Predicting the evolution of Escherichia coli by a data-driven approach

Proteome partitioning constraints in long-term laboratory evolution

Adaptive laboratory evolution experiments provide a controlled context in which the dynamics of selection and adaptation can be followed in real-time at the single-nucleotide level. And yet this precision introduces hundreds of degrees-of-freedom as genetic changes accrue in parallel lineages over generations. On short timescales, physiological constraints have been leveraged to provide a coarse-grained view of bacterial gene expression characterized by a small set of phenomenological parameters. Here, we ask whether this same framework, operating at a level between genotype and fitness, informs physiological changes that occur on evolutionary timescales. Using a strain adapted to growth in glucose minimal medium, we find that the proteome is substantially remodeled over 40 000 generations. The most striking change is an apparent increase in enzyme efficiency, particularly in the enzymes of lower-glycolysis. We propose that deletion of metabolic flux-sensing regulation early in the adaptation results in increased enzyme saturation and can account for the observed proteome remodeling.

Comparative genomics hints at dispensability of multiple essential genes in two Escherichia coli L-form strains

Despite the critical role of bacterial cell walls in maintaining cell shapes, certain environmental stressors can induce the transition of many bacterial species into a wall-deficient state called L-form. Long-term induced Escherichia coli L-forms lose their rod shape and usually hold significant mutations that affect cell division and growth. Besides this, the genetic background of L-form bacteria is still poorly understood. In the present study, the genomes of two stable L-form strains of E. coli (NC-7 and LWF+) were sequenced and their gene mutation status was determined and compared with their parental strains. Comparative genomic analysis between two L-forms reveals both unique adaptions and common mutated genes, many of which belong to essential gene categories not involved in cell wall biosynthesis, indicating that L-form genetic adaptation impacts crucial metabolic pathways. Missense variants from L-forms and Lenski's long-term evolution experiment (LTEE) were analyzed in parallel using an optimized DeepSequence pipeline to investigate predicted mutation effects (α) on protein functions. We report that the two L-form strains analyzed display a frequency of 6-10% (0% for LTEE) in mutated essential genes where the missense variants have substantial impact on protein functions (α<0.5). This indicates the emergence of different survival strategies in L-forms through changes in essential genes during adaptions to cell wall deficiency. Collectively, our results shed light on the detailed genetic background of two E. coli L-forms and pave the way for further investigations of the gene functions in L-form bacterial models.

Deep mutational scanning reveals the molecular determinants of RNA polymerase-mediated adaptation and tradeoffs

RNA polymerase (RNAP) is emblematic of complex biological systems that control multiple traits involving trade-offs such as growth versus maintenance. Laboratory evolution has revealed that mutations in RNAP subunits, including RpoB, are frequently selected. However, we lack a systems view of how mutations alter the RNAP molecular functions to promote adaptation. We, therefore, measured the fitness of thousands of mutations within a region of rpoB under multiple conditions and genetic backgrounds, to find that adaptive mutations cluster in two modules. Mutations in one module favor growth over maintenance through a partial loss of an interaction associated with faster elongation. Mutations in the other favor maintenance over growth through a destabilized RNAP-DNA complex. The two molecular handles capture the versatile RNAP-mediated adaptations. Combining both interaction losses simultaneously improved maintenance and growth, challenging the idea that growth-maintenance tradeoff resorts only from limited resources, and revealing how compensatory evolution operates within RNAP.

Laboratory evolution reveals general and specific tolerance mechanisms for commodity chemicals

Although strain tolerance to high product concentrations is a barrier to the economically viable biomanufacturing of industrial chemicals, chemical tolerance mechanisms are often unknown. To reveal tolerance mechanisms, an automated platform was utilized to evolve Escherichia coli to grow optimally in the presence of 11 industrial chemicals (1,2-propanediol, 2,3-butanediol, glutarate, adipate, putrescine, hexamethylenediamine, butanol, isobutyrate, coumarate, octanoate, hexanoate), reaching tolerance at concentrations 60%-400% higher than initial toxic levels. Sequencing genomes of 223 isolates from 89 populations, reverse engineering, and cross-compound tolerance profiling were employed to uncover tolerance mechanisms. We show that: 1) cells are tolerized via frequent mutation of membrane transporters or cell wall-associated proteins (e.g., ProV, KgtP, SapB, NagA, NagC, MreB), transcription and translation machineries (e.g., RpoA, RpoB, RpoC, RpsA, RpsG, NusA, Rho), stress signaling proteins (e.g., RelA, SspA, SpoT, YobF), and for certain chemicals, regulators and enzymes in metabolism (e.g., MetJ, NadR, GudD, PurT); 2) osmotic stress plays a significant role in tolerance when chemical concentrations exceed a general threshold and mutated genes frequently overlap with those enabling chemical tolerance in membrane transporters and cell wall-associated proteins; 3) tolerization to a specific chemical generally improves tolerance to structurally similar compounds whereas a tradeoff can occur on dissimilar chemicals, and 4) using pre-tolerized starting isolates can hugely enhance the subsequent production of chemicals when a production pathway is inserted in many, but not all, evolved tolerized host strains, underpinning the need for evolving multiple parallel populations. Taken as a whole, this study provides a comprehensive genotype-phenotype map based on identified mutations and growth phenotypes for 223 chemical tolerant isolates.

Towards evolutionary predictions: Current promises and challenges

Evolution has traditionally been a historical and descriptive science, and predicting future evolutionary processes has long been considered impossible. However, evolutionary predictions are increasingly being developed and used in medicine, agriculture, biotechnology and conservation biology. Evolutionary predictions may be used for different purposes, such as to prepare for the future, to try and change the course of evolution or to determine how well we understand evolutionary processes. Similarly, the exact aspect of the evolved population that we want to predict may also differ. For example, we could try to predict which genotype will dominate, the fitness of the population or the extinction probability of a population. In addition, there are many uses of evolutionary predictions that may not always be recognized as such. The main goal of this review is to increase awareness of methods and data in different research fields by showing the breadth of situations in which evolutionary predictions are made. We describe how diverse evolutionary predictions share a common structure described by the predictive scope, time scale and precision. Then, by using examples ranging from SARS-CoV2 and influenza to CRISPR-based gene drives and sustainable product formation in biotechnology, we discuss the methods for predicting evolution, the factors that affect predictability and how predictions can be used to prevent evolution in undesirable directions or to promote beneficial evolution (i.e. evolutionary control). We hope that this review will stimulate collaboration between fields by establishing a common language for evolutionary predictions.

Transcriptional Potential Determines the Adaptability of Escherichia coli Strains with Different Fitness Backgrounds

Adaptation through the fitness landscape may be influenced by the gene pool or expression network. However, genetic factors that determine the contribution of beneficial mutations during adaptive evolution are poorly understood. In this study, we experimentally evolved wild-type Escherichia coli K-12 MG1655 and its isogenic derivative that has two additional replication origins and shows higher background fitness. During the short time of experimental evolution, the fitness gains of the two E. coli strains with different fitness backgrounds converged. Populational genome sequencing revealed various mutations with different allele frequencies in evolved populations. Several mutations occurred in genes affecting transcriptional regulation (e.g., RNA polymerase subunit, RNase, ppGpp synthetase, and transcription termination/antitermination factor genes). When we introduced mutations into the ancestral E. coli strains, beneficial effects tended to be lower in the ancestor with higher initial fitness. Replication rate analysis showed that the various replication indices do not correlate with the growth rate. Transcriptome profiling showed that gene expression and gene ontology are markedly enriched in populations with lower background fitness after experimental evolution. Further, the degree of transcriptional change was proportional to the fitness gain. Thus, the evolutionary trajectories of bacteria with different fitness backgrounds can be complex and counterintuitive. Notably, transcriptional change is a major contributor to adaptability. IMPORTANCE Predicting the adaptive potential of bacterial populations can be difficult due to their complexity and dynamic environmental conditions. Also, epistatic interaction between mutations affects the adaptive trajectory. Nevertheless, next-generation sequencing sheds light on understanding evolutionary dynamics through high-throughput genome and transcriptome information. Experimental evolution of two E. coli strains with different background fitness showed that the trajectories of fitness gain, which slowed down during the later stages of evolution, became convergent. This suggests that the adaptability of bacteria can be counterintuitive and that predicting the evolutionary path of bacteria can be difficult even in a constant environment. In addition, transcriptional change is associated with fitness gain during the evolutionary process. Thus, the adaptability of cells depends on their intrinsic genetic capacity for a given evolutionary period. This should be considered when genetically engineered bacteria are optimized through adaptive evolution.

Whole-genome sequencing and gene sharing network analysis powered by machine learning identifies antibiotic resistance sharing between animals, humans and environment in livestock farming

Anthropogenic environments such as those created by intensive farming of livestock, have been proposed to provide ideal selection pressure for the emergence of antimicrobial-resistant Escherichia coli bacteria and antimicrobial resistance genes (ARGs) and spread to humans. Here, we performed a longitudinal study in a large-scale commercial poultry farm in China, collecting E. coli isolates from both farm and slaughterhouse; targeting animals, carcasses, workers and their households and environment. By using whole-genome phylogenetic analysis and network analysis based on single nucleotide polymorphisms (SNPs), we found highly interrelated non-pathogenic and pathogenic E. coli strains with phylogenetic intermixing, and a high prevalence of shared multidrug resistance profiles amongst livestock, human and environment. Through an original data processing pipeline which combines omics, machine learning, gene sharing network and mobile genetic elements analysis, we investigated the resistance to 26 different antimicrobials and identified 361 genes associated to antimicrobial resistance (AMR) phenotypes; 58 of these were known AMR-associated genes and 35 were associated to multidrug resistance. We uncovered an extensive network of genes, correlated to AMR phenotypes, shared among livestock, humans, farm and slaughterhouse environments. We also found several human, livestock and environmental isolates sharing closely related mobile genetic elements carrying ARGs across host species and environments. In a scenario where no consensus exists on how antibiotic use in the livestock may affect antibiotic resistance in the human population, our findings provide novel insights into the broader epidemiology of antimicrobial resistance in livestock farming. Moreover, our original data analysis method has the potential to uncover AMR transmission pathways when applied to the study of other pathogens active in other anthropogenic environments characterised by complex interconnections between host species.

Livestock have been suggested as an important source of antimicrobial-resistant (AMR) Escherichia coli, capable of infecting humans and carrying resistance to drugs used in human medicine. China has a large intensive livestock farming industry, poultry being the second most important source of meat in the country, and is the largest user of antibiotics for food production in the world. Here we studied antimicrobial resistance gene overlap between E. coli isolates collected from humans, livestock and their shared environments in a large-scale Chinese poultry farm and associated slaughterhouse. By using a computational approach that integrates machine learning, whole-genome sequencing, gene sharing network and mobile genetic elements analysis we characterized the E. coli community structure, antimicrobial resistance phenotypes and the genetic relatedness of non-pathogenic and pathogenic E. coli strains. We uncovered the network of genes, associated with AMR, shared across host species (animals and workers) and environments (farm and slaughterhouse). Our approach opens up new avenues for the development of a fast, affordable and effective computational solutions that provide novel insights into the broader epidemiology of antimicrobial resistance in livestock farming.

Escherichia coli Data-Driven Strain Design Using Aggregated Adaptive Laboratory Evolution Mutational Data

Microbes are being engineered for an increasingly large and diverse set of applications. However, the designing of microbial genomes remains challenging due to the general complexity of biological systems. Adaptive Laboratory Evolution (ALE) leverages nature’s problem-solving processes to generate optimized genotypes currently inaccessible to rational methods. The large amount of public ALE data now represents a new opportunity for data-driven strain design. This study describes how novel strain designs, or genome sequences not yet observed in ALE experiments or published designs, can be extracted from aggregated ALE data and demonstrates this by designing, building, and testing three novel Escherichia coli strains with fitnesses comparable to ALE mutants. These designs were achieved through a meta-analysis of aggregated ALE mutations data (63 Escherichia coli K-12 MG1655 based ALE experiments, described by 93 unique environmental conditions, 357 independent evolutions, and 13 957 observed mutations), which additionally revealed global ALE mutation trends that inform on ALE-derived strain design principles. Such informative trends anticipate ALE-derived strain designs as largely gene-centric, as opposed to noncoding, and composed of a relatively small number of beneficial variants (approximately 6). These results demonstrate how strain design efforts can be enhanced by the meta-analysis of aggregated ALE data.

Forecasting of phenotypic and genetic outcomes of experimental evolution in Pseudomonas protegens

Experimental evolution with microbes is often highly repeatable under identical conditions, suggesting the possibility to predict short-term evolution. However, it is not clear to what degree evolutionary forecasts can be extended to related species in non-identical environments, which would allow testing of general predictive models and fundamental biological assumptions. To develop an extended model system for evolutionary forecasting, we used previous data and models of the genotype-to-phenotype map from the wrinkly spreader system in Pseudomonas fluorescens SBW25 to make predictions of evolutionary outcomes on different biological levels for Pseudomonas protegens Pf-5. In addition to sequence divergence (78% amino acid and 81% nucleotide identity) for the genes targeted by mutations, these species also differ in the inability of Pf-5 to make cellulose, which is the main structural basis for the adaptive phenotype in SBW25. The experimental conditions were changed compared to the SBW25 system to test if forecasts were extendable to a non-identical environment. Forty-three mutants with increased ability to colonize the air-liquid interface were isolated, and the majority had reduced motility and was partly dependent on the Pel exopolysaccharide as a structural component. Most (38/43) mutations are expected to disrupt negative regulation of the same three diguanylate cyclases as in SBW25, with a smaller number of mutations in promoter regions, including an uncharacterized polysaccharide synthase operon. A mathematical model developed for SBW25 predicted the order of the three main pathways and the genes targeted by mutations, but differences in fitness between mutants and mutational biases also appear to influence outcomes. Mutated regions in proteins could be predicted in most cases (16/22), but parallelism at the nucleotide level was low and mutational hot spot sites were not conserved. This study demonstrates the potential of short-term evolutionary forecasting in experimental populations and provides testable predictions for evolutionary outcomes in other Pseudomonas species.

Genome-Scale Metabolic Models and Machine Learning Reveal Genetic Determinants of Antibiotic Resistance in Escherichia coli and Unravel the Underlying Metabolic Adaptation Mechanisms

Antimicrobial resistance (AMR) is becoming one of the largest threats to public health worldwide, with the opportunistic pathogen Escherichia coli playing a major role in the AMR global health crisis. Unravelling the complex interplay between drug resistance and metabolic rewiring is key to understand the ability of bacteria to adapt to new treatments and to the development of new effective solutions to combat resistant infections. We developed a computational pipeline that combines machine learning with genome-scale metabolic models (GSMs) to elucidate the systemic relationships between genetic determinants of resistance and metabolism beyond annotated drug resistance genes. Our approach was used to identify genetic determinants of 12 AMR profiles for the opportunistic pathogenic bacterium E. coli. Then, to interpret the large number of identified genetic determinants, we applied a constraint-based approach using the GSM to predict the effects of genetic changes on growth, metabolite yields, and reaction fluxes. Our computational platform leads to multiple results. First, our approach corroborates 225 known AMR-conferring genes, 35 of which are known for the specific antibiotic. Second, integration with the GSM predicted 20 top-ranked genetic determinants (including accA, metK, fabD, fabG, murG, lptG, mraY, folP, and glmM) essential for growth, while a further 17 top-ranked genetic determinants linked AMR to auxotrophic behavior. Third, clusters of AMR-conferring genes affecting similar metabolic processes are revealed, which strongly suggested that metabolic adaptations in cell wall, energy, iron and nucleotide metabolism are associated with AMR. The computational solution can be used to study other human and animal pathogens.

IMPORTANCEEscherichia coli is a major public health concern given its increasing level of antibiotic resistance worldwide and extraordinary capacity to acquire and spread resistance via horizontal gene transfer with surrounding species and via mutations in its existing genome. E. coli also exhibits a large amount of metabolic pathway redundancy, which promotes resistance via metabolic adaptability. In this study, we developed a computational approach that integrates machine learning with metabolic modeling to understand the correlation between AMR and metabolic adaptation mechanisms in this model bacterium. Using our approach, we identified AMR genetic determinants associated with cell wall modifications for increased permeability, virulence factor manipulation of host immunity, reduction of oxidative stress toxicity, and changes to energy metabolism. Unravelling the complex interplay between antibiotic resistance and metabolic rewiring may open new opportunities to understand the ability of E. coli, and potentially of other human and animal pathogens, to adapt to new treatments.

Biocide-Induced Emergence of Antibiotic Resistance in Escherichia coli

Biocide use is essential and ubiquitous, exposing microbes to sub-inhibitory concentrations of antiseptics, disinfectants, and preservatives. This can lead to the emergence of biocide resistance, and more importantly, potential cross-resistance to antibiotics, although the degree, frequency, and mechanisms that give rise to this phenomenon are still unclear. Here, we systematically performed adaptive laboratory evolution of the gut bacteria Escherichia coli in the presence of sub-inhibitory, constant concentrations of ten widespread biocides. Our results show that 17 out of 40 evolved strains (43%) also decreased the susceptibility to medically relevant antibiotics. Through whole-genome sequencing, we identified mutations related to multidrug efflux proteins (mdfA and acrR), porins (envZ and ompR), and RNA polymerase (rpoA and rpoBC), as mechanisms behind the resulting (cross)resistance. We also report an association of several genes (yeaW, pyrE, yqhC, aes, pgpA, and yeeP-isrC) and specific mutations that induce cross-resistance, verified through mutation repairs. A greater capacity for biofilm formation with respect to the parent strain was also a common feature in 11 out of 17 (65%) cross-resistant strains. Evolution in the biocides chlorophene, benzalkonium chloride, glutaraldehyde, and chlorhexidine had the most impact in antibiotic susceptibility, while hydrogen peroxide and povidone-iodine the least. No cross-resistance to antibiotics was observed for isopropanol, ethanol, sodium hypochlorite, and peracetic acid. This work reinforces the link between exposure to biocides and the potential for cross-resistance to antibiotics, presents evidence on the underlying mechanisms of action, and provides a prioritized list of biocides that are of greater concern for public safety from the perspective of antibiotic resistance.

SIGNIFICANCE STATEMENT

Bacterial resistance and decreased susceptibility to antimicrobials is of utmost concern. There is evidence that improper biocide (antiseptic and disinfectant) use and discard may select for bacteria cross-resistant to antibiotics. Understanding the cross-resistance emergence and the risks associated with each of those chemicals is relevant for proper applications and recommendations. Our work establishes that not all biocides are equal when it comes to their risk of inducing antibiotic resistance; it provides evidence on the mechanisms of cross-resistance and a risk assessment of the biocides concerning antibiotic resistance under residual sub-inhibitory concentrations.

Reflections on the Predictability of Evolution: Toward a Conceptual Framework

Evolution is generally considered to be unpredictable because genetic variations are known to occur randomly. However, remarkable patterns of repeated convergent evolution are observed, for instance, loss of pigments by organisms living in caves. Analogous phenotypes appear in similar environments, sometimes in response to similar constraints. Alongside randomness, a certain evolutionary determinism also exists, for instance, the selection of particular phenotypes subjected to particular environmental constraints in the “evolutionary funnel.” We pursue the idea that eco-evolutionary specialization is in some way determinist. The conceptual framework of phenotypic changes entailing specialization presented in this essay explains how evolution can be predicted. We also discuss how the predictability of evolution could be tested using the case of metabolic specialization through gene losses. We also put forward that microorganisms could be key models to test and possibly make headway evolutionary predictions and knowledge about evolution.

Accelerated knowledge discovery from omics data by optimal experimental design

How to design experiments that accelerate knowledge discovery on complex biological landscapes remains a tantalizing question. We present an optimal experimental design method (coined OPEX) to identify informative omics experiments using machine learning models for both experimental space exploration and model training. OPEX-guided exploration of Escherichia coli’s populations exposed to biocide and antibiotic combinations lead to more accurate predictive models of gene expression with 44% less data. Analysis of the proposed experiments shows that broad exploration of the experimental space followed by fine-tuning emerges as the optimal strategy. Additionally, analysis of the experimental data reveals 29 cases of cross-stress protection and 4 cases of cross-stress vulnerability. Further validation reveals the central role of chaperones, stress response proteins and transport pumps in cross-stress exposure. This work demonstrates how active learning can be used to guide omics data collection for training predictive models, making evidence-driven decisions and accelerating knowledge discovery in life sciences.

How to design experiments that accelerate knowledge discovery on complex biological landscapes remains a tantalizing question. Here, the authors present OPEX, an optimal experimental design method to identify informative omics experiments for both experimental space exploration and model training.

Predicting the decision making chemicals used for bacterial growth

Predicting the contribution of media components to bacterial growth was first initiated by introducing machine learning to high-throughput growth assays. A total of 1336 temporal growth records corresponding to 225 different media, which were composed of 13 chemical components, were generated. The growth rate and saturated density of each growth curve were automatically calculated with the newly developed data processing program. To identify the decision making factors related to growth among the 13 chemicals, big datasets linking the growth parameters to the chemical combinations were subjected to decision tree learning. The results showed that the only carbon source, glucose, determined bacterial growth, but it was not the first priority. Instead, the top decision making chemicals in relation to the growth rate and saturated density were ammonium and ferric ions, respectively. Three chemical components (NH₄⁺, Mg²⁺ and glucose) commonly appeared in the decision trees of the growth rate and saturated density, but they exhibited different mechanisms. The concentration ranges for fast growth and high density were overlapped for glucose but distinguished for NH₄⁺ and Mg²⁺. The results suggested that these chemicals were crucial in determining the growth speed and growth maximum in either a universal use or a trade-off manner. This differentiation might reflect the diversity in the resource allocation mechanisms for growth priority depending on the environmental restrictions. This study provides a representative example for clarifying the contribution of the environment to population dynamics through an innovative viewpoint of employing modern data science within traditional microbiology to obtain novel findings.