Mode switching characterizes the activity of large conductance potassium channels recorded from rat cortical fused nerve terminals

doi:10.1111/j.1469-7793.1998.733ba.x

Journal List > J Physiol > v.513(Pt 3); Dec 15, 1998

J Physiol. 1998 December 15; 513(Pt 3): 733–747.

doi: 10.1111/j.1469-7793.1998.733ba.x.

PMCID: PMC2231315

Mode switching characterizes the activity of large conductance potassium channels recorded from rat cortical fused nerve terminals

M A Smith and M L J Ashford

Department of Biomedical Sciences, University of Aberdeen, Institute of Medical Sciences, Foresterhill, Aberdeen AB25 2ZD, UK

Corresponding author M. L. J. Ashford: Department of Biomedical Sciences, University of Aberdeen, Institute of Medical Sciences, Foresterhill, Aberdeen AB25 2ZD, UK. Email: mike.ashford/at/abdn.ac.uk

Received June 22, 1998; Accepted September 16, 1998.

This article has been cited by other articles in PMC.

Abstract

Inside-out recordings from rat cortical fused nerve terminals indicate that the most common channel observed was a large conductance K⁺ (BK) channel with characteristics dissimilar to conventional cell body calcium-activated BK (BK_Ca) channels.
BK channels exhibit mode switching between low (mode 1) and high (mode 2) activity, an effect not influenced by membrane voltage. Increasing internal Ca²⁺ concentration increased time spent in mode 2 as did application of protein kinase A, an effect not mimicked by protein kinase C or protein kinase G.
Mode 1 activity was voltage independent although depolarization increased mode 2 channel activity. Global average channel activity was voltage and Ca²⁺ dependent.
Alkaline phosphatase treatment induced channel activity to reside permanently in mode 2, where activity was voltage and Ca²⁺ dependent but unaffected by protein kinases A, G or C.
Internal application of tetraethylammonium blocked BK channel activity in a manner identical to that reported for BK_Ca channels.
These results indicate that nerve terminal membranes have large conductance K⁺ channels with significant differences in gating kinetics and regulation of activity compared with BK_Ca channels of other neuronal preparations. The BK channel subtype may play a unique physiological role specific to the nerve terminal.

The majority of our current knowledge of presynaptic excitability derives from work on non-CNS models. In the mammalian CNS, indirect methods based on quantal analysis of postsynaptic responses have resulted in important insights into factors affecting the probability of transmitter release (Stevens, 1993). However these techniques cannot distinguish between modulation of presynaptic K⁺ or Ca²⁺ channels (Forsythe, 1994). Recently direct recordings from specialized large synapses have been reported: the Calyx of Held in the binaural auditory pathway (Forsythe, 1994; Borst et al. 1995) and the large isolated nerve terminal (synaptosome) of the posterior pituitary (Jackson et al. 1991; Bielefeldt et al. 1992; Kilic et al. 1996). Direct recordings from other isolated mammalian CNS synaptosomes have generally been unsuccessful because of their small size. However, fusion of synaptosome membranes into giant proteoliposomes (Hosokawa et al. 1994) or large lipid bilayers (Farley & Rudy, 1988) produces structures amenable to electrophysiological recordings. Possibly a less invasive method to increase the size of synaptosomes for electrophysiological recordings is to fuse many synaptosomes together (Lee et al. 1995a).

BK_Ca channels are present in presynaptic terminals from several types of synapses (Jackson, 1995) and contribute to the modulation of the nerve terminal action potential (Robitaille & Charlton, 1992). Neurotransmitter release is highly sensitive to nerve terminal Ca²⁺ entry (Augustine et al. 1985), a process regulated by changing the amplitude or duration of the nerve terminal action potential. For example, drugs that block large conductance calcium-activated potassium (BK_Ca) channels prolong the action potential allowing more Ca²⁺ channels to open and for a longer time (Robitaille & Charlton, 1992), consequently enhancing Ca²⁺ entry and increasing transmitter release (Llinás et al. 1981).

BK_Ca channels occur in a wide variety of tissues and act to reduce membrane excitability (Latorre et al. 1989). Their activity is primarily regulated by an increase in intracellular Ca²⁺ and membrane depolarization (Barrett et al. 1982) whereas phosphorylation modulates channel activity. Protein kinase A (PKA) (Farley & Rudy, 1988; Lee et al. 1995b) and protein kinase G (PKG) (Robertson et al. 1993) increase BK_Ca channel activity whereas protein kinase C (PKC) inhibits BK_Ca channel activity in rat pituitary tumour cells (Shipston & Armstrong, 1996), although it reactivates BK_Ca channels after run-down in posterior pituitary nerve terminals (Bielefeldt & Jackson, 1994). High BK_Ca channel activity is maintained by membrane depolarization and micromolar intracellular Ca²⁺, although there are some notable exceptions. Skeletal muscle BK_Ca channels exhibit long shut periods with exposure to high intracellular Ca²⁺ concentrations, behaviour attributed to either pore block by contaminating Ba²⁺ (Neyton, 1996) or Ca²⁺, or due to inherent channel gating (Rothberg et al. 1996). Cloned BK_Ca channels from Drosophila (dslo) expressed in Xenopus oocytes also display large, slow (seconds to minutes) changes in channel activity not described by a simple kinetic model (Silberberg et al. 1996).

In the present study we have recorded the activity of large conductance K⁺ channels from rat cortex fused synaptosomes and compared their properties with cell body BK_Ca channels. Some of the preliminary data have been reported in abstract form (Smith & Ashford, 1997, 1998).

METHODS

In a previous study enzymes, associated with presynaptic membranes such as acetylcholinesterase and the mitochondrial enzyme citrate synthase, were shown to be present in isolated rat cortical synaptosomes (isolated nerve terminals) fused with polyethylene glycol (PEG; Lee et al. 1995a). Briefly, two male Sprague-Dawley rats (50–100 g) were killed by cervical dislocation. Cerebral cortices were rapidly removed and submerged in ice-cold 50 mm Tris-HCl buffer (pH 7.2) containing 320 mm sucrose. The cortices were homogenized in 10 × volume of Tris-HCl buffer using 12 strokes from a Varilab homogenizer (Jencon Scientific Ltd, Leighton Buzzard, UK) at medium speed, and then centrifuged at 1600 g, at 4°C for 10 min. The resultant pellet and supernatant were resuspended, homogenized and then centrifuged, as above. The supernatant was layered onto a discontinuous Percoll gradient (24 %, 10 % and 7.5 % v/v) and centrifuged at 20 000 g, at 4°C for 20 min. The synaptosome layer was removed from the 24/10 % Percoll interface and diluted 1: 5 with warm (37°C) saline (mm: 135 NaCl, 5 KCl, 1 MgCl₂, 1 CaCl₂, 10 Hepes, pH 7.4) plus 10 mm D-glucose. Synaptosomes were centrifuged at 3000 g for 7 min at room temperature (20–25°C), the resultant pellet was resuspended in saline (37°C) plus glucose (final volume of 300 μl) and then incubated at 37°C for 5 min. Fusion was achieved by mixing 1450 molecular weight PEG (Sigma) with the synaptosomes (65 %, w/v final volume) at 37°C over 1 min and then left for a further minute. The synaptosome suspension was gradually diluted with 20 ml of saline over 6 min and then centrifuged at 4500 g, at 37°C for 2 min. The pellet was resuspended in saline plus glucose and plated on 35 mm diameter (Nuncon, Denmark) culture plates at room temperature for 1 h. Fused synaptosomes that did not adhere to the base of the culture plates were removed by superfusion with saline before recordings were made.

Inside-out patches were made from fused synaptosomes (between 2 and 10 μm in diameter) using patch electrodes (5–10 MΩ) backfilled with electrode solution (mm: 140 KCl, 1 MgCl₂, 1 CaCl₂, 10 Hepes buffer, pH 7.2). Voltage-clamp recordings were obtained with an Axopatch 200A amplifier (Axon Instruments); recordings were filtered at 2 kHz and stored on DAT tape for off-line analysis. Voltage steps, generated using a computer running Vgen 2.7 software (J. Dempster, University of Strathclyde, UK), were dispatched to the amplifier via a DT2801A digital-to-analog converter (Data Translation, Malboro, MA, USA). Inside-out patches were held at 0 mV and ± 20, ± 40 or ± 60 mV alternate 2 s voltage steps were applied 30 times per voltage protocol. The standard bath solution contained (mm): 140 KCl, 1 MgCl₂, 9.75 CaCl₂, 10 EGTA and 10 Hepes buffer (pH 7.2, 10 μm free Ca²⁺). To examine calcium sensitivity, free Ca²⁺ was changed by varying the CaCl₂ concentration (calculated by the ‘METLIG’ program; P. England & R. Denton, University of Bristol, UK). Potassium selectivity was examined by replacement of the bath solution with a solution containing 40 mm KCl and 100 mm NaCl. All chemicals, except protein kinase G (PKG, isoform 1α) and the catalytic subunit of protein kinase C (PKC, isolated from rat brain) which were acquired from Calbiochem, were purchased from Sigma. cGMP, ATP (Mg²⁺ salt), catalytic subunit of protein kinase A (PKA), PKG, PKC, tetraethylammonium (TEA), BaCl₂, 4-aminopyridine (4-AP) and alkaline phosphatase (Type V) were dissolved in distilled H₂O stocks whereas ATP was dissolved in 5 mm Hepes buffer (pH 7.2). All stocks except TEA, 4-AP and BaCl₂, which were kept at 4°C, were stored at −20°C and added at the appropriate concentrations to the bathing solution by a gravity flow superfusion system.

Vcan 4.0 (J. Dempster) was used to analyse currents from the voltage steps, whereas patches held at constant voltages were analysed by Pat 6.6 (J. Dempster) and a group of programs written in Axobasic (Dr N. W. Davis, University of Leicester, UK). Pre-recorded currents were played back through an analog-to-digital translator (DT2801A, Data Translation, for Pat 6.6 and Vcan 4.0 or DigiData 1200, Axon Instruments, for Axobasic programs), digitized at 10 kHz using Pat 6.6 and Axobasic programs or 300 Hz using Vcan 4.0. Single channel records were replayed into an Easygraph TA240 recorder (Gould) for illustrative purposes. The potential across the membrane is described following the usual sign convention for membrane potential (i.e. inside negative). All single channel recordings were carried out at room temperature (20–25°C). All data in the text are presented as mean values ± standard error of the mean (s.e.m.) and statistical significance was determined at the 95 % level of confidence using either analysis of variance (ANOVA) or Student's t test. Open-state probability (P_o) was calculated from total patch current (I) divided by the sum of the single channel current amplitudes (i) and number of functional channels (N_f), as determined from current-frequency histograms. However, in many patches N_f could not be accurately determined and therefore a derivation of P_o was used as a measurement of channel activity (N_fP_o), where I was divided by i.

RESULTS

Channel conductance and ion selectivity

Inside-out patch recordings (n = 157) were obtained from fused cortical isolated nerve terminals (between 2 and 10 μm in diameter) of male Sprague-Dawley rats. In 67 patches recorded under symmetrical (140 mm) KCl conditions, a large conductance channel was observed, characterized by a linear current-voltage (I–V) relationship (over the ± 60 mV range) and a mean slope conductance of 204.3 ± 7.4 pS (n = 23; Fig. 1). Under asymmetrical potassium conditions (i.e. 140 mm KCl in the pipette and 40 mm KCl, 100 mm NaCl in the bath) single channel currents displayed inward rectification at positive potentials and the I–V relationship closely followed that predicted by the Goldman-Hodgkin- Katz (GHK) constant current equation for this potassium gradient (Fig. 1B). Extrapolated reversal potentials were −1.1 ± 0.6 mV (n = 23), and +25.7 ± 2.3 (n = 7) for symmetrical and asymmetrical K⁺ gradients, respectively. The calculated Nernst potentials for the symmetrical and asymmetrical gradients are 0 and +34 mV, respectively. These data demonstrate that these large conductance (BK) channels selectively conduct potassium over sodium. In the majority of patches containing BK channels two other main channel types were also observed. In patches devoid of BK channels, these two channel types were shown to be non-selective cation channels, characterized by mean slope conductances of 50.2 ± 4.9 pS (n = 51) and 24.2 ± 3.3 pS (n = 39), respectively. These latter channels are not considered further in the present paper.

Figure 1

Single channel current-voltage relationship of the BK channel

Channel kinetics

A particularly striking feature of BK channel currents recorded from patches excised from rat fused cortical synaptosomes was the non-stationary behaviour of the channel kinetics. In all excised patches examined and maintained at a constant voltage in the presence of 10 μm intracellular Ca²⁺ the BK channels exhibited long periods (many seconds, up to several minutes) where channel open-state probability (P_o) was either high (P_o > 0.10, generally values were greater than 0.3) or extremely low (P_o < 0.05) in which case the kinetics were extremely fast and flickery (Fig. 2A). Such behaviour observed for other channel types (e.g. L-type Ca²⁺ channels) has been termed ‘mode switching’ (Catterall & Streissnig, 1992). This behaviour was observed throughout the lifetime of patches (up to 1 h) and furthermore there was no indication of decreased channel activity (‘run-down’) with time. Originally three modes were proposed to described the kinetics of this channel, such that activity was defined as a closed mode, a low activity mode (P_o < 0.2) and a high activity mode (P_o > 0.2; Smith & Ashford, 1997). However, in a recent study we show that, in concentrations of intracellular calcium required to activate this channel, only the high and low activity modes are apparent (Smith & Ashford, 1998). Binomial distribution can be used to ascertain whether channels, within a single membrane patch with equivalent open-state probabilities, open independently of each other (Blatz & Magleby, 1986):

A mathematical equation, expression, or formula that is to be displayed as a block (callout) within the narrative flow. The name of referred object is tjp0513-0733-m1.jpg

(1)

where n is the number of channels in the patch, r is the number of channels open, and p is the probability that any given channel is open. Consequently it was determined whether nerve terminal BK channels followed a binomial distribution by examining the percentage of time, P_o, that 0, 1 or 2 etc. channels were open in multichannel patches. In a set of experiments where two observable functional channels were present at +40 mV in 10 μm intracellular Ca²⁺, the percentage time the channels resided in the closed state, or a single, or a double open state did not follow a binomial distribution (n = 4). For example, channels with a P_o of 0.46 spent 7.69 % of time in the closed state, 92.26 % of time in the single open state and 0.05 % of time in the double open state which compares with binomial percentage predictions of 29.00 % of time in the closed state, 49.70 % of time in the single open state and 21.30 % of time in the double open state. The simplest explanation for this type of behaviour is that the average P_o of each channel in the patch was different, due to one of the channels residing in the high activity mode and the other in the low activity mode. It is therefore proposed that nerve terminal BK channels switch between low and high activity modes which will now be referred to as mode 1 and mode 2, respectively.

Figure 2

Mode switching of the nerve terminal BK channel

In three excised patches maintained at −20 mV in 10 μm free intracellular Ca²⁺, where only one observable functional channel was present, analysis of channel open-state probability was undertaken during mode 1 and 2 sojourns. In mode 1 the open-state probability was 0.01 ± 0.01 which was statistically different from the value for P_o, 0.43 ± 0.15, obtained for mode 2 (paired t test; P < 0.05). In order to determine whether the appearance of high levels of channel activity occurred in a random manner or were clustered in defined periods, due to mode switching, the data were examined for statistical correlation in the patterns of gating by using the procedures of ‘runs’ analysis (Horn et al. 1984) and Fisher's exact test (Plummer & Hess, 1991). The runs test was used as follows:

(2)

where Z is a normal random variable with mean 0 and standard deviation 1, R is the observed number of runs, N is the total number of time fragments (trials) examined and p is the probability of finding one of the two kinds of elements (i.e. the number of trials in mode 2). Inside-out patches held at −20 mV in 10 μm intracellular Ca²⁺, with only one observable functional channel present, were used to measure P_o in consecutive 15 s trials over 10 or 20 min (Fig. 2B). Each trial was classified as residing in a low activity (P_o < 0.05, mode 1) or high activity (P_o > 0.05, mode 2) state. Runs were defined as a succession of one or more identical trials followed and preceded by the alternate activity. The expected number of runs is 2Np(1 - p) if the switching between modes is random. From the observed number of runs, Z was evaluated as shown in eqn (2). As clustering of mode 2 activity was observed, a one-sided test for randomness was used, with a value of Z < −2.3 taken as significant (P < 0.01). The number of runs observed was much less than the number expected from a completely random process, the mean Z statistic being −4.39 ± 0.47 (n = 4). A large and negative value of Z is indicative of the BK channel modal activity in successive time intervals being significantly grouped. Further analysis of these data using Fisher's exact test where pairs of consecutive time intervals are classified according to their gating pattern into the four possible combinations (1–1, 1–2, 2–1, 2–2), resulted in probability values of less than 0.001 for each of the patches indicating non-random association of modes.

On more detailed analysis of single channel kinetics the mean closed but not open times were also demonstrated to be statistically different between modes 1 and 2 (paired t test; P < 0.02 and P < 0.04, respectively). Mean closed times were 428 ± 110 ms in mode 1 and 6.24 ± 2.67 ms in mode 2 whereas mean open times (corrected for missed events) were 2.03 ± 1.12 ms and 8.86 ± 6.25 ms in modes 1 and 2, respectively. Open and closed duration histograms from mode 2 were best fitted by the sum of three and four exponentials, respectively (Fig. 2C), with best-fit values of 0.36 ± 0.13, 11.45 ± 8.81 and 23.61 ± 3.06 ms for open times and 0.10 ± 0.02, 0.98 ± 0.25, 26.23 ± 10.39 and 241.88 ± 101.58 ms for closed times. It was not possible to acquire the required number of events for detailed analysis of mode 1 kinetics as long enough recordings of patches containing a single BK channel could not be obtained.

Voltage sensitivity

In order to determine an estimate of the voltage dependence of global channel activity, (N_fP_o), and for the channel activity associated with the individual modes, a voltage protocol covering a 5 min period was used which held inside-out patches (with 10 μm intracellular Ca²⁺) at 0 mV for 6 s with alternate 2 s duration voltage steps (in the range of −60 to +60 mV, in 20 mV steps) 30 times per protocol. Individual currents from voltage steps were defined as mode 1 (N_fP_o less than 0.1 over the step duration) or mode 2 (N_fP_o greater than 0.1 over the step duration). The percentage of voltage steps that reside either in mode 1 or 2 were not influenced by membrane voltage (Fig. 3A and C). However, when the global average channel activity was calculated from all the voltage steps over each voltage protocol and from steps only in mode 1 or 2, it was found that the global average channel activity and the channel activity associated with mode 2 were dependent on membrane voltage, with depolarization causing an increase in mean channel activity (Fig. 3A and B). In contrast mean activity in mode 1 was independent of membrane voltage. To obtain V_0.5 values (the membrane potential at which N_fP_o is half the maximum fitted value of N_fP_o), the change in N_fP_o observed when membrane voltage was depolarized was plotted in accordance with the Boltzmann equation for a 2-state model. The mean V_0.5 for the global average channel activity was −29.8 ± 16.6 mV (n = 5) and was −34.1 ± 5.1 mV (n = 5) when fitted to mode 2 mean channel activity. The mean channel activity calculated for steps only in mode 1 could not be plotted with the Boltzmann equation because there was no statistical (ANOVA; P > 0.25) change in activity over the membrane voltages used (± 60 mV range). The V_0.5 values obtained from both average and mode 2 channel activity in 10 μm internal Ca²⁺ contrast with values generally reported for large conductance calcium-activated (BK_Ca) channels which have a V_0.5 ranging between +11 and +16 mV in similar concentrations of internal Ca²⁺ (Latorre et al. 1989). However, BK_Ca channels in anterior pituitary cells have been reported to have a V_0.5 of −30 mV in 1 μm intracellular Ca²⁺ (Wong & Adler, 1986).

Figure 3

Channel activity is voltage dependent as a result of mode 2

Calcium sensitivity

The intracellular calcium dependence of nerve terminal BK channel activity was examined. Mean global channel activity in 10 and 100 μm internal Ca²⁺ displayed voltage dependence, with depolarization inducing an increase in channel activity. The BK channel activity observed was calcium dependent with activity greater (paired t test; P < 0.01) in 100 μm compared with 10 μm for all voltages examined, although the mean V_0.5 values were not statistically (paired t test) different; −24.6 ± 9.1 mV (n = 5) and −33.5 ± 9.1 (n = 5) for 10 and 100 μm Ca²⁺, respectively (Fig. 4B). However, at all voltages tested (± 40 mV range), application of 1 μm internal Ca²⁺ statistically reduced channel activity (paired t test; P < 0.03). In all Ca²⁺ concentrations examined, the global average channel activity did not increase significantly beyond a membrane potential of +20 mV (up to +60 mV, data not shown). However, the maximal attainable value for mean activity decreased (ANOVA; P < 0.002) with a reduction in internal Ca²⁺. It is possible that the reduction of mean channel activity by decreasing internal Ca²⁺ was simply due to its influence on mode switching, such that intracellular Ca²⁺ altered the frequency of modes 1 and 2. This is shown in Fig. 4C where it can be seen that increasing the concentration of calcium increased the proportion (paired t test; P < 0.04) of voltage steps showing mode 2 behaviour. Thus the reduction in maximally attainable channel activity with lower concentrations of intracellular Ca²⁺ is, at least partially, due to less time spent in mode 2 and more time in mode 1. The global average channel activity plotted against intracellular Ca²⁺ at +40 and −40 mV (Fig. 4D) shows that depolarization induced a greater degree of channel activity over the calcium range examined. However, the concentration of Ca²⁺ required to half-maximally activate the BK channel (EC₅₀) was not dependent on membrane voltage, with a mean EC₅₀ of 11.6 μm at +40 mV and 11.4 μm at −40 mV (n = 5). These data contrast with published values of between 1 nm and 1 μm Ca²⁺ (at similar voltages) for BK_Ca channels (Latorre et al. 1989). In contrast, the Hill slope for the Ca²⁺ activation curve was +1.6 at +40 mV and +3.8 at −40 mV, values within the published range for BK_Ca channels (Latorre et al. 1989), indicating that the nerve terminal BK channel binds a similar number of Ca²⁺ molecules to somatic BK_Ca channels (Lee et al. 1995b).

Figure 4

BK channel activity is dependent on intracellular Ca²⁺

Consequently these data indicate that the cortical nerve terminal BK channels have significantly different properties, particularly with respect to channel kinetics and calcium and voltage dependence, from BK_Ca channels generally described in other tissues.

Alkaline phosphatase abolishes modal switching

Alkaline phosphatase (10 U ml⁻¹) applied to the internal aspect of the isolated inside-out patch caused channel activity to increase abruptly within 5 min of application and this increased level of activity was sustained for periods up to 1 h (Fig. 5A and B), an effect which was not reversed by washing out the phosphatase (n = 22). For example, in patches stepped to a voltage of −20 mV in 10 μm intracellular Ca²⁺, the mean open-state probability was 0.01 ± 0.001 and 0.40 ± 0.13 for modes 1 and 2, respectively, and following 10 U ml⁻¹ alkaline phosphatase treatment the open-state probability was 0.43 ± 0.14 (n = 5) with no evidence of mode 1 activity discerned. Once activated by alkaline phosphatase treatment, the kinetics of the BK channel were compared with those observed during sojourns to mode 2 of BK channels in untreated patches. In three inside-out patches containing a single BK channel, maintained at −20 mV in 10 μm intracellular Ca²⁺, the mean open-state probability following alkaline phosphatase treatment was 0.43 ± 0.24 (Fig. 5B). Analysis of these patches gave values for mean open (corrected for missed events) and closed times of 5.34 ± 1.90 and 37.61 ± 15.54 ms, respectively. Open and closed duration histograms were best fitted by the sum of three and four exponentials, respectively (Fig. 5C), with best fitting values of 0.78 ± 0.58, 8.22 ± 3.47 and 17.35 ± 1.60 ms for open times and 0.14 ± 0.03, 0.69 ± 0.23, 22.00 ± 11.75 and 148.55 ± 30.62 ms for closed times. The mean values for P_o and the kinetic data were statistically (unpaired t test) identical to those obtained for the BK channel exhibiting mode 2 behaviour prior to alkaline phosphatase. Thus application of alkaline phosphatase converts channel activity exclusively into mode 2 type behaviour and not into a previously unobserved activity mode. Furthermore, using the voltage step protocols described above, channel activity resided in mode 2 for all voltages (± 60 mV) examined (n = 22).

Figure 5

Modal channel activity is irreversibly abolished by internal alkaline phosphatase treatment

By converting channel activity exclusively into mode 2 with alkaline phosphatase, the voltage and calcium sensitivity of BK channel activity could be examined without the complication induced by mode switching. Therefore voltage steps in the presence of different concentrations of internal Ca²⁺ were used to determine channel activity following alkaline phosphatase treatment. Nerve terminal BK channel activity was dependent on membrane potential, with membrane depolarization increasing N_fP_o (Fig. 6A and B). Membrane voltage was plotted against N_fP_o and fitted to the Boltzmann equation to obtain V_0.5 values of +63.1 ± 12.4 mV (n = 5) in 3 μm, −24.8 ± 19.5 mV (n = 5) in 10 μm and −20.4 ± 14.9 mV (n = 5) in 100 μm intracellular Ca²⁺. However, in contrast to global BK channel activity prior to alkaline phosphatase treatment where openings were observed (particularly in multichannel patches) in the presence of 1 μm internal Ca²⁺, following alkaline phosphatase, application of 1 μm internal Ca²⁺ abolished all channel activity at the voltages tested (± 60 mV range). Although the V_0.5 values in 10 and 100 μm Ca²⁺ contrast with those obtained for rat cortical cell body BK_Ca channels (Lee et al. 1995b) they statistically match (unpaired t test) the V_0.5 values obtained for the channel when residing in mode 2 prior to alkaline phosphatase treatment (−25 and −33 mV in 10 and 100 μm Ca²⁺, respectively). This indicates that the voltage sensitivity of the nerve terminal BK channel was not affected by the alkaline phosphatase treatment although reduction in intracellular Ca²⁺ concentration to low micromolar levels caused the Boltzmann curve to move rightwards (Fig. 6B) such that membrane depolarization could partially restore channel activity. One possible explanation for this difference is that after alkaline phosphatase treatment the channel exclusively remained in voltage-dependent mode 2 and consequently membrane depolarization could restore channel activity caused by the reduction in intracellular Ca²⁺. However, channel activity prior to alkaline phosphatase was not restored by membrane depolarization because the reduction in internal Ca²⁺ caused the channel to reside for a greater proportion of time in the voltage-independent mode 1. On examining BK channel activity at various intracellular Ca²⁺ concentrations for voltages of +40 and −40 mV (Fig. 6C) it can be seen that at both voltages the channel displayed identical sensitivity with mean values for EC₅₀ of 4.1 μm at +40 mV and 5.0 μm at −40 mV with Hill slopes of 3.2 and 10.8 at +40 and −40 mV, respectively. Consequently, in comparison with global mean channel activity prior to alkaline phosphatase, the BK channel displayed an enhanced sensitivity to calcium (compare Fig. 4B and D with Fig. 6B and C), again this was most likely to be due to the lack of mode 1 sojourns. These data indicate that the nerve terminal BK channel is inherently Ca²⁺ dependent when in mode 2 configuration, and consequently the decrease in global BK channel activity induced by reduction of intracellular Ca²⁺ is not solely due to an increased proportion of time spent in mode 1.

Figure 6

Channel activity after alkaline phosphatase treatment is voltage and calcium dependent

The role of phosphorylation on channel gating

Several studies have shown that BK_Ca channel activity can be modulated by the phosphorylation state of the channel (Farley & Rudy, 1988; Robertson et al. 1993; Bielefedt & Jackson, 1994; Lee et al. 1995b; Shipston & Armstrong, 1996). We therefore investigated the role of kinases on BK channel activity in cortical fused nerve terminals.

In 10 μm Ca²⁺, application of 2 mm Mg-ATP (n = 4) to the internal membrane surface did not statistically (paired t test) alter nerve terminal global BK channel activity compared with controls (data not shown) at all voltages tested (± 40 mV range), nor was there any significant alteration in the proportion of time the channel resided in modes 1 and 2. These data indicate that a protein kinase may not be present in the isolated nerve terminal membrane patch unlike that observed for the BK_Ca channel of cortical neurone somata (Lee et al. 1995b) or the BK_Ca channel reconstituted from rat brain and incorporated into lipid bilayers (Reinhart et al. 1991). However, 2 mm Mg-ATP in the presence of 12 U ml⁻¹ of the catalytic subunit of protein kinase A (PKA, n = 4) did increase global channel activity (Fig. 7A) by increasing the proportion of time spent in mode 2 (paired t test, P < 0.02; Fig. 7B). The increase in channel activity was voltage dependent with statistical significance reached at +20 mV (paired t test; P < 0.02) and +40 mV (paired t test; P < 0.005). The V_0.5 value for controls was not determined because of the lack of significant voltage dependence; however, a V_0.5 of 7.71 ± 2.36 mV was calculated following PKA treatment. After alkaline phosphatase treatment, 2 mm ATP in the presence of 12 U ml⁻¹ PKA did not change channel activity when in 3 μm (n = 3) or 10 μm intracellular Ca²⁺ (n = 4; Fig. 7C). For example, V_0.5 values for controls were −8.8 ± 25.1 and 46.1 ± 23.1 mV for 10 and 3 μm internal Ca²⁺, respectively, which compares with −6.0 ± 28.2 and 54.2 ± 11.2 mV for 10 and 3 μm internal Ca²⁺, respectively, when in the presence of PKA. In addition, channel activity did not change (paired t test) from controls in the presence of 2 mm Mg-ATP, with 200 U ml⁻¹ protein kinase G (PKG) and 200 μm cGMP or, with 0.05 U ml⁻¹ protein kinase C (PKC) before and after alkaline phosphatase treatment (n = 4; data not shown).

Figure 7

Protein kinase A (PKA) increases channel activity by increasing the proportion of time in mode 2

Tetraethylammonium channel blockade

In order to determine whether the BK channel described above displayed a similar pharmacological profile to somatic BK_Ca channels, we decided to examine the actions of a variety of non-specific K⁺ channel blockers. Unfortunately it was not possible to obtain outside-out patches from the fused nerve terminals to test the actions of externally applied agents (e.g. iberiotoxin). Consequently the actions of barium, 4-aminopyridine (4-AP) and tetraethylammonium (TEA) were examined on BK channel activity when applied to the intracellular aspect of the membrane.

Application of 10 mm 4-AP had no effect on channel activity or amplitude (n = 4), whereas the presence of 5 mm Ba²⁺ resulted in an overall reduction of global channel activity by 86 ± 1 % (n = 3). The addition of TEA (10 mm) caused a slight reduction in single channel amplitude, most noticeable during sojourns in mode 2 activity (n = 6). It is well documented that the majority of BK_Ca channels are blocked by high (20–70) millimolar intracellular concentrations of TEA (Latorre et al. 1989). Therefore the actions of internally applied TEA were examined in more detail and to allow easier analysis channels were converted to permanent residence in mode 2 by pre-treating inside-out patches with 10 U ml⁻¹ alkaline phosphatase. TEA at concentrations in the range 10–100 mm reduced the single channel current amplitude in a concentration-dependent manner (n = 4; Fig. 8A). A plot of the relationship between intracellular TEA concentration and single channel amplitude (i_TEA/i_C), determined as the current amplitude in the presence of TEA (i_TEA) divided by control amplitude (i_C) when at a membrane potential of −20 mV, resulted in a dissociation constant (K_d) of 67.1 mm (Fig. 8B). In addition, to assess the voltage dependence of TEA block, data were analysed by the Woodhull (1973) method (Fig. 8C):

(3)

where i_C is the single channel amplitude, i_TEA is the single channel amplitude in the presence of 50 mm TEA, z isthe valency of TEA, δ is the location of TEA block within the membrane electric field, F is Faraday's constant, V is membrane voltage, R is the gas constant, T is absolute temperature and K_d(0mV) is the dissociation constant at 0 mV. A plot of ln[(i_C/i_TEA) − 1] against membrane voltage yielded an estimated K_d(0mV) of 153 mm and a δ of 0.098, as calculated from the y-intercept and the gradient, respectively. These data show that the BK channel is blocked weakly by internal TEA in a relatively voltage-independent manner, such that the blocking site was only 10 % within the electric field of the membrane. Thus the data pertaining to internal TEA block of BK channels are consistent with those reported for BK_Ca channels in various other tissues (Latorre et al. 1989).

Figure 8

Single BK channels are blocked by internal tetraethylammonium (TEA)

DISCUSSION

Nerve terminal large conductance potassium channel

The nerve terminal large conductance (BK) channel reported in this study has a single channel slope conductance of 204 pS in symmetrical 140 mm KCl, which compares well with the unitary conductance of 230 pS for BK_Ca channels recorded from cortical cell bodies (Lee et al. 1995b) and other tissues (130–300 pS; Latorre et al. 1989) under similar ionic conditions. In addition, channel currents closely followed the Goldman-Hodgkin-Katz (GHK) current constants for a ‘pure potassium electrode’ when in both symmetrical and asymmetrical potassium conditions, thus indicating that the channel selectively passes K⁺ ions.

Modal switching

Many types of ion channel display changes in their gating behaviour (modes) that occur on time scales from milliseconds to minutes. Ligand-gated channels such as acetylcholine receptors show long silent periods due to desensitization (Sakmann et al. 1980). Voltage-gated Cl⁻ channels from rat skeletal muscle switch between a normal mode containing the majority of openings and closures, a buzz mode made up of brief bursts of rapid channel openings and closures, and a shut mode where no openings are observed for periods of seconds to minutes (Blatz & Magleby, 1986). Voltage-gated L-type Ca²⁺ channels also display clear transformations in kinetic activity, a property modulated by activators and inhibitors of this channel (Catterall & Striessnig, 1992). Voltage-gated cation channels from Aplysia neurones also switch between high activity and ‘buzz’ modes (Wilson & Kaczmarek, 1993).

In the present study we demonstrate that BK channels present in rat cortical fused nerve terminals display modal activity characterized by sudden and reversible transitions between periods of low or high channel open-state probability, which we have termed modes 1 and 2, respectively. Modal switching behaviour has not been observed for BK_Ca channel activity recorded from cell bodies of rat acutely dissociated cortical neurones (Lee et al. 1995b), although modal switching was reported as a rare event in BK_Ca channel activity recorded from cultured hippocampal neurones (Wann & Richards, 1994). It has been reported that BK_Ca channel activity from rat brain synaptosomes incorporated into lipid bilayers display long periods (seconds) of channel closures (Farley & Rudy, 1988) which may be reminiscent of mode switching. In addition, K⁺ channel activity recorded from fused Torpedo synaptosomes, which exhibited the kinetic properties of an A-type K⁺ channel, also displayed mode switching in a voltage-dependent manner (Butkevich et al. 1997). The reasons why these channels undergo sudden transitions in their kinetic behaviour is not well understood, but there may be physiological mechanisms that contribute to or control this activity. For example, modal switching in voltage-dependent cation channels from Aplysia neurones can be modulated by dephosphorylation of channel tyrosine residues by a protein kinase A-regulated tyrosine phosphatase (Wilson & Kaczmarek, 1993). For example, the proportion of time that BK channels spent in mode 2 was increased by incubation with PKA or by increasing the concentration of intracellular Ca²⁺. However, unlike the A-type K⁺ channel in Torpedo synaptosomes (Butkevich et al. 1997), mode switching of BK channel activity in the present study is independent of membrane voltage such that the frequency of either mode is the same at all voltages examined. The addition of alkaline phosphatase and consequent likely removal of a crucial phosphate group from the BK channel irreversibly switches its gating into mode 2. Moreover, protein kinases A, G and C do not reverse this effect, indicating that the site of phosphate group cleavage may either require a different type or isoform of kinase to restore modal activity, or alternatively, the site is not available for kinase action. These data constrast with phosphorylation- dephosphorylation studies reported previously. In rat pituitary nerve terminals, BK_Ca channel activity runs down following patch excision; a process thought to be mediated by enzymatic dephosphorylation of the channel, such that channel phosphorylation by PKC restores channel activity following run-down (Bielefeldt & Jackson, 1994). In addition BK_Ca channels from rat brain incorporated in lipid bilayers display two types of BK_Ca channel which are characterized by different channel kinetics, charybdotoxin (ChTX) sensitivity and sensitivity to PKA and protein phosphatase 2A (PP-2A, Reinhart et al. 1991). For example type 1 BK_Ca channels display ‘fast gating kinetics’, are blocked by ChTX, and channel activity is activated by PKA and reduced by PP-2A. In contrast, type 2 BK_Ca channels display ‘slow channel gating kinetics’, are ChTX insensitive and channel activity is reduced by PKA but reversed by PP-2A.

However, the physical nature of mode switching may not be due to dephosphorylation and phosphorylation of the channel. Mode switching occurs in the absence of ATP, and is not influenced by addition of ATP alone. Consequently an alternative explanation for mode switching is that an integral part of the protein structure acts to occlude ion flow, i.e. a ‘ball and chain’ mechanism of pore block (Toro et al. 1992). Such a mechanism may be consistent with the observation that during sojourns in mode 1, the channel displays fast ‘flickery-type’ kinetics, particularly at hyperpolarizing potentials (Hicks & Marrion, 1998). The ability of alkaline phosphatase to switch channel activity irreversibly into mode 2 suggests that, if a ‘ball and chain’ mechanism does cause mode switching, some intrinsic phosphate group is essential for this action.

Voltage and calcium sensitivity

The sequence of the pore forming subunit of BK_Ca channels, cloned and expressed from human brain (Tseng-Crank et al. 1994), indicates that BK_Ca channels are structurally related to voltage-gated K⁺ channels. Cloned and native BK_Ca channel activity is dependent on membrane potential and intracellular Ca²⁺, with depolarization and micromolar intracellular Ca²⁺ required for activity (Barrett, 1982; Tseng-Crank et al. 1994; Lee et al. 1995b). Nerve terminal BK channel activity is also dependent on membrane depolarization and intracellular Ca²⁺, consistent with this channel being a member of the BK_Ca channel family. This is further supported by the preliminary experiments describing the pharmacology of the BK channel which displayed sensitivity to intracellular Ba²⁺ and TEA and insensitivity to 4-AP. In particular, the sensitivity of BK channels to internal TEA was indistinguishable from that described for BK_Ca channels. However, the membrane voltage required for half-maximal activation of nerve terminal BK channels before and after alkaline phosphatase treatment is approximately −30 mV in 10 μm intracellular Ca²⁺, which is significantly more negative than for BK_Ca channels from other neuronal preparations and tissues (Latorre et al. 1989; Jackson, 1995). Thus, the BK channel present in rat cortex nerve terminals requires less depolarization for activation than BK_Ca channels. Furthermore, nerve terminal BK channels, before and after alkaline phosphatase treatment, require higher concentrations of intracellular Ca²⁺ (approximately 5–10 μm) for channel activation than BK_Ca channels from neuronal cell bodies and peptidergic nerve terminals (0.1–1 μm Ca²⁺) at similar membrane voltages (Reinhart et al. 1989; Bielefeldt & Jackson, 1994; Lee et al. 1995b). At present it is unclear why the nerve terminal BK channel should exhibit gating properties clearly different from that displayed by BK_Ca channels. One possible explanation is the presence of multiple BK_Ca channel isoforms expressed in brain (Tseng-Crank et al. 1994). It has been suggested that the C-terminal region of the BK_Ca channel is important in calcium gating (Wei et al. 1994). Furthermore, cloning and expression studies of human brain BK_Ca channels have shown that there are four sites within the C-terminal where RNA splicing occurs and preliminary studies indicate significant differences in their calcium sensitivity (Tseng-Crank et al. 1994). Alternative RNA splicing of dslo has also been reported and when expressed, channels display kinetic behaviour reminiscent of mode switching (Silberberg et al. 1996). Consequently, it is possible that the nerve terminal BK channel is an alternative splice variant of the cell body BK_Ca channel.

Many native BK_Ca channels consist of two distinct subunits, the pore-forming unit (α-subunit) and a smaller β-subunit (Garcia-Calvo et al. 1994). Co-expression studies demonstrate that the presence of the β-subunit enhances the calcium sensitivity of the channel. Thus the nerve terminal BK channel may result from a particular combination of α-subunit splice variant and β-subunit.

Physiological significance

The activation of presynaptic BK_Ca channels reduces the amplitude, duration and frequency of the nerve terminal action potential (Robitaille & Charlton, 1992; Sah, 1996). Neurotransmitter release is highly sensitive to nerve terminal Ca²⁺ entry (Augustine et al. 1985) via voltage-gated calcium channels, the opening and closing of which are closely associated with changes in the nerve terminal potential. Consequently alteration of the nerve terminal action potential properties would be expected to modulate neurotransmitter release. Thus when the nerve terminal BK channel is in mode 1 it would be expected that the action potential duration, amplitude and frequency will be greater in comparison with that observed when the channel is in mode 2. Therefore, whilst predominately in voltage-independent mode 1 operation, BK channels would be expected to facilitate high and sustained calcium influx and thus contribute to increased neurotransmitter release, whereas during voltage-dependent mode 2 properties, BK channels would facilitate a reduced and shortened calcium influx and thus decreased neurotransmitter release.

The resting concentration of internal Ca²⁺ in nerve terminals is between 20 and 60 nm (Yawo & Chuhma, 1993), thus at rest it is likely that the nerve terminal BK channel is not active. Calyces of Held, loaded with the Ca²⁺ indicator Calcium Green-5N, show a rise in intracellular Ca²⁺ within 30 ms following stimulation by a single action potential (Borst et al. 1995). Although the intracellular Ca²⁺ concentration was not quantified, their data indicate intracellular Ca²⁺ rises to micromolar concentrations. This is consistant with a similar evoked rise in intracellular Ca²⁺ measured in chick ciliary ganglion loaded with the Ca²⁺ indicator fura-2 (Larkum et al. 1994). Consequently the rise in intracellular Ca²⁺ following a single stimulus may be sufficient to stimulate BK channel activation, and regulate action potential duration and hence the amount of neurotransmitter released. However, it may be more likely that the high level of intracellular Ca²⁺ required to activate BK channels will be achieved preferentially during high frequency action potential activity in the terminal. Indeed evidence indicates that BK_Ca channels are co-localized with voltage-gated Ca²⁺ channels in presynaptic microdomains (Gola & Crest, 1993), allowing close functional coupling of Ca²⁺ entry and BK_Ca channels. Such a mechanism may be important in limiting the global calcium transient within the terminal and consequent transmitter release and may also be of pathophysiological significance, counteracting the generation of action potentials associated with increased nerve traffic during epileptic seizures. If the BK channel described in this study does play such roles in the control of cortical neurone presynaptic excitability, an understanding of the physiological mechanisms by which BK channel activity is controlled, in particular mode switching, may provide fundamental insights to central neurotransmission.

Acknowledgments

We thank The Wellcome Trust (grant no. 042726) for financial support and technical assistance from Dr N. W. Davis (University of Leicester, UK). M. A. S. is a Wellcome Trust Prize student (grant no. 048618).

References

Augustine, GJ; Charlton, MP; Smith, SJ. Calcium entry and transmitter release at voltage-clamped nerve terminals of squid. The Journal of Physiology. 1985;367:163–181. [PubMed]
Barrett, JN; Magleby, KL; Pallotta, BS. Properties of single calcium-activated potassium channels in cultured rat muscle. The Journal of Physiology. 1982;331:211–230. [PubMed]
Bielefeldt, K; Jackson, MB. Phosphorylation and dephosphorylation modulate a Ca²⁺-activated K⁺ channel in rat peptidergic nerve terminals. The Journal of Physiology. 1994;475:241–254. [PubMed]
Bielefeldt, K; Rotter, JL; Jackson, MB. Three potassium channels in rat posterior pituitary nerve terminals. The Journal of Physiology. 1992;458:41–67. [PubMed]
Blatz, AL; Magleby, KL. Quantitative description of three modes of activity of fast chloride channels from rat skeletal muscle. The Journal of Physiology. 1986;378:141–174. [PubMed]
Borst, JGG; Helmchen, F; Sakmann, B. Presynaptic and postsynaptic whole-cell recordings in the medial nucleus of the trapezoid body of the rat. The Journal of Physiology. 1995;489:825–840. [PubMed]
Butkevich, A; Ohana, O; Meir, A; Rahamimoff, R. Voltage dependent switch in the activity mode of the K⁺ channel in presynaptic nerve terminals. NeuroReport. 1997;8:2539–2545. [PubMed]
Catterall, WA; Striessnig, J. Receptor sites for Ca²⁺ channel antagonists. Trends in Pharmacological Sciences. 1992;13:256–262. [PubMed]
Farley, J; Rudy, B. Multiple types of voltage-dependent Ca²⁺-activated K⁺ channels of large conductance in rat brain synaptosomal membranes. Biophysical Journal. 1988;53:919–934. [PubMed]
Forsythe, ID. Direct patch recording from identified presynaptic terminals mediating glutamatergic EPSCs in the rat CNS, in vitro. The Journal of Physiology. 1994;479:381–387. [PubMed]
Garcia-Calvo, M; Knaus, H-G; McManus, OB; Giangiacomo, KM; Kaczorowski, GJ; Garcia, ML. Purification and reconstruction of the high-conductance calcium-activated potassium channel from tracheal smooth muscle. Journal of Biological Chemistry. 1994;269:676–682. [PubMed]
Gola, M; Crest, M. Colocalization of active K_ca and Ca²⁺ channels within Ca²⁺ domains in Helix neurones. Neuron. 1993;10:689–699. 10.1016/0896-6273(93)90170-V. [PubMed]
Hicks, GA; Marrion, NV. Ca²⁺-dependent inactivation of large conductance Ca²⁺-activated K⁺ (BK) channels in rat hippocampal neurones produced by pore block from an associated particle. The Journal of Physiology. 1998;508:721–734. [PubMed]
Horn, R; Vandenberg, CA; Lange, K. Statistical analysis of single sodium channels - Effects of N-bromoacetamide. Biophysical Journal. 1984;45:323–335. [PubMed]
Hosokawa, Y; Sandri, G; Panfili, E; Cherubini, E. Characterization of a voltage-dependent anionic channel in fused synaptosomes isolated from rat hippocampi. Neuroscience Letters. 1994;169:167–170. 10.1016/0304-3940(94)90382-4. [PubMed]
Jackson, MB. Presynaptic excitability. International Review of Neurobiolology. 1995;38:201–251.
Jackson, MB; Konnerth, A; Augustine, GJ. Action potential broadening and frequency-dependent facilitation of calcium signals in pituitary nerve terminals. Proceedings of the National Academy of Sciences of the USA. 1991;88:380–384. [PubMed]
Kilic, G; Stolpe, A; Lindau, M. A slowly activating voltage-dependent K⁺ current in rat pituitary nerve terminals. The Journal of Physiology. 1996;497:711–725. [PubMed]
Larkum, ME; Warren, DA; Bennett, MR. Calcium concentration changes in the calyciform nerve terminal of the avian ciliary ganglion after tetanic stimulation. Journal of the Autonomic Nervous System. 1994;46:175–188. 10.1016/0165-1838(94)90035-3. [PubMed]
Latorre, R; Oberhauser, A; Labarca, P; Alvarez, O. Varieties of calcium-activated potassium channels. Annual Review of Physiology. 1989;51:385–399. 10.1146/annurev.ph.51.030189.002125. [PubMed]
Lee, K; Dixon, AK; Rowe, ICM; Ashford, MLJ; Richardson, PJ. Direct demonstration of sulphonylurea-sensitive K_ATP channels on nerve terminals of the rat motor cortex. British Journal of Pharmacology. 1995a;115:385–387. [PubMed]
Lee, K; Rowe, ICM; Ashford, MLJ. Characterization of an ATP-modulated large conductance Ca²⁺-activated K⁺ channel present in rat cortical neurones. The Journal of Physiology. 1995b;488:319–337. [PubMed]
Llinás, R; Steinberg, IZ; Walton, K. Presynaptic calcium currents in squid giant synapse. Biophysical Journal. 1981;33:289–321. [PubMed]
Neyton, J. A Ba²⁺ chelator suppresses long shut events in fully activated high conductance Ca²⁺-dependent K⁺ channels. Biophysical Journal. 1996;71:220–226. [PubMed]
Plummer, MR; Hess, P. Reversible uncoupling of inactivation in N-type calcium channels. Nature. 1991;351:657–659. 10.1038/351657a0. [PubMed]
Reinhart, PH; Chung, S; Levitan, IB. A family of calcium-dependent potassium channels from rat brain. Neuron. 1989;2:1031–1041. 10.1016/0896-6273(89)90227-4. [PubMed]
Reinhart, PH; Chung, S; Martin, BL; Brautigan, DL; Levitan, IB. Modulation of calcium-activated potassium channels from rat brain by protein kinase A and phosphatase 2A. Journal of Neuroscience. 1991;11:1627–1635. [PubMed]
Robertson, BE; Schubert, R; Hescheler, J; Nelson, MT. cGMP-dependent protein kinase activates Ca-activated K channels in cerebral artery smooth muscle cells. American Journal of Physiology. 1993;265:C299–303. [PubMed]
Robitaille, R; Charlton, MP. Presynaptic calcium signals and transmitter release are modulated by calcium-activated potassium channels. Journal of Neuroscience. 1992;12:297–305. [PubMed]
Rothberg, BS; Ricardo, AB; Song, L; Magleby, KL. High Ca²⁺ concentrations induce a low activity mode and reveal Ca²⁺-independent long shut intervals in BK channels from rat muscle. The Journal of Physiology. 1996;493:673–689. [PubMed]
Sah, P. Ca²⁺-activated K⁺ currents in neurones: types, physiological roles and modulation. Trends in Neurosciences. 1996;19:150–154. 10.1016/S0166-2236(96)80026-9. [PubMed]
Sakmann, B; Patlak, J; Neher, E. Single acetylcholine-activated channels show burst-kinetics in presence of desensitising concentrations of agonist. Nature. 1980;286:71–73. [PubMed]
Shipston, MJ; Armstrong, DL. Activation of protein kinase C inhibits calcium-activated potassium channels in rat pituitary tumour cells. The Journal of Physiology. 1996;493:665–672. [PubMed]
Silberberg, SD; Lagrutta, A; Adelman, JP; Magleby, KL. Wanderlust kinetics and variable Ca²⁺-sensitivity of dslo, a large conductance Ca²⁺-activated K⁺ channel, expressed in oocytes. Biophysical Journal. 1996;70:2640–2651. [PubMed]
Smith, MA; Ashford, MLJ. Modal activity of a large conductance potassium channel recorded from rat motor cortex fused nerve terminals. The Journal of Physiology. 1997;504.P:160–161P.
Smith, MA; Ashford, MLJ. Alkaline phosphatase irreversibly abolishes modal activity of large conductance Ca²⁺-activated K⁺ channels in rat cortical fused nerve terminals. The Journal of Physiology. 1998;509.P:63P.
Toro, L; Stefani, E; Latorre, R. Internal blockade of a Ca(2+)-activated K⁺ channel by Shaker B inactivating ‘ball’ peptide. Neuron. 1992;9:237–245. 10.1016/0896-6273(92)90163-8. [PubMed]
Tseng-Crank, J; Foster, CD; Krause, JD; Mertz, R; Godinot, N; DiChaira, TJ; Reinhart, PH. Cloning, expression, and distribution of functionally distinct Ca²⁺-activated K⁺ channel isoforms from human brain. Neuron. 1994;13:1315–1330. 10.1016/0896-6273(94)90418-9. [PubMed]
Wann, KT; Richards, CD. Properties of single calcium-activated potassium channels of large conductance in rat hippocampal neurons in culture. European Journal of Neuroscience. 1994;6:607–617. [PubMed]
Wei, A; Solaro, C; Lingle, C; Salkoff, L. Calcium sensitivity of BK-type K_ca channels determined by a separable domain. Neuron. 1994;13:671–681. 10.1016/0896-6273(94)90034-5. [PubMed]
Wilson, GF; Kaczmarek, LK. Mode-switching of a voltage-gated cation channel is mediated by protein kinase A-regulated tyrosine phosphatase. Nature. 1993;366:433–438. 10.1038/366433a0. [PubMed]
Wong, BS; Adler, M. Tetraethylammonium blockade of calcium-activated potassium channels in clonal anterior pituitary cells. Pflügers Archiv. 1986;407:279–284.
Woodhull, AM. Ionic blockage of sodium channels in nerve. Journal of General Physiology. 1973;61:687–708. [PubMed]
Yawo, H; Chuhma, N. Preferential inhibition of ω-conotoxin-sensitive presynaptic Ca²⁺ channels by adenosine autoreceptors. Nature. 1993;365:256–258. [PubMed]