Changes in gill H⁺-ATPase and Na⁺/K⁺-ATPase expression and activity during freshwater acclimation of Atlantic salmon (Salmo salar)

Many species of salmonid fish exhibit an anadromous life history, with hatching and early development occurring in freshwater, followed by a migration to the sea where fish experience a significant growth phase before returning to freshwater to spawn. Iteroparous salmonid species (e.g. Salmo salar) are able to migrate back to the sea and may complete this anadromous cycle many times throughout their lives. During migration between freshwater and seawater, a salmon's gill must change between an ion-absorbing and an ion-secreting epithelium. Many studies have examined the physiology of salmonids during the process of smoltification, in which juvenile salmon make the transition from freshwater to seawater. These studies show an increase and then recovery in plasma sodium and chloride levels, coinciding with an increase in gill Na⁺/K⁺-ATPase activity and protein levels of Na⁺/K⁺-ATPase, cystic fibrosis transmembrane conductance regulator (CFTR) and Na⁺/K⁺/Cl^– cotransporter (NKCC) (Hoar, 1988). The observed changes in gill Na⁺/K⁺-ATPase expression involve switching between two distinct isoforms of this protein, as mRNA (Richards et al., 2003; Bystriansky et al., 2006) and protein (McCormick et al., 2009) expression of the α1b isoform increases during seawater acclimation, while expression of the α1a isoform decreases during the transition to seawater. In contrast, very few studies have examined the physiology of the gill during acclimation of salmonid fishes to freshwater.

Most of the few studies available on the seawater to freshwater transition have observed salmonids during their natural spawning migration into freshwater. During the natural migration of Arctic char (Bystriansky et al., 2007a) and sockeye salmon (Shrimpton et al., 2005) into freshwater, Na⁺/K⁺-ATPase ‘isoform switching’ is observed in the gill, but with α1a levels increasing and α1b levels decreasing. This observation supports the idea that α1a and α1b isoforms have specific roles for salmonids living in freshwater and marine environments, respectively. During natural freshwater migration, however, fish are exposed to a wide range of physiological challenges in addition to salinity change, including the cessation of feeding, exposure to changing environmental temperatures, constant and sometimes exhaustive exercise, and extensive changes to their reproductive physiology as they develop gametes and prepare to spawn. Added to this is the inherent metabolic cost associated with all of these processes and, for Oncorhynchus species, the unknown physiological basis of their inevitable mortality following spawning. As a result of this suite of naturally changing variables, a complex acclimation process, and the fact that most studies examining naturally migrating salmon cannot confirm how long the fish have been exposed to freshwater when they are sampled, reported osmoregulatory changes during the transition from seawater to freshwater have been inconsistent and difficult to interpret or explain. Our understanding of how salmonids acclimate to freshwater, a crucial moment in the anadromous salmonid life cycle, would thus be greatly enhanced by examining this process under controlled conditions. Therefore, in this study we examined the biochemical and molecular changes associated with freshwater acclimation under controlled conditions. We chose Atlantic salmon (Salmo salar) as a model species for this study as they are iteroparous and therefore should exhibit the capacity to fully acclimate to freshwater. Many stocks of Atlantic salmon return to freshwater months prior to spawning (Saunders, 1981) and may reside in this environment for some time before their next marine migration. This study is the first comprehensive examination of the physiology of a salmonid fish gill during freshwater acclimation under laboratory conditions.

Atlantic salmon (S. salar L.) were reared in seawater (32‰) for more than 1 year at the Centre for Aquaculture and Environmental Research (Department of Fisheries and Oceans Canada, West Vancouver, BC) and fed daily to satiation with a commercial salmon feed (Skretting Canada, Vancouver, BC, Canada). Salmon were approximately 2.5 years of age and had a mean length of 32.5±0.5 cm and a mean mass of 316.0±13.1 g. There were no statistical differences in salmon length or mass between each experimental and control group. Photoperiod and seawater temperature varied with season as tanks were located outdoors and supplied with ocean seawater. Salmon were fed to satiation daily with a commercial diet up until 1 day prior to the beginning of the experiment through to the 7 day sampling point. Feeding continued daily for the remainder of the experiment except for 1 day prior to the 14 and 30 day sampling times. The experiment began during May, when fish were transferred to either an identical seawater (control) tank or a tank containing a mixture of well water and Cypress Creek freshwater. Freshwater and seawater were maintained at a similar temperature throughout the experiment (8–10°C). After 4 and 8 h, and 1, 2, 4, 7, 14 and 30 days, eight fish were collected from each salinity (freshwater and seawater) and anaesthetized in a bath containing 100 mg l^–1 MS-222 buffered with 100 mg l^–1 sodium bicarbonate. After approximately 2 min, once fish had lost equilibrium and stopped swimming, they were removed from the bath, weighed and measured, and a 1 ml blood sample was taken by caudal puncture with a heparinized (500 U heparin ml^–1) syringe fitted with a 23 gauge needle. Fish were then killed by a blow to the head and subsequent severance of the spinal cord. Gill samples were quickly excised and frozen in liquid nitrogen. Blood samples were transferred to a microcentrifuge tube and centrifuged at 5000 g for 5 min. Plasma was then collected and frozen in liquid nitrogen. Plasma and gill samples were transferred to a –80°C freezer until analysed. All experimental procedures fully complied with Canadian Council of Animal Care guidelines (protocol no. A07-0288). Although control samples from fish remaining in seawater were taken at all eight time points in the experiment, each of the parameters measured was only determined for the 1, 4, 7 and 30 day control groups as a check to ensure that there were no substantial effects of holding conditions or season on the parameters measured.

Plasma osmolality was measured using a vapour pressure osmometer (Model 5520, Wescor, UT, USA). Plasma chloride levels were measured using a HBI digital chloridometer (Haake Buchler Instruments Inc., Sadlebrook, NJ, USA). Plasma sodium levels were determined using a flame atomic absorption spectrometer (Spectra AA; Varian, Mulgrave, VIC, Australia).

Gill tissue was scraped from filaments with a glass slide and homogenized on ice in SEID buffer (pH 7.5, 150 mmol l^–1 sucrose, 10 mmol l^–1 EDTA, 50 mmol l^–1 imidazole, 0.1% sodium deoxycholate) using a ground glass homogenizer. Homogenates were centrifuged for 1 min (4°C) at 5000 g to remove filaments and other insoluble material. The supernatant was used directly in the assay of gill Na⁺/K⁺-ATPase and H⁺-ATPase activity, which was determined at 10°C. Na⁺/K⁺-ATPase activity was measured by a method modified from Hu and Kaplan (Hu and Kaplan, 2000) by monitoring the difference in the amount of inorganic phosphate (P_i) liberated from gill homogenates in the presence and absence of the Na⁺/K⁺-ATPase-specific inhibitor ouabain (final concentration 1 mmol l^–1). A final concentration of 1 mmol l^–1 ouabain was previously determined to be optimal for full inhibition of Na⁺/K⁺-ATPase activity in Atlantic salmon gill tissue (Bystriansky et al., 2006). P_i liberated during ATP hydrolysis was determined by the method of Brotherus et al. (Brotherus et al., 1981). Homogenate protein concentrations were determined using the method of Bradford (Bradford, 1976). Gill H⁺-ATPase activity was measured as the bafilomycin-sensitive ATPase activity in the same manner as used for Na⁺/K⁺-ATPase, but through monitoring the P_i liberated from gill homogenates in the presence and absence of the inhibitor bafilomycin (final concentration 30 μmol l^–1). A final concentration of 30 μmol l^–1 bafilomycin was determined to be optimal for full inhibition of H⁺-ATPase activity prior to analysis of samples by testing a wide range of concentrations (0.01 μmol l^–1 to 10 mmol l^–1) of bafilomycin on 10 representative gill samples. The presence of EGTA and sodium azide in the assay mix was not found to alter measured H⁺-ATPase levels. All samples were run in duplicate.

Total RNA was extracted from gill samples using TriZol isolation reagent (Invitrogen, Carlsbad, CA, USA) using the guanidine thiocyanate method (Chomczynski and Sacchi, 1987). Isolated total RNA was quantified spectrophotometrically and run (2 μg) on an agarose gel (1%) to check for RNA quality and integrity. First strand cDNA was synthesized from 2 μg of total RNA using a high capacity cDNA reverse transcription kit (Applied Biosystems, Foster City, CA, USA). Quantitative RT-PCR (qRT-PCR) was performed on an ABI Prism 7000 sequence analysis system (Applied Biosystems). PCR reactions contained 1 μl of cDNA, 150 pmol of each primer and Universal SYBR green master mix (Applied Biosystems). Forward and reverse primers used for each gene were isoform specific and tested to ensure that they amplified only a single target gene. Primer sequences used were as follows: elongation factor 1α (EF1α) forward 5′ GAG ACC CAT TGA AAA GTT CGA GAA G 3′, EF1α reverse 5′ GCA CCC AGG CAT ACT TGA AAG 3′; Na⁺/K⁺-ATPase α1a forward 5′ GGC CGG CGA GTC CAA T 3′, Na⁺/K⁺-ATPase α1a reverse 5′ GAG CAG CTG TCC AGG ATC CT 3′; Na⁺/K⁺-ATPase β1b forward 5′ CTG CTA CAT CTC AAC CAA CAA CAT T 3′, Na⁺/K⁺-ATPase α1b reverse 5′ CAC CAT CAC AGT GTT CAT TGG AT 3′; Na⁺/K⁺-ATPase α1c, α3 and β1 primers were developed by Stefansson et al. (Stefansson et al., 2007); FXYD 5, 6, 9 and 11 primers were developed by Tipsmark (Tipsmark, 2008); CFTR I and II, and NKCC primers were developed by Nilsen et al. (Nilsen et al., 2007); V-H⁺-ATPase β-subunit primers were developed by Kiilerich et al. (Kiilerich et al., 2007). The primers used to determine Na⁺/K⁺-ATPase α1a expression were compared with those used by Madsen et al. (Madsen et al., 2009) and were found to yield statistically identical results (data not shown). qRT-PCR reaction conditions and analysis were conducted as described previously (Scott et al., 2004a). Negative control reactions for qRT-PCR were performed with original total RNA from several representative samples to determine potential genomic DNA contamination. For all genes monitored, genomic contamination was found to be negligible, for example consisting of a maximum of 1:1648 starting copies for the Na⁺/K⁺-ATPase isoform α1a gene. The relative quantity of each gene's mRNA in gill samples was normalized to an endogenous reference (EF1α) and expressed relative to the mean value for the Atlantic salmon acclimated to seawater (control) for 1 day.

Chemicals not identified above were purchased from Sigma Chemical Co. (Sigma-Aldrich Canada Ltd, Oakville, ON, Canada) and were of the highest available purity.

All data are presented as means ± s.e.m. of a sample size of eight individuals. Comparisons of plasma sodium and chloride levels and total osmolality, gill Na⁺/K⁺-ATPase and H⁺-ATPase activities, and gill mRNA levels for each of the genes tested were performed using analysis of variance (ANOVA). For most parameters examined, the control groups (1, 4, 7 and 30 day seawater groups) did not differ statistically. In these cases, all control data were pooled and a Dunnett's multiple comparison test was performed to compare the pooled control sample with each experimental group. In two cases the control groups changed over time (mRNA levels of H⁺-ATPase and CFTR II) so control data were not pooled, and all control and experimental groups were compared first using a one-way ANOVA followed by a Holm–Sidak test to make pair-wise comparisons. For all comparisons, P<0.05 was considered significant.

All salmon survived their exposure to freshwater and appeared to be in good health throughout the experiment.

Atlantic salmon held in seawater had stable plasma osmolality and sodium and chloride levels throughout the experiment. Following freshwater exposure, each of these plasma parameters decreased significantly. Plasma osmolality decreased ∼5.4% within 4 h of freshwater exposure (Fig. 1A) and then slowly increased over the next several days until it was not significantly different from control levels on day 14 of the experiment. Similarly, plasma sodium decreased significantly following freshwater exposure, but did not reach a minimum until day 4, when levels were ∼10.3% below control levels (Fig. 1B). Plasma chloride decreased significantly, reaching a minimum 4 days after exposure to freshwater, and did not fully recover to control levels (Fig. 1C).

Atlantic salmon gill H⁺-ATPase activity increased gradually over the first 14 days of freshwater exposure to a peak activity of ∼0.99±0.09 μmol P_i mg^–1 protein h^–1, an increase of more than 48% over that of control fish (Fig. 2A), and remained near peak levels through to the end of the experiment. Interestingly, expression of gill H⁺-ATPase mRNA did not change as a result of the change in salinity (Fig. 2B). Over the course of the experiment, mRNA levels were stable, until the 30 day mark when H⁺-ATPase mRNA levels of both freshwater- and seawater-acclimated individuals significantly increased compared with those of individuals sampled at previous time points.

Atlantic salmon maintained in seawater had a gill Na⁺/K⁺-ATPase activity of ∼1.42±0.06 μmol P_i mg^–1 protein h^–1, which was stable throughout the experiment (Fig. 3A). Four days following freshwater exposure, gill Na⁺/K⁺-ATPase activity significantly increased by ∼43% compared with control fish, and remained high through the 7 and 14 day sampling points. After 30 days of freshwater acclimation, gill Na⁺/K⁺-ATPase activity returned to control levels.

Expression of gill Na⁺/K⁺-ATPase isoform α1a increased significantly, by more than 7-fold, during freshwater acclimation, with levels peaking after 14 days (Fig. 3B). Conversely, gill α1b isoform expression decreased significantly in fish within 4 h of freshwater exposure (Fig. 3C) and levels remained depressed through to the 14 day sampling point. By the end of the experiment (30 days), gill α1b expression had returned to levels similar to those of control salmon. In contrast to the α1a and α1b isoforms, the α1c and α3 isoforms were unchanged by freshwater exposure (Table 1).

mRNA levels of the gill Na⁺/K⁺-ATPase β1-subunit also changed with freshwater acclimation (Fig. 4A). Expression of the β1-subunit mRNA increased transiently following freshwater acclimation, reaching a peak of ∼57% higher than control levels after 7 days. The expression pattern for FXYD protein 11 was very similar to that seen for the Na⁺/K⁺-ATPase β-subunit, with levels transiently higher (∼53%) in freshwater-exposed fish at 7 and 14 days (Fig. 4B). In contrast to FXYD 11, expression of FXYD proteins 5, 6 and 9 was not affected by the salinity change (Table 1).

The mRNA expression of gill NKCC was relatively stable, aside from a significant, but transient, increase in levels for the 14 day freshwater-acclimated group compared with control fish (Table 1). CFTR I mRNA levels significantly decreased following 2 and 4 days of freshwater exposure. However, after 7 days in freshwater there was no difference between control and experimental groups, and after 14 days, CFTR I levels were significantly higher in freshwater-exposed salmon than in control fish, a trend that remained significant through to the end of the experiment (Table 1). When only control groups were compared, mRNA expression of CFTR II decreased significantly from levels found 1 day into the experiment, reaching a minimum level on day 4 before increasing steadily throughout the remainder of the experiment. However, when compared with the 1 day group, control CFTR II levels remained significantly lower even after 30 days (Table 1).

Atlantic salmon held in seawater had stable plasma osmolality and sodium and chloride levels throughout the experiment, suggesting these fish were fully acclimated to seawater. The deviation in plasma osmolality (∼5.4%) from control levels seen in this study during freshwater acclimation was much less than that typically observed during seawater acclimation of Atlantic salmon. Several studies have shown that an increase of ∼10–20% is normally seen following seawater exposure (D'Cotta et al., 2000; Singer et al., 2002; Bystriansky et al., 2006). This difference suggests that either osmoregulation is less of a challenge in freshwater or the mechanisms needed to osmoregulate in freshwater can be brought online more quickly during this transition. Within 7 days of freshwater exposure, plasma sodium had returned to control levels. In the freshwater salmonid gill ion uptake model (Avella and Bornancin, 1989), the role of the Na⁺/K⁺-ATPase is to pump sodium across the basolateral membrane into the blood, thereby maintaining low intracellular sodium levels, which aids the inward movement of sodium into the cell across its apical surface. The gradient needed for sodium uptake is also linked to the action of the apical vacuolar proton ATPase (V-H⁺-ATPase). In trout, several studies have localized the V-H⁺-ATPase to both pavement cells and mitochondria rich cells, with the majority of protein located on the apical membrane (Lin et al., 1994; Wilson et al., 2000). Despite the importance of maintaining a low intracellular sodium concentration for the system to work, most researchers regard the apical H⁺-ATPase as the primary engine that drives Na⁺ uptake, and little attention has been paid to the importance of Na⁺/K⁺-ATPase in this model. In fact, recent reviews on fish osmoregulation state the role of gill Na⁺/K⁺-ATPase in the active ion uptake model as ‘uncertain’ (Perry, 1997) or ‘unclear’ (Evans et al., 2005). In contrast, Kirschner (Kirschner, 2004) suggested that the energy required for sodium uptake can be generated solely by Na⁺/K⁺-ATPase if freshwater sodium concentrations are in the 1 mmol l^–1 range. In freshwater, with its lower Na⁺ concentration (μmol l^–1 range), the combined action of both the Na⁺/K⁺-ATPase and H⁺-ATPase would be necessary to drive sodium uptake (reviewed by Parks et al., 2008). The relative roles of these two ion pumps in generating the energy needed to take up sodium against its concentration gradient are thus poorly understood and are rarely discussed. The results from the current study may help to shed some light on the relative roles of the H⁺-ATPase and Na⁺/K⁺-ATPase in the salmonid gill during freshwater acclimation.

Interestingly, to our knowledge, no previous study has examined changes in gill H⁺-ATPase activity following transfer of seawater-acclimated salmonids into freshwater. We found that Atlantic salmon gill H⁺-ATPase activity increased by ∼48% during acclimation to freshwater; however, a significant increase in activity did not occur until day 14 of exposure (Fig. 2). Considering the fact that plasma sodium had returned to control levels within 7 days of freshwater exposure (Fig. 1), the importance of the observed increase in gill H⁺-ATPase activity is difficult to justify. As marine fish express H⁺-ATPase in their gills (mainly for acid secretion) (Evans et al., 2005), basal levels of this protein found in seawater-acclimated Atlantic salmon gill may be sufficient to drive Na⁺ uptake following freshwater exposure. As individual gill cell types are known to vary in their H⁺-ATPase activity (Hawkings et al., 2004), it is possible that measuring activity in whole gill homogenates may have masked significant increases in individual cell types responsible for the bulk of sodium flux (i.e. PNA^– mitochondria rich cells). Other studies have also shown that changes in gill H⁺-ATPase appear to occur quite slowly in response to salinity change in salmonids. For example, Wilson and colleagues (Wilson et al., 2002) saw no change in H⁺-ATPase activity after 8 days of seawater acclimation, while Hawkings and colleagues (Hawkings et al., 2004) and Lin and Randall (Lin and Randall, 1993) found decreased activity in rainbow trout sampled after 2 weeks and 8–10 weeks of seawater exposure, respectively.

We also found that H⁺-ATPase mRNA levels did not change in Atlantic salmon as a result of freshwater exposure. This suggests that the observed increase in H⁺-ATPase activity was not the result of increased transcription. Increased H⁺-ATPase activity may be due to changes at the translational or post-translational level as certain modifiers are known to alter H⁺-ATPase activity (e.g. cAMP, protein kinase A) (O'Donnell et al., 1996; Voss et al., 2007). In addition, vesicular trafficking of endosomal H⁺-ATPase stores to the plasma membrane has been described in several epithelia including frog skin (Lacoste et al., 1993), turtle urinary bladder (Cannon et al., 1985) and kidney collecting tubule (Schwartz and Al Awqati, 1985). However, as such post-translational modifications and mobilization of intracellular stores to the plasma membrane surface normally occur over a much shorter time scale (hours), they do not explain the long lag time before a significant increase in activity is seen following freshwater exposure. Taking all this into consideration, we must also bear in mind that the slow induction of H⁺-ATPase following freshwater exposure may suggest that this is not the principle mechanism involved in driving sodium uptake during the initial stages of freshwater acclimation.

The 43% increase in gill Na⁺/K⁺-ATPase activity seen here following freshwater exposure is quite limited compared with the typical 200–400% increase in activity seen during the first month of seawater acclimation for Atlantic salmon (Singer et al., 2002; Bystriansky et al., 2006). However, gill Na⁺/K⁺-ATPase activity increased within 4 days of freshwater exposure, and remained high through to 14 days, corresponding to the time when plasma sodium levels returned to basal levels. Once plasma sodium levels were restored, Na⁺/K⁺-ATPase activity subsided with a concomitant increase in gill H⁺-ATPase activity. Although the percentage increase in H⁺-ATPase activity (48%) was larger than the increase in Na⁺/K⁺-ATPase activity (43%), the absolute increase in activity was not (0.32 vs 0.61 μmol P_i mg^–1 protein h^–1). The earlier and more robust increase in Na⁺/K⁺-ATPase activity is interesting and may support a role for Na⁺/K⁺-ATPase in at least partially powering sodium uptake in the early stages of freshwater acclimation.

Our observation that Atlantic salmon increase gill Na⁺/K⁺-ATPase activity following exposure to freshwater is somewhat surprising as the consensus among most researchers is that gill Na⁺/K⁺-ATPase activity is higher in seawater-acclimated salmonids, and decreases following migration into freshwater. There are dozens of studies showing that gill Na⁺/K⁺-ATPase activity increases during seawater acclimation of salmonids, so this is not an area of contention (reviewed by Folmar and Dickoff, 1980; McCormick and Saunders, 1987). However, there are significantly fewer studies that examine gill Na⁺/K⁺-ATPase regulation during freshwater acclimation of salmonids. Wild Arctic char migrating into freshwater have been reported to have significantly higher gill Na⁺/K⁺-ATPase activity immediately following freshwater entry (Bystriansky et al., 2007a). Shrimpton and colleagues showed that gill Na⁺/K⁺-ATPase activity decreased in 3 of 4 stocks of sockeye salmon monitored while still in seawater en route to enter freshwater (Shrimpton et al., 2005). However, gill Na⁺/K⁺-ATPase activity in these stocks then increased when fish were sampled at the spawning grounds. The regulation of gill Na⁺/K⁺-ATPase during the spawning migration appears to be quite complicated, and may be influenced by the many changes occurring over a range of physiological systems. After all, Na⁺/K⁺-ATPase has many physiological roles, and observed changes are not necessarily associated with whole body ion regulation. To our knowledge, the only other study to examine changes in gill Na⁺/K⁺-ATPase activity during transfer from seawater to freshwater under controlled conditions was performed by Tipsmark and colleagues, who transferred brown trout to seawater for 60 days, then transferred them back to freshwater for 10 days (Tipsmark et al., 2002). One day after transfer back to freshwater, no change in gill Na⁺/K⁺-ATPase was seen. However, after 10 days in freshwater, the activity of the enzyme was significantly lower, but still more than 4-fold higher than that observed in the original freshwater group. Taking into account data from non-salmonid fish, an increase in gill Na⁺/K⁺-ATPase activity following acclimation to freshwater has been widely observed in species including sea bass [Dicentrarchus labrax (Lassere, 1971; Jensen et al., 1998)], killifish [Fundulus heteroclitus (Scott et al., 2004a; Scott et al., 2004b)], pupfish [Cyprinodon salinus (Stuenkel and Hillyard, 1980)], mullet [Mugil cephalis (Ciccotti et al., 1994) and Chelon labrosus (Gallis et al., 1979)], flounder [Platichthys flesus (Stagg and Shuttleworth, 1982) and Gillichthys mirabilis (Doneen, 1981)], milkfish [Chanos chanos (Lin et al., 2003)], sea bream [Mylio macrocephalus (Kelly et al., 1999)] and striped bass [Morone saxatilis (Tipsmark et al., 2004)]. Many of these species also increase gill Na⁺/K⁺-ATPase activity when transferred to higher salinities. This phenomenon has been described by Lasserre (Lasserre, 1971) as a salinity-dependent ‘U-shaped’ regulation of gill Na⁺/K⁺-ATPase activity.

Another surprising result of the current study is the observation that mRNA levels for NKCC and CFTR I did not decrease as a result of freshwater acclimation, and in fact these levels increased, albeit transiently. NKCC and CFTR are thought to be central to the ion excretion model of marine fish gills (Silva et al., 1977; Foskett and Scheffey, 1982), and as such their expression might be expected to decrease following entry into freshwater. Our data do not support this hypothesis. Interestingly, NKCC has been suggested to play an absorptive role in freshwater-acclimated killifish gills as it is found in an apical position (Katoh et al., 2008). In contrast, gill NKCC is found in a basolateral position in seawater-acclimated killifish and in both freshwater- and seawater-acclimated trout where it is involved in salt secretion (Katoh et al., 2008). In addition, an apically localized Na⁺/Cl^– cotransporter (NCC) has been identified in freshwater tilapia (but not seawater tilapia) (Hiroi et al., 2008), suggesting that this family of proteins can play a role in ion uptake or secretion depending on their tissue localization. Finally, as Atlantic salmon are iteroparous and have the capacity to return to the ocean at any time, it is conceivable that levels of gill NKCC and CFTR do not decrease during freshwater entry to allow salmon to quickly acclimate to seawater during their next ocean migration

The results of the present study are consistent with previous reports of gill Na⁺/K⁺-ATPase isoform switching in salmonid fish in response to salinity change (Richards et al., 2003; Shrimpton et al., 2005; Bystriansky et al., 2006; Bystriansky et al., 2007a; Bystriansky et al., 2007b), and support the idea that the Na⁺/K⁺-ATPase α1a and α1b isoforms each have a specific function for fish living in freshwater and marine environments, respectively. However, the precise roles and functions of the various Na⁺/K⁺-ATPase isoforms in salmon gills remain to be elucidated. Overall, the available data so far do not support a simple model in which the Na⁺/K⁺-ATPase α1a isoform is important during Na⁺ uptake from freshwater, while the Na⁺/K⁺-ATPase α1b isoform is important during Na⁺ secretion in seawater. For example, in the current study we found that although Na⁺/K⁺-ATPase α1b levels initially decreased in response to freshwater exposure, they returned to levels similar to those in seawater-acclimated Atlantic salmon by 30 days.

A similar phenomenon occurs during the acclimation of rainbow trout to seawater, where α1b expression quickly increases and then subsides to control levels after approximately 15 days (Richards et al., 2003). Similarly, Shrimpton and colleagues reported that Na⁺/K⁺-ATPase α1b expression generally declines in sockeye salmon returning to freshwater, but then increases again during spawning (Shrimpton et al., 2005).

The functional Na⁺/K⁺-ATPase consists of two protein subunits, a catalytic α-subunit and a glycoprotein β-subunit. These two subunits assemble in a 1:1 ratio creating an αβ heterodimer. The Na⁺/K⁺-ATPase β-subunit is thus required for proper function of the Na⁺/K⁺-ATPase, playing a role in maturation, delivery and insertion of the αβ heterodimer into the plasma membrane (Bystriansky and Kaplan, 2007), and potentially influencing the catalytic activity of the enzyme (Geering, 2005). As increased Na⁺/K⁺-ATPase protein abundance requires both an α- and a β-subunit, any increase in α-subunit protein would require an equivalent increase in β-subunit synthesis. Thus, it is not surprising that we also observed increased levels of β1-subunit mRNA in Atlantic salmon gill following freshwater exposure. To our knowledge, gill expression of the β-subunit has not previously been examined in response to salinity change in salmonids; however, its expression has been shown to increase during smoltification of Atlantic salmon (while fish were still in freshwater) in a pattern similar to that observed for the α-subunit isoform α1b (Nilsen et al., 2007; Stefansson et al., 2007). As β1-subunit mRNA is increased both in freshwater acclimation and during smoltification (and presumably upon acclimation to seawater), it seems that the same β1-subunit may complex with both α1a and α1b isoforms of the α-subunit. Na⁺/K⁺-ATPase αβ subunit complexing in fishes may be important for the successful up-regulation of Na⁺K⁺-ATPase activity during acclimation to changing salinity and warrants further examination.

FXYD proteins are a family of small single membrane-spanning proteins that are known to interact with Na⁺/K⁺-ATPase and modulate its kinetic properties (Garty and Karlish, 2006). Tipsmark recently identified eight FXYD protein isoforms in various tissues of Atlantic salmon (Tipsmark, 2008); mRNAs for five of these were expressed at appreciable levels in gill. Although Tipsmark saw limited differences in gill expression between fully acclimated Atlantic salmon in freshwater and seawater (Tipsmark, 2008), we chose to examine these gill FXYDs over a time course during freshwater acclimation to determine whether transient changes in the expression of these genes might occur. Like Tipsmark (Tipsmark, 2008), we also saw no change in the expression of FXYD 5, 6 and 9. However, expression of FXYD 11 changed significantly following freshwater exposure. Tipsmark described FXYD 11 as a ‘teleost isoform’ as it contains the characteristic FXYD domain but is otherwise quite divergent from the remaining members of the gene family (Tipsmark, 2008). Interestingly, this gene is expressed almost exclusively in gill, where it is also the most abundant of the FXYD isoforms (by mRNA abundance) (Tipsmark, 2008). The time course of changes in FXYD 11 mRNA expression and in Na⁺/K⁺-ATPase β1-subunit mRNA expression showed strong similarity, increasing significantly 7 days following freshwater exposure before subsiding back to control levels after 30 days. Consistent with the observations of Tipsmark (Tipsmark, 2008), expression of FXYD 11 mRNA was not different between fully acclimated freshwater and seawater individuals. The specific influence of FXYD 11 on the kinetic properties of Na⁺/K⁺-ATPase is not known but is clearly another very important avenue for study. The similarity in expression patterns of β1 and FXYD 11 suggests they may complex with α-subunit isoform α1a to form the predominant Na⁺/K⁺-ATPase species in gills of freshwater-acclimated salmonids.

The results of this study add evidence to the idea that the α1a and α1b gill Na⁺/K⁺-ATPase α-subunit isoforms play different roles in salmonids acclimated to freshwater and marine environments. In this study, Atlantic salmon acclimating to freshwater showed the opposite pattern of gill Na⁺/K⁺-ATPase ‘isoform switching’ to what we previously reported during seawater acclimation of rainbow trout, Atlantic salmon and Arctic char (Richards et al., 2003; Bystriansky et al., 2006). Our observation that gill Na⁺/K⁺-ATPase activity increases during freshwater acclimation seems to agree with observations made in studies on most non-Salmonid fishes where Na⁺/K⁺-ATPase activity is seen to increase during acclimation to both higher and lower salinity (Lassere, 1971). This suggests that gill Na⁺/K⁺-ATPase may play a more important role in the active ion uptake model for salmonid fish than was previously appreciated. The exact role of gill Na⁺/K⁺-ATPase and the mechanisms that regulate its expression and activity in this process are still unclear and deserve further examination.

Fish were a kind gift of Dr Dave Higgs and Dr Terri Sutherland at the Centre for Aquaculture and Environmental Research, Department of Fisheries and Oceans Canada. We also thank Bob Devlin, Janice Oakes, Tanya Hollo, Rush Dhillon, Lisa Skinner, Tracy Pollack, Mike Jones, Rob Dominelli, Wendy Tymchuk and Carlo Biagi for their assistance with sampling and maintaining the Atlantic salmon used in this study.