Regulation of the K channels by cytoplasmic domains

Regulation of intracellular potassium levels is one of the basic functions of all cells, controlling cellular osmolarity and transmitting information. In higher organisms, elaborate control of transmembrane potassium flux has evolved to endow nervous systems with the remarkable ability to transmit electrical signals between cells. Multiple genes, gene splicing, mRNA editing, and selective tetrameric assembly of K channel genes provide the basis for creating distinct electrophysiological properties at varying developmental and cellular stages. This assembly mechanism, primarily governed by the T1 domain, is under the control of intracellular signals. Atomic structures of the isolated T1 domains of Shaker and Shaw subfamilies provided us with valuable structural insights into understanding both channel assembly and functional regulation of the entire channel molecule through conformational changes.     

Dynamic control of deactivation gating by a soluble amino-terminal domain in HERG K(+) channels

K(+) channels encoded by the human ether-à-go-go-related gene (HERG) are distinguished from most other voltage-gated K(+) channels by an unusually slow deactivation process that enables cardiac I(Kr), the corresponding current in ventricular cells, to contribute to the repolarization of the action potential. When the first 16 amino acids are deleted from the amino terminus of HERG, the deactivation rate is much faster (Wang, J., M.C. Trudeau, A.M. Zappia, and G.A. Robertson. 1998. J. Gen. Physiol. 112:637-647). In this study, we determined whether the first 16 amino acids comprise a functional domain capable of slowing deactivation. We also tested whether this "deactivation subdomain" slows deactivation directly by affecting channel open times or indirectly by a blocking mechanism. Using inside-out macropatches excised from Xenopus oocytes, we found that a peptide corresponding to the first 16 amino acids of HERG is sufficient to reconstitute slow deactivation to channels lacking the amino terminus. The peptide acts as a soluble domain in a rapid and readily reversible manner, reflecting a more dynamic regulation of deactivation than the slow modification observed in a previous study with a larger amino-terminal peptide fragment (Morais Cabral, J.H., A. Lee, S.L. Cohen, B.T. Chait, M. Li, and R. Mackinnon. 1998. Cell. 95:649-655). The slowing of deactivation by the peptide occurs in a dose-dependent manner, with a Hill coefficient that implies the cooperative action of at least three peptides per channel. Unlike internal TEA, which slows deactivation indirectly by blocking the channels, the peptide does not reduce current amplitude. Nor does the amino terminus interfere with the blocking effect of TEA, indicating that the amino terminus binding site is spatially distinct from the TEA binding site. Analysis of the single channel activity in cell-attached patches shows that the amino terminus significantly increases channel mean open time with no alteration of the mean closed time or the addition of nonconducting states expected from a pore block mechanism.We propose that the four amino-terminal deactivation subdomains of the tetrameric channel interact with binding sites uncovered by channel opening to specifically stabilize the open state and thus slow channel closing.     

Role of inwardly rectifying K(+) channels in K(+)-induced cerebral vasodilatation in vivo

We tested whether activation of inwardly rectifying K(+) (Kir) channels, Na(+)-K(+)-ATPase, or nitric oxide synthase (NOS) play a role in K(+)-induced dilatation of the rat basilar artery in vivo. When cerebrospinal fluid [K(+)] was elevated from 3 to 5, 10, 15, 20, and 30 mM, a reproducible concentration-dependent vasodilator response was elicited (change in diameter = 9 +/- 1, 27 +/- 4, 35 +/- 4, 43 +/- 12, and 47 +/- 16%, respectively). Responses to K(+) were inhibited by approximately 50% by the Kir channel inhibitor BaCl(2) (30 and 100 microM). In contrast, neither ouabain (1-100 microM, a Na(+)-K(+)-ATPase inhibitor) nor N(G)-nitro-L-arginine (30 microM, a NOS inhibitor) had any effect on K(+)-induced vasodilatation. These concentrations of K(+) also hyperpolarized smooth muscle in isolated segments of basilar artery, and these hyperpolarizations were virtually abolished by 30 microM BaCl(2). RT-PCR experiments confirmed the presence of mRNA for Kir2.1 in the basilar artery. Thus K(+)-induced dilatation of the basilar artery in vivo appears to partly involve hyperpolarization mediated by Kir channel activity and possibly another mechanism that does not involve hyperpolarization, activation of Na(+)-K(+)-ATPase, or NOS.     

Interactions of cations with the cytoplasmic pores of inward rectifier K(+) channels in the closed state

Ion channels gate at membrane-embedded domains by changing their conformation along the ion conduction pathway. Inward rectifier K(+) (Kir) channels possess a unique extramembrane cytoplasmic domain that extends this pathway. However, the relevance and contribution of this domain to ion permeation remain unclear. By qualitative x-ray crystallographic analysis, we found that the pore in the cytoplasmic domain of Kir3.2 binds cations in a valency-dependent manner and does not allow the displacement of Mg(2+) by monovalent cations or spermine. Electrophysiological analyses revealed that the cytoplasmic pore of Kir3.2 selectively binds positively charged molecules and has a higher affinity for Mg(2+) when it has a low probability of being open. The selective blocking of chemical modification of the side chain of pore-facing residues by Mg(2+) indicates that the mode of binding of Mg(2+) is likely to be similar to that observed in the crystal structure. These results indicate that the Kir3.2 crystal structure has a closed conformation with a negative electrostatic field potential at the cytoplasmic pore, the potential of which may be controlled by conformational changes in the cytoplasmic domain to regulate ion diffusion along the pore.     

Saxitoxin is a gating modifier of HERG K+ channels

Potassium (K+) channels mediate numerous electrical events in excitable cells, including cellular membrane potential repolarization. The hERG K+ channel plays an important role in myocardial repolarization, and inhibition of these K+ channels is associated with long QT syndromes that can cause fatal cardiac arrhythmias. In this study, we identify saxitoxin (STX) as a hERG channel modifier and investigate the mechanism using heterologous expression of the recombinant channel in HEK293 cells. In the presence of STX, channels opened slower during strong depolarizations, and they closed much faster upon repolarization, suggesting that toxin-bound channels can still open but are modified, and that STX does not simply block the ion conduction pore. STX decreased hERG K+ currents by stabilizing closed channel states visualized as shifts in the voltage dependence of channel opening to more depolarized membrane potentials. The concentration dependence for steady-state modification as well as the kinetics of onset and recovery indicate that multiple STX molecules bind to the channel. Rapid application of STX revealed an apparent "agonist-like" effect in which K+ currents were transiently increased. The mechanism of this effect was found to be an effect on the channel voltage-inactivation relationship. Because the kinetics of inactivation are rapid relative to activation for this channel, the increase in K+ current appeared quickly and could be subverted by a decrease in K+ currents due to the shift in the voltage-activation relationship at some membrane potentials. The results are consistent with a simple model in which STX binds to the hERG K+ channel at multiple sites and alters the energetics of channel gating by shifting both the voltage-inactivation and voltage-activation processes. The results suggest a novel extracellular mechanism for pharmacological manipulation of this channel through allosteric coupling to channel gating.     

Na(+) interaction with the pore of Shaker B K(+) channels: zero and low K(+) conditions.

The Shaker B K(+) conductance (G(K)) collapses (in a reversible manner) if the membrane is depolarized and then repolarized in, 0 K(+), Na(+)-containing solutions (Gómez-Lagunas, F. 1997. J. Physiol. 499:3-15; Gómez-Lagunas, F. 1999. Biophys. J. 77:2988-2998). In this work, the role of Na(+) ions in the collapse of G(K) in 0-K(+) solutions, and in the behavior of the channels in low K(+) was studied. The main findings are as follows. First, in 0-K(+) solutions, the presence of Na(+) ions is an important factor that speeds the collapse of G(K). Second, external Na(+) fosters the drop of G(K) by binding to a site with a K(d) = 3.3 mM. External K(+) competes, in a mutually exclusive manner, with Na(o)(+) for binding to this site, with an estimated K(d) = 80 microM. Third, NMG and choline are relatively inert regarding the stability of G(K); fourth, with [K(o)(+)] = 0, the energy required to relieve Na(i)(+) block of Shaker (French, R.J., and J.B. Wells. 1977. J. Gen. Physiol. 70:707-724; Starkus, J.G., L. Kuschel, M. Rayner, and S. Heinemann. 2000. J. Gen. Physiol. 110:539-550) decreases with the molar fraction of Na(i)(+) (X(Na,i)), in an extent not accounted for by the change in Delta(mu)(Na). Finally, when X(Na,i) = 1, G(K) collapses by the binding of Na(i)(+) to two sites, with apparent K(d)s of 2 and 14.3 mM.     

Oxidative modification of M-type K(+) channels as a mechanism of cytoprotective neuronal silencing

Voltage-gated K(+) channels of the Kv7 family underlie the neuronal M current that regulates action potential firing. Suppression of M current increases excitability and its enhancement can silence neurons. We here show that three of five Kv7 channels undergo strong enhancement of their activity by oxidative modification induced by physiological concentrations of hydrogen peroxide. A triple cysteine pocket in the channel S2-S3 linker is critical for this effect. Oxidation-induced enhancement of M current produced a hyperpolarization and a dramatic reduction of action potential firing frequency in rat sympathetic neurons. As hydrogen peroxide is robustly produced during hypoxia-induced oxidative stress, we used an oxygen/glucose deprivation neurodegeneration model that showed neuronal death to be severely accelerated by M current blockade. Such blockade had no effect on survival of normoxic neurons. This work describes a novel pathway of M-channel regulation and suggests a role for M channels in protective neuronal silencing during oxidative stress.     

Molecular architecture of a retinal cGMP-gated channel: the arrangement of the cytoplasmic domains

Cyclic nucleotide-gated (CNG) channels play a central role in the conversion of sensory information, such as light and scent, into primary electrical signals. We have purified the CNG channel from bovine retina and have studied it using electron microscopy and image processing. We present the structure of the channel to 35 A resolution. This three-dimensional reconstruction provides insight into the architecture of the protein, suggesting that the cyclic nucleotide-binding domains, which initiate the response to ligand, 'hang' below the pore-forming part of the channel, attached by narrow linkers. The structure also suggests that the four cyclic nucleotide-binding domains present in each channel form two distinct domains, lending structural weight to the suggestion that the four subunits of the CNG channels are arranged as a pair of dimers.     

Membrane-perturbing properties of two Arg-rich paddle domains from voltage-gated sensors in the KvAP and HsapBK K(+) channels

Voltage-gated K(+) channels are gated by displacement of basic residues located in the S4 helix that together with a part of the S3 helix, S3b, forms a "paddle" domain, whose position is altered by changes in the membrane potential modulating the open probability of the channel. Here, interactions between two paddle domains, KvAPp from the K(v) channel from Aeropyrum pernix and HsapBKp from the BK channel from Homo sapiens, and membrane models have been studied by spectroscopy. We show that both paddle domains induce calcein leakage in large unilamellar vesicles, and we suggest that this leakage represents a general thinning of the bilayer, making movement of the whole paddle domain plausible. The fact that HsapBKp induces more leakage than KvAPp may be explained by the presence of a Trp residue in HsapBKp. Trp residues generally promote localization to the hydrophilic-hydrophobic interface and disturb tight packing. In magnetically aligned bicelles, KvAPp increases the level of order along the whole acyl chain, while HsapBKp affects the morphology, also indicating that KvAPp adapts more to the lipid environment. Nuclear magnetic resonance (NMR) relaxation measurements for HsapBKp show that overall the sequence has anisotropic motions. The S4 helix is well-structured with restricted local motion, while the turn between S4 and S3b is more flexible and undergoes slow local motion. Our results indicate that the calcein leakage is related to the flexibility in this turn region. A possibility by which HsapBKp can undergo structural transitions is also shown by relaxation NMR, which may be important for the gating mechanism.     

alpha-helical structural elements within the voltage-sensing domains of a K(+) channel

Voltage-gated K(+) channels are tetramers with each subunit containing six (S1-S6) putative membrane spanning segments. The fifth through sixth transmembrane segments (S5-S6) from each of four subunits assemble to form a central pore domain. A growing body of evidence suggests that the first four segments (S1-S4) comprise a domain-like voltage-sensing structure. While the topology of this region is reasonably well defined, the secondary and tertiary structures of these transmembrane segments are not. To explore the secondary structure of the voltage-sensing domains, we used alanine-scanning mutagenesis through the region encompassing the first four transmembrane segments in the drk1 voltage-gated K(+) channel. We examined the mutation-induced perturbation in gating free energy for periodicity characteristic of alpha-helices. Our results are consistent with at least portions of S1, S2, S3, and S4 adopting alpha-helical secondary structure. In addition, both the S1-S2 and S3-S4 linkers exhibited substantial helical character. The distribution of gating perturbations for S1 and S2 suggest that these two helices interact primarily with two environments. In contrast, the distribution of perturbations for S3 and S4 were more complex, suggesting that the latter two helices make more extensive protein contacts, possibly interfacing directly with the shell of the pore domain.     

Merging functional studies with structures of inward-rectifier K(+) channels

Inwardly rectifying K(+) (Kir) channels have a wide range of functions including the control of neuronal signalling, heart rate, blood flow and insulin release. Because of the physiological importance of these channels, considerable effort has been invested in understanding the structural basis of their physiology. In this review, we use two recent, high-resolution structures as foundations for examining our current understanding of the fundamental functions that are shared by all K(+) channels, such as K(+) selectivity and channel gating, as well as characteristic features of Kir channel family members, such as inward rectification and their regulation by intracellular factors.     

Impairment of HERG K(+) channel function by tumor necrosis factor-alpha: role of reactive oxygen species as a mediator

Congestive heart failure (CHF) is associated with susceptibility to lethal arrhythmias and typically increases levels of tumor necrosis factor-alpha (TNF-alpha) and its receptor, TNFR1. CHF down-regulates rapid delayed-rectifier K(+) current (I(Kr)) and delays cardiac repolarization. We studied the effects of TNF-alpha on cloned HERG K(+) channel (human ether-a-go-go-related gene) in HEK293 cells and native I(Kr) in canine cardiomyocytes with whole-cell patch clamp techniques. TNF-alpha consistently and reversibly decreased HERG current (I(HERG)). Effects of TNF-alpha were concentration-dependent, increased with longer incubation period, and occurred at clinically relevant concentrations. TNF-alpha had similar inhibitory effects on I(Kr) and markedly prolonged action potential duration (APD) in canine cardiomyocytes. Immunoblotting analysis demonstrated that HERG protein level was slightly higher in canine hearts with tachypacing-induced CHF than in healthy hearts, and TNF-alpha slightly increased HERG protein level in CHF but not in healthy hearts. In cells pretreated with the inhibitory anti-TNFR1 antibody, TNF-alpha lost its ability to suppress I(HERG), indicating a requirement of TNFR1 activation for HERG suppression. Vitamin E or MnTBAP (Mn(III) tetrakis(4-benzoic acid) porphyrin chloride), a superoxide dismutase mimic) prevented, whereas the superoxide anion generating system xanthine/xanthine oxidase mimicked, TNF-alpha-induced I(HERG) depression. TNF-alpha caused robust increases in intracellular reactive oxygen species, and vitamin E and MnTBAP abolished the increases, in both HEK293 cells and canine ventricular myocytes. We conclude that the TNF-alpha/TNFR1 system impairs HERG/I(Kr) function mainly by stimulating reactive oxygen species, particularly superoxide anion, but not by altering HERG expression; the effect may contribute to APD prolongation by TNF-alpha and may be a novel mechanism for electrophysiological abnormalities and sudden death in CHF.     

Identification of an extracytoplasmic region of H+,K(+)-ATPase labeled by a K(+)-competitive photoaffinity inhibitor.

The photoaffinity reagent 8-[(4-azidophenyl)-methoxy]-1-tritiomethyl-2, 3-dimethylimidazo-[1,2-alpha]pyridinium iodide ([3H]mDAZIP) has been synthesized and used to photoinactivate and label purified hog gastric H+,K(+)-ATPase. The specific (K(+)-sensitive) components of both photoinactivation and labeling showed dependences on inhibitor concentration consistent with covalent modification at an extracytoplasmic site of reversible K(+)-competitive binding in the dark. The maximum amount of specific labeling (1.2 nmol/mg) was similar to the number of phosphorylation sites measured (1.0 +/- 0.14 nmol/mg). Specific labeling was distributed 76% on the alpha chain, 18% on the beta chain, and 6% on undefined peptides. Various digestions with trypsin, protease V8, and thermolysin were employed to fragment the labeled enzyme. Gasphase sequencing of the radioactive peptides identified the major site of specific labeling to be within a region where only two stretches of amino acids (Leu105 to Ile126 and Leu139 to Phe155, designated H1 and H2, respectively) are predicted to span the membrane. This in turn suggested that the labeling site was located within or close to the proposed loop between them (Gln127 to Asn138). A computer-driven energy minimization protocol yielded a loop structure to which SCH 28080 (the parent structure of [3H]mDAZIP) could be docked. Conversely, modeling of the corresponding region of Na+,K(+)-ATPase (a homologous enzyme with much lower affinity for SCH 28080) yielded no apparent binding site. Similarities in the inhibition of H+,K(+)-ATPase by SCH 28080 and of Na+,K(+)-ATPase by ouabain lead to the hypothesis that, in each case, inhibitor binding to E2-P is associated with an increase in the hydrophobicity of the environment of the loop between H1 and H2.     

In vivo analysis of cohesin architecture using FRET in the budding yeast Saccharomyces cerevisiae

Cohesion between sister chromatids in eukaryotes is mediated by the evolutionarily conserved cohesin complex. Cohesin forms a proteinaceous ring, large enough to trap pairs of replicated sister chromatids. The circumference consists of the Smc1 and Smc3 subunits, while Scc1 is thought to close the ring by bridging the Smc (structural maintenance of chromosomes) ATPase head domains. Little is known about two additional subunits, Scc3 and Pds5, and about possible conformational changes of the complex during the cell cycle. We have employed fluorescence resonance energy transfer (FRET) to analyse interactions within the cohesin complex in live budding yeast. These experiments reveal an unexpected geometry of Scc1 at the Smc heads, and suggest that Pds5 plays a role at the Smc hinge on the opposite side of the ring. Key subunit interactions, including close proximity of the two ATPase heads, are constitutive throughout the cell cycle. This depicts cohesin as a stable molecular machine undergoing only transient conformational changes during binding and dissociation from chromosomes. Using FRET, we did not observe interactions between more than one cohesin complex in vivo.     

Fast inactivation in potassium channels: An interplay of cytoplasmic domains

Fast inactivation in voltage-gated potassium channels has traditionally been associated exclusively with the N-terminus. Here, we explore the role of the T1 domain using a series of chimeric channels. A chimeric channel, 4N/2, (N-terminus from the rapidly inactivating hKv1.4, and the channel body from the non-inactivating hKv1.2), exhibited slower and incomplete inactivation as compared to the wild-type hKv1.4. Replacing the T1 domain of 4N2 with that from hKv1.2 (4N/2T1/2), restored inactivation, while that from hKv1.1 (4N/1T1/2) completely abolished inactivation. Based on these observations, we hypothesize a correlation between the tetramerization domain and the putative inactivation domain receptor in the process of rapid inactivation of hKv1 channels.     

Two N-terminal domains of Kv4 K(+) channels regulate binding to and modulation by KChIP1

The family of calcium binding proteins called KChIPs associates with Kv4 family K(+) channels and modulates their biophysical properties. Here, using mutagenesis and X-ray crystallography, we explore the interaction between Kv4 subunits and KChIP1. Two regions in the Kv4.2 N terminus, residues 7-11 and 71-90, are necessary for KChIP1 modulation and interaction with Kv4.2. When inserted into the Kv1.2 N terminus, residues 71-90 of Kv4.2 are also sufficient to confer association with KChIP1. To provide a structural framework for these data, we solved the crystal structures of Kv4.3N and KChIP1 individually. Taken together with the mutagenesis data, the individual structures suggest that that the Kv4 N terminus is required for stable association with KChIP1, perhaps through a hydrophobic surface interaction, and that residues 71-90 in Kv4 subunits form a contact loop that mediates the specific association of KChIPs with Kv4 subunits.     

The interaction of Na(+) and K(+) in the pore of cyclic nucleotide-gated channels

The permeability ratio between K(+) and Na(+) ions in cyclic nucleotide-gated channels is close to 1, and the single channel conductance has almost the same value in the presence of K(+) or Na(+). Therefore, K(+) and Na(+) ions are thought to permeate with identical properties. In the alpha-subunit from bovine rods there is a loop of three prolines at positions 365 to 367. When proline 365 is mutated to a threonine, a cysteine, or an alanine, mutant channels exhibit a complex interaction between K(+) and Na(+) ions. Indeed K(+), Rb(+) and Cs(+) ions do not carry any significant macroscopic current through mutant channels P365T, P365C and P365A and block the current carried by Na(+) ions. Moreover in mutant P365T the presence of K(+) in the intracellular (or extracellular) medium caused the appearance of a large transient inward (or outward) current carried by Na(+) when the voltage command was quickly stepped to large negative (or positive) membrane voltages. This transient current is caused by a transient potentiation, i.e., an increase of the open probability. The permeation of organic cations through these mutant channels is almost identical to that through the wild type (w.t.) channel. Also in the w.t. channel a similar but smaller transient current is observed, associated to a slowing down of the channel gating evident when intracellular Na(+) is replaced with K(+). As a consequence, a rather simple mechanism can explain the complex behavior here described: when a K(+) ion is occupying the pore there is a profound blockage of the channel and a potentiation of gating immediately after the K(+) ion is driven out. Potentiation occurs because K(+) ions slow down the rate constant K(off) controlling channel closure. These results indicate that K(+) and Na(+) ions do not permeate through CNG channels in the same way and that K(+) ions influence the channel gating.     

Localization of voltage-gated K(+) channels in squid giant axons

We have localized the classical voltage-gated K(+) channel within squid giant axons by immunocytochemistry using the Kv1 antibody of Rosenthal et al. (1996). Widely dispersed patches of intense immunofluorescence were observed in the axonal membrane. Punctate immunofluorescence was also observed in the axoplasm and was localized to approximately 25-50-microm-wide column down the length of the nerve (axon diameter approximately 500 microm). Immunoelectronmicroscopy of the axoplasm revealed a K(+) channel containing vesicles, 30-50 nm in diameter, within this column. These and other vesicles of similar size were isolated from axoplasm using a novel combination of high-speed ultracentrifugation and controlled-pore size, glass bead separation column techniques. Approximately 1% of all isolated vesicles were labeled by K(+) channel immunogold reacted antibody. Incorporation of isolated vesicle fractions within an artificial lipid bilayer revealed K(+) channel electrical activity similar to that recorded directly from the axonal membrane by Llano et al. (1988). These K(+) channel-containing vesicles may be involved in cycling of K(+) channel protein into the axonal membrane. We have also isolated an axoplasmic fraction containing approximately 150-nm-diameter vesicles that may transport K(+) channels back to the cell body.