Ma Z , Wong KY , Horrigan FT .

NS42901 NINDS NIH HHS , R01 NS042901 NINDS NIH HHS

	Figure 1. Extracellular Cu2+ inhibits mSlo1 activation. (A) Ik evoked in response to a pulse to +240 mV from a holding potential of −80 mV (0 Ca2+). Currents in 0 Cu2+ (control), 100 μM Cu2+, and washout with 5 mM EGTA were recorded from the same patch. Tail currents are shown on an expanded time scale, and have their own time scale bar. (B) The time course of inhibition by 100 μM Cu2+ was measured with a double pulse protocol. The ratio of steady-state IK during two 30-ms test pulses to +220 mV is plotted against the interval between pulses (t). Cu2+ was applied at −80 mV with rapid perfusion immediately following the first test pulse (t = 0). IK is inhibited rapidly (τ = 0.2 s, dotted curve) following a delay of ∼0.1 s likely representing the time for Cu2+ to reach the patch. (C) Single channel IK-V relation is unaffected by 100 μM Cu2+ (top). Representative traces at +160 mV from a single channel patch show a marked decrease in PO (bottom). (D) Normalized GK-V relations determined in 0 Cu2+ (•), 100 μM Cu2+ (▵), and 1000 μM Cu2+ (▪) for the same patch are fit by Boltzmann functions (0 Ca: GKmax = 1, V0.5 = 218 mV, Zapp = 0.82 e; 1000 Cu2+: GKmax = 0.81, V0.5 = 300 mV, Zapp = 0.57 e). (E) Normalized outward IK at +220 mV and (F) tail currents at −80 mV in 0 Cu2+ and 100 μM Cu2+ are fit by exponential functions. 100 μM Cu2+ increased the activation time constant 1.6-fold from 1.25 to 1.98 ms and decreased the deactivation time constant 1.4-fold from 153 to 108 μs.
	Figure 2. mSlo1 channels are extremely Cu2+ sensitive. (A) Mean GK-V relations with [Ca2+]i = 3 μM at different [Cu2+]: 0 (○), 0.1 (•), 0.2 (□), 0.5(▴), 1.0 (▵), 2.5 (▪), 5 (▿), 10 (▾), 20 (⋄), 50 (♦), 100 () and 200 μM (⋄), are normalized by GKmax in 0 Cu2+ and fit by Boltzmann functions. (B) Mean GK-V relations from panel A are normalized by GKmax at each [Cu2+] based on Boltzmann fits. (C) Cu2+ dose–response relations for ΔV0.5 = (V0.5 [Cu2+] − V0.5[0]) (•) and GK[Cu2+] at 110 mV (○), 150 mV (▵) and 190 mV (▿) were determined from fits in A. Solid lines are fits to a Hill equation. Thick dotted curve is a fit to Eq. 1 with most parameters set to previously determined values (zJ = 0.58 e, L0 = 10−6, zL = 0.3 e, KD(Ca2+) = 11 μM, C = 8, D = 25, E = 2.4) (Horrigan and Aldrich, 2002). VhC = 162 mV was adjusted to fit V0.5 of the 0 Cu2+ control, and Cu2+-dependent parameters (KDcu = 0.75 μM, Ecu = 0.124) were then varied to fit the V0.5-[Cu2+] relation. Thin dotted curve is a fit to Eq. 2 with KDcu = 2.1 μM, Ecu = 0.129, and all other parameters the same as for Eq. 1. (D) IC50 and (E) nH obtained by fitting GK-[Cu2+] relations in 3 μM Ca2+ (•) and 0 Ca2+ (□) are plotted at different voltages. Dashed lines are IC50 = 1.97 μM, nH = 1.01 from the ΔV0.5-[Cu2+] relation fit in panel C. (F) Apparent charge (Zapp) from Boltzmann GK-V fits in A are plotted versus [Cu2+] and fit by a Hill equation (IC50 = 0.72 ± 0.12 μM, nH = 1.1 ± 0.2). Curves A and B are the predictions of Eqs. 1 and 2, respectively, using parameters from C.
	Figure 3. The mechanism and state dependence of Cu2+ action. (A) Mean log(Po)-V relations for mSlo1 for 0 and 5 μM [Ca2+]i in 0 Cu2+(○) and 100 μM Cu2+(▪), respectively. Dotted lines are exponential fits to the limiting slope of log(Po) with partial charge zL = 0.3 e. (B) Log(Po)-V relations from A are superimposed by shifting the 0 Cu2+ curves along the voltage axis by +45 mV. (C) IK evoked by 30-ms test pulses to +120 mV before (IP1) and after (IP2) application of 100 μM Cu2+ using the illustrated protocol (5 μM [Ca2+]i), without leak subtraction. IP1 was evoked from a holding potential of −80 mV in 0 Cu2+ following perfusion with standard external solution for 5 s. IP2 was recorded in 100 μM Cu2+ following perfusion with 100 μM Cu2+ for 500 ms at different prepulse voltages (VPRE). Following the second pulse the patch was washed at −80 mV for 5 s with 5 mM EGTA and then for 5 s with standard external solution before repeating the protocol. Maximal inhibition of IP2 was observed with VPRE = −80 to +40 mV (green traces) and minimal inhibition with VPRE = +180 to +220 mV (red traces). (D) The fraction of channels not inhibited fNI = (IP2[VPRE] − IP2[−80])/(IP1 − IP2[−80]) from C (▪, fNI[100 Cu2+]) is plotted against VPRE and compared with the mean steady-state Po-V relation (○) in 0 Cu2+ and 5 μM [Ca2+]i estimated as GK/GKmax. A control experiment (□, fNI[0 Cu2+]) obtained from a different patch using the pulse protocol in C shows that fNI is voltage independent when the 100 μM Cu2+ solution is replaced with 0 Cu2+.
	Figure 4. Cu2+ selectively inhibits mSlo1 activation. (A) GK-V shift (ΔV0.5) produced by 100 μM of different transition metals (Cd2+, Mn2+, Fe3+, Co2+, Ni2+, Cu2+, Zn2+) (0 [Ca2+]i). Stars indicate a significant change in V0.5 relative to the control (P < 0.05, paired t test). (B) GK-V relations in the absence of metal ions (○), 100 μM Cu2+ (•), 100 μM Cd2+ (▪), or 4,000 μM Cd2+ (□) were normalized by the control and fit with Boltzmann functions (10 μM [Ca2+]i). (C) Dose–response relations for inhibition of GK[+220 mV] by different metal ions (0 [Ca2+]i) fit with Hill equations (Cu2+ (○) IC50 = 2.0 μM, nH = 0.98; Cd2+ (•) IC50 = 0.70 mM, nH = 1.0; Zn2+ (□) IC50 = 1.2 mM, nH = 0.87; Mn2+ (▪) IC50 = 6.7 mM, nH = 1.4; Ni 2+ (▵) IC50 = 7.8 mM, nH = 1.5).
	Figure 5. Cu2+ action is independent of native Cys and His residues but is pH dependent. (A) GK-V relations for WT (○: 0 Cu2+; •: 100 Cu2+) and C14A/C141A/C277A (“C3A”, ▵: 0 Cu2+; ▴: 100 Cu2+) show similar inhibition by 100 μM Cu2+. (B) GK-V relations for WT (○: 0 Cu2+; •: 100 Cu2+) and H254A (<: 0 Cu2+; ▾: 100 Cu2+) also respond similarly to 100 μM Cu2+ as indicated by the shift in V0.5. (C) Mean dose–response relations (GK-[Cu]2+) measured at voltages near V0.5 for WT (○), C14A/C141A/C277A (▵), and H254A (▾) are indistinguishable. The solid line is a fit by the Hill equation (IC50 = 2.0 μM, nH = 0.98) (D) Dependence of Cu2+ action on extracellular pH (pHO) is fit by Hill equation (pH0.5 = 6.0, nH = 1.15). GK-pHO relation for WT channels at +220 mV in 0 Ca2+ represents GK in 100 μM Cu2+ normalized by GK in 0 Cu2+ at each pHO.
	Figure 6. Acidic residues contribute to metal sensitivity of mSlo1 and other voltage-gated channels. (A) Multiple sequence alignment of voltage sensor domain from mSlo1, Kv channels (Shaker, Kv1.2, and Kv2.1), dEAG, and hERG, and other voltage-gated channels (Nav1.4, Nav1.5, Cav1.4, Cav3.2 Domain IV) (based on Liu et al., 2003; Long et al., 2007). Putative transmembrane segments are indicated by solid bars (Long et al., 2007). Charged residues that are highly conserved in mSlo1 and other voltage-gated channels are in red (acidic) or blue (basic). Sites that were mutated in mSlo1 are highlighted yellow with positions labeled. Putative metal coordinating residues in mSlo1, dERG, and hERG are indicated by boxes. (B) Dose–response relations (GK at +260 mV) fit by Hill equations for D153K (▪, IC50 = 16.7 μM, nH = 1.0) and D153H (▴, IC50 = 0.26 μM, nH = 0.73) (5 μM [Ca2+]i). Dashed curve represents fit to WT data from Fig. 5 C. (C) The dose–response relation (GK at +160 mV) for D186A (•, IC50 = 2.53 μM, nH = 0.97), D186H (▴, IC50 = 2.82 μM, nH = 0.95) (1 μM [Ca2+]i). (D) The dose–response relations (GK at +240 mV) for D133A (•, IC50 = 56.5 μM, nH = 0.97), D133C (<, IC50 = 1.5 μM, nH = 0.74) and D133H (▴, IC50 = 0.73 μM, nH = 0.63) (0 [Ca2+]i).
	Figure 7. Nonacidic residues also contribute to Cu2+ sensitivity. (A) IC50s of mutant channels are plotted at each position tested for Cu2+ (open symbols) or Cd2+ (filled symbols). Error bars were within the symbol size and therefore excluded. Symbols indicate different substitutions (WT: , ; Ala: ○, •; Lys: □; Gln: ⋄; Asn: ⋄; Cys: ▿,▾; His: ▵, ▴). Dashed line indicates the IC50 of the WT for Cu2+. Vertical solid lines indicate the differences in IC50 between mutants at putative coordination sites. Dose–response relations (GK-[Cu2+]) for putative coordination sites 151 and 207 in 0 Ca2+ are plotted and fit by Hill equations in (B) Q151A (○, IC50 = 23.8 μM, nH =1.0), Q151C (<, IC50 = 0.96 μM, nH = 0.97) at +240 mV and (C) R207Q (⋄, IC50 = 85.8 μM, nH = 0.77) and R207C (<, IC50 = 0.52 μM, nH = 0.77) at +180 mV.
	Figure 8. The role of R207 in Cu2+ sensitivity. (A) GK-V relations fit by Boltzmann functions for R207C (⋄) and R207Q (▿) in 0 Cu2+ (V0.5 = 149 mV, zAPP = 0.64 e) and in 100 μM Cu2+ (R207C, ▾) and 1000 μM Cu2+ (R207C: ○, V0.5 = 314.5 mV, zAPP = 0.64 e; R207Q: ♦, V0.5 = 245.6 mV, zAPP = 0.46 e). (B) Log(Po)-V relation for R207C in 0 (<) and 100 μM Cu2+ (▾). (C) Cd2+ dose–response relations (GK-[Cd2+]) fit by Hill equations for WT (•, IC50 = 720 μM, nH = 0.98) at +220 mV and R207C (▴, IC50 = 0.038 μM, nH = 0.73) at +180 mV (0 [Ca2+]i). (D) GK-V relations for R207C in 0 Cd2+ (<, V0.5 = 149 mV, zAPP = 0.64 e), 100 μM Cd2+ (▾), and 1000 μM Cd2+ (○, V0.5 = 316 mV, zAPP = 0.61 e) (0 [Ca2+]i).
	Figure 9. Effect of mutations on Cu2+-efficacy. (A) The maximal GK-V shift (ΔV0.5MAX) in response to saturating 1000 μM Cu2+ are plotted for the WT () and mutants at the four putative Cu2+-coordinating sites (Ala: •; Lys: ▪; Gln: ⋄; Cys: ▿; His: ▴). The dashed line indicates ΔV0.5MAX for the WT. The vertical solid lines indicate the range of ΔV0.5MAX for mutants at each position. (B) A speculative model of the Cu2+ binding site in the resting (R) and activated (A) state of the voltage sensor. In the resting state, Cu2+ is coordinated by D133(S1), Q151(S2), D153(S2), and R207(S4). The relative size and position of transmembrane segments correspond to those in the structure of a Kv1.2/Kv2.1 chimera (Long et al., 2007). During activation, the S2 residues interact with Cu2+ in a state-independent manner while interactions with D133 and R207 are weakened or disrupted, presumably by changes in the position or orientation of S1 and S4 relative to S2. In this way, mutation of any of these residues alter IC50, but only mutation of D133 or R207 alter efficacy.
	Figure 10. The Cu2+ coordination sites are conserved in Shaker channels. (A) IK evoked from WT Shaker channels in response to pulses to −10 mV from a holding potential of −90 mV at the indicated [Cu2+] (0–100 μM). (B) GK-V relations from the same patch in different [Cu2+] are normalized by GKmax in 0 Cu2+. (0 (○), 1 (•), 5 (▾), 20 (▴), and 100 μM Cu2+ (▪)). (C) IK at 0 mV from WT channels at the indicated [Zn2+] (0–10,000 μM). (D) GK-V relations in different [Zn2+], normalized by GKmax in 0 Zn2+ (0 (○), 100 (•), 1,000 (▾), 4,000 (▴), 10,000 μM Zn2+ ()). (E) IK at 0 mV from E247A channels in 0, 1, and 5 μM Cu2+. (F) IK at +30 mV from E247C in 0, 0.1, 1, and 5 μM Cu2+. (G and H) IK at +10 mV from E247A (G) or E247C (H) in 0, 100, and 1,000 μM Zn2+. (I and J) IK from R365A at 0 mV (I) or R365C at −40 mV(J) in 0, 1, and 5 μM Cu2+. All Shaker experiments were performed from a holding potential of −90 mV. Most data were obtained in standard Shaker internal and external solutions (see Materials and methods) with the exception of R365 mutant data in I and J, which were obtained in a low chloride external solution with all but 10 mM Cl− replaced by MeO3−, as in standard mSlo1 external solution.
	Figure 11. Mechanism and site of Cu2+ action. (A) Scheme 1 (see text), green indicates Cu2+-dependent mechanisms, black represents HA model (Horrigan and Aldrich, 2002). (B and C) Residues mutated in mSlo1 are mapped onto the structure of a Kv1.2/Kv2.1 chimera (Long et al., 2007). B and C show side and top views, respectively. Only the S1–S4 segments from one subunit are shown with S1–S2 linker removed for clarity. Putative Cu2+-coordinating residues were changed to the corresponding residue from mSlo1: D133(red), Q151(yellow), D153(red), R207(cyan). The backbone was colored dark blue at other positions that were tested. Dotted spheres indicate the position of water molecules in the structure. The green sphere illustrates a possible position for Cu2+.

Armstrong, Charge movement associated with the opening and closing of the activation gates of the Na channels. 1974, Pubmed

Armstrong, Charge movement associated with the opening and closing of the activation gates of the Na channels. 1974, Pubmed
Assaf, Release of endogenous Zn2+ from brain tissue during activity. , Pubmed
Berkovitch, Crystal structure of biotin synthase, an S-adenosylmethionine-dependent radical enzyme. 2004, Pubmed
Bewley, Crystal structures of Bacillus caldovelox arginase in complex with substrate and inhibitors reveal new insights into activation, inhibition and catalysis in the arginase superfamily. 1999, Pubmed
Boland, Cysteines in the Shaker K+ channel are not essential for channel activity or zinc modulation. 1994, Pubmed , Xenbase
Bush, Metals and neuroscience. 2000, Pubmed
Butler, mSlo, a complex mouse gene encoding "maxi" calcium-activated potassium channels. 1993, Pubmed , Xenbase
Campos, Two atomic constraints unambiguously position the S4 segment relative to S1 and S2 segments in the closed state of Shaker K channel. 2007, Pubmed , Xenbase
Castelli, Cu2+, Co2+, and Mn2+ modify the gating kinetics of high-voltage-activated Ca2+ channels in rat palaeocortical neurons. 2003, Pubmed
Colquhoun, Binding, gating, affinity and efficacy: the interpretation of structure-activity relationships for agonists and of the effects of mutating receptors. 1998, Pubmed
Cox, Allosteric gating of a large conductance Ca-activated K+ channel. 1997, Pubmed , Xenbase
Cui, Allosteric linkage between voltage and Ca(2+)-dependent activation of BK-type mslo1 K(+) channels. 2000, Pubmed
Di Costanzo, Stereochemistry of guanidine-metal interactions: implications for L-arginine-metal interactions in protein structure and function. 2006, Pubmed
Diaz, Interaction of internal Ba2+ with a cloned Ca(2+)-dependent K+ (hslo) channel from smooth muscle. 1996, Pubmed , Xenbase
Elinder, Metal ion effects on ion channel gating. 2003, Pubmed
Faber, Calcium-activated potassium channels: multiple contributions to neuronal function. 2003, Pubmed
Ferguson, Competitive Mg2+ block of a large-conductance, Ca(2+)-activated K+ channel in rat skeletal muscle. Ca2+, Sr2+, and Ni2+ also block. 1991, Pubmed
Fernandez, Molecular mapping of a site for Cd2+-induced modification of human ether-à-go-go-related gene (hERG) channel activation. 2005, Pubmed , Xenbase
Ferraroni, Crystal structure of a zinc-activated variant of human carbonic anhydrase I, CA I Michigan 1: evidence for a second zinc binding site involving arginine coordination. 2002, Pubmed
Forsyth, Empirical relationships between protein structure and carboxyl pKa values in proteins. 2002, Pubmed
Frederickson, The neurobiology of zinc in health and disease. 2005, Pubmed
Frederickson, Synaptic release of zinc from brain slices: factors governing release, imaging, and accurate calculation of concentration. 2006, Pubmed
Gilly, Slowing of sodium channel opening kinetics in squid axon by extracellular zinc. 1982, Pubmed
Gilly, Divalent cations and the activation kinetics of potassium channels in squid giant axons. 1982, Pubmed
Gordon, Subunit interactions in coordination of Ni2+ in cyclic nucleotide-gated channels. 1995, Pubmed , Xenbase
Gruss, The two-pore-domain K(+) channels TREK-1 and TASK-3 are differentially modulated by copper and zinc. 2004, Pubmed
Hamill, Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. 1981, Pubmed
Hartter, Evidence for release of copper in the brain: depolarization-induced release of newly taken-up 67copper. 1988, Pubmed
Holmgren, The activation gate of a voltage-gated K+ channel can be trapped in the open state by an intersubunit metal bridge. 1998, Pubmed
Hopt, Methods for studying synaptosomal copper release. 2003, Pubmed
Horrigan, Allosteric voltage gating of potassium channels I. Mslo ionic currents in the absence of Ca(2+). 1999, Pubmed , Xenbase
Horrigan, Coupling between voltage sensor activation, Ca2+ binding and channel opening in large conductance (BK) potassium channels. 2002, Pubmed , Xenbase
Horrigan, Heme regulates allosteric activation of the Slo1 BK channel. 2005, Pubmed
Horrigan, Allosteric voltage gating of potassium channels II. Mslo channel gating charge movement in the absence of Ca(2+). 1999, Pubmed
Horrigan, Mg2+ enhances voltage sensor/gate coupling in BK channels. 2008, Pubmed , Xenbase
Howell, Stimulation-induced uptake and release of zinc in hippocampal slices. , Pubmed
Hu, Effects of multiple metal binding sites on calcium and magnesium-dependent activation of BK channels. 2006, Pubmed , Xenbase
Huidobro-Toro, Trace metals in the brain: allosteric modulators of ligand-gated receptor channels, the case of ATP-gated P2X receptors. 2008, Pubmed
Jeong, Divalent metals differentially block cloned T-type calcium channels. 2003, Pubmed
Jiang, X-ray structure of a voltage-dependent K+ channel. 2003, Pubmed
Kang, A molecular determinant of nickel inhibition in Cav3.2 T-type calcium channels. 2006, Pubmed , Xenbase
Kardos, Nerve endings from rat brain tissue release copper upon depolarization. A possible role in regulating neuronal excitability. 1989, Pubmed
Kehl, Molecular determinants of the inhibition of human Kv1.5 potassium currents by external protons and Zn(2+). 2002, Pubmed
Kiss, Toxic effects of heavy metals on ionic channels. 1994, Pubmed
Kozma, Histochemical detection of zinc and copper in various neurons of the central nervous system. 1981, Pubmed
Li, Rapid translocation of Zn(2+) from presynaptic terminals into postsynaptic hippocampal neurons after physiological stimulation. 2001, Pubmed
Linder, Copper biochemistry and molecular biology. 1996, Pubmed
Liu, Negative charges in the transmembrane domains of the HERG K channel are involved in the activation- and deactivation-gating processes. 2003, Pubmed , Xenbase
Long, Voltage sensor of Kv1.2: structural basis of electromechanical coupling. 2005, Pubmed
Long, Atomic structure of a voltage-dependent K+ channel in a lipid membrane-like environment. 2007, Pubmed
Long, Crystal structure of a mammalian voltage-dependent Shaker family K+ channel. 2005, Pubmed
Ma, Role of charged residues in the S1-S4 voltage sensor of BK channels. 2006, Pubmed , Xenbase
MacKinnon, Role of surface electrostatics in the operation of a high-conductance Ca2+-activated K+ channel. 1989, Pubmed
Mash, Complexation of copper by zwitterionic aminosulfonic (good) buffers. 2003, Pubmed
Mathie, Zinc and copper: pharmacological probes and endogenous modulators of neuronal excitability. 2006, Pubmed
Misonou, Immunolocalization of the Ca2+-activated K+ channel Slo1 in axons and nerve terminals of mammalian brain and cultured neurons. 2006, Pubmed
Morera, External copper inhibits the activity of the large-conductance calcium- and voltage-sensitive potassium channel from skeletal muscle. 2003, Pubmed
Nelson, Molecular mechanisms of subtype-specific inhibition of neuronal T-type calcium channels by ascorbate. 2007, Pubmed
Neyton, A Ba2+ chelator suppresses long shut events in fully activated high-conductance Ca(2+)-dependent K+ channels. 1996, Pubmed
Oberhauser, Activation by divalent cations of a Ca2+-activated K+ channel from skeletal muscle membrane. 1988, Pubmed
Ono, Regional distribution of metallothionein, zinc, and copper in the brain of different strains of rats. 1999, Pubmed
Osterberg, Physiology and pharmacology of copper. 1980, Pubmed
Pathak, Closing in on the resting state of the Shaker K(+) channel. 2007, Pubmed
Pivovarov, Modeling of ionic equilibria of trace metals (Cu2+, Zn2+, Cd2+) in concentrated aqueous electrolyte solutions at 25 degrees C. 2005, Pubmed
Rothberg, Kinetic structure of large-conductance Ca2+-activated K+ channels suggests that the gating includes transitions through intermediate or secondary states. A mechanism for flickers. 1998, Pubmed
Rulísek, Coordination geometries of selected transition metal ions (Co2+, Ni2+, Cu2+, Zn2+, Cd2+, and Hg2+) in metalloproteins. 1998, Pubmed
Satin, A mutant of TTX-resistant cardiac sodium channels with TTX-sensitive properties. 1992, Pubmed , Xenbase
Sato, Localization of copper to afferent terminals in rat locus ceruleus, in contrast to mitochondrial copper in cerebellum. 1994, Pubmed
Schlief, Copper homeostasis in the CNS: a novel link between the NMDA receptor and copper homeostasis in the hippocampus. 2006, Pubmed
Silverman, Mg(2+) modulates voltage-dependent activation in ether-à-go-go potassium channels by binding between transmembrane segments S2 and S3. 2000, Pubmed , Xenbase
Silverman, Binding site in eag voltage sensor accommodates a variety of ions and is accessible in closed channel. 2004, Pubmed , Xenbase
Slomianka, Labeling of the neurons of origin of zinc-containing pathways by intraperitoneal injections of sodium selenite. 1990, Pubmed
Sokołowska, Cu(II) complexation by "non-coordinating" N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES buffer). 2005, Pubmed
Takeda, Movement of zinc and its functional significance in the brain. 2000, Pubmed
Teisseyre, The inhibitory effect of copper ions on lymphocyte Kv1.3 potassium channels. 2006, Pubmed
Thompson, Fluorescent zinc indicators for neurobiology. 2002, Pubmed
Webster, Intracellular gate opening in Shaker K+ channels defined by high-affinity metal bridges. 2004, Pubmed
Yang, Mg2+ mediates interaction between the voltage sensor and cytosolic domain to activate BK channels. 2007, Pubmed
Yellen, An engineered cysteine in the external mouth of a K+ channel allows inactivation to be modulated by metal binding. 1994, Pubmed