Folding units in globular proteins.

We present a method to identify all compact, contiguous-chain, structural units in a globular protein from x-ray coordinates. These units are then used to describe a complete set of hierarchic folding pathways for the molecule. Our analysis shows that the larger units are combinations of smaller units, giving rise to a structural hierarchy ranging from the whole protein monomer through supersecondary structures down to individual helices and strands. It turns out that there is more than one way to assemble the protein by self-association of its compact units. However, the number of possible pathways is small--small enough to be exhaustively explored by a computer program. The hierarchic organization of compact units in protein molecules is consistent with a model for folding by hierarchic condensation. In this model, neighboring hydrophobic chain sites interact to form folding clusters, with further stepwise cluster association giving rise to a population of folding intermediates.