Evolution underlies the entirety of Earth’s biodiversity, as all species diverged from the last universal common ancestor billions of years ago. Although able to effect incredible changes over long periods, the need for multiple generations of mutation and competition renders evolution nearly imperceptible, at the timescale of human observation, for all but the most quickly reproducing organisms. Thus microbial adaptation, given microbes’ rapid generation time and enormous population sizes, is perhaps most pressing to understand. This unavoidable evolutionary process facilitates the rise and spread of antibiotic resistance, and frequently countervails attempts to genetically engineer organisms for human purposes.

The bacterium Escherichia coli, easily the most highly studied microbe to date, is an ideal model by which to investigate evolution. With both clinical and biotechnological relevance, a thorough understanding of the adaptive principles governing E. coli evolution is of great importance. In this dissertation I seek to probe the adaptive capabilities of E. coli using custom robotics systems that function as ‘evolution machines.’ Enabled by this automation, adaptive walks along the fitness landscape can be tracked in real-time with experimental throughput, data quality, and environmental control impossible to replicate manually.

I subject E. coli to stressful perturbations and analyze the mechanisms by which it evolves to restore robust growth, using data types such as phenotypic characterization, whole genome sequencing, and transcriptomics. I demonstrate the remarkable adaptive flexibility of E. coli as it readily evolves to tolerate elevated temperatures, altered isotopic composition, rapidly fluctuating growth environments, and even replacement of important native genes with foreign DNA. Overall, these studies establish condition-specific evolutionary responses, general mechanisms for growth rate improvement, and guiding principles for the successful use of laboratory evolution experiments as a tool for biological discovery and engineering.

Much progress has been made in establishing the causality of mutations that occur during adaptive laboratory evolution (ALE) on organism physiology. In contrast, little progress has been made in detailing the mechanisms and overarching principles of evolution that govern the adaptive process. This dissertation describes the following three aims that sought to address the technical and scientific challenges required to better understand adaptive evolution in micro-organisms. First, an improved method to quantify intracellular metabolites from sampling and extraction to acquisition and quantitation was developed and validated. Second, an improved method to measure intracellular fluxes from sampling and extraction to acquisition and modeling was developed and validated. And finally, aims 1 and 2 were applied in order to uncover system level principles and mechanistic level network functions that were required to adapt to major metabolic perturbation.

Hybrid genomes, formed by high accuracy short-reads and long-reads with the ability to span even large repeat regions, allow for analysis of specific mutations that arise in longitudinal isolates of both clinical and lab-evolved samples. Presented here is an analysis of mutations that arise in a clinical isolate of methicillin-sensitive Staphylococcus aureus, Tx0117, which displays a high inoculum effect with cefazolin. The curing of the beta-lactamase gene created strain Tx0117c, which demonstrates an increased resistance to a variety of antimicrobial peptides. Analysis of the acquired mutations using Breseq uncovered five additional genes affected by the curing process, which may contribute to the increased resistance observed. Additionally, this thesis includes a study on automation of copy number variation detection in lab-evolved strains using the ALE system. The implementation of a python function, which takes a map of coverage depth per base position as input, is tested here in its ability to accurately extract amplification events from only short-read data and a reference genome of the parent strain. The results are validated using an automated pipeline which utilizes hybrid genome assemblies of the evolved strains and blastn (version 2.9.0) for identification of local alignments in the genomes. The sample sets analyzed suggest that short-read data alone is sufficient in identifying amplifications by size, depth, and location. Additionally, the new function allows for efficient and accurate analysis of the flanking regions of amplifications and the encoded genes, providing insight into the mechanisms of acquisition of such events.

Overflow metabolism can be explained by a pathway switch behavior in the cell based on the tradeoffs between metabolic efficiency and proteomic efficiency(enzymatic kinetics). This thesis simulates this behavior on the both small-scale and genome-scale model. By sweeping the enzyme kinetics parameters on multiple pathways or reactions, the simulation could potentially approach the experimental results, however, would converge to some bounds beyond. Therefore, hypotheses of some extra constraints, such as the limitation of oxygen uptake ,electron transport chain, as well as the function of motility protein are presented.

Constraint-based modeling of metabolism at the genome scale has existed as a successful field for two decades. However, genome-scale modeling has entered a new era, with models of ever-increasing size which account for every known reaction in the metabolic and gene expression networks, known as ME models. These models are able to integrate for both metabolic costs (operational expenses) and gene expression costs (capital expenses) in a quasi-econometric model of growing bacterial cells. This modeling type presents new challenges but also brings exciting novel capabilities to constraint-based modeling.

An initial challenge due to models of this size and scale is simply software. Existing software tools for constraint-based modeling simply do not scale to handle models at the size of ME models. Additionally, they make assumptions about the model structure which prevent them from modeling the nonlinearities present in ME models. To address this, a new software package, COBRApy was developed to meet these needs. Unfortunately, interoperability between constraint-based modeling frameworks was already difficult, often resulting in different modeling results. To address this shortcoming, a web-based model validator was deployed to aid the constraint-based modeling community in ensuring all models compute identically with different tools. Finally, a new framework was written on top of COBRApy to build and simulate ME models quickly, accurately, and reproducibly.

To exploit the descriptive nature of ME models, new algorithms were developed to gain increased biological insight from modeling simulations. To improve the understanding of the transcriptional regulatory network in E. coli, a method was developed to predict transcription factor activity from ME simulations, which were successfully used to guide experimental design in identifying novel transcription factor binding sites. To improve the ability of the E. coli ME model to directly predict differential gene expression, a novel method was developed to estimate ME model parameters from high-throughput omics data. Finally, the ME model of the E. coli K-12 MG1655 strain was extended to build strain-specific ME models for 40 different E. coli strains.

The microbial production of commodity chemicals is a promising avenue for the development of sustainable processes for the utilization of renewable resources and reducing our dependency on foreign oil. In order to become cost and energy effective, the process must utilize an organism that is optimized for production of a number of reduced by-products from variety of feedstocks. Escherichia coli is one of the most commonly used host organisms for metabolic engineering and overproduction of metabolites due to its metabolic versatility, amenability to genetic manipulation, and the ability to produce a wide variety of reduced by-products such as bio-ethanol and organic acids. E. coli has also been extensively characterized with respect to its metabolic physiology. It is capable of surviving in a variety of environmental conditions, such as oxic and anoxic; however the different growth rates and different secretion products under oxic anc anoxic conditions poses a significant challenge for metabolic engineering processes in which environmental perturbations will influence the outcome of the bio-catalytic process. Therefore, the utilization of the oxygen-independent strain for bio-catalysis eliminates the need for the stringent control over the fermentation environment with respect to oxygenation, thus significantly reducing the cost of the entire bio-catalytic process. Therefore, it is of interest to develop an E. coli strain incapable of oxygen utilization, to be used as a platform strain for metabolic engineering.

Here, I present the work aimed at the (i) development of an oxygen-independent platform strain, (ii) understanding of its physiological behavior, and (iii) utilization of this strain for metabolic engineering applications. Such a strain can be useful for the overproduction of commodity chemicals under various conditions independent of oxygen supply and optimization of anaerobic metabolic engineering designs using adaptive evolution under oxic conditions. The results show that upon the removal of the oxygen-utilization pathway, the ECOM4 (Escherichia coli Cytochrome Oxidase Mutant 4) strain was unable to undergo an aerobic-anaerobic shift and exhibited similar phenotypes under both conditions with D-lactic acid as a sole growth-associated by-product. Moreover, I show that the ECOM4 strain can be used for the overproduction of organic and amino acids from renewable resources.

The advent of single cell profiling technologies has brought unprecedented resolution to cell heterogeneity in key human tissues. A significant remaining challenge is to interpret this cell variation in terms of meaningful functional differences and interactions between cell types. In this study, we perform a metabolic reconstruction-based assessment of transcriptional heterogeneity in the human brain and kidney. We focus specifically on transporters as their expression is associated with both metabolic interactions between cell types and uptake of drugs. We find that: 1) we are able to identify drug-transporter relationships through structural homology between drugs and native transporter substrates, 2) we observe concomitant brain-region specific expression differences in key transporters that may impact therapeutic impact, 3) upstream metabolic genes have expression that correlates with the transporter expression, suggesting that transporter differences are likely physiologically significant, 4) examination of kidney data shows differential expression across different regions and cell types in the human kidney, 5) metabolic reconstructions provide useful interpretation to help understand the significance of single cell gene expression differences. Native metabolic activity of transporters is postulated through expression of metabolic genes with activities that are correlated as determined by metabolic flux modeling. This work illustrates the types of higher order interactions that can be elucidated from single cell profiling data and paves the way to clinical interventions targeting particular cell types and cell interactions.