Item

Abstract

The proteome of life is mainly formed by proteins belonging to a few thousand domain

families. However, besides these, each genome appears to encode proteins that have little or no sequence similarity to proteins in other genomes, often not even in different strains of the same species. These proteins, variously referred to as singletons or ORFans, may in some cases just correspond to untranscribed open reading frames, but proteomic studies show that many are in fact real proteins. Indeed, there is rapidly growing evidence that new proteins continuously arise in previously non-coding DNA sequence, particularly in

multicellular eukaryotes. So far, none of the ones that have been studied show a tendency to fold into a defined tertiary structure, but at least in prokaryotes, which are not particularly tolerant of unstructured polypeptides in their cytosol and typically have mutation rates that allow them to clear junk efficiently from their genomes, it is not

inconceivable that some are partly or entirely structured. Proceeding from the observation that the simplest mechanism to obtain a longer open reading frame is the amplification of a shorter sequence lacking stop codons and that the dominant mechanism to obtain a folded protein is the repetition of a shorter unstructured peptide, I have been collecting instances of such new repetitive proteins that are likely to fold. Analysis with AlphaFold unsurprisingly now indicates that most have the propensity to form fibrous folds.