Virtual screening to enrich hit lists from high-throughput screening: a case study on small-molecule inhibitors of angiogenin

"Hit lists" generated by high-throughput screening (HTS) typically contain a large percentage of false positives, making follow-up assays necessary to distinguish active from inactive substances. Here we present a method for improving the accuracy of HTS hit lists by computationally based virtual screening (VS) of the corresponding chemical libraries and selecting hits by HTS/VS consensus. This approach was applied in a case study on the target-enzyme angiogenin, a potent inducer of angiogenesis. In conjunction with HTS of the National Cancer Institute Diversity Set and ChemBridge DIVERSet E (approximately 18,000 compounds total), VS was performed with two flexible library docking/scoring methods, DockVision/Ludi and GOLD. Analysis of the results reveals that dramatic enrichment of the HTS hit rate can be achieved by selecting compounds in consensus with one or both of the VS functions. For example, HTS hits ranked in the top 2% by GOLD included 42% of the true hits, but only 8% of the false positives; this represents a sixfold enrichment over the HTS hit rate. Notably, the HTS/VS method was effective in selecting out inhibitors with midmicromolar dissociation constants typical of leads commonly obtained in primary screens.     

H‐bond network optimization in protein–protein complexes: Are all‐atom force field scores enough?

Structural prediction of protein-protein complexes given the structures of the two interacting compounds in their unbound state is a key problem in biophysics. In addition to the problem of sampling of near-native orientations, one of the modeling main difficulties is to discriminate true from false positives. Here, we present a hierarchical protocol for docking refinement able to discriminate near native poses from a group of docking candidates. The main idea is to combine an efficient sampling of the full system hydrogen bond network and side chains, together with an all-atom force field and a surface generalized born implicit solvent. We tested our method on a set of twenty two complexes containing a near-native solution within the top 100 docking poses, obtaining a near native solution as the top pose in 70% of the cases. We show that all atom force fields optimized H-bond networks do improve significantly state of the art scoring functions.     

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.     

Identification of Phytochemicals Targeting c-Met Kinase Domain using Consensus Docking and Molecular Dynamics Simulation Studies

c-Met receptor tyrosine kinase is a proto-oncogene whose aberrant activation is attributed to a lower rate of survival in most cancers. Natural product-derived inhibitors known as “fourth generation inhibitors” constitute more than 60% of anticancer drugs. Furthermore, consensus docking approach has recently been introduced to augment docking accuracy and reduce false positives during a virtual screening. In order to obtain novel small-molecule Met inhibitors, consensus docking approach was performed using Autodock Vina and Autodock 4.2 to virtual screen Naturally Occurring Plant-based Anti-cancer Compound–Activity–Target database against active and inactive conformation of c-Met kinase domain structure. Two hit molecules that were in line with drug-likeness criteria, desired docking score, and binding pose were subjected to molecular dynamics simulations to elucidate intermolecular contacts in protein–ligand complexes. Analysis of molecular dynamics simulations and molecular mechanics Poisson–Boltzmann surface area studies showed that ZINC08234189 is a plausible inhibitor for the active state of c-Met, whereas ZINC03871891 may be more effective toward active c-Met kinase domain compared to the inactive form due to higher binding energy. Our analysis showed that both the hit molecules formed hydrogen bonds with key residues of the hinge region (P1158, M1160) in the active form, which is a hallmark of kinase domain inhibitors. Considering the pivotal role of HGF/c-Met signaling in carcinogenesis, our results propose ZINC08234189 and ZINC03871891 as the therapeutic options to surmount Met-dependent cancers.     

Accurate decisions in an uncertain world: collective cognition increases true positives while decreasing false positives

In a wide range of contexts, including predator avoidance, medical decision-making and security screening, decision accuracy is fundamentally constrained by the trade-off between true and false positives. Increased true positives are possible only at the cost of increased false positives; conversely, decreased false positives are associated with decreased true positives. We use an integrated theoretical and experimental approach to show that a group of decision-makers can overcome this basic limitation. Using a mathematical model, we show that a simple quorum decision rule enables individuals in groups to simultaneously increase true positives and decrease false positives. The results from a predator-detection experiment that we performed with humans are in line with these predictions: (i) after observing the choices of the other group members, individuals both increase true positives and decrease false positives, (ii) this effect gets stronger as group size increases, (iii) individuals use a quorum threshold set between the average true- and false-positive rates of the other group members, and (iv) individuals adjust their quorum adaptively to the performance of the group. Our results have broad implications for our understanding of the ecology and evolution of group-living animals and lend themselves for applications in the human domain such as the design of improved screening methods in medical, forensic, security and business applications.     

How to choose relevant multiple receptor conformations for virtual screening: a test case of Cdk2 and normal mode analysis

Better treatment of protein flexibility is essential in structure-based drug design projects such as virtual screening and protein-ligand docking. Diversity in ligand-binding mechanisms and receptor conformational changes makes it difficult to treat dynamic features of the receptor during the docking simulation. Thus, the use of pregenerated multiple receptor conformations is applied today in virtual screening studies. However, generation of a small relevant set of receptor conformations remains challenging. To address this problem, we propose a new protocol for the generation of multiple receptor conformations via normal mode analysis and for the selection of several receptor conformations suitable for docking/virtual screening. We validated this protocol on cyclin-dependent kinase 2, which possesses a binding site located at the interface between two subdomains and is known to undergo significant conformational changes in the active site region upon ligand binding. We believe that the suggested rules for the choice of suitable receptor conformations can be applied to other targets when dealing with in silico screening on flexible receptors.     

Quantitative assessment of hit detection and confirmation in single and duplicate high-throughput screenings

The process of identifying active targets (hits) in high-throughput screening (HTS) usually involves 2 steps: first, removing or adjusting for systematic variation in the measurement process so that extreme values represent strong biological activity instead of systematic biases such as plate effect or edge effect and, second, choosing a meaningful cutoff on the calculated statistic to declare positive compounds. Both false-positive and false-negative errors are inevitable in this process. Common control or estimation of error rates is often based on an assumption of normal distribution of the noise. The error rates in hit detection, especially false-negative rates, are hard to verify because in most assays, only compounds selected in primary screening are followed up in confirmation experiments. In this article, the authors take advantage of a quantitative HTS experiment in which all compounds are tested 42 times over a wide range of 14 concentrations so true positives can be found through a dose-response curve. Using the activity status defined by dose curve, the authors analyzed the effect of various data-processing procedures on the sensitivity and specificity of hit detection, the control of error rate, and hit confirmation. A new summary score is proposed and demonstrated to perform well in hit detection and useful in confirmation rate estimation. In general, adjusting for positional effects is beneficial, but a robust test can prevent overadjustment. Error rates estimated based on normal assumption do not agree with actual error rates, for the tails of noise distribution deviate from normal distribution. However, false discovery rate based on empirically estimated null distribution is very close to observed false discovery proportion.     

Target-biased scoring approaches and expert systems in structure-based virtual screening

Structure-based virtual screening followed by selection of a top fraction of the rank-ordered result list suffers from many false positives and false negatives because the general scoring functions are not accurate enough. Many approaches have emerged to address this problem by including knowledge about the specific target in the scoring and selection steps. This target bias can include requirements for critical interactions, use of pharmacophore patterns or interaction patterns found in known co-crystal structures, and similarity to known ligands. Such biases are implemented in methods that vary from filtering tools for pre- or post-processing, to expert systems, quantitative (re)scoring functions, and docking tools that generate target-biased poses.     

Progress toward virtual screening for drug side effects

The development and application of a computational protocol for conducting virtual screens of drug side interactions is described. A conventional drug-docking algorithm (AutoDock) is used to conduct two separate studies. First, a series of docking simulations is performed by using guanosine diphosphate and adenosine diphosphate as prototype drugs with the goal of successfully differentiating known receptors from a large number of bait receptors. Using the electrostatic energy of the purine ring as a basis for discrimination allows the correct identification of receptors in blind studies with 100% specificity and 94% sensitivity. In a second study, similar methodology is used to investigate the binding of clinically relevant inhibitors (Gleevec, purvalanol A, and hymenialdisine) to a variety of protein kinase targets. Overall, excellent agreement between experimental and predicted preferences for kinase targets is obtained. An important conclusion from the latter study is that homology-modeled structures of putative receptors may reasonably be used as targets for docking when directly solved crystal structures are not available. The prospects for routine application of the methodology as a means of identifying potential side interactions of candidate drugs are discussed.     

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.     

Where have all the interactions gone? Estimating the coverage of two-hybrid protein interaction maps

Yeast two-hybrid screens are an important method for mapping pairwise physical interactions between proteins. The fraction of interactions detected in independent screens can be very small, and an outstanding challenge is to determine the reason for the low overlap. Low overlap can arise from either a high false-discovery rate (interaction sets have low overlap because each set is contaminated by a large number of stochastic false-positive interactions) or a high false-negative rate (interaction sets have low overlap because each misses many true interactions). We extend capture-recapture theory to provide the first unified model for false-positive and false-negative rates for two-hybrid screens. Analysis of yeast, worm, and fly data indicates that 25% to 45% of the reported interactions are likely false positives. Membrane proteins have higher false-discovery rates on average, and signal transduction proteins have lower rates. The overall false-negative rate ranges from 75% for worm to 90% for fly, which arises from a roughly 50% false-negative rate due to statistical undersampling and a 55% to 85% false-negative rate due to proteins that appear to be systematically lost from the assays. Finally, statistical model selection conclusively rejects the Erd?s-Rényi network model in favor of the power law model for yeast and the truncated power law for worm and fly degree distributions. Much as genome sequencing coverage estimates were essential for planning the human genome sequencing project, the coverage estimates developed here will be valuable for guiding future proteomic screens. All software and datasets are available in and , -, and -, and are also available from our Web site, http://www.baderzone.org.     

Biochemical profiling in silico--predicting substrate specificities of large enzyme families

A general high-throughput method for in silico biochemical profiling of enzyme families has been developed based on covalent docking of potential substrates into the binding sites of target enzymes. The method has been tested by systematically docking transition state--analogous intermediates of 12 substrates into the binding sites of 20 alpha/beta hydrolases from 15 homologous families. To evaluate the effect of side chain orientations to the docking results, 137 crystal structures were included in the analysis. A good substrate must fulfil two criteria: it must bind in a productive geometry with four hydrogen bonds between the substrate and the catalytic histidine and the oxyanion hole, and a high affinity of the enzyme-substrate complex as predicted by a high docking score. The modelling results in general reproduce experimental data on substrate specificity and stereoselectivity: the differences in substrate specificity of cholinesterases toward acetyl- and butyrylcholine, the changes of activity of lipases and esterases upon the size of the acid moieties, activity of lipases and esterases toward tertiary alcohols, and the stereopreference of lipases and esterases toward chiral secondary alcohols. Rigidity of the docking procedure was the major reason for false positive and false negative predictions, as the geometry of the complex and docking score may sensitively depend on the orientation of individual side chains. Therefore, appropriate structures have to be identified. In silico biochemical profiling provides a time efficient and cost saving protocol for virtual screening to identify the potential substrates of the members of large enzyme family from a library of molecules.     

Investigating combinatorial approaches in virtual screening on human inducible 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (PFKFB3): a case study for small molecule kinases

Efforts toward improving the predictiveness in tier-based approaches to virtual screening (VS) have mainly focused on protein kinases. Despite their significance as drug targets, small molecule kinases have been rarely tested with these approaches. In this paper, we investigate the efficacy of a pharmacophore screening-combined structure-based docking approach on the human inducible 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase, an emerging target for cancer chemotherapy. Six out of a total 1364 compounds from NCI’s Diversity Set II were selected as true actives via throughput screening. Using a database constructed from these compounds, five programs were tested for structure-based docking (SBD) performance, the MOE of which showed the highest enrichments and second highest screening rates. Separately, using the same database, pharmacophore screening was performed, reducing 1364 compounds to 287 with no loss in true actives, yielding an enrichment of 4.75. When SBD was retested with the pharmacophore filtered database, 4 of the 5 SBD programs showed significant improvements to enrichment rates at only 2.5% of the database, with a 7-fold decrease in an average VS time. Our results altogether suggest that combinatorial approaches of VS technologies are easily applicable to small molecule kinases and, moreover, that such methods can decrease the variability associated with single-method SBD approaches.     

Uncovering false positives on a virtual screening search for cruzain inhibitors

Some unexpected promiscuous inhibitors were observed in a virtual screening protocol applied to select cruzain inhibitors from the ZINC database. Physical-chemical and pharmacophore model filters were used to reduce the database size. The selected compounds were docked into the cruzain active site. Six hit compounds were tested as inhibitors. Although the compounds were designed to be nucleophilically attacked by the catalytic cysteine of cruzain, three of them showed typical promiscuous behavior, revealing that false positives are a prevalent concern in VS programs.     

Structure Prediction and Validation of the ERK8 Kinase Domain

Extracellular signal-regulated kinase 8 (ERK8) has been already implicated in cell transformation and in the protection of genomic integrity and, therefore, proposed as a novel potential therapeutic target for cancer. In the absence of a crystal structure, we developed a three-dimensional model for its kinase domain. To validate our model we applied a structure-based virtual screening protocol consisting of pharmacophore screening and molecular docking. Experimental characterization of the hit compounds confirmed that a high percentage of the identified scaffolds was able to inhibit ERK8. We also confirmed an ATP competitive mechanism of action for the two best-performing molecules. Ultimately, we identified an ERK8 drug-resistant “gatekeeper” mutant that corroborated the predicted molecular binding mode, confirming the reliability of the generated structure. We expect that our model will be a valuable tool for the development of specific ERK8 kinase inhibitors.     

A detailed comparison of current docking and scoring methods on systems of pharmaceutical relevance

A thorough evaluation of some of the most advanced docking and scoring methods currently available is described, and guidelines for the choice of an appropriate protocol for docking and virtual screening are defined. The generation of a large and highly curated test set of pharmaceutically relevant protein-ligand complexes with known binding affinities is described, and three highly regarded docking programs (Glide, GOLD, and ICM) are evaluated on the same set with respect to their ability to reproduce crystallographic binding orientations. Glide correctly identified the crystallographic pose within 2.0 A in 61% of the cases, versus 48% for GOLD and 45% for ICM. In general Glide appears to perform most consistently with respect to diversity of binding sites and ligand flexibility, while the performance of ICM and GOLD is more binding site-dependent and it is significantly poorer when binding is predominantly driven by hydrophobic interactions. The results also show that energy minimization and reranking of the top N poses can be an effective means to overcome some of the limitations of a given docking function. The same docking programs are evaluated in conjunction with three different scoring functions for their ability to discriminate actives from inactives in virtual screening. The evaluation, performed on three different systems (HIV-1 protease, IMPDH, and p38 MAP kinase), confirms that the relative performance of different docking and scoring methods is to some extent binding site-dependent. GlideScore appears to be an effective scoring function for database screening, with consistent performance across several types of binding sites, while ChemScore appears to be most useful in sterically demanding sites since it is more forgiving of repulsive interactions. Energy minimization of docked poses can significantly improve the enrichments in systems with sterically demanding binding sites. Overall Glide appears to be a safe general choice for docking, while the choice of the best scoring tool remains to a larger extent system-dependent and should be evaluated on a case-by-case basis.     

Toward the virtual screening of Cdc25A phosphatase inhibitors with the homology modeled protein structure

Cdc25 phosphatases have been considered as attractive drug targets for anticancer therapy due to the correlation of their overexpression with a wide variety of cancers. As a method for the discovery of novel inhibitors of Cdc25 phosphatases, we have evaluated the computer-aided drug design protocol involving the homology modeling of Cdc25A and virtual screening with the two docking tools: FlexX and the modified AutoDock program implementing the effects of ligand solvation in the scoring function. The homology modeling with the X-ray crystal structure of Cdc25B as a template provides a high-quality structure of Cdc25A that enables the structure-based inhibitor design. Of the two docking programs under consideration, AutoDock is found to be more accurate than FlexX in terms of scoring putative ligands. A detailed binding mode analysis of the known inhibitors shows that they can be stabilized in the active site of Cdc25A through the simultaneous establishment of the multiple hydrogen bonds and the hydrophobic interactions. The present study demonstrates the usefulness of the modified AutoDock program as a docking tool for virtual screening of new Cdc25 phosphatase inhibitors as well as for binding mode analysis to elucidate the activities of known inhibitors. Figure Structures and available IC50 values (in μM) of the twenty Cdc25 phosphatase inhibitors seeded in docking library