Mice deficient in TWIK-1 are more susceptible to kainic acid-induced seizures

Ajung Kim; Yeonju Bae; Changdev G Gadhe; Hyun-Gug Jung; Esther Yang; Hyun Kim; Jaekwang Lee; Chanseob Shim; Young Hoon Sung; Junyeol Noh; Eun-Jin Kim; Dawon Kang; Ae Nim Pae; Eun Mi Hwang; Jae-Yong Park

doi:10.1016/j.isci.2024.111587

. 2024 Dec 12;28(1):111587. doi: 10.1016/j.isci.2024.111587

Mice deficient in TWIK-1 are more susceptible to kainic acid-induced seizures

Ajung Kim ^1,¹¹, Yeonju Bae ^1,^2,¹¹, Changdev G Gadhe ^3,¹¹, Hyun-Gug Jung ¹, Esther Yang ⁴, Hyun Kim ⁴, Jaekwang Lee ⁵, Chanseob Shim ⁶, Young Hoon Sung ^7,⁸, Junyeol Noh ², Eun-Jin Kim ⁹, Dawon Kang ⁹, Ae Nim Pae ³, Eun Mi Hwang ^1,^∗, Jae-Yong Park ^2,^10,^12,^∗∗

PMCID: PMC11732521 PMID: 39811670

Summary

TWIK-1 belongs to the two-pore domain K⁺ (K2P) channel family, which plays an essential role in the background K⁺ conductance of cells. Despite the development of exon 2-deleted Twik-1 knockout (KO) mice, the physiological role of TWIK-1 has remained largely unknown. Here, we observed that the exon 2-deleted Twik-1 KO mice expressed an internally deleted TWIK-1 (TWIK-1 ΔEx2) protein, which unexpectedly acts as a functional K⁺ channel. The Twik-1 nKO mice in which exon 1 was targeted using the CRISPR-Cas9 technique provides strong evidence that TWIK-1 mediates K⁺ currents that are responsible for the background passive conductance in astrocytes. Deficiency of TWIK-1-mediated astrocytic passive conductance increased susceptibility to kainic acid-induced seizures. This study paves the way for functional studies on TWIK-1-mediated astrocytic passive conductance. In addition, the exon 1-targeted Twik-1 KO mice would help elucidate the physiological roles of TWIK-1.

Subject areas: Biological sciences, Molecular neuroscience, Natural sciences, Neuroscience, Systems neuroscience

Graphical abstract

Highlights

•
The exon 2-deleted Twik-1 KO mice express an internally deleted TWIK-1 mutant
•
The internally deleted TWIK-1 mutant unexpectedly acts as a functional K⁺ channel
•
The exon 1-targeted Twik-1 KO mice display reduced astrocytic passive conductance
•
Mice deficient in TWIK-1 increased susceptibility to kainic acid-induced seizures

Biological sciences; Molecular neuroscience; Natural sciences; Neuroscience; Systems neuroscience

Introduction

The first member of the two-pore domain K⁺ (K2P) channel family, K2P1 (also known as Twik-1; tandem of pore domains in a weak inward rectifying K⁺ channel, KCNK1) was cloned from a human kidney cDNA library. It is expressed in various tissues, including the heart, kidney, and brain.¹ In the brain, Twik-1 is highly expressed in the neurons of various brain regions including the hippocampus, thalamus, striatum, neocortex, and cerebellum.¹^,² The TWIK-1 channel shows the highest mRNA expression level among the K⁺ channels in astrocytes.³ Considering the wide expression of Twik-1 gene, TWIK-1 could have pivotal roles in various tissues. However, the electrophysiological properties and functional roles of TWIK-1 are poorly understood.

It is speculated that TWIK-1 is a non-functional channel owing to its very small or non-measurable currents in heterologous expression systems.⁴ Two distinct hypotheses attempt to explain the low or non-functional expression of TWIK-1 via sumoylation or constitutive endocytosis.⁵^,⁶ However, these silencing mechanisms could not explain how TWIK-1 functions as an ion channel and its physiological roles. Alternatively, TWIK-1 channel forms heterodimers with other K2P channels⁵^,⁷ to serve as functional heterodimeric channels.⁶^,⁷^,⁸ In cerebellar granule cells, TWIK-1 forms a heterodimer with TASK-3, a member of K2P channel family.⁵ Cell-type specific gene silencing using the specific short hairpin RNA (shRNA) against Twik-1 (TWIK-1 shRNA) showed that the TWIK-1/TASK-3 heterodimeric channel contributes to K⁺ conductance and intrinsic excitability in mouse hippocampal dentate gyrus (DG) granule cells.⁸ Therefore, heterodimerization of TWIK-1 with other K2P channels could be a possible alternative mechanism that explains how TWIK-1 functions in native cells.

Mature astrocytes have an unusually leaky membrane and extremely low membrane resistance. This electrophysiological property, called astrocytic passive conductance, results in a linear current-voltage relationship of K⁺ membrane conductance.⁷ The electrophysiological properties of conventional voltage-gated K⁺ channels and leaky K⁺ channels differ with the astrocytic passive conductance; therefore, the molecular identity of passive conductance remained unknown for a long time. We previously suggested that TWIK-1 can act as a main contributor for astrocytic passive conductance.⁷ TWIK-1 forms a heterodimeric channel with TREK-1 (K2P2, KCNK2), another K2P channel, via a disulfide bond-dependent interaction. Silencing Twik-1 using TWIK-1 shRNA induces a significant reduction in astrocytic passive conductance.⁷ In addition, we reported that spadin, a TREK-1 inhibitor, inhibits TWIK-1/TREK-1 heterodimeric channel and significantly reduces astrocytic passive conductance.⁹ The results from specific silencing of Twik-1 and spadin-mediated pharmacological inhibition suggested that TWIK-1 could be a key component in the molecular mechanism underlying astrocytic passive conductance; however, Twik-1 knockout (KO) mice do not exhibit altered astrocytic passive conductance.¹⁰ Therefore, this remains to be elucidated.

Twik-1 KO mice were generated by deleting exon 2 of the Twik-1 gene.¹¹ This led to a defect in phosphate transport in the proximal tubule and water transport in the medullary collecting duct of the kidney.¹¹^,¹² Interestingly, Twik-1 KO mice also exhibit hyperpolarization of the resting membrane potential (RMP) in pancreatic β cells and astrocytes.¹⁰^,¹² In general, since a deficient K⁺ channel induces depolarization of the RMP, therefore, the unexpected deviation of RMP in Twik-1 KO mice suggests that TWIK-1 may act as an unexpected K⁺ channel, or there are unknown reasons in Twik-1 KO mice. These mice express exon 2-deleted messenger RNA (mRNA) of Twik-1, where exon 1 is directly linked to exon 3 without a frameshift mutation. This results in the expression of an internal region-deleted TWIK-1 (TWIK-1 ΔEx2) protein.¹¹ Our earlier gene silencing studies indicated that TWIK-1 functions as a K⁺ channel in the DG cells and astrocytes of mouse hippocampus⁷^,¹³; therefore, we decided to examine the expression of TWIK-1 ΔEx2 protein and its role in Twik-1 KO mice.

In this study, we elucidated that the exon 2-deleted Twik-1 KO mice express the TWIK-1 ΔEx2 protein, which unexpectedly acts as a functional K⁺ channel. These findings prompted us to generate another strain of Twik-1 KO mice (Twik-1 nKO mice) where exon 1 is targeted using the Clustered Regularly Interspaced Short Palindromic Repeats/CRISPR-associated protein 9 (CRISPR/Cas9) technique. The results provide strong evidence that TWIK-1 mediates the K⁺ currents that are responsible for the background passive conductance in astrocytes, and TWIK-1-mediated astrocytic passive conductance is involved in susceptibility to kainic acid-induced seizures. The Twik-1 nKO mice could be a powerful animal model for investigating the enigmatic roles of TWIK-1.

Results

Twik-1 KO mice produce an internal region-deleted TWIK-1 mutant protein

Twik-1 KO mice, first described in 2005, were generated by deleting exon 2 of the Twik-1 gene.¹¹ Exon 2 of the Twik-1 gene encodes the pore domain, the transmembrane domain (TM)2, and TM3 of the channel. In Twik-1 KO mice, the deletion of exon 2 (396 nucleotides) of the Twik-1 gene was confirmed through Polymerase Chain Reaction (PCR) using their genomic DNA (Figure S1A). In addition, we confirmed the expression of exon 2-deleted Twik-1 (Twik-1 ΔEx2) mRNA using Reverse Transcription (RT)-PCR and sequencing analysis, in Twik-1 KO mice (Figure S1B). Twik-1 ΔEx2 mRNA results from the direct linkage of exon 1 to exon 3 without a frameshift mutation.¹¹ Exon 2 of the Twik-1 gene corresponds to the amino acid residues, 119−250 of the TWIK-1 channel protein. Therefore, it is possible that the Twik-1 ΔEx2 mRNA does not produce a premature termination codon; this results in exon 2 being skipped while the reading frame is retained, potentially producing an abnormal protein (an internally deleted protein) rather than undergoing nonsense-mediated decay. However, it is not confirmed whether Twik-1 ΔEx2 mRNA can be used for protein expression in Twik-1 KO mice.

We verified that the expression of the C-terminus of the TWIK-1 channel, which is encoded by exon 3, is not deleted in Twik-1 KO mice. We detected a TWIK-1 protein band in Twik-1 KO mice with a size approximately 14.87 kDa smaller than that in the wild type (WT) mice through western blotting (Figure S1C). Next, we generated a C-terminus-specific antibody against the mouse TWIK-1 channel (Figures S1D and S1E). Immunohistochemistry using the antibody indicated strong immunohistochemical signals against TWIK-1 protein in the various brain regions of Twik-1 KO mice; these include the hippocampus, cerebellum, striatum, and neocortex (Figures 1A and 1B). In addition, we observed immunohistochemical signals against the TWIK-1 protein in the hippocampal neurons and astrocytes of Twik-1 KO mice (Figure 1C). These data strongly suggest that the TWIK-1 ΔEx2 protein in Twik-1 KO mice, like the WT TWIK-1 channel protein, is highly expressed in the various previously reported brain regions and cell types.

*Twik-1* KO mice express internal region-deleted TWIK-1 mutant protein

(A and B) WT and *Twik-1* KO mice were stained using the C-terminus-specific antibody against TWIK-1 channel in whole or in several brain regions. *Twik-1* KO mice expressed TWIK-1 like protein. Scale bars and their size are indicated at the bottom of each image.

(C) Subregions of hippocampus from WT and *Twik-1* KO mice showing expression of MAP2 (red) and TWIK-1 (green) or GFAP (red) and TWIK-1 (green). Expression of TWIK-1-like protein in *Twik-1* KO mice merged with MAP2 and GFAP. Scale bar = 20 μm.

TWIK-1 ΔEx2 protein mediates potassium current in the primary cultured astrocytes from Twik-1 KO mice

Most potassium channels are composed of tetramers because four pore domains are required to make a functional channel. However, in the case of K2P channels, one protein with two pore domains can constitute a functional channel as a dimer.¹⁴ Twik-1 KO mice expressed the protein in a form lacking the second and third transmembrane domains and one pore domain of the TWIK-1 channel, which is encoded by exon 2 (Figure 1). Therefore, it is possible that the TWIK-1 ΔEx2 protein could function as one subunit if it has one remaining pore domain. In particular, the 119^th glycine (G) and 120^th tyrosine (Y) residues, which are part of the GYG motif also known as a potassium selectivity filter (SF), are still present in the TWIK-1 ΔEx2 protein,¹⁵ suggesting that it could mediate K⁺ current. To confirm this possibility, electrophysiological experiments were performed using specific shRNAs against each exon. shRNAs targeting exon 2 (sh-Ex2) and exon 3 (sh-Ex3) of the Twik-1 gene were constructed and validated (Figures 2A and 2B). The expression of TWIK-1 proteins was reduced in both sh-Ex2 and sh-Ex3 treatments in primary cultured astrocytes from WT mice; however, it was only reduced in the sh-Ex3 treatment in the primary cultured astrocytes from Twik-1 KO mice. We measured K⁺ currents in the presence of Twik-1 shRNAs through electrophysiological experiments. Under Twik-1 knockdown conditions, K⁺ currents, both inward and outward, were significantly reduced by sh-Ex2 and sh-Ex3, in the primary cultured astrocytes from WT mice (Figures 2C and 2D). In the primary cultured astrocytes from Twik-1 KO mice, there was no effect on K⁺ currents in the sh-Ex2 treatment; however, sh-Ex3 effectively reduced both inward and outward K⁺ currents (Figures 2E and 2F).

Astrocytic potassium conductance of TWIK-1 ΔEx2 reduced by shRNA targeting exon3 of TWIK-1

(A) Schematic illustration of TWIK-1 shRNAs targeting exon 2 and exon 3.

(B) Validation of TWIK-1 shRNA using western blotting. Sc shRNA or sh-Ex2/sh-Ex3 transfected into primary astrocytes. Actin was used as the loading control.

(C) *I-V* relationship of curves showed the average trace from the Sc shRNA or sh-Ex2/sh-Ex3 plasmid-transfected primary astrocytes of WT mice.

(D) Summary bar graph of Sc shRNA or sh-Ex2/sh-Ex3 currents plotted at −150 mV and +50 mV in primary astrocytes of WT mice. Number on each bar indicates ‘n’ for each condition.

(E) *I-V* relationship of curves showed the average trace from the Sc shRNA or sh-Ex2/sh-Ex3 plasmid-transfected primary astrocytes of *Twik-1* KO mice.

(F) Summary bar graph of Sc shRNA or sh-Ex2/sh-Ex3 currents plotted at −150 mV and +50 mV in primary astrocytes of *Twik-1* KO mice. Number on each bar indicates ‘n’ for each condition.

(G) The location of the stereotaxical injection of AAV-mChe-Sc shRNA or AAV-mChe-TWIK-1 Ex3 shRNA to the striatum radiatum.

(H) Immunohistochemical staining of the hippocampal slice with the anti-TWIK-1 antibody. Scale bar = 25 μm. Enlarged images show expression of TWIK-1 (green), mCherry (red), and GFAP (purple). Scale bar = 10 μm.

(I) The whole-cell currents from WT and *Twik-1* KO hippocampal astrocytes. Command voltage pulse protocol: voltages were stepped from −140 mV to +20 mV in increments of 20 mV from a holding potential of −60 mV. Dotted line indicates 0 pA.

(J) The current–voltage relationship (I–V) from −140 mV to +20 mV.

(K) The bar graph of current at −140 mV and +20 mV.

(L) The bar graph representing input resistance of the respective conditions. All Data are presented as mean ± s.e.m. p-values were obtained with two-way ANOVA followed by Tukey’s post hoc test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

To exclude the possibility of off-target effects by shRNA,¹⁶ we conducted a similar experiment using the CRISPR/Cas9 system to target each exon of the Twik-1 gene. The efficiency of the Twik-1 targeting guide RNA was confirmed through immunocytochemistry in primary cultured astrocytes from both WT and Twik-1 KO mice (Figure S2A). Using electrophysiological experiments, we confirmed that the K⁺ currents were significantly reduced by all CRISPR/Cas9 vectors targeting each exon of the Twik-1 gene in the astrocytes cultured from WT mice (Figures S2B and S2C). However, targeting only exon 2 did not result in any changes in the K⁺ current in the astrocytes cultured from Twik-1 KO mice (Figures S2D and S2E). Therefore, our data strongly suggested that the TWIK-1 ΔEx2 protein can function as a channel for mediating astrocytic K⁺ currents in Twik-1 KO mice.

TWIK-1 ΔEx2 protein mediates astrocytic passive conductance in hippocampal slices of Twik-1 KO mice

We assessed whether the knockdown of TWIK-1 using Adeno-Associated Virus (AAV)-expressing TWIK-1 Ex3 shRNA affected the astrocytic passive conductance of Twik-1 KO mice in vivo. We stereotaxically injected AAVs carrying either scramble (Sc) shRNA or TWIK-1 Ex3 shRNA into the hippocampal stratum radial region of WT and Twik-1 KO mice, respectively (Figures 2G and S2F). After 2 weeks, brains were harvested and stained with antibodies against TWIK-1 and the astrocytic marker Glial fibrillary acidic protein (GFAP). Astrocytic TWIK-1 immunoreactivity was significantly reduced when the mice were injected with AAV-mChe-TWIK-1 Ex3 shRNA compared to that in those injected with AAV-mChe-Sc shRNA, in both WT and Twik-1 KO groups (Figure 2H). In addition, densitometric analysis of the TWIK-1 staining intensity in AAV-mChe-Sc shRNA or AAV-mChe-TWIK-1 Ex3 shRNA treated astrocytes (mChe+ and GFAP+ cells) showed no difference between WT and Twik-1 KO mice (Figure S2G). We confirmed the effect of TWIK-1 Ex3 shRNA on astrocytic passive conductance in the hippocampal slices of WT and Twik-1 KO mice. TWIK-1 Ex3 shRNA significantly reduced astrocytic passive conductance in both WT and Twik-1 KO mice (Figure 2I). The outward currents at +20 mV and inward currents at −140 mV decreased by about 50% in both WT and Twik-1 KO mice (Figures 2J and 2K). Input resistance at −140 mV was significantly increased in AAV-mChe-TWIK-1 Ex3 shRNA-injected mice compared to that in AAV-mChe-Sc shRNA-injected mice, in both WT and Twik-1 KO groups (Figure 2L). Therefore, the TWIK-1 ΔEx2 protein mediates an astrocytic passive conductance, like the full-length TWIK-1 protein, even in vivo.

TWIK-1 ΔEx2 protein can form a disulfide-linked heteromer with TREK-1 protein through its cysteine 69

TWIK-1 ΔEx2 protein can mediate K⁺ currents like full-length TWIK-1 in both primary cultured and adult hippocampal astrocytes; therefore, we hypothesized that TWIK-1 ΔEx2 could function by forming a heteromer with TREK-1 in the astrocytes. This possibility is supported by the presence of the C69 residue, which is a critical residue for the formation of a heterodimer between TWIK-1 with TREK-1,⁷^,¹⁷^,¹⁸ in TWIK-1 ΔEx2. We examined whether TWIK-1 ΔEx2 interacts with TREK-1 by forming a disulfide-bond through the C69 residue in a heterologous system. Initially, we constructed cysteine–serine (CS) mutants into TWIK-1 ΔEx2 and examined the ability of these TWIK-1 ΔEx2 mutants to interact with TREK-1. Both TWIK-1 and TWIK-1 ΔEx2 with the CS mutation had a significantly reduced binding affinity to TREK-1 (Figure 3A). Bimolecular fluorescence complementation (BiFC) experiments confirmed that TWIK-1 ΔEx2, like WT TWIK-1, can closely bind to TREK-1. However, a significant decrease in the BiFC signal was observed in the CS mutant group of TWIK-1 ΔEx2 (Figure 3B).

TWIK-1 ΔEx2 can form a disulfide-linked heteromer with TREK-1 via the C69 residues

(A) Co-immunoprecipitation assay in HEK293T cells. HA-TREK-1 was co-transfected with con-GFP, GFP- TWIK-1, GFP-TWIK-1 C69S, GFP-TWIK-1 ΔEx2, or GFP-TWIK-1 ΔEx2 C69S. Immunoprecipitation was performed with anti-GFP antibody. Input represented 5% of total cell lysates.

(B) BiFC assay. VN and VC are the N-terminal and C-terminal fragments of the Venus protein, respectively. Intense Venus signals (green) indicate dimerization. VN-TWIK-1 and VN-TWIK-1 ΔEx2 strongly interact with TREK-1-VC. Upon mutation of the C69S of TWIK-1 (VN-TWIK-1 C69S, VN-TWIK-1 ΔEx2 C69S), Venus signals were not detected. Nuclei were visualized with DAPI (blue). Scale bar = 25 μm.

(C) Co-immunoprecipitation assay in the primary astrocytes. Immunoprecipitation was performed with anti-TREK-1 antibody. Input represented 5% of total lysates.

(D) Duolink PLA assay. Intense signals were detected in Sc shRNA-transfected astrocytes, but not in the TREK-1 shRNA-transfected astrocytes. Scale bar = 10 μm.

(E) The PLA signals were counted and the average number of spots per cell is presented in the graph. Spots on each bar indicates ‘n’ for each condition.

(F) Duolink PLA assay. Intense signals were detected in the Sc shRNA-transfected astrocytes, but not in the TWIK-1 Ex3 shRNA-transfected astrocytes. Scale bar = 20 μm.

(G) The PLA signals were counted and the average number of spots per cell is presented in the graph. Spots on each bar indicates ‘n’ for each condition. All values are presented as mean ± s.e.m. p-values were obtained with two-way ANOVA followed by Tukey’s post hoc test. ∗∗∗∗p < 0.0001.

We examined the endogenous interaction between TWIK-1 ΔEx2 and TREK-1 in the primary cultured astrocytes from Twik-1 KO mice through co-immunoprecipitation; there was a clear association between TREK-1 and TWIK-1 ΔEx2 (Figure 3C). Duolink Proximity Ligation Assay (PLA) confirmed that TWIK-1 ΔEx2 and TREK-1 bind endogenously in the astrocytes (Figure 3D). Positive PLA signals between TWIK-1 or TWIK-1 ΔEx2 and TREK-1 were observed in primary cultured astrocytes expressing Sc shRNA; however, these signals were significantly reduced in the astrocytes expressing TREK-1 shRNA (Figures 3D and 3E). Consistent with earlier results, the PLA signal identified in primary cultured astrocytes from Twik-1 KO mice was greatly reduced by TWIK-1 Ex3 shRNA and restored by TWIK-1 Ex3 shRNA-treated insensitive form of TWIK-1 ΔEx2 but not by its CS mutation (Figures 3F and 3G).

We examined the characteristics of TREK-1-dependent TWIK-1 expression on the cell surface. When Flag-TWIK-1 or Flag-TWIK-1 ΔEx2 alone was transfected into COS-7 cells, Flag-TWIK-1 or Flag-TWIK-1 ΔEx2 failed to co-localize with WGA488, a cell surface staining-dye (Figures S3A and S3D, upper panel). However, in the presence of HA-TREK-1, co-localizations of TWIK-1 or TWIK-1 ΔEx2 with WGA488 was markedly increased (Figures S3A and S3D, lower panel).

To test whether binding of TWIK-1 ΔEx2 to TREK-1 generates a linear-like K⁺ currents in a heterologous system, we performed electrophysiological experiments in COS-7 cells. Heterologous co-expression of TWIK-1 ΔEx2 and TREK-1 in COS-7 cells indicated the linearity of the I-V relationship (Figures S3E and S3F), like that in the co-expression of TWIK-1 and TREK-1 (Figures S3B and S3C). Therefore, the TWIK-1 ΔEx2 protein generated by deletion of TWIK-1 exon 2 also forms a disulfide-linked heteromer with TREK-1 protein.

Glycine 119 of TWIK-1 ΔEx2 works as a selectivity filter of potassium ion

To test whether the TWIK-1 ΔEx2 protein can structurally form a stable potassium channel with the TREK-1 protein and the glycine 119 can act as a selective filter, we performed homology modeling and molecular dynamics (MD) simulation using known human TWIK-1 and TREK-1 protein structures as templates. We first developed 50 candidate models each, for TWIK-1 and TREK-1; the results from the multiple sequence alignment between the human template protein and the mouse protein was used as the input for the Modeller 9v7 program. We calculated the root-mean-square deviation (RMSD) between Cα of the template and our models and checked the phi-psi torsion angle using the MolProbity server. We selected the final models based on the phi-psi torsion angle statistics. The suggested models were in good accordance with the crystal structure geometry. Three-dimensional (3D) representation of WT or TWIK-1 ΔEx2 and TREK-1 models are shown in Figures S4A and S4B; 98.6 and 99.8% residues of each model protein were in the allowed regions of the Ramachandran plot (Figures S4C and S4D). Therefore, we expected these models to have a stable structure.

To investigate whether glycine 119, located in the exon-junction of TWIK-1 ΔEx2, is at a critical site in determining the K⁺ ionic conductivity, like that in WT TWIK-1, we performed conventional MD simulations with all atoms unconstrained. Figures 4A and 4B represent the selective filter residues of TREK-1 (TTIG) and TWIK-1 (STTG) in the WT and TWIK-1 ΔEx2. However, inside ΔEx2-G119E TWIK-1, the selectivity filter residues become TWIK-1 (STTE) (Figure 4C). Principal component analysis (PCA) of the covariance matrix was used to characterize the dominant modes of motion underlying the protein dynamics. The interconnected motions inside the protein are directly associated with stability and subsequently, the protein function. Figures 4D–4F shows the PC1 versus PC2 graph as a 2D-projection of the trajectory plot which elucidates the overall collective motion of the TWIK-1/TREK-1 in terms of Gibbs free energy. The graphs show that the WT system is the most stable; it had two minima which were close enough for the protein to jump from one minimum to another in MD simulation (Figure 4D). In the case of TWIK-1 ΔEx2, the protein had several minima and traveled from local to global minima (Figure 4E). However, in the case of TWIK-1 ΔEx2-G119E, the protein had three significant minima, which indicates that the protein has very high flexibility and does not have any stable conformation throughout the MD simulation (Figure 4F). We calculated three topmost eigenvectors to identify the dominant mode of motion and plotted the first eigenvector as a porcupine plot for WT, ΔEx2, and ΔEx2-G119E (Figures 4G–4I). WT TWIK-1 had very few arrows, which is indicative of stability (Figure 4G). TREK-1 in WT TWIK-1 simulation is the least affected. In TWIK-1 ΔEx2, it is visible that the TM helices show moderate movement compared to that in the WT TWIK-1 (Figure 4H). The motion of helices shows that there is little opening at the top and bottom of the selective filter, and this could be the reason for the passage of potassium through selective filter. In contrast, TWIK-1 ΔEx2-G119E shows very high amplitude motions of the overall protein structure compared to that observed in WT and ΔEx2 of TWIK-1 (Figure 4I). In this case, the nearby helices of the selective filter appear to show more motion. The length and direction of the arrow indicated the amplitude of motion. We calculated the number of ions transported across the membrane through visual inspection of the MD simulation trajectories. In the case of the WT TWIK-1, there was only two potassium ions were transported in the inside-out direction (Figure S4E and Video S1). The stable nature of the channel inside the WT TWIK-1 leads to a very compact structure in the selective filter region and limited transport. However, in the case of TWIK-1 ΔEx2, a total of three potassium ions were transported in the inside-out direction, and one in the outside-in direction across selective filter (Video S2). Deletion of exon two from TWIK-1 resulted in a higher conduction of potassium ions compared to that in the WT because the architecture of the selective filter was destabilized inside TWIK-1 ΔEx2. In the case of ΔEx2-G119E, only one potassium ion was transported across the selective filter in the outside-in direction (Video S3). Several potassium ions accumulated above the selective filter region at the extracellular site; however, they could not be transported because of the mutant G119E.

TWIK-1 ΔEx2 rescues the K⁺ current mediated by TWIK-1

(A) TWIK-1/TREK-1, (B) TWIK-1 ΔEx2/TREK-1, and (C) TWIK-1 ΔEx2-G119E/TREK-1 models indicated as secondary structures. TWIK-1 and TREK-1 structures are indicated using smudge green and light pink color. Selectivity filter residues for TWIK-1 (STTG) and TREK-1 (TTIG) represented by the green and pink stick model, respectively. In case of TWIK-1 ΔEx2-G119E (STTE), E119 is visible at the top of selectivity filter. Principal component 1 (PC1) and PC2 were calculated from the full MD simulation trajectory using Gromacs utility package and plotted as a graph (D–F) for TWIK-1/TREK-1, TWIK-1 ΔEx2/TREK-1, and TWIK-1 ΔEx2-G119E/TREK-1. X axis represents PC1 and Y axis represents PC2. Backbone atoms were used to determine eigenvectors from the simulation trajectories, and first eigenvector is reported for G. TWIK-1/TREK-1, (H) TWIK-1 ΔEx2/TREK-1, and (I) TWIK-1 ΔEx2-G119E/TREK-1. Backbones of TWIK-1 and TREK-1 were depicted by green and light pink colors, respectively. White arrow shows the direction of movement for TWIK-1/TREK-1 during MD simulations. Figures were rendered using PyMOL2.2 (J) *I-V* relationship of curves showed the averaged current amplitudes from Sc shRNA, TWIK-1 shRNA, TWIK-1 shRNA + TWIK-1 ΔEx2, or TWIK-1 shRNA + TWIK-1 ΔEx2 G119E-transfected WT astrocytes.

(K) Summary bar graph of Sc shRNA, TWIK-1 shRNA, TWIK-1 shRNA + TWIK-1 ΔEx2, or TWIK-1 shRNA + TWIK-1 ΔEx2 G119E currents plotted at −150 mV and +50 mV. Number on each bar indicates ‘n’ for each condition.

(L) Reversal potential of each experimental conditions. All Data are mean ± s.e.m. p-values were obtained with two-way ANOVA followed by Tukey’s post hoc test. ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Video S1. Side view of the potassium ion conductance of WT TWIK1-TREK1 model

Coils show protein. The SF region of TWIK1-TREK is centered at the screen; K⁺ ions represented by the gold sphere. Water and lipid molecules were hidden for clarity. TWIK1 is front facing, and TREK1 is at the back side. The movie generated using VMD program.

Download video file^{(2MB, mp4)}

Video S2. Side view of potassium ion conductance from Ex2del TWIK1-TREK1 model

The secondary structure represents protein. TWIK1 is focused on the front side, and TREK1 is behind the TWIK1. K⁺ and Cl⁻ are represented by gold and cyan sphere, respectively. TWIK1 is front facing, and TREK1 is at the back side. Water and lipid molecules were hidden for clarity. The movie generated using VMD program.

Download video file^{(3.3MB, mp4)}

Video S3. Side view of the potassium ion conductance from Ex2del-G119E TWIK1-TREK1 model

SF residues and protein were shown by stick, and coils, respectively. G119E is visible at the top of the SF region. K⁺ and Cl⁻ are represented by gold and cyan sphere, respectively. TWIK1 is front facing and TREK1 is at back side. Water and lipid molecules are hidden for clarity. The movie generated using VMD program.

Download video file^{(2.5MB, mp4)}

We confirmed using electrophysiology experiments that the glycine 119 proposed through modeling functions as a potassium selective filter in TWIK-1 ΔEx2. First, we tested whether the decrease in astrocyte K⁺ current following the TWIK-1 knockdown using shRNA against exon 2 was rescued with the expression of TWIK-1 ΔEx2. When TWIK-1 ΔEx2 was co-expressed under the TWIK-1 knockdown condition with TWIK-1 Ex2 shRNA, the astrocytic K⁺ currents were fully rescued. However, the K⁺ currents in astrocytes co-expressing TWIK-1 ΔEx2 G119E, a pore mutant of TWIK-1 ΔEx2, were not recovered (Figures 4J and 4K). In addition, the depolarized reversal potential in TWIK-1 Ex2 shRNA-transfected astrocytes was rescued (TWIK-1 shRNA only; −43.98 ± 5.82 mV vs. TWIK-1 shRNA+ TWIK-1 ΔEx2; −69.84 ± 3.28 mV) in the co-expression with TWIK-1 ΔEx2. However, co-expression of TWIK-1 ΔEx2 G119E and TWIK-1 Ex2 shRNA in astrocytes did not restore the reversal potential of astrocytes (Figure 4L). Therefore, TWIK-1 ΔEx2 can completely rescue the deficient activity of WT TWIK-1 in cultured astrocytes, and its glycine 119 acts as a potassium selective filter.

Tyrosine 120 of TWIK-1 ΔEx2 also works as a selectivity filter of potassium ion

The GYG motif in potassium channels is known as a potassium selectivity filter (SF). Since 120th tyrosine (Y) residues, a part of the GYG motif, is also present in the TWIK-1 ΔEx2 protein, we decided to investigate whether tyrosine 120 located in the TWIK-1 ΔEx2 is also at a critical site in determining the K⁺ ionic conductivity. We constructed additional mutant of TWIK-1 ΔEx2 (TWIK-1 ΔEx2 Y120F), in which 120^th tyrosine (Y) residue is replaced by phenylalanine (F), and we performed whole-cell and single-cell patch-clamp electrophysiology experiments. When TREK-1, TWIK-1 WT, TWIK-1 ΔEx2, or TWIK-1 ΔEx2 Y120F alone were transfected into HEK293T cells, only TREK-1 showed an outwardly rectifying current while TWIK-1 WT, TWIK-1 ΔEx2, and TWIK-1 ΔEx2 Y120F showed insignificant current (Figures S5A and S5B). However, heterologous co-expression of TWIK-1 ΔEx2 and TREK-1 in HEK293T cells indicated the linearity of the I-V relationship, like that in the co-expression of TWIK-1 WT and TREK-1 (Figures S5C and S5D). However, the linearity of the I-V relationship had disappeared when co-expression of TWIK-1 ΔEx2 Y120F and TREK-1, in which an outwardly rectifying current had detected (Figures S5C and S5D).

Next, we analyzed the single-channel kinetics of HEK293T cells transfected with TREK-1 or TWIK-1 alone, as well as co-expression of TREK-1 and TWIK-1 channels (TWIK-1 WT, TWIK-1 ΔEx2, and TWIK-1 ΔEx2 Y120F). Single-channel recordings obtained at pipette potentials of −80 mV and +80 mV, using pipette and bath solutions containing 150 mM KCl, are presented in Figure 5. No single-channel currents were observed in cells transfected with TWIK-1 or its mutants alone. In contrast, cells expressing TREK-1 exhibited single-channel currents with the characteristic outward rectification of TREK-1 channels (Figure 5A). When comparing cells transfected with TREK-1 alone to those co-transfected with TREK-1 and TWIK-1, cells co-expressing TREK-1 and TWIK-1 exhibited a loss of outward rectification, indicating TWIK-1 regulates TREK-1 channel activity. Similarly, co-transfection of TREK-1 with TWIK-1 ΔEx2 produced linear-like currents (Figure 5C), suggesting that TWIK-1 ΔEx2 also modulates TREK-1 channels. Cells co-transfected with TREK-1 and TWIK-1ΔEx2 Y120F yielded single-channel properties like TREK1 alone, indicating that the TWIK1 ΔEx2 Y120F mutant does not have a regulatory effect on TREK1 channels (Figure 5C). To analyze the linearity, we measured the rectification index using single-channel conductance obtained from Figure 5C. Single-channel conductance were calculated at +80 mV and −80 mV. The average conductance of TREK-1, TREK-1/TWIK-1, TREK-1/TWIK-1 ΔEx2, TREK-1/TWIK-1 ΔEx2 G119E, and TREK1/TWIK1ΔEx2-Y120F expressed in HEK293T cells were 90.0 ± 5.0, 75.4 ± 7.1, 82.9 ± 5.2, and 91.3 ± 9.2 pS at +80 mV, and 70.2 ± 7.1, 80 ± 9.6, 82.3 ± 6.9, and 71.5 ± 5.8 pS at −80 mV, respectively (n = 6). As shown in Figure 5D, the rectification index (single-channel I-V relationships) of TREK-1/TWIK-1 and TREK-1/TWIK-1 ΔEx2 are 0.94 ± 0.10 and 1.00 ± 0.05, respectively (n = 6). Our present data showed that the TWIK-1 ΔEx2 can display the linearity of the I-V relationship with TREK-1 in both whole-cell current and single-cell current level, and tyrosine 120 of TWIK-1 ΔEx2 also can work as a potassium selective filter.

Tyrosine 120 of TWIK-1 ΔEx2 is critical for the K⁺ current mediated by TWIK-1 ΔEx2

(A) Single-channel recordings of mChe-TREK-1, TWIK-1-IRES2-GFP, TWIK-1 ΔEx2-IRES2-GFP, and TWIK-1 ΔEx2 Y120F-IRES2-GFP in cell-attached patches of HEK293 cells. Single-channel currents were recorded at pipette potentials of +80 mV and −80 mV using pipette and bath solutions containing 150 mM KCl.

(B) Single-channel recordings of mChe-TREK1/TWIK-1-IRES2-GFP, mChe-TREK-1/TWIK-1 ΔEx2-IRES2-GFP, and TREK-1/TWIK-1 ΔEx2 Y120F-IRES2-GFP in cell-attached patches of HEK293 cells.

(C) Unitary conductances at +80 mV and −80 mV (n = 6). p-values were derived from Mann-Whitney U test. ∗p < 0.05 compared to mChe-TREK-1. ^#p < 0.05 compared to mChe-TREK-1/TWIK-1-IRES2-GFP.

(D) Bar graph indicates rectification indexes obtained single-channel conductance at +80 mV and −80 mV shown in (C). p-values were derived from Mann-Whitney U test. ^$p < 0.05 compared to mChe-TREK-1/TWIK-1 ΔEx2-IRES2-GFP.

Twik-1 nKO mice with exon 1 targeting exhibit deficiency of TWIK-1 protein in vivo

Twik-1 KO mice express TWIK-1 ΔEx2, which can perform functions like that of full-length TWIK-1; therefore, the exon2 deleted Twik-1 KO mice would not be suitable for the functional studies of TWIK-1 channel. Therefore, we generated another strain of Twik-1 KO (nKO) mouse targeting the protein coding sequence (CDS) region of exon 1 using the CRISPR/Cas9 system. Targeting with dual single-guided RNAs (sgRNA) results in a higher frequency of indels than using single sgRNAs¹⁹; therefore, we designed two sgRNAs targeting the CDS region of exon 1 and the contiguous intron region (Figure 6A). We obtained a heterozygous KO founder (F1) and confirmed that the target site was effectively deleted in the genomic DNA obtained from the tail of Twik-1 nKO mice. The homozygous KO mice (F2) were generated from an intercross between F1 heterozygous KO mice and confirmed through PCR genotyping. Primers were designed to amplify a 788 bp or a 442 bp fragment from WT or Twik-1 nKO mice, respectively (Figure 6B). Analysis of the mRNA from the brain samples of Twik-1 nKO mice showed the deletion of the rear part of exon 1 and partial insertion of intron (Figure 6C). In situ hybridization analysis using the deleted part of exon 1 as a probe did not detect any signals in the brains of Twik-1 nKO mice (Figure 6C). Twik-1 mRNA band was not detected in the RT-PCR analysis using primers targeting the deleted region of exon 1 and exon 2 (Figure 6D). Consequently, sequence analysis of the mRNA showed that frameshift occurs in the Twik-1 mRNA of Twik-1 nKO mice, leading to early termination (Figures S6A and S6B). We verified the expression of TWIK-1 in Twik-1 nKO mice using the antibody against the C-terminus of TWIK-1 (Figures S1D and S1E). Immunohistochemistry analysis showed no TWIK-1 signal in the whole brain, including the hippocampus and cerebellum in the brain slice of Twik-1 nKO mice compared to that in the brain slice from the littermate controls (Figure 6E). This tendency was also observed in hippocampal neurons and astrocytes (Figure 6F). Therefore, Twik-1 nKO mice express a TWIK-1 protein prematurely terminated in the N-terminal region, which does not contain the pore domain required for forming the potassium ion channels.

Generation of the *Twik-1* nKO mice

(A) A representative view of the CRISPR/Cas9 targeting strategy used for generating the *Twik-1* nKO mice. The targeting sites are indicated by red line, and the binding sites for single guide RNAs (sgRNAs) are marked in black, with the protospacer adjacent motif (PAM) sequence in red.

(B) A representative PCR genotyping result for TWIK-1 WT, heterozygous (Het), and homozygous (KO) mice.

(C) Schematic illustrations of WT and *Twik-1* nKO mRNA sequences. The ISH probe is indicated using the red line and the PCR primers for genotyping are indicated using blue arrows. *In situ* hybridization analysis of whole brain from adult male WT and *Twik-1* nKO mice.

(D) RT-PCR analysis of *Twik-1* gene expression in the cerebral cortex of WT and *Twik-1* nKO mice.

(E) WT and *Twik-1* nKO mice were stained using the TWIK-1 C-terminus nanobody in whole or in several brain regions. Scale bars and their size are indicated at the bottom of each image.

(F) Subregions of hippocampus from WT and *Twik-1* nKO mice showing expression of MAP2 (red) and TWIK-1 (green) or GFAP (red) and TWIK-1 (green). Arrow heads indicate astrocytes expressing TWIK-1. Scale bar = 20 μm.

The Twik-1 nKO mice show reduction in astrocytic passive conductance and K⁺ buffering activity

Astrocytes contribute to K⁺ buffering against extracellular K⁺ increases in the brain.²⁰ The passive conductance of astrocytes is important in the K⁺ buffering.⁷^,²¹ Therefore, we investigated the passive conductance of hippocampal astrocytes in response to changes in external K⁺ concentration (from 3.5 to 7 mM and 15 mM) in the brain slices from WT and Twik-1 nKO mice. K⁺ currents were increased when the [K⁺]_o changed from 3.5 to 7 mM and 15 mM in the hippocampal astrocytes of WT mice. The K⁺ currents in hippocampal astrocytes of Twik-1 nKO were not altered with the change in [K⁺]_o. The I-V relationship was always linear regardless of the changes in the extracellular K⁺ concentration (Figures 7A and 7B). Comparing the I-V curves to K⁺ currents in WT mice, both outward and inward currents in Twik-1 nKO were dramatically reduced (Figure 7C). The exon 2 deleted Twik-1 KO mice showed slightly increased passive conductance, unlike our Twik-1 nKO mice (Figures S7A and S7B), in accordance with earlier reports.²² The amplitudes of K⁺ currents in the hippocampal astrocytes of Twik-1 nKO were the lowest (Figure 7D). Therefore, the hippocampal astrocytes from Twik-1 nKO mice displayed significant reduction in passive conductance and K⁺ buffering.

*Twik-1* nKO mice have altered basic electrophysiological properties in the hippocampal astrocytes

(A) Whole-cell currents evoked by command voltages stepped from −140 mV to +20 mV in 20 mV increments in WT and *Twik-1* nKO astrocytes when the bath solution K+ concentration was elevated from 3.5 mM to 7 mM and to 15 mM.

(B and C) The corresponding current–voltage relationships for the recordings shown in A as indicated. Symbols are defined in A. The currents used for these I–V plots were measured at the ends of the test pulses.

(D) Representative amplitude quantification of astrocytic passive conductance. All Data are indicated mean ± s.e.m. p-values were obtained with Student’s t test. ∗∗p < 0.01.

(E) Schematic representation of the experimental protocol: WT and *Twik-1* KO mice are systemically injected with KA (35 mg/kg, i.p). The behavioral seizures were evaluated.

(F) Time course of mean behavioral seizure score following KA injection.

(G) Latency to score 3 seizure for each group.

(H) Total cumulative seizure score for each group. All Data are presented as mean ± s.e.m. p-values were obtained with two-way ANOVA followed by Bonferroni’s post hoc test or Student’s t test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001.

The dysregulation in astrocytic potassium buffering in response to extracellular K⁺ concentration is involved in epileptic seizure.²³ In addition, the potassium ion channels in neuron are important in determining susceptibility to epileptic seizures.²⁴^,²⁵^,²⁶ To determine seizure severity in Twik-1 nKO mice, we observed their seizure behavior following the intraperitoneal injection of 35 mg/kg kainic acid (KA) in both the WT and Twik-1 nKO mice (Figure 7E). We scored the behavioral seizure responses using the modified Racine scale for 90 min following KA injection. Twik-1 nKO mice showed a higher seizure score throughout the observation period compared to WT mice (Figure 7F). In addition, the latency to score 3 was significantly shorter, and the cumulative score of seizure events was higher in Twik-1 nKO mice compared to that in WT mice (Figures 7G and 7H). Therefore, Twik-1 nKO mice are more susceptible to KA-induced seizures.

The deficiency of astrocytic TWIK-1 and the resulting dysfunction of passive conductance accelerate the onset of seizures

We aimed to determine if the sensitivity to seizures could be regulated by astrocytic TWIK-1 alone. AAV-R-CREon TWIK-1 shRNA was developed to deplete TWIK-1 exclusively in the astrocytes and was stereotaxically injected into the stratum radiatum of the hippocampal CA1 region, with the astrocyte-specific GFAP promoter-driven CRE virus²⁷ (Figure 8A). After 2 weeks, immunohistochemistry (IHC) confirmed TWIK-1 depletion in astrocytes (Figures 8B and 8C). The AAV-R-CREon system selectively expressed TWIK-1 shRNA and fluorescent proteins in astrocytes. The knockdown efficiency of TWIK-1 was high,⁷ and we measured the passive conductance in these astrocytes; it was significantly reduced (Figure 8D). The mean I-V relationship showed that the hippocampal astrocytes of mice infected with AAV-R-CREon TWIK-1 shRNA/AAV-GFAPp-BFP-Cre showed approximately 45% less passive conductance in both the outward and inward currents (Figures 8E and 8F). In addition, their input resistance was considerably increased compared to that in the hippocampal astrocytes of AAV-R-CREon TWIK-1 shRNA/AAV-GFAPp-BFP-infected mice (Figure 8G). Therefore, we used this AAV system to selectively knockdown TWIK-1 in the hippocampal astrocytes and monitor KA-induced seizure behavior (Figure 8H). The behavioral seizure responses during the 90 min observation indicated that AAV-R-CREon TWIK-1 shRNA/AAV-GFAPp-BFP-Cre-infected mice rapidly processed seizure onset compared to AAV-R-CREon TWIK-1 shRNA/AAV-GFAPp-BFP-treated mice (Figure 8I). Therefore, the time to reach score 3 decreased; however, there was no change in the cumulative score (Figures 8J and 8K). We evaluated the neuronal damage in the hippocampus using Fluoro-jade B (FJB) staining.²⁸ The positive signals for FJB were significantly increased in the CA1 of the hippocampus after KA administration in AAV-R-CREon TWIK-1 shRNA/AAV-GFAPp-BFP mice (Figures 8L and 8M). Therefore, TWIK-1 deficiency in astrocytes significantly affects the reduction in passive conductance as well as the decision of seizure onset.

Astrocytic ablation of TWIK-1 showed decreased passive conductance and KA-induced early onset of seizures

(A) Diagram showing the location of the stereotaxical injection of AAV-GFAPp-BFP/AAV-R-CREon TWIK-1 shRNA or AAV-GFAPp-BFP-Cre/AAV-R-CREon TWIK-1 shRNA to the stratum radiatum.

(B) Scheme representation of BFP, GFP, and mCherry under BFP or BFP-Cre virus infection.

(C) Immunohistochemical staining of the hippocampal slice with anti-TWIK-1 antibody. Scale bar = 25 μm. Enlarged images show expression of TWIK-1 (green), mCherry (red), GFAP (purple), and BFP (blue). Scale bar = 10 μm.

(D) The whole-cell currents from CREon TWIK-1 shRNA/GFAPp-BFP and CREon TWIK-1 shRNA/GFAPp-BFP-Cre-injected mice hippocampal astrocytes. Command voltage pulse protocol: voltages were stepped from −140 mV to +20 mV in increments of 20 mV from a holding potential of −60 mV. Dotted line indicates 0 pA.

(E) The current–voltage relationship (I–V) from −160 mV to +40 mV.

(F) The bar graph of current at −160 mV and +40 mV.

(G) The bar graph representing input resistance of respective conditions. All Data are presented as mean ± s.e.m. p-values were obtained with two-way ANOVA followed by Tukey’s post hoc test. ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

(H) Schematic representation of the experimental protocol: Mice are systemically injected with KA (35 mg/kg, i.p). Behavioral seizures were evaluated.

(I) Time course of mean behavioral seizure score following KA injection.

(J) Latency to score 3 seizure for each group.

(K) Total cumulative seizure score for each group.

(L) At three days following KA injection, there were Fluoro-Jade positive hippocampal CA1 neurons in the virus-injected mice. Scale bar = 50 μm.

(M) The graph represents the number of FJB⁺ neurons in the CA1 after a seizure. All Data are presented as mean ± s.e.m. p-values were obtained with two-way ANOVA followed by Bonferroni’s post hoc test or Student’s t test. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗∗p < 0.0001.

Discussion

The molecular mechanism underlying astrocytic passive conductance, and the importance of its physiological function have long been inconclusive. Earlier, we demonstrated through shRNA experiments that the TWIK-1/TREK-1 heterodimer was a very strong candidate for astrocytic passive conductance.⁷ However, the effect was not validated in TWIK-1/TREK-1 double knockout mice.²² Interestingly, while silencing of the Twik-1 gene using an shRNA system results in depolarization of membrane potentials in the hippocampal astrocytes and dentate gyrus granule cells,⁷^,⁸ renal principal cells and pancreatic β cells were hyperpolarized in Twik-1 KO mice.¹²^,²⁹ Although these studies have been performed in different cell types, the conflicting results in membrane potential in the Twik-1 knockdown and Twik-1 knockout mice prompted us to consider additional experimental validation method that considers both the off-target effect of knockdown system and compensatory effects of knockout systems. Therefore, in this study, we aimed to clarify that TWIK-1 is indeed a key molecule for passive conductance, using a variety of experimental methods (shRNA against each exons of Twik-1 gene, CRISPR/Cas9 system, and generation of another knockout mice) to overcome the experimental limitations, and to explore its physiological function using the exon 1-targeted Twik-1 KO mice.

First, we explored the reason underlying the discrepancy between the results from shRNA experiments and the existing Twik-1 KO mice. The Twik-1 gene consists of three exons; the Twik-1 KO mice were generated by deleting the second exon.¹¹ However, this could result in an incomplete protein as exon 1 and exon 3 of Twik-1 gene are linked in-frame. It was not tested whether this TWIK-1 ΔEx2 protein can be stably expressed and can replace the full-length TWIK-1 function. We constructed an antibody that recognizes the C-terminus of the TWIK-1 protein and can be used for immunostaining. Using this, we confirmed that the TWIK-1 ΔEx2 protein is stably expressed in the whole brain like the full-length TWIK-1 of WT mice (Figures 1 and S1). Using shRNA against different targets (exon 3 of Twik-1 gene) and the CRISPR/Cas9 system, we confirmed that the TWIK-1 ΔEx2 protein could act as one of the functional subunits of the potassium channel in primary cultured astrocytes and in in vivo systems (Figures 2 and S2). In addition, we verified that the TWIK-1 ΔEx2 protein can form a heteromer with TREK-1 via disulfide bonds, as previously reported by our group (Figures 3 and S3). Therefore, conventional Twik-1 KO mice have some functional replacement to mimic wild-type mice, at least in astrocytes; they can no longer be called ‘knockouts’ of Twik-1 gene.

Next, we investigated the occurrence of structural modifications at the filter site in the TWIK-1 ΔEx2 protein because the junction between exon 1 and exon 3 is critical for determining potassium permeability. To confirm that selectivity was maintained at this site, we developed 3D models for mouse TWIK-1/TREK-1 and TWIK-1 ΔEx2/TREK-1 using the reported human TWIK-1 and TREK-1 protein templates. These models demonstrated that 98.6 and 99.8% of the residues in the full-length TWIK-1 and TWIK-1 ΔEx2 models, respectively, were within the region of the Ramachandran plot, indicating the robustness and reliability of our models (Figures S4A–S4D). Earlier, we confirmed that the 119^th glycine is an important amino acid that determines the selectivity in the TWIK-1 potassium filter; potassium permeability was significantly reduced in the G119E mutant. We tested the concurrence of this result in the predicted 3D model. We assessed the stability of the simulations by calculating the backbone RMSDs; the wild-type TWIK-1 was the most stable, followed by TWIK-1 ΔEx2 and TWIK-1 ΔEx2-G119E (Figures 4A–4I). We artificially set up an environment with potassium and chlorine ions, performed MD simulations of each 3D model, and observed potassium conductance, which was consistent with the results of the electrophysiology experiments (Videos S1, S2, and S3). In addition, we elucidated that the TWIK-1 ΔEx2 can display a linear-like currents with TREK-1 using the both whole-cell and single-cell patch recordings (Figures 5 and S5). These data from molecular and electrophysiological experiments, and simulations strongly suggest that exon 2-deleted Twik-1 KO mice do not have functional deficits and are not suitable for use as proper Twik-1 KO mice (Figures 1, 2, 3, 4, and 5). Therefore, it is necessary to reinterpret the results of previous studies that used exon 2-deleted Twik-1 KO mice and to develop another Twik-1 KO mouse model with strategies to identify the role of Twik-1 KO mice.

In this study, we also generated a Twik-1 KO (Twik-1 nKO) mouse line that targets exon 1 instead of exon 2 using the dual CRISPR/Cas9 system (Figure 6). Twik-1 nKO mice produced a short truncated TWIK-1 protein that was terminated early, and the expression of TWIK-1 was not detected using the antibody against the C-terminus of TWIK-1 in the neurons and astrocytes (Figures 6E and 6F). The passive conductance in the hippocampal astrocytes of the Twik-1 nKO mice decreased markedly and remained unaffected by external potassium concentration, confirming that previous experiments with shRNA yielded more reliable results (Figures 6A–6D). Since the passive conductance of Twik-1 nKO mice was effectively reduced, we carefully analyzed the behavioral abnormalities of 8- to 10-week-old Twik-1 nKO mice. However, we could not find a significant difference in body weight and locomotion activity (Data not shown). If the residual passive conductance observed in Twik-1 nKO mice is sufficient to maintain normal brain function, it is also possible that this conductance might be supported by other ion channel proteins, albeit to a lesser extent than TWIK-1,³⁰ and that some compensation occurred during development. To rule out this possibility, further experiments with conditional KO mice targeting exon 1 of the Twik-1 gene will be needed. The exon 1-targeted Twik-1 KO mice, Twik-1 nKO mice, showed a reduction in effective passive conductance despite no apparent behavioral changes such as spontaneous seizures; therefore, we examined the activity of astrocytic potassium buffering in them. If the activity of potassium buffering in astrocytes has a defect, we assume that excessive potassium can accumulate outside the cells and induce neuronal excitability. Therefore, we established a KA-induced seizure model and compared the susceptibility of KA-induced seizure in wild-type and Twik-1 nKO mice. As expected, the Twik-1 nKO mice showed excessive seizure sensitivity (Figures 7E–7H). However, these results may also be due to the neuronal TWIK-1 ion channels, which contribute to intrinsic excitability and increase the excitability of neurons. Therefore, we tested passive conductance in astrocytes and KA-induced seizure susceptibility following astrocyte-selective removal of TWIK-1 only; we obtained similar results with that in Twik-1 nKO mice (Figure 8). Based on studies showing the interplay between impairments of astrocytic K⁺ clearance and networks of neuronal activity,³¹^,³² we propose that the deficiency of Twik-1 in astrocytes enhance severe results of KA-induced seizure behavior via dysfunction of astrocytic neuromodulation activity. Further studies will be needed to investigate detailed mechanisms of this TWIK-1-mediated astrocytic neuromodulation activity.

As the Twik-1 nKO mice appear to have normal growth (data not shown), it might be possible that astrocytic passive conductance per se could not be involved in pivotal functions under physiological conditions. Interestingly, astrocytic passive conductance is slightly reduced by knockdown of the Kir4.1 channel,⁷ while the astrocyte-specific deficit of kir4.1 channel results in spontaneous seizures.³³ Based on these previous studies, our cautious hypothesis is that the more critical factor determining potassium buffering may not be passive conductance per se, but rather a much larger role of Kir4.1, which keeps the membrane potential more negative and provides a driving force for cations such as potassium ions. We also could not exclude the possibility that the remaining astrocytic passive conductance shown in astrocytes of Twik-1 nKO mice, which are mediated by currently unidentified channels, could be sufficient for normal brain function. Although the answers to these questions are premature to estimate from our current results and further studies should be required, our data suggest that the Twik-1-mediated astrocytic passive conductance could play an important role in the setting of suprathreshold against KA-induced overstimulation in the brain.

Taken together, we have unequivocally identified TWIK-1 as the primary ion channel for passive conductance and confirmed that passive conductance is involved in determining KA-induced seizure susceptibility. In addition, our Twik-1 nKO mice can be used as invaluable experimental animals for interpreting the enigmatic physiological functions of TWIK-1 in the brain and other organs, such as the kidney, heart, and pancreas.

Limitations of the study

In this study, we demonstrate the exon 1-targeted Twik-1 KO mice display the reduced passive conductance and enhanced susceptibility to KA-induced seizure behavior. However, our experiments were conducted in astrocytes to examine the role of TWIK-1 channel. Thus, future studies of TWIK-1 in neurons and other cells with the Twik-1 KO mice will be informative on understanding the roles of TWIK-1.

Resource availability

Lead contact

Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Jae-Yong Park (jaeyong68@korea.ac.kr).

Materials availability

Specific requests about exon1-targeted Twik-1 knockout mice should be directed to Jae-Yong Park Jae-Yong Park (jaeyong68@korea.ac.kr).

Data and code availability

•
All data reported in this paper will be shared by the lead contact upon request.
•
This paper does not report the original code.
•
Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Acknowledgments

We are grateful to Dr. Florian Lesage for kindly providing the Twik-1 KO mice. This work was supported by the National Research Foundation (NRF) of Korea (2017R1A2B3012502 and 2022R1A2C1093143) to J.Y.P. This research was also supported by a grant of KIST intramural grant (2E32901), awarded to E.M.H.

Author contributions

E.M.H. and J.Y.P. designed the research and wrote the manuscript. A.K. and Y.B. conducted most of the experiments and acquired the data. C.G.G. and A.N.P. simulated the three-dimensional structure of the heteromer of TWIK-1 or its mutants and TREK-1 and analyzed the structural stability. Y.H.S. constructed Twik-1 nKO mice and H.G.J. maintained and validated this mouse line. J.W.L. measured the passive conductance of astrocytes according to the external potassium concentration. C.S. provided an antibody for the C-terminus of TWIK-1, and Y.E. and H.K. performed in situ hybridization. J.N., E.J.K, and D.K. performed additional construction of expression vectors and electrophysiology experiments for revision.

Declaration of interests

The authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies

Mouse anti TWIK-1	ELPIS Biotech	This paper
Alexa Fluor® 594 AffiniPure™ F(ab')₂ Fragment Donkey Anti-Mouse IgG (H + L)	Jackson ImmunoResearch	Cat# 715-586-150; RRID: AB_2340857
Mouse anti Flag	Sigma-Aldrich	Cat# F3165; RRID:AB_259529
Rat anti HA	Roche Life Science	Cat# 11867423001; RRID:AB_390918
Alexa Fluor® 594 AffiniPure™ F(ab')₂ Fragment Donkey Anti-Rat IgG (H + L)	Jackson ImmunoResearch	Cat# 712-586-150; RRID: AB_2340690
Alexa Fluor® 647 AffiniPure™ F(ab')₂ Fragment Donkey Anti-Rat IgG (H + L)	Jackson ImmunoResearch	Cat# 712-606-150; RRID: AB_2340695
DAPI	Sigma-Aldrich	D9542-10MG
wheat germ agglutinin and Alexa Fluor™ 488 Conjugate	ThermoFisher	W11261
Alexa Fluor® 647 AffiniPure™ F(ab')₂ Fragment Donkey Anti-Mouse IgG (H + L)	Jackson ImmunoResearch	Cat# 715-606-150; RRID: AB_2340865
Chicken anti MAP2	Abcam	Cat# ab5392; RRID:AB_2138153
Goat anti GFAP	Thermo Fisher Scientific	Cat# PA5-18598; RRID: AB_10984384
Alexa Fluor® 647 AffiniPure™ F(ab')₂ Fragment Donkey Anti-Chicken IgY (IgG) (H + L)	Jackson ImmunoResearch	Cat# 703-606-155; RRID: AB_2340380
Alexa Fluor® 594 AffiniPure™ F(ab')₂ Fragment Donkey Anti-Goat IgG (H + L)	Jackson ImmunoResearch	Cat# 705-586-147; RRID: AB_2340434
Rabbit anti TREK-1	Alomone Labs	Cat# APC-047; RRID: AB_2040136
Mouse anti GFP	Santa Cruz Biotechnology	Cat# sc-9996; RRID: AB_627695
Mouse anti TWIK-1	Santa Cruz Biotechnology	Cat# sc-517040
Goat anti TREK-1	Santa Cruz Biotechnology	Cat# sc-11556; RRID: AB_2280806
Mouse anti actin	Sigma-Aldrich	Cat# A5441; RRID: AB_476744
Mouse anti-rabbit IgG-HRP	Santa Cruz Biotechnology	Cat# sc-2357; RRID: AB_628497
mouse anti-goat IgG-HRP	Santa Cruz Biotechnology	Cat# sc-2354; RRID: AB_628490

Bacterial and virus strains

AAV-mChe-TWIK-1-Ex3-shRNA	This paper	N/A
AAV-mChe-Sc shRNA	Choi et al.⁸	N/A
AAV-GFAPp-BFP	Kim et al.²⁷	N/A
AAV-GFAPp-BFP-Cre	Kim et al.²⁷	N/A
AAV-R-CREon-TWIK-1 shRNA	This paper	N/A

Critical commercial assays

EzchangeTM site-directed Mutagenesis kit	Enzynomics	Cat#EZ004S
Gateway™ LR Clonase™ Enzyme Mix	Invitrogen	11791043
SensiFAST™ cDNA Synthesis Kit	BIOLINE	BIO-65054
Duolink® In Situ PLA® Probe Anti-Rabbit PLUS	Sigma-Aldrich	DUO92002
Duolink® In Situ PLA® Probe Anti-Mouse MINUS	Sigma-Aldrich	DUO92004
Duolink® In Situ Detection Reagents Green	Sigma-Aldrich	DUO92014
Neon® Transfection System	Thermo Fisher	MPP100, MPS100, MPK1025

Experimental models: Cell lines

Human: HEK293T cells	Korean Cell Line Bank	KCLB No.21573
Monkey: COS-7 cells	Korean Cell Line Bank	KCLB No.21651

Experimental models: Organisms/strains

B6; CrlOri	Orient Bio	N/A
B6. Twik-1 KO	Nie et al.¹¹	N/A
B6. Twik-1 New KO	This paper	N/A
Oligonucleotides used for genotyping	This paper	N/A
5′-TGCCTTTACGGTGTCTGCTC -3′	This paper	N/A
5′-GAGAAGGCAGAAATGGAAACT-3′	This paper	N/A
5′-CAGGAAGAGGAGGGTGAACG-3′	This paper	N/A
5′-ATTTGCACGAAACCAGGAA A-3′	This paper	N/A
5′- TCCAAGGGTCATTGCTAGGT-3′	This paper	N/A

Recombinant DNA

pDEST-GFP-C	Hwang et al.⁷	N/A
pDEST-HA-C	Hwang et al.⁷	N/A
pDEST-Flag-C	Hwang et al.⁷	N/A
pSicoR	Addgene	Addgene plasmid #11579
Control CRISPR/Cas9	Santa Cruz biotechnology	sc-421243
Twik-1 CRIPSR/Cas9	Santa Cruz biotechnology	Sc-418922
pBiFC-VN173	Hwang et al.⁷	N/A
pBiFC-VC155	Hwang et al.⁷	N/A

Software and algorithms

pClamp	Molecular devices	RRID:SCR_011323
NIS-Element	Nikon	RRID:SCR_014329
ImageJ	ImageJ	RRID: SCR_003070
Fiji	Fiji	RRID:SCR_002285
Clustal Omega	Clustal Omega	https://www.ebi.ac.uk/Tools/msa/clustalo/
MolProbity server	MolProbity server	http://molprobity.biochem.duke.edu/index.php
GraphPad Prism 9.2.0	Graphpad	RRID:SCR_002798

Open in a new tab

Experimental model and study participant details

Animals

Male C57BL/6 mice, 8–10 weeks old, were used for all experiments. Animal care and handling were performed according to the guidelines of the Institutional Animal Care and Use Committee (IACUC) at the Korea Institute of Science and Technology (Seoul, Korea, KIST-2020-067) and the Korea University (Seoul, Korea, KUIACUC-2019-0050). Exon 2-deleted Twik-1 KO mice were generated in previous studies,¹² and kindly gifted from Prof. Florian Lesage (Sophia Antipolis, France). Exon-1 targeted Twik-1 KO mice were generated using CRISPR/Cas9 through the customized genome-edited mouse service (ToolGen, Inc.). The sgRNA sequences were as follows: Twik-1-sgRNA1, 5′- TTATGAGGACCTGCTGCGCCAGG-3′ and Twik-1-sgRNA2, 5′-GTACTGCAGCTGTTACTCTTTGG-3′.

Cell line culture

HEK293T cells and COS-7 cells were purchased from the Korean Cell Line Bank (Seoul National University). These cells were free from mycoplasma. HEK293T cells were cultured in DMEM (Gibco) media supplemented with 10% fetal bovine serum (FBS) (Gibco) and 1% penicillin-streptomycin (Gibco). COS-7 cells were cultured in RPMI1640 (Gibco) media supplemented with 10% FBS and 1% penicillin-streptomycin at 37°C in a humidified atmosphere of 95% air and 5% CO₂.

Primary astrocyte culture

Cultures of primary astrocytes from p1 C57BL/6 mouse pups were prepared. In brief, the hippocampus was dissected free of adherent meninges, minced, and dissociated into a single-cell suspension by trituration. All experimental procedures were performed in accordance with the institutional guidelines of the Korea Institute of Science and Technology (KIST, Seoul, Korea). The cells obtained were seeded in a culture dish and grown in DMEM supplemented with 10% FBS and 1% penicillin–streptomycin. On the third day of culture, cells were washed by repeated pipetting, and the medium was replaced to remove debris and other floating cell types.

Method details

Plasmid and shRNA

To obtain the full-length cDNA of mouse Twik-1 (GeneBank accession no. NM_008430) and Trek-1 (NM_010607), constructs were generated using an RT-PCR-based gateway cloning method (Invitrogen). Twik-1 C69S, G119E, ΔEx2, ΔEx2-C69S, ΔEx2-G119E, and ΔEx2-Y120F mutants were obtained using EZchange site-directed mutagenesis kit. All constructs were cloned into pDEST-GFP-C, pDEST-mCherry-C, pDEST-HA-C, pDEST-Flag-C, and pDEST-IRES2-GFP vectors using Gateway cloning (Invitrogen). For gene knockdown in mouse brain and primary astrocytes, Twik-1 shRNA and Trek-1 shRNA were constructed. The target region of the shRNA against mouse Twik-1 exon 2: 5′- GCATCATCTACTCTGTCATCG-3′ and Trek-1: 5′-GCGTGGAGATCTACGACAAGT-3′ were described in our previous study.⁸ Exon 3 targeting Twik-1 shRNA (5′-GGAAGATGTTCTACGTG-AAGA-3′) was obtained through oligonucleotide-directed mutagenesis (Enzynomics). These target sequences were cloned into the pSicoR vector (Addgene, Cambridge, MA), and Sc shRNA was used as a control.

Gene knock-out using the CRISPR/Cas9 system

Twik-1 was knocked-out in primary astrocytes using the CRISPR/Cas9 system. Control and Twik-1 CRISPR/Cas9 plasmid were purchased from Santa Cruz biotechnology (sc-421243 and sc-418922, respectively). When different alleles of mutant clones had to be analyzed, target regions were amplified using PCR with 2X TOPsimpleTM DyeMIX-Forte (Enzynomics), amplicons were cloned using pTOP Blunt V2 (Enzynomics, Daejeon, Korea), and transformed into TOP10 competent cells. Bacterial colonies containing plasmids with inserts were selected through colony-direct PCR, and the PCR products were confirmed through DNA sequencing.

Genotyping

Twik-1 KO and Twik-1 nKO mice were genotyped using PCR. The primer sequences were as follows: Twik-1 KO IA34 (5′-TGCCTTTACGGTGTCTGCTC -3′), Twik-1 KO IA18 (5′-GAGAAGGCAGAAATGGAAACT-3′), Twik-1 KO A49 (5′-CAGGAAGAGGAGGGTGAACG-3′), Twik-1 nKO forward (5′-ATTTGCACGAAACCAGGAA A-3′), and Twik-1 nKO reverse (5′- TCCAAGGGTCATTGCTAGGT-3′).

RT-PCR

Total RNA was isolated from mouse brain tissues using an RNA purification Kit (GeneAll, Hybrid-R). cDNA was synthesized from 1000 ng total RNA; reverse transcription was performed using a SensiFAST cDNA Synthesis Kit (BIOLINE), according to the manufacturer’s instructions. For RT-PCR, 2x TOPsimpleTM DyeMIX-Tenuto (Enzynomics) was used. GAPDH was used as the loading control.

Gene	Sequence
Twik-1 Forward Primer (exon 1) Reverse Primer (exon 3)	5′- CAGCAATTATGGAGTGTCGG-3′ 5′-CTGCTCAGTGACGGAGGAGA-3′
GAPDH Forward Primer Reverse Primer	5′-GTCTTCACCACCATGGAGAA-3′ 5′-GCATGGACTGTGGTCATGAG-3′

Open in a new tab

In situ hybridization

In situ hybridization was performed as described by Kim et al.³⁴ In brief, frozen sections (14 μm thick) of WT and Twik-1 KO mice were cut coronally through the hippocampal formation. Sections were fixed in 4% formaldehyde, treated with 0.25% acetic anhydride in 0.1 M triethanolamine/0.9% NaCl (pH 8 0.0), dehydrated in increasing concentrations of ethanol, delipidated in chloroform, rinsed in ethanol, and finally air-dried. Hybridization probe specific for mouse Twik-1 mRNA was prepared using the following regions: nt 145−355 of Twik-1 (NM_008430). Antisense riboprobe was generated using 35S-uridine triphosphate (PerkinElmer #NEG039H250UC) and the Riboprobe system (Promega, P1440). The probes were labeled with α-35S-dATP deoxynucleotidyl transferase (GIBCO BRL). The sections were hybridized overnight at 37°C with 1 × 10⁶ cpm of labeled probe per slide. Radioactive signals were visualized by exposing the sections to β-max film (Kodak, Rochester, NY, USA) for 5 days.

Transfection

Transfections of expression plasmid into HEK293T or COS-7 cells were performed with Lipofectamine 2000 (Invitrogen) or polyethylenimine (PEI), according to the manufacturer instructions. In the case of primary astrocyte cells, on the fifth day of culture, cells were transfected with the indicated cDNAs and shRNAs through electroporation on a Neon Microporator (Invitrogen) using an optimized voltage protocol (1100 V, 20-ms pulse width, two pulses). The electroporated cells were seeded onto a culture plate and cultured for at least 24 h.

Immunocytochemical analysis and quantification

COS-7 cells seeded on coverslips were transfected with GFP-TWIK-1 or GFP-TWIK-1 ΔC for 24 h, and fixed with 4% PFA. Cells were permeabilized by treating with 0.3% Triton X-100 for 5 min. The coverslips were blocked using 3% BSA/5% normal donkey serum and incubated with TWIK-1 antibody (1:300, ELPIS Biotech) overnight at 4°C. On the next day, the cells were incubated with Alexa 594-conjugated secondary antibody (Jackson Immunoresearch) for 1 h and then stained with DAPI.

COS-7 cells were transfected as indicated with either Flag-TWIK-1, Flag-TWIK-1 ΔEx2 and HA-TREK-1 or empty vector – and Flag-expressing plasmids. After 24−72 h, the cells were fixed with 4% PFA, followed by staining with 5.0 μg/mL wheat germ agglutinin and Alexa Fluor 488 Conjugate (Thermofisher) for 10 min; the cells were then permeabilized with 0.3% Triton X-100 for 5 min. The coverslips were blocked using 3% BSA/5% normal donkey serum and incubated with anti-Flag Flag (M-2, Sigma-Aldrich, F3165) and anti-HA (3F10, Roche Applied Science) antibodies overnight at 4°C. On the next day, the cells were incubated with Alexa Fluor 594- or 647-conjugated secondary antibodies (Jackson Immunoresearch) for 1 h and then stained with DAPI.

WT and TWIK-1 KO mice primary astrocytes were transiently transfected with the Sc shRNA or TREK-1 shRNA construct using Neon Transfection System (Thermo Fisher). After 24−72 h, the cells were fixed with 4% PFA, followed by staining with 5.0 μg/mL wheat germ agglutinin and Alexa Fluor 488 Conjugate (Thermofisher) for 10 min; the cells were then permeabilized with 0.3% Triton X-100 for 5 min. The coverslips were blocked using 3% BSA/5% normal donkey serum and incubated with TWIK-1 nanobody overnight at 4°C. On the next day, the cells were incubated with Alexa 647-conjugated secondary antibody (Jackson Immunoresearch) for 1 h and then stained with DAPI. Imaging was performed on a confocal microscope (Nikon A1). TWIK-1 surface expression was assessed in 15–20 cells for each group by determining the Pearson’s coefficients using ImageJ software. The membrane fluorescence of TWIK-1 was assessed by comparing the mean gray value (fluorescence intensity)-distance (μm) profiles (line scans) in the acquired images of COS-7 and primary astrocytes. Lines were drawn across each cell, and linescan analysis was performed to identify TWIK-1 expression using the plot profile function on ImageJ.

Immunohistochemistry

Mice were anesthetized using avertin and subjected to intracardiac perfusion with saline, followed by 4% PFA solution in PBS. Brains were post-fixed in 4% PFA overnight at 4°C, and then 40 μm-thick sections were obtained using vibratome (Leica, VT1200). Immunohistochemistry was performed parallelly for all the groups. Free-floating sections were permeabilized with 0.5% Triton X-100 in PBS for 30 min at room temperature and subsequently incubated in blocking buffer (5% normal donkey serum, 3% BSA, and 0.2% Triton X-100 in PBS) for 1 h at room temperature. Sections were then incubated with primary antibodies, such as mouse TWIK-1 nanobody (Elpis-Biotech), Chicken anti-MAP2 antibody (Abcam cat# ab5392), Goat anti-GFAP antibody (Thermo fisher, cat# PA5-18598) overnight at 4°C. The sections were washed thrice in PBS and incubated with suitable fluorescence Alexa Fluor-tagged secondary antibodies (Jackson ImmunoResearch). The tissue sections were counter stained with DAPI and were mounted on glass slides for microscopy. Confocal images were obtained on a Nikon A1 confocal microscope.

Bimolecular fluorescence complementation assay

TWIK-1 and TREK-1 were cloned into the bimolecular fluorescence complement vectors, pBiFC-VN173 and pBiFC-VC155. To confirm the expression of each BiFC vector, additional Flag and HA tags were inserted in the C-terminal region of both BiFC vectors. HEK293T cells were co-transfected with the cloned BiFC vectors including VN-TREK-1 or VN-TWIK-1. The next day, these cells were fixed with 4% paraformaldehyde for 15 min at room temperature and permeabilized with 0.3% Triton X-100 for 5 min. After blocking for 1 h, the cells were incubated with rat anti-HA (1: 500, 3F10, Roche Applied Science) and mouse anti-Flag (1:500, M-2, Sigma-Aldrich, F3165) antibodies at 4°C overnight. Then, the cells were incubated with Alexa Fluor 594- or 647-conjugated secondary antibodies (Jackson ImmunoResearch Laboratories) for 1 h and then stained with DAPI to visualize nuclei. All images were acquired using confocal microscopy on a Nikon A1 confocal microscope.

Co-immunoprecipitation (Co-IP) and immunoblotting

For Co-IP, primary astrocytes and transfected HEK293T cells were lysed with lysis buffer (50 mM Tris, 150 mM NaCl, 1% NP-40, 1% Triton X-100, 2 μM EDTA, and 1 mM PMSF containing protease inhibitor cocktail, pH 7.4). The lysates were centrifuged for 20 min at 13,000 rpm at 4°C. The supernatant was transferred to a new tube and 5% of it was saved as the input control. 2 μg of Anti-TREK-1 (Alomone labs, APC-047) or anti-GFP (B-2, Santa Cruz) antibody were incubated with protein A/G PLUS Agarose (Santa Cruz) for 30 min and then washed thrice with ice-cold DPBS. The lysates were then incubated with antibody-A/G Agarose beads complex on a rotator overnight at 4°C. After three washes with ice-cold DPBS containing protease inhibitor cocktail, the lysate-bead complex was centrifuged at 2000 rpm. For western blotting, primary astrocytes and hippocampal tissues were lysed with lysis buffer (50 mM Tris, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.2% SDS, and 1 mM PMSF containing protease inhibitor cocktail, pH 7.4). Total protein (20 μg/lane) was subjected to SDS-PAGE and transferred to PVDF membranes. The membranes were blocked using 5% non-fat milk, and then, blotted with the appropriate antibodies: anti-TWIK-1 (Santa Crus, sc-517040), anti-TREK-1 (Santa Crus, sc-11556), anti-GFP (B-2, Santa Cruz, sc-9996), anti-HA (3F10, Roche Applied Science), and anti-actin (Sigma-Aldrich, A5441). The membranes were then washed and incubated with HRP-conjugated secondary antibodies. After a final wash, the immunoreactivity of the blots were evaluated using enhanced chemiluminescence (Thermo fisher).

Duolink-proximity ligation assay (PLA)

PLA was performed in WT and TWIK-1 KO mice primary astrocytes to determine whether ΔEx2 TWIK-1 interacts with TREK-1. Before the PLA assay, astrocytes were transiently transfected with Sc shRNA, TREK-1 shRNA, or TWIK-1 Ex3 shRNA construct using Neon Transfection System (Thermo Fisher). PLA was conducted according to the manufacturer’s instructions (Sigma-Aldrich). Briefly, cells were washed with PBS, fixed in 4% PFA for 15 min. After incubation with a blocking buffer for 1 h, the cells were incubated with the primary antibodies against TWIK-1 (1: 100, 4D7, Santa Cruz) and TREK-1 (1:200, APC-047, Alomone Labs). Donkey anti-rabbit PLAplus and donkey anti-mouse PLAminus probes were applied, followed by ligation and amplification with Duolink detection reagent Green, according to the manufacturer’s instructions.

Electrophysiological recording in cultured astrocytes

Cultured astrocytes were plated onto coverslips for electrophysiological experiments. The standard solution in the pipette contained 150 mM KCl, 1 mM CaCl₂, 1 mM MgCl₂, 5 mM EGTA, and 10 mM HEPES (pH 7.2, adjusted with KOH). Standard bath solution contained (in mM): 150 NaCl, 3 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, 5.5 D-glucose, and 20 sucrose (pH 7.4, adjusted with NaOH). Patch pipettes were made from borosilicate glass capillaries (Warner Instruments). The pipette resistance was 4–5 MΩ. Whole-cell currents were recorded using a patch clamp amplifier (Axon Instruments, Axopatch 700B). Current–voltage (I–V) curves were measured using ramped pulses (from −150 mV to +50 mV over 1000-ms) from a holding potential of −60 mV. A Digidata 1550 A interface (Axon Instruments) was used to convert digital– analog signals between the amplifier and the computer. Data were sampled at 5 kHz and filtered at 1 kHz. Currents were analyzed with Clampfit software (Axon Instruments). All experiments were conducted at room temperature.

Electrophysiological recording in COS-7 or HEK293T cells

Whole-channel patch-clamp recordings were performed to analyze I–V curves from COS-7 or HEK293T cells expressing TWIK-1-related constructs with or without TREK-1. The standard solution for the pipette contained (in mM): 150 KCl, 1 CaCl₂, 1 MgCl₂, 5 EGTA, and 10 HEPES (pH 7.2 was adjusted with KOH). Standard bath solution contained (in mM): 150 NaCl, 3 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES, 5.5 D-glucose, and 20 sucrose (pH 7.4 was adjusted with NaOH). The I–V relations were measured by applying 1-s ramp pulses (from −150 mV to +50 mV) from a holding potential of −60 mV. The Digidata 1550A interface was used to convert digital–analogue signals between the amplifier and the computer. Data were sampled at 5 kHz and filtered at 1 kHz. Currents were analyzed using Clampfit software (Axon Instruments, Union City, CA, USA). All experiments were conducted at approximately 25°C.

Single-channel patch-clamp recordings were conducted with an Axopatch 200B amplifier (Axon Instruments). Single-channel currents were filtered at 2 kHz with an 8-pole Bessel filter (−3 dB; Frequency Devices, Haverhill, MA) and digitized using the Axon Digidata 1550B interface (Molecular Devices, San Jose, CA, USA) at a sampling rate of 20 kHz. The threshold for channel openings was set at 50%. Single-channel current analysis was performed using pCLAMP software (version 10, Axon Instruments). The filter dead time for single-channel analysis was set at 100 μs (0.3/cutoff frequency), and events shorter than 50 μs were excluded from detection. Single-channel current traces shown in the figures were filtered at 2 kHz. For cell-attached patch experiments, both the pipette and bath solutions contained (in mM): 150 KCl, 1 MgCl₂, 5 EGTA, and 10 HEPES (pH 7.3), with pH adjusted using HCl or KOH as needed. All experiments were conducted at approximately 25°C. Mann-Whitney U tests were performed to compare the single channel conductances and rectification indexes.

Slice preparation and electrophysiological recording in hippocampal astrocytes

Hippocampal brain slices (thickness: 300 μm) were prepared from mice (8 weeks old). Following decapitation, the brain was rapidly removed and placed in cold artificial cerebrospinal fluid (ACSF) comprising (in mM): 130 NaCl, 24 NaHCO₃, 3.5 KCl, 1.25 NaH₂PO₄, 0.5 CaCl₂, 3.5 MgCl₂, and 10 glucose, pH 7.4, at room temperature with oxygenation (95% O₂ and 5% CO₂). For the recording, ASCF was changed to 1.5 CaCl₂ and 1.5 MgCl₂. The slices were prepared using an oscillating tissue slicer (Leica VT1000s) at 4°C and stored in room temperature with oxygenation (95% O₂, 5% CO₂); the prepared slices were allowed to recover for at least 1 h before recording. For passive conductance recording, patch pipettes had a resistance of 5–7 MΩ when filled with pipette solution composed of (in mM): 140 KCl, 10 HEPES, 5 EGTA, 2 Mg-ATP, 0.2 NaGTP, adjusted to pH 7.2 with KOH. Whole-cell patch recordings were performed in hippocampal astrocytes with a voltage-clamp configuration using a patch clamp amplifier (Axon Instruments, Axopatch 700B). Astrocytic passive currents were recorded using a step pulse protocol (from −160 mV to 40 mV with 10 mV increase for 1000-ms) from a holding potential of −60 mV. For recording the potassium buffering, each slice was transferred from a recovery/holding reservoir to the recording chamber of a fixed-stage upright microscope (Olympus BX51WI) and submerged in oxygenated ACSF that was supplied to the chamber at a rate of 1.5–2 mL/min. The submerged slice was visualized either directly via the microscope’s optics, or indirectly via a high resolution CCD camera system (optiMos, Qimaging) that received the output of a CCD camera attached to the microscope’s video port. Recordings were obtained using Multiclamp 700B (Axon Instruments) and were filtered at 1−2 kHz. Whole-cell recordings from the hippocampal astrocytes were documented with K-gluconate-based internal solution composed of (mM): 140 K-gluconate, 2 MgCl₂, 10 EGTA, 10 HEPES, 4 MgATP, 0.3 Na₃-GTP and pH 7.2 (OSM = 304). Holding potential was −70 mV. Astrocytic passive currents were recorded using the step protocol (from −150 mV to 50 mV with 10 mV increase for 1000-ms) and digitized at 5 kHz. All trace was analyzed using pCLAMP 10 software (Axon Instruments). Experiments with a holding current of more than −100 pA or in which there was a change >30% in the input resistance of the control were rejected.

Stereotaxic injection

WT and TWIK-1 KO mice (6–7 weeks old) were anesthetized with avertin (2,2,2-tribromethanol in 2-methyl 2-butanol) and placed in a stereotaxic frame (Kopf Instruments, Tujunga, CA, USA). Briefly, the scale was opened and two holes were drilled in the skull (−1.7 mm AP, ±1.4 mm ML from bregma). AAV-Sc shRNA and AAV-TWIK-1-Ex3-shRNA (2.21 × 10¹⁴ GC/mL) were packaged in the serotype DJ at KIST Virus Facility. These viruses were bilaterally injected (250 nL per side) into the Stratum radiatum area (1.55–1.6 mm DV from the dura) through a Hamilton Syringe with a syringe pump (KD Scientific, Holliston, MA, USA) that infused the virus at a speed of 0.1 μL/min. The Hamilton Syringe was left undisturbed at the injected points for 10 min.

Homology modeling

Full-length mouse TWIK-1 consists of 336 amino acid residues. After exon two deletion (ΔEx2) the resultant sequence consisted of 204 aa. The closest homolog of mouse TWIK-1 (UniProt accession number: O08581) is human TWIK-1 (UniProt accession number: O00180). The template structures for modeling mouse TWIK-1 and TREK-1 are human TWIK-1 (PDB code: 3UKM) and TREK-1 (PDB code: 4TWK) 3D-structures. Multiple sequence alignment of the query and the template protein was generated using the Clustal Omega program (https://www.ebi.ac.uk/Tools/msa/clustalo/). The resultant alignment was used in the Modelller9v7 program to develop 50 models each for TWIK-1 and TREK-1. A crucial disulfide bond between TREK-1/TWIK-1 (Cys108-Cys69) was generated in the heterotrimeric models. Five potassium ion coordinates were transformed from the crystal structure to the model. Heterotrimeric models were selected based on the root-mean-square deviation (RMSD) between the Cα of template and models. Selected models were further evaluated using the MolProbity server (http://molprobity.biochem.duke.edu/index.php) to check proper phi-psi torsion angle. We selected final models based on the Phi-psi torsion angle statistics. In addition, we generated the heterodimeric model of the G119E mutant for further simulation.

Root-Mean-Square Deviation (RMSD) analysis

To check the stability of TWIK-1/TREK-1 and TWIK-1 ΔEx2/TREK-1 models, we computed the RMSD for backbone atoms as a function of time using trajectories. RMSD graph for WT shows that the system took ∼10000ps time to equilibrate and later it formed a plateau throughout the simulation. The average RMSD is 0.37(±0.03) nm. Parenthesis include standard deviations. In the case of ΔEx2 TWIK1-TREK1, it is visible that the simulation system took comparatively longer time (∼40000ps) for equilibration than the WT system. After ∼50000ps, it formed a plateau and showed a decline in RMSD after ∼70000ps. Average RMSD for the ΔEx2 TWIK1-TREK1 system is 0.4 (±0.05) nm. Contrary to that observed for WT and ΔEx2, the ΔEx2-G119E mutant systems showed an initial ∼10000ps time for equilibration, and later formed a plateau up to ∼37000ps. However, later it showed consistent increment in the RMSDs throughout the simulation period. The average RMSD was 0.5 (±0.17) nm for ΔEx2-G119E TWIK1-TREK. From the RMSD graphs, it is evident that the WT system is very stable, followed by the ΔEx2 system. The ΔEx2-G119E mutant system was very unstable.

Molecular dynamics (MD) simulation

MD simulation is conducted using Gromacs 5.0.6 installed on Sun-server at Korea Institute of Science and Technology Information (KISTI) supercomputing center. Selected heterotrimeric models were placed into 150 dipalmitoylphosphatidylcholine (DPPC) lipid bilayers using CHARMM-GUI membrane builder. Membrane-embedded heterotrimeric models were explicitly solvated using a water box extended 10 Å from the protein. The total charge on the simulation systems was neutralized by adding a 0.15 mM KCl solution; total atomic sizes in the full-length and ΔEx2 of TWIK-1 simulation cells indicated 56,871 and 62,845 atoms, respectively. The atoms of proteins, lipids, and ions were modeled using the CHARMM36 all-atom force field, and the TIP3P model was assigned for water. Lennard-Jones and Coulombic interactions were calculated using a group-based scheme and smooth particle-mesh Ewald method at 1.2 nm cutoff. All the simulation systems were minimized to 50000 steps using the steepest-descent method. Heterotrimeric simulation systems equilibrate in six steps with reduced force constant and increased simulation time of 25.0, 25.0, 25.0, 100.0, 100.0, and 100.0 ps (ps). The first three steps of equilibration consist of system thermalization at 323.15 K and, the later three steps of equilibration consist of pressurization at 1 bar. For the six-step equilibration, the force restraint on the backbones was reduced from 4000.0, 2000.0, 1000.0, 500.0, 200.0, to 50.0 and on the sidechains from 2000.0, 1000.0, 500.0, 200.0, 50.0, to 0.0, kJ/mol/nm,² respectively, to adjust the protein inside membrane. During the equilibration phase, temperature and pressure were coupled using the Berendsen algorithm. Production simulation was primarily conducted in an isothermal-isobaric ensemble (NPT), where the temperature was coupled using the Nosé-Hoover thermostat at 323.15 K with a coupling constant of 0.5 ps; pressure was coupled semi-isotropically at 1 bar using the Parinello-Rahman barostat with a coupling constant of 2 ps. The H-bonds in water and other molecules were constrained using SHAKE and P-LINCS algorithms, respectively. Production simulation at 100 ns with a time step of 10 fs (fs) was carried out for the homodimeric and heterotrimeric systems.

Principal Component Analysis (PCA)

PCA is a standard mathematical tool used to calculate the correlations in protein. This method was used to detect the motion in the flexible regions of proteins. PCA combined with physical models of protein motions can discriminate against the relevant conformational changes in the protein from the background atomic fluctuations. PCA techniques convert a series of potentially coordinated observations into a set of orthogonal vectors called principal components (PCs). These PCs are used to illustrate the most dominant modes of motions underlying protein dynamics. PCs are calculated in two steps, first, the calculation of covariance matrix using simulation trajectory, and second, diagonalization of the covariance matrix. Calculations of covariance matrix involve the superimposition of simulation trajectories over starting structures. PCA calculations are performed on the backbone atoms to determine the eigenvectors and concomitant eigenvalues. Largest eigenvalues containing eigenvectors indicate the lowest mode of motion in the protein structure. To calculate PCs, we used the g_covar and g_anaeig functions of the Gromacs package.

Kainic acid-induced seizure behavior

Kainic acid (Milestone Pharmtech USA Inc.) dissolved in saline was administered via intraperitoneal injection at a dose of 35 mg/kg. Male mice, 8–10 weeks old, were monitored for 90 min after the injection. Seizure scores were monitored every 5 min using a Racine scale (0, normal behavior; 1, immobilization; 2, head nodding; 3, whole body myoclonus; 4, continuous rearing and falling; 5, clonic–tonic seizure; 6, death.

Fluoro-Jade B (FJB) staining

At 72 h after KA injection, the mice were anesthetized using avertin followed by intracardiac perfusion with saline and 4% PFA solution in PBS. The brain was rapidly removed, post-fixed in 4% PFA overnight at 4°C, and then 40 μm thick sections were obtained using vibratome (Leica, VT1200). For FJB staining, the brain sections were mounted on slides and air-dried overnight. Slides were rehydrated for 5 min in 100% ethanol, 2 min in 70% ethanol, and 2 min in distilled H₂O. The slides were transferred to 0.06% potassium permanganate solution, for 5 min, followed by rinsing in water and then immersing in a solution of 0.1% acetic acid and Fluoro-Jade B (Sigma Aldrich) 0.0004%. The slides were washed thrice in distilled water and then cleared by immersing in xylene for 5 min before closing it with a coverslip with a mounting solution. Confocal images were obtained on a Nikon A1 confocal microscope.

Quantification and statistical analysis

Statistical analysis was performed using the GraphPad Prism 9.2.0 software. Data were analyzed using a one-way or two-way ANOVA followed by Turkey’s post hoc test. Numerical data are presented as mean ± standard error of the mean (s.e.m). The variances were similar between the groups compared. The statistical significance of the data was assessed using unpaired Student’s t test or Mann-Whistney U test. The significance level was set at p < 0.05, p < 0.01, p < 0.001, or p < 0.0001.

Published: December 12, 2024

Footnotes

Supplemental information can be found online at https://doi.org/10.1016/j.isci.2024.111587.

Contributor Information

Eun Mi Hwang, Email: emhwang@kist.re.kr.

Jae-Yong Park, Email: jaeyong68@korea.ac.kr.

Supplemental information

Document S1. Figures S1–S7

mmc1.pdf^{(2.8MB, pdf)}

References

1.Aller M.I., Wisden W. Changes in expression of some two-pore domain potassium channel genes (KCNK) in selected brain regions of developing mice. Neuroscience. 2008;151:1154–1172. doi: 10.1016/j.neuroscience.2007.12.011. [DOI] [PubMed] [Google Scholar]
2.Talley E.M., Solorzano G., Lei Q., Kim D., Bayliss D.A. Cns distribution of members of the two-pore-domain (KCNK) potassium channel family. J. Neurosci. 2001;21:7491–7505. doi: 10.1523/JNEUROSCI.21-19-07491.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
3.Cahoy J.D., Emery B., Kaushal A., Foo L.C., Zamanian J.L., Christopherson K.S., Xing Y., Lubischer J.L., Krieg P.A., Krupenko S.A., et al. A transcriptome database for astrocytes, neurons, and oligodendrocytes: a new resource for understanding brain development and function. J. Neurosci. 2008;28:264–278. doi: 10.1523/JNEUROSCI.4178-07.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Rajan S., Plant L.D., Rabin M.L., Butler M.H., Goldstein S.A.N. SUMOylation silences the plasma membrane leak K+ channel K2P1. Cell. 2005;121:37–47. doi: 10.1016/j.cell.2005.01.019. [DOI] [PubMed] [Google Scholar]
5.Plant L.D., Zuniga L., Araki D., Marks J.D., Goldstein S.A.N. SUMOylation silences heterodimeric TASK potassium channels containing K2P1 subunits in cerebellar granule neurons. Sci. Signal. 2012;5 doi: 10.1126/scisignal.2003431. [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Plant L.D., Dementieva I.S., Kollewe A., Olikara S., Marks J.D., Goldstein S.A.N. One SUMO is sufficient to silence the dimeric potassium channel K2P1. Proc. Natl. Acad. Sci. USA. 2010;107:10743–10748. doi: 10.1073/pnas.1004712107. [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Hwang E.M., Kim E., Yarishkin O., Woo D.H., Han K.S., Park N., Bae Y., Woo J., Kim D., Park M., et al. A disulphide-linked heterodimer of TWIK-1 and TREK-1 mediates passive conductance in astrocytes. Nat. Commun. 2014;5:3227. doi: 10.1038/ncomms4227. [DOI] [PubMed] [Google Scholar]
8.Choi J.H., Yarishkin O., Kim E., Bae Y., Kim A., Kim S.C., Ryoo K., Cho C.H., Hwang E.M., Park J.Y. TWIK-1/TASK-3 heterodimeric channels contribute to the neurotensin-mediated excitation of hippocampal dentate gyrus granule cells. Exp. Mol. Med. 2018;50:1–13. doi: 10.1038/s12276-018-0172-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Bae Y., Choi J.H., Ryoo K., Kim A., Kwon O., Jung H.G., Hwang E.M., Park J.Y. Spadin modulates astrocytic passive conductance via inhibition of TWIK-1/TREK-1 heterodimeric channels. Int. J. Mol. Sci. 2020;21 doi: 10.3390/ijms21249639. [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Wang W., Putra A., Schools G.P., Ma B., Chen H., Kaczmarek L.K., Barhanin J., Lesage F., Zhou M. The contribution of TWIK-1 channels to astrocyte K(+) current is limited by retention in intracellular compartments. Front. Cell. Neurosci. 2013;7:246. doi: 10.3389/fncel.2013.00246. [DOI] [PMC free article] [PubMed] [Google Scholar]
11.Nie X., Arrighi I., Kaissling B., Pfaff I., Mann J., Barhanin J., Vallon V. Expression and insights on function of potassium channel TWIK-1 in mouse kidney. Pflügers Archiv. 2005;451:479–488. doi: 10.1007/s00424-005-1480-9. [DOI] [PubMed] [Google Scholar]
12.Chatelain F.C., Bichet D., Douguet D., Feliciangeli S., Bendahhou S., Reichold M., Warth R., Barhanin J., Lesage F. TWIK1, a unique background channel with variable ion selectivity. Proc. Natl. Acad. Sci. USA. 2012;109:5499–5504. doi: 10.1073/pnas.1201132109. [DOI] [PMC free article] [PubMed] [Google Scholar]
13.Yarishkin O., Lee D.Y., Kim E., Cho C.H., Choi J.H., Lee C.J., Hwang E.M., Park J.Y. TWIK-1 contributes to the intrinsic excitability of dentate granule cells in mouse hippocampus. Mol. Brain. 2014;7:80. doi: 10.1186/s13041-014-0080-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
14.Choe S. Potassium channel structures. Nat. Rev. Neurosci. 2002;3:115–121. doi: 10.1038/nrn727. [DOI] [PubMed] [Google Scholar]
15.Oakes V., Furini S., Pryde D., Domene C. Exploring the dynamics of the TWIK-1 channel. Biophys. J. 2016;111:775–784. doi: 10.1016/j.bpj.2016.07.009. [DOI] [PMC free article] [PubMed] [Google Scholar]
16.Echeverri C.J., Beachy P.A., Baum B., Boutros M., Buchholz F., Chanda S.K., Downward J., Ellenberg J., Fraser A.G., Hacohen N., et al. Minimizing the risk of reporting false positives in large-scale RNAi screens. Nat. Methods. 2006;3:777–779. doi: 10.1038/nmeth1006-777. [DOI] [PubMed] [Google Scholar]
17.Lesage F., Reyes R., Fink M., Duprat F., Guillemare E., Lazdunski M. Dimerization of TWIK-1 K+ channel subunits via a disulfide bridge. EMBO J. 1996;15:6400–6407. doi: 10.1002/j.1460-2075.1996.tb01031.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Miller A.N., Long S.B. Crystal structure of the human two-pore domain potassium channel K2P1. Science. 2012;335:432–436. doi: 10.1126/science.1213274. [DOI] [PubMed] [Google Scholar]
19.Zhou J., Wang J., Shen B., Chen L., Su Y., Yang J., Zhang W., Tian X., Huang X. Dual sgRNAs facilitate CRISPR/Cas9-mediated mouse genome targeting. FEBS J. 2014;281:1717–1725. doi: 10.1111/febs.12735. [DOI] [PubMed] [Google Scholar]
20.Ryoo K., Park J.Y. Two-pore domain potassium channels in astrocytes. Exp. Neurobiol. 2016;25:222–232. doi: 10.5607/en.2016.25.5.222. [DOI] [PMC free article] [PubMed] [Google Scholar]
21.Tang X., Taniguchi K., Kofuji P. Heterogeneity of Kir4.1 channel expression in glia revealed by mouse transgenesis. Glia. 2009;57:1706–1715. doi: 10.1002/glia.20882. [DOI] [PMC free article] [PubMed] [Google Scholar]
22.Du Y., Kiyoshi C.M., Wang Q., Wang W., Ma B., Alford C.C., Zhong S., Wan Q., Chen H., Lloyd E.E., et al. Genetic Deletion of TREK-1 or TWIK-1/TREK-1 Potassium Channels does not Alter the Basic electrophysiological Properties of Mature hippocampal Astrocytes in situ. Front. Cell. Neurosci. 2016;10:13. doi: 10.3389/fncel.2016.00013. [DOI] [PMC free article] [PubMed] [Google Scholar]
23.David Y., Cacheaux L.P., Ivens S., Lapilover E., Heinemann U., Kaufer D., Friedman A. Astrocytic dysfunction in epileptogenesis: consequence of altered potassium and glutamate homeostasis? J. Neurosci. 2009;29:10588–10599. doi: 10.1523/JNEUROSCI.2323-09.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Dong P., Zhang Y., Hunanyan A.S., Mikati M.A., Cui J., Yang H. Neuronal mechanism of a BK channelopathy in absence epilepsy and dyskinesia. Proc. Natl. Acad. Sci. USA. 2022;119 doi: 10.1073/pnas.2200140119. [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Kim E.C., Zhang J., Tang A.Y., Bolton E.C., Rhodes J.S., Christian-Hinman C.A., Chung H.J. Spontaneous seizure and memory loss in mice expressing an epileptic encephalopathy variant in the calmodulin-binding domain of Kv7.2. Proc. Natl. Acad. Sci. USA. 2021;118 doi: 10.1073/pnas.2021265118. [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Huang W., Ke Y., Zhu J., Liu S., Cong J., Ye H., Guo Y., Wang K., Zhang Z., Meng W., et al. TRESK channel contributes to depolarization-induced shunting inhibition and modulates epileptic seizures. Cell Rep. 2021;36 doi: 10.1016/j.celrep.2021.109404. [DOI] [PubMed] [Google Scholar]
27.Kim S.C., Kim A., Park J.Y., Hwang E.M. Improved AAV vector system for cell-type-specific RNA interference. J. Neurosci. Methods. 2022;368 doi: 10.1016/j.jneumeth.2021.109452. [DOI] [PubMed] [Google Scholar]
28.do Nascimento A.L., Dos Santos N.F., Campos Pelágio F., Aparecida Teixeira S., de Moraes Ferrari E.A., Langone F. Neuronal degeneration and gliosis time-course in the mouse hippocampal formation after pilocarpine-induced status epilepticus. Brain Res. 2012;1470:98–110. doi: 10.1016/j.brainres.2012.06.008. [DOI] [PubMed] [Google Scholar]
29.Millar I.D., Taylor H.C., Cooper G.J., Kibble J.D., Barhanin J., Robson L. Adaptive downregulation of a quinidine-sensitive cation conductance in renal principal cells of TWIK-1 knockout mice. Pflügers Archiv. 2006;453:107–116. doi: 10.1007/s00424-006-0107-0. [DOI] [PubMed] [Google Scholar]
30.Zhou M., Du Y., Aten S., Terman D. On the electrical passivity of astrocyte potassium conductance. J. Neurophysiol. 2021;126:1403–1419. doi: 10.1152/jn.00330.2021. [DOI] [PMC free article] [PubMed] [Google Scholar]
31.Purushotham S.S., Buskila Y. Astrocytic modulation of neuronal signalling. Front. Netw. Physiol. 2023;3:1205544. doi: 10.3389/fnetp.2023.1205544. [DOI] [PMC free article] [PubMed] [Google Scholar]
32.Bellot-Saez A., Stevenson R., Kékesi O., Samokhina E., Ben-Abu Y., Morley J.W., Buskila Y. Neuromodulation of astrocytic K+ clearance. Int. J. Mol. Sci. 2021;22:1–19. doi: 10.3390/ijms22052520. [DOI] [PMC free article] [PubMed] [Google Scholar]
33.Djukic B., Casper K.B., Philpot B.D., Chin L.S., McCarthy K.D. Conditional knock-out of Kir4.1 leads to glial membrane depolarization, inhibition of potassium and glutamate uptake, and enhanced short-term synaptic potentiation. J. Neurosci. 2007;27:11354–11365. doi: 10.1523/JNEUROSCI.0723-07.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]
34.Kim H., Ko J.P., Kang U.G., Park J.B., Kim H.L., Lee Y.H., Kim Y.S. Electroconvulsive shock reduces inositol 1,4,5-trisphosphate 3-kinase mRNA expression in rat dentate gyrus. J. Neurochem. 1994;63:1991–1994. doi: 10.1046/j.1471-4159.1994.63051991.x. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Video S1. Side view of the potassium ion conductance of WT TWIK1-TREK1 model

Download video file^{(2MB, mp4)}

Video S2. Side view of potassium ion conductance from Ex2del TWIK1-TREK1 model

Download video file^{(3.3MB, mp4)}

Video S3. Side view of the potassium ion conductance from Ex2del-G119E TWIK1-TREK1 model

Download video file^{(2.5MB, mp4)}

Document S1. Figures S1–S7

mmc1.pdf^{(2.8MB, pdf)}

Data Availability Statement

•
All data reported in this paper will be shared by the lead contact upon request.
•
This paper does not report the original code.
•
Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

[bib1] 1.Aller M.I., Wisden W. Changes in expression of some two-pore domain potassium channel genes (KCNK) in selected brain regions of developing mice. Neuroscience. 2008;151:1154–1172. doi: 10.1016/j.neuroscience.2007.12.011. [DOI] [PubMed] [Google Scholar]

[bib2] 2.Talley E.M., Solorzano G., Lei Q., Kim D., Bayliss D.A. Cns distribution of members of the two-pore-domain (KCNK) potassium channel family. J. Neurosci. 2001;21:7491–7505. doi: 10.1523/JNEUROSCI.21-19-07491.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib3] 3.Cahoy J.D., Emery B., Kaushal A., Foo L.C., Zamanian J.L., Christopherson K.S., Xing Y., Lubischer J.L., Krieg P.A., Krupenko S.A., et al. A transcriptome database for astrocytes, neurons, and oligodendrocytes: a new resource for understanding brain development and function. J. Neurosci. 2008;28:264–278. doi: 10.1523/JNEUROSCI.4178-07.2008. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib4] 4.Rajan S., Plant L.D., Rabin M.L., Butler M.H., Goldstein S.A.N. SUMOylation silences the plasma membrane leak K+ channel K2P1. Cell. 2005;121:37–47. doi: 10.1016/j.cell.2005.01.019. [DOI] [PubMed] [Google Scholar]

[bib5] 5.Plant L.D., Zuniga L., Araki D., Marks J.D., Goldstein S.A.N. SUMOylation silences heterodimeric TASK potassium channels containing K2P1 subunits in cerebellar granule neurons. Sci. Signal. 2012;5 doi: 10.1126/scisignal.2003431. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib6] 6.Plant L.D., Dementieva I.S., Kollewe A., Olikara S., Marks J.D., Goldstein S.A.N. One SUMO is sufficient to silence the dimeric potassium channel K2P1. Proc. Natl. Acad. Sci. USA. 2010;107:10743–10748. doi: 10.1073/pnas.1004712107. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib7] 7.Hwang E.M., Kim E., Yarishkin O., Woo D.H., Han K.S., Park N., Bae Y., Woo J., Kim D., Park M., et al. A disulphide-linked heterodimer of TWIK-1 and TREK-1 mediates passive conductance in astrocytes. Nat. Commun. 2014;5:3227. doi: 10.1038/ncomms4227. [DOI] [PubMed] [Google Scholar]

[bib8] 8.Choi J.H., Yarishkin O., Kim E., Bae Y., Kim A., Kim S.C., Ryoo K., Cho C.H., Hwang E.M., Park J.Y. TWIK-1/TASK-3 heterodimeric channels contribute to the neurotensin-mediated excitation of hippocampal dentate gyrus granule cells. Exp. Mol. Med. 2018;50:1–13. doi: 10.1038/s12276-018-0172-4. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib9] 9.Bae Y., Choi J.H., Ryoo K., Kim A., Kwon O., Jung H.G., Hwang E.M., Park J.Y. Spadin modulates astrocytic passive conductance via inhibition of TWIK-1/TREK-1 heterodimeric channels. Int. J. Mol. Sci. 2020;21 doi: 10.3390/ijms21249639. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib10] 10.Wang W., Putra A., Schools G.P., Ma B., Chen H., Kaczmarek L.K., Barhanin J., Lesage F., Zhou M. The contribution of TWIK-1 channels to astrocyte K(+) current is limited by retention in intracellular compartments. Front. Cell. Neurosci. 2013;7:246. doi: 10.3389/fncel.2013.00246. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib11] 11.Nie X., Arrighi I., Kaissling B., Pfaff I., Mann J., Barhanin J., Vallon V. Expression and insights on function of potassium channel TWIK-1 in mouse kidney. Pflügers Archiv. 2005;451:479–488. doi: 10.1007/s00424-005-1480-9. [DOI] [PubMed] [Google Scholar]

[bib12] 12.Chatelain F.C., Bichet D., Douguet D., Feliciangeli S., Bendahhou S., Reichold M., Warth R., Barhanin J., Lesage F. TWIK1, a unique background channel with variable ion selectivity. Proc. Natl. Acad. Sci. USA. 2012;109:5499–5504. doi: 10.1073/pnas.1201132109. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib13] 13.Yarishkin O., Lee D.Y., Kim E., Cho C.H., Choi J.H., Lee C.J., Hwang E.M., Park J.Y. TWIK-1 contributes to the intrinsic excitability of dentate granule cells in mouse hippocampus. Mol. Brain. 2014;7:80. doi: 10.1186/s13041-014-0080-z. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib14] 14.Choe S. Potassium channel structures. Nat. Rev. Neurosci. 2002;3:115–121. doi: 10.1038/nrn727. [DOI] [PubMed] [Google Scholar]

[bib15] 15.Oakes V., Furini S., Pryde D., Domene C. Exploring the dynamics of the TWIK-1 channel. Biophys. J. 2016;111:775–784. doi: 10.1016/j.bpj.2016.07.009. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib16] 16.Echeverri C.J., Beachy P.A., Baum B., Boutros M., Buchholz F., Chanda S.K., Downward J., Ellenberg J., Fraser A.G., Hacohen N., et al. Minimizing the risk of reporting false positives in large-scale RNAi screens. Nat. Methods. 2006;3:777–779. doi: 10.1038/nmeth1006-777. [DOI] [PubMed] [Google Scholar]

[bib17] 17.Lesage F., Reyes R., Fink M., Duprat F., Guillemare E., Lazdunski M. Dimerization of TWIK-1 K+ channel subunits via a disulfide bridge. EMBO J. 1996;15:6400–6407. doi: 10.1002/j.1460-2075.1996.tb01031.x. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib18] 18.Miller A.N., Long S.B. Crystal structure of the human two-pore domain potassium channel K2P1. Science. 2012;335:432–436. doi: 10.1126/science.1213274. [DOI] [PubMed] [Google Scholar]

[bib19] 19.Zhou J., Wang J., Shen B., Chen L., Su Y., Yang J., Zhang W., Tian X., Huang X. Dual sgRNAs facilitate CRISPR/Cas9-mediated mouse genome targeting. FEBS J. 2014;281:1717–1725. doi: 10.1111/febs.12735. [DOI] [PubMed] [Google Scholar]

[bib20] 20.Ryoo K., Park J.Y. Two-pore domain potassium channels in astrocytes. Exp. Neurobiol. 2016;25:222–232. doi: 10.5607/en.2016.25.5.222. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib21] 21.Tang X., Taniguchi K., Kofuji P. Heterogeneity of Kir4.1 channel expression in glia revealed by mouse transgenesis. Glia. 2009;57:1706–1715. doi: 10.1002/glia.20882. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib22] 22.Du Y., Kiyoshi C.M., Wang Q., Wang W., Ma B., Alford C.C., Zhong S., Wan Q., Chen H., Lloyd E.E., et al. Genetic Deletion of TREK-1 or TWIK-1/TREK-1 Potassium Channels does not Alter the Basic electrophysiological Properties of Mature hippocampal Astrocytes in situ. Front. Cell. Neurosci. 2016;10:13. doi: 10.3389/fncel.2016.00013. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib23] 23.David Y., Cacheaux L.P., Ivens S., Lapilover E., Heinemann U., Kaufer D., Friedman A. Astrocytic dysfunction in epileptogenesis: consequence of altered potassium and glutamate homeostasis? J. Neurosci. 2009;29:10588–10599. doi: 10.1523/JNEUROSCI.2323-09.2009. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib24] 24.Dong P., Zhang Y., Hunanyan A.S., Mikati M.A., Cui J., Yang H. Neuronal mechanism of a BK channelopathy in absence epilepsy and dyskinesia. Proc. Natl. Acad. Sci. USA. 2022;119 doi: 10.1073/pnas.2200140119. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib25] 25.Kim E.C., Zhang J., Tang A.Y., Bolton E.C., Rhodes J.S., Christian-Hinman C.A., Chung H.J. Spontaneous seizure and memory loss in mice expressing an epileptic encephalopathy variant in the calmodulin-binding domain of Kv7.2. Proc. Natl. Acad. Sci. USA. 2021;118 doi: 10.1073/pnas.2021265118. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib26] 26.Huang W., Ke Y., Zhu J., Liu S., Cong J., Ye H., Guo Y., Wang K., Zhang Z., Meng W., et al. TRESK channel contributes to depolarization-induced shunting inhibition and modulates epileptic seizures. Cell Rep. 2021;36 doi: 10.1016/j.celrep.2021.109404. [DOI] [PubMed] [Google Scholar]

[bib27] 27.Kim S.C., Kim A., Park J.Y., Hwang E.M. Improved AAV vector system for cell-type-specific RNA interference. J. Neurosci. Methods. 2022;368 doi: 10.1016/j.jneumeth.2021.109452. [DOI] [PubMed] [Google Scholar]

[bib28] 28.do Nascimento A.L., Dos Santos N.F., Campos Pelágio F., Aparecida Teixeira S., de Moraes Ferrari E.A., Langone F. Neuronal degeneration and gliosis time-course in the mouse hippocampal formation after pilocarpine-induced status epilepticus. Brain Res. 2012;1470:98–110. doi: 10.1016/j.brainres.2012.06.008. [DOI] [PubMed] [Google Scholar]

[bib29] 29.Millar I.D., Taylor H.C., Cooper G.J., Kibble J.D., Barhanin J., Robson L. Adaptive downregulation of a quinidine-sensitive cation conductance in renal principal cells of TWIK-1 knockout mice. Pflügers Archiv. 2006;453:107–116. doi: 10.1007/s00424-006-0107-0. [DOI] [PubMed] [Google Scholar]

[bib30] 30.Zhou M., Du Y., Aten S., Terman D. On the electrical passivity of astrocyte potassium conductance. J. Neurophysiol. 2021;126:1403–1419. doi: 10.1152/jn.00330.2021. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib31] 31.Purushotham S.S., Buskila Y. Astrocytic modulation of neuronal signalling. Front. Netw. Physiol. 2023;3:1205544. doi: 10.3389/fnetp.2023.1205544. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib32] 32.Bellot-Saez A., Stevenson R., Kékesi O., Samokhina E., Ben-Abu Y., Morley J.W., Buskila Y. Neuromodulation of astrocytic K+ clearance. Int. J. Mol. Sci. 2021;22:1–19. doi: 10.3390/ijms22052520. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib33] 33.Djukic B., Casper K.B., Philpot B.D., Chin L.S., McCarthy K.D. Conditional knock-out of Kir4.1 leads to glial membrane depolarization, inhibition of potassium and glutamate uptake, and enhanced short-term synaptic potentiation. J. Neurosci. 2007;27:11354–11365. doi: 10.1523/JNEUROSCI.0723-07.2007. [DOI] [PMC free article] [PubMed] [Google Scholar]

[bib34] 34.Kim H., Ko J.P., Kang U.G., Park J.B., Kim H.L., Lee Y.H., Kim Y.S. Electroconvulsive shock reduces inositol 1,4,5-trisphosphate 3-kinase mRNA expression in rat dentate gyrus. J. Neurochem. 1994;63:1991–1994. doi: 10.1046/j.1471-4159.1994.63051991.x. [DOI] [PubMed] [Google Scholar]

PERMALINK

Mice deficient in TWIK-1 are more susceptible to kainic acid-induced seizures

Ajung Kim

Yeonju Bae

Changdev G Gadhe

Hyun-Gug Jung

Esther Yang

Hyun Kim

Jaekwang Lee

Chanseob Shim

Young Hoon Sung

Junyeol Noh

Eun-Jin Kim

Dawon Kang

Ae Nim Pae

Eun Mi Hwang

Jae-Yong Park

Summary

Graphical abstract

Highlights

Introduction

Results

Twik-1 KO mice produce an internal region-deleted TWIK-1 mutant protein

Figure 1.

TWIK-1 ΔEx2 protein mediates potassium current in the primary cultured astrocytes from Twik-1 KO mice

Figure 2.

TWIK-1 ΔEx2 protein mediates astrocytic passive conductance in hippocampal slices of Twik-1 KO mice

TWIK-1 ΔEx2 protein can form a disulfide-linked heteromer with TREK-1 protein through its cysteine 69

Figure 3.

Glycine 119 of TWIK-1 ΔEx2 works as a selectivity filter of potassium ion

Figure 4.

Tyrosine 120 of TWIK-1 ΔEx2 also works as a selectivity filter of potassium ion

Figure 5.

Twik-1 nKO mice with exon 1 targeting exhibit deficiency of TWIK-1 protein in vivo

Figure 6.

The Twik-1 nKO mice show reduction in astrocytic passive conductance and K+ buffering activity

Figure 7.

The deficiency of astrocytic TWIK-1 and the resulting dysfunction of passive conductance accelerate the onset of seizures

Figure 8.

Discussion

Limitations of the study

Resource availability

Lead contact

Materials availability

Data and code availability

Acknowledgments

Author contributions

Declaration of interests

STAR★Methods

Key resources table

Experimental model and study participant details

Animals

Cell line culture

Primary astrocyte culture

Method details

Plasmid and shRNA

Gene knock-out using the CRISPR/Cas9 system

Genotyping

RT-PCR

In situ hybridization

Transfection

Immunocytochemical analysis and quantification

Immunohistochemistry

Bimolecular fluorescence complementation assay

Co-immunoprecipitation (Co-IP) and immunoblotting

Duolink-proximity ligation assay (PLA)

Electrophysiological recording in cultured astrocytes

Electrophysiological recording in COS-7 or HEK293T cells

Slice preparation and electrophysiological recording in hippocampal astrocytes

Stereotaxic injection

Homology modeling

Root-Mean-Square Deviation (RMSD) analysis

Molecular dynamics (MD) simulation

Principal Component Analysis (PCA)

Kainic acid-induced seizure behavior

Fluoro-Jade B (FJB) staining

Quantification and statistical analysis

Footnotes

Contributor Information

Supplemental information

The Twik-1 nKO mice show reduction in astrocytic passive conductance and K⁺ buffering activity