The role of environmental calcium in the extreme acid tolerance of northern banjo frog (Limnodynastes terraereginae) larvae

Life in freshwater environments is complicated by the fact that animals are hyper-ionic with respect to the environment and ions tend to move out of the animal along their diffusion gradients. In order to offset ion losses, freshwater animals actively take up ions from the environment via specialised cells in the gill epithelia, integument (in larvae and embryos) and across the gut (Edwards and Marshall, 2012; Evans et al., 1999). The physiological challenges of living in freshwater environments can be further compounded by low pH. Most aquatically respiring animals are intolerant of water pH <5 (low pH). In many animals, low pH water substantially impairs both ion uptake capacity and epithelial integrity, leading to a rapid loss of homeostasis that can be fatal (Freda and Dunson, 1984; McDonald et al., 1984; Meyer et al., 2010, 2020; Robinson, 1993).

The effect of low pH water on branchial epithelial integrity drives its toxicity in many species. In acid-sensitive animals, exposure to low pH water reduces transepithelial resistance through the disruption of paracellular tight junctions (Daye and Garside, 1976; Meyer et al., 2010; Rosseland and Staurnes, 1994). Tight junctions (TJs) are composed of a complex of transmembrane and associated proteins which function in cell–cell adhesion and control epithelial permeability by acting as a selective barrier to ions and small molecules (Schneeberger and Lynch, 2004). In doing so, TJs regulate and maintain key transcellular ion gradients necessary to facilitate transepithelial ion uptake. Low extracellular pH has been shown to reduce the resistance of the paracellular pathway by changing the conformation of TJ proteins, which alters their gating properties (Schneeberger and Lynch, 1992, 2004).

Environmental calcium concentration ([Ca²⁺]_e) is essential to maintaining the integrity of the epithelial TJ (Schneeberger and Lynch, 1992). Increased [Ca²⁺]_e has been shown to ameliorate the effects of low pH on epithelial integrity in some fish and amphibian species (Dalziel et al., 1986; Matsuo and Val, 2002; McDonald et al., 1983; Meyer et al., 2020). This may be because low [Ca²⁺]_e causes dissociation of the paracellular junctions, increasing epithelial permeability (Bhat et al., 1993; Ma et al., 2000; O'Keefe et al., 1987). In some fishes and larval amphibians acutely exposed to low pH, [Ca²⁺]_e correlates negatively with the rate of net Na⁺ loss (Cummins, 1988; Freda et al., 1991; Gascon et al., 1987; Gonzalez and Dunson, 1989; Kumai et al., 2011; McDonald and Rogano, 1986; McDonald et al., 1983). Ca²⁺ may act on the TJ directly, but also indirectly through Ca²⁺-dependent proteins in adherens junctions (AJs) which sit basally to the TJ (Brown and Davis, 2002). Removal of Ca²⁺ from the extracellular space has been shown to reduce cytosolic [Ca²⁺] (González-Mariscal et al., 1990), cause the detachment and internalisation of extracellular AJ and TJ proteins (Volberg et al., 1986), and reduce transepithelial resistance (González-Mariscal et al., 1990). AJs function primarily in cell–cell adhesion, but also in the regulation of the actin cytoskeleton and as transcriptional regulators (Hartsock and Nelson, 2008). Extracellular Ca²⁺ can regulate paracellular permeability by interacting directly with calcium-dependent junctional proteins on the cell surface and/or through active transcellular Ca²⁺ uptake and cytosolic Ca²⁺ signalling pathways (Stuart et al., 1996). Although [Ca²⁺]_e is important in determining the effects of low pH exposure in fish and amphibians, the specific mechanism through which this occurs remains unknown.

Transcellular uptake of Ca²⁺ from the surrounding water via the gills constitutes the major route by which Ca²⁺ is taken up in fish (Baldwin and Bentley, 1980; Chasiotis et al., 2012; Flik and Verbost, 1995). Active branchial Ca²⁺ uptake occurs at ionocytes in the branchial epithelium. Extracellular Ca²⁺ enters ionocytes through non-voltage-gated epithelial Ca²⁺ channels (ECaC) in the apical membrane (Edwards and Marshall, 2012; Flik and Verbost, 1995) and is then shuttled to the basolateral membrane by Ca²⁺-binding proteins such as calbindin for extrusion via Na⁺/Ca²⁺ exchangers (NCX) and plasma membrane Ca²⁺-ATPase transporters (PMCA). Once in the basolateral extracellular space, Ca²⁺ can directly interact with the extracellular Ca²⁺-binding domain of E-cadherin in AJs to influence its properties and those of the overlying TJ (Pokutta et al., 1994; Zhang et al., 2009). Branchial Ca²⁺ absorption is tightly controlled through the regulation of ECaC activity and changes in the abundance of intracellular Ca²⁺ transport proteins (Cai et al., 1993; Kelly and Wood, 2008; Khanal and Nemere, 2008; Shahsavarani and Perry, 2006; Verbost et al., 1993; Wongdee and Charoenphandhu, 2013). This, in turn, may reduce the transcellular movement of Ca²⁺ to the basolateral extracellular space, which could limit its availability to E-cadherin in the AJ and compromise the permeability of the overlying TJ. Factors that compromise the maintenance of transcellular Ca²⁺ transport pathways could influence the Ca²⁺-sensitive aspects of junctional stability.

Given that highly acidic water is toxic to most aquatic animals, naturally acidic freshwater bodies are surprisingly biodiverse. Among the most acidic freshwater ecosystems in the world is the Wallum along the eastern coast of southern Queensland and northern New South Wales in Australia. Wallum ecosystems are characterised by highly acidic waters, ranging from pH 2.8 to 5.5 (Hines and Meyer, 2011). Compounding the difficulties of living at low pH, Wallum waters are also dilute (low in salts) and soft (low in Ca²⁺ and Mg²⁺; Bayly, 1964). Despite these challenges, larvae of some Wallum frog species can tolerate exceptionally acidic waters (Hines and Meyer, 2011; Meyer, 2004; Meyer et al., 2020). One such species is the northern banjo frog, Limnodynastes terraereginae, populations of which can be found throughout eastern Australia inhabiting aquatic environments which range in pH from circumneutral to pH 3.0, making it one of the most highly acid-tolerant vertebrate species known. Meyer (2004) showed that L. terraereginae larvae from Wallum environments are highly acid tolerant in part as a result of their capacity to withstand the acute disruption of epithelial integrity and associated loss of body Na⁺. Moreover, [Ca²⁺]_e was implicated in underpinning the resistance of larvae to acid-induced Na⁺ efflux in acutely exposed L. terraereginae larvae but not in acid-acclimated larvae (Meyer, 2004). The role of [Ca²⁺]_e in facilitating acid tolerance in acid-acclimated larvae and the mechanistic basis by which this occurs remain unclear. Given that [Ca²⁺]_e is limited in Wallum environments and that Ca²⁺ uptake is typically inhibited by low pH (Malley, 1980; Yeh et al., 2003), understanding how amphibian larvae manage Ca²⁺ transport in low pH water is likely central to understanding the mechanistic basis of their tolerance to these extreme environments.

To determine the importance of [Ca²⁺] in Na⁺ homeostasis at low pH, we examined whole-animal Na⁺ and Ca²⁺ flux following both acute and chronic exposure to low pH and different [Ca²⁺]_e in L. terraereginae larvae. To understand the role of transcellular Ca²⁺ uptake for the maintenance of epithelial integrity at low pH, L. terraereginae larvae reared at low pH were exposed to the calcium channel antagonist ruthenium red (RR). We also measured gene expression patterns of four key Ca²⁺ transport proteins (ECaC, calbindin, NCX and PMCA) and E-cadherin in the gill epithelia. We hypothesised that acute exposure to low pH would result in increased net Na⁺ efflux under low [Ca²⁺]_e, but that chronic exposure (acclimation) to low [Ca²⁺]_e would reduce net Na⁺ efflux with low pH exposure and increase Ca²⁺ influx. In addition, we hypothesised that the acute impairment of apical Ca²⁺ uptake via ECaC (with RR) would increase net efflux of Na⁺ at low pH, consistent with a role for transcellular Ca²⁺ transport in the maintenance of intercellular junction integrity. Finally, acclimation to both low pH and low [Ca²⁺]_e was hypothesised to correspond to an increase in the expression of the four key Ca²⁺ transporters (ECaC, calbindin, NCX and PMCA) and E-cadherin, consistent with an increased rate of transcellular Ca²⁺ uptake and reinforcement of Ca²⁺-dependent AJs to protect junctional integrity in acid-acclimated larvae.

All animals were collected under the Queensland Department of Environment and Heritage Protection Scientific Purposes Permit (WITK15563515), and all procedures were approved by The University of Queensland's Animal Ethics Welfare Unit (SBS/484/17) and the University of Queensland's Animal Ethics Committee (approval number: SBS/460/14/ARC). Limnodynastes terraereginae Fry 1915 egg masses were collected in January 2018 from Bribie Island National Park (water pH 3.3–4.4), QLD, Australia. Eggs were allocated to circumneutral (pH 6.5) or low (pH 3.5) pH artificial soft water (ASW; Freda and Dunson, 1984): distilled water plus (in μmol l⁻¹) 40 CaCl₂·2H₂O, 40 MgSO₄·7H₂O, 120 NaCl, 50 NaOH and 20 KCl; pH was adjusted with 0.1 mol l⁻¹ H₂SO₄. Hatched tadpoles were housed in 5 l plastic tanks connected to two 200 l filtered recirculating aquarium systems (15 tanks per system). Each system was connected to a canister filter for biological, mechanical and chemical filtration (Fluval G6). System pH was monitored daily (LAQUA P-22, Horiba Instruments, Singapore) and regulated as necessary through the addition of 0.1 mol l⁻¹ H₂SO₄. Water [Na⁺]_e and [Ca²⁺]_e were measured weekly using flame photometry (BWB Technologies, Newbury, UK). Tadpoles were fed every second day with thawed frozen spinach, and each system underwent a 20% water change weekly. Room temperature was maintained at 22±1°C with fluorescent overhead lighting programmed to a 12 h:12 h light:dark photoperiod (06:00–18:00 h). The following experiments were replicated once within the laboratory.

To assess the effect of acute exposure to high or low [Ca²⁺]_e on acid-induced net Na⁺ and Ca²⁺ flux, L. terraereginae larvae (n=36; Gosner stages 26–38; Gosner, 1960) were randomly allocated to three 5 l tanks (n≤12 per tank to obtain adequate statistical power, based on previous ion flux work with the species in Meyer, 2004) containing 50 μmol l⁻¹ [Ca²⁺]_e ASW (Table 1). Water pH was maintained at pH 6.5 and larvae were acclimated to these conditions for 4 weeks. Water pH was monitored daily and regulated as necessary via addition of 0.1 mol l⁻¹ H₂SO₄. Tadpoles were fed every second day with thawed frozen spinach, and each system underwent a 100% water change weekly. Ammonia levels were monitored using an API^® Ammonia Test Kit (Mars Fishcare, Chalfont, PA, USA). Larvae were fasted for 2 days and then placed into individual 200 ml glass beakers containing 50 ml of 5, 50 or 250 μmol l⁻¹ [Ca²⁺]_e ASW (n=12 per treatment), 30 min prior to testing. Water pH in half of the beakers in each treatment (n=6) was acutely lowered to pH 3.5 through the addition of dilute H₂SO₄ (0.1 mol l⁻¹). A 5 ml water sample was collected from all beakers 1 and 7 h following pH adjustment for the measurement of Na⁺ and Ca²⁺ concentration. Net ion flux was determined by subtracting the ion concentration at the start of the exposure (1 h sample) from that at the end (7 h sample). Following experimentation, all larvae were lightly blotted dry, weighed (mean mass Na⁺ model 0.68 g, Ca²⁺ model 0.7 g), then euthanised by immersion in 0.25 mg l⁻¹ buffered MS222 (Ramlochansingh et al., 2014) and pithing for assessment of total body [Na⁺] and [Ca²⁺]. Larvae were dissolved in 2 ml of 70% NHO₃ and 5 ml 1:100 aliquots were prepared by filtering the sample through a 45 μmol l⁻¹ filter syringe (Fisherbrand™, Loughborough, UK) before [Na⁺] and [Ca²⁺] were measured in duplicate using a flame photometer as described above.

To assess the effect of chronic exposure (acclimation) to high or low [Ca²⁺]_e on whole-animal net Na⁺ and Ca²⁺ flux following acute exposure to low pH, L. terraereginae larvae (n=24; Gosner stages 26–38) were transferred to tanks containing either 5 or 250 μmol l⁻¹ [Ca²⁺]_e ASW 4 weeks prior to sampling. Control larvae were maintained in 50 µmol l⁻¹ ASW. Thirty minutes prior to testing, larvae were placed into individual 200 ml glass beakers containing 50 ml of ASW with the same [Ca²⁺ ] as their holding tanks (n=12 per treatment). Water pH in the beakers in each treatment was then acutely lowered to pH 3.5 through the addition of 0.1 mol l⁻¹ H₂SO₄. Water samples were then collected from all beakers at 1 and 7 h and analysed as detailed above for [Na⁺] and [Ca²⁺]. Larvae were blotted dry and weighed (mean mass Na⁺ model 0.64 g, Ca²⁺ model 0.59 g).

RR was used to inhibit apical Ca²⁺ transport. RR does not penetrate TJs and is commonly used as a histochemical marker of the barrier formed by epithelial TJs (González-Mariscal et al., 1989; West et al., 2002). Limnodynastes terraereginae larvae (n=12; Gosner stages 26–38) were randomly allocated to an isolated 5 l tank containing 50 µmol l⁻¹ [Ca²⁺]_e ASW at pH 3.5 and maintained for 4 weeks as detailed in experiment 1. Thirty minutes prior to testing, larvae (n=12) were placed into individual 200 ml glass beakers containing 50 ml of 50 μmol l⁻¹ [Ca²⁺]_e ASW at pH 3.5. RR was added to half of the beakers to a concentration proven to induce an approximately half-maximal response in ECaC activity in vitro (10 μmol l⁻¹; Hoenderop et al., 2001). Water samples were collected at 1 and 7 h post-exposure and analysed as described above. Larvae were then removed from beakers, blotted dry and weighed (mean mass 0.86 g).

To assess whether [Ca²⁺]_e exposure and low pH influences the expression of branchial Ca²⁺ transport proteins and E-cadherin, L. terraereginae larvae (n=36; Gosner stages 26–38) were randomly allocated to six isolated 5 l tanks (n=6 per tank) containing 5, 50 and 250 μmol l⁻¹ [Ca²⁺]_e ASW. Water pH was maintained at pH 6.5 (n=18) or reduced to pH 3.5 (n=18). Larvae were maintained under these conditions for 4 weeks. Limnodynastes terraereginae larvae were then euthanised by immersion in 0.25 mg l⁻¹ buffered MS222 (Ramlochansingh et al., 2014) and pithing. Both branchial baskets were dissected free and stored in RNAlater (Ambion Inc.) at 4°C for 24 h, before being moved into a −20°C freezer. Total RNA was extracted from L. terraereginae gills using an RNeasy Mini Kit following the manufacturer's instructions (Qiagen, Valencia, CA, USA). Total RNA was eluted from the silicon spin column in ultrapure water and its concentration quantified using a Qubit fluorometer (ThermoFisher Scientific, Waltham, MA, USA). Any residual genomic DNA contamination was removed, and RNA was reverse transcribed using an iScript gDNA Clear cDNA Synthesis Kit (Bio-Rad Laboratories Inc., Hercules, CA, USA) following the manufacturer’s guidelines. Appropriate no-reverse transcriptase controls were generated by replacing reverse transcriptase with water.

The transcripts for target genes (ECaC, Calb1, PMCA, NCX, E-cadherin) and house-keeping genes (β-actin, GAPDH, RPS, TUB) were identified using an in-house L. terraereginae transcriptome with homologous sequences from other amphibians as the reference query. Reference sequences were compared against the L. terraereginae transcriptome using the ‘blastn’ tool in Galaxy Australia (Jalili et al., 2020). Putative L. terraereginae gene sequences were then compared against the National Centre for Biotechnology Information (NCBI) database using the ‘blastn’ tools default parameters to confirm their identity. PrimerQuest (Integrated DNA Technologies, Caraville, IA, USA) was used to design specific qPCR primers (Table S1). All primer pairs were evaluated for specificity and to ensure that they produced only a single band of the appropriate length using MyTaq DNA Polymerase (Bioline, Alexandria, NSW, Australia) and agarose gel electrophoresis.

All analyses were conducted in the R statistical environment (http://www.R-project.org/). α was set at 0.05 for all statistical tests. Models were two-tailed and assumed a Gaussian error structure, and data satisfied assumptions of hypothesis tests. The effects of body size on rates of ion flux were accounted for by considering wet body mass as a covariate in statistical models.

Differences in net Na⁺ and Ca²⁺ flux between treatment groups were tested by fitting analysis of covariance (ANCOVA) models using the car package (Fox and Weisberg, 2018). For experiment 1, a two-way ANCOVA was fitted using exposure [Ca²⁺]_e and test pH as independent variables. A separate two-way ANCOVA was fitted for whole-body [Na⁺] and [Ca²⁺] in these larvae using exposure [Ca²⁺]_e and test pH as independent variables (Fig. S1). To determine whether acclimation to low or high Ca²⁺ affected Na⁺ and Ca²⁺ flux at low pH in experiment 2, larvae reared at 50 μmol l⁻¹ [Ca²⁺]_e and exposed to low pH at 5 and 250 μmol l⁻¹ [Ca²⁺]_e from experiment 1 were compared with larvae reared at 5 or 250 μmol l⁻¹ [Ca²⁺]_e and exposed to low pH at 5 and 250 μmol l⁻¹ [Ca²⁺]_e. A two-way ANCOVA was fitted modelling test [Ca²⁺]_e and rearing [Ca²⁺]_e (equimolar to test [Ca²⁺]_e versus 50 μmol l⁻¹ [Ca²⁺]_e control) as independent variables. For experiment 3, a one-way ANCOVA was fitted with RR treatment as the independent variable. For all experiments, one-sample Student's t-tests were used to test whether Na⁺ and Ca²⁺ flux in control groups was significantly different from zero, suggesting a departure from homeostasis. Net Na⁺ and Ca²⁺ flux was assessed in separate models. Water evaporation from beakers over the course of the experiment was small (0.5–1.5 ml in 7 h) and did not affect our measures of ion flux. Post hoc analyses for all ion flux experiments were conducted with the emmeans package (https://CRAN.R-project.org/package=emmeans) using the Tukey method of P-value adjustment for multiple comparisons. All data reported are estimated marginal means adjusted for the effect of the covariate.

Differences in target gene expression between treatment groups was analysed by comparing ΔCt values in analysis of variance models using the car package (Fox and Weisberg, 2018). A two-way analysis of variance model was fitted using acclimation [Ca²⁺]_e and pH as independent variables. This model was fitted for all genes of interest. Post hoc analyses were conducted using Tukey's honestly significant difference test for multiple comparisons.

Baseline net Na⁺ and Ca²⁺ flux in control larvae (pH 6.5, 50 μmol l⁻¹ [Ca²⁺]_e acclimated) at pH 6.5 was close to zero, although there was a small but significant net loss of Na⁺ (t₅=−5.21, P<0.01; Fig. 1). Net Ca²⁺ flux in control larvae was not significantly different from zero. There was a significant interaction between pH and [Ca²⁺]_e on net Na⁺ efflux in L. terraereginae larvae (F_29,2=6.18, P<0.01). Irrespective of [Ca²⁺]_e, all larvae exposed acutely to pH 3.5 water experienced a substantial net loss of Na⁺, indicating that Na⁺ efflux rates were considerably higher than Na⁺ uptake rates. The magnitude of the effect of acute low pH exposure on Na⁺ flux was greatest in larvae exposed to 5 μmol l⁻¹ [Ca²⁺]_e. There was less effect of low pH on Na⁺ efflux in larvae exposed to 50 μmol l⁻¹ (t₂₉=3.2, P<0.01) and 250 μmol l⁻¹ [Ca²⁺]_e (t₂₉=2.56, P=0.041). There was no effect of pH or [Ca²⁺]_e on net Ca²⁺ flux.

Relative to larvae simultaneously exposed to low pH and acute alterations to [Ca²⁺]_e, 4 weeks of acclimation to both low and high calcium levels reduced the impact of acute low pH exposure on net Na⁺ flux (F_19,1=9.59, P<0.01; Fig. 2). Similarly, there was a significant effect of acclimation to 250 μmol l⁻¹ [Ca²⁺]_e on net Ca²⁺ flux, with larvae acclimated to high [Ca²⁺]_e experiencing a significant net Ca²⁺ influx compared with larvae in 50 μmol l⁻¹ [Ca²⁺]_e (F_19,1=12.56, P<0.01).

In L. terraereginae larvae reared from hatching at pH 3.5 and with 50 μmol l⁻¹ [Ca²⁺]_e ASW, baseline net Na⁺ efflux was small but significantly lower than zero (t₅=−4.1833, P<0.01; Fig. 3). Larvae had a baseline net Ca²⁺ influx that was slightly but significantly higher than zero (t₅=2.7916, P=0.038). Acute exposure of L. terraereginae larvae to RR resulted in a substantial increase in both net Na⁺ efflux (F_9,1=71.174, P≤0.001) and net Ca²⁺ efflux (F_9,1=38.14, P<0.001) over the 6 h exposure period.

The expression of key Ca²⁺ transport proteins and E-cadherin in the gills of L. terraereginae was compared across larvae reared at both pH 6.5 and pH 3.5 in low, moderate or high [Ca²⁺]_e. There was no significant effect of acclimation pH or [Ca²⁺]_e on the gene expression of branchial ECaC, PMCA or NCX channels (Fig. 4). However, branchial calbindin gene expression was significantly higher in larvae reared at pH 3.5 versus pH 6.5 (F_1,24=5.640, P=0.026) but was not influenced by [Ca²⁺]_e.

Highly acidic waters pose a major threat to transcellular Ca²⁺ uptake pathways, which may play a role in the acute and potentially fatal branchial Na⁺ loss experienced by acid-sensitive aquatic animals exposed to low pH. Conversely, acid-tolerant animals may employ a suite of mechanisms that enable them to protect Ca²⁺ uptake capacity, which in turn allows them to resist the negative effects of low pH on junctional integrity. Consistent with our hypothesis, acute exposure to low pH water in the presence of low [Ca²⁺]_e increased net Na⁺ efflux, but not net Ca²⁺ flux in acid-tolerant L. terraereginae larvae. Provision of additional [Ca²⁺]_e reduced net Na⁺ efflux rates, but this effect was saturable. Acclimation to both low and high [Ca²⁺]_e improved the resistance of larvae to Na⁺ efflux at low pH and resulted in an increased net Ca²⁺ influx. Inhibition of apical epithelial calcium uptake by RR resulted in the complete loss of tolerance to low pH in larvae acclimated to low pH water, consistent with our hypothesis that acclimation to low pH involves the protection of Ca²⁺ uptake capacity. Acclimation to low pH water increased branchial gene expression of the intracellular Ca²⁺ transport protein calbindin independent of [Ca²⁺]_e. Given that 90% of the transcellular Ca²⁺ flux is transported by calbindin (Bronner, 2001), and the maximum Ca²⁺ flux through rat duodenal cells is a linear function of the cellular calbindin content (Bronner et al., 1986), we contend that the finding of calbindin upregulation at low pH is consistent with a role for increased transcellular Ca²⁺ trafficking in the tolerance of low pH water. These results establish a potential role for the security of Ca²⁺ uptake capacity in the tolerance of L. terraereginae larvae living in highly acidic waters.

Acid-naive L. terraereginae larvae reared at circumneutral pH and acutely exposed to low pH had a greater rate of net Na⁺ efflux than larvae reared and tested at pH 6.5, consistent with an acute negative effect of low pH on intercellular junction integrity. The high rate of net Na⁺ efflux was substantially reduced in animals acclimated to low pH, consistent with an acid-tolerant phenotype. In acid-naive L. terraereginae, the high rate of net Na⁺ efflux was exacerbated in the presence of low [Ca²⁺]_e, indicating a protective effect of [Ca²⁺]_e on epithelial junction integrity. This is consistent with previous studies showing that [Ca²⁺]_e plays a significant role in determining net Na⁺ efflux rates and mortality (acid tolerance) with low pH exposure in a range of fish and amphibian larvae (Brown, 1981, 1982, 1983; Cummins, 1988; Freda and Dunson, 1984; Freda et al., 1991; Gascon et al., 1987; Gonzalez and Dunson, 1989; Gonzalez et al., 1998; Kullberg et al., 1993; Kumai et al., 2011; McDonald and Rogano, 1986; McDonald et al., 1983; Meyer et al., 2010; Riesch et al., 2015; Wright and Snekvik, 1978). However, environmental Ca²⁺ was only beneficial for controlling Na⁺ efflux up to a point: exposure of larvae to 250 μmol l⁻¹ [Ca²⁺]_e did not further reduce net Na⁺ flux beyond that of larvae exposed to control (50 μmol l⁻¹) levels. This may indicate the influence of elevated Ca²⁺ on Na⁺ efflux is saturable, and 50 μmol l⁻¹ [Ca²⁺]_e completely saturates the gill epithelium. This is consistent with the findings of Meyer (2004), who demonstrated that 80 μmol l⁻¹ [Ca²⁺]_e reduced Na⁺ efflux during low pH exposure in L. terraereginae larvae, but raising [Ca²⁺]_e to 400 μmol l⁻¹ had no further effect. Caution must be taken in extrapolating this finding directly to the natural setting, where dissolved organic compounds may alter the effects of H⁺ on the ion transport processes (Picker et al., 1993; Wood et al., 2003).

The protective effect of [Ca²⁺]_e on branchial junction permeability has been attributed to extracellular Ca²⁺ bound to the gill epithelium, specifically to the TJ (Freda and McDonald, 1988; Gonzalez and Dunson, 1989; McDonald, 1983; McWilliams, 1983; Reid et al., 1991; Yu et al., 2010). However, these studies did not account for the possibility that low pH affects junctional stability by impairing transcellular Ca²⁺ uptake pathways. Exposure of L. terraereginae larvae to water containing 10 μmol l⁻¹ RR, a potent ECaC channel inhibitor (Nilius et al., 2001), resulted in a large (7500–17,500 nmol h⁻¹) increase in net Na⁺ efflux. This suggests that the inhibition of the transcellular Ca²⁺ uptake pathway has an immediate and severe effect on junctional stability, leading to increased junctional permeability to Na⁺. Similar findings have been observed in studies of goldfish and tetras exposed to La³⁺, a Ca²⁺ analogue that, unlike RR, also has the capacity to penetrate TJs and interact with AJs directly (Eddy and Bath, 1979; Gonzalez et al., 1997; Lacaz-Vieira and Marques, 2004). The inhibition of transcellular Ca²⁺ uptake using RR has not been performed previously on whole animals but does show that inhibition of Ca²⁺ uptake can have catastrophic impacts on homeostasis, consistent with the loss of junctional stability. We posit that the inhibition of apical Ca²⁺ uptake disrupts AJ stability, with consequences for the maintenance of TJ stability, and that any effect of acid on net Na⁺ flux is at least in part mediated by calcium ions. This is consistent with the finding that Ca²⁺ channel inhibitors increased amphibian embryo mortality at low pH, but Na⁺ channel blockers did not (Shu et al., 2015). As ECaC activity is inhibited by low pH (Vennekens et al., 2001), we propose that a loss of Ca²⁺ uptake capacity following acute exposure to low pH may underpin the loss of junctional stability and resulting Na⁺ efflux in acid-sensitive organisms. Conversely, adaptations that counter ECaC inhibition at low pH may protect Ca²⁺ uptake capacity in acidophilic species. While the effects of RR on Na⁺ efflux in acid-acclimated larvae were rapid and extreme, an investigation of its effects on junctional morphology would be needed to demonstrate that high Na⁺ losses were the result of junctional disruption and not the inhibition of other Na⁺ and Ca²⁺ transport pathways. While RR is a potent inhibitor of ECaC, it can also affect other Ca²⁺ transport proteins which may otherwise affect the maintenance of ion balance (Hajnóczky et al., 2006; Vincent and Duncton, 2011), or act as a systemic toxicant.

Unlike Na⁺ flux, acute exposure to low pH water did not affect net Ca²⁺ flux when larvae were acutely exposed to high or control [Ca²⁺]_e. In both circumneutral and low pH water, Ca²⁺ flux was not significantly different from zero, suggesting that rates of influx and efflux were balanced. This was unexpected as low pH has been shown to inhibit Ca²⁺ uptake (ECaC activity) in multiple cell lines (Bindels et al., 1994; Hoenderop et al., 1999; Vennekens et al., 2001). If low pH also inhibited ECaC activity in L. terraereginae, it should have manifested as a net increase in the rate of Ca²⁺ loss. In fact, L. terraereginae larvae reared in high [Ca²⁺]_e and acutely exposed to low pH experienced a net Ca²⁺ influx (∼250 nmol h⁻¹) compared with animals reared at control [Ca²⁺]_e levels. In low pH reared larvae, there was an apparent net increase in Ca²⁺ uptake. This influx suggests that ECaC in the branchial epithelium of L. terraereginae larvae is not substantially inhibited by protonation, and that transcellular environmental Ca²⁺ uptake is maintained at low pH and may play a role in facilitating acid tolerance in L. terraereginae. This may highlight an adaptation for the prevention of Ca²⁺ uptake inhibition at low pH and could be linked to the expression of a less pH sensitive ECaC isoform. Consistent with this hypothesis, inhibition of apical Ca²⁺ uptake by RR did result in a large increase in net Ca²⁺ efflux in larvae. Clearly, the disturbance of transcellular Ca²⁺ uptake has serious implications for the maintenance of transepithelial resistance via the loss of junction stability at low pH.

The maintenance of transcellular Ca²⁺ transport is potentially important in promoting acid tolerance of acidophilic animals. Ca²⁺ might influence tolerance of low pH via association with the Ca²⁺-dependent AJ, which is directly responsible for the stability of the TJ and thus epithelial permeability, a major factor in preventing Na⁺ loss at low pH in acidophilic species (Gumbiner et al., 1988; Kwong et al., 2014; Watabe-Uchida et al., 1998). Lowering intracellular [Ca²⁺] in MDCK cells has been shown to interfere with the formation of TJs (Stuart et al., 1996). Ca²⁺ also has many signalling functions such as hormone regulation (Clapham, 2007; D'Souza-Li, 2006), which may potentially alter the expression of genes involved in maintaining junctional integrity at low pH. Calbindin and other intracellular Ca²⁺-binding proteins function to buffer intracellular Ca²⁺ concentration by facilitating the basolateral extrusion of Ca²⁺ taken up across the apical membrane (Christakos et al., 1989). The finding that calbindin mRNA was upregulated in L. terraereginae larvae reared at pH 3.5 is suggestive of increased transcellular Ca²⁺ movement in the gill epithelium of L. terraereginae larvae acclimated to low pH. Interestingly, cytosolic Ca²⁺ is critical in regulating ECaC activity (Hoenderop et al., 1999). This finding is consistent with the idea that transcellular Ca²⁺ transport is involved in acid tolerance to some degree and that Ca²⁺ uptake is not inhibited by low pH in L. terraereginae larvae.

In contrast to our hypotheses, mRNA expression of the key transcellular Ca²⁺ transport proteins ECaC, NCX and PMCA was not influenced by environmental pH or [Ca²⁺]_e in L. terraereginae larvae despite the observation that acclimation to low pH was accompanied by an apparent net increase in Ca²⁺ uptake. If this effect was indeed the result of an increase in Ca²⁺ uptake and not a reduction in efflux, then it is possible that it was facilitated by an increase in the activity of the existing channels as opposed to the de novo production of more channels. Likewise, the lack of increase in E-cadherin mRNA expression with chronic low pH exposure suggests that E-cadherin function was unaffected by low pH. E-cadherins bind Ca²⁺ from the extracellular environment; our data suggest that maintenance of the transcellular Ca²⁺ transport pathway allows the maintenance of favourable Ca²⁺ concentrations in the extracellular space to prevent E-cadherin disfunction. Consistent with this idea, increased expression of calbindin mRNA in acid-acclimated larvae provides some evidence that increased transcellular Ca²⁺ transport plays a role in promoting extreme acid tolerance. Increased Ca²⁺ uptake at the apical membrane may be evidenced by increased intracellular Ca²⁺ shuttling rates (and an associated increased abundance of calbindin proteins) to maintain Ca²⁺-dependent junction dynamics in low pH environments. The finding that low pH-acclimated larvae upregulated calbindin mRNA and had a net Ca²⁺ influx supports a role for increased transcellular [Ca²⁺]_e movement in facilitating acid tolerance in L. terraereginae larvae. However, mRNA expression levels do not always correlate well with actual levels of protein expression, so care must be taken when interpreting mRNA expression patterns in the absence of corresponding protein expression levels. Differential post-translational processing of mRNA and other factors can be responsible for the low correlation between an organism's transcriptome and its proteome (Ghazalpour et al., 2011; Marguerat et al., 2012). Although we cannot rule out a paracellular route for the uptake of Ca²⁺ into the extracellular space, studies in fish suggest that more than 97% of branchial (gills) Ca²⁺ uptake is active (transcellular) (Flik et al., 1995). The hormonal control of Ca²⁺ transport protein function and abundance is a potentially overlooked factor in understanding how branchial Ca²⁺ transport is influenced by extracellular pH and [Ca²⁺]_e in L. terraereginae.

This study showed that environmental Ca²⁺ has a protective effect on the control of Na⁺ efflux at low pH in L. terraereginae larvae. Furthermore, it demonstrated that larvae have a capacity for acclimation to low pH via changes in Na⁺ and Ca²⁺ flux, which appear to involve the transcellular pathway (i.e. increased calbindin mRNA upregulation). Given that L. terraereginae larvae can maintain or increase Ca²⁺ uptake at low pH, this suggests that protonation of the branchial epithelium likely does not outwardly inhibit Ca²⁺ uptake. However, we only examined net Na⁺ and Ca²⁺ flux, which does not reveal details about the behaviour of separate uptake and efflux pathways. For example, increases in efflux rates that are compensated for by commensurate increases in uptake rates may be missed by measuring just net ion flux. Radioactive isotopes or fluorescent ion analogues could be used to better resolve changes in influx and efflux pathways and how they contribute to net ion flux. Inhibition of apical Ca²⁺ uptake by RR supports a role for the maintenance of transcellular Ca²⁺ uptake in the control of branchial junction stability at low pH in L. terraereginae but does not reveal the specific site of action for Ca²⁺ in preventing Na⁺ loss at low pH. Using intracellular Ca²⁺ markers to track Ca²⁺ movement during acclimation to low pH may help to elucidate this mechanism. As L. terraereginae larvae are exceptionally acid tolerant, their ability to maintain Ca²⁺ uptake in very soft and acidic waters may be a unique adaptation. Comparing transcellular Ca²⁺ transport capabilities with those of other acid-sensitive species might reveal the mechanistic adaptations employed by L. terraereginae in the maintenance of epithelial stability at low pH. The current study highlights a role for transcellular Ca²⁺ transport and the prevention of Ca²⁺ uptake inhibition by low pH in the extreme acid tolerance of L. terraereginae larvae.

The authors thank Dr Edward Meyer for useful discussions during the project, and Dr Nicholas Wu and Callum McKercher for assistance with gene expression studies.