Zn²⁺ Activates Large Conductance Ca²⁺-activated K⁺ Channel via an Intracellular Domain

Abstract

Zinc is an essential trace element and plays crucial roles in normal development, often as an integral structural component of transcription factors and enzymes. Recent evidence suggests that intracellular Zn²⁺ functions as a signaling molecule, mediating a variety of important physiological phenomena. However, the immediate effectors of intracellular Zn²⁺ signaling are not well known. We show here that intracellular Zn²⁺ potently and reversibly activates large-conductance voltage- and Ca²⁺-activated Slo1 K⁺ (BK) channels. The full effect of Zn²⁺ requires His³⁶⁵ in the RCK1 (regulator of conductance for K⁺) domain of the channel. Furthermore, mutation of two nearby acidic residues, Asp³⁶⁷ and Glu³⁹⁹, also reduced activation of the channel by Zn²⁺, suggesting a possible structural arrangement for Zn²⁺ binding by the aforementioned residues. Extracellular Zn²⁺ activated Slo1 BK channels when coexpressed with Zn²⁺-permeable TRPM7 (transient receptor potential melastatin 7) channels. The results thus demonstrate that Slo1 BK channels represent a positive and direct effector of Zn²⁺ signaling and may participate in sculpting cellular response to an increase in intracellular Zn²⁺ concentration.

Introduction

Zinc is the second most abundant transition metal in the human body, playing a pivotal role in the normal development and growth. The utmost importance of zinc is evidenced by the diverse array of symptoms that could result from a chronic dietary deficiency of zinc (1). Biochemically, zinc serves as an essential structural and a catalytic component in many metalloproteins (2), in which the metal is typically coordinated by four or five ligands (3). Multiple zinc coordination geometries are known, but histidine and cysteine typically act as essential ligands (4).

In addition to its role as an integral structural and catalytic factor, Zn²⁺ is increasingly recognized as a potential intracellular signaling molecule, similar to Ca²⁺ (5, 6). Like intracellular Ca²⁺, intracellular Zn²⁺ is normally kept to a very low concentration, from pm to nm (5). Although measurements of free intracellular Zn²⁺ concentrations ([Zn²⁺]_i) in living cells remain challenging, studies do suggest that [Zn²⁺]_i may significantly increase under some conditions. For example, a robust release of Zn²⁺ from the endoplasmic reticulum, termed “zinc wave,” has been observed in response to extracellular stimuli, further suggesting that Zn²⁺ may act as an intracellular second messenger (7). In addition, local [Zn²⁺]_i may be significantly higher near Zn²⁺-permeable channels (5, 8), analogous to the well known micro- and nano-domains of intracellular Ca²⁺ (9). Moreover, [Zn²⁺]_i may increase concomitantly with [Ca²⁺]_i under pathological conditions such as ischemia/hypoxia (5, 6, 10, 11), in which intracellular Ca²⁺ overload is suspected to contribute to cell death in these conditions (12). However, whether such increases in [Zn²⁺]_i contribute to the deleterious effect or play a compensatory cell-protective effect is not clear (5, 11, 13–16).

Large-conductance voltage- and Ca²⁺-activated K⁺ (BK_Ca, Slo1 BK or K_Ca1.1) channels are distinguished by their allosteric activation by voltage and intracellular Ca²⁺ (17–19). Like other voltage-gated K⁺ channels, a BK channel complex includes four pore-forming α (Slo1) subunits, each of which contains a voltage sensor domain (S1–S4) and one-fourth of the ion conduction pore (S5–S6) (20). In addition, each Slo1 subunit possesses the transmembrane segment S0 (21) and a large cytoplasmic area harboring two homologous domains termed “regulators of conductance of potassium” (RCK1 and RCK2) essential for activation by Ca²⁺ for the channel (22, 23). Functionally, BK channels participate in many crucial physiological phenomena including vasoregulation, synaptic transmission, and hormone secretion mainly by affecting membrane excitability (17). In addition, as a feedback controller of intracellular Ca²⁺, BK channel activation has been demonstrated to have a potent cell protection effect by limiting the influx of Ca²⁺ during hypoxia/ischemia (24, 25).

The concomitant increases in [Zn²⁺]_i and [Ca²⁺]_i in ischemia/hypoxia and the cytoprotective role of the BK channel under the pathological conditions prompted us to examine whether Zn²⁺ is also a physiological activator of the channel. The Slo1 protein indeed contains multiple putative Zn²⁺-binding amino acid sequences such as HXXXH (X represents any amino acid) identified in other metal-binding proteins including S100 proteins, the largest subgroup of the EF-hand Ca²⁺-binding protein family (26–28). Our excised patch clamp measurements from heterologously expressed human Slo1 (hSlo1)³ BK channels revealed that intracellular Zn²⁺ robustly activates the channel and that mutation of one histidine residue in the RCK1 domain fully abolished the stimulatory effect of Zn²⁺. Our results therefore suggest that Slo1 coordinates Zn²⁺ using amino acid ligands in the RCK1 domain and that the Slo1 BK channel is a positive effector of intracellular Zn²⁺ signaling.

EXPERIMENTAL PROCEDURES

Channel Expression

Human Slo1 (KCNMA1; U11058) and its mutants in the expression vector pCI-neo (Promega), HA-tagged rat TRPM7 (XP_001056331) in pTracer-CMV vector (Invitrogen), and rat SK2 (KCNN2; U69882) in the expression vector pcDNA3 (Invitrogen) were transiently expressed in HEK tsA cells using FuGENE 6 (Roche) as described (29). In some experiments, hSlo1 and β1 (KCNMB1; U38907) in pEGFP-N1 (Clontech) were transfected together with a weight ratio of 1:1. The mutant channels were constructed using a PCR-based mutagenesis method (Agilent), and the sequences were verified (University of Pennsylvania DNA Sequencing Facility).

Electrophysiology and Data Analysis

Ionic currents were recorded using the cell-attached or excised inside-out configuration at room temperature. Patch electrodes (Warner) had a typical initial resistance of 1.5–2.0 megohms. The series resistance, up to 90% of the initial input resistance, was electronically compensated in the macroscopic current measurements. Macroscopic capacitive and leak currents were subtracted using a P/6 protocol. The current signal was filtered at 10 kHz through the built-in filter of the patch clamp amplifier (AxoPatch 200A; MDS Analytical Technologies) and digitized at 100 kHz using an ITC-16 AD/DA interface (HEKA). Conductance-voltage (G-V) curves were generated from tail currents and fitted with a Boltzmann equation as described (29). The resulting half-activation voltage (V_0.5) was used to quantify the effect of Zn²⁺ on the channel. Both activation and deactivation time constants were obtained by fitting the currents with a single exponential excluding the initial 180 μs. The results were analyzed as described using IGOR Pro (WaveMetrics) (29). Statistical comparisons between two groups were performed using the unpaired or paired t test, as appropriate. Comparison of more than two groups was performed using analysis of variance followed by a Tukey HSD test as implemented in IGOR Pro. Statistical significance was assumed at p ≤ 0.05, and the data are presented as mean ± S.E. The number of samples in each group is shown in parentheses unless noted otherwise.

Chemicals and Solutions

All chemicals were from Sigma except for 2-aminoethyl methanethiosulfonate hydrobromide (MTSEA; Biotium). TPEN was dissolved in dimethyl sulfoxide and diluted with the internal recording solution to the final concentration of 10 μm. The final concentration of dimethyl sulfoxide (0.02%, v/v) did not affect Slo1 channel currents. For inside-out patch recording, the extracellular solution contained 140 mm KCl, 2 mm MgCl₂, 10 mm HEPES, pH 7.2, with N-methyl-d-glucamine (NMDG). The intracellular solution contained 140 mm KF, 10 mm HEPES, pH 7.2 or 6.2, with NMDG and a different concentration of ZnCl₂ or ZnSO₄. The use of KF in the solution limited [Ca²⁺] < 20 nm (30). In the experiments with high concentrations of Ca²⁺, the intracellular solution did not contain any chelator and the pH was adjusted to 7.2 with NMDG. For the cell-attached patch experiment, the electrode solution contained 140 mm KCl, 2 mm MgCl₂ or 2 mm ZnCl₂, 10 mm HEPES, pH 7.2, with NMDG. The bath solution contained 130 mm NaCl, 4.0 mm KCl, 2 mm CaCl₂, 2 mm MgCl₂, 10 mm HEPES, 15 mm glucose, pH 7.4, with NMDG.

RESULTS

Intracellular Zn²⁺ Activates hSlo1 Channels

To observe the effect of cytoplasmic Zn²⁺ on the Slo1 channel while maintaining a very low concentration of Ca²⁺, we used KF in the internal solution in which most of the contaminating Ca²⁺ precipitated due to the low solubility of CaF₂. In such an internal solution, the free Ca²⁺ concentration has been estimated to be <20 nm (30). Consistently, we found that the activity of the hSlo1 channel remained unaltered when the inside-out patches were transferred from the KF internal solution (see “Experimental Procedures”) to the KCl internal solution with 11 mm EGTA in which [Ca²⁺] is calculated to be <10 nm (WEBMAXC STANDARD; data not shown). In addition, we found that up to 300 μm of Zn²⁺ in the KF solution failed to activate the small-conductance Ca²⁺-activated channel 2 (SK2), which has higher Ca²⁺ sensitivity than the Slo1 channel (9) (data not shown). These observations together affirmed that [Ca²⁺]_i was appropriately buffered to a negligible level when Zn²⁺ was added into the KF internal solution.

Addition of Zn²⁺ (0.3–300 μm) quickly and reversibly increased hSlo1 BK currents (Fig. 1, A and C) in a concentration-dependent manner (Fig. 1B). TPEN, a Zn²⁺ chelator with low affinity for Ca²⁺, fully antagonized the stimulatory effect of the Zn²⁺ addition to the intracellular solution (Fig. 1, C and D), further confirming that it was Zn²⁺ that increased the hSlo1 current. In contrast, extracellular Zn²⁺, up to 2 mm, was without any stimulatory effect (see Fig. 6B; see also Ref. 31).

FIGURE 1.

Application of Zn²⁺ to the cytoplasmic side activates hSlo1 channels. A, representative hSlo1 currents at 100 mV without and with 1 μm Zn²⁺ (top). The currents were elicited by pulses from 0 to 100 and then to 0 mV. The peak outward current size at 100 mV is plotted as a function time (bottom). B, representative hSlo1 currents (top) and values of normalized currents (I/I_Con, bottom) at different concentrations of Zn²⁺ at 100 mV. The values of I/I_Con were 1.17 ± 0.14, 2.09 ± 0.31, 2.62 ± 0.62, 4.88 ± 0.68, 10.50 ± 1.19, 18.90 ± 1.40, and 18.41 ± 1.52 at 0.1, 0.3, 1, 10, 30, 100, and 300 μm Zn²⁺, respectively. The currents were elicited by pulses from 0 to 100 and then to −80 mV. C, Zn²⁺ reversibly and repeatedly increased hSlo1 channel currents, but 10 μm TPEN abolished the effect of Zn²⁺. The currents were elicited as in A. D, fractional increase in the peak current size by Zn²⁺ (10 μm) in the absence and presence of TPEN (10 μm). The currents were elicited as in C. *, p < 0.01 compared with control group (n = 3 in each group). E, I-V curves of hSlo1 channels in the absence (open circles) and presence of 10 μm (filled circles), 30 μm (filled squares), and 100 μm (filled diamonds) Zn²⁺. F, G-V curves of hSlo1 channels with different concentrations of Zn²⁺: 0 μm (no Zn²⁺ added; open circles), 10 μm (filled circles), 30 μm (filled squares), and 100 μm (filled diamonds). The steepness of the G-V curves with Zn²⁺ was not different to that in the control condition (p > 0.69). G, V_0.5 changes by different concentration of Zn²⁺. The concentration response was fitted by a Hill equation, ΔV_0.5(x) = ΔV_0.5(max)/[1 + (EC₅₀/x)ⁿ] where n is the Hill coefficient, x is the Zn²⁺ concentration, and ΔV_0.5(max) is the maximal shift in V_0.5. H, representative normalized currents recorded at 150 mV before (thin trace) and after (thick trace) application of 100 μm Zn²⁺ (left). Time constants of activation before and after application of 100 μm Zn²⁺ (n = 7 in each group) (right). I, representative normalized currents recorded at −80 mV after 40-ms pulses of 150 mV in the absence (thin trace) and presence (thick trace) of 100 μm Zn²⁺ (left). Time constants of deactivation before and after application of 100 μm Zn²⁺ (n = 8 in each group) (right). *, p < 0.05 compared with control group. J, representative single-channel currents from different patches at −100, −50, and 50 mV before and after application of Zn²⁺. 10 μm and 100 μm Zn²⁺ were used at positive and negative voltages, respectively. Con, control.

FIGURE 6.

Coexpression with TRPM7 channels facilitates opening of hSlo1 channels. A, representative hSlo1 currents recorded at the time indicated in the presence of extracellular Mg²⁺ (top) or Zn²⁺ (bottom). Single-channel currents were recorded at −80 mV in the cell attached configuration from cells transfected with hSlo1 and TRPM7. The time 0 min indicates seal formation. Similar results were obtained in 8 of 13 cells. B, mean time courses of changes in channel open probability in the presence of extracellular Mg²⁺ (open symbols) or Zn²⁺ (filled symbols) with hSlo1 channels alone (circles) or with hSlo1 and TRPM7 channels together (squares). Single channel currents were recorded as in A. N, number of channels; P_o, open probability.

The current-enhancing effect of Zn²⁺ was voltage-dependent (Fig. 1E) and accompanied by a shift in G-V to the hyperpolarized direction without any change in the steepness (Fig. 1F). Saturating concentrations of Zn²⁺ (≥100 μm) produced a shift in V_0.5 of about −75 mV. The Zn²⁺-dependent shift in G-V V_0.5 had an EC₅₀ value of 33.6 ± 12.2 μm and a Hill coefficient of 0.93 ± 0.22 (Fig. 1G).

We noticed that high concentrations of Zn²⁺ slightly diminished the peak outward currents at extreme positive voltages (e.g. 200 mV in Fig. 1E) without decreasing the inward tail current size. This small inhibitory effect, most probably reflecting voltage-dependent block of the channel pore by Zn²⁺ (30), was not investigated any further. In addition to the shift of voltage dependence of activation to the hyperpolarized direction, Zn²⁺ slowed the deactivation kinetics without affecting the activation kinetics (Fig. 1, H and I).

The stimulatory effect of Zn²⁺ was also observed at the single-channel level. Zn²⁺ drastically increased single-channel open probability in a wide range of voltages, including a physiologically relevant negative voltage (−50 mV) and an extreme negative voltage where the primary voltage sensors of the channel are not activated (Fig. 1J). Zn²⁺ had no noticeable effect on the unitary current size (Fig. 1J).

Zn²⁺-dependent Activation of the hSlo1 BK Channel Did Not Require the Conserved Zinc-binding Motifs

Structural studies suggest that histidine and cysteine are the two most frequently used zinc ligands in metalloproteins, in which zinc interacts with the imidazole nitrogen or thiol sulfur in the conserved zinc-binding motifs such as HXXXH, and CXXXH (X represents any amino acid) (2–4, 27, 28, 32). Mutation of either histidine residue in the conserved motif typically disrupts the Zn²⁺ coordination and reduces catalytic activity of metalloenzymes (26, 33). Inspection of the hSlo1 sequence shows that the cytoplasmic domain of the channel contains three putative zinc-binding motifs, ⁴⁶⁴HNKAH⁴⁶⁸, ⁷⁴⁹HELKH⁷⁵³, and ⁶¹²CKACH⁶¹⁶, localized in RCK1, RCK2, and the linker region between the two RCK domains, respectively (Fig. 2A). To assess the contributions of His and Cys to the Zn²⁺-induced Slo1 BK channel activation, we utilized diethyl pyrocarbonate, a histidine-modifying reagent (34), and a cysteine-modifying reagent, MTSEA (35). Our results showed that pretreatment of the channel with diethyl pyrocarbonate significantly attenuated the Zn²⁺-induced activation of the channel, decreasing the V_0.5 shift to ∼50% of that observed in the control group (Fig. 2, B and H). In contrast, MTSEA failed to alter the Zn²⁺-induced channel activation (Fig. 2, C and H). We thus reasoned that the Slo1 protein interacts with Zn²⁺ using histidine residues, possibly in the aforementioned zinc motifs (Fig. 2A).

FIGURE 2.

Zn²⁺ remains effective in mutant Slo1 channels with the conserved zinc-binding motifs disrupted. A, a schematic representation of the wild-type hSlo1 channel, showing three potential Zn²⁺-binding motifs in the RCK1 and RCK2 domains and the linker between of the two domains (top). B and C, representative G-V curves before (open circles) and after (filled circles) application of 100 μm Zn²⁺ following pretreatment with 2 mm diethyl pyrocarbonate for 5 min (B) and 1 mm MTSEA for 10 min (C). D–F, representative G-V curves of the His-to-Arg mutant channels indicated in the absence (open circle) and presence (filled circle) of 100 μm Zn²⁺. G, representative single-channel currents recorded from hSlo1 H616R in the absence (top) and presence (bottom) of 100 μm Zn²⁺. The single-channel openings were elicited by pulses to 100 mV from 0 mV. Similar results were observed in another three patches. H, changes in V_0.5 by 100 μm Zn²⁺ in the wild-type channels with and without diethyl pyrocarbonate or MTSEA pretreatment and also in the His-to-Arg mutants. *, p < 0.001 compared with the no treatment wild-type group.

The potential involvement of the histidine residues in the zinc-binding motifs in the channel was further tested by mutation of His⁴⁶⁴, His⁶¹⁶, His⁷⁴⁹, and His⁷⁵³. A robust stimulatory effect of Zn²⁺, indistinguishable from that in the wild-type channel, remained in these His-to-Arg mutants (Fig. 2, D–F). The mutant channel H616R (29) did not express well enough to record macroscopic currents; however, the mutant retained a Zn²⁺ sensitivity indistinguishable from that of the wild-type channel based on single-channel measurements (Fig. 2G). These results collectively indicated that the His residue(s) that coordinate Zn²⁺ are located elsewhere.

Zn²⁺ Is Less Effective at Low pH

The bound Zn²⁺ can be removed from the metalloenzymes in low pH conditions, possibly owing to the protonation of imidazole nitrogens (36). We therefore examined whether intracellular H⁺ affected the action of Zn²⁺ on the Slo1 channel. The Zn²⁺-induced shift in V_0.5 was indeed significantly reduced in the pH 6.2 internal solution to −25.1 ± 7.1 mV, less than a half of that at pH 7.2 (p < 0.01; Fig. 3, A and F).

FIGURE 3.

His³⁶⁵ in the RCK1 domain is required for Zn²⁺ action. A, G-V curves in the wild-type channel at pH 6.2 with (filled circles) and without (open circles) 100 μm Zn²⁺. B–E, typical G-V curves in the His mutants at pH 7.2 before (open circles) and after (filled circles) application of 100 μm Zn²⁺. F, changes in V_0.5 in the wild-type channel at pH 7.2 and 6.2 and in the His mutants at pH 7.2. *, p < 0.001 compared with the wild-type (WT) channel at pH 7.2.

Mutation of His³⁶⁵ Abolishes the Zn²⁺ Effect

We previously demonstrated that two His residues, His³⁶⁵ and His³⁹⁴, in the RCK1 domain serve as the primary H⁺ sensors of the hSlo1 channel and mediate pH-dependent activation of the channel (37, 38). The antagonistic effect of low pH on the Zn²⁺-dependent activation suggests that the same His residues may be required for the Zn²⁺ action. Consistent with this possibility, the double mutation H365R/H394R completely abolished the effect of Zn²⁺ on V_0.5; the ΔV_0.5 value was −2.8 ± 5.1 mV (p < 0.001 compared with the wild-type channel; Fig. 3, B and F). Of the two His residues, His³⁶⁵ clearly plays the most important role, for the single mutation H365R alone eliminated the Zn²⁺ sensitivity (ΔV_0.5 = −5.5 ± 2.9 mV; p < 0.0001 compared with the wild-type channel; Fig. 3, C and F). Mutation of His³⁶⁵ to neutral alanine (H365A) also completely disrupted the Zn²⁺ sensitivity of the channel (−6.0 ± 3.9 mV; p < 0.001 compared with the wild-type channel and p > 0.5 compared with H365R) (Fig. 3, D and F). In contrast, the mutant H394R remained fully Zn²⁺-sensitive (ΔV_0.5 = −55.2 ± 1.2 mV; p > 0.5; Fig. 3, E and F). While both His³⁶⁵ and His³⁹⁴ in the RCK1 domain are important for pH-dependent activation of the hSlo1 channel (37, 38), only His³⁶⁵ is required for the Zn²⁺-dependent activation of the channel.

Select Acidic Residues in the RCK1 Domain Implicated in the Ca²⁺ Sensitivity Are also Important for the Zn²⁺ Action

His³⁶⁵, required for the Zn²⁺-dependent activation of the hSlo1 channel (see Fig. 3) also participates in both Ca²⁺- and H⁺-dependent activation of the Slo1 channel such that the stimulatory effect of H⁺ is diminished at higher concentrations of Ca²⁺ (37, 38). We hypothesized that Ca²⁺ may also interfere with the Zn²⁺-dependent activation of the channel. As predicted by this idea, we found that in the presence of 100 μm Ca²⁺, which is a saturating concentration for the high-affinity Ca²⁺ sensors of the Slo1 channel (39–42), Zn²⁺ failed to alter G-V (Fig. 4, B and F), indicative of a functional competition between Zn²⁺ and Ca²⁺.

FIGURE 4.

Negatively charged residues in the RCK1 domain contribute to the Zn²⁺-dependent activation of the Slo1 channel. A, a schematic representation of the potential Zn²⁺/Ca²⁺ sites in the wild-type hSlo1 channel (top) and a homology model of the mouse Slo1 (mSlo1) RCK1 domain (bottom) based on the structure of MthK channel (62). The residues required for the effects of Zn²⁺, Ca²⁺, and Mg²⁺ are highlighted. The mSlo1 sequence is identical to that of hSlo1 in the RCK1 domain. The images were prepared with MacPyMOL. B, a representative G-V curve in the wild-type channel in the presence of 200 μm Ca²⁺ alone (open circles) and of 200 μm Ca²⁺ and 100 μm Zn²⁺ together (filled circles). C–E, representative G-V curves in the Ca²⁺ sensor mutants in the absence (open circles) and presence (filled circles) of 100 μm Zn²⁺. F, changes in V_0.5 caused by 100 μm Zn²⁺ in the wild-type (WT) and the mutant channels. The mutation Δ884–885 impairs the Ca²⁺ bowl function (39). *, p < 0.001 compared with the wild-type channel and #, p < 0.01 compared with the H365R channel.

Previous mutagenesis studies suggest the presence of at least three potential divalent cation sensors in each Slo1 subunit (18, 41–43) (Fig. 4A); a high-affinity sensor in the RCK1 domain, a high-affinity Ca²⁺ bowl sensor, and a low affinity sensor in the RCK1 domain that also mediates Mg²⁺-dependent activation of the channel (42). The charge-neutralization mutation D367A in the RCK domain is known to disrupt the high-affinity Ca²⁺-sensing by the RCK1 domain (41). We found that the mutation significantly decreased the shift in V_0.5 by 100 μm Zn²⁺ by ∼35% to −36.9 ± 5.6 mV (p < 0.01 compared with the wild-type channel; Fig. 4, C and F). The function of the high-affinity Ca²⁺ bowl sensor in the RCK2 domain is disrupted by the deletion mutation Δ884–885 (39). This deletion mutation, however, failed to alter the stimulatory effect of Zn²⁺ on the channel (Fig. 4, D and F).

The low-affinity divalent cation sensitivity of the Slo1 channel is in part mediated by Glu³⁹⁹ in the RCK1 domain (43). The mutation E399A, which impairs the stimulatory action of mm levels of Mg²⁺ on the channel (42, 43), noticeably attenuated the Zn²⁺-dependent shift in V_0.5 by ∼35% to −37.2 ± 1.8 mV (p < 0.01 compared with the wild-type channel; Fig. 4, E and F). The shifts in V_0.5 by Zn²⁺ in the D367A and E399A mutants were statistically indistinguishable (Fig. 4F).

Other transition metals, such as Mn²⁺, also activate the Slo1 channel (30, 42). We found that the effect of Mn²⁺ was completely disrupted by the mutation E399A but not by the mutation H365A, which eliminates the Zn²⁺ sensitivity (supplemental Fig. S1).

Coexpression of β1 Subunit Does Not Alter the Effect of Zn²⁺

In addition to the four pore-forming Slo1 subunits, a native BK channel complex may also include auxiliary β subunits in a tissue-dependent manner (44). Heterologous coexpression of the auxiliary subunit β1, predominantly expressed in the cardiovascular system, dramatically increases the overall Ca²⁺ sensitivity and slows both the activation and deactivation kinetics of the channel complex (44). The underlying mechanism is postulated to involve an increase in the Ca²⁺ affinity of the high-affinity Ca²⁺ sensors in the RCK1 domain and the Ca²⁺ bowl in the RCK2 domain (45). Because the stimulatory effect of Zn²⁺ on the Slo1 channel was in part dependent on Asp³⁶⁷, an established component in the high-affinity RCK1 Ca²⁺ sensor (41), we examined whether coexpression of β1 enhanced the effectiveness of Zn²⁺. Functional coexpression of β1 was verified by the characteristically slower activation and deactivation kinetics. We found that Zn²⁺ remained effective in enhancing the Slo1 current. The shift in V_0.5 (−56.8 ± 2.4 mV) was indistinguishable from that without coexpression of β1 (Fig. 5; p > 0.5).

FIGURE 5.

Coexpression of β1 does not alter the effectiveness of Zn²⁺. A, representative hSlo1+β1 currents at 100 mV before (thin trace) and after (thick trace) application of 100 μm Zn²⁺. The currents were elicited by pulses from 0 to 100 and then to −80 mV. B and C, typical I-V curves and G-V curves of hSlo1+β1 channels in the absence (open circles) and presence (filled circles) of 100 μm Zn²⁺. D, changes in V_0.5 caused by 100 μm Zn²⁺ in Slo1 channel with or without the β1 subunit.

Extracellular Zn²⁺ Activates Slo1 BK Channel when Coexpressed with TRPM7

Many membrane transport proteins including ion channels mediate translocation of the extracellular Zn²⁺ into intracellular space. Extracellular Zn²⁺ did not affect the Slo1 channel activity; however, it robustly activated the channels when they were coexpressed with TRPM7, a nonselective cation channel permeable to Zn²⁺ (46, 47). In contrast, extracellular Mg²⁺ did not alter Slo1 channel open probability (Fig. 6).

DISCUSSION

Zn²⁺ is well known for its structural role in a large number of metalloproteins, including some voltage-gated K⁺ channels in which the metal ion mediates tetramerization of the channel proteins (48, 49). As an important intracellular messenger, Zn²⁺ also modulates multiple signaling pathways, but yet only a small number of its direct effectors have been clearly identified (5, 7). Among ion channels, recent studies show that the TRPA1 (transient receptor potential channel A1) (50, 51) and the ATP-sensitive K⁺ channel (K_ATP) (52) are activated by intracellular Zn²⁺ at nm and μm concentrations, respectively. Our study now adds the Slo1 channel as a new member of the Zn²⁺ signaling cascades. Heterologously expressed Slo1 BK channels are robustly activated by μm levels of intracellular Zn²⁺ in cell-free membrane patches, independently of the auxiliary subunit β1. Moreover, mutation of His³⁶⁵ in the RCK1 domain or nearby Asp³⁶⁷ or Glu³⁹⁹ involved in the Ca²⁺ sensing fully or partially abolished the channel activation by Zn²⁺.

Our finding that Zn²⁺ activates heterologously expressed Slo1 channels is in contrast with a previous report that Zn²⁺ had no effect on rat skeletal muscle BK channels incorporated in planar lipid bilayers (30). The reason for the apparent discrepancy is not clear. It may be noted that the authors also failed to observe any stimulatory effect of Mg²⁺, an established activator of Slo1 channels (43, 53) in the same study (30).

The Mechanism of Channel Activation by Zn²⁺

The functional competition between Zn²⁺ and Ca²⁺ in activation of the Slo1 channel observed in this study is in line with the mutagenesis result that Asp³⁶⁷, essential for the normal high-affinity Ca²⁺ sensing of the channel (41), is also required for Zn²⁺ action. Accordingly, the mechanism of channel activation by Zn²⁺ may be similar to that by Ca²⁺. Although physical measurements of Ca²⁺ binding to the RCK1 sensor and the Ca²⁺ bowl sensor are preliminary (54–56), conformational changes in an isolated hSlo1 cytoplasmic domain induced by Ca²⁺ have been detected (54). Structural and functional studies of MthK and Slo1 suggest that Ca²⁺-dependent activation of the Slo1 channel may be accompanied by an expansion of the cytoplasmic domain termed a “gating ring” (22), the mechanical energy of which is further coupled to the channel pore (19, 57). Like Ca²⁺, Zn²⁺ may induce a similar expansion of the gating ring to promote activation of the channel. However, some differences between the effects of Ca²⁺ and Zn²⁺ exist. The maximal shift in V_0.5 by Zn²⁺, about −75 mV, is clearly smaller than that by Ca²⁺, which can produce a shift of −200 mV at 300 μm (42). One readily discernible reason for the difference is that the Ca²⁺ action is supported by both the sensor in the RCK1 domain and the Ca²⁺ bowl sensor in the RCK2 domain (39, 41, 54, 58). Even in the absence of the Ca²⁺ bowl sensor, 300 μm Ca²⁺ can produce a −125 mV shift (41), still greater than that by Zn²⁺. The maximal V_0.5 shift by Zn²⁺ is similar to that caused by H⁺, which also works via the RCK1 sensor and functionally competes with Ca²⁺ (37). The smaller shift by H⁺ is attributed to weaker allosteric coupling between the gate of the channel and the RCK1 sensor when H⁺ is bound as compared with that with Ca²⁺ bound (38). Thus, the coupling strength in the presence of Zn²⁺ may be similarly lower than that with Ca²⁺. Another difference between the effects of Ca²⁺ and Zn²⁺ relates to Glu³⁹⁹ in the RCK1 domain, a critical component in the low-affinity divalent ion sensing of the channel, and its neutralization impairs the channel activation by mm levels of Mg²⁺ (43). Whereas the stimulatory effect of μm levels of Ca²⁺ does not depend on Glu³⁹⁹, the effect of Zn²⁺ is diminished when Glu³⁹⁹ is neutralized (Fig. 4). The action of Zn²⁺ is thus influenced by the residues involved in both the high-affinity and low affinity divalent cation sensing mechanism (19). The biophysical mechanism of the channel activation by Zn²⁺ may be similar to that by Ca²⁺ because, unlike effect of Mg²⁺ (59), the Zn²⁺ action remains effective even at negative voltages where the voltage sensors of the channel are not activated. Finally, coexpression with β1 enhances the shift in V_0.5 by Ca²⁺ but does not alter that by Zn²⁺ (Fig. 5). A similar β1-indepenent effect is observed with intracellular H⁺ (37, 38) further supporting the idea that Zn²⁺ and H⁺ may share a similar mechanism in Slo1 channel activation.

Zinc Coordination by Slo1

In many metalloproteins that contain zinc as a stable cofactor, the metal is coordinated by a water molecule and three to four ligands provided by the amino acid residues, typically the side chains of His, Glu, Asp, and Cys (4). Some proteins coordinate zinc using His, Asp, and Glu (4). In Slo1, at least His³⁶⁵, Asp³⁶⁷, and Glu³⁹⁹ contribute to the stimulatory effect of Zn²⁺ and His365 is required. The lack of a high-resolution atomic structure of the channel, however, precludes a detailed inference on the zinc coordination geometry. Furthermore, unlike most other zinc-containing proteins, binding of Zn²⁺ to the channel is rapid and readily reversible, and it is not clear how applicable the structural information obtained from the metalloproteins that contain zinc as a stable cofactor may be to the Slo1 protein. Many intracellular EF-hand Ca²⁺-binding proteins also reversibly bind to Zn²⁺ at concentrations similar to those used to activate Slo1 BK channels (28). Structural studies suggest that Zn²⁺ is often located in close proximity to Ca²⁺ sites and that the two ions reciprocally modulate binding of the other (60, 61), in agreement with our finding that Ca²⁺ and Zn²⁺ competitively activate Slo1 BK channel. The RCK1 domain, which contains the His residue essential for the Zn²⁺ action, was once postulated to contain an EF-hand-like domain (55). However, subsequent structural studies on the prokaryotic channel MthK, which shares a high level of sequence similarity in this area with the Slo1 BK channel, did not support this idea (22, 23). The homology model of Slo1 (Fig. 4A) (62) based on the MthK structure clearly shows that Asp³⁶⁷ and Glu³⁹⁹, are located in the vicinity of His³⁶⁵, forming a potential ligand binding pocket that accommodates a Ca²⁺, H⁺, or carbon monoxide (37, 63). The requirement for His and the contributions from Asp and Glu in Zn²⁺ activation of the Slo1 channel are in line with the zinc coordination schemes found in metalloproteins such as an Escherichia coli rhamnose isomerase (4, 64). We therefore suggest that His³⁶⁵, Asp³⁶⁷, and Glu³⁹⁹ in the RCK1 sensor coordinate Zn²⁺, and the conformational change of the sensor promotes opening of the gate. In TRPA1 channels, which are also activated by intracellular Zn²⁺, His and Cys residues located some distance away in the primary sequence appear to play a critical role in the Zn²⁺ sensitivity (51).

Physiological and Pathophysiological Implications

Our study demonstrated that human Slo1 BK channels were activated by high nm to μm of intracellular Zn²⁺. Similar concentrations were also used in Zn²⁺ modulation of other intracellular proteins such as mitochondrial enzymes (65–67) and ion channels (52). For instance, the EC₅₀ for activation of recombinant K_ATP channels, sulfonylurea receptor (SUR)1/Kir6.2 and SUR2A/Kir6.2 are 1.8 and 60 μm, respectively (52). Such [Zn²⁺]_i may not be observed physiologically in the bulk intracellular compartment. However, local [Zn²⁺]_i may reach higher levels near intracellular Zn²⁺ stores or Zn²⁺ permeable channels and it plays important roles in normal neuronal transmission and immune response (5, 7, 68). Interestingly, some Zn²⁺-permeable ion channels may physically colocalize with Slo1 BK channels, potentially exposing the latter to a locally high level of Zn²⁺ (69). Although quantitative studies of such local Zn²⁺ domains are unavailable, functional analyses of local Ca²⁺ domains suggest that the [Ca²⁺]_i near Slo1 BK channels can be a few orders of magnitude greater than the mean bulk concentration (9, 69). Thus it is plausible that the local [Zn²⁺]_i increases transiently to a μm level to activate Slo1 BK channels. Our results (Fig. 6) show that such an increase in [Zn²⁺]_i could occur through an influx of Zn²⁺ from the extracellular compartments mediated by Zn²⁺-permeable TRPM7 channels (46, 47, 70). The extracellular concentration of Zn²⁺ in confined compartments such as synaptic clefts may reach several hundred μm (5). Because both TRPM7 and Slo1 BK channels are widely expressed, TRPM7 channels could inject enough Zn²⁺ to activate Slo1 BK channels. Along with TRPA1 (50, 51) and K_ATP channels (52), Slo1 BK channels now represent a family of intracellular Zn²⁺-activated ion channels that could play physiological roles. Increases in [Zn²⁺]_i may be even greater under some pathological conditions such as brain ischemia/reperfusion and epilepsy (5, 12). For example, in the experimental seizures induced by kainic acid, [Zn²⁺]_i may increase to hundreds of nm and several μm in hippocampal and cortical neurons, respectively (71, 72). A recent study suggests that the actual increase in [Zn²⁺]_i during brain ischemia and reperfusion may be significantly more than previously estimated because the divalent cation overload traditionally thought to be from Ca²⁺, is actually from Zn²⁺ (10). This interpretation and the observation that [Ca²⁺]_i may reach 30 μm during ischemia (12) together indicate the actual [Zn²⁺]_i may be in the μm range, sufficient to activate Slo1 BK channels, suggesting that Zn²⁺-dependent activation of Slo1 BK channels may play a role during cerebral ischemia. The finding that pharmacological activation of BK channels is cell protective during ischemic stroke (24, 25) indicates that the Zn²⁺-dependent activation of the channel probably represents a compensatory and adaptive response.

In summary, this study demonstrates that hSlo1 BK channels are intracellular Zn²⁺-activated channels and represent a new effector of intracellular Zn²⁺ signaling. The stimulatory effect of Zn²⁺ requires His, Asp and Glu in the RCK1 domain. As a member of the Zn²⁺-signaling cascade, Slo1 BK channels may participate in many phenomena mediated by intracellular Zn²⁺, particularly in some diseases associated with a significant increase in [Zn²⁺]_i.