How does consensus scoring work for virtual library screening? An idealized computer experiment

Contact-based ligand-clustering approach for the identification of active compounds in virtual screening

Evaluation of docking results is one of the most important problems for virtual screening and in silico drug design. Modern approaches for the identification of active compounds in a large data set of docked molecules use energy scoring functions. One of the general and most significant limitations of these methods relates to inaccurate binding energy estimation, which results in false scoring of docked compounds. Automatic analysis of poses using self-organizing maps (AuPosSOM) represents an alternative approach for the evaluation of docking results based on the clustering of compounds by the similarity of their contacts with the receptor. A scoring function was developed for the identification of the active compounds in the AuPosSOM clustered dataset. In addition, the AuPosSOM efficiency for the clustering of compounds and the identification of key contacts considered as important for its activity, were also improved. Benchmark tests for several targets revealed that together with the developed scoring function, AuPosSOM represents a good alternative to the energy-based scoring functions for the evaluation of docking results.

Novel cruzain inhibitors for the treatment of Chagas' disease

The protozoan parasite Trypanosoma cruzi, the etiological agent of Chagas’ disease, affects millions of individuals and continues to be an important global health concern. The poor efficacy and unfavorable side effects of current treatments necessitate novel therapeutics. Cruzain, the major cysteine protease of T. cruzi, is one potential novel target. Recent advances in a class of vinyl sulfone inhibitors are encouraging; however, as most potential therapeutics fail in clinical trials and both disease progression and resistance call for combination therapy with several drugs, the identification of additional classes of inhibitory molecules is essential. Using an exhaustive virtual-screening and experimental validation approach, we identify several additional small-molecule cruzain inhibitors. Further optimization of these chemical scaffolds could lead to the development of novel drugs useful in the treatment of Chagas’ disease.

Application of consensus scoring and principal component analysis for virtual screening against β-secretase (BACE-1)

Background

In order to identify novel chemical classes of β-secretase (BACE-1) inhibitors, an alternative scoring protocol, Principal Component Analysis (PCA), was proposed to summarize most of the information from the original scoring functions and re-rank the results from the virtual screening against BACE-1.

Method

Given a training set (50 BACE-1 inhibitors and 9950 inactive diverse compounds), three rank-based virtual screening methods, individual scoring, conventional consensus scoring and PCA, were judged by the hit number in the top 1% of the ranked list. The docking poses were generated by Surflex, five scoring functions (Surflex_Score, D_Score, G_Score, ChemScore, and PMF_Score) were used for pose extraction. For each pose group, twelve scoring functions (Surflex_Score, D_Score, G_Score, ChemScore, PMF_Score, LigScore1, LigScore2, PLP1, PLP2, jain, Ludi_1, and Ludi_2) were used for the pose rank. For a test set, 113,228 chemical compounds (Sigma-Aldrich® corporate chemical directory) were docked by Surflex, then ranked by the same three ranking methods motioned above to select the potential active compounds for experimental test.

Results

For the training set, the PCA approach yielded consistently superior rankings compared to conventional consensus scoring and single scoring. For the test set, the top 20 compounds according to conventional consensus scoring were experimentally tested, no inhibitor was found. Then, we relied on PCA scoring protocol to test another different top 20 compounds and two low micromolar inhibitors (S450588 and 276065) were emerged through the BACE-1 fluorescence resonance energy transfer (FRET) assay.

Conclusion

The PCA method extends the conventional consensus scoring in a quantitative statistical manner and would appear to have considerable potential for chemical screening applications.

First identification of small-molecule inhibitors of Pontin by combining virtual screening and enzymatic assay

The human protein Pontin, which belongs to the AAA+ (ATPases associated with various cellular activities) family, is overexpressed in several cancers and its silencing in vitro leads to tumour cell growth arrest and apoptosis, making it a good target for cancer therapy. In particular, high levels of expression were found in hepatic tumours for which the therapeutic arsenal is rather limited. The three-dimensional structure of Pontin has been resolved previously, revealing a hexameric assembly with one ADP molecule co-crystallized in each subunit. Using Vina, DrugScore and Xscore, structure-based virtual screening of 2200 commercial molecules was conducted into the ATP-binding site formed by a dimer of Pontin in order to prioritize the best candidates. Complementary to the in silico screening, a versatile and sensitive colorimetric assay was set up to measure the disruption of the ATPase activity of Pontin. This assay allowed the determination of inhibition curves for more than 20 top-scoring compounds, resulting in the identification of four ligands presenting an inhibition constant in the micromolar concentration range. Three of them inhibited tumour cell proliferation. The association of virtual screening and experimental assay thus proved successful for the discovery of the first small-molecule inhibitors of Pontin.

Structure-based virtual screening for drug discovery: a problem-centric review

Structure-based virtual screening (SBVS) has been widely applied in early-stage drug discovery. From a problem-centric perspective, we reviewed the recent advances and applications in SBVS with a special focus on docking-based virtual screening. We emphasized the researchers' practical efforts in real projects by understanding the ligand-target binding interactions as a premise. We also highlighted the recent progress in developing target-biased scoring functions by optimizing current generic scoring functions toward certain target classes, as well as in developing novel ones by means of machine learning techniques.

Identification by virtual screening and in vitro testing of human DOPA decarboxylase inhibitors

Dopa decarboxylase (DDC), a pyridoxal 5'-phosphate (PLP) enzyme responsible for the biosynthesis of dopamine and serotonin, is involved in Parkinson's disease (PD). PD is a neurodegenerative disease mainly due to a progressive loss of dopamine-producing cells in the midbrain. Co-administration of L-Dopa with peripheral DDC inhibitors (carbidopa or benserazide) is the most effective symptomatic treatment for PD. Although carbidopa and trihydroxybenzylhydrazine (the in vivo hydrolysis product of benserazide) are both powerful irreversible DDC inhibitors, they are not selective because they irreversibly bind to free PLP and PLP-enzymes, thus inducing diverse side effects. Therefore, the main goals of this study were (a) to use virtual screening to identify potential human DDC inhibitors and (b) to evaluate the reliability of our virtual-screening (VS) protocol by experimentally testing the "in vitro" activity of selected molecules. Starting from the crystal structure of the DDC-carbidopa complex, a new VS protocol, integrating pharmacophore searches and molecular docking, was developed. Analysis of 15 selected compounds, obtained by filtering the public ZINC database, yielded two molecules that bind to the active site of human DDC and behave as competitive inhibitors with K(i) values ≥10 µM. By performing in silico similarity search on the latter compounds followed by a substructure search using the core of the most active compound we identified several competitive inhibitors of human DDC with K(i) values in the low micromolar range, unable to bind free PLP, and predicted to not cross the blood-brain barrier. The most potent inhibitor with a K(i) value of 500 nM represents a new lead compound, targeting human DDC, that may be the basis for lead optimization in the development of new DDC inhibitors. To our knowledge, a similar approach has not been reported yet in the field of DDC inhibitors discovery.

Solvent interaction energy calculations on molecular dynamics trajectories: increasing the efficiency using systematic frame selection

End-point methods such as linear interaction energy (LIE) analysis, molecular mechanics generalized Born solvent-accessible surface (MM/GBSA), and solvent interaction energy (SIE) analysis have become popular techniques to calculate the free energy associated with protein-ligand binding. Such methods typically use molecular dynamics (MD) simulations to generate an ensemble of protein structures that encompasses the bound and unbound states. The energy evaluation method (LIE, MM/GBSA, or SIE) is subsequently used to calculate the energy of each member of the ensemble, thus providing an estimate of the average free energy difference between the bound and unbound states. The workflow requiring both MD simulation and energy calculation for each frame and each trajectory proves to be computationally expensive. In an attempt to reduce the high computational cost associated with end-point methods, we study several methods by which frames may be intelligently selected from the MD simulation including clustering and address the question of how the number of selected frames influences the accuracy of the SIE calculations.

CSAR benchmark exercise of 2010: combined evaluation across all submitted scoring functions

As part of the Community Structure-Activity Resource (CSAR) center, a set of 343 high-quality, protein–ligand crystal structures were assembled with experimentally determined K_d or K_i information from the literature. We encouraged the community to score the crystallographic poses of the complexes by any method of their choice. The goal of the exercise was to (1) evaluate the current ability of the field to predict activity from structure and (2) investigate the properties of the complexes and methods that appear to hinder scoring. A total of 19 different methods were submitted with numerous parameter variations for a total of 64 sets of scores from 16 participating groups. Linear regression and nonparametric tests were used to correlate scores to the experimental values. Correlation to experiment for the various methods ranged R² = 0.58–0.12, Spearman ρ = 0.74–0.37, Kendall τ = 0.55–0.25, and median unsigned error = 1.00–1.68 pK_d units. All types of scoring functions—force field based, knowledge based, and empirical—had examples with high and low correlation, showing no bias/advantage for any particular approach. The data across all the participants were combined to identify 63 complexes that were poorly scored across the majority of the scoring methods and 123 complexes that were scored well across the majority. The two sets were compared using a Wilcoxon rank-sum test to assess any significant difference in the distributions of >400 physicochemical properties of the ligands and the proteins. Poorly scored complexes were found to have ligands that were the same size as those in well-scored complexes, but hydrogen bonding and torsional strain were significantly different. These comparisons point to a need for CSAR to develop data sets of congeneric series with a range of hydrogen-bonding and hydrophobic characteristics and a range of rotatable bonds.

Systematic exploitation of multiple receptor conformations for virtual ligand screening

The role of virtual ligand screening in modern drug discovery is to mine large chemical collections and to prioritize for experimental testing a comparatively small and diverse set of compounds with expected activity against a target. Several studies have pointed out that the performance of virtual ligand screening can be improved by taking into account receptor flexibility. Here, we systematically assess how multiple crystallographic receptor conformations, a powerful way of discretely representing protein plasticity, can be exploited in screening protocols to separate binders from non-binders. Our analyses encompass 36 targets of pharmaceutical relevance and are based on actual molecules with reported activity against those targets. The results suggest that an ensemble receptor-based protocol displays a stronger discriminating power between active and inactive molecules as compared to its standard single rigid receptor counterpart. Moreover, such a protocol can be engineered not only to enrich a higher number of active compounds, but also to enhance their chemical diversity. Finally, some clear indications can be gathered on how to select a subset of receptor conformations that is most likely to provide the best performance in a real life scenario.

Formal Estimation of Errors in Computed Absolute Interaction Energies of Protein-ligand Complexes

A largely unsolved problem in computational biochemistry is the accurate prediction of binding affinities of small ligands to protein receptors. We present a detailed analysis of the systematic and random errors present in computational methods through the use of error probability density functions, specifically for computed interaction energies between chemical fragments comprising a protein-ligand complex. An HIV-II protease crystal structure with a bound ligand (indinavir) was chosen as a model protein-ligand complex. The complex was decomposed into twenty-one (21) interacting fragment pairs, which were studied using a number of computational methods. The chemically accurate complete basis set coupled cluster theory (CCSD(T)/CBS) interaction energies were used as reference values to generate our error estimates. In our analysis we observed significant systematic and random errors in most methods, which was surprising especially for parameterized classical and semiempirical quantum mechanical calculations. After propagating these fragment-based error estimates over the entire protein-ligand complex, our total error estimates for many methods are large compared to the experimentally determined free energy of binding. Thus, we conclude that statistical error analysis is a necessary addition to any scoring function attempting to produce reliable binding affinity predictions.

Structural ensemble in computational drug screening

IMPORTANCE OF THE FIELD

Structure-based in silico drug screening is now widely used in drug development projects. Structure-based in silico drug screening is generally performed using a protein-compound docking program and docking scoring function. Many docking programs have been developed over the last 2 decades, but their prediction accuracy remains insufficient.

AREAS COVERED IN THIS REVIEW

This review highlights the recent progress of the post-processing of protein-compound complexes after docking.

WHAT THE READER WILL GAIN

These methods utilize ensembles of docking poses of compounds to improve the prediction accuracy for the ligand-docking pose and screening results. While the individual docking poses are not reliable, the free energy surface or the most probable docking pose can be estimated from the ensemble of docking poses.

TAKE HOME MESSAGE

The protein-compound docking program provides an arbitral rather than a canonical ensemble of docking poses. When the ensemble of docking poses satisfies the canonical ensemble, we can discuss how these post-docking analysis methods work and fail. Thus, improvements to the docking software will be needed in order to generate well-defined ensembles of docking poses.

Virtual screening for ligands of the insect molting hormone receptor

Insect growth is regulated by the orchestrated event of ecdysteroids and their receptor proteins. Agonists/antagonists of ecdysteroid receptor are predicted to disrupt normal growth, providing good candidates of new insecticides. A database of over 2 million compounds was subjected to a shape-based virtual screening cascade to identify novel nonsteroidal hits similar to the known EcR ligand ponasterone A. Testing revealed micromolar hits against two strains of insect cells. Docking experiments against EcR were used to support the predicted binding mode of these ligands based on their overlay to ponasterone A.

Ligand supported homology modeling and docking evaluation of CCR2: docked pose selection by consensus scoring

Chemokine receptor 2 (CCR2) is a G-protein coupled receptor (GPCR) and a crucial target for various inflammatory and autoimmune diseases. The structure based antagonists design for many GPCRs, including CCR2, is restricted by the lack of an experimental three dimensional structure. Homology modeling is widely used for the study of GPCR-ligand binding. Since there is substantial diversity for the ligand binding pocket and binding modes among GPCRs, the receptor-ligand binding mode predictions should be derived from homology modeling with supported ligand information. Thus, we modeled the binding of our proprietary CCR2 antagonist using ligand supported homology modeling followed by consensus scoring the docking evaluation based on all modeled binding sites. The protein-ligand model was then validated by visual inspection of receptor-ligand interaction for consistency of published site-directed mutagenesis data and virtual screening a decoy compound database. This model was able to successfully identify active compounds within the decoy database. Finally, additional hit compounds were identified through a docking-based virtual screening of a commercial database, followed by a biological assay to validate CCR2 inhibitory activity. Thus, this procedure can be employed to screen a large database of compounds to identify new CCR2 antagonists.

Virtual screening for lead discovery

The identification of small drug-like compounds that selectively inhibit the function of biological targets has historically been a major focus in the pharmaceutical industry, and in recent years, has generated much interest in academia as well. Drug-like compounds are valuable as chemical genetics tools to probe biological pathways in a reversible, dose- and time-dependent manner for drug target identification. In addition, small molecule compounds can be used to characterize the shape and charge preferences of macromolecular binding sites, for both structure-based and ligand-based drug design. High-throughput screening is the most common experimental method used to identify lead compounds. Because of the cost, time, and resources required for performing high-throughput screening for compound libraries, the use of alternative strategies is necessary for facilitating lead discovery. Virtual screening has been successful in prioritizing large chemical libraries to identify experimentally active compounds, serving as a practical and effective alternative to high-throughput screening. Methodologies used in virtual screening such as molecular docking and scoring have advanced to the point where they can rapidly and accurately identify lead compounds in addition to predicting native binding conformations. This chapter provides instructions on how to perform a virtual screen using freely available tools for structure-based lead discovery.

Structure-based virtual screening and electrophysiological evaluation of new chemotypes of K(v)1.5 channel blockers

Atrial fibrillation (AF) is the most prevalent nonfatal cardiac rhythm disorder associated with an increased risk of heart failure and stroke. Considering the ventricular side effects induced by anti-arrhythmic agents in current use, K(v)1.5 channel blockers have attracted a great deal of deliberation owing to their selective actions on atrial electrophysiology. Herein we report new chemotypes of K(v)1.5 channel blockers that were identified through a combination of structure-based virtual screening and in silico druglike property prediction including six scoring functions, as well as electrophysiological evaluation. Among them, five of the 18 compounds exhibited >50 % blockade ratio at 10 microM, and have structural features different from conventional K(v)1.5 channel blockers. These novel scaffolds could serve as hits for further optimization and SAR studies for the discovery of selective agents to treat AF.

Scoring functions and their evaluation methods for protein-ligand docking: recent advances and future directions

The scoring function is one of the most important components in structure-based drug design. Despite considerable success, accurate and rapid prediction of protein-ligand interactions is still a challenge in molecular docking. In this perspective, we have reviewed three basic types of scoring functions (force-field, empirical, and knowledge-based) and the consensus scoring technique that are used for protein-ligand docking. The commonly-used assessment criteria and publicly available protein-ligand databases for performance evaluation of the scoring functions have also been presented and discussed. We end with a discussion of the challenges faced by existing scoring functions and possible future directions for developing improved scoring functions.

Small molecule correctors of F508del-CFTR discovered by structure-based virtual screening

Folding correctors of F508del-CFTR were discovered by in silico structure-based screening utilizing homology models of CFTR. The intracellular segment of CFTR was modeled and three cavities were identified at inter-domain interfaces: (1) Interface between the two Nucleotide Binding Domains (NBDs); (2) Interface between NBD1 and Intracellular Loop (ICL) 4, in the region of the F508 deletion; (3) multi-domain interface between NBD1:2:ICL1:2:4. We hypothesized that compounds binding at these interfaces may improve the stability of the protein, potentially affecting the folding yield or surface stability. In silico structure-based screening was performed at the putative binding-sites and a total of 496 candidate compounds from all three sites were tested in functional assays. A total of 15 compounds, representing diverse chemotypes, were identified as F508del folding correctors. This corresponds to a 3% hit rate, ~tenfold higher than hit rates obtained in corresponding high-throughput screening campaigns. The same binding sites also yielded potentiators and, most notably, compounds with a dual corrector-potentiator activity (dual-acting). Compounds harboring both activity types may prove to be better leads for the development of CF therapeutics than either pure correctors or pure potentiators. To the best of our knowledge this is the first report of structure-based discovery of CFTR modulators.

Advances and challenges in protein-ligand docking

Molecular docking is a widely-used computational tool for the study of molecular recognition, which aims to predict the binding mode and binding affinity of a complex formed by two or more constituent molecules with known structures. An important type of molecular docking is protein-ligand docking because of its therapeutic applications in modern structure-based drug design. Here, we review the recent advances of protein flexibility, ligand sampling, and scoring functions—the three important aspects in protein-ligand docking. Challenges and possible future directions are discussed in the Conclusion.

Improving predicted protein loop structure ranking using a Pareto-optimality consensus method

Background

Accurate protein loop structure models are important to understand functions of many proteins. Identifying the native or near-native models by distinguishing them from the misfolded ones is a critical step in protein loop structure prediction.

Results

We have developed a Pareto Optimal Consensus (POC) method, which is a consensus model ranking approach to integrate multiple knowledge- or physics-based scoring functions. The procedure of identifying the models of best quality in a model set includes: 1) identifying the models at the Pareto optimal front with respect to a set of scoring functions, and 2) ranking them based on the fuzzy dominance relationship to the rest of the models. We apply the POC method to a large number of decoy sets for loops of 4- to 12-residue in length using a functional space composed of several carefully-selected scoring functions: Rosetta, DOPE, DDFIRE, OPLS-AA, and a triplet backbone dihedral potential developed in our lab. Our computational results show that the sets of Pareto-optimal decoys, which are typically composed of ~20% or less of the overall decoys in a set, have a good coverage of the best or near-best decoys in more than 99% of the loop targets. Compared to the individual scoring function yielding best selection accuracy in the decoy sets, the POC method yields 23%, 37%, and 64% less false positives in distinguishing the native conformation, indentifying a near-native model (RMSD < 0.5A from the native) as top-ranked, and selecting at least one near-native model in the top-5-ranked models, respectively. Similar effectiveness of the POC method is also found in the decoy sets from membrane protein loops. Furthermore, the POC method outperforms the other popularly-used consensus strategies in model ranking, such as rank-by-number, rank-by-rank, rank-by-vote, and regression-based methods.

Conclusions

By integrating multiple knowledge- and physics-based scoring functions based on Pareto optimality and fuzzy dominance, the POC method is effective in distinguishing the best loop models from the other ones within a loop model set.

Limits of Free Energy Computation for Protein-Ligand Interactions

A detailed error analysis is presented for the computation of protein-ligand interaction energies. In particular, we show that it is probable that even highly accurate computed binding free energies have errors that represent a large percentage of the target free energies of binding. This is due to the observation that the error for computed energies quasi-linearly increases with the increasing number of interactions present in a protein-ligand complex. This principle is expected to hold true for any system that involves an ever increasing number of inter or intra-molecular interactions (e.g. ab initio protein folding). We introduce the concept of best-case scenario errors (BCS(errors)) that can be routinely applied to docking and scoring exercises and used to provide errors bars for the computed binding free energies. These BCS(errors) form a basis by which one can evaluate the outcome of a docking and scoring exercise. Moreover, the resultant error analysis enables the formation of an hypothesis that defines the best direction to proceed in order to improve scoring functions used in molecular docking studies.

Development of improved models for phosphodiesterase-4 inhibitors with a multi-conformational structure-based QSAR method

Phosphodiesterase-4 (PDE-4) is an important drug target for several diseases, including COPD (chronic obstructive pulmonary disorder) and neurodegenerative diseases. In this paper, we describe the development of improved QSAR (quantitative structure-activity relationship) models using a novel multi-conformational structure-based pharmacophore key (MC-SBPPK) method. Similar to our previous work, this method calculates molecular descriptors based on the matching of a molecule's pharmacophore features with those of the target binding pocket. Therefore, these descriptors are PDE4-specific, and most relevant to the problem under study. Furthermore, this work expands our previous SBPPK QSAR method by explicitly including multiple conformations of the PDE-4 inhibitors in the regression analysis, and thus addresses the issue of molecular flexibility. The nonlinear regression problem resulted from including multiple conformations has been transformed into a linear equation and solved by an iterative partial least square (iPLS) procedure, according to the Lukacova-Balaz scheme. 35 PDE-4 inhibitors have been analyzed with this new method, and predictive models have been developed. Based on the prediction statistics for both the training set and the test set, these new models are more robust and predictive than those obtained by traditional ligand-based QSAR techniques as well as that obtained with the SBPPK method reported in our previous work. As a result, multiple predictive models have been added to the collection of QSAR models for PDE4 inhibitors. Collectively, these models will be useful for the discovery of new drug candidates targeting the PDE-4 enzyme.

In silico screening for non-nucleoside HIV-1 reverse transcriptase inhibitors using physicochemical filters and high-throughput docking followed by in vitro evaluation

Reverse transcriptase, being the pivot in human immunodeficiency virus replication, is one of the most attractive targets for the development of new antiretroviral agents. We applied a virtual screening workflow based on a combination of physicochemical filters with high-throughput rigid molecular docking to discover novel efficient lead scaffolds for human immunodeficiency virus type 1 reverse transcriptase inhibition. In our protocol, different filters were employed to enrich the lead-likeness and improve the ligands efficiency of the filtered compounds. Out of the 238,819 compounds included in the National Cancer Institute database, 500 virtual screening hits were retrieved employing FILTER and FRED (molecular docking engine) softwares. Four compounds from the 20 highest ranking scored hits tested positive in human immunodeficiency virus type 1 reverse transcriptase using non-radioactive colorimetric assay method. These results demonstrate that our virtual screening protocol is able to enrich novel scaffolds for human immunodeficiency virus type 1 reverse transcriptase inhibition that could be useful for drug development in the area of acquired immune-deficiency syndrome treatment.

Structure-based virtual screening for novel inhibitors of Japanese encephalitis virus NS3 helicase/nucleoside triphosphatase

Japanese encephalitis (JE) is a significant cause of human morbidity and mortality throughout Asia and Africa. Vaccines have reduced the incidence of JE in some countries, but no specific antiviral therapy is currently available. The NS3 protein of Japanese encephalitis virus (JEV) is a multifunctional protein combining protease, helicase and nucleoside 5′‐triphosphatase (NTPase) activities. The crystal structure of the catalytic domain of this protein has recently been solved using a roentgenographic method. This enabled structure‐based virtual screening for novel inhibitors of JEV NS3 helicase/NTPase. The aim of the present research was to identify novel potent medicinal substances for the treatment of JE. In the first step of studies, the natural ligand ATP and two known JEV NS3 helicase/NTPase inhibitors were docked to their molecular target. The refined structure of the enzyme was used to construct a pharmacophore model for JEV NS3 helicase/NTPase inhibitors. The freely available ZINC database of lead‐like compounds was then screened for novel inhibitors. About 1 161 000 compounds have been screened and 15 derivatives of the highest scores have been selected. These compounds were docked to the JEV NS3 helicase/NTPase to examine their binding mode and verify screening results by consensus scoring procedure.

Discovery of inhibitors of lupin diadenosine 5',5'''-P(1),P(4)-tetraphosphate hydrolase by virtual screening

Novel inhibitors of lupin diadenosine 5',5'''-P(1),P(4)-tetraphosphate (Ap(4)A) hydrolase have been identified by in silico screening of a large virtual chemical library. Compounds were ranked on the basis of a consensus from six scoring functions. From the top 100 ranked compounds six were selected and initially screened for inhibitory activity using a single concentration isothermal titration calorimetry assay. Two of these compounds that showed excellent solubility properties were further analyzed, but only one [NSC51531; 2-((8-hydroxy-4-(4-methyl-2-sulfoanilino)-9,10-dioxo-9,10-dihydro-1-anthracenyl)amino)-5-methylbenzenesulfonic acid] exhibited competitive inhibition with a K(i) of 1 microM. A structural analogue of this compound also exhibited competitive inhibition with a comparable K(i) of 2.9 microM. (1)H, (15)N NMR spectroscopy was used to map the binding site of NSC51531 on lupin Ap(4)A hydrolase and demonstrated that the compound bound specifically in the substrate-binding site, consistent with the competitive inhibition results. Binding of NSC51531 to the human form of Ap(4)A hydrolase is nonspecific, suggesting that this compound may represent a useful lead in the design of specific inhibitors of the plant-like form of Ap(4)A hydrolases.

Comparative assessment of scoring functions on a diverse test set

Scoring functions are widely applied to the evaluation of protein-ligand binding in structure-based drug design. We have conducted a comparative assessment of 16 popular scoring functions implemented in main-stream commercial software or released by academic research groups. A set of 195 diverse protein-ligand complexes with high-resolution crystal structures and reliable binding constants were selected through a systematic nonredundant sampling of the PDBbind database and used as the primary test set in our study. All scoring functions were evaluated in three aspects, that is, "docking power", "ranking power", and "scoring power", and all evaluations were independent from the context of molecular docking or virtual screening. As for "docking power", six scoring functions, including GOLD::ASP, DS::PLP1, DrugScore(PDB), GlideScore-SP, DS::LigScore, and GOLD::ChemScore, achieved success rates over 70% when the acceptance cutoff was root-mean-square deviation < 2.0 A. Combining these scoring functions into consensus scoring schemes improved the success rates to 80% or even higher. As for "ranking power" and "scoring power", the top four scoring functions on the primary test set were X-Score, DrugScore(CSD), DS::PLP, and SYBYL::ChemScore. They were able to correctly rank the protein-ligand complexes containing the same type of protein with success rates around 50%. Correlation coefficients between the experimental binding constants and the binding scores computed by these scoring functions ranged from 0.545 to 0.644. Besides the primary test set, each scoring function was also tested on four additional test sets, each consisting of a certain number of protein-ligand complexes containing one particular type of protein. Our study serves as an updated benchmark for evaluating the general performance of today's scoring functions. Our results indicate that no single scoring function consistently outperforms others in all three aspects. Thus, it is important in practice to choose the appropriate scoring functions for different purposes.

Discovery of novel HIV entry inhibitors for the CXCR4 receptor by prospective virtual screening

The process of HIV entry begins with the binding of the viral envelope glycoprotein gp120 to both the CD4 receptor and one of CXCR4 or CCR5 chemokine coreceptors. There is currently considerable interest in developing novel ligands which can attach to these coreceptors and hence block virus-cell fusion. This article compares the application of structure-based (docking) and ligand-based (QSAR analyses, pharmacophore modeling, and shape matching) virtual screening tools to find new potential HIV entry inhibitors for the CXCR4 receptor. The comparison is based on retrospective virtual screening of a library containing different known CXCR4 inhibitors from the literature, a smaller set of active CXCR4 inhibitors selected from a large combinatorial virtual library and synthesized by us, and some druglike presumed inactive molecules as the reference set. The enrichment factors and diversity of the retrieved molecular scaffolds in the virtual hit lists was determined. Once the different virtual screening approaches had been validated and the best parameters had been selected, prospective virtual screening of our virtual library was applied to identify new anti-HIV compounds using the same protocol as in the retrospective virtual screening analysis. The compounds selected using these computational tools were subsequently synthesized and assayed and showed activity values ranging from 4 to 0.022 microg/mL.

FINDSITE: a threading-based approach to ligand homology modeling

Ligand virtual screening is a widely used tool to assist in new pharmaceutical discovery. In practice, virtual screening approaches have a number of limitations, and the development of new methodologies is required. Previously, we showed that remotely related proteins identified by threading often share a common binding site occupied by chemically similar ligands. Here, we demonstrate that across an evolutionarily related, but distant family of proteins, the ligands that bind to the common binding site contain a set of strongly conserved anchor functional groups as well as a variable region that accounts for their binding specificity. Furthermore, the sequence and structure conservation of residues contacting the anchor functional groups is significantly higher than those contacting ligand variable regions. Exploiting these insights, we developed FINDSITE(LHM) that employs structural information extracted from weakly related proteins to perform rapid ligand docking by homology modeling. In large scale benchmarking, using the predicted anchor-binding mode and the crystal structure of the receptor, FINDSITE(LHM) outperforms classical docking approaches with an average ligand RMSD from native of approximately 2.5 A. For weakly homologous receptor protein models, using FINDSITE(LHM), the fraction of recovered binding residues and specific contacts is 0.66 (0.55) and 0.49 (0.38) for highly confident (all) targets, respectively. Finally, in virtual screening for HIV-1 protease inhibitors, using similarity to the ligand anchor region yields significantly improved enrichment factors. Thus, the rather accurate, computationally inexpensive FINDSITE(LHM) algorithm should be a useful approach to assist in the discovery of novel biopharmaceuticals.

Development of a novel virtual screening cascade protocol to identify potential trypanothione reductase inhibitors

The implementation of a novel sequential computational approach that can be used effectively for virtual screening and identification of prospective ligands that bind to trypanothione reductase (TryR) is reported. The multistep strategy combines a ligand-based virtual screening for building an enriched library of small molecules with a docking protocol (AutoDock, X-Score) for screening against the TryR target. Compounds were ranked by an exhaustive conformational consensus scoring approach that employs a rank-by-rank strategy by combining both scoring functions. Analysis of the predicted ligand−protein interactions highlights the role of bulky quaternary amine moieties for binding affinity. The scaffold hopping (SHOP) process derived from this computational approach allowed the identification of several chemotypes, not previously reported as antiprotozoal agents, which includes dibenzothiepine, dibenzooxathiepine, dibenzodithiepine, and polycyclic cationic structures like thiaazatetracyclo-nonadeca-hexaen-3-ium. Assays measuring the inhibiting effect of these compounds on T. cruzi and T. brucei TryR confirm their potential for further rational optimization.

The identification of novel PLC-gamma inhibitors using virtual high throughput screening

Phospholipase C-gamma (PLC-gamma) has been identified as a possible biological target for anticancer drug therapy but suitable inhibitors are lacking. Therefore, in order to identify active compounds (hits) virtual high throughput screening was performed. The crystal structure of the PLC-delta isoform was used as a model docking scaffold since no crystallographic data are available on its gamma counterpart. A pilot screen was performed using approximately 9.2x10(4) compounds, where the robustness of the methodology was tested. This was followed by the main screening effort where approximately 4.4x10(5) compounds were used. In both cases, plausible compounds were identified (virtual hits) and a selection of these was experimentally tested. The most potent compounds were in the single digit micro-molar range as determined from the biochemical (Flashplate) assay. This translated into approximately 15 microM in a functional assay in cells. About 30% of the virtual hits showed activity against PLC-gamma (IC(50)<50 microM).

Virtual screening to identify lead inhibitors for bacterial NAD synthetase (NADs)

Virtual screening was employed to identify new drug-like inhibitors of NAD synthetase (NADs) as antibacterial agents. Four databases of commercially available compounds were docked against three subsites of the NADs active site using FlexX in conjunction with CScore. Over 200 commercial compounds were purchased and evaluated in enzyme inhibition and antibacterial assays. 18 compounds inhibited NADs at or below 100 microM (7.6% hit rate), and two were selected for future SAR studies.

A method to enhance the hit ratio by a combination of structure-based drug screening and ligand-based screening

We examined the procedures to combine two different in silico drug-screening results to achieve a high hit ratio. When the 3D structure of the target protein and some active compounds are known, both structure-based and ligand-based in silico screening methods can be applied. In the present study, the machine-learning score modification multiple target screening (MSM-MTS) method was adopted as a structure-based screening method, and the machine-learning docking score index (ML-DSI) method was adopted as a ligand-based screening method. To combine the predicted compound’s sets by these two screening methods, we examined the product of the sets (consensus set) and the sum of the sets. As a result, the consensus set achieved a higher hit ratio than the sum of the sets and than either individual predicted set. In addition, the current combination was shown to be robust enough for the structural diversities both in different crystal structure and in snapshot structures during molecular dynamics simulations.

MedusaScore: an accurate force field-based scoring function for virtual drug screening

Virtual screening is becoming an important tool for drug discovery. However, the application of virtual screening has been limited by the lack of accurate scoring functions. Here, we present a novel scoring function, MedusaScore, for evaluating protein-ligand binding. MedusaScore is based on models of physical interactions that include van der Waals, solvation, and hydrogen bonding energies. To ensure the best transferability of the scoring function, we do not use any protein-ligand experimental data for parameter training. We then test the MedusaScore for docking decoy recognition and binding affinity prediction and find superior performance compared to other widely used scoring functions. Statistical analysis indicates that one source of inaccuracy of MedusaScore may arise from the unaccounted entropic loss upon ligand binding, which suggests avenues of approach for further MedusaScore improvement.

Rescoring docking hit lists for model cavity sites: predictions and experimental testing

Molecular docking computationally screens thousands to millions of organic molecules against protein structures, looking for those with complementary fits. Many approximations are made, often resulting in low "hit rates." A strategy to overcome these approximations is to rescore top-ranked docked molecules using a better but slower method. One such is afforded by molecular mechanics-generalized Born surface area (MM-GBSA) techniques. These more physically realistic methods have improved models for solvation and electrostatic interactions and conformational change compared to most docking programs. To investigate MM-GBSA rescoring, we re-ranked docking hit lists in three small buried sites: a hydrophobic cavity that binds apolar ligands, a slightly polar cavity that binds aryl and hydrogen-bonding ligands, and an anionic cavity that binds cationic ligands. These sites are simple; consequently, incorrect predictions can be attributed to particular errors in the method, and many likely ligands may actually be tested. In retrospective calculations, MM-GBSA techniques with binding-site minimization better distinguished the known ligands for each cavity from the known decoys compared to the docking calculation alone. This encouraged us to test rescoring prospectively on molecules that ranked poorly by docking but that ranked well when rescored by MM-GBSA. A total of 33 molecules highly ranked by MM-GBSA for the three cavities were tested experimentally. Of these, 23 were observed to bind--these are docking false negatives rescued by rescoring. The 10 remaining molecules are true negatives by docking and false positives by MM-GBSA. X-ray crystal structures were determined for 21 of these 23 molecules. In many cases, the geometry prediction by MM-GBSA improved the initial docking pose and more closely resembled the crystallographic result; yet in several cases, the rescored geometry failed to capture large conformational changes in the protein. Intriguingly, rescoring not only rescued docking false positives, but also introduced several new false positives into the top-ranking molecules. We consider the origins of the successes and failures in MM-GBSA rescoring in these model cavity sites and the prospects for rescoring in biologically relevant targets.

Differentiation of AmpC beta-lactamase binders vs. decoys using classification kNN QSAR modeling and application of the QSAR classifier to virtual screening

The use of inaccurate scoring functions in docking algorithms may result in the selection of compounds with high predicted binding affinity that nevertheless are known experimentally not to bind to the target receptor. Such falsely predicted binders have been termed 'binding decoys'. We posed a question as to whether true binders and decoys could be distinguished based only on their structural chemical descriptors using approaches commonly used in ligand based drug design. We have applied the k-Nearest Neighbor (kNN) classification QSAR approach to a dataset of compounds characterized as binders or binding decoys of AmpC beta-lactamase. Models were subjected to rigorous internal and external validation as part of our standard workflow and a special QSAR modeling scheme was employed that took into account the imbalanced ratio of inhibitors to non-binders (1:4) in this dataset. 342 predictive models were obtained with correct classification rate (CCR) for both training and test sets as high as 0.90 or higher. The prediction accuracy was as high as 100% (CCR = 1.00) for the external validation set composed of 10 compounds (5 true binders and 5 decoys) selected randomly from the original dataset. For an additional external set of 50 known non-binders, we have achieved the CCR of 0.87 using very conservative model applicability domain threshold. The validated binary kNN QSAR models were further employed for mining the NCGC AmpC screening dataset (69653 compounds). The consensus prediction of 64 compounds identified as screening hits in the AmpC PubChem assay disagreed with their annotation in PubChem but was in agreement with the results of secondary assays. At the same time, 15 compounds were identified as potential binders contrary to their annotation in PubChem. Five of them were tested experimentally and showed inhibitory activities in millimolar range with the highest binding constant K(i) of 135 microM. Our studies suggest that validated QSAR models could complement structure based docking and scoring approaches in identifying promising hits by virtual screening of molecular libraries.