Function diversity within folds and superfamilies

Abstract

The structural genomics initiatives significantly increased the numbers of three-dimensional structures available for proteins of unknown function. However, the extent to which structural information helps understanding function is still a matter of debate. Here, the value of detecting structural relationships at different levels (typically, fold and superfamily) for transferring functional annotations between proteins is reviewed. First, function diversity of proteins sharing the same fold is investigated, and it is shown that although the identification of a fold can in some cases provide clues on functional properties, the diversity of functions within a fold can be such that this information is very limited for some particularly diverse folds (e.g. super-folds). Next, since structural data can help detecting homology in the absence of sequence similarity, function diversity between proteins from the same superfamily (homologous proteins) is analysed. The evolutionary causes and the mechanisms that have generated the observed functional diversity between related proteins are discussed, and helpful tools for the correlated analysis of structure, function and evolution are reviewed.