Are Deep Learning Structural Models Sufficiently Accurate for Virtual Screening? Application of Docking Algorithms to AlphaFold2 Predicted Structures.

Machine learning-based protein structure prediction algorithms, such as RosettaFold and AlphaFold2, have greatly impacted the structural biology field, arousing a fair amount of discussion around their potential role in drug discovery. While there are few preliminary studies addressing the usage of these models in virtual screening, none of them focus on the prospect of hit-finding in a real-world virtual screen with a model based on low prior structural information. In order to address this, we have developed an AlphaFold2 version where we exclude all structural templates with more than 30% sequence identity from the model-building process. In a previous study, we used those models in conjunction with state-of-the-art free energy perturbation methods and demonstrated that it is possible to obtain quantitatively accurate results. In this work, we focus on using these structures in rigid receptor-ligand docking studies. Our results indicate that using out-of-the-box Alphafold2 models is not an ideal scenario for virtual screening campaigns; in fact, we strongly recommend to include some post-processing modeling to drive the binding site into a more realistic holo model.

Structure-Activity Relationship Studies of Antimalarial Plasmodium Proteasome Inhibitors─Part II.

With increasing reports of resistance to artemisinins and artemisinin-combination therapies, targeting the Plasmodium proteasome is a promising strategy for antimalarial development. We recently reported a highly selective Plasmodium falciparum proteasome inhibitor with anti-malarial activity in the humanized mouse model. To balance the permeability of the series of macrocycles with other drug-like properties, we conducted further structure-activity relationship studies on a biphenyl ether-tethered macrocyclic scaffold. Extensive SAR studies around the P1, P3, and P5 groups and peptide backbone identified compound TDI-8414. TDI-8414 showed nanomolar antiparasitic activity, no toxicity to HepG2 cells, high selectivity against the Plasmodium proteasome over the human constitutive proteasome and immunoproteasome, improved solubility and PAMPA permeability, and enhanced metabolic stability in microsomes and plasma of both humans and mice.

Are Deep Learning Structural Models Sufficiently Accurate for Free-Energy Calculations? Application of FEP+ to AlphaFold2-Predicted Structures.

The availability of AlphaFold2 has led to great excitement in the scientific community─particularly among drug hunters─due to the ability of the algorithm to predict protein structures with high accuracy. However, beyond globally accurate protein structure prediction, it remains to be determined whether ligand binding sites are predicted with sufficient accuracy in these structures to be useful in supporting computationally driven drug discovery programs. We explored this question by performing free-energy perturbation (FEP) calculations on a set of well-studied protein-ligand complexes, where AlphaFold2 predictions were performed by removing all templates with >30% identity to the target protein from the training set. We observed that in most cases, the ΔΔG values for ligand transformations calculated with FEP, using these prospective AlphaFold2 structures, were comparable in accuracy to the corresponding calculations previously carried out using crystal structures. We conclude that under the right circumstances, AlphaFold2-modeled structures are accurate enough to be used by physics-based methods such as FEP in typical lead optimization stages of a drug discovery program.

Design, Synthesis, and Optimization of Macrocyclic Peptides as Species-Selective Antimalaria Proteasome Inhibitors.

With over 200 million cases and close to half a million deaths each year, malaria is a threat to global health, particularly in developing countries. Plasmodium falciparum, the parasite that causes the most severe form of the disease, has developed resistance to all antimalarial drugs. Resistance to the first-line antimalarial artemisinin and to artemisinin combination therapies is widespread in Southeast Asia and is emerging in sub-Saharan Africa. The P. falciparum proteasome is an attractive antimalarial target because its inhibition kills the parasite at multiple stages of its life cycle and restores artemisinin sensitivity in parasites that have become resistant through mutation in Kelch K13. Here, we detail our efforts to develop noncovalent, macrocyclic peptide malaria proteasome inhibitors, guided by structural analysis and pharmacokinetic properties, leading to a potent, species-selective, metabolically stable inhibitor.

Development of a Highly Selective Plasmodium falciparum Proteasome Inhibitor with Anti-malaria Activity in Humanized Mice.

Plasmodium falciparum proteasome (Pf20S) inhibitors are active against Plasmodium at multiple stages-erythrocytic, gametocyte, liver, and gamete activation stages-indicating that selective Pf20S inhibitors possess the potential to be therapeutic, prophylactic, and transmission-blocking antimalarials. Starting from a reported compound, we developed a noncovalent, macrocyclic peptide inhibitor of the malarial proteasome with high species selectivity and improved pharmacokinetic properties. The compound demonstrates specific, time-dependent inhibition of the ß5 subunit of the Pf20S, kills artemisinin-sensitive and artemisinin-resistant P. falciparum isolates in vitro and reduces parasitemia in humanized, P. falciparum-infected mice.

Rigorous Free Energy Simulations in Virtual Screening.

Virtual high throughput screening (vHTS) in drug discovery is a powerful approach to identify hits: when applied successfully, it can be much faster and cheaper than experimental high-throughput screening approaches. However, mainstream vHTS tools have significant limitations: ligand-based methods depend on knowledge of existing chemical matter, while structure-based tools such as docking involve significant approximations that limit their accuracy. Recent advances in scientific methods coupled with dramatic speedups in computational processing with GPUs make this an opportune time to consider the role of more rigorous methods that could improve the predictive power of vHTS workflows. In this Perspective, we assert that alchemical binding free energy methods using all-atom molecular dynamics simulations have matured to the point where they can be applied in virtual screening campaigns as a final scoring stage to prioritize the top molecules for experimental testing. Specifically, we propose that alchemical absolute binding free energy (ABFE) calculations offer the most direct and computationally efficient approach within a rigorous statistical thermodynamic framework for computing binding energies of diverse molecules, as is required for virtual screening. ABFE calculations are particularly attractive for drug discovery at this point in time, where the confluence of large-scale genomics data and insights from chemical biology have unveiled a large number of promising disease targets for which no small molecule binders are known, precluding ligand-based approaches, and where traditional docking approaches have foundered to find progressible chemical matter.

Mechanism of biomolecular recognition of trimethyllysine by the fluorinated aromatic cage of KDM5A PHD3 finger.

The understanding of biomolecular recognition of posttranslationally modified histone proteins is centrally important to the histone code hypothesis. Despite extensive binding and structural studies on the readout of histones, the molecular language by which posttranslational modifications on histone proteins are read remains poorly understood. Here we report physical-organic chemistry studies on the recognition of the positively charged trimethyllysine by the electron-rich aromatic cage containing PHD3 finger of KDM5A. The aromatic character of two tryptophan residues that solely constitute the aromatic cage of KDM5A was fine-tuned by the incorporation of fluorine substituents. Our thermodynamic analyses reveal that the wild-type and fluorinated KDM5A PHD3 fingers associate equally well with trimethyllysine. This work demonstrates that the biomolecular recognition of trimethyllysine by fluorinated aromatic cages is associated with weaker cation-π interactions that are compensated by the energetically more favourable trimethyllysine-mediated release of high-energy water molecules that occupy the aromatic cage.

Small-molecule targeting of MUSASHI RNA-binding activity in acute myeloid leukemia.

The MUSASHI (MSI) family of RNA binding proteins (MSI1 and MSI2) contribute to a wide spectrum of cancers including acute myeloid leukemia. We find that the small molecule Ro 08-2750 (Ro) binds directly and selectively to MSI2 and competes for its RNA binding in biochemical assays. Ro treatment in mouse and human myeloid leukemia cells results in an increase in differentiation and apoptosis, inhibition of known MSI-targets, and a shared global gene expression signature similar to shRNA depletion of MSI2. Ro demonstrates in vivo inhibition of c-MYC and reduces disease burden in a murine AML leukemia model. Thus, we identify a small molecule that targets MSI's oncogenic activity. Our study provides a framework for targeting RNA binding proteins in cancer.

Large-Scale Validation of Mixed-Solvent Simulations to Assess Hotspots at Protein-Protein Interaction Interfaces.

The ability to target protein-protein interactions (PPIs) with small molecule inhibitors offers great promise in expanding the druggable target space and addressing a broad range of untreated diseases. However, due to their nature and function of interacting with protein partners, PPI interfaces tend to extend over large surfaces without the typical pockets of enzymes and receptors. These features present unique challenges for small molecule inhibitor design. As such, determining whether a particular PPI of interest could be pursued with a small molecule discovery strategy requires an understanding of the characteristics of the PPI interface and whether it has hotspots that can be leveraged by small molecules to achieve desired potency. Here, we assess the ability of mixed-solvent molecular dynamic (MSMD) simulations to detect hotspots at PPI interfaces. MSMD simulations using three cosolvents (acetonitrile, isopropanol, and pyrimidine) were performed on a large test set of 21 PPI targets that have been experimentally validated by small molecule inhibitors. We compare MSMD, which includes explicit solvent and full protein flexibility, to a simpler approach that does not include dynamics or explicit solvent (SiteMap) and find that MSMD simulations reveal additional information about the characteristics of these targets and the ability for small molecules to inhibit the PPI interface. In the few cases were MSMD simulations did not detect hotspots, we explore the shortcomings of this technique and propose future improvements. Finally, using Interleukin-2 as an example, we highlight the advantage of the MSMD approach for detecting transient cryptic druggable pockets that exists at PPI interfaces.

Identification of a Mycothiol-Dependent Nitroreductase from Mycobacterium tuberculosis.

The success of Mycobacterium tuberculosis (Mtb) as a pathogen depends on the redundant and complex mechanisms it has evolved for resisting nitrosative and oxidative stresses inflicted by host immunity. Improving our understanding of these defense pathways can reveal vulnerable points in Mtb pathogenesis. In this study, we combined genetic, structural, computational, biochemical, and biophysical approaches to identify a novel enzyme class represented by Rv2466c. We show that Rv2466c is a mycothiol-dependent nitroreductase of Mtb and can reduce the nitro group of a novel mycobactericidal compound using mycothiol as a cofactor. In addition to its function as a nitroreductase, Rv2466c confers partial protection to menadione stress.

Conformational Landscape of the PRKACA-DNAJB1 Chimeric Kinase, the Driver for Fibrolamellar Hepatocellular Carcinoma.

In fibrolamellar hepatocellular carcinoma a single genetic deletion results in the fusion of the first exon of the heat shock protein 40, DNAJB1, which encodes the J domain, with exons 2-10 of the catalytic subunit of protein kinase A, PRKACA. This produces an enzymatically active chimeric protein J-PKAcα. We used molecular dynamics simulations and NMR to analyze the conformational landscape of native and chimeric kinase, and found an ensemble of conformations. These ranged from having the J-domain tucked under the large lobe of the kinase, similar to what was reported in the crystal structure, to others where the J-domain was dislodged from the core of the kinase and swinging free in solution. These simulated dislodged states were experimentally captured by NMR. Modeling of the different conformations revealed no obvious steric interactions of the J-domain with the rest of the RIIß holoenzyme.

Calculating Water Thermodynamics in the Binding Site of Proteins - Applications of WaterMap to Drug Discovery.

The ability to accurately characterize the solvation properties (water locations and thermodynamics) of biomolecules is of great importance to drug discovery. While crystallography, NMR, and other experimental techniques can assist in determining the structure of water networks in proteins and protein-ligand complexes, most water molecules are not fully resolved and accurately placed. Furthermore, understanding the energetic effects of solvation and desolvation on binding requires an analysis of the thermodynamic properties of solvent involved in the interaction between ligands and proteins. WaterMap is a molecular dynamics-based computational method that uses statistical mechanics to describe the thermodynamic properties (entropy, enthalpy, and free energy) of water molecules at the surface of proteins. This method can be used to assess the solvent contributions to ligand binding affinity and to guide lead optimization. In this review, we provide a comprehensive summary of published uses of WaterMap, including applications to lead optimization, virtual screening, selectivity analysis, ligand pose prediction, and druggability assessment.

Investigating Protein-Peptide Interactions Using the Schrödinger Computational Suite.

The Schrödinger software suite contains a broad array of computational chemistry and molecular modeling tools that can be used to study the interaction of peptides with proteins. These include molecular docking using Glide and Piper, relative binding free energy predictions with FEP+, conformational searches using MacroModel and Desmond, and structural refinement using Prime and PrimeX. In this review we provide a comprehensive overview of these tools and describe their potential application in the identification and optimization of peptide ligands for proteins.

Relative Binding Free Energy Calculations Applied to Protein Homology Models.

A significant challenge and potential high-value application of computer-aided drug design is the accurate prediction of protein-ligand binding affinities. Free energy perturbation (FEP) using molecular dynamics (MD) sampling is among the most suitable approaches to achieve accurate binding free energy predictions, due to the rigorous statistical framework of the methodology, correct representation of the energetics, and thorough treatment of the important degrees of freedom in the system (including explicit waters). Recent advances in sampling methods and force fields coupled with vast increases in computational resources have made FEP a viable technology to drive hit-to-lead and lead optimization, allowing for more efficient cycles of medicinal chemistry and the possibility to explore much larger chemical spaces. However, previous FEP applications have focused on systems with high-resolution crystal structures of the target as starting points-something that is not always available in drug discovery projects. As such, the ability to apply FEP on homology models would greatly expand the domain of applicability of FEP in drug discovery. In this work we apply a particular implementation of FEP, called FEP+, on congeneric ligand series binding to four diverse targets: a kinase (Tyk2), an epigenetic bromodomain (BRD4), a transmembrane GPCR (A2A), and a protein-protein interaction interface (BCL-2 family protein MCL-1). We apply FEP+ using both crystal structures and homology models as starting points and find that the performance using homology models is generally on a par with the results when using crystal structures. The robustness of the calculations to structural variations in the input models can likely be attributed to the conformational sampling in the molecular dynamics simulations, which allows the modeled receptor to adapt to the "real" conformation for each ligand in the series. This work exemplifies the advantages of using all-atom simulation methods with full system flexibility and offers promise for the general application of FEP to homology models, although additional validation studies should be performed to further understand the limitations of the method and the scenarios where FEP will work best.

In search of novel ligands using a structure-based approach: a case study on the adenosine A2A receptor.

In this work, we present a case study to explore the challenges associated with finding novel molecules for a receptor that has been studied in depth and has a wealth of chemical information available. Specifically, we apply a previously described protocol that incorporates explicit water molecules in the ligand binding site to prospectively screen over 2.5 million drug-like and lead-like compounds from the commercially available eMolecules database in search of novel binders to the adenosine A2A receptor (A2AAR). A total of seventy-one compounds were selected for purchase and biochemical assaying based on high ligand efficiency and high novelty (Tanimoto coefficient ≤0.25 to any A2AAR tested compound). These molecules were then tested for their affinity to the adenosine A2A receptor in a radioligand binding assay. We identified two hits that fulfilled the criterion of ~50 % radioligand displacement at a concentration of 10 µM. Next we selected an additional eight novel molecules that were predicted to make a bidentate interaction with Asn2536.55, a key interacting residue in the binding pocket of the A2AAR. None of these eight molecules were found to be active. Based on these results we discuss the advantages of structure-based methods and the challenges associated with finding chemically novel molecules for well-explored targets.

Molecular dynamics techniques for modeling G protein-coupled receptors.

G protein-coupled receptors (GPCRs) constitute a major class of drug targets and modulating their signaling can produce a wide range of pharmacological outcomes. With the growing number of high-resolution GPCR crystal structures, we have the unprecedented opportunity to leverage structure-based drug design techniques. Here, we discuss a number of advanced molecular dynamics (MD) techniques that have been applied to GPCRs, including long time scale simulations, enhanced sampling techniques, water network analyses, and free energy approaches to determine relative binding free energies. On the basis of the many success stories, including those highlighted here, we expect that MD techniques will be increasingly applied to aid in structure-based drug design and lead optimization for GPCRs.

Predicting Binding Affinities for GPCR Ligands Using Free-Energy Perturbation.

The rapid growth of structural information for G-protein-coupled receptors (GPCRs) has led to a greater understanding of their structure, function, selectivity, and ligand binding. Although novel ligands have been identified using methods such as virtual screening, computationally driven lead optimization has been possible only in isolated cases because of challenges associated with predicting binding free energies for related compounds. Here, we provide a systematic characterization of the performance of free-energy perturbation (FEP) calculations to predict relative binding free energies of congeneric ligands binding to GPCR targets using a consistent protocol and no adjustable parameters. Using the FEP+ package, first we validated the protocol, which includes a full lipid bilayer and explicit solvent, by predicting the binding affinity for a total of 45 different ligands across four different GPCRs (adenosine A2AAR, ß1 adrenergic, CXCR4 chemokine, and δ opioid receptors). Comparison with experimental binding affinity measurements revealed a highly predictive ranking correlation (average spearman ρ = 0.55) and low root-mean-square error (0.80 kcal/mol). Next, we applied FEP+ in a prospective project, where we predicted the affinity of novel, potent adenosine A2A receptor (A2AR) antagonists. Four novel compounds were synthesized and tested in a radioligand displacement assay, yielding affinity values in the nanomolar range. The affinity of two out of the four novel ligands (plus three previously reported compounds) was correctly predicted (within 1 kcal/mol), including one compound with approximately a tenfold increase in affinity compared to the starting compound. Detailed analyses of the simulations underlying the predictions provided insights into the structural basis for the two cases where the affinity was overpredicted. Taken together, these results establish a protocol for systematically applying FEP+ to GPCRs and provide guidelines for identifying potent molecules in drug discovery lead optimization projects.

Chemical basis for the recognition of trimethyllysine by epigenetic reader proteins.

A large number of structurally diverse epigenetic reader proteins specifically recognize methylated lysine residues on histone proteins. Here we describe comparative thermodynamic, structural and computational studies on recognition of the positively charged natural trimethyllysine and its neutral analogues by reader proteins. This work provides experimental and theoretical evidence that reader proteins predominantly recognize trimethyllysine via a combination of favourable cation-π interactions and the release of the high-energy water molecules that occupy the aromatic cage of reader proteins on the association with the trimethyllysine side chain. These results have implications in rational drug design by specifically targeting the aromatic cage of readers of trimethyllysine.

Docking and Virtual Screening Strategies for GPCR Drug Discovery.

Progress in structure determination of G protein-coupled receptors (GPCRs) has made it possible to apply structure-based drug design (SBDD) methods to this pharmaceutically important target class. The quality of GPCR structures available for SBDD projects fall on a spectrum ranging from high resolution crystal structures (<2 Å), where all water molecules in the binding pocket are resolved, to lower resolution (>3 Å) where some protein residues are not resolved, and finally to homology models that are built using distantly related templates. Each GPCR project involves a distinct set of opportunities and challenges, and requires different approaches to model the interaction between the receptor and the ligands. In this review we will discuss docking and virtual screening to GPCRs, and highlight several refinement and post-processing steps that can be used to improve the accuracy of these calculations. Several examples are discussed that illustrate specific steps that can be taken to improve upon the docking and virtual screening accuracy. While GPCRs are a unique target class, many of the methods and strategies outlined in this review are general and therefore applicable to other protein families.