Guiding lead optimization with GPCR structure modeling and molecular dynamics

Bartender: Martini 3 Bonded Terms via Quantum Mechanics-Based Molecular Dynamics

Coarse-grained (CG) molecular dynamics (MD) simulations have grown in applicability over the years. The recently released version of the Martini CG force field (Martini 3) has been successfully applied to simulate many processes, including protein-ligand binding. However, the current ligand parametrization scheme is manual and requires an a priori reference all-atom (AA) simulation for benchmarking. For systems with suboptimal AA parameters, which are often unknown, this translates into a CG model that does not reproduce the true dynamical behavior of the underlying molecule. Here, we present Bartender, a quantum mechanics (QM)/MD-based parametrization tool written in Go. Bartender harnesses the power of QM simulations and produces reasonable bonded terms for Martini 3 CG models of small molecules in an efficient and user-friendly manner. For small, ring-like molecules, Bartender generates models whose properties are indistinguishable from the human-made models. For more complex, drug-like ligands, it is able to fit functional forms beyond simple harmonic dihedrals and thus better captures their dynamical behavior. Bartender has the power to both increase the efficiency and the accuracy of Martini 3-based high-throughput applications by producing numerically stable and physically realistic CG models.

AI-based identification of therapeutic agents targeting GPCRs: introducing ligand type classifiers and systems biology

Identifying ligands targeting G protein coupled receptors (GPCRs) with novel chemotypes other than the physiological ligands is a challenge for in silico screening campaigns. Here we present an approach that identifies novel chemotype ligands by combining structural data with a random forest agonist/antagonist classifier and a signal-transduction kinetic model. As a test case, we apply this approach to identify novel antagonists of the human adenosine transmembrane receptor type 2A, an attractive target against Parkinson's disease and cancer. The identified antagonists were tested here in a radio ligand binding assay. Among those, we found a promising ligand whose chemotype differs significantly from all so-far reported antagonists, with a binding affinity of 310 ± 23.4 nM. Thus, our protocol emerges as a powerful approach to identify promising ligand candidates with novel chemotypes while preserving antagonistic potential and affinity in the nanomolar range.

L-DOPA and Droxidopa: From Force Field Development to Molecular Docking into Human β2-Adrenergic Receptor

The increasing interest in the molecular mechanism of the binding of different agonists and antagonists to β₂-adrenergic receptor (β₂AR) inactive and active states has led us to investigate protein–ligand interactions using molecular docking calculations. To perform this study, the 3.2 Å X-ray crystal structure of the active conformation of human β₂AR in the complex with the endogenous agonist adrenaline has been used as a template for investigating the binding of two exogenous catecholamines to this adrenergic receptor. Here, we show the derivation of L-DOPA and Droxidopa OPLS all atom (AA) force field (FF) parameters via quantum mechanical (QM) calculations, molecular dynamics (MD) simulations in aqueous solutions of the two catecholamines and the molecular docking of both ligands into rigid and flexible β₂AR models. We observe that both ligands share with adrenaline similar experimentally observed binding anchor sites, which are constituted by Asp113/Asn312 and Ser203/Ser204/Ser207 side chains. Moreover, both L-DOPA and Droxidopa molecules exhibit binding affinities comparable to that predicted for adrenaline, which is in good agreement with previous experimental and computational results. L-DOPA and Droxidopa OPLS AA FFs have also been tested by performing MD simulations of these ligands docked into β₂AR proteins embedded in lipid membranes. Both hydrogen bonds and hydrophobic interaction networks observed over the 1 μs MD simulation are comparable with those derived from molecular docking calculations and MD simulations performed with the CHARMM FF.

Identification of Highly Selective Orexin 1 Receptor Antagonists Driven by Structure-Based Design

OX1 receptor antagonists are of interest to treat, for example, substance abuse disorders, personality disorders, eating disorders, or anxiety-related disorders. However, known dual OX1/OX2 receptor antagonists are not suitable due to their sleep-inducing effects; therefore, we were interested in identifying a highly OX1 selective antagonist with a sufficient window to OX2-mediated effects. Herein, we describe the design of highly selective OX1 receptor antagonists driven by the X-ray structure of OX1 with suvorexant, a dual OX1/OX2 receptor antagonist. Moderately selective OX1 antagonists comprising a [2.2.1]-bicyclic scaffold served as our starting point. Based on our binding mode hypothesis, we postulated which part of the scaffold points toward one of the regions where the two binding pockets differ. Structural changes in this part resulted in a modified core with higher inherent selectivity compared to the [2.2.1]-bicyclic template. The structure-based design, synthesis, and hit-to-lead evaluation of this novel OX1 receptor-selective scaffold are discussed herein.

Analysis of L-DOPA and droxidopa binding to human β2-adrenergic receptor

Over the last two decades, an increasing number of studies has been devoted to a deeper understanding of the molecular process involved in the binding of various agonists and antagonists to active and inactive conformations of β2-adrenergic receptor (β2AR). The 3.2 Å x-ray crystal structure of human β2AR active state in combination with the endogenous low affinity agonist adrenaline offers an ideal starting structure for studying the binding of various catecholamines to adrenergic receptors. We show that molecular docking of levodopa (L-DOPA) and droxidopa into rigid and flexible β2AR models leads for both ligands to binding anchor sites comparable to those experimentally reported for adrenaline, namely D113/N312 and S203/S204/S207 side chains. Both ligands have a hydrogen bond network that is extremely similar to those of noradrenaline and dopamine. Interestingly, redocking neutral and protonated versions of adrenaline to rigid and flexible β2AR models results in binding poses that are more energetically stable and distinct from the x-ray crystal structure. Similarly, lowest energy conformations of noradrenaline and dopamine generated by docking into flexible β2AR models had binding free energies lower than those of best poses in rigid receptor models. Furthermore, our findings show that L-DOPA and droxidopa molecules have binding affinities comparable to those predicted for adrenaline, noradrenaline, and dopamine, which are consistent with previous experimental and computational findings and supported by the molecular dynamics simulations of β2AR-ligand complexes performed here.

Ligand Affinities within the Open-Boundary Molecular Mechanics/Coarse-Grained Framework (I): Alchemical Transformations within the Hamiltonian Adaptive Resolution Scheme

Our recently developed Open-Boundary Molecular Mechanics/Coarse Grained (OB-MM/CG) framework predicts ligand poses in important pharmaceutical targets, such as G-protein Coupled Receptors, even when experimental structural information is lacking. The approach, which is based on GROMOS and AMBER force fields, allows for grand-canonical simulations of protein-ligand complexes by using the Hamiltonian Adaptive Resolution Scheme (H-AdResS) for the solvent. Here, we present a key step toward the estimation of ligand binding affinities for their targets within this approach. This is the implementation of the H-AdResS in the GROMACS code. The accuracy of our implementation is established by calculating hydration free energies of several molecules in water by means of alchemical transformations. The deviations of the GROMOS- and AMBER-based H-AdResS results from the reference fully atomistic simulations are smaller than the accuracy of the force field and/or they are in the range of the published results. Importantly, our predictions are in good agreement with experimental data. The current implementation paves the way to the use of the OB-MM/CG framework for the study of large biological systems.

Hybrid MM/CG Webserver: Automatic Set Up of Molecular Mechanics/Coarse-Grained Simulations for Human G Protein-Coupled Receptor/Ligand Complexes

Hybrid Molecular Mechanics/Coarse-Grained (MM/CG) simulations help predict ligand poses in human G protein-coupled receptors (hGPCRs), the most important protein superfamily for pharmacological applications. This approach allows the description of the ligand, the binding cavity, and the surrounding water molecules at atomistic resolution, while coarse-graining the rest of the receptor. Here, we present the Hybrid MM/CG Webserver (mmcg.grs.kfa-juelich.de) that automatizes and speeds up the MM/CG simulation setup of hGPCR/ligand complexes. Initial structures for such complexes can be easily and efficiently generated with other webservers. The Hybrid MM/CG server also allows for equilibration of the systems, either fully automatically or interactively. The results are visualized online (using both interactive 3D visualizations and analysis plots), helping the user identify possible issues and modify the setup parameters accordingly. Furthermore, the prepared system can be downloaded and the simulation continued locally.

Ligand Pose Predictions for Human G Protein-Coupled Receptors: Insights from the Amber-Based Hybrid Molecular Mechanics/Coarse-Grained Approach

Human G protein-coupled receptors (hGPCRs) are the most frequent targets of Food and Drug Administration (FDA)-approved drugs. Structural bioinformatics, along with molecular simulation, can support structure-based drug design targeting hGPCRs. In this context, several years ago, we developed a hybrid molecular mechanics (MM)/coarse-grained (CG) approach to predict ligand poses in low-resolution hGPCR models. The approach was based on the GROMOS96 43A1 and PRODRG united-atom force fields for the MM part. Here, we present a new MM/CG implementation using, instead, the Amber 14SB and GAFF all-atom potentials for proteins and ligands, respectively. The new implementation outperforms the previous one, as shown by a variety of applications on models of hGPCR/ligand complexes at different resolutions, and it is also more user-friendly. Thus, it emerges as a useful tool to predict poses in low-resolution models and provides insights into ligand binding similarly to all-atom molecular dynamics, albeit at a lower computational cost.

Multiscale Molecular Modeling in G Protein-Coupled Receptor (GPCR)-Ligand Studies

G protein-coupled receptors (GPCRs) are major drug targets due to their ability to facilitate signal transduction across cell membranes, a process that is vital for many physiological functions to occur. The development of computational technology provides modern tools that permit accurate studies of the structures and properties of large chemical systems, such as enzymes and GPCRs, at the molecular level. The advent of multiscale molecular modeling permits the implementation of multiple levels of theories on a system of interest, for instance, assigning chemically relevant regions to high quantum mechanics (QM) level of theory while treating the rest of the system using classical force field (molecular mechanics (MM) potential). Multiscale QM/MM molecular modeling have far-reaching applications in the rational design of GPCR drugs/ligands by affording precise ligand binding configurations through the consideration of conformational plasticity. This enables the identification of key binding site residues that could be targeted to manipulate GPCR function. This review will focus on recent applications of multiscale QM/MM molecular simulations in GPCR studies that could boost the efficiency of future structure-based drug design (SBDD) strategies.

Fragment-based screening identifies molecules targeting the substrate-binding ankyrin repeat domains of tankyrase

The PARP enzyme and scaffolding protein tankyrase (TNKS, TNKS2) uses its ankyrin repeat clusters (ARCs) to bind a wide range of proteins and thereby controls diverse cellular functions. A number of these are implicated in cancer-relevant processes, including Wnt/β-catenin signalling, Hippo signalling and telomere maintenance. The ARCs recognise a conserved tankyrase-binding peptide motif (TBM). All currently available tankyrase inhibitors target the catalytic domain and inhibit tankyrase's poly(ADP-ribosyl)ation function. However, there is emerging evidence that catalysis-independent "scaffolding" mechanisms contribute to tankyrase function. Here we report a fragment-based screening programme against tankyrase ARC domains, using a combination of biophysical assays, including differential scanning fluorimetry (DSF) and nuclear magnetic resonance (NMR) spectroscopy. We identify fragment molecules that will serve as starting points for the development of tankyrase substrate binding antagonists. Such compounds will enable probing the scaffolding functions of tankyrase, and may, in the future, provide potential alternative therapeutic approaches to inhibiting tankyrase activity in cancer and other conditions.

Time-Domain Analysis of Molecular Dynamics Trajectories Using Deep Neural Networks: Application to Activity Ranking of Tankyrase Inhibitors

Molecular dynamics simulations provide valuable insights into the behavior of molecular systems. Extending the recent trend of using machine learning techniques to predict physicochemical properties from molecular dynamics data, we propose to consider the trajectories as multidimensional time series represented by 2D tensors containing the ligand-protein interaction descriptor values for each time step. Similar in structure to the time series encountered in modern approaches for signal, speech, and natural language processing, these time series can be directly analyzed using long short-term memory (LSTM) recurrent neural networks or convolutional neural networks (CNNs). The predictive regression models for the ligand-protein affinity were built for a subset of the PDBbind v.2017 database and applied to inhibitors of tankyrase, an enzyme of the poly(ADP-ribose)-polymerase (PARP) family that can be used in the treatment of colorectal cancer. As an additional test set, a subset of the Community Structure-Activity Resource (CSAR) data set was used. For comparison, the random forest and simple neural network models based on the crystal pose or the trajectory-averaged descriptors were used, as well as the commonly employed docking and molecular mechanics Poisson-Boltzmann surface area (MM-PBSA) scores. Convolutional neural networks based on the 2D tensors of ligand-protein interaction descriptors for short (2 ns) trajectories provide the best accuracy and predictive power, reaching the Spearman rank correlation coefficient of 0.73 and Pearson correlation coefficient of 0.70 for the tankyrase test set. Taking into account the recent increase in computational power of modern GPUs and relatively low computational complexity of the proposed approach, it can be used as an advanced virtual screening filter for compound prioritization.

Molecular Dynamics Simulations of Adenosine Receptors: Advances, Applications and Trends

:

Adenosine receptors (ARs) are transmembrane proteins that belong to the G protein-coupled receptors (GPCRs) superfamily and mediate the biological functions of adenosine. To date, four AR subtypes are known, namely A1, A2A, A2B and A3 that exhibit different signaling pathways, tissue localization, and mechanisms of activation. Moreover, the widespread ARs and their implication in numerous physiological and pathophysiological conditions had made them pivotal therapeutic targets for developing clinically effective agents.

:

The crystallographic success in identifying the 3D crystal structures of A2A and A1 ARs has dramatically enriched our understanding of their structural and functional properties such as ligand binding and signal transduction. This, in turn, has provided a structural basis for a larger contribution of computational methods, particularly molecular dynamics (MD) simulations, toward further investigation of their molecular properties and designing bioactive ligands with therapeutic potential. MD simulation has been proved to be an invaluable tool in investigating ARs and providing answers to some critical questions. For example, MD has been applied in studying ARs in terms of ligand-receptor interactions, molecular recognition, allosteric modulations, dimerization, and mechanisms of activation, collectively aiding in the design of subtype selective ligands.

:

In this review, we focused on the advances and different applications of MD simulations utilized to study the structural and functional aspects of ARs that can foster the structure-based design of drug candidates. In addition, relevant literature was briefly discussed which establishes a starting point for future advances in the field of drug discovery to this pivotal group of drug targets.

A Comparative Study of Modern Homology Modeling Algorithms for Rhodopsin Structure Prediction

Rhodopsins are seven α-helical membrane proteins that are of great importance in chemistry, biology, and modern biotechnology. Any in silico study on rhodopsin properties and functioning requires a high-quality three-dimensional structure. Due to particular difficulties with obtaining membrane protein structures from the experiment, in silico prediction of the three-dimensional rhodopsin structure based only on its primary sequence is an especially important task. For the last few years, significant progress was made in the field of protein structure prediction, especially for methods based on comparative modeling. However, the majority of this progress was made for soluble proteins and further investigations are needed to achieve similar progress for membrane proteins. In this paper, we evaluate the performance of modern protein structure prediction methodologies (implemented in the Medeller, I-TASSER, and Rosetta packages) for their ability to predict rhodopsin structures. Three widely used methodologies were considered: two general methodologies that are commonly applied to soluble proteins and a methodology that uses constraints that are specific for membrane proteins. The test pool consisted of 36 target-template pairs with different sequence similarities that was constructed on the basis of 24 experimental rhodopsin structures taken from the RCSB database. As a result, we showed that all three considered methodologies allow obtaining rhodopsin structures with the quality that is close to the crystallographic one (root mean square deviation (RMSD) of the predicted structure from the corresponding X-ray structure up to 1.5 Å) if the target-template sequence identity is higher than 40%. Moreover, all considered methodologies provided structures of average quality (RMSD < 4.0 Å) if the target-template sequence identity is higher than 20%. Such structures can be subsequently used for further investigation of molecular mechanisms of protein functioning and for the development of modern protein-based biotechnologies.

Emerging themes in heterotrimeric G-protein signaling in plants

Heterotrimeric G-proteins are key signaling components involved during the regulation of a multitude of growth and developmental pathways in all eukaryotes. Although the core proteins (Gα, Gβ, Gγ subunits) and their basic biochemistries are conserved between plants and non-plant systems, seemingly different inherent properties of specific components, altered wirings of G-protein network architectures, and the presence of novel receptors and effector proteins make plant G-protein signaling mechanisms somewhat distinct from the well-established animal paradigm. G-protein research in plants is getting a lot of attention recently due to the emerging roles of these proteins in controlling many agronomically important traits. New findings on both canonical and novel G-protein components and their conserved and unique signaling mechanisms are expected to improve our understanding of this important module in affecting critical plant growth and development pathways and eventually their utilization to produce plants for the future needs. In this review, we briefly summarize what is currently known in plant G-protein research, describe new findings and how they are changing our perceptions of the field, and discuss important issues that still need to be addressed.

Intracellular Transfer of Na+ in an Active-State G-Protein-Coupled Receptor

Playing a central role in cell signaling, G-protein-coupled receptors (GPCRs) are the largest superfamily of membrane proteins and form the majority of drug targets in humans. How extracellular agonist binding triggers the activation of GPCRs and associated intracellular effector proteins remains, however, poorly understood. Structural studies have revealed that inactive class A GPCRs harbor a conserved binding site for Na+ ions in the center of their transmembrane domain, accessible from the extracellular space. Here, we show that the opening of a conserved hydrated channel in the activated state receptors allows the Na+ ion to egress from its binding site into the cytosol. Coupled with protonation changes, this ion movement occurs without significant energy barriers, and can be driven by physiological transmembrane ion and voltage gradients. We propose that Na+ ion exchange with the cytosol is a key step in GPCR activation. Further, we hypothesize that this transition locks receptors in long-lived active-state conformations.

Computational Methods Used in Hit-to-Lead and Lead Optimization Stages of Structure-Based Drug Discovery

GPCR modeling approaches are widely used in the hit-to-lead (H2L) and lead optimization (LO) stages of drug discovery. The aims of these modeling approaches are to predict the 3D structures of the receptor-ligand complexes, to explore the key interactions between the receptor and the ligand and to utilize these insights in the design of new molecules with improved binding, selectivity or other pharmacological properties. In this book chapter, we present a brief survey of key computational approaches integrated with hierarchical GPCR modeling protocol (HGMP) used in hit-to-lead (H2L) and in lead optimization (LO) stages of structure-based drug discovery (SBDD). We outline the differences in modeling strategies used in H2L and LO of SBDD and illustrate how these tools have been applied in three drug discovery projects.

GPCR Homology Model Generation for Lead Optimization

The vast increase of recently solved GPCR X-ray structures forms the basis for GPCR homology modeling to atomistic accuracy. Nowadays, homology models can be employed for GPCR-ligand optimization and have been reported as invaluable tools for drug design in the last few years. Elucidation of the complex GPCR pharmacology and the associated GPCR conformations made clear that different homology models have to be constructed for different activation states of the GPCRs. Therefore, templates have to be chosen accordingly to their sequence homology as well as to their activation state. The subsequent ligand placement is nontrivial, as some recent X-ray structures show very unusual ligand binding sites and solvent involvement, expanding the space of the putative ligand binding site from the generic retinal binding pocket to the whole receptor. In the present study, a workflow is presented starting from the selection of the target sequence, guiding through the GPCR modeling process, and finishing with ligand placement and pose validation.

Dynamic roles for the N-terminus of the yeast G protein-coupled receptor Ste2p

The Saccharomyces cerevisiae α-factor receptor Ste2p has been used extensively as a model to understand the molecular mechanism of signal transduction by G protein-coupled receptors (GPCRs). Single and double cysteine mutants of Ste2p were created and served as surrogates to detect intramolecular interactions and dimerization of Ste2p using disulfide cross-linking methodology. When a mutation was introduced into the phylogenetically conserved tyrosine residue at position 26 (Y26C) in the N-terminus of Ste2p, dimerization was increased greatly. The amount of dimer formed by this Y26C mutant was greatly reduced by ligand binding even though the ligand binding site is far removed from the N-terminus; the lowering of the dimer formation was consistent with a conformational change in the N-terminus of the receptor upon activation. Dimerization was decreased by double mutations Y26C/V109C or Y26C/T114C indicating that Y26 is in close proximity to V109 and T114 of extracellular loop 1 in native Ste2p. Combined with earlier studies, these results indicate previously unrecognized roles for the N-terminus of Ste2p, and perhaps of GPCRs in general, and reveal a specific N-terminus residue or region, that is involved in GPCR signaling, intrareceptor interactions, and receptor dimerization.

Rapid and accurate assessment of GPCR-ligand interactions Using the fragment molecular orbital-based density-functional tight-binding method

The reliable and precise evaluation of receptor–ligand interactions and pair‐interaction energy is an essential element of rational drug design. While quantum mechanical (QM) methods have been a promising means by which to achieve this, traditional QM is not applicable for large biological systems due to its high computational cost. Here, the fragment molecular orbital (FMO) method has been used to accelerate QM calculations, and by combining FMO with the density‐functional tight‐binding (DFTB) method we are able to decrease computational cost 1000 times, achieving results in seconds, instead of hours. We have applied FMO‐DFTB to three different GPCR–ligand systems. Our results correlate well with site directed mutagenesis data and findings presented in the published literature, demonstrating that FMO‐DFTB is a rapid and accurate means of GPCR–ligand interactions. © 2017 Authors. Journal of Computational Chemistry Published by Wiley Periodicals, Inc.

Cell-impedance-based label-free technology for the identification of new drugs

INTRODUCTION

Drug discovery has progressed from relatively simple binding or activity screening assays to high-throughput screening of sophisticated compound libraries with emphasis on miniaturization and automation. The development of functional assays has enhanced the success rate in discovering novel drug molecules. Many technologies, originally based on radioactive labeling, have sequentially been replaced by methods based on fluorescence labeling. Recently, the focus has switched to label-free technologies in cell-based screening assays. Areas covered: Label-free, cell-impedance-based methods comprise of different technologies including surface plasmon resonance, mass spectrometry and biosensors applied for screening of anticancer drugs, G protein-coupled receptors, receptor tyrosine kinase and virus inhibitors, drug and nanoparticle cytotoxicity. Many of the developed methods have been used for high-throughput screening in cell lines. Cell viability and morphological damage prediction have been monitored in three-dimensional spheroid human HT-29 carcinoma cells and whole Schistosomula larvae. Expert opinion: Progress in label-free, cell-impedance-based technologies has facilitated drug screening and may enhance the discovery of potential novel drug molecules through, and improve target molecule identification in, alternative signal pathways. The variety of technologies to measure cellular responses through label-free cell-impedance based approaches all support future drug development and should provide excellent assets for finding better medicines.

Molecular Modelling Approaches for the Analysis of Histamine Receptors and Their Interaction with Ligands

Several experimental techniques to analyse histamine receptors are available, e.g. pharmacological characterisation of known or new compounds by different types of assays or mutagenesis studies. To obtain insights into the histamine receptors on a molecular and structural level, crystal structures have to be determined and molecular modelling studies have to be performed. It is widely accepted to generate homology models of the receptor of interest based on an appropriate crystal structure as a template and to refine the resulting models by molecular dynamic simulations. A lot of modelling techniques, e.g. docking, QSAR or interaction fingerprint methods, are used to predict binding modes of ligands and pharmacological data, e.g. affinity or even efficacy. However, within the last years, molecular dynamic simulations got more and more important: First of all, molecular dynamic simulations are very helpful to refine the binding mode of a ligand to a histamine receptor, obtained by docking studies. Furthermore, with increasing computational performance it got possible to simulate complete binding pathways of ions or ligands from the aqueous extracellular phase into the allosteric or orthosteric binding pocket of histamine receptors.