Regionality of short and long period oscillators in the suprachiasmatic nucleus and their manner of synchronization.

In mammals, the center of the circadian clock is located in the suprachiasmatic nucleus (SCN) of the hypothalamus. Many studies have suggested that there are multiple regions generating different circadian periods within the SCN, but the exact localization of the regions has not been elucidated. In this study, using a transgenic rat carrying a destabilized luciferase reporter gene driven by a regulatory element of Per2 gene (Per2::dLuc), we investigated the regional variation of period lengths in horizontal slices of the SCN. We revealed a distinct caudal medial region (short period region, SPR) and a rostro-lateral region (long period region, LPR) that generate circadian rhythms with periods shorter than and longer than 24 hours, respectively. We also found that the core region of the SCN marked by dense VIP (vasoactive intestinal peptide) mRNA-expressing neurons covered a part of LPR, and that the shell region of the SCN contains both SPR and the rest of the LPR. Furthermore, we observed how synchronization is achieved between regions generating distinct circadian periods in the SCN. We found that the longer circadian rhythm of the rostral region appears to entrain the circadian rhythm in the caudal region. Our findings clarify the localization of regionality of circadian periods and the mechanism by which the integrated circadian rhythm is formed in the SCN.

Introduction

The center of the mammalian circadian clock is located in the suprachiasmatic nucleus (SCN) of the hypothalamus, which consists of a bilateral pair of SCN across the third ventricle and contains approximately 10,000 cells on each side [1]. A mutual positive/negative feedback loop is formed by the regular expression of multiple clock genes: Per1, Per2, Cry1, Cry2, Bmal1, Clock, and their protein products [1, 2]. The clock gene expression rhythm within the SCN is robust, having a circadian rhythm of approximately 24 hours [2]. To maintain the integrated circadian rhythm as a single functional unit, the circadian rhythms in the oscillating neurons in the SCN must be synchronized.

Functionally the SCN is divided into two regions [3]. The ventrolateral core region receives direct projection from the retina, while the dorsomedial shell region does not [3, 4]. The core region is composed of photo-responsive retinorecipient neurons that deliver photic information to the shell [3, 5]. After an abrupt shift of the light/dark cycle (LD cycle), the locomotor activity of rodents shows a slow shift in locomotor activity that is observed as jet lag [6]. We previously found a slow shift of the circadian rhythm in the shell after an abrupt shift of the LD cycle, and supposed that the slow shift causes jet lag [3].

The SCN is a heterogeneous structure comprising many types of neurons [1, 4, 7, 8]. Most SCN neurons are GABAergic, which is an inhibitory neurotransmitter [1, 4, 9]. Many of these GABAergic neurons co-express neuropeptides such as vasoactive intestinal peptide (VIP), gastrin releasing polypeptide (GRP), and arginine vasopressin AVP) [1, 10]. AVP is expressed primarily in the shell of the SCN, while VIP and GRP are expressed in the core. VIP has been demonstrated to be particularly important for the maintenance and entrainment of cellular clocks in individual SCN neurons [11-14]. In addition, AVP-expressing neurons are densely expressed in the shell and have been shown to extend jet lag [15, 16]. Other neurotransmitters such as GABA and GRP may play additional roles for the maintenance of the circadian rhythm in the SCN [17-20].

Previous studies have suggested that each neuron in the SCN has a different cell-autonomous circadian rhythm [21, 22], and that there are regional period differences [21, 23–26]. Noguchi et al. [25, 26] dissected the SCN into dorsal-ventral and rostral-caudal coordination, and found differences in circadian period within the SCN. Koinuma et al. [21] revealed that there is a small region in the dorsomedial part of the ex-vivo coronal slices of the SCN, showing a shorter circadian period (short period region, SPR) than the rest of the SCN (long period region, LPR) and also revealed that a phase wave propagates from SPR to LPR. However, the localizations of these regions generating the various circadian rhythms within the SCN and how they are synchronized with each other has not been fully elucidated.

In this study, we investigated differences of circadian period in the rat SCN by monitoring the bioluminescence of coronal and horizontal slices. We observed and analyzed the regional period differences in the SCN, rostral-caudal coordination, and the relationship between the direction of the phase wave propagation and the period regionality. Furthermore, by dissecting the SCN slice into fragments, we investigated which region of the SCN determines the circadian period of the whole SCN and how the circadian rhythms are integrated within the SCN.

Materials and methods

Animals

Male transgenic rats of Wistar background carrying a bioluminescence reporter of Period2 (Per2) expression were used. In these rats, the rat Per2 promoter region was fused to a destabilized luciferase (dLuc) reporter gene [27]. The rats were bred and raised in our animal facility in Kindai University Faculty of Medicine under LD cycle with lights on/off at 7:00/19:00 or 19:00/7:00. Light intensity during the light phase was approximately 400 lux. Room temperature was 22 °C. The rats were fed commercial chow and tap water ad libitum. All the rats were two to three months old at the time of the experiments. All procedure was performed under isoflurane anesthesia, and all efforts were made to minimize suffering.

The experiments were conducted in accordance with the Kindai University Animal Experiment Regulations and the NIH Guidelines for the Care and Use of Laboratory Animals. All animal experimental procedures were approved by the Institutional Animal Experimentation Committee of Kindai University School of Medicine (Permission number: KAME-30-036).

Slice preparation for ex-vivo cultures

Under deep anesthesia, an animal was decapitated between ZT0 and ZT12 (ZT, zeitgeber time), and the brain was harvested in ice-cold Hanks’ balanced salt solution (pH 7.4, Sigma, USA). Coronal and horizontal brain slices were prepared by a Microslicer (Dosaka, Japan) at thicknesses of 200 μm and 150 μm, respectively. The region containing the SCN was dissected from the slices and placed on a culture insert (ORG50; Millipore, Germany) in 35-mm culture dishes with 1.3 mL of culture medium, DMEM (12100046, Gibco, USA) containing D-luciferin K salt (0.1 mM for PMT recording, 0.2 mM for imaging; DOJINDO, Japan) and supplemented with NaHCO3 (2.7 mM; Nacalai tesque, Japan), HEPES (10 mM; DOJINDO, Japan), kanamycin (20 mg/L; Gibco, Thermo Fisher Scientific, USA), insulin (5 μg/mL; Sigma, USA), putrescine (100 nM; Sigma, USA), apo-transferrin (100 mg/mL; Sigma, USA), progesterone (20 nM; Sigma, USA), and sodium selenite (30 nM; Gibco, USA) [28, 29].

In situ hybridization

Digoxigenin-labeled rVip (nucleotides 119–808; accession number X02341) cRNA probes were synthesized according to the manufacturer’s protocol (Roche Diagnostics Japan, Japan). Horizontal brain slices (150 μm) were prepared as described above, fixed by immersing in 4% paraformaldehyde solution overnight, and processed using the free floating in situ hybridization method as described in our previous studies [6, 30]. In the present study, using this in situ hybridization technique, we detected the core region by the localization of Vip mRNA-containing neurons in horizontal SCN slices.

Bioluminescence recording of coronal slices by PMT

Bioluminescence from cultured coronal slices of 200 μm thickness was measured using a photomultiplier tube (PMT, Kronos; ATTO, Japan) for 1 in every 10 min at 37 °C. The measurements were started immediately after the slice preparation and continued for 7–14 days. The data between 12 and 132 hours in culture were used for analysis.

Bioluminescence recording of horizontal slices by EMCCD camera

Horizontal slices of 150 μm thickness were cultured under the same conditions except that the concentration of D-luciferin K salt was 0.2 mM. Per2::dLuc luminescence was recorded by one of three imaging systems: Multiversa, (ATTO, Tokyo, Japan) with an EMCCD camera (iXon 897, Andor, Belfast, UK; Exposure: 59 min., Em gain value: 500, Binning: 1×1) cooled at -90 °C; LUMINOVIEW (LV200, OLYMPUS, Japan) with an EMCCD camera (C9100-23B, Hamamatsu Photonics, Japan; Exposure: 29–59 min., Sensitivity gain: 150–200 (exposure 29 min.), 100–150 (exposure 59 min.), Gain: 1) cooled at -80 °C; or Cellgraph with an EMCCD camera (AB-3000, Atto, Japan; Exposure: 59 min., Electron Multiplier Gain: 300, Pre-Amplifier Gain: 1.0) cooled at -70 °C. Bioluminescence was recorded every 30 or 60 min. The measurement was started immediately after slice preparation and continued for 7 to 10 days. After the measurement, an integrated image of 24 to 120 hours was created using Image J, and the outline of the SCN was obtained from this image.

Analysis of bioluminescence data to reveal the period and phase of the circadian rhythm

We set ROIs (ROI, regions of interest) dividing the SCN horizontal slices into several regions depending on the experiment. The average value of bioluminescence inside each ROIs was measured. The SCN horizontal slice was further divided into small square grids (grid size; 32 μm × 32 μm). The raw data from PMT and cooled CCD camera were detrended by subtracting the 24-hour moving average [31, 32] and smoothed by taking a 5-point moving average [33]. The detrended and smoothed data from 24 hours to 120 hours after the beginning of recording were fitted to a mathematically generated damped cosine curve [21] {y = a + b•exp(—c•t) •cos 2π[(t + d) / e], t: time, a: mesor, b, c, d: constants, e: period} using Excel Solver (Microsoft, USA). We calculated the period and phase of the circadian bioluminescence rhythm from each grid via the fitted curve. Grids with correlations <0.6 between detrended bioluminescence data and fitted curve were excluded from the analysis. Origin (OriginLab, USA) was used to visualize the circadian period from each grid.

Statistical analysis

Repeated measures one-way ANOVA with post-hoc Bonferroni test was used to analyze the period length measured by PMT in coronal slices. To analyze the effect of forskolin (FK, adenylate cyclase activator), we conducted repeated measures two-way ANOVA with post-hoc Bonferroni test, and multivariate comparison with post-hoc Tukey test. To analyze the effect of separation by a scalpel, we conducted repeated measures two-way ANOVA with post-hoc Bonferroni test.

Results

Phase mapping of circadian oscillations in horizontal slices of rat SCN

As shown in Fig 1A, horizontal SCN slices of 150 μm thickness were prepared (n = 4). In a slice, to examine how a phase-advanced or phase-delayed region relates to the core and the shell regions in the SCN, we investigated the localization of Vip-mRNA expressing neurons as a marker of the SCN core by using in situ hybridization [6] (Fig 1B). Then we examined the Per2::dLuc bioluminescence rhythm focusing on the difference along the rostro-caudal axis within the rat SCN, using a different individual from the one presented above. The phase wave of bioluminescence propagated from caudal to rostral and from medial to lateral in the SCN (Fig 1C, S1 Movie). The bioluminescence rhythm in the caudal area and in the medial area were the most advanced (Fig 1D and 1E, red) and those of the rostral and lateral areas were the most delayed (Fig 1D and 1E, blue). To examine this phase distribution within the SCN in more detail, we divided the SCN into small grids and analyzed them separately (Fig 1F). The phase was advanced in the medial-caudal area relative to the lateral-rostral area of the SCN. Comparing the phase map in the SCN horizontal slice (Fig 1F) with the localization of Vip mRNA expressing neurons (Fig 1B), Vip-expressing regions appeared to correspond to the regions with delayed phase within the SCN. In contrast, the shell included regions with both advanced and delayed phase.

Fig 1

Schema of the SCN slices and Per2::dLuc bioluminescence rhythm in the SCN.

(A) A schematic diagram of the sagittal section of the SCN (modified from Nagano et al., 2019). The outline of the SCN is denoted by a dashed line (considered as the shell region of the SCN), and the outline of the VIP region (considered as the core region of the SCN) by a dashed line. The green rectangle (*) in the picture indicates the location of the horizontal slice excision for the present study. (B) Representative horizontal section of the SCN showing the rVip-expressing neurons by in situ hybridization. The outline of the core region was identified by rVip-expressing neurons. Black dashed line: outline of the SCN, White dashed line: core region of the SCN, Scale bar: 250 μm. (C) Representative bioluminescence images of a horizontal slice of the rat SCN. White line indicates the outline of the SCN. The beginning of the first light period in the former light-dark cycle before decapitation was described as (projected) ZT0. Scale bar: 500 μm. Data shown in C-F are from a single slice. (D) three ROIs were set on a unilateral SCN with the same rostro-caudal width and same medial-lateral width. (E) The average of the bioluminescence intensity of each ROI was plotted against time. The phase in the caudal ROI (red) are advanced compared to those of the middle (yellow) and rostral (blue) fragments (upper panel). The phases in the medial ROIs (red) are advanced compared to those of the middle (yellow) and lateral (blue) fragments (lower panel). (F) Phase map of 1st acrophase (after the first pZT0) of the SCN. This figure indicates that the phase wave propagates from the green region to the brown region. The phase was advanced in caudal and medial relative to the rostral and lateral, indicating that the phase wave propagates from the caudal side to the rostral side, and from the medial side to the lateral side. Grid size: 32 μm.

Circadian period analysis on consecutive coronal slices of SCN

We next investigated the difference in circadian period using consecutive coronal sections containing the SCN. Six consecutive 200 μm coronal slices were prepared and were set into Kronos for examination of the Per2::dLuc circadian bioluminescence rhythm (n = 6). Among them, clear circadian rhythms were detected from three or four SCN slices per animal. The circadian rhythm showing the largest amplitude was selected (Middle) along with the adjacent rostral (Rostral) and caudal sections (Caudal) (Fig 2A, S1 Fig). The mean values of the period length of the Rostral, Middle, and Caudal were 23.8 ± 0.2, 23.8 ± 0.1, and 22.4 ± 0.2 hours, respectively (Mean ± SE, Fig 2B). The circadian period of Caudal was significantly shorter than those of the other two sections (Fig 2B), while no significant difference was detected between Rostral and Middle (p = 1.0) (one-way ANOVA, F(2,4) = 23, p = 0.0018; post-hoc Bonferroni test, caudal vs rostral, p = 0.0055; caudal vs middle, p = 0.0039; rostral vs middle, p = 1.0). These findings suggest that the caudal region of the SCN contains oscillators with shorter circadian periods than those in the middle and rostral regions.

Fig 2

Per2::dLuc bioluminescence rhythm in coronal slices of the SCN measured by PMT.

(A) Representative Per2::dLuc bioluminescence rhythm from consecutive slices of a single individual animal (Rostral, Middle, and Caudal). Left and right panels show raw data and detrended data, respectively. (B) Circadian periods from Rostral, Middle, and Caudal (Mean ± SE). The periods of each slice are superimposed. Period data from the same individuals are connected by lines. Repeated measures one-way ANOVA with Post-hoc Bonferroni test. ***: p < 0.001 vs Rostral and Middle, N.S.: not significant.

Effect of forskolin on circadian period in horizontal slice

To confirm the regional differences of circadian period, we disrupted the intracellular synchronization using forskolin (FK). Previously we found that FK disrupts the intercellular synchronization in the SCN [21, 34]. Horizontal slices of 150 μm thickness were cultured in medium containing 10 μM FK (n = 6) or vehicle (DMSO). One to three horizontal slices were placed on a single culture insert, and bioluminescence was recorded by one of the EMCCD cameras. We set regions of interest (ROI) on the rostral half and caudal half of the unilateral SCN (Fig 3A). and designated them as Rostral area and Caudal area, respectively (Fig 3A). In FK treated horizontal slices, we found the circadian period of the Caudal area to be significantly shorter than that of the Rostral area (Fig 3B, S2 Movie). In contrast, the periods of the two areas were comparable in vehicle-treated cultures (Repeated measures two-way ANOVA, Rostral area vs Caudal area; F(1,10) = 8.2, p = 0.017, Interaction; F(1,10) = 7.5, p = 0.021. Post-hoc Bonferroni test; Rostral area vs Caudal area, Vehicle; p = 1.0, FK; p = 0.0016). Further, we divided the SCN bioluminescence images into grids (64 × 64 μm) for detailed analyses (Fig 3C). In the vehicle-treated cultures, the phase differences among the acrophases of the circadian rhythm in each grid were maintained from day 1 to day 5 (Fig 3C, upper panel). In contrast, in FK-treated cultures the phase difference among grids gradually increased (Fig 3C, lower panel). This difference is quantitatively shown in Fig 3D, which compares acrophase SD (standard deviation) within the SCN slices for each day in culture. The SDs of the 2nd–5th acrophases of the Caudal area were significantly larger than those of the Rostral area (Multivariate comparison, Caudal vs Rostral; F(1,117) = 14, p = 0.0043, Peak 1st–5th; F(1,117) = 72, p < 0.001, Interaction; F(4,117) = 8.7, p < 0.001. Post-hoc Tukey test; Caudal vs Rostral; 1st, p = 0.1823; 2nd, p < 0.001; 3rd, p < 0.001; 4th, p < 0.001; 5th, p < 0.001, Vehicle vs FK; 1st, p = 0.9160; 2nd, p = 0.0296; 3rd, p < 0.001; 4th, p < 0.001; 5th, p < 0.001). This finding suggested that FK administration caused desynchrony among circadian rhythms in the SCN. We divided the SCN into smaller grids (32×32 μm) for further detailed analyses and visualized circadian period as a map. In FK-treated slices, we found the caudal region showed periods shorter than 24 hours (Fig 3E, S2B Fig). This region showing shorter periods (designated as short period region; SPR) occupied the caudal tip of the SCN and continued to the medial narrow area. In contrast, the circadian periods of other areas in the SCN were longer than 24 hours (designated as long period region; LPR).

Fig 3

Effect of forskolin on Per2::dLuc bioluminescence rhythm.

(A) Representative bioluminescence image of the SCN (left) and bioluminescence rhythm (right) of vehicle (upper panel) or FK-treated slices (lower panel). The average bioluminescence intensities of rostral (blue) and caudal (red) areas are plotted against time. Gray bars show the periods for curve fitting. The difference in period length between rostral and caudal areas increased in FK-treated slices. (B) Period lengths of the circadian rhythm in the rostral and caudal areas in the vehicle (n = 6) and FK-treated (n = 6) groups. The mean values of the bioluminescence periods of the rostral and caudal areas of the vehicle group were 24.0 ± 0.06 hours, 23.9 ± 0.1 hours, and in the FK-treated group, 24.1 ± 0.3 hours and 22.7 ± 0.5 hours, respectively (Mean ± SE). Repeated measures two-way ANOVA with post hoc Bonferroni test showed no significant difference between rostral and caudal areas of Vehicle group, but showed a significant difference between rostral and caudal areas of FK-treated group (*: p < 0.05, N.S.: not significant). (C) Small grids were used to divide the unilateral SCN, and bioluminescence data from all grids were plotted against time. With the vehicle, the period lengths were almost identical. In FK-treated slices, the circadian rhythms showed desynchrony compared with vehicle-treated slices. Gray bar: time used for curve fitting. Grid size: 64 μm. (D) The standard deviation of acrophase in each (1st–5th) cycle within a single SCN slice (mean ± SE). Acrophase was calculated by the fitted curve obtained from the bioluminescence of each grid. According to multivariate comparison and post-hoc Tukey test, there were significant differences between the rostral area and caudal area in phase variation of the 2nd–5th cycles (**: p < 0.01, ***: p < 0.001). (E) The period lengths of the circadian bioluminescence rhythms from grids were calculated by curve fitting. The period lengths of the vehicle-treated slice were similar, whereas the period lengths of the FK-treated slices showed shorter circadian rhythms in the caudal to medial grids than those in other grids. Med: medial, Lat: lateral. Grid size: 32 μm.

Effect of rostro-caudal separation on circadian period

Knowing the localization of SPR and LPR, we investigated which region is dominant when they are synchronized. Horizontal SCN slices with a thickness of 150 was also divided into rostral and caudal fragments by scalpel (n = 7, Fig 4A). All fragments were placed on one culture insert and bioluminescence was recorded by EMCCD cameras. We set ROIs on the edges of the caudal, rostral and intact SCN, designating them Rostral, Caudal, and Intact, respectively, and the circadian periods of the bioluminescence from each ROI was measured. The circadian periods of Caudal were significantly shorter than those of Rostral (Fig 4B–4D) and Intact (Repeated measures one-way ANOVA, Rostral vs Caudal; F(2,12) = 12, p = 0.0016, Post-hoc Bonferroni test; Intact vs Rostral, p = 1.0; Intact vs Caudal, p = 0.0042; Rostral vs Caudal, p = 0.0036). Simultaneously, we prepared SCN slices without dividing by scalpel, and compared the circadian period with the intact side of the SCN dissected by scalpel (S3A Fig). We found that midline dissection of the SCN had no significant effect on the period length in horizontal SCN slices (S3B Fig).

Fig 4

Effect of rostro-caudal separation on Per2::dLuc bioluminescence rhythm.

(A) Schematic diagram of the dissected SCN. ROIs were set by outlines of Caudal, Rostral and Intact, and the circadian period of the bioluminescence from each ROI was measured. (B) One representative Per2::dLuc bioluminescence rhythm from three ROIs of a single slice. Gray bar indicates the period for curve fitting. (C) Circadian period of each ROI (n = 7). The period lengths were 23.5 ± 0.2, 24.6 ± 0.2, 24.6 ± 0.1, (Mean ± SE) in Caudal, Rostral and Intact, respectively. Repeated measures one-way ANOVA with post-hoc Bonferroni test: **p < 0.01, N.S. = not significant. (D) Representative bioluminescence images of the SCN dissected by scalpel. White lines represent the outlines of Caudal, Rostral and Intact. White arrowheads indicate the peak phases of each fragment. Scale bar: 500 μm.

Discussion

In this study, by disrupting the synchrony among oscillating neurons, we found a region located mainly in the caudal area of the SCN showing circadian periods much shorter than 24 hours (SPR) in a horizontal slice culture of the SCN (Figs 2, 3 and S2 Fig). The direction of circadian phase wave propagation in the rat SCN detected by bioluminescence was from the SPR to other areas, which suggested that the SPR initiates the phase wave at the first step (Fig 1). Further, we dissected the SCN into rostral and caudal fragments by scalpel and found that the circadian period in the caudal fragment was much shorter than those in the rostral and intact SCN slices, which suggested that the circadian period of the rat SCN is determined by LPR rather than by SPR (Fig 4).

The localization of SPR in horizontal slices observed in the present study seems to be consistent with our previous study using coronal slice cultures [21]. In the previous study, we specified a narrow medial region of the SCN with a shorter Per2 expression period (SPR) and found that the phase wave propagated from SPR to LPR [21]. In the present experiment, we also observed that the phase wave of Per2 expression in SCN horizontal slices propagated from caudal to rostral and from medial to lateral (Fig 1C–1F), that is, from SPR to LPR. The narrow SPR at the middle of the SCN in the horizontal slice shown by FK treatment (Fig 3E) was consistent with the morphological analysis of SPR in coronal sections in our previous study [21]. In addition, the direction of phase wave propagation was consistent between studies in that the wave started at the SPR and ended at the LPR. It is highly probable that the SPR observed in our previous study is identical to the SPR shown by the horizontal SCN slice analysis in the present study.

What mechanism binds the regions with distinct circadian periods? VIP and AVP are peptides that have been thoroughly investigated as substances synchronizing the oscillating neurons in the SCN [16, 35–37]. VIP is richly expressed among the neurons in the core of the SCN, and its receptor VPAC2 is mainly localized in the shell region of the SCN [15, 16]. VPAC2 gene deficient mice show desynchrony among oscillating neurons in the SCN [38]. In contrast, AVP-expressing neurons are mainly localized in the shell [25] and work to synchronize the oscillating neurons in the shell [26]. When dissected by scalpel, the caudal fragments contained the shell region and the rostral fragment contained both shell and core regions as shown in the in situ hybridization study showing the localization of VIP neurons as a core region marker (Fig 1B). AVP is abundantly expressed in the shell region of the SCN and contributes to maintaining the synchrony and phase difference among oscillating neurons in the SCN [15, 16]. Considering our present findings that there are regions with distinct intrinsic circadian periods within the SCN, synchronization of circadian oscillators by VIP and AVP contributes substantially to the circadian period of the SCN. On the other hand, Shinohara et al. [23] revealed that in rat suprachiasmatic nucleus slice cultures treated with antimitotic drugs that decreased the number of glial cells, the release of arginine vasopressin and vasoactive intestinal polypeptide showed different circadian periods. The finding suggests that the glial cells are also involved in the synchrony of oscillating neurons in the SCN.

Differences in the circadian periods between VIP- and AVP-neurons might be involved in the regional differences in circadian period. Noguchi et al. [25, 26] reported that AVP cells have intrinsically short circadian periods and are entrained by VIP cells. These studies suggest that AVP- and VIP-expressing neurons have distinct circadian periods [23, 25, 26, 36]. In the present study, we divided the horizontal SCN slice into rostral and caudal fragments. As shown by the in-situ hybridization study (Fig 1B), most of the Vip-expressing neurons were contained in the rostral fragment. Therefore, it is possible that the shortening of the circadian period of the caudal fragment is due to the removal of the VIP-expressing neurons. However, this hypothesis that VIP neurons and AVP neurons respectively generate long and short circadian rhythms seems to be inconsistent with our present and previous findings [21] that the shell region contains both SPR and LPR. AVP are rich in the shell region, so the entire shell region would generate short circadian rhythms. However, this contradiction may be explained by the uneven localization of AVP-expressing neurons in the shell. AVP-expressing neurons are dense in the medial region and sparse in the lateral region of the shell [39, 40]. If AVP-neurons generate shorter circadian periods, it is possible that partial region of the shell containing dense AVP-neurons generates short circadian periods compared with those in other shell regions.

The SPR has similar characteristics to the morning oscillator (MO) in that it is localized in the caudal region and in that the phase wave propagation starts there. Many studies have suggested the existence of a distinct morning oscillator (MO) and evening oscillator (EO) within the SCN, as the activity rhythms of rodents are separated into two components under certain LD conditions [41-43]. Jagota et al. [44] measured electrophysiological activity in horizontal SCN slices of hamsters under varying LD cycles and found two distinct peaks, possibly representing the MO and EO. In more recent studies, a bioluminescence reporter has been used to investigate the localization of EO and MO in the SCN. Inagaki et al. [28] showed two groups of oscillators coupled to the onset and end of activity (indicating EO and MO) in the mouse SCN. Another study by Yoshikawa et al. [45] mapped the localization of the two oscillators on the horizontal SCN slices in which the MO is located in the caudal tip of the SCN. In the present study, we found a region with a short circadian period in the caudal region of the SCN, and this region seemed to initiate the phase wave propagation. The finding is consistent with the properties of the MO reported previously [45]. Therefore, the present study suggested that SPR observed in the present study might be identical with the MO shown in other studies [43, 45].

In the present study, the vehicle group also showed differences in the circadian periods within the SCN (S2B Fig). The region that showed the short circadian period was mainly located in the caudal region where the SPR is, indicating that there was desynchrony even without FK. It is possible that the fragments of the SCN lost by preparing horizontal slices might be also necessary for the synchrony of oscillating neurons in the horizontal slices of the SCN. As shown in Fig 1A, the SCN slice culture did not contain the entire SCN. The lost fragment of SCN was essential for keeping the entire SCN synchronized. In addition, the retinohypothalamic tract, a neural projection from the retina to SCN, was also lost in this slice culture. It is possible that this loss also contributed to the desynchrony with vehicle treatment. Such structural disruptions and loss of components might be a limitation of slice culture experiments.

In conclusion, we analyzed the regional circadian period difference in the rat SCN and found that the phase wave propagates from the SPR to LPR. Further, we found that the circadian period of the caudal region is entrained by the rostral region of the SCN, which constitutes the overall integrated period of the whole SCN. The localization of the SPR and the direction of the phase wave propagation suggested that the SPR in the caudal region of the SCN may be identical to the MO of the two-oscillator model. The relationships between the SPR/LPR and MO/EO should be investigated further.

Per2::dLuc bioluminescence rhythm by PMT from consecutive cultured coronal SCN slices.

A representative of the six specimens is shown in Fig 2A. Data from five other SCNs are shown. The gray lines and black dotted lines indicate the detrended wave forms and fitted curves, respectively.

(TIF)

Click here for additional data file.

Effect of FK on Per2::dLuc bioluminescence rhythm.

(A) Time series of bioluminescence images from SCN horizontal slices treated with vehicle (DMSO, upper panel) and forskolin (FK, lower panel). The numbers below the pictures indicate projected ZT (ZT, zeitgeber time). Scale bar: 500 μm. (B) Grid analysis of circadian periods of Per2::dLuc bioluminescence rhythms. Grid size: 32 μm.

(TIF)

Click here for additional data file.

Effect of separating right and left SCN in a horizontal slice.

(A) ROIs were set on the SCN with and without dissection. Mean bioluminescence rhythm from each ROIs were measured. The periods of bioluminescence rhythms from the SCN without dissection (left (a) and right (b), left picture, Dissection(-), n = 5) and from unilateral SCN with dissection along the midline but without rostro-caudal dissection ((c), right picture, Dissection(+), n = 7). Dashed lines in the right picture indicate the dissection lines. (B) Statistical analysis between Dissection(-) and Dissection(+) groups. Bioluminescence period was 24.8 ± 0.14, 24.9 ± 0.16 hours for Dissection(-) group (Mean ± SE, Left and Right respectively), and 24.6 ± 0.13 hours for Dissection(+) group (Intact). No significant difference was found by repeated measures one-way ANOVA with a post hoc Bonferroni test between right and left SCN without dissection (a and b) and with dissection (c) (repeated measures one-way ANOVA; F(2,8) = 1.4, p = 0.31: post-hoc Bonferroni test; a vs. b, p = 1.0, a vs. c, p = 1.0, b vs. c, p = 0.42).

(TIF)

Click here for additional data file.

Bioluminescence imaging of a horizontal SCN slice without any treatment.

The number in the lower right corner indicates the elapsed time from the start of the measurement.

(AVI)

Click here for additional data file.

Bioluminescence imaging of a horizontal SCN slice with FK treatment.

The number in the lower right corner indicates the elapsed time from the start of the measurement.

(AVI)

Click here for additional data file.

30 Jun 2022

31 Aug 2022

Author‘s Response to Reviewers

First of all, we appreciate reviewers for their constructive and meaningful suggestions. According to reviewers’ suggestion, we revised --manuscript and Fig 1D,1E and Fig 3D. Point to point responses are shown as follows. Comments by editor and reviewers are attached with ## in front of the paragraph and responses by authors are without any signs.

Response to Editor:

## I also noticed that the sample size (the number of biological replications) of the experiments shown in Fig. 1 and 4 is not clearly stated. Please state the sample size explicitly. Also, please follow the ARRIVE 2.0 guidelines

(Experimental animals | ARRIVE Guidelines

The ARRIVE guidelines 2.0: Updated guidelines for reporting animal research | PLOS Biology)

and provide the necessary information (e.g., age and sex of rat used, type of anesthesia used, etc.) in your manuscript.

Thank you very much for pointing out. We have corrected the manuscript. (Line 89, 178, 310)

Reviewer #1

## Fig 1C The authors should quantify the propagation of the phase wave in the horizontal slice along the rostral-caudal, medial-lateral axes.

Thank you very much for pointing out. Fig 1E, which we originally presented (upper panel), was dividing the horizontal SCN into three sections in the rostral-caudal directions, and was indicating by the phase shift of each bioluminescence that the phase wave was directed from the caudal to the rostral direction. Following reviewer’s suggestion, we have added a figure (Fig 1E, lower panel) that clarifies the progression of the phase wave in the medial-lateral direction. These figures clearly show the propagation of the phase wave from the medial to the lateral direction. In addition, we have added a comment to the manuscript that Fig 1F clearly shows that the phase wave propagates from the green region to the brown region regardless of how the region is divided.

Added as follows: (Line 211-216) The phase in the caudal ROI (red) are advanced compared to those of the middle (yellow) and rostral (blue) fragments (upper panel). The phases in the medial ROIs (red) are advanced compared to those of the middle (yellow) and lateral (blue) fragments (lower panel). (F) Phase map of 1st acrophase (after the first pZT0) of the SCN. This figure indicates that the phase wave propagates from the green region to the brown region.

## Figs 3A, B, E The authors use forskolin to disrupt intracellular synchrony in a horizontal SCN slice and observe that the caudal SCN has a shorter circadian period than the rostral SCN. However, in vehicle treated horizontal slices, the period in the rostral SCN is not significantly different from the caudal SCN. How does this fit with the conclusion from Fig. 2 that in coronal slices, the caudal SCN has a shorter circadian period than the rostral SCN? Are coronal slices “more desynchronized,” similar to the effect of adding forskolin to a horizontal slice?

Thank you for the valuable comment. In the study using coronal slices, rostrocaudal connection was physically disrupted in coronal slices. In such conditions, we found that the circadian period observed in the caudal SCN slice was shorter than in others (Fig. 2B). Therefore, it is highly probable that, in the horizontal SCN slices which showed synchrony among oscillating neurons, the rostrocaudal connection was preserved. Based on these findings, we have speculated that the communication between neurons in the horizontal plane plays an important role in maintaining synchrony within the SCN. We consider that, in the coronal slices, the rostral-caudal communication has been interrupted, revealing the endogenous intrinsic circadian periods in the caudal SCN.

## The phase maps shown in Supplementary Fig 2B suggest that in some horizontal slices treated (c, d, perhaps e), the caudal SCN has a shorter period than the rostral SCN, but this is not quantified.

Thank you for pointing this out. The observation of the regions showing short periods in the vehicle-treated horizontal SCN slices suggests that some caudal small regions of the SCN slices are out of synchrony with other regions. We speculate that the reason for this is that the slices do not contain the entire SCN as shown in Figure 1A; we believe that the loss of a portion of the SCN makes it difficult to achieve synchronization of the entire SCN contained in the slices. In vehicle-treated horizontal SCN slices, the desynchronized subregions vary widely from section to section, so we have not quantified them beyond the period mapping of each section to show the period. This is one of the limitations of ex vivo studies. We have added these contents above in the Discussion section.

Added as follows: (Line 414-419) In the present study, the vehicle group also showed differences in the circadian periods within the SCN (Supplementary Figure S2B). The region that showed the short circadian period was mainly located in the caudal region where the SPR is, indicating that there was desynchrony even without FK. It is possible that the fragments of the SCN lost by preparing horizontal slices might be also necessary for the synchrony of oscillating neurons in the horizontal slices of the SCN.

## Figs 3C,D Was acrophase synchronicity different between rostral and caudal ROIs? The authors clearly show that there is more phase dispersal in forskolin treated slices, but this is perhaps unsurprising given the authors’ previously published work (Koinuma et al. EJN 2013, Sujino et al. Sci Rep 2018). It would be exciting to see if the observed rostral-caudal period difference in the horizontal slice is accompanied by a rostral-caudal difference in phase dispersal.

Thank you for the valuable comment. According to the reviewer’s comment, we re-analyze the phase synchrony of rostral and caudal ROIs separately and found significant difference. As shown in revised Fig. 3D. it clearly showed that the short period region of the caudal side in the FK-treated slices is accompanied by a large SD of the phase compared to rostral side.

Reviewer #2

Major questions:

## The point raised by the authors that the caudal SCN region with a shorter intrinsic circadian period may coordinate the timing of circadian rhythms throughout the SCN deserves additional discussion. For example, what are the implications for entrainment, and are there studies in mice or other mammals supporting this rhythm pattern?

Thank you very much for the comment. We have data in the SCN of Per2::LUC knock-in mice, which indicates regional period differences within the SCN. However, the data is preliminary so far. We consider the analysis of other species an issue for further investigation.

## An additional point needing discussion is on the relevance of these in vitro results to the SCN of the intact animal and its integration with the rest of the brain. Are the authors certain that the differences in period and phase observed between SCN regions is not an artifact of the isolation of neural tissue, separating it from other SCN regions and SCN afferents that could serve in coordinating rhythms in SCN subregions?

Thank you very much for pointing this out. In SCN slice cultures, not all of the components of the SCN are included in one section. The input and output signals to the SCN, including the retinohypothalamic tract, are cut off. Therefore, we cannot rule out the possibility that the circadian period inside the SCN is influenced by the SCN cutting. We recognize this as a limitation of the present experiment. However, in our experiment, we mainly focused on the endogenous circadian period of neurons and the synchronization among neurons within the SCN. Therefore, we consider that the blockade of the synchronization inside the SCN not only by FK administration, but also by the physical blockade of intercellular signal transduction by slicing or knife-cut, have revealed the endogenous circadian rhythm and signaling pathway within the SCN. We have added these contents above in the Discussion section.

Added as follows: (Line 414-424) In the present study, the vehicle group also showed differences in the circadian periods within the SCN (Supplementary Figure S2B). The region that showed the short circadian period was mainly located in the caudal region where the SPR is, indicating that there was desynchrony even without FK. It is possible that the fragments of the SCN lost by preparing horizontal slices might be also necessary for the synchrony of oscillating neurons in the horizontal slices of the SCN. As shown in Figure 1A, the SCN slice culture did not contain the entire SCN. The lost fragment of SCN was essential for keeping the entire SCN synchronized. In addition, the retinohypothalamic tract, a neural projection from the retina to SCN, was also lost in this slice culture. It is possible that this loss also contributed to the desynchrony with vehicle treatment. Such structural disruptions and loss of components might be a limitation of slice culture experiments.

## Is there any evidence that the SCN in the animal also shows these patterns? Do any ablation studies in rats or hamsters support the connectivity of rhythmic regions and phase wave propagation described here?

Thank you very much for pointing this out. The phase wave propagation within the SCN has been extensively studied especially in Mouse. Mathematical models have also been used to analyze the results. On the other hand, we are not familiar with the results of studies on hamsters. We consider that the ablation studies of other species are an issue for further investigation.

## Line 121, What was the dorsal-ventral position for the horizontal sections? How much of the retinohypothalamic tract remained, or was the section above this tissue?

Thank you very much for the comment. When we prepared horizontal slices, we cut from the ventral side of the SCN and always cut sections in a thickness of 150 µm from where the optic nerve and optic chiasm just barely remained. Based on this, it is probable that the horizontal SCN slices are separated from the RHT. We added these contents in Discussion session.

Added as follows: (Line 414-424) In the present study, the vehicle group also showed differences in the circadian periods within the SCN (Supplementary Figure S2B). The region that showed the short circadian period was mainly located in the caudal region where the SPR is, indicating that there was desynchrony even without FK. It is possible that the fragments of the SCN lost by preparing horizontal slices might be also necessary for the synchrony of oscillating neurons in the horizontal slices of the SCN. As shown in Figure 1A, the SCN slice culture did not contain the entire SCN. The lost fragment of SCN was essential for keeping the entire SCN synchronized. In addition, the retinohypothalamic tract, a neural projection from the retina to SCN, was also lost in this slice culture. It is possible that this loss also contributed to the desynchrony with vehicle treatment. Such structural disruptions and loss of components might be a limitation of slice culture experiments.

## Line 159, Excluding grids from the analysis where rhythm data deviated from the fitted cosine wave seems to conflict with providing an accurate characterization of the SCN’s behavior. Why was this assumption that the circadian rhythm must match a cosine wave used here? Circadian rhythms have many different waveforms. It is important to justify why this step was done and indicate how much data was excluded. Are the authors certain that they are not missing important circadian properties? Imagine a cell type or group of cells that are only active briefly during the day but provide an important role. They would not easily match the fitted function.

Thank you for the valuable comment. We agree with the point that there could be cell types with non-cosine type circadian rhythms. In this study, however, we only focused on the cells that fit the cosine curve, since the percentage of ROIs excluded from the analysis was less than 1% for both Vehicle-treated and FK-treated groups. The low percentage of exclusion may be attributed to the size of the ROIs (32 µm × 32 µm) which is much bigger than a cell size, and the relatively moderate condition (r > 0.6). In addition, because we analyzed the bioluminescence intensities with grids larger than a single cell in size, it is possible that we were not able to recognize those cells which did not fit the cosine curve. Therefore, in this study, we only focused on the cells that fit the cosine curve and used the curve fitting including cosine within the equation. In the future, we believe that further improvement in spatial and temporal resolution may allow us to identify special groups of cells that do not fit the cosine wave.

## Line 238, Why was forskolin used to disrupt circadian rhythms in this study and the previous one? Why was it preferred instead of an agent that acts more specifically like known neurotransmitters or peptides used in intercellular communication between SCN cells?

We thank you for the comments. Attempts were made to obtain desynchrony in the rat SCN using MDLs (inactivators of adenylate cyclase) and TTX in the SCN, but desynchronization could not be achieved in the rat SCN even with MDL and TTX. It is because we could not determine the appropriate concentration to detect desynchronization. MDL and TTX made the circadian oscillations of each oscillating neuron of the rat SCN damp with low concentration compared with mouse SCN. Therefore, we used FK, which has been employed in previous reports, to attain desynchronization in the SCN. In a previous paper, we showed that FK administration does not affect the endogenous circadian rhythm of the cell line (Koinuma et al. European J Neurosci 2013 ). Further, in this paper, we also observed that there is a difference in the circadian periods in rostral and caudal coronal slices as shown in Fig.2B. We speculate that FK disrupts intercellular synchrony in the SCN, possibly via a sustained increase in cAMP concentration.

Minor concerns:

## Line 138, Provide EMCCD camera gain and exposure times used.

Thank you for pointing out. The cameras used and their gains are listed below. We have also added to the manuscript.

Multiversa with an EMCCD camera (iXon 897), Exposure：59 min., Em gain value : 500, Binning : 1×1

LUMINOVIEW with an EMCCD camera (C9100-23B), Exposure : 29～59min., Sensitivity gain : 150~200 (exposure 29min.), 100~150 (exposure 59min.), Gain : 1

Cellgraph with an EMCCD camera (AB-3000), Exposure : 59min., Electron Multiplier Gain : 300, Pre-Amplifier Gain : 1.0

## Line 269, This phrase seems vague and needs more clarity and description: “…showed different circadian periods by anti-mitotic treatment…”

Thank you for pointing out. Our quotation was not accurate. We have corrected as follows. (Line: 374-378)

(Before revision)

“Shinohara et al.[23] reported that AVP and VIP release from the cultured SCN showed different circadian periods by anti-mitotic treatment, and suggested that both AVP and VIP release from the SCN are under different circadian oscillators bearing different circadian periods.”

(After revision)

“Shinohara et al.[23] revealed that the rat suprachiasmatic nucleus slice culture with a treatment of antimitotic drugs that suppress the mitosis of glial cells, circadian rhythms in the release of arginine vasopressin and vasoactive intestinal polypeptide showed different circadian periods. The finding suggests that the glial cells are also involved in the synchrony of oscillating neurons in the SCN.”

## The study appears to be focused on certain SCN neurons. What might be the role of glial cells or other neuron types in these coordinated circadian rhythms? Could they have generated some of the bioluminescence signal measured, and how could that affect the seeming interpretation that VIP and AVP cells are most responsible for the observed phenomena?

It is highly probable that glial cells in the SCN have some functions as a part of the circadian oscillator. In the revised manuscript, we cited Shinohara’s study that shows the importance of glia for the synchronization. We added the sentences below.

(Line: 374-378) “Shinohara et al.[23] revealed that the rat suprachiasmatic nucleus slice culture with a treatment of antimitotic drugs that suppress the mitosis of glial cells, circadian rhythms in the release of arginine vasopressin and vasoactive intestinal polypeptide showed different circadian periods. The finding suggests that the glial cells are also involved in the synchrony of oscillating neurons in the SCN.”

Also, it is possible many factors are thought to be involved in the synchronization of the SCN other than the AVP and VIP neurons but also the GRP neuron and neurons expressing various neurotransmitters and neuromodulators. We consider that the analysis of the contributions of glial cells and other neurons to the circadian rhythm is an issue for further investigation.

We sincerely appreciate your kind but critical readings and meaningful suggestions.

Submitted filename: Response to Reviewers_final.docx

Click here for additional data file.

19 Sep 2022

4 Oct 2022

Author's Response to Reviewers

First of all, we appreciate reviewers for their constructive and meaningful suggestions. According to reviewers’ suggestion, we revised Abstract and added Data availability statement. Point to point responses are shown as follows. Comments by editor and reviewers are attached with ## in front of the paragraph and responses by authors are without any signs.

Reviewer #2

## Relative to my response to Review Question 2: To allow for other possible interpretations of the data, I suggest that the authors qualify or moderate the statement in the Abstract "We found that the longer circadian rhythm of the rostral region entrains the circadian rhythm in the caudal region". A better statement might read "... the rostral region appears to entrain the circadian rhythm in the caudal region". I am concerned that a rigorous test of entrainment was not performed, which would require more sophisticated control of the interactions between the oscillators including a repeated control and release from entrainment and assessment of subsequent phase.

Thank you for the valuable comment. We agree with your opinion and revised Abstract as follows: (Line 35-36) We found that the longer circadian rhythm of the rostral region appears to entrain the circadian rhythm in the caudal region.

## Also, I did not see a statement saying that all the data are available or would be made available in a data repository if the paper is accepted for publication.

Thank you very much for your comment. We added Data availability statement in the manuscript.

Added as follows: (Line 438 - 442) Data used to generate the figures and supplementary figures in this manuscript are accessible on a public figshare repository (10.6084/m9.figshare.21261324). Other data and further information are also available from the corresponding author on reasonable request.

We sincerely appreciate your kind but critical readings and meaningful suggestions.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

6 Oct 2022

Regionality of short and long period oscillators in the suprachiasmatic nucleus and their manner of synchronization

PONE-D-22-16293R2

Dear Dr. Shigeyoshi,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Shin Yamazaki, Ph.D.

Section Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

10 Oct 2022

PONE-D-22-16293R2

Regionality of short and long period oscillators in the suprachiasmatic nucleus and their manner of synchronization

Dear Dr. Shigeyoshi:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Shin Yamazaki

Section Editor

PLOS ONE