Examination of folding in random sequence proteins

Biological proteins are known to fold into specific three-dimensional conformations. A prevailing view among protein scientists has been that proteins fold because evolution has acted to select folding sequences from the vast ensemble of possible sequences, most of which, it was presumed, were incapable of folding. An alternative explanation of the existence of protein folding is that it results from inherent, self-organizing properties of linear, amino acid polymers. We have tested the second hypothesis by searching for structural characteristics commonly found in folded, biological proteins within samples of random-sequence polypeptides. Libraries of arbitrary-sequence polypeptides, expressed as cleavable fusions with ubiquitin, were constructed for use as a random sample of protein sequence space. The novel polypeptides, with average amino acid composition similar to the mean for natural proteins, were analyzed spectroscopically. Heat-labile secondary structure was demonstrated in the random proteins by circular dichroism experiments. Using fluorescence emission spectroscopy, fusions and a cleaved protein were shown to exist in denaturant-sensitive, collapsed conformations with a range of unfolding cooperativities. The collapsed state of one fusion was shown to interact with a fluorescent dye in a manner similar to that observed with molten globule forms of natural proteins, while another fusion demonstrated behavior resembling that of native protein conformations. At least three common structural properties found in evolved proteins, namely, collapsed conformation, secondary structure, and cooperative unfolding have been shown to be prevalent in stochastic polypeptides. It appears the ability to fold is a more common characteristic of polypeptides, unbiased by evolution, than previously imagined. These results impact theories of protein folding, protein evolution, and the origin of life