Drug repurposing is a promising field in drug discovery that identifies new therapeutic opportunities for existing drugs (e.g. Metformin and Raloxifene). High-content screens, new biomarkers, and noninvasive imaging techniques have created new capabilities for pursuing novel indications for approved compounds. Hits from our FDA Approved Drug Collection will provide a significant head start in any drug optimization program.

• Carefully collected and distilled collection of 1 123 FDA approved drugs
• Supported with full bioactivity annotation, pathway indications and all related chemical structure information (structure, CAS, smiles, molecular parameters)
• NMR and HPLC validated to ensure the highest purity
• Supplied in 96, 384 or 1536-well plates, ready for immediate shipping
• Minimal preparation – just peel, dilute & transfer to assay plate

To date no commercially available compound collection covered all drug classes, resulting in complications in combinatorial screening attempts. We collaborated with a group of researchers at CeMM and collected a condensed screening set of 299 small molecules representing the entire target and chemical space of all FDA-approved drugs. The CeMM Library of Unique Drugs (CLOUD) covers prodrugs and active forms at pharmacologically relevant concentrations and is ideally suited for combinatorial studies. Recent publication [1] provides evidences of successful drug repurposing by pairwise compound screening from CLOUD collection in cancer cell viability assay. The findings highlight that refined chemical library CLOUD is a powerful tool for drug repurposing and evaluation of new substances suitable for all high-content and high-throughput assays.

Key features

• Covers all therapeutically significant space of approved drugs.
• Contains prodrugs and active metabolites at pharmacologically relevant concentrations.
• Ideally applicable for combinatorial assays to study new drug combinations.
• Most structurally diverse drug library among commercially available collections.
• Immediately accessible in convenient pre-plated formats with carefully prepared detailed documentation. The library can be also made in any customized ready-to-screen formats.

Examples of drugs in the library

Selection and library evolution scheme

[1] Nature Chemical Biology 13, 771–778 (2017).

Fragment screening usually provides high numbers of hits and is one of the most affordable approaches to start drug discovery. However, subsequent steps involving hit-to-lead optimization may be challenging and require significant resources. It’s even more painful when efforts result in false positives.

We have designed a new fragment library specifically addressing the quality of decision making when evaluating hits. Our High Fidelity Fragment Library features high medchem tractability enabling researchers to grow interesting hits with confidence.

High medchem tractability of this set was achieved through structure review and selection by FBDD experts at Takeda and Carmot Therapeutics. In particular, we would like to thank Dr. Derek Cole, Dr. Dan Erlanson, Dr. David Lawson and Dr. Xiaolun Wang for their involvement in the design of our High Fidelity Fragment Library.

High quality: All compounds selected for this library passed turbidity tests to assure high solubility in water at 1 mM; all aggregators were filtered out. In addition, the fragments in this set were screened by surface plasmon resonance (SPR) to remove any false positive fragments. SPR screening was kindly provided by Dr. Delphine Collin at HarkerBio.

Optimal molecular properties: Fragments in this set all have 9–16 heavy atoms, are moderately complex and have suitable physiochemical and shape profiles.

Our High Fidelity Fragment Library is available for prompt delivery in various formats. The most popular library options are listed below:

Library design

A complex iterative approach has been applied in the design of this library. Fragment selection was made from over 260k Ro3-compliant compounds in Enamine’s stock collection. All industry recommended medchem filters, including PAINS, were applied. Diversity selection was performed using clustering algorithm and manual review of identified clusters to select the most representative structure within each cluster.

All compounds were tested for solubility in water and aggregation by laser nephelometry to remove compounds with solubility issues. The resulting set of fragments was then tested in clean SPR screens to exclude SPR “sticky” compounds.

The selection process and resulting molecular parameters of our High Fidelity Fragment Set are described in the following scheme:

The library was designed to be a universal tool to search potential JAK-STAT pathway modulators. Protein structure-based analysis and scaffold-hopping approach were used to create an optimal in silico screening models. Additionally, ligand-based approach has been applied to enrich the library with topological analogs and similar compounds to reported actives.

Resulted library is as a generic starting point for ligand search, regardless of particular SH2 domain of interest, with high probability of initial hit discovery. The library was validated with in vitro screening against SRC SH2 domain resulting in 3.5 % of identified hits.

Library design

All available structures of SH2 domain binding sites were analyzed and clustered based on their spatial molecular shape. After the alignment analysis and clustering structures with most different conformations were selected for virtual screening. Enamine MedChem filtered in-stock subset (~1 M compounds) with additional selection of compounds carrying peptidomimetic motifs was used for molecular docking calculations.

Multiple sequence alignment was applied to 120 SH2 domains contained within 110 proteins. Over 200 protein structures were extracted from Protein Data Bank (PDB): 66 NMR-based structures and 153 derived from X-ray crystallography experiments. As variation of the binding site conformation may significantly influence its binding properties, all files were split into individual structures, resulting in total structure count of 1633.

Key pharmacophore interaction points: pTyr binding pocket, carbonyl O in the binding site center, hydrophobic sub-pocket.

Solvent accessible molecular surface within 12 Å from the binding site center was used for calculation of shape-based numeric descriptors. 3D structures were clustered based on 3D shape similarity. 8 spatially diverse structures were selected for docking.

Virtual screening against Stat3beta:

• Collaborator: Gyeong Baeg, NYMC

Chemical compounds are significantly smaller in size than the natural peptide interacting with this target. Therefore, when creating the library, we tried to fill all potential sub-pockets available near the phosphotyrosine binding site. The entire available surface of the protein was conditionally broken down into 5 models and proceeding from this was carried out post-docking analysis. As an example, the first two models:

STAT3 models: Schematic representation of ligand binding pocket