Electrophysiological characterization of native Na⁺–HCO₃^– cotransporter current in bovine parotid acinar cells

Abstract

Using patch-clamp and molecular biological techniques, we identified and characterized membrane currents most likely generated by an electrogenic Na⁺–HCO₃⁻ cotransporter (NBCe) in acutely dissociated bovine parotid acinar (BPA) cells. When BPA cells were dialysed with a N-methyl-d-glucamine (NMDG)-glutamate-rich pipette solution, switching a Na-glutamate-rich, nominally HCO₃⁻-free bath solution to the one containing 25 mm HCO₃⁻, but not Cl⁻, elicited a whole-cell current with a linear current–voltage relation. The HCO₃⁻ evoked current was abolished by total replacement of extracellular Na⁺ (Na⁺_o) with NMDG or by 0.5 mm 4,4′-diisothiocyanato-stilbene-2,2′-disulfonic acid (DIDS), and was only partially supported by Li⁺_o, but not by K⁺_o, Cs⁺_o, and choline_o. The reversal potential shift of DIDS (0.5 mm)-sensitive current induced by a change of [Na⁺]_o corresponded to an apparent coupling ratio of HCO₃⁻ to Na⁺ of 2. RT-PCR analysis showed the presence of transcripts of NBCe1-B, but not NBCe1-A in BPA cells. Electrophysiological and pharmacological properties of whole-cell currents recorded from HEK293 cells transfected with the NBCe1-B, which was cloned from BPA cells resembled those of the native currents. Non-invasive measurements of membrane potential changes in the cell-attached patch configuration indicated that an NBCe activity is present in intact unstimulated BPA cells bathed in a 25 mm HCO₃⁻-containing solution. Collectively, these results not only suggest that an NBCe is present, functional and may be mediated, at least in part, by NBCe1-B in BPA cells, but also provide the first electrophysiological characterization of transport properties of NBCe expressed in native exocrine glands.

Unlike other well-studied mammalian salivary glands, the parotid glands of adult ruminants such as cattle and sheep secrete continuously and rapidly (especially during feeding), large volumes of a HCO₃⁻-rich, almost isotonic fluid, which serves both as a proper fermentation medium for the microorganisms and as a buffer the fermentation products, mostly short-chain fatty acids in the rumen (Coats & Wright, 1957; Kay, 1960; Stevens, 1988). The HCO₃⁻ concentration in final saliva secreted by the ruminant parotid glands ranges from about 100 mm even up to 140 mm (Coats & Wright, 1957; Kay, 1960). Therefore, the higher concentration is comparable with that found in the pancreatic juice of humans, dogs, and guinea pigs (Steward et al. 2005). Although the HCO₃⁻ concentration in the primary saliva of ruminant parotid glands has not been directly measured, it is believed that these glands form a HCO₃⁻-rich primary fluid, even if their ducts also secrete HCO₃⁻, and that the primary secretion is driven by the secondary active transport of HCO₃⁻ across the basolateral membrane (Cook et al. 1994).

Previous functional studies have suggested that HCO₃⁻ may accumulate in the parotid cells via at least two mechanisms. One is by diffusion of CO₂ across the basolateral membrane and its subsequent generation of intracellular HCO₃⁻ by the action of carbonic anhydrase (CA) along with the extrusion of H⁺ via basolateral transporters such as the Na⁺–H⁺ antiporter. However, experiments in sheep and bovine parotid glands have indicated that CA inhibitors only decrease both the salivary secretion rate and HCO₃⁻ output during secretomotor nerve stimulation by about 40% (Blair-West et al. 1980) and muscarinically stimulated ⁸⁶Rb⁺ efflux by about 55% (Lee & Turner, 1993), respectively, implying that the HCO₃⁻ accumulation is not exclusively mediated by the hydration of intracellular CO₂ by CA. Consistent with this implication, it has been shown in sheep parotid acinar cells that a component of muscarinically stimulated Na⁺ uptake across the plasma membrane is dependent on extracellular HCO₃⁻/CO₂, but not on Cl⁻, and inhibited by dihydrogen-4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid (H₂DIDS) (Poronnik et al. 1995) and that intracellular pH recovery from an acid load during a cholinergic stimulation is dependent on extracellular Na⁺ and HCO₃^-/CO₂ and can be inhibited by H₂DIDS (Steward et al. 1996). These observations together suggest that the basolateral accumulation of HCO₃⁻ into the parotid acinar cells occurs via a stilbene-sensitive, Na⁺–HCO₃⁻ cotransport. However, it remains largely unknown in the ruminant parotid whether the cotransport is mediated by the electrogenic Na⁺–HCO₃⁻ cotransporter (NBCe) as postulated in the pancreatic duct.

The NBCe1, a group of NBCe, is believed to play a key role in HCO₃⁻ secretion by mediating a pathway for HCO₃⁻ influx across the basolateral membrane of the pancreatic duct (see for reviews, Romero et al. 2004; Steward et al. 2005), that secretes an isotonic fluid rich in HCO₃⁻ in many species (Steward et al. 2005), because immunohistochemical studies have shown that the NBCe1, especially NBCe1-B (or pNBC1), is expressed in the basolateral membrane of the native pancreatic ducts (Abuladze et al. 1998; Marino et al. 1999; Thévenod et al. 1999; Satoh et al. 2003; Roussa et al. 2004). Its functional impact has been suggested by the observations that a stilbene-sensitive, basolateral Na⁺- and HCO₃⁻-dependent intracellular pH recovery from an acid load, which is also found in native pancreatic ducts (Zhao et al. 1994; Villanger et al. 1995; Ishiguro et al. 1996) is affected by alterations in the basolateral membrane potential in a human pancreatic duct cell line endogenously expressing NBCe1-B (Shumaker et al. 1999). Furthermore, an Ussing chamber study in a mouse pancreatic duct cell line has demonstrated that a stilbene-sensitive, Na⁺- and HCO₃⁻-dependent current across the basolateral membrane is generated by coupling the transport of two equivalents of bicarbonate per each sodium (i.e. 2 HCO₃⁻: 1 Na⁺ stoichiometry), which would favour the basolateral HCO₃⁻ uptake during HCO₃⁻ secretion in pancreatic duct cells (Gross et al. 2001). Although such a cotransporter current has been never reported in native pancreatic duct cells, a membrane potential measurement study has suggested that the basolateral cotransporter is electrogenic in guinea pig pancreatic ducts (Ishiguro et al. 2002).

Taking the apparent functional similarities between the pancreas and ruminant parotid together with a growing body of recent immunohistochemical evidence for the expression of NBCe1 (or NBCe1-B) in non-ruminant salivary glands (Roussa et al. 1999; Luo et al. 2001; Park et al. 2002; Kim et al. 2003), we hypothesized that ruminant parotid acinar cells may express a significant electrogenic Na⁺–HCO₃⁻ cotransport activity and thus could serve to provide detailed information about its transport properties in a native HCO₃⁻-secreting epithelium. In the present study, we first used the whole-cell patch-clamp technique to show that acutely dissociated bovine parotid acinar (BPA) cells indeed exhibit a membrane current, which is totally dependent upon the presence of both extracellular Na⁺ and HCO₃⁻, is blocked by DIDS, and has an apparent coupling ratio of 1 Na⁺ to 2 HCO₃⁻, characteristics that are consistent with those reported for NBCe(s) endogenously and heterologously expressed in various cell types (Boron & Boulpaep, 1983; Romero et al. 1997; also see for reviews, Boron & Boulpaep, 1989; Romero et al. 2004). Using molecular biological techniques, we then show that the cells express transcripts of NBCe1-B, but not NBCe1-A, and subsequently demonstrate that detailed biophysical and pharmacological properties of the native current are similar to those of current recorded from HEK293 cells transfected with NBCe1-B cloned from BPA cells. We finally show that the cotransporter activity can significantly contribute to the membrane potential of intact BPA cells. To the best of our knowledge, the present work provides the first electrophysiological characterization of transport properties of the electrogenic Na⁺–HCO₃⁻ cotransporter expressed in a native HCO₃⁻ -secreting exocrine gland.

Methods

Patch-clamp experiments

Bovine parotid tissue was obtained from a local slaughterhouse (Hokkaido Hayakita Meat Inspection Center, Hayakita, Japan). After the tissue was removed from the slaughtered animal, it was kept at 4°C in a standard NaCl-rich bath solution (Solution A, Table 1) usually for several hours until the animal was approved to be negative for testing bovine spongiform encephalopathy (BSE). Isolated acini and acinar cells were prepared as described elsewhere (Hayashi et al. 2003; Takahata et al. 2003).

Table 1.

Composition of experimental solutions (all concentrations are in mm)

	Pipette					Bath
Components	A	B	C	D	E	A	B	C	D	E	F	G	H	I	J	K	L	M	N
NaCl	10	10	—	10	—	145	—	—	25	—	—	—	—	—	—	—	120	—	—
KCl	—	—	—	—	135	5	—	—	—	—	—	—	—	—	—	—	5	5	5
Na-glutamate	—	—	—	—	—	—	120	145	120	—	120	—	—	—	—	—	—	—	—
NMDG-glutamate	120	100	25	50	—	—	—	—	—	120	—	—	—	—	—	—	—	—	—
NMDG-Cl	—	—	10	—	—	—	—	—	—	—	—	—	—	—	—	—	—	120	120
NaHCO3	—	—	10	—	—	—	25	—	—	—	—	—	—	—	25	—	25	—	—
CholineHCO3	—	—	15	—	—	—	—	—	—	25	25	25	25	25	—	—	—	25	—
KHCO3	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	25	—	—	—
Cs-glutamate	—	—	—	—	—	—	—	—	—	—	—	—	—	120	120	120	—	—	—
Li-glutamate	—	—	—	—	—	—	—	—	—	—	—	120	—	—	—	—	—	—	—
Choline-glutamate	—	—	—	—	—	—	—	—	—	—	—	—	120	—	—	—	—	—	—
CholineCl	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	25
Glucose	—	—	—	—	10	10	—	—	—	—	—	—	—	—	—	—	10	10	10
Hepes	10	10	100	100	10	10	10	10	10	10	10	10	10	10	10	10	10	10	10
EGTA	10	10	10	10	5	—	—	—	—	—	—	—	—	—	—	—	—	—	—
CaCl2	—	—	—	—	—	1 or 0	1 or 0	1 or 0	1 or 0	1 or 0	1 or 0	1 or 0	1 or 0	1 or 0	1 or 0	1 or 0	1	1	1
MgCl2	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1

The nominally HCO₃⁻-free bath solutions (pH 7.4; adjusted with NaOH or NMDG) were continuously gassed with 100% O₂. The HCO₃⁻-containing solutions were prepared by adding either KHCO₃, NaHCO₃, or (/and) cholineHCO₃ to the solutions that contained all other constituents (pH 7.4; adjusted with NMDG), and were equilibrated with 5% CO₂/95% O₂ (pH 7.4). The pipette solution containing HCO₃⁻ (25 mm) was continuously gassed with a mixture of 95% O₂ and 5% CO₂ until it was filled into the patch pipette.

The methods used for measurements of whole-cell and single-channel currents were similar to those described elsewhere (Hayashi et al. 2003; Takahata et al. 2003). In brief, an Axopatch-1D patch-clamp amplifier and the pCLAMP6 software (Axon Instruments, Union City, CA, USA) were used to perform voltage clamp and data storage and analysis. A reference Ag–AgCl electrode was connected to the bathing solution via an agar bridge filled with the standard NaCl-rich bath solution. The whole-cell and single-channel currents were filtered at 500 Hz with an internal four-pole Bessel filter, and sampled at 2 kHz. Current–voltage (I–V) relations for whole-cell currents were mainly studied using voltage ramps. Typically, the cell potential held at −80 mV, and the command voltage was varied from −100 to +50 mV over a duration of 800 ms every 10 s. Some experiments were performed using 10 mV voltage pulses, each of 400 ms duration, delivered at voltages ranging between −100 and +50 mV, and voltage pulses were separated by 7 s during which the cell potential held at −80 mV. The cell capacitance was 34.2 ± 1.0 pF (n = 67). The average series resistance (R_s), which was 26.7 ± 1.2 MΩ (n = 67), was not electrically compensated. The experimental solutions used in the present patch-clamp experiments are described in Table 1. The pipette potential was corrected for the liquid junction potential between the pipette solution and the external solution, and between the external solution and the agar bridge as described elsewhere (Barry & Lynch, 1991; Neher, 1992).

Ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid (EGTA), 4,4′-diisothiocyanato-stilbene-2,2′-disulfonic acid (DIDS), glybenclamide, 2-(4-(2-hydroxyethyl)-1-piperazinyl) ethanesulfonic acid (Hepes), niflumic acid (NFA), N-methyl-d-glucamine (NMDG), 5-nitro-2-(3-phenylpropylamino) benzoic acid (NPPB) and phloretin were obtained from Sigma (St Louis, MO, USA). Atropine and diphenylamine-2-carboxylate (DPC) were obtained from Wako Chemicals (Osaka, Japan), and collagenase type II was obtained from Worthington Biochemical Corporation (Freehold, NJ, USA). Other chemicals were of reagent grade. Stock solutions of DIDS (50 mm) and atropine (10 mm) were prepared in distilled water. Stock solutions of glybenclamide (0.3 m), NFA (0.1 m), NPPB (0.1 m), phloretin (0.3 m), DPC (0.1 m) were dissolved in dimethyl sulfoxide (DMSO).

All experiments were performed at room temperature. The results are reported as means ±s.e.m. of independent experiments (n), where n refers to the number of cells patched. Statistical significance was evaluated using Student's two-tailed paired or unpaired t test as appropriate. A value of P < 0.05 was considered significant.

Data fitting

To analyse titration curves for DIDS inhibition of the macroscopic NBC current, the ratio I/I₀ measured in the presence (I) of the blocker to that in its absence (I₀), was described by the following equation:

(1)

where K_i is the inhibitory constant of the blocker, A is the concentration of the blocker and n is the pseudo Hill coefficient. The K_m values for external Na⁺ were calculated from the macroscopic currents induced by varying the concentration (C) of Na⁺ in the bath solution from 0 to 145 mm. The data were fitted to the Michaelis–Menten equation:

(2)

where I_max is the maximal current and K_m is the half-activation constant using a nonlinear fit program (Igor Pro, Wave Metrics, Lake Oswego, OR, USA).

An apparent HCO₃⁻to Na⁺ transport ratio (q) was estimated by the following equation:

(3)

where F, R and T have their usual meaning, and brackets and subscripts i and o stand for the intra- (i.e. pipette solution) and extracellular concentration of the indicated ions, respectively. E_rev is the reversal potential that can be experimentally determined from I–V relationship.

Cloning of NBC from bovine parotid cells

mRNA was extracted from bovine parotid cells prepared as described above using TRIzol reagent (Life Technologies, Grand Island, NY, USA) and BioMag mRNA purification kit (Polysciences, Warrington, PA, USA) following the producer's instructions. First-strand cDNA was generated from mRNA using SuperScriptII RT (Life Technologies). The specific oligonucleotide primers for polymerase chain reaction (PCR) for NBCe1-B were derived from the published open reading frame sequences of the bovine NBCe1-B (GenBank accession No.AF308160). The NBCe1-B sense primer was 5′-ATG GAG GAT GAA GCT GTC CTG GAC AGA GGG-3′ (nt 1–30) and the antisense primer was 5′-CAG CAT GAT GTG TGG CGT TC-3′ (nt 3239–3220). Other specific primers for NBCe1-A were derived from the published sequences of the rat NBCe1-A (GenBank accession No.AF027362). The NBCe1-A sense primer was 5′-ATG TCC ACT GAA AAT GTG GAA GGG AAG CCC-3′ (nt 56–85) and the antisense primer was the same primer as the NBCe1-B (nt 3162–3143). The size of the expected fragments of NBCe1-B and NBCe1-A were 3239 bp and 3107 bp, respectively. The cDNA from the rat kidney was used as a positive control for NBCe1-A. The PCR reaction was performed with TaKaRa LA Taq (Takara Bio, Otsu, Japan). The PCR conditions were: denaturation 94°C/30 s; annealing 65°C/30 s; extension 72°C/4 min; 35 cycles. As a control, β-actin-cDNA was amplified using the primers 5′-GAC TAC CTC ATG AAG ATC CT-3′ (sense) and 5′-CCA CAT CTG CTG GAA GGT GG-3′ (antisense) and 510 bp product was obtained. The PCR products obtained were resolved in 1% agarose gels with ethidium bromide.

The sequence of the bovine NBCe1-B was extended by 5′-RACE and 3′-RACE (rapid amplification of cDNA ends). Briefly, for 5′-RACE first-strand cDNA synthesis was performed with the gene-specific antisense primer 5′-CGA TCT CAT GGT AGG ACT TGG CTT TC-3′ (nt 1021–996). The cDNA was purified and a poly(C) tail was added to the 5′ end using terminal deoxynucleotidyl transferase (Promega, Madison, WI, USA) and dCTP. The tailed cDNAs were amplified by PCR using the oligo(dG) anchor primer 5′-GGC CAC GCG TCG ACT AGT ACG GGI IGG GII GGG IIG-3′ and the gene-specific antisense primer 5′-AAG GTC AGG TTT CAA TAG GC-3′ (nt 600–581). The PCR products obtained were gel-purified, cloned into pGEM-T Easy Vector and sequenced. For 3′-RACE, first-strand cDNA synthesis was performed using an oligo(dT) anchor primer 5′-CCA GTG AGC AGA GTG ACG TTT TTT TTT TTT TTT TT-3′. The cDNA mixtures were amplified with PCR using the gene-specific sense primer 5′-TTC AAG CCA ACG AGT CCA AAC-3′ (nt 2275–2295) and the anchor primer 5′-CCA GTG AGC AGA GTG ACG-3′, and with nested PCR using the gene-specific sense primer 5′-CCT CAT GGT GGT GTG CTC-3′ (nt 2484–2501) and anchor primer. The PCR products obtained were gel-purified and directly sequenced.

Based on the obtained sequence information of the untranslated region and the published sequences of the open reading frame, we amplified five overlapping PCR fragments by RT-PCR with high-fidelity DNA polymerase (Pfu-Turbo, Stratagene, LaJolla, CA, USA) using the following primers:

1. fragment 1: 5′-CTC GCG CCG ACA ACT TC-3′ (5′ untranslated region, sense) 5′-CGA TCT CAT GGT AGG ACT TGG CTT TC-3′ (nt 1021–996, antisense)

2. fragment 2: 5′-GCT TGT TGG CGA GGT AGA CTT-3′ (nt 861–881, sense) 5′-AGT TGG AGT TGA TGG GGT AGT-3′ (nt 1849–1829, antisense)

3. fragment 3: 5′-ATT TGG AGT TTC GCC TTT GGA T-3′ (nt 1649–1670, sense) 5′-GAT GTG AGC GAT GGA GAT GAC-3′ (nt 2556–2536, antisense)

4. fragment 4: 5′-TTC AAG CCA ACG AGT CCA AAC-3′ (nt 2275–2295, sense) 5′-CAG CAT GAT GTG TGG CGT TC-3′ (nt 3239–3220, antisense)

5. fragment 5: 5′-CCT CAT GGT GGT GTG CTC-3′ (nt 2484–2501, sense) 5′-GAG GCA CAA CTT TTA CTG GA-3′ (3′ untranslated region, antisense)

The fragments were cloned into the pGEM-T Easy vector and sequenced. The different nucleotides from the published sequences were confirmed by sequencing both the full-length PCR fragment of NBCe1-B and the fragments 1–5. The fragments were assembled by ligation of 5 fragments EcoRI (in vector)/Eco81I (5′ untranslated region – nt 990), Eco81I/NheI (nt 990– 1819), NheI/EcoRV (nt 1819–2456), EcoRV/PstI (nt 2456–2871), PstI/EcoRI (in vector) (nt 2871–3′ untranslated region) into pCI-neo mammalian expression vector (Promega) linearized by EcoRI.

Transfection of HEK293 cells with NBCe1-B

HEK293 cells expressing NBCe1-B were generated by transfecting cells with a plasmid construct encoding NBCe1-B that was cloned from bovine parotid cells in the mammalian expression vector pCI-neo using lipofectamine (Life Technologies, Inc., Tokyo, Japan). Mock transfected cells were also prepared using the same vector. G418 (0.8 mg ml⁻¹)-resistant colonies were purified and tested for NBCe1-B expression by whole-cell patch clamp and RT-PCR. The cells were then maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, penicillin (100 U ml⁻¹)/streptomycin (100 µg ml⁻¹) and 0.2 mg ml⁻¹ G418. The cells were grown at 37°C in a water-saturated 5% CO₂, 95% air atmosphere. For patch-clamp experiments, the cells were seeded on glass coverslips and patched 2–4 days after seeding. For the purpose of the present study, we chose and characterized a clone of transfected cells, which exhibited a DIDS-sensitive, Na⁺- and HCO₃⁻-dependent whole-cell current. The experimental conditions were the same as those described for the acutely dissociated BPA cells unless otherwise stated. In whole-cell patch clamp experiments, the cell capacitance and the average series resistance (R_s) were 23.1 ± 1.7 pF (n = 31) and 16.9 ± 0.7 MΩ (n = 31), respectively.

Results

A DIDS-sensitive, Na⁺- and HCO₃⁻-dependent whole-cell current in bovine parotid acinar cells

Using the standard whole-cell patch-clamp technique (i.e. fast whole-cell), we first examined whether bovine parotid acinar (BPA) cells exhibit a membrane current attributable to electrogenic Na⁺–HCO₃⁻ cotransporter (NBC) activity. Figure 1Aa shows traces of instantaneous current-voltage (I–V) relations of whole-cell currents obtained from a BPA cell that was dialysed with a NMDG-glutamate-rich pipette solution (pipette solution A in Table 1). In this experiment, we began with the cell in a nominally HCO₃⁻-free, Na-glutamate-rich bath solution (bath solution C in Table 1), and then switched to a solution containing 25 mm NaCl (bath solution D in Table 1) to confirm that the cell exhibited little, if any, Cl⁻-permeable conductance. Once the cell was perfused with a bathing solution containing 25 mm NaHCO₃ (bath solution B in Table 1), a large whole-cell current with a reversal potential more negative than −80 mV was observed. The HCO₃⁻-induced current decreased to the initial level upon washing the cell with the nominally HCO₃⁻-free, Na-glutamate-rich solution (not shown). Similar results were obtained in 13 independent experiments (Fig. 1Ab). The HCO₃⁻-induced current was also dependent upon external Na⁺ (Na⁺_o). As shown in Fig. 1Ba and b, when Na⁺_o was replaced with NMDG in the presence of 25 mm HCO_3o⁻, the currents were largely reduced. The incidence of the Na⁺- and HCO₃⁻-dependent whole-cell conductance was 100% (n = 182) irrespective of the pipette solutions used (i.e. A–D; Table 1). The conductance was also inhibited by 0.5 mm DIDS, a well-known blocker of the electrogenic Na⁺–HCO₃⁻cotransporter (Fig. 1Cd).

Figure 1. A DIDS-sensitive, Na⁺- and HCO₃⁻-dependent whole-cell current (I_{Na⁺/HCO3⁻}) in bovine parotid acinar (BPA) cells.

Aa, instantaneous current–voltage (I–V) relationships of whole-cell currents obtained from a single bovine parotid acinar (BPA) cell. The pipette was filled with a NMDG-glutamate rich pipette solution (pipette solution A), and the bath contained either a Na-glutamate-rich solution (glutamate; bath solution C), a solution containing 25 mm NaCl (Cl⁻; bath solution D) or a solution containing 25 mm NaHCO₃ (HCO₃⁻; bath solution B). The currents were evoked by voltage ramps, each of 800 ms duration, from a holding potential of −78 mV (or −79 mV) between −98 mV (or −99 mV) and +52 mV (or +51 mV). Ab, summary of I–V relationships. Values are means ±s.e.m. of 13 independent experiments (•, HCO₃⁻; ▴, Cl⁻; ▪, glutamate). Error bars representing s.e.m. were omitted when so small as to lie within symbols. Ba, instantaneous I–V relationships obtained from a BPA cell bathed in solutions having 25 mm HCO₃⁻ in the presence (Na⁺; bath solution B) and absence of external Na⁺ (Na⁺_o) (NMDG; bath solution E). Currents were evoked by voltage ramps from a holding potential of −78 mV (or −80 mV) between −98 mV (or −100 mV) and +52 mV (or +50 mV). Bb, summary of I–V relationships. Values are means ±s.e.m. of nine independent experiments. (•, Na⁺; ♦, NMDG). Error bars representing s.e.m. were omitted when so small as to lie within symbols. Ca and b, traces of whole-cell currents associated with voltage steps for 400 ms from a holding potential of −78 mV between −98 mV and +52 mV in 10 mV intervals. Data were obtained from a BPA cell in the absence and presence of 0.5 mm DIDS in the bath solution (solution B). Cc, traces of the DIDS-sensitive currents obtained from traces of Fig. 1Ca and b. Cd, current traces obtained from a cell before (Control) and after (+DIDS) addition of 0.5 mm DIDS to the bath solution (bath solution B). The inhibitory effect of DIDS was completely reversible (not shown). Data from the same cell shown in Fig. 1Ca and b. Ce, comparison of the DIDS-sensitive I–V relationships obtained from ramp pulse (line; obtained from Fig. 1Cd) and voltage step pulse protocols (○, obtained from Fig. 1Cc).

We also studied the kinetics of the DIDS-sensitive, Na⁺_o- and HCO_3o⁻-dependent current using a voltage-pulse protocol (Fig. 1Ca–c). In the voltage range tested, the currents activated instantaneously and showed little, if any, time-dependent activation or inactivation (Fig. 1Cc). The DIDS-sensitive current had a linear I–V relationship, from which the reversal potential was estimated to be approximately −90 mV (Fig. 1Ce). These results together suggest that BPA cells express a DIDS-sensitive, Na⁺- and HCO₃⁻-dependent current. The current will hereafter be designated I_{Na⁺/HCO3⁻}. Furthermore, since there was little, if any, difference between I–V relationships determined by voltage ramp- and voltage step pulse-protocols (Fig. 1Ce), electrophysiological properties of I_{Na⁺/HCO3⁻} were characterized by using ramp voltage protocol in the following experiments, unless otherwise stated.

Monovalent cation dependence

Figure 2A shows families of whole-cell ramp currents recorded from a BPA cell bathed in a solution containing choline-HCO₃ (25 mm) and either Na⁺, Li⁺, Cs⁺, NMDG or choline as a major monovalent cation (120 mm). Among monovalent cations tested, only Li⁺ could support a minor fraction of the current (Fig. 2B). We also determined whether K⁺ could support the HCO_3o⁻-induced currents. To this end, we used a Cs⁺ (120 mm)-rich bath solution having 25 mm of test cations (Na⁺, K⁺ or choline) to completely block an inwardly rectifying K⁺ current (Hayashi et al. 2003). Under these experimental conditions, K⁺ as choline failed to support the I_{Na⁺/HCO3⁻} (Fig. 2C and D). When the outward current amplitudes at +50 mV were normalized to the values calculated from the difference between the corresponding values obtained in the presence of Na⁺ (120 or 25 mm) and choline, the relative selectivity sequence was: Na⁺ (by definition: 100%) ≫ Li⁺ (13.9 ± 1.2%) > K⁺ (4.7 ± 0.7%) > choline (by definition: 0%) ≈ NMDG (−0.1 ± 2.1%) ≈ Cs⁺ (−2.9 ± 1.3%).

Figure 2. Monovalent cation dependence.

A, traces of whole-cell currents obtained from a BPA cell bathed in a solution containing 25 mm HCO₃⁻ rich in either Na⁺ (bath solution F), Li⁺ (bath solution G), Cs⁺ (bath solution I), NMDG (bath solution E), or choline (bath solution H). The pipette was filled with a NMDG-glutamate-rich solution (pipette solution A). B, summary of I–V relationships. Values are means ±s.e.m. of five or six independent experiments. Error bars representing s.e.m. were omitted when so small as to lie within symbols. (•, Na⁺; ▵, Li⁺; □, Cs⁺; ▴, NMDG; ○, choline) C, traces of whole-cell currents obtained from a cell bathed in a Cs⁺ (120 mm)-rich solution having either 25 mm Na⁺ (bath solution J), K⁺ (bath solution K), or choline (bath solution I). The pipette solution and ramp pulse protocols were identical, and similar to those described in Fig. 2A, respectively. D, summary of I–V relationships. Values are means ±s.e.m. of six independent experiments. Error bars representing s.e.m. were omitted when so small as to lie within symbols. (•, Na⁺; ▴, K⁺; □, choline)

Extracellular Na⁺ dependence

Figure 3A shows traces of instantaneous I–V relations obtained from a BPA cell in the presence of different extracellular Na⁺ concentration ([Na⁺]_o) (145, 100, 25, 9, 3, and 0 mm), indicating that the I_{Na⁺/HCO3⁻} is dependent upon [Na⁺]_o (see also summary shown in Fig. 3B). To analyse the [Na⁺]_o dependence of the I_{Na⁺/HCO3⁻} in more detail, the [Na⁺]_o-dependent current amplitudes at different membrane potentials normalized to the corresponding values in the presence of 145 mm[Na⁺]_o were plotted against the [Na⁺]_o (Fig. 3C). The normalized currents at +50 mV were also fitted to the Michaelis–Menten equation (eqn (2), see Methods) with an apparent K_m of 20.8 ± 1.0 mm (n = 9), the K_m being independent of the membrane potential (Fig. 3C inset).

Figure 3. Extracellular Na⁺ (Na_o⁺) dependence of I_{Na⁺/HCO3⁻}.

A, tracings of instantaneous I–V relationships recorded from a BPA cell in the presence of different extracellular Na⁺ concentrations ([Na⁺]_o; 145, 100, 25, 9, 3, 0 mm). The solutions containing 145, 25 and 0 mm[Na⁺]_o were the bath solution B, J and I (Table 1), respectively, from which the solutions having 100, 9 and 3 mm[Na⁺]_o were prepared. The pipette was filled with a solution containing 25 mm HCO₃⁻ and 100 mm Hepes (pipette solution C). B, summary of I–V relations. Values are means ±s.e.m. of nine experiments (•, 145 mm; ○, 100 mm; ▴, 25 mm; ▵, 9 mm, □, 3 mm; ▪, 0 mm[Na⁺]_o). Error bars representing s.e.m. were omitted when so small as to lie within symbols. C, plot of the outward current amplitude at −40 mV (○), 0 mV (□), or +50 mV (▵) (n = 9) as a function of [Na⁺]_o. Data were normalized to the value obtained from those with 145 mm[Na⁺]_o and 0 mm[Na⁺]_o(I_xmM−I_0mM/I_145mM−I_0mM). Each continuous line represents the result of least-squares curve fit of the Michaelis–Menten equation (eqn (2)) to the data. Inset; voltage dependence of K_m values for [Na⁺]_o. The continuous line shows linear regression fit to the data. Values are means ±s.e.m. of nine experiments. Error bars representing s.e.m. were omitted when so small as to lie within symbols.

Transport stoichiometry

We next determined the value of the apparent coupling ratio (i.e. transport stoichiometry; q, see eqn (3) in Methods) of HCO₃⁻ to Na⁺ of the I_{Na⁺/HCO3⁻}. To this end, we measured the reversal potential of the DIDS-sensitive current in the presence of two different [Na⁺]_o (i.e. 3 and 9 mm). In these experiments we assumed that the concentrations of cytosolic Na⁺ and HCO₃⁻ were clamped with the pipette solution for two reasons. First, these [Na⁺]_o concentrations elicit a small current and thus should minimize possible changes in local [Na⁺]_i and [HCO₃⁻]_i underneath the plasma membrane due to the influx of these ions. In fact, such changes probably appear to take place when a large current is induced in the presence of a high [Na⁺]_o (e.g. see Fig. 1A and B; the HCO₃⁻-induced current is reversed at a V_m less negative than −100 mV). Second, [HCO₃⁻]_o was assumed to be equal to [HCO₃⁻]_i, because the pipette solutions were buffered with 100 mm Hepes (pipette solution C, Table 1) to help to hold [pH]_i and thus [HCO₃⁻]_i constant, and because the pipette and bath solutions were equilibrated with 5% CO₂. Therefore, the value of q was calculated from the shift of reversal potential of DIDS (0.5 mm)-sensitive currents upon changing [Na⁺]_o; that is, from a simplified form of the equation (eqn (3), see Methods):

Figure 4A illustrates traces of DIDS-sensitive currents obtained from a same cell bathed in two different [Na⁺]_o (3 and 9 mm). In these experiments, we also confirmed that the cells displayed little, if any, Cl⁻ current as mentioned before (not shown) and found that increase of [Na⁺]_o from 3 to 9 mm caused a shift of the reversal potential towards a negative potential. The mean values of the reversal potential shift and the corresponding q were 24.8 ± 2.6 mV (n = 4) and 2.2 ± 0.1 (n = 4) (Fig. 4B), respectively. We also performed additional experiments using a pipette solution containing 60 mm Na⁺ and 25 mm HCO₃⁻, and obtained the mean values of the reversal potential shift of 26.3 ± 2.7 mV (n = 4), corresponding to a q-value of 2.1 ± 0.1 (n = 4).

Figure 4. Transport stoichiometry.

A, representative traces of DIDS (0.5 mm)-sensitive currents obtained from a same BPA cell bathed in a HCO₃⁻ (25 mm)-containing solution having 3 or 9 mm[Na⁺]_o. The pipette and bath solutions were the same as described in Fig. 3A. B, the shift of the reversal potential of DIDS-sensitive current induced by increasing [Na⁺]_o from 3 mm to 9 mm, and the corresponding value of stoichiometry (Na⁺: HCO₃⁻= 1: q) calculated from the following equation: ΔE_rev=RT/F(q− 1) ln ([Na⁺]_o/[Na⁺]_o). Values are means ±s.e.m. of four experiments.

Pharmacological properties

To further characterize the inhibitory action of DIDS on the I_{Na⁺/HCO3⁻}, we studied the concentration- and the voltage-dependence of the DIDS inhibition. In these experiments, we examined the effect of [Cl⁻]_o replacement with glutamate on the whole-cell current and ensured that the cells showed little, if any, detectable Cl⁻ conductance (i.e. less than 15 pA at 0 mV). Figure 5A shows I–V relations obtained from a cell in the absence and the presence of DIDS (0.01–1 mm) and demonstrates that DIDS inhibits the I_{Na⁺/HCO3⁻} in a concentration-dependent manner. DIDS (0.01, 0.1, 0.5 and 1 mm) at +50 mV reduced the currents to 89.8 ± 1.8%, 48.1 ± 4.1%, 12.2 ± 3.4%, and 4.8 ± 3.5% of the control values, respectively (Figs 5B and C), inhibition being reversible (97.2 ± 3.5% of the control level upon washout of DIDS) (Fig. 5C). Interestingly, we also found that DIDS block was voltage dependent (Fig. 5D). The data at different membrane potentials were fitted to eqn (1) (see Methods), and the apparent half-maximal inhibition (K_i) and pseudo Hill coefficient were determined at each potential (Fig. 5E). The following K_i values and pseudo Hill coefficient were obtained, e.g. 0.41 ± 0.11 mm and 0.99 ± 0.07 at −40 mV, 0.27 ± 0.08 mm and 1.03 ± 0.05 at −10 mV, 0.16 ± 0.04 mm and 1.11 ± 0.04 at +20 mV, 0.11 ± 0.03 mm and 1.16 ± 0.06 at +50 mV (n = 9). Figure 5E (inset) also shows the plot of K_i values as a function of membrane potential. The K_i decreased e-fold per 67.9 ± 2.4 mV depolarization (n = 9).

Figure 5. Pharmacological properties of I_{Na⁺/HCO3⁻}.

A, effect of external DIDS on I_{Na⁺/HCO3⁻}. Typical I–V relationships obtained in the absence (control; bath solution B) or presence of 0.01, 0.1, 0.5, 1.0 mm DIDS are shown. The pipette was filled with a NMDG-glutamate-rich solution (pipette solution B). B, summary of I–V relationships. Values are means ±s.e.m. of nine independent experiments. Error bars representing s.e.m. were omitted when so small as to lie within symbols. (•, control; ○, 0.01 mm; ▴, 0.1 mm; ▵, 0.5 mm; ▪, 1 mm) C, dose–response relationship for DIDS block of I_{Na⁺/HCO3⁻} at +50 mV. Data are from Fig. 5B (n = 9). *P < 0.001. D, voltage dependence of the DIDS block. Fractional currents in the presence of 0.1 mm DIDS at −40, −10, +20, and +50 mV are shown. Data are from Fig. 5B (n = 9). E, plot of fractional current as a function of DIDS concentration at a membrane potential of −40 (•), −10 (▵), +20 (○), or +50 mV (▴). Each point represents the mean ±s.e.m. of nine experiments. The lines are fits to the Hill equation (eqn (1)). Error bars representing s.e.m. were omitted when so small as to lie within symbols. Inset, voltage dependence of K_i values for DIDS block. K_i values are plotted on a semilogarithmic scale. The continuous line shows the linear regression fit to the data. Values are means ±s.e.m. of nine experiments. F, effects of anion channel blockers (0.1 mm of DIDS (n = 9), phloretin (n = 5), niflumic acid (NFA) (n = 5), NPPB (n = 5), glybenclamide (n = 3), and DPC (n = 5)) on I_{Na⁺/HCO3⁻} at 0 mV. Both the pipette and bath solutions were the same as shown in Fig. 5A*P < 0.05. **P < 0.001.

We also extended our pharmacological characterization of the native I_{Na⁺/HCO3⁻} in BPA cells. Because anion channel blockers, including niflumic acid, NPPB, glybenclamide and DPC, which are less specific compared to cationic channel blockers (Jentsch et al. 2002) have been often used for functional experiments to elucidate the role of the anion channel in transepithelial HCO₃⁻ transport, we wanted to test whether they would affect an electrogenic Na⁺–HCO₃⁻ cotransporter function. The results are summarized in Fig. 5F. Phloretin, niflumic acid, NPPB, and glybenclamide (0.1 mm) significantly decreased the I_{Na⁺/HCO3⁻} to 54.5 ± 1.0% (n = 5, P < 0.001), 71.1 ± 0.8% (n = 5, P < 0.001), 70.6 ± 1.3% (n = 5, P < 0.001), and 85.7 ± 1.5% (n = 3, P < 0.02) of the control values at 0 mV (Fig. 5F), and to 57.0 ± 4.8% (n = 5, P < 0.001), 78.7 ± 1.8% (n = 5, P < 0.001), 79.1 ± 2.0% (n = 5, P < 0.001), and 84.5 ± 3.3% (n = 3, P < 0.05) of the control values at −40 mV, respectively. DPC also reduced the currents to 87.5 ± 1.4% (n = 5, P < 0.001) of the control values at 0 mV, but not at −40 mV (94.3 ± 2.2% (n = 5, P= 0.064) of the control). Vehicle alone had no effect on the currents (data not shown, n = 4).

Molecular characterization of NBCe from bovine parotid cells

Based on the functional data described above, we subsequently hypothesized that a variant of NBCe1 such as NBCe1-A and NBCe1-B may be present in BPA cells. In line with the hypothesis, RT-PCR with mRNA from bovine parotid cells yielded a clear band for NBCe1-B, but not for NBCe1-A (Fig. 6A). The cDNA of NBCe1-B contained a 3240 bp open reading frame (ORF) encoding a 1079-amino acid protein. The sequence comparisons of ORF showed the cDNA and corresponding amino acid sequence to be 99.8% (eight nucleotide differences) and 99.7% (three amino acid differences; 421P to H, 735P to S, 1027V to L) identical to those of the bovine corneal NBC (NBCe1-B) (AF308160, Sun et al. 2000) (Fig. 6B).

Figure 6. Expression of NBCe1-B mRNA in BPA cells.

A, RT-PCR analysis of NBCe1-A, NBCe1-B, and β-actin. mRNA transcripts detected by gel electrophoresis. See Methods for details of the primers and experimental conditions. Total RNA isolated from rat kidney was also used as a template as a positive control for NBCe1-A. Lane M: size markers (1 kbp DNA ladder, Life Technologies, Inc.). RT, reverse transcription. B, schematic of nucleotide and amino acid sequences of the bovine parotid NBCe1-B coding region. Portions of the sequences that are identical to those of the bovine corneal NBCe1-B (AF308160) are not shown. The regions where these NBCe1-B differ are indicated in bold.

We also transfected HEK293 cells with NBCe1-B cloned from bovine parotid cells and compared the electrophysiological properties of the expressed NBCe1-B currents with those of the native currents under similar ionic conditions. We first confirmed that HEK293 cells transfected with NBCe1-B, but not mock-transfected cells (n = 40, data not shown) displayed a DIDS-sensitive, Na⁺-dependent HCO₃⁻ current, whose kinetics were similar to those observed for native currents (Fig. 7A), and that NBCe1-B mRNA was expressed in the transfected cells (Fig. 7B). Unlike BPA cells where the incidence of the native cotransporter current was 100% irrespective of the pipette solutions employed, both the magnitude and incidence of the NBCe1-B currents were variable and dependent upon the pipette solution used. When the transfected cells were dialysed with the pipette solutions A, C, and D, the incidence of the currents was 30.1% (82 out of 272 cells tested), 0.6% (1 out of 154 cells tested) and 32.3% (20 out of 62 cells tested), respectively. We also noticed that stability of the currents was variable, so that the currents often rapidly ran down. The reason for these phenomena is still unclear at present. Therefore, the following results were obtained in the selected cells that expressed the cotransporter current, which was stable and ranged between 13.7 and 153.5 pA at +50 mV.

Figure 7. Biophysical and pharmacological properties of heterologously expressed NBCe1-B current.

A, traces of DIDS-sensitive Na⁺–HCO₃⁻-dependent whole-cell currents evoked by voltage steps for 400 ms from a holding potential of −78 mV between −98 mV and +52 mV in 10 mV intervals, which were obtained from a HEK293 cell transfected with bNBCe1-B. Experimental conditions were identical to those described in Fig. 1Ca–c. B, RT-PCR analysis of bNBCe1-B. HEK293 cells transfected with bNBCe1-B (Construct: bNBCe1-B), but not mock transfected cells (Construct: mock) express the mRNA for bNBCe1-B. See Methods for details of the primers and experimental conditions. C, monovalent cation dependence of bNBCe1-B current. Test cations were Cs⁺, Li⁺, NMDG, choline (a) and K⁺ (b). Solutions and pulse protocol were the same as described in Fig 2A and C. D, [Na⁺]_o dependence of bNBCe1-B. Tracings of instantaneous I–V relationships obtained from a cell in the presence of different [Na⁺]_o (145, 100, 25, 9, 3, 0 mm) are shown. The pipette was filled with the pipette solution D. Bath solutions and pulse protocol were the same as described in Fig. 3A. Inset, voltage dependence of K_m for [Na⁺]_o. The continuous line shows linear regression fit to the data. Values are means ±s.e.m. of five experiments. E, dose–response relationship for DIDS block of bNBCe1-B currents at +50 mV. The pipette was filled with the pipette solution A, and bath solutions were the same as described in Fig. 5A. Values are means ±s.e.m. of seven experiments. *P < 0.05; **P < 0.001. Inset, voltage dependence of K_i for DIDS block. K_i values were plotted on a semilogarithmic scale. The continuous line shows the linear regression fit to the data. Values are means ±s.e.m. of seven experiments. F, effects of anion channel blockers (0.1 mm of DIDS (n = 7), Niflumic acid (NFA) (n = 5), NPPB (n = 4), Glybendclamide (n = 3) and DPC (n = 5)) on bNBCe1-B currents at 0 mV. The pipette was filled with the pipette solution A, and the bath solution was the same as shown in Fig. 5F. *P < 0.05, **P < 0.01. Phloretin was not tested, because it was found to increase a background current, probably attributable to a cation conductance (data not shown).

The basic electrophysiological properties of NBCe1-B were quantitatively similar to those of native NBC in BPA cells (Fig. 7C–F), although we could not estimate the stoichiometry of the cloned NBCe1-B current, due to the problems mentioned above. When expressed in HEK293 cells, the NBCe1-B currents were not supported by Li⁺, Cs⁺, NMDG, choline or K⁺ (Fig. 7Ca and b). Given that HEK293 cells are likely to endogenously express a cation conductance, it is reasonable to conclude that the relative selectivity sequence was Na⁺ (by definition, 100%) ≫ K⁺ (0.5 ± 2.3%) ≈ choline (by definition, 0%) ≈ Li⁺ (−2.1 ± 2.2%) ≈ NMDG (−6.8 ± 2.4%) ≈ Cs⁺ (−16.1 ± 5.4%) (n = 6). Figure 7D shows the [Na⁺]_o dependence of the NBCe1-B current. The calculated apparent K_m for [Na⁺]_o at +50 mV was 21.1 ± 1.2 mm (n = 5), the value being independent of the membrane voltage (Fig. 7D inset). These results suggest that extracellular Na⁺ binding may be insensitive to membrane voltage. We confirmed that DIDS reversibly inhibited NBCe1-B currents in a concentration- and voltage-dependent manner (Fig. 7E). The K_i decreased e-fold per 54.3 ± 7.4 mV depolarization, K_i and Hill coefficient at 0 mV for the DIDS inhibition being 0.18 ± 0.05 mm and 0.97 ± 0.05, respectively (Fig. 7E inset). Other anion channel blockers (0.1 mm of niflumic acid, NPPB, and glybenclamide) also had significant inhibitory effects on the NBCe1-B current, so that they reduced the current to 77.9 ± 5.3% (n = 5), 85.6 ± 2.1% (n = 4), and 81.8 ± 2.9% (n = 3) of the control values at 0 mV, respectively (Fig. 7F), and to 88.2 ± 3.3%, 91.5 ± 1.8%, and 85.6 ± 1.5% of the control values at −40 mV, respectively. However, DPC (0.1 mm) had no significant effects on the current (90.5 ± 4.3% and 92.2 ± 4.2% of the control values (n = 5) at 0 mV and −40 mV, respectively). Vehicle alone had no effect on the current (n = 5). The biophysical and pharmacological properties of the NBCe1-B current and the native NBC current are summarized in Table 2.

Table 2.

Comparison of the biophysical and pharmacological properties of native NBCe current and NBCe1-B current

	Native NBCe current	NBCe1-B current
Relative cation selectivity	Na+≫ Li+ > K+ > choline ≈ NMDG ≈ Cs+	Na+≫ K+≈ choline ≈ Li+≈ NMDG ≈ Cs+
Na+o dependency	Yes	Yes
Apparent Km of Na+o at +50 mV	20.8 ± 1.0 mm	21.1 ± 1.2 mm
Voltage dependency of Km	No	No
Stoichiometry (q)	2.2 ± 0.1	—
DIDS block	Yes	Yes
Ki(0) of DIDS	0.22 ± 0.06 mm	0.18 ± 0.05 mm
Hill coefficient at 0 mV	1.05 ± 0.04	0.97 ± 0.05
Ki decrease e-fold per	67.9 ± 2.4 mV depolarization	54.3 ± 7.4 mV depolarization
Blocking of other anion channel
blockers (0.1 mm) at –40 mV
Phloretin	Yes	—
NFA	Yes	Yes
NPPB	Yes	Yes
Glybenclamide	Yes	Yes
DPC	No	No

Values are means ±s.e.m. of 4–9 experiments.

Role of NBC current in setting membrane potential of unstimulated BPA cells

Using a method of non-invasive measurement of membrane potential in the cell-attached patch configuration originally described for human T lymphocytes (Verheugen et al. 1995), we have previously shown that a Ba²⁺-sensitive K⁺ conductance is, at least in part, involved in setting resting membrane potential of BPA cells bathed in a nominally HCO₃⁻/CO₂-free solution (Hayashi et al. 2003). However, given that the BPA cells express a functional electrogenic Na⁺–HCO₃⁻ cotransporter, which can generate a significant outward current at a membrane potential more positive to the reversal potential of NBC current, we hypothesized that the cotransporter, if functional, would also play a role in setting the membrane potential (V_m) at rest in intact cells (i.e. non-dialysed cells) under quasi-physiological conditions (i.e. in the presence of 25 mm HCO₃⁻/5% CO₂). To test this hypothesis, we measured, with high K⁺ solution in the pipette, the reversal potential (E_rev) of a K⁺-selective single-channel current (i.e. maxi-K⁺ channel current) in cell-attached patches. Even though the intracellular K⁺ concentration ([K⁺]_i), which is not available in BPA cells would not be the same as that of pipette solution, the shift of E_rev should give a quantitative measure of change in V_m, assuming that [K⁺]_i is constant. As shown in Fig. 8A and B, removal of Na⁺ from the bathing solution in the continued presence of 25 mm HCO₃⁻/5% CO₂ caused a reversible depolarization of V_m by 13.0 ± 1.3 mV (n = 11), an effect that was abolished by addition of 1 mm DIDS to the bath (Fig. 8B). However, removing Na⁺_o in the absence of 25 mm HCO₃⁻/5% CO₂ induced only a small change of V_m (Fig. 8B). Furthermore, adding 25 mm HCO₃⁻/5% CO₂ caused a significant hyperpolarization of V_m by 11.3 ± 3.2 mV (n = 6) in the continued presence of Na⁺_o, an effect that was also abolished by DIDS (1 mm) (Fig. 8C). We also performed additional experiments to examine whether V_m would change even in the presence of a K⁺ channel blocker Ba²⁺ (1 mm) in the bathing solution. We first confirmed the previous finding (Hayashi et al. 2003) that extracellular Ba²⁺ depolarizes V_m in a nominally HCO₃⁻/CO₂-free bath solution (not shown). In the continued presence of Ba²⁺, switching to a 25 mm HCO₃⁻/5% CO₂-containing solution caused a hyperpolarization of V_m by 23.5 ± 3.4 mV (n = 5), and removal of Na⁺ from the bathing solution in the continued presence of 25 mm HCO₃⁻/5% CO₂ induced a reversible depolarization of V_m by 38.9 ± 7.2 mV (n = 5), respectively. These results together suggest that the cotransporter current can contribute to V_m in BPA cells in the presence of 25 mm HCO₃⁻/5% CO₂.

Figure 8. Role of the NBC current in setting membrane potential of unstimulated BPA cells.

Aa, typical traces of single maxi-K⁺ channel current in a cell-attached patch. The patch pipette was filled with the KCl-rich (150 mm K⁺) solution (pipette solution E) and the bath solution was the NaCl-rich solution containing 25 mm HCO₃⁻ (bath solution L: Control and After control) or NMDG-Cl-rich solution containing 25 mm HCO₃⁻ (bath solution M: Na⁺free). Voltage (−V_p) was +43 (control and after control) or +45 mV (Na⁺ free). Atropine (1 µm) was included in the bath solutions. Ab, I–V relations of maxi-K⁺ channel obtained from the patch shown in (a) (•, control; □, Na⁺ free; ▵, after control). B, the reversal potential shifts induced by replacement of Na⁺ with NMDG and choline in the absence (left; bath solution A and N) or the presence of 25 mm HCO₃⁻ (middle; bath solution L and M) or the presence of 25 mm HCO₃⁻ and 1 mm DIDS (right). Values are means ±s.e.m. of 5 or 11 experiments. **P < 0.001. C, the reversal potential shifts induced by addition of 25 mm HCO₃⁻ to the bath solution in the continued presence of Na⁺_o (bath solution A and L) without (left) or with 1 mm DIDS (right). Values are means ±s.e.m. of six experiments. *P < 0.05.

Discussion

The present study has provided several lines of evidence that a functional electrogenic Na⁺–HCO₃⁻ cotransporter is expressed in bovine parotid acinar (BPA) cells. First, whole-cell patch-clamp experiments have demonstrated that acutely dissociated BPA cells indeed express a membrane current (I_{Na⁺/HCO3⁻}), which is evoked by introducing either HCO₃⁻ or Na⁺ into the bath solution containing Na⁺ or HCO₃⁻, respectively, in 100% of the cells tested and is sensitive to DIDS. However, control experiments showed that such a current could not be elicited by Cl⁻_o, even in the presence of Na⁺_o. The shift of the reversal potential of I_{Na⁺/HCO3⁻} is also consistent with that predicted from a knowledge of the transmembrane HCO₃⁻ and Na⁺ gradients (i.e. by a 2 HCO₃⁻: 1 Na⁺ stoichiometry). Second, RT-PCR analysis showed that BPA cells express transcripts of NBCe1-B. Furthermore, heterologous expression in HEK293 cells of NBCe1-B (bNBCe1-B) cloned from BPA cells induced a DIDS-sensitive, Na⁺- and HCO₃⁻-dependent current, whose biophysical and pharmacological properties resemble those of the native currents in many respects. Finally, cell-attached patch experiments indicated that a manoeuvre that is expected to change NBCe activity was able to affect V_m of BPA cells. To the best of our knowledge, the present study is the first to demonstrate and characterize a NBCe current naturally expressed not only in mammalian salivary glands, but also in exocrine glands.

In the present study, the native parotid NBCe current amplitude was approximately 250 pA at 0 mV (see Figs 1, 2, 3), which yields 8.3 pA pF⁻¹ for a 30 pF cell. This current amplitude appears to be much larger than those previously reported for native NBCe-mediated whole-cell outward current in freshly dissociated rat ventricular myocytes (about 0.52 pA pF⁻¹ at 0 mV; Yamamoto et al. 2005) and salamander retinal Müller cells (about 7 pA cell⁻¹ at 0 mV; Newman, 1991). The native current is even larger than that observed in HEK293 cells expressing NBCe1-B (about 30 pA at 0 mV, which yields 2 pA pF⁻¹ for a 15 pF cell, see Fig. 7C and D) under similar experimental conditions, and may be also comparable to that reported for rNBCe1-A expressed in Xenopus oocytes (approximately several hundreds of nanoamps at 0 mV (Sciortino & Romero, 1999) with a capacitance of 100 nF (Stühmer, 1992)). Therefore, BPA cells must express the NBCe with a high overall activity that should be proportional to the product of the number of transporters and the overall transport rate of a single transporter molecule.

The present results indicate that the native parotid NBCe current is specific for Na⁺ over K⁺, Cs⁺, choline and NMDG, suggesting that Na⁺ is a physiological substrate. The results also indicate that Li⁺ can only partially substitute for Na⁺ on the native current (i.e. about 14% of the Na⁺). However, unlike the native current, bNBCe1-B when expressed in HEK293 cells could not transport Li⁺. The reason for this apparent discrepancy is unclear at present. Li⁺ transport through bNBCe1-B could be affected by the presence or absence of an as yet unknown cellular factor in BPA cells. From this point of view, it is interesting to note that such a variation has been also reported for hNBCe1-A and rNBCe1-A, which are 97% homologous (Amlal et al. 1998; Sciortino & Romero, 1999): Li⁺ elicits only about 3% of the Na⁺-induced current in Xenopus oocytes expressing rNBCe1-A (Sciortino & Romero, 1999), whereas Li⁺ supports 25% of the DIDS-sensitive, HCO₃⁻-dependent intracellular pH recovery from an acid load in the presence of Na⁺_o in HEK293 cells expressing hNBCe1-A (Amlal et al. 1998). Li⁺ is also suggested to be transported as well as Na⁺ by NBCe endogenously expressed in cultured corneal endothelial cells (Jentsch et al. 1984) and freshly isolated leech glial cells (Munsch & Deitmer, 1994). Further studies are indeed necessary to address the question as to whether diverse Li⁺ selectivity observed in various experimental systems is attributable to a different molecular basis or cellular environment, where a NBCe is expressed.

The stoichiometry of the cotransport, in conjunction with the electrochemical gradients for Na⁺ and HCO₃⁻, is critical for the transport direction, and thus its function (see for reviews, Gross & Kurtz, 2002; Kurtz et al. 2004). We determined the Na⁺ and HCO₃⁻ stoichiometry of the native parotid cotransporter from the shift of the reversal potential of DIDS-sensitive whole-cell current upon changing [Na⁺]_o between 3 mm and 9 mm. The results are consistent with an apparent stoichiometry that two HCO₃⁻ ions may accompany the influx of one Na⁺ ion, as described for a basolateral NBCe current endogenously expressed in a mouse pancreatic duct cell line (Gross et al. 2001). However, it remains unclear whether HCO₃⁻ is the actual transported substrate on the native NBCe, because this stoichiometry is also equivalent to alternative models involving cotransport of one Na⁺ with one CO₃²⁻ or a single NaCO₃⁻ ion pair (see for recent reviews, Kurtz et al. 2004; Romero et al. 2004).

It has been well established that stilbene derivatives, including DIDS, block the cotransporter activity (Romero & Boron, 1999). We have shown here that extracellular application of DIDS blocks both the native and bNBCe1-B currents in a similar concentration-dependent manner, the blockade being fast and reversible under the present experimental conditions. Intriguingly, our data also indicate that the DIDS blockade is voltage-dependent. Given that DIDS is a negatively charged molecule, the block may be interpreted as suggesting that the DIDS-binding site is located at the transmembrane electric field and might be close to a substrate-binding site or permeant path. In this regard, it is interesting to note that NBCe1-B cloned from BPA cells possess a predicted DIDS-binding motif (KMIK (603–606)) around the putative transmembrane-5 (Romero & Boron, 1999). We also found that the parotid NBCe current is significantly reduced by phloretin, niflumic acid, NPPB, and glybenclamide at 0.1 mm. The blockers' profile, which was also similar to that observed for the bNBCe1-B current, stress the need to interpret data with much caution when they are used to probe the role of an anion channel in physiological functions of the cells that express NBCe.

The results in the present cell-attached patch experiments leave little doubt that NBCe activity can also contribute to V_m of unstimulated BPA cells in a HCO₃⁻/CO₂-containing solution. Taken together with a previous study showing that a strong inwardly rectifying K⁺ current (probably mediated by Kir2.1) probably contributes to setting a resting V_m of unstimulated BPA cells bathed in a nominally HCO₃⁻/CO₂-free, Hepes-buffered solution (Hayashi et al. 2003), it is tempting to hypothesize that the NBCe with an equivalent stoichiometry of 2 HCO₃⁻: 1 Na⁺ may transport HCO₃⁻ in an inward direction, generating an outward current that tends to hyperpolarize the cell in cooperation with the K⁺ channel in BPA cells. Our observations that the V_m changes were more pronounced in the presence of external Ba²⁺ (1 mm) support this hypothesis. Similar observation has been made in leech glial cells, in which the cotransport in situ is suggested to possess an equivalent stoichiometry of 2 HCO₃⁻: 1 Na⁺ (Deitmer & Schlue, 1989).

Assuming that the native NBCe is localized in the basolateral membrane of BPA cells, as shown for NBCe1 in various epithelial cells so far studied (see for a recent review, Romero et al. 2004), including salivary acinar cells (Roussa et al. 1999; Park et al. 2002), an attractive possibility is, by analogy with a proposed model for the muscarinically stimulated HCO₃⁻ secretion in bovine and sheep parotid glands (Lee & Turner, 1993; Poronnik et al. 1995; Steward et al. 1996), that it might be also involved in resting HCO₃⁻ secretion not only by mediating basolateral HCO₃⁻ influx, but also by hyperpolarizing the cell to provide the driving force for eventual HCO₃⁻ efflux across the luminal membrane (via an anion channel) in BPA cells. Even though the basolateral HCO₃⁻ accumulation mediated by Na⁺–H⁺ exchange that has been present in unstimulated sheep parotid acinar cells (Poronnik et al.1993, 1995; Steward et al. 1996) might also occur in parallel in BPA cells, the electrogenicity of the NBCe could be supportive for the secretion. In this regard, it is worth mentioning that an electrical coupling between electrogenic efflux of anions across the apical membrane and electrogenic HCO₃⁻ uptake at the basolateral membrane has been proposed to play a role in secretin-stimulated HCO₃⁻ secretion in pancreatic ducts (Ishiguro et al. 1996; Shumaker et al. 1999), a model accounting for the secretion in the absence of any significant change in intracellular pH (Ishiguro et al. 1996).