Druggable Binding Sites Research Articles

Exploring Druggable Binding Sites on the Class A GPCRs Using the Residue Interaction Network and Site Identification by Ligand Competitive Saturation.

G protein-coupled receptors (GPCRs) play a central role in cellular signaling and are linked to many diseases. Accordingly, computational methods to explore potential allosteric sites for this class of proteins to facilitate the identification of potential modulators are needed. Importantly, the availability of rich structural data providing the locations of the orthosteric ligands and allosteric modulators targeting different GPCRs allows for the validation of approaches to identify new allosteric binding sites. Here, we validate the combination of two computational techniques, the residue interaction network (RIN) model and the site identification by ligand competitive saturation (SILCS) method, to predict putative allosteric binding sites of class A GPCRs. RIN analysis identifies hub residues that mediate allosteric signaling within a receptor and have a high capacity to alter receptor dynamics upon ligand binding. The known orthosteric (and allosteric) binding sites of 18 distinct class A GPCRs were successfully predicted by RIN through a dataset of 105 crystal structures (91 ligand-bound, 14 unbound) with up to 77.8% (76.9%) sensitivity, 92.5% (95.3%) specificity, 51.9% (50%) precision, and 86.2% (92.4%) accuracy based on the experimental and theoretical binding site data. Moreover, graph spectral analysis of the residue networks revealed that the proposed sites were located at the interfaces of highly interconnected residue clusters with a high ability to coordinate the functional dynamics. Then, we employed the SILCS-Hotspots method to assess the druggability of the novel sites predicted for 7 distinct class A GPCRs that are critical for a variety of diseases. While the known orthosteric and allosteric binding sites are successfully explored by our approach, numerous putative allosteric sites with the potential to bind drug-like molecules are proposed. The computational approach presented here promises to be a highly effective tool to predict putative allosteric sites of GPCRs to facilitate the design of effective modulators.

Combined Physics- and Machine-Learning-Based Method to Identify Druggable Binding Sites Using SILCS-Hotspots.

Identifying druggable binding sites on proteins is an important and challenging problem, particularly for cryptic, allosteric binding sites that may not be obvious from X-ray, cryo-EM, or predicted structures. The Site-Identification by Ligand Competitive Saturation (SILCS) method accounts for the flexibility of the target protein using all-atom molecular simulations that include various small molecule solutes in aqueous solution. During the simulations, the combination of protein flexibility and comprehensive sampling of the water and solute spatial distributions can identify buried binding pockets absent in experimentally determined structures. Previously, we reported a method for leveraging the information in the SILCS sampling to identify binding sites (termed Hotspots) of small mono- or bicyclic compounds, a subset of which coincide with known binding sites of drug-like molecules. Here, we build on that physics-based approach and present a ML model for ranking the Hotspots according to the likelihood they can accommodate drug-like molecules (e.g., molecular weight >200 Da). In the independent validation set, which includes various enzymes and receptors, our model recalls 67% and 89% of experimentally validated ligand binding sites in the top 10 and 20 ranked Hotspots, respectively. Furthermore, we show that the model's output Decision Function is a useful metric to predict binding sites and their potential druggability in new targets. Given the utility the SILCS method for ligand discovery and optimization, the tools presented represent an important advancement in the identification of orthosteric and allosteric binding sites and the discovery of drug-like molecules targeting those sites.

CC chemokine receptor 2 is allosterically modulated by sodium ions and amiloride derivatives through a distinct sodium ion binding site

CC chemokine receptor 2 and CCL2 are highly involved in cancer growth and metastasis, and immune escape. Raised sodium ion concentrations in solid tumours have also been correlated to metastasis and immune modulation. Sodium ions can modulate class A G protein-coupled receptors through the sodium ion binding site characterized by a highly conserved aspartic acid residue (D2.50), also present in CCR2. Hence, we further explored this binding site in CCR2 by radioligand binding studies and mutagenesis. Modulation of three distinctly binding radioligands by sodium ions and amiloride derivates was investigated. Sodium ions were observed to be relatively weak modulators of antagonist binding, but substantially increased 125I-CCL2 dissociation from CCR2. 6-Substituted Hexamethylene Amiloride (HMA) modulated all tested radioligands. Induced-fit docking of HMA in the presumed sodium ion binding site of CCR2 confirmed its binding site. Finally, investigation of (cancer-associated) mutations in the sodium ion binding site showed a markedly decreased expression compared to wild type. Only two mutants, G123A3.35 and G127K3.39, were able to be bound by [3H]INCB3344 and [3H]CCR2-RA-[R]. Thus, mutagenesis showed that the sodium ion binding site residues, which are distinct from other class A GPCRs and related to chemokine receptor evolution, are crucial for receptor integrity. Moreover, the tested mutations appeared to have no effect on modulation observed by HMA or a minor effect on sodium chloride modulation on the tested radioligands. All in all, these results invite further exploration of the CCR2 sodium ion binding site in (cancer) biology, and potentially as a third druggable binding site.

Unified Aedes aegypti Protein Resource Database (UAAPRD): An Integrated High-Throughput In Silico Platform for Comprehensive Protein Structure Modeling and Functional Target Analysis to Enhance Vector Control Strategies.

A comprehensive examination of Aedes aegypti's proteome to detect key proteins that can be targeted with small molecules can disrupt blood feeding and disease transmission. However, research currently only focuses on finding repellent-like compounds, limiting studies on identifying unexplored proteins in its proteome. High-throughput analysis generates vast amounts of data, raising concerns about accessibility and usability. Establishing a dedicated database is a solution, centralizing information on identified proteins, functions, and modeled structures for easy access and research. This study focuses on scrutinizing key proteins in A. aegypti, modeling their structures using RaptorX standalone tool, identification of druggable binding sites using BiteNet, validating the models via Ramachandran plot studies and refining them via 50-ns molecular dynamic simulations using Schrodinger Maestro. By analyzing ~ 18k proteins in the proteome of A. aegypti in our previous studies, all proteins involved in the light and dark circadian rhythm of the mosquito, inclusive of proteins in blood feeding, metabolism, etc. were chosen for the current study. The outcome is UAAPRD, a unique repository housing information on hundreds of previously unmodeled and un-simulated mosquito proteins. This robust MYSQL database ( https://uaaprd.onrender.com/user ) houses data on 309 modeled & simulated proteins of A. aegypti. It allows users to obtain protein data, view evolutionary analysis data of the protein categories, visualize proteins of interest, and send request to screen against the pharmacophore models present in UAAPRD against ligand of interest. This study offers crucial insights for developing targeted studies, which will ultimately contribute to more effective vector control strategies.

Deep attention network for identifying ligand-protein binding sites

One of the critical aspects of structure-based drug design is to choose important druggable binding sites in the protein's crystallography structures. As experimental processes are costly and time-consuming, computational drug design using machine learning algorithms is recommended. Over recent years, deep learning methods have been utilized in a wide variety of research applications such as binding site prediction. In this study, a new combination of attention blocks in the 3D U-Net model based on semantic segmentation methods is used to improve localization of pocket prediction. The attention blocks are tuned to find which point and channel of features should be emphasized along spatial and channel axes. Our model's performance is evaluated through extensive experiments on several datasets from different sources, and the results are compared to the most recent deep learning-based models. The results indicate the proposed attention model (Att-UNet) can predict binding sites accurately, i.e. the overlap of the predicted pocket using the proposed method with the true binding site shows statistically significant improvement when compared to other state-of-the-art models. The attention blocks may help the model focus on the target structure by suppressing features in irrelevant regions.

Identification of Druggable Binding Sites and Small Molecules as Modulators of TMC1.

Our ability to hear and maintain balance relies on the proper functioning of inner ear sensory hair cells, which translate mechanical stimuli into electrical signals via mechano-electrical transducer (MET) channels, composed of TMC1/2 proteins. However, the therapeutic use of ototoxic drugs, such as aminoglycosides and cisplatin, which can enter hair cells through MET channels, often leads to profound auditory and vestibular dysfunction. Despite extensive research on otoprotective compounds targeting MET channels, our understanding of how small molecule modulators interact with these channels remains limited, hampering the discovery of novel compounds. Here, we propose a structure-based screening approach, integrating 3D-pharmacophore modeling, molecular simulations, and experimental validation. Our pipeline successfully identified several novel compounds and FDA-approved drugs that reduced dye uptake in cultured cochlear explants, indicating MET modulation activity. Molecular docking and free-energy estimations for binding allowed us to identify three potential drug binding sites within the channel pore, phospholipids, and key amino acids involved in modulator interactions. We also identified shared ligand-binding features between TMC and structurally related TMEM16 protein families, providing novel insights into their distinct inhibition, while potentially guiding the rational design of MET-channel-specific modulators. Our pipeline offers a broad application to discover small molecule modulators for a wide spectrum of mechanosensitive ion channels.

Abstract 7073: Integrating fragment-based crystallographic screening with dTAG-PROTAC degradation and genetic rescue to explore the function of eIF4E

Abstract The translation initiation factor eIF4E is a rate-limiting factor for protein synthesis that binds the mRNA m7G-cap to initiate the recruitment and binding of eIF4F components such as eIF4G. Targeting eIF4E has long been considered a promising anticancer strategy but it has remained undruggable using conventional screening approaches. Fragment-based crystallographic screening is a powerful technique for probing the surface of proteins to identify potentially druggable binding sites. Here, we describe fragment-based screening using a combination of Xray crystallography and NMR to identify low affinity fragment hits binding to the mRNA cap-binding site or at an additional site of unknown functional relevance (site 2). Subsequent rounds of iterative structure-based design yielded a lead molecule, with 100nM binding potency for site 2, that could disrupt the eIF4E:eIF4G interaction and inhibit cap-dependent translation in cell lysates. To explore the functional relevance of site 2 we established a cell model expressing an eIF4E-FKBP12F36V fusion protein that was sensitive to degradation following treatment with the heterobifunctional dTAGV-1 molecule. Degradation of eIF4E resulted in inhibition of cell growth and reduced expression of MCL1 protein, that were rescued by re-expression of a non-degradable wild-type eIF4E. A mutation reported to block canonical eIF4G binding (W73F) disrupted binding to eIF4G but to our surprise retained the ability to rescue the eIF4E degradation phenotype. Mutation of site 2 (L85R and L134R) also disrupted the eIF4E:eIF4G interaction. The L85R mutant remained able to partially rescue the cellular responses, in contrast to the L134R mutant that could not rescue eIF4E loss. The W73F/L85R double mutant resulted in a combinatorial loss of eIF4E function. Despite disrupting eIF4E function in lysates, there was a significant drop-off of compound activity in cells. However, treatment of the W73F mutant with compound recapitulated the combination of the W73F canonical eIF4G binding mutation with the L85R site 2 mutation. This approach demonstrates the power of coupling fragment screening with target-degradation and genetic rescue approaches to find and explore novel functional pockets. Our data suggest it may be necessary to disrupt the extended protein-protein interaction made by both the canonical and non-canonical regions of eIF4E to drive a strong functional effect in cells. However, the discovery of a lead compound, the associated structural understanding, and the knowledge that one site 2 mutation inactivates eIF4E provides hope that a significantly more potent small molecule inhibitor or degrader could drive a more profound cellular effect which may lead to improved cellular activity and a deeper understanding of eIF4E biology. Citation Format: Paul A. Clarke, Swee Y. Sharp, Marianna Martella, Christopher I. Milton, Charlotte East, Mladen Vinkovic, Nicola Wallis, George Ward, Caroline Richardson, Andrew Woodhead. Integrating fragment-based crystallographic screening with dTAG-PROTAC degradation and genetic rescue to explore the function of eIF4E [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 7073.

Graph Attention Site Prediction (GrASP): Identifying Druggable Binding Sites Using Graph Neural Networks with Attention.

Identifying and discovering druggable protein binding sites is an important early step in computer-aided drug discovery, but it remains a difficult task where most campaigns rely on a priori knowledge of binding sites from experiments. Here, we present a binding site prediction method called Graph Attention Site Prediction (GrASP) and re-evaluate assumptions in nearly every step in the site prediction workflow from data set preparation to model evaluation. GrASP is able to achieve state-of-the-art performance at recovering binding sites in PDB structures while maintaining a high degree of precision which will minimize wasted computation in downstream tasks such as docking and free energy perturbation.

Understanding the mechanisms of green tea EGCG against amyloid β oligomer neurotoxicity through computational studies.

Oligomeric species of amyloid β peptide (Aβ) are pivotal in Alzheimer's disease (AD) pathogenesis, making them valuable therapeutic targets. Currently, there is no cure or preventive therapy available for AD, with only a few therapeutics offering temporary alleviation of symptoms. Natural products (NPs) are now considered promising anti-amyloid agents. Green tea catechins have garnered considerable attention due to their ability to remodel the toxic amyloid β peptide oligomers (AβOs) into non-toxic assemblies. Nevertheless, the precise molecular mechanism underlying their effects on AβOs remains unclear. In this study, we employ a combination of binding site prediction, molecular docking, and dynamics simulations to gain mechanistic insights into the binding of the potent anti-amyloid epigallocatechin-3-gallate (EGCG) and the less effective catechin, epicatechin (EC), on the structure of pore-forming Aβ tetramers (PDB ID 6RHY). This recently elucidated structure represents AβO(1-42) with two faces of the hydrophobic β-sheet core and hydrophilic edges. Our simulations revealed three potential druggable binding sites within the AβO: two in hydrophilic edges and one in the β-sheet core. Although both catechins bind via hydrogen bond (H-bond) and aromatic interactions to the three potential binding sites, EGCG interacted with key residues more efficiently than EC. We propose that EGCG may remodel AβOs preventing pore formation by binding to the hydrophilic edge binding sites. Additionally, EGCG interacts with key residues in the oligomer's β-sheet core binding site, crucial for fibrillar aggregation. A better understanding of how anti-amyloid compounds remodelling AβOs could be valuable for the development of new therapeutic strategies targeting Aβ in AD. Further experimental validation using point mutations involving key residues could be useful to define whether the establishment of these interactions is crucial for the EGCG remodelling effect.

Abstract A142: Fragment based discovery of inhibitors of the eIF4E:eIF4G interaction

Abstract eIF4E is the rate limiting factor for cap-dependent protein translation, forming an essential part of the translation initiation complex, eIF4F and is a key signalling convergence point for the major oncogenic pathways MAPK and PI3K/AKT/MTOR. eIF4E has been shown to be overexpressed in cancer, which has been associated with poor prognosis and resistance to chemotherapeutics and targeted agents such as mTOR and MAPK inhibitors. eIF4E is therefore a promising target for cancer treatment but has so far been difficult to target with small molecules due to a lack of druggable pockets. Fragment based crystallographic screening is a powerful technique for probing the surface of proteins to identify potentially druggable binding sites. Here, we describe our fragment-based approach and hit optimisation campaign to generate potent compounds binding to eIF4E For the fragment screen, we generated an engineered form of eIF4E, where the N-terminal (residues 1-35) were removed and replaced a short, tethered peptide based on the canonical binding sequence of the eIF4E binding protein 4E-BP1, and screened 1400 fragments using a combination of Xray crystallography and NMR. Several low affinity (mM) fragment hits were identified, binding to both the mRNA cap- binding site and at an additional binding site of unknown functional relevance. Using a combination of biophysical and biochemical assays, subsequent rounds of iterative structure-based drug design of yielded potent lead molecules against both binding sites. For the cap-binding site, lead compounds of low nM potency were generated but due to the polar nature of the binding site, the molecules had very poor permeability properties and did not inhibit cell proliferation. For the non-canonical binding site, lead compounds with potency of 100nM alongside less potent enantiomers were generated which provided useful tools for probing the function of the binding site in vitro. We developed a MSD (Meso-scale discovery) assay to determine that the lead compound and not the inactive enantiomer could disrupt the eIF4E:eIF4G interaction in HeLa cell lysates with an IC50 of 1uM. Using a HeLa extract in vitro translation assay, we could also demonstrate that both the cap site and non-canonical site lead compounds could inhibit cap-dependent but not IRES driven protein translation. This work demonstrates the power of fragment screening to identify novel functional pockets on difficult to drug proteins such as eIF4E. Citation Format: Caroline J Richardson, Mladen Vincovic, Charlotte East, Nicola Wallis, George Ward, Andrew Woodhead. Fragment based discovery of inhibitors of the eIF4E:eIF4G interaction [abstract]. In: Proceedings of the AACR-NCI-EORTC Virtual International Conference on Molecular Targets and Cancer Therapeutics; 2023 Oct 11-15; Boston, MA. Philadelphia (PA): AACR; Mol Cancer Ther 2023;22(12 Suppl):Abstract nr A142.

Probing a distinct druggable tubulin binding site with gatorbulins 1–7, their metabolic and physicochemical properties, and pharmacological consequences

Microtubules, consisting of α/β-tubulin heterodimers, are prime targets for anticancer drug discovery. Gatorbulin-1 (GB1, 1a) is a recently described marine natural product that targets tubulin at a new, seventh pharmacological site at the tubulin intradimer interface. Using our previously developed robust route towards GB1 (1a), we synthesized simplified, first-generation gatorbulins, GB2–7 (1b–1g) of this highly modified cyclodepsipeptide (GB1) that does not contain any proteinogenic amino acid. We systematically investigated the structure-activity relationship at the biochemical and cellular level using GB1-susceptible ovarian and cervical cancer cells. We validated that the hydroxamate moiety in the N-methyl-alanine residue is critical for activity. All other structural modifications present in GB1, including C-hydroxylation of asparagine, methylation at C-4 of proline, and sp2 hybridization in dehydro-alanine, were proven to be functionally relevant. Replacement of the primary amide with a methyl ester also resulted in reduced activity, indicating the intricate scaffold optimization by the GB1-producing cyanobacterium. Inhibition of tubulin polymerization in vitro and binding affinities correlated very well, translating into differentials in cellular efficacy. We used docking and molecular dynamics to evaluate the effects of the chemical simplification at the structural level, indicating that each modification resulted in loss of target interactions, although energetically modest. Similar to cevipabulin that targets two different sites on the tubulin dimer, GB1 promotes proteasome-mediated tubulin degradation but by an unknown mechanism, presumably distinct from that of cevipabulin. Comparison with cevipabulin indicated that this compound binds to the same tubulin region as GB1 (1a), although the binding mode is distinct. Cevipabulin almost exclusively interacts with α-tubulin, including nonexchangeable GTP. In contrast, GB1 (1a) makes extensive contact and hydrogen bonds with both α- and β-chains of tubulin. GB1-7 showed excellent solubility and much higher than that of paclitaxel. Hepatic microsome stability was excellent, human cytochrome P450s were not inhibited and plasma binding was minimal with high free fractions. Passive permeability was predicted to be high based on PAMPA. Parent compound GB1 (1a) was further evaluated using a cellular model with MDCK cells stably transduced with the human efflux transporter MDR1/P-gp, showing similar permeability with and against transporter gradient, indicating that GB1 (1a) is a poor P-gp substrate.

Suramin inhibits SARS-CoV-2 nucleocapsid phosphoprotein genome packaging function

The coronavirus disease 2019 (COVID-19) pandemic is fading, however its etiologic agent severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) continues posing - despite the availability of licensed vaccines – a global health threat, due to the potential emergence of vaccine-resistant SARS-CoV-2 variants. This makes the development of new drugs against COVID-19 a persistent urgency and sets as research priority the validation of novel therapeutic targets within the SARS-CoV-2 proteome. Among these, a promising one is the SARS-CoV-2 nucleocapsid (N) phosphoprotein, a major structural component of the virion with indispensable role in packaging the viral genome into a ribonucleoprotein (RNP) complex, which also contributes to SARS-CoV-2 innate immune evasion by inhibiting the host cell type-I interferon (IFN-I) response. By combining miniaturized differential scanning fluorimetry with microscale thermophoresis, we found that the 100-year-old drug Suramin interacts with SARS-CoV-2 N-terminal domain (NTD) and C-terminal domain (CTD), thereby inhibiting their single-stranded RNA (ssRNA) binding function with low-micromolar Kd and IC50 values. Molecular docking suggests that Suramin interacts with basic NTD cleft and CTD dimer interface groove, highlighting three potentially druggable ssRNA binding sites. Electron microscopy shows that Suramin inhibits the formation in vitro of RNP complex-like condensates by SARS-CoV-2 N with a synthetic ssRNA. In a dose-dependent manner, Suramin also reduced SARS-CoV-2-induced cytopathic effect on Vero E6 and Calu-3 cells, partially reverting the SARS-CoV-2 N-inhibited IFN-I production in 293T cells. Our findings indicate that Suramin inhibits SARS-CoV-2 replication by hampering viral genome packaging, thereby representing a starting model for design of new COVID-19 antivirals.

Identification of Potent Inhibitors Targeting EGFR and HER3 for Effective Treatment of Chemoresistance in Non-Small Cell Lung Cancer.

Non-small cell lung cancer (NSCLC) is the most common form of lung cancer. Despite the existence of various therapeutic options, NSCLC is still a major health concern due to its aggressive nature and high mutation rate. Consequently, HER3 has been selected as a target protein along with EGFR because of its limited tyrosine kinase activity and ability to activate PI3/AKT pathway responsible for therapy failure. We herein used a BioSolveIT suite to identify potent inhibitors of EGFR and HER3. The schematic process involves screening of databases for constructing compound library comprising of 903 synthetic compounds (602 for EGFR and 301 for HER3) followed by pharmacophore modeling. The best docked poses of compounds with the druggable binding site of respective proteins were selected according to pharmacophore designed by SeeSAR version 12.1.0. Subsequently, preclinical analysis was performed via an online server SwissADME and potent inhibitors were selected. Compound 4k and 4m were the most potent inhibitors of EGFR while 7x effectively inhibited the binding site of HER3. The binding energies of 4k, 4m, and 7x were -7.7, -6.3 and -5.7 kcal/mol, respectively. Collectively, 4k, 4m and 7x showed favorable interactions with the most druggable binding sites of their respective proteins. Finally, in silico pre-clinical testing by SwissADME validated the non-toxic nature of compounds 4k, 4m and 7x providing a promising treatment option for chemoresistant NSCLC.

Identifying promising druggable binding sites and their flexibility to target the receptor-binding domain of SARS-CoV-2 spike protein

The spike protein of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is crucial for viral infection. The interaction of its receptor-binding domain (RBD) with the human angiotensin-converting enzyme 2 (ACE2) protein is required for the virus to enter the host cell. We identified RBD binding sites to block its function with inhibitors by combining the protein structural flexibility with machine learning analysis. Molecular dynamics simulations were performed on unbound or ACE2-bound RBD conformations. Pockets estimation, tracking and druggability prediction were performed on a large sample of simulated RBD conformations. Recurrent druggable binding sites and their key residues were identified by clustering pockets based on their residue similarity. This protocol successfully identified three druggable sites and their key residues, aiming to target with inhibitors for preventing ACE2 interaction. One site features key residues for direct ACE2 interaction, highlighted using energetic computations, but can be affected by several mutations of the variants of concern. Two highly druggable sites, located between the spike protein monomers interface are promising. One weakly impacted by only one Omicron mutation, could contribute to stabilizing the spike protein in its closed state. The other, currently not affected by mutations, could avoid the activation of the spike protein trimer.

Revealing druggable cryptic pockets in the Nsp1 of SARS-CoV-2 and other β-coronaviruses by simulations and crystallography.

Non-structural protein 1 (Nsp1) is a main pathogenicity factor of α- and β-coronaviruses. Nsp1 of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) suppresses the host gene expression by sterically blocking 40S host ribosomal subunits and promoting host mRNA degradation. This mechanism leads to the downregulation of the translation-mediated innate immune response in host cells, ultimately mediating the observed immune evasion capabilities of SARS-CoV-2. Here, by combining extensive molecular dynamics simulations, fragment screening and crystallography, we reveal druggable pockets in Nsp1. Structural and computational solvent mapping analyses indicate the partial crypticity of these newly discovered and druggable binding sites. The results of fragment-based screening via X-ray crystallography confirm the druggability of the major pocket of Nsp1. Finally, we show how the targeting of this pocket could disrupt the Nsp1-mRNA complex and open a novel avenue to design new inhibitors for other Nsp1s present in homologous β-coronaviruses.

ProBiS-Fold Approach for Annotation of Human Structures from the AlphaFold Database with No Corresponding Structure in the PDB to Discover New Druggable Binding Sites.

ProBiS (Protein Binding Sites), a local structure-based comparison algorithm, is used in the new ProBiS-Fold web server to annotate human structures from the AlphaFold database without a corresponding structure in the Protein Data Bank (PDB) to discover new druggable binding sites. The ProBiS algorithm is used to compare each query protein structure predicted by the AlphaFold approach with the protein structures from the PDB to identify similarities between known binding sites found in the PDB and yet unknown binding sites in the AlphaFold database. Ligands bound in these identified similar PDB sites are then transferred to each query protein from the AlphaFold database, and binding sites are identified as ligand clusters on an AlphaFold protein. Small molecule binding sites are assigned druggability scores based on the similarity of their ligands to known drugs, allowing them to be ranked according to their perceived and actual potential for drug development. ProBiS-Fold provides interactive and downloadable binding sites for the entire human structural proteome, including more than 3000 new druggable binding sites that have no corresponding structure in the PDB, taking into account AlphaFold's model quality, to enable protein structure-function relationship studies and pharmaceutical drug discovery research. The web server is freely accessible to academic users at http://probis-fold.insilab.org.

Virtual Drug Repositioning as a Tool to Identify Natural Small Molecules That Synergize with Lumacaftor in F508del-CFTR Binding and Rescuing

Cystic fibrosis is a hereditary disease mainly caused by the deletion of the Phe 508 (F508del) of the cystic fibrosis transmembrane conductance regulator (CFTR) protein that is thus withheld in the endoplasmic reticulum and rapidly degraded by the ubiquitin/proteasome system. Cystic fibrosis remains a potentially fatal disease, but it has become treatable as a chronic condition due to some CFTR-rescuing drugs that, when used in combination, increase in their therapeutic effect due to a synergic action. Also, dietary supplementation of natural compounds in combination with approved drugs could represent a promising strategy to further alleviate cystic fibrosis symptoms. On these bases, we screened by in silico drug repositioning 846 small synthetic or natural compounds from the AIFA database to evaluate their capacity to interact with the highly druggable lumacaftor binding site of F508del-CFTR. Among the identified hits, nicotinamide (NAM) was predicted to accommodate into the lumacaftor binding region of F508del-CFTR without competing against the drug but rather stabilizing its binding. The effective capacity of NAM to bind F508del-CFTR in a lumacaftor-uncompetitive manner was then validated experimentally by surface plasmon resonance analysis. Finally, the capacity of NAM to synergize with lumacaftor increasing its CFTR-rescuing activity was demonstrated in cell-based assays. This study suggests the possible identification of natural small molecules devoid of side effects and endowed with the capacity to synergize with drugs currently employed for the treatment of cystic fibrosis, which hopefully will increase the therapeutic efficacy with lower doses.

Identification of a binding site on soluble RANKL that can be targeted to inhibit soluble RANK-RANKL interactions and treat osteoporosis

One of the major challenges for discovering protein-protein interaction inhibitors is identifying selective and druggable binding sites at the protein surface. Here, we report an approach to identify a small molecular binding site to selectively inhibit the interaction of soluble RANKL and RANK for designing anti-osteoporosis drugs without undesirable immunosuppressive effects. Through molecular dynamic simulations, we discovered a binding site that allows a small molecule to selectively interrupt soluble RANKL-RANK interaction and without interfering with the membrane RANKL-RANK interaction. We describe a highly potent inhibitor, S3-15, and demonstrate its specificity to inhibit the soluble RANKL-RANK interaction with in vitro and in vivo studies. S3-15 exhibits anti-osteoporotic effects without causing immunosuppression. Through in silico and in vitro experiments we further confirm the binding model of S3-15 and soluble RANKL. This work might inspire structure-based drug discovery for targeting protein-protein interactions.

Targeting tribbles homolog 3 (TRIB3) protein against type 2 diabetes for the identification of potential inhibitors by in silico screening

Tribbles homolog 3 (TRIB3) protein is inhibiting the insulin signaling by directly binding to the Akt/PKB leading to insulin resistance in the pancreas causing type 2 diabetes mellitus. Hence, TRIB3 protein is considered as a possible drug target for the new lead identification against type 2 diabetes. In the present study, the homology model of TRIB3 protein was generated to explore its biochemical function and molecular interactions in the new lead identification. The energy minimization of TRIB3 protein was carried out and evaluated by validation protocols for structure reliability. The druggable binding site of TRIB3 protein was identified for the virtual screening and molecular docking studies. The Asinex-fragments library of 22634 small molecules was docked at TRIB3 active site using the Glide module to identify new chemical entities. A total of 9 molecules were identified as final hits from virtual screening and their potency was ranked using Glide score, Glide energies, and residues interactions. The 6 prioritized lead molecules were further optimized using AutoDock, Prime MM/GBSA, and percentage of human oral absorption for the identification of potential leads. The molecules L2, L5, and L6 are identified as lead inhibitors and are showing consistent interactions with key residues Glu194 and Lys196 of TRIB3 protein. The identified potential leads were analyzed by ADME properties for their drug likeness and HergIC50 values are predicted for the prevention of preclinical failures. The present work sheds light on the identification of the best lead molecules against TRIB3 protein and offers a route to design as novel potential drug candidates for T2DM.

BiRDS - Binding Residue Detection from Protein Sequences Using Deep ResNets.

Protein-drug interactions play important roles in many biological processes and therapeutics. Predicting the binding sites of a protein helps to discover such interactions. New drugs can be designed to optimize these interactions, improving protein function. The tertiary structure of a protein decides the binding sites available to the drug molecule, but the determination of the 3D structure is slow and expensive. Conversely, the determination of the amino acid sequence is swift and economical. Although quick and accurate prediction of the binding site using just the sequence is challenging, the application of Deep Learning, which has been hugely successful in several biochemical tasks, makes it feasible. BiRDS is a Residual Neural Network that predicts the protein's most active binding site using sequence information. SC-PDB, an annotated database of druggable binding sites, is used for training the network. Multiple Sequence Alignments of the proteins in the database are generated using DeepMSA, and features such as Position-Specific Scoring Matrix, Secondary Structure, and Relative Solvent Accessibility are extracted. During training, a weighted binary cross-entropy loss function is used to counter the substantial imbalance in the two classes of binding and nonbinding residues. A novel test set SC6K is introduced to compare binding-site prediction methods. BiRDS achieves an AUROC score of 0.87, and the center of 25% of its predicted binding sites lie within 4 Å of the center of the actual binding site.