Retinal Development and The Expression Proles of Opsins Genes During Larvae Development in Takifugu Rubripes

Vision is the dominant sensory modality in sh and critical for the survival of sh larvae to detect predators or capture prey. The visual capacity of sh larvae is determined by the structure of the retina and the opsins expressed in the retinal and non-retinal photoreceptors. In this study, the retinal structure and opsin expression patterns during the early development stage of Takifugu rubripes larvae were investigated. At around two days after hatching days (dah), the yolk sac of T. rubripes disappeared, the mouth was clearly visible and the larvae started swimming and feeding on rotifers. Histological examination showed that at 1 dah, six layers were observed in the retina of T. rubripes larva, including the pigment epithelial layer (RPE), photoreceptor layer (PRos/is), outer nuclear layer (ONL), inner nuclear layer (INL), inner plexiform layer (IPL) and ganglion cell layer (GCL). At 2 dah, all eight layers were visible in the retina in T. rubripes larva, including RPE, PRos/is, ONL, outer plexiform layer (OPL), INL, IPL, GCL and optic ber layer (OFL). By measuring the thickness of each layer, opposing developmental trends were found in the thickness of ONL, OPL, INL and IPL, GCL and OFL. The nuclear density of ONL, INL and GCL and ratio of ONL/INL, ONL/GCL and INL/GCL were also measured and the ratio of ONL/GCL ranged from 1.9 at 2 dah to 3.4 at 8 dah and no signicant difference was observed between the different developmental stages (p > 0.05). No signicant difference was observed for the INL/GCL ratio between the different developmental stages, which ranged from 1.2 at 2 dah to 2.0 at 18 dah (p > 0.05). The results of quantitative real-time PCR showed that the expression of rhodopsin, LWS, SWS2, green opsin, rod opsin, opsin3 and opsin5 could be detected from 1 dah. These results suggested that the maturation of eye of T. rubripes occurred during the period of transition from endogenous to mixed feeding, explaining the need for vision-based survival skills during the early life stages after hatching and for the overall ecology and tness of T. rubripes.


es recognize
ifferent light wavelengths, while rods recognize the intensity of light.

In most vertebrates, light is captured through light-sensitive proteins called opsins, which are expressed in ciliary photoreceptor cells (c-opsins) and encode for G protein-coupled receptors that bind to a lightabsorbing, vitamin A-derived nonprotein retinal chromophore (Shichida et al. 1998;Palczewski et al. 2000; Collin et al. 2003;Cortesi et al. 2015; Kasagi et al. 2015).Depending on whether they are directly involved in visual imaging, opsins can be divided into two categories as visual and non-visual opsins (Terakita et al. 2005;Gao et al.2016).Visual opsin proteins are further grouped into ve classes spectrally tuned to absorb light at different wavelengths based on both phylogenetic and functional relationships, including medium to long wave-sensitive opsins detecting green to red (MWS or LWS), short wave-sensitive opsins sensitive to ultraviolet to blue (SWS1) and violet to blue (SWS2), rhodopsin which is only used in scotopic vision (RH1) and rhodopsin-like opsins that detect blue to green (RH2) (Hargrave et al. 1983 The puffer sh Takifugu rubripes of the order Tetraodontiformes, family Tetraodontidae and genus Takiugu, also known as Fugu, is one of the most important marine sh cultured in China, Japan, and Korea because of its delicious taste and high nutritional value.In China, about 70% of the annual product f T. rubripes is exported overseas and its domestic consumption is increasing since eating puffer sh became legal in China in 2016 (Katamachi et al. 2015;Kim et al. 2016;Zhang et al. 2019).At present, the scale of breeding of T. rubripes is expanding in the coastal area of China.With increasing customer demand, the arti cial breeding of T. rubripes has become particularly important and optimal lighting regimes for arti cial environments are needed based on species and developmental phases (Liu et al. 2019).More important for larval rearing, the lighting conditions need to be matched to the visual system of the target species.Thus, this study investigated the visual system of fugu, including the retinal development and the visual and non-visual protein expression during the early developmental stage of T. rubripes.


Materials And Methods


Larval rearing

The eggs of T. rubripes were obtained from Dalian Tianzheng Industrial Co., LTD in April 2020 and 2400 larvae were randomly distributed equally into three 300-L cylindrical tanks that were 80 cm high.From 2 days after hatching (dah), fugu larvae were fed with rotifer and artemia to satiation.To remove debris, excess feed, and dead larva and to maintain the quality of the rearing water, the bottom of the tank was cleaned and the water was renewed twice daily.Nitrites and ammonia were measured weekly, while temperature, salinity, pH, and dissolved oxygen were monitored daily.Mean values of nitrites and ammonia were always less than 0.05 mg L −1 and 0.2 mg L −1, respectively.Larva were maintained at a temperature of 19.0 to 21.0°C, pH was maintained at 7~8 and t

ined above 8 mg
L −1 .


Histology and retinal morphometric analyses

At 1, 2, 3, 4, 6, 8, 13, 18, and 26 dah, 20 fugu larva were anesthetized on ice and then xed in Bouin's uid for 24 to 28 h either as the whole sh at all stages up to 18 dah or just the heads at 26 dah.According to routine histological techniques, the samples were transferred into 70% ethanol until being prepared for histological analysis.After dehydration in a graded series of alcohol, the samples were embedded in para n.Four to six µm sections were cut and stained with hematoxylin-eosin staining (H&E).Sections were observed on a microscope (Leica DM4000 B LED, Leica, Wetzlar, GER) and photographed (Leica DFC450 C, Leica, Wetzlar, GER).Retinal measurements of the sections were made using image analysis software LAS X (Image Pro Plus, v. 4.5, Media Cybernetics Inc. Rockville, MD, USA) and involved measuring the thickness of each layer RPE, PRos/is, ONL, the ONL, OPL, INL, IPL, GCL, OFL, and the nuclei density of ONL, INL and GCL as the number of ONL nuclei/100 µm, INL nuclei/100 µm, and GCL/100 µm.For each of these parameters, six measurements were performed in the central, dorsal and ventral regions of each retina (n = 9/sampling point).The ratio of thickness of each retinal layer to total thickness (TT), the ratios of number of ONL nuclei to number of INL nuclei, the ratios of number of ONL nuclei to number of GCL nuclei and the ratios of number of INL nuclei to number of GCL nuclei were further calculated.

RNA extraction and Quantitative real-time PCR At 1, 2, 3, 4, 6, 8, 13, 18, and 26 dah, 20 fugu larva were anesthetized on ice, the whole sh up to 18 dah and the heads only at 26 dah were quickly immersed in ice-cold RNAlater RNA stabilization reagent (Ambion, Austin, TX, USA) and then were immediately transferred to an ultra-low temperature freezer at −80°C until processed.The RNA was extracted using Trizol reagent (Invitrogen, Carlsbad, CA, USA according to the manufacturer's instructions.The concentration and integrity of the RNA were checked using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) and a NanoDrop ND-1 ophotometer (Thermo Scienti c, Wilmington, DE, USA).

The speci c primers for reference genes and opsins were designed by primer premier 5.0 software (Table 1).Gene expression studies were performed by qPCR.The Applied Biosystems 7900 HT Real-Time PCR System (Applied Biosystems, Foster City, CA, USA) was used with the SYBR FAST qPCR Kit Master Mix (2×) Universal system (KAPA Biosystems, Boston, MA, USA), as recommended by the manufacturer.Brie y, an aliquot of 1 µg of RNA pretreated with DNase I (37°C, 30 min) was used as a template for cDNA synthesis with random hexamers according to the user information for 1st Strand cDNA Synthesis Kit (Takara Bio Inc., Otsu, Japan).Ampli cation was conducted with an initial denaturation at 95°C for 5 min, followed by 40 cycles of ampli cation at 95°C for 3 s and 60°C for 20 s.Melting curves were plotted to ensure that a single polymerase chain reaction (PCR) product ampli cation was obtained for each pair of primers.The stability of T. rubripes was veri ed using Norm nder (v 0.953) (17)(18)(19)(20) and the results showed that β-actin was comparatively more stable in the present study.The qPCRs were conducted in triplicates and the relative gene expression was calculated using 2 −ΔΔCT , where ΔCT = cycle threshold (CT) of the target gene minus the CT of β-actin, and ΔΔCT = ΔCT of an

us the calibrator sample method.Reverse GATCCCCAGATGCAA
AGAAC


Statistical analysis

One-way ANOVA followed by the Tukey's test (IBM SPSS statistics version 22.0, IBM, Chicago, IL, USA) was performed to examine the statistical signi cance of all data.A p value of < 0.05 was considered signi cant.


Results


Morphological observations of larvae in various stages

Hatching occurred over two days and hatching time was not synchronous (Fig. 1).At 1 dah, the brain appeared very large in lateral views and the larvae displayed the characteristic roundish body shape of a puffer sh.The larvae still had plenty of yolk and oil droplets within the yolk sac.The pigmentation of the larvae by both erythrophores (red), melanophores (black) and xanthophores (yellow) spread and covered the yolk sac and eye pigmentation emerged.The tail of the larvae was completely pigment free and there was a sharp margin where the pigmentation ended in front of the tail.Most of the larvae had the mouth located midventral between the eyes and punctate melanin on the yolk sac.At 2 dah, the tail of the larva was still completely pigment free, the mouth was clearly visible and the anterior tip of the mouth began protruding beyond the eyes.The larva seemed to have used up all its maternally supplied nutrients.Re ective iridophores became prominent in the eyes and the abdominal pigment gradually accumulated into stellate in most of the larvae.Most of the larvae started swimming, and feeding on rotifers.Pigmentation was more intense and extensive than was observed at 1 dah and especially increased at the dorsal region of the larva.At 3 dah, the dorsal n bud was clearly observed.At 8 dah, the teeth of larvae were formed, the air sac appeared and they began to attack each other.The pectoral ns were well developed and the caudal ns were clearly visible.From 18 dah, lamentous rays were observed on the caudal n and the melanophores began to appear in the tail.The n rays were observed on caudal ns at 26 dah.Larvae were also very similar to adult T. rubripes at that point (Fig. 1).


Observation of retinal microstructures at various stages

At 1 dah, six layers were observed in the retina in T. rubripes larva, including RPE, PRos/is, ONL, INL, IPL and GCL.At 2 dah, all eight layers were observed in the retina in T. rubripes larva, including RPE, PRos/is, ONL, OPL, INL, IPL, GCL and OFL (Fig. 2).The thickness of each layer was shown in The ratio of each layer thickness relative to the to

l thickness was also calculated in F
g.


Expression of retinal related genes

The results of qPCR showed that the expression of rhodopsin, LWS, SWS2, green opsin, rod opsin, opsin3 and opsin5 could be detected from 1 dah, as seen in Fig. 4. No signi cant difference was observed in the expression of rhodopsin from 1 to 4 dah, then the expression increased till 13 dah and then subsequently decreased.It again increased from 18 dah to a maximum at 26 dah (p 0.05).The gene expression of LWS increased gradually from 1 dah and reached at the maximum value at 26 dah.The gene expression of SWS2 initially increased from 1 dah, decreased from 4 and nally increased from 18 dah.The gene expression of green opsin initially increased from 1 dah and then decreased from 8 dah.The expression of rod opsin gradually increased from 1 dah, reached at the maximum value at 26 dah and no signi cant difference was observed from 1 to 8 dah (p > 0.05).No signi cant difference was observed in the expression of opsin3 from 1 to 13 dah (p > 0.05), with the maximum at 18 dah, which was signi cantly higher than other sampling points (p 0.05) and then decreased from 18 to 26 dah.The gene expression of Opsin5 increased from two dah and then decreased at 8 dah, and

hen increas
d from 13 dah again.


Discussion

Before rst feeding, the yolk is utilized for embryo development in most marine sh.As yolk reserves are gradually exhausted, larvae undergo the di cult transition from endogenous to mixed feeding period to obtain energy and required nutrients to support growth and development (Ma et al. 2010;Yúfera et al. 2014).Changes in the behavior of marine sh larvae are closely related to the development of the sensory organs, especially the visual system, which is crucial for feeding and predator defense in larval survival (Lim and Mukai 2014).It is very important for larvae to complete the differentiation of retinal functional cells and mouth-opening before the yolk sac is depleted because food intake requires coordination of food searching, detection, attract, capture and ingestion (Rønnestad et al. 2013, Lim and Mukai 2014, Hu et al. 2018).In this study, the yolk sac of T. rubripes disappeared and the larvae were mouth opening and feeding on rotifers at 2 dah.Histological observation showed that all the ten retinal layers were visible in T. rubripes from 2 dah, indicating that the well-developed visual system provided the necessary conditions for larval feeding.This result is consistent with other reports in teleosts, such as Danio rerio, where the iridophores are scattered over the retina, where the iris will develop and mouth is opening and feeding at 2 dah (Kimmel et al. 1995).In Epinephelus akaara larvae, yolk absorption was complete at 3 dah, and the mouth had opened at 4 dah.Hatched larvae had ONL, INL, and GCL starting from 2 to 3 dah, the retina was differentiated into PRE, PRos/is, ONL, OPL, INL, IPL, GCL and the choroid membrane was pigmented at 4 dah (Kim et al. 2013;Kim et al. 2019).In Engraulis anchoita, the eyes were pigmented and the GCL and the PR were visible in the retina, towards the end of yolk sac stage, when the larvae were 4 mm in length.This stage of acquisition of functionality coincided with the absorption of yolk and the beginning of exogenous feeding (Miranda et al. 2020).In Sparus aurata, the maturation of eye also occurred at 3 to 4 dah and it underwent profound anatomical and physiological alterations, such as the opening of the mouth and anus, the resorption of the yolk sac and functional differentiation of the alimentary canal, liver and pancreas (Parry et  The RPE layer and the PRos/is layer can be observed early from 1 dah in T. rubripes.Although the teleost eye is very similar to the mammalian eye, it is characterized by several unique structures, such as teleost eyes lake eyelids except for the nictitating membranes of certain sharks and most teleost cannot alter the size of their pupil (Kusmic and Gualtieri 2000;Reckel et al. 2002), so the sh retina is more susceptible to potential light-induced damage as they are continuously exposed to intense light.Consequently, alternative protective strategies have developed to cope with high light intensities, including migration of melanin granules and photoreceptor mobility.Acting as an anti-oxidant adjacent to the outer segments of photoreceptor cells, ocular melanin protects the retina against light-induced cell toxicity (Sanyal and Zeilmaker 1988), by migrating in an apical direction in response to light within processes of the RPE and enshroud photoreceptors compared to higher vertebrate (Allen and Hallows 1997).These photoreceptors can move into or out of the deep recesses of the RPE, so the RPE layer develops earlier in T. rubripes, which may be critical to protects the retina against light-induced cell toxicity.The results are consistent with some previous studies in other teleost.For example, in Acanthopagrus latus, Mugil cephalus and Alosa sapidissima, before the retina developed, a thin layer of RPE containing a few melanin particles was clearly observed at the edge of the retina (He et al. 1985;Xu et al. 1988, Gao et al. 2016).In this study, opposing developmental trends were found in the thickness of ONL and OPL, INL and IPL, GCL and OFL in T. takifugu.In teleost, the ONL is composed of the cell bodies of the cones and rods.The OPL contains the processes and synaptic terminals of rods, cones, horizontal cells and bipolar cells.The nuclei of the bipolar cells, amacrine cells, horizontal cells and Müller cells are found in the INL and the IPL consists of the connections between bipolar, amacrine and ganglion cells.The nuclei of ganglion cells form the GCL and the OFL contains the axons of ganglion cells as they collect to form the optic nerve (Fernald 1990 In a previous study, the ONL/INL ratio was used to estimate the degree of spatial summation of visual information at the rst synapse in the retina and compared to crepuscular at 1.4 to 1.7 or nocturnal 2.7 to 3.5 foraging species, diurnal feeding shes were shown to have a lower summation ratio of 0.5 to 1.4 (Munz and McFarland 1973;Schieber et al. 2012).In nocturnal shes, the higher ratio increases visual sensitivity by pooling the signals from many photoreceptors, at the expense of spatial resolving power, which re ects the adaptation to dim conditions (Munz and McFarland 1973).In the present study, the ONL/INL ratio of T. rubripes was 1.3 to 2.5 during the early developmental stage, suggesting tha the fugu larvae have higher visual sensitivity.As reported previously, the female T. rubripes lays demersal, adhesive eggs in coastal waters at a depth of 10 to 50 m during spring, and juveniles remain in nursery ground areas near the main spawning grounds from spring to summer, and then enter wider areas (Katamachi et al. 2015;Kim et al. 2016;Zha

et al. 2019)
The higher ratio of the ONL/INL in T. rubripes re ects an adaptation to their surrounding light conditions.However, Schieber et al. (2012) examined the retinal anatomy of four elasmobranch species with differing ecologies, including Port Jackson shark Heterodontus portusjacksoni, the bull shark Carcharhinus leucas, pink whip

y Himantu
a fai, and the epaulette shark Hemiscyllium ocellatum and found that the ONL:INL ratio may be a less robust indicator of diel activity patterns in elasmobranchs.The ratio of the nucleus number of the ONL layer to the GCL layer re ects the degree of retina network convergence (Xu et al. 1998), and this re ects the degree of visual sensitivity and light sensitivit