Research ArticlePhysiology

Involvement of the Protein Kinase CK2 in the Regulation of Mammalian Circadian Rhythms

See allHide authors and affiliations

Science Signaling  02 Jun 2009:
Vol. 2, Issue 73, pp. ra26
DOI: 10.1126/scisignal.2000305

Abstract

Posttranslational modifications of clock proteins are crucial to generating proper circadian rhythms of the correct length and amplitude. Here, we show that the protein kinase CK2 (casein kinase 2) plays a role in regulating the mammalian circadian clock. We found that inhibiting CK2 activity resulted in a decrease in the amplitude and an increase in the period of oscillations in circadian gene expression. CK2 specifically bound and phosphorylated PERIOD2 (PER2) and collaborated with the protein kinase CKIε to promote PER2 degradation. We also identified a CK2 phosphorylation site (serine-53) in PER2, whose phosphorylation played a role in fine-tuning circadian rhythms and regulating PER2 stability but was dispensable for the cooperative effect of CK2 and CKIε. Thus, our study identifies CK2 as a regulatory element of mammalian circadian rhythms and uncovers a role for CK2 in PER2 degradation.

Introduction

Circadian rhythms are daily biological cycles that control various physiological processes (1, 2). Transcriptional and translational feedback loops are believed to comprise the core molecular mechanisms that generate circadian rhythms (3, 4). In the primary autoregulatory feedback loop in mammals, the CLOCK-BMAL1 heterodimer promotes transcription of the Period (Per) and Cryptochrome (Cry) genes and, in turn, PER and CRY repress the transcriptional activity of CLOCK-BMAL1. CLOCK-BMAL1 also promotes transcription of the gene encoding REV-ERBα, which inhibits Bmal1 transcription, constituting the secondary feedback loop. In addition, RORα promotes Bmal1 transcription, and NPAS2 [neuronal PAS (Per-Arnt-Sim) domain protein 2], a close analog of CLOCK, functions as a CLOCK substitute to promote transcription of Per and Cry. The transcription factors DBP (D-element binding protein) and E4BP4 (E4 binding protein 4) have also been implicated in the regulation of circadian rhythms (5). Posttranslational modification of clock proteins plays a critical role in determining period length and rhythm robustness (6). In particular, PER phosphorylation by DOUBLETIME in Drosophila or by its mammalian orthologs casein kinase Iε and δ (CKIε/δ) in mammals has been well described as a critical regulatory mechanism for its degradation and subcellular localization (710). However, although the protein kinase CK2 has been implicated in the control of circadian rhythms of Arabidopsis (11), Drosophila (12, 13), and Neurospora (14), there is no evidence for its role in mammalian circadian rhythms. Here we present evidence for the involvement of CK2 in the mammalian circadian clock. Moreover, our analysis reveals a collaborative effect of CK2 and CKIε on PERIOD2 (PER2) degradation, suggesting that CK2 plays an important role in generating accurate circadian rhythms.

Results

Inhibition of CK2 activity affects circadian gene expression

We hypothesized that CK2, which exists as a holoenzyme with catalytic (α) and regulatory (β) subunits, might regulate mammalian circadian rhythms. To test this idea, we examined whether inhibiting its kinase activity affected the expression of genes that encode proteins involved in maintaining circadian rhythm (circadian genes). Using a Bmal1-Luciferase (Bmal1-Luc) reporter system (15), we recorded bioluminescence from cultured mouse fibroblasts by means of a photomultiplier tube to assess both the period and the amplitude of circadian rhythms. The addition of the CK2-specific inhibitor 2-dimethylamino-4,5,6,7-tetrabromo-benzimidazole (DMAT) at 0 or 12 hours after dexamethasone stimulation, which can entrain the rhythms of circadian gene expression, caused a marked decrease in rhythm amplitude and a small but significant increase in period (Fig. 1, A and C). Another compound that inhibits CK2 activity, 5,6-dichloro-1-β-d-ribofuranosylbenzimidazole (DRB), also dampened rhythm amplitude (fig. S1). Similarly, expression of a plasmid encoding CK2α-K68A, a catalytically inactive form of CK2α that is reported to act as a dominant negative (16), led to a small but significant decrease in the amplitude and also a small but significant increase in the period (Fig. 1, B and C). In contrast, overexpression of both the catalytic subunit α and the regulatory subunit β of CK2 resulted in increased amplitude and shortened period (Fig. 1, B and C). Although DMAT has been reported to inhibit other kinases than CK2 (17), these results together indicate that altered CK2 activity alters circadian rhythms in cultured cells.

Fig. 1

Alteration of CK2 activity affects circadian gene expression. (A) Circadian oscillation of Bmal1-Luc reporter activity is affected by DMAT. Bioluminescence from dexamethasone-treated NIH3T3 cells was recorded (cpm, counts per minute). DMSO or DMAT was added 0 hours (left) or 12 hours (right) after dexamethasone treatment. Representative raw data and detrended data are shown at the top and the bottom, respectively. Three independent experiments gave similar results. (B) Expression of CK2 subunits alters circadian oscillation of the Bmal1-Luc reporter activity. Cells were transfected with the Bmal1-Luc reporter plasmid and plasmids encoding the indicated CK2 subunits. Bioluminescence was recorded as in (A). Three independent experiments gave similar results. (C) Calculated periods and amplitudes of bioluminescence rhythms shown in (A) and (B). Means ± SD are shown for each trace in (A) (n = 4) and (B) (n = 3). P values as compared to DMSO or vector are indicated. (D) CK2 inhibition dampens amplitude of circadian oscillation of Bmal1 mRNA expression. Serum-stimulated cells were incubated in the presence of DMSO or DMAT. mRNA abundance was determined by real-time quantitative PCR. Data were normalized to 18S rRNA and values at time 0 were set to 1. Error bars indicate SD (n = 3).

The effect of inhibiting CK2 on circadian gene expression was further analyzed by monitoring the expression of endogenous clock genes. Consistent with the Bmal1-Luc reporter experiments, DMAT markedly dampened the amplitude of circadian oscillations of Bmal1 messenger RNA (mRNA) expression (Fig. 1D). The rhythm amplitude of the DMAT-treated group was estimated to be 67% of that of the dimethyl sulfoxide (DMSO)–treated group. In this experiment, however, it was difficult to estimate the period because of the long sampling intervals. DMAT also decreased the amplitude of the Per2 mRNA expression rhythm (fig. S2). We also examined the effect of DMAT on bioluminescence rhythms in suprachiasmatic nucleus (SCN) slices from Per2::Luciferase (Per2::Luc) knock-in mice (18) and found that DMAT significantly dampened reporter activity oscillation, which reflects oscillation of PER2::LUC expression, and lengthened period (Fig. 2, A to C). DMAT washout reinitiated a robust bioluminescence oscillation, suggesting that the reduced amplitude was not the consequence of unhealthy tissue. DMSO vehicle alone affected neither the rhythm amplitude nor the period (Fig. 2, A to C). These results strongly suggest that CK2 kinase activity is necessary for normal mammalian circadian rhythms in vivo.

Fig. 2

CK2 inhibition affects circadian rhythms in SCN slices. (A) CK2 inhibition affects circadian oscillation of the PER2::LUC reporter activity in SCN slices. Bioluminescence from the SCN slices excised from Per2::Luc knock-in mouse were recorded. Shown are the representative raw data. Red bar and gray bar indicate treatment with DMAT (40 μM) and DMSO, respectively. Filled arrowhead indicates the exchange to fresh medium. (B) Period lengths of individual slices before and after treatment with DMAT or DMSO are plotted on the left. Period lengths were determined from more than three cycles of bioluminescence rhythms. Means (n = 3) of period increases are shown on the right. *P < 0.005. (C) Means ± SD (n = 3) of amplitude changes were calculated as the ratio of the amplitude of bioluminescence rhythms from 24 to 48 hours after DMAT or DMSO treatment to that from 24 to 48 hours before treatment. *P < 0.05.

Phosphorylation of a CK2 substrate shows circadian oscillation in vivo

CDC37, a molecular chaperone, is a major CK2 substrate; Ser13 of CDC37, which is reported to be specifically phosphorylated by CK2, is the only CDC37 site that is phosphorylated in vivo (19). We found that phosphorylation of CDC37 Ser13 showed a circadian oscillation in both mouse liver (fig. S3A) and serum-stimulated cultured fibroblasts (fig. S3B), although the total amounts of CDC37 (phosphorylated and unphosphorylated), CK2α, and CK2β were almost constant (fig. S3, A and B). The ability of cell lysate obtained from serum-stimulated fibroblasts to phosphorylate the CK2 site of CDC37 in vitro was also nearly constant and did not show a circadian oscillation (fig. S3C). These results suggest that the phosphorylation and dephosphorylation events seen at CDC37 Ser13 in vivo may be regulated in a circadian manner, although circadian regulation of CK2 activity or phosphatase activity or both could not be preserved in the cell-free system.

CK2 specifically binds and phosphorylates PER2

To identify the CK2 target(s) involved in regulating the mammalian circadian clock, we examined a set of clock proteins for their ability to interact with CK2. COS7 cells were transfected with Myc-tagged clock proteins together with FLAG-tagged CK2 α or β subunits. Immunoprecipitation assays with antibody against FLAG (anti-FLAG) revealed that CK2β strongly interacted with PER2 (Fig. 3A). CK2β also interacted with DBP and interacted weakly with CRY2.

Fig. 3

CK2 binds and phosphorylates PER2. (A) Interaction between PER2 and CK2. COS7 cells were cotransfected with Myc-tagged clock proteins and FLAG-tagged CK2 subunit α or β. Cell extracts were subjected to IP with anti-FLAG and analyzed by immunoblotting with the indicated antibodies. Asterisks indicate nonspecific bands. (B) Interaction between PER2 and CK2α. COS7 cells were cotransfected with Myc-tagged PER2, HA-tagged CK2α, and FLAG-tagged CK2β. Cell extracts were subjected to immunoprecipitation with anti-HA and analyzed by immunoblotting with the indicated antibodies. (C) Interaction between PER2 and CK2β in vivo. Liver extracts of mice that were prepared at CT18 were subjected to immunoprecipitation with anti-PER2 or anti-HA (control) and analyzed by immunoblotting with the indicated antibodies. (D) CT-dependent interaction between PER2 and CK2β in vivo. Liver extracts were prepared at CT6 and CT18 and analyzed as in (C). (E) (Top) PER2 is phosphorylated by CK2. Myc-tagged clock proteins expressed in COS7 cells were immunoprecipitated with anti-Myc and incubated in the presence of [γ-32P]ATP with or without recombinant CK2α. Autoradiograph was analyzed by a phosphorimager. (Bottom) The indicated input proteins were analyzed by immunoblotting. Arrowheads indicate the positions of clock proteins. (F) In vitro translated (IVT) PER2 is phosphorylated by CK2. Myc-tagged PER2 was translated in vitro and immunoprecipitated with anti-Myc. In vitro kinase assay was performed as in (E).

We then examined the possibility that CK2α interacts with PER2 in the presence of CK2β. COS7 cells were transfected with plasmids encoding Myc-tagged PER2, hemagglutinin (HA)-tagged CK2α, and FLAG-tagged CK2β. Immunoprecipitation assays with antibody against HA (anti-HA) showed that, in the presence of CK2β, CK2α interacted with PER2 (Fig. 3B). These data indicate that PER2 can interact with the CK2 holoenzyme. To investigate in vivo interactions between endogenous CK2 and PER2, we prepared mouse liver extracts and performed immunoprecipitation assays with an antibody to Per2. As a result, CK2β was efficiently coimmunoprecipitated with PER2 (Fig. 3C). The interaction between CK2β and PER2 depended on circadian time (Fig. 3D), because PER2 abundance varied with circadian time.

These results show that PER2 interacts with CK2β in vivo. In vitro kinase assays with recombinant CK2α and Myc-tagged clock proteins expressed in COS7 cells showed that PER2 was the best CK2 substrate among the clock proteins tested (Fig. 3E). DBP was phosphorylated only weakly by CK2 (Fig. 3E). PER1 and PER2 were phosphorylated even in the absence of CK2, likely by some coimmunoprecipitated endogenous kinase(s). This raised the possibility that CK2 might enhance the activity of the coimmunoprecipitated kinase(s) that directly phosphorylates PER2. To determine whether CK2 phosphorylates PER2 directly, we used in vitro translated PER2 as a substrate and found that that it was phosphorylated by CK2 in vitro (Fig. 3F). Given the importance of PER2 to the mammalian circadian clock, these data identify Per2 as a strong candidate for the CK2 target involved in control of mammalian circadian rhythms.

CK2 regulates PER2 degradation

Because changes in PER2 stability have been reported to be associated with changes in the period length of circadian rhythms (20, 21), we examined whether forced expression of CK2 affected PER2 stability. PER2 was stable when expressed in COS7 cells, and expression of CK2 did not significantly affect PER2 stability (Fig. 4A). In contrast, expression of CKIε destabilized PER2 and coexpression of CK2 significantly enhanced CKIε-dependent PER2 degradation (Fig. 4A). Expression of the α subunit of CK2 enhanced CKIε-dependent PER2 degradation, whereas expression of the β subunit stabilized PER2 (Fig. 4B), suggesting that the catalytic subunit α is responsible for the potentiation of CKIε-dependent PER2 degradation. To evaluate the requirement for the kinase activity of CK2α, we used CK2α-K68A, a catalytically inactive form of CK2α, instead of wild-type CK2α and found that CK2α-K68A did not enhance CKIε-dependent PER2 degradation (Fig. 4C). These results show that the kinase activity of CK2 is required for its ability to enhance CKIε-dependent PER2 degradation. We next examined the effect of CK2 inhibition on PER2 stability. The cell-permeable phosphatase inhibitor calyculin A has been shown to elicit rapid degradation of PER2 (22), and we found that DMAT significantly inhibited this calyculin A–induced PER2 degradation (Fig. 4D). These results indicate that CK2 promotes PER2 degradation in a manner that depends on its kinase activity.

Fig. 4

CK2 regulates PER2 degradation. (A to C) Cells transfected with the indicated plasmids were treated with cycloheximide (CHX) (t = 0). Cell extracts were analyzed by immunoblotting with antibodies to the indicated proteins. The graphs show quantified PER2 abundance. The values were normalized to α-tubulin and the values at t = 0 were set to 1. Mean ± SEM (n = 3) are shown. *P < 0.05 compared to CKIε(+)/CK2(−). Three independent experiments gave similar results. (D) Cells transfected with Myc-tagged PER2 were pretreated with cycloheximide and either DMSO or DMAT and treated with calyculin A (t = 0). Cell extracts were analyzed as in (A). The graph shows the relative PER2 abundance at t = 60 min where the values at t = 0 were set to 1. Mean ± SEM (n = 3) are shown. *P < 0.05.

Identification of PER2’s CK2 binding region

We hypothesized that CK2-mediated phosphorylation of PER2 might regulate its degradation. To address this possibility, we first determined the CK2 binding region within PER2. We coexpressed a series of Myc-tagged PER2 deletion mutants with FLAG-tagged CK2β in COS7 cells and performed immunoprecipitation experiments with anti-FLAG. Both an N-terminal fragment (residues 1 to 330) and a C-terminal fragment (residues 1056 to 1257) of PER2 strongly interacted with CK2β (Fig. 5A). A shorter PER2 N-terminal fragment (residues 1 to 148) did not interact with CK2β, indicating that amino acid residues 149 to 330 are necessary for this interaction. We also found that the N-terminal fragment (residues 1 to 330) of PER2 coimmunoprecipitated with the C-terminal fragment (residues 1056 to 1257) of PER2 in the absence of CK2 overexpression (Fig. 5B). These results suggest that the N- and C-terminal regions of PER2 are juxtaposed in the protein and serve as a binding site for CK2.

Fig. 5

Identification of the CK2 binding regions and the CK2 phosphorylation site within PER2. (A) CK2β interacts with both the N-terminal and the C-terminal regions of PER2. COS7 cells were cotransfected with Myc-tagged PER2 deletion mutants and FLAG-tagged CK2β. Cell extracts were subjected to immunoprecipitation with anti-FLAG and analyzed by immunoblotting with the indicated antibodies. The structures of PER2 deletion constructs are shown on the left. (B) Interaction between the N-terminal and the C-terminal regions of PER2. COS7 cells were cotransfected with Myc-tagged PER2 (residues 1 to 330) and HA-tagged PER2 (residues 1056 to 1257). Cell extracts were subjected to immunoprecipitation with anti-HA and analyzed by immunoblotting with the indicated antibodies. (C) An N-terminal region of PER2 is phosphorylated by CK2. The indicated deletion mutants of Myc-tagged PER2 were used for the in vitro kinase assay. (D) (Top) Alignment of mouse (m), rat (r), and human (h) PER2 amino acid sequences corresponding to amino acid residues 45 to 80 of mouse PER2. Serine/threonine residues and acidic residues are indicated in blue and red, respectively. Filled green circles indicate serine/threonine residues conserved among these three species. (Bottom) Myc-tagged PER2 (residues 1 to 621) with the indicated alanine substitution were used for in vitro kinase assay. (E) DMAT inhibits CK2-mediated phosphorylation of PER2, and CK2 holoenzyme phosphorylates PER2. PER2 (residues 1 to 621) was used for in vitro kinase assay with the indicated concentrations of DMAT. CK2 holoenzyme was also used, instead of CK2α. The relative band intensities were quantified and normalized to CBB staining. (F) CK2, but not CKIε or GSK-3β, phosphorylates Ser53 of PER2. FLAG-tagged CKIε and recombinant GSK-3β were used for the in vitro kinase assay, instead of CK2α. (G) Mutation at Ser53 reduces phosphorylation of PER2 in vivo. The incorporation of phosphate in vivo into PER2 (1–621)–WT and PER2 (1–621)–S53A was visualized by 33P autoradiography after immunoprecipitation.

Identification of Per2’s CK2-mediated phosphorylation site

To determine the CK2-mediated phosphorylation site(s) within PER2, we used a series of mPER2 deletion mutants. These PER2 mutants were expressed in COS7 cells, immunoprecipitated, and used as substrates for the in vitro kinase assay. The N-terminal half of PER2 (residues 1 to 621), but not the C-terminal half (residues 618 to 1257), was phosphorylated by CK2 (Fig. 5C). A truncation mutant containing PER2 residues 45 to 621 was phosphorylated by CK2, whereas a mutant containing Per2 residues 81 to 621 was not (Fig. 5C). These results suggest that the CK2 phosphorylation site(s) of PER2 is within amino acid residues 45 to 80. This region contains five serine or threonine residues that are conserved among mouse, rat, and human PER2 (Fig. 5D). Using a series of mutants of mPER2 (1–621) in which these serine or threonine residues were substituted with alanine, we found that substitution at Ser53 specifically abolished phosphorylation in the CK2 kinase assay (Fig. 5D). DMAT inhibited CK2-mediated phosphorylation of PER2 (1–621) in a dose-dependent manner (Fig. 5E). In addition, the CK2 holoenzyme, like CK2α, was able to phosphorylate PER2 (1–621) (Fig. 5E).

Ser53 of PER2 is followed by a conserved stretch of acidic amino acids (Fig. 5D) that constitute a consensus sequence for CK2-mediated phosphorylation. Because acidic residue–directed phosphorylation is a feature of CKIε- and glycogen synthase kinase–3β (GSK-3β)–mediated phosphorylation as well, we investigated the possibility that CKIε and GSK-3β also phosphorylate PER2 at Ser53. In vitro kinase assays showed that CKIε and GSK-3β phosphorylated both PER2 (1–621)–WT (wild type) and Per2 (1–621)–S53A to essentially the same extent, whereas CK2α phosphorylated PER2 (1–621)–WT much more efficiently than PER2 (1–621)–S53A (Fig. 5F). Collectively, these results suggest that Ser53 of mouse PER2 is the major CK2 phosphorylation site. We next examined whether Ser53 of PER2 was phosphorylated in living cells. NIH3T3 cells were transfected with Myc-tagged PER2 (1–621)–WT or Myc-tagged PER2 (1–621)–S53A and labeled with [33P]orthophosphate. Myc-tagged PER2 (1–621)–WT or Myc-tagged PER2 (1–621)–S53A were then immunoprecipitated with antibody to Myc (anti-Myc), and the precipitates were analyzed by SDS–polyacrylamide gel electrophoresis (SDS-PAGE) and 33P autoradiography. Phosphorylation of PER2 (1–621)–S53A was markedly lower than that of PER2 (1–621)–WT (Fig. 5G), suggesting that PER2 Ser53 is indeed phosphorylated in mammalian cells.

Effect of CK2 phosphorylation site mutation on PER2 stability

To determine whether CK2-mediated phosphorylation of PER2 at Ser53 regulates PER2 degradation, we made a full-length PER2-S53A mutant and examined the effects of CK2 and CKIε on PER2-S53A stability. Coexpression of CK2 enhanced CKIε-dependent degradation of PER2-S53A (Fig. 6A) to almost the same extent as it did that of PER2-WT (see Fig. 4A). This indicates that Ser53 is dispensable for the cooperative action of CKIε and CK2. In contrast, compared to PER2-WT, the PER2-S53A mutant resisted calyculin A–induced degradation (Fig. 6B). This indicates that phosphorylation of PER2 at Ser53 could modulate PER2 stability. We used the N-terminal half of PER2 (residues 1 to 621), which lacks the CKIε-binding region (23), to determine whether calyculin A–induced degradation of PER2 depends on CKIε. Indeed, PER2 (1–621), in contrast to full-length PER2, was unable to associate with CKIε in cultured cells (fig. S4) and, unlike full-length PER2, did not undergo CKIε-dependent degradation (Fig. 6C). However, this truncated form of PER2 still underwent degradation after calyculin A treatment of cells, and the S53A mutant form of PER2 (1–621) resisted calyculin A–induced degradation (Fig. 6D). Thus, phosphorylation of PER2 at Ser53 affects PER2 stability independent of CKIε, although this effect is manifested only under specific conditions (calyculin A treatment). Together, these results suggest that CK2 promotes PER2 degradation through two mechanisms: One depends on PER2 phosphorylation at Ser53, and the other is independent of it (Fig. 6E). The latter involves CKIε as a critical regulator of PER2 degradation.

Fig. 6

The CK2 site mutation in PER2 suppresses its degradation. (A) Cells transfected with the indicated plasmids were treated with cycloheximide (t = 0). Cell extracts were analyzed by immunoblotting with antibodies to the indicated proteins. The graph shows quantified PER2 abundance. The values were normalized to α-tubulin and the values at t = 0 were set to 1. Means ± SEM (n = 3) are shown. *P < 0.01. Two independent experiments gave similar results. (B) Cells transfected with Myc-tagged PER2-WT or PER2-S53A were pretreated with cycloheximide and treated with calyculin A (t = 0). Cell extracts were analyzed as in (A). The graph shows the relative PER2 abundance at t = 60 min where the values at t = 0 were set to 1. Mean ± SEM (n = 3) are shown. *P < 0.05. (C) Cells transfected with the indicated plasmids were treated with cycloheximide (t = 0). Cell extracts were analyzed as in (A). Two independent experiments gave similar results. (D) Cells transfected with the indicated plasmids were pretreated with cycloheximide and treated with calyculin A (t = 0). Cell extracts were analyzed as in (A). The graph shows the relative PER2 abundance at t = 60 min where the values at t = 0 were set to 1. Mean ± SEM (n = 3) are shown. *P < 0.05. (E) Schematic model of the role of CK2 in the regulation of PER2 degradation. CK2 promotes PER2 degradation through two independent mechanisms, phosphorylation of PER2 at Ser53 and collaboration with CKIε.

The CK2 phosphorylation site mutant of PER2 impairs circadian gene expression

We next examined whether the PER2-S53A mutant affected the mammalian circadian clock. Bioluminescence recording from NIH3T3 cells transfected with plasmids encoding the Bmal1-Luc reporter and either PER2-WT or PER2-S53A showed that PER2-S53A dampened the amplitude of circadian oscillation more potently than did PER2-WT (Fig. 7). This difference was not due to differences in transfection efficiency because the steady-state abundance of PER2-WT and S53A was comparable (fig. S5). Intriguingly, forced expression of PER2-S53A slightly shortened rhythm period, indicating that the CK2 phosphorylation site mutant of PER2 (PER2-S53A) has the effect opposite to that of overall inhibition of CK2 activity on rhythm period. This might reflect two distinct mechanisms by which CK2 promotes PER2 degradation.

Fig. 7

The CK2 site mutant of PER2 impairs circadian gene expression. Bioluminescence from cells cotransfected with the Bmal1-Luc reporter plasmid and either Myc-tagged PER2-WT or PER2-S53A was recorded. Representative raw data and detrended, amplitude-compensated data are shown at the top and the bottom, respectively. Three independent experiments gave similar results. The bottom table shows the calculated periods and amplitudes of bioluminescence rhythms. Means ± SD are shown for each trace (n = 6). P values as compared to the vector are indicated.

Discussion

Here we show that CK2 is involved in control of the mammalian circadian clock and that it regulates PER2 stability. PER2 is an essential component of the mammalian circadian clock, and its posttranslational regulations such as protein degradation and nuclear translocation are critical to generating proper rhythms of the correct length and amplitude. Accumulating evidence suggests that CKIε/δ plays a central role in regulating PER2 degradation (2224); however, a PER2 site(s) whose phosphorylation by CKI promotes PER2 degradation has not yet been identified. Indeed, phosphorylation of human PER2 at Ser662 (Ser659 of mouse PER2), the best-characterized CKIε/δ phosphorylation site on PER2, increases PER2 abundance (21), and the half-life of the PER2-S662G mutant is shorter than that of PER2-WT likely because of enhanced nuclear clearance of the mutant protein (20). Thus, the collaborative effect of CKIε and CK2 on PER2 degradation, which we show here, sheds light on the mechanisms regulating the rapid degradation of PER2. We also identified Ser53 as the PER2 target site for CK2-mediated phosphorylation. Because Ser53 of mPER2 is not conserved in the other two mammalian PER homologs, regulation by CK2-mediated phosphorylation should be specific to PER2. In Drosophila, CK2 is also thought to be involved in PER regulation. The timing of PER translocation into the nucleus is delayed in CK2 mutant flies, and CK2 can modulate PER’s transcriptional repression activity (12, 13, 25). We found that Ser53 of mPER2 is not conserved in Drosophila PER. Moreover, CK2-mediated phosphorylation sites within Drosophila PER are not conserved in mammalian Per2 (26). Therefore, although CK2 phosphorylates PER in both Drosophila and mammals, the role of the phosphorylation in these organisms differs.

The functional importance on mPER2 stability of its phosphorylation on Ser53 is still not completely understood. For several reasons, it seems likely that the major function of CK2 with regard to PER2 degradation is to enhance CKIε-dependent degradation of PER2, rather than to promote degradation through phosphorylation of Ser53. First, forced expression of CK2 alone did not significantly affect PER2 stability. Second, although PER2-S53A shows increased stability in the presence of calyculin A, in the absence of calyculin A, the steady-state abundance of PER2-S53A is comparable to that of PER2-WT. Third, the effect of S53A mutation on protein stability in the presence of calyculin A is relatively modest (Fig. 6, B and D). These results suggest that Ser53 is quantitatively less important in regulating PER2 abundance than the collaborative effect of CKIε and CK2. Moreover, forced expression of PER2-S53A does not lengthen rhythm period but slightly shortens it, whereas overall inhibition of CK2 activity, similar to inhibition of CKIε/δ activity, does lengthen rhythm period. In this context, because it seems that phosphorylation of PER2 at Ser53 and the collaboration between CK2 and CKIε independently regulate PER2 stability, we speculate that acceleration of PER2 degradation could have distinct effects on the rhythm period in a stage-dependent manner. In the PER2 “accumulating stage,” when PER2 abundance is increasing, inhibition of PER2 degradation may promote PER2 accumulation and thereby shorten the rhythm period (Fig. 8). In contrast, in the PER2 “working stage,” when PER2 functions as a transcriptional regulator in the nucleus, inhibition of PER2 degradation may prolong PER2 activity and thereby lengthen the period. It is possible that CK2-mediated phosphorylation of PER2 at Ser53 moderately suppresses the accumulation of PER2 in the accumulating stage and thereby delays the nuclear translocation of PER2. On the other hand, collaboration between CK2 and CKIε could determine the timing of PER2 clearance from the nucleus by stimulating PER2 degradation. Given that phosphorylation of CDC37 on the CK2 site is low from CT (circadian time) 6 to 14 and reaches its peak at CT18 to 22 (fig. S3A), when PER2 abundance rapidly decreases in liver (27), it is likely that CK2 makes a relatively small contribution to the delay in PER2 accumulation but plays a more crucial role in triggering PER2 degradation to close the transcriptional feedback loop. It is also possible, however, that CK2 and phosphorylation of PER2 at Ser53 might have a role other than regulation of PER2 degradation in the mammalian circadian clockwork.

Fig. 8

A hypothetical model of how regulation of PER2 degradation could regulate circadian period. PER2 accumulates and functions in transcriptional regulation. Then PER2 undergoes CKIε/δ–dependent rapid degradation to close the transcriptional feedback loop. Impaired PER2 degradation may cause changes in period length in a stage-dependent manner.

Our results here show that CK2 promotes PER2 degradation by enhancing CKIε-dependent PER2 degradation in a manner independent of the phosphorylation of PER2 at Ser53. Because the kinase activity of CK2 is also important for its collaboration with CKIε in PER2 degradation, CK2 phosphorylation of some target proteins should play a key role in the collaborative effect of CKIε and CK2. It is possible that CK2 directly phosphorylates CKIε and thereby regulates CKIε activity. It is also possible that CK2 phosphorylates PER2 at a site other than Ser53 in a CKIε-dependent manner. In our preliminary experiment, CKIε was coimmunoprecipitated with CK2β only in the presence of PER2 (fig. S6), suggesting that PER2, CKIε, and CK2β can form a ternary complex in cells. Such a ternary complex could enable these two kinases to function coordinately. Thus, although the underlying mechanism is not yet clear, our results show that CK2 regulates rapid degradation of PER2 through a collaboration with CKIε. Further studies will be needed to elucidate the detailed mechanism underlying CK2-mediated enhancement of CKIε-dependent PER2 degradation.

In conclusion, this study has revealed an essential role for CK2 in fine-tuning the mammalian circadian clock. This study underscores the importance of CK2 in the regulation of circadian oscillators in diverse organisms ranging from fungi to mammals and suggests that CK2 is a common regulator of evolutionarily diversified circadian molecular oscillators. While this paper was undergoing final revision, two other papers appeared indicating that CK2 has a role in regulating mammalian circadian rhythms (28, 29).

Materials and Methods

Animals

C57BL/6 wild-type male mice were kept under a 12-hour light:12-hour dark cycle for 2 weeks before the start of experiments. Tissues were harvested on the first day of constant darkness at the indicated CT, immediately frozen in liquid nitrogen, and stored at −80°C before processing. Animal care was in accordance with institutional guidelines. Circadian time is the time under constant dark and free-running conditions, and CT12 corresponds to the time at which animals start their locomotor activity.

Cell culture and synchronization

NIH3T3 cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM) containing 10% calf serum (CS). Cells were seeded in 35-mm dishes, grown to confluence, and then serum-starved in DMEM containing 1% CS for 24 hours. The medium was changed to DMEM containing 50% CS with the indicated concentration of DMAT (Calbiochem) dissolved in DMSO or DMSO alone and cells were incubated for 1 hour, after which they were washed and maintained in medium containing 1% CS with DMAT or DMSO. Transfection was performed 1 day before cells reached confluence. COS7 cells were cultured in DMEM containing 10% fetal bovine serum (FBS). The cells were seeded in 60-mm dishes and incubated for 24 hours before transfection.

Materials

Mouse monoclonal anti-Myc [9E10; for immunoprecipitation (IP)], rabbit polyclonal anti-Myc [A-14; for immunoblotting (IB)], and mouse monoclonal anti-CDC37 (E4) were purchased from Santa Cruz Biotechnology. Mouse monoclonal anti-HA was from Covance. Mouse monoclonal anti-FLAG (M2; for IP), rabbit polyclonal anti-FLAG (for IB), and mouse monoclonal anti–α-tubulin (DM1A) were from Sigma. Rabbit anti-PER2 was from Alpha Diagnostic International. Rabbit polyclonal anti-CK2α and anti-CK2β were described previously (19). Rabbit polyclonal antibody directed against the Ser13-phosphorylated form of CDC37 (anti–[pSer13]-Cdc37) was described previously (30). mPer1, mPer2, mPer3, mCry1, mCry2, and mBmal1, mNpas2, mRev-erbα, mDbp, and mE4bp4 complementary DNAs (cDNAs) were subcloned into pcDNA3 (Myc). hRorα1 and hRorα4 cDNAs subcloned into pcDNA3 (Myc) and Bmal1-Luc reporter plasmid were described previously (15). hCK2α and hCK2β subcloned into p3xFLAG-CMV7.1 were described previously (19). Recombinant CK2α, CK2α2β2 holoenzyme, CDC37-WT, and CDC37-S13A proteins were described previously (19, 30). Recombinant GSK-3β was purchased from Sigma.

Preparation of liver extracts

Mouse liver tissues were homogenized in ice-cold incubation buffer [50 mM tris-Cl, 50 mM NaCl, 100 mM NaF, 10% glycerol, 2 mM EDTA, 2 mM sodium orthovanadate, 10 mM sodium pyrophosphate, 1 mM dithiothreitol (DTT), 1% Nonidet P-40 (pH 8.0); supplemented with aprotinin (0.1 trypsin inhibitor units/ml), leupeptin (10 μg/ml), and 1 mM phenylmethylsulfonyl fluoride], and the lysates were centrifuged at 12,000g for 20 min. The supernatants were used as liver extracts and analyzed by immunoblotting.

Bioluminescence recordings

NIH3T3 cells grown to confluence were treated with dexamethasone for 2 hours and then washed with DMEM containing 1% CS and 0.1 mM luciferin. SCN slices were excised from mPer2::Luciferase knock-in mice and placed on a culture membrane (Millicell-CM, PICM030-50; Millipore) in a covered and sealed petri dish. Bioluminescence was measured and integrated for 1 min at intervals of 15 min, with a photomultiplier tube (LM2400, Hamamatsu), according to the protocol described previously (18).

Analysis of bioluminescence data

Quantitative and statistical analyses of circadian parameters of the bioluminescence rhythms were performed basically as previously described (31). In brief, the raw data were detrended by subtracting 24-hour running means (moving average) from data. Next, to correct for damping effects, each value of detrended data was divided by the standard deviation within each corresponding 24-hour moving window. The obtained data that had been corrected for both the baseline drifting and damping effects were then analyzed for period by a linear least-squares estimation, according to the following equation:

y(t)=αsin(2πtτϕ)

where α = amplitude, t = time, τ = period, and ϕ = phase. This procedure gave period and phase information of the rhythms. Amplitude from the time range of 36 to 60 hours was calculated by a linear least-squares estimation of detrended data according to the obtained period and phase information. To compensate for the baseline variance of luminescence intensity among different data sets, the calculated amplitude was further normalized by dividing the calculated amplitude by the average of luminescence intensity between 36 and 60 hours in the raw data.

mRNA analysis

Preparation of RNA samples and real-time quantitative polymerase chain reaction (PCR) analysis were performed as described previously (32). The primers for the PCR analysis were as follows: mBmal1, 5′-GTAGTCCCAGTAACGATGAGGCAG-3′ and 5′-CTACAGCGGCCATGGCAAGTCACTA-3′; mPer2, 5′-CAGTGATGCCAAGTTTGTGGAGTTC-3′ and 5′-TGAGTGAAAGAATCTAAGCCGCTGC-3′; Myc-mPer2-WT and S53A, 5′-GGACTTGAATGAAGGATCCG-3′ and 5′-AGCAGTTCTCGTTTCC-3′; 18S rRNA, 5′-CGCCGCTAGAGGTGAAATTC-3′ and 5′-CGAACCTCCGACTTTCGTTCT-3′. The rhythm amplitude was calculated as the ratio of the mean value of the second peak to that of the first trough.

Mutagenesis

The mutants used were constructed by PCR-based mutagenesis. PCR was performed with KOD plus DNA polymerase (Toyobo). Dpn I restriction enzyme–treated PCR products were transformed into Escherichia coli. Positive clones were picked up, and mutagenesis was verified by sequencing.

Transfections

Transfection was performed by using LipofectAmine Plus (Invitrogen) according to the manufacturer’s instructions. For bioluminescence recordings of the Bmal1-Luc reporter activity, NIH3T3 cells were transfected with 100 ng of the Bmal1-Luc reporter, 300 ng each of CK2α, CK2β, and CK2α-K68A, and 10 or 40 ng of Per2-WT or Per2-S53A. For monitoring of PER2 abundance, NIH3T3 cells were transfected with 700 ng of Myc-tagged Per2-WT or Per2-S53A. COS7 cells were transfected with 200 ng of Myc-tagged Per2-WT or Per2-S53A, 100 ng of HA-tagged CKIε, and 200 ng of FLAG-tagged CK2α, α-K68A, and β. The total amount of DNA per dish was adjusted to 1 μg by adding pcDNA3 empty vector. For immunoprecipitation assays, COS7 cells were transfected with 700 ng each of the indicated plasmids. For the in vitro kinase assay, COS7 cells were transfected with 1.4 μg of the indicated plasmids. The total amount of DNA per dish was adjusted to 2.1 μg by adding pcDNA3 empty vector. For in vivo cell labeling, NIH3T3 cells were transfected with 4 μg of the indicated plasmids.

Coimmunoprecipitation assay

Thirty-six hours after transfection, COS7 cells were scraped into incubation buffer. Cell lysates were centrifuged at 12,000g for 20 min. The supernatant was then mixed and incubated with antibody and protein G–Sepharose beads (GE Healthcare) for 2 hours at 4°C. The beads were then washed twice with the incubation buffer. After resolution by SDS-PAGE, the precipitates were analyzed by immunoblotting.

In vitro kinase assay

Each of the Myc-tagged clock proteins was expressed in COS7 cells, and the cells were lysed in the incubation buffer. Myc-tagged proteins were immunoprecipitated with anti-Myc (9E10) and protein G–Sepharose beads for 2 hours at 4°C. The beads were washed twice with the incubation buffer and once with a kinase reaction buffer [50 mM tris-Cl (pH 7.5), 200 mM NaCl, 10 mM MgCl2, 15 mM β-glycerophosphate, 2 mM EGTA, 1 mM DTT, 50 μM adenosine triphosphate (ATP)]. The washed beads were mixed with a kinase reaction buffer supplemented with 0.1 MBq of [γ-32P]ATP and 200 ng of CK2α or CK2α2β2 holoenzyme and incubated for 15 min at 30°C. The reaction was stopped by the addition of Laemmli sample buffer. After resolution by SDS-PAGE, substrate phosphorylation was detected with a Bio-Rad phosphorimager. FLAG-CKIε was expressed in COS7 cells and immunopurified with anti-FLAG M2 affinity gel (Sigma). A kinase reaction buffer for CKIε contains 50 mM tris-Cl (pH 7.5), 10 mM MgCl2, 1 mM DTT, and 50 μM ATP. GSK-3β was also used according to the manufacturer’s instructions. For in vitro phosphorylation of CDC37, NIH3T3 cells were synchronized by serum stimulation and lysed in the incubation buffer at the indicated time points. Cell lysates were centrifuged at 12,000g for 20 min, and the supernatant was used as cell extracts. The in vitro kinase assay was performed by incubating 1 μg of purified CDC37-WT or CDC37-SA with 10 μg of the cell extracts in a kinase reaction buffer in the presence or absence of 0.1 MBq of [γ-32P]ATP for 30 min at 30°C. The reaction was stopped by the addition of Laemmli sample buffer. After resolution by SDS-PAGE, phosphorylation of CDC37 was analyzed with a phosphorimager or immunoblotting.

In vitro translation

Myc-tagged PER2 was synthesized in vitro by using TNT T7 Coupled Reticulocyte Lysate System (Promega) according to the manufacturer’s instructions. The reaction product was diluted in incubation buffer and immunoprecipitated with anti-Myc (9E10) as described above. The precipitate was used in the in vitro kinase assay.

Per2 degradation assay

Thirty-six hours after transfection, NIH3T3 cells were treated with cycloheximide (20 μg/ml) and either DMSO or DMAT (40 μM) for 30 min. Then calyculin A (Calbiochem) was added to the medium at a final concentration of 5 nM. The cells were frozen in liquid nitrogen at the indicated time points and stored at −80°C before processing. Cell extracts were resolved by SDS-PAGE and analyzed by immunoblotting.

In vivo cell labeling

Twenty-four hours after transfection of NIH3T3 cells, culture medium was replaced with phosphate-free DMEM containing 10% FBS and incubated for 4 hours. The medium was again replaced with fresh phosphate-free medium, and 1.85 MBq of [33P]orthophosphate (Perkin-Elmer) per 100-mm dish was added. Cells were incubated at 37°C for 6 hours with occasional swirling. 33P-labeled cells were lysed in incubation buffer and cell extracts were immunoprecipitated with anti-Myc.

Statistical analysis

Data comparisons were made by using the two-tailed Student’s t test for repeated measures.

Acknowledgments

We thank J. S. Takahashi for Per2::Luciferase knock-in mice and the members of our laboratory for their helpful suggestion. This work was supported by grants-in-aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan.

Supplementary Materials

www.sciencesignaling.org/cgi/content/full/2/73/ra26/DC1

Fig. S1. Circadian oscillation of the Bmal1-Luc reporter activity is affected by DRB.

Fig. S2. CK2 inhibition dampens circadian oscillation of Per2 mRNA expression.

Fig. S3. Circadian oscillation of phosphorylation of CDC37 at a CK2 site.

Fig. S4. The N-terminal half of PER2 does not bind to CKIε.

Fig. S5. Expression of Per2-WT and S53A are comparable.

Fig. S6. CKIε is coimmunoprecipitated with CK2β in the presence of PER2.

References and Notes

View Abstract

Navigate This Article