Cuticular pheromones stimulate hygienic behavior in the honey bee (Apis mellifera)

The health of western honey bee (Apis mellifera) colonies is challenged by the parasitic mite Varroa destructor and the numerous harmful pathogens it vectors. Selective breeding for the naturally occurring social immune trait “hygienic behavior” has emerged as one sustainable approach to reducing the mites’ impact on honey bees. To expand our understanding of hygienic triggers and improve hygienic selection tools, we tested the hypothesis that the cuticular compounds (Z)-10-tritriacontene and (Z)-6-pentadecene, previously associated with unhealthy honey bee brood and/or brood targeted for hygiene, are triggers of honey bee hygienic behavior independent of brood health. In support of our hypothesis, application of synthetic (Z)-10-tritriacontene and (Z)-6-pentadecene onto brood and brood cell caps significantly increased hygienic behavior compared to application of similarly structured hydrocarbon controls (Z)-16-dotriacontene and (Z)-7-pentadecene. Furthermore, we demonstrate a significant positive correlation between colony-level hygienic responses to (Z)-10-tritriacontene and the traditional freeze-killed brood assay for selection of hygienic honey bee stocks. These results confirm biological activity of (Z)-6-pentadecene and reveal (Z)-10-tritriacontene as a novel hygiene trigger. They also support development of improved tools for honey bee colony monitoring and hygienic selection, and thus may accelerate development of honey bee stocks with greater resistance to Varroa and associated pathogens.


Figure 1.
Hygienic responses to pupae treated with 1 μL of 1% Z10-C 33 in hexane, or 1 μL hexane, or left as untreated controls, in three colonies from different breeding backgrounds. Letters indicate significant differences in cell status between treatments for each colony. Compared to either control (hexane or untreated), significantly more cells treated with Z10-C 33 were uncapped and the pupae removed, at both 4 and 24 h post treatment, in all three colonies tested. Hexane treatment also elicited significantly greater hygienic responses than untreated controls at both time points and in all three colonies.

Figure 2.
Developmental status of in vitro reared pupae from three colonies with different breeding backgrounds treated with 1 μL of 1% Z10-C 33 in hexane, or 1 μL hexane, or left untreated as controls. Letters indicate significant differences in developmental status between treatments for each colony. Treatment had a significant effect on the number of deformed bees at the time of expected adult emergence in VSH, HYG, and UNS colonies. Compared to either control, significantly more pupae treated with Z10-C 33 exhibited deformities at the time of expected adult emergence in all three colonies tested. (2020) 10:7132 | https://doi.org/10.1038/s41598-020-64144-8 www.nature.com/scientificreports www.nature.com/scientificreports/

Discussion
Two major types of chemical cues and signals can guide honey bee workers to perform hygienic behavior. First, compounds not naturally found on honey bee brood may be perceived as "foreign" cues by workers, and could elicit some form of uncapping and/or removal behavior when detected in brood cells or on brood cell caps. This form of chemosensory processing is similar to the recognition of foreign invaders in social insect colonies 41,54-56 . Hygienic behavior may also be stimulated by intraspecific signals (pheromones) through changes  www.nature.com/scientificreports www.nature.com/scientificreports/ in amounts and ratios of naturally occurring native chemicals. This form of chemosensory processing is similar to the recognition of nestmates vs. conspecific non-nestmates in social insect colonies 54,[57][58][59] . Notably, cuticular hydrocarbons play important functions in both intra-and interspecific recognition in social insects [60][61][62] . Our and previous results indicate that the hydrocarbons Z10-C 33 and Z6-C 15 , when applied to brood or capped brood cells, induce hygienic behavior. In contrast, the structurally similar hydrocarbons Z16-C 32 and Z7-C 15 , which have not been associated with honey bee brood, elicited significantly less hygienic behavior. Thus, the chemical trigger for hygienic behavior is more similar to intraspecific communication than to a general recognition of a foreign stimulus. Furthermore, we show that hygienic responses to Z10-C 33 and Z6-C 15 may be useful as indicators of hygienic behavior at the colony level. Though previous studies have involved application of synthetic brood chemicals directly to pupae 49 and extracts of brood signals to brood cell caps 50 , to our knowledge, this is the first study to apply synthetic versions of honey bee brood compounds to brood cell caps to induce hygienic behavior. Therefore, results from this study not only confirm the biological activity of two cuticular hydrocarbons in triggering hygienic behavior towards otherwise healthy brood, but may also be useful in the development of improved tools for hygienic selection of pest-and disease-resistant honey bees, although further research related to quantitative and interactive effects of chemicals on hygienic behavior are needed.
Both Z10-C 33 and C 15 (isomers unidentified) occur naturally on honey bee cuticles, are found in higher quantities on Varroa-infested brood, and have been linked to hygienic behavior 47,49,50 . The experimental results support our hypothesis that Z10-C 33 and Z6-C 15 are triggers of honey bee hygienic behavior independent of brood health. Direct treatment of pupae with Z10-C 33 , and to a lesser extent hexane, elicited hygienic behavior in VSH, HYG, and UNS colonies. While these results are similar to those previously reported for Z6-C 15 -treated brood 49 , in vitro rearing indicated that increased hygienic behavior might have been partially due to a detrimental effect of direct exposure of brood to Z10-C 33 .
Hygienic behavior has been likened to programmed cell death 63 or "social apoptosis" 64 , with the idea that removal of unhealthy individuals may improve overall colony health. Although the effect of Z10-C 33 on brood susceptibility at quantities naturally produced by brood was not tested here, the considerable detrimental effect  www.nature.com/scientificreports www.nature.com/scientificreports/ of Z10-C 33 on pupal development combined with the lack of a hexane effect suggests that brood susceptibility to natural hygienic signals could play a role in hygiene at the colony level. It is unclear whether the hygienic signal could represent a coopted apoptosis mechanism 64 but this idea is consistent with recent findings that brood signaling differs by honey bee stock 50 , and that high brood susceptibility at the individual level may confer resistance to Varroa at the colony level 64 . Indeed, further studies regarding relationships among honey bee stock, quantitative brood signaling, and brood susceptibility are needed to clarify whether developmental interference caused by naturally produced brood signals plays a role in the apoptotic induction of hygienic behavior.
Because we were primarily interested in the function of applicable chemical signals (independent of induced mortality and/or developmental abnormalities), we shifted to treating brood cell caps rather than pupae directly. We found clear evidence that the hydrocarbons Z10-C 33 and Z6-C 15 , previously associated with Varroa and DWV-stressed brood, when applied to brood cell caps, elicited hygienic behavior. This effect was specific to these particular compounds because the structurally similar chemicals Z16-C 32 and Z7-C 15 , which have not previously been associated with honey bee brood, did not have a comparable effect. Because development of brood under Z10-C 33 and Z6-C 15 treated caps was not different from that of brood under hexane treated and untreated caps, we conclude that the hygienic behavior observed was not a result of brood abnormality related to cap treatment. Differential responses to Z10-C 33 and Z6-C 15 compared with structurally similar hydrocarbon controls may be related to two chemosensory mechanisms. First, control hydrocarbons may be detected but ignored as irrelevant. Some, albeit low, responses to high concentrations of control hydrocarbons support this idea. Second, honey bees may possess receptors specifically tuned to Z10-C 33 and Z6-C 15 , whereas they may be functionally anosmic to the structurally related but "non-natural" Z16-C 32 and Z7-C 15 . Nazzi et al. 49 tested honey bee responses to brood treated with Z6-, Z7-, and Z8-C 15 isomers, and found Z6-C 15 to be the most effective at triggering honey bee hygienic behavior. Additional structure-activity studies that expand analyses of bioactive natural hydrocarbons to multiple colonies could improve the conclusiveness of these findings, and provide critical data to discriminate between these two hypotheses.
Finally, we identified a significant positive correlation between hygienic responses to Z10-C 33 and to FKB across 10 honey bee colonies. This suggests that hygienic response to Z10-C 33 may be related to hygienic response to the similarly non-volatile necromone oleic acid, previously associated with FKB 53 . While our findings suggest that chemical hygiene triggers may be useful in measuring colony hygiene level, colony response 24 h after treatment with Z10-C 33 was relatively low and had a smaller range than response to FKB. Accordingly, and because natural signals likely involve mixtures rather than isolated compounds 50 , further studies involving shorter assay times that prevent misclassification of recapped cells, higher chemical concentrations, additional compounds, and mixtures of chemical stimuli should be conducted to optimize development of an improved assay for measuring hygienic behavior specific to Varroa and brood diseases.
Together, our findings provide support for the hypothesis that Z10-C 33 and Z6-C 15 are triggers of honey bee hygienic behavior independent of brood health. From a practical viewpoint, our results showed that synthetically produced compounds can be applied to capped brood cells to elicit honey bee hygienic behavior. This approach could be developed as a tool for evaluation of honey bee hygiene at the colony level. Such an assay may be useful for the improvement of selective breeding, because colony responses to actual brood stress signals may rely on different mechanisms or olfactory sensitivities than selection based on brood killed by freezing in the FKB assay [65][66][67] . Thus, semiochemical assays may be better suited to distinguish colonies with enhanced disease-and pest-resistance, facilitating honey bee management decisions such as which colonies may benefit most from chemical treatment to manage Varroa, or which breeders provide the most hygienic queens. A semiochemical hygiene assay may also be more rapid, and more beekeeper-friendly than current selection methods which require killing, infestation, and/or meticulous inspection of brood. Consequently, testing and breeding for hygienic behavior may become more widespread. Improved honey bee breeding and management decisions have potential to sustainably improve honey bee health, reducing the risks associated with many current Varroa management practices, such as evolution of pest and pathogen resistance, and contamination of commercial honey and beeswax. Thus, although further development is needed, our findings suggest that the exploitation of intrinsic signals associated with honey bee health may provide new tools and strategies of benefit to queen breeders, commercial beekeepers, farmers, and consumers.

Methods
A recent study linked Z10-C 33 to Varroa-infested, DWV-infected, and hygiene-targeted honey bee brood 50 . Given the structural similarity of Z10-C 33 to Z6-C 15 , previously linked to hygienic removal 49 , we decided to investigate the effectiveness of Z10-C 33 and Z6-C 15 in triggering hygienic behavior. For this purpose, Varroa-Sensitive Hygienic (VSH), Minnesota Hygienic (HYG), and unselected control (UNS) honey bee colonies were established at the University of North Carolina at Greensboro apiary in the Spring of 2017. The VSH queen was sourced from the USDA-ARS Honey Bee Breeding Laboratory in Baton Rouge, where queens are selected based on suppression of mite reproduction. The HYG queen was sourced from the well-established breeder Jeff Hull (Minnesota) and was selected based on removal of>95% freeze-killed brood. The UNS queen was sourced from Triad Bee Supply which obtains their queens from Gardner Apiaries in Baxley, Georgia. For each experimental hydrocarbon, a control hydrocarbon of similar size and structure, but not known to be a component of the honey bee's cuticular hydrocarbons, was also tested. Hydrocarbons (Z10-C 33 , Z16-C 32 , Z6-C 15, and Z7-C 15 ) were synthesized by Z-selective Wittig reactions between the appropriate aldehydes and phosphonium salts, or by Z-selective olefin metathesis reactions. Crude products were purified in two steps, by flash vacuum chromatography on silica gel, eluting with hexanes, followed by recrystallization from acetone at ~4 °C for longer chain compounds, or ~−20 °C for shorter chain compounds. Synthesis is described in further detail, below. Dilutions in hexane of 0.1%, 0.3%, and 1.0% (wt/vol) were prepared for each hydrocarbon. The lowest dilution is equivalent to that Scientific RepoRtS | (2020) 10:7132 | https://doi.org/10.1038/s41598-020-64144-8 www.nature.com/scientificreports www.nature.com/scientificreports/ previously reported 49 , and higher dilutions were chosen to approximate dose effects on a logarithmic scale. All sample collections and analyses were conducted at the University of North Carolina at Greensboro. 15 . A solution of 1-decyne (5.52 g, 40 mmol) and ~50 mg triphenylmethane indicator in dry THF under Ar was cooled to ~−15 °C in an ice/acetone bath, and butyllithium (2.6 M in hexanes) was added dropwise until the solution turned pink. An additional 15.4 ml of butyllithium solution (40 mmol) was then added over 30 min, and the resulting solution was warmed to room temperature and stirred 1 h. Powdered NaI (0.6 g, 4 mmol) was then added, followed by dropwise addition of bromopentane (3.93 g, 26 mmol). The mixture was heated to reflux and stirred 22 h, then cooled and quenched with 1 M aqueous NH 4 Cl solution, and extracted with hexane. The hexane layer was washed with saturated NaHCO 3 and brine, then dried and concentrated. The residue was purified by Kugelrohr distillation, taking a forerun of the excess 1-decyne (oven temp <40 °C, 0.05 mm Hg), then changing the collection bulb and collecting the desired product (2.8 g, bp~60 °C, 0.05 mm Hg).

Synthesis of alkenes tested in bioassays. Synthesis of Z6-C
The distilled product was flushed through a plug of silica gel with hexane and into a 200 ml round-bottomed flask with a magnetic stir bar. Lindlar catalyst (150 mg) and quinolone (1.5 ml) were added, and the flask was sealed and flushed with nitrogen, then hydrogen. With the sealed flask attached to a gas burette filled with hydrogen, stirring was then started, resulting in uptake of ~310 ml of hydrogen, at which point uptake virtually ceased. The flask was flushed with nitrogen, and the mixture was filtered through a plug of celite, rinsing well with hexane. The resulting solution was washed twice with 1 M HCl, then dried and concentrated. The residue was flushed through a pad of silica gel with hexane, then Kugelrohr distilled (2.82 g, bp~60 °C, 0.05 mm Hg). Because the resulting product was contaminated with about 4% of the alkyne starting material, a portion (1.2 g) was repurified by vacuum flash chromatography on silica gel in a 60 ml Buchner funnel. The silica was prewetted with hexane, then the impure alkene was loaded as a hexane solution, and the column was eluted with 5 ×30 ml hexane. Synthesis of Z10-C 33 . Triflic anhydride (2.2. ml, 12 mmol) was added dropwise to a slurry of docosanol (3.26 g, 10 mmol), pyridine (0.8 ml, 10 mmol), and ~50 mg dimethylaminopyridine catalyst in 50 ml methylene chloride, cooling as necessary to keep the reaction temperature <25 °C. When the addition was complete the mixture was stirred 1 h at room temperature, producing a pale brown, slightly cloudy solution. The solution was diluted with 100 ml hexane, and filtered through a pad of silica gel, rinsing the filter pad with 2:1 hexane in methylene chloride. The resulting clear solution was concentrated to a white solid which was taken up in ether and used immediately.
Butyllithium (2.24 M in hexanes) was added to an ice-bath cooled solution of 1-undecyne (2.28 g, 15 mmol) and ~50 mg triphenylmethane indicator in 50 ml dry THF under Ar until a pink color persisted (~7 ml, 15.7 mmol). The solution was stirred for 1 h, then the ether solution of the triflate was added dropwise at 0 °C, and the mixture was warmed to room temperature and stirred overnight. The reaction was then quenched with saturated aqueous NH 4 Cl, and extracted with hexane. The hexane layer was washed with brine, dried, and concentrated. The residue was flushed through a pad of silica gel with hexane, then Kugelrohr distilled (oven temp ~50 °C, 0.2 mm Hg) to remove excess 1-undecyne. The solid residue was then recrystallized from 50 ml acetone, warming to solubilize the product, then cooling to room temperature, yielding the alkyne product as a single peak (1.76 g), with additional impure alkyne in the filtrate.
The alkyne (1.7 g) was taken up in 30 ml hexane, and quinoline (0.75 ml) and Lindlar catalyst (75 mg) were added. The reaction flask was sealed and flushed sequentially with nitrogen, then hydrogen, then connected to a gas burette filled with hydrogen. The mixture was stirred until hydrogen uptake ceased. After flushing with nitrogen, the mixture was filtered through celite, most of the quinoline was removed under high vacuum, and the residue was recrystallized from 30 ml of hot acetone, after cooling to 4 °C. The resulting white solid was still contaminated with quinoline, and so a hexane solution was flushed through a plug of silica gel with hexane. After concentration, the residue was recrystallized again from acetone, yielding the alkene as a white solid (1. Hygienic response to treatment of pupae with Z10-C 33 . This experiment was conducted using one VSH, one HYG, and one UNS colony. To obtain a same-age cohort of honey bee brood, the locations of uncapped brood cells containing 5 th instar larvae were marked using a permanent marker on transparent plastic sheets secured above experimental cells with thumbtacks. Combs containing experimental cells were placed back into the colony and recollected within 8 h. Cells capped within that time were marked for experimental use, and frames were returned to their respective colonies. On day 6 post-capping, experimental cells were opened by cutting and lifting one side of the cell cap with a razor blade. The pupa underneath received either no treatment, or treatment with either 1 μL of hexane or with 1 μL of 1.0% Z10-C 33 in hexane. Cells were then resealed by gently pressing the cap against the cell wall with the side of a razor blade, and frames were returned to their respective colonies. Uncapping and removal of brood in experimental cells were recorded at 4 and 24 h after the frame was reintroduced. Sample sizes were 40 cells per treatment for UNS and HYG colonies, and 50 cells per treatment for the VSH colony. Effects of treatment of pupae with Z10-C 33 on development. This experiment was conducted using the same three VSH, HYG, and UNS colonies. As described above, capped cells containing brood 6-d post-capping were carefully opened, and the brood inside received either no treatment, or treatment with either 1 μL of hexane, or 1 μL of 1.0% Z10-C 33 in hexane. Pupae were then removed from the brood comb using flexible-tipped forceps, and gently placed onto fan-folded filter paper in Petri dishes. Petri dishes were placed in an incubator maintained at 34 °C and 50% RH. Brood was examined for injury (dark pigmentation) after 48 h in the incubator, and any injured brood were discarded. On the day after expected emergence, brood were examined for normal development, defined by typical adult pigmentation and shape with proper wing development. Any deviation from this was considered "deformed" and used to calculate developmental success. Brood sample sizes were 35, 29, and 27 individuals per treatment for VSH, HYG, and UNS colonies, respectively.
Hygienic response to wax cap treatment. We tested effects on hygienic behavior of application of test compounds to wax caps of brood cells. Hygienic assays were conducted by applying hexane, 0.1%, 0.3%, or 1.0% dilutions of Z10-C 33 , Z6-C 15 , or appropriate controls (Z16-C 32 and Z7-C 15 , respectively) in hexane to capped brood cells in a VSH colony. For each assay, 2 mL of solution were applied to a circular area of capped honey bee brood. Similar to the established freeze-killed brood assay 68 , the treated area was isolated using a piece of PVC pipe (7.5 cm inner diameter, approximately 8 cm long). Chemicals were applied using an H-100D Single Action airbrush and compressor (Paasche, Kenosha, WI), modified with glass bottles fitted with glass tubing for this application. For each assay, 2 mL of the solution were added to the bottle immediately before application to wax caps. Thus, 0.1%, 0.3%, and 1.0% solutions deposited approximately 45, 136, and 453 µg hydrocarbon/cm 2 of capped honey bee brood, respectively. Capped cells were counted directly after treatment, and frames were returned to the colony. After 24 h, frames were recollected, and capped cells in the treated region were recounted. For comparisons of Z10-C 33 , Z16-C 32 , and hexane, sample sizes were 8, 6, and 3 replicates, respectively. For comparisons of Z6-C 15 , Z7-C 15 , and hexane, sample sizes were 12, 6, and 5 replicates, respectively. Because brood availability was limited, priority was given to replication of Z10-C 33 and Z6-C 15 assays, followed by assays of the structural controls Z16-C 32 and Z7-C 15 . Assay scores were determined by dividing the total number of uncapped and removed cells after 24 h by the total number of capped cells in the circular assay area at the beginning of the assay.
Effects of wax cap treatment on development. Capped VSH brood cells in a 7.5 cm diameter circular area were left untreated (control) or treated with 2 mL of hexane, 0.1%, 0.3%, or 1.0% Z10-C 33 or Z6-C 15 solutions in hexane with the airbrush, as described above. After 30 min, white-eye pupae (aged 5-6 days post-capping) were removed from the brood comb using flexible-tipped forceps, and gently placed onto filter paper in Petri dishes held in an incubator (34 °C, 50% RH). Brood was examined for injury (dark pigmentation) after 48 h in the incubator, and any injured brood were discarded. On the day after expected emergence, brood were examined for normal development, as described above. Any deviation from this was considered "deformed" and the proportion of successful development calculated. Brood sample size was 40 pupae per treatment.
Comparison of freeze-killed brood (FKB) and Z10-C 33 -treatment assays. Hygienic responses to Z10-C 33 wax cap treatment and FKB assays were compared in ten colonies, including 2 HYG, 2 VSH, and 6 UNS. For the chemical assay, 1.0% Z10-C 33 was applied to wax caps as described above, and any uncapping or removal of a cell after 24 h was counted as a cell targeted by hygienic behavior. The percentage of such cells among all initially capped cells was calculated for the assay score, as above. For FKB assays, a 7.5 cm diameter PVC tube was placed on a section of capped pupae aged 3-10 d post-capping, and brood in the assay area were frozen using liquid nitrogen. Frames were returned to their colony of origin, and after 24 h, all cells within the test area that still contained any pupae were counted and recorded, according to standard practice 68 . Assay scores were determined by dividing the total number of cells containing any pupae after 24 h by the total number of capped cells in the circular assay area at the beginning of the assay and subtracting this number from 1.

Statistical analyses.
Pearson's Chi-square analysis with Bonferroni correction was used to test effects of pupal treatment on hygienic responses in VSH, HYG, and UNS colonies. All analyses were comparisons of manipulated (uncapped or removed) versus non-manipulated (capped) cells. Pearson's Chi-square analysis with Bonferroni correction was also used to test effects of pupal treatment on the numbers of successfully and unsuccessfully developing VSH, HYG, and UNS brood. Two-way ANOVAs with Bonferroni-corrected post-hoc Scientific RepoRtS | (2020) 10:7132 | https://doi.org/10.1038/s41598-020-64144-8 www.nature.com/scientificreports www.nature.com/scientificreports/ comparisons were used to test effects of chemical type and chemical concentration on hygienic response to wax cap treatments. Pearson's Chi-square analysis with Bonferroni correction was used to test effects of wax cap treatment on the development of VSH, HYG, and UNS brood. A Pearson's correlation coefficient was calculated to test for a positive correlation between hygienic responses to Z10-C 33 and FKB assays across colonies, and a one-tailed p-value is reported. Chi-square analyses were calculated based on raw data, while assay scores and related analyses were based on percentages. All statistics were performed using IBM SPSS Statistics, Version 25.

Data availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author upon request.