Using pH- and voltage-sensitive microelectrodes, as well as the two-electrode voltage-clamp and macropatch techniques, we compared the functional properties of the three NBCe1 variants (NBCe1-A, -B, and -C) with different amino and/or carboxy termini expressed in Xenopus laevis oocytes. Oocytes expressing rat brain NBCe1-B and exposed to a CO2/HCO3− solution displayed all the hallmarks of an electrogenic Na+/HCO3− cotransporter: (a) a DIDS-sensitive pHi recovery following the initial CO2-induced acidification, (b) an instantaneous hyperpolarization, and (c) an instantaneous Na+-dependent outward current under voltage-clamp conditions (−60 mV). All three variants had similar external HCO3− dependencies (apparent KM of 4–6 mM) and external Na+ dependencies (apparent KM of 21–36 mM), as well as similar voltage dependencies. However, voltage-clamped oocytes (−60 mV) expressing NBCe1-A exhibited peak HCO3−-stimulated NBC currents that were 4.3-fold larger than the currents seen in oocytes expressing the most dissimilar C variant. Larger NBCe1-A currents were also observed in current–voltage relationships. Plasma membrane expression levels as assessed by single oocyte chemiluminescence with hemagglutinin-tagged NBCs were similar for the three variants. In whole-cell experiments (Vm = −60 mV), removing the unique amino terminus of NBCe1-A reduced the mean HCO3−-induced NBC current 55%, whereas removing the different amino terminus of NBCe1-C increased the mean NBC current 2.7-fold. A similar pattern was observed in macropatch experiments. Thus, the unique amino terminus of NBCe1-A stimulates transporter activity, whereas the different amino terminus of the B and C variants inhibits activity. One or more cytosolic factors may also contribute to NBCe1 activity based on discrepancies between macropatch and whole-cell currents. While the amino termini influence transporter function, the carboxy termini influence plasma membrane expression. Removing the entire cytosolic carboxy terminus of NBCe1-C, or the different carboxy terminus of the A/B variants, causes a loss of NBC activity due to low expression at the plasma membrane.
Introduction
Na/HCO3 cotransporters (NBCs) are functionally diverse proteins that are involved in the regulation of intracellular pH (pHi), absorption or secretion of HCO3−, and maintenance of ion homeostasis in many tissues. The first Na+/HCO3− cotransporter was identified by function in the proximal tubule of the kidney (Boron and Boulpaep, 1983), where the transporter has a 1:3 Na+:HCO3− stoichiometry (Soleimani et al., 1987) and is responsible for reabsorbing as much as ∼90% of filtered bicarbonate (Boron et al., 1997). In glial cells of the central nervous system, an electrogenic Na+/HCO3− cotransporter with a 1:2 Na+:HCO3− stoichiometry (Deitmer and Schlue, 1989; O'Connor et al., 1994; Bevensee et al., 1997a,b) contributes to intracellular and extracellular pH changes that can influence neuronal activity (Chesler, 2003; McAlear and Bevensee, 2004).
At the molecular level, the first cDNA encoding an electrogenic NBC was identified by expression from salamander kidney (Romero et al., 1997). The cloning of salamander kidney NBC paved the way for homology cloning of both electrogenic and electroneutral NBCs, as well as other cation-coupled anion transporters (for review see Romero et al., 2004). Cation-coupled anion transporters in conjunction with anion exchangers (AEs) are members of a superfamily of bicarbonate transporters (BTs). Using the convention introduced by Choi et al. (2000) and expanded by Romero et al. (2004), we refer to the first cloned NBC as NBCe1-A, where “e” refers to electrogenic, “1” refers to the first gene cloned of this family, and “A” refers to the first splice variant identified. The putative membrane topology of NBCe1 shown in Fig. 1 is based on sequence alignment of NBCe1 and AE1 (Romero et al., 1998b). We mapped the NBCe1 sequence on a topology model of the related bicarbonate transporter anion exchanger 1 or AE1 (Taylor et al., 2001). As described by Taylor et al. (2001), the putative topology of AE1 is determined from proteolysis and cysteine accessibility data.
NBCe1 proteins arise from different splice variants of gene SLC4A4 (Abuladze et al., 2000). The electrogenic Na/HCO3 cotransporters can be categorized into one of three groups based on differences at the amino and/or carboxy termini (Fig. 1): NBCe1-A (“kidney” clone), NBCe1-B (“heart,” “pancreas,” and rat-brain 1 NBC clone), and NBCe1-C (rat-brain 2 NBC clone). At the amino acid level, NBCe1-B is identical to NBCe1-A except at the amino terminus where 85 residues replace the 41 residues of NBCe1-A. NBCe1-C is identical to NBCe1-B except at the carboxy terminus where 61 residues replace the 46 residues of NBCe1-B. As reported for the A and B clones from human (Choi et al., 1999), the unique amino terminus of the rat A variant is only 6% identical and 13% homologous to that of the rat B and C variants. Furthermore, while both termini are hydrophilic, ∼50% of the 85 amino-terminal residues of the B and C variants are charged residues in contrast to only ∼22% of the 41 unique amino-terminal residues of the A variant. Although NBCe1-C has only been reported from rat brain, exon excision between introns at positions 23 and 24 in human SLC4A4 (Abuladze et al., 2000) would create the unique carboxy terminus. Indeed, using RT-PCR techniques, we have recently identified the full-length cDNAs encoding both NBCe1-B and NBCe1-C from human brain (unpublished data).
Although the physiological significance of the three NBCe1 variants with different amino and/or carboxy termini is not fully known, several investigators have identified specific residues/regions within the cytoplasmic amino and carboxy termini of NBCe1 that contribute to protein function, regulation, and expression. For example, in Ussing-chamber experiments on transfected mouse renal proximal tubule cells, Gross et al. reported that PKA-dependent phosphorylation of the Ser at position 982 within the cytoplasmic carboxy terminus of the A variant (Gross et al., 2001), or position 1026 of the B variant (Gross et al., 2003), changes transporter stoichiometry from 1:3 to 1:2 Na+:HCO3−. This Ser within the cytoplasmic carboxy terminus is common in all three variants. As proposed by the authors, the phosphorylation of the carboxy terminus may induce a conformation change that blocks a HCO3− binding site on the transporter. The stoichiometry of the A and B variants appears to be cell type dependent, and not due to the different amino termini (Gross et al., 2001). On the other hand, the amino terminus does contribute to the regulation of transporter activity. For example, Gross et al. (2003) report that the Thr at position 49 of human NBCe1-B is required for the cAMP-stimulated increase in transporter activity. This increase does not appear to involve a change in the phosphorylation of the Thr. Curiously, this Thr is one of two residues that are different in the homologous amino-terminal region of rat NBCe1-B and -C. The carboxy terminus of NBCe1 can also influence expression of the protein. For instance, removing the carboxy-terminal 23 residues of NBCe1-A causes a mistargeting of the transporter to the apical instead of the basolateral membrane when transfected into kidney epithelial cells (Li et al., 2004).
In addition to structure–function data, there is also genetic information that highlights the importance of the cytosolic amino terminus of NBCe1 before the first predicted transmembrane domain. There are human patients with mutations in SLC4A4 who present primarily with proximal renal tubular acidosis (pRTA) and ocular abnormalities (Igarashi et al., 1999; Dinour, D., A. Knecht, I. Serban, and E.J. Holtzman. 2000. J. Am. Soc. Neurol. 11:3A; Igarashi, T., J. Inatomi, T. Sekine, Y. Yakeshima, N. Yoshikawa, and H. Endou. 2000. J. Am. Soc. Nephrol. 11:106A; Dinour et al., 2004). One patient has an inactivating homozygous missense mutation in which a Ser replaces Arg at position 298 in the cytoplasmic amino terminus of NBCe1-A (Igarashi et al., 1999). In expression studies using ECV304 cells, the authors found that this mutant NBC displayed only ∼50% of wild-type transporter activity. This substitution is found at position 342 in the B and C variants. A second patient has a homozygous missense mutation in which a Leu replaces Ser at position 427 at the beginning of the first predicted transmembrane domain of NBCe1-A (Dinour et al., 2004). When expressed in oocytes, the mutant NBC displayed only ∼10% of wild-type transporter activity. In a separate study, this mutant NBC was mistargeted to the apical instead of the basolateral membrane of polarized Madin-Darby canine kidney (MDCK) cells (Li et al., 2005).
According to the aforementioned studies, regions within the cytoplasmic amino and carboxy termini can influence the function, regulation, and expression of NBCe1 variants. However, a detailed comparison of the biophysical properties of all three variants with different amino and/or carboxy termini has yet to be performed. In the present manuscript, we used pH- and voltage-sensitive microelectrodes to characterize, for the first time, the function of rat brain NBCe1-B expressed in Xenopus oocytes. Subsequently, we used the two-electrode, voltage-clamp technique to compare the activities, as well as the ion and voltage dependencies of all three NBCe1 variants. Although all three variants have similar external ion and voltage dependencies, as well as plasma membrane expression levels as assessed by single oocyte chemiluminescence, the activity of the A variant is greater than the activities of the other two variants. According to structure–function analyses using whole-cell and macropatch recording techniques, the higher activity of A is due to its unique amino terminus. While the different amino termini influence function, one or more regions within the different carboxy termini contribute(s) to plasma membrane expression.
Portions of this work have been published in abstract form (Williams, J.B., and M.O. Bevensee. 2002. FASEB J. 16:A796; McAlear, S.D., J.B. Williams, and M.O. Bevensee. 2004. FASEB J. 18:A1022–A1023; Liu, X., S.D. McAlear, and M.O. Bevensee. 2005. FASEB J. 19:A142; McNicholas-Bevensee, C.M., X. Liu, S.D. McAlear, and M.O. Bevensee. 2006. FASEB J. 20:A1232).
Materials And Methods
NBC Constructs and Mutagenesis
Wild-type NBCe1 Variants.
We used cDNAs encoding NBCe1-A from rat kidney (Romero et al., 1998b), and NBCe1-C from rat brain (Bevensee et al., 2000) subcloned into the oocyte expression vector pTLNII as previously described. Full-length NBCe1-B, which was also previously identified from rat brain by RT-PCR (Bevensee et al., 2000), was constructed using convenient restriction enzymes to swap the unique carboxy terminus of full-length variant C in pTLNII with the carboxy terminus of variant B from a partial-length construct in pBluescript. For both wild-type and mutant constructs (see below), NBC expression was optimized by introducing a Kozak sequence (Kozak, 1986) at the initiator codon using PCR techniques and a Peltier thermal cycler (PTC-220 DNA Engine Dyad, MJ Research, Inc.). Sequence analysis and primer design were performed using either DNAsis (Hitachi Software) or Vector NTI Advance 9.0 (InforMax, Invitrogen), and all constructs were confirmed by bidirectional DNA sequencing (DNA Sequencing Core, Center for AIDS Research and the Genomics Core Facility, Heflin Center for Human Genetics, both at the University of Alabama at Birmingham).
HA-tagged NBCe1 Constructs.
To insert the nine-residue hemagglutinin (HA) epitope (YPYDVPDYA) into NBCe1-C, we used the QuikChange PCR-based mutagenesis kit (Stratagene) to create a unique restriction site (Bsu36I) at base pair 1941 (residue 647) within the extracellular loop between transmembrane domains 5 and 6. A double-stranded pair of 5′- and 3′-phosphorylated primers encoding the HA epitope was flanked by Bsu36I sticky ends. Equal concentrations of the two primers were first denatured in a thermocycler (Genius FGEN05TP, Techne) by heating to 94°C (15 s), and then allowed to anneal by slowly cooling the reaction in increments of 2°C (15 s each) to 37°C. This double-stranded primer pair was then ligated into the engineered Bsu36I site in NBCe1-C. Because of the introduced Bsu36I cut sites, the inserted HA epitope in the protein was flanked by Asp and Gly. HA-tagged NBCe1-A and -B were constructed from the tagged NBCe1-C using convenient restriction enzymes.
Truncated NBCe1 Variants.
Truncated NBCe1 constructs (see Fig. 1) were generated using PCR techniques and HA-tagged NBCe1 variants subcloned into pTLNII as templates. In generating amino-terminal truncations, we targeted residue 43 of the A variant and the homologous residue 87 of the C variant to optimize the Kozak sequence. NBCe1 constructs truncated at the carboxy terminus were generated by introducing a targeted stop codon using site-directed mutagenesis (QuikChange kit, Stratagene).
Functional Studies on NBCs
Generation of cRNA.
pTLNII plasmids containing NBCe1 constructs were linearized with the restriction enzyme MluI. The linearized cDNA was then transcribed from the SP6 promoter using the SP6 transcription kit (Ambion), and the resulting cRNA was purified using the RNeasy kit (QIAGEN).
Isolation and Injection of Oocytes.
Oocytes were harvested from female Xenopus laevis frogs using an approach very similar to that previously described (Romero et al., 1998b; Bevensee et al., 2000). A small incision was made in the abdominal cavity of the frog, and oocyte-containing segments of the ovarian lobe were removed. The segments were teased apart into small pieces and digested for 1.5–2 h in sterile Ca2+-free ND96 containing 2 mg ml−1 collagenase A (Roche Applied Science). Subsequently, the dissociated oocytes were first washed in Ca2+-free ND96, and then in Ca2+-containing ND96 before healthy-looking stage V/VI oocytes were separated under a dissecting microscope (GZ6, Leica). The oocytes were incubated at 18°C in sterile ND96 containing 10 mM Na/pyruvate and 10 mg ml−1 gentamycin (Mediatech Inc.).
Oocytes were injected with 48 nl of either RNase-free H2O or a cRNA solution using a “Nanoject II” microinjector (Drummond Scientific). Injected cells were incubated at 18°C in the aforementioned oocyte media, and experiments were performed at room temperature at least 2 d after injection.
pHi and Vm Experiments.
Injected oocytes were placed in a flowthrough chamber connected to a custom-designed, dual-bank, solution delivery system. Main solution lines from two banks, which are each connected to six solution lines via a six-way rotary manifold, were directed to a two-position, Eagle four-way miniature solenoid valve with five ports (Clippard Instrument Laboratory). During experiments, solution from one bank was directed to the chamber, and solution from the other bank was directed to waste for priming purposes. Changing solutions delivered to the chamber occurred by pneumatically alternating between the two valve positions.
For simultaneous pHi and voltage recordings, microelectrodes were pulled from borosilicate glass capillaries (G200F-4, Warner Instruments) with a Brown-Flaming micropipette puller (P-80 or P-97, Sutter Instruments). For pH electrodes, pulled acid-washed capillaries were subsequently baked at 200°C and silanized with bis-(methylamino)dimethylsilane (Fluka). Electrode tips were filled with hydrogen ionophore I-cocktail B (Fluka), and then the electrodes were backfilled with a pH 7.0 solution containing (in mM): 150 NaCl, 40 KH2PO4, and 23 NaOH. pH electrodes were then connected to one channel of a high-impedance electrometer (FD223, WPI). Voltage electrodes were filled with a saturated KCl solution and connected to a second channel. The microelectrodes typically had resistances of 1–3 MΩ. The pH signal was obtained with a four-channel electrometer (Biomedical Instrumentation Laboratory, Department of Cellular and Molecular Physiology, Yale University, New Haven, CT) that subtracts the potential of the voltage electrode from the potential of the pH electrode. A miniature calomel electrode (Accumet, Fisher Scientific) filled with saturated KCl served as the reference for the voltage electrode. Data were acquired and plotted using custom-designed software written by Mr. Duncan Wong for the Boron laboratory (Department of Cellular and Molecular Physiology, Yale University).
Two-electrode, Voltage-clamp Experiments.
Voltage-sensitive and current-passing microelectrodes were pulled from borosilicate glass capillaries (G200F-4 or G83165T-4, Warner Instruments) with a micropipette puller (P-80 or P-97, Sutter Instruments or PC-10, Narishige, Tokyo, Japan). The electrodes were filled with saturated KCl and attached to the OC-725C voltage-clamp apparatus (Warner Instruments). Electrode resistances were typically 1–3 MΩ for the voltage electrodes, and 0.1–0.6 MΩ for the current electrodes. For experiments at a fixed holding potential of −60 mV, data were obtained at a filtering frequency of either 8–10 or 800 Hz with an 8-pole Bessel filter (LFP-8, Warner Instruments) and a sampling frequency of 30 Hz or 2 kHz, respectively. For the current-voltage (I-V) relationships shown, the data were obtained at a filtering frequency of 800 Hz and a sampling frequency of 2 kHz. The voltage-step protocol for the I-V plots shown included 12 sweeps in which the voltage was held at −60 mV for 60 ms, then stepped to one of 12 voltages (-200 mV to 20 mV in increments of 20 mV) for 20 ms, and finally returned to −60 mV for 20 ms before the next sweep. After obtaining each I-V plot, we confirmed that the oocyte was electrically tight by turning off the voltage clamp and verifying that the spontaneous Vm was close to the acquired reversal potential (Virkki et al., 2002). Data acquired with the ClampEx software (Axon Instruments pClamp 8.2, Molecular Devices, San Jose, CA) were digitized with a 1322A interface (Axon Instruments), and then analyzed with ClampFit software (pClamp 8.2, Axon Instruments).
Inside-out macropatch experiments.
Macropatch studies were performed on oocytes using a modification of the technique described by Hilgemann (1995). Patch pipettes were pulled from N-51-A borosilicate glass capillaries (O.D. 0.084 in. or 0.064 in., Drummond) using a PC-10 micropipette puller (Narishige). Tips gently broken to 10–12 μm were plunged into a bead of melted 8161 Corning glass (G86165T-4, Warner Instruments) fixed to 30-gauge, MF-9 platinum wire (Technical Products International Inc.), which was transiently heated by passing current. A break at the tip resulted from the wire retracting as it cooled. The process of heating the glass bead, plunging the tip, and cooling the wire was repeated until a satisfactory jagged-free tip of ∼14 μm in diameter was obtained. Pipette resistances were ∼4 MΩ.
Experiments were performed at room temperature (∼23°C) in a flowthrough chamber on the stage of an inverted microscope (DMIRB, Leica). All solutions contained low Cl− (2 mM) to minimize contaminating endogenous Cl− currents in the oocyte (Machaca et al., 2002; Weber, 2002). Immediately before experiments, pipettes were backfilled with solution containing 5% CO2/33 mM HCO3−. Access to the plasma membrane was obtained by shrinking the oocyte in a hypertonic solution (see below) and then removing the vitelline membrane with fine forceps. GΩ seals were obtained at a negative holding potential (Vp = −50 mV), and solution flow was initiated after patch excision. NBC currents were similar from patches obtained from either the animal or vegetal pole of the oocyte.
Currents were obtained using an Axopatch 200B patch-clamp amplifier (Axon Instruments). Low-pass (1 kHz, internal filter) currents were digitized with a Digidata-1322A interface (Axon Instruments) at a sampling frequency of 5 kHz. Clampex software (pClamp 8.2, Axon Instruments) was used for data acquisition and analysis. Seal stability was routinely monitored throughout the experiment by measuring both membrane capacitance (using the membrane test function of Clampex) and seal resistance. For the figures shown, current recordings were further filtered at 100 Hz (8-pole Bessel filter), and subjected to data reduction (substitute average by a factor of four) using Clampfit (pClamp 8.2, Axon Instruments). Data were exported to Microsoft Excel 2002 for analysis and Origin 7.5 (OriginLab Corporation) for graphing.
Solutions
Two-electrode, Voltage-clamp Experiments.
The standard ND96 solution (pH 7.5) contained (in mM): 96 NaCl, 2 KCl, 1 MgCl2, 1.8 CaCl2, 5 HEPES, 2.5 NaOH. In the standard 5% CO2/33 mM HCO3− solution, 33 mM NaCl was replaced with an equimolar amount of NaHCO3, and the solution was equilibrated with 5% CO2/95% O2 to pH 7.5. In the low Cl− (7.6 mM)-containing ND96 and 5% CO2/33 mM HCO3− solutions used in the bicarbonate and voltage dependence experiments, the NaCl was replaced with equimolar amounts of sodium gluconate. 7.6 mM Cl− was kept constant in solutions containing different amounts of HCO3− by replacing sodium gluconate with equimolar amounts of NaHCO3. pH of the different HCO3−-containing solutions was maintained at 7.5 by equilibrating the solutions with appropriate mixtures of CO2/O2. In the Na+ dependence experiments, equimolar Na+ was replaced with NMDG+.
Inside-out Macropatch Experiments.
The hyperosmotic solution used to remove the vitelline membrane contained (in mM): 220 NMDG+, 220 aspartic acid, 2 MgCl2, 10 EGTA, and 10 HEPES. The solution was titrated to pH 7.2 with NMDG+.
For inside-out macropatch experiments, the 2 mM-Cl−, ND96 solution contained (in mM) 96 sodium cyclamate, 2 mM KOH, 2 mM cyclamic acid, 1 mM MgCl2, 1.8 mM calcium cyclamate, 5 mM HEPES, and was adjusted to pH 7.5 with NaOH. In the 2-mM Cl− solution containing 5% CO2/33 mM HCO3−, 33 mM sodium cyclamate was replaced with an equimolar amount of NaHCO3, and the solution was equilibrated with 5% CO2/95% O2 to pH 7.5. For all experiments, the 2 mM Cl− solution containing 5% CO2/33 mM HCO3− served as the patch pipette solution.
All chemicals were obtained from Sigma-Aldrich unless otherwise specified. According to packaging information, 4,4′-diisothiocyanatostilbene-2,2′ disulfonate (DIDS) was at least 80% pure, a value factored into making DIDS-containing solutions.
Expression Studies on NBCs
Immunoblotting.
Similar to previously described (Schmitt et al., 1999; Bevensee et al., 2000), each oocyte was homogenized in a buffer containing protease inhibitors, and the suspension was centrifuged to remove cell debris and nuclei. Proteins were separated by SDS-7.5% PAGE (Ready Gels Precast Gels, Bio-Rad Laboratories) and transferred to an Immobilin-P PVDF membrane (Millipore). The membrane was probed for 1–2 h with a 1:200 dilution of a rabbit polyclonal antibody (Rab3A) to the amino terminus of NBCe1 (Schmitt et al., 1999), and then for 1 h with a 1:10K dilution of the secondary antibody, goat αrabbit-IgG coupled to horseradish peroxidase (HRP) (Jackson ImmunoResearch Laboratories). Bound HRP was detected by chemiluminescence (SuperSignal, Pierce Chemical Co.) before being exposed to HXR Film (Hawkins X-Ray Supply).
Single-oocyte Chemiluminescence.
To evaluate the expression of NBCs at the oocyte plasma membrane, we used the single-oocyte chemiluminescence (SOC) technique pioneered by the Jan lab (Zerangue et al., 1999; Margeta-Mitrovic et al., 2000) and used by others (Yoo et al., 2003) to quantitate a hemagglutinin (HA)–tagged protein expressed at the cell surface. This technique employs enzyme amplification with a chemiluminescence substrate and sensitive, linear detection with a luminometer. As reported by Yoo et al. (2003), there is a linear relationship between surface expression detected by SOC and functional activity of the K+ channel, ROMK (Kir 1.1).
The following protocol with sterile ND96 solutions was performed on injected oocytes incubated at 4°C. Oocytes were fixed with 4% paraformaldehyde in ND96 for 15 min, rinsed 3×5 min with equal volumes of ND96, and then incubated for 30 min (or overnight) in a 1% BSA-ND96 blocking solution (used in subsequent antibody incubation steps). Fixing the oocytes with paraformaldehyde does not appreciably increase their permeability to the HA antibody. As shown in Results, there was no detectable labeling of truncated, HA-tagged NBCe1 proteins that fail to traffic to the plasma membrane, but were clearly present in a microsomal fraction based on immunoblot analyses. Fixed oocytes were incubated for 1 h in a 1:100 dilution of the rat monoclonal α-hemagglutinin antibody (Roche), and then for 1 h in a 1:400 dilution of the secondary antibody, goat α-rat IgG-HRP (Jackson ImmunoResearch Laboratories). For chemiluminescence readings, each oocyte was transferred to an eppendorf tube, and the transferred solution was replaced with 50 μl SuperSignal Elisa Femto substrate (Pierce Chemical Co.). The eppendorf tube was placed in a luminometer (TD-20/20, Turner Designs) and luminescence was measured 15 s later.
Statistics
Means between groups of data were compared using one-factor analysis of variance (ANOVA), as well as paired or unpaired forms of the Student's t test (Microsoft Excel 2002). P < 0.05 is considered significant. Rates of pHi recoveries were determined by linear fits to pHi vs. time data using a least-squares method. With Origin 7.5 software (OriginLab), Na+ dependence data were fit using Michaelis-Menten enzyme kinetics, and HCO3−-dependence data were fit using a modified form of the Michaelis-Menten equation that includes a linear component (Grichtchenko et al., 2000). Comparisons of fits were performed using Origin 7.5 software (OriginLab).
Results
Functional Expression of Rat Brain NBCe1-B in Oocytes
At the time the cDNA encoding rat brain NBCe1-C was cloned by homology and the protein characterized as an electrogenic NBC when expressed in oocytes, rat brain NBCe1-B was also identified (Bevensee et al., 2000). The amino acid sequence of rat brain NBCe1-B is 96% identical to that of human heart NBCe1-B, which has been characterized as a stilbene-sensitive, electrogenic NBC when expressed in oocytes (Choi et al., 1999). In the present manuscript, we expressed the rat brain version of NBCe1-B in oocytes and evaluated NBC function using pH- and voltage-sensitive microelectrodes.
At the onset of the experiment shown in Fig. 2 A, an oocyte injected with rat brain NBCe1-B and bathed in a nominally CO2/HCO3−-free, HEPES-buffered solution (ND96, pH 7.5) had a resting pHi of ∼7.32 (top trace) and a Vm of −55 mV (bottom trace). Exposing the oocyte to a solution containing 1.5% CO2/10 mM HCO3− (pH 7.5) elicited an initial decrease in pHi (ab) due to CO2 entry and subsequent formation of H+ in the cell. The pHi then increased (bc) due to NBC-mediated HCO3− transport into the oocyte. The mean pHi recovery rate at a pHi of 7.12 ± 0.02 (after the initial CO2-induced acidification) was 7.1 ± 0.8 ×10−5 pH units s−1 (n = 6). This pHi recovery was blocked (cd) by 200 μM DIDS, an inhibitor of bicarbonate transporters. As shown in Table I, stilbenes reduced the rate of pHi recovery under different CO2/HCO3− conditions by a mean of 76 ± 7.5% (range: 58–100%). Inhibition of the pHi recovery was at least partially reversible (de) when DIDS was removed. Returning the oocyte to ND96 caused pHi to increase rapidly (ef) due to the net conversion of intracellular H+ and HCO3− to H2O and CO2, which exited the cell. The final resting pHi was higher than the initial value, an observation consistent with net NBC-mediated HCO3− influx during segment be.
The voltage changes in the NBC-expressing oocyte exposed to CO2/HCO3− and then DIDS (Fig. 2 A, bottom trace) demonstrate that NBCe1-B is electrogenic. Applying the CO2/HCO3− solution elicited an ∼17-mV hyperpolarization (a′b′) that is consistent with NBC transporting more bicarbonate than sodium into the oocyte. In the same six experiments summarized above, the mean hyperpolarization elicited by 1.5% CO2/10 mM HCO3− was 25 ± 8 mV. Also consistent with electrogenic Na/HCO3− cotransport activity was the ∼10-mV depolarization seen when the oocyte in CO2/HCO3− was exposed to 200 μM DIDS (c′).
The experimental maneuvers shown in Fig. 2 A had markedly different effects on pHi and Vm of the H2O-injected control oocyte shown in Fig. 2 B. As shown in the top trace, the 1.5% CO2/10 mM HCO3− solution elicited the initial pHi decrease (ab), but no subsequent pHi recovery (bc). At a mean pHi of 7.08 ± 0.07, the mean pHi recovery rate of 2.1 ± 1.2 ×10−5 pH units s−1 (n = 4) was 3.4-fold less (P < 0.01) than the rate seen in NBCe1-B-expressing oocytes. Furthermore, 200 μM DIDS in the presence of CO2/HCO3− had little effect on the low, sustained pHi (cd). Returning the oocyte to ND96 caused pHi to increase and return to approximately the initial resting value (ef). The voltage changes in the control oocyte were equally unimpressive (bottom trace). CO2/HCO3− had no effect on Vm (a′b′), and DIDS only caused a small hyperpolarization (c′), consistent with inhibition of an endogenous Cl− conductance. In the same four experiments summarized above, the mean HCO3−-induced hyperpolarization of 1.5 ± 0.3 mV was 17-fold less than the hyperpolarization seen in NBCe1-B–expressing oocytes.
To increase the activity of NBCe1-B, we performed additional pHi/Vm experiments using solutions containing 33 mM HCO3− and equilibrated with 5% CO2 to maintain pH at 7.5. The mean pHi-recovery rate at a pHi of 6.87 ± 0.03 was 12.0 ± 2.5 ×10−5 pH units s−1 (n = 11). At a similar pHi (6.84 ± 0.03) in H2O-injected oocytes, the mean pHi recovery rate of 4.3 ± 1.9 ×10−5 pH units s−1 (n = 6) was 2.8-fold less (P = 0.01). In these same experiments, the mean hyperpolarization elicited by 5% CO2/33 mM HCO3− was 53 ± 4.8 mV in NBC-expressing oocytes, and only 0.5 ± 0.5 mV in H2O-injected oocytes. In summary, rat brain NBCe1-B expressed in oocytes is electrogenic, stilbene-sensitive, and activated by CO2/HCO3−. As described below in voltage-clamp experiments, the transporter is also Na+ dependent. All these characteristics are hallmarks of an electrogenic Na/HCO3 cotransporter (see Romero et al., 2004).
Voltage Dependencies of the NBCe1 Variants
To study the electrogenicity of the NBCe1 variants in more detail, we used the two-electrode, voltage-clamp technique to examine current–voltage (I-V) relationships of all three NBCe1 variants. In each experiment, I-V relationships were obtained from oocytes first bathed in a low Cl−, ND96 solution, and then after 10 min in 5% CO2/33 mM HCO3− (to allow for intracellular equilibration of the physiological buffer). In some experiments, I-V plots were subsequently obtained after 2 min in the CO2/HCO3− solution containing 200 μM DIDS. The HCO3−-dependent I-V plot for an NBC is the difference between the I-V plots obtained in the presence and absence of CO2/HCO3−, after subtracting the corresponding HCO3−-dependent I-V plot obtained from batch-matched H2O-injected eggs.
An experiment performed on an oocyte expressing NBCe1-C is shown in Fig. 3 A. NBC currents obtained at potentials from −200 to +20 mV were larger in the presence of 5% CO2/33 mM HCO3− (squares) than in ND96 (diamonds). The larger currents were particularly evident at voltages more positive than −80 mV. In the presence of DIDS, the I-V plot (triangles) reverted back to that seen with ND96. DIDS-sensitive HCO3− currents were not observed in an H2O-injected control oocyte (Fig. 3 B). I-V plots looked similar for the control oocyte exposed to ND96 and CO2/HCO3− ± DIDS.
In Fig. 4 A, the average I-V relationships are shown for NBCe1-A (closed diamonds), -B (open squares), and -C (closed triangles). There are four noteworthy observations. First, the currents for the A variant are the largest. Second, all three I-V plots display a slight outward rectification. Third, the I-V plots for the B and C variants are nearly identical. The final observation is that the Erev is ∼−85 mV for A (Fig. 4 A), but markedly more negative for B (∼−155 mV) and C (∼−170 mV) (Fig. 4 A, boxed inset). Although the more negative Erev values are consistent with different Na:HCO3− transport stoichiometries, a more likely explanation is different transmembrane Na+ and HCO3− gradients established by the NBCs. For example, compared to A, the less active B and C variants generate a smaller increase in [Na+] and [HCO3−] on the cytosolic side of the membrane, and therefore a more negative Vm for B and C is required to drive transport in the reverse direction out of the cell. We tested this possible explanation by injecting a smaller amount of NBCe1-A cRNA to reduce the magnitude of the whole-cell A current. Indeed, oocytes exhibiting smaller A currents displayed a negative shift in Erev (Fig. 4 B, open diamonds). For comparison, the I-V relationship for NBCe1-A shown in A is replotted in Fig. 4 B (closed diamonds).
To minimize the influence of different transmembrane gradients, we compared the voltage dependencies of the three variants under similar transporter-mediated whole-cell currents. NBCe1-B and -C currents were increased by injecting oocytes with NBC cRNAs containing a modified Kozak sequence, whereas NBCe1-A currents were decreased by injecting oocytes with less cRNA. As shown in Fig. 4 C, the voltage dependencies of the three variants exhibiting comparable NBC-mediated currents are very similar.
Expression of the NBCe1 Variants at the Plasma Membrane of the Oocyte
Activity of Hemagglutinin-tagged NBCe1 Variants.
The higher activity of NBCe1-A compared to the B and C variants shown in Fig. 4 A could be due to differences in plasma membrane expression. We therefore used the SOC technique with HA-tagged NBCe1 constructs to assess surface expression. We first tested the function of the tagged constructs expressed in oocytes. As shown in Fig. 4 B, there is no difference between the voltage dependencies of the untagged NBCe1-A (closed diamonds) redrawn from Fig. 4 A, and the HA-tagged transporter (NBCe1-AHA, closed squares). We performed additional studies to compare the function of tagged NBCe1-C. In pHi/Vm and voltage-clamp studies in which oocytes expressing either tagged or untagged NBCe1-C were exposed to 5% CO2/33 mM HCO3−, both groups displayed similar NBC-mediated hyperpolarizations, pHi recovery rates following the initial CO2-induced acidification, HCO3−-induced outward currents under voltage-clamp conditions, and I-V relationships. In conclusion, the introduced HA epitope does not alter NBCe1 activity.
Comparing Function and Surface Expression of NBCe1 Variants.
We expressed HA-tagged NBCe1-A and NBCe1-C in oocytes, and then evaluated both transporter function with the two-electrode, voltage-clamp technique, and expression with the SOC technique. In our functional assay, we monitored NBC-mediated outward currents in oocytes voltage clamped at −60 mV and exposed to 5% CO2/33 mM HCO3−. In contrast to an oocyte injected with H2O, an oocyte injected with either the A or C variant displayed an outward current when exposed to 5% CO2/33 mM HCO3− (Fig. 5 A). The A and C currents differed in the following three ways. First, the A-mediated current was ∼4.5-fold larger than the C-mediated current. Second, the A-mediated current peaked faster than the C-mediated current. Finally, in contrast to the C-mediated current, the A-mediated current decayed after its peak. The current decay is consistent with reduced transport in response to the buildup of substrate or a pHi increase at the inner surface of the oocyte membrane (and opposite changes at the outer surface of the membrane). According to subsequent data presented, the more active the transporter (as judged by the magnitude of the peak HCO3−-induced current), the faster the decay. As summarized in Fig. 5 B on two batches of day-matched oocytes, the mean H2O-subtracted, peak CO2/HCO3−-induced current was 4.3-fold larger in oocytes expressing NBCe1-A (874 ± 82 nA, n = 8) compared to NBCe1-C (201 ± 4 nA, n = 7). The larger NBC-mediated current seen with NBCe1-A compared to C is consistent with the I-V plot data shown in Fig. 4 A.
We used the same two batches of oocytes to examine plasma membrane NBC expression with the SOC technique. For each batch, the luminescence for each oocyte was normalized to the mean luminescence of oocytes expressing NBCe1-A. As expected, the mean normalized luminescence (Norm. Lum.) for A- or C- expressing oocytes was considerably higher than the mean value for H2O-injected oocytes (Fig. 5 C). These readings are specific for the anti-HA antibody because in a separate experiment (not depicted), luminescence was not observed when the anti-HA antibody was preabsorbed with an equal amount (1 mg ml−1) of the HA peptide. As shown by the second and third bars in Fig. 5 C, the mean Norm. Lum. was the same for the two NBCe1 variants (P > 0.3).
Similar functional and expression studies were performed on the HA-tagged NBCe1-B variant. The mean CO2/HCO3−-induced current was threefold smaller in oocytes expressing NBCe1-B (258 ± 10 nA, n = 6) compared to NBCe1-A (787 ± 57 nA, n = 7). According to SOC analysis, the mean Norm. Lum. was identical for oocytes expressing the B variant (1.01 ± 0.07, n = 5) and A variant (1.00 ± 0.09, n = 5). Therefore, the lower activity of B or C vs. A cannot be explained by lower plasma membrane expression.
External Ion Dependencies of the NBCe1 Variants
The increased activity of NBCe1-A compared to the B and C variants may be due to a higher affinity for HCO3− and/or Na+. In the following two sections, we compare –for each NBCe1 variant– the external HCO3− and Na+ dependencies.
External HCO3− Dependencies.
We used an approach introduced by Grichtchenko et al. (2000) to determine the HCO3− dependencies of NBCe1s expressed in oocytes voltage clamped at −60 mV. NBC-mediated outward currents were recorded when oocytes were exposed to either 5% CO2/33 mM HCO3− (standard HCO3−), or solutions containing different HCO3− concentrations and equilibrated with appropriate CO2/O2 mixtures to maintain a pH of 7.5. All solutions contained 7.6 mM Cl− as described above.
Results from an oocyte expressing NBCe1-C are shown in Fig. 6 A. The oocyte was initially bathed in the normal Cl− (103.6 mM), HEPES-buffered solution and voltage clamped at −60 mV. Exposing the oocyte to a solution containing 7.6 mM Cl− generated a slow outward current (ab), possibly due to a small gluconate conductance. Exposing the oocyte to solutions containing different [HCO3−]s elicited rapid NBC-mediated outward currents (c-i). Currents were larger at progressively higher [HCO3−]s (h vs. d vs. e vs. f). Each non-33 mM HCO3− exposure was flanked by standard HCO3− exposures. A similar experiment performed on an H2O-injected oocyte is shown in Fig. 6 B. The CO2/HCO3− solutions elicited only slow inward currents (c-i).
Using data from experiments similar to those shown in Fig. 6 A, we plotted the HCO3− dependencies of NBCe1-B and -C. To correct for current drifts during experiments, as well as any differences in NBC expression among oocytes, we normalized the current elicited by each HCO3− exposure to the mean current elicited by the flanking standard HCO3− exposures. Corresponding currents obtained from H2O-injected oocytes were subtracted before normalization. The external HCO3− dependencies are shown for NBCe1-B (Fig. 6 C) and NBCe1-C (Fig. 6 D). The data for both variants are well fit using a model that combines a Michaelis-Menten mechanism with a linear component, as has been reported for rat kidney NBCe1-A (Grichtchenko et al., 2000). The external HCO3− dependencies of the B and C variants are nearly identical. For NBCe1-B, the apparent KM for HCO3− is 4.68 ± 0.47 mM and the normalized Vmax is 0.92 ± 0.03 (n = 24 from six oocytes). For NBCe1-C, the apparent KM for HCO3− is 4.31 ± 0.46 mM and the normalized Vmax is 0.88 ± 0.03 (n = 24 from six oocytes). These values for the B and C variants are very similar to our values of apparent KM (6.90 ± 3.04 mM) and normalized Vmax (0.96 ± 0.12) (n = 10 from three oocytes) obtained for the A variant. Furthermore, our values for NBCe1-A are nearly identical to the values of apparent KM (6.5 mM) and normalized Vmax (0.97) previously reported for NBCe1-A (Grichtchenko et al., 2000).
We also examined the bicarbonate dependence of an NBCe1-A/C chimera that contains the 41 amino-terminal residues of the A variant and the 61 carboxy-terminal residues of the C variant. In these studies, the A/C chimera exhibited a mean outward current (1240 ± 103 nA, n = 5) similar to that of wild-type NBCe1-A (1096 ± 296 nA, n = 3) when the oocytes were exposed to 5% CO2/33 mM HCO3−. This chimera also had a similar apparent KM for HCO3− of 5.33 ± 0.54 mM and normalized Vmax of 1.00 ± 0.04 (n = 17 from five oocytes). In conclusion, the external HCO3− dependencies are no different (P ≥ 0.89) for all three NBCe1 variants and the NBCe1-A/C chimera. Therefore, the larger NBC-mediated current seen with NBCe1-A compared to the B and C variants is not due to a higher HCO3− affinity.
External Na+ Dependencies.
We used a similar approach to the one described above to determine the external Na+ dependencies of the NBCe1 variants. In our assay, we examined the magnitudes of the NBC-mediated outward currents elicited by 5% CO2/33 mM HCO3− solutions containing various [Na+]s. Our experimental protocol minimized changes in intracellular [Na+] ([Na+]i) by minimizing the time oocytes spent in reduced [Na+]s.
Results from an oocyte expressing NBCe1-C are shown in Fig. 7 A. The voltage-clamped oocyte was initially bathed in ND96 and then exposed to 5% CO2/33 mM HCO3− solutions containing different amounts of Na+. These HCO3− exposures elicited rapid NBC-mediated outward currents (a-i), which were smaller at progressively lower concentrations of Na+ (b vs. d vs. f vs. h). Using an approach similar to that described above for our bicarbonate-dependence experiments, each low-Na+, HCO3− exposure was flanked by full-Na+ (98.5 mM) HCO3− exposures. A similar experiment on an H2O-injected oocyte subjected to the same experimental protocol is shown in Fig. 7 B. The CO2/HCO3− solutions of different Na+ concentrations elicited only small outward currents.
One noteworthy observation (unpublished data) with oocytes expressing NBCe1-A was the presence of a Na+-independent, outward current elicited by CO2/HCO3− that was not seen in H2O-injected oocytes. In the absence of external Na+, the 33 mM HCO3− solution elicited a mean outward current in A-expressing oocytes (153 ± 48 nA, n = 4) that was 13-fold larger (P = 0.03) than that seen in H2O-injected oocytes (12 ± 3 nA, n = 4), and 18 ± 5% of the total HCO3− current in “full” 98.5 mM Na+. This Na+-independent HCO3− current, which was not seen in oocytes expressing NBCe1-B (12 ± 3 nA, n = 6) or NBCe1-C (19 ± 3 nA, n = 10), may be due to transporter slippage as described for H+ pumps, neurotransmitter transporters, and metal ion transporters (Nelson et al., 2002). Alternatively, a small NBC-mediated conductance (e.g., HCO3−) that is independent of Na+-coupled transport may be responsible. A Na+ conductance has been described for the electroneutral NBC, NBCn1 (Choi et al., 2000).
We used data from experiments similar to that shown in Fig. 7 A to determine the Na+ dependencies of NBCe1-A, -B, and -C. For each experiment, we normalized the current elicited by 5% CO2/33 mM HCO3− solutions containing different [Na+]s to the mean current elicited by a flanking CO2/HCO3− solution containing full, 98.5 mM Na+. HCO3− currents obtained in the absence of Na+ were subtracted before normalization. We plot the external Na+ dependencies for the two most dissimilar variants: NBCe1-A in Fig. 7 C and NBCe1-C in Fig. 7 D. The apparent Km and normalized Vmax values are similar for both the A and C variants. The apparent KM values are 20.6 ± 1.7 mM for NBCe1-A (n = 33 from five oocytes) and 31.7 ± 2.6 mM for NBCe1-C (n = 71 from 12 oocytes). Sciortino and Romero (1999) reported a similar apparent KM of 30 mM for NBCe1-A. Our corresponding normalized Vmax values are 1.21 ± 0.03 for NBCe1-A and 1.31 ± 0.04 for NBCe1-C. In additional studies, we also examined the external Na+ dependence of NBCe1-B and obtained an apparent KM of 35.5 ± 3.7 mM and normalized Vmax of 1.35 ± 0.05 (n = 23 from five oocytes). In conclusion, all three NBCe1 variants have similar external Na+ dependencies.
Role of the Amino Termini on the Function and Expression of NBCe1 Variants in Oocytes
According to the aforementioned HCO3− and Na+ dependence data, the higher NBC current seen with NBCe1-A vs. -B and -C (Figs. 4 and 5) is not due to higher substrate affinities. Therefore, the A variant must have a higher transport velocity compared to the B and C variants. As shown in Fig. 1, the differences in amino acid sequence among these three variants reside only at the amino and/or carboxy termini. The amino terminus is likely responsible for differences in NBC activity because the B and C variants (with low activity) share the same amino terminus, which is different than the unique amino terminus of the A variant (with high activity). Furthermore, the A/C chimera containing the unique carboxy terminus of C attached to the A variant displays NBC currents similar to those of the wild-type A variant.
Activity and Surface Expression of NBCe1 Variants Truncated before the First Transmembrane Domain.
We used the HCO3−-induced, outward-current assay shown in Fig. 5 A to examine the function of NBCe1 constructs with amino-terminal truncations. The three experimental traces shown in Fig. 8 A are from voltage-clamped oocytes injected with cRNA encoding wild-type NBCe1-C (CWT), NBCe1-C missing the first 426 residues before the first predicted transmembrane domain (CΔN426), or NBCe1-C missing the first 213 residues (CΔN213). Oocytes exposed to a solution containing 5% CO2/33 mM HCO3− elicited an outward current of ∼250 nA in the oocyte expressing CWT, but little/no current in the oocyte expressing either CΔN426 or CΔN213.
From experiments similar to those shown in A, the mean CO2/HCO3−-induced currents from oocytes expressing CΔN426 or CΔN213 were similar to those seen in day-matched, H2O-injected oocytes, and more than 90% less than those in day-matched, CWT-expressing eggs (Fig. 8 B). (Because we often examined several constructs on a given day, we necessarily used current and SOC data obtained from oocytes injected with H2O and/or wild-type NBC constructs in more than one analysis.) A similar loss of NBC activity was observed with the homologous truncation of NBCe1-A before the first transmembrane domain, AΔN382 (n = 2, not depicted).
SOC analysis was performed on the same two oocyte batches, and luminescence for each oocyte was normalized to the mean luminescence of oocytes expressing wild-type CWT. Compared with the mean Norm. Lum. for CWT, the mean values were actually 50% higher (P = 0.002) for CΔN426 and similar (P = 0.14) for CΔN213 (Fig. 8 C). In summary, the lost NBC activity of CΔN426 and CΔN213 is not due to the absence of expression at the plasma membrane.
In further structure–function studies, we examined the effects on transporter activity and plasma membrane expression of removing the NH2-terminal 43 residues of NBCe1-A (AΔN43) and 87 residues of NBCe1-C (CΔN87). Because the amino terminus of the B and C variants is identical, studies on CΔN87 provide NH2-terminal information on the B variant.
Activity and Surface Expression of NBCe1-A Lacking its Unique Amino Terminus.
At a holding potential of −60 mV, an oocyte expressing AΔN43 (Fig. 9 A, left) displayed an HCO3−-induced outward current that was ∼50% smaller than the current seen in the oocyte expressing wild-type NBCe1-A (AWT). Therefore, removing the amino-terminal 43 residues of the A variant decreases transporter activity. The summary data of HCO3−-mediated outward currents are shown in Fig. 9 B (left). From five batches of oocytes, the mean HCO3−-induced current for AΔN43 (443 ± 31 nA, n = 18) was 55% smaller (P < 0.001) than the mean current for AWT (990 ± 81 nA, n = 15). The mean HCO3−-induced current from batch-matched, H2O-injected oocytes was only 0.8% of the mean current from AWT-expressing oocytes.
In four of these five batches in which SOC data were obtained, the mean Norm. Lum. was 10% lower (marginal significance; P = 0.05) for oocytes injected with AΔN43 compared to those injected with AWT (Fig. 9 C, left). A 10% lower surface expression would not explain the 55% decrease in transporter activity of AΔN43 compared to AWT. In addition, in one of these four batches in which the surface expression of AΔN43 (n = 5) was 96 ± 14% of AWT (n = 5), the mean NBC- mediated current for AΔN43 (656 ± 105 nA, n = 3) was 46% smaller (P = 0.04) than for AWT (1220 ± 192 nA, n = 3). Therefore, the lower activity of AΔN43 compared to AWT cannot be explained by a difference in surface expression.
Activity and Surface Expression of NBCe1-C Lacking the Different Amino Terminus.
Different results were obtained with the homologous NH2-terminal truncation of NBCe1-C (CΔN87). At a holding potential of −60 mV, an oocyte expressing CΔN87 (Fig. 9 A, right) displayed an HCO3−-induced outward current that was 3.4-fold larger than the current seen in the oocyte expressing CWT. Therefore, removing the amino-terminal 87 residues of CWT increases transporter activity. The summary data of HCO3−-mediated outward currents from experiments similar to those shown in Fig. 9 A (right) are shown in B (right). From seven batches of oocytes, the mean HCO3−-induced current for CΔN87 (654 ± 48 nA, n = 23) was 2.7-fold larger (P < 0.001) than the mean current for CWT (242 ± 12 nA, n = 19). As expected from Fig. 5 data, the mean current for CWT was ∼25% of the mean current for AWT.
In five of the seven batches above in which SOC data were obtained, the mean Norm. Lum. was 33% higher (P < 0.001) for oocytes injected with CΔN87 compared to those injected with CWT (Fig. 9 C, right). A 33% increase in surface expression would not explain the 170% increase in transporter activity of CΔN87 compared with CWT. Moreover, in two of these five batches in which the surface expression of CΔN87 (n = 9) was 102 ± 4% of CWT (n = 10), the mean NBC-mediated current for CΔN87 (713 ± 137 nA, n = 6) was also 2.7- fold larger (P = 0.01) than for CWT (264 ± 20 nA, n = 5). Therefore, the higher activity of CΔN87 compared to CWT cannot be explained by a difference in surface expression. Data from Fig. 9 are consistent with the amino-terminal residues of NBCe1-A stimulating transporter activity, and the amino-terminal residues of NBCe1-B and C inhibiting transporter activity.
If the amino termini of the NBCe1 variants are solely responsible for the higher transporter activity seen with A compared to B and C, then the activities of AΔN43 and CΔN87 should be identical. However, as shown in Fig. 9, the mean CΔN87 current (654 nA) is ∼50% larger (P < 0.001) than the mean AΔN43 current (443 nA). The larger mean CΔN87 current could be explained by the slightly higher surface expression of CΔN87 vs. AΔN43 as revealed in the SOC analysis shown in Fig. 9 C. However, a difference in surface expression does not appear to be the only explanation according to current–voltage (I-V) analyses (Fig. 10).
I-V Relationships of Wild-type and Mutant NBCe1-A and C Constructs.
We next examined the I-V relationships of AWT, CWT, AΔN43, and CΔN87. A representative experiment on an oocyte expressing CΔN87 is shown in Fig. 10 A. I-V relationships were obtained from the oocyte first bathed in ND96 (diamonds), then after 1 min (circles) and 10 min (squares) in 5% CO2/33 mM HCO3−, and finally after 2 min in the physiological buffer containing 200 μM DIDS (triangles). The larger currents obtained with the oocyte in CO2/HCO3− were inhibited by the presence of 200 μM DIDS. The left-shifted I-V plot obtained after a 1 vs. 10 min exposure to CO2/HCO3− was due to the incomplete equilibration of CO2/HCO3− across the membrane, and consequently, a larger extracellular- to-intracellular HCO3− gradient.
From experiments similar to that shown in Fig. 10 A, mean HCO3−-dependent I-V plots were obtained from oocytes injected with wild-type NBCe1-A and -C, and the truncated constructs AΔN43 and CΔN87 (Fig. 10 B). The data were from day-matched experiments, and mean I-V data from H2O-injected oocytes were subtracted. Similar to the Fig. 9 results obtained at a fixed potential of −60 mV, the voltage-dependent AΔN43 currents (Fig. 10 B, open diamonds) were smaller than the corresponding wild-type A currents (closed diamonds), and the CΔN87 currents (open triangles) were larger than the corresponding wild-type C currents (closed triangles) at all holding potentials from −200 to +20 mV (except at −80 mV). For each construct, the NBC conductance (gNBC) was determined from the slope of a linear fit to either the mean inward currents (Vh from −200 to −100 mV), or the mean outward currents (Vh from −80 to +20 mV) plotted in Fig. 10 B. For the inward currents, the gNBC of AΔN43 was 64% lower than that of AWT, and the gNBC of CΔN87 was 2.3-fold higher than that of CWT. Expectedly, the gNBC of CWT was only 28% of the gNBC of AWT. Similar gNBC results were obtained for the outward currents.
CΔN87 currents were also found to be consistently larger than the AΔN43 currents. The mean CΔN87 current was approximately twofold larger than the mean AΔN43 current at −200 and +20 mV. Furthermore, gNBC of CΔN87 was 1.8-fold higher than the gNBC of AΔN43. These data are consistent with the unique carboxy terminus of the C variant vs. the carboxy terminus of the A/B variant contributing to higher NBC activity in the absence of the amino terminus.
Role of the Carboxy Termini on the Function and Expression of NBCe1 Variants in Oocytes
Activity and Surface Expression of NBCe1 Truncated after the Last Transmembrane Domain.
To determine the importance of the cytosolic carboxy terminus of NBCe1, we removed the 97 residues after the last predicted transmembrane domain of NBCe1-C (CΔC97). This construct is identical to the homologous truncation of NBCe1-B. After expressing CΔC97 in oocytes, we used pH-sensitive microelectrodes and the voltage-clamp technique to measure simultaneously pHi and membrane current of oocytes voltage clamped at −60 mV.
Records from simultaneous pHi and voltage-clamp experiments on oocytes expressing wild-type NBCe1-C (CWT) and CΔC97 are shown in Fig. 11. After stable pHi and Vm values were obtained, oocytes were voltage clamped at −60 mV and exposed to 5% CO2/33 mM HCO3−. As expected from Fig. 2 and Fig. 5 data, the CWT-expressing oocyte (panel A) exhibited an instantaneous HCO3−-stimulated outward current (point a′, bottom trace), as well as a pHi recovery following the initial CO2-induced acidification (abc, top trace). Removing external Na+ reversed the outward current (c′), and blocked the pHi recovery (cd). In contrast, the CΔC97-expressing oocyte (panel B) exposed to CO2/HCO3− exhibited neither a HCO3−-stimulated outward current (point a′, bottom trace), nor any appreciable pHi recovery after the CO2-induced acidification (abc, upper trace). Stars in the voltage traces represent time points during the experiment when I-V plots were obtained (unpublished data).
The summary data from experiments similar to those shown in Fig. 11 (A and B) are shown in Fig. 12 (A and B). Data were obtained from two batches of oocytes. The mean rate of pHi recovery (dpHi/dt) during the segment bc pHi recovery in oocytes expressing CΔC97 was fourfold less than the pHi recovery rate in oocytes expressing CWT, and similar (P > 0.14) to the rate in oocytes injected with H2O (panel A). The mean CO2/HCO3−-induced outward current (a′) was 14-fold smaller in oocytes expressing CΔC97 compared to CWT, and no different (P > 0.18) than the mean current in H2O-injected eggs (panel B).
To identify the explanation for the lost transporter activity of CΔC97, we used immunoblotting and SOC techniques to examine protein expression. According to the immunoblot data shown in Fig. 12 C, a polyclonal antibody to the amino terminus of NBCe1 recognized the expected-size protein from total microsomal protein isolated from an oocyte injected with either CΔC97 (lane 1) or CWT (lane 3). No labeling was observed in protein from an H2O-injected oocyte (lane 2). Thus, the injected CΔC97 cRNA is translated and the protein is expressed in a microsomal fraction. However, according to the SOC data (normalized to the mean luminescence of CWT-expressing oocytes), the normalized mean luminescence of CΔC97-injected oocytes is approximately threefold lower than that of CWT, and even slightly lower than that of H2O-injected oocytes (P < 0.01). Therefore, CΔC97 is not expressed at the plasma membrane of oocytes, a result that was confirmed by immunocytochemistry with the anti-HA antibody (unpublished data). In summary, the loss of CΔC97 activity in oocytes is due to the absence of CΔC97 protein at the plasma membrane.
Activity and Surface Expression of NBCe1 Variants Truncated at the Different Carboxy Termini.
To explore the contribution of the different carboxy termini of the NBCe1 variants to plasma membrane expression, and possibly function, we performed additional studies comparing the expression and function of wild-type and truncated variants. More specifically, we created two additional constructs: NBCe1-A truncated 46 residues from the carboxy terminus (AΔC46) and NBCe1-C truncated 61 residues from the carboxy terminus (CΔC61). The CΔC61 construct is identical to the homologous truncation of NBCe1-B.
NBC-mediated outward currents in response to CO2/HCO3− are shown in Fig. 13 A for two voltage-clamped oocytes; one expressing AWT and the other expressing AΔC46. The HCO3−-induced outward current of ∼0.43 μA in the AΔC46-expressing oocyte was 71% smaller than the current of ∼1.45 μA in the AWT-expressing oocyte. The summary data from similar experiments performed on Awt and AΔC46, as well as Cwt and CΔC61 are plotted in Fig. 13 B. The mean currents obtained from H2O-injected oocytes were subtracted from the corresponding currents seen in the NBC-injected oocytes. The mean NBC-mediated currents of the truncated NBCs were ∼30% of the corresponding wild-type NBCs. As shown in Fig. 13 C, both AWT and AΔC46 are expressed in a microsomal membrane fraction. Similar results were obtained with CWT and CΔC61 (not depicted). However, according to summary SOC data (Fig. 13 D) from the same batches of oocytes, AΔC46 and CΔC61 are poorly expressed at the plasma membrane compared to the corresponding wild-type NBCs. Therefore, the reduced activity of AΔC46 and CΔC61 in oocytes is predominantly due to low surface expression.
Transporter Activity of NBCe1 Variants in the Inside-out Macropatch
To examine the contribution of the amino termini to NBCe1 transporter activity in the absence of cytosolic factors, we next examined the function of CWT, AWT, CΔN87, and AΔN43 in excised macropatches in the inside-out configuration from oocytes. In our experimental protocol (Fig. 14 A), we excised patches into a low-Cl− (2 mM), ND96 solution and then activated NBCs by exposing the patch to the low-Cl− solution containing 5% CO2/33 mM HCO3−. NBC-mediated inward currents were obtained at a holding pipette potential (Vp or −Vm) of +60 mV.
CWT and CΔN87.
As shown in Fig. 14 B, there was no appreciable change in current when a macropatch from an H2O-injected oocyte was exposed first to CO2/HCO3−, and then to the HCO3− solution containing 200 μM DIDS. In contrast, exposing a macropatch from a CWT-expressing oocyte to CO2/HCO3− elicited an inward current of ∼5 pA (Fig. 14 C). This inward current is consistent with NBC- mediated net-negative charge moving from the bath to the pipette held at +60 mV. Switching back to the ND96 solution caused the current to return to the original level. In many experiments with either the C or A variant, we could reactivate this HCO3−-induced inward current by reexposing the patch to the CO2/HCO3− solution (unpublished data). However, the current change elicited by the second HCO3− exposure was often smaller—a result consistent with “rundown” of transporter activity. The phenomenon of rundown is observed with channels such as ROMK1 (McNicholas et al., 1994), and reflects events such as protein dephosphorylation (McNicholas et al., 1994) or loss of protein interaction with membrane phospholipids such as phosphatidylinositol 4,5-bisphosphate (PIP2) (Huang et al., 1998; Suh and Hille, 2005).
As shown in Fig. 14 D, exposing a macropatch from an oocyte expressing CΔN87 to the low-Cl−, CO2/HCO3− solution elicited a large inward current of ∼28 pA. The current was ∼5.5-fold larger than the HCO3−-induced current seen with CWT (Fig. 14 C). If these HCO3−-induced inward currents are NBC mediated, then they should be inhibited by stilbene derivatives such as DIDS. Indeed, the inward current seen with CΔN87 was completely reversed by exposing the patch to 200 μM DIDS in the continued presence of CO2/HCO3− (Fig. 14 D). DIDS inhibits NBCe1 activity from either the extracellular or intracellular side of an oocyte macropatch (Heyer et al., 1999). In a total of five experiments similar to that shown in D, the mean HCO3−-induced inward current was inhibited 89 ± 11% by 200 μM DIDS. The mean HCO3−-induced inward current seen with CWT (e.g., in Fig. 14 C) was also inhibited 103 ± 30% (n = 3) by DIDS. In summary, the larger HCO3−-stimulated currents elicited by removing the amino terminus of NBCe1-C reported above in the whole-cell experiments (Figs. 9 and 10) are also seen in macropatches excised from the oocyte.
AWT and AΔN43.
We performed macropatch experiments on AWT and AΔN43 using the aforementioned protocol. Based on the finding that macropatch currents for CΔN87 were larger than those of CWT, an observation that paralleled the whole-cell data, we anticipated AWT to exhibit the largest macropatch currents of all the NBC constructs in the present study. As shown in Fig. 14 E, an HCO3−-induced, NBC-mediated current of ∼7 pA was indeed obtained in a macropatch from an oocyte expressing AWT. Unexpectedly however, the mean AWT current was of similar magnitude to the CWT current (Fig. 14 C), and considerably less than the CΔN87 current (D). In additional studies (not depicted), we confirmed that this AWT-mediated HCO3−-induced current was eliminated with 200 μM DIDS. The inward current elicited by CO2/HCO3− in a macropatch expressing AΔN43 was absent (Fig. 14 F). The NBC-mediated current rose much faster for the AWT (Fig. 14 E) variant than the CWT variant (C)—an observation also made in whole-cell experiments (see Fig. 5). Therefore, the activation kinetics of A appear faster than those of C, even in the absence of the cytosol.
From experiments similar to those shown in Fig. 14 (B–F), we plotted the mean HCO3−-induced currents from a total of 68 inside-out macropatch experiments (Fig. 15). Similar to the results from the whole-cell data, the mean macropatch current for CΔN87 is 3.2-fold larger than the mean current for CWT (P = 0.03), whereas the mean current for AΔN43 is 71% less than that of AWT (P < 0.01). In fact, the mean macropatch currents from oocytes injected with AΔN43 and H2O are indistinguishable (P = 0.25). Macropatch data in disagreement with the whole-cell data include the absence of AΔN43 current, as well as similar mean currents for AWT and CWT (P = 0.32). These disagreements may reflect the absence of a cytosolic component in the macropatch experiments that is necessary for both the AΔN43 current and the large AWT current seen in the whole-cell experiments.
Discussion
Activity of NBCe1 Variants
In this study, we compared the function and expression of the three NBCe1 variants (A, B, and C) heterologously expressed in Xenopus oocytes. In whole-cell experiments, the A variant displayed larger NBC currents than the B and C variants; an observation not due to differences in plasma membrane expression, or differences in voltage or external ion dependencies. The higher transport velocity of the A variant is due to its unique amino terminus. In both whole-cell and macropatch studies, removing the unique amino terminus of the A variant decreases activity, whereas removing the different amino terminus of the C variant increases activity.
Voltage and Ion Dependencies of NBCe1 Variants in Xenopus Oocytes
While all three variants exhibited similar voltage and external ion dependencies, the reversal potential (Erev) of NBCe1-A was more positive (∼−85 mV) than the Erev values of the B and C variants (∼−160 mV) (Fig. 4 A). Working on rat kidney NBCe1-A expressed in oocytes, Sciortino and Romero (1999) obtained an Erev similar to ours, and calculated a corresponding 1:2 Na+:HCO3− stoichiometry for the transporter. As mentioned in Results, the more negative Erev values for the B and C variants are unlikely to reflect different Na+:HCO3− stoichiometries; but rather, different substrate and pH gradients established immediately across the plasma membrane by the variants. Indeed, we observed a similar negative shift of Erev in oocytes displaying smaller NBCe1-A currents in response to less cRNA injected (Fig. 4 B). Furthermore, the I-V relationships for the three variants looked similar when we adjusted expression levels and matched the magnitude of the whole-cell, transporter currents (Fig. 4 C).
For the NBCe1 variants, we calculated transporter stoichiometry using Erev values and the extracellular Na+ and HCO3− concentrations used in the study. The equilibrium potential of an electrogenic NBC (ENBC) is described by the following equation (Boron and Boulpaep, 1983; Deitmer and Schlue, 1989):
where R, T, and F have their usual meanings, and n is the HCO3−:Na+ stoichiometry. Solving Eq. 1 for n yields the transformed equation:
We determined the mean intracellular HCO3− concentration of ∼8.3 mM from simultaneous pHi and voltage-clamp experiments on NBCe1-C–expressing oocytes similar to those shown in Fig. 11 A. For the intracellular Na+ concentration, we used a measured value of 10.4 mM previously reported by Sciortino and Romero (1999) for unclamped oocytes expressing NBCe1-A. Our calculated Na+:HCO3− stoichiometries are 1:1.8 for the B variant and 1:1.7 for the C variant. These values are similar to the 1:2 Na+:HCO3− stoichiometry previously calculated for the A variant (Sciortino and Romero, 1999). Moreover, we used the Erev of −120/−140 mV from the I-V relationship shown in Fig. 4 (B and C) to calculate a similar Na+:HCO3− stoichiometry of ∼1:2 for the A variant. We used the lower-current I-V relationships for the analysis to minimize the complication of a large NBC-mediated substrate/pH gradient across the membrane (see above). In summary, all three variants expressed in oocytes appear to have the same 1:2 Na+:HCO3− stoichiometry.
Contribution of the Carboxy Termini to NBCe1 Activity
The carboxy-terminal truncations of NBCe1 variants examined in this study express poorly at the plasma membrane of oocytes. Based on our results, one or more regions within the different carboxy termini among the variants is/are required for proper plasma membrane expression in oocytes. Similarly, Cordat et al. (2003) reported that an AE1 construct truncated 11 residues from the carboxy terminus displayed reduced plasma membrane expression when expressed in HEK293 cells. Furthermore, Li et al. (2004) reported that GFP-tagged, human NBCe1-A mutants missing either the carboxy-terminal 26 or 50 residues retargets to the apical membrane of MDCK cells, and the latter construct displays significant intracytoplasmic expression. Our data with both AΔC46 and CΔC61 are consistent with this finding that the carboxy termini can influence trafficking. However, the effect of truncating the different carboxy termini found in A and C on plasma membrane expression is more dramatic in oocytes. The reduced surface expression could be due to the loss of a residue/region required for protein expression/stability, or an altered conformation that leads to organellar retention.
Contribution of the Amino Terminus to NBCe1 Activity
The following three general models would explain the contribution(s) of the different amino termini to NBC activity: (1) the unique amino terminus of A (AN41) is stimulatory for an otherwise low-activity transporter, (2) the amino terminus of B/C is inhibitory (CN85) for an otherwise high-activity transporter, or (3) both termini can influence function of an otherwise intermediate-activity transporter. Our observation that both whole-cell and macropatch NBC currents were inhibited to an intermediate level by AΔN43 and stimulated to an intermediate level by CΔN87 are consistent with the third possibility. The amino-terminal truncations likely lead to changes in transporter velocity. Indeed, the external Na+ and HCO3− dependencies for CΔN87 were similar to those for CWT (unpublished data). Therefore, the different amino termini of the variants can influence transporter activity two to threefold even though they only comprise a small fraction of the entire NBCe1 sequence (4% for A and 8% for C). While the different amino termini stimulate or inhibit transporter activity, one or more regions of the cytosolic amino terminus downstream of the variable region are required for NBC function. Removing nearly all of the cytosolic amino terminus of NBCe1-A or -C, or approximately half of this region of NBCe1-C, leads to a complete loss of transporter function without affecting plasma membrane expression.
Potential Mechanism(s) for the Amino-terminal Effects.
There are several possible mechanisms by which the amino terminus of an NBCe1 could affect transporter activity. An amino terminus may bind to a region of the same NBC protein and either elicit a change in protein conformation or directly interact with the ion translocation pathway. However, a specific carboxy terminus does not appear to be essential in light of the observation that the A variant and the A variant containing the carboxy terminus of C (A/C chimera) display a similar level of activity. Interaction with the translocation pathway may be similar to the “ball and chain” mechanism of inactivation of potassium channels (Zagotta et al., 1990; Tseng-Crank et al., 1993). Alternatively, the amino terminus of one NBC may exert its effect by binding to another NBC protein. Evidence for multimerization of NBCe1 has been reported in preliminary form (Espiritu, D.J.D., A.A. Bernardo, and J.A.L. Arruda. 2003. FASEB J. 17:A462; Espiritu, D.J.D., A.A. Bernardo, and J.A.L. Arruda. 2004. J. Am. Soc. Nephrol. 15:F-PO016.). Another mechanism for the amino-terminal effects may involve regulatory proteins/factors that are either cytosolic or membrane bound (see below). Regulatory proteins may serve as intermediate proteins that couple the amino termini to regions of an NBC protein.
NBCe1 Macropatch Currents
NBCe1-mediated HCO3− Transport across a patch of Oocyte Membrane.
We used the macropatch technique in the inside-out configuration to examine the potential involvement of a cytosolic component on the activity of NBCe1 constructs expressed in oocytes. The magnitudes of our NBC macropatch currents were similar to those previously reported for rat NBCe1-A (Heyer et al., 1999). In our studies, it was not uncommon to record from patches with little or no NBC-mediated currents. For all NBC groups, there was a broad range of HCO3−-mediated currents. For CWT, current responses ranged from an outward current of 4.2 pA to an inward current of −14 pA. The range of responses may reflect different plasma membrane expression levels of NBCs among patches. Indeed, AWT currents from three vegetal-pole patches obtained from each of two oocytes also exhibited a range of currents: 0.4, −5.1, and −3.3 pA for one oocyte, and −7.8, −3.3, and −7.4 pA for the other oocyte. NBCe1 expression at the oocyte surface may therefore be somewhat heterogeneous and reflect clustering of transporter proteins.
Influence of the Amino Termini on NBCe1 Activity.
There are three noteworthy observations regarding the macropatch results and the influence of the amino termini on NBC activity. First, the mean macropatch current for CΔN87 is 3.2-fold larger than the mean current for CWT. A similar 2.7-fold larger current for CΔN87 was observed in the whole-cell experiments. Second, the mean current for AΔN43 is less than that of AWT. Similar results were obtained in the whole-cell experiments: the mean AΔN43 current was 55% less than that of AWT. Both of these observations corroborate the whole-cell data regarding the functional impact of the different amino termini.
The third observation in the macropatch studies, which deviates from the whole-cell data, is the observation that the mean macropatch current for AWT is similar to the mean current for CWT. Furthermore, the mean current for AΔN43 is not only lower than for AWT, but is nonexistent and similar to that seen in patches from H2O-injected oocytes. These deviations from the whole-cell data may be due to different intracellular ion dependencies or sensitivities among the variants, and the fact that the macropatch experiments were performed with a symmetrical solution containing extracellular ion concentrations. High Na+ and HCO3− concentrations on the cytosolic side were used to maximize the transporter currents. Another explanation is the involvement of cytosolic factors in the whole-cell experiments that are absent in the macropatch experiments. While such factors could either stimulate A or inhibit C, the AΔN43 current observed in whole-cell but not macropatch experiments is consistent with a cytosolic factor stimulating the A variant.
Potential cytosolic factors include cytoskeletal elements such as actin or enzymes. Indeed, the amino-terminal domain of AE1 interacts with actin, protein 4.1, protein 4.2, as well as glycolytic enzymes (Reithmeier et al., 1996). Classic regulatory factors include cyclic AMP and ATP. Heyer et al. (1999) reported that 2 mM ATP increases the activity of rat NBCe1-A from inside-out macropatches by twofold. The effect of ATP on NBCe1 activity may be mediated by ATP-dependent lipid kinases and PIP2, as previously reported for the cardiac Na-Ca exchanger and KATP potassium channels (Hilgemann and Ball, 1996). There is considerable evidence highlighting the regulatory effects of membrane phospholipids such as PIP2 on ion channels and transporters (Suh and Hille, 2005). Future studies will be necessary to characterize potential regulatory factors such as ATP, PIP2, and kinases/phosphatases that may differentially modulate the NBCe1 variants.
Acknowledgments
We thank Mr. Albert Tousson (High Resolution Imaging Center, University of Alabama at Birmingham) for assistance with confocal microscopy used in the immunocytochemical analysis of NBC expression in oocytes. Dr. Peter R. Smith (Physiology and Biophysics, University of Alabama at Birmingham) provided assistance and the luminometer used for single-oocyte chemiluminescence. A special thanks goes to Dr. Chou-Long Huang (University of Texas Southwestern Medical Center at Dallas, Dallas, TX) for providing guidance with the macropatch technique. We thank Mrs. Debeshi Majumdar for her helpful suggestions and comments regarding the manuscript.
This work was supported by the American Heart Association, Southeast Affiliate (0265083B) and National Institutes of Health/National Institute of Neurological Disorders and Stroke (NS046653).
Olaf S. Andersen served as editor.
References
S.D. McAlear and X. Liu contributed equally to this work.
Abbreviations used in this paper: AE, anion exchanger; DIDS, 4,4′-diisothiocyanatostilbene-2,2′ disulfonate; HA, hemagglutinin; HRP, horseradish peroxidase; NBC, Na/bicarbonate cotransporter; pHi, intracellular pH; PIP2, phosphatidylinositol 4,5-bisphosphate; SOC, single-oocyte chemiluminescence.