Ligand Binding Pose Research Articles

A systematic methodology for large scale compound screening: A case study on the discovery of novel S1PL inhibitors

Decrease in sphingosine 1-phosphate (S1P) concentration induces migration of pathogenic T cells to the blood stream, disrupts the CNS and it is implicated in multiple sclerosis (MS), a progressive inflammatory disorder of the central nervous system (CNS), and Alzheimer’s disease (AD). A promising treatment alternative for MS and AD is inhibition of the activity of the microsomal enzyme sphingosine 1-phosphate lyase (S1PL), which degrades intracellular S1P. This report describes an integrated systematic approach comprising virtual screening, molecular docking, substructure search and molecular dynamics simulation to discover novel S1PL inhibitors. Virtual screening of the ZINC database via ligand-based and structure-based pharmacophore models yielded 10000 hits. After molecular docking, common substructures of the top ranking hits were identified. The ligand binding poses were optimized by induced fit docking. MD simulations were performed on the complex structures to determine the stability of the S1PL-ligand complex and to calculate the binding free energy. Selectivity of the selected molecules was examined by docking them to hERG and cytochrome P450 receptors. As a final outcome, 15 compounds from different chemotypes were proposed as potential S1PL inhibitors. These molecules may guide future medicinal chemistry efforts in the discovery of new compounds against the destructive action of pathogenic T cells.

Comparative assessment of machine-learning scoring functions on PDBbind 2013

Computational docking is the core process of computer-aided drug design (CADD); it aims at predicting the best orientation and conformation of a small molecule (drug ligand) when bound to a target large receptor molecule (protein) in order to form a stable complex molecule. The docking quality is typically measured by a scoring function: a mathematical predictive model that produces a score representing the binding free energy and hence the stability of the resulting complex molecule. An effective scoring function should produce promising drug candidates which can then be synthesized and physically screened using high throughput screening (HTS) process. Therefore, the key to CADD is the design of an efficient highly accurate scoring function. Many traditional techniques have been proposed, however, the performance was generally poor. Only in the last few years the application of the machine learning (ML) technology has been applied in the design of scoring functions; and the results have been very promising.In this paper, we propose 12 scoring functions based on a wide range of ML techniques. We analyze the performance of each on the scoring power (binding affinity prediction), ranking power (relative ranking prediction), docking power (identifying the native binding poses among computer-generated decoys), and screening power (classifying true binders versus negative binders) using the PDBbind 2013 database. We compare our results with the recently published comparative assessment of scoring functions (CASF-2013) of 20 classical scoring functions most of which are implemented in main-stream commercial software. For the scoring and ranking powers, the proposed ML scoring functions depend on a wide range of features (energy terms, pharmacophore, geometrical) that entirely characterize the protein–ligand complexes (about 108 features); these features are extracted from several docking software available in the literature; a feature-space reduction technique, namely, principal component analysis is then applied and the performance is studied accordingly. For the docking and screening powers, the proposed ML scoring functions depend on the geometrical features of the RF-Score (36 features) to utilize a larger number of training complexes (relative to the large number of decoys in the testing set). For the scoring power, the best ML scoring function (RF) achieves a Pearson correlation coefficient between the predicted and experimentally determined binding affinities of 0.704 versus 0.614 achieved by the best classical scoring function (X-ScoreHM). For the ranking power, the best ML scoring function (RF) achieves a Spearman correlation coefficient between the ranks based on the predicted and experimentally determined binding affinities of 0.697 versus 0.626 achieved by the best classical scoring function (X-ScoreHM). For the docking power, the best ML scoring function (BRT) has a success rate in identifying the top best-scored ligand binding pose within 2Å root-mean-square deviation from the native pose of 13.8% versus 81.0% achieved by the best classical scoring function (ChemPLP@GOLD). For the screening power, the best ML scoring function (SVM) has an average enrichment factor and success rate at the top 1% level of 3.76 and 6.45% versus 19.54 and 60% respectively achieved by the best classical scoring function (GlideScore-SP).

Ligand- and receptor-based docking with LiBELa.

Methodologies on molecular docking are constantly improving. The problem consists on finding an optimal interplay between the computational cost and a satisfactory physical description of ligand-receptor interaction. In pursuit of an advance in current methods we developed a mixed docking approach combining ligand- and receptor-based strategies in a docking engine, where tridimensional descriptors for shape and charge distribution of a reference ligand guide the initial placement of the docking molecule and an interaction energy-based global minimization follows. This hybrid docking was evaluated with soft-core and force field potentials taking into account ligand pose and scoring. Our approach was found to be competitive to a purely receptor-based dock resulting in improved logAUC values when evaluated with DUD and DUD-E. Furthermore, the smoothed potential as evaluated here, was not advantageous when ligand binding poses were compared to experimentally determined conformations. In conclusion we show that a combination of ligand- and receptor-based strategy docking with a force field energy model results in good reproduction of binding poses and enrichment of active molecules against decoys. This strategy is implemented in our tool, LiBELa, available to the scientific community.

Identification of potential Tpx inhibitors against pathogen-host interactions

Yersinia organisms cause many infectious diseases by invading human cells and delivering their virulence factors via the type three secretion system (T3SS). One alternative strategy in the fight against these pathogenic organisms is to interfere with their T3SS. Previous studies demonstrated that thiol peroxidase, Tpx is functional in the assembly of T3SS and its inhibition by salicylidene acylhydrazides prevents the secretion of pathogenic effectors. In this study, the aim was to identify potential inhibitors of Tpx using an integrated approach starting with high throughput virtual screening and ending with molecular dynamics simulations of selected ligands. Virtual screening of ZINC database of 500,000 compounds via ligand-based and structure-based pharmacophore models retrieved 10,000 hits. The structure-based pharmacophore model was validated using high-throughput virtual screening (HTVS). After multistep docking (SP and XP), common scaffolds were used to find common substructures and the ligand binding poses were optimized using induced fit docking. The stability of the protein–ligand complex was examined with molecular dynamics simulations and the binding free energy of the complex was calculated. As a final outcome eight compounds with different chemotypes were proposed as potential inhibitors for Tpx. The eight ligands identified by a detailed virtual screening protocol can serve as leads in future drug design efforts against the destructive actions of pathogenic bacteria.

Binding of the Antagonist Caffeine to the Human Adenosine Receptor hA2AR in Nearly Physiological Conditions.

Lipid composition may significantly affect membrane proteins function, yet its impact on the protein structural determinants is not well understood. Here we present a comparative molecular dynamics (MD) study of the human adenosine receptor type 2A (hA2AR) in complex with caffeine—a system of high neuro-pharmacological relevance—within different membrane types. These are POPC, mixed POPC/POPE and cholesterol-rich membranes. 0.8-μs MD simulations unambiguously show that the helical folding of the amphipathic helix 8 depends on membrane contents. Most importantly, the distinct cholesterol binding into the cleft between helix 1 and 2 stabilizes a specific caffeine-binding pose against others visited during the simulation. Hence, cholesterol presence (~33%-50% in synaptic membrane in central nervous system), often neglected in X-ray determination of membrane proteins, affects the population of the ligand binding poses. We conclude that including a correct description of neuronal membranes may be very important for computer-aided design of ligands targeting hA2AR and possibly other GPCRs.

Identification of STAT1 and STAT3 specific inhibitors using comparative virtual screening and docking validation.

Signal transducers and activators of transcription (STATs) facilitate action of cytokines, growth factors and pathogens. STAT activation is mediated by a highly conserved SH2 domain, which interacts with phosphotyrosine motifs for specific STAT-receptor contacts and STAT dimerization. The active dimers induce gene transcription in the nucleus by binding to a specific DNA-response element in the promoter of target genes. Abnormal activation of STAT signaling pathways is implicated in many human diseases, like cancer, inflammation and auto-immunity. Searches for STAT-targeting compounds, exploring the phosphotyrosine (pTyr)-SH2 interaction site, yielded many small molecules for STAT3 but sparsely for other STATs. However, many of these inhibitors seem not STAT3-specific, thereby questioning the present modeling and selection strategies of SH2 domain-based STAT inhibitors. We generated new 3D structure models for all human (h)STATs and developed a comparative in silico docking strategy to obtain further insight into STAT-SH2 cross-binding specificity of a selection of previously identified STAT3 inhibitors. Indeed, by primarily targeting the highly conserved pTyr-SH2 binding pocket the majority of these compounds exhibited similar binding affinity and tendency scores for all STATs. By comparative screening of a natural product library we provided initial proof for the possibility to identify STAT1 as well as STAT3-specific inhibitors, introducing the ‘STAT-comparative binding affinity value’ and ‘ligand binding pose variation’ as selection criteria. In silico screening of a multi-million clean leads (CL) compound library for binding of all STATs, likewise identified potential specific inhibitors for STAT1 and STAT3 after docking validation. Based on comparative virtual screening and docking validation, we developed a novel STAT inhibitor screening tool that allows identification of specific STAT1 and STAT3 inhibitory compounds. This could increase our understanding of the functional role of these STATs in different diseases and benefit the clinical need for more drugable STAT inhibitors with high specificity, potency and excellent bioavailability.

QM/MM Docking Strategy and Prime/MM-GBSA Calculation of Celecoxib Analogues as N-myristoyltransferase Inhibitors

Developing chemicals that inhibit N-myristoyltransferase (Nmt) is a promising adjuvant therapeutic to improve the efficacy and selectivity of antifungal agents. Reliable prediction of binding-free energy and binding affinity of Nmt inhibitors can provide a guide for rational drug design. In this study, Quantum Polarised Ligand Docking (QPLD) strategy and Prime/Molecular Mechanics Generalized Born Surface Area (Prime/MM-GBSA) calculations were applied to predict the binding mode and free energy for a series of celecoxib analogues as Nmt inhibitors which were also found to have good anti-inflammatory activity. In vitro antifungal assay indicated that these derivatives were also acting as potent antifungal agents. Reliable docking results showed superior performance on both ligand binding pose and docking score accuracy. Then, the Prime/MM–GBSA method based on the docking complex was used to predict the binding-free energy. The combined use of QM/MM docking and Prime/MM-GBSA method gave a good correlation between the predicted binding-free energy and experimentally determined zone of inhibition and Minimum Inhibitory Concentration (MIC) values. The molecular docking combined with Prime/MM-GBSA simulation can not only be used to rapidly and accurately predict the binding-free energy of novel Nmt inhibitors but also provide a novel strategy for lead discovery and optimization targeting Nmt.

Identification of new hit scaffolds by INPHARMA-guided virtual screening

INPHARMA, an NMR-based method that determines ligand binding poses, allows finding novel scaffolds as inhibitors of a target protein.

SuperPain—a resource on pain-relieving compounds targeting ion channels

Pain is more than an unpleasant sensory experience associated with actual or potential tissue damage: it is the most common reason for physician consultation and often dramatically affects quality of life. The management of pain is often difficult and new targets are required for more effective and specific treatment. SuperPain (http://bioinformatics.charite.de/superpain/) is freely available database for pain-stimulating and pain-relieving compounds, which bind or potentially bind to ion channels that are involved in the transmission of pain signals to the central nervous system, such as TRPV1, TRPM8, TRPA1, TREK1, TRESK, hERG, ASIC, P2X and voltage-gated sodium channels. The database consists of ∼8700 ligands, which are characterized by experimentally measured binding affinities. Additionally, 100 000 putative ligands are included. Moreover, the database provides 3D structures of receptors and predicted ligand-binding poses. These binding poses and a structural classification scheme provide hints for the design of new analgesic compounds. A user-friendly graphical interface allows similarity searching, visualization of ligands docked into the receptor, etc.

HotLig: A Molecular Surface-Directed Approach to Scoring Protein–Ligand Interactions

Accurate prediction of ligand-binding poses is crucial for understanding molecular interactions and is very important for drug discovery, structural design, and optimization. In this study, we developed a novel scoring program, HotLig, which applies the Connolly surface of a protein to calculate hydrophobic interaction and paired pharmacophore interactions with ligands. In addition to molecular surface distance, ligand-contacting areas and hydrogen-bond angles were also introduced to the scoring functions in HotLig. Four individual energy scoring functions for H-bonds, ionic pairs, metal coordination, and hydrophobic effects were derived from 600 protein-ligand complexes, and then, their weighting factors were optimized through an interaction-characterized training set. Success rates of ligand-binding-pose predictions (with a root mean squared deviation of ≤2 Å) for the Wang, GOLD, and Cheng data sets were respectively validated to be 91.0%, 87.0%, and 85.6%. HotLig was found to possess equally good predictive powers for the hydrophilic (88.6%) and hydrophobic subsets (87.5%), and the success rate for the mixed subset was as high as 96.9%. The Spearman correlation coefficients were as good as 0.609 to 0.668, which indicates HotLig also has satisfactory predictive power for binding affinities. These results suggested that the HotLig can analyze diverse ligands, including peptides, and is expected to be a powerful tool for drug design and discovery.

A Unified Model of the GABAA Receptor Comprising Agonist and Benzodiazepine Binding Sites

We present a full-length α1β2γ2 GABA receptor model optimized for agonists and benzodiazepine (BZD) allosteric modulators. We propose binding hypotheses for the agonists GABA, muscimol and THIP and for the allosteric modulator diazepam (DZP). The receptor model is primarily based on the glutamate-gated chloride channel (GluCl) from C. elegans and includes additional structural information from the prokaryotic ligand-gated ion channel ELIC in a few regions. Available mutational data of the binding sites are well explained by the model and the proposed ligand binding poses. We suggest a GABA binding mode similar to the binding mode of glutamate in the GluCl X-ray structure. Key interactions are predicted with residues α1R66, β2T202, α1T129, β2E155, β2Y205 and the backbone of β2S156. Muscimol is predicted to bind similarly, however, with minor differences rationalized with quantum mechanical energy calculations. Muscimol key interactions are predicted to be α1R66, β2T202, α1T129, β2E155, β2Y205 and β2F200. Furthermore, we argue that a water molecule could mediate further interactions between muscimol and the backbone of β2S156 and β2Y157. DZP is predicted to bind with interactions comparable to those of the agonists in the orthosteric site. The carbonyl group of DZP is predicted to interact with two threonines α1T206 and γ2T142, similar to the acidic moiety of GABA. The chlorine atom of DZP is placed near the important α1H101 and the N-methyl group near α1Y159, α1T206, and α1Y209. We present a binding mode of DZP in which the pending phenyl moiety of DZP is buried in the binding pocket and thus shielded from solvent exposure. Our full length GABAA receptor is made available as Model S1.

Mapping histamine H4receptor–ligand binding modes

Computational prediction of ligand binding modes in G protein-coupled receptors (GPCRs) remains a challenging task. Systematic consideration of different protein modelling templates, ligand binding poses, and ligand protonation states in extensive molecular dynamics (MD) simulation studies enabled the prediction of ligand-specific mutation effects in the histamine H4receptor, a key player in inflammation.

Novel Insights of Structure-Based Modeling for RNA-Targeted Drug Discovery

Substantial progress in RNA biology highlights the importance of RNAs (e.g., microRNAs) in diseases and the potential of targeting RNAs for drug discovery. However, the lack of RNA-specific modeling techniques demands the development of new tools for RNA-targeted rational drug design. Herein, we implemented integrated approaches of accurate RNA modeling and virtual screening for RNA inhibitor discovery with the most comprehensive evaluation to date of five docking and 11 scoring methods. For the first time, statistical analysis was heavily employed to assess the significance of our predictions. We found that GOLD:GOLD Fitness and rDock:rDock_solv could accurately predict the RNA ligand poses, and ASP rescoring further improved the ranking of ligand binding poses. Due to the weak correlations (R(2) < 0.3) of existing scoring with experimental binding affinities, we implemented two new RNA-specific scoring functions, iMDLScore1 and iMDLScore2, and obtained better correlations with R(2) = 0.70 and 0.79, respectively. We also proposed a multistep virtual screening approach and demonstrated that rDock:rDock_solv together with iMDLScore2 rescoring obtained the best enrichment on the flexible RNA targets, whereas GOLD:GOLD Fitness combined with rDock_solv rescoring outperformed other methods for rigid RNAs. This study provided practical strategies for RNA modeling and offered new insights into RNA-small molecule interactions for drug discovery.

Conformational Sampling of Flexible Ligand-binding Protein Loops

Protein loops are often involved in diverse biological functions, and some functional loops show conformational changes upon ligand binding. Since this conformational change is directly related to ligand binding pose and protein function, there have been numerous attempts to predict this change accurately. In this study, we show that it is plausible to obtain meaningful ensembles of loop conformations for flexible, ligand-binding protein loops efficiently by applying a loop modeling method. The loop modeling method employs triaxial loop closure algorithm for trial conformation generation and conformational space annealing for global energy optimization. When loop modeling was performed on the framework of ligand-free structure, loop structures within <TEX>$3\AA$</TEX> RMSD from the crystal loop structure for the ligand-bound state were sampled in 4 out of 6 cases. This result is encouraging considering that no information on the ligand-bound state was used during the loop modeling process. We therefore expect that the present loop modeling method will be useful for future developments of flexible protein-ligand docking methods.

Molecular modeling study of checkpoint kinase 1 inhibitors by multiple docking strategies and prime/MM–GBSA calculation

Developing chemicals that inhibit checkpoint kinase 1 (Chk1) is a promising adjuvant therapeutic to improve the efficacy and selectivity of DNA-targeting agents. Reliable prediction of binding-free energy and binding affinity of Chk1 inhibitors can provide a guide for rational drug design. In this study, multiple docking strategies and Prime/Molecular Mechanics Generalized Born Surface Area (Prime/MM-GBSA) calculation were applied to predict the binding mode and free energy for a series of benzoisoquinolinones as Chk1 inhibitors. Reliable docking results were obtained using induced-fit docking and quantum mechanics/molecular mechanics (QM/MM) docking, which showed superior performance on both ligand binding pose and docking score accuracy to the rigid-receptor docking. Then, the Prime/MM-GBSA method based on the docking complex was used to predict the binding-free energy. The combined use of QM/MM docking and Prime/MM-GBSA method could give a high correlation between the predicted binding-free energy and experimentally determined pIC(50) . The molecular docking combined with Prime/MM-GBSA simulation can not only be used to rapidly and accurately predict the binding-free energy of novel Chk1 inhibitors but also provide a novel strategy for lead discovery and optimization targeting Chk1.

Accommodating Protein Flexibility for Structure-Based Drug Design

Proper incorporation of protein flexibility for prediction of binding poses and affinities of small compounds has attracted increasing attention recently in computational drug design. Various approaches have been proposed to accommodate protein flexibility in the prediction of binding modes and the binding free energy of ligands in an efficient manner. In this review, the significance of incorporating protein flexibility is discussed from the structural biophysical point of view, and then various approaches of generating protein conformation ensembles, as well as their successes and limitations, are introduced and compared. Special emphasis is on how to generate a proper ensemble of conformation for a specific purpose, as well as the computational efficiency of various approaches. Different searching algorithms for the prediction of optimal binding poses of ligands, which are the core engines of docking programs, are accounted for. Scoring functions for evaluation of protein-ligand complexes are compared. Two end-point methods of free energy calculation, Molecular Mechanics/Poisson-Boltzmann Surface Area (MM/PBSA) and the Linear Interaction Energy (LIE) method, are briefly reviewed. Finally, we also provide an example for the extension of the conventional protein-ligand docking algorithm for prediction of multiple binding sites and ligand translocation pathways.

Post Processing of Protein-Compound Docking for Fragment-Based Drug Discovery (FBDD): In-Silico Structure-Based Drug Screening and Ligand-Binding Pose Prediction

For fragment-based drug development, both hit (active) compound prediction and docking-pose (protein-ligand complex structure) prediction of the hit compound are important, since chemical modification (fragment linking, fragment evolution) subsequent to the hit discovery must be performed based on the protein-ligand complex structure. However, the naïve protein-compound docking calculation shows poor accuracy in terms of docking-pose prediction. Thus, post-processing of the protein-compound docking is necessary. Recently, several methods for the post-processing of protein-compound docking have been proposed. In FBDD, the compounds are smaller than those for conventional drug screening. This makes it difficult to perform the protein-compound docking calculation. A method to avoid this problem has been reported. Protein-ligand binding free energy estimation is useful to reduce the procedures involved in the chemical modification of the hit fragment. Several prediction methods have been proposed for high-accuracy estimation of protein-ligand binding free energy. This paper summarizes the various computational methods proposed for docking-pose prediction and their usefulness in FBDD.

Conformational transitions upon ligand binding: holo-structure prediction from apo conformations.

Biological function of proteins is frequently associated with the formation of complexes with small-molecule ligands. Experimental structure determination of such complexes at atomic resolution, however, can be time-consuming and costly. Computational methods for structure prediction of protein/ligand complexes, particularly docking, are as yet restricted by their limited consideration of receptor flexibility, rendering them not applicable for predicting protein/ligand complexes if large conformational changes of the receptor upon ligand binding are involved. Accurate receptor models in the ligand-bound state (holo structures), however, are a prerequisite for successful structure-based drug design. Hence, if only an unbound (apo) structure is available distinct from the ligand-bound conformation, structure-based drug design is severely limited. We present a method to predict the structure of protein/ligand complexes based solely on the apo structure, the ligand and the radius of gyration of the holo structure. The method is applied to ten cases in which proteins undergo structural rearrangements of up to 7.1 Å backbone RMSD upon ligand binding. In all cases, receptor models within 1.6 Å backbone RMSD to the target were predicted and close-to-native ligand binding poses were obtained for 8 of 10 cases in the top-ranked complex models. A protocol is presented that is expected to enable structure modeling of protein/ligand complexes and structure-based drug design for cases where crystal structures of ligand-bound conformations are not available.

Conformational Transitions Upon Ligand Binding: Holo Structure Prediction from Apo Conformations

Biological function of proteins is frequently associated with the formation of complexes with small-molecule ligands. Experimental structure determination of such complexes at atomic resolution, however, can be time-consuming and costly. Computational methods for structure prediction of protein/ligand complexes, particularly docking, are as yet restricted by their limited consideration of receptor flexibility, rendering them not applicable for predicting protein/ligand complexes if large conformational changes of the receptor upon ligand binding are involved. Accurate receptor models in the ligand-bound state (holo structures), however, are a prerequisite for successful structure-based drug design.Hence, if only an unbound (apo) structure is available distinct from the ligand-bound conformation, structure-based drug design is severely limited.We present a method to predict the structure of protein/ligand complexes based solely on the apo structure, the ligand and the radius of gyration of the holo structure. The method is applied to ten cases in which proteins undergo structural rearrangements of up to 7.1 A backbone RMSD upon ligand binding. In all cases, receptor models within 1.5 A backbone RMSD to the target were predicted and close-to-native ligand binding poses were obtained for eight of ten cases in the top-ranked complex models.The developed protocol is expected to enable structure modeling of protein/ligand complexes and structure-based drug design for cases where crystal structures of ligand-bound conformations are not available.

Blind docking method combining search of low‐resolution binding sites with ligand pose refinement by molecular dynamics‐based global optimization

This study describes the development of a new blind hierarchical docking method, bhDock, its implementation, and accuracy assessment. The bhDock method uses two-step algorithm. First, a comprehensive set of low-resolution binding sites is determined by analyzing entire protein surface and ranked by a simple score function. Second, ligand position is determined via a molecular dynamics-based method of global optimization starting from a small set of high ranked low-resolution binding sites. The refinement of the ligand binding pose starts from uniformly distributed multiple initial ligand orientations and uses simulated annealing molecular dynamics coupled with guided force-field deformation of protein-ligand interactions to find the global minimum. Assessment of the bhDock method on the set of 37 protein-ligand complexes has shown the success rate of predictions of 78%, which is better than the rate reported for the most cited docking methods, such as AutoDock, DOCK, GOLD, and FlexX, on the same set of complexes.