No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Development and Validation of a Genetic Algorithm for Flexible Docking
Abstract
Prediction of small molecule binding modes to macromolecules of known three-dimensional structure is a problem of paramount importance in rational drug design (the "docking" problem). We report the development and validation of the program GOLD (Genetic Optimisation for Ligand Docking). GOLD is an automated ligand docking program that uses a genetic algorithm to explore the full range of ligand conformational flexibility with partial flexibility of the protein, and satisfies the fundamental requirement that the ligand must displace loosely bound water on binding. Numerous enhancements and modifications have been applied to the original technique resulting in a substantial increase in the reliability and the applicability of the algorithm. The advanced algorithm has been tested on a dataset of 100 complexes extracted from the Brookhaven Protein DataBank. When used to dock the ligand back into the binding site, GOLD achieved a 71% success rate in identifying the experimental binding mode.
... The filtering resulted in the remaining 2864 and 1196 molecules as potential inhibitors for TSLPR and TSLP, respectively. To assess whether virtual hits fulfill the previously defined interaction patterns represented by the 3D pharmacophore, molecular docking experiments were performed by docking them into the prepared protein structure (PDB (Berman et al, 2000) code: 5J11 ) using GOLD v5.2 (Jones et al, 1997). The resulting poses were compared with the 3D pharmacophore alignment from the virtual screening, and similar conformations were prioritized. ...
... To check whether virtual hits from 3D pharmacophore screening are actually able to fulfill the desired interaction pattern, a molecular docking step was included in the virtual screening workflow. Virtual hits were docked using GOLD v5.2 (Jones et al, 1997) with standard parameters and GoldScore (Verdonk et al, 2003) as a scoring function. Binding site residues were defined as all protein residues in 10 Å distance from the key residues Asp92 or Arg153 for TSLPR and TSLP, respectively. ...
Thymic stromal lymphopoietin (TSLP) is a key player in atopic diseases, which has sparked great interest in therapeutically targeting TSLP. Yet, no small-molecule TSLP inhibitors exist due to the challenges of disrupting the protein–protein interaction between TSLP and its receptor. Here, we report the development of small-molecule TSLP receptor inhibitors using virtual screening and docking of >1,000,000 compounds followed by iterative chemical synthesis. BP79 emerged as our lead compound that effectively abrogates TSLP-triggered cytokines at low micromolar concentrations. For in-depth analysis, we developed a human atopic disease drug discovery platform using multi-organ chips. Here, topical application of BP79 onto atopic skin models that were co-cultivated with lung models and Th2 cells effectively suppressed immune cell infiltration and IL-13, IL-4, TSLP, and periostin secretion, while upregulating skin barrier proteins. RNA-Seq analysis corroborate these findings and indicate protective downstream effects on the lungs. To the best of our knowledge, this represents the first report of a potent putative small molecule TSLPR inhibitor which has the potential to expand the therapeutic and preventive options in atopic diseases.
... These calculations were performed using GOLD program version 2020.3.0 (Jones et al., 1997), and the protein coordinates were taken from the crystal structures of HSA in complex with hemin and myristic acid (PDB ID 1O9X) (Zunszain et al., 2003) and of HAG in the unbound form (PDB ID 3KQ0) (Schönfeld et al., 2008). The experimental procedure was similar to that described for LAP, N-LAP and O-LAP in HSA . ...
... To provide a more in-depth understanding of the molecular bases responsible for the distinct photobehavior of the GFT metabolites, GFT-M1 and GFT-M2, relay on the protein that transport them, their binding modes were studied in silico. To this end, docking studies were first performed using the GOLD program version 2021.3.0 (Jones et al., 1997), followed by MD simulation studies to provide a more realistic picture of the ligand arrangement upon binding. The protein coordinates of the reported wild-type structure of HAG (PDB ID 3KQ0) and of HSA in complex with myristic acid and hemin (PDB ID 1O9X) were used for these studies (Zunszain et al., 2003;Schönfeld et al., 2008). ...
The photobiological damage that certain drugs or their metabolites can photosensitize in proteins is generally associated with the nature of the excited species that are generated upon interaction with UVA light. In this regard, the photoinduced damage of the anticancer drug gefitinib (GFT) and its two main photoactive metabolites GFT-M1 and GFT-M2 in cellular milieu was recently investigated. With this background, the photophysical properties of both the drug and its metabolites have now been studied in the presence of the two main transport proteins of human plasma, i.e., serum albumin (HSA) and α1-acid glycoprotein (HAG) upon UVA light excitation. In general, the observed photobehavior was strongly affected by the confined environment provided by the protein. Thus, GFT-M1 (which exhibits the highest phototoxicity) showed the highest fluorescence yield arising from long-lived HSA-bound phenolate-like excited species. Conversely, locally excited (LE) states were formed within HAG, resulting in lower fluorescence yields. The reserve was true for GFT-M2, which despite being also a phenol, led mainly to formation of LE states within HSA, and phenolate-like species (with a minor contribution of LE) inside HAG. Finally, the parent drug GFT, which is known to form LE states within HSA, exhibited a parallel behavior in the two proteins. In addition, determination of the association constants by both absorption and emission spectroscopy revealed that the two metabolites bind stronger to HSA than the parent drug, whereas smaller differences were observed for HAG. This was further confirmed by studying the competing interactions between GFT or its metabolites with the two proteins using fluorescence measurements. These above experimental findings were satisfactorily correlated with the results obtained by means of molecular dynamics (MD) simulations, which revealed the high affinity binding sites, the strength of interactions and the involved amino acid residues. In general, the differences observed in the photobehavior of the drug and its two photoactive metabolites in protein media are consistent with their relative photosensitizing potentials.
... Docking was performed with the GOLD software (Version 2022.1.0) employing the ASP scoring function [22], from which the highest-scoring poses were selected for analysis. ...
... Molecular docking simulations were performed using GOLD software [22], employing the default values and using the ASP scoring function. A search space of 20Å 3 , entered on the ligand's position in the crystallographic structure, was defined, and each ligand was docked 100 times. ...
Despite the efforts made to stop the tuberculosis (TB) epidemic, it still remains one of the leading causes of death from an infectious disease. Previous work in the group uncovered a new family of amides which showed promising activities against Mycobacterium tuberculosis. A closer look at the literature showed that these compounds are structurally related to the DNB family of DprE1 inhibitors, so, we decided to study a wide range of substituted amides and determine their activity, focusing on unexplored structures related to the dinitrobenzamides (DNB) found in the literature. To synthesize our library of compounds we started from 3,5-dinitrobenzoic acid to form the nitroaromatic core that is characteristic of the DNB’s, to which we then added linear or cyclic amine moieties. Additionally, the impact of terminal aromatic moieties was also assessed for some derivatives, via an ether, ester, or amide bond. In order to obtain the desired derivatives, multiple synthetic approaches were used, mainly focused on nucleophilic addition/elimination reactions, SN2 reactions and Mitsunobu reactions. The activity was impacted mainly by two structural features, the addition of an aromatic moiety as a terminal group and the type of the linker. The most interesting compounds exhibited activities comparable to isoniazid and DNB1. Computational studies were also performed aimed at understanding their possible interactions with DprE1. The most active compounds are a good starting point for further development, and we plan to study an extended family of compounds based in those structures.
... Typically, during docking, the protein structure is regarded as rigid, and the aim is to find the best fit between the ligand and the protein, akin to the lock and key analogy [10,17]. Some docking programs, such as GOLD, allow modeling of some aspect of induced fit "hand in glove" binding by permitting side chain flexibility of the amino acid residues in the binding site [102]. Other docking programs model conformational movement in the protein backbone during binding, a technique that has gained traction in protein-protein docking [103,104]. ...
Binding affinity is a fundamental parameter in drug design, describing the strength of the interaction between a molecule and its target protein. Accurately predicting binding affinity is crucial for the rapid development of novel therapeutics, the prioritization of promising candidates, and the optimization of their properties through rational design strategies. Binding affinity is determined by the mechanism of recognition between proteins and ligands. Various models, including the lock and key, induced fit, and conformational selection, have been proposed to explain this recognition process. However, current computational strategies to predict binding affinity, which are based on these models, have yet to produce satisfactory results. This article explores the connection between binding affinity and these protein-ligand interaction models, highlighting that they offer an incomplete picture of the mechanism governing binding affinity. Specifically, current models primarily center on the binding of the ligand and do not address its dissociation. In this context, the concept of ligand trapping is introduced, which models the mechanisms of dissociation. When combined with the current models, this concept can provide a unified theoretical framework that may allow for the accurate determination of the ligands’ binding affinity.
... The molecular docking calculations were carried out using the GOLD docking program (v 4.1.2) [62], selecting one active site shared by two neighboring protein subunits. As a first step, a flexible model of the DHAP molecule was docked into the active sites of both receptors, keeping the axial water molecule coordinated to the metal center to fulfil the sixth coordinating position. ...
Dihydroxyacetone phosphate (DHAP)-dependent aldolases catalyze the aldol addition of DHAP to a variety of aldehydes and generate compounds with two stereocenters. This reaction is useful to synthesize chiral acyclic nucleosides, which constitute a well-known class of antiviral drugs currently used. In such compounds, the chirality of the aliphatic chain, which mimics the open pentose residue, is crucial for activity. In this work, three DHAP-dependent aldolases: fructose-1,6-biphosphate aldolase from rabbit muscle, rhanmulose-1-phosphate aldolase from Thermotoga maritima, and fuculose-1-phosphate aldolase from Escherichia coli, were used as biocatalysts. Aldehyde derivatives of thymine and cytosine were used as acceptor substrates, generating new acyclic nucleoside analogues containing two new stereocenters with conversion yields between 70% and 90%. Moreover, structural analyses by molecular docking were carried out to gain insights into the diasteromeric excess observed.
... These training poses were generated using GOLD v.5.6.33 (ref. 56). Consequently, for both CASF-2007 and CASF-2013, there was a total of 365,000 training poses available for fine-tuning purposes. ...
Despite the success of pretrained natural language processing (NLP) models in various fields, their application in computational biology has been hindered by their reliance on biological sequences, which ignores vital three-dimensional (3D) structural information incompatible with the sequential architecture of NLP models. Here we present a topological transformer (TopoFormer), which is built by integrating NLP models and a multiscale topology technique, the persistent topological hyperdigraph Laplacian (PTHL), which systematically converts intricate 3D protein–ligand complexes at various spatial scales into an NLP-admissible sequence of topological invariants and homotopic shapes. PTHL systematically transforms intricate 3D protein–ligand complexes into NLP-compatible sequences of topological invariants and shapes, capturing essential interactions across spatial scales. TopoFormer gives rise to exemplary scoring accuracy and excellent performance in ranking, docking and screening tasks in several benchmark datasets. This approach can be utilized to convert general high-dimensional structured data into NLP-compatible sequences, paving the way for broader NLP based research.
... For several years, different theoretical methods have been developed that serve as tools for the synthesis or preparation of new drugs that interact with some biomolecules (Rarey et al., 1996;Jones et al., 1997;Österberg et al., 2002;Venkatachalam et al., 2003); in this way, the theoretical analysis on protein-ligand complex formation is necessary to predict the possible treatment of diseases. ...
Several drugs have been used to treat asthma diseases, such as salmeterol, ipratropium bromide, montelukast, and fluticasone; however, some of these drugs can cause side effects such as hypokalemia, lactic acidosis, and hypotension. Analyzing these data, this study aimed to evaluate the possible interaction of twenty-two carbazole derivatives with the M1-muscrinic receptor to provide a new therapeutic alternative against asthma. The theoretical interaction of carbazole derivatives with M1-muscrinic receptor surface was determined using 5cxv protein, pirenzepine, atropine, AF-150, and PD159714 drugs as theoretical tools in a DockingServer software. The results showed differences in the interaction of carbazole derivatives with the 5cxv protein surface compared with pyranzepine, atropine, AF-150, and PD159714 drugs. Besides, constant inhibition (Ki) for carbazole derivatives 11 and 22 was lower than for pirenzepine and AF-150 drugs. Other data indicate that Ki values for 11 and 22 were higher than atropine and ipratropium bromide. In addition, the Ki values for compounds 17 and 20 were like both atropine and PD150714 drugs. Finally, Ki values for carbazole derivatives 17 and 20 were lower than pyranzepine, ipratropium bromide, and AF-150 reagents. All these data suggest that carbazole derivatives 11, 17, 20, and 22 may act as M1-muscarinic receptor inhibitor agents; this phenomenon could result in the regulation of bronchial tone in asthma disease.
... To investigate whether the performance of binding pose prediction depends on the crystallographic structures, we analyzed the RMSD between the docked poses of active molecules and their corresponding X-ray crystallographic structures. Poses exhibiting an RMSD less than 2 Å were considered as an accurate and acceptable docking pose 27,28 . We utilized the "success rate" metric, employing a threshold of 2 Å. ...
The coronavirus disease 19 pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has led to a global health crisis with millions of confirmed cases and related deaths. The main protease (Mpro) of SARS-CoV-2 is crucial for viral replication and presents an attractive target for drug development. Despite the approval of some drugs, the search for effective treatments continues. In this study, we systematically evaluated 342 holo-crystal structures of Mpro to identify optimal conformations for structure-based virtual screening (SBVS). Our analysis revealed limited structural flexibility among the structures. Three docking programs, AutoDock Vina, rDock, and Glide were employed to assess the efficiency of virtual screening, revealing diverse performances across selected Mpro structures. We found that the structures 5RHE, 7DDC, and 7DPU (PDB Ids) consistently displayed the lowest EF, AUC, and BEDROCK scores. Furthermore, these structures demonstrated the worst pose prediction results in all docking programs. Two structural differences contribute to variations in docking performance: the absence of the S1 subsite in 7DDC and 7DPU, and the presence of a subpocket in the S2 subsite of 7DDC, 7DPU, and 5RHE. These findings underscore the importance of selecting appropriate Mpro conformations for SBVS, providing valuable insights for advancing drug discovery efforts.
... Docking) software version 2022.2.0 (Jones et al., 1997 ...
3,4‐Methylenedioxymethamphetamine (MDMA, ‘ecstasy’) is re‐emerging in clinical settings as a candidate for the treatment of specific neuropsychiatric disorders (e.g. post‐traumatic stress disorder) in combination with psychotherapy. MDMA is a psychoactive drug, typically regarded as an empathogen or entactogen, which leads to transporter‐mediated monoamine release. Despite its therapeutic potential, MDMA can induce dose‐, individual‐, and context‐dependent untoward effects outside safe settings. In this study, we investigated whether three new methylenedioxy bioisosteres of MDMA improve its off‐target profile. In vitro methods included radiotracer assays, transporter electrophysiology, bioluminescence resonance energy transfer and fluorescence‐based assays, pooled human liver microsome/S9 fraction incubations, metabolic stability studies, isozyme mapping, and liquid chromatography coupled to high‐resolution mass spectrometry. In silico methods included molecular docking. Compared with MDMA, all three MDMA bioisosteres (ODMA, TDMA, and SeDMA) showed similar pharmacological activity at human serotonin, dopamine, and norepinephrine transporters (hSERT, hDAT, and hNET, respectively) but decreased agonist activity at 5‐HT2A/2B/2C receptors. Regarding their hepatic metabolism, they differed from MDMA, with N‐demethylation being the only metabolic route shared, and without forming phase II metabolites. In addition, TDMA showed an enhanced intrinsic clearance in comparison to its congeners. Additional screening for their interaction with human organic cation transporters (hOCTs) and plasma membrane monoamine transporter (hPMAT) revealed a weaker interaction of the MDMA analogs with hOCT1, hOCT2, and hPMAT. Our findings suggest that these new MDMA bioisosteres might constitute appealing therapeutic alternatives to MDMA, sparing the primary pharmacological activity at hSERT, hDAT, and hNET, but displaying a reduced activity at 5‐HT2A/2B/2C receptors and alternative hepatic metabolism. Whether these MDMA bioisosteres may pose lower risk alternatives to the clinically re‐emerging MDMA warrants further studies. image
... For the second phase (re-dock) 100% efficiency was used in conjunction with fifty docking runs. The GoldScore(GS) 63 and ChemScore(CS), 64 improved Piecewise Linear Potential (ChemPLP) 65 and Astex Statistical Potential (ASP) 66 scoring functions were implemented to validate the predicted binding modes and relative energies of the ligands using the GOLD v5.2 software suite. STOCK1N compounds identified as p62 binders in silico (Table S3) were purchased from InterBioScreen Ltd and screened in cells expressing TET-inducible p62-fluc construct as previously described. ...
... Molecular docking calculations were performed using GOLD (CCDC Discovery 2020) and the ChemScore fitness function (Jones et al., 1997;Verdonk et al., 2003). The structure of human cannabinoid receptor CB1 was taken from PDB 5TGZ (Hua et al., 2016a). ...
The endocannabinoid (eCB) system regulates several brain functions and is implicated in neurological disorders. The pharmacological blockade of cannabinoid receptors has a therapeutic potential for various cognitive deficits, but also produces severe psychiatric side effects. Hence, new cannabinoid compounds that potentiate therapeutic effects, while minimizing toxicity, are required. In this study, we synthesized and characterized a novel antagonist/inverse agonist of CB 1 receptors. UVI3502 showed affinity for two [ ³ H]CP55,940 binding sites (IC 50Hi 0.47 ± 1.94 nM and IC 50Lo 1470 ± 1.80 nM). Subsequent binding assays performed in CB 1 and CB 2 overexpressing membranes determined that the low affinity binding site corresponded to CB 1 , but the high-affinity binding site of UVI3502 did not correspond to CB 2 and the possibility of it corresponding to GPR55 was analyzed. The affinity of UVI3502 for CB 1 receptors was further confirmed with neuroanatomical specificity by autoradiography in key brain areas, in which functional [ ³⁵ S]GTPɣS assays demonstrated that UVI3502 behaved as an antagonist/inverse agonist of CB 1 receptors, blocking the stimulation evoked by potent cannabinoid receptor agonist CP55,940 and decreasing basal [ ³⁵ S]GTPɣS binding. The in silico characterization of the binding to CB 1 receptor through molecular docking and molecular dynamics suggests that this activity is explained by the planar and rigid structure of UVI3502, which is optimal for interactions with the inactive state of the receptor. These results indicate that UVI3502 is a novel antagonist/inverse agonist of CB 1 receptors, making it a compelling candidate for pharmacologically blocking cannabinoid receptors in the central nervous system.
Significance Statement
UVI3502 is a novel antagonist/inverse agonist of CB 1 receptors, with almost no affinity for CB 2 receptors and an additional high-affinity binding site for a third, cannabinoid-like receptor, potentially GPR55. In relevant brain areas for learning and memory processes with a high expression of CB 1 , UVI3502 blocks the stimulation evoked by the cannabinoid receptor agonist CP55,940, rendering it as an interesting compound for the pharmacological blockade of cannabinoid receptors in the central nervous system.
... Docking was performed for 100 genetic algorithm runs with 100,000 iterations and early termination option was disabled. The GOLD fitness score was calculated from the contributions of hydrogen bond and van der Waals interactions between the protein and ligand ( 42 ). From the GOLD based docking, the best-sampled pose (highest GOLD score) was selected and analyzed for interactions. ...
Triple negative breast cancer (TNBC) is associated with a generally poor prognosis due to its highly aggressive and metastatic nature, lack of targetable receptors, as well as the frequent development of resistance to chemotherapy. We previously reported that AU1, a small molecule developed as an inhibitor of BPTF (bromodomain PHD finger containing transcription factor), was capable of sensitizing preclinical models of triple negative breast cancer to chemotherapy, in part via the promotion of autophagy. In studies reported here, we identify an additional property of this compound, specifically that sensitization is associated with inhibition of the P-glycoprotein efflux pump. In silico molecular docking studies indicate that AU1 binds to active regions of the efflux pump, in a manner consistent with inhibition of pump function. This work identifies a novel chemical structure that can influence multidrug efflux, an established mechanism of drug resistance in triple negative breast cancer, that has not yet been successfully addressed by clinical efforts.
... No [30,36] ChemProp No [27] ADMET-AI No [37] Docking Glide Yes [38] PLANTS Yes [39] GOLD Yes [40] OEDock Yes [41] Smina No [42] Gnina No [43] Vina No [44] rDock No [45] Synthesizability SA score No [46] RA Score No [25] AiZynthFinder No [26] Reaction Step threshold No evaluation metrics for comparison upon exit. A series of presets are already present including GuacaMol and MolOpt, where benchmark-specific metrics are computed. ...
Generative models are undergoing rapid research and application to de novo drug design. To facilitate their application and evaluation, we present MolScore. MolScore already contains many drug-design-relevant scoring functions commonly used in benchmarks such as, molecular similarity, molecular docking, predictive models, synthesizability, and more. In addition, providing performance metrics to evaluate generative model performance based on the chemistry generated. With this unification of functionality, MolScore re-implements commonly used benchmarks in the field (such as GuacaMol, MOSES, and MolOpt). Moreover, new benchmarks can be created trivially. We demonstrate this by testing a chemical language model with reinforcement learning on three new tasks of increasing complexity related to the design of 5-HT2a ligands that utilise either molecular descriptors, 266 pre-trained QSAR models, or dual molecular docking. Lastly, MolScore can be integrated into an existing Python script with just three lines of code. This framework is a step towards unifying generative model application and evaluation as applied to drug design for both practitioners and researchers. The framework can be found on GitHub and downloaded directly from the Python Package Index.
Scientific Contribution
MolScore is an open-source platform to facilitate generative molecular design and evaluation thereof for application in drug design. This platform takes important steps towards unifying existing benchmarks, providing a platform to share new benchmarks, and improves customisation, flexibility and usability for practitioners over existing solutions.
Graphical Abstract
... Thus, the search for candidate ligands of the studied targets occurs faster and easier (Neudert and Klebe 2011). Generic algorithms is a docking methodology that mimics the process of evolution, manipulating a collection of data structures called chromosomes, where each of these structures encodes a possible solution, that is, a possible orientation of the ligand within the binding site of a given protein and, thus, a suitability score can be assigned based on the relative merit of that solution (Jones et al. 1997). ...
Lignans and neolignans, commonly found in plants, are generally composed of two phenylpropane (C6-C3) units with a marked diversity of molecular structures and biological activities of therapeutic interest. They serve as lead compound for organic synthesis of derivatives intended to optimize the pharmacological activity, and, also, for studying structure-activity relationships. Piperaceae species are one of the richest natural sources of lignans and neolignans. In the leaves of Piper rivinoides Kunth, Piperaceae, some biologically active neolignans with leishmanicidal, antibacterial, and antitumor effects have been identified. Furthermore, it had been shown that the neolignans conocarpan, eupomatenoid-5, and eupomatenoid-6, isolated from P. rivinoides leaves inhibited CYP1A1/2 activity in the rat liver microsomes. The present study extends these observations and finds that the crude hydroalcoholic extract of P. rivinoides leaves (100 µg/ml) and conocarpan, eupomatenoid-5, and eupomatenoid-6 (100 µM) inhibited the 7-ethoxyresorufin-O-deethylase and 7-methoxyresorufin-O-demethylation activities catalyzed by the cytochrome P450 isoforms rhCYP1A1 and rhCYP1A2. Molecular docking analysis has suggested that π-π stacking interactions involving residues Phe123, Phe224, and Phe258 in CYP1A1, and Phe125, Phe226, Gly316, Ala317, Ile386, and Leu497 in CYP1A2, are crucial for the stabilizing interactions of these neolignans with the active sites of CYP1A1 and CYP1A2.
... Molecular docking calculations were performed using the GOLD 2022.3.0 package [37]. The score function used was ChemPLP due to its superior performance compared to the other functions available in the package, both in pose prediction and virtual screening [38]. ...
Brown seaweeds of the Fucus genus represent a rich source of natural antiviral products. In this study, a Fucus ceranoides hydroalcoholic extract (FCHE) was found to inhibit 74.2 ± 1.3% of the proteolytic activity of the free SARS-CoV-2 3CL protease (3CLpro), an enzyme that plays a pivotal role in polyprotein processing during coronavirus replication and has been identified as a relevant drug discovery target for SARS- and MERS-CoVs infections. To purify and identify 3CLpro ligands with potential inhibitory activity using a one-step approach, we immobilized the enzyme onto magnetic microbeads (3CLpro-MPs), checked that the enzymatic activity was maintained after grafting, and used this bait for a ligand-fishing strategy followed by a high-resolution mass spectrometry analysis of the fished-out molecules. Proof of concept for the ligand-fishing capacity of the 3CLpro-MPs was demonstrated by doping the FCHE extract with the substrate peptide TSAVLQ-pNA, resulting in the preferential capture of this high-affinity peptide within the macroalgal complex matrix. Ligand fishing in the FCHE alone led to the purification and identification via high-resolution mass spectrometry (HRMS) of seven hepta-, octa-, and decapeptides in an eluate mix that significantly inhibited the free 3CLpro more than the starting FCHE (82.7 ± 2.2% inhibition). Molecular docking simulations of the interaction between each of the seven peptides and the 3CLpro demonstrated a high affinity for the enzyme’s proteolytic active site surpassing that of the most affine peptide ligand identified so far (a co-crystallographic peptide). Testing of the corresponding synthetic peptides demonstrated that four out of seven significantly inhibited the free 3CLpro (from 46.9 ± 6.4 to 76.8 ± 3.6% inhibition at 10 µM). This study is the first report identifying peptides from Fucus ceranoides with high inhibitory activity against the SARS-CoV-2 3CLprotease which bind with high affinity to the protease’s active site. It also confirms the effectiveness of the ligand-fishing strategy for the single-step purification of enzyme inhibitors from complex seaweed matrices.
... The software GOLD 3.0.1 was employed and the conformations of binding for SA, RV and GA were procured by executing docking studies (Jones et al., 1997). The compounds SA, RV and GA were docked to the HMG-CoA reductase and LPL enzymes active sites to analyze the probability of regulatory activity of these compounds. ...
... Molecular docking experiments were performed with a monomer (chain A) of the human topoisomerase IIα ATPase domain (PDB: 1ZXM) [14] using the GOLD software [68]. We again followed our standardized and validated docking procedure to prepare the protein as well as the ATP binding site, which was validated in our studies on catalytic topo IIα inhibitors [22,45]. ...
... It is a convenient method of reproducing the secondary structure distribution of key residues in proteins and identifying residues implicated in cavities or clefts [38,39]. The molecular modeling software Chimera 1.15 was used for visualization [40]. ...
Aim: New microtubule-targeting agents are needed to improve cancer treatment. The recent characterization of the anticancer alkaloid securinine as a tubulin-binding agent prompted us to explore the interaction of related monomeric and dimeric analogues with tubulin. The interaction between the α/β-tubulin dimer and alkaloids fluevirines A–F and flueggenines A–I, isolated from the bush Flueggea virosa (Roxb. ex Willd.) Royle, was investigated using molecular docking. Methods: Two molecular models were initially compared for the binding of securinine to α/β-tubulin. The pironetin-binding site model (5FNV) was selected for the subsequent docking analysis with all compounds. Empirical energies of interaction (ΔE) were measured and compared. Results: Fluevirine A has been identified as a potent tubulin binder. This dimeric alkaloid formed more stable complexes with tubulin than the monomeric counterparts, such as fluevirines B–D. The bis-indole derivative fluevirine E also provided more stable complexes than (nor)securinine. The study was extended to the dimeric alkaloids flueggenines A–I and three compounds were identified as potential tubulin binders: the polycyclic product flueggenine B, the norsecurinine-indole hybrid flueggenine E, and the norsecurinine dimer flueggenine I. This later compound proved to be well adapted to fit into the pironetin site of tubulin, extending its two norsecurinine units between the colchicine-binding area and the pironetin site, in close proximity to the pironetin-reactive cysteine-316 residue. Structure-binding relationships were delineated. Conclusions: The study identifies the dimeric alkaloids fluevirine A and flueggenine I as potential α-tubulin binding agents. For the first time, dimeric alkaloids including two C-C connected norsecurinine units are characterized as tubulin ligands. The study contributes to a better understanding of the mechanism of action of Flueggea alkaloids and should help the design of anticancer analogues targeting the pironetin site of α-tubulin.
... On the other hand, structure-based methods use traditional global or local molecular docking to generate the structures of PpIs for a more comprehensive assessment of PpIs. 14 When the binding sites or interfaces for PpIs have been explicitly specified, local docking tools, such as AutoDock Vina 15 and GOLD, 16 would generate the binding poses of peptides within the user-defined binding sites. Global docking, such as pepATTRACT 17 and HPEPDOCK, 18 would exhaustively search the entire protein surface to identify potential binding sites, and then predict the binding structures of peptides. ...
... These clusters were obtained using the root mean square deviation (RMSD) of all nonhydrogen atoms as a similarity measure. To predict the binding conformation and affinity of a xylose-based substrate, the protein-ligand docking program GOLD was used (Jones et al. 1997). Since the xylan sugar chains are considered too long for molecular docking, the model substrate chosen as example for evaluation was xylohexaose (Xie et al. 2021). ...
Xylanases are key biocatalysts in the degradation of the β‐1,4‐glycosidic linkages in the xylan backbone of hemicellulose. These enzymes are potentially applied in a wide range of bioprocessing industries under harsh conditions. Metagenomics has emerged as powerful tools for the bioprospection and discovery of interesting bioactive molecules from extreme ecosystems with unique features, such as high temperatures. In this study, an innovative combination of function-driven screening of a compost metagenomic library and automatic extraction of halo areas with in-house MATLAB functions resulted in the identification of a promising clone with xylanase activity (LP4). The LP4 clone proved to be an effective xylanase producer under submerged fermentation conditions. Sequence and phylogenetic analyses revealed that the xylanase, Xyl4, corresponded to an endo-1,4-β-xylanase belonging to glycosyl hydrolase family 10 (GH10). When xyl4 was expressed in Escherichia coli BL21(DE3), the enzyme activity increased about 2-fold compared to the LP4 clone. To get insight on the interaction of the enzyme with the substrate and establish possible strategies to improve its activity, the structure of Xyl4 was predicted, refined, and docked with xylohexaose. Our data unveiled, for the first time, the relevance of the amino acids Glu133 and Glu238 for catalysis, and a close inspection of the catalytic site suggested that the replacement of Phe316 by a bulkier Trp may improve Xyl4 activity. Our current findings contribute to enhancing the catalytic performance of Xyl4 towards industrial applications.
Key points
• A GH10 endo-1,4-β-xylanase (Xyl4) was isolated from a compost metagenomic library
• MATLAB’s in-house functions were developed to identify the xylanase-producing clones
• Computational analysis showed that Glu133 and Glu238 are crucial residues for catalysis
Graphical abstract
... Luciferin-ME was docked into the binding pocket of CYP2A7-WT and Luciferin-3FBE was docked into the binding site of CYP2A7*1 using GOLD v. 5.8.1 (Genetic Optimisation for Ligand Docking; CCDC Software, Cambridge, UK) [36]. In total, 20 genetic algorithm (GA) runs with the coordinates of the heme iron serving as the operation center were used with a radius of 10 Å. ...
CYP2A7 is one of the most understudied human cytochrome P450 enzymes and its contributions to either drug metabolism or endogenous biosynthesis pathways are not understood, as its only known enzymatic activities are the conversions of two proluciferin probe substrates. In addition, the CYP2A7 gene contains four single-nucleotide polymorphisms (SNPs) that cause missense mutations and have minor allele frequencies (MAFs) above 0.5. This means that the resulting amino acid changes occur in the majority of humans. In a previous study, we employed the reference standard sequence (called CYP2A7*1 in P450 nomenclature). For the present study, we created another CYP2A7 sequence that contains all four amino acid changes (Cys311, Glu169, Gly479, and Arg274) and labeled it CYP2A7-WT. Thus, it was the aim of this study to identify new substrates and inhibitors of CYP2A7 and to compare the properties of CYP2A7-WT with CYP2A7*1. We found several new proluciferin probe substrates for both enzyme variants (we also performed in silico studies to understand the activity difference between CYP2A7-WT and CYP2A7*1 on specific substrates), and we show that while they do not act on the standard CYP2A6 substrates nicotine, coumarin, or 7-ethoxycoumarin, both can hydroxylate diclofenac (as can CYP2A6). Moreover, we found ketoconazole, 1-benzylimidazole, and letrozole to be CYP2A7 inhibitors.
... The advances in ML-based molecular docking methods have been described in Section 2.1. Since the physics-based (traditional) docking remains one of the most popular and easy-touse approaches, here we will only briefly mention a few docking programs: UCSF Dock [98], Autodock [99], Autodock Vina [100], GOLD [101], Glide [102], rDock [103], GalaxyDock2 [104], FlexAID [105], PLANTS [106], GeauxDock [107], and many others. Autodock, and especially Autodock Vina, are probably the most popular docking programs, likely because they are easy to install on most Linux distributions. ...
Introduction:
Modern drug discovery revolves around designing ligands that target the chosen biomolecule, typically proteins. For this, the evaluation of affinities of putative ligands is crucial. This has given rise to a multitude of dedicated computational and experimental methods that are constantly being developed and improved.
Areas covered:
In this review, the authors reassess both the industry mainstays and the newest trends among the methods for protein - small-molecule affinity determination. They discuss both computational affinity predictions and experimental techniques, describing their basic principles, main limitations, and advantages. Together, this serves as initial guide to the currently most popular and cutting-edge ligand-binding assays employed in rational drug design.
Expert opinion:
The affinity determination methods continue to develop toward miniaturization, high-throughput, and in-cell application. Moreover, the availability of data analysis tools has been constantly increasing. Nevertheless, cross-verification of data using at least two different techniques and careful result interpretation remain of utmost importance.
The discovery of antimicrobials with novel mechanisms of action is crucial to tackle the foreseen global health crisis due to antimicrobial resistance. Bacterial two-component signaling systems (TCSs) are attractive targets for the discovery of novel antibacterial agents. TCS-encoding genes are found in all bacterial genomes and typically consist of a sensor histidine kinase (HK) and a response regulator. Due to the conserved Bergerat fold in the ATP-binding domain of the TCS HK and the human chaperone Hsp90, there has been much interest in repurposing inhibitors of Hsp90 as antibacterial compounds. In this study, we explore the chemical space of the known Hsp90 inhibitor scaffold 3,4-diphenylpyrazole (DPP), building on previous literature to further understand their potential for HK inhibition. Six DPP analogs inhibited HK autophosphorylation in vitro and had good antimicrobial activity against Gram-positive bacteria. However, mechanistic studies showed that their antimicrobial activity was related to damage of bacterial membranes. In addition, DPP analogs were cytotoxic to human embryonic kidney cell lines and induced the cell arrest phenotype shown for other Hsp90 inhibitors. We conclude that these DPP structures can be further optimized as specific disruptors of bacterial membranes providing binding to Hsp90 and cytotoxicity are lowered. Moreover, the X-ray crystal structure of resorcinol, a substructure of the DPP derivatives, bound to the HK CheA represents a promising starting point for the fragment-based design of novel HK inhibitors.
IMPORTANCE
The discovery of novel antimicrobials is of paramount importance in tackling the imminent global health crisis of antimicrobial resistance. The discovery of novel antimicrobials with novel mechanisms of actions, e.g., targeting bacterial two-component signaling systems, is crucial to bypass existing resistance mechanisms and stimulate pharmaceutical innovations. Here, we explore the possible repurposing of compounds developed in cancer research as inhibitors of two-component systems and investigate their off-target effects such as bacterial membrane disruption and toxicity. These results highlight compounds that are promising for further development of novel bacterial membrane disruptors and two-component system inhibitors.
Enhancing virtual screening enrichment has become an urgent problem in computational chemistry, driven by increasingly large databases of commercially available compounds, without a commensurate drop in in vitro screening costs. Docking these large databases is possible with cloud-scale computing. However, rapid docking necessitates compromises in scoring, often leading to poor enrichment and an abundance of false positives in docking results. This work describes a new scoring function composed of two parts – a knowledge-based component that predicts the probability of a particular atom type being in a particular receptor environment; and a tunable weight matrix that converts the probability predictions into a dimensionless score suitable for virtual screening enrichment. This score, the FitScore, represents the compatibility between the ligand and the binding and is capable of a high degree of enrichment across standardized docking test sets.
The vast biological and biochemical diversity of the global ocean is the driver behind marine bioprospecting for novel bioproducts. As Marine Biotechnology is gaining momentum as one of the main pillars of the ‘Brue Growth’ revolution, the ability to screen for novel compounds of interest in species with little or no genomic resources is paramount. With this respect, proteins, which are easily metabolised, can be synthetised using convenient DNA recombinant methods and can easily be modified to better meet the needs of human society, making them prized targets. Evidently, proteins that hold natural bioactivity and specificity such as toxins and other venom components, have long captured the focus of biotechnologists, leading to the merger between environmental omics and toxinology termed as ‘venomics’. Indeed, bioactive proteins such as conopeptides, conotoxins, turripeptides and others are long deemed important subjects of research. Even though current mainstream paradigms set the focus on secondary metabolites from marine organisms, transcriptomics and proteomics approaches and their combination are rising strategies for screening for thousands of proteins and peptides in non-conventional biological models, emphasising, but not limited to, marine invertebrate animals due to their abundance, biodiversity and uncanny biochemical strategies to cope with selective pressure in literally every known marine habitat. Untargeted approaches, such as RNA-Seq – based transcriptomics and tandem mass spectrometry – based proteomics, can circumvent limitations related with absent or reduced genomic annotation. The present review will outline the main contributions of ‘omics’ and computational approaches for bioprospecting for proteinaceous marine bioactives. Despite the relatively low number of ‘omics’ studies with the main purpose of discover novel compounds, there is already important literature showcasing pipelines and approaches for revolutionising the exploration of the ocean.
Natural products have long been an important source of inspiration for medicinal chemistry and drug discovery. In the cosmetic field, they remain the major elements of the composition and serve as marketing asset. Recent research showed the implication of salt‐inducible kinases on the melanin production in skin via MITF regulation. Finding new potent modulators on such target could open the way to several cosmetic applications to attenuate visible signs of photoaging and improve the tan without sun. Since virtual screening can be a powerful tool for detecting hit compounds in the early stages of a drug discovery process, we applied this method on salt‐inducible kinase 2 to discover potential interesting compounds. Here, we present the different steps from the construction of a database of natural products, to the validation of a docking protocol and the results of the virtual screening. Hits from the screening were tested in vitro to confirm their efficiency and results are discussed.
Given the diverse pathophysiological mechanisms underlying Alzheimer’s disease, it is improbable that a single targeted drug will prove successful as a therapeutic strategy. Therefore, exploring various hypotheses in drug design is imperative. The sequestration of Fe(II) and Zn(II) cations stands out as a crucial mechanism based on the mitigation of reactive oxygen species. Moreover, inhibiting acetylcholinesterase represents a pivotal strategy to enhance acetylcholine levels in the synaptic cleft. This research aims to investigate the analogs of Huperzine A, documented in scientific literature, considering of these two hypotheses. Consequently, the speciation chemistry of these structures with Fe(II) and Zn(II) was scrutinized using quantum chemistry calculations, molecular docking simulations, and theoretical predictions of pharmacokinetics properties. From the pharmacokinetic properties, only two analogs, HupA-A1 and HupA-A2, exhibited a theoretical permeability across the blood-brain barrier; on the other hand, from a thermodynamic standpoint, the enantiomers of HupA-A2 showed negligible chelation values. The enantiomers with the most favorable interaction parameters were S′R′HupA-A1 (ΔGBIND = −40.0 kcal mol−1, fitness score = 35.5) and R′R′HupA-A1 (ΔGBIND = −35.5 kcal mol−1, fitness score = 22.61), being compared with HupA (ΔGBIND = −41.75 kcal mol−1, fitness score = 39.95). From this study, some prime candidates for promising drug were S′R′HupA-A1 and R′R′HupA-A1, primarily owing to their favorable thermodynamic chelating capability and potential anticholinesterase mechanism.
Quantum chemistry calculations were carried out at B3LYP/6-31G(d) level, considering the IEF-PCM(UFF) implicit solvent model for water. The coordination compounds were assessed using the Gibbs free energy variation and hard and soft acid theory. Molecular docking calculations were conducted using the GOLD program, based on the crystal structure of the acetylcholinesterase protein (PDB code = 4EY5), where the ChemScore function was employed with the active site defined as the region within a 15-Å radius around the centroid coordinates (X = −9.557583, Y = −43.910473, Z = 31.466687). Pharmacokinetic properties were predicted using SwissADME, focusing on Lipinski’s rule of five.
Cyclic tetrapeptides c(Pro-Phe-Pro-Phe) obtained by the mechanosynthetic method using a ball mill were isolated in a pure stereochemical form as a homochiral system (all L-amino acids, sample A) and as a heterochiral system with D configuration at one of the stereogenic centers of Phe (sample B). The structure and stereochemistry of both samples were determined by X-ray diffraction studies of single crystals. In DMSO and acetonitrile, sample A exists as an equimolar mixture of two conformers, while only one is monitored for sample B. The conformational space and energetic preferences for possible conformers were calculated using DFT methods. The distinctly different conformational flexibility of the two samples was experimentally proven by Variable Temperature (VT) and 2D EXSY NMR measurements. Both samples were docked to histone deacetylase HDAC8. Cytotoxic studies proved that none of the tested cyclic peptide is toxic.
Computational algorithms and tools have retrenched the drug discovery and development timeline.
Many drug formulations containing small active molecules are used for the treatment of coronary artery disease, which affects a significant part of the world’s population. However, the inadequate profile of these molecules in terms of therapeutic efficacy has led to the therapeutic use of protein and peptide-based biomolecules with superior properties, such as target-specific affinity and low immunogenicity, in critical diseases. Protein‒protein interactions, as a consequence of advances in molecular techniques with strategies involving the combined use of in silico methods, have enabled the design of therapeutic peptides to reach an advanced dimension. In particular, with the advantages provided by protein/peptide structural modeling, molecular docking for the study of their interactions, molecular dynamics simulations for their interactions under physiological conditions and machine learning techniques that can work in combination with all these, significant progress has been made in approaches to developing therapeutic peptides that can modulate the development and progression of coronary artery diseases. In this scope, this review discusses in silico methods for the development of peptide therapeutics for the treatment of coronary artery disease and strategies for identifying the molecular mechanisms that can be modulated by these designs and provides a comprehensive perspective for future studies.
Membrane permeability is a natural defense barrier that contributes to increased bacterial drug resistance, particularly for Gram-negative pathogens. As such, accurate delivery of the antibacterial agent to the target has become a growing research area in the infectious diseases field as a means of improving drug efficacy. Although the efficient transport of siderophore-antibiotic conjugates into the cytosol still remains challenging, great success has been achieved in the delivery of β-lactam antibiotics into the periplasmic space via bacterial iron uptake pathways. Cefiderocol, the first siderophore-cephalosporin conjugate approved by the US Food and Drug Administration, is a good example. These conjugation strategies have also been applied to the precise delivery of β-lactamase inhibitors, such as penicillin-based sulfone 1, to restore β-lactam antibiotic efficacy in multidrug-resistant bacteria. Herein, we have explored the impact on the bacterial activity of 1 by modifying its iron chelator moiety. A set of derivatives functionalized with diverse iron chelator groups and linkages to the scaffold (compounds 2–8) were synthesized and assayed in vitro. The results on the ability of derivatives 2–8 to recover β-lactam antibiotic efficacy in difficult-to-treat pathogens that produce various β-lactamase enzymes, along with kinetic studies with the isolated enzymes, allowed us to identify compound 2, a novel β-lactamase inhibitor with an expanded spectrum of activity. Molecular dynamics simulation studies provided us with further information regarding the molecular basis of the relative inhibitory properties of the most relevant compound described herein.
Molecular docking advances early-stage drug discovery by predicting the geometries and affinities of small-molecule compounds bound to drug-target receptors, predictions that researchers can leverage in prioritizing drug candidates for experimental testing. Unfortunately, existing docking tools often suffer from poor usability, data security, and maintainability, limiting broader adoption. Additionally, the complexity of the docking process, which requires users to execute a series of specialized steps, often poses a substantial barrier for non-expert users. Here, we introduce MolModa, a secure, accessible environment where users can perform molecular docking entirely in their web browsers. We provide two case studies that illustrate how MolModa provides valuable biological insights. We further compare MolModa to other docking tools to highlight its strengths and limitations. MolModa is available free of charge for academic and commercial use, without login or registration, at https://durrantlab.com/molmoda.
Angiotensin‐converting enzyme inhibitors are widely used in treating arterial hypertension, acting on the renin‐angiotensin‐aldosterone system and controlling blood pressure. We present a novel, greener, and faster methodology to assess the 1,2,4‐oxadiazol‐5‐one ring and perform molecular modifications to obtain angiotensin‐converting enzyme (ACE) inhibitors using this heterocyclic core. Molecular docking simulations indicate that the tested compounds exhibited an affinity for the ACE binding site, with scores comparable to the commercial inhibitor lisinopril. However, in vitro assays revealed that the compounds were ineffective in inhibiting ACE activity. The lack of inhibition may be related to the compounds' more apolar nature. These results emphasize the importance of integrating computational and experimental approaches in developing new drugs, providing valuable insights for planning future studies to optimize the activity of synthesized compounds.
An Artificial Metalloenzyme (ArM) built employing the streptavidin‐biotin technology has been used for the enantioselective synthesis of binaphthyls by means of asymmetric Suzuki‐Miyaura cross‐coupling reactions. Despite its success, it remains a challenge to understand how the length of the biotin cofactors or the introduction of mutations to streptavidin leads the preferential synthesis of one atropisomer over the other. In this study, we apply an integrated computational modeling approach, including DFT calculations, protein‐ligand dockings and molecular dynamics to rationalize the impact of mutations and length of the biotion cofactor on the enantioselectivities of the biaryl product. The results unravel that the enantiomeric differences found experimentally can be rationalized by the disposition of the first intermediate, coming from the oxidative addition step, and the entrance of the second substrate. The work also showcases the difficulties facing to control the enantioselection when engineering ArM to catalyze enantioselective Suzuki‐Miyaura couplings and how the combination of DFT calculations, molecular dockings and MD simulations can be used to rationalize artificial metalloenzymes.
The interactions with calf thymus DNA (CT-DNA) of three Schiff bases formed by the condensation of hesperetin with benzohydrazide (HHSB or L1H3), isoniazid (HIN or L2H3), or thiosemicarbazide (HTSC or L3H3) and their CuII complexes (CuHHSB, CuHIN, and CuHTSC with the general formula [CuLnH2(AcO)]) were evaluated in aqueous solution both experimentally and theoretically. UV–Vis studies indicate that the ligands and complexes exhibit hypochromism, which suggests helical ordering in the DNA helix. The intrinsic binding constants (Kb) of the Cu compounds with CT-DNA, in the range (2.3–9.2) × 106, from CuHTSC to CuHHSB, were higher than other copper-based potential drugs, suggesting that π–π stacking interaction due to the presence of the aromatic rings favors the binding. Thiazole orange (TO) assays confirmed that ligands and Cu complexes displace TO from the DNA binding site, quenching the fluorescence emission. DFT calculations allow for an assessment of the equilibrium between [Cu(LnH2)(AcO)] and [Cu(LnH2)(H2O)]+, the tautomer that binds CuII, amido (am) and not imido (im), and the coordination mode of HTSC (O−, N, S), instead of (O−, N, NH2). The docking studies indicate that the intercalative is preferred over the minor groove binding to CT-DNA with the order [Cu(L1H2am)(AcO)] > [Cu(L2H2am)(AcO)] ≈ TO ≈ L1H3 > [Cu(L3H2am)(AcO)], in line with the experimental Kb constants, obtained from the UV–Vis spectroscopy. Moreover, dockings predict that the binding strength of [Cu(L1H2am)(AcO)] is larger than [Cu(L1H2am)(H2O)]+. Overall, the results suggest that when different enantiomers, tautomers, and donor sets are possible for a metal complex, a computational approach should be recommended to predict the type and strength of binding to DNA and, in general, to macromolecules.
Antigenic similarities between Zika virus (ZIKV) and other flaviviruses pose challenges to the development of virus-specific diagnostic tools and effective vaccines. Starting with a DNA-encoded one-bead-one-compound combinatorial library of 508,032 synthetic, non-natural oligomers, we selected and characterized small molecules that mimic ZIKV epitopes. High-throughput fluorescence-activated cell sorter-based bead screening was used to select molecules that bound IgG from ZIKV-immune but not from dengue-immune sera. Deep sequencing of the DNA from the “Zika-only” beads identified 40 candidate molecular structures. A lead candidate small molecule “CZV1-1” was selected that correctly identifies serum specimens from Zika-experienced patients with good sensitivity and specificity (85.3% and 98.4%, respectively). Binding competition studies of purified anti-CZV1-1 IgG against known ZIKV-specific monoclonal antibodies (mAbs) showed that CZV1-1 mimics a nonlinear, neutralizing conformational epitope in the domain III of the ZIKV envelope. Purified anti-CZV1-1 IgG neutralized infection of ZIKV in cell cultures with potencies comparable to highly specific ZIKV-neutralizing mAbs. This study demonstrates an innovative approach for identification of synthetic non-natural molecular mimics of conformational virus epitopes. Such molecular mimics may have value in the development of accurate diagnostic assays for Zika, as well as for other viruses.
Schiff bases are most the widely used class of molecules and their metal complexes are utilized as medicinally important scaffolds due to their versatile biological profile. So herein, we have...
Nictaba is a (GlcNAc) n -binding, stress-inducible lectin from Nicotiana tabacum , that serves as the archetypical lectin for the family of Nictaba-related lectins. Nictaba and Nictaba-related lectins play pivotal roles in plant defense mechanisms and stress response pathways. Despite extensive research into the variety of biological activities and physiological role(s) of the lectin, the three-dimensional structure of Nictaba remained largely unknown. Here, we report crystal structures for Nictaba in the apo form and bound to chitotriose. The structures reveal a jelly-roll fold for the Nictaba protomer and the assembly thereof into a novel dimerization interface among lectins. The chitotriose binding mode centers around the central GlcNAc residue providing insights into the determinants of specificity of Nictaba towards carbohydrate structures. Indeed, by integrating such structural insights with inputs from glycan arrays, molecular docking, and molecular dynamics simulations, we propose that Nictaba employs a single carbohydrate-recognition domain within each of the two subunits in the dimer to display pronounced specificity towards GlcNAc-containing carbohydrates. Furthermore, we identified amino acid residues involved in the extended binding site capable of accommodating structurally diverse high-mannose and complex N -glycans, highlighting the lectin potential to recognize N -glycan structures. Glycan array and in silico analyses revealed interactions centered around the conserved Man3GlcNAc2 core, explaining the broad recognition of N -glycan structures. Collectively, the integrated structural and biochemical insights presented here fill a hitherto substantial void into the atlas of lectin structure-function relationships and pave the way for future developments in plant stress biology and lectin-based applications.
The proteins within the human epidermal growth factor receptor (EGFR) family, members of the tyrosine kinase receptor family, play a pivotal role in the molecular mechanisms driving the development of various tumors. Tyrosine kinase inhibitors, key compounds in targeted therapy, encounter challenges in cancer treatment due to emerging drug resistance mutations. Consequently, machine learning has undergone significant evolution to address the challenges of cancer drug discovery related to EGFR family proteins. However, the application of deep learning in this area is hindered by inherent difficulties associated with small‐scale data, particularly the risk of overfitting. Moreover, the design of a model architecture that facilitates learning through multi‐task and transfer learning, coupled with appropriate molecular representation, poses substantial challenges. In this study, we introduce GraphEGFR, a deep learning regression model designed to enhance molecular representation and model architecture for predicting the bioactivity of inhibitors against both wild‐type and mutant EGFR family proteins. GraphEGFR integrates a graph attention mechanism for molecular graphs with deep and convolutional neural networks for molecular fingerprints. We observed that GraphEGFR models employing multi‐task and transfer learning strategies generally achieve predictive performance comparable to existing competitive methods. The integration of molecular graphs and fingerprints adeptly captures relationships between atoms and enables both global and local pattern recognition. We further validated potential multi‐targeted inhibitors for wild‐type and mutant HER1 kinases, exploring key amino acid residues through molecular dynamics simulations to understand molecular interactions. This predictive model offers a robust strategy that could significantly contribute to overcoming the challenges of developing deep learning models for drug discovery with limited data and exploring new frontiers in multi‐targeted kinase drug discovery for EGFR family proteins.
Protein–ligand interactions (PLIs) are essential for cellular activities and drug discovery. But due to the complexity and high cost of experimental methods, there is a great demand for computational approaches to recognize PLI patterns, such as protein–ligand docking. In recent years, more and more models based on machine learning have been developed to directly predict the root mean square deviation (RMSD) of a ligand docking pose with reference to its native binding pose. However, new scoring methods are pressingly needed in methodology for more accurate RMSD prediction. We present a new deep learning-based scoring method for RMSD prediction of protein–ligand docking poses based on a Graphormer method and Shell-like graph architecture, named GSScore. To recognize near-native conformations from a set of poses, GSScore takes atoms as nodes and then establishes the docking interface of protein–ligand into multiple bipartite graphs within different shell ranges. Benefiting from the Graphormer and Shell-like graph architecture, GSScore can effectively capture the subtle differences between energetically favorable near-native conformations and unfavorable non-native poses without extra information. GSScore was extensively evaluated on diverse test sets including a subset of PDBBind version 2019, CASF2016 as well as DUD-E, and obtained significant improvements over existing methods in terms of RMSE, $R$ (Pearson correlation coefficient), Spearman correlation coefficient and Docking power.
The x-ray crystal structures of normal human transthyretin (prealbumin) and the amyloidogenic Val-30-Met variant have been refined at 1.7-A resolution to R-values of 0.168 and 0.179, respectively, for 19,882 and 20,362 reflections (Fobs > 2.0 sigma). Standard deviations for stereochemical parameters are 0.018 and 0.022 A for bond distances, 0.030 and 0.038 A for angle distances, and 0.035 and 0.070 A for planar 1-4 distances. The newly refined normal structure shows improvement over the original structure of Blake and Swan (Blake, C. C. F., and Swan, I. D. A. (1971) J. Mol. Biol. 61, 217-224) in stereochemistry and in the conformation of the loop regions. Residues Arg-103, Thr-123, Asn-124, and Pro-125 have now been resolved, and residues 1-9 and 126-127 have been modeled with the aid of simulated annealing refinement. The functional form of transthyretin is a tetramer, having a cylindrical cavity which will bind thyroxine and an exterior binding site for the complex of retinol with retinol-binding protein. The monomer is a beta barrel flattened to become more like a sandwich with residue 30 in the interior. The methionyl for valyl substitution forces the beta sheets of the monomer as much as 1 A apart, resulting in a distortion of the thyroxine-binding cavity, in agreement with the independent observations that the Met-30 variant has low affinity for thyroxine.
As part of a structure-based drug design program directed against enzyme targets in the human immunodeficiency virus (HIV), we have determined the three-dimensional structures of the HIV type 1 protease complexed with two hydroxyethylene-based inhibitors. The inhibitors (SKF 107457 and SKF 108738) are hexapeptide substrate analogues with the scissile bond being replaced by a hydroxyethylene isostere. The structures were determined using x-ray diffraction data to 2.2 A measured at the Cornell High Energy Synchrotron Source on hexagonal crystals of each of the complexes. The structures have been extensively refined using a reciprocal space least-squares method to conventional crystallographic R factors of 0.186 and 0.159, respectively. The protein structure differs from that in the unliganded state of the enzyme and is most similar to that of the structure of the other reported (Jaskolski, M., Tomasselli, A. G., Sawyer, T. K., Staples, D. G., Heinrikson, R. L., Schneider, J., Kent, S. B. H., and Wlodawer, A. (1990) Biochemistry 29, 5889-5907) hydroxyethylene-based inhibitor complex. Unlike in that structure, however, the inhibitors are observed, in the present crystal structures, in two equally abundant orientations that are a consequence of the homodimeric nature of the enzyme coupled with the asymmetric structures of the inhibitors. Although the differences between the two inhibitors used in the present study are confined to the P1' site, the van der Waals interactions made by the inhibitor atoms with the amino acid residues in the protein differ throughout the structures of the inhibitors.
The crystal structure of a class A beta-lactamase from Staphylococcus aureus PC1 has been refined at 2.0 A resolution. The resulting crystallographic R-factor (R = sigma h parallel Fo[-]Fc parallel/sigma h[Fo], where [Fo] and [Fc] are the observed and calculated structure factor amplitudes, respectively), is 0.163 for the 17,547 reflections with I greater than or equal to 2 sigma (I) within the 8.0 A to 2.0 A resolution range. The molecule consists of two closely associated domains. One domain is formed by a five-stranded antiparallel beta-sheet with three helices packing against a face of the sheet. The second domain is formed mostly by helices that pack against the second face of the sheet. The active site is located in the interface between the two domains, and many of the residues that form it are conserved in all known sequences of class A beta-lactamases. Similar to the serine proteases, an oxyanion hole is implicated in catalysis. It is formed by two main-chain nitrogen atoms, that of the catalytic seryl residue, Ser70, and that of Gln237 on an edge beta-strand of the major beta-sheet. Ser70 is interacting with another conserved seryl residue, Ser130, located between the two ammonium groups of the functionally important lysine residues, Lys73 and Lys234. Such intricate interactions point to a possible catalytic role for this second seryl residue. Another key catalytic residue is Glu166. There are several unusual structural features associated with the active site. (1) A cis peptide bond has been identified between the catalytic Glu166 and Ile167. (2) Ala69 and Leu220 have strained phi, psi dihedral angles making close contacts that restrict the conformation of the active site beta-strand involved in the formation of the oxyanion hole. (3) A buried aspartate residue, the conserved Asp233, is located next to the active site Lys234. It is interacting with another buried aspartyl residue, Asp246. An internal solvent molecule is also involved, but the rest of its interactions with the protein indicate it is not a cation. (4) Another conserved aspartyl residue that is desolvated is Asp131, adjacent to Ser130. Its charge is stabilized by interactions with four main-chain nitrogen atoms. (5) An internal cavity underneath the active site depression is filled with six solvent molecules. This, and an adjacent cavity occupied by three solvent molecules partially separate the omega-loop associated with the active site from the rest of the protein.(ABSTRACT TRUNCATED AT 250 WORDS)
A set of algorithms designed to enhance the display of protein binding cavities is presented. These algorithms, collectively entitled CAVITY SEARCH, allow the user to isolate and fully define the extent of a particular cavity. Solid modeling techniques are employed to produce a detailed cast of the active site region, which can then be color-coded to show both electrostatic and steric interactions between the protein cavity and a bound ligand.
Rat intestinal fatty acid binding protein (I-FABP) is a 131-residue protein composed of two short alpha-helices (alphaI and alphaII) and 10 anti-parallel beta-strands organized into two nearly orthogonal beta-sheets. The structure of crystalline I-FABP with bound tetradecanoate (myristate) has been refined to a resolution of 1.5 angstrom and compared to the 1.2 angstrom structure of apo-I-FABP, the 1.9 angstrom structure of I-FABP:hexadecanoate (palmitate) and the 1.75 angstrom structure of I-FABP:9Z-octadecanoate (oleate) to determine how this model fatty acid receptor accommodates changes in the length of its fatty acid ligand. Myristate is located in the interior of the protein. A highly ordered, electrostatic network containing 7 hydrogen (H)-bonds links the OE1 and OE2 atoms of myristate's carboxylate group, the indole nitrogen of Trp82, NH1, and NH2 of Arg106, NE2, and OE1 of Gln115, and 2 interior ordered waters. The hydrocarbon chain of the bound fatty acid is slightly bent. Its convex face lies in a crevice, forming van der Waals contacts with the side chains of several hydrophobic and aromatic residues. Its concave face is exposed to an array of 8 interior ordered waters whose positions are stabilized by H-bond interactions with other waters, H-bond interactions with the side chains of polar/ionizable residues, and van der Waals contacts with the surface of the fatty acid. Addition of 2 or 4 methylenes to myristate produces remarkably little change in the positions of I-FABP's main chain and side chain atoms and interior ordered waters. The principal alterations are in the conformation of a surface opening (portal) connecting external and internal solvent and in the position of the benzene side chain of Phe55. Changes in the conformation of the portal reflect movement of two of its components: the backbone of alphaII and a type I turn (Ala73, Asp74) connecting two beta-strands. The positions of the main chain atoms of Ala73 and Asp74 appear to be determined by their ability to form van der Waals contacts with the omega-terminus of the fatty acid. The side chain of Phe55 appears to function as an adjustable aromatic lid, located over the portal, whose position is dependent on an ability to form van der Waals contacts with a fatty acid's omega-terminus. The structure of I-FABPArg106-->Gln with bound oleate was refined to 1.74 angstrom and compared to the 1.75 angstrom structure of I-FABpArg106:oleate so that we could assess the contribution of the electrostatic network, which binds carboxylate, to the overall positioning of the fatty acid. This mutation results in destabilization of the electrostatic network as evidenced by the presence of several alternate positions for carboxylate in the electron density map. The discrete disorder of the carboxylate group is associated with movement of the hydrocarbon chain along the length of the binding pocket and extension of the methyl terminus beyond the boundary of the aromatic lid. Movement of the acyl chain does not perturb the positions of the protein's interior ordered waters or the side chain of its Phe55 residue. These findings provide insights about the potential manner in which I-FABP can accommodate fatty acids whose chain length exceeds 18 carbons. Together, our studies emphasize the contributions of ''feeble forces'' (ion pair, H-bond, and van der Waals interactions) to the binding enthalpy in this model fatty acid receptor system.
We describe a new paradigm for modeling proteins to interactive computer graphics systems-continual maintenance of a physically valid representation, combined with direct user control and visualization. This is achieved by a fast algorithm for energy minimization, capable of real-time performance on all atoms of a small protein, plus graphically specified user tugs. The modeling system, called Sculpt, rigidly constrains bond lengths, bond angles, and planar groups (similar to existing interactive modeling programs), while it applies elastic restraints to minimize the potential energy due to torsions, hydrogen bonds, and van der Waals and electrostatic interactions (similar to existing batch minimization programs), and user-specified springs. The graphical interface can show bad and/or favorable contacts, and individual energy terms can be turned on or off to determine their effects and interactions. Sculpt finds a local minimum of the total energy that satisfies all the constraints using an augmented Lagrange-multiplier method; calculation time increases only linearly with the number of atoms because the matrix of constraint gradients is sparse and banded. On a 100-MHz MIPS R4000 processor (Silicon Graphics Indigo), Sculpt achieves 11 updates per second on a 20-residue fragment and 2 updates per second on an 80-residue protein, using all atoms except non-H-bonding hydrogens, and without electrostatic interactions. Applications of Sculpt are described: to reverse the direction of bundle packing in a designed 4-helix bundle protein, to fold up a 2- stranded β-ribbon into an approximate β-barrel, and to design the sequence and conformation of a 30-residue peptide that mimics one partner of a protein subunit interaction. Computer models that are both interactive and physically realistic (within the limitations of a given force field) have 2 significant advantages: (1) they make feasible the modeling of very large changes (such as needed for de novo design), and (2) they help the user understand how different energy terms interact to stabilize a given conformation. The Sculpt paradigm combines many of the best features of interactive graphical modeling, energy minimization, and actual physical models, and we propose it as an especially productive way to use current and future increases in computer speed.
Changes of torsion angles and deformations of valence angles, especially in coordination polyhedra, are mainly responsible for the inherent conformational flexibility of molecules. They allow them to adopt various shapes and to adapt to the requirements in a particular molecular environment. The latter aspect makes these systems only tractable by computational methods with an enormous effort. Focusing on common molecular fragments imbedded in various environments in crystal structures allows one to study possible deformations, to elucidate conformational preferences (low-energy conformations), and to trace conformational interconversions. Large data sets of crystal structures can be analyzed by means of statistical methods to discover correlations and systematics among the molecular dimensions. In this overview several systems are described which show the scope of the approach. Conformational studies of rigid phenyl rotors reveal deviating properties depending on the actual constitution of the fragments (Ph-X-Ph, Ph-X-Y-Ph, Ph3X). In these and in more complex systems like metal phosphine complexes, the phenyl groups perform various concerted movements related to different gearing motions of interlocked cog-wheels. Conformational properties and low-energy conformers of ring systems are accessible from the evaluation of crystal data. Also larger and unsymmetric systems composed of cyclic and open-chain portions can be studied by these methods. Changes of the valence angle relationships at a particular atomic center enable molecules to access different areas of conformational space. Possible deformations of coordination geometries, observed in crystal structures, are discussed for examples with three- up to eightcoordination. Interconversion pathways among different coordination geometries can be mapped which explain putative reaction pathways or the fluxional behavior of coordination complexes that are known to occur under solution or gas-phase conditions.
A perturbational method is described for calculating the interaction energy of two molecules in the region where the overlap between their wave-functions is significant. By working directly with a basis of determinants constructed from the SCF orbitals of the separated molecules, without orthogonalization, it is possible to avoid many of the disadvantages of other methods.
Background: Bacterial 3α, 20β-hydroxysteroid dehydrogenase (3α, 20β-HSD) reversibly oxidizes the 3α and 20β hydroxyl groups of androstanes and pregnanes and uses nicotinamide adenine dinucleotide as a cofactor. 3α, 20β-HSD belongs to a family of short-chain dehydrogenases that has a highly conserved Tyr-X-X-X-Lys sequence. The family includes mammalian enzymes involved in hypertension, digestion, fertility and sperm atogenesis. Several members of the enzyme family, including 3α, 20β-HSD, are competitively inhibited by glycyrrhizic acid, a steroidal compound found in licorice, and its derivative, carbenoxolone, ananti-inflammatory glucocorticoid. Results The three-dimensional structure of the enzyme-carbenoxolone complex has been determined and refined at 2.2 å resolution to a crystallographic R-factor of 19.4%. The hemisuccinate side chain of carbenoxolone makes a hydrogen bond with the hydroxyl group of the conserved residue Tyr152 and occupies the position of the nicotinamide ring of the cofactor. The occupancies of the inhibitor in four independent catalytic sites refine to 100%, 95%, 54% and 36%. Conclusion The steroid binds at the catalytic site in a mode much like the previously proposed mode of binding of the substrate cortisone. No bound cofactor molecules were found. The varying occupancy of steroid molecules observed in the four catalytic sites is either due to packing differences or indicates a cooperative effect among the four sites. The observed binding accounts for the inhibition of 3α,20β-HSD.
The Protein Data Bank is the international depository for the results of structural studies of biological macromolecules. At present, some 1200 biological macromolecules are reported to have been crystallized, of which approximately 350 have structures determined to sufficient resolution that an atomic model has been generated. The Data Bank currently holds over 260 coordinate entries, with bibliographic entries for the remaining structures. For approximately one-third of the coordinate entries, associated structure factors are stored in the Bank. Most of the coordinate entries in the Bank are based directly on crystallographic studies. The Bank also includes a small number of model structures which have been generated either by theoretical analysis and computer simulation or by analogy with structures of chemically homologous molecules. Although the first structures deposited were proteins, we now distribute coordinates for DNA, tRNA, and polysaccharides as well. In quite a few instances, several data sets are held for a macromolecule obtained from different organisms, or in different states. 1 reference, 8 tables. (ACR)
The nature of intermolecular interactions between carbon-bonded halogens (C−X, X = F, Cl, Br, or I) and electronegative atoms (El = N, O and S) has been analysed, focusing on the role of specific attractive forces and the anisotropic repulsive wall around halogen atoms. Searches of the Cambridge Structural Database show that electronegative atoms in various hybridization states clearly prefer to form contacts to Cl, Br, and I (but not F) in the direction of the extended C−X bond axis, at interatomic distances less than the sum of the van der Waals radii. Ab initio intermolecular perturbation theory calculations show that the attractive nature of the X···El interaction is mainly due to electrostatic effects, but polarization, charge-transfer, and dispersion contributions all play an important role. The magnitude of the interaction for the chloro-cyanoacetylene dimer is about 10 kJ/mol, demonstrating the potential importance of these kinds of nonbonded interactions. The directionality of the interaction is explained by the anisotropic electron distribution around the halogen atom, causing a decreased repulsive wall and increased electrostatic attraction for electronegative atoms in the observed preferred position. In contrast, carbon-bonded hydrogens show no directionality in their contacts to the halogen atoms, because the angular dependence of the electrostatic energy is reversed and acts to counter rather then to reinforce the effect of the anisotropic repulsive wall.
Electronegativity is discussed on the basis of Mulliken's definition (χ = Ev + Iv), which leads to the conclusion, that it is not a property of atoms in their ground state, but of atoms in the same conditions in which they are found in molecules, the valence state. Valence state promotion energies are calculated and reported for a large variety of states of the atoms and ions of the first and second period. Combining these promotion energies with ionization potentials and electron affinities yields the electronegativities of a number of valence states. It is found that electronegativity can be defined in this way only for bonding orbitals, and the term "orbital electronegativity" is suggested for the values listed. The calculated orbital electronegativities for σ orbitals are found to be higher in every case than for π orbitals, and to be linearly related to the amount of s character in the hybrid orbitals. As expected, the electronegativity increases with increasing s character of the orbital considered.
The design and synthesis of high-affinity FKBP12 ligands is described. These compounds potently inhibit the cis-trans-peptidylprolyl isomerase (rotamase) activity catalyzed by FKBP12 with inhibition constants (K i,app ) as low as 1 nM, yet they possess remarkable structural simplicity relative to FK506 and rapamycin, from which they are conceptually derived. The atomic structures of three FKBP12-ligand complexes and of one unbound ligand were determined by X-ray crystallography and are compared to the FKBP12-FK506 and FKBP12-rapamycin complexes
The complex of ribonuclease A (RNase A) with uridine vanadate (U-V), a transition-state analogue, has been studied with 51V and proton NMR spectroscopy in solution and by neutron diffraction in the crystalline state. Upon the addition of aliquots of U-V at pH 6.6, the C'-H resonances of the two active-site histidine residues 119 and 12 decrease in intensity while four new resonances appear. Above pH 8 and below pH 5, these four resonances decrease in intensity as the complex dissociates. These four resonances are assigned to His- 1 19 and His- 12 in protonated and unprotonated forms in the RNase-U-V complex. These resonances do not titrate or change in relative area in the pH range 5-8, indicating a slow protonation process, and the extent of protonation remains constant with ca. 58% of His-12 and ca. 26% of His- 119 being protonated. The results of diffraction studies show that both His-1 2 and His-1 19 occupy well-defined positions in the RNase-U-V complex and that both are protonated. However, while the classic interpretation of the mechanism of action of RNase based on the proposal of Findlay et al. (Findlay, D., Herries, D. G., Mathias, A. P., Rabin, B. R., & Ross, C. A. (1962) Biochem. J. 85, 152-1531 requires both His-12 and His-1 19 to be in axial positions relative to the pentacoordinate transition state, in the diffraction structure His-12 is found to be in an equatorial position, while Lys-41 is close to an axial position. Hydrogen exchange data show that the mobility and accessibility of amides in the RNase-U-V complex do not significantly differ from what was observed in the native enzyme. The results of both proton NMR in solution and neutron diffraction in the crystal are compared and interpreted in terms of the mechanism of action of RNase.
The mechanism of irreversible inactivation of mandelate racemase (MR) from Pseudomonas putida by alpha-phenylglycidate (alpha PGA) has been investigated stereochemically and crystallographically. The (R) and (S) enantiomers of alpha PGA were synthesized in high enantiomeric excess (81% ee and 83% ee, respectively) using Sharpless epoxidation chemistry. (R)-alpha PGA was determined to be a stereospecific and stoichiometric irreversible inactivator of MR. (S)-alpha PGA does not inactivate MR and appears to bind noncovalently to the active site of MR with less affinity than that of (R)-alpha PGA. The X-ray crystal structure (2.0-A resolution) of MR inactivated by (R)-alpha PGA revealed the presence of a covalent adduct formed by nucleophilic attack of the epsilon-amino group of Lys 166 on the distal carbon on the epoxide ring of (R)-alpha PGA. The proximity of the alpha-proton of (S)-mandelate to Lys 166 [configurationally equivalent to (R)-alpha PGA] was corroborated by the crystal structure (2.1-A resolution) of MR complexed with the substrate analog/competitive inhibitor, (S)-atrolactate [(S)-alpha-methylmandelate]. These results support the proposal that Lys 166 is the polyvalent acid/base responsible for proton transfers on the (S) face of mandelate. In addition, the high-resolution structures also provide insight into the probable interactions of mandelate with the essential Mg2+ and functional groups in the active site.
A ribonuclease T1 homologue, ribonuclease Ms (RNase Ms) from Aspergillus saitoi, has been crystallized as a complex with a substrate analogue GfpC where the 2'-hydroxyl (2'-OH) group of guanosine in guanylyl-3',5'-cytidine (GpC) is replaced by the 2'-fluorine (2'-F) atom to prevent transesterification. The crystal structure of the complex was solved at 1.8-angstrom resolution to a final R-factor of 0.204. The role of His92 (RNase T1 numbering) as the general acid catalyst was confirmed. Of the two alternative candidates for a general base to abstract a proton from the 2'-OH group, His40 and Glu58 were found close to the 2'-F atom, making the decision between the two groups difficult. We then superposed the active site of the RNase Ms/GfpC complex with that of pancreatic ribonuclease S (RNase S) complexed with a substrate analogue UpcA, a phosphonate analogue of uridylyl-3',5'-adenosine (UpA), and found that His12 and His119 of RNase A almost exactly coincided with Glu58 and His92, respectively, of RNase Ms. Similar superposition with a prokaryotic microbial ribonuclease, RNase St [Nakamura, K. T., Iwahashi, K., Yamamoto, Y., Iitaka, Y., Yoshida, N., & Mitsui, Y. (1982) Nature 299, 564-566], also indicated Glu58 as a general base. Thus the present comparative geometrical studies consistently favor, albeit indirectly, the traditional as well as the most recent notion [Steyaert, J., Hallenga, K., Wyns, L., & Stanssens, P. (1990) Biochemistry 29, 9064-9072] that Glu58, rather than His40, must be the general base catalyst in the intact enzymes of the RNase T1 family.
A molecular mechanics force field implemented in the Sybyl program is described along with a statistical evaluation of its efficiency on a variety of compounds by analysis of internal coordinates and thermodynamic barriers. The goal of the force field is to provide good quality geometries and relative energies for a large variety of organic molecules by energy minimization. Performance in protein modeling was tested by minimizations starting from crystallographic coordinates for three cyclic hexapeptides in the crystal lattice with rms movements of 0.019 angstroms, 2.06 degrees, and 6.82 degrees for bond lengths, angles, and torsions, respectively, and an rms movement of 0.16 angstroms for heavy atoms. Isolated crambin was also analyzed with rms movements of 0.025 angstroms, 2.97 degrees, and 13.0 degrees for bond lengths, angles, and torsions respectively, and an rms movement of 0.42 angstroms for heavy atoms. Accuracy in calculating thermodynamic barriers was tested for 17 energy differences between conformers, 12 stereoisomers, and 15 torsional barriers. The rms errors were 0.8, 1.7, and 1.13 kcal/mol, respectively, for the three tests. Performance in general purpose applications was assessed by minimizing 76 diverse complex organic crystal structures, with and without randomization by coordinate truncation, with rms movements of 0.025 angstroms, 2.50 degrees, and 9.54 degrees for bond lengths, angles and torsions respectively, and an average rms movement of 0.192 angstroms for heavy atoms.
Many approaches have been developed for solving the docking problem: Predict the structure and binding free energy of a ligand-receptor complex given only the structures of the free ligand and receptor. We review major approaches for docking small-molecule ligands to receptors and focus on the successes and limitations of their application to drug design.
A computer algorithm is presented for calculating the part of the van der Waals surface of molecule that is accessible to solvent. The solvent molecule is modeled by a sphere. This sphere is, in effect, rolled over the molecule to generate a smooth outer-surface contour. This surface contour is made up of pieces of spheres and tori that join at circular arcs. The spheres, tori and arcs are defined by analytical expressions in terms of the atomic coordinates, van der Waals radii and the probe radius. The area of each surface piece may be calculated analytically and the surface may be displayed on either vector or raster computer-graphics systems. These methods are useful for studying the structure and interactions of proteins and nucleic acids.
An empirical solvation model that allows for the elimination of solvent degrees of freedom in molecular dynamics (MD) simulations of biomolecules is proposed. The potential of mean force due to the first solvation shell is approximated by means of a simple, easily derivable analytic function of the solvent-accessible surface area of the molecule. The solvent contribution to the free energy is evaluated by means of only two atomic solvation parameters. This approach requires about 30% more computational effort than an in vacuo simulation, but a factor of 10 to 50 less than a MD simulation involving solvation by explicit water molecules. The implicit solvation model is assessed by application to proteins of different size. Average structural properties are calculated and compared to values obtained from X-ray structures and from MD simulations using explicit water molecules. The complementarity of the implicit solvation force and the intra-solute force field has been checked. The artefacts induced by the use of a vacuum boundary condition without solvation force in a MD simulation are considerably reduced.
We describe a computational method for docking flexible molecules into protein binding sites. The method uses a genetic algorithm (GA) to search the combined conformation/orientation space of the molecule to find low energy conformations. Several techniques are described that increase the efficiency of the basic search method. These include the use of several interacting GA subpopulations or niches; the use of a “growing” algorithm that initially docks only a small part of the molecule, and the use of gradient minimization during the search. To illustrate the method, we dock Cbz-GlyP-Leu-Leu (ZGLL) into thermolysin. This system was chosen because a well refined crystal structure is available and because another docking method had previously been tested on this system. Our method is able to find conformations that lie physically close to and in some cases lower in energy than the crystal conformation in reasonable periods of time on readily available hardware.
We have performed Hayes—Stone intermolecular perturbation theory (IMPT) calculations on amide…water and amide…amide complexes in order to estimate the change ΔW in intermolecular interaction energy associated with the hydrogen bond exchange process amide(NH)…water+water…(OC)amide⇔amide(NH)…(OC)amide+water…water. ΔW is found to be small and varies by almost 5 kJ/mol and in sign for the amides formamide, acetamide, N-methyl formamide and N-methyl acetamide. The main variations in the amide hydrogen bond energies occur in the electrostatic and exchange-repulsion contributions. This reflects the variation in the charge distributions of the hydrogen bonding groups between the different amides. Thus, we cannot quantify an isolated hydrogen bond strength with any great accuracy, and care must be used in extrapolating model potentials based on small model systems to peptides and proteins.
The conformational space of a flexible three-dimensional (3-D) molecule can be represented for searching purposes by a smoothed bounded-distance matrix. Such matrices provide an effective way of carrying out flexible searching, but search times can be much greater than with comparable rigid searches that take no account of conformational flexibility. The most time-consuming part of a flexible search is the final conformational search, and in this paper we compare the efficiency and the effectiveness of the distance geometry, systematic search, random search, genetic-algorithm, and directed-tweak methods for conformational searching. Experiments with sets of 1538 and 9886 flexible 3-D structures suggest that the most effective of these are the genetic-algorithm and directed-tweak methods, both of which are efficient enough to enable the searching of databases containing nontrivial numbers of flexible 3-D structures.
The Cambridge Structural Database (CSD) records bibliographic, 2D chemical, and 3D structural results for organocarbon compounds studied by X-ray and neutron diffraction. In January 1991, the CSD contained 86 026 entries derived from 584 primary sources. The CSD system comprises the database, together with associated software for search, retrieval, analysis, and display of the stored information. Over the past few years, the CSD system has been considerably upgraded to provide efficient, integrated, and user-friendly search facilities. The development of two new systems, Version 3 and Version 4 are described here. Both systems operate from a completely restructured CSD search file (ASER) upgraded to include bit-screen heuristics (Version 3), and further upgraded by a complete set of fully digitized 2D chemical diagram representations of database entries (Version 4). A new program, QUEST, provides integrated search facilities for text, numeric, and 2D chemical information; 2D chemical similarity searching is also included. The alphanumeric query language of Version 3 QUEST is replaced by a fully interactive menu-driven graphical interface in Version 4 in which high-quality 2D chemical diagrams form part of standard system output. Subsequent 3D search and data analysis operations are performed by the program GSTAT, common to both new systems; 3D graphical output is generated by the PLUTO package. A "general" entry-sequential search file permits both versions to operate on a wide variety of hardware platforms with minimal installation problems. Special implementations have also been developed for DEC-VAX/VMS and Silicon Graphics (Unix) environments, to take advantage of more efficient, machine-specific search file organizations. Plans for a further database upgrade and for the integration of extended 3D search capabilities into QUEST (Version 5) are briefly discussed.
A genetic algorithm has been developed for molecular mechanics calculations. It has been proved to be a robust and efficient structure optimization technique. Because it uses randomly generated starting structures and stochastic operators, the resulting structures are not subjected to the chemist's bias. © 1994 by John Wiley & Sons, Inc.
A genetic algorithm (GA) conformation search method is used to dock a series of flexible molecules into one of three proteins. The proteins examined are thermolysin (tmn), carboxypeptidase A (cpa), and dihydrofolate reductase (dfr). In the latter two proteins, the crystal ligand was redocked. For thermolysin, we docked eight ligands into a protein conformation derived from a single crystal structure. The bound conformations of the other ligands in tmn are known. In the cpa and dfr cases, and in seven of the eight tmn ligands, the GA docking method found conformations within 1.6 Å root mean square (rms) of the relaxed crystal conformation. © 1995 John Wiley & Sons, Inc.
This book sets out to explain what genetic algorithms are and how they can be used to solve real-world problems. The first objective is tackled by the editor, Lawrence Davis. The remainder of the book is turned over to a series of short review articles by a collection of authors, each explaining how genetic algorithms have been applied to problems in their own specific area of interest. The first part of the book introduces the fundamental genetic algorithm (GA), explains how it has traditionally been designed and implemented and shows how the basic technique may be applied to a very simple numerical optimisation problem. The basic technique is then altered and refined in a number of ways, with the effects of each change being measured by comparison against the performance of the original. In this way, the reader is provided with an uncluttered introduction to the technique and learns to appreciate why certain variants of GA have become more popular than others in the scientific community. Davis stresses that the choice of a suitable representation for the problem in hand is a key step in applying the GA, as is the selection of suitable techniques for generating new solutions from old. He is refreshingly open in admitting that much of the business of adapting the GA to specific problems owes more to art than to science. It is nice to see the terminology associated with this subject explained, with the author stressing that much of the field is still an active area of research. Few assumptions are made about the reader's mathematical background. The second part of the book contains thirteen cameo descriptions of how genetic algorithmic techniques have been, or are being, applied to a diverse range of problems. Thus, one group of authors explains how the technique has been used for modelling arms races between neighbouring countries (a non- linear, dynamical system), while another group describes its use in deciding design trade-offs for military aircraft. My own favourite is a rather charming account of how the GA was applied to a series of scheduling problems. Having attempted something of this sort with Simulated Annealing, I found it refreshing to see the authors highlighting some of the problems that they had encountered, rather than sweeping them under the carpet as is so often done in the scientific literature. The editor points out that there are standard GA tools available for either play or serious development work. Two of these (GENESIS and OOGA) are described in a short, third part of the book. As is so often the case nowadays, it is possible to obtain a diskette containing both systems by sending your Visa card details (or $60) to an address in the USA.
The Protein Data Bank is a computer-based archival file for macromolecular structures. The Bank stores in a uniform format atomic co-ordinates and partial bond connectivities, as derived from crystallographic studies. Text included in each data entry gives pertinent information for the structure at hand (e.g. species from which the molecule has been obtained, resolution of diffraction data, literature citations and specifications of secondary structure). In addition to atomic co-ordinates and connectivities, the Protein Data Bank stores structure factors and phases, although these latter data are not placed in any uniform format. Input of data to the Bank and general maintenance functions are carried out at Brookhaven National Laboratory. All data stored in the Bank are available on magnetic tape for public distribution, from Brookhaven (to laboratories in the Americas), Tokyo (Japan), and Cambridge (Europe and worldwide). A master file is maintained at Brookhaven and duplicate copies are stored in Cambridge and Tokyo. In the future, it is hoped to expand the scope of the Protein Data Bank to make available co-ordinates for standard structural types (e.g. alpha-helix, RNA double-stranded helix) and representative computer programs of utility in the study and interpretation of macromolecular structures.
The principal protein excreted in male rat urine, urinary alpha 2-globulin and the homologous mouse protein, major urinary protein, have been well characterized, although their functions remain unclear. Male rat urine affects the behaviour and sexual response of female rats, leading to the proposal that rodent urinary proteins are responsible for binding pheromones and their subsequent release from drying urine. Urinary alpha 2-globulin is also involved in hyaline droplet nephropathy, an important toxicological syndrome in male rats resulting from exposure to a number of industrial chemicals and characterized by the accumulation of liganded urinary alpha 2-globulin in lysosomes in the kidney, followed by the induction of renal cancer. We now report the three-dimensional structures of mouse major urinary protein (at 2.4 A resolution) and rat urinary alpha 2-globulin (at 2.8 A resolution). The results corroborate the role of these proteins in pheromone transport and elaborate the structural basis of ligand binding.
A method for solid-filling protein cavities is presented. The method uses a pattern-recognition technique based on cellular logic operations to distinguish between convex and concave regions of a protein. In doing this it solid fills protein cavities and automatically defines a boundary between cavity and exterior free space. The operations used to fill the cavities also can be used to process the filler to filter out small-scale features. So far the main use of the method has been in visualizing protein active sites for docking. The method can be used to find cavities of a given size range and could be used to find novel protein binding sites.
The crystal structures of endothiapepsin, a fungal aspartic proteinase (EC 3.4.23.6), cocrystallized with two oligopeptide renin inhibitors, PD125967 and PD125754, have been determined at 2.0-A resolution and refined to R-factors of 0.143 and 0.153, respectively. These inhibitors, which are of the hydroxyethylene and statine types, respectively, possess a cyclohexylalanine side chain at P1 and have interesting functionalities at the P3 position which, until now, have not been subjected to crystallographic analysis. PD125967 has a bis(1-naphthylmethyl)acetyl residue at P3, and PD125754 possesses a hydroxyethylene analogue of the P3-P2 peptide bond for proteolytic stability. The structures reveal that the S3 pocket accommodates one naphthyl ring with conformational changes of the Asp 77 and Asp 114 side chains, the other naphthyl group residing in the S4 region. The P3-P2 hydroxyethylene analogue of PD125754 forms a hydrogen bond with the NH of Thr 219, thereby making the same interaction with the enzyme as the equivalent peptide groups of all inhibitors studied so far. The absence of side chains at the P2 and P1' positions of this inhibitor allows water molecules to occupy the respective pockets in the complex. The relative potencies of PD125967 and PD125754 for endothiapepsin are consistent with the changes in solvent-accessible area which take place on inhibitor binding.
Different hexameric forms of insulin have been crystallized from a variety of conditions. In the presence of 1% phenol, 1.0 M sodium chloride, and at a pH of 8.5, a rhombohedral form is produced with two monomers in the asymmetric unit, space group R3, a = 79.92 A and c = 40.39 A. The structure has been solved and refined, using data between 8.0 and 2.5 A resolution, to a residual of 0.157. Each of the monomers adopts an R conformation, that is residues B1-B8 are alpha-helical. As a result of the T to R transition, an elliptical cavity is created between symmetry-related monomers and is occupied by a phenol molecule. A region of density within bonding distance to one of the zinc ions has been interpreted as an additional phenol molecule.
Analogues of peptides ranging in size from three to six amino acids and containing the hydroxyethylene dipeptide isosteres Phe psi Gly, Phe psi Ala, Phe psi NorVal, Phe psi Leu, and Phe psi Phe, where psi denotes replacement of CONH by (S)-CH(OH)CH2, were synthesized and studied as HIV-1 protease inhibitors. Inhibition constants (Ki) with purified HIV-1 protease depend strongly on the isostere in the order Phe psi Gly greater than Phe psi Ala greater than Phe psi NorVal greater than Phe psi Leu greater than Phe psi Phe and decrease with increasing length of the peptide analogue, converging to a value of 0.4 nM. Ki values are progressively less dependent on inhibitor length as the size of the P1' side chain within the isostere increases. The structures of HIV-1 protease complexed with the inhibitors Ala-Ala-X-Val-Val-OMe, where X is Phe psi Gly, Phe psi Ala, Phe psi NorVal, and Phe psi Phe, have been determined by X-ray crystallography (resolution 2.3-3.2 A). The crystals exhibit symmetry consistent with space group P6(1) with strong noncrystallographic 2-fold symmetry, and the inhibitors all exhibit 2-fold disorder. The inhibitors bind in similar conformations, forming conserved hydrogen bonds with the enzyme. The Phe psi Gly inhibitor adopts an altered conformation that places its P3' valine side chain partially in the hydrophobic S1' pocket, thus suggesting an explanation for the greater dependence of the Ki value on inhibitor length in the Phe psi Gly series. From the kinetic and crystallographic data, a minimal inhibitor model for tight-binding inhibition is derived in which the enzyme subsites S2-S2' are optimally occupied. The Ki values for several compounds are compared with their potencies as inhibitors of proteolytic processing in T-cell cultures chronically infected with HIV-1 (MIC values) and as inhibitors of acute infectivity (IC50 values). There is a rank-order correspondence, but a 20-1000-fold difference, between the values of Ki and those of MIC or IC50. IC50 values can approach those of Ki but are highly dependent on the conditions of the acute infectivity assay and are influenced by physiochemical properties of the inhibitors such as solubility.
Metal ions serve a variety of functions in proteins. The most important function is to enhance the structural stability of the protein in the conformation required for biological function and to take part in the catalytic processes of enzymes. Metal ions can take part in trigger and control mechanisms by specifically altering or stabilizing a macromolecular conformation on binding. This chapter discusses the properties of metals that are useful in the structure and function of proteins, particularly enzymes, and the geometry of interaction of metals with the various chemical groups of proteins are emphasized. Ligands donate an electron pair to the bond and are generally negatively charged or neutral. Important in a study of metal-ligand interactions are the polarizability of both the metal ion and the ligand, the number of the ligands around each metal ion, and the stereochemistry of the resulting arrangement. The stereochemistry of liganding of metal ions in proteins is known for several proteins. Some selected examples follow with data derived from the Protein Data Bank are emphasized in the chapter. The metals in protein crystal structures discussed are— namely, (1) copper, (2) iron, (3) manganese, (4) zinc, (5) magnesium, and (6) calcium. In the cases in which two different metals are bound, information can be obtained on preferential sites for each metal in the presence of the other.
The crystal structure of the Fab fragment of the murine monoclonal anti-dinitrophenyl-spin-label antibody AN02 complexed with its hapten has been solved at 2.9 A resolution using a novel molecular replacement method. Prior to translation searches, a large number of the most likely rotation function solutions were subjected to a rigid body refinement against the linear correlation coefficient between intensities of observed and calculated structure factors. First, the overall orientation of the search model and then the orientations and positions of the four Fab domains (VH, VL, CH1 and CL) were refined. This procedure clearly identified the correct orientation of the search model. The refined search model was then subjected to translation searches which unambiguously determined the enantiomer and position in the unit cell of the crystal. The successful search model was refined 2.5 A crystal structure of the Fab fragment of HyHel-5 from which non-matching residues in the variable domains had been removed. HyHel-5 is a murine monoclonal antibody whose heavy and light chains are of the same subclass (gamma 1, kappa, respectively) as AN02. After molecular replacement the structure of the AN02 Fab has been refined using simulated annealing in combination with model building and conjugate gradient refinement to a current crystallographic R-factor of 19.5% for 12,129 unique reflections between 8.0 and 2.9 A. The root-mean-square (r.m.s.) deviation from ideal bond lengths is 0.014 A, and the r.m.s. deviation from ideal bond angles is 3.1 degrees. The electron density reveals the hapten sitting in a pocket formed by the loops of the complementarity determining region. The dinitrophenyl ring of the hapten is sandwiched between the indole rings of Trp96 of the heavy-chain and Trp91 of the light-chain. The positioning of the hapten and general features of the combining site are in good agreement with the results of earlier nuclear magnetic resonance experiments.
The crystal structure of the complex between adenylate kinase from bovine mitochondrial matrix and its substrate AMP has been refined at 1.85 A resolution (1 A = 0.1 nm). Based on 42,519 independent reflections of better than 10 A resolution, a final R-factor of 18.9% was obtained with a model obeying standard geometry within 0.016 A in bond lengths and 3.2 degrees in bond angles. There are two enzyme: substrate complexes in the asymmetric unit, each consisting of 226 amino acid residues, one AMP and one sulfate ion. A superposition of the two full-length polypeptides revealed deviations that can be described as small relative movements of three domains. Best superpositions of individual domains yielded a residual overall root-mean-square deviation of 0.3 A for the backbone atoms and 0.5 A for the sidechains. The final model contains 381 solvent molecules in the asymmetric unit, 2 x 72 = 144 of which occupy corresponding positions in both complexes.
The crystal structure of unliganded dihydrofolate reductase (DHFR) from Escherichia coli has been solved and refined to an R factor of 19% at 2.3-A resolution in a crystal form that is nonisomorphous with each of the previously reported E. coli DHFR crystal structures [Bolin, J. T., Filman, D. J., Matthews, D. A., Hamlin, B. C., & Kraut, J. (1982) J. Biol. Chem. 257, 13650-13662; Bystroff, C., Oatley, S. J., & Kraut, J. (1990) Biochemistry 29, 3263-3277]. Significant conformational changes occur between the apoenzyme and each of the complexes: the NADP+ holoenzyme, the folate-NADP+ ternary complex, and the methotrexate (MTX) binary complex. The changes are small, with the largest about 3 A and most of them less than 1 A. For simplicity a two-domain description is adopted in which one domain contains the NADP+ 2'-phosphate binding site and the binding sites for the rest of the coenzyme and for the substrate lie between the two domains. Binding of either NADP+ or MTX induces a closing of the PABG-binding cleft and realignment of alpha-helices C and F which bind the pyrophosphate of the coenzyme. Formation of the ternary complex from the holoenzyme does not involve further relative domain shifts but does involve a shift of alpha-helix B and a floppy loop (the Met-20 loop) that precedes alpha B. These observations suggest a mechanism for cooperativity in binding between substrate and coenzyme wherein the greatest degree of cooperativity is expressed in the transition-state complex. We explore the idea that the MTX binary complex in some ways resembles the transition-state complex.
High level bacterial resistance to chloramphenicol is generally due to O-acetylation of the antibiotic in a reaction catalysed by chloramphenicol acetyltransferase (CAT, EC 2.3.1.28) in which acetyl-coenzyme A is the acyl donor. The crystal structure of the type III enzyme from Escherichia coli with chloramphenicol bound has been determined and refined at 1.75 A resolution, using a restrained parameter reciprocal space least squares procedure. The refined model, which includes chloramphenicol, 204 solvent molecules and two cobalt ions has a crystallographic R-factor of 18.3% for 27,300 reflections between 6 and 1.75 A resolution. The root-mean-square deviation in bond lengths from ideal values is 0.02 A. The cobalt ions play a crucial role in stabilizing the packing of the molecule in the crystal lattice. CAT is a trimer of identical subunits (monomer Mr 25,000) and the trimeric structure is stabilized by a number of hydrogen bonds, some of which result in the extension of a beta-sheet across the subunit interface. Chloramphenicol binds in a deep pocket located at the boundary between adjacent subunits of the trimer, such that the majority of residues forming the binding pocket belong to one subunit while the catalytically essential histidine belongs to the adjacent subunit. His195 is appropriately positioned to act as a general base catalyst in the reaction, and the required tautomeric stabilization is provided by an unusual interaction with a main-chain carbonyl oxygen.
The Metropolis technique of conformation searching is combined with rapid energy evaluation using molecular affinity potentials to give an efficient procedure for docking substrates to macromolecules of known structure. The procedure works well on a number of crystallographic test systems, functionally reproducing the observed binding modes of several substrates.
The serine protease gamma-chymotrypsin was covalently inhibited with two different photoreversible cinnamate compounds, and the structures of the resulting complexes were determined to 1.9-A resolution. The inhibitors show different kinetics of binding, inhibition, and nonphotochemical deacylation relative to each other in solution activity assays. The crystal structures of the enzyme-cinnamate complexes show that both compounds acylate serine 195 and that the two molecules are bound in similar nonproductive conformations which have drastic effects on their ability to turn over. Substitution of a diethylamino group on the para position of the cinnamate ring causes a 1000-fold increase in the thermal stability of the inhibitor toward hydrolysis and deacylation.
The crystal structure of a fluorescein-Fab (4-4-20) complex was determined at 2.7 A resolution by molecular replacement methods. The starting model was the refined 2.7 A structure of unliganded Fab from an autoantibody (BV04-01) with specificity for single-stranded DNA. In the 4-4-20 complex fluorescein fits tightly into a relatively deep slot formed by a network of tryptophan and tyrosine side chains. The planar xanthonyl ring of the hapten is accommodated at the bottom of the slot while the phenylcarboxyl group interfaces with solvent. Tyrosine 37 (light chain) and tryptophan 33 (heavy chain) flank the xanthonyl group and tryptophan 101 (light chain) provides the floor of the combining site. Tyrosine 103 (heavy chain) is situated near the phenyl ring of the hapten and tyrosine 102 (heavy chain) forms part of the boundary of the slot. Histidine 31 and arginine 39 of the light chain are located in positions adjacent to the two enolic groups at opposite ends of the xanthonyl ring, and thus account for neutralization of one of two negative charges in the haptenic dianion. Formation of an enol-arginine ion pair in a region of low dielectric constant may account for an incremental increase in affinity of 2-3 orders of magnitude in the 4-4-20 molecule relative to other members of an idiotypic family of monoclonal antifluorescyl antibodies. The phenyl carboxyl group of fluorescein appears to be hydrogen bonded to the phenolic hydroxyl group of tyrosine 37 of the light chain. A molecule of 2-methyl-2,4-pentanediol (MPD), trapped in the interface of the variable domains just below the fluorescein binding site, may be partly responsible for the decrease in affinity for the hapten in MPD.