Search

Glutamate is the major excitatory neurotransmitter in the mammalian central nervous system. After release from presynaptic nerve terminals, glutamate is quickly removed from the synaptic cleft by a family of five glutamate transporters, the so-called excitatory amino acid transporters (EAAT1-5). EAATs are prototypic members of the growing number of dual-function transport proteins: they are not only glutamate transporters, but also anion channels. Whereas the mechanisms underlying secondary active glutamate transport are well understood at the functional and at the structural level, mechanisms and cellular roles of EAAT anion conduction have remained elusive for many years. Recently, molecular dynamics simulations combined with simulation-guided mutagenesis and experimental analysis identified a novel anion-conducting conformation, which accounts for all experimental data on EAAT anion currents reported so far. We here review recent findings on how EAATs accommodate a transporter and a channel in one single protein.

CLC-type chloride channels exhibit a unique double-barreled architecture with two independently functioning ion conduction pathways, the so-called protopores. There exist gating processes that open and close individual protopores as well as common processes that jointly mediate slow opening and closing of both protopores. Different isoforms exhibit distinct voltage dependences and kinetics of gating. Whereas opening of the individual and common gate of homo-dimeric ClC-1 is promoted by membrane depolarization, ClC-2 is closed at positive potentials and opens only at negative voltages. To characterize the functional interaction of protopores we engineered a concatameric construct linking the coding regions of ClC-1 and ClC-2 in an open reading frame, expressed it in mammalian cells and measured anion currents through whole-cell and single channel patch clamping. In the hetero-dimeric assembly, each protopore displayed two kinetically distinct gating processes. Fast gating of the ClC-1 protopore closely resembled fast protopore gating of homo-dimeric channels. The voltage dependence of ClC-2 fast gating was shifted to more positive potentials by the adjacent ClC-1 protopore, resulting in open ClC-2 protopores at positive voltages. We observed two slow gating processes individually acting on ClC-1 and ClC-2 protopores, with distinct time and voltage dependences. Single channel recordings demonstrated that hetero-dimerization additionally modified the unitary conductance of ClC-2 protopores. Our findings suggest that inter-subunit interactions do not only affect common gating, but also ion permeation and gating of individual protopores in hetero-dimeric ClC channels.

ClC-2 is a voltage-dependent chloride channel that activates slowly at voltages negative to the chloride reversal potential. Adenosine triphosphate (ATP) and other nucleotides have been shown to bind to carboxy-terminal cystathionine-ß-synthase (CBS) domains of ClC-2, but the functional consequences of binding are not sufficiently understood. We here studied the effect of nucleotides on channel gating using single-channel and whole-cell patch clamp recordings on transfected mammalian cells. ATP slowed down macroscopic activation and deactivation time courses in a dose-dependent manner. Removal of the complete carboxy-terminus abolishes the effect of ATP, suggesting that CBS domains are necessary for ATP regulation of ClC-2 gating. Single-channel recordings identified long-lasting closed states of ATP-bound channels as basis of this gating deceleration. ClC-2 channel dimers exhibit two largely independent protopores that are opened and closed individually as well as by a common gating process. A seven-state model of common gating with altered voltage dependencies of opening and closing transitions for ATP-bound states correctly describes the effects of ATP on macroscopic and microscopic ClC-2 currents. To test for a potential pathophysiological impact of ClC-2 regulation by ATP, we studied ClC-2 channels carrying naturally occurring sequence variants found in patients with idiopathic generalized epilepsy, G715E, R577Q, and R653T. All naturally occurring sequence variants accelerate common gating in the presence but not in the absence of ATP. We propose that ClC-2 uses ATP as a co-factor to slow down common gating for sufficient electrical stability of neurons under physiological conditions.

Ca(2+)-activated K+ channels of a large conductance (BKCa) in human skeletal muscle were studied by patch clamping membrane blebs and by using the three microelectrode voltage-clamp recording technique on resealed fibre segments. Single-channel recordings in bleb-attached and inside-out modes revealed BKCa conductances of 230 pS for symmetrical and 130 pS for physiological K+ distributions. Open probability increased with membrane depolarization and increasing internal [Ca2+]. The Hill coefficient was 2.0, indicating that at least two Ca2+ ions are required for full activation. Kinetic analysis revealed at least two open and three closed states. An additional long-lived inactivated state, lasting about 0.5-20 s, was observed following large depolarizations, when extracellular K+ was lowered to physiological values. BKCa were blocked by three means: (1) externally by tetraethylammonium which reduced single-channel amplitude (IC50 approx. 0.3 mM); (2) internally by polymyxin B which decreased the open probability (IC50 approx. 5 micrograms/ml); and (3) externally by charybdotoxin which caused long-lasting periods of inactivation (IC50 < 10 nM). Measurements on resealed fibre segments at physiological [K+] were in accordance with the single-channel data: only when intracellular [Ca2+] was elevated did charybdotoxin (50 nM) reduce the macroscopic membrane K+ conductance with depolarizing voltage steps.

The effects of volume change and cytoskeleton manipulation on the Cl- channels in human embryonic skeletal muscle cells were studied. Trypsination, used for production of myoballs, changes the channel properties only a little. When the external osmolarity was reduced from 300 to 270 mosmol/l, the specific Cl- conductance, gCl, (at -80 mV) of myoballs increased from 5.1 +/- 1.9 to 30.4 +/- 12.2 microS/cm2 (SD; n = 6) within 15 min. Concomitantly, the kinetics of Cl- currents, elicited by clamping the membrane potential from a negative to positive values, changed from activation and subsequent slow inactivation to instantaneous activation with fast inactivation. G protein activation, protein kinase action or [Ca2+]i elevation seemed not to be involved in these effects. Similar changes were produced in the absence of a transmembrane osmotic gradient by 500 nM intracellular cytochalasin D (gCl = 34.3 +/- 10.3 microS/cm2; n = 6) or 12.5 microM colchicine (gCl = 15.4 +/- 1.4 microS/cm2; n = 5). When the external osmolarity was increased to 418 mosmol/l, 1 microM cytochalasin D did not affect gCl. In four of six cell-attached patches the open probability of the intermediate Cl- channel was increased after reduction of the bath osmolarity. In inside-out patches, the drugs increased the open probability of the channels. It is concluded that the Cl- channels are under control of the cytoskeleton.

The regulation of Cl- channels in human myoballs by G proteins was studied using whole-cell and inside-out patch recordings. After perfusion of the cell with 0.1 mM GTP[gamma S], the specific Cl- conductance, GCl, at standard resting potential (-85 mV) was increased from 5.9 microS/cm2 to 103 microS/cm2, and the kinetics upon stepping the potential to positive values was changed from an activating current with very slow inactivation to a fast inactivating current with no potential-dependent activation. These effects were not affected by the simultaneous blockade of several signal cascades involving G proteins. Addition of the protein kinase blockers PKI (25 microM), H8 (10 microM), or of the phospholipase-A2-blocking agent quinacrine (10 microM), had not much influence on these GTP[gamma S] effects. Buffering of the intracellular Ca2+ concentration (0.1 microM) or addition of the Ca2+/calmodulin antagonist trifluoperazine (50 microM) was also without effect. Pre-incubation of the cells with pertussis toxin or with cholera toxin did not change GCl. In excised inside-out patches voltage-clamped at -85 mV, application of GTP[gamma S] influenced the "intermediate" Cl- channel, the Cl- channel type having the highest density in these cells, by increasing the number of transitions in a half-conductance state. The probability of the channel being in one of the two conducting states rose from 0.015 to 0.67, and the kinetics of the single-channel currents was changed so that, on average, it was similar to the whole-cell current kinetics seen after application of GTP[gamma S]. It is concluded that a G protein is directly interacting with these channels.

The Cl- channels in human myoballs were investigated with several recording techniques. Three types of channels were found and dubbed "small", "intermediate", and "large", according to their different conductance. The intermediate Cl- channel was observed most frequently. It was active at the resting potential immediately after seal formation in cell-attached as well as in excised patches. Its Cl- selectivity was rather high (PCl/PNa = 9.46; PCl/PMeSO4 = 7.85 where P denotes permeability) and the slope conductance at the reversal potential with [Cl-]o/[Cl-]i equal to 160 mM/42 mM was 31 pS. The channel showed an open-channel substructure with two subconductance levels having equal amplitudes. It can conduct two kinetically different currents that correspond to the activating and the inactivating Cl- current components described by Zachar et al. (1992). The small Cl- channel had a conductance of 10 pS at the reversal potential, a PCl/PNa of 2.7, and a PCl/PMeSO4 of 22.6. Its open probability was biggest negative to -85 mV, resulting in an inactivating whole-cell Cl- current component. Because of the small channel density and conductance the contribution of this channel type to the whole-cell current seems to be small. Patches with only one small channel were never observed which suggests that this channel type occurs in clusters. A third type of channel with very large conductance (250 pS) was seen only four times.

Chloride currents in human myoballs were investigated with the tight-seal whole-cell recording method in a wide range of membrane voltages (-125 to +145 mV). Two current components having different kinetics could be distinguished. In more than 90% of the myoballs the following results were obtained. At negative potentials, the amplitude of the Cl- current was small and independent of time. The amplitude of the current increased as the membrane potential was made more positive. At potentials positive to +75 mV, the current increased monoexponentially with time. Inactivation occurred only during very long (greater than 3 s) pulses. When such a test pulse was preceded by a conditioning pulse to +60 mV, the current at potentials more than +90 mV was markedly smaller than in the absence of a prepulse, and no activation was provoked by strongly pulses. Recovery from inactivation could only be measured at potentials negative to -40 mV. The Cl- conductance at -85 mV was 5.9 +/- 3.64 microS/cm2 (SD; n = 10). In about 5% of the myoballs a kinetically different current was visible, characterized by fast inactivation at highly positive potentials. The current amplitudes were substantially larger in such cases, the Cl- conductance at -85 mV being 12.2 +/- 9.02 microS/cm2 (n = 4).

The influence of giga-seal formation on the properties of the Na+ channels within the covered membrane patch was investigated with a whole-cell pipette and a patch pipette applied to the same cell. Current kinetics, current/voltage relation and channel densities were determined in three combinations: (i) voltage-clamping and current recording with the whole-cell pipette, (ii) voltage-clamping with the whole-cell pipette and current recording with the patch pipette and, (iii) voltage-clamping and current recording with the patch pipette. The Hodgkin-Huxley (1952) parameters tau m and tau h were smaller for the patch currents than for the whole cell, and the h infinity curve was shifted in the negative direction. The channel density was of the order of 10 times smaller. All effects were independent of the extracellular Ca2+ concentration. The capacitive current generated in the patch by the whole-cell Na+ current and its effect on the transmembrane voltage of the patch were evaluated. The kinetic parameters of the Na+ channels in the patch did not depend on whether the voltage was clamped with the whole-cell pipette or the patch pipette. Thus, the results are not due to spurious voltage.