Discovery of novel human intestinal maltase i nhibitors on WISDOM - PowerPoint PPT Presentation

Discovery of novel human intestinal maltase i nhibitors on WISDOM environment and in vitro confirmation Hwa-Ja Ryu, Hanh Thi Thanh Nguyen, Sehoon Lee, Soon wook Hwang, Ana Lucia Dacosta, Vincent Breton, Doman Kim

Phases of a pharmaceutical development Virtual screening: Use of high-performance computing to analyse large databases of chemical compounds in order t o identify possible candidates

Virtual Screening  Computational methods – Pharmacophore based search – Structure based docking  Requirements – 3D Structure of target – Databases of small molecules – A method to dock and score bound small molecule 3

What is docking? Protein + Inhibitor RECEPTOR Protein + Potential inhibitors LIGAND Future drug ?

Strategy of drug design Grid Autodock 3.0 Molecular docking 500 000 Complex Chimera visualization 329 in vitro Wet 50 Laboratory in vivo 10

High throughput virtual docking Millions of chemical compounds available in laboratories Chemical compounds : High Throughput Screening Chembridge – 500,000 1-10$/compound, nearly impossible Molecular docking (Autodock) Computational data challenge Targets : Hits screening Leads Human maltase glucoamylase using assays Clinical testing (2QMJ) performed on Drug living cells

Large Computational resources Grid computing is applying the resources of many computers in a network to a single pr oblem at the same time - usually to a scientific or technical problem that requires a g reat number of computer processing cycles or access to large amounts of data. Example : “EGEE (Enabling Grids for E- sciencE) is providing a produc tion quality grid infrastructure spanning more than 30 countri es with over 150 sites” 7

Grid-Enabled Virtual screening • The grid infrastructure allow to deploy large scale computer-based in-silico scre ening : – Docking programs are often restricted by the computational time, due to the eno rmous number of possibilities that should be examined; – More computational time = more accuracy for the scoring function – Efficiency is especially required for drug design • WISDOM initiative aims to demonstrate the relevance and the impact of the gri d approach to address drug discovery for neglected and emerging diseases. Use of GRID COMPUTING to speed up the whole process

Database of small molecules • Drug-like: MDDR (MDL Drug Data Report) >147,000 compounds, CMC (Comp rehensive Medicinal Chemistry) >8,600 compounds • Non-drug-like: ACD (Available Chemicals Directory) ~3 millions compounds • CSD (Cambridge Structural Database, www.ccdc.cam.ac.uk): 264,000 compound s • Corporates Databases: few millions in pharmaceuticals companies • Virtual libraries (Combinatorial chemistry) • ZINC, a free database of commercially-available compounds for virtu al screening. ZINC contains over 4.6 million compounds in ready-to- dock, 3D formats Chembridge database (www.chembridge.com): 454,000 compound • s 9

Finding Inhibitors of Human Intestinal Maltase

Human Intestinal Maltase (HMA)  α -glucosidase in the brush border of the small intestines responsible for digestion of maltose oligosaccharides into glucose  Inhibition of the enzyme activity → retardation of glucose absorption → decrease in postprandial blood glucose level  Important target in treatment of diabetes type 2 and obesity  α -glucosidase inhibitors – Acarbose (Glucobay), Miglitol (Glyset), Voglibose (Voglib) with side-effects  Need to discover alternative inhibitors with greater potency and fewer side-effects

Binding information of acarbose with human maltase Sim L et. al. 2008J Mol Biol. 375(3):782-92

Filtration process 454,000 chemical compounds from Chembridge Scoring based on docking score ( 308,307) 3016 compounds select ed Interaction with key residues 2616 compounds selected Key interactions binding models clustering 42 compoun d selected In vitro test

Statistics of datachallenge deployment on WISDOM production environment Total numbers of docking ¡ 308,307 ¡ Total size of output results ¡ 16.3 GBytes ¡ Estimated duration by 1CPU ¡ 22.4 years ¡ Duration of experiments ¡ 3.2 days ¡ Maximum numbers of concurrent CPUs ¡ 4700 CPUs ¡ Crunching Factor ¡ 2556 ¡ Distribution Efficiency ¡ 54.4 % ¡

Hydrogen bond interactions with Key residues of two hit compounds in active site of protein 18 17 In active site of HMA 17 18 Hydrogen bonding of 2 hit compounds

Cloning and expression of human maltase in Pichia pastoris C ontro l ¡ PCR ¡ Enzyme activity M ¡ ¡ ¡ ¡ ¡P ¡ 2.7Kb ¡ 0 ¡ ¡ ¡24 ¡40 ¡ ¡48 ¡96h ¡Glc ¡ ¡ ¡ 0 ¡ ¡ ¡24 ¡ ¡40 ¡ ¡ ¡48 ¡96h ¡ ¡ ¡ • Conditions for HMA expression → Culture 500 ml in 2 L flask at 30 ℃ and 200 rpm → 0.5% methanol → ~4 days → enzyme reaction : 90 min at 37 ℃ (50 mM maltose)

Enzyme Kinetics Analysis

Inhibitory activity of the indentified hits with HMA Compound Lowest M.W clogP Ki IC50 Type of No. energy ( µ M) inhibition (g/mol) (µM) 19.8 17 -16.43 473 3.04 58±4 competitive ±1.2 19.6± 18 -16.44 429 3.56 55±3 competitive 0.9 Acarbose -12.62 645.605 -6.655 19.4 52±4 competitive

Docking experiment of two hit compounds with human pancreatic α -amylase Active site pocket Human pancreatic α -amylase PDB ID: 1XCX Number Name of Binding energy compounds (kcal/mol) 1 IAB -15.69 2 7007617 -12.99 Active site poc ket 3 7002209 -12.89 Acarbose ¡ ¡ 17 ¡ 18 ¡

Inhibitory activities of the identified 2 hits on human pancreatic α -amylase

Conclusions  Identification ¡of ¡novel ¡HMA ¡inhibitors ¡through ¡structure-­‑based ¡virtual ¡ screening ¡ ¡  After ¡ ¡datachallenge ¡of ¡308,307 ¡compounds, ¡ ¡42 ¡compounds ¡were ¡selec ted ¡for ¡ in ¡vitro ¡ inhibition ¡assay. ¡  Inhibitory ¡activities ¡of ¡compound ¡No.17 ¡and ¡18 ¡showed ¡a ¡good ¡inhibiti on ¡comparable ¡to ¡that ¡of ¡acarbose. ¡  In ¡contrast ¡to ¡acarbose, ¡the ¡potent ¡inhibitors ¡show ¡no ¡inhibition ¡of ¡hu man ¡pancreatic ¡α-­‑amylase. ¡It ¡may ¡overcome ¡the ¡side-­‑effects ¡of ¡acarbo se. ¡  Cytotoxicity ¡will ¡be ¡examined. ¡ ¡

Acknowledgements