Article Information

• PDF (214 kb)
• Full Text with Large Figures

PubMed

• Articles by Birgit Hirschberg
• Articles by Neil V. Marrion

Related Articles

• Neurotransmitter Modulation of Small-Conductance Ca-Activate...

  Neurotransmitter Modulation of Small-Conductance Ca-Activated K Channels by Regulation of Ca Gating Neuron, Volume 59, Issue 3, 14 August 2008, Pages 439-449 François Maingret, Bertrand Coste, Jizhe Hao, Aurélie Giamarchi, Duane Allen, Marcel Crest, David W. Litchfield, John P. Adelman and Patrick Delmas Summary Small-conductance Ca-activated K (SK) channels are widely expressed in neuronal tissues where they underlie postspike hyperpolarizations, regulate spike-frequency adaptation, and shape synaptic responses. SK channels constitutively interact with calmodulin (CaM), which serves as Ca sensor, and with protein kinase CK2 and protein phosphatase 2A, which modulate their Ca gating. By recording coupled activities of Ca and SK2 channels, we showed that SK2 channels can be inhibited by neurotransmitters independently of changes in the activity of the priming Ca channels. This inhibition involves SK2-associated CK2 and results from a 3-fold reduction in the Ca sensitivity of channel gating. CK2 phosphorylated SK2-bound CaM but not KCNQ2-bound CaM, thereby selectively regulating SK2 channels. We extended these observations to sensory neurons by showing that noradrenaline inhibits SK current and increases neuronal excitability in a CK2-dependent fashion. Hence, neurotransmitter-initiated signaling cascades can dynamically regulate Ca sensitivity of SK channels and directly influence somatic excitability. Summary | Full Text | PDF (1466 kb)

• Inwardly Rectifying Current-Voltage Relationship of Small-Co...

  Inwardly Rectifying Current-Voltage Relationship of Small-Conductance Ca-Activated K Channels Rendered by Intracellular Divalent Cation Blockade Biophysical Journal, Volume 80, Issue 5, 1 May 2001, Pages 2207-2215 Heun Soh and Chul-Seung Park Abstract Small conductance Ca-activated K channels (SK channels) are a group of K-selective ion channels activated by submicromolar concentrations of intracellular Ca independent of membrane voltages. We expressed a cloned SK channel, rSK2, in oocytes and investigated the effects of intracellular divalent cations on the current-voltage (I-V) relationship of the channels. Both Mg and Ca reduced the rSK2 channel currents in voltage-dependent manners from the intracellular side and thus rectified the I-V relationship at physiological concentration ranges. The apparent affinity of Mg was changed as a function of both transmembrane voltage and intracellular Ca concentration. Extracellular K altered the voltage dependence as well as the apparent affinities of Mg binding from intracellular side. Thus, the inwardly rectifying I-V relationship of SK channels is likely due to the voltage-dependent blockade of intracellular divalent cations and that the binding site is located within the ion-conducting pathway. Therefore, intracellular Ca modulates the permeation characteristics of SK channels by altering the I-V relationship as well as activates the channel by interacting with the gating machinery, calmodulin, and SK channels can be considered as Ca-activated inward rectifier K channels. Abstract | Full Text | PDF (236 kb)

• Sodium Permeability of a Cloned Small-Conductance Calcium-Ac...

  Sodium Permeability of a Cloned Small-Conductance Calcium-Activated Potassium Channel Biophysical Journal, Volume 89, Issue 5, 1 November 2005, Pages 3111-3119 Narae Shin, Heun Soh, Sunghoe Chang, Do Han Kim and Chul-Seung Park Abstract Small-conductance Ca-activated potassium channels (SK channels) are heteromeric complexes of pore-forming main subunits and constitutively bound calmodulin. SK channels in neuronal cells are activated by intracellular Ca that increases during action potentials, and their ionic currents have been considered to underlie neuronal afterhyperpolarization. However, the ion selectivity of neuronal SK channels has not been rigorously investigated. In this study, we determined the monovalent cation selectivity of a cloned rat SK channel, rSK2, using heterologous expression and electrophysiological measurements. When extracellular K was replaced isotonically with Na, ionic currents through rSK2 reversed at significantly more depolarized membrane potentials than the value expected for a Nernstian relationship for K. We then determined the relative permeability of rSK2 for monovalent cations and compared them with those of the intermediate- and large-conductance Ca-activated K channels, IK and BK channels. The relative permeability of the rSK2 channel was determined as , indicating substantial permeability of small ions through the channel. Although a mutation near the selectivity filter mimicking other K-selective channels influenced the size-selectivity for permeant ions, Na permeability of rSK2 channels was still retained. Since the reversal potential of endogenous SK current is determined by Na permeability in a physiological ionic environment, the ion selectivity of native SK channels should be reinvestigated and their in vivo roles may need to be restated. Abstract | Full Text | PDF (234 kb)

• …more

Copyright © 1999 The Biophysical Society. All rights reserved.
Biophysical Journal, Volume 77, Issue 4, 1905-1913, 1 October 1999

doi:10.1016/S0006-3495(99)77032-3

Channels, Receptors, and Transporters

Gating Properties of Single SK Channels in Hippocampal CA1 Pyramidal Neurons

Birgit Hirschberg^*, 1, James Maylie^#, John P. Adelman^* and Neil V. Marrion^#, ,  

^* Vollum Institute, Oregon Health Sciences University, Portland, Oregon 97201 USA
^# Department of Obstetrics and Gynecology, Oregon Health Sciences University, Portland, Oregon 97201 USA

Address reprint requests to Neil V. Marrion, Department of Pharmacology, University of Bristol, School of Medical Sciences, University Walk, Bristol BS8 1TD, United Kingdom. Tel.: +44-117-928-7636; Fax: +44-117-925-0168.

¹ Birgit Hirschberg's present address is Merck & Co., RY80N-C42, P.O. Box 2000, Rahway NJ 07065-0900.

The slow afterhyperpolarization (AHP) in hippocampal neurons has an important influence on membrane excitability. It is activated after a burst of action potentials and underlies spike frequency adaptation, terminating burst firing (Madison and Nicoll, 1984). The slow AHP is generated by activation of small conductance calcium-activated potassium channels (SK), resulting from the entry of Ca²⁺ through voltage-gated Ca²⁺ channels (Lancaster and Nicoll, 1987). It has been assumed that multiple SK channels may exist because the AHP in some cell types is sensitive to the bee venom toxin, apamin (e.g., bullfrog sympathetic neurons, Pennefather et al) and is not in others (e.g., rat hippocampal CA1 pyramidal neurons, Lancaster and Adams, 1986). This assumption has been confirmed by the cloning of three distinct members of the SK channel family, SK1–3 (Köhler et al,Joiner et al). Both the apamin-insensitive SK1 and the apamin-sensitive SK2 channels exhibited a single channel conductance of ∼10pS in isotonic potassium and were half-activated by 0.6–0.7μM Ca²⁺ (Köhler et al,Hirschberg et al). These data are in agreement with the single SK channel properties observed in GH₃ anterior pituitary cells (Lang and Ritchie, 1987), T lymphocytes (Grissmer et al), and adrenal chromaffin cells (Park, 1994). However, the only study of single SK channels in hippocampal pyramidal neurons reported a larger single channel conductance of 18–20pS and a slightly lower sensitivity to Ca²⁺ [open probability (P_o) 0.5 with 1μM Ca²⁺; designated P(o) in figures] (Lancaster et al).

A significantly greater Ca²⁺ sensitivity of SK channels might be expected from intracellular Ca²⁺ imaging studies in hippocampal neurons, since it has been shown that during the slow AHP the somatic Ca²⁺ concentration increases only to 0.1μM (Knöpfel et al). However, the P_o of SK channels at the peak of the slow AHP has been estimated to be 0.4–0.6 (Sah and Issacson, 1995,Valiante et al), predicting an intracellular Ca²⁺ concentration of 0.5–1μM. It is possible that neither the channel described previously in hippocampal neurons (Lancaster et al) nor a homolog of the recombinant SK1–3 underlies the slow AHP in hippocampal neurons. Alternatively, hippocampal SK channels may only be activated by local increases in Ca²⁺ that are not reflected by the somatic measurements.

We have examined the gating properties of the hippocampal SK channel. By using inside-out membrane patches excised from acutely isolated hippocampal CA1 pyramidal neurons, we have determined that hippocampal SK channels exhibit a conductance of 10pS (in symmetrical potassium solutions), are half-activated by ∼0.6μM Ca²⁺, and display voltage-independent gating. In addition, hippocampal SK channels display nonstationary kinetics. These properties are very similar to cloned rSK2 and hSK1 channels (Köhler et al,Hirschberg et al) and suggest that activation of SK channels underlying generation of the slow AHP occurs by a local rise of submembrane Ca²⁺ to levels higher than those measured in the soma.

Acutely dissociated hippocampal CA1 neurons were obtained as described previously (Cloues et al). Briefly, Sprague-Dawley rats (9–14 days old) were anesthetized with halothane and decapitated. Hippocampi were rapidly dissected and cut into 300–400-μm slices. Slices were incubated at 37°C in a dissociation solution containing (in mM): Na₂SO₄, 82; K₂SO₄, 30; HEPES, 10; MgCl₂, 5; ethylene glycol bis (b-aminoethyl ether)-N, N,N′,N′-tetraacetic acid (EGTA), 1 (pH 7.4), with added protease type XXIII (3mg/ml) for 7–8min and bubbled with O₂. Tissue slices were then transferred to a solution containing trypsin inhibitor (1mg/ml) and bovine serum albumin (BSA) (1mg/ml) for 1min and finally rinsed in dissociation solution containing no enzyme. The CA1 region was microdissected and triturated into Falcon Primaria dishes as needed.

Cells were washed and superfused (15ml min⁻¹) with an external solution containing (in mM): KMeSO₄, 125; KCl, 35; HEPES(Na), 10; EGTA, 10; CaCl₂, 5.64 (to give an estimated free concentration of 0.06μM (Fabiato and Fabiato, 1979) (pH 7.4 with KOH). Cells in this solution had ∼0mV membrane potential. All potentials are expressed as the negative of the potential imposed on the pipette. Membrane patches were first excised to the inside-out patch configuration into the Mg²⁺-free superfusion solution containing 0.1μM Ca²⁺. The free Ca²⁺ concentration was raised up to 3μM (Fabiato and Fabiato, 1979) once the patch had stabilized. Changes in solutions containing different concentrations of free Ca²⁺ were achieved by bath superfusion and were complete within several seconds.

Excised inside-out patch recordings (Hamill et al) were made using thick-walled (1.5mm O.D., 0.5mm I.D.) quartz electrodes (7–10MΩ) containing the external solution described above, supplemented with 100nM charybdotoxin to prevent contamination by BK channel openings. Voltage-dependent potassium channel activity was prevented by including 4-aminopyridine (1mM), 3,4-diaminopyridine (1mM), α-dendrotoxin (200nM), and β-dendrotoxin (200nM) in the pipette solution. Single channel currents were recorded with an Axopatch 200 amplifier using a CV201A headstage (Axon Instruments, Foster City, CA), filtered at 1–4kHz with an 8-pole Bessel filter (Frequency Devices, Haverhill, MA) and acquired at 100-μs intervals for analysis using Pulse (Heka, dist. by Instrutech Corp., New York, NY) onto a Quadra 650 (Apple Computer, Cupertino, CA). No differences in open and closed times were seen when data were filtered either at 1kHz or 4kHz. Single channels were analyzed using MacTAC (Bruxton Corp., Seattle, WA, dist. by Instrutech Corp., NY). The “50% threshold” technique was used to estimate event amplitudes and durations, with each transition inspected visually before being accepted. Open and closed duration histograms were constructed with MacTacfit (Bruxton Corp., distributed by Instrutech Corp.), binned logarithmically (20 bins/decade) and plotted against the square root transformation of the ordinate (number of events/bin). The distribution was fitted by a sum of exponential probability density functions using the maximum-likelihood method. With this type of representation, peaks in the histogram correspond to the time constant of the exponential (Sigworth and Sine, 1987). A correction was made for the rise time of the filter (Colquhoun and Sigworth, 1983) and all bins were used for fitting. The number of statistically significant components was determined by the method of maximum likelihood ratios (Horn and Lange, 1983). Data were not corrected for missed events. Channel open probability (P_o) was estimated as NP_o, the product of the open probability times the number of channels. NP_o was calculated using ReadEvents v1.37 (Scott Eliasof, Portland, OR), as the sum of (dwell time×level number) divided by the total time. N was estimated as the number of simultaneously open channels at a P_o>0.5. Finally, P_o was obtained by dividing NP_o by N. Where applicable, values were expressed as mean±SD, and P values were derived from unpaired two-tailed Student's t-tests. Results were considered significantly different at P<0.01. Oocyte expression and recordings of cloned hSK1 channels were performed as described previously for rSK2 (Hirschberg et al).

We observed a variable degree of channel “run-down” in the absence of any treatment. In the presence of a fixed concentration of Ca²⁺, ∼30% of patches exhibited SK channels whose activity was lost within the first minute of recording. Of the remaining patches, ∼70% exhibited SK channels whose activity was lost within the first 10min of recording. Loss of activity occurred during high and low P_o behavior and was abrupt, complete, and irreversible; i.e., no channel openings were observed over minutes even in increased [Ca²⁺] or after removal and reapplication of Ca²⁺. Generally, there was no change in gating leading up to the loss of activity. The loss of SK channel activity in excised inside-out patches may suggest that channel activity requires cytoplasmic factors. All reagents were obtained from Sigma, except α- and β-dendrotoxin (Alomone, Israel), CaCl₂ (Fluka, NY), and HEPES (Calbiochem, CA).

In ∼90 of 600 patches, application of Ca²⁺ to the inner face of inside-out membrane patches caused activation of a small-amplitude channel. Increasing the concentration of Ca²⁺ from 0.1 to 1μM caused a progressive increase in channel P_o (Figure 1A). A plot of P_o as a function of Ca²⁺ concentration was described by the Hill equation, with an EC₅₀ of 0.56μM and a Hill coefficient (n_H) of 4.6 (Figure 1B). This observed Ca²⁺-dependence of channel P_o agrees well with the macroscopic Ca²⁺-dependence of both hSK1 and rSK2 (EC₅₀ values 0.7 and 0.63μM and n_H values of 3.9 and 4.8, respectively; Köhler et al).

Display large version of this figure

Figure 1

Ca²⁺-dependence of hippocampal SK channel activity. (A) Current traces from an inside-out patch excised from a hippocampal CA1 neuron and held at −60mV. As the Ca²⁺ concentration bathing the patch was raised from 0.1 to 1μM, SK channel activity increased. Channel openings are shown as downward deflections. (B) Open probabilities calculated from the patch shown in A are plotted as a function of the Ca²⁺ concentration bathing the patch. Least-squares fitting of the data to the Hill equation yielded NP_omax 0.74, EC₅₀ 0.56μM, and n_H 4.6, where NP_omax is the open probability in saturating Ca²⁺ concentrations.

The voltage-dependence of channel P_o was determined in the presence of a fixed Ca²⁺ concentration. Figure 2A shows traces from a patch recorded at three membrane potentials containing a single SK channel activated by 1μM Ca²⁺. Channel gating was not obviously dependent on membrane potential, with P_o values being similar over a membrane potential range of −100 to +60mV (Figure 2B). Measurement of the single channel amplitude at each voltage gave rise to the current-voltage relationship shown (Figure 2B, inset), yielding a slope conductance of 9.8 pS. Determination of slope conductance in six patches gave a mean value of 10.1±0.5 pS. The Ca²⁺ sensitivity, voltage-independence, and single channel conductance of this channel were not significantly different from those obtained from both hSK1 and rSK2 clones (see Köhler et al and Hirschberg et al), identifying it as an SK channel.

Display large version of this figure

Figure 2

Hippocampal SK channel activity as a function of membrane potential. (A) In the presence of 1μM Ca²⁺, SK channel activity is shown at the illustrated membrane potentials. Openings are upward at +60mV and downward at negative potentials. As is apparent, SK channel activity was not markedly voltage-sensitive. (B) SK channel open probability is plotted as a function of membrane potential for the patch shown in A. The solid line is fit by linear regression analysis, showing little voltage-dependence to SK channel P_o. Inset: single channel amplitudes obtained by the fitting of a single Gaussian function to amplitude histograms constructed from the patch shown in A are plotted as a function of membrane potential. Linear regression analysis yielded a slope conductance of 9.8 pS.

The decay of the slow AHP in hippocampal pyramidal neurons is insensitive to membrane potential (Lancaster and Adams, 1986). This may arise from the open-state kinetics of the underlying SK channel being insensitive to voltage, as occurs with cloned rSK2 channels (Hirschberg et al). In the presence of a fixed concentration of Ca²⁺, the P_o of hippocampal SK channels was voltage-independent (Figure 2B). Figure 3A shows open duration histograms at −60mV and +60mV constructed from a single channel patch bathed in 1μM Ca²⁺. Both open duration histograms were best fit by the sum of two exponentials with similar time constants (Figure 3A). The time constant of each exponential component is shown as a function of membrane potential in Figure 3BC, with each symbol reflecting data from one patch. Data from hippocampal neurons are shown as closed symbols, and data from cloned hSK1 channels expressed in Xenopus oocytes are shown for comparison (open symbols). The long and short open-time constants seen with hippocampal SK channels were indistinguishable from those obtained with hSK1, and both open-time constants were independent of voltage. In addition, voltage did not significantly affect the relative contribution of each exponent to the open duration histogram (Figure 3D).

Display large version of this figure

Figure 3

Voltage-independence of SK channel gating. (A) Open-time histograms from a patch exposed to 1μM Ca²⁺ and held either at −60mV (top) or at +60mV (bottom). Histograms were fit to the sum of two exponentials with illustrated time constants. (B) Time constants of the short open-time component are shown as a function of membrane potential for two patches excised from hippocampal CA1 neurons, one bathed in 1μM Ca²⁺ (closed circles) and the other in 0.6μM Ca²⁺ (closed squares). For comparison, open symbols represent data from three patches excised from Xenopus oocytes expressing recombinant hSK1 channels and bathed in 0.2μM Ca²⁺. The lines in panels B and C were drawn for display purposes only. (C) Time constants representing the long open-time component for the same patches as in B. (D) Data represent the percentage of openings contributing to the short open-time constant for the patches also used in B and C. For each patch the percentage of short openings was essentially voltage-independent. The absolute values differ from patch to patch due to differences in Ca²⁺ concentration and P_o behavior (see text).

rSK2 channels have been shown to exhibit and switch between two dominant modes of gating, low and high P_o behavior (Hirschberg et al). The open-time distribution of rSK2 channels exhibiting either behavior was best fit by the sum of two exponentials, with a larger fraction of short-duration events occurring during low P_o (Hirschberg et al). However, the main factor determining the observance of low P_o was the presence and weighting of a very slow closed-time component (Hirschberg et al). During high P_o activity this closed time was absent, being replaced by a closed time an order of magnitude faster (Hirschberg et al).

rSK2 channels were found to switch P_o behaviors most frequently in Ca²⁺ concentrations close to the EC₅₀ (Hirschberg et al). In 8 of 10 patches exposed to 0.6μM Ca²⁺ and three of three patches exposed to 1μM Ca²⁺, hippocampal SK channels were observed to spontaneously switch P_o behaviors during the time of recording. This is illustrated in Fig. 4. During the first 2min of recording in the presence of 1μM Ca²⁺, a patch containing a single SK channel exhibited relatively high P_o (Figure 4AB). Analysis of open and closed times showed that the majority of openings were of long duration (τ=6.3 and 1.2ms), with closed times being best described by the sum of two exponentials with time constants of 0.94 and 5.7ms (Figure 4C). After ∼114s of recording, SK channel activity switched to an extremely low P_o behavior (Figure 4AB). Observed openings were rare and of short duration, being described by a single exponential of time constant 0.64ms (Figure 4D, left). In contrast to high P_o behavior, the closed-time distribution was best fit by the sum of three exponentials (τ=1.7, 5.6, and 208ms) (Figure 4D, right). As with cloned rSK2 channels, the short and intermediate closed times were similar to those seen during high P_o behavior with the appearance of an additional very long closed time. After ∼60s of low P_o behavior, SK channel activity abruptly switched back to high P_o behavior (Figure 4AB). In three patches bathed in 1μM Ca²⁺, average P_o values were 0.67±0.13 for the high P_o behavior and 0.03±0.02 for the low P_o behavior. Therefore, in the presence of a fixed concentration of Ca²⁺, hippocampal SK channels can spontaneously and rapidly switch between two P_o behaviors, a property observed with cloned rSK2 channels (Hirschberg et al).

Display large version of this figure

Figure 4

Switches between two gating behaviors. (A) Current traces from a patch containing a single SK channel bathed in 1μM Ca²⁺, at a holding potential of −60mV. The parallel lines mark a break in the display during low P_o activity; 114s after patch excision, 25s after the start of the recording (downward arrow), channel activity spontaneously and abruptly decreased to a very low open probability. The rare openings were all of short duration. Channel activity abruptly increased back to its previous level 46s later (upward arrow). (B) Stability plot for the patch in A showing the open probability during 1-s intervals as a function of time after patch excision. The arrows indicate the abrupt switch between high P_o activity and low P_o behavior (see A). (C) Open- (left) and closed-time (right) histograms assembled from the two periods of high open probability for the recording in A. The open-time histogram was fit by two exponentials with time constants 6.3 and 1.2ms. The closed-time histogram was fit with two exponentials, with most closures being of very short duration. (D) Open- and closed-time histograms representing the period of low open probability shown in A and B. The open-time histogram was fit by a single exponential with a time constant of 0.7ms, and the closed-time histogram was fit by the sum of three exponentials with time constants 1.7ms, 5.6ms, and 208ms.

It has been reported that rSK2 channels spent more time in high P_o activity as the Ca²⁺ concentration was increased (Hirschberg et al). A similar relationship was observed with hippocampal SK channels. In the presence of 0.3μM Ca²⁺, SK channel activity was of very low P_o throughout, with only brief sojourns to an intermediate value (Figure 5A). Open- and closed-time analysis revealed that the majority of openings were of short duration (τ=0.8 and 6.5ms) and the closed-time distribution was best described by the sum of three exponentials (τ=1.1, 5.1, and 711ms) (Figure 5A). Increasing the concentration of Ca²⁺ bathing the patch to 3μM evoked predominantly high P_o activity (Figure 5B). The majority of openings were of long duration (τ=12.6ms) with a minor component of short-duration openings (τ=0.9ms). Most closures were described by exponentials with short and intermediate time constants (τ=0.68 and 3.3ms), with a few events of longer time constant (τ=150ms) corresponding to the very brief period of low P_o behavior shown in Figure 5B (asterisk; Figure 5B). Therefore, high and low P_o behaviors can be observed over a range of Ca²⁺ concentrations, and raising the Ca²⁺ concentration promotes high P_o SK channel gating.

Display large version of this figure

Figure 5

Ca²⁺-dependence of SK channel gating. (Ai) Open probability during 1-s intervals as a function of time during recording for a patch held at −60mV and bathed in 0.3μM Ca²⁺. At this concentration of Ca²⁺, SK channel activity was of low P_o behavior throughout. (Aii) Open- (left) and closed-time (right) histograms obtained from the recording illustrated in Ai. The open-time histogram was fit with the sum of two exponentials, with most openings being of short duration (τ=0.8ms and 6.5ms). The closed-time histogram was fit to the sum of three exponentials with time constants 1.1ms, 5.1ms, and 711ms. The low P_o activity was characterized by the presence of the long closed-time component in the histogram. (Bi) Open probability during 1-s intervals for the same patch as in A (holding potential −60mV), after the concentration of Ca²⁺ was increased to 3μM Ca²⁺. A very brief switch to low open probability gating was seen ∼26s into the recording (asterisk). (Bii) Open- and closed-time histograms obtained from channel activity promoted by 3μM Ca²⁺. The open-time histogram was fit with two exponentials with time constants 0.9 and 12.6ms. In contrast to low P_o activity, the closed-time histogram was fit to the sum of three exponentials, but with most of the closures being of short duration (τ=0.68ms and 3.3ms). The presence of the small number of events contributing to the slowest component (τ=150ms) likely arises from the short period of low P_o activity (see above).

A Ca²⁺-dependence of the long closed time is apparent from histograms shown in Figure 3 and Figure 4 and Figure 5. In contrast, the time constants describing channel openings were similar in different Ca²⁺ concentrations, and only their relative contribution varied. Hippocampal SK channel kinetics were determined over a range of 0.3 to 3μM Ca²⁺, and data from cloned hSK1 channels are shown as the open symbols for comparison (Fig. 6). The open-time constant of hippocampal SK channels did not change significantly over this range of Ca²⁺, and was similar to that seen with hSK1 (Figure 6AB). The Ca²⁺-dependence of channel P_o was in part due to Ca²⁺-dependent changes in the relative fractions of long and short openings. Figure 6C shows that the percentage of openings described by the short time constant exponent decreased monotonically as the Ca²⁺ concentration was increased.

Display large version of this figure

Figure 6

Effect of Ca²⁺ concentration on SK channel open- and closed-times. (A) Increasing the concentration of Ca²⁺ from 0.3 to 3μM had no obvious effect on the magnitude of the short open-time constant in patches excised from hippocampal neurons (closed circles), and values were similar to those seen in patches from Xenopus oocytes expressing recombinant hSK1 channels (open circles). Hippocampal patches were held at −60mV, and oocyte patches at −80mV. Lines in all panels were drawn for display purposes only and were not fit to the data. (B) Increasing the Ca²⁺ concentration had no apparent effect on the magnitude of the long open-time constant (same patches as A). (C) The increase in SK channel P_o promoted by increasing the Ca²⁺ concentration is partly the result of a loss of short-duration openings. As the concentration of Ca²⁺ was increased from 0.3 to 3μM, the percentage of short-duration openings decreased monotonically. (D) Raising the concentration of Ca²⁺ bathing patches excised from hippocampal CA1 neurons had no effect on the magnitude of either the short (circles) or intermediate (squares) closed-time components. (E) The Ca²⁺-dependence of SK channel P_o likely results from the reduction in the duration and number of long closures. Plotted in E is the magnitude of the long closed-time as a function of Ca²⁺ concentration, showing that the time constant describing these events decreased monotonically as the Ca²⁺ concentration was increased. (F) As the concentration of Ca²⁺ was increased, the number of long closures decreased (triangles), while the percentage of short-duration closures diametrically increased (circles).

The time constants reflecting short and intermediate closures were not dependent on Ca²⁺ concentration (Figure 6D), while the magnitude of the time constant describing long-duration closures was obviously sensitive to the concentration of Ca²⁺. The time constant describing long closures progressively declined as the concentration of Ca²⁺ was increased (Figure 6E). As the magnitude of this time constant decreased, the number of described closings also declined, with a concomitant increase in the number of short-duration closures (Figure 6F). These data are very similar to the Ca²⁺-dependence of open and closed states observed for rSK2 channels (Hirschberg et al). Therefore, the increase in SK channel P_o observed on increasing the concentration of Ca²⁺ is attributed to a decrease in the number and lifetime of long closures and an increase in the number of long-duration openings.

The conductance, Ca²⁺ sensitivity, and voltage-independent properties of SK channels in hippocampal CA1 pyramidal neurons were very similar to those observed for cloned apamin-sensitive rSK2 channels (Hirschberg et al). However, hippocampal SK channels are insensitive to the bee venom toxin, apamin (Lancaster and Adams, 1986). We have initiated a study of the properties of the apamin-insensitive cloned hSK1 channel. We have observed a single channel conductance of 9.2±0.3pS (N=3) in isotonic potassium, a value not statistically different from that observed with hippocampal SK channels (P>0.01; Köhler et al). hSK1 channels were half-activated by 0.7±0.06μM Ca²⁺, with an n_H value of 3.9±0.45 (Köhler et al), values very similar to hippocampal SK channels (see text). Combining data from both P_o behaviors, mean open-time constants for hippocampal SK channels were τ₁=1.1±0.3ms and τ₂=7.7±2.9ms (N=11), while hSK1 open times were τ₁=1.0±0.4ms and τ₂=6.5±2.3ms (N=10). These comparisons suggest that hippocampal SK channels may be the native correlate for the SK1 gene product. Support for this proposal comes from RT-pcr experiments on RNA isolated from acutely dissociated rat CA1 neurons. Primers were used to sequences within the core of the molecule that were conserved between rSK1 and hSK1, but divergent in rSK2. Control experiments showed that these primers were indeed subtype-specific (data not shown). These experiments show that an abundance of SK1 mRNA is present in hippocampal CA1 neurons, with only a minor presence of SK2 transcript (J. P. Adelman, unpublished observation). However, it should be noted that our comparison is between the SK channel native to rat hippocampal neurons and the heterologously expressed human SK1 channel. The biophysical properties of rSK1 are not known.

The slow rise of the slow AHP in hippocampal neurons has been proposed to arise from the diffusion of Ca²⁺ from its point of entry to the SK channel (Lancaster and Adams, 1986). The main support of this proposal is that the slow AHP can be eliminated by intracellular EGTA, a relatively slow Ca²⁺ chelator (Lancaster and Nicoll, 1987). The observed sensitivity of hippocampal SK channels to Ca²⁺ (EC₅₀ 0.56μM, see Fig. 1) demands that a substantial rise in intracellular Ca²⁺ concentration must underlie generation of the slow AHP. This argument is supported by the finding that the P_o of the SK channel at the peak of the slow AHP is 04–0.6 (Sah and Issacson, 1995,Valiante et al), values that can only be observed with an intracellular Ca²⁺ concentration of 0.6–1μM (see Fig. 1). However, bulk increases of intracellular Ca²⁺ only up to 0.1μM have been measured (Knöpfel et al).

The slow AHP in hippocampal neurons is blocked by nimodipine, implying that L-type channels provide Ca²⁺ for SK channel activation (Rascol et al,Moyer et al,Tanabe et al). The subcellular location of SK channels in hippocampal neurons is not known. It has been proposed that they may be somatic (Lancaster et al) or located in the proximal dendritic tree (Sah and Bekkers, 1996). SK channels recorded in this study were somatic, as is the distribution of L-type Ca²⁺ channels in these neurons (Hell et al). This distribution, within a soma of ∼10μm in diameter, appears inconsistent with the slow rise of the slow AHP. It has been suggested that the characteristic time course of the slow AHP results from SK channels activating slowly in response to an increase in cytosolic Ca²⁺ (Sah and Clements, 1999). This was suggested because it was assumed that SK channels are located in the proximal apical dendrite (Sah and Bekkers, 1996) and the kinetics of the intracellular Ca²⁺ transient observed in this region are too rapid (Sah and Clements, 1999). However, this proposal assumes that SK channels are not somatic in their distribution, contradicting evidence from single channel recording (Lancaster et al; this study). In addition, to permit a model to be constructed reproducing the time course of the slow AHP, it was assumed that SK channels are activated by Ca²⁺ with an EC₅₀ of 150nM (Sah and Clements, 1999). This assumption is not supported by measurements of both cloned (Köhler et al,Hirschberg et al) and native (Lang and Ritchie, 1987,Lancaster et al,Grissmer et al,Park, 1994,Selyanko et al; this study) SK channels. Finally, this suggestion is not in agreement with the rapid activation of cloned SK channels by intracellular Ca²⁺ (Xia et al) and the finding in this study that the closed-time kinetics of hippocampal SK channels are consistent with them responding rapidly to a rise of intracellular Ca²⁺ (see Figure 4 and Figure 5 and Figure 6).

Recently, L-type Ca²⁺ and SK channels have been observed within the same patch. These experiments have indicated that L and SK channels are colocalized, being separated by only ∼100–150nm (Marrion and Tavalin, 1998). This finding is consistent with requiring ∼1μM Ca²⁺ to be present at the SK channel during the peak of the slow AHP (see above). However, it does not explain the kinetics of activation of the slow AHP. SK channels activate quite rapidly. For example, fast-flow application of 10μM Ca²⁺ to excised inside-out macropatches found that hSK1 and rSK2 channels activated with time constants of 5.8 and 6.3ms, respectively (Xia et al; see also Lancaster and Zucker, 1994). Modeling SK channel gating has predicted that rSK2 channels would activate with a time constant of ∼20ms in the presence of 1μM Ca²⁺ (Hirschberg et al). Therefore, the relative proximity would permit a high enough concentration of Ca²⁺ to be present at the SK channel, but the predicted rate of activation does not allow for a slow rise of the slow AHP. It has been proposed that delayed facilitation of L-type Ca²⁺ channels underlies the time course of the slow AHP (Cloues et al). Delayed facilitation is induced by a train of action potential waveforms and is characterized by prolonged L-type channel activity at membrane potentials negative to −60mV (Cloues et al). The time course of delayed facilitation is the same as the slow AHP, with both exhibiting a slow-rising phase and decay (Pedarzani and Storm, 1993,Cloues et al). Therefore, it is possible that delayed facilitation dictates the time course of SK channel activation, producing the characteristic slow rise and decay of the slow AHP.