Computational Design of Peptide Inhibitors for Class B G Protein-coupled Receptors

Project: Research project

Project Details

Description

Computational Design of Peptide Inhibitors for Class B G Protein-coupled Receptors Computational Design of Peptide Inhibitors for Class B G Protein-coupled Receptors G protein-coupled receptors (GPCRs) are integral membrane proteins that transfer signals across cell membranes. They are arguably one of the most pharmaceutically important protein families and are the targets of a large fraction (-30%) of all current drugs. The so-called class B GPCRs are a subgroup that bind peptide hormones via a large extracellular domain (ECD). Class B GPCRs are implicated in many human disease including diabetes, osteoporosis, neurodegeneration, inflammation, dwarfism, and psychiatric disorders. We propose to develop new computational methods to design peptide inhibitors of class B GPCRs and to validate the methodology by experimentally testing designed inhibitors of the prototypical class B GPCR, the secretin receptor. Because of the enormous number of potential amino acid sequences of the peptide inhibitor, a two-stage computational design approach will be employed. First, a fast but approximate residue pair empirical potential will be combined with Loopy Belief Propagation optimization to find a limited set of peptides that are predicted to bind strongly to the ECD. The structure of the ECD for the human secretin receptor will be generated using comparative modeling based on the experimental structure of the PACAP receptor. Second, the structure of these peptides bound to the ECD will be predicted using all-atom molecular mechanics followed by scoring using physical energy components. The ICM program, which uses an efficient Monte Carlo algorithm to minimize a physical energy function will be used to optimize the structure of the complex by sampling the conformations of all peptide residues and contacting ECD residues. A harmonic restraint potential will be used to maintain the peptide backbone near that in the template structure. The differences of the energy components between the bound and unbound components will then be calculated and used as input to a Random Forest machine learning classifier to predict which peptides bind strongly to the ECD. A set of peptides predicted to bind strongly to the secretin receptor ECD will then be experimentally tested using competition binding assays. The results will be used to validate the computational method and form the basis for an NIH grant proposal incorporating these methods.
StatusFinished
Effective start/end date1/1/0112/31/10

Funding

  • Mayo Clinic Arizona: $17,021.00

Fingerprint

Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.