MDFit: automated molecular simulations workflow enables high throughput assessment of ligands-protein dynamics.

Molecular dynamics (MD) simulation is a powerful tool for characterizing ligand-protein conformational dynamics and offers significant advantages over docking and other rigid structure-based computational methods. However, setting up, running, and analyzing MD simulations continues to be a multi-step process making it cumbersome to assess a library of ligands in a protein binding pocket using MD. We present an automated workflow that streamlines setting up, running, and analyzing Desmond MD simulations for protein-ligand complexes using machine learning (ML) models. The workflow takes a library of pre-docked ligands and a prepared protein structure as input, sets up and runs MD with each protein-ligand complex, and generates simulation fingerprints for each ligand. Simulation fingerprints (SimFP) capture protein-ligand compatibility, including stability of different ligand-pocket interactions and other useful metrics that enable easy rank-ordering of the ligand library for pocket optimization. SimFPs from a ligand library are used to build & deploy ML models that predict binding assay outcomes and automatically infer important interactions. Unlike relative free-energy methods that are constrained to assess ligands with high chemical similarity, ML models based on SimFPs can accommodate diverse ligand sets. We present two case studies on how SimFP helps delineate structure-activity relationship (SAR) trends and explain potency differences across matched-molecular pairs of (1) cyclic peptides targeting PD-L1 and (2) small molecule inhibitors targeting CDK9.

Rationalizing PROTAC-Mediated Ternary Complex Formation Using Rosetta.

Proteolysis-targeting chimaeras (PROTACs) are molecules that combine a target-binding warhead with an E3 ligase-recruiting moiety; by drawing the target protein into a ternary complex with the E3 ligase, PROTACs induce target protein degradation. While PROTACs hold exciting potential as chemical probes and as therapeutic agents, development of a PROTAC typically requires synthesis of numerous analogs to thoroughly explore variations on the chemical linker; without extensive trial and error, it is unclear how to link the two protein-recruiting moieties to promote formation of a productive ternary complex. Here, we describe a structure-based computational method for evaluating the suitability of a given linker for ternary complex formation. Our method uses Rosetta to dock the protein components and then builds the PROTAC from its component fragments into each binding mode; complete models of the ternary complex are then refined. We apply this approach to retrospectively evaluate multiple PROTACs from the literature, spanning diverse target proteins. We find that modeling ternary complex formation is sufficient to explain both activity and selectivity reported for these PROTACs, implying that other cellular factors are not key determinants of activity in these cases. We further find that interpreting PROTAC activity is best approached using an ensemble of structures of the ternary complex rather than a single static conformation and that members of a structurally conserved protein family can be recruited by the same PROTAC through vastly different binding modes. To encourage adoption of these methods and promote further analyses, we disseminate both the computational methods and the models of ternary complexes.

Exquisite ligand stereoselectivity of a Drosophila juvenile hormone receptor contrasts with its broad agonist repertoire.

The sesquiterpenoid juvenile hormone (JH) is vital to insect development and reproduction. Intracellular JH receptors have recently been established as basic helix-loop-helix transcription factor (bHLH)/PAS proteins in Drosophila melanogaster known as germ cell-expressed (Gce) and its duplicate paralog, methoprene-tolerant (Met). Upon binding JH, Gce/Met activates its target genes. Insects possess multiple native JH homologs whose molecular activities remain unexplored, and diverse synthetic compounds including insecticides exert JH-like effects. How the JH receptor recognizes its ligands is unknown. To determine which structural features define an active JH receptor agonist, we tested several native JHs and their nonnative geometric and optical isomers for the ability to bind the Drosophila JH receptor Gce, to induce Gce-dependent transcription, and to affect the development of the fly. Our results revealed high ligand stereoselectivity of the receptor. The geometry of the JH skeleton, dictated by two stereogenic double bonds, was the most critical feature followed by the presence of an epoxide moiety at a terminal position. The optical isomerism at carbon C11 proved less important even though Gce preferentially bound a natural JH enantiomer. The results of receptor-ligand-binding and cell-based gene activation assays tightly correlated with the ability of different geometric JH isomers to induce gene expression and morphogenetic effects in the developing insects. Molecular modeling supported the requirement for the proper double-bond geometry of JH, which appears to be its major selective mechanism. The strict stereoselectivity of Gce toward the natural hormone contrasts with the high potency of synthetic Gce agonists of disparate chemistries.

Substituents at the C3' and C3'N positions are critical for taxanes to overcome acquired resistance of cancer cells to paclitaxel.

We tested the role of substituents at the C3' and C3'N positions of the taxane molecule to identify taxane derivatives capable of overcoming acquired resistance to paclitaxel. Paclitaxel-resistant sublines SK-BR-3/PacR and MCF-7/PacR as well as the original paclitaxel-sensitive breast cancer cell lines SK-BR-3 and MCF-7 were used for testing. Increased expression of the ABCB1 transporter was found to be involved in the acquired resistance. We tested three groups of taxane derivatives: (1) phenyl group at both C3' and C3'N positions, (2) one phenyl at one of the C3' and C3'N positions and a non-aromatic group at the second position, (3) a non-aromatic group at both C3' and C3'N positions. We found that the presence of phenyl groups at both C3' and C3'N positions is associated with low capability of overcoming acquired paclitaxel resistance compared to taxanes containing at least one non-aromatic substituent at the C3' and C3'N positions. The increase in the ATPase activity of ABCB1 transporter after the application of taxanes from the first group was found to be somewhat higher than after the application of taxanes from the third group. Molecular docking studies demonstrated that the docking score was the lowest, i.e. the highest binding affinity, for taxanes from the first group. It was intermediate for taxanes from the second group, and the highest for taxanes from the third group. We conclude that at least one non-aromatic group at the C3' and C3'N positions of the taxane structure, resulting in reduced affinity to the ABCB1 transporter, brings about high capability of taxane to overcome acquired resistance of breast cancer cells to paclitaxel, due to less efficient transport of the taxane compound out of the cancer cells.

Artificial proteins as allosteric modulators of PDZ3 and SH3 in two-domain constructs: A computational characterization of novel chimeric proteins.

Artificial multidomain proteins with enhanced structural and functional properties can be utilized in a broad spectrum of applications. The design of chimeric fusion proteins utilizing protein domains or one-domain miniproteins as building blocks is an important advancement for the creation of new biomolecules for biotechnology and medical applications. However, computational studies to describe in detail the dynamics and geometry properties of two-domain constructs made from structurally and functionally different proteins are lacking. Here, we tested an in silico design strategy using all-atom explicit solvent molecular dynamics simulations. The well-characterized PDZ3 and SH3 domains of human zonula occludens (ZO-1) (3TSZ), along with 5 artificial domains and 2 types of molecular linkers, were selected to construct chimeric two-domain molecules. The influence of the artificial domains on the structure and dynamics of the PDZ3 and SH3 domains was determined using a range of analyses. We conclude that the artificial domains can function as allosteric modulators of the PDZ3 and SH3 domains. Proteins 2016; 84:1358-1374. © 2016 Wiley Periodicals, Inc.

Pharmacophore modeling, 3D-QSAR and DFT studies of IWR small-molecule inhibitors of Wnt response.

In this study, a 5-point pharmacophore model was developed and the model was used to generate a predictive atom-based 3D quantitative structure activity relationship (3D-QSAR) analysis for the studied dataset of 50 compounds. The obtained 3D-QSAR model shows correlation coefficient (R(2)) of 0.87 for training set compounds and excellent predictive power (Q(2)) of 0.81 for cross-validated test set compounds. External validation indicated that our 3D-QSAR model has high predictive power with [Formula: see text] and [Formula: see text] values of 0.99 and 0.65, respectively. The most active and least active compounds were further optimized using density functional theory at B3LYP/3-21*G level. Further, pharmacophoric model was employed for pharmacophore-based screening to identify potential inhibitors against Wnt/ß-catenin pathway. Hence, these molecules could act as selective inhibitors of Wnt/ß-catenin pathway which can be experimentally validated. The backbone of these inhibitors could serve as templates for designing drug-like molecules specifically targeting Wnt/ß-catenin pathway.

Poor Generalization by Current Deep Learning Models for Predicting Binding Affinities of Kinase Inhibitors.

The extreme surge of interest over the past decade surrounding the use of neural networks has inspired many groups to deploy them for predicting binding affinities of drug-like molecules to their receptors. A model that can accurately make such predictions has the potential to screen large chemical libraries and help streamline the drug discovery process. However, despite reports of models that accurately predict quantitative inhibition using protein kinase sequences and inhibitors' SMILES strings, it is still unclear whether these models can generalize to previously unseen data. Here, we build a Convolutional Neural Network (CNN) analogous to those previously reported and evaluate the model over four datasets commonly used for inhibitor/kinase predictions. We find that the model performs comparably to those previously reported, provided that the individual data points are randomly split between the training set and the test set. However, model performance is dramatically deteriorated when all data for a given inhibitor is placed together in the same training/testing fold, implying that information leakage underlies the models' performance. Through comparison to simple models in which the SMILES strings are tokenized, or in which test set predictions are simply copied from the closest training set data points, we demonstrate that there is essentially no generalization whatsoever in this model. In other words, the model has not learned anything about molecular interactions, and does not provide any benefit over much simpler and more transparent models. These observations strongly point to the need for richer structure-based encodings, to obtain useful prospective predictions of not-yet-synthesized candidate inhibitors.

Docking Studies of Adenosine Analogues with NS5 Methyltransferase of Yellow Fever Virus.

Yellow fever virus (YFV) is caused by single stranded positive RNA virus called Flavivirus. Till now no specific antiviral agents are available for the treatment of YFV, and despite a commercial YFV vaccine, there are still approximately 30,000 deaths worldwide each year and cases have been increasing in the last 20 years. Here, the effects of adenosine analogues and commercially available adenosine derivative drugs on NS5 methyltransferase of YFV have been performed by the comparative docking study. Based on the docking score, the glide energy and the number of interactions of the adenosine analogues with the Pubchem ID 13792 and 1077 showed the better scoring function than the best ranked commercially available adenosine analogue derived antiviral drug Cc3ado. From the docking result it reveals that these adenosine analogues can bind to the active site of NS5 methyltransferase protein and inhibit the viral replication.

The identification of new substrates of human DHRS7 by molecular modeling and in vitro testing.

Human DHRS7 (SDR34C1) is one of insufficiently described enzymes of the short-chain dehydrogenase/reductase superfamily. The members of this superfamily often play an important pato/physiological role in the human body, participating in the metabolism of diverse substrates (e.g. retinoids, steroids, xenobiotics). A systematic approach to the identification of novel, physiological substrates of DHRS7 based on a combination of homology modeling, structure-based virtual screening and experimental evaluation has been used. Three novel substrates of DHRS7 (dihydrotestosterone, benzil and 4,4'-dimetylbenzil) have been described.

Computational insights into the inhibition of influenza viruses by rupestonic acid derivatives: pharmacophore modeling, 3D-QSAR, CoMFA and COMSIA studies.

The pharmacophore modeling and 3D-QSAR studies were performed on a series of amino alkyl rupestonates (Rupestonic Acid) derivatives reported for H1N1, H3N2 and Influenza B virus, NA inhibition. In order to improve the efficacy of amino alkyl rupestonates derivatives, a four point pharmacophore model with one acceptor and three hydrophobic regions was developed. Furthermore, the 3D-QSAR model was generated based on the pharmacophore hypothesis (AHHH) for each subtype. The hypothesis was more significant with R(2)=0.9204, Q(2)=0.917 for H1N1, R(2)=0.8911, Q(2)=0.8905 for H3N2 and R(2)=0.8385, Q(2)=0.7043 for Influenza B virus. The 3D-QSAR results provided an invaluable insight into structure activity correlation and it was shown that the hydrophobic regions were crucial for inhibitory activity. CoMFA and COMSIA validation had been done by leave one out and no validation methods.

Binding mode analysis of anti-influenza drugs in H1N1 (2009) and H5N1 influenza A virus and designing of potential H1N1 inhibitors.

The main goal of this study is to understand the molecular-level interactions of neuraminidase inhibitor. The molecular docking, molecular dynamics and binding energy calculation analyses were carried out and the results revealed that the 150-cavitiy in the active site may play an important role in binding of drugs. Free energy calculations revealed that electrostatic interaction is more favourable for Oseltamivir interaction with H1N1 and van der Waals interaction is more favourable for H5N1, whereas Zanamivir favours the electrostatic interaction in both the strains (H1N1 and H5N1). Energy-optimised pharmacophore mapping was created using Oseltamivir drug. The pharmacophore model contained two hydrogen-bond acceptor and two hydrogen bond donor sites. Using these pharmacophore features, we screened a compound database to find a potential ligand that could inhibit the H1N1 protein. The current research will pave the way to find potent neuraminidase inhibitors against H1N1 (2009) strain.

Sighting of tankyrase inhibitors by structure- and ligand-based screening and in vitro approach.

Tankyrase 1 and 2 (TNKS) are promising and attractive therapeutic targets in anticancer drug development. Herein, we report the findings of structure- and ligand-based virtual screening for novel TNKS1 inhibitors using iterative rounds of in silico studies and subsequent biological evaluation methods. Upon screening of three compound databases, a final set of five molecules were selected for experimental validation. In order to prove our in silico findings, tankyrase activity was assessed by a calorimetric assay with the five identified lead molecules. Out of five, only C1 (7309981) showed significant inhibition of TNKS1 enzyme. Furthermore, the toxicity of the selected 5 compounds was measured using cytotoxicity experiments and inhibition of cell growth, and it was more pronounced in C1, followed by C5 and C3 (7309981 > 7245236 > 7275738). The morphological assessment, DNA damage and chromatin condensation and fragmentation results also confirmed that C1 has enhanced activity against MCF-7 cells.

Understanding the evolutionary relationship of hemagglutinin protein from influenza viruses using phylogenetic and molecular modeling studies.

The existing H1N1 (2009) swine flu is pandemic in nature and is responsible for global economic losses and fatalities. Among the eight gene segments of H1N1, hemagglutinin (HA) plays a major role in the attachment of the virus to the host cell surface and entry of viral RNA into the host cell leads to infection. In this study, sequence and phylogenetic analysis of the H1N1 (2009) HA, from Mexico City along with 1952 sequences, from different subtypes of pandemic influenza A virus were studied and results showed that the closest relationship of H1N1 (2009) Mexico strain was with the H1N1 (2007) Mallard Norway strain. Analysis of secondary structures predicted from the protein sequence revealed that diminishing of alpha helixes was observed in many areas of the sequences between the years 2005 to 2010. Conversely, analysis at the structural level is necessary to critically assess the functional significance. Structural level investigation was therefore done for the above said proteins by constructing the 3D structure of these proteins through homology modeling. The models were validated and structural level similarities were evaluated through superimposition. Subsequently, docking studies were done to find the binding mode of the sialic acid (SA) with influenza HA. Molecular dynamics simulations were executed to study the interactions of SA molecule with the HA. Energetic analysis reveals that van der Waal interaction is more favorable for binding of HA with SA of the whole influenza virus. Binding pocket analysis shows that intensities of H-bond donor and acceptor are more in H1N1 (2009).

Molecular insights on TNKS1/TNKS2 and inhibitor-IWR1 interactions.

Tankyrases (TNKS) belong to the poly(ADP-ribose)polymerase (PARP) protein super family and play a vital role in the Wnt/ß-catenin signaling pathway. TNKS is a potential target for therapeutic intervention against various cancers, heritable diseases (e.g. cherubism) and implications in the replication of herpes simplex virus (HSV). The recent discovery of the structure of TNKS with an IWR1 inhibitor has provided insight into the binding modes which are specific for the TNKS protein which will aid in the development of drugs that are specific for the TNKS protein. The current study investigates molecular interactions between the induced pocket of TNKS1 and TNKS2 with an IWR1 compound using computational approaches. Molecular docking analysis of IWR1 at the induced pocket of TNKS1 and TNKS2 was performed. The resulting protein-ligand complexes were simulated for a timescale of 100 ns. Results revealed the stable binding of IWR1 at the induced pocket of TNKS1 and TNKS2 proteins. Apart from active site amino acids, π-π stack paring interactions were also crucial for the protein-ligand binding and stability of the complex. Further, energy-optimized pharmacophore mapping was performed and the resulting pharmacophore model contained a four (TNKS1-IWR1) and five (TNKS2-IWR1) featured sites. Based on the pharmacophore models, the best inhibitors were screened from the ZINC natural product compound database and these could be used as potential drugs against TNKS1 and TNKS2.

In silico structural and functional analysis of the human TOPK protein by structure modeling and molecular dynamics studies.

Over expression of T-lymphokine-activated killer cell-originated protein kinase (TOPK) has been associated with leukemia, myeloma tumors and various other cancers. The function and regulatory mechanism of TOPK in tumor cells remains unclear. Structural studies that could reveal the regulatory mechanism have been a challenge because of the unavailabity of TOPK's crystal structure. Hence, in this study, the 3D structure of TOPK protein has been constructed by using multiple templates. The quality and reliability of the generated model was checked and the molecular dynamics method was utilized to refine the model. APBS method was employed to know the electrostatic potential surface of the modeled protein and it was found that the optimum pH for protein stability is 3.4 which will further help in mechanistic hypothesis of TOPK protein. Active site of TOPK was identified from available literature and HTVS was employed to identify the lead molecules. The expected binding modes of protein-ligand complexes were reproduced in the MD simulation which indicates that the complex is relatively stable. The pharmacokinetic properties of the lead molecules are also under acceptable range. TOPK act as a substrate for CDK1 and the protein-protein docking and dynamics studies were carried out to analyze the effect of Thr9Ala mutation of TOPK in the two protein complex formation. It shows that the wild type complex is more stable when compared with the mutant type. Such structural information at atomic level not only exhibits the action modes of TOPK inhibitors but also furnishes a novel starting point for structure based drug design of TOPK inhibitors.

Homology modeling, molecular dynamics, e-pharmacophore mapping and docking study of Chikungunya virus nsP2 protease.

To date, no suitable vaccine or specific antiviral drug is available to treat Chikungunya viral (CHIKV) fever. Hence, it is essential to identify drug candidates that could potentially impede CHIKV infection. Here, we present the development of a homology model of nsP2 protein based on the crystal structure of the nsP2 protein of Venezuelan equine encephalitis virus (VEEV). The protein modeled was optimized using molecular dynamics simulation; the junction peptides of a nonstructural protein complex were then docked in order to investigate the possible protein-protein interactions between nsP2 and the proteins cleaved by nsP2. The modeling studies conducted shed light on the binding modes, and the critical interactions with the peptides provide insight into the chemical features needed to inhibit the CHIK virus infection. Energy-optimized pharmacophore mapping was performed using the junction peptides. Based on the results, we propose the pharmacophore features that must be present in an inhibitor of nsP2 protease. The resulting pharmacophore model contained an aromatic ring, a hydrophobic and three hydrogen-bond donor sites. Using these pharmacophore features, we screened a large public library of compounds (Asinex, Maybridge, TOSLab, Binding Database) to find a potential ligand that could inhibit the nsP2 protein. The compounds that yielded a fitness score of more than 1.0 were further subjected to Glide HTVS and Glide XP. Here, we report the best four compounds based on their docking scores; these compounds have IDs of 27943, 21362, ASN 01107557 and ASN 01541696. We propose that these compounds could bind to the active site of nsP2 protease and inhibit this enzyme. Furthermore, the backbone structural scaffolds of these four lead compounds could serve as building blocks when designing drug-like molecules for the treatment of Chikungunya viral fever.

Pharmacophore filtering and 3D-QSAR in the discovery of new JAK2 inhibitors.

Janus kinase 2 (JAK2) plays a crucial role in the patho-mechanism of cardiovascular pathologies, myeloproliferative disorders and many other diseases. Thus, effective JAK2 kinase inhibitors may be of significant therapeutic importance. In this study, a pharmacophore mapping studies were undertaken for a series of phenylaminopyrimidines derivatives. A five point pharmacophore with two hydrogen bond donors (D), two hydrogen bond acceptors (A) and one aromatic ring (R) as pharmacophoric features were developed. The pharmacophore hypothesis yielded a statistically significant 3D-QSAR model, with a correlation coefficient of R²=0.970 for training set compounds. The model generated showed excellent predictive power, with a correlation coefficient of Q²=0.822. The external validation indicated that our QSAR models possessed high predictive powers with r²(0) value of 0.999 and r²(m) value of 0.637 respectively. The model was then employed as 3D search query to screen against public compound libraries (Asinex, TOSLab, Maybride and Binding database) in-order to identify a new scaffold. We have identified thirteen distinct drug-like molecules binding to the JAK2. Interestingly, some of the compounds show activity against JAK2 by PASS biological activity prediction. Hence, these molecules could be potential selective inhibitors of JAK2 that can be experimentally validated and their backbone structural scaffold could serve as building blocks in designing drug-like molecules for JAK2.

In silico studies on marine actinomycetes as potential inhibitors for Glioblastoma multiforme.

Glioblastoma multiforme (GBM) is considered to be the most common and often deadly disorder which affects the brain. It is caused by the over expression of proteins such as ephrin type-A receptor 2 (EphA2), epidermal growth factor receptor (EGFR) and EGFRvIII. These 3 proteins are considered to be the potential therapeutic targets for GBM. Among these, EphA2 is reported to be over-expressed in ˜90% of GBM. Herein we selected 35 compounds from marine actinomycetes, 5 in vitro and in vivo studied drug candidates and 4 commercially available drugs for GBM which were identified from literature and analysed by using comparative docking studies. Based on the glide scores and other in silico parameters available in Schrödinger, two selected marine actinomycetes compounds which include Tetracenomycin D and Chartreusin exhibited better binding energy among all the compounds studied in comparative docking. In this study we have demonstrated the inhibition of the 3 selected targets by the two bioactive compounds from marine actinomycetes through in-silico docking studies. Furthermore molecular dynamics simulation were also been performed to check the stability and the amino acids interacted with the 3 molecular targets (EphA2 receptor, EGFR, EGFRvIII) for GBM. Our results suggest that Tetracinomycin D and Chartreusin are the novel and potential inhibitor for the treatment of GBM.