A STRATEGY FOR DETECTING THE CONSERVATION OF FOLDING-NUCLEUS RESIDUESIN PROTEIN SUPERFAMILIES

Background: Nucleation-growth theory predicts that fast-folding peptid e sequences fold to their native structure via structures in a transit ion-state ensemble that share a small number of native contacts (the f olding nucleus). Experimental and theoretical studies of proteins sugg est that residues participating in folding nuclei are conserved among homologs. We attempted to determine if this is true in proteins with h ighly diverged sequences but identical folds (superfamilies). Results: We describe a strategy based on comparisons of residue conservation i n natural superfamily sequences with simulated sequences (generated wi th a Monte-Carlo sequence design strategy) for the same proteins. The basic assumptions of the strategy were that natural sequences will con serve residues needed for folding and stability plus function, the sim ulated sequences contain no functional conservation, and nucleus resid ues make native contacts with each other. Based on these assumptions, we identified seven potential nucleus residues in ubiquitin superfamil y members, Non-nucleus conserved residues were also identified; these are proposed to be involved in stabilizing native interactions, We fou nd that all superfamily members conserved the same potential nucleus r esidue positions, except those for which the structural topology is si gnificantly different, Conclusions: Our results suggest that the conse rvation of the nucleus of a specific fold can be predicted by comparin g designed simulated sequences with natural highly diverged sequences that fold to the same structure, We suggest that such a strategy could be used to help plan protein folding and design experiments, to ident ify new superfamily members, and to subdivide superfamilies further in to classes having a similar folding mechanism.