Structure-based tailoring of compound libraries for high-throughput screening: discovery of novel EphB4 kinase inhibitors

High-throughput docking is a computational tool frequently used to discover small-molecule inhibitors of enzymes or receptors of known three-dimensional structure. Because of the large number of molecules in chemical libraries, automatic procedures to prune multimillion compound collections are useful for high-throughput docking and necessary for in vitro screening. Here, we propose an anchor-based library tailoring approach (termed ALTA) to focus a chemical library by docking and prioritizing molecular fragments according to their binding energy which includes continuum electrostatics solvation. In principle, ALTA does not require prior knowledge of known inhibitors, but receptor-based pharmacophore information (hydrogen bonds with the hinge region) is additionally used here to identify molecules with optimal anchor fragments for the ATP-binding site of the EphB4 receptor tyrosine kinase. The 21,418 molecules of the focused library (from an initial collection of about 730,000) are docked into EphB4 and ranked by force-field-based energy including electrostatic solvation. Among the 43 compounds tested in vitro, eight molecules originating from two different anchors show low-micromolar activity in a fluorescence-based enzymatic assay. Four of them are active in a cell-based assay and are potential anti-angiogenic compounds.     

Ligand-induced conformational changes: improved predictions of ligand binding conformations and affinities

A computational docking strategy using multiple conformations of the target protein is discussed and evaluated. A series of low molecular weight, competitive, nonpeptide protein tyrosine phosphatase inhibitors are considered for which the x-ray crystallographic structures in complex with protein tyrosine phosphatase 1B (PTP1B) are known. To obtain a quantitative measure of the impact of conformational changes induced by the inhibitors, these were docked to the active site region of various structures of PTP1B using the docking program FlexX. Firstly, the inhibitors were docked to a PTP1B crystal structure cocrystallized with a hexapeptide. The estimated binding energies for various docking modes as well as the RMS differences between the docked compounds and the crystallographic structure were calculated. In this scenario the estimated binding energies were not predictive inasmuch as docking modes with low estimated binding energies corresponded to relatively large RMS differences when aligned with the corresponding crystal structure. Secondly, the inhibitors were docked to their parent protein structures in which they were cocrystallized. In this case, there was a good correlation between low predicted binding energy and a correct docking mode. Thirdly, to improve the predictability of the docking procedure in the general case, where only a single target protein structure is known, we evaluate an approach which takes possible protein side-chain conformational changes into account. Here, side chains exposed to the active site were considered in their allowed rotamer conformations and protein models containing all possible combinations of side-chain rotamers were generated. To evaluate which of these modeled active sites is the most likely binding site conformation for a certain inhibitor, the inhibitors were docked against all active site models. The receptor rotamer model corresponding to the lowest estimated binding energy is taken as the top candidate. Using this protocol, correct inhibitor binding modes could successfully be discriminated from proposed incorrect binding modes. Moreover, the ranking of the estimated ligand binding energies was in good agreement with experimentally observed binding affinities.     

Applications of the ArgusLab4/AScore protocol in the structure-based binding affinity prediction of various inhibitors of group-1 and group-2 influenza virus neuraminidases (NAs)

Using the crystal structures of inhibitors bound to either group-2 or group-1 neuraminidases (NAs), AScore/ShapeDock (GaDock) scoring was shown to identify the binding modes in agreement with the experiment for all inhibitors docked in their own NA/inhibitor crystal structures. To investigate the effect of small changes in protein structure on predicted binding modes, in a set of 132 docking experiments (11 inhibitors docked in 12 group-2 NA structures), AScore/ShapeDock (GaDock) identified the correct binding modes of 116 complexes. In a total of 88 docking experiments (8 inhibitors docked in 11 group-1 NA structures), AScore/ShapeDock predicted 80 binding modes correctly. Flexible AScore/ShapeDock docking, as quite reproducible, is suggested to be convenient for designing novel H5N1 inhibitors.     

Discovery of Fragment Molecules That Bind the Human Peroxiredoxin 5 Active Site

The search for protein ligands is a crucial step in the inhibitor design process. Fragment screening represents an interesting method to rapidly find lead molecules, as it enables the exploration of a larger portion of the chemical space with a smaller number of compounds as compared to screening based on drug-sized molecules. Moreover, fragment screening usually leads to hit molecules that form few but optimal interactions with the target, thus displaying high ligand efficiencies. Here we report the screening of a homemade library composed of 200 highly diverse fragments against the human Peroxiredoxin 5 protein. Peroxiredoxins compose a family of peroxidases that share the ability to reduce peroxides through a conserved cysteine. The three-dimensional structures of these enzymes ubiquitously found throughout evolution have been extensively studied, however, their biological functions are still not well understood and to date few inhibitors have been discovered against these enzymes. Six fragments from the library were shown to bind to the Peroxiredoxin 5 active site and ligand-induced chemical shift changes were used to drive the docking of these small molecules into the protein structure. The orientation of the fragments in the binding pocket was confirmed by the study of fragment homologues, highlighting the role of hydroxyl functions that hang the ligands to the Peroxiredoxin 5 protein. Among the hit fragments, the small catechol molecule was shown to significantly inhibit Peroxiredoxin 5 activity in a thioredoxin peroxidase assay. This study reports novel data about the ligand-Peroxiredoxin interactions that will help considerably the development of potential Peroxiredoxin inhibitors.     

Prediction of ligand-binding sites of proteins by molecular docking calculation for a random ligand library

A new approach to predicting the ligand-binding sites of proteins was developed, using protein-ligand docking computation. In this method, many compounds in a random library are docked onto the whole protein surface. We assumed that the true ligand-binding site would exhibit stronger affinity to the compounds in the random library than the other sites, even if the random library did not include the ligand corresponding to the true binding site. We also assumed that the affinity of the true ligand-binding site would be correlated to the docking scores of the compounds in the random library, if the ligand-binding site was correctly predicted. We call this method the molecular-docking binding-site finding (MolSite) method. The MolSite method was applied to 89 known protein-ligand complex structures extracted from the Protein Data Bank, and it predicted the correct binding sites with about 80-99% accuracy, when only the single top-ranked site was adopted. In addition, the average docking score was weakly correlated to the experimental protein-ligand binding free energy, with a correlation coefficient of 0.44.     

Molecular determinants of sarco/endoplasmic reticulum calcium ATPase inhibition by hydroquinone-based compounds

The ion transport activity of the sarco/endoplasmic reticulum calcium ATPase (SERCA) is specifically and potently inhibited by the small molecule 2,5-di-tert-butylhydroquinone (BHQ). In this study, we investigated the relative importance of the nature and position of BHQ's four substituents for enzyme inhibition by employing a combination of experimental and computational techniques. The inhibitory potencies of 21 commercially available or synthesized BHQ derivatives were determined in ATPase activity assays, and 11 compounds were found to be active. Maximum inhibitory potency was observed in compounds with two para hydroxyl groups, whereas BHQ analogues with only one hydroxyl group were still active, albeit with a reduced potency. The results also demonstrated that two alkyl groups were an absolute requirement for activity, with the most potent compounds having 2,5-substituents with four or five carbon atoms at each position. Using the program GOLD in conjunction with the ChemScore scoring function, the structures of the BHQ analogues were docked into the crystal structure of SERCA mimicking the enzyme's E(2) conformation. Analysis of the docking results indicated that inhibitor binding to SERCA was primarily mediated by a hydrogen bond between a hydroxyl group and Asp-59 and by hydrophobic interactions involving the bulky inhibitor alkyl groups. Attempts to dock BHQ into crystal structures corresponding to the E(1) conformation of the enzyme failed, because the conformational changes accompanying the E(2)/E(1) transition severely restricted the size of the binding site, suggesting that BHQ stabilizes the enzyme in its E(2) form. The potential role of Glu309 in enzyme inhibition is discussed in the context of the computational results. The docking scores correlated reasonably well with the measured inhibitory potencies and allowed the distinction between active and inactive compounds, which is a key requirement for future virtual screening of large compound databases for novel SERCA inhibitors.     

Evaluation of docking programs for predicting binding of Golgi alpha-mannosidase II inhibitors: a comparison with crystallography

Golgi alpha-mannosidase II (GMII), a zinc-dependent glycosyl hydrolase, is a promising target for drug development in anti-tumor therapies. Using X-ray crystallography, we have determined the structure of Drosophila melanogaster GMII (dGMII) complexed with three different inhibitors exhibiting IC50's ranging from 80 to 1000 microM. These structures, along with those of seven other available dGMII/inhibitor complexes, were then used as a basis for the evaluation of seven docking programs (GOLD, Glide, FlexX, AutoDock, eHiTS, LigandFit, and FITTED). We found that small inhibitors could be accurately docked by most of the software, while docking of larger compounds (i.e., those with extended aromatic cycles or long aliphatic chains) was more problematic. Overall, Glide provided the best docking results, with the most accurately predicted binding around the active site zinc atom. Further evaluation of Glide's performance revealed its ability to extract active compounds from a benchmark library of decoys.     

Accounting for ligand-bound metal ions in docking small molecules on adenylyl cyclase toxins

The adenylyl cyclase toxins produced by bacteria (such as the edema factor (EF) of Bacillus anthracis and CyaA of Bordetella pertussis) are important virulence factors in anthrax and whooping cough. Co-crystal structures of these proteins differ in the number and positioning of metal ions in the active site. Metal ions bound only to the ligands in the crystal structures are not included during the docking. To determine what effect these "missing" metals have on docking results, the AutoDock, LigandFit/Cerius2, and FlexX programs were compared for their ability to correctly place substrate analogues and inhibitors into the active sites of the crystal structures of EF, CyaA, and mammalian adenylate cyclase. Protonating the phosphates of substrate analogues improved the accuracy of docking into the active site of CyaA, where the grid did not account for one of the three Mg2+ ions in the crystal structure. The AutoDock ranking (based on docking energies) of a test group of compounds was relatively unaffected by protonation of carboxyl groups. However, the ranking by FlexX-ChemScore varied significantly, especially for docking to CyaA, suggesting that alternate protonation states should be tested when screening compound libraries with this program. When the charges on the bound metal were set correctly, AutoDock was the most reliable program of the three tested with respect to positioning substrate analogues and ranking compounds according to their experimentally determined ability to inhibit EF.     

In silico development of new acetylcholinesterase inhibitors

In this work, we made use of fragment-based drug design (FBDD) and de novo design to obtain more powerful acetylcholinesterase (AChE) inhibitors. AChE is associated with Alzheimer’s disease (AD). It was found that the cholinergic pathways in the cerebral cortex are compromised in AD and the accompanying cholinergic deficiency contributes to the cognitive deterioration of AD patients. In the FBDD approach, fragments are docked into the active site of the protein. As fragments are molecular groups with a low number of atoms, it is possible to study their interaction with localized amino acids. Once the interactions are measured, the fragments are organized by affinity and then linked together to form new molecules with a high degree of interaction with the active site. In the other approach, we used the de novo design technique starting from reference drugs used in the AD treatment. These drugs were broken into fragments (seeds). In the growing strategy, fragments were added to each seed, growing new molecules. In the linking strategy, two or more separated seeds were linked with different fragments. Both strategies combined produced a library of more than 2 million compounds. This library was filtered using absorption, distribution, metabolism, and excretion properties. The resulting library with around six thousand compounds was filtered again. In this case, structures with Tanimoto coefficients >.85 were discarded. The final library with 1500 compounds was submitted to docking studies. As a result, 10 compounds with better interaction energy than the reference drugs were obtained.     

Towards in silico lead optimization: scores from ensembles of protein/ligand conformations reliably correlate with biological activity

Accurately ranking protein/ligand interactions and distinguishing subtle differences between homologous compounds in a virtual focused library in silico is essential in a structure-based drug discovery program. In order to establish a predictive model to design novel inhibitors of dihydrofolate reductase (DHFR) from the parasitic protozoa, Cryptosporidium hominis, we docked a series of 30 DHFR inhibitors with measured inhibition constants against the crystal structure of the protein. By including protein flexibility and averaging the energies of the 25 lowest protein/ligand conformers we obtained more accurate total nonbonded energies from which we calculated a predicted biological activity. The calculated and measured biological activities showed reliable correlations of 72.9%. Additionally, visual analysis of the ensemble of protein/ligand conformations revealed alternative ligand binding pockets in the active site. Using the same principles we then created a homology model of DHFR from Toxoplasma gondii and docked 11 inhibitors. A correlation of 50.2% between docking score and activity validates both the method and the model. The correlations presented here are particularly compelling considering the high structural similarity of the ligands and the fact that we have used structures derived from crystallographic data and homology modeling. These docking principles may be useful in any lead optimization study where accurate ranking of similar compounds is desired.     

Discovery of new MurA inhibitors using induced-fit simulation and docking

We report on the successful application of ProBiS-CHARMMing web server in the discovery of new inhibitors of MurA, an enzyme that catalyzes the first committed cytoplasmic step of bacterial peptidoglycan synthesis. The available crystal structures of Escherichia coli MurA in the Protein Data Bank have binding sites whose small volume does not permit the docking of drug-like molecules. To prepare the binding site for docking, the ProBiS-CHARMMing web server was used to simulate the induced-fit effect upon ligand binding to MurA, resulting in a larger, more holo-like binding site. The docking of a filtered ZINC compound library to this enlarged binding site was then performed and resulted in three compounds with promising inhibitory potencies against MurA. Compound 1 displayed significant inhibitory potency with IC₅₀ value of 1 μM. All three compounds have novel chemical structures, which could be used for further optimization of small-molecule MurA inhibitors.     

Pre-docking filter for protein and ligand 3D structures

Virtual drug screening using protein-ligand docking techniques is a time-consuming process, which requires high computational power for binding affinity calculation. There are millions of chemical compounds available for docking. Eliminating compounds that are unlikely to exhibit high binding affinity from the screening set should speed-up the virtual drug screening procedure. We performed docking of 6353 ligands against twenty-one protein X-ray crystal structures. The docked ligands were ranked according to their calculated binding affinities, from which the top five hundred and the bottom five hundred were selected. We found that the volume and number of rotatable bonds of the top five hundred docked ligands are similar to those found in the crystal structures and corresponded with the volume of the binding sites. In contrast, the bottom five hundred set contains ligands that are either too large to enter the binding site, or too small to bind with high specificity and affinity to the binding site. A pre-docking filter that takes into account shapes and volumes of the binding sites as well as ligand volumes and flexibilities can filter out low binding affinity ligands from the screening sets. Thus, the virtual drug screening procedure speed is increased.     

Flexible ligand docking using conformational ensembles.

Molecular docking algorithms suggest possible structures for molecular complexes. They are used to model biological function and to discover potential ligands. A present challenge for docking algorithms is the treatment of molecular flexibility. Here, the rigid body program, DOCK, is modified to allow it to rapidly fit multiple conformations of ligands. Conformations of a given molecule are pre-calculated in the same frame of reference, so that each conformer shares a common rigid fragment with all other conformations. The ligand conformers are then docked together, as an ensemble, into a receptor binding site. This takes advantage of the redundancy present in differing conformers of the same molecule. The algorithm was tested using three organic ligand protein systems and two protein-protein systems. Both the bound and unbound conformations of the receptors were used. The ligand ensemble method found conformations that resembled those determined in X-ray crystal structures (RMS values typically less than 1.5 A). To test the method's usefulness for inhibitor discovery, multi-compound and multi-conformer databases were screened for compounds known to bind to dihydrofolate reductase and compounds known to bind to thymidylate synthase. In both cases, known inhibitors and substrates were identified in conformations resembling those observed experimentally. The ligand ensemble method was 100-fold faster than docking a single conformation at a time and was able to screen a database of over 34 million conformations from 117,000 molecules in one to four CPU days on a workstation.     

Receptor based 3D-QSAR to identify putative binders of Mycobacterium tuberculosis Enoyl acyl carrier protein reductase

In the current study, the applicability and scope of 3D-QSAR models (CoMFA and CoMSIA) to complement virtual screening using 3D pharmacophore and molecular docking is examined and applied to identify potential hits against Mycobacterium tuberculosis Enoyl acyl carrier protein reductase (MtENR). Initially CoMFA and CoMSIA models were developed using series of structurally related arylamides as MtENR inhibitors. Docking studies were employed to position the inhibitors into MtENR active site to derive receptor based 3D-QSAR models. Both CoMFA and CoMSIA yielded significant cross validated q² values of 0.663 and 0.639 and r² values of 0.989 and 0.963, respectively. The statistically significant models were validated by a test set of eight compounds with predictive r² value of 0.882 and 0.875 for CoMFA and CoMSIA. The contour maps from 3D-QSAR models in combination with docked binding structures help to better interpret the structure activity relationship. Integrated with CoMFA and CoMSIA predictive models structure based (3D-pharmacophore and molecular docking) virtual screening have been employed to explore potential hits against MtENR. A representative set of 20 compounds with high predicted IC₅₀ values were sorted out in the present study.     

3D Model of the Acetylcholinesterase Catalytic Cavity Probed by Stereospecific Organophosphorous Inhibitors

Automated docking was performed for stereospecific and quasi-irreversible organophosphorous acetylcholinesterase (AChE) inhibitors. Twelve chiral inhibitor structures, corresponding to six enantiomeric pairs, each with a phosphorus atom as a stereocentre, were docked to the crystal structure of mouse AChE. This study gives evidence that in inhibitors with different aromatic and cationic leaving groups these groups are oriented towards the entry of the active site, as recently suggested by Hosea et al[1] for inhibitors with a thiocholine leaving group. The results of the docking were used to establish a three dimensional model of the volume sterically available to the inhibitors within the AChE active site.     

New leads of metallo-beta-lactamase inhibitors from structure-based pharmacophore design

We have applied pharmacophore generation, database searching, docking methodologies, and experimental enzyme kinetics to discover new structures for design of di-zinc metallo-beta-lactamase inhibitors. Based on crystal structures of class B1 metallo-beta-lactamases with a succinic acid and a mercapto-carboxylic acid inhibitor bound to the enzyme, two pharmacophore models were constructed. With the Catalyst program, these pharmacophores were used to search the ACD database, which provided a total of 74 hits representing four different chemical classes of compounds: Dicarboxylic acids, phosphonic and sulfonic acid derivatives, and mercapto-carboxylic acids. All hits were docked into different metallo-beta-lactamases (from classes B1 and B3) using the GOLD docking program. A selection scheme based on the GOLD scores, the Catalyst fit and shape values, and the size of the compounds (molecular weight, surface area, and number of rotatable bonds) was developed and thirteen compounds representing all four chemical classes were selected for experimental studies. Three compounds with new scaffolds hitherto not present in metallo-beta-lactamase inhibitors have IC50 values less than 15 microM and may serve as starting points in the design of metallo-beta-lactamase inhibitors.     

High-Throughput Screening for Novel Inhibitors of Neisseria gonorrhoeae Penicillin-Binding Protein 2

The increasing prevalence of N. gonorrhoeae strains exhibiting decreased susceptibility to third-generation cephalosporins and the recent isolation of two distinct strains with high-level resistance to cefixime or ceftriaxone heralds the possible demise of β-lactam antibiotics as effective treatments for gonorrhea. To identify new compounds that inhibit penicillin-binding proteins (PBPs), which are proven targets for β-lactam antibiotics, we developed a high-throughput assay that uses fluorescence polarization (FP) to distinguish the fluorescent penicillin, Bocillin-FL, in free or PBP-bound form. This assay was used to screen a 50,000 compound library for potential inhibitors of N. gonorrhoeae PBP 2, and 32 compounds were identified that exhibited >50% inhibition of Bocillin-FL binding to PBP 2. These included a cephalosporin that provided validation of the assay. After elimination of compounds that failed to exhibit concentration-dependent inhibition, the antimicrobial activity of the remaining 24 was tested. Of these, 7 showed antimicrobial activity against susceptible and penicillin- or cephalosporin-resistant strains of N. gonorrhoeae. In molecular docking simulations using the crystal structure of PBP 2, two of these inhibitors docked into the active site of the enzyme and each mediate interactions with the active site serine nucleophile. This study demonstrates the validity of a FP-based assay to find novel inhibitors of PBPs and paves the way for more comprehensive high-throughput screening against highly resistant strains of N. gonorrhoeae. It also provides a set of lead compounds for optimization of anti-gonococcal agents.     

Identification of selective MMP-9 inhibitors through multiple e-pharmacophore,ligand-based pharmacophore,molecular docking,and density functional theory approaches

Matrix metalloproteinase-9 (MMP-9) is a significant target for the development of drugs for the treatment of arthritis, CNS disorders, and cancer metastasis. The structure-based and ligand-based methods were used for the virtual screening (VS) of database compounds to obtain potent and selective MMP-9 inhibitors. Experimentally known MMP-9 inhibitors were used to grow up ligand-based three pharmacophore models utilizing Schrodinger suite. The X-ray crystallographic structures of MMP-9 with different inhibitors were used to develop five energy-optimized structure-based (e-pharmacophore) models. All developed pharmacophores were validated and applied to screen the Zinc database. Pharmacophore matched compounds were subjected to molecular docking to retrieve hits with novel scaffolds. The molecules with diverse structures, high docking scores and low binding energies for various crystal structures of MMP-9, were selected as final hits. The Induced fit docking (IFD) analysis provided significant information about the driving of inhibitor to approve a suitable bioactive conformational position in the active site of protein. Since charge transfer reaction occurs during receptor–ligand interaction, therefore, electronic features of hits (ligands) are interesting parameters to explain the binding interactions. Density functional theory (DFT) at B3LYP/6-31G* level was utilized to explore electronic features of hits. The docking study of hits using AutoDock was helpful to establish the binding interactions. The study illustrates that the combined pharmacophore approach is advantageous to identify diverse hits which have better binding affinity to the active site of the enzyme for all possible bioactive conformations. The approach used in the study is worthy to design drugs for other targets.     

Design,synthesis, biological evaluation and computational investigation of novel inhibitors of dihydrofolate reductase of opportunistic pathogens

The present work deals with design, synthesis and biological evaluation of novel, diverse compounds as potential inhibitors of dihydrofolate reductase (DHFR) from opportunistic microorganisms; Pneumocystis carinii (pc), Toxoplasma gondii (tg) and Mycobacterium avium (ma). A set of 14 structurally diverse compounds were designed with varying key pharmacophoric features of DHFR inhibitors, bulky distal substitutions and different bridges joining the distal part and 2,4-diaminopyrimidine nucleus. The designed compounds were synthesized and evaluated in enzyme assay against pc, tg and ma DHFR. The rat liver (rl) DHFR was used as mammalian standard. As the next logical step of the project, flexible molecular docking studies were carried out to predict the binding modes of these compounds in pcDHFR active site and the obtained docked poses were post processed using MM-GBSA protocol for prediction of relative binding affinity. The predicted binding modes were able to rationalize the experimental results in most cases. Of particular interest, both the docking scores and MM-GBSA predicted ΔG_bind were able to distinguish between the active and low active compounds. Furthermore, good correlation coefficient of 0.797 was obtained between the IC₅₀ values and MM-GBSA predicted ΔG_bind. Taken together, the current work provides not only a novel scaffold for further optimization of DHFR inhibitors but also an understanding of the specific interactions of inhibitors with DHFR and structural modifications that improve selectivity.