Adjacent pore-lining residues within sodium channels identified by paired cysteine mutagenesis.

The pores of voltage-gated ion channels are lined by protein loops that determine selectivity and conductance. The relative orientations of these "P" loops remain uncertain, as do the distances between them. Using site-directed mutagenesis, we introduced pairs of cysteines into the P loops of micro1 rat skeletal muscle sodium channels and sought functional evidence of proximity between the substituted residues. Only cysteinyl residues that are in close proximity can form disulfide bonds or metal-chelating sites. The mutant Y401C (domain I) spontaneously formed a disulfide bond when paired with E758C in the P loop of domain II; the same residue, when coupled with G1530C in domain IV, created a high-affinity binding site for Cd2+ ions. The results provide the first specific constraints for intramolecular dimensions of the sodium channel pore.

Receptor stimulation causes slow inhibition of IRK1 inwardly rectifying K+ channels by direct protein kinase A-mediated phosphorylation.

Strongly rectifying IRK-type inwardly rectifying K+ channels are involved in the control of neuronal excitability in the mammalian brain. Whole-cell patch-clamp experiments show that cloned rat IRK1 (Kir 2.1) channels, when heterologously expressed in mammalian COS-7 cells, are inhibited following the activation of coexpressed serotonin (5-hydroxytryptamine) type 1A receptors by receptor agonists. Inhibition is mimicked by internal perfusion with GTP[gamma-S] and elevation of internal cAMP concentrations. Addition of the catalytic subunits of protein kinase A (PKA) to the internal recording solution causes complete inhibition of wild-type IRK1 channels, but not of mutant IRK1(S425N) channels in which a C-terminal PKA phosphorylation site has been removed. Our data suggest that in the nervous system serotonin may negatively control IRK1 channel activity by direct PKA-mediated phosphorylation.

Voltage gating and permeation in a gap junction hemichannel.

Gap junction channels are formed by members of the connexin gene family and mediate direct intercellular communication through linked hemichannels (connexons) from each of two adjacent cells. While for most connexins, the hemichannels appear to require an apposing hemichannel to open, macroscopic currents obtained from Xenopus oocytes expressing rat Cx46 suggested that some hemichannels can be readily opened by membrane depolarization [Paul, D. L., Ebihara, L., Takemoto, L. J., Swenson, K. I. & Goodenough, D. A. (1991), J. Cell Biol. 115, 1077-1089]. Here we demonstrate by single channel recording that hemichannels comprised of rat Cx46 exhibit complex voltage gating consistent with there being two distinct gating mechanisms. One mechanism partially closes Cx46 hemichannels from a fully open state, gammaopen, to a substate, gammasub, about one-third of the conductance of gammaopen; these transitions occur when the cell is depolarized to inside positive voltages, consistent with gating by transjunctional voltage in Cx46 gap junctions. The other gating mechanism closes Cx46 hemichannels to a fully closed state, gammaclosed, on hyperpolarization to inside negative voltages and has unusual characteristics; transitions between gammaclosed and gammaopen appear slow (10-20 ms), often involving several transient substates distinct from gammasub. The polarity of activation and kinetics of this latter form of gating indicate that it is the mechanism by which these hemichannels open in the cell surface membrane when unapposed by another hemichannel. Cx46 hemichannels display a substantial preference for cations over anions, yet have a large unitary conductance (approximately 300 pS) and a relatively large pore as inferred from permeability to tetraethylammonium (approximately 8.5 angstroms diameter). These hemichannels open at physiological voltages and could induce substantial cation fluxes in cells expressing Cx46.

Spatial localization of calcium channels in giant fiber lobe neurons of the squid (Loligo opalescens).

Whole-cell voltage clamp was used to investigate the properties and spatial distribution of fast-deactivating (FD) Ca channels in squid giant fiber lobe (GFL) neurons. Squid FD Ca channels are reversibly blocked by the spider toxin omega-Agatoxin IVA with an IC50 of 240-420 nM with no effect on the kinetics of Ca channel gating. Channels with very similar properties are expressed in both somatic and axonal domains of cultured GFL neurons, but FD Ca channel conductance density is higher in axonal bulbs than in cell bodies at all times in culture. Channels presumably synthesized during culture are preferentially expressed in the growing bulbs, but bulbar Ca conductance density remains constant while Na conductance density increases, suggesting that processes determining the densities of Ca and Na channels in this extrasomatic domain are largely independent. These observations suggest that growing axonal bulbs in cultured GFL neurons are not composed entirely of "axonal" membranes because FD Ca channels are absent from the giant axon in situ but, rather, suggest a potential role for FD Ca channels in mediating neurotransmitter release at the motor terminals of the giant axon.

N-methyl-D-aspartate receptor-induced proteolytic conversion of postsynaptic class C L-type calcium channels in hippocampal neurons.

Ca2+ influx controls multiple neuronal functions including neurotransmitter release, protein phosphorylation, gene expression, and synaptic plasticity. Brain L-type Ca2+ channels, which contain either alpha 1C or alpha 1D as their pore-forming subunits, are an important source of calcium entry into neurons. Alpha 1C exists in long and short forms, which are differentially phosphorylated, and C-terminal truncation of alpha 1C increases its activity approximately 4-fold in heterologous expression systems. Although most L-type calcium channels in brain are localized in the cell body and proximal dendrites, alpha 1C subunits in the hippocampus are also present in clusters along the dendrites of neurons. Examination by electron microscopy shows that these clusters of alpha 1C are localized in the postsynaptic membrane of excitatory synapses, which are known to contain glutamate receptors. Activation of N-methyl-D-aspartate (NMDA)-specific glutamate receptors induced the conversion of the long form of alpha 1C into the short form by proteolytic removal of the C terminus. Other classes of Ca2+ channel alpha1 subunits were unaffected. This proteolytic processing reaction required extracellular calcium and was blocked by inhibitors of the calcium-activated protease calpain, indicating that calcium entry through NMDA receptors activated proteolysis of alpha1C by calpain. Purified calpain catalyzed conversion of the long form of immunopurified alpha 1C to the short form in vitro, consistent with the hypothesis that calpain is responsible for processing of alpha 1C in hippocampal neurons. Our results suggest that NMDA receptor-induced processing of the postsynaptic class C L-type Ca2+ channel may persistently increase Ca2+ influx following intense synaptic activity and may influence Ca2+-dependent processes such as protein phosphorylation, synaptic plasticity, and gene expression.

Evidence that free polyunsaturated fatty acids modify Na+ channels by directly binding to the channel proteins.

The effects of free polyunsaturated fatty acids (PUFA) on the binding of ligands to receptors on voltage-sensitive Na+ channels of neonatal rat cardiac myocytes were assessed. The radioligand was [benzoyl-2,5-(3)H] batrachotoxinin A 20alpha-benzoate ([(3)H]BTXB), a toxin that binds to the Na+ channel. The PUFA that have been shown to be antiarrhythmic, including eicosapentaenoic acid (EPA; C20:5n-3), docosahexaenoic acid (DHA; C22:6n-3), eicosatetraynoic acid (ETYA), linolenic acid (C18:3n-3), and linoleic acid (C18:2n-6), inhibited [(3)H]BTXB binding in a dose-dependent fashion with IC50 values of 28-35 microM, whereas those fatty acids that have no antiarrhythmic effects including saturated fatty acid (stearic acid, C18:0), monounsaturated fatty acid (oleic acid; C18:1n-9), and EPA methyl ester did not have a significant effect on [(3)H]BTXB binding. Enrichment of the myocyte membrane with cholesterol neither affected [(3)H]BTXB binding when compared with control cells nor altered the inhibitory effects of PUFA on [(3)H]BTXB binding. Scatchard analysis of [(3)H]BTXB binding showed that EPA reduced the maximal binding without altering the Kd for [(3)H]BTXB binding, indicating allosteric inhibition. The inhibition by EPA of [(3)H]BTXB binding was reversible (within 30 min) when delipidated bovine serum albumin was added. The binding of the PUFA to this site on the Na+ channel is reversible and structure-specific and occurs at concentrations close to those required for apparent antiarrhythmic effects and a blocking effect on the Na+ current, suggesting that binding of the PUFA at this site relates to their antiarrhythmic action.

P2X4: an ATP-activated ionotropic receptor cloned from rat brain.

Extracellular ATP exerts pronounced biological actions in virtually every organ or tissue that has been studied. In the central and peripheral nervous system, ATP acts as a fast excitatory transmitter in certain synaptic pathways [Evans, R.J., Derkach, V. & Surprenant, A. (1992) Nature (London) 357, 503-505; Edwards, F.A., Gigg, A.J. & Colquhoun, D. (1992) Nature (London) 359, 144-147]. Here, we report the cloning and characterization of complementary DNA from rat brain, encoding an additional member (P2X4) of the emerging multigenic family of ligand-gated ATP channels, the P2X receptors. Expression in Xenopus oocytes gives an ATP-activated cation-selective channel that is highly permeable to Ca2+ and whose sensitivity is modulated by extracellular Zn2+. Surprisingly, the current elicited by ATP is almost insensitive to the common P2X antagonist suramin. In situ hybridization reveals the expression of P2X4 mRNA in central nervous system neurons. Northern blot and reverse transcription-PCR (RT-PCR) analysis demonstrate a wide distribution of P2X4 transcripts in various tissues, including blood vessels and leukocytes. This suggests that the P2X4 receptor might mediate not only ATP-dependent synaptic transmission in the central nervous system but also a wide repertoire of biological responses in diverse tissues.

An ATP-activated, ligand-gated ion channel on a cholinergic presynaptic nerve terminal.

ATP has recently been identified as a fast neurotransmitter in both the central and peripheral nervous systems. Several studies have suggested that ATP can also affect the release of classical neurotransmitters, including acetylcholine with which it is co-released. We have searched for ATP receptors on a cholinergic presynaptic nerve terminal using the calyx-type synapse of the chicken ciliary ganglion. ATP was pulsed onto the terminals under voltage clamp and induced a short latency cation current that exhibited inward rectification and marked desensitization. This current was not seen with adenosine but was mimicked by several sterically restricted ATP analogs and was blocked by suramin. ATP-activated single ion channels exhibited prominent flickering and had a conductance of approximately 17 pS. Our results demonstrate a ligand-gated P2X-like purinergic receptor on a cholinergic presynaptic nerve terminal.

Determinants of the G protein-dependent opioid modulation of neuronal calcium channels.

The modulation of a family of cloned neuronal calcium channels by stimulation of a coexpressed mu opioid receptor was studied by transient expression in Xenopus oocytes. Activation of the morphine receptor with the synthetic enkephalin [D-Ala2,N-Me-Phe4,Gly-ol5]enkephalin (DAMGO) resulted in a rapid inhibition of alpha1A (by approximately 20%) and alpha1B (by approximately 55%) currents while alpha1C and alpha1E currents were not significantly affected. The opioid-induced effects on alpha1A and alpha1B currents were blocked by pertussis toxin and the GTP analogue guanosine 5'-[beta-thio]diphosphate. Similar to modulation of native calcium currents, DAMGO induced a slowing of the activation kinetics and exhibited a voltage-dependent inhibition that was partially relieved by application of strong depolarizing pulses. alpha1A currents were still inhibited in the absence of coexpressed Ca channel alpha2 and beta subunits, suggesting that the response is mediated by the alpha1 subunit. Furthermore, the sensitivity of alpha1A currents to DAMGO-induced inhibition was increased approximately 3-fold in the absence of a beta subunit. Overall, the results show that the alpha1A (P/Q type) and the alpha1B (N type) calcium channels are selectively modulated by a GTP-binding protein (G protein). The results raise the possibility of competitive interactions between beta subunit and G protein binding to the alpha1 subunit, shifting gating in opposite directions. At presynaptic terminals, the G protein-dependent inhibition may result in decreased synaptic transmission and play a key role in the analgesic effect of opioids and morphine.

Stimulation of growth factor receptor signal transduction by activation of voltage-sensitive calcium channels.

To understand the mechanisms by which electrical activity may generate long-term responses in the nervous system, we examined how activation of voltage-sensitive calcium channels (VSCCs) can stimulate the Ras/mitogen-activated protein kinase (MAPK) signaling pathway. Calcium influx through L-type VSCCs leads to tyrosine phosphorylation of the adaptor protein Shc and its association with the adaptor protein Grb2, which is bound to the guanine nucleotide exchange factor Sos1. In response to calcium influx, Shc, Grb2, and Sos1 inducibly associate with a 180-kDa tyrosine-phosphorylated protein, which was determined to be the epidermal growth factor receptor (EGFR). Calcium influx induces tyrosine phosphorylation of the EGFR to levels that can activate the MAPK signaling pathway. Thus, ion channel activation stimulates growth factor receptor signal transduction.

Heteromeric connexons in lens gap junction channels.

Gap junction channels are formed by paired oligomeric membrane hemichannels called connexons, which are composed of proteins of the connexin family. Experiments with transfected cell lines and paired Xenopus oocytes have demonstrated that heterotypic intercellular channels which are formed by two connexons, each composed of a different connexin, can selectively occur. Studies by Stauffer [Stauffer, K. A. (1995) J. Biol. Chem. 270, 6768-6772] have shown that recombinant Cx26 and Cx32 coinfected into insect cells may form heteromeric connexons. By solubilizing and subfractionating individual connexons from ovine lenses, we show by immunoprecipitation that connexons can contain two different connexins forming heteromeric assemblies in vivo.

Channel behavior in a gamma-aminobutyrate transporter.

Current produced by a gamma-aminobutyrate (GABA) transporter stably transfected into a mammalian cell line was observed in cell-attached and excised membrane patches. When GABA was absent, a fraction of the transporters produced cation-permeable channels. When GABA plus Na+ was on either side of the membrane, the majority of transporters produced a high-frequency current noise attributed to the movement of ions in an occluded pore.

cAMP compartmentation is responsible for a local activation of cardiac Ca2+ channels by beta-adrenergic agonists.

The role of cAMP subcellular compartmentation in the progress of beta-adrenergic stimulation of cardiac L-type calcium current (ICa) was investigated by using a method based on the use of whole-cell patch-clamp recording and a double capillary for extracellular microperfusion. Frog ventricular cells were sealed at both ends to two patch-clamp pipettes and positioned approximately halfway between the mouths of two capillaries that were separated by a 5-micron thin wall. ICa could be inhibited in one half or the other by omitting Ca2+ from one solution or the other. Exposing half of the cell to a saturating concentration of isoprenaline (ISO, 1 microM) produced a nonmaximal increase in ICa (347 +/- 70%; n = 4) since a subsequent application of ISO to the other part induced an additional effect of nearly similar amplitude to reach a 673 +/- 130% increase. However, half-cell exposure to forskolin (FSK, 30 microM) induced a maximal stimulation of ICa (561 +/- 55%; n = 4). This effect was not the result of adenylyl cyclase activation due to FSK diffusion in the nonexposed part of the cell. To determine the distant effects of ISO and FSK on ICa, the drugs were applied in a zero-Ca solution. Adding Ca2+ to the drug-containing solutions allowed us to record the local effect of the drugs. Dose-response curves for the local and distant effects of ISO and FSK on ICa were used as an index of cAMP concentration changes near the sarcolemma. We found that ISO induced a 40-fold, but FSK induced only a 4-fold, higher cAMP concentration close to the Ca2+ channels, in the part of the cell exposed to the drugs, than it did in the rest of the cell. cAMP compartmentation was greatly reduced after inhibition of phosphodiesterase activity with 3-isobutyl-methylxanthine, suggesting the colocalization of enzymes involved in the cAMP cascade. We conclude that beta-adrenergic receptors are functionally coupled to nearby Ca2+ channels via local elevations of cAMP.

Adenosine activates ATP-sensitive potassium channels in arterial myocytes via A2 receptors and cAMP-dependent protein kinase.

The mechanism by which the endogenous vasodilator adenosine causes ATP-sensitive potassium (KATP) channels in arterial smooth muscle to open was investigated by the whole-cell patch-clamp technique. Adenosine induced voltage-independent, potassium-selective currents, which were inhibited by glibenclamide, a blocker of KATP currents. Glibenclamide-sensitive currents were also activated by the selective adenosine A2-receptor agonist 2-p-(2-carboxethyl)-phenethylamino-5'-N- ethylcarboxamidoadenosine hydrochloride (CGS-21680), whereas 2-chloro-N6-cyclopentyladenosine (CCPA), a selective adenosine A1-receptor agonist, failed to induce potassium currents. Glibenclamide-sensitive currents induced by adenosine and CGS-21680 were largely reduced by blockers of the cAMP-dependent protein kinase (Rp-cAMP[S], H-89, protein kinase A inhibitor peptide). Therefore, we conclude that adenosine can activate KATP currents in arterial smooth muscle through the following pathway: (i) Adenosine stimulates A2 receptors, which activates adenylyl cyclase; (ii) the resulting increase intracellular cAMP stimulates protein kinase A, which, probably through a phosphorylation step, opens KATP channels.

Susceptibility of cloned K+ channels to reactive oxygen species.

Free radical-induced oxidant stress has been implicated in a number of physiological and pathophysiological states including ischemia and reperfusion-induced dysrhythmia in the heart, apoptosis of T lymphocytes, phagocytosis, and neurodegeneration. We have studied the effects of oxidant stress on the native K+ channel from T lymphocytes and on K+ channels cloned from cardiac, brain, and T-lymphocyte cells and expressed in Xenopus oocytes. The activity of three Shaker K+ channels (Kv1.3, Kv1.4, and Kv1.5), one Shaw channel (Kv3.4), and one inward rectifier K+ channel (IRK3) was drastically inhibited by photoactivation of rose bengal, a classical generator of reactive oxygen species. Other channel types (such as Shaker K+ channel Kv1.2, Shab channels Kv2.1 and Kv2.2, Shal channel Kv4.1, inward rectifiers IRK1 and ROMK1, and hIsK) were completely resistant to this treatment. On the other hand tert-butyl hydroperoxide, another generator of reactive oxygen species, removed the fast inactivation processes of Kv1.4 and Kv3.4 but did not alter other channels. Xanthine/xanthine oxidase system had no effect on all channels studied. Thus, we show that different types of K+ channels are differently modified by reactive oxygen species, an observation that might be of importance in disease states.