Fish

One hundred and eighty immature upper-mode Atlantic salmon, Salmo salar, parr (1 year old, >13 cm in length, mixed sex, first generation hatchery fish of the Irish Burrishoole River stock) were obtained in January 1997 from the Foslaks Hatchery (Randers, Denmark) where they had been hatched and reared in in-door tanks under simulated natural photoperiod and water temperature (minimum temperatures during winter 4°C). The fish were brought to the Odense University Campus and held in outdoor 500-l flow-through freshwater (FW) tanks supplied with Odense tap water (1.4 mM Cl⁻, 1.5 mM SO₄²⁻, 1.5 mM Na⁺, 0.16 mM K⁺, 3 mM Ca²⁺, 0.6 mM Mg²⁺, pH 8.3 and total CO₂ content of 5.5 mM). They were fed a 2% (body weight)⁻¹ diet of commercial trout pellets three times a week.

Sampling

Smolt development was assessed by regular 2–3 week interval samplings of fish from February through June (Table 1). On each date, two groups of fish were sampled: a group of 10 fish from the FW tank and a group of 10 fish which had been exposed to SW for 24 hr in order to assess SW-tolerance. Feeding was stopped four days prior to each sampling. Twenty-four hours prior to sampling, 10 randomly selected fish were transferred from the FW stock tank to a 35 ppt SW tank (400-l; 10°C; 12 hr light :12 hr dark cycle). The fish was stunned by a blow to the head. Blood was drawn from the caudal vessels into heparinized syringes, and the plasma was immediately separated by centrifugation at 5000×g for 3 min. The length and weight of the fish were measured. The fish was then decapitated, and additional sampling occurred as outlined below.

Table 1

Mean daily water temperature ± SD since the previous sampling, mean fork length ±SD of FW and SW-challenged salmon, mean body weight ± SD of FW salmon, and condition factor (100 × weight* length⁻³) ± SD of FW salmon.

From 4 FW-fish per group one 1st, two 3rd, and two 4th gill arches were pooled and immediately homogenized in 2.5 ml ice-cold denaturing solution for subsequent mRNA analyses (4 M guanidinium thiocyanate, 25 mM sodium citrate, 0.5% sarcosyl, 0.1 M β-mercaptoethanol, and 0.3% antifoam, pH 7.0).

From all FW- and SW-challenged fish one 2nd gill arch was dissected and placed in SEI-buffer for subsequent Na⁺,K⁺-ATPase activity analysis (300 mM sucrose, 20 mM EDTA, 50 mM imidazole, pH 7.3). All samples were immediately frozen in liquid nitrogen and stored at −80°C until analyzed. A piece of paraxial muscle was dissected and immediately weighed and dried to constant weight to determine total water content.

Cloning of gill ion transporters

The overall cloning procedure followed the description by Cutler et al. (1995b) as modified by Cutler et al. (1997). In short, total RNA was extracted from gills of FW-acclimated S. salar. First-strand cDNA synthesis from 5 μg total RNA was done using Superscript II reverse transcriptase (Gibco BRL, Gaithersburg, MD, USA) for 5 h at 45°C. Using degenerate primer pairs (Table 2), cDNA fragments of the gill Na⁺,K⁺-ATPase α and β subunits and the vacuolar-type H⁺-ATPase B subunit were amplified by Polymerase Chain Reaction (PCR). Forty PCR cycles were performed. Each PCR cycle consisted of 5 s denaturation at 92°C, 30 s annealing at 51–60°C, followed by 60 s primer-extension at 72°C. The primer-pairs used were designed as degenerate primers, the sequences of which were taken from two regions of amino acids which were identical between all published vertebrate sequences of the Na⁺,K⁺-ATPase α and β subunits, and the V-H⁺-ATPase B (56 kDa) subunit, respectively. All degenerate primers had inosine/cytosine wobbles incorporated at positions of nucleotide uncertainty as previously described by Cutler et al. (1995b). Positive fragments were purified from agarose gels with a Geneclean kit (Bio101, CA, USA), ligated into a pCR-Blunt II-TOPO vector using the Zero Blunt TOPO PCR Cloning kit (Invitrogen, CA, USA), and sequenced by a dideoxy chain termination method using the Big Dye Terminator sequencing kit (Perkin Elmer, CA, USA). The sequences of 3 clones from each individual fragment were compared using GeneJockey II software (Premier Biosoft Int., CA, USA) to give the precise sequences shown in Figs. 1 and 2. Comparison to known DNA sequences were performed using the BLAST algorithm ( http://www.ncbi.nlm.nih.gov/blast/).

Table 2

The three pairs of degenerate primers used during the PCR amplification of the Atlantic salmon Na⁺,K⁺-ATPase α subunit, Na⁺,K⁺-ATPase β subunit, and the vacuolar-type H⁺-ATPase B subunit cDNA fragments.

Fig. 1

Nucleotide and deduced amino acid sequences of the Atlantic salmon gill Na⁺,K⁺-ATPase α subunit (A) and Na⁺,K⁺-ATPase β subunit (B) cDNA fragments. The cDNA fragments were amplified by PCR using the degenerate primers provided in Table 2. Nucleotides are numbered from 1 on the left hand side, amino acids from 1 on the right hand side. The cDNA sequences have been submitted to the EMBL Data Library under the accession numbers: Na⁺,K⁺-ATPase α subunit ( AJ250809) and Na⁺,K⁺-ATPase β subunit ( AJ250810).

Fig. 2

Nucleotide and deduced amino acid sequence of the Atlantic salmon gill vacuolar-type H⁺-ATPase B subunit cDNA fragment. The cDNA fragment were amplified by PCR using the degenerate primers provided in Table 2. Nucleotides are numbered from 1 on the left hand side, amino acids from 1 on the right hand side. The cDNA sequence has been submitted to the EMBL Data Library under the accession number:  AJ250811.

Analyses

Plasma [Cl⁻] was measured by coulometric titration (Radiometer CMT 10, Copenhagen, Denmark). Specific Na⁺,K⁺-ATPase activity was analyzed at 25°C in crude homogenates of gill by the method of McCormick (1993) using a plate reader (Spectramax, Molecular Devices, Sunnyvale, CA, USA). Protein content was measured by the method of Lowry et al. (1951) modified for plate reader.

For Northern blotting, total RNA was isolated as described by Madsen et al. (1995). Total RNA (20 μg) from the gills of 4 FW-salmon from each of the 9 sampling dates were analyzed by formal-dehyde gel electrophoresis on the same gel and transferred by capillary blotting onto a nylon membrane (Zeta probe, Bio-Rad, Hercules, CA, USA). The Northern blot membrane was pre-hybridized for 4 hr at 47°C in 10 ml of pre-hybridization buffer containing 50% deionised formamide, 5×SSC (0.75 M NaCl, 75 mM Na₃citrate, pH 7.0), 1% SDS (sodium dodecyl sulphate), 5×Denhardt's (0.1% Ficoll, 0.1% polyvinyl pyrrolidone, 0.1% bovine serum albumin), 0.1% NaP_iP_i (tetrasodium pyrophosphate), 1 mM EDTA (ethylenediamine tetraacetic acid), 5 mg denatured calf thymus DNA, and 2 mg denatured yeast transfer RNA. Complementary DNA probes were radio-labelled (α³²P-dCTP) by random primer extension (Oligolabelling kit, Pharmacia, Uppsala, Sweden) and separated from the unincorpo-rated nucleotides on a G-50 micro column (ProbeQuant™, Pharmacia), denatured, and added to the prehybridization-buffer, and hybridized for 16 hr. Radioactivity was detected by phosphor imaging (Storm, Molecular Dynamics, Sunnyvale, CA, USA), and relative band intensities were analyzed by the ImageQuaNT 4.1 software (Molecular Dynamics). The membrane was stripped and reprobed using the specific cloned fragments amplified by PCR. To adjust for unequal loading, data are presented as the ratio of specific mRNA of interest to β -actin mRNA content. The approximate sizes of hybridization bands were evaluated by including a 0.24–9.5 kb RNA ladder (Gibco BRL).

Statistics

Statistical differences were analyzed using SYSTAT 5.03 (Systat, 1991, Evanston, IL, USA). When necessary, square root- or log-transformations of data were performed to meet the parametric ANOVA assumption of homogeneity of variances (evaluated by residual-plots). For each individual parameter, data were analyzed by one-way ANOVA. Differences among individual groups were analyzed by Tukey's Honestly Significant Difference Test. A significance level of α =0.05 was used.

Cloning and sequencing of gill ion transporter cDNA fragments

The PCR-amplifications on Atlantic salmon gill tissue cDNA using the degenerate primers shown in Table 2 resulted in single cDNA fragments with the predicted sizes: i) Na⁺,K⁺-ATPase α₁ subunit app. 890 base pairs (bp), ii) Na⁺,K⁺-ATPase β₁ subunit app. 233 bp, and iii) V-H⁺-ATPase B subunit app. 510 bp. Three clones of each cDNA fragment were completely sequenced on both strands and compared to obtain the Taqpolymerase error-free sequences provided in Figs. 1 and 2.

Size estimation of specific gill ion transporter mRNAs

The sizes of the prominent mRNAs of the three ion transporter subunits were: 3.8 kb (Na⁺,K⁺-ATPase α₁ subunit, Fig. 3A), 2.4 kb (Na⁺,K⁺-ATPase β₁ subunit, Fig. 3B), and 3.0 kb (V-H⁺-ATPase B subunit, Fig. 3C). The intensity of the hybridization signal of all mRNA species increased with increasing amount of total RNA loaded: 1, 5, or 20 μg total gill RNA.

Fig. 3

Northern Blots of 1, 5, and 20 μg of total RNA extracted from the gills of a FW-acclimated Atlantic salmon pre-smolt. The same blot was hybridized with a radio-labelled cloned cDNA fragment of the a) Na⁺,K⁺-ATPase α subunit, b) Na⁺,K⁺-ATPase β subunit, or c) the Vacuolar-type H⁺-ATPase B subunit (see Figs. 1 and 2). Autoradiographic films were exposed for 4 days at −80°C. Molecular size markers (kb) are indicated. The estimated sizes of the specific mRNA's were: a) 3.8 kb, b) 2.4 kb, and c) 3.0 kb, respectively.

Smolt development

The salmon showed clear signs of smoltification during the course of this study. There was a markedly improved salinity tolerance during the spring as judged by decreased deflection in plasma [Cl⁻] and muscle water content when challenged with SW from February through April (Fig. 4). Optimal performance in the SW-test was seen in May. The improved hypo-osmoregulatory ability coincided with a steady increase in gill Na⁺,K⁺-ATPase activity from February until May (Fig. 5). This was followed by an abrupt decrease in both enzyme activity and salinity tolerance in the salmon sampled in June. The salmon thus showed typical physiological smolt development in April-May.

Fig. 4

Deflections in plasma [Cl⁻] (A) and muscle water (B) in 24-hr 35 ppt SW-challenged Atlantic salmon during smoltification. Data are shown as mean difference ± SEM of 10 SW-challenged salmon to the mean of 10 FW-fish. Values with shared letters are not significantly different (P>0.05).

Fig. 5

Gill Na⁺,K⁺-ATPase activity in Atlantic salmon during smoltification. Data are shown as mean ± SEM of the pooled values of 10 FW- and 10 SW-challenged salmon. Values with shared letters are not significantly different (P>0.05).

Gill ion transporter mRNA levels

Gill Na⁺,K⁺-ATPase α₁ and β₁ subunit mRNA levels both increased during the spring reaching their maximum values by April (α₁ subunit 2.3-fold and β₁ subunit 1.8-fold the values in early February, Fig. 6A,B). In June, both mRNA species were back to the low levels observed in February. The α₁ and β subunit mRNA levels changed similarly during smoltification, as indicated by the stable ratio between levels of the two mRNA species (Fig. 6C; linear correlation analysis using individual paired values: R=0.92).

Fig. 6

Beta-actin normalized levels of gill Na⁺,K⁺-ATPase α subunit (A) and Na⁺,K⁺-ATPase β subunit (B) in Atlantic salmon during smoltification. In (C) the ratio of α :β subunit levels is shown. Data are shown as mean ± SEM of 4 FW-salmon. Values with shared letters are not significantly different (P>0.05). The Na⁺,K⁺-ATPase α subunit to Na⁺,K⁺-ATPase β subunit ratio did not change during smoltification (C, P>0.13).

Gill V-H⁺-ATPase B subunit mRNA levels changed significantly during smoltification (Fig. 7A). Transcript levels were high in the beginning of February followed by low levels in late February–early March. It then increased 2-fold to reach a plateau by late March–April. The level of B subunit mRNA then fell to low levels through May and June.

Fig. 7

Beta-actin normalized levels of gill vacuolar-type H⁺-ATPase B subunit mRNA (A) in Atlantic salmon through smoltification. In (B) the ratio of Na⁺,K⁺-ATPase α subunit to vacuolar-type H⁺-ATPase B subunit mRNA ratio is shown. Data are shown as mean ± SEM of 4 FW-salmon. Values with shared letters are not significantly different (P>0.05).

Gill Na⁺,K⁺-ATPase subunit mRNAs and V-H⁺-ATPase B subunit mRNA levels did not change in parallel through smoltification, i.e. the ratios varied (Fig. 7B; data for the Na⁺,K⁺ATPase β₁ subunit : V-H⁺-ATPase B subunit mRNA ratio not shown). The ratio of Na⁺,K⁺-ATPase α₁ subunit and β₁ sub-unit, respectively, to V-H⁺-ATPase B subunit mRNA levels increased through April and May. This was followed by a sharp decline to pre-smolt values in June.

Gill Na⁺,K⁺-ATPase α and β subunit cDNA fragments

The α Na⁺,K⁺-ATPase cDNA fragment cloned in this study is most likely an α₁ isoform based on the high identity of the deduced amino acid sequence with other cloned teleost α₁–subunits: 81% with the α₁ Na⁺,K⁺-ATPase of white sucker (Schönrock et al., 1991) and 85% with European eel (Cutler et al., 1995a). The size of the prominent α subunit mRNA species in the Atlantic salmon gill (3.8 kb) is similar to those reported earlier in gills of Atlantic salmon using either rainbow trout (D'Cotta et al., 1996, 2000) or Xenopus (Madsen et al., 1997) α subunit cDNA probes. Apart from a major 3.7 kb α subunit transcript, D'Cotta et al. (1996) also reported a minor transcript of 1.8 kb. We did not observe a similar size transcript. The existence of such a minor, possibly truncated form of the same α subunit expressed in salmon gills may, however, still be possible as the present α subunit isoform could differ from the one cloned by D'Cotta et al. (1996). Alternatively, the α subunits may be the same, and the truncated mRNA species was just not detectable using our cDNA probe. This is possible as the 670 bp cDNA fragment that D'Cotta et al. (1996) cloned, corresponds to the eel α Na⁺,K⁺-ATPase amino acids 615-826 (Cutler et al., 1995a) and thus is located further towards the 3′ end of the transcript than the present cDNA. The present 3.8 kb α transcript is also comparable to α subunit transcripts reported from other teleosts (e.g. white sucker: 3.8–4.15 kb, Schönrock et al., 1991; rainbow trout: 3.7 kb, Kisen et al., 1994; eel: 3.5 kb, Cutler et al., 1995a; brown trout: 3.8 kb, Madsen et al., 1995). Even though highly stringent hybridization conditions were used, there is a slight possibility that additional (similar sized) isoforms were detected by the present cDNA. In future studies it is crucial that multiple isoform expression is carefully investigated using isoform specific clones.

The cloned β Na⁺,K⁺-ATPase cDNA fragment is most likely a β₁ isoform as the deduced amino acid sequence shares 79% identity with European eel β₁ Na⁺,K⁺-ATPase sequence (Cutler et al., 1995b) and because the last 5 amino acids (Gly, Phe, Pro, Leu, Gln) in the deduced sequence are conserved in β₁ isoforms from different species but not in other β Na⁺,K⁺ATPase isoforms. Identities with β₁ isoforms of other vertebrate species are low though similar to that generally shared among β₁ orthologs (e.g. 43%: Torpedo californica: Noguchi et al., 1986; 48% Homo sapiens: Kawakami et al., 1986). The size of the prominent β₁ subunit mRNA species in the Atlantic salmon gill (2.4 kb) is similar to the single transcript sizes reported for both the β₁ isoform (eel gills: 2.35 kb; Cutler et al., 1995b) and for the brain-specific β₃ isoform (zebrafish and eel: 2.4 kb; Appel et al., 1996; Cutler et al., 1997).

Gill V-H⁺-ATPase 56 kDa B subunit cDNA fragment

The cloned V-H⁺-ATPase B subunit cDNA fragment is from the middle region of the coding sequence which is the most conserved between species and subunit isoforms (Nelson et al., 1992). The deduced 153 amino acid sequence is identical and at the nucleotide level has 97% identity with the recently cloned V-H⁺-ATPase B-subunit fragment from rainbow trout gill (Perry et al., 2000). The deduced amino acid sequence also shows a very high degree of identity with VH⁺-ATPase B subunits from other vertebrate species such as chicken (97%: Bartkiewicz et al., 1995), human and bovine B2 isoform (96%: Bernasconi et al., 1990; Nelson et al., 1992), and human and bovine B1 isoform (94%: Südhof et al., 1989; Nelson et al., 1992).

The size of the B subunit mRNA species detected in the gills of Atlantic salmon is 3.0 kb which is in good accordance with the 3.0-3.2 kb found for rainbow trout gill (Perry et al., 2000) and for both human and bovine B1 and B2 subunit transcripts (Südhof et al., 1989; Bernasconi et al., 1990; Nelson et al., 1992). Interestingly, Niederstätter and Pelster (2000) recently cloned both a V-H⁺-ATPase B1 and a B2 isoform from A. anguilla swimbladder gas gland which shared 97 and 98% identity, respectively, with the fragment cloned from S. Salar in the present study. The eel gas gland B1 transcript was 2.9 kb and the B2 transcript was 3.5 kb in size. In the present study, all the isolated plasmid clones had identical inserts, and we detected only one size mRNA species. Although the sequence identity suggests that the salmonid gill V-H⁺ATPase B-subunit is a B2 isoform (Perry et al., 2000; this study), Niederstätter and Pelster (2000) based on transcript size, concluded that the B-subunit isoform isolated from the gills of O. mykiss most likely is a B1 isoform. Furthermore, 2 poly A signals were shown to exist in the A. anguilla B2 gene although only the second poly A site seemed to be in use. In contrast, using low-stringency hybridization, Puopolo et al. (1992) detected 2 transcripts of the sizes 2.0 and 3.2 kb in bovine tissues using a cDNA fragment of the B2 subunit isoform. Similarly, two transcripts with sizes of 1.7 and 3.5 kb were observed in chicken tissues (Bartkiewicz et al., 1995). The high-stringency hybridization procedure used in the present study prevented the cross-hybridization of the specific probe to any possible isoform gene product of different size or differentially processed product of the same gene, suggesting that only one V-H⁺-ATPase B-subunit gene is expressed in the gills of S. salar.

Gill Na⁺,K⁺-ATPase expression during smoltification

The most characteristic biochemical and physiological signs of salmonid smoltification are the spring surge in gill Na⁺,K⁺-ATPase activity and the associated increase in SW-tolerance (cf. introduction). In the present study, the smolt climax was reached in April-May, after which the process of de-smoltification was abrubtly initiated (Figs. 4A,B and 5). The data suggest that the increased Na⁺,K⁺-ATPase activity (i.e. pump abundance when measured under the present conditions of V_max) at least in part is caused by the simultaneous increase of both α and β subunit transcripts in the gill through the course of smoltification (Fig. 6A, B). There is an overall picture that mRNA expression and enzyme activity change in parallel. However, there is one exception. The almost 3-fold increase in Na⁺,K⁺-ATPase activity in late February occurs without any detectable change in mRNA levels. Thus, there is not a consistent 1:1 relationship between changes in subunit mRNA expression and enzyme activity. Uncoupling between α subunit mRNA levels and overall Na⁺,K⁺-ATPase activity during smoltification is also evident in previous studies of smolting Atlantic salmon (D'Cotta et al., 1996, 2000) and brown trout (Nielsen et al., 1999). This is not surprising, since additional factors such as translational efficiency/rate, and post-translational processing and protein stability may affect functional enzyme assembly maturation or degradation. Further, there is a possibility that additional isoforms which are not detected by the Northern cDNA probes may contribute to enzymatic activity. The very abrupt decrease in enzyme mRNA and activity at the onset of de-smoltification is, however, characterized by a strong parallelism indicating that mRNA expression is one tool of regulating synthesis and abundance of protein. During FW- to SW-acclimation of euryhaline teleosts, similar increases to the ones seen during salmon smoltification have been observed at the level of both Na⁺,K⁺ATPase α (Cutler et al., 1995a; Kisen et al., 1994; Madsen et al., 1995) and β subunit (Cutler et al., 1995b) mRNA. Increased transcription of the subunit genes and/or stabilization of both subunit mRNAs thus appear to be a common part of the molecular mechanism of Na⁺,K⁺-ATPase mobilization during smoltification and FW- to SW-acclimation.

Nielsen et al. (1999) speculated that one mechanism of controlling functional enzyme synthesis during smoltification could occur at the level of differential synthesis of α and β subunits. The present study provides evidence that such regulation does not occur in Atlantic salmon at least at the level of mRNA expression, as these were regulated similarly during smoltification (Fig. 6C). Such highly coordinated synthesis seems to be the typical case in mammalian systems also (review by Geering, 1990), even though cases of non-coordinated changes in the expression ratio have also been documented (Lavoi et al., 1997).

Gill V-type H⁺-ATPase expression during smoltification

Interestingly, V-H⁺-ATPase B subunit mRNA levels were increased at intermediate stages of smoltification (March-April), concurrently with increased levels of α and β subunit mRNAs. However, at the peak of smoltification (in May), when both α and β subunit mRNAs, Na⁺,K⁺-ATPase activity, and SW-tolerance reached their maximal levels, there was a simultaneous drop in H⁺-ATPase B subunit expression and hence an increase in the Na⁺,K⁺-ATPase to V-H⁺-ATPase B-subunit mRNA ratio (Fig. 7A,B). There is increasing evidence that VH⁺-ATPase plays an important role for branchial Na⁺ uptake in FW-teleosts (cf. Introduction). Recent in vivo experiments using the specific V-H⁺-ATPase activity inhibitor Bafilomycin A (Fenwick et al., 1999) strongly suggests that the original model by Krogh (1938), including an apical entry of Na⁺ via Na⁺/H⁺ (NH₄⁺) should be replaced by the model proposed by Avella and Bornancin (1989) in which apical Na⁺ entry occurs through Na⁺ channels, driven by active pumping of protons from the epithelial cell cytosol into the surrounding medium. Since its first purification from vacuolar membranes of yeast (Kakinuma et al., 1981), the vacuolar-type H⁺-ATPase has been shown also to be the proton pump located in the apical membrane of many absorptive epithelia (e.g. mammalian kidney: Brown et al., 1988; osteoclast bone resorptive cells: Blair et al., 1989; frog skin: Klein et al., 1997) including the fish gill epithelium (Lin et al., 1994). In these tight epithelia, the V-H⁺ATPase is targeted to the apical plasma membrane (Nelson et al., 1992; Lin et al., 1994). Thus the present changes in B subunit mRNA levels are likely to translate into changes in expression of the B subunit protein, which makes up part of the V₁ cytosolic domain of the plasma membrane V-H⁺ATPase (review by Forgac, 1998). Mature proton pumps hence are inserted directly into the gill apical plasma membrane or in cytoplasmic vesicles which subsequently may fuse with the apical plasma membrane, as seen in catfish (Ictalurus nebulosus) experiencing a hypercapnic acidosis (Laurent et al., 1994).

The gradual build-up of gill Na⁺,K⁺-ATPase activity and hypo-osmoregulatory ability may be viewed as a preparation to encounter the marine environment. Thus, the developing smolt should be considered “a SW-fish residing in FW”, and there is evidence that this is indeed a functional implication of the physiological transformation. Concurrent with the development of SW-type chloride cell-accessory cell complexes, the gill epithelium gradually becomes more leaky, and the developing smolt enters a state of negative sodium balance (Primmett et al., 1988) with an associated decline in plasma ion levels (e.g. Houston, 1959; Madsen and Naamansen, 1989). This indicates that a gradual mal-adaptation to the FW environment may take place as a consequence of smolt development. Providing that the messenger is translated into functional protein, one plausible explanation for the present high expression of V-H⁺-ATPase during smoltification is, that it is needed in order to counteract excess salt loss while in FW. This hypothesis is in accordance with the elevated unidirectional branchial Na⁺ influx observed during smoltification in Atlantic salmon (Primmett et al., 1988). The H⁺-ATPase expression restores to pre-smolt values around the time where the peak smolt stage is reached. This is the time where the sea is normally encountered during migration, and in this way, the drop in H⁺-ATPase expression may be seen as the last in a series of preparative changes minimizing Na⁺ uptake and thus facilitating the “anticipated” SW-encounter.

Conclusion and perspectives

This study shows that the expression of gill Na⁺,K⁺ATPase α and β subunit genes are increased to a similar extent during S. salar smoltification, which at least partly causes the well-known build-up of gill Na⁺,K⁺-ATPase activity levels. The expression of the gill V-H⁺-ATPase B-subunit also changes during smoltification in a manner which suggests that the V-H⁺-ATPase enzyme may be an important mechanism to secure ion-uptake during smoltification in FW until the time where the V-H⁺-ATPase expression drops and the fish becomes a “fully prepared” smolt. Alternatively, changes in V-H⁺-ATPase may be associated with changes in acid-base regulation occurring during smoltification. It has been demonstrated that V-H⁺-ATPase subunits (mRNA and protein) are expressed in teleost osmo- and acid-base regulatory epithelia such as the gill (Lin et al., 1994; S et al., 1995; Perry et al., 2000; Wilson et al., 2000; this study), kidney (Perry and Fryer, 1997; Perry et al., 2000), and intestine (Perry et al., 2000; M. Seidelin, unpublished), but further studies are needed to elucidate subunit expression patterns and regulation as well as functional maturation of V-H⁺-ATPase enzyme at the cellular level in fish.