Characterizations of botanical attractant of Halyomorpha halys and selection of relevant deorphanization candidates via computational approach

Halyomorpha halys has been recognized as a global cross-border pest species. Along with well-established pheromone trapping approaches, there have been many attempts to utilize botanical odorant baits for field monitoring. Due to sensitivity, ecological friendliness, and cost-effectiveness for large-scale implementation, the selection of botanical volatiles as luring ingredients and/or synergists for H. halys is needed. In the current work, botanical volatiles were tested by olfactometer and electrophysiological tests. Results showed that linalool oxide was a potential candidate for application as a behavioral modifying chemical. It drove remarkable attractiveness toward H. halys adults in Y-tube assays, as well as eliciting robust electroantennographic responsiveness towards antennae. A computational pipeline was carried out to screen olfactory proteins related to the reception of linalool oxide. Simulated docking activities of four H. halys odorant receptors and two odorant binding proteins to linalool oxide and nerolidol were performed. Results showed that all tested olfactory genes were likely to be involved in plant volatile-sensing pathways, and they tuned broadly to tested components. The current work provides insights into the later development of field demonstration strategies using linalool oxide and its molecular targets.


Materials and methods
Insects. Nymphs and adults of H. halys were obtained from laboratory colonies at the Institute of Plant Protection and Agro-products Safety, Anhui Academy of Agricultural Sciences, Hefei. They were continuously reared on a diet of organic green beans (Phaseolus vulgaris L.) and corn (Zea mays L.) in rearing cages (60 × 60 × 60 cm) at 25 ± 1 °C, 65 ± 5% RH and 16 L: 8 D photoperiod. Insects were fasted for 1-2 h prior to the tests. Olfactometer assay. Olfactometer assays were done using a Y-tube system which was previously described 17 . Parameters for the Y-tube were: stem length at 30 cm, arm length at 20 cm, stem diameter at 3 cm, arm diameter at 2.5 cm, and arm angle at 90°. Airflow was constantly fixed at 0.5 L/min with purification and humidify done by successive connections to activated carbon and double distilled water. Laddered solutions were done by mixing standard chemicals with n-hexane solvent. The control arm was set by providing the same volume of n-hexane solvent to compare with each treatment. Single H. halys adult was introduced from the end of the stem tube and allowed to choose within 5 min during each trial. All tests were done during scotophase under infrared light at 25 ± 2 °C and 40-60 RH. A total of 30 replicates were done for each chemical at each dosage toward each gender. H. halys to linalool oxide. Each H. halys antenna was prepared following standard procedures by cutting the tip and base of the antenna and immediately mounting the excised antenna between two ends of a recording probe (Ockenfels SYNTECH GmbH, Buchenbach, Germany). The other end of the recording probe was directly connected via an interface box to a signal acquisition interface board (IDAC 2; Ockenfels SYNTECH GmbH, Buchenbach, Germany). Stimulations were manually driven by a gas stimulator (CS-55, Ockenfels SYNTECH GmbH, Buchenbach, Germany). Linalool oxide was tested at dosages of 1 μg, 10 μg, and 100 μg, respectively, and n-hexane was used as control. Each stimulation record contained successive measurements of air-controltreatment-control-air. Continuous air flow was set at 150 mL/min, and stimulate flow velocity was 20 mL/min for 0.1 s. n-Hexane (95%, Sigma-Aldrich, St. Louis, MO, USA) was used as control. One μg/μl (E)-2-decenal (95%, Sigma-Aldrich, St. Louis, MO, USA) and the measure dosage was 10 μg, and the measure order of antenna was air-n-hexane-(E)-2-decenal. The replication was 10, direct voltage was 2 mv, continuous flow velocity was 150 ml/min, stimulate flow velocity was 20 ml/min, stimulation time was 0.1 s, and stimulus intervals was 10 s. Raw data of voltages were transferred by: Relative response value (mV) = sample response value (mV)-control response value (mV) before statistical analysis was done.

Chemicals. Synthetic standard chemicals used within the study for bioassays and electrophysiological
Olfactory gene characterization. Gene

Simulated molecular docking.
Docking studies were done to predict binding of selected H. halys olfactory proteins toward linalool oxide and nerolidol, respectively. All known 30 OBPs and 4 ORs of H. halys were SWISS-MODEL-ed and assessed by GMQE and QMEAN values. Specifically, the confidential interval for GMQE was set at 0-1, and QMEAN was set at [− 4, 0]. Higher GMQE values indicated more promising modeling, and lower QMEAN values indicated better binding possibilities of ligands and the selected proteins. For OBPs, a lower cutting threshold of 30% identity was used to initially screen from 30 proteins before docking was done.
Three-dimensional structures of tested volatiles were downloaded from Pubchem (https:// pubch em. ncbi. nlm. nih. gov). Data were transferred to PDB formats via OpenBabel V3.0 66 before energies of ligand structures were minimized by Molecular Operating Environment (MOE; CCG ULC., Montreal, Canada). The docking algorithm was conducted by using AutoDockTools (ADT) 67 . Docking affinities of selected proteins to ligands were automatically evaluated by ADT. Docking results were visualized and exported as vector images by PyMol (The PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC.) before edited in Adobe Illustrator CS6 software (Adobe, San Jose, CA, USA).
Statistics and data processing. Comparison of means was done using SPSS with GLM and Tukey HSD at α = 0.05. Counts data was compared with Chi-square test at α = 0.05. All statistics were carried out using IBM SPSS Statistics 22.0.0 (SPSS, Chicago, IL, USA). Bar and plot charts were developed using Prism 5 for Windows ver. 5.01 (GraphPad software, San Diego, CA, USA). Correlation matrix was developed with Statgraphics Centurion XVII (Statpoint Technologies, Inc., VA, USA). Data accessibility. All described data in this work have been included in the manuscript and online supplementary materials (Table S1-S3, Dataset S1).

Results
Behavioral valence of H. halys to selected allelochemicals. Within all four tested chemicals, linalool oxide elicited significant choice behaviors at the lowest dosages ( Fig. 1). Both female and male adults of H. halys chose linalool oxide in the Y-tube assays when the chemical was applied at 1 µg and 10 µg. While higher dosages of 100 µg drove contrary choice results in both genders, reflecting the potential repellency role of linalool oxide at high dosages. Other tested chemicals did not drive observable choice behaviors until they were applied at 40 µg dosages. Furthermore, linalool oxide at 10 μg and nerolidol at 40 μg stimulated significant gender-biased behavioral preferences, showing that female H. halys adults were more sensitive than males under these tested dosages (Fig. 1). In sum, H. halys adults are most sensitive toward linalool oxide within all tested chemicals.
Electrophysiological responses of H. halys antennae toward linalool oxide. The resolutions of EAG tests were lower for linalool oxide compared with the behavioral assays ( Fig. 2A). Female antennae showed responsiveness to linalool oxide at 10 µg dosages and the responding level increased dramatically along with the increase in dosage. On the other hand, male antennae only started to respond to linalool oxide when applied at 100 µg ( Fig. 2A). Overall, gender bias was observed in the EAG assays, as also shown in the Y-tube assays that females were more sensitive than male H. halys adults (Figs. 1, 2A 68,69 . The selected HhalOR24a, HhalOr45b, HhalOr82a, and HhalOr4-like were separated into four major clusters. Among them, HhalOr82a was more similar with lepidopteran ORs other than hemipteran ones (Fig. 2B). Structural predictions of these four proteins all showed a representative 7-TMD structure of insect ORs (Fig. 2C), and they were able to form the tetramer structure which was proved to be the functional basis of ORs. When the four ORs were aligned with reference ORs, it showed that they were not conserved in the N-part. However, much more conservation was observed during the C-end (Fig. 2D). Since all referred insect ORs to linalool oxide were broadly tuning ORs toward plant odorants, we carried out docking simulations using the four HhalORs in order to investigate H. halys attractiveness to this botanical volatile at the molecular level.
Simulated docking to linalool oxide and nerolidol. Simulated docking studies were conducted using HhalOR4-like, HhalOR24a, HhalOR45b, and HhalOR82a, comparing with the parameters of HarmOR12 ( Fig. 3A-F, Table S2). The tested ligands were the tetrahydrofuran linalool oxide and the sesquiterpene alcohol nerolidol (Fig. 3A). The H. armigera OR12 was checked for binding affinity to linalool oxide by stimulated docking as reference. It showed that binding energies for HarmOr12 with nerolidol was − 4.25, and it was − 3.26 with linalool-oxide (Fig. 3B, Table S 2). Similar results were observed for HhalOR24a, HhalOR45b, and www.nature.com/scientificreports/ HhalOR82a, which presented lower binding energies for nerolidol than for linalool-oxide ( Fig. 3D-F, Table S 2). While HhalOR4-like exhibited better predicted binding affinity with linalool-oxide (− 3.22) than nerolidol (− 2.97) (Fig. 3C). Among, HhalOR82a was predicted to be the best matched receptor for nerolidol with binding energy at − 4.49. For linalool-oxide, the best predicted receptors were HhalOR4-like and HhalOR82a, both with a binding energy at − 3.22 (Table S 2). In order to draw a promising conclusion, investigations on OBPs were also done with similar protocols and algorithms. However, among 30 HhalOBPs, we only identified two qualified OBPs namely HhalOBP8 and HhalOBP30 which had > 30% identities with reference model (Chrysopa pallens OBP4: 6jpm.1.A; CpalOBP4). Both HhalOBP8 and HhalOBP30 were typical 6-C OBPs as CpalOBP4, and the identities for comparing with CpalOBP4 were 36.97% for HhalOBP8 and 33.04% for HhalOBP30, respectively (Fig. 3G). We have observed that binding energies for HhalOBP8 with nerolidol was − 4.81, and it was − 3.71 for HhalOBP30 with nerolidol. For another ligand linalool oxide, simulated binding energies were − 3.17 with HhalOBP8 and HhalOBP30, respectively (Fig. 3H,I, Table S3). Due to the results, it was suggested that both HhalOBP8 and HhalOBP30 could be general OBPs which were involved in plant odorant sensation. HhalOBP8 had better binding potential to nerolidol than HhalOBP30 did.   62,73,74 . Future studies may involve identifications of key host cues, their molecular targets, and reliable implementation methods, e.g., the RNA interference approach 75 . Furthermore, the mechanism by which key volatile signals are coded through central olfactory systems of this stink bug is still unknown and a fascinating area to explore.

Linalool oxide as additional ingredient for odorant bait. The tetrahydrofuran linalool oxide was
reported to have enriched botanical bait formulations, and originated from plant volatiles/essence 76,77 . Lures containing linalool oxide have been widely applied for trapping insects including moths, mosquitoes, and beetles 47,48,78 . As a common emission from natural botanical products, linalool oxide was found to be related to plant injury 77,78 . The secondary metabolite role of linalool oxide has implied its potential functions in plant defense 79 . In fact, this component has also been used as a control reagent for houseflies and coffee bugs 80,81 .
For H. halys and other stink bugs, attraction by linalool oxide has not yet been reported. Literature has shown that this species is behaviorally modified by plant essential oils, which have presented this chemical in mixed volatile blends 12 . Some works have reported that selected key ingredients of volatiles may work better than a full spectrum of plant volatile blend 82 . It may support that linalool oxide has the potential to be a vital addition for optimization of current commercial luring recipes for H. halys as this chemical has outperformed other tested components in the Y-tube assays. The results from the current study have raised the possibility that linalool oxide can be used as an ecological insect behavioral modifying chemical in stink bugs, as is revealed in lepidopteran and dipteran species. Future field trials and implementation of fully-established botanical blend recipes for testing the final effectiveness of artificial baiting approaches for H. halys could benefit from screening on more terpenoids, tetrahydrofurans, esters, and aromatics.
Narrowed-down spectra for receptor protein deorphanization. One important move in the field of chemical ecology is to identify functional genes (mostly receptors) for understanding the olfactory reception of selected bioactive odorants, or "deorphanization" 55 . The so-called reversed chemical ecology sought to solve this matter by providing peripheral coding information for later sorting signaling from brain innervated patterns in higher neuropils of insects 83 . However, most of the volatile ligands did not activate all receptors in a species. A sensing spectrum provided by an OR limits the firing pattern of the corresponding odorant sensory neuron and thus influenced behavioral outputs of insects 84  Combinatorial coding by multiple-receptor decision toward allelochemicals. It was intriguing that linalool oxide attracted H. halys with lower to middle range dosages while repelling them when applied at a high dosage. This phenomenon was identified in Drosophila, which employed a sensory switch to drive contrary behaviors in sour and salt reception 87,88 . It is possible that linalool oxide was coded via a combinatorial pathway through the olfactory systems in H. halys, as it has been shown in the vinegar fly. As a common botanical volatile component, linalool oxide was not likely to be involved in the labeled line circuits, which were www.nature.com/scientificreports/ mostly used by insects to decide life-and-death issues 89 . While higher concentrations in the air of linalool oxide could mean damage already done to the plants 77 , and this repellent modality may help this species to balance their populations within the distributed areas 78,90 . This potential ecological significance of dose may be useful for development of linalool oxide-and nerolidol-based push-pull strategies in a more precise way. On the other side, multiple pathways decision system by the insects' olfaction involved several ORs and glomeruli in the neuropil 91,92 . In this work, selected OBPs and ORs shared similar docking affinities with linalool oxide and nerolidol ligands, which also implied that the olfactory reception pathways for these components involved more than one receptor to be functionally operated 69 . Furthermore, tuning spectra for linalool oxide of the H. halys ORs may also be broad, as it has been shown in docking simulation results, indicating that additional botanical behavioral modification chemicals may exist.