The Benefits of Mycorrhizae are Frequency-Dependent: A Case Study With a Non-Mycorrhizal Mutant of Pisum Sativum

ABSTRACTMutualisms are remarkably common in the plant kingdom. The mycorrhizal association which involves plant roots and soil fungi is particularly common, and found among members of the majority of plant families. This association is a resource-resource mutualism, where plants trade carbon-based compounds for nutrients, such as phosphorus and nitrogen, mined by the fungi.Evolutionary models usually assume that a mutation grants a small number of individual plants the ability to associate with mycorrhizal fungi, and that this subsequently spreads through the population resulting in the evolution of mutualism. This frequency-dependent hypothesis has been difficult to test, because it is rare to have members of the same species that are capable and incapable of forming the mutualism.Here we describe the results of an experiment that took advantage of a mutant pea (Pisum sativum L. R25) that is incapable of forming mycorrhizal (or rhizobial) associations, and differs from the wildtype (P. sativum cv. Sparkle) by a single recessive Mendelian allele (Pssym8). We grew each genotype either alone or in every combination of pairwise mixed- or same-genotype. We also present an evolutionary matrix game, which we parameterize from the experimental 15N results, that allows us to estimate the costs and benefits of the mutualism.We find that there was no difference between R25 and WT when grown with a competitor of the same genotype, but when R25 and WT compete, WT has a significant fitness advantage. From the model, we estimate that the benefit in units of fitness (g pod mass) obtained from direct plant nitrogen uptake is 22.2 g, and mycorrhizae increase this by only 0.6 g. The costs of plant nitrogen uptake are 9.4 g, while the cost of trade with mycorrhizae is 0.1g.From the model and experiment, we conclude that this relatively small cost-benefit ratio of the mycorrhizal association is enough to drive the evolution of mutualism in frequency-dependent selection. However, without the mutant R25 genotype we would not have been able to draw this conclusion. This validation of frequency-dependent evolutionary models is important for continued theoretical development.