Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Jun 29.
Published in final edited form as: J Biol Rhythms. 2018 Jun 11;33(4):363–375. doi: 10.1177/0748730418779958

Rapid Response and Slow Recovery of the H3K4me3 Epigenomic Marker in the Liver after Light-mediated Phase Advances of the Circadian Clock

Dmytro Grygoryev *,1, Michael R Rountree †,1, Furaha Rwatambuga *, Anna Ohlrich *, Ayaka Kukino *, Matthew P Butler *,‡, Charles N Allen *,‡, Mitchell S Turker *,§,2
PMCID: PMC9241578  NIHMSID: NIHMS1816743  PMID: 29888643

Abstract

Mammalian tissues display circadian rhythms in transcription, translation, and histone modifications. Here we asked how an advance of the light-dark cycle alters daily rhythms in the liver epigenome at the H3K4me3 (trimethylation of lysine 4 on histone 3) modification, which is found at active and poised gene promoters. H3K4me3 levels were first measured at 4 time points (zeitgeber time [ZT] 3, 8, 15, and 20) during a normal 12L:12D light-dark cycle. Peak levels were observed during the early dark phase at ZT15 and dropped to low levels around lights-on (ZT0) between ZT20 and ZT3. A 6-h phase advance at ZT18 (new lights-on after only 6 h of darkness) led to a transient extension of peak H3K4me3 levels. Although locomotor activity reentrained within a week after the phase advance, H3K4me3 rhythms failed to do so, with peak levels remaining in the light phase at the 1-week recovery time point. Eight weekly phase advances, with 1-week recovery times between each phase advance, further disrupted the H3K4me3 rhythms. Finally, we used the mPer2Luc knockin mouse to determine whether the phase advance also disrupted Per2 protein expression. Similar to the results from the histone work, we found both a rapid response to the phase advance and a delayed recovery, the latter in sync with H3K4me3 levels. A model to explain these results is offered.

Keywords: phase advance, circadian epigenome, histone methylation, peripheral tissue, circadian disruption

A network of circadian clocks in the body governs daily rhythms in physiology and behavior (Maywood et al., 2007; Barclay et al., 2012; Mohawk et al., 2012). Disruptions of these clocks—for example, by shift work, jetlag, or social jetlag—cause increased body mass index, insulin resistance, and impaired glucose tolerance in humans (Karlsson et al., 2001; Suwazono et al., 2006; Dochi et al., 2008; Suwazono et al., 2008; Pan et al., 2011; Buxton et al., 2012; Monk and Buysse, 2013; Leproult et al., 2014; Vimalananda et al., 2015). In animal models, environmental or genetic disruptions of the circadian clock likewise promote weight gain and compromise glucose tolerance and insulin sensitivity (Laposky et al., 2005; Turek et al., 2005; Scheer et al., 2009; Marcheva et al., 2010; Karatsoreos et al., 2011; Leproult et al., 2014). Important questions in circadian biology research, therefore, are how clock disruptions affect daily rhythms and how they lead to pathological outcomes.

The principal entraining cue for circadian rhythms is light transmitted directly from the retina to the suprachiasmatic nucleus (SCN) (Golombek and Rosenstein, 2010; Besharse and McMahon, 2016). Neural and humoral outputs from the SCN, as yet poorly defined, in turn entrain local clocks in other brain areas and in peripheral tissues of the body (Dibner et al., 2010). The normal temporal organization of clocks within the body is perturbed during phase shifting as the system reentrains to a new light-dark cycle. An advance of the light-dark cycle immediately begins to shift the clock in the SCN, but peripheral oscillators lag behind, creating temporary desynchrony among organ systems (Yamazaki et al., 2000; Davidson et al., 2009; Yamaguchi et al., 2013). In general, the SCN and locomotor activity reentrain faster than peripheral organ systems after a phase advance or phase delay (Yamazaki et al., 2000; Davidson et al., 2009; Kiessling et al., 2010; Yamaguchi et al., 2013; Tahara et al., 2017).

Circadian changes in histone modifications are collectively termed the circadian epigenome (Masri and Sassone-Corsi, 2010; Koike et al., 2012; Aguilar-Arnal and Sassone-Corsi, 2013); however, it is unknown how and even whether rhythms in the epigenome react to and then reentrain to phase shifts. The trimethylation of lysine 4 on the histone H3 protein subunit (H3K4me3) is an important histone modification found predominantly near active and poised promoters (Vermeulen and Timmers, 2010; Le Martelot et al., 2012), and its levels change rhythmically in the liver during a daily cycle (Koike et al., 2012; Le Martelot et al., 2012; Valekunja et al., 2013). We therefore tested how H3K4me3 rhythms respond to a 6-h phase advance, both immediately and after 1 week. These experiments were repeated in the mPer2Luc knockin mouse model (Yoo et al., 2004), which expresses the functional Per2 protein fused to a reporter luciferase construct, to determine whether H3K4me3 rhythms would parallel Per2 expression rhythms in the liver. Finally, we tested whether repeated phase shifts would further compromise H3K4me3 rhythms because extended circadian disruption is associated with poor outcomes (Karatsoreos et al., 2011).

MATERIALS AND METHODS

General Animal Information

The OHSU Institutional Animal Care and Use Committee approved all experimental procedures involving animals, and all efforts were made to minimize pain and the number of animals used. All mice were housed in 12-h light/12-h dark cycles (LD) in polycarbonate cages with pelleted cellulose bedding (BioFresh Performance Bedding; Absorption Corp., Ferndale, WA). Times are reported in zeitgeber time (ZT), with ZT0 and ZT12 defined as lights-on and lights-off, respectively. Food (LabDiet 5L0D) and water were available ad libitum in all experiments because food restriction can alter daily rhythms (Damiola et al., 2000; Mukherji et al., 2015a; Mukherji et al., 2015b).

Animals and Groups (H3K4me3 Experiments)

Adult male C57BL/6J mice at 3 months of age were purchased from Jax Labs (Bar Harbor, ME) and allowed to acclimate for 1 month in LD (lights on 0600–1800 h PST) before the start of experiments. Mice were housed 4 to a cage in an environmental chamber (Percival Scientific, Perry, IA) maintained at 20 to 21 °C. Housing light intensity was approximately 300 lux.

H3K4me3 enrichment was measured at 4 time points (ZT3, ZT8, ZT15, and ZT20) in each condition. Four mice were kept in each cage, and all mice from a given cage were used for a single time point (see “Sequence Analysis of DNA Samples for H3K4me3 after ChIP,” below). Four time points were used per experiment; thus 16 mice were used per condition. At the designated collection times, mice were deeply anesthetized with isoflurane (Hospira, Inc., Lake Forest, IL) and killed by cervical dislocation; the eyes were removed to prevent any postmortem light effect, and the livers were removed for processing.

Mice (N = 80) were evenly assigned to the following treatments:

Baseline 1: Upon arrival, half of the mice were housed in LD (lights-on 0600-1800 h) and half in a reversed cycle (1800-0600 h) to facilitate tissue collection at all time points during normal work hours. After 1 month of acclimation, livers were collected as above. Reversed light cycles were not used thereafter, and all mice acclimated in LD (0600-1800 h) for a month before treatments commenced.

Baseline 2: After acclimation in LD, livers were collected around the clock at the ZTs above.

Immediate response: The light cycle was advanced 6 h by turning on the lights at ZT18. Livers were collected over the next 24 h (new ZT3, 8, 15, and 20, corresponding to the old ZT21, 2, 9, and 14, respectively).

Recovery: Livers were collected over 24 h at 1 week after a 6-h phase advance of the light cycle.

Chronic: The light cycle was advanced by imposing a short 6-h dark period once per week for 8 weeks. One week after the final advance, livers were collected around the clock.

Animals and Groups (PER2::LUC Experiments)

Male and female mPer2Luc knockin mice on a C57BL/6 background (Yoo et al., 2004) were purchased from Jackson Laboratories (B6.129S6-Per2tm1Jt/J, Strain Code: 006852), bred locally, and housed 3 to 5 per cage in LD in light-tight chambers (Phenome Technologies) with light provided by monochromatic LEDs (130 lux; λpeak 525 nm, 25-nm half band width). Background dim red light remained on continuously (0.2 lux; λpeak 625 nm, 25-nm half band width).

Mice were assigned to 2 separate treatments. The recovery from a phase advance was studied in 17 mice (4 males, 13 females). Baseline rhythms were measured in LD (lights-on 0600-1800 h). One week later, the light cycle was advanced by 6 h, and the recovery rhythms were measured after one more week. The immediate response to a phase advance was studied in a crossover design (n = 14 males) so that PER2::LUC bioluminescence rhythms were measured (details below) in each mouse during control conditions and in the 24 h after a 6-h advance. After acclimating for at least 1 month in LD (lights-on 0600-1800 h PST), half of the mice were transferred at 1800 h to an adjacent chamber with lights on 0000-1200 h. PER2::LUC bioluminescence was measured in all mice every 4 h starting at 0400 h (the original ZT20 and the new ZT4). After 3 weeks, the remaining mice were transferred to the advanced light cycle, and PER2::LUC bioluminescence was measured again in both groups, but beginning at the opposite times (ZT4 or ZT20 of the 0000-1200 h light cycle).

Locomotor Activity Recording

General locomotor activity was measured by passive infrared sensors (5-min bins; Vitalview, Minimitter, Bend, OR) from a cage of 4 mice, housed in identical conditions to those used for the H3K4me3 experiments. Time series data were imported into Clocklab (Actimetrics, Wilmette, IL) for analysis of behavioral rhythmicity. Actograms were double-plotted, and the display setting was set to “scaled.”

Chromatin Immunoprecipitation for H3K4me3

Chromatin immunoprecipitation (ChIP) assays were performed as previously reported by others (Nakayama et al., 2000) with modification. Liver tissue from each mouse was cut with a scalpel into small pieces (~ 0.5 mm3 size) and crosslinked with a freshly made 1% formaldehyde solution in PBS with protease inhibitor (PI cocktail tablet, Complete Mini, EDTA-free; Roche, Indianapolis, IN) at a ratio of 10 mL of solution per 1 g of tissue by rotation (Mini-Tube Rotator; Fisher Scientific) at room temperature for 5 min. The reaction was quenched by adding 2.5 M glycine to a final concentration of 0.125 M and rotating at room temperature for another 5 min. The suspension was centrifuged for 5 min at 4 °C in Beckman Coulter GPKR centrifuge (84 g); the precipitate was washed with 10 mL of cold PBS with PI and was spun down, and the supernatant was removed. The tissue was then homogenized on ice with a glass Dounce homogenizer (loose) for cell disruption in 5 mL of PBS with PI, centrifuged 5 min at 162 g at 4 °C, resuspended in 5 mL of Cell Lysis Buffer (5 mM PIPES, 85 mM KCL, 0.5% IGEPAL, 1 PI tablet), and incubated on ice for 15 min. The tissue was then homogenized on ice with Dounce glass homogenizer (tight) with 10 strokes. The suspension was centrifuged for 5 min at 2000 g at 4 °C, the supernatant was removed, and isolated nuclei were checked under a microscope and then frozen at −80 °C until sonicated.

Isolated nuclei were resuspended in SDS Shearing Buffer (10 mM Tris HCl, 0.25% SDS, 1 mM EDTA, 1× Halt protease inhibitor cocktail [Thermo Scientific, Rockford, IL, USA]), approximately 0.15 mg of tissue per 1 μL of the buffer, and sonicated for 10 min (duty cycle 2%, intensity 3, peak incident power 105 W, cycles per burst 200). Initial experiments were performed with isolated DNA to ensure a size range of approximately 200 to 700 bp. After sonication, the sheared chromatin was centrifuged 5 min at 4 °C at 10,000 g (Forma Scientific Micromax benchtop centrifuge), and supernatant was transferred to a new tube.

Next, 100 μL of the chromatin prep (approximately 17 μg of DNA) was diluted with 200 μL of 1.5× ChIP buffer (75 mM Tris HCl, 0.21 M NaCl, 1.5 mM EDTA, 1.5% Triton-X, 0.15% deoxycholate) with PI, and 3 μL of each sample was saved as input. Then 3 μL of histone H3 trimethyl Lys4 antibody (H3K4me3; Active Motif, Carlsbad, CA) was added to each experimental sample, and the samples were incubated overnight at 4 °C with rotation. The chromatin prep from each liver was processed separately.

Magnetic beads (Magna ChIP Protein A+G Magnetic Beads; Millipore), 20 μL per sample, were washed twice with 2 mL of 1× ChIP buffer with PI and returned to an original volume of 20 μL. The magnetic beads were added to each 100-μL sample and incubated for 2 h at 4 °C with rotation. The beads were then washed twice with 1 mL of ChIP buffer (without PI) and incubated for 10 min at 4 °C with rotation, followed by 2 washes with NaCl ChIP buffer (1× ChIP buffer with 0.5 M NaCl), 2 washes with LiCl wash buffer (10 mM Tris HCl, 0.25 M LiCl, 1 mM EDTA, 0.5% NP-40, 0.5% deoxycholate), and 2 washes with TE buffer (10 mM Tris HCl, 1 mM EDTA). After washing, 62.5 μL of TES buffer (50 mM Tris HCl, 1% SDS, 10 mM EDTA) was added to each sample, and the samples were incubated at 65 °C for 10 min with mixing. The supernatants from each sample were saved in screw-top cap vials. The incubation step was then repeated, and the supernatants were combined. TES (113 μL) was added to the input samples (total of 12 μL after combining samples from 4 livers).

Next, 5.2 μL of 5 M NaCl and 6.25 μL of RNase A (4 μg/μL) were added to 125 μL of immunoprecipitated and input samples. The samples were then incubated for 30 to 60 min at 37 °C. After incubation, 5 μL of proteinase K (20 μg/μL) was added to all samples, and the samples were decrosslinked overnight at 65 °C. After decrosslinking DNA from proteins, each DNA sample was purified with the MinElute PCR Purification Kit (Qiagen, Valencia, CA), according to the manufacturer’s protocol.

Sequence Analysis of DNA Samples for H3K4me3 after ChIP

After obtaining liver DNA samples from each mouse, as described above, we first performed a quality control experiment based on promoter region enrichment to ensure that the ChIP had worked for each liver sample. Real-time (RT)-PCR was used to examine the DNA samples from each mouse for H3K4me3 promoter enrichment for the Tcf25 transcription factor gene, relative to an intronic region, which served as a negative control because it showed no enrichment for the H3K4me3 modification. The mouse liver Tcf25 promoter has been shown by others to exhibit a circadian rhythm (Valekunja et al., 2013). We used the reported promoter primer sequences for Tcf25 (GAGGAAGAGGGACCAAAACC; GGTTGTTGACACGGACTCCT) and designed primer sequences for an intronic region upstream of exon 13 (GAGGAAGAGGGACCAAAACC; GGTTGTTGA CACGGACTCCT). For each experimental condition, all ChIPs were performed together, and the DNA samples were sequenced in parallel to reduce background variability. When all 4 DNA samples from a given time point were found to exhibit promoter enrichment (see Suppl. Fig. S1 for a representative example), they were pooled (~5 ng per sample) and sent to the OHSU Massively Parallel Sequencing Shared Resource (MPSSR) for library preparation and sequencing. Libraries were made using the Illumina TruSeq ChIP Library Prep kit (catalog number IP-202-1012), following the manufacturer’s protocol, and 5 to 10 ng of DNA was used as input followed by 16 to 18 rounds of amplification. Samples were run on the HiSeq2500, single read with 100 cycles. The sequencing results were then transmitted from the MPSSR to the authors for bioinformatics analysis, as described below.

Bioinformatics

Data files of H3K4me3 ChIP-seq have been deposited to GEO (http://www.ncbi.nlm.nih.gov/geo/) under the accession number (GSE108481). H3K4me3 ChIP-seq reads were mapped to the mouse genome (version mm10) using Bowtie 2 with default settings (http://bowtie-bio.sourceforge.net/bowtie2/index.shtml) (Langmead and Salzberg, 2012) (Suppl. Table S1). The output files in SAM format were converted to BAM format, PCR duplicates were removed (rmdup tool), the files were sorted, and index files were created using SAMtools (http://samtools.sourceforge.net) (Li et al., 2009). BAM files were converted to read density tiled data files (TDF) for viewing in the IGV genome browser using igvtools (www.broadinstitute.org/igv). BED files were generated from BAM files using BEDTools (http://bedtools.readthedocs.io/en/latest/content/bedtools-suite.html) (Quinlan and Hall, 2010). The BED files were used to identify enriched H3K4me3 domains using the RSEG program with a bin size (-b) of 750 (http://smithlabresearch.org/software/rseg/) (Song and Smith, 2011). Peak annotation was performed using the annotatePeaks.pl script from the HOMER software package (http://homer.ucsd.edu/homer/ngs/annotation.html). Enriched H3K4me3 peaks with a minimum size of 1.5 kb and within 2 kb of an annotated transcription start site (TSS) were kept to create a BED file of 10,719 unique enriched peak regions found in the control baseline samples. This BED file was then used as the reference coordinates in the seqMINER program (http://bips.u-strasbg.fr/seqminer/) (Ye et al., 2011). H3K4me3 read enrichment scores (log2 transformed) were generated for each H3K4me3 ChIP-seq mapped sample (BED file) using the Enrichment Based Method of the seqMINER program and normalized with an IgG ChIP-seq control file (Ye et al., 2011). An IgG ChIP-seq input file was used as the control.

Significant enrichment of H3K4me3 under baseline conditions was detected in the promoter regions of 10,719 genes. Each gene was assigned a baseline oscillation amplitude (maximum minus minimum enrichment across the 4 zeitgeber times). To best visualize the heat map analysis, the 1,072 genes representing the top 10% of Baseline 2 amplitudes were selected and ordered by peak time (27 genes with a peak at ZT8, all others with a peak at ZT15) and then by amplitude. Data from the other conditions (Baseline 1, Immediate, 1 week Recovery, and Chronic) were plotted in the same order. Heat maps were generated by transforming all data (20 data points) to a z-score within a gene and pseudocoloring the output using HeatMapper (Babicki et al., 2016).

Enrichment scores were compared across zeitgeber time and condition by mixed-model analysis with promoter region as a random factor and with condition and time as categorical fixed factors (REML method, JMP Pro 11.0; SAS Institute, Inc., Cary, NC). A significant interaction of condition × time indicates a change in rhythm shape or phase. Data were also converted to z-score across all conditions.

PER2::LUC In Vivo Imaging and Analysis

Clock phase in the liver and submandibular gland was estimated from PER2::LUC bioluminescence measured every 4 h (Tahara et al., 2012). Mice were lightly anesthetized with isoflurane, injected with D-luciferin (Promega, Madison, WI; 15 mg/kg, injected as a sterile 3-mg/mL solution in PBS). The neck and abdomen were shaved prior to the first measurement. Bioluminescence was measured 10 min after injection using an Electron Magnified (EM) CCD camera (ImageEM, Hamamatsu, Japan, controlled by Piper software version 2.6.89.18, Stanford Photonics, Stanford, CA) in an ONYX dark box (Stanford Photonics). Euthermia was maintained by a temperature-controlled stage (mTCII micro-Temperature Controller, Cell MicroControls; Norfolk, VA). Bioluminescence was captured by the camera in EM mode (sum of eight 125-msec exposures, gain 500). A brightfield reference image of the anesthetized mouse was taken each time under dim red light (1 lux, 633 nm, 15 nm half band width). Bioluminescence was scored offline (ImageJ, NIH). Each image was opened in ImageJ, and the intensity of the 24-bit grayscale image was quantified using elliptical regions of interest centered on the brightest areas of the tissue (liver: 20 mm × 15 mm; submandibular gland: 11 mm × 11 mm).

Bioluminescence rhythms were assessed by mixed-model linear regression with zeitgeber time and condition (control or advanced) as fixed factors and mouse as a random factor as each mouse was its own control. A significant interaction of time and condition was interpreted as a change in the shape of the rhythm. At each time point, pairwise comparisons were conducted by paired t test. All tests were conducted in JMP Pro 11.0.

RESULTS

H3K4me3 Rhythms under a Normal (Baseline) 12:12 LD Regimen

The first goal of this project was to use chromatin immunoprecipitation sequencing (ChIP-seq) for H3K4me3 to measure the effects of a light-mediated phase advance in the liver epigenome. Effective H3K4me3 ChIP pull-down was verified prior to sequencing by testing each liver DNA sample for H3K4me3 enrichment at the promoter region of the Tcf25 transcription factor gene by PCR compared with a negative control intronic region. The Tcf25 promoter is enriched for H3K4me3, and it exhibits daily changes for this modification (Valekunja et al., 2013). A representative promoter versus intron comparison for the Baseline 2 samples (see below) is shown in Supplementary Figure S1. The 4 samples per timepoint were then pooled and sequenced at the OHSU Massively Parallel Sequencing Shared Resource.

Baseline rhythms in H3K4me3 levels for mice on undisrupted LD were measured twice, the first time (Baseline 1) with one half of the mice entrained to a dark schedule during our working day and the second time (Baseline 2) with mice that remained on the normal light-dark regimen and were collected around the clock, as for all experiments described below (see Materials and Methods). Both baselines (Fig. 1A, B) showed similar patterns of H3K4me3 enrichment. Relative read density changes for the 10,719 promoter-associated enriched H3K4me3 peaks across the 4 time points were calculated. We then determined the top 10% of genes with the greatest amplitude rhythms (maximum minus minimum H3K4me3 enrichment across the 4 time points, based on Baseline 2 data). A heat map for this analysis shows daily rhythmicity (Fig. 1A, B) with low H3K4me3 levels early in the light phase (ZT3), when mice are less active, followed by increasing H3K4me3 levels that peak during the dark phase (ZT15) when mice are most active. Our observation of peak H3K4me3 levels during the dark phase of an unperturbed daily rhythm is consistent with work by others (Koike et al., 2012; Valekunja et al., 2013). The only difference between the two baseline measurements was at ZT20 because the drop-off from the peak at ZT15 was more rapid for Baseline 2 than for Baseline 1; however, overall the rhythmic patterns for H3K4me3 were similar for the 2 sets of baseline control samples (Fig. 1A, B; also see Suppl. Fig. S2).

Figure 1.

Figure 1.

Open in a new tab

Heat maps of liver H3K4me3 enrichment (top 10% by amplitude) at the zeitgeber times indicated. Genes are ordered by the time of peak and then by oscillation amplitude according to Baseline 2; the same order of genes is preserved in other panels. White and black bars indicate the light-dark cycle for each condition. Colors represent z-score transformation within a gene across all 20 samples shown in A-E. For each condition, H3K4me3 levels were measured at 4 time points: ZT3, ZT8, ZT15, ZT20, and 4 mice were collected per timepoint. The H3K4me3 ChIP was performed for each mouse and, after a quality control check (see Materials and Methods section “Sequence Analysis of DNA Samples for H3K4me3 after ChIP“ and Suppl. Fig. S1), pooled prior to library preparation for sequencing. (A) Baseline 1. derived from animals evenly distributed between a standard and a reverse light-dark cycle. (B) Baseline 2. derived from animals that all remained on a single standard light-dark cycle. (C) Immediate. Mice examined the day immediately following the 6-h phase advance (F shows the actual advance). (D) Recovery. Mice examined after 1 week of recovery from the phase advance. (E) Chronic. Mice examined after 8 weekly phase advances, followed by an additional 1-week recovery. (F) The baseline and immediate response of a subset of 200 genes (indicated by “F” in C and D) are replotted together to show the immediate effect of light. Top: Baseline 2 data are plotted twice for better visualization of the rhythm. Bottom: The ZT3, ZT8, and ZT15 time points of Baseline 2 and the new ZT3, ZT8, ZT15, and ZT20 of the immediate response are shown. Note that instead of dropping, high levels of H3K4me3 are maintained after the lights are turned on at ZT18.

Supplementary Figure S3 shows the enrichment density heat map data for all transcription start sites (TSS) associated with enriched H3K4me3 peaks for Baseline 2 and confirms significant clustering of the H3K4me3 around the TSS. Additional evidence for the quality of the ChIP-seq analysis was visualized by plotting all H3K4me3 peaks relative to promoter for the Baseline 2 data and observing bimodal peaks surrounding the TSS (Suppl. Fig. S4), as is commonly observed (Quinodoz et al., 2014; Hansen et al., 2015). A genome browser view for a region of chromosome 5 demonstrated reproducible H3K4me3 peaks localized to gene promoters in all experiments for a given region of the genome (Suppl. Fig. S5). Supplementary Table S1 provides the total mapped reads for Baselines 1 and 2 (and the other experiments discussed below), and Supplementary Table S2 lists all 10,719 promoter associated enriched peaks ranked highest to lowest for H3K4me3 read density for the Baseline 2 samples at ZT15.

H3K4me3 Rhythms after a Single 6-h Phase Advance

H3K4me3 levels were measured in the first day of the new LD cycle after a 6-h advance (Fig. 1C, new ZT3, 8, 15, and 20, labeled “Immediate”), with the new ZT0 (lights-on) at the prior ZT18 (see Fig. 1F). If the phase-advanced lights-on lacked a rapid effect on H3K4me3 rhythmicity, we would expect a drop in H3K4me3 levels at the new ZT3 (equivalent in time to ZT21 if the phase advance had not occurred), consistent with the drop that occurs from ZT15 to ZT20 during an unperturbed cycle (Fig. 1A, B). Instead, H3K4me3 levels remained high for at least 3 h after the phase advance (ZT3 in Fig. 1C) demonstrating an acute effect of light on the liver circadian epigenome. This effect is best observed by merging the Baseline and Immediate heat maps in Fig. 1F, which shows little change in H3K4me3 levels from the Baseline ZT15 to the new ZT3 6 h later (also see Suppl. Fig. S2).

The extended peak in H3K4me3 levels after the phase advance was not followed by a systematic delay, however, of the rhythm. H3K4me3 levels increased again at the new ZT20, similar to the rise in H3K4me3 levels that would occur without the phase advance, although not to the same extent (Fig. 1F, compared with previous ZT15; Suppl. Fig. S2). This suggests that light during the dark phase may have a positive masking effect on H3K4me3 levels but a smaller phase-shifting effect. As a result, however, the rise in H3K4me3 levels occurred toward the end of the dark phase of the light-dark cycle (i.e., shortly before the lights turned on anew), as opposed to earlier in the dark phase during a normal daily cycle.

We therefore asked whether the liver circadian epigenome would reentrain to the light-dark cycle after a week (Recovery Group), a time point shown by others (LeGates et al., 2009; Kiessling et al., 2010; Summa et al., 2012) to be sufficient for behavioral reentrainment, as defined by locomotor activity rhythms. We verified in identical conditions that locomotor activity reentrains within a week after a 6-h phase advance (Fig. 2 for group-housed mice; similar results were observed for singly housed mice, not shown). However, most H3K4me3 levels still peaked during the light phase at ZT3 or ZT8 (Fig. 1D; Suppl. Fig. S2). This result shows that the H3K4me3 rhythms failed to reentrain by 1 week after the phase advance despite full behavioral reentrainment, as measured by locomotor activity.

Figure 2.

Figure 2.

Open in a new tab

Double-plotted actogram of group-housed mice. Hatch marks along each line show locomotor activity as measured by passive infrared sensors mounted above the cage. Successive days are plotted from top to bottom. Yellow and gray indicate lights-on and lights-off, respectively. After a phase advance accomplished by turning on lights at ZT18, locomotor activity reentrains in 1 week. Red line through onsets shows the pattern of reentrainment on the right.

Phase Advance Effect on H3K4me3 Enrichment at Clock Gene Promoters

We further investigated this rapid response and delay in recovery by analyzing the data from 6 clock genes: Per1, Per2, Cry1, Cry2, Bmal1, and Clock. All 6 responded immediately to the early lights-on and maintained higher levels than would have occurred in darkness (Fig. 3A). One week after the advance, the rhythms were still out of phase with Baseline conditions, now peaking 6 to 12 h earlier than predicted (Fig. 3B). Peak H3K4me3 levels still occurred during the light phase instead of during the dark phase. H3K4me3 levels for the Cry1, Cry2, Bmal1, and Clock promoters at 1 week showed less change from the peak to trough levels than observed for the unperturbed samples and in each case resembled the bulk patterns observed with the heat maps (Fig. 1D and Suppl. Fig. S2). Similar results to the sequence data were provided by a qRT-PCR analysis for H3K4me3 levels at the Per2 and Bmal1 promoters using DNA from a second round of H3K4me3 ChIP from the liver samples (Suppl. Fig. S6).

Figure 3.

Figure 3.

Open in a new tab

Immediate and 1-week effects for H3K4me3 levels at clock promoters from a single 6-h phase advance. (A) Baseline rhythms are plotted twice (black) to provide a visual context for the immediate response to a phase shift (red). The baseline LD cycle is shown by white/black bars, and the new shifted schedule is shown by white/red bars. All data are plotted by baseline zeitgeber time. (B) Daily rhythms of H3K4me3 enrichment at clock gene promoter regions at Baseline (black) and 1 week after the 6-h phase advance (red). The H3K4me3 rhythm fails to reentrain to the light-dark cycle 1 week after the phase advance. The data are plotted twice to illustrate the rhythms.

H3K4me3 Rhythms after Multiple 6-h Phase Advances

The lack of full recovery for H3K4me3 rhythmicity after a week suggested a surprisingly long-lasting effect of H3K4me3 rhythms after a single advance. Importantly, it also suggested that in chronic phase-shifting experiments, each new shift impinges on a circadian epigenome that is still out of sync with the prevailing light-dark cycle, even if behavior has recovered. Therefore, we measured H3K4me3 rhythms after multiple weekly phase advances, a paradigm that is linked to a variety of pathologies (Knutsson, 2003; Davidson et al., 2006; Martino et al., 2008; Scheer et al., 2009; Marcheva et al., 2010; Lahti et al., 2012; Evans and Davidson, 2013; Moller-Levet et al., 2013; West and Bechtold, 2015; Kettner et al., 2016). Two relevant observations were made. The first observation was a clearly disrupted daily rhythm for H3K4me3, with a delayed peak compared with baseline and much greater variability in phase among different gene promoters (compare Fig. 1E with Figs. 1A and B). The second observation was a dampening in rhythm amplitude for H3K4me3 levels (see Suppl. Fig. S2). On a final note, when the data from the different experiment conditions were analyzed together by mixed-model regression, with gene promoter as a random factor, significant effects of condition (F4, 20026 = 1726, p < 0.001), ZT (F3, 20026 = 521, p < 0.001), and their interaction (F12, 20026 = 2120, p < 0.001) were observed.

The Effects of a Phase Advance on PER2::LUC Protein Expression

The sequencing (Fig. 3) and confirmatory PCR data (Suppl. Fig. S6) for H3K4me3 levels at the Per2 promoter showed a slowed recovery from the phase advance, relative to locomotor activity. To determine whether recovery of the daily rhythm for the Per2 protein was also delayed, we repeated the 1-week recovery experiment using mPer2Luc knockin mice. Figure 4A shows the daily rhythm for PER2::LUC bioluminescence in the liver and submandibular gland of an individual mouse before and a week after the phase advance. Under baseline conditions, liver PER2::LUC rose during the dark phase. One week after the advance, PER2::LUC rhythms were altered, with an earlier rise and peak compared with baseline (Fig. 4A, B; significant effects of condition: F1,176 = 54, p < 0.001; time: F5,176 = 127, p < 0.001; and their interaction: F5,176 = 17, p < 0.001). Further, peak levels were higher in the postshift mice at ZT14, 18, and 22 (t16 > 2.8, p < 0.05, for all). We also note that in both the preshift and postshift conditions, the peak of PER2::LUC protein expression lagged behind the peak of H3K4me3 levels at the Per2 promoter, suggesting a retention of this relationship. Similar patterns were observed in both males and females (Suppl. Fig. S7).

Figure 4.

Figure 4.

Open in a new tab

PER2::LUC bioluminescence in the liver and submandibular gland. (A) Representative images of a single mouse measured 1 week before and 1 week after a 6-h phase advance of the light-dark cycle. PER2::LUC bioluminescence is pseudocolored and plotted on the reference images of the anesthetized mouse taken under dim red light. S, submandibular gland; L, liver. (B) Liver and (C) submandibular gland (SMG) PER2::LUC bioluminescence rhythms (mean±SE, n = 17) plotted against zeitgeber time for baseline conditions (black line) and 1 week (dashed orange line) after the phase advance. (D) Liver and (E) submandibular gland PER2::LUC bioluminescence measured in baseline LD conditions (black) and in the first 24 h after a 6-h phase advance (dashed red) in 14 mice. For illustration, the control conditions are replotted to the left in gray to indicate the expected baseline conditions prior to the experiment. *significant differences from paired t test, p < 0.05.

For a second tissue, we measured PER2::LUC bioluminescence in the submandibular gland because it has a daily rhythm similar to that of the liver under LD conditions in the mPer2Luc mouse (Tahara et al., 2012; Hamaguchi et al., 2015a; Hamaguchi et al., 2015b), although daily rhythms for this gland are sensitive to different entrainment cues (Vujovic et al., 2008). As for liver, there were significant effects of condition (F1,176 = 12, p < 0.001), time (F5,176 = 102, p < 0.001), and their interaction (F5,176 = 4.3, p = 0.001). But there was no change in the timing of peak levels pre- and postshift for the submandibular gland, although the peak was higher postshift (t16 = 3.8, p = 0.001; Fig. 4A, C). Similar patterns were observed in both males and females (Suppl. Fig. S7).

For a final experiment, we tested whether Per2 protein levels respond rapidly after the phase advance, as we showed for the H3K4me3 modification. In this crossover experiment, liver PER2::LUC bioluminescence was significantly higher 4, 8, 12, and 20 h after lights-on in the shifted mice (Fig. 4D: paired t test, t13 ≥ 2.4, p < 0.05 for all). Overall, there were significant effects of condition (F1,143 = 10.7, p < 0.01), time (F5,143 = 54, p < 0.001), and their interaction (F5,143 = 4.0, p < 0.01). In contrast to the liver, no immediate response was detectable in the submandibular gland, although shifted mice had lower PER2::LUC bioluminescence at the last time point measured (Fig. 4E: t13 = 3.5, p < 0.01). This contributed to a significant interaction of condition by time (F5,143 = 3.8, p < 0.01).

DISCUSSION

Our results demonstrate that a circadian histone modification, H3K4me3, in the liver responded rapidly to a light-mediated circadian disruption. In contrast to this immediate response to light at night, reentrainment was slow. One week after the phase advance, peak H3K4me3 levels were significantly out of phase with the normal daily rhythm despite the mice displaying reentrained locomotor activity rhythms. Consistent with the lagging recovery of the H3K4me3 levels, we also found incomplete recovery for rhythmic expression of PER2::LUC in the liver 1 week after the phase advance that remained in sync with the incomplete recovery of H3K4me3. Recovery of liver Per2 mRNA and PER2::LUC expression has previously been shown to take more than a week to recover after an 8-h phase advance (Yamaguchi et al., 2013; Tahara et al., 2017); Per1-Luc expression reentrainment is also slower in rat liver as measured ex vivo (Yamazaki et al., 2000). Our additional work with the mPer2Luc knockin mice showed a rapid response to the phase advance, as revealed by increased protein levels in the first 12 h after the advance, which is similar in appearance to the short-term response of the H3K4me3 levels. Others have shown an increase in expression of a luciferase reporter from the Per1 promoter in mouse adrenal glands shortly after a light exposure during the dark phase (Ishida et al., 2005).

Our observations of perturbations at both the epigenomic and protein level from a phase advance bring up a broader point, which is that even under normal conditions, complex relationships exist in the liver (and assuredly other tissues) for the multiple clock components. These components include rhythmic clock proteins binding to DNA, histone modifications (Koike et al., 2012), DNA methylation (Xia et al., 2015; Oh et al., 2018), RNA expression, and clock protein abundance (Panda et al., 2002; Ko and Takahashi, 2006; Koike et al., 2012; Vollmers et al., 2012; Duong and Weitz, 2014; Reinke and Asher, 2016; Sobel et al., 2017), which completes a loop back to DNA binding. Despite these well-documented rhythms, how environmental perturbations like phase shifts affect the relationships among these components is not known. Desynchrony within the cells of a tissue could contribute to the pathological outcomes associated with circadian disruptions (Knutsson, 2003; Davidson et al., 2006; Martino et al., 2008; Scheer et al., 2009; Marcheva et al., 2010; Lahti et al., 2012; Evans and Davidson, 2013; Moller-Levet et al., 2013; West and Bechtold, 2015; Kettner et al., 2016).

Here we showed that Per2 protein levels responded to the light advance with similar kinetics to H3K4me3 levels, both in the immediate response and in the incomplete recovery after 1 week. These apparently similar observations, however, do not necessarily reflect similar mechanisms. The lag between Per2 promoter H3K4me3 and Per2 protein peaks was consistent in the Baseline and Recovery conditions. We note, however, that unexpected complexity could exist even for this mechanism because under normal conditions H3K4me3 changes follow, instead of precede, RNA Polll binding (Koike et al., 2012). In contrast, a transcription-translation mechanism is unlikely to explain the rapid rise in Per2 protein levels shortly after the phase advance. If so, the rapid increase in Per2 protein levels is unrelated to the seemingly similar response for H3k4me3 shortly after the phase advance. Our study, therefore, represents just the first step in determining whether the multilayered components of the circadian clock respond as a unit, and likewise recover as a unit, after a circadian disruption. If the answer is mixed, as suggested by a recent study with a ketogenic diet (Tognini et al., 2017), future work will be required to determine how the circadian epigenome fractures after a circadian disruption, the link between these changes and those for RNA and protein expression, and whether this putative fracturing has a link to disease outcomes.

We propose a model to explain both the rapid response and slow recovery of the H3K4me3 rhythm based on the timing of the transition from dark to light relative to H3K4me3 levels (Fig. 5). The normal daily transition from dark to light (ZT24/ZT0) occurs when H3K4me3 levels are low. In contrast, lights-on for the phase advance occurs at ZT18, a time when H3K4me3 levels are at or near their peak. Our data show that rather than decreasing, H3K4me3 enrichment stays at relatively high levels (see Fig. 1F), suggesting an acute effect of light on H3K4me3 levels, potentially via the SCN’s autonomic efferent pathways, as observed in the adrenal clock’s response to light (Kiessling et al., 2010). To explain the slow recovery, we hypothesize that the daily acute stimulation of H3K4me3 after the phase advance continues to occur when its levels remain high, instead of the normal low levels at ZT0. This mismatch would extend the peak of the H3K4me3 rhythm and thereby delay the circadian epigenome instead of advancing it (Fig. 5). Indeed, 1 week after the phase advance, the H3K4me3 rhythm may be reentraining by delay. An ordered advance of the peak would draw it from ZT21 (original ZT15) to ZT15. Instead, the Recovery peak occurs at ZT3.

Figure 5.

Figure 5.

Open in a new tab

Schematic model of overall H3K4me3 enrichment rhythms showing a delaying effect of light exposure. In normal conditions, H3K4me3 enrichment is at a minimum at the beginning of the light phase. If light, via as-yet unknown pathways, has a stimulatory effect (arrows) on H3K4me3 methylation, the model predicts that this effect would extend the peak when lights are turned on in the middle of the night, thereby delaying the rhythm. The black trace shows the expected stable rhythm in unperturbed conditions; the red dashed line shows the hypothetical progressive delays in response to the advanced LD cycle.

An interesting aspect of this model is that the time and duration of a phase advance could play an important role in the length of time required for full recovery of H3K4me3 rhythms. In mouse liver, our data suggest entrainment by delays after a 6-h advance in both PER2::LUC and H3K4me3 rhythms (in 2 separate groups of mice). In contrast, the liver appears to reentrain via advances to an 8-h advance as reported by PER2::LUC (Tahara et al., 2017) or Per2 mRNA rhythms (Yamaguchi et al., 2013). The phase and period response of activity rhythms depend on both the timing and duration of a light pulse (Comas et al., 2006).

The extent to which food consumption affects baseline and reentrainment of the circadian epigenome is not known, and in our experiment food consumption during the phase shifts was not measured. The timing of food availability is known to play an important role in regulating the liver clock, including the observation that feeding restricted to the light phase will cause the liver clock to “flip” to being most active during this time (Damiola et al., 2000; Mukherji et al., 2015a). Thus, if the mice continue to eat out of phase despite being provided food ad libitum, this could contribute to the delayed recovery. Future work will be necessary to identify the specific roles of both light and food in slowing recovery of H3K4me3 rhythms, as well as other epigenetic modifications.

Finally, we note that while our study demonstrates that the liver circadian epigenome can respond to perturbed lighting conditions, it does not address the signal that is transmitted presumably from the SCN to the liver. Transmission via autonomic nervous system (ANS) projections is a good candidate mechanism given similar acute effects of light on adrenal oscillators by the ANS (Ishida et al., 2005) and evidence for a multisynaptic ANS pathway between the SCN and the liver (Buijs et al., 2003). Additional hormonal or behavioral pathways may contribute to entrainment of the circadian clock in liver cells (Guo et al., 2005; Guo et al., 2006). Because light during the dark phase can modify the liver’s epigenome, this system may allow us to dissect the action and mechanisms by which the SCN entrains the circadian epigenome.

Supplementary Material

suppl_material1
suppl_material2

ACKNOWLEDGMENTS

This work was supported by Oregon Institute of Occupational Sciences pilot funds (M.S.T., C.N.A., and M.P.B.) and NIH RO1-NS036607 (C.N.A.). Illumina sequencing was performed by the OHSU Massively Parallel Sequencing Shared Resource with valuable assistance from Dr. Robert Searles and Ms. Amy Carlos.

Footnotes

CONFLICT OF INTEREST STATEMENT

The author(s) have no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Supplemental material is available for this article online.

REFERENCES

  1. Aguilar-Arnal L and Sassone-Corsi P (2013) The circadian epigenome: how metabolism talks to chromatin remodeling. Curr Opin Cell Biol 25:170–176. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Babicki S, Arndt D, Marcu A, Liang Y, Grant JR, Maciejewski A, and Wishart DS (2016) Heatmapper: web-enabled heat mapping for all. Nucleic Acids Res 44:W147–153. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Barclay JL, Tsang AH, and Oster H (2012) Interaction of central and peripheral clocks in physiological regulation. Prog Brain Res 199:163–181. [DOI] [PubMed] [Google Scholar]
  4. Besharse JC and McMahon DG (2016) The retina and other light-sensitive ocular clocks. J Biol Rhythms 31:223–243. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Buijs RM, la Fleur SE, Wortel J, Van Heyningen C, Zuiddam L, Mettenleiter TC, Kalsbeek A, Nagai K, and Niijima A (2003) The suprachiasmatic nucleus balances sympathetic and parasympathetic output to peripheral organs through separate preautonomic neurons. J Comp Neurol 464:36–48. [DOI] [PubMed] [Google Scholar]
  6. Buxton OM, Cain SW, O’Connor SP, Porter JH, Duffy JF, Wang W, Czeisler CA, and Shea SA (2012) Adverse metabolic consequences in humans of prolonged sleep restriction combined with circadian disruption. Sci Transl Med 4:129ra143. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Comas M, Beersma DG, Spoelstra K, and Daan S (2006) Phase and period responses of the circadian system of mice (Mus musculus) to light stimuli of different duration. J Biol Rhythms 21:362–372. [DOI] [PubMed] [Google Scholar]
  8. Damiola F, Le Minh N, Preitner N, Kornmann B, Fleury-Olela F, and Schibler U (2000) Restricted feeding uncouples circadian oscillators in peripheral tissues from the central pacemaker in the suprachiasmatic nucleus. Genes Dev 14:2950–2961. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Davidson AJ, Castanon-Cervantes O, Leise TL, Molyneux PC, and Harrington ME (2009) Visualizing jet lag in the mouse suprachiasmatic nucleus and peripheral circadian timing system. Eur J Neurosci 29:171–180. [DOI] [PubMed] [Google Scholar]
  10. Davidson AJ, Sellix MT, Daniel J, Yamazaki S, Menaker M, and Block GD (2006) Chronic jet-lag increases mortality in aged mice. Curr Biol 16:R914–916. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dibner C, Schibler U, and Albrecht U (2010) The mammalian circadian timing system: organization and coordination of central and peripheral clocks. Annu Rev Physiol 72:517–549. [DOI] [PubMed] [Google Scholar]
  12. Dochi M, Sakata K, Oishi M, Tanaka K, Kobayashi E, and Suwazono Y (2008) Relationship between shift work and hypercholesterolemia in Japan. Scand J Work Environ Health 34:33–39. [DOI] [PubMed] [Google Scholar]
  13. Duong HA and Weitz CJ (2014) Temporal orchestration of repressive chromatin modifiers by circadian clock Period complexes. Nat Struct Mol Biol 21:126–132. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Evans JA and Davidson AJ (2013) Health consequences of circadian disruption in humans and animal models. Prog Mol Biol Transl Sci 119:283–323. [DOI] [PubMed] [Google Scholar]
  15. Golombek DA and Rosenstein RE (2010) Physiology of circadian entrainment. Physiol Rev 90:1063–1102. [DOI] [PubMed] [Google Scholar]
  16. Guo H, Brewer JM, Champhekar A, Harris RB, and Bittman EL (2005) Differential control of peripheral circadian rhythms by suprachiasmatic-dependent neural signals. Proc Natl Acad Sci U S A 102:3111–3116. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Guo H, Brewer JM, Lehman MN, and Bittman EL (2006) Suprachiasmatic regulation of circadian rhythms of gene expression in hamster peripheral organs: effects of transplanting the pacemaker. J Neurosci 26:6406–6412. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Hamaguchi Y, Tahara Y, Hitosugi M, and Shibata S (2015a) Impairment of circadian rhythms in peripheral clocks by constant light is partially reversed by scheduled feeding or exercise. J Biol Rhythms 30:533–542. [DOI] [PubMed] [Google Scholar]
  19. Hamaguchi Y, Tahara Y, Kuroda H, Haraguchi A, and Shibata S (2015b) Entrainment of mouse peripheral circadian clocks to <24 h feeding/fasting cycles under 24 h light/dark conditions. Sci Rep 5:14207. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hansen P, Hecht J, Ibrahim DM, Krannich A, Truss M, and Robinson PN (2015) Saturation analysis of ChIP-seq data for reproducible identification of binding peaks. Genome Res 25:1391–1400. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Ishida A, Mutoh T, Ueyama T, Bando H, Masubuchi S, Nakahara D, Tsujimoto G, and Okamura H (2005) Light activates the adrenal gland: timing of gene expression and glucocorticoid release. Cell Metab 2:297–307. [DOI] [PubMed] [Google Scholar]
  22. Karatsoreos IN, Bhagat S, Bloss EB, Morrison JH, and McEwen BS (2011) Disruption of circadian clocks has ramifications for metabolism, brain, and behavior. Proc Natl Acad Sci U S A 108:1657–1662. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Karlsson B, Knutsson A, and Lindahl B (2001) Is there an association between shift work and having a metabolic syndrome? Results from a population based study of 27,485 people. Occup Environ Med 58:747–752. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kettner NM, Voicu H, Finegold MJ, Coarfa C, Sreekumar A, Putluri N, Katchy CA, Lee C, Moore DD, and Fu L (2016) Circadian homeostasis of liver metabolism suppresses hepatocarcinogenesis. Cancer Cell 30:909–924. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kiessling S, Eichele G, and Oster H (2010) Adrenal glucocorticoids have a key role in circadian resynchronization in a mouse model of jet lag. J Clin Invest 120:2600–2609. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Knutsson A (2003) Health disorders of shift workers. Occup Med (Lond) 53:103–108. [DOI] [PubMed] [Google Scholar]
  27. Ko CH and Takahashi JS (2006) Molecular components of the mammalian circadian clock. Hum Mol Genet 15 Spec No 2:R271–277. [DOI] [PubMed] [Google Scholar]
  28. Koike N, Yoo SH, Huang HC, Kumar V, Lee C, Kim TK, and Takahashi JS (2012) Transcriptional architecture and chromatin landscape of the core circadian clock in mammals. Science 338:349–354. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Lahti T, Merikanto I, and Partonen T (2012) Circadian clock disruptions and the risk of cancer. Ann Med 44:847–853. [DOI] [PubMed] [Google Scholar]
  30. Langmead B and Salzberg SL (2012) Fast gapped-read alignment with Bowtie 2. Nat Methods 9:357–359. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Laposky A, Easton A, Dugovic C, Walisser J, Bradfield C, and Turek F (2005) Deletion of the mammalian circadian clock gene BMAL1/Mop3 alters baseline sleep architecture and the response to sleep deprivation. Sleep 28:395–409. [DOI] [PubMed] [Google Scholar]
  32. Le Martelot G, Canella D, Symul L, Migliavacca E, Gilardi F, Liechti R, Martin O, Harshman K, Delorenzi M, Desvergne B, et al. (2012) Genome-wide RNA polymerase II profiles and RNA accumulation reveal kinetics of transcription and associated epigenetic changes during diurnal cycles. PLoS Biol 10:e1001442. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. LeGates TA, Dunn D, and Weber ET (2009) Accelerated reentrainment to advanced light cycles in BALB/cJ mice. Physiol Behav 98:427–432. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Leproult R, Holmback U, and Van Cauter E (2014) Circadian misalignment augments markers of insulin resistance and inflammation, independently of sleep loss. Diabetes 63:1860–1869. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G, Durbin R and Genome Project Data Processing Subgoup (2009) The Sequence Alignment/Map format and SAMtools. Bioinformatics 25:2078–2079. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Marcheva B, Ramsey KM, Buhr ED, Kobayashi Y, Su H, Ko CH, Ivanova G, Omura C, Mo S, Vitaterna MH, et al. (2010) Disruption of the clock components CLOCK and BMAL1 leads to hypoinsulinaemia and diabetes. Nature 466:627–631. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Martino TA, Oudit GY, Herzenberg AM, Tata N, Koletar MM, Kabir GM, Belsham DD, Backx PH, Ralph MR, and Sole MJ (2008) Circadian rhythm disorganization produces profound cardiovascular and renal disease in hamsters. Am J Physiol Regul Integr Comp Physiol 294:R1675–1683. [DOI] [PubMed] [Google Scholar]
  38. Masri S and Sassone-Corsi P (2010) Plasticity and specificity of the circadian epigenome. Nat Neurosci 13:1324–1329. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Maywood ES, O’Neill JS, Reddy AB, Chesham JE, Prosser HM, Kyriacou CP, Godinho SI, Nolan PM, and Hastings MH (2007) Genetic and molecular analysis of the central and peripheral circadian clockwork of mice. Cold Spring Harb Symp Quant Biol 72:85–94. [DOI] [PubMed] [Google Scholar]
  40. Mohawk JA, Green CB, and Takahashi JS (2012) Central and peripheral circadian clocks in mammals. Annu Rev Neurosci 35:445–462. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Moller-Levet CS, Archer SN, Bucca G, Laing EE, Slak A, Kabiljo R, Lo JC, Santhi N, von Schantz M, et al. (2013) Effects of insufficient sleep on circadian rhythmicity and expression amplitude of the human blood transcriptome. Proc Natl Acad Sci U S A 110:E1132–1141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Monk TH and Buysse DJ (2013) Exposure to shift work as a risk factor for diabetes. J Biol Rhythms 28:356–359. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Mukherji A, Kobiita A, and Chambon P (2015a) Shifting the feeding of mice to the rest phase creates metabolic alterations, which, on their own, shift the peripheral circadian clocks by 12 hours. Proc Natl Acad Sci U S A 112:E6683–6690. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Mukherji A, Kobiita A, Damara M, Misra N, Meziane H, Champy MF, and Chambon P (2015b) Shifting eating to the circadian rest phase misaligns the peripheral clocks with the master SCN clock and leads to a metabolic syndrome. Proc Natl Acad Sci U S A 112:E6691–6698. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Nakayama J, Klar AJ, and Grewal SI (2000) A chromodomain protein, Swi6, performs imprinting functions in fission yeast during mitosis and meiosis. Cell 101:307–317. [DOI] [PubMed] [Google Scholar]
  46. Oh G, Ebrahimi S, Carlucci M, Zhang A, Nair A, Groot DE, Labrie V, Jia P, Oh ES, Jeremian RH, et al. (2018) Cytosine modifications exhibit circadian oscillations that are involved in epigenetic diversity and aging. Nat Commun 9:644. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Pan A, Schernhammer ES, Sun Q, and Hu FB (2011) Rotating night shift work and risk of type 2 diabetes: two prospective cohort studies in women. PLoS Med 8:e1001141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Panda S, Antoch MP, Miller BH, Su AI, Schook AB, Straume M, Schultz PG, Kay SA, Takahashi JS, and Hogenesch JB (2002) Coordinated transcription of key pathways in the mouse by the circadian clock. Cell 109:307–320. [DOI] [PubMed] [Google Scholar]
  49. Quinlan AR and Hall IM (2010) BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26:841–842. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Quinodoz M, Gobet C, Naef F, and Gustafson KB (2014) Characteristic bimodal profiles of RNA polymerase II at thousands of active mammalian promoters. Genome Biol 15:R85. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Reinke H and Asher G (2016) Circadian clock control of liver metabolic functions. Gastroenterology 150:574–580. [DOI] [PubMed] [Google Scholar]
  52. Scheer FA, Hilton MF, Mantzoros CS, and Shea SA (2009) Adverse metabolic and cardiovascular consequences of circadian misalignment. Proc Natl Acad Sci U S A 106:4453–4458. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Sobel JA, Krier I, Andersin T, Raghav S, Canella D, Gilardi F, Kalantzi AS, Rey G, Weger B, Gachon F, et al. (2017) Transcriptional regulatory logic of the diurnal cycle in the mouse liver. PLoS Biol 15:e2001069. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Song Q and Smith AD (2011) Identifying dispersed epigenomic domains from ChIP-seq data. Bioinformatics 27:870–871. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Summa KC, Vitaterna MH, and Turek FW (2012) Environmental perturbation of the circadian clock disrupts pregnancy in the mouse. PLoS One 7:e37668. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Suwazono Y, Dochi M, Sakata K, Okubo Y, Oishi M, Tanaka K, Kobayashi E, Kido T, and Nogawa K (2008) A longitudinal study on the effect of shift work on weight gain in male Japanese workers. Obesity (Silver Spring) 16:1887–1893. [DOI] [PubMed] [Google Scholar]
  57. Suwazono Y, Sakata K, Okubo Y, Harada H, Oishi M, Kobayashi E, Uetani M, Kido T, and Nogawa K (2006) Long-term longitudinal study on the relationship between alternating shift work and the onset of diabetes mellitus in male Japanese workers. J Occup Environ Med 48:455–461. [DOI] [PubMed] [Google Scholar]
  58. Tahara Y, Kuroda H, Saito K, Nakajima Y, Kubo Y, Ohnishi N, Seo Y, Otsuka M, Fuse Y, Ohura Y, et al. (2012) In vivo monitoring of peripheral circadian clocks in the mouse. Curr Biol 22:1029–1034. [DOI] [PubMed] [Google Scholar]
  59. Tahara Y, Takatsu Y, Shiraishi T, Kikuchi Y, Yamazaki M, Motohashi H, Muto A, Sasaki H, Haraguchi A, Kuriki D, et al. (2017) Age-related circadian disorganization caused by sympathetic dysfunction in peripheral clock regulation. NPJ Aging Mech Dis 3:16030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Tognini P, Murakami M, Liu Y, Eckel-Mahan KL, Newman JC, Verdin E, Baldi P, and Sassone-Corsi P (2017) Distinct circadian signatures in liver and gut clocks revealed by ketogenic diet. Cell Metab 26:523–538.e5. [DOI] [PubMed] [Google Scholar]
  61. Turek FW, Joshu C, Kohsaka A, Lin E, Ivanova G, McDearmon E, Laposky A, Losee-Olson S, Easton A, Jensen DR, et al. (2005) Obesity and metabolic syndrome in circadian Clock mutant mice. Science 308:1043–1045. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Valekunja UK, Edgar RS, Oklejewicz M, van der Horst GT, O’Neill JS, Tamanini F, Turner DJ, and Reddy AB (2013) Histone methyltransferase MLL3 contributes to genome-scale circadian transcription. Proc Natl Acad Sci U S A 110:1554–1559. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Vermeulen M and Timmers HT (2010) Grasping trimethylation of histone H3 at lysine 4. Epigenomics 2:395–406. [DOI] [PubMed] [Google Scholar]
  64. Vimalananda VG, Palmer JR, Gerlovin H, Wise LA, Rosenzweig JL, Rosenberg L, and Ruiz Narvaez EA (2015) Night-shift work and incident diabetes among African-American women. Diabetologia 58:699–706. [DOI] [PMC free article] [PubMed] [Google Scholar]
  65. Vollmers C, Schmitz RJ, Nathanson J, Yeo G, Ecker JR, and Panda S (2012) Circadian oscillations of protein-coding and regulatory RNAs in a highly dynamic mammalian liver epigenome. Cell Metab 16:833–845. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Vujovic N, Davidson AJ, and Menaker M (2008) Sympathetic input modulates, but does not determine, phase of peripheral circadian oscillators. Am J Physiol Regul Integr Comp Physiol 295:R355–360. [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. West AC and Bechtold DA (2015) The cost of circadian desynchrony: evidence, insights and open questions. Bioessays 37:777–788. [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Xia L, Ma S, Zhang Y, Wang T, Zhou M, Wang Z, and Zhang J (2015) Daily variation in global and local DNA methylation in mouse livers. PLoS One 10:e0118101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  69. Yamaguchi Y, Suzuki T, Mizoro Y, Kori H, Okada K, Chen Y, Fustin JM, Yamazaki F, Mizuguchi N, Zhang J, et al. (2013) Mice genetically deficient in vasopressin V1a and V1b receptors are resistant to jet lag. Science 342:85–90. [DOI] [PubMed] [Google Scholar]
  70. Yamazaki S, Numano R, Abe M, Hida A, Takahashi R, Ueda M, Block GD, Sakaki Y, Menaker M, and Tei H (2000) Resetting central and peripheral circadian oscillators in transgenic rats. Science 288:682–685. [DOI] [PubMed] [Google Scholar]
  71. Ye T, Krebs AR, Choukrallah MA, Keime C, Plewniak F, Davidson I, and Tora L (2011) seqMINER: an integrated ChIP-seq data interpretation platform. Nucleic Acids Res 39:e35. [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Yoo SH, Yamazaki S, Lowrey PL, Shimomura K, Ko CH, Buhr ED, Siepka SM, Hong HK, Oh WJ, Yoo OJ, et al. (2004) PERIOD2::LUCIFERASE real-time reporting of circadian dynamics reveals persistent circadian oscillations in mouse peripheral tissues. Proc Natl Acad Sci U S A 101:5339–5346. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

suppl_material1
NIHMS1816743-supplement-suppl_material1.pdf (3.9MB, pdf)
suppl_material2
NIHMS1816743-supplement-suppl_material2.xlsx (2.3MB, xlsx)

RESOURCES