WHEN FLOWERS PLAY DEAD: MICROBES AS ARCHITECTS OF A ‘DECEPTIVE’ FLORAL PHENOTYPE

Abstract

Like most entities in the phyllosphere, flowers are universally colonized by microorganisms. Because flowers are the reproductive structures of Angiosperms, microbial residence on or within floral tissues may have the potential to impact floral trait evolution and floral reproductive success by altering floral rewards and influencing pollinator behavior. To date, studies approaching these questions have mostly focused on floral microbes as larcenists of nectar rewards or as potential floral pathogens disruptive to flower-pollinator interactions. Thus, the unexplored potential for floral microbes to fundamentally contribute to extended floral phenotypes or drive evolutionary shifts to new pollinator classes constitutes a major gap in our understanding of the ecology and evolution of floral phenotypes. In this dissertation, I addressed this gap by testing the microbial impact on pollinator attraction and reward production in the flowers of the common pawpaw, Asimina triloba, an understory tree native to the deciduous forests of eastern North America. The small, maroon flowers of A. triloba produce a floral scent that is reminiscent of fermenting fruit or yeasty bread dough. Not surprisingly, the floral scent bouquet is dominated by short-chain aliphatic alcohols, acids, and esters typical of microbial metabolism, known to attract saprophilic beetles (Glischrochilus; Nitidulidae) and drosophilid flies that are more commonly associated with decaying organic substrates, not flowers. To test the hypothesis of microbial-mediated pollinator attraction within A. triloba flowers, I investigated the floral microbiome using both culture-independent and classic microbiological methods. In addition, I performed manipulative experiments including flower sterilization, floral visitor exclusion via micromesh bagging, electroantennographic examination of pollinator chemosensory capabilities, and arthropod trap bioassays in the field. The microbiome results revealed a dynamic pattern of microbial residence that is correlated with increased fermentative volatile production in flowers, while the manipulative experiments and bioassays confirmed that microbial residence on flowers influences the attraction of wild Glischrochilus beetles, the most likely pollinators at my study site. Multiple sources of evidence support the hypothesis that microbial decay within pawpaw flowers represents a unique floral reward to its saprotrophic pollinators, which challenges the notion that the flowers engage in mimicry or deception for their floral advertisement.