J Comp Physiol A (2003) 189: 769–775 DOI 10.1007/s00359-003-0451-x
O R I GI N A L P A P E R
B. O. Zimma Æ M. Ayasse Æ J. Tengo¨ Æ F. Ibarra C. Schulz Æ W. Francke
Do social parasitic bumblebees use chemical weapons? (Hymenoptera, Apidae)
Received: 17 April 2003 / Revised: 26 July 2003 / Accepted: 27 July 2003 / Published online: 3 September 2003 Springer-Verlag 2003
Abstract The bumblebee Bombus (Psithyrus) norvegicus Sp.-Schn. is an obligate social parasite of B. (Pyrobombus) hypnorum L. Behavioural observations indicated that nest-invading B. norvegicus females may use allomones to defend themselves against attacking host workers. However, so far no defensive chemicals used by social parasitic bumblebee females have been identified. We analysed volatile constituents of the cuticular lipid profile of B. norvegicus females. Furthermore, we performed electrophysiological studies and behavioural experiments in order to identify possible chemical weapons. Coupled gas chromatography-electroantennography showed 15 compounds to trigger responses in antennae of the host workers. Using gas chromatography–mass spectrometry, the main compound among the cuticular volatiles of B. norvegicus females was found to be dodecyl acetate. A corresponding mixture of synthetic volatiles as well as pure dodecyl acetate showed a strong repellent effect on starved host workers. B. norvegicus females use dodecyl acetate to repel attacking B. hypnorum workers during nest usurpation and subsequently during colony development. Dodecyl acetate is the first repellent allomone identified in bumblebees.
B. O. Zimma Æ M. Ayasse (&) Department of Experimental Ecology, University of Ulm, Albert-Einstein-Allee 11, 89069 Ulm, Germany E-mail:
[email protected] Tel.: +49-731-5022663 Fax: +49-731-5022683 J. Tengo¨ Ecological Research Station of Uppsala University, O¨lands Skogsby 6280, 38693 Fa¨rjestaden, Sweden F. Ibarra Æ C. Schulz Æ W. Francke Institute of Organic Chemistry, University of Hamburg, Martin-Luther-King-Platz 6, 20146 Hamburg, Germany
Keywords Bombus norvegicus Æ Chemical weapon Æ Dodecyl acetate Æ Repellent allomone Æ Social parasitic bumblebees
Introduction Social insect colonies with their brood, adults, and food stores are highly attractive for many organisms that act as scavengers, parasites, predators or even competitors (Ayasse and Paxton 2002). In ants, wasps and bees, a wide range of social parasitism evolved (Wilson 1971; Schmid-Hempel 1998). Michener (2000) defined a social parasite as a female that enters a nest of the social host and in some way replaces the queen, so that host workers thereafter rear offspring of the parasite. Nest defence is, therefore, of extreme importance to social insects, and various mechanisms have evolved to enhance colony survival (Schmid-Hempel 1998; Ayasse and Paxton 2002). A successful parasitism depends on a successful invasion of the host nest, as well as on the ability to remain in the host colony without being expelled or killed (Lenoir et al. 2001). One of the major problems for the parasite is to overcome or evade their host’s defence system. This could be achieved with offensive chemicals that may influence host behaviour. Most research on semiochemicals, employed by social parasites during nest invasion, has been conducted on ants (Lenoir et al. 2001). In the slave-making species Polyergus rufescens the usurping queen uses the Dufour’s gland secretion as chemical weapon while entering a host colony. Since the aggressiveness of Formica cunicularia host workers dramatically decreased towards alien ants experimentally treated with Polyergus queens’ Dufour’s gland secretion, it was assumed that the semiochemicals produced in this gland function as an appeasement allomone to the residents of the host colony (Mori et al. 2000). On the other hand, D’Ettorre et al. (2000) argued that the Dufour’s gland secretion of
770
this species acts more like a ‘‘repellent’’ than an appeasement allomone. In bees much less information exists. Bumblebees of the subgenus Psithyrus are obligate brood parasites of nest-building bumblebees of the genus Bombus. The lack of a pollen-collecting apparatus on the posterior tibia of the parasitic females makes them totally dependent on the host colony to rear their own offspring. After finding and invading a host nest the Psithyrus females treat their host queens in two different fashions: (1) if the Psithyrus females hide inside the nest for a while, they usually either approach the host queen unthreateningly and cohabitate with her without obvious aggression (Fisher 1988; Ku¨pper and Schwammberger 1995; Ku¨pper 1996), and (2) if attacked by the host, they show aggressive behaviour, sometimes attack and grasp host workers and queen, pull them under their body and show stinging movements, called ‘‘mauling’’ (van Honk et al. 1981; Fisher 1983a, 1983b, 1984a, 1984b, 1988). It was assumed that chemical mimicry and/or repellent allomones are responsible for the differences between these two types of nest invasion tactics by a social parasite (Dettner and Liepert 1994). Fisher (1984a) hypothesized that host-specific parasites may be able to evolve a species-specific appeasement substance to invade a colony more efficiently than the host generalist species which violently assault host colonies, since they are unable to produce a pheromone that is effective on a variety of host species which are often not even closely related. Like slave making ants, social parasitic bumblebees have enlarged Dufour’s glands (Fisher and Sampson 1992). Because of the large size and other functions of this gland in various Bombus species (Ayasse et al. 1995, 1999), it is hypothesized that it may play also an important role during a parasite’s nest invasion. Parasitic Bombus females may smear glands’ contents on their cuticle, but so far very few chemical data on the Dufour’s gland secretion of Psithyrus females are available (Taghizadeh 1996). No behavioural experiments to test the role of parasite semiochemicals, used during nest invasion or later in colony development, have been performed. The aim of the present work was to investigate whether semiochemicals are involved during nest invasion of females of the social parasitic bumblebee B. (Psithyrus) norvegicus Sp.-Schn., that exclusively parasitize nests of B. (Pyrobombus) hypnorum L. (Løken 1984). We used gas chromatography coupled with electroantennographic detection (GC-EAD), chemical analyses (GC, GC-MS) and behavioural experiments to identify semiochemicals used as chemical weapons acting against host females. The following questions were addressed: (1) which cuticular volatiles of B. norvegicus females can be detected by workers of their host, B. hypnorum?; (2) are these compounds produced in the Dufour’s gland?; and (3) do the substances that are electrophysiologically active have a repellent effect on host workers?
Materials and methods Rearing of bumblebees In spring 2001 two colonies of B. hypnorum were obtained by offering modified bird nest boxes to nesting site searching queens on the island of O¨land, Sweden and in Carinthia, Austria. After the first worker-brood had emerged, both colonies were transferred to the laboratory in Vienna, Austria and confined to a 30·20·18 cm plastic box without the possibility to forage in the field. The bees were kept in darkness and under conditions of controlled temperature (24C) and relative humidity (50–70%). They were fed with sugar water and pollen ad libitum. Since it was not possible to distinguish between foragers and in-nest workers, the workers were randomly selected from the nest for behavioural tests and electrophysiological investigations. Sample collection B. norvegicus females that were actively breeding in nests of their host species, B. hypnorum, reared on O¨land, were removed from the nests and killed by freezing at )50C. For the odour samples their abdomina were cut and immersed for 30 s in 1 ml pentane for extraction. After dissection under insect Ringer solution the Dufour’s glands were removed, extracted in 100 ll pentane for 24 h at room temperature and the extract stored at )50C. The crude extracts of the Dufour’s glands were concentrated in microvials at 40C to 30 ll, the crude surface extracts of the abdomina had appropriate concentrations for further chemical or electrophysiological analyses. For quantitative analyses 1 lg octadecane was added as internal standard to all samples before GC-EAD recordings.
Electrophysiological recording (GC with GC-EAD) The GC-EAD analyses were performed with a HP 6890 gas chromatograph equipped with a DB5-MS column (30 m·0.32 mm i.d., 0.25 lm film; J&W Scientific, Folsom, Calif., USA) that was temperature programmed from 50C to 310C at a rate of 10C min)1. Helium was used as the carrier gas; a GC effluent splitter (split ratio 1:1) was used, and the outlet was added to a purified and humidified airstream, directed over the excised antenna of B. hypnorum workers. The tip of the antenna was cut off and the antenna mounted between two glass electrodes filled with insect Ringer solution. The electrode holding the base was connected to a grounded Ag-AgCl wire, the electrode at the antenna’s tip was connected via an interface box to a signal acquisition interface board (IDAC; Syntech, Hilversum) for signal transfer to a PC. The responses of the flame ionization detector (FID) and the EAD signals were recorded simultaneously. The GC-EAD analyses were performed with cuticular surface extracts of B. norvegicus females. Only such odour compounds were termed ‘‘active’’ which showed reproducible EAG responses in at least four GC-EAD runs.
Chemical analyses of volatiles Chemical analyses of volatiles were performed by gas chromatography (Hewlett Packard 5890 equipped with a 30 m DB5-MS column, 0.32 mm i.d., 0.25 lm film; J&W Scientific, Folsom, Calif., USA), operated splitless at 120C for 1 min then programmed to 300C at a rate of 4C min)1 (Ayasse et al. 2000). Structure elucidation of active cuticle constituents was based on combined GC-MS at 70 eV using a VG70/250 SE instrument (Vacuum Generators, Manchester, England), linked to a HP 5890 gas chromatograph. Analytical conditions were the same as mentioned above. Mass spectra were compared with those reported in the literature (McLafferty and Stauffer 1989) and gas chromatographic
771 retention times (coinjection) with those of authentic reference samples. To identify the unknown compounds, a databank of retention times of known chemical compounds was used. Double bond positions in unsaturated compounds were assigned according to Buser et al. (1983) and Dunkelblum et al. (1985).
Repellent tests Synthetic blends of the electrophysiologically active compounds were prepared according to the relative proportions of each of the compounds in the natural cuticular extracts, giving a concentration of one animal-equivalent per 20 ll. Solutions of citral (reported as a repellent in stingless bees; Blum et al. 1970) and of farnesol (a known flower odour; Knudsen et al. 1993) were prepared at a concentration of 7.5 lg per 20 ll solvent. The main compound dodecyl acetate was tested at a concentration of 150 lg per 20 ll solvent, corresponding to its concentration in the natural sample. The repellent effect of the cuticle surface extract of B. norvegicus females was tested by offering a droplet of sugar water (1 kg sugar per 1 l water), surrounded with solvent or chemical substances, to workers of B. hypnorum that had been starved for 6–7 h (modified after D‘Ettorre et al. 2000). Altogether 59 tests were carried out in round plastic dishes (8.5 cm in diameter), the sugar water was applied in the centre on a glass cover slip (2 cm·2 cm) and surrounded with 20 ll of a solution of (1) nine of the major substances identified among the 15 GC-EAD active compounds (Fig. 1), (2) dodecyl acetate, (3) citral, (4) farnesol (control), or (5) pentane as a solvent control. All test were carried out under red light at 24C between 6 p.m. and 8 p.m. Prior to testing, the bumblebees were allowed to acclimate for 1 min in a small glass tube in the arena next to the cover slip. After removal of the glass tube, we recorded the time it took the bees to reach the droplet of sugar water and registered the retreat behaviour when approaching the area impregnated with test solution. If the bumblebee worker did not feed until the end of the experiment
(3 min), it was put in a new plastic-dish with an untreated droplet of sugar water to test its starved status. The data were used for statistical analyses only when the bumblebee fed on the sugar water during this control test. In this case the time limit until feeding was set at 180 s. In four cases the workers never came close to the droplet and stayed away from the impregnated area during the whole experiment. These bees drank in the control test with untreated sugar water. In these cases the time until feeding (180 s) and not the number of retreats was used for statistical analysis because only the behaviour tests were used when the worker finally reached the droplet, otherwise the results could be distorted.
Statistics Absolute and relative amounts of GC-EAD-active compounds were tested for the significance of differences between cuticle surface extracts and Dufour’s gland extracts of B. norvegicus females using a Mann-Whitney U-test. The time until feeding and the behaviour during repellent tests were compared among test groups using analyses of variances (ANOVA) followed by a multiple comparison test (LSD-test; Norusis 1993a, 1993b).
Results Perception of parasite’s volatiles by host workers Fifteen compounds from the cuticular extract of parasitic B. norvegicus females triggered receptor potentials in worker antennae of their host B. hypnorum (Fig. 1). Volatiles from three classes of chemical compounds were identified: eight esters (decyl acetate, (Z)-3-dodecenyl acetate, dodecyl acetate, tridecyl acetate, tetradecyl acetate, (Z)-7-hexadecenyl acetate, oleic acid ethylester and (Z)-9-octadecenyl acetate), two alcohols (dodecanol and (Z)-9-octadecenol), and four aldehydes (tetradecanal, (Z)-7-hexadecenal, hexadecanal and (Z)-9-octadecenal). One compound (retention time 19.56 min) remained unidentified. Although the EAG responses of the active compounds were rather weak, the reproducibility of all reactions was proven in repeated GC-EAD runs. Chemical composition of Psithyrus scent
Fig. 1 Coupled gas chromatography-electroantennography (GCEAD) recording of the cuticular extract of a Bombus norvegicus female using an antenna of a worker of B. hypnorum (an asterisk denotes compounds used in the repellent tests)
The main classes of chemical compounds present in the cuticular extracts of B. norvegicus females were found to be alkanes, alkenes, and esters. Dodecyl acetate was the main component. Tetradecyl acetate as well as straight chain alkanes and alkenes with an uneven number of carbons and a length between 23 and 29 carbons occurred in smaller amounts. The uneven numbered alkanes always showed higher concentrations than the adjacent even-numbered ones (Hefetz et al. 1996). The mean absolute amount of dodecyl acetate was 162.0±65.6 lg (SE) per female. This compound comprised 85% of all GC-EAD-active compounds. Tetradecyl acetate occurred in much smaller amounts (x±SE=12.1±3.9 lg/female).
772
Fig. 2 Comparison of relative proportions (mean+SE) of GCEAD-active compounds in cuticular extracts (n=8) and Dufour’s gland extracts (n=4) of B. norvegicus females; for names of compounds see Fig. 1 (an asterisk denotes significant differences, Mann-Whitney U-test, P
