An allosteric modulator activates BK channels by perturbing coupling between Ca2+ binding and pore opening

Zhang, Guohui; Xu, Xianjin; Jia, Zhiguang; Geng, Yanyan; Liang, Hongwu; Shi, Jingyi; Marras, Martina; Abella, Carlota; Magleby, Karl L.; Silva, Jonathan R.; Chen, Jianhan; Zou, Xiaoqin; Cui, Jianmin

doi:10.1038/s41467-022-34359-6

Download PDF

Article
Open access
Published: 09 November 2022

An allosteric modulator activates BK channels by perturbing coupling between Ca²⁺ binding and pore opening

Nature Communications volume 13, Article number: 6784 (2022) Cite this article

2732 Accesses
6 Citations
2 Altmetric
Metrics details

Subjects

Abstract

BK type Ca²⁺-activated K⁺ channels activate in response to both voltage and Ca²⁺. The membrane-spanning voltage sensor domain (VSD) activation and Ca²⁺ binding to the cytosolic tail domain (CTD) open the pore across the membrane, but the mechanisms that couple VSD activation and Ca²⁺ binding to pore opening are not clear. Here we show that a compound, BC5, identified from in silico screening, interacts with the CTD-VSD interface and specifically modulates the Ca²⁺ dependent activation mechanism. BC5 activates the channel in the absence of Ca²⁺ binding but Ca²⁺ binding inhibits BC5 effects. Thus, BC5 perturbs a pathway that couples Ca²⁺ binding to pore opening to allosterically affect both, which is further supported by atomistic simulations and mutagenesis. The results suggest that the CTD-VSD interaction makes a major contribution to the mechanism of Ca²⁺ dependent activation and is an important site for allosteric agonists to modulate BK channel activation.

The binding and mechanism of a positive allosteric modulator of Kv3 channels

Article Open access 21 March 2024

Calcium-gated potassium channel blockade via membrane-facing fenestrations

Article 31 August 2023

A small-molecule activation mechanism that directly opens the KCNQ2 channel

Article 02 January 2024

Introduction

BK type Ca²⁺ activated K⁺ channels are important in regulating neural excitability^1,2,3, neurotransmitter release⁴, and muscle contraction^5,6. Aberrant function of BK channels is associated with various human diseases such as neurological disorders^7,8,9, hypertension¹⁰ and impaired urinary control¹¹. The function of BK channels relies on their activation by both elevated intracellular Ca²⁺ and depolarization of the membrane voltage. The opening of BK channels provides a negative feedback mechanism to hyperpolarize the membrane and reduce intracellular Ca²⁺ by deactivating voltage-dependent Ca²⁺ channels¹². The Ca²⁺- and voltage-dependent activation of BK channels has been an important model system for understanding allosteric mechanisms of ion channel gating¹³. Unlike canonical voltage gated K⁺ channels, such as Shaker, BK channels exhibit an intrinsic open probability (P_o ~ 10⁻⁶) in the absence of stimuli. Ca²⁺ binding or voltage sensor activation allosterically activates the channel by favoring the open state^13,14. The mechanisms of voltage and Ca²⁺-dependent BK channel activation have been described by an allosteric model¹³, which has successfully accounted for the functional properties of BK channel activation, pharmacology¹⁵, modulation by regulatory subunits¹⁶ and mutations^17,18. The molecular bases of these allosteric activation mechanisms are not clear. Understanding these molecular mechanisms may help reveal the physiological role of BK channels, facilitate the development of therapeutic agents that modulate BK channel activation, and elucidate fundamental mechanisms of allosteric modulation.

BK channels are formed by four Slo1 subunits with structural domains that are associated with distinct functions^19,20,21 (Fig. 1a). In the membrane, a central pore is formed by the transmembrane segments S5–S6 from all four subunits. Surrounding the pore are the voltage sensor domains (VSDs) from the four subunits, comprised of the transmembrane segments S1–S4. Slo1 also contains an additional transmembrane segment S0. The cytosolic tail domain (CTD) of each Slo1 subunit harbors two Ca²⁺-binding sites, and the four CTDs, each containing two structural motifs known as the regulator of potassium conductance (RCK), form a ring-like structure, known as the gating ring. Studies have shown that Ca²⁺ and voltage activate BK channels with distinct molecular mechanisms such that depolarization can activate the channel in the absence of Ca²⁺ binding²², while Ca²⁺ binding can activate the channel when the voltage sensor is at the resting state^13,14,23. In these activation mechanisms Ca²⁺ or voltage alter the conformation of their respective sensors, and the conformational change is then coupled to the opening of the activation gate via pathways of intramolecular interactions in the channel protein²⁴. Although the sensors for Ca²⁺ and voltage²⁵ and the properties of the activation gate^25,26 have been identified, the allosteric coupling mechanisms between the sensors and the activation gate in BK channels are still under investigation^27,28. One of the long-standing issues is to identify the paths of intramolecular interactions within the channel protein by which cytosolic Ca²⁺ binding or voltage sensor activation opens the channel pore across the membrane^{19,20,23,27,28,29,30,31,32,33}. Another issue is that both Ca²⁺ and voltage open the same activation gate. Therefore, the respective paths for mediating voltage and Ca²⁺-dependent channel opening must merge eventually. Although Ca²⁺ and voltage activate BK channels via distinct mechanisms, the two activation mechanisms influence each other allosterically^13,34. Where and how the allosteric paths for Ca²⁺- and voltage-dependent activation merge remains a fundamental question.

In BK channels while the CTD serves as the Ca²⁺ sensor it was also shown to contribute to voltage-dependent gating^28,31,34, and CTD conformation was altered by voltage-dependent activation³⁵. For Ca²⁺ binding at the CTD and for the VSD activation to open the pore, interactions among the CTD, VSD and the pore domains are essential. Although the structural domain interactions that are required for voltage and Ca²⁺-dependent activation have been intensively studied, mechanistic roles of these domain interactions remain poorly understood. In the Slo1 subunit the CTD is covalently connected to the S6 by a peptide of 15 amino acids, known as the C-linker. A previous study showed that deletion or insertion of C-linker residues altered Ca²⁺- and voltage-dependent activation in a manner that correlated with the C-linker length, suggesting that Ca²⁺ binding to the CTD may open the channel by “tugging” the C-linker²⁹. On the other hand, the CTD also has a non-covalent interaction with the VSD of its neighboring subunit (Fig. 1a). This interaction was first revealed by the discovery that the CTD and VSD form a close interface because residues from both the CTD and VSD of the neighboring subunits form a Mg²⁺-binding site, which differs from the Ca²⁺-binding sites in CTD. Mg²⁺ binding activates the channel by enhancing voltage-dependent activation via an electrostatic interaction with the voltage sensor³⁶. The interactions between CTD and VSD were then more clearly shown by BK channel structures^19,20. Based on the structural evidence, it was hypothesized that the CTD-VSD interaction may contribute to the coupling between Ca²⁺ binding and pore opening³⁰. This notion was strongly supported by the fact that the comparison between the structure of the BK channel with Ca²⁺/Mg²⁺ bound and that of metal-free showed a large Ca²⁺/Mg²⁺-dependent change in the interface between CTD and VSD^19,20. This hypothesis is also consistent with an earlier finding that the region of the CTD located close to the VSD is important in determining the different Ca²⁺ sensitivities among various BK channel homologs³⁷. Furthermore, mutations at the CTD-VSD interface were also found to alter Ca²⁺ sensitivity^27,38. However, these mutations either had small effects on Ca²⁺-dependent activation³⁸ or, besides reducing Ca²⁺-dependent activation, also altered voltage-dependent activation and resulted in an increase of the intrinsic open probability of the channel²⁷. Did these mutations alter the Ca²⁺-dependent activation or did some of these mutations alter the voltage-dependent activation by affecting the voltage sensor, which then indirectly affect Ca²⁺-dependent activation? Thus, whether the CTD-VSD interaction as revealed by structural data mediates coupling of Ca²⁺ binding to pore opening remains unclear.

In this study we use in silico screening to identify a compound that binds to a site at the CTD–VSD interface and enhances BK channel activation. We find that the compound specifically enhances Ca²⁺-dependent activation by perturbing the pathway for coupling between Ca²⁺ binding and pore opening. These results support the idea that the non-covalent CTD–VSD interaction makes a major contribution to the coupling between Ca²⁺ binding and pore opening that is independent of voltage-dependent activation, and that the CTD–VSD interface is an important site for allosteric agonists to enhance Ca²⁺-dependent activation of BK channels.

Results

In order to probe if the VSD–CTD interaction of BK channels is required for Ca²⁺-dependent activation, we initiated an in silico search for compounds that may interfere with BK channel Ca²⁺-dependent activation. Specifically, we docked a subset of the Available Chemical Database (ACD, Molecular Design Ltd.) of about 4 × 10⁴ compounds, in which each compound carries one or two charged groups with a net charge of 0, ±1, or ±2, to a site near the VSD-CTD interface (Fig. 1a, Supplementary Fig. 1). Patch-clamp recordings of BK channels expressed in Xenopus oocytes with and without the candidate compounds from the in silico screening showed that one of the compounds, BC5 applied at the cytosolic side of the membrane patch (Fig. 1b), activated the channel as measured from the inward tail currents at a voltage (−80 mV) more negative than the equilibrium potential of the K⁺ (0 mV) (Fig. 1c, d and Supplementary Fig. 2a). BC5 shifted the voltage-dependent activation (GV relation) of BK channels to more negative voltages (Fig. 1d), with the concentration at the half-maximum GV shift (EC₅₀) of 2.5 ± 0.3 µM (Fig. 1e). In addition, BC5, applied in the cytosolic side of the membrane patch, also inhibited the outward BK currents at positive voltages (Fig. 1c and Supplementary Fig. 2b).

We found that the activation and inhibition of BK channels by BC5 were independent molecular processes. BC5 inhibited the channel at a lower concentration than activation, with the half-maximum inhibition concentration (IC₅₀) of 1.1 ± 0.1 µM (Fig. 1e, Supplementary Fig. 2b). The inhibition was dependent on voltage, with weaker inhibition at less positive potentials (Fig. 1f) and no inhibition at a negative potential (Fig. 1c). These characteristics of inhibition are consistent with the mechanism that the positively charged BC5 acts as a pore blocker from the cytosolic side, and can be flushed out by negative voltages and inward K⁺ currents, reminiscent of Ca²⁺ and Mg²⁺ block of the channel^39,40. Single channel recordings showed that BC5 caused brief closures during channel opening without altering the single channel conductance (Fig. 1g, h), consistent with the mechanism that BC5 acts as a discrete pore blocker with relatively fast on- and off-rates, but not fast enough to be a fast blocker that reduces single channel conductance. In addition, unlike BC5 activation, BC5 inhibition of the channel showed little dependence on Ca²⁺ (Fig. 1f) or Mg²⁺ (Supplementary Fig. 3a, b), and the mutations that reduced BC5 activation had little effect on BC5 inhibition (Supplementary Fig. 4a, b). These results suggest that BC5 inhibition of the channel is an off-target effect, namely, in addition to interacting with our target docking site for channel activation (Fig. 1a), BC5 may interact with the pore (Supplementary Fig. 4c) to block the outward K⁺ currents.

Our target docking site of BC5 in the BK channel is located near the Mg²⁺-binding site for channel activation, which is formed by residues from both the VSD and CTD^36,41 (Fig. 2a). As shown in Fig. 2a, d, residue E399 is shared by both the Mg²⁺- and BC5-binding sites. If BC5 interacts with the target docking site to activate the channel, we expected that an electrostatic repulsion between Mg²⁺ and the charges in BC5 and a direct competition between Mg²⁺ and BC5 for binding would reduce BC5 activation. Consistent with this idea, we found that BC5 at the same concentration (100 µM) induced less shift in GV relations in the presence of 10 mM Mg²⁺ (Fig. 2b, c and Supplementary Fig. 3a). Likewise, a charge reversal mutation of the Mg²⁺-binding residue, D99R¹⁹, also reduced BC5 activation of the channel (Fig. 2c), although D99 is not part of the BC5-binding pocket, supporting that BC5 binds in the vicinity of the Mg²⁺-binding site.

**Fig. 2: BC5 interactions with the targeted site at the CTD-VSD interface.**

To further characterize BC5 interactions with the target docking site, 200 ns molecular dynamics (MD) simulations of the open and closed states of BK channel with and without docked BC5 were performed in explicit membrane and water (see “Methods”) (Supplementary Table 1). The results showed multiple binding modes between BC5 and the channel that interchanged dynamically with time, and in the process BC5 interacted with different residues (Fig. 2d). As summarized in Supplementary Table 2, the hydrophobic naphthalenyl group of BC5 was buried in contact with the intracellular side of S0, S1, S3, and S4 as well as the residue F395 from αB of CTD. The charged arginine part of BC5 interacted with the negatively charged residue E399 (Fig. 2d). The binding poses of BC5 were more narrowly confined in the open conformation, suggesting that BC5 binding is stronger and may favor the open state of BK channels. We mutated the BC5 interacting residues individually or in combination to alanine or other amino acids that changed the charge of the BC5 interacting residue, and measured dose responses of the mutant channel activation to BC5 (Fig. 2e). These mutant channels showed increased EC₅₀ (Fig. 2f), suggesting that the mutations reduced BC5 binding by interrupting the interactions between those residues and BC5. Since BC5 binds to the channel with multiple modes, a mutation to a single residue in the binding pocket may not be sufficient to abolish BC5 binding. Instead, the mutation may prevent some of the binding modes but reinforce others. This may have happened to some of the mutations, which, while enhancing EC50, also increased the slope of the BC5 dose response and the maximum channel activation (Fig. 2e, f). To further disrupt BC5 binding, we made double, triple or quadruple mutations that combined these individual mutations. However, except for the two shown in Figs. 2e and 2f, the combined mutations abolished channel function. These results suggest that this pocket of residues are important for channel function. Consistently, we found that some of the individual mutations altered channel function (Supplementary Fig. 5). These results support that BC5 enhances BK channel activation by binding to the target docking site.

BK channels are activated by voltage and Ca²⁺ with distinct mechanisms⁷. Thus, we asked whether BC5 shifted the GV relation of BK channels by modulating the voltage or Ca²⁺-dependent activation mechanism. To address this question, we first examined whether BC5 modulated VSD activation in BK channels by measuring gating currents with and without BC5 (Fig. 3a). Unlike the GV relationship, the voltage dependence of gating charge movement (QV) was not shifted to more negative potentials (Fig. 3a). Furthermore, measurements of intrinsic openings of the channel at negative voltages (<−100 mV) where the voltage sensors were at the resting state^14,42 showed that BC5 enhanced open probability (Fig. 3b). These results suggest that BC5 did not activate the channel by altering the voltage-dependent mechanism. Since the BC5- and Mg²⁺-binding sites are located closely and share E399 (Fig. 2a), we examined if BC5 activates the channel via a similar mechanism as Mg²⁺ by modulating VSD activation³⁶. Mg²⁺ activates BK channels by interacting with the voltage sensor when the VSD is at the activated conformation^43,44,45, which results in a reduction of the off-gating current amplitude at the end of the depolarizing voltage pulse⁴⁶. The Mg²⁺-VSD interaction also shifts GV relation to more negative potentials. However, Mg²⁺ cannot alter open probabilities at negative potentials when the VSD is not activated⁴⁶, differing from the BC5 effect (Fig. 3b). Unlike Mg²⁺, BC5 also had no effect on the amplitude of off-gating currents (Fig. 3a). Therefore, BC5 did not alter voltage-dependent activation indirectly by a mechanism similar to that of Mg²⁺. Previous studies show that Ca²⁺ binding also shifts the GV relation to more negative potentials and enhances open probability at negative potentials^13,23. Interestingly, the amount of GV shift and the enhancement of open probability at negative voltages caused by 100 µM BC5 were comparable to those caused by 1 µM Ca²⁺^22,23, suggesting that BC5 activated the channel by specifically modulating the Ca²⁺-dependent mechanism.

**Fig. 3: BC5 perturbs Ca²⁺-dependent activation.**

To further examine the BC5 regulation of the Ca²⁺-dependent activation mechanism, we measured BC5 activation of BK channels at various Ca²⁺ concentrations. We found that Ca²⁺ reduced BC5 activation and the reduction was dependent on Ca²⁺ concentration. At 100 µM Ca²⁺, which saturated Ca²⁺-dependent activation¹¹, BC5 no longer activated the channel (Fig. 3c, d and Supplementary Fig. 6). The effect of Ca²⁺ on BC5 activation is specifically associated with Ca²⁺-dependent activation of the channel, because mutations that abolished Ca²⁺-binding sites in the CTD for channel activation⁴⁷ also abolished Ca²⁺ effects on BC5 activation of the channel (Fig. 3e, Supplementary Fig. 7). We studied a truncated BK channel in which the entire CTD was replaced with a short exogenous peptide, known as the Core-MT⁴⁸. The Core-MT channel was no longer activated by Ca²⁺ because the deletion of all Ca²⁺-binding sites. However, it was still activated by BC5, and 100 µM Ca²⁺ had no effect on BC5 activation (Fig. 3f and Supplementary Fig. 8a–c). These results suggest that Ca²⁺ competed with BC5 for activating BK channels.

Ca²⁺-binding sites within the CTD are not spatially close to the BC5-binding site at the VSD-CTD interface (Fig. 1a), and further, BC5 activated the mutant channels in which Ca²⁺-binding sites or the entire CTD were eliminated (Fig. 3e, f). Therefore, BC5 likely does not interact with Ca²⁺ binding directly, but rather the competition between Ca²⁺ and BC5 is allosteric. In other words, Ca²⁺ binding to the channel may alter the conformation at the BC5-binding site. For BC5 to open the activation gate independently of voltage on one hand and compete with Ca²⁺ binding on the other hand, it may interact with the molecular pathway that couples Ca²⁺ binding to channel opening. That is, BC5 perturbs the pathway to promote pore opening, which can be overwhelmed by a stronger perturbation at high Ca²⁺ concentrations due to Ca²⁺ binding. To further explore this idea, we performed dynamic network (Fig. 4) and information flow analysis (Supplementary Fig. 9) to examine the coupling pathway from the Ca²⁺-binding residue D367 to the pore lining F315. Specifically, the network was constructed in which each node represented a single residue (or BC5) and the lengths of edges captured the dynamic correlation calculated from the MD trajectories (see “Methods”). Nodes that were more strongly coupled would have shorter connecting edges. We then calculated a set of “optimal” and “suboptimal” paths between the selected residues, which were the shortest paths and thus the most strongly coupled⁴⁹. The top 20 paths with the strongest coupling strengths are shown in Fig. 4a, b (green sticks) to illustrate the allosteric coupling pathways connecting the Ca²⁺-binding site (D367) to the pore residues (F315) in the same subunit and in a neighboring subunit. The Ca²⁺-binding site is in the CTD, while the activation gate is located in the membrane spanning domain. In both open and closed conformations, the allosteric pathway crossed these domains via two major branches, with one branch via the non-covalent VSD-CTD interface from αA/αB in the CTD to the S4-S5 linker/S6 in the VSD, and the other branch via the covalent C-linker to S6. To complement the pathway analysis, we analyzed the information flow^50,51 between D367 (“source”) and F315 (“sink”), which quantifies the contributions of all the residues to dynamic coupling. The results show that the information flow through VSD-CTD interface is higher or similar to that through the C-linker/S6 connection (Supplementary Fig. 9), suggesting that this interface is critical for the coupling between Ca²⁺ binding and pore opening in addition to the C-linker/S6 pathway²⁹. Importantly, the network analysis shows that BC5 lies adjacent to the key allosteric pathways at the VSD-CTD interface, interacting with residues F395 and E399, which are part of both the BC5-binding site and the allosteric pathway in the closed structure bound with BC5 (Fig. 4a). In addition, in other closed and open structures F395 and E399 are in the vicinity of the allosteric pathway within a distance of < 6 Å (Fig. 4a, b and Supplementary Table 2). These results suggest that BC5 binding likely perturbs the allosteric pathway to interact with both Ca²⁺ binding and pore opening.

**Fig. 4: BC5 interactions with the allosteric pathway for coupling Ca²⁺ binding to pore opening.**

We performed a mutational scanning of selected residues on or in the vicinity of the allosteric pathway and found that the mutation T245W reduced whereas E219R enhanced the sensitivity of the channel to Ca²⁺ for activation¹⁸ (Fig. 4c, d). T245 is part of the allosteric pathway, while E219 is in the vicinity of the allosteric pathway in the open state with or without BC5 (Fig. 4b). E219 and T245 are in S4 and S5, respectively, and are downstream from BC5 in the allosteric pathway, being closer to the activation gate (Fig. 4a, b). If BC5 perturbs the allosteric pathway to activate the channel as shown above (Fig. 4a, b; Supplementary Fig. 9) we expected that these two mutations would alter BC5 activation as well. Indeed, similar to the effects on Ca²⁺-dependent activation, T245W reduced BC5 activation of the channel such that BC5 shifted the GV less in the mutant than in the WT channel (Fig. 4c, d), while E219R enhanced BC5 activation (Fig. 4c, d). These results support the notion that the allosteric pathway couples Ca²⁺ binding to pore opening (Fig. 4a, b) and that BC5 activates the channel by perturbing the pathway. Taken together, our results demonstrate the importance of the molecular interactions at the VSD-CTD interface of BK channels in coupling Ca²⁺ binding to pore opening. One of the two major allosteric pathways mediating Ca²⁺ binding in the CTD to open the channel gate in the membrane is via the VSD-CTD interface. It is not clear where the allosteric path eventually merges with that for voltage-dependent activation. However, since both Ca²⁺ and BC5 can activate the channel independently of voltage^13,23 (Fig. 3a, b) the allosteric path for Ca²⁺-dependent activation starting from Ca²⁺-binding sites remains distinct from voltage-dependent activation at least down to the BC5 site.

Discussion

We identified BC5 using in silico screening in order to modify Ca²⁺-dependent activation of BK channels (Fig. 1a, b). Our results showed that BC5 activated BK channels by selectively altering the Ca²⁺-dependent activation mechanism in the absence of voltage sensor activation and without indirectly modifying voltage-dependent gating (Fig. 3a, b). Molecular docking, atomistic simulations and mutagenesis studies together suggest that BC5 binds at the CTD-VSD interface (Fig. 2), away from Ca²⁺-binding sites in the CTD and the pore in the membrane. Yet it not only promotes pore opening (Fig. 1d) but also interacts with Ca²⁺ binding such that its effect on channel activation is inhibited by Ca²⁺ binding (Fig. 3c–f). Thus, BC5 perturbs both ends of the Ca²⁺-dependent activation mechanism, the sensor and the pore, allosterically by interacting with the pathway that couples Ca²⁺ binding to pore opening. This pathway is built intrinsically in the BK channel protein, since BC5 still activates the Core-MT channel (Fig. 3f), in which the entire CTD is removed and replaced with an exogenous short peptide, indicating that the pathway downstream from the BC5-binding site is preserved despite the removal of Ca²⁺ sensors.

We found that BC5 activated BK channels by shifting the GV relation (Fig. 2f) and increasing open probability at negative voltages (Fig. 3b) to a similar extent as 1 µM Ca²⁺^47,52. When the dependence of BK channel activation on Ca²⁺ was fit with the MWC model, the Ca²⁺-binding affinity at the open state was ~1 µM^47,52. Thus, BC5 activation of the channel only reached a part of the full capacity of Ca²⁺-dependent activation (Fig. 4d). This is not considered surprising because Ca²⁺ binding is coupled to pore opening via two major pathways, one is through the CTD-VSD interface, which is affected by BC5, and the other is through the C-linker^{19,20,21,27,29,30,36,38} (Fig. 4a, b). However, in our dynamic network and information flow analyses, the highest information flow residues were observed at the VSD-CTD interface (Fig. 4a, b and Supplementary Fig. 9). These results suggest that the pathway through the VSD-CTD interface may make more contributions to Ca²⁺-dependent activation than the C-linker. However, since the maximal activation by BC5 was less than half of that by Ca²⁺ (Fig. 4d), BC5 interaction with the pathway may not be strong enough to even achieve the full capacity of Ca²⁺-dependent activation via the CTD-VSD pathway. This idea is consistent with the result that Ca²⁺ binding inhibited BC5 effects on channel activation (Fig. 3c, d), suggesting that the perturbation of the pathway by Ca²⁺ binding was stronger and overrode that by BC5.

Each Slo1 subunit of BK channels contains two Ca²⁺-binding sites in the CTD^41,47,53,54, and we found that Ca²⁺ binding to both sites shared the same CTD-VSD allosteric pathway for coupling to the pore. We measured Ca²⁺ effects on BC5 activation of the channel with either Ca²⁺-binding site removed by mutation 5D5N or D367N⁴⁷. BC5 activated the channel similarly to that in the WT BK in either of these mutant channels in the absence of Ca²⁺. In 100 µM Ca²⁺ BC5 activation was not abolished as in the WT channel (Fig. 3c, d and Supplementary Fig. 10), indicating that the disrupted Ca²⁺ binding to either of the removed site by 5D5N or D367N could no longer inhibit BC5 effects via the CTD–VSD pathway. In other words, Ca²⁺ binding to both sites is required to inhibit BC5 activation effects. These results indicate that both Ca²⁺ binding sites are coupled to the activation gate via the same allosteric pathway that interacts with BC5. Since BC5 activated the 5D5N or D367N channels to a similar extent by a similar GV shift (Supplementary Fig. 10), the two Ca²⁺-binding sites may activate the channel with a roughly equal amount via the allosteric pathway that interacts with BC5. Our dynamic pathway analysis shows that the part of the pathway downstream from the BC5 site on the VSD–CTD interface to the activation gate is common to both Ca²⁺-binding sites, but the part of the pathway upstream from the BC5 site to the two Ca²⁺-binding sites diverges to connect individual Ca²⁺ sites (Supplementary Fig. 11).

BC5 activated BK channels as an allosteric agonist, but it also inhibited the channel as an off-target effect (Fig. 1c–e). Opposite to its activation of the channel, which was dependent on Ca²⁺ (Fig. 3c, d) but not on voltage (Fig. 3a, b), the inhibition by BC5 was dependent on voltage but not Ca²⁺ (Fig. 1f). While BC5 can be used as an excellent tool to study Ca²⁺-dependent activation mechanisms, it can also be used as an effective BK channel inhibitor in physiological conditions, where the membrane voltage remains more positive than the K⁺ equilibrium potential so that BK channels are always inhibited. Our results revealed the CTD-VSD interface as an excellent site for allosteric agonists (Fig. 2) and the inner pore as a site for inhibitors (Supplementary Fig. 4) of BK channels. It should be noted that our current study suggests that BC5 adopts multiple binding modes in the CTD-VSD interface and most contacts are transient. Thus, further chemical modification of BC5 may yield new compounds with more stable interactions with BK channels, which may possess stronger allosteric activation activity. These results also suggest that potentially other sites along the allosteric pathways (Fig. 4a, b) can be used for allosteric agonists or antagonists to modulate BK channel function. These sites can be targeted for the development of therapeutic molecules to treat BK channel associated pathologies including drug therapies for BK channel associated diseases such as neurological disorders associated with either gain-of-function or loss-of-function BK mutations⁷ or impaired urinary bladder control¹¹.

Methods

Molecular docking and in silico screening

A structure-based in silico screening strategy was employed in this study. This strategy was previously applied to the KCNQ1 channel, resulting in the discovery of the novel channel modulators^55,56. Specifically, the in silico screening method used our molecular docking software MDock^57,58,59,60 as the docking engine, and each small molecule in a chemical library was docked to a selected site on a target protein. The protein structure was treated as a rigid body during docking and other parameters in the docking program were set as default. Then, compounds in the library were ranked by their binding scores evaluated by ITScore2 (implemented in MDock)⁵⁹, which is a knowledge-based atomic pairwise scoring function derived by using a statistical mechanics-based iterative method based on a training set of experimentally determined protein–ligand complex structures (a refined set of PDBbind 2012 containing 2897 protein–ligand complexes^61,62).

As for the protein structure for docking, both open (Ca²⁺-bound, PDB: 6v38) and closed (Ca²⁺-free, PDB: 6v3g) conformations of human BK channels were solved using Cryo-EM²¹. A major difference between these two conformations is the relative position between the membrane spanning region (S0-S6) and the cytosolic domain (CTD), as shown in Supplementary Fig. 1a. Because we aimed to discover chemical compounds stabilizing BK channels during their open state, the interface between the voltage sensor domain (S1-S4) and CTD was selected for the target site and the structure in the open state (PDB: 6v38) was used in the in silico screening study (see Fig. 1a and Supplementary Fig. 1b). Because the residues interacting with Mg²⁺ are located at the VSD-CTD interface and were included in our target site, the Mg²⁺ ion in the channel structure was removed before docking. The protein structure was prepared using UCSF Chimera^63,64. Moreover, the experimental hit, BC5, was docked to both open and closed conformations of BK channels. The resulting binding modes were used as starting points for atomistic simulations.

Next, to prepare for the chemical library in this proof-of-concept study, considering that the VSD-CTD interface contains many charged residues (e.g., K98, D99, D173, and E219 on VSD, and R342, E374, K392, and E399 on CTD), we screened a subset of the Available Chemical Database (ACD version 1999, Molecular Design Ltd., currently is a part of BIOVIA in Dassault Systèmes, https://www.3ds.com/) in which each compound contains either one charged group (carrying a charge of +1, −1, +2 or −2) or two charged groups (with one group carrying +1 charge and one group carrying −1 charge). There were a total of ~4 × 10⁴ compounds in this chemical sub-database for screening. Because accurately predicting protein–ligand binding energies remains a challenging task for scoring functions in molecular docking programs^64,65, current in silico screening methods are in pursuit of an enriched subset of potential candidates^66,67. Therefore, we further visually inspected the docking modes of the top 2% compounds ranked by the scoring function, ITScore2⁵⁹. Compounds forming strong interactions (e.g., hydrogen bonds or salt bridges) with both VSD and CTD in their docking modes were selected. One of the selected compounds, BC5, forms hydrogen bonds with N172 and D173 from VSD and S337 and E399 from CTD, respectively. At the end, 9 purchasable compounds (reported in Supplementary Table 3) were selected for experimental assays.

Mutations and expression

All mutations in this study were made by using overlap-extension polymerase chain reaction (PCR) with Pfu polymerase (Stratagene) on the template of the mbr5 splice variant of mslo1 (Uniprot ID: Q08460)⁶⁸. Sequencing of the PCR-amplified regions is used to confirm the mutation⁴¹. mRNA was synthesized in vitro from linearized cDNA using T3 polymerase kits (Ambion, Austin, TX). mRNA of mutations was injected into oocytes (stage IV–V) from female Xenopus laevis at an amount of 0.05–50 ng/oocyte. The injected oocytes were incubated at 18 °C for 2–7 days before electrophysiology recordings. We have complied with all relevant ethical regulations and animal protocol is approved by IACUC at Washington University in St. Louis, USA.

Electrophysiology

All experimental data were collected from inside-out patches. The set-up included an Axopatch 200-B patch-clamp amplifier (Molecular Devices, Sunnyvale, CA), ITC-18 interface and Pulse acquisition software (HEKA Electronik, Lambrecht, Germany). Borosilicate pipettes used for inside-out patches were pulled using a Sutter P-1000 (Sutter Instrument, Novato, CA) and then wax coated and fire polished with a resistance of 0.5–1.5 MΩ.

The macroscopic currents were collected at the 50 kHz sampling rate (20-μs intervals) with low-pass-filtered at 10 kHz. A P/4 leak subtraction protocol with a holding potential of –120 mV was applied to remove capacitive transients and leak currents (except for experiments in Fig. 3b). The solutions used in recording macroscopic ionic currents include: (1) Pipette solution (in mM): 140 potassium methanesulfonic acid, 20 HEPES, 2 KCl, 2 MgCl₂, pH 7.2. (2) The nominal 0 µM [Ca²⁺]_i solution contained about 0.5 nM free [Ca²⁺]_i (in mM): 140 Potassium methanesulfonate, 20 HEPES, 2 KCl, 5 EGTA, pH 7.1. (3) Basal bath (intracellular) solution to make different [Ca²⁺]_i (in mM): 140 Potassium methanesulfonate, 20 HEPES, 2 KCl, 1 EGTA, and 22 mg/L 18C6TA, pH 7.2. CaCl₂ was then added into the basal solution to obtain the desired free [Ca²⁺]_i that was measured by a Ca²⁺-sensitive electrode (Thermo Electron, Beverly, MA).

Gating currents (Fig. 3a) were recorded from inside-out patches at the 200 kHz sampling rate and 20 kHz filtering with the same leak subtraction protocol. (4) Pipette solution for gating currents (in mM): 125 tetraethylammonium (TEA) methanesulfonic, and 2 TEA chloride, 2 MgCl₂, 20 HEPES, pH 7.2 and (5) bath solution for gating currents (in mM): 135 N-methyl-D-glucamine (NMDG) methanesulfonic, 6 NMDG chloride, 20 HEPES, 5 EGTA, pH 7.2.

Single-channel currents (Fig. 1g, h) were collected using an Axopatch 200B and sampled at 200 kHz with a Digidata 1322A (Molecular Devices). Filtering for analysis was at 5 kHz⁶⁹.

BC5 (ARG-4-methoxy-2-naphthylamine, ordered from Medchemexpress LLC Monmouth Junction, NJ) was dissolved into DMSO to make a 100 mM stock solution and then added to recording solutions to achieve the indicated concentrations. All other chemicals were purchased from Sigma-Aldrich. All the experiments were performed at room temperature (22–24 °C).

Data analysis

Conductance–voltage (GV) relationship was determined by measuring macroscopic tail current amplitudes at –80 or −120 mV. The GV relationship was normalized to the maximum G value (G_Max) and was fitted with the Boltzmann function:

$$G/{G}_{{{{{{\rm{Max}}}}}}}=1/(1+{{{{{\rm{exp}}}}}}(-z{e}_{{{{{{\rm{o}}}}}}}(V-{V}_{1/2})/kT))=1/(1+{{{{{\rm{exp}}}}}}(({V}_{1/2}-V)/b))$$

(1)

where z is the number of equivalent charges that move across the membrane electric field, e_o is the elementary charge, V is membrane potential, V_1/2 is the voltage where G/G_Max reaches 0.5, k is Boltzmann’s constant, T is absolute temperature, and b is the slope factor with units of mV. Each GV relationship presented in figures was the averaged results of 3–9 patches and error bars in the figures indicate the standard error of mean (SEM).

In order to measure limiting slope the open probability (P_O) and the total open probability for all the channels N in a patch (NP_O) at negative voltages (Fig. 3b) were first measured. The currents of all opening events from a patch containing hundreds or thousands of channels during a long pulse (5 or 10 s) at each voltage were integrated, and then divided by the single channel current amplitude and total time to obtain NP_O.

The estimation of the total number of channels expressed in a membrane is from the equation

$$I=N\gamma {P}_{{{{{{\rm{O}}}}}}}(V-{E}_{{{{{{\rm{K}}}}}}})=Ni$$

(2)

where i represents single channel current and i = P_O(V − E_K). N was estimated by measuring the macroscopic current at 100 mV in the presence of 100 μM [Ca²⁺]_i, where the single channel conductance is 273 pS and P_O is about 1.0.

Dose response curve of G–V shifts (∆V_1/2) or current inhibition on BC5 concentration (EC₅₀ in Figs. 1e, 2e, f, IC₅₀ in Fig. 1e, and Supplementary Fig. 4b) was obtained from the fitting to the Hill equation:

$${\theta }=\Delta {V}_{1/2}^{{\max }}/[1+{({{{{{{\rm{EC}}}}}}}_{50}/X)}^{n}]$$

(3)

Where θ is the expected response at dosage X, $\triangle {V}_{1/2}^{{{\max }}}$ is the maximum response for an infinite dosage, EC₅₀ is the dosage at which the response is 50% of maximum, and n is the Hill Coefficient. In all of our final fittings Hill Coefficient was set to 1 since it was close to 1 during the initial fits. The curve was fitted using the “curve_fit” fuction Python Scipy module. The standard errors of EC₅₀ and $\triangle {V}_{1/2}^{{{\max }}}$ were estimated directly from the SEMs of the input data and validated by numerical simulations.

Reproducibility

The number of experiments (patches) n for some figure panels are listed below. Figure 1d, n = 12 for control, n = 3 for 0.01 and 0.03 µM BC5; n = 10 for 0.1 µM BC5; n = 4 for 0.3 and 1 µM BC5; n = 5 for 2 µM BC5; n = 6 for 5 µM BC5; n = 3 for 10 µM BC5; n = 5 for 30 and 100 µM BC5; n = 4 for 300 µM BC5.

Figure 1e, for GV shift, n is the same as Fig. 1d; For current inhibition by BC5, n = 12 for control, n = 4 for 0.01, 0.1 and 0.3 µM BC5; n = 5 for 1 and 2 µM BC5; n = 4 for 5 µM BC5; n = 3 for 10 µM BC5; n = 5 for 30 and 100 µM BC5; n = 4 for 300 µM BC5;

Figure 2e, for WT, n is the same as shown in Fig. 1d; for I105E, n = 4; for D173C, n = 5; for E180A, n = 6; for N182A, n = 8; for E395A, n = 5; for E399R, n = 3; for DM1 (E180AN182A), n = 3; for DM2 (I105EE399R), n = 3.

Atomistic simulations

The cryo-EM structures of hBK channel in the Ca²⁺-bound (PDB: 6v38) or Ca²⁺-free (PDB: 6v3g) states were used in all simulations reported in this work²¹. Several short loops absent only in one structure were rebuilt using the other structure as template using the Swiss-PDB server⁷⁰. For example, A567-E576 and N585-E591 missing in Ca²⁺-bound structure were rebuilt using the Ca²⁺-free structure. Missing N- and C-terminal segments as well as several long loops (e.g., V53-G92, A614-V683, and D834-I870) are presumably dynamic and thus not included in the current simulations. Residues before and after the missing segments are capped with either an acetyl group (for N-terminus) or a N-methyl amide (for C-terminus). Standard protonation states under neutral pH were assigned for all titratable residues.

As summarized in Supplementary Table 1, four sets of atomistic simulations were performed: The Ca²⁺-bound state of BK channel with/without BC5 (sim 1-3 and 4-6, respectively) and the Ca²⁺-free state of BK channel with/without BC5 (sim 7-9 and 10-12, respectively). The initial binding pose of BC5 was identified with docking study. In simulations 1-3 and 6-9, a BC5 molecule is placed in binding pocket in each of the four subunits.

All initial structures were first inserted in model POPC lipid bilayers and then solvated in TIP3P water using the CHARMM-GUI web server⁷¹. The systems were neutralized and 150 mM KCl added. The final simulation boxes contain about ~900 lipid molecules and ~130,000 water molecules and other solutes, with a total of ~680,000 atoms and dimensions of ~185 × 185 × 160 Å³. The CHARMM36m all-atom force field⁷² and the CHARMM36 lipid force field⁷³ were used. The interaction parameters of BC5 were then assigned by identifying similar atom, bond, angle, and dihedral-angle types from similar small molecules within the CHARMM36m all-atom force field and CHARMM General Force Field (Charmn36cgenff)⁷² (see Supplementary Fig. 12). All simulations were performed using CUDA-enabled versions of Gromacs 2018^74,75. The default CHARMM-GUI generated configuration was used except otherwise noted here. Electrostatic interactions were described by using the Particle Mesh Ewald (PME) algorithm⁷⁶ with a cutoff of 12 Å. Van der Waals interactions were cutoff at 12 Å with a smooth switching function starting at 10 Å. Covalent bonds to hydrogen atoms were constrained by the SHAKE algorithm⁷⁷, and the MD time step was set at 2 fs. The temperature was maintained at 298 K using the Nose-Hoover thermostat^78,79. The pressure was maintained semi-isotopically at 1 bar at membrane lateral directions using the Parrinello–Rahman barostat algorithm⁸⁰.

All systems were first minimized for 5000 steps using the steepest descent algorithm, followed by a series of equilibration steps where the positions of heavy atoms of the protein, BC5 and lipid were harmonically restrained per CHARMM-GUI Membrane Builder protocol⁷¹. Briefly, 6 equilibration step (25 ps for steps 1–3, 100 ps for steps 4–5 and 10 ns for step 6) were performed, where the restrained force constant for proteins and BC5 molecules were set to 10, 5, 2.5, 1.0, 0.5 and 0.1 kcal mol^–1 Å⁻², respectively. For lipids, the phosphorus is restrained with force constants of 2.5, 2.5, 1.0 and 0.5, 0.1 and 0.0 kcal mol⁻¹ Å⁻², respectively. In the last equilibration step, only protein heavy atoms were harmonically restrained and the system was equilibrated for 10 ns under NPT (constant particle number, pressure and temperature) conditions at 298 K and 1 bar. All production simulations were performed under NPT conditions at 298 K and 1 bar.

Analysis

Unless stated otherwise, snapshots were extracted every 50 ps for all equilibrium MD trajectories for calculation of statistical distributions. Molecular illustrations were prepared using VMD⁸¹. Contact probability was calculated with MDAnalysis⁸². Specifically, a residue is considered to have contact with BC5 if any of its non-hydrogen atoms is found within 6 Å of any heavy atoms of BC5. The results were first averaged over the four BC5 in four BK subunits, then averaged over three simulations in each condition.

Dynamic network analysis

The analysis identifies the most probable pathways of dynamic coupling between two selected residues using the Floyd-War-shall algorithm⁴⁹. For this, each residue (and BC5) is represented as a node of the network. If two residues have a contact (identified with a minimal non-hydrogen atom distance cutoff of 5 Å) for greater than 75% of the simulation time, an edge connecting the two corresponding nodes is added to the network. The resulting contact matrix was weighted based on the covariance of dynamic fluctuation (C_ij) calculated from the same MD trajectory as w_ij = −log(|C_ij|). The length of a possible pathway D_ij between distant nodes i and j is defined as the sum of the edge weights between consecutive nodes along this path. The optimal pathway is identified as the shortest path, thus having the strongest dynamic coupling. Suboptimal paths between i and j are identified as additional top paths ranked using the path length. The analysis was performed using the Network View⁸³. The first 20 optimal and suboptimal paths were selected for analysis presented in this work.

Information flow

Information flow provides a global assessment of the contributions of all nodes to the dynamic coupling between selected “source” and “sink” nodes^50,51. This analysis complements the dynamic pathway analysis to provide additional insights on how different residues may contribute to sensor–pore coupling. For this, we first construct a network similar to the one described above. Pairwise mutual information was calculated between node i and j as follows: M_ij = H_i + H_j − H_ij. H_i is calculated as $\frac{1}{N}\,{\sum }_{n=1}^{N}[-{{{{{\rm{ln}}}}}}{\rho }_{i}(x)]$, where ρ_i(x) is the fluctuation density and x is the distance to the equilibrium position⁵¹. Gaussian mixture model (GMM)⁸⁴ is used to estimate the density. The residue network is then defined as A_ij = C_ij M_ij, in which C_ij is the contact map. To analysis the information flow from source (S₀) to sink (S_I) nodes, the network Laplacian, is calculated as L = D − A, where D is diagonal degree matrix: ${D}_{ii}={\sum }_{j}{A}_{ij}$. The information flow through a given node (residue) is defined as ${f}_{i}=\,\frac{1}{2}({\sum }_{j}|({P}_{i}-{P}_{j})|{A}_{ij})$. The potentials P is given by $P={\widetilde{L}}^{-1}{b}$, where ${\widetilde{L}}^{-1}$ is inverse reduced Laplacian and b is b is the supply vector which corresponds to one unit of current entering at the source node that will exit at sink nodes. The magnitude of f_i thus quantifies the contribution of residue i to dynamic coupling between the source and sink nodes.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

The data that support this study are available from the corresponding authors upon reasonable request. Previously published accession codes used in this study: 6V38 and 6V3G. Source data underlying Figs. 1–4 and Supplementary Figs. 3–5, 9, and 10 are available as a Source data file. Source data are provided with this paper.

Code availability

Custom code used in this study is publicly available at GitHub (https://github.com/zhiguangjia/Ananlysis-code-for-BC5-Nature-communication-).

References

Adams, P. R., Constanti, A., Brown, D. A. & Clark, R. B. Intracellular Ca²⁺ activates a fast voltage-sensitive K⁺ current in vertebrate sympathetic neurones. Nature 296, 746–749 (1982).
Article ADS CAS PubMed Google Scholar
Lancaster, B. & Nicoll, R. A. Properties of two calcium-activated hyperpolarizations in rat hippocampal neurones. J. Physiol. 389, 187–203 (1987).
Article CAS PubMed PubMed Central Google Scholar
Robitaille, R., Garcia, M. L., Kaczorowski, G. J. & Charlton, M. P. Functional colocalization of calcium and calcium-gated potassium channels in control of transmitter release. Neuron 11, 645–655 (1993).
Article CAS PubMed Google Scholar
Davies, A. G. et al. A central role of the BK potassium channel in behavioral responses to ethanol in C. elegans. Cell 115, 655–666 (2003).
Article CAS PubMed Google Scholar
Brayden, J. E. & Nelson, M. T. Regulation of arterial tone by activation of calcium-dependent potassium channels. Science 256, 532–535 (1992).
Article ADS CAS PubMed Google Scholar
Wellman, G. C. & Nelson, M. T. Signaling between SR and plasmalemma in smooth muscle: sparks and the activation of Ca²⁺-sensitive ion channels. Cell Calcium 34, 211–229 (2003).
Article CAS PubMed Google Scholar
Cui, J. BK channel gating mechanisms: progresses toward a better understanding of variants linked neurological diseases. Front. Physiol. 12, 762175 (2021).
Article PubMed PubMed Central Google Scholar
Zhang, G. et al. A gain-of-function mutation in KCNMA1 causes dystonia spells controlled with stimulant therapy. Mov. Disord. 35, 1868–1873 (2020).
Article CAS PubMed PubMed Central Google Scholar
Du, W. et al. Calcium-sensitive potassium channelopathy in human epilepsy and paroxysmal movement disorder. Nat. Genet. 37, 733–738 (2005).
Article CAS PubMed Google Scholar
Brenner, R. et al. Vasoregulation by the beta1 subunit of the calcium-activated potassium channel. Nature 407, 870–876 (2000).
Article ADS CAS PubMed Google Scholar
Petkov, G. V. Central role of the BK channel in urinary bladder smooth muscle physiology and pathophysiology. Am. J. Physiol. Regulat. Integr. Comp. Physiol. 307, R571–R584 (2014).
Article CAS Google Scholar
Contreras, G. F. et al. A BK (Slo1) channel journey from molecule to physiology. Channels 7, 442–458 (2013).
Article CAS PubMed PubMed Central Google Scholar
Horrigan, F. T. & Aldrich, R. W. Coupling between voltage sensor activation, Ca²⁺ binding and channel opening in large conductance (BK) potassium channels. J. Gen. Physiol. 120, 267–305 (2002).
Article CAS PubMed PubMed Central Google Scholar
Horrigan, F. T., Cui, J. & Aldrich, R. W. Allosteric voltage gating of potassium channels I. mSlo ionic currents in the absence of Ca2+. J. Gen. Physiol. 114, 277–304 (1999).
Article CAS PubMed PubMed Central Google Scholar
Rockman, M. E., Vouga, A. G. & Rothberg, B. S. Molecular mechanism of BK channel activation by the smooth muscle relaxant NS11021. J. Gen. Physiol. https://doi.org/10.1085/jgp.201912506 (2020).
Bao, L. & Cox, D. H. Gating and ionic currents reveal how the BKCa channel’s Ca²⁺ sensitivity is enhanced by its beta1 subunit. J. Gen. Physiol. 126, 393–412 (2005).
Article CAS PubMed PubMed Central Google Scholar
Wang, B., Rothberg, B. S. & Brenner, R. Mechanism of increased BK channel activation from a channel mutation that causes epilepsy. J. Gen. Physiol. 133, 283–294 (2009).
Article CAS PubMed PubMed Central Google Scholar
Zhang, G. et al. A charged residue in S4 regulates coupling among the activation gate, voltage, and Ca2+ sensors in BK channels. J. Neurosci. 34, 12280–12288 (2014).
Article PubMed PubMed Central Google Scholar
Tao, X., Hite, R. K. & MacKinnon, R. Cryo-EM structure of the open high-conductance Ca2+-activated K+ channel. Nature 541, 46–51 (2017).
Article ADS CAS PubMed Google Scholar
Hite, R. K., Tao, X. & MacKinnon, R. Structural basis for gating the high-conductance Ca2+-activated K+ channel. Nature 541, 52–57 (2017).
Article ADS CAS PubMed Google Scholar
Tao, X. & MacKinnon, R. Molecular structures of the human Slo1 K(+) channel in complex with beta4. eLife https://doi.org/10.7554/eLife.51409 (2019).
Cui, J., Cox, D. H. & Aldrich, R. W. Intrinsic voltage dependence and Ca²⁺ regulation of mslo large conductance Ca-activated K⁺ channels. J. Gen. Physiol. 109, 647–673 (1997).
Article CAS PubMed PubMed Central Google Scholar
Yang, J. et al. An epilepsy/dyskinesia-associated mutation enhances BK channel activation by potentiating Ca2+ sensing. Neuron 66, 871–883 (2010).
Article CAS PubMed PubMed Central Google Scholar
Zhou, Y., Yang, H., Cui, J. & Lingle, C. J. Threading the biophysics of mammalian Slo1 channels onto structures of an invertebrate Slo1 channel. J. Gen. Physiol. 149, 985–1007 (2017).
Article CAS PubMed PubMed Central Google Scholar
Yang, H., Zhang, G. & Cui, J. BK channels: multiple sensors, one activation gate. Front. Physiol. 6, 29 (2015).
Article PubMed PubMed Central Google Scholar
Jia, Z., Yazdani, M., Zhang, G., Cui, J. & Chen, J. Hydrophobic gating in BK channels. Nat. Commun. 9, 3408 (2018).
Article ADS PubMed PubMed Central Google Scholar
Geng, Y. et al. Coupling of Ca(2+) and voltage activation in BK channels through the alphaB helix/voltage sensor interface. Proc. Natl Acad. Sci. USA 117, 14512–14521 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Zhang, G. et al. Deletion of cytosolic gating ring decreases gate and voltage sensor coupling in BK channels. J. Gen. Physiol. 149, 373–387 (2017).
Article CAS PubMed PubMed Central Google Scholar
Niu, X., Qian, X. & Magleby, K. L. Linker-gating ring complex as passive spring and Ca²⁺-dependent machine for a voltage- and Ca²⁺-activated potassium channel. Neuron 42, 745–756 (2004).
Article CAS PubMed Google Scholar
Lee, U. S. & Cui, J. BK channel activation: structural and functional insights. Trends Neurosci. 33, 415–423 (2010).
Article CAS PubMed PubMed Central Google Scholar
Savalli, N., Pantazis, A., Yusifov, T., Sigg, D. & Olcese, R. The contribution of RCK domains to human BK channel allosteric activation. J. Biol. Chem. 287, 21741–21750 (2012).
Article CAS PubMed PubMed Central Google Scholar
Carrasquel-Ursulaez, W. et al. Hydrophobic interaction between contiguous residues in the S6 transmembrane segment acts as a stimuli integration node in the BK channel. J. Gen. Physiol. 145, 61–74 (2014).
Article Google Scholar
Tian, Y., Heinemann, S. H. & Hoshi, T. Large-conductance Ca2+- and voltage-gated K+ channels form and break interactions with membrane lipids during each gating cycle. Proc. Natl Acad. Sci. USA 116, 8591–8596 (2019).
Article ADS CAS PubMed PubMed Central Google Scholar
Lorenzo-Ceballos, Y., Carrasquel-Ursulaez, W., Castillo, K., Alvarez, O. & Latorre, R. Calcium-driven regulation of voltage-sensing domains in BK channels. eLife 8, e44934 (2019).
Article CAS PubMed PubMed Central Google Scholar
Miranda, P., Holmgren, M. & Giraldez, T. Voltage-dependent dynamics of the BK channel cytosolic gating ring are coupled to the membrane-embedded voltage sensor. eLife https://doi.org/10.7554/eLife.40664 (2018).
Yang, H. et al. Activation of Slo1 BK channels by Mg2+ coordinated between the voltage sensor and RCK1 domains. Nat. Struct. Mol. Biol. 15, 1152–1159 (2008).
Article CAS PubMed PubMed Central Google Scholar
Krishnamoorthy, G., Shi, J., Sept, D. & Cui, J. The NH2 terminus of RCK1 domain regulates Ca²⁺-dependent BK(Ca) channel gating. J. Gen. Physiol. 126, 227–241 (2005).
Article CAS PubMed PubMed Central Google Scholar
Yang, J. et al. Interaction between residues in the Mg2+-binding site regulates BK channel activation. J. Gen. Physiol. 141, 217–228 (2013).
Article ADS CAS PubMed PubMed Central Google Scholar
Cox, D. H., Cui, J. & Aldrich, R. W. Separation of gating properties from permeation and block in mslo large conductance Ca-activated K⁺ channels. J. Gen. Physiol. 109, 633–646 (1997).
Article CAS PubMed PubMed Central Google Scholar
Morales, E., Cole, W. C., Remillard, C. V. & Leblane, N. Block of large conductance Ca2+-activated K+ channels in rabbit vascular myocytes by internal Mg2+ and Na+. J. Physiol. 495, 701–716 (1996).
Article CAS PubMed PubMed Central Google Scholar
Shi, J. et al. Mechanism of magnesium activation of calcium-activated potassium channels. Nature 418, 876–880 (2002).
Article ADS CAS PubMed Google Scholar
Shi, J. & Cui, J. Intracellular Mg²⁺ enhances the function of BK-type Ca²⁺-activated K⁺ channels. J. Gen. Physiol. 118, 589–606 (2001).
Article ADS CAS PubMed PubMed Central Google Scholar
Chen, R.-S., Geng, Y. & Magleby, K. L. Mg2+ binding to open and closed states can activate BK channels provided that the voltage sensors are elevated. J. Gen. Physiol. 138, 593–607 (2011).
Article CAS PubMed PubMed Central Google Scholar
Horrigan, F. T. & Ma, Z. Mg²⁺ enhances voltage sensor/gate coupling in BK channels. J. Gen. Physiol. 131, 13–32 (2008).
Article CAS PubMed PubMed Central Google Scholar
Hu, L. et al. Participation of the S4 voltage sensor in the Mg²⁺-dependent activation of large conductance (BK) K⁺ channels. Proc. Natl Acad. Sci. USA 100, 10488–10493 (2003).
Article ADS CAS PubMed PubMed Central Google Scholar
Yang, H. et al. Mg²⁺ mediates interaction between the voltage sensor and cytosolic domain to activate BK channels. Proc. Natl Acad. Sci. USA 104, 18270–18275 (2007).
Article ADS CAS PubMed PubMed Central Google Scholar
Zhang, G. et al. Ion sensing in the RCK1 domain of BK channels. Proc. Natl Acad. Sci. USA 107, 18700–18705 (2010).
Article ADS CAS PubMed PubMed Central Google Scholar
Budelli, G., Geng, Y. Y., Butler, A., Magleby, K. L. & Salkoff, L. Properties of Slo1 K+ channels with and without the gating ring. Proc. Natl Acad. Sci. USA 110, 16657–16662 (2013).
Article ADS CAS PubMed PubMed Central Google Scholar
Floyd, R. W. Algorithm 97: shortest path. Commun. ACM 5, 345 (1962).
Article Google Scholar
Kang, P. W. et al. Calmodulin acts as a state-dependent switch to control a cardiac potassium channel opening. Sci. Adv. 6, eabd6798 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Westerlund, A. M., Fleetwood, O., Pérez-Conesa, S. & Delemotte, L. Network analysis reveals how lipids and other cofactors influence membrane protein allostery. J. Chem. Phys. 153, 141103 (2020).
Article ADS CAS PubMed Google Scholar
Cox, D. H., Cui, J. & Aldrich, R. W. Allosteric gating of a large conductance Ca-activated K⁺ channel. J. Gen. Physiol. 110, 257–281 (1997).
Article CAS PubMed PubMed Central Google Scholar
Xia, X.-M., Zeng, X. & Lingle, C. J. Multiple regulatory sites in large-conductance calcium-activated potassium channels. Nature 418, 880–884 (2002).
Article ADS CAS PubMed Google Scholar
Schreiber, M. & Salkoff, L. A novel calcium-sensing domain in the BK channel. Biophys. J. 73, 1355–1363 (1997).
Article CAS PubMed PubMed Central Google Scholar
Lin, Y. et al. Modulating the voltage sensor of a cardiac potassium channel shows antiarrhythmic effects. Proc. Natl Acad. Sci. USA 118, e2024215118 (2021).
Article CAS PubMed PubMed Central Google Scholar
Liu, Y. et al. A PIP2 substitute mediates voltage sensor-pore coupling in KCNQ activation. Commun. Biol. 3, 385 (2020).
Article CAS PubMed PubMed Central Google Scholar
Huang, S. Y. & Zou, X. An iterative knowledge-based scoring function to predict protein-ligand interactions: II. Validation of the scoring function. J. Comput. Chem. 27, 1876–1882 (2006).
Article CAS PubMed Google Scholar
Xu, X., Ma, Z., Duan, R. & Zou, X. Predicting protein–ligand binding modes for CELPP and GC3: workflows and insight. J. Computer Aided Mol. Des. 33, 367–374 (2019).
Article ADS Google Scholar
Yan, C., Grinter, S. Z., Merideth, B. R., Ma, Z. & Zou, X. Iterative knowledge-based scoring functions derived from rigid and flexible decoy structures: evaluation with the 2013 and 2014 CSAR benchmarks. J. Chem. Inf. modeling 56, 1013–1021 (2016).
Article CAS Google Scholar
Yan, C. & Zou, X. In Methods in Pharmacology and Toxicology (ed. Zhang, W.) 153–166 (Springer Press, 2016).
Wang, R., Fang, X., Lu, Y., Yang, C.-Y. & Wang, S. The PDBbind Database: methodologies and updates. J. Med. Chem. 48, 4111–4119 (2005).
Article CAS PubMed Google Scholar
Cheng, T., Li, X., Li, Y., Liu, Z. & Wang, R. Comparative assessment of scoring functions on a diverse test set. J. Chem. Inf. Modeling 49, 1079–1093 (2009).
Article CAS Google Scholar
Pettersen, E. F. et al. UCSF Chimera–a visualization system for exploratory research and analysis. J. Comput. Chem. 25, 1605–1612 (2004).
Article CAS PubMed Google Scholar
Huang, S. Y., Grinter, S. Z. & Zou, X. Scoring functions and their evaluation methods for protein-ligand docking: recent advances and future directions. Phys. Chem. Chem. Phys. 12, 12899–12908 (2010).
Article CAS PubMed Google Scholar
Brooijmans, N. & Kuntz, I. D. Molecular recognition and docking algorithms. Annu. Rev. Biophys. Biomol. Struct. 32, 335–373 (2003).
Article CAS PubMed Google Scholar
Grinter, S. Z. & Zou, X. Challenges, applications, and recent advances of protein-ligand docking in structure-based drug design. Molecules 19, 10150–10176 (2014).
Article PubMed PubMed Central Google Scholar
Xu, X., Huang, M. & Zou, X. Docking-based inverse virtual screening: methods, applications, and challenges. Biophys. Rep. 4, 1–16 (2018).
Article PubMed PubMed Central Google Scholar
Butler, A., Tsunoda, S., McCobb, D. P., Wei, A. & Salkoff, L. mSlo, a complex mouse gene encoding “maxi” calcium-activated potassium channels. Science 261, 221–224 (1993).
Article ADS CAS PubMed Google Scholar
Nimigean, C. M. & Magleby, K. L. Functional coupling of the beta₁ subunit to the large conductance Ca²⁺-activated K⁺ channel in the absence of Ca²⁺. Increased Ca²⁺ sensitivity from a Ca²⁺-independent mechanism. J. Gen. Physiol. 115, 719–736 (2000).
Article CAS PubMed PubMed Central Google Scholar
Schwede, T., Kopp, J. R., Guex, N. & Peitsch, M. C. SWISS-MODEL: an automated protein homology-modeling server. Nucleic Acids Res. 31, 3381–3385 (2003).
Article CAS PubMed PubMed Central Google Scholar
Lee, J. et al. CHARMM-GUI input generator for NAMD, GROMACS, AMBER, OpenMM, and CHARMM/OpenMM simulations using the CHARMM36 additive force field. J. Chem. Theory Comput. 12, 405–413 (2016).
Article CAS PubMed Google Scholar
Huang, J. et al. CHARMM36m: an improved force field for folded and intrinsically disordered proteins. Nat. Methods 14, 71–73 (2017).
Article CAS PubMed Google Scholar
Klauda, J. B. et al. Update of the CHARMM all-atom additive force field for lipids: validation on six lipid types. J. Phys. Chem. B 114, 7830–7843 (2010).
Article CAS PubMed PubMed Central Google Scholar
Hess, B., Kutzner, C., van der Spoel, D. & Lindahl, E. GROMACS 4: algorithms for highly efficient, load-balanced, and scalable molecular simulation. J. Chem. Theory Comput. 4, 435–447 (2008).
Article CAS PubMed Google Scholar
Abraham, M. J. et al. GROMACS: High performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX 1-2, 19–25 (2015).
Article ADS Google Scholar
Darden, T., York, D. & Pedersen, L. Particle mesh Ewald: an N⋅log(N) method for Ewald sums in large systems. J. Chem. Phys. 98, 10089–10092 (1993).
Article ADS CAS Google Scholar
Ryckaert, J.-P., Ciccotti, G. & Berendsen, H. J. C. Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes. J. Comput. Phys. 23, 327–341 (1977).
Article ADS CAS Google Scholar
Nosé, S. A unified formulation of the constant temperature molecular dynamics methods. J. Chem. Phys. 81, 511–519 (1984).
Article ADS Google Scholar
Hoover, W. G. Canonical dynamics: equilibrium phase-space distributions. Phys. Rev. A 31, 1695–1697 (1985).
Article ADS CAS Google Scholar
Parrinello, M. & Rahman, A. Polymorphic transitions in single crystals: a new molecular dynamics method. J. Appl. Phys. 52, 7182–7190 (1981).
Article ADS CAS Google Scholar
Humphrey, W., Dalke, A. & Schulten, K. VMD: visual molecular dynamics. J. Mol. Graph 14, 27–38 (1996). 33-38.
Article Google Scholar
Michaud-Agrawal, N., Denning, E. J., Woolf, T. B. & Beckstein, O. MDAnalysis: a toolkit for the analysis of molecular dynamics simulations. J. Comput. Chem. 32, 2319–2327 (2011).
Article CAS PubMed PubMed Central Google Scholar
Eargle, J. & Luthey-Schulten, Z. NetworkView: 3D display and analysis of protein·RNA interaction networks. Bioinformatics 28, 3000–3001 (2012).
Article CAS PubMed PubMed Central Google Scholar
Dempster, A. P., Laird, N. M. & Rubin, D. B. Maximum likelihood from incomplete data via the EM algorithm. J. R. Stat. Soc. Ser. B 39, 1–22 (1977).
MathSciNet MATH Google Scholar

Download references

Acknowledgements

This work was supported by grants R01 HL142301 (J. Cui and J.R.S.), R35 GM136409 (X.Z.), R35 GM144045 (J. Chen), and R01 GM114694 (K.L.M.) from NIH.

Author information

These authors contributed equally: Guohui Zhang, Xianjin Xu, Zhiguang Jia.

Authors and Affiliations

Department of Biomedical Engineering, Center for the Investigation of Membrane Excitability Disorders, Cardiac Bioelectricity and Arrhythmia Center, Washington University, St. Louis, MO, USA
Guohui Zhang, Hongwu Liang, Jingyi Shi, Martina Marras, Carlota Abella, Jonathan R. Silva & Jianmin Cui
Dalton Cardiovascular Research Center, University of Missouri – Columbia, Columbia, MO, USA
Xianjin Xu & Xiaoqin Zou
Department of Physics and Astronomy, University of Missouri – Columbia, Columbia, MO, USA
Xianjin Xu & Xiaoqin Zou
Department of Biochemistry, University of Missouri – Columbia, Columbia, MO, USA
Xianjin Xu & Xiaoqin Zou
Institute for Data Science and Informatics, University of Missouri – Columbia, Columbia, MO, USA
Xianjin Xu & Xiaoqin Zou
Department of Chemistry, University of Massachusetts, Amherst, MA, USA
Zhiguang Jia & Jianhan Chen
Department of Biochemistry and Molecular Biology, University of Massachusetts, Amherst, MA, USA
Zhiguang Jia & Jianhan Chen
Department of Physiology and Biophysics, University of Miami Miller School of Medicine, Miami, FL, USA
Yanyan Geng & Karl L. Magleby

Authors

Guohui Zhang
View author publications
You can also search for this author in PubMed Google Scholar
Xianjin Xu
View author publications
You can also search for this author in PubMed Google Scholar
Zhiguang Jia
View author publications
You can also search for this author in PubMed Google Scholar
Yanyan Geng
View author publications
You can also search for this author in PubMed Google Scholar
Hongwu Liang
View author publications
You can also search for this author in PubMed Google Scholar
Jingyi Shi
View author publications
You can also search for this author in PubMed Google Scholar
Martina Marras
View author publications
You can also search for this author in PubMed Google Scholar
Carlota Abella
View author publications
You can also search for this author in PubMed Google Scholar
Karl L. Magleby
View author publications
You can also search for this author in PubMed Google Scholar
Jonathan R. Silva
View author publications
You can also search for this author in PubMed Google Scholar
Jianhan Chen
View author publications
You can also search for this author in PubMed Google Scholar
Xiaoqin Zou
View author publications
You can also search for this author in PubMed Google Scholar
Jianmin Cui
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

J. Cui, X.Z., and J. Chen conceived and initiated the studies. J. Cui, X.Z., J. Chen, J.R.S., and K.L.M. provided supervision and management. G.Z., H.L., J.S., M.M., C.A., J.R.S., and J. Cui performed electrophysiological experiments and analyzed data. X.X. and X.Z. performed molecular modeling and in silico screening. Z.J. and J. Chen performed molecular dynamic (MD) simulations, dynamic network and information flow analyses. Y.G. and K.L.M. did experiments and analyzed data on single channels. G.Z., Z.J., J. Chen, X.X, X.Z., and J. Cui wrote the manuscript, with input from K.L.M, J.R.S., and C.A.

Corresponding authors

Correspondence to Jonathan R. Silva, Jianhan Chen, Xiaoqin Zou or Jianmin Cui.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks the anonymous reviewers for their contribution to the peer review of this work. Peer reviewer reports are available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Peer Review File

Reporting Summary

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Zhang, G., Xu, X., Jia, Z. et al. An allosteric modulator activates BK channels by perturbing coupling between Ca²⁺ binding and pore opening. Nat Commun 13, 6784 (2022). https://doi.org/10.1038/s41467-022-34359-6

Download citation

Received: 25 May 2022
Accepted: 21 October 2022
Published: 09 November 2022
DOI: https://doi.org/10.1038/s41467-022-34359-6

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.