Functional diversity of electrogenic Na⁺–HCO₃⁻ cotransport in ventricular myocytes from rat, rabbit and guinea pig

Abstract

The Na⁺–HCO₃⁻ cotransporter (NBC) is an important sarcolemmal acid extruder in cardiac muscle. The characteristics of NBC expressed functionally in heart are controversial, with reports suggesting electroneutral (NBCn; 1HCO₃⁻ : 1Na⁺; coupling coefficient n = 1) or electrogenic forms of the transporter (NBCe; equivalent to 2HCO₃⁻ : 1Na⁺; n = 2). We have used voltage-clamp and epifluorescence techniques to compare NBC activity in isolated ventricular myocytes from rabbit, rat and guinea pig. Depolarization (by voltage clamp or hyperkalaemia) reversibly increased steady-state pH_i while hyperpolarization decreased it, effects seen only in CO₂/HCO₃⁻-buffered solutions, and blocked by S0859 (cardiac NBC inhibitor). Species differences in amplitude of these pH_i changes were rat > guinea pig ≈ rabbit. Tonic depolarization (−140 mV to −0 mV) accelerated NBC-mediated pH_i recovery from an intracellular acid load. At 0 mV, NBC-mediated outward current at resting pH_i was +0.52 ± 0.05 pA pF⁻¹ (rat, n = 5), +0.26 ± 0.05 pA pF⁻¹ (guinea pig, n = 5) and +0.10 ± 0.03 pA pF⁻¹ (rabbit, n = 9), with reversal potentials near −100 mV, consistent with n = 2. The above results indicate a functionally active voltage-sensitive NBCe in these species. Voltage-clamp hyperpolarization negative to the reversal potential for NBCe failed, however, to terminate or reverse NBC-mediated pH_i-recovery from an acid load although it was slowed significantly, suggesting electroneutral NBC may also be operational. NBC-mediated pH_i recovery was associated with a rise of [Na⁺]_i at a rate ∼25% of that mediated via NHE, and consistent with an apparent NBC stoichiometry between n = 1 and n = 2. In conclusion, NBCe in the ventricular myocyte displays considerable functional variation among the three species tested (greatest in rat, least in rabbit) and may coexist with some NBCn activity.

Cardiac contractile and electrical activity is sensitive to changes of intracellular pH (Orchard & Kentish, 1990; Orchard & Cingolani, 1994). This sensitivity accounts for part of the depression of ventricular function during myocardial ischaemia, a condition associated with low pH_i (see e.g. Allen & Orchard, 1987; Gettes & Casio, 1992). Sarcolemmal ion transporters regulate pH_i in cardiac cells by importing or exporting excess acid or base. The major transporters responsible for acid extrusion are Na⁺–HCO₃⁻ cotransport (NBC) and Na⁺–H⁺ exchange (NHE) (Lagadic-Gossmann et al. 1992; Leem et al. 1999), although a monocarboxylic acid transporter (MCT) also extrudes lactic acid during periods of high glycolytic activity (Poole & Halestrap, 1993). There is universal agreement that cardiac NHE and MCT are electroneutral transporters.

In contrast to NHE and MCT, expression studies of the gene products of cardiac NBCs have identified at least two electrogenic (Choi et al. 1999; Pushkin et al. 2000; Sassani et al. 2002; Virkki et al. 2002) as well as one electroneutral (Pushkin et al. 1999; Choi et al. 2000) isoform (defined by Boron and coworkers as NBCe1-B, NBCe2-c (NBC4c) and NBCn1, respectively; see Romero et al. 2004, for a review of NBC classification). Evidence, however, that electrogenic and electroneutral NBCs are functional in intact heart tissue is less clear. For example, while NBC was proposed to be electrogenic in cat papillary muscle (Camilión de Hurtado et al. 1995, 1996) and rat ventricular myocytes (Aiello et al. 1998), earlier work on guinea pig ventricular myocytes (Lagadic-Gossmann et al. 1992) and sheep Purkinje strands (Dart & Vaughan-Jones, 1992) was suggestive of electroneutral transport. This raises the possibility of species and/or cell-type differences in the functional expression of NBCe and NBCn isoforms.

The presence of NBCe in heart cells (e.g. with an equivalent stoichiometric coupling, n HCO₃⁻ : 1 Na⁺, of n = 2 or 3) means that, providing the membrane potential is positive to the transporter's equilibrium potential (E_NBC), acid extrusion (i.e. HCO₃⁻ influx) is favoured thermodynamically. The resulting outward ionic current (I_NBC) has been suggested to hyperpolarize the resting membrane potential in cat and rat ventricular tissue, and to shorten the evoked action potential in rat (Camilión de Hurtado et al. 1995; Aiello et al. 1998). NBCe activity may therefore affect cardiac excitation and contraction. Electrogenicity of NBC also implies that the transporter's activity may be sensitive to changes in the membrane potential (Aiello et al. 1998) and hence to sarcolemmal electrical activity (Camilión de Hurtado et al. 1996). Interest in the functional consequences of NBC activity has been heightened by reports that stimulation of an NBCe transporter (NBCe1-B, otherwise known as hhNBC) contributes to myocardial reperfusion injury and perhaps to the generation of postischaemic related arrhythmias (Khandoudi et al. 2001).

In view of the uncertainty regarding the functional expression of NBCe in heart, we have compared evidence for its operation in ventricular myocytes isolated enzymatically from three commonly used species (rabbit, guinea pig, rat). Voltage-clamp and epifluorescence measurements of membrane current, intracellular pH and intracellular Na⁺ have been used to assess NBC activity. We have examined the voltage sensitivity of acid efflux through NBC during intracellular acidosis and determined whether NBC activity results in significant Na⁺ ion influx. We consider the possibility that NBCe and NBCn isoforms may be coactive within a single ventricular myocyte and that the flux density of ion transport through NBC may differ among species.

Methods

Myocyte isolation

The experiments were performed on adult ventricular myocytes isolated from rabbit, guinea pig and rat by enzymatic digestion. The isolation procedures for rabbit and guinea pig myocytes have been previously described (e.g. Xu & Spitzer, 1994; Skolnick et al. 1998) and were also used for rats. All procedures involving animals conformed to the UK Animals (Scientific Procedures) Act 1986/local named Committee guidelines. In brief, animals were anaesthetized with sodium pentobarbital (50 mg kg⁻¹, i.v. rabbit; i.p. guinea pig, rat) then the excised heart was attached to an aortic cannula and perfused with solutions gassed with 100% O₂ and held at 37°C, pH 7.3. Perfusion with a 0 mm Ca²⁺ solution for 5 min was followed by 15 min of perfusion with the same solution containing 1 mg ml⁻¹ collagenase (type II, Worthington Biochemical, Freehold, NJ, USA), 0.1 mg ml⁻¹ protease (type XIV, Sigma Chemical, St Louis, MO, USA), and 0.1 mm CaCl₂. The heart was then perfused for 5 min with the same solution containing no enzymes. The left ventricle was minced and shaken for 10 min, and then filtered through a nylon mesh. Cells were stored at room temperature in normal Hepes-buffered solution. All myocytes used in this study were rod-shaped, had well-defined striations, and did not spontaneously contract. Experiments were performed within 10 h of isolation.

Cell superfusion chamber

Bathing solutions were held at 37 ± 0.1°C in glass reservoir bottles that were sealed except for a small vent at the top and an exit port at the bottom. Solutions were delivered by gravity from the bottles to the cell bath through water-jacketed, gas-impermeable stainless steel tubing. The temperature of the solutions in the superfusion chamber was 36 ± 0.3°C. The 1-ml Plexiglas cell bath had a clear glass bottom and was mounted on the stage of an inverted microscope (Diaphot, Nikon, Japan). Bathing solutions flowed continuously through the bath at 4–6 ml min⁻¹, and solution depth was held at approximately 3 mm. The bottom of the bath was coated with laminin (BDBiosciences, Bedford, MA, USA) to improve cell adhesion.

Solutions

Two systems were used to buffer H⁺ in the myocyte bathing solutions: 20–24 mm Hepes with no added CO₂/HCO₃⁻, and 5% CO₂/HCO₃⁻ buffer with no added Hepes. The pH of all myocyte bathing solutions was 7.4, and adjusted to that value using 1 m NaOH (Hepes buffer) or NaHCO₃ (CO₂/HCO₃⁻ buffer). In the latter case, final NaHCO₃ concentration was 22 mm (Oxford, UK) or 18.5 mm (Salt Lake City, USA). The lower value reflects the fact that barometric pressure in Salt Lake City is lower (typically 649 mmHg), yielding a P_CO2 of approximately 30 mmHg in solutions saturated with 5% CO₂–95% O₂.

In addition to the buffer itself, Hepes-buffered solution contained (mm): 126 NaCl, 11 dextrose, 4.4 KCl, 1.0 MgCl₂, 1.08 CaCl₂. CO₂/HCO₃⁻-buffered solution was continuously gassed with 5.0% CO₂–95% O₂ and, in addition to the NaHCO₃ buffer, contained (mm): 120 NaCl, 11 dextrose, 4.4 KCl, 1.0 MgCl₂, 1.08 CaCl₂.

For Na⁺-free CO₂/HCO₃⁻-buffered solution, NaCl and NaHCO₃ were replaced completely with an equimolar concentration of n-methyl-d-glucamine (NMDG) and the pH was adjusted with HCl. Chloride-free CO₂/HCO₃⁻-buffered solution contained (mm): 120 sodium glucuronate (or sodium gluconate), 11 dextrose, 1.0 MgSO₄, 3.0 calcium gluconate, 4.4 potassium gluconate and 18.5–22 NaHCO₃ (5.0% CO₂–95.0% O₂).

When using osmotically compensated hyperkalaemic solutions, the control CO₂/HCO₃⁻-buffered solution contained (mm): 100 NaCl, 11 dextrose, 4.4 KCl, 1.0 MgCl₂, 40 NMDG, 1.08 CaCl₂ and 18.5–22 NaHCO₃ (gassed with 5.0% CO₂–95.0% O₂, pH adjusted to 7.4 with HCl). Hepes-buffered solution contained (mm): 100 NaCl, 11 dextrose, 4.4 KCl, 1.0 MgCl₂, 40 NMDG, 1.08 CaCl₂, 24 Hepes and 12.9 NaOH (pH adjusted to 7.4 with HCl). For high (44.4 mm)-K⁺ solution, 40 mm NMDG was replaced with the same concentration of KCl.

The normal pipette filling solution contained (mm): 123 potassium glutamate, 15 NaCl, 10 KCl, 10 Hepes, 12 KOH and 4 HCl (pH 7.1). In the Cl⁻-free experiments, the pipette solution contained (mm): 132 potassium glutamate, 1.0 KCl, 15 sodium glucuronate, 10 Hepes and 6 KOH.

Measurement of pH_i

The pH_i was measured in single myocytes with an epifluorescence system coupled to the Nikon inverted microscope. Carboxy-seminaphthorhodafluor-1 (carboxy-SNARF-1), a pH-sensitive fluorophore, was used as the fluorescent pH indicator as previously described in detail (e.g. Spitzer & Bridge, 1992; Skolnick et al. 1998; Leem et al. 1999). In brief, myocytes were equilibrated at 37°C for 10 min in the normal Hepes-buffered solution containing 10–13 μm of the acetoxymethyl ester of SNARF-1, SNARF-AM (Molecular Probes, Eugene, OR, USA). They were then placed in the cell bath, where they were bathed in the normal solution (Hepes or CO₂/HCO₃⁻-buffered) for at least 20 min before pH_i measurements began. Excitation at 515 nm was provided by a mercury arc lamp, and was directed to the bath via a 40× oil-immersion objective lens (NA 1.3). Emitted fluorescence was simultaneously collected by two photomultiplier tubes equipped with band-pass filters centred at 640 ± 20 nm and 580 ± 20 nm. The fluorescence emission ratio (640/580) was digitized at 5–10 kHz (Digidata 1322A, Axon Instruments, Union City, CA, USA). The emission ratio was calibrated as previously described (Buckler & Vaughan-Jones, 1990; Spitzer & Bridge, 1992; Leem et al. 1999) using solutions of varying pH that also contained 10 μm nigericin.

Determination of efflux of acid equivalents via NBC and NHE

Intracellular acid loading was achieved with ammonium prepulses (10–30 mm NH₄⁺) (see e.g. Leem et al. 1999). When NBC activity was examined, CO₂/HCO₃⁻-buffered solutions were superfused while NHE was inhibited using either 30 μm HOE 694 (Scholz et al. 1993) or HOE 642 (cariporide, Scholz et al. 1995) as previously described (Lagadic-Gossmann et al. 1992; Leem et al. 1999). Acid flux (J_NBC) was estimated as −β_totdpH_i/dt during pH_i recovery from the acid load, where β_tot is total intracellular buffering power. When NHE activity was examined, Hepes-buffered solutions were superfused, thus inactivating NBC (Leem et al. 1999); J_NHE was estimated as −β_intdpH_i/dt during recovery from an acid load where β_int is intrinsic intracellular buffering power.

Determination of intracellular buffering capacity

The intrinsic (non-CO₂) buffering capacity (β_int) of rabbit, rat and guinea pig myocytes over the pH_i range 6.40–7.50 was taken from data published previously (Zaniboni et al. 2003). Intracellular buffering due to CO₂ (β_CO2) was calculated as, β_CO2= 2.3[HCO₃⁻]_i, where [HCO₃⁻]_i=[HCO₃⁻]_o10^pH_i−pH_o (Roos & Boron, 1981; Leem et al. 1999). Total buffering capacity (β_tot) is equal to the sum of β_int and β_CO2. It was assumed that β_tot = β_int when myocytes were bathed in Hepes-buffered solutions containing no added CO₂/HCO₃⁻.

Electrophysiological techniques

All electrophysiological measurements were made with whole-cell perforated patch pipettes (amphotericin B-perforated, 240 μg ml⁻¹) to minimize intracellular dialysis. Pipettes were constructed from borosilicate capillary glass (Corning 8250 glass) and had resistances of approximately 2 MΩ when filled. Resting membrane potential in unclamped cells was recorded with an Axoclamp-2A amplifier system (Axon Instruments) in bridge-mode and voltage-clamping was achieved with an Axopatch 200B clamp system using a CV203BU headstage. Membrane potential (E_m) and membrane current (I_m) were filtered at 5 kHz and digitized at 10–20 kHz with a 16-bit A/D converter (Digidata 1322A) and analysed using pCLAMP 8 software (Axon Instruments). The reference electrode was a flowing 3 m KCl bridge for experiments involving changes in bath Cl⁻ and a Ag–AgCl pellet was used when [Cl⁻]_o was held constant.

Current–voltage relationships for I_NBC were measured by voltage clamping myocytes with a ramp protocol. Cells were held initially at an E_m of −80 mV and then clamped to +50 mV for 1 s followed by a linear ramp to −140 mV over 30 s. The ramp was applied first during superfusion with Hepes buffered and then after 4 min in CO₂/HCO₃⁻-buffered solution, with the difference current taken as I_NBC. The reversal potential (E_NBC) was determined from the best fit of I_NBC. pH_i was also measured during these experiments. In an earlier study, Aiello et al. (1998) used the same technique for measuring I_NBC in rat ventricular myocytes and obtained results very similar to ours. The previous observation that the pH buffer, Hepes, has no direct effect on cardiac action potential configuration (Spitzer & Hogan, 1979) suggests that Hepes per se did not affect our measurements of I_NBC.

All bathing solutions used in the voltage-clamp experiments (ramps and steps) contained 0.5 mm Ba²⁺ to block inward I_K1 (e.g. Cordeiro et al. 1998) and 10 μm nifedipine to reduce Ca²⁺ entry that may subsequently affect pH_i (e.g. Vaughan-Jones et al. 1983). Nifedipine did not alter pH_i regulatory systems, as judged from its lack of effect on pH_i recovery from an intracellular acid load under CO₂/HCO₃⁻-buffered conditions (rabbit ventricular myocytes, n = 14; data not shown).

Measurement of [Na⁺]_i with SBFI fluorescence

Dye loading and ratiometric measurement

Intracellular Na⁺ concentration, [Na⁺]_i, was measured in single myocytes using the fluorophore benzofuram isophthalate (SBFI). To load cells with the dye, 0.3 μl of 1 mm stock solution of the AM ester of the dye (in DMSO) was mixed with 2 μl of 20% (v/v) pluronic acid F127 (in DMSO) and added to 300 μl of ventricular myocytes, suspended in culture medium. The mixture was incubated in darkness for 2 h at room temperature. Cells were then transferred to the experimental chamber and allowed to settle for a few minutes whereupon they were superfused with normal Hepes or CO₂/HCO₃⁻ solution at 37°C. It has been estimated that 70% of intracellular SBFI is cytoplasmic, the rest being located within intracellular organelles (Harootunian et al. 1989).

The SBFI signal was quantified ratiometrically in an epifluorescence system built around a Nikon Diaphot 200B inverted microscope. Excitation light was provided by a 75 W Xe UV lamp. As SBFI is a dual-excitation dye, excitation light was filtered alternately at 340 ± 5 nm and 380 ± 5 nm, using a rocking filter-wheel (see e.g. Donoso et al. 1992). Each filter was presented for 0.5 s at 1 Hz. The emission signal from a cell was directed through a 510 ± 10 nm interference filter and measured using a photomultiplier tube (Thorn EMI; Fact 50 MK III electron tubes). Current from the PMT was converted to voltage and digitized (Cambridge Electronic Design, CED 1401). The ratio of emission elicited at the two excitation wavelengths (F₃₄₀/F₃₈₀) was stored and converted later to a measure of [Na⁺]_i using a ratiometric calibration.

Calibration of ratiometric signal

This was done by superfusing cells with solutions containing various known concentrations of sodium: 0, 5, 10, 15 and 20 mm (see e.g. Donoso et al. 1992). Solutions were obtained by appropriate mixing of two solutions containing 135 mm NaCl or 135 mm KCl. The other ingredients of calibration solutions were 2 mm EGTA, two ionophores to assist equilibration of intracellular with extracellular Na⁺ (2 μm gramicidin D and 40 μm monesin), 100 μm strophanthidin to inhibit Na⁺–K⁺ ATPase, and 10 mm Hepes; pH adjusted to 7.4 with known quantities of NaOH or KOH. Over the [Na⁺] range 0 mm to 20 mm, the relationship between fluorescence ratio and [Na⁺] was linear.

Correcting for pH sensitivity of ratiometric signal

The pH sensitivity of SBFI is modest (Minta & Tsien, 1989), typically leading to small artifactual excursions in the [Na⁺]_i signal during large changes of pH_i. These are characterized by a small step decrease in the [Na⁺]_i signal when there is a large step decrease in pH_i, and a step increase in response to a step rise of pH_i (Harrison et al. 1992). Such artifacts are evident in the experiment illustrated in Fig. 8A. This shows a recording of [Na⁺]_i in a guinea pig ventricular myocyte superfused with Hepes-buffered solution. The light-grey trace shows the uncorrected recording of [Na⁺]_i. Note the small, initial step-changes in the trace upon addition and removal of extracellular NH₄⁺, a procedure that induces large step-changes of pH_i (typical pH_i changes are plotted below the [Na⁺]_i trace in Fig. 8A, derived from a computational model of pH_i regulation in this cell type; Leem et al. 1999). The step artifacts were removed by applying a correction factor to the calibrated [Na⁺]_i trace. This was estimated from the step change in the [Na⁺]_i trace when extracellular ammonium concentration was reduced to zero in the presence of 30 μm cariporide (as shown in the latter part of Fig. 8A), using the following equation:

The best-fitting value of coefficient α for the data was 3. The corrected trace (black) is superimposed on the uncorrected trace in Fig. 8A, showing that the slower, physiological changes of [Na⁺]_i were little affected by the correction. A similar approach was used to correct the [Na⁺]_i signal on addition or removal of ammonium under CO₂/HCO₃⁻-buffered conditions. Again, the best-fit value for α was 3 (n = 14 cells).

Figure 8. Change in intracellular sodium induced by activation of NBC and NHE.

A, response of [Na⁺]_i in a guinea pig myocyte to intracellular acidosis imposed in Hepes-buffered bathing solution. The transient rise in [Na⁺]_i was completely blocked by inhibition of NHE with cariporide. The [Na⁺]_i signal shown in black was corrected for changes in pH_i as described in the Methods. B, response of [Na⁺]_i in a rat myocyte to intracellular acidosis imposed in CO₂/HCO₃⁻-buffered solution. Following the second and third prepulses, 30 μm cariporide and 10 μm S0859 were present in the superfusates, respectively. Both agents reduced the time course and magnitude of the transient rise in [Na⁺]_i associated with pH_i recovery. The bottom trace shows the simulated pH_i time course necessary to correct the SBFI signal for its pH_i sensitivity (see Methods). By relating the Na⁺ rise after the second or third prepulse to the control rise, the percentage of Na⁺ influx on NHE and NBC can be estimated.

Statistics

Summarized results are expressed as means ± s.e.m. A paired Student's t test was used to test significance between results obtained with each cell serving as its own control. An unpaired t test was used to test significance between results obtained on different cells. P < 0.05 was considered significant.

Results

Evidence for NBC in rabbit heart

Although NBC has been identified in cardiac tissue from guinea pig (Lagadic-Gossmann et al. 1992), sheep (Dart & Vaughan-Jones, 1992), rat (Ng et al. 1993; Kohout & Rogers, 1995; Le Prigent et al. 1997; Aiello et al. 1998; Khandoudi et al. 2001), cat (Camilión de Hurtado et al. 1995), ferret (Grace et al. 1993; Vandenberg et al. 1993) and human (Loh et al. 2002), its existence in rabbit ventricle is unresolved. Rabbit NBC was thus evaluated in the first series of experiments. Figure 1A shows that in the presence of a normal CO₂/HCO₃⁻ buffer system, the potent NHE1 inhibitor, HOE 694 (30 μm), only partially blocked pH_i recovery from intracellular acidosis. This contrasts with the essentially complete inhibition of pH_i recovery by this agent or by cariporide (HOE 642) in cells bathed in Hepes-buffered, CO₂/HCO₃⁻-free solution (Fig. 1B). NHE is therefore the principal acid extrusion mechanism in the absence of carbonic buffer while, in its presence, a HCO₃⁻-dependent acid extrusion mechanism is also evident. This latter mechanism requires extracellular Na⁺, as illustrated in Fig. 1C, where restoring [Na⁺]_o to an acid-loaded myocyte exposed to cariporide activated pH_i recovery. In contrast, HCO₃⁻-dependent pH_i recovery did not appear to require Cl⁻ since there was no significant difference between acid extrusion estimated at a common pH_i (6.85 units) in myocytes equilibrated with normal or zero Cl⁻ solution (Fig. 1D). Collectively, these results are consistent with the presence of a functional NBC in rabbit ventricular myocytes.

Figure 1. Evidence for NBC in rabbit ventricular myocytes.

A, inhibition of NHE1 with HOE 694 in CO₂/HCO₃⁻-buffered solution did not completely block recovery of pH_i from the intracellular acidosis elicited by NH₄Cl prepulses (10 mm). B, summary of acid extrusion in Hepes-buffered solution without HOE (n = 31 cells) and with HOE compounds (30 μm, 642, n = 12 cells; 694, n = 6 cells). The inset shows a representative experiment. C, example of the requirement for external sodium to induce pH_i recovery in a rabbit myocyte bathed in CO₂/HCO₃⁻-buffered solution with NHE inhibited. One of four similar experiments. D, example experiment showing that acid extrusion, in the presence of CO₂/HCO₃⁻ with NHE inhibited, persists in chloride-free solution. One of 11 similar experiments. The results are summarized in the inset, which shows mean net acid extrusion measured between pH_i values of 6.8 and 6.9 (filled bars: control, n = 13; open bars: chloride free, n = 8). In this protocol cells were bathed in Cl⁻-free solution for at least 25 min prior to applying the ammonium prepulse, a protocol sufficient to remove [Cl⁻]_i from rabbit myocytes.

Comparison of pH_i dependence of NBC among three species

Intracellular acidosis is a major activator of acid extrusion via both NHE and NBC. The relationship between pH_i and acid efflux through NBC (J_NBC) in rat, rabbit and guinea pig ventricular myocytes has been plotted in Fig. 2. J_NBC was measured from pH_i recovery in CO₂/HCO₃⁻-buffered solution with NHE blocked using either 30 μm cariporide or HOE 694 (e.g. Fig. 1A). As individual cells were neither voltage clamped nor field stimulated, membrane potential would have been close to its normal resting value of approximately −85 mV. At any given pH_i, mean values for J_NBC were essentially identical in the presence of cariporide or HOE 694 (not shown) and so have been combined in the data plotted in Fig. 2A.

Figure 2. Effect of pH_i on acid efflux.

A, summary of the relationship between J_NBC and pH_i in rat (n = 66, black), rabbit (n = 26, blue), and guinea pig (n = 73, red) myocytes. Continuous lines are best-fits of data to the equation: J_NBC = J_max·[H⁺]^m/(K^m+[H⁺]^m), where J_max is the maximum flux, K is the K_m for protons, [H⁺] is proton concentration and m is the Hill coefficient. The respective values for J_max, K and m for NBC in rat were: 6.315, 10^−6.506 and 1.677. For guinea pig they were: 4.861, 10^−6.47 and 1.913. For rabbit they were: 5.0, 10^−6.427 and 1.816. B, summary of the relationship between J_NHE and pH_i in rat (n = 70, black), rabbit (n = 31, blue), and guinea pig (n = 118, red) myocytes. J_NHE in guinea pig myocytes was greater than that in rat or rabbit cells at pH_i values less than ∼6.8. The dashed lines are the best-fit results for J_NBC given in panel A. Continuous lines are best-fits of data to the equation: J_NHE = J_max[H⁺]^m/(K^m+[H⁺]^m). The respective values for J_max, K and m for NHE in rat were: 34.188, 10^−6.380 and 1.933. For guinea pig they were: 47.110, 10^−6.376 and 2.134. For rabbit they were: 8.011, 10^−6.667 and 2.544.

J_NBC was similar in rabbit and guinea pig myocytes over the pH_i range of 6.4–7.2, but higher in rat at pH_i values below approximately 6.9 (Fig. 2A). For example, at pH_i 6.6 the mean J_NBC in rat myocytes was approximately 50% greater than that in rabbit and guinea pig cells. Figure 2B presents data for the pH_i dependence of acid extrusion via NHE for all three species. Results were derived from pH_i recovery in CO₂/HCO₃⁻-free solution (pH_o 7.4) buffered with Hepes.

Voltage sensitivity of steady-state pH_i

Studies of NBCe1 and NBCe2 gene products, when expressed in oocytes, have indicated that activity of both transporters is voltage sensitive (Choi et al. 1999; Virkki et al. 2002) with I_NBC displaying a reversal potential (E_NBC, equilibrium potential) in accordance with equilibrium thermodynamics:

(1)

where R, T, and F have their usual meaning, and n is the coupling coefficient (n HCO₃⁻ ions cotransported with each Na⁺ ion). If NBCe1 and/or NBCe2 gene products are functionally active in ventricular myocytes, they should mediate an acid extrusion that is sensitive to membrane potential, with depolarization enhancing and hyperpolarization reducing or even reversing the extrusion, thereby affecting resting pH_i. It is not known if acid efflux via NBCn1 is voltage sensitive, but the flux would not be expected to display a reversal potential, as the equilibrium condition for such a transporter is defined only by the sarcolemmal gradients of Na⁺ and HCO₃⁻ ions.

Changing membrane potential under voltage clamp conditions

Figure 3A illustrates the response of resting pH_i (rat, carbonic buffer) to a voltage-clamp hyperpolarization from a holding potential of −80 mV to −140 mV followed by a step to 0 mV. At −80 mV the calculated E_NBC (eqn (1)) was −96 mV, assuming equality of the pipette and intracellular sodium concentrations. Hyperpolarization decreased pH_i while depolarization increased it. Similar changes in pH_i occurred in rabbit and guinea pig myocytes subjected to the same clamp protocol, but the changes were smaller in magnitude (Fig. 3B). When voltage steps were applied to rabbit myocytes superfused with Hepes solution, nominally free of CO₂/HCO₃⁻, there were no significant changes in pH_i (Fig. 3C), indicating that they are bicarbonate dependent.

Figure 3. Response of pH_i to changing E_m with voltage clamp.

A, example of the large changes in pH_i elicited by the clamp protocol when applied to a rat myocyte bathed in CO₂/HCO₃⁻ solution. Cells were initially held at −80 mV, then clamped to −140 mV for 3 min followed by 10 min at 0 mV, then back to −140 mV. B, summary of E_m-induced changes in pH_i in rat (n = 5), guinea pig (n = 8) and rabbit (n = 12). *P-values (paired) ranged between 0.01 and 0.004. C, summary of the lack of response of pH_i to changing E_m in rabbit myocytes (n = 7) bathed in Hepes-buffered solution containing no added CO₂/HCO₃⁻. In contrast, when cells (n = 12) were bathed in CO₂/HCO₃⁻-buffered solution, hyperpolarization caused a significant fall in pH_i (*P < 0.05, paired) while depolarization induced a small increase. D, in the absence of the NBC blocker, S0859, rat myocytes (n = 6) displayed a large rapid increase in pH_i in response to a 10 min step voltage change from −80 mV to 0 mV. *P-values ranged from 0.02 to 0.009, paired. This increase was completely blocked in myocytes (n = 4) bathed for 5 min in S0859 (10 μm). E, the pattern of E_m-induced changes in rabbit pH_i (n = 5) was not altered by lowering [Cl⁻]_o to 1 mm for at least 25 min (paired P-values).

The voltage sensitivity of pH_i was inhibited by 10 μm S0859, a selective cardiac NBC inhibitor (Ch'en & Vaughan-Jones, 2001b), as shown in Fig. 3D which pools data gathered from rat myocytes. The voltage-dependent changes of pH_i may therefore be attributed to a modulation of NBC activity. An alternative explanation is that the changes in pH_i were secondary to voltage-dependent changes in [Cl⁻]_i (Vaughan-Jones, 1979). Alterations in [Cl⁻]_i could, in principle, change pH_i by varying net membrane fluxes of HCO₃⁻ and OH⁻ anions via Cl⁻–HCO₃⁻ exchange (Vaughan-Jones, 1979) and Cl⁻–OH⁻ exchange (Sun et al. 1996), respectively. To test this hypothesis, the clamp protocol was applied to rabbit cells bathed in very low [Cl⁻]_o solution (present as 0.5 mm BaCl₂) using a pipette filling solution that also contained 1 mm Cl⁻ (Fig. 3E). The depolarization-induced changes in pH_i were very similar to those in normal [Cl⁻]_o, strongly suggesting that the anion exchangers were not involved.

Changing membrane potential by varying [K⁺]_o (constant osmolarity)

In myocytes not subjected to voltage clamp, increasing [K⁺]_o from 4.4 to 44.4 mm in CO₂/HCO₃⁻-buffered solution (pH_o 7.4) depolarized E_m from approximately −85 mV to −25 mV (rabbit, n = 2). Superfusate osmolarity, [Na⁺]_o and [Cl⁻]_o were held constant at 317 mosmol l⁻¹, 118.5 mm and 148 mm, respectively, by replacing NMDG-Cl with KCl (see Methods). Figure 4A shows that, in rat myocytes, pH_i increased under these conditions. In contrast, as summarized in Fig. 4B, the same changes of [K⁺]_o made in Hepes-buffered solution produced no change of pH_i. The hyperkalaemia-induced increase of pH_i was inhibited by 10 μm S0859, indicating that it was mediated by acid extrusion on NBC (Fig. 4C). Figure 4D summarizes data from all three species, showing that similar, but smaller, effects were also evident in rabbit and guinea pig myocytes. The rank order for effects on pH_i was rat > rabbit ≈ guinea pig, paralleling the effect on pH_i of voltage-clamp steps (Fig. 3B), and confirming that pH_i in the presence of CO₂/HCO₃⁻ is voltage sensitive.

Figure 4. Effect of potassium-induced depolarization on pH_i with bathing solution osmolarity held constant.

A, example of the rise in pH_i that occurred in rat myocytes bathed in CO₂/HCO₃⁻-buffered solution and exposed to high [K⁺]_o for 10 min. B, summary of results from rat myocytes bathed in CO₂/HCO₃⁻-buffered (n = 7) and Hepes-buffered (n = 5) solution. C, mean change in rat pH_i following 10 min exposure to high [K⁺]_o (CO₂/HCO₃⁻ buffer) in the absence (n = 7) and presence of S0859 (n = 4). The NBC inhibitor completely blocked the rise in pH_i. D, summary of the rise in pH_i elicited by 10 min exposure to high [K⁺]_o in CO₂/HCO₃⁻-buffered solution (rat, n = 7; rabbit, n = 13; guinea pig, n = 13). *P < 0.05 (paired).

Changing membrane potential by varying [K⁺]_o (uncompensated osmolarity)

Much larger effects of hyperkalaemia on resting pH_i (0.1–0.2 pH_i units) have been reported previously for cat papillary muscle (Camilión de Hurtado et al. 1995, 1996) but in those cases [K⁺]_o was elevated (4.5 mm to 45 mm) without osmotic compensation. We therefore examined the effect of [K⁺]_o-induced depolarization on pH_i in myocytes under conditions in which additional KCl was added to the bathing solution without osmotic compensation. In contrast to the small alkalosis shown in Fig. 4, uncompensated hyperkalaemia elicited much larger increases of pH_i in all three species (Fig. 5).

Figure 5. Response of pH_i to potassium-induced depolarization without osmotic compensation.

A, guinea pig myocyte: response of pH_i to superfusion with hypertonic solution (generated by directly adding either 40 mm NaCl or 40 mm KCl to Tyrode solution, denoted as 40 Na and 40 K, respectively) in the absence and then in the presence of 250 μm DIDS. B, rat myocyte: response of pH_i to isotonic hyperkalaemia (80 mm) or hypertonic hyperkalaemia (80 mm), denoted by 80K(iso) and 80K, respectively. These solutions were first applied in the absence of CO₂/HCO₃⁻ (Hepes-buffered), then in CO₂/HCO₃⁻-buffered solution. C, rabbit myocytes: pooled data showing effect on pH_i of hypertonic hyperkalaemia (44.4 mm) without (n = 5) and with (n = 7) 30 μm HOE 642. The absence of full pH_i recovery after removal of hyperkalaemia was evident in all cells tested.

For example, as shown in Fig. 5A, an 8 min exposure of guinea pig myocytes to osmotically uncompensated high K⁺ solution (44.5 mm) increased pH_i by 0.09 ± 0.02 units (n = 9), instead of the 0.01 unit alkalosis seen with osmotically compensated hyperkalaemia in this species (Fig. 4D). This figure also shows that a similar alkalosis occurred (0.06 ± 0.01 units; n = 12) when an equivalent rise of extracellular tonicity was induced at constant 4.5 mm [K⁺]_o (achieved in this case by adding 40 mm NaCl to the bathing solution). These hypertonically induced increases in pH_i persisted in the presence of 250 μm DIDS (a non-selective inhibitor of NBC; Lagadic-Gossmann et al. 1992) ruling out any major role for NBC activation. In the presence of DIDS, pH_i increased by 0.09 ± 0.02 (n = 9) during hypertonic addition of 40 mm KCl, and by 0.05 ± 0.01 (n = 9) during hypertonic addition of 40 mm NaCl.

Similar results were observed in experiments on rat ventricular myocytes (n = 6). Figure 5B illustrates one such experiment. The myocyte was exposed to hyperkalaemia (80 mm) under conditions where the solution osmolarity was kept constant by an equivalent reduction of [Na⁺]_o, or under hypertonic conditions where 75.5 mm KCl was added to the Tyrode solution. Once again, a large alkalosis was only evident during exposure to hypertonic solution. In rabbit myocytes, osmotically uncompensated hyperkalaemia (4.4 mm to 44.4 mm) elicited an increase of pH_i that was inhibited markedly by cariporide, suggesting that much of the alkalosis was secondary to an activation of NHE (Fig. 5C). In addition, exposure of rabbit myocytes (n = 7) to an equivalent osmotic step (288 to 347 mosmol l⁻¹) by adding sucrose, while keeping [K⁺]_o and [Na⁺]_o constant, elicited an alkalosis similar to that shown in Fig. 5C in the absence of cariporide (data not shown). These results demonstrate that the use of hypertonic hyperkalaemia to explore the voltage dependence of NBC induces significant pH_i changes that are unrelated to the transporter's activity.

Voltage sensitivity of pH_i recovery from intracellular acidosis

We assessed NBC's voltage sensitivity during its activation by intracellular acidosis. Figure 6A illustrates a typical experiment on a guinea pig ventricular myocyte. E_m was clamped initially at −80 mV in the presence of 30 μm cariporide, while an acid load was imposed by prepulsing the cell with extracellular ammonium. During the ensuing pH_i recovery, E_m was hyperpolarized to −140 mV for 200 s, and then depolarized to 0 mV. Hyperpolarization roughly halved the rate of pH_i recovery while the subsequent depolarization restored it. These findings are consistent with hyperpolarization decreasing and depolarization increasing HCO₃⁻ influx via a voltage-sensitive NBC. Similar results were obtained in six other experiments.

Figure 6. Effect of E_m on NBC during recovery from intracellular acidosis.

A, hyperpolarization slows and depolarization speeds the rate of pH_i recovery mediated by NBC in a guinea pig ventricular myocyte. Only the expanded recovery phase of pH_i is illustrated while the inset shows the entire experimental protocol. The calculated values of E_NBC for n values of 2 and 3 are superimposed on the voltage-clamped E_m signal. B, mean changes in J_NBC (▵J_NBC) in guinea pig myocytes (n = 7) elicited by the two clamp steps (−80 to −140 mV; −140−0 mV). The ▵J_NBC that would have occurred due to pH_i recovery alone over this time interval, with E_m kept at its normal resting value, is also shown. C, response of pH_i recovery rate (following an ammonium prepulse) to K⁺-induced depolarization in a rat myocyte. D, response of pH_i recovery rate to K⁺-induced depolarization in a guinea pig myocyte. E, comparison of J_NBC–pH_i curves from rat myocytes at normal resting E_m (•, same data as Fig. 2A) and in cells depolarized with 44.4 mm [K⁺]_o (▴, n = 11) during pH_i recovery from an acid-load (induced by ammonium prepulse). All bathing solutions used in this figure were buffered with CO₂/HCO₃⁻ and contained 30 μm HOE 642. Nifedipine (10 μm) was also included during the voltage clamp and high [K⁺]_o experiments.

At pH_i 6.80 during the acid load shown in Fig. 6A, E_NBC (estimated from eqn (1)) would have been −85 mV and −130 mV, for an ionic coupling coefficient of n = 3 and 2, respectively (assuming [Na⁺]_pip=[Na⁺]_i= 15 mm). Thus, at an E_m of −140 mV, acid efflux via an electrogenic NBC (n = 2 or 3) would have been terminated or even reversed (thus mediating HCO₃⁻ ion efflux, equivalent to acid influx into the cell). In none of the seven experiments, however, was pH_i recovery prevented, although in all cases it was slowed significantly. The magnitude of the change in net acid efflux (▵J_NBC) elicited by the switch between the two holding potentials is summarized in Fig. 6B. ▵J_NBC was calculated as the difference in J_NBC measured 30 s before and after the voltage step. For purposes of comparison, the small decrease in J_NBC (▵J_NBC) that may be attributed simply to the pH_i change during this 1 min period (independent of the voltage jump), is also shown. At pH_i∼6.80, all acid extrusion is mediated via NBC (with NHE blocked by cariporide, pH_i recovery is dependent on extracellular HCO₃⁻ and Na⁺; Lagadic-Gossmann et al. 1992, see also Fig. 1C). The failure of a large hyperpolarization to terminate or reverse pH_i recovery therefore suggests that, in addition to NBCe, an NBCn may be operating across the sarcolemma, as the thermodynamic energy source for this would be voltage independent and would still have favoured acid extrusion.

In another series of experiments, a smaller depolarization to approximately −25 mV was induced in unclamped rat (n = 11) and guinea pig (n = 3) myocytes by applying 44.4 mm [K⁺]_o (with solution osmolarity and [Na⁺]_o held constant; Fig. 6C, rat myocyte; Fig. 6D, guinea pig myocyte). In this case, the driving force for acid efflux through both electrogenic and electroneutral NBCs would remain outward throughout. In the rat myocyte there was a stimulatory effect on NBC-mediated pH_i recovery, but in the guinea pig myocyte (Fig. 6D) no effect was detected. This species difference in the responsiveness of pH_i recovery to modest hyperkalaemic depolarization is consistent with the findings for steady-state pH_i shown in Fig. 4D. Figure 6E plots the relationship between J_NBC and pH_i in rat myocytes exposed to a [K⁺]_o of 4.4 mm (normal E_m) and 44.4 mm (depolarized E_m). Depolarization induced a large upward shift in the J_NBCversus pH_i curve, and increased its slope. This indicates that the fractional increase in J_NBC in 44.4 mm [K⁺]_o (about 100%) was the same over the range of pH_i values tested.

In summary, depolarization can enhance the ability of intracellular acidosis to activate NBC activity.

Voltage dependence of I_NBC

Given that the voltage-dependent changes in pH_i observed in CO₂/HCO₃⁻-buffered solution appeared to be mediated by NBC, it was of interest to determine the voltage dependence of I_NBC (Fig. 7). Our approach was first to apply a voltage-clamp ramp while measuring membrane current in Hepes-buffered solution. NBC was then activated by switching to a CO₂/HCO₃⁻-buffered solution and, after 4 min, the ramp was reapplied. The difference current on switching to CO₂/HCO₃⁻ was taken as I_NBC (see Methods, and see Aiello et al. 1998). After 4 min, the intracellular acidosis produced initially by the switch to CO₂/HCO₃⁻ had been attenuated by sarcolemmal acid extrusion and was relatively modest (pH_i 7.0–7.1; n = 19). By abbreviating the time between the first and second voltage ramps, the likelihood of I_NBC rundown or time-dependent changes in other currents activated by the ramp was minimized. In control experiments using ammonium prepulses to induce intracellular acidification, we found no effect of changing pH_iper se (over the pH_i range from 7.0 to 7.3) on I–V curves from myocytes (rabbit, n = 6–13) bathed continuously in Hepes-buffered solution. Similarly, successive ramps applied to myocytes (rabbit, n = 5) bathed in Hepes-buffered solution over a 4 min interval produced virtually identical I–V curves.

Figure 7. E_m dependence of I_NBC in all species.

A, example of I–V curves recorded in a rat myocyte bathed first in Hepes (pH_i 7.29, t= 0) then in CO₂/HCO₃⁻(pH_i 7.11, t= 4 min). B, difference current (I_NBC) calculated from the I–V curves in panel A. C, summary of difference currents (I_NBC) from all species: (•, rat, n = 5) (▴, guinea pig, n = 9) (▪, rabbit, n = 5). The mean pH_i values measured during the ramps, first in Hepes- and then 4 min later in CO₂/HCO₃⁻-buffered solution were: rat 7.25 ± 0.01 (Hepes), 7.05 ± 0.02 (CO₂/HCO₃⁻); rabbit 7.19 ± 0.02 (Hepes), 7.00 ± 0.03 (CO₂/HCO₃⁻); guinea pig 7.23 ± 0.02 (Hepes), 7.07 ± 0.02 (CO₂/HCO₃⁻).

Current–voltage relationships for I_NBC are summarized in Fig. 7. A representative rat myocyte experiment is shown in Fig. 7A and includes I–V curves recorded first in Hepes- and then after 4 min in CO₂/HCO₃⁻-buffered solution, as well as the difference current (Fig. 7B). A voltage-sensitive, bicarbonate-dependent outward current was clearly evident in the voltage range positive to about −100 mV, reversing at more negative voltages. This is similar to that previously attributed to electrogenic NBC in this cell-type (Aiello et al. 1998). The current cannot be ascribed to HCO₃⁻ flowing through an anion channel as, at the pH_i for the measurement (about 7.05), any outward HCO₃⁻ current would have reversed at about −40 mV (e.g. Spitzer & Hogan, 1979), similar to that for Cl⁻ (Harvey & Hume, 1989). Current reversal at −100 mV, however, is consistent with I_NBC for a coupling coefficient of n = 2 (estimated from eqn (1)). Results from similar experiments are summarized in Fig. 7C. In all cases, there was a significant HCO₃⁻-activated current, displaying a similar reversal potential (E_NBC) of −98 ± 12 mV for rat (n = 5), −103 ± 20 mV for rabbit (n = 5), and −106 ± 14 mV for guinea pig (n = 9). There were, however, marked species differences in I_NBC density, with rat > guinea pig > rabbit. For example, at an E_m of 0 mV, I_NBC at resting pH_i was +0.52 ± 0.05 pA pF⁻¹ (rat), +0.26 ± 0.05 pA pF⁻¹ (guinea pig) and +0.10 ± 0.03 pA pF⁻¹ (rabbit).

Effect of NBC on the resting membrane potential

Activation of electrogenic NBC when membrane potential is positive to E_NBC will be associated with outward current and may therefore hyperpolarize the resting membrane potential. We evaluated this prediction in quiescent myocytes using patch-pipettes filled with normal solution, containing either 5 mm or 15 mm[Na⁺]. E_m and pH_i were recorded simultaneously as the superfusate was switched from Hepes-buffered to CO₂/HCO₃⁻-buffered solution (pH_o 7.4) for 10 min. In all three species, the switch to CO₂/HCO₃⁻-buffered solution elicited a small depolarization rather than the predicted hyperpolarization. This ranged from approximately 0.4 mV to 2 mV (rat n = 5; guinea pig, n = 9; rabbit, n = 11) and was significant only in rabbit and rat myocytes at 1 min, suggesting that resting membrane potential is little affected by NBCe activity.

Changes in [Na⁺]_i associated with NBC activation

Acid efflux via NBC should be associated with Na⁺ influx. Figure 8 shows representative examples of changes in [Na⁺]_i recorded from guinea pig and rat myocytes during activation of NBC and NHE. In Fig. 8A, a guinea pig myocyte bathed in Hepes-buffered solution was subjected to an intracellular acid load (induced by ammonium prepulse). The subsequent pH_i recovery was accompanied by a transient rise of [Na⁺]_i that was blocked by cariporide, as shown in the latter part of the trace, indicating that it was mediated by a stimulation of NHE. Plotted below the [Na⁺]_i traces in Fig. 8 are the computed pH_i changes expected during the protocols (see Methods).

Figure 8B shows a recording of [Na⁺]_i in a rat myocyte superfused with carbonic buffered solution. In this case cariporide substantially reduced but did not abolish the transient rise of [Na⁺]_i following an ammonium prepulse, suggesting that the remaining rise was mediated by Na⁺ influx on NBC. This was supported by the observation that, later in the same experiment, 10 μm S0859 also attenuated the post-ammonium rise of [Na⁺]_i. The predicted time course for pH_i is plotted beneath the [Na⁺]_i trace and was calculated assuming that cariporide selectively blocks NHE (Scholz et al. 1995) and that S0859 blocks NBC activity (Ch'en & Vaughan-Jones, 2001b). In guinea pig myocytes subjected to the same experimental protocol as shown in Fig. 8B, similar changes of [Na⁺]_i were seen.

The rate of rise of [Na⁺]_i 0.5–2 min after ammonium removal (dNa⁺_i/dt) has been averaged for several cells in Fig. 9A (for guinea pig) and Fig. 9B (for rat), normalized to the rate seen initially in the same cell after ammonium-removal in the absence of cariporide or S0859. There was roughly a 20–30% reduction of dNa⁺_i/dt in the presence of S0859, while an 80–90% reduction occurred with cariporide. This indicates that, at the pH_i at which the measurements were made (∼6.80, see Fig. 8B), about 20% of total Na⁺ influx associated with pH_i regulation was trafficked on the NBC transport system. Reference to the acid-efflux curves determined for NBC and NHE (Fig. 2) permits an estimate of the fraction of total acid efflux through each transporter under these conditions. These fractions have been plotted in Fig. 9C (guinea pig) and Fig. 9D (rat), along with the corresponding fractions for total Na⁺ influx. It is apparent that fractional Na⁺ influx is similar to fractional acid efflux. The apparent coupling coefficient for total NBC, n, was estimated in two separate ways, using experimental data for NHE and NBC kinetics (Figs 9C and D):

and where F^H_NBC and F^Na_NBC are, respectively, the fraction of total acid efflux and Na⁺ influx carried on NBC (ratio of flux in the presence and absence of cariporide), and where F^H_NHE and F^Na_NHE are, respectively, the fraction of acid efflux and Na⁺ influx carried on NHE (ratio of flux in presence and absence of S0859). The data shown in Fig. 9C and D yield values for n of 1.53 ± 0.37 (n = 17) for guinea pig and 1.40 ± 0.42 (n = 23) for rat. These are non-integer numbers although neither was significantly different statistically from n = 1 or n = 2 (P > 0.05).

Figure 9. Sodium influx associated with NBC and NHE activation.

A and B, guinea pig and rat, respectively: data for net Na⁺ influx (estimated from dNa⁺_i/dt measured following 20 mm ammonium removal) obtained in the absence of drug or in the presence of S0859 or cariporide. Data have been normalized to a control dNa⁺_i/dt obtained following NH₄⁺ removal in the absence of drug. In the case of the control Na⁺ influx, two successive ammonium prepulses were performed, and dNa⁺_i/dt after the second was normalized to that after the first. Figure 8B illustrates one experimental protocol. C, flux of Na⁺, J_Na (filled bars) or acid, J_H (open bars, estimated from Fig. 2) mediated by NHE or NBC, expressed as a fraction of total flux (through both NHE and NBC). D, analogous data for rat myocytes. Sample sizes (n) in panels C and D shown above bars.

Discussion

Functional diversity of electrogenic NBC

A major aim of this study was to establish whether electrogenicity is a general property of Na⁺–HCO₃⁻ cotransport in commonly used ventricular myocyte preparations (rat, rabbit and guinea pig). Using electrophysiological and fluorescence techniques we have obtained evidence for NBCe in all three species. However, I_NBC density (Fig. 7) and the responsiveness of resting pH_i to changes in E_m (Figs 3 and 4) varied considerably among species, in the order: rat > guinea pig ≃ rabbit. This sequence is also similar to that for the pH_i dependence of acid efflux via total NBC (all isoforms) shown in Fig. 2A and may reflect a lower level of NBCe expression in guinea pig and rabbit compared with rat.

Our identification of NBCe in rat ventricular myocytes corroborates a previous report (Aiello et al. 1998). Thus, it seems likely that NBCe is a common pH_i regulatory system in mammalian ventricles. This is consistent with mRNA, protein measurement, and oocyte expression studies of NBCe1 and NBCe2 isoforms from human, rat and guinea pig heart (Choi et al. 1999; Choi et al. 2000; Virkki et al. 2002).

Comparison of acid efflux through NHE and NBC (cf. Fig. 2A with Fig. 2B) reveals that, over much of the pH_i range, flux through NHE is usually larger in all three species. Nevertheless, the different pH_i sensitivity of the two types of transporter means that fractional acid extrusion through NBC (F^H_NBC) in the three species increases from 15 to 20% at pH_i 6.50 to about 50% at pH_i 7.1. This confirms a previous report (Leem et al. 1999) that NBC is an important controller of steady-state pH_i in heart.

The NBC substrates

While one HCO₃⁻ anion may accompany the influx of one Na⁺ ion on electroneutral NBC in heart (i.e. n = 1), it is not certain that HCO₃⁻ is the transported substrate on electrogenic NBC. Alternative models involve coinflux of Na⁺ with CO₃²⁻ anions, or the uncoupled inward transport of a NaCO₃⁻ ion pair (see e.g. Romero et al. 2004). These alternative models are equivalent to conventional electrogenic Na⁺–HCO₃⁻ co-influx, with an apparent coupling coefficient (n·HCO₃⁻ : Na⁺) equal to or greater than 2. For simplicity, we have expressed our analysis of cardiac NBC in terms of equivalent movements of HCO₃⁻ and Na⁺ ions.

Voltage sensitivity of pH_i

The presence of NBCe in non-cardiac tissues is characterized by voltage dependence of both pH_i and I_NBC (Siebens & Boron, 1989; Deitmer & Szatkowski, 1990; Grichtchenko & Chesler, 1994; Munsch & Deitmer, 1994; Shumaker et al. 1999; Gross et al. 2001). Depolarization positive to E_NBC favours HCO₃⁻ influx, outward I_NBC and an increase in pH_i. Conversely, hyperpolarization negative to E_NBC favours HCO₃⁻ efflux, inward I_NBC and a reduction of pH_i. In all three species, we detected E_m sensitivity of pH_i attributable to NBC both under steady-state conditions and during pH_i recovery from intracellular acidosis. Indeed, our work using the perforated patch technique is the first to show a direct E_m dependence of pH_i in heart. Previous work utilized either hyperkalaemia or electrical pacing to modulate E_m (Camilión de Hurtado et al. 1995, 1996).

Voltage-dependent changes of resting pH_i were observed under voltage-clamp conditions and during manipulation of [K⁺]_o. While our results point to NBCe activity in the heart, it is important to stress that E_m sensitivity alone is not proof of electrogenicity of a transporter. However, the fall of resting pH_i observed in all three species upon hyperpolarization to −140 mV (Fig. 3B) is consistent with reversed NBCe since it also occurred in Cl⁻-free conditions (Fig. 3E). In this setting the other acid-loading transporters, Cl⁻–HCO₃⁻ exchange and Cl⁻–OH⁻ exchange, cannot function (Leem et al. 1999). NBCe reversal is also in agreement with the reversal of I_NBC observed in all three species at potentials negative to −100 mV (Fig. 7C) and would indicate that, over the voltage range tested, the changes of resting pH_i are due to both forward and backward modes of NBCe activity.

Two previous reports have shown that hyperkalaemia when applied with an accompanying rise of solution tonicity elicits a large alkalosis (∼0.2 units for a 10-fold rise in [K⁺]_o) in cat papillary muscle in vitro (Camilión de Hurtado et al. 1995, 1996). This change of pH_i is much greater than that observed in isolated ventricular myocytes exposed to hyperkalaemia under isotonic conditions (Fig. 4, see also Aiello et al. 1998). In addition, we found that NHE blockade markedly attenuated the alkalosis elicited by hypertonic hyperkalaemia (Fig. 5C). This finding is consistent with earlier work showing that exposure of guinea pig ventricular muscle to hypertonic solutions activates NHE (Whalley et al. 1991). Thus it seems possible that much of the hyperkalaemia-induced modulation of pH_i reported previously in cat papillary muscle was secondary to osmotic rather than voltage-dependent effects.

Voltage dependence of pH_i recovery

The present work provides the first demonstration that NBC-mediated pH_i recovery from intracellular acidosis in cardiac tissue is voltage sensitive, with depolarization speeding and hyperpolarization slowing acid extrusion (Fig. 6). Human pancreatic duct cells also display depolarization-induced stimulation of electrogenic NBC during acid load recovery (Shumaker et al. 1999).

In rat myocytes, depolarization from approximately −85 mV to −25 mV with 44.4 mm [K⁺]_o nearly doubled NBC-mediated acid extrusion (Fig. 6). This demonstrates dramatically that, in this species, in addition to pH_i and pH_o (Ch'en & Vaughan-Jones, 2001a), membrane potential is a key regulator of NBC activity. In contrast, although pH_i recovery from acidosis in guinea pig myocytes was responsive to relatively large depolarizing voltage-clamp steps (140 mV), it did not appear to be affected by the smaller depolarization induced by 44.4 mm [K⁺]_o (∼60 mV). Lagadic-Gossmann et al. (1992) also failed to detect depolarization-induced changes of pH_i recovery in guinea pig ventricular myocytes subjected to smaller voltage-clamp steps than used here, and to 45 mm [K⁺]_o (constant osmolarity). Similarly, pH_i recovery in sheep cardiac Purkinje fibres (Dart & Vaughan-Jones, 1992) and in cat papillary muscle (Camilión de Hurtado et al. 1995) was unresponsive to depolarization induced by hyperkalaemia. A low voltage sensitivity of pH_i recovery in certain species is similar to the reduced responsiveness of steady-state pH_i to depolarization in guinea pig and rabbit myocytes (Figs 3 and 4), and may reflect lower levels of I_NBC (Fig. 7C). It may also reflect the coexpression of a functionally active electroneutral NBC (see section on NBC stoichiometry).

Effect of NBC activation on resting membrane potential

Hyperpolarization of resting E_m is a frequently observed response of non-cardiac cells to activation of NBCe using application of extracellular CO₂/HCO₃⁻-buffered solution (Deitmer & Schlue, 1989; Bevensee et al. 1997; Choi et al. 1999) Similarly, a switch from Hepes to CO₂/HCO₃⁻-buffered solution (pH_o 7.4) was reported to induce a small (1–4 mV) hyperpolarization in canine cardiac Purkinje strands (Spitzer & Hogan, 1979; Dart & Vaughan-Jones, 1992), cat papillary muscle (Camilión de Hurtado et al. 1995), and rat ventricular myocytes (Aiello et al. 1998). In contrast, we found that resting E_m either depolarized slightly or did not change in any of the three species following the switch to CO₂/HCO₃⁻-buffered solution. Examination of the I–V curves (Fig. 7C) indicates the predicted hyperpolarization is likely to be small. For example, the average outward I_NBC at normal resting values of E_m was approximately 0.03 pA pF⁻¹, which for a 150 pF cell yields 4.5 pA. Assuming a normal resting input resistance of between 30 and 50 MΩ (Brown et al. 1981; Hume & Uehara, 1985; Huelsing et al. 1998) a hyperpolarization of less than 1 mV would be predicted. Such a small hyperpolarization may easily be obscured by pH_i-induced changes in other membrane currents, for example, by an acidosis-induced decrease in inward rectifier K⁺ current, as occurs in guinea pig ventricular myocytes (Ito et al. 1992).

Stoichiometry of NBC

Electrogenic NBC

Since NBC clearly supports acid extrusion under normal conditions, its coupling coefficient, n, must be ≤ 2 (Lagadic-Gossmann et al. 1992; Dart & Vaughan-Jones, 1992; Leem et al. 1999). In addition, E_NBC, measured as the reversal potential for electrogenic NBC-current, matches closely the value predicted from equilibrium thermodynamics for n =2. Thus E_NBC was measured to be −103 mV (calculated −108 mV) for rabbit, −106 mV (calculated −99 mV) for guinea pig, and −98 mV (calculated −102 mV) for rat. Aiello et al. (1998) also concluded that the stoichiometry of NBCe in rat ventricular myocytes was n = 2. The present work, however, gives no clue as to the NBCe isoform (e.g. NBCe1-B or NBCe2-c) responsible for I_NBC in heart. Indeed mRNA for both isoforms has been identified in mammalian cardiac tissue (Choi et al. 1999; Pushkin et al. 2000).

Electroneutral NBC

Recently, a novel NBC has been cloned from human heart (Pushkin et al. 1999; Choi et al. 2000) that, when expressed in Xenopus oocytes, mediates electroneutral NBC (NBCn1, known variously as NBC2 or NBC3; see Romero et al. 2004 for a review of NBC classification). The mRNA and protein product for NBCn1 has been detected in human and rat heart (Pushkin et al. 1999; Choi et al. 2000; Khandoudi et al. 2001).

Although the voltage-dependent changes in pH_i and membrane current reported here are consistent with expression of one or more NBC electrogenic gene products, the experiments do not provide direct evidence concerning the operation of NBCn. However, the following three observations suggest that NBCn may operate in parallel with NBCe in ventricular myocytes.

1. Voltage clamping E_m negative to E_NBC (calculated for n≥ 2) in a guinea pig ventricular myocyte reduced but did not terminate or reverse NBC-mediated pH_i recovery from an intracellular acid load. This suggests that, although NBCe was expected to have reversed, as predicted from the voltage dependence of I_NBC (Fig. 7), NBCn (whose thermodynamic energy source is independent of membrane potential) may have continued to operate in the forward mode, producing acid extrusion. Thus, the effect on pH_i recovery under these conditions would reflect the net activities of NBCe and NBCn.

2. From our SBFI-fluorescence studies, the fraction of Na⁺ influx attributable to NBC during pH_i regulation suggests an apparent NBC coupling coefficient of n≈ 1.4, i.e. a non-integer number (note that a range of values between n = 1 and n = 2 was statistically feasible). While this may be consistent with an ionic coupling of 3HCO₃⁻ : 2Na⁺, our values of E_NBC indicated n = 2 in all three species (Fig. 7C). The alternative interpretation of apparently non-integer values for ‘n’ is that a fraction of the myocyte's NBC population may operate with n = 1 while the rest operates with n = 2.

3. We have compared total acid efflux through NBC (J^tot_NBC) derived from the acid-efflux curves in Fig. 2A, with acid-efflux derived from I_NBC (i.e. electrogenic acid efflux, J_NBCe; Fig. 7C). Results for each species are summarized in Table 1. Values derived for J_NBCe are greatly influenced by the value assumed for the surface area to volume ratio (S/V) of the myocyte. Nevertheless, irrespective of various assumed S/V values, the results in all three species indicate that J^tot_NBC > J_NBCe, suggesting that both NBCe and NBCn transporters may be coactive within a ventricular myocyte. One may criticize these particular calculations on the grounds that they are based on measurement of low levels of NBC flux and I_NBC estimated at resting pH_i. Nevertheless, the resting value for total J_NBC presented in Fig. 2 is very similar to that reported previously in guinea pig myocytes (Leem et al. 1999), while the I_NBC value reported here for rat is in line with that of Aiello et al. (1998).

Table 1.

Estimate of J_NBCe/J^tot_NBC in resting myocytes under normal conditions

	INBC (pA pF−1)	volume (pl)	JNBCe (mm min−1)	JtotNBC (mm min−1)	JNBCe/JtotNBC
Rat	0.04	18a	0.40	0.68	0.6
	(pHi 7.05)	32b	0.22	0.68	0.3
		44c	0.16	0.68	0.2
Rabbit	0.04	29d	0.28	0.42	0.7
	(pHi 7.00)	31e	0.26	0.42	0.6
		38f	0.22	0.42	0.5
Guinea pig	0.02	18a	0.18	0.32	0.6
	(pHi 7.07)	46c	0.08	0.32	0.3
		28g	0.12	0.32	0.4

It was assumed that electrogenic NBC had a 1 : 2 stoichiometry and J_NBCe (mm min⁻¹ l⁻¹) was given as: J_NBCe= 2(I_NBC× C_m)/(F× cell volume), where C_m is membrane capacitance (pF) and F is the Faraday constant. Values for C_m were 149 pF (rat), 175 pF (rabbit) and 155 pF (guinea pig). Values of I_NBC at −85 mV were taken from Fig. 7C and those for J^tot_NBC at the same pH_i were taken from Fig. 2A. Surface/volume ratios (μm²μm⁻³) were:

^a0.84 (Satoh et al. 1996),

^b0.46 (Page, 1978),

^c0.34 (Page et al. 1971),

^d0.60 (Page & Surdyk-Droske, 1979),

^e0.56 (Page, 1978),

^f0.46 (Satoh et al. 1996).

^gFixed volume of 28 pl (Campbell et al. 1987).

Functional significance of Na⁺ influx through cardiac NBC

In addition to restoring pH_i following an acute intracellular acid load, activation of NBC and NHE induces Na⁺ influx (Fig. 8). The increase in [Na⁺]_i that accompanies NHE activation increases [Ca²⁺]_i (via effects on sarcolemmal Na⁺–Ca²⁺ exchange) and this contributes to the recovery of contraction during a sustained acidosis (Bountra & Vaughan-Jones, 1989; Harrison et al. 1992). It is well documented, however, that an overdrive of NHE, such as that accompanying postischaemic reperfusion in the heart, can cause [Ca²⁺]_i overload resulting in arrhythmia and myocardial injury (Dennis et al. 1990; Karmazyn et al. 1999). Inhibition of NHE reduces these effects by prolonging acidosis and by attenuating the [Ca²⁺]_i overload, through blockade of the NHE-induced rise in [Na⁺]_i (Scholz et al. 1993; An et al. 2001).

In addition to NHE, NBC activation may also significantly contribute to reperfusion injury. For example, inhibition of the transporter is reported to be cardioprotective in rat myocytes subjected to anoxic acidosis (Schäfer et al. 2000) and during postischaemic reperfusion of ferret (Vandenberg et al. 1993) and rat heart (Khandoudi et al. 2001). The latter study suggested a specific role for NBCe in promoting injury. The authors also reported significant up-regulation of NBCe in human cardiomyopathy. The present work is the first to show a net Na⁺ influx associated with NBC activity in a cardiac cell. Figure 8 shows that, while this is less than that mediated through NHE activity, it nevertheless drives a significant rise of [Na⁺]_i in the order of 1 mm min⁻¹ at pH_i 6.70. Low pH_i activation of NBC during reperfusion events in the heart may therefore promote [Ca²⁺]_i loading by inducing a rise of [Na⁺]_i, as occurs with NHE.

Significant NBCe activity in tissue exposed to the hyperkalaemia that typically accompanies ischaemia (e.g. Allen & Orchard, 1987) may serve to exacerbate [Ca²⁺]_i overload, as elevating [K⁺]_o considerably enhances activation of NBCe by intracellular acidosis (Fig. 6E). The Ca²⁺-overloading properties of NBCe activation will, however, be mitigated by the transporter's Na⁺-sparing properties. On NBCe (n = 2), an average of 0.5 Na⁺ ions will accompany the influx of one equivalent HCO₃⁻ ion, compared with 1.0 Na⁺ ion on NBCn, or 1.0 Na⁺ ion on NHE (in exchange for one H⁺ ion). Because of this Na⁺-sparing property, selective NBCe inhibitor drugs may be less cardioprotective than NBCn inhibitors or NHE inhibitors (like cariporide) during episodes of ischaemia–reperfusion. Exploration of the pharmacological properties and overall expression of different cardiac NBC gene products may form a useful approach to exploring the cardioprotective potential of NBC inhibition.