Intrinsic Disorder and the Evolution of Viral Overlapping Genes

Abstract

Overlapping genes result from frameshifts over pre-existing genetic sequences by a process called overprinting, yielding two or more different proteins encoded by the same nucleotides. In this situation, single base changes would frequently alter the amino acids in two or more proteins simultaneously, but such mutations would need to fit two or more sets of structural constraints. One way around such double or multiple constraints would be for overlapping genes to be enriched in structural disorder, which is more tolerant of mutations than is structure. To test this idea, we predicted structure/disorder in the protein products of manually curated overlapping genes from unspliced RNA viruses. Overlapping regions were found to be significantly more disordered than non-overlapping regions and to have a sequence composition biased towards disorder-promoting amino acids. For a subset of the overlapping genes in our dataset, the genetic sequences that were created de novo by overprinting of ancestral genes were determined. Most of the protein products of these novel genes are disordered and have unusual amino acid compositions. Furthermore, almost all of these gene products are accessory proteins rather than replicases or other proteins fundamental to viral replication or structure, and these proteins are orphans without homologues. Proteins that have been created by overprinting different homologues of the same genes display a diversity of functional and structural features, facts that are consistent with their de novo origin. Our results offer a glimpse of the structural and functional characteristics of protein regions encoded by genes created de novo by overprinting events in viruses. In most cases, intrinsically disordered gene products seem to help alleviate both the difficulty of generating structured proteins de novo, and the increased evolutionary constraints expected for multiple-coding genetic sequences.