Geographic variation and coevolution in a fungus-growing ant (Apterostigma dentigerum) quadripartite symbiosis

The fungus-growing ant (Apterostigma dentigerum) is engaged in an obligate mutualism with fungi the ants cultivate as their primary food source. These fungal cultivars are parasitized by an ascomycete fungi (Escovopsis) capable of rapidly consuming the fungal garden. To manage these garden parasites, the ants employ antibiotic-producing actinobacteria (Pseudonocardia), which have the potential to inhibit Escovopsis growth. Coevolutionary theory predicts that host-parasite interactions will vary over geographic space; however, previous work found little variation in the fungal cultivar’s ability to inhibit the most common Escovopsis parasite across Central and South America. Moreover, population genetic structure analyses found no evidence that cultivar and Escovopsis genotypes are tracking one another. These results suggest that fungal cultivars are not co-adapting to Escovopsis parasites. Here, I examine whether the symbiotic actinobacteria (Pseudonocardia) are instead evolving in response to Escovopsis parasites in a coevolutionary arms race. In contrast to cultivar-Escovopsis interactions, Pseudonocardia-Escovopsis bioassay experiments showed that the symbiotic bacteria’s ability to inhibit Escovopsis is highly variable. The bacteria were also significantly better at inhibiting the most common Escovopsis morphotype, whereas less common morphotypes were better at overgrowing the bacteria. Additional bioassay experiments found no evidence for Pseudonocardia-Escovopsis local adaptation among individual ant colonies or across geographic space, even though Pseudonocardia show a degree of population genetic structure across geographic space. Overall, it appears that Pseudonocardia may be functioning as the indirect ‘host’ responsible for defending cultivars against Escovopsis within Apterostigma dentigerum.