Local pollinator adaptation in deceptive Arum maculatum

Visual and olfactory (scent) floral cues are means by which plants communicate with their animal pollinators and are thus essential in plant-pollinator interactions. The relative importance of these cues varies among systems, but specialized pollination systems are typically mediated by floral scents. Although several thousand scent compounds have been described so far, little is known about the individual compounds responsible for pollinator attraction. Floral scents are known to vary within species among populations, but there are still large gaps in our knowledge about the evolutionary forces driving this variability, which may be due to genetic drift and/or local selective adaptation in response to a geographically variable pollinator climate. The deceptive Lords and Ladies, Arum maculatum, with its moth fly pollinators is our model system to identify the mechanisms of pollinator attraction and test whether variability in floral scent is adaptive. This plant attracts and lures specifically two psychodid moth flies (Psychoda phalaenoides and Psycha grisescens) that both oviposit on cow and horse dung, but have different olfactory preferences. In Europe, there is a strong geographical pattern of pollinator climate in that populations from north vs. south of the Alps are mainly pollinated by P. phalaenoides and P. grisescens, respectively, although both insect species occur in either region. Based on recent genetic reports, this north-south pattern matches a phylogeographic subdivision of A. maculatum in DNA fingerprints (AFLPs) but is not reflected in the mitochondrial DNA of the moth flies. Although the inflorescence scent of A. maculatum has been documented to vary intraspecifically, this has never been studied at a population level in relation to the north-south genetic/pollinator pattern. It is thus still unknown whether (1) local populations of A. maculatum from north vs. south of the Alps have adaptively diverged in scent traits to attract the most abundant and/or efficient regional pollinators; or (2) cis- trans Alpine variation in pollinator climate merely reflects local insect abundance rather than differences in scent traits. We use a novel combination of ecological, chemical-ecological and genomic methods, and propose to study local adaptation in floral scents. For the first time, we take the opportunity to distinguish between different molecular signatures of neutral variation versus adaptive change in floral scents. Specifically, we will identify the compounds responsible for pollinator attraction in A. maculatum and study the evolutionary forces driving inter-population variability in floral scents. Obtained data will contribute to our understanding of the evolution of intraspecific variability in floral scents as a category of traits implicated in plant reproductive isolation and speciation.

Floral scents are important means for communication between plants and their animal pollinators. They are known to vary intraspecifically among populations, but there are large gaps in our knowledge about the evolutionary forces driving this chemical variability, which may be due to genetic drift and/or local adaptation in response to a geographically variable pollinator climate. Here, we used the brood-site deceptive lords-and-ladies, Arum maculatum, with its fly pollinators to study mechanisms of pollinator attraction and to test whether variability in floral scent is adaptive. Our study was performed in populations north and south of the European Alps. First, we compared the geographic patterns of pollinators and their local habitat availability (as determined by light trap catches) as well as of the inflorescence scents emitted. Second, we sought to identify the mechanisms of pollinator attraction by investigating which molecule(s) within the blend is (are) responsible for tricking the main pollinator species. Finally, we tested for the involvement of inflorescence scent in local adaptation of A. maculatum. We found that female moth flies (Psychoda phalaenoides) were the predominant visitors north of the Alps, while small dung flies (Sphaeroceridae) dominated the visitor assemblages in the south, followed by females of P. grisescens, male moth flies, and chironomids. Light trap catches north of the Alps consisted mostly of moth flies and few other fly families (e.g., Ceratopogonidae, Chironomidae), while the latter were predominant in those from the south. By comparing floral visitors and light trap catches, we found that local pollinator availability only partly explains geographic differences in visitor assemblages of A. maculatum north vs. south of the Alps. This suggests that floral visitors are also differentially attracted to A. maculatum inflorescences in these two regions. We recorded the highest number of floral scent compounds ever found in a single plant species (289 in total), and detected pronounced differences in scent composition between the two regions. In electrophysiological measurements, we identified a high number of floral volatiles eliciting an antennal response in pollinators, albeit the importance of single compounds in pollinator attraction remains to be assessed. We found different signs of phenotypic selection north vs. south of the Alps, but most of the differences in scent between these regions could not be explained by differential phenotypic selection. Finally, we are assessing whether differentiation in floral scent between the regions is reflected in the genetic structure of A. maculatum, and whether there are genomic signatures of selection associated with scent. Overall, we demonstrated that scent and floral visitors differ between north and south of the Alps, and that some of these regional scent differences is likely adaptive and thus contributes to shaping differential pollinator communities.