The Renal Electrogenic Na⁺:HCO₃⁻ Cotransporter

One of the major functions of the renal proximal tubule is to reclaim HCO₃⁻ that has been filtered in the glomerulus. This reabsorption of HCO₃⁻ from the proximal-tubule lumen to the blood helps to maintain an appropriately high [HCO₃⁻] in the blood and thus helps to stabilize blood pH. Failure to reabsorb sufficient HCO₃⁻ would lead not only to the loss of HCO₃⁻ in the urine but also to the loss of Na⁺ (which normally accompanies reabsorbed HCO₃⁻) and osmotically obligated water. Thus, the consequences of reduced Na⁺ reabsorption would be metabolic acidosis and volume depletion.

As shown in Fig. 1, HCO₃⁻ reabsorption by the proximal tubule is a multistep process. First, H⁺ is secreted into the lumen by Na⁺/H⁺ exchangers (antiporters) and presumably vacuolar-type H⁺ pumps (V-type ATPases). Second, this H⁺ titrates luminal HCO₃⁻ to CO₂ and H₂O, a process accelerated by carbonic anhydrase IV (CA IV), which is tethered to the extracellular surface of the apical membrane. Third, the newly formed CO₂ and H₂O enter the proximal-tubule cell. Fourth, cytoplasmic CA II accelerates the regeneration of H⁺ and HCO₃⁻. Finally, this HCO₃⁻ exits the proximal-tubule cell across the basolateral membrane, completing the movement of HCO₃⁻ from lumen to blood.

In the portion of the proximal tubule furthest downstream from the glomerulus, the S3 segment, the efflux of HCO₃⁻ across the basolateral membrane is mediated by both a Cl⁻/HCO₃⁻ exchanger and an electrogenic Na⁺:HCO₃⁻ cotransporter (Nakhoul et al. 1990). However, in more proximal segments of the proximal tubule (S2 and S1 segments), where the bulk of HCO3⁻ is in fact reabsorbed, the relative contribution of Cl⁻/HCO₃⁻ exchange decreases and that of electrogenic Na⁺:HCO₃⁻ cotransport increases (Kondo and Frömter, 1987). All told, the proximal tubule reabsorbs some 90 % of the filtered HCO3⁻ and 70 % of the filtered Na⁺. The electrogenic Na⁺:HCO₃⁻ cotransporter carries the vast majority of this HCO₃⁻ and approximately 25 % of the Na⁺.

First described by Boron and Boulpaep (1983) in the proximal tubule of the salamander Ambystoma tigrinum, electrogenic Na⁺:HCO₃⁻ cotransporters have subsequently been studied in several epithelial cell lines (Jentsch et al. 1984, 1985), intact rat and rabbit proximal tubules (Alpern, 1985; Sasaki et al. 1987), as well as in a number of other cell types (Deitmer and Schlue, 1989; Gleeson et al. 1989; Fitz et al. 1989; Weintraub and Machen, 1989; Rajendran et al. 1991; M. O. Bevensee and W. F. Boron, in preparation; M. O. Bevensee, M. Apkon and W. F. Boron, in preparation). The four hallmarks of the renal Na⁺:HCO₃⁻ cotransporter are its electrogenicity, Na⁺-dependence, HCO₃⁻-dependence and blockade by stilbene derivatives such as DIDS. Probably the most reliable assay for detecting the electrogenic Na⁺:HCO₃⁻ cotransporter in intact cells is to remove extracellular Na⁺. This maneuver causes an abrupt positive shift (i.e. depolarization) in the cell membrane voltage (V_m). In the absence of HCO₃⁻, in the presence of DIDS or in cells lacking the cotransporter, removing Na⁺ generally causes a small negative shift in V_m, reflecting the normal Na⁺ conductance of the cell. In the presence of a functional electrogenic Na⁺:HCO₃⁻ cotransporter, Na⁺ removal also causes a slower fall in intracellular pH (pHi) as well as a DIDS-sensitive fall in [Na⁺]_i.

The electrogenicity of the cotransporter implies that each Na⁺ moves with two or more HCO₃⁻ or HCO₃⁻ equivalents (e.g. CO₃²⁻). The Na⁺:HCO₃⁻ stoichiometry has not been directly measured. However, it has been deduced by Soleimani et al. (1987) on the basis of thermodynamic data obtained from membrane vesicles prepared from the basolateral membranes of rabbit proximal tubules. The principle underlying these experiments is that the net direction of cotransport depends on the stoichiometry as well as on the Na⁺ and HCO₃⁻ electrochemical gradients. Varying these gradients while measuring the net direction of transport, the authors concluded that the renal electrogenic Na⁺:HCO₃⁻ cotransporter has a Na⁺:HCO₃⁻ stoichiometry of 1:3. Further work (Soleimani and Aronson, 1989) was consistent with the hypothesis that the renal electrogenic Na⁺:HCO₃⁻ cotransporter has binding sites for 1 Na⁺, 1 HCO₃⁻ and 1 CO₃²⁻, and that sulphite (SO₃²⁻) can replace CO₃²⁻ at the CO₃²⁻ site (Fig. 2).

The HCO₃⁻ transporters shown in Fig. 3 might be thought of as a functional family of transporters that may or may not be related molecular genetically. The renal electrogenic Na⁺:HCO₃⁻ cotransporter, with a Na⁺:HCO₃⁻ stoichiometry of 1:3, is one of as many as four Na⁺-coupled HCO₃⁻ transporters, as well as a Cl⁻/HCO₃⁻ exchanger and a K⁺:HCO₃^−?cotransporter. There is at least one, and perhaps two, other Na⁺:HCO₃⁻ cotransporter that normally mediates a net influx, rather than a net efflux, of HCO₃⁻ equivalents (see Boron and Boulpaep, 1989). That is, they are acid extruders rather than acid loaders. In invertebrate glial cells (Deitmer and Schlue, 1989), mammalian astrocytes (M. O. Bevensee and W. F. Boron, in preparation; M. O. Bevensee, M. Apkon and W. F. Boron, in preparation) and other cells, an electrogenic Na⁺:HCO₃⁻ cotransporter mediates a net influx of Na⁺ and HCO₃⁻. On the basis of analyses of electrochemical gradients of Na⁺ and HCO₃⁻, it is believed that such acid-extruding electrogenic Na⁺:HCO₃⁻ cotransporters have a 1:2 stoichiometry. Thus, it is the stoichiometry of the cotransporter, as well as the electrochemical gradients of the transported ions, that determines the net direction of transport and whether the cotransporter behaves as an acid loader (as is generally the case with a 1:3 stoichiometry) or as an acid extruder (as is generally the case with a 1:2 stoichiometry). In addition to cotransporters with stoichiometries of 1:3 and 1:2, work by Vaughan-Jones and colleagues (Dart and Vaughan-Jones, 1992) on heart cells suggests that these cells have an electroneutral Na⁺:HCO₃⁻ cotransporter with a Na⁺:HCO₃⁻ stoichiometry of 1:1. Camilion de Hurtado et al. (1995), however, concluded that this cotransporter is an electrogenic 1:2 Na⁺:HCO₃⁻ cotransporter.

The Na⁺-driven Cl⁻/HCO₃⁻ exchanger is the major acid extruder in a wide variety of animal cells. In contrast to the two electrogenic Na⁺:HCO₃⁻ cotransporters, the Na⁺-driven Cl⁻/HCO₃⁻ exchanger is electroneutral and has an absolute requirement for Cl⁻. Like the 1:2 and 1:1 Na⁺:HCO₃⁻ cotransporters, the Na⁺-driven Cl⁻/HCO₃⁻ exchanger normally moves HCO₃⁻ into cells. Aside from the work described here on the cloning of the 1:3 electrogenic Na⁺:HCO₃⁻ cotransporter, there is no published molecular information on any of the Na⁺-coupled HCO₃⁻ transporters.

The Cl⁻/HCO₃⁻ exchanger, sometimes referred to as the Na⁺-independent Cl⁻/HCO₃⁻ exchanger to distinguish it from the Na⁺-driven Cl⁻/HCO₃⁻ exchanger, was the first HCO₃⁻ transporter to be described. There is now considerable information on the molecular biology and biochemistry of Cl⁻/HCO₃⁻ exchangers. The first of these to be cloned was the band 3 protein of red blood cells, also known as AE-1 (for Anion Exchanger) (Kopito and Lodish, 1985). AE-1 plays a major role in CO₂ carriage by erythrocytes. Two related Cl⁻/HCO₃⁻ exchangers, AE-2 and AE-3, are present in a wide variety of non-erythroid cells. Because they normally mediate the exchange of extracellular Cl⁻ for intracellular HCO₃⁻, AE-2 and AE-3 function as intracellular acid loaders.

There are no molecular tools available for uniquely identifying any of the cation-coupled HCO₃⁻ transporters. By forming covalent derivatives with its targets, the stilbene derivative DIDS usually blocks all of the cation-coupled HCO₃⁻ transporters other than the K⁺:HCO₃⁻ cotransporter. DIDS also inhibits a host of other transporters, channels and proteins. It is clear that DIDS, as well as the related compound 4-acetamido-4′-isothiocyano-2,2′-stilbenedisulfonate (SITS), is hardly specific. Therefore, in embarking on our attempt to clone the renal electrogenic Na⁺:HCO₃⁻ cotransporter, we decided on an expression-cloning approach. Others have used Xenopus laevis oocytes to expression-clone cDNAs as widely ranging as those that encode the Na⁺:glucose cotransporter (Hediger et al. 1987), the H⁺:oligopeptide cotransporter (Fei et al. 1994), the epithelial amiloride-sensitive Na⁺ channel (Canessa et al. 1994) and the Ca²⁺ receptor from the parathyroid gland (Brown et al. 1993).

We isolated poly(A)+ RNA from either the cortex of rabbit kidneys or whole kidneys of tiger salamanders, and microinjected this RNA into stage 5/6 oocytes from the frog Xenopus laevis. After waiting several days, we assayed the oocytes for expression of an electrogenic Na⁺:HCO₃⁻ cotransporter, using the approach schematized in Fig. 4. An oocyte, which is approximately 1.2 mm in diameter, is impaled with both a conventional voltage electrode filled with 3 mol l⁻¹ KCl and a pH-sensitive microelectrode filled with a liquid pH sensor.

Perhaps the two most important aspects of an expression-cloning project are to identify a good source of the mRNA that encodes the protein of interest (see below) and to develop a good assay. Our assay was to monitor the V_m and pHi changes caused by removing extracellular Na⁺ in the presence of CO₂/HCO₃⁻. In the presence of an active electrogenic Na⁺:HCO₃⁻ cotransporter, this maneuver is expected to cause Na⁺, HCO₃⁻ and net negative charge to leave the cell via the electrogenic Na⁺:HCO₃⁻ cotransporter and thus to cause a rapid depolarization (i.e. positive shift in V_m) and a slower pHi decrease. To make the assay more specific, we required that these V_m and pHi changes be blocked by DIDS and be dependent upon the presence of CO₂/HCO₃⁻. By itself, the depolarization induced by Na⁺ removal is nearly diagnostic of electrogenic Na⁺:HCO₃⁻ cotransport. Bear in mind that a conventional Na⁺ conductance would produce a hyperpolarization, not the depolarization characteristic of the electrogenic Na⁺:HCO₃⁻ cotransporter. A decrease in electrogenic Na⁺ pumping will also produce a depolarization, albeit a slow one that should be prevented by pretreating the cell with ouabain (Boron and Boulpaep, 1983). Indeed, we conducted all of our screening experiments in the presence of ouabain to eliminate the potential complication coming from the Na⁺/K⁺ pump. A Na⁺-dependent K⁺ conductance (Kameyama et al. 1984) would lead to a depolarization in response to Na⁺ removal, and this effect may have complicated the identification of a 1:2 electrogenic Na⁺:HCO₃⁻ cotransporter in mammalian astrocytes (M. O. Bevensee, M. Apkon and W. F. Boron, unpublished results). However, by requiring that the depolarization induced by Na⁺ removal be blocked both by DIDS and by removing CO₂/HCO₃⁻, we should have eliminated potential confusion from Na⁺-dependent K⁺ conductances or any other known mechanisms. By further requiring that the V_m changes be accompanied by HCO₃⁻-dependent and DIDS-sensitive pHi changes, we should have made the assay virtually specific for either a 1:3 or 1:2 electrogenic Na⁺:HCO₃⁻ cotransporter.

We emphasize the specificity of the assay because others have attempted to identify electrogenic Na⁺:HCO₃⁻ cotransporters using much less specific physiological criteria. For example, applying DIDS would be expected to cause a depolarization if an electrogenic Na⁺:HCO₃⁻ cotransporter were operating in the inward direction and a hyperpolarization if the cotransporter were operating in the outward direction. However, DIDS is notoriously non-specific; it can inhibit a wide variety of membrane proteins, including certain Cl⁻ channels, that could influence V_m. Thus, a DIDS-induced V_m change can hardly be regarded as specific for an electrogenic Na⁺:HCO₃⁻ cotransporter. Another approach that has been used to identify electrogenic Na⁺:HCO₃⁻ cotransporters is to monitor the V_m changes that occur when a CO₂/HCO₃⁻ solution is applied. In principle, applying CO₂/HCO₃⁻ should cause an initial hyperpolarization (due to electrogenic Na⁺/HCO₃⁻ influx) that wanes as [HCO₃⁻]_i slowly increases, secondary to the influx of CO₂. In the new steady state, the predicted V_m would be either more negative or more positive than the initial V_m, depending on whether the net direction of transport was inward or outward, respectively. However, the application of CO₂/HCO₃⁻ can cause non-specific V_m changes. In the first place, CO₂ diffuses rapidly across most cell membranes and could cause pHi decreases in the bulk cytoplasm, preceded by perhaps even larger pH changes in the unstirred layer on the inner surface of the cell membrane. Because virtually every ion channel and transporter is at least somewhat pH-sensitive, and some exquisitely so, these pH changes are expected to cause V_m changes. A second mechanism for CO₂-induced V_m changes is the presence of conductance pathways for HCO₃⁻. A third potential mechanism for V_m effects is carbamino formation, analogous to the rapid and reversible formation of carbamino hemoglobin, which leads to major conformational changes. Any or all of these CO₂/HCO₃⁻-induced effects could lead to V_m changes, some of which would probably be blocked or masked by the effects of DIDS. For these reasons, we believe that, when working with an undefined system, it is extremely dangerous to employ non-specific assays for electrogenic Na⁺:HCO₃⁻ cotransport, such as simply the application of DIDS and/or CO₂/HCO₃⁻.

Our initial approach was to isolate poly(A)+ RNA from the renal cortex of rabbits. Unfortunately, the V_m record in our Na⁺-removal assay showed no sign of a transient depolarization. Instead, removing extracellular Na⁺ simply caused V_m to become more negative, rapidly and reversibly. These V_m changes are characteristic of a classical Na⁺ conductance.

Reasoning that mRNA isolated from another amphibian might express better in Xenopus oocytes, our next approach was to isolated poly(A)+ RNA from the kidneys of the tiger salamander Ambystoma tigrinum. This was the species in which the electrogenic Na⁺:HCO₃⁻ cotransporter was first identified (Boron and Boulpaep, 1983). Indeed, we found that Na⁺ removal led to a rapid, but transient, depolarization and a slow pHi decrease. These effects were blocked by DIDS and did not occur in the nominal absence of CO₂/HCO₃⁻. Nor did these effects occur until several days after the injection of the RNA.

After screening a cDNA library by expression in Xenopus oocytes, we identified a single clone, which we have named NBC (for Na⁺:Bicarbonate Cotransporter). It is a full-length clone with a start Met and a poly(A)+ tail. The deduced amino-acid sequence of the encoded protein is 1025 amino acids long. Hydropathy analysis predicts multiple membrane-spanning domains. Comparing the DNA sequence of NBC with the GenBank data base reveals that NBC has no substantial homology to known proteins and is thus the first in a new family of Na⁺-coupled HCO₃⁻ transporters.

When cRNA made from the pure NBC clone is injected into Xenopus oocytes, the V_m and pHi changes are those predicted above. However, the magnitudes of the changes are substantially larger than when we injected oocytes with poly(A)+ RNA.

Using the NBC clone from Ambystoma, which we refer to as aNBC, we cloned its homologue from a rat kidney library. The rat clone, which we refer to as rNBC, is 86 % identical to aNBC. It encodes a protein with a predicted length of 1035 amino acids.

We predict that the 1:3 and 1:2 electrogenic Na⁺:HCO₃⁻ cotransporters will be sufficiently homologous at the base-pair level for it to be possible to use aNBC or rNBC (which presumably represent 1:3 electrogenic Na⁺:HCO₃⁻ cotransporters) to identify various 1:2 electrogenic Na⁺:HCO₃⁻ cotransporters. If there is a discrete 1:1 Na⁺:HCO₃⁻ cotransporter, then aNBC and rNBC may aid in the identification of that cDNA as well. We would expect that each of the major classes of Na⁺:HCO₃⁻ cotransporter (i.e. 1:3, 1:2 and 1:1) would have multiple isoforms, just as the Na⁺/H⁺ exchangers (NHEs) and AEs have multiple isoforms. The other two uncloned HCO₃⁻ transporters, the Na⁺-driven Cl⁻/HCO₃⁻ exchanger and the K⁺:HCO₃⁻ cotransporter, are more different, physiologically, from the various Na⁺:HCO₃⁻ transporters. In particular, the Na⁺-driven Cl⁻/HCO₃⁻ exchanger differs from the Na⁺:HCO₃⁻ cotransporter in that it has an absolute requirement for Cl⁻. The K⁺:HCO₃⁻ cotransporter differs from the Na⁺:HCO₃⁻ cotransporters in that it transports K⁺ rather than Na⁺ and that it is inhibited by quaternary alkyl ammonium ions, such as tetraethylammonium (B. A. Davis, E. M. Hogan, J. Zhao and W. F. Boron, unpublished observations), but is insensitive to DIDS. Thus, it will be interesting to see whether the NBCs can be used to identify the cDNAs encoding either of these other two HCO₃⁻ transporters.

As the cDNAs encoding the various HCO₃⁻ transporters become identified, comparing the structures and functions of the wild-type transporters may yield important clues as to which parts of the respective molecules might be important for determining such parameters as Na⁺:HCO₃⁻ stoichiometry and Cl⁻ binding. Truncations, deletions and point mutations to the cDNAs encoding these transporters, as well as chimeras made among their cDNAs, could yield additional insights into structure–function relationships of the HCO₃⁻ transporters. It is intriguing to speculate that, if the 1:3 electrogenic Na⁺:HCO₃⁻ cotransporter has separate binding sites for Na⁺, CO₃²⁻ and HCO₃⁻, the 1:2 cotransporter might lack the HCO₃⁻ binding site and the 1:1 cotransporter might lack the CO₃²⁻ binding site. Chimera experiments may make it possible to test this hypothesis.

As the cDNAs encoding the HCO₃⁻ transporters become available, it will be possible to establish the tissue distribution of their mRNAs. The development of antibodies specific for various HCO₃⁻ transporters and their isoforms may make it possible to determine the tissue distribution of the transporter proteins and to determine their topology. Antibodies suitable for immunocytochemistry could establish the cellular distribution (e.g. apical versus basolateral membrane of epithelial cells) of the transporter proteins. Finally, antibodies suitable for immunoprecipitation could permit biochemical analyses of transporter proteins.

In short, the next few years could be an exciting time for those interested in HCO₃⁻ transport, with the promised development of a host of molecular tools making it possible to probe the molecular physiology of HCO₃⁻ transporters to an unprecedented degree.

This work was supported by NIH grants DK30344, NS18400, DKI7433 and DK43171. M.F.R. was supported by grants from the National Kidney Foundation and the NIH (NRSA DK09342).