A simple model for calculating the kinetics of protein folding from three-dimensional structures

An elementary statistical mechanical model was used to calculate the foldin g rates for 22 proteins from their known three-dimensional structures. In t his model, residues come into contact only after ail of the intervening cha in is in the native conformation. An additional simplifying assumption is t hat native structure grows from localized regions that then fuse to form th e complete native molecule, The free energy function for this model contain s just two contributions-conformational entropy of the backbone and the ene rgy of the inter-residue contacts. The matrix of interresidue interactions is obtained from the atomic coordinates of the three-dimensional structure. For the 18 proteins that exhibit two-state equilibrium and kinetic behavio r, profiles of the free energy versus the number of native peptide bonds sh ow two deep minima, corresponding to the native and denatured states, For f our proteins known to exhibit intermediates in folding, the free energy pro files show additional deep minima. The calculated rates of folding the two- state proteins, obtained by solving a diffusion equation for motion on the free energy profiles, reproduce the experimentally determined values surpri singly well, The success of these calculations suggests that folding speed is largely determined by the distribution and strength of contacts in the n ative structure. We also calculated the effect of mutations on the folding kinetics of chymotrypsin inhibitor 2, the most intensively studied two-stat e protein, with some success.