Computer-aided Drug Discovery Research Articles

Computer aided drug discovery of highly ligand efficient, low molecular weight imidazopyridine analogs as FLT3 inhibitors

The FLT3 kinase represents an attractive target to effectively treat AML. Unfortunately, no FLT3 targeted therapeutic is currently approved. In line with our continued interests in treating kinase related disease for anti-FLT3 mutant activity, we utilized pioneering synthetic methodology in combination with computer aided drug discovery and identified low molecular weight, highly ligand efficient, FLT3 kinase inhibitors. Compounds were analyzed for biochemical inhibition, their ability to selectively inhibit cell proliferation, for FLT3 mutant activity, and preliminary aqueous solubility. Validated hits were discovered that can serve as starting platforms for lead candidates.

LIGSIFT: an open-source tool for ligand structural alignment and virtual screening.

Shape-based alignment of small molecules is a widely used approach in computer-aided drug discovery. Most shape-based ligand structure alignment applications, both commercial and freely available ones, use the Tanimoto coefficient or similar functions for evaluating molecular similarity. Major drawbacks of using such functions are the size dependence of the score and the fact that the statistical significance of the molecular match using such metrics is not reported. We describe a new open-source ligand structure alignment and virtual screening (VS) algorithm, LIGSIFT, that uses Gaussian molecular shape overlay for fast small molecule alignment and a size-independent scoring function for efficient VS based on the statistical significance of the score. LIGSIFT was tested against the compounds for 40 protein targets available in the Directory of Useful Decoys and the performance was evaluated using the area under the ROC curve (AUC), the Enrichment Factor (EF) and Hit Rate (HR). LIGSIFT-based VS shows an average AUC of 0.79, average EF values of 20.8 and a HR of 59% in the top 1% of the screened library. LIGSIFT software, including the source code, is freely available to academic users at http://cssb.biology.gatech.edu/LIGSIFT. Supplementary data are available at Bioinformatics online. skolnick@gatech.edu.

GalaxySite: ligand-binding-site prediction by using molecular docking.

Knowledge of ligand-binding sites of proteins provides invaluable information for functional studies, drug design and protein design. Recent progress in ligand-binding-site prediction methods has demonstrated that using information from similar proteins of known structures can improve predictions. The GalaxySite web server, freely accessible at http://galaxy.seoklab.org/site, combines such information with molecular docking for more precise binding-site prediction for non-metal ligands. According to the recent critical assessments of structure prediction methods held in 2010 and 2012, this server was found to be superior or comparable to other state-of-the-art programs in the category of ligand-binding-site prediction. A strong merit of the GalaxySite program is that it provides additional predictions on binding ligands and their binding poses in terms of the optimized 3D coordinates of the protein–ligand complexes, whereas other methods predict only identities of binding-site residues or copy binding geometry from similar proteins. The additional information on the specific binding geometry would be very useful for applications in functional studies and computer-aided drug discovery.

The Bitter Barricading of Prostaglandin Biosynthesis Pathway: Understanding the Molecular Mechanism of Selective Cyclooxygenase-2 Inhibition by Amarogentin, a Secoiridoid Glycoside from Swertia chirayita

Swertia chirayita, a medicinal herb inhabiting the challenging terrains and high altitudes of the Himalayas, is a rich source of essential phytochemical isolates. Amarogentin, a bitter secoiridoid glycoside from S. chirayita, shows varied activity in several patho-physiological conditions, predominantly in leishmaniasis and carcinogenesis. Experimental analysis has revealed that amarogentin downregulates the cyclooxygenase-2 (COX-2) activity and helps to curtail skin carcinogenesis in mouse models; however, there exists no account on selective inhibition of the inducible cyclooxygenase (COX) isoform by amarogentin. Hence the computer-aided drug discovery methods were used to unravel the COX-2 inhibitory mechanism of amarogentin and to check its selectivity for the inducible isoform over the constitutive one. The generated theoretical models of both isoforms were subjected to molecular docking analysis with amarogentin and twenty-one other Food and Drug Authority (FDA) approved lead molecules. The post-docking binding energy profile of amarogentin was comparable to the binding energy profiles of the FDA approved selective COX-2 inhibitors. Subsequent molecular dynamics simulation analysis delineated the difference in the stability of both complexes, with amarogentin-COX-2 complex being more stable after 40ns simulation. The total binding free energy calculated by MMGBSA for the amarogentin-COX-2 complex was −52.35 KCal/mol against a binding free energy of −8.57 KCal/mol for amarogentin-COX-1 complex, suggesting a possible selective inhibition of the COX-2 protein by the natural inhibitor. Amarogentin achieves this potential selectivity by small, yet significant, structural differences inherent to the binding cavities of the two isoforms. Hypothetically, it might block the entry of the natural substrates in the hydrophobic binding channel of the COX-2, inhibiting the cyclooxygenation step. To sum up briefly, this work highlights the mechanism of the possible selective COX-2 inhibition by amarogentin and endorses the possibility of obtaining efficient, futuristic and targeted therapeutic agents for relieving inflammation and malignancy from this phytochemical source.

Progress towards Automated Kepler Scientific Workflows for Computer-aided Drug Discovery and Molecular Simulations

We describe the development of automated workflows that support computed-aided drug discovery (CADD) and molecular dynamics (MD) simulations and are included as part of the National Biomedical Computation Resource (NBCR). The main workflow components include: file-management tasks, ligand force field parameterization, receptor-ligand molecular dynamics (MD) simulations, job submission, serial and parallel execution, and monitoring on relevant high-performance computing (HPC) resources, receptor structural clustering, virtual screening (VS), and statistical analyses of the VS results. The workflows aim to standardize simulation and analysis and promote best practices within the molecular simulation and CADD communities. Each component is developed as a stand-alone workflow, which should allow for easy integration into larger frameworks built suiting user needs, while remaining intuitive and easy to extend.

SWEETLEAD: an In Silico Database of Approved Drugs, Regulated Chemicals, and Herbal Isolates for Computer-Aided Drug Discovery

In the face of drastically rising drug discovery costs, strategies promising to reduce development timelines and expenditures are being pursued. Computer-aided virtual screening and repurposing approved drugs are two such strategies that have shown recent success. Herein, we report the creation of a highly-curated in silico database of chemical structures representing approved drugs, chemical isolates from traditional medicinal herbs, and regulated chemicals, termed the SWEETLEAD database. The motivation for SWEETLEAD stems from the observance of conflicting information in publicly available chemical databases and the lack of a highly curated database of chemical structures for the globally approved drugs. A consensus building scheme surveying information from several publicly accessible databases was employed to identify the correct structure for each chemical. Resulting structures are filtered for the active pharmaceutical ingredient, standardized, and differing formulations of the same drug were combined in the final database. The publically available release of SWEETLEAD (https://simtk.org/home/sweetlead) provides an important tool to enable the successful completion of computer-aided repurposing and drug discovery campaigns.

SAR and Computer-Aided Drug Design Approaches in the Discovery of Peroxisome Proliferator-Activated Receptor γ Activators: A Perspective

Activators of PPARγ, Troglitazone (TGZ), Rosiglitazone (RGZ), and Pioglitazone (PGZ) were introduced for treatment of Type 2 diabetes, but TGZ and RGZ have been withdrawn from the market along with other promising leads due cardiovascular side effects and hepatotoxicity. However, the continuously improving understanding of the structure/function of PPARγ and its interactions with potential ligands maintain the importance of PPARγ as an antidiabetic target. Extensive structure activity relationship (SAR) studies have thus been performed on a variety of structural scaffolds by various research groups. Computer-aided drug discovery (CADD) approaches have also played a vital role in the search and optimization of potential lead compounds. This paper focuses on these approaches adopted for the discovery of PPARγ ligands for the treatment of Type 2 diabetes. Key concepts employed during the discovery phase, classification based on agonistic character, applications of various QSAR, pharmacophore mapping, virtual screening, molecular docking, and molecular dynamics studies are highlighted. Molecular level analysis of the dynamic nature of ligand-receptor interaction is presented for the future design of ligands with better potency and safety profiles. Recently identified mechanism of inhibition of phosphorylation of PPARγ at SER273 by ligands is reviewed as a new strategy to identify novel drug candidates.

Modeling G protein-coupled receptors and their interactions with ligands

The recent boom of G protein-coupled receptor (GPCR) crystallography is currently revolutionizing the way modulators of these highly druggable targets are discovered. Not only are these structures directly applicable to computer-aided drug discovery, but they also provide templates for the construction of homology models of other receptors. The study of the binding mode of GPCR modulators through docking experiments remains challenging. In addition to an expert use of advanced modeling tools, the application of experimental knowledge derived from site-directed mutagenesis data is fundamental for the generation of accurate receptor-ligand complexes applicable to drug discovery. We expect that the growing number of experimental and computational GPCR structures will boost the rational discovery of novel modulators in coming years.

Screening inhibitors against h1n1 neuraminidase using molecular docking

Prediction of protein in large databases is one of the major research objectives in structural and functional proteomics. Therefore, its amino acid sequence is compared with the sequences of structural database using computational techniques. It needs to be design drugs for diseases which are being caused by proteins whose structure is unknown; In order to streamline computer aided drug discovery and development in the spirit of scientific method. The outbreak of H1N1 influenza pandemic strains that are resistant to currently prescribed antivirals has earmarked the need to develop/screen novel compounds that target viral gene products such as Neuraminidase. Preliminary investigation revealed active site residue dissimilarity with H5N1 Neuraminidase. Screening and comparison of various antiviral drugs by using Bioinformatics techniques. Original anti-flu drugs such as Oseltamivir, Zanamivir etc were allowed to bind within the active region of H5N1 as well as H1N1 neuraminidase to study the affinities. Screenings of various other drugs were performed using Drug Bank database to assess the ability of binding of such drugs and to check the possibility of them as Flu inhibitors. Affinity of binding was performed by GOLD, PATCH Dock and ME dock softwares.

Benchmarking Ligand-Based Virtual High-Throughput Screening with the PubChem Database

With the rapidly increasing availability of High-Throughput Screening (HTS) data in the public domain, such as the PubChem database, methods for ligand-based computer-aided drug discovery (LB-CADD) have the potential to accelerate and reduce the cost of probe development and drug discovery efforts in academia. We assemble nine data sets from realistic HTS campaigns representing major families of drug target proteins for benchmarking LB-CADD methods. Each data set is public domain through PubChem and carefully collated through confirmation screens validating active compounds. These data sets provide the foundation for benchmarking a new cheminformatics framework BCL::ChemInfo, which is freely available for non-commercial use. Quantitative structure activity relationship (QSAR) models are built using Artificial Neural Networks (ANNs), Support Vector Machines (SVMs), Decision Trees (DTs), and Kohonen networks (KNs). Problem-specific descriptor optimization protocols are assessed including Sequential Feature Forward Selection (SFFS) and various information content measures. Measures of predictive power and confidence are evaluated through cross-validation, and a consensus prediction scheme is tested that combines orthogonal machine learning algorithms into a single predictor. Enrichments ranging from 15 to 101 for a TPR cutoff of 25% are observed.

Patents of bio-active compounds based on computer-aided drug discovery techniques

In recent times, there has been an increased use of Computer-Aided Drug Discovery (CADD) techniques in Medicinal Chemistry as auxiliary tools in drug discovery. Whilst the ultimate goal of Medicinal Chemistry research is for the discovery of new drug candidates, a secondary yet important outcome that results is in the creation of new computational tools. This process is often accompanied by a lack of understanding of the legal aspects related to software and model use, that is, the copyright protection of new medicinal chemistry software and software-mediated discovered products. In the center of picture, which lies in the frontiers of legal, chemistry, and biosciences, we found computational modeling-based drug discovery patents. This article aims to review prominent cases of patents of bio-active organic compounds that involved/protect also computational techniques. We put special emphasis on patents based on Quantitative Structure-Activity Relationships (QSAR) models but we include other techniques too. An overview of relevant international issues on drug patenting is also presented.

2012 Late-Breaking Basic Science Oral Session

### Cardiac Myosin Binding Protein C is an Ultra-early and Cardiac-specific Biomarker of Myocardial Necrosis Diederik W Kuster1, Adriana Cardenas-Ospina2, Lawson Miller2, Christian Troidl3, Holger M Nef3, Christoph Liebetrau3, Mollmann Helge3, Karen S Pieper4, Kenneth W Mahaffey4, Neal S Kleiman5, Bruno D Stuyvers2, Ali J Marian6, Sakthivel Sadayappan1; 1Loyola Univ Chicago, Maywood, IL, 2Memorial Univ, St John’s, Canada, 3Kerckhoff Heart and Thorax Cntr, Bad Neuheim, Germany, 4Duke Univ Med Cntr, Durham, NC, 5The Methodist DeBakey Heart and Vascular Cntr, Houston, TX, 6Univ of Texas Health Sciences, Houston, TX Rationale: We recently demonstrated that proteolytic cleavage fragments of cardiac myosin binding protein-C (cMyBP-C) can be detected in the serum after myocardial infarction (MI) in a rat model and in patients with MI. The findings implied that serum cMyBP-C might be useful as cardiac-specific biomarker for MI. The release kinetics of cMyBP-C in the circulation post-MI remains to be elucidated. Objective: Determine the release kinetics of cMyBP-C as an ultra-early and cardiac-specific biomarker of myocardial necrosis. Method and Results: To determine the exact timing of cMyBP-C release in the bloodstream post-acute MI, left anterior descending (LAD) coronary artery was ligated in adult swine (n=6). ECG showed significant ST elevation. Infarct size represented 12.4 ± 1.9% of total ventricular mass. Blood samples were collected before and at predetermined time points between 30 min and 14 days after LAD ligation. Plasma cMyBP-C level was quantified using a highly sensitive and rapid sandwich enzyme-linked immunosorbent assay. Compared with baseline, cMyBP-C levels were increased in post-MI serum within 45 min (0.64 ± 0.52 ng/ml) after LAD ligation and declined after 16 hrs to the baseline level (0.01 ± 0.00 ng/ml). In contrast, cardiac troponin I (cTnI) level peaked after …

Review on Chemogenomics Approach: Interpreting Antagonist Activity of Secreted Frizzled-Related Protein 1 in Glaucoma Disease with In-Silico Docking

Computer-aided drug discovery is a growing frontier in science. It covers different sub areas like chemoinformatics and chemogenomics. Chemogenomics is one of the emerging inter-disciplinary approaches in drug discovery, which combines conventional ligand based approach with biological information of drug targets. The main goal of this review is to check effective application of chemogenomics in understanding interactions between all possible ligands and their potential drug targets at molecular level. Recent studies revealed that increased expression of sFRP1an inhibitor of Wnt signalling pathway, seems to be responsible for Elevated Intracellular Pressure (IOP) in glaucoma patients. Glaucoma is a worldwide spread disease. Here, secreted frizzled-related protein-1 (sFRP1) has been used as a target protein. An important role of sFRP1, an antagonist of Wnt signalling pathway, has been found in regulating IOP. Wnt3a ligand protein and a natural compound from marine source Mycaperoxide H - have been used as ligands. In-silico docking of these ligands with sFRP family implies answers to many intricate queries in drug development field. Using above mentioned ligand-protein model in this study, application of chemogenomics has tried to explore the interaction of active site of proteins with the novel ligands. Henceforth, the present review will focus on predictive in-silico chemogenomic approaches with computer aided drug design could be used in drug design domain in identifying new targets in various diseases, in time and cost effective manner.

Call for Papers: GRC, CADD, and statistics, and all that

Molecular modeling and the art of computer-aided drug discovery seldom make much use of statistics, despite being fields that can not calculate important properties with great reliability. The 2013 CADD Gordon conference intends to examine what prevents a more effective use of statistics in routine modeling and to raise consciousness as to what is possible. Practical methods will be discussed, deeper issues in applying standard approaches addressed and research on successes and failures in other disciplines presented by invited experts.

Editorial[Hot Topic ; Special Issue: Multi-Criteria Approaches in Computer-Aided Drug Discovery

Editorial[Hot Topic ; Special Issue: Multi-Criteria Approaches in Computer-Aided Drug Discovery

Calculation of Host-Guest Binding Affinities Using a Quantum-Mechanical Energy Model.

The prediction of protein-ligand binding affinities is of central interest in computer-aided drug discovery, but it is still difficult to achieve a high degree of accuracy. Recent studies suggesting that available force fields may be a key source of error motivate the present study, which reports the first mining minima (M2) binding affinity calculations based on a quantum mechanical energy model, rather than an empirical force field. We apply a semi-empirical quantum-mechanical energy function, PM6-DH+, coupled with the COSMO solvation model, to 29 host-guest systems with a wide range of measured binding affinities. After correction for a systematic error, which appears to derive from the treatment of polar solvation, the computed absolute binding affinities agree well with experimental measurements, with a mean error 1.6 kcal/mol and a correlation coefficient of 0.91. These calculations also delineate the contributions of various energy components, including solute energy, configurational entropy, and solvation free energy, to the binding free energies of these host-guest complexes. Comparison with our previous calculations, which used empirical force fields, point to significant differences in both the energetic and entropic components of the binding free energy. The present study demonstrates successful combination of a quantum mechanical Hamiltonian with the M2 affinity method.

The Molecular Dynamics of Trypanosoma brucei UDP‐Galactose 4′‐Epimerase: A Drug Target for African Sleeping Sickness

During the past century, several epidemics of human African trypanosomiasis, a deadly disease caused by the protist Trypanosoma brucei, have afflicted sub-Saharan Africa. Over 10 000 new victims are reported each year, with hundreds of thousands more at risk. As current drug treatments are either highly toxic or ineffective, novel trypanocides are urgently needed. The T. brucei galactose synthesis pathway is one potential therapeutic target. Although galactose is essential for T. brucei survival, the parasite lacks the transporters required to intake galactose from the environment. UDP-galactose 4′-epimerase (TbGalE) is responsible for the epimerization of UDP-glucose to UDP-galactose and is therefore of great interest to medicinal chemists. Using molecular dynamics simulations, we investigate the atomistic motions of TbGalE in both the apo and holo states. The sampled conformations and protein dynamics depend not only on the presence of a UDP-sugar ligand, but also on the chirality of the UDP-sugar C4 atom. This dependence provides important insights into TbGalE function and may help guide future computer-aided drug discovery efforts targeting this protein.

Recent technological advancements in ligand-based and structure-based modeling: BCL: Cheminfo and rosetta ligand

B ligand-based and structure-based computational methods are invaluable tools in computer aided drug discovery. This presentation will introduce technological advancements in both disciplines. First, we introduce BCL: ChemInfo, a comprehensive machine learning-based quantitative structure activity relationship (QSAR) modeling framework featuring novel molecular descriptors, diverse automated feature selection, GPU acceleration, and consensus model analysis. Public availability of high-throughput screening (HTS) data is rapidly increasing highlighting the need for ligand-based computational methods, such as BCL: ChemInfo, to accelerate probe development and drug discovery while reducing costs. The framework was benchmarked on publically available HTS data (PubChem) for nine targets, selected as representatives for the major protein families most commonly targeted by therapeutics, to investigate the influence of size and composition of training data, the choice of objective function, effectiveness of several feature selection techniques, as well as the predictive power of consensus predictors using different machine learning techniques. Second, recently advancements in the small molecule docking using RosettaLigand will be presented including simultaneous docking of explicit interface water molecules, and small molecule docking into comparative models. A study performed on protein-centric water docking shows an improvement in ligand placement at a ratio of 9:1 while ligand-centric water docking allows for the recovery of up to 56% of failed docking studies using 341 structures from the CSAR benchmark of diverse protein-ligand complexes. A study on docking into comparative models found RosettaLigand was successful in recovering a native-like binding mode among the top ten scoring binding modes for 21 of 30 cases while template selection based on ligand occupancy rather than template-target identity was discovered to increase success.

Discrete Molecular Dynamics Distinguishes Nativelike Binding Poses from Decoys in Difficult Targets

Virtual screening is one of the major tools used in computer-aided drug discovery. In structure-based virtual screening, the scoring function is critical to identifying the correct docking pose and accurately predicting the binding affinities of compounds. However, the performance of existing scoring functions has been shown to be uneven for different targets, and some important drug targets have proven especially challenging. In these targets, scoring functions cannot accurately identify the native or near-native binding pose of the ligand from among decoy poses, which affects both the accuracy of the binding affinity prediction and the ability of virtual screening to identify true binders in chemical libraries. Here, we present an approach to discriminating native poses from decoys in difficult targets for which several scoring functions failed to correctly identify the native pose. Our approach employs Discrete Molecular Dynamics simulations to incorporate protein-ligand dynamics and the entropic effects of binding. We analyze a collection of poses generated by docking and find that the residence time of the ligand in the native and nativelike binding poses is distinctly longer than that in decoy poses. This finding suggests that molecular simulations offer a unique approach to distinguishing the native (or nativelike) binding pose from decoy poses that cannot be distinguished using scoring functions that evaluate static structures. The success of our method emphasizes the importance of protein-ligand dynamics in the accurate determination of the binding pose, an aspect that is not addressed in typical docking and scoring protocols.

Absolute Binding Free Energy Calculations to Improve the Accuracy of Near-Native Ligand Pose Predictions

When the high-resolution structure of a protein target is available, molecular docking experiments are commonly used for computer-aided drug discovery. Molecular docking experiments are advantageous since they can provide the binding mode of a molecule in a given target protein as well as the binding affinity. However, despite considerable efforts, accurate prediction of ligand poses bound to a target is still challenging due to the protein's structural flexibility. Recent free energy perturbation molecular dynamics simulation (FEP/MD) calculations have shown that the calculated binding free energies are in good agreement with experimental data for co-crystal benchmark targets. Here, we present an integrated methodology in which initial candidate selection from pose decoys obtained by docking is followed by FEP/MD calculations to improve the accuracy of near-native ligand pose prediction. Our approach is evaluated for the small molecule α-helix mimetics inhibiting protein-protein interactions such as p53-MDMX/MDM2 and BAK-MCL-1. The results demonstrate that using the centroid models of the most populated clusters of docking decoys is an efficient approach to select a small set of ligand conformations in which a near-native pose may be included and applying the FEP/MD method enhances ability in discriminating the near-native ligand conformation from the candidates.