In silico Druggability Assessment of the NUDIX Hydrolase Protein Family as a Workflow for Target Prioritization

Unexpected Noncovalent Off-Target Activity of Clinical BTK Inhibitors Leads to Discovery of a Dual NUDT5/14 Antagonist

Cofactor mimicry represents an attractive strategy for the development of enzyme inhibitors but can lead to off-target effects due to the evolutionary conservation of binding sites across the proteome. Here, we uncover the ADP-ribose (ADPr) hydrolase NUDT5 as an unexpected, noncovalent, off-target of clinical BTK inhibitors. Using a combination of biochemical, biophysical, and intact cell NanoBRET assays as well as X-ray crystallography, we confirm catalytic inhibition and cellular target engagement of NUDT5 and reveal an unusual binding mode that is independent of the reactive acrylamide warhead. Further investigation of the prototypical BTK inhibitor ibrutinib also revealed potent inhibition of the largely unstudied NUDIX hydrolase family member NUDT14. By exploring structure-activity relationships (SARs) around the core scaffold, we identify a potent, noncovalent, and cell-active dual NUDT5/14 inhibitor. Cocrystallization experiments yielded new insights into the NUDT14 hydrolase active site architecture and inhibitor binding, thus providing a basis for future chemical probe design.

Advances in Drug Discovery and Design using Computer-aided Molecular Modeling

Computer-aided molecular modeling is a rapidly emerging technology that is being used to accelerate the discovery and design of new drug therapies. It involves the use of computer algorithms and 3D structures of molecules to predict interactions between molecules and their behavior in the body. This has drastically improved the speed and accuracy of drug discovery and design. Additionally, computer-aided molecular modeling has the potential to reduce costs, increase the quality of data, and identify promising targets for drug development. Through the use of sophisticated methods, such as virtual screening, molecular docking, pharmacophore modeling, and quantitative structure-activity relationships, scientists can achieve higher levels of efficacy and safety for new drugs. Moreover, it can be used to understand the activity of known drugs and simplify the process of formulating, optimizing, and predicting the pharmacokinetics of new and existing drugs. In conclusion, computer-aided molecular modeling is an effective tool to rapidly progress drug discovery and design by predicting the interactions between molecules and anticipating the behavior of new drugs in the body.

Antileishmanial Activity and In Silico Molecular Docking Studies of Malachra alceifolia Jacq. Fractions against Leishmania mexicana Amastigotes

Malachra alceifolia Jacq. (family Malvaceae), known as "malva," is a medicinal plant used as a traditional therapy in many regions of America, Africa and Asia. Traditionally, this plant is used in the form of extracts, powder and paste by populations for treating fever, stomachache, inflammation, and parasites. However, the ethnopharmacological validation of M. alceifolia has been scarcely researched. This study showed that the chloroform fraction (MA-IC) and subfraction (MA-24F) of the leaves of M. alceifolia exhibited a potential antileishmanial activity against axenic amastigotes of Leishmania mexicana pifanoi (MHOM/VE/60/Ltrod) and had high and moderate cytotoxic effects on the viability and morphology of macrophages RAW 264.7. This study reports, for the first time, possible terpenoid metabolites and derivatives present in M. alceifolia with activity against some biosynthetic pathways in L. mexicana amastigotes. The compounds from the subfractions MA-24F were highly active and were analyzed by gas chromatography-mass spectrometry (GC-MS) and by a molecular docking study in L. mexicana target protein. This study demonstrates the potential modes of interaction and the theoretical affinity energy of the metabolites episwertenol, α-amyrin and methyl commate A, which are present in the active fraction MA-24F, at allosteric sites of the pyruvate kinase, glyceraldehyde-3-phosphate dehydrogenase, triose phosphate isomerase, aldolase, phosphoglucose isomerase, transketolase, arginase and cysteine peptidases A, target proteins in some vital biosynthetic pathways were responsible for the survival of L. mexicana. Some phytoconstituents of M. alceifolia can be used for the search for potential new drugs and molecular targets for treating leishmaniases and infectious diseases. Furthermore, contributions to research and the validation and conservation of traditional knowledge of medicinal plants are needed globally.

A Guide to In Silico Drug Design

The drug discovery process is a rocky path that is full of challenges, with the result that very few candidates progress from hit compound to a commercially available product, often due to factors, such as poor binding affinity, off-target effects, or physicochemical properties, such as solubility or stability. This process is further complicated by high research and development costs and time requirements. It is thus important to optimise every step of the process in order to maximise the chances of success. As a result of the recent advancements in computer power and technology, computer-aided drug design (CADD) has become an integral part of modern drug discovery to guide and accelerate the process. In this review, we present an overview of the important CADD methods and applications, such as in silico structure prediction, refinement, modelling and target validation, that are commonly used in this area.

Druggable protein prediction using a multi-canal deep convolutional neural network based on autocovariance method

Drug targets must be identified and positioned correctly to research and manufacture new drugs. In this study, rather than using traditional methods for drug expansion, the drug target is determined using machine learning. Machine learning has generated significant interest and desire in recent years and extensive research due to its low cost and speed of operation. As a result, it is critical to develop an intelligent classification system for drug proteins. This study proposes two distinct models for the prediction of druggable protein classes based on the deep learning method. The translation of drug-protein sequences is based on six physicochemical properties of amino acids. Following the application of the autocovariance method, converted sequences are used as fixed-length input vectors in deep stacked sparse auto-encoders (DSSAEs) network. The coded protein sequences are also considered and utilized as a six-channel input vector for the deep convolutional neural network model. The experimental results contributing to the deep convolution model are more efficient than previous studies for classifying druggable proteins. The proposed approach achieved a sensitivity of 96.92%, a specificity of 99.51%, and an accuracy of 98.29%.

In Silico Methods for Identification of Potential Active Sites of Therapeutic Targets

Target identification is an important step in drug discovery, and computer-aided drug target identification methods are attracting more attention compared with traditional drug target identification methods, which are time-consuming and costly. Computer-aided drug target identification methods can greatly reduce the searching scope of experimental targets and associated costs by identifying the diseases-related targets and their binding sites and evaluating the druggability of the predicted active sites for clinical trials. In this review, we introduce the principles of computer-based active site identification methods, including the identification of binding sites and assessment of druggability. We provide some guidelines for selecting methods for the identification of binding sites and assessment of druggability. In addition, we list the databases and tools commonly used with these methods, present examples of individual and combined applications, and compare the methods and tools. Finally, we discuss the challenges and limitations of binding site identification and druggability assessment at the current stage and provide some recommendations and future perspectives.

In-Silico Characterization of Estrogen Reactivating β-Glucuronidase Enzyme in GIT Associated Microbiota of Normal Human and Breast Cancer Patients

Estrogen circulating in blood has been proved to be a strong biomarker for breast cancer. A β-glucuronidase enzyme (GUS) from human gastrointestinal tract (GIT) microbiota including probiotics has significant involvement in enhancing the estrogen concentration in blood through deconjugation of glucuronidated estrogens. The present project has been designed to explore GIT microbiome-encoded GUS enzymes (GUSOME) repertoire in normal human and breast cancer patients. For this purpose, a total of nineteen GUS enzymes from human GIT microbes, i.e., seven from healthy and twelve from breast cancer patients have been focused on. Protein sequences of enzymes retrieved from UniProt database were subjected to ProtParam, CELLO2GO, SOPMA (secondary structure prediction method), PDBsum (Protein Database summaries), PHYRE2 (Protein Homology/AnalogY Recognition Engine), SAVES v6.0 (Structure Validation Server), MEME version 5.4.1 (Multiple Em for Motif Elicitation), Caver Web server v 1.1, Interproscan and Predicted Antigenic Peptides tool. Analysis revealed the number of amino acids, isoelectric point, extinction coefficient, instability index and aliphatic index of GUS enzymes in the range of 586−795, 4.91−8.92, 89,980−155,075, 25.88−40.93 and 71.01−88.10, respectively. Sub-cellular localization of enzyme was restricted to cytoplasm and inner-membrane in case of breast cancer patients’ bacteria as compared to periplasmic space, outer membrane and extracellular space in normal GIT bacteria. The 2-D structure analysis showed α helix, extended strand, β turn and random coil in the range of 27.42−22.66%, 22.04−25.91%, 5.39−8.30% and 41.75−47.70%, respectively. The druggability score was found to be 0.05−0.45 and 0.06−0.80 in normal and breast cancer patients GIT, respectively. The radius, length and curvature of catalytic sites were observed to be 1.1−2.8 Å, 1.4−15.9 Å and 0.65−1.4, respectively. Ten conserved protein motifs with p < 0.05 and width 25−50 were found. Antigenic propensity-associated sequences were 20−29. Present study findings hint about the use of the bacterial GUS enzymes against breast cancer tumors after modifications via site-directed mutagenesis of catalytic sites involved in the activation of estrogens and through destabilization of these enzymes.

Nudix hydrolase 18 catalyzes the hydrolysis of active triphosphate metabolites of the antivirals remdesivir, ribavirin, and molnupiravir

Remdesivir and molnupiravir have gained considerable interest because of their demonstrated activity against SARS-CoV-2. These antivirals are converted intracellularly to their active triphosphate forms remdesivir-TP and molnupiravir-TP. Cellular hydrolysis of these active metabolites would consequently decrease the efficiency of these drugs; however, whether endogenous enzymes that can catalyze this hydrolysis exist is unknown. Here, we tested remdesivir-TP as a substrate against a panel of human hydrolases and found that only Nudix hydrolase (NUDT) 18 catalyzed the hydrolysis of remdesivir-TP with notable activity. The kcat/Km value of NUDT18 for remdesivir-TP was determined to be 17,700 s-1M-1, suggesting that NUDT18-catalyzed hydrolysis of remdesivir-TP may occur in cells. Moreover, we demonstrate that the triphosphates of the antivirals ribavirin and molnupiravir are also hydrolyzed by NUDT18, albeit with lower efficiency than Remdesivir-TP. Low activity was also observed with the triphosphate forms of sofosbuvir and aciclovir. This is the first report showing that NUDT18 hydrolyzes triphosphates of nucleoside analogs of exogenous origin, suggesting that NUDT18 can act as a cellular sanitizer of modified nucleotides and may influence the antiviral efficacy of remdesivir, molnupiravir, and ribavirin. As NUDT18 is expressed in respiratory epithelial cells, it may limit the antiviral efficacy of remdesivir and molnupiravir against SARS-CoV-2 replication by decreasing the intracellular concentration of their active metabolites at their intended site of action.

Computational modeling and druggability assessment of Aggregatibacter actinomycetemcomitans leukotoxin

The leukotoxin (LtxA) of Aggregatibacter actinomycetemcomitans (A. actinomycetemcomitans) is a protein exotoxin belonging to the repeat-in-toxin family (RTX). Numerous studies have demonstrated that LtxA may play a critical role in the pathogenicity of A. actinomycetemcomitans since hyper-leukotoxic strains have been associated with severe disease. Accordingly, considerable effort has been made to elucidate the mechanisms by which LtxA interacts with host cells and induce their death. However, these attempts have been hampered by the unavailability of a tertiary structure of the toxin, which limits the understanding of its molecular properties and mechanisms. In this paper, we used homology and template free modeling algorithms to build the complete tertiary model of LtxA at atomic level in its calcium-bound Holo-state. The resulting model was refined by energy minimization, validated by Molprobity and ProSA tools, and subsequently subjected to a cumulative 600ns of all-atom classical molecular dynamics simulation to evaluate its structural aspects. The druggability of the proposed model was assessed using Fpocket and FTMap tools, resulting in the identification of four putative cavities and fifteen binding hotspots that could be targeted by rational drug design tools to find new ligands to inhibit LtxA activity.

An Evolutionary Conservation and Druggability Analysis of Enzymes Belonging to the Bacterial Shikimate Pathway

Enzymes belonging to the shikimate pathway have long been considered promising targets for antibacterial drugs because they have no counterpart in mammals and are essential for bacterial growth and virulence. However, despite decades of research, there are currently no clinically relevant antibacterial drugs targeting any of these enzymes, and there are legitimate concerns about whether they are sufficiently druggable, i.e., whether they can be adequately modulated by small and potent drug-like molecules. In the present work, in silico analyses combining evolutionary conservation and druggability are performed to determine whether these enzymes are candidates for broad-spectrum antibacterial therapy. The results presented here indicate that the substrate-binding sites of most enzymes in this pathway are suitable drug targets because of their reasonable conservation and druggability scores. An exception was the substrate-binding site of 3-deoxy-D-arabino-heptulosonate-7-phosphate synthase, which was found to be undruggable because of its high content of charged residues and extremely high overall polarity. Although the presented study was designed from the perspective of broad-spectrum antibacterial drug development, this workflow can be readily applied to any antimicrobial target analysis, whether narrow- or broad-spectrum. Moreover, this research also contributes to a deeper understanding of these enzymes and provides valuable insights into their properties.

Structure-based assessment and druggability classification of protein-protein interaction sites

The featureless interface formed by protein–protein interactions (PPIs) is notorious for being considered a difficult and poorly druggable target. However, recent advances have shown PPIs to be druggable, with the discovery of potent inhibitors and stabilizers, some of which are currently being clinically tested and approved for medical use. In this study, we assess the druggability of 12 commonly targeted PPIs using the computational tool, SiteMap. After evaluating 320 crystal structures, we find that the PPI binding sites have a wide range of druggability scores. This can be attributed to the unique structural and physiochemical features that influence their ligand binding and concomitantly, their druggability predictions. We then use these features to propose a specific classification system suitable for assessing PPI targets based on their druggability scores and measured binding-affinity. Interestingly, this system was able to distinguish between different PPIs and correctly categorize them into four classes (i.e. very druggable, druggable, moderately druggable, and difficult). We also studied the effects of protein flexibility on the computed druggability scores and found that protein conformational changes accompanying ligand binding in ligand-bound structures result in higher protein druggability scores due to more favorable structural features. Finally, the drug-likeness of many published PPI inhibitors was studied where it was found that the vast majority of the 221 ligands considered here, including orally tested/marketed drugs, violate the currently acceptable limits of compound size and hydrophobicity parameters. This outcome, combined with the lack of correlation observed between druggability and drug-likeness, reinforces the need to redefine drug-likeness for PPI drugs. This work proposes a PPI-specific classification scheme that will assist researchers in assessing the druggability and identifying inhibitors of the PPI interface.

Virtual Screening Inhibitors of Ubiquitin-specific Protease 7 combining Pharmacophore Modeling and Molecular Docking

Ubiquitin-specific protease 7 (USP7) is one of the most extensively studied deubiquitinases. USP7 exhibits a high expression signature in various malignant tumors, suggesting that it is a marker of tumor prognosis and a potential drug target for anti-tumor therapy. In this study, virtual screening based on pharmacophore model and biological evaluation have been applied for the discovery of novel USP7 inhibitors targeting the catalytic active site. The TS-4 was screened from 215,480 small molecules and was found to have USP7 inhibitory activity. Preliminary in vitro studies disclosed its antiproliferative activity on human colon cancer cell lines (HCT-116 and RKO), compared with normal colon cell line (CCD841CoN). Molecular dynamics (MD) simulation revealed the combine mechanism between USP7 with the TS-4. The TS-4 formed stable interactions with Asp295, Phe409 and Tyr514, which were critical to enhance its biological activity. This compound will serve as a promising hit compound for facilitating the further design of novel USP7 inhibitors.

The applications of deep learning algorithms on in silico druggable proteins identification

•

We developed the first deep learning-based druggable protein classifier for fast and accurate identification of potential druggable proteins.

•

Experimental results on a standard dataset demonstrate that the prediction performance of deep learning model is comparable to those of existing methods.

•

We visualized the representations of druggable proteins learned by deep learning models, which helps us understand how they work.

•

Our analysis reconfirms that the attention mechanism is especially useful for explaining deep learning models.

Introduction

The top priority in drug development is to identify novel and effective drug targets. In vitro assays are frequently used for this purpose; however, traditional experimental approaches are insufficient for large-scale exploration of novel drug targets, as they are expensive, time-consuming and laborious. Therefore, computational methods have emerged in recent decades as an alternative to aid experimental drug discovery studies by developing sophisticated predictive models to estimate unknown drugs/compounds and their targets. The recent success of deep learning (DL) techniques in machine learning and artificial intelligence has further attracted a great deal of attention in the biomedicine field, including computational drug discovery.

Objectives

This study focuses on the practical applications of deep learning algorithms for predicting druggable proteins and proposes a powerful predictor for fast and accurate identification of potential drug targets.

Methods

Using a gold-standard dataset, we explored several typical protein features and different deep learning algorithms and evaluated their performance in a comprehensive way. We provide an overview of the entire experimental process, including protein features and descriptors, neural network architectures, libraries and toolkits for deep learning modelling, performance evaluation metrics, model interpretation and visualization.

Results

Experimental results show that the hybrid model (architecture: CNN-RNN (BiLSTM) + DNN; feature: dictionary encoding + DC_TC_CTD) performed better than the other models on the benchmark dataset. This hybrid model was able to achieve 90.0% accuracy and 0.800 MCC on the test dataset and 84.8% and 0.703 on a nonredundant independent test dataset, which is comparable to those of existing methods.

Conclusion

We developed the first deep learning-based classifier for fast and accurate identification of potential druggable proteins. We hope that this study will be helpful for future researchers who would like to use deep learning techniques to develop relevant predictive models.

Insight into the Binding and Hydrolytic Preferences of hNudt16 Based on Nucleotide Diphosphate Substrates

Nudt16 is a member of the NUDIX family of hydrolases that show specificity towards substrates consisting of a nucleoside diphosphate linked to another moiety X. Several substrates for hNudt16 and various possible biological functions have been reported. However, some of these reports contradict each other and studies comparing the substrate specificity of the hNudt16 protein are limited. Therefore, we quantitatively compared the affinity of hNudt16 towards a set of previously published substrates, as well as identified novel potential substrates. Here, we show that hNudt16 has the highest affinity towards IDP and GppG, with K_d below 100 nM. Other tested ligands exhibited a weaker affinity of several orders of magnitude. Among the investigated compounds, only IDP, GppG, m⁷GppG, AppA, dpCoA, and NADH were hydrolyzed by hNudt16 with a strong substrate preference for inosine or guanosine containing compounds. A new identified substrate for hNudt16, GppG, which binds the enzyme with an affinity comparable to that of IDP, suggests another potential regulatory role of this protein. Molecular docking of hNudt16-ligand binding inside the hNudt16 pocket revealed two binding modes for representative substrates. Nucleobase stabilization by Π stacking interactions with His24 has been associated with strong binding of hNudt16 substrates.

Druggable hot spots in trypanothione reductase: novel insights and opportunities for drug discovery revealed by DRUGpy

Assessment of target druggability guided by search and characterization of hot spots is a pivotal step in early stages of drug-discovery. The raw output of FTMap provides the data to perform this task, but it relies on manual intervention to properly combine different sets of consensus sites, therefore allowing identification of hot spots and evaluation of strength, shape and distance among them. Thus, the user's previous experience on the target and the software has a direct impact on how data generated by FTMap server can be explored. DRUGpy plugin was developed to overcome this limitation. By automatically assembling and scoring all possible combinations of consensus sites, DRUGpy plugin provides FTMap users a straight-forward method to identify and characterize hot spots in protein targets. DRUGpy is available in all operating systems that support PyMOL software. DRUGpy promptly identifies and characterizes pockets that are predicted by FTMap to bind druglike molecules with high-affinity (druggable sites) or low-affinity (borderline sites) and reveals how protein conformational flexibility impacts on the target's druggability. The use of DRUGpy on the analysis of trypanothione reductases (TR), a validated drug target against trypanosomatids, showcases the usefulness of the plugin, and led to the identification of a druggable pocket in the conserved dimer interface present in this class of proteins, opening new perspectives to the design of selective inhibitors.