The stoichiometry of the electrogenic sodium bicarbonate cotransporter pNBC1 in mouse pancreatic duct cells is 2 HCO3:1 Na+


Corresponding author E. Gross: VA Medical Center, MS151(W), 10701 East Boulevard, Cleveland, OH 44106, USA. Email:


  • 1The electrogenic sodium bicarbonate cotransporter pNBC1 is believed to play a major role in the secretion of bicarbonate by pancreatic duct cells, by transporting bicarbonate into the cell across the basolateral membrane. Thermodynamics predict that this function can be achieved only if the reversal potential of the cotransporter is negative to the cell's membrane potential, or equivalently that the HCO3:Na+ stoichiometry is not larger then 2:1. However, there are no data available on either the reversal potential or the HCO3:Na+ stoichiometry of pNBC1 in pancreatic cells.
  • 2We studied pNBC1 function in mouse pancreatic duct cells. RT-PCR analysis of total RNA revealed that these cells contain the message for pNBC1, but not for kNBC1, NBC2 or NBC3.
  • 3To measure cotransporter activity, mouse pancreatic duct cells were grown to confluence on a porous substrate, mounted in an Ussing chamber, and the apical plasma membrane permeabilized with amphotericin B. Ion flux through pNBC1 was achieved by applying Na+ concentration gradients across the basolateral plasma membrane. The current through the cotransporter was isolated as the difference current due to the reversible inhibitor dinitrostilbene disulfonate (DNDS).
  • 4Current-voltage relationships for the cotransporter, measured at three different Na+ concentration gradients, were linear over a range of about 100 mV. The reversal potential data, obtained from these current-voltage relationships, all corresponded to a 2 HCO3:1 Na+ stoichiometry.
  • 5The data indicate that pNBC1 is functionally expressed in mouse pancreatic duct cells. The cotransporter operates with a 2 HCO3:1 Na+ stoichiometry in these cells, and mediates the transport of bicarbonate into the cell across the basolateral membrane.

One of the major functions of the exocrine pancreas is the secretion of large amounts of bicarbonate into the pancreatic juice. The high concentration of bicarbonate is required to dissolve digestive enzymes, secreted by the acinar glands, and to protect the duodenum and small intestine from the gastric acid. In humans, pigs and guinea-pigs, the bicarbonate concentration in the pancreatic duct can increase to 150 mm (Padfield et al. 1989) – about 5-fold greater than the bicarbonate concentration in the blood. This implies that bicarbonate accumulates in the pancreatic duct cells against its concentration gradient across the basolateral membrane. One mechanism for bicarbonate accumulation in these cells is by diffusion of CO2 across the basolateral membrane, and its subsequent conversion into HCO3 and H+ by carbonic anhydrase. However, experiments in rabbits (Swanson & Solomon, 1975; Kuijpers et al. 1984), cats (Case et al. 1969, 1970), dogs (Banks & Sum, 1971), pigs (Raeder & Mathisen, 1982) and humans (Dyck et al. 1972; Anand et al. 1994) have indicated that inhibition of carbonic anhydrase with acetazolamide can only decrease bicarbonate secretion by ≈50 %. These observations have led to the suggestion that the concentrative uptake of bicarbonate across the basolateral membrane in these cells is coupled to the transport of sodium down its concentration gradient.

Micropuncture experiments in the cat provided the first evidence that bicarbonate secretion in the pancreatic duct is coupled to chloride uptake across the apical membrane of duct cells (Case et al. 1969; Lightwood & Reber, 1977). This transport mode is consistent with the existence of a Cl-HCO3 exchanger (AE) in the apical membrane of these cells, although the exact isoform(s) of AE has not been identified. Chloride exits the cell via an apical cystic fibrosis transmembrane conductance regulator (CFTR) (Gray et al. 1989; Crawford et al. 1991; Marino et al. 1991; Trezise & Buchwald, 1991). More recent studies with guinea-pig pancreatic ducts have demonstrated Cl independence of apical bicarbonate exit (Ishiguro et al. 1996).

Functional evidence for the existence of basolateral sodium-dependent bicarbonate transport in these cells comes from functional studies which have demonstrated that intracellular pH recovery is dependent on basolateral Na+ and HCO3 and can be inhibited by DIDS (Zhao et al. 1994; Villanger et al. 1995; Ishiguro et al. 1996; de Ondarza & Hootman, 1997). These results are consistent with either the presence of a Na+-dependent Cl-HCO3 exchanger or a sodium bicarbonate cotransporter. Additional support for the existence of a sodium bicarbonate cotransporter in the pancreas has recently been provided by the cloning and characterization of the human pancreatic variant of the NBC1 gene, pNBC1 (Abuladze et al. 1998, 2000). pNBC1 has been localized to both ductal and acinar cells using in situ hybridization (Abuladze et al. 1998). More recent immunocytochemistry experiments have localized NBC1 to the basolateral surface of human duct cells (Marino et al. 1999), although the antibody used in the latter study cannot distinguish between kNBC1 and pNBC1.

As a secondary active, electrogenic, transport system, thermodynamic considerations suggest that the sodium concentration gradient could drive the transport of bicarbonate into the cell via an electrogenic sodium bicarbonate cotransporter only if the reversal potential of the cotransporter is more negative then the resting membrane potential. The reversal potential of the cotransporter is determined by the HCO3:Na+ stoichiometry. Although pNBC1 has been shown to transport Na+ and HCO3 (Abuladze et al. 1998), and is electrogenic (Choi et al. 1999), no data are available on the stoichiometry or reversal potential of pNBC1 in pancreatic duct cells.

In the present study we demonstrate for the first time that pNBC1 in mammalian pancreatic duct cells is electrogenic with a stoichiometry of 2 HCO3:1 Na+. The current generated by the cotransporter in these cells can be blocked by DNDS, and is Na+ dependent and Cl independent. These results suggest that pNBC1 mediates cellular bicarbonate uptake in these cells.


Cell culture

Experiments were carried out with the cell line mPEC1, derived from ductal fragments of an ImmortaMouse (Takacs-Jarrett et al. 2001). Passages 30-35 were used for the present study. Cells were grown on Ethicon-coated (20 % Ethicon in 60 % ethanol) Millicell-CM filters (diameter, 30 mm) in a 1:1 mixture of Dulbecco's Modified Eagle's Medium and Ham's F12, supplemented with 10 ng ml−1 epithelial growth factor, 0.5 mm isobutyl methyl xanthine (IBMX), 30 μg ml−1 penicillin G, 50 μg ml−1 streptomycin sulfate, 50 units ml−1 nystatin, 10 units ml−1 interferon γ, 2 mm glutamine and 2.5 % fetal bovine serum (exocrine medium). Usually, about 6 x 105 cells were seeded and grown to confluence in 5 days.

Pancreatic duct isolation

Six mice were killed by exposure to a rising concentration of CO2 and their pancreases removed and placed in ice-cold Hepes-buffered Ringer solution (HBR). The tissue was cut into small fragments and digested at 37°C in a solution of HBR that contained 0.25 mg ml−1 collagenase I, 0.25 mg ml−1 collagenase IV, 0.1 mg ml−1 DNase and 0.025 mg ml−1 soybean trypsin inhibitor. At 15 min intervals the digest was dispersed with a 10 ml pipette. After 60 min the digest was filtered through 149 μm nylon mesh. The pancreatic fragments retained on the filter were collected and resuspended in a solution of rat tail collagen. After the collagen solution had formed a gel, tissue culture medium (exocrine medium supplemented with 10 % fetal bovine serum) was added and the gels were placed in a tissue culture incubator. During the next 24-48 h the ends of the ductal fragments sealed and the ducts became inflated. At this point the gel was digested with collagenase (1 mg ml−1 each of collagenase I and IV) and the inflated ducts were collected with a glass micropipette.


Filter inserts were mounted vertically in a thermostatically controlled Ussing chamber equipped with gas inlets for CO2 bubbling, and separate reservoirs for the perfusion of apical and basolateral compartments. Electrophysiological measurements were carried out with an epithelial voltage-clamp amplifier (EC825, Warner Instruments, Hamden, CT, USA). Data were digitized at 100 kHz and recorded through an A/D converter (PowerLab/400, AD Instruments, Castle Hill, Australia) on a Pentium PC for further analysis. Data were filtered at 0.5 Hz. Current-voltage relationships were measured by stepping the voltage command from -100 mV to +100 mV in 10 mV steps, using the stimulator utility of the Chart program (AD Instruments). Cells were continuously perfused with modified Ringer solution (for composition of solutions see Table 1), at a rate of about 3 ml min−1. All solutions were kept under constant CO2 pressure throughout the experiment. To measure the cotransporter-related current, cell monolayers were permeabilized with 20 μm apical amphotericin B, from a 100 mm stock solution in DMSO, as described previously (Gross & Hopfer, 1996). Given that H+ readily equilibrates across the permeabilized membrane and assuming the cell membrane is permeable to CO2, the concentration of HCO3, at equilibrium, can be calculated from the Henderson-Hasselbalch relation: [HCO3]= 3 x 10−5x 760 xfx 10(pH -6.1), where f is the partial pressure of CO2(PCO2).

RT-PCR analysis of NBC isoforms in pancreatic duct cells

Total RNA was isolated from mouse cell lines and tissues with the Trizol reagent (Gibco BRL) as per the manufacturer's protocol. Random-primed first strand cDNA was synthesized with AMV reverse transcriptase. PCR analysis was carried out with the Expand Long Template PCR System (Roche), using specific primers for amplifying mouse pNBC1, kNBC1, NBC2 and NBC3. The following conditions were used for PCR: denature 94°C, 30 s; anneal 58°C, 30 s; extend 68°C, 3 min; for 30 cycles. The following primers were used for amplifying pNBC1: sense, 5′-ATGTGTGTGATGAAGAAGAAGTAGAAG-3′; antisense, 5′- GACCGAAGGTTGGATTTCTTG-3′. For amplifying kNBC1 specifically, the following primers were used: sense, 5′-CACTGAAAATGTGGAAGGGAAG-3′; antisense, 5′-GACCGAAGGTTGGATTTCTTG-3′. For NBC2: sense: 5′-CTTCGCACCTCCCATTCCTAAGAG-3′; antisense: 5′-CATCTTCGAAATTTATTGTCACACACTCACAG-3′. For NBC3: sense: 5′-GCAGGCTCAAGGTGTACAACC-3′; antisense: 5′-CTGGTTGGAACAGGGACCTCAG-3′. The NBC2 primers were based on the mouse expressed sequence tag (EST) clone (GenBank accession number AV275974). The NBC3 primers were based on the human NBC3 sequence (GenBank accession number AF047033).


Amphotericin B, Hepes, d-glucose, N-methyl-d-glucamine (NMDG), gluconic acid and all salts were purchased from Sigma Chemical Co. (St Louis, MO, USA). Dinitrostilbene disulfonate (DNDS) was obtained from Pfaltz & Bauer, Inc. (Waterbury, CT, USA).


Experiments were performed at least 4 times. The results for the reversal potential and stoichiometry are presented as means ±s.e.m.


In addition to the pancreatic (pNBC1) and kidney (kNBC1) variants of the NBC1 gene, two other isoforms have recently been cloned: NBC2 from retina (Ishibashi et al. 1998), and a non-electrogenic NBC3 from skeletal muscle (Pushkin et al. 1999a,b). To check which of these clones are present in mPEC1 cells, total RNA was extracted, reverse transcribed and amplified with primers specific for pNBC1, kNBC1, NBC2 or NBC3. As illustrated in Fig. 1, mPEC1 cells express pNBC1, but not any of the other three proteins.

Figure 1.

RT-PCR amplification of mPEC1 cells reveals the message for pNBC1

Total RNA from mPEC1 cells or mouse pancreatic duct fragments was reversed transcribed using AMV reverse transcriptase (RT). Specific regions of pNBC1, kNBC1, NBC2 or NBC3 were amplified using primers as described in the Methods section. M, 1 kb ladder; lane 1 (mPEC1), pNBC1 primers, +RT; lane 2 (mPEC1), pNBC1 primers, -RT; lane 3 (mPEC1), kNBC1 primers, +RT; lane 4 (mouse kidney), kNBC1 primers, +RT; lane 5 (mouse kidney), kNBC1 primers, -RT; lane 6 (mPEC1), NBC2 primers, +RT; lane 7 (mouse testes), NBC2 primers, +RT; lane 8 (mouse testes), NBC2 primers, -RT; lane 9 (mPEC1), NBC3 primers, +RT; lane 10 (mouse heart), NBC3 primers, +RT; lane 11 (mouse heart), NBC3 primers, -RT; lane 12 (mouse pancreas), pNBC1 primers, +RT; lane 13 (mouse pancreas), pNBC1 primers, -RT.

Current-voltage relationships of pNBC1

The ionic permeability of the apical membrane of mPEC1 cells was increased with 20 μm amphotericin B. This manipulation effectively removes the electrical resistance of the apical plasma membrane and renders the membrane permeable to small monovalent ions, but not to those with higher valencies, and remains restricted for several hours to the plasma membrane to which it was applied (Kirk & Dawson, 1983; Backman et al. 1992; Illek et al. 1993; Acevedo, 1994; Gross & Hopfer, 1996). Under these conditions, the electrical properties of the basolateral membrane can be monitored with extracellular electrodes.

Figure 2 shows a typical protocol for collecting current- voltage relationships. After permeabilization of the apical membrane with amphotericin B, a 5-fold sodium concentration gradient (high on the basolateral side) was applied across the basolateral membrane, while the concentration of HCO3 across that membrane was symmetrical. Amphotericin B permeabilizes the membrane to H+, and CO2 equilibrates across the plasma membrane (both apical and basolateral) by free diffusion. Under these circumstances the concentration of bicarbonate in the cytosolic face of the basolateral membrane is clamped to that of the apical compartment. Transepithelial voltage was then stepped from -100 to +100 mV in 10 mV steps and the resulting current at each voltage measured (continuous line in Fig. 2a). The NBC inhibitor DNDS (Coppola & Fromter, 1994; Gross & Hopfer, 1996; Devor et al. 1999) was then added to the basolateral solution for 10 min before a second current- voltage relationship was collected (dotted line in Fig. 2a). The DNDS-sensitive current, i.e. the net current through pNBC1, is the difference between the current measured in the presence of DNDS and that measured in the absence of DNDS.

Figure 2.

Bicarbonate dependence of pNBC1 current

A 5-fold sodium concentration gradient was applied across the cell monolayer, which had been apically permeabilized with 20 μm amphotericin B, in the presence of CO2/bicarbonate (A and B) by perfusing the basolateral compartment with solution 50Na and the apical compartment with solution 10Na. A, voltage pulse protocol used to collect I-V relationships in the absence (continuous line) and presence (dotted line) of 2 mm DNDS. Voltage was stepped from -100 to +100 mV in 10 mV steps. The current at each voltage was recorded for 5 s. B, I-V relationships in the absence (○) and presence (○) of DNDS obtained by averaging the current at each voltage over 5 s. C, I-V relationships in the absence (○) and presence (○) of DNDS for a 5-fold sodium concentration gradient but in the nominal absence of CO2/bicarbonate.

Figure 2B shows the current-voltage relationships in the absence and presence of DNDS. The current at each voltage was obtained from the voltage pulse protocol shown in Fig. 2A, by averaging the current at each voltage step for 5 s.

To test whether the DNDS-sensitive current depends on the presence of bicarbonate, we repeated the experiment, this time perfusing the cells with the same solutions as in Fig. 2B but in the nominal absence of CO2/HCO3. As can be seen in Fig. 2C, withdrawal of HCO3 from the solutions abolished the DNDS-sensitive component of the current.

There was some concern as to whether addition of amphotericin B to the cell monolayers might destroy the polarity of the epithelial monolayer by disrupting the tight junction integrity. We thus repeated the experiment by perfusing the cells with Na+- and K+- containing solution (solution 4K in Table 1) and measured the activity of the Na+,K+-ATPase, a functional marker of the basolateral membrane, following permeabilization with amphotericin B. The Na+,K+-ATPase is a widely accepted basolateral marker in epithelial transport studies, and should thus co-localize with pNBC1 on the basolateral membrane. Figure 3 shows the effect of apical application of 20 μm amphotericin B. The permeabilization process can be conveniently followed by the short-circuit current (ISC).

After permeabilization, ISC (equivalent to cation flux from the apical to basal compartment) increased steeply and then levelled off. Based on several criteria, the steady-state current largely represents Na+,K+-ATPase turnover. (1) The current could be reversibly inhibited by replacement of K+ with NMDG+ in both apical and basal solution (not shown). (2) It could also be reversibly inhibited by 1 mm basal ouabain (Fig. 3). (3) The current also depended on the presence of Na+ in the perfusion solutions (data not shown). These results demonstrate that the amphotericin-treated monolayers retained a substantial degree of cellular and monolayer integrity; however, the apical membrane became permeable to monovalent ions and thus electrically conductive.

Figure 3.

Polarized expression of the basolateral marker Na+,K+-ATPase in apically permeabilized pancreatic duct cells

Cells were mounted in an Ussing chamber and the transepithelial voltage was clamped to zero. Cells were perfused initially with a solution containing K+ and Na+ (solution 4K in Table 1). At the indicated time the apical surface was permeabilized by adding 20 μm amphotericin B, which resulted in an increase in short-circuit current, ISC. The increased ISC following apical application of amphotericin B reflects the activity of the Na+,K+-ATPase, as demonstrated by a decrease in ISC upon basolateral application of 1 mm of the enzyme's inhibitor ouabain, or upon removal of K+ (not shown).

To check whether any primary active transporter systems (i.e. ATPases) might contribute to the measured currents in our experiments, we also measured the DNDS-sensitive current-voltage relationship under symmetrical substrate concentrations, by perfusing both sides with solution 20Na (Table 1) containing 20 mm Na+ and 9 mm HCO3. The results of this experiment are depicted in Fig. 4 (⋄). As can be seen from the figure, the current- voltage relationship of pNBC1 under these symmetrical conditions goes through the origin (0,0), arguing against the contribution of any active transport system to the measured currents.

Figure 4.

I-V relationships for pNBC1 in apically permeabilized monolayers

Cell membranes were permeabilized with 20 μm amphotericin B. ΔI is the DNDS-sensitive current. The potential of the basal compartment was taken as zero. The ratio of apical to basolateral Na+ concentration (AP/BL) was as follows: ▪, 10 mm/80 mm (solutions 10Na/80Na); ○, 10/50 (10Na/50Na); ▴, 10/20 (10Na/20Na); ⋄, 20/20 (20Na/20Na); ▵, 20/10; ○, 50/10; and □, 80/10. The reversal potentials for the different gradients were evaluated graphically from the intersection of the lines with the X-axis and are tabulated in Table 2.

The stoichiometry of the Na+-HCO3 cotransporter

The HCO3:Na+ stoichiometry of pNBC1 was calculated from the reversal potential (Vrev) and eqn (1), as described previously (Gross & Hopfer, 1996):

display math(1)

where n is the number of bicarbonate anions cotransported with each sodium cation, and the subscripts i and o represent intra- and extracellular concentrations of the indicated ion. R, T and F have their usual meanings.

The reversal potential

For a symmetrical HCO3 concentration, the reversal potential depends logarithmically on the magnitude and direction of the Na+ concentration gradient. We, thus, measured Vrev for pNBC1 in pancreatic duct cells from its current-voltage relationship for three different sodium concentration gradients (Fig. 4 and Table 2). Figure 4 shows the change in current (ΔI) versus voltage for apically permeabilized monolayers, with Na+ concentration gradients of 20/10 (solutions 20Na/10Na in Table 1), 50/10 (solutions 50Na/10Na) and 80/10 (solutions 80Na/10Na), applied in either the AP→BL or the BL→AP direction. ΔI is the difference in current in the absence and presence of 2 mm DNDS and V is the transepithelial voltage. The reversal potential was determined from the intercept of the plot with the voltage axis.

Table 2 summarizes the results for the reversal potentials and the calculated stoichiometries (n). The data in Table 2 are consistent with a 2 HCO3:1 Na+ stoichiometry for each of the three concentration gradients tested.


The exocrine pancreas secretes a bicarbonate-rich fluid into the pancreatic duct. The large concentrations of bicarbonate (up to 150 mm in humans), necessary to dissolve digestive enzymes secreted by the acinar glands, is also thought to protect the small intestine from the gastric acid that enters the duodenum. The current model for bicarbonate secretion across these cells calls for bicarbonate entry into the cell across the basolateral membrane via pNBC1. RT-PCR analysis of total RNA revealed that our cells contain the message for pNBC1 but not for kNBC1, NBC2 or NBC3. This result establishes the fact that our cell line is a suitable model for studying the biophysical and physiological properties of pNBC1 in its native, endogenous, environment. For pNBC1 to transport bicarbonate into the cell, its reversal potential should be more negative than the basolateral membrane potential. With [Na+]i of 14 mm, [Na+]o of 140 mm, [HCO3]o of 25 mm and [HCO3]i of 15 mm, eqn (1) predicts a reversal potential of -86 mV, for a 2 HCO3:1 Na+ stoichiometry and -50 mV for a 3:1 stoichiometry. Thus, at a membrane potential range of -30 to -65 mV (Novak & Pahl, 1993), a cotransporter with a 2:1 stoichiometry would transport bicarbonate into the cell, while a 3:1 cotransporter (e.g. in the renal proximal tubule) would mainly transport bicarbonate out of the cell. Our results, showing a 2:1 stoichiometry for pNBC1, are consistent with the transport of bicarbonate into pancreatic duct cells by pNBC1.

In a more recent study, Shumaker et al. (1999) have suggested that impaired bicarbonate secretion in the pancreas of cystic fibrosis patients results from the inability of the duct cells to depolarize the membrane potential sufficiently in order to drive enough bicarbonate influx through pNBC1, due to the lack of active apical CFTR Cl channels. This proposition is based on the assumption that pNBC1 in pancreatic duct cells is electrogenic. Our results support this assumption.

In the present study, we studied the stoichiometry of pNBC1 in mammalian pancreatic epithelial cells. In a previous study we employed a similar approach to measure the stoichiometry of the renal electrogenic sodium bicarbonate cotransporter (Gross & Hopfer, 1996) and found that the cotransporter exhibited a 3 HCO3:1 Na+ stoichiometry. The results of the present study suggest that pNBC1 in pancreatic cells operates with a different stoichiometry from that of kNBC1 in kidney cells. It is possible that the different stoichiometries exhibited by the two NBC1 variants are sequence related. Human pNBC1 is a 1079 amino acid polypeptide (Abuladze et al. 1998), while human kNBC1 is a 1035 amino acid polypeptide (Burnham et al. 1997). The primary structure of the two variants is almost identical except for a small region in their amino-termini. Two recent studies (Heyer et al. 1999; Sciortino & Romero, 1999) have reported that the stoichiometry of kNBC1 from rat kidney expressed in Xenopus oocytes is 2 HCO3:1 Na+. As discussed above, a 2:1 stoichiometry would not be able to support bicarbonate reabsorption from the glomerular filtrate. Furthermore, this finding is in contrast to the 3:1 stoichiometry reported previously for kNBC1 in rat (Gross & Hopfer, 1996), rabbit (Yoshitomi & Fromter, 1984, 1985) and Necturus (Lopes et al. 1987) proximal tubule. It seems likely that the oocyte expression system does not mirror the environment in the mammalian proximal tubule.

Alternatively, the different stoichiometries could result from interaction(s) of the cotransporter proteins with some, as yet unknown, cell-specific regulatory factors that might be present in one organ but absent in the other. Studies are currently in progress to distinguish and test these possibilities.


The authors wish to thank Mrs Karen Hawkins and Drs Anoop A. Kumar and Michael B. Ganz for technical assistance and valuable comments on the manuscript. This work was supported by grants from the American Heart Association and Kidney Foundation of Ohio to Eitan Gross; NIH grants DK46976, DK58563 and DK07789, the Iris and B. Gerald Cantor Foundation, the Max Factor Family Foundation, the Verna Harrah Foundation, the Richard and Hinda Rosenthal Foundation, and the Fredericka Taubitz Foundation to Ira Kurtz; and NIH grant DK53318 to Calvin U. Cotton. Natasha Abuladze is supported by a training grant from the National Kidney Foundation of Southern California (J891002).