, Volume 96, Issue 2, pp 309–314

Octanoic acid confers to royal jelly varroa-repellent properties

  • Francesco Nazzi
  • Renzo Bortolomeazzi
  • Giorgio Della Vedova
  • Fabio Del Piccolo
  • Desiderato Annoscia
  • Norberto Milani
Short Communication

DOI: 10.1007/s00114-008-0470-0

Cite this article as:
Nazzi, F., Bortolomeazzi, R., Della Vedova, G. et al. Naturwissenschaften (2009) 96: 309. doi:10.1007/s00114-008-0470-0


The mite Varroa destructor Anderson & Trueman is a parasite of the honeybee Apis mellifera L. and represents a major threat for apiculture in the Western world. Reproduction takes place only inside bee brood cells that are invaded just before sealing; drone cells are preferred over worker cells, whereas queen cells are not normally invaded. Lower incidence of mites in queen cells is at least partly due to the deterrent activity of royal jelly. In this study, the repellent properties of royal jelly were investigated using a lab bioassay. Chemical analysis showed that octanoic acid is a major volatile component of royal jelly; by contrast, the concentration is much lower in drone and worker larval food. Bioassays, carried out under lab conditions, demonstrated that octanoic acid is repellent to the mite. Field studies in bee colonies confirmed that the compound may interfere with the process of cell invasion by the mite.


Varroa destructor Repellent Cell invasion Behaviour Royal jelly Octanoic acid 


Recent mysterious die-offs of honeybees have brought back to light the importance that domesticated bees have got for agriculture underlining the importance of effective honeybee protection (Stokstad 2007). The mite Varroa destructor Anderson & Trueman represents a major threat for apiculture in the Western world (Sammataro et al. 2000); due to rapid selection of acaricide-resistant strains, the control is a major challenge. This makes the search for alternative methods very urgent. A better understanding of the biology of V. destructor, particularly, the chemical communication between the parasite and its host, could help in the development of novel techniques for the control and could provide new selection criteria for breeding mite-tolerant bees.

Entering a brood cell containing a bee larva is a crucial step in the biological cycle of V. destructor, as reproduction can only take place inside sealed brood cells. However, brood cells from different sex and castes are subject to different invasion rates by the mite. In particular, drone cells are much more likely to be invaded than worker cells, whereas queen cells are not normally infested (Calderone et al. 2002). Signals involved in the process of cell invasion have been investigated; most of research carried out so far has focused on semiochemicals released by the bee larvae leading to the identification of several compounds. Unfortunately, none of them proved to be active under field conditions (Milani 2002). More recently, attention has been paid to stimuli from the larval food, and an attractant semiochemical was identified from drone food (Nazzi et al. 2004). The food which is provided to queen larvae throughout their pre-imaginal life is royal jelly. Its repellent properties, accounting for the lower invasion rates of queen cells, have been reported (Calderone et al. 2002). Compounds affecting the arresting response of the mite were isolated from this source (Drijfhout et al. 2005), but no volatile chemicals influencing the orientation of the invading mite have been identified yet.

Mites are carried to brood cells by infested nurse bees; when the bee is in front of the cell opening, the mite leaves the bee and enters the cell, walks on the bee larva and crawls between larva and cell wall until the cell bottom is reached (Boot et al. 1994). Thus, when the mite leaves the bee for invading the brood cell, it does so without having a prior contact either with the larva or the cell content. In conclusion, information to decide whether to stay on the bee or to invade the cell is likely to be acquired from the distance (Boot et al. 1994); this suggests that volatile chemicals are involved.

Volatile compounds from royal jelly have not been studied in detail. Some free acids including octanoic were identified (Boch et al. 1979), but to our knowledge, no data about their concentration in the larval food of queens, workers and drones or their biological activity on the varroa mite were available until now.

The objective of this study was to investigate the semiochemicals affecting the process of host finding by the reproducing mites. In particular, we aimed at identifying possible volatile repellent compounds from royal jelly diverting the mites from queen cells.

Materials and methods

Biological material

The worker bee larvae and the adult females of V. destructor used in all the experiments came from Apis mellifera colonies maintained in Udine (northeastern Italy). Previous studies indicated that the local bee population consists of hybrids between A.m. ligustica and A.m. carnica. The mites and bee larvae were obtained from brood cells capped 0–15 h before using standard methods (Nazzi et al. 2001).

Royal jelly was collected from colonies maintained in the experimental apiary. Larvae were transferred into queen cups when they were 1 day old. Royal jelly was collected 3 days later and kept at −20°C in sealed vials until used.

Extraction of royal jelly

Five microlitres of a water solution of heptanoic acid (1,000 μg/mL) as internal standard (I.S.) and 1 mL of 0.1 M phosphate buffer pH 2 were added to 50–100 mg of royal jelly in a test tube. After vortexing, the mixture was extracted with 5 mL of diethyl ether by vigorously shaking. The organic phase was collected into a test tube and dried over anhydrous magnesium sulphate for 30 min. The diethyl ether extract was then transferred into a 4-mL vial and reduced to dryness under a gentle stream of nitrogen, stopping as soon as the solvent was completely evaporated to avoid loss of volatile acids. The magnesium sulphate was washed once with 2 mL of diethyl ether. Then, 500 μL of ethyl acetate and 100 μL of N,O-bis(trimethylsilyl)trifluoroacetamide were added and the mixture silylated at 50°C for 1 h in a water bath under magnetic stirring. One microlitre was used for the gas chromatography–mass spectrometry (GC-MS) analysis.

GC-MS analysis

Extracts were analyzed by gas chromatography coupled to quadrupole mass spectrometry (Shimadzu, Q2010). The injection was in splitless mode (2 min) with helium as carrier gas at a flow rate of 0.9 mL/min. The fused silica column was a 30-m × 0.25-mm i.d., 0.25-μm film thickness SPB5 (Supelco). The initial column temperature was set at 50°C for 2 min, then programmed to 150°C at 5°C/min and then to 280°C at 10°C/min, maintaining the final temperature for 10 min. The injector, interface and ion source temperatures were at 280°C, 280°C and 200°C, respectively. The instrument operated in the electron impact mode (70 eV). To avoid saturation of the detector caused by high boiling point compounds, the mass acquisition was stopped at 25 min, leaving the GC column oven temperature program to go to completeness. The samples were analysed as trimethylsilyl (TMS) derivatives.

To determine the overall composition of the lipidic fraction of royal jelly, some samples were analysed also in split mode (split ratio 1:100) under the same chromatographic conditions extending the mass acquisition to the end of the chromatographic analysis.

Quantification of octanoic acid

For the quantitative analysis of the TMS derivatives of the samples, the signal of extracted ions at m/z 201 (M-CH3)+ and 187 (M-CH3)+ were used for TMS-octanoate and TMS-heptanoate, respectively. The relative response factor of octanoic acid with respect to I.S. was calculated daily by analysing a standard solution of the acids in triplicate. The linearity of the method was tested between 0.2 and 40 μg/mL. The repeatability (n = 8) was determined on two different samples of royal jelly, and the results, expressed as %RSD, were 4.3% and 5.3%. The recovery of the analytical method was evaluated at two spiking levels by preparing three samples of royal jelly and three samples of larval food fortified with 10.3 and 1.0 μg, respectively, of octanoic acid. The percentage recovery was of 97 ± 6 for the royal jelly and of 93 ± 1 in the case of larval food.

Seven samples of royal jelly were analysed with this method; each sample was analysed in triplicate.

The same method was applied for the determination of octanoic acid in three samples of worker food and three samples of drone food. These were collected from natural brood cells 0–12 h before capping.

Lab bioassay

A glass arena with four wells (7-mm diameter, 8-mm depth) equidistant (1 cm) from the centre was used. The treatment was applied to two opposite wells, whilst the others were used as controls. When extracts or pure compounds were assayed, control wells received 1 μl of the solvent alone. After evaporation of the solvent, one bee larva was placed into each well, then one adult female mite was placed in the centre of the arena. The position of the mite was noted every 5 min. The bioassays lasted 30 min. Twenty arenas were used at a time. Each stimulus was tested four times. Experiments were carried out at 35°C and 75% R.H.. The following stimuli were tested with the bioassay:
  1. (a)

    royal jelly: 10 mg;

  2. (b)

    royal jelly acetone extract: 10 mg equivalents (for extraction, 1 g of royal jelly was smeared with a spatula inside a conical test tube, and about 5 mL of acetone was added; after 1 h, the tube was centrifuged until the royal jelly residue settled and the supernatant was transferred to another tube for bioassay);

  3. (c)

    octanoic acid: 1 μg in diethyl ether (this dose was selected as, according to our determination, it roughly resembles the amount present in 10 mg of royal jelly);

  4. (d)

    heptanoic acid: 1 μg in diethyl ether (this was tested for comparison to assess the specificity of octanoic acid);

  5. (e)

    nonanoic acid: 1 μg in diethyl ether (see above);

  6. (f)

    octanoic acid at lower (1, 10 and 100 ng) concentrations; and

  7. (g)

    octanoic and nonanoic acid at a higher (10 μg) concentration; in this case, acetone was used as a solvent.


Field bioassay

Octanoic acid was tested in the hive as following.

One hundred nanograms or 1 μg of the compound dissolved in 1 μl of deionised water was applied to worker brood cells containing l5 larvae in an infested colony using a 10-μl Hamilton syringe. An equal number of cells was treated with 1 μl of water and used as a control. The cells were marked on an acetate sheet for subsequent control. After 12 h, the sealed cells were opened and inspected, and the number of infested cells was noted.

The experiment was replicated five times in two successive years. On the whole, about 600 brood cells were considered for each experimental group.

Statistical analysis

For each arena, the number of times the mite was observed in the treated and control wells, respectively, over the 30-min period was calculated for being used as a score for the statistical analysis; thus, for a given arena, the score of treated and control wells can vary between 0 and 6. For each replication, the scores of single arenas were summed; with 20 arenas, the scores can vary between 0 to 120. The outcome of an experiment consisting of several replications was represented with the average score per replication.

Repellency of different treatments was calculated as the ratio between mites observed on bee larvae in control wells and those observed in treated wells (repellency = average score control/average score treated).

For the statistical analysis, a matrix was constructed with as many rows as the mites used in the bioassay and two columns containing the scores for treated and control wells for each of the tested mites. The treatment and control scores in a given set of data were compared by a sampled randomisation test; the randomisation distribution was resampled 106 times with a computer program written for this purpose. A sampled randomisation test was chosen because the distribution of the variables to be compared is unknown.

The proportion of treated and control cells that were infested were compared using the Mantel–Haenszel method after testing the homogeneity in the odds ratios of the replicated 2 × 2 tables.


In the lab bioassay, royal jelly appeared to be repellent for V. destructor, and so was the acetone extract (Table 1).
Table 1

Response of V. destructor to different stimuli in a bioassay




P (sampled randomised test)

Royal jelly

10 mg



Royal jelly acetone extract

10 mg equiv



Heptanoic acid

1 μg



Octanoic acid

1 μg



Nonanoic acid

1 μg



Repellency is the ratio between mites observed on bee larvae in the control wells of an observation arena and those observed in wells treated with the stimulus. The figure is the average of four replicates; in each replicate, 20 mites were used

The GC-MS chromatograms of the silylated diethyl ether extract of a sample of royal jelly both in split (split ratio 1:100) and splitless mode are reported in Fig. 1a,b, respectively. The identities of the compounds are reported in Table 2.
Fig. 1

GC-MS chromatograms of the silylated diethyl ether extract of a sample of royal jelly

Table 2

Volatile compounds identified in a royal jelly sample as TMS derivatives

Peak number

Retention time






2-Methylbutanoic acida




3-Methylbutanoic acid




Pentanoic acida




Hexanoic acida





2-Hexenoic acida




Heptanoic acid





2-Hydroxy-4-methylpentanoic acida




Octanoic acida





2-Hydroxyhexanoic acida




Phenylacetic acid




2-Octenoic acid





Nonanoic acida









Unknown 1




Unknown 2




3-Hydroxyoctanoic acid









Butylated hydroxytoluene (BHT)

From solvent




7-Hydroxyoctanoic acid





8-Hydroxyoctanoic acid





3-Hydroxydecanoic acid





9-Hydroxydecanoic acid





9-Hydroxy-2-decenoic acid





10-Hydroxydecanoic acid





10-Hydroxy-2-decenoic acid





Decanedioic acid





11-Hydroxydodecanoic acid





Decenedioic acid





3,10-Dihydroxydecanoic acid





Hexadecanoic acid



Peak number refers to Fig. 1

aMarks the compounds whose identification was confirmed by injection of an authentic standard; other compounds were tentatively identified from the mass spectrum and literature data when available.

References: 1, Boch et al. 1979; 2, Drijfhout et al. 2005; 3, Lercker et al. 1981

In Fig. 1a, lower boiling compounds, which elute in the first section of the GC trace, are hardly detected, whilst in Fig. 1b, octanoic acid appears to be the most abundant volatile compound from royal jelly.

Concentration of octanoic acid in the analysed royal jelly samples varied from 113 ± 2 to 252 ± 8 μg/g, with most samples having less than 150 μg/g. By contrast, worker and drone food, which are attractive to the mite, contained only 3.2–7.6 and 2.1–7.3 μg of octanoic acid per gram of larval food, respectively. Conversely, other compounds that had already been identified from bee larval food (Nazzi et al. 2004) showed a similar concentration in royal jelly compared to drone food.

Octanoic acid, tested with the standard bioassay at a concentration similar to that found in royal jelly, appeared to be as repellent as royal jelly itself, whereas similar compounds that were tested for comparison failed to give a comparable response (Table 1). At lower doses, octanoic acid was not repellent (repellency of 1, 10 and 100 ng: 1.41, 1.11 and 0.85, respectively). Repellency increased at 10 μg (repellency, 19.67), although at this concentration, the otherwise inactive nonanoic acid appeared to be repellent as well (repellency, 5.42), suggesting that toxic effects may be involved when the dose is beyond biological levels.

In the hive, brood cells treated with 100 or 1,000 ng of octanoic acid were significantly less infested than control cells by 30% and 33%, respectively (infestation in control cells = 0.17%, infestation in cells treated with 100 ng of octanoic acid = 0.12%, infestation in cells treated with 1,000 ng of octanoic acid = 0.11%; P = 0.024 and P = 0.020, respectively), confirming the biological activity of octanoic acid and suggesting a possible practical use of the compound.


The repellency of royal jelly for V. destructor, which was first noted by Calderone et al. (2002), was confirmed here. The biological activity of the royal jelly acetone extract suggested that repellency is mediated by polar semiochemicals as previously noted by other authors (Drijfhout et al. 2005). Abundance of octanoic acid in the volatile fraction of royal jelly confirmed early studies by Boch et al. (1979) and suggested a possible biological role for the compound; in fact, drone and worker food, which are attractive to the mite, contained only negligible amounts of it. Successive biological assays demonstrated the repellency of this compound for V. destructor both under lab and field conditions, indicating that the compound can affect the process of brood cell invasion by the mite. Although the differences obtained in the field trial between infestation in control cells and those treated with octanoic acid seem to be small, such differences, over time, can lead to drastic differences in the size of mite populations.

Together with the evidence that V. destructor is attracted to brood cells for reproduction by a compound from larval food (Nazzi et al. 2001, 2004), the results reported here, about a fatty acid from royal jelly which contributes to queen cells repellency, indicate that signals from the host food are exploited by V. destructor throughout the host-finding process. Localisation of a suitable host for reproduction is of crucial importance for the mite, but the cues emitted by the host itself are probably not reliable enough for this task. In fact, most compounds, identified so far from bee larvae, are widespread inside the beehive (see Nazzi et al. 2001 for a discussion on this issue). For this reason, more specific compounds such as those found in the food provided by nurse bees to larvae may represent a suitable cue for the parasite looking for its host.

Foraging parasitoids often use stimuli derived from their host as well as from the food of the host. In this case, host-derived stimuli are most reliable but hard to detect, whereas those from the host food are easier to detect but less reliable indicators. To describe such a situation, the expression “reliability–detectability problem” was coined (Vet and Dicke 1992). In the case of the varroa mite and possibly other parasites living in close contact with their host, detectability may not be a major problem; thus, reliability becomes the driving force in the process of host finding, and the very peculiar composition of bee larval food seems to provide an easy solution to the problem.

In conclusion, our results suggest that octanoic acid from royal jelly is involved in the repellency of queen cells. Ironically, a trivial compound, octanoic acid, also named caprylic acid by being a component of goat’s milk (Capra hircus), seems to be involved in the bioactivity of royal jelly and gives to honeybee young queens protection against V. destructor.

Despite initial great expectations, semiochemicals have not yet found a use for the control of the varroa mite; repellent compounds shaped by co-evolution may open up new perspectives.


We dedicate this work to the memory of Norberto Milani who inspired this research and to whom we are all greatly indebted for friendship and generous support. We gratefully thank Prof. J.A. Pickett for revising an earlier version of the manuscript.

Copyright information

© Springer-Verlag 2008

Authors and Affiliations

  • Francesco Nazzi
    • 1
  • Renzo Bortolomeazzi
    • 2
  • Giorgio Della Vedova
    • 1
  • Fabio Del Piccolo
    • 1
  • Desiderato Annoscia
    • 1
  • Norberto Milani
    • 1
  1. 1.Dipartimento di Biologia e Protezione delle PianteUniversità degli Studi di UdineUdineItaly
  2. 2.Dipartimento di Scienze degli AlimentiUniversità degli Studi di UdineUdineItaly

Personalised recommendations