Discovery of PRMT5 N-Terminal TIM Barrel Ligands from Machine-Learning-Based Virtual Screening

Area of Science:

• Oncology
• Biochemistry
• Computational Biology

Background:

• Protein arginine methyltransferase 5 (PRMT5) is a key cancer therapeutic target due to its role in symmetric dimethylation of proteins.
• Current PRMT5 inhibitors primarily target the catalytic domain, but inhibiting the PRMT5/pICLn protein-protein interface is also crucial.
• Developing small molecules to disrupt protein-protein interactions presents a significant therapeutic opportunity.

Purpose of the Study:

• To develop and apply a machine learning-based virtual screening method to identify inhibitors of the PRMT5/pICLn protein-protein interaction.
• To screen for novel small molecules that can disrupt the PRMT5/pICLn complex.

Main Methods:

• A machine learning-based virtual screening pipeline was designed and utilized.
• 18 compounds were selected from the virtual screen for experimental validation.
• Surface plasmon resonance (SPR) was used to assess binding affinity.

Main Results:

• One compound, Z319334062, demonstrated binding affinity to the PRMT5/pICLn target with a K_D of 21.5 μM.
• Z319334062 exhibited dose-dependent inhibition of symmetric dimethylation levels.
• The compound's efficacy was observed in patient-derived glioblastoma cell lines.

Conclusions:

• The developed machine learning pipeline is effective for identifying inhibitors of the PRMT5/pICLn interaction.
• Compound Z319334062 represents a promising lead for developing novel PRMT5-targeted cancer therapies.
• Disrupting the PRMT5/pICLn interface is a viable strategy for cancer treatment.