Adaptive evolution of nontransitive fitness in yeast.

A common misconception is that evolution is a linear 'march of progress', where each organism along a line of descent is more fit than all those that came before it. Rejecting this misconception implies that evolution is nontransitive: a series of adaptive events will, on occasion, produce organisms that are less fit compared to a distant ancestor. Here we identify a nontransitive evolutionary sequence in a 1000-generation yeast evolution experiment. We show that nontransitivity arises due to adaptation in the yeast nuclear genome combined with the stepwise deterioration of an intracellular virus, which provides an advantage over viral competitors within host cells. Extending our analysis, we find that nearly half of our ~140 populations experience multilevel selection, fixing adaptive mutations in both the nuclear and viral genomes. Our results provide a mechanistic case-study for the adaptive evolution of nontransitivity due to multilevel selection in a 1000-generation host/virus evolution experiment.

It is widely accepted in biology that all life on Earth gradually evolved over billions of years from a single ancestor. Yet, there is still much about this process that is not fully understood. Evolution is often thought of as progressing in a linear fashion, with each new generation being better adapted to its environment than the last. But it has been proposed that evolution is also nontransitive: this means even if each generation is 'fitter' than its immediate predecessor, these series of adaptive changes will occasionally result in organisms that are less fit than their distant ancestors. Laboratory experiments of evolution are a good way to test evolutionary theories because they allow researchers to create scenarios that are impossible to observe in natural populations, such as an organism competing against its extinct ancestors. Buskirk et al. set up such an experiment using yeast to determine whether nontransitive effects can be observed in the direct descendants of an organism. At the start of the experiment, the yeast cells were host to a non-infectious 'killer' virus that is common among yeast. Cells containing the virus produce a toxin that destroys other yeast that lack the virus. The populations of yeast were given a nutrient-rich broth in which to grow and subjected to a simple evolutionary pressure: to grow fast, which limits the amount of resources available. As the yeast evolved, they gained beneficial genetic mutations that allowed them to outcompete their neighbors, and they passed these traits down to their descendants. Some of these mutations occurred not in the yeast genome, but in the genome of the killer virus, and this stopped the yeast infected with the virus from producing the killer toxin. Over time, other mutations resulted in the infected yeast no longer being immune to the toxin. Thus, when Buskirk et al. pitted these yeast against their distant ancestors, the new generation were destroyed by the toxins the older generation produced. These findings provide the first experimental evidence for nontransitivity along a line of descent. The results have broad implications for our understanding of how evolution works, casting doubts over the idea that evolution always involves a direct progression towards new, improved traits.

Computational Basis for On-Demand Production of Diversified Therapeutic Phage Cocktails.

New therapies are necessary to combat increasingly antibiotic-resistant bacterial pathogens. We have developed a technology platform of computational, molecular biology, and microbiology tools which together enable on-demand production of phages that target virtually any given bacterial isolate. Two complementary computational tools that identify and precisely map prophages and other integrative genetic elements in bacterial genomes are used to identify prophage-laden bacteria that are close relatives of the target strain. Phage genomes are engineered to disable lysogeny, through use of long amplicon PCR and Gibson assembly. Finally, the engineered phage genomes are introduced into host bacteria for phage production. As an initial demonstration, we used this approach to produce a phage cocktail against the opportunistic pathogen Pseudomonas aeruginosa PAO1. Two prophage-laden P. aeruginosa strains closely related to PAO1 were identified, ATCC 39324 and ATCC 27853. Deep sequencing revealed that mitomycin C treatment of these strains induced seven phages that grow on P. aeruginosa PAO1. The most diverse five phages were engineered for nonlysogeny by deleting the integrase gene (int), which is readily identifiable and typically conveniently located at one end of the prophage. The Δint phages, individually and in cocktails, killed P. aeruginosa PAO1 in liquid culture as well as in a waxworm (Galleria mellonella) model of infection.IMPORTANCE The antibiotic resistance crisis has led to renewed interest in phage therapy as an alternative means of treating infection. However, conventional methods for isolating pathogen-specific phage are slow, labor-intensive, and frequently unsuccessful. We have demonstrated that computationally identified prophages carried by near-neighbor bacteria can serve as starting material for production of engineered phages that kill the target pathogen. Our approach and technology platform offer new opportunity for rapid development of phage therapies against most, if not all, bacterial pathogens, a foundational advance for use of phage in treating infectious disease.

Genome Sequences of Six Cluster N Mycobacteriophages, Kevin1, Nenae, Parmesanjohn, ShrimpFriedEgg, Smurph, and SpongeBob, Isolated on Mycobacterium smegmatis mc2155.

The annotation of six cluster N Mycobacterium smegmatis phages (Kevin1, Nenae, Parmesanjohn, ShrimpFriedEgg, Smurph, and SpongeBob) reveals regions of genomic diversity, particularly within the central region of the genome. The genome of Kevin1 includes two orphams (genes with no similarity to other phage genes), with one predicted to encode an AAA-ATPase.