First, Do No Harm: Constraint to Be Benign in De Novo Protein-Coding Evolution

Abstract

François Jacob famously referred to evolution as a tinkerer who works with no end goal, cobbling together existing material to make new functions, slowly retouching projects over eons. Jacob claimed that evolution takes a long time to produce complex novelty, such as eyes and brains, and the “probability that a functional protein would appear de novo by random association of amino acids is practically zero.” With the discovery of proteins born de novo from non-coding DNA, Jacob’s claims have come under scrutiny. In this dissertation, I re-evaluate Jacob’s idea of evolution as a tinkerer in the light of de novo evolution. Random peptide studies suggest that random amino acid sequences are expected to produce harmful peptides, in large part due to being aggregation-prone. In contrast, newly born proteins are very unlikely to be aggregation-prone, even more so than old proteins, suggesting that to be beneficial, a peptide must “first, do no harm.” To understand how new proteins that “do no harm” are able to emerge de novo from non-coding regions, I first investigate the properties that make a peptide less likely to be harmful, and then I investigate whether selection is able to act on non-coding regions via translation errors to enrich for those properties. I show that small and disorder-promoting amino acids make a random peptide less likely to be harmful in Escherichia coli. Moreover, the same amino acids that help a random peptide do no harm in bacteria are enriched in young animal Pfam domains, suggesting a deep concordance across the tree of life in the forces that govern gene birth, although the same trend is not found in plants. I also show that evolution is able to act on 3ʹUTRs in Saccharomyces cerevisiae through readthrough errors to make C-terminal extensions do no harm when translated, providing a pool of non-coding sequences that evolution can draw from for de novo evolution. My findings suggest that proteins born de novo from non-coding DNA are the result of an evolutionary process of selective purging. My work expands Jacob’s tinkerer analogy to include both functional and non-functional material. The tinkerer still tinkers continuously, with even greater breadth of material than originally imagined.