Abstract Details


Poster 42: Visually Interpretable Analysis of Kinase Selectivity Related Features Derived from Field-based Proteochemometrics

Vigneshwari Subramanian1, 2, Peteris Prusis1, Lars-Olof Pietilä1, Henri Xhaard2, Gerd Wohlfahrt1
1Orion Pharma, Orionintie 1, 02101 Espoo, Finland.
2Centre for Drug Research, Faculty of Pharmacy, University of Helsinki, FI-00014 Helsinki, Finland.
Protein kinases are amongst the most important drug targets as they play crucial roles in various processes such as cell growth, differentiation, apoptosis and intracellular signal transmission. Several hundred diseases are related to dysregulation of kinases [1]. Most of the kinase inhibitors currently available in the market are known to interact with multiple kinases and induce toxic side effects [2]. Achieving selectivity towards a biological target has been a main focus of pharmaceutical research, but has been proven difficult e.g. for kinases due to high similarity between their ATP binding pockets. Therefore structural characterization of the ATP binding sites of kinases is an integral part in the development of more selective inhibitors/drugs.

Comparison of molecular interaction fields of binding sites within a protein family is a valuable tool to (qualitatively) interpret the selectivity of ligands [3, 4]. A more quantitative approach to address selectivity issues of receptors is proteochemometrics, a multivariate statistics method, which aims to correlate both ligand and protein description with affinity [5, 6]. Unlike conventional QSAR models, proteochemometric models provide good predictability and interpretability for both activity and specificity simultaneously for ligands and for targets [6]. Employing molecular interaction fields to describe proteins in combination with 2D and 3D ligand descriptors in proteochemometric models provides a way for visualizing, understanding and modifying selectivity profiles of small-molecule inhibitors.

The method is demonstrated for 50 kinases with ~2600 activity values collected from the Protein Data Bank and literature [7, 8, 9]. Knowledge-based and electrostatic protein fields describing the binding sites were calculated with SVL scripts in MOE [4]. Molecular properties, SMARTS based fingerprints and field-based numerical descriptors were calculated for 80 ligands using Mold2 [11], Open Babel [12] and Volsurf [13] programs. Proteochemometric models using field-based protein descriptors, ligand descriptors and experimentally measured affinity values were generated by Partial Least Squares methods (PLS) implemented in SIMCA [10]. Visual interpretation of the models with MOE highlights protein field points and ligand functional groups which influence binding affinity and selectivity.

To date, proteochemometrics has been used only with sequence-based descriptors. Proteochemometric models based on protein-fields contain more structural details than sequence-based models, which can be visualized and used to support the design of selective inhibitors.

References
[1] Melnikova, I. and Golden, J. (2004) Targeting protein kinases. Nature Reviews Drug Discovery, 3, 993-994.
[2] Lapins, M. and Wikberg, J. E. S. (2010) Kinome-wide interaction modelling using alignment-based and alignment-independent approaches for kinase description and linear and non-linear data analysis techniques. BMC Bioinformatics, 11, 339-354.
[3] Hoppe, C., Steinbeck, C., Wohlfahrt, G. (2006) Classification and comparison of ligand-binding sites derived from grid-mapped knowledge-based potentials. J. Mol. Graph. Model., 24, 328-340.
[4] Wohlfahrt, G., Sipila, J., Pietila, L. O. (2009) Field-Based Comparison of ligand and coactivator binding sites of nuclear receptors. Biopolymers, 91, 884-894.
[5] Prusis, P., Muceniece, R., Andersson, P., Post, C., Lundstedt, T. and Wikberg, J.E.S. (2001) PLS modeling of chimeric MS04/MSH-peptide and MC1/MC3-receptor interactions reveals a novel method for the analysis of ligand-receptor interactions. Biochim. Biophys. Acta, 1544, 350-357.
[6] Wikberg, J. E. S., Lapinsh, M. and Prusis, P. (2004) Proteochemometrics: A Tool for Modeling the Molecular Interaction Space, in Chemogenomics in Drug Discovery: A Medicinal Chemistry Perspective (eds H. Kubinyi and G. Muller), Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim, FRG.
[7] Davis, M. I., et al. (2011) Comprehensive analysis of kinase inhibitor selectivity. Nature Biotechnology, 29, 1046 1051.
[8] Karamen, M. W., et al. (2008) A quantitative analysis of kinase inhibitor selectivity. Nature Biotechnology, 26, 127 -132.
[9] Metz, J.T., et al. (2011) Navigating the kinome. Nature Chemical Biology, 7, 200 202.
[10] SIMCA-P version 12, Umetrix AB, Box 7960, SE -907, 19 Umea, Sweden, 2011.
[11] Hong, H., et al. (2008) Mold2, Molecular Descriptors from 2D structures for Chemoinformatics and Toxicoinformatics. J. Chem. Inf. Model., 48, (7) 1337 1344.
[12] Boyle, N. M. O. et al. (2011) Open Babel : An open chemical toolbox. J. ChemInf., 3 33.
[13] Cruciani, G. et al.(2000) VolSurf: a new tool for the pharmacokinetic optimization of lead compounds. Eur. J. Pharm. Sci., 11, S29S39.

Return to Programme