Cloning and characterization of a human electrogenic Na+- Formula cotransporter isoform (hhNBC)

Inyeong Choi, Michael F. Romero, Nassirah Khandoudi, Antoine Bril, Walter F. Boron


Our group recently cloned the electrogenic Na+- HCO3 cotransporter (NBC) from salamander kidney and later from mammalian kidney. Here we report cloning an NBC isoform (hhNBC) from a human heart cDNA library. hhNBC is identical to human renal NBC (hkNBC), except for the amino terminus, where the first 85 amino acids in hhNBC replace the first 41 amino acids of hkNBC. About 50% of the amino acid residues in this unique amino terminus are charged, compared with ∼22% for the corresponding 41 residues in hkNBC. Northern blot analysis, with the use of the unique 5′ fragment of hhNBC as a probe, shows strong expression in pancreas and expression in heart and brain, although at much lower levels. InXenopus oocytes expressing hhNBC, adding 1.5% CO2/10 mM HCO3 hyperpolarizes the membrane and causes a rapid fall in intracellular pH (pHi), followed by a pHi recovery. Subsequent removal of Na+ causes a depolarization and a reduced rate of pHi recovery. Removal of Cl from the bath does not affect the pHi recovery. The stilbene derivative DIDS (200 μM) greatly reduces the hyperpolarization caused by adding CO2/ HCO3 . In oocytes expressing hkNBC, the effects of adding CO2/ HCO3 and then removing Na+ were similar to those observed in oocytes expressing hhNBC. We conclude that hhNBC is an electrogenic Na+- HCO3 cotransporter and that hkNBC is also electrogenic.

  • intracellular pH
  • acid extruder
  • oocytes
  • bicarbonate

because almost every cellular process is sensitive to changes in intracellular pH (pHi), the regulation of pHi is important for normal cell function (9). Steady-state pHi is determined by the balance between acid extruders (i.e., transporters that move acid out of cells or alkali into cells) and acid loaders. Acid extruders include Na+/H+exchangers (the NHEs) (6), various H+ pumps, and the Na+-driven Cl/ HCO3 exchanger, which is present in invertebrate cells (34, 39) as well as cultured fibroblasts (28), cultured mesangial cells (13), and central nervous system neurons (37). Acid loaders include Cl/ HCO3 exchangers (the AEs) (3). In addition, a K+- HCO3 cotransporter, which would function as an acid loader, has been reported in squid giant axons (21, 22).

Electrogenic Na+- HCO3 cotransporters (10) can function either as acid extruders or acid loaders. In renal proximal tubules, the electrogenic Na+- HCO3 cotransporter has a Na+ to HCO3 stoichiometry of 1:3 and moves HCO3 out of the cell (4, 12); i.e., the cotransporter functions as an acid loader. In glial cells (8, 19), however, the electrogenic Na+- HCO3 cotransporter appears to have a stoichiometry of 1:2 and moves HCO3 into the cell (i.e., it is an acid extruder). Finally, electroneutral Na+- HCO3 cotransporters have been reported in vascular smooth muscle cells (1) and in the heart (18, 27). These transporters would have a stoichiometry of 1:1 and mediate HCO3 uptake. On the other hand, there is also a report that the cardiac cotransporter is electrogenic (15), presumably with a stoichiometry of 1:2.

Romero and co-workers (32) expression cloned the cDNA encoding an electrogenic Na+- HCO3 cotransporter (NBC) in the kidney of the tiger salamanderAmbystoma tigrinum. Others subsequently cloned similar cDNAs from mammalian kidneys (14, 33). The open reading frame of the NBC cDNA is ∼3.2 kb in length, corresponding to 1,035 amino acids. The amino acid sequences of the NBC are ∼30% identical to those of the AEs, and these two families of HCO3 transporters have very similar hydropathy profiles. Expression inXenopus oocytes ofAmbystoma NBC (aNBC) and rat kidney NBC (rkNBC) confirms that both clones encode electrogenic Na+- HCO3 cotransporters that are blocked by the stilbene DIDS (32, 33). These two NBCs have similar apparentK m values for extracellular HCO3 , ∼10 mM (20). Antibodies specific for rkNBC localize the cotransporter to the basolateral membrane of the renal proximal tubules of rat and rabbit (36).

As noted above, unlike the Na+- HCO3 cotransporters identified in kidney and glial cells, the cardiac Na+- HCO3 cotransporter described by one group is electroneutral (18, 27). Therefore, we decided to attempt to clone the cardiac NBC by screening a human heart cDNA library with rkNBC. We identified a cDNA from human heart (hhNBC) that is identical to the one from human kidney (hkNBC), except that the 85 amino-terminal amino acids in hkNBC are replaced by 41 amino-terminal amino acids in hhNBC. While our paper was being submitted, a paper appeared by Abuladze et al. (2), reporting the cloning of a cDNA with an identical sequence from pancreas. We found that both hhNBC and hkNBC, when expressed inXenopus oocytes, are electrogenic. This work is the first demonstration that a human NBC is electrogenic.


Cloning hhNBC. The open reading frame of rkNBC (GenBank Accession no. AF004017) was cut into three pieces with BstX I, radiolabeled with [α-32P]dCTP by random priming, pooled, and used to screen a human heart λ-ZAPII cDNA library (Stratagene, La Jolla, CA). The titrated plaques (0.64 × 106) were plated and blotted on nitrocellulose filters. Hybridization was performed in 0.5 M sodium phosphate, 7% SDS, 1% BSA at 65°C overnight. The membranes were washed at 50°C in 1× standard sodium citrate (SSC; 150 mM NaCl/15 mM sodium citrate)/0.1% SDS and then autoradiographed. We then performed a PCR with pBluescript vector primers adjacent to the cloning sites to amplify the cDNA inserts from primary positive plaques.1PCR products were separated on 1.0% agarose gel, transferred via capillary to a Hybond membrane (Clontech, Palo Alto, CA), and separately hybridized with each probe used for the screening library. Positive plaques were plated for a second screening with the same probes. We isolated four positive phage clones and excised the inserts by rescuing the plasmids. Sequencing was performed by the Keck Sequencing Center at Yale University. We analyzed the sequence using DNAsis (Hitachi Software, San Bruno, CA).

The above approach yielded nucleotide sequence data from both the 5′ end and the 3′ end of the putative clone. However, we were missing the middle of the putative clone. We performed RT-PCR with human heart poly(A)+ RNA Clontech (catalog no. 6533–1; a pool of whole heart from Caucasians ages 20–78 who had died of trauma) to obtain the missing middle portion. The upstream primer sequence was 5′-CCG GAG AAG GAC CAG CTG AAG-3′, corresponding to a region near the 3′ end of clone 15.1, which contains the 5′ fragment of hhNBC. The downstream primer sequence was 5′-ATC AGA GTA GGG AGG AAA GAG-3′, corresponding to a sequence at the TAG stop codon of clone 4.1, which contains the 3′ fragment of hhNBC. The result of the PCR2 was a DNA fragment that represented the middle and 3′ end of the putative clone, which we sequenced. We ligated this PCR product to clone 15.1 (the 5′ end) at the Pml I site, to obtain the putative full-length clone of hhNBC (GenBank Accession no. AF069510). We confirmed the sequence of this full-length clone near the ligation site.

To confirm that clones 15.1 and 4.1 represent, respectively, the 5′ and 3′ ends of the same clone, we ran a PCR between the ATG start codon of clone 15.1 and the TAG stop codon of clone 4.1. The result was a 3.2-kb PCR product, the sequence of which corresponded to the ligation product discussed above.

Cloning hkNBC. The 5′ fragment of hkNBC cDNA was isolated from human pancreas poly(A)+ RNA by RT-PCR with primers designed on the basis of the published hkNBC sequence (14). The upstream primer was 5′-TTG GGA GGC TTA GCA GGA AAG ATG TCC-3′ (−21 to +6), and the downstream primer was 5′-TTT CTT GGT TTG ATG CCG GTG CTT CCG-3′(+496 to +522). PCR was performed as described above,2 and the final single PCR product was cloned into the pCRII vector (Invitrogen, Carlsbad, CA). To obtain a full-length hkNBC, we ligated the above PCR product to hhNBC at the Afl II site. The inserted DNA fragment, as well as the ligation site, were sequenced to confirm that the ligation product is identical to the published hkNBC sequence (GenBank accession no. AF007216).

Northern blots. A human multiple-tissue Northern blot (catalog no. 7760–1) was purchased from Clontech. A 32P-labeled, random-hexamer-primed cDNA probe (GIBCO BRL, Gaithersburg, MD) was made from the unique 5′ region of the hhNBC (7–271 bp of the coding region of hhNBC). Hybridization was performed in the ExpressHyb hybridization buffer (Clontech) at 68°C for 2 h, at a probe concentration of 0.83 × 106counts ⋅ min−1 ⋅ ml−1. The membrane was washed at 37°C in 2× SSC/0.1% SDS for 40 min and then at 50°C in 0.1× SSC/0.1% SDS for 1.5 h. The membrane was exposed to Kodak X-Omat film (Kodak, Rochester, NY), which was developed 1 day later for the detection of high-intensity signals. For the detection of low-intensity signals, the film was autoradiographed for 7 days.

Membrane isolation and Western blot analysis. Oocyte plasma membranes were isolated as described by Preston et al. (30), with some modification. Groups of four oocytes were washed with the PBS and homogenized in 0.5 ml of fresh hypotonic lysis buffer (7.5 mM potassium phosphate, pH 7.4, 1 mM EDTA, 1 mM PMSF, 1 μg/ml pepstatin A, 1 μg/ml leupeptin). The cellular debris was removed by centrifuging at 810g for 5 min. For the collection of membranes, the supernatant was then centrifuged at 15,000g for 30 min at 4°C. The pellets were gently washed with the lysis buffer and were dissolved in 40 μl of a sample-loading buffer containing 2% SDS. The samples were separated on a 7.5% SDS-polyacrylamide gel and transferred to a polyvinylidene difluoride membrane. Blots were preincubated for 1 h in blocking buffer containing 0.2% I-block (Tropix, Bedford, MA) and 0.05% Tween 20 in Tris-buffered saline (TBS; 50 mM Tris, pH 7.4; 150 mM NaCl) and then were incubated with an antibody specific to the carboxy terminus of rkNBC (1:500 dilution) (36). After several washes with TBS containing 0.05% Tween 20, blots were incubated with anti-rabbit IgG conjugated to horseradish peroxidase (HRP; Sigma, St. Louis, MO) for 1 h (1:2,000 dilution in blocking buffer). Blots were washed and developed by HRP/hydrogen peroxide-catalyzed oxidation of luminol under alkaline conditions (Pierce, Rockford, IL).

pHi and membrane potential measurements in oocytes.

The hhNBC insert in the plasmid described above contains 44 bp of the 5′-untranslated region (UTR), as well as the entire coding region of hhNBC. The insert was digested withEcoR I and ligated into pGH19, an oocyte expression vector (40). The resulting construct and the hkNBC construct in pGH19 were linearized withNot I before transcription and then in vitro transcribed with the mMessage mMachine kit (Ambion, Austin, TX) using T7 RNA polymerase. The ratio of cap analog to GTP was decreased by increasing the final GTP concentration to 3 mM, thereby maximizing the synthesis of full-length transcripts. Oocytes ofX. laevis were obtained as described by Romero et al. (32). Defolliculated oocytes (stages V and VI) were injected with 50 nl of RNA (0.5 mg/ml) or water and incubated in OR3 buffer (50% Leibovitz L-15 media withl-glutamine, 5 mM HEPES, pH 7.5) supplemented with 5 U/ml penicillin-streptomycin (33). Injected oocytes were maintained for 3–7 days at 18°C before use.

pH-sensitive microelectrodes were made from borosilicate glass capillaries (2.0-mm outer diameter; Warner Instrument, Hamden, CT) that were pulled on a microelectrode puller (model P-97; Sutter, Novato, CA), dried in an oven at 200°C for at least 2 h, and silanized in the vapor of bis(dimethylamino)-dimethyl silane (Fluka, Ronkonkoma, NY) in a closed vessel. The electrode tip was then filled with hydrogen ionophore 1 cocktail B (Fluka) and backfilled with a pH 7.0 phosphate buffer (38). The electrode was connected to a high-impedance electrometer (model FD-223; WPI, Sarasota, FL) and calibrated in standard solutions of pH 6 and 8. The slope was generally 55–57 mV/pH unit.

Physiological solutions. Table1 summarizes the composition of the solutions used in the electrophysiological experiments. Choline, gluconate, and DIDS were all obtained from Sigma.

View this table:
Table 1.

Experimental solutions

Statistics. Initial rates of change of pHi(dpHi/dt) were determined by using a computer to fit a line to the pHi vs. time record. Statistical significance was judged from Student’st-tests. Data are expressed as means ± SE.


Sequence analysis of hhNBC cDNA. By screening a human heart cDNA library and subsequently employing a PCR approach, we cloned a full-length cDNA, hhNBC. The open reading frame of hhNBC encodes a protein of 1,079 amino acids, with a predicted molecular mass of 121 kDa. As shown in Fig.1 A, this clone has a 5′ coding region that is different from that of the previously published human kidney clone, hkNBC (14). The first 255 bp (i.e., 85 amino acids) of the coding region of hhNBC replaces the first 123 bp (41 amino acids) of hkNBC (Fig.1 B). In addition, hhNBC has a phenylalanine at amino acid position 255 instead of the serine reported for hkNBC. We obtained the portion of hhNBC that encodes the F255 by screening a cDNA library. Except for these differences, hhNBC is identical to hkNBC, including the 3′-UTR.

Fig. 1.

A: nucleotide and amino acid sequences of the unique 5′ end of human heart Na+- Formula cotransporter (hhNBC). The 5′ untranslated region is shown in lower case letters. B: comparison of deduced amino-terminal amino acid sequences for hhNBC and human kidney NBC (hkNBC). First 85 amino acids of hhNBC replace first 41 amino acids of hkNBC. Gray regions are conserved residues.C: hydropathy plot of hhNBC. About 50% of the 85 amino acids of hhNBC are charged and polar, and this becomes hydrophilic. Inset: first 100 amino acids of hhNBC are compared with corresponding regions of hkNBC.

The unique amino terminus of hhNBC consists mainly of charged amino acids, such as aspartate, glutamate, histidine, lysine, and arginine. These charged residues comprise ∼50% of the unique 85 amino acids in hhNBC, considerably higher than the ∼22% of residues that are charged among the corresponding 41 amino acids of hkNBC. Thus the amino-terminal region of hhNBC is very hydrophilic (Fig.1 C). This region is reminiscent of AE2 and AE3, which also have large, amino-terminal domains consisting of highly charged amino acids. There is 20–30% sequence similarity among the amino-terminal charged domains of hhNBC, AE2, and AE3, but not AE1. Finally, the unique amino terminus of hhNBC contains a second putative consensus protein kinase A (PKA) phosphorylation at amino acid position 49, in addition to a conserved PKA site at amino acid position 1023.

Tissue distribution of hhNBC mRNA.Using the unique 5′ region (nucleotide positions 7–271 of the coding regions) of hhNBC as a probe, we performed a high-stringency Northern blot analysis in various human tissues. After a 1-day exposure, the probe showed substantial hybridization to an ∼9-kb transcript in pancreas but not to any of the other tissues on the blot (Fig.2 A). Because we expected to detect hhNBC mRNA in the heart, having cloned hhNBC from human heart tissues, we exposed the film for 7 days to determine whether the signal could be detected in heart. We found that hhNBC mRNA is strongly expressed in pancreas and weakly expressed in heart and brain (Fig. 2 B). hhNBC mRNA is undetectable in the kidney. It is interesting to note that, inAmbystoma (i.e., tiger salamander) tissues, Romero et al. (32) did not detect signals in pancreas, even though they probed with a full-length cDNA.

Fig. 2.

Northern blot of human tissue. cDNA probe corresponded to the 5′ coding region (nucleotides 7–271) of hhNBC. Probe recognizes an ∼9-kb transcript. A: 1 day of exposure. B: 7 days of exposure. Tissues examined include pancreas (P), kidney (K), skeletal muscle (S), liver (L), lung (Lu), placenta (Pl), brain (B), and heart (H).

Abuladze et al. (2) probed a Northern blot, obtained from the same source as ours, with part of the unique 5′ region of hhNBC. They detected strong expression in pancreas and moderate expression in brain, but not heart. This disparity may be due to the probe specificity and/or stringency. We used nearly the entire unique nucleotide (7–271) as a probe, whereas Abuladze et al. used a synthetic oligonucleotide probe corresponding to nucleotides 118–212 (95 bp).

Immunoblot experiments on expression of hhNBC in oocytes. To study the functional properties of hhNBC, we expressed hhNBC inXenopus oocytes by injecting them with hhNBC cRNA. We first tested whether the open reading frame of hhNBC could be translated into protein with the expected molecular mass. We used a rabbit reticulocyte system (without microsomes) to perform an in vitro translation of hhNBC cRNA and observed a 120-kDa product (not shown), as expected from the deduced amino acid sequence of hhNBC cDNA, which predicts a molecular mass of 121 kDa. We then injected the same cRNA into the oocytes and tested the expression of hhNBC by Western blot analysis, using an antibody against the carboxy terminus of rkNBC (36). Because the deduced amino acid sequences of hhNBC and rkNBC are ∼98% identical at their carboxy termini, we expected the antibody against rkNBC to react with the hhNBC protein. Figure3 shows that the antibody reacted with a 130-kDa protein in oocytes injected with rkNBC cRNA (lane 3). The same antibody also recognized an ∼135-kDa band in oocytes injected with hhNBC cRNA (lane 2). The slightly greater molecular mass is expected because the 85 amino acids at the amino terminus of hhNBC replace the 41 amino acids at the amino terminus of rkNBC. In addition to the major band, the antibody detects a smaller band (∼130 kDa), which may be the result of proteolysis or a less extensive glycosylation. Water-injected control oocytes showed no immunoreactivity (lane 1).

Fig. 3.

Immunoblot of hhNBC in oocytes. Membrane preparations from oocytes injected with cRNA for hhNBC, rat kidney NBC (rkNBC), and control water were separated by 7.5% SDS-PAGE and immunoblotted with an antibody (MBP-NBC-5) specific to rkNBC (36). In oocytes injected with rkNBC cRNA, the antibody detected a single band of ∼130 kDa, whereas in oocytes injected with hhNBC cRNA, the antibody detected a major band at ∼135 kDa as well as a slightly smaller minor band.

Functional properties of hhNBC expressed in Xenopus oocytes. To study the functional properties of hhNBC, we used microelectrodes to monitor pHi and membrane potential (V m) in control oocytes as well as in oocytes expressing hhNBC. Figure4 Aillustrates the effects of exposing a water-injected oocyte to a solution buffered with 1.5% CO2/10 mM HCO3 . Figure4 B is a similar experiment, but on an oocyte expressing hhNBC. In both oocytes, applying CO2/ HCO3 caused a decrease in pHi. In the case of the control oocyte, pHifell rapidly and by a relatively large amount and then recovered very slowly. In the case of the hhNBC oocyte, the pHi decrease was smaller and slower. Moreover, pHi recovered rather rapidly. Thus the pHiprofile in the hhNBC oocyte is consistent with an extremely robust expression of the cotransporter, which not only produced a pHi recovery but also blunted the initial CO2-induced acidification. Applying CO2/ HCO3 also caused characteristic changes inV m, a slowly developing depolarization in the water-injected oocyte but a very large and rapid hyperpolarization in the hhNBC oocyte. The hyperpolarization wanes in the case of the hhNBC oocyte because, for three reasons, the hhNBC-mediated inward transport of Na+ and HCO3 gradually slows down:1) CO2 diffuses into the cell and produces HCO3 ,2) hhNBC actively transports HCO3 into the cell, and3) hhNBC transports Na+ into the oocyte. In each case, an hhNBC substrate slowly builds up inside the cell, slowing inward transport. Note, however, thatV m is always more negative in the presence than in the absence of CO2/ HCO3 . Thus, as previously shown for aNBC (32) and rkNBC (33), the direction of electrogenic Na+- HCO3 cotransport by hhNBC is inward in oocytes.

Fig. 4.

Physiological characterization of the hhNBC clone expressed in oocytes.A: water-injected, control oocyte. Three days after injection of water, intracellular pH (pHi) and membrane potential (V m) were monitored with microelectrodes before and during application of a solution containing 1.5% CO2/10 mM Formula (choline was Na+ substitute; typical of 5 experiments). B: hhNBC-expressing oocyte. Protocol was same as in A. Note that application of CO2/ Formula caused a slower and smaller pHidecrease, from which the oocyte recovered rather robustly. Also, application of CO2/ Formula caused a large and rapid hyperpolarization (rather than a slow depolarization). Finally, removal of Na+ caused a robust acidification (rather than no change) and depolarization (rather than a hyperpolarization); typical of 8 experiments.C: effect of Cl removal on pHi in oocyte expressing hhNBC. Gluconate replaced Cl. Removing extracellular Cl in presence of CO2/ Formula had no significant effect on rate of pHi recovery and change inV m (period between two arrows); typical of 4 experiments.D: effect of DIDS on pHi andV m in oocyte expressing hhNBC. Before exposing the oocyte to 200 μM DIDS, the oocyte was briefly pulsed with 1.5% CO2/10 mM Formula (arrow; segmentabc). In continued presence of DIDS, second exposure to CO2/ Formula elicited a reduced hyperpolarization (def). DIDS reducedV m elicited by CO2/ Formula by ∼70%; typical of 4 experiments with 200 μM DIDS. Not shown are 3 other experiments with 500 μM DIDS, in which DIDS reducedV m by >90%.

To test whether the expressed transporter is coupled to Na+, we next removed Na+ from the bath in the continued presence of CO2/ HCO3 . Removing extracellular Na+(replacing Na+ with choline) had little effect on pHi in the control oocyte (Fig. 4 A) but produced a pHi decrease in the hhNBC oocyte. This intracellular acidification is consistent with the “reversal” of HCO3 transport (i.e., Na+-coupled HCO3 efflux rather than influx). Removing extracellular Na+ also produced characteristicV m changes in the two oocytes. In the control oocyte, Na+ removal elicited a small hyperpolarization, reflecting a small Na+ conductance, as previously observed (32, 33). In the hhNBC oocyte, however, Na+ removal produced a large depolarization (from −80 to −10 mV), consistent with the electrogenic efflux of Na+ with more than one HCO3 . Returning Na+ to the outside of the oocytes reversed the pHi andV m changes.

To test the unlikely possibility that hhNBC encodes a Na+-driven Cl/ HCO3 exchanger, we removed Clfrom the bath (replacing Clwith gluconate). As shown in Fig. 4 C, removing extracellular Clin the presence of CO2/ HCO3 had no significant effect on either the rate of pHi recovery or the membrane potential (n = 4). If the pHi recovery had been due to a Na+-driven Cl/ HCO3 exchanger, then Cl removal should have speeded up the pHirecovery, at least transiently. Thus our results indicate that hhNBC does not transport Cl, although we cannot rule out the unlikely possibility that hhNBC may require intracellular Cl.

Finally, we addressed the question of whether this electrogenic, Na+-dependent (and Cl-independent) HCO3 cotransport process can be blocked by the stilbene derivative DIDS, which is known to inhibit electrogenic Na+- HCO3 cotransport in other systems (11). Before exposing the oocyte to DIDS, we briefly pulsed the cell with CO2/ HCO3 and then returned it to the CO2/ HCO3 -free ND96 media (arrow in Fig. 4 D). This brief CO2/ HCO3 pulse produced a hyperpolarization of nearly 60 mV in this experiment. We then applied 200 μM DIDS in the absence of CO2/ HCO3 . In the continued presence of DIDS, the second exposure to CO2/ HCO3 elicited a hyperpolarization of >20 mV. Three other similar experiments showed that 200 μM DIDS blocked, on average, 70 ± 4% of the voltage change elicited by CO2/ HCO3 (P < 0.01). Figure4 D also shows that DIDS failed to block completely the depolarization and acidification elicited by removing extracellular Na+. In contrast, 200 μM DIDS completely blocks aNBC expressed in oocytes (31, 32). Not shown are experiments in which we found that 500 μM DIDS reduced the hyperpolarization by ∼90% (n = 3).

Effect of substituting the amino terminus of hkNBC for the amino terminus of hhNBC. Our functional studies described above indicate that, despite its unique amino terminus, hhNBC is an electrogenic Na+- HCO3 cotransporter, similar to rat kidney NBC or salamander kidney NBC. However, the electrogenicity had yet to be examined for human kidney NBC, which is nearly identical to hhNBC, except for a different amino terminus. To address the functional impact of the charged amino terminus of hhNBC, we therefore constructed hkNBC by performing RT-PCR with primers specific to the amino terminus of hkNBC and replacing the unique amino terminus of the hhNBC cDNA with the one for hkNBC. We then expressed the hkNBC in oocytes and monitored the pHi andV m as described above.

Figure 5 shows that applying CO2/ HCO3 caused an acidification, followed by a pHi recovery. The application of CO2/ HCO3 also causes a rapid hyperpolarization that slowly wanes (n = 5), as in hhNBC-expressing oocytes. Removing extracellular Na+ (replacing Na+ with choline) produced an intracellular acidification, as well as a rapid depolarization. The pHi andV m effects of adding CO2/ HCO3 and removing Na+ were fully reversible. Thus, as is the case for the other NBCs, hkNBC is electrogenic.

Fig. 5.

Electrophysiological characterization of the “kidney” clone (i.e., hkNBC) expressed in an oocyte. Applying 1.5% CO2/10 mM Formula caused an acidification, followed by a pHi recovery. Application of CO2/ Formula also caused a rapid hyperpolarization that slowly waned. Removing extracellular Na+ (replacing Na+ with choline) produced a reversible intracellular acidification, as well as a rapid depolarization; typical of 5 experiments.

We noticed that the CO2/ HCO3 pulse generally produced a smaller hyperpolarization in hkNBC-expressing oocytes than hhNBC-expressing oocytes. hhNBC and hkNBC differ in the nucleotide sequence surrounding the initiator methionine. The hhNBC cDNA has a perfect Kozak sequence (A GGATG̲ G), whereas hkNBC cDNA has an imperfect sequence (A AGATG̲ T). The highly conserved G in position +4 immediately following the ATG codon in hhNBC is critically important for halting the scanning 40S ribosomal subunit and thus to initiate translation (26). Thus hhNBC and hkNBC may be translated with different efficiencies.


Cloning of hhNBC from human heart. In this study, we report the cloning of an NBC-related cDNA from human heart tissues. Studies on pHiregulation of heart cells have provided conflicting evidence for both an electroneutral Na+- HCO3 cotransporter (18, 27) and an electrogenic Na+- HCO3 cotransporter (15). The evidence for the electroneutral cotransporter was obtained in guinea pig myocytes and sheep Purkinje fibers, whereas the data supporting the presence of an electrogenic Na+- HCO3 cotransporter was obtained in cat myocardium. Thus it is possible that the conflicting conclusions of the two groups reflect a species difference. Another possibility is that both transporters are present in the same tissue.

We cloned hhNBC from heart tissue of humans. hhNBC has a unique amino-terminal domain, which consists mainly of charged or polar amino acid residues, and is definitely electrogenic. Because there are no functional studies of the Na+- HCO3 cotransporter in human cardiac tissue, it is possible that hhNBC represents the (or at least one of the) Na+- HCO3 cotransporters in human heart tissue. A question that arises is why the level of mRNA is apparently so low in cardiac tissue.

One possibility is that the heart mRNA used in our Northern blot was partially degraded. Another possibility is that hhNBC mRNA levels are indeed low in human heart. However, even low mRNA levels may be sufficient to maintain high levels of the hhNBC protein. On the other hand, even if the hhNBC protein levels are low, the activity of each hhNBC molecule may be sufficiently high to produce a high functional level of hhNBC activity in human cardiac cells. Yet a third possibility is that hhNBC is expressed at high levels, but only in specialized cell types (e.g., endothelium, vascular smooth muscle cells, sinoatrial nodal cells). Finally, it is possible that hhNBC is normally expressed at relatively low levels but is upregulated in response to certain stimuli (e.g., training, hypoxia).

hhNBC is an isoform of electrogenic renal NBC. Except for the first 85 amino acid residues at the cytoplasmic amino terminus of hhNBC, hhNBC and hkNBC are identical at the amino acid level. It is likely that hhNBC and hkNBC represent alternative splice products. Supporting this hypothesis is the observation that the partial fragment (1.6 kb) of the 3′-UTR of hhNBC is identical to the corresponding region of hkNBC (data not shown). The UTRs are under less selective pressure during evolution and thus are more prone to mutations. If hhNBC and hkNBC originate from separate genes, we should expect sequence variations between hhNBC and hkNBC in the 3′-UTR, which is not the case.

Abuladze et al. (2) reported a cDNA from pancreas (hpNBC) that, in its coding region, has a sequence that is identical to that of hhNBC. Interestingly, the 5′-UTR of hhNBC is different from that reported for hpNBC. One possible explanation for this difference is that hhNBC and hpNBC are encoded by the same gene but have different tissue-specific 5′-UTRs and thus different regulation at the mRNA level. There is evidence that alternatively spliced 5′-UTRs can play a role in posttranscriptional regulation in other genes, including theN-methyl-d-aspartate receptor subunit 2 (25, 35), human erythrocyte protein 4.1 (7), human inducible nitric oxide synthase (17), and the excitatory amino acid transporter EAAT2 (29). The genomic structure(s) of NBC should clarify this issue. It is interesting to note that the 5′-UTR of hpNBC is 100% identical to the known 29 bp of the 5′-UTR of the NBC cloned from mouse pancreas (GenBank accession no. AF020195, cited in Ref. 2). Moreover, the 5′-UTR of hpNBC is 88% identical to the 5′-UTR of a rat NBC-related cDNA (accession no. AI144995).

In the present study, we report that both hhNBC and hkNBC are electrogenic. Because both cotransporters are electrogenic, despite the differences in their amino termini, our results indicate that different regions of the respective amino termini are not responsible for the electrogenicity of NBC. However, hhNBC appears to be less sensitive to DIDS than AmbystomaNa+- HCO3 cotransport (31, 32) or rkNBC (M. F. Romero, unpublished results). This result suggests that the amino terminus, which is believed to be cytoplasmic, may be able to influence the binding of DIDS to the extracellular surface.

Electrogenicity of hkNBC. Burnham et al. (14) cloned hkNBC from human kidney and verified that it cotransports Na+ and HCO3 and that it is blocked by DIDS. We have now demonstrated that hkNBC, like the renal NBC clones from salamander and rat, is electrogenic. Our results on hhNBC and hkNBC thus provide the first molecular evidence for the existence of an electrogenic Na+- HCO3 cotransporter from a human source. Moreover, our data demonstrate that more than one variant of an electrogenic Na+- HCO3 cotransporter exists in humans.

Possible physiological function of hhNBC. We found that the unique amino terminus of hhNBC, which encodes 85 amino acids, is strongly expressed in the pancreas, as shown by Northern blot analysis. Indeed, Abuladze et al. cloned from the pancreas a cDNA that is identical to hhNBC (2). In the amino terminus of hkNBC, 41 amino acids replace the 85 amino acids of hhNBC. What might be the significance for the pancreas of the unique amino terminus of hhNBC? The pancreatic ducts rapidly secrete HCO3 into the duct lumen, producing a luminal HCO3 concentration that is fivefold higher than the HCO3 concentration in the blood plasma. This HCO3 secretion neutralizes a large portion of the gastric acid that reaches the duodenum. Physiological experiments have shown that pancreatic duct cells have a Na+-dependent HCO3 transporter at the basolateral membrane that mediates net HCO3 uptake into the duct cells (23, 24). This pancreatic transporter differs functionally from hkNBC in at least one crucial respect: the renal cotransporter mediates net HCO3 efflux, whereas the pancreatic cotransporter mediates net HCO3 influx. The difference between net HCO3 efflux (kidney) and influx (pancreas) could be dictated simply by the ion and voltage gradients in the respective cell type or by differences in posttranslational modification. However, it is possible that the unique 85 amino acids of hhNBC change the stoichiometry of the cotransporter from 1:3 to 1:2.

Another possible role for the unique 85 amino acids at the amino terminus of hhNBC is suggested by the observation that these 85 amino acids include a novel consensus phosphorylation site for PKA. Both hhNBC and hkNBC have consensus PKA sites near the carboxy terminus, but only hhNBC has such a site near the amino terminus. It is interesting to recall that secretin strongly stimulates both net HCO3 secretion by pancreatic ducts (5) as well as HCO3 uptake across the basolateral membrane via the putative pancreatic NBC (23). Secretin interacts with specific membrane receptors on pancreatic duct cells and activates the cAMP/PKA cascade. Thus the presence of a second PKA site in the amino terminus of hhNBC could be important for activation of HCO3 transport in the pancreas by secretin.

In summary, an NBC isoform was cloned from human heart tissues (hhNBC). The hhNBC clone, when expressed inXenopus oocytes, reveals physiological properties similar to the renal electrogenic Na+- HCO3 transporter.


We thank Duncan Wong for computer programming support and Dr. Gritchenko for technical support.


  • Address for reprint requests: I. Choi, Dept. of Cellular and Molecular Physiology, Yale Univ. School of Medicine, 333 Cedar St., New Haven, CT 06520 (E-mail: inyeong.choi{at}

  • This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-30344. For part of the time, I. Choi was supported by a postdoctoral fellowship from the American Heart Association. M. F. Romero was supported by a National Research Service Award (DK-09342) and a grant from the American Heart Association.

  • Portions of this work have been published in preliminary form (16).

  • Present address of M. F. Romero: Dept. of Physiology and Biophysics, Case Western Reserve University, School of Medicine, 2119 Abington Rd., Cleveland, OH 44106-4790.

  • 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 We used 35 cycles of PCR, with denaturation at 94°C for 1 min, annealing and extension at 68°C for 3 min, and 7 min at 68°C for the last extension.

  • 2 In the first 10 PCR cycles, we denatured at 94°C for 30 s, annealed at 65°C for 1 min, and extended at 68°C for 2 min. The second 20 cycles were identical, except that we lengthened the duration of the extension by 20 s in each cycle. After the 30th cycle, we allowed an additional 7 min for extension at 68°C.


View Abstract