Light-induced phase-delay of the chicken pineal circadian clock is associated with the induction of cE4bp4, a potential transcriptional repressor of cPer2 gene

Abstract

The chicken pineal gland contains the autonomous circadian oscillator together with the photic-input pathway. We searched for chicken pineal genes that are induced by light in a time-of-day-dependent manner, and isolated chicken homolog of bZIP transcription factor E4bp4 (cE4bp4) showing high similarity to vrille, one of the Drosophila clock genes. cE4bp4 was expressed rhythmically in the pineal gland with a peak at very early (subjective) night under both 12-h light/12-h dark cycle and constant dark conditions, and the phase was nearly opposite to the expression rhythm of cPer2, a chicken pineal clock gene. Luciferase reporter gene assays showed that cE4BP4 represses cPer2 promoter through a E4BP4-recognition sequence present in the 5′-flanking region, indicating that cE4BP4 can down-regulate the chick pineal cPer2 expression. In vivo light-perturbation studies showed that the prolongation of the light period to early subjective night maintained the high level expression of the pineal cE4bp4, and presumably as a consequence delayed the onset of the induction of the pineal cPer2 expression in the next morning. These light-dependent changes in the mRNA levels of the pineal cE4bp4 and cPer2 were followed by a phase-delay of the subsequent cycles of cE4bp4/cPer2 expression, suggesting that cE4BP4 plays an important role in the phase-delaying process as a light-dependent suppressor of cPer2 gene.

Daily rhythms of biological activities observed in a variety of organisms from bacteria to humans are driven by endogenous oscillators called circadian clocks (1). The circadian clock autonomously oscillates even in the absence of external time cues, although the period length differs slightly from 24 h. Under natural conditions, the circadian clock is entrained (synchronized) to the 24-h day by environmental time cues, most commonly by light (2), but the molecular nature of the light-entrainment of the clock is not fully understood. In vertebrates, the central circadian clocks are located in several neuronal tissues such as the suprachiasmatic nucleus (SCN), the retina, and the pineal gland (3–7), and the oscillation seems to be generated by cell-autonomous mechanisms (8, 9). Among clock-containing cells, the chick pinealocyte is unique in that it retains not only the circadian oscillator but also intrinsic phototransduction pathway for the light-entrainment within a single cell (3, 4, 9, 10–14), providing a prominent model for the study of the light-entrainment mechanism, especially at the cellular and molecular levels (15).

Recent molecular and genetic analyses of clock genes have demonstrated that the transcription/translation-based autoregulatory feedback loop plays a central and common role in generating the circadian rhythmicity among various organisms (16, 17). In the case of Drosophila, clock genes such as period, timeless, cycle (bmal), clock, cryptochrome, double-time, and vrille were identified as necessary for normal circadian function (18). Although the role of vrille remains to be elucidated, the products of many clock genes seem to form interlocked negative feedback loops for stable oscillation of the Drosophila clock (19). Mammalian homologs of all of the Drosophila clock genes except vrille have been studied in relation to their roles in the circadian clock system. In mice, mCLOCK:mBMAL1 heteromer is suggested to drive transcription of three period genes (designated mPer1, mPer2, and mPer3 in the mouse), two cryptochrome genes (mCry1 and mCry2), and output (clock-controlled) genes such as vasopressin gene (20–22). mPER and mCRY proteins thus induced in turn inhibit the mCLOCK:mBMAL1-mediated transcription, and hence this feedback cycle generates rhythmic expression of several clock genes and output genes. The molecular framework of the circadian oscillator established in the mouse SCN appears to be applicable to that in the chicken pineal gland, in which the mRNA levels of chicken homologs, cPer2, cClock, cBmal1, and cBmal2 exhibit diurnal fluctuations (ref. 23; T.O., K. Yamamoto, K. Okano, T. Hirota, T. Kasahara, M. Sasaki, Y. Takanaka, and Y.F., unpublished results). Functionally, both cCLOCK:cBMAL1 and cCLOCK:cBMAL2 complexes up-regulate cPer2 transcription, which is subject to down-regulation by cPER2 protein (T.O., K. Yamamoto, K. Okano, T. Hirota, T. Kasahara, M. Sasaki, Y. Takanaka, and Y.F., unpublished results), suggesting that these chicken clock genes form the autoregulatory feedback loop as well. Now, the chicken pineal clock system provides a new avenue to understanding of a molecular link between the oscillator and the intrinsic photic-input pathway.

As a general feature of the light-entrainment of the clock, a light stimulus given at early and late subjective night induces the phase-delay and advance of the clock, respectively, whereas that at (early) subjective day gives little effect on the clock phase (2, 12). Because the product levels of many clock genes likely define, rather than simply reflect, the phase of the clock, the light-induced phase-shift is closely associated with light-dependent changes in the mRNA or protein levels of a subset of clock genes (16). In the case of Drosophila, the clock phase is reset by light-induced degradation of dTIM protein (24–27). On the other hand, in the mammalian SCN, the mRNA levels of Per1 and Per2 genes rapidly increase in response to a brief light stimulus that induces a phase-shift of the clock (28–32). In contrast to the short-light-pulse-induced phase-shift, the phase-shift induced by a nonphotic stimulus such as a wheel-running action is accompanied by a rapid decrease in the mRNA levels of the SCN Per1 and Per2 genes (33, 34). Thus, the changes (rise and fall) of the mRNA levels of Per genes seem to be important for the phase-shift of the SCN clock (30, 31). However, little is known about a molecular system that induces the changes of the expression levels of Per genes in response to the external time cues.

In the present study, we searched for chick pineal genes that are induced by light in a time-of-day-dependent manner, and isolated chicken homolog of bZIP transcription factor E4bp4 (cE4bp4), which shows high similarity to vrille, one of the Drosophila clock genes (35). Here, we show several lines of evidence for an important role of cE4BP4 serving as a transcriptional repressor of cPer2 gene in the chick pineal clock system. Noticeably, the prolonged light period, which induced a phase-delay of the chick pineal clock, up-regulates cE4bp4 expression and delayed the onset of the cPer2 induction in the next morning, suggesting that cE4BP4 contributes to the phase-delaying process as a light-responsive repressor of cPer2 gene.

Materials and Methods

Animals.

Animals were treated in accordance with the guidelines of the University of Tokyo. Newly hatched chicks were purchased from local suppliers, and housed under various light/dark conditions with a constant light intensity of ≈300 lux at the level of chicks. Pineal glands were isolated from killed chicks, frozen in liquid nitrogen, and kept at −80°C until subsequent analyses. All of the procedures during the dark-period were performed under dim red light (>640 nm).

Differential Display Analysis.

One-day-old chicks were maintained in light/dark (LD) cycles for 7 days and transferred to constant darkness (DD) on day 8, when animals were exposed to light for 1 h from circadian time (CT) 0, CT14, or CT20, and their pineal glands were isolated at CT1, CT15, or CT21, respectively (40 chicks each). On the other hand, control animals were kept in the dark on day 8, and their pineal glands were isolated at CT1.5, CT15.5, or CT21.5 (40 chicks each). Total RNA extracted from each pool of the isolated pineal glands was reverse transcribed by using ThermoScript (Life Technologies, Grand Island, NY) with an anchored oligo(dT) primer at 55°C for 60 min. Briefly, a fixed amount (1.5 μg) of the total RNA preparation was incubated in a reaction mixture (30 μl) composed of 50 mM Tris acetate (pH 8.4), 75 mM potassium acetate, 8 mM magnesium acetate, 5 mM DTT, 1 mM each of the dNTPs, 2.5 μM anchored oligo(dT) primer (RNAimage kit, GenHunter, Nashville, TN), 60 units RNase Inhibitor (Life Technologies), and 22.5 units ThermoScript (Life Technologies). One-hundredth of this reaction mixture was subjected to PCR in a reaction mixture (10 μl) composed of 10 mM Tris⋅HCl (pH 8.3), 50 mM KCl, 1.5 mM MgCl₂, 200 μM each of the dNTPs, 1 μM anchored oligo(dT) primer, 1 μM H-AP primer (RNAimage kit, GenHunter), and 1 unit AmpliTaq (Applied Biosystems) under the following cycle conditions: 27 cycles of 94°C (30 s), 34°C (2 min), 72°C (30 s) for amplification, followed by 72°C (7 min) for final extension. The PCR products were subjected to 7.5% native PAGE, stained with SYBR Green I (Molecular Probes), and quantified by FLA-2000 bioimage analyzer (Fuji Film, Tokyo, Japan).

Cloning of cE4bp4 cDNA.

The PCR products selected in the differential display analysis were cut out from the gel and eluted by boiling in 50 μl distilled water at 95°C for 5 min. An aliquot (2 μl) of the eluted DNA solution was subjected to the secondary PCR under the same PCR conditions as those in the initial analysis. The reamplified DNA fragment was subcloned into pCR2.1TOPO vector (Invitrogen) and sequenced.

A cDNA clone including the entire coding region of cE4bp4 gene was obtained by RACE (rapid amplification of cDNA ends), and the coding sequence of cE4bp4 (GenBank accession no. AF335427) was amplified by PCR with a pair of primers, 5′-TGTGTGTGGATTTAAAACTACTG-3′ and 5′-ACTTTACTGTAGTACTCAGGTC-3′, which correspond to the 5′- and 3′-untranslated regions of cE4bp4 gene, respectively. Possible PCR errors were identified and eliminated by sequencing a total of seven clones obtained from three independent PCRs.

Quantitative Reverse Transcription (RT)-PCR Analysis.

Quantitative RT-PCR analyses were performed as previously described (36). PCR primers used were as follows: for cE4bp4, 5′-CCTTTCTCAGTTCAGGTGAC-3′ (forward) and 5′-TGAAATGACATCATGAGTCCAG-3′ (reverse); for cPer2, 5′-GGAAGTCCTTGCAGTGCATAC-3′ (forward) and 5′-ACAGGAAGCGGATATGCAG-3′ (reverse); for chicken TATA-box binding protein (cTbp; GenBank accession no. D83135), 5′-GTCGAATATAATCCCAAGCG-3′ (forward) and 5′-TCTGCTCGAACTTTAGCACC-3′ (reverse). The optimal cycle numbers for quantitative analyses were 19 for cE4bp4, 19 for cPer2, and 21 for cTbp under our experimental conditions.

Transcriptional Assay.

LMH chicken hepatoma cells were grown in Waymouth's MB752/1 medium (Life Technologies) supplemented with 10% FBS, and plated at 5 × 10⁶ cells per well in six-well plates. On the next day, cells in each well were transfected by using Lipofectamine plus (Life Technologies) with (i) indicated amount of expression plasmid containing the coding sequence of cE4bp4 in pcDNA3.1/V5/His (Invitrogen) without introduction of any tags, (ii) 300 ng of firefly luciferase reporter (derivative of pGL3-Basic or pGL3-Promoter), and (iii) 1 ng of Renilla luciferase reporter, pRL-CMV (Promega), as an internal control. The total amount of expression plasmid DNA added to each well was adjusted to 2.0 μg by mixing pcDNA3.1/V5/His empty vector. Two days after transfection, the extract prepared from harvested cells was subjected to luminometry-based dual-luciferase assay (Promega), and the firefly luciferase activity was normalized by Renilla luciferase activity for each extract.

Reporter plasmids used were as follows: (i) cPer2 promoter-luc: A 674-bp genomic DNA fragment of the 5′-flanking region of cPer2 gene (GenBank accession no. AF335428; −607 to +67; +1 indicates a putative transcription initiation site estimated by 5′-RACE) was obtained by using LA PCR in vitro Cloning Kit (Takara Shuzo, Kyoto, Japan) according to the manufacturer's protocol. This fragment was subcloned into pGL3-Basic vector (Promega). (ii) mut. distal site-luc: The “distal site”-containing DNA fragment (−340 to −309; Bsu36I-NotI-digested fragment) of cPer2 promoter-luc construct was substituted by a synthesized fragment in which the distal site (5′-GTGATGTAAC-3′; −334 to −325) was mutated to 5′-GGAGACGCTC-3′. (iii) mut. proximal site-luc: The “proximal site”-containing DNA fragment (−102 to −52; PmlI-AscI-digested fragment) of cPer2 promoter-luc was substituted by a synthesized fragment in which the proximal site (5′-CTTATGTAAA-3′; −96 to −87) was mutated to 5′-CGAGACGCTA-3′. (iv) Distal site × 3-luc: Three copies of a 20-bp piece, the distal binding site with its flanking sequences (5′-GAGGCGTGATGTAACCTCTG-3′; −339 to −320) were linked in tandem, and the 60-bp fragment was inserted into an SV40-driven luciferase reporter (pGL3-Promoter, Promega). (v) mut. distal site × 3-luc: The same as iv, except that every E4BP4-binding site was mutated to 5′-GGAGACGCTC-3′.

Results

Identification of a Novel Light-Controlled Gene.

To approach the molecular basis of the phase-dependent phase-shifting mechanism in the chicken pineal clock system, we used a differential display-based cloning of pineal genes that are induced by light in a CT-dependent manner. We compared the effects of 1-h light stimulus given at CT0, CT14, and CT20 on the band intensities of PCR products (Fig. 1). Screening of ≈6,000 bands of the PCR products revealed that most of the light-induced products responded to light at every circadian time point tested. A phase-dependent light-induction was observed only in a few products, among which one (indicated by an arrow in Fig. 1) showed a remarkable light-induction at CT14, a time when a light stimulus induces a phase-delay of the chick pineal clock (12), whereas far less light-induction of the product was observed at both CT0 and CT20 (Fig. 1). The cloning and sequencing of the product (393-bp cDNA fragment) revealed that it encoded a peptide (60 aa) showing high similarity (70% identity) to the N-terminal region (Met-1–Cys-60) of human E4BP4 (hE4BP4), which was previously isolated as a transcriptional repressor recognizing the adenovirus E4 promoter region (37). The full-length coding sequence obtained by 3′-RACE encoded a protein of 458 aa residues 83% identical to hE4BP4 (Fig. 2A). This chicken homolog of E4bp4 (cE4bp4) shows the highest similarity to vrille (Vri, Fig. 2B) among all of the coding sequences of Drosophila genome, and the DNA-binding domain (basic region) of E4BP4 shows high similarity not only with VRI but also with the PAR subfamily of bZIP transcription factors, including albumin gene D-site binding protein (DBP, Fig. 2B).

Figure 1.

A time-of-day-specific light-induced gene identified by differential display analyses. The pineal glands were isolated from chicks exposed to a 1-h light pulse [L(+)] or kept in the dark [L(−)] at indicated CT points of a day. Duplicate amplification reactions were performed with a pair of H-T11C primer (5′-AAGCTTTTTTTTTTC-3′) and H-AP21 primer (5′-AAGCTTTCTCTGG-3′, both primers from RNAimage kit, GenHunter), and the amplified products were loaded on the gel side by side. The arrow indicates the position of the amplified DNA fragment derived from cE4bp4 transcript.

Figure 2.

Amino acid sequence of chicken E4BP4. (A) The deduced amino acid sequence of chicken E4BP4 is aligned with that of human E4BP4 (37), and the conserved residues are shown with white characters on black backgrounds. Asterisks and a horizontal line above the sequence represent a heptad leucine repeat and a basic domain, respectively. (B) Comparison of bZIP domain of chicken E4BP4 with other typical bZIP proteins; human E4BP4 (GenBank accession no. U26173), VRI (Y11837), CES-2 (U60979), VBP (U09222), TEF (U44059), HLF (M95585), DBP (U06936), GIANT (X05426), CREB (X14788), cFos (M18043), and cJun (X15547). VBP, TEF, HLF, and DBP are members of PAR bZIP protein family (38, 39). Residues identical to those of chicken E4BP4 are shown with white characters on black backgrounds. The amino acid identity (%) of each bZIP protein to chicken E4BP4 is shown on the right of the sequence.

Light-Dependent and Time-of-Day-Dependent Expression of cE4bp4.

A time course of light-dependent cE4bp4 induction was investigated by quantitative RT-PCR analyses, in which the mRNA levels of the chick pineal cE4bp4 were pursued during and after the 1-h light exposure from CT14 (Fig. 3A). The cE4bp4 mRNA level began to increase within 30 min after the light onset, and at 60 min it reached a peak level ≈2-fold higher than the basal level. After the light was turned off, the cE4bp4 mRNA level began to decrease and reached a level close to that of the control (dark-kept) animals within 3 h after the light offset. The light-dependent induction of cE4bp4 was observed at any circadian time within a day (Fig. 3B), but this is not contradictory to the observation in Fig. 1 because both the basal levels in the dark and the light-induced levels heavily depended on circadian time (Fig. 3B; see also Fig. 1). A light-dependent induction of cE4bp4 was also observed in an in vitro experiment with cultured chick pinealocytes (see Fig. 8, which is published as supplemental data on the PNAS web site, www.pnas.org), indicating that the pineal photoreceptive molecule(s) is responsible for the induction of cE4bp4.

Figure 3.

Light-dependent change in cE4bp4 mRNA levels. One-day-old chicks were maintained in LD cycles for 8 days and then transferred to DD on day 9, when animals were exposed to light for 1 h from each indicated time points [Light (+)] or kept in the dark for control [Light (−)]. (A) Time course of cE4bp4 induction by light given at CT14–15. The pineal glands were isolated at indicated time points, and the relative mRNA levels of cE4bp4 and cTbp in the isolated pineal glands were evaluated by RT-PCR (Upper). The band intensities of the amplified products of cE4bp4 were normalized to those of cTbp, and the value of CT14 (0 min) was set to 1 (Lower). (B) Light-induction of cE4bp4 at various circadian time points. The pineal glands were isolated from light-exposed animals (open bars) or dark-kept animals (filled bars) 1 h after each indicated time point. Relative mRNA levels of cE4bp4 and cTbp were evaluated as described in A, and the value of dark-kept animals at CT22 was set to 1. All of the values are the mean ± SEM from three independent experiments.

The circadian fluctuation of the chick pineal cE4bp4 expression under DD conditions had a peak around the day-night transition phase (between CT11 and CT15; Figs. 3B and 4B), and a similar or more pronounced fluctuation with a higher peak/trough ratio was observed in animals maintained in 12-h light/12-h dark (LD) cycles (Fig. 4A). These results indicate that the expression of cE4bp4 is controlled not only by the external light but also by the circadian clock. In both LD and DD, the phase of the rhythmic expression of cE4bp4 was constantly delayed by ≈8 h relative to that of cPer2 (Fig. 4).

Figure 4.

Temporal changes in mRNA levels of cE4bp4 and cPer2 in the chicken pineal gland. One-day-old chicks were maintained in LD cycles for 8 days and then transferred to DD. The pineal glands were isolated at indicated time points of day 7–8 (A, in LD) and day 9 (B, in DD). Relative mRNA levels of cE4bp4, cPer2, and cTbp in the isolated pineal glands were evaluated by RT-PCR (Upper). The band intensities of the amplified products of cE4bp4 and cPer2 were normalized to those of cTbp, and the peak value for each gene was set to 1 (Lower). All of the values are the mean ± SEM from three independent experiments.

cE4BP4 Represses Transcription from cPer2 Promoter.

The rhythmic expression of cE4bp4 with the phase nearly opposite to that of cPer2 (Fig. 4) suggests that cE4BP4 down-regulates cPer2 transcription because of a possible function of cE4BP4 as a transcriptional repressor (37). To test this idea, we isolated a 5′-flanking region of cPer2 gene (607 bp) and found two potential cE4BP4-binding sites (Fig. 5A). One, termed “distal site,” at −334 to −325 (+1 indicates a putative transcription initiation site) matched 9 of 10 bp with the consensus hE4BP4-binding sequence (37), and the other, termed “proximal site,” at −96 to −87 matched 8 of 10 bp (Fig. 5A). To examine whether cE4BP4 regulates cPer2 transcription through the potential cE4BP4-binding sites, we performed transcriptional assays in chicken hepatoma LMH cells using a luciferase reporter construct that contains the 5′-flanking region of cPer2 gene (−607 to +67, Fig. 5A). The transcriptional activity of the cPer2-promoter-containing construct was repressed by transfection of cE4BP4 expression construct in a dose-dependent manner (Fig. 5B, Left). This cE4BP4-dependent repression was completely abolished by mutation of the distal site, but not by mutation of the proximal site (Fig. 5B, Center and Right), indicating that the distal site is responsible for the repressing activity of cE4BP4 on cPer2 promoter. To investigate whether the distal site is sufficient for the repressing activity of cE4BP4, we examined the transcriptional activity of an SV40-driven reporter construct containing three copies of the distal site with its short flanking sequences (see Materials and Methods). As shown in Fig. 5C, cE4BP4 repressed strikingly the transcriptional activity of the tandem-repeat construct, and this repression was entirely abolished when the distal site sequences were all mutated. These results indicate that cE4BP4 represses transcription from the cPer2 promoter through the distal E4BP4 recognition sequence.

Figure 5.

cE4BP4 represses transcription from cPer2 promoter. (A) cPer2 promoter. The solid box and the horizontal line show the 5′-untranslated region and the 5′-flanking region of cPer2 gene, respectively, with numbers representing the position relative to the transcription initiation site. The nucleotide sequences of the two potential cE4BP4 binding sites (open boxes; “distal site” and “proximal site”) are compared with the hE4BP4 recognition sequence (R, purine; Y, pyrimidine) (37), and the bases conserved in the cPer2 gene were marked with asterisks. Filled circles and a vertical bar indicate Sp1-binding sites and a CACGTG E-box, respectively. (B) Transcriptional assay was performed with a luciferase reporter construct containing the 5′-flanking region (−607 to +67) of cPer2 gene (cPer2 promoter-luc). The construct, “mut. distal site-luc” or “mut. proximal site-luc”, is the same with “cPer2 promoter-luc,” except that the distal site (−334 to −325) or the proximal site (−96 to −87) was mutated, respectively (see Materials and Methods). (C) Transcriptional assay was performed with an SV40-driven luciferase reporter construct that contains three copies of the distal site sequence or the mutated sequence (“distal site × 3-luc” or “mut. distal site × 3-luc,” respectively). Each value of luciferase activity was expressed relative to that of controls transfected with pcDNA3.1/V5/His empty vector alone without the cE4BP4 expression construct. All of the values are the mean ± SEM of three independent experiments.

Effect of the Prolongation of the Light Period on the Expression of cE4bp4/cPer2.

Given that cE4BP4 would repress cPer2 transcription in vivo, cPer2 expression should be down-regulated when chicks were exposed to light that induces a sufficient amount of cE4bp4 expression. Then, we investigated effects on the chick pineal cE4bp4/cPer2 expression of a prolonged light period (18 h) extended to early subjective night, a period when cE4bp4 mRNA levels are elevated strikingly by light (Fig. 3B). As shown in Fig. 6, the prolonged light period maintained the high-level expression of cE4bp4, and the subsequent decrease of the cE4bp4 mRNA level was delayed relative to that in control animals exposed to the 12-h light period. In marked contrast, the pineal cPer2 expression was maintained at low levels until the next morning, and the onset of the subsequent cPer2 increase was delayed by ≈2 h. Then, the effect of the light prolongation on the chick pineal clock phase was evaluated by cPer2/cE4bp4 expression cycles in chicks that were maintained in the dark after the prolonged light period. The phases of the rhythmic expression of both cPer2 and cE4bp4 were clearly delayed by ≈2 h relative to those in controls exposed to the 12-h light period (Fig. 6). These results indicate that the phase-delay induced by the light prolongation is closely associated with the light-dependent up-regulation of cE4bp4 and with the subsequent delay of the morning induction of cPer2.

Figure 6.

Effect of the prolongation of the light period on the expression of cE4bp4 and cPer2. One-day-old chicks were maintained in 12-h light/12-h dark cycles for 14 days. On day 15, a group of animals (18L) were maintained under the 18-h light/6-h dark condition, whereas control animals (12L) were maintained under the 12-h light/12-h dark condition. All of the animals were transferred and maintained in DD from day 16. The pineal glands were isolated at indicated time points of days 15–17. Relative mRNA levels of cE4bp4, cPer2, and cTbp were evaluated by RT-PCR (Upper), and each value is expressed as described in Fig. 4 (Lower). All of the values are the mean ± SEM from three independent experiments.

Discussion

In the present study, we have identified a bZIP transcription factor, cE4bp4, as a potential transcriptional repressor of the chick pineal cPer2 gene. Rhythmic expression of cE4bp4 with a phase almost opposite to that of cPer2 was observed in LD (Figs. 4A and 6), in DD (Figs. 4B and 6), and also under the prolonged light period condition (Fig. 6). Taken together with the repressing activity of cE4BP4 on cPer2 promoter (Fig. 5), it is strongly suggested that cE4BP4 down-regulates the chick pineal cPer2 transcription in light-dependent and time-of-day-dependent manners (see a model in Fig. 7).

Figure 7.

A model for the role of cE4BP4 in the chicken pineal clock system.

Here, we would stress the physiological importance of cE4BP4 in the light-entrainment, especially as a key player in the phase-delaying mechanism. As shown in Fig. 6, the prolonged light period, which induced a phase-delay of the clock, maintained cE4bp4 expression at high levels and delayed the rising phase of cPer2 in the next morning. This observation suggests that the delay of cPer2 induction was due to a delay of timing when its transcription became free from the suppression by cE4BP4 in the morning. Thus, cE4BP4 may play a very important role in the phase-delaying process, serving as a light-dependent transcriptional suppressor of cPer2 gene (Fig. 7). Here, it is noteworthy that the observed suppressing effect of the light prolongation on the pineal cPer2 expression (Fig. 6) contrasts strikingly with the effect of a 1-h light stimulus in DD that induces an acute increase of the pineal cPer2 mRNA level (T.O., K. Yamamoto, K. Okano, T. Hirota, T. Kasahara, M. Sasaki, Y. Takanaka, and Y.F., unpublished results). These opposite effects on cPer2 expression let us hypothesize that the phase-shift induced by the “prolonged light period” and the “short light pulse” is elicited through molecular processes totally distinct from each other.

The diurnal fluctuation of cE4bp4 expression in LD (Figs. 4A and 6) and DD (Figs. 4B and 6) suggests that cE4BP4 protein down-regulates cPer2 transcription in a time-of-day-dependent manner. On the assumption that E4BP4 protein levels rapidly reflect E4bp4 mRNA levels in the chicken pineal cells as seen in other cells (40, 41), we can speculate that cE4BP4 proteins accumulate during the (subjective) day and go into the nucleus to repress the transcription of cPer2 gene. This repression would lead to the gradual decrease in the mRNA level of cPer2 in the evening. Before the dawn, in contrast, the cE4BP4 protein level may decline along with the decrease in the cE4bp4 mRNA level, allowing the morning induction of cPer2. Thus, the time-of-day-specific rise and fall of cE4BP4 may contribute to the rhythmic expression of cPer2 gene (Fig. 7). Such a circadian function of cE4BP4 could be evaluated by the temporal change in the pineal cE4BP4 protein level. Unfortunately, however, our efforts to detect the pineal cE4BP4 proteins were unsuccessful probably because of the low-level expression to be detected by antibodies that were raised against its amino-terminal or carboxyl-terminal part fused to glutathione S-transferase (GST; data not shown).

We found that the 5′-flanking region of cPer2 gene contains not only the E4BP4 recognition sequence but also CACGTG E-box (Fig. 5A), which was demonstrated to be a functional binding site of cCLOCK:cBMAL1/2 heteromers that positively regulate cPer2 transcription (T.O., K. Yamamoto, K. Okano, T. Hirota, T. Kasahara, M. Sasaki, Y. Takanaka, and Y.F., unpublished results). Taken together with the negative regulation by cE4BP4, the time-of-day-dependent transcription of cPer2 gene is likely regulated by multiple factors in a well-concerted manner (Fig. 7). Interestingly, the promoter region of mPer1 gene contains the E4BP4 recognition sequence as well, and this site is likely a target site of transcriptional activator DBP (42), which shares very similar recognition sequences with E4BP4 (37, 39). Considering that E4BP4 and DBP could compete with each other for a single binding site, a time-of-day-specific regulation exerted by two factors with apparently opposite functions may serve as a molecular switch for the robust rhythmic expression of Per gene.

The rhythmic expression of cE4bp4 may contribute to the rhythmic expression of clock controlled (output) genes as well. Mouse vasopressin gene, one of the output genes, is subject to rhythmic transactivation by CLOCK:BMAL1 heteromers through CACGTG E-box found in the promoter region (21). It is also possible that cE4BP4 regulates a subset of output genes in a manner different from E-box-mediated one, and thereby cE4BP4 may diversify the output pathways from the circadian oscillator.

E4bp4 shows the highest similarity to Vri among all of the coding sequences of Drosophila genome. On the other hand, a blast database search (in December 2000) demonstrated that VRI protein showed higher similarity to E4BP4 than to any other members of bZIP proteins when the 70-aa stretches including the bZIP domain were compared (Fig. 2B). The structural similarity between E4BP4 and VRI strongly suggests that E4bp4 is a counterpart of Vri, which has been identified as a Drosophila clock gene by genetic studies (35). Although it is unclear at present whether VRI functions as a transcriptional repressor of dPer gene, the observed suppression of dPER protein level because of continuous expression of Vri (35) implies that the time-keeping mechanism mediated by a bZIP transcriptional repressor is conserved among animal clock systems.

Abbreviations

LD

light/dark

DD

constant darkness

SCN

suprachiasmatic nucleus

CT

circadian time

RACE

rapid amplification of cDNA ends

RT

reverse transcription

DBP

D-site binding protein