Deubiquitinase Targeted and Focused Libraries

Deubiquitinases (DUB) play an important role in the ubiquitin pathway. This class of enzymes provides the removal of monoubiquitin and polyubiquitin chains from proteins. Their catalytic activity includes thiol-dependent hydrolysis of ester, thioester, amide, peptide and isopeptide bonds formed by the C-terminal Gly of ubiquitin. Usually a single ubiquitin protein or chains of ubiquitin are added to lysine residues of a substrate protein. These post-translational modifications are added to proteins by the ubiquitination machinery: ubiquitin-activating enzymes (E1), ubiquitin-conjugating enzymes (E2) and ubiquitin ligases (E3).

We focused our attention on the first pair of representatives (USP1/USP2) of the main ubiquitin-specific protease superfamily and UBA5 of E1 class. Our in silico platform was implemented to predict new types of active compounds. All supporting data was taken from ChEMBL DB and RCSB protein data bank. Two different approaches were applied to carry out both structure-based and ligand-based search and design of our new Deubiquitinase Targeted and Focused Libraries.

Docking of the entire Life Chemicals Stock Collection in the active site of E1 activating protein was performed with Glide from Schrödinger Suite. To improve docking results, a set of constraints was defined (Fig. 1). The resulting Deubiquitinase E1 Targeted Library includes 417 stock-available compounds with potential activity towards E1 enzymes.

Pharmacophore modeling was employed for the search of DUB-specific compounds with the aid of both Cresset and Schrödinger modules (Fig. 2). As a result, 1,239 prospective small-molecule inhibitors of ubiquitin-specific proteases (USP1,2) were identified for Life Chemicals Deubiquitinase USP1,2 Focused Library.

K-mean clustering, QSAR and Pharmacophore models used for design of the Deubiquitinase librariy

 

Fig.2. A. K-mean clustering method was implemented to select the best cluster by its total activity (colored 3D diagram is a visualized distribution of 150 clusters). B. QSAR field-based analysis for 3D bioactive structure prediction by conformer search and cross-alignment of reference compounds. C. Pharmacophore-based search of compounds against USP