Structural Comparison Allowing

We present an e cient method for the comparison of protein structures, allowing swiveling motions. In all currently available methodologies developed and applied to the comparisons of protein structures, the molecules are considered to be rigid objects. The method described here extends and generalizes current approaches to searches for structural similarity between molecules by viewing proteins as objects consisting of rigid parts connected by rotary joints. During the matching, the rigid subparts are allowed to be rotated with respect to each other around pre-de ned swiveling points in one of the molecules. This technique straightforwardly detects structural motifs having hinge(s) between their domains. Whereas other existing methods detect hinge-bent motifs by initially nding the matching rigid parts and subsequently merging these together, our method automatically detects recurring substructures, allowing full 3 dimensional rotations about their swiveling points. Yet, the method is extremely fast, avoiding the time consuming full conformational space search. Comparison of two protein structures, without a prede nition of the motif, takes only seconds to few minutes on a workstation. This algorithm provides a highly e cient, fully automated tool. Its complexity is only O(n2), where n is the number of C atoms in the compared molecules. As in our previous methodologies, the matching is independent of the order of the amino acids in the polypeptide chain. Here we illustrate the performance of this highly powerful tool on a large number of proteins exhibiting hinge-bending domain movements. Despite the motions, known hinge-bent domains/motifs which have been assembled and classi ed, are correctly identi ed. Additional matches are detected as well. This approach has been motivated by a technique for model based recognition of articulated objects originating in computer vision and robotics.