The PDBbind database: collection of binding affinities for protein-ligand complexes with known three-dimensional structures

We have screened the entire Protein Data Bank (Release No. 103, January 2003) and identified 5671 protein-ligand complexes out of 19 621 experimental structures. A systematic examination of the primary references of these entries has led to a collection of binding affinity data (K(d), K(i), and IC(50)) for a total of 1359 complexes. The outcomes of this project have been organized into a Web-accessible database named the PDBbind database.     

A general and fast scoring function for protein-ligand interactions: a simplified potential approach

A fast, simplified potential-based approach is presented that estimates the protein-ligand binding affinity based on the given 3D structure of a protein-ligand complex. This general, knowledge-based approach exploits structural information of known protein-ligand complexes extracted from the Brookhaven Protein Data Bank and converts it into distance-dependent Helmholtz free interaction energies of protein-ligand atom pairs (potentials of mean force, PMF). The definition of an appropriate reference state and the introduction of a correction term accounting for the volume taken by the ligand were found to be crucial for deriving the relevant interaction potentials that treat solvation and entropic contributions implicitly. A significant correlation between experimental binding affinities and computed score was found for sets of diverse protein-ligand complexes and for sets of different ligands bound to the same target. For 77 protein-ligand complexes taken from the Brookhaven Protein Data Bank, the calculated score showed a standard deviation from observed binding affinities of 1.8 log Ki units and an R2 value of 0.61. The best results were obtained for the subset of 16 serine protease complexes with a standard deviation of 1.0 log Ki unit and an R2 value of 0.86. A set of 33 inhibitors modeled into a crystal structure of HIV-1 protease yielded a standard deviation of 0.8 log Ki units from measured inhibition constants and an R2 value of 0.74. In contrast to empirical scoring functions that show similar or sometimes better correlation with observed binding affinities, our method does not involve deriving specific parameters that fit the observed binding affinities of protein-ligand complexes of a given training set. We compared the performance of the PMF score, B?hm's score (LUDI), and the SMOG score for eight different test sets of protein-ligand complexes. It was found that for the majority of test sets the PMF score performs best. The strength of the new approach presented here lies in its generality as no knowledge about measured binding affinities is needed to derive atomic interaction potentials. The use of the new scoring function in docking studies is outlined.     

Specific noncovalent interactions at protein-ligand interface: implications for rational drug design

Specific noncovalent interactions that are indicative of attractive, directional intermolecular forces have always been of key interest to medicinal chemists in their search for the "glue" that holds drugs and their targets together. With the rapid increase in the number of solved biomolecular structures as well as the performance enhancement of computer hardware and software in recent years, it is now possible to give more comprehensive insight into the geometrical characteristics and energetic landscape of certain sophisticated noncovalent interactions present at the binding interface of protein receptors and small ligands based on accumulated knowledge gaining from the combination of two quite disparate but complementary approaches: crystallographic data analysis and quantum-mechanical ab initio calculation. In this perspective, we survey massive body of published works relating to structural characterization and theoretical investigation of three kinds of strong, specific, direct, enthalpy-driven intermolecular forces, including hydrogen bond, halogen bond and salt bridge, involved in the formation of protein-ligand complex architecture in order to characterize their biological functions in conferring affinity and specificity for ligand recognition by host protein. In particular, the biomedical implications of raised knowledge are discussed with respect to potential applications in rational drug design.     

A knowledge-based scoring function for protein-ligand interactions: Probing the reference state

Knowledge-based scoring functions have recently emerged as an alternative and very promising way of ranking protein-ligand complexes with known 3D structure according to their binding affinities. Theses implified potential-based approaches use the structural information stored in databases of protein-ligand complexes to derive atom pair interaction potentials also known as potentials of mean force (PMF). The derived PMF depend on the definition of a suitable reference state. The reference states vary among suggested knowledge-based scoring functions. Therefore, we attempt here to shed some light on the influence of different reference state definitions on the predictive power of a knowledge-based scoring function that has been introduced by us very recently [J. Med. Chem., 42 (1999) 791]. It is shown that a reference state that implicitly and more comprehensively accounts for protein and ligand solvation gives the most consistent scoring results for four test sets of diverse protein-ligand complexes taken from the Brookhaven Protein Data Bank. It is also shown that a reference sphere radius of at least 7–8 Å is needed to effectively capture solvation effects that are treated implicitly in the scoring function.     

PDBcal: a comprehensive dataset for receptor-ligand interactions with three-dimensional structures and binding thermodynamics from isothermal titration calorimetry

Compounds designed solely based on structure often do not result in any improvement of the binding affinity because of entropy-enthalpy compensation. Thermodynamic data along with structure provide an opportunity to gain a deeper understanding of this effect and aid in the refinement of scoring functions used in computational drug design. Here, we scoured the literature and constructed the most comprehensive hand-curated calorimetry dataset to date. It contains thermodynamic and structural data for more than 400 receptor-ligand complexes. The dataset can be accessed through a web interface at http://www.pdbcal.org. The thermodynamic data consists of free energy, enthalpy, entropy and heat capacity as measured by isothermal titration calorimetry (ITC). The dataset also contains the experimental conditions that were used to carry out the ITC experiments. The chemical structures of the ligands are also provided. Analysis of the data confirms the existence of enthalpy-entropy compensation effect for the first time using strictly ITC data.     

A new concept for multidimensional selection of ligand conformations (MultiSelect) and multidimensional scoring (MultiScore) of protein-ligand binding affinities.

In this work, eight different scoring functions have been combined with the aim of improving the prediction of protein-ligand binding conformations and affinities. The obtained scores were analyzed using multivariate statistical methods to generate expressions, with the ability (1) to select the best candidate between different docked conformations of an inhibitor (MultiSelect) and (2) to quantify the protein-ligand binding affinity (MultiScore). By use of the docking program GOLD, 40 different inhibitors were docked into the active site of three matrix metalloproteinases (MMP's), yielding a total of 120 enzyme-inhibitor complexes. For each complex, a single conformation of the inhibitor was selected using principal component analysis (PCA) for the scores obtained by the eight functions SCORE, LUDI, GRID, PMF_Score, D_Score, G_Score, ChemScore, and F_Score. Binding affinities were estimated based on partial least-squares projections onto latent structures (PLS) on the eight scores of each selected inhibitor conformation. By use of this procedure, R(2) = 0.78 and Q(2) = 0.78 were obtained when comparing experimental and calculated binding affinities. MultiSelect was evaluated by applying the same method for selecting docked conformations for 18 different protein-ligand complexes of known three-dimensional structure. In all cases, the selected ligand conformations were found to be very similar to the experimentally determined ligand conformations. A more general evaluation of MultiScore was performed using a set of 120 different protein-ligand complexes for which both the three-dimensional structures and the binding affinities were known. This approach allowed an evaluation of MultiScore independently of MultiSelect. The generality of the method was verified by obtaining R(2) = 0.68 and Q(2) = 0.67, when comparing calculated and experimental binding affinities for the 120 X-ray structures. In all cases, LUDI, SCORE, GRID, and F_Score were included as important functions, whereas the fifth function was PMF_Score and ChemScore for the MMP and X-ray models, respectively.     

Probing hot spots at protein-ligand binding sites: a fragment-based approach using biophysical methods

Mapping interactions at protein-ligand binding sites is an important aspect of understanding many biological reactions and a key part of drug design. In this paper, we have used a fragment-based approach to probe "hot spots" at the cofactor-binding site of a model dehydrogenase, Escherichia coli ketopantoate reductase. Our strategy involved the breaking down of NADPH (Kd = 300 nM) into smaller fragments and the biophysical characterization of their binding using WaterLOGSY NMR spectroscopy, isothermal titration calorimetry (ITC), and inhibition studies. The weak binding affinities of fragments were measured by direct ITC titrations under low c value conditions. The 2'-phosphate and the reduced nicotinamide groups were found to contribute a large part of the binding energy. A combination of ITC and site-directed mutagenesis enabled us to locate the fragments at separate hot spots on opposite ends of the cofactor-binding site. This study has identified structural determinants for cofactor recognition that represent a blueprint for future inhibitor design.     

Predicting binding modes,binding affinities and `hot spots' for protein-ligand complexes using a knowledge-based scoring function

The development of a new knowledge-based scoring function (DrugScore) and its power to recognize binding modes close to experiment, to predict binding affinities, and to identify ‘hot spots’ in binding pockets is presented. Structural information is extracted from crystallographically determined protein-ligand complexes using ReLiBase and converted into distance-dependent pair-preferences and solvent-accessible surface (SAS) dependent singlet preferences of protein and ligand atoms. The sum of the pair preferences and the singlet preferences is calculated using the 3D structure of protein-ligand complexes either taken directly from the X-raystructure or generated by the docking tool FlexX. DrugScore discriminates efficiently between well-docked ligand binding modes (root-mean-squaredeviation <2.0 Å with respect to a crystallographically determined reference complex) and computer-generated ones largely deviating from the native structure. For two test sets (91 and 68 protein-ligand complexes, taken from the PDB) the calculated score recognizes poses deviating <2 Å from the crystal structure on rank 1 in three quarters of all possible cases. Compared to the scoring function in FlexX, this is a substantial improvement. For five test sets ofcrystallographically determined protein-ligand complexes as well as for two sets of ligand geometries generated by FlexX, the calculated score is correlated with experimentally determined binding affinities. For a set of 16 crystallographically determined serine protease inhibitor complexes, a R² value of 0.86 and a standard deviation of 0.95 log units is achievedas best result; for a set of 64 thrombin and trypsin inhibitors docked into their target proteins, aR² value of 0.48 and a standard deviation of 0.7 log units is calculated. DrugScore performs better than other state-of-the-art scoring functions. To assess DrugScore's capability to reproduce the geometry of directional interactions correctly, ‘hotspots’ are identified and visualized in terms of isocontour surfaces inside the binding pocket. A dataset of 159 X-ray protein-ligand complexes is used to reproduce and highlight the actually observed ligand atom positions. In 74% of all cases, the actually observed atom type corresponds to an atom type predicted by the most favorable score at the nearest grid point. The prediction rate increases to 85% ifat least an atom type of the same class of interaction is suggested. DrugScore is fast to compute and includes implicitly solvation and entropy contributions. Small deviations in the 3D structureare tolerated and, since only contacts to non-hydrogenatoms are regarded, it does not require any assumptions on protonation states.     

Structural requirements for imidazobenzodiazepine binding to GABA(A) receptors

Several structural subclasses of ligands bind to the benzodiazepine (BZD) binding site of the GABA(A) receptor. Previous studies from this laboratory have suggested that imidazobenzodiazepines (i-BZDs, e.g., Ro 15-1788) require domains in the BZD binding site for high-affinity binding that are distinct from the requirements of classic BZDs (e.g., flunitrazepam). Here, we used systematic mutagenesis and the substituted cysteine accessibility method to map the recognition domain of i-BZDs near two residues implicated in BZD binding, gamma(2)A79 and gamma(2)T81. Both classic BZDs and i-BZDs protect cysteines substituted at gamma(2)A79 and gamma(2)T81 from covalent modification, suggesting that these ligands may occupy common volumetric spaces during binding. However, the binding of i-BZDs is more sensitive to mutations at gamma(2)A79 than classic BZDs or BZDs that lack a 3'-imidazo substituent (e.g., midazolam). The effect that gamma(2)A79 mutagenesis has on the binding affinities of a series of structurally rigid i-BZDs is related to the volume of the 3'-imidazo substituents. Furthermore, larger amino acid side chains introduced at gamma(2)A79 cause correspondingly larger decreases in the binding affinities of i-BZDs with bulky 3' substituents. These data are consistent with a model in which gamma(2)A79 lines a subsite within the BZD binding pocket that accommodates the 3' substituent of i-BZDs. In agreement with our experimental data, computer-assisted docking of Ro 15-4513 into a molecular model of the BZD binding site positions the 3'-imidazo substituent of Ro 15-4513 near gamma(2)A79.     

Determination of protein-ligand binding modes using complexation-induced changes in (1)h NMR chemical shift

A new method for determining three-dimensional solution structures of protein-ligand complexes using experimentally determined complexation-induced changes in (1)H NMR chemical shift (CIS) is introduced. The method has been validated using the complex formed between the protein antitumor antibiotic neocarzinostatin (NCS) and a synthetic chromophore analogue. The X-ray crystal structure of the unbound protein and the backbone amide proton CIS were the input data used in the determination of the three-dimensional structure of the complex. The experimental CIS values were used in a continuous direct structure refinement process based on genetic algorithms to sample conformational space. The calculated structure of the complex agrees well with the NMR solution structure, indicating the potential of this approach for structure determination.     

Structural interaction fingerprints: a new approach to organizing, mining, analyzing, and designing protein-small molecule complexes

The combination of advances in structure-based drug design efforts in the pharmaceutical industry in parallel with structural genomics initiatives in the public domain has led to an explosion in the number of structures of protein-small molecule complexes structures. This information has critical importance to both the understanding of the structural basis for molecular recognition in biological systems and the design of better drugs. A significant challenge exists in managing this vast amount of data and fully leveraging it. Here, we review our work to develop a simple, fast way to store, organize, mine, and analyze large numbers of protein-small molecule complexes. We illustrate the utility of the approach to the management of inhibitor complexes from the protein kinase family. Finally, we describe our recent efforts in applying this method to the design of target-focused chemical libraries.     

Quisi——一种快速检测分析睡眠脑电的新技术

Quisi是一种测量、记录并加以自动分析的便携式睡眠脑电新技术。本文介绍了德国近期有关Quisi在精神科研究中的应用资料,包括问题的提出、何谓Quisi、原理与方法、Quisi与经典PSG比较、Quisi 1.0版本与传统定量分析比较、远程医疗和临床应用,并就国内开展Quisi技术的前景及相关问题作一简短评价。     

A statistical method for analyzing data collected by a creel/angler survey (part 2)

This article describes a unique analytical method employed to characterize angler activities on the lower 6-mile stretch of the Passaic River in New Jersey. The method used data collected by a creel/angler survey that was designed to capture the information necessary to calculate the exposure factors needed to characterize the fish consumption pathway for recreational anglers in a human health risk assessment for the river. The survey used two methods to address the challenges of conducting a creel/angler survey in an urban and industrial setting with limited river access. While unique, the analytical method described in this article is based upon accepted methods of interpreting survey data and basic laws of probability. This article was written as a companion to two other articles, also in this issue and cited here, of which one describes in detail the survey methodology designed for the lower Passaic River creel/angler survey to meet various challenges unique to conducting such a survey in urban and industrialized rivers, and the other presents, validates, and interprets the results of the lower Passaic River work relating to human exposure factors using the methodology described in this article.     

Structural and enzymatic analyses reveal the binding mode of a novel series of Francisella tularensis enoyl reductase (FabI) inhibitors

Because of structural and mechanistic differences between eukaryotic and prokaryotic fatty acid synthesis enzymes, the bacterial pathway, FAS-II, is an attractive target for the design of antimicrobial agents. We have previously reported the identification of a novel series of benzimidazole compounds with particularly good antibacterial effect against Francisella tularensis, a Category A biowarfare pathogen. Herein we report the crystal structure of the F. tularensis FabI enzyme in complex with our most active benzimidazole compound bound with NADH. The structure reveals that the benzimidazole compounds bind to the substrate site in a unique conformation that is distinct from the binding motif of other known FabI inhibitors. Detailed inhibition kinetics have confirmed that the compounds possess a novel inhibitory mechanism that is unique among known FabI inhibitors. These studies could have a strong impact on future antimicrobial design efforts and may reveal new avenues for the design of FAS-II active antibacterial compounds.     

Selective elimination of interactions: a method for assessing thermodynamic contributions to ligand binding with application to rhinovirus antivirals

A new method for evaluating the free energy of various physical interactions, such as hydrogen-bond, electrostatic, or van der Waals interactions, is presented. Rather than destroying or creating whole groups, selective (pairwise) interactions are eliminated from the total potential energy and the energy difference with the fully interacting system is evaluated. The exponential ensemble average of such an energy difference is then directly related to the corresponding free energy difference. This procedure is then applied to a rather large protein-ligand system involving the coat proteins of a human rhinovirus and an antiviral ligand. The results seem to indicate that a particular bent hydrogen bond between the ligand and protein system may not be favorable for binding. The method presented gives an estimate of the hydrogen bond free energy contribution with an available trajectory that was previously computed without the expenditure of sizeable computational resources such as recomputing a trajectory. This procedure is effective and efficient for computing the free energy for a given type of physical interaction. It can be used for calculating the binding energy differences for various interactions which can be used to guide the search for isosoluble synthetic targets.     

Vancomycin binding to low-affinity ligands: delineating a minimum set of interactions necessary for high-affinity binding

Bacterial resistance to vancomycin has been attributed to the loss of an intermolecular hydrogen bond between vancomycin and its peptidoglycan target when cell wall biosynthesis proceeds via depsipeptide intermediates rather than the usual polypeptide intermediates. To investigate the relative importance of this hydrogen bond to vancomycin binding, we have determined crystal structures at 1.0 A resolution for the vancomycin complexes with three ligands that mimic peptides and depsipeptides found in vancomycin-sensitive and vancomycin-resistant bacteria: N-acetylglycine, D-lactic acid, and 2-acetoxy-D-propanoic acid. These, in conjunction with structures that have been reported previously, indicate higher-affinity ligands elicit a structural change in the drug not seen with these low-affinity ligands. They also enable us to define a minimal set of drug-ligand interactions necessary to confer higher-affinity binding on a ligand. Most importantly, these structures point to factors in addition to the loss of an intermolecular hydrogen bond that must be invoked to explain the weaker affinity of vancomycin for depsipeptide ligands. These factors are important considerations for the design of vancomycin analogues to overcome vancomycin resistance.     

Some novel observations of a drug (indoprofen)—albumin interaction

Indoprofen, a new-steroidal anti-inflammatory agent, binds to human serum albumin in an endothermic reaction at low degrees of saturation and with an exothermic reaction at higher drug to protein ratios. Although indoprofen appears to bind to the same primary site as diazepam, dialysis studies show an increased binding of both drugs in the presence of the other. This novel observation is probably due to the effect of the drugs on the N → B conformational change of human serum albumin.     

Application of high-frequency rheology measurements for analyzing protein-protein interactions in high protein concentration solutions using a model monoclonal antibody (IgG2)

The purpose of this work was to explore the utilization of high-frequency rheology analysis for assessing protein-protein interactions in high protein concentration solutions. Rheology analysis of a model monoclonal immunoglobulin G2 solutions was conducted on indigenously developed ultrasonic shear rheometer at frequency of 10 MHz. Solutions at pH 9.0 behaved as most viscous and viscoelastic whereas those at pH 4.0 and 5.4 exhibited lower viscosity and viscoelasticity, respectively. Intrinsic viscosity, hydrophobicity, and conformational analysis could not account for the rheological behavior of IgG2 solutions. Zeta potential and light scattering measurements showed the significance of electroviscous and specific protein-protein interactions in governing rheology of IgG2 solutions. Specific protein-protein interactions resulted in formation of reversible higher order species of monomer. Solution storage modulus (G'), and not loss modulus or complex viscosity, was the more reliable parameter for predicting protein-protein interactions. Predictions about the nature of protein-protein interactions made on the basis of solution G' were found to be consistent with observed effect of pH and ionic strength on zeta potential and scattered intensity of IgG2 solutions. Results demonstrated the potential of high-frequency storage modulus measurements for understanding behavior of proteins in solutions and predicting the nature of protein-protein interactions.