Voltage-gated anion channel of the electric organ of Narke japonica incorporated into planar bilayers

Voltage-gated anion channels in vesicles prepared from the electric organ of Narke japonica were studied using two methods. Ionic permeability was measured by the light scattering method, which could be used to measure the ion permeation of whole vesicles but only at a time scale of slower than about 0.1 s. The single channel conductances and permeability ratios for various ions were measured after fusing the vesicles to phospholipid bilayers. Both sets of results coincided, indicating that the anion channels observed with the planar bilayer method are the major route for anion passage in these vesicles. The channels showed anion selectivity and did not allow the permeation of cations such as K+ and choline+. The single channel conductance was 18 pS in 0.1 M Cl-. SCN- inhibited the conductance in a voltage-dependent reversible manner on both sides of a channel. SCN- may bind to the Cl- binding site in a channel and thus block it. 4,4'-Diisothiocyanostilbene-2,2'-disulfonic acid (DIDS) blocked a channel on the cis (extracellular) side irreversibly. The number of anion channels per vesicle was estimated to be about 50. It was also shown that all anion channels in the vesicles were open at the very instance of fusion with planar membranes.     

Ionic selectivity, saturation, and block in a K+-selective channel from sarcoplasmic reticulum

The open-channel conductance properties of a voltage-gated channel from sarcoplasmic reticulum were studied in planar phospholipid membranes. The channel is ideally selective for K+ over Cl- and for K+ over Ca++. In symmetrical 1 M solutions, the single-channel conductance (in pmho) falls in the order: K+ (214) > NH4+ (157) > Rb+ (125) > Na+ (72) > La+ (8.1) > Cs+ (< 3). In neutral bilayers, the channel conductance saturates with ion activity according to a rectangular hyperbolic relation, with half-saturation activities of 54 mM for K+ and 34 mM for Na+. Under symmetrical salt conditions, the K+:Na+ channel conductance ratio increases with salt activity, but the permeability ratio, measured by single-channel bi-ionic potentials, is constant between 20 mM and 2.5 M salt; the permeability ratio is equal to the conductance ratio in the limit of low-salt concentration. The channel conductance varies < 5% in the voltage range -100 to +70 mV. The maximum conductance varies K+ and Na+ is only weakly temperature dependent (delta H++ = 4.6 and 5.3 kcal/mol, respectively), but that of Li+ varies strongly with temperature (delta H++ = 13 kcal/mol). The channel's K+ conductance is blocked asymmetrically by Cs+, and this block is competitive with K+. The results are consistent with an Eyring-type barriers as it permeates the channel. The data conform to Lüger's (1973. Biochem. Biophys. Acta. 311:423-441) predictions for a "pure" single-ion channel.     

Channel properties of the purified acetylcholine receptor from Torpedo californica reconstituted in planar lipid bilayer membranes

The electrophysiological properties of the cation channel of the purified nicotinic acetylcholine receptor (AChR) reconstituted in planar lipid bilayers were characterized. Single-channel currents were activated by acetylcholine, carbamylcholine and suberyldicholine. The single channel conductance (28 pS in 0.3 M NaCl) was ohmic and independent of the agonist. Single channel currents increased with Na+ concentration to a maximum conductance of 95 pS and showed a half-saturation point of 395 mM. The apparent ion selectivity sequence, derived from single-channel current recordings, is: NH+4 greater than Cs+ greater than Rb+ greater than or equal to Na+ Cl-, F-, SO2-(4). The distribution of channel open times was fit by a sum of two exponentials, reflecting the existence of at least two distinct open states. The time constants depend on the choice of agonist, being consistently longer for suberyldicholine than for carbamylcholine. Similar channel properties were recorded in bilayers formed from monolayers at the tip of patch pipets . Single-channel currents occur in paroxysms of channel activity followed by quiescent periods. This pattern is more pronounced as the agonist concentration increases, and is reflected in histograms of channel-opening frequencies. Computer simulations with a three-state model, consisting of two closed (unliganded and liganded) and one open state, do not resemble the recorded pattern of channel activity, especially at high agonist concentration. Inclusion of a desensitized liganded state reproduces the qualitative features of channel recordings. The occurrence of paroxysms of channel activity thus seems to result from the transit of AChR through its active conformation, from which it can open several times before desensitizing.     

Cyclic AMP-dependent protein kinase activation of cystic fibrosis transmembrane conductance regulator chloride channels in planar lipid bilayers.

Membrane vesicles, prepared from mouse NIH-3T3 fibroblasts and Chinese hamster ovary cells expressing high levels of cystic fibrosis transmembrane conductance regulator (CFTR), were fused with Mueller-Rudin planar lipid bilayers. Upon addition of the catalytic subunit of cAMP-dependent protein kinase and ATP, low conductance Cl(-)-selective ion channels were observed in 10 of 16 experiments. The channels had a linear current-voltage relationship and a unitary conductance of approximately 6.5 pS. The channels were more permeable to Cl- than to I- and showed no appreciable time-dependent voltage activation. In contrast, addition of cAMP-dependent protein kinase and ATP to lipid bilayers fused with vesicles prepared from mock transfected (n = 14) cells failed to activate Cl- channels. These data support the conclusion that CFTR is a Cl- channel. They indicate that it can be reconstituted in a planar lipid bilayer and that the biophysical and regulatory properties are very similar to those observed in the native cell membrane. These data also argue against the requirement for loosely associated factors for regulation or function of the channel.     

Probing an open CFTR pore with organic anion blockers

The cystic fibrosis transmembrane conductance regulator (CFTR) is an ion channel that conducts Cl- current. We explored the CFTR pore by studying voltage-dependent blockade of the channel by two organic anions: glibenclamide and isethionate. To simplify the kinetic analysis, a CFTR mutant, K1250A-CFTR, was used because this mutant channel, once opened, can remain open for minutes. Dose-response relationships of both blockers follow a simple Michaelis-Menten function with K(d) values that differ by three orders of magnitude. Glibenclamide blocks CFTR from the intracellular side of the membrane with slow kinetics. Both the on and off rates of glibenclamide block are voltage dependent. Removing external Cl- increases affinity of glibenclamide due to a decrease of the off rate and an increase of the on rate, suggesting the presence of a Cl- binding site external to the glibenclamide binding site. Isethionate blocks the channel from the cytoplasmic side with fast kinetics, but has no measurable effect when applied extracellularly. Increasing the internal Cl- concentration reduces isethionate block without affecting its voltage dependence, suggesting that Cl- and isethionate compete for a binding site in the pore. The voltage dependence and external Cl- concentration dependence of isethionate block are nearly identical to those of glibenclamide block, suggesting that these two blockers may bind to a common binding site, an idea further supported by kinetic studies of blocking with glibenclamide/isethionate mixtures. By comparing the physical and chemical natures of these two blockers, we propose that CFTR channel has an asymmetric pore with a wide internal entrance and a deeply embedded blocker binding site where local charges as well as hydrophobic components determine the affinity of the blockers.     

A chloride channel from lobster walking leg nerves. Characterization of single-channel properties in planar bilayers

A novel, small conductance of Cl- channel was characterized by incorporation into planar bilayers from a plasma membrane preparation of lobster walking leg nerves. Under conditions of symmetrical 100 mM NaCl, 10 mM Tris-HCl, pH 7.4, single Cl- channels exhibit rectifying current-voltage (I-V) behavior with a conductance of 19.2 +/- 0.8 pS at positive voltages and 15.1 +/- 1.6 pS in the voltage range of -40 to 0 mV. The channel exhibits a negligible permeability for Na+ compared with Cl- and displays the following sequence of anion permeability relative to Cl- as measured under near bi-ionic conditions: I- (2.7) greater than NO3- (1.8) greater than Br- (1.5) greater than Cl- (1.0) greater than CH3CO2- (0.18) greater than HCO3- (0.10) greater than gluconate (0.06) greater than F- (0.05). The unitary conductance saturates with increasing Cl- concentration in a Michaelis-Menten fashion with a Km of 100 mM and gamma max = 33 pS at positive voltage. The I-V curve is similar in 10 mM Tris or 10 mM HEPES buffer, but substitution of 100 mM NaCl with 100 mM tetraethylammonium chloride on the cis side results in increased rectification with a 40% reduction in current at negative voltages. The gating of the channel is weakly voltage dependent with an open-state probability of 0.23 at -75 mV and 0.64 at +75 mV. Channel gating is sensitive to cis pH with an increased opening probability observed for a pH change of 7.4 to 11 and nearly complete inhibition for a pH change of 7.4 to 6.0. The lobster Cl- channel is reversibly blocked by the anion transport inhibitors, SITS (4-acetamido, 4'-isothiocyanostilbene-2,2'-disulfonic acid) and NPPB (5-nitro-2-(3-phenylpropylamino)benzoic acid). Many of these characteristics are similar to those previously described for small conductance Cl- channels in various vertebrate cells, including epithelia. These functional comparisons suggest that this invertebrate Cl- channel is an evolutionary prototype of a widely distributed class of small conductance anion channels.     

Anion competition for a volume-regulated current.

We have examined whether the anionic amino acids, glutamate and aspartate, permeate through the same volume-regulated conductance permeant to Cl- ions. Cell swelling was initiated in response to establishing a whole-cell configuration in the presence of a hyposmotic gradient. Volume-regulated anion currents carried by Cl-, glutamate, or aspartate developed with similar time courses and showed similar voltage-dependent inactivation. Permeability ratios (Paa/PCl) calculated from measured reversal potentials were dependent on the mole fraction ratio (MFR) of the permeant anions ([aa]/([aa] + [Cl-])). MFR was varied from 0.00 to 0.97. As the fraction of amino acid increased, Paa/PCl decreased. Current amplitude was similarly dependent on MFR. These results show that the permeation of anionic amino acids and that of Cl- ions are not independent of each other, indicating that the ion channel underlying the volume-regulated conductance can be occupied by more than one ion at a time. Application of Eyring rate theory indicated that the major barrier to Cl- ion permeation is at the intracellular side of the membrane, and that the major barrier to amino acid permeation is at the extracellular side of the membrane. The interactions between these permeant ions may have a physiological modulatory role in volume regulation through a volume-regulated anion conductance.     

Properties of the cGMP-activated channel of retinal on-bipolar cells.

Whole-cell patch-clamp recordings were obtained from on-bipolar cells in, or isolated from, retinal slices prepared from dogfish retina. The properties of the cGMP-activated conductance of on-bipolar cells were compared with that of rod photoreceptors. The on-bipolar cell cGMP-activated channel was blocked by L-cis-diltiazem, a block which was strongly voltage dependent. However, this channel is not identical with that of photoreceptors. The location of the L-cis-diltiazem blocking site and its accessibility in the channel are not the same as in rods. The voltage dependence of block suggests that the blocking site, although near the intracellular side of the channel, is accessible to the positively charged form of L-cis-diltiazem only from the outward facing side of the channel. Furthermore, in contrast to rod channels, the conductance of the on-bipolar cell channels is unaltered by the removal of external divalent cations.     

Asymmetry of the gramicidin channel in bilayers of asymmetric lipid composition: I. Single channel conductance

Summary Gramicidin-doped asymmetric bilayers made by the Montal-Mueller method exhibited an asymmetric current-voltage relationship. The asymmetric conductance was shown to be the product of two components, a rectifying single-channel conductance and an asymmetric voltage dependence of the reaction which leads to the conducting channel. The single-channel conductance was asymmetric in both asymmetric bilayers made of charged lipids and asymmetric bilayers made only of neutral lipids. The single-channel asymmetry decreased with increasing ion concentration. From the comparison of the singlechannel conductance in symmetric and asymmetric bilayers and the dependence of the asymmetry on the solution ion concentrations, it was concluded that (1) the rate of ion entry into the channel is dependent on the lipid composition of the membrane and is asymmetric in asymmetric bilayers; (2) the entry step is rate determining at low ion concentrations; and (3) at higher ion concentrations the rate-determining step is the translocation across the main barrier in the membrane; and this translocation appears insensitive to lipid asymmetry.     

Antibacterial peptide pleurocidin forms ion channels in planar lipid bilayers

Pleurocidin, a 25-residue alpha helical cationic peptide, isolated from skin mucous secretions of the winter flounder, displays a strong anti-microbial activity and appears to play a role in innate host defence. This peptide would be responsible for pore formation in the membrane of bacteria leading to lysis and therefore death. In this study, we investigated the behaviour of pleurocidin in different planar lipid bilayers to determine its mechanism of membrane permeabilisation. Macroscopic conductance experiments showed that pleurocidin did not display a pore-forming activity in neutral phosphatidylcholine/phosphatidylethanolamine (PC/PE) lipid bilayers. However, in 7:3:1 PC/PE/phosphatidylserine (PS) lipid bilayers, pleurocidin showed reproducible I/V curves at different peptide concentrations. This activity is confirmed by single-channel experiments since well-defined ion channels were obtained if the lipid mixture was containing an anionic lipid (PS). The ion channel characteristics such as-no voltage dependence, only one unitary conductance, linear relation ship current-voltage-, are not in favour of the membrane permeabilisation according to the barrel model but rather by the toroidal pore formation.     

Immunological identification of a Cl- channel protein in electric organs of Narke japonica.

A Cl- channel in electric organs of Narke japonica was reconfirmed, using a polyclonal antibody, to be a 180k protein composed of two identical 90k units. The specific antibody against the 180k protein from electric organs reduced, when added in the cis side solution, the open probability of the Cl- channel in a planar bilayer membrane without affecting the single channel conductance, ion selectivity nor voltage dependency. Furthermore, the antibody added in the trans side also affected the channel to increase the open probability. The eluate of the immuno-affinity chromatography was found to contain only the 180k protein. These results indicate that 180k protein to be an integral membrane protein and to form the Cl- channel.     

Low-conductance chloride channel from crayfish skeletal muscle incorporated into planar lipid bilayers.

Low-conductance chloride channel from skeletal muscle SR vesicles of the crayfish Astacus fluviatilis was incorporated into planar lipid bilayers and its basic characteristics were investigated. The channel has a relatively low unitary conductance of 26 pS in symmetrical 160 mmol/l choline-chloride. The dependence of the channel conductance on Cl- concentration shows saturating behavior with a maximum conductance of 37 pS and an ionic activity for half-maximum conductance Km = 75 mmol/l. The channel exhibits a complex kinetics with several modes of activity. Open state probability slightly decreases with the increasing absolute value of voltage. The channel activity does not appear to be dependent on the presence of Ca2+ ions. The channel is effectively inhibited by DIDS, a stilbene derivative. The permeability properties of the channel are similar to the specific behavior of the "double-barrelled" channel from Torpedo electroplax described by Miller and White (1984).     

Schistosoma mansoni: patch-clamp study of a nonselective cation channel in the outer tegumental membrane of females.

An apparent ion channel with a conductance of 295 pS is present in isolated inside-out patches of outer tegumental membrane taken from female Schistosoma mansoni. With positive voltages applied to the intracellular face of the patch, percentage open time for the channel was 0 to 50; with negative voltages applied, percentage open time was greater than 99. Step changes in applied voltage characteristically induced opening-closing activity. However, there was no maintained applied voltage at which there was a high level of sustained opening-closing activity. The 295 pS conductance was by far the most commonly occurring conductance but it appears to result from cooperativity among several channels, the unitary conductance for the channel averaging 95 pS. Alterations in the Na+ or K+ concentration ratios changed the reversal potential for this conductance but alterations in the Cl- concentration did not. From this it is concluded that this channel is selective for Na+ or K+ over Cl- and it appears to be a nonselective cation channel.     

Single chloride-selective channels active at resting membrane potentials in cultured rat skeletal muscle.

The patch-clamp technique was used to characterize channels that could contribute to the resting Cl-conductance in the surface membrane of cultured rat skeletal muscle. Two Cl- -selective channels, in addition to the Cl- -selective channel of large conductance described previously (Blatz and Magleby, 1983), were observed. One of these channels had fast kinetics and a conductance of 45 +/- 1.8 pS (SE) in symmetrical 100 mM KCl. The other had slow kinetics and a conductance of 61 +/- 2.4 pS. The channel with fast kinetics typically closed within 1 ms after opening and flickered between the open and shut states. The channel with slow kinetics typically closed within 10 ms after opening and displayed less flickering. Both channels were active in excised patches of membrane held at potentials similar to resting membrane potentials in intact cells, and both were open a greater percentage of time with depolarization. Under conditions of high ion concentrations, both channels exhibited nonideal selectivity for Cl- over K+ with the permeability ratio PK/PCl of 0.15-0.2. Additional experiments on the fast Cl- channel indicated that its activity decreased with lowered pHi and that SO2-4 and CH3SO-4 were ineffective charge carriers. These findings, plus the observation that the fast Cl- channel was also active in membrane patches on intact cells, suggest that the fast Cl- channel provides a molecular basis for at least some of the resting Cl- conductance. The extent to which the slow Cl- channel contributes is less clear as it was typically active only after excised patches of membrane had been exposed to high concentrations of KCl at the inner membrane surface.     

Voltage-dependent block of anthrax toxin channels in planar phospholipid bilayer membranes by symmetric tetraalkylammonium ions. Effects on macroscopic conductance

In a recent paper (Blaustein, R. O., T. M. Koehler, R. J. Collier, and A. Finkelstein, 1989. Proc. Natl. Acad. Sci. USA. 86:2209-2213) we described the general channel-forming properties of the PA65 fragment of anthrax toxin in planar phospholipid bilayer membranes. In the present paper we extend our previous studies of the permeability properties of this channel, using a series of symmetric tetraalkylammonium (TAA) ions. Our main finding is that at micromolar concentrations on either the cis (toxin-containing) or trans side of a membrane containing many (greater than 1,000) channels, these ions, ranging in size from tetramethylammonium to tetrahexylammonium, induce a voltage-dependent reduction of membrane conductance. (We attribute a similar voltage-dependent reduction of membrane conductance by millimolar concentrations of HEPES to a cationic form of this buffer present at micromolar concentrations.) In going from large negative to large positive voltages (on the TAA side) one sees that the conductance first decreases from its value in the absence of TAA, reaches a minimum, and then rises back at larger positive voltages toward the level in the absence of TAA. Our interpretation of this behavior is that these symmetric TAA ions block the cation-selective PA65 channel in a voltage-dependent manner. We postulate that there is a single site within the channel to which TAA ions can bind and thereby block the passage of the major current-carrying ion (potassium). A blocking ion is driven into the site by modest positive voltages, but is driven off the site and through the channel by larger positive voltages, thus explaining the relief of block. (In the accompanying paper [Blaustein, R. O., E. J. A. Lea, and A. Finkelstein. 1990. J. Gen. Physiol. 96:921-942] we confirm this interpretation of the data by analysis at the single-channel level.) This means that these blocking ions can pass through the channel; the permeability to tetrahexylammonium, the largest ion studied, implies that the narrowest part of the channel has a diameter of at least 11 A.     

Conduction and blocking properties of a predominantly anion-selective channel from human platelet surface membrane reconstituted into planar phospholipid bilayers

We have investigated the basic properties of a predominantly anion-selective channel derived from highly purified human platelet surface membrane. Single channels have been reconstituted into planar phospholipid bilayers by fusion of membrane vesicles and recorded under voltage-clamp conditions. The channel is found to have the following properties: (i) Channel activity occurs in bursts of openings separated by long closed periods. (ii) The current-voltage relationship is nonlinear. Channel current is seen to rectify, with less current flowing at positive than at negative voltages. Rectification may be due to asymmetric block by HEPES/Tris buffers. In 450 mM KCl, 5 mM HEPES/Tris, pH 7.2, the single channel conductance at -40 mV is approximately 160 pS and at +40 mV is approximately 90 pS. (iii) The conductance-concentration relationship follows a simple saturation curve. Half maximal conductance is achieved at a concentration of approximately 1000 mM KCl, and the curve saturates at a conductance of approximately 500 pS. (iv) Reversal potentials interpreted in terms of the Goldman-Hodgkin-Katz equation indicate a Cl: K permeability ratio of 4:1. (v) The channel accepts all of the halides as well as a number of other anions. The following sequence of relative anion permeabilities (in the presence of K+) is obtained: F- less than acetate- less than gluconate- less than Cl- less than Br- less than I- less than NO3- less tha SCN-.(vi) Cations as large as TEA+ are permeant. (vii) Current through the channel is blocked in the presence of DIDS, SITS and ATP, but not by Zn2+.     

Antibacterial peptide pleurocidin forms ion channels in planar lipid bilayers

Pleurocidin, a 25-residue α helical cationic peptide, isolated from skin mucous secretions of the winter flounder, displays a strong anti-microbial activity and appears to play a role in innate host defence. This peptide would be responsible for pore formation in the membrane of bacteria leading to lysis and therefore death. In this study, we investigated the behaviour of pleurocidin in different planar lipid bilayers to determine its mechanism of membrane permeabilisation. Macroscopic conductance experiments showed that pleurocidin did not display a pore-forming activity in neutral phosphatidylcholine/phosphatidylethanolamine (PC/PE) lipid bilayers. However, in 7:3:1 PC/PE/phosphatidylserine (PS) lipid bilayers, pleurocidin showed reproducible I/V curves at different peptide concentrations. This activity is confirmed by single-channel experiments since well-defined ion channels were obtained if the lipid mixture was containing an anionic lipid (PS). The ion channel characteristics such as—no voltage dependence, only one unitary conductance, linear relation ship current-voltage—, are not in favour of the membrane permeabilisation according to the barrel model but rather by the toroidal pore formation.     

Functional influence of the pore helix glutamate in the KcsA K+ channel

The E71 residue is buried near the selectivity filter in the KcsA K+ channel and forms a carboxyl-carboxylate bridge with D80. We have investigated the importance of E71 by examining neutralization mutants at this position using biochemical and electrophysiological methods. E71 mutations differentially destabilize the detergent-solubilized tetramer; among them, the E71V neutralization mutant has a relatively subtle effect. The E71V channel displays electrical activity when reconstituted into planar lipid bilayers. In single channel recordings, the mutant retains K+/Na+ selectivity, and its conductance in the outward direction is unaltered. Some conduction properties are changed: inward conductance is increased. Our results show that that the E71 side chain is not a primary determinant of ion selectivity or conduction in the wild-type channel, either directly or through the E71:D80 carboxyl-carboxylate bridge.     

Side-chain charge effects and conductance determinants in the pore of ClC-0 chloride channels

The charge on the side chain of the internal pore residue lysine 519 (K519) of the Torpedo ClC-0 chloride (Cl-) channel affects channel conductance. Experiments that replace wild-type (WT) lysine with neutral or negatively charged residues or that modify the K519C mutant with various methane thiosulfonate (MTS) reagents show that the conductance of the channel decreases when the charge at position 519 is made more negative. This charge effect on the channel conductance diminishes in the presence of a high intracellular Cl- concentration ([Cl-]i). However, the application of high concentrations of nonpermeant ions, such as glutamate or sulfate (SO42-), does not change the conductance, suggesting that the electrostatic effects created by the charge at position 519 are unlikely due to a surface charge mechanism. Another pore residue, glutamate 127 (E127), plays an even more critical role in controlling channel conductance. This negatively charged residue, based on the structures of the homologous bacterial ClC channels, lies 4-5 A from K519. Altering the charge of this residue can influence the apparent Cl- affinity as well as the saturated pore conductance in the conductance-Cl- activity curve. Amino acid residues at the selectivity filter also control the pore conductance but mutating these residues mainly affects the maximal pore conductance. These results suggest at least two different conductance determinants in the pore of ClC-0, consistent with the most recent crystal structure of the bacterial ClC channel solved to 2.5 A, in which multiple Cl--binding sites were identified in the pore. Thus, we suggest that the occupancy of the internal Cl--binding site is directly controlled by the charged residues located at the inner pore mouth. On the other hand, the Cl--binding site at the selectivity filter controls the exit rate of Cl- and therefore determines the maximal channel conductance.     

Properties of Voltage-gated Ion Channels Formed by Syringomycin E in Planar Lipid Bilayers

Using the planar lipid bilayer technique we demonstrate that the lipodepsipeptide antibiotic, syringomycin E, forms voltage-sensitive ion channels of weak anion selectivity. The formation of channels in bilayers made from dioleoylglycerophosphatidylserine doped with syringomycin E at one side (1–40 μg/ml) was greatly affected by cis-positive voltage. A change of voltage from a positive to a negative value resulted in (i) an abrupt increase in the single channel conductance (the rate of increase was voltage dependent) simultaneous with (ii) a closing of these channels and an exponential decrease in macroscopic conductance over time. The strong voltage dependence of multichannel steady state conductance, the single channel conductance, the rate of opening of channels at positive voltages and closing them at negative voltages, as well as the observed abrupt increase of single channel conductance after voltage sign reversal suggest that the change of the transmembrane field induces a significant rearrangement of syringomycin E channels, including a change in the spacing of charged groups that function as voltage sensors. The conductance induced by syringomycin E increased with the sixth power of syringomycin E concentration suggesting that at least six monomers are required for channel formation. Received: 3 April 1995/Revised: 24 August 1995