The mammalian clock component PERIOD2 coordinates circadian output by interaction with nuclear receptors

Abstract

Mammalian circadian clocks provide a temporal framework to synchronize biological functions. To obtain robust rhythms with a periodicity of about a day, these clocks use molecular oscillators consisting of two interlocked feedback loops. The core loop generates rhythms by transcriptional repression via the Period (PER) and Cryptochrome (CRY) proteins, whereas the stabilizing loop establishes roughly antiphasic rhythms via nuclear receptors. Nuclear receptors also govern many pathways that affect metabolism and physiology. Here we show that the core loop component PER2 can coordinate circadian output with the circadian oscillator. PER2 interacts with nuclear receptors including PPARalpha and REV-ERBalpha and serves as a coregulator of nuclear receptor-mediated transcription. Consequently, PER2 is rhythmically bound at the promoters of nuclear receptor target genes in vivo. In this way, the circadian oscillator can modulate the expression of nuclear receptor target genes like Bmal1, Hnf1alpha, and Glucose-6-phosphatase. The concept that PER2 may propagate clock information to metabolic pathways via nuclear receptors adds an important facet to the clock-dependent regulation of biological networks.

Figure 1.

PER2 interacts with nuclear receptors. (A) Structural organization of mouse PER proteins. Gray boxes represent the PAS A, PAS B, and PAC domains. Corepressor-like (CoRNR, yellow) and coactivator like (LXXLL, pink) protein–protein interaction motifs are highlighted. Alignment of the LXXLL sequences of PER1 (blue) and PER2 (red) is shown. Numbers indicate the amino acid positions in the primary structure. (B) PPARα was immunoprecipitated from mouse liver nuclear extracts. Extracts derived from Per2^Brdm1 mice were used to monitor the specificity of the antibody against PER2. (C) BMAL1 was immunoprecipitated from mouse liver nuclear extracts. (D) HA-tagged nuclear receptors were expressed in NIH 3T3 cells and were immunoprecipitated from nuclear extracts. Expression vectors used for cotransfection are indicated. (E) PER2 was immunoprecipitated from mouse liver nuclear extracts. An extract derived from Rev-Erbα^−/− mice demonstrates the specificity of the antibody against REV-ERBα. Immunoprecipitated proteins were detected by Western blot analysis with the indicated antibodies (on the right). The input is shown in the left panels. On the left side of the panels, the positions of marker bands are indicated (relative molecular weight). Reactions with beads and extract alone were used as controls for nonspecific binding. (ZT) Zeitgeber time.

Figure 2.

LXXLL-related motifs in PER2 are important for its physical interaction with nuclear receptors. (A–C) Immunoprecipitation of HA-tagged PPARα (A,C) or REV-ERBα (B,C) from NIH 3T3 nuclear extracts (see Fig. 1D). (mLCCLL) Mutation of N-terminal motif; (mLLNLL) mutation of C-terminal motif; (mLXXLL) mutation of N- and C-terminal motifs. (D,E) NIH 3T3 cells were transfected with Rev-Erbα (D) or Bmal1 (E) luciferase reporter vector either alone or together with an expression vector for PER1 (blue), PER2 (red), or mutated PER2 (mLXXLL; green). Amounts of expression vectors were adjusted to yield similar repression of Rev-Erbα luc. Cells were synchronized with dexamethasone, and luciferase activity of each culture was recorded. Data are plotted as mean ± SD (n = 2, representative experiment out of three independent experiments).

Figure 3.

PER2 binding to the regulatory regions of circadian target genes in vivo. (A–D) Chromatin from wild-type (solid lines) and Rev-Erbα^−/− mice (dashed lines) was prepared at 4-h intervals from mice held in a 12-h light/12-h dark cycle (LD 12:12). The binding of the indicated proteins to their regulatory region was analyzed by ChIP. Specific TaqMan probes were used to detect the Rev-Erbα promoter region (A), the Bmal1 promoter region (+50, B), the Bmal1 upstream regulatory region (−1600, C), and the Hnf1α promoter region (D). Plotted are the mean values ± SEM from three independent experiments; ZT0 is double-plotted. (*) The regulatory region preceding the Rev-Erbα knockout allele.

Figure 4.

Constant expression of Bmal1 in livers of Rev-Erbα^−/−/Per2 mutant mice. Total RNA, protein extracts, and chromatin from wild-type (black), Rev-Erbα^−/− (green), Per2^Brdm1 (blue), and Rev-Erbα^−/−/Per2 mutant (red) animals maintained in LD 12:12 were prepared. (A,C) The relative amounts of Bmal1 (A) and Per1 (C) mRNA were measured by TaqMan real-time RT–PCR. Plotted are the mean values ± SEM from three independent experiments; ZT0 is double-plotted. (B) NIH 3T3 cells were transfected with Bmal1 ext. luciferase reporter, HA-Rxrα expression vector (RXRα as a heterodimerization partner of PPARα was used to enhance the effects of PPARα), and the indicated expression vectors. Cells were synchronized with dexamethasone, and bioluminescence was recorded. The top panel shows the cotransfections with Per2-V5 (red), and the bottom panel shows the cotransfections with mutated Per2-V5 (mLXXLL, green). Dashed lines or dotted lines represent cotransfections of the increasing amounts of HA-Pparα together with the constant amount of Per2-V5 or mLXXLL-Per2-V5 expression vector, respectively. Amounts of expression vectors were adjusted to yield similar repression of Rev-Erbα luc (data not shown). Data are plotted as mean only (n = 2, representative experiment out of three independent experiments). Note that the LXXLL sequences are involved in the activation of the reporter gene by PER2 and PPARα. (D) Nuclear extracts were analyzed by Western blot analysis using the indicated antibodies. Arrows indicate differently migrating forms of the corresponding protein. The accumulation of RNA polymerase II (αPOL2) is shown as a loading control. (E) ChIP analysis with αBMAL1 using a specific TaqMan probe for the promoter region of the Rev-Erbα gene or the Rev-Erbα^−/− knockout allele (Rev-Erbα*) (mean ± SEM from three independent experiments; ZT0 is double-plotted).

Figure 5.

Analysis of circadian mRNA in the SCN and free-running activity of wild-type and Rev-Erbα^−/−/Per2 mutant animals. (A–C) Relative expression of Bmal1 (A), Per1 (B), and Cry1 (C) mRNA in SCN tissue under constant (DD) conditions from wild-type (black), Rev-Erbα^−/− (green), Per2^Brdm1 (blue), and Rev-Erbα^−/−/Per2 mutant (red) animals as revealed by in situ hybridization. Plotted are the mean values ± SEM from three independent experiments; CT0 is double-plotted. (D) Representative locomotor activity records of wild-type, rhythmic, and arrhythmic Rev-Erbα^−/−/Per2 mutant animals as double plots. The top bar indicates two consecutive days, and the dark shaded areas represent phases of darkness. The side bar indicates the days when the animal displays arrhythmic locomotor activity.

Figure 6.

Rhythmic glycogen accumulation is abolished in Rev-Erbα^−/−/Per2 mutant mice. (A,B) mRNA analysis in liver. The relative amounts of Pepck1 (C) and G6Pase (D) mRNA were measured by TaqMan real-time RT–PCR. (C) Cytosolic PEPCK activity. (D) G6Pase activity measured in liver microsomes. (E) Liver glycogen content. Values represent the mean ± SEM. (F) PAS staining of liver tissue in wild-type and Rev-Erbα^−/−/Per2 mutant animals at ZT0 and ZT12. Bar, 200 μm. Plotted are the mean values ± SEM from three independent experiments derived from wild-type (black), Rev-Erbα^−/− (green), Per2^Brdm1 (blue), and Rev-Erbα^−/−/Per2 mutant (red) animals; ZT0 is double-plotted.

Figure 7.

Model of how PER2 may couple E-box-driven and nuclear receptor-regulated gene expression. In the oscillator, both PER proteins act as repressors of E-box-mediated circadian transcription via interaction with BMAL1 and CLOCK. In addition, PER2 can modulate NRE-mediated transcription via interaction with nuclear receptors. This affects expression of Bmal1 in the oscillator and target genes in the output such as Hnf1α may be regulated. These output targets can be modulated via E-boxes, NREs, or both. PER2–nuclear receptor interactions may be involved in this regulatory process (hatched arrows) to coordinate clock output processes.