Abstract Book - RACI Centenary Congress | RACI … Book 1 Structure ... This project has received...

Abstract Book

1

Structure-based Precision Design: Discovery of Potent and Selective Inhibitors of Cdc2-like Kinase 1 as Autophagy Inducers

Shengyong Yang, Qizheng Sun, Guifeng Lin, Linli Li, Xiting Jin, Luyi Huang

Presenting author’s e-mail: [email protected]

State Key Laboratory of Biotherapy and Cancer Center, West China Hospital, and Collaborative Innovation Center for Biotherapy, Sichuan University. Chengdu, Sichuan,

China, 610041.

Autophagy inducers represent new promising agents for the treatment of a wide range of medical illnesses including neurodegenerative, and drug-induced organ injury. However, just very few autophagy inducers have been reported, and particularly there is a lack of safe agents for clinical application. Recent studies have demonstrated that inhibition of the cdc2-like kinase 1 (CLK1) could specifically and efficiently induce autophagy. However, as far as we know, all of the known CLK1 inhibitors belong to the miscellaneous type, which means a poor kinase selectivity and being not suitable for clinical use. To discover new potent and selective CLK1 inhibitors, we, in this investigation, utilized a structure-based precision design strategy, which involves highly accurate virtual screening and structural optimization based on the co-crystal structures of CLK1-ligand complexes. A number of new CLK1 inhibitors were obtained and the most potent and selective one is compound 25, which showed an IC50 value of 2 nM against CLK1. It exhibited very weak or no activity against majority of other kinases. The crystal structure of CLK1 in complex with 25 was resolved, with which the potency and kinase selectivity of 25 are well interpreted. In in vitro assays, 25 showed strong ability to induce autophagy. In vivo, it displayed significant hepatoprotective effect in the acetaminophen (APAP)-induced liver injury mouse model. Collectively, this study provides so far the most potent and selective CLK1 inhibitor, which could be used as a chemical probe or agent in future mechanism of action or disease therapy studies.

2

Structure-based Approaches to THe DISCOVERY of Anti-trypanosomatid Folate pathway inhibitors

Wade, RC1,2,3

E-mail: [email protected]

1 Molecular and Cellular Modeling Group, Heidelberg Institute for Theoretical Studies (HITS), Heidelberg, Germany

2 Center for Molecular Biology (ZMBH), DKFZ-ZMBH Alliance, Heidelberg University, Heidelberg, Germany

3 Interdisciplinary Center for Scientific Computing (IWR), Heidelberg University, INF 368, 69120 Heidelberg, Germany

Trypanosomatids are parasitic species that cause serious human diseases such as leishmaniasis, Chagas' disease and sleeping sickness. Currently available treatments are hindered by such problems as side-effects, parasitic resistance and the high cost of medicines. There is therefore a strong need for improved medicines and new types of drugs. Here, I will discuss computational aspects of our structure-based strategies to targeting the trypanosomatid folate pathway protein, particularly pteridine reductase 1 (PTR1), which is involved in the salvage of pteridines in trypanosomatids, and does not have a human homologue. We have conducted a systematic comparative mapping of folate pathway on- and off-targets and evaluated the possibilities for multiple-target inhibition. We compared the structural, physico-chemical and dynamic [1] properties of the binding sites of parasitic and human enzymes to identify features to optimize inhibitor selectivity. Besides folate-like compounds, we identified inhibitors by optimizing compounds found by virtual screening, fragment-based design and screening of natural product libraries [2-4]. Crystal structures permitted validation of docking protocols and the development of a quantum chemical approach to ligand fragment scoring [5]. This project has received funding from the European Union’s Seventh Framework Programme for research, technological development and demonstration under grant agreement n°603240 (NMTrypI - New Medicine for Trypanosomatidic Infections). http://fp7-nmtrypi.eu/ [1] Stank, A et al. TRAPP webserver: predicting protein binding site flexibility and detecting transient binding pockets. Nucl. Acids Res (2017) gkx277. [2] Ferrari, S et al. Virtual Screening Identification of Nonfolate Compounds, Including a CNS Drug, as Antiparasitic Agents Inhibiting Pteridine Reductase. J. Med. Chem. (2011) 54, 211-221. [3] Borsari, C et al. Profiling of flavonol derivatives for the development of antitrypanosomatidic drugs. J. Med. Chem., (2016) 59:7598–7616. [4] Di Pisa, F et al. Chroman-4-One Derivatives Targeting Pteridine Reductase 1 and Showing Anti-Parasitic Activity. Molecules, (2017) 22: 426. [5] Jedwabny W, Panecka J, Dyguda-Kazimierowicz E, Wade RC, Sokalski WA. Application of a simple quantum chemical approach to ligand fragment scoring for Trypanosoma brucei pteridine reductase 1 inhibition J. Comp. Aided Mol. Des. In press.

3

Knowledge Based Mass Spectroscopic Natural Product Research Platform, Flora Genesis System

Sangwon Lee1, Kibeom Shin1, Myungwon Seo1, Sungbo Hwang1, Hwan You2,

Kyoung Tai No1,2

1Department of Biotechnology, Yonsei University, Seoul 03722, Korea 2Bioinformaics & Molecular Design Research Center, Seoul 03722, Korea


Background and Motivations

Structure elucidation of natural product has been an extremely challenging and lengthy task for a long time. There were plenty of attempts to automate the structure elucidation process. Automated structure elucidation from spectral data was, however, based on similarity calculation using massive database which regards tandem mass spectrometry only with a limiting capability. The increasing availability and capability of machine learning algorithms, along with high-resolution tandem mass spectrometry, is enabling high-throughput label-free system for identification of known chemical entities. High resolution FT-ICR MS and MS/MS coupled with machine-learning algorithms have achieved remarkable advance. Detecting each precursor ion through high resolution MS/MS is now available and extremely exact molecular weight information is given for each precursor. Kernel-based learning algorithms revolutionized the applicability of database. In this context, we suggest the integrative pipeline for structure elucidation which name is Flora genesis (FG). FG pipeline contains MS/MS along with FT-ICR MS instrument. These instruments paired with natural compound focused library will accumulate knowledge, and accelerate the finding of novel natural products in cost effective way.

Methodology

To elucidate a structure of unknown, various instruments from each of partners are lined in the FG instrument pipeline to get higher performance. There are high resolution FT-ICR MS, LC-MSn. DB is inevitable in high throughput natural product identification process. Accumulation of natural product focused library would offer higher efficiency of screening for the Novel Chemical Entities.

4

Therapeutic Strategies for Treatment Of Flt3-Itd+ Aml ——From Drug Discovery to Precise Clinical Application: Opportunities, Challenges and Solutions

Qingsong Liu *

e-mail: [email protected]

High Magnetic Field Laboratory, Chinese Academy of Sciences, Anhui, Hefei, P.R.China

LT3-ITD mutant has been observed in about 30% AML patients and extensively studied as a drug discovery target. Based on the structure of PCI-32765 (Ibrutinib), a BTK kinase inhibitor that was recently reported to bear FLT3 kinase activity, through a structure-guided drug design approach, we have discovered novel FLT3 inhibitor CHMFL-FLT3-122, which displayed great in vitro/ in vivo efficacies in the preclinical models. In addition, the extensive preclinical drug like properties and safety evaluation have moved this compound to IND application. However, the preclinical drug sensitivity genomic (DSG) study in the in vitro “virtual clinical trial” revealed a series of problems for the precise clinical application of these target therapies. We will present some preliminary pharmacogenomic data and describe our primary thoughts for using HDGS technology as one of the solutions for the current clinical precise drug treatment.

5

“Enzyme Kinetics-Driven” Design, And Synthesis of Dc511020, One Nce Clinical Candidate for Anti- Alzheimer’s Disease(Ad) Treatment

Zhou, Y, Li, J, Liu, H

[email protected]

Shanghai Institute of Materia Medica, Chinese Academy of Science, Shanghai 201203, China

Alzheimer's disease (AD) is a devastating and progressive neurodegenerative disease. According to the information from “World Alzheimer Report 2016”, 47 million people currently live with dementia in worldwide, and the projected number will increase in further to more than 131 million by 2050. The global estimated cost of dementia will be a trillion dollar disease by 2018. Currently, cholinergic replacement therapy especially acetylcholinesterase inhibitors (AChEIs) are still the most popular and widely accepted drugs for AD therapy. DC511020 (DC20) is one of the next generation of AChE inhibitor with wide safety window, targeting to develop and become the New Chemical Entity for AD treatment. The hits and leads of novel AChE inhibitors was generated from an “enzyme kinetics-driven” drug design strategy in Shanghai Institute of Materia Medica (SIMM), Shanghai, China. With the systematic and comprehensive lead optimization and pharmaceutical profiling, compound DC20 was finally selected as clinical candidate for preclinical and clinical development. DC20 shows its very well drug-like properties and developability, highlights, but not limites, as: 1) Favorable enzymatic mechanism.DC20 possesses a novel kinetic enzymatic mechanism of high affinity and slow dissociation through introduction of fluorine atom, which might contribute to the safety window of keeping the high potency and reducing the side effects; 2) High potent and good drug-like properties. DC20 orally administration significantly improves the cognitive impairments of three AD animal models, including scopolamine or intracerebroventricular injection of Aβ-induced learning deficits models and APP/PS1 transgenic mouse model. DC20 shows high oral bioavailability and brain targeted distribution in both rodent and non-rodent animal; 3) Good safety profile and safety window; 4) Chemistry, Manufacture and Control.

6

Virtual Screening Approaches for Identification of Novel Scaffolds for Cypd and SETDB1/ESET Inhibitors as Potential Treatments for Neurodegenerative Diseases

Ae Nim Pae

Convergence Research Center for Dementia, Korea Institute of Science and Technology, 5, Hwarang-ro 14-gil, Seongbuk-gu, Seoul 02792, Korea


Cyclophilin D (CypD) is a mitochondria-specific cyclophilin that is known to play a pivotal role in the formation of the mitochondrial permeability transition pore (mPTP).The formation and opening of the mPTP disrupt mitochondrial homeostasis, cause mitochondrial dysfunction and eventually lead to cell death. Several recent studies have found that CypD promotes the formation of the mPTP upon binding to β amyloid (Aβ) peptides inside brain mitochondria, suggesting that neuronal CypD has a potential to be a promising therapeutic target for Alzheimer’s disease (AD). In this study, we generated an energy-based pharmacophore model by using the crystal structure of CypD – cyclosporine A (CsA) complex, and performed virtual screening of ChemDiv database, which yielded forty-five potential hit compounds with novel scaffolds. We further tested those compounds using mitochondrial functional assays in neuronal cells and identified fifteen compounds with excellent protective effects against Aβ-induced mitochondrial dysfunction. We believe that this study offers new insights into the rational design of small molecule CypD inhibitors, and provides a promising lead for future therapeutic development. ERG-associated protein with SET domain (ESET/SET domain bifurcated 1/SETDB1/KMT1E) is a histone lysine methyltransferase (HKMT) and it preferentially tri-methylates lysine 9 of histone H3 (H3K9me3). SETDB1/ESET leads to heterochromatin condensation and epigenetic gene silencing. These functional changes are reported to correlate with Huntington’s disease (HD) progression and mood-related disorders which make SETDB1/ESET a viable drug target. The present investigation was performed to identify novel peptide-competitive small molecule inhibitors of the SETDB1/ESET by a combined in silico-in vitro approach.

7

Computer-aided Design of Inhibitors Targeting Human Sialyltransferases

Montgomery, A.P.1, Szabo, R.1, Yu, H.1,2, Skropeta, D.1,2,3


1School of Chemistry, University of Wollongong, NSW, Australia 2Centre for Medical and Molecular Bioscience, University of Wollongong, Wollongong,

NSW, Australia 3Illawarra Health and Medical Research Institute, University of Wollongong, NSW, Australia

Sialyltransferases (STs) catalyse the synthesis of sialylated glycoconjugates involved in cell-cell interactions. Overexpression of STs is observed in many different types of cancers and is thought to promote metastasis through altered sialylation patterns of the cell. A wide range of ST inhibitors have been developed based on the natural donor, CMP-Neu5Ac, as potential new antimetastatic agents (Szabo and Skropeta 2017). To improve selectivity, pharmacokinetic properties and overall ease of synthesis of these inhibitors, we have investigated the replacement of the charged phosphodiester linker present in many ST inhibitors with a neutral isostere such as a carbamate or triazole, in a combined computational and experimental study. Molecular docking, molecular dynamics simulations and binding free energy calculations have been undertaken with the human ST6Gal I crystal structure (Kuhn et al 2013). These calculations have indicated that compounds containing either carbamate or triazole linkers can potentially bind to human ST6Gal I comparable to their phosphodiester-linked counterparts, suggesting that these linkers are suitable neutral isosteres (Montgomery et al 2016). These findings are surprising as the phosphodiester linker was previously believed to be important for binding. Further analyses has indicated that there is a strong enthalpyentropy compensation contributing to these findings. The synthetic component of this study, includes the preparation of required αhydroxyphosphonate, alkyne and nucleoside synthons along with the successful coupling of these to generate the target compounds. Overall this work has laid the foundation for the further rational design of novel carbamate and triazole-linked ST inhibitors. Szabo, R.; Skropeta, D., Advancement of Sialyltransferase Inhibitors: Therapeutic

Challenges and Opportunities. Med. Res. Rev. 2017, 37, 219-270. Kuhn, B.; Benz, J.; Greif, M.; Engel, A. M.; Sobek, H.; Rudolph, M. G., The Structure of

Human α-2,6-Sialyltransferase Reveals the Binding Mode of Complex Glycans. Acta Crystallogr. D 2013, 69, 1826-1828.

Montgomery, A.; Szabo, R.; Skropeta, D.; Yu, H., Computational Characterisation of the Interactions between Human ST6Gal I and Transition-state Analogue Inhibitors: Insights for Inhibitor Design. J. Mol. Recogn 2016, 29, 210-222.

8

Natural Products used as Chemical Library for Protein-Protein Interactions (PPIs) Targeted Drug Discovery

Xuemei, J1, Kyungro, L1,2, Nam Hee, K3, Hyun Sil, K3, Jong In, Y3, Jiwon, C2, Kyoung Tai,

N1,2


1 Yonsei University, Seoul, Korea 2 Bioinformatics & Molecular Design Research Center (BMDRC), Yonsei University, Seoul,

Korea 3Yonsei University College of Dentistry, Seoul, Korea

Protein-protein interactions (PPIs) are essential for cellular processes and have been recognized as attractive targets for therapeutic intervention. Therefore, the construction of PPI-focused chemical library is inevitable necessity for future drug discovery. Natural products have been used as traditional medicines to treat human diseases for millennia and their molecular scaffolds have been identified in a diverse number of approved drugs and drug candidates. The recent successful inhibition of PPIs by natural products give us an opportunity to use natural products as a chemical library for PPI targeted drug discovery. In this work, we compared the distribution of small molecular PPI inhibitors database (iPPIs), FDA-approved drugs and natural products database (NPDB) by using eight molecular descriptors to explore the potential of natural product library for PPI inhibitor drug discovery. Then we evaluated the distribution of NPDB and iPPIs in the chemical space represented by the molecular fingerprint and molecular scaffolds to identify the useful scaffolds which would have effects on PPIs. To estimate the effects of natural products to inhibit PPI targets, the molecular docking was employed. After that, we predicted a set of high potent natural products by using the iPPI-likeness score based on a docking score-weighted model. These selected natural products showed high binding affinities to the PPI target that have been validated by in vitro experiment. Overall, our study illustrates the potency of natural products to target PPIs and accelerates the design of the PPI-focused chemical library in drug discovery.

9

Revisiting the Inter-relationship Between Correlation Coefficient, Sample Size, and Confidence Level

Wang, R.-X.* Yang, Q.-F.


Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences Shanghai, People's Republic of China

The quality of a computational model, e.g. a scoring function for computing protein-ligand binding affinity, is often measured by the correlation coefficient (R) between experimental data and the computed values. In order to compare the performance of two models, one needs to consider the R values produced by them as well as the sample size to derive a meaningful hypothesis. Recently, Carlson et al. published an analysis on this issue (J. Chem. Inf. Model. 2013, 53, 1837-1841), and they came to an astonishing conclusion that current available data sets were simply too small for testing scoring functions at an acceptable confidence level. In this study, we examined this issue and found that Carlson et al had misinterpreted their analysis results. Besides, it is more appropriate to apply Meng's method for comparing correlated correlation coefficients (Psychological Bulletin, 1992, 111, 172-175) to this analysis. Our analysis reveals that: (1) A larger data set is required if there is only a subtle difference between the R values of two models. (2) A larger data set is required if the R values of two models are at a lower level. (3) If the inter-correlation between the outcomes of two models is low, a large data set of several hundred to several thousand samples is required; otherwise, a smaller data set of several dozen samples is required. A numerical simulation was conducted, which verified that our results were consistent with the theoretical prediction; whereas Carlson's results were not.

10

MAPPING OF DRUG-LIKE CHEMICAL UNIVERSE WITH REDUCED COMPLEXITY MOLECULAR FRAMEWORKS

Kontijevskis, A., Gouliaev, A. H., Franch, T., Nørregaard-Madsen, M.


Nuevolution A/S, Copenhagen, Denmark

The emergence of DNA-encoded chemical libraries (DEL) field in past decade has attracted attention of pharmaceutical industry as a powerful mechanism for the discovery of novel drug-like hits for various biological targets. Nuevolution Chemetics technology1 enables DNA encoded synthesis of billions of chemically diverse drug-like small molecule compounds, and the efficient screening and optimization of these, facilitating effective identification of drug candidates at an unprecedented speed and scale. Although many approaches have been developed by the cheminformatics community for the analysis and visualization of drug-like chemical space, most of them are restricted to the analysis of maximum few millions of compounds and cannot handle collections of 108-1012 compounds typical for DELs. To address this big chemical data challenge, we developed Reduced Complexity Molecular (RCM) frameworks2 methodology as an abstract and very general way of representing chemical structures. By further introducing RCM framework descriptors we constructed a global framework map of all drug-like chemical space and demonstrate how chemical space occupied by multi-million-member drug-like Chemetics DNA-encoded libraries and virtual combinatorial libraries with >1012 members could be analyzed and mapped without a need for library enumeration. We further validate the approach by performing RCM framework-based searches in drug-like chemical universe and mapping Chemetics library selection outputs for LSD1 target on a global framework chemical space map. References

1. Nuevolution A/S, https://nuevolution.com 2. This work is submitted and is being currently under review in the Journal of

Chemical Information and Modelling.

11

Discovering binding pathways and metastable states with Markov state models

Thomas, T1, Yuriev, E1, Chalmers, DK1


1 Monash Institute of Pharmaceutical Science, Melbourne, Australia Structure-based drug design is largely reliant on the approximation that drug binding follows a simple two-state model, bound and unbound. In reality, there are many shorter-lived metastable states along the pathway to the bound state, these metastable states are themselves interesting targets for drug design, they are known to correlate with allosteric sites and are promising for bitopic ligands.

Molecular dynamics (MD) simulations allow us to examine the binding pathways of a drug in a detail unreachable by experimental methods, metastable sites can be discovered, and the primary barriers to drug binding identified. Although binding to many drug targets now falls into timescales that can be comfortably simulated with MD, many other targets, including G protein-coupled receptors (GPCRs), are still out of reach of conventional simulations.

Markov state models (MSMs) are a statistics based method that have proven effective at extending the timescales reachable by MD simulations in protein folding and simple ligand binding events. By observing individual transitions in many short MD simulations, the probability of transitioning from one state to another can be extrapolated to much longer timescales.

We have developed a methodology that has allowed us to construct MSMs of drug binding to a GPCR, allowing us to identify the metastable sites and binding pathways of the antipsychotic haloperidol binding to the D3 dopamine receptor.

12

Less is more: distilling human-interpretable simplicity from models of large qHTS result matrices.

Brown, J.B.1


1 Kyoto University Graduate School of Medicine, Kyoto, Japan

Chemogenomics is a field of chemical bioscience which attempts to find the proteins involved in a phenotype (forward chemogenomics), and chemical modulators of the identified proteins (reverse chemogenomics). Advances in high-throughput screening and combinatorial chemistry have yielded enormous datasets of ligand-receptor bioactivity matrices, but it is still a major challenge to effectively build extrapolating models from these matrices. Further, the majority of machine-learning models are almost impossible to interpret by humans. In a very recent work [1], we have established a technique for identifying a greatly reduced subset of ligand-target pairs which still function to be highly predictive on an entire family-wide ligand-target matrix. In addition to the ability to be predictive, the reduced set can maintain an effective balance between strong and weak binding pairs, as well as display predictive ability on the level of individual targets. The ability is poised to assist pharmaceutical companies and chemical biology labs that generate large ligand-target matrices and are struggling to interpret the results. The method could also be exploited to select experiments to run rather than screening en masse.

Importantly, the technique calls into question the current hype in drug discovery that algorithms such as deep learning can somehow explain large SAR data in a way that cannot be modeled by simpler algorithms. While there is no detriment in having big data per se, its utility might be subjective. Drug discovery based on robust statistics continues to hold the brightest prospect. In this keynote, I will cover many aspects of this exciting new technology. [1] Daniel Reker*, Petra Schneider, Gisbert Schneider, J.B. Brown* Active learning for computational chemogenomics Future Medicinal Chemistry, 9(4):381-402 DOI: 10.4155/fmc-2016-0197 -- PubMed ID 28263088

13

Method Development for Prediction of Drug-Target Interactions

Yun Tang, Feixiong Cheng, Zengrui Wu

Shanghai Key Laboratory of New Drug Design, School of Pharmacy, East China University

of Science and Technology, Shanghai 200237, China


Drug-target interaction (DTI) is the basis of drug discovery. It can be used in virtual screening of active compounds for a specific target, or target prediction for an active compound. However, it is time-consuming and costly to determine potential DTIs experimentally. In silico methods are hence very helpful and have shown great advantages in drug discovery.

In recent years, we developed a method named network-based inference (NBI) to

predict new DTIs for old drugs.1 Excellent performance was obtained for the method on the benchmark datasets. Five old drugs identified via NBI were experimentally validated to have potential binding effects on estrogen receptors or dipeptidyl peptidase IV.1

However, the NBI method cannot predict potential targets for any new compounds

outside the DTI networks, due to lack of connections between new compounds and existing DTI networks. To address this deficiency, we proposed a new method, named substructure-drug-target network-based inference (SDTNBI), for DTI prediction.2 SDTNBI employed chemical substructures to bridge the gap between new chemical entities and known DTI networks. By utilizing a way of resource diffusion, it can prioritize potential targets for old drugs and new chemical entities in a large scale. Recently, we further made an improvement on SDTNBI by introducing three parameters into it, namely balanced substructure-drug-target network based inference (bSDTNBI), to identify potential targets for both old drugs and new chemical entities.3 High performance was yielded in both 10-fold and leave-one-out cross validations, outperforming the original NBI and SDTNBI methods. The method was also validated by in vitro assays.

References:

1) Cheng F, Liu C, Jiang J, Lu W, Li W, Liu G, Zhou W, Huang J, Tang Y. Prediction of drug-target interactions and drug repositioning via network-based inference. PLoS Comput Biol 2012, 8, e1002503.

2) Wu Z, Cheng F, Li J, Li W, Liu G, Tang Y. SDTNBI: an integrated network and chemoinformatics tool for systematic prediction of drug-target interactions and drug repositioning. Briefings in Bioinformatics, 2017, 18(2): 333-347.

3) Wu Z, Lu W, Wu D, Luo A, Bian H, Li J, Li W, Liu G, Huang J, Cheng F, Tang Y. In silico prediction of chemical mechanism-of-action via an improved network-based inference method. British J. Pharmacol. 2016, 173(23): 3372-3385.

14

PreMetabo for prediction of drug metabolism in hepatocytes: phase I, phase II, and transporters

Hwang, SB1, Shin, SE2, Seo, MW1, Shin, HK1, No, KT1,2


1Computational Systems Biology Lab., Department of Biotechnology, Yonsei University,

Seoul, Republic of Korea

2Bioinformatics and Molecular Design Research Center, Yonsei Engineering Research Park, Seoul, Republic of Korea

Toxicity and efficacy of drug can be altered through metabolism reactions in hepatocyte; therefore, prediction on drug metabolism in early phase can give benefits in drug discovery process. Researches on drug metabolism revealed the essential role of cytochrome P450 enzymes (CYPs) for phase I reactions, UDP-glucuronosyltransferases (UGTs) and sulfotransferases (SULTs) for phase II reactions, and P-glycoprotein (P-gp) for efflux of drug molecules.

PreMetabo (prediction + metabolism) was implemented by combining prediction models developed on drug metabolism. The software covers various endpoints such as 1) prediction of Site Of Metabolism (SOM) by CYP1A2, CYP2D6, CYP2C9, and CYP3A4, 2) classification of substrate to CYP2D6, CYP3A4, UGT, SULT, and P-gp, and 3) classification of inhibitor to CYP2C9, CYP2C19, CYP2D6, and CYP3A4. SOM of a drug molecule was predicted with docking and activation energy. Classification models aimed to screen molecules vulnerable to metabolism reactions in hepatocytes. PreMetabo is not only available as GUI version, but also as web version (https://premetabo.bmdrc.kr/).

15

Cyclic D/L Peptide Nanotubes Towards Functional Bionanomaterials

Silk, MR1, Newman, J2, Caradoc-Davies, T3, Price, JR3, Ratcliffe, J4, White, J4, Perrier, S5, Thompson, PE1, Chalmers, DK1


1Monash Institute of Pharmaceutical Sciences, Melbourne, Australia 2CSIRO Biomedical Program, Parkville, Australia

3Australian Synchrotron, Clayton, Australia 4CSIRO Manufacturing, Clayton, Australia 5The University of Warwick, Coventry, UK

Cyclic peptides comprised of an even number of alternating D/L amino acids can self-assemble into nanotubes through backbone hydrogen bonding. This class of nanomaterial offers immense chemical and structural diversity that can be tailored through amino acid composition. Their porous nature and high biological compatibility grants potential applications as drug-delivery vehicles, synthetic membranes and protein/ion-channels. However, their development into functional nanomaterials is hindered by difficulties in structure-characterisation, poor control over the nature of nanotube assembly and low aqueous solubility. In order to develop nanotubes of high stability and aqueous solubility, we have studied a wide-range of cyclic D/L peptide nanotubes through an iterative approach involving computational design, solid-phase peptide synthesis and structural analysis. The nanostructures are characterised in-solution by CD, DLS, NMR and cryo-TEM as well as in the solid-state by X-ray crystallography and FT-IR. We have obtained the first crystal structures of cyclic peptide nanotubes in both the parallel and antiparallel configuration, allowing for in-depth analyses of the H-bond network and side-chain interactions associated with nanotube stability. New peptides have been designed and studied by molecular dynamics as we seek methods for controlling and directing the nanotube assembly process to develop complex, functionalised materials for application in biomedicine.

A) B) C) Figure 1. (A) Chemical structure of a cyclic D/L octapeptide subunit. (B) Basic depiction of the backbone hydrogen-bond network between peptide subunits, with dotted lines for H-bonds. (C) Molecular model of a cyclic D/L peptide nanotube comprised of eight cyclo[(Asp-D-Leu-Lys-D-Leu)2] subunits.

16

Forging the Basis for Developing Protein−Ligand Interaction Scoring Functions

Liu, Z1, Su, M1, Han, L1, Liu, J1, Yang, Q1, Li, Y*1, Wang, R*1, 2


1 Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences, Shanghai, China 2 Macau University of Science and Technology, Macau, People’s Republic of China

In structure-based drug design, scoring functions are widely used for fast evaluation of protein-ligand interactions. Regardless of their technical difference, scoring functions all need data sets combining protein-ligand complex structures and binding affinity data for parametrization and validation. However, data sets of this kind used to be rather limited in terms of size and quality. On the other hand, standard metrics for evaluating scoring function used to be ambiguous, which do not directly reflect the genuine quality of scoring functions. In a recently published paper (Acc. Chem. Res., 2017, DOI: 10.1021/acs.accounts. 6b00491), we describe our long-lasting efforts to overcome these obstacles, which involve two related projects. On the first project, we have created the PDBbind database. It is the first database that systematically annotates the protein-ligand complexes in the Protein Data Bank (PDB) with experimental binding data. This database has been updated annually since its first public release in 2004. The latest release (v2016) provides binding data for 16179 biomolecular complexes in PDB. Data sets provided by PDBbind have been applied to many computational and statistical studies. In particular, it has become a major data resource for scoring function development. On the second project, we have established the Comparative Assessment of Scoring Functions (CASF) benchmark for scoring function evaluation. Our key idea is to decouple the “scoring” process from the “sampling” process, so scoring functions can be tested in a relatively pure context to reflect their quality. Importantly, CASF is designed to be an open-access benchmark.

17

Free Energy Methods to Expedite Drug Discovery

Williams-Noonan, BJ,1 Chalmers, DK,1 Yuriev, E1

[email protected] 1 Monash University, Melbourne, Australia

Underpinning all rational design-based drug discovery projects is the interaction between a drug and its target. So, improved methods of determining the magnitude of protein-ligand interactions have the potential to expedite drug discovery processes and reduce the cost.1 Presently, computational free energy methods offer an appealing means to assist in existing drug discovery projects. In this research, we first look at cyclic peptides that bind to the allosteric site of Human Immunodeficiency Virus Integrase (HIV-IN) and use free energy techniques to understand the reason behind their modest mM binding affinity. Using free energy techniques it was found that the global minimum conformer of one cyclic peptide in solution was more energetically favourable than the crystal-bound pose in solution, thereby imposing a large entropic penalty to binding. Therefore, it was shown using molecular dynamics (MD) that peptides constrained into the correct bound-pose in solution should resultantly bind with higher affinity. Here we also present a study aimed at comparing the effectiveness of four common free energy methods at replicating known binding affinities. As before, the specific target was the allosteric site of HIV-IN, and the ligands in the study were those published by Fader, et al (2014).2 We utilised four free energy techniques and compared them in terms of their ability to reproduce the rank order of the ligands in the sample. Relative Free Energy Perturbation (FEP) outperformed all other in silico methods trialled. As computers and algorithms continue to improve, free energy techniques should become more mainstream in drug design. REFERENCES

[1] DiMasi, J. A.; Grabowski, H. G.; Hansen, R. W. Innovation in the pharmaceutical industry: new estimates of R&D

costs. Journal of health economics 2016, 47, 20-33.

[2] Fader, L. D.; Malenfant, E.; Parisien, M.; Carson, R.; Bilodeau, F.; Landry, S.; Pesant, M.; Brochu, C.; Morin, S.;

Chabot, C. Discovery of BI 224436, a noncatalytic site integrase inhibitor (NCINI) of HIV-1. ACS medicinal

chemistry letters 2014, 5, 422-427.

18

Function Prediction of an Herb Based on its Components

Xu, WL1, Jiang SY1, Hu J1, Li, JJ2, Yao, JH1,2*

[email protected]

1 Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences, Shanghai, China 2 Zhengzhou Institute of Technology, Zhengzhou, Henan, China It is well known that herbs take a very important role in human health and agricultural production. In general, there is information about functions of herbs in classic herb dictionaries. These functions were tested and verified in preventing and curing diseases. Some reports showed that functions of an herb had close relationship with properties of its components. In the last thirty years, Chemoinformatics has been used in drug discovery and pesticide design successfully. These years, it has been employed in study of herbs. For example, 1) Construction of an information system, which includes herbs data and their compounds data. This system could be used to retrieve information about an herb and its compounds’ data by experiment and prediction from a query herb name, retrieve experimental and predictive information about compounds including the query fragment and data of herbs including these compounds from a query compound. This information system organically links information about herbs and chemical information about its components. 2) Revealing reasons about functions of herbs in preventing and treating diseases, discovering new functions of herbs by properties of compounds in herbs. The results will be helpful to rationally use and understand the mechanism of herbs in treating diseases. In parallel, pesticide functions of herbs would be discovered based on bioactivities of components by prediction or test. Herein, study of the herb, Apium graveolen L. will be introduced. The study results showed that Chemoinformatics played a positive role in understanding and discovering new functions of herbs. Acknowledgement

This work is supported by the project: cooperation between SIOC and ZIT

19

Heterocycle-Containing Macrocyclic β-strand mimetics

Schumann, N1, Kirby, S1, Abell, AD1

[email protected]

1The University of Adelaide, Adelaide, Australia Protease substrates almost universally adopt a β-strand conformation on active site binding to peptides and inhibitors. There is therefore a need to design structurally simple inhibitors which define this conformation. The introduction of a macrocyclic constraint linking the P1 and P3 residues of peptidomimetic-based protease inhibitors is one such way that has been demonstrated to define this secondary structure and to give rise to potent inhibitors of Calpain and other proteases.1 Second-generation macrocyclic inhibitors have been produced by replacing two amino acid residues in the peptidomimetic backbone by a planar pyrrole via Friedel-Crafts acylation. This insertion maintains the appropriate β-strand conformation while reducing peptide character and results in pM inhibitors of Cathepsin L and S.2 Herein is described work towards a third-generation macrocyclic protease inhibitor where a range of heterocycles is introduced into the backbone by simple amide bond formation and again to define a β-strand geometry. The synthesis proceeds via preparation of the linear peptide sequence by solution-phase peptide couplings, macrocyclisation by ring-closing metathesis between the P1 and P3 residues, and finally introduction of a terminal nitrile-containing residue by another peptide coupling reaction. Ongoing work is concerned with investigating the effects of the nature of the macrocyclic tether and heterocycle unit to geometry and activity.

1. A. D. Abell, M. A. Jones, J. M. Coxon, J. D. Morton, S. G. Aitken, S. B. McNabb, H. Y. Lee, J. M. Mehrtens, N. A. Alexander, B. G. Stuart, A. T. Neffe, R. Bickerstaffe, Angew. Chem. Int. Ed. 2009, 48, 1455. 2. K. C. H. Chua, M. Pietsch, X. Zhang, S. Hautmann, H. Y. Chan, J. B. Bruning, M. Gutschow, A. D. Abell, Angew. Chem. Int. Ed. 2014, 53, 7828.

20

In silico drug discovery targeting Hippo pathway and YAP-TEAD protein-protein interactions for cancer therapy

Kim, JW1, Choi, JW2, No, KT1,2


1Computational Systems Biology Lab., Department of Biotechnology, Yonsei University,

Seoul, Republic of Korea

2Bioinformatics and Molecular Design Research Center, Yonsei Engineering Research Park, Seoul, Republic of Korea

Hippo pathway is one of important pathways regulating tissue growth and proliferation. This pathway is a complex signalling network consisted of many components. Dysregulation on this pathway can result in overgrowth of phenotypes because of a malfunction of stem cell proliferation, differentiation and apoptosis, which are directly involved with components to cancer cell developments. Protein-Protein interaction (PPI) in YAP (yes-associated protein) and TEAD (transcriptional enhancer associate domain) is a key interaction which regulate cancer cell growth in hippo pathway. Regulating YAP-TEAD Protein-Protein interaction by small molecules could play an important role in cancer therapy. In this study, we identified small molecules inhibiting YAP-TEAD interactions by competitively binds to TEAD instead of YAP. Molecular hits were identified by computer based screening pipeline composed of pharmacophore based virtual screening, and conventional molecular docking with PPI specific molecular force field. Several compounds were selected by visual inspection from inhibitors with proper pharmacophore fit score and higher docking scores in in silico procedure. In vitro test of luciferase activity experiment was conducted on lead compounds selected by above process. As a result, several lead compounds were identified to inhibit YAP-TEAD protein-protein interactions.

Abstract Book - RACI Centenary Congress | RACI … Book 1 Structure ... This project has received...

Documents

Transcript of Abstract Book - RACI Centenary Congress | RACI … Book 1 Structure ... This project has received...