Innate positive chemotaxis to paeonal from highly attractive Chinese medicinal herbs in the cigarette beetle, Lasioderma serricorne

Lasioderma serricorne, also known as cigarette beetle, can exploit a wide variety of stored materials as foods, but it is particularly common on tobacco and herbs. This beetle is a dominant pest species of stored Chinese medicinal materials (CMMs) causing high economic damages, making effective control strategies urgently needed. Behavioural manipulation is an important component of Integrated Pest Management. To the best of our knowledge, plant-borne volatile organic compounds (VOCs) have never been explored to develop lures for managing L. serricorne. In this study, the behavioural responses of L. serricorne to VOCs from four selected CMMs (Euphorbia kansui, Aconitum carmichaelii, Eucommia ulmoides and Pinellia ternata) were studied and their components analysed. Then, the olfactory responses of L. serricorne to the most abundant VOC identified in the preferred CMM, i.e., paeonal, was tested. L. serricorne showed significant differences in its preferences for the VOCs from the four CMMs, i.e, E. kansui > A. carmichaelii > E. ulmoides > P. ternata. From the VOCs of E. kansui, A. carmichaelii, E. ulmoides, and P. ternata, 77, 74, 56, and 81 molecules, were identified, respectively. Paeonal (23.5%), junipene (17.2%), hexanal (17.1%), and benzeneacetonitrile (14.0%) were the most abundant, respectively. Since paeonal dominated the VOC spectrum of the most preferred CMM, this compound was selected for further studies. L. serricorne showed significant positive responses to paeonal tested at various doses, with the most attractive ones being 100 μg and 500 μg. Our findings shed light on the olfactory cues routing the food searching behaviour in the cigarette beetle, providing important information on how L. serricorne targets particular CMMs. The high attractiveness of paeonal at low doses tested here may be exploited further to develop novel monitoring and control tools (e.g., lure-and-kill strategies) against this important stored product pest.

were used in these experiments. Air flow was set as 200 mL/min/arm, the air flow was passing through activated charcoal (for purification) and distilled water (for humidification) before entering in each odour chamber and then in each arm of the Y-tube 51 . In total, 50 adults were used to test each odour treatment, with the olfactometer cleaned after each tested beetle, as described by Carpita et al. 52 . The arm positions were reversed to avoid positional bias after 5 individuals were tested, and the Y-tube olfactometer was replaced with a fresh one after 10 beetles were tested. All bioassays were conducted between 9:00 a.m. and 5:00 p.m.
To understand whether there was any intrinsic bias within the olfactometer, 24 runs were carried out with CA passing through each empty chamber and with unmated insects, 24 different L. serricorne adults (2-3 days old) were tested individually. The time required for the insect to make a choice was recorded. If the beetle had not made a choice after 30 min, another beetle was tested. The longest time taken for all the responding insects was recorded and used as the maximum run time in subsequent tests.
To understand whether L. serricorne could distinguish between odour stimuli and CA, 30 runs were carried out with 25 g of A. sinensis (feeding materials used for rearing L. serricorne) in one branch and CA in the other branch. Equal numbers of male and female beetles were tested.
Collection and analysis of volatiles. The collection and analysis of CMM VOCs were conducted as recently detailed by Cao et al. 51 . Different CMMs (each 2.0 g) were placed in glass bottles for 2 h to confine the odour, and then headspace volatiles were collected using a solid-phase microextraction (SPME) fiber. A manual injector and a ~50/30 µm DVB/CAR/PDMS StableFlex fiber head was inserted into the mouth of the bottle. VOCs were extracted for 40 min at 80 °C, then the fibre head was quickly removed and inserted into the injection port of the gas chromatograph (GC) (250 °C, run in splitless sampling mode). The collected VOCs were analyzed by GC-MS (HP6890/5975 C, Agilent Technologies). An apolar chromatographic column (ZB-5MSI 5% phenyl-95% dimethylpolysiloxane 30 m × 0.25 mm, film thickness 0.25 μm) was used. Temperature was programmed to rise from 40 to 255 °C with a 5 °C/min rate, and then maintained for 2 min. The temperatures of the vaporizing chamber, interface, and quadrupole rod were set at 250 °C, 280 °C, and 150 °C, respectively. The chemical identities of the peaks that were mainly present in CMMs were determined by comparing the mass spectra of compounds with those in databases (NIST 2017 and WILEY 275). Moreover, the coherence of the temperature-programmed retention indices (RIs) with respect to those recorded in Adams 53 and NIST 17 54 was used as an additional criterion for peak assignment. Spectra and retention times were also compared with those of authentic standards, which were purchased from Sigma-Aldrich, Germany. However, the tentative identification of compounds was not confirmed in this study. Additional parameters were as follows: delay time of solvent, 1.0 min; ion source, EI; ionization potential, 70 eV; emission current, 34.6 μA; voltage of the multiplier, 1671 V; and scanning from 29 to 500 atomic mass units.
Behavioural responses of L. serricorne to paeonal from E. kansui VoCs. The VOC mixture from E. kansui was the most attractive to L. serricorne, as assessed by the Y-tube olfactometer bioassays. Therefore, since paeonal was the most abundant compound identified from the VOCs of E. kansui, the behavioral responses of L. serricorne to paeonal were tested as follows.
Y-tube bioassays. The Y-tube olfactometer described above was also used here to test the olfactory responses of L. serricorne to paeonal. Mineral oil was used as a control, a test (at the concentrations detailed above, ranging from 0.1 to 100 μg μl −1 ) or control (10 μl of paeonal solution or mineral oil, respectively) stimulus was adsorbed onto a filter paper disk (1.0-cm diameter) placed in the odor chamber. L. serricorne beetles can choose between a specific dose of the tested stimulus (1, 10, 100, 500, and 1000 μg) and the mineral oil. Bioassays were conducted using the method detailed above; 2-3 days old unmated L. serricorne adults were tested. In total, 50 adults were tested per each cue and compared with mineral oil.
Six-arm olfactometer bioassays. The behavioural responses of adult L. serricorne to different doses of paeonal were also evaluated in a six-arm olfactometer in agreement with the method reported in Liu et al. 55 . Briefly, the six-arm olfactometer consisted of a central chamber (12-cm internal diameter) with six arms (6-cm length and 1.5-cm internal diameter), each connected to a glass tube (20-cm length and 1.5-cm diameter) that projected outwards at an equidistance, with 60° angles between pairs of tubes. Each arm was connected through Teflon tubing to a glass vessel, which was used to contain a test or control stimulus (10 μl). Paeonal doses of 1, 10, 100, 500, and 1000 μg were used as test stimuli, and the airflow was set at 200 ml/min to drive the odour source to insects. L. serricorne (2-3 days old adults after emergence that were starved for 3 h) were introduced in groups (150 individuals per group) into the central chamber with a brush. L. serricorne that entered an arm of the olfactometer within 20 min were counted as having made a choice for a particular odour source. The beetles that did not enter an arm within this time were considered 'non-responders' . After each test, the olfactometer was cleaned, dried and the arms were rotated (60°). Bioassays were replicated six times and were carried out between 9:00 a.m. and 5:00 p.m. A 25-W light was also placed in the centre, 60 cm above the chamber to eliminate any light bias. statistical analysis. The null hypothesis that L. serricorne adults showed no preference for either Y-tube arm (a response equal to 50:50) was analysed using a chi-square goodness-of-fit test. Choices made by male and female L. serricorne were examined separately and, if there was no significant difference in the variances between the male and female beetles, then the data were pooled, and the choices were considered, regardless of sex 51,56 . The www.nature.com/scientificreports www.nature.com/scientificreports/ number of insects found in the different arms of the six-arm olfactometer were subjected to Friedman two-way ANOVA by ranks and in the case of significance (P < 0.05) the Wilcoxon signed ranks test was used for separation of means. All statistical analyses were performed using SPSS 18.0 for Windows (SPSS Inc., Chicago, IL, USA).
Human and animal rights. This research did not involve any human participants and/or animals, only the cigarette beetle, L. serricorne.

Results
Behavioural responses of L. serricorne to CMMs-borne volatiles. Y-tube bioassays. When L. serricorne were tested for their relative preferences between the two arms of the Y-tube olfactometer linked to chambers with CA, there was no bias in a series of 24 experiments; 12 beetles chose each arm. The mean time for a beetle to make a choice was 132.3 ± 10.7 s, and the maximum time was 231 s. When L. serricorne were presented to A. sinensis versus CA, 21 females chose A. sinensis and 9 CA, while 19 males chose A. sinensis and 11 CA. Because there was no difference in the response to A. sinensis between male and female L. serricorne, these results were pooled. L. serricorne significantly preferred A. sinensis (40) to CA (20) (χ 2 = 6.67, df = 1, P = 0.01). The mean time for L. serricorne to make a choice was 103.0 ± 6.1 s, and the maximum time was 223 s. Therefore, L. serricorne could distinguish between the CMMs odours and CA, and the maximum run time was 231 s. Thus, the responses of L. serricorne to volatiles were observed for 240 s (4 min) in the Y-tube olfactometer in following test.

Discussion
In Y-tube olfactometer bioassays, L. serricorne adults were strongly attracted by VOCs emitted by all CMMs tested in this study, suggesting a strong plasticity of the insect olfactory system in locating suitable food sources. In the Anobiidae family, attraction to a range of CMMs was also reported for the adults of the drugstore beetle, Stegobium paniceum (L.), and it is probably related to the high polyphagia of these species 51 .
Furthermore, in our behavioural bioassays, L. serricorne adults exhibited significantly different preferences among the four CMMs examined, with the ranking E. kansui > A. carmichaelii > E. ulmoides > P. ternata. Since experiments were carried out in the absence of visual and contact stimuli, this clear ranking preference showed by L. serricorne adults indicated that plant-borne VOCs provide important cues in the selection of a preferred host by this pest. Thus, the results might partially explain why the damage caused by L. serricorne differs in degrees depending on the CMMs species involved 57 . Different stored-product pests, such as Trogoderma granarium Everts (Coleoptera: Dermestidae), Tribolium castaneum Herbst (Coleoptera: Tenebrionidae), C. maculatus, S. zeamais and S. paniceum showed different preferences for the volatiles of several stored products 27,29,51,58 .
Earlier research showed that there were only 13 components identified from the VOCs of E. ulmoides bark, with the most abundant compound nonanal (17.5%) 59 . However, 56 VOCs were detected in E. ulmoides bark in the present study, with hexanal (17.1%) as the most abundant compound. This discrepancy may be linked to the differences among experimental protocols used in the two studies, as well as to different growing conditions and harvesting time of E. ulmoides plant material, which may significantly affect the VOC bouquet, as already noted for other botanical species. Concerning the other three CMMs in our study, GC-MS analysis revealed that P. ternata had the richest volatile profile with 81 components, followed by E. kansui and A. carmichaelii, with 77, and  www.nature.com/scientificreports www.nature.com/scientificreports/ 74 components, respectively. Furthermore, the most abundant compounds varied among the four CMMs species. Thus, the differing olfactory responses of L. serricorne among CMMs species may be related to differences in both the types and air concentrations of volatile components. However, only a limited number of the VOCs emitted by host plants are crucial for host location by phytophagous insects 60,61 . Herein, paeonal -identified from the preferred CMM E. kansui -was attractive to L. serricorne adults at various doses. Furthermore, it showed the highest relative preferences for 100-and 500-μg paeonal doses, indicating that both the types and doses of VOCs from CMMs may influence the beetle olfactory responses 62 . Besides, the attractiveness of junipene, hexanal, benzeneacetonitrile (each was the most abundant compound in one of the other CMMs), or other compounds identified in the study also needs to be further studied using electroantennographic assays, behavioural, and field trapping tests 29,63 . Potential lures could be developed from these tests, which will aid in the development of new safe and effective trapping strategies to monitor and control this pest. Information about the bioactivity of these compounds on other insects are still scarce. However, it is worthy to note that benzeneacetonitrile, identified from the volatiles released by adult mature male Schistocerca gregaria Forskal (Orthoptera: Acrididae), had strong repellent effects on gregarious mature males 64 , outlining that the same VOCs can lead to extremely diverse behavioural responses in insects, therefore potential non-target insects arising from the employ of VOCs for monitoring and control should be always considered 65 .
Besides, it has been also showed that an important parasitoid of L. serricorne, Lariophagus distinguendus Forster (Hymenoptera: Pteromalidae) 66 , mainly rely to the VOCs emitted by adult beetle feces when locating its host, stressing that the potential impact of the VOCs proposed for real-world use should be evaluated also on biocontrol agents of the targeted pest, to shed light on potential behavioural changes 67,68 . In this framework, the attractiveness of VOCs from the CMMs to L. serricorne parasitoids deserves further research. Such information would be useful to boost biocontrol management strategies based on the employ of hymenopteran parasitoids to control L. serricorne attacking CMMs.
In conclusion, there were three key findings arising from this study. First, Y-tube olfactometer bioassays showed that VOCs from all four CMMs strongly attract L. serricorne adults. L. serricorne exhibited a significant preference for E. kansui VOCs among the four CMMs tested. Second, VOCs differed both in the type and relative content of their components among the CMMs, which may account for the different behavioural responses of L. serricorne adults to various CMMs. Finally, L. serricorne showed a significant preference for paeonal tested at various doses, with 100-and-500 μg being the most attractive. These results indicated that this olfactory cue plays an important role for food location in cigarette beetles. From an applied point of view, basic knowledge reported here may be useful -pending proper field evaluation in real-world conditions -to develop novel monitoring and control (e.g., "lure and kill" technique) tools for this pest.
In particular, paeonal identified from E. kansui VOCs could be exploited alone, as an highly effective lure for both sexes, or in combination with sex pheromone to improve current trapping system for male cigarette beetles [34][35][36] . Finally, a further intriguing perspective is to use this food-borne VOC attractant to enhance the effectiveness of selected physical cues (UV black light) exploited for monitoring of L. serricorne beetles 30,69,70 .