Renal Physiology

Cation and voltage dependence of rat kidney electrogenic Na+- Formula cotransporter, rkNBC, expressed in oocytes

Christopher M. Sciortino, Michael F. Romero


Recently, we reported the cloning and expression of the rat renal electrogenic Na+- HCO3 cotransporter (rkNBC) in Xenopusoocytes [M. F. Romero, P. Fong, U. V. Berger, M. A. Hediger, and W. F. Boron. Am. J. Physiol. 274 (Renal Physiol. 43): F425–F432, 1998]. Thus far, all NBC cDNAs are at least 95% homologous. Additionally, when expressed in oocytes the NBCs are1) electrogenic,2) Na+ dependent,3) HCO3 dependent, and4) inhibited by stilbenes such as DIDS. The apparent HCO3 :Na+coupling ratio ranges from 3:1 in kidney to 2:1 in pancreas and brain to 1:1 in the heart. This study investigates the cation and voltage dependence of rkNBC expressed inXenopus oocytes to better understand NBC’s apparent tissue-specific physiology. Using two-electrode voltage clamp, we studied the cation specificity, Na+ dependence, and the current-voltage (I-V) profile of rkNBC. These experiments indicate that K+ and choline do not stimulate HCO3 -sensitive currents via rkNBC, and Li+ elicits only 3 ± 2% of the total Na+ current. The Na+ dose response studies show that the apparent affinity of rkNBC for extracellular Na+ (∼30 mM [Na+]o) is voltage and HCO3 independent, whereas the rkNBCI-Vrelationship is Na+ dependent. At [Na+]o v max (96 mM), theI-Vresponse is approximately linear; both inward and outward Na+- HCO3 cotransport are observed. In contrast, only outward cotransport occurs at low [Na+]o(<1 mM [Na+]o). All rkNBC currents are inhibited by extracellular application of DIDS, independent of voltage and [Na+]o. Using ion-selective microelectrodes, we monitored intracellular pH and Na+ activity. We then calculated intracellular [ HCO3 ] and, with the observed reversal potentials, calculated the stoichiometry of rkNBC over a range of [Na+]ovalues from 10 to 96 mM at 10 and 33 mM [ HCO3 ]o. rkNBC stoichiometry is 2 HCO3 :1 Na+ over this entire Na+ range at both HCO3 concentrations. Our results indicate that rkNBC is highly selective for Na+, with transport direction and magnitude sensitive to [Na+]oas well as membrane potential. Since the rkNBC protein alone in oocytes exhibits a stoichiometry of less than the 3 HCO3 :1 Na+ thought necessary for HCO3 reabsorption by the renal proximal tubule, a control mechanism or signal that alters its in vivo function is hypothesized.

  • sodium/bicarbonate cotransport
  • NBC
  • Xenopus oocyte expression
  • intracellular pH
  • sodium transport
  • bicarbonate transport
  • kinetics
  • voltage clamp

the electrogenicNa+- HCO3 cotransporter was first described in the renal proximal tubule (8) and later cloned from the salamander kidney by functional expression usingXenopus oocytes (34). By homology, several Na+- HCO3 cotransporter (NBC) cDNA isoforms have been cloned from different mammalian tissues including rat kidney (33), human kidney (9), human heart (13), human pancreas (2, 29), and brain (5). These proteins are either 1,035 (kidney) or 1,079 (other organs) amino acids in length and are at least 95% identical to one another. Variation in sequence occurs predominantly in the NH2-terminal 45–85 amino acids. Expression studies in oocytes (13, 33, 34) show that the basic functions of these NBC isoforms are similar: they are1) electrogenic,2) Na+ dependent,3) HCO3 dependent, and4) inhibited by stilbenes such as DIDS.

Although Na+- HCO3 cotransport systems are physiologically implicated in many tissues, these cotransporters appear to function differently depending on the tissue. NBC is located at the basolateral membrane of renal proximal tubule cells (37) and electrogenically moves Na+/ HCO3 out of the cell into the blood. This mechanism is responsible for 80–90% of HCO3 reabsorbed in the kidney. However, in pancreatic ductal cells, Na+/ HCO3 influx is thought to occur (19). In cultured hippocampal glial cells (6, 36) and in the eye (21, 23), both HCO3 influx and efflux have been measured and are electrogenic (6, 26, 27). Yet recovery from an acid load elicited by increasing heart rate in cat papillary muscle is also attributed to electrogenic Na+/ HCO3 influx (3, 10, 11), whereas experiments in guinea pig ventricular myocytes indicate that the Na+- HCO3 cotransport is electroneutral (24).

Electrogenic NBC transport moves a net negative charge in the direction of transport (i.e., a ratio of HCO3 :Na+> 1:1). Both electrochemical gradients and cell membrane potential dictate the direction of ion flux. The HCO3 :Na+coupling ratio has been used as a predictor of transport direction. The larger this ratio, the more effective the cell potential acts as a driving force to move Na+/ HCO3 out of the cell against the Na+gradient. Studies of the electrogenic Na+/ HCO3 cotransporter in renal tubules (48), proximal tubule cell lines (16), and vesicles made from rabbit basolateral membranes (39) predict a HCO3 :Na+coupling of 3:1. It is thought that this ratio is necessary to move HCO3 out of the proximal tubule cell across the basolateral membrane. However, the Na+- HCO3 cotransporter in the pancreas and brain is thought to have a coupling ratio of 2:1 (6, 15, 19) and the heart may either be 2:1 (11) or 1:1 (14, 24).

Although different physiological characteristics have been attributed to these different tissues, the NBC clones are greater than 95% homologous. Thus the purpose of the present study was to elucidate NBC transport characteristics to understand how and if NBC can fulfill all of these roles. Specifically, we expressed rkNBC inXenopus oocytes to1) determine the monovalent cation specificity, 2) define the voltage dependence, and 3) determine apparent affinity for extracellular Na+([Na+]o). For rkNBC specificity, all ionic and current changes attributed to rkNBC are inhibitable by the stilbene, DIDS. From measurements of the [Na+]odose response of rkNBC-stimulated HCO3 current, and with measurements of intracellular pH (pHi) and sodium activity (aNai), we directly calculate the stoichiometry of rkNBC-mediated Na+- HCO3 cotransport over a range of [Na+]olevels. Interestingly, our studies indicate that the stoichiometry of rkNBC is 2 HCO3 :1 Na+, rather than 3:1 as previously reported for in vitro tissue studies, and is independent of the Na+ gradient and [ HCO3 ]o.


Oocyte Experiments

Oocyte isolation and injection. Xenopus laevis were purchased from Xenopus Express (Beverly Hills, FL). Oocytes were removed and collagenase dissociated as previously described (32, 33). To optimize rkNBC expression, we used a rkNBC-cDNA construct in theXenopus expression vector pTLN2 (33). Capped cRNA was synthesized using a linearized cDNA template and the SP6 mMessage mMachine kit (Ambion, Austin, TX). Oocytes were injected with 50 nl of rkNBC cRNA (0.2 μg/μl) or water and incubated at 18°C in OR3 media (32). Oocytes were studied 3–10 days after injection. Each experimental procedure was studied on at least two batches of oocytes from differentXenopus to account for possible biologic variations between animals.


Solutions. Experimental solutions are detailed in Table 1.

View this table:
Table 1.

Experimental solutions

Two-electrode voltage clamp. Oocyte currents were recorded with OC-720C voltage clamp (Warner Instruments, Hamden, CT). Electrodes were fashioned from borosilicate glass using a model P-97 puller (Sutter, Novato, CA). Electrode tips were filled with 1% agarose/3 M KCl and backfilled with 3 M KCl. Current and voltage electrodes had resistances of 0.5–1 MΩ. Current signals were filtered with an eight-pole Bessel filter (−3-dB cutoff, frequency of 2–5 kHz) and digitized at 10 kHz. Current and voltage signals were acquired via an EPC-16 I/O interface using Pulse software, and data were analyzed using the PulseFit program (HEKA). Oocytes were clamped at a holding potential (V h) of −60 mV; and current was constantly monitored and recorded at 1 Hz. Initial experiments determined that currents elicited by voltage steps saturated in 50 ms to steady-state levels. Thus current-voltage (I-V) protocols consisted of 70-ms steps fromV h to potentials from −160 mV to 60 mV in 20-mV steps. The mean steady-state current is plotted against voltage (Fig. 2,A–E).

Ion-selective microelectrode.Ion-selective microelectrodes were used to monitor pHi andaNaiof rkNBC and water-injected oocytes as previously described (33). Intracellular ion activity was measured as the difference between the ion-selective electrode (pH or Na+) and a KCl voltage electrode impaled into the oocyte; membrane potential (V m) was measured as the potential difference between the KCl microelectrode and an extracellular calomel (33, 34). Briefly, ion-selective microelectrodes were fabricated using filamented borosilicate glass pulled to 0.5-μm tips and silanized at 210°C withbis-(dimethylamino)dimethylsilane (Fluka, Ronkonkoma, NY), and the shanks were coated with Sylgard (Dow Corning, Midland, MI). Micropipettes were cooled under vacuum, and the tips were filled with either Na+ionophore cocktail A or H+ionophore I-cocktail B ion-selective resin (Fluka Chemical). Na+electrodes were backfilled with 150 mM NaCl. H+ electrodes were backfilled with (in mM) 40 KH2PO4, 23 NaOH, and 15 NaCl, pH 7.0. pH electrodes were calibrated using pH 6.0 and 8.0 (traceable National Bureau Standards; Fisher Scientific, Pittsburgh, PA) followed by point calibration in ND96 (pH 7.50,solution 1, Table 1). Na+ electrodes were calibrated with 10 mM and 100 mM NaCl, and the specificity was checked using 100 mM KCl, followed by point calibration in ND96 (96 mM Na+). Na+ electrodes had a selectivity1of 52 ± 5 for Na+ over K+ (calculated as described by Abdulnour-Nakhoul and coworkers Ref. 1). Both types of ion-selective electrodes had slopes of at least −56 mV/decade change.

Buffering power was calculated as previously reported (35). Briefly, the total apparent buffering power (βT, see Table2) is defined as the change in [ HCO3 ] before and after application of CO2/ HCO3 (once steady-state is reached) divided by the change in pHi elicited from the same solution changes, i.e., βT = Δ[ HCO3 ]steady-state/ΔpHi.

View this table:
Table 2.

Intracellular ion activity measurements

Cation selectivity and apparent extracellular Na+vmax.

The cation selectivity of rkNBC-mediated cotransport and the apparent maximal extracellular Na+-stimulated HCO3 current (v max) were measured using the bath solution protocol shown in Fig.4 A. The ability of K+, Li+, and choline to stimulate HCO3 -dependent current via rkNBC is tested by measuring cation current responses ± HCO3 . Briefly, an oocyte was perfused with a test cation (solution 3,4, or5, Table 1) non- HCO3 solution and anI-Vrelation was recorded. The bath solution was switched to the respective HCO3 /test cation solution (1.5% CO2/10 mM HCO3 /pH 7.5,solutions 8–12 and5, Table 1), i.e., Li+/ND96 (solution 3, Table 1) to Li+/1.5% CO2/ HCO3 (solution 10, Table 1). AnI-Vrelation was measured at the peak HCO3 -stimulated current. The HCO3 -dependent response was calculated as the difference between the twoI-Vrelationships (Fig. 4). The [Na+]oat which rkNBC-mediated HCO3 current saturates (i.e., attains an apparent currentv max) was determined using the solution protocol in Fig.4 A. The HCO3 -mediatedI-Vresponses for [Na+]oranging from 84 to 120 mM were tested. Solution osmolalities were matched with choline (solutions 2,6, 9, 13, Table 1).

[Na+]oconcentration dependence.

Figure1 A shows a schematic of the experimental protocol used to study the extracellular Na+ dependence of rkNBC. Our protocol was designed to monitor Na+-induced currents while maintaining pHi andaNaiat steady state. An oocyte was put into an ∼100-μl perfusion chamber and perfused at 8–10 ml/min. Dye exchange experiments showed that the entire chamber volume was exchanged in <1 s. Therefore, the initial current response of rkNBC cotransport can be measured without cell rundown due to long mixing times. TheV h (−60 mV) current of rkNBC-expressing oocytes was recorded for the duration of the experiment at 1 Hz. A baseline non- HCO3 I-Vrelation was recorded once current stabilized after initial electrode impalements. The bath solution was changed to 1.5% CO2/10 mM HCO3 /96 mM Na+ (solution 8, Table 1) or 5% CO2/33 mM HCO3 /96 mM Na+ (solution 14, Table 1) for 10 min to allow a steady-state current and pHi to be reached (compare Figs. 1 B and3 B). Unless otherwise specified, Na+ replacement was with choline. The bath solution was changed to 0 Na+ (solution 12, Table 1) for 24 s, pulsed for 8 s to a test Na+ (in mM 0, 1, 10, 24, 36, 48, 72, 84, 96), then to 96 mM Na+ for 24 s. AnI-Vrelation was recorded at the peak current induced by each Na+ change. This quick pulse protocol maintains a steady-state pHi andaNai, and baseline 0 Na+ current, allowing three to six randomly ordered test measurements of [Na+]oper oocyte to be taken and ensures that test Na+ responses are measured from the same “fully outward” transport state of rkNBC (Figs.1 B and2 B).

Fig. 1.

Current-voltage (I-V) response protocol.A: model of Na+ dose protocol. Rat renal electrogenic Na+- Embedded Image cotransporter (rkNBC) cRNA or water-injected (control) oocytes were voltage clamped to a holding potential (V h) of −60 mV in ND96 (solution 1, Table 1). Bath solution was then switched to 1.5% CO2/10 mM Embedded Image or 5% CO2/33 mM Embedded Image at pH 7.5/96 mM Na+ (solutions 8 and 14, respectively, Table 1) for 10 min. Extracellular Na+([Na+]o) was then removed (0 Na+, choline replacement solutions 9 and15, respectively, Table 1) for 24 s, followed by a 12-s perfusion in test Na+ (in mM: 1, 10, 24, 36, 48, 72, 84, 96). AnI-Vresponse was recorded at the peak current induced by the Na+ solution (x). Na+ was reapplied for 24 s. The 0-test-96 mM Na+ pulse sequence was used to maintain a stable baseline current and was repeated for 3–6 test Na+ concentrations per oocyte. B:V h recorded during a Na+ dose response experiment. V hcurrent response of an rkNBC-expressing oocyte resulting from the solution protocol in A was monitored at 1 Hz. A steady-state current was reached in 8–10 min after 1.5% CO2/10 mM Embedded Image administration. Test pulses are to 96, 48, and 24 mM Na+. Peak current stimulated by these test pulses was Na+ concentration dependent.

Fig. 2.

I-Vresponse of rkNBC. A: steady-state current sweeps ± [Na+]o± DIDS. Unsubtracted current sweep data recorded from of a rkNBC-expressing oocyte are shown in presence of extracellular 1.5% CO2/10 mM Embedded Image with either 96 mM Na+ (black squares,solution 8) or Na+ replaced with choline (green diamonds, solution 12). Addition of 200 μM DIDS to the CO2/ Embedded Image -containing bath solution irreversibly inhibited current in presence (red square) or absence (red diamond) of extracellular Na+.B: rkNBCI-Vrelationship ± [Na+]o± DIDS. ND96 (non- Embedded Image )I-Vrelationship is subtracted fromI-Vrelations measured from current sweep data inA to yield the Embedded Image -stimulated current. Current response in presence of extracellular Na+ (black squares) is approximately linear, whereas in absence of Na+ (green diamonds) only negative currents are observed. After a 5-min exposure to 200 μM DIDS (red), while still in CO2/ Embedded Image , all current was blocked. C: control oocyteI-Vrelationships. A water-injected control oocyte (H2O) was voltage clamped, and anI-Vresponse was recorded in ND96 (blue triangles,solution 1), 1.5% CO2/10 mM Embedded Image /pH 7.5 with 96 mM Na+ (black squares,solution 8), 0 Na+ ND96 (green diamonds,solution 5), and in presence of 200 μM DIDS (red solutions 1 and8 plus DIDS, respectively). The fiveI-Vresponses of the water-injected oocyte are superimposable, indicating absence of an endogenous Embedded Image -stimulated current. On the same axis, a rkNBC-expressing oocyte (rkNBC) was voltage clamped, and anI-Vrelationship was recorded in non- Embedded Image Ringer solution of ND96 (black squares, solution 1), ND120 (blue diamonds, solution 2), 0 Na+ ND120 (green diamonds, solution 6), and in presence of 200 μM DIDS (red circle, solution 2 plus DIDS). In the absence of Embedded Image , the fourI-Vresponses of the rkNBC-expressing oocyte to these test solutions (solutions 1,2 ± DIDS, and6) are also superimposable. No current was recorded by either increasing the Na+ concentration or by complete Na+ removal. In addition, there is not a DIDS-inhibitable current present in absence of Embedded Image .D: Embedded Image -dependent steady-state current sweeps. Unsubtracted current sweep data recorded from a rkNBC-expressing oocyte are shown in presence of extracellular 1.5% CO2/10 mM Embedded Image with 96 mM Na+ (black squares, solid line) or Na+ replaced with choline (green diamonds, solid line). The same oocyte was then exposed to 5% CO2/33 mM Embedded Image with 96 mM Na+ (black squares, dotted line) or Na+ replaced with choline (green diamonds, dotted line). E: [ Embedded Image ]o- and [Na+]o-dependentI-Vrelationships. A rkNBC-expressing oocyte was voltage clamped and equilibrated in 1.5% CO2/10 mM Embedded Image /96 mM Na+ for 10 min (see Fig. 1). AnI-Vrelation was then recorded in presence of 96 mM Na+ (black squares, solid line) and 0 Na+ (choline; green diamonds, solid line). This protocol was then repeated using 5% CO2/33 mM Embedded Image .I-Vrelations were again recorded with the oocyte exposed to 96 mM (black square, dotted line) and 0 mM (green diamonds, dotted line) extracellular Na+.F: transport convention. From theI-Vrelationships of rkNBC-expressing oocytes, both the direction and magnitude of cotransport were measured. Positive current indicates a net negative charge movement into the cell. This is the unbalanced inward movement of Embedded Image /Na+via rkNBC. Negative currents indicate the outward movement of Na+/ Embedded Image as net negative charge exits the cell.

DIDS inhibition. The stilbene sulfonate DIDS (Sigma, St. Louis, MO), a known inhibitor of rkNBC (33), was used to determine 1) the non- HCO3 current via rkNBC,2) the voltage dependence of DIDS inhibition, and 3) the [Na+]odependence of DIDS block. Preliminary experiments showed that 100 μM DIDS was sufficient to block rkNBC-mediated cotransport (Fig. 5,bottom). To determine the non- HCO3 current through rkNBC, oocytes were bathed in ND96 or 0 Na+ ND96 (solutions 1 and5, Table 1) andI-Vrelationships were recorded. DIDS, 200 μM, was then added to the bath solution, and the oocyte was perfused for 5 min.I-Vrelations were recorded in ND96 and 0 Na+ ND96 in the presence of DIDS. Voltage sensitivity of DIDS inhibition was tested by first recording anI-Vrelation in the presence of CO2/ HCO3 . DIDS, 200 μM, was then added to the CO2/ HCO3 solution, and the oocyte was incubated for 5 min. AnI-Vrelationship was then recorded in the presence of DIDS. The extracellular Na+ dependence of DIDS inhibition was measured by first recording a Na+ dose response experiment as described above, perfusing the oocyte for 5 min in 200 μM DIDS/CO2/ HCO3 , and repeating the Na+ dose solution protocol in the continued presence of 200 μM DIDS.


Voltage Dependence of rkNBC

In the presence of extracellular CO2/ HCO3 , the electrogenic Na+- HCO3 cotransporter expressed in Xenopusoocytes moves a net negative charge into the cell, and this movement is blocked by extracellular DIDS (13, 33, 34). Using the two-electrode voltage clamp, we studied the voltage dependence of rkNBC expressed inXenopus oocytes. First we identified the contribution of rkNBC expression on the basal currents of oocytes (Fig. 2 C). TheI-Vrelationships in ND96 (non- HCO3 ,solution 1, Table 1) are similar in the presence (black) and absence (green) of extracellular Na+. Addition of 200 μM DIDS (red) to the bath solution does not change this basal current. Therefore the oocyte current in the absence of HCO3 is not due to charge movement via rkNBC. Moreover, Fig. 2 C illustrates that control oocytes show no endogenous electrogenic Na+- HCO3 cotransporter, i.e., no Na+-dependent, DIDS-inhibitable current with or without HCO3 . Consequently, the ND96 (non- HCO3 )I-Vresponse can be subtracted from the HCO3 elicitedI-Vrelations to yield the rkNBC-dependent current.

Figure 2, A andB, illustrates that addition of extracellular CO2/ HCO3 stimulates a DIDS-inhibitable current in rkNBC oocytes not observed in water-injected oocytes (Fig. 2 C). In the presence of 96 mM extracellular Na+, theI-Vresponse of rkNBC oocytes is almost linear, i.e., the direction and magnitude of current (Na+/ HCO3 transport) depends onV m. Positive currents represent the net inward movement of Na+/ HCO3 (Fig. 2 F), whereas negative currents represent the outward movement of Na+/ HCO3 . At the reversal potentials,E rev, there is no net current (transport).E rev values for 10 and 33 mM extracellular HCO3 and [Na+]ofrom 10 to 96 mM are listed in Table 3. These data show that decreasing extracellular Na+ shiftsE rev more positive. The positive shift ofE rev does not mirror the calculated shift in the Nernst potential for Na+ (Table 3), indicating that transport direction is not dependent on the Na+ gradient alone. Furthermore, for our conditions (Fig. 1), pHi and thus [ HCO3 ]ido not change (Fig.3 B), indicating that for the brief solution changes, the HCO3 gradient is approximately static.

View this table:
Table 3.

Erev and stoichiometry

Fig. 3.

Ion-selective microelectrode experiment.A: water, intracellular pH (pHi) experiment. An experiment monitoring pHi and membrane potential (V m) of a control oocyte is shown. Addition of 1.5% CO2/10 mM Embedded Image to bath elicited no change inV m and resulted in cell acidification of 0.195 ± 0.016. This is due to the diffusion of CO2 across the membrane, formation of H2CO3, and subsequent generation of H+and Embedded Image from the fast dissociation.B: rkNBC, pHi experiment. pHi andV m of a rkNBC-expressing oocyte is shown. Initial pH is ∼0.2 pH units higher than the water control. Addition of 1.5% CO2/10 mM Embedded Image resulted in a large, transient hyperpolarization and fall of pHi(0.145 ± 0.010 pH units).V m and pHi reached a steady state in 8–10 min; pHi was 7.28 ± 0.02 (n = 9, 4 frogs) and intracellular [ Embedded Image ] was 4.8 ± 0.2 mM (calculated from the Henderson-Hasselbach equation). After 10-min perfusion in CO2/ Embedded Image , bath Na+ is removed (replaced with choline at bars) for 30-s intervals. The oocyte quickly depolarizes and then repolarizes upon readdition of Na+ without change in pHi. This experiment mimics the Na+ dose response protocol in Fig.1, illustrating that brief changes in extracellular Na+ do not significantly alter steady-state pHi, and therefore Embedded Image , of the oocyte.C: water,aNaiexperiment. Intracellular Na+activity (aNai) and V m were monitored in control oocyte. CO2/ Embedded Image addition resulted in no changes in eitherV m oraNai.D: rkNBC,aNaiexperiment. As seen in B, 1.5% CO2/10 mM Embedded Image addition results in a hyperpolarization of an rkNBC oocyte.aNaiincreases and is ∼9 mM at steady state. Extracellular Na+ removal decreasesaNaiand depolarizes the cell. Chloride removal does not effectaNai.

When Na+ is removed from the bath solution, only negative current is observed (Fig. 2,B andE). These data indicate that NBC-mediated influx of negative charge does not occur without extracellular Na+. Additionally, without extracellular Na+, the negative currents are augmented, indicating increased outward Na+- HCO3 cotransport. Taken together, these data indicate that neither Na+ nor HCO3 alone is capable of inducing rkNBC currents. Moreover, all of these currents are inhibited by 200 μM bath DIDS (Fig. 2, A andB), independent of extracellular Na+ andV m.

Intracellular Ion Effects of rkNBC Expression

CO2/ HCO3 addition to the bath solution results in the acidification of oocytes due to CO2 diffusion across the plasma membrane, intracellular hydration, and subsequent hydrolysis to form H+ and HCO3 . pHi andaNaichanges are quantified in Table 2 for rkNBC and water-injected oocytes perfused with 1.5% CO2/10 mM HCO3 (pH 7.5) or 5% CO2/33 mM HCO3 (pH 7.5). Water-injected oocytes exposed to 1.5% CO2/10 mM HCO3 /pH 7.5, acidify by 0.195 ± 0.016 (n = 5, 3 frogs, means ± SE) (Fig. 3 A). In contrast, the initial pHi of rkNBC oocytes is ∼0.2 pH units more alkaline than water-injected controls (Fig. 3,A vs.B, and Table 2), and the addition of 1.5% CO2/10 mM HCO3 elicits an immediate hyperpolarization of −72 ± 1.8 mV (n = 9, 4 frogs) concomitant with the fall in pHi (0.145 ± 0.010 pH units, Fig. 3 B). The apparent buffering power (βT) (35) of rkNBC-expressing oocytes at 1.5% CO2/33 mM HCO3 is 35.7 ± 3.1 mM/pH unit (n = 9), whereas that of water-injected controls is 16.7 ± 1.8 mM/pH unit (n = 5) (Table 2). The increased β of rkNBC-expressing oocytes (βrkNBC) indicates that an additional buffering system (i.e., acid-base transporter) is present. The increased movement of HCO3 across the plasma membrane increases [ HCO3 ]iand thus buffers more of the cytosolic H+ generated by the hydration of CO2 entering the oocyte.

CO2/ HCO3 addition to rkNBC-oocytes also results in anaNaiincrease in concert with the hyperpolarization (Fig.3 D) but not in water-injected controls (Fig. 3 C). Nominal air CO2 over several days (days 0–3) provides sufficient solution HCO3 to raise initialaNaiby ∼2 mM in rkNBC oocytes over water controls. rkNBC oocytes exhibit a pHi recovery ( HCO3 influx) over 5–10 min in the continued presence of CO2/ HCO3 . Removal of extracellular Na+causes an immediate and reversible depolarization (Na+/ HCO3 efflux), followed by a delayed fall in pHi (33) andaNai(Fig. 3 D). Both pHi andaNairesponses of rkNBC-expressing oocytes are blocked by 200 μM DIDS (not shown).

Figure 3 B also illustrates that pHi can remain at steady state upon short and repeated removals and replacement of extracellular Na+ (every 30 s). Thus we can measure changes in whole cell current on the second time scale without significantly disturbing the steady-state [ HCO3 ]ioraNai.

Cation Dependence of rkNBC

Li+ appears to substitute for Na+ for transport via the Na+ HCO3 cotransporter when assayed in basolateral membrane vesicles from rabbit kidney (40). Studies using human kidney NBC (hkNBC) expressed in HEK-293 cells also indicate that Li+ induces a 25% DIDS-sensitive HCO3 pHi recovery from an acid load (4). We used two-electrode voltage clamp to determine whether cations other than Na+(K+, Li+, and choline) could be transported in Xenopus oocytes expressing rkNBC. No inward transport (positive current) was observed when Na+ is replaced by choline in the bath solution. To ensure that any current measured was due to the test cation, we pulsed solutions from non- HCO3 to HCO3 solution with constant 96 mM cation present (Fig.4 A). The HCO3 -stimulated current is the difference between theI-Vrelationships in these two solutions. As illustrated in Fig. 4,B andC, K+ and choline do not facilitate HCO3 transport via rkNBC. Li+ is only capable of 3 ± 2% of the Na+ response (n = 5, 2 frogs) over a voltage range from −160 to +60 mV.

Fig. 4.

Cation dependence. A: solution pulse protocol. This solution protocol is used to test whether K+, Li+, or choline are capable of stimulating a Embedded Image -dependent current in rkNBC oocytes. Oocytes were voltage clamped and bathed in ND96 (solution 1, Table 1) for 5 min before switching to test cation/non- Embedded Image Ringer (solutions 3 to5, Table 1) for 5 min. Solution was then switched to the corresponding 1.5% CO2/10 mM Embedded Image solution for 2 min (solutions 10 to12, Table 1, i.e., Li+-ND96 to 1.5% CO2/10 mM Embedded Image /96 mM Li+) and returned to non- Embedded Image Ringer for 2 min. AnI-Vrelation was recorded before and after each solution change. Embedded Image -stimulated current for each cation was taken as the difference between the non- Embedded Image and Embedded Image I-Vresponses. Cation solutions were tested in random order.B: current sweeps of cation replacement. The unsubtracted current sweep data from a rkNBC-expressing oocyte exposed to extracellular K+- Embedded Image (red), Li+- Embedded Image (blue), Na+- Embedded Image (black), and choline- Embedded Image (green) using the protocol described in A.C: rkNBCI-Vresponse of cations ± Embedded Image . SubtractedI-Vresponse curves from the current sweeps inB show that only Na+ (black) stimulates a strong Embedded Image -dependent current. Extracellular K+ (red) and choline (green) haveI-Vrelations that lie on the voltage axis, indicating no transport. Li+ (blue) shows only a slight current response of a maximal 3 ± 2% of the Na+ response over the voltage range tested.

Extracellular Na+Dependence of rkNBC

Determination of experimental vmax.

We next determined the maximal Na+response (v max) of rkNBC-expressing oocytes in constant to CO2/ HCO3 . Recording theI-Vresponses after varying [Na+]obetween 84 and 120 mM (osmolalities matched with choline), we found no difference between the 96 and 120 mM Na+ I-Vresponse (Fig. 5,top). Thus we conclude that the [Na+]oat which v maxcurrent is observed is ∼96 mM. Even at supermaximal Na+ concentrations, DIDS inhibits all Na+/ HCO3 current responses (Fig. 5, top).

Fig. 5.

Experimental determination of apparent rkNBC v max(maximal Na+ and Embedded Image stimulated current).Top:v max ± DIDS. To avoid changes in solution osmolality that could activate endogenous acid-base transporters or Cl channels, experiments depicted at top used derivatives ofsolutions 2,9, and13 (Table 1). Shown are eightI-Vrelationships. TheI-Vprofile of a rkNBC-expressing oocyte does not change between ND96 (yellow diamonds, solution 1) and ND120 (yellow squares, solution 2). Subsequently, all solutions in topwere based on the 120 mM cation formulation, i.e., made by mixingsolutions 9 and13 in appropriate ratios. Thev max for extracellular Na+ was determined using the solution protocol described in Fig.4 A. CO2/ Embedded Image I-Vrelations for 96 mM (black squares) and 120 mM (blue triangles) extracellular Na+ overlap (<4% divergence at +60 mV, n = 5, 2 frogs). Therefore, v maxoccurs at ∼96 mM Na+. In absence of extracellular Na+, no outward current was observed (120 mM choline, green diamonds). DIDS, 200 μM, inhibited rkNBC current for all Na+ concentrations (120 mM Na+, red circles; 96 mM Na+ red squares; and 0 Na+, red diamonds).Middle: rkNBCI-Vresponse, extracellular Na+dependence. Experimental protocol outlined in Fig. 1 was used to determine the extracellular Na+dependence of a rkNBC-expressing oocyte.I-Vresponses for 96 mM (black squares), 72 mM (blue diamonds), 24 mM (blue open circles), 10 mM (blue open diamonds), 1 mM (inverted blue open triangles), and 0 mM (green diamonds) extracellular Na+ are shown (all solutions are 96 mM cation based). Bottom: rkNBC extracellular Na+ dependence ± DIDS. Na+ concentration dependenceI-Vrelationships of a rkNBC-expressing oocyte were measured at 96 mM (black squares), 48 mM (blue triangles), 24 mM (blue open circles), and 0 mM (green diamonds) extracellular Na+. Oocyte was then incubated for 5 min in 1.5% CO2/10 mM Embedded Image with 100 μM DIDS, and experiment was repeated (extracellular Na+: 96 mM, red squares; 48 mM, red triangles; 24 mM, red circles; and 0 mM, red diamonds). DIDS, 100 μM, inhibits inward and outward Na+/ Embedded Image transport at all Na+concentrations and voltages.

Extracellular Na+concentration dependence.

We developed the solution protocol in Fig.1 A to study the extracellular Na+ dose response profile of rkNBC. As described above, a rkNBC oocyte reaches a pHi and voltage steady state within 8–10 min perfusion with CO2/ HCO3 at saturating [Na+]o. Figure 1 B illustrates that the resulting current, atV h = −60 mV, also plateaus within the 10-min initial perfusion. Figure 5,middle, shows a set ofI-Vrelations measured for a dose response experiment. The magnitude of outward current (NaHCO3 influx) is decreased, and inward current (NaHCO3 efflux) is augmented with decreasing extracellular Na+. As noted above, E revshifts in the positive direction, but is not equated with the Na+ reversal potential (Table 3). We also found that 100 μM DIDS inhibits rkNBC current at all Na+ concentrations tested, 0 to 120 mM (Fig. 5, top andbottom), independent of Na+ and voltage.

Calculation of apparent K0.5.

Figures 5, middle, and Fig.6 Aillustrate that rkNBC transport direction is voltage and extracellular Na+ dependent. From these experiments, we calculated the apparent affinity coefficient of rkNBC for extracellular Na+(K 0.5) of Na+- HCO3 cotransport at each voltage measured. To plot the rkNBC-specific currents, we subtracted the current at a given [Na+]oand V m from the current elicited at that sameV m by the 0 Na+/CO2/ HCO3 solution, i.e., II 0 Na+(Fig. 6 B). This subtraction yields the full range of rkNBC activity (maximal outwardI to maximal inwardI). We fit these data at each pulseV m with a right rectangular hyperbolic function2(Michaelis-Menten) and calculated the apparentK 0.5 as ∼30 mM (Fig. 6). Remarkably, these data indicate that at every testV m (−160 to +60 mV) that the apparentK 0.5 for extracellular Na+ is ∼30 mM. Since all curves were similar, after normalizing currents, we grouped all voltages for each test [Na+] to generate a composite graph (Fig. 6 C).

Fig. 6.

Apparent extracellular Na+affinity for rkNBC. A: rkNBCI-Vresponse.I-Vresponses elicited using the solution protocol outline in Fig. 1 are shown for 1.5% CO2/10 mM Embedded Image . As previously, current responses are subtracted from the non- Embedded Image current. Subtracted currents at 96 mM (•, solid line), 84 mM (•, broken line), 72 mM (♦, broken line), 48 mM (▴, broken line), 36 mM (●, broken line), and 0 mM (♦, solid line) extracellular Na+ are shown.B: Na+-stimulated current measurement. Na+-stimulated current at each voltage tested (−160 to 60 mV in 20-mV steps) inA was determined by subtracting the 0 Na+/CO2/ Embedded Image current (I 0 Na+) from the test Na+/CO2/ Embedded Image current (I). The “II 0 Na+” current is plotted vs. extracellular [Na+] for each voltage. These points were fit with a two-parameter right rectangular hyperbolic function of generalized formv =a(b[Na+]o)/(1 +b[Na+]o) (Michaelis-Menten). All lines fit with anR 2 value >0.99.C: apparentK 0.5 of rkNBC for extracellular Na+. Na+-stimulated currents (ΔI =II 0 Na+) were determined as illustrated in B, normalized to the 96 mM Na+response (experimentalv max, i.e.,I max) and plotted vs. [Na+]oand fit as described above. ApparentK 0.5 was calculated from the resulting equation as the [Na+]othat elicits the half-maximal current (0.5). ApparentK 0.5 values were similar for the 12 test voltages (not shown). Thus we grouped the voltage data together and calculated the overall apparentK 0.5 of rkNBC for extracellular Na+ as 30 mM (solid points, n = 180, 6 frogs). Open circles represent data from 5% CO2/33 mM Embedded Image .

HCO3 :Na+stoichiometry of rkNBC.

Cotransport of Na+ and HCO3 through rkNBC can be described as a coupled transport process by the chemical equation[Na+]i+b·[HCO3]i[Na+]o+b·[HCO3]o Equation 1The stoichiometry of transport (b) can be determined from the concentration gradients across the cell membrane andE rev, i.e., whenΔμNai­o=bΔμHCO3o­i Equation 2where Δ μNai­o is the in-to-out electrochemical potential difference for Na+, Δ μHCO3o­i is the out-to-in electrochemical potential difference for HCO3 , andb is the cotransport ratio of HCO3 :Na+. By substituting measured values, the expression becomes[Na+]i/[Na+]o=([HCO3]o/[HCO3]i)b Equation 3 ×exp[(b1)FVm/RT] whereR, T, and F have their usual meaning (gas constant, absolute temperature, and Faraday constant, respectively). TheV m at which there is no net ionic flux across the cell membrane is the reversal potential, orE rev. The values of E rev, [ HCO3 ]i, andaNaiwere measured or determined from the previously described experiments (Tables 2 and 3). Solving for b inEq. 4, we calculate the transport stoichiometry of rkNBC for our conditions to be 2 HCO3 :1 Na+ at 10 and 33 mM [HCO3 ]o(Fig. 7). These results are obtained between 10 and 96 mM extracellular Na+. Thus rkNBC stoichiometry does not appear sensitive to the cellular Na+ gradient.

Fig. 7.

Stoichiometry of rkNBC. Embedded Image :Na+stoichiometry was calculated from experimentally determined reversal potentials (Table 3) and the measured values of intracellular Embedded Image andaNaias outlined in the text (Equation3 ). Stoichiometry for [Na+]oranging from 10 to 96 mM is 2 Embedded Image :1 Na+. Solid bars represent data from 1.5% CO2/10 mM Embedded Image , and open bars represent data from 5% CO2/33 mM Embedded Image .


The process of electrogenic Na+- HCO3 cotransport is functionally described in a variety of tissues, e.g., kidney, brain, heart, eye, and pancreas. Cloning of these NBC isoforms reveals that they are 95% homologous at the amino acid level (30). However, function in these tissues ranges from strict HCO3 excretion (kidney), to mainly HCO3 influx (pancreas),3and both modes of transport (eye and heart). To begin to define the determinants of NBC function, we studied the voltage- and cation-dependent properties of Na+- HCO3 cotransport expressed in Xenopusoocytes. Our studies indicate that rkNBC is specific for Na+ over K+ and choline. These results are consistent with studies of the human kidney isoform, indicating that K+ is not a substrate for NBC (4). However, our Li+ results differ, i.e., Li+ minimally stimulates rkNBC-mediated HCO3 transport. Yet the hkNBC and rkNBC isoforms are 97% homologous. Of the 22 amino acid difference between the two clones, 10 alter charge. H502 in rat is switched to N502 in human and has the predicted location at the putative extracellular portion of transmembrane span 3 (TM-3, L482 to F498). Such charge switches may sufficiently alter a Na+ selectivity filter of NBC in the human clone to allow Li+ but not K+ to be transported. This amino acid is conserved between the human kidney (9, 13, 29), human heart (13), and human pancreas (2) NBC isoforms. The ability of human NBC to cotransport Li+ could have therapeutic consequences resulting from Li+ administration. Such Li+ cotransport capacity might not be predicted to alter NBC renal function, but might affect both the brain and the heart where NBC-mediated cotransport is bidirectional.

Since extracellular Na+ is high and cytosolic Na+ is low, the normal, physiological, electrochemical gradient for Na+ is inward. Under this condition, we find theI-Vresponse of rkNBC-expressing oocytes is roughly linear. Extracellular Na+ of 96 mM is the experimentally determined saturation point of rkNBC current with bath CO2/ HCO3 , or the functionalv max for our conditions. At this apparentv max, theE rev is −94 mV and −100 mV for 10 and 33 mM extracellular HCO3 , respectively. Only at voltages more negative thanE rev is outward cotransport observed. At −73 mV (49), the approximate basolateral membrane potential of mammalian, renal proximal tubule cells, rkNBC-mediated cotransport in oocytes is inward (raising both pHi andaNaicompared with controls), not outward as expected for electrogenic Na+- HCO3 cotransporter function of the proximal tubule basolateral membrane. The observed rkNBC function in oocytes is more comparable to that described in tissues where NBC functions as an acid extruder, such as the brain and pancreas. Our results suggest that a mechanism of directional control for the cotransporter is present in the kidney and not found in these other tissues. Additionally, our data predict that variable modes of cotransport would occur with changes in cell membrane potential, physiological or pathological. For example, Camilion de Hurtado and coworkers (10, 11) reported that cat papillary muscle could recover from an acid load caused by increased heart rate. These authors argued that depolarization during cardiac contraction increased the activity of the electrogenic Na+- HCO3 cotransporter causing HCO3 influx. This finding is consistent with our data which show a depolarization could move the membrane potential throughE rev of rkNBC reversing the cotransport mode from outward to inward and thus increase pHi.

Our data indicate that addition of the stilbene DIDS to the bath solution inhibits both the inward and outward Na+- HCO3 cotransport modes of rkNBC. Blockade is not overcome by increasing extracellular Na+ or by alteringV m, and with time the blockade is irreversible (not shown). In the murine band 3 protein (AE1), H2-DIDS covalently binds to Lys539 and Lys851 (28). The NBCs and the anion exchangers (AE1–3) are the initial members of a HCO3 superfamily of transporters (30,33) for which a predicted DIDS binding motif [K-(Y)-(X)-K for Y = M, L, and X = I, V, Y] seems to exist (30). rkNBC has two putative motifs: 1) KMIK at 558–561 (near H502N described above) and2) KLKK at 768–771. Perhaps these sites are close to a substrate binding site or permeant path. Upon DIDS binding, this path may be occluded or the regional structure altered, leading to complete abolition of transport.

Using BSC1 cells to examine the native electrogenic Na+- HCO3 cotransporter, Jentsch and associates (22) found that for 10 mM HCO3 , the apparentK 0.5 for extracellular Na+ was between 35 and 45 mM. To study the native electrogenic Na+- HCO3 cotransporter, presumably NBC, Gross and Hopfer (17) used a rat proximal tubular cell line (SKPT-0193). Employing a “zero-trans” condition (15 mM intracellular HCO3 with no basolateral Na+ present), these investigators calculated the apparentK 0.5 for intracellular Na+ binding as ∼18 mM. Nevertheless, it is unclear whether intra- and extracellular Na+ affinities of NBC are similar. For rkNBC expressed inXenopus oocytes, we calculate the apparent K 0.5 for extracellular Na+ to be ∼30 mM at a “normal”V m (−60 mV). The Na+ sensitivity of rkNBC cotransport appears unaffected by membrane potential. That is, over the entire voltage, range the apparentK 0.5 for extracellular Na+ is ∼30 mM (Fig. 6). However, although the apparentK 0.5 is stable over the entire voltage range tested (−160 to +60 mV), the current magnitude increases asV m becomes more positive. Thus either more rkNBC protein is inserted into the plasma membrane or the rate of cotransport is increasing. Since insertion and retrieval of protein is unlikely to occur in the millisecond-to-second time scale, the cotransport-mediated current increase likely indicates that the overall transport rate is voltage sensitive, yet extracellular Na+ binding is not. Moreover, with physiological Na+ concentration greater than 30 mM, it is unlikely that extracellular Na+ binding is a rate-limiting step of cotransport.

This study was designed to elucidate some of the fundamental properties of rkNBC in an isolated system. A HCO3 :Na+stoichiometry of 3:1 is thought to be necessary for HCO3 efflux from renal proximal tubule cells. This stoichiometry is necessary to overcome the Na+ gradient and use the favorable movement of charge down the voltage gradient. However, here we have determined the stoichiometry of rkNBC expressed inXenopus oocytes as 2:1. This stoichiometry is [Na+]oindependent (Fig. 7), as evidenced by a stoichiometry of 2:1 between 10 and 96 mM extracellular Na+. For normal physiological conditions (high extracellular Na+) andV m more positive than −80 mV, we find only inward Na+- HCO3 cotransport (outward current). This same stoichiometry is found at both 10 and 33 mM HCO3 (Fig. 7). Taken together, these results indicate that rkNBC alone can mediate HCO3 influx as described for some tissues (e.g., brain, heart, liver, and pancreas). However, these data alone cannot explain the HCO3 reabsorption by the renal proximal tubule cell. A recent report examining rkNBC function in giant patches has also found a 2 HCO3 :1 Na+ stoichiometry (18). Additionally, several groups have measured or calculated a stoichiometry of 3 HCO3 :1 Na+ for native membranes of basolateral membrane vesicles (39), intact proximal tubules (48), or proximal tubular cell monolayers (16). Although formally one could hypothesize another NBC isoform in the kidney, Schmitt and associates (37) in a recent immunolocalization study have found that two different antibodies, specific for NBC, recognize a major protein at the basolateral membrane of mammalian proximal tubules. The same study explicitly demonstrated that these antibodies recognize the functional and recombinant rkNBC expressed inXenopus oocytes. Moreover, in the several years since NBC was cloned by expression (31, 34), no other electrogenic Na+- HCO3 cotransporter has been reported for the kidney nor have similar renal clones appeared in the Expressed Sequence Tag (EST) databases. Consequently, although formally possible, it seems unlikely that another (novel and major protein) electrogenic Na+- HCO3 cotransporter mediating HCO3 reabsorption in the mammalian proximal tubule, will be found. It is possible that a factor(s) is absent in theXenopus oocyte but present endogenously in HEK-293 cells or native proximal tubule membranes. Such a factor might be regulated such that it, in turn, could shift the activity of NBC from an acid extruder to that of an acid loader, i.e., HCO3 reabsorption. Alternatively, an as yet unknown endogenous protein of theXenopus oocyte could interact with rkNBC to mask the “true” stoichiometry of rkNBC transport. Additional studies using other heterologous expression systems or planar lipid bilayers will be required to entirely rule out this latter possibility.

Our data suggest that in the intact proximal tubule or vesicles there are likely other factors, such as binding partners or cellular factors, which modify the rkNBC function allowing NaHCO3 efflux ( HCO3 reabsorption). Regulation of rkNBC activity or coupling by protein kinase A (PKA) and/or PKC might be important, since there is a predicted PKA phosphorylation site and seven PKC consensus phosphorylation sites (30, 33). Such activation could shift the cotransporter voltage dependence such that Na+/ HCO3 efflux occurs. Currently, we have no information indicating that other proteins associate with rkNBC. However, accessory proteins, such as those recently illustrated with the Na/H exchangers (25, 46, 47), could modify rkNBC function. Moreover, AE1 has long been known to form dimers (7, 20, 43-45), associate with cytoskeletal proteins (12), and has recently been shown to associate with carbonic anhydrase II (42). Although it is unclear whether such modifications or protein associations alter the transport function of either the AEs or the NBCs, it is attractive speculation. By extension, tissue-specific expression of a modulator could allow the same NBC protein to be used for both HCO3 influx and efflux in different locations. The specific regulator(s) of rkNBC function that allow HCO3 efflux in the renal proximal tubule is (are) yet unknown. Further studies that explore the isolated protein properties will help to determine whether factors or binding partners can alter NBC function.


We thank Drs. Eberhard Frömter, Ulrich Hopfer, and Suzanne Müller-Berger for helpful suggestions on the manuscript. We thank Dr. Stephen Jones for suggesting how to present the Na+ current response data.


  • Address for reprint requests and other correspondence: M. F. Romero, Dept. of Physiology & Biophysics, Case Western Reserve Univ. School of Medicine, 2119 Abington Rd., Cleveland, OH 44106-4970 (E-mail:mfr2{at}

  • Portions of this work have been presented in preliminary form (38).

  • This work was supported by a grant from the American Heart Association (to M. F. Romero) and a HHMI-institutional grant (to Case Western Reserve University). C. M. Sciortino was supported by National Institute of Diabetes and Digestive and Kidney Diseases Predoctoral Fellowship DK-07678.

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

  • 1 Selectivity for the ion of interest (i) over the interfering ion (j) is calculated asK ij =a i/[a jexp(z i/z j)] × 10exp[(E jE i)/m], where a is activity anm is the slope of the electrode. Na+ electrodes maintain their calibrated slope even at 5 mM NaCl.

  • 2 Note that similar results are obtained if the data are fit by a second order polynomial.

  • 3 A preliminary report from our laboratory (41) indicates that models of pancreatic HCO3 absorption and secretion may not include all of the relevant transporters. In rat pancreatic ductal epithelia, NBC protein is mainly basolateral but is also present apically. Moreover, the basolateral membranes of rat pancreatic acinar cells also contain NBC protein.


View Abstract