Determination of volatile compounds that elicit removal of diseased brood by hygienic honey bees.

Breeding
The breeding program for colonies that display hygienic behavior began in 1993 at the University of Minnesota. Rapid-hygienic and slow-hygienic lines of bees (sensu Wilson-Rich et al., 2009) were selected from colonies derived from the Italian subspecies Apis mellifera ligustica. Colonies were initially selected based on their ability to survive winter in Minnesota, their ability to increase in population rapidly in spring, to produce large amounts of honey, and to be gentle (not sting) during normal management. If the colonies displayed the above set of traits they were then tested for hygienic behavior, using a freeze-killed brood assay (Spivak and Reuter, 1998), which is an indirect measure of a colony’s hygienic behavior. The removal of freeze-killed brood is correlated with the removal rates of diseased brood; colonies that uncap and remove approximately 160 freeze-killed pupae within 24 hr will tend to remove diseased brood rapidly. Colonies that exhibited the most rapid removal rates were chosen to raise queens for the next generation. Daughter queens were then instrumentally inseminated with sperm from drones from other rapid-hygienic colonies. The criterion for choosing rapid-hygienic colonies for breeding is an arbitrary cutoff established at 95% removal of freeze killed brood within 24 hr.

Volatile Collection
Volatiles were collected from bees for two purposes. The first purpose identified the changes in volatiles between healthy and diseased brood. Three types of larvae were used, including:
a) healthy fifth instar larvae,
b) fifth instar larvae in the early stage of chalkbrood infection, and
c) larvae fully overcome with fungal mycelia (mummies).
Each type of larvae was removed from the comb with sterile forceps and placed in quick-fit glass aeration chambers. The disease was confirmed by fungal growth through the larvae after the volatile collections were completed. Collections were determined to be from healthy larvae if no larvae showed clinical symptoms of disease at the conclusion of the volatile collection period. Each collection consisted of either
a) 25 healthy fifth instar larvae,
b) 25 early-stage infection larvae, or
c) 75 mummies.
For use with the electroantennography, the second collection of volatiles employed a large quantity of healthy larvae or diseased larvae placed in a larger aeration chamber.

For each experiment, the volatile collection apparatus was constructed as follows: charcoal-filtered and humidified air at 0.5 L/min was passed over the larvae and then through pre-conditioned Super-Q adsorbent traps for 24 hr at room temperature. Each trap was eluted with 150 l methylene chloride forced through the absorbent with filtered nitrogen gas. The elutions stored at -70°C until used.

Chemical Analysis
To identify the disease-related volatiles, Samples were analyzed by gas chromatography and the chromatograms were compared from the healthy, early stage infected and fully diseased brood. Volatile compounds were confirmed by comparison of their chromatographic retention times and mass spectra with commercially available standards analyzed on the same instrument. To confirm detection by bees of these identified compounds, gas chromatography linked electroantennal detection (GC-EAD) was used for the diseased-brood samples, the healthy larvae samples and the standards. The effluent was split to allow simultaneous recording of the flame ionization detector (FID) and the antennal response (EAD). Electroantennogram recordings were obtained from excised antennae of bees that were age 15-20 days old as this is the age of bee that would be performing hygienic behavior (Arathi et al., 2000). Three source colonies used for the GC-EAD tests were from the University of Minnesota breeding program and were deemed greater than 95% hygienic based on freeze killed brood tests.

Field Bioassays:
I. Topical Applications. The topical application assay tested the hypothesis that hygienic behavior of honey bees can be elicited in field colonies by experimental application of appropriate olfactory stimuli to healthy larvae. The assays were conducted in August, 2007. Twelve colonies were chosen for the bioassay based on the freeze-kill removal method. A comb containing larvae was removed from each of the twelve colonies and the position of 25 fifth instar larvae on the comb was marked on a transparency overlaying the comb. A Picospritzer II (General Valve Corporation) was used to dispense 0.5 l of each treatment compound, onto healthy larvae. Six treatments were applied to separate groups of 25 healthy larvae. The treatment compounds were benzyl alcohol, 2-phenylethanol (phenethyl acetate) and deionized water. The sixth treatment was 0.5 l of a mixture of equal parts benzyl alcohol, 2-phenylethanol, and phenethyl acetate. The comb was then returned to the colony from which it was taken. At 4 and 24 hr the transparency was replaced on the frame and the number of larvae that the adult bees had removed was recorded.

II. Paraffin larval dummies. The assay utilizing paraffin larval dummies tested the hypothesis that hygienic behavior is based on a threshold response in which bees with the highest olfactory sensitivity to diseased brood initiate the hygienic response more quickly as compared to bees with a lower olfactory sensitivity. Paraffin larval dummies were made by first melting paraffin wax at 60-65oC. A 10-2 dilution of methyl linolenate was made with the wax. Methyl linolenate is one of several compounds present in brood pheromone of fifth instar larvae. When this compound is added to a paraffin wax dummy and placed in a larval cell, adult bees cap the cell containing the dummy (Le Conte, 1990). The addition of chalkbrood volatile compounds to wax dummies containing brood pheromone would result in a decrease in capping of these dummies if the adult bees detect the disease compounds. The response threshold model can be tested by measuring the differential abilities of adult bees to detect these chalkbrood compounds at variable concentrations.

From this mixture of wax and methyl linolenate a total of eight treatment compound /concentrations were made: phenethyl acetate at 10-2 ml/ml and 10-9 ml/ml, 2-phenylethanol at 10-2 ml/ml and 10-9 ml/ml, benzyl alcohol at 10-2 ml/ml and 10-9 ml/ml and a mixture of the previous three compounds at 10-2 ml/ml and 10-9 ml/ml. Additional treatments were made of methyl linolenate alone and paraffin wax without the addition of methyl linolenate or disease compounds. The liquid wax containing each treatment was poured into plastic drinking straws and allowed to cool and harden. The straw was then cut away from the wax and the wax cut into 25 mm sections to create larval dummies.

The assays were conducted in August 2008. Six colonies that displayed varying degrees of hygienic behavior, as determined through freeze-killed brood assays, were chosen for this bioassay. A comb containing fifth instar larvae was removed from the colony. Combs with fifth instar larvae were chosen so that the larval dummies would be in context with actual fifth instars that were about to be capped with wax by adult bees. Twenty larval dummies of each treatment were placed into empty cells and their location was marked on a transparency. No more than two treatments were tested on a single comb at a time. The combs were then returned to the colony and the number of dummies the bees capped with wax was recorded after 24 hr.

Statistical Analysis
Binomial regressions were performed to compare the proportion of healthy topically treated larvae removed or paraffin dummies that were left uncapped in response to each treatment with the degree of hygienic behavior of the colony as a covariate. The delta method was used to calculate the student’s t value for contrasts between treatments with differences declared at  =0.05. Analyses were done with Arc (Cook and Weisberg, 1999).