Escherichia coli adaptation under prolonged resource exhaustion is characterized by extreme parallelism and frequent historical contingency

• Published in: PLoS genetics Vol. 20; no. 6; p. e1011333
• Main Authors: Zion, Shira, Katz, Sophia, Hershberg, Ruth
• Format: Journal Article
• Language: English
• Published: United States Public Library of Science 17.06.2024
• Subjects: Adaptation; Adaptation (Physiology); Adaptation, Physiological - genetics; Carbon; Cloning; E coli; Escherichia coli; Escherichia coli - genetics; Evolution, Molecular; Experiments; Genetic aspects; Genome, Bacterial; Genomes; Genomics; Mutation; Physiological aspects; Soil productivity; Stationary phase; Whole Genome Sequencing; Israel
• ISSN: 1553-7404; 1553-7390; 1553-7404
• DOI: 10.1371/journal.pgen.1011333

Like many other non-sporulating bacterial species, Escherichia coli is able to survive prolonged periods of resource exhaustion, by entering a state of growth called long-term stationary phase (LTSP). In July 2015, we initiated a set of evolutionary experiments aimed at characterizing the dynamics of E . coli adaptation under LTSP. In these experiments populations of E . coli were allowed to initially grow on fresh rich media, but were not provided with any new external growth resources since their establishment. Utilizing whole genome sequencing data obtained for hundreds of clones sampled at 12 time points spanning the first six years of these experiments, we reveal several novel aspects of the dynamics of adaptation. First, we show that E . coli continuously adapts genetically, up to six years under resource exhaustion, through the highly convergent accumulation of mutations. We further show that upon entry into LTSP, long-lasting lineages are established. This lineage structure is in itself convergent, with similar lineages arising across independently evolving populations. The high parallelism with which adaptations occur under LTSP, combined with the LTSP populations’ lineage structure, enable us to screen for pairs of loci displaying a significant association in the occurrence of mutations, suggestive of a historical contingency. We find that such associations are highly frequent and that a third of convergently mutated loci are involved in at least one such association. Combined our results demonstrate that LTSP adaptation is characterized by remarkably high parallelism and frequent historical contingency.