Introduction

The pore-forming α-subunits of KCa2.x/KCa3.1 channels are encoded by the KCNN genes, including KCNN1 for KCa2.1 (SK1), KCNN2 for KCa2.2 (SK2), KCNN3 for KCa2.3 (SK3), and KCNN4 for KCa3.1 (IK or SK4) channels. The single-channel conductance values of KCa2.x and KCa3.1 channels are ∼10 pS and ∼40 pS in symmetrical K+ solutions, which are much smaller than that of KCa1.1 (BK) channels (~200 pS) channels [1]. These channels were hence named small-conductance (KCa2.x also called SK) and intermediate-conductance (KCa3.1 also called IK) Ca2+-activated K+ channels. KCa2.x/KCa3.1 channels are voltage independent and are activated exclusively by intracellular Ca2+. The sub-micromolar sensitivity to Ca2+ of KCa2.x/KCa3.1 channels is attributed to the Ca2+-binding protein calmodulin (CaM) constitutively associated with the channels [2]. KCa2.x channels are activated by Ca2+ with EC50 values ranging from 300 to 750 nM, while KCa3.1 channels exhibit apparent Ca2+ sensitivity of 100–400 nM [3]. Elevated intracellular Ca2+ levels cause conformational changes in the channel-CaM complex and result in K+ outflow from excitable and non-excitable cells.

Channel structure

KCa2.x/KCa3.1 channels assemble as a homotetramer of four α-subunits of each subtype, although functional heterotetrameric channels among subtypes have also been reported [4, 5]. Each pore-forming α-subunit consists of six transmembrane domains denoted S1–S6. The selective permeability of K+ ions is attributed to the selectivity filter that sits between the S5 and S6 transmembrane domains in the channel pore. The KCa2.x/KCa3.1 channel subtypes are highly homologous in their six transmembrane domains, but the amino acid sequence and length at their cytoplasmic N- and C- termini vary among the subtypes [6, 7].

Within the four KCa2.x/KCa3.1 channel subtypes, atomistic structures are only available for the KCa3.1 channel. The full-length cryogenic electron microscopy (cryo-EM) structures of KCa3.1 channels have been determined in the absence and presence of Ca2+, shedding light on a Ca2+/CaM gating mechanism for these channels [8]. In the absence of Ca2+, the C-lobe of CaM binds to the HA/HB helices in the proximal channel C-terminus, the N-lobe of CaM is highly flexible, and the channel pore is closed. When bound with Ca2+, the N-lobe of CaM becomes well-structured and interacts with the linker between the S4 and S5 transmembrane domains (S4–S5 linker) of a neighboring α-subunit. The interaction between the Ca2+-bound CaM N-lobe and the S4-S5 linker triggers movement of the S6 transmembrane domain and the opening of the channel pore.

Channel physiology

KCa2.x channel subtypes expressed in neurons contribute to the medium afterhyperpolarization and regulate neuronal firing frequency [9]. Intracellular Ca2+ levels may increase as a result of Ca2+ release from the endoplasmic reticulum through ryanodine receptors or IP3 receptors, as well as Ca2+ influx through the voltage- or ligand-gated Ca2+-permeable channels in neurons [9]. KCa2.x channels activated by the elevated Ca2+ levels dampen the neuronal firing frequency, as a mechanism for regulating neuronal excitability by Ca2+ signaling [10, 11].

One well-studied role of KCa2.x channels in the central nervous system is their involvement in the long-term potentiation (LTP). In hippocampal neurons, activating muscarinic acetylcholine receptors increases phosphorylation of the KCa2.2-CaM complex, reduces the channel activity, thus lifting their negative influence on LTP [12, 13]. In animal studies, inhibition of KCa2.x channels improves, while positive modulation of KCa2.x channels impairs learning and memory [9], echoing the essential role of KCa2.x channels in LTP. Detailed information on the physiological role of KCa2.x channels in learning and memory can be found in an excellent review by ref. [9].

KCa2.x channels also play a vital role in controlling the regular tonic firing in cerebellar Purkinje neurons [14]. The KCa2.2 channel is the predominant subtype expressed in cerebellar Purkinje neurons [15,16,17]. The KCa2.2 channel is one of the principal ion channels involved in cerebellar Purkinje neuron pacemaking [14]. Dysfunctional KCa2.2 channel activity has been linked with movement phenotypes including tremors and ataxias in rodents [18] and humans [19]. Positive modulation of KCa2.2 channels has been proposed as a potential therapeutic strategy for ataxias [20].

KCa2.x channels are also expressed in the heart. A higher level of KCa2.x channel protein is expressed in the atria, pace-making cells, and Purkinje fibers than in the ventricles of a normal heart [21,22,23]. KCa2.2 knockout mice exhibit significantly prolonged action potential duration (APD), while overexpression of KCa2.2 channels in atrioventricular nodal cells shortens APD [23]. KCa2.2 [24] and KCa2.3 [25] gene polymorphisms have also been linked with atrial fibrillation through genome-wide association studies (GWAS). There is a very recent review on cardiac KCa2.x channels by ref. [26].

In the vascular endothelium, two-channel subtypes, KCa2.3 and KCa3.1 are expressed [27]. Intracellular Ca2+ levels increase in response to mechano- or chemo-stimuli, activating KCa2.3 and KCa3.1 channels. The opening of KCa2.3 and KCa3.1 channels allows K+ efflux and hyperpolarizes vascular endothelial cells. The hyperpolarization then spreads to the underlying vascular smooth muscle, leading to blood vessel dilation, a phenomenon called endothelium-dependent hyperpolarization (EDH) mediated vasodilation. Mice with a genetic deficit of KCa2.3 and KCa3.1 channels exhibit hypertension [28, 29]. There is a detailed review article by Wulff et al. on the endothelial KCa2.3 and KCa3.1 channels in blood pressure regulation [27].

KCa3.1 channel subtype is different from KCa2.x channels in many ways, including their peripheral expression including the erythrocytes [30] and lymphocytes [31]. In T-lymphocyte, the KCa3.1 channel contributes to electrochemical gradients for Ca2+ influx, which is critical for the proliferation of T cells [32]. Together with the mechanosensitive PIEZO1 channel, the KCa3.1 channel (also called the Gardos channel) regulates red blood cell volume [33].

Channel mutations and genetic disorders

To the best of our knowledge, there is no reported human mutation of KCa2.1 channels linked with genetic disorders. Loss-of-function (LOF) KCa2.2 mutations are associated with neurodevelopmental disorders including cerebellar ataxias [19] and tremors [18, 34]. Gain-of-function (GOF) KCa2.3 mutations are linked with Zimmermann-Laband syndrome (ZLS) [35,36,37] and idiopathic non-cirrhotic portal hypertension (INCPH) [38], while GOF KCa3.1 mutations are associated with a subset of hereditary xerocytosis (HX) [39,40,41,42,43].

LOF KCa2.2 mutations and neurodevelopmental disorders

A LOF rKCa2.2_I289N mutation that diminishes KCa2.2 channel activity has been linked with tremors in rats [18]. Meanwhile, its corresponding hKCa2.2_I288S mutation is causative for neurodevelopmental disorders in humans [19]. In addition, hKCa2.2_L321del, hKCa2.2_I359M, hKCa2.2_Y361C, hKCa2.2_G362S, hKCa2.2_L388V, and hKCa2.2_L432P mutations of the human KCNN2 gene cause neurodevelopmental disorders including cerebellar ataxia, motor and language developmental delay, and intellectual disability [19]. A hKCa2.2_G371E mutation has been linked with tremulous myoclonus-dystonia in humans [34].

Among these mutations, rKCa2.2_I289N [18], hKCa2.2_L321del [19], hKCa2.2_I359M [19], hKCa2.2_G362S [19], hKCa2.2_L388V [19], and hKCa2.2_L432P [19] lead to either reduced or complete loss of channel activity. This became clear when examined via heterologous expression of these mutant channels in cell lines (Table 1). The remaining mutations, hKCa2.2_I288S [19], hKCa2.2_Y361C [19], and hKCa2.2_G371E [34] have not been studied for their effects on channel activity. Mutations that introduce early stop codons, hKCa2.2_Y160* [19], and hKCa2.2_Y267* [19] are assumed to cause premature truncation of the channel protein and are thus classified as LOF.

Table 1 Effects of pathogenic KCa2.2 mutations on channel activity.
Full size table

Most human KCa2.2 mutations are de novo variants (except hKCa2.2_G371E [34] and hKCa2.2_L432P [19]). Patients have heterozygous KCNN2 alleles. KCa2.2 channels are tetramers that consist of four subunits encoded by the KCNN2 gene. rKCa2.2_I289N mutation has been found to suppress the activity of co-expressed wild-type (WT) rKCa2.2 channels and thus deemed dominant-negative [18]. The dominant-negative effect of rKCa2.2_I289N and its equivalent human hKCa2.2_I288S mutation may contribute in part to the autosomal dominance of these mutant alleles [18]. Other human KCa2.2 mutations in Table 1 are also autosomal dominant. It is unknown whether the subunits carrying other mutations in Table 1 can co-assemble with WT subunits in patients with one WT and one mutant allele, let alone the impact of the mutant channel subunits may have when co-expressed with KCa2.2_WT channels.

There are two pathogenic KCa2.2 mutations reported in mice, one missense mKCa2.2_L168P (jitter mice, Mutagenetix database) and one deletion encompassing exon 1 and exon 2 [44] (frissonnant mice). Unlike in rats and humans, these two pathogenic KCa2.2 variants are recessive in mice, which could be explained by distinct baseline KCa2.2 expression levels, gene regulation, or compensatory mechanisms in response to KCa2.2 haploinsufficiency in different species. It is not clear whether the subunits carrying these mouse mutations can co-assemble with WT subunits.

Movement disorders, including cerebellar ataxia, tremor, or extrapyramidal symptoms, are present in patients carrying hKCa2.2_Y267*, hKCa2.2_I288S, hKCa2.2_I359M, hKCa2.2_Y361C, hKCa2.2_L388V, and hKCa2.2_L432P mutations [19]. The mechanism by which LOF KCa2.2 mutations cause cerebellar ataxia symptoms lies in the cerebellar Purkinje neurons. The KCa2.2 channel is one of the principal ion channels involved in the pace-making of cerebellar Purkinje neurons [14]. Dysfunction of KCa2.2 channels may lead to the loss of firing precision in cerebellar Purkinje neurons, similar to what has been reported in the Purkinje neurons of episodic ataxia (EA2) [45, 46], spinocerebellar ataxias (SCA1 [47], SCA2 [48], SCA3 [49], and SCA6 [50]), and Huntington’s disease [51,52,53] mouse models. Irregular neuronal firing in cerebellar Purkinje neurons may subsequently lead to Ca2+ overload and neurodegeneration [20, 54], which may underlie the pathogenesis of ataxias.

Intellectual disability symptoms are present in patients carrying the mutations in Table 1, except hKCa2.2_G371E [34]. In animal studies, overexpression of KCa2.2 channels affects hippocampal synaptic plasticity and impairs learning and memory [55]. Therefore, it is surprising for the KCa2.2 LOF mutations to cause intellectual disability as well. One possible explanation might be that a compensatory event to the LOF KCa2.2 mutations is responsible for the intellectual disability symptoms of these patients, which will require future studies.

Patients carrying the mutations in Table 1 (except hKCa2.2_L321del [19], hKCa2.2_L432P [19], and hKCa2.2_G371E [34]) show psychiatric symptoms, including autistic features, attention deficit hyperactivity disorder, and even psychotic episodes. GWAS has associated the KCNN2 gene encoding the KCa2.2 channel with autistic spectrum disorder (ASD) [56, 57]. Expression meta-analysis has shown increased KCNN2 gene expression in ASD individuals compared with controls [56]. Overexpression of the KCa2.2 channel also underlies cortical dysfunction in a model of PTEN-associated autism [58]. It is thus puzzling that both increased expression and LOF mutations of the KCa2.2 channel are associated with ASD.

GOF KCa2.3 mutations, Zimmermann-Laband syndrome, and idiopathic non-cirrhotic portal hypertension

ZLS [OMIM: 135500] is a rare genetic disorder characterized by gingival enlargement, developmental delay, intellectual disability, together with abnormal fingers, fingernails, nose, and ears. ZLS has been associated with genetic mutations in KCNH1 (encoding an Eag1 K+ channel) [59], KCNK4 (encoding a K2P K+ channel) [60], and most recently KCNN3 (encoding the KCa2.3 channel) [35,36,37] genes. The hKCa2.3_K269E [35], hKCa2.3_A287S [36], hKCa2.3_G350D [35], hKCa2.3_S436C [35], hKCa2.3_A536T [37], hKCa2.3_V539del [36] and hKCa2.3_V555F [36] mutations of the human KCNN3 gene cause ZLS [35,36,37]. The hKCa2.3_V450L [38] mutation is associated with INCPH but not ZLS.

Among these mutations, hKCa2.3_K269E, hKCa2.3_G350D, hKCa2.3_S436C, and hKCa2.3_V450L mutant channels exhibited faster kinetics of current activation by Ca2+ upon break-in with whole-cell patch-clamp recordings than the hKCa2.3_WT [35]. Our group quantitatively determined the apparent Ca2+ sensitivity of these four mutant channels and the hKCa2.3_V555F mutant. They all exhibited increased apparent Ca2+ sensitivity compared with the hKCa2.3_WT in inside-out patch-clamp recordings [61], when examined via heterologous expression of these mutant channels in HEK293 cells. The remaining mutations hKCa2.3_A287S [36], hKCa2.3_A536T [37], and hKCa2.3_V539del [36] have not been studied for their effects on channel activity (Table 2).

Table 2 Effects of pathogenic KCa2.3 mutations on channel activity.
Full size table

The hKCa2.3_V450L [38] genetic mutation in humans is familial, unlike the hKCa2.3_A287S [36] mutation which is unknown. Every other human KCa2.3 mutations in Table 2 are de novo variants. Notably KCNN3 alleles in the patients are heterozygous. Theoretically, the KCa2.3 subunits carrying the mutations in Table 2 can co-assemble with KCa2.3_WT subunits, and form channel tetramers. It is currently unknown what impact the co-assembling of mutant and WT subunits has on channel function.

The expression of KCa2.3 channels in portal veins has not been reported and the mechanism for hKCa2.3_V450L [38] to cause INCPH is unclear. It is well established that KCa2.3 channels are expressed in arterial endothelial cells [27]. The role of KCa2.3 channels in the regulation of veins is less known [62]. It was also speculated that the KCa2.3 channel might play a role in the homeostasis of liver cells, like what had been reported for the KCa3.1 channel [63]. Increased K+ channel activity in liver cells may cause stress and portal hypertension [35].

Based on the contribution of KCa2.3 channels to the EDH-mediated vasodilation, it has been speculated that mutant KCa2.3 channels expressed in vascular endothelium may be related to vascular damage during limb development of ZLS patients [35]. Fluid shear stress may trigger exaggerated vasodilation during human embryonic development because of the excessive hyperpolarization due to hypersensitivity to Ca2+ of the ZLS-related mutant KCa2.3 channels. In critical phases of embryonic development, the consequent edema and vascular ruptures may lead to distal digital hypoplasia with aplastic or hypoplastic nails and terminal phalanges [35].

Intellectual disability and developmental delay symptoms are also reported in patients carrying the KCa2.3 mutations in Table 2, even though the development delay symptom was reported to be mild in patients carrying the hKCa2.3_ A536T mutation [34]. Overexpression of KCa2.3 channels in mice causes hippocampal shrinkage associated with cognitive impairment [64], suggesting the role of KCa2.3 channels in the central nervous system. It is not a surprise to see intellectual disability and developmental delay in patients carrying GOF hKCa2.3 mutations.

Gingival hyperplasia was reported in patients carrying mutations in Table 2, except hKCa2.3_ A287S [36], hKCa2.3_ G350D [35], and hKCa2.3_ A536T [37]. GOF mutations in genes encoding K+ channels including KCNQ1 [65], KCNH1 [59], KCNJ8 [66], and KCNK4 [60] have been associated with hereditary gingival overgrowth. A recent mechanistic study revealed that activation of K+ channels promotes fibrogenic response in hereditary gingival overgrowth via clustering and activation of the small GTP-binding protein Ras [67].

GOF KCa3.1 mutations and hereditary xerocytosis

HX (OMIM 194380) also known as dehydrated hereditary stomatocytosis, is an autosomal dominant congenital hemolytic anemia characterized by erythrocyte dehydration. The majority of HX cases have been linked to GOF mutations of the mechanosensitive cationic PIEZO1 channel in erythrocytes [68]. A small subset (~10%) of HX (also called the Gardos channelopathy) has been linked with hKCa3.1_V282M [69], hKCa3.1_V282E [69], hKCa3.1_S314P [40], hKCa3.1_A322V [41], and hKCa3.1_R352H [39, 43, 70] mutations in the KCa3.1 channel encoded by the KCNN4 gene.

Among these mutations, hKCa3.1_S314P, hKCa3.1_A322V, and hKCa3.1_R352H cause hypersensitivity to Ca2+ of the mutant channels (Table 3) [43, 61]. The remaining mutations hKCa3.1_V282M [69], and hKCa3.1_V282E [69] have not been studied for their effects on the channel’s sensitivity to Ca2+.

Table 3 Effects of pathogenic KCa3.1 mutations on channel activity.
Full size table

All the human KCa3.1 mutations in Table 3 are inherited familial variants. All patients carry one WT and one mutant KCNN4 allele. The KCa3.1 subunits carrying mutations in Table 3 may co-assemble with KCa3.1_WT subunits in the tetrameric channel assembly. The activity of such channels containing both WT and mutant subunits is unknown and requires future studies.

The excessive opening of these mutant KCa3.1 channels could lead to increased K+ efflux, followed by water loss and erythrocyte dehydration [39, 40, 42, 71]. Erythrocytes from patients carrying these KCa3.1 mutations often exhibit decreased K+ content and increased Na+ content, accompanied by increased mean corpuscular hemoglobin concentration resulting from cell dehydration.

Structure–function relationship of the mutant channels

Many of the KCa2.x/3.1 mutation hot spots are located at regions essential for channel gating, including the S4-S5 linker, the selectivity filter, the pore-lining transmembrane S6 domain, and the HA/HB helices. In the S4-S5 linker, hKCa2.2_I288S [19] in humans and its corresponding rKCa2.2_I289N [18] in rats are LOF mutations, while hKCa2.3_S436C and hKCa2.3_V450L are GOF mutations that increase the apparent Ca2+ sensitivity [61]. In the cryo-EM structure of KCa3.1 channels, the N-lobe of CaM forms contacts with the S4-S5 linker when bound with Ca2+, which pulls the pore-forming transmembrane domains to open the gate [8]. Similar interactions between the CaM N-lobe and the S4-S5 linker are also predicted in the homology models of KCa2.2 [72] and KCa2.3 [61] channels. The binding interfaces between CaM and its substrates are often hydrophobic [73]. As such, hKCa2.2_I288S and rKCa2.2_I289N mutations may decrease hydrophobicity at the interface and impair the interactions between the CaM N-lobe and the S4-S5 linker, leading to LOF mutant KCa2.2 channels (Fig. 1). In contrast, hKCa2.3_V450L increases hydrophobicity at the interface and strengthens the interactions between the CaM N-lobe and the S4-S5 linker (Fig. 2), leading to more efficient channel opening and GOF mutant KCa2.3 channels. It is still not clear how hKCa2.3_S436C mutation causes GOF. But mutating its corresponding serine residue in hKCa3.1 (hKCa3.1_S181) to tryptophan or tyrosine amino acid residue increases the hydrophobicity at the interface and causes hypersensitivity to Ca2+ [74].

Fig. 1: KCa2.2 channel structure and LOF mutations.
figure 1

a Human KCa2.2 channel homology model generated using human KCa3.1 channel (PDB: 6cnn) as a template. Pore-forming channel α-subunits are shown in salmon and its accessory CaM is shown in gray. Mutations are shown as spheres with carbon atoms in green, oxygen atoms in red, and nitrogen atoms in dark blue. Mutations in only one of the four α-subunits are shown for clarity. In the inset, the relative position of CaM N-lobe, the S4-S5 linker (S45A and S45B helices), and HA helix is shown. b Mutations in selectivity filter and transmembrane S6 domain of channel pore. Mutations in only one of two opposite α-subunits are shown for clarity. c Schematic representation of one KCa2.2 channel subunit. Pathogenic LOF mutations are shown as green circles. a, b were generated using Pymol (Schrödinger LLC). c was generated using Biorender.com

Full size image
Fig. 2: KCa2.3 channel structure and GOF mutations.
figure 2

a Human KCa2.3 channel homology model generated using human KCa3.1 channel (PDB: 6cnn) as a template. Pore-forming channel α-subunits are shown in pale blue and its accessory CaM is shown in gray. Mutations are shown as spheres with carbon atoms in magenta, oxygen atoms in red and nitrogen atoms in dark blue. Mutations in only one of the four α-subunits are shown for clarity. In the inset, the relative position of CaM N-lobe, the S4-S5 linker (S45A and S45B helices), and HA helix are shown. b Mutations in the transmembrane S6 domain of channel pore. Mutations in only one of two opposite α-subunits are shown for clarity. c Schematic representation of one KCa2.3 channel subunit. Pathogenic GOF mutations are shown as red circles. a, b were generated using Pymol (Schrödinger LLC). c was generated using Biorender.com

Full size image

In the selectivity filter, hKCa2.2_I359M, hKCa2.2_Y361C, hKCa2.2_G362S are LOF mutations (Fig. 1B) [19]. The selectivity filter is well conserved between different K+ channels and mutations in the selectivity filter can often lead to LOF [75], by disrupting the K+ passage.

In the transmembrane S6 domain, hKCa2.2_L388V [19] is a LOF mutation, while hKCa2.3_A536T [37], hKCa2.3_V539del [36], and hKCa3.1_V282M/E [69] are GOF mutations. The V539 residue in hKCa2.3 (hKCa2.3_V539) corresponds to the V282 residue in hKCa3.1 (hKCa3.1_V282). In the cryo-EM structure of KCa3.1, V282 defines the narrowest constriction site of the cytoplasmic gate (Fig. 3) [8]. The replacement of V282 by a glycine residue generates a “leaky” channel that conducts K+ current in the absence of Ca2+ [76]. Theoretically, the GOF hKCa3.1_V282E mutation introduces negatively charged residues at the cytoplasmic gate. This may cause electrostatic repulsion and enlargement of the gate, leading to constitutively active channels that leak K+. How hKCa3.1_V282M mutation causes GOF channel activity still requires investigation.

Fig. 3: KCa3.1 channel structure and GOF mutations.
figure 3

a Human KCa3.1 channel cryo-EM structure (PDB: 6cnn). Pore-forming channel α-subunits are shown in pale green, and its accessory CaM is shown in gray. Mutations are shown as spheres with carbon atoms in magenta, oxygen atoms in red and nitrogen atoms in dark blue. Mutations in only one of the four α-subunits are shown for clarity. In the inset, the relative position of CaM C-lobe and HA/HB helices is shown. b The V282 residue defines the narrowest site of the cytoplasmic gate. c Schematic representation of one KCa3.1 channel subunit. Pathogenic GOF mutations are shown as red circles. a and b were generated using Pymol (Schrödinger LLC). c was generated using Biorender.com

Full size image

In the HA/HB helices, hKCa2.2_L432P [19] is a LOF mutation, while hKCa2.3_V555F, hKCa3.1_S314P, hKCa3.1_A322V, and hKCa3.1_R352H increase the apparent Ca2+ sensitivity [61]. The most studied HA/HB helices mutation is the rKCa2.2_V407F mutation corresponding to hKCa2.3_V555F. The rKCa2.2_V407F mutation increases the apparent Ca2+ sensitivity by enhancing the hydrophobic interactions between the proximal end of HA helix, the S4-S5 linker, and the CaM N-lobe, which may pull the transmembrane S6 domain more efficiently during the pore opening [72]. The hKCa2.3_V555F mutation may increase the hydrophobicity at the proximal end of the HA helix of KCa2.3 channels and enhance its interactions with the S4-S5 linker and the CaM N-lobe in a similar fashion (Fig. 2A). The hKCa2.2_L432P, hKCa3.1_S314P, and hKCa3.1_A322V mutations in the distal HA helix, as well as the hKCa3.1_R352H mutation in the HB helix, are at the interface between the HA/HB helices and the CaM C-lobe (Figs. 1 and 3). Their roles in the channel activation by Ca2+ are less understood. It seems that changes at the interface between the HA/HB helices and the CaM C-lobe may also affect channel activity.

Two GOF mutations, hKCa2.3_K269E and hKCa2.3_G350D, are speculated to interact with casein kinase 2 (CK2) [35]. The KCa2.3_K269E mutation is equivalent to K121 in rKCa2.2 channels that are essential for the CK2 phosphorylation of the rKCa2.2-CaM complex [77]. The positively charged K121 residue in rKCa2.2 channels is not the phosphorylation site. Without the positively charged residue, CK2 cannot phosphorylate the rKCa2.2-CaM complex effectively [77]. The rKCa2.2_K121A mutation diminished the phosphorylation of the rKCa2.2-CaM complex [77]. These two GOF hKCa2.3 mutations may reduce the phosphorylation and negative modulation by CK2 and thus cause Ca2+-hypersensitivity.

One GOF mutation hKCa2.3_A287S is in the transmembrane S1 domain (Fig. 2). The mKCa2.2_L168P mutation identified in jitter mice equivalent to the hKCa2.2_L173P in humans is in the S2 domain (Fig. 1). Transmembrane S1–S4 domains in the voltage-gated K+ (Kv) channels are referred to as the voltage-sensing domain [78]. Unlike the Kv channels, KCa2.x/ KCa3.1 channels are voltage independent. Even though a voltage-sensing role of the S1–S4 domains is not expected in KCa2.x/ KCa3.1 channels, their regulatory role in these voltage-independent channels may still need to be elucidated.

Heteromultimer formed by different channel subtypes

Different KCa2.1, KCa2.2, and KCa2.3 channel subtypes can form heteromultimers in human and mouse atrial myocytes [5]. Heteromultimerization of KCa2.1/KCa3.1 subtypes [4], KCa2.1/KCa2.2 subtypes [79, 80], or KCa2.1/KCa2.2/KCa2.3 subtypes [81] has also been reported in heterologous expression systems. Truncated channel fragments of KCa2.2 [82] or KCa2.3 [83, 84] can suppress the activity of other co-expressed KCa2.x/KCa3.1 subtypes in a dominant-negative fashion, implying the potential heteromultimerization between subtypes.

KCa2.1, KCa2.2, and KCa2.3 subtypes are expressed in the central nervous system [9]. Expression of KCa2.1 and KCa2.2 channels exhibits a partially overlapping distribution pattern in the neocortex and hippocampus. KCa2.3 subtype is predominantly expressed in the basal ganglia, thalamus, and various brain stem nuclei [17]. The distinct and yet partially overlapping expression profiles of the subtypes imply possible heteromultimerization between KCa2.x subtypes in the central nervous system. The LOF KCa2.2 mutations may impact more than the KCa2.2 subtype itself in human brains. Other KCa2.x subtypes expressed in the same type of cells or tissues may be dominant-negatively affected as well. The psychiatric and neurological symptoms observed in patients carrying KCa2.2 LOF mutations may arise from the impaired activity of both KCa2.2 and other co-assembled KCa2.x subtypes. Similarly, the central nervous system symptoms of patients carrying GOF KCa2.3 mutations may be attributed to the elevated activity of both KCa2.3 and other co-assembled KCa2.x subtypes, which will require future studies.

Potential pharmacological therapeutic strategy

Genomic editing done by CRISPR/Cas9 may offer the ultimate cure for these genetic disorders when the technology matures [85]. Until then, pharmacological therapy may fill in the gap. The pharmacology for KCa2.x/KCa3.1 channel subtype has been well developed [3]. Small molecule positive and negative modulators with differential subtype-selectivity are available (Table 4). Senicapoc [86] inhibits KCa3.1 channels with IC50 values of ~11 nM, and selectivity of ~1000-fold for KCa3.1 channels over KCa2.x channel subtypes. Senicapoc has exhibited excellent pharmacokinetic properties in humans [87] and is being studied in a clinical trial (ClinicalTrials.gov Identifier: NCT04372498) for HX patients carrying GOF KCa3.1 channel mutations.

Table 4 Small molecule modulators of KCa2.x and KCa3.1 channels.
Full size table

For treating ZLS and INCPH related to GOF KCa2.3 mutations, negative modulators will be needed. AP14145 is equipotent in inhibiting KCa2.2 and KCa2.3, but is not effective on KCa3.1 channels [88]. We tested AP14145 on the ZLS- and INCPH-related mutant KCa2.3 channels. The inhibitory effect of AP14145 on the mutant channels is somewhat weaker than on the KCa2.3_WT channels [62].

For neurodevelopmental disorders related to LOF KCa2.2 mutations, positive modulators might be beneficial. There is a prototype positive modulator, CyPPA, that potentiates the activity of KCa2.2 and KCa2.3 channels selectively [89]. CyPPA binds to a putative binding pocket at the interface between the HA/HB helices and the constitutively associated CaM [90]. We performed chemical modification of CyPPA and developed several more potent and more selective positive modulators [91]. It is unknown how useful these positive modulators are. More research is needed to determine their effectiveness.