Laboratory Evaluation of Natural and Synthetic Aromatic Compounds as Potential Attractants for Male Mediterranean fruit Fly, Ceratitis capitata†

Ceratitis capitata, the Mediterranean fruit fly, is one of the most serious agricultural pests worldwide responsible for significant reduction in fruit and vegetable yields. Eradication is expensive and often not feasible. Current control methods include the application of conventional insecticides, leading to pesticide resistance and unwanted environmental effects. The aim of this study was to identify potential new attractants for incorporation into more environmentally sound management programs for C. capitata. In initial binary choice bioassays against control, a series of naturally occurring plant and fungal aromatic compounds and their related analogs were screened, identifying phenyllactic acid (7), estragole (24), o-eugenol (21), and 2-allylphenol (23) as promising attractants for male C. capitata. Subsequent binary choice tests evaluated five semisynthetic derivatives prepared from 2-allylphenol, but none of these were as attractive as 2-allylphenol. In binary choice bioassays with the four most attractive compounds, males were more attracted to o-eugenol (21) than to estragole (24), 2-allylphenol (23), or phenyllactic acid (7). In addition, electroantennography (EAG) was used to quantify antennal olfactory responses to the individual compounds (1–29), and the strongest EAG responses were elicited by 1-allyl-4-(trifluoromethyl)benzene (11), estragole (24), 4-allyltoluene (14), trans-anethole (9), o-eugenol (21), and 2-allylphenol (23). The compounds evaluated in the current investigation provide insight into chemical structure–function relationships and help direct future efforts in the development of improved attractants for the detection and control of invasive C. capitata.


Introduction
Tropical fruit flies (Diptera: Tephritidae) are among the most economically important pests that threaten global agriculture [1]. Among the exotic fruit flies, the Mediterranean fruit fly (medfly), Ceratitis capitata (Wiedemann), is considered one of the most destructive agricultural pests because of its direct damage to many varieties of fruits and vegetables [2][3][4]. The species apparently originated in sub-Saharan Africa; however, increased human mobility and trade in agricultural commodities have increased the incidence of the introduction of exotic fruit flies to the United States [5]. Medfly was first Molecules 2019, 24 4 Me Two short-range bioassays were carried out. In Experiment 1, paired t-tests found that male response to compounds 7, 9, 11, 13, 21, 23, 24, and 27 was higher than the response to the associated solvent control ( Table 1). Response of flies was converted to Attraction Index for comparisons among these eight compounds. There was an effect of type of chemical on Attraction Index. (F7,32 = 16.69, p < 0.0001). The highest index was found with chemical 7 (0.63 ± 0.26), intermediate indexes
In particular, 2-allylphenol (23), by reaction with Ac 2 O and pyridine, was converted into the corresponding 1-O-acetylderivative (25). Its 1 H-NMR spectrum differed from that of 23 for the presence of the typical singlet of the acetyl group at δ 2.32 and was very similar to that previously reported Gresser et al. [24]. Further confirmation was given by an ESI-MS spectrum, which showed the protonated form [M + H] + at m/z 177.
By reaction with an ethereal solution of diazomethane, 23 was converted into the corresponding methyl ether (allylanisole, 26). Its 1 H NMR spectrum differed from that of 23 for the significant presence of the singlet of methoxy group at δ 3.83 and was very similar to that previously reported [25]. Furthermore, its ESI-MS showed the protonated form [M + H] + at m/z 149.
2-Allylphenol (23), by reaction with 4-bromobenzoic acid, yielded its corresponding p-bromobenzoyl ester (27 (23), by reaction with mesyl chloride in pyridine, afforded the corresponding mesyl ester (28). Its 1 H-NMR spectrum was differed from 23 by the significant presence of the singlet of the methyl group at δ 3.20 and was very similar to that previously reported by Lei et al. [26]. Its ESI-MS spectrum showed the protonated form [M + H] + at m/z 213.
These results demonstrated the importance of the functional groups and their position on the benzene ring. In particular, the side chain of 7 was the most important moiety to stimulate attraction of males in Experiment 1. The presence of the allyl group was also a critical factor, as evidenced by little or no attraction observed with compounds lacking this moiety, as in 1-6, 16, and 17. The same result was obtained when the allylic chain was modified by isomerization of the double bond, as in 9 and 18, or by the presence of a propyl residue, as in 19. However, other substitutions on the aromatic ring of allylbenzene (10) were needed to confer activity, as 10 elicited no response. When the number of substituents on the aromatic ring increased, those with three functional groups lost activity, as in 13, 15, and 22. Among compounds having two substituents, as in 20 and 21, only the latter, having a substituent ortho-located with respect to the allyl group, was active. Although methyl eugenol (20) is a male attractant for Bactrocera dorsalis (Diptera: Tephritidae) [27], male medflies showed poor response to compound 20. Among the orthoand para-monosubstituted allyl benzenes, only 23 and 24 displayed activity while the others (8, 11, 12, and 14) were inactive. This suggests that the efficacy was dependent upon the substituents and their volatility and fragrance. Furthermore, comparing the activity of the four esters (25,(27)(28)(29) and one ether of 23 with that of the parent compound, they were all less attractive. Thus, the presence of a free hydroxyl and the ortho-allyl group was important for the activity. It should also be noted that some of the compounds tested could be potential repellents; however, that determination would require bioassays with a different experimental design (e.g., a known attractant vs. a combination of attractant plus potential repellent). Further investigations are needed to assess potential repellent properties of compounds deemed non-attractive in this study. Table 1. Log P values and number (mean ± std dev) of sterile male C. capitata attracted to compounds 1-29 presented in binary choice bioassays against control (Experiment 1).
Lipophilicity, expressed as the logarithm of the octanol-water partition coefficient (log P), is often correlated with biological activity. Log P values of a series of test chemicals often follow predictable trends in biological assays [28]. Compounds with lower log P values are classified as polar, while those with higher log P values are considered more lipophilic with better membrane permeability [29]. However, we did not observe any correlation between Log P values for compounds 1-28 [30] and 29 [31] and male response in Experiment 1 (r = 0.03037, n = 29, p = 0.8757) ( Table 1).
When pairwise comparisons of the four most attractive compounds from Experiment 1 were tested in Experiment 2, there were clear choices among most chemicals (Table 2). More males were attracted to compound 21, and fewer males were attracted to compound 7 in all bioassays. Attraction to compounds 23 and 24 was intermediate, with no difference in attraction when 23 and 24 were tested together. Phenyllactic acid (7) has been isolated from cultures of Lactobacillus plantarum [32] and may function as a food-based attractant. Estragole (24) is more likely a kairomone, since it occurs in a variety of essential oil-bearing plants such as basil, tarragon, chervil, fennel, clary sage, anise, and rosemary [33], as well as in the leaves of various avocado cultivars [34], ripe apple [35] and citrus fruits [36]. These results support that both the presence of the allyl residue, and substituents on the aromatic ring are key structural features that confer attraction of male C. capitata.
In EAG analyses (Figure 2), there were significant differences in mean olfactory response to test chemicals observed in all four groupings (Group 1: F 6, 68 = 185.02, p < 0.001; Group 2: F 6, 68 = 95.97, p < 0.001; Group 3: F 7, 76 = 375.321, p < 0.001; and Group 4: F 6, 68 = 78.33, p < 0.001). In general, strong EAG responses were elicited by compounds that were observed to be attractive in short range bioassays (indicated by black bars in Figure 2). Of the nine highest-ranked chemicals in Experiment 1 (Table 1), only compound 7 had a low amplitude depolarization peak (Figure 2, Group 3). This may have been due to differences in sample preparation between the bioassays and EAG analyses. The former used 10% dilutions in acetone, whereas the latter used chemicals in their neat form. With compound 7, the neat material at 24 • C was in solid state, and the dry crystals may not have generated significant volatiles in the headspace of the EAG sample bottle. It is also possible that compound 7, when presented in bioassays, was detected by contact chemoreceptors on the tarsi rather than by antennal olfactory receptors.
In addition, there were two chemicals that elicited higher than expected EAG responses. Of the twenty low-ranked (i.e., non-attractive) chemicals from the behavioral assays (indicated by gray bars in Figure 2), all displayed weak EAG responses except for compounds 26 and 28, which produced relatively high amplitude EAG peaks (Figure 2, Group 3). Potential explanations for this observation are that these two chemicals may be (i) true attractants, but the insects are not at the proper physiological stage to respond appropriately (all test insects were sterile virgin males of the same age); (ii) synergistic attractants, not behaviorally active alone but increasing response when combined with primary attractants (e.g., putrescine, a synergist when combined with ammonia as a tephritid food-based attractant [37]; (iii) repellents, which would also be detected by antennal olfactory receptors (e.g., ammonia, a protein feeding cue attractive at low doses but repellent at high dose) [38]; or (iv) other biologically relevant compounds in the environment but unrelated to attraction behavior. age); (ii) synergistic attractants, not behaviorally active alone but increasing response when combined with primary attractants (e.g., putrescine, a synergist when combined with ammonia as a tephritid food-based attractant [37]; (iii) repellents, which would also be detected by antennal olfactory receptors (e.g., ammonia, a protein feeding cue attractive at low doses but repellent at high dose) [38]; or (iv) other biologically relevant compounds in the environment but unrelated to attraction behavior.

Natural and Synthetic Compounds
The compounds used for the study (Figure 1) are natural and synthetic aromatic derivatives with different functional groups. In particular, piceol (1) and acetovanillone (2) were purified from the organic extract of Crinum buphanoides bulbs, a native South African Amaryllidaceae plant [21]. Tyrosol (3) and resorcinol (4) were isolated from the culture filtrate of the fungus Dothiorella vidmadera, a pathogen involved in the Botryosphaeria dieback of grapevine [22]. 4-Hydroxybenzaldehyde (5) was isolated from the solid cheatgrass (Bromus tectorum) culture of a Fusarium strain belonging to the F. tricinctum species complex [23]. 2-allylanisole (26). An ethereal solution of CH 2 N 2 was added to a solution of 2-allylphenol (23, 120 µL) in MeOH (120 µL) to obtain a persistent yellow color. The reaction was carried out at room temperature under stirring and was stopped after 24 h by evaporation under an N 2 stream. The crude residue (130.1 mg) was purified by CC, using n-hexane:EtOAc (9:1) as eluent, to give 100 mg of 2-allylanisole (23) as a homogeneous compound. Its 1 H NMR data were very similar to those already reported in literature [25];

Insects
Sterile male C. capitata were obtained from the Programa Moscamed mass rearing facility (El Pino, Guatemala), where they were irradiated as pupae 2 d prior to emergence with 95 Gy of gamma radiation from a Co60 source. These are the temperature-sensitive lethal strain flies [39] that are used for the preventative release program [40] in Florida. Thus, only males were obtained, and only virgin males were used for testing. Irradiated pupae were shipped initially to the USDA-APHIS Medfly Project (Sarasota, FL, USA) and then to the USDA-ARS Subtropical Horticulture Research Station in Miami, FL. Holding conditions at Miami consisted of a 12/12 h L/D photoperiod, 25 ± 2 • C, and 75 ± 5% RH. Pupae were placed in collapsible cages (30.5 × 30.5 × 30.5 cm). After eclosion, adult flies were provided with water (2% agar blocks) and food (3:1 mixture of cane sugar and yeast hydrolysate). Flies used for all studies were 5 to 10 d-old, sexually mature sterile virgin males. Only sterile flies were available for use in this research because there are no wild populations in Florida. Previous research, however, has found that response of sterile males to semiochemicals is similar to response of wild males (e.g., Reference [15]).

Short-Range Bioassays
Small cage bioassays were used to quantify the short-range attraction of sterile male C. capitata using a modified version of the binary choice tests [41]. All observations were carried out at room temperature as described above in small collapsible cages (20.3 × 20.3 × 20.3 cm) into which 50 flies were introduced 1 h prior to the start of each experiment. Tests were initiated by introducing two Petri dishes (53 mm diameter and 12 mm height) with substrates positioned symmetrically (37 mm apart). After 30 min, the number of flies at each dish was recorded. Experiment 1 compared the response to each individual chemical (10 µL of a 10% dilution in acetone) with the response to a paired solvent control (10 µL acetone). Test substrate or control was added to the center of a filter paper disk (Whatman #1, 3.5 cm diam). The filter paper disk was air-dried briefly to allow the solvent to evaporate and was placed into the middle of a Petri dish. Bioassays were replicated five times, and the position of substrates reversed between replicates. Flies and Petri dishes were used only once, and cages were washed with acetone between experiments to eliminate potential residual chemicals.
Pairwise comparisons of the chemicals that elicited the highest response in the initial tests were then conducted in Experiment 2 (compounds 7, 21, 23, and 24). The response to each individual chemical (10 µL of a 10% dilution in acetone) was compared with the response to each other selected chemicals in this two-choice test bioassay with all possible combinations tested (7 vs. 21, 7 vs. 23, 7 vs. 24, 21 vs. 23, 21 vs. 24, 23 vs. 24). There were ten replicates of the pairwise comparisons, with each pair tested in separate cages at the same time.

Electroantennography (EAG) Analysis
Peripheral olfactory responses were recorded from antennae of male C. capitata using a Syntech EAG system (Syntech Original Research Instruments, Hilversum, Netherlands) and methods developed by Kendra et al. [37,38,42]. Test substrates consisted of the 29 compounds, each 20 mg neat material. The standard reference sample (positive control) was tea tree oil, 20 mg (Essential Oil India-SAT Group, Kannauj, India), shown previously to elicit strong EAG responses in male medflies [15]. Each substrate was placed into a separate 250 mL hermetic glass bottle equipped with a lid containing a short thru-hull port (Swagelok, Solon, OH, USA) and silicone septum (Alltech, Deerfield, IL, USA). Sample bottles were sealed and equilibrated overnight at 24 • C to allow for headspace saturation with volatiles.
Freshly dissected antennal preparations (whole head mounts) were secured between electrodes using salt-free conductive gel (Spectra 360, Parker Laboratories, Fairfield, NJ, USA) and placed under a stream of humidified air, purified with activated charcoal granules, at a flow rate of 400 mL/min. Using gas-tight syringes (SGE Analytical Science, Victoria, Australia), samples of saturated vapor were withdrawn from the test bottles, injected into the airstream, and presented to the antennae. In each recording session, samples (fixed 1 mL doses) were delivered in the following order: the tea tree standard, test chemicals in random order, a clean air injection (negative control), and a final standard injection. There was a 2-min interval (clean air flush) between injections to prevent antennal adaptation (diminished EAG response resulting from repeated exposure to chemical stimuli). Due to the large number of test chemicals, EAG analyses were conducted using four groupings; each group compared olfactory responses to seven or eight chemicals, randomly chosen, and responses were measured from ten replicate females.
EAG responses to test substrates were measured initially in millivolts (peak height of depolarization) and then normalized to percentages relative to the EAG response obtained with the reference sample. Normalization corrects for time-dependent variability (gradual decline) in antennal performance and allows for comparison of relative EAG responses obtained with different substrates [37,[43][44][45] and with different cohorts of insects [37,38]. Finally, any response recorded with the negative control was subtracted from the normalized test responses to correct for "pressure shock" caused by injection volume. All statistical analyses were performed using the corrected normalized EAG values.

Statistical Analysis
Pair t-tests were used to test for differences in number of males attracted to each choice in the binary choice tests in Experiments 1 and 2 (Proc TTEST; SAS Institute, 2016) [46]. Male response was converted to Attraction Index (number attracted to the compound minus number attracted to the control divided by total number of males tested) [47] to compare the eight compounds that attracted more males than the paired control in Experiment 1. One-way ANOVA and Tukey's mean separation tests were used to determine effect of chemical on Attraction Index in Experiment 1 and on olfactory responses in EAG analyses. When necessary, data were transformed prior to ANOVA to satisfy conditions of equal variance [48]; non-transformed means ± standard deviations are presented.

Conclusions
In an effort to find effective new attractants for C. capitata, we investigated 29 structurally related aromatic compounds in short range bioassays and EAG analyses. The combined results identified phenyllactic acid (7), estragole (24), o-eugenol (21), and 2-allylphenol (23) as promising candidates for sexually mature males. Of these four compounds, o-eugenol (21) was observed to be the most attractive in binary choice tests. The presence of the allyl residue and substituents on the aromatic ring appear to be key structural features that confer attraction to these compounds. This study provides insight into the attractiveness of structural variants of aromatic compounds with various substituent groups to male C. capitata. Another promising approach could be the synthesis of estragole analogs with allyl groups at different sites on the aromatic ring. In addition, further studies are needed to evaluate these compounds, alone and in combination, to determine their efficacy in the field.