- Email: [email protected]
K+-ATPase in ionocytes during the development of euryhaline medaka (Oryzias latipes) embryos
Gene Expression Patterns 10 (2010) 185–192
Contents lists available at ScienceDirect
Gene Expression Patterns journal homepage: www.elsevier.com/locate/gep
Expression of Ol-foxi3 and Na+/K+-ATPase in ionocytes during the development of euryhaline medaka (Oryzias latipes) embryos Violette Thermes a,*, Chia-Cheng Lin b, Pung-Pung Hwang b a b
INRA, SCRIBE, Campus de Beaulieu, 35042 Rennes, France Institute of Cellular and Organismic Biology, Academia Sinica, Nangang, Taipei, Taiwan, ROC
a r t i c l e
i n f o
Article history: Received 1 February 2010 Received in revised form 30 March 2010 Accepted 2 April 2010 Available online 11 April 2010 Keywords: Fish Teleost Osmoregulation Gills Yolk sac Epidermis Skin Mitochondria-rich cells Chloride cells Ionocytes Progenitors atp1a1a.1 p63 Pharyngeal endoderm Epibranchial placodes Vitellin Zone Lateral Zone
a b s t r a c t Osmoregulation is a vital function that is essential to all vertebrates. Ionocytes are epithelial cells responsible for this function and have been extensively studied in adult teleost fish gills. The euryhaline medaka (Oryzias latipes) has recently emerged as an investigative model because of its ability to acclimatize easily to water presenting various salinities. However, no studies to date have focused on the development of ionocytes in medaka embryos. We first analyzed the distribution of ionocytes in the skin and gills during development, using a specific marker of differentiated ionocytes (the Na+/K+-ATPase pump, or NKA). Strikingly, we were able to identify two ionocyte domains on the yolk surface ectoderm, that we named the Vitellin Zone (VZ) and the Lateral Zone (LZ). In zebrafish, ionocyte differentiation has been shown to be controlled by two forkhead-box genes, foxi3a and foxi3b. We cloned the medaka foxi3 ortholog which appeared to be highly similar to foxi3b. Whole-mount in situ hybridizations performed on medaka embryos revealed that Ol-foxi3 is expressed in differentiated ionocytes of the pharyngeal endoderm, the branchial arches and the yolk epidermis, as well as in epibranchial placode territories. We further focused on the expression patterns of the yolk epidermis and compared the expression of Ol-foxi3 with that of the non-neural progenitor marker p63. We evidenced that Ol-foxi3 is expressed in progenitor cells which are first of all located uniformly in the VZ and then transitorily clustered in the LZ. Taken together, these data contribute to a clearer understanding of osmoregulatory tissue ontogenesis in euryhaline fish. Ó 2010 Elsevier B.V. All rights reserved.
1. Results and discussion Ion transport is a vital function that plays an essential role in the osmoregulation of body fluids in all vertebrates. The epithelia responsible for this function are located in several organs, including the kidney collecting duct, the amphibian urinary bladder, the gut, the inner ear, the fish gills and the skin (Takagi, 1997, for a review see Brown and Breton (1996), Evans et al. (2005) and Hwang and Lee (2007)). All transporting epithelia contain a subset of specialized cells called ionocytes (formally called mitochondria-rich cells or chloride cells) (Keys and Willmer, 1932), that transport various ions across the epithelia down gradients created by the strongly expressed plasma membrane enzymatic complex Na+/K+-ATPase (NKA) pump (Blanco and Mercer, 1998; Hwang and Lee, 2007; Therien and Blostein, 2000). Medaka (Oryzias lati-
* Corresponding author. Tel.: (+33) 2 23 48 57 20; fax: (+33) 2 23 48 50 20. E-mail address: [email protected] (V. Thermes). 1567-133X/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.gep.2010.04.001
pes) is a small fish model (for a review see Wittbrodt et al. (2002)) that displays particular adaptability to various water salinities (Inoue and Takei, 2002, 2003). The medaka can survive and develop normally when transferred from fresh water to sea water. For this reason, this euryhaline fish has recently emerged as an interesting model for osmoregulation studies (Kang et al., 2008; Wu et al., 2009). However, no studies to date have focused on the development of skin and gill epithelial cells during embryonic development in the medaka. The yolk skin epidermis contains several types of cells, including ionocytes, and develops after gastrulation from the non-neural ectoderm. In the pharyngeal embryonic region, the ectoderm contributes to formation of the branchial apparatus, which develops as a result of coordinated interplay between the pharyngeal ectoderm, the pharyngeal endoderm, the mesoderm and the epibranchial placodes. More specifically, the pharyngeal ectoderm gives rise to the gill epithelium, whereas the pharyngeal endoderm forms the epithelium that lines the pharynx and contributes to the formation of associated
186
V. Thermes et al. / Gene Expression Patterns 10 (2010) 185–192
organs. Both the gill epithelium and pharynx epithelia contain ionocytes. During vertebrate evolution, development of the pharyngeal apparatus has undergone a number of modifications. In terrestrial vertebrates, the parathyroid gland evolved as a result of the internalization of fish gills during the transition from an aquatic to a terrestrial environment (Graham et al., 2005). Nevertheless, it has been shown that the foxi class genes constitute a conserved feature of pharyngeal development and play a significant role in the establishment of branchial development (Graham, 2001; Solomon et al., 2003b). foxi class genes belong to a large and highly conserved gene family that encodes winged helix/forkhead-box transcription factors. They have been analyzed in detail in mouse and zebrafish. In mouse, three foxi genes have been identified and shown to display distinct and dynamic expression patterns in early craniofacial epithelia, including the otic vesicles (foxi1), cranial ectoderm (foxi2), early placodes of the surface ectoderm and the pharyngeal endoderm (foxi3) (Ohyama and Groves, 2004). Gene knock-out in the mouse revealed that Foxi1(/) mice display no abnormal development of the inner ear, but an absence of ionocyte differentiation (Hulander et al., 2003). In the zebrafish, two genes have been identified as being expressed in the epibranchial placodes and otic vesicles (foxi1), or in the chorda-mesoderm and pharyngeal arches at a later stage (foxi2). Zebrafish also present two members of the foxi3 gene, foxi3a and foxi3b, that are exclusively expressed in yolk epidermis ionocytes. Recent studies have found that foxi3 genes control ionocyte differentiation (Chang et al., 2009; Esaki et al., 2009; Hsiao et al., 2007; Janicke et al., 2007, for review see also Hwang and Perry (2010)). In order to gain a clear understanding of skin and gill ionocyte development in the medaka, we first of all analyzed the distribution of ionocytes during the development of skin and gill epithelia, using NKA immunostaining. We then cloned the medaka ortholog of zebrafish foxi3 key genes and examined its spatio-temporal expression during development, with a specific reference to ionocytes. 1.1. Na+/K+-ATPase-rich cell distribution during medaka embryonic development In order to determine the stages of emergence of ionocytes in embryonic skin and gills, the time course expression of the NKA pump was analyzed by immunostaining on whole mount embryos. NKA immunoreactivity was first observed at st.20, in a few scattered cells on the yolk epidermis (Fig. 1A). As development proceeded, an increasing number of NKA-immunoreactive cells (NKA-ir cells) could be detected on the yolk epidermis (Fig. 1A– F). Interestingly, at st.30 (35-somite stage), a higher density of NKA-ir cells was detected in a salt-and-pepper pattern in the axial yolk epidermis on both sides of the trunk (Fig. 1C). This area was named the Lateral Zone (LZ), while the surface covering the whole yolk was named the Vitellin Zone (VZ). Subsequently, LZ and VZ could easily be distinguished from the difference in NKA-ir cell density (Fig. 1C–F). Furthermore, at st.25 (18-somite stage) and st.30 (35-somite stage), NKA-ir cells could also be detected in the pharyngeal endoderm and in the first gill slit (also called the lateral gill chamber, Fig. 1G and H). Subsequently, from st.35 onwards, NKA-ir cells were detected on the epithelium of branchial arches as they develop (Fig. 1I and J). At the larva stage, numerous NKAir cells were found to be located on the developing primary gill filaments (Fig. 1K). 1.2. Cloning and expression analysis of Ol-foxi3 during medaka embryonic development We then focused our study on the foxi3 transcription factor and its expression in the skin and gills during development. For this
purpose, we cloned the medaka foxi3 ortholog detected by analyzing the medaka genome. Only one sequence related to zebrafish foxi3 genes could be identified and further sequence alignments enabled the construction of a phylogenetic tree (Fig. 2A). Fulllength sequences of the zebrafish FOXI3a and FOXI3b displayed 56.5% amino acid identity. The Fugu FOXI3 (Solomon et al., 2003b) is 55.3% and 71.5% identical to the zebrafish FOXI3a and FOXI3b, respectively. Although the Ol-FOXI3 protein resembles the zebrafish FOXI3a (56.5% identity), it is clearly more closely related to the Danio FOXI3b and Fugu FOXI3 (70.8% and 81.3% identity at the amino acid level, respectively). A multiple alignment of FOXI3 proteins (Fig. 2B) showed the forkhead conserved domains and confirmed that it is the medaka Ol-foxi3 ortholog. The expression pattern of Ol-foxi3 was analyzed by wholemount in situ hybridization (WMISH) in medaka embryos and larvae. At st.18 (late neurula stage), Ol-foxi3 mRNA expression was first detected in the anterior embryonic ectoderm flanking the midbrain–hindbrain boundary (MHB) region, which corresponds to the epibranchial (EB) placodes (Fig. 3A and I). This expression domain is maintained at st.20 (4-somite stage, Fig. 3B) but was no longer detected at st.25 (18-somite stage, Fig. 3C). From that stage, embryos displayed Ol-foxi3 expression in the pharyngeal endoderm (Fig. 3C–F). Indeed, lateral views of embryos at different stages display a high level of expression in the pharyngeal endoderm that gradually diminishes until st.35, where it is eventually only expressed in scarce cells (Fig. 3J–N). Conversely, from st.30, a punctuated pattern of Ol-foxi3 was detected in the developing pharyngeal ectoderm (i.e. branchial arches epithelium) that progressively extends from the rostral to the caudal extremities as the pharyngeal arches develop (Fig. 3K–N). Finally, from st.18, Olfoxi3 was also expressed in single cells scattered in the VZ domain of the yolk epidermis (Fig. 3A and H). This spotted and specific staining is congruent with an expression in ionocytes (see Fig. 1). Consistently with the ionocyte distribution described in Fig. 1, additional foxi3-positive cells were also observed at st.25 in the inner most part of the LZ (Fig. 3C). This staining gradually spreads as the embryo develops, whereas the Ol-foxi3 expression level in the VZ slightly diminishes. At the larva stage, Ol-foxi3 expression was observed in the ionocytes of both the gill epithelium and the yolk epidermis (Fig. 3O and H). In order to confirm that Ol-foxi3 is expressed in NKA-ir cells, we performed double fluorescent WMISH using Ol-foxi3 and atp1a1a.1 probes (Fig. 4). Indeed, the NKA complex contains a catalytic a subunit and a stabilizing b subunit (Jorgensen, 1974). In order to detect NKA at the gene level, the atp1a1a.1 gene encoding the NKA a1 subunit was cloned. Further expression analysis on adult gills enabled to demonstrate that atp1a1a.1 does indeed reflect NKA distribution (data not shown). At st.27–28 and at larva stage, double fluorescent WMISH confirmed a co-expression of Ol-foxi3 and atp1a1a.1 in the pharyngeal endoderm and the first gill slit (Fig. 4A–C), the skin ectoderm (Fig. 4D and F) and the gills (Fig. 4D and E). At st.20, no expression of atp1a1a.1 was detected in Ol-foxi3-expressing cells in the EB domain (data not shown). 1.3. The skin epidermis displays two territories of ionocyte progenitors One of the most striking features of Ol-foxi3 and NKA expression patterns is the existence of two ionocyte domains on the yolk epidermis, the VZ and LZ. The expression of NKA in the VZ, which was apparent at st.20, is delayed when compared with the onset of Olfoxi3 expression (compare Fig. 1A and Fig. 3A). This delay was also observed in the LZ, where NKA-ir cells appeared at st.30 and Olfoxi3 expression was seen as early as st.25 (compare Fig. 1C and Fig. 3C). Starting from this observation, we analyzed the co-expression of Ol-foxi3 and atp1a1a.1 in the LZ in more detail, between st.25 and st.30 (Fig. 5). As expected, we observed that Ol-foxi3 is
V. Thermes et al. / Gene Expression Patterns 10 (2010) 185–192
187
Fig. 1. Distribution of NKA-immunoreactive cells during medaka embryonic development. NKA-ir cells are detected in embryos (A–E and G–J) and in larvae (F and K) by whole-mount immunostaining. (A and F) From st.20 onwards, NKA-ir cells are detected on the yolk epidermis. The LZ and the VZ domains display different densities of NKA-ir cells. (G–J) Ventral views of embryos at different developmental stages. The development of branchial arches can be followed by nuclear DAPI staining. (G and H) At st.25 and st.30, NKA-ir cells are detected in the pharyngeal endoderm and the first gill slit. (I and J) At st.35 and st.39, NKA-ir cells are detected in the branchial ectoderm. (K) At the larval stage, NKA-ir cells are located on primary gill filaments. Note that NKA-ir cells are mainly located on the internal face (or efferent face, dotted line). (A) Embryo in ventral view, animal pole to the top. (G–K) Embryo and larva in ventral view, anterior to the top. (B–D) Embryo in dorsal view, anterior to the top. (E) Embryo in lateral view, anterior to the top. (F) Larva in lateral view, anterior to the left. LZ, Lateral Zone; VZ, Vitellin Zone; ba, branchial arches; ey, eye; gs, gill slit; and pe, pharyngeal endoderm. Scale bars: 100 lm in A–F and 50 lm in G–K.
first expressed at st.25 in a few cells lined up in the LZ on both sides of the body axis (Fig. 5A, arrows). Atp1a1a.1 mRNA was not detected in these cells, although atp1a1a.1 was highly expressed in other Ol-foxi3-positive cells of the yolk sac epidermis (i.e. in the VZ domain). At st.27–28, the number of Ol-foxi3-positive/atp1a1a.1-negative cells in the LZ increases (Fig. 5B). Then, at st.30, these latter cells gradually start to express atp1a1a.1, which ultimately leads to only a few Ol-foxi3-positive/atp1a1a.1-negative cells left in the LZ (Fig. 5C). Interestingly, the same gene activation sequence was observed in the VZ, between st.18 and st.25 (data not shown). Previous zebrafish studies had reported that foxi3a is expressed in both undifferentiated ionocyte progenitors (i.e. NKA-negative
cells) and differentiated ionocytes (i.e. NKA-positive cells) (Hsiao et al., 2007; Janicke et al., 2007). To investigate the possibility that transitory Ol-foxi3-positive/atp1a1a.1-negative cells of the LZ also correspond to ionocyte progenitors, we tested whether p63, an epidermal progenitor marker (Bakkers et al., 2002), is expressed in these cells. We conducted double fluorescent WMISH for Ol-foxi3 and atp1a1a.1 in combination with anti-p63 immunostaining, at st.27–28 (Fig. 5D). A high-magnification view showed that each of the Ol-foxi3-positive/atp1a1a.1-negative cells located in the LZ was p63-positive. Taken together, these data suggest that the LZ is a secondary site of ionocyte differentiation that appears at st.25, after a first wave of differentiation initiated in the VZ at st.18. We infer that the higher density of ionocytes in the LZ may
188
V. Thermes et al. / Gene Expression Patterns 10 (2010) 185–192
Fig. 2. Sequence analyses of Oryzias latipes FOXI3. (A) A phylogenetic tree of vertebrate foxi3 related genes based on a maximum likelihood distance analysis of the amino acid alignment of nine homologs. Numbers above branches indicate the bootstrap percentage. (B) Multiple alignment of full-length amino acid sequences of members of the Foxi3 family. The forkhead conserved domains are underlined and conserved residues are highlighted in black.
be due to the reduced surface of this territory compared with the large VZ surface that covers the yolk sac.
1.4. Concluding remarks Osmoregulation is a vital function that has been studied extensively in the skin and gills of teleost fish using morphological and physiological techniques. Despite the major significance of these mechanisms in the fish embryo, little is known about the occurrence of ionocytes at this stage of the life cycle. The present study thus focused on NKA expression in the skin and branchial region during medaka development. We observed that NKA-ir cells first appear on the yolk skin at st.20 (4-somite stage), on the pharyngeal endoderm and gill slit at st.25 (18-somite stage), and on the gill epithelium of late embryos (at st.35). Most studies that have investigated the embryonic ionocytes distribution were focused on the
yolk skin. Among them, only one study has investigated the enteric ionocyte distribution in the gut during embryonic development (Sucre et al., 2010). The ionocyte distribution was described in the first gill slits and gut of the sea bass embryo (Dicentrarchus labrax). The findings of our study revealed that the embryonic gut in medaka contains a high density of NKA-ir cells that also express the Ol-foxi3 gene. This result reinforces the idea that these are ion-transporting cells, although the biological significance of the ion transport triggered by these cells in the embryo still remains unclear. Moreover, the expression of Ol-foxi3 strongly suggests a function for this gene in the specification and differentiation of ionocytes in the gut, which is similar to that of zebrafish foxi3a and foxi3b in the yolk epidermis (Hsiao et al., 2007). This study also reports in detail on the expression pattern of Olfoxi3 during embryogenesis. Ol-foxi3 was found to be expressed in the epibranchial region at st.18 (late neurula stage), in ionocytes of the pharyngeal endoderm at st.25 (18-somite stage) and in ionocytes of the skin and gills at st.18 (late neurula stage) and st.30 (35-somite stage), respectively. Thus, Ol-foxi3 expression in ionocytes is not restricted to the yolk sac like the zebrafish foxi3a and foxi3b genes, which strongly suggests that Ol-foxi3 is involved in the development of other transporting epithelia such as the pharynx and gills (Hsiao et al., 2007; Janicke et al., 2007; Solomon et al., 2003b). Interestingly, Ol-foxi3 expression in the epibranchial region and pharyngeal endoderm is reminiscent of that of the zebrafish foxi1, expressed in epibranchial and otic placodes (Nechiporuk et al., 2007; Solomon et al., 2003a), and of the mouse foxi2 and foxi3 expressed in the surface ectoderm and pharyngeal endoderm of branchial pouches (Ohyama and Groves, 2004). However, phylogenetic analyses indicated that mouse foxi2 and foxi3 genes are not the orthologs of zebrafish foxi3a/3b, but that mouse foxi1 is the ortholog of these latter and thus also of Ol-foxi3 (Solomon et al., 2003b). This discrepancy between the expression patterns of medaka foxi3, zebrafish foxi3a/foxi3b and mouse foxi1, which all belong to the same foxi subgroup, may indicate a shift of function between paralogs in these different species. Finally, the expression of Ol-foxi3 in the pharyngeal region at st.25 suggests that the development of future gills is specified at this stage in medaka embryos, and that Ol-foxi3 is an early marker of the branchial apparatus. Finally, on the yolk epidermis, the Ol-foxi3 expression pattern is different from that of the zebrafish foxi3a and foxi3b genes (Hsiao et al., 2007; Janicke et al., 2007; Solomon et al., 2003b). In the zebrafish, specified epidermal ionocytes appear in the ventral ectoderm, from 90% epiboly to the 14-somite stage (Hsiao et al., 2007) or from the 3- to 17-somite stages as described in (Janicke et al., 2007). Three main ionocyte subtypes are produced, the NaR (Na+/K+-ATPase-Rich), the HR (H+-ATPase-Rich) and the NCC (N+/Cl Cotransporter) cells (Hwang, 2009). After these stages, no more ionocytes are generated on the yolk skin. It is accepted that foxi3a and foxi3b are responsible for the segregation and differentiation of these subpopulations. In the present case, we evidenced that medaka displays two waves of ionocyte differentiation in the embryonic yolk sac. ‘Ionopoiesis’ is first initiated in the VZ (at st.18, late neurula stage) and then reactivated in the LZ of the yolk sac epidermis (at st.25, 18-somite stage). Although it is not yet known whether different subtypes of ionocytes exist in medaka, it is possible to hypothesize that two undifferentiated ionocyte subpopulations may be generated successively within these two sites. On the other hand, we assumed that only one foxi3 gene, highly similar to zebrafish foxi3b, is present in the medaka genome, which apparently does not favor ionocyte diversity. However, one can envisage a scenario under which possible ionocyte subpopulations might be under the control of key factors other than Ol-foxi3 that are perhaps specific to each domain. Further molecular and cellular analyses will be necessary to understand the physiological outcome of this cell organization in the medaka yolk sac epidermis.
V. Thermes et al. / Gene Expression Patterns 10 (2010) 185–192
189
Fig. 3. Ol-foxi3 expression during medaka embryonic development. Whole-mount in situ hybridization (WMISH) on medaka embryos and larvae at the stages indicated. (A and B) At st.18 and st.20, Ol-foxi3 mRNA are detected in a salt-and-pepper pattern on the yolk epidermis (arrowheads, inset). (C–H) From st.25 onwards, Ol-foxi3 is detected in scattered cells in the LZ and VZ domains. (A and I, inset) Ol-foxi3 mRNA are detected in the rostral ectoderm in the epibranchial region. (J–O) From st.27 to larva stage, Ol-foxi3 is expressed in the pharyngeal endoderm and branchial arches. At st.35, scare cells in the pharyngeal endoderm and intestine express Ol-foxi3. (A–G) dorsal views, (O–H) lateral views, (N) ventral view. Anterior to the left. EB, epibranchial region; LZ, Lateral Zone; VZ, Vitellin Zone; f, fin; g, gills; int, intestine; pa, pharyngeal arches; and pe, pharyngeal endoderm. Scale bars: 50 lm.
To conclude, our data show that Ol-foxi3 is expressed in both the yolk epidermis and the early craniofacial epithelia. Moreover, we show that ionocyte development (or ‘‘ionopoiesis”) in medaka embryos occurs at two sites of the yolk sac, thus differing from zebrafish embryos. The present study shows that the medaka constitutes a new embryological model to study skin/gill development as well as ionocyte ontogeny, and should pave the way for new physiological studies of osmoregulation in euryhaline embryos.
2. Experimental procedures 2.1. Medaka breeding strain Medaka embryos and larvae of the CAB strain were raised at 26 °C under a reproduction regime (14 h light/10 h dark). Fertilized eggs were collected after spawning (at the onset of light) and incubated in Yamamoto’s embryo rearing medium (Yamamoto, 1975).
190
V. Thermes et al. / Gene Expression Patterns 10 (2010) 185–192
Fig. 4. Co-localization of Ol-foxi3 and atp1a1a.1 in embryo and larva. Double fluorescent WMISH with Ol-foxi3 (in green) and atp1a1a.1 (in red) probes. (A–C) St.27–28 embryos. Ol-foxi3 and atp1a1a.1 are co-expressed in the pharyngeal endoderm (A and B) and the first gill slit (C). Ol-foxi3-positive/atp1a1a.1-negative cells are detected in the LZ (A, arrow). (D–F) Ol-foxi3 and atp1a1a.1 are co-expressed in ionocytes of the gill ectoderm (E) and yolk skin (F). (B) Longitudinal optic section as indicated in (A). Embryo and larva are orientated with the anterior to the top. (C) Transversal optic section as indicated in (B). LZ, Lateral Zone; VZ, Vitellin Zone; gs; gill slit; and pe, pharyngeal endoderm. Scale bars: 100 lm in A, B, and D and 50 lm in C, E, and F.
Fig. 5. Medaka embryos display a secondary site of ionocyte progenitors in the epidermis. Double fluorescent WMISH with Ol-foxi3 (in green) and atp1a1a.1 (in red) were performed between st.25 and st.30 (A–C). Triple staining with p63 antibody (in blue) was performed at st.27–28 (D). (A) St.25 embryos show Ol-foxi3-positive/atp1a1a.1negative cells in the LZ (arrows) and Ol-foxi3-positive/atp1a1a.1-positive cells in the VZ. (B) At st.27–28, embryos display a higher number of Ol-foxi3-positive/atp1a1a.1negative cells in the LZ. (C) At st.30, only scarce Ol-foxi3-positive/atp1a1a.1-negative cells are observed, while the number of Ol-foxi3-positive/atp1a1a.1-positive cells increases in the LZ. (D) At st. 27–28, Ol-foxi3-positive/atp1a1a.1-negative cells of the LZ co-express the p63 progenitor marker (asterisks and inset) indicating that these cells are ionocyte progenitors. Ol-foxi3-positive/atp1a1a.1-positive staining is always excluded from p63-positive cells. Embryo and larva are in dorsal view, anterior to the top. LZ, Lateral Zone and VZ, Vitellin Zone. Scale bars: 130 lm in A–D and 10 lm in D, inset.
During all experiments, the embryos were placed at 26 °C and staged according to the developmental table described by Iwamatsu (1994, 2004) and Furutani-Seiki and Wittbrodt (2004). 2.2. Sequence analysis and cDNA cloning In order to identify medaka foxi3 and atp1a1a.1 sequences we used the zebrafish FOXI amino acid sequence for a BLAST search
on the Medaka Genome project database (http://dolphin.lab.nig.ac.jp/medaka/). The partial sequences identified were used to design specific primers for 50 and 30 RACE PCR. To clone the cDNA sequence, total RNA was extracted from adult medaka gills using Trizol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions, and reverse transcription was further performed following the instructions in the SMART RACE cDNA Amplification Kit (Clontech, CA, USA). To synthesize Ol-foxi3
V. Thermes et al. / Gene Expression Patterns 10 (2010) 185–192
and atp1a1a riboprobes, cDNA fragments of 754 pb and 722 pb were isolated by PCR on the synthesized SMART cDNA, using the following primers: Ol-foxi3 forward 50 -TGTACCTGGAT 0 CGTCCTCGTCTCGG and Ol-foxi3 reverse 5 - GGTTGGTAGTTTAGA GACGCCTG; atp1a1a forward 50 -AGGAGGAGCCA GCAAACGA TAAT-30 and atp1a1a reverse 50 -TTCCAGGTTCTTCACCAAGCAAT 30 . PCR fragments were subcloned into pGEM-T Easy (Promega, WI, USA) and then sequenced. We carefully compared the Ol-FOXI3 amino acid sequence with the Danio rerio FOXI3a (NP_944599) and FOXI3b (NP_944600), Xenopus tropicalis FOXI3 (Q6P8A3), Takifugu rubripes FOXI3 (CAG10122), mouse FOXI1 (NP_001094934), rat FOXI1 (NP_001099246), human FOXI1 (NP_001129121) and Ciona intestinalis FOXI (BAE06443) sequences available on the NCBI server and the Ensemble genome browser. CLUSTALX version 2.0.11 (Larkin et al., 2007) was used for multiple alignments and the tree was constructed by Neighbor-Joining using MEGA version 4.1 (Larkin et al., 2007).
2.3. Whole-mount immunostaining and in situ hybridization Embryos were fixed with 4% paraformaldehyde (PFA) in phosphate buffer 0.12 M (PBS, pH7.4), overnight at room temperature. Larvae were anesthetized on ice and fixed overnight at 4 °C. Embryos and larvae were rinsed in PBS. Embryos were manually dechlorinated with fine forceps. The specimens were then dehydrated in 100% methanol and stored at 20 °C for subsequent analyses. For whole-mount immunostaining, the specimens were rehydrated in PBST (0.1% Tween 20 in PBS), incubated for 1 h with 3% bovine serum albumin (BSA) and then incubated overnight at 4 °C with the primary antibody. Late embryos were subjected to mild proteinase K treatment (10 lg/ml) for 30 min for st.35, and for 10 min for st.39. The Na+/K+-ATPase pump was detected with the mouse monoclonal anti-a5 (Developmental Studies Hybridoma Bank, University of Iowa, 1:300). The DNp63 protein was detected with the mouse monoclonal anti-p63 (4A4, Santa Cruz, 1:200). After rinsing for 3 h in PBST, the samples were incubated for 1 h with the secondary antibody. A goat anti-mouse IgG-Alexa 568 (1:200 in PBS) and a goat anti-mouse IgG-Alexa 594 (1:200 in PBS) were used to detect anti-a5 and anti-p63, respectively. Embryos were finally stained with DAPI (Invitrogene) for global morphology. Fluorescence was detected under a confocal microscope (SP5 and SP3, Leica). For single-color whole mount in situ hybridization, Ol-foxi3 antisense digoxygenine-UTP probes were generated by in vitro transcription. Embryos were subjected to a proteinase K treatment (10 lg/ml) for the following times: 5 min for embryos up to the 4-somite stage (st.20), 10 min for 19-somite embryos (st.25), 20 min for 24-somite embryos (st.27), 30 min for later stage until st.35 embryos, 1 h 30 min for larvae. Whole embryos were cleared and mounted in 80% glycerol for observation under a microscope (Leica Z16 APO). Images were generated using a digital camera (Leica DFC420C). For fluorescent double-color in situ hybridization, the Ol-foxi3 riboprobe was labeled with digoxygenin, and atp1a1a.1 with fluorescein, and used at 1 ng/lL. Fluorescent in situ hybridizations were performed using the TSA™ PLUS Cy3/Fluorescein System (Perkin Elmer). The fluorescein-labeled probe was detected first with a rabbit anti-Fluorescein HRP-conjugate antibody (Roche), overnight at 4 °C. Embryos were processed using the Tyramid Signal Amplification Cy3 system for 10 min (dilution 1:50), extensively washed in PBST and incubated for 10 min at 65 °C in 50% formamide, 2 SSC, 0.1% Tween 20 prior to incubation with the anti-digoxygenin HRP-conjugate antibody. The digoxygenin-labeled probe was detected using the TSA-fluorescein system for 10 min (dilution 1:50). For triple staining, the embryos were not subjected to a proteinase K treatment.
191
Acknowledgements We wish to thank Pia Kiilerich for careful reading of the manuscript and helpful comments, Cécile Melin for her skilful maintenance of the fish and Patrick Prunet for his support. We thank the Core Facility of the Institute of Cellular and Organismic Biology for use of the SP5 confocal microscope. We are grateful to Fabrice Senger and François Tiaho from the PIXEL platform of Rennes 1 University for their assistance with the SP2 confocal microscope. This work was supported by the French Ministry of Foreign and European Affairs (PHC ORCHID 2009) and grants to P.P. Hwang from Academia Sinica, Taïwan, ROC (Thematic Project 2009).
References Bakkers, J., Hild, M., Kramer, C., Furutani-Seiki, M., Hammerschmidt, M., 2002. Zebrafish DeltaNp63 is a direct target of Bmp signaling and encodes a transcriptional repressor blocking neural specification in the ventral ectoderm. Dev. Cell 2, 617–627. Blanco, G., Mercer, R.W., 1998. Isozymes of the Na-K-ATPase: heterogeneity in structure, diversity in function. Am. J. Physiol. 275, F633–F650. Brown, D., Breton, S., 1996. Mitochondria-rich, proton-secreting epithelial cells. J. Exp. Biol. 199, 2345–2358. Chang, W.J., Horng, J.L., Yan, J.J., Hsiao, C.D., Hwang, P.P., 2009. The transcription factor, glial cell missing 2, is involved in differentiation and functional regulation of H+-ATPase-rich cells in zebrafish (Danio rerio). Am. J. Physiol. Regul. Integr. Comp. Physiol. 296, R1192–R1201. Esaki, M., Hoshijima, K., Nakamura, N., Munakata, K., Tanaka, M., Ookata, K., Asakawa, K., Kawakami, K., Wang, W., Weinberg, E.S., et al., 2009. Mechanism of development of ionocytes rich in vacuolar-type H(+)-ATPase in the skin of zebrafish larvae. Dev. Biol. 329, 116–129. Evans, D.H., Piermarini, P.M., Choe, K.P., 2005. The multifunctional fish gill: dominant site of gas exchange, osmoregulation, acid-base regulation, and excretion of nitrogenous waste. Physiol. Rev. 85, 97–177. Furutani-Seiki, M., Wittbrodt, J., 2004. Medaka and zebrafish, an evolutionary twin study. Mech. Dev. 121, 629–637. Graham, A., 2001. The development and evolution of the pharyngeal arches. J. Anat. 199, 133–141. Graham, A., Okabe, M., Quinlan, R., 2005. The role of the endoderm in the development and evolution of the pharyngeal arches. J. Anat. 207, 479–487. Hsiao, C.D., You, M.S., Guh, Y.J., Ma, M., Jiang, Y.J., Hwang, P.P., 2007. A positive regulatory loop between foxi3a and foxi3b is essential for specification and differentiation of zebrafish epidermal ionocytes. PLoS One 2, e302. Hulander, M., Kiernan, A.E., Blomqvist, S.R., Carlsson, P., Samuelsson, E.J., Johansson, B.R., Steel, K.P., Enerback, S., 2003. Lack of pendrin expression leads to deafness and expansion of the endolymphatic compartment in inner ears of Foxi1 null mutant mice. Development 130, 2013–2025. Hwang, P.P., 2009. Ion uptake and acid secretion in zebrafish (Danio rerio). J. Exp. Biol. 212, 1745–1752. Hwang, P.P., Lee, T.H., 2007. New insights into fish ion regulation and mitochondrion-rich cells. Comp. Biochem. Physiol. A; Mol. Integr. Physiol. 148, 479–497. Hwang, P.P., Perry, S.F., 2010. Ionic and acid–base regulation. In: Perry, S.F., Ekker, M., Farrell, A.P., Brauner, C.J. (Eds.), Fish Physiology: Zebrafish, vol. 28. Elsevier, San Diego. Inoue, K., Takei, Y., 2002. Diverse adaptability in Oryzias species to high environmental salinity. Zoolog. Sci. 19, 727–734. Inoue, K., Takei, Y., 2003. Asian medaka fishes offer new models for studying mechanisms of seawater adaptation. Comp. Biochem. Physiol. B; Biochem. Mol. Biol. 136, 635–645. Iwamatsu, T., 1994. Stages of normal development in the medaka Oryzias latipes. Zoolog. Sci. 11, 825–839. Iwamatsu, T., 2004. Stages of normal development in the medaka Oryzias latipes. Mech. Dev. 121, 605–618. Janicke, M., Carney, T.J., Hammerschmidt, M., 2007. Foxi3 transcription factors and Notch signaling control the formation of skin ionocytes from epidermal precursors of the zebrafish embryo. Dev. Biol. 307, 258–271. Jorgensen, P.L., 1974. Purification and characterization of (Na+ plus K+)-ATPase. IV. Estimation of the purity and of the molecular weight and polypeptide content per enzyme unit in preparations from the outer medulla of rabbit kidney. Biochim. Biophys. Acta 356, 53–67. Kang, C.K., Tsai, S.C., Lee, T.H., Hwang, P.P., 2008. Differential expression of branchial Na+/K(+)-ATPase of two medaka species, Oryzias latipes and Oryzias dancena, with different salinity tolerances acclimated to fresh water, brackish water and seawater. Comp. Biochem. Physiol. A; Mol. Integr. Physiol. 151, 566–575. Keys, A., Willmer, E.N., 1932. Chloride secreting cells in the gills of fishes, with special reference to the common eel. J. Physiol. 76, 368–378. Larkin, M.A., Blackshields, G., Brown, N.P., Chenna, R., McGettigan, P.A., McWilliam, H., Valentin, F., Wallace, I.M., Wilm, A., Lopez, R., et al., 2007. Clustal W and Clustal X version 2.0. Bioinformatics 23, 2947–2948.
192
V. Thermes et al. / Gene Expression Patterns 10 (2010) 185–192
Nechiporuk, A., Linbo, T., Poss, K.D., Raible, D.W., 2007. Specification of epibranchial placodes in zebrafish. Development 134, 611–623. Ohyama, T., Groves, A.K., 2004. Expression of mouse Foxi class genes in early craniofacial development. Dev. Dyn. 231, 640–646. Solomon, K.S., Kudoh, T., Dawid, I.B., Fritz, A., 2003a. Zebrafish foxi1 mediates otic placode formation and jaw development. Development 130, 929–940. Solomon, K.S., Logsdon Jr., J.M., Fritz, A., 2003b. Expression and phylogenetic analyses of three zebrafish FoxI class genes. Dev. Dyn. 228, 301–307. Sucre, E., Charmantier-Daures, M., Grousset, E., Charmantier, G., Cucchi-Mouillot, P., 2010. Embryonic occurrence of ionocytes in the sea bass Dicentrarchus labrax. Cell Tissue Res.
Takagi, Y., 1997. Meshwork arrangement of mitochondria-rich, Na+, K+-ATPase-rich cells in the saccular epithelium of rainbow trout (Oncorhynchus mykiss) inner ear. Anat. Rec. 248, 483–489. Therien, A.G., Blostein, R., 2000. Mechanisms of sodium pump regulation. Am. J. Physiol. Cell Physiol. 279, C541–C566. Wittbrodt, J., Shima, A., Schartl, M., 2002. Medaka–a model organism from the far East. Nat. Rev. Genet. 3, 53–64. Wu, S.C., Horng, J.L., Liu, S.T., Hwang, P.P., Wen, Z.H., Lin, C.S., Lin, L.Y., 2009. Ammonium-dependent sodium uptake in mitochondrion-rich cells of Medaka (Oryzias latipes) Larvae. Am. J. Physiol. Cell Physiol. Yamamoto, T., 1975. Medaka (Killifish) Biology and Strains. Keigaku, Tokyo.
Contents lists available at ScienceDirect
Gene Expression Patterns journal homepage: www.elsevier.com/locate/gep
Expression of Ol-foxi3 and Na+/K+-ATPase in ionocytes during the development of euryhaline medaka (Oryzias latipes) embryos Violette Thermes a,*, Chia-Cheng Lin b, Pung-Pung Hwang b a b
INRA, SCRIBE, Campus de Beaulieu, 35042 Rennes, France Institute of Cellular and Organismic Biology, Academia Sinica, Nangang, Taipei, Taiwan, ROC
a r t i c l e
i n f o
Article history: Received 1 February 2010 Received in revised form 30 March 2010 Accepted 2 April 2010 Available online 11 April 2010 Keywords: Fish Teleost Osmoregulation Gills Yolk sac Epidermis Skin Mitochondria-rich cells Chloride cells Ionocytes Progenitors atp1a1a.1 p63 Pharyngeal endoderm Epibranchial placodes Vitellin Zone Lateral Zone
a b s t r a c t Osmoregulation is a vital function that is essential to all vertebrates. Ionocytes are epithelial cells responsible for this function and have been extensively studied in adult teleost fish gills. The euryhaline medaka (Oryzias latipes) has recently emerged as an investigative model because of its ability to acclimatize easily to water presenting various salinities. However, no studies to date have focused on the development of ionocytes in medaka embryos. We first analyzed the distribution of ionocytes in the skin and gills during development, using a specific marker of differentiated ionocytes (the Na+/K+-ATPase pump, or NKA). Strikingly, we were able to identify two ionocyte domains on the yolk surface ectoderm, that we named the Vitellin Zone (VZ) and the Lateral Zone (LZ). In zebrafish, ionocyte differentiation has been shown to be controlled by two forkhead-box genes, foxi3a and foxi3b. We cloned the medaka foxi3 ortholog which appeared to be highly similar to foxi3b. Whole-mount in situ hybridizations performed on medaka embryos revealed that Ol-foxi3 is expressed in differentiated ionocytes of the pharyngeal endoderm, the branchial arches and the yolk epidermis, as well as in epibranchial placode territories. We further focused on the expression patterns of the yolk epidermis and compared the expression of Ol-foxi3 with that of the non-neural progenitor marker p63. We evidenced that Ol-foxi3 is expressed in progenitor cells which are first of all located uniformly in the VZ and then transitorily clustered in the LZ. Taken together, these data contribute to a clearer understanding of osmoregulatory tissue ontogenesis in euryhaline fish. Ó 2010 Elsevier B.V. All rights reserved.
1. Results and discussion Ion transport is a vital function that plays an essential role in the osmoregulation of body fluids in all vertebrates. The epithelia responsible for this function are located in several organs, including the kidney collecting duct, the amphibian urinary bladder, the gut, the inner ear, the fish gills and the skin (Takagi, 1997, for a review see Brown and Breton (1996), Evans et al. (2005) and Hwang and Lee (2007)). All transporting epithelia contain a subset of specialized cells called ionocytes (formally called mitochondria-rich cells or chloride cells) (Keys and Willmer, 1932), that transport various ions across the epithelia down gradients created by the strongly expressed plasma membrane enzymatic complex Na+/K+-ATPase (NKA) pump (Blanco and Mercer, 1998; Hwang and Lee, 2007; Therien and Blostein, 2000). Medaka (Oryzias lati-
* Corresponding author. Tel.: (+33) 2 23 48 57 20; fax: (+33) 2 23 48 50 20. E-mail address: [email protected] (V. Thermes). 1567-133X/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.gep.2010.04.001
pes) is a small fish model (for a review see Wittbrodt et al. (2002)) that displays particular adaptability to various water salinities (Inoue and Takei, 2002, 2003). The medaka can survive and develop normally when transferred from fresh water to sea water. For this reason, this euryhaline fish has recently emerged as an interesting model for osmoregulation studies (Kang et al., 2008; Wu et al., 2009). However, no studies to date have focused on the development of skin and gill epithelial cells during embryonic development in the medaka. The yolk skin epidermis contains several types of cells, including ionocytes, and develops after gastrulation from the non-neural ectoderm. In the pharyngeal embryonic region, the ectoderm contributes to formation of the branchial apparatus, which develops as a result of coordinated interplay between the pharyngeal ectoderm, the pharyngeal endoderm, the mesoderm and the epibranchial placodes. More specifically, the pharyngeal ectoderm gives rise to the gill epithelium, whereas the pharyngeal endoderm forms the epithelium that lines the pharynx and contributes to the formation of associated
186
V. Thermes et al. / Gene Expression Patterns 10 (2010) 185–192
organs. Both the gill epithelium and pharynx epithelia contain ionocytes. During vertebrate evolution, development of the pharyngeal apparatus has undergone a number of modifications. In terrestrial vertebrates, the parathyroid gland evolved as a result of the internalization of fish gills during the transition from an aquatic to a terrestrial environment (Graham et al., 2005). Nevertheless, it has been shown that the foxi class genes constitute a conserved feature of pharyngeal development and play a significant role in the establishment of branchial development (Graham, 2001; Solomon et al., 2003b). foxi class genes belong to a large and highly conserved gene family that encodes winged helix/forkhead-box transcription factors. They have been analyzed in detail in mouse and zebrafish. In mouse, three foxi genes have been identified and shown to display distinct and dynamic expression patterns in early craniofacial epithelia, including the otic vesicles (foxi1), cranial ectoderm (foxi2), early placodes of the surface ectoderm and the pharyngeal endoderm (foxi3) (Ohyama and Groves, 2004). Gene knock-out in the mouse revealed that Foxi1(/) mice display no abnormal development of the inner ear, but an absence of ionocyte differentiation (Hulander et al., 2003). In the zebrafish, two genes have been identified as being expressed in the epibranchial placodes and otic vesicles (foxi1), or in the chorda-mesoderm and pharyngeal arches at a later stage (foxi2). Zebrafish also present two members of the foxi3 gene, foxi3a and foxi3b, that are exclusively expressed in yolk epidermis ionocytes. Recent studies have found that foxi3 genes control ionocyte differentiation (Chang et al., 2009; Esaki et al., 2009; Hsiao et al., 2007; Janicke et al., 2007, for review see also Hwang and Perry (2010)). In order to gain a clear understanding of skin and gill ionocyte development in the medaka, we first of all analyzed the distribution of ionocytes during the development of skin and gill epithelia, using NKA immunostaining. We then cloned the medaka ortholog of zebrafish foxi3 key genes and examined its spatio-temporal expression during development, with a specific reference to ionocytes. 1.1. Na+/K+-ATPase-rich cell distribution during medaka embryonic development In order to determine the stages of emergence of ionocytes in embryonic skin and gills, the time course expression of the NKA pump was analyzed by immunostaining on whole mount embryos. NKA immunoreactivity was first observed at st.20, in a few scattered cells on the yolk epidermis (Fig. 1A). As development proceeded, an increasing number of NKA-immunoreactive cells (NKA-ir cells) could be detected on the yolk epidermis (Fig. 1A– F). Interestingly, at st.30 (35-somite stage), a higher density of NKA-ir cells was detected in a salt-and-pepper pattern in the axial yolk epidermis on both sides of the trunk (Fig. 1C). This area was named the Lateral Zone (LZ), while the surface covering the whole yolk was named the Vitellin Zone (VZ). Subsequently, LZ and VZ could easily be distinguished from the difference in NKA-ir cell density (Fig. 1C–F). Furthermore, at st.25 (18-somite stage) and st.30 (35-somite stage), NKA-ir cells could also be detected in the pharyngeal endoderm and in the first gill slit (also called the lateral gill chamber, Fig. 1G and H). Subsequently, from st.35 onwards, NKA-ir cells were detected on the epithelium of branchial arches as they develop (Fig. 1I and J). At the larva stage, numerous NKAir cells were found to be located on the developing primary gill filaments (Fig. 1K). 1.2. Cloning and expression analysis of Ol-foxi3 during medaka embryonic development We then focused our study on the foxi3 transcription factor and its expression in the skin and gills during development. For this
purpose, we cloned the medaka foxi3 ortholog detected by analyzing the medaka genome. Only one sequence related to zebrafish foxi3 genes could be identified and further sequence alignments enabled the construction of a phylogenetic tree (Fig. 2A). Fulllength sequences of the zebrafish FOXI3a and FOXI3b displayed 56.5% amino acid identity. The Fugu FOXI3 (Solomon et al., 2003b) is 55.3% and 71.5% identical to the zebrafish FOXI3a and FOXI3b, respectively. Although the Ol-FOXI3 protein resembles the zebrafish FOXI3a (56.5% identity), it is clearly more closely related to the Danio FOXI3b and Fugu FOXI3 (70.8% and 81.3% identity at the amino acid level, respectively). A multiple alignment of FOXI3 proteins (Fig. 2B) showed the forkhead conserved domains and confirmed that it is the medaka Ol-foxi3 ortholog. The expression pattern of Ol-foxi3 was analyzed by wholemount in situ hybridization (WMISH) in medaka embryos and larvae. At st.18 (late neurula stage), Ol-foxi3 mRNA expression was first detected in the anterior embryonic ectoderm flanking the midbrain–hindbrain boundary (MHB) region, which corresponds to the epibranchial (EB) placodes (Fig. 3A and I). This expression domain is maintained at st.20 (4-somite stage, Fig. 3B) but was no longer detected at st.25 (18-somite stage, Fig. 3C). From that stage, embryos displayed Ol-foxi3 expression in the pharyngeal endoderm (Fig. 3C–F). Indeed, lateral views of embryos at different stages display a high level of expression in the pharyngeal endoderm that gradually diminishes until st.35, where it is eventually only expressed in scarce cells (Fig. 3J–N). Conversely, from st.30, a punctuated pattern of Ol-foxi3 was detected in the developing pharyngeal ectoderm (i.e. branchial arches epithelium) that progressively extends from the rostral to the caudal extremities as the pharyngeal arches develop (Fig. 3K–N). Finally, from st.18, Olfoxi3 was also expressed in single cells scattered in the VZ domain of the yolk epidermis (Fig. 3A and H). This spotted and specific staining is congruent with an expression in ionocytes (see Fig. 1). Consistently with the ionocyte distribution described in Fig. 1, additional foxi3-positive cells were also observed at st.25 in the inner most part of the LZ (Fig. 3C). This staining gradually spreads as the embryo develops, whereas the Ol-foxi3 expression level in the VZ slightly diminishes. At the larva stage, Ol-foxi3 expression was observed in the ionocytes of both the gill epithelium and the yolk epidermis (Fig. 3O and H). In order to confirm that Ol-foxi3 is expressed in NKA-ir cells, we performed double fluorescent WMISH using Ol-foxi3 and atp1a1a.1 probes (Fig. 4). Indeed, the NKA complex contains a catalytic a subunit and a stabilizing b subunit (Jorgensen, 1974). In order to detect NKA at the gene level, the atp1a1a.1 gene encoding the NKA a1 subunit was cloned. Further expression analysis on adult gills enabled to demonstrate that atp1a1a.1 does indeed reflect NKA distribution (data not shown). At st.27–28 and at larva stage, double fluorescent WMISH confirmed a co-expression of Ol-foxi3 and atp1a1a.1 in the pharyngeal endoderm and the first gill slit (Fig. 4A–C), the skin ectoderm (Fig. 4D and F) and the gills (Fig. 4D and E). At st.20, no expression of atp1a1a.1 was detected in Ol-foxi3-expressing cells in the EB domain (data not shown). 1.3. The skin epidermis displays two territories of ionocyte progenitors One of the most striking features of Ol-foxi3 and NKA expression patterns is the existence of two ionocyte domains on the yolk epidermis, the VZ and LZ. The expression of NKA in the VZ, which was apparent at st.20, is delayed when compared with the onset of Olfoxi3 expression (compare Fig. 1A and Fig. 3A). This delay was also observed in the LZ, where NKA-ir cells appeared at st.30 and Olfoxi3 expression was seen as early as st.25 (compare Fig. 1C and Fig. 3C). Starting from this observation, we analyzed the co-expression of Ol-foxi3 and atp1a1a.1 in the LZ in more detail, between st.25 and st.30 (Fig. 5). As expected, we observed that Ol-foxi3 is
V. Thermes et al. / Gene Expression Patterns 10 (2010) 185–192
187
Fig. 1. Distribution of NKA-immunoreactive cells during medaka embryonic development. NKA-ir cells are detected in embryos (A–E and G–J) and in larvae (F and K) by whole-mount immunostaining. (A and F) From st.20 onwards, NKA-ir cells are detected on the yolk epidermis. The LZ and the VZ domains display different densities of NKA-ir cells. (G–J) Ventral views of embryos at different developmental stages. The development of branchial arches can be followed by nuclear DAPI staining. (G and H) At st.25 and st.30, NKA-ir cells are detected in the pharyngeal endoderm and the first gill slit. (I and J) At st.35 and st.39, NKA-ir cells are detected in the branchial ectoderm. (K) At the larval stage, NKA-ir cells are located on primary gill filaments. Note that NKA-ir cells are mainly located on the internal face (or efferent face, dotted line). (A) Embryo in ventral view, animal pole to the top. (G–K) Embryo and larva in ventral view, anterior to the top. (B–D) Embryo in dorsal view, anterior to the top. (E) Embryo in lateral view, anterior to the top. (F) Larva in lateral view, anterior to the left. LZ, Lateral Zone; VZ, Vitellin Zone; ba, branchial arches; ey, eye; gs, gill slit; and pe, pharyngeal endoderm. Scale bars: 100 lm in A–F and 50 lm in G–K.
first expressed at st.25 in a few cells lined up in the LZ on both sides of the body axis (Fig. 5A, arrows). Atp1a1a.1 mRNA was not detected in these cells, although atp1a1a.1 was highly expressed in other Ol-foxi3-positive cells of the yolk sac epidermis (i.e. in the VZ domain). At st.27–28, the number of Ol-foxi3-positive/atp1a1a.1-negative cells in the LZ increases (Fig. 5B). Then, at st.30, these latter cells gradually start to express atp1a1a.1, which ultimately leads to only a few Ol-foxi3-positive/atp1a1a.1-negative cells left in the LZ (Fig. 5C). Interestingly, the same gene activation sequence was observed in the VZ, between st.18 and st.25 (data not shown). Previous zebrafish studies had reported that foxi3a is expressed in both undifferentiated ionocyte progenitors (i.e. NKA-negative
cells) and differentiated ionocytes (i.e. NKA-positive cells) (Hsiao et al., 2007; Janicke et al., 2007). To investigate the possibility that transitory Ol-foxi3-positive/atp1a1a.1-negative cells of the LZ also correspond to ionocyte progenitors, we tested whether p63, an epidermal progenitor marker (Bakkers et al., 2002), is expressed in these cells. We conducted double fluorescent WMISH for Ol-foxi3 and atp1a1a.1 in combination with anti-p63 immunostaining, at st.27–28 (Fig. 5D). A high-magnification view showed that each of the Ol-foxi3-positive/atp1a1a.1-negative cells located in the LZ was p63-positive. Taken together, these data suggest that the LZ is a secondary site of ionocyte differentiation that appears at st.25, after a first wave of differentiation initiated in the VZ at st.18. We infer that the higher density of ionocytes in the LZ may
188
V. Thermes et al. / Gene Expression Patterns 10 (2010) 185–192
Fig. 2. Sequence analyses of Oryzias latipes FOXI3. (A) A phylogenetic tree of vertebrate foxi3 related genes based on a maximum likelihood distance analysis of the amino acid alignment of nine homologs. Numbers above branches indicate the bootstrap percentage. (B) Multiple alignment of full-length amino acid sequences of members of the Foxi3 family. The forkhead conserved domains are underlined and conserved residues are highlighted in black.
be due to the reduced surface of this territory compared with the large VZ surface that covers the yolk sac.
1.4. Concluding remarks Osmoregulation is a vital function that has been studied extensively in the skin and gills of teleost fish using morphological and physiological techniques. Despite the major significance of these mechanisms in the fish embryo, little is known about the occurrence of ionocytes at this stage of the life cycle. The present study thus focused on NKA expression in the skin and branchial region during medaka development. We observed that NKA-ir cells first appear on the yolk skin at st.20 (4-somite stage), on the pharyngeal endoderm and gill slit at st.25 (18-somite stage), and on the gill epithelium of late embryos (at st.35). Most studies that have investigated the embryonic ionocytes distribution were focused on the
yolk skin. Among them, only one study has investigated the enteric ionocyte distribution in the gut during embryonic development (Sucre et al., 2010). The ionocyte distribution was described in the first gill slits and gut of the sea bass embryo (Dicentrarchus labrax). The findings of our study revealed that the embryonic gut in medaka contains a high density of NKA-ir cells that also express the Ol-foxi3 gene. This result reinforces the idea that these are ion-transporting cells, although the biological significance of the ion transport triggered by these cells in the embryo still remains unclear. Moreover, the expression of Ol-foxi3 strongly suggests a function for this gene in the specification and differentiation of ionocytes in the gut, which is similar to that of zebrafish foxi3a and foxi3b in the yolk epidermis (Hsiao et al., 2007). This study also reports in detail on the expression pattern of Olfoxi3 during embryogenesis. Ol-foxi3 was found to be expressed in the epibranchial region at st.18 (late neurula stage), in ionocytes of the pharyngeal endoderm at st.25 (18-somite stage) and in ionocytes of the skin and gills at st.18 (late neurula stage) and st.30 (35-somite stage), respectively. Thus, Ol-foxi3 expression in ionocytes is not restricted to the yolk sac like the zebrafish foxi3a and foxi3b genes, which strongly suggests that Ol-foxi3 is involved in the development of other transporting epithelia such as the pharynx and gills (Hsiao et al., 2007; Janicke et al., 2007; Solomon et al., 2003b). Interestingly, Ol-foxi3 expression in the epibranchial region and pharyngeal endoderm is reminiscent of that of the zebrafish foxi1, expressed in epibranchial and otic placodes (Nechiporuk et al., 2007; Solomon et al., 2003a), and of the mouse foxi2 and foxi3 expressed in the surface ectoderm and pharyngeal endoderm of branchial pouches (Ohyama and Groves, 2004). However, phylogenetic analyses indicated that mouse foxi2 and foxi3 genes are not the orthologs of zebrafish foxi3a/3b, but that mouse foxi1 is the ortholog of these latter and thus also of Ol-foxi3 (Solomon et al., 2003b). This discrepancy between the expression patterns of medaka foxi3, zebrafish foxi3a/foxi3b and mouse foxi1, which all belong to the same foxi subgroup, may indicate a shift of function between paralogs in these different species. Finally, the expression of Ol-foxi3 in the pharyngeal region at st.25 suggests that the development of future gills is specified at this stage in medaka embryos, and that Ol-foxi3 is an early marker of the branchial apparatus. Finally, on the yolk epidermis, the Ol-foxi3 expression pattern is different from that of the zebrafish foxi3a and foxi3b genes (Hsiao et al., 2007; Janicke et al., 2007; Solomon et al., 2003b). In the zebrafish, specified epidermal ionocytes appear in the ventral ectoderm, from 90% epiboly to the 14-somite stage (Hsiao et al., 2007) or from the 3- to 17-somite stages as described in (Janicke et al., 2007). Three main ionocyte subtypes are produced, the NaR (Na+/K+-ATPase-Rich), the HR (H+-ATPase-Rich) and the NCC (N+/Cl Cotransporter) cells (Hwang, 2009). After these stages, no more ionocytes are generated on the yolk skin. It is accepted that foxi3a and foxi3b are responsible for the segregation and differentiation of these subpopulations. In the present case, we evidenced that medaka displays two waves of ionocyte differentiation in the embryonic yolk sac. ‘Ionopoiesis’ is first initiated in the VZ (at st.18, late neurula stage) and then reactivated in the LZ of the yolk sac epidermis (at st.25, 18-somite stage). Although it is not yet known whether different subtypes of ionocytes exist in medaka, it is possible to hypothesize that two undifferentiated ionocyte subpopulations may be generated successively within these two sites. On the other hand, we assumed that only one foxi3 gene, highly similar to zebrafish foxi3b, is present in the medaka genome, which apparently does not favor ionocyte diversity. However, one can envisage a scenario under which possible ionocyte subpopulations might be under the control of key factors other than Ol-foxi3 that are perhaps specific to each domain. Further molecular and cellular analyses will be necessary to understand the physiological outcome of this cell organization in the medaka yolk sac epidermis.
V. Thermes et al. / Gene Expression Patterns 10 (2010) 185–192
189
Fig. 3. Ol-foxi3 expression during medaka embryonic development. Whole-mount in situ hybridization (WMISH) on medaka embryos and larvae at the stages indicated. (A and B) At st.18 and st.20, Ol-foxi3 mRNA are detected in a salt-and-pepper pattern on the yolk epidermis (arrowheads, inset). (C–H) From st.25 onwards, Ol-foxi3 is detected in scattered cells in the LZ and VZ domains. (A and I, inset) Ol-foxi3 mRNA are detected in the rostral ectoderm in the epibranchial region. (J–O) From st.27 to larva stage, Ol-foxi3 is expressed in the pharyngeal endoderm and branchial arches. At st.35, scare cells in the pharyngeal endoderm and intestine express Ol-foxi3. (A–G) dorsal views, (O–H) lateral views, (N) ventral view. Anterior to the left. EB, epibranchial region; LZ, Lateral Zone; VZ, Vitellin Zone; f, fin; g, gills; int, intestine; pa, pharyngeal arches; and pe, pharyngeal endoderm. Scale bars: 50 lm.
To conclude, our data show that Ol-foxi3 is expressed in both the yolk epidermis and the early craniofacial epithelia. Moreover, we show that ionocyte development (or ‘‘ionopoiesis”) in medaka embryos occurs at two sites of the yolk sac, thus differing from zebrafish embryos. The present study shows that the medaka constitutes a new embryological model to study skin/gill development as well as ionocyte ontogeny, and should pave the way for new physiological studies of osmoregulation in euryhaline embryos.
2. Experimental procedures 2.1. Medaka breeding strain Medaka embryos and larvae of the CAB strain were raised at 26 °C under a reproduction regime (14 h light/10 h dark). Fertilized eggs were collected after spawning (at the onset of light) and incubated in Yamamoto’s embryo rearing medium (Yamamoto, 1975).
190
V. Thermes et al. / Gene Expression Patterns 10 (2010) 185–192
Fig. 4. Co-localization of Ol-foxi3 and atp1a1a.1 in embryo and larva. Double fluorescent WMISH with Ol-foxi3 (in green) and atp1a1a.1 (in red) probes. (A–C) St.27–28 embryos. Ol-foxi3 and atp1a1a.1 are co-expressed in the pharyngeal endoderm (A and B) and the first gill slit (C). Ol-foxi3-positive/atp1a1a.1-negative cells are detected in the LZ (A, arrow). (D–F) Ol-foxi3 and atp1a1a.1 are co-expressed in ionocytes of the gill ectoderm (E) and yolk skin (F). (B) Longitudinal optic section as indicated in (A). Embryo and larva are orientated with the anterior to the top. (C) Transversal optic section as indicated in (B). LZ, Lateral Zone; VZ, Vitellin Zone; gs; gill slit; and pe, pharyngeal endoderm. Scale bars: 100 lm in A, B, and D and 50 lm in C, E, and F.
Fig. 5. Medaka embryos display a secondary site of ionocyte progenitors in the epidermis. Double fluorescent WMISH with Ol-foxi3 (in green) and atp1a1a.1 (in red) were performed between st.25 and st.30 (A–C). Triple staining with p63 antibody (in blue) was performed at st.27–28 (D). (A) St.25 embryos show Ol-foxi3-positive/atp1a1a.1negative cells in the LZ (arrows) and Ol-foxi3-positive/atp1a1a.1-positive cells in the VZ. (B) At st.27–28, embryos display a higher number of Ol-foxi3-positive/atp1a1a.1negative cells in the LZ. (C) At st.30, only scarce Ol-foxi3-positive/atp1a1a.1-negative cells are observed, while the number of Ol-foxi3-positive/atp1a1a.1-positive cells increases in the LZ. (D) At st. 27–28, Ol-foxi3-positive/atp1a1a.1-negative cells of the LZ co-express the p63 progenitor marker (asterisks and inset) indicating that these cells are ionocyte progenitors. Ol-foxi3-positive/atp1a1a.1-positive staining is always excluded from p63-positive cells. Embryo and larva are in dorsal view, anterior to the top. LZ, Lateral Zone and VZ, Vitellin Zone. Scale bars: 130 lm in A–D and 10 lm in D, inset.
During all experiments, the embryos were placed at 26 °C and staged according to the developmental table described by Iwamatsu (1994, 2004) and Furutani-Seiki and Wittbrodt (2004). 2.2. Sequence analysis and cDNA cloning In order to identify medaka foxi3 and atp1a1a.1 sequences we used the zebrafish FOXI amino acid sequence for a BLAST search
on the Medaka Genome project database (http://dolphin.lab.nig.ac.jp/medaka/). The partial sequences identified were used to design specific primers for 50 and 30 RACE PCR. To clone the cDNA sequence, total RNA was extracted from adult medaka gills using Trizol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions, and reverse transcription was further performed following the instructions in the SMART RACE cDNA Amplification Kit (Clontech, CA, USA). To synthesize Ol-foxi3
V. Thermes et al. / Gene Expression Patterns 10 (2010) 185–192
and atp1a1a riboprobes, cDNA fragments of 754 pb and 722 pb were isolated by PCR on the synthesized SMART cDNA, using the following primers: Ol-foxi3 forward 50 -TGTACCTGGAT 0 CGTCCTCGTCTCGG and Ol-foxi3 reverse 5 - GGTTGGTAGTTTAGA GACGCCTG; atp1a1a forward 50 -AGGAGGAGCCA GCAAACGA TAAT-30 and atp1a1a reverse 50 -TTCCAGGTTCTTCACCAAGCAAT 30 . PCR fragments were subcloned into pGEM-T Easy (Promega, WI, USA) and then sequenced. We carefully compared the Ol-FOXI3 amino acid sequence with the Danio rerio FOXI3a (NP_944599) and FOXI3b (NP_944600), Xenopus tropicalis FOXI3 (Q6P8A3), Takifugu rubripes FOXI3 (CAG10122), mouse FOXI1 (NP_001094934), rat FOXI1 (NP_001099246), human FOXI1 (NP_001129121) and Ciona intestinalis FOXI (BAE06443) sequences available on the NCBI server and the Ensemble genome browser. CLUSTALX version 2.0.11 (Larkin et al., 2007) was used for multiple alignments and the tree was constructed by Neighbor-Joining using MEGA version 4.1 (Larkin et al., 2007).
2.3. Whole-mount immunostaining and in situ hybridization Embryos were fixed with 4% paraformaldehyde (PFA) in phosphate buffer 0.12 M (PBS, pH7.4), overnight at room temperature. Larvae were anesthetized on ice and fixed overnight at 4 °C. Embryos and larvae were rinsed in PBS. Embryos were manually dechlorinated with fine forceps. The specimens were then dehydrated in 100% methanol and stored at 20 °C for subsequent analyses. For whole-mount immunostaining, the specimens were rehydrated in PBST (0.1% Tween 20 in PBS), incubated for 1 h with 3% bovine serum albumin (BSA) and then incubated overnight at 4 °C with the primary antibody. Late embryos were subjected to mild proteinase K treatment (10 lg/ml) for 30 min for st.35, and for 10 min for st.39. The Na+/K+-ATPase pump was detected with the mouse monoclonal anti-a5 (Developmental Studies Hybridoma Bank, University of Iowa, 1:300). The DNp63 protein was detected with the mouse monoclonal anti-p63 (4A4, Santa Cruz, 1:200). After rinsing for 3 h in PBST, the samples were incubated for 1 h with the secondary antibody. A goat anti-mouse IgG-Alexa 568 (1:200 in PBS) and a goat anti-mouse IgG-Alexa 594 (1:200 in PBS) were used to detect anti-a5 and anti-p63, respectively. Embryos were finally stained with DAPI (Invitrogene) for global morphology. Fluorescence was detected under a confocal microscope (SP5 and SP3, Leica). For single-color whole mount in situ hybridization, Ol-foxi3 antisense digoxygenine-UTP probes were generated by in vitro transcription. Embryos were subjected to a proteinase K treatment (10 lg/ml) for the following times: 5 min for embryos up to the 4-somite stage (st.20), 10 min for 19-somite embryos (st.25), 20 min for 24-somite embryos (st.27), 30 min for later stage until st.35 embryos, 1 h 30 min for larvae. Whole embryos were cleared and mounted in 80% glycerol for observation under a microscope (Leica Z16 APO). Images were generated using a digital camera (Leica DFC420C). For fluorescent double-color in situ hybridization, the Ol-foxi3 riboprobe was labeled with digoxygenin, and atp1a1a.1 with fluorescein, and used at 1 ng/lL. Fluorescent in situ hybridizations were performed using the TSA™ PLUS Cy3/Fluorescein System (Perkin Elmer). The fluorescein-labeled probe was detected first with a rabbit anti-Fluorescein HRP-conjugate antibody (Roche), overnight at 4 °C. Embryos were processed using the Tyramid Signal Amplification Cy3 system for 10 min (dilution 1:50), extensively washed in PBST and incubated for 10 min at 65 °C in 50% formamide, 2 SSC, 0.1% Tween 20 prior to incubation with the anti-digoxygenin HRP-conjugate antibody. The digoxygenin-labeled probe was detected using the TSA-fluorescein system for 10 min (dilution 1:50). For triple staining, the embryos were not subjected to a proteinase K treatment.
191
Acknowledgements We wish to thank Pia Kiilerich for careful reading of the manuscript and helpful comments, Cécile Melin for her skilful maintenance of the fish and Patrick Prunet for his support. We thank the Core Facility of the Institute of Cellular and Organismic Biology for use of the SP5 confocal microscope. We are grateful to Fabrice Senger and François Tiaho from the PIXEL platform of Rennes 1 University for their assistance with the SP2 confocal microscope. This work was supported by the French Ministry of Foreign and European Affairs (PHC ORCHID 2009) and grants to P.P. Hwang from Academia Sinica, Taïwan, ROC (Thematic Project 2009).
References Bakkers, J., Hild, M., Kramer, C., Furutani-Seiki, M., Hammerschmidt, M., 2002. Zebrafish DeltaNp63 is a direct target of Bmp signaling and encodes a transcriptional repressor blocking neural specification in the ventral ectoderm. Dev. Cell 2, 617–627. Blanco, G., Mercer, R.W., 1998. Isozymes of the Na-K-ATPase: heterogeneity in structure, diversity in function. Am. J. Physiol. 275, F633–F650. Brown, D., Breton, S., 1996. Mitochondria-rich, proton-secreting epithelial cells. J. Exp. Biol. 199, 2345–2358. Chang, W.J., Horng, J.L., Yan, J.J., Hsiao, C.D., Hwang, P.P., 2009. The transcription factor, glial cell missing 2, is involved in differentiation and functional regulation of H+-ATPase-rich cells in zebrafish (Danio rerio). Am. J. Physiol. Regul. Integr. Comp. Physiol. 296, R1192–R1201. Esaki, M., Hoshijima, K., Nakamura, N., Munakata, K., Tanaka, M., Ookata, K., Asakawa, K., Kawakami, K., Wang, W., Weinberg, E.S., et al., 2009. Mechanism of development of ionocytes rich in vacuolar-type H(+)-ATPase in the skin of zebrafish larvae. Dev. Biol. 329, 116–129. Evans, D.H., Piermarini, P.M., Choe, K.P., 2005. The multifunctional fish gill: dominant site of gas exchange, osmoregulation, acid-base regulation, and excretion of nitrogenous waste. Physiol. Rev. 85, 97–177. Furutani-Seiki, M., Wittbrodt, J., 2004. Medaka and zebrafish, an evolutionary twin study. Mech. Dev. 121, 629–637. Graham, A., 2001. The development and evolution of the pharyngeal arches. J. Anat. 199, 133–141. Graham, A., Okabe, M., Quinlan, R., 2005. The role of the endoderm in the development and evolution of the pharyngeal arches. J. Anat. 207, 479–487. Hsiao, C.D., You, M.S., Guh, Y.J., Ma, M., Jiang, Y.J., Hwang, P.P., 2007. A positive regulatory loop between foxi3a and foxi3b is essential for specification and differentiation of zebrafish epidermal ionocytes. PLoS One 2, e302. Hulander, M., Kiernan, A.E., Blomqvist, S.R., Carlsson, P., Samuelsson, E.J., Johansson, B.R., Steel, K.P., Enerback, S., 2003. Lack of pendrin expression leads to deafness and expansion of the endolymphatic compartment in inner ears of Foxi1 null mutant mice. Development 130, 2013–2025. Hwang, P.P., 2009. Ion uptake and acid secretion in zebrafish (Danio rerio). J. Exp. Biol. 212, 1745–1752. Hwang, P.P., Lee, T.H., 2007. New insights into fish ion regulation and mitochondrion-rich cells. Comp. Biochem. Physiol. A; Mol. Integr. Physiol. 148, 479–497. Hwang, P.P., Perry, S.F., 2010. Ionic and acid–base regulation. In: Perry, S.F., Ekker, M., Farrell, A.P., Brauner, C.J. (Eds.), Fish Physiology: Zebrafish, vol. 28. Elsevier, San Diego. Inoue, K., Takei, Y., 2002. Diverse adaptability in Oryzias species to high environmental salinity. Zoolog. Sci. 19, 727–734. Inoue, K., Takei, Y., 2003. Asian medaka fishes offer new models for studying mechanisms of seawater adaptation. Comp. Biochem. Physiol. B; Biochem. Mol. Biol. 136, 635–645. Iwamatsu, T., 1994. Stages of normal development in the medaka Oryzias latipes. Zoolog. Sci. 11, 825–839. Iwamatsu, T., 2004. Stages of normal development in the medaka Oryzias latipes. Mech. Dev. 121, 605–618. Janicke, M., Carney, T.J., Hammerschmidt, M., 2007. Foxi3 transcription factors and Notch signaling control the formation of skin ionocytes from epidermal precursors of the zebrafish embryo. Dev. Biol. 307, 258–271. Jorgensen, P.L., 1974. Purification and characterization of (Na+ plus K+)-ATPase. IV. Estimation of the purity and of the molecular weight and polypeptide content per enzyme unit in preparations from the outer medulla of rabbit kidney. Biochim. Biophys. Acta 356, 53–67. Kang, C.K., Tsai, S.C., Lee, T.H., Hwang, P.P., 2008. Differential expression of branchial Na+/K(+)-ATPase of two medaka species, Oryzias latipes and Oryzias dancena, with different salinity tolerances acclimated to fresh water, brackish water and seawater. Comp. Biochem. Physiol. A; Mol. Integr. Physiol. 151, 566–575. Keys, A., Willmer, E.N., 1932. Chloride secreting cells in the gills of fishes, with special reference to the common eel. J. Physiol. 76, 368–378. Larkin, M.A., Blackshields, G., Brown, N.P., Chenna, R., McGettigan, P.A., McWilliam, H., Valentin, F., Wallace, I.M., Wilm, A., Lopez, R., et al., 2007. Clustal W and Clustal X version 2.0. Bioinformatics 23, 2947–2948.
192
V. Thermes et al. / Gene Expression Patterns 10 (2010) 185–192
Nechiporuk, A., Linbo, T., Poss, K.D., Raible, D.W., 2007. Specification of epibranchial placodes in zebrafish. Development 134, 611–623. Ohyama, T., Groves, A.K., 2004. Expression of mouse Foxi class genes in early craniofacial development. Dev. Dyn. 231, 640–646. Solomon, K.S., Kudoh, T., Dawid, I.B., Fritz, A., 2003a. Zebrafish foxi1 mediates otic placode formation and jaw development. Development 130, 929–940. Solomon, K.S., Logsdon Jr., J.M., Fritz, A., 2003b. Expression and phylogenetic analyses of three zebrafish FoxI class genes. Dev. Dyn. 228, 301–307. Sucre, E., Charmantier-Daures, M., Grousset, E., Charmantier, G., Cucchi-Mouillot, P., 2010. Embryonic occurrence of ionocytes in the sea bass Dicentrarchus labrax. Cell Tissue Res.
Takagi, Y., 1997. Meshwork arrangement of mitochondria-rich, Na+, K+-ATPase-rich cells in the saccular epithelium of rainbow trout (Oncorhynchus mykiss) inner ear. Anat. Rec. 248, 483–489. Therien, A.G., Blostein, R., 2000. Mechanisms of sodium pump regulation. Am. J. Physiol. Cell Physiol. 279, C541–C566. Wittbrodt, J., Shima, A., Schartl, M., 2002. Medaka–a model organism from the far East. Nat. Rev. Genet. 3, 53–64. Wu, S.C., Horng, J.L., Liu, S.T., Hwang, P.P., Wen, Z.H., Lin, C.S., Lin, L.Y., 2009. Ammonium-dependent sodium uptake in mitochondrion-rich cells of Medaka (Oryzias latipes) Larvae. Am. J. Physiol. Cell Physiol. Yamamoto, T., 1975. Medaka (Killifish) Biology and Strains. Keigaku, Tokyo.