INTRODUCTION

From intracellular pH (pH_i), variations in cell membrane potential difference, and potassium transport measurements, Amlal et al. (1) described recently in this laboratory a new electroneutral K⁺/NH₄⁺(H⁺) antiport mechanism in intact medullary thick ascending limb (MTAL)¹ cells; the K⁺/NH₄⁺(H⁺) antiport mechanism was blocked by 10 mM barium and 0.1 mM verapamil, whereas it was unaffected by other K⁺ channel blockers such as quinidine and millimolar concentrations of amiloride or by the H⁺/K⁺-ATPase blocker SCH 28080. Because K⁺ conductances appeared unable to transport NH₄⁺ (1), which was in agreement with previous membrane vesicles (2) and electrophysiological (3, 4) studies of the MTAL, it was proposed (1) that the K⁺/NH₄⁺(H⁺) antiport mechanism could account for the luminal barium-sensitive NH₄⁺ transport that had been demonstrated in isolated and perfused MTAL tubules and hypothetically attributed to NH₄⁺ transport by MTAL apical K⁺ channels (5-7). However, because the tissue used in the previous work consisted of MTAL fragments in suspension, the apical or basolateral location of the K⁺/NH₄⁺(H⁺) antiport mechanism could not be assessed and remained uncertain (1). Yet, the knowledge of the cellular location of this NH₄⁺ transport pathway is necessary to understand the mechanisms of NH₄⁺ transepithelial transport by the MTAL, which is critical to ammonia accumulation in the renal medulla and subsequent secretion in adjacent medullary collecting ducts (8); this specialized NH₄⁺ renal pathway is thus of major importance with regard to urinary net acid excretion by which the kidney regulates acid-base balance.

The aims of the present work were therefore to establish directly the presence of a K⁺/H⁺ antiport mechanism in isolated MTAL membrane vesicles and to localize to the apical or basolateral plasma membrane this transport mechanism. To these purposes, we have adapted to a MTAL homogenate the principle of the Mg²⁺-EGTA aggregation plus differential centrifugation method to isolate a MTAL membrane vesicle preparation highly enriched in apical membrane; the preparation, however, also contained noticeable amounts of basolateral membrane. The results demonstrate that a barium- and verapamil-sensitive electroneutral K⁺/H⁺ antiport mechanism is located in the apical and not in the basolateral plasma membrane of the preparation. In addition, we show that the K⁺/H⁺ antiport mechanism is inhibited by arginine vasopressin (AVP) and 8-bromo-cAMP through the cAMP-dependent protein kinase (PKA) pathway.

EXPERIMENTAL PROCEDURES

The method used was that described previously in detail (9, 10), with some modifications designed to improve the MTAL fragment recovery. In brief, 14-22 kidneys of anesthetized male Sprague-Dawley rats (200-300 g in body weight) were bathed in situ for 1 min with ice-cold dissecting solution before rapidly removing them to avoid anoxic damage to medullary tissues and improve cell viability. The kidneys were then cut into thin slices along the corticopapillary axis into ice-cold Hanks' solution supplemented with 24 mM bicarbonate and bubbled with 95% O₂, 5% CO₂, pH 7.38. Small tissue pieces of inner stripes of outer medulla were then subjected at 37 °C to successive 6-min periods of collagenase digestion (0.40 g/liter in the above solution). To separate long MTAL fragments from isolated cells and small fragments of other medullary tissues, the pooled supernatants were first centrifuged at 150 × g for 2 min, and the resulting supernatant was discarded; the pellet was suspended in a medium containing 300 mM mannitol, 6 mM EGTA, 0.1 mM AEBSF, 12 mM Trizma (Tris base), pH 7.4, centrifuged at 230 × g for 2 min, and the supernatant was discarded. This maneuver was repeated, and the final tubular fragment pellet was resuspended in the latter medium, which gave the final MTAL suspension. The latter was made almost exclusively of MTAL tubules (>95%), occasional thin descending limb fragments, few medullary collecting tubules, and no isolated cells, as observed frequently by light microscopy. These characteristics were checked by electron microscopy, which confirmed that there was no proximal tubule fragment (S3) in this suspension obtained from carefully dissected inner stripes of outer medulla.²

MTAL cells were loaded with the pH-sensitive fluorescent probe BCECF, and pH_i was monitored at 37 °C as described in detail previously (9, 10).

Membrane vesicles were prepared from the MTAL suspension by modifications of the Mg²⁺-EGTA-aggregation method used by Biber et al. (11) to prepare proximal brush-border membranes. All steps of the fractionation procedure were performed at 2-4 °C and are schematized in Fig. 1. The MTALs suspension (21.4 ml) was homogenized first five times in a Dounce homogenizer and second in a Waring Blendor homogenizer at 22,000 rpm for 1.5 min; 30 ml of bi-distilled water with MgCl₂ was then added to the homogenate to obtain a final concentration of 12 mM MgCl₂ and 2.5 mM EGTA and a final osmolarity of ~170 mosmol/liter. After 20 min, the homogenate (H in Fig. 1) was divided into two ~25-ml fractions that were processed further in parallel; each fraction was centrifuged at 10,600 × g for 15 min with a Beckman J2-MC centrifuge and JS-13.1 rotor, which gave the pellet P1 and the supernatant S1 in Fig. 1. S1 was centrifuged at 48,000 × g for 24 min with a Sorvall RC-5B centrifuge and SS-34 rotor, which gave the vesicle-containing pellet P2 and the supernatant S2. The pooled membrane vesicles were resuspended in 15 ml of one of the loading solutions described below, and the resulting vesicle suspension was centrifuged at 48,000 × g for 40 min; this loading maneuver was repeated to ensure complete loading of the vesicles with the appropriate solution; the final membrane vesicle pellet was suspended manually and homogenized by repeated passage through a 27-gauge needle in a small volume (150-300 µl) of the loading solution. The membrane vesicle suspension was stored at 80 °C and used within 2 weeks for transport studies and enzyme assays.

All transport studies in membrane vesicles were performed at ambient temperature (22-25 °C) by a filtration technique using 0.45-µm cellulose acetate-nitrate filters (Millipore, HAWP). Membrane vesicles were thawed and kept on ice until use. When transport inhibitors were used, controls were performed in the presence of the vehicles of these inhibitors.

K⁺/H⁺ antiport activity was assessed by measuring the uptake of ⁸⁶Rb⁺, a substitute for K⁺, stimulated by an in > out H⁺ gradient. Membrane vesicles were loaded as described above with solution A, pH 6.6, or A, pH 8.0, listed in Table I. The uptake reaction was initiated by diluting and mixing 15-40-µg proteins of the membrane vesicles (10 µl) in 65 µl of isoosmolar solution B, pH 6.6, or C, pH 8.0 (Table I), containing in final concentrations 0.25-0.40 mM ⁸⁶RbCl (0.20-0.35 µCi/ml), 0.1 mM quinidine, and 1.5 mM furosemide to minimize background ⁸⁶Rb⁺ uptake by other MTAL K⁺ carriers such as quinidine-sensitive K⁺ channels (1, 3) and furosemide-sensitive K⁺(NH₄⁺)-Cl cotransport (12). The uptake reaction was terminated after timed periods by rapid dilution of the reaction mixture with 2 ml of an isoosmolar ice-cold "stop solution" containing 275.5 mM mannitol, 16 mM Hepes, 16.9 mM Tris, 10 mM BaCl₂, 0.1 mM quinidine, and 1.5 mM furosemide, pH 8.0; the latter transport inhibitors were present in the stop solution to prevent some release of ⁸⁶Rb⁺ from the vesicles into the cold stop solution from occurring. The diluted sample was poured immediately on a filter kept under suction and washed rapidly with an additional 12 ml of the ice-cold stop solution. The filter was placed in 4 ml of Ultimagold MV scintillation fluid (Packard) and counted by scintillation spectroscopy. Filter blanks were determined by diluting 10 µl of the loading solution in 65 µl of the reaction medium, and the resulting mixture was processed exactly as above. For one membrane vesicle preparation, each uptake condition and its filter blank were performed in triplicate; the mean filter blank value was subtracted from the experimental values to determine the amounts taken up by the vesicles. We first tested whether any binding of ⁸⁶Rb⁺ occurred in the membrane vesicles. For this purpose we have measured ⁸⁶Rb⁺ uptake at equilibrium (120-150 min) at various medium osmolarities that were adjusted with mannitol. As shown in Fig. 2, extrapolation of the regression line to infinite osmolarity (1/osmolarity approaching zero), a point at which the intravesicular free space should be negligible, indicates the presence of a binding component for ⁸⁶Rb⁺ at low Rb⁺ concentration (0.38 mM); the ⁸⁶Rb⁺ binding component accounted for 60% of the equilibrium value observed at standard osmolarity (340 mosmol/liter). As will be shown below (see Fig. 5B), ⁸⁶Rb⁺ binding was only intravesicular because there was no evidence for immediate binding; ⁸⁶Rb⁺ binding was thus secondary to ⁸⁶Rb⁺ transport inside the vesicles. Thus correction for a 60% binding component was made for equilibrium values but not for values at early times since the latter truly represented ⁸⁶Rb⁺ transport. At standard osmolarity (340 mosmol/liter), the distribution space of free ⁸⁶Rb⁺ was 2.6 µl/mg of protein.

Table I. Composition of experimental solutions used in vesicle studies Values are in mM and represent final concentrations. Osmolarity of all solutions was 340 mosmol/liter. In some experiments, solutions B and C contained up to 0.40 mM RbCl, and solution G contained 1 mM bumetanide replacing mannitol. Experiment A loada A loada B uptakeb C uptakeb D uptakec E uptakec F loadd G uptaked Mannitol 175.75 174.5 173.65 172.4 173.75 172.5 286 83 Hepes 120 65.5 120 65.5 120 65.5 16 16 Tris 14.25 70 14.25 70 14.25 70 8 8 MgCl2 10 10 10 10 10 10 Mg-(gluconate)2 10 10 RbCl 0.25 0.25 NaCl 0.25 0.25 0.5 KCl (or choline-Cl) 100 Quinidine 0.1 0.1 Furosemide 1.5 1.5 1.5 1.5 Amiloride 2 pH 6.6 8.0 6.6 8.0 6.6 8.0 7.4 7.4 a,b Rb+/H+ antiport. a,c Na+/H+ antiport. d Na+-K+-2Cl cotransport.

Na⁺/H⁺ antiport activity was assessed by measuring the uptake of ²²Na⁺ stimulated by an in > out H⁺ gradient. Membrane vesicles were loaded as described above with solution A, pH 6.6 (Table I). The uptake reaction was initiated by diluting and mixing 15-30-µg proteins of the membrane vesicles (10 µl) in 65 µl of isoosmolar solution D, pH 6.6, or E, pH 8.0 (Table I), containing in final concentrations 0.25 mM ²²NaCl (1.75 µCi/ml) and 1.5 mM furosemide to minimize background ²²Na⁺ uptake by Na⁺-K⁺-2Cl cotransport. The uptake reaction was terminated after timed periods by rapid dilution of the reaction mixture with 2 ml of an isoosmolar ice-cold stop solution containing 273.6 mM mannitol, 16 mM Hepes, 16.9 mM Tris, 10 mM MgCl₂, 1.5 mM furosemide, and 2 mM amiloride, pH 8.0. The diluted sample was poured immediately on a filter kept under suction and washed rapidly with an additional 9 ml of the ice-cold stop solution. Scintillation counting, determinations of filter blanks, and calculations were made as described above for ⁸⁶Rb⁺.

Na⁺-K⁺-2Cl cotransport activity was assessed by measuring the bumetanide-sensitive uptake of ²²Na⁺ stimulated by an inwardly directed 100 mM KCl gradient. Membrane vesicles were loaded as described above with solution F (Table I). The uptake reaction was initiated by diluting and mixing 10-20-µg proteins of membrane vesicles (10 µl) in 40 µl of isoosmolar solution G (Table I) containing 0.5 mM ²²NaCl (5.25 µCi/ml) and 2 mM amiloride to minimize background ²²Na⁺ uptake by Na⁺ channels and Na⁺/H⁺ exchange. The uptake reaction was terminated after timed periods by rapid dilution of the reaction mixture with 2 ml of an isoosmolar ice-cold stop solution containing 293 mM mannitol, 16 mM Hepes, 8 mM Tris, 5 mM NaCl, 5 mM KCl, 2 mM amiloride, and 1 mM bumetanide, pH 7.4; 5 mM NaCl and KCl were present in the stop solution to optimize bumetanide binding to the Na⁺-K⁺-2Cl cotransporter (for review, see Ref. 13). The diluted sample was poured immediately on a filter kept under suction and washed rapidly with an additional 6 ml of the ice-cold stop solution. Scintillation counting, determinations of filter blanks, and calculations were made as described above for ⁸⁶Rb⁺.

Protein amounts were determined by the method of Bradford (14); alkaline phosphatase activity was measured by the colorimetric determination of p-nitrophenol produced after sample addition by the hydrolysis of p-nitrophenyl phosphate used as a substrate (Sigma Diagnostics, Technical Bulletin 104). Na⁺/K⁺-ATPase activity was measured at 37 °C, pH 7.4, by a radioisotopic method after treatment of the various tissue fractions by sodium dodecyl sulfate (SDS) with bovine serum albumin as a detergent buffer, as described by Forbush (15). The assay was performed as follows. 9.5 µg/ml tissue proteins were preincubated for 10 min in the presence of 943.2 µg/ml bovine serum albumin and 67 µg/ml SDS (bovine serum albumin/SDS ratio = 14.1 and SDS/protein ratio = 7.1; osmolarity = 315 mosmol/liter); then the mixture was incubated for 15 min in the additional presence of 9 mM Tris-ATP and a tracer amount of [-³²P]ATP (final osmolarity ~ 195 mosmol/liter). The reaction was terminated by the addition of activated charcoal diluted in 1 N HCl (25%, w/v) and cooling to 2-4 °C. After centrifugation, the radioactivity of an aliquot of the supernatant containing ³²P was determined by standard liquid scintillation techniques. Samples without tissue were processed in parallel to determine the spontaneous breakdown of ATP, which was subtracted from the experimental values. Na⁺/K⁺-ATPase activity was calculated as the difference in ATP hydrolysis between that observed with 2.5 mM KCl plus 50 mM NaCl and that observed with 2.5 mM ouabain in the nominal absence of Na⁺ and K⁺. All media contained 1 mM EGTA, 10 mM MgCl₂, and 0.1 mM SCH 28080 to avoid any contribution of the MTAL ouabain-sensitive K⁺-ATPase activity (16).

Carrier-free ²²NaCl, ⁸⁶RbCl, and [-³²P]ATP were obtained from Amersham, Buckinghamshire, U. K. Collagenase CH grade II was obtained from Boehringer Mannheim. BCECF-AM ester was from Molecular Probes, Eugene, OR. D()Mannitol was from Merck, Darmstadt, Germany. H-89 was from Calbiochem. AEBSF, amiloride, bumetanide, furosemide, quinidine, verapamil, Tris-ATP, AVP, 8-bromo-cAMP, and all other chemicals were obtained from Sigma-chimie S.A.R.L., LaVerpillière, France. SCH 28080 was a gift from Schering-Plough Laboratories; HOE 694 was a gift from L. Counillon and J. Pouysségur.

Results are expressed as means ± S.E. Statistical significance between experimental groups was assessed by Student's t test or by one-way analysis of variance (ANOVA) completed by a t test using the within-groups residual variance of ANOVA, as appropriate. In some figures, curves fitting the data points were drawn by computer-assisted nonlinear regression (SigmaPlot, Jandel Scientific).

RESULTS

The specific activities, enrichment factors, and yields of marker enzymes for the apical membrane (alkaline phosphatase) and the basolateral membrane (Na⁺/K⁺-ATPase) are summarized in Table II. Alkaline phosphatase was selected as an enzyme marker for the MTAL apical membrane because, first, this enzyme copurifies with [³H]bumetanide binding sites in canine medullary membranes (17); and, second, Na⁺-K⁺-2Cl cotransport activity is strongly linearly correlated with the alkaline phosphatase enrichment factor of rat MTAL membranes (r = 0.91, p < 0.0001), a fact that we have established in another work (18). As shown in Table II, the membrane vesicle fraction P2 had a high enrichment factor (23.9 ± 2.6, ranging from 14 to 36) and a high yield (32.7 ± 4.7%, ranging from 15 to 56%) in the apical marker alkaline phosphatase. The enrichment factor (3.4 ± 0.2) and yield (4.7 ± 0.6%) in Na⁺/K⁺-ATPase, although seemingly low compared with those of alkaline phosphatase, must be considered as representing a relatively important amount of basolateral membranes because the total surface area of the basolateral membrane is ~24 times larger than that of the apical membrane of the rat MTAL in the inner stripe of outer medulla (19). However, the sealing of basolateral membranes is usually incomplete (for review, see Ref. 20); it was established in a previous work from this laboratory that only 30% of MTAL basolateral membranes isolated with a Ca²⁺ aggregation method were sealed and thus could contribute to solute transport by the vesicles (21). Electron microscopy examination showed that the preparation was composed essentially of round or oval vesicles having a mean diameter of ~0.3 µm². From these considerations we conclude that both apical and basolateral membrane vesicles were present in substantial amounts in the membrane vesicles used in the present study.

Table II. Specific activities, enrichment factors, and yields of marker enzymes in fractions obtained during the membrane isolation procedure Values are means ± S.E. obtained from nine preparations. Specific activities are in nmol/min/mg of protein. Fractions are those of Fig. 1. Fraction Protein Alkaline phosphatase Na+/K+-ATPase Quantity Yield Specific activity Enrichment factor Yield Specific activity Enrichment factor Yield mg % % % H 24.2  ± 3.5 25.9  ± 3.3 1,373  ± 67 P1 15.4  ± 2.4 62.9  ± 1.7 16.5  ± 2.5 0.7  ± 0.1 42.2  ± 4.9 1,572  ± 117 1.2  ± 0.1 71.9  ± 3.1 S2 6.9  ± 0.8 29.3  ± 1.2 12.8  ± 1.2 0.53  ± 0.05 15.6  ± 1.6 98  ± 19 0.08  ± 0.02 2.3  ± 0.6 P2 0.34  ± 0.07 1.4  ± 0.1 644  ± 140 23.9  ± 2.6 32.7  ± 4.7 4,776  ± 583 3.4  ± 0.2 4.7  ± 0.6 Total recovery 93.5  ± 2.2 90.5  ± 3.6 78.9  ± 3.0

The membrane vesicles displayed strong Na⁺-K⁺-2Cl cotransport activity, which is clear evidence for the presence of apical membrane vesicles in important amounts. As shown in Fig. 3A, most of 0.5 mM ²²Na⁺ uptake during early times was 1 mM bumetanide-sensitive in the presence of an inwardly directed 100 mM KCl gradient; the KCl gradient-stimulated ²²Na⁺ uptake probably reached after 30 s a value higher than the eventual equilibrium value (overshoot), whereas ²²Na⁺ uptake in the presence of 1 mM bumetanide increased almost linearly toward the equilibrium value. Thus the 1 mM bumetanide-sensitive component of 0.5 mM ²²Na⁺ uptake (Na⁺-K⁺-2Cl cotransport activity) reached by 30 s a maximum value of 320 ± 32 pmol/mg of protein (Fig. 3B). By contrast, in the presence of an inwardly directed 100 mM choline-Cl gradient, ²²Na⁺ uptake was ~75% lower than with KCl, which was not different from the uptake observed in the absence of a Cl gradient and was not affected by 1 mM bumetanide (not shown); thus there was no bumetanide-sensitive K⁺-independent Na⁺-Cl cotransport. Extrapolation of the curves in Fig. 3A to time zero indicates that an almost immediate ²²Na⁺ uptake occurred similarly in the presence or absence of bumetanide, probably because of rapid external binding of ²²Na⁺. Finally, ²²Na⁺ uptake equilibrium values were identical in the presence or absence of bumetanide (Fig. 3A), which indicates that bumetanide did not modify ²²Na⁺ binding. These observations demonstrate that the bumetanide-sensitive component of the KCl gradient-stimulated ²²Na⁺ uptake was mediated by the activity of the apical MTAL Na⁺-K⁺-2Cl cotransporter.

Na⁺-K⁺-2Cl cotransport.

Na⁺/H⁺ exchange (NHE) activity was demonstrated as the amiloride-sensitive H⁺ gradient-stimulated ²²Na⁺ uptake by the membrane vesicles. As shown in Fig. 4A, 0.25 mM ²²Na⁺ uptake at 10 s was stimulated more than 2-fold by an in (i) > out (o) H⁺ gradient (pH_i = 6.6/pH_o = 8.0); the H⁺ gradient-stimulated ²²Na⁺ uptake was abolished by 2 mM amiloride, a high amiloride concentration that completely blocks the NHE-1, NHE-2, and NHE-3 isoforms of the rat Na⁺/H⁺ antiport at this low Na⁺ concentration (22). It must be emphasized that 25 µM amiloride and 50 µM HOE 694 inhibited to similar extents (34.6 ± 8.0 and 39.9 ± 6.3%, respectively) NHE activity estimated as the difference in ²²Na⁺ uptake between the presence and absence of a H⁺ gradient (Fig. 4A); the latter effects could be explained by inhibition of the amiloride- and HOE 694-sensitive NHE-1 isoform but not by inhibition of the amiloride- and HOE 694-resistant NHE-3 isoform (23, 24). Indeed, Fig. 4B shows that ~90% of the 50 µM HOE 694-sensitive Na⁺/H⁺ antiport activity was blocked by concentrations of HOE 694 as low as 0.1 and 1 µM, which demonstrates blockade by HOE 694 of the NHE-1 isoform only (23, 24). Similarly, 25 µM amiloride inhibits NHE-1 by ~94% and NHE-3 by only ~20% in the presence of 0.25 mM Na⁺.³ These data thus indicate that both basolateral NHE-1 and apical NHE-3 activities of the rat MTAL (21, 25, 26) were detectable in our membrane preparation, which confirms that both basolateral and apical membrane vesicles were functionally present, as mentioned above. It may be noted that NHE-2 activity, if present in MTAL plasma membranes as it was suggested in preliminary studies (27), might hardly be detected under the present experimental conditions (0.25 mM external Na⁺) because the Michaelis constant for Na⁺ of rat NHE-2 is very high at 30-50 mM, whereas those of NHE-1 and NHE-3 are much lower at 10 and 5 mM, respectively (22, 28, 29); any weak NHE-2 activity should, nevertheless, also be blocked completely by 25 µM amiloride or 50 µM HOE 694 (23).

As shown in Fig. 5A, an in > out H⁺ gradient (pH_i = 6.6/pH_o = 8.0) stimulated 0.27 mM ⁸⁶Rb⁺ uptake by the membrane vesicles at 10 s compared with ⁸⁶Rb⁺ uptake observed in the absence of a H⁺ gradient at either pH_i = pH_o = 6.6 or pH_i = pH_o = 8.0. The H⁺ gradient caused an intravesicular accumulation of ⁸⁶Rb⁺ from 40 to 180 s to levels higher than the eventual equilibrium value (overshoot) (Fig. 5B). Extrapolation of the curve in Fig. 5B to time zero showed crossing of the y axis at 0 uptake, which indicates the absence of external binding of ⁸⁶Rb⁺. Because the equilibrium ⁸⁶Rb⁺ uptake values were identical at pH 6.6 or 8.0 (Fig. 5A), intravesicular ⁸⁶Rb⁺ binding, which can be deduced from the comparison of data in Figs. 2 and 5B, was constant and independent of the pH value. These results thus provide evidence that ⁸⁶Rb⁺ transport inside the vesicles was coupled by a Rb⁺/H⁺ exchange mechanism to the movement of H⁺ outside the vesicles. The difference in ⁸⁶Rb⁺ uptake observed in the presence and absence of a H⁺ gradient will be thereafter referred to as the H⁺ gradient-stimulated ⁸⁶Rb⁺ uptake.

As shown in Fig. 6, the Rb⁺/H⁺ exchange mechanism displayed the same pharmacological profile as that of the K⁺/NH₄⁺(H⁺) antiport described previously in intact MTAL cells (1). Indeed, the H⁺ gradient-stimulated ⁸⁶Rb⁺ uptake was abolished by 10 mM barium or 0.1 mM verapamil but was unaffected by 1 mM amiloride and 0.1 mM SCH 28080, a H⁺/K⁺-ATPase blocker; up to 2 mM amiloride, which blocks Na⁺ carriers and K⁺ channels as well at this concentration (1, 30, 31), did not affect the H⁺ gradient-stimulated ⁸⁶Rb⁺ uptake (see Fig. 10 below). Because 0.1 mM quinidine and 1.5 mM furosemide were present in these experiments, these results exclude K⁺ and Na⁺ channels, Na⁺/H⁺ exchangers, Na⁺-K⁺-2Cl and K⁺-Cl cotransports, and H⁺/K⁺-ATPase as having contributed to the H⁺ gradient-stimulated ⁸⁶Rb⁺ uptake, which will be thereafter referred to as a Rb⁺/H⁺ antiport mechanism.

It appeared interesting to specify the kinetics of the inhibition by verapamil because this agent is, beside barium, the sole known inhibitor of the K⁺/NH₄⁺(H⁺) antiport mechanism. As shown in Fig. 7, verapamil was inhibitory at concentrations  30 µM; a Dixon plot of the data (1/uptake against verapamil concentrations; not shown) yielded a straight line (r = 0.99) consistent with a single inhibitory site having an apparent inhibitory constant (K_i) of 55 µM.

To check whether the H⁺ gradient-stimulated ⁸⁶Rb⁺ uptake was electroneutral, as expected (1), we used the proton ionophore FCCP in the presence of an in > out H⁺ gradient to impose a conductive outwardly directed H⁺ flux and attendant inside negative membrane potential difference. As shown in Fig. 8, ⁸⁶Rb⁺ uptake was enhanced in the presence of 80 µM FCCP, which was thus due to stimulation of conductive ⁸⁶Rb⁺ transport pathways. The conductive ⁸⁶Rb⁺ transport pathways were partially inhibited to the same extent by 10-100 µM amiloride and probably blocked completely by 2 mM amiloride (Fig. 8). This amiloride sensitivity profile suggests that at least two amiloride-sensitive transport pathways accounted for conductive ⁸⁶Rb⁺ uptake. First, the previously described MTAL NH₄⁺(Na⁺) channel (1), which is sensitive to µM amiloride concentrations, transports K⁺ as well⁴ and thus probably Rb⁺. Second, residual quinidine-insensitive K⁺ channels, which may be blocked by high amiloride concentrations (1, 30, 31), could be responsible for the second component of conductive ⁸⁶Rb⁺ uptake which was blocked by 2 mM amiloride (Fig. 8). In any case, that up to 2 mM amiloride had no effect on the H⁺ gradient-stimulated ⁸⁶Rb⁺ uptake (see Figs. 6 and 10) indicates that ⁸⁶Rb⁺ uptake by conductive pathways was undetectable in the absence of FCCP and thus that there was no frank variation in the membrane potential difference under the latter condition. These results were consistent with the electroneutrality of the Rb⁺/H⁺ antiport mechanism and suggest that the spontaneous proton conductance in the membrane vesicles was low enough to generate no noticeable membrane potential difference in response to a H⁺ gradient. It is worth noting that when ⁸⁶Rb⁺ transport by conductive pathways was blocked maximally by 2 mM amiloride, ⁸⁶Rb⁺ uptake in the presence of a H⁺ gradient was reduced significantly in the presence of FCCP compared with controls in the absence of FCCP (p < 0.05; Fig. 8) to a level similar to those observed in the absence of a H⁺ gradient; this could be explained by rapid collapse of the proton gradient by FCCP.

Thus the results summarized in Figs. 5, 6, 7, 8 demonstrate the presence in the membrane vesicles of an electroneutral barium- and verapamil-sensitive Rb⁺/H⁺ antiport mechanism.

To localize the Rb⁺/H⁺ antiport mechanism to the apical or basolateral MTAL membranes that were both present in the preparation, as mentioned above, correlations were prospectively established between the Rb⁺/H⁺ antiport activity and the enrichment factors in alkaline phosphatase (apical marker enzyme) and Na⁺/K⁺-ATPase (basolateral marker enzyme) in nine successive membrane vesicle preparations; as mentioned above, spontaneous variations in the enrichment factors were large enough to serve this purpose. In the presence of a H⁺ gradient, ⁸⁶Rb⁺ uptake was strongly positively correlated with the alkaline phosphatase enrichment factor (r = 0.875, p < 0.01; Fig. 9A) and was negatively correlated with that of Na⁺/K⁺-ATPase (r = 0.665, p < 0.05; Fig. 9B). There was no significant correlation between ⁸⁶Rb⁺ uptake and the marker enzymes in the absence of a H⁺ gradient (r < 0.35 with n = 5 for both enzymes). These results clearly demonstrate that the Rb⁺/H⁺ antiport mechanism is located in the apical membranes and not in the basolateral membranes of the preparation.

To confirm the conclusion that the Rb⁺/H⁺ antiport mechanism is located in the MTAL apical membrane by another experimental approach, we have looked for a functional interaction between Rb⁺/H⁺ antiport and the rat MTAL apical NHE-3 isoform (21, 32). Indeed, the presence of Na⁺/H⁺ antiport activity in the vesicles that contain the Rb⁺/H⁺ antiport mechanism could affect ⁸⁶Rb⁺ transport by the latter by collapsing the H⁺ gradient, like FCCP as suggested above. Thus the H⁺ gradient-stimulated ⁸⁶Rb⁺ uptake was measured in the presence or absence in the uptake medium of 5 mM Na⁺; measurements were also made in the presence or absence of 5 mM K⁺ for comparison and to check the expected competition between K⁺ and ⁸⁶Rb⁺. We used amiloride instead of HOE 694 as the inhibitor of Na⁺/H⁺ antiporters in most of the experiments because HOE 694 cannot be used at concentrations high enough to block NHE-3. As shown in Fig. 10, the H⁺ gradient-stimulated ⁸⁶Rb⁺ uptake was abolished in the presence of 5 mM K⁺ and incompletely but significantly reduced in the presence of 5 mM Na⁺. On the one hand, both the H⁺ gradient-stimulated ⁸⁶Rb⁺ uptake and the K⁺ effect on the latter were unaffected by 2 mM amiloride (compare Fig. 10, A and C), which is consistent with competition between K⁺ and ⁸⁶Rb⁺ on the amiloride-insensitive K⁺(Rb⁺)/H⁺ antiport mechanism. On the other hand, 25 µM amiloride, an amiloride concentration that blocks by ~91% the amiloride-sensitive NHE-1 isoform, and Na⁺ channels as well, but only by ~11% the NHE-3 isoform in the presence of 5 mM sodium,³ did not prevent 5 mM Na⁺ from significantly reducing the H⁺ gradient-stimulated ⁸⁶Rb⁺ uptake (Fig. 10B); this, which was also observed with 50 µM HOE 694 (not shown), indicates that the inhibitory effect of 5 mM Na⁺ on ⁸⁶Rb⁺ uptake could not be explained by an interaction between basolateral NHE-1 and the K⁺(Rb⁺)/H⁺ antiport mechanism. By contrast, the Na⁺ effect on the H⁺ gradient-stimulated ⁸⁶Rb⁺ uptake was abolished by 2 mM amiloride, which blocks NHE-3 at this Na⁺ concentration³ (Fig. 10C). These results demonstrate a functional interaction between the K⁺(Rb⁺)/H⁺ antiport mechanism and NHE-3 and thus show that these carriers are located in the same apical membrane vesicles. It may be noted that 0.1 mM quinidine, a weak inhibitor of Na⁺/H⁺ antiport (33), was present in the experiments illustrated in Fig. 10. We have thus checked the extent of inhibition of Na⁺/H⁺ antiport in the membrane vesicles by quinidine in 5 mM ²²Na⁺ uptake experiments. In three membrane vesicle preparations, 0.1 mM quinidine inhibited Na⁺/H⁺ antiport activity by 3.2 ± 2.6 and 1.2 ± 1.2% in the absence and presence of 50 µM HOE 694, respectively. Thus 0.1 mM quinidine had negligible effects on Na⁺/H⁺ antiporters under these experimental conditions.

To assess possible effects of stimulation of the PKA pathway on the K⁺/NH₄⁺(H⁺) antiport mechanism, studies were performed on MTAL suspensions with pH_i measurements. As described previously (1), MTALs were preincubated in a Na⁺-free high K⁺ medium, pH 7.4 (solution A in Table III) in which Na⁺-dependent transporters including some NH₄⁺ carriers are not operational. When these K⁺-loaded cells were diluted in the fluorometer cuvette in the same Na⁺-free high K⁺ medium, pH_i was 7.35-7.45 (Fig. 11), i.e. almost equal to the extracellular pH; this was because of equilibration of the transmembrane H⁺ gradient on the K⁺ concentration gradient ([K⁺]_i ~ [K⁺]_o) in this medium in which K⁺/NH₄⁺(H⁺) antiport is the sole functional H⁺ transport mechanism (MTAL cells are normally acidified in a Na⁺-free low K⁺ medium (9)). There was no difference in the preincubation pH_i values whether the cells were treated or not by 8-bromo-cAMP or AVP.

Table III. Composition of experimental solutions used in pHi measurements Concentrations are in mM. All solutions were bubbled with 100% O2, adjusted to pH 7.4 with Tris, and contained 0.1 g/liter bovine serum albumin. A B mM NMGa-Cl 143 KCl 143 K2HPO4 0.8 0.8 KH2PO4 0.2 0.2 CaCl2 1 1 MgCl2 1 1 NH4Cl 4 4 Hepes 10 10 Glucose 5 5 L-Leucine 5 5 a N-Methyl-D-glucamine.

By contrast, when the K⁺-loaded cells were diluted abruptly into a Na⁺-free low K⁺ medium (solution B in Table III, final K⁺ concentration of ~8.5 mM) to impose an outward directed K⁺ concentration gradient, an intracellular acidification occurred (Fig. 11) caused by NH₄⁺(H⁺) influx coupled to K⁺ efflux through the K⁺/NH₄⁺(H⁺) antiport mechanism (the cell acidification was blocked completely by 10 mM barium; not shown), as demonstrated previously (1). 0.1 mM quinidine, 1 µM amiloride, and 1.5 mM furosemide were present in these experiments to block K⁺ and NH₄⁺ conductances and K⁺(NH₄⁺)-Cl cotransport, respectively (1, 12). Thus only direct, not indirect, effects on K⁺/NH₄⁺(H⁺) antiport can be observed with this experimental protocol.

Preincubation of the cells with 0.5 mM 8-bromo-cAMP for 4 min inhibited the K⁺/NH₄⁺(H⁺) antiport activity by 52% (Fig. 11); the initial rate of the cell acidification (dpH_i/dt)⁵ decreased from 1.09 ± 0.12 pH unit/min in control (n = 14) to 0.52 ± 0.21 in 8-bromo-cAMP-treated tubules (n = 15; p < 0.03). Of note, there was no more difference in the pH_i values after 30 s probably because the K⁺ concentration gradient was not maintained in the Na⁺-free low K⁺ medium. The inhibitory effect of 8-bromo-cAMP was reproduced by AVP, a peptide hormone known to control MTAL transports essentially through the cAMP pathway. As shown in Fig. 12, 10⁸ M AVP decreased the dpH_i/dt from 0.87 ± 0.10 pH unit/min in control tubules (n = 17) to 0.58 ± 0.09 (n = 18; p < 0.04). This inhibitory AVP effect was abolished by 15 µM H-89 (Fig. 12), which completely blocks PKA in MTAL cells at this concentration (34). H-89 also abolished the inhibitory effect of 8-bromo-cAMP (dpH_i/dt was 0.86 ± 0.27 pH unit/min in control (n = 8) versus 0.80 ± 0.20 (n = 10); not shown). Thus, activation of PKA by 8-bromo-cAMP or AVP inhibited K⁺/NH₄⁺(H⁺) antiport activity.

8

DISCUSSION

Present results demonstrate for the first time the presence of a K⁺/H⁺ antiport mechanism in membrane vesicles of the rat MTAL, localize this transport mechanism to the apical membrane of MTAL cells, and show that AVP inhibits K⁺(NH₄⁺)/H⁺ antiport through the cAMP/PKA pathway.

In the proximal tubule, the apical brush-border membrane is considerably developed so that its surface area is nearly twice that of the basolateral membrane (19); thus almost pure preparations of proximal brush-border membranes can be obtained in amounts sufficient for transport studies (35). By contrast, the apical membrane surface of the MTAL of the inner stripe of the kidney outer medulla (iMTAL) is small compared with that of other membranes of iMTAL cells; in particular, the basolateral membrane surface area is ~24 times larger than that of the apical membrane in the rat iMTAL (19). One may calculate that a pure MTAL apical membrane fraction should have an enrichment factor of 65 in an ideal apical marker (35), which probably would imply, with use of current purification methods, a recovery of very low amounts of apical membrane from a reasonable number of rat kidneys because the iMTAL is a very short nephron segment. With a simple Mg²⁺-EGTA aggregation plus differential centrifugation method applied to a homogenate of MTALs harvested from 14-22 rat kidneys, we have obtained a membrane fraction with relatively high mean enrichment factor (~24) and yield (~33%) in alkaline phosphatase, which is an apical enzyme marker in the MTAL as discussed above. Indeed, the membrane vesicles displayed strong Na⁺-K⁺-2Cl cotransport activity that compares favorably with results obtained previously by others under the same experimental conditions of KCl gradient and ²²Na⁺ concentration; Na⁺-K⁺-2Cl cotransport was responsible for an uptake of 320 ± 32 pmol of Na⁺/mg protein for 30 s in this study versus 95 ± 13 in the rat medullary membrane study of Reeves et al. (36). This difference is quantitatively attributable to the relatively low enrichment in MTAL apical membranes of the membrane preparation of the latter work; indeed, both Na⁺-K⁺-2Cl cotransport activity and alkaline phosphatase enrichment factor (7.5 ± 0.6) were reduced ~4-fold (36) compared with the corresponding values of the present study. This correlation between Na⁺-K⁺-2Cl cotransport activity and alkaline phosphatase enrichment factor derived from results obtained by different laboratories further establishes alkaline phosphatase as being a marker enzyme of the MTAL apical membrane. In a previous work in this laboratory (21), the MTAL apical membrane fraction obtained by a Ca²⁺ aggregation method had an enrichment factor of 10 in alkaline phosphatase, but Na⁺-K⁺-2Cl cotransport activity had not been measured. We have used in the present study Mg²⁺-EGTA instead of Ca²⁺ because membranes obtained after Mg²⁺ precipitation in Ca²⁺-free solutions generally have low leak permeabilities, particularly for protons (20, 37), which was desirable in the present work aimed at studying a proton carrier; as mentioned above, the present membrane vesicles should have a low proton leak permeability that was undetectable by usual transport studies. The vesicle preparation contained, however, noticeable amounts of MTAL basolateral membranes; this was demonstrated by the observation, based on results obtained with HOE 694 and amiloride, that the total Na⁺/H⁺ exchange activity of the vesicles was accounted for by both basolateral NHE-1 (~40%) and apical NHE-3 (~60%).

The presence of a K⁺/H⁺ antiport mechanism in the rat MTAL is demonstrated in the present study of MTAL membranes. Indeed, ⁸⁶Rb⁺ uptake, which occurred competitively with K⁺, was stimulated to levels higher than equilibrium values by an in > out H⁺ gradient. The H⁺ gradient-stimulated ⁸⁶Rb⁺ uptake was abolished by 10 mM barium and 0.1 mM verapamil but unaffected by quinidine and mM concentrations of amiloride that are K⁺ channels blockers; verapamil inhibited the K⁺(Rb⁺)/H⁺ antiport mechanism with an apparent K_i of 55 µM under the present experimental conditions. Furthermore, the H⁺ gradient-stimulated ⁸⁶Rb⁺ uptake was electroneutral because it contained no conductive component, as demonstrated in experiments using the proton ionophore FCCP and amiloride that inhibited conductive transport pathways of Rb⁺. Thus the K⁺(Rb⁺)/H⁺ antiport mechanism detected in the membrane vesicles displayed the same properties as those described previously for the K⁺/NH₄⁺(H⁺) antiport in intact MTAL cells (1), which strongly suggests that these studies were dealing with the same carrier. The transport activity of the K⁺(Rb⁺)/H⁺ antiport mechanism appeared substantial because amounts of ⁸⁶Rb⁺ uptake were similar to those of ²²Na⁺ uptake by Na⁺/H⁺ antiport under similar experimental conditions of pH gradient and extracellular cation concentration. However, since the kinetic characteristics (Michaelis constant and maximum velocity) of the K⁺(Rb⁺)/H⁺ antiport mechanism are unknown, a precise quantitative comparison between K⁺(Rb⁺)/H⁺ and Na⁺/H⁺ antiport activities cannot be done. The molecular identity of the K⁺(Rb⁺)/H⁺ antiport mechanism is unknown at present; this transport mechanism is not a H⁺/K⁺-ATPase because it is insensitive to SCH 28080.

The K⁺(Rb⁺)/H⁺ antiport mechanism, which was first described in intact MTAL cells and was thus necessarily located in the plasma membrane of these cells (1), has been localized to the apical and not basolateral membranes of the vesicle preparation in the present study. First, the K⁺(Rb⁺)/H⁺ antiport activity was positively correlated with the apical marker alkaline phosphatase and negatively with basolateral Na⁺/K⁺-ATPase; these findings provide strong evidence for the apical location of the K⁺(Rb⁺)/H⁺ antiport because alkaline phosphatase activity colocalizes with the MTAL Na⁺-K⁺-2Cl cotransporter (17, 18). Second, a functional interaction has been demonstrated between K⁺(Rb⁺)/H⁺ antiport and NHE-3 but not NHE-1 based on results obtained with NHE inhibitors (amiloride and HOE 694); this observation is also of strong significance because NHE-3 and NHE-1 are established as apical and basolateral carriers, respectively, in the MTAL (21, 32). Because the membrane preparation had a high yield in apical membranes (32.7 ± 4.7%), it can be concluded that a K⁺(Rb⁺)/H⁺ antiport mechanism is present in the apical membrane of the rat MTAL. On the other hand, although our results exclude the presence of K⁺/H⁺ antiport in the basolateral vesicles of the preparation, they do not rule out completely that K⁺/H⁺ antiport might also be present, like Na⁺/H⁺ exchange, in the basolateral membrane of the MTAL because the membrane preparation used in the present study had a low yield in basolateral membranes (4.7 ± 0.6%), which might not have been representative of the whole MTAL basolateral membrane.

The apical MTAL K⁺(Rb⁺)/H⁺ antiport mechanism was previously demonstrated to be actually an NH₄⁺ carrier because it functions at rates higher with NH₄⁺ (in the K⁺/NH₄⁺ mode) than with H⁺ at physiological concentrations of these ions (1). K⁺/NH₄⁺ antiport, driven by the outward directed K⁺ concentration gradient, should provide physiologically an efficient means of luminal NH₄⁺ uptake as well as of K⁺ recycling in the tubular fluid and could account for a substantial part of the luminal NH₄⁺ transport in the MTAL. Indeed, ~40% of active ammonia absorption is not attributable to Na⁺-NH₄⁺-2Cl cotransport activity in the rabbit MTAL (38). The luminal barium-sensitive NH₄⁺ transport, which is attributable to the K⁺/NH₄⁺(H⁺) antiport mechanism and not to K⁺ channels as discussed above, was responsible for 30-59% and ~45% of total apical NH₄⁺ transport in the isolated and perfused iMTAL of the rat (7) and mouse (6), respectively, as assessed by pH_i measurements. These results show that K⁺/NH₄⁺(H⁺) antiport contributes significantly to MTAL NH₄⁺ transport; considering the activity of the apical K⁺/NH₄⁺(H⁺) antiport mechanism is thus necessary when studying bicarbonate and NH₄⁺ absorption by the MTAL, all the more so because this transport mechanism may be regulated by hormonal intracellular messengers. Indeed, we have shown in the present study that PKA, activated by 8-bromo-cAMP or AVP, inhibits the K⁺/NH₄⁺(H⁺) antiport mechanism. Inhibition of the apical K⁺/NH₄⁺(H⁺) antiport mechanism by AVP could explain, at least in part, why arginine vasopressin does not enhance transepithelial ammonia absorption in the rat MTAL (39) despite an expected stimulation of Na⁺-NH₄⁺-2Cl cotransport, the other main MTAL apical NH₄⁺ transport pathway. Thus the apical K⁺/NH₄⁺(H⁺) antiport mechanism may also serve to dissociate MTAL ammonia and NaCl absorption under certain physiological conditions.