Host foraging for differentially adapted brassica-feeding aphids by the braconid parasitoid Diaeretiella rapae

Abstract

Interactions occurring in a tritrophic system comprising plants, aphids and parasitoids are of great complexity. The generalist endoparasitoid Diaeretiella rapae (McIntosh) (Hymenoptera: Aphidiidae) displays specialist characteristics on brassica feeding aphids. Previously, we studied differential signalling to D. rapae by specialist and generalist Brassicaceae feeding aphids on turnip. We reported no differences in the attractiveness of volatile compounds from the two turnip/aphid complexes. However, we reported a significantly greater D. rapae attack rate on the specialist Lipaphis erysimi (Kaltenbach) than the generalist Myzus persicae (Sulzer). As a consequence we predicted that D. rapae would forage more efficiently and produce more offspring on L. erysimi. We present here some additional data collected in a more complex spatial/temporal environment in large experimental chambers and discuss this, drawing attention to the need for careful interpretation of mechanistic information in predicting the overall foraging process.

Plants are a principal component of parasitoid-host interactions.¹ Plant derived chemical cues may be exploited by parasitoids to locate successfully host habitats, hosts and to assess host quality either positively or negatively.² However, there are a multitude of other direct and indirect interactions that structure aphid-parasitoid communities.³ Plants of the family Brassicaceae have a diverse and interesting secondary chemistry, which forms an important component of their defence against herbivores. This secondary chemistry includes glucosinolates located within plant cells, which upon damage are hydrolysed by enzymes known as myrosinases.^4,5 Products of these reactions include nitriles and volatile isothiocyanates. Lipaphis erysimi (Kaltenbach), the turnip aphid, is a specialist on plants of the order Capparales, particularly the Brassicaceae, and like its fellow specialist, Brevicoryne brassicae,⁶ sequesters glucosinolates using them as defence against natural enemies,⁷ whereas Myzus persicae (Sulzer), the peach-potato aphid, does not sequester glucosinolates⁷ but excretes large quantities in its honeydew.⁸ We hypothesised that Diaeretiella rapae (McIntosh) would display different foraging behaviours toward plant-derived chemical signals from these differentially adapted aphids feeding on turnip.

We have studied some of the behavioural responses of D. rapae to these two aphid species on a single cultivar of turnip (Brassica rapa var rapifera cv Tokyo Cross). Volatile compounds from turnip infested by L. erysimi or M. persicae are both attractive to D. rapae, but not differentially so. The volatile composition of both turnip/aphid complexes (TAC) includes two isothiocyanates, of which one, 3-butenyl isothiocyanate, was shown to be attractive in Y-tube bioassays.⁷ In no choice tests, attack rates, defined as the number of full contacts between the ovipositor of a parasitoid and its potential aphid hosts in a minute (counting multiple stabbings repetitively on the same aphid as one attack), were significantly greater on L. erysimi than on M. persicae, irrespective of which of these aphids was the original host.⁹ This is possibly due to glucosinolate derived kairomones that may be detected either by antennation or ovipositor probing on the potential host.¹⁰ This behaviour is primed by cues received by the parasitoid during emergence from its mummy case.⁹ Taking these mechanistic results together, we hypothesised that given an equal supply of potential hosts, D. rapae would parasitize more L. erysimi aphids than M. persicae, irrespective of the host in which they developed.

We tested this hypothesis in four Perspex chambers (L 1.7 m, W 1.2 m, H 1 m) with each containing four potted turnip plants infested with 250 L. erysimi aphids and four with 250 M. persicae aphids. Plants infested with the different aphid species were placed alternately within the chamber in two equally spaced parallel rows. A total of 15 naive D. rapae parasitoids were released from three points in each chamber. A total of four one-chamber replicates were made using D. rapae reared on L. erysimi and four with D. rapae reared on M. persicae. Eleven days after the release of parasitoids the number of mummified aphids on each plant was recorded.

We found that D. rapae reared on L. erysimi produced significantly more mummies on L. erysimi than M. persicae, as we hypothesised. However, D. rapae reared on M. persicae produced significantly more mummies on M. persicae than L. erysimi (Fig. 1). This was contrary to our hypothesis. In addition, there were significantly more mummies produced by parasitoids reared on L. erysimi than those reared on M. persicae.

Figure 1.

Average number of mummified aphids recovered per species (LE = Lipaphis erysimi; MP = Myzus persicae) per cage. * indicates a significant difference by t-test in the number of mummies produced on LE and MP, (t = 2.39, p = 0.048) and (t = 2.40, p = 0.048) for D. rapae reared on M. persicae and L. erysimi respectively.

In an attempt to identify possible reasons for these observations we recorded different aphid patch parameters for M. persicae and L. erysimi including patch area, aphids per patch and patch density. M. persicae formed larger less dense patches than L. erysimi (Fig. 2). As well as forming different patch structures the aphids are different in colour, M. persicae being generally yellow, while L. erysimi are a dark grey/green. Green and yellow pigments reflect distinctly different spectra of light, which can be detected by the trichromatic vision common to most hymenoptera.¹⁰ Thus colonies of the different species present different visual cues, which may provide essential foraging cues to parasitoids.^11,12 However, it seems unlikely that the parasitoids used in this experiment, naive but for the conditioning received upon emergence from their mummy case, would display different responses to visual cues. We propose that the chemistry of the system is likely to hold the key to the behaviour of D. rapae, and put forward the following explanation.

Figure 2.

Aphid patch parameters, including patch area (cm²), number of aphids per patch, and density of aphids within the patches (Aphids/cm²) for Lipaphis erysimi (LE) and Myzus persicae (MP).

D. rapae does not have a preference for the volatile compounds associated with either TAC. So, though not necessarily random, it is probable that there is an equal chance of encountering each aphid species first. It is evident that one of the main differences between the two TAC is the chemical constituents and quantity of honeydew, which has been shown to alter on-plant foraging for B. brassicae by D. rapae.^13–15 The honeydew composition of aphids is species and plant specific¹⁵ and has been shown to provide parasitoids with information on host species identity through species-specific kairomones.¹⁷ M. persicae feeding on the Brassicaceae produce a honeydew rich in glucosinolates.⁸ A likely consequence is that the cuticle of the M. persicae mummy case will retain some of this chemistry which will then be received by the parasitoid upon emergence. This stage of a parasitoid's life has been shown to influence adult responses to volatiles^18–20 and attack rate.⁴ Subsequent location of this same substance whilst foraging could prompt parasitoids to spend an increased time residing in these patches. It has been shown in the hymenopterous whitefly parasitoid Encarsia formosa (Gahan) that honeydew of a range of species including non-hosts will have an arrestant effect, but that host honeydew results in longer searching times.¹⁵ This same effect could account for our findings, and the lower total number of mummies produced could be explained by the innately lower attack rate on M. persicae compared to L. erysimi.

In conclusion we draw attention to the following points.

1. It is evident that for some parasitoid species emergence from their mummy case provides information that can manifest itself in long lived host foraging decisions. We hypothesise that aphid honeydew provides an important contact cue that may influence the residence time of D. rapae, and could for other specialist parasitoids.

2. Elucidation of parasitoid host foraging mechanisms provides important information, but increasing experimental complexity may not reveal intuitive results. This is likely to become even more important under heterogeneous field conditions, highlighting the need for corroborating the significance of laboratory studies through well designed field or field-simulation studies.

Addendum to: Blande JD, Pickett JA, Poppy GM. A Comparison of Semiochemically Mediated Interactions Involving Specialist and Generalist Brassica-feeding Aphids and the Braconid Parasitoid Diaeretiella rapae. J Chem Ecol. 2007;33:767–779. doi: 10.1007/s10886-007-9264-7.