Birth, Death, and Diversification of Mobile Promoters in Prokaryotes

• Published in: Genetics (Austin) Vol. 197; no. 1; pp. 291 - 299
• Main Authors: van Passel, Mark W J, Nijveen, Harm, Wahl, Lindi M
• Format: Journal Article
• Language: English
• Published: United States Genetics Society of America 01.05.2014
• Subjects: branching-process model; Chromosome Mapping; conversion; dna; escherichia-coli; Gene Duplication - genetics; Gene Transfer, Horizontal; Genetic Variation - genetics; Genomes; Genomics; innovation; Interspersed Repetitive Sequences - genetics; Investigations; Models, Genetic; population-genetics; Prokaryotes; Prokaryotic Cells; Promoter Regions, Genetic - genetics; Reservoirs; selection; sequence evolution; transposable elements; singleton; horizontal gene transfer; putative mobile promoters (PMP); transcriptional rewiring
• ISSN: 1943-2631; 0016-6731; 1943-2631
• DOI: 10.1534/genetics.114.162883

A previous study of prokaryotic genomes identified large reservoirs of putative mobile promoters (PMPs), that is, homologous promoter sequences associated with nonhomologous coding sequences. Here we extend this data set to identify the full complement of mobile promoters in sequenced prokaryotic genomes. The expanded search identifies nearly 40,000 PMP sequences, 90% of which occur in noncoding regions of the genome. To gain further insight from this data set, we develop a birth–death–diversification model for mobile genetic elements subject to sequence diversification; applying the model to PMPs we are able to quantify the relative importance of duplication, loss, horizontal gene transfer (HGT), and diversification to the maintenance of the PMP reservoir. The model predicts low rates of HGT relative to the duplication and loss of PMP copies, rapid dynamics of PMP families, and a pool of PMPs that exist as a single copy in a genome at any given time, despite their mobility. We report evidence of these “singletons” at high frequencies in prokaryotic genomes. We also demonstrate that including selection, either for or against PMPs, was not necessary to describe the observed data.