Glucocorticoid receptor suppresses gene expression of Rev‐erbα (Nr1d1) through interaction with the CLOCK complex
Abstract
Glucocorticoids have various medical uses but are accompanied by side effects. The glucocorticoid receptor (GR) has been reported to regulate the clock genes, but the underlying mechanisms are incompletely understood. In this study, we focused on the suppressive effect of the GR on the expression of Rev‐erbα (Nr1d1), an important component of the clock regulatory circuits. Here we show that the GR suppresses Rev‐erbα expression via the formation of a complex with CLOCK and BMAL1, which binds to the E‐boxes in the Nr1d1 promoter. In this GR‐CLOCK‐BMAL1 complex, the GR does not directly bind to DNA, which is referred to as tethering. These findings provide new insights into the role of the GR in the control of circadian rhythm.
Abbreviations
ADX, adrenalectomy
AP‐1, activator protein 1
DEX, dexamethasone
GC, glucocorticoid
GR, glucocorticoid receptor
GRE, glucocorticoid response element
NF‐κB, nuclear factor‐κB
PVDF, polyvinylidene fluoride
TFEL, Transcription Factor Expression Library
Adrenal glucocorticoids (GCs) are essential hormones for survival, and clinically it is also an important medicine used to treat a wide range of diseases such as asthma and rheumatic diseases mainly for the purpose of suppressing inflammation [1, 2]. However, side effects of GC treatment include insomnia, abnormal lipid metabolism, impaired glucose tolerance, bone loss and other side effects [3, 4]. Thus, clarifying unknown mechanisms of GC functions is expected to lead to the development of GC therapies with fewer side effects.
The glucocorticoid receptor (GR) is one of the most well‐studied nuclear receptors. It is translocated to the nucleus when bound to GCs and typically promotes transcription by binding to the glucocorticoid response element (GRE) as a dimer [5]. In addition to its transactivation activity, the GR can also suppress the expression of specific genes. Even though several transcriptional repression mechanisms mediated by the GR have been described, including the presence of a negative GRE where the GR represses transcription via directly binding to DNA sequences [6], many others remain unknown.
Rev‐erbα (Nr1d1) is one of the clock genes whose expression is regulated by CLOCK (NPAS2, MOP4) [7, 8], which has been identified as the first circadian clock gene in mammals [9]. CLOCK dimerizes with BMAL1 and binds to the E‐box [10], which induces transcriptional activation of Rev‐erbα, which in turn represses Bmal1 transcription [7]. This pathway comprises one of the core circadian feedback loops [11]. Since the endogenous secretion of GCs is characterized by a prominent circadian oscillation, the relationship between the circadian rhythm and the GR signaling pathway has been intensively investigated. Extensive evidence indicated that the GR signaling pathway plays a crucial role in the regulation of clock genes [12]. In particular, it has been known that the GR affects the expressions of Per genes [13, 14], which are components of another clock regulatory circuit. In addition, the GR also regulates the expression of Rev‐erbα [15, 16]. Thus, the hypothalamic–pituitary–adrenal axis is considered to act as a synchronizer of the circadian system in peripheral organs [17].
Although it has been reported that Rev‐erbα is negatively regulated by the GR [15], the mechanism of the transcriptional repression is still unknown. Here, we demonstrate that the GR suppresses Rev‐erbα gene expression in the liver via the formation of a complex with CLOCK and BMAL1. This complex binds to the E‐boxes in the promoter region of Rev‐erbα. In this GR‐CLOCK‐BMAL1 DNA bound complex, the GR does not directly bind to the DNA. This binding mechanism is often referred to as tethering. These observations provide novel insights into the GR‐mediated regulation of the clock genes.
Materials and methods
Animals
Six‐week‐old ICR male mice were purchased from SLC (Tokyo, Japan), and adapted to the experimental environment for 1 week prior to the study. All of the mice were maintained in a temperature controlled environment with 14 h light/10 h dark cycles. All mice had free access to a standard diet (MF, Oriental Yeast, Tokyo, Japan) and water. The mice were sacrificed at 14:00 (ZT 9:00).
Adrenalectomy (ADX) was performed in male ICR mice using the dorsal approach under pentobarbital and isoflurane anesthesia. Both adrenal glands were removed from the mice and the surgical incisions were sutured with 5‐0 silk (cat. No. K890H; Ethicon, Bridgewater, NJ, USA). After surgery, the mice were given free access to a standard diet, water and saline. All of the studied animals were anesthetized and euthanized according to the protocol approved by the Tsukuba University Animal Care and Use Committee.
Materials
Dexamethasone (DEX) was purchased from Wako Pure Chemical Industries (Osaka, Japan) and the Corticosterone ELISA kit was purchased from Cayman Chemical (Ann Arbor, MI, USA).
Plasmids
The Nr1d1‐promoter‐luc reporter plasmid was engineered to contain a fragment of the mouse Nr1d1 promoter region from nucleotides −1473 to +1938 in the pGL3‐basic vector (Promega, Madison, WI, USA). The Nr1d1 E‐box‐luc plasmid was engineered to contain fragments of the Nr1d1's E‐box sequences with the neighboring six nucleotides in the pGL3‐promoter vector (Promega). The sequences comprise the following: CTCAGACACGTGTGTGAGCCCTCGCACGTGGCACCCATGCAGCACGTGGGCCCGCTAGCAAACGTGAGAGCTTCACGTGATTGGA, where the nucleotides in italic bold indicate the E‐boxes.
The mouse Clock expression plasmid pCMV10/3xflag‐Clock (#47334) was purchased from Addgene (Watertown, MA, USA). The mouse Bmal1 expression plasmid was included in the Transcription Factor Expression Library (TFEL) [18]. The mouse Bmal1 dominant negative (Bmal1‐DN) expression plasmid was engineered as described previously [19] from a TFEL Bmal1 plasmid to contain 1 to 444 bases of mouse Bmal1 with a HA‐tag at the N‐terminus that was cloned into the pcDNA3.1 vector (Thermo Fisher Scientific, Waltham, MA, USA). The mouse myc‐GR expression plasmid was engineered from a TFEL GR plasmid to contain full length mouse GR with a myc‐tag at the N‐terminus that was cloned into the pcDNA3.1 vector (Thermo Fisher Scientific).
Recombinant adenoviruses
Recombinant adenoviruses were constructed as described previously using the Gateway system (Invitrogen, Carlsbad, CA, USA) [20]. The Nr1d1‐promoter‐luc adenovirus was engineered to contain the Nr1d1‐pomoter‐luc in the pENTR4 vector. The Per1 knock down adenovirus was constructed in the pENTR‐U6 vector (Invitrogen) in order to produce previously reported siRNA [21]. The sequences of the DNA oligos comprise the following: Sense, caccgCGCTCGCCCTGGCCAATAAtgtgaagccacagatggTTATTGGCCAGGGCGAGCG; Antisense, aaaaCGCTCGCCCTGGCCAATAAccatctgtggcttcacaTTATTGGCCAGGGCGAGCGc, where the uppercase nucleotides represent the target sequences.
In vivo imaging of luciferase activity
In vivo imaging was performed as described previously [20]. The adenovirus was transduced by injection into the jugular vein and after 3–14 days the D‐luciferin potassium salt (Wako Pure Chemical Industries, Osaka, Japan) was injected intraperitoneally into mice, and luminescence in the liver was captured 48 h continuously every 4 h from 8:00 (ZT 3:00) to 8:00 (ZT 3:00) for the circadian rhythm check, and at 8:00 (ZT 3:00) and 16:00 (ZT 11:00) to determine the effect of ADX surgery and DEX treatment using an IVIS imaging system (PerkinElmer, Waltham, MA, USA). The relative photon emission over the liver region was quantified using the living image software (PerkinElmer).
Transfection and luciferase assays
HEK293 cells and HepG2 cells were cultured in DMEM (Nacalai Tesque, Kyoto, Japan) containing 4.5 g·L−1 glucose, 100 U·mL−1 penicillin, and 100 μg·mL−1 streptomycin sulfate supplemented with 10% FBS.
Luciferace assays were performed as described previously [22]. Briefly, cells were seeded in a 48 well plate and cultured until they reached 60–80% confluency. Each expression plasmid, luciferase reporter plasmid and pSV40‐Renilla plasmid were co‐transfected into cells using the SuperFect Transfection Reagent (Qiagen, Hilden, Germany) for HEK293 cells and the Lipofectamine 3000 Reagent (Thermo Fisher Scientific) for HepG2 cells according to the manufacturer's protocols. After transfection, 1 mm DEX in EtOH or EtOH alone was added to the medium and the cells were cultured. After 24 h, the cells were harvested. Luciferase assays were performed according to the manufacturer's protocol and the results were normalized to Renilla luciferase activity.
Antibodies
Anti human/mouse Clock (C‐8, sc‐271603: mouse monoclonal), anti‐GR (G‐5, sc‐393232: mouse monoclonal), anti‐GR (H‐300, sc‐8992: rabbit polyclonal), anti‐c‐Myc (9E10, sc‐40: mouse monoclonal), anti‐Per1 (E‐8, sc‐398890: mouse monoclonal) and anti‐GAPDH (6C5, sc‐32233: mouse monoclonal) antibodies and normal mouse IgG (sc‐2025) were purchased from Santa Cruz Biotechnology (Dallas, TX, USA). Anti‐Flag antibody (M2: mouse monoclonal) was purchased from Sigma‐Aldrich (St. Louis, MO, USA).
Immunoprecipitation
We performed immunoprecipitation experiments as described previously [23]. Nuclear protein extracts from mouse liver were prepared as described previously [24, 25].
Western blots
Proteins were separated by SDS/PAGE and electrically blotted onto a polyvinylidene fluoride (PVDF) membrane utilizing a semi‐dry method. Blocking was conducted with 5% skim milk. Each antibody was used at a 1 : 1000 dilution, and incubated with the PVDF membrane overnight at 4 °C with shaking. The membranes were then washed and the secondary antibody was incubated at room temperature for 1 h. The bands were then visualized using the ECL western blotting detection system (GE Healthcare, Chicago, IL, USA).
ChIP assay
Isolation of hepatic nuclei from mouse liver and ChIP assays were performed as described previously [26, 27]. Chromatin DNA was used as a template for quantitative PCR (q‐PCR). The primer sets are listed in the Table S1. TAT primers were used as a positive control and primers that amplify a region located upstream of the Nr1d1 gene with no ChIP‐seq peak were used as a negative control.
Quantitative PCR
Total RNA was prepared using the Sepasol reagent (Nacalai Tesque), and total RNA (1 μg) was reverse transcribed using the Perfect Real Time kit (Takara, Tokyo, Japan) as described previously [28]. q‐PCR was performed with SYBR green dye (Kapa Biosystems, Wilmington, MA, USA) on a 7300 Real‐Time PCR system. Primer sets are listed in the Table S1.
Statistical analysis
The data were presented as the mean ± SEM. The data were compared by the unpaired two‐tailed Student's t‐test for the comparison of two groups. The paired data from the ChIP assay were compared by the paired one‐tailed Student's t‐test. Data sets involving more than two groups were assessed by the Holm's post‐hoc test. Differences were considered statistically significant at P < 0.05 (*P < 0.05 and **P < 0.01). All of the experiments were repeated at least twice.
Results
Comprehensive search for genes regulated by GCs and GR
Similar to humans, endogenous murine GCs exhibit diurnal variation and are induced by environmental stresses [29]. It has also been reported that GCs are involved in the diurnal variation in several genes and 100 genes have been reported to fluctuate daily in response to GCs in the liver [30]. Therefore, we examined the previously reported microarray data that identified genes involved in the GR‐mediated pathways (GEO No: GSE34229, GSE24256, GSE21048, GSE13461 and GSE24255) and searched for genes that were altered in response to GCs in the liver. We also examined the ChIP‐seq data (GSM1122512 and GSM1122513) on the UCSC genome browser, and took the presence of ChIP‐seq peaks into consideration (Fig. S1A). As a result, we decided to focus on Rev‐erbα (Nr1d1), which was markedly suppressed by GCs and had significant binding peaks of GR in the gene (Fig. S1B).
GR agonist suppresses Nr1d1 gene expression in the liver
Next, we checked the effect of a GR agonist on Nr1d1 gene expression in the liver. In order to exclude the influence of endogenous fluctuation of GCs, the adrenal glands were removed. Bilateral ADX or a sham operation was performed on 6–7 week‐old male mice, and the reduction in corticosterone in the ADX group was confirmed (Fig. S1C). After 2–3 days of postoperative stabilization, mice were injected intraperitoneally with 1 mg·kg−1 of DEX, a GR agonist, or a vehicle control. At 6 h after injection, we examined the expression of Nr1d1 in the liver using q‐PCR. In the DEX group, Nr1d1 gene expression was markedly suppressed compared with the vehicle group (Fig. 1A). Thus, it was confirmed that a GR agonist suppressed Nr1d1 gene expression in the liver.
In vivo Ad‐luc analysis showing Nr1d1 promoter sufficient for circadian oscillation
Based on the available GR ChIP‐seq data (ChIP‐Atlas database [31] and GSM1122512 and GSM1122513), we cloned the promoter region of the mouse Nr1d1 gene (nucleotides −1473 to +1938) and transferred it to a luciferase reporter vector. This region included all of the ChIP‐seq peaks of GR binding as well as the E‐boxes reported to regulate Nr1d1 gene expression (Fig. S1B) [32].
In order to evaluate whether this promoter region is sufficient to produce circadian oscillation in vivo, we performed animal experiments using the in vivo Ad‐luc analytical system originally developed by our laboratory [18, 20, 28, 33]; Nr1d1‐promoter‐luc adenovirus was administered to mice, and luciferase reporter activities were measured in the liver using an IVIS imaging system. As shown in Fig. 1B and Fig. S1D, the Nr1d1‐promoter‐luc construct exhibited the circadian expression pattern, which showed the peak value at 16:00–20:00 (ZT 11:00–15:00) and the bottom value at 8:00 (ZT 3:00). Moreover, DEX administration at 8:00 (ZT 3:00) markedly lowered luciferase activities at 16:00 in both ADX and sham operation groups (Fig. 1B–E). These results indicated that the 3 kb promoter region is sufficient to confer the circadian oscillation of the Nr1d1 gene in vivo. It was also notable that in the ADX group, the luciferase activities were significantly elevated at both 8:00 (ZT 3:00) and 16:00 (ZT 11:00) following ADX compared with pre‐ADX. These results suggested that GC deficiency due to ADX can upregulate Nr1d1 promoter activity.
Inhibitory effects of the GR on Nr1d1 promoter activity mediated by Clock/Bmal1
Using the 3 kb Nr1d1 promoter linked to luciferase, we next examined the mechanism by which the GR suppresses Nr1d1 gene transcription. When the GR expression plasmid and the GR ligand DEX were tested in HEK293 cells, no suppressive effects on Nr1d1 promoter activity were observed (Fig. S1E). Since Nr1d1 is known to be upregulated by Clock and Bmal1 via E‐box sequences [8], co‐transfection of expression plasmids encoding Clock and Bmal1 were tested with the GR expression plasmid and the Nr1d1 promoter linked to luciferase. As shown in Fig. 2A, suppression of the Nr1d1 promoter activity by the GR occurred in the presence of Clock and Bmal1. Notably, in the presence of a dominant‐negative form of Bmal1 (Bmal1‐DN), the suppression of the Nr1d1 promoter activity by the GR was not observed. These results demonstrated that the GR cannot suppress Nr1d1 promoter activity without Clock and Bmal1. Thus, the inhibitory effects of the GR on Nr1d1 expression are mediated by Clock and Bmal1.
GR suppresses Nr1d1 promoter without direct GR‐binding sites
Next, we examined the presence or absence of direct GR‐binding sites within the 3 kb Nr1d1 promoter region. Since we found five E‐boxes in this region, these E‐boxes were extracted and tandemly connected. Although the E‐box comprises a sequence of six nucleotides, it is known that its effect varies depending on the surrounding sequence [32], so we included the surrounding six nucleotides of each E‐box in our reporter construct. When the E‐box‐luc reporter construct lacking GR‐binding sites was transfected into HEK293 cells, the effect of co‐transfected Clock and Bmal1 was similar to the native 3 kb Nr1d1‐promoter‐luc reporter construct (Fig. 2B). Moreover, the inhibitory effect of the GR on the E‐box‐luc reporter construct was also quite similar to the native promoter, despite its lack of direct GR‐binding sites (Fig. 2A–C). These results indicated that the GR suppresses Nr1d1 promoter activity through its interaction with Clock/Bmal1, without direct binding to DNA.
Exclusion of secondary suppression due to GR effects on other clock genes
It is known that clock genes form several feedback loops, and among them several genes vary in a GR‐dependent manner. These include the clock genes Per and Cry that suppress Clock/Bmal1 [34]. In addition, Per1 is known to be increased by GCs [13]. Therefore, it is possible that the GR may indirectly suppress the E‐box via the elevation of Per1. To assess these secondary influences, we examined several clock genes and found that only Per1 expression was increased upon GR activation in HepG2 cells as well as in the liver (Fig. S2). Next, we examined the effects of the GR on Nr1d1 promoter activity while abolishing the contribution from Per1 by using a knockdown adenovirus directed against Per1. The HepG2 cells were infected with the Per1 knockdown adenovirus and the Per1 knockdown was confirmed (Fig. S3D–F). In the absence of Per1, the suppression of the Nr1d1 promoter activity upon GR activation remained unchanged (Fig. S3A,B). These results indicated that Per1 did not play a role in GR‐mediated inhibition of Nr1d1 expression.
Molecular interaction between GR and CLOCK complex
In order to examine the direct molecular interaction between the GR and the Clock proteins, an immunoprecipitation experiment was conducted. This interaction has previously been reported elsewhere [35]. The immunoprecipitation experiments were performed both in vitro and in vivo. In the in vitro experiment, we found that the GR interacted with Clock proteins and this interaction was independent of GCs in whole cell lysates (Fig. 3A). In the in vivo experiment, all of the mice were adrenalectomized and the interaction between the GR and Clock in liver nuclei was examined with or without DEX administration. Our results showed that the GR is translocated to the nucleus in a ligand‐dependent manner, where it binds to Clock proteins (Fig. 3B).
Ligand‐dependent DNA binding of GR to E‐box
In order to confirm the binding of the GR to the E‐boxes on chromatin, GR ChIP analysis was performed in this region containing the E‐boxes. Mice were adrenalectomized and compared between groups administered 1 mg·kg−1 DEX and vehicle at 8:00 (ZT 3:00). Primers from the E‐boxes E1 to E4/5 were prepared to quantify each E‐box in the Nr1d1 promoter region. Negative control primers were utilized to amplify a region located upstream of the Nr1d1 gene with no ChIP‐seq peaks. Primers that amplify the tyrosine aminotransferase (TAT) gene enhancer were used as a positive control that detected the GRE site in a previous GR ChIP study [36]. As shown in Fig. 4A, significant increases in GR binding upon GR activation with DEX were observed in all of the E‐boxes from E1 to E4/5. These results indicated that the GR indirectly interacted with the E‐boxes in a ligand‐dependent manner. In contrast, the results from Clock ChIP experiments showed that the Clock protein was apparently bound to the E‐boxes and that GR activation did not affect its binding (Fig. 4B). These results are consistent with our model that the GR is translocated into the nucleus in a ligand‐dependent manner and suppresses Nr1d1 transcription by forming a complex with the Clock protein on the E‐box.
Discussion
In this study, we demonstrated that the GR suppresses Rev‐erbα (Nr1d1) gene expression through interaction with the Clock/Bmal1 complex on the E‐boxes in the Nr1d1 promoter (Fig. 5).
The GR can associate with specific genomic sites in multiple ways. Currently, the following four modes of site‐specific GR‐genome interactions are known [37]: (a) two GR monomers bind to a canonical GR‐binding sequence present in a GRE in a head‐to‐head fashion, (b) the GR binds to inverted‐repeat GR‐binding sequences and two GR monomers are bound to opposite sides of the DNA in a head‐to‐tail fashion, (c) the GR interacts with GR‐binding half sites operating in conjunction with proximal non‐GR transcriptional regulatory factors and (d) the GR interacts at specific genomic sites without directly binding to DNA. Our results showed that among the four types of sequence‐specific GR‐DNA interactions, the GR‐Nr1d1 promoter interaction can be classified as type 4. In this GR‐DNA interaction mode, the GR can occupy specific genomic regions without directly binding to DNA, which is often referred to as tethering. In this mechanism, the DNA‐binding domain of the GR forms protein‐protein contacts with other transcription factors, such as activator protein 1 (AP‐1) or nuclear factor‐κB (NF‐κB), which are specifically bound at their cognate sequence motifs. Thus, Clock/Bmal1 now joins a growing list of transcription factors, including NF‐κB, AP‐1, CREB, NF‐AT, STAT6, IRF3, STAT3, GATA‐3 and t‐Bet, whose activities are known to be suppressed by the tethering mode of GR action [38].
As for the biological relevance of GR‐mediated repression of Rev‐erbα, it is known that GR activation is an entraining signal for peripheral circadian oscillators [17], and it can reset the phase of the clock by regulating Rev‐erbα expression as well as Pers [16]. The GC‐mediated periodical resetting of the peripheral clock can be particularly important during stress, when organisms need to adjust the circadian rhythm‐linked activity of their bodies to react to stressors [12]. In addition, the endogenous secretion of GCs, which peaks during the habitual sleep‐wake transition [39], might partly contribute to the suppression of Rev‐erbα during the active awake period (Fig. 1D). However, this mechanism can be compensated by other mechanisms as it is obvious from the fact that adrenalectomized mice also exhibited circadian oscillation pattern of Rev‐erbα expression similar to normal mice (Fig. 1D,E).
In conclusion, the mechanism by which the GR reduces Rev‐erbα expression in the liver was clarified; the direct molecular interaction between the GR and the CLOCK complex on E‐boxes suppresses the transcription of Rev‐erbα (Nr1d1) gene.
Acknowledgements
This work was supported by MEXT/JSPS KAKENHI Grant Numbers 23116006 (Grant‐in‐Aid for Scientific Research on Innovative Areas: Crosstalk of transcriptional control and energy pathways by hub metabolites), 15H03092 (Grant‐in‐Aid for Scientific Research (B)), 21591123 and 18590979 (Grant‐in‐Aid for Scientific Research (C)), 26560392 and 16K13040 (Grant‐in‐Aid for Challenging Exploratory Research), and 03J10558 (Grant‐in‐Aid for JSPS Fellows) (to NY). It was also supported by research grants from the Uehara Memorial Foundation, Nakatani Foundation, ONO Medical Research Foundation, Takeda Science Foundation, Suzuken Memorial Foundation, Japan Heart Foundation, Kanae Foundation for the Promotion of Medical Science, Senri Life Science Foundation, Japan Foundation for Applied Enzymology, and Okinaka Memorial Institute for Medical Research (to NY).
Author contributions
YM and NY conceived the experiments. YM performed the experiments under the guidance of YT and analyzed the data together with NY. YM and NY co‐wrote the paper. All authors discussed the results and commented on the manuscript.
