Transcriptomic analysis of deceptively pollinated Arum maculatum (Araceae) reveals association between terpene synthase expression in floral trap chamber and species-specific pollinator attraction

Abstract

Deceptive pollination often involves volatile organic compound emissions that mislead insects into performing nonrewarding pollination. Among deceptively pollinated plants, Arum maculatum is particularly well-known for its potent dung-like volatile organic compound emissions and specialized floral chamber, which traps pollinators-mainly Psychoda phalaenoides and Psychoda grisescens-overnight. However, little is known about the genes underlying the production of many Arum maculatum volatile organic compounds, and their influence on variation in pollinator attraction rates. Therefore, we performed de novo transcriptome sequencing of Arum maculatum appendix and male floret tissue collected during anthesis and postanthesis, from 10 natural populations across Europe. These RNA-seq data were paired with gas chromatography-mass spectrometry analyses of floral scent composition and pollinator data collected from the same inflorescences. Differential expression analyses revealed candidate transcripts in appendix tissue linked to malodourous volatile organic compounds including indole, p-cresol, and 2-heptanone. In addition, we found that terpene synthase expression in male floret tissue during anthesis significantly covaried with sex- and species-specific attraction of Psychoda phalaenoides and Psychoda grisescens. Taken together, our results provide the first insights into molecular mechanisms underlying pollinator attraction patterns in Arum maculatum and highlight floral chamber sesquiterpene (e.g. bicyclogermacrene) synthases as interesting candidate genes for further study.

Keywords: Chemical ecology; GC–MS; RNA-seq; molecular ecology; plant-pollinator interactions; sesquiterpenes.

Fig. 1.

Overview of the sampling design, bioinformatic pipeline, and differential expression analyses in this study. Inset photo: A. maculatum inflorescence, with the lower spathe chamber dissected to reveal the male and female florets. Inset map: population codes, placed over their approximate locations.

Fig. 2.

Significantly enriched GO terms when comparing A. maculatum appendix (n = 7) and male floret (n = 16) transcript expression during anthesis. Nonredundant GO terms are visualized in semantic similarity space (allowed similarity = 0.8); the full list of GO terms represented above is available in Supplementary Appendix 2. [*] indicates parent groups linked to the biosynthesis of volatile compounds and [^] indicates parent groups linked to pollen production

Fig. 3.

Significantly enriched GO terms when comparing A. maculatum transcript expression in male floret tissue during anthesis (n = 14), vs. approximately 18 h after anthesis (n = 14). Nonredundant GO terms are visualized in semantic similarity space (allowed similarity = 0.8); the full list of GO terms represented above is available in Supplementary Appendix 2. [*] indicates parent groups linked to the biosynthesis of volatile compounds

Fig. 4.

A simplified overview of key proteins involved in producing aromatic amino acid-derived A. maculatum VOCs, and a heatmap of their vst-transformed expression during anthesis. Color scale represents whether expression in a given group (i.e. tissue type and population) is above or below the transcript’s mean expression across all samples. Genes in gray are absent from the A. maculatum transcriptome. Gene names: BEBT = benzyl alcohol O-benzoyltransferase; BSMT, benzoic acid/salicylic acid carboxyl methyltransferase; DHQ, 3-dehydroquinate dehydratase/shikimate dehydrogenase; EGS, eugenol synthase; MKS, methylketone synthase; PAAS, phenylacetaldehyde synthase; TSA/B, tryptophan synthase alpha/beta subunit; ?, unknown gene(s).

Fig. 5.

Heatmap of vst-transformed expression of putative terpene synthases in A. maculatum. Colour scale represents whether expression in a given group (i.e. tissue type, stage of anthesis, and population) is above or below the transcript’s mean expression across all samples.