The Extracellular K+ Concentration Dependence of Outward Currents through Kir2.1 Channels Is Regulated by Extracellular Na+ and Ca2+*

It has been known for more than three decades that outward Kir currents (IK1) increase with increasing extracellular K+ concentration ([K+]o). Although this increase in IK1 can have significant impacts under pathophysiological cardiac conditions, where [K+]o can be as high as 18 mm and thus predispose the heart to re-entrant ventricular arrhythmias, the underlying mechanism has remained unclear. Here, we show that the steep [K+]o dependence of Kir2.1-mediated outward IK1 was due to [K+]o-dependent inhibition of outward IK1 by extracellular Na+ and Ca2+. This could be accounted for by Na+/Ca2+ inhibition of IK1 through screening of local negative surface charges. Consistent with this, extracellular Na+ and Ca2+ reduced the outward single-channel current and did not increase open-state noise or decrease the mean open time. In addition, neutralizing negative surface charges with a carboxylate esterifying agent inhibited outward IK1 in a similar [K+]o-dependent manner as Na+/Ca2+. Site-directed mutagenesis studies identified Asp114 and Glu153 as the source of surface charges. Reducing K+ activation and surface electrostatic effects in an R148Y mutant mimicked the action of extracellular Na+ and Ca2+, suggesting that in addition to exerting a surface electrostatic effect, Na+ and Ca2+ might inhibit outward IK1 by inhibiting K+ activation. This study identified interactions of K+ with Na+ and Ca2+ that are important for the [K+]o dependence of Kir2.1-mediated outward IK1.

Inward rectifier K ϩ channels (Kir) 3 channels are important in maintaining stable resting membrane potentials, controlling excitability, and shaping the initial depolarization, as well as the final repolarization of action potentials in many cell types, including heart cells (1)(2)(3)(4)(5)(6). The physiological functions of Kir channels are closely related to their unique inward rectification mechanism, which allows inward currents to pass through the channel more easily than outward currents (6). Although small, the outward I K1 plays a crucial role in controlling membrane excitability and action potential duration. The gain or loss of outward I K1 in the heart can lead to re-entry or arrhythmia, respectively (7). In contrast to normal conditions, in which the extracellular K ϩ concentration ([K ϩ ] o ) is ϳ5 mM, under pathological cardiac conditions, such as ischemia, tachycardia, and fibrillation, [K ϩ ] o can be as high as 18 mM (8,9). Of particular relevance, extracellular K ϩ ions accumulate in intercellular clefts and t-tubules of cardiac myocytes, where Kir2.x channels are expressed (10). When [K ϩ ] o is increased, the outward I K1 increases, despite a reduction in the electrochemical gradient (1,11). Accordingly, increases in I K1 result in reduced excitability, slow conductance, and abbreviation of the refractory period, and thereby predispose the heart to re-entrant ventricular arrhythmias, the leading cause of death from coronary artery disease (12). The [K ϩ ] o dependence of the outward I K1 thus might be important for regulating the physiological and pathological functions of the heart. Despite the importance of this long known [K ϩ ] o dependence of the outward I K1 , the underlying mechanism has remained unclear.
Previous studies have shown that the dependence of Kir channel activity on membrane potential (V m ) shifts in parallel with V m -E K (1)(2)(3)(4)(5)(6), where E K is the equilibrium potential for K ϩ . Thispropertyhasbeenpreviouslyattributedtothedrivingforcedependent block of Kir2.1 channels by intracellular blockers (13,14). According to this hypothesis, known as the blockingparticle model, cytoplasmic blockers are dragged into or pushed out of the channel pore by the outward or inward flux of K ϩ , respectively. However, the results of experiments performed at various [K ϩ ] o and intracellular K ϩ concentrations indicate that Kir channels detect a combination of V m and [K ϩ ] o rather than just V m -E K (2)(3)(4)(5)(6)15). Moreover, the Kir channel cannot conduct outward current in the absence of extracellular K ϩ (16). To account for the regulation of Kir activity by [K ϩ ] o , an alternative K ϩ -activated K ϩ channel model has been proposed (3,13,17). Thus, although previous studies have provided insights into the regulation of Kir channel functions by [K ϩ ] o , exactly how outward I K1 is increased by increases in [K ϩ ] o has remained unclear.
In this study, we examined the [K ϩ ] o dependence of the Kir2.1-mediated outward I K1 . Our results show that outward I K1 is steeply dependent on [K ϩ ] o in the pathophysiological range because of the [K ϩ ] o -dependent inhibition of outward I K1 by extracellular Na ϩ and Ca 2ϩ , which act by screening surface charges and, possibly, also by inhibiting the K ϩ activation mechanism. Our study has identified the roles of extracellular Na ϩ and Ca 2ϩ as well as surface charges in the mechanism underlying the regulation of Kir-mediated outward I K1 by

EXPERIMENTAL PROCEDURES
Preparation of Xenopus Oocytes-Xenopus oocytes were isolated by partial ovariectomy from frogs anesthetized with 0.1% tricaine (3-aminobenzoic acid ethyl ester). After suturing the surgical incision, animals were monitored during the recovery period before being returned to their tank. All surgical and anesthesia procedures were reviewed and approved by the Academia Sinica Institutional Animal Care and Utilization Committee.
Molecular Biology-cRNAs for expression in Xenopus oocytes were obtained by in vitro transcription (mMessage mMachine, Ambion, Dallas, TX). Site-directed mutations were generated using the PCR. Correct mutations were confirmed by sequencing cDNAs using the ABI Prism TM dRhodamine Terminator Cycle Sequencing Ready Reaction Kit (PE Applied Biosystems, Foster City, CA).
Electrophysiological Recordings-Currents were recorded at room temperature (21-24°C) using the patch clamp technique (18,19) (Fig. 1), 280 mM sucrose was added to solutions containing [K ϩ ] o Յ 20 mM to maintain osmolarity. Single-channel recordings were carried out in inside-out patches at a sampling rate of 10 kHz and a filtering rate of 2 kHz. The [K ϩ ] of the intracellular solution was 150 mM, and the [K ϩ ] of the pipette solution was 5 or 20 mM (Table 1). For pretreatment of oocytes with trimethyloxonium (TMO), de-vitallined oocytes were placed in a solution containing 90 mM KCl and 100 mM HEPES (pH 8.0) (21). TMO was then added to the solution to yield a final TMO concentration of 50 mM, and the solution was gently mixed by pipetting to allow efficient reaction of TMO with the surface of oocytes. TMO-treated oocytes were washed before transferring to the recording chamber. Command voltage pulses were controlled and data were acquired using pClamp10 software (Molecular Devices).
Data Analysis-Due to channel rundown in single-channel patches, the numbers of open and closed events were not sufficiently large to allow construction of open and closed histograms from each individual patch. Therefore, all single-channel events at each V m were pooled together on square root-log coordinates to construct single open and closed time distributions, which were best-fitted with two and three exponential components, respectively, using the maximum-likelihood method (22). The mean open and closed times obtained from this method have been shown to be similar to those determined by averaging the mean open and mean closed times of each patch (23). Opening events were sufficiently long and well resolved (mean open time Ͼ5 ms) that data correction for missed events was not required (dead time ϭ 360 s). Singlechannel current variances () in open and closed states were calculated using Clampfit basic statistics. Maximum outward I K1 was defined as the peak outward I K1 . Averaged data were presented as mean Ϯ S.E. Student's t tests for independent sam- ples were used to assess the statistical significance of differences between groups. A p value Յ 0.05 was considered statistically significant. Fig. 1A shows I-V m relationships recorded from one outside-out patch expressing Kir2.1 channels with K ϩ (1-150 mM) as the only extracellular cation. I-V m curves were almost linear at V m negative to E K . In contrast, at V m positive to E K , I K1 reached a maximum value, which we refer to as peak outward I K1 . Consistent with previous findings (6) Fig. 1D shows that the effects of [K ϩ ] o on the peak outward I K1 ratio and normalized I K1 at a fixed V m Ϫ E K were not significantly different (p Ͼ 0.05). Thus, in subsequent experiments, we focused on the effects of [K ϩ ] o on the peak outward I K1 ratio.

The Steep Dependence of I K1 on [K ϩ ] o Requires the Presence of Extracellular Na ϩ and Ca 2ϩ -
The results shown in Fig. 1A were obtained from experiments performed by varying [K ϩ ] o without adjusting osmolarity. To investigate whether an osmolarity imbalance contributed to the [K ϩ ] o dependence of the peak outward I K1 ratio, we performed experiments in which 280 mM sucrose was added to solutions containing [K ϩ ] o Յ 20 mM. Fig. 1D shows that the [K ϩ ] o dependence of the peak outward I K1 ratio was the same in the presence and absence of sucrose, indicating that an osmolarity imbalance does not account for the [K ϩ ] o dependence of the peak outward I K1 ratio. Therefore, in subsequent experiments osmolarity was not adjusted in solutions containing [K ϩ ] o Յ 20 mM. In summary, Fig. 1D shows that the [K ϩ ] o dependence of the peak outward I K1 ratio was shifted to the right by extracellular Na ϩ , either alone or in combination with Ca 2ϩ . As a result, the amplitude of the outward I K1 markedly depends on [K ϩ ] o in the pathophysiological range (3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18) Table 1 Is Independent of Intracellular Block-Next, we investigated the mechanisms underlying the inhibition of outward I K1 by extracellular Na ϩ and Ca 2ϩ . It has been proposed that extracellular K ϩ may bind to a high affinity site in the Kir2.1 channel pore and electrostatically weaken the pore blockade by intracellular Mg 2ϩ or polyamines such that inward rectification is weaker at higher [K ϩ ] o (13,14). The experiments shown in Fig. 1 were conducted in outside-out patches where perfusion was limited so that intracellular blockers could not be completely removed. Therefore, it is possible that extracellular Na ϩ and Ca 2ϩ might inhibit or weaken the K ϩ -blocker interaction, thereby increasing the apparent block by intracellular blockers and shifting the [K ϩ ] o dependence of the peak outward I K1 ratio to the right. If such K ϩ -blocker interactions indeed determine the outward I K1 amplitude, the [K ϩ ] o dependence of the peak outward I K1 ratio should also be shifted to the right when the intracellular blockage is stronger. In this study, we found that the degrees of inward rectification (1 Ϫ (peak outward I K1 /I K1 recorded at Ϫ40 mV)) varied in outside-out patches exposed to 150 mM [K ϩ ] o , indicating various degrees of washout of endogenous intracellular blockers (see "Discussion"). To examine whether blockade by intracellular blockers plays a role in the inhibition of I K1 by extracellular Na ϩ and Ca 2ϩ , we divided the data into two groups based on the degree of inward rectification at 150 mM [K ϩ ] o . Fig. 2A shows I-V m relationships recorded at the indicated [K ϩ ] o from two different outside-out patches; in one, the degree of inward rectification was 0.88 (upper panel), and in the other, the degree of rectification was 0.78 (lower panel). Fig. 2C shows that the averaged [K ϩ ] o dependence of the peak outward I K1 ratio in the presence of 140 mM [Na ϩ ] o was not significantly different between the two groups with different degrees of inward rectification, indicating that the inhibition caused by extracellular Na ϩ is not related to a K ϩ -blocker interaction mechanism. To further test this idea, we performed experiments using a D172N mutant with reduced sensitivity to intracellular block (24,25).  lower panel). Fig. 2C shows that the [K ϩ ] o /peak outward I K1 ratio relationships were similar (p Ͼ 0.05) in the wild-type Kir2.1 channel and D172N mutant with and without spermine included in the intracellular solution. These results show that the reduction of I K1 by extracellular Na ϩ is not related to inhibition of K ϩ -blocker interactions in the pore.
Pore  (32). Because the open channel probability of outward i K1 through the D172N mutant was higher than through the wild-type Kir2.1 channel (33), and because the effects of extracellular Na ϩ and Ca 2ϩ on the wild-type and D172N mutant were the same (Fig. 2), we examined the effects of extracellular Na ϩ /Ca 2ϩ on i K1 in the D172N mutant. Fig. 3A shows outward i K1 recorded from two different inside-out patches (used to allow efficient washout of intracellular blockers), one at V m ϭ ϩ30 mV (V m Ϫ E K ϭ ϩ70 mV) and   (Fig. 3C, panel a). Furthermore, extracellular Na ϩ and Ca 2ϩ did not decrease the open probability (Fig. 3C, panel  b) or affect the mean open time at V m ϭ 0 or ϩ20 mV, but did increase the mean open time at V m ϭ ϩ30 and ϩ40 mV (Fig.  3C, panel c). The closed time was decreased at 0 mV by extracellular Na ϩ and Ca 2ϩ , but was increased at ϩ30 and ϩ40 mV (Fig. 3C, panel d). Taken together, the results suggest that even if extracellular Na ϩ and Ca 2ϩ block the channels, the effect is overshadowed by a surface electrostatic mechanism as discussed below.
Extracellular Na ϩ and Ca 2ϩ Inhibit Outward I K1 in Part via a Surface Electrostatic Effect-In addition to blocking channels, ions such as H ϩ , Na ϩ , and Ca 2ϩ have been shown to exert a surface electrostatic effect on various ion channels (34,35). We next examined whether extracellular Na ϩ and Ca 2ϩ regulate the outward I K1 through such a surface electrostatic effect. Demonstrating that I K1 inhibition is dependent on the valence of the screening ion, but not on the particular ionic species, is a good test for the existence of a surface electrostatic effect (35). Fig. 4 shows the effects of channel-impermeant monovalent and divalent cations on the normalized peak outward I K1 (peak outward I K1 recorded in the test cation divided by that of control I K1 ) at 5 mM [K ϩ ] o . Monovalent N-methyl-D-glucamine (NMG ϩ ), Li ϩ , and Na ϩ had similar effects on the peak outward I K1 . Ca 2ϩ and Sr 2ϩ strongly decreased the peak outward I K1 in the same concentration-dependent manner, reducing it by about 75% at saturating concentrations. The observations that divalent cations were much more potent than monovalent cations, and that different sized cations with the same valence had the same effect are consistent with a surface-electrostatic mechanism, although block at a superficial site without access restrictions from the extracellular side cannot be ruled out.  Table 1. Peak outward I K1 recorded from outside-out patches were normalized to that obtained in the absence of Na ϩ , Li ϩ , NMG ϩ , Sr 2ϩ , and Ca 2ϩ . To further investigate the idea that surface potential is responsible for the [K ϩ ] o dependence of the outward I K1 , we examined the effect of extracellular TMO, a carboxylate esterifying agent that neutralizes negative surface charges (35,36). The I-V m relationship recorded with K ϩ as the only extracellular cation in an outside-out patch from an oocyte pretreated with 50 mM TMO is presented in Fig. 5A. Outward currents in TMO-treated oocytes were more sensitive to changes in [K ϩ ] o than were those recorded from oocytes not treated with TMO (Fig. 5C). When [K ϩ ] o was decreased from 150 to 1 mM in patches from TMO-treated oocytes, outward I K1 decreased by 57% compared with a 27% reduction without TMO treatment. These results support the conclusion that screening of negative surface charges can indeed decrease outward I K1 , and indicate that extracellular Na ϩ and Ca 2ϩ may inhibit outward I K1 by a surface-charge effect. In the presence of physiological [Na ϩ ] o and [Ca 2ϩ ] o , outward I K1 in TMO-treated oocytes was further reduced (Fig. 5, B and C). In fact, [K ϩ ] o dependence of the outward peak I K1 ratio in the presence of extracellular Na ϩ and Ca 2ϩ was the same with and without TMO treatment. The fact the [K ϩ ] o dependence of the peak outward I K1 ratio was further decreased by extracellular Na ϩ and Ca 2ϩ in the TMO-treated Kir2.1 channel indicates that extracellular Na ϩ /Ca 2ϩ have effects in addition to surface-charge screening effects.
Next, we sought to identify the source of the surface charges. There are five negatively charged residues (Asp 112 , Asp 114 , Glu 125 , Asp 152 , and Glu 153 ) located at the extracellular surface of the Kir2.1 channel (26,37,38). One of these residues, Glu 125 , has been shown to be involved in the block of inward I K1 by extracellular Ba 2ϩ and Mg 2ϩ , and in the permeation of K ϩ (26,39), probably via a surface charge-screening effect (26). Glu 153 has also been shown to be the primary source of surface charges for the inward K ϩ conductance of the Kir2.1 channel (38). To gain a better understanding of the physical location of these residues, we built a homology model of the Kir2.1 channel based on the crystal structure of the closely related Kir2.2 channel (40). The homology model (Fig. 6) suggests that Asp 114 , Glu 125 , and Glu 153 are most likely exposed to the extracellular milieu. In addition, the model predicts that Asp 112 is buried within each subunit and Glu 152 is buried at the interface of adjacent subunits. Therefore, we used site-directed mutagenesis to examine whether the Asp 114 , Glu 125 , and Glu 153 sites are involved in regulating the outward I K1 of the Kir2.1 channel. Fig. 5C shows that with K ϩ as the only extracellular cation, the [K ϩ ] o dependence of the peak outward I K1 ratio was not changed in D114N, E125N, or E153Q mutants. Because it is possible that more than one surface charge is involved in the surface electrostatic effects, we examined whether the D114N/ E153Q double mutation and/or D114N/E125N/E153Q triple mutation had a greater effect on outward I K1 . Fig. 7, A and B, show the I-V m relationships for Kir2.1 mutants D114N/E153Q and D114N/E125N/E153Q, respectively, recorded with K ϩ as the only extracellular cation. Outward I K1 in each of these mutants was more sensitive to changes in [K ϩ ] o than recorded from the wild-type (Fig. 1A). In fact, as shown in Fig. 7C, the [K ϩ ] o dependence of the peak outward I K1 ratio in the D114N/ E153Q and D114N/E125N/E153Q mutants was statistically indistinguishable from recordings of the TMO-treated wildtype Kir2.1 channel in the absence of extracellular Na ϩ and Ca 2ϩ . These results indicate that both Asp 114 and Glu 153 are involved in the surface electrostatic effects of extracellular Na ϩ and Ca 2ϩ . Fig. 7C also shows that the [K ϩ ] o /peak outward I K1 ratio relationships for the D114N/E153Q and D114N/E125N/ E153Q mutants were the same as for the wild-type Kir2.1 channel in the presence of extracellular Na ϩ and Ca 2ϩ . These results indicate that extracellular Na ϩ and Ca 2ϩ inhibit the Kir2.1 channel by an additional mechanism in which Asp 114 , Glu 125 , and Glu 153 are not involved.
K ϩ Activation Contributes to the [K ϩ ] o Dependence of Outward I K1 and Its Regulation by Extracellular Na ϩ and Ca 2ϩ -Results presented in Figs. 5 and 7 show that although a surface electrostatic mechanism is engaged, additional mechanisms are involved in the action of extracellular Na ϩ and Ca 2ϩ . Fig. 1D shows that increases in [K ϩ ] o can relieve the inhibition of outward I K1 by extracellular Na ϩ and Ca 2ϩ , suggesting that Na ϩ and Ca 2ϩ may compete with K ϩ for the same binding site(s) and generate opposite effects. It is possible that Na ϩ and Ca 2ϩ compete with K ϩ for a channel-activation site. Previously, it has been shown that [K ϩ ] o -dependent increases in inward I K1 were reduced in R148Y mutants compared with those in wildtype Kir2.1 channels (3). In addition, we have shown that neutralization of Arg 148 reduced the increases in the inward singlechannel conductance of Kir2.1 channels induced by elevated (41). These results indicate that a K ϩ activation mecha- nism exists in the Kir2.1 channel and that residue Arg 148 may be involved in the mechanism (3). Thus, to examine whether Na ϩ and Ca 2ϩ compete with K ϩ for an activation site, we examined the [K ϩ ] o /peak outward I K1 ratio relationships in the R148Y mutant. Fig. 8A shows that increases in [K ϩ ] o led to decreases in outward I K1 in the R148Y mutant; this was in contrast to the wild-type channel, where increases in [K ϩ ] o increased outward I K1 . This observation suggests that K ϩ -mediated channel activation exists in the Kir2.1 channel and that this mechanism is greatly reduced in the R148Y mutant. Therefore, when [K ϩ ] o is increased, the surface charge screening effect of extracellular K ϩ is manifested and outward I K1 decreases. Increased [K ϩ ] o may produce a great surface charge screening effect and thus substantially reduce outward I K1 . If this argument is correct, removing the surface charge in the R148Y mutant with TMO should reduce the inhibitory effect of high [K ϩ ] o on outward I K1 . Indeed, as shown in Fig. 8B, [K ϩ ] o dependence of outward I K1 was much reduced in the R148Y mutant pretreated with TMO, supporting our view that the surface charge effect is dominant in the R148Y mutant. Fig. 8C shows the [K ϩ ] o dependence of peak outward I K1 in the presence of Na ϩ and Ca 2ϩ in the R148Y mutant pretreated with TMO. The humpshaped outward I K1 that occurred near E K was completely abolished, and the outward I K1 was about the same when [K ϩ ] o was increased. In addition, the outward I K1 formed a plateau that extended over a broad V m range. The mechanism for this plateauing of outward I K1 is unknown, but it was also observed in the absence of extracellular Na ϩ and Ca 2ϩ . Therefore, the effect is likely related to the R148Y mutation rather than the presence of Na ϩ and Ca 2ϩ . Fig. 8E shows the averaged effects of [K ϩ ] o on the peak outward I K1 ratio in the R148Y mutant. Because a sharp peak outward I K1 was not obvious in the presence of Na ϩ and Ca 2ϩ , the peak outward I K1 ratio was calculated by normalizing the outward I K1 at V m Ϫ E K ϭ ϩ20 mV at different [K ϩ ] o to that at 150 mM [K ϩ ] o . Both the presence of Na ϩ and Ca 2ϩ and pretreatment with TMO greatly reduced the dependence of the peak outward I K1 ratio on [K ϩ ] o elevation.
It is noted that the outward I K1 -reducing effects of Na ϩ and Ca 2ϩ were slightly, but significantly, larger than that of TMO pretreatment in the R148Y mutant (Fig. 8E). Several previous studies have shown that extracellular Na ϩ could induce timedependent decay of inward currents through Kir channels of skeletal muscle (28), egg cells (29), and cardiac myocytes (27,30,31). The inhibitory effect of Na ϩ has been attributed to V m -dependent block of the Kir channel by extracellular Na ϩ (28 -31) or to stabilizing the channel at a closed state (27) (this may be relevant to K ϩ -activation mechanisms). Indeed, we found that, extracellular Na ϩ /Ca 2ϩ reduced the inward K ϩ conductance in both the wild-type and R148Y channel (Fig. 8, D and F). It is possible that extracellular Na ϩ /Ca 2ϩ may also inhibit outward K ϩ conductance by blocking the channel pore and the effect is preserved in the R148Y mutant.
In summary, these results show that Arg 148 is crucial for K ϩ activation of the Kir2.1 channel, and demonstrate that both K ϩ activation and surface charge effects regulate the amplitude of the outward I K1 . In addition, reducing K ϩ activation and surface electrostatic effects mimicked the action of extracellular Na ϩ and Ca 2ϩ , suggesting that Na ϩ and Ca 2ϩ may inhibit outward I K1 by affecting both K ϩ activation and surface electrostatic mechanisms.

DISCUSSION
The modulation of Kir channels can profoundly affect health, in some cases causing diseases, but it also provides a possible   . Effects of extracellular K ؉ , Na ؉ , and Ca 2؉ on the I-V m curves of R148Y mutants. A, I-V m curves for the R148Y mutant recorded using a ramp protocol (V m ϭ Ϫ150 to ϩ50 mV over 3 s) from an outside-out patch exposed to different [K ϩ ] o , with K ϩ as the only cation. B, I-V m curves for the R148Y mutant recorded from patches with TMO treatment in the absence of extracellular Na ϩ and Ca 2ϩ . C, I-V m curves for the R148Y mutant recorded from patches with TMO treatment in the presence of extracellular Na ϩ and Ca 2ϩ . D, I-V m curves for the R148Y mutant without TMO treatment in the presence of extracellular Na ϩ and Ca 2ϩ . E, [K ϩ ] o dependence of the peak outward I K1 ratio recorded from the R148Y mutant under the indicated conditions. F, [K ϩ ] o dependence of normalized inward K ϩ conductance (normalized G) in the wild-type and R148Y mutant. Normalized G was obtained by dividing inward K ϩ conductance at various [K ϩ ] o to that obtained at 150 mM [K ϩ ] o in the absence of extracellular Na ϩ /Ca 2ϩ . Inward K ϩ conductance was measured as the slope of I-V m relationship at V m Ϫ E K ϭ Ϫ10 to Ϫ35 mV. *, p Ͻ 0.05; **, p Ͻ 0.005; and ***, p Ͻ 0.001 for comparisons of data obtained with TMO pretreatment in the absence (OE) and presence (‚) of Na ϩ and Ca 2ϩ ; n ϭ 2-7.
Regulation of Outward I K1 by Extracellular Na ϩ and Ca 2ϩ -We present several lines of evidence supporting the conclusion that Na ϩ and Ca 2ϩ inhibit outward I K1 by a surface charge screening effect. First, extracellular Na ϩ and Ca 2ϩ reduced the outward single-channel current, an effect that was greater at lower ionic strength. Second, extracellular Na ϩ and Ca 2ϩ did not decrease the mean open time. Third, TMO modification and neutralizing both Asp 114 and Glu 153 by site-directed mutagenesis decreased outward I K1 in the absence of extracellular Na ϩ and Ca 2ϩ .
In addition to a surface-electrostatic effect, extracellular Na ϩ and Ca 2ϩ also inhibit the outward I K1 through at least another mechanism (Fig. 5). Because the inhibitory effect of reducing both K ϩ activation and surface charges (R148Y mutant pretreated with TMO) on the outward I K1 was similar to the effects of extracellular Na ϩ and Ca 2ϩ , we propose that extracellular Na ϩ and Ca 2ϩ may also reduce outward I K1 by inhibiting the K ϩ activation mechanism. Finally, we cannot rule out that extracellular Na ϩ /Ca 2ϩ might also inhibit outward K ϩ conductance by blocking the channel pore. The multiple mechanisms involved in extracellular Na ϩ and Ca 2ϩ inhibition may explain why V m dependence of the mean open and mean closed times of the i K1 are complicated.
It is noted that various degrees of inward rectification were recorded in outside-out patches, indicating that washout of endogenous intracellular blockers was variable even though the pipette solution did not contain Mg 2ϩ or polyamines. However, it should be emphasized that the incomplete washout of endogenous blockers would not affect our interpretations of the inhibitory effects of extracellular Na ϩ and Ca 2ϩ on outward I K1 . First, we found that varying degrees of intracellular block and addition of 100 M spermine did not affect the inhibition of outward I K1 by extracellular Na ϩ . Second, we showed that extracellular Na ϩ and Ca 2ϩ reduced single-channel conductance and slightly increased open probability. In contrast, it has been shown that increases in the intracellular spermine concentration decrease the open probability of outward singlechannel currents through the Kir2.1 channel but do not change the single-channel conductance (33). The simplest interpretation of these results is that the regulation of [K ϩ ] o -dependent outward I K1 by extracellular Na ϩ and Ca 2ϩ is not related to intracellular block.
Regulation of Outward I K1 by Extracellular K ϩ -This study shows that extracellular K ϩ has two opposite effects on the outward I K1 through the Kir2.1 channel. First, increases of [K ϩ ] o increase outward K ϩ conductance (as shown in Fig. 1A). Direct activation of K ϩ channels by K ϩ has been proposed as an explanation for the increase in K ϩ channel activity caused by increased [K ϩ ] o in various types of K ϩ channels (3,3,13,17,43,44). Based on site-directed mutagenesis studies, it has been suggested that K ϩ occupancy at extracellular site(s) is involved in the K ϩ activation of inward rectifier K ϩ channels and V m -gated K ϩ channels (3,17,26,44). It is possible that extracellular K ϩ binds to a site (or sites) stabilized by salt bridges formed by residues Glu 138 and Arg 148 (45) and then activate channel activity in the Kir2.1 channel. Indeed, the recently resolved crystal structure of the Kir2.2 channel shows that such an ion pair "staples" the pore region together (40). The homology model of the Kir2.1 channel also revealed that Arg 148 and Glu 138 are located close to each other (Fig. 6B). Another hypothesis to account for the [K ϩ ] o dependence of Kir channel activation is that the block of outward I K1 by polyamines is relieved by increased [K ϩ ] o (14). It has been shown that changes in [K ϩ ] o shift the V m dependence of polyamine block (14). This finding explains why the voltages at which the peak outward I K1 occur depend on V m Ϫ E K when [K ϩ ] o is changed (6).
Second, increases of [K ϩ ] o decrease outward K ϩ conductance by a surface electrostatic mechanism. Because extracellular Na ϩ /Ca 2ϩ can screen extracellular surface charges and thus reduce outward I K1 , extracellular K ϩ should also have such an effect. Indeed, this effect was demonstrated by the steep reduction of outward I K1 when [K ϩ ] o were increased in the R148Y mutant (Fig. 7A). Usually, this mechanism is concealed by the K ϩ activation effect but is unraveled in the R148Y mutant where [K ϩ ] o dependence of Kir channel activation is reduced.
Conclusion-K ϩ activation has been described for several K ϩ channels, including Kir and V m -gated K ϩ channels, and is an important mechanism for regulating K ϩ channel function. Likewise, surface electrostatic effects modulate many properties of ion channels. However, few studies have addressed the relevance of these effects in terms of the pathophysiological functions of ion channels. In this study, we showed that the [K ϩ ] dependence of the outward I K1 is controlled by both K ϩ activation and surface electrostatic mechanisms. In addition, extracellular Na ϩ and Ca 2ϩ inhibit outward I K1 in a [K ϩ ] o -dependent fashion, at least in part by a surface electrostatic effect, resulting in a steep [K ϩ ] o dependence of the outward I K1 in the pathophysiological [K ϩ ] o range. Our work suggests that surface charge screeners by inhibiting the outward I K1 may be potential candidates for diseases related to elevated outward I K1 .