Common structural folds in several protein pairs searched by an iterative superposition algorithm

The iterative superposition algorithm described by Rossmann and Argos (J. Biol. Chem. 1975; 250: 7525) for comparing protein structures has been used to search some common structural folds for several protein pairs. Each protein sequence is divided into numerous overlapping blocks of residues and the initial seed residues are chosen randomly from residues that are assigned within each block. The number of blocks and the number of residues selected for each block are both dependent on the number of maximum residues allowed for each block set beforehand. Residues in each block generated for each protein are treated as a series of seed residues and then used to align with those generated for each block of the second protein. A weighting scheme is employed in the refinement step for finding the best set of aligned coordinates. By using a variable threshold for selection of the C^x-C^x distances calculated between each residue pair and by searching the largest sequential residue pairs that are matched within a selected tolerance, we find that the method is capable of identifying common structural folds for either smaller proteins for which higher degree of structural similarity is present in between, or for larger proteins that are far more dissimilar in structure to each other. The protein pairs aligned are not confined to a particular structural class. Therefore, these results indicate that it is feasible to use the iterative superposition method to perform more extensive searches for common structural folds between proteins.