Identification of a druggable pocket of the calcium-activated chloride channel TMEM16A in its open state

The calcium-activated chloride channel TMEM16A is a potential drug target to treat hypertension, secretory diarrhea, and several cancers. However, all reported TMEM16A structures are either closed or desensitized, and direct inhibition of the open state by drug molecules lacks a reliable structural basis. Therefore, revealing the druggable pocket of TMEM16A exposed in the open state is important for understanding protein–ligand interactions and facilitating rational drug design. Here, we reconstructed the calcium-activated open conformation of TMEM16A using an enhanced sampling algorithm and segmental modeling. Furthermore, we identified an open-state druggable pocket and screened a potent TMEM16A inhibitor, etoposide, which is a derivative of a traditional herbal monomer. Molecular simulations and site-directed mutagenesis showed that etoposide binds to the open state of TMEM16A, thereby blocking the ion conductance pore of the channel. Finally, we demonstrated that etoposide can target TMEM16A to inhibit the proliferation of prostate cancer PC-3 cells. Together, these findings provide a deep understanding of the TMEM16A open state at an atomic level and identify pockets for the design of novel inhibitors with broad applications in chloride channel biology, biophysics, and medicinal chemistry. The calcium-activated chloride channel TMEM16A is a potential drug target to treat hypertension, secretory diarrhea, and several cancers. However, all reported TMEM16A structures are either closed or desensitized, and direct inhibition of the open state by drug molecules lacks a reliable structural basis. Therefore, revealing the druggable pocket of TMEM16A exposed in the open state is important for understanding protein–ligand interactions and facilitating rational drug design. Here, we reconstructed the calcium-activated open conformation of TMEM16A using an enhanced sampling algorithm and segmental modeling. Furthermore, we identified an open-state druggable pocket and screened a potent TMEM16A inhibitor, etoposide, which is a derivative of a traditional herbal monomer. Molecular simulations and site-directed mutagenesis showed that etoposide binds to the open state of TMEM16A, thereby blocking the ion conductance pore of the channel. Finally, we demonstrated that etoposide can target TMEM16A to inhibit the proliferation of prostate cancer PC-3 cells. Together, these findings provide a deep understanding of the TMEM16A open state at an atomic level and identify pockets for the design of novel inhibitors with broad applications in chloride channel biology, biophysics, and medicinal chemistry. TMEM16A (calcium-activated chloride channel, CaCC) is an anion channel broadly expressed in a variety of cells and is activated by the rise of intracellular Ca2+ (1Caputo A. Caci E. Ferrera L. Pedemonte N. Barsanti C. Sondo E. et al.TMEM16A, a membrane protein associated with calcium-dependent chloride channel activity.Science. 2008; 322: 590-594Crossref PubMed Scopus (1031) Google Scholar, 2Yang Y.D. Cho H. Koo J.Y. Tak M.H. Cho Y. Shim W.S. et al.TMEM16A confers receptor-activated calcium-dependent chloride conductance.Nature. 2008; 455: 1210-1215Crossref PubMed Scopus (1052) Google Scholar, 3Schroeder B.C. Cheng T. Jan Y.N. Jan L.Y. Expression cloning of TMEM16A as a calcium-activated chloride channel subunit.Cell. 2008; 134: 1019-1029Abstract Full Text Full Text PDF PubMed Scopus (944) Google Scholar). TMEM16A plays important roles in a variety of physiological functions, including the transmission of nerve signals, the control of smooth muscle contraction, and the regulation of epithelial cell secretion (4Ji Q. Guo S. Wang X. Pang C. Zhan Y. Chen Y. et al.Recent advances in TMEM16A: structure, function, and disease.J. Cell Physiol. 2019; 234: 7856-7873Crossref PubMed Scopus (79) Google Scholar). TMEM16A has nine related family members in eukaryotes, of which the functionally clearer ones are TMEM16A and B that form CaCCs and TMEM16D-F and J-K that function as lipid scramblases (5Oh U. Jung J. Cellular functions of TMEM16/anoctamin.Pflugers Arch. 2016; 468: 443-453Crossref PubMed Scopus (121) Google Scholar). The structures of TMEM16A and its family members have been recently determined with the calcium-activated gating mechanism elucidated (6Shi S. Pang C. Guo S. Chen Y. Ma B. Qu C. et al.Recent progress in structural studies on TMEM16A channel.Comput. Struct. Biotechnol. J. 2020; 18: 714-722Abstract Full Text Full Text PDF PubMed Scopus (18) Google Scholar). High-throughput drug screening and computer-aided drug design have identified an array of TMEM16A activators and potentiators, among which ETX001, a potentiator of TMEM16A, has entered clinical trials as therapeutic agent for cystic fibrosis (7Al-Hosni R. Ilkan Z. Agostinelli E. Tammaro P. The pharmacology of the TMEM16A channel: therapeutic opportunities.Trends Pharmacol. Sci. 2022; 43: 712-725Abstract Full Text Full Text PDF PubMed Scopus (4) Google Scholar). Meanwhile, a large number of TMEM16A inhibitors have been identified, including natural products Cepharanthine (8Zhang X. Zhang G.H. Zhao Z.J. Xiu R.L. Jia J. Chen P.P. et al.Cepharanthine, a novel selective ANO1 inhibitor with potential for lung adenocarcinoma therapy.Biochim. Biophys. Acta Mol. Cell Res. 2021; 1868119132Crossref Scopus (17) Google Scholar), Evodiamine (9Zhao Z.J. Xue Y.R. Zhang G.H. Jia J. Xiu R.L. Jia Y.G. et al.Identification of evodiamine and rutecarpine as novel TMEM16A inhibitors and their inhibitory effects on peristalsis in isolated Guinea-pig ileum.Eur. J. Pharmacol. 2021; 908174340Crossref Scopus (6) Google Scholar), Honokiol (10Wang T. Wang H. Yang F. Gao K. Luo S. Bai L. et al.Honokiol inhibits proliferation of colorectal cancer cells by targeting anoctamin 1/TMEM16A Ca(2+) -activated Cl(-) channels.Br. J. Pharmacol. 2021; 178: 4137-4154Crossref PubMed Scopus (10) Google Scholar), and Luteolin (11Zhang X. Li H. Zhang H. Liu Y. Huo L. Jia Z. et al.Inhibition of transmembrane member 16A calcium-activated chloride channels by natural flavonoids contributes to flavonoid anticancer effects.Br. J. Pharmacol. 2017; 174: 2334-2345Crossref PubMed Scopus (46) Google Scholar) and synthetic compounds Ani9 (12Seo Y. Lee H.K. Park J. Jeon D.K. Jo S. Jo M. et al.Ani9, a novel potent small-molecule ANO1 inhibitor with negligible effect on ANO2.PLoS One. 2016; 11e0155771Crossref Scopus (119) Google Scholar), CaCCinh-A01 (13Bill A. Hall M.L. Borawski J. Hodgson C. Jenkins J. Piechon P. et al.Small molecule-facilitated degradation of ANO1 protein: a new targeting approach for anticancer therapeutics.J. Biol. Chem. 2014; 289: 11029-11041Abstract Full Text Full Text PDF PubMed Scopus (66) Google Scholar), T16Ainh-A01 (14Xiang-qin T. Ke-tao M. Xian-wei W. Yang W. Zhi-kun G. Jun-qiang S. Effects of the calcium-activated chloride channel inhibitors T16Ainh-A01 and CaCCinh-A01 on cardiac fibroblast function.Cell Physiol. Biochem. 2018; 49: 706-716Crossref PubMed Scopus (18) Google Scholar), Monna (15Boedtkjer D.M. Kim S. Jensen A.B. Matchkov V.M. Andersson K.E. New selective inhibitors of calcium-activated chloride channels - T16A(inh) -A01, CaCC(inh) -A01 and MONNA - what do they inhibit?.Br. J. Pharmacol. 2015; 172: 4158-4172Crossref PubMed Scopus (61) Google Scholar), and 4-arylthiophene-3-carboxylic acid (16Wang Y. Gao J. Zhao S. Song Y. Huang H. Zhu G. et al.Discovery of 4-arylthiophene-3-carboxylic acid as inhibitor of ANO1 and its effect as analgesic agent.Acta Pharm. Sin. B. 2021; 11: 1947-1964Crossref PubMed Scopus (12) Google Scholar). These inhibitors have shown potential therapeutic effects in the laboratory against TMEM16A-related diseases such as hypertension, cancer, and secretory diarrhea (7Al-Hosni R. Ilkan Z. Agostinelli E. Tammaro P. The pharmacology of the TMEM16A channel: therapeutic opportunities.Trends Pharmacol. Sci. 2022; 43: 712-725Abstract Full Text Full Text PDF PubMed Scopus (4) Google Scholar). However, TMEM16A inhibitors have not progressed as rapidly as potentiators, without any inhibitors been brought to market. The relationship between TMEM16A and cancer and the role TMEM16A plays in malignant development have drawn much attention from the pharmaceutical community (4Ji Q. Guo S. Wang X. Pang C. Zhan Y. Chen Y. et al.Recent advances in TMEM16A: structure, function, and disease.J. Cell Physiol. 2019; 234: 7856-7873Crossref PubMed Scopus (79) Google Scholar). On the one hand, TMEM16A has been shown to be highly expressed in many cancer cells and upregulation of TMEM16A is thought to be associated with reduced overall survival in patients with breast (17Luo S.Y. Wang H. Bai L.C.A. Chen Y.W. Chen S. Gao K. et al.Activation of TMEM16A Ca2+-activated Cl- channels by ROCK1/moesin promotes breast cancer metastasis.J. Adv. Res. 2021; 33: 253-264Crossref PubMed Scopus (16) Google Scholar), pancreatic (18Kim Y. Park H.S. The function and mechanism of TMEM16A/ANO1 in pancreatic cancer.Biophys. J. 2017; 112: 548aAbstract Full Text Full Text PDF Google Scholar), or gastric cancer (19Liu F. Cao Q.H. Lu D.J. Luo B. Lu X.F. Luo R.C. et al.TMEM16A overexpression contributes to tumor invasion and poor prognosis of human gastric cancer through TGF-beta signaling.Oncotarget. 2015; 6: 11585-11599Crossref PubMed Scopus (73) Google Scholar). In gastric cancer, TMEM16A overexpression promotes tumor invasion and predicts poor prognosis by affecting TGF-β signaling function (19Liu F. Cao Q.H. Lu D.J. Luo B. Lu X.F. Luo R.C. et al.TMEM16A overexpression contributes to tumor invasion and poor prognosis of human gastric cancer through TGF-beta signaling.Oncotarget. 2015; 6: 11585-11599Crossref PubMed Scopus (73) Google Scholar). On the other hand, numerous studies have shown that TMEM16A inhibitors have significant therapeutic effects in multiple cancers including lung cancer (20Guo S. Bai X. Liu Y. Shi S. Wang X. Zhan Y. et al.Inhibition of TMEM16A by natural product silibinin: potential lead compounds for treatment of lung adenocarcinoma.Front. Pharmacol. 2021; 12643489Google Scholar), colorectal cancer (10Wang T. Wang H. Yang F. Gao K. Luo S. Bai L. et al.Honokiol inhibits proliferation of colorectal cancer cells by targeting anoctamin 1/TMEM16A Ca(2+) -activated Cl(-) channels.Br. J. Pharmacol. 2021; 178: 4137-4154Crossref PubMed Scopus (10) Google Scholar), oral squamous cell carcinoma (21Kashyap M.K. Marimuthu A. Kishore C.J.H. Peri S. Keerthikumar S. Prasad T.S.K. et al.Genomewide mRNA profiling of esophageal squamous cell carcinoma for identification of cancer biomarkers.Cancer Biol. Ther. 2009; 8: 36-46Crossref PubMed Scopus (103) Google Scholar), and breast cancer (17Luo S.Y. Wang H. Bai L.C.A. Chen Y.W. Chen S. Gao K. et al.Activation of TMEM16A Ca2+-activated Cl- channels by ROCK1/moesin promotes breast cancer metastasis.J. Adv. Res. 2021; 33: 253-264Crossref PubMed Scopus (16) Google Scholar). Among them, the effect of modulators on the MAPK signaling pathway is considered to be crucial to inhibit the proliferation of lung adenocarcinoma cells (22Guo S. Chen Y. Shi S. Wang X. Zhang H. Zhan Y. et al.Arctigenin, a novel TMEM16A inhibitor for lung adenocarcinoma therapy.Pharmacol. Res. 2020; 155104721Crossref Scopus (41) Google Scholar). Liu et al. (23Liu W. Lu M. Liu B. Huang Y. Wang K. Inhibition of Ca(2+)-activated Cl(-) channel ANO1/TMEM16A expression suppresses tumor growth and invasiveness in human prostate carcinoma.Cancer Lett. 2012; 326: 41-51Crossref PubMed Scopus (148) Google Scholar) found that TMEM16A was highly expressed in some prostate cancer cells and downregulation or inhibition of TMEM16A helped to reduce prostate cancer cell viability and to inhibit tumor growth. Further studies revealed that TMEM16A could regulate TNF-α signaling and promote apoptosis in prostate cancer PC-3 cells (24Song Y. Gao J. Guan L. Chen X. Gao J. Wang K. Inhibition of ANO1/TMEM16A induces apoptosis in human prostate carcinoma cells by activating TNF-alpha signaling.Cell Death Dis. 2018; 9: 703Crossref PubMed Scopus (47) Google Scholar). In conclusion, the roles of TMEM16A in cancer are relatively clear and TMEM16A inhibitors are important candidates for the treatment of tumors. Structure-based drug design and screening facilitated the discovery of TMEM16A inhibitors, which relied heavily on the ligand-binding pocket in the calcium-bound or calcium-free structure of the channel. Previously, we identified a nonselective inhibitor-binding pocket located in the extracellular vestibule region in the calcium-bound state of TMEM16A (25Shi S. Guo S. Chen Y. Sun F. Pang C. Ma B. et al.Molecular mechanism of CaCCinh-A01 inhibiting TMEM16A channel.Arch. Biochem. Biophys. 2020; 695108650Crossref Scopus (11) Google Scholar, 26Shi S. Ma B. Sun F. Qu C. An H. Theaflavin binds to a druggable pocket of TMEM16A channel and inhibits lung adenocarcinoma cell Viability.J. Biol. Chem. 2021; 297101016Abstract Full Text Full Text PDF Scopus (15) Google Scholar). We performed a virtual screening based on this binding pocket and screened two inhibitors, Theaflavin (26Shi S. Ma B. Sun F. Qu C. An H. Theaflavin binds to a druggable pocket of TMEM16A channel and inhibits lung adenocarcinoma cell Viability.J. Biol. Chem. 2021; 297101016Abstract Full Text Full Text PDF Scopus (15) Google Scholar) and Zafirlukast (27Shi S. Ma B. Sun F. Qu C. Li G. Shi D. et al.Zafirlukast inhibits the growth of lung adenocarcinoma via inhibiting TMEM16A channel activity.J. Biol. Chem. 2022; 298101731Abstract Full Text Full Text PDF Scopus (13) Google Scholar), from the natural product library and the marketed drug library. Meanwhile, Zhang and Bai found that mutations in several residues (D383, R515, R535, K603, E623, E624) of the extracellular vestibule significantly attenuated the potency of several TMEM16A inhibitors, including rutecarpine (9Zhao Z.J. Xue Y.R. Zhang G.H. Jia J. Xiu R.L. Jia Y.G. et al.Identification of evodiamine and rutecarpine as novel TMEM16A inhibitors and their inhibitory effects on peristalsis in isolated Guinea-pig ileum.Eur. J. Pharmacol. 2021; 908174340Crossref Scopus (6) Google Scholar), cepharanthine (8Zhang X. Zhang G.H. Zhao Z.J. Xiu R.L. Jia J. Chen P.P. et al.Cepharanthine, a novel selective ANO1 inhibitor with potential for lung adenocarcinoma therapy.Biochim. Biophys. Acta Mol. Cell Res. 2021; 1868119132Crossref Scopus (17) Google Scholar), and nuciferine (28Bai X. Liu X.Y. Li S.T. An H.L. Kang X.J. Guo S. Nuciferine inhibits TMEM16A in dietary adjuvant therapy for lung cancer.J. Agric. Food Chem. 2022; 70: 3687-3696Crossref PubMed Scopus (20) Google Scholar), further confirming the hypothesis that the extracellular vestibule is a druggable pocket. In 2022, the structure of the complex of TMEM16A with the inhibitor 1PBC was resolved, which showed that 1PBC binds below the extracellular vestibule and blocks the ion conduction pore, revealing a new inhibitor binding pocket (29Lam A.K.M. Rutz S. Dutzler R. Inhibition mechanism of the chloride channel TMEM16A by the pore blocker 1PBC.Nat. Commun. 2022; 13: 2798Crossref PubMed Scopus (7) Google Scholar). To inhibit an ion channel, the inhibitor either stabilizes the closed state of the channel or blocks the ion conduction pore of the channel in its open state. However, thus far, all reported TMEM16A structures, including the complex structure, are either in closed or desensitized states. Therefore, revealing the druggable pocket of TMEM16A exposed in the open state is important to understand protein–ligand interactions and facilitate rational drug design. In this study, we used enhanced sampling algorithm and segmental modeling method to reconstruct the open-state structure of TMEM16A and identified a druggable pocket that unifies the two identified inhibitor-binding pockets in protein space. Moreover, we screened a traditional Chinese medicine monomer derivative, Etoposide, based on this binding pocket and revealed its molecular mechanism of TMEM16A inhibition. Finally, we tested the inhibitory efficacy of Etoposide on TMEM16A highly expressing prostate cancer cells. These results conceptually show that the conformational changes of ion channels can directly influence their interactions with inhibitors and contribute to drug design targeting TMEM16A. Thus far, cryo-EM structures are available for the calcium-bound and the calcium-free states of TMEM16A (Fig. 1A). In the calcium-free structure (PDB ID: 5OYG) (30Paulino C. Kalienkova V. Lam A.K.M. Neldner Y. Dutzler R. Activation mechanism of the calcium-activated chloride channel TMEM16A revealed by cryo-EM.Nature. 2017; 552: 421-425Crossref PubMed Scopus (182) Google Scholar), the TM6 helix adopts a relaxed conformation, allowing cytoplasmic Ca2+ to access the ligand-binding site. However, all calcium-bound structures were observed to exist in a desensitized state (Fig. 1B) (30Paulino C. Kalienkova V. Lam A.K.M. Neldner Y. Dutzler R. Activation mechanism of the calcium-activated chloride channel TMEM16A revealed by cryo-EM.Nature. 2017; 552: 421-425Crossref PubMed Scopus (182) Google Scholar, 31Dang S. Feng S. Tien J. Peters C.J. Bulkley D. Lolicato M. et al.Cryo-EM structures of the TMEM16A calcium-activated chloride channel.Nature. 2017; 552: 426-429Crossref PubMed Scopus (208) Google Scholar, 32Lam A.K.M. Rheinberger J. Paulino C. Dutzler R. Gating the pore of the calcium-activated chloride channel TMEM16A.Nat. Commun. 2021; 12: 785Crossref PubMed Scopus (24) Google Scholar). Superimposition of the calcium-bound structure of TMEM16A with the open-state structure of nhTMEM16 (PDB ID: 6QM9) (33Kalienkova V. Mosina V.C. Bryner L. Oostergetel G.T. Dutzler R. Paulino C. Stepwise activation mechanism of the scramblase nhTMEM16 revealed by cryo-EM.Elife. 2019; 8e44364Crossref PubMed Scopus (64) Google Scholar) revealed that the RMSD values of the TM4 backbone of both structures exceeded 4 Å, while the extracellular portion reached 7 Å (Fig. 1, C and D). This indicates that activation of the desensitized state structure requires TM4 to complete conformational rearrangement. To construct the open-state structure of TMEM16A, we performed accelerated molecular dynamics (aMD) simulations. This method has been shown to enable a much faster evolution of the calculations by reducing the local energy barrier, thus capturing events on the millisecond scale (34Pierce L.C.T. Salomon-Ferrer R. de Oliveira C.A.F. McCammon J.A. Walker R.C. Routine access to millisecond time scale events with accelerated molecular dynamics.J. Chem. Theory Comput. 2012; 8: 2997-3002Crossref PubMed Scopus (383) Google Scholar). In our aMD simulation, we used the calcium-bound structure of TMEM16A (PDB ID: 5OYB) as the starting model. The aMD system was sampled for a total of 400 ns, and the RMSD of the protein backbone atoms in the system converged to 1.5 to 2.4 Å (Fig. S1). To clearly describe the conformational change of TM4, we set the radius of gyration of the outer structure of TM4 and TM6 and the distance between TM4 and TM6 at the entrance of the pore as two reaction coordinates and then reconstructed the free energy surface of the TMEM16A conformation. As shown in Figure 1E and Fig. S3A, two low-energy conformations of TM4 and TM6 helices appeared during aMD. The overlay of the simulated snapshots shows that the most dominant low-energy conformation of TM4 is to maintain an overall approach to TM6 (c state), while the other low-energy conformation is the upper part of the conformational changes of TM4 in the direction away from TM6 (b state, similar to nhTMEM16 open state) (Figs. 1F and S3B). During these processes, the TM6 conformation is more stable overall. Since the aMD method reduces the energy barrier for conformational changes of the whole protein, this may have a potential impact on the structure of the support regions (except TM3-TM7) of the protein. To reduce the impact of conformational change on protein stability, we performed structural reassembly using the captured b-state structure and cryo-EM structure (Fig. S3C). Specifically, TM3-TM5 of the newly assembled structures were derived from the b state structure obtained by aMD, while the rest of the protein body adopted the conformation of the cryo-EM structure of the desensitized state (PDB ID: 5OYB). The advantage of this segmental modeling is that it allows for local conformational rearrangement of the protein while minimizing the impact on the whole (Fig. S3D). Finally, we obtained the open-state structure of TMEM16A, and the local energy estimate of the membrane protein showed that the structure is reasonable. TM3 and TM4 have more obvious conformational rearrangements in the open-state structure compared with the desensitized state structure (Fig. 1G). Especially, the RMSD value of the upper part of TM4 exceeds 5 Å (Fig. 1H). The pore was further analyzed, and as shown in Figure 1I, the open state of TMEM16A has a funnel-shaped pore, and the pore radius exceeds 2 Å except for the chloride-binding site (K645) (35Shi S. Pang C. Ren S. Sun F. Ma B. Guo S. et al.Molecular dynamics simulation of TMEM16A channel: linking structure with gating.Biochim. Biophys. Acta Biomembr. 2022; 1864183777Crossref Scopus (4) Google Scholar). These data indicate that the open-state structure of TMEM16A constructed by the aMD simulation and segmental modeling method is a reasonable one. The calcium-bound state of TMEM16A has a nonselective inhibitor-binding pocket, the extracellular vestibule (26Shi S. Ma B. Sun F. Qu C. An H. Theaflavin binds to a druggable pocket of TMEM16A channel and inhibits lung adenocarcinoma cell Viability.J. Biol. Chem. 2021; 297101016Abstract Full Text Full Text PDF Scopus (15) Google Scholar) (Fig. 2A). From desensitized to open state, the TM3, TM4, and TM6 helices of TMEM16A move away from each other, significantly changing the extracellular vestibule. Therefore, we speculate that the channel may expose a new pocket of blocker when it is open. We predicted the ligand-binding pocket in the extracellular vestibule region using POCASA (36Yu J. Zhou Y. Tanaka I. Yao M. Roll: a new algorithm for the detection of protein pockets and cavities with a rolling probe sphere.Bioinformatics. 2010; 26: 46-52Crossref PubMed Scopus (241) Google Scholar) and found that the space between TM4 and TM6 increases and interconnects with the original extracellular vestibule, forming a potentially druggable pocket with a space of 1488 Å3 (Fig. 2B). To verify the modeling prediction, we performed a virtual screening targeting the druggable pocket with the traditional Chinese medicine monomer and derivative library (Fig. 2C). We selected compounds with high affinity for whole-cell membrane clamp electrophysiology testing with human embryonic kidney 293T (HEK293T) cells overexpressing TMEM16A. First, HEK293T cells transfected with TMEM16A plasmid have recordable currents in the presence of 600 nM Ca2+, and this current could be completely inhibited by T16Ainh-A01 of TMEM16A-specific inhibitor, indicating that the recorded current comes from TMEM16A (Fig. 2E). We identified a derivative of podophyllotoxin, Etoposide, that could inhibit TMEM16A current in a concentration-dependent manner, with a half-maximal inhibition concentration (IC50) of 13.6 ± 1.3 μM (Fig. 2, C–F). Current–voltage analysis shows that Etoposide does not affect the outward rectification current–voltage relationship characteristic of TMEM16A (Fig. 2G). To understand the molecular mechanism of TMEM16A inhibition by Etoposide, we constructed two molecular dynamics (MD) simulation systems based on open-state structures: Apo and Holo states, corresponding to the ligand-free and ligand-bound systems, respectively (Fig. 3, A and B). First, we calculated the root-mean-square fluctuation of the protein in both systems to assess the flexibility of the protein structure. As shown in Figure 3C, the Loops in both systems are more flexible, and the root-mean-square fluctuations of the pore region helices are less than 2 Å, indicating that these helices are more stable. Then, we calculated the radius of gyration of the binding pocket region to check the contraction and expansion of this region. As shown in Figure 3D, the Rg values of Apo and Holo states converge at 13.10 Å and 13.45 Å, respectively, indicating that the contraction of the binding pocket is more pronounced in the Apo state. While in the Holo state, the movement of TM4 to TM6 is blocked because of the binding of Etoposide. Finally, we superimposed snapshots of the Apo and Holo states at 200 ns with the open state and the desensitized state (PDB ID: 5OYB) of TMEM16A. Compared with the simulated initial open-state structure, the RMSD values of TM4 in the Apo and Holo states are greater than 2 Å, and the RMSD of the Apo state structure is even higher at 6 Å (Fig. 3, E and F). This indicates that TM4 is not stable away from TM6. Compared with the desensitized state structure (PDB ID: 5OYB), the RMSD value of TM4 in the Apo and Holo states is close to 2 Å. This suggests that the cryo-EM conformation of TMEM16A is an energy stable state and explains the difficulty to obtain the open-state conformation (Fig. 3, G and H). In all Holo systems, the RMSD of Etoposide was less than 1.5 Å and there were no large fluctuations, indicating that Etoposide binding in this pocket is very stable (Fig. S1). We speculate that Etoposide can enter the pocket when the channel is open and block the ion conduction pore of the channel thereby inhibiting TMEM16A current. We then explored the interaction mechanism between Etoposide and TMEM16A. As a blocking inhibitor, Etoposide binds to the pocket in the open state. Based on the Etoposide/TMEM16A binding pose, the dihedral angle between the Etoposide e/c ring and the angle between the g/a/e rings were set as reaction coordinates to construct a free-energy landscape of molecular conformations (Fig. 4A). As shown in Figure 4B and Fig. S4, the free energy landscape of Etoposide reveals two different modes of interaction with TMEM16A. In all two modes, the a/b/c/d rings of Etoposide are interposed between TM4 and TM6 and the e ring is bound near T518. The difference is that the g/f rings of Etoposide in the most stable state contact the hydrophobic gate, which is more favorable to fill the lower part of the pocket (Fig. 4C). The interaction analysis showed that Etoposide formed four weak interactions with the 21 amino acids of TMEM16A, mainly van der Waals and hydrogen bonding interactions (Fig. 4D). We performed the necessary quantum chemical calculations to characterize the electrostatic surface potential and electron localization function of Etoposide. As shown in Figure 4, E and F, three sites of Etoposide molecule have strong electrostatic features and high electron activity, which are the carboxyl and ester groups on the d/e/f ring, respectively. T518/N546/E633 of TMEM16A are close to the three active sites of Etoposide, respectively (Fig. 4, C and F). The results of site-directed mutagenesis showed that the TMEM16A channel could still be inhibited by 1 mM Etoposide after mutating T518/N546/E633 to alanine, respectively, but the potency of Etoposide to inhibit the channel was significantly reduced (Fig. 4, G–I). As shown in Figure 4, H and I, the half effective inhibitory concentration of etoposide on T518A/N546A/E633A reached 61.4 ± 4.8 μM, 73.8 ± 7.4 μM, 92.8 ± 7.3 μM, respectively, which was significantly higher than that of wildtype (13.6 ± 1.3 μM). In conclusion, the computational and mutagenesis experiments showed that the d/e/f rings of Etoposide are the key sites for binding TMEM16A, and the lateral insertion mode of the a/b/c/d rings further stabilizes the ligand binding. Previously, inhibitors of TMEM16A are widely believed to bind to the extracellular vestibule of the desensitized state of TMEM16A (9Zhao Z.J. Xue Y.R. Zhang G.H. Jia J. Xiu R.L. Jia Y.G. et al.Identification of evodiamine and rutecarpine as novel TMEM16A inhibitors and their inhibitory effects on peristalsis in isolated Guinea-pig ileum.Eur. J. Pharmacol. 2021; 908174340Crossref Scopus (6) Google Scholar, 27Shi S. Ma B. Sun F. Qu C. Li G. Shi D. et al.Zafirlukast inhibits the growth of lung adenocarcinoma via inhibiting TMEM16A channel activity.J. Biol. Chem. 2022; 298101731Abstract Full Text Full Text PDF Scopus (13) Google Scholar, 28Bai X. Liu X.Y. Li S.T. An H.L. Kang X.J. Guo S. Nuciferine inhibits TMEM16A in dietary adjuvant therapy for lung cancer.J. Agric. Food Chem. 2022; 70: 3687-3696Crossref PubMed Scopus (20) Google Scholar). To understand the interaction between Etoposide and the desensitized state, we performed docking and MD simulations based on the cryo-EM structure of the desensitized state (PDB ID: 5OYB) (30Paulino C. Kalienkova V. Lam A.K.M. Neldner Y. Dutzler R. Activation mechanism of the calcium-activated chloride channel TMEM16A revealed by cryo-EM.Nature. 2017; 552: 421-425Crossref PubMed Scopus (182) Google Scholar). First, we tested the binding affinity of Etoposide to the desensitized state of TMEM16A with a binding free energy of −10 kcal/mol (Fig. 5A). The binding mode of Etoposide was similar to that of the inhibitors Theaflavin and Zafirlukast, which were previously screened in this binding pocket (Fig. 5B). Then, simulations with a total time of 600 ns were performed to assess the stability of Etoposide in this binding pocket. As shown in Fig. S2, the RMSD of Etoposide during the simulation converged to 0.5 to 1.4 Å. Meanwhile, we constructed the free-energy landscape of Etoposide conformation using the distance between Etoposide and R535 and the angle between e/d/f rings as reaction coordinates (Fig. 5, C and D). We extracted the most stable state of Etoposide for interaction analysis. As shown in Figure 5, D and E, the a/e/d rings of Etoposide form H-bonds with R535/K603/R621/E633 and there is extensive van der Waals interactions with other surrounding re