Channelopathy-causing mutations in the S₄₅A/S₄₅B and HA/HB helices of K_Ca2.3 and K_Ca3.1 channels alter their apparent Ca²⁺ sensitivity

Abstract

Small- and intermediate-conductance Ca²⁺-activated potassium (K_Ca2.x and K_Ca3.1) channels are activated exclusively by a Ca²⁺-calmodulin gating mechanism. Wild-type K_Ca2.3 channels have a Ca²⁺ EC₅₀ value of ~0.3 μM, while the apparent Ca²⁺ sensitivity of wild-type K_Ca3.1 channels is ~0.27 μM. Heterozygous genetic mutations of K_Ca2.3 channels have been associated with Zimmermann-Laband syndrome and idiopathic noncirrhotic portal hypertension, while K_Ca3.1 channel mutations were reported in hereditary xerocytosis patients. K_Ca2.3_S436C and K_Ca2.3_V450L channels with mutations in the S₄₅A/S₄₅B helices exhibited hypersensitivity to Ca²⁺. The corresponding mutations in K_Ca3.1 channels also elevated the apparent Ca²⁺ sensitivity. K_Ca3.1_S314P, K_Ca3.1_A322V and K_Ca3.1_R352H channels with mutations in the HA/HB helices are hypersensitive to Ca²⁺, whereas K_Ca2.3 channels with the equivalent mutations are not. The different effects of the equivalent mutations in the HA/HB helices on the apparent Ca²⁺ sensitivity of K_Ca2.3 and K_Ca3.1 channels may imply distinct modulation of the two channel subtypes by the HA/HB helices. AP14145 reduced the apparent Ca²⁺ sensitivity of the hypersensitive mutant K_Ca2.3 channels, suggesting the potential therapeutic usefulness of negative gating modulators.

Graphical Abstract

Introduction

Small- and intermediate-conductance Ca²⁺-activated potassium (K_Ca2.x and K_Ca3.1) channels are voltage independent. Mammalian KCNN genes encode three K_Ca2.x channel subtypes, including K_Ca2.1 (SK1), K_Ca2.2 (SK2), K_Ca2.3 (SK3), and one K_Ca3.1 (IK or SK4) channel subtype. These channels are constitutively associated with the Ca²⁺-binding protein calmodulin (CaM). Both K_Ca2.x and K_Ca3.1 channels are activated by a unique Ca²⁺-CaM gating mechanism.¹

Human heterozygous genetic mutations of K_Ca2.3 channels have been linked to Zimmermann-Laband syndrome (ZLS)² and idiopathic noncirrhotic portal hypertension (INCPH)³. The ZLS-related (K269E, G350D and S436C) and INCPH-related (V450L) mutant channels exhibited faster kinetics of current activation by Ca²⁺ upon break-in with whole-cell patch-clamp recordings, and thus were inferred to have elevated apparent Ca²⁺ sensitivity². The apparent Ca²⁺ sensitivity has not been determined quantitatively on any of these ZLS-related or INCPH-related mutant K_Ca2.3 channels. Meanwhile, heterozygous genetic mutations in K_Ca3.1, which underlies the so-called Gardos channel in erythrocytes, lead to hereditary xerocytosis. The Gardos channelopathy-related (V282M/E⁴, S314P⁵, A322V⁶ and R352H^7–9) mutations were also speculated to cause hypersensitivity to Ca²⁺ of K_Ca3.1 channels, mostly based on more prominent responses of mutant channel current to an inhibitor TRAM-34^5,7,8 or a positive modulator NS309^5,6. The apparent Ca²⁺ sensitivity has only been determined quantitatively on one of these Gardos channelopathy-related mutant channels, K_Ca3.1_R352H⁸.

A recent cryo-EM study revealed that the pore forming K_Ca3.1 channel subunits form a tetrameric structure, with the Ca²⁺ sensor protein CaM constitutively bound (Fig. 1A).¹⁰ In each channel subunit, there are six transmembrane α-helical domains that are denoted S1–S6. CaM interacts with the HA/HB helices, in addition to the linker between the S4 and S5 transmembrane domains (S4–S5 linker).¹⁰ The S4–S5 linker consists of two α-helices, S₄₅A and S₄₅B (Fig. 1B). The HA/HB helices from one channel subunit (green), the S₄₅A and S₄₅B helices from a neighboring channel subunit (yellow) and CaM (grey) closely interact with each other. Three of the Gardos channelopathy-related (S314P⁵, A322V⁶ and R352H^7–9) missense mutations are located in the HA/HB helices, while the fourth one (V282M/E⁴) is located in the transmembrane S5 domain close to the pore.

Fig. 1. Channelopathy-causing mutations in the S₄₅A/S₄₅B helices and HA/HB helices of human K_Ca2.3 and human K_Ca3.1 channels.

(A) The S₄₅A/S₄₅B helices and HA/HB helices of human K_Ca3.1 channel cryo-EM structure with Ca²⁺ (PDB: 6cnn). Four CaM are shown in grey. Four channel subunits are shown in different colors (green, yellow, cyan, and magenta). (B) Ca²⁺ bound CaM (grey), HA and HB helices from one channel subunit (green), the S₄₅A and S₄₅B helices from a neighboring channel subunit (yellow) become a stable structural group. The Gardos channelopathy-related (S314P, A322V and R352H) mutations are shown as spheres (carbon atoms shown in red and nitrogen atoms shown in blue). (C) The S₄₅A/S₄₅B helices and HA/HB helices of a homology model of the human K_Ca2.3 channel was generated using the human K_Ca3.1 channel cryo-EM structure with Ca²⁺. Four CaM are shown in grey. Four channel subunits are shown in different colors (green, yellow, cyan, and magenta). (D) The ZLS-related (S436C) and INCPH-related (V450L) mutations are shown as spheres (carbon atoms shown in red).

Utilizing this cryo-EM structure of human K_Ca3.1 (PDB code: 6CNN) as a template, we generated a homology model of the human K_Ca2.3 channel (Fig. 1C). The INCPH-related (V450L) and one of the ZLS-related (S436C) mutations are located in the S₄₅A/S₄₅B helices (Fig. 1D), while the other two ZLS-related (K269E and G350D) mutations are found in the N-terminus and transmembrane domains, respectively.

We previously identified a valine-to-phenylalanine (V-to-F) mutation in the HA helix that causes Ca²⁺ hypersensitivity in the K_Ca2.1, K_Ca2.2 and K_Ca2.3 subtypes¹¹, but the equivalent V-to-F mutation did not alter the apparent Ca²⁺ sensitivity in K_Ca3.1 channels.¹² In contrast, an asparagine-to-alanine (N-to-A) mutation in the S₄₅A/S₄₅B helices increased the apparent Ca²⁺ sensitivity in both K_Ca2.2a and K_Ca3.1 channel subtypes.¹² Thus, we speculate that the differential modulation of K_Ca2.3 and K_Ca3.1 channels may result more from effects mediated by the HA/HB helices than the S₄₅A/S₄₅B helices. Although K_Ca2.x and K_Ca3.1 channels are activated by the same Ca²⁺-CaM gating mechanism, they exhibit differences in their function, modulation and pharmacology.¹ The K_Ca3.1 structures¹⁰ and K_Ca2.x homology models^12–15 offer an opportunity to perform structure-function studies and examine the differences between the channel subtypes, which might reveal structural information useful for the design of subtype-selective biophysical tools¹² and pharmacological tools¹⁶. To the best of our knowledge, Ca²⁺-hypersensitive human genetic mutations have only been reported in K_Ca2.3 and K_Ca3.1 channels, but not in K_Ca2.1 and K_Ca2.2 channels. Utilizing these disease-causing mutations in the HA/HB helices and the S₄₅A/S₄₅B helices as biophysical tools, we examined whether mutations in corresponding positions might influence K_Ca2.3 and K_Ca3.1 channels differentially.

Here, we quantified the elevated apparent Ca²⁺ sensitivity of K_Ca2.3 channels with the INCPH-related (V450L) and the ZLS-related (S436C) mutations in the S₄₅A/S₄₅B helices. Their equivalent mutations in the S₄₅A/S₄₅B helices of K_Ca3.1 channels also increased the apparent Ca²⁺ sensitivity. On the other hand, the Gardos channelopathy-related (S314P, A322V and R352H) mutations in the HA/HB helices elevated the apparent Ca²⁺ sensitivity of K_Ca3.1 channels, whereas their equivalent mutations in K_Ca2.3 channels did not change the apparent Ca²⁺ sensitivity. A negative gating modulator, AP14145, reduced the apparent Ca²⁺ sensitivity of the INCPH- and ZLS-related mutant K_Ca2.3 channels, suggesting its potential therapeutic usefulness.

Results

Mutations in the S₄₅A/S₄₅B helices of K_Ca2.3 and K_Ca3.1 channels

We previously identified an asparagine-to-alanine (N-to-A) mutation in the S₄₅A/S₄₅B helices that increased the apparent Ca²⁺ sensitivity in both K_Ca2.2a and K_Ca3.1 channel subtypes.¹² We first investigated the equivalent N441A mutation, together with the ZLS-related S436C and INCPH-related V450L mutations in the S₄₅A/S₄₅B helices of K_Ca2.3 channels (Fig. 2A). In the study first describing these disease-associated mutations², the kinetics of current activation by Ca²⁺ upon break-in during whole-cell patch-clamp recordings was used to infer an increase in apparent Ca²⁺ sensitivity. In order to quantify this increase in apparent Ca²⁺ sensitivity, we used the inside-out patch-clamp to measure the Ca²⁺-dependent activation of the mutant channels heterologously expressed in HEK293 cells (Fig. S1). With the inside-out patch configuration, we were able to measure the responses of the channels to various Ca²⁺ concentrations from the same patch and determine the apparent Ca²⁺ sensitivity of the mutant channels. Compared with wild-type (WT) K_Ca2.3 channel (EC₅₀ = 0.30 ± 0.025 μM, n = 8), the S436C, N441A and V450L mutations effectively elevated the K_Ca2.3 channel apparent Ca²⁺ sensitivity (Fig. 2B). EC₅₀ values for Ca²⁺ were 0.087 ± 0.0058 μM (n = 6, P < 0.0001), 0.064 ± 0.0098 μM (n = 6, P < 0.0001), and 0.15 ± 0.013 μM (n = 6, P < 0.0001), respectively (Fig. 2C and Table 1).

Fig. 2. The ZLS-related (S436C) and INCPH-related (V450L) mutations caused Ca²⁺ hypersensitivity of K_Ca2.3 channels.

(A) Amino acid sequence alignment of human K_Ca2.3 [GenBank: NP_002240.3] and human K_Ca3.1 [GenBank: NP_002241.1] at the S₄₅A and S₄₅B helices (highlighted in yellow). Amino acid residues mutated are shown in red font. (B) Activation by Ca²⁺ of the WT and mutant K_Ca2.3 channels carrying mutations in the S₄₅A/S₄₅B helices. (C) EC₅₀ values for activation by Ca²⁺ of the WT and mutant K_Ca2.3 channels. *** P < 0.001 compared with WT. No asterisk means no statistical significance compared with WT.

Table 1.

Apparent Ca²⁺ sensitivity of WT and mutant channels.

KCa2.3	EC50 for Ca2+ (μM)	Hill coefficient	Statistics	KCa3.1	EC50 for Ca2+ (μM)	Hill coefficient	Statistics
KCa2.3_WT	0.30 ± 0.025	3.88 ± 0.59	NA	KCa3.1_WT	0.27 ± 0.012	2.87 ± 0.48	N/A
KCa2.3_WT + AP14145	1.08 ± 0.10	1.34 ± 0.14	P < 0.0001 vs KCa2.3_WT
KCa2.3_K269E	0.086 ± 0.013	3.45 ± 0.46	P < 0.0001 vs KCa2.3_WT
KCa2.3_K269E + AP14145	0.14 ± 0.011	3.15 ± 0.72	P < 0.05 vs KCa2.3_K269E
KCa2.3_G350D	0.12 ± 0.013	4.19 ± 0.43	P < 0.0001 vs KCa2.3_WT
KCa2.3_G350D + AP14145	0.27 ± 0.025	3.86 ± 1.09	P < 0.0001 vs KCa2.3_G350D
KCa2.3_S436C	0.087 ± 0.0058	4.37 ± 0.85	P < 0.0001 vs KCa2.3_WT	KCa3.1_S181C	0.10 ± 0.011	2.78 ± 0.25	P < 0.0001 vs KCa3.1_WT
KCa2.3_S436C + AP14145	0.19 ± 0.026	2.28 ± 0.45	P < 0.01 vs KCa2.3_S436C
KCa2.3_N441A	0.064 ± 0.0098	4.21 ± 1.09	P < 0.0001 vs KCa2.3_WT	KCa3.1_N186A	0.10 ± 0.0055	2.66 ± 0.45	P < 0.0001 vs KCa3.1_WT
KCa2.3_V450L	0.15 ± 0.013	4.15 ± 0.58	P < 0.001 vs KCa2.3_WT	KCa3.1_V195L	0.20 ± 0.020	4.08 ± 0.89	P < 0.01 vs KCa3.1_WT
KCa2.3_V450L + AP14145	0.31 ± 0.061	3.01 ± 0.62	P < 0.05 vs KCa2.3_V450L
KCa2.3_V555F	0.067 ± 0.0047	3.77 ± 0.20	P < 0.0001 vs KCa2.3_WT	KCa3.1_V298F	0.32 ± 0.032	3.39 ± 0.43	P > 0.05 vs KCa3.1_WT
KCa2.3_V555F + AP14145	0.14 ± 0.015	2.81 ± 0.33	P < 0.001 vs KCa2.3_V555F
KCa2.3_A571P	0.30 ± 0.027	2.75 ± 0.25	P > 0.9999 vs KCa2.3_WT	KCa3.1_S314P	0.064 ± 0.0033	4.07 ± 0.41	P < 0.0001 vs KCa3.1_WT
KCa2.3_T579V	0.25 ± 0.032	2.79 ± 0.47	P = 0.5666 vs KCa2.3_WT	KCa3.1_A322V	0.059 ± 0.0039	3.94 ± 0.29	P < 0.0001 vs KCa3.1_WT
KCa2.3_R612H	0.27 ± 0.014	2.78 ± 0.23	P = 0.8339 vs KCa2.3_WT	KCa3.1_R352H	0.085 ± 0.0081	3.66 ± 0.55	P < 0.0001 vs KCa3.1_WT

We then tested their equivalent mutations in K_Ca3.1 channels (Fig. S2). Three equivalent mutations S181C, N186A and V195L also increased apparent Ca²⁺ sensitivity of K_Ca3.1 channels (Fig. 3A). EC₅₀ values for Ca²⁺ of the three mutants were 0.10 ± 0.011 μM (n = 7, P < 0.0001), 0.10 ± 0.0055 μM (n = 5, P < 0.0001), and 0.20 ± 0.020 μM (n = 5, P < 0.01), compared with WT K_Ca3.1 channels (EC₅₀ = 0.27 ± 0.012 μM, n = 8, Fig. 3B and Table 1).

Fig. 3. The mutations equivalent to K_Ca2.3_S436C and K_Ca2.3_V450L caused Ca²⁺ hypersensitivity of K_Ca3.1 channels.

(A) Concentration-dependent activation by Ca²⁺ of the WT and mutant K_Ca3.1 channels carrying mutations in the S₄₅A/S₄₅B helices. (B) EC₅₀ values for the activation of the WT and mutant K_Ca3.1 channels by Ca²⁺. ** P < 0.01, *** P < 0.001 compared with WT channels. No asterisk means no statistical significance compared with respective WT channels.

Mutations in the HA/HB helices of K_Ca2.3 and K_Ca3.1 channels

In our previous study, a valine-to-phenylalanine (V-to-F) mutation in the HA helix caused Ca²⁺ hypersensitivity in the K_Ca2.1, K_Ca2.2 and K_Ca2.3 subtypes¹¹, but the equivalent V-to-F mutation did not alter the apparent Ca²⁺ sensitivity in K_Ca3.1 channels.¹² Interestingly, a heterozygous V-to-F mutation in K_Ca2.3 (V555F) in a young child linked to ZLS was subsequently reported.¹⁷ We here investigated this newly reported ZLS-related mutation hypothesizing that it would display increased apparent Ca²⁺ sensitivity.

We investigated the K_Ca3.1_V298F mutation, together with the Gardos channelopathy-related (S314P⁵, A322V⁶ and R352H^7–9) missense mutations located in the HA/HB helices of K_Ca3.1 channels (Fig. 4A and Fig. S3). Three Gardos channelopathy-related S314P, A322V and R352H mutations increased the apparent Ca²⁺ sensitivity of K_Ca3.1 channels (Fig. 4B). The EC₅₀ values for Ca²⁺ of the three mutants were 0.064 ± 0.0033 μM (n = 9, P < 0.0001), 0.059 ± 0.0039 μM (n = 8, P < 0.0001), and 0.085 ± 0.0081 μM (n = 8, P < 0.0001), compared with WT K_Ca3.1 channels (EC₅₀ = 0.27 ± 0.012 μM, n = 8, Fig. 4C and Table 1). Consistent with our previous report¹², the V298F mutation did not change the apparent Ca²⁺ sensitivity compared with the WT K_Ca3.1 channels (Fig. 4C and Table 1).

Fig. 4. The Gardos channelopathy-related (S314P, A322V and R352H) mutations elevated Ca²⁺ sensitivity of K_Ca3.1 channels.

(A) Amino acid sequence alignment of human K_Ca2.3 [GenBank: NP_002240.3] and human K_Ca3.1 [GenBank: NP_002241.1] at the HA and HB helices (highlighted in green). Amino acid residues mutated are shown in red font. (B) Concentration-dependent activation by Ca²⁺ of the WT and mutant K_Ca3.1 channels carrying mutations in the HA/HB helices. (C) EC₅₀ values for the activation of the WT and mutant K_Ca3.1 channels by Ca²⁺. *** P < 0.001 compared with WT channels. No asterisk means no statistical significance compared with respective WT channels.

We next tested the equivalent mutations in K_Ca2.3 channels (Fig. S4). Three equivalent mutations A571P, T579V and R612H did not alter the K_Ca3.1 channel apparent Ca²⁺ sensitivity (Fig. 5A). The EC₅₀ values for Ca²⁺ of the three mutants were 0.30 ± 0.027 μM (n = 8, P > 0.9999), 0.25 ± 0.032 μM (n = 6, P = 0.5666), and 0.27 ± 0.014 μM (n = 11, P = 0.8339), compared with WT K_Ca2.3 channels (EC₅₀ = 0.30 ± 0.025 μM, n = 8, Fig. 5B and Table 1). On the other hand, the V555F mutation increased the apparent Ca²⁺ sensitivity compared with the WT K_Ca2.3 channels (Fig. 5B and Table 1).

Fig. 5. The mutations equivalent to K_Ca3.1_S314P, K_Ca3.1_A322V and K_Ca3.1_R352H did not change apparent Ca²⁺ sensitivity of K_Ca2.3 channels.

(A) Activation by Ca²⁺ of the WT and mutant K_Ca2.3 channels carrying mutations in the HA/HB helices. (B) EC₅₀ values for activation by Ca²⁺ of the WT and mutant K_Ca2.3 channels. *** P < 0.001 compared with WT. No asterisk means no statistical significance compared with WT.

We then tested another two ZLS-related mutations, K269E in the N-terminus (Fig. 6A) and G350D in the transmembrane domains (Fig. 6B) of K_Ca2.3 channels. Both K269E and G350D mutations increased the apparent Ca²⁺ sensitivity of K_Ca2.3 channels, with EC₅₀ values of 0.086 ± 0.013 μM (n = 6, P < 0.0001) and 0.12 ± 0. 013 μM (n = 8, P < 0.0001), respectively (Table 1).

Fig. 6. The effects of AP14145 on the Ca²⁺-dependent activation of WT and mutant K_Ca2.3 channels.

Activation by Ca²⁺ of the K_Ca2.3_K269E (A), K_Ca2.3_G350D (B), K_Ca2.3_S436C (C), K_Ca2.3_V450L (D), K_Ca2.3_V555F (E) and K_Ca2.3_WT (F) channels.

Effects of a negative gating modulator on the mutant K_Ca2.3 channels

AP30663¹⁸, a negative gating modulator of K_Ca2.3 channels, is currently in early clinical development for conversion of atrial fibrillation¹⁹. AP30663 is not commercially available. To explore potential pharmacological therapies for the diseases caused by mutant K_Ca2.3 channels, we tested the effects of a structurally related compound, AP14145²⁰, on the WT and mutant K_Ca2.3 channels (Fig. 6A–F). AP14145 (10 μM) right-shifted the Ca²⁺ concentration-response curves of all the mutant and WT channels, indicating its effectiveness as a negative gating modulator. AP14145 induced a ~3.6-fold decrease in the apparent Ca²⁺ sensitivity of K_Ca2.3_WT channels (Fig. 6F, Table 1). Its effects on the mutant channels were somewhat less prominent. AP14145 changed the EC₅₀ values for Ca²⁺ of K_Ca2.3_K269E, K_Ca2.3_G350D, K_Ca2.3_S436C, K_Ca2.3_V450L and K_Ca2.3_V555F channels by ~1.63, ~2.25, ~2.18, ~2.07 and ~2.09 folds, respectively (Fig. 6A–E, Table 1).

Discussion

K_Ca2.x/K_Ca3.1 channels are a unique group of potassium channels that are voltage-independent.²¹ Mutations in both K_Ca2.3 and K_Ca3.1 channels have been linked with genetic human disorders. We here quantified the effects of these mutations on the apparent Ca²⁺ sensitivity of the channels. The Gardos channelopathy-related K_Ca3.1_S314P⁵, K_Ca3.1_A322V⁶ and K_Ca3.1_R352H^7–9 mutations in the HA/HB helices caused hypersensitivity to Ca²⁺, which could lead to increased K⁺ outflow and erythrocyte dehydration^5,7,22,23.

The ZLS- and INCPH-related mutations in the S₄₅A/S₄₅B helices increased the apparent Ca²⁺ sensitivity of K_Ca2.3 channels. Exactly how the Ca²⁺ hypersensitivity of K_Ca2.3 channels leads to ZLS and INCPH is unclear. Increased K_Ca2.3 channel function in the central nervous system might be related to the defective cognitive function.²⁴ Elevated K_Ca2.3 channel activity in the vascular endothelium is speculated to be involved in distal digital hypoplasia². A previous report² demonstrated faster time courses of channel activation upon breakthrough of the cell membrane to the whole-cell configuration with pipette solution containing 1 μM of Ca²⁺, implying higher apparent Ca²⁺ sensitivity of the ZLS- and INCPH-related mutant K_Ca2.3 channels. In this study, we quantitatively determined the EC₅₀ values for Ca²⁺ of these mutant channels, using the inside-out configuration (Table 1). They are indeed hypersensitive to Ca²⁺, with the ZLS-related mutants being more sensitive to Ca²⁺ than the INCPH-related mutant channel. The reported ZLS² and INCPH³ patients have heterozygous mutations in their KCNN3 alleles. Their K_Ca2.3 channels might be comprised of both WT and mutant subunits in the tetrameric channel architecture, which may be speculated to have apparent Ca²⁺ sensitivity higher than the WT but lower than the mutant channels reported in this study (Table 1).

The amino acid sequences of K_Ca2.3 at the S₄₅A and S₄₅B helices shares ~59% identity with the K_Ca3.1 channel (highlighted in yellow, Fig. 2A). These two channel subtypes share ~60% identity at the HA/HB helices (highlighted in green, Fig. 4A). When bound with Ca²⁺, CaM forms a stable structural group with the HA/HB helices, and the S₄₅A/S₄₅B helices from a neighboring K_Ca3.1 channel subunit (Fig. 1A,B)¹⁰. A homology model of the human K_Ca2.3 channels showed similar interactions (Fig. 1C,D). Our previous study addressed the crucial role of the interactions between CaM, the HA/HB helices and the S₄₅A/S₄₅B helices for the apparent Ca²⁺ sensitivity of the channels.¹²

In this study, we demonstrated that equivalent mutations in the S₄₅A/S₄₅B helices of K_Ca2.3 (K_Ca2.3_S436C, K_Ca2.3_N441A and K_Ca2.3_V450L) and K_Ca3.1 (K_Ca3.1_S181C, K_Ca3.1_N186A and K_Ca3.1_V195L) channels elevated the apparent Ca²⁺ sensitivity of both channel subtypes (Fig. 2 and Fig. 3). In contrast, mutations in the HA/HB helices of K_Ca3.1 (K_Ca3.1_S314P, K_Ca3.1_A322V and K_Ca2.3_R352H) channels caused Ca²⁺ hypersensitivity (Fig. 4), whereas their equivalent mutations in K_Ca2.3 (K_Ca2.3_A571P, K_Ca2.3_T579V and K_Ca2.3_R612H) did not alter the apparent Ca²⁺ sensitivity (Fig. 5). Furthermore, a mutation in the HA/HB helices of K_Ca2.3 (K_Ca2.3_V555F) channels caused Ca²⁺ hypersensitivity (Fig. 5), whereas its equivalent mutations in K_Ca3.1 (K_Ca3.1_V298F) did not change the apparent Ca²⁺ sensitivity (Fig. 4). As such, more synchronized changes in the apparent Ca²⁺ sensitivity of K_Ca2.3 and K_Ca3.1 channel subtypes were observed in response to equivalent mutations in the S₄₅A/S₄₅B helices. In contrast, these two channel subtypes often exhibited differential responses to equivalent mutations in the HA/HB helices. Given the involvement of K_Ca2.3 and K_Ca3.1 channels in genetic disorders, understanding the regulation of their apparent Ca²⁺ sensitivity is critical. With cryo-EM structures only available for K_Ca3.1 channels¹⁰, caution will be needed when performing structure-function studies of K_Ca2.x channel subtypes, especially for the HA/HB helices.

It was reported that the apparent Ca²⁺ sensitivity of K_Ca2.2 channels can be negatively regulated by the phosphorylation of CaM.²⁵ Indeed, the apparent Ca²⁺ sensitivity of all K_Ca2.x/K_Ca3.1 channel subtypes can be reduced by the phosphorylation of CaM.²⁶ As such, the EC₅₀ values for Ca²⁺ determined in HEK293 cells may deviate from the apparent Ca²⁺ sensitivity of the channels expressed in other cell types (e.g. endothelial cells), as a result of the phosphorylation status.²⁶ K_Ca3.1 is unique among these four subtypes in that its activation can be enhanced by the phosphorylation of a histidine residue (H358) in the HB helix²⁷, possibly by antagonizing copper-mediated inhibition of the channel.²⁸ This unique regulation of K_Ca3.1 by the phosphorylation of H358 in the HB helix may also echo this notion that K_Ca3.1 channels are modulated differently by the HA/HB helices from K_Ca2.x channels.

The pharmacology for K_Ca3.1 channel subtype has been well developed¹. Senicapoc²⁹ and TRAM-34³⁰ inhibit K_Ca3.1 channels with IC₅₀ values of ~11 nM and ~20 nM, respectively, with selectivity (200–1000-fold) towards IK channels over other channel subtypes. With excellent pharmacokinetic properties in humans, senicapoc has become a potential candidate for the treatment of Gardos channelopathy, a subset of hereditary xerocytosis.³¹ A clinical trial (ClinicalTrials.gov Identifier: NCT04372498) is studying senicapoc in patients with hereditary xerocytosis caused by K_Ca3.1 channel mutations. On the other hand, no pharmacological treatment has been explored for ZLS and INCPH caused by K_Ca2.3 channel mutations. AP14145, which is equipotent in modulating K_Ca2.2 and K_Ca2.3²⁰, with no effect on K_Ca3.1, has been reported to treat atrial fibrillation in pigs³² and goats³³. A structurally related compound, AP30663¹⁸, is currently in early clinical development for conversion of atrial fibrillation¹⁹. Our results indicated that AP14145 right-shifted the Ca²⁺ concentration-response curves (Fig. 6) and reduced the apparent Ca²⁺ sensitivity (Table 1) of the mutant K_Ca2.3 channels. The K_Ca2.3_V450L mutant channel with an EC₅₀ value for Ca²⁺ of 0.15 ± 0.013 μM (Table 1) is associated with INCPH. This patient with K_Ca2.3_V450L mutation was not classified as ZLS³. Among the ZLS-related mutant channels, K_Ca2.3_G350D is the least sensitive to Ca²⁺ with an EC₅₀ value of 0.12 ± 0.013 μM (Table 1). As such, it can be inferred that mutations that elevate apparent Ca²⁺ sensitivity to < 0.12 μM can be associated with ZLS. AP14145 was able to reduce the apparent Ca²⁺ sensitivity of the ZLS-related mutant channels to 0.14 ± 0.011 μM (K_Ca2.3_K269E + AP14145), 0.27 ± 0.025 μM (K_Ca2.3_G350D + AP14145), 0.19 ± 0.026 μM (K_Ca2.3_S436C + AP14145) and 0.14 ± 0.015 μM (K_Ca2.3_V555F + AP14145, Table 1), implying the potential therapeutic usefulness of negative K_Ca2 channel gating modulators. The KCNN3 gene encoding K_Ca2.3 channels has only very recently been associated with ZLS² and INCPH³. To the best of our knowledge, there are currently no transgenic mouse models for these disorders. Future development of animal models will enable preclinical studies for therapeutic options.

Materials and Methods

Electrophysiology

The effect of mutations on the apparent Ca²⁺ sensitivity of K_Ca2.3 and K_Ca3.1 channels was investigated as previously described.^12,15,26 Briefly, mutations were introduced to human K_Ca2.3 and human K_Ca3.1 channels using QuickChange II site-directed mutagenesis kit (Agilent) or through molecular cloning services (Genscript). The WT and mutant channel cDNAs, constructed in the pIRES2-AcGFP1 vector (Clontech) were transfected into HEK293 cells by the calcium–phosphate method. AP14145 was purchased from Tocris Bioscience. K_Ca2.3 and K_Ca3.1 currents were recorded 1–2 days after transfection, with an Axon200B amplifier (Molecular Devices) at room temperature.

pClamp 10.5 (Molecular Devices) was used for data acquisition and analysis. The resistance of the patch electrodes ranged from 3–5 MΩ. The pipette solution contained (in mM): 140 KCl, 10 Hepes (pH 7.4), 1 MgSO₄. The bath solution containing (in mM): 140 KCl, 10 Hepes (pH 7.2), 1 EGTA, 0.1 Dibromo-BAPTA, and 1 HEDTA was mixed with Ca²⁺ to obtain the desired free Ca²⁺ concentrations, calculated using the software by Chris Patton of Stanford University (https://somapp.ucdmc.ucdavis.edu/pharmacology/bers/maxchelator/webmaxc/webmaxcS.htm). The Ca²⁺ concentrations were verified using a Ca²⁺ calibration buffer kit (Thermo Fisher Scientific). Briefly, a standard curve was generated using the Ca²⁺ buffers from the kit and a fluorescence Ca²⁺ indicator. Then the Ca²⁺ concentrations of our bath solutions were determined through interpolation on the standard curve.

Currents were recorded using an inside-out patch configuration. The inside-out patch configuration allows us to measure the response of the channels to various Ca²⁺ concentrations from the same patch and effectively compare the apparent Ca²⁺ sensitivity of WT and mutant channels. The intracellular face was initially exposed to a zero-Ca²⁺ bath solution, and subsequently to bath solutions with a series of Ca²⁺ concentrations. Currents were recorded by repetitive 1-s-voltage ramps from − 100 mV to + 100 mV from a holding potential of 0 mV. One minute after switching of bath solutions, ten sweeps with a 1-s interval were recorded. The integrity of the patch was examined by switching the bath solution back to the zero-Ca²⁺ buffer. Data from patches, which did not show significant changes in the seal resistance after solution changes, were used for further analysis. To construct the concentration-dependent positive modulation of channel activities, the current amplitudes at − 90 mV in response to various concentrations of Ca²⁺ were normalized to that obtained at maximal concentration of Ca²⁺. The normalized currents were plotted as a function of the concentrations of Ca²⁺. EC₅₀ values and Hill coefficients were determined by fitting the data points to a standard concentration–response curve (Y = 100/(1 + (X/EC₅₀)^ − Hill)). The data analysis was performed using pClamp 10.5 (Molecular Devices) in a blinded fashion. All data are shown as mean ± SEM unless otherwise indicated. One-way ANOVA and Tukey’s post hoc tests were used for data comparison of three or more groups. The Student’s t-test was used for data comparison if there were only two groups.

Computational modeling

We built homology model for the human K_Ca2.3 channel based on a cryo-EM structure of K_Ca3.1 channel (PDB Code: 6CNN). Sequence alignment among K_Ca2.3 and K_Ca3.1 channels was generated by Clustal Omega server (https://www.ebi.ac.uk/Tools/msa/clustalo/). The sequence identity between K_Ca2.3 and K_Ca3.1 is 46.6%, which makes the K_Ca3.1 Cryo-EM structure an excellent structural template to generate homology models for the K_Ca2.3 channels. We used the MODELLER program³⁴ to generate 10 initial homology models for K_Ca2.3 channel based on the K_Ca3.1 structural template, and selected the one with the best internal DOPE score from the program.

Highlights.

• Disease-causing mutations elevated the apparent Ca²⁺ sensitivity of K_Ca2.3 channels

• Disease-causing mutations elevated the apparent Ca²⁺ sensitivity of K_Ca3.1 channels

• K_Ca2.3 and K_Ca3.1 channels are regulated differentially by the HA/HB helices

• AP14145 reduced the Ca²⁺ sensitivity of disease related mutant K_Ca2.3 channels

Abbreviations

N-to-A

Asparagine-to-alanine

Ca²⁺

Calcium

CaM

Calmodulin

INCPH

Idiopathic noncirrhotic portal hypertension

K_Ca3.1 or IK

Intermediate-conductance Ca²⁺-activated K⁺ channels

K_Ca2.x or SK

Small-conductance Ca²⁺-activated K⁺ channels

V-to-F

Valine-to-phenylalanine

WT

Wild-type

ZLS

Zimmermann-Laband syndrome