Calcium-activated chloride channels (CaCCs) mediate anionic currents in response to intracellular calcium levels [1-4], are widely expressed in various tissues, and play a crucial role in diverse physiological functions [5]. Anoctamin 1 (ANO1 or TMEM16A) has been initially cloned as a CaCCs that was activated by calcium and generated outwardly rectified currents blocked by anion channel blockers [6-8]. Anoctamin 2 (ANO2 or TMEM16B) was first identified as a chloride channel in mouse ano2 cRNA-injected oocyte [7] and has been found to be widely expressed in various tissues. It was first observed in the retina [9] and olfactory epithelium [10]. In the olfactory receptor neurons of the olfactory epithelium, ANO2 generates an inward current in response to calcium influx of the cyclic nucleotide-gated ion channel, which contributes to depolarization and action potential [10-12]. In photoreceptor cells of retinal pigment epithelium, ANO2 is responsible for ATP-induced inward current and depolarization [9, 13, 14]. ANO2 is also expressed in the central nervous system and is expected to produce an outward current, as the intracellular chloride is maintained at low levels in mature neurons. The outward current of ANO2 performs diverse functions in neurons throughout the brain, including reducing action potential duration in the hippocampus [15], amygdala [16], and cerebellum [17-19]. It also decreases excitability and induces spike frequency adaptation by afterhyperpolarization in the thalamus [20] and cerebellum [17-19]. Meanwhile, ANO2 also alters the hyperpolarization of the resting membrane potential in the pineal gland [21] and widens continuous potential oscillation variance in the striatum [22]. Also, ANO2 contributes to synaptic transmission between neurons such as reducing temporal summation of excitatory synaptic transmission in the hippocampus [15], reducing inhibitory synaptic transmission in the amygdala [16], attenuating inhibitory synaptic transmission in the cerebellum [19], and reducing excitatory synaptic transmission in the lateral septum [23]. Although the ANO2 channel is involved in several distinct roles for neuronal signaling depending on the brain region or cell type, assuming the existence of a single subtype of the ANO2 channel limits the understanding of its functional diversity with different Ca²⁺-sensitivity observed in the brain. Remarkably, ANO2 channels reportedly have multiple splicing variants that can affect distinct expression and function in the brain.

Although ANO2 is known to have several splice variants [10, 24-26], the functional implications of ANO2 isoforms are not yet well understood. ANO2 channels have a homodimer structure, with each dimer consisting of 10 transmembrane α-helices with cytosolic N- and C-terminal domains [27]. Their pore-like structure and calcium-binding sites across the transmembrane domains α6, α7, and α8 make the general architecture of the ANO (TMEM16) family, and amino acid sequences of the calcium-binding sites in these transmembrane domains are conserved in the ANO family [27-29]. As the first known CaCC, ANO1 (TMEM16A) channels, which are primarily expressed in sensory neurons, share characteristics with ANO2 [6-8]. The structural and functional characteristics of ANO1 have been extensively studied, whereas those of ANO2 have mostly been depicted by the sequence homology of ANO1. ANO1 channels have been reported to have multiple protein isoforms generated by alternative splicing of four segments (a, b, c, and d) [28]. Alternative splicing of ANO1 segments b or c, but not segment d, regulates rectification and calcium sensitivity [8, 30, 31]. In addition, the EF-hand-like region of ANO1 corresponding to exon 9 in ano1 (NC_000073.7, NCBI), whose acidic amino acids are well conserved in ANO2 [29], is involved in the modulation of calcium sensitivity in both ANO1 and ANO2 [32].

Alternative splicing of exon 14 of ANO2 channels has been reported to have regional specificity in different tissues, including the brain [25]. Ano2 exon 14 encodes for four amino-acid residues (⁴⁵⁶ERSQ, so-called segment c) within the intracellular loop between transmembrane domains 2 and 3 [10, 25]. Thus, alternative splicing of ANO2 can contribute to calcium-dependent channel activity and account for the regional and cellular diversity of function and expression of channels in the brain.

To date, alternative splicing of ano2 has been revealed to occur on exons 1 (a or b), 2, 4, and 14, resulting in multiple isoforms of ANO2. However, isoform-specific expression and neuronal function in different brain regions remain unclear. In this study, the regional distribution of ANO2 isoforms in the mouse brain and their channel properties in HEK293T cells were investigated.

Animals

Male C57BL/6J mice were maintained under a 12:12-light–dark cycle (lights on at 7:00 A.M.) and had ad libitum access to food and water. Animal care and handling were performed following the guidelines of the Institutional Animal Care and Use Committee at Yonsei University (Seoul, Korea). Animal experiments were approved by the Institutional Animal Care and Use Committee of Korea (approval number: IACUC-A-201911-983-02).

Reverse transcription-polymerase chain reaction (RT-PCR)

RNA was extracted from the adult C57BL/6L mouse brain and NIH3T3 cell line using TRIzol reagent (Invitrogen, Waltham, MA, USA). 1 µg RNA was reverse-transcribed using SuperScript™ III Reverse Transcriptase (Thermo Fisher Scientific, Waltham, MA, USA) to generate cDNA. cDNA from the adult mouse brain and NIH3T3 cell line was used as a template for PCR analysis. The primers were: ano2-F, 5’-GTC ATT GCT GTC ACC TCC GA-3’; ano2-R, 5’-TCA TCC AGT CCA CCA GGA CA-3’; exon 2-F, 5’-ATG GGC AGC AGT ACC TCA AAG-3’; exon 2-R2, 5’-ACT TTC CTC TGG TTG TCG TGA-3’; exon 1b-F, 5’-ATG GAT CCA GAA CAC CTG CC-3’; exon 1b-R, 5’-AAC AGC CAG AGA GTG TCC AG-3’; exon 4-F, 5’-TCT GTC TTT GTC CGG ATC CAC-3’; exon 4-R, 5’-GGT CCC AGA AGT AGC CCA AC-3’; exon 14-F, 5’-AAC GTT CCC AGG AAC ACT-3’; exon 14-R, 5’-CAA TGG AGA ACG TCA GGG C-3’; mouse β-actin-F, 5’-CGT GCC GCC TGG AGA AAC C-3’; and mouse β-actin-R, 5’-TGG AAG AGT GGG AGT TGC TGT TG-3’. PCR was carried out with Quest Taq PCR Kit I (Bioquest, Seoul, Korea) using the T100 thermal cycler (Bio-Bad, Hercules, CA, USA). To investigate the expression of ANO2 isoforms in the mouse brain, RT-PCR was performed under the following conditions: 95°C for 3 min for the initial denaturation; 95°C for 30 s, 55°C for 30 s, 72°C for 30 s for 32 cycles; and 72°C for 5 min for the final extension.

ANO2 isoform expression constructs

The expression constructs were all derived from mouse isoform 1 cDNA (NM_153589.2, gifted by KIST). An AflII site was introduced at the 5’ end and a NotI at the 3’ end of the fragments to fuse it into a multi-cloning site of the pcDNA5/FRT vector (Thermo Fisher Scientific, Waltham, MA, USA). To clone the mouse isoform 2 construct, mutagenesis was performed using overlapping PCR (Cosmogenetech, Seoul, Korea). By using the previously cloned isoform 1 construct, the following deletion was made in isoform 1 using primers: BamHI-F, 5’-GTC TTT GTC CGG ATC CAC GCC CCA TGG CAG GTG C-3’; HindIII-R, 5’-TTG AGC AAG AAA GCT TTG AGG ATC AAC CGT TCT TC-3’; exon 14 deletion-F, 5’-GAT CGA AGA GGA AGA AGA ACA CTC TCG GCC TGA ATA T-3’; and exon 14 deletion-R, 5’-CAG GCC GAG AGT GTT CTT CTT CCT CTT CGA TCC CAG T-5’. Conditions for PCR were as follows: 95°C for 5 min for the initial denaturation, 95°C for 60 s, 56°C for 45 s, 72°C for 80 s for 35 cycles, and 72°C for 5 min for the final extension. PCR products were digested with BamHI and HindIII. The ANO2 sequence containing BamHI and HindIII sites was eluted and infused into the pcDNA5/FRT vector. The resulting constructs were confirmed by sequencing.

Stable cell line generation

Flp-HEK293T cells (Invitrogen, Waltham, MA, USA) were transfected with isoform 1 or 2 constructs in 24-well plates, using Lipofectamine 2000 reagent (Thermo Fisher Scientific, Waltham, MA, USA). The pOG44 constructs (Addgene, Watertown, MA, USA) were also transfected to express flp recombinase in cells transfected with either isoform 1 or 2. To generate the stable cell line, the transfected cells were selected using 100 μg/ml hygromycin (Sigma Aldrich, Burlington, MA, USA) in DMEM (Thermo Fisher Scientific, Waltham, MA, USA) with 10% FBS (Thermo Fisher Scientific, Waltham, MA, USA) and 1% penicillin/streptomycin (Thermo Fisher Scientific, Waltham, MA, USA) for weeks. The stable cell line was stored in freezing media with 50% DMEM (Thermo Fisher Scientific, Waltham, MA, USA), 40% FBS (Thermo Fisher Scientific, Waltham, MA, USA), and 10% DMSO (Tocris, Bristol, UK).

To confirm whether transfected cell lines expressed the correct ano2 mRNA, RT-PCR was performed on cDNA from Flp-HEK293T and two stable cell lines (isoforms 1 and 2) using the following primers: ano2-F; ano2-R; exon14-F; exon14-R; human β-actin-F, 5’-CCG CGA GAA GAT GAC CCA GAT-3’; and human β-actin-R, 5’-GGA TAG CAC AGC CTG GAT AGC A-3’.Conditions for RT-PCR were as follows: 95°C for 3 min for initial denaturation, 95°C for 30 s, 55°C for 30 s, 72°C for 30 s for 30 cycles, and 72°C for 5 min for the final extension.

Quantitative reverse transcription-polymerase chain reaction (RT-qPCR)

To analyze ano2 and exon14 expression levels in the transfected cell lines, a separate quantitative RT-PCR was performed on cDNA from the two stable cell lines (isoforms 1 and 2) using the following primers: ano2-F; ano2-R; exon14-F; exon14-R; human β-actin-F; and human β-actin-R. PCR was carried out with TB Green Premix Ex Taq II (Takara, Kusatsu City, Japan) using the CFX Connect Real-Time PCR System (Bio-Bad, Hercules, CA, USA). Three-step procedure for qPCR was as follows: 95°C for 3 min for the initial denaturation, 95°C for 30 s, 55°C for 30 s, 72°C for 30 s for 40 cycles, and 72°C for 5 min for the final extension. The relative mRNA expression levels were quantified using the ΔΔCq method and normalized to β-actin.

Patch clamp recording

All electrophysiological traces were measured using the whole-cell patch clamp technique. Patch electrodes (3~6 MΩ) were pulled on a Sutter P-97 horizontal puller (Sutter Instrument Company, Novato, CA, USA) from borosilicate glass capillaries (World Precision Instruments, Sarasota, FL, USA). The recorded signals were amplified using a MultiClamp 700 B amplifier (Axon Instruments, Union City, CA, USA). Data acquisition was performed using Digitizer 1550 B and pClamp10 software (Axon Instruments, Union City, CA, USA). Cells with a series resistance <25 MΩ were used for statistical analysis.

HEK293T cell lines with stable expression of ANO2 isoforms 1 and 2 were used for all patch clamp recordings. Cells were grown on poly-D-lysine-coated glass or plastic coverslips and perfused with solution (in mM): 10 HEPES, 139 NaCl, 3 KCl, 2 CaCl₂, 2 MgCl₂, 6 glucose, 22 sucrose at 26°C and 36°C. The temperature was controlled using a TC-324B automatic temperature controller (Warner Instruments, Holliston, MA, USA). For the whole-cell current recording of the cell lines, the control intracellular solution contained 150 mM CsCl, 10 mM HEPES, and 10 mM HEDTA. [Cl^-]_extra/[Cl^-]_intra=150 mM/150 mM to adjust the reversal potential to 0 mV was calculated by the Nernst equation:

Erev=RT/zF ln([Cl-]extra/[Cl-]intra)(1)

where E_rev is the reversal potential, z is the valence of the ionic species, F is Faraday’s constant, R is the gas constant, and T is the absolute temperature.

To make the free-Ca²⁺ solution, CaCl₂ was added to the control solution. CsCl and CaCl₂ were added at a total Cl^- concentration of 150 mM. WEBMAXC STANDARD calculator (https://somapp.ucdmc.ucdavis.edu) was used to determine the free Ca²⁺ concentration.

To record the non-stationary current, the holding potential was -70 mV in voltage-clamp mode during 90~150 s of the current recording. To record the steady-state current, the holding potential was set to -70 mV. The current was recorded by 1 s voltage steps from -100 mV to 100 mV, followed by -100 mV for 90~150 s after membrane disruption.

Data analysis

The calculated conductance versus calcium was fitted into Hill’s equation:

G=Gmax[Ca2+]intranH/([Ca2+]intranH+EC50nH)(2)

where G is the conductance, G_max is the maximal conductance, K_1/2 is the half-maximal [Ca²⁺]_intra (also called EC₅₀), and n_H is the Hill coefficient.

Electrophysiological data analysis was performed using the Clampfit (Molecular device). The numbers and individual dots refer to the number of cells unless otherwise indicated in the figure legends. GraphPad Prism (GraphPad Software) was used for the data presentation and statistical analysis. Statistical significance was set at *p<0.05, **p<0.01, and ***p<0.001. Data are presented as the mean±SEM.

Molecular modeling of ANO2 channels

The ANO2 sequence of Mus musculus was obtained from the UniProt database (UniProt ID: Q8CFW1) [33]. Protein structures with and without exon 14 were predicted using AlphaFold2, which has been shown to have high accuracy [34]. In homology modeling, refined structure models were generated with AlphaFold2 and were superimposed with the protein structure of ANO1 (PDB ID: 5OYB) to form a dimer structure. Hydrogen atoms were re-added at pH 7.0, and their positions were optimized using PROPKA implemented in the Maestro program [35]. Restrained energy minimization was performed with OPLS3e within 0.3 Å root mean square deviation [36].

Cavity analysis was performed using CAVER 2.0 [37, 38]. The cavities in the ANO2 dimer structure were identified using a hydrophobic probe with 1.4 Å van der Waals radius. In CAVER, the protein body is modeled with a discrete 3-dimensional grid space of all atoms with an atomic van der Waals radius, and the most accessible paths are identified using graph-searching algorithms to find the short direct paths from the active sites to the external environment [38]. To compare the cavity volumes between ANO2 structures with and without exon 14, the most promising clustered cavities with the calcium-binding sites in a dimer state were selected from the external environment. ⁴⁷⁹ERSQ residues were positioned at the site of entry to the calcium-binding site. The entrance cavity was defined as the cavity closest to the ⁴⁷⁹ERSQ residues.

To analyze protein–protein interactions in ANO2 dimer structures, the quantum-mechanical analysis was performed with fragment molecular orbitals (FMO) and three-dimensional scattered pair interaction energies (3D-SPIEs) method to sort the significant interactions more stable than -3.0 kcal/mol and within 6 Å between two residues [39].

All FMO calculations were performed using the version of February 14, 2018, GAMESS [40]. The two-body FMO method was applied to the ANO2 dimer structure to investigate the key interactions at the second-order Møller–Plesset perturbation (MP2) [41] and polarizable continuum model (PCM) [42] with the 6-31G** basis set (FMO2-MP2/6-31G**/PCM level). In the FMO calculations, only residues within 10 Å from ⁴⁷⁹ERSQ residues in the dimer state were included, which were defined as T472–S485 in the structure with ⁴⁷⁹ERSQ residues. The residues without ⁴⁷⁹ERSQ were defined as T472–S481. All input files were prepared in compliance with the hybrid orbital projection scheme fragmentation [43].

The two-body FMO calculation consists of fragmentation, fragment self-consistent field calculation, fragment pair self-consistent field calculation, and total property evaluation [39]. Detailed information and state-of-the-art methodologies are described in the references [44, 45]. In the total property evaluation, FMO provides the pair interaction energies (PIE) between two fragments that correspond to the residues in this work. In the PIE decomposition analysis (PIEDA), the PIE between two fragments indicates the sum of five energies from electrostatic (∆E^es), exchange-repulsion (∆E^ex), charge-transfer, and mixed (∆E^ct+mix), dispersion (∆E^di), and solvation energy terms (∆G^sol).

PIE=ΔEes+ΔEex+ΔEct+mix+ΔEdi+ΔGsol(Equation 3)

The electrostatic term (∆E^es) arises from the Coulomb interaction between polarized charge distributions of fragments, the exchange-repulsion term (∆E^ex) arises from the interaction between fragments closely situated and is related to the overlap of two occupied orbitals, the charge-transform and mixed term (∆E^ct+mix) is derived from the interaction between the occupied orbitals of a donor and unoccupied orbitals of an acceptor, the dispersion term (∆E^di) is obtained from instantaneous dipole moments of fragments and is hydrophobic, and the solvation term (∆G^sol) arises from solute-solvent interactions with implicit solvation.

Two types of splice variants of ano2 are expressed in distinct brain regions

ANO2 has been reported to have several alternative splice variants [10, 24, 25]. Ano2 exon 2 (with 1a), 1b, and 4 were located in the N-terminal region, and exon 14 was located in the first intracellular loop (Fig. 1A, 1B). The distribution of splice variants encoded by exon 2 (with 1a), 1b, and 4 have not been reported, although splice variants of ano2 exon 14 have been examined in the rat brain. To investigate the expression patterns of ano2 splice variants in the mouse brain by targeting known alternative splice sites, primers were first designed to detect exons 1b, 2, 4, and 14 (Fig. 1B). In addition, exon 24 was targeted to detect the common ano2 exon (see Methods section for details).

To determine the distribution of ano2 splice variants in the mouse brain, mRNAs collected from the cerebral cortex, hippocampus, and thalamus, where the functional channel activity of ANO2 has been previously demonstrated, were compared with those from NIH3T3 cells (for the negative control), the olfactory epithelium (positive control for the variants with exons 1b and 4), and retina (positive control for the variants with exons 4 and 14). Remarkably, ano2 with exon 2 (with 1a) and exon 4 was detected in all the targeted brain regions, whereas mRNA with exon 1b was not detected in any of the targeted brain regions (Fig. 1C). Exons 1b and 4 were detected in the olfactory epithelium and exons 4 and 14 were detected in the retina (Fig. 1C), which is consistent with previous studies [10]. Notably, ano2 with exon 14 was expressed only in the hippocampal region but not in the cerebral cortex and thalamus (Fig. 1C). The distribution of splice variants of ano2 in the mouse brain is summarized in Fig. 1D. Given the alternative splicing of exons 1b, 2, 4, and 14, these ano2 variants can encode three ANO2 isoforms, particularly isoforms 1 and 2 in the brain (Fig. 1D). The splice variant without exon 14, encoding isoform 2, was more broadly detected in the thalamus and cerebral cortex, whereas the splice variant with exon 14, encoding isoform 1, was expressed in the hippocampus (Fig. 1D). ANO2 is an important CaCC for neuronal signaling, with diverse functions in the cerebral cortex, thalamus, and hippocampus [15, 17, 20], which coincides with the multiple alternative splicing of ano2. Notably, ANO2, but not calcium-activated potassium channels, contributes to spike frequency adaptation in the thalamus, whereas calcium-activated potassium channels control spike frequency adaptation in the hippocampus [20, 46]. Meanwhile, ANO2 in hippocampal neurons was shown to regulate action potential duration [15] which has not been observed in neurons in the thalamus [20]. These results demonstrate that the alternative splicing of ano2 can be attributed to the regional diversity of ANO2 channels, which may be associated with the distinct channel properties shown in these brain regions.

Alternative splicing of ANO2 alters calcium-dependent channel kinetics

To understand whether the identified alternative splicing of ano2 is attributable to the regional diversity of ANO2 functions, the biophysical properties of the two isoforms of ANO2 channels were examined in HEK293T cells expressing both isoforms individually by site-directed mutagenesis.

In stable cell lines derived from HEK293T cells transfected with isoform 1 or 2 constructs, the expression of each ANO2 isoform was confirmed by detecting the common exon and exon 14 of ano2 variants (Fig. 2A, 2B). cDNA from the cell line transfected with the isoform 1 construct had both the common ano2 exon and exon 14, whereas those with isoform 2 lacked exon 14 (Fig. 2B). An RT-qPCR analysis demonstrated that the expression levels of the common ano2 exon between two cell lines with isoforms 1 and 2 were similar (isoform 1: 1.03±0.10, n=6; isoform 2: 0.94±0.23, n=4; p=0.8892, unpaired t-test), however those of exon 14 was significantly lower in isoform 2 compared in isoform 1 (isoform 1: 1.06±0.15, n=6; isoform 2: 0.001±0.0001, n=5; p=0.0008, unpaired t-test) (Fig. 2C).

Using stable cell lines expressing ANO2 isoforms 1 or 2, a non-stationary current analysis was conducted to examine the activation kinetics of CaCCs. To induce a calcium-activated current, [Ca²⁺]_i was varied from 0 μM to 1 μM and 5 μM. Neither 0 μM nor 1 μM [Ca²⁺]_i activated non-stationary current; however, at 5 μM, the current was induced in both cell lines transfected with isoforms 1 and 2 (Fig. 2D). The time constant of activation of robust current at 5 μM of [Ca²⁺]_i was measured (Fig. 2E), and the current activation of cells expressing ANO2 isoform 1 was faster than that of isoform 2 (isoform 1: 3.53±0.44 s, n=10; isoform 2: 5.60±0.74 s, n=9; p =0.0339, unpaired t-test) (Fig. 2F). Application of various intracellular calcium concentrations ranging from 10 to 500 μM significantly affected current activation in both HEK293T cell lines transfected with isoforms (isoform 1: R²=0.2037, p=0.0024; isoform 2: R²=0.1809, p=0.0097 by simple linear regression) (Fig. 2G, 2H). When the amplitudes of the calcium-activated currents measured at each [Ca²⁺]_i were fitted by the Hill equation to calculate the dose–response relationship, EC₅₀ of isoform 1 was 7.1 μM, which was slightly smaller than that of isoform 2 (isoform 1: EC₅₀=7.1 μM, n_H=1.4; isoform 2, EC₅₀=12.0 μM, n_H=1.5) (Fig. 2I). Therefore, these results demonstrate that alternative splicing of ano2 exon 14 affected calcium-dependent activation and calcium sensitivity of ANO2 channels, and further suggest that splice variants of ANO2 would contribute to the functional diversity of ANO2 with different levels of Ca²⁺-dependent activation.

Isoforms 1 and 2 of ANO2 channels have similar voltage- and temperature-dependent kinetics

Although ANO2 is known as a calcium-activated channel, other factors, such as the voltage or electrical gradient, are also involved in controlling the current of ANO2. At low [Ca²⁺]_i, the voltage modulates the calcium-activated current, which drives ANO2 to make an outwardly rectified current at last [47]. In the brain, the outwardly rectified current of ANO2 is essential to modulate the spike adaptation or temporal summation of synaptic transmission by outward current [15, 20]. Thus, to compare the voltage dependency of isoforms 1 and 2, the rectification of isoforms 1 and 2 was examined after calcium-dependent activation. Steady-state currents of isoforms 1 and 2 in HEK293T cells were recorded using step-voltage protocols (-100 to 100 mV, increasing 20 mV per step) at intracellular Ca²⁺ levels of 0, 1, and 5 μM (Fig. 3A). Zero and 1 μM [Ca²⁺]_i did not induce a steady-state current; however, 5 μM [Ca²⁺]_i did (Fig. 3A). The current–voltage relationship of the steady-state current from both cell lines showed that isoforms 1 and 2 mediate the outwardly rectified current (Fig. 3B). The rectification calculated from two HEK293T cell lines expressing either isoform 1 or 2 were not significantly different (isoform 1: 4.7±1.1, n=12; isoform 2: 6.6±0.9, n=8; p=1.989, unpaired t-test) (Fig. 3C).

Neither ANO1 nor ANO2 has been reported to conduct current under a moderate heat range (23°C~42°C) in the absence of intracellular calcium, confirming that these are CaCCs [48]. In the presence of intracellular Ca²⁺, ANO1 conductance is regulated in a temperature-dependent manner within similar temperatures (25, 35, and 45°C) [48]. However, the temperature dependency of ANO2 and its isoforms has not yet been examined. To investigate whether Ca²⁺-dependent activation of ANO2 is modulated by temperature and if two isoforms have different temperature dependencies, the current–voltage (I~V) relationship of two ANO isoforms was measured by the current amplitude at the end of each step at three different intracellular [Ca²⁺]_i (0, 1, and 5 μM) and temperatures 26°C (room temperature) and 36°C (body temperature). Both isoforms 1 and 2 had outwardly rectified currents in the presence of 5 μM [Ca²⁺]_i (Fig. 3A), which is similar to the current–voltage relationship measured at both 26°C and 36°C (Fig. 3B, 3D). The ratio of current at 36°C and 26°C with 5 μM [Ca²⁺]_i was not significantly different between ANO2 isoforms 1 and 2 (isoform 1: 1.02±0.2, n=8; isoform 2: 1.31±0.4, n=9; p=0.5408, unpaired t-test) (Fig. 3E).

Collectively, the results from the steady-state currents that occur after calcium-dependent activation indicate that regulatory factors, such as voltage and temperature, did not differently affect calcium-activated currents in isoforms 1 and 2.

Molecular structure estimation of ANO2 isoforms predicts calcium-dependent channel kinetics

ANO2 and ANO1 share conserved calcium-binding sites (N650, E654, E702, E705, E734, and D738 in mice) in these transmembrane domains, and the structural change in ANO1 by the mutation of the calcium-binding site alters its calcium sensitivity [27, 28]. Moreover, the EF-hand-like region of ANO1, which is also well-conserved in ANO2, has been reported to be associated with calcium sensitivity in ANO1 [32].

Molecular modeling analyses were performed to determine whether the structural variance due to alternative splicing of ano2 exon 14, resulting in ANO2 isoforms 1 and 2, contributed to calcium-dependent activation. First, the protein structures in the absence and presence of exon 14, encoding the residues of ⁴⁷⁹ERSQ in the amino acid sequence (Fig. 4A, 4B), were predicted by AlphaFold2 [34]. Furthermore, the dimer structures of ANO2 isoforms 1 and 2 were generated using homology modeling. In the predicted protein structure with ⁴⁷⁹ERSQ, the dimer of the ANO2 isoform 1 has a helical structure in the first intracellular loop where the exon 14 sequence is located (Fig. 4C, 4D). However, in the predicted protein structure without exon 14, the dimer of ANO2 isoform 2 has a loop-like structure in the first intracellular loop where the exon 14 sequence is located (Fig. 4E, 4F). Additionally, the quantum-mechanical analysis was performed at FMO-MP2/6-31G**/PCM level in the pair interaction energies decomposition analysis (PIEDA) to identify significant interactions between residues. Stable interactions with energy levels greater than -3.0 kcal/mol and distance of up to 6 Å between two residues were consdiered as significant [39]. Isoform 1 had 14 intra-protein interactions in ⁴⁷⁹ERSQ residues, which plays a critical role in maintaining the helix structure (Supplementary Table 1). In contrast, isoform 2 had only four intra-protein interactions, which were insufficient to maintain the helix structure (Supplementary Table 2). Thus, the deletion of exon 14 transforms the helix structure of ANO2 isoform 1 into a loop-like structure of isoform 2, as shown in Fig. 4F. Remarkably, the orientation of the E477 residue in the cytoplasmic region of the ANO2 channel was altered by the alternative splicing of exon 14 in the modeling structure (Fig. 4D, 4F).

Next, to explore the effect of altered orientation of the E477 residue of intracellular regions of ANO2, cavity analysis was performed with ANO2 isoform dimer structures of isoforms 1 and 2 using CAVER 2.0 [37, 38] and then the calculated cavities were clustered into three parts: (1) calcium-binding, (2) remaining, and (3) entrance space. The calculated volumes of isoform 1 cavity were 674.8 ų in the calcium-binding space, 1754.4 ų in the remaining space, and 2012.5 ų in the entrance space (Fig. 5A). The first intracellular loop, containing ⁴⁷⁹ERSQ, was located near the entrance space of the cavities (Fig. 5C, 5E). However, the calculated volumes of isoform 2 cavities were 594.0 ų in the calcium-binding space, 1798.4 ų in the remaining space, and 1253.4 ų in the entrance space (Fig. 5B). Notably, the entrance space of the ANO2 isoform 2 cavity were much smaller than those of the ANO2 isoform 1 (Fig. 5A, 5B). The altered orientation of the E477 residue of isoform 2 from the cytoplasmic side seems to obstruct the entrance of the cavity, unlike the intracellular structure of isoform 1 (Fig. 5E, 5F), thereby hindering cytosolic calcium from entering the channel and then reaching the calcium-binding site inside the ANO2 channels (Fig. 6). These 3D molecular modeling analyses support the idea that slower activation and lower calcium sensitivity of ANO2 isoform 2 might be due to the smaller volume size of the channel entrance to the calcium-binding site when compared to those in the 3D modeling structure of isoform 1.

Here, two different splice variants of ANO2 were found to be expressed in the brain with distinct regional distributions: isoform 1 in the hippocampus and isoform 2 in the cerebral cortex and thalamus. The calcium-activated current of ANO2 isoform 1 had faster activation to intracellular calcium concentration, as compared with isoform 2. In addition, in molecular modeling analyses, the cavity entrance for calcium ions was demonstrated to be smaller in ANO2 isoform 2, through which Ca²⁺ enters the calcium-binding site from the cytoplasm than in isoform 1. Collectively, the two ANO2 isoforms found in the mouse brain have different calcium-dependent activation kinetics, with distinct structures in the intracellular monomeric cavity.

Ano2 mRNA, first discovered in the olfactory epithelium and retina, is expressed in various brain regions, including the cerebral cortex, thalamus, hippocampus, amygdala, and basal ganglia. CaCC, presumably ANO2, has been associated with various neuronal functions in the central nervous system [9, 10, 15]. More importantly, its functions in neurons vary in different brain regions. For instance, ANO2 regulates the action potential width in the hippocampus [15], amygdala [16], and cerebellum [17-19], whereas it accommodates spike frequency in the thalamus without affecting spike width [20]. However, the causes of various channel properties in these brain regions have rarely been described. ANO2 has various splicing isoforms, such as ano2 transcript exon 14 encoding segments c in the cytosolic N-terminus of ANO2 channel is lacking in the olfactory cilial membranes and is present in the retinal pigment epithelium [10, 24, 25]. The olfactory isoform of ANO2 channels lacking ⁴⁷⁹ERSQ residues is the major isoform of the brain, expressed in the cerebral cortex, hippocampus, cerebellum, and brain stem of the rat [25]. Consistent with a previous report, ano2 exon 14 was found to be lacking in the olfactory epithelium (Fig. 1C). However, our findings showed that the ANO2 isoform without exon 14 was further divided into two subtypes, isoform 2 (with ano2 exon 1a and 2, and without 1b) exclusively detected in the cerebral cortex and thalamus, and isoform 3 (without 1a and 2, and with 1b) expressed only in the olfactory epithelium (Fig. 1D). Moreover, ANO2 isoform 1 (with exons 1a, 2, and 14) is expressed in the hippocampus and retina. Notably, RT-PCR results of this study showed that ano2 exon 2 encoding the end region of the N-terminus of ANO2 channels is present in various brain regions along the retina, but not in the olfactory epithelium, whereas ano2 exon 1b has the opposite expression pattern to exon 2 (Fig. 1C). The function of splicing variant exon 2 in ANO2 channels needs to be further investigated. Accounted together with previous reports and our findings, multiple splice variants of ano2 transcripts would underpin the functional diversity of ANO2 channels in distinct regions of the brain.

ANO2 is activated by the intracellular calcium ion by the calcium- or calmodulin-binding sites, leading to the flux of chloride ions through the pore structure and the generation of an outwardly rectified chloride current as observed in the current–voltage relationship [10, 26]. Reportedly, the calcium-dependent activation occurred in over 1.5 μM intracellular calcium in ANO2-expressing cells [26]. In the present study, isoform 2, lacking exon 14, was found to have slower calcium-dependent activation than isoform 1, with exon 14, which was measured in ANO2-transfected stable HEK cells (Fig. 2). ANO2 has been reported to mediate spike frequency adaptation in the thalamus [20], in which isoform 2 is exclusively expressed (Fig. 1D). Spike adaptation occurs gradually in hundreds of milliseconds by continuous intrinsic potentiation, which induces a calcium rise by the voltage-gated calcium channel that activates ANO2. In isoform 2-transfected HEK293T cells, Ca²⁺-dependent activation was relatively slower (Fig. 2F), which indicates that ANO2 gradually slowed spike generation in thalamocortical neurons. In addition, in hippocampal neurons, blocking ANO2 channels reportedly increases the action potential width [15], which might be attributed to the faster Ca²⁺ activation of isoform 1 (Fig. 1D, 2F). Both isoforms showed similar calcium sensitivity, rectification index, and temperature-dependent current profile (Fig. 3C, 3E), which agreed well with previous reports [10, 26, 47, 48]. Altogether, these results suggest that splice variants of ANO2 channels can generate distinct channel kinetics, which would result in various regional properties of signal transmission of neurons in the brain.

Besides, the rundown effect of ANO2 has been observed in previous studies using inside-out patch clamp recordings, where the current was induced by calcium application at a high concentration for 10 minutes and the peak amplitude of the current was measured [10, 24-26]. Specifically, the rundown was observed in cells expressing retinal or olfactory isoforms, corresponding to isoform 1 and 2 in the present study [10, 24, 26, 47]. In whole-cell patch clamp recordings of ANO2-expressing cells, however, the rundown effect was not observed [26]. These findings implicate that exon 14 may not contribute to the rundown of the ANO2 channels in the excised patch clamp recording.

The increase in cytosolic calcium ion concentration activates the ANO2 channel by Ca²⁺ binding to the calcium-binding sites of ANO2, which is putatively located in the transmembrane region (α6, α7, and α8; Fig. 4) of the channel. In a previous report, mutation of the EF-hand-like region in the N-terminus of ANO1, which is well conserved in ANO2, altered calcium responses and voltage-dependent biophysical properties, and is located at the intracellular entrance towards the calcium-binding site in the modeled channel structure [29]. These findings suggest that the molecular architecture of these channels contributes to the modulation of calcium-dependent channel activity. Modeling data of this study demonstrated that the volume of the entrance cavity of ANO2 channels reaching the calcium-binding site differed between isoforms 1 and 2 (Fig. 5). The larger cavity entrance modeled in isoform 1 may allow intracellular Ca²⁺ to easily access the calcium-binding site located within the transmembrane domains of ANO2 channels, leading to faster calcium-dependent activation kinetics than that of isoform 2, as shown in Fig. 2F, and vice versa. In addition to the cavity volume of ANO2, the molecular architecture of ANO2, encoded by exon 14, was found to affect calcium-dependent channel activation. Exon 14 of the ano2 variant encodes for ⁴⁷⁹ERSQ residues of the channel, which is a part of the negatively charged acidic amino acid sequence, ⁴⁷⁵EEEEERSQ, in the first intracellular loop between transmembrane domains 1 and 2. Five consecutive E sequences have been reported to increase voltage dependency without altering the calcium sensitivity [47]. The RSQ sequence was found to contribute to slow kinetics in the bi-exponential gating of ANO2-transfected HEK cells [49]. In particular, the ERSQ sequence was reported to produce a long-lasting calcium-activated current in ANO2-transfected HEK cells [25]. In our modeled structure of ANO2, the altered orientation of E477 in the cytosol resulting from the absence of the ERSQ sequence (Fig. 4D, 4F) is likely to reduce the volume of cavity entrance in isoform 2 (Fig. 5, 6), which can affect the kinetics of calcium-dependent activation. Therefore, alternative splicing of ano2 exons in the first intracellular loop of ANO2 can modify the molecular structure of the channel, leading to the modulation of calcium-dependent channel activity.

These findings suggest that alternative splicing variants of ANO2 can alter the molecular structure of the channels depending on the regional properties of the brain, thus varying the activation kinetics of the channel, leading to diverse functions in the brain.

This work was supported by the Samsung Science and Technology Foundation under Project Number SSTF-BA2201-12 and by the National Research Foundation of Korea (NRF) under Project Number 2021R1A2C3007164.