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

A practical guide to machine-learning scoring for structure-based virtual screening

Structure-based virtual screening (SBVS) via docking has been used to discover active molecules for a range of therapeutic targets. Chemical and protein data sets that contain integrated bioactivity information have increased both in number and in size. Artificial intelligence and, more concretely, its machine-learning (ML) branch, including deep learning, have effectively exploited these data sets to build scoring functions (SFs) for SBVS against targets with an atomic-resolution 3D model (e.g., generated by X-ray crystallography or predicted by AlphaFold2). Often outperforming their generic and non-ML counterparts, target-specific ML-based SFs represent the state of the art for SBVS. Here, we present a comprehensive and user-friendly protocol to build and rigorously evaluate these new SFs for SBVS. This protocol is organized into four sections: (i) using a public benchmark of a given target to evaluate an existing generic SF; (ii) preparing experimental data for a target from public repositories; (iii) partitioning data into a training set and a test set for subsequent target-specific ML modeling; and (iv) generating and evaluating target-specific ML SFs by using the prepared training-test partitions. All necessary code and input/output data related to three example targets (acetylcholinesterase, HMG-CoA reductase, and peroxisome proliferator-activated receptor-α) are available at https://github.com/vktrannguyen/MLSF-protocol , can be run by using a single computer within 1 week and make use of easily accessible software/programs (e.g., Smina, CNN-Score, RF-Score-VS and DeepCoy) and web resources. Our aim is to provide practical guidance on how to augment training data to enhance SBVS performance, how to identify the most suitable supervised learning algorithm for a data set, and how to build an SF with the highest likelihood of discovering target-active molecules within a given compound library.

Challenges in natural product-based drug discovery assisted with in silico-based methods

The application of traditional medicine by humans for the treatment of ailments as well as improving the quality of life far outdates recorded history. To date, a significant percentage of humans, especially those living in developing/underprivileged communities still rely on traditional medicine for primary healthcare needs. In silico-based methods have been shown to play a pivotal role in modern pharmaceutical drug discovery processes. The application of these methods in identifying natural product (NP)-based hits has been successful. This is very much observed in many research set-ups that use rationally in silico-based methods in combination with experimental validation techniques. The combination has rendered the use of in silico-based approaches even more popular and successful in the investigation of NPs. However, identifying and proposing novel NP-based hits for experimental validation comes with several challenges such as the availability of compounds by suppliers, the huge task of separating pure compounds from complex mixtures, the quantity of samples available from the natural source to be tested, not to mention the potential ecological impact if the natural source is exhausted. Because most peer-reviewed publications are biased towards "positive results", these challenges are generally not discussed in publications. In this review, we highlight and discuss these challenges. The idea is to give interested scientists in this field of research an idea of what they can come across or should be expecting as well as prompting them on how to avoid or fix these issues.

EmbedDTI: Enhancing the Molecular Representations via Sequence Embedding and Graph Convolutional Network for the Prediction of Drug-Target Interaction

The identification of drug-target interaction (DTI) plays a key role in drug discovery and development. Benefitting from large-scale drug databases and verified DTI relationships, a lot of machine-learning methods have been developed to predict DTIs. However, due to the difficulty in extracting useful information from molecules, the performance of these methods is limited by the representation of drugs and target proteins. This study proposes a new model called EmbedDTI to enhance the representation of both drugs and target proteins, and improve the performance of DTI prediction. For protein sequences, we leverage language modeling for pretraining the feature embeddings of amino acids and feed them to a convolutional neural network model for further representation learning. For drugs, we build two levels of graphs to represent compound structural information, namely the atom graph and substructure graph, and adopt graph convolutional network with an attention module to learn the embedding vectors for the graphs. We compare EmbedDTI with the existing DTI predictors on two benchmark datasets. The experimental results show that EmbedDTI outperforms the state-of-the-art models, and the attention module can identify the components crucial for DTIs in compounds.

FEPrepare: A Web-Based Tool for Automating the Setup of Relative Binding Free Energy Calculations

Relative binding free energy calculations in drug design are becoming a useful tool in facilitating lead binding affinity optimization in a cost- and time-efficient manner. However, they have been limited by technical challenges such as the manual creation of large numbers of input files to set up, run, and analyze free energy simulations. In this Application Note, we describe FEPrepare, a novel web-based tool, which automates the setup procedure for relative binding FEP calculations for the dual-topology scheme of NAMD, one of the major MD engines, using OPLS-AA force field topology and parameter files. FEPrepare provides the user with all necessary files needed to run a FEP/MD simulation with NAMD. FEPrepare can be accessed and used at https://feprepare.vi-seem.eu/.

Unraveling the unbinding pathways of SARS-CoV-2 Papain-like proteinase known inhibitors by Supervised Molecular Dynamics simulation

The COVID-19 disease has infected and killed countless people all over the world since its emergence at the end of 2019. No specific therapy for COVID-19 is not currently available, and urgent treatment solutions are needed. Recent studies have found several potential molecular targets, and one of the most critical proteins of the SARS-CoV-2 virus work machine is the Papain-like protease (Plpro). Potential inhibitors are available, and their X-ray crystallographic structures in complex with this enzyme have been determined recently. However, their activities against this enzyme are insufficient and need to be characterized and improved to be of clinical values. Therefore, in this work, by utilizing the Supervised Molecular Dynamics (SuMD) simulation method, we achieved multiple unbinding events of Plpro inhibitors, GRL0617, and its derivates, and captured and understood the details of the unbinding pathway. We found that residues of the BL2 loop, such as Tyr268 and Gln269, play major roles in the unbinding pathways, but the most important contributing factor is the natural movements and behavior of the BL2 loop, which can control the entire process. We believe that the details found in this study can be used to refine and optimize potential inhibitors like GRL0617 and design more efficacious inhibitors as a treatment for the SARS-CoV-2 virus.

Intermolecular interaction among Remdesivir, RNA and RNA-dependent RNA polymerase of SARS-CoV-2 analyzed by fragment molecular orbital calculation

COVID-19, a disease caused by a new strain of coronavirus (SARS-CoV-2) originating from Wuhan, China, has now spread around the world, triggering a global pandemic, leaving the public eagerly awaiting the development of a specific medicine and vaccine. In response, aggressive efforts are underway around the world to overcome COVID-19. In this study, referencing the data published on the Protein Data Bank (PDB ID: 7BV2) on April 22, we conducted a detailed analysis of the interaction between the complex structures of the RNA-dependent RNA polymerase (RdRp) of SARS-CoV-2 and Remdesivir, an antiviral drug, from the quantum chemical perspective based on the fragment molecular orbital (FMO) method. In addition to the hydrogen bonding and intra-strand stacking between complementary strands as seen in normal base pairs, Remdesivir bound to the terminus of an primer-RNA strand was further stabilized by diagonal π-π stacking with the -1A' base of the complementary strand and an additional hydrogen bond with an intra-strand base, due to the effect of chemically modified functional group. Moreover, stable OH/π interaction is also formed with Thr687 of the RdRp. We quantitatively revealed the exhaustive interaction within the complex among Remdesivir, template-primer-RNA, RdRp and co-factors, and published the results in the FMODB database.

Assessing multiple score functions in Rosetta for drug discovery

Rosetta is a computational software suite containing algorithms for a wide variety of macromolecular structure prediction and design tasks including small molecule protocols commonly used in drug discovery or enzyme design. Here, we benchmark RosettaLigand score functions and protocols in comparison to results of other software recently published in the Comparative Assessment of Score Functions (CASF-2016). The CASF-2016 benchmark covers a wide variety of tests including scoring and ranking multiple compounds against a target, ligand docking of a small molecule to a target, and virtual screening to extract binders from a compound library. Direct comparison to the score functions provided by CASF-2016 results shows that the original RosettaLigand score function ranks among the top software for scoring, ranking, docking and screening tests. Most notably, the RosettaLigand score function ranked 2/34 among other report score functions in CASF-2016. We additionally perform a ligand docking test with full sampling to mimic typical use cases. Despite improved performance of newer score functions in canonical protein structure prediction and design, we demonstrate here that more recent Rosetta score functions have reduced performance across all small molecule benchmarks. The tests described here have also been uploaded to the Rosetta scientific benchmarking server and will be run weekly to track performance as the code is continually being developed.

Discovery and optimizing polycyclic pyridone compounds as anti-HBV agents

INTRODUCTION

Hepatitis B disease is caused by the hepatitis B virus (HBV), which is a DNA virus that belongs to the Hepadnaviridae family. It is a considerable health burden, with 257 million active cases globally. Long-standing infection may create a fundamental cause of liver disease and chronic infections, including cirrhosis, hepatocellular, and carcinoma liver failure. There is an urgent need to develop novel, safe, and effective drug candidates with a novel mechanism of action, improved activity, efficacy, and cure rate.

AREAS COVERED

Herein, the authors provide a concise report focusing on a general and cutting-edge overview of the current state of polycyclic pyridone-related anti-HBV agent patents from 2016 to 2018 and some future perspectives.

EXPERT OPINION

In medicinal chemistry, high-throughput screening (HTS), hit-to-lead optimization (H2L), bioisosteric replacement, and scaffold hopping approaches are playing a major role in the discovery and development of HBV inhibitors. Developing polycyclic pyridone-related anti-HBV agents that could target host factors has attracted significant interest and attention in recent years.

Case studies on computer-based identification of natural products as lead molecules

The development and application of computer-aided drug design/discovery (CADD) techniques (such as structured-base virtual screening, ligand-based virtual screening and neural networks approaches) are on the point of disintermediation in the pharmaceutical drug discovery processes. The application of these CADD methods are standing out positively as compared to other experimental approaches in the identification of hits. In order to venture into new chemical spaces, research groups are exploring natural products (NPs) for the search and identification of new hits and more efficient leads as well as the repurposing of approved NPs. The chemical space of NPs is continuously increasing as a result of millions of years of evolution of species and these data are mainly stored in the form of databases providing access to scientists around the world to conduct studies using them. Investigation of these NP databases with the help of CADD methodologies in combination with experimental validation techniques is essential to identify and propose new drug molecules. In this chapter, we highlight the importance of the chemical diversity of NPs as a source for potential drugs as well as some of the success stories of NP-derived candidates against important therapeutic targets. The focus is on studies that applied a healthy dose of the emerging CADD methodologies (structure-based, ligand-based and machine learning).

Conformational Searching with Quantum Mechanics

Estimating the range of three-dimensional structures (conformations) that are available to a molecule is a key component of computer-aided drug design. Quantum mechanical simulation offers improved accuracy over forcefield methods, but at a high computational cost. The question is whether this increased cost can be justified in a context in which high-throughput analysis of large numbers of molecules is often key. This chapter discusses the application of quantum mechanics to conformational searching, with a focus on three key challenges: (1) the generation of ensembles that include a good approximation to a molecule's bioactive conformation at as prominent a ranking as possible; (2) rational analysis and modification of a pre-established bioactive conformation in terms of its energetics; and (3) approximation of real solution-phase conformational ensembles in tandem with NMR data. The impact of QM on the high-throughput application (1) is debatable, meaning that for the moment its primary application is still lower-throughput applications such as (2) and (3). The optimal choice of QM method is also discussed. Rigorous benchmarking suggests that DFT methods are only acceptable when used with large basis sets, but a trickle of papers continue to obtain useful results with relatively low-cost methods, leading to a dilemma that the literature has yet to fully resolve.

DeepBindRG: a deep learning based method for estimating effective protein-ligand affinity

Proteins interact with small molecules to modulate several important cellular functions. Many acute diseases were cured by small molecule binding in the active site of protein either by inhibition or activation. Currently, there are several docking programs to estimate the binding position and the binding orientation of protein–ligand complex. Many scoring functions were developed to estimate the binding strength and predict the effective protein–ligand binding. While the accuracy of current scoring function is limited by several aspects, the solvent effect, entropy effect, and multibody effect are largely ignored in traditional machine learning methods. In this paper, we proposed a new deep neural network-based model named DeepBindRG to predict the binding affinity of protein–ligand complex, which learns all the effects, binding mode, and specificity implicitly by learning protein–ligand interface contact information from a large protein–ligand dataset. During the initial data processing step, the critical interface information was preserved to make sure the input is suitable for the proposed deep learning model. While validating our model on three independent datasets, DeepBindRG achieves root mean squared error (RMSE) value of pKa (−logK_d or −logK_i) about 1.6–1.8 and R value around 0.5–0.6, which is better than the autodock vina whose RMSE value is about 2.2–2.4 and R value is 0.42–0.57. We also explored the detailed reasons for the performance of DeepBindRG, especially for several failed cases by vina. Furthermore, DeepBindRG performed better for four challenging datasets from DUD.E database with no experimental protein–ligand complexes. The better performance of DeepBindRG than autodock vina in predicting protein–ligand binding affinity indicates that deep learning approach can greatly help with the drug discovery process. We also compare the performance of DeepBindRG with a 4D based deep learning method “pafnucy”, the advantage and limitation of both methods have provided clues for improving the deep learning based protein–ligand prediction model in the future.

Computational prediction of GPCR oligomerization

There has been a recent and prolific expansion in the number of GPCR crystal structures being solved: in both active and inactive forms and in complex with ligand, with G protein and with each other. Despite this, there is relatively little experimental information about the precise configuration of GPCR oligomers during these different biologically relevant states. While it may be possible to identify the experimental conditions necessary to crystallize a GPCR preferentially in a specific structural conformation, computational approaches afford a potentially more tractable means of describing the probability of formation of receptor dimers and higher order oligomers. Ensemble-based computational methods based on structurally determined dimers, coupled with a computational workflow that uses quantum mechanical methods to analyze the chemical nature of the molecular interactions at a GPCR dimer interface, will generate the reproducible and accurate predictions needed to predict previously unidentified GPCR dimers and to inform future advances in our ability to understand and begin to precisely manipulate GPCR oligomers in biological systems. It may also provide information needed to achieve an increase in the number of experimentally determined oligomeric GPCR structures.

Characterising GPCR-ligand interactions using a fragment molecular orbital-based approach

There has been fantastic progress in solving GPCR crystal structures. However, the ability of X-ray crystallography to guide the drug discovery process for GPCR targets is limited by the availability of accurate tools to explore receptor-ligand interactions. Visual inspection and molecular mechanics approaches cannot explain the full complexity of molecular interactions. Quantum mechanical approaches (QM) are often too computationally expensive, but the fragment molecular orbital (FMO) method offers an excellent solution that combines accuracy, speed and the ability to reveal key interactions that would otherwise be hard to detect. Integration of GPCR crystallography or homology modelling with FMO reveals atomistic details of the individual contributions of each residue and water molecule towards ligand binding, including an analysis of their chemical nature.

Beyond the Flavour: The Potential Druggability of Chemosensory G Protein-Coupled Receptors

G protein-coupled receptors (GPCRs) belong to the largest class of drug targets. Approximately half of the members of the human GPCR superfamily are chemosensory receptors, including odorant receptors (ORs), trace amine-associated receptors (TAARs), bitter taste receptors (TAS2Rs), sweet and umami taste receptors (TAS1Rs). Interestingly, these chemosensory GPCRs (csGPCRs) are expressed in several tissues of the body where they are supposed to play a role in biological functions other than chemosensation. Despite their abundance and physiological/pathological relevance, the druggability of csGPCRs has been suggested but not fully characterized. Here, we aim to explore the potential of targeting csGPCRs to treat diseases by reviewing the current knowledge of csGPCRs expressed throughout the body and by analysing the chemical space and the drug-likeness of flavour molecules.