Improved approaches to ligand growing through fragment docking and fragment-based library design

Die Fragment-basierte Wirkstoffforschung (“fragment-based drug discovery“ – FBDD) hat in den vergangenen zwei Jahrzehnten kontinuierlich an Beliebtheit gewonnen und sich zu einem dominanten Instrument der Erforschung neuer chemischer Moleküle als potentielle bioaktive Modulatoren entwickelt. FBDD is...

Full description

Saved in:
Bibliographic Details
Main Author: Chevillard, Florent
Contributors: Kolb, Peter (Dr.) (Thesis advisor)
Format: Doctoral Thesis
Published: Philipps-Universität Marburg 2016
Online Access:PDF Full Text
Tags: Add Tag
No Tags, Be the first to tag this record!

In the past two decades, fragment-based drug discovery (FBDD) has continuously gained popularity in drug discovery efforts and has become a dominant tool in order to explore novel chemical entities that might act as bioactive modulators. FBDD is intimately connected to fragment extension approaches, such as growing, merging or linking. These approaches can be accelerated using computational programs or semi-automated workflows for textit{de novo} design. Although computers allow for the facile generation of millions of suggestions, this often comes at a price: uncertain synthetic feasibility of the generated compounds, potentially leading to a dead end in an optimization process. In this manuscript we developed two computational tools which could support the FBDD elaboration cycle: PINGUI and SCUBIDOO. PINGUI is a semi-automated workflow for fragments growing guided by both the protein structure and synthetic feasibility. SCUBIDOO is a freely accessible database which currently holds 21 M virtual products. This database was created by combining commercially available building blocks with robust organic reactions. Thus, every virtual product comes with synthetic instructions. Most of the crucial functions of PINGUI (creation of derived libraries or applying organic reaction) were then implemented in the SCUBIDOO website. PINGUI and SCUBIDOO were then applied to fragment-based ligand discovery efforts targeting the $beta_{2}$-adrenergic receptor ($beta_{2}$AR) and the PIM1 kinase. In a first study focusing on the $beta_{2}$AR, we suggested a total of eight diverse extensions for different fragment hits using PINGUI. The eight compounds were successfully synthesized and further assays showed that four products had an improved affinity compared to their respective initial fragment. In a second study, SCUBIDOO was applied in order to quickly identify fragments and suggest extensions that could bind to PIM1. This study yielded a fragment hit and its associated crystal structure. Synthesis of derived products is in progress. Lastly, SCUBIDOO was coupled with automated robotic synthesis in order to synthesize hundreds of compounds in parallel. 127 products among the 240 suggested were synthesized (53%). Those compounds were designed so they are likely to bind to the $beta_{2}$AR and will be tested in the near future. The aforementioned computational tools could improve early fragment-based drug discovery projects, especially in the realm of fragment growing strategies. For instance, PINGUI suggests extensions that are very likely to be attachable, making it a useful creative tool for medicinal chemists during structure-activity relationship (SAR) studies. With so far 53% success synthesis rate, SCUBIDOO has shown that it is amenable to be integrated to automated robotic synthesis. Every synthesis attempt is prone to improve the knowledge contained within the database and thus increase the synthesis success rate over time. Furthermore, all synthesized product were novel compounds, thus demonstrating how SCUBIDOO could explore new quadrants of the chemical space.​