The evolutionary success of lytic bacteriophages depends on key life-history traits, including adsorption, lysis time, burst size, and persistence in the environment. However, how these traits evolve to allow adaptation to different environments remains poorly understood. Here, we explored this question in ΦX174, combining experimental evolution and mathematical modelling. By investigating how serial transfer conditions shape evolutionary outcomes in liquid culture, we found that the time between transfers imposes divergent selection on adsorption and context-dependent directional selection on persistence. Longer transfer intervals, which allow multiple infection cycles until host depletion, favoured fast-adsorbing, highly persistent mutants that could rapidly initiate infections and remained viable in the absence of the host. In contrast, shorter transfer intervals selected for slower adsorption without substantially altering persistence. Mathematical modelling of phage population dynamics predicted that adsorption evolution during short transfers reflects a trade-off between two opposing selective forces within each transfer: an early phase in which susceptible hosts are abundant and adsorption is productive, favouring fast adsorption, and a later phase in which most hosts are already infected and adsorption primarily removes phage particles via attachment to already infected cells, favouring slower adsorption. A single-point mutation in the major capsid protein was sufficient to drive these changes in adsorption. In the case of fast-adsorbing mutants, this mutation was positively pleiotropic and also enhanced environmental persistence. Our findings show how simple changes in propagation conditions can steer phage phenotypes, providing insights relevant to evolutionary biology and phage therapy.

Understanding why some traits are maintained whereas others are repeatedly lost is a central question in evolutionary biology. Here, we address this question through the evolutionary dynamics of autoregulation of prfB, which encodes peptide-chain release factor 2 (RF2), a factor in bacterial translation termination. RF2 recognizes UGA and UAA stop codons and catalyzes polypeptide release. In many species, prfB contains an internal UGA stop codon that causes premature termination by RF2. Full RF2 synthesis depends on a + 1 programed ribosomal frameshifting (PRF) event at this stop codon, which occurs more frequently when RF2 levels are low, resulting in autoregulation of prfB expression. While widespread, this mechanism has been lost repeatedly across bacteria. We combined phylogenetics, experimental evolution, and molecular genetics to investigate the evolutionary forces underlying this loss. Phylogenetically informed analyses revealed no significant correlation between autoregulation and UGA stop codon usage, and autoregulation elimination in Pseudomonas fluorescens SBW25 had no detectable fitness effect. However, engineered mutations that reduced frameshifting at the prfB autoregulatory site caused fitness defects that were compensated by two classes of mutation: mutations affecting ribosome-associated proteins (RsmA, RsmH, RplI), and single-nucleotide deletions in prfB that adjusted the reading frame to bypass the internal stop codon, eliminating autoregulation. These results suggest that loss of prfB autoregulation can be facilitated by compensatory mutations when frameshifting at the prfB autoregulatory site is compromised and RF2 production is insufficient. Our findings illustrate how compensatory evolution can favor trait loss when the fitness benefit of losing the trait outweighs its cost.

Bacteriophages infect gram-negative bacteria by attaching to molecules present on the bacterial surface, often lipopolysaccharides (LPS). Modification of LPS can lead to resistance to phage infection. In addition, LPS modifications can impact antibiotic susceptibility, allowing for phage-antibiotic synergism. The evolutionary mechanism(s) behind such synergistic interactions remain largely unclear. Here, we show that the presence of antibiotics can affect the evolution of resistance to phage infection, using phage ΦX174 and Escherichia coli C. We use a collection of 34 E. coli C LPS strains, each of which is resistant to ΦX174, and has either a "rough" or "deep rough" LPS phenotype. Growth of the bacterial strains with the deep rough phenotype is inhibited at low concentrations of chloramphenicol and, to a much lesser degree, gentamicin. Treating E. coli C wild type with ΦX174 and chloramphenicol eliminates the emergence of mutants with the deep rough phenotype, and thereby slows the evolution of resistance to phage infection. At slightly lower chloramphenicol concentrations, phage resistance rates are similar to those observed at high concentrations; yet, we show that the diversity of possible mutants is much larger than at higher chloramphenicol concentrations. These data suggest that specific antibiotic concentrations can lead to synergistic phage-antibiotic interactions that disappear at higher antibiotic concentrations. Overall, we show that the change in survival of various ΦX174-resistant E. coli C mutants in the presence of antibiotics can explain the observed phage-antibiotic synergism.

I nsertion S equences (ISs) are mobile pieces of DNA that are widespread in bacterial genomes. IS movements typically involve (i) excision of the IS element, (ii) cutting of the target site DNA, and (iii) IS element insertion. This process generates a new copy of the IS element, as well as a short duplication at the target site. It has been noted that, when observing extant IS element copies in a genome, occasionally no T arget S ite D uplication (TSD) is readily identifiable. This has been attributed to degeneration of the TSD at some point after the insertion event. Here, we provide evidence that some IS movement events – namely, those that occur in association with large-scale genome rearrangements – occur without generating TSDs. In support of this hypothesis, we provide two direct, empirical observations of such IS transposition events: an IS 481 movement occurring with a large duplication in Pseudomonas fluorescens SBW25, and an IS 5 /IS 1182 movement plus a large deletion in Escherichia coli C. Additionally, we use sequencing data from the Lenski long-term evolution experiment to provide a further 14 examples of IS 150 movements in E. coli B that are associated with large deletions and do not carry TSDs. Overall, our results indicate that some IS elements can insert into, and thus repair, existing DNA breaks in bacterial genomes. ### Competing Interest Statement The authors have declared no competing interest.

Integrative and conjugative elements (ICEs) are self-transmissible mobile elements that transfer functional genetic units across broad phylogenetic distances. Accessory genes shuttled by ICEs can make significant contributions to bacterial fitness. Most ICEs characterized to date encode readily observable phenotypes contributing to symbiosis, pathogenicity, and antimicrobial resistance, yet the majority of ICEs carry genes of unknown function. Recent observations of rapid acquisition of ICEs in a pandemic lineage of Pseudomonas syringae pv. actinidae led to investigation of the structural and functional diversity of these elements. Fifty-three unique ICE types were identified across the P. syringae species complex. Together they form a distinct family of ICEs (PsICEs) that share a distant relationship to ICEs found in Pseudomonas aeruginosa . PsICEs are defined by conserved backbone genes punctuated by an array of accessory cargo genes, are highly recombinogenic, and display distinct evolutionary histories compared to their bacterial hosts. The most common cargo is a recently disseminated 16-kb mobile genetic element designated Tn 6212 . Deletion of Tn 6212 did not alter pathogen growth in planta, but mutants displayed fitness defects when grown on tricarboxylic acid (TCA) cycle intermediates. RNA-seq analysis of a set of nested deletion mutants showed that a Tn 6212 -encoded LysR regulator has global effects on chromosomal gene expression. We show that Tn 6212 responds to preferred carbon sources and manipulates bacterial metabolism to maximize growth.

Repetitive sequences are a common feature of bacterial genomes. Some repetitive sequences such as REPINs are mobile within the genome but inherited only vertically from mother to daughter across bacterial genomes. Selfish elements in contrast are mobile within the genome but also travel horizontally from genome to genome. Yet, no matter the nature of the association between the repetitive elements and the host, it is difficult to study the evolutionary dynamics of repetitive sequences across genomes. If it is unclear whether two sequences in two different genomes are in orthologous positions, it is difficult to infer parameters like the replication rate, horizontal transfer rate and rate of loss. Here we present a tool to facilitate these analyses called REPORTH. REPORTH determines whether repetitive sequences in different but closely related bacterial genomes occur in orthologous genomic positions. Whether a position is orthologous or not depends on the orthology of flanking sequences. Flanking sequences are deemed orthologous if they are bidirectional best hits. All repetitive sequences that are found in orthologous positions across different genomes are grouped together. Analyses of these groups can be used to study the evolutionary dynamics of selfish repetitive elements such as insertion sequences, but also for mutualistic repetitive elements such as REPINs. ### Competing Interest Statement The authors have declared no competing interest.

With the emergence of widespread antibiotic resistance, phages are an appealing alternative to antibiotics in the fight against multidrug-resistant bacteria. Over the past few years, many phages have been isolated from various environments to treat bacterial pathogens. While isolating novel phages for treatment has had some success for compassionate use, developing novel phages into a general therapeutic will require considerable time and financial resource investments. These investments may be less significant for well-established phage model systems. The knowledge acquired from decades of research on their structure, life cycle, and evolution ensures safe application and efficient handling. However, one major downside of the established phage model systems is their inability to infect pathogenic bacteria. This problem is not insurmountable; phage host range can be extended through genetic engineering or evolution experiments. In the future, breeding model phages to infect pathogens could provide a new avenue to develop phage therapeutic agents.

Insertion sequences (ISs) are mobile pieces of DNA that are widespread in bacterial genomes. IS movements typically involve (i) excision of the IS element, (ii) cutting of target site DNA, and (iii) IS element insertion. This process generates a new copy of the IS element and a short duplication at the target site. It has been noted that, for some extant IS copies, no target site duplications (TSDs) are readily identifiable. TSD absence has been attributed to degeneration of the TSD after the insertion event, recombination between identical ISs, or adjacent deletions. Indeed, the latter two—recombination between ISs and adjacent deletions—are frequent causes for the absence of TSDs, which we demonstrate here in an analysis of genome sequence data from the Lenski long-term evolution experiment. Furthermore, we propose that some IS movements—namely, those that occur in association with large-scale genomic rearrangements—do not generate TSDs, and occur without evidence for recombination between ISs or adjacent deletions. In support of this hypothesis, we provide two direct, empirical observations of such IS transposition events: an IS5 movement plus a large deletion in Escherichia coli C, and an IS481 movement occurring with a large duplication in Pseudomonas fluorescens SBW25. Although unlikely, it is possible that the observed deletion and associated IS movement occurred in two successive events in one overnight culture. However, an IS at the center of a large-scale duplication is not readily explained, suggesting that IS element activity may promote both large-scale deletions and duplications. IMPORTANCE Insertion sequences are the most common mobile genetic elements found in bacterial genomes, and hence they significantly impact bacterial evolution. We observe insertion sequence movement at the center of large-scale deletions and duplications that occurred during laboratory evolution experiments with Escherichia coli and Pseudomonas fluorescens, involving three distinct types of transposase. We raise the possibility that the transposase does not mediate DNA cleavage but instead inserts into existing DNA breaks. Our research highlights the importance of insertion sequences for the generation of large-scale genomic rearrangements and raises questions concerning the mechanistic basis of these mutations.

Compared to eukaryotes, repetitive sequences are rare in bacterial genomes and usually do not persist for long. Yet, there is at least one class of persistent prokaryotic mobile genetic elements: REPINs. REPINs are non-autonomous transposable elements replicated by single-copy transposases called RAYTs. REPIN-RAYT systems are mostly vertically inherited and have persisted in individual bacterial lineages for millions of years. Discovering and analyzing REPIN populations and their corresponding RAYT transposases in bacterial species can be rather laborious, hampering progress in understanding REPIN-RAYT biology and evolution. Here we present RAREFAN, a webservice that identifies REPIN populations and their corresponding RAYT transposase in a given set of bacterial genomes. We demonstrate RAREFAN’s capabilities by analyzing a set of 49 Stenotrophomonas maltophilia genomes, containing nine different REPIN-RAYT systems. We guide the reader through the process of identifying and analyzing REPIN-RAYT systems across S. maltophilia , highlighting erroneous associations between REPIN and RAYTs, and providing solutions on how to find correct associations. RAREFAN enables rapid, large-scale detection of REPINs and RAYTs, and provides insight into the fascinating world of intragenomic sequence populations in bacterial genomes. RAREFAN is available at http://rarefan.evolbio.mpg.de .

Phage therapy is a promising method for the treatment of multidrug-resistant bacterial infections. However, its long-term efficacy depends on understanding the evolutionary effects of the treatment. Current knowledge of such evolutionary effects is lacking, even in well-studied systems. We used the bacterium Escherichia coli C and its bacteriophage ΦX174, which infects cells using host lipopolysaccharide (LPS) molecules. We first generated 31 bacterial mutants resistant to ΦX174 infection. Based on the genes disrupted by these mutations, we predicted that these E. coli C mutants collectively produce eight unique LPS structures. We then developed a series of evolution experiments to select for ΦX174 mutants capable of infecting the resistant strains. During phage adaptation, we distinguished two types of phage resistance: one that was easily overcome by ΦX174 with few mutational steps ("easy" resistance) and one that was more difficult to overcome ("hard" resistance). We found that increasing the diversity of the host and phage populations could accelerate the adaptation of phage ΦX174 to overcome the hard resistance phenotype. From these experiments, we isolated 16 ΦX174 mutants that, together, can infect all 31 initially resistant E. coli C mutants. Upon determining the infectivity profiles of these 16 evolved phages, we uncovered 14 distinct profiles. Given that only eight profiles are anticipated if the LPS predictions are correct, our findings highlight that the current understanding of LPS biology is insufficient to accurately forecast the evolutionary outcomes of bacterial populations infected by phage.

Integrative mobile genetic elements (MGEs), such as transposons and insertion sequences, propagate within bacterial genomes, but persistence times in individual lineages are short. For long-term survival, MGEs must continuously invade new hosts by horizontal transfer. Theoretically, MGEs that persist for millions of years in single lineages, and are thus subject to vertical inheritance, should not exist. Here we draw attention to an exception — a class of MGE termed REPIN. REPINs are non-autonomous MGEs whose duplication depends on non-jumping RAYT transposases. Comparisons of REPINs and typical MGEs show that replication rates of REPINs are orders of magnitude lower, REPIN population size fluctuations correlate with changes in available genome space, REPIN conservation depends on RAYT function, and REPIN diversity accumulates within host lineages. These data lead to the hypothesis that REPINs form enduring, beneficial associations with eubacterial chromosomes. Given replicative nesting, our hypothesis predicts conflicts arising from the diverging effects of selection acting simultaneously on REPINs and host genomes. Evidence in support comes from patterns of REPIN abundance and diversity in two distantly related bacterial species. Together this bolsters the conclusion that REPINs are the genetic counterpart of mutualistic endosymbiotic bacteria.

AbstractPhage therapy is a promising method for the treatment of multi-drug-resistant bacterial infections. However, its long-term efficacy depends on understanding the evolutionary effects of the treatment. Current knowledge of such evolutionary effects is lacking, even in well-studied systems. We used the bacteriumEscherichia coliC and its bacteriophage ΦX174, which infects cells using host lipopolysaccharide (LPS) molecules. We first generated 31 bacterial mutants resistant to ΦX174 infection. Based on the genes disrupted by these mutations, we predicted that theseE. coliC mutants collectively produce eight unique LPS structures. We then developed a series of evolution experiments to select for ΦX174 mutants capable of infecting the resistant strains. During phage adaptation, we distinguished two types of phage resistance: one that was easily overcome by ΦX174 with few mutational steps (“easy” resistance), and one that was more difficult to overcome (“hard” resistance). We found that increasing the diversity of the host and phage populations could accelerate the adaptation of phage ΦX174 to overcome the hard resistance phenotype. From these experiments, we isolated 16 ΦX174 mutants that, together, can infect all 31 initially resistantE. coliC mutants. Upon determining the infectivity profiles of these 16 evolved phages, we uncovered 14 distinct profiles. Given that only eight profiles are anticipated if the LPS predictions are correct, our findings highlight that the current understanding of LPS biology is insufficient to accurately forecast the evolutionary outcomes of bacterial populations infected by phage.

This dataset contains the genome sequences of 49 S. maltophilia strains (input.zip) and the processed data from four individual RAREFAN runs. RAREFAN identifies RAYTs REPINS for each supplied genome and a reference genome. Four different strains where used as a reference genome: AA1, AB550 FDARGOOS_649, ISMMS3, and Sm53. The data were processed with default RAREFAN job parameters. Links to the original RAREFAN jobs are given below.

Eukaryotes and prokaryotes have distinct genome architectures, with marked differences in genome size, the ratio of coding/non-coding DNA, and the abundance of transposable elements (TEs). As TEs replicate independently of their hosts, the proliferation of TEs is thought to have driven genome expansion in eukaryotes. However, prokaryotes also have TEs in intergenic spaces, so why do prokaryotes have small, streamlined genomes? Using an in silico model describing the genomes of single-celled asexual organisms that coevolve with TEs, we show that TEs acquired from the environment by horizontal gene transfer can promote the evolution of genome streamlining. The process depends on local interactions and is underpinned by rock-paper-scissors dynamics in which populations of cells with streamlined genomes beat TEs, which beat non-streamlined genomes, which beat streamlined genomes, in continuous and repeating cycles. Streamlining is maladaptive to individual cells, but improves lineage viability by hindering the proliferation of TEs. Streamlining does not evolve in sexually reproducing populations because recombination partially frees TEs from the deleterious effects they cause. This article is part of the theme issue 'The secret lives of microbial mobile genetic elements'.

Relationships among organisms, in which one lives inside of another, with benefits accruing to both partners, are referred to as endosymbiotic. Such relationships are common in the biological world, yet descriptions are confined to organismal life. Here we argue that short sequence repeats known as REPINs – whose replication is dependent on a non-jumping RAYT transposase – form a similar facultative symbiotic relationship with the bacterial chromosome. Evidence stems from distribution patterns across the eubacteria: persistence times of many millions of years, exceedingly rare duplication rates, vertical transmission, and population biology typical of living organisms, including population size fluctuations that correlate with available genome space. Additional analysis of patterns of REPIN evolution conform with theoretical predictions of conflict (and resolution) arising from the effects of selection operating simultaneously on REPINs and host cells. A search for similar kinds of genomic symbionts suggests that the REPIN-RAYT system is not unique.

Compared to their eukaryotic counterparts, bacterial genomes are small and contain extremely tightly packed genes. Repetitive sequences are rare but not completely absent. One of the most common repeat families is REPINs. REPINs can replicate in the host genome and form populations that persist for millions of years. Here, we model the interactions of these intragenomic sequence populations with the bacterial host. We first confirm well-established results, in the presence and absence of horizontal gene transfer (hgt) sequence populations either expand until they drive the host to extinction or the sequence population gets purged from the genome. We then show that a sequence population can be stably maintained, when each individual sequence provides a benefit that decreases with increasing sequence population size. Maintaining a sequence population of stable size also requires the replication of the sequence population to be costly to the host, otherwise the sequence population size will increase indefinitely. Surprisingly, in regimes with high hgt rates, the benefit conferred by the sequence population does not have to exceed the damage it causes to its host. Our analyses provide a plausible scenario for the persistence of sequence populations in bacterial genomes. We also hypothesize a limited biologically relevant parameter range for the provided benefit, which can be tested in future experiments.

A recommendation – based on reviews by three anonymous reviewers – of the article: Cury J, Haller BC, Achaz G, Jay F (2021) Simulation of bacterial populations with SLiM. bioRxiv, 2020.09.28.316869, version 5 peer-reviewed and recommended by Peer community in Evolutionary Biology. https://doi.org/10.1101/2020.09.28.316869

Convergent evolution describes the process of different populations acquiring similar phenotypes or genotypes. Complex organisms with large genomes only rarely and only under very strong selection converge to the same genotype. In contrast, independent virus populations with very small genomes often acquire identical mutations. Here we test the hypothesis of whether convergence in early HIV-1 infection is common enough to serve as an indicator for selection. To this end, we measure the number of convergent mutations in a well-studied dataset of full-length HIV-1 env genes sampled from HIV-1 infected individuals during early infection. We compare this data to a neutral model and find an excess of convergent mutations. Convergent mutations are not evenly distributed across the env gene, but more likely to occur in gp41, which suggests that convergent mutations provide a selective advantage and hence are positively selected. In contrast, mutations that are only found in an HIV-1 population of a single individual are significantly affected by purifying selection. Our analysis suggests that comparisons between convergent and private mutations with neutral models allow us to identify positive and negative selection in small viral genomes. Our results also show that selection significantly shapes HIV-1 populations even before the onset of the adaptive immune system.