Small Molecule Ligands Research Articles

Improvements in Precision of Relative Binding Free Energy Calculations Afforded by the Alchemical Enhanced Sampling (ACES) Approach.

Accurate in silico predictions of how strongly small molecules bind to proteins, such as those afforded by relative binding free energy (RBFE) calculations, can greatly increase the efficiency of the hit-to-lead and lead optimization stages of the drug discovery process. The success of such calculations, however, relies heavily on their precision. Here, we show that a recently developed alchemical enhanced sampling (ACES) approach can consistently improve the precision of RBFE calculations on a large and diverse set of proteins and small molecule ligands. The addition of ACES to conventional RBFE calculations lowered the average hysteresis by over 35% (0.3-0.4 kcal/mol) and the average replicate spread by over 25% (0.2-0.3 kcal/mol) across a set of 10 protein targets and 213 small molecules while maintaining similar or improved accuracy. We show in atomic detail how ACES improved convergence of several representative RBFE calculations through enhancing the sampling of important slowly transitioning ligand degrees of freedom.

Enabling target-aware molecule generation to follow multi objectives with Pareto MCTS

Target-aware drug discovery has greatly accelerated the drug discovery process to design small-molecule ligands with high binding affinity to disease-related protein targets. Conditioned on targeted proteins, previous works utilize various kinds of deep generative models and have shown great potential in generating molecules with strong protein-ligand binding interactions. However, beyond binding affinity, effective drug molecules must manifest other essential properties such as high drug-likeness, which are not explicitly addressed by current target-aware generative methods. In this article, aiming to bridge the gap of multi-objective target-aware molecule generation in the field of deep learning-based drug discovery, we propose ParetoDrug, a Pareto Monte Carlo Tree Search (MCTS) generation algorithm. ParetoDrug searches molecules on the Pareto Front in chemical space using MCTS to enable synchronous optimization of multiple properties. Specifically, ParetoDrug utilizes pretrained atom-by-atom autoregressive generative models for the exploration guidance to desired molecules during MCTS searching. Besides, when selecting the next atom symbol, a scheme named ParetoPUCT is proposed to balance exploration and exploitation. Benchmark experiments and case studies demonstrate that ParetoDrug is highly effective in traversing the large and complex chemical space to discover novel compounds with satisfactory binding affinities and drug-like properties for various multi-objective target-aware drug discovery tasks.

Capture of RNA G-quadruplex structures using an l-RNA aptamer.

G-quadruplexes (dG4 and rG4) are nucleic acid secondary structures formed by the self-assembly of certain G-rich sequences, and they have distinctive chemical properties and play crucial roles in fundamental biological processes. Small molecule G4 ligands were shown to be crucial in characterizing G4s and understanding their functions. Nevertheless, concerns regarding the specificity of these synthetic ligands for further investigation of G4s, especially for rG4 isolation purposes, have been raised. In comparison to G4 ligands, we propose a novel magnetic bead-based pulldown assay that enables the selective capture of general rG4s using functionalized l-Apt.4-1c from both simple buffer and complex media, including total RNA and the cell lysate. We found that our l-RNA aptamer can pulldown general rG4s with a higher efficiency and specificity than the G4 small molecule ligand BioTASQ v.1 in the presence of non-target competitors, including dG4 and non-G4 structures. Our findings reveal that biotinylated l-aptamers can serve as effective molecular tools for the affinity-based enrichment of rG4 of interest using this new assay, which was also verified by quantitative reverse transcription-polymerase chain reaction (RT-qPCR) on endogenous transcripts. This work provides new and important insights into rG4 isolation using a functionalized l-aptamer, which can potentially be applied in a transcript-specific or transcriptome-wide manner in the future.

Ligand Identification using Deep Learning.

Accurately identifying ligands plays a crucial role in the process of structure-guided drug design. Based on density maps from X-ray diffraction or cryogenic-sample electron microscopy (cryoEM), scientists verify whether small-molecule ligands bind to active sites of interest. However, the interpretation of density maps is challenging, and cognitive bias can sometimes mislead investigators into modeling fictitious compounds. Ligand identification can be aided by automatic methods, but existing approaches are available only for X-ray diffraction and are based on iterative fitting or feature-engineered machine learning rather than end-to-end deep learning. Here, we propose to identify ligands using a deep learning approach that treats density maps as 3D point clouds. We show that the proposed model is on par with existing machine learning methods for X-ray crystallography while also being applicable to cryoEM density maps. Our study demonstrates that electron density map fragments can be used to train models that can be applied to cryoEM structures, but also highlights challenges associated with the standardization of electron microscopy maps and the quality assessment of cryoEM ligands. Code and model weights are available on GitHub at https://github.com/jkarolczak/ligands-classification . Datasets used for training and testing are hosted at Zenodo: 10.5281/zenodo.10908325 . dariusz.brzezinski@cs.put.poznan.pl.

Development of Ligands and Degraders Targeting MAGE-A3.

Type I melanoma antigen (MAGE) family members are detected in numerous tumor types, and expression is correlated with poor prognosis, high tumor grade, and increased metastasis. Type I MAGE proteins are typically restricted to reproductive tissues, but expression can recur during tumorigenesis. Several biochemical functions have been elucidated for them, and notably, MAGEs regulate proteostasis by serving as substrate recognition modules for E3 ligase complexes. The repertoire of E3 ligase complexes that can be hijacked for targeted protein degradation continues to expand, and MAGE-E3 complexes are an especially attractive platform given their cancer-selective expression. Additionally, type I MAGE-derived peptides are presented on cancer cell surfaces, so targeted MAGE degradation may increase antigen presentation and improve immunotherapy outcomes. Motivated by these applications, we developed novel, small-molecule ligands for MAGE-A3, a type I MAGE that is widely expressed in tumors and associates with TRIM28, a RING E3 ligase. Chemical matter was identified through DNA-encoded library (DEL) screening, and hit compounds were validated for in vitro binding to MAGE-A3. We obtained a cocrystal structure with a DEL analog and hypothesize that the small molecule binds at a dimer interface. We utilized this ligand to develop PROTAC molecules that induce MAGE-A3 degradation through VHL recruitment and inhibit the proliferation of MAGE-A3 positive cell lines. These ligands and degraders may serve as valuable probes for investigating MAGE-A3 biology and as foundations for the ongoing development of tumor-specific PROTACs.

CG-SLENP: A Chemical Genetics Strategy To Selectively Label Existing Proteins and Newly Synthesized Proteins.

Protein synthesis and subsequent delivery to the target locations in cells are essential for their proper functions. Methods to label and distinguish newly synthesized proteins from existing ones are critical to assess their differential properties, but such methods are lacking. We describe the first chemical genetics-based approach for selective labeling of existing and newly synthesized proteins that we termed as CG-SLENP. Using HaloTag in-frame fusion with lamin A (LA), we demonstrate that the two pools of proteins can be selectively labeled using CG-SLENP in living cells. We further employ our recently developed selective small molecule ligand LBL1 for LA to probe the potential differences between newly synthesized and existing LA. Our results show that LBL1 can differentially modulate these two pools of LA. These results indicate that the assembly states of newly synthesized LA are distinct from existing LA in living cells. The CG-SLENP method is potentially generalizable to study any cellular proteins.

Computational Analysis of a Marine-Derived Drug From Rhizophora mucronata Against the Capsid Protein of Rubella Virus.

Background Rubella, commonly known as German measles, is caused by a single-stranded RNA genome. Vaccination is currently the most effective method for preventing rubella and its complications. Molecular docking, a computer-based technique used in drug discovery and development, is used to investigate the interactions between potential drug candidates and their target proteins. It predicts the binding interactions between small molecules (ligands) and the target protein. In this study, we examined a marine-derived drug from Rhizophora mucronata for its potential antiviral properties against the rubella capsid virus. Our objective was to identify the active inhibitory sites of the capsid virus. Materials and methods Protein and ligand molecules were retrieved from Protein Data Bank(PDB) and PubChem databases. The Lamarckian genetic algorithm was used to calculate molecular docking using Autodock Tools 1.5.7. The docking parameters used for each docked molecule were determined from 100 separate docking experiments with a maximum of 2.5×10-6 energy and a mutation rate of 2.0 and mass over ratio of 0.8. The results were recorded as docking parameter files (DPF). PyMOL was used to view and investigate the interactions between ligand fragments and rubella capsid protein. Results This approach plays a crucial role in the development of structure-based drugs. The results of the molecular docking suggest that Rhizophorin has the potential to bind with the rubella capsid protein. The strong binding affinity of -6.05 kcal/mol between the ligand and the protein further supports the potential of Rhizophorin as a therapeutic agent. The formation of hydrogen bonds between the ligand and amino acid residues Glu79, Arg82, and Thr118 indicates the significance of electrostatic interactions in the binding process. Furthermore, the hydrophobic interactions between the ligand and residues Ala81, Val84, Leu87, and Ile119 suggest the role of non-polar interactions in stabilizing the complex. The identified amino acid residues involved in these binding interactions could serve as potential targets for drug development. In future studies, experimental validation of the predicted interactions could provide further insights into the potential of Rhizophorin as an antiviral agent. Conclusion According to the findings of this study, the in silico investigation successfully identified a target for inhibiting the rubella virus (RuV) capsid receptor molecule. Future investigations on these compounds will require in vitro and in vivo studies using models that are more relevant to the medicinal potential of the capsid protein molecule.

The C-terminal self-binding helical peptide of human estrogen-related receptor γ can be druggably targeted by a novel class of rationally designed peptidic antagonists.

Orphan nuclear estrogen-related receptor γ (ERRγ) has been recognized as a potential therapeutic target for cancer, inflammation and metabolic disorder. The ERRγ contains a regulatory AF2 helical tail linked C-terminally to its ligand-binding domain (LBD), which is a self-binding peptide (SBP) and serves as molecular switch to dynamically regulate the receptor alternation between active and inactive states by binding to and unbinding from the AF2-binding site on ERRγ LBD surface, respectively. Traditional ERRγ modulators are all small-molecule chemical ligands that can be classified into agonists and inverse agonists in terms of their action mechanism; the agonists stabilize the AF2 in ABS site with an agonist conformation, while the inverse agonists lock the AF2 out of the site to largely abolish ERRγ transcriptional activity. Here, a class of ERRγ peptidic antagonists was described to compete with native AF2 for the ABS site, thus blocking the active state of AF2 binding to ERRγ LBD domain. Self-inhibitory peptide was derived from the SBP-covering AF2 region and we expected it can rebind potently to the ABS site by reducing its intrinsic disorder and entropy cost upon the rebinding. Hydrocarbon stapling was employed to do so, which employed an all-hydrocarbon bridge across the [i, i + 4]-anchor residue pair in the N-terminal, middle or C-terminal region of the self-inhibitory peptide. As might be expected, it is revealed that the stapled peptides are good binders of ERRγ LBD domain and can effectively compete with the native AF2 helical tail for ERRγ ABS site, which exhibit a basically similar binding mode with AF2 to the site and form diverse noncovalent interactions with the site, thus conferring stability and specificity to the domain-peptide complexes.

The Discovery of a Specific CKIP-1 Ligand for the Potential Treatment of Disuse Osteoporosis.

Bone homeostasis relies on the delicate balance between osteoblast-mediated bone formation and osteoclast-mediated bone resorption. The casein kinase 2 interacting protein-1 (CKIP-1), a specific CK2α subunit-interacting protein, has been documented as one of the crucial negative regulators of bone formation. CKIP-1 siRNA therapy has constraints that limit its use in clinical applications. Therefore, it is necessary to explore effective targeting strategies for CKIP-1. In this study, we observed an upregulation of CKIP-1 protein expression in the microgravity environment, while its ubiquitination levels decreased. We further investigated the interaction between CKIP-1 and VHL and found that VHL enhanced CKIP-1 degradation through the ubiquitylation-proteasome system (UPS). Additionally, we discovered a small molecule ligand, named C77, through DNA-encoded library (DEL) screening, which binds to CKIP-1 both in vivo and in vitro, as confirmed by Surface Plasmon Resonance (SPR) and the Cellular Thermal shift assay (CETSA), respectively. Our findings demonstrated the potential of VHL and C77 as guiding factors in the development of CKIP-1-based Proteolysis-Targeting Chimeras (PROTACs), which could be future therapeutic interventions in disuse osteoporosis.

TetR family regulator AbrT controls lincomycin production and morphological development in Streptomyces lincolnensis

BackgroundThe TetR family of transcriptional regulators (TFRs), serving as crucial regulators of diverse cellular processes, undergo conformational changes induced by small-molecule ligands, which either inhibit or activate them to modulate target gene expression. Some ligands of TFRs in actinomycetes and their regulatory effects have been identified and studied; however, regulatory mechanisms of the TetR family in the lincomycin-producing Streptomyces lincolnensis remain poorly understood.ResultsIn this study, we found that AbrT (SLCG_1979), a TetR family regulator, plays a pivotal role in regulating lincomycin production and morphological development in S. lincolnensis. Deletion of abrT gene resulted in increased lincomycin A (Lin-A) production, but delayed mycelium formation and sporulation on solid media. AbrT directly or indirectly repressed the expression of lincomycin biosynthetic (lin) cluster genes and activated that of the morphological developmental genes amfC, whiB, and ftsZ. We demonstrated that AbrT bound to two motifs (5′-CGCGTACTCGTA-3′ and 5′-CGTACGATAGCT-3′) present in the bidirectional promoter between abrT and SLCG_1980 genes. This consequently repressed abrT itself and its adjacent gene SLCG_1980 that encodes an arabinose efflux permease. D-arabinose, not naturally occurring as L-arabinose, was identified as the effector molecule of AbrT, reducing its binding affinity to abrT-SLCG_1980 intergenic region. Furthermore, based on functional analysis of the AbrT homologue in Saccharopolyspora erythraea, we inferred that the TetR family regulator AbrT may play an important role in regulating secondary metabolism in actinomycetes.ConclusionsAbrT functions as a regulator for governing lincomycin production and morphological development of S. lincolnensis. Our findings demonstrated that D-arabinose acts as a ligand of AbrT to mediate the regulation of lincomycin biosynthesis in S. lincolnensis. Our findings provide novel insights into ligand-mediated regulation in antibiotic biosynthesis.

Regulation of Cannabinoid and Opioid Receptor Levels by Endogenous and Pharmacological Chaperones.

Cannabinoid and opioid receptor activities can be modulated by a variety of post-translational mechanisms including the formation of interacting complexes. This study examines the involvement of endogenous and exogenous chaperones in modulating the abundance and activity of cannabinoid CB1 receptor (CB1R), δ opioid receptor (DOR), and CB1R-DOR interacting complexes. Focusing on endogenous protein chaperones, namely receptor transporter proteins (RTPs), we examined relative mRNA expression in the mouse spinal cord and found RTP4 to be expressed at higher levels compared with other RTPs. Next, we assessed the effect of RTP4 on receptor abundance by manipulating RTP4 expression in cell lines. Overexpression of RTP4 causes an increase and knock-down causes a decrease in the levels of CB1R, DOR, and CB1R-DOR interacting complexes; this is accompanied by parallel changes in signaling. The ability of small molecule lipophilic ligands to function as exogenous chaperones was examined using receptor-selective antagonists. Long-term treatment leads to increases in receptor abundance and activity with no changes in mRNA supporting a role as pharmacological chaperones. Finally, the effect of cannabidiol (CBD), a small molecule ligand and a major active component of cannabis, on receptor abundance and activity in mice was examined. We find that CBD administration leads to increases in receptor abundance and activity in mouse spinal cord. Together, these results highlight a role for chaperones (proteins and small molecules) in modulating levels and activity of CB1R, DOR, and their interacting complexes potentially through mechanisms including receptor maturation and trafficking. SIGNIFICANCE STATEMENT: This study highlights a role for chaperones (endogenous and small membrane-permeable molecules) in modulating levels of cannabinoid CB1 receptor, delta opioid receptor, and their interacting complexes. These chaperones could be developed as therapeutics for pathologies involving these receptors.

Development of a fluorescent ligand that specifically binds to the c-MYC G-quadruplex by migrating the benzene group on a carbazole-benzothiazolium scaffold

c-MYC is one of the most important oncogenes, which is overexpressed in many cancers, and is highly related to development, metastasis, and drug resistance of cancers. The G4 structure in the promoter of c-MYC oncogene contributes a lot to the gene transcriptional mechanism. Small-molecule ligands binding to the c-MYC G4 appear to be a new class of anticancer agents. However, selective ligands for the c-MYC G4 over other G4s have been rarely reported. In this study, we reported a novel fluorescent ligand by migrating the benzene group on a carbazole-benzothiazolium scaffold, which was demonstrated to exhibit considerable specificity to the c-MYC G4, which was distinguished from other small-molecule ligands. The further cellular experiments suggested that this ligand may indeed target the promoter G4 and cause apparent transcriptional inhibition of the c-MYC oncogene instead of other G4-mediated oncogenes, which thereby resulted in cancer cell growth inhibition. Collectively, this study provided a good example for developing specific c-MYC G4 ligands, which may further develop into an effective anticancer agent that inhibit the c-MYC expression.

MultiModRLBP: A Deep Learning Approach for Multi-Modal RNA-Small Molecule Ligand Binding Sites Prediction.

This study aims to tackle the intricate challenge of predicting RNA-small molecule binding sites to explore the potential value in the field of RNA drug targets. To address this challenge, we propose the MultiModRLBP method, which integrates multi-modal features using deep learning algorithms. These features include 3D structural properties at the nucleotide base level of the RNA molecule, relational graphs based on overall RNA structure, and rich RNA semantic information. In our investigation, we gathered 851 interactions between RNA and small molecule ligand from the RNAglib dataset and RLBind training set. Unlike conventional training sets, this collection broadened its scope by including RNA complexes that have the same RNA sequence but change their respective binding sites due to structural differences or the presence of different ligands. This enhancement enables the MultiModRLBP model to more accurately capture subtle changes at the structural level, ultimately improving its ability to discern nuances among similar RNA conformations. Furthermore, we evaluated MultiModRLBP on two classic test sets, Test18 and Test3, highlighting its performance disparities on small molecules based on metal and non-metal ions. Additionally, we conducted a structural sensitivity analysis on specific complex categories, considering RNA instances with varying degrees of structural changes and whether they share the same ligands. The research results indicate that MultiModRLBP outperforms the current state-of-the-art methods on multiple classic test sets, particularly excelling in predicting binding sites for non-metal ions and instances where the binding sites are widely distributed along the sequence. MultiModRLBP also can be used as a potential tool when the RNA structure is perturbed or the RNA experimental tertiary structure is not available. Most importantly, MultiModRLBP exhibits the capability to distinguish binding characteristics of RNA that are structurally diverse yet exhibit sequence similarity. These advancements hold promise in reducing the costs associated with the development of RNA-targeted drugs.

MAGPIE: An interactive tool for visualizing and analyzing protein-ligand interactions.

Quantitative tools to compile and analyze biomolecular interactions among chemically diverse binding partners would improve therapeutic design and aid in studying molecular evolution. Here we present Mapping Areas of Genetic Parsimony In Epitopes (MAGPIE), a publicly available software package for simultaneously visualizing and analyzing thousands of interactions between a single protein or small molecule ligand (the "target") and all of its protein binding partners ("binders"). MAGPIE generates an interactive three-dimensional visualization from a set of protein complex structures that share the target ligand, as well as sequence logo-style amino acid frequency graphs that show all the amino acids from the set of protein binders that interact with user-defined target ligand positions or chemical groups. MAGPIE highlights all the salt bridge and hydrogen bond interactions made by the target in the visualization and as separate amino acid frequency graphs. Finally, MAGPIE collates the most common target-binder interactions as a list of "hotspots," which can be used to analyze trends or guide the de novo design of protein binders. As an example of the utility of the program, we used MAGPIE to probe how different antibody fragments bind a viral antigen; how a common metabolite binds diverse protein partners; and how two ligands bind orthologs of a well-conserved glycolytic enzyme for a detailed understanding of evolutionarily conserved interactions involved in its activation and inhibition. MAGPIE is implemented in Python 3 and freely available at https://github.com/glasgowlab/MAGPIE, along with sample datasets, usage examples, and helper scripts to prepare input structures.

Isoquinoline small molecule ligands are agonists and probe-dependent allosteric modulators of the glucagon subfamily of GPCRs

Class B1 G protein-coupled receptors (GPCRs) are peptide hormone receptors and well validated therapeutic targets, however development of non-peptide drugs targeting this class of receptors is challenging. Recently, a series of isoquinoline-based derivates were reported in the patent literature as allosteric ligands for the glucagon receptor subfamily, and two compounds, LSN3451217 and LSN3556672, were used to facilitate structural studies with the glucagon-like peptide-1 receptor (GLP-1R) and glucose dependent insulinotropic peptide receptor (GIPR) bound to orthosteric agonists. Here we pharmacologically characterized stereoisomers of LSN3451217 and LSN3556672, across the class B1 GPCR family. This revealed LSN3556672 isomers are agonists for the glucagon receptor (GCGR), GLP-1R, GIPR and the calcitonin receptor (CTR), albeit the degree of agonism varied at each receptor. In contrast, LSN3451217 isomers were more selective agonists at the GLP-1R, with lower potency at the GCGR and CTR and no activity at the GIPR. All compounds also modulated peptide-mediated cyclic adenosine monophosphate (cAMP) signaling at the GIPR, and to a lesser extent the GLP-1R, in a probe-dependent manner, with modest positive allosteric modulation observed for some peptides, and negligible effects observed with other peptides. In contrast neutral or weak negative/positive allosteric modulation was observed with peptides assessed at the GCGR and CTR. This study expands our knowledge on class B1 GPCR allosteric modulation and may have implications for future structural and drug discovery efforts targeting the class B1 GPCR subfamily.

RiboScreenTM Technology Delivers a Ribosomal Target and a Small-Molecule Ligand for Ribosome Editing to Boost the Production Levels of Tropoelastin, the Monomeric Unit of Elastin.

Elastin, a key structural protein essential for the elasticity of the skin and elastogenic tissues, degrades with age. Replenishing elastin holds promise for anti-aging cosmetics and the supplementation of elastic activities of the cardiovascular system. We employed RiboScreenTM, a technology for identifying molecules that enhance the production of specific proteins, to target the production of tropoelastin. We make use of RiboScreenTM in two crucial steps: first, to pinpoint a target ribosomal protein (TRP), which acts as a switch to increase the production of the protein of interest (POI), and second, to identify small molecules that activate this ribosomal protein switch. Using RiboScreenTM, we identified ribosomal protein L40, henceforth eL40, as a TRP switch to boost tropoelastin production. Drug discovery identified a small-molecule hit that binds to eL40. In-cell treatment demonstrated activity of the eL40 ligand and delivered increased tropoelastin production levels in a dose-dependent manner. Thus, we demonstrate that RiboScreenTM can successfully identify a small-molecule hit capable of selectively enhancing tropoelastin production. This compound has the potential to be developed for topical or systemic applications to promote skin rejuvenation and to supplement elastic functionality within the cardiovascular system.

A novel heuristic of rigid docking scores positively correlates with full-length nuclear receptor LRH-1 regulation

The nuclear receptor Liver Receptor Homolog-1 (LRH-1, NR5A2) is a ligand-regulated transcription factor and validated drug target for several human diseases. LRH-1 activation is regulated by small molecule ligands, which bind to the ligand binding domain (LBD) within the full-length LRH-1. We recently identified 57 compounds that bind LRH-1, and unexpectedly found these compounds regulated either the isolated LBD, or the full-length LRH-1 in cells, with little overlap. Here, we correlated compound binding energy from a single rigid-body scoring function with full-length LRH-1 activity in cells. Although docking scores of the 57 hit compounds did not correlate with LRH-1 regulation in wet lab assays, a subset of the compounds had large differences in binding energy docked to the isolated LBD vs. full-length LRH-1, which we used to empirically derive a new metric of the docking scores we call "ΔΔG". Initial regressions, correlations and contingency analyses all suggest compounds with high ΔΔG values more frequently regulated LRH-1 in wet lab assays. We then docked all 57 compounds to 18 separate crystal structures of LRH-1 to obtain averaged ΔΔG values for each compound, which robustly and reproducibly associated with full-length LRH-1 activity in cells. Network analyses on the 18 crystal structures of LRH-1 suggest unique communication paths exist between the subsets of LRH-1 crystal structures that produced high vs. low ΔΔG values, identifying a structural relationship between ΔΔG and the position of Helix 6, a previously established regulatory helix important for LRH-1 regulation. Together, these data suggest rigid-body computational docking can be used to quickly calculate ΔΔG, which positively correlated with the ability of these 57 hit compounds to regulate full-length LRH-1 in cell-based assays. We propose ΔΔG as a novel computational tool that can be applied to LRH-1 drug screens to prioritize compounds for resource-intense secondary screening.

A comprehensive exploration of the druggable conformational space of protein kinases using AI-predicted structures.

Protein kinase function and interactions with drugs are controlled in part by the movement of the DFG and ɑC-Helix motifs that are related to the catalytic activity of the kinase. Small molecule ligands elicit therapeutic effects with distinct selectivity profiles and residence times that often depend on the active or inactive kinase conformation(s) they bind. Modern AI-based structural modeling methods have the potential to expand upon the limited availability of experimentally determined kinase structures in inactive states. Here, we first explored the conformational space of kinases in the PDB and models generated by AlphaFold2 (AF2) and ESMFold, two prominent AI-based protein structure prediction methods. Our investigation of AF2's ability to explore the conformational diversity of the kinome at various multiple sequence alignment (MSA) depths showed a bias within the predicted structures of kinases in DFG-in conformations, particularly those controlled by the DFG motif, based on their overabundance in the PDB. We demonstrate that predicting kinase structures using AF2 at lower MSA depths explored these alternative conformations more extensively, including identifying previously unobserved conformations for 398 kinases. Ligand enrichment analyses for 23 kinases showed that, on average, docked models distinguished between active molecules and decoys better than random (average AUC (avgAUC) of 64.58), but select models perform well (e.g., avgAUCs for PTK2 and JAK2 were 79.28 and 80.16, respectively). Further analysis explained the ligand enrichment discrepancy between low- and high-performing kinase models as binding site occlusions that would preclude docking. The overall results of our analyses suggested that, although AF2 explored previously uncharted regions of the kinase conformational space and select models exhibited enrichment scores suitable for rational drug discovery, rigorous refinement of AF2 models is likely still necessary for drug discovery campaigns.

Ligand Gaussian Accelerated Molecular Dynamics 3 (LiGaMD3): Improved Calculations of Binding Thermodynamics and Kinetics of Both Small Molecules and Flexible Peptides.

Binding thermodynamics and kinetics play critical roles in drug design. However, it has proven challenging to efficiently predict ligand binding thermodynamics and kinetics of small molecules and flexible peptides using conventional molecular dynamics (cMD), due to limited simulation time scales. Based on our previously developed ligand Gaussian accelerated molecular dynamics (LiGaMD) method, we present a new approach, termed "LiGaMD3″, in which we introduce triple boosts into three individual energy terms that play important roles in small-molecule/peptide dissociation, rebinding, and system conformational changes to improve the sampling efficiency of small-molecule/peptide interactions with target proteins. To validate the performance of LiGaMD3, MDM2 bound by a small molecule (Nutlin 3) and two highly flexible peptides (PMI and P53) were chosen as the model systems. LiGaMD3 could efficiently capture repetitive small-molecule/peptide dissociation and binding events within 2 μs simulations. The predicted binding kinetic constant rates and free energies from LiGaMD3 were in agreement with the available experimental values and previous simulation results. Therefore, LiGaMD3 provides a more general and efficient approach to capture dissociation and binding of both small-molecule ligands and flexible peptides, allowing for accurate prediction of their binding thermodynamics and kinetics.

GOLEM: Automated and Robust Cryo-EM-Guided Ligand Docking with Explicit Water Molecules.

A detailed understanding of ligand-protein interaction is essential for developing rational drug-design strategies. In recent years, technological advances in cryo-electron microscopy (cryo-EM) brought a new era to the structural determination of biological macromolecules and assemblies at high resolution, marking cryo-EM as a promising tool for studying ligand-protein interactions. However, even in high-resolution cryo-EM results, the densities for the bound small-molecule ligands are often of lower quality due to their relatively dynamic and flexible nature, frustrating their accurate coordinate assignment. To address the challenge of ligand modeling in cryo-EM maps, here we report the development of GOLEM (Genetic Optimization of Ligands in Experimental Maps), an automated and robust ligand docking method that predicts a ligand's pose and conformation in cryo-EM maps. GOLEM employs a Lamarckian genetic algorithm to perform a hybrid global/local search for exploring the ligand's conformational, orientational, and positional space. As an important feature, GOLEM explicitly considers water molecules and places them at optimal positions and orientations. GOLEM takes into account both molecular energetics and the correlation with the cryo-EM maps in its scoring function to optimally place the ligand. We have validated GOLEM against multiple cryo-EM structures with a wide range of map resolutions and ligand types, returning ligand poses in excellent agreement with the densities. As a VMD plugin, GOLEM is free of charge and accessible to the community. With these features, GOLEM will provide a valuable tool for ligand modeling in cryo-EM efforts toward drug discovery.