The Role of Honey Bee Derived Aliphatic Esters in the Host-Finding Behavior of Varroa destructor

Simple Summary Honey bees provide essential pollination services for many agricultural crops and wild plants. However, honey bee health is at risk. As an obligate ectoparasite of honey bees, Varroa destructor has been causing huge damage to the worldwide Western honey bee colonies since the host shift from its original host, the Eastern honey bees. For years, a lot of effort has been made to control this parasite; however, more efficient and safer methods are still urgently needed. The current study also aimed to contribute to the Varroa mitigation by seeking the potential compounds that affected the host-finding behavior of Varroa mites. By profiling the headspace volatiles of honey bee larvae, four aliphatic esters were commonly identified from both worker and drone larvae of the original and new hosts. Among the esters, ethyl myristate was a novel compound and able to significantly attract the Varroa mites. The findings presented in this paper will deepen our understanding of the Varroa host finding behavior and provide new chemicals for the monitoring and control of V. destructor in the future. Abstract Varroa destructor is an obligate ectoparasite of honey bees and shifted from its original host Apis cerana to the new host Apis mellifera in the first half of the twentieth century. The host shift has resulted in a great threat to the health and survival of A. mellifera colonies worldwide. Chemical signals play a crucial role in all aspects of the Varroa life cycle, including host finding. However, the chemical cues that affect the host finding behavior of Varroa mites are still not fully understood. In this study, we systematically profiled the headspace volatiles of both worker and drone larvae of the two honey bee species by using solid phase micro-extraction coupled to gas chromatography-mass spectrometry (SPME-GC-MS), and then used electrophysiological recording and Y-tube olfactometer bioassay to study the potential roles of the selected compounds. The chemical profiling showed that there were four aliphatic esters, ethyl myristate (EM), methyl palmitate (MP), ethyl palmitate (EP), and ethyl oleate (EO) commonly detected from all four types of larval hosts. Among them, EM was a new substance identified from honey bee headspace volatiles. Results from electrophysiological recordings indicated that all the aliphatic esters could elicit significant responses of Varroa pit organs on its forelegs. Moreover, behavioral analyses revealed that EM could significantly attract V. destructor at a medium dosage (10 µg), while MP had no observable effect on the mites and both EP and EO were able to repel the parasites. Our findings suggest an important role of host-derived aliphatic esters in Varroa host finding, and provide new chemicals for Varroa monitoring and control.

larvae [32]. Moreover, the fourth and fifth instar larvae had different CHC profiles and distinct attractiveness to Varroa mites, indicating that the parasites may determine the host age by sensing CHCs [33]. More recently, Li et al. (2022) [34] revealed that the larval CHC profiles between the original host (A. cerana) and new host (A. mellifera) were significantly different, which gave rise to the host preference of V. destructor. Varroa mites can be attracted by not only host-derived compounds, but also the semiochemicals from brood-related sources such as cocoon and larval food. On the other hand, there are also chemicals inside the honey bee hives that show a repellent effect on the parasites [11,17]. In brief, chemical stimuli that affect the host finding behavior of Varroa mites are complex and diverse, and despite the known compounds it is likely to identify novel chemicals that may have stronger attractiveness/repellences to V. destructor.
In the current study, we assume that there are still some new compounds that can be derived from honey bee hosts and affect Varroa host finding behavior. We used a more recently-developed method to enrich and analyze the headspace volatiles from the worker and drone larvae of both the original and new host species. Since both drone and worker larvae in both species of honey bees can effectively attract Varroa mites [34,35], we compared and selected the common volatiles and then performed several electrophysiological analyses and Y-tube olfactometer bioassays. The forelegs of V. destructor were used to record the electrophysiological responses as the pit organ located on the foreleg tarsi was found to sense the headspace volatiles of adult honey bees [36,37]. We found out one novel and three known compounds, and revealed their attractive/repellent effect on Varroa mites during host finding.

Biological Samples
Both A. mellifera (Am) and A. cerana (Ac) colonies used in this study were kept in the same experimental apiary of the Institute of Zoology, Guangdong Academy of Sciences, Guangzhou, China. The A. mellifera colonies were further classified and labeled as Varroafree or Varroa-infested colonies. Early fifth instar larvae were sampled by uncapping the sealed brood cells and collecting those with bowed backs and not spinning silk yet. Both worker and drone larvae were obtained from Varroa-free A. mellifera and A. cerana colonies.
The female adults of V destructor were taken from sealed brood cells in heavily-infested A. mellifera colonies. All Varroa mites were maintained on the newly emerged bees for 48 h before physiological or behavioral tests.

Chemical Analysis
Headspace volatiles on the body surface of honey bee larvae were sampled using solid phase micro-extraction (SPME), and analyzed by gas chromatography-mass spectrometry (GC-MS). Briefly, for the SPME sampling, individual honey bee larvae sampled as described above were sealed in a 20 mL glass vial and 1 µL of 1-nonene ethanol solution (1:10,000, v/v) was also added to the vial as an internal standard. The headspace extraction was performed with a 50/30 µm DVB/CAR/PDMS stableflex SPME fiber (Supelco, Bellefonte, PA, USA) at 34 • C for 6 h. The enriched volatile compounds were then subjected to GC-MS analysis on an Agilent 7890B GC system (equipped with an HP-5MS UI column; film thickness: 0.25 µm; length: 30 m; inner diameter: 0.250 mm) connected to an Agilent 5977 MSD (quadrupole mass spectrometer with 70-eV electron impact ionization). The back inlet was at 260 • C and 7.06991 psi, and the carrier gas helium was provided at a constant Insects 2023, 14, 24 4 of 13 flow rate of 34 mL/min. Samples were added in splitless mode. Oven temperature was programmed at the following temperatures: 40 • C to 48 • C at 4 • C/min, 48 • C to 60 • C at 1.5 • C/min, 60 • C to 82 • C at 4 • C/min, 82 • C to 100 • C at 1.5 • C/min, 100 • C to 200 • C at 4 • C/min, and 200 • C to 260 • C at 6 • C/min. The mass spectrometer was set to scan from 33-390 m/z. The MS source and quadrupole were maintained at 230 • C and 150 • C, respectively. GC-MS data were analyzed using MSD ChemStation (G1701FA F. 01. 03. 2357) in MassHunter software suites (Agilent Technologies, Palo Alto, CA, USA). Compounds were identified by searching spectra against the NIST mass spectral libraries (NIST17) in MassHunter. The peak area of each compound was obtained by automatic integration of each total ion chromatogram (TIC) using a consistent baseline setting. The concentration of every compound was calculated by comparing with the peak area of the known amount of internal standard 1-nonene [38].

Electrotarsogram (ETG) Recordings
ETG recordings were performed on V. destructor foreleg exposed to test chemicals at room temperature, according to methods described in previous studies with modifications [36,39]. Briefly, the foreleg was excised at the base under a Zeiss Discovery V8 dissection microscope (Zeiss, Jena, Germany) and mounted to each side of a two-pronged electrode "fork" recording probe with a small amount of electrode gel (Spectra 360, Parker Laboratories, Fairfield, NJ, USA). A humidified and charcoal-filtered air flow provided by a stimulus controller (Syntech CS-55, Hilversum, The Netherlands) was constantly blown towards the organ at 30 mL/s at a distance of 1 cm.
To prepare odor cartridges, 10 µL of test or control solution was added to a piece of Whatman No. 1 filter paper (1 cm × 1 cm) which was placed in a glass Pasteur pipette and exposed to air for 30 s for evaporation of the solvent. Chemicals were dissolved in hexane to prepare the test solutions (100 mg/mL). In order to assess dosage response, a serial dilution was carried out to generate different concentrations of stimuli (0.01, 0.1, 1, 10 and 100 mg/mL). Therefore, dosages of each stimulation were 0.1, 1, 10, 100, and 1000 µg.
Each odor was puffed at the foreleg for 200 ms, followed by a two-minute interval to allow recovery of the organ. The resulting response was amplified and recorded by a signal acquisition interface board (IDAC-2, Syntech, Hilversum, The Netherlands), and then the recordings were analyzed by AutoSpike software (Syntech, Irvine, CA, USA). The olfactory response for each test stimuli was calibrated using the mean value of the two responses of control (solvent alone).

Y-Tube Olfactometer Bioassays
A Y-tube olfactometer bioassay was designed to test the behavioral response of V. destructor to the volatiles derived from honey bee larvae, modified from Pernal et al. (2005) [40]. The main part of the olfactometer was a glass Y-tube (stem and arms: length 5 cm, diameter 2 cm). A marker was made on the stem of the Y-tube, 3 cm from the junction of the "Y", where a mite would be introduced. Each arm of the Y-tube was connected to a cylindrical glass odor chamber. The anterior end of each odor chamber could be removed and replaced, thus allowing sample insertion and removal. Ambient air was charcoal-filtered, moistened with distilled water, and then blown into the two arms using a vacuum pump. The velocity of the airflow passing through the glass odor chambers was adjusted to 100 mL/min using an airflow meter (Kean Labor Insurance, Beijing, China). The experimental condition was kept at around 32 • C and 60% RH.
The odorants of different dosages were prepared as mentioned in Section 2.4. An aliquot (10 µL) of the test sample was added onto a piece of filter paper, exposed to the air for 30 s to allow solvent evaporation, and then placed in one odor chamber of the olfactometer, while the same amount of hexane (solvent control) was applied in the other odor chamber. In each test, a single Varroa mite was introduced into the marked point of the olfactometer and allowed 6 min to choose between the two branches. The movement of a tested mite was immediately recorded for 6 min using an iPhone 6 s (Apple Inc., Cupertino, CA, USA). The branch that the mite walked into was first noted, and then the distance between the mite's stop point and the odor chamber was measured. One mite might alternate its walking between branches many times, however, only its first choice was taken into account. A choice was scored positive when a mite entered the arm and either stayed in the odor chamber or walked within 3 cm of the odor chamber. If one mite remained in the stem and did not make a choice within 6 min, it was recorded as "no choice" and not included in the following data analysis. Individual mites were used only once, and at least 30 mites were tested for each treatment. The two odor chambers were randomly reversed to avoid any positional bias. The Y-tube and glass odor chambers were cleaned with ultrapure water, acetone and hexane and then baked in an oven at 80°C for at least 60 min before each trial.

Statistical Analyses
Since the assumptions of analysis of variance (ANOVA) were not met with the data for the amount of four esters and the ETG responses, the nonparametric Kruskal-Wallis test was performed to determine the significant differences, and Dunn's test was used to carry out the post hoc pairwise comparisons. A Chi-square test was applied to analyze the behavioral data from Y-tube olfactometer bioassays. In all cases, a p value < 0.05 was taken to be significant. All statistical analyses were performed using GraphPad Prism software version 8.0 (GRAPH PAD Software Inc, San Diego, CA, USA).

Identification and Comparison of Body Surface Volatiles of Four Kinds of Honey Bee Larvae
In total, there were 63 volatile compounds identified from the early fifth instar worker and drone larvae of the two honey bee species (Table S1). Compared to A. cerana, many more compounds were detected from both worker (36 vs. 17) and drone (36 vs. 20) larvae of A. mellifera, suggesting that this new host expressed more complicated odorant stimuli. Additionally, while the majority (>90%) of total volatiles were not constantly presented in all four types of honey bee larvae, there were four aliphatic esters, ethyl myristate (EM), methyl palmitate (MP), ethyl palmitate (EP), and ethyl oleate (EO) shared by them ( Figure 1).
Further quantitative and comparative analyses were carried out on the levels of these four shared substances among worker and drone larvae of both A. mellifera and A. cerana. Generally, a significant difference was detected for all four chemicals ( Figure 2. Kruskal-Wallis test. For EM: H = 21.39, df = 3, p < 0.0001; For MP: H = 32.29, df = 3, p < 0.0001; for EP: H = 37.00, df = 3, p < 0.0001; for EO: H = 15.49, df = 3, p = 0.0014). For intraspecific comparisons, we found that Ac worker larvae had equal amounts of EM, EP, and EO to Ac drone larvae, but significantly higher level of MP than Ac drone larvae (Dunn's test. MP: p = 0.0032), while Am worker larvae displayed significantly lower levels of all four compounds than Am drone larvae (Dunn's test. EM: p = 0.0028; MP: p = 0.0031; EP: p < 0.0001; EO: p = 0.0022). For interspecific comparisons, Ac worker larvae showed similar levels of EM to Am worker larvae, but significantly higher amount of MP, EP, and EO than Am worker larvae (Dunn's test. MP: p < 0.0001; EP: p = 0.0002; EO: p = 0.0159), while Ac drone larvae had similar levels of MP, EP, and EO to Am drone larvae, but significantly lower amount of EM than Am drone larvae (Dunn's test. EM: p = 0.0127). Insects 2023, 14, x FOR PEER REVIEW 6 of 13  levels of EM to Am worker larvae, but significantly higher amount of MP, EP, and EO than Am worker larvae (Dunn's test. MP: p < 0.0001; EP: p = 0.0002; EO: p = 0.0159), while Ac drone larvae had similar levels of MP, EP, and EO to Am drone larvae, but significantly lower amount of EM than Am drone larvae (Dunn's test. EM: p = 0.0127).

Figure 2.
Comparison of the amount of four aliphatic esters ethyl myristate (A), methyl palmitate (B), ethyl palmitate (C), and ethyl oleate (D) among the worker and done larvae in both Apis mellifera (Am) and Apis cerana (Ac). Boxes include the 25th-75th percentile. Horizontal bars denote the median value. Whiskers indicate the values that are no further than 1.5 × IQR from either the upper or the lower hinge. Each dot represents one larval sample. The sample size (n) is 19, 23, 22, and 26 for Ac worker, Ac drone, Am worker, and Am drone, respectively. Statistical significance is determined by Kruskal-Wallis test followed by Dunn's test for post hoc comparison. n.s. not significant, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

Electrophysiological Responses of Varroa destructor to the Identified Aliphatic Esters
ETG recordings showed that V. destructor responded significantly to all the four aliphatic esters (  for Ac worker, Ac drone, Am worker, and Am drone, respectively. Statistical significance is determined by Kruskal-Wallis test followed by Dunn's test for post hoc comparison. n.s. not significant, * p < 0.05, ** p < 0.01, *** p < 0.001, **** p < 0.0001.

Electrophysiological Responses of Varroa destructor to the Identified Aliphatic Esters
ETG recordings showed that V. destructor responded significantly to all the four aliphatic esters (Figure 3   All the responses to test compounds were calibrated to control (0.5). Each red dot represented one replicate (one mite) and 60 replicates (n = 60) were included for every single dosage and solvent control. Boxes include the 25th-75th percentile. Horizontal bars inside each box show the median value. Whiskers suggest values that are no further than 1.5 × IQR from either the upper or the lower hinge. Kruskal-Wallis tests were employed to determine the significant differences and Dunn's tests were used for post hoc comparison. Every response to the test compounds was compared to solvent control, and a different lowercase letter was labeled to indicate significant differences (p < 0.05).

Olfactory Responses of Varroa destructor to the Identified Aliphatic Esters
We used Y-tube olfactometer assays to detect the olfactory responses of V. destructor to EM, MP, EP, and EO. Behavioral data showed that EM significantly attracted V. destructor at a medium dosage of 10 µg ( Figure 4A. Chi-square test, χ 2 = 9.135, df = 1, p = 0.0025), but repelled the parasites at an ultra-high dosage of 1000 µg (χ 2 = 7.333, df = 1, p = 0.0068). On the other hand, V. destructor had no preference for MP regardless of tested dosages ( Figure 4B. Chi-square test, p > 0.05 for all five treatment concentrations). In contrast to the results observed with EM and MP, EP significantly repelled V. destructor at dosages of 10 µg and 100 µg ( Figure 4C. Chi-square test, 10 µg: χ 2 = 4.500, df = 1, p = 0.0339; 100 µg: χ 2 = 8.000, df = 1, p = 0.0047). Similarly, V. destructor was significantly repelled by EO from low to high dosages ( Figure 4D. Chi-square test, 0.1 µg: χ 2 = 16.00, df = 1, p < 0.0001; 10 µg: χ 2  1, 1, 10, 100, 1000 µg) and solvent control (s) were included. All the responses to test compounds were calibrated to control (0.5). Each red dot represented one replicate (one mite) and 60 replicates (n = 60) were included for every single dosage and solvent control. Boxes include the 25th-75th percentile. Horizontal bars inside each box show the median value. Whiskers suggest values that are no further than 1.5 × IQR from either the upper or the lower hinge. Kruskal-Wallis tests were employed to determine the significant differences and Dunn's tests were used for post hoc comparison. Every response to the test compounds was compared to solvent control, and a different lowercase letter was labeled to indicate significant differences (p < 0.05).

Discussion
Varroa destructor originally parasitized the Eastern honey bee Apis cerana, and then made a host shift to the Western honey bee A. mellifera in East Asia in the first half of the twentieth century, and spread rapidly across the world via global trade [11,41]. Both worker and drone larvae of the old and new hosts can attract the parasite [34,35], which suggests that common stimuli may exist among different host larvae. It was expected that the active chemicals affecting Varroa host finding would be identified more efficiently by screening the shared compounds by honey bee larvae of distinct species and sexes. To this end, we first collected both worker and drone larvae of the two honey bee species from the same apiary, and then profiled the headspace volatiles individually.
Once coming out from brood cells together with emerging bees, Varroa mites enter the phoretic stage and usually perform host-finding behavior as well, since they prefer nurse bees over both newly-emerged bees and foragers [42]. However, it seems more important for Varroa mites to find out the right larval hosts than the adult hosts, because successfully entering a larval cell is a key step for their reproduction and population development. In fact, much more efforts have been made to understand the details of this process. On the other hand, the body surface compounds of adult bees could potentially attract Varroa mites as well, and it would be interesting to see if the acting chemicals are kept the same between larval and adult bees.
In this study, healthy and normal worker and drone larval samples were required for profiling the headspace volatiles. Generally, larval samples from Varroa-free colonies are much healthier than those from Varroa-infested colonies [9,10]. Therefore, no Varroa infestation was taken as a standard of healthy colonies, which were then selected for sample collection. It does not mean that individuals collected from Varroa-free colonies are less attractive than those from Varroa-infested colonies. In addition, it has been shown that the cuticular hydrocarbon profile of worker bees can be modified by the presence of V. destructor in brood cells [43]. As A. cerana has become resistant to Varroa mites after a long term of co-evolution [44], no or very few mites can be seen in A. cerana colonies, so all the A. cerana colonies were considered Varroa-free.
There are two widely used methods for collecting chemicals from the body surface of honey bee larvae: hexane extraction [24,32,34,45], and headspace volatile enrichment

Discussion
Varroa destructor originally parasitized the Eastern honey bee Apis cerana, and then made a host shift to the Western honey bee A. mellifera in East Asia in the first half of the twentieth century, and spread rapidly across the world via global trade [11,41]. Both worker and drone larvae of the old and new hosts can attract the parasite [34,35], which suggests that common stimuli may exist among different host larvae. It was expected that the active chemicals affecting Varroa host finding would be identified more efficiently by screening the shared compounds by honey bee larvae of distinct species and sexes. To this end, we first collected both worker and drone larvae of the two honey bee species from the same apiary, and then profiled the headspace volatiles individually.
Once coming out from brood cells together with emerging bees, Varroa mites enter the phoretic stage and usually perform host-finding behavior as well, since they prefer nurse bees over both newly-emerged bees and foragers [42]. However, it seems more important for Varroa mites to find out the right larval hosts than the adult hosts, because successfully entering a larval cell is a key step for their reproduction and population development. In fact, much more efforts have been made to understand the details of this process. On the other hand, the body surface compounds of adult bees could potentially attract Varroa mites as well, and it would be interesting to see if the acting chemicals are kept the same between larval and adult bees.
In this study, healthy and normal worker and drone larval samples were required for profiling the headspace volatiles. Generally, larval samples from Varroa-free colonies are much healthier than those from Varroa-infested colonies [9,10]. Therefore, no Varroa infestation was taken as a standard of healthy colonies, which were then selected for sample collection. It does not mean that individuals collected from Varroa-free colonies are less attractive than those from Varroa-infested colonies. In addition, it has been shown that the cuticular hydrocarbon profile of worker bees can be modified by the presence of V. destructor in brood cells [43]. As A. cerana has become resistant to Varroa mites after a long term of co-evolution [44], no or very few mites can be seen in A. cerana colonies, so all the A. cerana colonies were considered Varroa-free.
There are two widely used methods for collecting chemicals from the body surface of honey bee larvae: hexane extraction [24,32,34,45], and headspace volatile enrichment [30,[46][47][48]. SPME has evolved into a versatile sample preparation tool, supporting headspace volatile enrichment with high efficiency [49]. Here, we applied SPME technology to sample honey bee larval volatiles, which were analyzed by GC-MS.
We identified more than 60 compounds from all larval headspace volatile samples. The composition of larval volatiles is quite different among the four types of honey bee larvae. In contrast, the majority of the cuticular hydrocarbons (CHCs) were shared by distinct larval types [34]. Despite this significant differentiation of larval volatiles, we found that four straight-chain fatty acid esters were constantly detected in all four larval types, the peaks of which were concentrated in the right end of total ion chromatograms (TICs). Among these, three esters, MP, EP, and EO are well known compounds serving as components of a brood pheromone [28,[50][51][52][53], while EM is newly identified in this study. In addition to these esters, there were some other compounds shared by different larval types, whose function in Varroa host finding is still needing to be explored.
It has long been established that the infestation rate of Varroa mites in A. mellifera drone cells is much higher than that in worker cells [25,26,54], which can be partially explained by the fact that drone larvae are more attractive than worker larvae [24]. Our results demonstrated that A. mellifera drone larvae contained a significantly higher amount of all four esters than worker larvae, which is consistent with previous studies [24,28] and supports the preferred invasion of drone cells by Varroa.
Since V. destructor has no eyes and lives in a dark hive environment, chemical signals play an important role in keeping its tight link with the honey bee hosts [11,23]. The mite also lacks antennae, but its forelegs function as such [23]. Similarly, our ETG recording analyses also showed that Varroa forelegs could respond to all the identified fatty acid esters from larval headspace volatiles, offering a physiological basis for the next behavioral assays. Generally, the strength of the ETG responses goes up with the increase of testing concentrations within a range of dosages, which is determined by the innate characteristics of olfactory neurons [55]. We did see that different dosages of esters had different effects on Varroa behavior. The reasons for this phenomenon are complicated. For one compound, not all concentrations have a consistent behavioral effect, and one dosage can attract Varroa, but the higher dosages just repel the mites. This is partially related to the chemical itself. Y-tube olfactometer bioassay is a classical and widely-used method for investigating the behavioral responses of insects and mites to test odors [56,57]. Here, this method was also applied to determine how Varroa mites responded behaviorally to the four fatty acid esters. We tested the identified compounds individually. It is important to first figure out the role of the identified compounds in affecting Varroa host finding behavior when they were applied solely. It was expected that a chemical alone was able to highly attract/repel the mites. Then, some combinations among the compounds can be carried out in future to obtain higher efficiency in the attraction/repellence of Varroa. Moreover, it would be interesting to know if one identified compound is better than another one in affecting the host finding of V. destructor.
For the new compound EM, we found that it significantly attracted Varroa mites at medium dosage, while it repelled the mites at extra high dosage. In regard to MP, no response was displayed by Varroa mites, which is consistent with earlier findings [29,30,58], but is inconsistent with that of Calderone and Lin (2001) [59] and Le Conte et al. (1989) [24]. Surprisingly, both EP and EO showed a repellent effect on the mites, which also contradicts the findings from Calderone and Lin (2001) [59] and Le Conte et al. (1989) [24]. These differences could be due to different testing instruments, testing conditions, or Varroa mites of different stages or physiology. Therefore, it is necessary to have standards for Varroa olfactory-related behavioral bioassays, so that results from different laboratories can be compared. The behavioral assay indicated that the aliphatic esters were not always attractive to Varroa mites, and some of them showed repellent effect on the mites. It makes it difficult and sometimes confusing to predict the total effect of the esters on the mites. Actual honey bee hives are a complex odorant world, and these volatiles are always emitted and function together, which is not like the behavioral assay conducted in laboratory where the odors can be tested one by one. We speculate that when combined volatiles work together, one compound generates the main influence while others have no or weak effect on the mites. However, further verifications are needed when using different chemical combinations.
Chemical stimuli that affect Varroa host-finding behavior have drawn long-term and intensive studies around the world. However, considering the complexity and diversity of chemical cues, we believe that there will be more chemical cues identified in the future, especially with the rapid development of chemical analysis technologies. Despite more standardized behavioral bioassays conducted in the laboratories, further field tests for the effectiveness and applicability of the signaling compounds and their potential combinations should be encouraged to produce attractants/repellents for Varroa monitoring and control. Especially as a novel compound, EM provides excellent potential for developing Varroa attractants, which would be safe and useful for monitoring and controlling the Varroa population by efficiently attracting and arresting the parasites. Funding: This work was supported by the GDAS Special Project of Science and Technology Development (grand numbers 2019GDASYL-0104019; 2022GDASZH-2022010106), the National Natural Science Foundation of China (grand numbers 32000347), the Guangzhou Science and Technology Plan Project (grand numbers 202103000009; 20212100047), and the Talent Project of Guangdong Institute of Applied Biological Resources (grand numbers GIABR-gjrc201702). The funding providers had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.