AI-Based Protein Structure Prediction in Drug Discovery: Impacts and Challenges.

Proteins are the molecular machinery of the human body, and their malfunctioning is often responsible for diseases, making them crucial targets for drug discovery. The three-dimensional structure of a protein determines its biological function, its conformational state determines substrates, cofactors, and protein binding. Rational drug discovery employs engineered small molecules to selectively interact with proteins to modulate their function. To selectively target a protein and to design small molecules, knowing the protein structure with all its specific conformation is critical. Unfortunately, for a large number of proteins relevant for drug discovery, the three-dimensional structure has not yet been experimentally solved. Therefore, accurately predicting their structure based on their amino acid sequence is one of the grant challenges in biology. Recently, AlphaFold2, a machine learning application based on a deep neural network, was able to predict unknown structures of proteins with an unprecedented accuracy. Despite the impressive progress made by AlphaFold2, nature still challenges the field of structure prediction. In this Perspective, we explore how AlphaFold2 and related methods help make drug design more efficient. Furthermore, we discuss the roles of predicting domain-domain orientations, all relevant conformational states, the influence of posttranslational modifications, and conformational changes due to protein binding partners. We highlight where further improvements are needed for advanced machine learning methods to be successfully and frequently used in the pharmaceutical industry.

Recent applications of computational methods to allosteric drug discovery.

Interest in exploiting allosteric sites for the development of new therapeutics has grown considerably over the last two decades. The chief driving force behind the interest in allostery for drug discovery stems from the fact that in comparison to orthosteric sites, allosteric sites are less conserved across a protein family, thereby offering greater opportunity for selectivity and ultimately tolerability. While there is significant overlap between structure-based drug design for orthosteric and allosteric sites, allosteric sites offer additional challenges mostly involving the need to better understand protein flexibility and its relationship to protein function. Here we examine the extent to which structure-based drug design is impacting allosteric drug design by highlighting several targets across a variety of target classes.

Selective, Small-Molecule Co-Factor Binding Site Inhibition of a Su(var)3-9, Enhancer of Zeste, Trithorax Domain Containing Lysine Methyltransferase.

The first chemical probe to primarily occupy the co-factor binding site of a Su(var)3-9, enhancer of a zeste, trithorax (SET) domain containing protein lysine methyltransferase (PKMT) is reported. Protein methyltransferases require S-adenosylmethionine (SAM) as a co-factor (methyl donor) for enzymatic activity. However, SAM itself represents a poor medicinal chemistry starting point for a selective, cell-active inhibitor given its extreme physicochemical properties and its role in multiple cellular processes. A previously untested medicinal chemistry strategy of deliberate file enrichment around molecules bearing the hallmarks of SAM, but with improved lead-like properties from the outset, yielded viable hits against SET and MYND domain-containing protein 2 (SMYD2) that were shown to bind in the co-factor site. These leads were optimized to identify a highly biochemically potent, PKMT-selective, and cell-active chemical probe. While substrate-based inhibitors of PKMTs are known, this represents a novel, co-factor-derived strategy for the inhibition of SMYD2 which may also prove applicable to lysine methyltransferase family members previously thought of as intractable.

Millisecond dynamics of BTK reveal kinome-wide conformational plasticity within the apo kinase domain.

Bruton tyrosine kinase (BTK) is a key enzyme in B-cell development whose improper regulation causes severe immunodeficiency diseases. Design of selective BTK therapeutics would benefit from improved, in-silico structural modeling of the kinase's solution ensemble. However, this remains challenging due to the immense computational cost of sampling events on biological timescales. In this work, we combine multi-millisecond molecular dynamics (MD) simulations with Markov state models (MSMs) to report on the thermodynamics, kinetics, and accessible states of BTK's kinase domain. Our conformational landscape links the active state to several inactive states, connected via a structurally diverse intermediate. Our calculations predict a kinome-wide conformational plasticity, and indicate the presence of several new potentially druggable BTK states. We further find that the population of these states and the kinetics of their inter-conversion are modulated by protonation of an aspartate residue, establishing the power of MD & MSMs in predicting effects of chemical perturbations.

Markov modeling reveals novel intracellular modulation of the human TREK-2 selectivity filter.

Two-pore domain potassium (K2P) channel ion conductance is regulated by diverse stimuli that directly or indirectly gate the channel selectivity filter (SF). Recent crystal structures for the TREK-2 member of the K2P family reveal distinct "up" and "down" states assumed during activation via mechanical stretch. We performed 195 µs of all-atom, unbiased molecular dynamics simulations of the TREK-2 channel to probe how membrane stretch regulates the SF gate. Markov modeling reveals a novel "pinched" SF configuration that stretch activation rapidly destabilizes. Free-energy barrier heights calculated for critical steps in the conduction pathway indicate that this pinched state impairs ion conduction. Our simulations predict that this low-conductance state is accessed exclusively in the compressed, "down" conformation in which the intracellular helix arrangement allosterically pinches the SF. By explicitly relating structure to function, we contribute a critical piece of understanding to the evolving K2P puzzle.

Estimation of relative free energies of binding using pre-computed ensembles based on the single-step free energy perturbation and the site-identification by Ligand competitive saturation approaches.

Accurate and rapid estimation of relative binding affinities of ligand-protein complexes is a requirement of computational methods for their effective use in rational ligand design. Of the approaches commonly used, free energy perturbation (FEP) methods are considered one of the most accurate, although they require significant computational resources. Accordingly, it is desirable to have alternative methods of similar accuracy but greater computational efficiency to facilitate ligand design. In the present study relative free energies of binding are estimated for one or two non-hydrogen atom changes in compounds targeting the proteins ACK1 and p38 MAP kinase using three methods. The methods include standard FEP, single-step free energy perturbation (SSFEP) and the site-identification by ligand competitive saturation (SILCS) ligand grid free energy (LGFE) approach. Results show the SSFEP and SILCS LGFE methods to be competitive with or better than the FEP results for the studied systems, with SILCS LGFE giving the best agreement with experimental results. This is supported by additional comparisons with published FEP data on p38 MAP kinase inhibitors. While both the SSFEP and SILCS LGFE approaches require a significant upfront computational investment, they offer a 1000-fold computational savings over FEP for calculating the relative affinities of ligand modifications once those pre-computations are complete. An illustrative example of the potential application of these methods in the context of screening large numbers of transformations is presented. Thus, the SSFEP and SILCS LGFE approaches represent viable alternatives for actively driving ligand design during drug discovery and development. © 2016 Wiley Periodicals, Inc.

Computational Modeling of ß-Secretase 1 (BACE-1) Inhibitors Using Ligand Based Approaches.

The binding affinities (IC50) reported for diverse structural and chemical classes of human ß-secretase 1 (BACE-1) inhibitors in literature were modeled using multiple in silico ligand based modeling approaches and statistical techniques. The descriptor space encompasses simple binary molecular fingerprint, one- and two-dimensional constitutional, physicochemical, and topological descriptors, and sophisticated three-dimensional molecular fields that require appropriate structural alignments of varied chemical scaffolds in one universal chemical space. The affinities were modeled using qualitative classification or quantitative regression schemes involving linear, nonlinear, and deep neural network (DNN) machine-learning methods used in the scientific literature for quantitative-structure activity relationships (QSAR). In a departure from tradition, ∼20% of the chemically diverse data set (205 compounds) was used to train the model with the remaining ∼80% of the structural and chemical analogs used as part of an external validation (1273 compounds) and prospective test (69 compounds) sets respectively to ascertain the model performance. The machine-learning methods investigated herein performed well in both the qualitative classification (∼70% accuracy) and quantitative IC50 predictions (RMSE ∼ 1 log). The success of the 2D descriptor based machine learning approach when compared against the 3D field based technique pursued for hBACE-1 inhibitors provides a strong impetus for systematically applying such methods during the lead identification and optimization efforts for other protein families as well.

Recent developments in targeting protein misfolding diseases.

Protein misfolding is an emerging field that crosses multiple therapeutic areas and causes many serious diseases. As the biological pathways of protein misfolding become more clearly elucidated, small molecule approaches in this arena are gaining increased attention. This manuscript will survey current small molecules from the literature that are known to modulate misfolding, stabilization or proteostasis. Specifically, the following targets and approaches will be discussed: CFTR, glucocerebrosidase, modulation of toxic oligomers, serum amyloid P (SAP) sections and HSF1 activators.

Small molecules that target protein misfolding.

Protein misfolding is a process in which proteins are unable to attain or maintain their biologically active conformation. Factors contributing to protein misfolding include missense mutations and intracellular factors such as pH changes, oxidative stress, or metal ions. Protein misfolding is linked to a large number of diseases such as cystic fibrosis, Alzheimer's disease, Parkinson's disease, amyotrophic lateral sclerosis, and less familiar diseases such as Gaucher's disease, nephrogenic diabetes insipidus, and Creutzfeldt-Jakob disease. In this Perspective, we report on small molecules that bind to and stabilize the aberrant protein, thereby helping it to attain a native or near-native conformation and restoring its function. The following targets will be specifically discussed: transthyretin, p53, superoxide dismutase 1, lysozyme, serum amyloid A, prions, vasopressin receptor 2, and α-1-antitrypsin.

Discovery of indazoles as inhibitors of Tpl2 kinase.

Synthesis, modeling and structure-activity relationship of indazoles as inhibitors of Tpl2 kinase are described. From a high throughput screening effort, we identified an indazole hit compound 5 that has a single digit micromolar Tpl2 activity. Through SAR modifications at the C3 and C5 positions of the indazole, we discovered compound 31 with good potency in LANCE assay and cell-based p-Erk assay.

Tetrahydrocarbazole-based serotonin reuptake inhibitor/dopamine D2 partial agonists for the potential treatment of schizophrenia.

A 5-fluoro-tetrahydrocarbazole serotonin reuptake inhibitor (SRI) building block was combined with a variety of linkers and dopamine D2 receptor ligands in an attempt to identify potent D2 partial agonist/SRI molecules for treatment of schizophrenia. This approach has the potential to treat a broader range of symptoms compared to existing therapies. Selected compounds in this series demonstrate high affinity for both targets and D2 partial agonism in cell-based and in vivo assays.