AJP-Heart Articles in PresS. Published on April 11, 2002 as DOI /ajpheart

Transcription

1 AJP-Heart Articles in PresS. Published on April 11, 2002 as DOI /ajpheart MS H /R1 THE ELECTROPHYSIOLOGICAL RESPONSE OF RAT ATRIAL MYOCYTES TO ACIDOSIS Kimiaki KOMUKAI, Fabien BRETTE and Clive H. ORCHARD School of Biomedical Sciences, University of Leeds, Leeds LS2 9NQ, UK Running title: Atrial myocytes and acidosis Correspondence should be sent to Professor C.H. Orchard School of Biomedical Sciences, University of Leeds, Leeds LS2 9NQ, UK Tel: +44-(0) Fax: +44-(0) Copyright 2002 by the American Physiological Society.

2 MS H /R1 1 SUMMARY The effect of acidosis on the electrical activity of isolated rat atrial myocytes was investigated using the patch clamp technique. Reducing the ph of the bathing solution from 7.4 to 6.5 shortened the action potential. Acidosis had no significant effect on transient outward or inward rectifier currents, but increased steady-state outward current. This increase was still present, although reduced, when intracellular Ca 2+ was buffered by BAPTA; BAPTA also inhibited acidosis-induced shortening of the action potential. 5 mm Ni 2+ had no significant effect on the acidosis-induced shortening of the action potential. Acidosis also increased inward current at 80 mv and depolarised the resting membrane potential. Acidosis activated an inwardly rectifying Cl - current that was blocked by DIDS, which also inhibited the acidosis-induced depolarisation of the resting membrane potential. It is concluded that an acidosis-induced increase in steady-state outward K + current underlies the shortening of the action potential and that an acidosis-induced increase in inwardly rectifying Cl - current underlies the depolarisation of the resting membrane potential during acidosis. Key words: action potential, potassium current, chloride current, perforated patch

3 MS H /R1 2 INTRODUCTION Cardiac muscle becomes acidic in a number of pathological conditions (27). Acidosis decreases the strength of contraction of the heart [for review see Orchard & Kentish (27)] and can induce arrhythmias [for review see Orchard & Cingolani (26)] both of which impair cardiac function. There are many mechanisms whereby acidosis can induce arrhythmias, among which is changing action potential duration: longer action potentials can produce triggered activity, shorter action potentials decrease the refractory period and can lead to premature contraction, and regional effects on the action potential can produce QT dispersion and re-entry tachyarrhythmia. Previous studies have shown marked effects of acidosis on the action potential in ventricular cells (3,19,21) that could produce arrhythmias by these mechanisms. However, despite the prevalence of atrial arrhythmias, the effect of acidosis on the atrial action potential is unknown. We have recently shown that acidosis prolongs the action potential in rat ventricular cells by inhibition of the steady state K + current (I SS ). However the response in atrial cells may be different. These cells have three distinct Ca 2+ -independent depolarization-induced outward K + currents: a rapidly activating, rapidly inactivating current (I TO,f ), a rapidly activating slowly inactivating current (I TO,s ), and a rapidly activating non-inactivating current (I SS ) (5,6). Although the kinetics of atrial I SS are similar to those of ventricular I SS, which is inhibited by acidosis, rat atrial I SS is carried by Kv1.5 (1,2,4,25) whereas rat ventricular I SS appears to be carried by Kv1.2 or Kv2.1 (29). Thus, the response of the action potential to acidosis may be different in rat atrial and ventricular cells. Acidosis also depolarises the resting membrane potential in ventricular cells, but the mechanism is unknown (26). Although a change in resting membrane potential will contribute to the electrophysiological response to acidosis, the response of the resting membrane potential to acidosis in atrial cells is unknown. In the present study we investigated the effect of acidosis on the action potential and resting membrane potential, and the underlying currents, in rat atrial cells. METHODS Cell isolation. Male Wistar rats weighing 220~250 g were stunned and then killed by cervical dislocation. The heart was quickly removed and washed with isolation solution (see below for composition) containing 0.5 mm CaCl 2. The aorta was cannulated and retrogradely perfused with isolation solution containing 0.5 mm CaCl 2, at ml/min and a temperature of 36.5±0.5 C. After confirming good contraction of the heart, it was perfused with isolation

4 MS H /R1 3 solution containing 0.1 mm EGTA (Sigma Chemical Co, Missouri, USA) for 4 min, and then with isolation solution containing 0.8 mg/ml collagenase (Type I, Worthington Biochemical Co, New Jersey, USA) and 0.08 mg/ml protease (Type XIV, Sigma) for 10 min. At the end of the perfusion, both atrial appendages were dissected from the heart and cut into small pieces, which were incubated with isolation solution containing 1% bovine serum albumin in addition to the above enzymes, for 10 min at 37 C. The tissue was filtered and the filtrate centrifuged at 40 g for 60 s. The supernatant was removed and the cells re-suspended in isolation solution containing 0.5 mm CaCl 2 and stored at room temperature. This process was repeated 4-5 times with the remaining tissue. Experimental set up. An aliquot of cells was transferred to the experimental chamber, which was mounted on the stage of an inverted microscope (Diaphot TMD, Nikon, Tokyo, Japan). In some experiments, the cells were illuminated with long-wavelength (>600 nm) light and the image detected using a CCD camera (WAT-92A, Watec, Japan) and monitor (Model PM-950, Ikegami Tsushinki Co Ltd, Utsunomiya, Japan). The patch clamp amplifier (8800 Total Clamp System, Dagan, Minnesota, USA) was controlled by a personal computer through an analogue/digital interface (Model 1401, Cambridge Electronic Design, Cambridge, England) using CED Patch and Voltage Clamp Software (Cambridge Electronic Design), which was also used for data acquisition. Voltage and current were also monitored using an oscilloscope (OS4100, Gould Inc, Hainault, England). Measurement of action potential and K + currents. The perforated patch clamp technique was used to monitor the action potential and K + currents (14,15,19,20). Electrodes (2-4 MΩ), were made from glass capillaries (GC150F-15, Clark Electromedical Instruments, Reading, England) using a vertical puller (PP-83, Narishige, Tokyo, Japan). The tip of the electrode was filled with pipette solution (see below for composition) and back-filled with pipette solution containing µg/ml amphotericin B. When the electrode tip was in the bath, junction potential was electronically offset to zero. After making a seal (>1 GΩ), capacitance was compensated and holding potential set to -40 mv. Five mv depolarising steps (20 ms in duration) were applied to monitor pore formation. The pipette solution contained 1 mm CaCl 2 to ensure that accidental rupture of the membrane resulted in cell death. Electrical access was usually obtained within min. After the capacity transients became constant, series resistance was compensated and measurements were made. Changing between control and acid solutions while the electrode tip was in the bath solution was used to ensure the absence of artefactual changes of potential. Action potentials were measured in current clamp mode. 3-5 ms current pulses, sufficient to trigger an action potential, were injected every 2 s

5 MS H /R1 4 from the amplifier. The duration of the action potential was measured at 25, 50 and 90% repolarization (APD 25, APD 50 and APD 90 respectively). Depolarisation-induced K + currents were recorded in the presence of 0.1 mm CdCl 2 (as were the recordings shown in figure 3 when K + was replaced with Cs + ) to block I Ca. 4-aminopyridine (4-AP) was not used for these experiments because of its non-specific inhibitory effects on K + currents (5). To measure transient outward current (I TO,f ), holding potential was set to -80 mv. Following a 40 ms prepulse to -40 mv [to inactivate Na + current (I Na )], 500 ms test pulses to voltages between -40 and +40 mv were applied; pulses were applied every 2 s. I TO,f was measured as the difference between the peak outward current and the current remaining at the end of pulse. Steady state outward current (I SS ) was monitored from a holding potential of -20 mv to avoid contamination by transient outward current, which is inactivated at this potential (ref. 6, see also Fig. 3A). A series of 500-ms test pulses to voltages between -40 and +40 mv was applied; pulses were applied every 2 s. I SS was measured at the end of the pulses. Inward rectifier current (I K1 ) was monitored from a holding potential of -40 mv. A series of 500-ms pulses to voltages between and -30 mv was applied; pulses were applied every 2 s. I K1 was measured at the end of the pulses. All experiments were performed at room temperature. Measurement of Cl - current. Conventional whole cell patch clamp was used to measure Cl - currents because Cl - diffusion between the pipette and the cell interior is restricted in the perforated patch configuration. A 3M KCl-Agar bridge/agcl pellet was used as the bath electrode. Holding potential was set to -40 mv. Two s test pulses to voltages between -120 and +40 mv (in 20 mv increments) followed by 400 ms pulses to +40 mv were applied; pulses were applied every 10 s (10). Current was measured at 2 s. All experiments were performed at room temperature. Solutions and chemicals. During isolation and cell storage, the isolation solution used contained (in mm): NaCl, 130; KCl, 5.4; MgCl 2.6H 2 O, 1.4; NaH 2 PO4, 0.4; creatine (Sigma), 10; taurine (Sigma), 20; HEPES ((N-[2-hydroxyethyl]piperazine-N -[2-ethanesulphonic acid]); Sigma), 5; glucose, 10. ph was adjusted to 7.30 using NaOH. When the action potential and K + currents were measured using the perforated patch clamp technique, the composition of the extracellular solution was (in mm): NaCl, 108; KCl, 5; Na 2 HPO 4.12H 2 O, 1; MgSO 4.7H 2 O, 1; Na acetate, 20; glucose 10; CaCl 2, 1; HEPES, 10; insulin, 5U/l. ph was first adjusted to 7.40 using NaOH, then HCl was used to decrease the ph of an aliquot of this solution to The composition of the pipette solution was (in mm): KCl, 130; NaCl, 10; MgCl 2.6H 2 O, 1.4; CaCl 2, 1; HEPES, 5. ph was adjusted to 7.10 using KOH. When Cl - current was measured using conventional whole cell clamp, the composition of the extracellular solution was (in mm):

6 MS H /R1 5 TEACl (Sigma), 120; CsCl (Sigma), 10; MgCl 2.6H 2 O, 1; HEPES, 10; BaCl 2 (Sigma), 2; glucose, 10; CaCl 2, 1; 4-aminopyridine (4-AP; Sigma), 2; nifedipine (Sigma), ph was first adjusted to 6.5 using TEAOH (Sigma) then an aliquot of the solution was adjusted to ph of 7.4 by further addition of TEAOH. The composition of the pipette solution was (in mm): CsCl, 110; TEACl, 20; MgATP (Sigma), 5; EGTA, 5; HEPES, 5. ph was adjusted to 7.1 using CsOH. All chemicals were purchased from BDH Laboratory Supplies (Poole, England) unless otherwise mentioned. Amphotericin-B was purchased from Sigma and a 50 mg/ml stock solution was made with dimethylsulfoxide (DMSO; Sigma) immediately before the experiments and diluted into the pipette solution before use. 1,2-Bis(2-aminophenoxy)-ethane-N,N,N,N - tetraacetic acid (BAPTA)-AM was purchased from Molecular Probes (Eugine, Oregon, US) and a 1 mm stock solution was made with dimethylsulfoxide (DMSO). A 10 mm stock solution of nifedipine and a 100 mm stock solution of 4,4 -disothiocyanato-stilbene-2,2 -disulfonic acid (DIDS; Sigma) were made with DMSO and kept in light-resistant containers. A 10 mm stock solution of strophanthidin (Sigma) was made with ethanol. All stock solutions were diluted into the perfusate immediately before use. Addition of DIDS (final concentration, up to 0.1 mm) did not alter ph of perfusate. Statistical analysis. Data are expressed as mean±sem for n cells. Paired or unpaired t tests (two-tailed) were performed as appropriate. When an unpaired t test was used, the variances were tested by F test. Statistically significance was taken as p<0.05. RESULTS The effect of acidosis on the action potential in atrial myocytes. Isolated atrial myocytes were current-clamped at 0.5 Hz. Fig. 1 shows representative action potentials recorded at control ph (open circle) and during acidosis (filled circle). Acidosis significantly shortened APD 90, from 79±7 ms to 48±6 ms (p<0.05, n=5), but did not alter APD 25 (4±0 ms at control ph, 4±0 ms in acidosis) or APD 50 (12±1 ms at control ph, 11±1 ms in acidosis). Acidosis also caused depolarization of the resting membrane potential, from 79±1 mv at control ph to 76±2 mv during acidosis (p<0.05, n=5). This change in resting potential is similar to that reported previously in ventricular cells. However, the abbreviation of the action potential by acidosis is in contrast to the prolongation reported in ventricular cells (see INTRODUCTION). Subsequent experiments were designed to investigate the current(s) responsible for these changes. Currents were normalised to cell capacitance and presented as current density.

7 MS H /R1 6 Mean cell capacitance, calculated from capacitance transients during a 10 mv hyperpolarising pulse from -80 mv, was 53±3 pf (n=38). The effect of acidosis on transient outward current in atrial myocytes. The transient outward current in rat atrial myocytes has two components: I TO,f and I TO,s (6), which have time constants of inactivation of ~180 ms and ~3 s, respectively (6). Thus, at the end of the 500 ms voltage clamp pulses used to monitor transient outward current (see METHODS), I TO,f will have decayed by ~94%, whereas I TO,s will have decreased by only ~15%. In addition, at the stimulation frequency used in the present study I TO,s will be mostly inactivated due to its slow recovery from inactivation (6). Thus the current measured as the difference between peak current and current at the end of the pulse (see METHODS) will reflect predominantly I TO,f. Figs. 2A and 2B show original traces of total outward current in a representative cell, elicited using the protocol described in the METHODS and shown in the inset in Fig. 2A, at control ph (A) and during acidosis (B). Fig. 2C shows the current at +60 mv in control and acidosis. When the currents are superimposed by offsetting one of the currents it is apparent that the amplitude and time course of I TO,f were identical in control and acidosis (Fig. 2C inset); because the currents can be superimposed in this way, this suggests that the acidosis-induced increase in outward current (compare fig. 2B with 2A) is a rapidly activating, non-inactivating current (see below), but that I TO,f is unchanged during acidosis. Fig. 2D shows mean current density-voltage relationships of I TO,f showing that acidosis did not alter I TO,f : current at +60 mv was 8.0±1.0 pa/pf in control and 6.7±1.5 pa/pf in acidosis (NS, n=6). At the end of the voltage clamp pulse used to elicit I TO,f (Fig. 2), membrane current is due to a small residual I TO,f (see above), I TO,s and the steady-state current (see below). Thus although I TO,s will be small under the conditions of the present experiments (above), it is possible to estimate this current by subtracting the steady-state current at the end of a 500 ms test pulse from a holding potential of -20 mv (see below) from the current at the end of the 500 ms test pulse to the same potential, but from a holding potential of -80 mv, used to elicit I TO (see also Ref. 6). Although small, the current measured in this way was not altered by acidosis: the current at +60 mv was 2.0±0.7 pa/pf in control and 2.5±0.6 pa/pf in acidosis (NS, n=6). Thus acidosis does not appear to alter I TO,f or I TO,s making it unlikely that changes in I TO underlie the observed changes in the action potential. The effect of acidosis on steady state outward current in atrial myocytes. I SS was evoked from a holding potential of -20 mv, to avoid contamination by I TO,f and I TO,s, which are inactivated at this potential, and was measured at the end of 500 ms test pulses (Fig. 3A inset, and see METHODS). Fig. 3A shows original traces of I SS in a representative cell at control ph.

8 MS H /R1 7 The cell is the same as presented in Figs. 2A-C: note that I TO is absent at this holding potential (c.f. Fig. 2). Fig. 3B shows original traces from the same cell during acidosis showing that acidosis increased this current. Fig. 3C shows the effect of acidosis on the current densityvoltage relationship showing that acidosis significantly increased I SS at all potentials: at +60 mv the current increased from 10.5±1.8 to 14.7±2.2 pa/pf (p<0.01, n=6). To determine whether this current was carried by K +, the effect of replacing the K + in the pipette and bathing solutions with Cs + was investigated. Fig. 3D shows that Cs + almost completely abolished this current at control ph, and that in the presence of Cs + acidosis did not increase outward current. Thus it appears that I SS is carried by K + and increased by acidosis. The effect of acidosis on steady state outward current in the presence of intracellular Ca 2+ buffering. In rat ventricular cells, acidosis increases intracellular [Ca 2+ ] and hence Ca 2+ - sensitive outward current (15). It seemed possible, therefore, that the increase in steady state outward current observed during acidosis in the present study was due to Ca 2+ -sensitive outward current. The above experiment was therefore repeated in the presence of BAPTA, to buffer intracellular Ca 2+. After establishing the perforated patch configuration, 200-ms depolarising pulses from -40 to 0 mv were applied every 2 sec and calcium current and cell contraction monitored. The cell was then exposed to 5 µm BAPTA/AM for 10 min; contraction was completely abolished within 5 min, although I Ca remained. The perfusate was changed to BAPTA-free Tyrode solution, and measurements made as described above from a holding potential of 20 mv. Fig. 4A shows the mean current-voltage relationship obtained at control ph (open squares) and during acidosis (filled squares) in the presence of BAPTA. At control ph, BAPTA did not significantly alter steady state outward current at any test potential (compare Fig. 3C and Fig. 4A, open symbols). In the presence of BAPTA, acidosis still increased I SS : at +60 mv the current increased from 8.0±1.1 to 9.3±1.7 pa/pf (n=7, p<0.01) and returned to after acidosis (NS versus pre-control). However Fig. 4B shows the difference current (acidosiscontrol; i.e. the increase in current produced by acidosis) in the absence and presence of BAPTA, showing that the increase induced by acidosis was significantly smaller in the presence of BAPTA. Because BAPTA reduced the acidosis-induced increase of I SS, its effect on the response of the action potential to acidosis was investigated, to test whether the increase in I SS might underlie the acidosis-induced abbreviation of the action potential. Fig 4C shows that BAPTA altered the configuration of the action potential at control ph (compare Fig 4C control with Fig 1 control), presumably by buffering the Ca 2+ transient and hence inhibiting inward Na + /Ca 2+

9 MS H /R1 8 exchange current (I Na/Ca ) during the action potential (inhibition of the exchange using Ni 2+ had a similar effect on action potential configuration, see below). In the presence of BAPTA acidosis caused a smaller decrease of action potential duration (Fig. 4C; APD 90 decreased from 53±10 ms to 41±9 ms, n=6, p<0.01, i.e. by 11±3 ms, compared with a decrease of 31±10 ms in the absence of BAPTA). These data are compatible with the idea that an acidosis-induced increase in Ca 2+ - sensitive and Ca 2+ -insensitive steady-state outward current underlie the abbreviation of the action potential observed during acidosis. The effect of acidosis on action potential in the presence of Ni 2+. It seems possible that changes in I Na/Ca during acidosis might also contribute to the abbreviation of the action potential (in which case inhibition of I Na/Ca by BAPTA would be expected to inhibit action potential shortening during acidosis). This seems feasible because Na + /Ca 2+ exchange is inhibited by acidosis (14,20,28,34) and modulated by intracellular Ca 2+. To test this idea, we investigated the effect of acidosis on the action potential in the presence of 5 mm NiCl 2, to block I Na/Ca (8). In the presence of Ni 2+ acidosis still shortened the action potential (APD 90 decreased from 72±17 to 53±17 ms, n=3, p<0.05). This decrease was not significantly different from that observed in the absence of Ni 2+, making it unlikely that changes in I Na/Ca contribute significantly to the abbreviation of the action potential observed during acidosis (although the decrease observed in the presence of Ni 2+ was slightly smaller than that observed in control, suggesting that I Na/Ca may play a small role). The effect of acidosis on I K1 in atrial myocytes. I K1 is responsible for late repolarisation and maintenance of the resting membrane potential. Thus I K1 was measured at control ph and during acidosis to investigate whether it could contribute to the observed changes in the action potential or resting potential. However acidosis had no significant effect on I K1 (not shown): at -120 mv the current measured at the end of the 500 ms test pulses used was -10.5±1.1 pa/pf at control ph, -10.4±1.2 pa/pf in acidosis (NS, n=5). It thus appears unlikely that I K1 plays a role in the observed changes in the action potential and resting potential. The mechanism of the resting membrane potential depolarisation during acidosis. To investigate the mechanism responsible for the depolarisation of the resting membrane potential during acidosis, the inward shift of holding current at -80 mv at control ph and during acidosis was monitored. Fig. 6C shows that acidosis significantly increased inward current at - 80 mv, and that this shift was not significantly different in the presence of BAPTA, when K + in

10 MS H /R1 9 the pipette and bathing solutions was replaced with Cs + or when pipette and bathing K + was replaced with Cs +, and 5 mm Ni 2+, 0.1 mm Ba 2+ and 20µM strophanthidin were present in the bathing solution (extracellular MgSO 4 was replaced with MgCl 2 and Na 2 HPO 4 was omitted to avoid precipitation with Ba 2+ ). Thus it appears unlikely that changes in Ca 2+ sensitive currents (including I Na/Ca ), L-type Ca current, K + currents or Na pump current (I Na/K ), underlie the acidosis-induced depolarisation of resting membrane potential. Fig. 6B shows that acidosis significantly depolarised the resting membrane potential in control, in the presence of Ni 2+ and BAPTA, as expected from the results above, but that 50 µm DIDS inhibited the acidosis-induced depolarisation of the resting membrane potential (Figs. 6A and 6B), suggesting that the depolarisation is due to acidosis-induced activation of a DIDSsensitive Cl - current. To test this idea further, the effect of acidosis on Cl - current was investigated. Figs. 7A and B show original traces of Cl - current recorded from a representative cell at control ph (A) and during acidosis (B) in the absence of Na + and K + and the presence of Cs +, TEA +, Ba 2+, nifedipine, 4-AP and internal EGTA as described in METHODS. Under these conditions (pipette [Cl - ] = 130 mm) an inwardly rectifying current was recorded with a reversal potential of 10 mv (Fig. 7C, open circles); reducing pipette [Cl - ] to 20 mm (replaced with aspartate - ) markedly decrease the measured current and shifted the reversal potential to 46 mv (Fig. 7C, open triangles). The similarity of the measured reversal potentials to the calculated Cl - reversal potentials (-2 mv and 49 mv respectively) suggests that the measured current is carried predominantly by Cl - ; the difference between the calculated and measured reversal potentials may represent imperfect equilibration between pipette and intracellular solutions. Acidosis increased inward current (Fig. 7C, filled symbols): when pipette [Cl - ] was 130 mm, current at -120 mv increased from -4.0±2.1 to -6.5±2.5 pa/pf (n=5, p<0.05). Fig. 7D shows that the current induced by acidosis (i.e. current in acidosis that at control ph) shows strong inward rectification and is decreased when pipette [Cl - ] is decreased: the current induced by acidosis at -120 mv was reduced from -2.5±0.6 pa/pf (n=5) to -0.4±0.2 pa/pf (n=3; p<0.05) when pipette [Cl - ] was reduced from 130 to 20 mm. The acidosis-induced increase in current was completely inhibited by DIDS (Fig. 7E), although DIDS had no significant effect on membrane current at control ph under these conditions or on resting membrane potential under control conditions (not shown). These data suggest, therefore, that the acidosis-induced depolarisation of the resting membrane potential is due to activation of a DIDS-sensitive, inwardly rectifying, Cl - current. The relationship between resting membrane potential and action potential duration.

11 MS H /R1 10 To test whether the observed change in action potential duration might be secondary to the acidosis-induced change in resting membrane potential per se, current injection was used to depolarise the resting membrane potential by 5 mv. However this manoeuvre tended to prolong action potential duration (not shown), making this unlikely. DISCUSSION In the present study, acidosis was produced by decreasing the ph of the perfusate (ph o ) from 7.4 to 6.5. Measurement of intracellular ph (ph i ) using the fluorescent dye BCECF has shown that ph i decreases by ~0.5 ph units in response to this change in ph o (20). These changes are, therefore, within the range observed pathophysiologically (27). ph i reaches a new steady state within 3 minutes; measurements in the present study were therefore carried out after 5 min exposure to the acid solution (15). In ventricular cells, the presence of a perforated patch electrode containing 5 mm HEPES does not affect the change of ph i that occurred on exposure to the acid solution (not shown); it therefore seems unlikely that it would do so in atrial cells. The perforated patch technique was used because acidosis alters the intracellular environment, increasing intracellular [Ca 2+ ] (12), intracellular [Na + ] (12), CaMKII activity via the increase in intracellular [Ca 2+ ] (13) and inhibiting protein phosphatase activity (23). We therefore used the perforated patch clamp technique where possible to minimise disruption to these normal responses, to investigate the physiological response to acidosis. The effect of acidosis on the atrial action potential The role of I Ca and I Na/Ca. We have previously shown that acidosis has little effect on I Ca in rat ventricular cells when the perforated patch technique is used (14,19,20). Because the channel that carries this current in atrial cells appears to be the same as in ventricular cells, it seems likely that acidosis has little effect on I Ca in atrial cells. Although inhibition of I Ca by Ni 2+ may contribute to the change in action potential configuration observed in the presence of Ni 2+, the observation that acidosis-induced shortening of the action potential still occurred in these conditions suggests that changes in I Ca are not necessary to account for the observed abbreviation of the action potential. Inhibition of Na + /Ca 2+ exchange shortens action potential duration (30). Acidosisinduced inhibition of Na + /Ca 2+ exchange (14,20,28,34) could, therefore, explain the observed abbreviation of the action potential during acidosis. However the present study shows that acidosis shortened the action potential duration even after the exchange had been inhibited by 5 mm Ni 2+ (Fig. 5), making it unlikely that inhibition of Na + /Ca 2+ exchange plays a major role in

12 MS H /R1 11 the acidosis-induced shortening of the action potential, although it is possible that it plays a small role (see RESULTS). The role of I TO,f and I TO,s. In the present study, acidosis did not alter I TO,f (Fig. 2) consistent with the observation that acidosis did not alter early repolarization (Fig. 1). Although atrial I TO,f is distinct from ventricular I TO in respect to its time course of activation and inactivation, the same channels (Kv4.2/4.3) are thought to underlie both currents. The present result is therefore consistent with the observation that acidosis does not alter I TO in rat ventricular cells when holding potential is negative to 60 mv (15). These data make it unlikely that the effect of acidosis on the action potential is mediated by a change in I TO,f. It is also unlikely that I TO,s is involved in the acidosis-induced shortening of the action potential because I TO,s is mostly inactivated in the conditions used in the present study, and appears to be unaffected by acidosis (see RESULTS), consistent with previous reports that Kv1.2, which appears to carry I TO,s, is unaffected by acidosis (32). The role of steady-state outward current. Kv1.5 is the most likely candidate for rat atrial I SS (25) because the time and voltage-dependent properties of heterologously expressed Kv1.5 channels are similar to those of rat atrial I SS, anti-kv1.5 antibody shows high levels of binding in rat atrial cells (2) and exposure of rat atrial cells to antisense oligodeoxynucleotides for Kv1.5 significantly inhibits I SS (4). The present data show that I SS is inhibited by replacing K + with Cs +, consistent with a K + current. I SS also appeared to have a Ca 2+ -insensitive component and a Ca 2+ -sensitive component that was only observed during acidosis (BAPTA decreased I SS during acidosis, but had no effect on I SS at control ph; Figs. 3 & 4). We are unaware of any previous work showing that I SS is Ca 2+ dependent, or has two components. However it is not clear whether these are two different components of I SS or whether the apparently Ca 2+ -insensitive component is due to incomplete buffering of Ca 2+ by BAPTA. In support of the latter suggestion, both components showed the same kinetics (rapidly activating, non-inactivating; not shown), and acidosis increased both components of I SS (Fig. 4B). Previous work has shown that acidosis decreases I SS in rat ventricular cells (see INTRODUCTION), although this may be because in rat ventricle I SS is carried by Kv1.2 or Kv2.1 (29). However acidosis (ph 7.3 to 6.3) also inhibits Kv1.5 current expressed in Xenopus oocytes (32), although this difference may be due to differences in the channel s environment and regulation in the two cell types (e.g. 9). The mechanism of action of acidosis on I SS is unclear, but is likely to be an effect on channel conductance rather than inactivation; the observation that the increase in steady-state current produced by acidosis is voltage independent

13 MS H /R1 12 (Fig. 3C) also suggests that it is unlikely that the effect of protons is within the channel pore. The observation that BAPTA inhibited the acidosis-induced decrease in action potential duration (Fig 4C), is compatible with the idea that an acidosis-induced increase in the Ca 2+ - sensitive component of I SS is responsible for much of the action potential shortening observed during acidosis. The role of I K1. In the present study, acidosis did not alter I K1 making it unlikely that changes in I K1 are responsible for the changes in the action potential observed during acidosis. I K1 in rat ventricle appears to be carried by Kir2.1 (24). Although the identity of rat atrial I K1 has not been established, Kir2.1 is ph-insensitive (37), consistent with the present data. The mechanism of acidosis-induced shortening of the atrial action potential. In the present study acidosis shortened the action potential of rat atrial cells. This is in contrast to previous work, which showed that acidosis prolongs the action potential of rat ventricular cells under identical experimental conditions. The abbreviation observed in the present study is unlikely to be due to the effects of acidosis on I Na/Ca, I Ca, I TO,f, I TO,s I K1 or I Cl,ir, or secondary to the depolarisation of the resting membrane potential (see above), but could be due predominantly to the effect of acidosis on I SS, which is increased in atrial cells, but inhibited in ventricular cells and which could, therefore, account for the different response in the two cell types. Unfortunately, there is no specific inhibitor of I SS with which to test this hypothesis although the observation that BAPTA inhibited the increase in I SS and the abbreviation of the action potential supports the idea. One problem with this hypothesis is that the effect of acidosis on I SS is greater at more positive potentials (Fig. 3) whereas the effect on the action potential is greater at more negative potentials (Fig. 1). Because many of the repolarising currents are voltage and time-dependent, the comparison of the action potential and the currents monitored under voltage clamp is not straightforward. However, a similar relationship between the configuration of the action potential and a change in I SS to that observed in the present study has been reported in response to phenylephrine (11), which produced a prolongation of late depolarisation which was ascribed to a decrease in I SS. It seems possible that the effect of the acidosis-induced increase in I SS on the action potential is masked during the early phase (APD 25 ; APD 50 ) of the action potential by the presence of other, large, currents (e.g. I Na, I TO,f ), and thus that an increase in I SS, possibly with a small contribution by altered I Na/Ca (see RESULTS), could underlie the abbreviation of the atrial action potential during acidosis. The effect of acidosis on the resting potential.

14 MS H /R1 13 Acidosis-induced depolarisation of the resting membrane potential has been observed by many investigators [for review see Orchard & Cingolani (26)] although the mechanism has remained obscure. The present study shows that the acidosis-induced increase in inward current, and hence resting membrane potential, is unlikely to be due to changes in Ca 2+ sensitive currents (including I Na/Ca ), Ca 2+ current, K + currents or Na + pump current (I Na/K ) (fig. 6B and C). However acidosis markedly increased inwardly rectifying Cl - current (I Cl,ir ), monitored using whole cell clamp (Fig. 7). This current, and the acidosis-induced increase in current, were reduced when pipette [Cl - ] was decreased (Fig. 7C,D): when pipette [Cl - ] was 20 mm, within the physiological range (10-20 mm; ref. 16), the acidosis-induced increase in current at 80 mv was pa/pf, similar to the increase in inward current observed during acidosis using perforated patch recording under more physiological conditions (~0.5 pa/pf; Fig. 6). The slightly higher current observed during perforated patch recording might be due to the high [Cl - ] in the pipette and a small permeability of the perforated patch to Cl - causing a small increase in intracellular [Cl - ]. The observation that the increase in inward current induced by acidosis was inhibited by decreasing pipette [Cl - ] and by DIDS (Figs. 7D and 7E), which also abolished the depolarisation of the resting membrane potential (Fig. 6) suggests that the depolarisation of the resting membrane potential is due to acidosis-induced activation of a DIDS-sensitive, inwardly rectifying, Cl - current. Assuming a membrane resistance of 118 MΩ, (range in rat ventricular myocytes MΩ, ref. 36), the small (28 pa) inward shift of current at 80 mv could account for the 3.3 mv depolarisation observed during acidosis. I Cl,ir does not appear to be due to protein kinase A-dependent Cl - current, Ca 2+ -activated Cl - current or swelling-induced Cl - current, because these do not show inward rectification (18, 31, 16). However this current shows similarities to that carried by ClC-2 channels, which have recently been found in atrial and ventricular cells (7,10); the Cl - current carried by these channels shows inward rectification (10,17), is of comparable magnitude to that recorded in the present study (10) and is increased by acidosis in Xenopus oocytes (17). The presence of Cd 2+, which blocks ClC-2, during measurement of depolarisation-induced K + currents (see METHODS) could then explain why this current was not observed during these K + current measurements, although the amplitude of I Cl,ir at physiological intracellular [Cl - ] is sufficiently small (Figs. 6 and 7) that it would be difficult to detect during measurement of relatively large K + currents, even in the absence of Cd 2+. One problem with this hypothesis is that these channels have been reported to be insensitive to stilbene derivatives such as DIDS, which agrees with the observation in the present study that DIDS did not alter I Cl,ir at control ph. It is possible, therefore, that acidosis alters the response to DIDS, either by affecting DIDS itself, or

15 MS H /R1 14 DIDS may block the stimulatory action of H + on ClC-2. Alternatively acidosis may increase I Cl,ir, not by acting on the channel, but by causing a DIDS-sensitive increase of intracellular [Cl - ], although the possible mechanism is unclear: the Cl - /OH - exchange (or H + -Cl - co-influx) mechanism described by Sun et al (33) is DIDS insensitive, while the DIDS-sensitive Na + - HCO3 - co-transporter and Cl - /HCO - 3 exchanger (22) would be inhibited in the HEPES buffered (HCO 3 -free) solutions used in the present study (33). In either case, it appears likely that ClC-2 channels carry the acidosis-induced I Cl,ir that underlies the depolarisation of the resting membrane potential during acidosis.

16 MS H /R1 15 Acknowledgments. This work was supported by British Heart Foundation and Wellcome Trust.

17 MS H /R1 16 REFERENCES 1. Barry, D.M., and J.M. Nerbonne. Myocardial potassium channels: electrophysiological and molecular diversity. Ann Rev Physiol 58: , Barry, D.M., J.S. Trimmer, J.P. Merlie, and J.M. Nerbonne. Differential expression of voltage-gated K channel subunits in adult rat heart. Circ Res 77: , Bethell, H.W., J.I. Vandenberg, G.A. Smith, and A.A. Grace. Changes in ventricular repolarization during acidosis and low-flow ischaemia. Am J Physiol 275:H551- H561, Bou-Abboud, E., and J.M. Nerbonne. Molecular correlates of the calcium-independent, depolarization-activated K currents in rat atrial myocytes. J Physiol (Lond) 517: , Boyle, W.A., and J.M. Nerbonne. A novel type of depolarization-activated K current in isolated adult rat atrial myocytes. Am J Physiol 260:H , Boyle, W.A., and J.M. Nerbonne. Two functionally distinct 4-aminopyridine-sensitive outward K currents in rat atrial myocytes. J Gen Physiol 100: , Britton, F.C., W.J. Hatton, C.F. Rossow, D. Duan, J.R. Hume, and B. Horowitz. Molecular distribution of volume-regulated chloride channels (ClC-2 and ClC-3) in cardiac tissues. Am J Physiol 279:H2225-H2233, Convery, M.K., and J.C. Hancox. Comparison of Na-Ca exchange current elicited from isolated rabbit ventricular myocytes by voltage ramp and step protocols. Pflugers Arch 437: , Deal, K.K., S.K. England, and M. Tamkun. Molecular physiology of cardiac potassium channels. Physiol Rev 76:49-67, Duan, D, L. Ye, F. Britton, B. Horowitz, and J.R. Hume. A novel anionic inward rectifier in native cardiac myocytes. Circ Res 86:e63-e71, Ertl, R., U. Jahnel, H. Nawrath, E. Carmeliet, and J. Vereecke. Differential electrophysiologic and inotropic effects of phenylephrine in atrial and ventricular heart muscle preparations from rats. Naunyn-Schmiedeberg s Arch Pharmacol 344: , Harrison, S.M., J.E. Frampton, E. McCall, M.R. Boyett, C.H. Orchard. Contraction and intracellular Ca 2+, Na + and H + during acidosis in rat ventricular myocytes. Am J Physiol 262:C348-C357, Hulme, J.T., J. Colyer, and C.H. Orchard. Acidosis alters the phosphorylation of Ser16 and Thr17 of phospholamban in rat cardiac muscle. Pflugers Arch 434: ,1997.

18 MS H /R Hulme, J.T., and C.H. Orchard. Effect of acidosis on Ca 2+ uptake and release by sarcoplasmic reticulum of intact rat ventricular myocytes. Am J Physiol 275:H977- H987, Hulme, J.T., and C.H. Orchard. Effect of acidosis on transient outward potassium current in isolated rat ventricular myocytes. Am J Physiol 278:H50-H59, Hume, J.R., D. Duan, M.L. Collier, J. Yamazaki, and B. Horowitz. Anion transport in heart. Physiol Rev 80:31-81, Jordt, S.-E., and T.J. Jentsch. Molecular dissection of gating in the ClC-2 chloride channel. EMBO J 16: , Kocic, I., Y. Hirano, M. Hiraoka. Ionic basis for membrane potential changes induced by hypoosmotic stress in guinea-pig ventricular myocytes. Cardiovasc Res 51:59-70, Komukai, K., C. Pascarel, and C.H. Orchard. Effect of acidosis on the action potential, I Ca and background steady-state current in isolated sub-endocardial and sub-epicardial rat ventricular myocytes (abstract). J Physiol (Lond) 526P:97P, Komukai, K., C. Pascarel, and C.H. Orchard. Compensatory role of CaMKII on I Ca and SR function during acidosis on rat ventricular myocytes. Pflugers Arch 442: , Kurachi, Y. The effects of intracellular protons on the electrical activity of single ventricular cells. Pflugers Arch 394: , Leem, C.H., D. Lagadic-Gossmann, and R.D. Vaughan-Jones. Characterization of intracellular ph regulation in the guinea-pig ventricular myocytes. J Physiol (Lond) 517: , Mundina-Weilenmann, C., L. Vittone, H.E. Cingolani, and C.H. Orchard. Effects of acidosis on phosphorylation of phospholamban and troponin I in rat cardiac muscle. Am J Physiol 270:C107-C114, Nakamura, T.Y., M. Artman, B. Rudy, and W.A. Coetzee. Inhibition of rat ventricular IK1 with antisense oligonucleotides targeted to Kir2.1 mrna. Am J Physiol 274:H892- H900, Nerbonne, J.M. (2000). Molecular basis of functional voltage-gated K channel diversity in the mammalian myocardium. J. Physiol (Lond) 525: , Orchard, C.H., and H.E. Cingolani (1994). Acidosis and arrhythmias in cardiac muscles. Cardiovasc Res 28: , Orchard, C.H., and J.C. Kentish (1990). The effects of changes of ph on the contractile function of cardiac muscle. Am J Physiol 258:C967-C981, Philipson, K.D., M.M. Bersohn, A.Y. Nishimoto (1982). Effects of ph on Na + -Ca 2+

19 MS H /R1 18 exchange in canine cardiac sarcolemma. Circ Res 50: , Scamps, F. Characterization of a beta-adrenergically inhibited K + current in rat cardiac ventricular cells. J Physiol (Lond) 491:81-97, Schouten, V.J.A., H.E.D.J. ter Keurs. The slow repolarization phase of the action potential in rat heart. J Physiol (Lond) 360:13-25, Sorota, S. Insights into the structure, distribution and function of the cardiac chloride channels. Cardiovasc Res 42, , Steidl, J.V., and A.J. Yool. Differential sensitivity of voltage-gated potassium channels Kv1.5 and Kv1.2 to acidic ph and molecular identification of ph sensor. Mol Pharmacol 55: , Sun, B., C.H. Leem, and R.D. Vaughan-Jones. Novel chloride-dependent acid loader in the guinea-pig ventricular myocytes: part of a dual acid-loading mechanism. J Physiol (Lond) 495:65-82, Teracciano, C.M.N., and K.T. MacLeod. Effects of acidosis on Na + /Ca 2+ exchange and consequences for relaxation in guinea pig cardiac myocytes. Am J Physiol 267:H477- H487, Wang, Z., L. Yue, M. White, G. Pelletier, and S. Nattel. Differential distribution of inward rectifier potassium channel transcripts in human atrium versus ventricle. Circulation 98: , White, R.L., D.C. Spray, A.C. Campos de Carvalho, B.A. Wittenberg, and M.V.L. Bennett. Some electrical and pharmacological properties of gap junctions between adult ventricular myocytes. Am J Physiol 249: C447-C455, Zhu, G., C. Liu, Z. Qu, S. Chanchevalap, H. Xu, and C. Jiang. CO 2 inhibits specific inward rectifier K+ channels by decrease in intracellular and extracellular ph. J Cell Physiol 183:53-64,2000.

20 MS H /R1 19 FIGURE LEGENDS Fig. 1. The effect of acidosis on the atrial action potential. Action potentials recorded from a representative atrial cell at control ph ( ) and during acidosis ( ). Dashed line indicates 0 mv. Fig. 2. The effect of acidosis on depolarization-induced outward current from a holding potential of -80 mv. A,B: original traces of depolarisation-induced outward current in a representative atrial cell in control (A) and acidosis (B). The protocol is shown schematically in the inset: the cell was held at -80 mv, and depolarised to a series of test voltages between -40 and +60 mv for 500 ms at 0.5 Hz. The arrow indicates zero current level. The capacitance of the cell was 40 pf. C: original traces of the current at +60 mv in control ( ) and acidosis ( ). When the currents are superimposed it is clear that the amplitude and time course were identical (inset). D: mean (±SEM) current density-voltage relationship of I TO,f in atrial cells at control ph ( ) and during acidosis ( ). Transient outward current was measured as the difference between the peak outward current and the current remaining at the end of pulses (see text for details). Fig. 3. The effect of acidosis on depolarization-induced outward current from a holding potential of -20 mv. A,B: original traces of depolarisation-induced outward current in the same cell as in Figs. 2A-C in control (A) and during acidosis (B). The inset shows the protocol: the cell was held at -20 mv, and hyperpolarised or depolarised to a series of test voltages between -40 and +60 mv for 500 ms at 0.5 Hz. The arrow indicates zero current level. C: the mean (±SEM) current density-voltage relationship of steady state outward current in atrial cells at control ph ( ) and during acidosis ( ). D: mean (±SEM) current density-voltage relationship of steady state outward current in atrial cells at control ph ( ) and during acidosis ( ) when K + in the pipette solution and perfusate was replaced with Cs +. Steady state outward current was measured at the end of pulse as described in the text. *, P<0.05 versus control. **, p<0.01 versus control. Fig. 4. The effect of acidosis on depolarization-induced outward current in the presence of intracellular Ca 2+ buffering. A: mean (±SEM) current density-voltage relationship of steady state outward current in atrial cells in the presence of BAPTA at control ph ( ) and during acidosis ( ). Following treatment with 5 µm BAPTA/AM, the same protocol was used as in Fig. 3 (see text for details). *, p<0.05 versus control. **, p<0.01 versus control. B: mean (±SEM) difference currents (acidosis-control) obtained in the absence ( ) and presence ( ) of

21 MS H /R1 20 BAPTA/AM. *, p<0.05 versus in the presence of BAPTA/AM. C: Action potentials recorded from a representative atrial cell after treatment with BAPTA/AM at control ph ( ) and during acidosis ( ). Dashed line indicates 0 mv. Fig. 5. The effect of acidosis on action potentials in the presence of Ni 2+. Action potentials recorded from a representative atrial cell in the presence of 5 mm Ni 2+ at control ph ( ) and during acidosis ( ). Dashed line indicates 0 mv. Figure 6. The effect of acidosis on resting membrane potential. A: action potentials recorded from representative atrial cell in the presence of 50 µm DIDS at control ph ( ) and during acidosis ( ). B: mean (±SEM) acidosis-induced depolarisation of resting membrane potential (from top to bottom): in control; +5 mm Ni 2+ ; +intracellular BAPTA; +50 µm DIDS. C: mean (±SEM) acidosis-induced inward shift of holding current at -80 mv (from top to bottom) in control; + intracellular BAPTA; with replacement of K + in the pipette solution and perfusate with Cs + ; with K + in the pipette solution and perfusate replaced with Cs +, and 5 mm Ni 2+, 0.1 mm Ba 2+, 20 µm strophanthidin present in the perfusate. N numbers shown in the parenthesis. *, p<0.05 versus control. All changes in resting membrane potential and holding current were significantly different from baseline (P<0.05) except that in the presence of DIDS (NS). Fig. 7. The effect of acidosis on Cl - current. A, B: original traces of Cl - current using ruptured patch in a representative atrial cell at control ph (A) and during acidosis (B); pipette [Cl - ] 130 mm. The cell was held at -40 mv and hyperpolarised to a series of test voltages between -120 and +40 mv in 20 mv increments for 2 s followed by 400 ms pulses to +40 mv at 0.1 Hz. The arrow indicates zero current level. C: mean (±SEM) current density-voltage relationships of Cl - current at control ph (open symbols) and during acidosis (filled symbols) when pipette [Cl - ] was 130 mm (circles) or 20 mm (triangles). *, P<0.05 versus control ph. D: mean (±SEM) difference current (acidosis-control) when pipette [Cl - ] was 130 mm ( ) or 20 mm (filled triangles). *, P<0.05 versus 130 mm pipette [Cl - ]. E: mean (±SEM) difference current in the absence ( ) and presence ( ) of 0.1 mm DIDS; pipette [Cl - ] 130 mm. *, P<0.05 versus in the absence of DIDS.

22 MS H /R1 21

23 MS H /R1 22

24 MS H /R1 23

25 MS H /R1 24

26 MS H /R1 25

27 MS H /R1 26

28 MS H /R1 27