When food is scarce, being able to anticipate the time when it is available has great survival benefits. The timing of food-seeking behaviour is thought to be driven by circadian body clocks that can be reset by nutritional cues, but how these food-entrainable oscillators work at the molecular level is unknown. The two period genes, Per1 and Per2, are known to be crucial for light-induced resetting of the main body clock, which is housed in the hypothalamic suprachiasmatic nucleus (SCN). Reporting in Current Biology, Feillet and colleagues investigated the role of these genes in food anticipation.

Nocturnal animals, which are normally most active at night, will shift their activity pattern if their only meal occurs in daytime. Indeed, wild-type mice exposed to restricted feeding times or to hypocaloric food show bouts of locomotor activity just before mealtimes. The researchers found that while mice lacking Per1 displayed similar food-anticipatory behaviour, Per2 mutants were unable to predict the time of food access. When tested in constant light or constant dark conditions, Per2 mutants still could not anticipate food availability, indicating that this inability was independent of light cues and that Per2 is crucial for the capacity to anticipate the timing of meals.

Next, the researchers set out to test whether Per1 and Per2 have a role in the food-induced resetting of the SCN. This central oscillator regulates the free-running rhythms of, amongst others, locomotor activity and temperature that become apparent when animals are housed in constant darkness. In wild-type mice, food restriction causes a phase advance in these rhythms. However, the free-running rhythm of Per1 mutants' locomotor activity displayed a phase delay, whereas that of Per2 mutants showed a large phase advance.

At the molecular level, the expression of clock genes Bmal1 and Cry1 in the SCN in response to limited food availability was also different in Per1 and Per2 mutants compared with wild-type mice. Furthermore, in Per2 mutants, expression of the clock-controlled genes Dbp and Avp in the SCN was greatly reduced.

Although these results indicated that Per1 and Per2 have a role in the resetting of the central body clock by nutritional cues, peripheral food-entrainable oscillators did not require Per1 and Per2. Using quantitative PCR, the researchers showed that food restriction caused phase shifts in the expression of clock genes Bmal1 and Rev-Erbα and the clock-controlled gene Dbp in the liver and kidney that were similar in Per1 and Per2 mutants and in wild-type mice.

This study provides evidence that that Per1 and Per2 are involved in the mechanism by which nutritional cues entrain the central, but not peripheral clocks and shows for the first time that Per2 is crucial for the ability to anticipate circadian cycles of food availability.