INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

THE RAPID COMPONENT of delayed rectifier K channel current (I_Kr) has been described for ventricular and atrial myocytes from many species, including humans (13, 28). I_Kr has unique gating properties that are important for its role in determining the cardiac action potential configuration. It has a fast C-type inactivation process probably due to a constriction of its outer mouth during membrane depolarization (14, 22, 23). Membrane repolarization causes an equally fast recovery from inactivation. Together, the voltage-dependent fast inactivation and reactivation processes of I_Kr channel create an inwardly rectifying current-voltage (I-V) relationship. The decrease in outward current at depolarized voltages economizes the amount of inward current needed to maintain the plateau amplitude during phase two of the cardiac action potential (6, 32). This is especially important for the ventricular action potential, of which the positive plateau phase is crucial for Ca influx and excitation-contraction coupling. The increase in outward current on repolarization, on the other hand, ensures that sufficient outward current is provided to support proper phase three repolarization (6, 32).

Another unique aspect of I_Kr gating is its slow rate of deactivation (5). This is especially the case when the resting membrane potential is partially depolarized (e.g., for human cardiac I_Kr deactivation at 60 mV and 36°C see Fig. 1 of Ref. 28 and Fig. 5 of Ref. 13). This may cause a "use-dependent" contribution of I_Kr to action potential repolarization during certain conditions. The I_Kr channels may be fully activated during a single action potential when there is a long and positive plateau phase, e.g., normal ventricular action potentials (6, 13, 28, 32). However, when the action potential duration is short and/or when the plateau phase is suppressed [e.g., some atrial action potentials (3, 27) or ventricular action potentials under abnormal conditions such as ischemia and reperfusion (9)], I_Kr channels may not reach a fully activated state during a single action potential. This, in conjunction with the slow rate of I_Kr deactivation at partially depolarized resting membrane potential, may lead to an accumulation of I_Kr channel activation at fast heart rates.

The issue of how pathological conditions of the heart can modulate the I_Kr channel function is important for an understanding of mechanisms for arrhythmias that occur under these abnormal conditions. This information is also important for the design of antiarrhythmic strategies. The focus of our work here was how extracellular H⁺ can modulate I_Kr function. Extracellular acidosis occurs during ischemia and as a result of inflammation (9). Its effects on Na, Ca, and some K channels have been examined (17, 18, 25, 30), but the information on possible changes in I_Kr is lacking. We used human ether-à-go-go-related gene (HERG) expressed in Xenopus oocytes as a model for I_Kr. HERG encodes an inwardly rectifying K channel similar to I_Kr in cardiac myocytes, although the kinetics of channel gating may be influenced by the expression system and the recording temperature (6, 19, 32). HERG has been a valuable tool in mechanistic studies of the unique gating characteristics of I_Kr or its modulation by drugs or by cellular milieu. Our results showed that the most prominent effect of extracellular acidification (in the range of pH_o 8.5-6.5) was an acceleration of HERG-channel deactivation. This occurred with little or no changes in other gating kinetics or in the voltage dependence of activation. Extrapolating these observations to the situation in the heart suggests that extracellular acidosis may reduce the use-dependent contribution of I_Kr to action potential repolarization under pathological conditions.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Site-directed mutagenesis and cRNA in vitro transcription. HERG wild-type (WT) in pAlter-Max was subjected to oligonucleotide-directed mutagenesis according to the manufacturer's instructions (Altered Sites II Mammalian Mutagenesis System, Promega, Madison, WI). An NH₂-terminal deletion mutant of HERG (HERG N, lacking amino acids 2-354) (23) was also used. Plasmid DNAs were linearized for in vitro transcription. The transcription reactions were performed with a commercial kit (mMessage mMachine, Ambion, Austin, TX) and T7 RNA polymerase. Denaturing agarose gel electrophoresis was used to check the quality of cRNA product of each transcription reaction and to quantify the yield. cRNA was dissolved in RNase-free water for oocyte injection.

Oocyte preparation and injection. The oocytes of Xenopus laevis were isolated by partial ovariectomy. Follicular cell layers were removed mechanically after mild digestion with collagenase (type B, Boehringer Mannheim, Indianapolis, IN). Four to six hours after isolation, oocytes were injected with cRNA solutions using a Drummond digital microdispenser. The volume injected was 30-50 nl per oocyte. The oocytes were incubated at 16°C in ND-96 solution (in mM: 96 NaCl, 2 KCl, 1.8 CaCl₂, 1 MgCl₂, 5 HEPES, and 2.5 Na pyruvate, pH 7.5 with NaOH), supplemented with penicillin (50 U/ml), streptomycin (50 µg/ml), gentamycin (10 µg/ml), and horse serum (4%). The oocytes were studied 2-6 days after injection.

Electrophysiological experiments. The oocytes were placed in a tissue chamber and superfused at room temperature (23-25°C) with a low-Cl solution to minimize interference from endogenous Ca-activated Cl currents. The solution had the following composition (in mM): 96 NaOH, 2 KOH, 1 MgSO₄, 1.8 CaCl₂, 5 HEPES, and 2.5 Na pyruvate. The pH was titrated to 6.5, 7.5, or 8.5 with methanesulfonic acid. The flow rate was maintained at ~10 ml/min, allowing a total exchange of bath solution in 15-30 s after the valve that controlled the solution flowing into the bath was switched. Membrane currents were studied with the two-microelectrode voltage-clamp technique, with an oocyte clamp amplifier (model OC-725B, Warner, Hamden, CT). Both the voltage-recording and the current-passing electrodes were made of "agarose-cushion pipettes" of low tip resistance (0.1-0.2 M) to improve the quality of voltage clamp (20). In some experiments, diethylpyrocarbonate (DEPC, Sigma) was added to the bath solution to reach a final concentration of 0.2-2 mM, and the solution was immediately used to superfuse the oocyte.

Data acquisition and analysis. The generation of voltage-clamp protocols and data acquisition was controlled by an IBM AT-compatible computer with Clampex of pCLAMP via a 12-bit digital-to-analog and analog-to-digital converter (TL-1 DMA Interface, Axon Instruments, Foster City, CA). Currents were low-pass filtered with an eight-pole Bessel filter (Frequency Devices, Haverhill, MA) at 2 kHz, digitized online, and stored on diskettes for offline analysis. The sampling interval for whole cell currents ranged from 0.1 to 1 ms.

Voltage-clamp protocols and methods of data analysis will be described in figure legends. In all our voltage-clamp protocols, there was a 20-ms prepulse from the holding voltage of 80 to 100 mV. The resulting current step was used for linear leak subtraction during data analysis. Data analysis was mainly carried out using Clampfit (version 6.1). PeakFit (Jandel Scientific, Corte Madera, CA) was used to fit the activation curves. When appropriate, data are presented as means ± SE. Statistical analysis of paired or unpaired t-test was performed using SigmaStat (Jandel Scientific Software, San Rafael, CA). Statistical significance is determined at a P value of 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effects of changing pH_o on HERG WT channel. Figure 1 illustrates the effects of changing pH_o from 7.5 to 6.5 (A) or from 7.5 to 8.5 (B) on the HERG WT currents. The most dramatic change was an accelerated decay of tail currents at 80 mV (deactivation) at the acidic extracellular pH (pH_o) (6.5). At pH_o 8.5, the deactivation rate was modestly slowed relative to that at pH_o 7.5. Figure 2A clearly shows a slowing of tail currents recorded at 60, 70, and 80 mV from the same oocyte when pH_o was changed from 6.5 to 7.5 and then 8.5. Figure 2B summarizes the kinetics of HERG deactivation over a voltage range of 50 to 120 mV at various pH_o levels. The decay phase of HERG tail currents could be well described by a double-exponential function. Changing pH_o from 7.5 to 6.5 shortened both the fast (_fast) and the slow (_slow) time constants and increased the fraction of the fast-decay component. These acidification-induced changes were less prominent as the voltage approached the reversal potential (98.2 ± 0.7 mV) and disappeared at 120 mV. Figure 2B also shows that changing pH_o from 7.5 to 8.5 slowed the HERG deactivation in the voltage range positive or equal to 100 mV by lengthening both _fast and _slow and by reducing the fast component of decay.

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Effects of changing extracellular pH (pHo) on wild-type human ether-à-go-go-related gene (HERG WT) current. A: effects of changing pHo from 7.5 to 6.5. B: effects of changing pHo from 7.5 to 8.5 in another oocyte. Currents were elicited by membrane depolarization from a holding voltage (Vh) of 80 mV to various test voltages (Vt; from 70 to +60 mV in 10-mV increments) for 1 s once every 15 s.

View larger version (29K): [in this window] [in a new window]   Fig. 2.   Changing pHo markedly affected time course of HERG WT deactivation. A: representative current traces recorded from one oocyte at specified pHo. Currents were elicited by depolarization to +60 mV followed by repolarization to 60, 70, and 80 mV. B: summary data of kinetics of HERG deactivation at different levels of pHo. Decay phase of HERG tail currents was fit with a double exponential function to estimate fast and slow time constants of decay (fast and slow, respectively, plotted on a logarithmic scale) and fraction of fast-decay component. These are shown as solid (pHo 6.5) and shaded (pHo 8.5) symbols with SE bars and connected by solid lines. Numbers in parentheses are measurements. When no SE bars are seen, they are smaller than symbols. Open symbols connected by dotted lines are data corrected for voltage shift associated with changing pHo. This was based on shift in mean value of half-maximal activation voltage (V0.5, see Fig. 4B): changing pHo from 7.5 to 6.5 shifted V0.5 by +2.3 mV and changing pHo to 8.5 shifted V0.5 by 3.6 mV. In this figure and Figs. 3, 4, 7-9, and 11, statistical analysis was carried out on data before correction for voltage shift, using paired t-test. * P < 0.05 for pHo 6.5 vs. 7.5; # P < 0.05 for pHo 8.5 vs. 7.5.

We further examined whether there were other changes in the HERG kinetics associated with alterations in pH_o. The results are summarized in Fig. 3. Changing pH_o between 6.5 and 8.5 did not alter the time constants of activation in the voltage range from 40 to +60 mV (Fig. 3A). There were no changes in the time constants of inactivation in the voltage range from 40 to +30 mV (Fig. 3B). The time constants of reactivation estimated in the voltage range from 30 to 80 mV were not altered by elevating pH_o from 7.5 to 8.5. Lowering pH_o to 6.5, however, induced a modest acceleration of the rate of reactivation (Fig. 3B).

View larger version (29K): [in this window] [in a new window]   Fig. 3.   Effects of changing pHo on kinetics of activation, inactivation and reactivation of HERG WT. A: time constants () of activation at specified pHo. Values of activation () were determined by one of two methods: by fitting current traces recorded during depolarization pulses with a single exponential function ("direct fit," open symbols) or by fitting envelope of peak tail currents at 80 mV following depolarization pulses of different durations ("tail envelope," filled symbols; depolarization pulse durations ranging from 50 to 1,550 ms). B:  of inactivation (open symbols) and recovery from inactivation (reactivation, solid symbols) at specified pHo. The  values of inactivation were determined using a 3-step protocol: from Vh 80 mV, channels were activated and inactivated by a depolarization pulse to +60 mV for 300 ms. Membrane was then repolarized to 80 mV for 20 ms to allow channels to recover from inactivation without appreciable deactivation. Membrane was stepped to +30 to 40 mV for 80 ms during which fully activated channels inactivated. Decay phase of current during third step was fit with a single exponential function to estimate  of inactivation. Time constants of reactivation were determined using following voltage-clamp protocol: from Vh 80 mV, membrane was depolarized to +60 mV for 100 ms followed by repolarization to various voltages (Vr) ranging from 30 to 80 mV. Fastest  values estimated from rising phase of tail currents were used as a measurement of  of reactivation. * P < 0.05 for pHo 6.5 vs. 7.5.

Another noticeable change in the HERG WT currents when pH_o was switched from 7.5 to 6.5 was an increase in the outward currents during depolarizing test pulses [test-pulse current (I_t; Fig. 1A)]. The I-V relationships of I_t measured at the end of 1-s test pulses at various pH_o levels are summarized in Fig. 4A. Extracellular acidification led to an increase in the I_t amplitudes in the voltage range from 10 to +30 mV. At more depolarized voltages, lowering pH_o enhanced the initial peak outward current and accentuated the subsequent decay phase (e.g., current traces recorded at +60 mV shown in Fig. 4A, inset), although the I_t amplitudes measured at 1 s were little affected. These observations suggest that H⁺ in the outer-mouth region could interfere with C-type inactivation of the HERG channel (11, 22). This is consistent with the observation shown in Fig. 3B that lowering pH_o accelerated the transition from the C-type inactivated state to the open state (reactivation). The I_t amplitudes at pH_o 8.5 were modestly increased relative to those at pH_o 7.5 in the voltage range from 40 to 10 mV. This was most likely due to a negative shift in the voltage dependence of channel activation (Fig. 4B, discussed below).

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Effects of changing pHo on current-voltage (I-V) relationships and voltage dependence of activation of HERG WT. A: I-V relationships of test-pulse current (It) at specified pHo. Voltage-clamp protocol was same as that described for Fig. 1. For each oocyte, It amplitudes were measured at end of 1-s test pulses, leak subtracted, and normalized by maximal It at pHo 7.5 and Vt 10 or 0 mV (normalized It). Inset: superimposed test-pulse currents recorded at +60 mV from one cell at three levels of pHo. B: voltage-dependence of activation at specified pHo. Activation curves were constructed by normalizing peak amplitudes of tail currents measured at 80 mV following 1-s test pulses to different test voltages by maximal tail current following Vt to +60 mV. This gave an estimate of fractions of channels activated. Relationship between fraction activated and Vt was fit with a simple Boltzmann function to give V0.5 and slope factor (k): fraction activated = 1/[1 + exp(V0.5Vt)/k]. Curves superimposed on data points were calculated from this equation, where pHo 6.5, V0.5 = 10 mV, k = 9.6 mV; pHo 7.5, V0.5 = 12.3 mV, k = 9.0 mV; and pHo 8.5, V0.5 = 15.9 mV, k = 9.1 mV. C: I-V relationships of fully activated current (Ifa) at specified pHo. Voltage-clamp protocol was the following: from a Vh of 80 mV, membrane was depolarized to +60 mV for 100 ms to fully activate channels. This was followed by repolarization to different voltages (Vr) for 150 ms, during which channels recovered from inactivation and (at Vr negative to 30 mV) deactivated. Plateau or peak amplitudes of tail currents at various Vr were measured, leak subtracted, and normalized by peak tail current measured at pHo 7.5 and Vr 60 mV (normalized Ifa). * P < 0.05 for pHo 6.5 vs. 7.5; # P < 0.05 for pHo 8.5 vs. 7.5.

Therefore, the most pronounced change in the HERG-channel function when pH_o was altered between 6.5 and 8.5 was an acceleration of deactivation associated with acidification. Similar observations were obtained when the bath solution contained Cl as the major anion (bath solution titrated with HCl instead of methanesulfonic acid, data not shown). This rules out the possibility that the observed effects of extracellular acidification on HERG depended on the anion species present in the bath solution. We consider several possible mechanisms by which extracellular acidification could modulate HERG-channel function. However, the data presented so far can rule out two of them.

Conventional "surface potential" theory could not explain effects of extracellular acidification on HERG. The conventional surface potential theory says that there are uniform, fixed negative charges on the external surface of the cell membrane (7). Increasing the extracellular H⁺ concentration ([H⁺]_o) can titrate these negative charges, causing a positive shift in the external surface potential. This will affect the electric field sensed by the voltage sensor within the membrane lipid bilayer, leading to a parallel shift in the positive direction of voltage dependencies of all channel gating functions (7). Figure 4B shows that, indeed, changing pH_o from 7.5 to 6.5 shifted the half-maximal activation voltage (V_0.5) from 12.3 ± 0.4 to 10.0 ± 1.0 mV (P = 0.046 by paired t-test). However, as shown by the data points in Fig. 2B that have been corrected for the voltage shift, the amount of voltage shift seen at the acidic pH_o (<3 mV) was much less than that required to account for the alterations in the rate of channel deactivation. Furthermore, the observations that extracellular acidification markedly accelerated the rate of deactivation with little or no effects on other kinetic parameters were not compatible with predictions of the conventional surface potential theory.

Acceleration of HERG deactivation by extracellular acidification was not due to pore blockade by H⁺. We explain the acidification-induced increase in I_t amplitude by an interference of C-type inactivation by H⁺ in the outer-mouth region (Figs. 1A and 4A) (11). There was no sign of pore blockade by H⁺ at depolarized voltages. However, on repolarization, the membrane electric field might move H⁺ in the outer mouth deeper into the pore, which could then impede ion flux through the pore. Pore blockade might develop in a time-dependent fashion and manifest as an apparent acceleration of decay of tail currents, although the peak amplitudes of tail currents at 80 mV were little affected (Fig. 1A). This scheme was not compatible with two of our observations. First, the degree of acidification-induced changes in the kinetics of deactivation was reduced at more hyperpolarized voltages (Fig. 2B). This was opposite the prediction based on a voltage-dependent pore blockade by external H⁺. Furthermore, if there were a voltage-dependent pore blockade by H⁺, stepping the membrane voltage to very negative levels should have revealed a reduction in the peak amplitudes of tail currents. Figure 4C shows that the peak amplitudes of tail currents were little affected at voltages as negative as 150 mV.

Did extracellular acidification alter rate of deactivation by protonating histidine residues on extracellular domain of HERG? Extracellular H⁺ may affect ion channel function by titrating the side chains of amino acids that are critical for channel function. The pKa value of the side chain of histidine is 6.2, making it the most likely candidate for variations in side-chain protonation in the pH_o range tested here. The third possibility we explored was that the effects of extracellular acidification on HERG were due to a protonation of histidine residues exposed to the extracellular solution. This might affect the rate of channel deactivation by an unidentified mechanism. This would be similar to the situation of delayed rectifier K channels in squid axon, in which histidine residues exposed to the extracellular solution are involved in determining the kinetics of channel gating (although in the squid axon K channels the histidine residues are important for the rate of activation but not that of deactivation) (24).

There are only two histidine residues in the extracellular domain of HERG. Both are in the loop between the S5 and the pore region (S5-P loop, Fig. 5). We mutated both histidine residues to negatively charged glutamate (H578E/H587E). The H578E/H587E mutant encoded an inwardly rectifying K channel similar to the HERG WT channel (Fig. 6), except that at voltages positive to +30 mV, the rate and steady-state level of C-type inactivation were less than those of the WT channel. The voltage dependence of activation of H578E/H587E [V_0.5 = 7.6 ± 1.6 and slope factor (k) = 8.1 ± 0.4 mV, n = 7] was similar to that of the WT channel, indicating that adding negative charges to the S5-P loop did not perturb the electric field sensed by the S4 domain. Similar to the WT channel, changing pH_o from 7.5 to 6.5 and to 8.5 only induced small degrees of shift in the V_0.5 of H578E/H587E activation (mean value of V_0.5 shifted by +3.3 and 2.0 mV, respectively). Importantly, lowering pH_o from 7.5 to 6.5 markedly accelerated the rate of deactivation of the H578E/H587E mutant, whereas elevating pH_o to 8.5 induced a modest slowing. The data are summarized in Fig. 7. These changes in the kinetics of deactivation could not be accounted for by the voltage shift associated with alterations in pH_o. Therefore, the two histidine residues on the extracellular loop of HERG were not involved in the acidification-induced alterations in deactivation kinetics.

View larger version (34K): [in this window] [in a new window]   Fig. 5.   Transmembrane topology of a HERG subunit and charge distribution on extracellular surface. Transmembrane segments are labeled as S1-S6. Pore region is denoted. Mutation sites are highlighted: H578 and H587 are in extracellular loop connecting S5 and pore region; G628 and S631 are in outer-mouth region of pore; N represents a deletion of amino acids 2-354 from cytoplasmic NH2 terminus.

View larger version (34K): [in this window] [in a new window]   Fig. 6.   Removing the only two histidine residues from extracellular domain of HERG did not prevent effects of changing pHo on rate of deactivation. Both histidine residues at positions 578 and 587 were replaced by glutamate (H578E/H587E). Voltage-clamp protocol was same as that described for Fig. 1. Currents in A and B were from two oocytes.

View larger version (23K): [in this window] [in a new window]   Fig. 7.   Summary data of kinetics of HERG H578E/H587E deactivation at different levels of pHo. Data analysis and figure format were same as those described for Fig. 2B. Decay phase of tail currents was fit with a double exponential function to estimate fast, slow, and fast component of deactivation. These are shown as solid (pHo 6.5) and shaded (pHo 8.5) symbols connected by solid lines. Corresponding open symbols are data corrected for shift in V0.5 of activation associated with changing pHo: +3.3 mV for pHo 6.5 and 2.0 mV for pHo 8.5. * P < 0.05 for pHo 6.5 vs. 7.5; # P < 0.05 for pHo 8.5 vs. 7.5.

Did extracellular acidification alter rate of deactivation by destabilizing C-type inactivated state? The fourth possibility we considered was that the transition from the C-type inactivated state to the open state was the rate-limiting step in determining the apparent time course of deactivation. In this scheme, external H⁺ might accelerate the deactivation process by destabilizing the HERG channel in the C-type inactivated state and expediting its transition into the open state. If this were the case, a mutation that disrupted the C-type inactivation process might eliminate the effects of changing pH_o on HERG deactivation.

This was tested by making a double mutation in the outer-mouth region of the pore (G628C/S631C; Fig. 5) (22). Figure 8A shows that test-pulse currents through this mutant channel continued to increase up to +60 mV, reflecting a lack of C-type inactivation. Furthermore, at 80 mV the tail current of G628C/S631C was in the inward direction. This was due to an increase in the Na permeability (P_K/P_Na from 406 ± 58 in WT to 10.9 ± 0.9 in G628C/S631C), shifting the reversal potential from 98.2 ± 0.7 to 58.1 ± 2.1 mV. The simultaneous changes in C-type inactivation and in ion selectivity are consistent with the notion that channel domains that mediate C-type inactivation are close to, or coincide with, the ion selectivity filter (11).

View larger version (31K): [in this window] [in a new window]   Fig. 8.   An outer-mouth mutation that disrupted C-type inactivation and K selectivity of the pore did not prevent effects of changing pHo on rate of deactivation. Both G628 and S631 were replaced by cysteine (G628C/S631C). A: original current traces recorded from one oocyte at different levels of pHo. Voltage-clamp protocol was same as that described for Fig. 1. Bottom inset in A: superimposed tail currents shown at expanded time and current scales (after Vt to +20 mV). B: summary data of kinetics of HERG G628C/S631C deactivation at different levels of pHo. Data analysis and figure format were same as those described for Fig. 2B. Decay phase of tail currents was fit with a double exponential function to estimate fast, slow, and fast component of deactivation. These data are shown as filled (pHo 6.5) and shaded (pHo 8.5) symbols connected by solid lines. Open symbols are data corrected for shift in V0.5 of activation associated with changing pHo: +6.6 mV for pHo 6.5 and 1.3 mV for pHo 8.5. * P < 0.5 for pHo 6.5 vs. 7.5; # P < 0.05 for pHo 8.5 vs. 7.5.

Figure 8A also shows that in the absence of C-type inactivation, changing pH_o from 7.5 to 6.5 still accelerated the deactivation rate. Changing pH_o to 8.5 caused an opposite effect. Figure 8B summarizes the results. The change in pH_o from 7.5 to 6.5 shifted V_0.5 by 6.6 ± 0.9 mV, whereas changing pH_o to 8.5 shifted V_0.5 by 1.3 ± 0.2 mV. Figure 8B also shows that the changes in the kinetics of deactivation could not be accounted for by the voltage shift associated with alterations in pH_o. Therefore, the effect of external acidification was independent of the C-type inactivation process. Data from the G628C/S631C mutant further show that these acidification-induced changes were not affected by alterations in the outer-mouth conformation, in the ion selectivity, or in the direction of ion flux through the pore when the deactivation rate was measured. These observations suggest that the outer mouth probably was not where external H⁺ bound and affected the deactivation rate.

Did extracellular acidification alter rate of deactivation as a result of an interaction between an external H⁺ binding site and cytoplasmic domains that control kinetics of HERG deactivation? HERG channel manifests an unusually slow rate of deactivation among voltage-gated K channels, and it has been shown that the cytoplasmic NH₂-terminal domain plays a key role in determining the deactivation kinetics of this channel (23, 26). The proposed mechanism is that the NH₂-terminal domain can bind to the cytoplasmic S4-S5 loop and stabilize the channel in the open state (26). The fifth possibility we considered was that there was an allosteric interaction between an extracellular H⁺ binding site and the intracellular domains that control the deactivation rate so that the manifestation of effects of external acidification required the presence of an intact NH₂-terminal domain. We explored this possibility by testing the effects of changing pH_o on an NH₂-terminal deletion mutant (HERG N, amino acids 2-354 deleted) (23).

Figure 9A shows original current traces of the HERG N mutant recorded at various pH_o levels. The deactivation rate was greatly accelerated relative to that of the WT channel. Selected tail-current traces of HERG N are shown at an expanded time scale in Fig. 9B. The decay phase of the tail currents could be well described by a single exponential function. At 80 mV the  of deactivation of HERG N was only 9.6 ± 0.6 ms, much shorter than the _fast of deactivation of the WT channel (158 ± 10 ms, Fig. 2B). With such a dramatic increase in the rate of deactivation, changing pH_o from 7.5 to 6.5 could still significantly shorten the  of HERG N deactivation (Fig. 10C). This change in  of deactivation could not be accounted for by the positive shift in V_0.5 of activation accompanying the same pH_o change (9.6 ± 1.3 mV). On the other hand, although changing pH_o to 8.5 appeared to induce a slowing of HERG N deactivation, the difference disappeared after the data were corrected for the negative shift in V_0.5 of activation (9.0 ± 3.2 mV).

View larger version (33K): [in this window] [in a new window]   Fig. 9.   Accelerating rate of HERG deactivation by removing most of cytoplasmic NH2-terminal domain did not prevent effects of changing pHo on rate of deactivation. A: current traces recorded from HERG N when pHo was switched from 7.5 to 6.5 (top) or from 7.5 to 8.5 (bottom). Voltage-clamp protocol was same as that described for Fig. 1 except that sampling time for tail currents was shortened from 1 to 0.1 ms to better resolve time courses of fast deactivation. B: tail currents of HERG N at specified pHo shown at an expanded time scale and superimposed on single exponential time course. Voltage-clamp protocol is shown in inset. Best-fit values of  are marked. For clarity, only every fifth data point is shown. C: summary data of  of deactivation at different levels of pHo. Data points at pHo 6.5 and 8.5 are shown as observed values (solid and shaded symbols connected by solid lines) or after correction for shift in V0.5 of activation (open symbols connected by dotted lines). Changing pHo from 7.5 to 6.5 shifted V0.5 by +9.6 mV and changing pHo to 8.5 shifted V0.5 by 9.0 mV. * P < 0.05 for pHo 6.5 vs. 7.5; # P < 0.05 for pHo 8.5 vs. 7.5.

Side-chain modification by extracellular DEPC accelerated HERG deactivation and prevented further acceleration by external acidification. The above data ruled out five possible mechanisms by which extracellular H⁺ could modify the gating kinetics of the HERG channel. We then used DEPC to test the idea that properties of amino acid side chains on the HERG channel exposed to the extracellular solution could affect channel function and, in particular, the rate of deactivation. DEPC is a hydrophilic reagent that can covalently modify amino acid side chains (16). Although its primary target is histidine, at high concentrations DEPC can also modify the side chains of cysteine, tyrosine, serine, lysine, arginine, and tryptophan (16). In these experiments, a high concentration of DEPC (2 mM) was used. The DEPC solution was freshly made after control data were obtained and was applied to the cell immediately. We waited for the DEPC-induced changes in HERG-channel function to reach a steady state and washed out this reagent for >10 min. The effects of DEPC on HERG were not reversible after this prolonged wash period, indicating that they resulted from a covalent modification of side chains on the channel. Data are summarized in Fig. 10. DEPC induced a reduction of HERG-current amplitude (more prominent at 80 mV, Fig. 10A, than at depolarized voltages, Fig. 10B) and a small degree of positive shift in the voltage dependence of activation (Fig. 10C, V_0.5 = 11.8 ± 3.7 mV in control and 7.1 ± 1.4 mV after DEPC treatment). By far, the most dramatic effect of DEPC treatment was an acceleration of HERG deactivation (Fig. 10D): both _fast and _slow of deactivation were shortened 3- to 10-fold in the voltage range from 50 to 120 mV. The fast component of deactivation was increased at voltages positive to 80 mV.

View larger version (30K): [in this window] [in a new window]   Fig. 10.   Effects of diethylpyrocarbonate (DEPC, 2 mM) on HERG WT. A: superimposed current traces recorded under control conditions, at steady state of DEPC effects (DEPC), and after washout of DEPC (Wash). Voltage-clamp protocol is shown at top. Dotted trace is "wash" tail current scaled so that its peak amplitude matches that of control tail current. B: I-V relationships of normalized It before and after DEPC treatment. Voltage-clamp protocol and data analysis were same as those described for Figs. 1 and 4A. It was normalized to control It recorded at 10 mV. C: effects of DEPC on voltage dependence of activation. Voltage-clamp protocol and data analysis were same as those described for Fig. 4B. Superimposed curves were calculated from equation in Fig. 4 with following parameter values: control, V0.5 = 11.8 mV, k = 9.1 mV; after DEPC, V0.5 = 7.1 mV, k = 8.7 mV. D: effects of DEPC on kinetics of deactivation. Data analysis and figure format were same as those described for Fig. 2B. In B to D, control data are shown as open circles and data after DEPC treatment are shown as solid squares (n = 4 and 6 respectively). Statistical analysis was carried out using unpaired t-test (+P < 0.05). In D, open squares are data after DEPC treatment that were corrected for shift in V0.5.

DEPC treatment also modified the response of HERG to extracellular acidification. Figure 11 summarizes the effects of lowering pH_o from 7.5 to 6.5 on HERG in oocytes that had been treated with 2 mM DEPC. Under these conditions, extracellular acidification induced a prominent decrease in the current amplitude (Fig. 11A). The test-pulse current measured after 1 s at 0 mV was reduced by 65 ± 6% (n = 6). Furthermore, Fig. 11B shows that lowering pH_o caused a marked positive shift in the activation curve (V_0.5 shifted from 7.1 ± 1.4 to +11.1 ± 2.9 mV) along with a decrease in its steepness (k increased from 8.7 ± 0.3 to 12.2 ± 0.5 mV). Although in the DEPC-treated oocytes there was an apparent acceleration of deactivation at the acidic pH_o, this change could be entirely accounted for by the positive shift (by 18.3 ± 1.9 mV) in the voltage dependence of activation (Fig. 11C). Therefore, extracellular DEPC-mediated covalent modification of side chains on the HERG channel accelerated channel deactivation and prevented further acceleration by extracellular acidification.

View larger version (27K): [in this window] [in a new window]   Fig. 11.   Effects of external acidification on DEPC (2 mM)-modified HERG WT. A: original current traces recorded from a DEPC-treated oocyte at pHo 7.5 and then 6.5. Voltage-clamp protocol was same as that described for Fig. 2A. B: activation curves of WT HERG from DEPC-treated oocytes at pHo 7.5 and 6.5. Voltage-clamp protocol and data analysis were same as those described for Fig. 4B. Superimposed curves were calculated from equation in Fig. 4 with following mean parameter values: pHo 7.5, V0.5 = 7.1 mV, k = 8.7 mV; pHo 6.5, V0.5 = +11.1 mV, k = 12.3 mV. C: deactivation kinetics of HERG WT at pHo 7.5 and 6.5 from DEPC-treated oocytes. Data analysis and figure format are same as those described for Fig. 2B. * P < 0.05 for pHo 6.5 vs. 7.5.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Major findings of this study. We observed a marked acceleration of HERG deactivation associated with extracellular acidification in the pH_o range of 8.5-6.5. This occurred with little or no changes in other gating kinetics. Our experiments further ruled out the following possible mechanisms for pH_o-induced changes in the deactivation kinetics: 1) extracellular H⁺ neutralized uniform negative surface charges and caused a general positive shift in the external surface potential; 2) extracellular H⁺ blocked the pore in a voltage- and time-dependent fashion and caused an apparent acceleration of deactivation; 3) extracellular H⁺ protonated the two histidine residues on the extracellular domain of HERG and accelerated deactivation by an unidentified mechanism; 4) extracellular H⁺ expedited channel transition from the C-type inactivated state to the open state, and this was the rate-limiting step in determining the rate of deactivation; and 5) extracellular H⁺ bound to a site that could interact with the cytoplasmic NH₂-terminal domain and affect the rate of deactivation. Our data from an outer-mouth mutation (G628C/S631C) suggest that the outer-mouth region was not the site where external H⁺ bound and affected deactivation (although H⁺ could bind to the outer mouth of the pore and affect the stability of the C-type inactivated state). Finally, extracellular application of a nonspecific amino acid modifying reagent, DEPC, induced a marked and irreversible acceleration of HERG deactivation. After DEPC modification, extracellular acidification could not induce an acceleration of the deactivation process other than that associated with the acidification-induced positive shift in channel activation.

Could the observed acceleration of HERG deactivation result from intracellular changes secondary to alterations in pH_o? Channel deactivation has been viewed as a "cytoplasmic" event. For example, studies on the Shaker, Kv2.1, and Kv1.4 channels have suggested that the cytoplasmic halves of S5 and S6 are important parts of the activation gates in voltage-gated K channels (8, 21, 33). These domains surround the cytoplasmic entrance to the pore. It has been suggested that changes in the membrane voltage initiate movements in the S4 domain (12, 29). These lead to conformational changes in the cytoplasmic halves of S5 and S6 and the opening (activation) or closing (deactivation) of the cytoplasmic entrance to the pore (activation gate). Furthermore, the cytoplasmic loop between S4 and S5 may serve a linker function: transmitting the signals of S4 movements to the activation gate (15). In the HERG channel, cytoplasmic domains seem to play an especially important role in the channel's slow deactivation process. For example, it has been proposed that the cytoplasmic NH₂-terminal domain can bind to the cytoplasmic S4-S5 loop and stabilize the HERG channel in the open state (26). Therefore, deletions in the NH₂-terminal domain of HERG can markedly accelerate the rate of deactivation (23, 26).

Therefore, we need to consider whether the changes in HERG deactivation observed in our experiments actually resulted from intracellular changes secondary to extracellular acidification. The following observations argue against this possibility. First, we used a membrane-impermeable pH buffer, HEPES, in our bath solutions. Therefore, extracellular acidification should not affect intracellular pH. Second, intracellular acidification slows the deactivation rate of HERG channels expressed in HEK-293 cells, an effect opposite to our finding here (31). Third, the fast onset and reversal of the acidification-induced changes in the HERG current and the fact that similar changes could be induced repetitively in the same oocyte further argue that these changes resulted from extracellular (pH_o) changes but not from secondary changes that occurred intracellularly. Fourth, extracellular application of membrane-impermeable DEPC induced a marked and irreversible acceleration of HERG deactivation. This was most likely due to a covalent modification of amino acid side chains exposed to the extracellular aqueous solution (16). Although we could not be sure of what residues were modified by DEPC, the data support the notion that there are residues accessible to the external aqueous phase (and thus to extracellular H⁺), whose side-chain properties can affect the rate of HERG deactivation.

Comparison with other studies. There are two other recent reports describing the effects of extracellular acidification on HERG function (10, 31). Although there are differences in the expression system (oocyte expression in this study and that of Ref. 10, HEK-293 cell expression in Ref. 31), and the recording conditions (two-microelectrode voltage clamp in oocytes and whole cell patch clamp with intracellular dialysis in HEK-293 cells), the findings are similar: the major effect of external acidification is an acceleration of HERG deactivation. Jo et al. (10) attributed this effect to a time- and voltage-dependent blockade of the HERG pore by external H⁺. Our data suggest that factor(s) other than pore blockade must contribute to the acceleration of HERG deactivation at acidic pH_o. In the study by Zhang et al. (31), the authors proposed that the NH₂-terminal domain plays a key role in the change in HERG kinetics. We used the same mutant (HERG N) in our experiments. Although in HERG N the degree of acceleration of deactivation associated with acidification of pH_o was reduced relative to that of the WT channel, the effect was not prevented by this mutation. This suggests that another factor(s) must be involved.

How did extracellular H⁺ or DEPC modify rate of HERG deactivation? Because our experiments showed that the only two histidine residues on the extracellular loop of HERG were not involved in pH_o modulation of deactivation, other amino acids that are accessible to the extracellular aqueous phase must be involved. Because there is no evidence that DEPC can modify carboxylate side chains, glutamate and aspartate could be excluded as candidates. It has been suggested that external H⁺ reduce the current amplitude through an inwardly rectifying K channel by protonating cysteine residues on the extracellular linker between the pore region and the second transmembrane domain (2). This is possible because the pKa value of cysteine side chain depends on the local environment, such as dielectric properties of the medium and nearby charges. DEPC could modify the cysteine side chain (16), and there are two cysteine residues in the extracellular loop between S1 and S2 of HERG. Whether one or both of these cysteines, or other residues that can be modified by DEPC, are involved in pH_o modulation of HERG deactivation can be tested by future mutagenesis experiments.

Once the residues involved in pH_o modulation of HERG deactivation are identified, we need to deal with the issue of why modification of these side chains can alter the deactivation rate of the channel. It is possible that these side chains can interact with the voltage sensor of the channel, the S4 domain, in a state-dependent manner so that their modification by external H⁺ can specifically affect the deactivation kinetics. Gilly and Armstrong (4) showed that extracellular Zn²⁺ could markedly delay the activation of Na channel in the squid axon with little effects on the rate of deactivation. They hypothesized that the channel's voltage sensor has negative charges that are exposed to the extracellular solution when the channel is in the resting state. Zn²⁺ binding to these negative charges stabilizes the channel in the resting state and slows the activation rate. Once the channel is activated, external Zn²⁺ cannot gain access to these negative charges and thus does not affect the deactivation rate. One can envision that negative charges (e.g., unprotonated cysteine side chains) on the external surface of the HERG channel can interact with the positive charges on the S4 domain when the channel is activated, and thus some of the S4 positive charges are exposed to the extracellular solution (1, 12, 29). These interactions stabilize the channel in the open state and slow the deactivation rate. Elevating [H⁺]_o can accelerate HERG deactivation by protonating these side chains and preventing salt-bridge formation in the open state.