Stabilizing protein-protein interface

Background

A complex of two proteins with unknown binding sites.
A binding pocket was identified by Receptor.AI at the protein-protein interface using a proprietary AI model for the binding site identification.
Two low-quality ligands are known with only mediocre binding affinity.
The goal is to design “molecular glue” to stabilize the membrane complex of 2 proteins associated with autophagy and neurodegeneration.

Enamine stock collection of 3.2M compounds was subject to virtual screening.
1000 top-ranked candidate compounds were selected.
387 compounds were used for experimental validation.

The experimental validation was conducted on the HEK 293 cell line to assess modulation of target protein expression.
Protein expression level was assessed by the Western blot. The hit criteria for the screening was set as a 120% increase in target protein concentration against the control, which resulted in the identification of seven hits.
The activity of all seven identified hit compounds, along with the absence of cytotoxicity, was substantiated through dose-response analysis.
Four active scaffolds were chosen for future series expansion, specifically the scaffolds of compound 1, 3, 4 and the shared scaffolds of compounds 2, 5-7.
The most promising hit compound 2 exhibited a nearly sub-micromolar EC₅₀ in a cellular functional assay with a 400% higher potency compared to the best existing rivals.

The binding pose for the Compound 2 mentioned above

Western blot, Compound 2, HEK 293 cell line, 24h after treatment

Target protein concentration increase: assay Round 1, 50 uM

Dose-Response relationship for Compound 2. Treatment time 24 and 48 hours