当前位置: 首页 > 期刊 > 《普通生理学杂志》 > 2005年第3期 > 正文
编号:11324906
Divalent Cation Sensitivity of BK Channel Activation Supports the Existence of Three Distinct Binding Sites
http://www.100md.com 《普通生理学杂志》
     1 Department of Anesthesiology, Washington University School of Medicine, St. Louis, Missouri 63110

    2 Department of Anatomy and Neurobiology, Washington University School of Medicine, St. Louis, Missouri 63110

    ABSTRACT

    Mutational analyses have suggested that BK channels are regulated by three distinct divalent cation-dependent regulatory mechanisms arising from the cytosolic COOH terminus of the pore-forming subunit. Two mechanisms account for physiological regulation of BK channels by μM Ca2+. The third may mediate physiological regulation by mM Mg2+. Mutation of five aspartate residues (5D5N) within the so-called Ca2+ bowl removes a portion of a higher affinity Ca2+ dependence, while mutation of D362A/D367A in the first RCK domain also removes some higher affinity Ca2+ dependence. Together, 5D5N and D362A/D367A remove all effects of Ca2+ up through 1 mM while E399A removes a portion of low affinity regulation by Ca2+/Mg2+. If each proposed regulatory effect involves a distinct divalent cation binding site, the divalent cation selectivity of the actual site that defines each mechanism might differ. By examination of the ability of various divalent cations to activate currents in constructs with mutationally altered regulatory mechanisms, here we show that each putative regulatory mechanism exhibits a unique sensitivity to divalent cations. Regulation mediated by the Ca2+ bowl can be activated by Ca2+ and Sr2+, while regulation defined by D362/D367 can be activated by Ca2+, Sr2+, and Cd2+. Mn2+, Co2+, and Ni2+ produce little observable effect through the high affinity regulatory mechanisms, while all six divalent cations enhance activation through the low affinity mechanism defined by residue E399. Furthermore, each type of mutation affects kinetic properties of BK channels in distinct ways. The Ca2+ bowl mainly accelerates activation of BK channels at low [Ca2+], while the D362/D367-related high affinity site influences both activation and deactivation over the range of 10–300 μM Ca2+. The major kinetic effect of the E399-related low affinity mechanism is to slow deactivation at mM Mg2+ or Ca2+. The results support the view that three distinct divalent-cation binding sites mediate regulation of BK channels.

    Key Words: BK channels Ca2+ regulatory sites divalent cations Slo activation

    Abbreviation used in this paper: RCK, regulatory of conductance for K+.

    INTRODUCTION

    The sites and mechanisms by which Ca2+ binding results in regulation of Ca2+-activated and voltage-dependent BK type K+ channels (Barrett et al., 1982; Moczydlowski and Latorre, 1983; Horrigan and Aldrich, 2002; Lingle, 2002; Magleby, 2003) remain incompletely understood. Available evidence suggests that the Ca2+ regulatory machinery is contained in the extensive cytosolic COOH terminus of each BK channel subunit (Schreiber and Salkoff, 1997; Schreiber et al., 1999; Bian et al., 2001; Bao et al., 2002; Shi et al., 2002; Xia et al., 2002, 2004), but that physiological regulation by Ca2+ is defined by multiple determinants in the COOH terminus. Specifically, mutational analyses indicate that at least three distinct Ca2+ regulatory mechanisms in each subunit may influence activation of BK channels, two of μM affinity mediating physiological regulation by Ca2+ (Schreiber and Salkoff, 1997; Bao et al., 2002; Xia et al., 2002) and one of lower affinity perhaps providing the basis for channel regulation by cytosolic Mg2+ (Shi et al., 2002; Xia et al., 2002). However, except perhaps for the so-called Ca2+ bowl residues (Schreiber and Salkoff, 1997; Bian et al., 2001; Bao et al., 2004), the extent to which these mutationally defined loci correspond to specific divalent cation binding sites remains unclear. In addition, the extent to which the three distinct regulatory effects are coupled or independent remains unknown.

    Explicit structural information about the BK channel COOH terminus remains unavailable. However, the recognition that domains within the BK channel COOH terminus exhibit homology to ligand-binding RCK (regulatory of conductance for K+) domains of bacterial K+ channels (Jiang et al., 2001, 2002) provides a substantive guide to evaluating the BK channel cytosolic structure. Specifically, the Ca2+-regulated MthK bacterial K+ channel exhibits an octameric arrangement of RCK domains, with each RCK domain proposed to contain a Ca2+-binding site (Jiang et al., 2002). Similarly, the COOH terminus of each BK channel subunit has also been proposed to contain a pair of RCK domains (Jiang et al., 2002; Roosild et al., 2004). Whereas the homology of the first RCK domain in the BK channel subunit to bacterial RCK domains is fairly well defined, the second BK RCK domain is less clear (Roosild et al., 2004). Yet, the model provided by the MthK channel raises the possibility that two ligand binding sites per BK subunit may be required to produce the conformational changes required for channel opening.

    The mutational analysis, in fact, can be viewed as supporting the idea that there are separate Ca2+ regulatory elements at least loosely associated with each of the RCK-containing motifs of each subunit. One of these elements is the Ca2+ bowl, an aspartate-rich region near the COOH terminus (Schreiber and Salkoff, 1997). The fact that Ca2+ bowl mutations remove only a portion of Slo1 Ca2+ dependence suggested that there must be additional Ca2+ regulatory sites. Subsequently, two separate loci in the first RCK domain were shown to disrupt a second component of Slo1 regulation by micromolar Ca2+ (Bao et al., 2002; Xia et al., 2002). Residues D362 and D367 were identified by a search for negatively charged residues in the first RCK domain that were shared among all Slo1 homologues, but not found in the Ca2+-independent, pH-regulated Slo3 channel (Xia et al., 2002). The D362A/D367A mutation, in conjunction with mutation of the Ca2+ bowl, results in complete abolition of Slo1 regulation up to 1 mM Ca2+. Another mutation, M513I, behaved in a fashion similar to the D362A/D367A mutation (Bao et al., 2002), both mutations removing a similar component of regulation by Ca2+. Thus, these results suggest that distinct Ca2+ binding sites, associated with two distinct modules in the COOH terminus of the BK channel subunit, may account for regulation of Slo1 current by micromolar Ca2+. Finally, a third set of residues, also in the first RCK domain, participates in regulation by mM Ca2+ or Mg2+ (Shi and Cui, 2001; Xia et al., 2002). While Ca2+ overlay assays have supported the idea that the Ca2+ bowl may bind divalent cations directly (Bian et al., 2001; Braun and Sy, 2001; Bao et al., 2004), at present there is no information about Ca2+ binding to the first RCK domain. Thus, in the absence of such information, it remains possible that the mutations in the first RCK domain do not represent a distinct binding site, but simply disruption of a component of Ca2+ regulation initiated by a single Ca2+ binding site elsewhere on the protein.

    We hypothesize that if each regulatory mechanism involves a distinct divalent cation binding site, each mechanism might exhibit differences in selectivity among different divalent cations. In fact, it has been previously observed that the functional disruption of the calcium bowl does not disrupt BK channel regulation produced by Cd2+, leading to the suggestion that the BK channel may contain additional Ca2+ regulatory sites (Schreiber and Salkoff, 1997). A number of other divalent cations, such as Mn2+, Ni2+, and Co2+, are also known to regulate BK activation (Oberhauser et al., 1988), but how their effects relate to known regulatory mechanisms is unknown. We have therefore evaluated the divalent cation sensitivity of each of the three mutationally defined processes contributing to regulation of BK channels. Furthermore, to assess whether the regulatory sites act independently or in concert to regulate BK gating, we also examined the kinetic consequences of mutation of each regulatory site.

    Our results show that each mutationally defined regulatory mechanism exhibits a sensitivity to a set of divalent cations distinct from each of the other two regulatory mechanisms. This argues strongly that each regulatory mechanism involves a specific and unique divalent cation binding site located on each subunit, although it does not establish that the mutated residues contribute to cation coordination. Furthermore, each regulatory site exerts distinct effects on channel kinetic behavior. The results support the idea that there are three distinct divalent cation binding sites that contribute in independent ways to regulation of BK channel activation.

    MATERIALS AND METHODS

    Generation and Expression of mSlo1 Mutants

    The wild-type construct for these studies was mouse Slo1 (mSlo1) (Butler et al., 1993), generously provided by L. Salkoff (Washington University, St. Louis, MO). Numbering used here for mSlo1 residues begins with the second potential initiation site. Construction of mutations was as previously described (Xia et al., 2002). Disruption of Ca2+ bowl function was accomplished by neutralization of five consecutive aspartic acid residues in that region (5D5N; mSlo1 residues 897 through 901). Mutations in the first RCK domain disrupted another high affinity Ca2+ regulatory effect (D362A/D367A), while a mutation in a second location in the first RCK domain (E399A) disrupted low affinity regulation by Mg2+ and Ca2+. We have also confirmed that mutation of E374A, another residue implicated in low affinity Mg2+ regulation (Shi et al., 2002), produces effects on regulation by Mg2+ identical to those of E399A (not depicted). Simultaneous mutation of both E374A and E399A produced no additional effect beyond that occurring with each mutation of each residue alone.

    Methods of expression in Xenopus oocytes were as previously described (Zeng et al., 2001, 2003). SP6 RNA polymerase was used to synthesize cRNA for oocyte injection after DNA was linearized with M1uI. 10–50 nl of cRNA (10–20 ng/μl) was injected into stage IV Xenopus oocytes harvested 1 d before. Oocytes were used 2–7 d after injection of cRNA.

    Electrophysiological Recordings

    Channel currents were recorded from inside-out patches (Hamill et al., 1981) using Clampex (Axon Instruments), as performed routinely in our laboratory (Lingle et al., 2001; Zeng et al., 2003). During electrophysiological recordings, oocytes were maintained in ND96 bath solution (96 mM NaCl, 2.0 mM KCl, 1.8 mM CaCl2, 1.0 mM MgCl2, 5.0 mM HEPES, pH 7.5). The pipette extracellular solution was 140 mM potassium methanesulfonate, 20 mM KOH, 10 mM HEPES, 2 mM MgCl2. Cd2+, Mn2+, Co2+, and Ni2+ solutions bathing the cytoplasmic face of the patch membrane contained 155 mM KF and 10 mM HEPES. Such solutions were typically prepared by adding aliquots of metal-chloride solutions to the KF solution. As noted previously (Oberhauser et al., 1988), because of the lack of solubility of CaF2, F– effectively removes free Ca2+ from such solutions such that the effective free Ca2+ concentrations can be considered <20 nM. In contrast, Cd2+, Mn2+, Co2+, and Ni2+ remain soluble. Since Sr2+ is insoluble in the presence of F–, Sr2+ solutions were made with 140 mM potassium methanesulfonate, 20 mM KOH, 10 mM HEPES, and buffered with 10 mM EGTA. Appropriate additions of Sr2+ were determined with EQCALWIN (Biosoft). Assuming a contaminant Ca2+ concentration of 15 μM, free Ca2+ in the 2, 5, 10, and 20 mM Sr2+ solutions should be 106, 54, 28, and 11 nM, respectively. None of the effects described here for mM Sr2+ are attributable to Ca2+ at these concentrations. The pH values of all these solutions were adjusted to 7.0 with methanesulfonic acid or KOH, as required. The methods of calibration and preparation of Ca2+ and Mg2+ solutions have been described (Zeng et al., 2001). Concentrations of Mn2+ >5 mM were not employed because of the formation of precipitates in the MnF solutions. Cd2+ concentrations of 1 mM and higher produced essentially irreversible effects on Slo1 currents, perhaps reflecting essentially irreversible binding of Cd2+ to particular cytosolic cysteine residues in the cytosolic face of the Slo1 subunit.

    Excised inside-out patches were locally perfused with solutions flowing from a blunt perfusion pipette with an 150–200-μm tip (Herrington et al., 1995; Solaro et al., 1997). The shank of the perfusion pipette was packed with seven PE10 tubes that were connected to reservoirs containing cytosolic solutions of specific divalent cation composition. Solution flow was permitted from only a single PE10 line at any time, and was activated either by a hand-controlled valve or a computer-controlled solenoid valve. Additional upstream valves allowed a given patch to be tested with up to 10 different solutions. A complete concentration change between a pair of solutions was generally complete at the cytoplasmic face of the patch within several hundred milliseconds.

    Current values were measured using ClampFit (Axon Instruments) and converted to conductance. Conductance–voltage (G–V) curves were generated from tail currents following standard procedures (Zeng et al., 2001). Families of G–V curves represent averages from a set of patches at specific divalent cation concentrations. For each patch, G–V curves were fit with the following:

    (1)

    to provide estimates of Vh, the voltage of half activation, and k, the slope factor describing the voltage dependence of the closed–open equilibrium. For comparison of the magnitude of Vh shifts, Vh values were obtained by subtracting the Vh in the absence of specific divalent ions. Activation and deactivation time constants were obtained by fitting the original current traces by exponential functions. In all cases, error bars show SEM.

    Experiments were done at room temperature (21–24°C). 1 N KOH solutions for preparation of physiological solutions was obtained from Fisher Scientific. KF, MnCl2, and NiCl2 were obtained from Alfa-Aesar. All other salts and chemicals were obtained from Sigma-Aldrich.

    RESULTS

    The results in this paper address two topics. First, the relative divalent cation selectivity of each of the mutationally defined divalent cation regulatory mechanisms is addressed. Second, the impact of each of the mutationally defined regulatory mechanisms on activation and deactivation time course is examined.

    Cd2+ Activates BK Channels by Acting on the D362/D367 Site but not the Ca2+ Bowl

    BK currents exhibit a characteristic shift in activation with elevations in cytosolic [Ca2+]i (Cox et al., 1997). Typically, the voltage of half activation (Vh) obtained from conductance–voltage (G–V) curves shifts to more negative potentials as Ca2+ is elevated (Xia et al., 2002). Here, we compare the Vh shifts caused by specific divalent cations for wild-type mSlo1 BK channels and mSlo1 mutants with selective disruption of each of three putative regulatory mechanisms. The ability of mutation of a specific regulatory mechanism to alter the Vh shift produced by a given divalent cation can be taken as evidence for interaction between this cation and the binding site that mediates the regulatory effect. It does not prove that the mutated residues participate in binding site.

    It has been previously reported that 60 μM Cd2+ activates BK currents, producing a negative Vh shift (Schreiber and Salkoff, 1997). This effect was not altered by mutation of residues in the Ca2+ bowl, leading to the proposal that there was a second divalent cation regulatory site (Schreiber and Salkoff, 1997). We therefore examined whether Cd2+ selectively activates BK channels through either of the other putative divalent cation regulatory sites.

    Representative currents recorded from wild-type mSlo1, D362A/D367A, and 5D5N over a range of [Cd2+] up to 300 μM are shown in Fig. 1 (A–C). At [Cd2+] >300 μM, both steady-state current and tail current amplitudes are largely blocked and the recovery is very slow. The blocking effect of Cd2+ on 5D5N seems weaker than that on mSlo1, although Cd2+ produces similar Vh shifts for both.

    For mSlo1 and 5D5N, as [Cd2+] is increased over the range of 0 through 300 μM, current activation is shifted to more negative activation potentials. In contrast, this is not the case for D362A/D367A. For each construct, G–V curves at different [Cd2+] were obtained from tail currents and fit with a Boltzmann function (Eq. 1) to obtain an estimate of Vh. Within each patch, conductance estimates were normalized to the maximal conductance observed at 30 μM Cd2+. The relative Vh shift (Vh), obtained by subtracting the Vh value at 0 Cd2+ from that at given [Cd2+], was used as a relative indication of the ability of Cd2+ to enhance activation.

    In wild-type mSlo1 channels, Cd2+ shifts G–V curves leftward (Fig. 1 D) and produces negative Vh shifts (Fig. 1 G). In contrast, the ability of Cd2+ to activate BK channels is largely abolished in D362A/D367A relative to mSlo1, especially when [Cd2+] is 100 μM or below (Fig. 1, E and G). On the other hand, the ability of [Cd2+] to cause current activation is well preserved in 5D5N (Fig. 1, F and H), consistent with previous observations (Schreiber and Salkoff, 1997). Up through 100 μM, Cd2+ shows little effect on the voltage dependence of channel gating. However, a more shallow G–V curve is observed at 300 μM Cd2+ for both mSlo1 (Fig. 1 D) and D362A/D367A (Fig. 1 E), probably arising in part from the voltage dependence of the mechanism of Cd2+ block.

    We next examined whether the E399 low affinity site might contribute to the additional Vh shift observed with 300 μM Cd2+ in D362A/D367A (Fig. 1 G). For the E399A channel, Cd2+ up through 30 μM had effects similar to its effects on wild-type mSlo1. However, Vh shifts at 100 and 300 μM [Cd2+] are largely reduced in the E399A channel compared with results obtained from mSlo1 (Fig. 1 I). Furthermore, for the double mutant, D362A/D367A+E399A, the ability of Cd2+ to shift current activation is essentially abolished over the entire test concentration range (Fig. 1 I). The apparently additive effects of D362A/D367A and E399A suggest that these two sites act independently in regards to the ability of Cd2+ to regulate BK channels.

    In conclusion, it is the regulatory mechanism defined by the D362/D367 site, and not that defined by the Ca2+ bowl, which mediates activation of BK channels by lower concentrations of Cd2+. At higher concentrations, Cd2+ also can enhance channel activation through the low affinity site defined by E399.

    Mn2+ Enhances Activation of BK Channels through the E399 Low Affinity Site

    Mn2+ has been shown to activate muscle membrane BK channels expressed in bilayers (Oberhauser et al., 1988). We therefore tested whether the ability of Mn2+ to modulate BK channels might be selectively associated with any of the three putative regulatory sites. Fig. 2 (A–C) shows currents activated in the presence of 0 through 5 mM Mn2+ for mSlo1, E399A, and the triple mutation (5D5N+D362A/D367A+E399A). For mSlo1, Mn2+ causes a leftward shift in the G–V curves with no change in voltage dependence (Fig. 2 D). The negative Vh shift produced by 5 mM Mn2+ relative to 0 divalent is >100 mV (Fig. 2 G), while 100 μM Mn2+ produces an 50-mV leftward shift. Mutation of the low affinity site alone is sufficient to abolish most of the effect of Mn2+ (Fig. 2, E and G). Moreover, additional mutation of the two high affinity sites together with E399A does not produce any additional effect other than that caused by E399A alone (Fig. 2, F and H). In fact, Mn2+ loses its ability to activate current only for those constructs in which the E399A mutation is present (Fig. 2 H), while Mn2+ produces similar current activation for all constructs containing an intact E399 site (Fig. 2 I). These results show clearly that Mn2+ modulates BK channels predominantly through the low affinity site. Furthermore, although we were only able to examine Mn2+ concentrations up through 5 mM, the E399A mutation appears to completely remove the approximately –90-mV shift produced by 5 mM Mn2+, whereas E399A removes about half of the approximately –60-mV shift caused by 10 mM Mg2+ (Xia et al., 2002).

    The cumulative data in Fig. 2 I suggest that mutation of either 5D5N or D362A/D367A alone may produce a small reduction in the ability of Mn2+ to shift activation at 100 μM Mn2+, raising the possibility of a small effect involving the higher affinity Ca2+ regulatory mechanisms. However, these small effects of 5D5N and D362A/D367A do not appear to be additive, and estimates of Vh for constructs with very positive Vh are subject to uncertainties (Fig. 2 I). Furthermore, simultaneous mutation of all three sites (the triple mutation: 5D5N/D362AD367A/E399A) shows a behavior similar to E399A alone (Fig. 2 H). Although we cannot completely exclude the possibility that Mn2+ may produce a small activation effect through one or both of the higher affinity sites, it is clear that in comparison to Ca2+ or Cd2+, the ability of Mn2+ to shift activation via these sites is either absent or markedly reduced.

    The E399 Low Affinity Site Is Partially Responsible for the Ability of Ni2+ and Co2+ To Activate BK Channels

    We next tested the selectivity of the three sites for Ni2+. In contrast to its effects on activation by Mn2+, the E399A mutation reduces, but does not fully abolish, the negative Vh shifts caused by Ni2+ on mSlo1, which is about –160 mV with 50 mM Ni2+ (Fig. 3 A). However, simultaneous mutation of the two high affinity sites has no effect on the ability of Ni2+ to produce Vh shifts (Fig. 3 B). The important role of E399 in the effects of Ni2+ is also supported by the fact that all constructs containing E399A exhibit Ni2+-induced shifts in Vh similar to those seen in E399A alone (Fig. 3 C). Furthermore, all constructs containing wild-type E399 exhibit Ni2+-induced Vh shifts similar to that of wild-type mSlo1 (Fig. 3 D).

    The experimental results from Co2+ are similar to those with Ni2+ (Fig. 3, E–H). In brief, 20 mM Co2+ produces an 150-mV negative Vh shift of mSlo1 BK channels (Fig. 3 E). All constructs containing E399A show a Vh–[Co2+] relationship similar to that of the E399A mutant, in which the ability of Co2+ to produce a Vh shift is partially removed (Fig. 3, E and G). Single or double mutation of high affinity sites does not change the Vh shifts caused by Co2+ on BK channels (Fig. 3, F and H).

    The results from Ni2+ and Co2+ suggest that, among the previously defined three divalent cation regulatory mechanisms, only the E399 low affinity site contributes appreciably to the activation effect conferred by Ni2+ and Co2+. However, it should be noted that for both mM Ni2+ and Co2+ some ability to shift Vh persists even after mutation of E399, similar to the effects of Mg2+, but distinct from that of Mn2+.

    Activation of BK Channels by Sr2+ Involves All Three Divalent Cation Regulatory Sites

    Representative current traces from wild-type mSlo1, D362A/D367A+E399A, D362A/D367A, and E399A over a range of [Sr2+] are shown in Fig. 4 (A–D). For [Sr2+] >300 μM, a strong voltage-dependent block was observed that influenced current estimates for both steady state and tail currents for all tested constructs. At very positive potentials, Sr2+ produces a slow, time-dependent blockade that can result in complete block of steady-state current. Unblock following repolarization is sufficiently slow that only partial recovery from block is observed during tail currents. Intriguingly, at higher [Sr2+] and strong depolarizations, the tail current G–V curves show an unusual N shape, which cannot be simply described by a combination of activation and blocking processes (Fig. 4, E and F). Previous work has demonstrated that Sr2+ reduces BK current both by slow and fast blocking mechanisms probably involving two different positions of occupancy within the pore (Sugihara, 1998). How these two different facets of Sr2+ block may relate to the apparent N shape seen here is not clear. However, the N shape phenomenon is suggestive of a presumed punchthrough mechanism (Nimigean and Miller, 2002), in which depolarization might drive the blocking ion through the ion channel. An alternative explanation has also been suggested in recent work pointing out that an apparent punchthrough phenomenon may arise from interactions between blocker and permeant ions with the permeation pathway (Heginbotham and Kutluay, 2004).

    Because of the N shape of the G–V curves with Sr2+, we were unable to define curves reflecting solely the ability of Sr2+ to activate BK channels. Manipulations of stimulation protocols such as prolongation of the time between sweeps, varying the activation step duration, or changing the tail current potential had minimal effect on the essential properties of the tail current G–V curves. As a consequence, here we simply show the relative tail current G–V shifts among various Sr2+ concentrations. G–V curves were normalized to the conductance at 10 μM Sr2+ for each patch. Relative shifts were measured at the nominal half activation position. We must emphasize that, because of the blocking effects and the possibility that channel opening has not reached saturation at 10 μM Sr2+, the real maximum conductance activated by 300 μM Ca2+ is much larger than the peak conductance shown in these figures. Despite the uncertainties involved in determining the true Vh for activation in Sr2+, the relative ability of Sr2+ to shift gating in different constructs seems readily apparent.

    For wild-type mSlo1, there are significant shifts in the G–V curves between [Sr2+] of 0 to 10 μM (blue bars), 10 to 300 μM (red bars), and 300 μM to 20 mM (green bars) (Fig. 4 E). For D362A/D367A+E399A, in which only the calcium bowl site is intact, the relative shift between 0 to 10 μM Sr2+ is similar to the shift in G–V seen in mSlo1, but little additional shift is observed between 10 μM to 20 mM Sr2+ (Fig. 4 F). This result indicates that the calcium bowl is mainly responsible for the effect of Sr2+ <10 μM.

    When only the D362/D367 residues are mutated, the relative G–V shift between 10 and 300 μM Sr2+ is substantially reduced compared with mSlo1 (Fig. 4 G). In contrast, with the E399A mutation alone, only the relative shift between 300 μM and 20 mM Sr2+ is significantly decreased, with little change in the shifts that occur over the range of 0 to 300 μM Sr2+ (Fig. 4 H). These results suggest that the D362/D367 site accounts for Sr2+ effects from 10 to 300 μM, and E399 participates in a low affinity effect of Sr2+.

    Because the calcium bowl accounts for Sr2+ effects at 10 μM Ca2+ and below, constructs in which the calcium bowl is mutated require higher Sr2+ before activation is observed. As a consequence, for constructs containing the 5D5N mutations, the severe blocking effects of Sr2+ at the voltages required for activation precluded our ability to study such constructs. Despite this limitation, the available results all support the conclusion that, similar to Ca2+, Sr2+ activates BK channels through all three divalent cation regulatory mechanisms.

    One final aspect of the effects of Sr2+ requires mention. In both construct D362A/D367A and also D362A/D367A+E399A, there appears to be a more robust block of tail currents than in either wild-type Slo1 or in E399A, when compared at a voltage (e.g., +100 to +150 mV) where 10 μM Sr2+ appears to produce near maximal activation. This difference occurs despite the fact that both the onset of block by Sr2+ at these voltages and recovery from block at more negative potentials is largely similar for all constructs. This suggests that the voltage- and time-dependent block by Sr2+, probably involving pore blockade, does not differ among the constructs. Yet there appears to be an additional reduction in relative tail current amplitude at higher Sr2+ that may not be explained solely by such a block. At present, we do not have an explanation for the additional blockade produced in constructs containing the D362A/D367A mutations.

    In sum, each of the three mutationally defined regulatory mechanisms exhibits a unique profile of sensitivity to divalent cations. The E399A mutation defines a process that is acted upon by all divalent cations we have examined. The D362A/D367A mutations participate in a mechanism that can be regulated by Ca2+, Cd2+, and Sr2+, while activation mediated by the Ca2+ bowl only occurs with Ca2+ and Sr2+.

    The Three Regulatory Sites Affect Channel Gating Kinetics in Distinct Ways

    Despite the number of distinct kinetic states that are hypothesized to contribute to BK channel gating (Rothberg and Magleby, 1999, 2000; Zhang et al., 2001; Horrigan and Aldrich, 2002), the kinetic relaxations associated with BK current activation and deactivation are remarkably simple, being reasonably well characterized by a single exponential time course (Horrigan and Aldrich, 2002). Increases in [Ca2+] typically increase current activation rate, up to some limiting activation rate at high [Ca2+] (Zhang et al., 2001). In contrast, increases in [Ca2+] slow deactivation, with little hint of any limiting deactivation rate even at mM Ca2+ or Mg2+ (Zhang et al., 2001). The existence of a single primary kinetic relaxation in the gating behavior of BK channels has been generally explained by the idea that the rate-limiting relaxation in the gating process is a central closed–open equilibrium (C-O) (Horrigan and Aldrich, 2002).

    Given that three distinct divalent cation-dependent mechanisms appear to regulate BK gating, here we evaluate whether any of these mechanisms are associated with distinct effects on channel gating kinetics. Because Ca2+ interacts with both high affinity sites and Mg2+ only acts at the low affinity site (Shi and Cui, 2001; Zhang et al., 2001; Shi et al., 2002; Xia et al., 2002), we have compared gating behavior at 0, 1, 10, and 300 μM Ca2+ to examine the kinetic roles of the calcium bowl and D362/D367 residues, while 20 mM Mg2+ was applied to evaluate the kinetic effects of the E399 residue.

    For channels in which the calcium bowl is mutated (5D5N), the increase in activation rate between 0 and 10 μM Ca2+ is largely reduced (compare sample traces on the right of Fig. 5, C and D), while the deactivation process is less strongly influenced by Ca2+ bowl mutation at Ca2+ up to 300 μM (Fig. 5, compare C with D). In contrast, for channels with simultaneous mutation of D362A/D367A+E399A, such that only the calcium bowl site is functional, the increase in activation rate between 0 and 10 μM Ca2+ is comparable with that in wild-type mSlo1 BK channels (Fig. 5, compare C with E). Yet, in this construct, the increase in activation rate between 10 and 300 μM Ca2+ and the slowing of deactivation is almost abolished (Fig. 5 E). These results convincingly demonstrate that the main kinetic contribution of the calcium bowl is to increase the activation rate for <10 μM Ca2+, while the Ca2+ bowl plays little role in effects of Ca2+ on deactivation.

    In contrast, for channels in which the D362/D367 site alone is mutated, increases in [Ca2+] from 0 to 300 μM produce only a small slowing of deactivation in marked contrast to wild-type Slo1 (compare Fig. 6 A with Fig. 5 C). Similarly, the increase in activation rate seen in wild-type Slo1 between 10 and 300 μM Ca2+ is also largely reduced by the D362A/D367A mutation (Fig. 6 A). On the other hand, for channels in which only the D362D367 site is functional (e.g., 5D5N+E399A; Fig. 6 B), the extent of slowing of deactivation produced by Ca2+ up through 300 μM and the acceleration of activation between 10 and 300 μM Ca2+ are comparable with those in mSlo1 (compare Fig. 5 C with Fig. 6 B). Consistent with these observations, 300 μM Ca2+ confers little kinetic effect on channels in which both high affinity sites are mutated simultaneously (5D5N/D362AD367A; Fig. 6 C). Taken together, these results indicate that the D362/D367 site is primarily responsible not only for the slowing of deactivation caused by Ca2+ up through 300 μM, but also for the acceleration of activation from 10 to 300 μM Ca2+. Thus, the Ca2+ bowl site and the D362/D367A site each influence distinct kinetic aspects of BK channel function. It should also be noted that the slight slowing of activation that is observed in wild-type Slo1 by increases in Ca2+ from 0 to 1 μM appears to be absent in constructs containing the D362A/D367A mutations.

    Mg2+, which enhances BK channel activation through the E399 low affinity site (Shi et al., 2002; Xia et al., 2002), only contributes to a slowing of deactivation of Slo1, with little effect on the activation kinetics of channel gating (Fig. 7 A). The slowing of deactivation is completely abolished in channels with the E399A mutation (Fig. 7 B). Therefore, the mechanism of channel regulation involving the low affinity Mg2+ site involves a slowing of deactivation, with little effect on activation. This conclusion is also supported by the properties of the construct in which the E399 site is intact, but both higher affinity sites are mutated, i.e, 5D5N+D362A/D367A. In the absence of the two high affinity sites, Mg2+ exerts the same effects on channel gating behavior as on wild-type BK channels (Fig. 7 C). Thus, the slowing of deactivation exerted by mM Mg2+ or Ca2+ is disrupted solely by mutation of the E399 residue.

    DISCUSSION

    The main conclusions of this work are the following. (1) The calcium bowl site is activated by Ca2+ and Sr2+, but not Cd2+ or smaller divalent cations. Disruption of this site has little effect on current deactivation, but this site plays a role in Ca2+-dependent increases in activation rate <10 μM Ca2+. (2) The D362/D367-related site is activated by Ca2+, Sr2+, and Cd2+, but not smaller divalent cations. This site contributes to a slowing of deactivation produced by Ca2+ up to 300 μM, and it also contributes to Ca2+-dependent increases in activation rate between 10 and 300 μM Ca2+. (3) The E399 low affinity site plays a role in slowing of deactivation, without affecting activation; this site can be activated by all the divalent cations tested. Overall the results suggest that mutation of a given regulatory site has largely similar effects irrespective of whether either of the other two sites are functional. The unique divalent cation selectivity of these regulatory mechanisms supports the existence of three distinct divalent cation binding sites. Moreover, the specific kinetic effects of each mechanism imply that these mechanisms act in independent ways to regulate BK channels.

    At Least Three Divalent Cation Binding Sites Regulate BK Gating

    The possible presence of two RCK domains associated with each BK subunit (Jiang et al., 2002; Roosild et al., 2004) naturally suggests that a Ca2+ regulatory site on each RCK domain on a single BK channel subunit may be required for proper function of the cytosolic structure of the BK channel. As summarized in INTRODUCTION, mutational analyses have supported this idea that two distinct Ca2+-dependent regulatory mechanisms sensitive to μM Ca2+ may participate in BK channel regulation. For the Ca2+ bowl regulatory mechanism, Ca2+ binding studies to fusion peptides containing the Ca2+ bowl (Bian et al., 2001; Braun and Sy, 2001) and more detailed mutagenesis in conjunction with both physiological and binding assays (Bao et al., 2004) have bolstered the idea that the Ca2+ bowl does, in fact, represent a likely Ca2+ binding site. However, for the high affinity mechanism identified by D362AD367A (or M513), no specific evidence supports the idea that any of these residues participate in coordination of Ca2+ ions. Based on mutagenesis alone, it has remained a possibility that each subunit contains only a single Ca2+ binding site, but that mutations at different positions in the COOH terminus can disrupt independently acting components of the regulatory apparatus.

    To provide independent evidence that each mutationally defined regulatory effect corresponds to a distinct divalent cation binding site, the present results have shown that each higher affinity Ca2+ regulatory site exhibits a unique profile of divalent cation sensitivity. An important point is that, for either the 5D5N mutations or D362AD367A mutations, the consequences of a given mutation are the same, irrespective of whether other sites are simultaneously mutated. This argues against the possibility that there is only a single high affinity site that exhibits altered binding/selectivity properties after mutation. Furthermore, the results suggest that the particular profile of divalent cation sensitivity observed for each putative regulatory mechanism reflects the intrinsic properties of a unique binding site and that each site acts largely independently of each other. Thus, two distinct higher affinity Ca2+ binding sites on each Slo1 subunit contribute to regulation of BK channels by divalent cations. The specific residues that coordinate ion binding for each of these sites remain to be identified.

    The presence of two higher affinity Ca2+ regulatory domains in each subunit supports the idea that COOH terminus involves a modular structure with a Ca2+-dependent regulatory site in association with each of the two RCK domains. The D362D367 residues are, in fact, positioned in the first BK RCK domain within loops homologous to those known to contain ligand coordinating residues in the RCK domains of bacterial K+ channels (Jiang et al., 2001). The position of the Ca2+ bowl with respect to the hypothesized second RCK domain is less clear, but is consistent with the idea that a second RCK-containing module in the BK channel COOH terminus also contains a Ca2+-binding site that regulates channel gating. Intriguingly, the idea that the cytosolic structure of each BK channel subunit contains two modular elements, perhaps similar to the MthK bacterial channel (Jiang et al., 2002), was supported by earlier work indicating that functionally normal channels could arise from the separate expression of cDNA encoding a peptide corresponding to the second RCK domain (and Ca2+ bowl) along with cDNA encoding a peptide corresponding to the first RCK domain attached to the pore module (Wei et al., 1994). The point of separation between the two components of the COOH terminus in these earlier experiments was essentially a linker between the two RCK domains.

    The present results also suggest that there is a second low affinity site in addition to that defined by E374 and E399. Even at 10 mM Mg2+, an appreciable Vh shift (20–30 mV) is still observed with either the E374A (Shi et al., 2002) or E399A (Xia et al., 2002) mutations. Here, following mutation of E399, similar residual shifts were observed with mM concentrations of Ni2+ and Co2+. Furthermore, with the E399A mutation, effects of Ni2+ and Co2+ were observed at concentrations even as low as 1 mM, despite the fact that neither cation has appreciable effects on the higher affinity sites, suggesting the presence of an additional low affinity site. Given that the E399A mutation completely abolishes the effects of 5 mM Mn2+, this putative second low affinity site may exhibit a selectivity distinct from that of E399. Thus, whatever the basis for the residual effect of mM concentrations of divalent cations following mutation of E399, there are probably specific residues on the BK subunit that mediate this effect.

    Defining the Ion Selectivity of an Allosteric Regulatory Site

    A previous study (Oberhauser et al., 1988) of skeletal muscle BK channels in bilayers reported that Ca2+, Cd2+, Sr2+, Mn2+, Fe2+, and Co2+ were effective at activating channels while Mg2+, Ni2+, Ba2+, Cu2+, Zn2+, Hg2+, and Sn2+ were ineffective. However, Ni2+ and Mg2+, when studied in the presence of Ca2+, produced an additional enhancement of activation, while Ba2+, Cu2+, Zn2+, Hg2+, and Sn2+ remained ineffective. These results were interpreted in terms of two kinds of effects: first, direct channel activation at a Ca2+ activation site and, second, an allosteric effect mediated by some divalent cations that enhanced the effectiveness of Ca2+. Now, within the three binding site perspective, those cations that were observed to cause activation (Oberhauser et al., 1988) correspond either to those acting on one or both of the high affinity sites (Ca2+, Sr2+, and Cd2+) or to those (Mn2+ and Co2+) that were studied at concentrations at which effects on the low affinity site would begin to be observed. Here, because of higher concentrations and more positive activation potentials, Mg2+ and Ni2+ were also observed to activate channels directly through the lower affinity site.

    Our results also support the earlier conclusion (Oberhauser et al., 1988) that selectivity for channel activation is correlated with cation size. The Ca2+ bowl interacts with the two largest cations we have examined, Ca2+ and Sr2+. The D362/D367 site interacts with the three largest cations (Ca2+, Sr2+, and Cd2+), while the lower affinity site interacts with all cations we have examined. On the other hand, the differential effects of Ca2+ and Cd2+ on the Ca2+ bowl points out that size per se is not sufficient for defining the effects of a given cation, but that other physico-chemical properties of a given ion are important. Yet, the fundamental importance of ionic radius is clear and, in general, it is the case that the binding sites of greater specificity are those that prefer ions of larger radii. This seems somewhat surprising, given that, in a site preferring larger ions, ions of smaller radii are also likely to pass in and out of the general coordinates of that site. Two factors may contribute to the lack of effect of smaller ions. One possibility is that, similar to one model of ion permeation (Doyle et al., 1998), precise coordination distances between ion occupancy positions and oxygen donors may be critical for defining the residence time of an ion at a given site. Thus, the specific chemical nature and resulting strength of interactions between a given divalent cation and ion-coordinating residues would impact on this residence time. A second factor is that our analysis compares selectivity for activation, and not selectivity for binding. When selectivity is being compared by a measurement of a response that depends on an allosteric mechanism, the ability of a cation to cause activation depends not only on its ability to occupy the binding site, but also on its ability to bind more tightly to the open state of the channel. This latter point is, in fact, the crux of the idea of allosteric regulation. The measurements used here provide no direct information about the ability of any given ion to occupy the presumed binding sites, except as indicated indirectly by current activation. Thus, some cations might effectively occupy a binding site, but exhibit no difference in binding between closed and open conformations. Cations acting in such a fashion would be expected to exert inhibitory effects on activation by Ca2+.

    An interesting case is the apparent lack of effect of Cd2+ on the Ca2+ bowl site, despite their similarity in ionic radii (Table I). Does Cd2+ act as a blocker of the Ca2+ bowl site Although specific experiments to test this possibility will need to be done, some aspects of the data seem consistent with this possibility. In the D362A/D367A construct containing a functional Ca2+ bowl, Cd2+ produces a small rightward shift in the G–V curve at 100 μM (Fig. 1 E). One interpretation of such an effect is that Cd2+ may bind with reasonable affinity to the binding site corresponding to calcium bowl, but the binding affinity is slightly higher in the closed state than in open state. Despite their similarity in ionic radii, it has been noted that Ca2+ and Cd2+ differ appreciably in their coordination chemistry (Oberhauser et al., 1988), with Ca2+ more likely to form ionic bonds with carboxyl and carbonyl groups, whereas Cd2+ has a higher affinity for nitrogen and sulfur moieties. Perhaps the Ca2+ bowl lacks the groups appropriate for tighter binding of Cd2+ to the open channel conformation. Experiments that examine competition between Cd2+ and Ca2+ in particular constructs might illuminate this issue.

    Similar considerations to those just mentioned pertain to evaluation of the selectivity of the site defined by the D362/D367 residues. The D362/D367 related site responds to Ca2+, Cd2+, and Sr2+, but not appreciably to four cations, Ni2+, Co2+, Mn2+, and Mg2+, all with ionic radii appreciably smaller than Ca2+ (Table I). Again, it seems likely that smaller cations would enter and transiently occupy the D362/D367 site, but that the closed and open channel conformations bind these ions with a similar affinity, such that binding does not favor channel opening but perhaps may inhibit activation by Ca2+. Previous results have, in fact, suggested that Mg2+ may competitively inhibit some component of activation produced by Ca2+ perhaps by acting at the D362/D367 site (Shi and Cui, 2001; Zhang et al., 2001).

    Multiple Binding Sites and Allosteric Models of BK Channel Activation

    BK channel activation at μM Ca2+ has been effectively described by allosteric gating models (Horrigan et al., 1999; Rothberg and Magleby, 2000; Horrigan and Aldrich, 2002; Magleby, 2003) in which a central closed–open equilibrium is independently regulated by either voltage-sensor movement or by Ca2+ binding. The ability of Ca2+ binding to favor channel activation has been proposed to involve a simple mechanical coupling mediated by a linker connecting the cytosolic structure to the pore-linking S6 helix (Niu et al., 2004). To date, the most extensive models have considered a single high affinity Ca2+ regulatory site per subunit (Rothberg and Magleby, 2000; Horrigan and Aldrich, 2002; Magleby, 2003), although the consequences of including a lower affinity site on each subunit have also been considered (Shi and Cui, 2001; Zhang et al., 2001; Horrigan and Aldrich, 2002). The present results require an extension of the previous model (Horrigan and Aldrich, 2002) to include two high affinity sites along with the low affinity site. Although the apparent independence of these putative binding sites suggests that the allosteric coupling between each of the binding constants should be close to 1, further work will be required to define the allosteric interaction factors between each of the components of this complex channel.

    The present results also show that mutation of a given regulatory site alters channel gating kinetics in a stereotypical fashion irrespective of whether any or all of the other sites have also been mutated. This also supports the idea that each regulatory mechanism acts independently of each other to influence distinct aspects of the overall conformational equilibria between closed and open states. For example, both high affinity sites influence transitions that contribute to Ca2+-dependent increases in channel activation rate, although over different concentration ranges. In contrast, the low affinity site defined by E399 has no effects on channel activation rates, but plays a prominent role in slowing the return of channels to closed states. Together, these results support the view that there is a complex set of conformational steps in the cytosolic structure of the BK channel that are regulated in a largely independent fashion by ligation of divalent cations to three distinct sites.

    ACKNOWLEDGMENTS

    We thank Lynn Lavack and Yimei Yue for preparation and injection of oocytes.

    This work was supported by the National Institutes of Health grant GM066215 to C. Lingle, by the Digestive Diseases Research Core Center at Washington University (DK52574), and an AHA-Heartland affiliate fellowship to X. Zeng.

    Olaf S. Andersen served as editor.

    Submitted: 13 December 2004

    Accepted: 21 January 2005

    REFERENCES

    Bao, L., C. Kaldany, E. Holmstrand, and D. Cox. 2004. Mapping the BKCa channel's "Ca2+ bowl": side-chains essential for Ca2+ sensing. J. Gen. Physiol. 123:475–489.

    Bao, L., A. Rapin, E. Holmstrand, and D. Cox. 2002. Elimination of the BKCa channel's high affinity Ca2+ sensitivity. J. Gen. Physiol. 120:173–189.

    Barrett, J.N., K.L. Magleby, and B.S. Pallotta. 1982. Properties of single calcium-activated potassium channels in cultured rat muscle. J. Physiol. 331:211–230.

    Bian, S., I. Favre, and E. Moczydlowski. 2001. Ca2+-binding activity of a COOH-terminal fragment of the Drosophila BK channel involved in Ca2+-dependent activation. Proc. Natl. Acad. Sci. USA. 98:4776–4781.

    Braun, A., and L. Sy. 2001. Contribution of potential EF hand motifs to the calcium-dependent gating of a mouse brain large conductance, calcium-sensitive K+ channel. J. Physiol. 533:681–695.

    Butler, A., S. Tsunoda, D.P. McCobb, A. Wei, and L. Salkoff. 1993. mSlo, a complex mouse gene encoding "maxi" calcium-activated potassium channels. Science. 261:221–224.

    Cox, D.H., J. Cui, and R.W. Aldrich. 1997. Allosteric gating of a large conductance Ca-activated K+ channel. J. Gen. Physiol. 110:257–281.

    Doyle, D.A., J. Morais Cabral, R.A. Pfuetzner, A. Kuo, J.M. Gulbis, S.L. Cohen, B.T. Chait, and R. MacKinnon. 1998. The structure of the potassium channel: molecular basis of K+ conduction and selectivity. Science. 280:69–77.

    Hamill, O.P., A. Marty, E. Neher, B. Sakmann, and F.J. Sigworth. 1981. Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflugers Arch. 391:85–100.

    Heginbotham, L., and E. Kutluay. 2004. Revisiting voltage-dependent relief of block in ion channels: a mechanism independent of punchthrough. Biophys. J. 86:3663–3670.

    Herrington, J., C.R. Solaro, A. Neely, and C.J. Lingle. 1995. The suppression of Ca2+- and voltage-dependent outward K+ current during mAChR activation in rat adrenal chromaffin cells. J. Physiol. 485:297–318.

    Horrigan, F., and R. Aldrich. 2002. Coupling between voltage-sensor activation, Ca2+ binding and channel opening in large conductance (BK) potassium channels. J. Gen. Physiol. 120:267–305.

    Horrigan, F.T., J. Cui, and R.W. Aldrich. 1999. Allosteric voltage gating of potassium channels I. Mslo ionic currents in the absence of Ca2+. J. Gen. Physiol. 114:277–304.

    Jiang, Y., A. Lee, J. Chen, M. Cadene, B.T. Chait, and R. MacKinnon. 2002. Crystal structure and mechanism of a calcium-gated potassium channel. Nature. 417:515–522.

    Jiang, Y., A. Pico, M. Cadene, B.T. Chait, and R. MacKinnon. 2001. Structure of the RCK domain from the E. coli K+ channel and demonstration of its presence in the human BK channel. Neuron. 29:593–601.

    Lingle, C.J. 2002. Setting the stage for molecular dissection of the regulatory components of BK channels. J. Gen. Physiol. 120:261–265.

    Lingle, C.J., X.-H. Zeng, J.-P. Ding, and X.-M. Xia. 2001. Inactivation of BK channels mediated by the N-terminus of the 3b auxiliary subunit involves a two-step mechanism: possible separation of binding and blockade. J. Gen. Physiol. 117:583–605.

    Magleby, K.L. 2003. Gating mechanism of BK (Slo1) channels: so near, yet so far. J. Gen. Physiol. 121:81–96.

    Moczydlowski, E., and R. Latorre. 1983. Gating kinetics of Ca2+-activated K+ channels from rat muscle incorporated into planar lipid bilayers. Evidence for two voltage-dependent Ca2+ binding reactions. J. Gen. Physiol. 82:511–542.

    Nimigean, C.M., and C. Miller. 2002. Na+ block and permeation in a K+ channel of known structure. J. Gen. Physiol. 120:323–335.

    Niu, X., X. Qian, and K.L. Magleby. 2004. Linker-gating ring complex as passive spring and Ca2+-dependent machine for a voltage- and Ca2+-activated potassium channel. Neuron. 42:745–756.

    Oberhauser, A., O. Alvarez, and R. Latorre. 1988. Activation by divalent cations of a Ca2+-activated K+ channel from skeletal muscle membrane. J. Gen. Physiol. 92:67–86.

    Pauling, L. 1960. Nature of the Chemical Bond and Structure of Molecules and Crystals. Third edition. Cornell University Press, Ithaca, NY. 644 pp.

    Roosild, T.P., K.T. Le, and S. Choe. 2004. Cytoplasmic gatekeepers of K+-channel flux: a structural perspective. Trends Biochem. Sci. 29:39–45.

    Rothberg, B.S., and K.L. Magleby. 1999. Gating kinetics of single large-conductance Ca2+-activated K+ channels in high Ca2+ suggest a two-tiered allosteric gating mechanism. J. Gen. Physiol. 114:93–124.

    Rothberg, B.S., and K.L. Magleby. 2000. Voltage and Ca2+ activation of single large-conductance Ca2+-activated K+ channels described by a two-tiered allosteric gating mechanism. J. Gen. Physiol. 116:75–99.

    Schreiber, M., and L. Salkoff. 1997. A novel calcium-sensing domain in the BK channel. Biophys. J. 73:1355–1363.

    Schreiber, M., A. Yuan, and L. Salkoff. 1999. Transplantable sites confer calcium sensitivity to BK channels. Nat. Neurosci. 2:416–421.

    Shi, J., and J. Cui. 2001. Intracellular Mg2+ enhances the function of BK-type Ca2+-activated K+ channels. J. Gen. Physiol. 118:589–606.

    Shi, J., G. Krishnamoorthy, Y. Yang, L. Hu, N. Chaturvedi, D. Harilal, J. Qin, and J. Cui. 2002. Mechanism of magnesium activation of calcium-activated potassium channels. Nature. 418:876–880.

    Solaro, C.R., J.P. Ding, Z.W. Li, and C.J. Lingle. 1997. The cytosolic inactivation domains of BKi channels in rat chromaffin cells do not behave like simple, open-channel blockers. Biophys. J. 73:819–830.

    Sugihara, I. 1998. Activation and two modes of blockade by strontium of Ca2+-activated K+ channels in goldfish saccular hair cells. J. Gen. Physiol. 111:363–379.

    Wei, A., C. Solaro, C. Lingle, and L. Salkoff. 1994. Calcium sensitivity of BK-type KCa channels determined by a separable domain. Neuron. 13:671–681.

    Xia, X.-M., X.-H. Zeng, and C.J. Lingle. 2002. Multiple regulatory sites in large-conductance calcium-activated potassium channels. Nature. 418:880–884.

    Xia, X.-M., X. Zhang, and C.J. Lingle. 2004. Ligand-dependent activation of Slo family channels is defined by interchangeable cytosolic domains. J. Neurosci. 24:5585–5591.

    Zeng, X.H., X.M. Xia, and C.J. Lingle. 2003. Redox-sensitive extracellular gates formed by auxiliary subunits of calcium-activated potassium channels. Nat. Struct. Biol. 10:448–454.

    Zeng, X.-H., X.-M. Xia, and C.J. Lingle. 2001. Gating properties conferred on BK channels by the 3b auxiliary subunit in the absence of its NH2 and COOH termini. J. Gen. Physiol. 117:607–627.

    Zhang, X., C.R. Solaro, and C.J. Lingle. 2001. Allosteric regulation of BK channel gating by Ca2+ and Mg2+ through a nonselective, low affinity divalent cation site. J. Gen. Physiol. 118:607–635.(Xu-Hui Zeng, Xiao-Ming Xi)