AlphaFold2-Based Method Designs Cyclic Peptides That Stabilize Protein-Protein Interactions | Research

The takeaway

A computational protocol combining AlphaFold2 with MD simulations rapidly designs cyclic peptides that simultaneously bind two protein partners, enabling PPI stabilization and targeted protein degradation.

AI designs dual binders
AlphaFold2-guided cyclic peptides simultaneously optimize binding to both partners of a protein-protein complex

What the researchers found

AlphaFold2-based cyclic peptide design achieved similar or better calculated interaction scores than known PPI binders, with well-balanced dual binding and applicability to bifunctional protein degradation targeting.

Why it matters

PPIs are involved in most diseases but are "undruggable" by small molecules. AI-designed cyclic peptides could unlock these targets for the first time.

How the study worked

Modified AlphaFold2 peptide design with confidence + force field scoring, MD simulations for dual-binding optimization, validation on known PPI systems, and application to bifunctional degradation targeting.

What this study cannot tell us

Computational design only; no experimental synthesis or validation yet. AlphaFold2 predictions have known limitations for peptide-protein interactions.

How to read the evidence

Computational method development with validation on known systems. Strong in silico performance needing experimental confirmation.

When this study was published

Published in 2025.

The bigger picture

The combination of AI protein structure prediction with peptide design could crack the "undruggable" PPI problem, one of the biggest challenges in drug discovery.

Questions still open

Will the designed cyclic peptides maintain predicted binding in experimental testing?
How does the design accuracy compare across different PPI types?
Can this pipeline be automated for high-throughput PPI drug discovery?

Common questions

What are cyclic peptides and why are they useful?

Cyclic peptides are ring-shaped molecules larger than typical drugs but smaller than antibodies. Their size allows them to interact with the large surfaces where proteins contact each other — interactions that small drugs cannot disrupt.

How does AlphaFold2 help design drugs?

AlphaFold2 predicts protein structures with remarkable accuracy. This study adapts it to design cyclic peptides that fit precisely between two interacting proteins, like designing a key that fits two locks simultaneously.

Read the original research

AlphaFold2-Guided Cyclic Peptide Stabilizer Design to Target Protein-Protein Interactions.

Proteins

View on PubMed ↗Publisher’s paper ↗

Citation

Halbwedl, Niklas; Zacharias, Martin. (2026). AlphaFold2-Guided Cyclic Peptide Stabilizer Design to Target Protein-Protein Interactions.. Proteins. https://doi.org/10.1002/prot.70123