Zebrafish Dark-Dependent Behavior Requires Phototransduction by the Pineal Gland

Abstract

Located dorsally underneath a thin translucent skull in many teleosts, the pineal gland is a photoreceptive organ known as a key element of the circadian clock system. Nevertheless, the presence of additional routes of photoreception presents a challenge in determining its specific roles in regulating photic-related behavior. Here, we show the importance of the pineal gland in mediating a prolonged motor response of zebrafish larvae to sudden darkness, both as a photodetector and as a circadian pacemaker. This was evident by a reduced motor response of Bsx-deficient larvae, lacking a pineal gland, to sudden darkness. Moreover, the typical daily rhythm of the intensity of this response was lost in the pineal-less larvae. In contrast, motor response to a sudden increase in illumination was unaffected. Furthermore, we show that the pineal-mediated behavioral response to darkness requires two elements: the photoreceptor cells and the projecting neurons. Dark response was impaired in larvae whose pineal photoreceptor cells were genetically ablated and in larvae whose pineal projecting neurons had undergone laser-axotomy. This study thus establishes the pineal gland as a mediator of dark-dependent behavior and reveals underlying cellular components involved in transducing information about darkness to the brain.

Keywords: VMR; brain‐specific homeobox (Bsx); photoreceptors; pineal; projecting neurons; zebrafish.

Figure 1

Prolonged behavioral response to darkness is attenuated in pineal‐less Bsx‐deficient larvae. (A) 3 day‐activity of 6–8 dpf bsx ^−/− mutant larvae (n = 21, dark gray) and bsx ^+/− control siblings (n = 22, light gray) under 12–12 h LD cycles (marked by alternating white and black horizontal boxes). The bold line marks the average smoothed activity within the group, in cm/10 min, and the light‐colored background marks SEM. (B) Average velocity (cm/sec) of each genotype 5 s before and after transitions to light (left) or darkness (right). There was no difference between genotypes in the magnitude of either startle responses (p > 0.05), measured by the displacement during the 1‐second‐long spike of high velocity movement that follows the transition. (C) No difference between genotypes in prolonged light response (p > 0.05). Left—activity (cm/min) 10 min before and after transition to light. Arrows represent the maximum activity per group following the change in illumination. Right—average prolonged light response per group, measured as the maximum activity (marked by the arrow on the left panel) normalized by the baseline activity 2 min before transition (see Supporting Information S1: Figure 1). Bars represent group means ± SEM. (D) Prolonged dark response was severely impaired in bsx ^−/− mutants compared with control siblings (p << 0.001***). Left—activity (cm/min) 10 min before and after transition to dark. Arrows represent the maximum activity per group following the change in illumination. Right—mean prolonged light response per group, measured as the maximum activity (marked by the arrow on the left panel) normalized by the baseline activity 2 min before transition. (E) 24 h‐activity of bsx ^−/− mutants (n = 10, dark gray) and control siblings (n = 39, light gray) under 2–2 h LD cycles. (F) Prolonged dark response was severely attenuated in bsx ^−/− mutants compared with that of control siblings (Mixed model genotype effect, p << 0.001***). The magnitude of response differed across transitions in control siblings with a daily rhythm, but attenuation in bsx ^−/− mutants was observed in all time points. Group statistics, All p values were obtained by post hoc analysis tests of a linear mixed model with BH correction (see Materials and methods). Asterisks represent p values, where “**” is under 0.01 and “***” is under 0.001. Full statistics are detailed in Supporting Information S2: Table 1, experiments 1–4.

Figure 2

Attenuation of prolonged dark response in pineal‐less Bsx‐deficient larvae cannot be attributed to visual impairment. (A) Comparison of average prolonged dark response between, from left to right, WT siblings (controls, solid gray), pineal‐less bsx ^−/− mutants, eyeless rx3 ^−/− mutants (chokh) and bsx ^−/− ;rx3 ^−/− double‐mutant siblings. Under both 12–12 h LD cycles (left and Supporting Information S1: Figure 2A) and 2–2 h LD cycles (right), eyeless rx3 ^−/− mutants exhibited equivalent average prolonged dark response to that of controls (p > 0.05 in both experiments). Double‐mutants exhibited impaired prolonged dark response compared with controls (p < 0.001*** in both experiments), consistent with Bsx‐deficiency. Bars represent group means ± SEM. (B) Under 12–12 h LD cycles (see Supporting Information S1: Figure 3A), blind atoh5 ^−/− mutants (lakritz, dark gray) exhibited average prolonged dark response equivalent to that of controls (light gray, atoh7 ^+/− or atoh7 ^+/+ , p> 0.05), despite lack of startle response (Supporting Information S1: Figure 3B). Sample sizes: (A left) 20 controls, 11 bsx ^−/− , 44 rx3 ^−/− , 17 bsx ^−/− ;rx3 ^−/− . (A right) 39 controls, 10 bsx ^−/− , 34 rx3 ^−/− , 13 bsx ^−/− ;rx3 ^−/− . (B) 23 atoh7−/−, 25 controls. All p values were obtained by post hoc analysis tests of a linear mixed model with BH correction. Asterisks represent p values, where “*” is under 0.05 and “***” is under 0.001. Full statistics are detailed in Supporting Information S2: Table 1, experiments 2–4 and 7.

Figure 3

Prolonged dark response is attenuated in pineal photoreceptors‐ablated larvae (“PRC‐ablated”). (A) Expression of nitroreductase and mCherry in Tg(tph2:NfsB‐mCherry)^y227 and Tg(aanat2:Gal4);Tg(UAS:NfsB‐mCherry)^c264 larvae is restricted to the pineal gland photoreceptors. (B) 24 h‐activity of Mtz‐treated tph2:NfsB (n = 48, dark gray) and WT sibling (control, n = 48, light gray) larvae under 2–2 h LD cycles (marked by alternating white and black boxes). Mtz‐treated larvae were less active and less responsive than previous controls (see Supporting Information S1: Figure 4). The bold line marks the average smoothed activity within the group, in cm/10 min, and the light‐colored background marks SEM. (C) Activity (cm/min) 10 min before and after a selected transition to darkness. (D) Prolonged dark response is impaired in both PRC‐ablated lines. Bars represent group means ± SEM. All p values were obtained by post hoc analysis tests of a linear mixed model with BH correction. Asterisks represent p‐values, where “**” is under 0.01. Full statistics are detailed in Supporting Information S2: Table 1, experiments 8–11.

Figure 4

Laser‐ablation of pineal projecting neurons' axons (“PN‐axotomized”) impairs prolonged dark response. (A) Exemplary PN‐axotomized Tg(foxd3:eGFP) larva showed severe degeneration, followed by partial regeneration of the PN axons. Left: immediately before laser‐ablation of the PN axons. The pineal and right eye are marked. Middle: 24 h postablation. The PN axons were severely degenerated. Right: 48 h postablation. The PN axons were partially regenerated. (B) Activity of 23 5 dpf PN‐axotomized Tg(foxd3:eGFP) larvae (dark gray) and 23 untreated control siblings (light gray) under 14–10 h LD cycles. Activity, in seconds moved per minute, is presented 60 min before and after the transition to darkness. The bold line marks the average smoothed activity within the group, and the light‐colored background marks SEM. (C) PN‐axotomized larvae exhibited an average reduction of prolonged dark response, 25% compared with controls in normalized activity during the 5 min after transition to darkness (p = 0.0063**, Wilcoxon's rank sum test). Activity after the transition was normalized by the baseline as before. Bars represent group means ± SEM. Asterisks represent p values, where “**” is under 0.01. Full statistics are detailed in Supporting Information S2: Table 1, experiments 12 and 13.