Electrophysiological and Behavioral Responses of Apis mellifera and Bombus terrestris to Melon Flower Volatiles

Simple Summary Honeybees (Apis mellifera) and bumblebees (Bombus terrestris) are often used to pollinate melon flowers in facilities. The researchers identified the volatiles in male and female flowers of a common melon species (Cucumis melo) grown in facilities and measured the EAG and behavioral responses of honeybees and bumblebees when exposed to isolated volatiles from the melon flowers. These results provide basic data for the rational utilization of bees. Abstract As important pollinators, honeybees and bumblebees present a pollination behavior that is influenced by flower volatiles through the olfactory system. In this study, volatile compounds from melon flowers were isolated and identified by headspace solid-phase microextraction (HS-SPME) and gas chromatography–mass spectrometry (GC-MS), and their effects on Apis mellifera and Bombus terrestris were investigated by electroantennogram (EAG) and behavior tests (Y-tube olfactometer). The results showed that 77 volatile compounds were detected in melon flowers, among which the relative content of aldehydes was the highest (61.34%; 82.09%). A. mellifera showed a strong EAG response to e-2-hexenal, e-2-octenal, and 1-nonanal. B. terrestris showed a strong EAG response to e-2-hexenal, e-2-octenal, 2,5-dimethyl-benzaldehyde, benzaldehyde and benzenepropanal. In behavior tests, the volatiles with the highest attractive rate to A. mellifera were e-2-hexenal (200 μg/μL, 33.33%) and e-2-octenal (300 μg/μL, 33.33%), and those to B. terrestris were e-2-hexenal (10 μg/μL, 53.33%) and 2,5-dimethyl-benzaldehyde (100 μg/μL, 43.33%). E-2-hexenal and e-2-octenal were more attractive to A. mellifera than B. terrestris, respectively (10 μg/μL, 10 μg/μL, 200 μg/μL). In conclusion, the volatiles of melon flowers in facilities have certain effects on the electrophysiology and behavior of bees, which is expected to provide theoretical and technical support for the pollination of A. mellifera and B. terrestris in facilities.


Introduction
Honeybees and bumblebees are regarded as the ideal pollinators in the world, playing an important role in improving ecological economic benefits and protecting the balance of the ecosystem [1][2][3]. They pollinate fruit and vegetable crops and play an important role in increasing yields and incomes [4][5][6].
Plant flowers regulate the interaction between pollinators and plants by releasing volatiles. Pollinators can identify volatiles of flowers according to their nutrition and reproductive needs, judge distance and location, and conduct pollen collection [7,8]. The volatile compounds of flowers have important effects on pollinator searching behavior [9,10]. Pollinators and flowering plants interact [11]. Pollinator collection relies not only on vision but also on olfactory signals. It has been found that pollinator visitors rely more on olfactory signals when visiting flowers [12,13]. Therefore, flower volatiles play a key role in the interaction between pollinators and plants.

EAG Recording
The standard compounds for EAG recording were determined according to the identification results of volatile compounds in female and male flowers of melon. The standard compounds for EAG recording are shown in Table 1. The antennae of A. mellifera and B. terrestris were stimulated by volatiles, and antennae potentiometers (Ockenfels Syntech GMBH, Buchenbach, Germany) were used to measure potential changes to illustrate the sensitivity of antennae to volatiles.
A pair of bee antennas were cut with a surgical blade and connected to an electrode with conductive adhesive (SpectraR360). The electrode was connected to a DC/AC amplifier (Syntech IDAC-4), and the output end of the amplifier signal was connected to a computer. The filter paper (3 cm × 1.5 cm) was folded into a "V" shape, 10 µL of the sample to be measured was dropped onto the filter paper and put into the sample tube, and the end of the sample tube was connected with the odor stimulation control device. The antennae potentiometer stimulus gas controller (Syntech CS-55) sent out two test air streams blowing together to the separated antennae: one was a continuous air stream directly blown out after filtration and humidification, and the other was a stream carrying the pungent odor of the sample. The airflow port and antennae were vertical, about 9 mm apart. The stimulated flow was 40 mL/min and the continuous flow was 500 mL/min.
With liquid paraffin as the solvent, the standard compound was prepared into 6 concentration gradients of 10, 100, 200, 300, 400, and 500 µg/µL to be tested, and the test was carried out with paraffin as control. Each antenna was alternately tested by liquid paraffin and sample, with the sample concentration from low to high. The antennae were replaced after one repeat of each sample, and the test was repeated 3 times in total. To prevent the antennae from being stimulated by the sample, bees should be kept away from the sample before each cutting of the antennae. The response of bees to volatiles was where EAG(X) is the amplitude (mV) of the EAG response to a compound and EAG(std) is the amplitude (mV) of the EAG response to the reference liquid paraffin of each recording session.

Behavior Tests
Based on EAG determination, the samples that could induce a strong reaction of the antennae of A. mellifera and B. terrestris were selected as the test compounds for olfactory behavior response ( Table 1). The concentration gradient of the compounds to be measured is the same as that at the EAG recording.
Behavior experiments were carried out using a Y-tube olfactometer (Nanjing Possum Instrument Co., Ltd., Nanjing, Jiangsu, China) in a dark room. The tests were conducted between 08:00 and 12:00 am. The sequence of gas passing through the device was as follows: air pump (Beijing Kean Labor Bao New Technology Company, Beijing, China) → activated carbon drying tube → distilled water humidifying bottle → gas flow meter → volatile matter flavor source bottle → Y-tube olfactometer. The devices were connected by silica gel tubes. Then 10 µL samples and liquid paraffin (control) were added to two pieces of filter paper prepared in advance (3 × 1.5 cm 2 ), and the two pieces of filter paper were placed in the flavor source bottle as the odor source, and the gas flow rate was 200 mL/min. The air pump was turned on for about 10 s to fill the Y-tube olfactometer with gas. The bees were then placed in the middle of the Y-tube olfactometer, and when the bees were familiar with the environment around the Y-tube olfactometer, the air pump was turned on again to observe the bees' selection of smell and record it.
The evaluation conditions of bees' odor selection were as follows: the observation time of each collecting bee was at least 10 min, and if the collecting bee entered the odor bottle or stayed in the front 1/3 area of the odor bottle for more than 4 min, it was regarded as the selection of samples in the odor bottle. Samples and paraffin in the vials should be replaced every time they are used (5 honeybees per test). The test was repeated for 3 groups with 10 honeybees in each group. After testing two honeybees, the olfactory machine was reversed or replaced to eliminate interference. After one sample was tested, the odor bottle and olfactory apparatus were cleaned with 75% ethanol and distilled water to eliminate residual odors for future use.
The results of behavioral tests were expressed by the repellent rate and attractive rate. The following equation is used to calculate the attractive rate and the repellent rate: Attractive Rate (%) = total number of bees in treatment arm/total number of bees tested × 100% Repellent Rate (%) = total number of bees in control arm/total number of bees tested × 100% 2.6. Statistical Analysis EAG relative response values (rEAG), attractive rate, and repellent rate were calculated by the formula as described in the Method section. The results of EAG tests were analyzed by one-way ANOVA, Fisher's protected least-significant-difference (LSD) multiple-comparison procedure was used to analyze the EAG response of A. mellifera and B. terrestris to volatiles at different concentrations. Independent sample t-test was used to analyze the difference in EAG response among different volatiles at the same concentration and different bee species. The results of behavioral tests were analyzed by the chi-square test. SPSS21 was used for data analysis, and Excel and GraphPad5.0 were used for drawing.

Identification of Volatile Compounds
The results of volatile compound identification showed that 77 volatiles were detected in the female and male flowers of melon, including aldehyde, alkane, ether, alkyne, alkene, ester, alcohol, aromatics, ketone, and others ( Table 2). There were similarities and differences in the types and contents of volatiles in flowers.

EAG Response of A. mellifera
A. mellifera had different rEAG of the same volatiles in different concentrations. The rEAG of A. mellifera to 3-phenyl-2-propenal and e-2-decenal at high concentrations were significantly higher than that at low concentrations (p < 0.05). The rEAG of dimethyl phthalate and heptadecane at low concentrations were significantly higher than that at high concentrations (p < 0.05). The rEAG of benzaldehyde and e-2-hexenal showed an increasing trend and then decreased with the increase in the concentration. For 1-nonanal and phytol, their rEAG increased with the increase in the concentration and reached the maximum (Table A1).
For A. mellifera, there were some volatiles with high rEAG, including 1-nonanal, e-2hexenal, e-2-octenal, and benzaldehyde. The rEAG of A. mellifera to 1-nonanal was the highest among all the concentrations. At the concentration of 10 μg/μL, the rEAG of 1nonanal was significantly higher than that of phytol (p < 0.05). At the concentration of 100 μg/μL to 300 μg/μL, the rEAG of 1-nonanal was significantly higher than that of e-2-hexenal and benzaldehyde (p < 0.05). (Table A1). Fifteen compounds were detected in both female and male flowers of melon. The relative contents of these volatiles accounted for 64.87% and 83.02% of the total volatiles in female and male flowers, respectively. Benzaldehyde had the highest relative content in both female and male flowers of melon (48.83% and 74.39%, respectively) (Table 3, Figure 1).

EAG Response of A. mellifera
A. mellifera had different rEAG of the same volatiles in different concentrations. The rEAG of A. mellifera to 3-phenyl-2-propenal and e-2-decenal at high concentrations were significantly higher than that at low concentrations (p < 0.05). The rEAG of dimethyl phthalate and heptadecane at low concentrations were significantly higher than that at high concentrations (p < 0.05). The rEAG of benzaldehyde and e-2-hexenal showed an increasing trend and then decreased with the increase in the concentration. For 1-nonanal and phytol, their rEAG increased with the increase in the concentration and reached the maximum (Table A1). For A. mellifera, there were some volatiles with high rEAG, including 1-nonanal, e-2-hexenal, e-2-octenal, and benzaldehyde. The rEAG of A. mellifera to 1-nonanal was the highest among all the concentrations. At the concentration of 10 µg/µL, the rEAG of 1-nonanal was significantly higher than that of phytol (p < 0.05). At the concentration of 100 µg/µL to 300 µg/µL, the rEAG of 1-nonanal was significantly higher than that of e-2-hexenal and benzaldehyde (p < 0.05). (Table A1). terrestris had different rEAGs of the same volatiles in different concentrations. The rEAGs of 2,5-dimethyl-benzaldehyde, 1,2-dimethoxy-4-(1-propenyl)-benzene, and dodecanal at high concentrations were significantly higher than those at low concentrations (p < 0.05). The rEAG of 1-nonanal increased with the increase in the concentration and reached the maximum when the concentration was 400 µg/µL (Table A2).
The rEAG of A. mellifera to 1-nonanal was the highest among all the concentrations. At the concentration of 10 µg/µL, the rEAG of B. terrestris to 3-phenyl-2-propenal was significantly higher than that of A. mellifera. At the concentration of 300 µg/µL, the rEAG of B. terrestris to 2,5-dimethyl-benzaldehyde was significantly higher than that of A. mellifera. At the concentration of 500 µg/µL, the rEAG of B. terrestris to dodecanal was significantly higher than that of A. mellifera. For other compounds, there were no significant differences between the two bee species (Figure 4).

Behavior Tests of A. mellifera
According to the EAG response test results of A. mellifera, three compounds with higher relative reaction values of EAG were selected for the behavior test: e-2-hexenal, e-2-octenal, and 1-nonanal.

Behavior Tests of B. terrestris
According to the EAG response test results of B. terrestris, five compounds with higher relative reaction values of EAG were selected for the behavior test: e-2-hexenal, e-2-octenal, benzaldehyde, 2,5-dimethyl-benzaldehyde, and benzenepropanal.
At the concentration of 10 μg/μL, the rEAG of B. terrestris to 3-phenyl-2-propenal was significantly higher than that of A. mellifera. At the concentration of 300 μg/μL, the rEAG of B. terrestris to 2,5-dimethyl-benzaldehyde was significantly higher than that of A. mellifera. At the concentration of 500 μg/μL, the rEAG of B. terrestris to dodecanal was significantly higher than that of A. mellifera. For other compounds, there were no significant differences between the two bee species (Figure 4).

Behavior Tests of A. mellifera
According to the EAG response test results of A. mellifera, three compounds with higher relative reaction values of EAG were selected for the behavior test: e-2-hexenal, e-2-octenal, and 1-nonanal.

Behavior Tests of A. mellifera
According to the EAG response test results of A. mellifera, three compounds with higher relative reaction values of EAG were selected for the behavior test: e-2-hexenal, e-2-octenal, and 1-nonanal.

Comparison of Behavior Tests between A. mellifera and B. terrestris
Both e-2-hexenal and e-2-octenal were used in behavior tests of A. mellifera and B. terrestris. The attractive rate of e-2-hexenal to B. terrestris at 10 µg/µL (53.3%) was higher than that of A. mellifera (23.3%), but at other concentrations, A. mellifera was higher. The maximum attractive rate of e-2-hexenal to B. terrestris was at a low concentration of 10 µg/µL, but to A. mellifera was found at 200 µg/µL (Figure 7a). E-2-octenal was more attractive to B. terrestris than A. mellifera at concentrations of 10 µg/µL and 200 µg/µL. E-2-octenal had the maximum attractiveness to B. terrestris at 200 µg/µL and A. mellifera at 300 µg/µL (Figure 7b).

Discussion
In the nature, pollinators are attracted or shunted by factors such as the scent, shape, and color of flowers [16]. Floral fragrances are secondary metabolites released by flowering plants and are composed of volatile organic compounds (VOCs) mixed in varying proportions [17]. Volatiles released by flowers can influence the pollination behavior of pollinators such as bees [10,[18][19][20]. The flower smells mainly include fragrant and malodorous types; however, both of these can help pollinate plants. For example, honeybees and butterflies can identify floral aromas to find reward food, and fetid odors will attract flies or beetles [8,18,[20][21][22].

Identification of Volatile Compounds from Melon Flowers
In this study, the volatiles of male and female melon flowers were analyzed by SPME and GC-MS. A total of 77 compounds in 10 groups were detected in the female and male flowers of melon: ethers, alkynes, alkanes, alkenes, esters, aldehydes, alcohols, aromatics, ketones, and others. Among these volatiles, aldehydes, such as benzaldehyde and e-2hexenal, had the most kinds and the highest relative content. Aldehydes are ubiquitous in flowering plants and play an important role in helping insects locate plants, avoiding pests, and attracting pollinators [23]. The relative content of benzaldehyde in female and male melon flowers was the highest (48.83%; 74.39%). Benzaldehyde is a common volatile organic compound and is the main component of flower fragrance in several crop plants [24], with an extremely complex chemical structure and a special aroma, mostly found in plant stems, leaves, and seeds. Fernandes identified both male and hermaphrodite flowers of five commercial types of melon Cucumis melo (Cantaloupe, Charentais, Galia, Piel de Sapo, and Yellow) and found only 1-2 aldehydes, of which benzaldehyde was identified only in Cantaloupe and Galia, and benzaldehyde was the most abundant among all volatiles in Cantaloupe hermaphrodite flowers [25]. The contents and types of volatile compounds in melon flowers of these varieties are different, which may be due to the genotypes, or environmental factors such as temperature and humidity, soil, water conditions,

Discussion
In the nature, pollinators are attracted or shunted by factors such as the scent, shape, and color of flowers [16]. Floral fragrances are secondary metabolites released by flowering plants and are composed of volatile organic compounds (VOCs) mixed in varying proportions [17]. Volatiles released by flowers can influence the pollination behavior of pollinators such as bees [10,[18][19][20]. The flower smells mainly include fragrant and malodorous types; however, both of these can help pollinate plants. For example, honeybees and butterflies can identify floral aromas to find reward food, and fetid odors will attract flies or beetles [8,18,[20][21][22].

Identification of Volatile Compounds from Melon Flowers
In this study, the volatiles of male and female melon flowers were analyzed by SPME and GC-MS. A total of 77 compounds in 10 groups were detected in the female and male flowers of melon: ethers, alkynes, alkanes, alkenes, esters, aldehydes, alcohols, aromatics, ketones, and others. Among these volatiles, aldehydes, such as benzaldehyde and e-2-hexenal, had the most kinds and the highest relative content. Aldehydes are ubiquitous in flowering plants and play an important role in helping insects locate plants, avoiding pests, and attracting pollinators [23]. The relative content of benzaldehyde in female and male melon flowers was the highest (48.83%; 74.39%). Benzaldehyde is a common volatile organic compound and is the main component of flower fragrance in several crop plants [24], with an extremely complex chemical structure and a special aroma, mostly found in plant stems, leaves, and seeds. Fernandes identified both male and hermaphrodite flowers of five commercial types of melon Cucumis melo (Cantaloupe, Charentais, Galia, Piel de Sapo, and Yellow) and found only 1-2 aldehydes, of which benzaldehyde was identified only in Cantaloupe and Galia, and benzaldehyde was the most abundant among all volatiles in Cantaloupe hermaphrodite flowers [25]. The contents and types of volatile compounds in melon flowers of these varieties are different, which may be due to the genotypes, or environmental factors such as temperature and humidity, soil, water conditions, analysis conditions of volatile compounds, and precision of GC-MS instruments [26].

EAG Response of Bees to Volatiles
Examination of the EAG responses of A. mellifera and B. terrestris to the volatiles of male and female melon flowers showed that all the volatiles could induce electrophysiological responses of both species in different concentrations.
The olfactory system of A. mellifera and B. terrestris is sensitive, and they can show different responses to different concentrations of the same volatiles. The level of odor concentration has a great influence on the electrophysiological responses of the two bee species. In this experiment, the electrophysiological responses of the two bee species to the volatiles of melon flowers showed different trends. The rEAG of A. mellifera to benzaldehyde, e-2-hexenal and e-2-octenal decreased with the increase in the concentration. The relative EAG response of A. mellifera and B. terrestris to 1-nonanal increased with the increase in the concentration. The recognition of odor by A. mellifera and B. terrestris would be affected by the concentration of volatile substances, possibly because the electrophysiological response of A. mellifera and B. terrestris has a certain threshold [36]. The optimal response concentration of A. mellifera and B. terrestris to different volatiles is different.
A. mellifera and B. terrestris on the same concentration of different volatiles have different EAG responses, possibly because of the different physical properties of volatiles, such as the different volatilization rate, which leads to the different time required for antennae to receive chemical signals. It is also possible that honeybees differ in their sensitivity to these volatiles [37], such as honeybees being more sensitive to 1-nonanal than other odors [27].
When the antennae of bees were stimulated by the same volatiles, the EAG response induced by some volatiles was stronger in B. terrestris than in A. mellifera, while others were just the opposite. This is possibly caused by the difference in the physiological structure of smell and smell sensitivity between A. mellifera and B. terrestris [37,38].

Behavior Tests
Behavior tests with the volatiles which could induce the strong EAG response of A. mellifera and B. terrestris showed the reaction of A. mellifera and B. terrestris to different concentrations of volatiles and showed different degrees of attraction and repellency.
The EAG response of A. mellifera to e-2-hexenal was strongest at 200 µg/µL, and the attractive rate (33.33%) was higher than the repellent rate (30.00%) (p > 0.05). The EAG responses of A. mellifera to e-2-octenal and 1-nonanal were the highest at 400 µg/µL, but the attractive rate to liquid paraffin was higher. In the behavior test of B. terrestris, at the concentration of 10 µg/µL, the attractive rate of e-2-hexenal and 2,5-dimethyl-benzaldehyde was higher than the repellent rate (p > 0.05). However, the EAG responses to e-2-hexenal and 2,5-dimethyl-benzaldehyde were the strongest at 100 µg/µL and 300 µg/µL, respectively. The electrophysiological response is not always presented in the behavioral response, and the phenomenon of inconsistency between the EAG results of antennae and the behavioral response often appears in the study of insect olfactory behavior [39][40][41]. It is possible that monomer compounds cannot cause directional reactions, and the combined effects of various odor volatiles should be considered [42,43]. The EAG response of all neurons in the antenna to the compound was measured by the potentiometer, and the response of the sensors in the antenna to the compound was measured by a monocesthesia recorder. Compared with the former, the EAG response values measured by the monocesthesia recorder were more similar to the results of the behavioral experiment [44]. The behavior selection of volatiles by insects is affected by many factors, such as environmental factors and their physiological conditions, rather than by a single active substance, which may be the main reason for the different results.
The attractive rate of e-2-hexenal to B. terrestris (53.3%) at 10 µg/µL was higher than that of A. mellifera (23.3%), but at other concentrations, the attractive rate to A. mellifera was higher. It suggests that the concentration of volatiles could affect the selection behavior of bees. The highest attractive rate of e-2-hexenal to B. terrestris was at the low concentration of 10 µg/µL, but to A. mellifera was found at 200 µg/µL. E-2-octenal had the maximum attraction to B. terrestris at 200 µg/µL and A. mellifera at 300 µg/µL. This may be caused by the difference in odor sensitivity between A. mellifera and B. terrestris [37]. Different volatiles or different volatile concentration can affect the behavior selection of bees, which is speculated to be one of the reasons why bees have a preference for different flowers. At the same time, different bee species have different preferences for flowers, which may also result in different bee species' ability to collect the same crops. It was found that the appropriate concentration of 2,5-dimethyl-benzaldehyde and e-2-hexenal could attract A. mellifera and B. terrestris.
Insects are closely related to the volatiles of flowers, studies have shown that honeybees are attracted to flowers by flowers' color and size, honeybees' olfactory responses, and food rewards [45,46]. Insects may respond to visual stimuli or olfactory stimuli [12,47,48], while others may respond to both [49,50], e.g., honeybees. Moreover, odors were more attractive to honeybees than visual cues [12,51]. In our experiments, we collected and identified plant volatiles, of which we selected monomers and a variety of compounds to conduct electrophysiological and behavioral experiments, then screened out monomers with strong electrophysiological activity and mixed compounds in different proportions. However, since an EAG test can determine whether a single compound has reaction activity to insects and the optimal concentration range, which can reduce the range for the behavioral test and reduce blindness. Only a single compound was used to measure the EAG and behavioral responses of A. mellifera and B. terrestris. Moreover, the indoor behavioral environment is different from the natural environment. Therefore, compounds should be mixed in different proportions in the future to verify the pollination of A. mellifera and B. terrestris under greenhouse conditions, and to provide theoretical and technical support for the pollination of A. mellifera and B. terrestris in the facility cultivation of melon.

Conclusions
Floral fragrances play an important role in the pollination process of honeybees. Thirty-nine compounds (9 groups) and fifty-two compounds (10 groups) were isolated and identified from female and male melon flowers, respectively, among which aldehydes accounted for the most. Different concentrations of volatiles could induce EAG responses of A. mellifera and B. terrestris. A. mellifera showed strong EAG response to e-2-hexenal, e-2-octenal, and 1-nonanal. B. terrestris showed a strong EAG reaction to e-2-hexenal, e-2-octenal, benzaldehyde, 2,5-dimethyl-benzaldehyde, and benzenepropanal. In behavior tests, it was found that the appropriate concentration of e-2-hexenal and 2,5-dimethylbenzaldehyde could attract A. mellifera and B. terrestris.