INVERSE SEQUENCE SIMILARITY IN PROTEINS AND ITS RELATION TO THE 3-DIMENSIONAL FOLD

Nowadays the most successful strategy for the prediction of the tertia ry structure of proteins is the homology-based modelling using known s tructures. A real chance to predict the general fold of a protein aris es only in cases with a sufficient sequence homology (e.g. 27% over 10 0 residues). In this analysis we examine the phenomenon of inverse seq uence similarity (ISS) in proteins and its structural meaning. In sequ ence data bases me found a lot of examples for ISS up to 34% identity over 204 residues and a surprisingly large number of self-inverse prot ein sequences. By inspection of inverse similar sequence pairs with kn own tertiary structures we observe that inverse sequence alignments ab ove the threshold indicating structural similarity generally do not im ply comparable folds for both. From our analysis we conclude that the straightforward employment of ISS for protein structure prediction fai ls even above the known threshold for 'safe similarity'. (C) 1997 Fede ration of European Biochemical Societies.