In mammals, day-length-sensitive (photoperiodic) seasonal breeding cycles depend on the pineal hormone melatonin, which modulates secretion of reproductive hormones by the anterior pituitary gland [1]. It is thought that melatonin acts in the hypothalamus to control reproduction through the release of neurosecretory signals into the pituitary portal blood supply, where they act on pituitary endocrine cells [2]. Contrastingly, we show here that during the reproductive response of Soay sheep exposed to summer day lengths, the reverse applies: Melatonin acts directly on anterior-pituitary cells, and these then relay the photoperiodic message back into the hypothalamus to control neuroendocrine output. The switch to long days causes melatonin-responsive cells in the pars tuberalis (PT) of the anterior pituitary to increase production of thyrotrophin (TSH). This acts locally on TSH-receptor-expressing cells in the adjacent mediobasal hypothalamus, leading to increased expression of type II thyroid hormone deiodinase (DIO2). DIO2 initiates the summer response by increasing hypothalamic tri-iodothyronine (T3) levels. These data and recent findings in quail [3] indicate that the TSH-expressing cells of the PT play an ancestral role in seasonal reproductive control in vertebrates. In mammals this provides the missing link between the pineal melatonin signal and thyroid-dependent seasonal biology.

Seasonal synchronization based on day length (photoperiod) allows organisms to anticipate environmental change. Photoperiodic decoding relies on circadian clocks, but the underlying molecular pathways have remained elusive [1]. In mammals and birds, photoperiodic responses depend crucially on expression of thyrotrophin β subunit RNA (TSHβ) in the pars tuberalis (PT) of the pituitary gland [2-4]. Now, using our well-characterized Soay sheep model [2], we describe a molecular switch governing TSHβ transcription through the circadian clock. Central to this is a conserved D element in the TSHβ promoter, controlled by the circadian transcription factor thyrotroph embryonic factor (Tef). In the PT, long-day exposure rapidly induces expression of the coactivator eyes absent 3 (Eya3), which synergizes with Tef to maximize TSHβ transcription. The pineal hormone melatonin, secreted nocturnally, sets the phase of rhythmic Eya3 expression in the PT to peak 12 hr after nightfall. Additionally, nocturnal melatonin levels directly suppress Eya3 expression. Together, these effects form a switch triggering a strong morning peak of Eya3 expression under long days. Species variability in the TSHβ D element influences sensitivity to TEF, reflecting species variability in photoperiodic responsiveness. Our findings define a molecular pathway linking the circadian clock to the evolution of seasonal timing in mammals.

The 24-h expression of seven clock genes (Bmal1, Clock, Per1, Per2, Cry1, Cry2, and CK1 ) was assayed by in situ hybridization in the suprachiasmatic nucleus (SCN) and the pars tuberalis (PT) of the pituitary gland, collected every 4 h throughout 24 h, from female Soay sheep kept under long (16-h light͞8-h dark) or short (8-h light͞16-h dark) photoperiods. Locomotor activity was diurnal, inversely related to melatonin secretion, and prolactin levels were increased under long days. All clock genes were expressed in the ovine SCN and PT. In the SCN, there was a 24-h rhythm in Clock expression, in parallel with Bmal1, in antiphase with cycles in Per1 and Per2; there was low-amplitude oscillation of Cry1 and Cry2. The waveform of only Per1 and Per2 expression was affected by photoperiod, with extended elevated expression under long days. In the PT, the high-amplitude 24-h cycles in the expression of Bmal1, Clock, Per1, Per2, Cry1, and Cry2, but not CK1 , were influenced by photoperiod. Per1 and Per2 peaked during the day, whereas Cry1 and Cry2 peaked early in the night. Hence, photoperiod via melatonin had a marked effect on the phase relationship between Per͞Cry genes in the PT. This supports the conclusion that an ''external coincidence model'' best explains the way photoperiod affects the waveform of clock gene expression in the SCN, the central pacemaker, whereas an ''internal coincidence model'' best explains the way melatonin affects the phasing of clock gene expression in the PT to mediate the photoperiodic control of a summer or winter physiology.

Many species express endogenous cycles in physiology and behavior that allow anticipation of the seasons. The anatomical and cellular bases of these circannual rhythms have not been defined. Here, we provide strong evidence using an in vivo Soay sheep model that the circannual regulation of prolactin secretion, and its associated biology, derive from a pituitary-based timing mechanism. Circannual rhythm generation is seen as the product of the interaction between melatonin-regulated timer cells and adjacent prolactin-secreting cells, which together function as an intrapituitary "pacemaker-slave" timer system. These new insights open the way for a molecular analysis of long-term timing mechanisms.