A Novel Bmal1 Mutant Mouse Reveals Essential Roles of the C-Terminal Domain on Circadian Rhythms

To generate Bmal1^GTΔC mice, ES cells harboring C-terminus truncated Bmal1 gene were obtained from Sanger Institute Gene Trap Resource (SIGTR, Cambridge, UK) [15]. The ES cells were injected into blastocystes and chimeric mice were generated as described previously [16]. The genotypes were determined by using three primer polymerase chain reaction (PCR) method (WT-F, 5’-CCTCTCCAAGGCTGTTTCTG-3’, WT-R, 5’-TGAGCCTGCCCTGGTAATAG-3’ and Bmal1^GTΔC-R, 5’-GGCCAAGTTTGTTTCCTTGA-3’). The mice were backcrossed for six generations on C57BL/6 mice purchased from the Orient Bio (Orient Bio Inc., Seongnam, Korea). For generating littermates, ten weeks-old male mice were paired at a ratio of 1:2 with female mice.

After genotype selections, male mice were transferred to a conventional animal facility and housed individually in Nalgene cages (Nalge Nunc, Rochester, NY). All mice were maintained in a specific pathogen free (SPF) animal facility and a sound-proof isolated room with a constant temperature (22–23°C) and humidity (50 ± 10%). Lights were on at 08:00 AM and off at 20:00 PM (L:D = 12h:12h). Illumination was provided by 32 W cool white fluorescence bulbs adjusted to 350 lux at the bottom of cage and completely blocked during the scotophase (< 0.5 lux). Mice were housed with a wood-chip bedding [17] and fed ad libitum with NIH-31 rodent chow (Zeigler Brothers, Gardners, PA) and tap water. All animal experiments were made to minimize suffering and approved by Seoul National University Institutional Animal Care and Use Committee. For wheel running activity experiments, the mice were transferred to a cage supplemented with a wheel running assembly at 8 weeks of age (Mini-Mitter, Bend, OR). After 1 week of adaptation, the wheel running activity was recorded at 6 min intervals and the raw data files were analyzed as described previously [18]. The fast Fourier transform (FFT) and free running period (FRP) were analyzed by ClockLab data analysis program (Actimetrics, Wilmette. IL). For the circadian tissue sampling, mice were sacrificed every 6hr from 30hr to 48hr in a constant darkness condition.

NIH-3T3 and mouse embryonic fibroblasts (MEFs) were maintained in DMEM containing 10% FBS, 100 units penicillin and 100 g/ml streptomycin. For MEF preparations, 13.5 day old mouse embryos were collected and dissociated by mincing and Trypsin-EDTA treatment [19]. Transfections were performed by using Lipofectamine and Plus reagent according to the manufacturer’s instructions (Invitrogen, Carlsbad, CA). For luciferase assays, 10 ng of pmPer1 6.8Kbp-luciferase construct was co-transfected with other effectors and total quantity of DNA was adjusted by pcDNA 3.0. Cells were harvested and luciferase activities were measured by dual luciferase assay system. The transfection efficiency was adjusted by renilla luciferase activity of pRL-TK (Promega, Madison, WI). All assays were performed in triplicate and repeated three times independently.

Mouse brains were immediately frozen in liquid nitrogen-chilled isopentane. Frozen sections (12mM) were cut coronally and fixed with 4% paraformaldehyde and dehydrated. Riboprobes were produced from the plasmids containing Per2 (GenBank accession number: AF036893↗, n.t. 186–555) and Bmal1 (GenBank accession number: AF022992↗, n.t. 1261–1614) cDNA. The probes for Bmal1 mRNA were designed to detect both Bmal1^wt and Bmal1^GTΔC. Antisense riboporbes were prepared by in vitro transcription using α-[³⁵S] UTP as manufacturer's instruction (Promega, Madison, WI). Sections were hybridized with the riboprobes overnight at 52°C and washed with sodium-saline citrate. Then, the sections were treated with RNase A (20 mg/ml) for 30 min at 37°C, and rinsed. Radioactive signals were visualized by exposing the sections to β-max film for 7 days (Kodak, Rochester, NY).

Mouse liver samples were immediately frozen by liquid nitrogen at the indicated times and stored at -80°C until the analysis. For mRNA analysis, total RNAs were extracted from the liver by the single step acid guanidinium thiocyanate-phenol-chloroform method. The purified RNA was reverse-transcribed according to the manufacturer’s instructions (Promega). The mRNA levels were quantified by the real-time PCR using following primer sets and normalized by GAPDH levels. Per1-F, 5′-GTGTCGTGATTAAATTAGTCAG-3′, Per1-R, 5′-ACCACTCATGTCTGGGCC-3′; Per2-F, 5′-GCGGATGCTCGTGGAATCTT-3′, Per2-R, 5′-GCTCCTTCAGGGTCCTTATC-3′; Bmal1-F, 5’-CCTAATTCTCAGGGCAGCAGAT-3’, Bmal1-R, 5’-TCCAGTCTTGGCATCAATGAGT-3’, Clock-F, 5’-TTGCTCCACGGGAATCCTT-3’, Clock-R, 5’-GGAGGGAAAGTGCTCTGTTGTAG-3’, Rev-erbα-F, 5’-AAGACATGACGACCCTGGAC-3’, Rev-erbα-R, 5’-GAGTCAGGGACTGGAAGCTG-3’, Cry1-F, 5’-CGAATGAATGCAAACTCCCT-3’, Cry1-R, 5’-AAAAATTCACGCCACAGGAG-3’, Dbp-F, 5’-CAAGAACAATGAAGCAGCCAAGAG-3’, Dbp-R, 5’-AGGGCACAAGCAACATTACC-3’, GAPDH-F, 5′-CATGGCCTTCCGTGTTCCTA-3′, GAPDH-R, 5′-CCTGCTTCACCACCTTCTTGA-3′. The probes for Bmal1 mRNA were designed to detect both Bmal1^wt and Bmal1^GTΔC.

For IB experiments, mouse liver tissues and cells were lysed with radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris [pH 8.0], 1% Triton X-100, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 1 mM NaF, 1 mM Na₃VO₄, and 1X protease inhibitor cocktail (Sigma, St. Louis, MO). Proteins were resolved on 6% or 8% SDS-polyacrylamide gels and transferred onto polyvinylidene difluoride membranes (Millipore, Billerica, MA). The membrane was blocked for 1 h at room temperature in 10% skim milk solution and incubated with several primary antibodies such as anti-BMAL1 [20], CLOCK (SantaCruz, Dallas, TX), β-GAL (Promega, Madison, WI), MYC (SantaCruz, Dallas, TX), ACTIN (SantaCruz, Dallas, TX). Immunoreactive bands were visualized with ECL reagents (Thermo Scientific, Waltham, MA) according to manufacturer's instructions. For IP experiments, NIH3T3 cells were transfected with the indicated plasmids. At 36 hours posttransfection, the cells were lysed with RIPA buffer and centrifuged at maximum speed for 20 min at 4°C. Equal amounts of total protein were incubated with 2μg of anti-Flag M2 (Sigma, St. Louis, MO) for 1.5 h at 4°C and then added protein A/G-Sepharose bead slurry. The final immune complexes were analyzed by immunoblotting with indicated antibodies [21].

To observe endogenous circadian rhythms of Bmal1^+/GTΔC and Bmal1^GTΔC/GTΔC, MEFs were generated from embryos of mutant mice that were bred with PER2::LUC knock-in mice [22]. At 6^th passage, those MEFs were incubated with 1μM dexamethasone (DEX) for 2hr, then, media were changed with a recording media containing 100μM D-luciferin (Promega). To analyze the effect of Bmal1^GTΔC on the circadian rhythm by the transient transfection, immortalized wild type (WT) MEFs were transfected with mutant clones and dual-color luciferases (Per2-SLR2 and Bmal1-Eluc). After 36hr of transfection, cells were synchronized by 1μM DEX treatment and bioluminescence was measured for 1min at 10min interval using KronosDio (ATTO Corporation, Tokyo, Japan) [21]. Each luciferase activity was calculated as previously reported [23]. The period length and amplitude were calculated as previous described [21].

To analyze the effect of Bmal1 C-terminal region on the localization of BMAL1 and CLOCK simultaneously, VENUS protein was fused to BMAL1^wt and BMAL1^GTΔC, and CERULEAN (CERUL) to CLOCK as indicated [24]. For BiFC assay, amino acid residues 1 to 172 (VN) and 173 to 238 (VC) of VENUS protein was fused to BMAL1 and CLOCK, respectively. NIH-3T3 cells were transfected with the indicated constructs, incubated at 37°C for 36hr and visualized by using Delta-Vision fluorescence microscope (Applied Precision, Isaquah, WA).

Data were analyzed by one-way or two-way analysis of variance (ANOVA) with Tukey post-tests using GraphPad PRISM (GraphPad Prism Software, La Jolla, CA). A p value less than 0.05 was considered as a significant difference.

We obtained ES cells from SIGTR where the gene trap vector was inserted in the Bmal1 gene locus. To identify the precise insertion site, we sequenced the 18^th intronic DNA region of Bmal1 based on SIGTR sequence tag information (#CE0167, Arntl^{Gt(CE0167)Wtsi}) [15]. The gene trap vector containing Engrailed 2 (En2), self-cleaving sequence, β-Geo (β-Galactosidase + neomycin) and poly(A) signal was found to be inserted at 3704 bp away from the 18^th exon (Fig 1A and S1A Fig). cDNA sequencing revealed that the 18^th intron of Bmal1^GTΔC was successfully spliced out and the 17^th exon was connected to the En2 exon, self-cleaving sequence 2A and β-Geo to generate a functional mRNA (Fig 1A and S1A Fig). Domain structures of WT and mutant BMAL1 proteins are illustrated (Fig 1B and S1B Fig). Although the hypothetical size of the fusion protein is approximately 180 kDa, β-GALACTOSIDASE (β-GAL) antibody detected 110 kDa proteins, indicating a functional self-cleaving 2A peptide sequence (Fig 1D). The genotypes were further determined by three primers PCR and immunoblotting (Fig 1C and 1D). Collectively, these results indicated that the gene trap vector was inserted at the 18^th intronic region of Bmal1 generating C-terminus truncated BMAL1 (residues 1–538) plus 52 amino acids from EN2 and the self-cleaving sequence. Using these heterozygous male and female mutant mice, we generated WT, heterozygous and homozygous mice for the experiments. The genotypes of littermates exhibited a 1:2:1 ratio, indicating no overt effect of Bmal1^GTΔC allele on embryonic development (Table 1).

We next examined the circadian behavior of Bmal1^GTΔC mice by monitoring wheel-running activities. To directly compare with Bmal1 null mutants, we also examined the behavior of Bmal1^+/- and Bmal1^-/- mice. Bmal1^+/GTΔC mice showed unstable circadian periodicity for ~20 days under the constant darkness (DD) condition. Compared with WT and Bmal1^+/- mice (Fig 2A), 90% (10/12) of the Bmal1^+/GTΔC mice showed arrhythmic behaviors after ~20 days in DD. The remaining mice also lost their rhythms after showing extremely long FRP (S2A Fig). Initially, Bmal1^+/GTΔC mice did not have significantly altered FRP during first 5 days in DD. However, their FRPs were lengthened continuously followed by rhythm loss. The FRP in the last 5 days before losing the rhythms was significantly increased (Fig 2B). However, the amplitude of Bmal1^+/GTΔC mice in the last 5 days was not significantly changed (S2C Fig). These unusual circadian behaviors of Bmal1^+/GTΔC mice were not found in WT and Bmal1^+/- mice (Fig 2A). As expected, Bmal1^GTΔC/GTΔC immediately became arrhythmic under DD. Furthermore, Bmal1^-/GTΔC mice also lost behavioral rhythms (S2B Fig). Interestingly, the total locomotor activities of Bmal1^GTΔC/GTΔC mice were comparable to those of WT mice, whereas Bmal1^-/- mice manifested significantly decreased activities regardless of LD cycles as reported previously (Fig 1C and S2D Fig) [9]. These results strongly suggested that the truncation of BMAL1 C-terminal region is sufficient to disrupt circadian rhythm, which may involve a different mechanism compared to the null mutant mice.

To understand the arrhythmic circadian behavior of Bmal1^GTΔC mice, we examined the mRNA and protein expression profiles of clock genes. As shown in Fig 3A, Per2 mRNA expression of WT mice showed peak expression at CT12 in the SCN, while reaching nadir expression at CT24. These patterns were also observed in Bmal1^+/GTΔC mice. However, Per2 mRNA expression was arrhythmic in Bmal1^GTΔC/GTΔC mice and remained at low levels. Interestingly, although Bmal1 mRNA expression was rhythmic in both WT and Bmal1^+/GTΔC mice, greater levels were observed in Bmal1^+/GTΔC mice. The increase in mRNA levels was more pronounced in Bmal1^GTΔC/GTΔC mice that showed arrhythmic expression.

To further substantiate and quantify these patterns, we examined the mRNA expression profiles of several clock genes in the liver. The oscillation of Per1 and Per2 mRNA was abrogated in Bmal1^GTΔC/GTΔC mice, but there was no difference in their rhythmic expression between WT and Bmal1^+/GTΔC mice (Fig 3B). However, those of Rev-erbα and Dbp mRNA were significantly decreased in both heterozygous and homozygous mutant mice. Interestingly, the levels of Bmal1 and Cry1 mRNAs were increased in both Bmal1^+/GTΔC and Bmal1^GTΔC/GTΔC mice, whereas Clock mRNA expression showed no difference (two-way ANOVA with Tuckey’s post-tests, **p<0.01). Furthermore, these observed increases were coupled with decreased Rev-erbα mRNA expression. In light of the reduced clock gene mRNA levels in Bmal1^-/- mice [3], these results implied that the rhythm disruption in Bmal1^GTΔC mice entails a different mechanism compared with the null mutants.

In addition to mRNA profiles, we also examined levels of clock proteins in the liver. As expected, β-GEO proteins were only found in Bmal1^+/GTΔC and Bmal1^GTΔC/GTΔC mice, and the size of BMAL1^GTΔC proteins was slightly reduced compared with BMAL1^wt. Despite robust circadian behavior during the first several days in DD, Bmal1^+/GTΔC mice did not exhibit robust oscillations of BMAL1^wt and BMAL1^GTΔC proteins compared with WT and Bmal1^+/- mice. In accord with behavioral and mRNA data, there was no circadian oscillation of BMAL1^GTΔC proteins in Bmal1^GTΔC/GTΔC mice, and no BMAL1 protein was detected in Bmal1^-/- mice. In contrast to the increased Bmal1 mRNA levels of the liver of Bmal1^+/GTΔC and Bmal1^GTΔC/GTΔC mice, there were only subtle changes of BMAL1^wt and BMAL1^GTΔC protein levels (S3A Fig). Furthermore, no difference in CLOCK level was observed among the different mouse genotypes, consistent with its mRNA patterns (Figs 3B and 4 and S3A Fig).

The above data from locomotor activities, mRNA and protein expressions clearly demonstrated circadian arrhythmicity of Bmal1^GTΔC/GTΔC mice. To determine whether the arrhythmicity resulted from disrupted molecular clock or deficits at in vivo systemic level, we generated Per2::luc knock-in MEFs harboring WT, Bmal1^+/GTΔC, Bmal1^GTΔC/GTΔC, Bmal1^+/- and Bmal1^-/- alleles. Despite the fact that Bmal1^+/GTΔC mice gradually lost the circadian locomotor rhythm, the rhythms of PER2::LUC expression were maintained. The period length of Bmal1^+/GTΔC mice was significantly decreased, but there was no change in the amplitude, compared with those of WT and Bmal1^+/- MEFs (Fig 5B and S3B Fig). As expected, the circadian oscillation in Bmal1^GTΔC/GTΔC MEFs was disrupted to similar degrees as in Bmal1^-/- MEFs, indicating cellular defects in rhythm generation (Fig 5). To substantiate these results, we introduced a BMAL1^GTΔC expressing vector to WT MEFs, and examined circadian oscillation of Per2 and Bmal1 promoter activity by using the dual-color luciferase technique [23]. Using this experimental system, we also tested N- and C-terminal truncated Bmal1 constructs (Fig 6A). The overexpression of Bmal1^GTΔC and Bmal1^ΔC clearly disrupted circadian rhythms in WT MEFs, whereas no change was observed in WT and Bmal1^ΔN transfected cells (Fig 6B). These results indicated that the loss of rhythms in Bmal1^GTΔC/GTΔC mice resulted from the defective molecular clock machinery at the cellular and molecular level.

Next, we examined the transcriptional activity of BMAL1^GTΔC on Per1 promoter. Overexpression of BMAL1^GTΔC with CLOCK failed to activate Per1 promoter activity. Interestingly, co-transfection of Bmal1^wt and Bmal1^GTΔC partially blocked WT CLOCK:BMAL1 activity, suggesting competition between intact and mutant BMAL1 as reported previously (Fig 7A) [10]. To determine whether the negative role of Bmal1^GTΔC allele resulted from protein cellular localization, we examined sub-cellular expression of BMAL1^wt and BMAL1^GTΔC in combination with CLOCK. Either alone or co-expressed with CLOCK, the fluorescent protein tagged BMAL1^GTΔC protein showed similar expression with BMAL1^wt (Fig 7B). We also did not observe any difference in the dimerization with CLOCK in BiFC assays (Fig 7C). To analyze these results quantitatively, we performed IP experiments. As shown in Fig 7D, CLOCK proteins in BMAL1^GTΔC-transfected cells showed significantly higher input and IP levels, perhaps as a result of activity-dependent degradation of the CLOCK: BMAL1^GTΔC dimer or uneven expression levels. To distinguish these possibilities, we conducted dose-dependency test [25]. The level of CLOCK was rapidly decreased with increased BMAL1^wt amount, as reported previously (Fig 7E). However, increased BMAL1^GTΔC expression did not attenuate the level of CLOCK. These results are consistent with the compromised transcriptional activation as shown in Fig 7A and the previous report [26].

All relevant data are within the paper and its Supporting Information files.