, CT15, CT21 and decapitated after light pulse soon. Wild-type mice with no light pulse were decapitated at the same time. The tissues from 10 mice were pooled, frozen in liquid nitrogen and stored at −80 °C until RNA extraction. Sampling for pineal, after 2 days in DD, Wistar rats were decapitated at a specified CT. The tissues from five rats were pooled, frozen in liquid nitrogen and stored at −80 °C until RNA extraction.

Total RNA was extracted from the mouse tissues using acid guanidinium phenol chloroform (Chomczynski & Sacchi 1987). A cDNA probe was generated from a pTRI-p53-mouse antisense probe template (Ambion, Austen, TX, USA) containing 265 bp of a highly conserved region spanning exons 4 through 7 of the mouse p53 gene for Ribonuclease Protection Assays (RPA). For Northern blotting analysis, 32P-labeled random primed probe was generated from rat p53 cDNA fragment (nt 514–808 in open reading frame).

RPA were carried out using RPA kits (Ambion) according to the manufacturer's instructions (Hamada et al. 1999). The RPA control was pTRI-Actin (mouse β-actin DNA) provided in the RPA kit. Hybridizations proceeded as follows. Total RNA from the eyes and SCN (10 μg) was hybridized with the p53 probe (1 × 105 cpm/μL), and 5 μg each from other tissues was hybridized with the mouse ß-actin probe (1 × 104 cpm/μL) at 43 °C for 12–15 h. RNase digestion proceeded using RNase A/T1 (diluted 1: 100) at 37 °C for 30 min and then stop buffer was added. Protected fragments were then precipitated with ethanol (100 mL). The sizes of the protected p53 and the β-actin cDNA transcripts were 265 and 126 bp, respectively. The fragments were precipitated for 30 min at −20 °C, pelleted by centrifugation (TOMY Eppendorf) at 12 000 × g for 20 min, dissolved in the buffer provided with the kit (Ambion) and resolved by 5% urea gel electrophoresis. The gel was dried onto Whatman 3 MM paper, exposed to an imaging plate (Fuji Film, Tokyo, Japan) and quantified using a Fujix Bio-imaging analyzer BAS 2000 (Fuji Film; Hamada et al. 1999). In Northern blotting analysis, total RNA from pineal (10 μg) was hybridized with 32P-labeled random primed probe, and following steps were carried out as described in the previous reports (Sakamoto et al. 1998).

Morphological staining with cresyl violet

The right auricle of anesthetized p53 KO mice (8 weeks of age) was cut, and the left side of the heart was perfused with 4% paraformaldehyde in phosphate-buffered saline (PBS) (−), pH 7.4. The brain was removed and fixed in 4% paraformaldehyde in PBS(−) for 3 h at 4 °C, then, paraffin sections (8 μm) were prepared as described (Matsui et al. 1996), and the SCN was stained with cresyl violet.


We thank Dr Shinichi Aizawa for a generous gift of p53 KO mice. This study was supported by a project grant for the Competitive Research Program from MITI, Japan, and was partially supported by Special Expenditures of ‘Reverse Translational Research from Advanced Medical Technology to Advanced Life Science; and from Real-time Tracking Technology to Real-time Tracking Life Science’ funded from Japanese Ministry of Education, Culture, Sports, Science and Technology.