Host-parasite coevolution promotes innovation through deformations in fitness landscapes

Inferring strain-level mutational drivers of phage-bacteria interaction phenotypes arising during coevolutionary dynamics

The enormous diversity of bacteriophages and their bacterial hosts presents a significant challenge to predict which phages infect a focal set of bacteria. Infection is largely determined by complementary-and largely uncharacterized-genetics of adsorption, injection, cell take-over, and lysis. Here we present a machine learning approach to predict phage-bacteria interactions trained on genome sequences of and phenotypic interactions among 51 Escherichia coli strains and 45 phage λ strains that coevolved in laboratory conditions for 37 days. Leveraging multiple inference strategies and without a priori knowledge of driver mutations, this framework predicts both who infects whom and the quantitative levels of infections across a suite of 2,295 potential interactions. We found that the most effective approach inferred interaction phenotypes from independent contributions from phage and bacteria mutations, accurately predicting 86% of interactions while reducing the relative error in the estimated strength of the infection phenotype by 40%. Feature selection revealed key phage λ and Escherchia coli mutations that have a significant influence on the outcome of phage-bacteria interactions, corroborating sites previously known to affect phage λ infections, as well as identifying mutations in genes of unknown function not previously shown to influence bacterial resistance. The method's success in recapitulating strain-level infection outcomes arising during coevolutionary dynamics may also help inform generalized approaches for imputing genetic drivers of interaction phenotypes in complex communities of phage and bacteria.

Inferring strain-level mutational drivers of phage-bacteria interaction phenotypes arising during coevolutionary dynamics

The enormous diversity of bacteriophages and their bacterial hosts presents a significant challenge to predict which phages infect a focal set of bacteria. Infection is largely determined by complementary - and largely uncharacterized - genetics of adsorption, injection, cell take-over and lysis. Here we present a machine learning approach to predict phage-bacteria interactions trained on genome sequences of and phenotypic interactions amongst 51 Escherichia coli strains and 45 phage λ strains that coevolved in laboratory conditions for 37 days. Leveraging multiple inference strategies and without a priori knowledge of driver mutations, this framework predicts both who infects whom and the quantitative levels of infections across a suite of 2,295 potential interactions. We found that the most effective approach inferred interaction phenotypes from independent contributions from phage and bacteria mutations, accurately predicting 86 % of interactions while reducing the relative error in the estimated strength of the infection phenotype by 40 % . Feature selection revealed key phage λ and E. coli mutations that have a significant influence on the outcome of phage-bacteria interactions, corroborating sites previously known to affect phage λ infections, as well as identifying mutations in genes of unknown function not previously shown to influence bacterial resistance. The method's success in recapitulating strain-level infection outcomes arising during coevolutionary dynamics may also help inform generalized approaches for imputing genetic drivers of interaction phenotypes in complex communities of phage and bacteria.

Environment-independent distribution of mutational effects emerges from microscopic epistasis

Predicting how new mutations alter phenotypes is difficult because mutational effects vary across genotypes and environments. Recently discovered global epistasis, in which the fitness effects of mutations scale with the fitness of the background genotype, can improve predictions, but how the environment modulates this scaling is unknown. We measured the fitness effects of ~100 insertion mutations in 42 strains of Saccharomyces cerevisiae in six laboratory environments and found that the global epistasis scaling is nearly invariant across environments. Instead, the environment tunes one global parameter, the background fitness at which most mutations switch sign. As a consequence, the distribution of mutational effects is predictable across genotypes and environments. Our results suggest that the effective dimensionality of genotype-to-phenotype maps across environments is surprisingly low.

Environment-independent distribution of mutational effects emerges from microscopic epistasis

Predicting how new mutations alter phenotypes is difficult because mutational effects vary across genotypes and environments. Recently discovered global epistasis, where the fitness effects of mutations scale with the fitness of the background genotype, can improve predictions, but how the environment modulates this scaling is unknown. We measured the fitness effects of ~100 insertion mutations in 42 strains of Saccharomyces cerevisiae in six laboratory environments and found that the global-epistasis scaling is nearly invariant across environments. Instead, the environment tunes one global parameter, the background fitness at which most mutations switch sign. As a consequence, the distribution of mutational effects is predictable across genotypes and environments. Our results suggest that the effective dimensionality of genotype-to-phenotype maps across environments is surprisingly low.

More evolvable bacteriophages better suppress their host

The number of multidrug-resistant strains of bacteria is increasing rapidly, while the number of new antibiotic discoveries has stagnated. This trend has caused a surge in interest in bacteriophages as anti-bacterial therapeutics, in part because there is near limitless diversity of phages to harness. While this diversity provides an opportunity, it also creates the dilemma of having to decide which criteria to use to select phages. Here we test whether a phage's ability to coevolve with its host (evolvability) should be considered and how this property compares to two previously proposed criteria: fast reproduction and thermostability. To do this, we compared the suppressiveness of three phages that vary by a single amino acid yet differ in these traits such that each strain maximized two of three characteristics. Our studies revealed that both evolvability and reproductive rate are independently important. The phage most able to suppress bacterial populations was the strain with high evolvability and reproductive rate, yet this phage was unstable. Phages varied due to differences in the types of resistance evolved against them and their ability to counteract resistance. When conditions were shifted to exaggerate the importance of thermostability, one of the stable phages was most suppressive in the short-term, but not over the long-term. Our results demonstrate the utility of biological therapeutics' capacities to evolve and adjust in action to resolve complications like resistance evolution. Furthermore, evolvability is a property that can be engineered into phage therapeutics to enhance their effectiveness.

Structural mechanism of bacteriophage lambda tail's interaction with the bacterial receptor

Bacteriophage infection, a pivotal process in microbiology, initiates with the phage's tail recognizing and binding to the bacterial cell surface, which then mediates the injection of viral DNA. Although comprehensive studies on the interaction between bacteriophage lambda and its outer membrane receptor, LamB, have provided rich information about the system's biochemical properties, the precise molecular mechanism remains undetermined. This study revealed the high-resolution cryo-electron microscopy (cryo-EM) structures of the bacteriophage lambda tail complexed with its irreversible Shigella sonnei 3070 LamB receptor and the closed central tail fiber. These structures reveal the complex processes that trigger infection and demonstrate a substantial conformational change in the phage lambda tail tip upon LamB binding. Providing detailed structures of bacteriophage lambda infection initiation, this study contributes to the expanding knowledge of lambda-bacterial interaction, which holds significance in the fields of microbiology and therapeutic development.

Genomic and phenotypic signatures of bacteriophage coevolution with the phytopathogen Pseudomonas syringae

The rate and trajectory of evolution in an obligate parasite is critically dependent on those of its host(s). Adaptation to a genetically homogeneous host population should theoretically result in specialization, while adaptation to an evolving host population (i.e., coevolution) can result in various outcomes including diversification, range expansion, and/or local adaptation. For viruses of bacteria (bacteriophages, or phages), our understanding of how evolutionary history of the bacterial host(s) impacts viral genotypic and phenotypic evolution is currently limited. In this study, we used whole genome sequencing and two different metrics of phage impacts to compare the genotypes and phenotypes of lytic phages that had either coevolved with or were repeatedly passaged on an unchanging (ancestral) strain of the phytopathogen Pseudomonas syringae. Genomes of coevolved phages had more mutations than those of phages passaged on a constant host, and most mutations were in genes encoding phage tail-associated proteins. Phages from both passaging treatments shared some phenotypic outcomes, including range expansion and divergence across replicate populations, but coevolved phages were more efficient at reducing population growth (particularly of sympatric coevolved hosts). Genotypic similarity correlated with infectivity profile similarity in coevolved phages, but not in phages passaged on the ancestral host. Overall, while adaptation to either host type (coevolving or ancestral) led to divergence in phage tail proteins and infectivity patterns, coevolution led to more rapid molecular changes that increased bacterial killing efficiency and had more predictable effects on infectivity range. Together, these results underscore the important role of hosts in driving viral evolution and in shaping the genotype-phenotype relationship.

Competition-driven eco-evolutionary feedback reshapes bacteriophage lambda's fitness landscape and enables speciation

A major challenge in evolutionary biology is explaining how populations navigate rugged fitness landscapes without getting trapped on local optima. One idea illustrated by adaptive dynamics theory is that as populations adapt, their newly enhanced capacities to exploit resources alter fitness payoffs and restructure the landscape in ways that promote speciation by opening new adaptive pathways. While there have been indirect tests of this theory, to our knowledge none have measured how fitness landscapes deform during adaptation, or test whether these shifts promote diversification. Here, we achieve this by studying bacteriophage [Formula: see text], a virus that readily speciates into co-existing receptor specialists under controlled laboratory conditions. We use a high-throughput gene editing-phenotyping technology to measure [Formula: see text]'s fitness landscape in the presence of different evolved-[Formula: see text] competitors and find that the fitness effects of individual mutations, and their epistatic interactions, depend on the competitor. Using these empirical data, we simulate [Formula: see text]'s evolution on an unchanging landscape and one that recapitulates how the landscape deforms during evolution. [Formula: see text] heterogeneity only evolves in the shifting landscape regime. This study provides a test of adaptive dynamics, and, more broadly, shows how fitness landscapes dynamically change during adaptation, potentiating phenomena like speciation by opening new adaptive pathways.

Rapid bacteria-phage coevolution drives the emergence of multiscale networks

Interactions between species catalyze the evolution of multiscale ecological networks, including both nested and modular elements that regulate the function of diverse communities. One common assumption is that such complex pattern formation requires spatial isolation or long evolutionary timescales. We show that multiscale network structure can evolve rapidly under simple ecological conditions without spatial structure. In just 21 days of laboratory coevolution, Escherichia coli and bacteriophage Φ21 coevolve and diversify to form elaborate cross-infection networks. By measuring ~10,000 phage-bacteria infections and testing the genetic basis of interactions, we identify the mechanisms that create each component of the multiscale pattern. Our results demonstrate how multiscale networks evolve in parasite-host systems, illustrating Darwin's idea that simple adaptive processes can generate entangled banks of ecological interactions.

Evolutionary "Crowdsourcing": Alignment of Fitness Landscapes Allows for Cross-species Adaptation of a Horizontally Transferred Gene

Genes that undergo horizontal gene transfer (HGT) evolve in different genomic backgrounds. Despite the ubiquity of cross-species HGT, the effects of switching hosts on gene evolution remains understudied. Here, we present a framework to examine the evolutionary consequences of host-switching and apply this framework to an antibiotic resistance gene commonly found on conjugative plasmids. Specifically, we determined the adaptive landscape of this gene for a small set of mutationally connected genotypes in 3 enteric species. We uncovered that the landscape topographies were largely aligned with minimal host-dependent mutational effects. By simulating gene evolution over the experimentally gauged landscapes, we found that the adaptive evolution of the mobile gene in one species translated to adaptation in another. By simulating gene evolution over artificial landscapes, we found that sufficient alignment between landscapes ensures such "adaptive equivalency" across species. Thus, given adequate landscape alignment within a bacterial community, vehicles of HGT such as plasmids may enable a distributed form of genetic evolution across community members, where species can "crowdsource" adaptation.

Competition-driven eco-evolutionary feedback reshapes bacteriophage lambda's fitness landscape and enables speciation

A major challenge in evolutionary biology is explaining how populations navigate rugged fitness landscapes without getting trapped on local optima. One idea illustrated by adaptive dynamics theory is that as populations adapt, their newly enhanced capacities to exploit resources alter fitness payoffs and restructure the landscape in ways that promote speciation by opening new adaptive pathways. While there have been indirect tests of this theory, none have measured how fitness landscapes deform during adaptation, or test whether these shifts promote diversification. Here, we achieve this by studying bacteriophage λ, a virus that readily speciates into co-existing receptor specialists under controlled laboratory conditions. We used a high-throughput gene editing-phenotyping technology to measure λ's fitness landscape in the presence of different evolved-λ competitors and found that the fitness effects of individual mutations, and their epistatic interactions, depend on the competitor. Using these empirical data, we simulated λ's evolution on an unchanging landscape and one that recapitulates how the landscape deforms during evolution. λ heterogeneity only evolved in the shifting landscape regime. This study provides a test of adaptive dynamics, and, more broadly, shows how fitness landscapes dynamically change during adaptation, potentiating phenomena like speciation by opening new adaptive pathways.

Evolutionary repeatability of emergent properties of ecological communities

Most species belong to ecological communities where their interactions give rise to emergent community-level properties, such as diversity and productivity. Understanding and predicting how these properties change over time has been a major goal in ecology, with important practical implications for sustainability and human health. Less attention has been paid to the fact that community-level properties can also change because member species evolve. Yet, our ability to predict long-term eco-evolutionary dynamics hinges on how repeatably community-level properties change as a result of species evolution. Here, we review studies of evolution of both natural and experimental communities and make the case that community-level properties at least sometimes evolve repeatably. We discuss challenges faced in investigations of evolutionary repeatability. In particular, only a handful of studies enable us to quantify repeatability. We argue that quantifying repeatability at the community level is critical for approaching what we see as three major open questions in the field: (i) Is the observed degree of repeatability surprising? (ii) How is evolutionary repeatability at the community level related to repeatability at the level of traits of member species? (iii) What factors affect repeatability? We outline some theoretical and empirical approaches to addressing these questions. Advances in these directions will not only enrich our basic understanding of evolution and ecology but will also help us predict eco-evolutionary dynamics. This article is part of the theme issue 'Interdisciplinary approaches to predicting evolutionary biology'.

Experimental Evolution of the TolC-Receptor Phage U136B Functionally Identifies a Tail Fiber Protein Involved in Adsorption through Strong Parallel Adaptation

Bacteriophages have received recent attention for their therapeutic potential to treat antibiotic-resistant bacterial infections. One particular idea in phage therapy is to use phages that not only directly kill their bacterial hosts but also rely on particular bacterial receptors, such as proteins involved in virulence or antibiotic resistance. In such cases, the evolution of phage resistance would correspond to the loss of those receptors, an approach termed evolutionary steering. We previously found that during experimental evolution, phage U136B can exert selection pressure on Escherichia coli to lose or modify its receptor, the antibiotic efflux protein TolC, often resulting in reduced antibiotic resistance. However, for TolC-reliant phages like U136B to be used therapeutically, we also need to study their own evolutionary potential. Understanding phage evolution is critical for the development of improved phage therapies as well as the tracking of phage populations during infection. Here, we characterized phage U136B evolution in 10 replicate experimental populations. We quantified phage dynamics that resulted in five surviving phage populations at the end of the 10-day experiment. We found that phages from all five surviving populations had evolved higher rates of adsorption on either ancestral or coevolved E. coli hosts. Using whole-genome and whole-population sequencing, we established that these higher rates of adsorption were associated with parallel molecular evolution in phage tail protein genes. These findings will be useful in future studies to predict how key phage genotypes and phenotypes influence phage efficacy and survival despite the evolution of host resistance. IMPORTANCE Antibiotic resistance is a persistent problem in health care and a factor that may help maintain bacterial diversity in natural environments. Bacteriophages ("phages") are viruses that specifically infect bacteria. We previously discovered and characterized a phage called U136B, which infects bacteria through TolC. TolC is an antibiotic resistance protein that helps bacteria pump antibiotics out of the cell. Over short timescales, phage U136B can be used to evolutionarily "steer" bacterial populations to lose or modify the TolC protein, sometimes reducing antibiotic resistance. In this study, we investigate whether U136B itself evolves to better infect bacterial cells. We discovered that the phage can readily evolve specific mutations that increase its infection rate. This work will be useful for understanding how phages can be used to treat bacterial infections.

Comparison of bacterial suppression by phage cocktails, dual-receptor generalists, and coevolutionarily trained phages

The evolution and spread of antibiotic-resistant bacteria have renewed interest in phage therapy, the use of bacterial viruses (phages) to combat bacterial infections. The delivery of phages in cocktails where constituent phages target different modalities (e.g., receptors) may improve treatment outcomes by making it more difficult for bacteria to evolve resistance. However, the multipartite nature of cocktails may lead to unintended evolutionary and ecological outcomes. Here, we compare a 2-phage cocktail with a largely unconsidered group of phages: generalists that can infect through multiple, independent receptors. We find that λ phage generalists and cocktails that target the same receptors (LamB and OmpF) suppress Escherichia coli similarly for ~2 days. Yet, a "trained" generalist phage, which previously adapted to its host via 28 days of coevolution, demonstrated superior suppression. To understand why the trained generalist was more effective, we measured the resistance of bacteria against each of our phages. We find that, when bacteria were assailed by two phages in the cocktail, they evolved mutations in manXYZ, a host inner-membrane transporter that λ uses to move its DNA across the periplasmic space and into the cell for infection. This provided cross-resistance against the cocktail and untrained generalist. However, these mutations were ineffective at blocking the trained generalist because, through coevolutionary training, it evolved to bypass manXYZ resistance. The trained generalist's past experiences in training make it exceedingly difficult for bacteria to evolve resistance, further demonstrating the utility of coevolutionary phage training for improving the therapeutic properties of phages.