Key message

  • How gustatory habituation affects feeding deterrent effect is less understood in tobacco cutworm.

  • Lemongrass oil was the most effective feeding deterrent, followed by fennel sweet and clove bud.

  • Fennel sweet oil and trans-anethole had feeding deterrence without exhibiting habituation.

  • Larvae exposed to lemongrass oil and citral exhibited habituation effect.

  • Concentration-dependent electrophysiological response was observed in the maxillary palps.

Introduction

Insects employ a variety of sensory mechanisms to locate and identify their foods. The primary senses involved are visual and chemosensory cues, including olfactory and gustatory signals (Anderson and Anton 2014). Gustatory sensilla, key components in the chemosensory system, play a crucial role in encoding the chemical properties of potential food sources, such as nutritional content, minerals, and toxins, before transmitting the information to the central neuron (Ohla et al. 2019). These sensilla are strategically located on different parts of insects, including mouthparts, legs, wing margins, and ovipositors (Dahanukar et al. 2005). In lepidopteran larvae, food perception is primarily facilitated through the maxillary palps on the head capsule and the medial/lateral sensilla styloconica on the maxillary galeae (Schoonhoven and Van Loon 2002; Zhou et al. 2021b). The gustatory receptors are tuned to distinct taste modalities, such as salty, sweet, sour, and bitter (Schoonhoven and Van Loon 2002). When gustatory receptors encounter deterrent substances typically associated with a bitter taste, insects often demonstrate avoidance behaviors as a defensive response (Glendinning et al. 1999; Pontes et al.2014; Zhou et al. 2021a; Chen et al. 2022).

The tobacco cutworm, Spodoptera litura (Fabricius) (Lepidoptera: Noctuidae), is a polyphagous pest attacking more than 300 crop plants. Its habitat ranges throughout tropical and temperate regions of Asia, Australia, and the Pacific Islands (Tuan et al. 2015; Fand et al. 2015). Depending on the crop and larval stage, this species can cause losses ranging from 25.8% to 100% (Dhir et al. 1992). Most control measures for the tobacco cutworms rely on chemical insecticides (Huang et al. 2006; Ahmad 2009; Ramzan et al. 2019). However, resistance problems to synthetic insecticides are emerging in many countries such as Pakistan, China, and India (Ahmad et al. 2007; Tong et al. 2013; Gandhi et al. 2016).

Dealing with insecticidal resistance in the field often results in more pesticide usage, posing risks to humans and the environment (Mallet 1989). This challenge has shifted focus towards Integrated Pest Management (IPM) for more sustainable solutions. The European Union’s ‘Farm to Fork Strategy’ exemplifies this shift, aiming to reduce the use of chemical pesticides by half by 2030 (European Commission 2020). In this context, essential oils are gaining attention as safer alternatives due to their low mammalian toxicity and environmental impact, which can be utilized as biorational pesticides for various purposes, such as attractants, repellents, antifeedants, fumigants, and insecticides (Mossa 2016). Many studies explored essential oils as potential antifeedants (Hummelbrunner and Isman 2001; González-Coloma et al. 2006; Akhtar et al. 2012; Kostić et al. 2021), with neem and its active constituent, azadirachtin, being prominent ones (Bomford and Isman 1996; Aerts and Mordue 1997).

The efficacy of these plant-derived antifeedants may diminish in field environments as the target insects are repeatedly exposed to feeding deterrents, which is called habituation. It is a form of non-associative learning where the response to repetitive stimuli decreases (Rankin et al. 2009), and it can alter the food selection behavior of insects. Taste neuron sensitivity, typically associated with gustatory habituation, involves increased sensitivity upon initial contact with specific plant-derived chemicals (Miles et al. 2005; Chen et al. 2022) but diminishes with prolonged exposure (Zhou et al. 2021a). The occurrence of habituation can vary depending on exposure time, substance, composition, concentration, and insect species. For example, the fifth instar larvae of S. litura did not exhibit any habituation effect on the crude neem extract. Still, they did against the corresponding amount of the single active constituent, azadirachtin (Bomford and Isman 1996). The response to deterrents is generally less pronounced in generalist insects than in specialists (Bernays et al. 2000), highlighting the need for extensive research on feeding deterrents, particularly in generalist species. For instance, a specialist herbivore, Helicoverpa assulta, exhibited habituation to the plant alkaloid strychnine in 48 h, which was faster than the generalist, H. armigera (72 h), with both species showing consistent electrophysiological responses in taste neurons corresponding to the results of bioassays (Zhou et al. 2021b). Despite many previous studies on the habituation of plant secondary metabolites, the mechanism of gustatory habituation in essential oils remains poorly understood.

This study examined the antifeedant and insecticidal activity of 29 commercially available essential oils and explored the habituation effect. The blending effects of essential oils in feeding deterrence and habituation were also examined based on their chemical analysis results. Additionally, electrophysiological studies were conducted to investigate the mechanism of gustatory habituation on three sensilla in the mouthparts of the larvae.

Materials and methods

Test insects

Eggs of the tobacco cutworm were obtained from Crop Protection Center, Farm Hannong Co., Nonsan, South Korea. The colony was maintained successively in an insectary of Seoul National University without exposure to any known pesticides for at least five years under 25 ± 2 °C, 60 ± 5% RH, and a 14:10 h L:D photoperiod. The larvae were reared on an artificial diet based on white bean and wheat bran in an insect breeding dish (⌀100 × 40 mm height). The adults were reared in a 300 × 300 × 300 mm cage and provided with a 10% sugar solution. Cabbage leaves were used for oviposition.

Test materials

Commercially available twenty-nine essential oils were purchased from various sources, including Absolute Aromas (Alton, Hampshire, England), Kanta Enterprises (Noida, India), Klimtech (Dimitrovgrad, Bulgaria), Neumond (Raisting, Germany), Plant Therapy (Twin Falls, ID, USA), and Sun Essential Oils (Phoenix, AZ, USA). Details regarding their scientific names, family names, parts of the plants used, extraction methods, and manufacturers are presented in Table S1. Pure standard compounds of citral (95%), eugenol (99%), (R)-( +)-limonene (97%) were purchased from Sigma-Aldrich (St. Louis, MO, USA) and trans-anethole (> 98%) from Tokyo chemical industry (Tokyo, Japan). Commercial neem extract and transfluthrin (analytical standard), positive controls, were obtained from Kanta Enterprises (Noida, U.P., India) and Sigma-aldrich, respectively.

GC–MS analyses

The chemical composition of the three most effective oils (clove bud, fennel sweet, and lemongrass) were analyzed using an ISQTM LT gas chromatography-mass spectrometer (Thermo Scientific, Waltham, MA, USA). A VF-5 ms column (60 m × 0.25 mm ID, 0.25 µm thickness) was used in electron ionization mode. The injection volume was 1 µL, and helium (99.999%) was a carrier gas at a 1.0 mL/min flow rate. The initial oven temperature was 50℃, which was then increased to 65, 120, 180, 210, and 310℃ at the rates of 10, 5, 5, 5, and 20℃/min, respectively. The total runtime of the analysis was 100 min. The analysis results were confirmed using the NIST MS Search program (version 2.0) and NIST/EPA/NIH Mass Spectral Libraries.

Bioassays

Feeding deterrent effect of essential oils

A no-choice test was conducted to examine the feeding deterrent effects of twenty-nine essential oils on S. litura larvae (Huang et al. 2014). Third instar larvae within the 9–12 mg body weight range were individually weighed and selected, then starved for 4 h prior to the test. Fresh leaf discs were cut from 4-week-old cabbage (Brassica campestris L.) using a #8 cork borer (⌀16 mm). Test oils or compounds were dissolved in a 5% tween 80 carrier solution, and the leaf discs were immersed for 5 s and allowed to dry for 30 min. Each treated disc was gently introduced into a Petri dish (⌀60 × 15 mm, SPL Life Sciences, Pocheon, Gyeonggi, South Korea) along with five starved larvae (Fig. S1A). Simultaneously, another set of five starved larvae was introduced to a negative control disc, which was only treated with the carrier solution. The tests were terminated when the consumption on the negative control disc reached approximately 50–70% (typically within 3–5 h). Larvae were removed from each Petri dish, and the consumed areas of the discs were photographed and measured using Adobe Photoshop software. The measurement results were used to calculate the Feeding Deterrence Index (FDI) using a formula (Akhtar et al. 2010);

$$ FDI{ }\left( \% \right) = \left\{ {\left( {C - T} \right)/\left( {C + T} \right)} \right\} \times 100 $$

where C and T represent the consumed areas of the control and treatment leaf discs by five larvae each, respectively. The FDI values ranged from -100 to 100, with zero indicating equal consumption of control and treatment leaf discs. All experiments were conducted using at least five different concentrations and three replications to identify the concentration causing 50% feeding deterrence relative to controls.

Contact and fumigation toxicity test

The insecticidal activity of the oils was evaluated via a contact and fumigation assay (Wang et al. 2011). The test oils and compounds were dissolved in acetone, and 200 µL of the test solution was applied to a 55 mm diameter filter paper. After drying for 1 min, the filter paper was placed in a Petri dish (60 × 15 mm). A group of ten 3rd instar larvae was transferred into the dish, and approximately 0.5 g of artificial diet was provided. The diet was cut into small pieces to prevent larvae from climbing and escaping from the treated surface (Fig. S1B). Mortality was recorded after 24 h, and at least five different doses were used to determine the LD50 values. The test was repeated at least three times.

Synergistic feeding deterrence effect of binary mixtures of essential oils

Among the twenty-nine oils tested, fennel sweet oil exhibited the strongest insecticidal and feeding deterrent activity. To examine the interaction of fennel sweet oil and others, binary mixtures were prepared in a 1:1 (w:w) ratio at the FDI20 concentration of fennel sweet oil (20.7 + 20.7 mg/mL). The feeding deterrence activity of the binary mixtures was examined via a no-choice feeding assay against the 3rd instar larvae, as mentioned above. The observed and expected feeding deterrence values were compared using the Bliss independence model (Bliss 1939);

$$E={O}_{a} + {O}_{b}\left(1 - {O}_{a}\right)$$

where E is the expected feeding deterrence of binary mixtures, and Oa and Ob are the observed feeding deterrence of fennel sweet and other oils, respectively. A Chi-square comparison determined the synergistic effect of the mixtures;

$$ {\rm X}^{2} = \left\{ {\left( {O_{m} - E} \right)^{2} } \right\}/E $$

where Om is observed feeding deterrence of binary mixtures. With df = 1 and α = 0.05, X2 is 3.84. If X2 > 3.84, the joint effect is determined to be synergistic when the Om surpasses E; conversely, it is deemed antagonistic. If X2 < 3.84, it is considered an additive interaction (Hummelbrunner and Isman 2001).

Synergistic effect among the oils and major active constituents

Of the remaining 28 essential oils, 13 exhibited synergistic interaction with fennel sweet oil in the feeding deterrence activity. We selected one (clove bud oil) among the synergizing combinations and another displaying an additive effect (lemongrass oil) for further investigation. The two chosen oils were the most active, followed by fennel sweet oil when tested individually.

The oils were blended in a 1:1 (w:w) ratio, and the dose-dependent responses in feeding deterrence and insecticidal activity were examined. Their interaction was further examined using their major constituents identified in GC–MS analyses. For fennel sweet and clove bud oils, trans-anethole (94.2%) and eugenol (90.3%) were selected, and for lemongrass oil, a combination of citral (63.1%) + limonene (16.2%) was chosen, considering the relatively lower concentration of the most abundant compound, citral.

For estimating the expected FDI50 and LC50 values of the combinations, we employed the Wadley’s model for calculation (Gisi et al. 1985);

$$E = \frac{a+b+\cdots +n}{\frac{a}{{V}_{A}}+\frac{b}{{V}_{B}}+\cdots +\frac{n}{{V}_{N}}}$$

where E is the expected FDI50 or LC50 value of the mixture, a is the proportion of compound A in the mixture, and VA is the FDI50 or LC50 value of the compound. To determine their relationship, we employed the R-value;

$$ R{ } = E/O_{m} $$

where Om is the observed value of the mixture, and we determined a synergistic effect when R ≥ 1.5, an additive effect when 1.5 > R ≥ 0.5, and an antagonistic interaction when R < 0.5.

Habituation tests

Pre-exposure to feeding deterrents

The larvae from the 2nd to 3rd instar stages (approximately 48 h old) were exposed to four different artificial diets to determine the habituation effects of the active oils. The larvae received an artificial diet mixed with four individual oils (lemongrass, fennel sweet, clove bud, neem) or three mixtures in a 1:1 ratio (lemongrass + fennel sweet, lemongrass + clove bud, fennel sweet + clove bud, w:w). To prevent volatilization of the oils by heat during preparation, the oils were added directly to the artificial diet and stirred vigorously for about 5 min. The concentration of the oils in the artificial diet was set at 1 mg/g, and it did not have any observable impact on the survival of the 2nd instar larvae. Exposure was conducted for 48 h. A freshly prepared diet was replaced in 24 h to prevent dehydration. Naive larvae were fed only a regular artificial diet (Fig. S2).

Two-choice test for habituation assay

A two-choice test was conducted to identify the preferences of experienced larvae (Hummelbrunner and Isman 2001). The 3rd instar larvae were starved for 4 h prior to the assay. Leaf discs were cut using a #6 cork borer (⌀13 mm), dipped into test solutions prepared in a 5% Tween 80 solution, and subsequently dried for 30 min. Four different sets of tests were conducted as follows: (1) at FDI90 concentrations of individual fennel sweet, clove bud, and lemongrass oils, (2) individual major constituents of each oil at their equivalent concentrations present in the FDI90 of the corresponding oils, (3) FDI90 concentrations of binary mixtures prepared at a 1:1 (w:w) ratio, and (4) binary mixtures of each oil at their individual FDI90 concentrations. The control disc only received the surfactant solution.

Two leaf discs were positioned in a 90 mm Petri dish containing 2% agar (Fig. S3). A single larva was then placed at the center of the medium, and its response was recorded for 4 h (until approximately 50% of the control leaf disc was consumed) using an APC1000 4 K Ultra HD camera (ABKO Co., Ltd, Seoul, South Korea). After recording, the larva was removed, and the remaining areas of each leaf disc were measured, similarly to the no-choice assay. The accumulated time that the larvae spent around each leaf disc was analyzed using Ethovision XT 15 software (Noldus, Wageningen, Netherlands). For the behavioral analysis, 0.5 cm of a buffer zone around each leaf disc was established, considering the lengths of the larvae from preliminary tests. Over 10 replications were performed for both naive and experienced larvae, and unresponsive larvae were excluded from the analyses.

Electrophysiological recording of gustatory sensillum

We conducted electrophysiological experiments using citral, an active constituent of lemongrass oil, which exhibited habituation, and trans-anethole, an active constituent of fennel sweet oil that did not show a habituation effect, as the stimulants. The compounds were diluted in 0.1 M NaCl using DMSO (final concentration of 0.1%) at three concentrations (50, 100, and 200 µM). DMSO (0.1%) was used as a negative control.

The recordings were performed using a tip recording technique (Hodgson et al. 1955) for the maxillary palp, lateral styloconica, and medial styloconica of the 3rd to 4th instar larvae. Borosilicate capillaries (TW150-4, World Precision Instruments, Sarasota, FL, USA) were pulled using a Dual-Stage Glass Micropipette puller (PC-100, Narishige, Tokyo, Japan). The impedance of electrolyte-filled micropipettes was measured using an Omega-Tip-Z (World Precision Instruments, Sarasota, FL, USA). The heads of naive and experienced larvae were excised, mounted on a reference pipette electrode filled with 0.1 M NaCl, and connected to a pre-amplifier (Taste Probe DTP-1, Syntech, Buchenbach, Germany). The stimulus solutions were loaded into the recording glass micropipette. Silver wire coated with silver chloride (AgCl) was used to connect the reference and recording electrodes (Fig. S4). The amplified signal was filtered (bass band filter; 100–3000 Hz) by an A/D interface (IDAC-4, Syntech) and then sampled. Each sensillum was exposed to the stimulus solution for 2 s, with more than 3 min intervals to prevent adaption. All experiments were performed over 9 repetitions per stimulus solution. The amplitude of the negative control was set as the baseline, and the number of spikes above the baseline was counted using an Autospike 32 software (Syntech).

Statistical analysis

A Probit analysis was conducted to determine LC50 values of individual essential oils and mixtures of oils or their major constituents. A simple linear regression was used to estimate FDI50 values in no-choice tests. Pearson’s, Spearman’s, and Kendall’s correlation analyses were performed to examine the relationship between LC50 and FDI50 values. The results from two-choice bioassays and electrophysiological studies were compared using an independent sample t-test. All statistical analyses were performed using an SPSS software (version 2.5, IBM, Armonk, NY, USA).

Results

Feeding deterrence and insecticidal activity of essential oils

The feeding deterrent effect and contact-fumigation toxicity of 29 essential oils were examined against the 3rd instar larvae of S. litura (Tables 1, S2, and S3). In the no-choice test, neem oil, a positive control, exhibited the most potent feeding deterrent effect (FDI50 = 26.31 and FDI90 = 78.14 mg/mL). Among the essential oils tested, lemongrass was the most active deterrent (FDI90 = 88.38 mg/mL), followed by fennel sweet and clove bud oils with their FDI90 values of 122.73 and 129.98 mg/mL, respectively. Dose-dependent antifeedant effects were statistically confirmed by the linear regression between FDIs and the concentrations. A higher antifeedant effect is observed as the concentration of the oil increases.

Table 1 Feeding deterrence and insecticidal activity of 29 essential oils against the 3rd instar larvae of the tobacco cutworm, Spodoptera litura
Full size table

In terms of insecticidal activity, fennel sweet oil stood out (LC50 of 8.4 mg/mL), slightly overlapping the 95% confidence limits with tea tree oil (7.2–9.7 and 9.6–12.0 mg/mL, respectively), but differed from the other remaining oils. Clove bud, basil, cinnamon, spearmint, tea tree, marjoram, thyme, and peppermint oils were the second most active oil group, exhibiting similar toxicity based on their 95% confidence limits. Lemongrass and neem oils, the two most potent feeding deterrents, failed to display corresponding insecticidal activity. Moreover, neem oil showed no insecticidal activity at the highest concentration tested (> 1000 mg/mL), while transfluthrin exhibited the greatest insecticidal effect with an LC50 value of 0.011 mg/mL.

To examine the relationship between feeding deterrence and acute toxicity, three types of correlation analyses (Pearson’s, Kendall’s, Spearman’s) were conducted on FDI50 and LC50 values of 28 essential oils (Fig. S5). No notable correlation was found between the two bioactivities, indicating that feeding deterrence does not necessarily predict or correlate with insecticidal activity in these oils. Specifically, Pearson’s correlation coefficient (R = 0.69) suggests a moderate positive relationship, but the lack of statistical significance (p = − 0.08) implies that this correlation might be due to random variation rather than a true association. Similarly, Kendall’s (R = 0.30, p = 0.14) and Spearman’s (R = 0.27, p = 0.22) analyses also showed positive but weak correlations with p-values above the conventional thresholds for significance (p = 0.05), reinforcing the conclusion that the feeding deterrence and insecticidal activity of these oils are not significantly correlated.

Feeding deterrent effect of binary mixtures of essential oils

To explore the combination effects of fennel sweet with other essential oils, the feeding deterrence was assessed in binary mixtures at a 1:1 (w:w) ratio using no-choice tests (Table 2). Out of 29 mixtures, 13 demonstrated synergistic effects, 8 were additive, and the remaining 8 combinations showed antagonistic interactions. Mixture of fennel sweet with marjoram exhibited the highest synergistic effect (X2 = 258.6) among the tested mixtures, followed by cinnamon, cypress, sandal wood, ylang ylang, and basil oils (X2 > 40.0). Regarding the two active antifeedants/toxicants, clove bud oil synergized the feeding deterrence of fennel sweet oil (X2 = 34.8), but an additive interaction was observed with lemongrass oil (X2 = 1.4). Their relationships were further confirmed in dose–response tests where the two oils and their major constituents were isovolumically mixed with fennel sweet oil and trans-anethole. Interestingly, neem oil, which had a notable feeding deterrent effect, showed an antagonistic effect when blended with fennel sweet oil.

Table 2 Feeding deterrence index of binary mixtures of fennel sweet and the other oils
Full size table

Chemical composition of essential oils

The chemical compositions of the essential oils were analyzed by GC–MS analyses (Table 3). The analysis revealed that citral (63.02%), which includes E- and Z- isomers, was the most abundant compound, followed by limonene (16.23%) in lemongrass oil. Fennel sweet and clove bud oils had trans-anethole (94.23%) and eugenol (90.34%) as the predominant constituents, respectively. At least 94.3% of the constituents were identified.

Table 3 Chemical constituents of three most active essential oils
Full size table

Dose-dependent response of the binary mixtures of essential oils and their major constituents

Based on the chemical analysis results, citral + limonene (4:1, summing up to 79.3%), trans-anethole (94.2%), and eugenol (90.3%) were selected as representative constituents of lemongrass, fennel sweet, and clove bud oils, respectively. Their feeding deterrence and insecticidal activity were then examined (Table 4). Citral and limonene, the most abundant constituents of lemongrass oil, showed additive effects in both feeding deterrence and toxicity assays against the 3rd instar larvae of the tobacco cutworm (R = 1.36 and 0.93, respectively) in their proportional ratio in the oil (4:1, w:w). The mixture exhibited similar activity to lemongrass oil in both feeding deterrent (FDI50 = 59.41 and 44.35 mg/mL for lemongrass oil and citral + limonene mixture, respectively) and insecticidal activity (LC50 = 24.3 and 16.8, respectively, overlapping within the 95% confidence limits). Trans-anethole showed similar toxicity to fennel sweet oil on the larvae (LC50 = 7.7 and 8.4 mg/mL for the compound and oil, respectively) but was a more effective antifeedant than the crude oil (FDI50 = 29.26 and 64.43 mg/mL for the compound and oil, respectively), suggesting the other remaining constituents negatively affect the feeding deterrence. Eugenol, constituting 94.2% of clove bud oil, demonstrated more potent activity in both feeding deterrence and toxicity compared to the crude oil. Since these compounds exhibited similar or greater bioactivity than the crude oils, they were designated as the major active constituents.

Table 4 Feeding deterrence and insecticidal activity of binary mixtures of essential oils and their major constituents against 3rd instar larvae of the tobacco cutworm
Full size table

Among the binary mixtures of three essential oils, the combination of fennel sweet and clove bud oils exhibited a synergistic feeding deterrence effect (R = 2.47) but an additive interaction in toxicity (R = 1.19). The mixture of their corresponding active ingredients, trans-anethole and eugenol, also displayed positive interaction in feeding deterrence, but it was rather additive (R = 1.39), suggesting the crucial roles of minor constituents of the oils. Meanwhile, the blend of lemongrass and clove bud oils displayed a relatively lower but additive relationship on toxicity (R = 0.80). However, the combination of their major active compounds, citral + limonene and eugenol, displayed an antagonistic effect (R = 0.41), indicating that other constituents in the oils may compensate for the negative interaction among the most abundant constituents. The rest of the combinations showed additive interactions.

Similar to the individual oils, the correlation analysis revealed no direct correspondence between the FDI50 and LC50 for the combinations of the oils and the major constituents. This was consistent across Pearson’s (R = 0.07, p = 0.60), Kendall’s (R = 0.09, p = 0.42) and Spearman’s (R = 0.08, p = 0.58) correlation coefficients.

Habituation test for essential oils and their mixtures

To assess the occurrence of habituation to lemongrass, fennel sweet, and clove bud oils, a two-choice test was conducted on naive and experienced larvae. For the individual oils, the larvae experienced with lemongrass and clove bud oils significantly consumed more leaf areas on the treated sides (p = 0.005 and 0.001) compared to the naive ones. The experienced larvae were also more attracted to the treated discs, as indicated by the in-zone time analyses (p = 0.007 and 0.003), suggesting habituation effects to those two oils both in gustatory response and repellency (Fig. 1).

Fig. 1
figure 1

Feeding deterrent and habituation effects of crude oils and their major constituents in the two-choice test. The larvae were pre-exposed to a crude oil at 1 mg/g.Asterisks indicate significant differences between naive group and experienced group in t-test (*p < 0.05, **p < 0.01, ***p < 0.001)

Full size image

Interestingly, the larvae previously exposed to fennel sweet oil selected the treated disc more than the untreated disc (p = 0.050). Still, the feeding deterrence activity remained active based on the leaf consumption area (p = 0.104). Neem oil, the positive control, displayed no sign of habituation in the experienced larvae (p = 0.530).

The major active ingredients of each oil exhibited not identical but rather similar responses to their crude oils. Only the larvae experienced with trans-anethole failed to show habituation effect, whereas those previously exposed to the other major compounds showed habituated responses. Interestingly, while habituation in feeding deterrence was similar between the oils and their major constituents regarding statistical significance, the repellent behaviors (in-zone time) were rather different. This may suggest that the larvae of the tobacco cutworm use different cues for initiating the feeding process and scouting for food sources.

In the following test, the occurrence of habituation in the oil mixtures was tested in either the equivalent amounts (1:1, w:w) or at physiologically the same levels of deterrence (FDI90 + FDI90, Fig. 2). In the treatment with equivalent amounts, all three combinations exhibited relatively increased consumption of treated leaf discs by experienced larvae. However, there was no statistical difference between the treated and untreated discs, which may suggest a suppression of the habituation effect. Among the three oil blends, those containing clove bud oil (lemongrass + clove bud oils and fennel sweet + clove bud oils) demonstrated strong attraction to the treated side by experienced larvae. In the mixtures with FDI90 levels of each oil, a strong preference for the untreated leaf disc was observed in all three combinations, for both naive and experienced groups, as was the feeding deterrence.

Fig. 2
figure 2

Feeding deterrent and habituation effects of binary mixtures of the oils in the two-choice test. Leaf discs were treated with a the FDI90 concentration of a binary mixture (1:1, w:w) of the oils, or b the combined FDI90 + FDI90 concentrations of individual oils. All the larvae were pre-exposed to a 1:1 (w:w) binary mixture at 1 mg/g. Asterisks indicate significant differences between naive group and experienced group in t-test (*p < 0.05, **p < 0.01, ***p < 0.001)

Full size image

Electrophysiological response of gustatory sensillum

The electrophysiological responses of the gustatory sensilla of the tobacco cutworm to citral and trans-anethole were examined in naive and experienced larvae exposed to each compound (Fig. 3). Different patterns were observed across substances, concentrations, sensilla, and pre-exposure. Whereas a concentration-dependent response to both compounds was observed in the maxillary palps, the responses in the lateral and medial styloconica were maintained at levels similar to the control (Figs. S2 and S3).

Fig. 3
figure 3

Electrophysiological responses of maxillary palps of the naive and experienced larvae. a At 200 μM of citral, a significant difference was observed in the number of spikes compared to the threshold, between c naive and e experienced larvae. However, b at 200 μM of trans-anethole, no significant difference was observed between d naive and f experienced larvae. Similar results were obtained with 100 μM of g citral and h trans-anethole. Asterisks indicate significant differences between naive group and experienced group in t-test (*p < 0.05, **p < 0.01, ***p < 0.001)

Full size image

The experienced group exhibited inhibitory activity in the sensory response to citral at both 100 and 200 µM, compared to the naive group, at every recording time (100 µM: df > 9, p = 0.00 at 500 ms; p = 0.19 at 1000 ms; p = 0.03 at 1500 ms; p = 0.10 at 2000 ms, 200 µM: df > 7, p = 0.003 at 500 ms; p = 0.001 at 1000 ms; p = 0.00 at 1500 ms; p = 0.00 at 2000 ms). The significant difference in spike counts between naive larvae and experienced larvae confirmed alignment with the gustatory habituation results observed in the two choice tests.

In contrast, there was no statistical difference between the naive and experienced groups in their responses to trans-anethole at both concentrations (100 µM: df > 11, p = 0.54 at 500 ms; p = 0.54 at 1000 ms; p = 0.24 at 1500 ms; p = 0.47 at 2000 ms, 200 µM: df = 10, p = 0.71 at 500 ms; p = 0.91 at 1000 ms; p = 0.47 at 1500 ms; p = 0.51 at 2000 ms), similar to the behavioral responses in the bioassay.

Discussion

The tobacco cutworm is a migratory pest that damages economically significant crops such as sweet peppers, tomatoes, tobacco, and cabbage during its larval stage (Bragard et al. 2019; Hamzah and Norsyazwina 2021). While synthetic insecticides are the most commonly used control measures, resistance to the major insecticide classes, including organophosphates, pyrethroids, and carbamates, was reported in field populations (Shad et al. 2012; Tong et al. 2013; Babu and Singh 2023), which can potentially lead to massive crop losses (Melanie et al. 2022). Focus has been given to antifeedant research as one of the potential candidates to reduce synthetic pesticide usage. Naturally occurring antifeedants, such as azadirachtin, caffeine, strychnine, and tannic acid, demonstrated efficacy against chewing insects (Thomas et al. 2003). Azadiractin, the main component of neem extract and oil, has a strong feeding deterrent effect, promoting it as the most commercially successful antifeedant. However, relatively lower insecticidal activity than feeding deterrence at the corresponding concentration is one of the limitations for using them in the field, along with the low solubility (Campos et al. 2016).

The feeding deterrent and repellent activity of plant essential oils were well-characterized in various insect pests, including Tribolium castaneum (Coleoptera: Tenebrionidae), Sitophilus oryzae (Coleoptera: Curculionidae), Trichoplusia ni (Lepidoptera: Noctuidae), and Aphis punicae (Hemiptera: Aphididae) (Stefanazzi et al. 2011; Akhtar et al. 2012; Sayed et al. 2022). Among the 29 essential oils tested in the present study, lemongrass oil showed the greatest feeding deterrent activity, followed by fennel sweet and clove bud oils. Fennel sweet oil also exhibited the greatest insecticidal activity, suggesting its potential for pest control agents. A previous study (Kostić et al. 2021) showed a similar result, that dill and fennel oils, which had high proportions of trans-anethole (87.5 and 65.1%, respectively), exhibited not only feeding deterrent effects but also superior physiological inhibition of Lymantria dispar (Lepidoptera: Erebidae) larvae compared to neem. Compared to the increasing interest in using plant essential oils as feeding deterrents, their antifeedant mechanism is poorly understood. The feeding deterrent activity indicated no correlation to their insecticidal activity examined by contact and fumigant bioassay, consistent with a similar finding in the cabbage looper (Jiang et al. 2012). The low correlation between feeding deterrence and acute toxicity might be an understandable interpretation that in insecticidal activity, cuticular penetration in contact toxicity and respiratory absorption in fumigation would be the main routes of uptake (Plata-Rueda et al. 2018). In contrast, the chemoreceptors will be directly involved in feeding deterrence and behavioral attraction/repellency. Nevertheless, fennel sweet oil exhibited both notable feeding deterrent effect and acute toxicity.

The bioactivity of the plant essential oils can be influenced significantly by their chemical composition. For example, in fennel sweet oil, trans-anethole was the most abundant constituent, comprising 94.2% of the oil. Other major constituents were estragole, limonene, and methyl undecanoate. Although the chemical composition was different, fennel oil also had similar constituents, including trans-anethole (44.6%), methyl undecanoate (21.2%), and limonene (2.6%). The difference in the chemical composition significantly affected both feeding deterrence (FDI50 values of fennel sweet and fennel oils were 64.4 and 690.8 mg/mL, respectively) and toxicity (LC50 = 8.4 and 186.7 mg/mL, respectively). The complex nature of the chemical composition of plant essential oils may result in wide differences in biological activity, possibly via complicated synergistic, additive, or antagonistic interactions among the constituents. Identifying the major active ingredients and understanding the intertwined relationship among them are crucial for assessing their potential for success in commercialization and utilization for pest control agents (Isman 2000, 2020).

Previous studies documented behavioral habituation to feeding deterrents in various herbivorous insect pests, including T. ni (Lepidoptera: Noctuidae), Agrotis ipsilon (Lepidoptera: Noctuidae), and Rhyzopertha dominica (Coleoptera: Bostrichidae) (Akhtar and Isman 2004; Zhou et al. 2021a; Giunti et al. 2021). However, compared to many reports on the habituation effects of plant essential oils, the underlying physiological mechanisms remain less understood. Insects use various cues to search for their food or hosts, and diverse receptors are involved in olfactory or gustatory chemoreception (Dahanukar et al. 2005). The present study suggests that different olfactory or gustatory receptors are engaged in feeding deterrence and habituation of plant essential oils. For example, the pre-exposed larvae to fennel sweet oil in an earlier stage displayed a significant behavioral preference for the oil-treated leaf discs compared to the untreated control in the two-choice test. However, those larvae that selected the treated leaf did not consume the food to the same extent, indicating that the feeding deterrent effect was not diminished (i.e., no habituation). Moreover, electrophysiological recording on maxillary palps of the tobacco cutworm using trans-anethole and citral demonstrated the physiological response in gustatory habituation. Through these results, we confirmed that the orientation towards food may contribute by the olfactory organ, whereas the decision to consume the food upon reaching it would be governed by the gustatory organ. Therefore, further research on antennal reception is necessary, with separate examination of host preference and gustatory discrimination, in order to comprehensively understand the feeding deterrent activity and repellent behaviors.

Among the three active essential oils of fennel sweet, lemongrass, and clove bud oils, only fennel sweet oil displayed a low tendency of habituation, along with the positive control, neem oil in the larvae of the tobacco cutworm. A similar result was observed in a previous study (Bomford and Isman 1996), which found that neem oil did not show habituation to the same species. Mordue (2004) suggested that azadirachtin, the major active constituent of neem oil, stimulates specific inhibitory cells of chemoreceptors and blocks the firing of sugar-responding cells, leading to feeding deterrence. Therefore, it is imperative to investigate whether trans-anethole and neem oil operate through the same feeding deterrence mechanism at the single-cell level. On the other hand, habituation was observed in citral, which exhibited remarkable feeding deterrence along with trans-anethole, and a decrease in response was observed in experienced larvae during electrophysiological experiments. These results, as confirmed by previous studies (Sfara et al. 2011; Stuck et al. 2014), indicate that gustatory habituation occurs as a result of sensory adaptation mediated by neuronal desensitization. Consequently, considering that citral and trans-anethole exhibit different desensitization patterns, their feeding deterrence mechanisms may differ.

The chemical complexity of plant essential oils and extracts often leads to results where the inactive constituents are involved in not only enhancing insecticidal activity (Miresmailli et al. 2006; Kim et al. 2021), but also decelerating the development of habituation and resistance (Bomford and Isman 1996; Feng and Isman 1995). For example, whereas the crude neem oil did not exhibit habituation in the tobacco cutworm larvae, pure azadirachtin displayed desensitization of feeding deterrence (Bomford and Isman 1996). Interestingly, the present study found that trans-anethole did not lead to habituation in the bioassay and electrophysiological recording even when treated alone, suggesting its potential as a novel feeding deterrent of fennel sweet oil. Several hypotheses could be proposed for the continual response of the maxillary palps to trans-anethole in the experienced larvae. First of all, habituation might occur differently based on the structural differences of the stimuli. Glendinning and Gonzalez (1995) showed that habituation patterns can vary within the same alkaloid (lobeline, nicotine, and hyoscyamine) depending on their chemical structures. Citral comprises a complex diene structure with an aldehyde group, while trans-anethole comprises a simpler benzene compound with an alkene and a methoxy group. Secondly, Sinding (2017) found that odorant compounds with lower habituation effects tend to have lower carbon chains and molecular weight but higher vapor pressure. Compared to citral (C10H16O, 152.2 g/mol, 0.29 hPa), trans-anethole (C10H12O, 148.2 g/mol, 1.33 hPa) satisfies the conditions for lower habituation, but the difference may not be significant. Moreover, as these structural differences primarily affect olfactory recognition, it is speculated that they may not be substantially relevant to gustatory habituation. Rather, trans-anethole might have a higher supra-threshold for habituation (Glendinning and Gonzalez 1995). As shown in Fig. 1, despite being treated with the same FDI90 concentration as other compounds, there was no statistical difference (p = 0.393) in the consumption of treated leaf discs with trans-anethole between experienced larvae and naive larvae. Nevertheless, experienced larvae showed a tendency for increased feeding area and in-zone time compared to naive larvae. This could be attributed to deterrent receptors adapting relatively slowly, allowing for the accumulation of sensory information after feeding, which might result in subsequent refusal responses (Schoonhoven et al. 2005). Nonetheless, an increase in food intake was observed in lemongrass and clove bud-experienced larvae, which may suggest sensitization or associative learning rather than habituation in a strict sense. Further studies are planned to provide a comprehensive understanding of the feeding behaviors of those active oils. Alternatively, only partial habituation may occur because the experienced larvae did not surpass the supra-threshold for complete habituation, resulting in some feeding behavior. Considering the complexity of gustatory habituation mechanisms, additional research on the sensitivity and specificity of gustatory receptors is required for a more comprehensive understanding of the mechanism.

Although promising, using chemically synthesized trans-anethole via mass production as a stand-alone feeding deterrent in the field may not be an ideal pest control strategy. This approach would face challenges, not only regarding concerns about its ‘natural’ origin but also due to cost-effectiveness issues (Forim et al. 2013). More importantly, the feeding deterrent effects of trans-anethole and fennel sweet oil were found to be approximately 2–3 times less active than neem oil. To boost the effectiveness of fennel sweet oil, we explored various binary mixtures with other essential oils. Marjoram oil significantly enhanced the feeding deterrence of fennel sweet oil, followed by cinnamon, cypress, and sandalwood oils. Despite these boosting effects, the oils mentioned were less effective as feeding deterrents when tested individually. Clove bud oil individually deterred the feeding behavior and improved the antifeedant activity in the mixture with fennel sweet oil, suggesting it could be a potential partner. Nonetheless, an increased preference for treated leaves among experienced larvae, though not statistically significant, suggests a potential for cross-habituation at the mixture level. Mixtures may also induce habituation, particularly when the substances impact the same deterrent neurons (Zhou et al. 2010, 2021a). Therefore, it is essential to ensure that deterrent neurons responding to each major constituent do not overlap to facilitate dishabituation. Future research should focus on dishabituation strategies rather than solely on reducing habituation at the mixture level. Various studies showed that habituation ceased when different substances were presented to habituated individuals (Epstein et al. 1992; Akhtar et al. 2003; Shikano et al. 2010).

Research on the mechanism of habituation at the mixture level remains less explored. Our present study suggests that concentrations and specific combinations can influence habituation in mixtures. To the best of our knowledge, this is the first study to delve into the gustatory habituation mechanisms of S. litura in response to essential oils and to investigate habituation within mixtures. In future studies, we plan to examine various aspects, including the responses of deterrent neurons, the optimal combination ratios, chemical structures, and other factors associated with the synergistic compounds of fennel sweet and strategies for dishabituation.

Author contributions

HJ and JHT conceived and designed the research. HJ conducted experiments. HJ and JHT analyzed the data and wrote the manuscript. All authors read and approved the manuscript.