Olfactory specialization in Drosophila suzukii supports an ecological shift in host preference from rotten to fresh fruit

Abstract

It has been demonstrated that Drosophila suzukii is capable of attacking ripening fruit, making it a unique species within a fly family named for their attraction towards the fermentation products associated with rotten fruits, vinegar, and yeast. It also has been hypothesized that D. suzukii is more attracted to the volatiles associated with the earlier ripening stages of fruit development, and in turn, that D. suzukii is less attracted to fermented food resources, especially when compared with D. melanogaster. Here, we demonstrate that D. suzukii and its close relative D. biarmipes are in fact more sensitive to volatiles associated with the fruit-ripening process; however, in choice-assays, both spotted-wing species are more attracted to fermented fruit than to earlier stages of fruit development, which is similar to the behavioral preferences of D. melanogaster, and thus, fruit developmental stage alone does not explain the ecological niche observed for D. suzukii. In contrast, we show that both D. suzukii and D. biarmipes are more attracted to leaf odors than D. melanogaster in behavioral trials. For D. suzukii, this differential behavioral preference towards leaves appears to be linked to β-cyclocitral, a volatile isoprenoid that we show is most likely a novel ligand for the "ab3A" neuron. In addition, this compound is not detected by either of the other two tested fly species.

Fig. 1

Diagnostic response profiles for the olfactory sensory neurons (OSNs) housed within the large basiconic sensilla of the three Drosophila species as measured from single sensillum recordings. Data are presented as spikes per second (+/− SEM), with the largest neuron amplitude being designated A, and B neuron the second largest, and so forth. (a) Response profiles of the 4 OSNs housed in the “ab1” neuron. (Green = “ab1A”; Yellow = “ab1B”; Blue = “ab1C”; Purple = “ab1D”). (b) Response profiles of the 2 OSNs housed in the ab2 neuron. (Green = “ab2A”; Yellow = “ab2B”). (c) Response profiles of the 2 OSNs housed in the ab3 neuron. (Green = “ab3A”; Yellow = “ab3B”)

Fig. 2

Electrohysiological and behavioral responses using fruit chemistry. (a) GC/coupled single sensillum recordings (“ab3” sensillum) with headspace samples from eight distinct stages of fruit development. Headspace collections are shown above (FID) with the respective A and B neuron response for each species shown below (SSR) (N = 3). (Grey = Drosophila melanogaster; Blue = D. biarmipes; Orange = D. suzukii). (b) Trap-capture rates for the three Drosophila species using the stages of strawberry development. An asterisk denotes a significant differences (α = 0.05). Only one stage was significantly different among the species (Green, P = 0.019; two-tailed, paired t-test). No other stages were significantly different among the species (Flower, P = 0.63; White, P = 0.08; Blush = 0.42; Red, P = 0.12, Rotten, P = 0.23). (c) GC-couple electroantennogram (GC/EAD) recordings with the headspace of full red strawberry fruit (N = 3). GC peaks were identified using GC/MS (and confirmed with synthetic standards) as (1) Methyl butyrate, (2) methyl isovalerate, (3) butyl acetate, (4) isopropyl butyrate, (5) isopentyl acetate, (6) 2-butoxy ethanol, (7) methyl hexanaote, (8) ethyl hexanoate, (9) hexyl acetate, (10) linalool, (11) benzyl acetate, (12) methyl salicylate. (d) Dose response curves for the “ab3A” OSN to several compounds identified from strawberry fruit headspace for which D. suzukii is more sensitive than D. melanogaster

Fig. 3

Electrophysiological and behavioral responses using leaf chemistry. (a) GC/coupled electroantennogram recordings (GC/EAD) with leaf headspace (strawberry) (N = 3). Top graph, GC trace (FID) of leaf headspace; bottom graphs, EAD responses (Grey = Drosophila melanogaster; Blue = D. biarmipes; Orange = D. suzukii). Inset figure depicts the D. suzukii-specific response to β-cyclocitral. GC peaks were identified (and confirmed with synthetic standards) as (1) Z-3-hexenol, (2) E-2-hexenol, (3) 1-octen-3-ol, (4) 6-methyl-5-hepten-2-ol, (5) Z-3-hexenyl acetate, (6) E-2-nonenol, (7) phenethyl alcohol, (8) 2-nitrophenol, (9) methyl salicylate, (10) β-cyclocitral, (11) eugenol, (12) β-ionone, (13) unknown. (b) Trap-capture rates of the three Drosophila species using either whole leaf, the fruit compound ethyl hexanoate (EH), or the leaf compound β-cyclocitral (β-Cyclo). An asterisk denotes significant differences at α ≤ 0.05 between the treatment and control or between the species tested, while two asterisks denote significance at α ≤ 0.01 (Two-tailed, paired t-test). (c) GC/coupled single sensillum recordings (GC/SSR) with β-cyclocitral across all large basiconic sensilla of the three Drosophila species (N = 3). (d) Dose response curves (SSR) for each fly species to the best “ab3A” OSN ligands for D. melanogaster (methyl and ethyl hexanoate), as well as to the best ligand for D. suzukii “ab3A” (β-cyclocitral). (e) Microhabitats where the three Drosophila species usually occur, showing separation in preference