INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

FUNCTIONAL PROPERTIES of the mammalian circadian system have been investigated extensively in nocturnal rodents (7, 14, 25, 26). However, it is not known whether they are a proper model for the mammalian circadian system in general, especially for humans. There seems to be a systematic difference in the properties of the circadian system between nocturnal and diurnal animals, as suggested by Aschoff (3). For example, responsiveness to light is considerably different. When the circadian rhythm is free running under constant illumination, the period becomes longer in nocturnal animals as the background light intensity is increased, whereas it becomes shorter in diurnal animals (3, 8, 9). On the other hand, previous observation showed that the circadian rhythms of a diurnal chipmunk have several characteristics common to those of humans. Asian chipmunks (also called Siberian chipmunks) are hibernators and live mainly in forests in a wide area of northern Asia. They spend most of the day time above ground without returning to their burrows in trees or underground. In both chipmunks and humans, the rest and activity times are consolidated under the light-dark cycle (2, 6, 13). A circabidian rhythm has been observed under constant conditions (2, 12). The thresholds of light intensity for suppressing the nocturnal melatonin secretion and for phase shifting the circadian rhythms are relatively high in chipmunks (29) as in humans (22). Diurnal chipmunks seem to be a better animal model for the human circadian system than nocturnal rodents.

However, the responsiveness of the circadian pacemaker to light has not been systematically examined in chipmunks. Previously, we examined the light responsiveness of the circadian locomotor rhythm in chipmunks by giving a single light pulse at four different circadian times (CT) on the first day of free running in constant darkness (DD) (2). The light pulse induced phase-dependent expression of Fos-like immunoreactivity in the suprachiasmatic nucleus (SCN), but did not produce significant phase shifts in the locomotor rhythm. A lack of light-induced phase shift on the first day of free running was ascribed to the after-effect of light entrainment on the circadian pacemaker. In the present study, a phase response curve (PRC) for a single light pulse was constructed in the chipmunk for a better understanding of the circadian system in diurnal mammals.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Animals and housing. Wild-caught chipmunks (Eutamias asiaticus; also called Eutamias sibiricus) of both sexes were used. Before the experiment, they were housed individually in a transparent polycarbonate cage (31 × 36 × 17.5 cm) in a room where environmental conditions were controlled [12:12-h light-dark cycle (LD), white fluorescent lights on 0600-1800, light intensity in the cage was ~100 lx, room temperature was 23 ± 2°C, humidity was 55 ± 5%]. They were released to constant dim light (dim LL) to allow the circadian system to free run. The light intensity in dim LL was ~3 lx at the level of the cage floor.

Locomotor activity measurement. Spontaneous locomotor activity was measured by an Animex (Animex Type III, Shimadzu), a condenser type actograph. The method for activity measurement is described in detail elsewhere (14, 15). Briefly, voluntary movements of an animal change the capacitance of condensers located below the plate on which the animal cage is placed. The change in capacitance was transformed into a current change, and the number of current changes was fed into computer every 15 min. Exchange of a cage for cleaning was done every 3-4 wk.

Phase responses to a 1-h light pulse. A single light pulse (white fluorescent light) of ~2,000 lx at the floor of the cage was given for 1 h to the chipmunks that showed a stable free- running locomotor rhythm for >2 wk under dim LL. The light pulses were given to a single chipmunk at least 16 days apart. In total, 74 pulses were given to 31 chipmunks. In three chipmunks, locomotor activity was decreased under dim LL and the rhythm was disrupted after the first light pulse. These results were omitted from analyses, and the animals were not used for further experiments. Five chipmunks did not respond to light pulses at any tested phase (12 pulses), and their results were excluded from the construction of a PRC. These five chipmunks were subjected to a 3-h pulse experiment.

Phase responses to 3-h light pulse. A single 3-h light pulse of 2,000 lx was given to 11 chipmunks whose circadian rhythm did not respond at all (5 animals) or responded markedly during the subjective night (6 animals) to the 1-h light pulse. A 3-h pulse was given at two different CT, 16.5 and 19.5, where CT0 was defined as the time of activity onset.

Data analyses. To construct a steady-state PRC, free-running period was calculated from a regression line fitted to the activity onsets of at least 10 consecutive cycles under steady-state free running. The transient cycles after the light pulse (4-8 days) were omitted from the calculation of a regression line. Phase-shifts were calculated on the next day of the light pulse by forward and backward extrapolations of two regression lines: one immediately before a light pulse and the other after the pulse. The midpoint of a light pulse was selected as the reference phase in both 1- and 3-h light pulse experiments. When a large phase shift of ~12 h was produced by a light pulse, the direction of the phase shift was decided from the direction of the phase shift during transients. In the case of phase shift accompanied by splitting of the activity band, the direction of the phase shift was decided to be the shift with a smaller difference when calculated as advance or delay. A PRC for a 1-h light pulse was constructed with the mean phase responses calculated every two CT bins (1 CT = free-running period/24 h). Changes in the free-running period () were calculated by subtracting the period before the light pulse from that after the pulse. Mean  responses for 1- and 3-h light pulses were also calculated every two CT bins.

Statistical analyses. Differences in the amount of phase shifts between 1- and 3-h light pulses were analyzed by a paired t-test. Linear regression was obtained between the amount of phase shift and change in the circadian period, and the correlation coefficient was statistically evaluated by t-test.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Phase-dependent phase shift by a single 1-h light pulse. Chipmunks showed robust free-running rhythms in locomotor activity with a consolidated activity band and a continuous rest period. As shown in Fig. 1, the light pulses given at the subjective evening phase delayed and at the subjective morning phase advanced the locomotor rhythm with transients of several days. Almost no phase shift was observed after the light pulses during the subjective day. When a light pulse was given around the midpoint of the subjective night (between CT16 and CT20), the activity rhythm showed a large phase shift up to 12 h (Fig. 2), which was often accompanied by splitting of the activity band (Fig. 2A). In five animals, the circadian rhythm did not significantly respond to a light pulse given at any tested phase. These results were not included in the PRC. Figure 3A summarizes the 47 phase responses in 23 chipmunks. The PRC (Fig. 3B) had a phase-delay portion between CT12 and CT18, an advance portion between CT20 and CT24, and a dead zone between CT4 and CT10. The phase-transition curve (Fig. 3C) had an average slope of zero. The PRC is categorized as a type 0 PRC.

View larger version (57K): [in this window] [in a new window]   Fig. 1.   Double-plotted spontaneous locomotor activity rhythms of chipmunks before and after a single light pulse (2,000 lx, 1 h) showing no response by pulse at subjective day (R9-1229), phase delay at subjective evening (V7-102), and phase advance at subjective morning (V8-907).  and arrows, onset phase and date of light pulse, respectively.

View larger version (58K): [in this window] [in a new window]   Fig. 2.   Double-plotted spontaneous locomotor activity rhythms of chipmunks showing a large phase shift by a light pulse at middle of subjective night. Four representative phase shifts accompanied by splitting of activity band (A) and by a marked change in free-running period (B).  and arrow, onset phase and date of light pulse, respectively.

View larger version (13K): [in this window] [in a new window]   Fig. 3.   Phase responses of chipmunks for 1 h-light pulses (n = 47) plotted against circadian phase of light pulse (A) and a phase response curve (PRC) constructed by mean phase responses in every 2 circadian time (CT) bin with SE (B). A phase transition curve (C) shows average slope of 0. Abscissa of phase-transition curve indicates phase of light pulse in prepulse free-running rhythm (initial phase), and ordinate indicates phase at which light pulse hit with respect to postpulse free-running rhythm (final phase).

Eradication of circadian rhythm by the light pulse. After a single light pulse given at the midpoint of the rest period, the circadian rhythm was abolished in some trials (Fig. 4). Rhythm disappearance was detected in 5 of 19 trials with the 1-h light pulse at CT16-CT21 and 3 of 18 trials with the 3-h pulse given either at CT16.5 or CT19.5. The frequency of rhythm disappearance was not significantly different between the two types of light pulse. In a chipmunk that became arrhythmic, effects of additional pulses were examined. A 1-h light pulse given 21 days after the first pulse did not produce any significant change in the locomotor activity pattern (Fig. 4). On the other hand, when a 6-h light pulse was given 9 days after the second pulse, the activity band became consolidated and the circadian rhythm was restored with a significant period (P < 0.01) of 23.3 h (Fig. 5). The effect, however, persisted for only 8 days. A 12-h light pulse given 13 days later restored again the locomotor rhythm with a period (P < 0.01) of 24.1 h. When the circadian rhythm was restored by an additional light pulse, the end of the activity band of the restored rhythm coincided approximately with the end of light pulse.

View larger version (64K): [in this window] [in a new window]   Fig. 4.   Double-plotted locomotor activity of representative rhythm eradication produced by 1-h light pulse at middle of subjective night.  and arrows, onset phase and date of light pulse, respectively.

View larger version (48K): [in this window] [in a new window]   Fig. 5.   Double plotted locomotor activity showing restoration of circadian rhythm in light-induced arrhythmic chipmunks after 6- and 12-h light pulse. , time of light exposure. Arrows at left, day of light exposure.

Light-induced splitting of activity band. A phase shift of ~12 h was often accompanied by a splitting of an activity band. In such cases, one activity component phase advanced and the other phase delayed. Both components were fused again after transients of several days (Fig. 2A). In some cases, split components did not fuse but kept free running with different periods (Fig. 6).

View larger version (77K): [in this window] [in a new window]   Fig. 6.   Double-plotted spontaneous locomotor rhythm showing continuation of splitting after a light pulse. A 2 periodogram analysis revealed significant circadian period (P < 0.01) of 23.3 and 23.9 h in chipmunk R5-1-11, and 23.8 and 24.8 h in R9-428.  and arrows, onset phase and date of light pulse, respectively.

Phase shifts by a 3-h light pulse. The phase responses to the 3-h light pulses were compared with those to the 1-h pulses in 11 chipmunks, in which five did not respond to the 1-h light pulse (nonresponders) and the remaining six responded markedly (high responders). The 3-h light pulses were given at CT16.5 and CT19.5. Figure 7 shows locomotor rhythm of one representative high responder and one nonresponder. The 3-h light pulse failed to yield a significant phase shift in the five nonresponders at both CT16.5 and CT19.5 (Figs. 8B). In all high responders, large phase-delay shifts were observed after the 3-h light pulse at CT16.5 (Figs. 8A). After the 3-h pulse at CT19.5, the locomotor rhythm became aperiodic in three and produced a large phase-advance shift in the remaining three chipmunks. Statistical significance was not detected in the amount of phase shift between the 1- and 3-h light pulses at either CT16.5 or CT19.5.

View larger version (73K): [in this window] [in a new window]   Fig. 7.   Double-plotted actograms of 2 chipmunks exposed to 3-h light pulses at CT16.5 and CT19.5, which demonstrated large phase shifts (R11) and essentially no phase shifts (R91) by light pulse. Note an immediate phase delay after pulse at CT16.5 (first pulse) and a phase advance with transients of 7 days after pulse at CT19.5 (second pulse) in R11. First pulse was given at CT19.5 and second at CT16.5 in R91.  and arrows, onset phase and date of light pulse, respectively.

View larger version (18K): [in this window] [in a new window]   Fig. 8.   Comparison of phase shifts by 1- and 3-h light pulse at CT16.5 and CT19.5 in high responders (A; n = 6 for CT16.5, n = 3 for CT19.5, remaining 3 became aperiodic) and nonresponders (B; n = 5). No significant difference was detected between 1- and 3-h pulses at either CT. See RESULTS for details.

Free-running rhythms of chipmunks. Free-running periods were calculated before and after the light pulses. The mean and standard deviation of the free- running period was 23.95 and 0.48 h for the responders (104 analyses in 23 chipmunks) and 23.94 and 0.21 h for the nonresponders (44 analyses in 5 chipmunks). A large phase shift induced by the light pulse was often accompanied by a change in free-running period (Fig. 2, A and B). The change in the period varied from 1.0 to +2.5 h. As shown in Fig. 9A, significant negative correlation (P < 0.01) was detected between the phase shifts () and change in the period (), indicating that phase-delay shifts were accompanied by lengthening of the free-running period and phase-advance shifts were accompanied by the shortening of the period. When was plotted against the phase of light pulse, the  response curve had a dead zone between CT22 and CT8 and significant period lengthening was detected between CT16 and CT18 (mean ± SE = 0.54 ± 0.16 h).

View larger version (15K): [in this window] [in a new window]   Fig. 9.   A: relation between amount of phase shifts () and change in circadian period (tau) after a light pulse in light responders. Positive and negative numbers of mean phase advance and delay shifts of circadian rhythms, respectively. Tau indicates difference between steady-state free-running periods before and after a light pulse. Positive and negative numbers of tau mean increase and decrease of free-running period, respectively. Oblique line in graph is a linear regression. Significant negative correlation (P < 0.01) was detected between and tau. B: tau plotted against circadian phase of light pulse. Mean tau responses in 2 CT bins revealed significant lengthening of tau (P < 0.05) at CT16-18.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The present study clearly showed a type 0 PRC for a brief light pulse in the circadian locomotor rhythm of a diurnal rodent, Asian chipmunk. The PRC has a large phase-delay portion in the early subjective night and an advance portion in the late subjective night (Fig. 3). The break point of the PRC was located at CT 18.5. A few PRCs were previously reported in diurnal rodents, such as antelope ground squirrels (Ammospermophilus leucurus) (27), eastern chipmunks (Tamias striatus) (20), and palm squirrels (Funambulus palmarum) (23). These are all type 1 PRCs, with a relatively large phase-delay portion and a small phase-advance portion (20, 23) or with similar amounts of phase shifts in both directions (27). Most of them have no dead zone. The PRC shape is known to be influenced by the strength of resetting stimulus (18, 21). Increasing the intensity or the duration of the light pulse changed the PRC shape from type 1 to type 0 (10, 18). Significant lengthening of pulse duration could cover either delay or advance phases of PRC, resulting in a lack of dead zone. The difference between the previous results and ours may be due to differences in the intensity and duration of the light pulse (duration varied from 15 min to 6 h and intensity from 100 to 1,000 lx) and background illumination (DD or dim LL). The light intensity in the present experiment was 2,000 lx and comparable to that used in the human experiments (12, 13, 22). In humans, a diurnal mammal, a type 1 PRC was constructed by using single light pulses of 2,500-5,000 lx of free-running circadian rhythms under temporal isolation (13), and a type 0 PRC was produced by three consecutive light pulses of ~10,000 lx (6). In rodents, the type of actograph used for behavioral data collection may also change the shape of the PRC, because access to a wheel is known to affect the free-running period (15, 35). In the previous studies, PRCs were all constructed using the circadian rhythm in wheel running activity, whereas in the present study, we measured the circadian rhythm in spontaneous locomotor activity.

A light pulse around the break point (CT18.5) abolished the circadian rhythm in 8 of 37 trials (Fig. 4). These animals were aperiodic for more than several weeks. The phenomenon is regarded as a singularity in which the circadian oscillation terminates (24). The circadian rhythm was not restored by the second light pulse of the same duration as the first (1 h), but it was restored by the third pulse of longer duration (6 h). Similar abolishment of circadian rhythmicity by a single light pulse was also reported in a palm squirrel (23). In humans, Jewett et al. (17) reported that circadian rhythms in rectal temperature and plasma cortisol were abolished by a single bright light pulse of long duration given in one or two successive circadian cycles. In the present study, circadian rhythmicity was lost in some animals for a few days after a light pulse but reappeared with a large phase shift. One potential mechanism for this eradication of the circadian rhythm is the uncoupling and internal desynchronization of constituent multiple oscillators of the circadian pacemaker. The multiple oscillator structure was recently revealed in the SCN of rats, in which a number of neurons exhibited a circadian rhythm in neuronal activity with a different period (16, 34). When the coupling among these neurons are disrupted, the circadian rhythm may disappear in the levels of the SCN, while persisting in individual neurons. Aperiodism due to a lack of neural cell adhesion molecule isoform suggested the importance of cell communication in the SCN for the circadian rhythm expression (30).

The present results are apparently different from our previous observations in which significant phase shifts were not produced by a light pulse given on the first day of free running (2). Similar discrepancy was also observed in humans (11). A light pulse in the subjective morning produced a significant phase-advance shift in the free-running rhythms under temporal isolation, whereas significant phase shift was not detected when pulsed on the first day of free running. This discrepancy might be explained by an after-effect of previous entrainment. It is well established that the free-running period is gradually deviated from the period of entrained rhythm until it reaches a steady-state value (25). In terms of a two-oscillator theory, the coupling between two constituent oscillators of the circadian pacemaker is influenced by the LD cycle to which the pacemaker entrains (7). It was demonstrated that the closer the period was to 24 h, the smaller the phase shift a light pulse produced (7). As a result, the light-induced phase shift is smaller on the first day than after many cycles of free running. A similar phenomenon was observed by Shimomura and Menaker (31) in the tau-mutant hamster, in which a type 1 PRC was detected 7 days after being released into DD, whereas a type 0 PRC was observed after 49 days in DD. They also explained the difference in terms of reduced coupling intensity between two constitutive oscillators during free running.

In the present experiment, a large phase shift was often accompanied by splitting of the activity band into two components that free ran with different periods for a few cycles until they fused (Fig. 2A). Split components did not fuse readily but continued to free run separately for >10 cycles in some chipmunks (Fig. 6). Furthermore, there was a significant correlation between the amount of phase shift and the change in the circadian period after a light pulse (Fig. 9). Splitting of the activity band and drastic change in  are explained by a hypothesis that the circadian pacemaker in chipmunks is composed of two oscillators having different periods. Two major oscillators have been suggested to exist within the rat SCN (32). A lack of daily resetting by light may reduce the coupling intensity between the oscillators and result in the change of phase relation between the PRCs of each oscillator (1); when a strong light pulse hits different phases of the two PRCs, a large phase shift, splitting, or a marked change of free- running period may occur. In the two-oscillator system, the looser the coupling, the higher the PRC amplitude is predicted (19).

Effects of 1- and 3-h pulses on the circadian phase were essentially the same. A 1-h light pulse was already enough to produce the maximum phase shifts in both directions for the high responders, while even a 3-h light pulse was still not enough to yield a significant phase shift in the nonresponder (Fig. 8). Nonresponders showed robust and stable free-running locomotor rhythm during the course of experiment as shown in Fig. 8. But many of the high responders also showed robust and stable locomotor activity (Figs. 1, 2, and 7) before the large phase shifts. Significant difference was not detected between the high and nonresponders in the free-running period, activity time, and activity levels. The difference in the light responsiveness could be explained by interindividual variation of the light sensitivity either at the level of the retina or at the pacemaker.

Free-running period was changed depending on the direction of the phase shift. Phase-delay shifts were accompanied by lengthening of the period, and advance shifts were accompanied by shortening. Light pulse significantly lengthened the free-running period at the phase-delaying phase of the PRC. At the late subjective day, light pulse affected the period, although it did not affect the phase. The results suggest that in addition to the phase response, phase-dependent  response is involved in the light entrainment of the chipmunks' circadian system.

In rodents, single gene mutation is reported to change the free-running period in two species, hamsters (28) and mice (33). The circadian period of the homozygote tau-mutant hamsters is shorter and that of the clock-mutant mice is longer than that of the respective wild-type animals by ~4 h. Furthermore, the clock-mutant mice become arrhythmic soon after they are released into DD (33). In chipmunks, a single light pulse changed the period of free-running rhythm up to 3.5 h. The difference was as large as the difference in the period between these mutant animals and their wild-type counterparts. Type 0 PRC is reported in both mutants, and type 1 PRC is found in the respective wild-type animals (5, 31). A lack of daily phase resetting by lights may disorganize the circadian system in tau and clock mutants and in chipmunks as well.

Chipmunks have similar characteristics in the circadian system to humans. The similar shape of the PRC (6), singularity (17), a high threshold of light for phase shifting (6, 14), uncoupling of two rhythm components with different periods under constant conditions (4), and circabidian rhythm (2, 12) are commonly observed in the two diurnal species. Chipmunks may provide a better animal model for the study of the human circadian system than nocturnal rodents.

It is concluded that a single light pulse induced type 0 resetting, singularity, and/or splitting of the activity band in the circadian system of diurnal chipmunks. The results suggest that the circadian pacemaker of chipmunks is composed of two major oscillators. These oscillators are further constituted by multiple circadian oscillators. The oscillatory coupling not only between the two major oscillators but also among the constitutive multiple oscillators is considerably influenced by light.