In silico discovery of nanobody binders to a G-protein coupled receptor using AlphaFold-Multimer

Area of Science:

• Immunology
• Computational Biology
• Structural Biology

Background:

• Antibodies are crucial for adaptive immunity, research, and therapeutics.
• Traditional antibody discovery involves extensive experimental work like immunization or library screening.
• Predicting antibody-antigen interactions computationally remains a significant challenge.

Purpose of the Study:

• To explore the potential of in silico antibody discovery using machine learning.
• To prospectively identify nanobody binders for the therapeutic target MRGPRX2.
• To validate a computational pipeline for antibody discovery that reduces experimental effort.

Main Methods:

• Utilized AlphaFold-Multimer (AF-M) for virtual screening of nanobody-GPCR interactions.
• Developed criteria from existing nanobody-GPCR structures to assess AF-M predictions.
• Performed a prospective in silico screen to identify high-affinity nanobody binders to MRGPRX2.

Main Results:

• Successfully identified nanobodies predicted to bind MRGPRX2 with high affinity.
• Confirmed the activity of identified nanobodies in cellular signaling and functional assays.
• Demonstrated the efficacy of AF-M in discriminating between interacting and non-interacting nanobody-GPCR pairs.

Conclusions:

• Provided a proof of concept for fully computational antibody discovery.
• Showcased the ability of AlphaFold-Multimer to facilitate in silico antibody discovery pipelines.
• Highlighted the potential to circumvent traditional laboratory experiments in antibody discovery.