Binary discontinuous compact protein domains

Few methods exist that identify discontinuous protein domains containing more than one polypeptide chain. This paper describes a new method for locating such discontinuous domains based on their compactness, and applies the methodology to locate the most compact domains in bovine pancreatic trypsin inhibitor, ribonuclease, cytochrome c and myoglobin. The compactness of all binary discontinuous peptide combinations is first exhaustively evaluated. Several screening steps are then used to locate those compact units that represent global minima of compactness. Since domains are generally taken to be large, mutually exclusive structures that span most of the protein's sequence, compact domains were found by examining all compact units (both continuous and discontinuous) to locate two or three units that span most of the protein's sequence, have little mutual overlap and good overall compactness. Compact domains compare well with domains found by other methods and with experimental evidence that may differentiate domain structure. The strongest experimental evidence for the existence of compact discontinuous domains comes from the work of Oas and Kim [(1988) Nature, 336, 42-48] where a peptide that corresponds almost exactly to a compact domain has been synthesized and shown to have native-like structure in solution.