A short half-life GFP mouse model for analysis of suprachiasmatic nucleus organization

2.1. Animals and housing

The mPer1::d2EGFP transgenic mouse was made with the B6C3F1 hybrid mouse as described in Kuhlman et al. [10]. The mice used in these studies were hemizygous for the d2EGFP transgene, which is an unstable form of GFP (we use hemizygotes in all our studies. In the few homozygotes we have processed, we do not detect any obvious difference in the amount of GFP expression). At about 3 months of age, animals were housed individually in translucent propylene cages (11.5×7.5×5 inches), in either a light–dark cycle (LD 12:12, n=33) or in constant darkness (DD, n=20) for 2 days (anatomical studies), or for 4–6 weeks (behavioral studies). The room was maintained at 21±1 °C. For animals in DD, a white noise generator (91 spl) masked environmental noise and a dim red light (0.5–1 lux) allowed for animal maintenance. All animals were provided with food and water ad libitum and cared for in accordance with IACUC Animal Welfare regulations at Columbia University and University of Kentucky.

2.2. Measurement of circadian locomotor activity

For behavioral studies in constant darkness (DD) cages (35×21×18 cm) were equipped with a running wheel (diameter 12.7 cm). Locomotor activity was monitored continuously using a computer-based data acquisition system (Dataquest, Data Sciences, St. Paul, MN, USA). Analysis of rhythm amplitude, period, and profile was determined by using the Clocklab program in Matlab 5.3.1 from Mathworks (Natick, MA, USA).

For the phase shift studies, animals were housed individually in constant darkness in cages equipped with running wheels. Wheel running activity was monitored using Chronobiology Kit (Stanford) software. After reaching stable free-running conditions, mice were subjected to a 30-min light pulse (about 800 lux) at a given phase. Animals’ free-running period (Tau) was calculated before and after light treatment (starting on the second cycle of activity onset following light treatment), using 6–10 cycles. The phase shift in response to light treatment was calculated as the difference in projected phase of activity onset prior to light treatment and phase of activity onset after light treatment. Phase of activity onset was determined using Chronobiology Kit by aligning a straight line to time of activity onset on a raster plot of activity (wheel revolutions per minute were binned at 2-min intervals) by eye, the slope of the line was set to Tau as calculated above. In some cases Tau varied slightly before and after light treatment (there was not a consistent direction), the two phase shift values were determined and values were averaged (for a given animal, the difference was less than 20 min). Animals were housed in a light-tight box, given animal to animal differences in period of free-running activity, animals received light treatment at different phases when they had reached a stable free-running period between treatments.

2.3. Conventional time-lapse imaging of SCN slices

Animals were housed in a 12-h light–dark cycle. Hypothalamic slices, 300 μm in thickness, containing the SCN were taken from animals 3–5 weeks old and placed in a 35 °C heated either open or closed chamber (Warner Instruments) and perfused with Earle’s balanced salts solution (Sigma) that had been supplemented with 2.2 g/l NaHCO₃, 0.8 g/l glucose, and 50 mg/l gentamicin. The solution was bubbled with 95% O₂/5% CO₂. Images of the SCN were captured every 30 min with the use of a narrow band GFP filter set (Chroma, Brattleboro, VT, USA) on either an upright microscope equipped with an Axon Imaging Workbench- (Axon, Foster City, CA, USA) controlled intensified CCD camera (Atto Instruments, Rockville, MD, USA) or an inverted microscope using IP Lab (Scanalytics, Fairfax, VA, USA) to control a cooled CCD camera (Roper Scientific, Trenton, NJ, USA). Using Axon Imaging Workbench or IP Lab, GFP fluorescence values were measured by delimiting areas inside and outside of the SCN using a region of interest tool and subtracting the non-SCN value from the SCN area value at each time point.

2.4. Perfusion and immunocytochemistry

For immunocytochemistry, hemizygous mice housed in LD were sacrificed at ZT 2, 6, 10, 14, 18 and 22, those in DD were sacrificed at the projected CT 2, 6, 8, 10, 14, 18 and 22. Animals sacrificed in the dark were deeply anesthetized (pentobarbital: 200 mg/kg) under the dim red light illumination and their heads were completely covered with a light proof hood to prevent light from reaching the eyes. Mice were perfused intracardially with 50 ml 0.9% saline followed by 100 ml of 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.3. Brains were post-fixed for 18–24 h at 4 °C, cryoprotected in 20% sucrose in 0.1 M phosphate buffer overnight. The 40-μm sections were processed free-floating. For single label immunocytochemistry, alternate or two of three sections were incubated with either rabbit polyclonal GFP (1:40 000, Molecular Probes), PER1 (1:5000, generous gift from Drs S.M. Reppert and D.R. Weaver, University of Massachusetts, MA, USA), or PER2 (1:5000, Alpha Diagnostic), or GFP, PER1 and Per1 mRNA (in situ procedure below) and processed with a modified avidin–biotin–immunoperoxidase technique [11] using DAB as the chromogen. For double label immunocytochemistry, alternate sections were incubated in PER1 antibody, used at 1:50 000, amplified with biotinylated tyramide and stained with Cy3 (Jackson Immunoresearch, West Grove, PA, USA). Sections were then incubated with the GFP antibody which was then labeled with Cy2. The sections were coverslipped with permount for DAB or Krystalon (EM Science, Gibbstown, NJ, USA) for Cy2–Cy3. In each immunocytochemistry run, sections from each time point were processed together to minimize variability attributable to handling conditions.

2.5. In situ hybridization

The mPer1 cDNA fragment-containing vectors (the gift of Dr H. Okamura, Kobe University, Japan) were linearized with restriction enzymes and then used as templates for sense or anti-sense complementary RNA (cRNA) probes. Digoxigenin (Dig)-labeled probe for mPer1 was made using Dig-UTP (Boehringer Mannheim, Mannheim, Germany) with a standard protocol for cRNA synthesis [23]. Sense mPer1 cRNA probe revealed no specific hybridization signals in brain sections.

Every third coronal section (40 μm) was processed simultaneously for in situ hybridization histochemistry. Tissue sections were processed with proteinase K (1 mg/ml, 0.1 M Tris buffer pH 8.0; 50 mM EDTA; 10 min) at 37 °C and 0.25% acetic anhydride in 0.1 M triethanolamine for 10 min. The sections were then incubated in hybridization buffer [60% formide, 10% dextran sulfate, 10 mM Tris–HCl (pH 8.0), 1 mM EDTA (pH 8.0), 0.6 M NaCl, 0.2% N-laurylsarcosine, 500 mg/ml, 200 mg/ml tRNA, 1× Denhardt’s, 0.25% SDS and 10 mM dithiothreitol (DTT)] containing the Dig-labeled mPer1 anti-sense cRNA probes for 16 h at 60 °C. After a high-stringency post-hybridization wash, sections were treated with RNase A, and then were further processed for immunodetection with a nucleic acid detection kit (Boehringer Mannheim). Sections were incubated in a solution containing nitroblue tetrazolium salt (0.34 mg/ml) and 5-bromo-4-chloro-3-indolyl phosphate toluidinium salt (0.18 mg/ml) (Boehringer Mannheim) for 16 h. The colorimetric reaction was stopped by immersing the sections in buffer 4 (10 mM Tris–HCl containing 1 mM EDTA, pH 8.0).

2.6. Immunocytochemical data analysis

Images of brain sections were captured using a CCD video camera (Sony XC77) attached to a light microscope (Olympus BH-2) using NIH Image (1.61). Three to five sections (depending whether alternate or every third sections were labeled, respectively), including rostral, mid, and caudal regions of the SCN, were analyzed per animal. Using the NIH Image program, measurements of relative optical density (ROD) were used to quantify the intensity of the Per1 mRNA, GFP, PER1 and PER2 signals in the SCN. Non-specific staining (background) outside the SCN was subtracted from staining within the SCN. The staining for each section was expressed as a percentage of the maximum staining in that ICC run, and the staining for each animal was expressed as the average of its section percentages. One-way ANOVA was used to test for time of day effects.

3.1. GFP transgenic mouse expresses precise circadian rhythms

As can be seen in Fig. 1 the circadian rhythms of the Per1::GFP mice are precise and have a high amplitude, with an average free-running period of 23.06±0.11 h (n=11). The wheel running activity profile shows most of the running occurring at the beginning of subjective night, and a smaller bout at the end of subjective night. The phase response curve resembles that of normal mice with phase advances and delays to light pulses presented in late or early subjective night, and no phase shifting during mid-subjective day.

3.2. GFP rhythms in vitro

Hypothalamic slices containing the SCN were taken between ZT0.25 and ZT1, and time-lapse imaging of the Per1-driven GFP fluorescence in the SCN was monitored for 24 h. The in vitro GFP relative fluorescence intensity in the SCN rises during the subjective day, exhibits a broad peak after mid-day (around ZT 7.0 based on the waveform average, n=4), and falls during the subjective night (Fig. 2A). Fig. 2B illustrates, for one recording, the change in GFP fluorescence in the SCN compared to background, among images captured at ZT1.5, ZT7.5, and ZT15.5. Early in the day, the SCN has detectable levels of GFP fluorescence that become more intense during the course of the day as the fluorescence peaks in the middle of the subjective day. Following the peak, GFP fluorescence in the SCN decreases during the subjective night.

3.3. GFP rhythm in vivo

In tissue taken from animals housed in LD, a circadian rhythm in intensity of GFP-ir staining was seen (ANOVA for LD, F(5)=4.8, P=0.004), with a peak around ZT 10 and a trough at ZT 22. With the 4-h sampling interval, the timing of the GFP-ir rhythm is similar to that of the fluorescence rhythm described in Fig. 2.

Fig. 3 shows rhythmic expression of Per1 mRNA, GFP, and PER1 protein from the same animals and PER2-ir in a different group, in animals held in DD. The left panel shows quantification of the data, and the right panels, photomicrographs depicting distinct high and low points of immunoreactivity.

In animals kept in DD for 2 days, circadian rhythms are seen in Per1 mRNA (F(6)=41.2, P=0.0001), GFP (F(6)=22, P=0.0003), and PER1 protein (F(6)=23.5, P=0.0003). Per1 mRNA appears to peak at around CT6–8, GFP and PER1 at about ZT10 (all quantified in serial sections from the same animals). PER2-ir measured in a different group of animals also shows a circadian rhythm (F(5)=15.3, P=0.002) peaking around CT 14 and with a trough at CT 02. There is a small population of residual Per1 mRNA cells seen at CT 18 (see Fig. 3), similar to that previously reported for PER1 protein by Field et al. [5]. In some animals depending on the precise plane of section, GFP, PER1 and PER2 proteins are also seen in this region.

3.4. Distribution and localization of GFP Per1 mRNA, PER1 and PER2

Double-label ICC for GFP and PER1 at the time of high expression reveals that both proteins have the same regional distribution within the SCN, and are co-localized in many of the same cells. The distribution of Per1 mRNA and PER2-ir at their peak expression times is also similar. Fig. 4 shows the regional distribution of GFP, PER1, Per1 mRNA and PER2-ir in five levels of the SCN, from rostral to caudal. All three proteins and Per1 mRNA expression are seen throughout the rostral SCN (levels 1 and 2). An interesting finding is that neither are expressed throughout the SCN. At the 3rd level, they are absent in a small ‘hole’ in the middle of the SCN. At the 4th level, most cells are concentrated within the medial SCN, the lateral SCN is almost devoid of Per1 mRNA GFP, PER1 or PER2-ir. We also noted that not all GFP and PER1-ir cells are double-labeled. Some cells contain GFP alone, and some cells contain PER1 alone.

Fig. 5 shows GFP immunoreactivity in the caudal SCN at the time of peak expression, CT 10. Double-labeled staining for GFP and the nuclear stain bis-benzimide shows that the region devoid of GFP is a cell-rich zone of the SCN.

4.1. GFP expression is rhythmic

The rhythm in GFP expression in the SCN, reflecting Per1 promoter activity, is similar in animals taken from LD or DD conditions and shows a peak around ZT and CT 7–10 and a trough around ZT and CT 22–24. The rhythm is similar in vitro, in cultured slices taken from animals all sacrificed at ZT0–1 (Fig. 2) and in vivo, in animals sacrificed at different time points (Fig. 3). Dynamics of Per1 gene expression tested in the same animals in DD show that Per1 mRNA expression peaks first, then, the increase in GFP expression precedes that of PER1 protein, although peak expression time is similar for both.

In summary, the data on intensity of the GFP signal, whether measured by fluorescence or by immunocytochemistry (ICC) indicates a true circadian rhythm of GFP expression. The results establish the usefulness of the Per1::GFP transgenic mouse as a tool to analyze circadian gene dynamics in the SCN. The mPer1 promoter sequences driving the short half-life GFP transgene contain three E-box elements and a CRE, and our data indicate that the dynamics of the reporter faithfully reflect the influence of transcription factors on these elements. The rhythmic patterns of Period genes expression in the transgenic Per1::GFP mouse is similar to what has been previously reported in other mice [5,18,25].

4.2. GFP distribution is topographically organized

As can be seen in Fig. 4, the overall distribution of Per1 mRNA, PER1, PER2 and GFP is similar from the rostral to the caudal aspect of the SCN: most cells colocalize PER1 and GFP, although with different intensities of staining for each protein. The occurrence of some single-labeled GFP and single-labeled PER1 cells is difficult to interpret. Possibly, these cells really don’t express both proteins. Alternatively, one of the proteins may be undetectable, as suggested by different staining intensities among cells for each protein. GFP and PER 1 proteins likely have a different time course of expression, with GFP accumulating faster than PER1. Another possible explanation is that, depending on the geometry of sectioning, single staining profiles may be observed since PER1 is a nuclear antigen, whereas GFP is in both compartments.

A surprising finding in this study was the absence of Per1-driven Per1 mRNA, GFP, PER1 and PER2 expression in some regions of the SCN. Specifically, cells of the mid-ventrolateral and caudo-lateral SCN (Fig. 4, 3rd and 4th level, respectively) do not express these mRNAs or proteins at any ZT or CT time examined. Nevertheless, this region is a neuron-rich zone of the SCN (Fig. 5). Previous work suggested that at its peak, mPer1 is expressed in all parts of the SCN, although an area devoid of Per1 expression would be difficult to detect with the autoradiographic methods used [19]. Furthermore most studies of mouse SCN, using ICC or digoxigenin labeling of PER proteins or Per mRNA, respectively, focus largely on the rostral to mid region of the nucleus (approximately levels 2–3 in Fig. 4). Importantly, at their peak Per1 mRNA, PER1 and PER2 proteins are not detected in parts of the mid to caudal SCN. At times of lowest expression, a few residual Per1 mRNA, PER1, PER2 and GFP labeled cells are seen is this region. These neurons account for a small fraction of the cells labeled with bis-benzimide (Fig. 5). Consistent with our observation, the occurrence of PER1 and PER2 protein has previously been reported in this zone [5,8]. It is not known whether these cells are rhythmic, in antiphase with the larger population, or whether they constitutively express Per genes.

It has been suggested that the organization of the mouse SCN differs from that of the rat and hamster in that light induced c-Fos is distributed extensively throughout the full extent of the SCN [2]. Other studies show that the mouse SCN can be delineated into a core and a shell according to its chemoarchitecture, and that the RHT terminates bilaterally over much of the SCN core [1]. Hastings et al. [8] showed that in the mouse SCN, rhythmic PER1 expression occurs mostly in vasopressin neurons, while the residual Per1 seen at ZT0–2 is not in vasopressin cells. The present results support the notion of a universal organization in the mammalian SCN. We have previously shown a highly topographical and compartmentalized expression of Per1, Per2, and Per3 mRNAs in the hamster SCN [7]. The rostral and dorsomedial vasopressin-rich regions are characterized by rhythmic expression of Per1 and Per2 mRNA. In contrast, in the caudal region marked by calbindin-immunoreactive cells, rhythmic expression is not detected. In the rat SCN, where clear markers for distinct SCN zones are not available, rhythms of Per1 and Per2 mRNA are strong in the dorsomedial and weak in the ventrolateral SCN [23], and the percentage of neurons showing rhythmic firing rate is much larger in the dorsomedial region than in the ventrolateral region [13].

In summary, the present results reveal that the molecular organization of clock genes in the mouse SCN is similar to that of the hamster, with a specific zone of rhythmic expression of clock genes. The mouse is becoming increasingly important for the behavioral and molecular analyses of the circadian system. The present results indicate that the degradable GFP construct in the Per1::GFP transgenic mouse is an excellent tool for this analysis.