AMMOS2: a web server for protein-ligand-water complexes refinement via molecular mechanics

Dynamics simulation, energetics calculation and experimental analysis of the intermolecular interaction between human neonatal ABL SH3 domain and its N-substituted peptoid ligands

Non-receptor tyrosine kinase of neonatal ABL (nABL) is distributed in the nucleus and cytoplasm of proliferating cells in embryo and neonate, and has been implicated in the pathogenesis of neonatal leukemia and other hematological diseases. The kinase contains a regulatory Src homology 3 (SH3) domain that can specifically recognize proline-rich peptide segments on its partner protein surface. In this study, we systematically investigated the N-substitution effect on the binding of an empirically designed proline-rich peptide p9 to nABL SH3 domain by integrating dynamics simulations, energetics calculations and fluorescence affinity assays. The p9 is an almost all proline-composed decapeptide, with only a sole tyrosine at its residue 4, which has been found to bind nABL SH3 domain at a micromolar level in a class I mode. Here, the non-key residues of p9 peptide were independently replaced by various N-substituted amino acids to create a systematic N-substitution profile, from which we can identify those favorable, neutral and unfavorable substitutions at each peptide residue. On this basis a combinatorial peptoid library was rationally designed by systematically combining the favorable N-substituted amino acids at non-key residues of p9 peptide, thus resulting in a number of its peptoid counterparts. The binding affinity of top peptoid hits was observed to be comparable with or improved moderately relative to p9 peptide, with Kd ranging between 3.1 and 76 μM. Structural analysis revealed that the peptoids can be divided into exposed, polar and hydrophobic regions from N- to C-termini, in which the polar and hydrophobic regions confer specificity and stability to the domain-peptoid interaction, respectively. In addition, a designed peptoid was also observed to exhibit 5.3-fold SH3-selectivity for nABL over cSRC, suggesting that the N-substitution can be used to improve not only binding affinity but also recognition specificity of SH3 binders.Communicated by Ramaswamy H. Sarma.

Trends in application of advancing computational approaches in GPCR ligand discovery

G protein-coupled receptors (GPCRs) comprise the most important superfamily of protein targets in current ligand discovery and drug development. GPCRs are integral membrane proteins that play key roles in various cellular signaling processes. Therefore, GPCR signaling pathways are closely associated with numerous diseases, including cancer and several neurological, immunological, and hematological disorders. Computer-aided drug design (CADD) can expedite the process of GPCR drug discovery and potentially reduce the actual cost of research and development. Increasing knowledge of biological structures, as well as improvements on computer power and algorithms, have led to unprecedented use of CADD for the discovery of novel GPCR modulators. Similarly, machine learning approaches are now widely applied in various fields of drug target research. This review briefly summarizes the application of rising CADD methodologies, as well as novel machine learning techniques, in GPCR structural studies and bioligand discovery in the past few years. Recent novel computational strategies and feasible workflows are updated, and representative cases addressing challenging issues on olfactory receptors, biased agonism, and drug-induced cardiotoxic effects are highlighted to provide insights into future GPCR drug discovery.

Scaffold-based analysis of nonpeptide oncogenic FTase inhibitors using multiple similarity matching, binding affinity scoring and enzyme inhibition assay

Oncogenic protein farnesyltransferase (FTase) is a key enzyme responsible for the lipid modification of a large and important number of proteins including Ras, which has been recognized as a druggable target of diverse cancers. Here, we report a systematic scaffold-based analysis to investigate the affinity, selectivity and cross-reactivity of nonpeptide inhibitors across ontology-enriched, disease-associated FTase mutants, by integrating multiple similarity matching, binding affinity scoring and enzyme inhibition assay. It is revealed that nonpeptide inhibitors are generally insensitive to FTase mutations; many of them cannot definitely select for wild-type target over mutant enzymes. Therefore, off-target is observed as a common phenomenon for the untargeted consequence of targeted therapies with FTase inhibition. This is not unexpected if considering that the enzyme active site is highly conserved in composition, configuration and function. The off-target, on the one hand, causes nonpeptide inhibitors with adverse drug reactions and, on the other hand, makes the inhibitors as promising candidates for the new use of old drugs. To practice the latter, a number of unexpected mutant-inhibitor interactions involved in cancer signaling pathways are uncovered in the created profile, from which several nonpeptide inhibitors are identified as insensitive to a drug-resistant mutation. Structural analysis suggests that the inhibitor ligands can bind to the mutant active site in a similar manner with wild-type target, although their nonbonded interactions appear to be impaired moderately upon the mutation.

New machine learning and physics-based scoring functions for drug discovery

Scoring functions are essential for modern in silico drug discovery. However, the accurate prediction of binding affinity by scoring functions remains a challenging task. The performance of scoring functions is very heterogeneous across different target classes. Scoring functions based on precise physics-based descriptors better representing protein–ligand recognition process are strongly needed. We developed a set of new empirical scoring functions, named DockTScore, by explicitly accounting for physics-based terms combined with machine learning. Target-specific scoring functions were developed for two important drug targets, proteases and protein–protein interactions, representing an original class of molecules for drug discovery. Multiple linear regression (MLR), support vector machine and random forest algorithms were employed to derive general and target-specific scoring functions involving optimized MMFF94S force-field terms, solvation and lipophilic interactions terms, and an improved term accounting for ligand torsional entropy contribution to ligand binding. DockTScore scoring functions demonstrated to be competitive with the current best-evaluated scoring functions in terms of binding energy prediction and ranking on four DUD-E datasets and will be useful for in silico drug design for diverse proteins as well as for specific targets such as proteases and protein–protein interactions. Currently, the MLR DockTScore is available at www.dockthor.lncc.br.

Prioritizing potential ACE2 inhibitors in the COVID-19 pandemic: Insights from a molecular mechanics-assisted structure-based virtual screening experiment

Angiotensin-converting enzyme 2 (ACE2) is a membrane-bound zinc metallopeptidase that generates the vasodilatory peptide angiotensin 1-7 and thus performs a protective role in heart disease. It is considered an important therapeutic target in controlling the COVID-19 outbreak, since SARS-CoV-2 enters permissive cells via an ACE2-mediated mechanism. The present in silico study attempted to repurpose existing drugs for use as prospective viral-entry inhibitors targeting human ACE2. Initially, a clinically approved drug library of 7,173 ligands was screened against the receptor using molecular docking, followed by energy minimization and rescoring of docked ligands. Finally, potential binders were inspected to ensure molecules with different scaffolds were engaged in favorable contacts with both the metal cofactor and the critical residues lining the receptor's active site. The results of the calculations suggest that lividomycin, burixafor, quisinostat, fluprofylline, pemetrexed, spirofylline, edotecarin, and diniprofylline emerge as promising repositionable drug candidates for stabilizing the closed (substrate/inhibitor-bound) conformation of ACE2, thereby shifting the relative positions of the receptor's critical exterior residues recognized by SARS-CoV-2. This study is among the rare ones in the relevant scientific literature to search for potential ACE2 inhibitors. In practical terms, the drugs, unmodified as they are, may be introduced into the therapeutic armamentarium of the ongoing fight against COVID-19 now, or their scaffolds may serve as rich skeletons for designing novel ACE2 inhibitors in the near future.

Machine learning classification can reduce false positives in structure-based virtual screening

With the recent explosion in the size of libraries available for screening, virtual screening is positioned to assume a more prominent role in early drug discovery's search for active chemical matter. In typical virtual screens, however, only about 12% of the top-scoring compounds actually show activity when tested in biochemical assays. We argue that most scoring functions used for this task have been developed with insufficient thoughtfulness into the datasets on which they are trained and tested, leading to overly simplistic models and/or overtraining. These problems are compounded in the literature because studies reporting new scoring methods have not validated their models prospectively within the same study. Here, we report a strategy for building a training dataset (D-COID) that aims to generate highly compelling decoy complexes that are individually matched to available active complexes. Using this dataset, we train a general-purpose classifier for virtual screening (vScreenML) that is built on the XGBoost framework. In retrospective benchmarks, our classifier shows outstanding performance relative to other scoring functions. In a prospective context, nearly all candidate inhibitors from a screen against acetylcholinesterase show detectable activity; beyond this, 10 of 23 compounds have IC₅₀ better than 50 μM. Without any medicinal chemistry optimization, the most potent hit has IC₅₀ 280 nM, corresponding to K_i of 173 nM. These results support using the D-COID strategy for training classifiers in other computational biology tasks, and for vScreenML in virtual screening campaigns against other protein targets. Both D-COID and vScreenML are freely distributed to facilitate such efforts.

Computational systematic selectivity of the Fasalog inhibitors between ROCK-I and ROCK-II kinase isoforms in Alzheimer's disease

Human Rho-associated coiled-coil forming kinase (ROCK) is a class of essential neurokinases that consists of two structurally conserved isoforms ROCK-I and ROCK-II; they have been revealed to play distinct roles in the pathogenesis of Alzheimer's disease (AD) and other neurological disorders. Selective targeting of the two kinase isoforms with small-molecule inhibitors is a great challenge due to the surprisingly high homology in kinase domain (92 %) and the full identity in kinase active site (100 %). Here, we describe a computational protocol to systematically profile the selectivity of Fasudil and its 25 analogs (termed as Fasalogs) between the two kinase isoforms. It is suggested that the substitution of Fasudil's 1,4-diazepane moiety with rigid ring such as Ripasudil and Dimehtylfasudil would render the resulting inhibitors of ROCK-II over ROCK-I (II-o-I) selectivity, while the substitution with long, flexible group such as H-89 and BDBM92607 tends to have I-o-II selectivity. Structural analysis reveals that the inhibitor affinity is not only determined by the identical active site, but also contributed from the non-identical first and second shells of the site as well as other non-conserved kinase regions, which can indirectly influence the active site and inhibitor binding through allosteric effect. A further kinase assay basically confirms the computational findings, which also exhibits a good consistence with theoretical selectivity over 10 tested samples (Rp = 0.89). In particular, the Fasalog compounds Dimehtylfasudil and H-89 are identified as II-o-I and I-o-II selective inhibitors. They can be considered as promising lead molecular entities to develop new specific ROCK isoform-selective Fasalog inhibitors.

Molecular Structure, Binding Affinity, and Biological Activity in the Epigenome

Development of valid structure–activity relationships (SARs) is a key to the elucidation of pathomechanisms of epigenetic diseases and the development of efficient, new drugs. The present review is based on selected methodologies and applications supplying molecular structure, binding affinity and biological activity data for the development of new SARs. An emphasis is placed on emerging trends and permanent challenges of new discoveries of SARs in the context of proteins as epigenetic drug targets. The review gives a brief overview and classification of the molecular background of epigenetic changes, and surveys both experimental and theoretical approaches in the field. Besides the results of sophisticated, cutting edge techniques such as cryo-electron microscopy, protein crystallography, and isothermal titration calorimetry, examples of frequently used assays and fast screening techniques are also selected. The review features how different experimental methods and theoretical approaches complement each other and result in valid SARs of the epigenome.

Virtual screening web servers: designing chemical probes and drug candidates in the cyberspace

The interplay between life sciences and advancing technology drives a continuous cycle of chemical data growth; these data are most often stored in open or partially open databases. In parallel, many different types of algorithms are being developed to manipulate these chemical objects and associated bioactivity data. Virtual screening methods are among the most popular computational approaches in pharmaceutical research. Today, user-friendly web-based tools are available to help scientists perform virtual screening experiments. This article provides an overview of internet resources enabling and supporting chemical biology and early drug discovery with a main emphasis on web servers dedicated to virtual ligand screening and small-molecule docking. This survey first introduces some key concepts and then presents recent and easily accessible virtual screening and related target-fishing tools as well as briefly discusses case studies enabled by some of these web services. Notwithstanding further improvements, already available web-based tools not only contribute to the design of bioactive molecules and assist drug repositioning but also help to generate new ideas and explore different hypotheses in a timely fashion while contributing to teaching in the field of drug development.

A Free Web-Based Protocol to Assist Structure-Based Virtual Screening Experiments

Chemical biology and drug discovery are complex and costly processes. In silico screening approaches play a key role in the identification and optimization of original bioactive molecules and increase the performance of modern chemical biology and drug discovery endeavors. Here, we describe a free web-based protocol dedicated to small-molecule virtual screening that includes three major steps: ADME-Tox filtering (via the web service FAF-Drugs4), docking-based virtual screening (via the web service MTiOpenScreen), and molecular mechanics optimization (via the web service AMMOS2 [Automatic Molecular Mechanics Optimization for in silico Screening]). The online tools FAF-Drugs4, MTiOpenScreen, and AMMOS2 are implemented in the freely accessible RPBS (Ressource Parisienne en Bioinformatique Structurale) platform. The proposed protocol allows users to screen thousands of small molecules and to download the top 1500 docked molecules that can be further processed online. Users can then decide to purchase a small list of compounds for in vitro validation. To demonstrate the potential of this online-based protocol, we performed virtual screening experiments of 4574 approved drugs against three cancer targets. The results were analyzed in the light of published drugs that have already been repositioned on these targets. We show that our protocol is able to identify active drugs within the top-ranked compounds. The web-based protocol is user-friendly and can successfully guide the identification of new promising molecules for chemical biology and drug discovery purposes.

Bone formation activity of an osteogenic dodecapeptide from blue mussels (Mytilus edulis)

A novel osteogenic dodecapeptide peptide (PIE), IEELEEELEAER, was purified from the protein hydrolysate of blue mussels (Mytilus edulis). PIE was identified using a capillary electrophoresis electrospray ionization-quadrupole-time of flight mass spectrometer. PIE showed a good reduction in the bone loss in ovariectomized mice, and it also increased the bone mineral density of the ovariectomized mice. PIE has a high affinity with integrins (PDB: , ). There are 8 and 12 amino acid residues from PIE that interact with integrins and , respectively. PIE accelerates the transformation of G0/G1 phase cells into G2 M phase cells, which promotes the growth of osteoblasts. PIE (100 μg mL^-1) can enhance alkaline phosphatase (ALP) activity by 26.48% compared with the control, and it also inhibits the growth of osteoclasts and tartrate resistant acid phosphatase (TRAP) activity. Therefore, PIE may contribute to preventing osteoporosis both in vitro and in vivo.

DPubChem: a web tool for QSAR modeling and high-throughput virtual screening

High-throughput screening (HTS) performs the experimental testing of a large number of chemical compounds aiming to identify those active in the considered assay. Alternatively, faster and cheaper methods of large-scale virtual screening are performed computationally through quantitative structure-activity relationship (QSAR) models. However, the vast amount of available HTS heterogeneous data and the imbalanced ratio of active to inactive compounds in an assay make this a challenging problem. Although different QSAR models have been proposed, they have certain limitations, e.g., high false positive rates, complicated user interface, and limited utilization options. Therefore, we developed DPubChem, a novel web tool for deriving QSAR models that implement the state-of-the-art machine-learning techniques to enhance the precision of the models and enable efficient analyses of experiments from PubChem BioAssay database. DPubChem also has a simple interface that provides various options to users. DPubChem predicted active compounds for 300 datasets with an average geometric mean and F₁ score of 76.68% and 76.53%, respectively. Furthermore, DPubChem builds interaction networks that highlight novel predicted links between chemical compounds and biological assays. Using such a network, DPubChem successfully suggested a novel drug for the Niemann-Pick type C disease. DPubChem is freely available at www.cbrc.kaust.edu.sa/dpubchem .

Identification by Molecular Docking ofHomoisoflavones from Leopoldia comosa as Ligands of Estrogen Receptors

The physiological responses to estrogen hormones are mediated within specific tissues by at least two distinct receptors, ERα and ERβ. Several natural and synthetic molecules show activity by interacting with these proteins. In particular, a number of vegetal compounds known as phytoestrogens shows estrogenic or anti-estrogenic activity. The majority of these compounds belongs to the isoflavones family and the most representative one, genistein, shows anti-proliferative effects on various hormone-sensitive cancer cells, including breast, ovarian and prostate cancer. In this work we describe the identification of structurally related homoisoflavones isolated from Leopoldia comosa (L.) Parl. (L. comosa), a perennial bulbous plant, potentially useful as hormonal substitutes or complements in cancer treatments. Two of these compounds have been selected as potential ligands of estrogen receptors (ERs) and the interaction with both isoforms of estrogen receptors have been investigated through molecular docking on their crystallographic structures. The results provide evidence of the binding of these compounds to the target receptors and their interactions with key residues of the active sites of the two proteins, and thus they could represent suitable leads for the development of novel tools for the dissection of ER signaling and the development of new pharmacological treatments in hormone-sensitive cancers.

Annotation of Peptide Structures Using SMILES and Other Chemical Codes-Practical Solutions

Contemporary peptide science exploits methods and tools of bioinformatics, and cheminformatics. These approaches use different languages to describe peptide structures—amino acid sequences and chemical codes (especially SMILES), respectively. The latter may be applied, e.g., in comparative studies involving structures and properties of peptides and peptidomimetics. Progress in peptide science “in silico” may be achieved via better communication between biologists and chemists, involving the translation of peptide representation from amino acid sequence into SMILES code. Recent recommendations concerning good practice in chemical information include careful verification of data and their annotation. This publication discusses the generation of SMILES representations of peptides using existing software. Construction of peptide structures containing unnatural and modified amino acids (with special attention paid on glycosylated peptides) is also included. Special attention is paid to the detection and correction of typical errors occurring in SMILES representations of peptides and their correction using molecular editors. Brief recommendations for training of staff working on peptide annotations, are discussed as well.

Why do antifreeze proteins require a solenoid?

Proteins whose presence prevents water from freezing in living organisms at temperatures below 0 °C are referred to as antifreeze proteins. This group includes molecules of varying size (from 30 to over 300 aa) and variable secondary/supersecondary conformation. Some of these proteins also contain peculiar structural motifs called solenoids. We have applied the fuzzy oil drop model in the analysis of four categories of antifreeze proteins: 1 - very small proteins, i.e. helical peptides (below 40 aa); 2 - small globular proteins (40-100 aa); 3 - large globular proteins (>100 aa) and 4 - proteins containing solenoids. The FOD model suggests a mechanism by which antifreeze proteins prevent freezing. In accordance with this theory, the presence of the protein itself produces an ordering of water molecules which counteracts the formation of ice crystals. This conclusion is supported by analysis of the ordering of hydrophobic and hydrophilic residues in antifreeze proteins, revealing significant variability - from perfect adherence to the fuzzy oil drop model through structures which lack a clearly defined hydrophobic core, all the way to linear arrangement of alternating local minima and maxima propagating along the principal axis of the solenoid (much like in amyloids). The presented model - alternative with respect to the ice docking model - explains the antifreeze properties of compounds such as saccharides and fatty acids. The fuzzy oil drop model also enables differentiation between amyloids and antifreeze proteins.