Calcium-Activated Potassium Currents In Mammalian Neurons


  • Pankaj Sah,

    1. *Division of Neuroscience, John Curtin School of Medical Research, Australian National University, Canberra and Prince of Wales Medical Research Institute, University of NSW, Randwick, New South Wales, Australia
    Search for more papers by this author
  • Phil Davies

    1. *Division of Neuroscience, John Curtin School of Medical Research, Australian National University, Canberra and Prince of Wales Medical Research Institute, University of NSW, Randwick, New South Wales, Australia
    Search for more papers by this author

Correspondence: PankajSah Division of Neuroscience, John Curtin School of Medical Research, GPO Box 334, Canberra, ACT 2601, Australia. Email:


1. Influx of calcium via voltage-dependent calcium channels during the action potential leads to increases in cytosolic calcium that can initiate a number of physiological processes. One of these is the activation of potassium currents on the plasmalemma. These calcium-activated potassium currents contribute to action potential repolarization and are largely responsible for the phenomenon of spike frequency adaptation. This refers to the progressive slowing of the frequency of discharge of action potentials during sustained injection of depolarizing current. In some cell types, this adaptation is so marked that despite the presence of depolarizing current, only a single spike (or a few spikes) is initiated. Following cessation of current injection, slow deactivation of calcium-activated potassium currents is also responsible for the prolonged hyperpolarization that often follows.

2. A number of macroscopic calcium-activated potassium currents that can be separated on the basis of kinetic and pharmacological criteria have been described in mammalian neurons. At the single channel level, several types of calcium-activated potassium channels also have been characterized. While for some macroscopic currents the underlying single channels have been unambiguously defined, for other currents the identity of the underlying channels is not clear.

3. In the present review we describe the properties of the known types of calcium-activated potassium currents in mammalian neurons and indicate the relationship between macroscopic currents and particular single channels.


An increase in potassium permeability in response to changes in cytosolic calcium was first identified in red blood cells. 1 Subsequently, a potassium current activated by intracellular calcium was recorded in snail neurons by Meech and Strümwasser. 2 This current is due to the opening of potassium channels gated by calcium from the cytoplasmic face. It is now clear that virtually all cell types express at least some type of calcium-activated potassium channel. In neurons, two broad families of calcium-activated potassium channel are known to exist. The large conductance, BK, or Maxi K-type channels, and the small conductance or SK-type channels. These two types of channel can be distinguished on the basis of both biophysical and pharmacological criteria.

The BK channels typically have single channel conductances greater than 100 pS (in symmetrical K+ solutions). They are voltage sensitive, with their probability of opening larger at depolarized potentials, and are blocked by low concentrations (0.5–1 mmol/L) of tetraethylammonium ions (TEA) and by the specific toxins charybdotoxin and iberiotoxin. 3,4 These channels are gated by calcium with a Kd that is steeply voltage sensitive, being several micromolar near resting membrane potentials and in the nanomolar range at depolarized potentials (+20 to +40 mV). 5 They open rapidly when calcium is raised on the cytosolic face and, upon removal of calcium, they close rapidly. It should be noted that while charybdotoxin is also known to block some voltage-dependent potassium channels as well as the intermediate conductance channels (vide infra), iberiotoxin appears to be selective for large conductance calcium-activated potassium channels.

The SK or small conductance channels have single channel conductances in the 5–20 pS range. They are voltage independent and are blocked by high concentrations of TEA (10–20 mmol/L); however, they are unaffected by charybdotoxin and iberiotoxin. They are blocked by the specific toxin apamin. 4 Like BK channels, these channels also respond rapidly to changes in free calcium and typically have Kd values below 1 μmol/L. 6

A third family of calcium-activated potassium channels, the intermediate conductance or IK channels, is also known to exist in some cells. These channels are voltage insensitive and are also blocked by charybdotoxin. 4 While these channels have been proposed to exist in magnocellular neurons of the supraoptic nucleus, 7 their role, if any, in other neurons is unclear. We will not discuss these channels further in the present review.

Molecular studies have also identified two broad families of calcium-activated potassium channel. One family, initially identified from Drosophila, constitutes the BK-type channels. 8 These are large proteins that show homology to the voltage-dependent potassium channels. As with some other ion channels, a large gene family has been identified with many individual proteins due to alternate splicing. 9,10 For BK channels, variation in the calcium and voltage sensitivity of different splice variants has clear physiological implications: the regulated expression of different splice variants plays a key role in tuning the oscillation frequency in individual hair cells. 11

Recently, a second gene family has been found that codes for SK-type channels. 12 Three members have been identified so far: SK1, SK2 and SK3. These proteins, when expressed in oocytes, form calcium-activated potassium channels with conductances compatible with SK channels. The SK2 and SK3 channels are apamin sensitive, while SK1 is not. Thus, SK2 and SK3 are likely candidates for forming SK channels. Interestingly, these channels do not have a calcium-binding motif and gating of these channels is thought to occur by the binding of calcium to calmodulin, which, in turn, is covalently linked to the potassium channel. 13


The macroscopic current that corresponds to activation of BK channels in neurons has been named IC. 14,15 As expected from the properties of BK channels, this current activates rapidly following calcium influx and, upon removal of calcium, rapidly deactivates. In contrast, due to their higher affinity for calcium at hyperpolarized membrane potentials, the activity of SK channels following action potentials leads to more long-lasting currents than those due to BK channels. However, the identity of the macroscopic current(s) associated with SK channels has been a subject of some debate.

Slow calcium-activated potassium currents activated by action potentials in neurons can be separated into two distinct types. These are distinguishable on both kinetic ( Fig. 1) and pharmacological ( Fig. 2) grounds. 16 One of these, IAHP, activates rapidly (1–5 msec) following calcium influx and decays with time constants of 50– 100s msec. This current is voltage insensitive, it is unaffected by low concentrations of TEA and is potently blocked by apamin. The IAHP was originally recorded from bullfrog sympathetic neurons. 15 A similar current has been termed gKCa,1 in mammalian autonomic neurons 17,18 and m-IAHP in cortical neurons. 19 The properties of this current are similar to those of SK channels and, therefore, it is likely that activation of these channels underlies IAHP. 12,16

Figure 1.

Outward tail currents in a guinea-pig coeliac ganglion cell (a) and a rat CA1 hippocampal pyramidal neuron (b) recorded in voltage clamp. Following a brief depolarizing voltage step sufficient to initiate a single action current (a) or several action currents (b), an outward tail current consisting of two components is observed. One component (IAHP) is rapidly activated and decays with a fast time constant, while a second component (sIAHP) has a slower onset and a slower time constant of decay. For initial descriptions of these data see Cassell and McLachlan 17 and Jobling et al.33

Figure 2.

Pharmacological separation of IAHP and sIAHP. Current recordings in voltage clamp showing the effects of addition of noradrenaline (10 μmol/L) and apamin on slow outward tail currents in a rat hippocampal pyramidal neuron (a) and a guinea-pig coeliac ganglion cell (b). Records taken in control solution and in the presence of the indicated drugs have been superimposed. Addition of noradrenaline abolishes the slower component of the outward current, revealing an underlying faster component of outward current. In contrast, addition of 100 nmol/L apamin blocks the fast component of the outward tail current (IAHP) but has relatively little effect on the slower component of the outward current (sIAHP). For initial descriptions of these data see Sah and Clements 22 and Jobling et al.33

The other type of current, sIAHP, activates much more slowly (100s msec) following calcium influx and decays with a time constant in the order of seconds. This current is also voltage insensitive and is unaffected by low concentrations of TEA. Unlike IAHP, it is not blocked by apamin, but is modulated by a range of transmitter systems, all of which reduce its amplitude. 16 This reduction is most likely by a direct action on the potassium channels. 20–22 This current was first recorded in myenteric neurons 23,24 and hippocampal pyramidal 25 cells, where it was also called IAHP. However, it is clear that this slower current is kinetically and pharmacologically distinct from the apamin-sensitive IAHP. Therefore, this slower current is now referred to as either sIAHP or slow-IAHP. It has been called gKCa,2 in sympathetic neurons 17 and sAHP in cortical neurons. 19 The channels underlying sIAHP have never been studied at the single channel level; however, noise analysis indicates that their unitary conductance is in a range compatible with SK channels. 21,26


The distinguishing feature of sIAHP is its characteristic slow time course. Following calcium influx, it rises to a peak with a time constant of approximately 100–300 msec and then decays with a time constant on the order of 1 s. Several explanations have been suggested for this slow time course.

1. The slow rise and decay may be due to the buffered diffusion of calcium from its point of influx to where the potassium channels are located. 27,28 This idea seems unlikely because, initially, following membrane depolarization, calcium rises quickly in all parts of the cell. 22,29,30 Furthermore, the rise and decay of the current has a high temperature sensitivity, 18,25 suggesting that it cannot be due to simple diffusion of calcium. Finally, exogenous addition of mobile calcium buffers, such as EGTA, which should accelerate the delivery of calcium to a distant site, 31 slow both the rise and decay of this current. 22,28,32

2. The slow time course of sIAHP may be due to calcium-induced calcium release (CICR). 18,33,34 This mechanism appears to be true for sensory vagal neurons; 35 however, there is little evidence for CICR in CA1 pyramidal neurons, which also express a similar macroscopic current. 28

3. The channels underlying sIAHP may require a second messenger, besides calcium, for their activation. 30,35,36 While there is no direct evidence to rule out this possibility, this explanation also seems unlikely because there is no delay to the foot of the potassium current that is activated by photorelease of caged calcium. 22 Second messenger-mediated responses typically exhibit latencies of > 20 msec to the foot of the response. 37

4. The slow activation of sIAHP may be due to delayed facilitation of the calcium channels that supply the calcium to activate sIAHP. 38,39 This hypothesis requires that the K channels underlying sIAHP respond rapidly to changes in [Ca2+]i. While this has been suggested to be the case in one study, 27 more recent results indicate that the channels underlying sIAHP do not respond quickly to calcium, but open with a distinct delay. 22 Furthermore, rapidly buffering calcium with a caged calcium chelator does not lead to a fast turnoff of sIAHP. This result indicates that the slow kinetics of sIAHP represent the intrinsic kinetic properties of the underlying potassium channels. 36,40,41

With the cloning of a family of small conductance calcium- activated potassium channels, 12 it has been suggested that the apamin-insensitive channel SK1 may underlie sIAHP. However, the kinetic properties of SK1 channels are similar to those of the apamin-sensitive members of this family, 6 making it unlikely that it underlies this current. 22 If the SK1 protein is a component of channels that underlie sIAHP, then there are two possibilities. First, sIAHP channels may simply be a heteromer SK1, SK2 and SK3, the properties of which have not as yet been explored. Second, it has been shown that calmodulin is an integral component of SK channels and gating of SK channels is mediated by binding of calcium to calmodulin. 13 Therefore, it is possible that a calcium- binding protein may be interacting with SK1 channels, thus conferring on them slow kinetics. Third, it is possible that the channels underlying sIAHP are formed from as yet undescribed subunits.


Following calcium influx via voltage-gated channels, due to the presence of intracellular calcium buffers, the concentration of cytosolic free calcium decays with several kinetic components. 42,43 A local ‘domain’ of high concentration is restricted to the region in the immediate vicinity of the calcium channels. As a consequence of this, for potassium channels with a low affinity for calcium, only those calcium-activated potassium channels lying in relatively close proximity to the calcium channels will be activated. 44,45 Several studies have demonstrated that calcium-activated potassium channels activated by calcium influx are indeed in close proximity to the sites of calcium influx. 44,46 More recently, it has been revealed that even within the same patch of somatic membrane, calcium influx through specific types of voltage-gated channels selectively activates different types of calcium-activated potassium channels. 39,47

Voltage-gated calcium gated channels can be separated on the basis of both biophysical and pharmacological sensitivity into at least six major groups (N-, L-, P-, Q-, T- and ‘resistant’ or R-type). Different types of neurons express some or all of these different channel types in varying proportions. 48 Because specific calcium channels may be associated with specific potassium channels and different calcium channels may be spatially segregated over the surface of a neuron, the possibility exists that different types of calcium channel may activate different calcium-activated potassium currents.

Many studies have used two properties to investigate the coupling between calcium and potassium channels: (i) as described below, different calcium-activated potassium channels underlie discrete temporal events (action potential repolarization and afterhyperpolarization (AHP)); and (ii) many different calcium channels are selectively blocked by a number of different pharmacological agents. Thus, using different antagonists, the roles of different types of calcium channels on the activation of calcium-activated potassium channels underlying action potential repolarization and the AHP have been investigated.

In a diverse range of neurons, addition of blockers of N-type calcium channels results in a reduction in the activation of the channels underlying the apamin-sensitive AHP. 49–57 Blockade of N-type channels can also block the activation of BK channels 39 and action potential repolarization 56 in some cell types, suggesting that a single class of calcium channel may play multiple roles in regulating neuronal excitability.

The effects of blocking L-type calcium channels on the activation of calcium-activated potassium channels is more diverse than that seen following blockade of N-type channels. The effects vary between neurons, ranging from no effects at all to blocking activation of BK- and/or SK-type channels. In sympathetic neurons, blockade of L-type channels slows action potential repolarization, 52 consistent with the notion that it reduces activation of BK-type channels, although this effect only occurs within a single electrophysiologically defined class of sympathetic neuron, the phasic class. 52 In other neuron types, L-type channel blockade has little or no effect on action potential repolarization, but has more pronounced effects on SK-type or apamin-resistant channels underlying the AHP. 34,39,52

Blockade of P-type channels has also been shown to block SK-type or apamin-resistant channels underlying sIAHP. 55,58 In single channel recordings from hippocampal somata, it was noted that openings of P/Q channels do not activate either BK- or SK-type channels, despite their presence being confirmed within the same patch. 39 It is also worth noting that, in many studies, there is a component remaining that is resistant to blockade using known specific antagonists. This resistant component suggests that R-type channels also play a role in the activation of these conductances.

The different linkages observed between calcium and potassium channels suggest there is a diverse array of possible combinations used by different neurons and that these linkages can vary between neuron type, species or even electrophysiologically defined phenotypes ( Table 1). However, the existence of these specific channel associations suggests that the spatial distribution of calcium and potassium channels must be organized to ensure that potassium channels are located within close proximity to particular calcium channels but also distant from other types of calcium channels.

Table 1.  Associations between calcium influx through different classes of voltage-gated channels and calcium-activated potassium channels
Calcium channel typeCa2+-activated K+ channelNeuron typeReference
  1. SK, slow conductance K+ channel; BK, large conductance K+ channel; SCG, superior cervical ganglion.

  Lumbar sympathetic chain56
  Nucleus basalis55
  Lamprey spinal57
  Hypoglossal50, 68
  CA1 pyramidal67
 sIAHP channelsNeocortical pyramidal58
  Nucleus basalis55
 BKLumbar sympathetic chain56
  Sympathetic nerve terminals69
  Chick sympathetic47
  Frog neuromuscular junction46
  Sympathetic (phasic)70
  Chick sympathetic47
  CA1 pyramidal39
  CA3 pyramidal34
P/QSKNeocortical pyramidal58
  Lamprey spinal57
 sIAHP channelsNeocortical pyramidal58
  Nucleus basalis55
  Hypoglossal50, 68
TSKNucleus basalis55
 sIAHP channelsCA3 pyramidal34


Because of their rapid activation and relatively large amplitude, activation of IC contributes to repolarization of action potentials in a number of cell types and is largely responsible for the fast hyperpolarization that follows action potentials in these neurons ( Fig. 3). 16 Due to their steep voltage dependence, BK channels activated during the upstroke of the action potential close rapidly following return of the membrane potential to negative values. 59 Thus, BK channels have little role in the prolonged AHP that follows action potentials. Because it is partially blocked by charybdotoxin and iberiotoxin, it has been suggested that BK channels may contribute to the slow AHP in myenteric neurons. 60 However, the slow AHP in these cells is voltage insensitive, 24 making it unlikely that BK channels are active during the AHP. 61

Figure 3.

Voltage and current records showing the effects of blockade of large conductance (BK)-type Ca2+-activated K+ channels in a rat superior cervical ganglion neuron. Records taken in control solution (thin trace) and in the presence of 20 nmol/L iberiotoxin (thicker trace) have been superimposed. (a) Iberiotoxin slows action potential repolarization but has little effect on the amplitude and time course of the fast component of the afterhyperpolarization (b) or underlying outward current (c) recorded in voltage clamp. For initial descriptions of these results see Davies et al.52

Activation of the IAHP and sIAHP is responsible for spike frequency adaptation, but these currents are not active in the action potential ( Fig. 4). However, cells that only express IAHP continue to fire tonically and activation of IAHP simply slows the maximal firing frequency. 62 The presence of sIAHP leads to a progressive slowing of the discharge frequency and eventual cessation of action potentials. 16,63,64 The different effects of the two currents can be understood by the different kinetic properties of the two currents. The IAHP activates rapidly following a single action potential and its slow decay leads to a prolonged slow AHP that immediately follows the action potential. As IAHP inactivates, the cell slowly reaches threshold and can fire a second action potential. Thus, activation of IAHP leads to a reduction in action potential frequency. In contrast, following calcium influx, sIAHP activates slowly, which allows the cell to spike several times, leading to summation of sIAHP and termination of spiking.

Figure 4.

The IAHP currents contribute to the afterhyperpolarization (ahp), but not to action potential repolarization. Voltage and current records showing the effects of blockade of small conductance (SK)-type Ca2+-activated K+ channels in a rat superior cervical ganglion neuron. Records taken in control solution (thin trace) and in the presence of 100 nmol/L apamin (thicker trace) have been superimposed. (a) Apamin had little effect on action potential repolarization, but it reduced both the amplitude and time course of the AHP (b) and the underlying outward tail current (IAHP; c) recorded in voltage clamp. For initial descriptions of these results see Davies et al.52

The sIAHP current is modulated by a range of transmitter systems, all of which reduce its amplitude. 65 This modulation dramatically changes the repetitive discharge properties of neurons that express sIAHP. Thus, the presence of sIAHP allows for a greater level of control over the firing properties of neurons. In hippocampal pyramidal cells, the sIAHP current has been suggested to be localized to the apical dendritic tree, where its activation selectively shunts synaptic inputs arising in the apical dendrites. 16


In conclusion, calcium-activated potassium currents form a diverse array of potassium currents that play distinct roles in shaping the electrical properties of neurons. The biophysical and molecular details of the channels underlying these currents is now beginning to emerge. These data are showing us the rich diversity that is available in this class of potassium channels.


PS is supported by the Sylvia and Charles Viertel Foundation. PD is supported by a grant from the National Health and Medical Research Council of Australia (970852). The authors thank Professor Elspeth McLachlan for her support and David Ireland and Juan Martínz-Pinna (Prince of Wales Medical Research Institute) for their discussion and helpful comments.