Characterization of single non-inactivating potassium channels in primary neuronal cultures of Drosophila

Within the last five years, dramatic advances in the study of membrane channels have been achieved by combining molecular genetics and patch-clamp technology (Sakmann et al. 1985; Claudio et al. 1987). In Drosophila, attempts to identify genes encoding potassium channels by means of transposon tagging and chromosome walking have met with great success (Schwarz et al. 1988; Timpe et al. 1988). However, there is limited information concerning the physiological and biophysical properties of neuronal channels in Drosophila, as the small diameter of Drosophila neurons precludes application of conventional electrophysiological methods (Tanouye et al. 1981; Wu et al. 1983b). The present work provides detailed information about single voltage-gated potassium channels in Drosophila neurons by using the gigaohm-seal patch-clamp technique.

All experiments were carried out with primary cultured neurons of Drosophila (Wu et al. 1983b). Mature third-instar larvae of wild-type (Canton-S) or eag (Ganetzky & Wu, 1985) Drosophila melanogaster were used. Before dissection, larvae were surface-sterilized by brief immersion in 70 % ethanol and then rinsed three times with sterile distilled water. The brain and ventral ganglion were dissected from larvae in tissue culture medium, and segmental nerves, imaginai discs and other tissues were removed. The culture medium was a revised Schneider medium (Schneider, 1964) from Grand Island Biological Co. (GIBCO; Grand Island, NY), supplemented with heat-inactivated (56°C, 30min) fetal bovine serum (10%), penicillin G (50 units ml⁻¹) and streptomycin sulfate (50 μg ml⁻¹) (all obtained from GIBCO). After rinsing, the brains and ganglia were first torn into small fragments in the culture medium with fine dissecting needles and then incubated for 1 h in a saline similar to Rinaldini’s (Schneider & Blumenthal, 1978) Ca²⁺-and Mg²⁺-free solution (800 mg of NaCl, 20 mg of KC1, 55 mg of NaH₂PO₄. H₂O, 100 mg of NaHCO₃, and 100 mg of glucose in 100 ml of distilled water) containing 10 units ml⁻¹ of papain (type III, Sigma Chemical Co., St Louis, MO). The fragments were rinsed three times with culture medium supplemented with 10 % fetal bovine serum and then dispersed into single cells in the same medium by repeated flushing through a siliconized glass pipet. To obtain monolayer cultures, 80–200 μl of cell suspension was placed on round glass coverslips (15 mm diameter) in dry Petri dishes. The coverslips were precoated with poly-L-lysine (M_r ≈ 14000; Sigma) to facilitate cell adhesion. Cultures with densities ranging from 0·5 to 5×10⁴cells cm^-2 were maintained at 25°C in humidified air containing 5 % CO₂. Cells cultured for 1 day to 2 weeks were used for experiments.

Membrane currents were recorded either from excised inside-out membrane patches or in the whole-cell configuration using the gigaohm-seal patch-clamp technique (Hamill et al. 1981). The cells were initially immersed in normal saline containing (in mmol l⁻¹); NaCl, 125; KC1, 5·5; CaCl₂,1·8; MgCl₂,0·8; glucose, 25; sucrose, 20; and Hepes, 10 (pH adjusted to 7·3 with NaOH). For recordings from cell-free patches, the pipet was filled with an ‘extracellular solution’ containing (in mmol l⁻¹): choline-Cl, 145; MgCl₂, 3; ethylenebis(oxyethylenenitrilo)-tetraacetic acid (EGTA), 2; and Hepes, 10 [pH adjusted to 7·3 with tetramethylammonium hydroxide (TMA-OH)]. This Ca²⁺-, K⁺-and Na⁺-free extracellular solution was used throughout the study to eliminate Ca²⁺-activated conductances and Na⁺-and K⁺-dependent inward currents (Yamamoto & Suzuki, 1987a,b). The composition of the internal solution used for the excised patch experiments was: 145 mmol l⁻¹ KC1, l·13 mmol l⁻¹ MgCl₂, l·92 mmol l⁻¹ EGTA and 10 mmol l·¹ Hepes. The pH of the internal solution was adjusted to 7·3 with KOH. In experiments to examine ionic selectivity of the channel, KC1 was totally replaced with either RbCl, NaCl or NH₄CI. Some experiments were performed with internal solutions containing either F⁻ or NO₃⁻ substituted for Cl⁻. In the case of whole-cell recordings, the pipet was filled with an ‘internal’ solution containing: 145 mmol l⁻¹ potassium glutamate, 1·13 mmol l⁻¹ MgCl₂, 1·92 mmol l⁻¹ EGTA and 10 mmol l⁻¹ Hepes. All solutions were filtered immediately before use through a membrane filter with 0·45 μm pore diameter (Toyo membrane filters, type TM-2). The temperature of the chamber was controlled by a Peltier device and measured with a thermocouple probe, and was kept within ± 0·5°C of the desired setting. Experiments were performed at 10°C, unless otherwise indicated.

To record single-channel currents, a Dagan model 8900 patch-clamp/whole-cell-clamp amplifier (Dagan, Minneapolis, MN) with probe type 8930 (feed-back resister: 10 GΩ) was used. All records were filtered at 3 kHz with a six-pole Bessel filter, except when mentioned. The bath potential was measured using a separate reference microelectrode filled with 3 mol l⁻¹ KC1. The interior of the pipet was clamped at a command voltage which was generated by a function generator (model 275, Wavetek Inc., San Diego, CA) under microcomputer (model 9816S; Hewlett-Packard Co., Fort Collins, CO) control (Yamamoto, 1986). The current signals were digitized with a digital oscilloscope (2090-3C; Nicolet Scientific Corp., Northvale, NJ) at the sampling frequency of 5 or 20 kHz, and the digitized data were stored on floppy disks. Leakage and capacitative currents were eliminated from the records by subtracting averaged records with no channel openings.

Using half-amplitude threshold analysis (Colquhoun & Sigworth, 1983), we determined the transitions between the open and closed states. Single-channel kinetics were analyzed from records in which overlapping of events due to simultaneous opening of two or more channels did not occur. Probability distribution histograms (Ehrenstein et al. 1974) were constructed from open and closed times, and decay rate constants and amplitudes for distributions were calculated by least-squares regression. The fitting was done between the second bin and the bin in which the number of events decayed to 3 % of the total sample number, as the first bin could be underestimated because of the limited frequency response of the recording system (Yamamoto & Yeh, 1984).

Under our experimental conditions, only one type of channel carrying outward currents was activated by depolarizing voltage steps (Fig. 1A). The outward currents were discernible at −30 mV or at more positive voltages. The channel opened repeatedly throughout the duration of the depolarizing steps (180 ms), producing sporadic bursts of activity separated by long interburst intervals. In the illustrated example, the mean amplitude of the single-channel current at +90 mV was 2·3pA, as estimated by amplitude histogram analysis.

We are not certain whether the channel has subconductance states, because the rapid kinetics of this channel prevented us from distinguishing possible submaximal levels from very short full openings that were damped by the limited frequency response of the recording system.

A typical example of whole-cell currents from a Drosophila neuron is shown in Fig. IB. The whole-cell current rose just after the onset of the depolarizing voltage step, and stayed at that level during depolarization or even increased slightly towards the end of the step. Thus, the outward current did not show any sign of inactivation, at least over the time scale examined. Ensemble averages of singlechannel currents have a similar time course to the whole-cell current (see Fig. 5B).

The amplitude of single-channel currents was temperature-sensitive (Fig. 2). The value of Q₁₀ estimated in two different patches was 1·49 and 1·39 between 10 and 20 °C.

The overall kinetics of the channel was speeded by increasing temperature, and many of the openings had a spike-like appearance at temperatures over 25 °C. The amplitude of the single-channel current in this temperature range was estimated from selected open events with prolonged lifetimes.

The amplitude of single-channel current increased on depolarization. The slope conductance calculated between +10 and + 90 mV was 17·6 ± 3·70pS (mean ± S.E., N = 7) at 10°C, when K⁺ was the major cation in the cytoplasmic side (see Fig. 4). The single-channel current was not seen when K⁺ was totally replaced with Cs⁺ (not shown). Substitution of Cl⁻ in internal solutions by F⁻ or NO₃⁻ had no appreciable effect on the single-channel current (not shown). These results indicate that the single-channel current under study is carried by K⁺. However, smaller outward currents were seen when K⁺ was replaced with either Rb⁺, NH₄⁺ or Na⁺ (Fig. 3). The slope conductances were 8·2 ± 1·7pS with Rb⁺ (N = 3), 9·3 ± l·1 pS with NH₄⁺ (N = 3), and 7·7 ± l·2 pS with Na⁺(N = 3) (Fig. 4).

One question which is of interest in this experiment is whether these permeant cations affect the gating kinetics of the channel. Although channels apparently had longer open times in Na⁺, NH₄⁺ and Rb⁺ solutions (see Fig. 3), the amplitudes of single-channel currents in these solutions were not large enough to permit further investigation of kinetic properties.

Tetraethylammonium (TEA⁺) is an effective blocker of potassium channels in other systems (Armstrong, 1975). We examined whether application of TEA⁺ to the cytoplasmic side of the membrane had effects on the non-inactivating single K⁺ channel currents of Drosophila. TEA⁺ had two effects on the channel. First, it dramatically increased the frequency of occurrence of blank traces which contained no channel openings when the membrane patch was repeatedly depolarized at 0·4 Hz. In a typical experiment, the frequency of occurrence of blank traces increased from 39 % to 88 % of total trials after the addition of TEA⁺ (at 50 mV). A detailed analysis of this action will be presented elsewhere (Yamamoto & Suzuki, 1989). Second, TEA⁺ decreased the amplitude of the single-channel current with minimal effect on its kinetics (Fig. 5A). Fig. 5B illustrates the average current of 32 successive single-channel records in the presence and absence of TEA⁺, and shows that the combination of these two actions of TEA⁺ produced a marked suppression of macroscopic outward currents. The inhibitory actions were readily detected following washing in of TEA⁺, and reversed quickly upon washing out (Fig. 5C).

The probability of a channel being in the open state (P_o) greatly increased as the membrane was depolarized (Fig. 6). P_o never exceeded 0·25 and did not saturate, even at 90 mV.

To characterize kinetic properties of the non-inactivating K⁺ channel, the open-and closed-time histograms were analyzed (Fig. 7). The open-time distribution could be fitted by a single exponential, whereas the closed-time distribution typically required three exponentials to obtain a good fit. The time constant of the open-time distribution as well as the time constants for three components of the closed-time distribution were voltage-dependent (Table 1).

In Drosophila, several single-gene mutants with some defects in K⁺ channel functions have been reported (Salkoff & Tanouye, 1986; Wu & Ganetzky, 1986). Mutations at the Shaker locus alter or reduce the transient outward current mediated by the A channel in larval and adult muscle (Sole et al. 1987). Mutations at the eag locus produce marked reduction of the steady-state outward current with little change in the transient current (Wu et al. 1983a). The single noninactivating potassium channel current could be recorded from neurons dissociated from an eag mutant (Fig. 8). A detailed analysis of conductance as well as gating properties failed to detect any significant difference between the wild-type and the eag mutant. No evidence was obtained for differences in the density of the channel in neuronal membranes of normal and mutant Drosophila.

The present experiments reveal that the ion channel mediating the non-inactivating outward current in Drosophila neurons shares some characteristics with delayed rectifier potassium channels in other systems. Our estimate of the single-channel conductance, 17·6 pS, with 145 mmol l⁻¹ cytoplasmic K⁺, is very close to that reported in frog skeletal muscle (Standen et al. 1985), human T lymphocytes (Cahalan et al. 1985; Bregestovski et al. 1986), turtle colon epithelial cells (Richards & Dawson, 1986), bovine adrenal chromaffin cells (Marty & Neher, 1985) and chick ciliary ganglion neurons (Gardner, 1986). Our experiments indicate that the non-inactivating channel is highly selective for K⁺, showing lower permeability to Rb⁺, NH₄⁺ and Na⁺. Cs⁺ appears not to permeate the channel. The ratio of slope conductances for these cations relative to K⁺ was: K⁺(l) > NH₄⁺(0·53) > Rb⁺(0·47) > Na⁺(0·43) ⪢ Cs⁺(0). A permeability ratio based on the reversal potential measurement could not be determined in our experiments since the single-channel current did not reverse its polarity within a manageable potential range (because the practically impermeant choline ion was used as the external cation). The selectivity sequence, based on permeability ratio, for the squid delayed rectifier K⁺ channel is: K⁺(l) > Rb⁺ > NH₄⁺(0·3) ⪢ Cs⁺, Na⁺ (Conti et al. 1975), and a similar sequence has been reported for delayed rectifier K⁺ channels in several other preparations (see Latorre & Miller, 1983). The discrepancy in the selectivity sequence between the present experiments and previous work, particularly the findings for Na⁺, may arise from the difference in the parameters (i.e. slope conductance vs reversal potential permeability ratio) used to estimate the selectivity sequence (see Eisenman & Horn, 1983). It may be noted that Na⁺ can be punched through the delayed K⁺ channel of squid axons at very positive potentials, carrying significantly large outward currents (French & Shoukimas, 1985).

Another similarity of the non-inactivating K⁺ channel of Drosophila to the delayed rectifier resides in its sensitivity to TEA⁺. The observation that cytoplasmic TEA⁺ reduces the single-channel amplitude is in keeping with previous reports in other preparations (Coronado et al. 1984; Standen et al. 1985). In addition, we have found an increased frequency of traces with no channel openings during repetitive pulses in the presence of TEA⁺ (Yamamoto & Suzuki, 1989).

Voltage-dependent gating of the non-inactivating K⁺ channel was clearly demonstrated (Fig. 6; Table 1). The finding that the closed-time distribution is fitted by three exponential functions indicates that the non-inactivating K⁺ channel possesses at least three distinct closed conformations.

The permeability and gating properties of the non-inactivating K⁺ channel of Drosophila neurons, described in this study, are typical of the delayed rectifier class of K⁺ channel. We did not find any defect in channel function in neurons of eag, which was believed to be a mutant of the delayed rectifier K⁺ channel in Drosophila. The eag mutation causes hyperexcitability of larval presynaptic terminals and reduction of outward currents in the larval muscle fiber (Wu et al. 1983a), suggesting that the channel associated with the eag locus may be present in some axons and terminals of Drosophila neurons. If the channel is confined to this region, this could explain why we did not detect it in the somatic membrane. It is also possible that Drosophila has several types of delayed rectifier K⁺ channel and that our preparation lacks the particular channel type encoded by the eag gene because of inadequate culture conditions. Indeed, cell bodies of neurons cultured by our method appear to lack Na⁺ channels, which are inducible by adding insulin to the culture media (Sole et al. 1986). Alternatively, the eag gene might not be a structural gene encoding the delayed rectifier K⁺ channel protein but, rather, might act as a regulator controlling expression of the channel in vivo.

Recently, Sole & Aldrich (1988) have reported that larval Drosophila neurons possess at least three distinct types of potassium channels. The channel reported here resembles their Ki-channel.

We would like to thank Dr Yoshiki Hotta for providing Canton-S flies, and Dr Alberto Ferrus for eag flies. We also thank Mrs Yasuko Matsumura and Miss Sachiyo Ishikawa for secretarial assistance.