Parallel Antiport Mechanisms for Na⁺ and Cl^– Transport in Herbivorous Teleost Intestine

The intestine of marine teleost fish actively absorbs NaCl and water from the lumen as a part of the osmoregulatory process that takes place in these animals (Ando, 1981; Skadhauge, 1974). The transport mechanism by which NaCl traverses the brush border membrane of intestinal epithelial cells has not been studied in a wide variety of marine teleosts. There are currently four mechanisms by which the electroneutral transport of NaCl across epithelial cell membranes is known to occur. One mechanism involves the cotransport of Na⁺, K⁺ and CP and has been found in the shark rectal gland (Hannafin, Kinne-Saffran, Friedman & Kinne, 1983), flounder intestine (Eveloff, Field, Kinne & Murer, 1980; Musch et al. 1982) and other tissues (Epstein & Silva, 1985). A second mechanism of NaCl transport is K⁺-independent cotransport of Na⁺ and Cl^–. This mechanism has been found in flounder urinary bladder (Stokes, 1984) and Necturus gallbladder (Ericson & Spring, 1982). A third mechanism of NaCl transport involves the action of dual antiporters: Na⁺/H⁺ antiport and C1⁻/OH^– or C1⁻/HCO₃^– antiport. This mechanism has been found in rat and rabbit small intestines (Knickelbein, Aronson, Atherton & Dobbins, 1983 ; Knickelbeinet al. 1985; Liedtke & Hopfer, 1982a,b) and Necturus proximal tubules (Seifter & Aronson, 1984). The final mechanism for NaCl transport involves the action of Na⁺/H⁺ antiport and Cl^–/formate^– antiport. This mechanism has been found in rabbit renal cortex (Karniski & Aronson, 1985).

The transport of NaCl across the brush border membrane of the intestinal epithelium of marine teleosts could occur by any or all of these four mechanisms. The only transport mechanism that has been documented in marine teleost fish is the Na⁺/K⁺/2C1⁻ mechanism in the winter flounder, Pseudopleuronectes americanus (Eveloff et al. 1980; Musch et al. 1982; Halm, Krasny & Frizzell, 1985). Whether other marine teleosts use this mechanism for NaCl transport or one of the other three has not been documented. This study was undertaken to determine the mechan-ism(s) of NaCl transport across the brush border membrane of intestinal epithelial cells in Oreochromis mossambicus, a herbivorous, euryhaline species of tilapia.

Seawater-adapted adult tilapia (100—150g) were captured locally and kept for 1 day prior to being used. Each animal was killed by a blow to the head, and the upper intestine removed and placed in 15 ml of buffer 1 [300 mmol 1⁻¹ mannitol; 60 mmol 1⁻¹ Hepes/Tris, 6 mmol 1⁻¹ EGTA, 0·1 mmol 1⁻¹ phenylmethylsulphonyl-fluoride (PMSF), pH 7·4]. After the tissue had been removed from the animals, 60 ml of ice-cold deionized water was added to the 15 ml of buffer 1. This mixture was immediately homogenized for 15 s in a polytron. Magnesium chloride was added to the homogenate to give a final concentration of 15 mmol 1⁻¹, and the resulting mixture was kept on ice for 15 min. After 15 min the mixture was centrifuged at 3000 g for 15 min. The supernatant was then centrifuged at 27 000g for 30 min. This pellet was resuspended in 30 ml of ice-cold buffer 2 (60 mmol 1⁻¹ mannitol, 12mmoll^–1 Hepes/Tris, 6mmoll^–1 EGTA, pH 7·4) using a Potter-Elvehjem homogenizer. The magnesium chloride precipitation and two centrifugations were repeated on the resuspended pellet. The pellet resulting from this second set of spins was resuspended in 30 ml of ice-cold transport buffer (see figure legends for composition), and this mixture was centrifuged at 27 000g’ for 30min. The final pellet was resuspended in transport buffer at a protein concentration of approxi-mately 10 mg ml^–1.

Purity of the brush border membrane vesicles (BBMV) obtained from this procedure was determined by measuring the activity of membrane marker enzymes in the final pellet and in the homogenate. The activities of alkaline phosphatase, Na⁺/K⁺-stimulated ATPase, K⁺-stimulated p-nitrophenylphosphatase, and cyto-chrome c oxidase were determined. Alkaline phosphatase activity was determined using Sigma chemical assay kit no. 104. Na⁺/K⁺-stimulated ATPase activity was determined using the coupled assay reaction of Berner, Kinne & Murer (1976). K⁺-stimulated p-nitrophenylphosphatase activity was determined using the procedures of Arvan & Castle (1982). Cytochrome c oxidase activity was determined using the procedures of Cooperstein & Lazarow (1951).

The uptake of ²²Na⁺ and ³⁶C1∼ by BBMV was measured by the Millipore membrane filtration technique of Hopfer, Nelson, Perrotto & Isselbacher (1973). The compositions of the transport buffer inside the vesicles and the incubation medium outside the vesicles are indicated in each figure. Transport experiments were started by adding a volume of membrane vesicles (e.g. 20 μl) to a volume of radiolabelled incubation medium (e.g. 180μ1) to make a reaction mixture. At pre-determined incubation times (i.e. 15 s, 1·5, 3, 5 and 120 min) a known volume of the reaction mixture (20 μl) was removed and injected into 2 ml of ice-cold stop solution to stop the reaction. The composition of the stop solution varied according to the experimental conditions. Its composition was usually the same as the composition of the external control medium. The stop solution and reaction mixture sample were immediately filtered through a 0·65 μm Millipore filter and washed with 8 ml of ice-cold stop solution. The filter was then placed in a scintillation vial and 5 ml of Beckman Ready Solv HP scintillation cocktail was added. All samples were counted in a Beckman LS-8100 scintillation counter. Each experiment consisted of a minimum of three replicates per time point per treatment. Each experiment was repeated at least once to verify the results. The results of only one of these repeated experiments are presented here as means ± Is.E. for each time point. Treatments were compared using Duncan’s multiple range test.

The amount of protein present in each experiment was determined by the Coomassie blue dye binding assay (Bio-Rad) as developed by Bradford (1976). A mixture of human albumin (50 mg ml^–1) and human globulin (30 mg ml^–1) was used for the protein standard.

²²NaCl was obtained from New England Nuclear (Boston, MA). H³⁶C1 was obtained from ICN Biomedicals, Inc. (Irvine, CA). Valinomycin, carbonyl cyanide zw-chlorophenylhydrazone (CCCP), amiloride, furosemide and the protein standard were obtained from Sigma Chemical (St Louis, MO).

Comparisons of the activities of alkaline phosphatase, Na⁺/K⁺-stimulated ATPase, K⁺-stimulated p-nitrophenylphosphatase and cytochrome c oxidase in the final pellet and the homogenate are presented in Table 1. Alkaline phosphatase is a marker for the brush border membrane, and its activity was increased 12-fold in the final pellet. Na⁺/K⁺-stimulated ATPase and K⁺-stimulated p-nitrophenylphospha-tase are markers for basolateral membranes, and their activities were increased 1·62-and 1·02-fold, respectively, in the final pellet. Cytochrome c oxidase is a marker for mitochondrial membranes, and its activity was increased 0·87-fold in the final pellet. These results indicate that the final pellet was enriched in brush border membranes and was not appreciably contaminated with either basolateral or mitochondrial membranes. It should be noted here that, since the final pellet was not 100% brush border membranes, the transport mechanisms that were being investigated here could be resident in some other minor contaminating membrane population (basolateral, mitochondrial, etc.), as well as being in the brush border membrane population.

To substantiate that BBMV were closed and osmotically reactive, the equilibrium uptake of 1 mmol 1⁻¹ Na⁺ was measured in the presence of increasing external raffinose concentrations (Fig. 1). If the vesicles were osmotically reactive, then as the external concentration of raffinose was increased the vesicles should shrink and hold less isotope. As shown in Fig. 1, the regression of Na⁺ uptake on the reciprocal of medium osmolarity was significant (P<0·01), indicating that the vesicles were osmotically reactive. The estimate of the slope is 0·64 ± 0·04 nmol mg protein^–1 mosmol^–1), and the estimate of the intercept is 1·04 ± 0·09 nmol mg protein^–1. The intercept indicates the amount of Na⁺ uptake that is independent of intravesicu-lar space (binding). A considerable amount of the Na⁺ uptake in Fig. 1 is due to the binding of Na⁺ to the vesicles. For example, at 300mosmoll⁻¹, 33% of the Na⁺ uptake is due to binding.

To test for the existence of the Na⁺/Cl^– cotransport model or the Na⁺/K⁺/2C1⁻ cotransport model in BBMV of tilapia upper intestine, the uptake of Na⁺ was measured in the presence of different ion gradients. As shown in Fig. 2, the uptake of Na⁺ was not stimulated by a Cl– gradient or by a KCl gradient into the vesicles when compared with the control (gluconate). These gradients would be expected to stimulate Na⁺ uptake if the mechanism of Na⁺ uptake were Na⁺/Cl^– cotransport or Na⁺/K⁺/2C1∼ cotransport. Also, the uptake of Na⁺ in the presence of a KCl gradient was not significantly affected by furosemide, which is known to affect the Na⁺/K⁺/2C1⁻ cotransport mechanism (Epstein & Silva, 1985; Hannafin et al. 1983). These results provide strong evidence against the presence of these two cotransport mechanisms for NaCl in the apical membrane of tilapia intestinal epithelial cells.

With evidence to suggest that the cotransport models discussed above were not the mechanisms of NaCl transport in tilapia intestinal BBMV, experiments to test for the existence of ion antiporters in tilapia intestinal BBMV were performed next. To test for the presence of a Na⁺/H⁺ antiporter, the uptake of Na⁺ was measured in the presence and absence of a proton gradient directed out of the vesicles. The internal and external solutions for this experiment did not contain Cl^– salts. Fig. 3 shows that the uptake of Na⁺ in the presence of an outwardly directed proton gradient (pH 6·0 in, pH 7·5 out) was much greater than that of the control (pH 7·5 in, pH 7·5 out). This stimulation was inhibited by 1 mmol 1⁻¹ amiloride which is known to inhibit Na⁺/H⁺ antiporters. Apparent influxes of 1 mmol 1⁻¹ Na⁺ were (nmol mgprotein^–1 15s^–1 ± 1 S.E.): control, l·46±0·09; pH gradient, 3·36± 0·14; pH gradient plus amiloride, 1·55 ±0·05. At 15s the pH gradient treatment was significantly different (P<0·05, Duncan’s multiple range test) from the control, while the pH gradient plus amiloride treatment was not. The various treatments in this experiment were short-circuited with valinomycin so that it could not be argued that the stimulation was due to diffusion potentials created by the diffusion of protons out of the vesicles, which in turn set up an electrical gradient that attracted Na⁺ into the vesicles. These results support the existence of a Na⁺/H⁺ antiporter in tilapia intestinal BBMV.

To test for the existence of a Cl^–/anion antiporter in tilapia intestinal BBMV, the uptake of Cl^– into vesicles containing different outwardly directed anion gradients was measured. The internal and external solutions for this experiment did not contain Na⁺ salts. Fig. 4 indicates that the uptake of Cl^– into these vesicles was stimulated by both an outwardly directed HCO₃^– gradient and an outwardly directed Cl^– gradient. Apparent influxes of 10 mmol 1⁻¹ Cl^– were (nmol mg protein^–1 15s^–1 ± 1 S.E.): gluconate, 12·90 ±0·19; HCO₃⁻, 15·83 ±0·34; Cl^–, 14·62 ± 0·42. At 15 s the HCO₃^– and Cl^– treatments were significantly different (P < 0·05, Duncan’s multiple range test) from the control. As discussed above for the Na⁺/H⁺ antiport experiment, the treatments were short-circuited with valinomycin so that electrical coupling can be ruled out as the reason for the stimulation of Cl^– uptake in the HCO₃^– and Cl^– treatments. These results support the existence of a C1⁻/HCO₃^– antiporter in tilapia intestinal BBMV. Overall, this supports the Na⁺/H⁺ : CC/HCO₃^– antiport model as the mechanism of NaCl transport in tilapia intestinal BBMV.

Early experiments which strongly suggested the existence of a mechanism capable of mediating electroneutral NaCl absorption in epithelia came from the work of Diamond (1962, 1964), Wheeler (1963) and Dietschy (1964). Their studies showed that Na⁺ and Cl^– were both actively absorbed by fish and rabbit gallbladder at nearly equal rates. They also found that the replacement of Na⁺ by nonabsorbed cations abolished active Cl^– absorption, and that the replacement of Cl^– by nonabsorbed anions abolished active Na⁺ absorption. This early work did not determine the site(s) of interaction between Na⁺ and Cl^– on the gallbladder epithelial cell membrane. Studies by Frizzell, Dugas & Schultz (1975) and Cremaschi & Henin (1975) determined that the site of coupling was at the brush border membrane. These early studies were made using whole-tissue techniques (e.g. Ussing chambers). A drawback of whole-tissue techniques is that they cannot easily be used to distinguish between symport and antiport mechanisms because the investigator has little, if any, control over what is inside the cell. Later studies used membrane vesicle techniques to delineate further the mechanisms of vertebrate NaCl transport. One of the first transport mechanisms to become widely documented in vesicles from a number of different tissues was Na⁺/K⁺/2C1⁻ transport (Epstein & Silva, 1985; Eveloff et al. 1980; Hannafin et al. 1983; Musch et al. 1982; Halm et al. 1985). Since the discovery of the Na⁺/K⁺/2C1⁻ mechanism other investigators have proposed different models for NaCl transport in epithelial tissues. These models involve the action of Na⁺/H⁺, Cl^–/OH^–, CK/HCO₃^– or Cl^–/formate^– antiperters (Karniski & Aronson, 1985 ; Knickelbein et al. 1983, 1985; Liedtke & Hopfer, 1982a,b; Seifter & Aronson, 1984). Both symport and antiport models can be used to explain coupled electroneutral NaCl transport in the tissues that have been studied.

For marine teleosts the transport of NaCl across the intestinal epithelium is an important part of the osmoregulatory process that these animals perform in order to survive in the marine environment. Prior to this study, the only membrane vesicle work which had been conducted on NaCl transport in the intestine of a marine teleost had been done on the winter flounder, Pseudopleuronectes americanus (Eveloff et al. 1980). This study disclosed that Na⁺ uptake into intestinal BBMV was stimulated by an inwardly directed Cl– gradient, and this stimulation was inhibited by furosemide. It did not determine if this NaCl cotransport was K⁺-dependent or K⁺-independent because all the treatments had K⁺ present. Later studies by Musch et al. (1982) and Halm et al. (1985) have shown that a Na⁺/K⁺/2C1⁻ cotransport mechanism is present in winter flounder intestine. To test for NaCl cotransport in tilapia intestinal BBMV, Na⁺ uptake was measured in the presence of inwardly directed Cl^– and KCl gradients. The conditions used were similar to those used by Knickelbein el al. (1985) to test for NaCl cotransport in rabbit ileal BBMV. The results in Fig. 2 indicate that neither Na⁺/Cl^– nor Na⁺/K⁺/2C1⁻ cotransport was present in tilapia intestinal BBMV, since Na⁺ uptake was not stimulated by an inwardly directed Cl^– or KCl gradient. Furosemide also had no effect on Na⁺ uptake (Fig. 2).

From Fig. 3 it can be seen that a Na⁺/H⁺ antiport mechanism exists in tilapia intestinal BBMV. This transporter has been documented in a wide variety of cells, including erythrocytes, lymphocytes, ova, sperm, skeletal muscle, cardiac muscle, neurones, capillary endothelium, fibroblasts, renal tubular cells, and intestinal and gallbladder epithelial cells (Mahnensmith & Aronson, 1985). The Na⁺/H⁺ anti-porter has also been implicated in the following physiological processes: regulation of intracellular pH, control of cell growth and proliferation, metabolic responses to hormones such as insulin and glucocorticoids, regulation of cell volume, and the transepithelial transport of NaCl via Na⁺/H⁺:C1 /HCO₃ antiport (Mahnensmith & Aronson, 1985).

Fig. 4 indicates that a CP/HCO₃^– antiport mechanism exists in tilapia intestinal BBMV. This transporter has been documented in rabbit small intestine by Knickelbein et al. (1985). A Cl⁻/OH⁻ antiporter has been documented by Liedtke & Hopfer (1982b) in rat small intestine. Both CK/HCO₃^– and C1⁻/OH^– exchange have also been documented in erythrocytes (Scarpa, Cecchetto & Azzone, 1970).

A possible mechanism for the coupling of the Na⁺/H⁺ antiporter and the C1⁻/HCO₃^– antiporter could be via the activity of carbonic anhydrase. Carbonic anhydrase facilitates the production of carbonic acid (H₂CO₃) from CO₂ and H₂O. Carbonic acid then dissociates into HCO₃^– and H ⁺, thus providing substrates for the antiporters. The net effect of this coupling mechanism is to move CO₂ out of the cell into the intestinal lumen and to move Na⁺ and Cl^– electroneutrally from the lumen into the cell. This provides a means for the cell to get rid of metabolic CO₂ while absorbing NaCl. A model of this coupling mechanism can be found in the paper by Knickelbein et al. (1985).

The reason for the presence of Na⁺/K⁺/2C1⁻ symport in winter flounder and Na⁺/H⁺: C1⁻/HCO₃⁻ antiport in tilapia is not yet understood. It should be noted here that the presence of Na⁺/H⁺: C1⁻/HCO₃⁻ antiport in winter flounder has not been ruled out. In fact, the paper by Halm et al. (1985) suggests that there is some other mechanism for the transport of NaCl present in the brush border membrane of winter flounder intestinal epithelial cells. This mechanism could be Na⁺/H⁺: C1⁻/HCO₃^– antiport. Eveloff et al. (1980) have already shown that Na⁺/H⁺ antiport is present in winter flounder intestinal BBMV. All that remains to be shown is that C1-/HCO₃^– antiport is present too. While the exact physiological details of NaCl transport in winter flounder intestinal epithelial cells are still somewhat unclear, our present experiments suggest that NaCl cotransport mechanisms are not present in tilapia intestinal epithelial cells. The lack of cotransport mechanisms in tilapia intestinal BBMV could be related to the diet and/or salinity-tolerance of this species. Tilapia are euryhaline herbivores, whereas winter flounder are stenohaline carnivores.

This study was supported by NSF grant number PCM 83-19973.