Volatile compounds were collected from apple branches (Malus domestica) at different developmental stages, and the antennal response of codling moth females (Cydia pomonella) to these compounds was recorded by electroantennography coupled to gas chromatography. Presence of a range of terpenoid compounds, many of which had antennal activity, was characteristic for volatile collections from branches with leaves, and from small green apples. Nine compounds from branches with leaves and green fruit consistently elicited an antennal response: methyl salicylate, (E)-β-farnesene, β-caryophyllene, 4,8-dimethyl-1,3(E),7-nonatriene, (Z)3-hexenol, (Z,E)-α-farnesene, linalool, germacrene D, and (E,E)-α-farnesene. The bouquet emitted from flowering branches contained in addition several benzenoid compounds which were not found after bloom. Small green apples, which are the main target of codling moth oviposition during the first seasonal flight period, released very few esters. In comparison, fully grown apples released a large number of esters, but fewer terpenoids. The study of apple volatiles eliciting an antennal response, together with a survey of the seasonal change in the release of these compounds, is the first step toward the identification of volatiles mediating host-finding and oviposition in codling moth females.
Keywords: Host plant volatiles; phenology; headspace collection; electroantennography; apple; Malus domestica; codling moth; Cydia pomonella; Tortricidae; Lepidoptera
Codling moth putative PRs and Orco
The complete open reading frames (ORFs) for CpomOR1 (JN836674.1), CpomOR3 (KJ420588), and CpomOrco (JN836672) were sequenced in prior studies10,25. For CpomOR6a, a 306 bp sequence was identified in the codling moth antennal transcriptome10. We performed 5′ RACE-PCR to complete the ORF of 1248 bp (updated to Genbank; accession number JN836671). Then, we compared the ORF sequence with the updated transcript variants found by Walker et al. (2016)11, and determined that this sequence corresponds to CpomOR6a. Analysing the translated protein, seven transmembrane domains with an intracellular localization for the N-terminus were predicted for CpomOR6a, CpomOR3, CpomOR1 and CpomOrco (Fig. 1), reflecting the typical topology of insect ORs27.
Heterologous expression of codling moth putative PRs and Orco
Prior to physiological tests of the codling moth receptors in the HEK293T heterologous expression system, we first established that the receptors were expressed at the cell membrane. Each of the receptors was cloned into the mammalian expression vector pcDNA5/TO with an N-terminal V5 epitope tag. Twelve to 16 h after transfection, the cells were permeabilized and fixed for immunolabeling with an anti-V5 antibody. Untransfected cells were used as a negative control. Distinct labelling of the plasma membranes of transfected cells indicated that the receptors were properly expressed in the system (Fig. 1). The putative PRs were also co-expressed with CpomOrco, resulting in a localization similar to that found when they were expressed alone (Fig. 1).
Physiological properties of the heterologously expressed codling moth putative PRs
The codling moth receptors were co-transfected with a plasmid carrying a gene for nucleus-targeted Blue Fluorescent Protein (BFP) as a control for transfection efficiency (Fig. 2). We began by testing the codling moth putative PRs coexpressed with Orco for functional expression using VUAA1 (acetamide,N-(4-ethylphenyl)-2-[[4-ethyl-5-(3-pyridinyl)-4H-1,2,4-triazol-3-yl]thio]-) or VUAA3 (acetamide,2-[[4-ethyl-5-(4-pyridinyl)-4H-1,2,4-triazol-3-yl]thio]-N-[4-(1-methylethyl)phenyl]-) agonists that are known to activate both heteromeric and homomeric Orco-containing receptor channels from other insect species28. HEK293T cells transfected with CpomOrco alone or CpomOrco in combination with either CpomOR6a, CpomOR3 or CpomOR1 generated calcium signals upon application of VUAA1 or VUAA3 (Fig. 2). Neither un-transfected nor transfected cells were sensitive to the maximum solvent concentration used (DMSO, 1%), suggesting that the agonist effects were not solvent dependent. Additionally, the overall sensitivity of the cells to the agonists correlated with BFP expression (Fig. 2). In the initial series of experiments, we examined whether the activation of ion channels underlies the generation of the agonist dependent calcium signals (Fig. 2). Whole-cell voltage clamp mode was used to record from cells sensitive to VUAAs. As shown for one cell (Fig. 2), at the holding potential −50 mV, VUAA1 activated an inward current (black line) that kinetically preceded the calcium signal (red line) and thus may determine the agonist mediated calcium influx. Interestingly, the kinetics of the calcium responses mediated by the activity of either homomeric CpomOrco or heteromeric CpomOrco + OR complexes were different (Fig. 2). For example, cells transfected with CpomOrco + OR1 could be characterized by faster activation/deactivation kinetics (Fig. 2). Similar results were observed for the two other heteromeric complexes: CpomOrco + OR6a and CpomOrco + OR3 (Fig. 2).
The response amplitude of both individual cells and cell populations depended on the agonist concentration. The population (cumulative) responses were used to generate the agonist concentration dependences (Fig. 2). For quantitative analysis and comparison, the average peak amplitudes of the responses of different cells were normalized to the responses elicited by application of a saturating concentration (1000 μM) of VUAA1 or VUAA3 (Fig. 2). The comparison of the dose-response characteristics obtained for different OR complexes suggests that homomeric CpomOrco is less sensitive to these agonists than its heteromeric counterparts are. Furthermore, the parameters of the calcium responses mediated by the activity of heteromeric CpomOrco + OR complexes are characterized by faster activation/deactivation kinetics (Fig. 2).
Amiloride derivatives (ADs) have been shown to block both heteromeric and homomeric currents during VUAA1 activation29,30. In our experiments with the putative Cpom PRs and Orco, the AD 5-(N-methyl-N-isobutyl) amiloride (MIA, 100 μM) applied extracellularly almost completely blocked the VUAA3 activated calcium signal (Fig. 3). Similar results were observed for both homomeric and all heteromeric complexes. In all cases, the effects were only partially reversible.
We next used whole-cell and outside-out patch-clamp recordings to characterize the basic electrophysiological properties of the CpomOrco + OR1 complex. All cells expressing CpomOrco + OR1 responded to VUAA3 (200 μM), generating an inward current that varied in amplitude from −49 to −1165 pA, with a mean amplitude of −431 ± 54 pA, n = 37. Un-transfected cells were not sensitive to VUAA3 (12 cells tested). The whole-cell currents gradually increased in a stimulus intensity dependent manner (Fig. 4). In all cases, when cells were stimulated multiple times, the responses were characterized by constant amplitudes and stable kinetic parameters, indicative of the ionotropic nature of the receptors under the current experimental conditions. The results of subsequent experiments using outside-out patch-clamp recordings further support these observations (Fig. 4). VUAA3 (200 μM) applied repeatedly to the extracellular surface of membrane patch reversibly increased the membrane current noise likely associated with the activity of ion channels. As found for the whole-cell currents, the VUAA3 activated ion channel noise of membrane patches demonstrated little if any rundown (Fig. 4). Note, the low single channel conductance and fast gating likely make the unitary currents undistinguishable.
Whole-cell recordings were also performed to estimate the selectivity of the Orco-based channel to a selection of inorganic monovalent cations (Fig. 4; see Methods for details). The permeability ratio sequence for the inorganic monovalent cations tested was (PX+/PNa+):
The sequence is consistent with the selectivity sequences previously reported for Orco-based channels from other insects31,32.
Deorphanization of codling moth putative PRs
As the physiological and pharmacological tests demonstrated that both the homomeric and heteromeric PR and Orco complexes were functionally expressed in the HEK293T cell system, we next used the system to look for natural ligands for the receptors. A library of potential codling moth pheromones and synergist compounds (Table 1), as well as additional compounds, such as plant volatiles and volatiles from fermentation and commercial substances (Supplementary Table S1) were screened against CpomOrco + OR6a, CpomOrco + OR3 and CpomOrco + OR1. Treatment with VUAA3 (250 μM) was used as a positive control. In tests with CpomOrco + OR6a, we observed clear activation in response to stimulation with (E,E)-8,10-dodecadien-1-yl acetate ((E,E)-codlemone acetate; Fig. 5). Subsequent experiments were used to calculate an EC50 of 51.84 ± 13.21 μM for codlemone acetate (Fig. 5); however, the amplitude at saturating concentrations (18.91 ± 10.31, dF) was ~27% of the positive control amplitude (69.71 ± 27.29, dF; Fig. 5). Interestingly, compared to the positive control, we observed a long lasting codlemone acetate activation of transfected HEK293T cells, which led to a delayed recovery after stimulation (Fig. 5).
Once we determined the response of CpomOR6a towards codlemone acetate using the HEK293T cell system, in parallel series of experiments we tested activation of CpomOR6a expressed in Drosophila at1 OSNs to a new panel of ligands, including codlemone, pear ester, codlemone acetate and structurally related compounds (Supplementary Dataset S2). As in the HEK293T cell system, the strongest response significantly different from the solvent was elicited by (E,E)-codlemone acetate. However, the receptor was also significantly activated by the isomers (E,Z)-8,10-dodecadien-1-yl acetate and (Z,Z)-8,10-dodecadien-1-yl acetate along with (E)-10-12-dodecadien-1-yl acetate (Fig. 5 and Supplementary Fig. S5). Furthermore, dose response experiments with CpomOR6a towards (E,E)-codlemone acetate demonstrated that the threshold for detection was 1.0 microgram (Fig. 5).
Previously we demonstrated that CpomOR3 expressed in ab3A and at1 OSNs is sensitive to pear ester25, and here we confirmed this result using the HEK293T cell expression system: pear ester [(E,Z)-ED] elicited a response from cells expressing CpomOrco + OR3 (Fig. 6). Furthermore, we found that an analogous ester emitted by pear (methyl (E,Z)-2,4-decadienoate [(E,Z)-MD])33, also activates CpomOrco + OR3. EC50 estimations (EC50HEK-(E,Z)-ED = 453.60 ± 119.6 μM; EC50HEK-(E,Z)-MD = 1082.08 ± 112.8 μM) and dose-response plots (Fig. 6) suggest that CpomOrco + OR3 has a lower specificity for (E,Z)-MD than for (E,Z)-ED. As with CpomOrco + OR6a responses to (E,E)-codlemone acetate, we observed a slow recovery after CpomOrco + OR3 stimulation with (E,Z)-MD (Fig. 6).
For the dose response of CpomOR3 when expressed in Drosophila ab3A OSNs (Fig. 6), a minimum dose of 100 ng loaded in the stimulus cartridge was required to elicit a response significantly different from the solvent for (E,Z)-ED and of 10 μg for (E,Z)-MD (Fig. 6). Application of a different heterologous expression system confirmed CpomOR3 sensitivity to both pear ester and its analogous (E,Z)-MD, with further suggesting lower specificity for (E,Z)-MD than (E,Z)-ED.
In contrast to CpomOR3 and CpomOR6a, we were unable to identify any ligands that activated CpomOrco + OR1. While calcium imaging experiments with VUAA1 and VUAA3 suggest that CpomOR1 is likely functionally expressed in the HEK293T system (Fig. 2), the receptor failed to respond to any of the ligands tested. In agreement with these results, CpomOR1 also failed to respond in Drosophila at1 OSNs to pheromones and synergists (spikes/s = 0.61, n = 5; response minus basal activity, see methods section for detail) or their combinations with codlemone (spikes/s = 0.00, n = 5; Supplementary Dataset S3).