SỞ KHOA HỌC VÀ CƠNG NGHỆ TP HỒ CHÍ MINH VIỆN KHOA HỌC VÀ CƠNG NGHỆ TÍNH TỐN BÁO CÁO TỔNG KẾT NGHIÊN CỨU CƠ CHẾ ỨC CHẾ SỰ TẠO SỢI AMYLOID CỦA TINH CHẤT TRÀ XANH (EGCG) Chủ nhiệm đề tài: TS Ngơ Sơn Tùng TP HỒ CHÍ MINH, THÁNG 7/2017 SỞ KHOA HỌC VÀ CÔNG NGHỆ TP HỒ CHÍ MINH VIỆN KHOA HỌC VÀ CƠNG NGHỆ TÍNH TỐN BÁO CÁO TỔNG KẾT NGHIÊN CỨU CƠ CHẾ ỨC CHẾ SỰ TẠO SỢI AMYLOID CỦA TINH CHẤT TRÀ XANH (EGCG) Viện trưởng: Nguyễn Kỳ Phùng Chủ nhiệm đề tài: TS Ngơ Sơn Tùng Ngơ Sơn Tùng TP HỒ CHÍ MINH, THÁNG 7/2017 Nghiên cứu chế ức chế tạo sợi Amyloid tinh chất trà xanh (EGCG) MỤC LỤC Trang MỞ ĐẦU ĐƠN VỊ THỰC HIỆN KẾT QUẢ NGHIÊN CỨU I Báo cáo khoa học II Tài liệu khoa học xuất III IChương trình giáo dục đào tạo IV Hội nghị, hội thảo 10 V File liệu 11 TÀI LIỆU THAM KHẢO CÁC PHỤ LỤC PHỤ LỤC 1: Bài báo “EGCG inhibits the oligomerization of amyloid beta (16-22) hexamer: Theoretical studies” PHỤ LỤC 2: Chứng lập trình Python Ô Nguyễn Thanh Nguyên Viện Khoa học Công nghệ Tính tốn TP Hồ Chí Minh Page Nghiên cứu chế ức chế tạo sợi Amyloid tinh chất trà xanh (EGCG) MỞ ĐẦU Như ta biết, Azheimer, biết đến biểu hội chứng suy giảm chức thần kinh, bệnh ảnh hưởng lên phần ba người cao tuổi (15) Tiếc đến bệnh lý AD chưa hiểu rõ Rất nhiều mơ hình đưa nhằm lý giải cho tác động AD lên não người bệnh Trong giả thuyết Amyloid mơ hình phổ biến kiểm chứng thơng qua nhiều cơng trình khoa học (6-8) Giả thuyết cho trình kết tập Aβ peptide hình thành nên cấu hình gây độc mà từ làm tổn thương đến não bệnh nhân AD (9-14) Do đó, nghiên cứu tập trung vào việc tìm kiếm chất ức chế tiềm tác động lên cấu hình Aβ dạng tiền sợi để điều trị AD Không may thông tin cấu trúc phân tử Aβ dạng tiền sợi cịn giới hạn chúng tồn mơi trường hỗn hợp gồm dạng sợi trưởng thành tiền sợi có bậc khác (15, 16) Việc tìm kiếm hợp chất ức chế hiệu chưa thành cơng có nhiều nghiên cứu cấu trúc Aβ dạng tiền sợi đoạn ngắn, sợi đơn sợi đôi (17-19) Mặc dù có nhiều cơng trình nghiên cứu thực để tìm hiểu cấu trúc Aβ dạng sợi đơn sợi đôi, kết gần lại hai dạng sợi không gây tổn thương cho tế thần kinh (20) Não bệnh nhân AD bị ảnh hưởng phân tử Aβ dạng tiền sợi có khối lượng lớn dạng sợi ba, sợi bốn v.v… Quá trình kết tập Aβ biết thường lên tới nhiều tuần nên việc mô phân tử Aβ tiền sợi dài đầy đủ có khối lượng lớn máy tính khơng thể giới hạn máy tính Bên cạnh đó, đoạn ngắn cắt từ Aβ Aβ16-22 Aβ10-35 chứng minh đại diện cho tính chất hoạt động sợi Aβ dài đầy đủ (17, 21, 22) Vì vậy, nghiên cứu tính tốn ảnh hưởng chất ức chế lên q trình tiền sợi hóa đoạn ngắn Aβ peptide đưa tranh tương tự cho tiền sợi dài đầy đủ (23) Chiến lược để tìm kiếm chất ức chế Aβ dạng tiền sợi bậc cao hiệu quy vào hai hướng Thứ thiết kế sàng lọc hợp chất có khả ngăn cản trình kết tập tạo thành Aβ dạng tiền sợi bậc cao (24-27) Thứ hai tìm hợp chất có khả làm giảm thời gian tồn Aβ dạng tiền sợi bậc cao (28) Có thể lý luận hợp chất đẩy nhanh trình kết tụ Aβ để tạo dạng Viện Khoa học Cơng nghệ Tính tốn TP Hồ Chí Minh Page Nghiên cứu chế ức chế tạo sợi Amyloid tinh chất trà xanh (EGCG) sợi trưởng thành Dựa hai phương hướng trên, nhiều chất ức chế tiềm thiết kế tìm bao gồm cách hợp chất tự nhiên (25, 29), β-sheet breaker peptides (30-32), dẫn suất vitamin K3 (26), carvedilol (33), adenosine triphosphate (34), propafenone (27), v.v Một số số chúng thử nghiệm lâm sàng adenosine triphosphate, carvedilol, nilvadipine, curcumin (35) Mặc dù vậy, cách thức mà hợp chất kiềm chế thúc đẩy hoạt động Aβ dạng tiền sợi bậc cao chưa hiểu biết rõ ràng Bên cạnh cấu trúc vịng thơm phát đóng vai trị quan trọng việc hình thành liên kết π-π stacking chuỗi Aβ peptide liền kề (36) Hợp chất epigallocatechin-3-gallate (EGCG), phân tử nhỏ chiết xuất từ trà xanh, mang vòng thơm cấu trúc, tạo nên liên kết π với residue kỵ nước Aβ Do đó, EGCG trở thành chức ức chế tiềm Aβ peptide (37) Cơ chế vật lý trình EGCG cản trở hoạt động Aβ dạng sợi đơn sợi đôi nghiên cứu mô máy tính (38-40), sau sợi đơn đôi phát không gây độc cho thần kinh Khả ức chế Aβ dạng sợi ba EGCG xác định phương pháp docking với thụ thể linh động (41) Kết docking EGCG ưu tiên tương tác với lõi kỵ nước trung tâm (residues 17-21) (41), nhờ làm giảm cấu hình beta vùng (42) Mặc dù docking với thụ thể linh động (flexible) có độ xác tốt so với với thụ thể tĩnh (rigid), phương pháp docking gặp hạn chế bỏ qua động học phân tử thụ thể lẫn phối tử Bên cạnh đó, thực nghiệm EGCG làm tăng số cấu hình cuộn ngẫu nhiên (ít độc) Aβ42 (43) Việc nghiên cứu sâu rộng chế EGCG ức chế cấu hình Aβ dạng tiền sợi có bậc cao hai lý thú EGCG qua phase II thử nghiệm lâm sàng chống AD (44) Tuy kết thử nghiệm lâm sàng chưa công bố, biết đến khả chống lão hóa tuổi già trà xanh y học cổ truyền Do đó, đề tài này, sử dụng mô động học trao đổ mơ hình (REMD) để nghiên cứu chế việc EGCG ức chế tiền sợi hóa Aβ dạng đa sợi Những kiến thức giúp gia tăng khả chữa trị AD Viện Khoa học Cơng nghệ Tính tốn TP Hồ Chí Minh Page Nghiên cứu chế ức chế tạo sợi Amyloid tinh chất trà xanh (EGCG) Lời cảm ơn đến ICST Công việc tài trợ Sở Khoa học & Công nghệ Viện Khoa học & Cơng nghệ Tính tốn (ICST) thành phố Hồ Chí Minh với đề tài có Hợp đồng số 186/2016/HD-SKHCN ngày 16 tháng 12 năm 2016 Viện Khoa học Công nghệ Tính tốn TP Hồ Chí Minh Page Nghiên cứu chế ức chế tạo sợi Amyloid tinh chất trà xanh (EGCG) ĐƠN VỊ THỰC HIỆN Phịng thí nghiệm: Chủ nhiệm đề tài: TS Ngô Sơn Tùng Thành viên đề tài: GS.TS Nguyễn Minh Thọ TS Nguyễn Minh Tâm Ô Nguyễn Thanh Nguyên Cơ quan phối hợp: Viện Khoa học Cơng nghệ Tính tốn TP Hồ Chí Minh Page Nghiên cứu chế ức chế tạo sợi Amyloid tinh chất trà xanh (EGCG) KẾT QUẢ NGHIÊN CỨU I BÁO CÁO KHOA HỌC Thông qua 24.000 ns mô MD, chứng minh nghiên cứu lý thuyết hiệu ứng EGCG ức chế tiền sợi hóa Aβ16-22 hexamer dung dịch, với làm rõ ý nghĩa vật lý khả nhờ tăng cường khả điều trị AD Nhìn chung, EGCG phát ngăn chặn hình thành tiền sợi hóa Aβ thơng qua cách khác cách gia tăng tiền sợi hóa Aβ cách giảm nội dung β hexamer Từ phân tích bề mặt lượng tự do, peptide Aβ chèn EGCG có cực tiểu cục với rào cản lượng lớn liên kết chúng Thông qua tương tác với peptide, EGCG có xu hướng đẩy nhanh tiến trình tiền sợi hóa làm ổn định cấu hình tiền sợi Aβ việc ổn định cấu trúc bậc hai peptide Aβ Ngoài ra, chất ức chế làm giảm số lượng nhóm cấu hình trung gian Aβ16-22 hexamer hịa tan Phát phù hợp với thí nghiệm sẵn có (45) Sự diện EGCG làm giảm kích thước β-sheet tiền sợi có bậc cao hai Mặc dù hình ảnh tương tác làm rõ, bị thay đổi xem xét Aβ40/42 có chiều dài đầy đủ Hơn nữa, phân tử chèn vào sợi lục Aβ16-22, nhờ làm tăng kích thước tham số đại diện bao gồm Rg CCS, thay đổi đồ liên kết khơng hóa trị peptide Aβ thơng qua hình thành tương tác khơng hóa trị với sợi đơn cấu thành Các phép tính hóa học lượng tử (DFT) cho thấy việc xếp chồng π đóng vai trị quan trọng tương tác EGCG peptide Aβ16-22 Xác định lượng tự ràng buộc lực liên kết EGCG với peptide Aβ có độ lớn tương tự chất ức chế có sẵn khác curcumin Tuy nhiên, phân tử EGCG tạo tương tác tĩnh điện van der Waals (vdW) với peptide tương tự lượng, chất curcumin lại có tương tác vdW mạnh mẽ hẳn (46) Nhìn chung, nghiên cứu đề nghị sử dụng EGCG điều trị AD Các kết nghiên cứu trình bày cách chi tiết báo khoa học đề cập Phụ lục I Viện Khoa học Công nghệ Tính tốn TP Hồ Chí Minh Page Nghiên cứu chế ức chế tạo sợi Amyloid tinh chất trà xanh (EGCG) II CÁC TÀI LIỆU KHOA HỌC ĐÃ XUẤT BẢN Các kết nghiên cứu công bố tạp chí Journal of Molecular Graphics and Modelling: Ngo, S T.; Truong, D T.; Tam, N M.; Nguyen, M T., EGCG inhibits the oligomerization of amyloid beta (16-22) hexamer: theoretical studies J Mol Graph Model 2017, 76, 1-10 Viện Khoa học Cơng nghệ Tính tốn TP Hồ Chí Minh Page Nghiên cứu chế ức chế tạo sợi Amyloid tinh chất trà xanh (EGCG) III CHƯƠNG TRÌNH GIÁO DỤC VÀ ĐÀO TẠO Ơ Nguyễn Thanh Nguyên hỗ trợ đào tạo khoa học Python (Phụ lục II) Viện Khoa học Cơng nghệ Tính tốn TP Hồ Chí Minh Page Nghiên cứu chế ức chế tạo sợi Amyloid tinh chất trà xanh (EGCG) 20 Sun Y, Xi W, Wei G Atomic-Level Study of the Effects of O4 Molecules on the Structural Properties of Protofibrillar Aβ Trimer: β-Sheet Stabilization, Salt Bridge Protection, and Binding Mechanism J Phys Chem B 2015;119(7):2786-94 21 Klimov DK, Thirumalai D Dissecting the assembly of A beta(16-22) amyloid peptides into antiparallel beta sheets Structure 2003;11:295-307 22 Wu C, Wang ZX, Lei HX, Duan Y, Bowers MT, Shea JE The binding of Thioflavin T and its neutral analog BTA-1 to protofibrils of the Alzheimer's disease A beta(16-22) peptide probed by molecular dynamics simulations J Mol Biol 2008;384(3):718-29 23 Xie L, Luo Y, Lin D, Xi W, Yang X, Wei G The molecular mechanism of fullerene-inhibited aggregation of Alzheimer's beta-amyloid peptide fragment Nanoscale 2014;6(16):9752-62 24 Ngo ST, Li MS Curcumin Binds to Abeta1-40 Peptides and Fibrils Stronger than Ibuprofen and Naproxen J Phys Chem B 2012;116(34):10165-75 25 Ngo ST, Li MS Top-leads from Natural Products for Treatment of Alzheimer's Disease: Docking and Molecular Dynamics Study Mol Sim 2013;39(4):279-91 26 Huy PD, Yu YC, Ngo ST, Thao TV, Chen CP, Li MS, et al In silico and in vitro Characterization of Anti-Amyloidogenic Activity of Vitamin K3 Analogues for Alzheimer's Disease Biochim Biophys Acta 2013;1830(4):2960-9 27 Ngo ST, Fang S-T, Huang S-H, Chou C-L, Huy PDQ, Li MS, et al Antiarrhythmic medication Propafenone is potential drug for Alzheimer’s disease by inhibiting aggregation of Aβ: in silico and in vitro studies J Chem Inf Model 2016 28 Bieschke J, Herbst M, Wiglenda T, Friedrich RP, Boeddrich A, Schiele F, et al Small-Molecule Conversion of Toxic Oligomers to Nontoxic β-Sheet–Rich Amyloid Fibrils Nat Chem Biol 2012;8(1):93-101 29 Lim G, Chu T, Yang F, Beech W, Frautsch S, Cole G The curry spice curcumin reduces oxidative damage and amyloid pathology in an Alzheimer transgenic mouse J Neurosci 2001;21(21):8370-7 30 Tjerngber LO, Naslund J, Lindqvist F, Jahannson J, Karlstrom AR, Thyberg J, et al Arrest of beta-amyloid beta-peptide fibril formation by a pentapeptide J Biol Chem 1996;271:8545-8 31 Wu C, Murray MM, Summer SLBL, Condron MM, Bitan G, Shea JE, et al The structure of A beta 42 C-terminal fragments probed by a combined experimental and theoretical study J Mol Biol 2009;387:492-501 32 Li HY, Monien MB, Lomakin A, Zemel R, Fradinger EA, Tan MA, et al Mechanistic investigation of the inhibition of A beta 42 assembly and neurotoxicity by A beta 42 C-terminal Biochemistry 2010;49:6358-64 33 Wang J, Ono K, Dickstein DL, Arrieta-Cruz I, Zhao W, Qian X, et al Carvedilol as a potential novel agent for the treatment of Alzheimer’s disease Neurobiol Aging 2011;32(12):2321.e1-.e12 34 Madrid PB, Chopra S, Manger ID, Gilfillan L, Keepers TR, Shurtleff AC, et al A Systematic Screen of FDA-Approved Drugs for Inhibitors of Biological Threat Agents PLoS ONE 2013;8(4):e60579 35 Paris D, Quadros A, Humphrey J, Patel N, Crescentini R, Crawford F, et al Nilvadipine antagonizes both Aβ vasoactivity in isolated arteries, and the reduced cerebral blood flow in APPsw transgenic mice Brain Res 2004;999(1):53-61 Viện Khoa học Cơng nghệ Tính tốn TP Hồ Chí Minh Page 13 Nghiên cứu chế ức chế tạo sợi Amyloid tinh chất trà xanh (EGCG) 36 Gazit E A possible role for π-stacking in the self-assembly of amyloid fibrils FASEB J 2002;16(1):77-83 37 Kristen AV, Lehrke S, Buss S, Mereles D, Steen H, Ehlermann P, et al Green tea halts progression of cardiac transthyretin amyloidosis: an observational report Clin Res Cardiol 2012;101(10):805-13 38 Zhang T, Zhang J, Derreumaux P, Mu Y Molecular Mechanism of the Inhibition of EGCG on the Alzheimer Aβ1–42 Dimer J Phys Chem B 2013;117(15):3993-4002 39 Wang S-H, Liu F-F, Dong X-Y, Sun Y Thermodynamic Analysis of the Molecular Interactions between Amyloid β-Peptide 42 and (−)-Epigallocatechin-3gallate J Phys Chem B 2010;114(35):11576-83 40 Liu F-F, Dong X-Y, He L, Middelberg APJ, Sun Y Molecular Insight into Conformational Transition of Amyloid β-Peptide 42 Inhibited by (−)-Epigallocatechin-3gallate Probed by Molecular Simulations J Phys Chem B 2011;115(41):11879-87 41 Chebaro Y, Jiang P, Zang T, Mu Y, Nguyen PH, Mousseau N, et al Structures of Aβ17–42 Trimers in Isolation and with Five Small-Molecule Drugs Using a Hierarchical Computational Procedure J Phys Chem B 2012;116(29):8412-22 42 Lopez del Amo JM, Fink U, Dasari M, Grelle G, Wanker EE, Bieschke J, et al Structural Properties of EGCG-Induced, Nontoxic Alzheimer's Disease Aβ Oligomers J Mol Biol 2012;421(4–5):517-24 43 Lorenzen N, Nielsen SB, Yoshimura Y, Vad BS, Andersen CB, Betzer C, et al How epigallocatechin gallate can inhibit α-synuclein oligomer toxicity in vitro J Biol Chem 2014 44 Sunphenon EGCg (Epigallocatechin-Gallate) in the Early Stage of Alzheimer´s Disease (SUN-AK) In: ClinicalTrials.gov [Internet] Bethesda (MD): National Library of Medicine (US) 2000-: Charite University, Berlin, Germany; [NLM Identifier: NCT00951834:[Available from: https://clinicaltrials.gov/ct2/show/NCT00951834 45 Ehrnhoefer DE, Bieschke J, Boeddrich A, Herbst M, Masino L, Lurz R, et al EGCG redirects amyloidogenic polypeptides into unstructured, off-pathway oligomers Nat Struct Mol Biol 2008;15(6):558-66 46 Ngo ST, Fang S-T, Huang S-H, Chou C-L, Huy PDQ, Li MS, et al Antiarrhythmic Medication Propafenone a Potential Drug for Alzheimer’s Disease Inhibiting Aggregation of Aβ: In Silico and in Vitro Studies J Chem Inf Model 2016;56(7):134456 Viện Khoa học Công nghệ Tính tốn TP Hồ Chí Minh Page 14 Nghiên cứu chế ức chế tạo sợi Amyloid tinh chất trà xanh (EGCG) CÁC PHỤ LỤC PHỤ LỤC 1: Bài báo : EGCG inhibits the oligomerization of amyloid beta (16-22) hexamer: theoretical studies Tạp chí : Journal of Molecular Graphics and Modelling Tác giả : Son Tung Ngo, Duc Toan Truong, Nguyen Minh Tam, Minh Tho Nguyen Viện Khoa học Cơng nghệ Tính tốn TP Hồ Chí Minh Page 15 Nghiên cứu chế ức chế tạo sợi Amyloid tinh chất trà xanh (EGCG) PHỤ LỤC 2: Chứng lập trình Python Ơ Nguyễn Thanh Ngun Viện Khoa học Cơng nghệ Tính tốn TP Hồ Chí Minh Page 16 Journal of Molecular Graphics and Modelling 76 (2017) 1–10 Contents lists available at ScienceDirect Journal of Molecular Graphics and Modelling journal homepage: www.elsevier.com/locate/JMGM EGCG inhibits the oligomerization of amyloid beta (16-22) hexamer: Theoretical studies Son Tung Ngo a,b,∗ , Duc Toan Truong a,b , Nguyen Minh Tam a,b,∗ , Minh Tho Nguyen c,d a Computational Chemistry Research Group, Ton Duc Thang University, Ho Chi Minh City, Viet Nam Faculty of Applied Sciences, Ton Duc Thang University, Ho Chi Minh City, Viet Nam Institute for Computational Science and Technology (ICST) Quang Trung Software City, Ho Chi Minh City, Viet Nam d Department of Chemistry, KU Leuven, B-3001 Leuven, Belgium b c a r t i c l e i n f o Article history: Received May 2017 Received in revised form 16 June 2017 Accepted 19 June 2017 Available online 21 June 2017 Keywords: EGCG Replica exchange molecular dynamics Amyloid beta oligomer Free energy perturbation -stacking Quantum calculation a b s t r a c t An extensive replica exchange molecular dynamics (REMD) simulation was performed to investigate the progress patterns of the inhibition of (−)-epigallocatechin-3-gallate (EGCG) on the A16-22 hexamer Structural variations of the oligomers without and with EGCG were monitored and analyzed in detail It has been found that EGCG prevents the formation of A oligomer through two different ways by either accelerating the A oligomerization or reducing the -content of the hexamer It also decreases the potential “highly toxic” conformations of A oligomer, which is related to the conformations having high order -sheet sizes Both electrostatic and van der Waals interaction energies are found to be involved to the binding process Computed results using quantum chemical methods show that the - stacking is a critical factor of the interaction between EGCG and the peptides As a result, the binding free energy of the EGCG to the A peptides is slightly larger than that of the curcumin © 2017 Elsevier Inc All rights reserved Introduction It has been known that about a third of the elder population suffer a neurodegenerative disorder including, among others, the Alzheimer’s disease (AD) [1–5] Until recently, the working mechanisms of AD have however not been well understood, and many paradigms have been proposed to explain how AD affects the human brain Of the most popular ones, the Amyloid hypothesis is supported by several recent studies [6–9] Accordingly, the Amyloid beta (A) peptides deposit to each other giving the oligomer forms, and then damage the brain of AD patients in the extracellular region [10–15] Therefore, the search for new potential inhibitors for formation A oligomers is of great current interest Unfortunately, because A oligomers usually exist in an environmental mixture of various types of oligomers and the A mature fibrils, experimental information [16,17] and theoretical studies [18,19] on such oligomers remain limited Although numerous designs have been proposed in both previous computational and experimental studies [20–23], efficient A therapies have been not completely suc- ∗ Corresponding authors at: Computational Chemistry Research Group, Ton Duc Thang University, Ho Chi Minh City, Vietnam E-mail addresses: ngosontung@tdt.edu.vn (S.T Ngo), nguyenminhtam@tdt.edu.vn (N.M Tam) http://dx.doi.org/10.1016/j.jmgm.2017.06.018 1093-3263/© 2017 Elsevier Inc All rights reserved cessful yet In this context, computational investigations of the A oligomer conformations including the A fragments, including their monomers and dimers [24–26] remain of great current interest, in view of a deeper understanding of the oligomerization processes Numerous investigations of the conformations of monomer and dimer forms of A peptides have been carried out, and recent studies indicated that these are of low toxicity [27] The patient’s brain is most likely to be injured by higher order oligomers such as the trimers, tetramers, etc In view of the fact that the aggregation process is very slow, which undergoes from several hours to several weeks in nature, it is not possible at this moment to carry out computational investigations of the misfolding of high order full-length A peptides starting from random initial conformations, simply due to the limitation of the computing resource Moreover, the short A peptide fragments, consisting of the central hydrophobic core of the peptides, are often used to target the behaviour of full-length counterparts such as A16-22 , A10-35 , etc [24,28,29] There are two main strategies in the search for a new potential candidate to prevent formation of A oligomers Search or design of a compound, which is able to inhibit the A oligomer formation, is the first strategy [20–22] Reducing the lifetimes of A oligomers or decreasing their amount constitutes the second strategy which implies discovering of the compounds having a possibility to enhance or slow down the aggregation process [30] S.T Ngo et al / Journal of Molecular Graphics and Modelling 76 (2017) 1–10 Several potential inhibitors were thus designed following these strategies, including natural compounds [21,31], -sheet breaker peptides [32–34], carvedilol [35], adenosine triphosphate [36], propafenone [22], etc Some of them have subsequently been under clinical trial, namely adenosine triphosphate, carvedilol, and nilvadipine [37] However, the molecular mechanism exerted by these compounds in preventing formation of A oligomer(s) or in accelerating this process is still not clearly understood According to the hypothesis put forward in a previous study by Gazit et al [38], the aromatic rings might play a crucial role in the amyloidogenesis by creating - stacking between neighboring chains of A peptides The (−)-epigallocatechin-3-gallate (denoted hereafter as EGCG), which is a small compound derived from green tea, consists of three aromatic rings and can thereby form - stacking contacts with A hydrophobic residues that may lead to the collapse of A oligomer structures In this regard, EGCG has emerged as a potential inhibitor of the A peptides [39] Physical insights into the actions of EGCG preventing A monomers to get together were clarified through computational studies [40–42], despite the fact that both A monomers and dimers are low toxicity forms [27] Inhibition of EGCG on the A trimer was determined through flexible receptor molecular docking method [43] Although the latter method produced more accurate results in comparison to the rigid receptor molecular docking method, the flexible docking approach remains limited due to the lack of the actual dynamics of both receptor and ligand Several earlier studies indicated that high concentration of EGCG is able to prevent the formation of A fibrils [44–47], but the relevant mechanism is unknown [45] However, it has been shown that EGCG favorably binds to the hydrophobic region of A peptides [45,48] Moreover, recent experiments suggested that the fibril derivatives of A peptides are not toxic, whereas the A oligomers could induce neurotoxicity [10–15,27] Therefore, a comprehensive study of the ways in which EGCG inhibits the formation of the high order of A oligomers (e.g pentamer, hexamer, etc.) is also of great current interest, because it has actually passed the phase II clinical trial [49] A deep understanding of the mechanism of binding between EGCG molecule and A oligomer molecule may enlarge the search scope for A therapies Due to the fact that the computing (CPU) time consumption is greatly prohibited, it is not possible yet to perform computations on the oligomerization of full-length A peptides An alternative approach is an investigation on their smaller but representative fragments Following this approach, we attempt in the present work to clarify the prevention of A oligomer formation by the action of EGCG For this purpose, we set out to investigate the effects of a single EGCG molecule on the oligomerization of the hexamer A16-22 peptides using both replica exchange molecular dynamics (REMD) simulations and quantum chemical calculations In this scheme, the randomly initial structures of two soluble systems are generated including 6A16-22 and 6A16-22 + EGCG Through more than 24,000 ns of MD simulations, the effects of EGCG on the oligomerization of the A16-22 hexamer are probed The structural change of the hexamers with the presence and absence of EGCG during the oligomerization process is monitored and analyzed Numerous series of evaluation are carried out such as the secondary structure, collision cross section, non-bonded contacts, free energy surfaces, clustering conformations, and binding free energy calculations The contribution of the - stacking to the interaction is investigated through quantum chemical calculation using density functional theory (DFT) with the B97-D functional and the 6-311++G(d,p) basis set Our purpose is to probe the molecular mechanism of the EGCG inhibition on the oligomerization processes of the high order A oligomers, and thereby to contribute to the understanding of the AD functioning Such a knowledge is imperative for an improvement of the AD therapy Fig (A) is a three-dimensional structure of EGCG (B) + (C) is the starting structures of the A16-22 hexamer with and without the presence of the EGCG compound, respectively The initial conformations of the hexamer are randomly generated using PyMOL package with starting helical structures of 6A Solvation are hidden from the figure Materials and methods 2.1 Replica exchange molecular dynamics (REMD) simulations In the current work, the effects of EGCG on the conformations of the A16-22 hexamer are investigated using the REMD method The impact of EGCG is thus described through a comparison of two systems of hexamers without and with the presence of the inhibitor The primary structure of the A16-22 fragment is generated through the PyMOL package [50] with helical structure base, then they are randomly generated to create hexamer systems using VMD application [51], as shown in Fig 1A Their parameters are treated using the all-atom Amber99SB-ILDN force field [52] The latter is well tested to be one of the most suitable force fields to represent A peptides [53] The three dimensional structure of EGCG is taken from PubChem database [54], and then it is randomly inserted in the 6A16-22 hexamer system (Fig 1B) The EGCG is represented using the Generalized Amber force field (GAFF) [55], implying a combination of the ACPYPE package and Ambertools 1.4 [56] The information on the net atomic charges of EGCG is obtained from molecular orbital (MO) quantum chemical (QM) computations at the Hartree-Fock (HF) level with the 6-31G(d,p) basis set Both hexamers without and with the presence of EGCG are solvated with a dodecahedron periodic boundary conditions (PBC) box with the volume of approximately 250 nm3 In this, water molecules are represented through the TIP3P water model [57] according to the recommendation of GROMACS Therefore, the solvated hexamer systems include the six A16-22 molecules, ∼8.000 water molecules, and one or zero EGCG molecule corresponding to the presence or absence of the EGCG compound, respectively Two of the solvated systems constitute the initial conformations of the REMD simulations utilizing GROMACS packaged 5.0.6 [58] For each system, an energy minimization is first carried out using the steepest descent method The minimized simulation is finished when the maximum of force is smaller than 10−6 kJ/(mol nm) Consequently, the energy minimized A hexamers are relaxed during 500 ps canonical ensemble (NVT) simulations at 300 K During the NVT simulation, the systems are positionally restrained with small harmonic forces The REMD simulations are then performed with the starting conformation of two solvated hexamers taking from the last snapshots of the NVT simulations In REMD simulations, the systems are simulated over 48 different temperatures, which are described in the Supplementary Information (SI) file These temperatures are determined using the generator webserver [59] The replica acceptance ratio is selected to be ∼20%, and the exchange S.T Ngo et al / Journal of Molecular Graphics and Modelling 76 (2017) 1–10 between two adjacent replicas is checked every ps Overall, more than 250,000 exchanged times are attempted over REMD simulations for each system The monitored data are performed every ps MD simulations are established using the parameters referred the previous studies [18] These include a time step of fs, a relaxation time of 0.1 ps, constraining covalent bonds through the LINCS method with an order 4, the non-bonded cutoff of 0.9 nm, calculated electrostatics interaction utilizing the fast smooth Particle-Mesh Ewald electrostatics method, and estimated van der Waals (vdW) interactions with cutoff of 0.9 nm 2.2 Free energy perturbation method The binding free energy of a ligand to a receptor can accurately be determined through the double-annihilation binding free energy method [22,60,61] In this scheme, a ligand molecule is decomposed into two systems including solvated complex and solvated ligand in such a way that the difference in free energy for annihilating the ligand from these systems is equal to the binding free energy of a ligand to a receptor The free energy differences of these processes are evaluated through the free energy perturbation (FEP) method [60] In this method, the coupling parameter is employed to alter the Hamiltonian of the solvated system from the full interaction state of the ligand with surrounding atoms ( = 0) to non-interaction state of the ligand with surrounding atoms ( = 1) The sum of these changes is actually the free energy difference between two states ( = and 1), evaluated by the Bennet’s acceptance ratio (BAR) method [62] In the present work, eight values of are used to adjust the Coulombic interaction including 0.0, 0.10, 0.20, 0.35, 0.50, 0.65, 0.80 and 1.00 Moreover, nine values of are chosen to alter the van der Waals (vdW) interactions including 0.00, 0.10, 0.25, 0.35, 0.50, 0.65, 0.75, 0.90 and 1.00 2.3 Measures used in data analysis The secondary structure of A16-22 hexamer is investigated utilizing the DSSP tool [63] The non-bonded and hydrogen bonded contacts between two molecules are evaluated The hydrogen bond is available when the acceptor-donor distance is smaller than 0.35 nm and the acceptor-hydrogen-donor angle larger than 135◦ The non-bonded interaction between two heavy atoms is counted when the mutual distance is smaller than 0.45 nm The GROMACS tool “sham” is employed to generate the free energy surface (FES) [64,65] The reaction coordinates are chosen as the gyrate radius (Rg ) and the root mean square deviation (RMSD) of the A16-22 hexamers The clustering method is carried out using GROMACS tool [66,67] The collision cross section (CCS) is a high impact parameter that can be determined through the ion mobility projection approximation calculation tool (IMPACT) [68] The - stacking contact between EGCG and A peptide is examined using the PLIP approach [69] 2.4 Quantum mechanical interaction energy The - stacking interaction energy between EGCG and A16-22 peptides is estimated from the interaction between aromatic rings of EGCG and part of A peptides, whose modeling structures are shown in Fig S1 (SI file) The corresponding binding energy is calculated by quantum chemical methods using density functional theory (DFT) with the B97-D functional in conjunction with the 6-311++G(d,p) basis set, which has been calibrated in a previous study [70] The Gaussian 09 package [71] is employed for DFT computations 3 Results and discussion 3.1 REMD simulations As mentioned above, the aggregation process of full-length A peptides is up to several days that leads to an extremely high CPU consumption when the high order full-length A oligomers are considered using MD computations [18,19] The A16-22 fragment, which is the central hydrophobic core of A peptide, can be used to represent the properties of the full-length one [72,73] Therefore, the A16-22 hexamer is randomly generated to investigate the effects of EGCG on the hexamer (Fig 1) The hexamer is duplicated, and then an EGCG molecule is randomly inserted into one soluble hexamer system Both hexamer systems with the presence and absence of EGCG are solvated using dodecahedron boxes, which are large enough to prevent any interaction contacts of an isolated molecule to itself through the PBC Two solvated systems are minimized and relaxed over energy minimizations and NVT simulations The equilibrated snapshots are then used as initial conformations of REMD simulations, which form an extremely efficient sampling method to investigate the A aggregation [74–77] For details of computations, the recorded exchange rates are ∼29% (Fig S2 of SI) over 250 ns of REMD simulations, the replicas are thus diffused over the entire temperature space as shown in Fig S3 (SI) as for an example of the track change of the A conformations at the lowest and highest temperatures during the computations In order to avoid any initial bias, the first 150 ns of REMD simulations of both soluble hexamer systems are discarded from any analysis Both hexamer secondary structures are investigated through the DSSP approach [63] and the results of 48 replicas at 150 ns are shown in Fig In particular, the hexamer without the presence of EGCG is formed with ∼0–48% of -content, ∼48–100% of random coil structure, ∼0–12% of turn structure, and ∼0–7% of ␣-content Moreover, the A hexamer in the presence of EGCG is formed ∼0–50% of -content, ∼50–100% of coil-content, ∼0–19% turn structure, and ∼0–7% helix structure The high diffusion of these metrics indicates that our computations are not targeted to any distinct structures Our REMD simulations are well converged, and both solvated hexamer systems reach equilibrium after 150 ns All investigated metrics are measured during different time intervals including 150–230 and 150–250 ns of REMD simulations such as radius of gyration, RMSD, -content and solvent access surface area (SASA) The consistence of these values implies the stability of soluble systems (Fig 3) Four metrics of the solvated hexamer are obtained without the presence of EGCG, namely, the average of radius of gyration of 1.33 ± 0.36 nm, the mean of the RMSD to 3.70 ± 0.32 nm, the average of -content to 31 ± 11%, and the mean of the SASA to 45.73 ± 4.53 nm2 The corresponding metrics of the hexamer with the presence of EGCG are found to be significantly shifted, including the Rg with the mean value of 1.15 ± 0.16 nm, the mean RMSD with an amount of 2.91 ± 0.15 nm, the mean of -content of 26 ± 11%, and a slightly larger SASA value of 45.98 ± 3.14 nm2 In addition, a comparison the data of two systems obtained in the different case reveals that the radius of gyration of the hexamer is much reduced when EGCG is induced, even though, as we can see in Fig 3, the distribution of Rg attains the maximum peak at the same time in both cases However, the maximum of distribution of system without EGCG is consistently smaller, since numerous values of radius of gyration are diffused in the range from 1.5 to 2.5 nm It is known that the high gyration radius (Rg > 1.5 nm) corresponds to the A conformation, which exists with at least one chain being non-touched to the rest (for example in Fig S4) Comparison of the distributions of Rg indicates that the hexamer with the presence of EGCG is more compact, but it is contrasted with the finding S.T Ngo et al / Journal of Molecular Graphics and Modelling 76 (2017) 1–10 Fig The secondary structure of 48 replicas of the two solvated hexamer systems at 150 ns of REMD simulations The metrics are estimated through DSSP tool The red colour highlights the hexamer without EGCG and the blue represents the hexamer with the presence of an EGCG molecule (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Table The average of the secondary structure of the solvated A16-22 hexamer in the absence and presence of EGCG The values are in unit of percentage (%) Systems Beta Coil Turn Helix 6A16-22 6A16-22 + EGCG 31 ± 11 26 ± 11 66 ± 11 71 ± 10 2±4 3±4 1±1 0±1 over the representative conformations of both systems As stated above, the mean gyration radius of 6A16-22 + EGCG is smaller than that of the isolated system by an amount of 16% The significant larger value of RMSD of the isolated hexamer suggests that the isolated system is more fluctuating Because the isolated oligomer is less compact and more unstable during the simulations, it may be argued that the isolated hexamer is aggregated in “slower” rate compared to 6A16-22 + EGCG system In other words, the EGCG molecule is accelerating the aggregation of A peptides that the lifetime and/or the number of A oligomeric conformations are reduced The AD patient can thus be protected from A oligomers 3.2 Secondary structure of the Aˇ16-22 hexamer The secondary structure of the solvated A16-22 hexamer is shifted upon addition of EGCG and the effects of EGCG altering the 6A16-22 oligomer are clarified The averages of secondary structure terms are listed in Table As mentioned above, the average of -content of the hexamer interacting with EGCG is smaller than that of the hexamer without EGCG by an amount of 5% Let us mention again that in particular the content of  structure of isolated A16-22 hexamer is ∼31 ± 11%, whereas this value is of ∼26 ± 11% when the EGCG is incorporated Although these averages are overlapped with each other within the error margins, this overlap is caused by a structural change of both hexamers over the phase space The average of -content of distinct A16-22 monomer in isolated hexamer systems is significantly smaller than the metric of the one constituting the isolated A16-22 octamer, which is up to 44.5% [73] The variation of -contents is probably caused by the effects of the different force fields representing the A hexamer and octamer The hexamer considered in the present work is treated by the Amber99SB-ILDN force field [52], while the A16-22 octamer is represented by the GROMOS96 43a1 force field [78] The GROMOS96 43a1 force field adopts the richest -content as compared to the others [53] However, the Amber99SB-ILDN is demonstrated to provide the best ensemble that is in good agreement with experiment [53] Moreover, the averaged randomly coiled structure of isolated hexamer is of ∼66 ± 11% which is much smaller than the corresponding metric of the solvated 6A16-22 + EGCG with an amount of ∼71 ± 10% The turn structure of the isolated 6A16-22 peptide is ∼2 ± 4% as compared to ∼3 ± 4% of the 6A16-22 + EGCG system The helix structure of the isolated 6A16-22 peptide is ∼1 ± 1% which is in good agreement with previous investigations that the ␣-structure is an intermediate step of A oligomerization process [17,24,28,79,80] However, the helix structure of the A16-22 + EGCG is ∼0 ± 1% implying that the hexamer in the presence of EGCG is probably aggregated faster than itself when it exists in the isolated state Overall, the EGCG reduces both the -structure and intermediate structure of the A16-22 hexamer The secondary structures of the A oligomers per residue are determined and presented in Fig In average over all constituting monomers, residues 17–21 form rigid -content, and the highest probability is found at the residue F19 The residues 16 and 22 are extremely flexible with 100% of random coil structure Both turn and helix structures are found at position 18–20 of both hexamer systems In agreement with the analysis of the secondary structures of all systems given above, the secondary structures of all hexamer residues, which consist of -structure and ␣-structure, are significantly reduced when the EGCG compound is inserted S.T Ngo et al / Journal of Molecular Graphics and Modelling 76 (2017) 1–10 Fig The convergence of REMD simulations The dot line with magenta and black colours represent to the metrics estimated over time interval of 150–230 ns of the hexamer with and without EGCG, respectively The continuous line with blue and red colours represent to the distribution of the investigated values during simulation interval 150–250 ns of the oligomers with and without the presence of EGCG, respectively (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) into the solvated system Consequently, both the random coil and turn structure of the systems containing EGCG are increased in this region In addition, the probability of the maximum  size of the peptide is evaluated The results shown in Fig reveal that EGCG tends to enhance the -sheet size, constructed by the two A peptides, that is implied as a low neurotoxic dimeric form of A oligomer [27] However, the EGCG decreases the percentage of  sheets, which are formed by three, four, five, or six of A peptides, and this is related to the high cytotoxicity conformations [27] Fig The distribution of secondary structure per residue of two A16-22 hexamers with the presence (blue color) and absence (red color) of EGCG (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) in average, the EGCG forms more intermolecular contacts with the sequences 18–20 of the A16-22 oligomer due to the formation of the non-bonded contacts including - stacking interaction (cf Fig 6) from their aromatic rings to these residues The EGCG thus alters the contact maps of the hexamer, which will be further discussed in following sections The contact maps of the A16-22 hexamers are constructed as the same way of the previous study [81] and presented in Fig The results obtained from analysis of the contact maps of constituting chains of the solvate A16-22 hexamers are in good agreement with a previous study of the A16-22 octamers [82] Although the A16-22 octamer is treated by the GROMOS96 43a1 force field [78] in the earlier study, the interaction picture is similar The obtained contact maps imply that the contact F19–F19 is the most favorable one 3.3 Intermolecular contacts The intermolecular contacts between EGCG and the A16-22 hexamer are investigated by measurement of the distance between heavy atoms of EGCG compound to A peptides This contact is counted when the distance is getting smaller than 0.45 nm The ligand is separated from three domains to identify the important part of EGCG contributing to the binding (Fig S5 of SI) The probability of the intermolecular contact between three domains of EGCG to each residue of the hexamer is shown in Fig S6 (SI) The first domain of EGCG favors interaction with residues 18–21 of the peptide as compared to other domains The third domain interacts with residues K16, L17, and E22 with the highest probability Moreover, Fig The -sheet sizes of the solvated A oligomers in the presence (blue color) and absence (red color) of EGCG (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) S.T Ngo et al / Journal of Molecular Graphics and Modelling 76 (2017) 1–10 Fig Representative - stacking contacts between EGCG and A16-22 hexamer The result involves the conformation M3 of the 6A16-22 + EGCG system using the PLIP protocol [69] The interaction energy is evaluated using DFT calculations with the B97-D functional and the 6-311++G(d,p) basis set Fig Free energy surface (FES) of two solvated A16-22 systems (A) is represented for the FES of the A peptide without EGCG, and (B) is FES of the hexamer when EGCG is inserted is an important element contributing to the interaction of EGCG and A peptides It may be argued that the interaction between EGCG to A peptide is not completely recovered using a classical approach In addition, the high probability of non-bonded contacts between K16 and E22 indicates that the A16-22 hexamer mostly forms antiparallel structures This is consistent with findings in previous studies [28] 3.4 Free energy surface (FES) of the Aˇ16-22 hexamers Fig Contact maps between two monomers of isolated A16-22 hexamer (upper panel) and A16-22 + EGCG hexamer (lower panel) With the appearance of EGCG, the average of non-bonded contacts between constituting chains is obviously decreased The non-bonded contact between residues F19 and F19 is decreased because they form contact to the EGCG aromatic rings, and thus the -structure of the oligomer is reduced The typical - stacking contacts between EGCG and the A peptide are shown in Fig Although the dispersion effects on the interaction energy cannot be estimated using classical mechanics, the stacking interaction, however, is more properly investigated using DFT calculations (using the B97-D functional where D stands for dispersion) [70] The interaction energy between aromatic rings of the EGCG and A16-22 peptide is calculated using the same functional with the 6–311++G(d,p) basis set and counterpoise corrections [83], for which the modeling is shown in Fig S1 (SI) The results with counterpoise corrections are smaller than the previous ones as by an amount of ∼2.5 to 5.0 kJ/mol The terms are ranged from −12.3 to −18.2 kJ/mol, which are shown in Fig 6, indicating that -stacking The representative conformations of both solvated A hexamers are determined through a combination of free energy surface and clustering methods The GROMACS tool “sham” is employed to construct the FES of two solvated systems, whose reaction coordinates are the radius of gyration (Rg ) and the RMSD of simulated systems The obtained results are displayed in Fig (the FES with the same scale of the reaction coordinates is described in Fig S7) Fig 8A denotes the free energy surface of the isolated A16-22 hexamer in solvation, whereas Fig 8B is an illustration of the FES when the EGCG is introduced Actually, the hexamer without the presence of EGCG adopts more energy minima, that are noted from M1 to M5 in Fig 8A, in comparison with that with the presence of EGCG whose FES forms three minima, noted from M1 to M3 The lowest free energy of the EGCG-hexamer is approximately of −14 kJ/mol, whereas this term, in the absence of the inhibitor, amounts to −12 kJ/mol Overall, the solvated A peptide system is slightly more stable when the EGCG is introduced Subsequently, the clustering method is carried out with the C˛ RMSD cutoff of 0.1 nm to determine the representative conformations of the hexamers The structures of the A16-22 peptides corresponding to the FES minima are found and shown in Fig Details of these conformations are given in Table In average, the appearance of EGCG alters the size of the A16-22 hexamer, which is increased by ∼4% of Rg and ∼2% of CCS (Table 2) It can be argued that the solvated system becomes more voluminous when the inhibitor is added The EGCG molecule stabilizes the A oligomer structure The secondary structure of the 6A16-22 + EGCG is more stable with a considerably smaller error bar (cf Table 2), and this is consistent with a previous study about the soluble A1-42 S.T Ngo et al / Journal of Molecular Graphics and Modelling 76 (2017) 1–10 Fig Representative structures of the solvated A16-22 hexamer in the absence, noted from M1 to M5, and presence, noted from M1 to M3 of EGCG Values given in the brackets represent the populations of the corresponding conformations dimer under effects of EGCG [40] Moreover, the helix structure is observed in stable conformations of A16-22 hexamer but is not appeared in the optimized conformations of 6A16-22 + EGCG, thus suggesting that the EGCG accelerates the oligomerization process of A peptide since the helix structure is an intermediate step of the A aggregation [17,24,28,79,80] This argument is confirmed by the clustering results of both solvated A peptides Accordingly, the conformations of the 6A16-22 and 6A16-22 + EGCG can arrange to 4423 and 3219 clusters with the cutoff = 0.1 nm, respectively The values of 1237 and 579 correspond to the number of clusters of 6A16-22 and 6A16-22 + EGCG systems, respectively, when the evaluation is carried out with the C˛ RMSD cut-off of 0.2 nm There is no change with a clustering cut-off of 0.3 nm that we can separate the conformations of two sysTable Details of the representative structures of the A16-22 hexamer in the presence and absence of EGCG, which are predicted by FES and clustering methods Secondary structure is evaluated using DSSP tool; Rg is estimated utilizing GROMACS tool, and collision cross section (CCS) is estimated through IMPAC protocol Minima Beta (%) Coil (%) Turn (%) Helix (%) Rg (nm) CCS (Å2 ) M1 M2 M3 M4 M5 Average M1 M2 M3 Average 43 29 52 31 40 39 ± 36 38 43 39 ± 57 71 48 62 60 60 ± 59 62 57 59 ± 0 0 0±0 0 2±2 0 1±3 0 0±0 0.96 1.10 1.06 1.05 0.94 1.02 ± 0.06 1.02 1.05 1.10 1.06 ± 0.03 831 915 884 912 814 871 ± 42 834 876 941 884 ± 44 Fig 10 The population of the first 20th 6A16-22 clusters in present (blue color) and absent (red color) of the EGCG (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) tems without and with the presence of EGCG to 656 and 276 groups The smaller number of clusters characterizes the higher population of each group (Fig 9) Obviously, the conformational states of the hexamer are seriously decreased by the EGCG, the population of the representative structure of the 6A16-22 + EGCG is thus higher compared to the isolated one (Fig 10) As the -sheet size of the A oligomer is altered, the neurotoxicity of A oligomer may be modified When the optimized structures of two soluble systems are considered, the total -content of the oligomers remains unchanged when the EGCG is introduced in such a way that the metrics are the same amount of 39% for both situations However, as shown in Fig 9, the representative conformations of the A16-22 hexamer consist of four conformations with the -sheet sizes being larger than two, including the oligomers with high neurotoxic ability [27] Conversely, in the presence of EGCG, two of three representative conformations of the 6A16-22 + EGCG containing the -sheet sizes are getting smaller S.T Ngo et al / Journal of Molecular Graphics and Modelling 76 (2017) 1–10 Table Binding free energy of EGCG to A peptides provided using double-annihilation binding free energy method (kJ/mol) System Ecou EvdW Gbind EGCG + 6A16-22 Curcumin + 12A11-40 a −47.2 13.4 −58.9 −104.9 −106.2 ± 7.3 −91.5 ± 18.0 a Taken from Ref [22] indicating that it may relate to the low cytotoxic form [27], due to the fact that the -sheet is separated by the EGCG (conformations M2 and M3 in Fig 9) Overall, it can be concluded that the EGCG reduces the high order conformation of A oligomer in reducing the number of conformations established by high order oligomers [84] This view is in good agreement with the analysis of the -sheet size of all snapshots over the REMD simulations described above 3.5 The binding free energy of EGCG to the hexamer The binding free energy provides us with the nature of binding of constituting molecules of a complex system Many approaches have recently been proposed to resolve this problem such as molecular docking [85], quantitative structure-activity relationship [86], linear interaction energy [87], molecular mechanics-PoissonBoltzmann surface area [88], fast pulling of ligand [89–91], thermodynamics integration [92], and free energy perturbation [60] methods The free energy can be determined through the doubleannihilation binding free energy method [60,61], which remains one of the most accurate approaches to this property The representative structures of the hexamers with and without inhibitor located as the lowest free energy minima are used as initial conformations of free energy perturbation determination The binding free energy of the constituting monomers to the others is evaluated for the first time to clarify the effects of EGCG on the structure of the A16-22 hexamer In this scheme, the isolated monomer is decomposed two times from the solvated complexes and solvated monomer systems, and then the free energy difference of two processes is evaluated as being equal to the binding free energy of the isolated monomer to the others Totally, the A16-22 monomer is annihilated 12 times from the solvated hexamer with and without EGCG However, the decomposition of the monomer from solvated A16-22 monomer is required just one time due to the uniform peptide There are 17 values of the coupling parameter , which are used to remove the non-bonded interactions between the distinguished A16-22 monomer with the surrounding atoms Each free energy calculation is of ns long In total 208 ns of MD simulations are performed to evaluate the average of binding free energy of the isolated monomers to the others [22,61], The free energy terms is determined from 500 to 1000 ps of MD simulations to avoid the initial fluctuations due to altered non-bonded interactions Therefore, the final binding free energy is computed through the difference of mean free energies after 500 ps Overall, the binding free energy is calculated to be −106.2 ± 7.3 kJ/mol (Table 3) This value is larger than the corresponding term of curcumin to the 12A9-40 peptide of −91.5 ± 18.0 kJ/mol obtained from the same method in a previous study [22] Both EGCG and curcumin are under clinical trials as potential inhibitors preventing the A oligomers formation, as their binding free energies to A peptides are roughly ranged in the same order of magnitude Particularly, the electrostatic interaction energy contribution to the binding affinity amounted to −58.9 ± 6.7 kJ/mol and the vdW term is slightly larger at −47.2 ± 2.9 kJ/mol It is consistent with the hydrogen bonded and non-bonded protein-ligand contacts (Fig S8 of SI), in which EGCG forms hydrogen bond to ∼4.5 residues of the hexamer Consequently, the hexamer had ∼21.6 residues forming non-bonded contacts to the ligand A hydrogen bonded interaction may be about ten times as strong as a nonbonded contact These results are significantly different from a previous component analysis of binding free energies between curcumin and A peptides, which are determined to be −104.9 and 13.4 kJ/mol corresponding to vdW and electrostatic energies, respectively [22] Obviously, in contrast to the nature of binding of curcumin to A peptides, the electrostatic contribution is found to be more important than the vdW counterpart in the binding process of EGCG to A peptides Conclusions Through more than 24,000 ns of MD simulations, we demonstrated in the present theoretical study the EGCG effects in inhibiting the oligomerization of the A16-22 hexamer in solution, along with a physical insight into its ability to enhance the AD therapy Overall, it has been found that EGCG can prevent the A oligomeric conformations through different ways either by an acceleration of the A oligomerization by a reduction of the content of the hexamer From a FES analysis, the A peptides find more local minima with larger free energy barriers connecting them, when the EGCG is inserted In addition, through interaction with the peptides, EGCG tends to accelerate the oligomerization process and to stabilize the conformations of the A oligomer in stabilizing the secondary structure of the A peptide, and reduce the number of clusters and intermediate conformations of the solvated A16-22 hexamer The former finding is in good agreement with available experiment [93] The presence of EGCG decreases the -sheet sizes with an order higher than two of the oligomers Although the interaction picture is now clarified, it may be altered when the full-length A40/42 can be considered Furthermore, the molecule can be inserted itself into the A16-22 hexamer It increases the size of the representative parameters including the Rg and CCS, and alters the non-bonded contact maps of the A peptides through formation of the nonbonded interactions to the constituent monomers Quantum chemical (DFT) calculations indicate that the stacking plays an important role in the interaction between EGCG and A16-22 peptides The binding free energy determination points out that the binding affinity of EGCG to A peptides is roughly at the same order of magnitude as those with other available inhibitors such as curcumin However, the EGCG molecule induces both electrostatic and van der Waals (vdW) interactions, almost with similar amount to the peptide, whereas curcumin favors rather vdW interactions [22] Overall, the present study suggests the use of EGCG in the AD treatment Acknowledgements This work is supported by Department of Science and Technology and Institute for Computational Science & Technology (ICST) at Ho Chi Minh City, Vietnam, under the grant number 186/2016/HDSKHCN References [1] A.S Henderson, A.F Jorm, Dementia, John Wiley & Sons Ltd, 2002 [2] D.J Selkoe, The molecular pathology of Alzheimer’s disease, Neuron (1991) 487–498 [3] J.L Cummings, Alzheimer’s disease, N Engl J Med 351 (2004) 56–67 [4] H.W Querfurth, F.M LaFerla, Alzheimer’s disease, N Engl J Med 362 (2010) 329–344 [5] Alzheimer’s association, Alzheimer’s Disease Facts and Figs., 2015 S.T Ngo et al / Journal of Molecular Graphics and Modelling 76 (2017) 1–10 [6] J Hardy, D.J Selkoe, Medicine – the amyloid hypothesis of Alzheimer’s disease: progress and problems on the road to therapeutics, Science 297 (2002) 353–356 [7] M Citron, Strategies for disease modification in Alzheimer’s disease, Nat Rev Neurosci (2004) 677–685 [8] A Aguzzi, T O’Connor, Protein aggregation diseases: pathogenicity and therapeutic perspectives, Nat Rev Drugs (2010) 237–248 [9] D.J Selkoe, J Hardy, The amyloid hypothesis of Alzheimer’s disease at 25 years, EMBO Mol Med (2016) 595–608 [10] D.M Walsh, D.J Selkoe, A oligomers——a decade of discovery, J Neurochem 101 (2007) 1172–1184 [11] M Bucciantini, E Giannoni, F Chiti, F Baroni, et al., Inherent toxicity of aggregates implies a common mechanism for protein misfolding diseases, Nature 416 (2002) 507–511 [12] A Quist, I Doudevski, H Lin, R Azimova, et al., Amyloid ion channels: a common structural link for protein-misfolding disease, Proc Natl Acad Sci U S A 102 (2005) 10427–10432 [13] T.L Williams, L.C Serpell, Membrane and surface interactions of Alzheimer’s A peptide–insights into the mechanism of cytotoxicity, FEBS J 278 (2011) 3905–3917 [14] Y Nakazawa, Y Suzuki, M.P Williamson, H Saitô, et al., The interaction of amyloid A(1-40) with lipid bilayers and ganglioside as studied by 31P solid-state NMR, Chem Phys Lipids 158 (2009) 54–60 [15] H Jang, J Zheng, R Nussinov, Models of -amyloid ion channels in the membrane suggest that channel formation in the bilayer is a dynamic process, Biophys J 93 (2007) 1938–1949 [16] G Bitan, M.D Kirkitadze, A Lomakin, S.S Vollers, et al., Amyloid -protein (A) assembly: A40 and A42 oligomerize through distinct pathways, Proc Natl Acad Sci U S A 100 (2003) 330–335 [17] M.D Kirkitadze, M.M Condron, D.B Teplow, Identification and characterization of key kinetic intermediates in amyloid -protein fibrillogenesis, J Mol Biol 312 (2001) 1103–1119 [18] S.T Ngo, H.M Hung, D.T Truong, M.T Nguyen, Replica exchange molecular dynamics study of the truncated amyloid beta (11-40) trimer in solution, Phys Chem Chem Phys 19 (2017) 1909–1919 [19] S.T Ngo, H.M Hung, K.N Tran, M.T Nguyen, Replica exchange molecular dynamics study of the amyloid beta (11–40) trimer penetrating a membrane, RSC Adv (2017) 7346–7357 [20] S.T Ngo, M.S Li, Curcumin binds to abeta1-40 peptides and fibrils stronger than ibuprofen and naproxen, J Phys Chem B 116 (2012) 10165–10175 [21] S.T Ngo, M.S Li, Top-leads from natural products for treatment of Alzheimer’s disease: docking and molecular dynamics study, Mol Sim 39 (2013) 279–291 [22] S.T Ngo, S.-T Fang, S.-H Huang, C.-L Chou, et al., Anti-arrhythmic medication propafenone a potential drug for Alzheimer’s disease inhibiting aggregation of A: in silico and in vitro studies, J Chem Inf Model 56 (2016) 1344–1356 [23] G Wen, W Qin, D Chen, Y Wang, et al., Stabilizing the monomeric amyloid-[small beta] peptide by tyrocidine A prevents and reverses amyloidogenesis without the accumulation of oligomers, Chem Com (2017) [24] S Jang, S Shin, Computational study on the structural diversity of amyloid beta peptide A(10-35) oligomers, J Phys Chem B 112 (2008) 3479–3484 [25] M.H Viet, P.H Nguyen, S.T Ngo, M.S Li, et al., Effect of the tottori familial disease mutation (D7N) on the monomers and dimers of abeta40 and abeta42, ACS Chem Neurosci (2013) 1446–1457 [26] A Yano, A Okamoto, K Nomura, S.i Higai, et al., Difference in dimer conformation between amyloid-(1-42) and (1-43) proteins: replica exchange molecular dynamics simulations in water, Chem Phys Lett 595–596 (2014) 242–249 [27] Y Sun, W Xi, G Wei, Atomic-level study of the effects of O4 molecules on the structural properties of protofibrillar A trimer: -sheet stabilization, salt bridge protection, and binding mechanism, J Phys Chem B 119 (2015) 2786–2794 [28] D.K Klimov, D Thirumalai, Dissecting the assembly of A beta(16–22) amyloid peptides into antiparallel beta sheets, Structure 11 (2003) 295–307 [29] C Wu, Z.X Wang, H.X Lei, Y Duan, et al., The binding of Thioflavin T and its neutral analog BTA-1 to protofibrils of the Alzheimer’s disease A beta(16–22) peptide probed by molecular dynamics simulations, J Mol Biol 384 (2008) 718–729 [30] J Bieschke, M Herbst, T Wiglenda, R.P Friedrich, et al., Small-molecule conversion of toxic oligomers to nontoxic -sheet-rich amyloid fibrils, Nat Chem Biol (2012) 93–101 [31] G Lim, T Chu, F Yang, W Beech, et al., The curry spice curcumin reduces oxidative damage and amyloid pathology in an Alzheimer transgenic mouse, J Neurosci 21 (2001) 8370–8377 [32] L.O Tjerngber, J Naslund, F Lindqvist, J Jahannson, et al., Arrest of beta-amyloid beta-peptide fibril formation by a pentapeptide, J Biol Chem 271 (1996) 8545–8548 [33] C Wu, M.M Murray, S.L.B.L Summer, M.M Condron, et al., The structure of A beta 42 C-terminal fragments probed by a combined experimental and theoretical study, J Mol Biol 387 (2009) 492–501 [34] H.Y Li, M.B Monien, A Lomakin, R Zemel, et al., Mechanistic investigation of the inhibition of A beta 42 assembly and neurotoxicity by A beta 42 C-terminal, Biochemistry 49 (2010) 6358–6364 [35] J Wang, K Ono, D.L Dickstein, I Arrieta-Cruz, et al., Carvedilol as a potential novel agent for the treatment of Alzheimer’s disease, Neurobiol Aging 32 (2321) (2011) e1–e12 [36] P.B Madrid, S Chopra, I.D Manger, L Gilfillan, et al., A systematic screen of FDA-approved drugs for inhibitors of biological threat agents, PLoS One (2013) e60579 [37] D Paris, A Quadros, J Humphrey, N Patel, et al., Nilvadipine antagonizes both A vasoactivity in isolated arteries, and the reduced cerebral blood flow in APPsw transgenic mice, Brain Res 999 (2004) 53–61 [38] E Gazit, A possible role for -stacking in the self-assembly of amyloid fibrils, FASEB J 16 (2002) 77–83 [39] A.V Kristen, S Lehrke, S Buss, D Mereles, et al., Green tea halts progression of cardiac transthyretin amyloidosis: an observational report, Clin Res Cardiol 101 (2012) 805–813 [40] T Zhang, J Zhang, P Derreumaux, Y Mu, Molecular mechanism of the inhibition of EGCG on the Alzheimer A1-42 dimer, J Phys Chem B 117 (2013) 3993–4002 [41] S.-H Wang, F.-F Liu, X.-Y Dong, Y Sun, Thermodynamic analysis of the molecular interactions between amyloid -peptide 42 and (−)-epigallocatechin-3-gallate, J Phys Chem B 114 (2010) 11576–11583 [42] F.-F Liu, X.-Y Dong, L He, A.P.J Middelberg, et al., Molecular insight into conformational transition of amyloid -Peptide 42 inhibited by (−)-epigallocatechin-3-gallate probed by molecular simulations, J Phys Chem B 115 (2011) 11879–11887 [43] Y Chebaro, P Jiang, T Zang, Y Mu, et al., Structures of A17-42 trimers in isolation and with five small-Molecule drugs using a hierarchical computational procedure, J Phys Chem B 116 (2012) 8412–8422 [44] M Miyata, T Sato, M Kugimiya, M Sho, et al., The crystal structure of the green tea polyphenol (−)-epigallocatechin gallate-transthyretin complex reveals a novel binding site distinct from the thyroxine binding site, Biochemistry 49 (2010) 6104–6114 [45] F.L Palhano, J Lee, N.P Grimster, J.W Kelly, Toward the molecular mechanism(s) by which EGCG treatment remodels mature amyloid fibrils, J Am Chem Soc 135 (2013) 7503–7510 [46] A.E Pelaez-Aguilar, L Rivillas-Acevedo, L French-Pacheco, G Valdes-Garcia, et al., Inhibition of light chain 6aJL2-R24G amyloid fiber formation associated with light chain amyloidosis, Biochemistry 54 (2015) 4978–4986 [47] R Malishev, S Nandi, S Kolusheva, Y Levi-Kalisman, et al., Toxicity inhibitors protect lipid membranes from disruption by A42, ACS Chem Neurosci (2015) 1860–1869 [48] S.-H Wang, X.-Y Dong, Y Sun, Thermodynamic analysis of the molecular interactions between amyloid -protein fragments and (−)-epigallocatechin-3-gallate, J Phys Chem B 116 (2012) 5803–5809 [49] Sunphenon EGCg (Epigallocatechin-Gallate) in the early stage of Alzheimer´ıs disease (SUN-AK) Charite University, Berlin, Germany ClinicalTrials.gov [Internet], vol 2015, National Library of Medicine (US), Bethesda (MD), 2000 [50] L Schrödinger, The PyMOL Molecular Graphics System, Version 1.5.0.4, Schrödinger, LLC, 2010 [51] W Humphrey, A Dalke, K Schulten, VMD: visual molecular dynamics, J Mol Graph 14 (1996) [52] A.E Aliev, M Kulke, H.S Khaneja, V Chudasama, et al., Motional timescale predictions by molecular dynamics simulations: case study using proline and hydroxyproline sidechain dynamics, Proteins: Struct Funct Bioinform 82 (2014) 195–215 [53] A.K Somavarapu, K.P Kepp, The dependence of amyloid- dynamics on protein force fields and water models, ChemPhysChem 16 (2015) 3278–3289 [54] Y Wang, J Xiao, T.O Suzek, J Zhang, et al., PubChem’s BioAssay database, Nucleic Acids Res 40 (2012) D400–D412 [55] J Wang, R.M Wolf, J.W Caldwell, P.A Kollman, et al., Development and testing of a general amber force field, J Comput Chem 25 (2004) [56] A.W Sousa da Silva, W.F Vranken, ACPYPE – AnteChamber PYthon parser interfacE, BMC Res Notes (2012) 1–8 [57] W.L Jorgensen, J Chandrasekhar, J.D Madura, R.W Impey, et al., Comparison of simple potential functions for simulating liquid water, J Chem Phys 79 (1983) 926–935 [58] M.J Abraham, T Murtola, R Schulz, S Páll, et al., GROMACS: high performance molecular simulations through multi-level parallelism from laptops to supercomputers, SoftwareX 1–2 (2015) 19–25 [59] A Patriksson, D.A van der Spoel, Temperature predictor for parallel tempering simulations, Phys Chem Chem Phys 10 (2008) 2073–2077 [60] R.W Zwanzig, High-Temperature equation of state by a perturbation method I Nonpolar gases, J Chem Phys 22 (1954) 1420–1426 [61] S.T Ngo, B.K Mai, D.M Hiep, M.S Li, Estimation of the binding free energy of AC1NX476 to HIV-1 protease wild type and mutations using free energy perturbation method, Chem Biol Drugs Des 86 (2015) 546–558 [62] C.H Bennett, Efficient estimation of free energy differences from Monte Carlo data, J Comput Phys 22 (1976) 245–268 [63] W.G Touw, C Baakman, J Black, T.A.H te Beek, et al., A series of PDB-related databanks for everyday needs, Nucleic Acids Res 43 (2015) D364–D368 [64] Y Mu, P.H Nguyen, G Stock, Energy landscape of a small peptide revealed by dihedral angle principal component analysis, Proteins: Struct Funct Bioinform 58 (2005) 45–52 [65] E Papaleo, P Mereghetti, P Fantucci, R Grandori, et al., Free-Energy landscape, principal component analysis, and structural clustering to identify representative conformations from molecular dynamics simulations: the myoglobin case, J Mol Graph Model 27 (2009) 889–899 [66] X Daura, K Gademann, B Jaun, D Seebach, et al., Peptide folding: when simulation meets experiment, Angew Chem Int Ed 38 (1999) 236–240 10 S.T Ngo et al / Journal of Molecular Graphics and Modelling 76 (2017) 1–10 [67] E Papaleo, P Mereghetti, P Fantucci, R Grandori, et al., Free-energy landscape, principal component analysis, and structural clustering to identify representative conformations from molecular dynamics simulations: the myoglobin case, J Mol Graph Model 27 (2009) 889–899 [68] G Marklund Erik, T Degiacomi Matteo, V Robinson Carol, J Baldwin Andrew, et al., Collision cross sections for structural proteomics, Structure 23 (2015) 791–799 [69] S Salentin, S Schreiber, V.J Haupt, M.F Adasme, et al., PLIP: fully automated protein-ligand interaction profiler, Nucleic Acids Res 43 (2015) W443–W447 [70] Y An, J.W.G Bloom, S.E Wheeler, Quantifying the -stacking interactions in nitroarene binding sites of proteins, J Phys Chem B 119 (2015) 14441–14450 [71] M.J Frisch, G.W Trucks, H.B Schlegel, G.E Scuseria, et al., Gaussian∼09 Revision E.01, Gaussian Inc., Wallingford CT, 2009, pp 2016 [72] M.H Viet, S.T Ngo, N.S Lam, M.S Li, Inhibition of aggregation of amyloid peptides by beta-sheet breaker peptides and their binding affinity, J Phys Chem B 115 (2011) 7433–7446 [73] L Xie, Y Luo, D Lin, W Xi, et al., The molecular mechanism of fullerene-inhibited aggregation of Alzheimer’s beta-amyloid peptide fragment, Nanoscale (2014) 9752–9762 [74] Y Sugita, Y Okamoto, Replica-exchange molecular dynamics method for protein folding, Chem Phys Lett 314 (1999) 141–151 [75] C.H Davis, M.L Berkowitz, Structure of the amyloid- (1-42) monomer absorbed to model phospholipid bilayers: a molecular dynamics study, J Phys Chem B 113 (2009) 14480–14486 [76] A Morriss-Andrews, J.E Shea, Computational studies of protein aggregation: methods and applications, in: M.A Johnson, T.J Martinez (Eds.), Annu Rev Phys Chem., Vol 66, 2015, pp 643–666 [77] L Tran, N Basdevant, C Prévost, T Ha-Duong, Structure of ring-shaped A42 oligomers determined by conformational selection, Sci Rep (2016) 21429 [78] W.F van Gunsteren, S.R Billeter, A.A Eising, P.H Hunenberger, et al., Biomolecular Simulation: The GROMOS96 Manual and User Guide, Vdf Hochschulverlag AG an der ETH, Zurich, 1996 [79] S Jang, S Shin, Amyloid -peptide oligomerization in silico: dimer and trimer, J Phys Chem B 110 (2006) 1955–1958 [80] Y Fezoui, D.B Teplow, Kinetic studies of amyloid -protein fibril assembly: differential effects of -helix stabilization, J Biol Chem 277 (2002) 36948–36954 [81] H.L Chiang, S.T Ngo, C.J Chen, C.K Hu, et al., Oligomerization of peptides LVEALYL and RGFFYT and their binding affinity to insulin, PLoS One (2013) e65358 [82] L Xie, Y Luo, G Wei, A(16–22) peptides can assemble into ordered -barrels and bilayer -sheets, while substitution of phenylalanine 19 by tryptophan increases the population of disordered aggregates, J Phys Chem B 117 (2013) 10149–10160 [83] F.B van Duijneveldt, J.G.C.M van Duijneveldt-van de Rijdt, J.H van Lenthe, State of the art in counterpoise theory, Chem Rev 94 (1994) 1873–1885 [84] P Jiang, W Li, J.-E Shea, Y Mu, Resveratrol inhibits the formation of multiple-layered -sheet oligomers of the human islet amyloid polypeptide segment 22-27, Biophys J 100 (2011) 1550–1558 [85] G.M Morris, R Huey, W Lindstrom, M.F Sanner, et al., AutoDock4 and AutoDockTools4: automated docking with selective receptor flexibility, J Comput Chem 30 (2009) 2785–2791 [86] H.H Maw, L.H Hall, E-State modeling of HIV-1 protease inhibitor binding independent of 3D information, J Chem Inf Comput Sci 42 (2002) 290–298 [87] J Aqvist, C Medina, J.-E Samuelsson, A new method for predicting binding affinity in computer-aidded drug design, Protein Eng (1994) 385–391 [88] J Srinivasan, T.E Cheatham, P Cieplak, P.A Kollman, et al., Continuum solvent studies of the stability of DNA, RNA, and phosphoramidate-DNA helices, J Am Chem Soc 120 (1998) 9401–9409 [89] S.T Ngo, H.M Hung, M.T Nguyen, Fast and accurate determination of the relative binding affinities of small compounds to HIV-1 protease using non-equilibrium work, J Comput Chem 37 (2016) 2734–2742 [90] D.T Truong, M.T Nguyen, V.V Vu, S.T Ngo, Fast pulling of ligand approach for the design of -secretase inhibitors, Chem Phys Lett 671 (2017) 142–146 [91] S.T Ngo, M.T Nguyen, M.T Nguyen, Determination of the absolute binding free energies of HIV-1 protease inhibitors using non-equilibrium molecular dynamics simulations, Chem Phys Lett 676 (2017) 12–17 [92] J.G Kirkwood, Statistical mechanics of fluid mixtures, J Chem Phys (1935) 300–313 [93] D.E Ehrnhoefer, J Bieschke, A Boeddrich, M Herbst, et al., EGCG redirects amyloidogenic polypeptides into unstructured, off-pathway oligomers, Nat Struct Mol Biol 15 (2008) 558–566