On the value of homology models for virtual screening: discovering hCXCR3 antagonists by pharmacophore-based and structure-based approaches

Human chemokine receptor CXCR3 (hCXCR3) antagonists have potential therapeutic applications as antivirus, antitumor, and anti-inflammatory agents. A novel virtual screening protocol, which combines pharmacophore-based and structure-based approaches, was proposed. A three-dimensional QSAR pharmacophore model and a structure-based docking model were built to virtually screen for hCXCR3 antagonists. The hCXCR3 antagonist binding site was constructed by homology modeling and molecular dynamics (MD) simulation. By combining the structure-based and ligand-based screenings results, 95% of the compounds satisfied either pharmacophore or docking score criteria and would be chosen as hits if the union of the two searches was taken. The false negative rates were 15% for the pharmacophore model, 14% for the homology model, and 5% for the combined model. Therefore, the consistency of the pharmacophore model and the structural binding model is 219/273 = 80%. The hit rate for the virtual screening protocol is 273/286 = 95%. This work demonstrated that the quality of both the pharmacophore model and homology model can be measured by the consistency of the two models, and the false negatives in virtual screening can be reduced by combining two virtual screening approaches.     

A False‐Positive Screening Hit in Fragment‐Based Lead Discovery: Watch out for the Red Herring

With the rising popularity of fragment‐based approaches in drug development, more and more attention has to be devoted to the detection of false‐positive screening results. In particular, the small size and low affinity of fragments drives screening techniques to their limit. The pursuit of a false‐positive hit can cause significant loss of time and resources. Here, we present an instructive and intriguing investigation into the origin of misleading assay results for a fragment that emerged as the most potent binder for the aspartic protease endothiapepsin (EP) across multiple screening assays. This molecule shows its biological effect mainly after conversion into another entity through a reaction cascade that involves major rearrangements of its heterocyclic scaffold. The formed ligand binds EP through an induced‐fit mechanism involving remarkable electrostatic interactions. Structural information in the initial screening proved to be crucial for the identification of this false‐positive hit.     

Understanding false positives in reporter gene assays: in silico chemogenomics approaches to prioritize cell-based HTS data

High throughput screening (HTS) data is often noisy, containing both false positives and negatives. Thus, careful triaging and prioritization of the primary hit list can save time and money by identifying potential false positives before incurring the expense of followup. Of particular concern are cell-based reporter gene assays (RGAs) where the number of hits may be prohibitively high to be scrutinized manually for weeding out erroneous data. Based on statistical models built from chemical structures of 650 000 compounds tested in RGAs, we created "frequent hitter" models that make it possible to prioritize potential false positives. Furthermore, we followed up the frequent hitter evaluation with chemical structure based in silico target predictions to hypothesize a mechanism for the observed "off target" response. It was observed that the predicted cellular targets for the frequent hitters were known to be associated with undesirable effects such as cytotoxicity. More specifically, the most frequently predicted targets relate to apoptosis and cell differentiation, including kinases, topoisomerases, and protein phosphatases. The mechanism-based frequent hitter hypothesis was tested using 160 additional druglike compounds predicted by the model to be nonspecific actives in RGAs. This validation was successful (showing a 50% hit rate compared to a normal hit rate as low as 2%), and it demonstrates the power of computational models toward understanding complex relations between chemical structure and biological function.     

Biosensor-based small molecule fragment screening with biolayer interferometry

Biosensor-based fragment screening is a valuable tool in the drug discovery process. This method is advantageous over many biochemical methods because primary hits can be distinguished from non-specific or non-ideal interactions by examining binding profiles and responses, resulting in reduced false-positive rates. Biolayer interferometry (BLI), a technique that measures changes in an interference pattern generated from visible light reflected from an optical layer and a biolayer containing proteins of interest, is a relatively new method for monitoring small molecule interactions. The BLI format is based on a disposable sensor that is immersed in 96-well or 384-well plates. BLI has been validated for small molecule detection and fragment screening with model systems and well-characterized targets where affinity constants and binding profiles are generally similar to those obtained with surface plasmon resonsance (SPR). Screens with challenging targets involved in protein–protein interactions including BCL-2, JNK1, and eIF4E were performed with a fragment library of 6,500 compounds, and hit rates were compared for these targets. For eIF4E, a protein containing a PPI site and a nucleotide binding site, results from a BLI fragment screen were compared to results obtained in biochemical HTS screens. Overlapping hits were observed for the PPI site, and hits unique to the BLI screen were identified. Hit assessments with SPR and BLI are described.     

ALARM NMR: a rapid and robust experimental method to detect reactive false positives in biochemical screens

High-throughput screening (HTS) of large compound collections typically results in numerous small molecule hits that must be carefully evaluated to identify valid drug leads. Although several filtering mechanisms and other tools exist that can assist the chemist in this process, it is often the case that costly synthetic resources are expended in pursuing false positives. We report here a rapid and reliable NMR-based method for identifying reactive false positives including those that oxidize or alkylate a protein target. Importantly, the reactive species need not be the parent compound, as both reactive impurities and breakdown products can be detected. The assay is called ALARM NMR (a La assay to detect reactive molecules by nuclear magnetic resonance) and is based on monitoring DTT-dependent (13)C chemical shift changes of the human La antigen in the presence of a test compound or mixture. Extensive validation has been performed to demonstrate the reliability and utility of using ALARM NMR to assess thiol reactivity. This included comparing ALARM NMR to a glutathione-based fluorescence assay, as well as testing a collection of more than 3500 compounds containing HTS hits from 23 drug targets. The data show that current in silico filtering tools fail to identify more than half of the compounds that can act via reactive mechanisms. Significantly, we show how ALARM NMR data has been critical in identifying reactive compounds that would otherwise have been prioritized for lead optimization. In addition, a new filtering tool has been developed on the basis of the ALARM NMR data that can augment current in silico programs for identifying nuisance compounds and improving the process of hit triage.     

Evaluation of potential inhibitors of squalene synthase based on virtual screening and in vitro studies

Squalene synthase (SQS) is a potential target for hyperlipidemia treatment. To identify novel chemical scaffolds of SQS inhibitors, we generated 3D-QSAR pharmacophore models using HypoGen. The best quantitative pharmacophore model, Hypo 1, was selected for virtual screening using two chemical databases, Specs and Traditional Chinese Medicine database (TCM). The best-mapped hit compounds were then subjected to filtering by Lipinski^’s rule of five and docking studies to refine the hits. Finally, five compounds were selected from the top-ranked hit compounds for SQS inhibitory assay in vitro. Three of these compounds could inhibit SQS in vitro, and should be further evaluated pre-clinically as a treatment for hyperlipidemia.     

Comparison of luminescence ADP production assay and radiometric scintillation proximity assay for Cdc7 kinase

Several assay technologies have been successfully adapted and used in HTS to screen for protein kinase inhibitors; however, emerging comparative analysis studies report very low hit overlap between the different technologies, which challenges the working assumption that hit identification is not dependent on the assay method of choice. To help address this issue, we performed two screens on the cancer target, Cdc7-Dbf4 heterodimeric protein kinase, using a direct assay detection method measuring [(33)P]-phosphate incorporation into the substrate and an indirect method measuring residual ADP production using luminescence. We conducted the two screens under similar conditions, where in one, we measured [(33)P]-phosphate incorporation using scintillation proximity assay (SPA), and in the other, we detected luminescence signal of the ATP-dependent luciferase after regenerating ATP from residual ADP (LUM). Surprisingly, little or no correlation were observed between the positives identified by the two methods; at a threshold of 30% inhibition, 25 positives were identified in the LUM screen whereas the SPA screen only identified two positives, Tannic acid and Gentian violet, with Tannic acid being common to both. We tested 20 out of the 25 positive compounds in secondary confirmatory study and confirmed 12 compounds including Tannic acid as Cdc7-Dbf4 kinase inhibitors. Gentian violet, which was only positive in the SPA screen, inhibited luminescence detection and categorized as a false positive. This report demonstrates the strong impact in detection format on the success of a screening campaign and the importance of carefully designed confirmatory assays to eliminate those compounds that target the detection part of the assay.     

Selection of in silico drug screening results for G-protein-coupled receptors by using universal active probes

We developed a new protocol for in silico drug screening for G-protein-coupled receptors (GPCRs) using a set of "universal active probes" (UAPs) with an ensemble docking procedure. UAPs are drug-like compounds, which are actual active compounds of a variety of known proteins. The current targets were nine human GPCRs whose three-dimensional (3D) structures are unknown, plus three GPCRs, namely β(2)-adrenergic receptor (ADRB2), A(2A) adenosine receptor (A(2A)), and dopamine D3 receptor (D(3)), whose 3D structures are known. Homology-based models of the GPCRs were constructed based on the crystal structures with careful sequence inspection. After subsequent molecular dynamics (MD) simulation taking into account the explicit lipid membrane molecules with periodic boundary conditions, we obtained multiple model structures of the GPCRs. For each target structure, docking-screening calculations were carried out via the ensemble docking procedure, using both true active compounds of the target proteins and the UAPs with the multiple target screening (MTS) method. Consequently, the multiple model structures showed various screening results with both poor and high hit ratios, the latter of which could be identified as promising for use in in silico screening to find candidate compounds to interact with the proteins. We found that the hit ratio of true active compounds showed a positive correlation to that of the UAPs. Thus, we could retrieve appropriate target structures from the GPCR models by applying the UAPs, even if no active compound is known for the GPCRs. Namely, the screening result that showed a high hit ratio for the UAPs could be used to identify actual hit compounds for the target GPCRs.     

Enhancement of chemical rules for predicting compound reactivity towards protein thiol groups

Non-specific chemical modification of protein thiol groups continues to be a significant source of false positive hits from high-throughput screening campaigns and can even plague certain protein targets and chemical series well into lead optimization. While experimental tools exist to assess the risk and promiscuity associated with the chemical reactivity of existing compounds, computational tools are desired that can reliably identify substructures that are associated with chemical reactivity to aid in triage of HTS hit lists, external compound purchases, and library design. Here we describe a Bayesian classification model derived from more than 8,800 compounds that have been experimentally assessed for their potential to covalently modify protein targets. The resulting model can be implemented in the large-scale assessment of compound libraries for purchase or design. In addition, the individual substructures identified as highly reactive in the model can be used as look-up tables to guide chemists during hit-to-lead and lead optimization campaigns.     

Discovery of novel Pim-1 kinase inhibitors by a hierarchical multistage virtual screening approach based on SVM model, pharmacophore, and molecular docking

In this investigation, we describe the discovery of novel potent Pim-1 inhibitors by employing a proposed hierarchical multistage virtual screening (VS) approach, which is based on support vector machine-based (SVM-based VS or SB-VS), pharmacophore-based VS (PB-VS), and docking-based VS (DB-VS) methods. In this approach, the three VS methods are applied in an increasing order of complexity so that the first filter (SB-VS) is fast and simple, while successive ones (PB-VS and DB-VS) are more time-consuming but are applied only to a small subset of the entire database. Evaluation of this approach indicates that it can be used to screen a large chemical library rapidly with a high hit rate and a high enrichment factor. This approach was then applied to screen several large chemical libraries, including PubChem, Specs, and Enamine as well as an in-house database. From the final hits, 47 compounds were selected for further in vitro Pim-1 inhibitory assay, and 15 compounds show nanomolar level or low micromolar inhibition potency against Pim-1. In particular, four of them were found to have new scaffolds which have potential for the chemical development of Pim-1 inhibitors.     

NIR-mbc94, a fluorescent ligand that binds to endogenous CB(2) receptors and is amenable to high-throughput screening

High-throughput screening (HTS) of chemical libraries is often used for the unbiased identification of compounds interacting with G protein-coupled receptors (GPCRs), the largest family of therapeutic targets. However, current HTS methods require removing GPCRs from their native environment, which modifies their pharmacodynamic properties and biases the screen toward false positive hits. Here, we developed and validated a molecular imaging (MI) agent, NIR-mbc94, which emits near infrared (NIR) light and selectively binds to endogenously expressed cannabinoid CB(2) receptors,?a recognized target for treating autoimmune diseases, chronic pain and cancer. The precision and ease of this assay allows for the HTS of compounds interacting with CB(2) receptors expressed in their native environment.     

Assessing the lipophilicity of fragments and early hits

A key challenge in many drug discovery programs is to accurately assess the potential value of screening hits. This is particularly true in fragment-based drug design (FBDD), where the hits often bind relatively weakly, but are correspondingly small. Ligand efficiency (LE) considers both the potency and the size of the molecule, and enables us to estimate whether or not an initial hit is likely to be optimisable to a potent, druglike lead. While size is a key property that needs to be controlled in a small molecule drug, there are a number of additional properties that should also be considered. Lipophilicity is amongst the most important of these additional properties, and here we present a new efficiency index (LLE_AT) that combines lipophilicity, size and potency. The index is intuitively defined, and has been designed to have the same target value and dynamic range as LE, making it easily interpretable by medicinal chemists. Monitoring both LE and LLE_AT should help both in the selection of more promising fragment hits, and controlling molecular weight and lipophilicity during optimisation.     

Pharmacophore Modeling and Virtual Screening of Novel Inhibitors for c-Kit Kinase

The stem cell factor receptor (c‐Kit) has been known to play critical roles in regulating numerous aspects of cellular behavior including cell growth, differentiation, migration and metabolism. In this investigation, a three‐dimensional pharmacophore model of c‐Kit inhibitors has been established by using the HypoGen algorithms implemented in the catalyst software package. The best quantitative pharmacophore model, hypothesis 1, which has the highest correlation coefficient (0.989), consists of one hydrogen bond acceptor, two hydrogen bond donors and one hydrophobic feature. To our knowledge, this is the first report on the pharmacophore modeling study of c‐Kit inhibitors. The best hypothesis, hypothesis 1, was used to screen molecular structural databases, including Specs and China Natural Products Database for potential lead compounds. The hit compounds were subsequently subjected to filtering by Lipinski's rules and docking study to refine the retrieved hits and as a result to reduce the rate of false positive. Finally 28 compounds were purchased or synthesized for further in vitro assay against several human tumour cell lines including A549, MCF‐7, HepG2 and PC‐3, in which c‐Kit is overexpressed. Two compounds show very low micromolar inhibition potency against the PC‐3 and HepG2 cell lines respectively. And they were selected for further modification and testing.     

Structure-based screening as applied to human FABP4: a highly efficient alternative to HTS for hit generation

The time-limiting step in HTS often is the development of an appropriate assay. In addition, hits from HTS fairly often turn out to be false positives and generally display unfavorable properties for further development. Here we describe an alternative process for hit generation, applied to the human adipocyte fatty acid binding protein FABP4. A small molecular ligand for FABP4 that blocks the binding of endogenous ligands may be developed into a drug for the treatment of type-2 diabetes. Using NMR spectroscopy, we screened FABP4 for low-affinity binders in a diversity library consisting of small soluble scaffolds, which yielded 52 initial hits in total. The potencies of these hits were ranked, and crystal structures of FABP4 complexes for two of the hits were obtained. The structural data were subsequently used to direct similarity searches for available analogues, as well as chemical synthesis of 12 novel analogues. In this way, a series of three selective FABP4 ligands with attractive pharmacochemical profiles and potencies of 10 microM or better was obtained.     

Biosensor analysis of beta-lactams in milk: comparison with microbiological, immunological, and receptor-based screening methods

Two recently developed surface plasmon resonance biosensor assays for detection of beta-lactams in milk were used to screen raw producer milk samples. Both assays use a beta-lactam receptor protein with carboxypeptidase activity for detection. The results of the biosensor assays were compared with those of various commercial screening tests, i.e., the Delvotest SP, Penzym S, Beta-STAR, SNAP, and Parallux. The results of the 2 biosensor assays showed good agreement with those of the other screening tests. Of 195 analyzed milk samples, the results of only 5 samples differed between the assays. Additionally, 30 milk samples with both negative and positive results in the screening assays were analyzed by liquid chromatography for identification and quantification of any beta-lactam residues. All screening tests showed 0% false-negative results with 15 incurred samples containing between 4.0 and 268 microg/kg penicillin G. The biosensor assays showed 27% positive results (false violatives) with 15 producer milk samples containing penicillin G concentrations between 0 and 3.6 microg/kg, i.e., below maximum residue limit. This figure varied between 27 and 53% for the other screening tests.     

Novel statistical approach for primary high-throughput screening hit selection

The standard activity threshold-based method (the "top X" approach), currently widely used in the high-throughput screening (HTS) data analysis, is ineffective at identifying good-quality hits. We have proposed a novel knowledge-based statistical approach, driven by the hidden structure-activity relationship (SAR) within a screening library, for primary hit selection. Application to an in-house ultrahigh-throughput screening (uHTS) campaign has demonstrated it can directly identify active scaffolds containing valuable SAR information with a greatly improved confirmation rate compared to the standard "top X" method (from 55% to 85%). This approach may help produce high-quality leads and expedite the hit-to-lead process in drug discovery.     

How large does a compound screening collection need to be?

Increasingly, chemical libraries are being produced which are focused on a biological target or group of related targets, rather than simply being constructed in a combinatorial fashion. A screening collection compiled from such libraries will contain multiple analogues of a number of discrete series of compounds. The question arises as to how many analogues are necessary to represent each series in order to ensure that an active series will be identified. Based on a simple probabilistic argument and supported by in-house screening data, guidelines are given for the number of compounds necessary to achieve a "hit", or series of hits, at various levels of certainty. Obtaining more than one hit from the same series is useful since this gives early acquisition of SAR (structure-activity relationship) and confirms a hit is not a singleton. We show that screening collections composed of only small numbers of analogues of each series are sub-optimal for SAR acquisition. Based on these studies, we recommend a minimum series size of about 200 compounds. This gives a high probability of confirmatory SAR (i.e. at least two hits from the same series). More substantial early SAR (at least 5 hits from the same series) can be gained by using series of about 650 compounds each. With this level of information being generated, more accurate assessment of the likely success of the series in hit-to-lead and later stage development becomes possible.     

Congestion game scheduling for virtual drug screening optimization

In virtual drug screening, the chemical diversity of hits is an important factor, along with their predicted activity. Moreover, interim results are of interest for directing the further research, and their diversity is also desirable. In this paper, we consider a problem of obtaining a diverse set of virtual screening hits in a short time. To this end, we propose a mathematical model of task scheduling for virtual drug screening in high-performance computational systems as a congestion game between computational nodes to find the equilibrium solutions for best balancing the number of interim hits with their chemical diversity. The model considers the heterogeneous environment with workload uncertainty, processing time uncertainty, and limited knowledge about the input dataset structure. We perform computational experiments and evaluate the performance of the developed approach considering organic molecules database GDB-9. The used set of molecules is rich enough to demonstrate the feasibility and practicability of proposed solutions. We compare the algorithm with two known heuristics used in practice and observe that game-based scheduling outperforms them by the hit discovery rate and chemical diversity at earlier steps. Based on these results, we use a social utility metric for assessing the efficiency of our equilibrium solutions and show that they reach greatest values.     

On‐Bead Screens Sample Narrower Affinity Ranges of Protein–Ligand Interactions Compared to Equivalent Solution Assays

Conceptually, on‐bead screening is one of the most efficient high‐throughput screening (HTS) methods. One of its inherent advantages is that the solid support has a dual function: it serves as a synthesis platform and as a screening compartment. Compound purification, cleavage and storage and extensive liquid handling are not necessary in bead‐based HTS. Since the establishment of one‐bead one‐compound library synthesis, the properties of polymer beads in chemical reactions have been thoroughly investigated. However, the characterization of the kinetics and thermodynamics of protein–ligand interactions on the beads used for screening has received much less attention. Consequently, the majority of reported on‐bead screens are based on empirically derived procedures, independent of measured equilibrium constants and rate constants of protein binding to ligands on beads. More often than not, on‐bead screens reveal apparent high affinity binders through strong protein complexation on the matrix of the solid support. After decoding, resynthesis, and solution testing the primary hits turn out to be unexpectedly weak binders, or may even fall out of the detection limit of the solution assay. Only a quantitative comparison of on‐bead binding and solution binding events will allow systematically investigating affinity differences as function of protein and small molecule properties. This will open up routes for optimized bead materials, blocking conditions and other improved assay procedures. By making use of the unique features of our previously introduced confocal nanoscanning (CONA) method, we investigated the kinetic and thermodynamic properties of protein–ligand interactions on TentaGel beads, a popular solid support for on‐bead screening. The data obtained from these experiments allowed us to determine dissociation constants for the interaction of bead‐immobilized ligands with soluble proteins. Our results therefore provide, for the first time, a comparison of on‐bead versus solution binding thermodynamics. Our data indicate that affinity ranges found in on‐bead screening are indeed narrower compared to equivalent interactions in homogeneous solution. A thorough physico‐chemical understanding of the molecular recognition between proteins and surface bound ligands will further strengthen the role of on‐bead screening as an ultimately cost‐effective method in hit and lead finding.     

Identification of small molecule aggregators from large compound libraries by support vector machines

Small molecule aggregators non‐specifically inhibit multiple unrelated proteins, rendering them therapeutically useless. They frequently appear as false hits and thus need to be eliminated in high‐throughput screening campaigns. Computational methods have been explored for identifying aggregators, which have not been tested in screening large compound libraries. We used 1319 aggregators and 128,325 non‐aggregators to develop a support vector machines (SVM) aggregator identification model, which was tested by four methods. The first is five fold cross‐validation, which showed comparable aggregator and significantly improved non‐aggregator identification rates against earlier studies. The second is the independent test of 17 aggregators discovered independently from the training aggregators, 71% of which were correctly identified. The third is retrospective screening of 13M PUBCHEM and 168K MDDR compounds, which predicted 97.9% and 98.7% of the PUBCHEM and MDDR compounds as non‐aggregators. The fourth is retrospective screening of 5527 MDDR compounds similar to the known aggregators, 1.14% of which were predicted as aggregators. SVM showed slightly better overall performance against two other machine learning methods based on five fold cross‐validation studies of the same settings. Molecular features of aggregation, extracted by a feature selection method, are consistent with published profiles. SVM showed substantial capability in identifying aggregators from large libraries at low false‐hit rates. © 2009 Wiley Periodicals, Inc.J Comput Chem, 2010