Daily light and feeding cycles act as powerful synchronizers of circadian rhythmicity. Ultimately, these external cues entrain the expression of clock genes, which generate daily rhythmic behavioral and physiological responses in vertebrates. In the present study, we investigated clock genes in a marine teleost (gilthead sea bream). Partial cDNA sequences of key elements from both positive (Bmal1, Clock) and negative (Per2, Cry1) regulatory loops were cloned before studying how feeding time affects the daily rhythms of locomotor activity and clock gene expression in the central (brain) and peripheral (liver) oscillators. To this end, all fish were kept under a light-dark (LD) cycle and were divided into three experimental groups, depending on the time of their daily meal: mid-light (ML), mid-darkness (MD), or at random (RD) times. Finally, the existence of circadian control on gene expression was investigated in the absence of external cues (DD + RD). The behavioral results showed that seabream fed at ML or RD displayed a diurnal activity pattern (>91% of activity during the day), whereas fish fed at MD were nocturnal (89% of activity during the night). Moreover, seabream subjected to regular feeding cycles (ML and MD groups) showed food-anticipatory activity (FAA). Regardless of the mealtime, the daily rhythm of clock gene expression in the brain peaked close to the light-dark transition in the case of Bmal1 and Clock, and at the beginning of the light phase in the case of Per2 and Cry1, showing the existence of phase delay between the positive and negative elements of the molecular clock. In the liver, however, the acrophases of the daily rhythms differed depending on the feeding regime: the maximum expression of Bmal1 and Clock in the ML and RD groups was in antiphase to the expression pattern observed in the fish fed at MD. Under constant conditions (DD + RD), Per2 and Cry1 showed circadian rhythmicity in the brain, whereas Bmal1, Clock, and Per2 did in the liver. Our results indicate that the seabream clock gene expression is endogenously controlled and in liver it is strongly entrained by food signals, rather than by the LD cycle, and that scheduled feeding can shift the phase of the daily rhythm of clock gene expression in a peripheral organ (liver) without changing the phase of these rhythms in a central oscillator (brain), suggesting uncoupling of the light-entrainable oscillator (LEO) from the food-entrainable oscillator (FEO). These findings provide the basis and new tools for improving our knowledge of the circadian system and entraining pathways of this fish species, which is of great interest for the Mediterranean aquaculture

Highlights d During evolution, the cavefish P. andruzzii has lost photoreactivation DNA repair d Only P. andruzzii and placental mammals are known to lack photoreactivation d The D-box enhancer coordinates DNA repair in response to ROS, UV, and visible light d Loss of D-box function in P. andruzzii underlies the lack of photoreactivation SUMMARY How the environment shapes the function and evolution of DNA repair systems is poorly understood. In a comparative study using zebrafish and the Somalian blind cavefish, Phreatichthys andruzzii, we reveal that during evolution for millions of years in continuous darkness, photoreactivation DNA repair function has been lost in P. andruzzii. We demonstrate that this loss results in part from loss-of-function mutations in pivotal DNA-repair genes. Specifically, C-terminal truncations in P. andruzzii DASH and 6-4 photolyase render these proteins predominantly cytoplasmic, with consequent loss in their functionality. In addition, we reveal a general absence of light-, UV-, and ROS-induced expression of P. andruzzii DNA-repair genes. This results from a loss of function of the D-box enhancer element, which coordinates and enhances DNA repair in response to sunlight. Our results point to P. andruzzii being the only species described, apart from placental mammals, that lacks the highly evolutionary conserved photoreactivation function. We predict that in the DNA repair systems of P. andruzzii, we may be witnessing the first stages in a process that previously occurred in the ancestors of placental mammals during the Mesozoic era.

Light represents the principal signal driving circadian clock entrainment. However, how light influences the evolution of the clock remains poorly understood. The cavefish Phreatichthys andruzzii represents a fascinating model to explore how evolution under extreme aphotic conditions shapes the circadian clock, since in this species the clock is unresponsive to light. We have previously demonstrated that loss-of-function mutations targeting non-visual opsins contribute in part to this blind clock phenotype. Here, we have compared orthologs of two core clock genes that play a key role in photic entrainment, cry1a and per2, in both zebrafish and P. andruzzii. We encountered aberrantly spliced variants for the P. andruzzii per2 transcript. The most abundant transcript encodes a truncated protein lacking the C-terminal Cry binding domain and incorporating an intronic, transposon-derived coding sequence. We demonstrate that the transposon insertion leads to a predominantly cytoplasmic localization of the cavefish Per2 protein in contrast to the zebrafish ortholog which is distributed in both the nucleus and cytoplasm. Thus, it seems that during evolution in complete darkness, the photic entrainment pathway of the circadian clock has been subject to mutation at multiple levels, extending from opsin photoreceptors to nuclear effectors.

During early development, most organisms display rhythmic physiological processes that are shaped by daily changes in their surrounding environment (i.e., light and temperature cycles). In fish, the effects of daily photocycles and their interaction with temperature during early developmental stages remain largely unexplored. We investigated the existence of circadian rhythms in embryonic development and hatching of three teleost species with different daily patterns of behavior: diurnal (zebrafish), nocturnal (Senegalese sole), and blind, not entrained by light (Somalian cavefish). To this end, fertilized eggs were exposed to three light regimes: 12 h of light: 12 h of darkness cycle (LD), continuous light (LL), or continuous darkness (DD); and three species-appropriate temperature treatments: 24°C, 28°C, or 32°C for zebrafish and cavefish and 18°C, 21°C, or 24°C for sole. The results pointed to the existence of daily rhythms of embryonic development and hatching synchronized to the LD cycle, with different acrophases, depending on the species: zebrafish embryos advanced their developmental stage during the light phase, whereas sole did so during the dark phase. In cavefish, embryogenesis occurred within 24 h post fertilization (hpf) at the same pace during day or night. The hatching rhythms appeared to be controlled by a clock mechanism that restricted or "gated" hatching to a particular time of day/night (window), so that embryos that reached a certain developmental state by that time hatch, whereas those that have not wait until the next available window. Under LL and DD conditions, hatching rhythms and the gating phenomenon persisted in cavefish, in zebrafish they split into ultradian bouts of hatching occurring at 12-18-h intervals, whereas in sole DD and LL produced a 24-h delay and advance, respectively. Hatching rates were best under the LD cycle and the reported optimal temperature for each species (95.2±2.7% of the zebrafish and 83.3±0.1% of the cavefish embryos hatched at 28°C, and 93.1±2.9% of the sole embryos hatched at 21°C). In summary, these results revealed that hatching rhythms in fish are endogenously driven by a time-keeping mechanism, so that the day and time of hatching are determined by the interplay between the developmental state (temperature-sensitive) and the circadian clock (temperature-compensated), with the particular phasing being determined by the diurnal/nocturnal behavior of the species.