Zonation of Ca²⁺ transport and enzyme activity in the caeca of rainbow trout – a simple structure with complex functions

Blind-ended digitations known as caeca are associated with the gastrointestinal tract of slightly more than half of known fish species (Hossain and Dutta, 1996). These projections are composed of four tissue layers: the predominant mucosa (the epithelia; Anderson, 1986), a reduced submucosa (the lamina propria and stratum compactum; Williams and Nickol, 1989), muscularis (the longitudinal and circular muscle) and serosa (the connective tissue). Moreover, the mucosa contains several cell types: goblet cells responsible for mucous secretion, absorptive columnar cells (Anderson, 1986; Williams and Nickol, 1989; Abaurrea-Equisoaín and Ostos-Garrido, 1996), secretory cells (Anderson, 1986; Williams and Nickol, 1989) and immune cells (leucocytes, macrophages; Williams and Nickol, 1989). The tissue of the caeca and the intestine (from where they originate) is generally similar (e.g. Burnstock, 1959; Kapoor et al., 1976; Buddington and Diamond, 1987; Chen et al., 2006; Faulk et al., 2007; Zizza and Desantis, 2011). However, the stratum compactum and muscularis are thinner (∼30 µm) compared with the intestine proper (∼45 µm; Anderson, 1986; Williams and Nickol, 1989). In addition, the number of goblet cells in the caeca can be dissimilar from the intestine proper (Williams and Nickol, 1989; Hossain and Dutta, 1996) depending on species. Heterogeneity of dispersion, from proximal base to distal tip, can also be present for both goblet cells (Jansson and Olsson, 1960) and absorptive cells (Anderson, 1986). However, when comparing anterior with posterior caeca, a single investigation failed to reveal structural zonation (Williams and Nickol, 1989).

Caeca structure has informed hypotheses about their function, which have included digestion and nutrient absorption (e.g. Jansson and Olsson, 1960; Bauermeister et al., 1979; Sire et al., 1981), as well as fermentation and storage (Buddington and Diamond, 1987). Other suggested roles include ion and water absorption (e.g. Bogé et al., 1988; Grosell et al., 2009). In fact, cells in the caeca of rainbow trout have lamellar structures, which appear to be in-foldings of the plasma membrane, associated with mitochondria as observed in other ion-transporting cells (Ezeasor and Stokoe, 1981; Giffard-Mena et al., 2006). As such, physiological and morphological studies have associated the caeca with ion and water transport, particularly during salinity acclimation [i.e. freshwater (FW) versus seawater (SW); e.g. Bogé et al., 1988; Abaurrea-Equisoaín and Ostos-Garrido, 1996; Veillette et al., 2005; Grosell et al., 2009]. Of particular interest to salt and water balance, Na⁺/K⁺-ATPase (NKA) uses ATP to generate electrochemical gradients, enabling ionic and osmotic regulation required to maintain homeostasis (Karnaky, 1986; McCormick, 1995; Brijs et al., 2015). In fact, enterocyte NKA activity is upregulated in SW compared with FW (Evans et al., 2005; Edwards and Marshall, 2012), directly fueling increased transport across the intestine; a phenomenon present in the caeca themselves (e.g. Veillette et al., 2005; Schwarz and Allen, 2014). In contrast, the impact of transitions to ion-poor water (IPW), which results in several osmoregulatory compensations, on the function of caeca is unknown. This is despite variations in FW ion levels (e.g. 4 mmol Ca²⁺ l⁻¹ versus 10 µmol Ca²⁺ l⁻¹; Marier et al., 1979) based on numerous geological and environmental factors. In fact, depending on water body location, fish may encounter a range of water compositions from very hard to naturally soft throughout their life cycle.

Finally, the gut of fish has documented zonation in terms of salt and water balance (e.g. Bucking and Wood, 2006a; Whittamore, 2012), transporter expression (e.g. Grosell et al., 2007; Bucking and Wood, 2012; Li et al., 2014), and nutrient transport and enzyme activities (e.g. Collie, 1985; Mommsen et al., 2003; Bucking and Wood, 2012; Rubino et al., 2014; Turner and Bucking, 2017). Indeed, the caeca of rainbow trout display the highest activities of several metabolic enzymes, including glutamate dehydrogenase (GDH), lactate dehydrogenase (LDH) and citrate synthase (CS) (Mommsen et al., 2003). However, it is unclear whether there is zonation of enzyme activity or transport within caeca and what responses these appendages have to abiotic factors such as salinity. FW–SW transitions correlate with alterations in gut metabolic enzyme activities in teleost fish [e.g. GDH, Chew et al., 2010; LDH and pyruvate kinase (PK), Ransberry et al., 2015], suggesting an environmentally dependent role for caeca in ATP production. Whether these effects are correlated with ion transport and energy demand is also unclear.

Despite previous assertions that the caeca were mere extensions of the intestine proper, detailed histological investigations suggest a heterogeneity of cells along individual caeca from the proximal base to the distal tip. Hence, our study investigated Ca²⁺ transport zonation within and between individual caeca using the scanning ion electrode technique (SIET) as well as pharmacological interventions such as LaCl₃ or W7 [N-(6-aminohexyl)5-chloro-1-naphthalene sulfonamide (a calmodulin antagonist that inhibits Ca²⁺-ATPase activity; Garrahan, 2018; Perry and Flik, 1988)]. Owing to their structure and function, we predicted region-specific ion transport along the individual caeca, mirroring epithelial cell distributions, but no zonation along the anterior–posterior axis. Furthermore, we predicted that zonation of enzyme activity (NKA and metabolic enzymes) would be present, correlating with ATP and transport demand within the organs. Finally, we tested the impact of environmental salinity to observe whether the function of caeca themselves was responsive. The effect of IPW on the caeca is unknown, but due to their role in salt and water balance in SW, they may be the source of needed salt absorption to compensate for lowered environmental concentrations. Hence, we predicted that IPW exposure would affect NKA and metabolic enzyme activity to support the increase in observed ion transport.

Experiments were conducted under approved institutional animal care guidelines in accordance with federal mandates. All chemicals were purchased from Sigma-Aldrich unless otherwise noted.

Juvenile rainbow trout [Oncorhynchus mykiss (Walbaum 1792), ∼45 g] were obtained from Humber Springs Trout Hatchery (Orangeville, ON, Canada). Fish were transported to the laboratory and housed in 200 l opaque tanks supplied with flow-through, aerated, dechlorinated City of Toronto tap water ([Na⁺]=590 µmol l⁻¹, [Cl⁻]=920 µmol l⁻¹, [Ca²⁺]=900 µmol l⁻¹, [K⁺]=50 µmol l⁻¹, pH 7.4). Fish were fed daily with commercial trout pellets (∼3% body mass ration; Martin Mills Profishent, Elmira, ON, CA; ∼1.4 g per fish) and were held under a constant photoperiod (14 h:10 h light:dark). Following a period of laboratory acclimation, fish were fed daily at a set time and animals were randomly sampled from the tank 24 h following feeding (referred to as fed FW fish). Additionally, animals were also fasted in FW for 1 week and then sampled (referred to as unfed FW fish).

For IPW exposures, animals were fed and the flow-through water supply to the tanks was abruptly changed to IPW ([Na⁺]=20 µmol l⁻¹, [Cl⁻]=40 µmol l⁻¹, [Ca²⁺]=2 µmol l⁻¹, [K⁺]=0.4 µmol l⁻¹, pH 7.4, temperature=13°C; measured at inflow). IPW water was generated through reverse osmosis of dechlorinated City of Toronto tap water. Removal of uneaten food (if any) 5 min following the designated feeding time eliminated any potential leaching of ions and subsequent alterations of IPW parameters. IPW-exposed animals were sampled following 24 h exposure (referred to as fed IPW fish).

An additional fed FW control group was used for comparison with the IPW animals. This was to determine whether the feeding results that we observed were repeatable. For this control group, the rainbow trout were held in FW, fed and then sampled 24 h following feeding as for the IPW fish.

During sampling, all animals were terminally anesthetised with pH buffered tricaine methasulfonate [MS-222 (Sundel Laboratories Ltd, Canada); 0.5 g l⁻¹, buffered to pH 7.4 with NaOH prior to use]. Following this, the body wall was opened and the entire gut removed. The anterior intestine was identified based on the presence of ∼75 caecum structures and divided into three equal and proportional sections: the anterior, the middle and the posterior. Individual caeca were removed from the anterior, middle and posterior areas of the anterior intestine at the junction with the anterior intestine itself. These intact caeca from each section were then used for ion flux measurements, described below. The remaining caeca from each section were removed from the anterior intestine, visually divided in half and subsequently dissected into proximal (previous site of attachment to the anterior intestine) and distal (at the caeca tip) portions, which were each pooled within each section. These pooled tissue samples were immediately freeze-clamped in liquid nitrogen and placed in a −80°C freezer until enzyme analysis. The remaining anterior intestine was sampled with the absence of caeca (to obtain tissue from the intestine proper), as were gill filaments from the entire second and third gill arch and the whole kidney, freeze-clamped and stored at −80°C until analysis.

The scanning ion electrode technique (SIET) was used to measure ion flux across the mucosa as previously described (Rheault and O'Donnell, 2001; Nguyen and Donini, 2010). Each caecum was everted before measurement, filled with saline (described briefly in Table S1) delivered via PE-50 tubing attached to a 22-G needle and 1 ml syringe, and tied with thread, creating a taut, sealed caecum. The presence of leaks was excluded through visual inspection. During preparation, the tissues were incubated in ice-cold Cortland saline [123 mmol l⁻¹ NaCl, 5 mmol l⁻¹ KCl, 1.9 mmol l⁻¹ CaCl₂, 1.0 mmol l⁻¹ MgSO₄, 11.9 mmol l⁻¹ NaHCO₃, 2.9 mmol l⁻¹ NaH₂PO₄, 5.5 mmol l⁻¹ glucose; pH 7.4 (adjusted to final pH with 1 mol l⁻¹ NaOH)], but ion transport measurements were obtained at room temperature (22°C) and each individual caecum was filled and incubated in a series of solutions containing various concentrations of Ca²⁺ (Table S1). A schematic layout of Ca²⁺ measurements is found in Fig. S1.

For asymmetrical treatments, the internal serosal saline that filled the sac contained 0 µmol l⁻¹ Ca²⁺. The external bath CaCl₂ concentrations (0, 4, 8, 16, 32, 46, 68, 136, 275, 500 and 1000 µmol l⁻¹) were manipulated to determine transport kinetics (Table S1). The osmotic pressure between the mucosal and serosal baths was maintained through the addition of mannitol [3.2–1753 µmol l⁻¹; measured with an osmometer (Advanced Analytics, Fisher Scientific)]. Additionally, using unfed tissues, 200 µmol l⁻¹ LaCl₃ or 1 µmol l⁻¹ W7 was added to the serosal and mucosal baths. W7 was dissolved in ethanol (100%; Fisher Scientific) and application of W7 never exceeded 0.1% vol/vol ethanol in the tissue bath. Ion transport measurements were made at the proximal portion of each caecum.

For symmetrical treatments, the mucosal and serosal bath Ca²⁺ concentrations were matched (0, 4, 8, 16, 32, 46, 68, 136, 275, 500 and 1000 µmol l⁻¹; Table S1). For this experiment, the PE-50 tubing was tied into the sac and heat-sealed between each measurement. The tubing was then re-opened, the sac drained and re-filled with a new Ca²⁺ concentration. This was repeated for each of the concentrations, which were applied in a random order. During the measurements of kinetic flux in unfed tissues, at 500 µmol l⁻¹ mucosal and serosal Ca²⁺, 200 µmol l⁻¹ LaCl₃, 1 µmol l⁻¹ W7 or 100 µmol l⁻¹ ouabain was added in a random order and flux was measured following 10 min of incubation. Following each inhibitor, the tissues were bathed in inhibitor-free mucosal saline for 10 min and recovery of transport rates to pre-exposure values was verified. Interference from La³⁺, W7 and ouabain on the ion-selective microelectrode (ISME) was not detected. Finally, ion transport measurements were made at the proximal and distal points of each caecum (along with a mid-point in the anterior caeca owing to their length) for tissues obtained from unfed FW animals only. For all other treatments, ion measurement were made at the proximal portion.

All intestinal enzymes were measured according to previously published methods. Before assays were run, pooled proximal and distal tissues from the anterior, middle and posterior caeca were individually ground into a fine powder with a mortar and pestle under liquid nitrogen and stored again at −80°C. Care was taken to ensure that the tissue did not thaw until placed in extraction buffer for enzyme analysis. Reactions were started by addition of tissue homogenization preparation and enzyme activity rates were determined as the maximal change in extinction at 340 nm owing to the oxidation or reduction of NADH or NAD using microplates read on a spectrophotometer (Biotek, Fisher Scientific) unless otherwise stated. Enzyme activity, defined as micromoles of substrate converted to product per minute (IU), at 25°C, was expressed per gram of intestinal tissue protein (determined by the Bradford method using a commercial kit; Bio-Rad) unless otherwise noted. Maximal activities were determined for each enzyme during optimization.

NKA (E.C. 3.6.3.9; e.g. Bucking et al., 2013; Turner and Bucking, 2017) activity was measured by placing an aliquot of ground tissue into homogenization buffer which was homogenized with a Wheaton overhead stirrer (with a modified pestle probe; Rochdale, UK) for 30 s. A supernatant was obtained following centrifugation (5000 g) at 4°C for 30 s. The homogenization buffers [SEI buffer (150 mmol l⁻¹ sucrose, 10 mmol l⁻¹ EDTA, 50 mmol l⁻¹ imidazole, pH 7.5); SEID buffer (0.5 g sodium deoxycholate in 100 ml SEI buffer)] and reaction buffers [salt solution (50 mmol l⁻¹ NaCl, 10.5 mmol l⁻¹ MgCl₂, 42 mmol l⁻¹ KCl, 50 mmol l⁻¹ imidazol), Solution A (50 mmol l⁻¹ imidazole, 2.8 mmol l⁻¹ PEP, 0.22 mmol l⁻¹ NADH, 0.7 mmol l⁻¹ ATP, 4 units ml⁻¹ LDH, 5 units ml⁻¹ PK), and Solution B (0.5 mmol l⁻¹ ouabain in 10 ml Solution A)] were made fresh daily before use. Solutions and homogenized tissues were kept on ice until the assay was read. NKA activity was expressed in µmol ADP mg⁻¹ protein h⁻¹.

An additional aliquot of the ground frozen tissue was added to four volumes of extract buffer [50 mmol l⁻¹ Hepes (pH 7.5), 50 mmol l⁻¹ KCl, 1 mmol l⁻¹ dithiothreitol and 0.5 mmol l⁻¹ EDTA] and homogenized as above. The homogenate was then centrifuged at 14,500 g for 5 min at 4°C to pellet cellular material before removing aliquots of the supernatant for each enzymatic assay. GDH (EC 1.4.1.3) activity was measured using previously published methods (e.g. Rubino et al., 2014). The formation rate of glutamate, coupled to the oxidation of NADH in the presence of ADP, was measured with 14 mmol l⁻¹ α-ketoglutarate as the substrate (omitted for control). L-type PK (EC 2.7.1.40) activity was measured as previously described (Mommsen et al., 1980). The reaction buffer consisted of imidazole (50 mmol l⁻¹), KCl (100 mmol l⁻¹), MgCl₂ (10 mmol l⁻¹), NADH (0.15 mmol l⁻¹), ADP (5 mmol l⁻¹), phosphoenol pyruvate (5.0 mmol l⁻¹; PEP; omitted for control), fructose-1,6 biphosphate (0.1 mmol l⁻¹; FBP) and excess lactic dehydrogenase (∼5 U). LDH (EC 1.1.1.27) was measured in a reaction buffer consisting of imidazole (50 mmol l⁻¹), NADH (0.15 mmol l⁻¹) and Na⁺-pyruvate (0.2 mmol l⁻¹), which was omitted for control. CS (EC 2.3.3.1) was assayed in Tris–HCl buffer (75 mmol l⁻¹ pH 8.0) containing DTNB [5,5′-dithio-bis (2-nitrobenzoic acid); 0.1 mmol l⁻¹] and oxaloacetate (OXA; 0.5 mmol l⁻¹) and monitored at 412 nm.

All statistical tests were carried out in SigmaStat 3 and plots were constructed in SigmaPlot10 (Systat). Data were first examined for normality and homogeneity of variance before statistical testing. Kinetic data were modeled using SigmaPlot for line of best fit with linear regression or Michaelis–Menten kinetics with single site saturation. Michaelis–Menten kinetics were determined as: J_Abs=J_Max×[X]/([X]+K_M), where J_Abs is the measured Ca²⁺ influx rate (nmol cm⁻² min⁻¹), [X] is the Ca²⁺ concentration (in μmol l⁻¹), J_Max is the maximal unidirectional flux rate (nmol cm⁻² min⁻¹) and the K_M value is the Ca concentration (in μmol l⁻¹) providing an uptake rate equal to half J_Max. Paired and unpaired Student’s t-tests were used to determine the difference between unfed and fed transport rates, as well as the average transport rates for each section following La³⁺ application, under asymmetrical conditions. The linear rates of transport following W7 application were compared using an ANCOVA. A one-way repeated-measures ANOVA (treatment as the factor) determined the differences between J_Max and K_M values under symmetrical conditions. Likewise, zonation of transport along the proximal–distal axis was examined using a one-way repeated-measures ANOVA (with section as the factor). The impact of inhibitors on the relative ion transport within each section was examined using a two-way repeated-measures ANOVA (inhibitor and caeca portions as factors), whereas the enzyme activities were examined with a two-way repeated-measures ANOVA (with section and treatment as factors). Individual enzyme activities were compared between the two freshwater feeding trials using an unpaired t-test. There were no differences found between the two feeding trials (P>0.05), thus the results from each were pooled and analyzed together. Unfed NKA activities across sections were compared using a one-way repeated-measures ANOVA, whereas NKA activities across tissue or tissue section and treatment were compared using a two-way repeated-measures ANOVA. All were followed by a Holm–Šidák post hoc test. Significance was assessed at P<0.05. Values are presented as means±s.e.m. (N=individual preparations from different individuals).

Under asymmetrical conditions, the three caeca sections displayed a net positive flux, indicating absorption into the mucosa (Fig. 1). Unfed animal tissues from the anterior (Fig. 1A), middle (Fig. 1B) or posterior caeca (Fig. 1C) did not display Michaelis–Menten kinetics as the values failed to converge to satisfy the conditions of the model. Instead, there was a combination of sigmoidal or curvilinear absorption at lower Ca²⁺ concentrations followed by a linear increase in absorption with the mucosal bath Ca²⁺ concentrations above 200 µmol l⁻¹ for all sections (Fig. 1). Feeding did not alter the general kinetics of transport, with all three sections again failing to meet the conditions for Michaelis–Menten transport; however, the linear transport rates above 200 µmol l⁻¹ Ca²⁺ were significantly elevated above the corresponding unfed values for all sections (Fig. 1).

Application of LaCl₃ to the serosal and mucosal baths eliminated the correlation between bath Ca²⁺ concentration and mucosal Ca²⁺ transport. The average transport rate for the anterior caeca across all calcium concentrations (9.4±0.4 nmol Ca²⁺ cm⁻² min⁻¹; N=6; Fig. 2A) was similar to that observed in the middle caeca (10.5±0.6 nmol Ca²⁺ cm⁻² min⁻¹; N=6; Fig. 2B) as well as the posterior caeca (9.7±0.7 nmol Ca²⁺ cm⁻² min⁻¹; N=6; Fig. 2C). In contrast, mucosal and serosal W7 application resulted in linear, diffusive transport in all three caeca sections. The slope of the rate of Ca²⁺ transport in the anterior caeca tissue (0.17; Fig. 2A) was 1.7-fold greater than that observed in the middle caeca tissue (0.1; Fig. 2B), which was greater again (2-fold) than the slope detected in the posterior caeca tissue (0.05; Fig. 2C).

Ca²⁺ transport by the proximal portion of each caeca section displayed typical Michaelis–Menten kinetics under symmetrical Ca²⁺ concentrations, as well as indicating net absorption into the mucosa (Fig. 3). The anterior caeca of unfed fish displayed a J_Max of 47.6±2.0 nmol Ca²⁺ cm⁻² min⁻¹ and a K_M of 118.1±2.3 µmol l⁻¹ Ca²⁺ (model R²=0.982; Fig. 3A). Feeding significantly increased the J_Max 2.1-fold (101.5±9.3 nmol Ca²⁺ cm⁻² min⁻¹) and the K_M 1.2-fold to 142.0±3.1 µmol l⁻¹ Ca²⁺ (model R²=0.975; Fig. 3A). The middle caeca of unfed fish exhibited a significantly lower J_Max (38.7±3.1 nmol Ca²⁺ cm⁻² min⁻¹) and similar K_M (113.6±2.3 µmol l⁻¹ Ca²⁺) compared with the anterior caeca (model R²=0.975; Fig. 3B). Feeding once again increased the J_Max (1.6-fold increase; 61.1±5.2 nmol Ca²⁺ cm⁻² min⁻¹) and K_M 1.2-fold (132.8±2.7 µmol l⁻¹ Ca²⁺), although both the J_Max and K_M remained significantly lower than in the anterior caeca of fed fish (model R²=0.962; Fig. 3B). In the posterior caeca, the J_Max for unfed fish was the lowest of all sections (33.4±3.9 nmol Ca²⁺ cm⁻² min⁻¹), as was the K_M (106.5±4.2 µmol l⁻¹ Ca²⁺; model R²=0.917; Fig. 3C). Feeding in FW nearly doubled the J_Max, significantly elevating the rate to 63.8±9.1 nmol Ca²⁺ cm⁻² min⁻¹, while the K_M significantly increased 1.4-fold to 142.3±5.8 µmol l⁻¹ Ca²⁺ (model R²=0.897; Fig. 3C). The corresponding values for unfed and fed fish were not statistically different from those from the middle caeca (Fig. 3B); however, the J_Max for both unfed and fed fish was significantly lower than the J_Max in the anterior caeca, while the K_M for unfed fish alone was lower (Fig. 3A).

Repeating the feeding experiment verified the elevated J_Max (103.8±1.5 nmol Ca²⁺ cm⁻² min⁻¹) and K_M (139.3±3.8 µmol l⁻¹ Ca²⁺) in the anterior caeca (model R²=0.945; Fig. 4A). Moreover, exposure to IPW further significantly increased the J_Max 1.4-fold to 148.4±2.1 nmol Ca²⁺ cm⁻² min⁻¹ and K_M 1.1-fold to 155.7±4.1 µmol l⁻¹ Ca²⁺ (model R²=0.971; Fig. 4A). In comparison, the increased J_Max and K_M were also present in a repeated experiment (J_Max=66.5±6.3 nmol Ca²⁺ cm⁻² min⁻¹; K_M=129.0±3.3 µmol l⁻¹ Ca²⁺) for the middle caeca obtained from fed animals (model R²=0.950; Fig. 4B). Additionally, IPW exposure similarly increased J_Max 1.4-fold (101.9±1.4 nmol Ca²⁺ cm⁻² min⁻¹) and K_M 1.1-fold (149.1±2.3 µmol l⁻¹ Ca²⁺; model R²=0.942; Fig. 4B). J_Max observed in the middle caeca tissue once again remained significantly lower than in the anterior caeca, while the K_M was not significantly different (Fig. 4B). Finally, repeating the feeding experiment once again resulted in similar transport characteristics of fed tissue (J_Max=67.8±4.3 nmol Ca²⁺ cm⁻² min⁻¹, K_M=137.8±2.5 µmol l⁻¹ Ca²⁺) in the posterior caeca (model R²=0.976; Fig. 4C). However, unlike the middle and anterior caeca, exposure to IPW did not further alter the J_Max (77.1±5.9 nmol Ca²⁺ cm⁻² min⁻¹) or the K_M (143.4±3.3 µmol l⁻¹ Ca²⁺; model R²=0.948; Fig. 4C). The K_M of the IPW posterior caeca was significantly lower than that in the anterior caeca, while the J_Max of the IPW posterior caeca was lower than both the middle and anterior caeca.

There was further zonation within each caeca section under symmetrical conditions. Moving from the proximal portion adjacent to the intestine proper towards the distal tip of each caeca section, there was a significant decrease in J_Max, while K_M was unaffected in unfed animal tissues (Table 1). Once again, there was a significant reduction in J_Max and K_M from the anterior to posterior sections, although this trend was only significant in the proximal portions; the distal portions showed no zonation along the anterior–posterior axis (Table 1).

When analyzing the impact of inhibitors on Ca²⁺ transport using a two-way repeated-measures ANOVA (inhibitor and portion of caeca as factors), there was no significant interaction (P=0.108) nor was there an effect of portion of caeca (P=0.422), whereas there was a significant effect of inhibitor application (P<0.001). Under symmetrical conditions, tissue exposed to ouabain significantly decreased Ca²⁺ transport across all sections and portions of the caeca to values ∼33% lower than those of the control (Table 2). Similarly, application of La³⁺ to the mucosal and serosal surfaces significantly reduced Ca²⁺ transport similarly across all three caeca sections and portions, decreasing transport by approximately 68% compared with control values (Table 2). Finally, application of W7 likewise significantly diminished Ca²⁺ transport consistently across portions and sections; however, transport was reduced by the greatest proportion to ∼12% of control values (Table 2).

Statistical analysis revealed a lack of significance in caeca portion (P=0.892) and a lack of interaction between caeca portion and treatment (P=0.708), allowing for direct interpretation of the significant effect (P<0.01) of treatment on GDH activity (Table 3). There was a significant increase in GDH activity in the IPW treatment over the unfed and fed treatments; however, there was not a significant difference between unfed and fed animals. PK activity likewise showed a significant effect of treatment (P<0.01), whereas caeca portion did not influence activity rates (P=0.152) and there was a lack of interaction between the two factors (P=0.107). As with GDH, IPW exposure elevated PK activity rates over those observed in fed and unfed animals; however, in contrast, there was also a significant increase in fed animals when compared with unfed animals (Table 3). LDH activity was significantly reduced below unfed values in both the fed and IPW-exposed animals (P<0.01); however, fed and IPW animals were not different (Table 3). There was again a lack of significant effect of caeca portion for LDH activity (P=0.099) and a lack of interaction between factors (P=0.422). As with PK, feeding significantly elevated CS levels above those observed in unfed animals, and IPW exposure significantly elevated activity further (P<0.05; Table 3). In contrast, however, there was a significant effect of section (P=0.02), with anterior portions displaying elevated CS activity over posterior portions, while no consistent differences appeared on the proximal–distal axis (Table 3). A lack of interaction between the treatment and portion of caeca factors (P=0.072) allowed for direct interpretation of these effects. Finally, NKA activity in unfed animal tissues showed a strong significant effect of section (anterior to posterior) as well as portion (proximal to distal), with the anterior proximal area exhibiting the highest activity rates and the posterior distal area the lowest (Table 3).

Further exploration of NKA activity rates confirmed a gradient along the anterior to posterior axis, with anterior caeca displaying the highest NKA activity rates and the posterior caeca the lowest (Fig. 5A). The anterior intestine proper presented NKA activity rates that were similar to those of the middle and posterior caeca (average across sections 3.5±0.4 µmol ADP mg⁻¹ protein h⁻¹; N=8; Fig. 5A). Feeding increased NKA activity rates in the anterior intestine proper, anterior caeca, middle caeca and posterior caeca (Fig. 5A). IPW exposure likewise increased NKA activity rates above fed values in the anterior intestine proper and anterior caeca; however, the middle and posterior caeca were unaffected (Fig. 5A). In contrast, feeding and IPW exposure did not elevate kidney NKA rates above those observed in unfed animals (Fig. 5B), whereas feeding elevated branchial NKA rates above unfed rates, and IPW exposure further elevated rates above feeding (Fig. 5B). There was no interaction between factors (P=0.791).

Ca²⁺ transport in the intestine displayed elements of both paracellular and transcellular transport. Under asymmetrical conditions, all three sections of the caeca showed diffusive absorption at higher Ca²⁺ concentrations (Fig. 1), indicative of paracellular movement. In addition, under symmetrical conditions, all three sections have shown typical transport kinetics of transcellular absorption (Figs 2 and 3). Based on these observations and previous research (Perry and Flik, 1988; Flik and Verbost, 1993; Klaren et al., 1993; Flik et al., 1996), we propose the following transport model within the enterocytes of the caeca (Fig. 6). Here, Ca²⁺ enters the cell through an apical channel and is extruded across the basolateral membrane by a plasma membrane Ca²⁺-ATPase (PMCA) and/or an Na⁺-Ca²⁺ exchanger (NCX), which is dependent on the NKA to create a favorable Na⁺ gradient (Fig. 6). Furthermore, paracellular diffusion likely occurs between the cells through the intracellular tight junctions (Fig. 6). Inhibition by ouabain supports the role of the NKA in the observed Ca²⁺ transport (Table 2), while W7 inhibition supports the role of a PMCA (Garrahan, 2018). La³⁺ inhibition of transport (Fig. 2, Table 2) suggests a blockade of paracellular diffusion as well as transcellular transport, possibly through an epithelial Ca²⁺ channel (Verbost et al., 1987; Perry and Flik, 1988; Hogstrand et al., 1995, 1996; Fig. 6). Low to negligible expression of the epithelial Ca²⁺ channel (ECaC) was detected in the intestine of rainbow trout (Shahsavarani, 2006), suggesting an alternative channel (Larsson et al., 1998). Paracellular inhibition of transport may be accounted for by the preference of claudin-2 (a tight-junction forming protein) for La³⁺ (Yu et al., 2010). Interestingly, Abaurrea-Equisoaín and Ostos-Garrido (1996) suggested two ion-transporting cell types in the caeca of rainbow trout. The role of distinct cell types in the caeca is unclear, but examination of the ion transport model of the rainbow trout gill suggests that the cells may individually contribute specific ion-transport pathways (Galvez et al., 2002; Hwang et al., 2011).

Feeding increased transport rates in all sections, with the largest effect present in the anterior caeca (Fig. 6), through increased paracellular diffusion (Fig. 1) as well as increased transcellular capacity and decreased specificity (Fig. 3). Increases in paracellular transport rates suggest an increase in tissue permeability, likely via alterations in the tight junctions between cells. Direct evidence that digesting a meal alters the permeability of the intestine is lacking in fish; however, increased water absorption in the intestine of fed FW fish (Bucking and Wood, 2006b) provides indirect evidence of increased permeability. Should this be occurring, tight-junction proteins such as those from the occludin and claudin family would likely reveal alterations in expression and/or function during feeding. In contrast, the increase in transcellular transport during digestion may have been accomplished by increasing the number of high-capacity, low-affinity transporters (i.e. NCX; Fig. 6), and/or increasing the activity of said transporters through post-translational modifications. This is the first evidence of such compensation during digestion; however, this has been observed in response to environmental manipulations (Schoenmakers et al., 1993). The NCX predominated over the PMCA, which is a comparatively low-capacity, high-affinity transporter (e.g. Blaustein and Lederer, 1999), even in the unfed tilapia intestine (Flik et al., 1996). This may indicate species-specific regulation of intestinal Ca²⁺ transport.

Exposure to IPW further increased ion transport rates above those seen during digestion. Again, this occurred primarily through an increased J_Max and slightly increased K_M, suggesting a similar increase in transporter capacity either through an increase in number and/or a prevalence of the NCX (Figs 3 and 6). In contrast to feeding, this response was not evident in the posterior portions of the caeca (Fig. 4). The opposite occurs during a FW–SW transfer, when demand for Ca²⁺ uptake is reduced, as activities of the PMCA and the NCX decrease; suggested to be an adjustment of Ca²⁺ carrier densities in the basolateral plasma membrane (Schoenmakers et al., 1993). Interestingly, the tubular networks found within the suspected ion-transporting cells are more developed in SW-exposed fish (Giffard-Mena et al., 2006), suggesting an environmental control over caeca morphology and subsequent physiological role. This may reflect intestinal water and salt transport required by SW teleost fish whereby active transport of solutes, sodium and chloride in particular as well as the excretion of bicarbonate (e.g. Ando, 1988; Usher et al., 1991; Wilson et al., 1996; Grosell et al., 2009), creates an osmotic drag to pull water across the intestine (e.g. reviewed by Whittamore, 2012). Combined with this is the precipitation of bicarbonate with calcium, lowering the osmotic pressure of the fluid in the intestine, aiding water uptake into the plasma (Wilson et al., 1996) and resulting in a reduced need for Ca²⁺ absorption. Feeding was an important control for the IPW exposure, as animals faced a risk of mortality if food was withheld during exposure (based on previous laboratory observations). It is possible that exposure to lowered salinity is compensated for directly through digestion, by providing either ions and/or energy. Indeed, the importance of dietary ions has been shown with decreasing salinity in numerous species (Wood and Laurent, 2003; Scott et al., 2006; Marshall and Grosell, 2006).

It is important to note that the observed K_M values (110 µmol l⁻¹ in the present study) vary from published values from other fish species (e.g. 0.19–2.3 µmol l⁻¹ in the tilapia intestine). Furthermore, the K_M for the PMCA is roughly 0.3 µmol l⁻¹ in mammalian studies (Carafoli, 1992), whereas NCX is approximately 10-fold less (approximately 3 µmol l⁻¹; Blaustein and Lederer, 1999). In general, we underestimated the K_M by ∼50-fold; however, other studies have over/underestimated these values using traditional radioisotopes by more than 1000-fold (Klinck et al., 2012). This strongly suggests that while using SIET to measure ion transport, relative changes in magnitude with treatment are indicative of effect, whereas absolute values may not be representative. Remarkably, the J_Max of Ca²⁺ uptake into the mucus binding compartment of the anterior intestine proper of freshwater rainbow trout was ∼8 nmol min⁻¹ cm⁻² (Klinck et al., 2012), roughly similar to the average uptake rate during La³⁺ exposure (Fig. 2), suggesting this may reflect background absorption into the mucous and/or bacteria.

Feeding and IPW exposure increased NKA activity in parallel with ion transport (Fig. 5), possibly aiding increased Ca²⁺ absorption (Fig. 6). Increased enterocyte NKA activity during digestion has been previously observed (e.g. Turner and Bucking, 2017) and likely supports increased ion transport associated with digestion (Bucking and Wood, 2006a). The few studies examining the transition from FW to IPW have revealed physiological changes that mirror those exhibited by euryhaline fishes during SW–FW transfer, i.e. lowered serum osmolality and circulating ion levels, as well as increased circulating glucose and intestinal NKA activity (Chasiotis et al., 2009). Although abrupt transfer has not been examined in the gut, longer-term acclimation to IPW also showed a variable decrease in intestinal occludin expression (Chasiotis et al., 2009), suggesting both paracellular and transcellular Ca²⁺ absorption increases during IPW exposure (Fig. 6). Scott et al. (2006) proposed that increased ion uptake across the intestine with decreasing salinity was driven by paracellular transport, not transcellular transport. Interestingly, long-term acclimation (28 days) of goldfish suggested that following an initial increase in paracellular transport, this is decreased and followed by an increase in transcellular transport (Chasiotis et al., 2009).

Feeding and IPW exposure altered enzymatic activities in a manner paralleling ATP demand through increased NKA activity (Table 3). In particular, CS, a key enzyme in the Krebs cycle and an indicator of mitochondria function and hence ATP production, mirrored NKA activity and ion transport patterns (Fig. 6). Traditionally, the effect of salinity changes on metabolic pathways has been examined in the liver (e.g. PK activity increases ∼1.5-fold, LDH increases 2- to 3-fold; Soengas et al., 1995a; Vijayan et al., 1996). Alterations in hepatic metabolism likely act to provide glucose and lactate for oxidation in the gill (Mommsen, 1984; Perry and Walsh, 1989), supporting the increased energy requirement of the gills as a result of altered ion transporter activity (e.g. Vijayan et al., 1996; Chang et al., 2007). In fact, during SW adaptation, the gills, kidney and intestine can increase total energy consumption (e.g. Evans, 1984; Soengas et al., 1995b); however, they lack extensive energetic reserves, such as glycogen or lipids (Perry and Walsh, 1989; Soengas et al., 1995b), which are rapidly depleted during acclimation (Chang et al., 2007). It is possible that the liver is not the only organ that alters metabolic pathways, as numerous reaction pathways were enhanced in IPW-acclimated fish (Table 3). When fed, SW transfer resulted in no apparent increase in substrate mobilization from the liver (Vijayan et al., 1996). Perhaps the intestine was supplying more substrates in the plasma, avoiding an increase in energy mobilization from the liver. The intestine of rainbow trout is notable for its capacity to store glucose as glycogen in the muscle layers, to oxidize glucose into lactate, and to regulate its own glucose homeostasis and glycolytic capacity (Polakof et al., 2010).

There was clear zonation in ion transport and NKA activity, along both the proximal–distal (Table 1) and anterior–posterior axes (Figs 3 and 6). Anterior–posterior zonation was also seen with CS activity (Figs 5 and 6); however, the other metabolic enzymes examined did not display zonation (Table 3; Fig. 6). The proximal–distal zonation was expected based on indications of cell-type zonation, with absorptive (Anderson, 1986) and possible ion-transporting cells (Abaurrea-Equisoaín and Ostos-Garrido, 1996). However, zonation along the anterior–posterior axis was unexpected (Williams and Nickol, 1989). Broad zonation along the anterior–posterior axis of the entire gut has been observed. Generally, the anterior intestines of teleosts display histological features that suggest lipid and amino acid absorption (Sire et al., 1981; Ostos Garrido et al., 1993), whereas the posterior intestine demonstrates macromolecule absorption via pinocytosis (Ezeasor and Stokoe, 1981; Anderson, 1986; Abaurrea et al., 1993). For example, intestinal nutrient transport showed zonation along the gut of coho salmon (proline uptake: caeca>>anterior intestine>posterior intestine; Collie, 1985), driven by an increased J_Max and similar K_M, suggesting an increased number of transporters as opposed to an isoform gradient. The fine-scale zonation in ion transport within the caeca themselves is a novel observation and the implication requires further research. It is interesting to observe that the anterior intestine proper had slightly lower NKA activities compared with the anterior caeca (Fig. 5A), despite the well-known role in nutrient absorption and digestion. This may be a result of increased muscle thickness around the intestine proper diluting the detected activity of NKA.

We have used a novel technique to establish the kinetics of Ca²⁺ transport in the caeca of rainbow trout, as well as responses to pharmacological inhibitors and environmental manipulation and digestion. Ca²⁺ transport in the caeca matches existing transport models with both paracellular and transcellular pathways. We have discovered zonation in this function, with proximal and anterior portions displaying the highest level of activity, coinciding with ultrastructural evidence. Ion transport patterns correlated with NKA and CS activity, suggesting an increase in ATP use and production. Increased carbohydrate and amino acid metabolism may have provided the ATP; however, zonation was not apparent. This supports the hypothesis that the diet contributes a substantial amount of ions required to maintain homeostasis, and during IPW exposure the contribution may further increase to compensate for a limited ability to acquire ions from the surrounding environment. Overall, our results suggest that the caeca of trout are heterogeneous appendages, with functional zones dedicated to active ion transport supported by energy production through local enzyme activity.

Thanks are given to A. Donini for his support and access to the SIET system. Sampling help from D. Ahkmen and H. Groser was appreciated. We thank the reviewers for their helpful comments, and appreciate their efforts.