Ion channel activation involves three major molecular processes: sensor activation, sensor-pore coupling, and pore opening. Sensors in channel proteins change conformation upon the stimulation of signals, such as changes in the membrane potential and ligand binding. The conformational change of sensors is propagated to the pore via interactions between sensors and the pore, known as sensor-pore coupling. Finally, the pore opens to allow ionic flow across the membrane. BK channel activation depends on both membrane potential and intracellular Ca²⁺. The studies of BK channel activation have revealed the following frameworks for understanding how the two stimuli open the same pore.

First, Ca²⁺ and voltage activate the channel with distinct mechanisms. In early studies the relation between Ca²⁺ and voltage in activating BK channels was an important question. The identification of Slo1 gene KCNMA1 and the availability of cDNA of Slo1 for functional expression of BK channels in exogenous cells (Atkinson et al., 1991; Adelman et al., 1992; Butler et al., 1993) allowed the studies to distinguish the distinct Ca²⁺ and voltage dependent activation mechanisms (Cui et al., 1997). When the intracellular Ca²⁺ concentration was kept low (≤0.5 nM) BK channels opened in response to membrane depolarization with a rate that exceeded the diffusion limit for Ca²⁺ to bind, suggesting that the channel can open by voltage without Ca²⁺ binding (Cui et al., 1997; Figures 1A,B). On the other hand, when the voltage sensor was kept at the resting state by negative membrane potentials (≤−140 mV) (Horrigan et al., 1999; Cui and Aldrich, 2000) the open probability of BK channels increased by 4 orders of magnitude when Ca²⁺ concentration was elevated to 100 μM (Horrigan and Aldrich, 2002; Yang et al., 2010; Figures 1C,D), suggesting that the channel can open by Ca²⁺ binding without voltage. Subsequent studies identified the voltage sensor and Ca²⁺ binding sites in different structural domains of the channel. While the residues responsible for voltage sensing are located in the transmembrane segments (Diaz et al., 1998; Cui and Aldrich, 2000; Ma and Horrigan, 2005; Pantazis et al., 2010a; Zhang et al., 2014), the Ca²⁺ binding sites are found to reside in the cytosolic domain (Schreiber and Salkoff, 1997; Shi et al., 2002; Xia et al., 2002; Bao et al., 2004; Yusifov et al., 2008; Wu et al., 2010; Yuan et al., 2010; Zhang et al., 2010; Javaherian et al., 2011; Yuan et al., 2012; Hite et al., 2017; Tao et al., 2017; Tao and MacKinnon, 2019). Thus, voltage and Ca²⁺ activate the channel by perturbing different structure domains that harbor the respective sensors and can open the channel independent of each other.

Second, Ca²⁺ and voltage activate the channel with allosteric mechanisms. In the extreme experimental conditions with low Ca²⁺ concentrations (≤0.5 nM) and negative membrane potentials (≤−140 mV) BK channels were observed to open with a small but measurable open probability (∼10^–7) (Horrigan et al., 1999; Cui and Aldrich, 2000; Figure 1C). In these experimental conditions the channel opening was not controlled by Ca²⁺ binding or voltage sensor activation, but was an intrinsic spontaneous event. The open probability of BK channels increases with Ca²⁺ binding or voltage sensor activation, since at the open conformation the channel has a higher Ca²⁺ affinity and a facilitated voltage sensor activation (McManus and Magleby, 1991; Cox et al., 1997; Horrigan et al., 1999; Horrigan and Aldrich, 2002). Each BK channel contains eight high-affinity Ca²⁺ binding sites (Schreiber and Salkoff, 1997; Shi et al., 2002; Xia et al., 2002; Bao et al., 2004; Wu et al., 2010; Yuan et al., 2010, 2012; Zhang et al., 2010; Hite et al., 2017; Tao et al., 2017; Tao and MacKinnon, 2019) and four voltage sensors (Hite et al., 2017; Tao et al., 2017; Tao and MacKinnon, 2019; Figure 2), and the open probability of the closed-open transitions increases with each Ca²⁺ binding and voltage sensor activation. This mechanism of Ca²⁺ and voltage dependent activation can be quantitatively described by a model that accounts for BK channel activation with the changes of Ca²⁺ concentration and voltage at a broad range that includes and extends beyond physiological conditions (Horrigan and Aldrich, 2002; Figure 2A). This mechanism is in contrast to Shaker K⁺ channels, for which the opening was tightly controlled by voltage sensor activation even when the open probability was at 10^–7 (Islas and Sigworth, 1999). Thus, the mechanism of coupling between sensor activation and pore opening in Shaker K⁺ channels is thought to be obligatory, while in BK channels is allosteric. The allosteric mechanism makes it possible for BK channels to sense both Ca²⁺ binding and voltage sensor activation for opening. In a channel with the pore tightly controlled by the voltage sensor, such as Shaker, Ca²⁺ binding would not have been able to open the channel without voltage sensor activation.

Third, interactions among structural domains are important for Ca²⁺ and voltage dependent activation of BK channels (Figure 2B). The primary sequence of Slo1 channels, in comparison with voltage gated K⁺ (Kv) channels, indicated that the channel included three structural domains, the transmembrane pore-gate domain (PGD), the transmembrane voltage sensor domain (VSD), and the cytosolic domain (CTD) (Cui, 2010; Lee and Cui, 2010). The structure of MthK (Jiang et al., 2002a,b) and subsequently the structures of the BK CTD (Wu et al., 2010; Yuan et al., 2010; Yuan et al., 2012) showed that the CTD contains two RCK (Regulator of K⁺ Conductance) domains in each Slo1 subunit, and the eight RCK domains of the four subunits form a ring-like structure called the gating ring. Based on the sequence homology with Kv channels, the structures of the VSD, which contains the transmembrane segments S1–S4, and PGD, which contains S5–S6, were modeled after the structure of Kv1.2 (Long et al., 2005a; Lee and Cui, 2009). Two Ca²⁺ binding sites, the Ca²⁺ bowl that is primarily formed by the RCK2 residues (Schreiber and Salkoff, 1997; Bao et al., 2004; Yusifov et al., 2008; Wu et al., 2010; Hite et al., 2017; Tao et al., 2017; Tao and MacKinnon, 2019) and the site that is located in RCK1 (Shi et al., 2002; Xia et al., 2002; Wu et al., 2010; Zhang et al., 2010; Yuan et al., 2012; Hite et al., 2017; Tao et al., 2017; Tao and MacKinnon, 2019), were found in the CTD of each Slo1. Residues in the VSD that are important for voltage sensing were identified (Ma and Horrigan, 2005; Zhang et al., 2014). Thus, the VSD and CTD harbor the voltage and Ca²⁺ sensors and propagate the stimuli to open the pore in PGD via interactions among these structure domains.

Prior to the publication of the cryo-EM structures of the whole BK channel (Hite et al., 2017; Tao et al., 2017; Tao and MacKinnon, 2019; Figures 2B,C) the coupling of Ca²⁺ binding in the CTD to the opening of the PGD had been proposed to involve two types of interactions among the structure domains. A study Niu et al. (2004) showed that voltage and Ca²⁺ dependent activation of BK channels was changed by altering the covalent link between the PGD and the CTD via a peptide of 15 residues, known as the C-linker. The changes in activation depended on the changes in C-linker length by addition or deletion of amino acids in a relationship as if the interaction between the PGD and CTD were like pulling a spring for channel activation. Thus, it was proposed that Ca²⁺ may activate the channel by causing a conformational change in the CTD that directly tugs the pore to open via the C-linker. On the other hand, the CTD was proposed to also interact with the VSD via non-covalent interactions, and such interactions may indirectly open the pore via VSD-PGD interactions (Lee and Cui, 2009). This proposal derived from the finding that the residues from both the VSD and CTD form a Mg²⁺ binding site that modulates voltage dependent activation (Shi and Cui, 2001; Hu et al., 2003; Zeng et al., 2005; Yang et al., 2007; Yang et al., 2008), and thus VSD and CTD are located closely. This mechanism was consistent with an earlier finding that the part of the CTD that is located close to the VSD is important in determining the different Ca²⁺ sensitivities between different BK channel homologs (Krishnamoorthy et al., 2005), and subsequent studies showed that alterations of the VSD-CTD interactions indeed affected Ca²⁺ dependent activation (Yang et al., 2013; Geng et al., 2020).

For voltage dependent activation, it was found that a negatively charged residue E219 in the VSD, which contributed to voltage sensing, interacted with charged residues E321 and E324 in the PGD to mediate the coupling of voltage sensor activation to channel opening (Zhang et al., 2014). The interaction, however, was also found to be involved in the coupling between Ca²⁺ bindings and pore opening. Similar to Ca²⁺ dependent activation that involves interactions among all CTD, VSD, and PGD, voltage dependent activation also involves all three structure domains. A comparison between the voltage dependent activation of the wild type BK channel with that of a truncated BK channel with CTD deletion (Budelli et al., 2013) showed that the CTD was important in the coupling between VSD activation and pore opening (Zhang et al., 2017). The CTD undergoes conformational changes during voltage dependent activation as detected by fluorescence signals (Miranda et al., 2018). Ca²⁺ binding to the CTD also alters voltage dependence of channel activation (Horrigan and Aldrich, 2002; Sweet and Cox, 2008; Savalli et al., 2012; Lorenzo-Ceballos et al., 2019). These results suggest that the interactions among all three structure domains are involved in voltage and Ca²⁺ dependent activation.

The CTD-VSD non-covalent interactions were more clearly shown by the cryo-EM structures of the whole BK channels (Hite et al., 2017; Tao et al., 2017; Tao and MacKinnon, 2019; Figures 2B,C). The CTD gating ring is located closely to the membrane spanning domains of the channel, with an extensive interface between the N-terminus of the CTD with the cytosolic part of the VSD (585 Ų per subunit in the Ca²⁺ structure) (Hite et al., 2017). Comparing the structure with Ca²⁺/Mg²⁺ bound to the metal-free structure the N-lobes of the RCK1 domain, which face the VSD, tilted in a rigid body fashion away from the pore axis. In the Ca²⁺/Mg²⁺ bound structure the cytosolic part of the VSD also showed a corresponding outward displacement, and the movements of both the RCK1 and VSD alter the interface between these domains. This Ca²⁺ dependent change in the interactions between CTD and VSD suggests that these interactions are part of the coupling mechanism for Ca²⁺ binding to open the channel, and is consistent with the studies to show that mutations in the VSD-CTD interface alter Ca²⁺ sensitivity (Yang et al., 2013; Geng et al., 2020). The C-linker in the cryo-EM structures is partly α-helical and partly extended and nearly identical in both the structures with and without Ca²⁺/Mg²⁺ bound, but undergoes a large positional displacement laterally as a rigid unit. Recent studies showed that the C-linker can interact with the membrane directly to alter BK channel activation (Tian et al., 2019; Yazdani et al., 2020b). These results suggest that the C-linker may not affect channel activation simply as a passive link between the CTD and PGD, but directly interact with the membrane and other parts of the channel protein such as the gating ring to mediate Ca²⁺ and voltage dependent activation.

In the tetrameric structure of the Kv1.2 channel the VSD of each subunit is adjacent to the PGD of its neighboring subunit, showing a domain-swapped configuration (Long et al., 2005b). The BK channel, however, does not show such a domain-swapping. Instead, the VSD and PGD of the Slo1 subunit interact within each subunit (Hite et al., 2017; Tao et al., 2017; Tao and MacKinnon, 2019; Figure 2C). This difference as revealed by the cryo-EM structure of BK channels suggests that the coupling between VSD activation to pore opening differs from the canonical electromechanical coupling mechanism for the domain-swapped Kv channels. In domain swapped Kv channels the peptide linking the VSD and the PGD of a subunit, the S4–S5 linker, interacts with the cytosolic part of the transmembrane helix S6 (Long et al., 2005b). S6 lines the inner pore, and the helices from four subunits cross at the cytosolic side of the membrane to form the activation gate that controls ionic flow during the opening and closing of these Kv channels (Liu et al., 1997; Long et al., 2005a,b). The interactions between the S4–S5 linker and S6 has been shown to mediate electromechanical coupling, which open the activation gate in response to VSD movements (Lu et al., 2001; Long et al., 2005b; Hou et al., 2020; Cowgill and Chanda, 2021). However, in BK channels, the S4–S5 linker of each subunit is short and not located close to S6, but interacts with the N-terminus of RCK1 (Hite et al., 2017; Tao et al., 2017; Tao and MacKinnon, 2019; Figure 2C). The VSD seems to be in close contact with the PGD within the same Slo1 subunit only with the interface between S4 and S5 helices (Figures 2B,C). The contact between S4 and S5 helices is extensive, which suggests that the interactions between the two helices are important for mediating the coupling between VSD activation and pore opening.

The extensive interactions between the S4 and S5 would also restrict the motion of S4, which may be responsible for the unique VSD movements during BK channel activation. In domain swapped Kv channels the charged residues that sense voltage for channel activation, known as gating charges, are primarily found in the S4 transmembrane helix (Aggarwal and MacKinnon, 1996; Mannuzzu et al., 1996; Seoh et al., 1996). S4 moves across the membrane with a large distance upon membrane potential depolarization, resulting in a transfer of total 12–16 equivalent gating charges across the electric filed in the membrane (Zagotta et al., 1994a,b; Aggarwal and MacKinnon, 1996; Mannuzzu et al., 1996; Seoh et al., 1996; Bezanilla, 2000; Gandhi and Isacoff, 2002). However, in BK channels gating charges are found in S2 (D153 and R167), S3 (D186), and S4 (R213, E219) segments (Ma and Horrigan, 2005; Pantazis et al., 2010a; Zhang et al., 2014). During voltage dependent activation of BK channels a total 2.32 equivalent gating charges are moved across the electric filed in the membrane (Horrigan and Aldrich, 1999; Ma and Horrigan, 2005), much smaller as compared to those in domain swapped Kv channels. These results suggest that the VSD may undergo small movements that involve many transmembrane segments during voltage dependent activation (Ma and Horrigan, 2005; Pantazis et al., 2010a,b; Pantazis and Olcese, 2012; Pantazis et al., 2018). Furthermore, the pore opening in BK channels also retrospectively control VSD activation, such that VSD activation increases with multiple time courses, and some of which are in parallel with channel opening (Horrigan and Aldrich, 1999; Savalli et al., 2006), consistent with the idea that the S4–S5 interactions may restrict VSD movements during BK channel activation.

Cryo-EM structures of BK channels reveal that the S6 helices from the four Slo1 subunits do not cross at the cytosolic side, either with or without Ca²⁺/Mg²⁺ bound (Figure 2B). If the metal-free structure represents the closed state of the channel, while the Ca²⁺/Mg²⁺ bound structure represents an open state (Hite et al., 2017), the cytosolic side of the pore is thus wide open at both the open and closed states of the channel. This result is consistent with the suggestion from the studies of BK channel block by quaternary ammonium (QA) molecules (Li and Aldrich, 2004; Wilkens and Aldrich, 2006; Tang et al., 2009) and other smaller molecules (Zhou et al., 2011; Zhou et al., 2015) that the channels do not have an intracellular gate so that the blockers can enter the pore even at the closed state. The study of a peptide blocker suggests that the cytosolic part of the pore may undergo a conformational change to reduce the pore size during BK channel closing (Li and Aldrich, 2006), but the change is not sufficient to close the pore to restrict the entrance of K⁺ ions or QA blockers. These results indicate that the activation gate of BK channels may be located above the cytosolic side, possibly at the selectivity filter, which is the narrowest part of the pore (Figures 2B,C) and with a conserved function among K⁺ channels to select K⁺ ion over other ions for permeation (Hite et al., 2017; Tao et al., 2017; Tao and MacKinnon, 2019).

Recently, an alternative mechanism for the activation gate was proposed. The structures of the metal free and Ca²⁺/Mg²⁺ bound BK channels suggested that the inner pore of the channel underwent a conformational change in the absence of metal binding, which mainly involved an amphipathic segment of S6, V₃₁₉PEIIE₃₂₄, to expose hydrophobic residues V319 and I323 to the pore inner surface (Hite et al., 2017; Tao et al., 2017; Tao and MacKinnon, 2019). As a result of these changes the inner pore beneath the selectivity filter becomes narrower, more elongated and hydrophobic. A molecular dynamics simulation study found that these changes promote dewetting transitions that completely deplete the inner pore of liquid water, giving rise to a vapor barrier to block the ion flow (Jia et al., 2018). Importantly, the dry pore remains physically open with an average diameter of ∼6 Å, allowing QA blockers to access the deep-pore region to block the channel even in the closed state. Such a hydrophobic gate depends on the hydrophobicity of the surface residues and the geometry of the inner pore (Aryal et al., 2015; Yazdani et al., 2020a). A functional study showed that the mutations in the BK channel inner pore A316D and A316V made the channel constitutively open and harder to open, respectively (Chen et al., 2014). Consistent with these results molecular dynamics simulations found that A316D reduced hydrophobicity of the pore and prevented dewetting transitions, while A316V enhanced hydrophobicity of the pore and made dewetting transitions faster with less water molecules remaining in the pore (Jia et al., 2018).

At present the hydrophobic gating mechanism is primarily based on molecular dynamic simulations, which can explain the experimental results of QA blockers on BK channel and some mutations. There has not been a direct experimental validation of this mechanism, partly due to the dilemma that once the vapor barrier forms the channel is closed and devoid of any functional detection. Comparing to the hydrophobic gate, the alternative mechanism, the selectivity filter acting as the activation gate, is not clearly defined with computational or experimental evidence. Where is the activation gate of BK channels and how it operates are not only part of the fundamental mechanism of voltage and Ca²⁺ dependent activation but also important for understanding BK channel modulation by compounds that act in the pore of BK channels, such as the QA blockers, paxilline (Zhou et al., 2020), and NS11021 (Rockman et al., 2020; Schewe et al., 2019). Some of these modulators have been excellent tools for the studies of the roles of BK channels in cellular and tissue function (Kaczorowski and Garcia, 2016; Cui, 2020), and may serve as the lead for therapies for BK variants linked neurological diseases as described in the following sections.

In a study of a family with inherited generalized epilepsy and coexistent paroxysmal dyskinesia (GEPD), a heterozygous A-G transition in exon 10 of KCNMA1 was identified, which results in the D434G mutation in the BK channel subunit Slo1 (Du et al., 2005; Figure 3A). The inheritance followed an autosomal dominant pattern, and sixteen members of the family were affected. Among these individuals nine had early childhood onset of absence epilepsy, showing the characteristic episodes of loss of awareness, with vacant staring and unresponsiveness, and electroencephalography (EEG) showing synchronous spike-and-wave discharges (SWDs). Three of these individuals also developed generalized tonic-clonic seizures. The seizures in some of these individuals were responsive to anti-epilepsy medicine valproate, lamotrigine, or clonazepam. Among the sixteen affected individuals twelve had paroxysmal dyskinesia including five with both seizures and paroxysmal dyskinesia. These individuals were described to have involuntary dystonic or choreiform movements of the mouth, tongue, and extremities. The episodes were not induced by sudden movements, but induced by alcohol, fatigue, and stress, so that the paroxysmal dyskinesia was classified as paroxysmal non-kinesigenic dyskinesia (PNKD). It was noted that the episodes of PNKD developed around the same age as the onset of the seizures, but during the episodes the individual had preserved consciousness.

Comparison of the currents of the D434G BK channels with the wild type (WT) BK currents showed that the D434G currents were increased with a faster activation time course at the same voltage and intracellular Ca²⁺ concentrations ≥1 μM (Du et al., 2005; Díez-Sampedro et al., 2006; Lee and Cui, 2009; Wang et al., 2009; Yang et al., 2010; Moldenhauer et al., 2020), which is the cytosolic Ca²⁺ concentration at neuronal excitations (Figures 3B,C). These results suggest that more currents flow through the D434G BK channels during an action potential. At these Ca²⁺ concentrations, the voltage dependence of channel conductance (GV relation) of D434G BK channels was shifted to more negative voltages (Figure 3C), indicating that channel activation was increased. The increased channel activation was actually primarily because of an increased Ca²⁺ sensitivity without changes in voltage dependent activation, because the mutation increased the apparent affinity for Ca²⁺ in channel activation (Díez-Sampedro et al., 2006; Wang et al., 2009; Yang et al., 2010), while at low Ca²⁺ concentrations (∼0.5 nM) with few Ca²⁺ binding to the channel the D434G mutation did not alter GV relation (Díez-Sampedro et al., 2006; Yang et al., 2010). The residue D434 is located close to the Ca²⁺ binding site in RCK1 (Figure 3A). The mutation D434G neutralizes a negative charge close to the Ca²⁺ binding site and makes the protein structure more flexible around the site, which may be responsible for the mutation to specifically enhance the affinity of Ca²⁺ for the RCK1 site and strengthen the coupling between Ca²⁺ binding to the site with pore opening (Yang et al., 2010).

Recently a knock-in mouse model carrying BK D434G mutation was characterized (Dong et al., 2021). The BK D434G mutation mice manifested symptoms resembling the human patients affected by the BK D434G mutation. Simultaneous video-EEG recordings showed that the BK-D434G mice had frequent episodes of spontaneous, generalized SWDs, and during which behavioral arrests, characteristic of absence seizures. These mice were also more susceptible to the induction of generalized seizures by injection of low dose of pentylenetetrazole (PTZ), a convulsant, than the WT mice, consistent with that some BK-D434G patients developed generalized tonic-clonic seizures (Du et al., 2005). Consistent with the human patients, the mutant animals exhibited severe locomotive defects in a battery of motor tests when there was no absence seizure. However, due to the limitations of mouse models, it is difficult to clearly tell if the animals also had PNKD. Both the heterozygous BK-D434G mutation (BK^DG/WT) and the homozygous mutation (BK^DG/DG) mice exhibited these symptoms, but the incidents were more frequent or severe in BK^DG/DG mice. All these symptoms could be suppressed by the treatment with a BK channel inhibitor, paxilline (PAX). These results support the idea that the D434G mutation in BK channels causes a gain of function of the channel that results in epilepsy and movement disorders.

Acute brain slice recordings from the BK D434G mutation mice further showed that the mutation caused hyperexcitability in both the cortical pyramidal neurons and the cerebellar Purkinje cells (Dong et al., 2021), which play essential roles in the pathogenesis of absence seizures (Crunelli et al., 2020) and contribute to the motor defects (Sausbier et al., 2004), respectively. These neurons from the BK^DG/WT mice exhibited a significantly increased action potential frequency compared with the BK^WT/WT mice. The action potentials in BK^DG/WT neurons showed a much faster repolarization with a shortened duration and augmented after-hyperpolarization amplitude (AHP). These results are consistent with that the BK-D434G channels have a higher Ca²⁺ sensitivity (Díez-Sampedro et al., 2006; Wang et al., 2009; Yang et al., 2010), allowing the channels to activate more during an action potential following membrane depolarization and the opening of voltage gated Ca²⁺ channels. The enhanced BK currents more efficiently hyperpolarize the membrane, enabling a faster recovery of the voltage-gated sodium channels from inactivation and potentially facilitating the activation of the hyperpolarization-activated cation (HCN) channels (Lancaster and Nicoll, 1987; Storm, 1987; Shah, 2014), which allow the neurons fire at a higher frequency. The application of BK channel inhibitor PAX widened the action potential, reduced AHP and decreased firing frequency.

The seizures in patients affected by BK D434G were responsive to general anti-epilepsy medicine (Du et al., 2005), and so were in the BK D434G mutation mice (Dong et al., 2021). Both ethosuximide and valproate suppressed generalized SWDs (Dong et al., 2021). It was shown that the anti-epilepsy medicine acetazolamide had no direct effect on BK channels (Moldenhauer et al., 2020), suggesting that BK channels were not the target of these general anti-epilepsy medicines. It is reported that 30% of absence epilepsy patients are pharmaco-resistant with the first-line anti-absence medicines (Crunelli et al., 2020). The results of BK channel inhibitor PAX (Dong et al., 2021) suggest that BK channels as a new target for anti-absence therapy may be developed to treat those pharmaco-resistant patients. However, the effects of PAX in suppressing generalized SWDs in BK D434G mutation mice only lasted for ∼30 min, which indicates that novel BK channel inhibitors with appropriate efficacy, specificity and pharmacokinetics need to be developed. The development of such novel drugs will benefit from the understanding of molecular mechanisms of BK channel gating.

The second disease associated KCNMA1 variant was identified in 2015, which resulted in the N1053S mutation in Slo1 (Zhang et al., 2015; Figure 3A). Subsequently more patients with the same Slo1 mutation were identified [also called N995S (Li et al., 2018), or N999S (Heim et al., 2020)]. These mutations were all de novo in unrelated patients. Some of these patients suffered movement disorders, diagnosed with PNKD (Zhang et al., 2015) or cataplexy (Heim et al., 2020), while some only suffered absence epilepsy or mixed with myoclonic seizures (Li et al., 2018). Only one patient suffered both symptoms (Heim et al., 2020). These patients also showed various degrees of developmental delay or intellectual disability. Whether symptoms respond to anti-epileptic medicines also varied among these patients.

As the symptoms of BK-N1053S patients showed some resemblance to those of BK-D434G patients, N1053S mutation is also a gain of function (GOF) mutation. Similar to D434G, the N1053S mutation increased BK currents and the activation kinetics at physiological intracellular Ca²⁺ concentrations with all depolarizing voltages (Li et al., 2018; Moldenhauer et al., 2020), suggesting that the mutation in neurons would shorten action potentials, increase after hyperpolarization and enhance spike firing frequencies (Li et al., 2018). However, the molecular mechanism of N1053S in causing functional changes differs from that of D434G. While G434G primarily enhances Ca²⁺ sensitivity, N1053S does not seem to alter Ca²⁺ dependent activation. The G-V relation of the N1053S BK channels shifts to negative voltages with a similar amount at different Ca²⁺ concentrations, and the deletion of Ca²⁺ binding sites does not prevent the G-V shift of the N1053S BK channels (Li et al., 2018).

Recently another de novo GOF KCNMA1 variant that resulted in a Slo1 N536H mutation (Figure 3A) was identified, with which the patient suffered frequent dystonic/atonic spells (Zhang et al., 2020). She also had disorders of autism spectrum, attention deficit hyperactivity, and intellectual disability. No obvious seizure was observed in the patients. The symptoms were not responsive to antiepileptic medicines, but dextroamphetamine, a central nervous system stimulant, completely controlled her dystonic episodes from >100 per day to none. N536H is also a GOF mutation that increases BK currents at 1 μM Ca²⁺ concentration by shifting G-V relation to negative voltages (Figures 3B,C). Similar to N1053S, but unlike G434G, N536H does not alter Ca²⁺ dependent activation of the BK channel.

The findings so far suggest that the BK channel GOF mutations are associated with absence epilepsy with possible development of myoclonic seizures, movement disorders, or both symptoms. However, it needs to be emphasized that these symptoms vary from patient to patient. The patients may also have cerebellar atrophy, development delays, autism spectrum and intellectual disability to various degrees. These different symptoms derive from increased BK currents either due to increased Ca²⁺ sensitivity or shift of G-V relation with an intact Ca²⁺ dependent activation. The larger BK currents may result in hyperactivity in different neuron types to cause various symptoms. Future studies are needed to dissect how the GOF BK channel mutations with increased Ca²⁺ sensitivity or shift of G-V relation with an intact Ca²⁺ dependent activation lead to different neurological symptoms.

A report in Tabarki et al. (2016) described two sisters harboring a homozygous KCNMA1 variant that resulted in the frameshift mutation T676Lfs^∗7, which changed T676 to L and caused a frame shift after 7 residues downstream (Figure 3A). The patients suffered development delays, severe cerebellar atrophy and seizures of myoclonic type, which progressed to tonic seizures with one of the siblings. The homozygous frameshift mutation presumably abolished the BK channel function although no functional studies were conducted. Several other loss of function (LOF) de novo BK channel mutations that either completely destroyed BK channel function or reduce BK channel currents were subsequently identified in unrelated patients. The mutations that destroyed BK channel function included G375R, S351Y, G356R, N449fs, I663V (Liang et al., 2019) and the truncation mutation R458ter (Yesil et al., 2018). Other mutations, including C413Y and P805L, decrease BK currents either by reduced protein expression or a shift of the G-V relation to the positive voltages (Liang et al., 2019; Figure 3A). These patients suffered neurological diseases that vary in symptoms and severity, most had development delay, intellectual disability, movements disorders (such as ataxia and axial hypotonia), and cerebellar atrophy. Some of the patients also had seizures.

These results reveal that the normal neurological functions are sensitive to BK channel function. Either a GOF or LOF BK channel mutation could tip the balance of neural function and lead to neurological diseases. While the symptoms of both GOF and LOF BK channel mutations show some overlap, such as development delay, intellectual disability, movement disorder, and epilepsy, the LOF BK channel mutations result in more severe cerebellar atrophy. The patients with BK G375R mutation even had visceral and cardiac malformations, connective tissue symptoms, and dysmorphic features, suggesting that BK channel mutations can impact organs beyond the nervous system (Liang et al., 2019). In a recent study a BK LOF mutation, G354S (Figure 3A), was identified in association with cerebellar degeneration, ataxia, developmental delay and intellectual disability in a young girl (Du et al., 2020). The mutation caused a shift of G-V relation of BK channel activation to more negative voltages. However, due to its location in the selectivity filter of the channel the G354S mutation reduced single channel conductance and increased Na⁺ permeability, thereby decreasing the macroscopic currents and making the currents less effective in repolarizing the membrane potential. Viral transfection of the G354S BK in mouse brain induced ataxia in the animals. Transfection of the G354S BK into dividing PC12 cells, the mutant channel suppressed outgrowth of neurites by reducing the neurite length. The mutant BK channel was also toxic to mitochondria, with the cells expressing the G354S BK showing reduced mitochondria content, disrupted mitochondria superstructure, altered mitochondria dynamics to increase fragmented fission forms, and decreased mitochondria membrane potential. NS1619, a BK channel activator, protected the cells expressing G354S BK from the reduced neurite outgrowth, cell death, and changes in mitochondria, suggesting that the recovery of BK channel function can be a therapy for the neurological diseases associated with BK channel LOF mutations.