Nature’s Shortcut to Protein Folding

Major advances in computational biology enable prediction of protein structures based on given amino-acyl sequences. This is based on simulation of the stochastic search for vast number of interactions in the system that includes the polypeptide and the solvent. However, this outstanding achievement is still limited to solution of the protein folding problem of polypeptides shorter that ca. 100 aminoacyl residues. What are the physical principles that enable the ultrafast folding of long polypeptides? We hypothesize that nature’s shortcut to fast folding of most proteins is fast formation of few non-local contacts between clusters of aminoacyl residues that are separated by segments of tens of residues at the initiation of the folding pathway (the “loop hypothesis”). Such interactions form few closed long loops and thus reduce the size of the search. By this mechanism stochastic search of small units is achievable.

In order to test this hypothesis we use time resolved FRET combined with site specific labeling and microfluidic based fast mixing in order to map fast formed contacts in unfolded polypeptide at the initiation of its folding pathway.

Application of this method testing the folding of two model proteins shows that indeed few long loops are formed within few microseconds after the initiation of refolding, while the rate limiting step for the folding is much slower. These experiments support the “loop hypothesis”. We suggest that implementation of this principle in future search algorithm might overcome the current size barrier of computational structure prediction.