Circadian Behavioral and Bioluminescence Rhythms in the Per2 LUC Syrian Hamster

Animals and Husbandry Conditions

Hamsters were bred and raised under a 24 h light:dark cycle with 14 h of light and 10 h of darkness (LD14:10, lights on defined as zeitgeber time 1 [ZT1]). Syrian hamsters used for generating Per2^LUC hamsters were produced from a wildtype breeding colony derived from Charles River stock (LVG Golden Syrian Hamster, Strain Code: 049). After generation, Per2^LUC hamsters were competent breeders, with estrous cycles and aggression starting in females as young as 2-3 months. For breeding, Per2^LUC female hamsters were staged to determine estrous state using vaginal smears and/or lordosis testing, then left with males for a short period of time (2-8 h) prior to female separation for gestation and lactation. The Per2^LUC mice used here were originally derived from the stock produced by Yoo et al. (2004), bred from monogamous pairs, and raised in a separate colony room programmed to LD12:12 (lights off: 1800 CST). At weaning, animals of both species were group-housed in same-sex cages without running wheels. Throughout life, ambient temperature in all animal rooms was maintained at 22 ± 2 °C, with ad libitum access to water and food. All procedures were conducted according to the NIH Guide for the Care and Use of Animals and approved by the Institutional Animal Care and Use Committee at each respective site (Utah State University IACUC protocol 2091, 13758; Marquette University IACUC protocol 3924; University Massachusetts, Amherst IACUC protocol 3537).

sgRNA Selection and Donor Plasmid Construction

Given the potential for single nucleotide polymorphisms between the published sequence of the hamster Per2 locus (Gene ID: 101834497) and the sequences from hamsters utilized here, isogenic genomic DNA approximately 2 kb upstream and 2 kb downstream of the stop codon of Per2 was polymerase chain reaction (PCR) amplified from genomic DNA isolated from hamster breeders, subcloned into TA cloning vectors and sequenced. Using Benchling software (San Francisco, CA), 3 sgRNAs were designed (Figure 1a, sgRNA1: 5’-AGATTGCTGGCTGGGGGACA-3’; sgRNA2: 5’-TAGATTGCTGGCTGGGGGAC-3’; sgRNA3: 5’-TAATGTAGATTGCTGGCTGG-3’), synthesized by Integrated DNA Technologies (IDT; Coralville, IA), and tested individually for on-target activity. Ribonucleoprotein complexes (RNPs) comprising sgRNAs and Cas9 protein (sgRNAs/Cas9) were assembled by mixing equal concentrations of sgRNAs and Cas9 protein to a final concentration of 1 μg/μL each in TE buffer (10 mM Tris, 0.1 mM ethylenediaminetetraacetic acid [EDTA] solution, pH 7.0). BHK-21 cells were nucleofected with sgRNAs/Cas9 RNP (3 μg/3 μg) using a P3 Primary Cell 4D-Nucleofector X kit (Lonza, Basel, Switzerland). Cells were harvested 72 h after nucleofection for gDNA extraction. Whether sgRNAs introduced modifications (indels) into the hamster genome was assessed by PCR restriction fragment length polymorphism (PCR-RFLP). PCR was performed with Ex Taq (Takara, San Jose, CA): initial denaturation (94 °C, 5 min), followed by 32 cycles of 30-sec denaturation (94 °C), 30-sec annealing (58 °C), and 30-sec extension (72 °C), with a final extension step for 5 min (72 °C). Primers for amplifying the region targeted by the sgRNA were Per 2F: 5’-TCCCTTGGGACTCGCTGATA-3’ and R: 5’-ACGACAAATCTCACCCCGTC-3’. The restriction enzyme BseYI (New England Biolabs, NEB, Ipswich, MA) was used for PCR-RFLP. The donor plasmid pKO2.1-Per2-Luc (Figure 1a) used for pronuclear injection was constructed by using a NEBuilder HiFi DNA Assembly Cloning Kit (NEB, Ipswich, MA). Briefly, left homologous arm (~1 kb), T2A (5’-GAGGGCAGAGGAAGTCTGCTAACATGCGGTGACGTCGAGGAGAATCCTGGCCCA-3’, as a synthetic DNA oligo), the firefly luciferase gene (from pGL4.20 [luc2], Promega #E6751), and right homologous arm (~1 kb) were PCR amplified with a 25-bp overlap on each side and assembled with plasmid pKO2.1 (Addgene plasmid #22674) linearized with MscI. The resultant plasmid was confirmed by restriction enzyme digestion mapping and Sanger sequencing (Figure S1).

OPEN IN VIEWER

Figure 1.

Generation of Per2^LUC Syrian hamsters using the CRISPR/Cas9-mediated gene knock-in strategy. (a) Schematic diagram illustrating the Syrian hamster Per2 gene (Gene ID: 101834497), the sgRNA sequences (in green, with the Protospacer Adjacent Motif, PAM, highlighted in red), and PCR primers (Per2F-R) and the restriction enzyme (BseYI, green underlined) used for detecting DNA cleavage efficiency. Also depicted are the donor plasmid carrying the T2A-Luciferase cassette, the Per2-Luc allele after the intended T2A-luciferase cassette integration immediately before the stop codon of Per2, and the junction PCR primers, F11-R12 and F13-R13, that were used to detect the genomic junctions generated by transgene cassette integration. Diagrams are not to scale. See also Figure S1. (b) Genotyping results of F0 hamsters by junction PCR to detect the Per2^LUC allele. (c) Genotyping results of F0 hamsters using Per2F-R to distinguish Per2^LUC/+ and Per2^LUC/LUC knock-in alleles.

Generation of Per2^LUC Hamsters

As in previous work (Fan et al., 2014), female Syrian hamsters were superovulated by an intraperitoneal injection of 10-25 IU of pregnant mare serum gonadotropin (BioVendor, Ashville, NC) based on body weight on the morning of Day 1 of the estrous cycle (ZT1-4). These females were mated to fertile males on Day 4 of the estrous cycle (ZT13) and were euthanized approximately 18 h after mating for zygote isolation. Pronuclear injections were performed in a dark room, and red filters were used for all microscope light sources. Zygotes were flushed from oviducts with warmed and equilibrated HECM-9 medium supplemented with 0.5 mg/ml human serum albumin, then washed twice with HECM-9, transferred into 20-µL drops of HECM-9 covered by mineral oil, and cultured at 37.5 ℃ under 10% CO₂, 5% O₂, and 85% N₂ for pronuclear injection. The sgRNA/Cas9 ribonucleoprotein and donor plasmid were diluted to 60 ng/µL:60 ng/µL using Tris/EDTA buffer (pH 7.0). For pronuclear injections, a group of 15-20 hamster zygotes were transferred to a 200-µL HECM-9 drop in the microinjection dish and covered with mineral oil. The ribonucleoprotein/donor was then injected into the male pronucleus. After injection, embryos were washed twice with equilibrated HECM-9 and cultured at 37.5 ℃ under 10% CO₂, 5% O₂, and 85% N₂. Embryos with normal morphology were transferred into the oviducts of pseudopregnant recipients that had been mated with vasectomized males on the prior day. Embryos were transferred bilaterally with 10-15 embryos per oviduct. Pups were genotyped at 2 weeks of age, with genomic DNA isolated from toe-clipping using a Qiagen Blood and Tissue Kit (Qiagen; Hilden, Germany). PCR was performed to screen T2A-luciferase cassette insertion with Ex Taq and the following parameters: initial denaturation at 94 °C for 1 min, followed by 32 cycles of 10-sec denaturation at 98 °C, 20-sec annealing at 56 °C, and 2-min extension at 68 °C, with a final extension step of 72 °C for 10 min. Hamsters were genotyped to identify animals carrying Per2^LUC by genomic PCR using primers specific for the new genomic junctions formed by the T2A-luciferase cassette insertion. Primers for amplifying the left junction are F11: 5’-CAGCGACTGTGTCCTGTTAC-3’ and R12: 5’-AAGTCCACCACCTTAGCCTC-3’. Primers for the right junction are F13: 5’-ACGTGCAAAAGAAGCTACCG-3’ and R13: 5’-GAGGTTCCCATCTCCCCAAA-3’ (Figure 1a and b). In addition, PCR with primers, Per2F and R, which were used in the DNA cleavage efficiency assays mentioned earlier, was also performed to further confirm genotyping results (Figure 1a and c). PCR was performed with Ex Taq using the following parameters: initial denaturation at 94 °C for 1 min followed by 32 cycles of 10-sec denaturation at 98 °C, 20-sec annealing at 56 °C, and 2 min extension at 68 °C, with a final extension step of 72 °C for 10 min.

Behavioral Analyses

Adult Per2^LUC and wildtype hamsters of both sexes were singly housed in cages equipped with running wheels. Locomotor rhythms were visualized and analyzed with Clocklab software (Actimetrics, Evanston, IL). To test effects of the Per2^LUC transgene on photoentrainment, hamsters were entrained to LD14:10 for at least 10 days (Per2^{LUC/ LUC} n = 11, wildtype n = 17). To evaluate effects on the free-running period, hamsters were released into DD for 9-12 days (Per2^{LUC LUC} n = 20, wildtype n = 14). Free-running period was measured with linear regression fit to activity onsets. To determine effects on photic re-entrainment, hamsters were exposed to an 8 h shift of the LD14:10 cycle in each direction (i.e., advance, then delay). Only hamsters with uninterrupted data collection were used in analyses for the advance (Per2^LUC/LUC n = 18, wildtype n = 6) and delay directions (Per2^LUC/LUC n = 15, wildtype n = 9). Rate of re-entrainment was quantified by visually determining the number of days required to realign the onset of locomotor activity to the new time of lights off in each hamster for each shift direction. Last, Per2^LUC hamsters were transferred from LD14:10 to LD6:18 for 8-12 weeks to track photoperiodic changes in behavior and physiology. Estrous cycles were monitored by daily vaginal smears taken at ZT0-2 under both LD14:10 and LD6:18, as in Kent et al. (1968). Reproductive tissue weight was collected after 9 weeks of LD6:18 for each sex (Per2^LUC/LUC female n = 4, Per2^LUC/LUC male n = 6).

Bioluminescent Tissue Collection

Per2^LUC hamsters were deeply anesthetized with pentobarbital (80 mg/kg, University of Massachusetts) or isoflurane anesthesia (Marquette), then euthanized by decapitation. Dissections were performed near dawn (ZT0-4), midday (ZT5-8), afternoon (ZT9-12), or night (ZT13-20, n = 5-16/tissue/ZT, with only 1 sample/tissue collected from each animal). Tissues collected were the SCN, pineal (PIN), pituitary (PIT), adrenal (ADR), lung, spleen (SPLN), white peri-adrenal adipose (ADI), liver, ovary (OVR), and testes. As in previous work (Evans et al., 2015), the brain and other tissues were removed and placed in chilled Hank’s Balanced Salt Solution supplemented with HEPES, NaHCO₃, and penicillin-streptomycin. SCN slices were collected in the coronal plane with a motorized vibratome (Leica Biosystems, VT1200S), and other tissue samples were collected manually with a scalpel. After collection, samples were cultured on a Millicell membrane (Millipore, Cat# PICM01250) with 1.2 ml of serum-free, air-buffered Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco 12100-046) supplemented with HEPES, NaHCO₃, penicillin-streptomycin, and D-luciferin (<0.2 mM, potassium salt, Gold Biotechnologies). For a subset of hamster tissues, dexamethasone (100 nM, Sigma, D1756) or forskolin (50 uM, Sigma, F6886) was applied for 1 h after a minimum of 7 days of recording to test if Per2-driven bioluminescence rhythms could be restored after damping, as in the works by Balsalobre et al. (2000) and Yamazaki et al. (2000). Tissues were collected at ZT6-10 from Per2^LUC mice at Marquette with isoflurane and other procedures described earlier, then compared to hamster tissues collected at the same ZT using identical procedures.

Bioluminescent Recording

Per2-driven bioluminescence rhythms were measured at the whole dish level for at least 6 days using a luminometer (Hamamatsu LM2400, Hamamatsu City, Japan; or Actimetrics) housed inside a light-tight incubator set to 36 °C. For SCN imaging, slices were imaged using a Stanford Photonics XR Mega-10Z cooled intensified CCD camera on a Zeiss AxioObserver Z1 microscope with a 10x Fluar objective lens. Images were collected with Piper Software (Stanford Photonics, Palo Alto, CA) at 15 frames/sec, filtered in real time to eliminate single-image noise events (i.e., cosmic rays), and stored as 90-sec images collected once every 10 min. A 2-h moving average was then applied, images were converted to 8-bit, pixel dimensions were reduced in half, and 3 consecutive images were summed to produce a series of 30-min images. Time in vitro was normalized across samples by expressing the start time of recording relative to ZT (e.g., if a recording started 1 h after lights off, the ZT start would be ZT13).

Luminometry Analyses

Per2-driven bioluminescence rhythms were analyzed using the Lumicycle analysis software (Actimetrics). Each time series was detrended by subtracting the 24-h running average from the raw data before fitting a damped sine wave to the first 5 cycles in vitro, starting with the time of the first trough. Period, damping rate (the number of days required for the rhythm amplitude to decrease to 1/e of the initial value, ca. 37%), and the goodness of fit of the cosine wave (% variance accounted by the fitted model) were recorded for each sample. Tissue samples were deemed rhythmic when goodness of fit exceeded 65%, which was selected to detect less-precise oscillations observed across mouse strains (Kim et al., 2021). Finally, the time and value for peak and trough bioluminescence levels were recorded for each sample over each cycle in vitro, from which amplitude (i.e., difference between peak and trough values) and peak-to-peak period length (i.e., difference between 2 consecutive peak times) were quantified for each cycle in vitro. For each sample, we calculated the average period and the precision of period (i.e., 1/standard deviation of peak-to-peak period length) over the first 5 cycles in vitro.

Imaging Analyses

Per2-driven bioluminescence rhythms were extracted and analyzed from imaging data using Matlab-based scripts, as in the study by Evans et al. (2011). For phase maps, a time series was generated for each 12-pixel-diameter region-of-interest (ROI) on a uniform grid with 2-pixel spacing. The time series for each ROI was judged circadian if the autocorrelation coefficient with a lag of 24 h was significant at α = 0.05, a local maximum occurred in the autocorrelation corresponding to a lag between 18 h and 30 h, and the signal-to-noise ratio of the time series was at least 0.75. The linear trend was eliminated, and a Butterworth filter was applied once forward and once backward to remove high- and low-frequency interference. Phase was calculated as the number of hours between the start of the recording and the time of peak bioluminescence expression on the first cycle in vitro. Because time in vitro is corrected for the ZT start time of recording, this measure is referred to as “ZT Peak Time.” To examine phase relationships across the SCN network in average phase maps, regional ZT peak time was expressed relative to the peak time of the whole anterior slice on the first cycle in vitro and referred to as “Relative Peak Time.”

Statistical Analyses

Statistical analyses were performed with JMP software (SAS Institute, Cary, NC). Data are represented in figures and tables as mean ± standard error of mean. Where appropriate, full factorial analysis of variance (FF ANOVA) was performed with the following factors: genotype, sex, tissue, and/or time of day, along with the corresponding interactions. The criterion for statistical significance was set at α = 0.05.

Generation of Per2^LUC Hamsters

The Per2^LUC construct was knocked into the endogenous Per2 locus using CRISPR/Cas9-mediated genetic engineering, as previously established for the Syrian hamster (Fan et al., 2014). The reporter gene luciferase (Luc) was inserted between the last codon and the stop codon of the Per2 gene with a viral T2A peptide linker between the coding sequences of Per2 and Luc to produce independent proteins from a single mRNA transcript (Figure 1a, Figure S1). Three sgRNAs were designed to target the stop codon region of the hamster Per2 locus (Figure 1a) and were 20 bp in length. After synthesis of sgRNA and assembly with Cas9 into the RNP complex, a PCR-RFLP assay was performed on transfected hamster BHK-21 cells to evaluate DNA cleavage (measured as indels). sgRNA1 exhibited the highest cleavage activity, resulting in 35% of indels in the transfected cells at its targeting site (data not shown). Pronuclear injection was then performed in fertilized hamster zygotes with sgRNA1/Cas9 RNP and the donor plasmid.

Pronuclear injected zygotes with normal morphology (363 total) were transferred into pseudopregnant female hamsters (14 total), resulting in 4 litters with 26 pups. Pups were genotyped by PCR using genomic DNA isolated from toe-clipping using primers to amplify the left genomic junction (primers F11-R12) and right genomic junction (primers F13-R13), created by integrating the T2A-luciferase cassette into the hamster Per2 locus (Figure 1a and b). Among the 26 pups screened, 3 pups (F0M6, F0M7, and F0F8) were positive for both left and right junctions. However, the right junctions in pups F0M6 and F0M7 were only weakly amplified (Figure 1b). Sanger sequencing of the junction PCR products indicated that the T2A-luciferase cassette was successfully integrated before the Per2 stop codon in all three pups, but indels were introduced into the right homologous arm regions in pups F0M6 and F0M7 (data not shown). These indels were most likely introduced by the non-homologous end joining (NHEJ) process, which are DNA mutations commonly observed with CRISPR-mediated gene knock-in (Merkle et al., 2015). Further genotyping demonstrated that F0F8 was homozygous for the T2A-luciferase cassette, while F0M6 and F0M7 were heterozygous (Figure 1c). Based on these results, we selected F0F8 to breed with a wildtype male, producing F1 Per2^LUC/+ hamsters used for breeding to establish the Per2^LUC/LUC colony. All the experiments described in this manuscript utilized Per2^LUC hamsters produced from this line.

To further evaluate the inserted sequence, we conducted detailed sequence analysis on the Per2^LUC genomic locus in the F0F8 progeny (n = 9). Genomic DNA was PCR amplified with primers F11 and R13 (Figure 1a) and used for Sanger sequencing (Supplemental Figure S1). Consistent with genotyping results using junction primers, DNA sequencing confirmed that the T2A-Luc cassette was integrated immediately before the Per2 stop codon as intended (Supplemental Figure S1). However, a C nucleotide insertion was identified 7 nucleotides downstream of the stop codon of the inserted Luc gene in each hamster (Supplemental Figure S1), which was likely introduced into the F0F8 founder DNA during CRISPR-mediated transgenic cassette integration.

Behavior and Physiology of Per2^LUC/LUC Hamsters

Photic entrainment was examined in wildtype and Per2^LUC/LUC Syrian hamsters of both sexes held under a standard photocycle (Figure 2a). Under LD14:10, wildtype and Per2^LUC/LUC hamsters did not differ in the phase of locomotor onset relative to lights off or the phasing of locomotor offset relative to lights on (Figure 2b). Similarly, the duration of the active phase did not differ by genotype (Figure 2c), but overall wheel revolutions were higher in Per2^LUC/LUC Syrian hamsters (data not shown, FF ANOVA omnibus: F(3,24) = 2.8, p = 0.06; GT: F(1,24) < 7.9, p < 0.01; sex: F(1,24) < 0.1, p = 0.8; GT*Sex: F(1,24) = 1.5, p = 0.2).

OPEN IN VIEWER

Figure 2.

Photoentrained and free-running locomotor activity rhythms in Per2^LUC/LUC (L/L) and Per2^+/+ hamsters (+/+). (a) Representative, doubled-plotted wheel-running actograms illustrating entrained locomotor rhythms in both genotypes. (b) Phase angle of entrainment for activity onset and offset in each genotype. PAE_on FF ANOVA omnibus: F(3,24) = 0.2, p = 0.8; GT: F(1,24) < 0.1, p = 0.8; sex: F(1,24) < 0.1, p = 0.8; GT*Sex: F(1,24) = 0.5, p = 0.5. PAE_off FF ANOVA omnibus: F(3,24) = 0.2, p = 0.8; GT: F(1,24) < 0.1, p = 0.9; sex: F(1,24) = 0.4, p = 0.5; GT*Sex: F(1,24) = 0.4, p = 0.5. (c) Activity duration did not differ by genotype. Alpha FF ANOVA omnibus: F(3,24) = 0.3, p = 0.8; GT: F(1,24) < 0.1, p = 0.8; sex: F(1,24) = 0.3, p = 0.5; GT*Sex: F(1,24) = 0.9, p = 0.3. (d) Representative, doubled-plotted wheel-running actograms illustrating free-running locomotor rhythms in both genotypes. (e and f) Period and error of period in each genotype. Note: Period was measured with a linear regression fit to activity onsets, and error of period was calculated from the deviation of daily onsets around the linear fit for each hamster. Period FF ANOVA omnibus: F(3,30) = 1.2, p = 0.3; GT: F(1,30) = 2.5, p = 0.12; sex: F(1,30) < 0.1, p = 0.9; GT*Sex: F(1,30) = 1.1, p = 0.3. Error FF ANOVA omnibus: F(3,30) = 3.8, p < 0.05; GT: F(1,30) = 0.1, p = 0.7; sex: F(1,30) = 10.2, p < 0.003; GT*Sex: F(1,30) < 0.1, p = 0.9.

Next, free-running rhythms were examined under DD (Figure 2d). Wildtype and Per2^LUC/LUC Syrian hamsters of both sexes were entrained to LD14:10, then released into DD for 9-12 days. The period of free-running rhythms did not differ between Per2^LUC/LUC and wildtype hamsters (Figure 2e). Variance in period did not differ by genotype, but this measure of circadian clock function did differ by sex (Figure 2f). Specifically, female Syrian hamsters of both genotypes displayed greater variance in free-running period than their male counterparts.

Re-entrainment to simulated jetlag was examined first by advancing the LD14:10 photocycle by 8 h, and then by delaying light offset by 8 h after a stable phase angle had been achieved (Figure 3a). Non-transgenic and Per2^LUC/LUC Syrian hamsters re-entrained within 2-3 weeks (Figure 3b and c), consistent with previous work in wildtype hamsters (Sisson et al., 2022). Following the 8-h LD advance, the number of days to re-entrainment did not differ by genotype or sex (Figure 3b). Similarly, re-entrainment to the 8-h LD delay did not differ by genotype or sex (Figure 3c).

OPEN IN VIEWER

Figure 3.

Re-entrainment to simulated jetlag in Per2^LUC/LUC (L/L) and Per2^+/+ hamsters (+/+). (a) Representative, doubled-plotted wheel-running actograms illustrating re-entrainment of locomotor rhythms during simulated jetlag in both genotypes. (b and c) Days to re-entrain in the eastward (8 h advance, FF ANOVA omnibus: F(3,20) = 0.4, p = 0.7; GT: F(1,20) = 0.3, p = 0.6; sex: F(1,20) < 0.1, p = 0.9; GT*Sex: F(1,30) = 0.9, p = 0.3) and westward directions (8 h delay, FF ANOVA omnibus: F(3,20) = 1.4, p = 0.2; GT: F(1,20) < 0.1, p = 0.8; sex: F(1,20) = 2.22, p = 0.15; GT*Sex: F(1,30) < 0.1, p = 0.7).

Last, Per2^LUC/LUC hamsters were transferred to short-day photocycles to monitor photoperiodic changes in daily rhythms and reproduction (Figure 4). Female Per2^LUC/LUC hamsters re-entrained to LD6:18 by delaying the phasing of activity onset (Figure 4a), consistent with a DD period >24 h (Figure 2e). Regular 4-day estrous cycles in behavioral and physiological indices were observed in females entrained to LD14:10 (Figure 4a, n = 7). After transfer to LD6:18, Per2^LUC/LUC female hamsters displayed a positive smear every 4 days for 47.9 ± 1.6 days (Figure 4a, n = 8), after which estrous discharges ceased and the corresponding fluctuations in locomotor activity diminished (Figure 4a). Postmortem organ weights from a separate group of LD6:18 hamsters also indicated that Per2^LUC/LUC hamsters were capable of regressing gonadal function after 9 weeks of short photoperiod entrainment (Figure 4b).

OPEN IN VIEWER

Figure 4.

Photoperiodic modulation of reproduction in Per2^LUC/LUC (L/L) hamsters. (a) Representative, doubled-plotted wheel-running actograms illustrating locomotor rhythms in female Syrian hamsters housed under LD14:10 (top) and LD6:18 (bottom). Plotted to the right of each actogram is daily total wheel revolutions, and asterisks indicate the day of a positive smear. (b and c) Paired testis and uterine weights after 9 weeks of LD6:18 entrainment in males and females, respectively. The dashed line on each panel indicates previously reported tissue weights in non-regressed hamsters of each sex maintained in long photoperiods.

Bioluminescence Luminometry and Imaging

Per2^LUC/LUC and Per2^LUC/+ hamsters were entrained to LD14:10 prior to PIT collection to test the degree to which Per2^LUC copy number influences the properties of bioluminescence rhythms measured ex vivo (Figure 5a). The PIT was selected for initial examination due to the ease of collecting this tissue, with previous work indicating that would oscillate ex vivo, and the physiological significance of this organ for hamster physiology controlled by the circadian system (e.g., photoperiodism). Tissue samples were extracted during the middle of the photophase based on previous work, suggesting this time of tissue collection minimally resets PER2::LUC rhythms in the Per2^LUC/LUC mouse (Davidson et al., 2009). Bioluminescence rhythms were monitored for at least 5 cycles, and measures of phase, period, amplitude, variability, and damping were extracted. As expected, Per2^LUC copy number influenced the amplitude of bioluminescence rhythms (Figure 5b), with lower signal detected in Per2^LUC/+ samples. Damping of bioluminescence rhythms also differed by Per2^LUC copy number (Figure 5c), but the percentage of variance accounted for by a fitted cosine curve did not differ by genotype (data not shown, t(23) = 0.7, p = 0.48). The period length of bioluminescence rhythms did not differ by Per2^LUC copy number, but cycle-to-cycle period length varied more in Per2^LUC/+ samples than in Per2^LUC/LUC samples (Figure 5d). Although bioluminescence rhythms damped in Per2^LUC/LUC PIT samples over time in culture, higher-amplitude expression could be restored by application of dexamethasone or forskolin (Figure 5e), consistent with previous work (Balsalobre et al., 2000; Yamazaki et al., 2000). Given the impact of Per2^LUC copy number on signal strength and precision without effects on period length, we used Per2^LUC/LUC hamsters for further study.

OPEN IN VIEWER

Figure 5.

Per2-driven bioluminescence rhythms of pituitary samples collected from Per2^LUC/LUC (L/L) and Per2^LUC/+ hamsters (L/+). (a) Representative times series illustrating bioluminescence levels varied by transgene copy number. Left: Baseline-subtracted time series; inset illustrates raw bioluminescence with the upper value of the y-axis indicated (i.e., 2400). Two samples are illustrated for each genotype with different line colors, as represented in the symbol legend. Right: Baseline-subtracted times series for Per2^LUC/+ pituitary samples, with y-axis adjusted to account for lower level of expression. Along the x-axis of each time series, white and gray boxes represent ex vivo projected times of light and darkness, respectively. (b and c) Amplitude and damping rate of bioluminescence rhythms varied by transgene copy number. Amplitude: Mixed Linear RM Model Copy#: F(1,86) = 5.4, p < 0.05; Cycle#: F(4,86) = 3.8, p < 0.01; Copy#*Cycle#: F(4,86) = 3.0, p < 0.05. Damping: t(23) = 2.9, p < 0.01; *Least Squared Means Contrasts post hoc comparisons, p < 0.05. (d) Period of bioluminescence rhythms did not differ by genotype (t(23) = 1.6, p = 0.12), but there was larger cycle-to-cycle deviation in period in Per2^LUC/+ pituitary samples (t(23) = −2.3, p < 0.05). (e) Damped bioluminescence rhythms could be reinitiated with 1-h administration of 100-nM dexamethasone or 50-µM forskolin.

A variety of tissues were collected from Per2^LUC/LUC hamsters of both sexes at different times of day to test whether bioluminescence rhythms monitored ex vivo are altered by dissection (Figures 6-8). Most hamster tissues examined here displayed at least 3 cycles of fluctuating bioluminescence levels (Figure 6). The exception was the testes, which failed to consistently oscillate (data not shown) despite favorable results from ovarian samples (Figure 7a). Overall, at least 50% of samples of SCN, PIN, PIT, ADR, lung, SPLN, white adipose (ADI), OVR, and liver were classified as rhythmic (Figure 7a). The percentage of tissues deemed rhythmic differed by tissue type and by time of dissection (Figure 7a). Overall, the percentage of rhythmic SCN and PIN was lower when collected at night. In addition, the goodness of fit, damping rate, and amplitude on the first cycle in vitro were affected by tissue type and/or time of dissection (goodness [tissue: F(9,364) = 6.53, p < 0.0001; time: F(3,364) = 4.75, p < 0.005]; damping [tissue: F(9,333) = 9.12, p < 0.0001; time: F(3,333) = 1.01, p > 0.3]; amplitude [tissue: (9,368) = 8.43, p < 0.0001; time: (3,368) = 0.53, p > 0.6]). Last, the period length of Per2-driven bioluminescence rhythms differed across organ types, but not by time of dissection (Figure 7c).

OPEN IN VIEWER

Figure 6.

Per2-driven bioluminescence rhythms across tissues collected from Per2^LUC/LUC (L/L) hamsters. Representative times series illustrating bioluminescence rhythms from each tissue when collected at different times of day. Conventions as in Figure 5.

OPEN IN VIEWER

Figure 7.

Percentage of rhythmic samples, damping, and period of Per2-driven bioluminescence rhythms over time in vitro. (a) Percentage of samples deemed to be rhythmic based on goodness of fit > 60% for each tissue type and time of dissection assessed. (a) “X” along the x-axis indicates no samples were collected for that tissue type at that time of day. FF ANOVA tissue: F(9,370) = 27.3, p < 0.005; time: F(3,370) = 16.9, p < 0.001. (b) Damping rate varied across tissue type, but not the time of dissection. FF ANOVA tissue: F(9,333) = 9.12, p < 0.0001; time: F(3,333) = 1.01, p > 0.3). (c) Period length differed by tissue type, but not time of dissection (FF ANOVA tissue: F(9,357) = 9.73, p < 0.0001; time: F(3,357) = 1.74, p > 0.15. For group sample sizes, c.f., Supplemental Table S1).

OPEN IN VIEWER

Figure 8.

Phasing of Per2-driven bioluminescence rhythms ex vivo is affected by time of dissection in specific tissues. (a) Zeitgeber time of peak bioluminescence expression divided by tissue type and time of dissection. FF ANOVA tissue: F(9,354) = 60.4, p < 0.0001; time: F(3,354) = 8.4, p < 0.0001. (b) Time of peak bioluminescence expression ex vivo relative to time of dissection (latency to peak). FF ANOVA tissue: F(9,354) = 38.7, p < 0.0001; time: F(3,354) = 35.1, p < 0.0001. *Tukey’s HSD post hoc, p < 0.05, with subscript indicating group differences (i.e., *D = Dawn group differs from others; *DvAN = Dawn group differs from afternoon and night groups; *ABC/ABCD = all groups differ; for group sample sizes, c.f., Supplemental Table S1).

Potential resetting effects of dissection on the phase of ex vivo bioluminescence rhythms were examined using a few approaches. First, we calculated the ZT of peak bioluminescence expression on the first cycle measured ex vivo (Figure 8). Second, we examined the latency between time of dissection and peak bioluminescence expression (Figure 8). Third, we evaluated the correlation between ZT peak time and time of dissection for each tissue (Figure 9). Tissues reset by dissection are expected to differ in the ZT peak time, but not latency to peak, which would be associated with a weak correlation between ZT peak time and the time of dissection. In contrast, tissues not reset by dissection would be expected to differ in latency but not ZT peak, with a strong correlation between ZT peak time and time of dissection. Time of dissection influenced the ZT of peak bioluminescence expression in a tissue-specific manner (Figure 8a). Specifically, morning dissections appeared to reset the ADR and PIT, afternoon dissections caused the liver to peak at a ZT different from other collection times, and all dissection times differed in the ZT of peak bioluminescence expression for the SPLN (Figure 8a). Complementing these results, latency to peak bioluminescence expression was affected by tissue type and time of dissection (Figure 8b). Most tissues displayed systematic differences in the latency to the time of peak bioluminescence expression consistent with the time interval between collection time points, which suggests that these tissues are resistant to dissection-induced resetting. Latency to peak bioluminescence expression varied for each tissue, except the liver and ATR (Figure 8b), indicating that these latter two tissues were each reset by dissection. Last, we assessed the relationship between ZT peak time and the ZT of dissection through correlational analyses (Figure 9, Supplemental Table S1). Given the time-specific nature of resetting in some tissues described earlier, we fit both a linear and polynomial model to the dataset for each tissue (Supplemental Table S1). The ADR, PIT, SPLN, and liver displayed a significant linear correlation between ZT peak time and the time of dissection, with the ADR and PIT dawn samples driving this trend. Non-significant trends were also detected for the lung and SCN (Supplemental Table S1), driven by samples collected during night and dawn. Overall, these results suggest that most tissues in the Per2^LUC/LUC hamster are robust to the resetting effects of dissection when collected during the afternoon (except the liver).

OPEN IN VIEWER

Figure 9.

Correlational analyses of the relationship between the dissection time and the ZT peak time of Per2-driven bioluminescence in each tissue type. Each individual datapoint represents a tissue sample, with the linear regression illustrated by the dashed line fit to datapoints of each tissue type. The linear fit is represented below the x-axis for each tissue. See also Supplemental Table S1.

Next, we compared bioluminescence rhythms from the SCN, ADR, PIN, PIT, SPLN, and lung of Per2^LUC/LUC hamsters to those of Per2^LUC/LUC mice collected during the afternoon using identical procedures. Overall, the period length of bioluminescence rhythms was similar across species, with the exception that the Per2^LUC/LUC mouse SCN displayed longer period (Figure 10a) and the Per2^LUC/LUC mouse PIN displayed shorter period (Figure 10A) than observed in Per2^LUC/LUC hamster samples. Phasing of peak bioluminescence expression across tissues was also comparable across species (Figure 10b and c).

OPEN IN VIEWER

Figure 10.

Comparison of Per2-driven bioluminescence rhythms in Per2^LUC/LUC hamster and mouse tissues. (a and b) Period and ZT peak of bioluminescence rhythms in each species. (SCN period: t(19) = 2.8, p < 0.05, PIN period: t(24) = -5.9, p < 0.0001). (c) Radial plots representing phasing of bioluminescence rhythms at the systems level in each species. Note: Hamster bioluminescence data from tissues collected at midday are replotted from Figure 8.

Last, we examined the spatiotemporal organization of the hamster SCN using bioluminescence imaging (Figure 11). Four consecutive slices were taken in the coronal plane across the anteroposterior SCN of 6 hamsters (4 females, 2 males) and imaged for at least 5 cycles ex vivo. When the cumulative bioluminescence signal was evaluated relative to brightfield images taken of each slice, the dominant signal observed was derived from the dorsal SCN (Figure 11a). Lower signal could be detected in ventral SCN near the optic chiasm (e.g., Slice3), with fewer bioluminescent cell-like objects in this region. More signal was detected from middle SCN slices than in other planes (Figure 11b, left). As observed in the mouse Per2^LUC/LUC SCN (Evans et al., 2011), spatiotemporal gradients in bioluminescence were observed in the hamster Per2^LUC/LUC SCN across both the anteroposterior and dorsoventral axes (Figure 11b, right and c). Specifically, bioluminescence rhythms peaked earlier in the posterior SCN and dorsal SCN relative to anterior and ventral SCN (Figure 11b and c). These results suggest that this pattern of spatiotemporal organization is a conserved feature of SCN timekeeping in both hamsters and mice.

OPEN IN VIEWER

Figure 11.

Spatiotemporal organization of Per2-driven bioluminescence rhythms in the hamster SCN. (a) Representative images of coronal SCN slices (150 μm) collected across the anteroposterior axis, with total bioluminescence over the first 24 h pseudocolored and superimposed onto bright field images. 3V: third ventricle, OC: optic chiasm, Scale bar: 100 μm. (b) Representative time series of bioluminescence rhythms measured from the SCN at each anteroposterior position. Left: raw time series. Right: baseline-subtracted time series. Color-coded arrowheads on the right plot indicate peak time for each slice. (c) Color-coded phase maps represent regional peak Per2-driven bioluminescence on the first cycle in vitro. Warm and cool colors represent late and early peaking regions, respectively. Individual phase maps are representative of n = 6 (4 females and 2 males).