Floral volatiles in Mimulus: chemical ecology, insect olfaction, genetics, and reproductive isolation

Abstract

Floral diversity is immense, with more than 250,000 species of angiosperms known. The number of phylogenetically diverse floral species that share common characteristics implies a background pattern of selection acting on floral traits. Unique combinations of these floral traits, or "pollination syndromes," are hypothesized to reflect selective pressure imposed by certain classes of pollinators. One flower character in particular, scent, has been hypothesized to operate as an unseen signal to attract certain pollinators, particularly when combined with other signals such as color and shape. However, the contribution of scent in pollinator-mediated selection between sister taxa has nearly always been inferred and rarely directly tested, and pollinator sensory mechanisms that drive attraction often remain unclear. The genus <italic>Mimulus</italic> (Phrymaceae) forms a developing model system for studying floral diversity and pollinator-driven speciation using a combination of genetic tools and field ecology. Two sister species of <italic>Mimulus</italic>, <italic>M. cardinalis</italic> and <italic>M. lewisii</italic>, are pollinated by hummingbirds and bumblebees respectively, and present a unique system in which to examine the sensory mechanisms and signals that might mediate pollinator-driven speciation. Using a combination of scent collection, gas chromatograph-coupled multi-unit recording, and behavioral experimentation, we investigated the role of multiple <italic>Mimulus</italic> volatiles on bumblebees (<italic>Bombus vosnesenskii</italic>). Three key compounds are found at different concentrations in the two species, with notable effects on electrophysiology and behavioral responses by bumblebees. Using linkage mapping and <italic>in vitro</italic> assays, we found that species-specific differences in abundance of these three compounds are due to two terpene synthase genes, which are expressed and functional in the bumblebee-pollinated species but are nonfunctional or not expressed in the hummingbird-pollinated species. RNAi knockdowns of these two terpene synthases in the bumblebee-pollinated species indicated the functional role that these two genes play in species-specific differences. When the RNAi knockdown plants were introduced to free-flying <italic>Bombus impatiens</italic> in a greenhouse experiment, one knockdown (<italic>LIMONENE-MYRCENE SYNTHASE</italic>) had no effect on bumblebee visitation, but the other (<italic>OCIMENE SYNTHASE</italic>) had a significant effect, decreasing bumblebee visitation compared with the wild-type bumblebee-pollinated flowers. By integrating chemistry, electrophysiology, behavior, and genetics, we have identified a significant phenotypic trait (<italic>E</italic>-β-ocimene levels), identified its genetic underpinnings (<italic>OCIMENE SYNTHASE</italic>), and have shown that it has a significant effect on pollination. Together, these results suggest that scent alone may be a sufficient force to drive differential pollinator attraction to sister species, providing a mechanism for speciation and maintenance of reproductive isolation in angiosperm taxa.