Uptake of Na+ from the environment is an indispensable strategy for the survival of freshwater fish, as they easily lose Na+ from the plasma to a diluted environment. Nevertheless, the location of and molecules involved in Na+ uptake remain poorly understood. In this study, we utilized Sodium Green, a Na+-dependent fluorescent reagent, to provide direct evidence that Na+ absorption takes place in a subset of the mitochondria-rich (MR) cells on the yolk sac surface of zebrafish larvae. Combined with immunohistochemistry, we revealed that the Na+-absorbing MR cells were exceptionally rich in vacuolar-type H+-ATPase (H+-ATPase) but moderately rich in Na+-K+-ATPase. We also addressed the function of foxi3a, a transcription factor that is specifically expressed in the H+-ATPase-rich MR cells. When foxi3a was depleted from zebrafish embryos by antisense morpholino oligonucleotide injection, differentiation of the MR cells was completely blocked and Na+ influx was severely reduced, indicating that MR cells are the primary sites for Na+ absorption. Additionally, foxi3a expression is initiated at the gastrula stage in the presumptive ectoderm; thus, we propose that foxi3a is a key gene in the control of MR cell differentiation. We also utilized a set of ion transport inhibitors to assess the molecules involved in the process and discuss the observations.
- chloride cell
- Sodium Green
- Na+/H+ exchanger; vacuolar-type H+-ATPase
survival of freshwater fish requires active Na+ uptake from an extremely hypoosmotic environment to maintain plasma Na+ levels at a concentration at least 100-fold higher than the environment (10). The gills of freshwater fish are known to play critical roles not only in gas exchange but also in the uptake of a variety of ions. Among the epithelial cells on the gills, the mitochondria-rich (MR) cells, also called chloride cells or ionocytes, have been presumed to be the primary sites for ion uptake (8, 10, 19, 25, 37, 43, 48). The cytoplasm of MR cells is rich in mitochondria, and the basolateral membrane is extensively invaginated into the cytoplasm to form a tubular structure on which Na+-K+-ATPase accumulates. These properties enable the cells to produce abundant chemical energy and an extensive electrical gradient, which allows the cells to incorporate ions against a steep concentration gradient. MR cells also contain an abundance of several ion transporters, further support for their role as the primary site of ion uptake (6, 7, 18, 33, 41, 56, 57). Additionally, a recent study showed that among the dispersed cell fractions from the gill epithelium of freshwater fish, the MR cell fraction has a higher activity of Na+ absorption than the pavement cell fraction, another major epithelial cell type on the gills (46).
Since the first suggestion by Krogh (28) that Na+ uptake is likely coupled to the excretion of H+ to maintain an electroneutral state, two different molecular mechanisms have been proposed. In the first model, the Na+/ H+ exchanger (NHE) is considered a key molecule for the incorporation of ambient Na+ by electroneutrally exchanging Na+ with H+ inside the cells (61). However, in the gill epithelium, this mechanism has been questioned on thermodynamic grounds, as the gill epithelium is surrounded by an extremely dilute environment (2, 25, 26, 32, 44). An alternative model postulates Na+ uptake through the epithelial Na+ channel (ENaC) that is electrogenically coupled to vacuolar-type H+-ATPase (H+-ATPase) (2, 31). In this model, energy consumptive excretion of H+ by H+-ATPase creates an electrical gradient that provides the driving force to draw in Na+ through ENaC from the diluted environment. This model has been supported by several pharmacological studies with selective inhibitors of ENaC and H+-ATPase (9, 11, 31, 32, 45, 46). Currently, identification of an ENaC homologue from fish has not been successful. However, immunolocalization of ENaC with a heterologous antibody has implied its existence in fish (17, 52, 56). Studies of the immunolocalization of H+-ATPase have revealed a subcellular localization to the apical membranes and have also shown that it is expressed in MR cells of mudskipper (Periophthalmodon schlosseri) (57), coho salmon (Oncorhynchus kisutch) (58), and zebrafish (Danio rerio) (33); in pavement cells of tilapia (O. mossambicus) (17, 56); and in both cell types of rainbow trout (O. mykiss) (30, 56). Although MR cells have been assumed to be the primary sites of Na+ absorption, these observations suggest that pavement cells may also have the potential to absorb Na+, at least in some species.
Notably, Kaneko and colleagues (22) have demonstrated that the H+-ATPase is localized to the basolateral, but not apical membrane of MR cells in killifish (Fundulus heteroclitus), suggesting that molecules other than H+-ATPase may be primarily involved in the Na+ uptake process in this species. On the other hand, isoforms of NHE have been shown to be localized to the apical side of MR cells in mudskipper (57) and Osorezan dace (Tribolodon hakonensis) (16). Moreover, Choe et al. (6) found that in an elasmobranch, the Atlantic stingray (Dasyatis sabina), NHE3 is localized to the apical membrane of the MR cells and expression is enhanced by acclimation to freshwater. In addition, selective inhibitors of NHE block Na+ uptake in goldfish (Carassius auratus) (45). These observations suggest that NHE, rather than ENaC coupled to H+-ATPase, may be critically involved in Na+ uptake in some species or in some circumstances, although the thermodynamic problem has yet to be resolved.
Zebrafish is a stenohaline freshwater fish. It has been extensively utilized as a model vertebrate for studies of embryogenesis and organogenesis, but is also increasingly recognized as a useful model for the investigation of physiological traits (3, 33, 35, 41, 47). During zebrafish development, the MR cells first appear on the body surface, especially on the yolk sac, prior to formation of the gills. The larval MR cells seem to be functionally equivalent to those on the gills at later stages (17, 18, 21, 33, 34). Furthermore, their localization to the spherical yolk sac, a much simpler structure than the gills, provides better opportunities to analyze morphology, gene expression, and functions of the MR cells in vivo.
In this study, we addressed the location of, and molecules involved in, Na+ absorption in zebrafish larvae. First, using Sodium Green, a Na+-dependent fluorescent reagent, we succeeded in visualizing the Na+-absorbing cells and obtained direct evidence that Na+ absorption takes place in a subset of the MR cells on the body surface, especially those on the yolk sac. Notably, this cell population was not so rich in Na+-K+-ATPase, but contained abundant H+-ATPase. Second, we examined the function of foxi3a, a transcription factor that is sporadically expressed on the yolk sac surface postgastrulation (51). We found that it was specifically expressed in the H+-ATPase-rich MR cells. Furthermore, when foxi3a was depleted from embryos by injection of an antisense morpholino oligonucleotide, differentiation of MR cells was inhibited and Na+ absorption was severely reduced, indicating that MR cells are the primary sites for Na+ absorption. Finally, we examined molecules involved in the process by using of a set of ion transport inhibitors.
MATERIALS AND METHODS
Zebrafish were maintained as described previously (55). The wild-type zebrafish TL line was generously provided by Dr. A. Kawakami (Yokohama, Japan). Fertilized eggs were incubated in 1× freshwater [1× FW: 60 mg ocean salt (Rohtomarine) per 1 liter of distilled water] at 28.5°C unless otherwise mentioned. Here 1× FW was used as our standard freshwater and corresponded to “Fish Water” as described in “The Zebrafish Book” (55). The concentration of major ions, pH, and osmolarity in 1× FW are listed in Supplemental Table S1 (the online version of this article contains the supplemental table). Hatched or dechorionated larvae were used for the following experiments without feeding. The animal protocols and procedures were approved by the Institutional Animal Care and Use Committee of Tokyo Institute of Technology and conform to the American Physiological Society’s “Guiding Principles in the Care and Use of Laboratory Animals” (1).
Na+ accumulation analyses.
Na+-dependent fluorescent reagent, Sodium Green tetra-acetate cell permeant, was purchased from Invitrogen and a 10-mM stock solution was prepared in DMSO. Sodium Green has been well characterized as a reagent for measuring intracellular Na+ concentration, especially in cell culture (1a, 29, 49, 59). On binding Na+, it exhibits an increase in fluorescence intensity in the range of 0 to 135 mM Na+ (13). At ∼55 h and 7 days postfertilization (hpf or dpf), living larvae were incubated in 10 μM Sodium Green diluted in 1× FW for 1 h. Subsequently, larvae were washed briefly with 1× FW, anesthetized in 0.2 mg/ml tricaine (3-amino benzoic acid ethylester, Nacalai Tesque), 1 mM Tris·HCl, pH 7.0, and embedded in 2% methylcellulose in 1× FW for photography with an Axiovert 200 inverted fluorescent microscope (Zeiss). Sodium Green was excited by a mercury vapor light source through an excitation filter with maximum transmission at 485 nm and a 20-nm bandwidth. The fluorescence was collected using a 515- to 565-nm band-pass emission filter. We have confirmed that the tricaine treatment did not noticeably affect illumination of Sodium Green. Single focused images were created from a series of partially focused images using software from HeliconFocus (Helicon, Ukraine). Mitochondrial-staining reagent, MitoTracker (Invitrogen), was used to illuminate the MR cells (12, 14, 33, 38). For double staining with Sodium Green and MitoTracker, ∼55-hpf larvae were first incubated for 30 min in 10 μM Sodium Green and then in 10 μM Sodium Green and 500 nM MitoTracker Orange CMTMRos for an additional 30 min. In double stains of Sodium Green with concanavalin A (ConA; a lectin protein capable of selectively binding α-mannopyranosyl and α-glucopyranosyl residues), larvae were first incubated in 10 μM Sodium Green for 30 min and subsequently in 10 μM Sodium Green and 50 μg/ml of Alexa Fluor 594-conjugated ConA (Invitrogen) for 30 min.
To examine the effects of Na+ and K+ ionophores on Sodium Green illumination, we used monensin (WAKO) as the Na+ ionophore and valinomycin (Invitrogen) as the K+ ionophore. Larvae (∼55 hpf) were incubated for 40 min in 10 μM Sodium Green diluted in 1× FW in the presence of 1 μM monensin or 1 μM valinomycin. The larvae were subsequently anesthetized and embedded in 1.5% methylcellulose in the presence of the ionophore.
To analyze Na+ accumulation in zebrafish larvae reared in a range of Na+ concentrations, we prepared 100× FW (6 g ocean salt per liter of distilled water, pH 6.1–6.8), 1× FW (as described above), 1/20× FW (3 mg ocean salt per liter, pH 7.2) and distilled water. Larvae were reared in the different mediums from fertilization. At ∼55 hpf, larvae were incubated for 1 h in 10 μM Sodium Green diluted in each medium. They were washed with each medium and anesthetized in tricaine to be photographed. We note that the zebrafish larvae were able to grow in distilled water at least until 4 dpf without any malformation (K. Hoshijima, unpublished observations).
Blockage of Na+ accumulation with ion transport inhibitors.
EIPA (Sigma) is used as a selective inhibitor of the NHE (3, 27, 45). Amiloride (Sigma) targets both the ENaC and the NHE, but ENaC is more efficiently inhibited than NHE at low concentrations (3, 27, 60). Bafilomycin A1 (Wako) is an inhibitor of H+-ATPase (4) and has been shown to effectively inhibit H+ efflux from the MR cells on the yolk sac of zebrafish larvae at 10 μM (33). Metolazone (WAKO), an analog of thiazide, is used as an inhibitor of Na+-K+-2Cl− cotransporter (NKCC) (54). Ouabain (WAKO) is an inhibitor of Na+-K+-ATPase. Ethoxzolamide (Sigma) is an inhibitor of carbonic anhydrase (CA) (3). It is noteworthy that the inhibitory effects of these reagents have been tested in zebrafish or other fish (3, 11, 33, 45, 46, 54, 60, 62). The test concentrations used in this study were determined according to the references.
To examine blockage of Na+ absorption by the ion transport inhibitors, larvae were reared in 20-fold diluted FW, which reduces Na+ in the MR cells. At ∼55 hpf, larvae were preincubated for 1.5 h in 1× FW in the presence of the inhibitors, followed by incubation in 1× FW for 1 h with the inhibitors and 10 μM Sodium Green. The Sodium Green reaction was terminated by washing with 1× FW containing the inhibitors. The inhibitors were also added during anesthesia.
At ∼55 hpf, larvae were fixed with 4% paraformaldehyde in PBS (in mM: 137 NaCl, 27 KCl, 10 Na2HPO4, 2 KH2PO4, pH 7.4) at 4°C overnight and dehydrated in 100% methanol. For double staining with ConA, the larvae were incubated with 50 μg/ml Alexa Fluor 594-conjugated ConA for 30 min prior to fixation. After rehydration in PBS with 0.1% Tween-20 (PBT) and incubation in PBT containing 10% sheep serum (Sigma) for 1 h, the larvae were incubated at 25°C overnight with an antibody against the B subunit of dace (Tribolodon hakonensis) H+-ATPase (diluted 1:500 with PBT containing 10% sheep serum) (16) or with an antibody against the α-subunit of eel (Anguilla japonica) Na+-K+-ATPase (diluted 1:1,000) (39). Following a wash with PBT, the larvae were further incubated for 2 h at 25°C with donkey anti-rabbit IgG conjugated with Alexa Fluor 488 (diluted 1:1,000, Invitrogen) and then washed with PBT and embedded in 100% glycerol to photograph. We note that the peptide sequence of the B subunit of dace H+-ATPase (GenPept accession no. BAC75967; amino acid residues 1–160) used for antibody preparation exhibits 99% identity to that of zebrafish (NP_878299; residues 201–360) and that the peptide sequence of the α-subunit of eel Na+-K+-ATPase (2122241A; residues 469–773) used for antibody preparation exhibits 92% identity to that of zebrafish (NP_571761; residues 473–779).
Whole mount in situ hybridization.
Whole mount in situ hybridization was performed as previously described with brief modifications (53), using a digoxigenin-labeled RNA probe for foxi3a corresponding to a 543-bp fragment of cDNA (51) or a probe for CA isoform 2 (ca2) corresponding to a 583-bp fragment of the cDNA (GenBank accession no. NM_199215; nt 57–639). We note that the ca2 is different from the paralog cahz (GenBank accession no. NM_131110) that is expressed in blood. To combine with immunohistochemistry, larvae hybridized with digoxigenin-labeled RNA probe were incubated with anti-digoxigenin Fab fragment conjugated with alkaline phosphatase (diluted 1:4,000; Roche) and anti-dace H+-ATPase antibody (diluted 1:500) simultaneously. After incubation with donkey anti-rabbit IgG conjugated with Alexa Fluor 488, larvae were photographed as described above. Larvae were then washed with PBT and incubated in BCIP/NBT color development substrate (Promega) to visualize the in situ hybridization. In double stains with Sodium Green, living larvae were first stained with Sodium Green and photographed as described above. Subsequently, they were fixed with 4% paraformaldehyde in PBS at 4°C overnight and dehydrated in 100% methanol and were then processed according to the in situ hybridization protocol.
Measurement of Na+ influx with 22Na+.
Ten ∼55-hpf larvae were transferred into a 1.5-ml microcentrifuge tube. After being rinsed with 1 ml of 1× FW, the larvae were incubated in 200 μl of 1× FW containing 7.4 kBq 22Na+ for 1 h. Subsequently, the larvae were rinsed three times with 1 ml of high-salt buffer containing 112 mM of nonradioactive Na+, pH 7.7, and then placed into 500 μl of 1× FW and transferred into test tubes to analyze incorporated 22Na+ with a gamma counter (model ARC-300; Aloka). To examine the effects of ion transport inhibitors on Na+ influx, 10 larvae in a microcentrifuge tube were preincubated for 1.5 h in 1× FW in the presence of the inhibitors and then incubated for 1 h in 1× FW with the inhibitors and 7.4 kBq 22Na+. We note that because Na+ influx activity varied among larvae laid by different parents, we prepared experimental and control larvae from siblings in any series of experiments and calculated the activity relative to the control; comparisons across experiments were then made using the relative activities rather than absolute values of the incorporated Na+ amounts.
Morpholinos are chemically modified oligonucleotides that are more stable in, and less toxic to, living cells. Antisense morpholino oligonucleotides specific for a target gene are designed to complementarily anneal to the mRNA around the translation start site and knockdown gene function by repressing translation (40). Morpholino injection was performed as previously described (20). Two antisense morpholino oligonucleotides against foxi3a were synthesized (Gene Tools): FOXI3A-MO#1 (5′-ATGCTTTCTTCCCGTTTCTCTTTGT-3′) and FOXI3A-MO#2 (5′-GAGAAAATGCCTCCTGCTCTGCTGC-3′). Approximately 1 nl of 0.5 mg/ml FOXI3A-MO#1 or 2 mg/ml FOXI3A-MO#2 in 1× Danieau buffer [in mM: 58 NaCl, 0.7 KCl, 0.4 MgSO4, 0.6 Ca(NO3)2, 5 HEPES at pH 7.6] was injected into the yolk of one-cell embryos. The injected embryos were reared in 1× FW and analyzed as described above.
Analysis of Sodium Green intensity with imageJ.
Fluorescent intensities with Sodium Green were analyzed with imageJ software version 1.36 (http://rsb.info.nih.gov/ij/index.html). Prior to the analysis, single-focused images with Sodium Green were converted to grayscale but not modified at all. Within imageJ, a Sodium Green-positive MR cell was encircled completely with the peripheral area and the maximum (max) and the minimum (min) intensity values in the circle were measured. We considered the minimum value to be background in the circle and defined the MR cell intensity as (max − min). We analyzed 40 positive cells in the middle of each image or all cells if there were < 40 detectable Sodium Green-positive cells. We note that for any series of experiments, we prepared individual images under identical conditions, i.e., identical excitation light intensity, exposure time, and magnification.
Na+ accumulation occurs in a subset of the MR cells on the zebrafish larvae.
To examine Na+ metabolism in zebrafish larvae, we stained the body with Sodium Green, a Na+-dependent fluorescent reagent. We observed bright fluorescence in round-to-oval shaped cells that were sporadically distributed on the yolk sac surface (Fig. 1A). The fluorescence in these cells was severely reduced in the presence of 1 μM monensin, a Na+ ionophore (Fig. 2, A and B, and Table 1), but was much less affected in the presence of valinomycin, a K+ ionophore (Fig. 2, A and C, and Table 1). These results indicate that the fluorescence in these cells is sensitive to intracellular Na+. Sporadic staining also occurred on the developing gills at a later stage (Fig. 1B). The distribution of these cells resembled that of the larval MR cells that first differentiate on the body surface, especially the yolk sac and then on the gills coincident with their development. To examine whether these Na+ accumulating cells correspond to the MR cells, we used MitoTracker, a mitochondrial-staining reagent that has previously been shown to illuminate the MR cells (12, 14, 33, 38). In a double stain of Sodium Green with MitoTracker, we found that Na+ accumulation occurred only in a subset of the MR cells [Sodium Green+/MR cell = 37.6%, calculated from a total of 218 cells in four larvae (Fig. 1, C–E)].
Hwang and colleagues (33) have definitively classified the MR cells on the zebrafish larvae into two groups by gene expression: cells that are rich in Na+-K+-ATPase and those that are rich in vacuolar-type H+-ATPase. They have also shown that the H+-ATPase-rich MR cells exhibited a strong affinity for ConA on the apical membrane (15, 33, 34). Using this classification, we further characterized the Na+-accumulating MR cells. When we costained the larvae with Sodium Green and ConA, most of the signals colocalized, indicating that Na+ accumulated almost exclusively in the H+-ATPase-rich MR cells (Sodium Green+ and ConA+ = 83.8%, Sodium Green+ and ConA− = 6.6%, Sodium Green− and ConA+ = 9.6%, calculated from a total of 136 cells in four larvae, Fig. 3, A–C). We confirmed that the ConA-positive cells were rich in H+-ATPase rather than Na+-K+-ATPase by staining with ConA followed by immunohistochemistry with antibodies against H+-ATPase (Fig. 3, D–F) and Na+-K+-ATPase (Fig. 3, G–I). The ConA-positive cells exhibited an abundance of H+-ATPase but only moderate expression of Na+-K+-ATPase. On the other hand, Na+-K+-ATPase was more highly expressed in the ConA-negative cells, corresponding to the second population of MR cells described by Hwang and colleagues (33). Immunohistochemistry also revealed that H+-ATPase was localized to a subapical region in the MR cells, as reported previously (15, 33, 34).
Na+ absorption requires differentiation of the MR cells.
To examine further the role of H+-ATPase-rich MR cells in Na+ absorption from the environment, we focused on a transcription factor that is specifically expressed in the cells (Fig. 4, A–E). Solomon et al. (51) described that among FoxI class transcription factors, foxi3a/b are sporadically expressed on the body surface during embryogenesis and on the gills at later stages. They noted that the expression was in the mucous cells, but we have found that foxi3a is specifically expressed in H+-ATPase-rich MR cells (Fig. 4, C–E). To analyze foxi3a function in these cells, we depleted foxi3a gene product from the embryo by injection of foxi3a-specific morpholinos. In noninjected larvae (n = 5), the number of cells expressing H+-ATPase and stained with ConA was more than 70 (Fig. 4, G and I). However, in all of the morpholino-injected larvae (n = 5), the H+-ATPase-positive and ConA-positive cells were reduced to < 5 (Fig. 4, F and H). These results indicate that differentiation of H+-ATPase-rich MR cells was blocked by foxi3a depletion. In the foxi3a-depleted larvae, Na+ accumulation was not observed on the yolk surface (Fig. 4, J and K, and Table 1). Moreover, when we examined Na+ influx with radioisotope, it was severely reduced in the foxi3a-depleted larvae compared with the noninjected control sibling (Fig. 5). These results suggest that H+-ATPase-rich MR cells are the primary sites for Na+ absorption from the environment in zebrafish larvae.
Na+ accumulation is dependent on Na+ concentration in the environment.
We next examined whether Na+ accumulation in MR cells is dependent on the Na+ concentration in the environment. When zebrafish larvae were reared in 100-fold concentrated medium (100× FW), Na+ accumulation was observed only in H+-ATPase-rich MR cells, similar to the standard 1× FW condition (Fig. 6, A and B, and Table 1). In 20-fold diluted medium (1/20× FW), we still detected Na+ accumulation, although the intensity was decreased (Fig. 6C and Table 1). However, when we reared larvae in distilled water, we were not able to detect Na+ accumulation with the fluorescent reagent (Fig. 6D and Table 1). Thus, Na+ accumulation is dependent on the Na+ concentration in the environment.
NHE- and H+-ATPase-dependent Na+ accumulation.
To address which molecules are involved in Na+ accumulation, we examined the efficiency of accumulation against a variety of ion transport inhibitors. In this series of experiments, to sensitively detect the effect of each inhibitor, we first reared the zebrafish larvae in 20-fold diluted medium (1/20× FW) to reduce Na+ in MR cells, followed by incubation under standard conditions in the presence or the absence of inhibitors (see materials and methods). In the presence of EIPA, a selective inhibitor of NHE, we observed near total blockage of Na+ accumulation (Fig. 7, A and B, and Table 1). At a low concentration of amiloride, where ENaC but not NHE is inhibited, Na+ accumulated normally (Fig. 7C and Table 1). However, at a higher concentration of amiloride, one that also blocks NHE activity, Na+ accumulation was almost completely inhibited (Fig. 7D and Table 1). Bafilomycin A1, an inhibitor of H+-ATPase, also exhibited a strong inhibitory effect on Na+ accumulation in the MR cells (Fig. 7E and Table 1). In Mozambique tilapia (Oreochromis mossambicus), it is speculated that NKCC is involved in ion uptake in the MR cells (18). When we used metolazone as a selective inhibitor of NKCC, Na+ accumulation was not substantially reduced (Fig. 7F and Table 1). The presence of 500 μM ouabain, an inhibitor of Na+-K+-ATPase, did not strongly affect Na+ accumulation either (Fig. 7G and Table 1). These results suggest that the NHE and H+-ATPase play critical roles in Na+ absorption in the MR cells.
Involvement of CA in Na+ accumulation in the MR cells.
It has been proposed that CA, which catalyzes production of H+ and HCO3− from H2O and CO2 (36) is involved in Na+ uptake in adult zebrafish (3), as well as in rainbow trout (Salmo gairdneri) (23) and the small-spotted dogfish shark (Scylorhinus canicula) (42), since application of the selective inhibitors of CA blocked Na+ uptake in these fish. Here, we observed a severe reduction of Na+ accumulation in the MR cells on the zebrafish larvae in the presence of ethoxzolamide, a CA inhibitor (Fig. 7H and Table 1). In addition, we also found that an isoform of CA, ca2, was specifically expressed in the Na+-accumulating MR cells (Fig. 8). These results strongly suggest that CA2 is critically involved in Na+ absorption in the MR cells.
Blockage of Na+ influx with ion transport inhibitors.
Finally, we examined the effects of ion transport inhibitors on Na+ influx with the radioisotope 22Na+ (Fig. 9). As was observed with Sodium Green, EIPA and bafilomycin A1, but not metolazone, inhibited Na+ influx, although EIPA was less efficient compared with its effect on Na+ accumulation in the MR cells. The inhibitory effects of amiloride were similar to the effects of EIPA, even at low concentrations. Interestingly, ouabain inhibited Na+ influx even though it did not affect Sodium Green staining. Finally, ethoxzolamide did not have a significant inhibitory effect on Na+ influx. Thus, while the effects of some inhibitors on Na+ influx are consistent with their effects on Na+ accumulation as measured by Sodium Green staining, there are differences seen with other inhibitors.
Using Sodium Green, a Na+-dependent fluorescent reagent, we have succeeded in visualizing Na+ accumulation in a subset of the MR cells on the zebrafish larvae in vivo. In addition, we have demonstrated that these cells are rich in H+-ATPase rather than Na+-K+-ATPase. Although Sodium Green was described to show ∼40-fold greater selectivity for Na+ than K+ (13), it may form a fluorescent complex with K+ since the intracellular K+ concentration is much higher than that of Na+. This possibility, however, seems to be ruled out since 1) fluorescence was much more sensitive to a Na+ ionophore than a K+ ionophore (Fig. 2 and Table 1), and 2) the lack of fluorescence correlates well with the severe reduction of Na+ influx in larvae in which differentiation of MR cells was blocked (Figs. 4 and 5 and Table 1). Reduction of fluorescence by several Na+ transport inhibitors also supports the claim that Sodium Green fluorescence reflects Na+ accumulation in these cells (Fig. 6).
The correlation between Na+ accumulation in the MR cells and MR cell-dependent Na+ influx also indicates that accumulation reflects a Na+ absorption process through these cells. This hypothesis is further supported by our observations. First, Na+ accumulation was also observed in the gills, which are primary tissues for ion uptake in freshwater fish (Fig. 1B). Second, Na+ accumulation was dependent on ambient Na+ concentration (Fig. 6). Third, the H+-ATPase-rich MR cells, where Na+ accumulation occurred, have previously been shown to excrete H+ (33), consistent with models for Na+ uptake coupled to H+ excretion. Finally, we also confirmed that Na+ accumulation was not the result of apoptotic cells and that Sodium Green did not induce apoptosis (Supplemental Fig. 1; the online version of this article contains supplemental Fig. 1). Taken together, we conclude that Na+ absorption occurs in the H+-ATPase-rich MR cells on the zebrafish larvae.
The MR cells in which we observed Na+ accumulation were not only rich in H+-ATPase but also expressed Na+-K+-ATPase (Fig. 3). This observation is in contrast to the findings from Hwang and colleagues (33) in which Na+-K+-ATPase was not detectable in the H+-ATPase-rich MR cells. We assume that this difference reflects different staining conditions, including antibodies and detection systems. On the other hand, we observed a similar localization of H+-ATPase (33). Expression was very abundant near the apical surface, i.e., in a “subapical” region (as revealed by costaining with ConA), but also weakly distributed throughout the cells. As H+-ATPase is a transmembrane protein complex, the cytoplasmic distribution likely reflects the localization to tubular structures invaginated from the basolateral membrane.
Depletion of foxi3a from the zebrafish embryo caused a blockage of MR cell differentiation and severe reduction of Na+ influx (Figs. 4 and 5). These results indicate not only that the MR cells are the primary sites for Na+ influx, but also that foxi3a plays a critical role in MR cell differentiation. The foxi3a was first reported to be expressed in mucous cells (51), but here we demonstrate that it is specifically expressed in H+-ATPase-rich MR cells. Interestingly, foxi3a expression is initiated on the ventral side of the animal hemisphere at the late gastrula stage (Fig. 4B) (51), a region where the epidermis differentiates (24). Even at first appearance, the expression was already sporadic, suggesting that it is expressed in the MR cell precursors. Thus, we propose that foxi3a functions as a master control gene of MR cell differentiation.
Blockage by bafilomycin A1 of Na+ accumulation in the MR cells and Na+ influx supports the hypothesis that H+-ATPase is critically involved in Na+ absorption by creating an electrical gradient that provides the driving force (Figs. 7 and 9). EIPA, a selective inhibitor of NHE, also showed blockage of Na+ accumulation in the MR cells, suggesting that the protein is also involved in the Na+ absorption process (Fig. 7). However, the effect of the inhibitor on Na+ influx was not as strong as the effect on Na+ accumulation (Fig. 9). Likewise, ouabain, an inhibitor of Na+-K+-ATPase, had little effect on Na+ accumulation in the MR cells but obviously reduced Na+ influx (Figs. 7 and 9). We assume that the discrepancy between the results is likely because radioisotope influx and Sodium Green monitor different aspects of Na+ absorption. The former reflects total incorporated Na+ in larvae. The latter, however, should represent Na+ accumulated transiently in the MR cells during transport from the environment to the plasma and should reflect a limited portion of total Na+ incorporated in the body. We also anticipate that inhibition of Na+ uptake from the environment at the apical surface of MR cells gives different results from inhibition of Na+ export to the plasma at the basolateral membrane. For example, if ouabain blocked Na+ export to the plasma from the MR cells but not Na+ uptake at the apical surface, Sodium Green would reveal the Na+ absorbed into the MR cells even though the excess Na+ would soon leak out; however, in this case, total Na+ incorporated in the plasma would be significantly reduced. We should note that Grosell and colleagues (3) first examined the effects of ion transport inhibitors on Na+ (and Cl−) uptake in zebrafish and found that the effectiveness was different in fish acclimated to different ambient salinities. In this study, we have used younger larvae that were acclimated to different ambient salinities than those used in the Grosell report. With some of the inhibitors, our results are different than those reported by Grosell and colleagues (3), suggesting that zebrafish may use different Na+ (and Cl−) uptake mechanisms depending on the environmental salinity, as well as the developmental stage.
Further understanding of the mechanism of Na+ uptake through the MR cells will require identification of the genes expressed in these cells and subsequent functional analyses. In the present study, we have shown that ca2 as well as H+-ATPase and Na+-K+-ATPase are expressed in the MR cells. Currently, we have not identified MR cell-specific nhe genes. Although nhe2, nhe3a/b, and nhe5 are expressed in ∼55-hpf larvae, the expression level is too low to identify the cell-specific localization by whole mount in situ hybridization (K. Hoshijima, unpublished observations). The heterologous NHE antibodies we used did not recognize the zebrafish homologs. To identify MR cell-specific nhe genes, we would need to improve the sensitivity of in situ hybridization and/or develop NHE antibodies specific to the zebrafish proteins. In addition, preparation of transgenic zebrafish in which GFP is expressed under the control of nhe gene promoters could be a useful approach. To analyze gene function, gene-specific depletion in the zebrafish larvae would be powerful. As described here for foxi3a, depletion of candidate genes by antisense morpholino oligonucleotides and subsequent analyses with Sodium Green in combination with measurement of Na+ influx with the radioisotope 22Na+ could clarify their involvement in the Na+ uptake process.
Identification by this study of a Na+-absorbing site in H+-ATPase-rich MR cells simultaneously opens a new question as to the function of the other population of MR cells, i.e., the Na+-K+-ATPase-rich MR cells. Because this population of MR cells possess typical features of MR cells (see introduction), they likely absorb ambient ions, possibly major ions other than Na+, such as Ca2+ and Cl−. Indeed, Hwang and colleagues (41) have recently demonstrated the presence of the epithelial Ca2+ channel (ECaC), which is considered to be involved in Ca2+ uptake, in a subpopulation of the Na+-K+-ATPase-rich MR cells. The rest of the population may be involved in Cl− uptake (5). These possibilities may also be explored as described here, i.e., by immunohistochemical localization of key molecules and by their depletion with morpholinos, followed by analysis with fluorescent dye sensitive to target ions, and with radioisotopes. We also predict that foxi3b (51), a paralog of foxi3a, might play an important role in MR cell differentiation in combination with foxi3a.
This work was supported by Ministry of Education, Culture, Sport, Science and Technology of Japan (MEXT) Grants-in-Aid for Scientific Research 14104002, 17570003, and 18059010 and the 21st Century Center of Excellence Program of MEXT and National Institute of Genetics Cooperative Research Program Grant 2005-B3.
We thank Drs. Atsushi Kawakami and Makoto Kobayashi for providing zebrafish and advising on management of the zebrafish facility and Dr. Lisa Goering for copyediting the manuscript.
↵* M. Esaki and K. Hoshijima contributed equally to this work.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society