Identify drug repurposing candidates by mining the protein data bank.

Predicting off-targets by computational methods is gaining increasing interest in early-stage drug discovery. Here, we present a computational method based on full 3D comparisons of 3D structures. When a similar binding site is detected in the Protein Data Bank (PDB) (or any protein structure database), it is possible that the corresponding ligand also binds to that similar site. On one hand, this target hopping case is probably rare because it requires a high similarity between the binding sites. On the other hand, it could be a strong rational evidence to highlight possible off-target reactions and possibly a potential undesired side effect. This target-based drug repurposing can be extended a significant step further with the capability of searching the full surface of all proteins in the PDB, and therefore not relying on pocket detection. Using this approach, we describe how MED-SuMo reproduces the repurposing of tadalafil from PDE5A to PDE4A and a structure of PDE4A with tadalafil. Searching for local protein similarities generates more hits than for whole binding site similarities and therefore fragment repurposing is more likely to occur than for drug-sized compounds. In this work, we illustrate that by mining the PDB for proteins sharing similarities with the hinge region of protein kinases. The experimentally validated examples, biotin carboxylase and synapsin, are retrieved. Further to fragment repurposing, this approach can be applied to the detection of druggable sites from 3D structures. This is illustrated with detection of the protein kinase hinge motif in the HIV-RT non-nucleosidic allosteric site.

Fast and automated functional classification with MED-SuMo: an application on purine-binding proteins.

Ligand-protein interactions are essential for biological processes, and precise characterization of protein binding sites is crucial to understand protein functions. MED-SuMo is a powerful technology to localize similar local regions on protein surfaces. Its heuristic is based on a 3D representation of macromolecules using specific surface chemical features associating chemical characteristics with geometrical properties. MED-SMA is an automated and fast method to classify binding sites. It is based on MED-SuMo technology, which builds a similarity graph, and it uses the Markov Clustering algorithm. Purine binding sites are well studied as drug targets. Here, purine binding sites of the Protein DataBank (PDB) are classified. Proteins potentially inhibited or activated through the same mechanism are gathered. Results are analyzed according to PROSITE annotations and to carefully refined functional annotations extracted from the PDB. As expected, binding sites associated with related mechanisms are gathered, for example, the Small GTPases. Nevertheless, protein kinases from different Kinome families are also found together, for example, Aurora-A and CDK2 proteins which are inhibited by the same drugs. Representative examples of different clusters are presented. The effectiveness of the MED-SMA approach is demonstrated as it gathers binding sites of proteins with similar structure-activity relationships. Moreover, an efficient new protocol associates structures absent of cocrystallized ligands to the purine clusters enabling those structures to be associated with a specific binding mechanism. Applications of this classification by binding mode similarity include target-based drug design and prediction of cross-reactivity and therefore potential toxic side effects.

Analysis of HSP90-related folds with MED-SuMo classification approach.

Three-dimensional structural information is critical for understanding functional protein properties and the precise mechanisms of protein functions implicated in physiological and pathological processes. Comparison and detection of protein binding sites are key steps for annotating structures with functional predictions and are extremely valuable steps in a drug design process. In this research area, MED-SuMo is a powerful technology to detect and characterize similar local regions on protein surfaces. Each amino acid residue's potential chemical interactions are represented by specific surface chemical features (SCFs). The MED-SuMo heuristic is based on the representation of binding sites by a graph structure suitable for exploration by an efficient comparison algorithm. We use this approach to analyze one particular SCOP superfamily which includes HSP90 chaperone, MutL/DNA topoisomerase, histidine kinases, and alpha-ketoacid dehydrogenase kinase C (BCK). They share a common fold and a common region for ATP-binding. To analyze both similar and differing features of this fold, we use a novel classification method, the MED-SuMo multi approach (MED-SMA). We highlight common and distinct features of these proteins. The different clusters created by MED-SMA yield interesting observations. For instance, one cluster gathers three types of proteins (HSP90, topoisomerase VI, and BCK) which all bind the drug radicicol.

A review of MED-SuMo applications.

Resolved three-dimensional protein structures are a major source of information for understanding protein functional properties. The current explosive growth of publicly available protein structures is producing large volumes of data for computational modelling and drug design methods. Target-based in silico drug design tools aid design and optimize compounds to bind to specific targets. MED-SuMo is a powerful technology for comparing local regions on protein surfaces, allowing similarities to be discovered and explored. This is a target-based tool that can exploit all available macromolecule structures. Its computational efficiency differentiates its approach from widely used methods such as docking and scoring, or map-based methods. As a result, MED-SuMo contributes to a large variety of real-world drug discovery applications. We review specific applications where MED-SuMo performed a significant role. These examples include functional annotation, pocket profiling, structural superposition, and functional binding site classification. We also review cases where MED-SuMo provided an innovative solution to frequent undertakings of the medicinal chemist and molecular modeller during lead discovery and lead optimization. These further cases include drug repurposing and fragment-based drug design.

Computational fragment-based drug design to explore the hydrophobic sub-pocket of the mitotic kinesin Eg5 allosteric binding site.

Eg5, a mitotic kinesin exclusively involved in the formation and function of the mitotic spindle has attracted interest as an anticancer drug target. Eg5 is co-crystallized with several inhibitors bound to its allosteric binding pocket. Each of these occupies a pocket formed by loop 5/helix alpha2 (L5/alpha2). Recently designed inhibitors additionally occupy a hydrophobic pocket of this site. The goal of the present study was to explore this hydrophobic pocket with our MED-SuMo fragment-based protocol, and thus discover novel chemical structures that might bind as inhibitors. The MED-SuMo software is able to compare and superimpose similar interaction surfaces upon the whole protein data bank (PDB). In a fragment-based protocol, MED-SuMo retrieves MED-Portions that encode protein-fragment binding sites and are derived from cross-mining protein-ligand structures with libraries of small molecules. Furthermore we have excluded intra-family MED-Portions derived from Eg5 ligands that occupy the hydrophobic pocket and predicted new potential ligands by hybridization that would fill simultaneously both pockets. Some of the latter having original scaffolds and substituents in the hydrophobic pocket are identified in libraries of synthetically accessible molecules by the MED-Search software.

Computational fragment-based approach at PDB scale by protein local similarity.

The large volume of protein-ligand structures now available enables innovative and efficient protocols in computational FBDD (Fragment-Based Drug Design) to be proposed based on experimental data. In this work, we build a database of MED-Portions, where a MED-Portion is a new structural object encoding protein-fragment binding sites. MED-Portions are derived from mining all available protein-ligand structures with any library of small molecules. Combined with the MED-SuMo software to superpose similar protein interaction surfaces, pools of matching MED-Portions can be retrieved from any binding surface query. The rapidity of this technology allows its application to a diverse set of 107 protein binding sites. The selectivity of the protocol is shown by a qualitative correlation between the average hydrophobicity of the pools of MED-Portions and those of the binding sites. To generate hitlike molecules, MED-Portions are combined in 3D with the MED-Hybridise toolkit. Our MED-Portion/MED-SuMo/MED-Hybridise protocol is applied to two targets that represent important protein superfamilies in drug design: a protein kinase and a G-Protein Coupled Receptor (GPCR). We retrieved actives molecules of PubChem bioassays for the two targets. The results show the potential for finding relevant leads from any protein 3D structure since the occurrence of interfamily MED-Portions is 25% for protein kinase and almost 100% for the GPCR.

MED-3DMC: a new tool to generate 3D conformation ensembles of small molecules with a Monte Carlo sampling of the conformational space.

Obtaining an efficient sampling of the low to medium energy regions of a ligand conformational space is of primary importance for getting insight into relevant binding modes of drug candidates, or for the screening of rigid molecular entities on the basis of a predefined pharmacophore or for rigid body docking. Here, we report the development of a new computer tool that samples the conformational space by using the Metropolis Monte Carlo algorithm combined with the MMFF94 van der Waals energy term. The performances of the program have been assessed on 86 drug-like molecules that resulted from an ADME/tox profiling applied on cocrystalized small molecules and were compared with the program Omega on the same dataset. Our program has also been assessed on the 85 molecules of the Astex diverse set. Both test sets show convincing performance of our program at sampling the conformational space.

MED-SuMoLig: a new ligand-based screening tool for efficient scaffold hopping.

The identification of small molecules with selective bioactivity, whether intended as potential therapeutics or as tools for experimental research, is central to progress in medicine and in the life sciences. To facilitate such study, we have developed a ligand-based program well-suited for effective screening of large compound collections. This package, MED-SuMoLig, combines a SMARTS-driven substructure search aiming at 3D pharmacophore profiling and computation of the local atomic density of the compared molecules. The screening utility was then investigated using 52 diverse active molecules (against CDK2, Factor Xa, HIV-1 protease, neuraminidase, ribonuclease A, and thymidine kinase) merged to a library of about 40,000 putative inactive (druglike) compounds. In all cases, the program recovered more than half of the actives in the top 3% of the screened library. We also compared the performance of MED-SuMoLig with that of ChemMine or of ROCS and found that MED-SuMoLig outperformed both methods for CDK2 and Factor Xa in terms of enrichment rates or performed equally well for the other targets.

Receptor-based computational screening of compound databases: the main docking-scoring engines.

The processes used by academic and industrial scientists to discover new drugs have recently experienced a true renaissance with many new and exciting techniques. The number of protein structures and/or chemical ligands is constantly growing, through the use of parallel chemistry, X-ray crystallography, NMR or homology modeling methods and so is the theoretical understanding of protein-ligand interactions. As such, structure-based approaches to drug-design and in silico screening are becoming routine part of most modern lead discovery programs. Prioritization of compound libraries is an extremely important task that aims at the rapid identification of tight-binding ligands and ultimately new therapeutic compounds. These in silico approaches combined with other experimental methods facilitate the design of new medicines to treat cardiovascular, degenerative, infectious, and neoplastic diseases, among others. Here, we review key concepts and specific features of several selected ligand-receptor docking/scoring methods while several other topics pertaining to the field of in silico screening are reviewed in the following articles of this special issue of Current Protein and Peptide Science.

The SuMo server: 3D search for protein functional sites.

UNLABELLED: We provide the scientific community with a web server which gives access to SuMo, a bioinformatic system for finding similarities in arbitrary 3D structures or substructures of proteins. SuMo is based on a unique representation of macromolecules using selected triplets of chemical groups having their own geometry and symmetry, regardless of the restrictive notions of main chain and lateral chains of amino acids. The heuristic for extracting similar sites was used to drive two major large-scale approaches. First, searching for ligand binding sites onto a query structure has been made possible by comparing the structure against each of the ligand binding sites found in the Protein Data Bank (PDB). Second, the reciprocal process, i.e. searching for a given 3D site of interest among the structures of the PDB is also possible and helps detect cross-reacting targets in drug design projects. AVAILABILITY: The web server is freely accessible to academia through http://sumo-pbil.ibcp.fr and full support is available from MEDIT (http://www.medit.fr). CONTACT: mjambon@burnham.org.