Odorant Receptors for Detecting Flowering Plant Cues Are Functionally Conserved across Moths and Butterflies

Abstract

Odorant receptors (ORs) are essential for plant-insect interactions. However, despite the global impacts of Lepidoptera (moths and butterflies) as major herbivores and pollinators, little functional data are available about Lepidoptera ORs involved in plant-volatile detection. Here, we initially characterized the plant-volatile-sensing function(s) of 44 ORs from the cotton bollworm Helicoverpa armigera, and subsequently conducted a large-scale comparative analysis that establishes how most orthologous ORs have functionally diverged among closely related species whereas some rare ORs are functionally conserved. Specifically, our systematic analysis of H. armigera ORs cataloged the wide functional scope of the H. armigera OR repertoire, and also showed that HarmOR42 and its Spodoptera littoralis ortholog are functionally conserved. Pursuing this, we characterized the HarmOR42-orthologous ORs from 11 species across the Glossata suborder and confirmed the HarmOR42 orthologs form a unique OR lineage that has undergone strong purifying selection in Glossata species and whose members are tuned with strong specificity to phenylacetaldehyde, a floral scent component common to most angiosperms. In vivo studies via HarmOR42 knockout support that HarmOR42-related ORs are essential for host-detection by sensing phenylacetaldehyde. Our work also supports that these ORs coevolved with the tube-like proboscis, and has maintained functional stability throughout the long-term coexistence of Lepidoptera with angiosperms. Thus, beyond providing a rich empirical resource for delineating the precise functions of H. armigera ORs, our results enable a comparative analysis of insect ORs that have apparently facilitated and currently sustain the intimate adaptations and ecological interactions among nectar feeding insects and flowering plants.

Keywords: Helicoverpa armigera; Glossata; odorant receptor; phenylacetaldehyde; plant volatile.

Fig. 1.

Responses of Helicoverpa armigera ORs (HarmORs) to plant volatiles by in vitro functional scanning. (A) The heat map was generated based on the mean current response of 28 functional HarmORs to 67 odorants at the dosage of 10⁻⁴ M. HarmORs are ordered based on the hierarchical cluster analysis of the mean responses. The current response was depicted by color intensity. The response values of each OR were acquired by testing four to eight oocytes. Three types of odorants are colored in orange (aromatics), blue (aliphatics), and green (terpenes). (B) Tuning breadth columns of 28 functional HarmORs to 67 odorants at the dosages of 10⁻⁴ M. Columns were generated by putting the largest value in the middle. Kurtosis values (K) were calculated to represent the degree of, where greater K values represent narrow tuning spectra and smaller ones represent broader spectra. The graphs were arranged by the K value of each HarmOR in descending order.

Fig. 2.

Dose–response profiles of the functional HarmORs toward their main ligands. (A) Dose–response curves of the functional HarmORs. The HarmOR repertoire was highly efficient in detecting terpenes and aromatics. Twenty-two functional HarmORs were arranged in an ascending order based on the EC₅₀ value of the most sensitive ligand (listed in supplementary table S2, Supplementary Material online). Only a subset of the main ligands of HarmOR59 is shown. The most sensitive ligand(s) of each HarmOR are listed at the top and the font line color represents its chemical family (terpenes, green; aromatics, orange; aliphatics, blue). The current response values of each HarmOR to their ligands were acquired by testing four to eight oocytes. (B) Principal component analysis of current response values of the HarmOR repertoire to plant volatiles at different doses. The HarmOR repertoire was efficient for sensing certain terpenes and aromatics. In all graphs, vectors quantifying the responses of the 28 ORs to each odorant were projected onto a 3D region. These 3D representations capture 78.34% (10⁻⁴ M), 65.11% (10⁻⁵ M), 68.89% (10⁻⁶ M), 84.52% (10⁻⁷ M) of the variation in the original 28D data set across different odorant concentrations. Colors represent the different chemical classes: orange (aromatics), green (terpenes), and blue (aliphatics).

Fig. 3.

Phylogenetic analysis of ORs from Lepidoptera. The phylogenetic tree was built using the amino acid sequences of 461 ORs from eight species based on the maximum likelihood algorithm. Twenty-three major clades with high node support (bootstrap values >80) were labeled with red or pink dots and blue letters (A–W). The main ligand of each functional HarmORs as well as other reported data were arranged around the phylogenetic tree, and is colored using the same color code as in figure 1 according to its chemical family. Each branch of the OR gene tree was colored differently according to species: red (Helicoverpa armigera, Harm), dark blue (Spodoptera littoralis, Slit), black (Helicoverpa assulta, Hass), sky blue (Bombyx mori, Bmor), green (Manduca sexta, Msex), brown (Ostrinia furnacalis, Ofur), violet (Chilo suppressalis, Csup), and gray (Grapholitha molesta, Gmol). Similarly, the functional identified ORs were colored using the same scheme.

Fig. 4.

Functional differentiation and conservation of orthologous ORs between Helicoverpa armigera and Spodoptera littoralis. (A) Tuning spectra comparison between H. armigera (red) and S. littoralis (blue). The responses of fifteen orthologous OR pairs against 21 overlapping odorants are shown. Twelve of the 15 pairs that share high sequence identities (from 63.33% to 88.49%) present nonoverlapping response spectra. The first three pairs of orthologous ORs showed identical response spectra. The mean response values of each OR were normalized by defining the maximal response as 100. The amino acid sequence identity of each orthologous OR pair is shown in each diagram. (B) Selection pressure analysis on four orthologous ORs clades (A, Orco; H, OR42; G, OR59; M, OR52) in the phylogenetic tree of figure 3. The dN/dS ratios of clades A, H, G, and M are listed behind the branches. The value with dN/dS <1 indicates the genes within one clade have evolved under purifying selection.

Fig. 5.

Unique lineage of OR42 orthologs across moths and butterflies. (A) Orthologous ORs of HarmOR42 in Lepidoptera species constitute a unique lineage with high node support and the functional identified OR42 orthologs in twelve species respond to PAA. The OR42 clade was picked from a phylogenetic tree of 1,619 ORs from 31 species belonging to the Lepidoptera (30 species from 13 families) and the Trichoptera (1 species) orders. The color dots on the nodes indicate bootstrap support values. Ditrysia and Glossata crown groups are highlighted in green and yellow. Functionally characterized orthologous ORs are marked by cyan dots, and those tuned to PAA (phenylacetaldehyde) are marked by red dots. A blue triangle behind EsemOR19 represents its lack of response for PAA, but responsivity for another ligand. The cross for RnubOR13 denotes that this OR is not activated by any of the tested ligands. The asterisk indicates that the function of SlitOR14 for sensing PAA was reported by de Fouchier et al. (2017). (B) Tuning spectra of OR42 orthologs from 11 Lepidoptera species, and one homolog (EsemOR19) from outside the OR42-lineage. All the orthologs in the 11 Lepidoptera species have similar response profiles, which robustly tuned to PAA. The tuning spectra of the 11 orthologs completely overlap based on a screen of six compounds at 10⁻⁴ M. EsemOR19 tuned to cinnamaldehyde, with a very weak current. Families are indicated beneath each histogram. The current response value of each OR42 ortholog was acquired by testing four to eight oocytes.

Fig. 6.

In vivo functional studies indicating the key role of HarmOR42 in sensing floral scent and host plant cues. (A) (Left) Behavioral responses of wild-type and HarmOR42-lacking mutants of Helicoverpa armigera moths to odorant or blends in two-choice olfactometers. Both of the female and male wild-type moths were significantly attracted by the floral scent mixture IAC that mainly contain PAA. The mutants lost their preference to IAC compared with wild type. PAA, phenylacetaldehyde; IAC blend (50% PAA, 20% salicylaldehyde, 10% methyl 2-methoxybenzoate, 10% linalool, and 10% (R)-(+)-limonene); IAC-p, IAC blend minus PAA; Blank, solvent. Asterisks in black font represent the statistical differences in distribution compared with solvent blank groups (*0.01 < P < 0.05; **P < 0.01). Four to six repetitions were performed for each chemical or blend. Thirty moths (either males or females) were tested in each repetition. (Right) Schematic representation of the two-choice olfactometer. (B) CRIPSPR/Cas9-based knock-out of HarmOR42 gene in H. armigera. Target sequences of the two sgRNAs (black) and the PAM sequences (blue) are shown under the exon cluster. The obtained mutation consisted of a 203-bp nucleotide fragment deletion and a 6-bp nonhomologous insertion (green) in the genome. (C) EAG responses of wild-type and HarmOR42 mutant moths to four major ligands (1 μg) of HarmOR42 (Left) and different dosages of the blend IAC (Right). The EAG response of HarmOR42 knock-out mutants to PAA or IAC was significantly decreased compared with wild-type moths. 2P, 2-phenylethanol; HA, heptanal; PAA, phenylacetaldehyde; CA, cinnamaldehyde; PH, Z11-16: Ald. Fifteen repetitions were performed for each odorant or blend. Asterisks represent the statistical difference by Student’s t-test (*0.01 < P < 0.05, **0.001 < P < 0.01, ***P < 0.001). (D) Petri dish behavioral experiments of H. armigera larvae to PAA. The larvae of wild-type H. armigera were strongly attracted by PAA. The HarmOR42 knock-out mutants lost their preference to PAA significantly. Ten repetitions were performed for either wild type or mutants. Ten third instar larvae were used in each repetition. Asterisks represent the statistical difference by Student’s t-test with both strains (*P < 0.05).