Spatial competition of toxin-secreting strains of yeast

Antagonistic interactions are widespread among microbes and can affect the structure and composition of microbial communities. Theoretical models of well-mixed population genetics with antagonistic interactions predict that a stronger competitor can invade a weaker one only if its initial population is larger than a critical inoculum size. In spatially-extended populations, the invasion of one competitor by another can be mapped to a nucleation problem and the invasion is predicted to be successful only above a critical nucleation size. To test these predictions, we have genetically engineered two strains of the baker’s yeast Saccharomyces cerevisiae to release two different toxins, whose production rates we can vary independently. These strains allowed us to study how antagonistic interactions affect the population dynamics and population genetics of spatially-structured populations at different levels of toxin production, i.e. at different relative strengths of the two competitors. We show that, both in spatially-structured and in well-mixed populations, a toxin-producing strain can displace another toxin-producing strain only if the initial inoculum is larger than a critical threshold, even if the invader strain enjoys a selective advantage.