Mammalian Peripheral Circadian Oscillators Are Temperature Compensated

Variations in ambient temperature present a unique obstacle to the timekeeping function of circadian clocks. Most biological reactions proceed with a temperature coefficient (Q10) ∼ 2 or 3, so that with every 10 °C increase in temperature, the reaction rate approximately doubles or triples. Therefore, a temperature-dependent clock would run faster at high temperatures than at low temperatures, and would not be a reliable predictor of time of day. So it is not surprising that mechanisms have evolved to ensure that the length of the circadian period remains relatively constant over a wide range of temperatures, a phenomenon known as temperature compensation. In mammals, self-sustaining rhythms have been measured in the master circadian clock, located in the SCN, and in peripheral tissues in vitro (Yamazaki et al., 2000; Yoo et al., 2004). When separated from the entrainment of the SCN, each peripheral tissue expresses tissue-specific differences in circadian period and phase (Yoo et al., 2004). While previous studies have shown that the circadian period in the SCN, retina, and in fibroblast cell lines remains relatively constant across a range of temperatures, it is unknown whether mammalian peripheral clocks are temperature compensated (Tosini and Menaker, 1998; Ruby et al., 1999; Izumo et al., 2003; Tsuchiya et al., 2003). To examine the effect of temperature on circadian oscillations, we measured PERIOD2::LUCIFERASE (PER2::LUC) rhythms in explants of central and peripheral tissues from mPer2Luc mice at 29, 31, 33, 35, and 37 °C. At temperatures ranging from 31 to 37 °C, robust rhythms of PER2::LUC were measured in the SCN, pituitary gland, cornea, adrenal gland, and lung. In the liver, the rhythm of PER2::LUC was robust at 37 °C, but damped within 2 cycles at all other temperatures examined. At 29 °C, only the pituitary consistently maintained a robust PER2::LUC rhythm. To determine if circadian clocks in mammalian central and peripheral tissues were temperature compensated, we calculated the average period of each tissue at various temperatures. The Q10 of the SCN, pituitary gland, cornea, adrenal gland, and lung were variable, but all were close to 1, suggesting that these tissues were temperature compensated (Fig. 1). Because the PER2::LUC rhythm of liver explants was not robust at temperatures below 37 °C, the Q10 of the liver could not be calculated. Figure 1 Central and peripheral oscillators are temperature compensated. The mean periods (hours) of SCN, pituitary, cornea, adrenal glands, lung, and liver harvested from mPer2Luc mice are expressed as a function of temperature (°C). Data are expressed ... Each peripheral tissue had a different Q10, ranging from 0.89 to 0.96. Recent studies have reported tissue-specific expression patterns of circadian genes and mutating these genes results in variable, tissuespecific effects on the expression of other clock components (Lowrey and Takahashi, 2004; Sato et al., 2004; Akashi and Takumi, 2005; Guillaumond et al., 2005; Noshiro et al., 2005; Ko and Takahashi, 2006). Therefore, while the molecular mechanisms controlling temperature compensation are unknown, it is possible that tissue-dependent differences in the Q10 of the mammalian periphery results from variations in the expression and regulation of circadian genes. Since Per2 participates in the transcriptional/translational feedback loops that regulate the expression of other circadian genes, temperature-induced changes in the bioluminescent waveform of PER2::LUC expression could reflect variable processing of the components of the feedback loops. However, we found no differences in the waveforms of PER2::LUC expression between tissues cultured at 31 and 37 °C (Fig. 2). It is possible that the effects of temperature on the molecular mechanism of the clock are not reflected in rhythmic PER2::LUC expression and other circadian genes should also be assessed. Alternatively, our method may not be sensitive enough to detect temperature-induced changes in waveform. Figure 2 The waveform of PER2::LUC expression in various tissues is not dependent on temperature. SCN, pituitary, lung, and cornea from mPer2Luc mice were cultured at 31 or 37 °C. The baseline-subtracted data for each tissue was averaged, normalized, and ... Temperature-dependent variations in the molecular timekeeping mechanism could also be reflected in the phase of PER2::LUC expression. When we compared tissues cultured at 31 or 37 °C, we found that the phase of the PER2::LUC rhythm was temperature dependent in SCN slices, but not in the pituitary, lung, or cornea (Fig. 2). Upon further analysis, we found that the circadian phase of PER2::LUC expression was gradually delayed with decreasing temperature (Fig. 3). The phase shift induced by changing the ambient temperature in our experiments is consistent with a previous study that demonstrated that the SCN entrains to temperature cycles (Herzog and Huckfeldt, 2003). Figure 3 Temperature-dependent phase of PER2::LUC expression in the SCN. The circadian time in hours (CT12 is lights-off) of the first peak of PER2::LUC expression (phase) in the SCN was averaged (expressed as the mean ± SEM) and plotted as a function ... In summary, our results demonstrate that the periods of mammalian peripheral tissues are temperature compensated over the range of 29 to 37 °C. We find tissuespecific differences in the Q10 and in temperature-induced shifts of circadian phase.