Detailed analysis of scoring functions for virtual screening

In silico identification of cyclosporin derivatives as potential inhibitors for RdRp of rotavirus by molecular docking and molecular dynamic studies

Rotavirus is one of the most common gastrointestinal viral diseases. Till date, there are only two vaccines available in the markets, which are specifically to be administered to young babies. In this study, VP1 RdRp is selected as potential target to carry out inhibitory activities. Cyclosporin A (Cys A) derivatives were designed via FBDD, pharmacokinetics, molecular docking, molecular dynamics (MD) simulation and molecular mechanics generalized born surface area was applied on these compounds. The results from these investigations were analyzed and it was found that the considered derivatives in this study were nontoxic and docking results revealed that the derivatives made some important bonds inside the active site of the receptors within a catalytic triad (Serine-Histidine-Aspartate). After analyzing the mean values of root mean square density (RMSD), root mean square fluctuation (RMSF), radius of gyration (RoG) and solvent accessible surface area (SASA) at 100 ns MD simulation of the selected compounds, it was found that compound 1 exhibits RMSD of 0.74 ± 0.10 Å, RMSF of 0.85 ± 0.15 Å, RoG of 16.45 ± 0.40 Å, SASA of 66.55 ± 0.35 nm2 and ΔGbind of -32.76 ± 0.02 kcal/mol. Therefore, the study revealed that amongst the designed and reported compounds, compound 1 was more stable within the active region of the RdRp and also this compound possesses lower binding free energy as compared to other selected compounds and Cys A as well.Communicated by Ramaswamy H. Sarma.

Novel PLGA-encapsulated-nanopiperine promotes synergistic interaction of p53/PARP-1/Hsp90 axis to combat ALX-induced-hyperglycemia

The present study predicts the molecular targets and druglike properties of the phyto-compound piperine (PIP) by in silico studies including molecular docking simulation, druglikeness prediction and ADME analysis for prospective therapeutic benefits against diabetic complications. PIP was encapsulated in biodegradable polymer poly-lactide-co-glycolide (PLGA) to form nanopiperine (NPIP) and their physico-chemical properties were characterized by AFM and DLS. ∼ 30 nm sized NPIP showed 86.68% encapsulation efficiency and - 6 mV zeta potential, demonstrated great interactive stability and binding with CT-DNA displaying upsurge in molar ellipticity during CD spectroscopy. NPIP lowered glucose levels in peripheral circulation by > 65 mg/dL compared to disease model and improved glucose influx in alloxan-induced in vivo and in vitro diabetes models concerted with 3-folds decrease in ROS production, ROS-induced DNA damage and 27.24% decrease in nuclear condensation. The 25% increase in % cell viability and inhibition in chromosome aberration justified the initiation of p53 and PARP DNA repairing protein expression and maintenance of Hsp90. Thus, the experimental study corroborated well with in silico predictions of modulating the p53/PARP-1/Hsp90 axis, with predicted dock score value of - 8.72, - 8.57, - 8.76 kcal/mol respectively, validated docking-based preventive approaches for unravelling the intricacies of molecular signalling and nano-drug efficacy as therapeutics for diabetics.

Prioritizing Small Sets of Molecules for Synthesis through in-silico Tools: A Comparison of Common Ranking Methods

Prioritizing molecules for synthesis is a key role of computational methods within medicinal chemistry. Multiple tools exist for ranking molecules, from the cheap and popular molecular docking methods to more computationally expensive molecular-dynamics (MD)-based methods. It is often questioned whether the accuracy of the more rigorous methods justifies the higher computational cost and associated calculation time. Here, we compared the performance on ranking the binding of small molecules for seven scoring functions from five docking programs, one end-point method (MM/GBSA), and two MD-based free energy methods (PMX, FEP+). We investigated 16 pharmaceutically relevant targets with a total of 423 known binders. The performance of docking methods for ligand ranking was strongly system dependent. We observed that MD-based methods predominantly outperformed docking algorithms and MM/GBSA calculations. Based on our results, we recommend the application of MD-based free energy methods for prioritization of molecules for synthesis in lead optimization, whenever feasible.

Exploring the inhibitory mechanisms of indazole compounds against SAH/MTAN-mediated quorum sensing utilizing QSAR and docking

The world is under the great threat of antimicrobial resistance (AMR) leading to premature deaths. Microorganisms can produce AMR via quorum sensing mechanisms utilizing S-adenosyl homocysteine/methylthioadenosine nucleosidase (SAH/MTAN) biosynthesis. But there is no specific drug developed to date to stop SAH/MTAN, which is a crucial target for the discovery of anti-quorum sensing compound. It has been shown that indazole compounds cause inhibition of SAH/MTAN-mediated quorum sensing, but the biochemical mechanisms have not yet been explored. Therefore, in this original research, an attempt has been made to explore essential structural features of these compounds by quantitative structure-activity relationship (QSAR) and molecular docking of indazole compounds having inhibition of SAH/MTAN-mediated quorum sensing. The validated QSAR predicted five essential descriptors and molecular docking helps to identify the active binding amino acid residues involved in ligand-receptor interactions that are responsible for producing the quorum sensing inhibitory mechanisms of indazole compounds against SAH/MTAN-mediated AMR.

Assessment of the Generalization Abilities of Machine-Learning Scoring Functions for Structure-Based Virtual Screening

In structure-based virtual screening (SBVS), it is critical that scoring functions capture protein-ligand atomic interactions. By focusing on the local domains of ligand binding pockets, a standardized pocket Pfam-based clustering (Pfam-cluster) approach was developed to assess the cross-target generalization ability of machine-learning scoring functions (MLSFs). Subsequently, 12 typical MLSFs were evaluated using random cross-validation (Random-CV), protein sequence similarity-based cross-validation (Seq-CV), and pocket Pfam-based cross-validation (Pfam-CV) methods. Surprisingly, all of the tested models showed decreased performances from Random-CV to Seq-CV to Pfam-CV experiments, not showing satisfactory generalization capacity. Our interpretable analysis suggested that the predictions on novel targets by MLSFs were dependent on buried solvent-accessible surface area (SASA)-related features of complex structures, with greater predicted binding affinities on complexes owning larger protein-ligand interfaces. By combining buried SASA-related features with target-specific patterns that were only shared among structurally similar compounds in the same cluster, the random forest (RF)-Score attained a good performance in the Random-CV test. Based on these findings, we strongly advise assessing the generalization ability of MLSFs with the Pfam-cluster approach and being cautious with the features learned by MLSFs.

Consensus scoring evaluated using the GPCR-Bench dataset: Reconsidering the role of MM/GBSA

The recent availability of large numbers of GPCR crystal structures has provided an unprecedented opportunity to evaluate their performance in virtual screening protocols using established benchmarking datasets. In this study, we evaluated the ability of MM/GBSA in consensus scoring-based virtual screening enrichment together with nine classical scoring functions, using the GPCR-Bench dataset consisting of 24 GPCR crystal structures and 254,646 actives and decoys. While the performance of consensus scoring was modest overall, combinations which included MM/GBSA performed relatively well compared to combinations of classical scoring functions. Combinations of MM/GBSA and good-performing scoring functions provided the highest proportion of improvements, with improvements observed in 32% and 19% of all combinations across all targets at the EF_1% and EF_5% levels respectively. Combinations of MM/GBSA and poor-performing scoring functions still outperformed classical scoring functions, with improvements observed in 26% and 17% of all combinations at the EF_1% and EF_5% levels. In comparison, only 14-22% and 6-11% of combinations of classical scoring functions produced improvements at EF_1% and EF_5% respectively. Efforts to improve performance by increasing the number of scoring functions in consensus scoring to three were mostly ineffective. We also observed that consensus scoring performed better for individual scoring functions possessing initially low enrichment factors, potentially implying their benefits are more relevant in such scenarios. Overall, this study demonstrated the first implementation of MM/GBSA in consensus scoring using the GPCR-Bench dataset and could provide a valuable benchmark of the performance of MM/GBSA in comparison to classical scoring functions in consensus scoring for GPCRs.

Applications of machine learning in computer-aided drug discovery

Machine learning (ML) has revolutionised the field of structure-based drug design (SBDD) in recent years. During the training stage, ML techniques typically analyse large amounts of experimentally determined data to create predictive models in order to inform the drug discovery process. Deep learning (DL) is a subfield of ML, that relies on multiple layers of a neural network to extract significantly more complex patterns from experimental data, and has recently become a popular choice in SBDD. This review provides a thorough summary of the recent DL trends in SBDD with a particular focus on de novo drug design, binding site prediction, and binding affinity prediction of small molecules.

An antigen display system of GEM nanoparticles based on affinity peptide ligands

Gram-positive enhancer matrix (GEM) nanoparticles are often used in mucosal immunity, preparation of subunit vaccines or as an immune adjuvant due to its good immunological activities in recent years. Here, we designed and screened out a high affinity peptide ligand PL23, which could specifically target the non-epitope region of Classic Swine Fever Virus (CSFV) E2 protein, by virtual screening technology, enzyme linked immunosorbent assay (ELISA) and surface plasmon resonance (SPR) test. The OD value of PL23 at 450 nm was reached 1.982, and the K_D value of it was 90.12 nM. Its binding capacity to protein was verified by SDS-PAGE as well. PL23 was subsequently conjugated to GEM nanoparticles by dehydration synthesis generating GEM-PL23 particles, and the GEM-PL-E2 particles were assembled after incubated with CSFV E2 protein. The cytotoxic test indicated that PL23, CSFV E2 protein, GEM nanoparticles, GEM-PL23 particles and GEM-PL-E2 particles were not toxic to cells and GEM nanoparticles could significantly promote the growth of APCs at high concentration for 1 h, p<0.001. In addition, GEM nanoparticles could promote the uptake of antigen by APCs. The cytokines tests suggested that GEM-PL-E2 particles could promote innate immune responses, regulate adaptive immune responses generated by T cells and APCs, and promote the differentiation and maturation of dendritic cells without producing inflammasomes. The results of immunological activity identification showed GEM-PL-E2 particles induced higher levels of both neutralizing antibodies and anti-CSFV antibodies than CSFV E2 protein in mice. This strategy provided a new, simpler, faster and cheaper method for assembling GEM nanoparticles, using an affinity peptide ligand replaced the protein anchor (PA), and provided a better application prospect for the application of GEM particles.

Identification of novel P2X7R antagonists by using structure-based virtual screening and cell-based assays

In the tumor microenvironment, inflammation and necrosis cause the accumulations of ATP extracellularly, and high concentrations of ATP can activate P2X7 receptors (P2X7R), which leads to the influx of Na⁺ , K⁺ , or Ca²⁺ into cells and trigger the downstream signaling pathways. P2X7R is a relatively unique ligand-gated ion channel, which is over-expressed in most tumor cells. The activated P2X7R facilitates the tumor growth, invasion, and metastasis. Inhibition of the P2X7R activation can be applied as a potential anti-tumor therapy strategy. There are currently no anti-tumor agents against P2X7R, though several P2X7R antagonists for indications such as anti-inflammatory and anti-depression were reported. In this study, we combined homology modeling (HM), virtual screening, and EB intake assay to characterize the structural features of P2X7R and identify several novel antagonists, which were chemically different from any other known P2X7R antagonists. The identified antagonists could effectively prevent the pore opening of P2X7R with IC50 values ranging from 29.14 to 35.34 μM. HM model showed the area between ATP-binding pocket, and allosteric sides were hydrophobic and suitable for small molecule interaction. Molecular docking indicated a universal binding mode, of which residues R294 and K311 were used as hydrogen bond donors to participate in antagonist interactions. The binding mode can potentially be utilized for inhibitor optimization for increased affinity, and the identified antagonists can be further tested for anti-cancer activity or may serve as chemical agents to study P2X7R related functions.

In silico data mining of large-scale databases for the virtual screening of human interleukin-2 inhibitors

Interleukin-2 (IL-2) is involved in the activation and differentiation of T-helper cells. Uncontrolled activated T cells play a key role in the pathophysiology by stimulating inflammation and autoimmune diseases like arthritis, psoriasis and Crohn's disease. T cells activation can be suppressed either by preventing IL-2 production or blocking the IL-2 interaction with its receptor. Hence, IL-2 is now emerging as a target for novel therapeutic approaches in several autoimmune disorders. This study was carried out to set up an effective virtual screening (VS) pipeline for IL-2. Four docking/scoring approaches (FRED, MOE, GOLD and Surflex-Dock) were compared in the re-docking process to test their performance in producing correct binding modes of IL-2 inhibitors. Surflex-Dock and FRED were the best in predicting the native pose in its top-ranking position. Shapegauss and CGO scoring functions identified the known inhibitors of IL-2 in top 1, 5 and 10 % of library and differentiated binders from non-binders efficiently with average AUC of > 0.9 and > 0.7, resp. The applied docking protocol served as a basis for the VS of a large database that will lead to the identification of more active compounds against IL-2.

Application and assessment of deep learning for the generation of potential NMDA receptor antagonists

Uncompetitive antagonists of the N-methyl d-aspartate receptor (NMDAR) have demonstrated therapeutic benefit in the treatment of neurological diseases such as Parkinson's and Alzheimer's, but some also cause dissociative effects that have led to the synthesis of illicit drugs. The ability to generate NMDAR antagonists in silico is therefore desirable for both new medication development and preempting and identifying new designer drugs. Recently, generative deep learning models have been applied to de novo drug design as a means to expand the amount of chemical space that can be explored for potential drug-like compounds. In this study, we assess the application of a generative model to the NMDAR to achieve two primary objectives: (i) the creation and release of a comprehensive library of experimentally validated NMDAR phencyclidine (PCP) site antagonists to assist the drug discovery community and (ii) an analysis of both the advantages conferred by applying such generative artificial intelligence models to drug design and the current limitations of the approach. We apply, and provide source code for, a variety of ligand- and structure-based assessment techniques used in standard drug discovery analyses to the deep learning-generated compounds. We present twelve candidate antagonists that are not available in existing chemical databases to provide an example of what this type of workflow can achieve, though synthesis and experimental validation of these compounds are still required.

Discovery of potent α-glucosidase inhibitors through structure-based virtual screening of an in-house azole collection

Diabetes mellitus, a chronic disorder characterized by hyperglycemia, is considered a pandemic of modern times. α-Glucosidase inhibitors emerged as a promising class of antidiabetic drugs with better tolerability compared with its alternatives. Azoles, although widely preferred in drug design, have scarcely been investigated for their potential against α-glucosidase. In this study, we evaluated α-glucosidase inhibitory effects 20 azole derivatives selected out of an in-house collection via structure-based virtual screening (VS) with consensus scoring approach. Seven compounds were identified with better IC₅₀ values than acarbose (IC₅₀  = 68.18 ± 1.01 µM), a well-known α-glucosidase inhibitor drug, which meant 35% success for our VS methodology. Compound 52, 54, 56, 59, and 81 proved highly potent with IC₅₀ values in the range of 40-60 µM. According to the enzyme kinetics study, four of them were competitive, 56 was non-competitive inhibitor. Structure-activity relationships, quantum mechanical, and docking analyses showed that azole rings at ionized state may be key to the potency observed for the active compounds and modifications to shift the balance between the neutral and ionized states further to the latter could yield more potent derivatives.

Rational drug repurposing for cancer by inclusion of the unbiased molecular dynamics simulation in the structure-based virtual screening approach: Challenges and breakthroughs

Managing cancer is now one of the biggest concerns of health organizations. Many strategies have been developed in drug discovery pipelines to help rectify this problem and two of the best ones are drug repurposing and computational methods. The combination of these approaches can have immense impact on the course of drug discovery. In silico drug repurposing can significantly reduce the time, the cost and the effort of drug development. Computational methods such as structure-based drug design (SBDD) and virtual screening can predict the potentials of small molecule binders, such as drugs, for having favorable effect on a particular molecular target. However, the demand for accuracy and efficiency of SBDD requires more sophisticated and complicated approaches such as unbiased molecular dynamics (UMD) simulation that has been recently introduced. As a complementary strategy, the knowledge acquired from UMD simulations can increase the chance of finding the right candidates and the pipeline of its administration is introduced and discussed in this review. An elaboration of this pipeline is also made by detailing an example, the binding and unbinding pathways of dasatinib-c-Src kinase complex, which shows that how influential this method can be in rational drug repurposing in cancer treatment.

Predicting or Pretending: Artificial Intelligence for Protein-Ligand Interactions Lack of Sufficiently Large and Unbiased Datasets

Predicting protein-ligand interactions using artificial intelligence (AI) models has attracted great interest in recent years. However, data-driven AI models unequivocally suffer from a lack of sufficiently large and unbiased datasets. Here, we systematically investigated the data biases on the PDBbind and DUD-E datasets. We examined the model performance of atomic convolutional neural network (ACNN) on the PDBbind core set and achieved a Pearson R² of 0.73 between experimental and predicted binding affinities. Strikingly, the ACNN models did not require learning the essential protein-ligand interactions in complex structures and achieved similar performance even on datasets containing only ligand structures or only protein structures, while data splitting based on similarity clustering (protein sequence or ligand scaffold) significantly reduced the model performance. We also identified the property and topology biases in the DUD-E dataset which led to the artificially increased enrichment performance of virtual screening. The property bias in DUD-E was reduced by enforcing the more stringent ligand property matching rules, while the topology bias still exists due to the use of molecular fingerprint similarity as a decoy selection criterion. Therefore, we believe that sufficiently large and unbiased datasets are desirable for training robust AI models to accurately predict protein-ligand interactions.

Effective estimation of the inhibitor affinity of HIV-1 protease via a modified LIE approach

The inhibition of the Human Immunodeficiency Virus Type 1 Protease (HIV-1 PR) can prevent the synthesis of new viruses. Computer-aided drug design (CADD) would enhance the discovery of new therapies, through which the estimation of ligand-binding affinity is critical to predict the most efficient inhibitor. A time-consuming binding free energy method would reduce the usefulness of CADD. The modified linear interaction energy (LIE) approach emerges as an appropriate protocol that performs this task. In particular, the polar interaction free energy, which is obtained via numerically resolving the linear Poisson–Boltzmann equation, plays as an important role in driving the binding mechanism of the HIV-1 PR + inhibitor complex. The electrostatic interaction energy contributes to the attraction between two molecules, but the vdW interaction acts as a repulsive factor between the ligand and the HIV-1 PR. Moreover, the ligands were found to adopt a very strong hydrophobic interaction with the HIV-1 PR. Furthermore, the results obtained corroborate the high accuracy and precision of computational studies with a large correlation coefficient value R = 0.83 and a small RMSE δ_RMSE = 1.25 kcal mol⁻¹. This method is less time-consuming than the other end-point methods, such as the molecular mechanics Poisson–Boltzmann surface area (MM/PBSA) and free energy perturbation (FEP) approaches. Overall, the modified LIE approach would provide ligand-binding affinity with HIV-1 PR accurately, precisely, and rapidly, resulting in a more efficient design of new inhibitors.

The inhibition of the Human Immunodeficiency Virus Type 1 Protease (HIV-1 PR) can prevent the synthesis of new viruses.

Virtual screening of novel compounds as potential ER-alpha inhibitors

Majority of breast cancers diagnosed today are estrogen receptor (ER)-positive, however, progesterone receptor-positive (PR-positive) is also responsible for breast cancer. Tumors that are ER/PR-positive are much more likely to respond to hormone therapy than tumors that are ER/PR-negative. Nearly 105 ERa inhibitors from literature when docked resulted in 31 compounds (pyrazolo[1,5-a]pyrimidine analogs and chromen-2-one derivatives) with better binding affinities. The maximum score obtained was -175.282 kcal/mol for compound, [2-(4- Fluoro-phenylamino)-pyridin-3-yl]-{4-[2-phenyl-7- (3, 4, 5-trimethoxy-phenyl)-pyrazolo[1,5-a]pyrimidine-5-carbonyl]-piperazin-1-yl}-methanone. The major H-bond interactions are observed with Thr347. In pursuit to identify novel ERa inhibitory ligands, virtual screening was carried out by docking pyrazole, bipyrazole, thiazole, thiadiazole etc scaffold analogs from literature.34 bipyrazoles from literature revealed Compound 2, ethyl 5-amino-1-(5-amino-3-anilino-4-ethoxycarbonyl-pyrazol-1-yl)-3-anilino-pyrazole-4-carboxylate, with -175.9 kcal/mol binding affinity with the receptor, where a favourable H-bond was formed with Thr347.On the other hand, screening 2035 FDA approved drugs from Drug Bank database resulted in 11 drugs which showed better binding affinities than ERa bound tamoxifen. Consensus scoring using 5 scoring schemes such as Mol Dock score, mcule, SwissDock, Pose&Rank and DSX respectively resulted in better rank-sumsfor Lomitapide, Itraconazole, Cobicistat, Azilsartanmedoxomil, and Zafirlukast.

Adequate prediction for inhibitor affinity of Aβ40 protofibril using the linear interaction energy method

The search for efficient inhibitors targeting Aβ oligomers and fibrils is an important issue in Alzheimer's disease treatment. As a consequence, an accurate and computationally cheap approach to estimate the binding affinity for many ligands interacting with Aβ peptides is very important. Here, the calculated binding free energies of 30 ligands interacting with 12Aβ11-40 peptides using the linear interaction energy (LIE) approach are found to be in good correlation with experimental data (R = 0.79). The binding affinities of these complexes are also calculated by using free energy perturbation (FEP) and molecular mechanic/Poisson-Boltzmann surface area (MM/PBSA) methods. The time-consuming FEP method provides results with similar correlation (R = 0.72), whereas MM/PBSA calculations show very low correlation with experimental data (R = 0.27). In all complexes, van der Waals interactions contribute much more than electrostatic interactions. The LIE model, which is much less time-consuming than both the FEP and MM/PBSA methods, opens the door to accurate and rapid affinity prediction of ligands with Aβ peptides and the design of new ligands.

In silico Prediction, Characterization, Molecular Docking, and Dynamic Studies on Fungal SDRs as Novel Targets for Searching Potential Fungicides Against Fusarium Wilt in Tomato

Vascular wilt of tomato caused by Fusarium oxysporum f.sp. lycopersici (FOL) is one of the most devastating diseases, that delimits the tomato production worldwide. Fungal short-chain dehydrogenases/reductases (SDRs) are NADP(H) dependent oxidoreductases, having shared motifs and common functional mechanism, have been demonstrated as biochemical targets for commercial fungicides. The 1,3,6,8 tetra hydroxynaphthalene reductase (T4HNR) protein, a member of SDRs family, catalyzes the naphthol reduction reaction in fungal melanin biosynthesis. We retrieved an orthologous member of T4HNR, (complexed with NADP(H) and pyroquilon from Magnaporthe grisea) in the FOL (namely; FOXG_04696) based on homology search, percent identity and sequence similarity (93% query cover; 49% identity). The hypothetical protein FOXG_04696 (T4HNR like) had conserved T-G-X-X-X-G-X-G motif (cofactor binding site) at N-terminus, similar to M. grisea (1JA9) and Y-X-X-X-K motif, as a part of the active site, bearing homologies with two fungal keto reductases T4HNR (M. grisea) and 17-β-hydroxysteroid dehydrogenase from Curvularia lunata (teleomorph: Cochliobolus lunatus PDB ID: 3IS3). The catalytic tetrad of T4HNR was replaced with ASN¹¹⁵, SER¹⁴¹, TYR¹⁵⁴, and LYS¹⁵⁸ in the FOXG_04696. The structural alignment and superposition of FOXG_04696 over the template proteins (3IS3 and 1JA9) revealed minimum RMSD deviations of the C alpha atomic coordinates, and therefore, had structural conservation. The best protein model (FOXG_04696) was docked with 37 fungicides, to evaluate their binding affinities. The Glide XP and YASARA docked complexes showed discrepancies in results, for scoring and ranking the binding affinities of fungicides. The docked complexes were further refined and rescored from their docked poses through 50 ns long MD simulations, and binding free energies (ΔG_bind) calculations, using MM/GBSA analysis, revealed Oxathiapiprolin and Famoxadone as better fungicides among the selected one. However, Famoxadone had better interaction of the docked residues, with best protein ligand contacts, minimum RMSD (high accuracy of the docking pose) and RMSF (structural integrity and conformational flexibility of docking) at the specified docking site. The Famoxadone was found to be acceptable based on in silico toxicity and in vitro growth inhibition assessment. We conclude that the FOXG_04696, could be employed as a novel candidate protein, for structure-based design, and screening of target fungicides against the FOL pathogen.

What do docking and QSAR tell us about the design of HIV-1 reverse transcriptase nonnucleoside inhibitors?

Despite vigorous studies, effective nonnucleoside inhibitors of HIV-1 reverse transcriptase (NNRTIs) are still in demand, not only due to toxicity and detrimental side effects of currently used drugs but also because of the emergence of multidrug-resistant viral strains. In this contribution, we present results of docking of 47 inhibitors to 107 allosteric centers of HIV-1 reverse transcriptase. Based on the average binding scores, we have constructed QSAR equations to elucidate directions of further developments in the inhibitor design that come from this structural data.

Microwave-assisted synthesis and antitumor evaluation of a new series of thiazolylcoumarin derivatives

A new series of thiazolylcoumarin derivatives was synthesized. The designed strategy embraced a molecular hybridization approach which involves the combination of the thiazole and coumarin pharmacophores together. The new hybrid compounds were tested for in vitro antitumor efficacy over cervical (Hela) and kidney fibroblast (COS-7) cancer cells. Compounds 5f, 5h, 5m and 5r displayed promising efficacy toward Hela cell line. In addition, 5h and 5r were found to be the most active candidates toward COS-7 cell line. The four active analogs, 5f, 5h, 5m and 5r were screened for in vivo antitumor activity over EAC cells in mice, as well as in vitro cytotoxicity toward W138 normal cells. Results illustrated that 5r has the highest in vivo activity, and that the four analogs are less cytotoxic than 5-FU toward W138 normal cells. In this study, 3D pharmacophore analysis was performed to investigate the matching pharmacophoric features of the synthesized compounds with trichostatin A. In silico studies showed that the investigated compounds meet the optimal needs for good oral absorption with no expected toxicity hazards.

Binding mode prediction and MD/MMPBSA-based free energy ranking for agonists of REV-ERBα/NCoR

The knowledge of the free energy of binding of small molecules to a macromolecular target is crucial in drug design as is the ability to predict the functional consequences of binding. We highlight how a molecular dynamics (MD)-based approach can be used to predict the free energy of small molecules, and to provide priorities for the synthesis and the validation via in vitro tests. Here, we study the dynamics and energetics of the nuclear receptor REV-ERBα with its co-repressor NCoR and 35 novel agonists. Our in silico approach combines molecular docking, molecular dynamics (MD), solvent-accessible surface area (SASA) and molecular mechanics poisson boltzmann surface area (MMPBSA) calculations. While docking yielded initial hints on the binding modes, their stability was assessed by MD. The SASA calculations revealed that the presence of the ligand led to a higher exposure of hydrophobic REV-ERB residues for NCoR recruitment. MMPBSA was very successful in ranking ligands by potency in a retrospective and prospective manner. Particularly, the prospective MMPBSA ranking-based validations for four compounds, three predicted to be active and one weakly active, were confirmed experimentally.

Machine Learning Consensus Scoring Improves Performance Across Targets in Structure-Based Virtual Screening

In structure-based virtual screening, compound ranking through a consensus of scores from a variety of docking programs or scoring functions, rather than ranking by scores from a single program, provides better predictive performance and reduces target performance variability. Here we compare traditional consensus scoring methods with a novel, unsupervised gradient boosting approach. We also observed increased score variation among active ligands and developed a statistical mixture model consensus score based on combining score means and variances. To evaluate performance, we used the common performance metrics ROCAUC and EF1 on 21 benchmark targets from DUD-E. Traditional consensus methods, such as taking the mean of quantile normalized docking scores, outperformed individual docking methods and are more robust to target variation. The mixture model and gradient boosting provided further improvements over the traditional consensus methods. These methods are readily applicable to new targets in academic research and overcome the potentially poor performance of using a single docking method on a new target.

Allosteric inhibition of topoisomerase I by pinostrobin: Molecular docking, spectroscopic and topoisomerase I activity studies

Cancer, the second major cause of mortality trailing the cardiovascular diseases, is a multifactorial heterogeneous disease and growing public health problem worldwide. Owing to severe adverse effects of currently available therapies, there is a growing interest in natural compounds present in our daily diet. Among various natural products, flavonoids-polyphenolic compounds have attracted much attention and have been well-documented for their biological activities. Some flavonoids may inhibit cancer cell proliferation by modulating the action of different enzymes and signal transduction pathways. In the present study, we have evaluated the effect of pinostrobin, a natural flavonoid, on the catalytic activity of topoisomerase I, an essential enzyme for normal DNA replication. Catalytic inhibition of topoisomerase I activity would impair DNA replication of rapidly dividing cancer cells and hence inhibits tumor progression. Pinostrobin interaction with the topoisomerase I and DNA assessed in silico indicated it to form a ternary complex with both. In silico data also suggested pinostrobin to be an effective allosteric inhibitor for topoisomerase I. Further, in vitro investigations such as ethidium bromide displacement assay and spectroscopic studies supported in silico results on the binding of pinostrobin at the interface of topoisomerase I and DNA. Pinostrobin effectively inhibited topoisomerase I activity in vitro further confirming our in silico and in vitro findings. Since topoisomerase I is essential for DNA replication, inhibition of its activity by pinostrobin highlights the therapeutic potential of pinostrobin as an anti-proliferative agent.

High-throughput detection of antioxidants in mulberry fruit using correlations between high-resolution mass and activity profiles of chromatographic fractions

BACKGROUND

Plant extracts contain a huge variety of pharmacologically active substances. Conventionally, various chromatographic methods must be applied several times to purify functional compounds to measure their functional activity. However, conventional purification methods are time-consuming and expensive due to the laborious purification process. Recently, a high-throughput discovery method that replaces such time-consuming purification processes was introduced; this method uses 15 T ultra-high-resolution Fourier transform ion cyclotron resonance mass spectrometry (15 T FT-ICR MS) and a high-throughput screening method. This 15 T FT-ICR MS provides unparalleled resolution and sub-ppm accuracy in mass measurements, while simultaneously detecting multiple compounds without separation. The high-throughput, simultaneous multi-component discovery method known as Scaling of Correlations between Activity and Mass Profiles (SCAMP) was used to detect functional compounds in a plant extract. We validated the performance of SCAMP using 33 fractions from antioxidant-rich mulberry ethyl acetate extract and known standard antioxidants.

RESULTS

The mulberry fruit was first separated into 33 fractions by LC and analyzed using high-resolution mass spectrometry. The antioxidative strength of the 33 fractions and standard antioxidants was measured. To validate the efficiency of this antioxidant discovery method, correlations between the antioxidation activity profile and changes in mass intensity of components within the 33 fractions were calculated to provide relative scores for the antioxidant candidate list. Enrichment curves and area under the curve (AUC) values were then calculated to compare the performance of the methods. Using this improved scoring method, five strong antioxidants, chlorogenic acid (14.2 ng), dihydoxy quercetin (46.2 ng), rutin (154.0 ng), quercetin (71.7 ng) and luteolin (3.5 ng) in 2 kg mulberry fruit, were found within the top 20 candidates.

CONCLUSIONS

We calculated AUCs in order to compare scoring methods quantitatively. Scoring systems were compared and calculated AUCs, where the AUCs for new scoring systems (0.98 and 0.99) were higher than the previously used correlation coefficient (AUC = 0.89). Using the new scoring algorithms, we successfully enriched thirteen unknown strong antioxidant candidates in addition to known antioxidants, methyl syringin and naringenin (3.5 ng) in mulberry extract. Targeted purification of these unknown candidates will significantly reduce purification time and labor.

The Comparison of Docking Search Algorithms and Scoring Functions

This chapter, composed of two parts, firstly provides molecular docking overview and secondly two molecular docking case studies are described. In overview, basic information about molecular docking are presented such as description of search algorithms and scoring functions applied in various docking programs. Brief description of methods utilized in some of the most popular docking programs also applied in our experimental work is provided. AutoDock, CDOCKER, GOLD, FlexX and FRED were used for docking studies of the DC-SIGN protein, while GOLD was further used for docking studies of cathepsin K protein. Methods and results of our studies with their contribution to science and medicine are presented. Content of the chapter is therefore appropriate for public of Medicinal and Organic Chemistry as an overview of docking studies, and also for Computational Chemists at the beginning of their work as the introduction to application of molecular docking programs.

Computational methods in drug discovery

The process for drug discovery and development is challenging, time consuming and expensive. Computer-aided drug discovery (CADD) tools can act as a virtual shortcut, assisting in the expedition of this long process and potentially reducing the cost of research and development. Today CADD has become an effective and indispensable tool in therapeutic development. The human genome project has made available a substantial amount of sequence data that can be used in various drug discovery projects. Additionally, increasing knowledge of biological structures, as well as increasing computer power have made it possible to use computational methods effectively in various phases of the drug discovery and development pipeline. The importance of in silico tools is greater than ever before and has advanced pharmaceutical research. Here we present an overview of computational methods used in different facets of drug discovery and highlight some of the recent successes. In this review, both structure-based and ligand-based drug discovery methods are discussed. Advances in virtual high-throughput screening, protein structure prediction methods, protein-ligand docking, pharmacophore modeling and QSAR techniques are reviewed.

Benchmark of four popular virtual screening programs: construction of the active/decoy dataset remains a major determinant of measured performance

Background

In a structure-based virtual screening, the choice of the docking program is essential for the success of a hit identification. Benchmarks are meant to help in guiding this choice, especially when undertaken on a large variety of protein targets. Here, the performance of four popular virtual screening programs, Gold, Glide, Surflex and FlexX, is compared using the Directory of Useful Decoys-Enhanced database (DUD-E), which includes 102 targets with an average of 224 ligands per target and 50 decoys per ligand, generated to avoid biases in the benchmarking. Then, a relationship between these program performances and the properties of the targets or the small molecules was investigated.

Results

The comparison was based on two metrics, with three different parameters each. The BEDROC scores with α = 80.5, indicated that, on the overall database, Glide succeeded (score > 0.5) for 30 targets, Gold for 27, FlexX for 14 and Surflex for 11. The performance did not depend on the hydrophobicity nor the openness of the protein cavities, neither on the families to which the proteins belong. However, despite the care in the construction of the DUD-E database, the small differences that remain between the actives and the decoys likely explain the successes of Gold, Surflex and FlexX. Moreover, the similarity between the actives of a target and its crystal structure ligand seems to be at the basis of the good performance of Glide. When all targets with significant biases are removed from the benchmarking, a subset of 47 targets remains, for which Glide succeeded for only 5 targets, Gold for 4 and FlexX and Surflex for 2.

Conclusion

The performance dramatic drop of all four programs when the biases are removed shows that we should beware of virtual screening benchmarks, because good performances may be due to wrong reasons. Therefore, benchmarking would hardly provide guidelines for virtual screening experiments, despite the tendency that is maintained, i.e., Glide and Gold display better performance than FlexX and Surflex. We recommend to always use several programs and combine their results. Graphical Abstract

Attracting cavities for docking. Replacing the rough energy landscape of the protein by a smooth attracting landscape

Molecular docking is a computational approach for predicting the most probable position of ligands in the binding sites of macromolecules and constitutes the cornerstone of structure-based computer-aided drug design. Here, we present a new algorithm called Attracting Cavities that allows molecular docking to be performed by simple energy minimizations only. The approach consists in transiently replacing the rough potential energy hypersurface of the protein by a smooth attracting potential driving the ligands into protein cavities. The actual protein energy landscape is reintroduced in a second step to refine the ligand position. The scoring function of Attracting Cavities is based on the CHARMM force field and the FACTS solvation model. The approach was tested on the 85 experimental ligand-protein structures included in the Astex diverse set and achieved a success rate of 80% in reproducing the experimental binding mode starting from a completely randomized ligand conformer. The algorithm thus compares favorably with current state-of-the-art docking programs.

Computational Approaches for Decoding Select Odorant-Olfactory Receptor Interactions Using Mini-Virtual Screening

Olfactory receptors (ORs) belong to the class A G-Protein Coupled Receptor superfamily of proteins. Unlike G-Protein Coupled Receptors, ORs exhibit a combinatorial response to odors/ligands. ORs display an affinity towards a range of odor molecules rather than binding to a specific set of ligands and conversely a single odorant molecule may bind to a number of olfactory receptors with varying affinities. The diversity in odor recognition is linked to the highly variable transmembrane domains of these receptors. The purpose of this study is to decode the odor-olfactory receptor interactions using in silico docking studies. In this study, a ligand (odor molecules) dataset of 125 molecules was used to carry out in silico docking using the GLIDE docking tool (SCHRODINGER Inc Pvt LTD). Previous studies, with smaller datasets of ligands, have shown that orthologous olfactory receptors respond to similarly-tuned ligands, but are dramatically different in their efficacy and potency. Ligand docking results were applied on homologous pairs (with varying sequence identity) of ORs from human and mouse genomes and ligand binding residues and the ligand profile differed among such related olfactory receptor sequences. This study revealed that homologous sequences with high sequence identity need not bind to the same/ similar ligand with a given affinity. A ligand profile has been obtained for each of the 20 receptors in this analysis which will be useful for expression and mutation studies on these receptors.

Fragment-based similarity searching with infinite color space

Fragment-based searching and abstract representation of molecular features through reduced graphs have separately been used for virtual screening. Here, we combine these two approaches and apply the algorithm RedFrag to virtual screens retrospectively and prospectively. It uses a new type of reduced graph that does not suffer from information loss during its construction and bypasses the necessity of feature definitions. Built upon chemical epitopes resulting from molecule fragmentation, the reduced graph embodies physico-chemical and 2D-structural properties of a molecule. Reduced graphs are compared with a continuous-similarity-distance-driven maximal common subgraph algorithm, which calculates similarity at the fragmental and topological levels. The performance of the algorithm is evaluated by retrieval experiments utilizing precompiled validation sets. By predicting and experimentally testing ligands for endothiapepsin, a challenging model protease, the method is assessed in a prospective setting. Here, we identified five novel ligands with affinities as low as 2.08 μM.

Benchmarking Data Sets for the Evaluation of Virtual Ligand Screening Methods: Review and Perspectives

Virtual screening methods are commonly used nowadays in drug discovery processes. However, to ensure their reliability, they have to be carefully evaluated. The evaluation of these methods is often realized in a retrospective way, notably by studying the enrichment of benchmarking data sets. To this purpose, numerous benchmarking data sets were developed over the years, and the resulting improvements led to the availability of high quality benchmarking data sets. However, some points still have to be considered in the selection of the active compounds, decoys, and protein structures to obtain optimal benchmarking data sets.

MoDock: A multi-objective strategy improves the accuracy for molecular docking

BACKGROUND

As a main method of structure-based virtual screening, molecular docking is the most widely used in practice. However, the non-ideal efficacy of scoring functions is thought as the biggest barrier which hinders the improvement of the molecular docking method.

RESULTS

A new multi-objective strategy for molecular docking, named as MoDock, is presented to further improve the docking accuracy with available scoring functions. Instead of simple combination of multiple objectives with fixed weight factors, an aggregate function is adopted to approximate the real solution of the original multi-objective and multi-constraint problem, which will simultaneously smooth the energy surface of the combined scoring functions. Then, method of centers and genetic algorithm are used to find the optimal solution. Tests of MoDock against the GOLD test data set reveal the multi-objective strategy improves the docking accuracy over the individual scoring functions. Meanwhile, a 70% ratio of the good docking solutions with the RMSD value below 1.0 Å outperforms other 6 commonly used docking programs, even with a flexible receptor docking program included.

CONCLUSIONS

The results show MoDock is an effective strategy to overcome the deviations brought by single scoring function, and improves the prediction power of molecular docking.

Comparing sixteen scoring functions for predicting biological activities of ligands for protein targets

Accurately predicting relative binding affinities and biological potencies for ligands that interact with proteins remains a significant challenge for computational chemists. Most evaluations of docking and scoring algorithms have focused on enhancing ligand affinity for a protein by optimizing docking poses and enrichment factors during virtual screening. However, there is still relatively limited information on the accuracy of commercially available docking and scoring software programs for correctly predicting binding affinities and biological activities of structurally related inhibitors of different enzyme classes. Presented here is a comparative evaluation of eight molecular docking programs (Autodock Vina, Fitted, FlexX, Fred, Glide, GOLD, LibDock, MolDock) using sixteen docking and scoring functions to predict the rank-order activity of different ligand series for six pharmacologically important protein and enzyme targets (Factor Xa, Cdk2 kinase, Aurora A kinase, COX-2, pla2g2a, β Estrogen receptor). Use of Fitted gave an excellent correlation (Pearson 0.86, Spearman 0.91) between predicted and experimental binding only for Cdk2 kinase inhibitors. FlexX and GOLDScore produced good correlations (Pearson>0.6) for hydrophilic targets such as Factor Xa, Cdk2 kinase and Aurora A kinase. By contrast, pla2g2a and COX-2 emerged as difficult targets for scoring functions to predict ligand activities. Although possessing a high hydrophobicity in its binding site, β Estrogen receptor produced reasonable correlations using LibDock (Pearson 0.75, Spearman 0.68). These findings can assist medicinal chemists to better match scoring functions with ligand-target systems for hit-to-lead optimization using computer-aided drug design approaches.