Effect of short- and long-term melatonin treatments on the reproductive activity of the tropical damselfish Chrysiptera cyanea

Photoperiod plays a role in controlling the initiation and termination of reproduction in fish. Melatonin is an internal transducer of environmental photoperiod and is involved in regulating reproduction. The present study aimed to examine how melatonin impacts the transcript levels of kisspeptin (kiss1 and kiss2), gonadotropin-releasing hormones (gnrh1), and the β-subunit of gonadotropins (fshβ and lhβ) in the brain of the sapphire devil, a tropical damselfish with long photoperiod preference. Feeding mature females with melatonin-containing pellets inhibited increases in the transcript levels of kiss1, gnrh1, and lhβ within 3 h. Continuous melatonin treatment for 1 week resulted in oocyte regression and downregulation of kiss2, gnrh1, fshβ, and lhβ. When the transcript levels of kiss1 and gnrh1 were measured at 4-h intervals in the brain of sapphire devil, a day–high/night–low fluctuation was observed. The hypothalamic–pituitary–gonadal (HPG) axis may be influenced by melatonin, exerting a negative effect at night because the transcript levels of aralkylamine N-acetyltransferase (aanat2) increased during the scotophase. The expression of aanat2 was higher under short-day than long-day conditions, suggesting that there is a seasonal change in melatonin levels at night. It was concluded that change in photoperiod becomes a key factor for controlling the hormone synthesis in the HPG axis through melatonin.


Introduction
Melatonin is an indoleamine synthesized mainly in the pineal organ and the retina and plays a crucial role in various physiological processes including development (Danilova et al. 2004), growth (Falcón et al. 2003), and reproduction (Amano et al. 2000(Amano et al. , 2004bVera et al. 2007). Melatonin levels fluctuate following a day-low/night-high cycle, which is controlled by the rate-limiting enzyme aralkylamine N-acetyltransferase (AANAT), which has a photodegradable characteristic and is affected by clock genes (Coon et al. 1999;Falcón et al. 2007). In addition to a daily cycle, certain fish also exhibit seasonal changes to melatonin secretion according to day length (García-Allegue et al. 2001;Bromage et al. 2001). Seasonal change in melatonin secretion is likely an endogenous cue for the initiation and termination of seasonal reproduction in fish since melatonin levels increase during the winter months, corresponding to the longer length of the night (Porter 2000;Porter et al. 2001;Bromage et al. 2001). Melatonin has previously been reported as a suppressor of reproductive activity in long-day breeders. For example, the gonadal development of the spotted snakehead Channa punctatus was inhibited after immersion in water containing melatonin, or after injection of melatonin (Renuka and Joshi 2010). On the other hand, melatonin acts as a stimulator of reproductive activity in short-day breeders. In underyearling precocious male masu salmon Oncorhynchus masou, testicular development was stimulated when fish were fed with melatonin-containing pellets, mimicking short-day conditions (Amano et al. 2000). However, the role of melatonin in reproductive activity may not be generalizable because of contradicting reports of the effect of melatonin. In the zebrafish Danio rerio, a long-day breeder, melatonin had a stimulatory effect on ovarian development when fish were reared in melatonin-containing water for 10 days (Carnevali et al. 2011). Interestingly, low-dose melatonin administration showed an opposite effect (stimulation) on testicular maturation in the masu salmon (Amano et al. 2000). Melatonin may impact neurons synthesizing gonadotropin-releasing hormone (GnRH) in the brain as well as cells synthesizing gonadotropin (GtH) in the pituitary gland (Amano et al. 2000(Amano et al. , 2004aGhosh and Nath 2005;Carnevali et al. 2011). Since the timing and amplitude of the changes in melatonin levels differ among fish species (Bromage et al. 2001), it is hypothesized that there is diverse adaptation in melatonin action on reproduction. In addition, it is not known how fish inhabiting low-latitude shallow waters can utilize melatonin for regulation of reproduction because little change in day length as well as water temperature occurs at these latitudes.
The sapphire devil Chrysiptera cyanea is a tropical damselfish that is widely found in shallow reefs of the Indo-West Pacific region. Previous reports showed that this species adapted to subtropical environments undergoes active spawning from May through August (Bapary et al. 2009) and that it can utilize changes in day length and food availability for the initiation and termination of seasonal reproduction (Bapary and Takemura 2010;Bapary et al. 2012). Immersing fish in melatonin-containing seawater inhibited the ovarian development of the sapphire devil (Badruzzaman et al. 2013), although how melatonin acts on the hypothalamic-pituitary-gonadal (HPG) axis remains unclear. The present study aimed to evaluate the short-term (days) and long-term (weeks) effects of melatonin treatment on mRNA levels of kisspeptin (kiss1 and kiss2), GnRH (gnrh1), and the β-subunit of GtH (fshβ and lhβ) in the brain of the sapphire devil. Diurnal profiles of kiss1, gnrh1, and aanat2 in the brain were also studied to confirm the relationship between their abundance and the effect of melatonin. The present study focused on aanat2 because it expresses specifically in the pineal organ of fish (Coon et al. 1999;Isorna et al. 2011).

Animals
Mature sapphire devils (1.35-2.83 g body weight) were collected on coral reefs around Sesoko Island, Okinawa, Japan, using a hand net during the daytime at low tide. The fish were acclimated in stock tanks with running seawater and ambient aeration under natural photoperiod and ambient water temperature at Sesoko Station, Tropical Biosphere Research Center, University of the Ryukyus, Okinawa, Japan. Fish were fed daily at 0900 h with commercial pellets (EP1; Nisshin-Marubeni, Tokyo, Japan).

Melatonin treatment
Melatonin was purchased from Sigma-Aldrich (St. Luis, MO, USA) and dissolved with absolute ethanol. Melatonin-containing pellets were prepared by 1 3 Vol.: (0123456789) spraying pellets with a melatonin-containing ethanol solution, as previously described (Badruzzaman et al. 2013). The final concentration of melatonin was adjusted to 500 μg/g dry pellets. Vehicle without melatonin was sprayed onto pellets as a control (melatonin-free pellets).
The experiments were conducted to evaluate the effects of short-and long-term treatment with melatonin on the mRNA levels of reproductionrelated genes (kiss1, kiss2, gnrh1, fshβ, and lhβ) in the brain of the sapphire devil. The experiments were carried out in May and June when this species undergoes active reproduction (Bapary et al. 2009). Fish were randomly taken from the stock tanks and transferred to four glass aquaria (60 L capacity) equipped with running seawater and ambient aeration under natural photoperiod and water temperature. Plastic pipes were provided as shelter and spawning nest and were placed on the bottom of each aquarium. The two aquaria, which contained 50 individuals (40 females and 10 males per tank) and 40 individuals (32 females and 8 males per tank) were prepared for the control and treatment group, respectively. After acclimation for several days, fish were fed with pellets twice per day, at 0900 h and 1500 h. The daily ration of food was adjusted to 2.0% of body weight throughout the experimental period.
For the short-term melatonin treatment, females (n = 8) were sampled from the aquarium of the control group at 0900 h (initial control), and then from the aquaria of both groups at 1200 h and 1500 h. For the long-term melatonin treatment, females (n = 8) in each aquarium were sampled at 1200 h after 1 and 2 weeks. Following anesthesia with 2-phenoxyethanol (Kanto Chemical, Tokyo, Japan) and euthanasia by decapitation, body mass and ovarian mass of all individuals were recorded. Then, pieces of the ovary were fixed in Bouin's solution for subsequent histological observation. Gonadosomatic index (GSI) was calculated using the following formula: GSI = (ovarian mass/body mass) × 100. The whole brain was excised, immediately frozen in liquid nitrogen, and stored at −80°C until extraction of total RNA from the tissue samples using TriPure Isolation Reagent (Roche Diagnostics, Indianapolis, IN, USA) according to the manufacturer's instructions.
Daily variation of sapphire devil kiss1 and gnrh1 mRNA In June, females (n = 42) were housed in filtered aquaria (60 L capacity) with ambient aeration under natural photoperiodic conditions (long-day; 0538-1926 h). After acclimation, fish (n = 6 per sampling time) were taken from the aquarium at 4-h intervals. After anesthesia, the whole brain was excised, immediately frozen in liquid nitrogen, and stored at −80°C until total RNA extraction. Sample collections during the scotophase were performed under dim red light.
Daily variation of sapphire devil aanat2 mRNA under programmed photoperiod This experiment was conducted in February using females (n = 72) with body mass ranging from 1.25 to 1.85 g, which were housed in filtered aquaria with ambient aeration under short-day (LD = 10:14; lighton from 0800-1800 h) and long-day (LD = 14:10; light-on from 0600-1800 h), which were set using a 20 W fluorescent light (1200 lx at water surface). After acclimation under these conditions for at least 1 week, fish (n = 5 per sampling time) were taken from aquaria at 4-h intervals. The following procedures were followed by the abovementioned methods.

Histological procedures
Following dehydration with a series of ethanol and permutation with xylene, ovaries were embedded in histoparaffin (m.p. 56-58°C; Merck, Darmstadt, Germany), serially sectioned at 5 μm, and then stained with hematoxylin and eosin for microscopic observation. According to oocyte staging of the white-spotted spinefoot Siganus canaliculatus (Hoque et al. 1998), oocytes were classified into the following six stages: the peri-nucleolus stage (PNS), oil droplet stage (ODS), yolk vesicle stage (YVS), primary yolk stage (PYS), secondary yolk stage (SYS), and tertiary yolk stage (TYS). Atretic oocyte (AO) was also observed in the present study.

Real-time quantitative PCR (qPCR)
The transcript levels of kiss1, kiss2, gnrh1, fshβ, lhβ, and aanat2 in the brain of the sapphire devil were 1 3 Vol:. (1234567890) assayed using the CFX96 Real Time System (Bio-Rad, Hercules, CA, USA) and GoTaq qPCR Master Mix (Promega, Madison, WI, USA). Total RNA was extracted from the whole brain of the sapphire devil. The first strand of cDNA was reverse-transcribed from 1 μg of total RNA using the PrimeScript RT Reagent Kit with gDNA Eraser (Takara Bio, Otsu, Japan). Specific qPCR primers were designed based on the full-length of the cDNA fragments (Table 1), which were previously cloned and characterized (Imamura et al. 2020). Each PCR was carried out in a final volume of 10 μl, containing 5 μl of 2× GoTaq qPCR Master Mix, 0.3 μl of forward and reverse primers, 2.4 μl of nuclease-free water, and 2 μl of cDNA template. The qPCR cycling conditions were 95°C for 2 min, then 40 cycles at 95°C for 15 sec and 60°C for 1 min. Plasmid DNA as standard was produced using the gene-specific primers. cDNA at 10-fold dilution was used for qPCR to construct the standard curve. The transcript levels of the target genes were normalized to those of elongation factor 1 alpha (ef1α), which served as an internal control.

Statistical analyses
All data are expressed as the mean ± standard error of the mean (SEM). Changes in GSI, as well as mRNA levels of reproduction-related genes (kiss1, kiss2, gnrh1, fshβ, and lhβ), were compared by twoway analysis of variance (ANOVA) followed by Tukey's test. Diurnal variation of kiss1, gnrh1, and aanat were compared by one-way ANOVA followed by Tukey's test using P < 0.05 as the level of statistical significance.

Effects of melatonin treatment on ovarian development and reproduction-related genes
The mature female sapphire devils were fed with melatonin-containing pellets (treatment group) or melatonin-free pellets (control group) at 0900 h and sacrificed at 0, 3, and 6 h after treatment. Females of the initial control had high GSI values. This value did not change within 6 hours after treatment (Fig. 1A). Histological observation revealed that ovaries of both groups contained vitellogenic oocytes at TYS (Supplementary data 1). The transcript levels of reproduction-related genes (kiss1, kiss2, gnrh1, fshβ, and lhβ) were determined in the brain of the sapphire devil ( Fig. 1B-F). The transcript levels of kiss1 and gnrh1 increased significantly (P < 0.05) in the brain of the control fish at 3 h. The transcript levels of kiss1, gnrh1, and lhβ decreased significantly in the melatonin-treated fish at 3 h (P < 0.05). However, there was little difference in the transcript levels of these genes between the groups at 6 h ( Fig. 1B, D, and F). Melatonin treatment did not affect the transcription levels of kiss2 and fshβ in the brain of sapphire devil females ( Fig. 1C and E).
Feeding with melatonin-containing or melatoninfree pellets lasted for 2 weeks (Fig. 2). High GSI was maintained in the control group throughout the experimental period. When fish were treated with melatonin, GSI significantly decreased at 1 and 2 weeks ( Fig. 2A). Melatonin treatment had an impact on the transcript levels of kiss2, gnrh1, fshβ, and lhβ, but not kiss1. Significantly lower mRNA levels of kiss2 at 1 week, and gnrh1, fshβ, and lhβ at 1 and 2 weeks were observed (Fig. 2B-F). Histological observation revealed that all ovaries at 1 and 2 weeks after treatment contained vitellogenic oocytes at TYS. AO and immature oocytes at PNS were observed in ovaries of the melatonin-treated fish (Supplementary data 2).
Diurnal variation of kiss1 and gnrh1 mRNA during spawning season Fish were sampled at 4-h intervals to evaluate the daily profiles of kiss1 and gnrh1 in the brain (Fig. 3). Both  genes showed clear daily variation with an increase during photophase and a decrease during scotophase. A peak of mRNA abundance occurred at 1200 h for kiss1 (Fig. 3A) and 1600 h for gnrh1 (Fig. 3B).
Diurnal variation of aanat2 mRNA under long-and short-day conditions Fish were reared under short-and long-day conditions to investigate the daily profile of aanat2 in the brain. The transcript levels of aanat2 fluctuated with an increase toward the start of scotophase in both conditions. Statistically higher transcript levels of aanat2 were observed at 1700 h and 2100h for short-day condition (Fig. 4A) and 2300 h for long-day conditions (Fig. 4B).

Discussion
Administration of melatonin has previously been reported to affect-either positively or negatively-the gonadal activity of fishes (Urasaki 1972;Sundararaj and Keshavanath 1976;Ghosh and Nath 2005;Renuka and Joshi 2010). However, the results obtained may be due to intraspecies differences in the effect of melatonin and differences in experimental design. For example, intraperitoneal injections of melatonin inhibited the ovarian development of the female walking catfish Clarias batrachus (Ghosh and Nath 2005) and medaka Oryzias latipes (a long-day spawner) (Urasaki 1972). On the other hand, ovarian development of the zebrafish was stimulated when fish were reared in melatonin-containing water under long-day conditions (Carnevali et al. 2011). Our results support an inhibitory effect of melatonin treatment on ovarian development including vitellogenesis of the sapphire devil because feeding with melatonin-containing pellets for 1 to 2 weeks resulted in the induction of atresia and the arrest of recruitment in vitellogenic oocytes in the ovaries, as previously reported (Badruzzaman et al. 2013). Therefore, it is likely that melatonin acts on the  Fig. 1 Short-term effect of melatonin treatment on gonadosomatic index (GSI) (A) and relative expression levels of reproduction-related genes (kiss1, kiss2, gnrh1, fshβ, and lhβ) (B-F) in the female sapphire devil. Fish were fed with melatonin-free pellets or melatonin-containing pellets at 0900 h. Ovary and brain were taken from fish at 0, 3, and 6 h after feeding. White and black columns indicate the data from the fish feeding with melatonin-free pellets and melatonin-containing pellets, respectively. Each value represents mean ± SEM. Different letters indicate significant difference at P < 0.05 Physiol Biochem (2022) 48:253-262 257 synthesis and/or secretion of hormones in the HPG axis because the direct action of melatonin is exerted through the binding of G-protein-coupled melatonin receptors (Dubocovich et al. 2010), which are expressed in the diencephalon (Ikegami et al. 2009), pituitary gland (Gaildrat and Falcón 2000), and ovary of fish (Jin et al. 2013;Chai et al. 2013;Hong et al. 2014). The present study also revealed that treatment with melatonin for at least 1 week is needed to retract ovarian features, because no change was observed in vitellogenic oocytes after melatonin treatment for 3 or 6 h. The present study revealed that there is a negative impact of melatonin on the transcript levels of certain reproduction-related genes in the HPG axis. Effects of melatonin on their expressions seem to be marked during the reproductive season because low transcription was recorded in developing ovaries with vitellogenic oocytes (Imamura et al. 2020). Unlike GSI and histological features of the ovary, changes in the transcript levels of kiss1, gnrh1, and lhβ, but not kiss2 and fshβ, occurred within 3 h of melatonin treatment.

Fish
These results demonstrate that melatonin suppresses a diurnal increase in the mRNA levels of melatonintreated fish, suggesting that melatonin acts as a diurnal driver of kiss1 and gnrh1 transcription. Previous studies have reported diurnal variation of kiss1 and gnrh1 transcription levels in the brain of certain teleosts; sbGnRH decreased during the dark period in the pituitary of the European seabass Dicentrarchus labrax reared under both long-day and short-day conditions (Bayarri et al. 2004). The transcript levels of kiss2 and gnrh1 had a marked day-night difference with an increase at 1200 h in the brain of orange-spotted grouper Epinephelus coioides in developing and mature stages (Chai et al. 2013). In the latter case, the melatonin receptor (MT1) had a reciprocal relationship in regulating the day-night variation of gnrh1 expression directly or indirectly through Kiss2 (Chai et al. 2013). A similar regulation is likely to occur in the sapphire devil because a day-high/night-low fluctuation of kiss1 and gnrh1 transcription was observed in the brain of female sapphire devils (Fig. 3). An Long-term effect of melatonin treatment on gonadosomatic index (GSI) (A) and relative expression levels of reproduction-related genes (kiss1, kiss2, gnrh1, fshβ, and lhβ) (B-F) in the female sapphire devil. Fish were fed with melatonin-free pellets or melatonin-containing pellets at 0900 h. Ovary and brain were taken from fish at 0, 1, and 2 weeks after feeding. White and black columns indicate the data from the fish feeding with melatonin-free pellets and melatonin-containing pellets, respectively. Each value represents mean ± SEM. Different letters indicate significant difference at P < 0.05 Physiol Biochem (2022) 48:253-262 258 inverse profile of aanat2 (Fig. 4) also supports the involvement of melatonin in the downregulation of kiss1 and gnrh1, since AANAT is the rate-limiting enzyme in melatonin synthesis in the pineal organ in fish (Coon et al. 1999;Isorna et al. 2011). The expression of lhβ was also downregulated in the brain of the sapphire devil, suggesting that lhβ transcription is a downstream target of GnRH1. In addition, there is likely a direct action of melatonin through its receptor in the pituitary gland because the expression of melatonin receptors was evidenced in certain teleosts including the pike Esox lucius (Gaildrat et al. 2002), the European sea bass (Sauzet et al. 2008), and the grass puffer Takifugu niphobles (Ikegami et al. 2009). Little difference in the transcript levels of kiss1, gnrh1, and lhβ was observed in the brain of the sapphire devil at 6 h after melatonin treatment, suggesting clearance of exogenous melatonin and decrement of its effectiveness. In this regard, melatonin was reported to be eliminated within 2 h from the blood circulation of the threespot wrasse Halichoeres trimaculatus when melatonin was injected by intraperitoneal injection ). We also evaluated the effect of weekly treatment with melatonin on ovarian development and gene expression in the sapphire devil. This experiment was founded on the previous finding that vitellogenic oocytes could be induced in the ovary when fish out of the breeding season were reared under long-day conditions within a defined range of water temperature (Bapary and Takemura 2010). The present study revealed that, in agreement with the regression and immature features of the ovary, melatonin treatment decreases kiss2, gnrh1, fshβ, and lhβ transcription levels in the brain of the sapphire devil. Thus, long-term treatment with melatonin has a negative impact on the transcript levels of reproductive genes in the HPG axis. Similar findings have been shown in certain fish species undergoing seasonal reproduction (Ghosh and Nath 2005;Sébert et al. 2008;Fukunaga et al. 2019). For example, daily injections of melatonin for 2 weeks decreased the plasma levels of GtH in the walking catfish C. batrachus (Ghosh and Nath 2005), and an implanted osmotic pump releasing physiological levels of melatonin downregulated lhβ in the pituitary of the honeycomb grouper Epinephelus merra (Fukunaga et al. 2019). Since the honeycomb grouper and the sapphire devil are both tropical species with long-day preference, it is likely that a decrease in photoperiod (an increase in melatonin effect at night) is related to the termination of breeding activity. It was previously reported that immersing female sapphire devils in estradiol-17β (E2)-containing water resulted in increased expression levels of kiss2 and gnrh1 in the brain, suggesting that E2 has a positive feedback on kiss-expressing neurons in the diencephalon, and in turn, upregulates gnrh1 in the preoptic region (Imamura et al. 2020). The inhibition of this melatonin feedback control may accelerate regression of gonadal development in the sapphire devil.

Fish
The present study showed that high transcription of aanat2 (over 1.2) was maintained longer in the brain of the sapphire devil when fish were reared under short-day conditions. This result indicates that the period of melatonin synthesis is longer under shortday conditions than long-day conditions since melatonin synthesis is dependent upon the expression of aanat2 in the pineal gland (Herrera-Pérez et al. 2011;McStay et al. 2014). Since melatonin levels in the pineal organ vary according to photoperiodic conditions, seasonal changes in melatonin may be related to the recognition of the season (García-Allegue et al. 2001).
In conclusion, the ovarian development of the sapphire devil is influenced by melatonin, which has a negative impact on hormone synthesis at the HPG axis. Gradual increases in the nocturnal duration of melatonin synthesis in the seasonal progression toward winter may trigger the termination of the spawning season in long-day spawners, including the sapphire devil. Further studies would be needed to clarify how melatonin is involved in the initiation and termination of ovarian development through the circadian system in the brain.