In silico prediction of human carboxylesterase-1 (hCES1) metabolism combining docking analyses and MD simulations Giulio Vistoli a, * , Alessandro Pedretti a , Angelica Mazzolari a , Bernard Testa b a Dipartimento di Scienze Farmaceutiche ‘Pietro Pratesi’, Facoltà di Farmacia, Università degli Studi di Milano, Via Mangiagalli, 25, I-20133 Milano, Italy b Department of Pharmacy, University Hospital Centre (CHUV), Rue du Bugnon, CH-1011 Lausanne, Switzerland article info Article history: Received 14 August 2009 Revised 26 October 2009 Accepted 27 October 2009 Available online 31 October 2009 Keywords: Esterases Human carboxylesterase-1 Metabolism prediction Docking analysis Molecular dynamics abstract Metabolic problems lead to numerous failures during clinical trials, and much effort is now devoted in developing in silico models predicting metabolic stability and metabolites. Such models are well known for cytochromes P450 and some transferases, whereas little has been done to predict the hydrolytic activ- ity of human hydrolases. The present study was undertaken to develop a computational approach able to predict the hydrolysis of novel esters by human carboxylesterase hCES1. The study involves both docking analyses of known substrates to develop predictive models, and molecular dynamics (MD) simulations to reveal the in situ behavior of substrates and products, with particular attention being paid to the influ- ence of their ionization state. The results emphasize some crucial properties of the hCES1 catalytic cavity, confirming that as a trend with several exceptions, hCES1 prefers substrates with relatively smaller and somewhat polar alkyl/aryl groups and larger hydrophobic acyl moieties. The docking results underline the usefulness of the hydrophobic interaction score proposed here, which allows a robust prediction of hCES1 catalysis, while the MD simulations show the different behavior of substrates and products in the enzyme cavity, suggesting in particular that basic substrates interact with the enzyme in their unprotonated form. Ó 2009 Elsevier Ltd. All rights reserved. 1. Introduction More than half of drug candidates fail during clinical trials due to an unsuitable pharmacokinetic profile. For this reasons, the most recent strategies in drug discovery and development give attention to an ADMET profiling of new molecules with the clear aim to se- lect (and develop) only drug-like compounds with an optimal pharmacokinetic profile. Such an early pharmacokinetic analysis requires the assessment of a maximum of useful and relevant molecular properties that allow a reliable prediction of drug-like- ness also for large molecular libraries. 1 As a result, much research effort has been invested in develop- ing in silico tools to predict physicochemical properties relevant to pharmacokinetic profile, for example, aqueous solubility and lipo- philicity. Similarly, predictive tools for metabolic stability are clearly useful since inadequate metabolism represents one of the most problematic failures during clinical trials. For metabolism, however, global quantitative approaches predictive of heteroge- neous molecular datasets are virtually impossible, given the large arsenal of diverse and not fully investigated enzymes involved in human xenobiotic metabolism. And indeed, most computational approaches in the recent literature are restricted to specific metab- olizing enzymes. 2,3 The metabolizing activity of the most relevant red-ox enzymes 4 (i.e., cytochromes P450, 5 monoamine oxidases, 6 and alcohol dehy- drogenases 7 ) and those of some significant conjugating enzymes 8 (e.g., UDP-glucuronosyltransferases, 9 sulfotransferases, 10 methyl- transferases, 11 and glutathione-S-transferases 12 ) have been inves- tigated in great detail by computational techniques and can be successfully predicted by several approaches. Conversely, little has been done to rationalize and predict in silico the hydrolyzing activity of the human hydrolases, 13 although they play key roles in the hydrolytic metabolism of xenobiotics as well as in the acti- vation of most prodrugs. 14 Prodrug strategies 15,16 allow, for example, the protection of some critical function in active agents and yield compounds with, for example, increased solubility, better bioavailability, or targeted active absorption. Clearly, such prodrugs must be promptly trans- formed into the active agent, a reaction that in most cases is one of hydrolysis. Thus, the possibility to predict the likelihood of a pro- drug being efficiently transformed into the corresponding active form becomes of crucial importance when designing optimal promoieties. Carboxylesterases 17,18 play a pivotal role in the hydrolysis of a variety of drugs or prodrugs featuring an ester, amide or carbamate function. Mammalian carboxylesterases (CESs) are members of the 0968-0896/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.bmc.2009.10.052 * Corresponding author. E-mail address: giulio.vistoli@unimi.it (G. Vistoli). Bioorganic & Medicinal Chemistry 18 (2010) 320–329 Contents lists available at ScienceDirect Bioorganic & Medicinal Chemistry journal homepage: www.elsevier.com/locate/bmc