The animal husbandry and experimental protocols of the present study were reviewed and approved by the Animal Care and Utilization Committee of National Taiwan Normal University. Adult medaka (Oryzias latipes) were reared in 1200 (length) × 450 (width) × 600 (height)-mm FW tanks equipped with water filters. We refreshed a quarter of the tank water every week. The water temperature of the tanks was controlled to 28°C by a thermal controller and the photoperiod of the animal room was controlled 14 h of light and 10 h of the dark. Fish were fed at least twice a day. In the morning of ten to ten-thirty, the fertilized eggs were collected from the belly of the female medaka. For SW acclimation, the collected eggs were directly transferred to Petri dishes containing 30 ppt SW and incubated at 28°C. We added the appropriate amounts of sea salt (Instant Ocean, Aquarium System, Mentor, OH, United States) to FW to prepare SW. SW-acclimated adult medaka were obtained by transferring adult fish to 30 ppt SW tanks for 2 weeks.

Coding sequences of known and predicted medaka NHE homologs were obtained from GenBank and Ensembl databases. For the phylogenetic analysis of NHEs candidates, the multiple sequence alignment program “ClustalW2” (European Molecular Biology Laboratory, Heidelberg, Germany) was used to compare the deduced amino acid sequence of medaka NHE homologs with all known data available in the European Institute of Bioinformatics (EBI) database and were subjected to phylogenetic inferences using the Neighbor-joining (NJ) method. The MEGA 6.0 (Molecular Evolutionary Genetics Analysis Version 6.0) software was used to perform 1,000 bootstrap replicate analyses. Based on the assemblies of the Ensembl Genome Browser, the physical gene map of the verified slc9a2 (NHE2) and slc9a4 (NHE4) loci was scaled. Genes located up-and downstream of slc9a2 (NHE2) and slc9a4 (NHE4) genes in these loci were compared to genomes of fish, amphibians, reptilian and mammalian to determine the highest score.

mRNA levels of NHE2 in different tissues (brain, gills, eyes, heart, liver, intestines, spleen, kidneys, muscles, and fins) of SW-acclimated adult medaka were determined using an RT-PCR. Tissues isolated from three to seven individuals were pooled as one sample. We used Trizol reagent (Invitrogen, Carlsbad, CA, United States) to homogenize all samples according to a volume ratio of 1:10 (sample: Trizol). After homogenizing, 0.1 ml of 1-bromo-3-chloropropane (BCP) solution (Sigma-Aldrich, St. Louis, MO, United States) was added to each sample. After centrifugation at 4°C and 14,000 rpm for 30 min, the supernatant was extracted and taken into the same volume of 100% isopropanol for RNA precipitation. Subsequently, total RNA was purified and DNA contamination was removed by a MasterPure™ RNA Purification Kit (Epicentre Biotechnologies, Madison, WI, United States). After purifying total RNA, the amount and the quality of total RNA were determined by a NanoDrop 1000 spectrophotometer (Thermo Scientific, Wilmington, DE, United States). The integrity of total RNA was checked by electrophoresis through an RNA-denaturing gel.

For the reverse-transcription reaction, 3 μg of total RNA was added into a final volume of 13 μl containing 1 μl of 10 mM dNTPs, 1 μl of 50 mM oligo(dT)₂₀ at 65°C for 5 min, and then the samples immediately moved onto the ice for at least 1 min. After that, each sample was supplemented to a final volume of 20 μl containing 5 mM dithiothreitol (DTT), 40 units of an RNase inhibitor (RNAaseOUT) and 200 units of SuperScript IV (SSIV) (Invitrogen, Carlsbad, CA, United States), and then incubated at 55°C for 15 min. For PCR amplification, 0.5 μl of complementary (c)DNA (dilute 1:200 by diethylpyrocarbonate water) from various tissues, including the brain, gills, eyes, heart, liver, intestines, spleen, kidneys, muscles and fins, were used as a template in a 25-μl final reaction volume containing 1 μl of 10 mM dNTP, 1 μl of 40 units of Gen-Taq polymerase (Genemark, Taipei, Taiwan), and 1 μl of 10 mM of NHE2 primers (forward: 5′-ATCGTCTGTTGTGCCCTC-3′; reverse: 5′-CAGTTCCACTCGTGCTCT-3′).

An antisense morpholino-modified oligonucleotides (MOs; Open Biosystems, Huntsville, AL) with the sequence 5′-TTCTCGTGGACCGAGTACATGATGC-3′ was used to target −6 to + 19 of the coding region of medaka NHE2; a sense photo-activated (photo)-MOs (Open Biosystems, Huntsville, AL), 5′-GCATCATGTACT-photo-GGTCCAC-3′, was used to target normal MOs with the middle of the photo-sensitive subunit. Standard control oligonucleotides (5′-ATCCATCTTGTGTGTTAGAAAACTG-3′; Gene Tools, Philomath, OR, United States) were used as the control MO, based on its no target and no significant biological activity.

The MOs were resuspended in 1X Danieau’s solution, stored at −20°C as a stock solution. Before use, we diluted the stock to desired concentrations (0.5, 1.0, 2.0, or 4.0 ng/embryo). In the present study, both MOs (0.5 ng/embryo) and photo-MOs (0.6 ng/embryo) were incubated at 50°C for 10 mins. Subsequently, both MOs and photo-MOs were co-injected into embryos at one to the two-cell stage by a gas-driven microinjection apparatus (ASI, Eugene, OR). After injection with both NHE2 MOs and photo-MOs, 5-day post-fertilization (dpf) embryos were stimulated by UV light (365 nm) for 1 min from UV Lightbox (Gene Tools, Philomath, OR, United States), and then 8-dpf embryos were sampled to measure proteins expression. The 5-dpf embryos not stimulated with UV light (365 nm) served as the control photo-MOs.

We collected 8-dpf embryos and fixed them in 4% paraformaldehyde (PFA) at 4°C overnight. The protocol for NHE2 IHC followed by that of a previous report (Inoue and Wittbrodt, 2011). For the immunoreaction, a polyclonal antibody against the domain (VAYPGRRSRFGNRSS; Genomics, Taipei, Taiwan) of medaka NHE2 (diluted 1: 300) and goat anti-rabbit immunoglobulin G (IgG) conjugated with Alexa Fluor 488 (Molecular Probes, Carlsbad, CA) (diluted 1: 500) were, respectively, used as the primary and secondary antibody, respectively. For Na⁺/K⁺-ATPase (NKA) staining, an α5-monoclonal antibody against the α-subunit of the avian NKA (Developmental Studies Hybridoma Bank, University of Iowa, Ames, IA) was used to detect NKA. The protocol was conducted as described in previous reports (Wu et al., 2010; Liu et al., 2013). Finally, images were captured by using a fluorescence microscope (Axioplan 2 Imaging, Carl Zeiss Oberkochen, Germany).

Ten 8-dpf embryos were collected as one sample and washed twice with phosphate-buffered saline (PBS). Afterward, homogenization buffer (100 mM imidazole, 5 mM EDTA, 200 mM sucrose, and 0.1% sodium deoxycholate) (with the pH adjusted to 7.6) with 10 units of protease inhibitors (Roche, Indianapolis, IN, United States) was used to homogenize all samples. After centrifuging at 4°C and 10,000 rpm for 10 min, the supernatant was extracted for analysis by using BCA protein assay reagents (Pierce Chemical Company, Rockford, IL, United States). Each sample (equivalent to a volume of 20 μg protein) was incubated at 95°C for 10 min for denaturing and then subjected to 10% sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE). After electrophoresis, the gel was transferred to polyvinylidene difluoride membranes (Millipore, Billerica, CA, United States). For the blocking reaction, the membranes were transferred to 5% non-fat milk at room temperature for 2 h. Afterward, blots were incubated with a rabbit anti-medaka NHE2 polyclonal antibody (diluted 1:1000) and a rabbit anti-human β-actin (Abcam, Tokyo, Japan, diluted 1:1000) at 4°C overnight. Signals of blots were enhanced with a chemiluminescence system (Millipore, Billerica, CA, United States). The membranes were incubated with goat anti-rabbit IgG (H + L) HRP (dilute 1:10000; Invitrogen, Carlsbad, CA, United States). The ImageQuant 4000 system (GE Healthcare, Buckinghamshire, United Kingdom) was used to capture the clear images for analysis. Finally, we quantitated the intensities of immunoreactive bands using ImageJ software¹.

The scanning ion-selective electrode technique (SIET; also called the non-invasive micro-test technique, NMT) was used to measure H⁺ and Cl^– secretion by skin ionocytes of medaka larvae. H⁺-selective and Cl^–-selective microelectrodes were prepared as described in previous studies (Shen et al., 2011; Liu et al., 2016). The signal of the microelectrode was connected to the main amplifier through an Ag/AgCl wire electrode holder and preamplifier (Younger United States, Amherst, MA, United States). The movement and positioning of the microelectrodes were performed with a three-dimensional (3D) positioner (Younger United States) driven by a step motor. Imfluxes V2.0 software (youngerusa.com; xuyue.net) was used to control the system and calculate ion fluxes.

Data are presented as the mean ± standard error of the mean (SEM). Student’s unpaired t-test (two-tailed) was used for comparisons of the means of two groups. The significance level was set to p < 0.05.

The sequence of medaka NHE2 (ENSORLG00000012399) was predicted using the Ensembl database and was further cloned and sequenced. The NJ method was used to generate the phylogenetic tree of protein sequences of NHEs from human (Homo sapiens), mouse (Mus musculus), rat (Rattus norvegicus), medaka (Oryzias. latipes), zebrafish (Danio rerio), tilapia (Oreochromis niloticus), stickleback (Gasterosteus aculeatus), and pufferfish (Takifugu rubripes and Tetraodon nigroviridis). The phylogenetic analysis demonstrated that the eight members of medaka NHEs (NHE1, NHE2, NHE3, NHE5, NHE6, NHE7, NHE8, and NHE9) were classified into different groups and clustered with orthologues from mammals (Figure 1). According to the phylogenetic analysis, the fish NHE2 clade was most closely related to the mammalian NHE2 and NHE4 clades. Further, gene arrangements in the genomic regions encompassing NHE2 and NHE4 were compared. Results showed that NHE2, TMEM182, and MFSD9 were evolutionarily located on the same chromosome from fish to mammals except for INPP4AA (Figure 2). The NHE4 gene occurred beside NHE2 on the same chromosome of amphibians, reptiles, birds, and mammals but not in fishes, suggesting that NHE2 duplication had occurred in the chromosome of amphibians and that had formed NHE4.

The putative transmembrane (TM) protein prediction of medaka NHE2 was predicted using TMHMM (server v. 2.0)² (Figure 3). Results showed that 12 TMs (black line) were found in medaka NHE2. In addition, compared to amino acid sequences of mammalian NHE2, we found that fish NHE2 lacked the two proline-rich sequences (Pro-1 and Pro-2) (dashed line) (Figure 3). It was demonstrated that Pro-1 and Pro-2 are necessary for appropriate subcellular targeting of rat NHE2 (Chow et al., 1999).

To examine the distribution of slc9a2 in medaka tissues, mRNA expression of slc9a2 in various tissues including the brain, gills, eyes, heart, liver, intestines, spleen, kidneys, muscles, and fins, were examined by an RT-PCR. The result showed that slc9a2 was abundantly expressed in the gills, intestines, and muscles (Figure 4). In contrast, slc9a2 expression was quite weak in the liver, spleen, and kidneys (Figure 4). In this study, rpl7 mRNA was used as an internal control.

IHC was applied with a medaka NHE2-specific antibody to localize NHE2 protein in ionocytes of SW acclimated embryos (Figure 5). NKA immunostaining was used as a marker of ionocytes. Confocal images showed that NHE2 signals (green; Figures 5A,C) were colocalized with NKA signals in the basolateral membrane of ionocytes (red; Figures 5B,C). The basolateral localization of NHE2 was also confirmed by the confocal z-scanning images (Figure 5D).

To knockdown NHE2 protein expression, traditional MOs were injected into embryos at the 2-cell stage. Injection with a low dose of NHE2-MOs (0.5 ng/embryo) dramatically increased the mortality rate to 83% at 5 dpf which reached 90% at 8 dpf (Figure 6A), while control-MOs showed only 12% mortality at 8 dpf. We also observed severe deformation in embryos, including edema and tail bending (Figure 6B), suggesting that NHE2 might be required for early embryo development. To reduce the impact on early development, a photo-MO technique was applied to knock-down NHE2 expression at a later stage (5 dpf). Results showed that the mortality of NHE2-photo-MOs was about 35% at 5 dpf and gradually increased to 55% at 8 dpf, while control-photo-MOs only caused 35% mortality at 8 dpf. Morphological defects were barely observed in NHE2-photo-MOs, with only slight edema or tail bending exhibited (Figure 6A).

An immunocytochemical analysis was used to compare NHE2 protein expression in 8-dpf embryos with or without NHE2-knockdown (KD). In the control-photo-MO group, green spots (NHE2 immunosignals) were observed on the yolk sac (Figure 7A), while those spots were not evident in the NHE2-photo-MO group. In addition, we also used a Western blot analysis to confirm NHE2-KD. Results showed a predicted band at 91 kDa in the control-photo-MO group, while the band was very faint in the NHE2-photo-MO group (Figure 7B).

To analyze the effects of NHE2 knockdown on H⁺ and Cl^– secretion by ionocytes, the SIET was used to measure H⁺ and Cl^– effluxes at individual ionocytes on the yolk sac skin of medaka. Results showed that the H⁺ efflux and Cl^– efflux significantly decreased by 60 and 34%, respectively, following NHE2 knockdown (Figures 8A,B).