State-specific Peptide Design Targeting G Protein-coupled Receptors

G protein-coupled receptors (GPCRs) play critical roles in various physiological processes and serve as pivotal targets for drug discovery. Peptides are particularly compelling therapeutic agents for targeting GPCRs, as they frequently exhibit superior affinity, selectivity, and potency compared to small molecule drugs. However, the dynamic conformational states of GPCRs profoundly influence their interactions with ligands, highlighting the need for state-specific strategies in peptide design. To address this, we developed an efficient state-specific peptide design framework targeting GPCRs. This approach includes an analysis of GPCR state-transition mechanisms and a specially optimized peptide structure prediction model for GPCRs, HelixFold-Multistate, allowing us to evaluate state-specific scores for the designed peptides. On recent GPCR-peptide PDBs, HelixFold-Multistate not only maintains the GPCR active state, similar to AlphaFold-Multistate, but also demonstrates high interaction prediction capabilities comparable to AlphaFold-Multimer in terms of DockQ and iRMS. Experiments on CXCR4-peptide affinity data indicate that its confidence scores exhibit a higher correlation and screening capability than those of AlphaFold-Multistate. By employing this design method, we successfully identified both the agonist and antagonist peptides for the Growth Hormone Secretagogue Receptor (GHSR) and Apelin Receptor (APJ), exhibiting EC50 values below 100 nM and 10 nM respectively. Despite the challenges with antagonist peptides, our method also identified inhibitors for GHSR and APJ, with respective IC50 values of 3.3 μM and 20.3 μM.