INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

BY MAINTAINING PLASMA OSMOLALITY (~330 mosmol/kgH₂O) below that of seawater (~1,000 mosmol/kgH₂O), marine teleosts constantly lose water to their surrounding environment. These animals, therefore, drink seawater to prevent dehydration. Fluid entering the intestine is isosmotic with plasma due to partial desalination of ingested fluid by the esophagus and the dilution of stomach contents with body water (16, 25). Active absorption of Na⁺ and Cl by the intestine drives water absorption (37). In addition to monovalent ions, divalent ions abundant in seawater (SO, Ca²⁺, and Mg²⁺) are absorbed after ingestion. Excess salts obtained from drinking are excreted by the gill (monovalent ions) and kidney (divalent ions).

The active absorption of NaCl by the intestinal epithelium is driven by the basolateral Na⁺-K⁺ pump, which creates an electrochemical gradient that drives electroneutral uptake (lumen to cell) of Na⁺, K⁺, and Cl by the loop-diuretic-sensitive Na⁺-K⁺-2Cl cotransporter (22). Sodium is pumped across the basolateral membrane via Na⁺-K⁺-ATPase, and Cl exits via Cl channels (4, 11, 12) and an electroneutral K⁺-Cl cotransporter (11). Subsequently, K⁺ is pumped back into the cell via Na⁺-K⁺-ATPase and moves across the apical and basolateral membranes via conductive pathways (12). K⁺ movement (cell to lumen) across the apical membrane and Cl movement (cell to interstitium) across the basolateral membrane generates a transepithelial electrical potential difference (PD) that is serosal negative (7), and this PD is a direct measure of NaCl absorption (23). Although the data have provided overwhelming evidence that the marine teleost intestine is involved in NaCl absorption, there have been few studies to indicate that the intestine may function in ion secretion as well (9, 10, 20, 24, 40).

While the mechanism of NaCl absorption is largely known, divalent ion transport by the marine teleost intestine is not well understood. Divalent ion levels (SO, Ca²⁺, and Mg²⁺) throughout the gastrointestinal tracts of both Southern flounder (Paralichthys lethostigma) (13) and winter flounder (25) after seawater ingestion have been measured. SO (29 mM), Mg²⁺ (49 mM), and Ca²⁺ (12 mM) concentrations are high in seawater compared with plasma (~1 mM). After ingestion, fluid entering the stomach is diluted with body water, resulting in slightly reduced [SO] and [Mg²⁺] with no change in [Ca²⁺]. In contrast, as fluid travels from the anterior intestine to the hindgut, both [SO] (50-80 mM) and [Mg²⁺] (80-160 mM) increase approximately two- to threefold. By assuming that all SO and Mg²⁺ are excreted at the kidneys and rectum, it was determined that 10-15% of ingested SO and Mg²⁺ are absorbed by the gastrointestinal tract (13, 25). However, the locations and mechanisms of divalent ion absorption by the gastrointestinal tract of marine teleosts remain unknown.

The purpose of the present investigation was to determine the direction, mechanism, and regulation of SO transport by the winter flounder intestine. Although the intestine is a site of net SO absorption (lumen to blood), we have determined that there exists an active secretory (blood to lumen) component through which cellular SO exchanges for luminal Cl at the apical membrane. This transepithelial anion-exchange mechanism is inhibited by sodium cyanide, ouabain, 4,4'-diisothiocyanatostilbene-2-2'-disulfonic acid, and satiety and may play roles in water absorption and in the maintenance of SO homeostasis.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals. Winter flounder, Pleuronectes americanus, were obtained by otter trawl in Frenchman's Bay, ME, or in Long Island Sound, CT. Animals (250-400 g) were held in flowing seawater (17-19°C) or in living stream units (Toledo) filled with artificial seawater (Utikem) at 12°C. Animal use followed the newest guiding principles for research (2).

Ussing chamber studies. After dissection, four to six pieces of intestine were removed from an area ~4 cm below the stomach and washed in ice-cold flounder saline. After removing the adventitia, tissues, supported by 150-µm nylon mesh, were mounted in Ussing chambers. Aperture size was 0.332 cm². Fluid volume was 1.2 ml/hemichamber. The temperature was maintained at a constant 20°C with water circulated on the outside surface of the chambers by a Lauda RM6 Electronic water bath. Fluid inside the chambers was constantly and vigorously stirred with small magnetic stir bars turned by external stir plates. Unless indicated otherwise, chambers were insufflated with humidified 99% O₂-1% CO₂.

Determination of transepithelial SO fluxes. Tissues were continuously short circuited during flux determinations when identical solutions bathed the serosal and mucosal sides of the tissue. Tissues were not short circuited under asymmetric conditions. Unidirectional tracer fluxes were initiated by the addition of 1.0-2.0 µCi ³⁵SO to the appropriate hemichamber. Duplicate 50-µl samples were taken from the unlabeled side at 30-min intervals over a period of 1.5 h and replaced with equal volumes of unlabeled solution. The specific activity of the labeled solution was determined at the beginning and end of each experiment.

Solutions. The serosal surface was always bathed with flounder saline (FS) containing (in mM) 150.0 NaCl, 4.0 KCl, 1.9 CaCl₂, 1.0 MgSO₄, 0.4 NaH₂PO₄, 4.2 NaHCO₃, 25.0 HEPES, and 5.5 glucose (pH 7.5). The mucosal surface was bathed with FS, Cl-free FS, HCO-free FS, Cl- and HCO-free FS, or artificial intestinal fluid containing 10, 25, or 50 mM SO. In the Cl-free FS solution, NaCl, KCl, and CaCl₂ were replaced with gluconate salts. In the HCO-free solution, NaHCO₃ was replaced with Na⁺ gluconate. As above, gluconate salts were used for substitution in the Cl- and HCO-free FS solution. To limit the production of HCO from CO₂, HCO-free FS and Cl- and HCO-free FS and their paired controls were gassed with 100% O₂. The intestinal fluid solutions were modifications of intestinal fluid contents observed in winter flounder (25). The 10 mM SO intestinal fluid solution contained (in mM) 95 NaCl, 2.0 KCl, 23 CaCl₂, 10 MgSO₄, 28.3 MgCl₂, 4.2 NaHCO₃, 25 HEPES, and 5.5 glucose (pH 7.5). The 25 mM SO intestinal fluid solution contained (in mM) 82.8 NaCl, 2.0 KCl, 24.3 CaCl₂, 25.0 MgSO₄, 29.2 MgCl₂, 4.2 NaHCO₃, 25.0 HEPES, and 5.5 glucose (pH 7.5). The 50 mM SO intestinal fluid solution contained (in mM) 59.0 NaCl, 2.0 KCl, 25.0 CaCl₂, 40 MgSO₄, 30.0 MgCl₂, 10.0 Na₂SO₄, 4.2 NaHCO₃, 25.0 HEPES, and 5.5 glucose (pH 7.5). The osmolality of all solutions was that of the FS, 340 mosmol/kgH₂O.

Ouabain-linked Sephadex bead preparation. In preparing brush-border membrane vesicles (BBMV), ouabain-linked Sephadex beads were used to limit basolateral membrane contamination. Synthesis of ouabain-linked beads was achieved by cross-linking oxidized ouabain to oxidized Sephadex. The method for Sephadex oxidation was adopted from Ref. 14. The reaction was inititated by adding 5 g of Sephadex (QAE A25) and 0.8 g of sodium periodate to 25 ml of 0.1 M sodium acetate adjusted to pH 5.0 with 1 M HCl. The mixture was stirred for 1.5 h at room temperature. The mixture was dialyzed six times with 1 liter of water (total of 6 liters) over a 1-h period. 1,6-Hexane diamine was added until 0.5 M, and the pH was adjusted to 9.5 with 2 M HCl. After 1 h of stirring, 0.36 g of sodium borohydride was added and stirred overnight (~12 h) at room temperature. The oxidized Sephadex was dialyzed six times with 1 liter of water (total of 6 liters) over a 2-h period. The oxidized Sephadex was lyophilized and stored at 4°C.

Preparation of BBMV. This method is a modification of Refs. 28 and 36. The intestine, from ~3 cm below the stomach and 5 cm above the anus, was removed from four to five flounder and placed in ice-cold FS. The tissue was rinsed with ice-cold FS, and the mucosa was removed by scraping with a glass microscope slide. The intestinal scrapings were placed in a small volume of FS and quickly frozen in liquid nitrogen. On the day of preparation, 4-5 g of intestinal scrapings were thawed and diluted 1:30 (wt/vol) with ice-cold mannitol solution that contained the following protease inhibitors (in µM): 100 4-(2-aminoethyl) benzenesulfonylfluoride HCl, 0.08 aprotinin, 5.0 bestatin, 1.5 E-64, 2.0 leupeptin, and 1.0 pepstatin A. The epithelium was homogenized in a Polytron (Brinkman) equipped with a PT-20-ST probe at speed 8 for three 20-s bursts separated by two 1-min rest periods. The homogenate was diluted 1:2 (vol/vol) with ice-cold mannitol solution containing the above protease inhibitors. CaCl₂ (10 mM) was added, and the homogenate was stirred on ice for 30 min. The homogenate was centrifuged at 3,000 g for 10 min, and the pellet was discarded. The supernatant was centrifuged at 39,500 g for 45 min, and the pellet (crude brush-border membrane fraction) was suspended in 150 ml of mannitol solution. The crude brush-border membrane fraction was stirred on ice with ouabain beads (0.1 ml of packed beads/mg protein) for 1 h followed by 9,000 g centrifugation for 10 min. The pellet was discarded, and the supernatant was homogenized (10 strokes) with a glass teflon homogenizer. Ouabain beads were added (0.1 ml packed beads/mg protein), and the mixture was stirred for 1 h followed by centrifugation at 9,000 g for 10 min. The pellet was discarded, and the supernatant was centrifuged at 39,500 g for 45 min. The pellet (brush-border membranes) was suspended in the appropriate intravesicular buffer with a syringe fitted with a 23-gauge needle. BBMV were centrifuged at 39,500 g for 45 min and then suspended in the appropriate intravesicular buffer using a 23-gauge needle. The vesicles were stored in liquid nitrogen until the day of flux measurements.

Measurement of vesicle uptake. ³⁵SO and D-[³H]glucose uptake were determined by adding 90 µl of external buffer to 10 µl of vesicles at room temperature (23°C). Two hundred microliters of external buffer were used when vesicles were loaded with 25 mM HCO. Uptake was "stopped" by adding ice-cold stop solution containing (in mM) 300 sucrose, 3 Tris-HEPES (pH 7.8), and 0.1 HgCl₂. Vesicles were collected and washed on Fisher brand 0.45-µm filters as previously described (28). Transport assays were done in triplicate on at least three separate membrane preparations. Compositions of internal and external buffers varied experimentally and are detailed in the figure legends.

Chemicals. ³⁵SO (H₂SO₄, >99% purity) and D-[³H]glucose were purchased from ICN Radiochemicals. Amiloride, DIDS, 1,6-hexane diamine, ouabain, Sephadex QAE A25, sodium borohydride, sodium cyanoborohydride, sodium periodate, and valinomycin were obtained from Sigma Chemical.

Statistics. Experimental results are expressed as means ± SE. Paired and unpaired comparisons of sample means were done using a Student's t-test. Differences were judged significant if P < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Under short-circuited conditions with identical FS bathing serosal and mucosal sides, winter flounder intestine actively secreted SO at a net rate of 8.55 ± 0.96 nmol · cm² · h¹ (1.5 h, Fig. 1). Fluxes were initiated at t = 0 with the addition of ³⁵SO tracer. SO secretion had not reached steady state after 1.5 h, indicating that the tracer is slow to equilibrate with intracellular SO pools. The absorptive flux was 26% of the secretory flux at 1.5 h. TER, PD, and I_sc averaged 47.2 ± 2.0  · cm², 1.82 ± 0.12 mV (mucosal side positive), and 11.6 ± 0.76 µA/cm², respectively (Table 1).

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Plot of unidirectional and net SO fluxes across winter flounder intestine as a function of time. Tissues were maintained under short-circuited conditions throughout the experiment. Unidirectional fluxes include serosal-to-mucosal secretory flux (S to M) and mucosal-to-serosal absorptive flux (M to S). Net flux represents the difference between the unidirectional fluxes. Values are means ± SE of 39 tissues.

Under physiological conditions, SO concentration in the anterior portion of the winter flounder intestine can be >50-fold higher than plasma. Figure 2 shows the effect of three experimental intestinal solutions containing varying concentrations of SO (10, 25, and 50 mM). With 10 mM SO bathing the luminal surface, the unidirectional absorptive flux increased to 21.3 ± 2.65 nmol · cm^{2 ·} h¹ at 1.5 h (Fig. 2B), which was sixfold higher than its paired control (Fig. 2A). The large increase in the unidirectional absorptive flux caused the net flux to reverse in the direction of absorption (3.36 ± 7.56 nmol · cm² · h¹). Surprisingly, the unidirectional absorptive flux only increased to 21.4 ± 6.0 nmol · cm² · h¹ in the presence of 25 mM luminal SO (1.5 h, Fig. 2D). However, with 50 mM luminal SO, the unidirectional absorptive flux was 71.4 ± 19.3 nmol · cm² · h¹ at 1.5 h (Fig. 2F) and 36-fold higher than the control value (Fig. 2E). Net absorption at a rate of 56.1 ± 16.5 nmol · cm² · h¹ (1.5 h) occurred with 50 mM SO bathing the luminal surface. The unidirectional secretory flux was not altered (compared with the paired control) when high levels of SO bathed the luminal side.

View larger version (27K): [in this window] [in a new window]   Fig. 2.   SO transport rate in the presence of luminal intestinal solutions with varying levels of SO and their paired controls (1 mM luminal SO). A: paired control for 10 mM luminal SO. B: 10 mM luminal SO. C: paired control for 25 mM luminal SO. D: 25 mM luminal SO. E: paired control for 50 mM luminal SO. F: 50 mM luminal SO. Unidirectional fluxes include S-to-M secretory flux and M-to-S absorptive flux. Net flux represents the difference between the unidirectional fluxes. Tissues were maintained under open-circuited conditions throughout the experiment. Values are presented as means ± SE of 5 preparations. * Significantly different from paired control (P < 0.05).

Figure 3 shows the effects of sodium cyanide (NaCN, 10 mM; Fig. 3A) and ouabain (10⁴ M; Fig. 3B) on the unidirectional secretory (serosal to mucosal), absorptive (mucosal to serosal), and net SO flux across the intestinal epithelium of winter flounder under short-circuited conditions. NaCN abolished net flux, whereas ouabain reduced net SO secretion to 29% of the control value. The effects of NaCN and ouabain appeared to be similar with both treatments inhibiting the unidirectional secretory flux ~50%, while having no effect on the unidirectional absorptive flux. Unlike NaCN, ouabain treatment caused a 42% significant increase in TER (Table 1). NaCN and ouabain treatment both abolished the PD and I_sc.

View larger version (14K): [in this window] [in a new window]   Fig. 3.   Effects of 10 mM sodium cyanide (NaCN; A) and ouabain (104 M; B) on transepithelial secretory (S to M), absorptive (M to S), and net SO fluxes across winter flounder intestine. Tissues were incubated with the respective treatment for 0.5 h before beginning the experiment. Values shown were obtained at t = 1.5 h and are means ± SE (vertical line) of 3 (NaCN) and 6 (ouabain) preparations. * Significantly different from paired control (P < 0.05).

When applied to the luminal bath solution, the anion-exchange inhibitor DIDS (0.2 mM) significantly reduced net SO secretion by 46% (Fig. 4). The effect was due to simultaneous small reductions in the unidirectional secretory flux and increases in unidirectional absorptive flux. In contrast, serosal application of DIDS had no effect on the unidirectional secretory or net SO flux. Similarly, TER, PD, and I_sc were not affected by serosal or luminal DIDS treatment (Table 1).

View larger version (17K): [in this window] [in a new window]   Fig. 4.   Effects of serosal and luminal application of DIDS (0.2 mM) on transepithelial secretory (S to M), absorptive (M to S), and net SO fluxes across winter flounder intestine. Values shown were obtained at t = 1.5 h and are means ± SE (vertical line) of 4 preparations. * Significantly different from paired control (P < 0.05).

The luminal DIDS effect prompted a series of anion removal (luminal bath) studies to probe the mechanism of anion exchange (Fig. 5). Removal of Cl and HCO (Cl-HCO) from the luminal bath solution eliminated net SO secretion through a 28% reduction in the secretory flux and a 79% increase in the absorptive flux (Fig. 5A). Net SO flux was reduced to 15% of the control value after luminal Cl removal (Fig. 5B). The effect (Cl removal) was similar to Cl-HCO removal in that the unidirectional secretory and absorptive fluxes were inhibited and stimulated, respectively. Removal of HCO from the luminal bath solution resulted in a 50% increase in net SO secretion caused by an increase in the unidirectional secretory flux (Fig. 5C). The unidirectional absorptive flux was not altered by HCO removal. TER was not affected by any of the three ion-substitution treatments (data not shown).

View larger version (14K): [in this window] [in a new window]   Fig. 5.   Effects of anion removal on transepithelial secretory (S to M), absorptive (M to S), and net SO fluxes. Flounder saline always bathed the serosal side. A: removal of luminal Cl and HCO (Cl and HCO free). B: removal of luminal Cl (Cl free). C: removal of luminal HCO (HCO free). Tissues were maintained under open-circuited conditions throughout the experiment. Values shown were obtained at t = 1.5 h and are means ± SE (vertical line) of 6 preparations. * Significantly different from paired control (P < 0.05).

Na⁺-dependent glucose transport is a diagnostic transport process in intestinal brush-border membranes. For this reason, we examined Na⁺-dependent glucose uptake into flounder foregut BBMV as a quality control measure. Figure 6 shows that imposition of a 100 mM Na⁺ gradient (out > in) caused concentrative glucose uptake into flounder foregut BBMV. Uptake at 1 min was 31% higher than at equilibrium (60 min). Glucose uptake, in the presence of the same Na⁺ gradient, was 61% higher than at equilibrium in vesicles that were short-circuited with 100 mM K⁺ (out = in) and valinomycin. There was no concentrative glucose uptake in the absence of a Na⁺ gradient.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   Glucose uptake by flounder foregut brush-border membrane vesicles in presence or absence of 100 mM Na+ gradient (out > in). Internal buffer contained (in mM) 100 K+ gluconate, 100 mannitol, 2 Ca2+ gluconate, 3 Mg2+ gluconate, 20 Tris-HEPES (pH 7.8), and 3 NaN3. Vesicles were incubated in the same medium (without 100 mM mannitol) containing 0.l mM glucose and either 100 mM Na+ gluconate or 200 mM mannitol, and where indicated, 20 µg valinomycin/mg protein. Each point represents mean ± SE of 3 preparations (n = 3). * Significantly different from no gradient (P < 0.05).  Significantly different from Na+ gradient (valinomycin, P < 0.05).

Figure 7 shows the effect of Cl and HCO gradients on SO uptake into flounder foregut BBMV. Imposition of a 100 mM Cl gradient (in > out) caused concentrative SO uptake, which was 68% higher at 5 s than equilibrium (Fig. 7A). Short-circuiting the vesicles with 100 mM K⁺ (in = out) and valinomycin had no effect on Cl-stimuluated SO uptake. There was no stimulation of SO uptake in the absence of a Cl gradient. Imposition of a 23 mM HCO gradient (in > out) did not produce concentrative SO uptake in either open- or short-circuited vesicles (Fig. 7B).

View larger version (19K): [in this window] [in a new window]   Fig. 7.   A: SO uptake by flounder foregut brush-border membrane vesicles in presence or absence of a 100 mM Cl gradient (in > out). Internal buffer contained either 100 mM KCl or 100 mM K+ gluconate and (in mM) 100 mannitol, 2 Ca2+ gluconate, 3 Mg2+ gluconate, 20 Tris-HEPES (pH 7.8), and 3 NaN3. B: SO uptake by brush-border membrane vesicles in presence or absence of a 23 mM HCO gradient (in > out). Internal buffer contained either 25 mM NaHCO3 and 50 mM mannitol or 100 mM mannitol, and (in mM) 100 K+ gluconate, 2 Ca2+ gluconate, 3 Mg2+ gluconate, 20 Tris-HEPES (pH 7.8), and 3 NaN3. Vesicles were incubated in external buffer containing (in mM) 100 K+ gluconate, 200 mannitol, 2 Ca2+ gluconate, 3 Mg2+ gluconate, 20 Tris-HEPES (pH 7.8), 0.l mM NaSO4, and, where indicated, 20 µg valinomycin/mg protein. Each point represents mean ± SE of 3 preparations (n = 3). * Significantly different from no gradient (P < 0.05).

Satiety (i.e., full stomach) abolished net SO transport by the winter flounder intestine (Fig. 8). The strong effect of feeding on net SO transport was due to a 3.9-fold increase in the unidirectional absorptive flux. Feeding had no effect on the unidirectional secretory flux. PD and I_sc were five- and threefold lower among fed than unfed fish, respectively (Table 1). TER was not different among fed and unfed fish.

View larger version (14K): [in this window] [in a new window]   Fig. 8.   Effect of satiety on transepithelial secretory (S to M), absorptive (M to S), and net SO fluxes. Values shown were obtained at t = 1.5 h and are means ± SE (vertical line) of 39 unfed and 10 fed tissues. * Significantly different from control (P < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Transport processes for SO are exaggerated in marine teleosts due to the necessity of seawater ingestion. SO secretion is accomplished in part by renal proximal tubule active secretion (29). While the gastrointestinal tract has been proposed to be a site of SO absorption (13, 25), its role in transporting SO has not been defined. In this study, we have examined SO transport by the winter flounder intestine to determine its role in absorption. In the absence of chemical and electrical gradients, however, the intestine of winter flounder actively secretes (blood to lumen) SO. There have been numerous studies to characterize the mechanisms of active NaCl absorption and the electrical characteristics generated from this transport (6, 17), and the electrophysiological parameters (TER, PD, and I_sc) obtained for the control tissues in the present study are consistent with the prior observations.

SO concentration in the lumen of the anterior intestine can be >50-fold higher than plasma (13, 25). For this reason, a series of experiments using modified intestinal solutions (see METHODS) was conducted to determine the effects of SO gradients (10, 25, or 50 mM SO) on SO transport by the intestine. The dramatic increase in unidirectional absorptive flux with increasing luminal [SO] suggests that, under physiological conditions, the intestine is a site of SO absorption. Of interest, however, is the observation that the increase in SO absorption is not linear with increasing luminal [SO]. This nonlinearity could be explained by a model in which the movement of SO from cell to blood across the basolateral membrane occurs on a relatively low-affinity SO transporter. The unidirectional secretory flux was maintained (not different from paired control), even in the presence of large trans-SO gradients. The maintenance of the unidirectional secretory flux with increasing luminal [SO] may suggest that intracellular [SO] does not change and that the secretory flux reflects the affinity of SO for the intracellular transport site.

Treatment with NaCN abolished net SO secretion, indicating metabolic dependence. The effect of NaCN on net flux under short-circuited conditions was manifested through a reduction in the unidirectional secretory flux; the unidirectional absorptive flux was not altered. By inhibiting the unidirectional secretory flux, ouabain reduced net secretion to 29% of the control value, strongly implicating the plasma membrane Na⁺ gradient in SO secretion. Transepithelial SO secretion by the marine teleost renal proximal tubule is also Na⁺ gradient dependent, i.e., inhibited by low-Na⁺ media and ouabain (29, 30). Elimination of the Na⁺ gradient apparently has indirect effects leading to an inhibition in SO secretion. For instance, secondary effects may arise from changes in intracellular Ca²⁺ and pH due to changes in Na⁺/Ca²⁺ and Na⁺/H⁺ exchange, respectively. In the eel intestine, intracellular pH is regulated by a basolateral Na⁺/H⁺ exchanger, which is inhibited by amiloride, resulting in intracellular acidification (39). Serosal treatment of the flounder intestine with amiloride (10⁴ M, n = 3), however, had no effect on the secretory (20.0 ± 4.85 vs. 16.2 ± 11.3; control vs. amiloride), absorptive (3.21 ± 0.06 vs. 2.47 ± 0.34), or net SO fluxes (16.8 ± 4.92 vs. 13.8 ± 11.0 nmol · cm² · h¹). Treatment of the intestine from other marine species with ouabain inhibits NaCl absorption and abolishes the PD and I_sc (27). Similarly, in winter flounder we found that ouabain treatment eliminated the PD and I_sc. In contrast, ouabain treatment increased TER. Ouabain inhibition of Na⁺-K⁺-ATPase in the goldfish (Carassius auratus) intestine increases intracellular Na⁺ and Cl, which results in cell swelling and increased TER (1).

Anion exchangers present in the basolateral (26) and brush border (31) membranes of marine teleost renal proximal tubule cells facilitate SO secretion. This prompted the investigation of similar mechanisms of SO secretion in the winter flounder intestine. The anion-exchange inhibitor DIDS reduced net SO secretion when applied to the luminal surface only, indicating that anion exchange possibly facilitates movement of SO from cell to lumen. The DIDS effect on net secretion (45% inhibition) was less than expected. The SO transport mechanism described here may be more like that in human ureteral epithelial cells where there exists both DIDS-sensitive and -insensitive SO transport (5). In the rabbit renal tubule connecting segment (segment between distal tubule and collecting duct), HCO secretion is facilitated by an apical Cl/HCO exchanger that is also DIDS insensitive (38). In the present study there was no effect of DIDS when applied to the serosal surface, and thus if this intestine contains a basolateral anion exchanger that transports SO, it is DIDS insensitive. On the other hand, it is possible that movement of SO from blood to cell requires an entirely different mechanism. In the mammalian intestine, SO absorption is made possible by a Na⁺-dependent SO cotransporter (19). Whereas the latter has only been localized to the brush-border membrane in mammals, its presence in the basolateral membrane is possible and could further explain why SO secretion in the winter flounder intestine is Na⁺ gradient dependent.

Removal of Cl and HCO together (Cl-HCO free) and Cl alone (Cl free) from the luminal bath solution almost completely inhibited net SO secretion by decreasing the secretory flux and increasing the absorptive flux. The inhibition of net secretion after luminal Cl removal alone suggests that luminal Cl is required for movement of SO from cell to lumen, and hence SO secretion. Cl-dependent concentrative SO uptake into foregut BBMV confirmed the presence of SO/Cl exchange at the brush-border membrane. As mentioned previously, ouabain treatment increases intracellular Cl concentration, which would decrease the driving force for exchange of luminal Cl for cellular SO, and, therefore, helps explain the decrease in net SO secretion after ouabain treatment. The increased unidirectional absorptive flux after removal of Cl alone and Cl and HCO together would be expected to occur via the downhill movement of Cl from blood to lumen driving the exchanger in reverse. Luminal HCO removal (HCO free) caused stimulation in net SO secretion through an increase in the unidirectional secretory flux. This observation may suggest that HCO competes for a site on the exchanger but is not effective in facilitating transport. There was no HCO-dependent concentrative SO uptake into BBMV, indicating that HCO does not facilitate SO secretion. More recently, Cl/HCO exchangers facilitating HCO secretion have been localized to the apical membrane of marine teleost intestinal epithelial cells (10, 40). The SO transporter identified in this study and HCO transporter in these other studies are likely two different transporters; however, their functions could be interrelated because they both use Cl to facilitate transport of their respective counteranions. In this case, the removal of HCO from the luminal bath solution would increase the rate of Cl/HCO exchange and potentially reduce the amount of Cl used for SO/Cl exchange. This scenario would result in reduced SO secretion, which is opposite to the observed increase. Luminal HCO removal could also act indirectly to increase SO secretion through changes in intracellular pH. Taken together, these results suggest that the brush border anion exchanger is most efficient when exchanging cellular SO for luminal Cl, and HCO effects are indirect.

Anion exchangers that exhibit high affinity for SO and Cl have been identified in several tissues including lobster (Homarus americanus) hepatopancreas (8) and mammlian ureter (5), intestine, and cartilage (34, 35). Similarly, we have identified a transporter in the brush-border membrane of the winter flounder intestine that functions best when exchanging SO (cellular) for Cl (luminal). However, the purpose of this transporter remains uncertain. As mentioned above, SO secretion by the intestinal epithelium could assist the regulation of plasma SO. Under physiological conditions (50 mM luminal SO), the unidirectional secretory flux was 21% of the unidirectional absorptive flux at 1.5 h. Therefore, increasing or decreasing the secretory flux could influence plasma SO levels. In addition, if the stoichiometry is such that two Cl are absorbed for one SO secreted (1SO:2Cl), then the transport process could generate an osmotic driving force for water absorption. In the rat (26) and Southern flounder (31) kidney, SO/HCO exchange is electroneutral. Similarly, SO/Cl exchange by band 3 in the human erythrocyte is electroneutral (18). However, in the lobster hepatopancreas Cl/SO exchange is electrogenic (1Cl:1SO) (3). Short circuiting BBMV had no effect on Cl-stimulated SO uptake, indicating that SO/Cl exchange in the flounder intestinal brush-border membrane is electroneutral. Further work will be required to determine the stoichiometry of this transport process.

Net SO transport by the intestine of fed fish was reduced to zero. Reasons for the effect of feeding on SO transport, however, are not entirely clear. Stomach distension, as occurs during feeding, has been shown to inhibit drinking in seawater-adapted eel, Anguilla japonica (15). Feeding, therefore, would reduce the absorption of excess SO. Inhibition of SO secretion during periods of limited drinking would contribute to regulation of plasma SO levels. Another possibility is that SO transport by the intestinal epithelium is altered in preparation for nutrient absorption. Both the PD and I_sc were greatly reduced in fed fish, indicating that the ability to absorb NaCl and water may have been downregulated. Slightly positive I_scs were observed in winter flounder that had food present in the intestine (24). If fluid ingestion were reduced during feeding, continued activity by mechanisms for NaCl and water absorption would be energetically expensive. The dramatic effect of feeding on SO transport and electrical properties suggests that hormones are involved in the regulation of salt and water transport by the marine teleost intestine.

In conclusion, this investigation has demonstrated that the intestine is a site of SO absorption in marine teleosts and has identified active SO secretion that is dependent on metabolism and the Na⁺ gradient. SO secretion (cell to lumen) is facilitated by a DIDS-sensitive anion exchanger at the brush-border membrane. This anion exchanger facilitates the exchange of luminal Cl for cellular SO, a process that is electroneutral. SO secretion by the intestine is strongly inhibited by satiety. Further work is required to determine whether this anion-exchange mechanism is important for regulating plasma SO and water absorption.