Development and validation of a genetic algorithm for flexible docking

Comparative evaluation of 11 scoring functions for molecular docking

Eleven popular scoring functions have been tested on 100 protein-ligand complexes to evaluate their abilities to reproduce experimentally determined structures and binding affinities. They include four scoring functions implemented in the LigFit module in Cerius2 (LigScore, PLP, PMF, and LUDI), four scoring functions implemented in the CScore module in SYBYL (F-Score, G-Score, D-Score, and ChemScore), the scoring function implemented in the AutoDock program, and two stand-alone scoring functions (DrugScore and X-Score). These scoring functions are not tested in the context of a particular docking program. Instead, conformational sampling and scoring are separated into two consecutive steps. First, an exhaustive conformational sampling is performed by using the AutoDock program to generate an ensemble of docked conformations for each ligand molecule. This conformational ensemble is required to cover the entire conformational space as much as possible rather than to focus on a few energy minima. Then, each scoring function is applied to score this conformational ensemble to see if it can identify the experimentally observed conformation from all of the other decoys. Among all of the scoring functions under test, six of them, i.e., PLP, F-Score, LigScore, DrugScore, LUDI, and X-Score, yield success rates higher than the AutoDock scoring function. The success rates of these six scoring functions range from 66% to 76% if using root-mean-square deviation < or =2.0 A as the criterion. Combining any two or three of these six scoring functions into a consensus scoring scheme further improves the success rate to nearly 80% or even higher. However, when applied to reproduce the experimentally determined binding affinities of the 100 protein-ligand complexes, only X-Score, PLP, DrugScore, and G-Score are able to give correlation coefficients over 0.50. All of the 11 scoring functions are further inspected by their abilities to construct a descriptive, funnel-shaped energy surface for protein-ligand complexation. The results indicate that X-Score and DrugScore perform better than the other ones at this aspect.

Predicting drug metabolism: a site of metabolism prediction tool applied to the cytochrome P450 2C9

The aim of the present study is to develop a method for predicting the site at which molecules will be metabolized by CYP 2C9 (cytochrome P450 2C9) using a previously reported protein homology model of the enzyme. Such a method would be of great help in designing new compounds with a better pharmacokinetic profile, or in designing prodrugs where the compound needs to be metabolized in order to become active. The methodology is based on a comparison between alignment-independent descriptors derived from GRID Molecular Interaction Fields for the CYP 2C9 active site, and a distance-based representation of the substrate. The predicted site of metabolism is reported as a ranking list of all the hydrogen atoms of each substrate molecule. Eighty-seven CYP 2C9-catalyzed oxidative reactions reported in the literature have been analyzed. In more than 90% of these cases, the hydrogen atom ranked at the first, second, or third position was the experimentally reported site of oxidation.

A priori molecular descriptors in QSAR: a case of HIV-1 protease inhibitors. II. Molecular graphics and modeling

Molecular graphics and modeling methods illustrated the chemical background of the a priori approach from part I, and visualized steric and electronic enzyme-inhibitor relationships at qualitative and quantitative level for 34 and its derivatives. The enzyme-inhibitor electron density overlap occurs at 1.5-5.5A cut-off distance, beyond van der Waals radii. Derivatives of 34 exhibit linear relationships between biological activity, molecular size and number of intermolecular interactions.

Information-based methods in the development of antiparasitic drugs

The discovery and optimization of antiparasitic compounds has profited by information-based methods newly emerged in the modern drug development process. The generation of computer models enables the cost-efficient and fast computational screening of virtual compound libraries for biologically active molecules. Two sources of information are available: structure-based drug design utilizes information about the disease target. We describe two different computational approaches, realized as the fast, flexible docking program FlexX and as the de novo design program LUDI. Ligand-based drug design, on the other hand, requires the structures and experimental data from biologically active compounds. Parasitic targets and antiparasitic compounds studied by various information-based methods include trypanosomal trypanothione reductase, antiprotozoal bisphosphonates, and trypanosomal glycosomal glyceraldehyde-3-phosphate dehydrogenase.

Molecular shape diversity of combinatorial libraries: a prerequisite for broad bioactivity

A computational method to rapidly assess and visualize the diversity in molecular shape associated with a given compound set has been developed. Normalized ratios of principal moments of inertia are plotted into two-dimensional triangular graphs and then used to compare the shape space covered by different compound sets, such as combinatorial libraries of varying size and composition. We have further developed a computational method to analyze interset similarity in terms of shape space coverage, which allows the shape redundancy between the different subsets of a given compound collection to be analyzed in a quantitative way. The shape space coverage has been found to originate mainly from the nature and the 3D-geometry (but not the size) of the central scaffold, while the number and nature of the peripheral substituents and conformational aspects were shown to be of minor importance. Substantial shape space coverage has been correlated with broad biological activity by applying the same shape analysis to collections of known bioactive compounds, such as MDDR and the GOLD-set. The aggregate of our results corroborates the intuitive notion that molecular shape is intimately linked to biological activity and that a high degree of shape (hence scaffold) diversity in screening collections will increase the odds of addressing a broad range of biological targets.

Pharmacophore-based molecular docking to account for ligand flexibility

Rapid computational mining of large 3D molecular databases is central to generating new drug leads. Accurate virtual screening of large 3D molecular databases requires consideration of the conformational flexibility of the ligand molecules. Ligand flexibility can be included without prohibitively increasing the search time by docking ensembles of precomputed conformers from a conformationally expanded database. A pharmacophore-based docking method whereby conformers of the same or different molecules are overlaid by their largest 3D pharmacophore and simultaneously docked by partial matches to that pharmacophore is presented. The method is implemented in DOCK 4.0.

Mechanism of action of the diazabicyclononanone-type kappa-agonists

The 2,4-di-2-pyridyl-3,7-dimethyl-3,7-diazabicyclo[3.3.1]nonan-9-one 1,5-diester HZ2 was recently found to exhibit high affinity and selectivity to the kappa-opioid receptor (KOR) in combination with an unusually long duration of action. Docking of HZ2 to the putative binding site model of the KOR revealed HZ2 to be tightly sitting in the binding pocket. Strong interactions, especially salts bridges between the protonated nitrogens of HZ2 and the glutamic acids 209 and 297, nicely explain the high affinity of HZ2 to the KOR. A formation of a hemiaminal bond between the keto carbonyl group of HZ2 and a lysine residue (Lys200) may explain the long duration of action.

Synthesis and evaluation of a novel nonsteroidal-specific endothelial cell proliferation inhibitor

The identification of agents with specific antiproliferative or cytostatic activity against endothelial cells has significant value for the treatment of pathologies associated with angiogenesis, including solid tumors. Here, we describe a novel substituted dibenzo[b,d]pyran-6-one scaffold, exemplified by structures 9a and 10, and report preliminary in vitro activity data indicating that this scaffold is a promising lead for the development of specific inhibitors of endothelial cell proliferation.

Ligand-induced conformational changes and a mechanism for domain closure in Aspergillus nidulans dehydroquinate synthase

In order to investigate systematically substrate and cofactor-induced conformational changes in the enzyme dehydroquinate synthase (DHQS), eight structures representing a series of differently liganded states have been determined in a total of six crystal forms. DHQS in the absence of the substrate analogue carbaphosphonate, either unliganded or in the presence of NAD or ADP, is in an open form where a relative rotation of 11-13 degrees between N and C-terminal domains occurs. Analysis of torsion angle difference plots between sets of structures reveals eight rearrangements that appear relevant to domain closure and a further six related to crystal packing. Overlapping 21 different copies of the individual N and C-terminal DHQS domains further reveals a series of pivot points about which these movements occur and illustrates the way in which widely separated secondary structure elements are mechanically inter-linked to form "composite elements", which propagate structural changes across large distances. This analysis has provided insight into the basis of DHQS ligand-initiated domain closure and gives rise to the proposal of an ordered sequence of events involving substrate binding, and local rearrangements within the active site that are propagated to the hinge regions, leading to closure of the active-site cleft.

Efficient conformational sampling of local side-chain flexibility

Side-chain flexibility of ligand-binding sites needs to be considered in the rational design of novel inhibitors. We have developed a method to generate conformational ensembles that efficiently sample local side-chain flexibility from a single crystal structure. The rotamer-based approach is tested here for the S1' pocket of human collagenase-1 (MMP-1), which is known to undergo conformational changes in multiple side-chains upon binding of certain inhibitors. First, a raw ensemble consisting of a large number of conformers of the S1' pocket was generated using an exhaustive search of rotamer combinations on a template crystal structure. A combination of principal component analysis and fuzzy clustering was then employed to successfully identify a core ensemble consisting of a low number of representatives from the raw ensemble. The core ensemble contained geometrically diverse conformers of stable nature, as indicated in several cases by a relative energy lower than that of the minimised template crystal structure. Through comparisons with X-ray crystallography and NMR structural data we show that the core ensemble occupied a conformational space similar to that observed under experimental conditions. The synthetic inhibitor RS-104966 is known to induce a conformational change in the side-chains of the S1' pocket of MMP-1 and could not be docked in the template crystal structure. However, the experimental binding mode was reproduced successfully using members of the core ensemble as the docking target, establishing the usefulness of the method in drug design.

A priori molecular descriptors in QSAR: a case of HIV-1 protease inhibitors. I. The chemometric approach

A quantitative structure-activity relationship (QSAR) study on 48 peptidic HIV-1 protease inhibitors was performed. Fourteen a priori molecular descriptors were used to build QSAR models. Hierarchical cluster analysis (HCA), principal component analysis (PCA) and partial least squares (PLS) regression were employed. PLS models with 32/16 (model I) and 48/0 (model II) molecules in the training/external validation set were constructed. The a priori molecular descriptors were related to two energetic variables using PLS. HCA and PCA on data from model II classified the inhibitors as slightly, moderately and highly active; three principal components, the chemical nature of which has been highlighted, are enough to describe the enzyme-inhibitor binding. Model I (r(2)=0.91, q(2)=0.84) is comparable to literature models obtained by various QSAR softwares, which justified the use of a priori descriptors.

Molecular modeling and metabolic studies of the interaction of catechol-O-methyltransferase and a new nitrocatechol inhibitor

Catechol-O-methyltransferase (COMT, EC 2.1.1.6) plays a central role in the metabolic inactivation of neurotransmitters and neuroactive xenobiotics possessing a catechol motif. 1-(3,4-Dihydroxy-5-nitrophenyl)-2-phenyl-ethanone (BIA 3-202) is a novel nitrocatechol-type inhibitor of COMT, the potential clinical benefit of which is currently being evaluated in the treatment of Parkinson's disease. In the present work we characterize the molecular interactions of BIA 3-202 within the active site of COMT and discuss their implication on the regioselectivity of metabolic O-methylation. Unrestrained flexible-docking simulations suggest that the solution structure of this complex is better described as an ensemble of alternative binding modes, in contrast to the well defined bound configuration revealed by the X-ray structures of related nitrocatechol inhibitors, co-crystallized with COMT. The docking results wherein presented are well supported by experimental evidence, where the pattern of in vitro enzymatic O-methylation and O-demethylation reactions are analyzed. We propose a plausible explanation for the paradoxical in vivo regioselectivity of O-methylation of BIA 3-202, as well as of its related COMT inhibitor tolcapone. Both compounds undergo in vivo O-methylation by COMT at either meta or para catechol hydroxyl groups. However, results herein presented suggest that, in a subsequent step, the p-O-methyl derivatives are selectively demethylated by a microsomal enzyme system. The overall balance is the accumulation of the m-O-methylated metabolites over the para-regioisomers. The implications for the general recognition of nitrocatechol-type inhibitors by COMT and the regioselectivity of their metabolic O-methylation are discussed.

Structure of a human carcinogen-converting enzyme, SULT1A1. Structural and kinetic implications of substrate inhibition

Sulfonation catalyzed by sulfotransferase enzymes plays an important role in chemical defense mechanisms against various xenobiotics but also bioactivates carcinogens. A major human sulfotransferase, SULT1A1, metabolizes and/or bioactivates many endogenous compounds and is implicated in a range of cancers because of its ability to modify diverse promutagen and procarcinogen xenobiotics. The crystal structure of human SULT1A1 reported here is the first sulfotransferase structure complexed with a xenobiotic substrate. An unexpected finding is that the enzyme accommodates not one but two molecules of the xenobiotic model substrate p-nitrophenol in the active site. This result is supported by kinetic data for SULT1A1 that show substrate inhibition for this small xenobiotic. The extended active site of SULT1A1 is consistent with binding of diiodothyronine but cannot easily accommodate beta-estradiol, although both are known substrates. This observation, together with evidence for a disorder-order transition in SULT1A1, suggests that the active site is flexible and can adapt its architecture to accept diverse hydrophobic substrates with varying sizes, shapes and flexibility. Thus the crystal structure of SULT1A1 provides the molecular basis for substrate inhibition and reveals the first clues as to how the enzyme sulfonates a wide variety of lipophilic compounds.

Surflex: fully automatic flexible molecular docking using a molecular similarity-based search engine

Surflex is a fully automatic flexible molecular docking algorithm that combines the scoring function from the Hammerhead docking system with a search engine that relies on a surface-based molecular similarity method as a means to rapidly generate suitable putative poses for molecular fragments. Results are presented evaluating reliability and accuracy of dockings compared with crystallographic experimental results on 81 protein/ligand pairs of substantial structural diversity. In over 80% of the complexes, Surflex's highest scoring docked pose was within 2.5 A root-mean-square deviation (rmsd), with over 90% of the complexes having one of the top ranked poses within 2.5 A rmsd. Results are also presented assessing Surflex's utility as a screening tool on two protein targets (thymidine kinase and estrogen receptor) using data sets on which competing methods were run. Performance of Surflex was significantly better, with true positive rates of greater than 80% at false positive rates of less than 1%. Docking time was roughly linear in number of rotatable bonds, beginning with a few seconds for rigid molecules and adding approximately 10 s per rotatable bond.

Design, synthesis and structure-activity relationships of benzoxazinone-based factor Xa inhibitors

A series of benzoxazinone derivatives was designed and synthesized as factor Xa inhibitors. We demonstrated that the naphthyl moiety in the aniline-based compounds 1 and 2 can be replaced with benzene-fused heterobicycles and biaryls to give factor Xa inhibitors with improved trypsin selectivity. The P4 modifications lead to monoamidines which are moderately active. The benzoxazinones 41-45 are potent against factor Xa, retain the improved trypsin selectivity of the corresponding aniline-based compounds, and show strong antithrombotic effect dose responsively.

Residues glutamate 216 and aspartate 301 are key determinants of substrate specificity and product regioselectivity in cytochrome P450 2D6

Cytochrome P450 2D6 (CYP2D6) metabolizes a wide range of therapeutic drugs. CYP2D6 substrates typically contain a basic nitrogen atom, and the active-site residue Asp-301 has been implicated in substrate recognition through electrostatic interactions. Our recent computational models point to a predominantly structural role for Asp-301 in loop positioning (Kirton, S. B., Kemp, C. A., Tomkinson, N. P., St.-Gallay, S., and Sutcliffe, M. J. (2002) Proteins 49, 216-231) and suggest a second acidic residue, Glu-216, as a key determinant in the binding of basic substrates. We have evaluated the role of Glu-216 in substrate recognition, along with Asp-301, by site-directed mutagenesis. Reversal of the Glu-216 charge to Lys or substitution with neutral residues (Gln, Phe, or Leu) greatly decreased the affinity (K(m) values increased 10-100-fold) for the classical basic nitrogen-containing substrates bufuralol and dextromethorphan. Altered binding was also manifested in significant differences in regiospecificity with respect to dextromethorphan, producing enzymes with no preference for N-demethylation versus O-demethylation (E216K and E216F). Neutralization of Asp-301 to Gln and Asn had similarly profound effects on substrate binding and regioselectivity. Intriguingly, removal of the negative charge from either 216 or 301 produced enzymes (E216A, E216K, and D301Q) with elevated levels (50-75-fold) of catalytic activity toward diclofenac, a carboxylate-containing CYP2C9 substrate that lacks a basic nitrogen atom. Activity was increased still further (>1000-fold) upon neutralization of both residues (E216Q/D301Q). The kinetic parameters for diclofenac (K(m) 108 microm, k(cat) 5 min(-1)) along with nifedipine (K(m) 28 microm, k(cat) 2 min(-1)) and tolbutamide (K(m) 315 microm, k(cat) 1 min(-1)), which are not normally substrates for CYP2D6, were within an order of magnitude of those observed with CYP3A4 or CYP2C9. Neutralizing both Glu-216 and Asp-301 thus effectively alters substrate recognition illustrating the central role of the negative charges provided by both residues in defining the specificity of CYP2D6 toward substrates containing a basic nitrogen.

Binding site characteristics in structure-based virtual screening: evaluation of current docking tools

Two new docking programs FRED (OpenEye Scientific Software) and Glide (Schrödinger, Inc.) in combination with various scoring functions implemented in these programs have been evaluated against a variety of seven protein targets (cyclooxygenase-2, estrogen receptor, p38 MAP kinase, gyrase B, thrombin, gelatinase A, neuraminidase) in order to assess their accuracy in virtual screening. Sets of known inhibitors were added to and ranked relative to a random library of drug-like compounds. Performance was compared in terms of enrichment factors and CPU time consumption. Results and specific features of the two new tools are discussed and compared to previously published results using FlexX (Tripos, Inc.) as a docking engine. In addition, general criteria for the selection of docking algorithms and scoring functions based on binding-site characteristics of specific protein targets are proposed. Figure Enrichment factors obtained with FlexX, Glide and FRED docking engines in combination with different scoring functions for seven selected targets with highly variable binding sites

New amide-bearing benzolactam-based protein kinase C modulators induce enhanced secretion of the amyloid precursor protein metabolite sAPPalpha

Protein kinase C (PKC) is known to participate in the processing of the amyloid precursor protein (APP). Abnormal processing of APP through the action of the beta- and gamma-secretases leads to the production of the 39-43 amino acid Abeta fragment, which is neurotoxic and which is believed to play an important role in the etiology of Alzheimer's disease. PKC activation enhances alpha-secretase activity, which results in a decrease of the amyloidogenic products of beta-secretase. In this article, we describe the synthesis of 10 new benzolactam V8 based PKC activators having side chains of varied saturation and lipophilicity linked to the aromatic ring through an amide group. The K(i) values measured for the inhibition of phorbol ester binding to PKCalpha are in the nanomolar range and show some correlation with their lipophilicity. Compounds 5g and 5h show the best binding affinity among the 10 benzolactams that were synthesized. By use of a cell line derived from an AD patient, significant enhancement of sAPPalpha secretion was achieved at 1 microM concentration for most of the compounds studied and at 0.1 microM for compounds 5e and 5f. At 1 microM the enhancement of sAPPalpha secretion for compounds 5c-h is higher than that observed for the control compound 8-(1-decynyl)benzolactam (BL). Of interest is the absence of activity found for the highly lipophilic ligand 5i, which has a K(i) of 11 nM. On the other hand, its saturated counterpart 5j, which possesses a comparable K(i) and ClogP, retains activity in the secretase assay. In the hyperplasia studies, 5f showed a modest response at 100 microg and 5e at 300 microg, suggesting that 5f was approximately 30-fold less potent than the PKC activator mezerein and 100-fold less potent than TPA. 5e was approximately 3-fold less active than 5f. On the basis of the effect of unsaturation for other potent PKC ligands, we would predict that 5e would retain biological activity in most assays but would show a marked loss of tumor-promoting activity. Compound 5e thus becomes a viable candidate compound in the search for Alzheimer's therapeutics capable of modulating amyloid processing.

Rational approaches towards reversible inhibition of type B monoamine oxidase. Design and evaluation of a novel 5H-Indeno[1,2-c]pyridazin-5-one derivative

The stereoelectronic properties of several potent reversible monoamine oxidase B (MAO-B) inhibitors were studied with a view to develop a pharmacophore model for reversible MAO-B inhibition. This study suggested that important specific H-bond and hydrophobic interactions are required for potent and selective MAO-B inhibition. These requirements were applied in the design and synthesis of a novel reversible and selective MAO-B inhibitor, 3-methyl-8-(4,4,4-trifluoro-butoxy)indeno[1,2-c]pyridazin-5-one, that is ca. 7000 times more selective as an inhibitor for MAO-B than for MAO-A, with K(i(MAO-B)) in the low nanomolar range.

Protein-based virtual screening of chemical databases. II. Are homology models of G-Protein Coupled Receptors suitable targets?

The aim of the current study is to investigate whether homology models of G-Protein-Coupled Receptors (GPCRs) that are based on bovine rhodopsin are reliable enough to be used for virtual screening of chemical databases. Starting from the recently described 2.8 A-resolution X-ray structure of bovine rhodopsin, homology models of an "antagonist-bound" form of three human GPCRs (dopamine D3 receptor, muscarinic M1 receptor, vasopressin V1a receptor) were constructed. The homology models were used to screen three-dimensional databases using three different docking programs (Dock, FlexX, Gold) in combination with seven scoring functions (ChemScore, Dock, FlexX, Fresno, Gold, Pmf, Score). Rhodopsin-based homology models turned out to be suitable, indeed, for virtual screening since known antagonists seeded in the test databases could be distinguished from randomly chosen molecules. However, such models are not accurate enough for retrieving known agonists. To generate receptor models better suited for agonist screening, we developed a new knowledge- and pharmacophore-based modeling procedure that might partly simulate the conformational changes occurring in the active site during receptor activation. Receptor coordinates generated by this new procedure are now suitable for agonist screening. We thus propose two alternative strategies for the virtual screening of GPCR ligands, relying on a different set of receptor coordinates (antagonist-bound and agonist-bound states).

Virtual screening to enrich hit lists from high-throughput screening: a case study on small-molecule inhibitors of angiogenin

"Hit lists" generated by high-throughput screening (HTS) typically contain a large percentage of false positives, making follow-up assays necessary to distinguish active from inactive substances. Here we present a method for improving the accuracy of HTS hit lists by computationally based virtual screening (VS) of the corresponding chemical libraries and selecting hits by HTS/VS consensus. This approach was applied in a case study on the target-enzyme angiogenin, a potent inducer of angiogenesis. In conjunction with HTS of the National Cancer Institute Diversity Set and ChemBridge DIVERSet E (approximately 18,000 compounds total), VS was performed with two flexible library docking/scoring methods, DockVision/Ludi and GOLD. Analysis of the results reveals that dramatic enrichment of the HTS hit rate can be achieved by selecting compounds in consensus with one or both of the VS functions. For example, HTS hits ranked in the top 2% by GOLD included 42% of the true hits, but only 8% of the false positives; this represents a sixfold enrichment over the HTS hit rate. Notably, the HTS/VS method was effective in selecting out inhibitors with midmicromolar dissociation constants typical of leads commonly obtained in primary screens.

LigandFit: a novel method for the shape-directed rapid docking of ligands to protein active sites

We present a new shape-based method, LigandFit, for accurately docking ligands into protein active sites. The method employs a cavity detection algorithm for detecting invaginations in the protein as candidate active site regions. A shape comparison filter is combined with a Monte Carlo conformational search for generating ligand poses consistent with the active site shape. Candidate poses are minimized in the context of the active site using a grid-based method for evaluating protein-ligand interaction energies. Errors arising from grid interpolation are dramatically reduced using a new non-linear interpolation scheme. Results are presented for 19 diverse protein-ligand complexes. The method appears quite promising, reproducing the X-ray structure ligand pose within an RMS of 2A in 14 out of the 19 complexes. A high-throughput screening study applied to the thymidine kinase receptor is also presented in which LigandFit, when combined with LigScore, an internally developed scoring function, yields very good hit rates for a ligand pool seeded with known actives.

Pin1 and Par14 peptidyl prolyl isomerase inhibitors block cell proliferation

Disruption of the parvulin family peptidyl prolyl isomerase (PPIase) Pin1 gene delays reentry into the cell cycle when quiescent primary mouse embryo fibroblasts are stimulated with serum. Since Pin1 regulates cell cycle progression, a Pin1 inhibitor would be expected to block cell proliferation. To identify such inhibitors, we screened a chemical compound library for molecules that inhibited human Pin1 PPIase activity in vitro. We found a set of compounds that inhibited Pin1 PPIase activity in vitro with low microM IC50s and inhibited the growth of several cancer lines. Among the inhibitors, PiB, diethyl-1,3,6,8-tetrahydro-1,3,6,8-tetraoxobenzo[lmn] phenanthroline-2,7-diacetate ethyl 1,3,6,8-tetrahydro-1,3,6,8-tetraoxo-benzo[lmn] phenanthroline-(2H,7H)-diacetate, had the least nonspecific toxicity. These results suggest that Pin1 inhibitors could be used as a novel type of anticancer drug that acts by blocking cell cycle progression.

Structure-based analysis of the ultraspiracle protein and docking studies of putative ligands

The ultraspiracle protein (USP) is the insect ortholog of the mammalian retinoid X receptor (RXR). Fundamental questions concern the functional role of USP as the heterodimerization partner of insect nuclear receptors such as the ecdysone receptor. The crystallographic structures of the ligand binding domain of USPs of Heliothis virescens and Drosophila melanogaster solved recently show that helix 12 is locked in an antagonist conformation raising the question whether USPs could adopt an agonist conformation as observed in RXRalpha. In order to investigate this hypothesis, a homology model for USP is proposed that allows a structural analysis of the agonist conformation of helix 12 based on the sequence comparison with RXR. For USP, one of the main issues concerns its function and in particular whether its activity is ligand independent or not. The x-ray structures strongly suggest that USP can bind ligands. Putative ligands have therefore been docked in the USP homology model. Juvenile hormones and juvenile hormone analogs were chosen as target ligands for the docking study. The interaction between the ligand and the receptor are examined in terms of the pocket shape as well as in terms of the chemical nature of the residues lining the ligand binding cavity.

Enhanced docking with the mining minima optimizer: acceleration and side-chain flexibility

The ligand-protein docking algorithm based on the Mining Minima method has been substantially enhanced. First, the basic algorithm is accelerated by: (1) adaptively determining the extent of each energy well to help avoid previously discovered energy minima; (2) biasing the search away from ligand positions at the surface of the receptor to prevent the ligand from staying at the surface when large sampling regions are used; (3) quickly testing multiple different ligand positions and orientations for each ligand conformation; and (4) tuning the source code to increase computational efficiency. These changes markedly shorten the time needed to discover an accurate result, especially when large sampling regions are used. The algorithm now also allows user-selected receptor sidechains to be treated as mobile during the docking procedure. The energies associated with the mobile side chains are computed as if they belonged to the ligand, except that atoms at the boundary between side chains and the rigid backbone are treated specially. This new capability is tested for several well-known ligand/protein systems, and preliminary application to an enzyme whose substrate is unknown--the recently solved hypothetical protein YecO (HI0319) from Haemophilus influenzae--indicates that side-chains relaxations allow candidate substrates of various sizes to be accommodated.

Three-dimensional molecular descriptors and a novel QSAR method

A novel set of molecular descriptors suitable for use in quantitative structure-activity relationships and related methods is described. These descriptors are a smooth and interpretable representation of atomic physicochemical property values and intramolecular atom pair distances. Distance atomic physicochemical parameter energy relationships (DAPPER), a novel structure-activity relationship (QSAR) method using these descriptors, is validated on standard datasets.

Distilling the essential features of a protein surface for improving protein-ligand docking, scoring, and virtual screening

For the successful identification and docking of new ligands to a protein target by virtual screening, the essential features of the protein and ligand surfaces must be captured and distilled in an efficient representation. Since the running time for docking increases exponentially with the number of points representing the protein and each ligand candidate, it is important to place these points where the best interactions can be made between the protein and the ligand. This definition of favorable points of interaction can also guide protein structure-based ligand design, which typically focuses on which chemical groups provide the most energetically favorable contacts. In this paper, we present an alternative method of protein template and ligand interaction point design that identifies the most favorable points for making hydrophobic and hydrogen-bond interactions by using a knowledge base. The knowledge-based protein and ligand representations have been incorporated in version 2.0 of SLIDE and resulted in dockings closer to the crystal structure orientations when screening a set of 57 known thrombin and glutathione S-transferase (GST) ligands against the apo structures of these proteins. There was also improved scoring enrichment of the dockings, meaning better differentiation between the chemically diverse known ligands and a approximately 15,000-molecule dataset of randomly-chosen small organic molecules. This approach for identifying the most important points of interaction between proteins and their ligands can equally well be used in other docking and design techniques. While much recent effort has focused on improving scoring functions for protein-ligand docking, our results indicate that improving the representation of the chemistry of proteins and their ligands is another avenue that can lead to significant improvements in the identification, docking, and scoring of ligands.

Pattern recognition and massively distributed computing

A feature of Peter Kollman's research was his exploitation of the latest computational techniques to devise novel applications of the free energy perturbation method. He would certainly have seized upon the opportunities offered by massively distributed computing. Here we describe the use of over a million personal computers to perform virtual screening of 3.5 billion druglike molecules against protein targets by pharmacophore pattern matching, together with other applications of pattern recognition such as docking ligands without any a priori knowledge about the binding site location.

A new test set for validating predictions of protein-ligand interaction

We present a large test set of protein-ligand complexes for the purpose of validating algorithms that rely on the prediction of protein-ligand interactions. The set consists of 305 complexes with protonation states assigned by manual inspection. The following checks have been carried out to identify unsuitable entries in this set: (1) assessing the involvement of crystallographically related protein units in ligand binding; (2) identification of bad clashes between protein side chains and ligand; and (3) assessment of structural errors, and/or inconsistency of ligand placement with crystal structure electron density. In addition, the set has been pruned to assure diversity in terms of protein-ligand structures, and subsets are supplied for different protein-structure resolution ranges. A classification of the set by protein type is available. As an illustration, validation results are shown for GOLD and SuperStar. GOLD is a program that performs flexible protein-ligand docking, and SuperStar is used for the prediction of favorable interaction sites in proteins. The new CCDC/Astex test set is freely available to the scientific community (http://www.ccdc.cam.ac.uk).

Sensitivity of molecular docking to induced fit effects in influenza virus neuraminidase

Many proteins undergo small side chain or even backbone movements on binding of different ligands into the same protein structure. This is known as induced fit and is potentially problematic for virtual screening of databases against protein targets. In this report we investigate the limits of the rigid protein approximation used by the docking program, GOLD, through cross-docking using protein structures of influenza neuraminidase. Neuraminidase is known to exhibit small but significant induced fit effects on ligand binding. Some neuraminidase crystal structures caused concern due to the bound ligand conformation and GOLD performed poorly on these complexes. A 'clean' set, which contained unique, unambiguous complexes, was defined. For this set, the lowest energy structure was correctly docked (i.e. RMSD < 1.5 A away from the crystal reference structure) in 84% of proteins, and the most promiscuous protein (1mwe) was able to dock all 15 ligands accurately including those that normally required an induced fit movement. This is considerably better than the 70% success rate seen with GOLD against general validation sets. Inclusion of specific water molecules involved in water-mediated hydrogen bonds did not significantly improve the docking performance for ligands that formed water-mediated contacts but it did prevent docking of ligands that displaced these waters. Our data supports the use of a single protein structure for virtual screening with GOLD in some applications involving induced fit effects, although care must be taken to identify the protein structure that performs best against a wide variety of ligands. The performance of GOLD was significantly better than the GOLD implementation of ChemScore and the reasons for this are discussed. Overall, GOLD has shown itself to be an extremely good, robust docking program for this system.

Design, synthesis, and tripeptidyl peptidase II inhibitory activity of a novel series of (S)-2,3-dihydro-2-(4-alkyl-1H-imidazol-2-yl)-1H-indoles

Butabindide, 1, was previously reported as a potent inhibitor (IC50 = 7 nM) of the serine protease enzyme tripeptidyl peptidase II (TPPII), an endogenous protease that degrades cholecystokinin-8 (CCK-8). We found that 1 has some inherent chemical instability, yielding diketopiperazine 2 fairly readily under mimicked physiological conditions. We therefore prepared imidazoles 3, which are void of 1's inherent instability, and have found that our novel analogues maintained comparable TPPII inhibitory activity (e.g.,for 3c, IC50 = 4 nM) as 1.

Impact of incorporating the 2C5 crystal structure into comparative models of cytochrome P450 2D6

Cytochrome P450 2D6 (CYP2D6) metabolizes approximately one third of the drugs in current clinical use. To gain insight into its structure and function, we have produced four different sets of comparative models of 2D6: one based on the structures of P450s from four different microorganisms (P450 terp, P450 eryF, P450 cam, and P450 BM3), another on the only mammalian P450 (2C5) structure available, and the other two based on alternative amino acid sequence alignments of 2D6 with all five of these structures. Principal component analysis suggests that inclusion of the 2C5 crystal structure has a profound effect on the modeling process, altering the general topology of the active site, and that the models produced differ significantly from all of the templates. The four models of 2D6 were also used in conjunction with molecular docking to produce complexes with the substrates codeine and 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP); this identified Glu 216 [in the F-helix; substrate recognition site (SRS) 2] as a key determinant in the binding of the basic moiety of the substrate. Our studies suggest that both Asp 301 and Glu 216 are required for metabolism of basic substrates. Furthermore, they suggest that Asp 301 (I-helix, SRS-4), a residue thought from mutagenesis studies to bind directly to the basic moiety of substrates, may play a key role in positioning the B'-C loop (SRS-1) and that the loss of activity on mutating Asp 301 may therefore be the result of an indirect effect (movement of the B'-C loop) on replacing this residue.

CLiBE: a database of computed ligand binding energy for ligand-receptor complexes

Consideration of binding competitiveness of a drug candidate against natural ligands and other drugs that bind to the same receptor site may facilitate the rational development of a candidate into a potent drug. A strategy that can be applied to computer-aided drug design is to evaluate ligand-receptor interaction energy or other scoring functions of a designed drug with that of the relevant ligands known to bind to the same binding site. As a tool to facilitate such a strategy, a database of ligand-receptor interaction energy is developed from known ligand-receptor 3D structural entries in the Protein Databank (PDB). The Energy is computed based on a molecular mechanics force field that has been used in the prediction of therapeutic and toxicity targets of drugs. This database also contains information about ligand function and other properties and it can be accessed at http://xin.cz3.nus.edu.sg/group/CLiBE.asp. The computed energy components may facilitate the probing of the mode of action and other profiles of binding. A number of computed energies of some PDB ligand-receptor complexes in this database are studied and compared to experimental binding affinity. A certain degree of correlation between the computed energy and experimental binding affinity is found, which suggests that the computed energy may be useful in facilitating a qualitative analysis of drug binding competitiveness.

The molecular and enzyme kinetic basis for the diminished activity of the cytochrome P450 2D6.17 (CYP2D6.17) variant. Potential implications for CYP2D6 phenotyping studies and the clinical use of CYP2D6 substrate drugs in some African populations

In this study, the basis for the diminished capacity of CYP2D6.17 to metabolise CYP2D6 substrate drugs and the possible implications this might have for CYP2D6 phenotyping studies and clinical use of substrate drugs were investigated in vitro. Enzyme kinetic analyses were performed with recombinant CYP2D6.1, CYP2D6.2, CYP2D6.17 and CYP2D6.T107I using bufuralol, debrisoquine, metoprolol and dextromethorphan as substrates. In addition, the intrinsic clearance of 10 CYP2D6 substrate drugs by CYP2D6.1 and CYP2D6.17 was determined by monitoring substrate disappearance. CYP2D6.17 exhibited generally higher K(m) values compared to CYP2D6.1. The V(max) values were generally not different except for metoprolol alpha-hydroxylation with the V(max) value for CYP2D6.17 being half that of CYP2D6.1. CYP2D6.1 and CYP2D6.2 displayed similar kinetics with all probe drugs except for dextromethorphan O-demethylation with the intrinsic clearance value of CYP2D6.2 being half that of CYP2D6.1. CYP2D6.17 exhibited substrate-dependent reduced clearances for the 10 substrates studied. In a clinical setting, the clearance of some drugs could be affected more than others in individuals with the CYP2D6(*)17 variant. The CYP2D6(*)17 allele might, therefore, contribute towards the poor correlation of phenotyping results when using different probe drugs in African populations. To investigate effects of CYP2D6(*)17 mutations on the structure of the enzyme, a homology model of CYP2D6 was built using the CYP2C5 crystal structure as a template. The results suggest an alteration in position of active-site residues in CYP2D6.17 as a possible explanation for the reduced activity of the enzyme.

Docking of flexible molecules using multiscale ligand representations

Structural genomics will yield an immense number of protein three-dimensional structures in the near future. Automated theoretical methodologies are needed to exploit this information and are likely to play a pivotal role in drug discovery. Here, we present a fully automated, efficient docking methodology that does not require any a priori knowledge about the location of the binding site or function of the protein. The method relies on a multiscale concept where we deal with a hierarchy of models generated for the potential ligand. The models are created using the k-means clustering algorithm. The method was tested on seven protein-ligand complexes. In the largest complex, human immunodeficiency virus reverse transcriptase/nevirapin, the root mean square deviation value when comparing our results to the crystal structure was 0.29 A. We demonstrate on an additional 25 protein-ligand complexes that the methodology may be applicable to high throughput docking. This work reveals three striking results. First, a ligand can be docked using a very small number of feature points. Second, when using a multiscale concept, the number of conformers that require to be generated can be significantly reduced. Third, fully flexible ligands can be treated as a small set of rigid k-means clusters.

Targeting thrombin and factor VIIa: design, synthesis, and inhibitory activity of functionally relevant indolizidinones

Guided by molecular modeling, docking experiments, and available X-ray crystal structure data on the serine protease Factor VIIa and thrombin, a series of indolizidinone derivatives was designed and synthesized having diverse functionality at the P1, P2, and P3 sites.

Structure activity relationship by NMR and by computer: a comparative study

There has recently been considerable interest in using NMR spectroscopy to identify ligand binding sites of macromolecules. In particular, a modular approach has been put forward by Fesik et al. (Shuker, S. B.; Hajduk, P. J.; Meadows, R. P.; Fesik, S. W. Science 1996, 274, 1531-1534) in which small ligands that bind to a particular target are identified in a first round of screening and subsequently linked together to form ligands of higher affinity. Similar strategies have also been proposed for in silico drug design, where the binding sites of small chemical groups are identified, and complete ligands are subsequently assembled from different groups that have favorable interactions with the macromolecular target. In this paper, we compare experimental and computational results on a selected target (FKBP12). The binding sites of three small ligands ((2S)1-acetylprolinemethylester, 1-formylpiperidine, 1-piperidinecarboxamide) in FKBP12 were identified independently by NMR and by computational methods. The subsequent comparison of the experimental and computational data showed that the computational method identified and ranked favorably ligand positions that satisfy the experimental NOE constraints.

A model binding site for testing scoring functions in molecular docking

Prediction of interaction energies between ligands and their receptors remains a major challenge for structure-based inhibitor discovery. Much effort has been devoted to developing scoring schemes that can successfully rank the affinities of a diverse set of possible ligands to a binding site for which the structure is known. To test these scoring functions, well-characterized experimental systems can be very useful. Here, mutation-created binding sites in T4 lysozyme were used to investigate how the quality of atomic charges and solvation energies affects molecular docking. Atomic charges and solvation energies were calculated for 172,118 molecules in the Available Chemicals Directory using a semi-empirical quantum mechanical approach by the program AMSOL. The database was first screened against the apolar cavity site created by the mutation Leu99Ala (L99A). Compared to the electronegativity-based charges that are widely used, the new charges and desolvation energies improved ranking of known apolar ligands, and better distinguished them from more polar isosteres that are not observed to bind. To investigate whether the new charges had predictive value, the non-polar residue Met102, which forms part of the binding site, was changed to the polar residue glutamine. The structure of the resulting Leu99Ala and Met102Gln double mutant of T4 lysozyme (L99A/M102Q) was determined and the docking calculation was repeated for the new site. Seven representative polar molecules that preferentially docked to the polar versus the apolar binding site were tested experimentally. All seven bind to the polar cavity (L99A/M102Q) but do not detectably bind to the apolar cavity (L99A). Five ligand-bound structures of L99A/M102Q were determined by X-ray crystallography. Docking predictions corresponded to the crystallographic results to within 0.4A RMSD. Improved treatment of partial atomic charges and desolvation energies in database docking appears feasible and leads to better distinction of true ligands. Simple model binding sites, such as L99A and its more polar variants, may find broad use in the development and testing of docking algorithms.

Insights from a three-dimensional model into ligand binding to constitutive active receptor

Two orphan nuclear receptors, constitutive active (or androstane) receptor (CAR) and pregnane X receptor (PXR), are among the most important mediators of ligand-activated transcriptional induction of liver microsomal cytochrome P450 drug-metabolizing enzymes. CAR and PXR belong to the same NR1I receptor subfamily and show high sequence homology to each other. The vitamin D receptor (VDR) also belongs to the NR1I subfamily and has the second highest homology to CAR in the ligand binding domain. A 3D model of the ligand binding domain of human CAR (hCAR) was constructed based on the available X-ray structures of human PXR (hPXR) and VDR (hVDR). The model shows that the size of the ligand binding cavities of hCAR and hPXR are similar, but larger than that of hVDR. hPXR's capability of binding to extremely large ligands, such as rifampicin, implies that its binding cavity may be able to expand further through the flexibility of a surface loop. In contrast, hCAR does not have this loop so that its cavity cannot expand, suggesting that hCAR would not bind to the largest hPXR ligands. Docking calculations of selected ligands to hCAR, based on the structural model, are consistent with previously reported receptor binding data. The results from this study indicate that structural modeling will be a useful tool for understanding ligand binding to hCAR and for design of drugs free of hCAR-mediated enzyme induction.

Mutations in cytochrome b resulting in atovaquone resistance are associated with loss of fitness in Plasmodium falciparum

Drug resistance in malarial parasites has become a major obstacle in the control of the disease. Strategies are urgently needed to control the development of resistance and to possibly reverse existing resistance. One key element required to reverse malaria drug resistance is for the parasites to "pay" a biological "cost" or suffer a loss of fitness when acquiring resistance to antimalarial drugs. Such a situation would be a disadvantage to the resistant parasites in the absence of drug pressure. We compared here the relative fitness of atovaquone-resistant Plasmodium falciparum K1 clones with single and double base mutations in their cytochrome b genes to their parent clones during erythrocytic stages in the absence of drug pressure. We found that the double amino acid mutation (M133I and G280D) is associated with a 5 to 9% loss of fitness and that the single amino acid change of M133I did not result in any detectable loss of fitness. Molecular modeling of the interaction of P. falciparum cytochrome b with ubiquinone led to the prediction that a loss of fitness of the malaria parasites would result from the G280D mutation due to its close proximity to the putative ubiquinone-binding site. This appears to have resulted in a weakening of the cytochrome b-ubiquinone complex, thereby causing the electron transport chain to become less efficient. Our results suggest that the prevalence of resistant parasites may decrease after the drug usage is discontinued.

Predicting the structure of protein cavities created by mutation

To assist in the efficient design of protein cavities, we have developed a minimization strategy that can predict with accuracy the fate of cavities created by mutation. We first modelled, under different conditions, the structures of six T4 lysozyme and cytochrome c peroxidase mutants of known crystal structure (where long, hydrophobic, buried side chains have been replaced by shorter ones) by minimizing the virtual structures derived from the corresponding wild-type co-ordinates. An unconstrained pathway together with an all-atom atom representation and a steepest descent minimization yielded modelled structures with lower root mean square deviations (r.m.s.d) from the crystal structures than other conditions. To test whether the method developed was generally applicable to other mutations of the kind, we have then modelled eighteen additional T4 lysozyme, barnase and cytochrome c peroxidase mutants of known crystal structure. The models of both cavity expanding and cavity collapsing mutants closely fit their crystal structures (average r.m.s.d. 0.33 +/- 0.25 A, with only one poorer prediction: L121A). The structure of protein cavities generated by mutation can thus be confidently simulated by energy minimization regardless of the tendency of the cavity to collapse or to expand. We think this is favoured by the fact that the typical response observed in these proteins to cavity-creating mutations is to experience only a limited rearrangement.

Crystal structures of two homologous pathogenesis-related proteins from yellow lupine

Pathogenesis-related class 10 (PR10) proteins are restricted to the plant kingdom where they are coded by multigene families and occur at high levels. In spite of their abundance, their physiological role is obscure although members of a distantly related subclass (cytokinin-specific binding proteins) are known to bind plant hormones. PR10 proteins are of special significance in legume plants where their expression patterns are related to infection by the symbiotic, nitrogen-fixing bacteria. Here we present the first crystal structures of classic PR10 proteins representing two homologues from one subclass in yellow lupine. The general fold is similar and, as in a birch pollen allergen, consists of a seven-stranded beta-sheet wrapped around a long C-terminal helix. The mouth of a large pocket formed between the beta-sheet and the helix seems a likely site for ligand binding. The shape of the pocket varies because, in variance with the rigid beta-sheet, the helix shows unusual conformational variability consisting in bending, disorder, and axial shifting. A surface loop, proximal to the entrance to the internal cavity, shows an unusual structural conservation and rigidity in contrast to the high glycine content in its sequence. The loop is different from the so-called glycine-rich P-loops that bind phosphate groups of nucleotides, but it is very likely that it does play a role in ligand binding in PR10 proteins.

Designing focused libraries using MoSELECT

When designing a combinatorial library it is usually desirable to optimise multiple properties of the library simultaneously and often the properties are in competition with one another. For example, a library that is designed to be focused around a given target molecule should ideally have minimum cost and also contain molecules that are bioavailable. In this paper, we describe the program MoSELECT for multiobjective library design that is based on a multiobjective genetic algorithm (MOGA). MoSELECT searches the product-space of a virtual combinatorial library to generate a family of equivalent solutions where each solution represents a combinatorial subset of the virtual library optimised over multiple objectives. The family of solutions allows the relationships between the objectives to be explored and thus enables the library designer to make an informed choice on an appropriate compromise solution. Experiments are reported where MoSELECT has been applied to the design of various focused libraries.