User:Pakpoom Subsoontorn/Notebook/general reading/2008/10/22

From OpenWetWare

Jump to: navigation, search
Project name Main project page
Previous entry      Next entry

Computational design of receptor and sensor proteins with novel functions(BioE300a presentation)

Outline

  • Challenges in designing protein -1
    • Complexity of protein-ligand interaction, too many possibility
    • Two approaches
      • Rational Design
      • Directed Evolution
  • What's new here?
    • Specific goals: changes the specificity of receptors/sensor protein
      • Precursors/ligand (from E.coli PBP): GBP, RBP, ABP, QBP, HBP
      • Targeted ligands: TNT, Serotonin, L-lactate
    • Designing Algorithms
      • Identify amino acid sequences that can form a complementary surface between the protein and the target ligand
      • Dead-end elimination theorem (DEE)
        • Deterministically identifies the global minimum of a semi-empirical potential function of molecular interactions
        • Modified Lennard-Jones potential
        • rapid computation
    • Building new proteins
      • PCR mutagenesis
      • Design outcomes
    • Testing Results:
      • QSAR
      • binding affinities/selectivities
      • Rewiring Pathways
  • Summary/Critique
  • More recent papers

Design Algorithm

  1. Define PCS that ford van de Waals interaction with the wild-type ligand, replace with ala
  2. Place a cubic grid for docking ligands with in a convex hull encompassing with wild-type ligand
  3. Construct an ensemble representing all rotational and internal degrees of freedom of ligand
  4. Place the rotational ensemble onto the grid and select pose that do not form unfavorable steric interactions and are partially confined within the convex hull. This will be the "doced ligand conformations" (Parallelized)
  5. Rank the docked ensemble by the interaction energy between the ligands and the receptor
  6. Calculate PCS of the top N (~10,000) docked ligands, using DEE (Parallelized)
    • Place side-chain rotamer library representing all possible mutation and side-chain conformation (except cys, pro) at all position in the PCS
    • Identify optimal sequence by determining the global minimum energy functions of interaction
    • Potential function is based on semi-empirical force field, a modified Lennard-Jones potential representing van der Waals interaction. Satisfy potential hydrogen bonds donors and acceptors in the ligand
    • As additional filter in DEE: should all non-eliminated hydrogen bond partners for a demanded atom arise from a single residue position, all non-hydrogen-bonding side-chains at this position are eliminated
    • Keep protein backbone fix
  7. Rank predicted design based on should all non-eliminated hydrogen bond partners for a demanded atom arise from a single residue position, all non-hydrogen-bonding side-chains at this position are eliminated.
  8. a small number of the top-ranked designs are experimentally tested.

DEE for protein foldings


Personal tools