Morphological and functional development of the spiral intestine in cloudy catshark (Scyliorhinus torazame)

Cartilaginous fish (sharks, rays and chimaeras) have evolved many unusual physiological features, including aspects of osmoregulation (Smith, 1936; Hyodo et al., 2014), reproduction (Musick and Ellis, 2005) and gastrointestinal morphology (Wilson and Castro, 2010). The spiral intestine is considered to be a plesiomorphic feature of vertebrate and is found in the non-teleostean actinopterygii, non-tetrapod sarcopterygii, elasmobranchii and holocephalii (Argyriou et al., 2016). The comparatively short intestine in these fishes is offset by the spiral arrangement of the mucosa (Khanna, 2004), significantly increasing the surface area of the intestine up to 3- and 6-fold depending on morphology and species (Bertin, 1958). One explanation for the physiological advantage of the spiral intestine in elasmobranchs is to maximize intestinal surface area in a limited body space that is dominated by a large lipid-dense liver (White, 1937). Digesta rotate along the longitudinal axis of the intestine, resulting in a slower evacuation time and further increasing absorptive capacity (Wetherbee et al., 1987). In cartilaginous fishes, the anatomy of the intestine has been well described in several species of shark and ray (Perrault, 1671; Parker, 1880; Khanna, 2004; Chatchavalvanich et al., 2006), and the spiral structures have been classified as spiral, ring or scroll type (White, 1937; Compagno et al., 2005; Leigh et al., 2017). However, we know little regarding the morphological and functional development of the intestine in elasmobranchs.

Unlike teleosts, cartilaginous fish are characterized by an extremely long developmental period in ovo or in utero. During early development, the richly vascularized yolk sac membrane contributes to absorption of nutrients by the developing embryos (Jollie and Jollie, 1967; Hamlett et al., 2011; Conrath and Musick, 2012). In viviparous species, embryos begin to use the gastrointestinal tract around the middle of gestation. In southern stingray (Dasyatis americana), lipid-rich histotroph known as uterine milk is secreted by the uterine villi and embryos drink the milk in utero (Hamlett et al., 1996), and in the crocodile shark (Pseudocarcharias kamoharai), embryos ingest trophic eggs in utero (Fujita, 1981). Conversely, oviparous embryos depend entirely on yolk; however, this has been shown to be present within the lumen of the embryonic intestine during development (Lechenault et al., 1993; Haines et al., 2005). In the oviparous lesser spotted dogfish (Scyliorhinus canicula), yolk internalization occurred approximately after the first third of embryonic life (Ballard et al., 1993) and in the ovoviviparous spiny dogfish (Squalus acanthias) yolk was firstly observed in the intestinal lumen of 70 mm length embryo, and then fat droplets were observed in intestinal epithelial cells when the embryos were approximately 150 mm in length (Tewinkel, 1943). These findings raise the hypothesis that, even in oviparous elasmobranchs, the spiral intestine is formed and probably functional for nutrient absorption well before hatching (Tewinkel, 1943; Rodda and Seymour, 2008). In the present study, we focused on the timing of the ‘pre-hatching’ event in the onset of nutrient absorption by the embryonic intestine. Pre-hatching occurs in oviparous elasmobranchs at the end of the first third of development and is a period when the egg casing is known to open to the environment to facilitate embryonic respiration and growth (Mellinger et al., 1986; Diez and Davenport, 1987; Mellinger et al., 1989). This hypothesized developmental trajectory is distinctly different from oviparous ray-finned fishes that typically hatch before complete development of the gastrointestinal tract (Ng et al., 2005). Unravelling the functional development of the spiral intestine is thus important for our understanding of embryonic growth in cartilaginous fish.

In the present study, we describe the functional development of the spiral intestine in the oviparous cloudy catshark [Scyliorhinus torazame (Tanaka 1908)]. To this end, we investigated (1) formation of the spiral structure, esophageal opening and rectum opening, (2) formation of the secondary folds and microvilli that are related to nutrient absorption and (3) the timing of when yolk first appears in the embryonic intestine. In addition, as protein is a major diet component in fishes (Clements and Raubenheimer, 2006), and peptide transporter 1 (PepT1) and the neutral amino acid transporter Slc6a19 are known to contribute to nutrient absorption across the teleost intestine (Verri et al., 2003; Rønnestad et al., 2007; Terova et al., 2009; Rimoldi et al., 2015; Xu et al., 2016; Orozco et al., 2017, 2018), we also examined spatiotemporal expression of mRNAs encoding PepT1 and Slc6a19 in the intestine as functional gene expression markers. We provide evidence of a functional intestine in the oviparous cloudy catshark shortly after the pre-hatching event, suggesting that pre-hatching represents a physiological turning point in the embryonic development of oviparous cartilaginous fish.

Spawned cloudy catshark (S. torazame) eggs collected from captive individuals were transported from Ibaraki Prefectural Oarai Aquarium to the Atmosphere and Ocean Research Institute at the University of Tokyo. They were kept in a 1000-liter tank with recirculating natural seawater under a constant photoperiod (12 h:12 h light:dark) at 16°C.

Embryos were staged according to the developmental stages established in lesser spotted dogfish, S. canicula (Ballard et al., 1993). Approximately 2 months post-fertilization at the end of the first third of development classified as the middle of developmental stage 31, the anterior part of the egg capsule opens and is known as the ‘pre-hatching’ or ‘eclosion’ (Ballard et al., 1993; Takagi et al., 2017). For our experiment, we further subdivided this stage into two phases: the first half prior to pre-hatch (stage 31 early or 31E) and the second half following pre-hatch (stage 31 late or 31L). Occurrence of pre-hatching was confirmed by gentle pressure on the egg capsule; if intracapsular fluid was released, pre-hatching had occurred. Once hatched, juveniles were maintained under similar environmental conditions and fed a diet of chopped squid twice a week ad libitum. Embryos and juveniles were sampled throughout the year from 2015 to 2020. For sampling, both embryos and juveniles were anesthetized in 0.02% ethyl 3-aminobenzoate methanesulfonate (Sigma-Aldrich, St Louis, MO, USA). All animal experiments were conducted according to the Guidelines for Care and Use of Animals approved by the ethics committee of the University of Tokyo (A16-13).

The embryos from developmental stage 23 to 34 and juveniles were used for histological observation. For juveniles and embryos at developmental stages 32, 33 and 34, spiral intestines were dissected out, and were fixed in formalin with picric acid (saturated picric acid:formalin 3:1) at 4°C overnight. Spiral intestines of embryos earlier than stage 32 were fixed in toto. Fixed tissues were washed twice in 70% ethanol, dehydrated in graded concentrations of ethanol, cleared with benzene, and embedded in Paraplast (McCormick Scientific, Richmond, IL, USA). Serial sections were cut at 6 µm thickness and mounted onto glass slides (Matsunami Glass, Osaka, Japan). The sections were stained with hematoxylin and eosin. First, rehydrated tissue sections were treated with hematoxylin solution (Wako Pure Chemical Industries, Osaka, Japan) for 5 min and then washed in running tap water for 30 min. The sections were then stained with eosin (Chroma Gesellschaft Schmid, Koengen, Germany) for 10 min, dehydrated through graded concentrations of ethanol, cleared with xylene, and mounted with Permount (Thermo Fisher Scientific, Waltham, MA, USA). Micrographs were obtained using a digital camera (DSRi1; Nikon, Tokyo, Japan). For three-dimensional (3D) reconstruction of the developing spiral intestine, sequential images of serial sections of the intestine were aligned using ImageJ (Fiji) (Schindelin et al., 2012) and the luminal space of intestine was filled using Photoshop (Adobe Systems, San Jose, CA, USA) before a 3D image was created using the 3D viewer of the ImageJ.

For scanning electron microscopy observation, embryos of stages 27, 31E and 32, and juveniles, were examined. After dissection, spiral intestines of juveniles and embryos of stages 31E and 32 were pre-fixed in a fixative containing 2% paraformaldehyde and 2.5% glutaraldehyde in 0.05 mol l⁻¹ cacodylate buffer (CB, pH 7.4) (PG solution) for 3 min to harden the surface of the tissue. Tissue samples were cut using a razor blade and then fixed again in the PG solution for 2 h. After washing in 0.05 mol l⁻¹ CB for 5 min, samples were fixed in 1% osmium tetroxide for 1 h, washed in 0.05 mol l⁻¹ CB, dehydrated through a graded concentration of ethanol, immersed in butyl alcohol, and then dried in a vacuum freeze-dryer (JFD-300, JEOL Ltd, Tokyo, Japan) overnight. Tissues mounted on a sample holder were coated with platinum palladium using ion sputtering apparatus (E-1030, Hitachi High-Technologies Corporation, Tokyo, Japan) and were kept in a vacuum desiccator until use.

Due to the small size of the samples, the intestines from stage 27 embryos were cryosectioned to expose the gut cavity. The embryos were pre-fixed in the PG solution for 30 min. After washing in 0.05 mol l⁻¹ CB, samples were immersed at 4°C in 10% glucose in 0.05 mol l⁻¹ CB for 10 min, 20% glucose for 30 min, 30% glucose for 40 min, optical cutting temperature (OCT) compound (Sakura Finetek Japan, Tokyo, Japan) for 30 min, and were then frozen in new OCT compound. Blocks were sliced at 5 µm using a cryostat until the intestinal lumen was exposed. After rinsing the residual OCT compound in 0.05 mol l⁻¹ CB and post-fixation in the PG solution for 2 h, tissues were processed for scanning electron microscopy as described above. All observations were made using a scanning electron microscope (s-4800, Hitachi High-Technologies Corporation) at a magnification of ×10,000 or ×20,000.

Complementary DNA cloning was performed as previously described in detail (Takagi et al., 2017) using whole spiral intestine dissected from juvenile cloudy catsharks. Total RNA was extracted using a commercially available kit (ISOGEN, Nippon Gene, Toyama, Japan). DNase-treated total RNA (1 µg) was reverse-transcribed to first-strand cDNA using an iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA, USA), following the manufacturer's instructions. Putative nucleotide sequences of mRNAs encoding catshark PepT1, Slc6a19 and β-actin were obtained by BLAST search on catshark transcriptome database Squalomix (https://transcriptome.riken.jp/squalomix/) (Hara et al., 2018). Partial cDNA fragments, which were amplified with Kapa Taq Extra DNA polymerase (Kapa Biosystems, Boston, MA, USA) and gene specific primer sets (Table S1), were ligated into pGEM-T Easy plasmid vector (Promega, Madison, WI, USA). The nucleotide sequences were determined by a DNA sequencer (ABI PRISM 3130, Life Technologies, Carlsbad, CA, USA).

Molecular phylogenetic analyses were conducted in MEGA7 (Kumar et al., 2016). The deduced amino acid sequences of catshark PepT1 and Slc6a19 were aligned with those of other animals using MUSCLE (Edgar, 2004). Molecular phylogenetic trees were inferred using the maximum likelihood method based on the Le_Gascuel_2008 model (Le and Gascuel, 2008). A discrete gamma distribution was used to model evolutionary rate differences among sites (five categories; +G parameter=0.4843). The bootstrap number was set to 1000.

Whole intestine was obtained from juveniles and embryos of stages 31E, 31L, 32, 33 and 34. Intestines were divided equally along the longitudinal axis into three parts: anterior, middle and posterior. Total RNA extraction and first-strand cDNA synthesis were conducted as described above. Quantitative PCR (qPCR) was performed using a 7900 HT Sequence Detection System (Life Technologies) with a Kapa SYBR Fast qPCR Kit (Kapa Biosystems), as previously described in detail (Hasegawa et al., 2016). The primers used for real-time qPCR were designed using PrimerQuest (http://sg.idtdna.com/primerquest/home/index) (Table S1). β-Actin mRNA was used as an internal control (Takagi et al., 2017). The plasmids containing cloned cDNA fragments were used as the known amounts of standard templates for absolute quantification in qPCR analyses.

Digoxigenin (DIG)-labeled cRNA probes were synthesized from plasmids containing PepT1 and Slc6a19 coding sequences with DIG RNA Labeling Kit (Roche Applied Science, Mannheim, Germany). Tissue sections were mounted onto MAS-GP-coated glass slides (Matsunami Glass). Hybridization was conducted using a previously described protocol (Takabe et al., 2012). In brief, deparaffinized sections were treated with 5 µg ml⁻¹ proteinase K (Sigma-Aldrich) in PBS for 10 min at 37°C, immersed in pre-hybridization buffer (5×SSC: 0.75 mol l⁻¹ NaCl and 83.3 mmol l⁻¹ sodium citrate; pH 7.0; 50% formamide) at 58°C and hybridized with DIG-labeled cRNA probes in hybridization buffer (50% formamide, 5×SSC, 40 µg ml⁻¹ bovine calf thymus DNA) at 58°C for 40 h. Sections were then washed in 2×SSC for 30 min at room temperature, 2×SSC for 1 h at 65°C, and 0.1×SSC for 1 h at 65°C. After incubation with alkaline phosphatase-conjugated anti-DIG antibody (Roche) for 2 h at room temperature, signals were visualized with 4-nitro blue tetrazolium chloride (450 mg ml⁻¹) and X-phosphate/5-bromo-4-chloro-3-indolyl-phosphate (175 mg ml⁻¹) until sections were colored at 25°C in the dark. Sections were counterstained with Nuclear Fast Red (Vector Laboratories, Burlingame, CA, USA).

Data are expressed as means±s.e.m. Statistical analyses were conducted using Kyplot5.0 software (KyensLab Inc., Tokyo, Japan). Normality of data was checked by the Shapiro–Wilk test, followed by a Steel–Dwass or Tukey–Kramer multiple comparison test. P-values less than 0.05 were considered statistically significant.

Catshark embryos continued to develop inside the egg capsule for approximately 6 months prior to hatching (Fig. 1A–E). Histological investigation revealed that the spiral structure of intestine was fully formed prior to the pre-hatching period. At stage 23 of embryonic development (approximately 1 month post-fertilization), a straight intestinal tract was visible in the body cavity, which consisted of pseudostratified epithelium and mesenchyme (Fig. 1F). At stage 24, two separated intestinal lumens were observed in sagittal sections, showing the presence of one or two spiral turns (Fig. 1G,O). The number of turns continued to increase after stage 24 (Fig. 1H–J). At stage 31E, the number of spiral turns reached seven (Fig. 1K), which is equal to that of juveniles (Fig. 1L). The number of spiral turns was confirmed by 3D reconstruction using serial sections (Fig. 1M,N).

At stage 23 in development, the yolk inside the external yolk sac (EYS) was black, and the embryo and the yolk stalk were transparent under transmitted light (Fig. 2A). At stage 24, sagittal sections revealed that the lumen of the yolk stalk was directly connected to the intestinal lumen (Fig. 2B, dotted line); however, a monolayer cellular membrane prevented yolk from flowing into the yolk stalk (Fig. 2C, arrows). At development stage 31E, yolk was not observed in the intestinal lumen (Fig. 2D); however, at stage 31L, white yolk was visible inside the yolk stalk (Fig. 2E, arrow) and eosinophilic yolk granules were observed in the intestine (Fig. 2F, arrow).

To determine the precise timing of yolk inflow into the embryonic intestine, we further examined four individuals, the developmental stages of which were within 48 h after pre-hatching. Within 24 h after pre-hatching, one embryo had no yolk (Fig. S1A), while a second individual showed yolk granules inside the intestinal lumen (Fig. S1B). At 36–48 h after pre-hatching, two embryos had yolk granules inside the intestinal lumen (Fig. S1C,D), implying that yolk inflow into the intestine is likely to occur within 48 h after pre-hatching.

Intestinal epithelium remained smooth without folding until pre-hatching (Fig. 3A,C), but in some individuals a slight protrusion of the intestinal epithelium was observed around the pre-hatching event, and obvious secondary folds were confirmed at stage 32 (Fig. 3E, arrows). The length of secondary folds gradually increased as the animal grew (Fig. 3E,G) and elongated secondary folds (Fig. 3G) and dense microvilli (Fig. 3H) were observed in the juvenile stage post-hatch. Microvillus formation began prior to secondary fold formation. Short microvilli were sparsely distributed in the stage 27 embryo (Fig. 3B, arrows). At stage 31E, microvilli covered the apical surface of the intestinal lumen (Fig. 3D) and both density and length of the microvilli increased in the juvenile stage (Fig. 3F,H).

At stage 25, the esophagus was cylindrical in structure, filled with mesenchyme, and could be distinguished from the oral cavity and stomach (Fig. 4A,B). At stage 27, the luminal epithelium was differentiated, but the esophagus remained closed (Fig. 4C,D, arrowheads). By stage 33, the esophagus appeared closed by adhering the dorsal and ventral sides of epithelia (Fig. 4E,F); however, this adhesion seemed not to be tight as some gaps were sparsely observed (Fig. 4F, arrows). The leaky but closed esophagus was observed also in the stage 34 embryo (Fig. 4H, arrows), and the closed portion became narrower longitudinally at this stage (Fig. 4G). We confirmed the opening of esophagus in fed juveniles (Fig. 4I,J).

In the lesser spotted dogfish embryo, the rectum was shown to be closed until hatch (Mellinger et al., 1986, 1987). The serial transverse sections from the end of the spiral intestine to the rectum revealed that the lumen of the rectal canal, which forms the junction between the spiral intestine (Fig. 5C) and the rectum (Fig. 5E), was not fully developed and closed at stage 34 (Fig. 5D). In the hatched juvenile, formation of the rectal canal is completed, and the lumen was continuously formed from the spiral intestine to the rectum (Fig. 5H–J).

By searching the catshark transcriptome database, contigs showing high homology to pept1 and slc6a19 were found, and fragments of cDNAs encoding the catshark PepT1 and Slc6a19 were cloned. The accession numbers for the partial sequences of catshark pept1 and slc6a19 are LC507941 and LC507942, respectively. Maximum-likelihood analysis clearly shows that the catshark cDNAs belong to the vertebrate pept1 (slc15a1) and slc6a19 lineages, respectively (Fig. 6).

To evaluate nutrient absorptive function of embryonic intestine, expression of pept1 and slc6a19 mRNAs were examined by quantitative RT-PCR (qPCR) and in situ hybridization (Figs 7–9). Low levels of pept1 mRNA expression were detected at stages 31E and 31L (Fig. 7A). The expression levels significantly increased at stage 32, and thereafter remained high until the juvenile stage. slc6a19 mRNA showed relatively low levels of expression during the embryonic period, and showed an increasing trend in expression in juveniles, although the changes between stages were statistically insignificant (Fig. 7B).

In the Mozambique tilapia, expression of pept1 mRNA was shown to be rich in the anterior region of the intestine (Orozco et al., 2017), therefore we divided the spiral intestine into anterior, middle and posterior sections, and examined regional expression of pept1 and slc6a19 mRNA. pept1 mRNA was uniformly distributed throughout the spiral intestine at and before stage 33, whereas from stage 34 onwards, expression of pept1 mRNA was higher in the anterior region of the spiral intestine compared with the posterior region (Fig. 8A). Consistent with the results of qPCR, moderate signals were found in the entire region of intestine at stage 32 by in situ hybridization, and anteriorly biased signals were shown as developmental stage progressed (Fig. 9A,E,I). In the juvenile intestine, pept1 mRNA showed intense signals in secondary folds constituting the first and second spiral turns (Fig. 9M) and moderate levels of pept1 mRNA signals were observed in the epithelial cells of third to fifth spiral turns, but signals were faint in the posterior part of spiral intestine. The pept1 mRNA signals were limited to the epithelial cells (Fig. 9B,F,J,N, arrowheads) and no expression was observed in submucosa.

Expression of slc6a19 mRNA exhibited different patterns compared with pept1 mRNA (Fig. 8B). At stage 32, mRNA levels of slc6a19 in the anterior region were significantly higher compared with levels in the middle region. At stage 33, expression of slc6a19 increased in the posterior region gradually and became significantly higher at stage 34. In the juvenile intestine, no difference was observed in slc6a19 mRNA levels among the three regions. In agreement with the results of qPCR, the posterior-most region showed strong signals for slc6a19 mRNA in the intestine of stage 34 embryos by in situ hybridization (Fig. 9K, arrow). Intense mRNA signals for slc6a19 were shown in secondary folds throughout the juvenile intestine (Fig. 9O). The slc6a19 mRNA signals were limited to the epithelial cells (Fig. 9D,H,L,P, arrowheads).

Here we have demonstrated that spiral formation of the intestine was completed before pre-hatching in the cloudy catshark. We further revealed the presence of yolk in the intestine and expression of key nutrient transporters after pre-hatching. Our data suggest a switch from nutrient absorption via the yolk sac membrane to the intestine at the pre-hatching stage, which probably facilitates growth and development of the embryo inside the egg case.

The presence of a spiral intestine in developing embryos has been shown in several elasmobranch species including the round stingray Urobatis halleri (Babel, 1966), lesser spotted dogfish (Ballard et al., 1993) and tropical whitespotted bamboo shark Chiloscyllium plagiosum (Xu et al., 2015), but detailed spiral formation and functional relevance of the spiral intestine during embryonic development were unknown. Recently, detailed morphogenesis of the spiral intestine was reported in little skate Leucoraja erinacea using histology and micro-computed tomography scan (Theodosiou and Oppong, 2019). In the little skate, spiral formation began at stage 25, and the number of spiral turns reached 6.5 by stage 30. The number of spiral turns in embryos at the final developmental stage (stage 34; more than 14 weeks after stage 30) was eight, equivalent to that in adult individuals (Theodosiou and Oppong, 2019); however, the number of spiral turns at pre-hatching was not reported so we are unable to compare the exact timing of complete spiral formation between the cloudy catshark and little skate. Currently, the present study is the only study showing that spiral formation of the intestine is completed before pre-hatching in an elasmobranch. Further studies in other species are needed to confirm whether our finding on the spiral formation is common in all cartilaginous fish.

In addition to spiral formation, microvillus formation was observed prior to pre-hatching. Short microvilli were sparsely distributed from stage 27, and microvilli covered the apical surface of the intestinal lumen at stage 31E, but intestinal epithelium remained smooth without folding until pre-hatching. In vertebrate intestines, it is generally accepted that three morphogenetic processes occur continuously to maximize surface area for nutrient absorption: (1) elongation of the gut tube, (2) microvilli formation, and (3) villi or secondary fold formation (Walton et al., 2016). In the zebrafish embryo, microvilli formation starts at 74 h post-fertilization (Wallace et al., 2005), whereas secondary folds formed around 96 h post-fertilization (Ng et al., 2005). Spiral formation in the cartilaginous fish intestine could be considered to be functionally equivalent to the elongation of gut tube in terms of an increase in luminal surface area. Therefore, it could be hypothesized that the described series of morphogenetic processes in the intestine may be common across jawed vertebrates.

After the pre-hatching event, obvious secondary folds were found at stage 32, and the length of secondary folds, as well as the length and the density of microvilli, increased as the embryo grew. However, the most prominent change that occurred before and after the pre-hatching event was the presence of yolk in the embryonic intestine. Lechenault et al. (1993) reported that yolk transfer begins within 1 week of pre-hatching in the lesser spotted dogfish embryo. In our detailed observation, yolk transfer occurred within 48 h after pre-hatching. This variation may be due to a difference in incubation temperature of eggs (14°C in the previous study, and 16°C in the present study). Yolk transfer from the EYS into the embryonic intestine is considered to be driven by cilia on the epithelium of EYS and the yolk stalk (Tewinkel, 1943); however, we identified a membrane preventing the movement of yolk into the intestine prior to pre-hatching. As the appearance of yolk occurs at pre-hatching, it can be assumed that the collapse of this membrane may be tightly regulated, possibly by signals released from the maturing intestine.

Our morphological findings are consistent with the previous histological observation showing that intestinal epithelial cells of elasmobranch embryos have the ability to absorb nutrients (Tewinkel, 1943; Okano et al., 1981; Teshima and Tomonaga, 2016). To further confirm nutrient absorption from the embryonic intestine, we examined expression of mRNAs encoding a peptide transporter (PepT1) and a neutral amino acid transporter (Slc6a19) as functional markers for nutrient absorption (Clements and Raubenheimer, 2006). The cDNA encoding PepT1 was previously cloned from the intestine of the bonnethead shark, Sphyrna tiburo (Hart et al., 2016), and we show the anterior-biased expression of pept1 mRNA at stage 34 and juvenile spiral intestines of the cloudy catshark. This bias in expression is consistent with the distribution of pept1 mRNAs in teleosts, such as European sea bass Dicentrarchus labrax (Terova et al., 2009), turbot Scophthalmus maximus (Xu et al., 2016) and tilapia Oreochromis mossambicus (Orozco et al., 2017). Zebrafish Danio rerio and grass carp Ctenopharyngodon idella show proximally restricted expression of pept1 mRNA (Verri et al., 2003; Liu et al., 2013). These results imply that pept1 mRNA in the intestine generally shows expression pattern in an anterior-biased fashion in fish. However, ubiquitous pept1 mRNA expression has been shown in Atlantic cod Gadus morhua and killifish Fundulus heteroclitus macrolepidotus (Bucking and Schulte, 2012; Rønnestad et al., 2007). In our observations, pept1 mRNA was initially expressed broadly throughout the entire intestine in early developmental stages, but the expression of pept1 was restricted to the anterior intestine as development proceeded. A similar change in mRNA expression patterns (broader in larva) was also reported in the Japanese eel Anguilla japonica (Ahn et al., 2013). The expression pattern of pept1 mRNA seems to be plastic depending on ontogeny, physiological condition or species.

Transporters contributing to amino acid absorption include a number of proteins that have distinct substrate specificities to neutral, basic and acidic amino acids (Bröer, 2008). In the present study, we selected a neutral amino acid transporter, Slc6a19, as the indicator of amino acid absorption due to the importance of neutral amino acids for maintaining body homeostasis (Sveier et al., 2001; Bröer et al., 2004; Craig and Moon, 2013; Belghit et al., 2014). Slc6a19 is a major neutral amino acid transporter expressed in the vertebrate intestine, and mutations in the slc6a19 gene cause an inherited autosomal recessive defect known as Hartnup disorder (Kleta et al., 2004). Expression of slc6a19 mRNA was found throughout the juvenile intestine in a pattern similar to that shown in European sea bass and turbot (Rimoldi et al., 2015; Xu et al., 2016). Aminopeptidase activity was shown to peak in the mid-spiral intestine of young adults or late juveniles of bonnethead sharks (Jhaveri et al., 2015), suggesting that peptides not absorbed in the anterior region may be digested by aminopeptidases and then absorbed in the form of free amino acids across the posterior region. In accordance with this notion, a remarkable expression of slc6a19 mRNA was observed in the most posterior part of the intestine at stage 34. In the stage 34 embryo, the rectal canal that connects the spiral intestine to the rectum was closed, so the embryo will retain yolk digesta until hatching. This result was consistent with that reported in the lesser spotted dogfish, where the rectum of the embryo was closed until hatching (Mellinger et al., 1986, 1987). It is possible that embryos may achieve maximum nutrient absorption from yolk digesta by up-regulating transport expression such as Slc6a19 in the posterior region of spiral intestine. In Mozambique tilapia, the neutral amino acid transporter slc6a18 mRNA was expressed in the posterior part of intestine to complete the absorption of neutral amino acids, supporting our notion (Orozco et al., 2018).

As yolk is directly transferred into the spiral intestine, it is assumed that anterior parts of the digestive tract, such as the esophagus and stomach, are not used during embryogenesis in the lecithotrophic cloudy catshark. Our histological observations revealed that the position equivalent to the esophagus was closed throughout the embryonic period. Consistent with our results, the digestive tract anterior to the spiral intestine appears to be inactive in the spiny dogfish embryo (Tewinkel, 1943). Immunoreactive signals for the gastric proton pump (H⁺–K⁺-ATPase), which represents the functional marker of stomach activity, were first detected at stage 34 in lesser spotted dogfish embryos (Gonçalves et al., 2019). In contrast, embryos of many viviparous species drink or eat nutrient materials supplied from the mother (Fujita, 1981; Hamlett et al., 2011), meaning that their esophagus and stomach must be active even during the embryonic period. Collectively, cartilaginous fish embryos during the yolk-dependent period probably use only the posterior part of digestive tract, and may close a connection between spiral intestine and the oral cavity. This strategy may have an advantage to reduce the cost of innate immunity inside the digestive tract, as seawater, rich in microorganisms and pathogens, would not come into the digestive tract from the oral cavity. Indeed, it has been reported that gut-associated lymphomyeloid tissue develops before the onset of feeding in the lesser spotted dogfish embryo, which is much later compared with the development of other immune systems (Lloyd-Evans, 1993).

Pre-hatching is a unique feature of cartilaginous fish among metazoans (Lechenault et al., 1993). Previously, our understanding of the physiological purpose of pre-hatching or opening of the eggshell to the environment was to facilitate circulation of seawater and improve respiration (Diez and Davenport, 1987), and/or to make room for the growing embryo within the egg case (Mellinger et al., 1986, 1989). The immune system is considered to be functional before the pre-hatching event, as immunoglobulin-positive cells are detected in most of the lymphomyeloid system (liver, interrenal, thymus, spleen and Leydig organ) in the lesser spotted dogfish embryo (Lloyd-Evans, 1993). In relation to respiratory function, the cloudy catshark and port Jackson shark (Heterodontus portusjacksoni) embryos begin buccal pumping, just after the pre-hatching event (Rodda and Seymour, 2008; Tomita et al., 2014).

In the present study, we clearly showed that pre-hatching is a turning point for embryonic nutritional absorption; after pre-hatching, embryos incorporate yolk into their intestine to digest and absorb nutrient from their yolk sac. A similar ‘shift’ has been described for the production of urea from the extra-embryonic yolk sac membrane to the embryonic liver, a crucial osmolyte for marine cartilaginous fish necessary for survival in seawater (Takagi et al., 2014, 2017). Lechenault et al. (1993) suggested that pre-hatching is comparable to hatching of anamniotes because pre-hatching occurs as a result of the degradation of a solid layer. Our findings on changes in physiological function after pre-hatching in the embryo support this notion. It is probable that pre-hatching in oviparous cartilaginous fish is a key window in embryonic development as they are for the first time exposed to the environment, but remain within the eggshell for further development and growth prior to emergence.

We sincerely thank all the staff of Oarai Aquarium for their kind support in providing experimental animals, and Professor S. Kuraku and staff of the Laboratory for Phyloinformatics, RIKEN Center for Biosystems Dynamics Research (BDR), for providing nucleotide sequences of cloudy catshark. We also thank Professor W. G. Anderson at the University of Manitoba for critical comments on the manuscript.