A direct repeat of E-box-like elements is required for cell-autonomous circadian rhythm of clock genes

Abstract

Background: The circadian expression of the mammalian clock genes is based on transcriptional feedback loops. Two basic helix-loop-helix (bHLH) PAS (for Period-Arnt-Sim) domain-containing transcriptional activators, CLOCK and BMAL1, are known to regulate gene expression by interacting with a promoter element termed the E-box (CACGTG). The non-canonical E-boxes or E-box-like sequences have also been reported to be necessary for circadian oscillation.

Results: We report a new cis-element required for cell-autonomous circadian transcription of clock genes. This new element consists of a canonical E-box or a non-canonical E-box and an E-box-like sequence in tandem with the latter with a short interval, 6 base pairs, between them. We demonstrate that both E-box or E-box-like sequences are needed to generate cell-autonomous oscillation. We also verify that the spacing nucleotides with constant length between these 2 E-elements are crucial for robust oscillation. Furthermore, by in silico analysis we conclude that several clock and clock-controlled genes possess a direct repeat of the E-box-like elements in their promoter region.

Conclusion: We propose a novel possible mechanism regulated by double E-box-like elements, not to a single E-box, for circadian transcriptional oscillation. The direct repeat of the E-box-like elements identified in this study is the minimal required element for the generation of cell-autonomous transcriptional oscillation of clock and clock-controlled genes.

Figure 1

The "EE-element" possesses transcriptional activity displaying a circadian rhythm. (A) Sequence alignments of the promoter region, containing each EE-element of Per1, Per2 and Per3 gene, among human, rhesus, rat, mouse and opossum (The NCBI accession numbers are indicated in Additional file 6). The core sequences (E1 and E2) within the EE-element are shown in capital letters. Numbers at both sides of alignments indicate the position from the transcription start site (TSS). (B) EE-element-driven luciferase bioluminescence by IV-ROMS. Black lines indicate bioluminescence after serum shock; and gray lines show the negative control (replacement with same medium). Abscissa presents "day"; and ordinate, "relative luciferase intensity". First peak values of the curves were set to 1.

Figure 2

Effects of EE-element mutants on circadian expression of luciferase. (A) Wild-type (upper, gray letters) and mutant (bottom, black letters) sequences of hPer2 E1 and E2 and their flanking sequences are shown. Underlines indicate sequences changed from the wild-type one. The core sequence within the EE-element is shown in capital letters. (B) EE-element-driven luciferase bioluminescence by IV-ROMS. Gray lines indicate bioluminescence of wild-type hPer2 EE-element; and black lines, those of mutants. The abscissa presents "day", and the ordinate indicates "relative luciferase intensity". First peak values of the curves were set to 1.

Figure 3

Pull-down assays by using oligonucleotides. (A) Both BMAL1 and CLOCK proteins bind to Per2 oligonucleotide. BMAL1 is phospholrylated in a circadian manner, whereas no binding to the random oligonucleotide occurs. Mouse liver lysates at CT 16 were electrophoresed and immunoblotted with anti-BMAL1 and anti-CLOCK (PC) and immunoprecipitated by anti-CLOCK and blotted by either BMAL1 or CLOCK (IP). WT; wild-type hPer2 oligonucleotide, E1; E1 mutant oligonucleotide, E2; E2 mutant oligonucleotide, E1mE2m; E1, E2 double mutant oligonucleotide. (B) As for BMAL1, the binding to E2m is clearly observed; whereas no binding to either E1m or E1mE2m is detected. For CLOCK, the binding to E2m is observed; however, the binding to E1m is hardly seen and no binding to E1mE2m is evident.

Figure 4

The space between the 2 E-box-like sequences is critical for cell-autonomous circadian transcription. (A) Wild-type (sp6; gray letters) and mutant (sp4, 5, 7, 8; solid letters) sequences of the hPer2 EE-element and their flanking sequences are shown. Underlines indicate nucleotide(s) inserted into the space of the EE-element. The core sequence within the EE-element is shown in capital letters. (B) EE-element-driven luciferase bioluminescence detected by IV-ROMS. Black lines indicate bioluminescence of mutant hPer2 EE-elements; and gray lines, those of the wild type. The abscissa presents "day"; the ordinate shows "relative luciferase intensity". First peak values of the curves were set to 1.

Figure 5

Structural model for the binding of bHLH transcription factors to DNA. (A) Since 1 turn of double helix DNA is approximately 34 Å and requires 10.4 nucleotides, 1 EE-element, which consists of 12 nucleotides, spans 39.2 Å (34/10.4 × 12). (B) Two binding faces form a 55.8-degree angle (360/10.4 × 12–360). (C) The maximum dimension of the bHLH structure, which lies along the DNA, is about 32 Å, which is within the 39 Å calculated above and allows the binding. Green indicates a hetero complex of transcription factors such as BMAL1 and CLOCK; and light blue, unknown factors.