Conformational energy calculations are used to analyze the interactions of structural substructures in subtilisin BPN. These substructures are kept fixed or-"rigid" so that the only variables in the calculations are the backbone segments that separate them. The flexible segments are assumed to be free turns. Using this representation of the protein it is possible to predict both a likely order of events along a folding pathway and preferred modes of conformational changes of the native protein. Moreover, when the native structure has been perturbed by moving the substructures apart, it is possible to assess the range of interactions that return the protein, upon energy minimization, to its original conformation. These results suggest an approach to the folding problem based on the piecemeal formation of tertiary structure from smaller prefolded fragments.The information necessary to determine the three-dimensional structure of a folded protein is contained in its sequence of amino acids (1). Thus, it should be possible in principle to predict the geometry of a protein from its primary structure. However, it does not seem possible that the final conformation of a protein is established by selecting the lowest energy state from among all possible conformations. The time required for random motion to produce all possible states would be many orders of magnitude longer than the observed folding time (2). In addition, a search of all possible states would presumably result in the structure's being trapped in some "incorrect" local minimum (unless biological evolution specifically selected against this possibility). Thus, an elucidation of the folding pathway is an essential step in discovering the relationship between the primary and tertiary structure of a protein.We assume that proteins fold by following a multiply. branched pathway in which the first stage is the formation of local secondary structure governed by interactions between amino acids that are near each other in the peptide chain. Subsequently, these structures, such as a-helices and antiparallel fl-strands, would interact, perhaps being modified in the process, to produce larger structural fragments which then undergo further assembly to yield the native conformation, which we assume to correspond to the crystal conformation. In fact, there is considerable evidence for the existence of "independent" structural fragments of tertiary structure. For example, Rao and Rossman (3) have noted similar patterns of secondary structure organization in a number of proteins. Wetlaufer (4) has emphasized the existence of large globular domains that have been noticed in many proteins although it has not yet been possible to establish unique criteria for their identification. However, in a number of cases (e.g., immunoglobulins, serine proteases) the proteins are clearly formed of distinct domains.The existence of independent structural regions of various sizes provides an enormous simplification for computational efforts to predict structure from sequence. I...

No abstract

No abstract

No abstract

No abstract