Design of a Protein Potential Energy Landscape by Parameter Optimization

We propose an automated protocol for designing the energy landscape of aprotein energy function by optimizing its parameters. The parameters areoptimized so that not only the global minimum energy conformation becomesnative-like, but also the conformations distinct from the native structure havehigher energies than those close to the native one. We successfully apply our protocol to the parameter optimization of the UNRESpotential energy, using the training set of betanova, 1fsd, the 36-residuesubdomain of chicken villin headpiece (PDB ID 1vii), and the 10-55 residuefragment of staphylococcal protein A (PDB ID 1bdd). The new protocol of theparameter optimization shows better performance than earlier methods where onlythe difference between the lowest energies of native-like and non-nativeconformations was adjusted without considering various degrees ofnative-likeness of the conformations. We also perform jackknife tests on otherproteins not included in the training set and obtain promising results. Theresults suggest that the parameters we obtained using the training set of thefour proteins are transferable to other proteins to some extent.