Enamine - smart chemistry solutions

Protein-Protein Interaction Libraries

Download General Library SDF
Download Bromodomain Targeted Set SDF
Download PDZ-domains target Set SDF

Targeting protein-protein interactions (PPIs) is a perfect way to avoid numerous side effects that are often present when substrate binding site is blocked. Various examples of drug candidates that were withdrawn from clinical trials because of unexpected side effects of direct antagonist have proved importance of development of new protein-protein interactions inhibitors.

Importantly, our team of experienced computational, synthetic, medicinal chemists and biologists is ready to address your specific needs in tackling protein-protein interaction of your interest. Please, challenge us with your biological concepts, computational ideas and synthetic designs.


General Library

28 000 compounds

Systemic analysis of published data on thermodynamic, structural and pharmacological aspects of numerous PPIs spanning pro- and eukaryotic proteomes allowed us to develop dedicated approach to the library design. We have analyzed more than 20 different protein-protein complexes to highlight specific features of majority of potent inhibitors in this area. Several specific recognition patterns that received attention are exemplified by α-helix, β-sheet, PDZ-, Armadillo/HEAT-, Polo-box- domains. Numerous bacterial-, parasitic- and viral PPIs have been modeled as well. As a result of Ligand- and Structure based analyses we came up with the selection of compounds that are:

  • likely to target specific recognition patterns including key amino acids, amino acid sequences, secondary/tertiary structures, domains, stable and transient folds, α-helices and ‘hot loops’;
  • possess drug/lead-like properties (MedChem filters and molecular specified parameters);
  • synthetically feasible, yet structurally unique.

Small molecule mimetics of key recognition amino acids, peptides reported to form ‘hot loops’ and ‘hot sequences’, spiro-heterocycles, bi-/tri-cyclic turn and hairpin mimetics, medium-size rings exhibiting high Fsp3character are components that form the core of this pharmacologically relevant selection of over 28 000 shelf-available compounds.

Examples of chemotypes populating our PPI modulators library

Examples of chemotypes populating our PPI modulators set 


Enamine Bromodomain Targeted Set

2 700 compounds

Targeting bromodomains is of considerable interest and important task of medicinal chemistry because of the proved importance of these protein modules as therapeutic targets [1, 2]. The object of bromodomains is “reading” imply in their capability to recognize and bind to acetyllysine moieties of histones that play a key role in the epigenetic regulation of gene transcription [3]. Compounds acting as anticancer, anti-inflammotary and antiviral agents were identified among effective bromodomains inhibitors [2, 3].

Library Design

Enamine MedChem stock subset, Ro5 compliant and filtered through carefully combined structural filters, was used for subsequent selection of compounds. As far as bromodomains form interactions with acetylated lysine residues number of scaffolds mimic to this motive such as oxazoles, oxadiazoles and other related were picked out as privileged moieties (Fig. 1). Further, algorithm of ligand-based approach accounting known chemotypes and known pharmacophores modulating bromodomain functions [2] was implemented in compound selection procedures.

Figure 1. Examples of chemotypes and structural features used in Library design.

To create and successful validate in silico screening models careful analysis of over 50 reported bromodomain-ligand complexes was carried out (reported in PDB). Basing on the results of binding sites alignment and analysis of key residues positions we have defined four different groups of bromodomain-ligand interaction. This exploration allowed us for separation of four distinct classes of active ligands depending on their chemical structure and respectively types of interaction with bromodomains:

  • methylisoxazoles,
  • benzodiazepines and their analogues,
  • bis-N-phenylpyrimidine diamines,
  • “highly aromatic” flat ligands.

For each corresponding group 3D pharmacophore models were built, validated with reference compounds set and then used to search of new perspective bromodomain ligands (Fig. 2).

a) bis-N-phenylpyrimidinediamines, pink - HBA, orange - aromatic features, blue - HBD

b) methylisoxazoles, pink - Nitrogen in cycle, blue - methyl group, orange - aromatic cycle;

c) benzodiazepines, pink - HBA, blue - Nitrogen in cycle, orange - aromatic cycle;

d) ) “high aromatic” ligands, pink - HBA, orange - aromatic cycle.

Figure 2. Ligand superpositions and proposed pahrmacophore models for each group of ligands.

Molecular profiling

Set of selected compounds has attractive drug-like molecular properties, compliant with Lipinski’s Rules, 83.52 % correspond to lead-like criteria.

Parameters Range
MW ≤ 450
logD at pH7.4 -4 … 4
Hb Donors ≤ 5
Hb Acceptors ≤ 8
Ring count ≤ 4
RotBonds ≤ 10


Enamine PDZ-domains target Set

2 300 compounds

PDZ domain-mediated interactions are of significant interest over the last decades and are indicated as key to cell life in drug development. Many of these domains are involved in the modulation of numerous signaling pathways which play an important role in fundamental cellular function. Versatile therapeutic potential of those targets includes neurology, cancer and immunology as well as other pathway event coordination. Furthermore, PDZ domains are convenient targets for drug discovery intention due to highly conserved binding site.

Library Design

Since structural data for the most of PDZ-domains are well-established receptor-based approach was the primary method for library formation. The docking calculations were run on a large sets of PDZ-domains including PDS-95, DVL1, DVL3, Shank3, AF6, Erbin, etc. Meanwhile different reported conformation of the proteins’ binging sites have been taken in account and simultaneously compared to shapes of other proteins in the family (recorded in PDB). General analysis of geometric shapes of protein binding pockets enabled us to construct shape descriptors using calculation and comparison of protein solvent accessible surfaces.

Over 2 800 conformations from 163 PDZ domains were used to compare of their binding sites. Six general types PDZ domains (Fig.3) were selected as representatives to be in silico screened against Enamine MedChem compound collection. After processing of docking results the library was enriched with compounds bearing bioisosteric chemotypes to known highly active inhibitors.  

Figure 3. Space of shape descriptors: clustering and representatives

Schematic representation of the results obtained after sequential comparison of geometric shapes of PDZs binding pockets. Examples of docking results and protein binding pocket surface are shown below.

Figure 4. Examples of selected chemotypes and privileged moieties.

© 1991 - 2018 ENAMINE
Terms of Use | Privacy Policy | Cookies Policy