Modulation of NaCl absorption by [HCO₃⁻] in the marine teleost intestine is mediated by soluble adenylyl cyclase

Abstract

Intestinal HCO₃⁻ secretion and NaCl absorption are essential for counteracting dehydration in marine teleost fish. We investigated how these two processes are coordinated in toadfish. HCO₃⁻ stimulated a luminal positive short-circuit current (I_sc) in intestine mounted in Ussing chamber, bathed with the same saline solution on the external and internal sides of the epithelium. The I_sc increased proportionally to the [HCO₃⁻] in the bath up to 80 mM NaHCO₃, and it did not occur when NaHCO₃ was replaced with Na⁺-gluconate or with NaHCO₃ in Cl⁻-free saline. HCO₃⁻ (20 mM) induced a ∼2.5-fold stimulation of I_sc, and this [HCO₃⁻] was used in all subsequent experiments. The HCO₃⁻-stimulated I_sc was prevented or abolished by apical application of 10 μM bumetanide (a specific inhibitor of NKCC) and by 30 μM 4-catechol estrogen [CE; an inhibitor of soluble adenylyl cyclase (sAC)]. The inhibitory effects of bumetanide and CE were not additive. The HCO₃⁻-stimulated I_sc was prevented by apical bafilomycin (1 μM) and etoxolamide (1 mM), indicating involvement of V-H⁺-ATPase and carbonic anhydrases, respectively. Immunohistochemistry and Western blot analysis confirmed the presence of an NKCC2-like protein in the apical membrane and subapical area of epithelial intestinal cells, of Na⁺/K⁺-ATPase in basolateral membranes, and of an sAC-like protein in the cytoplasm. We propose that sAC regulates NKCC activity in response to luminal HCO₃⁻, and that V-H⁺-ATPase and intracellular carbonic anhydrase are essential for transducing luminal HCO₃⁻ into the cell by CO₂/HCO₃⁻ hydration/dehydration. This mechanism putatively coordinates HCO₃⁻ secretion with NaCl and water absorption in toadfish intestine.

the intestine of marine fish absorbs NaCl and water to compensate for dehydration in a hyperosmotic environment. This hypoosmoregulatory mechanism relies in part on HCO₃⁻ secretion, in exchange for Cl⁻ uptake, resulting in intestinal luminal concentration of HCO₃⁻ {[HCO₃⁻]} of > 100 mM (20, 65). The resulting alkaline conditions lead to the formation of Ca²⁺ and Mg²⁺ carbonate precipitates and a reduction of osmotic pressure (20, 65). Water is absorbed following transepithelial NaCl absorption, and the carbonate precipitates are expelled with the feces. Intestinal carbonate precipitation takes place in all marine bony fish species investigated so far, and it has been argued that the combined piscine carbonates output contributes up to 15% of total oceanic carbonate production (64).

In the intestine of marine fish, the major pathway for NaCl absorption from the lumen and into the intestinal cells is through apical Na⁺-K⁺-2Cl⁻ cotransporters (NKCC), also known as bumetanide-sensitive cotransporter. In mammalian epithelia, the apical NKCC isoform is predominantly present in the thick ascending loop of Henle, and it is termed NKCC2 (10, 19, 28, 32, 39, 48, 49, 66). Several studies have confirmed the involvement of an NKCC2-like protein in fish intestine. The evidence includes a mucosal-positive transepithelial potential difference under in vivo-like conditions, related to Cl⁻- dependent short-circuit current (I_sc) (1), codependence of Na⁺ and Cl⁻ transport (12, 15, 38), dependence on an apical K⁺ conductance (14, 38), and, importantly, sensitivity to the loop diuretic bumetanide (15, 38, 42, 43). The presence of intestinal NKCC2 was confirmed by RT-PCR in seawater eels (7) and by immunohistochemistry in mudskipper from brackish water (65), killifish (34), and sea bass (31). Like many epithelia, the driving force for NaCl absorption is provided by basolateral Na⁺/K⁺-ATPases.

An additional pathway for apical Cl⁻ entry is apical anion exchange, which can, in some cases, account for as much as 70% of the overall Cl⁻ absorption by the intestinal epithelium (20). Interestingly, elevation of luminal NaCl results in a reduction of HCO₃⁻ secretion, despite the increased concentration of luminal substrate (Cl⁻) for the process (23). This response (reduction in anion exchange) to elevated luminal NaCl, which would favor salt absorption by cotransport pathways, suggests that coordination of the multiple Cl⁻ uptake pathways takes place in the marine teleost intestine.

Intestinal bicarbonate secretion and NaCl and water absorption are regulated by several factors. In the winter flounder, atrial natriuretic factor and its downstream intracellular second messenger cGMP inhibit the ion-absorbing I_sc (41, 42), and they also reduce the net inward flux of Na⁺ and Cl⁻ (41). These effects may be related to the inhibition of NaCl and water absorption when euryhaline fish enter freshwater (15, 37), but they also could be involved in acclimation to seawater in eel (58). At least part of the cGMP effect is attributable to inhibition of apical NKCC and K⁺ channels activity (51), a regulation that also takes place in the mammalian thick ascending loop of Henle (2).

Addition of 8-Br-cAMP (an exogenous, phosphodiesterase-resistant cAMP analog) also affects I_sc, but the inhibition is incomplete and the underlying cellular mechanisms are poorly understood. Some studies have applied 8-Br-cAMP in combination with theophylline, and they have found robust inhibition of I_sc and Na⁺ and Cl⁻ absorption (13, 15). However, theophylline is a nonspecific phosphodiesterase inhibitor that prevents both cAMP and cGMP degradation. When applied alone, 8Br-cAMP increased both Cl⁻ absorption and secretion in a manner that resulted in a reduction (although not statistically significant) of net Cl⁻ absorption, together with inhibitions of the outside-positive I_sc and net K⁺ secretion and an increase in transepithelial conductance (G_te) (51). Thus, it appears that regulation of ion transport across fish intestine by cAMP involves multiple, possibly opposing, pathways.

The anion bicarbonate is an overlooked factor that regulates ion absorption in fish intestine. Reducing the saline's [HCO₃⁻] from 20 mM to 5 mM induces an ∼60% reduction of Cl⁻ absorption and I_sc, a 30% reduction of Na⁺ absorption, and a significant increase in G_te (12, 13). HCO₃⁻ stimulation of ion absorption has been attributed to increases in pH (12), which stimulates NaCl and water absorption on its own (44). However, it is possible that both pH and HCO₃⁻ have synergistic stimulatory effects via independent mechanisms not yet identified.

One of such potential mechanisms is the newly identified soluble adenylyl cylase (sAC). sAC is an evolutionary conserved signaling enzyme whose activity is directly modulated by HCO₃⁻ to produce cAMP (3, 6, 63). Unlike the classic, hormonal- and G protein-regulated transmembrane ACs (tmACs), sAC is present in the cell cytoplasm and in organelles (67, 68). sAC has been proposed to be a universal sensor of HCO₃⁻, and indirectly, of pH and Pco₂ (6). Furthermore, sAC has recently been shown to be present in an elasmobranch fish, the dogfish Squalus acanthias, where it is essential for activating branchial H⁺ absorption and HCO₃⁻ secretion to counteract blood alkalosis (63). Like mammalian and cyanobacterial sACs (6, 25, 55), dogfish sAC (dfsAC) is bicarbonate responsive and inhibited by two unrelated small molecules, 4-catechol estrogen (CE) and KH7 (63).

In the present paper, we have investigated whether sAC is involved in regulating NaCl absorption across the intestinal epithelium of a marine fish in response to elevated HCO₃⁻, a feature that might explain previous results and may also apply to other NaCl-absorbing epithelia.

METHODS

Animals.

Gulf toadfish (Opsanus beta) 20–30 g in size, from Biscayne Bay, FL, were obtained from shrimp fishermen during the winter and spring of 2008 and 2009 and transferred to holding facilities at the University of Miami Rosenstiel School of Marine and Atmospheric Sciences. Procedures were approved by the University of Miami Animal Care and Use Committee. On arrival, the fish received a prophylactic treatment for ectoparasites, as previously described (35). Fish were provided short lengths of polyvinylchloride tubing as shelters in 80-liter glass aquaria that received a continuous flow of filtered seawater (salinity 27–30 ppt, 22–25°C, from Bear Cut, Florida). Fish were fed frozen squid twice weekly, but food was withheld 48 h before experimentation. Only tissues from males and nonovigerous females were used for the procedures outlined below.

I_sc and G_te measurements.

Toadfish were anaesthetized with MS-222 (0.5 g/l; Argent Laboratories, Redmond, WA), and the spinal chord was severed. After opening the abdominal cavity, the anterior intestine (from the pyloric sphincter to ∼1.5 cm distally) was dissected, cut open, and mounted in a tissue holder that exposed 0.7 cm⁻² epithelial surface area (not considering villi and microvilli). The tissue-containing holder was inserted in an Ussing chamber (model P2300; Physiological Instruments, San Diego, CA) containing 2 ml of pregassed saline. During the experiments, gas mixes were delivered to each hemichamber through airlifts to maintain O₂, CO₂ and HCO₃⁻ levels constant and promote mixing in the bathing solutions. The Ussing chambers were placed in chamber holders maintained at 25 ± 1°C throughout experimentation. Current and voltage electrodes (model P2020-S; Physiological Instruments) connected to amplifiers (model VCC600; Physiological Instruments) recorded the I_sc under voltage-clamp conditions (0 mV). At 60-s intervals, the G_te was calculated from an imposed 1-mV voltage pulse from the mucosal to the serosal side and the resulting current deflection. I_sc measurements were logged on a personal computer using a data acquisition system (BIOPAC systems interface with AcqKnowledge software, version 3.8.1).

The bathing saline was applied symmetrically into both hemichambers, and its composition was (in mM) 151 NaCl, 3 KCl, 0.88 MgSO₄, 4.6 Na₂HPO₄, 0.48 K₂HPO₄, 1 CaCl₂, 11 HEPES-free acid, 11 HEPES-free salt and 4.5 urea. All saline solutions were bubbled with 100% O₂ or 0.3% CO₂-99.7 O₂ before adjusting osmolarity to 330 mOsm and pH to 7.8. Serosal saline also contained 3 mM glucose. Mounted epithelia were initially bathed in this saline (notice the absence of HCO₃⁻) gassed with 100% O₂ until a stable I_sc was reached. To elevate [HCO₃⁻], different volumes of a 1 M NaHCO₃ stock solution were pipetted into both hemichambers to avoid diffusive fluxes, and the gas was changed to 0.3% CO₂-99.7 O₂. Pharmacological inhibitors were injected into the luminal bath from concentrated stocks in DMSO, and the added volume of drug was 2 μl in all cases. DMSO alone had no effect on I_sc or G_te. After each change of saline or addition of drug, it typically took 15–20 min for I_sc and G_te to reach new stable values. The I_sc and G_te after at least additional 20 min of stable readings were used for statistical analyses.

Western blot analysis and immunofluorescence.

The following antibodies were used: mouse monoclonal α5 (against α₁- and β₁-subunits of chicken Na⁺/K⁺-ATPase; 0.01 μg/ml in Western blot analysis, 0.05 μg/ml in immunofluorescence), mouse monoclonal T4 (against the carboxy-term of human NKCC Met-902 to Ser-1212; 0.5 μg/ml in Western blot analysis, 0.2 μg/ml in immunofluorescence), rabbit polyclonal dfsAC antibodies [against an epitope in dfsAC: INNEFRNYQGRINKC (63); 0.15 μg/m1 in Western blot analysis, 70 μg/ml in immunofluorescence], and mouse monoclonal R7 [against an epitope in mammalian sAC: EIERIPDQRAVKV (27); 0.1 μg/ml in Western blot analysis, 20 μg/ml in immunofluorescence]. The α5 and T4 antibodies developed by Douglas Fambourgh (John Hopkins University, Baltimore, MD) and Christian Lytle (University of California Riverside, Riverside, CA), respectively, were obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the National Institute of Child Health and Human Development and maintained by Department of Biological Sciences, The University of Iowa, Iowa City, IA.

For Western blot analysis, the anterior intestine was dissected, flash frozen in liquid N₂ and stored at −80°C. Samples were homogenized in lysis buffer (in mM: 150 NaCl, 5 EDTA, 50 Tris, 1 PMSF, 1 DTT, and 10 μg/ml aprotinin, 10 μg/ml leupeptin, pH 7.5) by using a glass Dounce homogenizer and were spun down (10 min, 16,000 g, 4°C). Approximately 10 μl of the supernatant (15 μg total protein) was combined with 2× Laemmli buffer (30), heated (70°C, 10 min), and used in PAGE-Western blot analysis. Purified recombinant rat sAC protein, produced as previously described (6) was used as control for the R7 antibody.

For immunofluorescence, the anterior intestine was dissected, cut into ∼2 mm sections, and immersed in fixative (4% paraformaldehyde in 0.1 M PBS) overnight at 4°C. The fixative was replaced by ice-cold 15% (wt/vol) sucrose, incubated for 4 h at 4°C, then replaced by 30% (wt/vol) sucrose, and stored at 4°C. Fixed tissues were sectioned at 4–8 μm using a cryostat (−15°C) and immunostained. Sections were treated with 1% SDS in PBS, followed by 2× 5 min washes in PBS and blocking in 2% normal goat serum in PBS for 1 h at room temperature. Incubation with the antibodies (in 2% normal goat serum in PBS) was performed overnight at 4°C. After 3× 5 min washes in PBS, the secondary antibodies Alexa Fluor 488-conjugated anti-mouse IgG or Alexa Fluor 568-conjugated anti-rabbit IgG were applied for 1 h at room temperature and subsequently washed 3× 5 min in PBS. For double labeling, sections were first labeled with dfsAC (overnight) and fluorescent anti-rabbit IgG (1 h), followed by blocking and 1-h incubation at room temperature (T4) or overnight incubation at 4^°C (R7) and anti-mouse fluorescent IgG (1-h room temperature) with 3× 5 min washes in PBS in between. All sections were mounted using 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI)-containing medium (Vector Labs).

Statistical analyses.

All data sets were analyzed using one-way repeated-measures ANOVA, followed by Tukey's multiple-comparisons test using GraphPad Prism version 4.0a for Macintosh (GraphPad Software, San Diego, CA). Differences between means were considered statistically significant when P < 0.05.

RESULTS

Bicarbonate stimulates I_sc and G_te.

Increasing the [NaHCO₃] in the bathing solutions resulted in dose-dependent increases in the outside-positive I_sc and in G_te (Fig. 1A). These effects were not due to the extra Na⁺ or to increases in osmotic pressure, since Na⁺-gluconate has no effect on I_sc or G_te (Fig. 1B). Therefore, the observed stimulations were due to HCO₃⁻ itself and/or to the associated increase in pH. For subsequent experiments, 20 mM NaHCO₃ was used because it induced significant changes on I_sc and G_te and it was within or near physiological values for both the luminal and the serosal sides. Addition of 20 mM NaHCO₃ raised the saline's pH from 7.80 to 8.15. Since a pH of 8.15 did not change I_sc significantly and it only produced a small increase in G_te (Fig. 1C), the HCO₃⁻ stimulation of I_sc by 20 mM HCO₃⁻ is independent of pH.

Fig. 1.

HCO₃⁻ stimulation of short-circuit current (I_sc) and transepithelial conductance (G_te) across toadfish anterior intestine: dose-response to NaHCO₃ (A), dose-response to Na⁺-gluconate (B), and effect of 20 mM NaHCO₃ and the equivalent change in pH (C) . Bic, bicarbonate. Letters indicate different levels of statistical significance (repeated-measures ANOVA, Tukey's multiple-comparisons test, P < 0.05, n = 4–5).

sAC in toadfish intestine.

We hypothesized that sAC, a recently discovered HCO₃⁻-responsive signaling enzyme, mediates the HCO₃⁻-stimulation of I_sc in toadfish intestine. Polyclonal antibodies against dfsAC recognized a single specific band of ∼100 kDa in anterior intestine samples from toadfish (Fig. 2A), which was absent in the peptide preabsorption and preimmune serum controls (not shown). A mouse monoclonal antibody (R7), raised against a different epitope in mammalian sAC, also recognized a specific ∼100 kDa band (Fig. 2B). sAC in the toadfish intestine was detected by immunofluorescence using dfsAC (Fig. 3, A, C, and E) and R7 (Fig. 3, D and E) antibodies. sAC immunoreactivity was present throughout the epithelial intestinal cells but not directly on the apical membrane. Incubation with preimmune serum (Fig. 3B), mouse IgG, or dfsAC antibodies preabsorbed with specific peptide (not shown) resulted in no signal.

Fig. 2.

Soluble adenylyl cyclase (sAC) in toadfish (TF) intestine. Western blot analysis using antibodies (Ab) against dogfish sAC (dfsAC; A) and mammalian sAC mouse monoclonal antibody R7 (B). A purified, 50-kDa isoform of rat sAC (sAC_t) was used as control for R7. The second lane in each panel shows preabsorption with peptides (pept) corresponding with the respective epitopes.

Fig. 3.

sAC in toadfish intestine. Immunofluorescence detection using antibodies against dfsAC (A and C; red) and mammalian sAC (D; R7, green). B: incubation with dfsAC preimmune serum. Brightfield and fluorescent pictures (A and B) are overlaid. E: double staining using dfsAC and R7 antibodies. Yellow indicates colocalization of signals. Nuclei are stained with 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI; blue) (A, B, and E). Scale bars: 10 μM. F: sequences and positions of antigen peptides in rat and dfsAC. Conserved amino acid residues are shaded in black, similar residues are in grey.

Involvement of sAC in the stimulation of I_sc by HCO₃⁻.

Micromolar concentrations of CE specifically inhibits all known sAC and sAC-related enzymes, including CyaC from cyanobacteria (55), dfsAC (63), and mammalian sAC (47, 55). CE (30 μM) significantly reduced the I_sc in the absence of HCO₃⁻ in the bathing saline and slightly increased G_te (Fig. 4A). Subsequent addition of 20 mM HCO₃⁻ in the presence of CE failed to stimulate I_sc, but G_te increased similar to controls; the CE inhibition of I_sc was only partially reversible (Fig. 4A). Application of CE after I_sc had been stimulated by 20 mM HCO₃⁻ completely abolished the HCO₃⁻ stimulation (Fig. 4B). 17β-estradiol (30 μM), a compound related to CE that is inert toward sAC-like cyclases (47, 55) did not significantly affect the HCO₃⁻-stimulated I_sc (Fig. 4C). Although similar concentrations of KH7, the most specific sAC inhibitor available (25), failed to affect I_sc, most KH7 precipitated immediately after entering the bathing saline. Precipitation was exacerbated by the continuous gas bubbling necessary during these experiments. To increase the KH7 concentration ([KH7]) in solution, the nominal [KH7] was raised to 200 μM, which resulted in significant inhibition of the HCO₃⁻-stimulated I_sc (Fig. 4D). The real [KH7] in solution during these experiments is unknown, but based on the visible precipitate it certainly was much lower than 200 μM.

Fig. 4.

Involvement of sAC in the HCO₃⁻ stimulation (20 mM) of I_sc and G_te across toadfish anterior intestine. A: sAC inhibitor 4-catechol estrogen (CE; 30 μM) prevents the HCO₃⁻ stimulation of I_sc. B: CE inhibits the I_sc previously stimulated by HCO₃⁻. C: 17β-estradiol (Est) does not affect the stimulated I_sc. D: KH7 (200 μM), another sAC inhibitor, also inhibits the I_sc previously stimulated by HCO₃⁻. The real KH7 concentration is not known because KH7 was only partially soluble in the bathing saline. E: 2′5′-dideoxyadenosine (Dideox, 50 μM), a specific inhibitor of transmembrane adenylyl cyclases, has the opposite effect on I_sc compared with CE and KH7. Letters indicate different levels of statistical significance (repeated-measures ANOVA, Tukey's multiple-comparisons test, P < 0.05; n = 5–6).

Although the specificity of CE for sAC is much higher than for tmACs, high concentrations of CE can also inhibit tmACs (55). To test the potential involvement of tmACs, we applied 2′5′-dideoxyadenosine (50 μM), which inhibits tmAC-dependent cellular processes without affecting sAC (50). Unlike CE, 2′5′-dideoxyadenosine increased the outside-positive HCO₃⁻-stimulated I_sc (Fig. 4E). These results suggest that the CE effect is not due to cross-reaction with tmACs.

NKCC and Na⁺-K⁺-ATPase in toadfish intestine.

Two major players in fish intestinal NaCl absorption are apical NKCC and basolateral Na⁺/K⁺-ATPase. However, these transporters have never been shown to be present in toadfish intestine. T4 (against human NKCC) and α5 (against avian Na⁺/K⁺-ATPase) monoclonal antibodies recognized single bands of ∼180 and ∼100 kDa, respectively (Fig. 5, A and D), which match previously published results from fish (9, 26, 60, 61). In immunostained sections, both transporters were found in almost all epithelial cells, with the exception of larger, round cells that are presumably goblet (mucus secreting) cells. NKCC was predominantly present in the apical membrane and subapical area (Fig. 5E), indicative of an NKCC2 isoform. Weaker NKCC immunostaining was also found throughout the cell (presumably an NKCC1 isoform), but the signal was only observed after increasing the sensitivity of the detector significantly. Na⁺/K⁺-ATPase was present throughout the cells, presumably in basolateral infoldings typical of marine fish intestine (12, 40, 53); Na⁺/K⁺-ATPase immunoreactivity was absent from the apical membrane (Fig. 5, B and C). Because NKCC and sAC immunoreactivities overlapped in the subapical region (Fig. 5F), we next tested whether the HCO₃⁻ and sAC stimulation of I_sc was due to activation of NKCC.

Fig. 5.

Na⁺/K⁺-ATPase (NKA), Na⁺-K⁺-2Cl⁻ cotransporter (NKCC) and sAC immunoreactivity in toadfish intestine. NKA: Western blot analysis (n = 3) (A); immunofluorescence (B) and brightfield overlay (C). Nuclei have been stained with DAPI (blue). Notice that NKA is present in basolateral infoldings and absent in the apical membrane. NKCC: Western blot analysis (n = 3) (D); immunofluorescence (E). Notice that NKCC signal is strongly apical and subapical. F: overlay of NKCC signal (green) and sAC signal (red). Scale bars: 10 μM.

Involvement of NKCC in the stimulation of I_sc by HCO₃⁻.

Sensitivity to bumetanide is a hallmark of NKCC (66). Apical application of bumetanide (10 μM) completely prevented (Fig. 6A) or abolished (Fig. 6B) the HCO₃⁻-stimulated I_sc. The inhibitory effects of bumetanide and of the sAC inhibitor CE were not additive, regardless of the order of application (Fig. 6, A–C), suggesting that sAC and NKCC belong in the same pathway. To confirm that the I_sc was indeed carried by Cl⁻ ions, we conducted experiments in chloride-free bathing saline (chlorides substituted by gluconates). In this saline, 20 mM HCO₃⁻ did not stimulate I_sc or G_te (Fig. 6D). Although 40 mM HCO₃⁻ did stimulate I_sc in Cl⁻-free conditions, the stimulation was minor, and it was not inhibited by bumetanide (Fig. 6D), indicating the effect is different from that studied in normal, Cl⁻-containing saline.

Fig. 6.

Involvement of NKCC in the HCO₃⁻-stimulation (20 mM) of I_sc and G_te across toadfish anterior intestine. A: NKCC inhibitor bumetanide (Bum; 10 μM) prevents the HCO₃⁻ stimulation of I_sc, the sAC inhibitor CE (30 μM) does not induce any further inhibition. B: bumetanide inhibits the I_sc previously stimulated by HCO₃⁻; CE does not have any further effect. C: same as in B, but the order of application of CE and bumetanide are reversed. D: in the absence of chloride ions in the bathing saline (replaced by gluconate), 20 mM NaHCO₃ does not stimulate I_sc or G_te. NaHCO₃ (40 mM) induced a small, significant stimulation but it was not dependent on NKCC. Letters indicate different levels of statistical significance (repeated-measures ANOVA, Tukey's multiple-comparisons test, P < 0.05; n = 5–6).

Involvement of V-H⁺-ATPase and carbonic anhydrase in the stimulation of I_sc by HCO₃⁻.

Apical addition of the V-H⁺-ATPase inhibitor bafilomycin (1 μM) prevented the HCO₃⁻-stimulated I_sc (Fig. 7A). Etoxolamide (1 mM), a permeable inhibitor of carbonic anhydrase, also prevented the HCO₃⁻-stimulated I_sc, and it additionally reduced the basal I_sc under HCO₃⁻-free conditions (Fig. 7B).

Fig. 7.

Involvement of V-H⁺-ATPase (VHA) and carbonic anhydrase (CA) in the HCO₃⁻ stimulation (20 mM) of I_sc and G_te across toadfish anterior intestine. A: VHA inhibitor bafilomycin (Baf; 1 μM) prevents the HCO₃⁻ stimulation of I_sc. VHA inhibition was not reversible. B: permeable CA inhibitor etoxolamide (Et; 1 mM) inhibits the basal I_sc and prevents the stimulation by HCO₃⁻. Bumetanide does not have any further effect. Letters indicate different levels of statistical significance (repeated-measures ANOVA, Tukey's multiple-comparisons test, P < 0.05; n = 5).

DISCUSSION

The intestine of marine fish is an excellent model to study epithelial NaCl absorption because of the homogeneity of cell types and its well-characterized electrophysiology. In fact, the luminal-positive I_sc across fish intestine has been established to be equivalent to the asymmetrical net Na⁺ and Cl⁻ absorption (reviewed in Ref. 37).

Our results demonstrate that the anion HCO₃⁻ stimulates NaCl absorption across the anterior intestine of toadfish, an effect that is independent of alkaline pH. Regulation of NaCl absorption by HCO₃⁻/pH in the marine fish intestine is not surprising considering the processes of carbonate precipitation and NaCl-driven water absorption. While a portion of the HCO₃⁻ secreted into the intestinal lumen by apical anion exchangers of the Slc26a6 family (29) precipitates as CaCaO₃ and MgCaO₃, there is still significant (up to 100 mM) HCO₃⁻ dissolved in the luminal fluids (65). Our proposed model for HCO₃⁻-stimulated NaCl absorption is depicted in Fig. 8. Part of the dissolved HCO₃⁻ combines with H⁺ pumped into the lumen by V⁺-H⁺-ATPases, which is known to be present in the apical membrane of toadfish intestinal cells (22). This reaction yields CO₂ and H₂O, and, while it might be catalyzed by extracellular carbonic anhydrase, the complete inhibition by apical bafilomycin suggests enzymatic catalysis is not necessary for the stimulation of I_sc. According to our proposed model, luminal CO₂ then diffuses inside the intestinal cells where it is hydrated back into H⁺ and HCO₃⁻ by cytoplasmic carbonic anhydrase. Interestingly, in intestinal cells of seawater-acclimated rainbow trout, cytoplasmic carbonic anhydrase is located in the subapical region (21), which is also where sAC is most abundantly present in toadfish intestinal cells. Therefore, it is likely that intracellular HCO₃⁻ generated by carbonic anhydrase activates sAC, which, in turn, activates apical NaCl absorption across NKCCs in a yet undetermined manner, but which should depend on cAMP. The driving force for NaCl absorption is provided by basolateral Na⁺/K⁺-ATPases, which we have confirmed to be at the toadfish intestinal epithelium by Western blot analysis and immunohistochemistry. The Na⁺/K⁺-ATPase is another potential target of sAC activity.

Fig. 8.

Model for the activation of I_sc by HCO₃⁻ in toadfish intestinal cells. The anion exchanger secrets intracellular HCO₃⁻ [produced by cellular metabolism and imported by NBC from the blood (59)] into the intestinal lumen. Luminal HCO₃⁻ results in Ca²⁺ and Mg²⁺ precipitation (not shown), a fraction of the excess HCO₃⁻ combines with H⁺ pumped by the VHA to form CO₂, which diffuses into the intestinal cells. CA_cyt hydrates CO₂ back into HCO₃⁻ and H⁺ (not shown). HCO₃⁻ activates sAC, which stimulates (+) the activity of apical NKCC to absorb NaCl via a cAMP-dependent mechanism yet to be determined. Putative apical K⁺ and basolateral Cl⁻ channels are not shown. CA_cyt, cytoplasmic carbonic anhydrase; NBC, Na⁺/HCO₃⁻ cotransporter.

sAC in toadfish intestine.

Based on immunodetection with two different antibodies, an sAC-like protein is present in the toadfish intestine. Unfortunately, we were unable to clone toadfish sAC by a variety of approaches, including bioinformatic methods and homology cloning using degenerate primers. Based on our experience working with mammalian (3, 11) and dfsAC (63), cloning the toadfish sAC ortholog is challenging because of the relatively low conservation of the sAC genes between species and the typically low sAC mRNA abundance. Nonetheless, immunodetection of a protein band of identical molecular size and colocalization in microscopy sections using two different antibodies, although heterologous, is good evidence for the presence of sAC in toadfish. The alternatives, that a protein unrelated to sAC contains the two different epitopes or that the antibodies recognize different proteins of identical molecular size, are highly unlikely.

To our knowledge, this is the first study to attempt selective inhibition of sAC and tmACs in an ion-transporting epithelium. Inhibition of sAC (with CE and KH7) had the opposite effect of inhibition of tmACs (with 2′5′-dideoxyadenosine). This finding has two profound implications. First, it suggests that cAMP signaling in toadfish intestinal cells occurs in defined intracellular microdomains (i.e., basolateral and apical membranes, cytoplasm, etc.) and that each of them may exert different, even opposing physiology. Second, it shows that studies using permeable cAMP analogs should be interpreted with caution, as the cAMP analogs are likely to have multiple effects on cellular physiology through at least two different cAMP signaling pathways. This situation is further complicated by the common use of nonspecific inhibitors of phosphodiesterases, which could not only prevent the degradation of cAMP from multiple sources, but also of cGMP.

Targets of the presumptive HCO₃⁻/sAC pathway.

The two most likely mechanisms of NKCC activation are PKA-mediated insertion of NKCC from subapical vesicles into the apical membrane and PKA-phosphorylation of NKCC already present at the apical membrane. There is abundant evidence supporting these two mechanisms, which are not mutually exclusive, in other systems. cAMP has been shown to promote the insertion of mammalian NKCC2 into the cell membrane both in heterologous expression experiments in Xenopus oocytes (36) and in the native thick ascending loop of Henle (4, 46). The cAMP effect has been traditionally linked to hormones, such as vasopressin (reviewed in Ref. 16), but the potential involvement of sAC has not yet been considered, likely due to this signaling enzyme being discovered recently. Notably, sAC mRNA is present in several NaCl and water absorbing mammalian epithelia, including small intestine (17), and sAC has been immunolocalized in the thick ascending loop of Henle in the kidney (47).

Vasopressin induces NKCC2 activation in at least two ways: increased translocation into the membrane and direct NKCC2 phosphorylation (18). However, a direct link between vasopressin, cAMP, PKA, and NKCC2 has not yet been established. Our immunohistochemical pictures from toadfish intestinal fragments (which were immediately fixed after dissection) clearly show that NKCC is present both in the apical membrane and in the subapical region of intestinal epithelial cells, leaving open both trafficking and direct phosporylation as possible NKCC activating mechanisms in response to HCO₃⁻ via sAC.

Alternatively or additionally, the pathway might involve activation of Na⁺/K⁺-ATPase. Because this enzyme is the main driving force behind NaCl absorption via NKCC (20, 23), a putative modulation of Na⁺/K⁺-ATPase by sAC would explain why sAC and NKCC inhibitors did not show additive effects. Interestingly, sAC has been shown to regulate Na⁺/K⁺-ATPase catalytic activity in kidney collecting duct cells (24).

Some of our experimental conditions, inhibitors, and results require a few clarifications. First, to avoid diffusive ion movement across intestinal segments mounted in Ussing chamber, it is essential to apply symmetrical conditions in the saline at the serosal and mucosal sides. We mimicked the composition of the serosal side (blood) found in live fish, reasoning that it is best for cell function. A [HCO₃⁻] of 80 mM, although normal for the intestinal lumen, is much higher than the normal [HCO₃⁻], pH, and osmolarity in the serosal side. Therefore, we used 20 mM HCO₃⁻ in all subsequent experiments, which also stimulated I_sc significantly. We predict the greater I_sc stimulated by higher [HCO₃⁻] to be at least partially dependent on sAC. Second, HCO₃⁻ not only stimulated the absorptive I_sc, but it also increased G_te in a dose-dependent manner. None of the inhibitors affected the stimulated G_te, although 20 mM HCO₃⁻ failed to stimulate G_te in Cl⁻-free conditions. It is difficult to speculate about these results, but the most likely candidates are transporters in the apical membrane (which typically dominates G_te) such as secretory K⁺ channels and electrogenic Cl⁻/HCO₃⁻ exchangers (SLC26a6 and possibly SLC26a3), which might be directly or indirectly modulated by HCO₃⁻ and/or pH . Third, inhibition of sAC or carbonic anhydrase inhibited the basal I_sc in HCO₃⁻-free saline. This points to a role of endogenous metabolic CO₂ as an additional modulator of NaCl transport in toadfish intestine via carbonic anhydrase and sAC.

Interaction between intestinal NaCl absorption and anion exchange.

Earlier observations demonstrate that conditions of low luminal NaCl, which are unfavorable for salt absorption via NKCC, results in elevated Cl⁻ absorption by anion exchange as evident from elevated HCO₃⁻ secretion (23). This enhanced anion exchange appears to be, at least in part, via SLC26a6 proteins (22, 29). The present study, on the other hand, demonstrates that conditions more adverse to Cl⁻ absorption via anion exchange (high-luminal HCO₃⁻) stimulate NaCl absorption via NKCC. It thus appears that the pathways for Cl⁻ absorption by the intestinal epithelium in marine fish are coordinated, depending on luminal conditions to allow for Cl⁻ and thus water absorption. While sAC appears to be the link between luminal HCO₃⁻ and NKCC activation, it remains to be established how low NaCl stimulates anion exchange (or how high NaCl inhibits anion exchange).

HCO₃⁻ as a modulator of epithelial ion transport.

From an ion-transporting perspective, HCO₃⁻ is typically only considered as a substrate for anion exchangers and Na⁺/HCO₃⁻ cotransporters (57), and any additional effects are usually attributed to indirect changes in pH. However, HCO₃⁻ can also act as a signaling molecule, which can gain specificity and be amplified through HCO₃⁻-responsive soluble adenylyl cyclase (3, 6, 45, 56, 63).

Several studies have reported modulation of transepithelial ion transport by HCO₃⁻ and alkaline pH in a variety of epithelia, including rat small intestine (52), rat colon (8), human jejunum (54), crab gills (62), fish intestine (5, 12, 13), and frog stomach (33). These unexplained effects of HCO₃⁻ could depend on sAC, similar to the toadfish intestine, or to other yet unidentified HCO₃⁻ modulation.

Perspectives and Significance

We have identified a novel regulatory pathway for epithelial NaCl absorption, which depends on HCO₃⁻, and it may involve sAC and NKCC. Our results also suggest the existence of cAMP signaling microdomains in ion-transporting epithelia. Activation of NaCl absorption by HCO₃⁻/sAC/NKCC may be essential for intestinal carbonate precipitation and water absorption, a process that is essential for marine fish osmoregulation and which has implications for the ocean carbon cycle. Moreover, the literature suggests this novel regulation may exist in other epithelia.

GRANTS

This work was supported by the Journal of Experimental Biology Traveling Fellowship to M. Tresguerres and National Science Foundation Grant IOS0743903 (to M. Grosell) and National Institutes of Health Grants GM-62328 and NS-55255 to L. R. Levin and J. Buck.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).