Exp Neurobiol 2023; 32(3): 133-146
Published online June 30, 2023
© The Korean Society for Brain and Neural Sciences
Dongsu Lee1†, Hocheol Lim1,2†, Jungryun Lee1, Go Eun Ha1, Kyoung Tai No1,2 and Eunji Cheong1*
1Department of Biotechnology, College of Life Science and Biotechnology, Yonsei University, Seoul 03722, 2The Interdisciplinary Graduate Program in Integrative Biotechnology & Translational Medicine, Yonsei University, Incheon 21983, Korea
Correspondence to: *To whom correspondence should be addressed.
TEL: 82-2-2123-5885, FAX: 82-2-2123-8284
†These authors contributed equally to this article.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/4.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.
Anoctamin 2 (ANO2 or TMEM16B), a calcium-activated chloride channel (CaCC), performs diverse roles in neurons throughout the central nervous system. In hippocampal neurons, ANO2 narrows action potential width and reduces postsynaptic depolarization with high sensitivity to Ca2+ at relatively fast kinetics. In other brain regions, including the thalamus, ANO2 mediates activity-dependent spike frequency adaptations with low sensitivity to Ca2+ at relatively slow kinetics. How this same channel can respond to a wide range of Ca2+ levels remains unclear. We hypothesized that splice variants of ANO2 may contribute to its distinct Ca2+ sensitivity, and thus its diverse neuronal functions. We identified two ANO2 isoforms expressed in mouse brains and examined their electrophysiological properties: isoform 1 (encoded by splice variants with exons 1a, 2, 4, and 14) was expressed in the hippocampus, while isoform 2 (encoded by splice variants with exons 1a, 2, and 4) was broadly expressed throughout the brain, including in the cortex and thalamus, and had a slower calcium-dependent activation current than isoform 1. Computational modeling revealed that the secondary structure of the first intracellular loop of isoform 1 forms an entrance cavity to the calcium-binding site from the cytosol that is relatively larger than that in isoform 2. This difference provides structural evidence that isoform 2 is involved in accommodating spike frequency, while isoform 1 is involved in shaping the duration of an action potential and decreasing postsynaptic depolarization. Our study highlights the roles and molecular mechanisms of specific ANO2 splice variants in modulating neuronal functions.
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 . 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
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 . 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) . 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 , is involved in the modulation of calcium sensitivity in both ANO1 and ANO2 .
Alternative splicing of exon 14 of ANO2 channels has been reported to have regional specificity in different tissues, including the brain .
To date, alternative splicing of
Male C57BL/6J mice were maintained under a 12:12-light–dark cycle (lights on at 7:00 A.M.) and had
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:
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’;
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
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:
where Erev 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.
The calculated conductance versus calcium was fitted into Hill’s equation:
where G is the conductance, Gmax is the maximal conductance, K1/2 is the half-maximal [Ca2+]intra (also called EC50), and nH 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.
The ANO2 sequence of
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 . 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 .
All FMO calculations were performed using the version of February 14, 2018, GAMESS . 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)  and polarizable continuum model (PCM)  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 .
The two-body FMO calculation consists of fragmentation, fragment self-consistent field calculation, fragment pair self-consistent field calculation, and total property evaluation . 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 (∆Ees), exchange-repulsion (∆Eex), charge-transfer, and mixed (∆Ect+mix), dispersion (∆Edi), and solvation energy terms (∆Gsol).
The electrostatic term (∆Ees) arises from the Coulomb interaction between polarized charge distributions of fragments, the exchange-repulsion term (∆Eex) arises from the interaction between fragments closely situated and is related to the overlap of two occupied orbitals, the charge-transform and mixed term (∆Ect+mix) is derived from the interaction between the occupied orbitals of a donor and unoccupied orbitals of an acceptor, the dispersion term (∆Edi) is obtained from instantaneous dipole moments of fragments and is hydrophobic, and the solvation term (∆Gsol) arises from solute-solvent interactions with implicit solvation.
ANO2 has been reported to have several alternative splice variants [10, 24, 25].
To determine the distribution of
To understand whether the identified alternative splicing of
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
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, nH=1.4; isoform 2, EC50=12.0 μM, nH=1.5) (Fig. 2I). Therefore, these results demonstrate that alternative splicing of
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 . 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 . In the presence of intracellular Ca2+, ANO1 conductance is regulated in a temperature-dependent manner within similar temperatures (25, 35, and 45°C) . 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.
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 .
Molecular modeling analyses were performed to determine whether the structural variance due to alternative splicing of
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 Å3 in the calcium-binding space, 1754.4 Å3 in the remaining space, and 2012.5 Å3 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 Å3 in the calcium-binding space, 1798.4 Å3 in the remaining space, and 1253.4 Å3 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 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 . 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 , 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 , 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 . 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 . 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
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.