Pollinator-mediated selection on floral traits is thought to have influenced the evolution and diversification of angiosperms. Orchids exhibit highly specialized pollinator associations that are thought to promote and maintain reproductive isolation between sympatric lineages. However, the mechanisms by which angiosperms adapt to and shift among different pollinators remain poorly understood. Around 10% of neotropical orchid species are pollinated by scent-collecting male euglossine bees. In this system, floral scents simultaneously attract and reward bee pollinators. Here I used male euglossine bee pollinated orchid species from genus Gongora to determine the underlying mechanisms of pollinator attraction and reproductive isolation between sympatric lineages.

In Chapter 1 I examined the variation in volatile organic compounds (VOCs) emitted by the inflorescences of Gongora plants found in two natural populations from Costa Rica. Floral scent varied discretely in both populations with each chemically distinct group (chemotype) attracting a different subset of euglossine species. To test whether these differences in pollinator attraction contribute to reproductive isolation, we genotyped pollen masses recovered directly from male bees caught in the field to reconstruct pollinator networks and perform population genetic analyses. The results of this chapter revealed one population to be structured by chemotype despite varying levels of pollinator overlap. The second population showed little evidence of genetic differentiation between chemotypes despite high degrees of floral trait and pollinator attraction divergence. These findings suggest that natural selection may be driving the observed patterns of divergence in floral phenotypes despite gene flow.

To further explore the genetic basis of floral scent biosynthesis and differentiation, in Chapter 2 I assembled and annotated the reference genome of one of the characterized Gongora chemotypes. Terpenoids are the most common class of compounds found in the scent of male euglossine bee pollinated orchids and their biosynthesis and diversity is mainly mediated by terpene synthases (TPS). I manually annotated a total of 21 TPS genes in Gongora. This number is comparable to the number of TPSs found in other sequenced orchid genomes. Finally, I sequenced individuals from two sympatric chemotypes with low levels of genomic differentiation to explore patterns of genomic differentiation across the genome. Genome-wide differentiation was low with several small and scattered regions showing high levels of differentiation. These results suggest that the genetic architecture might be complex involving many loci of small effect.

In conclusion, our results demonstrate the capacity of floral scent to finely control specific pollinator attraction, pollination network architecture, and their role in mediating reproductive isolation. The genome assembly will serve as the foundation for future research aimed at understanding the evolution of floral scent.