Does a more precise chemical description of protein-ligand complexes lead to more accurate prediction of binding affinity?

Predicting the binding affinities of large sets of diverse molecules against a range of macromolecular targets is an extremely challenging task. The scoring functions that attempt such computational prediction are essential for exploiting and analyzing the outputs of docking, which is in turn an important tool in problems such as structure-based drug design. Classical scoring functions assume a predetermined theory-inspired functional form for the relationship between the variables that describe an experimentally determined or modeled structure of a protein-ligand complex and its binding affinity. The inherent problem of this approach is in the difficulty of explicitly modeling the various contributions of intermolecular interactions to binding affinity. New scoring functions based on machine-learning regression models, which are able to exploit effectively much larger amounts of experimental data and circumvent the need for a predetermined functional form, have already been shown to outperform a broad range of state-of-the-art scoring functions in a widely used benchmark. Here, we investigate the impact of the chemical description of the complex on the predictive power of the resulting scoring function using a systematic battery of numerical experiments. The latter resulted in the most accurate scoring function to date on the benchmark. Strikingly, we also found that a more precise chemical description of the protein-ligand complex does not generally lead to a more accurate prediction of binding affinity. We discuss four factors that may contribute to this result: modeling assumptions, codependence of representation and regression, data restricted to the bound state, and conformational heterogeneity in data.

CREDO: a structural interactomics database for drug discovery.

CREDO is a unique relational database storing all pairwise atomic interactions of inter- as well as intra-molecular contacts between small molecules and macromolecules found in experimentally determined structures from the Protein Data Bank. These interactions are integrated with further chemical and biological data. The database implements useful data structures and algorithms such as cheminformatics routines to create a comprehensive analysis platform for drug discovery. The database can be accessed through a web-based interface, downloads of data sets and web services at http://www-cryst.bioc.cam.ac.uk/credo. Database URL: http://www-cryst.bioc.cam.ac.uk/credo.

What can we learn from the evolution of protein-ligand interactions to aid the design of new therapeutics?

Efforts to increase affinity in the design of new therapeutic molecules have tended to lead to greater lipophilicity, a factor that is generally agreed to be contributing to the low success rate of new drug candidates. Our aim is to provide a structural perspective to the study of lipophilic efficiency and to compare molecular interactions created over evolutionary time with those designed by humans. We show that natural complexes typically engage in more polar contacts than synthetic molecules bound to proteins. The synthetic molecules also have a higher proportion of unmatched heteroatoms at the interface than the natural sets. These observations suggest that there are lessons to be learnt from Nature, which could help us to improve the characteristics of man-made molecules. In particular, it is possible to increase the density of polar contacts without increasing lipophilicity and this is best achieved early in discovery while molecules remain relatively small.

USRCAT: real-time ultrafast shape recognition with pharmacophoric constraints.

BACKGROUND: Ligand-based virtual screening using molecular shape is an important tool for researchers who wish to find novel chemical scaffolds in compound libraries. The Ultrafast Shape Recognition (USR) algorithm is capable of screening millions of compounds and is therefore suitable for usage in a web service. The algorithm however is agnostic of atom types and cannot discriminate compounds with similar shape but distinct pharmacophoric features. To solve this problem, an extension of USR called USRCAT, has been developed that includes pharmacophoric information whilst retaining the performance benefits of the original method. RESULTS: The USRCAT extension is shown to outperform the traditional USR method in a retrospective virtual screening benchmark. Also, a relational database implementation is described that is capable of screening a million conformers in milliseconds and allows the inclusion of complex query parameters. CONCLUSIONS: USRCAT provides a solution to the lack of atom type information in the USR algorithm. Researchers, particularly those with only limited resources, who wish to use ligand-based virtual screening in order to discover new hits, will benefit the most. Online chemical databases that offer a shape-based similarity method might also find advantage in using USRCAT due to its accuracy and performance. The source code is freely available and can easily be modified to fit specific needs.

Annular tautomerism: experimental observations and quantum mechanics calculations.

The use of MP2 level quantum mechanical (QM) calculations on isolated heteroaromatic ring systems for the prediction of the tautomeric propensities of whole molecules in a crystalline environment was examined. A Polarisable Continuum Model was used in the calculations to account for environment effects on the tautomeric relative stabilities. The calculated relative energies of tautomers were compared to relative abundances within the Cambridge Structural Database (CSD) and the Protein Data Bank (PDB). The work was focussed on 84 annular tautomeric forms of 34 common ring systems. Good agreement was found between the calculations and the experimental data even if the quantity of these data was limited in many cases. The QM results were compared to those produced by much faster semiempirical calculations. In a search for other sources of the useful experimental data, the relative numbers of known compounds in which prototropic positions were often substituted by heavy atoms were also analysed. A scheme which groups all annular tautomeric transformations into 10 classes was developed. The scheme was designed to encompass a comprehensive set of known and theoretically possible tautomeric ring systems generated as part of a previous study. General trends across analogous ring systems were detected as a result. The calculations and statistics collected on crystallographic data as well as the general trends observed should be useful for the better modelling of annular tautomerism in the applications such as computer-aided drug design, small molecule crystal structure prediction, the naming of compounds and the interpretation of protein-small molecule crystal structures.

Atomic interactions and profile of small molecules disrupting protein-protein interfaces: the TIMBAL database.

Growing evidence of the possibility of modulating protein-protein interactions with small molecules is opening the door to new approaches and concepts in drug discovery. In this paper, we describe the creation of TIMBAL, a hand-curated database holding an up to date collection of small molecules inhibiting multi-protein complexes. This database has been analysed and profiled in terms of molecular properties. Protein-protein modulators tend to be large lipophilic molecules with few hydrogen bond features. An analysis of TIMBAL's intersection with other structural databases, including CREDO (protein-small molecule from the PDB) and PICCOLO (protein-protein from the PDB) reveals that TIMBAL molecules tend to form mainly hydrophobic interactions with only a few hydrogen bonding contacts. With respect to potency, TIMBAL molecules are slightly less efficient than an average medicinal chemistry hit or lead. The database provides a resource that will allow further insights into the types of molecules favoured by protein interfaces and provide a background to continuing work in this area. Access at http://www-cryst.bioc.cam.ac.uk/timbal.

Structural interactomics: informatics approaches to aid the interpretation of genetic variation and the development of novel therapeutics.

Here we review the use of informatics in structural interactomics, with particular emphasis on understanding interfacial contacts in the development of novel therapeutics and the interpretation of genetic variation. We describe the availability and applicability of structural databases of protein interactions which facilitate this endeavour. We demonstrate the applicability of a structural interactomics approach to the study of the fibroblast growth factor (FGF)-stimulated mitogen-activated protein kinase (MAPK) pathway.

On the origins of enzyme inhibitor selectivity and promiscuity: a case study of protein kinase binding to staurosporine.

Relationships between ligand binding and the shapes of the binding sites in families of homologous enzymes are investigated by comparing matrices of distances between key binding site atoms. Multiple linear regression is used to help identify key distances that influence ligand binding affinity. In order to illustrate the utility of this generic approach, we study protein kinase binding sites for ATP and the promiscuous competitive inhibitor, staurosporine. We show that the size of the gatekeeper residue and the closure between the first glycine of the GXGXXG motif and the aspartate of the DFG loop act together to promote tight binding. Our web-based tool, 'mapping analogous hetero-atoms onto residue interactions' (MAHORI), indicates that the greater the number of hydrogen bonds made by the kinase around the methylamine group of staurosporine, the tighter the binding. The conservation of surrounding atoms identified using our novel grid-based method clearly demonstrates that the most structurally conserved part of the binding site for staurosporine is the main chain of the hinge region. The critical role of interactions that are not dependent on side-chain identities is consistent with the promiscuous nature of this inhibitor.

CREDO: a protein-ligand interaction database for drug discovery.

Harnessing data from the growing number of protein-ligand complexes in the Protein Data Bank is an important task in drug discovery. In order to benefit from the abundance of three-dimensional structures, structural data must be integrated with sequence as well as chemical data and the protein-small molecule interactions characterized structurally at the inter-atomic level. In this study, we present CREDO, a new publicly available database of protein-ligand interactions, which represents contacts as structural interaction fingerprints, implements novel features and is completely scriptable through its application programming interface. Features of CREDO include implementation of molecular shape descriptors with ultrafast shape recognition, fragmentation of ligands in the Protein Data Bank, sequence-to-structure mapping and the identification of approved drugs. Selected analyses of these key features are presented to highlight a range of potential applications of CREDO. The CREDO dataset has been released into the public domain together with the application programming interface under a Creative Commons license at http://www-cryst.bioc.cam.ac.uk/credo. We believe that the free availability and numerous features of CREDO database will be useful not only for commercial but also for academia-driven drug discovery programmes.

A structural bioinformatics approach to the analysis of nonsynonymous single nucleotide polymorphisms (nsSNPs) and their relation to disease.

The prediction of the effects of nonsynonymous single nucleotide polymorphisms (nsSNPs) on function depends critically on exploiting all information available on the three-dimensional structures of proteins. We describe software and databases for the analysis of nsSNPs that allow a user to move from SNP to sequence to structure to function. In both structure prediction and the analysis of the effects of nsSNPs, we exploit information about protein evolution, in particular, that derived from investigations on the relation of sequence to structure gained from the study of amino acid substitutions in divergent evolution. The techniques developed in our laboratory have allowed fast and automated sequence-structure homology recognition to identify templates and to perform comparative modeling; as well as simple, robust, and generally applicable algorithms to assess the likely impact of amino acid substitutions on structure and interactions. We describe our strategy for approaching the relationship between SNPs and disease, and the results of benchmarking our approach -- human proteins of known structure and recognized mutation.