Pairwise contact potentials are unsuitable for protein folding

We demonstrate that pairwise contact potentials alone cannot be used to pre dict the native fold of a protein. Ideally, one would hope that a universal energy function exists, for which the native folds of all proteins are the respective ground states. Here we pose a much more restricted question: Is it possible to find a set of contact parameters for which the energy of th e native contact map of a single protein (crambin) is lower than that of al l possible physically realizable decoy maps? The set of maps we used was de rived by energy minimization (not by threading). We seek such a set of para meters by perceptron learning, a procedure which is guaranteed to find such a set if it exists. We found that it is impossible to fine-tune contact pa rameters that will assign all alternative conformations higher energy than that of the native map. This finding proves that there is no pairwise conta ct potential that can be used to fold any given protein. Inclusion of addit ional energy terms, such as hydrophobic (solvation), hydrogen bond, or mult ibody interactions may help to attain foldability within specific structura l families. (C) 1998 American Institute of Physics. [S0021-9606(98)50247-4] .