Effects of dialkoxybenzenes against Varroa destructor and identification of 1-allyloxy-4-propoxybenzene as a promising acaricide candidate

The honey bee is responsible for pollination of a large proportion of crop plants, but the health of honey bee populations has been challenged by the parasitic mite Varroa destructor. Mite infestation is the main cause of colony losses during the winter months, which causes significant economic challenges in apiculture. Treatments have been developed to control the spread of varroa. However, many of these treatments are no longer effective due to acaricide resistance. In a search of varroa-active compounds, we tested the effect of dialkoxybenzenes on the mite. A structure–activity relationship revealed that 1-allyloxy-4-propoxybenzene is most active of a series of dialkoxybenzenes tested. We found that three compounds (1-allyloxy-4-propoxybenzene, 1,4-diallyloxybenzene and 1,4-dipropoxybenzene) cause paralysis and death of adult varroa mites, whereas the previously discovered compound, 1,3-diethoxybenzene, which alters host choice of adult mites in certain conditions, did not cause paralysis. Since paralysis can be caused by inhibition of acetylcholinesterase (AChE), a ubiquitous enzyme in the nervous system of animals, we tested dialkoxybenzenes on human, honey bee and varroa AChE. These tests revealed that 1-allyloxy-4-propoxybenzene had no effects on AChE, which leads us to conclude that 1-allyloxy-4-propoxybenzene does not exert its paralytic effect on mites through AChE. In addition to paralysis, the most active compounds affected the ability of the mites to find and remain at the abdomen of host bees provided during assays. A test of 1-allyloxy-4-propoxybenzene in the field, during the autumn of 2019 in two locations, showed that this compound has promise in the treatment of varroa infestations.


Isolation of AChE from Apis mellifera
Several hundred honey bees were collected off combs of multiple hives from Honey Bee Zen apiary (New Westminster, BC, Canada). The bees were anesthetized with CO2 before being frozen and stored at -80˚C. The heads were removed from 280 nurse bees, and their antennae discarded. Tissues and buffer solutions were kept on ice for as long as possible during the extraction. The phosphate buffer used contained 1% Triton-X-100 (v/v) to dissolve membranebound AChE, and filter-sterilized aprotinin (0.1μ g/mL) to protect isolated proteins from degradation. The bee heads were thoroughly homogenized in 20 mL of phosphate buffer. The homogenate was then centrifuged at 15,000 × g for 40 min, 4ºC. The supernatant was removed and centrifuged for an additional 20 min at 20,000 × g, 4ºC. The supernatant was removed and filter-sterilized (0.45 μm), and 10% sterile glycerol was added before the solution was aliquoted and flash frozen. The AChE isolate was then stored at -80˚C until it was used in assays.
The concentration of the isolated protein solution was determined using the Bradford Assay. Bovine serum albumin (BSA) was obtained from Sigma-Aldrich Canada Co. (Oakville, Ontario), and used as a standard. Eleven concentrations of BSA were prepared between 0.238 μg/mL and 14.29 μg/mL, and their absorbance was measured at 595 nm in the presence of Bradford Reagent. Absorbance over concentration was plotted, and the linear part of the curve (R 2 > 0.99) was determined. This linear portion was used to calculate the concentration of protein isolate based on its absorbance in the presence of Bradford Reagent. The concentration was calculated as 22.25 μg/mL; the value which was used to determine Specific Activity of Honey Bee AChE assays.

Isolation of AChE from Varroa mites
Adult female Varroa mites were collected from mite infested comb from Langley, BC, Canada. Mites were removed from emerging bees or from brood cells with a fine paint brush. All the mites were directly placed in Eppendorf tubes, flash frozen and stored at -80˚C until use. We used 380 mites in this assay.
Whole varroa mites were used for extracting Varroa AChE (VdAChE) because of very small size of the mite. The mites and extracts were kept on ice. First, varroa were homogenized in a sterilized Eppendorf tube using a sealed glass pipette as a pestle. Phosphate buffer pH 7.6 in 1% Triton X-100 was used as a homogenizing solution (5 µl homogenizing solution per mite). The homogenate in the Eppendorf tube was sonicated (Branson Ultrasonic, fitted with a micro tip) and centrifuged at 15000 × g for 5 min at 4˚C for the pellets to settle down. The supernatant was transferred using micropipettes into sterile Eppendorf tubes. The first extraction was done with one third of the total homogenizing solution. The above procedure was repeated for the second extraction and the pellets were placed in the 3 mL glass homogenizer for the last extraction. All the supernatants were collected in one tube and centrifuged again at 15000 × g for 5 min at 4 ˚C. The brown extract solution was aliquoted (100 L) in sterile Eppendorf tubes, by filtration through a 0.45 μm filter. Aliquots were flash frozen in liquid nitrogen and stored at -40° C until further analyses.
Bradford assay: Ten concentrations (0 to 20 µg/ml) of BSA were prepared from 0.1 mg/ml of stock BSA to determine the unknown concentration of the varroa protein extract. Absorbance was measured at 595 nm. Varroa mite extract (10 µl) at dilutions (1:2, 1:4 and 1:10) were used to give absorbance within the BSA absorbance range. The average of protein concentration calculated from all three dilutions was 6.86 µg/mL. This value was used to calculate specific AChE activity of varroa extracts. Figure S1. A. Reaction catalyzed by acetylcholinesterase (AChE). B. Coupled reactions used to detect AChE activity in the Ellman assay.

Determination of the Extinction Coefficient (ε)
We used 2-mercaptoethanol (βME) as a reducing agent for DTNB to determine ε. A 10 mM stock solution of βME in phosphate buffer was prepared. 892 μL of 100 mM Phosphate buffer was combined in a cuvette with 100 μL of 10 mM DTNB and 8 μL of 10 mM βME. The solution was inverted, and absorbance was measured from 300-500 nm. Three runs were performed, and the average wavelength which gave the highest absorbance of TNB 2was 426.3 nm. Consequently, a λmax of 426 nm was used in all AChE assays. The absorbance was then measured at 426 nm for varying βME concentrations from 0.045 mM to 0.005 mM. Three replicate measures of each concentration were performed, and the absorbance was averaged. Absorbance was plotted against βME concentration. The linear portion (R 2 > 0.99) of the data points yielded an ε value of 11792 cm -1 M -1 . This extinction coefficient was therefore used to calculate the concentration of product formation for all AChE assays conducted.

Optimization of Tween 20 Concentration:
Optimization of Tween 20 concentration in our compound stocks was necessary to prevent the precipitation of the compounds during assay runs. Twelve Tween 20 concentrations between 0.02% -1.4% (v/v) were tested over the course of 5 minutes to ensure a lack of precipitation. For each Tween 20 concentration, 0.5 mM, 1 mM, and 3 mM of test compound were tested, and three replicates were performed. The optimal Tween 20 concentration was 1.1% in the assay mixture (stock inhibitor concentration of 55%). This concentration did not result in any precipitation over the course of 5 minutes for 0.5 mM, 1 mM, and 3 mM of any test compound.

Field trials
Placement of release devices Figure S2. Placement of release devices with 3c{3,6} (treatment) or no compound (control). The craft sticks placed horizontally across the top bars of the frames were used to suspend the treatment devices (insert on left) between the combs within the frame inter-spaces (see photo on right, blue arrows).
Analyses of Porapak devices. Porapak devices were retrieved from the colonies, wrapped individually with Al foil, placed in plastic Ziplock bags and taken to SFU Burnaby for analysis.
Devices were stored at -70 o C until extraction. Devices were extracted with HPLC-grade hexane:EtOAc 4:1 with 1,4-dimethoxybenzene (Sigma), 20 ng/µL (= 2 mg/100 mL) as internal standard. Each Porapak-containing pipette was placed above a pipette fitted with a glass wool plug and loaded with silica gel (~ 2 cm high) and a layer of Na2SO4 (as drying agent). Solvent (5 mL) was added to the Porapak pipette and allowed to run through both pipettes prior to collection. The silica gel was used to remove any polar compounds deposited by the bees and the drying agent removed any moisture captured in the Porapak device. The volume of the solvent collected was determined.
Samples were run on a Perkin Elmer Clarus 690 GC interfaced with a Clarus SQ8T MS, equipped with a Velocity-5 30 m column (0.25 mm i. d., 0.2 µm film thickness). The GC was programmed 80 o C (5 min), 10 o /min to 250 o C (10 min). Samples were injected using an automatic injector (1 µL), operated in splitless mode. Mass spectra were collected in EI+ mode from 50-350 amu, from 5 min to 25 min. The scan time was 0.35 s and the inter scan delay was 0.05 s.
Calibration was done with a pure standard of 3c{3,6} in the extraction solvent with internal standard.
Analyses of alcohol from washes. Two mL of alcohol from the wash was taken and diluted with 2 mL of distilled water. The mixture was extracted with 2×2 mL of solvent (see above hex:EtoAc 4:1 with 20 ng/µL internal standard). Combined organic phases were dried over Na2SO4 and the volume of dry extract was determined. Samples were run by GC-MS as described above for the Porapak devices.
Analyses of wax samples. To the wax collected from the hives (~ 1 g) was added 2 mL of isopropanol. The samples were left to extract overnight at 4 o C. Next, the isopropanol layer was collected, diluted with 2 mL of water and extracted as described above for the washes.
Analyses of honey samples. To the honey (with wax cells) collected from the hives was added 2 mL of isopropanol. Samples were left to extract overnight at 4 o C. The remaining steps were as described above for the other samples.

Supplemental results
Structure-activity study (2018) Figure S3. Results from a limited structure-activity survey of dialkoxybenzenes against varroa mites, done in 2018 (see text, under "Results Structure-activity relationship). A. Host choice after 3 hours of treatment: nurse (solid), forager (hatched), no choice (stippled). B. Host choice after 5 hours of treatment. There were no significant differences in host choice, except for 3c{2,6} in which mites not having made a choice were significantly higher than in all others (Kruskal-Wallis, p < 0.05). C. Number of mites lost (stippled), dead (hatched) or paralyzed (solid gray) after 3 hours of treatment. † Compound 3c{3,6} differed marginally from the blank with regard to paralysis and death+paralysis (Kruskal-Wallis, p < 0.1) D. Number of mites lost (stippled), dead (hatched) or paralyzed (solid gray) after 5 hours of treatment. * Compound 3c{3,6} differed significantly from the blank with regard to paralysis and death+paralysis (Kruskal-Wallis, p < 0.05).

Structure-activity study 2020
Please see the main text for the experiments done. Acaricidal activity at 5 h was analyzed by quantitative structure-activity relationship (QSAR) in Molecular Operating Environment (Chemical Computing). The following descriptors were computed and validated: dipole moment, total energy, electrostatic energy, angular energy, non-bonding energy, van der Waals energy, highest occupied molecular orbital, lowest unoccupied molecular orbital, total accessible surface * * area (ASA), positive ASA, negative ASA, hydrophobic ASA, polar ASA and van der Waals surface area. Figure S4. Results from quantitative structure-activity relationship (QSAR) analysis. A. Correlation of Log Pow (octanol water partition coefficient) vs. activity (percentage of net dead + paralyzed mites at 5 h). B. Correlation of the hydrophobic accessible molecular surface area vs. activity. C. Correlation of the total accessible molecular surface area vs. activity. D. Plot of the dipole vs. activity (there was no correlation).

Volatility assay
For the evaporation experiment, 2.3 µL of the compound stock solution (0.5 µmol/µL) was dispensed onto the parafilm square (15 mm x 15 mm). The parafilm square was already placed at the bottom of the 4-dram vial before addition of compound solution. The vial was sealed with a plug-type rubber stopper. The vial setup was placed in the incubator (37 o C) for 3 hours. For the gas chromatography-mass spectrometry (GC-MS) run, 1 mL of the headspace was collected and injected. The GC-MS instrument was a Perkin Elmer Claurs 690 GC interfaced with a Perkin Elmer Clarus SQ8T MS, and equipped with a Velocity-5 30 m column (0.25 mm i. d., 0.2 µm film thickness). Please see above for MS parameters and calibration. Samples were injected manually through a 1 mL gastight syringe.

Testing of 3c{3,6}, 3c{6,6} and the 1:1 blend on Varroa mites by themselves
The assay was set up as described in the methods, in glass dishes with a 2×2 cm 2 piece of Parafilm in the lid. The Parafilm received the treatment in hexane (10 L). Blanks received only hexane (10 L). One mite was placed in the center of the dish, and the setup was incubated at 32 o C, as described in the main article. Four doses were tested: 200 g, 500 g, 1.0 mg and 10 mg. Five mites were placed in each dish. Compounds 3c{3,6} and 3c{6,6} were tested by themselves and in a 1:1 blend. The results (Fig. S6) show that compound 3c{3,6} is more active than 3c{6,6} and that blending results in dilution of the more active compound. Figure S6. Effect of blending compounds 3c{3,6} and 3c{6,6} on the net number of mites paralyzed or dead. A. Experiment performed in 2018, with 500 g of 3c{3,6}, 3c{6,6} and the 1:1 blend (500 g each compound), as well as 1 mg of each pure compound. Bars represent the total of paralyzed + dead mites, after subtraction of the number obtained in the paired blank, averages ± S. E. of 4 replicates.  There is no significant difference in kinetics of AmAChE at any concentration or DEET. c) and d) Kinetics of AmAChE in the presence of 3c{3,6}; the black line represents 0 mM, darkest grey (square points) 0.5 mM, next darkest grey (diamond points) 1 mM, and lightest grey (triangle points) 3 mM of 3c{3,6}. e) and f) AmAChE in the presence of 3c{6,6}; no significant difference between treatments. g) and h) AmAChE in the presence of 3b{2,2}. The black line represents the average of 0 mM, 0.5 mM, and 1 mM. The grey represents 3 mM which had a significant increase in Vmax.