Synthesis and Antifeedant Activity of Racemic and Optically Active Hydroxy Lactones with the p-Menthane System

Two racemic and two enantiomeric pairs of new δ-hydroxy-γ-lactones based on the p-menthane system were prepared from racemic and optically active cis- and trans-piperitols. The Johnson-Claisen rearrangement of the piperitols, epoxidation of the γδ-unsaturated esters, and acidic lactonization of the epoxy esters were described. The structures of the compounds were confirmed spectroscopically. The antifeedant activities of the hydroxy lactones and racemic piperitone were evaluated against three insect pests: lesser mealworm, Alphitobius diaperinus (Panzer); Colorado potato beetle, Leptinotarsa decemlineata (Say); and peach-potato aphid, Myzus persicae (Sulz.). The chemical transformation of piperitone by the introduction of a lactone moiety and a hydroxy group changed its antifeedant properties. Behavioral bioassays showed that the feeding deterrent activity depended on the insect species and the structure of the compounds. All hydroxy lactones deterred the settling of M. persicae. Among chewing insects, the highest sensitivity showed A. diaperinus adults.

Our interest in terpenoid lactones is inspired mainly by their activity as insect feeding deterrents; many natural antifeedants contain the lactone moiety and have isoprenoid structures [17][18][19][20]. However, their low concentrations in plants and their typically complex syntheses have limited the large-scale application of natural antifeedants. Therefore, in our opinion, synthetic feeding deterrents with simple structures offer better potential for practical use in insect pest population control. Thus, we have synthesized a number of terpenoid lactones by using optically active monoterpenoids as the starting materials [13-15, 21, 22].
In many syntheses, we obtained the final products as enantiomeric pairs. Our studies confirmed that the biological effects of the optically active substances depend on their configuration. Moreover, structural modifications of natural monoterpenoids, such as lactonization, iodolactonization, and incorporation of hydroxyl and carbonyl groups, can cause changes in their biological activities. Some of the prepared compounds were very effective antifeedants against selected insect pests. Their activities are comparable with that of the most active known antifeedant, azadirachtin [11,15,16].
We studied three insect pests: the lesser mealworm, Alphitobius diaperinus (Panzer) is a common cosmopolitan insect pest of poultry houses, particularly in broiler sheds. Its mass occurrence in poultry farms creates serious veterinary and economic problems for poultry breeders: the beetles damage the insulation of the houses and have the potential to act as reservoirs for poultry parasites and pathogens [23,24]. The beetle is also known to be a pest in animal feeds, especially in neglected storage rooms.
The Colorado potato beetle, Leptinotarsa decemlineata (Say), is widely regarded as the most important insect defoliator of plants in the family Solanaceae in North America, Europe, and Asia. High selection pressure, together with a natural propensity to adapt to toxic substances, has finally resulted in a large number of insecticide-resistant L. decemlineata populations. The beetle has developed resistance to insecticides from different chemical groups [25,26]. Consequently, alternative methods for the control of this pest are needed.
Aphids, in addition to their direct detrimental effects on host plants due to the uptake of phloem sap, are the most important vectors of plant viruses. They transmit nearly 30% of all hitherto described plant virus species, which account for nearly 50% of all insect-borne viruses [27]. Therefore, it is crucial to deter aphid probing and feeding at pre-ingestional, ingestional, and post-ingestional phases. In our former study, we found that the chlorinated, brominated, and iodinated lactone derivatives of piperitone deterred the probing, feeding, and settling of the peach-potato aphid Myzus persicae (Sulz.), in contrast to piperitone itself, which appeared to be a weak attractant [13].

Materials and Methods General
Analytical TLC was performed on Merck Kieselgel 60 F 254 plates with mixtures of hexane and diethyl ether in various ratios. Compounds were detected by spraying the plates with 1% Ce (SO 4 ) 2 and 2% H 3 [P(Mo 3 O 10 ) 4 ] in 10% H 2 SO 4 , followed by heating to 120-200°C. Column chromatography was performed on silica gel (Kieselgel 60, 230-400 mesh ASTM, Merck) with mixtures of hexane and diethyl ether (in various ratios) as eluents.

Chemical synthesis
Synthesis of epoxy esters (3a-c, 4a-c, and 9a-c). All epoxy esters were obtained according to the following general procedure. A solution of m-chloroperbenzoic acid (3 mmol) in dry methylene chloride (15 mL) was added dropwise to a stirred solution of γ,δ-unsaturated ester (3 mmol) in methylene chloride (20 mL). The mixture was stirred at room temperature for 24 h. Then the reaction mixture was washed with saturated Na 2 S 2 O 3 solution, and extracted with methylene chloride. The separated organic layer was washed with 0.5 M NaHCO 3 solution, dried over anhydrous MgSO 4 , and concentrated in vacuo. The crude product (or mixture of products) was purified by column chromatography (hexane/diethyl ether, gradient from 9:1 to 4:1). The yields of the reactions and physical and spectral data for the epoxy esters obtained are given below.
Synthesis of δ-hydroxy-γ-lactones (6a-c and 11a-c). All δ-hydroxy-γ-lactones were obtained according to the following general procedure. To the solution of epoxy ester (or mixture of epoxy esters, 2 mmol) in THF (10 mL) were added water (5 mL), and 5 drops of HClO 4 (70%). The mixture was stirred at room temperature for 24 h, and extracted with diethyl ether. The separated ethereal solution was washed with saturated NaHCO 3 solution, brine, dried over anhydrous MgSO 4 , and concentrated in vacuo. The crude product was purified by column chromatography (hexane/diethyl ether, 4:1). The yields of the reactions and physical and spectral data of the hydroxy lactones obtained are given below.
( . Its IR and NMR spectra were identical to those of (±)-6c.
Chewing insects (A. diaperinus and L. decemlineata). From the test data for chewing insects (A. diaperinus and L. decemlineata), the deterrence coefficients (relative R, and absolute A) were calculated using Eqs (1) and (2) [36,37]: where C and CC are the weights of the control consumed by the insects in the choice and nochoice tests, respectively, and similarly, E and EE are the weights of the treated food consumed, respectively. The measure of the deterrent activity of the tested compounds is the total coefficient of deterrence, Eq (3): The total coefficient of deterrence, T, which ranged from −200 to 200, serves as an index of activity. Compounds with T ranging from 151 to 200 are classified as very strong deterrents, and those with values from 101 to150 and from 51 to 100 are classified as good deterrents and compounds with medium activity, respectively. Compounds with T values lower than 50 are poor antifeedants. Negative T values indicate attraction. In practice, the deterrent activity of the chemical compounds is significant in a no-choice situation.
To estimate and compare larval and adult feeding levels in the no-choice tests, the amount of treated food consumed was expressed as a percentage of the consumption in the control according to Eq (4): The mean values of the deterrence coefficients were compared using one-way analysis of variance (ANOVA) followed by Tukey's test at a significance level of p < 0.05 [38]. The mean values of consumption by the larvae and adults in the no-choice test were compared using Student's t-test.
Aphids (M. persicae). In the experiment with M. persicae, the settling choice-test was applied as described previously [13]. From the test data, the relative index of deterrence (DI) was calculated using Eq (5): where C is the number of aphids that settled on the control leaf and T is the number of aphids that settled on the leaf treated using the studied compound. The values of DI range between +1 (ideal deterrent) and −1 (ideal attractant). The test data were analyzed using the Student's ttest. Aphid probing and, especially, the phloem sap uptake by M. persicae was monitored in the no-choice situation by using the electrical penetration graph (EPG) technique, i.e., the electronic registration of aphid stylet penetration in plant tissues, according to the procedure described previously [13,39]. The parameters derived from the EPGs, related to activities in the peripheral and vascular tissues, were analyzed according to their frequency and duration. The results were statistically analyzed using the Mann-Whitney U test at p < 0.05.
The epoxidation of cis-esters (2a-c) with m-chloroperbenzoic acid afforded a mixture of epoxy esters 3a-c and 4a-c, respectively (Fig 1). GC analysis of these mixtures indicated, that one epoxy ester was formed in a slight diastereoisomeric excess (53:47%). Unfortunately, these mixtures were inseparable by column chromatography. The structures of the products were confirmed by 1 H NMR spectroscopy. According to analysis by Dreiding models, the doublets of doublets at 3.24 and 2.96 ppm were ascribed to H-2' and H-3', respectively, of the trans-isomer of epoxy ester (3c), and the broad doublets at 3.00 and 2.94 ppm to these protons (H-2' and H-3', respectively) in the cis-isomer (4c). Protons H-2' and H-3' in the trans-isomer (3c) are more deshielded (in comparison to cis-isomer, 4c) because they are located on the same side of the cyclohexane ring as the carboethoxymethylene group. From the integration of these signals, it follows that the mixture contains 53% 3c and 47% 4c. Two doublets at 2.21 and 2.57 ppm (J = 14.6 Hz) were ascribed to the methylene protons (CH 2 -2) of the carboethoxymethylene group of the cis-isomer of epoxide (4c), and the two-proton singlet at 2.34 ppm to these protons in the trans-isomer (3c). In epoxy ester 4c, the isopropyl group is oriented cis to the oxirane ring. Taking this into consideration, the two doublets (J = 6.8 Hz) at 0.98 and 1.02 ppm were ascribed to the methyl protons of the isopropyl group of diastereoisomer 4c.
The six-proton doublet at 0.95 ppm (J = 6.8 Hz) was ascribed to the isopropyl methyl protons of epoxy ester 3c.
The key step of the synthesis was the acidic lactonization of the obtained epoxy esters. The reaction of mixtures of 3a and 4a, 3b and 4b, and 3c and 4c, catalyzed by HClO 4 in THF/H 2 O solution, gave only the single corresponding δ-hydroxy-γ-lactone (6a, 6b, and 6c, respectively, Fig 1). The structures of these products were confirmed by their spectral data. In the IR spectrum of lactone 6c, the characteristic absorption bands of the γ-lactone moiety (1778 cm −1 ) and the hydroxy group (3372 cm −1 ) were observed. The presence of a doublet (δ = 3.91) for H- The formation of only a single hydroxy lactone from both isomers of the epoxy esters can be explained on the basis of the mechanism suggested for lactonization of another epoxy ester with the p-menthane framework [22]. The product of the proposed first step, the opening of the epoxide to the diol ester, was not observed in the reaction mixture, because it most likely cyclized immediately to the γ-lactone. The diol results from the nucleophilic attack of H 2 O on C-2' or C-3' from the opposite side of the oxonium ion which is formed after H + addition to the oxirane oxygen. In the case of the trans-epoxides (3a-c), attack at C-2' leads to trans- diequatorial diols 5 with the carboethoxymethylene group in the axial position (Fig 1). The same dihydroxy esters (5) can be obtained in the case of nucleophile attack on C-3' in the cisepoxides (4a-c). These diols (5) undergo lactonization to the corresponding δ-hydroxy-γ-lactones (6a-c). The enantiomeric excesses of hydroxy lactones 6a and 6b, determined by chiral GC, were the same as the starting γ,δ-unsaturated esters (2a and 2b): 98% and 91%, respectively (S7 Fig). Oxidation of γ,δ-unsaturated esters 8a-c with m-chloroperbenzoic acid gave only one epoxy ester (9a-c, respectively, Fig 2). The comparison of the 1 H NMR spectra of 9c and the mixture of epoxides 3c and 4c allows the determination of the configuration (cis or trans) of the oxirane ring relative to the carboethoxymethylene group. Two broad doublets (J = 3.7 Hz) at 2.93 and 2.98 ppm for H-2' and H-3', respectively, in the 1 H NMR spectrum of 9c were observed. The shapes and chemical shifts of these signals are analogous to the multiplets for H-2' and H-3' in epoxide 4c. Additionally, two doublets (δ = 2.31 and 2.44, J = 13.8 Hz) from the methylene protons (CH 2 -2) of the carboethoxymethylene group, similarly to 4c, were observed. These data indicate that, in epoxy ester 9c, the oxirane ring is oriented cis to the carboethoxymethylene group. In the same manner as in the synthesis of racemic epoxy ester 9c, the oxidation of the optically active γ,δ-unsaturated esters (8a and 8b) afforded the corresponding epoxides (9a and 9b, respectively).
The acidic lactonization of epoxy esters 9a-c gave δ-hydroxy-γ-lactones 11a-c, respectively. The presence of the hydroxy group and the γ-lactone ring in lactone 11c was confirmed by the absorption bands in the IR spectrum at 3478 and 1772 cm −1 , respectively. In the 1 H NMR spectrum of 11c, multiplets due to H-5 (4.29 ppm) and a doublet due to H-6 (4.12 ppm) were present. The small value of the coupling constant, J H5-H6 = 2.8 Hz, proved the trans-diequatorial orientation of the coupled protons. Therefore, it can be concluded that the C−OH as well as the lactone C−O bonds are located in axial positions. The enantiomeric excesses of lactones 11a (ee = 94%) and 11b (ee = 98%), determined by chiral GC (S8 Fig), were identical to the enantiomeric excesses of the γ,δ-unsaturated esters (8a and 8b, respectively) which were used for the reaction.
Similarly to the synthesis of lactones 6a-c, we assume that the lactonization of epoxy esters 9a-c proceeds via the diols (10, Fig 2). In the case of epoxides 9a-c, nucleophilic attack at C-3' leads to trans-diaxial diols 10 with the carboethoxymethylene group in the equatorial position.

Biological activities
Lesser Mealworm, Alphitobius diaperinus. The antifeedant activity results for piperitone and the hydroxy lactones (6a-c and 11a-c) toward A. diaperinus are presented in Table 1. Values represent the means of four replicates, each set by using ten larvae or adults (n = 40). A, absolute coefficient (no-choice test); R, relative coefficient (choice test); T, total coefficient (T = A + R). Means followed by the same letters within each column are not significantly different (one-way ANOVA and Tukey's tests, p < 0.05).
The starting piperitone showed weak deterrent activity for both developmental stages, especially in the no-choice test. What is more, in the trials using larvae, piperitone stimulated higher food intake compared to that by the control, and therefore, it is defined as an attractant. Likewise, its lactone derivative 11a showed attractant properties. The consumption of food treated with this compound (11a) was the highest (152.43% compared to the control) (Fig 3). Among the tested hydroxy lactones, only 6b was a very good antifeedant against A. diaperinus larvae, but only in the choice test. In the no-choice situation, its activity was moderate. Larvae feeding decreased by approximately 40% compared to the control (Fig 3). The remaining compounds were weak antifeedants.
The sensory sensitivity of A. diaperinus adults was significantly higher compared with the larvae. All the hydroxy lactones were better antifeedants with rescpect to piperitone. The most active compounds were 11a and 11c, with activities comparable to that of the most active known antifeedant, azadirachtin [30]. In the presence of these lactones (11a and 11c), the food consumed by the adults in the no-choice test represented only around 3% of the consumption in the control (Fig 3). Isomer 11b, with the 1R,4R,5S,6S configuration, was a weaker antifeedant than its enantiomer 11a or racemic lactone 11c. High activity in the no-choice test was observed for hydroxy lactone 6a. The low value of the relative coefficient R for this compound was because of the low level of feeding in the control. This may suggest that lactone 6a operates as a repellent. In this case, a strong reduction in treated food consumption in both tests and only low discrimination between the treated and control foods in the choice test was observed [40]. Thus, the change in configuration from 1S,4R,5R,6R (6a) to 1S,4S,5R,6R (11a) leads to change in properties and model of operation. The same situation was observed in the case of the pair of lactones, 6a and 6b. The coefficients A and R obtained for hydroxy lactones 6b and 11a clearly indicate their antifeedant properties, with a somewhat weaker activity for isomer 6b. The activity of 6c in terms of the model of operation was similar to that of lactone 6b. Despite the fact that compounds 6a and 6c had very similar T-values, their activities and action models were different. Hydroxy lactones 6b and 11b showed similar activities (consumption in relation to the control was about 13% in the no-choice tests) (Fig 3). In the choice test, the activity of isomer 6b was slightly higher, but without significant difference.
Colorado Potato Beetle, Leptinotarsa decemlineata. The tested compounds demonstrated weak antifeedant activity against L. decemlineata (Table 2). Piperitone did not reduce the feeding behavior, especially of larvae in the no-choice test. No significant difference was noted between area consumed of the piperitone-treated leaf disks and untreated ones. Disks treated with hydroxy lactones 6b, 6c, and 11a in the no-choice tests were eaten at a higher proportion than the control disks. The disk treated with lactone 6b were the most willingly consumed (consumption in relation to the control was 160.6%) (Fig 4).
The change in configuration of the chiral centers from 1R,4S,5S,6S (6b) to 1S,4S,5R,6R (11a) only slightly decreased the value of the coefficient A, and these compounds remained attractants in the no-choice test ( Table 2). In this study, L. decemlineata larvae were deterred only by hydroxylactone 11b during the choice tests, where they consumed significantly more control disks than the treated ones. However, in the no-choice bioassays, this compound (11b) reduced herbivory of the treated leaf disks by approximately 40% (Fig 4).
The test compounds also showed low antifeedant activity against L. decemlineata adults. Thus, although comparison of the total deterrence coefficients revealed the strongest activities for the hydroxy lactones 11a and 11b, from a practical point of view, they are rather weak deterrents (Table 2).
Aphids, Myzus persicae. The relationship between aphids and plants is in some ways unique among insects. Aphids feed on phloem sap using their specialized sucking-piercing mouthparts, which penetrate plant tissues intercellularly until they reach the vascular tissue [41]. Plant penetration can be divided into three distinct phases: the pathway, xylem, and phloem phases. During pathway, the insects' stylets (aphid mouthparts) pierce through the epidermis and non-vascular tissues; the xylem phase involves the uptake of xylem sap; and the phloem phase comprises the main feeding, but is always preceded by sieve element salivation, presumably to suppress phloem wound responses [42]. Of the three probing phases, the pathway and phloem phases are crucial steps in the chemosensory-based host-plant selection and host-plant acceptance processes. The parameters derived from the EPG describe aphid behavior during probing and feeding and are good indicators of plant suitability or the interference of probing by chemical or physical factors, including exogenously applied chemicals, in individual plant tissues [43]. Accordingly, the alteration of aphid behavior during the pathway phase may reflect the hindrance of probing at a pre-ingestional level; changes in behavior during contact with the phloem elements may denote an ingestional effect; and the refusal to settle on plants even if the feeding process has not been impeded may be a symptom of post-ingestional deterrence [13,14,44]. All hydroxy lactones (6a-c and 11a-c) evoked negative responses in the M. persicae, and the intensity of the effect increased over time. However, the deterrent activity of the individual compounds varied in potency, time of expression, and duration of the effect, depending on the spatial structure of the lactones and biological background of their activity (Fig 5).
Hydroxy lactones 6a-c and 11a-c deterred the settling of M. persicae, in contrast to piperitone, which showed weak attractant properties (DI 24 = −0.43) (Fig 5). The lactone 6c prevented aphid settling as soon as 1 h after application, and the index of deterrence remained at a relatively high level until the end of the 24 h experiment (DI 1 = 0.36, DI 2 = 0.38, and DI 24 = 0.44). The deterrent effect of the remaining hydroxy lactones occurred later than 2 h after application. The relative indices of deterrence DI 24 of these lactones ranged from 0.23 (11a) to 0.64 and 0.61 (6b and 11c, respectively) ( Fig 5).
The monitoring of aphid probing behavior via EPG revealed that lactone 6b was an active deterrent at the pre-ingestional and ingestional levels ( Table 3, S9-S11 Figs, S2 File). The preingestional activity was manifested during the pathway phase, that is, during aphid probing in non-phloem tissues. A reduction of the average probing time and increase in the number of short probes that did not reach beyond the epidermis occurred, and fewer aphids reached phloem vessels after the application of 6b. The ingestional activity was demonstrated in the considerable reduction of sap ingestion activities in both total time and individual ingestion periods and the failure to sustain the ingestion after sampling phloem sap in a number of individuals. The activity of lactone 11c was expressed after the ingestion of phloem sap. There was no significant decrease in the ingestion time, either total or individual. However, the considerably higher number of epidermal and mesophyll probes following ingestion periods may indicate negative changes in the phloem sap quality due to the application of 11c. None of the studied behavioral aspects of aphid probing was affected after the application of lactones 11a and 11b. Nevertheless, the avoidance of 11aand 11b-treated leaves by aphids was found 24 h after application. This may suggest post-ingestional effects of these lactones, probably due to metabolic reasons.
The incorporation of the γ-lactone moiety and hydroxy group into the piperitone molecule changed its biological activity. Behavioral bioassays showed that the antifeedant activity of the studied lactones was related to the insect species and structure of the compounds. Among the  chewing insects, A. diaperinus adults showed the highest sensitivity, with the hydroxy lactones, except for 6c, strongly limiting their feeding. All the piperitone derivatives were much stronger antifeedants for adults than for larvae, which is a quite commonly noted phenomenon. This is because the adults possess a larger number of mouthpart chemoreceptors and choose specific host plants and sites for oviposition [45]. All the hydroxy lactones deterred the settling of M. persicae; this was in contrast to piperitone, which showed weak attractant properties. Lactone 6b was the strongest deterrent at the pre-ingestional and ingestional levels. Among the piperitone derivatives, no clear effect was noted related to the configuration of the chiral centers on the deterrent activity of the tested compounds.