Towards bioengineered ecosystems: Three studies in invasion biology

An important feature of ecosystems is their invasibility, i.e. their resistance to invasion by nonnative organisms. Invasion is a driver of change at scales as small as a bioreactor and as large as the scale of biogeography. Invasion can be either a desired outcome, for example in the use of probiotics to replace healthful gut microflora after a disturbance to the gut microbial ecosystem, or an undesired outcome, for example in the contamination of open culture systems such as algal raceway ponds. Biological invasion theory attempts to predict what makes a community vulnerable to invasion, what characteristics make an invader successful, and what consequences following invasion might arise. In this thesis, a review of recent literature on microbial invasion biology was conducted, and three separate studies in invasion biology were pursued. In Chapter 1, two overarching yet poorly understood factors determing microbial community invasibility were identified: environmental modulation, and community structure. The first factor, community structure, describes the set of species in an ecosystem and how they are trophically and otherwise related. While the majority of work in invasion biology has focused on the relationship between diversity and invasibility, we argue that it is imperative to move beyond simple diversity metrics towards more information-rich descriptors of community structure, by casting traditional ecological concepts like plasticity, redundancy, dormancy, and diversity into the language of networks. The second factor, environmental modulation, describes how communities can alter their own environment, deterring potential invaders by making the environment less consumable or habitable. We argue that this process has received too little attention in invasion biology and that insights can be drawn from the field of biopreservation, which utilizes beneficial microorganisms to preserve food environments from unwanted invaders. Chapter 2 introduces wild sourdough fermentations as a model system for the study of invasion biology through the lens of biopreservation. An experimental investigation of the robustness of traditional sourdough cultures to invasion was performed, exploring the role of environmental modulation of pH in deterring invaders. Two invader organisms were tested for their ability to invade, and a mathematical model was developed describing the bacterial fraction of a wheat flour dough. First, a simple experiment with a laboratory strain of Escherichia coli introduced into a traditionally propagated sourdough culture indicated that the production of organic acids is the primary mechanism of invasion resistance of sourdoughs against E. coli, which is ubiquitous in human environments. Second, experiments with the acidophile Alicyclobacillus acidoterrestris showed that while acid-tolerant invaders exist, the presence of environmental hurdles such as low temperature and high osmolarity are major factors in preventing their establishment in sourdough cultures. Finally, a mathematical model for the growth, metabolic output, and self-inhibition of the dominant sourdough organism Lactobacillus sanfranciscensis in wheat flour dough was developed. This model was able to predict population density, pH, lactate concentration, and maltose concentration in a manner consistent with experimental data. Chapter 3 considers the general problem of invasion in a three-species ecosystem, which was simulated mathematically assuming that the community structure of the ecosystem is completely defined. The aim of this chapter was to determine any relationships between the invasibility of a particular ecosystem and its community structure. We considered the scenario in which a rare species in an ecosystem suddenly becomes abundant as a result of a discrete mutation in one or more ecological parameters describing the system, representing either a sudden change in ecological strategy, an actual mutation, or a sudden change in the environment. The concepts of invasion distance and invasion direction were formulated, the former referring to the magnitude of the change in parameters required for a rare species to become abundant, and the latter referring to the direction of that parameter change. By performing simulations for 44 out of the 138 possible community structures in a simple 3-species ecosystem, we found that the invasion direction varied significantly with community structure but the invasion distance did not. This result set up future work investigating analytically how invasion direction varies with community structure. Chapter 4 introduces a mathematical model for a phage-bacterium ecosystem. From an invasion biology standpoint, either the phage or the bacterium can be considered as an invader, depending on the ecosystem they inhabit and also on one's point of view. A model was developed to describe an ecosystem containing three populations: phage-infected bacteria, uninfected bacteria, and an environmental reservoir of free bacteriophages. Five system parameters were used as inputs to the model: intrinsic host death rate, virus degradation rate, environmental transmission rate, lysis rate, and burst size. The model was set up to capture logarithmic growth of bacteria, as well as both both lytic and lysogenic life stages of the phage. All pairwise relationships between the five system parameters were investigated, by examining two payoff functions plotted as contour surfaces over each pair of parameters. We found that a high burst size, low virus degradation rate, and low host intrinsic death rate always resulted in the hiest payoff to the phage. In contrast, payoffs always reached a maximum when lysis rate and environmental transmission rate took on intermediate values. Moreover, a third payoff, the basic epidemiological reproduction number R0, was shown to be limited in expressivity as it could only capture monotonic payoff relationships. These results were relevant to the classic "trade-off" hypothesis in virulence theory, which states that lysis rate (virulence) and transmission rate are coupled. While no coupling relationship was assumed in our model, the relationship between lysis rate and transmission rate was markedly the most complex out of all pairwise relationships considered, thus validating the substantial amount of interest historically in this relationship as opposed to other relationships between system parameters. Finally, conditions were found for a stable steady-state solution representing a desirable outcome in a phage therapy context, which attempts to control unwanted bacterial populations by using bacterophages as therapeutic agents. These conditions were overlayed upon the phage payoff surfaces, and it was found that the phage therapy stability region was always located at higher lysis rate and host intrinsic death rate values than the payoff optima. Taken together, the work here contributes to the fields of biological and ecological engineering in the following ways. First, the role of ecosystem structure and environmental modulation were elucidated, and gaps in the literature in both of these areas were identified. Principles were then abstracted from both physical (e.g. sourdough fermentation) and simulated systems (e.g. phage-bacterium interaction) towards a greater understanding of robustness towards invasion and evolutionary trade-off relationships, respectively. Finally, several future directions for research were identified, including the generation of mathematical models to describe yeast-bacterium interactions in sourdoughs, further work on the effect of community structure on invasion direction, and additional experimental and theoretical work towards the realization of phage therapy