The influence of population size on host-parasite coevolution
Host-parasite coevolution is defined as reciprocal adaptation between coexisting hosts and parasites. It associates with strong antagonistic selection, leading to fast changes in fitness-related traits, such as host resistance or parasite virulence, which in turn affects overall host population performance and parasite prevalence. The resulting dynamics of the interaction inherently causes population size fluctuations. Infection outbreaks followed by parasite disappearance, host mass extinctions and periodic oscillations in host and parasite abundance are all examples of changes in population size under natural conditions. These demographic variations enhance stochasticity and affect the process of evolution. Despite a large body of evidence suggesting that fluctuating population size is an inevitable consequence of host-parasite interplay, population size is usually assumed to be constant or infinite (ignoring stochasticity) in current studies on coevolution(Chapter I). The main goal of this thesis is to enhance a more realistic view on host-parasite coevolution by theoretically and experimentally testing the influence of fluctuating population size and associated stochasticity. First, together with colleagues, I examined the consequences of changing population size in a theoretical model by relaxing conventionally made assumptions of infinite and constant population size (Chapter II). We found that fluctuating population size combined with stochasticity dramatically changed host-parasite coevolution dynamics by (i) greatly increasing fixation rates and, therefore, (ii) preventing continuous genotype oscillations, which in case of infinite or constant population size would sustain (in accordance with negative-frequency dependent selection). As an experimental approach, central for this thesis, I carried out an evolution experiment with two interacting model organisms - the nematode Caenorhabditis elegans and the pathogenic bacterium Bacillus thuringiensis (Chapter III). I developed for this purpose a high-throughput protocol, which allowed propagation of many replicate populations for 23 host generations under three different demographic regimes: small populations (increased stochasticity), large populations (“deterministic” situation), and populations periodically forced to bottlenecks (fluctuating population size). After the experiment, I phenotypically characterized evolved host and parasite populations by exposing them to the ancestral antagonist and found the following evolutionary changes in fitness related-traits: (i) an increase in host fecundity, (ii) a decrease in host survival, and (iii) the accumulation of population divergence in parasite virulence. Additionally, I performed a time-shift experiment by confronting coevolved host and parasite populations from three different time points of the coevolution experiment in all possible combinations in order to infer the temporal dynamics of coevolution. The time-shift experiment revealed (iv) a striking pattern of negative frequency-dependent selection providing the first experimental demonstration of this type of dynamics for experimentally coevolved host and parasite. Moreover, (v) negative frequency-dependent selection was found in large populations and only partially in populations subjected to bottlenecks but not in small populations, suggesting that fluctuating population size and increased stochasticity can alter coevolutionary dynamics, in accordance with the results of the theoretical model. Finally, I performed a functional analysis of two toxin genes of B. thuringiensis, which had been identified as candidate genes in a separate evolution experiment, and confirmed their contribution to the pathogenicity of this bacterium (Chapter IV). Taken together, my PhD project emphasizes the selective impact of coevolution on trait evolution in both antagonists, especially under large population sizes.