Two helices in the third intracellular loop determine anoctamin 1 (TMEM16A) activation by calcium
- 1.7k Downloads
Anoctamin 1 (ANO1)/TMEM16A is a Cl− channel activated by intracellular Ca2+ mediating numerous physiological functions. However, little is known of the ANO1 activation mechanism by Ca2+. Here, we demonstrate that two helices, “reference” and “Ca2+ sensor” helices in the third intracellular loop face each other with opposite charges. The two helices interact directly in a Ca2+-dependent manner. Positively and negatively charged residues in the two helices are essential for Ca2+-dependent activation because neutralization of these charges change the Ca2+ sensitivity. We now predict that the Ca2+ sensor helix attaches to the reference helix in the resting state, and as intracellular Ca2+ rises, Ca2+ acts on the sensor helix, which repels it from the reference helix. This Ca2+-dependent push-pull conformational change would be a key electromechanical movement for gating the ANO1 channel. Because chemical activation of ANO1 is viewed as an alternative means of rescuing cystic fibrosis, understanding its gating mechanism would be useful in developing novel treatments for cystic fibrosis.
KeywordsAnoctamin 1 Anoctamin 2 Calcium Activation Helix Structure
Ca2+-activated Cl− channels (CaCCs) mediate trans-epithelial fluid movements and thereby enable secretions in salivary glands, pancreas, intestines, and airways [12, 8, 6]; regulate vascular smooth muscle tone and cardiac myocyte excitability [50, 23, 24]; modulate neuronal cell excitability; and amplify sensory signals in retinal or olfactory sensory neurons [8, 20, 18]. Because of this ability to regulate epithelial secretion, chemical activation of CaCCs is viewed as an alternative means of rescuing disabled fluid movement in the epithelia of cystic fibrosis patients [26, 25, 42].
Anoctamin 1 (ANO1; also known as TMEM16A) is a candidate for Ca2+-activated chloride channels and has biophysical and pharmacological properties similar to those of identified CaCCs [47, 3, 31]. As expected for a CaCC candidate gene, ANO1 is expressed in the salivary gland for salivation [47, 30], in the airway epithelium for controlling mucin secretion , in the blood vessels for vascular tone , and in the pacemaker cells for gastrointestinal smooth muscle contraction . Additionally, ANO1 is known as a heat sensor that mediates thermal pain in sensory neurons . ANO2 is expressed in olfactory sensory neurons, suggesting its role in olfactory signal amplification [2, 35]. ANO2 is also expressed in the hippocampus and controls synaptic excitability . ANO5 is linked to a rare skeletal syndrome, gnathodiaphyseal dysplasia [21, 41, 40]. ANO6 is also known to be a Ca2+-activated Cl− channel associated with scramblase activity essential for blood clotting. Its channelopathy is associated with Scott syndrome, a rare disease of hemorrhage [46, 36, 37, 33, 11].
Among the 10 isoforms of anoctamin channel family, ANO1 and ANO2 are activated by physiological concentrations of Ca2+ at the resting membrane potential [47, 29]. Despite these multifunctional roles of the anoctamin channel family, the Ca2+ activation mechanisms remain elusive. Recently, a putative Ca2+-acting site was suggested . Yu and colleagues proposed a revised topology of ANO1 using fluorescent tags at different key positions that differs from the conventional topology of ANO1 . Furthermore, two Glu residues (702-EYME-705) in the third intracellular loop were suggested to be essential for the Ca2+ activation. This has been further confirmed by Scudieri and colleagues that a chimera of ANO1 replacing the third intracellular loop with that of ANO2 lowers the Ca2+ sensitivity . Although these reports suggest that a region in the third intracellular loop is important for the ANO1 or ANO2 activation by Ca2+, molecular insights for the activation is largely unknown. Thus, this study pursues to determine how Ca2+ activates ANO1.
The third intracellular loop is essential for Ca2+-induced ANO1 activation
In order to find potential Ca2+-binding sites in the ANO1 sequence, we first measured the Ca2+ sensitivities of ANO1 and ANO2. To this end, different concentrations of free Ca2+ were applied to the bath of inside-out membrane patches excised from human embryonic kidney (HEK) 293T cells that were transfected with Ano1 and Ano2 tagged on the C-terminus with enhanced green fluorescence protein (EGFP). Among splice variants of ANO1, we used an a,c-splice variant that has 116 residues in the N-terminus and the 448-EAVK-451 insert .
Then, a mutagenesis study was performed to determine a potential Ca2+ activation site. ANO1 contains three putative sites that have multiple acidic amino acids. The first site is the Glu- and Asp-rich (E/D rich) region spanning aa121 ~ 147 in the N-terminus. A deletion mutant (aa121 ~ 148) in the E/D rich region was not expressed in the membrane and no currents were observed at a maximal Ca2+. However, a mutant having nine Asp or Glu residues replaced with Asn or Gln in the E/D rich region (120-HQNNKRFRRQQYQGNLLEAGLQLQNDQDT-148, ED9) elicited currents with EC50 value of 2.4 μM comparable to wild type (Fig. 1c). The second site (285-DGDYEGDNVEFND-297) in the N-terminus shares high homology with the Ca2+-binding signature sequence of the EF-hand (DxDxDGxxxxxE) of calmodulin [45, 10]. Replacing all four Asp or Glu residues with Ala shifted the EC50 for Ca2+ to 17.8 μM (n = 6). The third potential Ca2+-binding site is located in the first intracellular loop (ICL1) between transmembrane domain 2 and 3 (TM2 and 3). This site has five consecutive Glu residues (444-EEEEEAVKDHPRAE-457), which are loosely aligned with the Ca2+-bowl region of large conductance Ca2+-activated K+ channel [1, 49]. A deletion mutant (444-EEEEE-448) in this region had an EC50 of 2.1 μM (n = 8)  (Fig. 1b, c).
Recently, a revised topology of ANO1 has been proposed in which a loop that was previously thought to be extracellular forms an intracellular loop (Fig. 1a) . A deletion mutant (aa 653–711) of the third intracellular loop (ICL3) in the newly proposed topology was extremely insensitive to Ca2+. The dose-response curve was shifted dramatically to the right (EC50 = 2.5 mM, n = 10) and currents rarely showed a saturation up to 10 mM Ca2+ (Fig. 1c). Thus, the ICL3 region appears important for activation of ANO1 by Ca2+.
Structural prediction of the ICL3 region
Gating mechanism by Ca2+
Because the two parallel helices in the ICL3 region of ANO1 have amino acids with opposite charges, these opposing charges may engage in the activation of ANO1 by Ca2+. We predicted that, in the closed state, the two helices hold together primarily due to the ionic interactions between positive and negative charges in the two helices (Fig. 7f). However, as intracellular Ca2+ increases, Ca2+ binds and covers the Ca2+ sensor helix around the two Glu residues. These positive charges from the Ca2+ sensor helix may break the ionic interaction between the two helices and repel the reference helix away from the Ca2+ sensor helix (Fig. 7f). When the Ca2+ ions are removed, the Ca2+ sensor helix again moves closer to the reference helix. This conformational change of the two helices would lead to a conformational change in the channel gate.
Ca2+-dependent interaction of the reference and Ca2+-sensor helices
We then applied the reference helix and Ca2+ sensor helix peptides to isolated inside-out patches of HEK 293T cells transfected with ANO1, to see if the peptides antagonized Ca2+ in activating ANO1. Indeed, application of 4-μM reference helix peptide or 10-μM Ca2+ sensor helix peptide (700-TPEYMEMIIQFGF-712) markedly blocked Ca2+-induced ANO1 currents (Fig. 4c, d). In contrast, the Ala-mutant reference helix peptide failed to block the Ca2+-induced ANO1 currents.
Mutations in the two helices of ICL3 alter the Ca2+ sensitivity
Eact acts on the reference helix peptide
We then applied the two helix peptides to see if these peptides compete with Eact in activating ANO1. When 10 μM Ca2+-sensor helix peptide was applied together with 1 μM Eact to inside-out patches from ANO1-transfected cells, the peptide failed to inhibit Eact-activated currents (Fig. 6b, c). In contrast, when 4-μM reference helix peptide was applied together with Eact, the peptide significantly inhibited the Eact-activated ANO1 currents (p < 0.05, one-way ANOVA, Newman-Keuls post hoc test, n = 5 ~ 14). However, the Ala-substituted reference helix peptide failed to inhibit the Eact-activated ANO1 currents (Fig. 6c). These results clearly suggest that Eact acts on the reference helix in the ICL3 region.
The two helices in the ICL3 region are dispensable for voltage- and heat-induced ANO1 activation
Another activation stimulus for ANO1 is heat. Bath temperature greater than 44 °C is sufficient to open the ANO1 channel . Because the activation by Ca2+ and heat are synergistic, we tested whether the heat-induced ANO1 activation is also mediated by the ICL3 region. To eliminate activation by Ca2+, we removed Ca2+ from the pipette solution (0 Ca2+ with 5 mM ethylene glycol tetraacetic acid (EGTA)) in whole-cell recording. Surprisingly, a mutant (ΔICL3) which hardly responded to sub-millimolar Ca2+ concentration elicited currents when the temperature of bath solution was raised over 44 °C (Fig. 7d, e). Thus, these results suggest that the activation mechanism of ANO1 by heat is different from that by Ca2+.
With homology modeling, we predict that the ICL3 region in the revised topology of ANO1 has a unique structure composed of two parallel helices that interact with each other in a Ca2+-dependent manner and cause rightward shifts in the Ca2+ sensitivity upon mutation. Furthermore, Eact, a synthetic agonist of ANO1, also acts on the ICL3 region. However, the two helices apparently do not control voltage-induced activation because mutagenesis did not change the sensitivity to voltage. Furthermore, the ICL3 region does not account for heat-induced activation. Although this study is based on homology modeling, present experimental data including surface plasmon resonance (SPR) assay, circular dichroism analysis, as well as mutational study along with electrophysiological experiments support the structure model.
According to the model of ICL3 of ANO1, several structural elements appear important for the activation by Ca2+. The two helices are oppositely charged. The role of positive charges in the reference helix appears to hold the Ca2+ sensor helix in the resting state. We imagine that the reference helix pushes the Ca2+-sensor helix away when Ca2+ ions attach to it (Fig. 7f). Thus, removal of these positive charges in the reference helix decreases the direct binding and sensitivity to Ca2+ concentrations. However, because the removal of positive charges in the reference helix showed two-order differences in Ca2+ sensitivity that is smaller than the four-order change with the mutant in the Ca2+ sensor helix (Fig. 5a), factors other than the ionic interaction may also contribute to the movement of two helices by Ca2+. In line with this, positively charged amine residues of Lys and Arg in the reference helix do not align along one side facing the Ca2+ sensor helix. Instead, these positive charges distribute in many directions in a circular fashion (Fig. 3). In addition, only ANO1 and ANO2 are sensitive to physiological range of Ca2+ at the resting membrane potential and have the helix-loop-helix structure in the ICL3 region. It is uncertain whether other ANO family members have this structural element. It is conceivable that these other ANO channels may have different mechanisms for channel gating if they are channels. However, as ANO6 has been suggested as a scramblase [46, 36], other ANO family members may function other than channels.
Many studies tried to define the Ca2+ activation site in ANO1. Initially, the Ca2+ bowl-like region in the ICL1 was studied for the Ca2+ sensitivity . However, the deletion of the region failed to show a shift in the Ca2+ sensitivity. The lack of change in the Ca2+ sensitivity of mutants in this region of ANO1 and ANO2 was also observed by others . Furthermore, Tien and colleagues observed that an “a” variant of ANO1 lacking the “EAVK” residues retains the EC50 of Ca2+ less than 1 μM, comparable to that of the “a,c” variant . Similarly, we also failed to see a discernible change in EC50 after the deletion of 444-EEEEE-448 (Fig. 1b, c). In contrast, Xiao and Cui reported that the deletion of 444-EEEEEAVK-451 shifted ~40-fold increase in EC50 . Yu and colleagues found the acidic residues, E702 and E705, in the ICL3 region of the revised topology that were critical for the Ca2+ sensitivity of ANO1 . Scudieri and colleagues also confirmed that this region is sensitive to the Ca2+-dependent activation . The present study also confirmed that this region is important for the ANO1 activation by Ca2+. Recently, Tien and colleagues also suggest that the ICL3 region is an important site for the Ca2+ sensitivity . They extended mutation to E734 (equivalent to E730 in “a” variant) and D738 (equivalent to D734 in “a” variant) residues beyond the E702 and E705 residues and found that the E734 and D738 are involved in the Ca2+-dependent gating, because mutations in these residues made a large shift in EC50. Thus, Tien and colleagues proposed that the Ca2+ caging with the four Glu and Asp residues is essential for the ANO1 activation. It is not clear how Ca2+ binding to this site gates ANO1. It is conceivable that these additional Glu734 and Asp738 residues may affect the interaction between Ca2+ and the reference Ca2+ sensor helices, because the two Glu and Asp residues are close to the Ca2+ sensor helix. However, to identify the interactions between the two helices and the E734 and D738 residues requires further study.
In the present study, high Ca2+ evoked currents in ΔICL3 mutant-expressing HEK cells. This remnant current is probably caused by endogenous ANO isoforms present in native HEK cells. HEK cells are known to have ANO6 and 8 . Thus, the current activated by high Ca2+ in ΔICL3 mutant-expressing cells would be currents of endogenous ANO6 or other isoforms in HEK cells. Another possibility is that there are multiple sites other than the ICL3 region in ANO1 that controls the Ca2+-dependent activation. So, the deletion of the ICL3 region may not remove all the currents activated by Ca2+.
In summary, this study demonstrates the structural basis of channel gating by Ca2+ and voltage in ANO1. Although electrophysiological, protein chemistry, and structure modeling have predicted that opposite charges between the two helices in the ICL3 play critical roles in channel gating, the precise mechanism underlying the activation requires crystal structure analysis. Because ANO1 and other family members mediate numerous physiologic functions, elucidation of the activation mechanism by Ca2+ is essential for a detailed understanding of ANO-related pathophysiology. ANO1 agonists are now considered to be useful in the management of cystic fibrosis. Understanding structural and molecular mechanisms for ANO1 activation is useful in designing anti-cystic fibrosis drugs.
Materials and methods
Mutagenesis and gene expression
All mutants were generated from the wild-type construct, mouse ANO1 (pEGFP-N1-mANO1). Amino acid substitution or deletion mutants were prepared using a site-directed mutagenesis kit (Muta-direct, iNtRON Biotech) or by using the overlap PCR method. Mutations in all mutants were confirmed by sequencing whole nucleotide sequences. HEK 293T cells were transfected with 1 μg of pEGFP-N1-mANO1 or pEGFP-N1-mutants, 0.05 μg pEGFP-N1, and FuGENE (Roche Diagnostics, Penzberg, Germany) in 35-mm dishes. Transfected cells were incubated in DMEM supplemented with 10 % fetal bovine serum (GIBCO) and penicillin-streptomycin at 37 °C in a 5 % CO2 incubator. Cells were used 1 or 2 days after transfection.
Borosilicate glass pipettes (World Precision Instruments, Sarasota, FL) with tip resistances of ~2 Mohms were used to form gigaseals on HEK cells. Inside-out membrane patches were excised by pulling the pipette away from the cell. Currents were recorded with a patch-clamp amplifier (Axopatch 200B, Molecular Probes) with a 1-KHz filter. Data from the amplifier were digitized with Digidata 1440A (Molecular Probes) and stored on a computer.
The control pipette solution contained 140 mM N-methyl d-glucamine, 2 mM MgCl2, and 10 mM HEPES adjusted to pH 7.2 with HCl. For Ca2+-free solution, 10 mM ethylene glycol tetraacetic acid (EGTA) was added to the control solution. The bath solution contained (in mM) 140 N-methyl d-glucamine, 2 MgCl2, 10 HEPES, 10 chelator (EGTA, N(2-hydroxyethyl) ethylenediamine triacetic acid (HEDTA), or nitrilotriacetic acid (NTA)), and a calculated amount of CaCl2 (adjusted with HCl to pH 7.2). EGTA was used for 0.1 ~ 1.0 μM free Ca2+, whereas HEDTA and NTA were used for 3.0 ~ 30 and 100 ~ 1,000 μM free Ca2+, respectively . No chelator was added to solutions with free Ca2+ greater than 1 mM. To calculate free Ca2+ in pipette solutions, the WEBMAXC program was used (http://www.stanford.edu/~cpatton/webmaxc/webmaxcS.htm). Free Ca2+ less than 30 μM in each solution was monitored with Fura-2 fluorometry .
Surface plasmon resonance measurement
A Biacore 3000 was used for kinetic surface plasmon resonance measurement as previously described . Briefly, a biotinylated peptide (692-NLEPFAGLTPEYMEM-706) was immobilized on a SA chip or custom streptavidin-coated CM5 sensor chip (GE Healthcare), with one flow cell with biotin alone as a reference cell. For the binding assay, HEPES-buffered saline containing 0.005 % of Tween 20 was used as a running buffer. Various concentrations of the reference helix peptide (aa 651–671) were injected for 120 s and dissociated for 240 s at a flow rate of 20 μL/min. Binding curves were analyzed using BIAevaluation 3.1.
Structure homology modeling
The amino acid residues of third intracellular loop (residues 603–663) in ANO2 were selected as the target sequence for homology modeling. Automated homology modeling was implemented by MODWEB (http://salilab.org/modweb), a web server for automated comparative protein structure modeling.
This research was supported by a grant from the National Research Foundation of Korea (no. 20110018358), a grant from BK21+ program of Ministry of Education of Korea, and NIH grants GM60448 and EY114852 (HCH)
JL recorded channel currents and carried out structural modeling; JJ carried out the Ca2+ overlay and SPR assays; JW, BL, YJ, and YDY worked on mutagenesis; MHT, HC, DJY, and HwC carried out current recordings; SHP and BWH worked on structural modeling; SH and JY worked on SPR assay; HCH planned and wrote the manuscript; and OU planned, supervised experiments, and wrote the manuscript.
Conflict of interest
The authors declare that they have no conflict of interest.
- 13.Huang F, Zhang H, Wu M, Yang H, Kudo M, Peters CJ, Woodruff PG, Solberg OD, Donne ML, Huang X, Sheppard D, Fahy JV, Wolters PJ, Hogan BL, Finkbeiner WE, Li M, Jan YN, Jan LY, Rock JR (2012) Calcium-activated chloride channel TMEM16A modulates mucin secretion and airway smooth muscle contraction. Proc Natl Acad Sci U S A 109(40):16354PubMedCentralPubMedCrossRefGoogle Scholar
- 15.Hwang SJ, Blair PJ, Britton FC, O'Driscoll KE, Hennig G, Bayguinov YR, Rock JR, Harfe BD, Sanders KM, Ward SM (2009) Expression of anoctamin 1/TMEM16A by interstitial cells of Cajal is fundamental for slow wave activity in gastrointestinal muscles. J Physiol 587(Pt 20):4887–4904PubMedCentralPubMedCrossRefGoogle Scholar
- 28.Pieper U, Webb BM, Dong GQ, Schneidman-Duhovny D, Fan H, Kim SJ, Khuri N, Spill YG, Weinkam P, Hammel M, Tainer JA, Nilges M, Sali A (2014) ModBase, a database of annotated comparative protein structure models and associated resources. Nucleic Acids Res 42(Database issue):D336–346PubMedCentralPubMedCrossRefGoogle Scholar
- 30.Romanenko VG, Catalan MA, Brown DA, Putzier I, Hartzell HC, Marmorstein AD, Gonzalez-Begne M, Rock JR, Harfe BD, Melvin JE (2010) Tmem16A encodes the Ca2+-activated Cl- channel in mouse submandibular salivary gland acinar cells. J Biol Chem 285(17):12990–13001PubMedCentralPubMedCrossRefGoogle Scholar
- 38.Tien J, Peters CJ, Wong XM, Cheng T, Jan YN, Jan LY, Yang H (2014) A comprehensive search for calcium binding sites critical for TMEM16A calcium-activated chloride channel activity. Elife:e02772Google Scholar
- 41.Tsutsumi S, Kamata N, Vokes TJ, Maruoka Y, Nakakuki K, Enomoto S, Omura K, Amagasa T, Nagayama M, Saito-Ohara F, Inazawa J, Moritani M, Yamaoka T, Inoue H, Itakura M (2004) The novel gene encoding a putative transmembrane protein is mutated in gnathodiaphyseal dysplasia (GDD). Am J Hum Genet 74(6):1255–1261PubMedCentralPubMedCrossRefGoogle Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.