Costs and benefits of iridoid glycosides in multitrophic systems

Because plants cannot run away from their attackers, move to more favourable locations or hide, they have to be able to tolerate, adapt and/or defend themselves. Plants have evolved an enormous array of mechanical and chemical defences against herbivores. One can distinguish three types of defence strategies: direct defences that directly affect the herbivores, indirect defences that attract the enemies of the herbivores and tolerance, which reduces the fitness consequences of herbivore damage. It is unlikely that direct and indirect defences act independently in plants. Natural enemies of the herbivores (the indirect defence) can be negatively affected when attacking larvae which feed on plants defended by high levels of allelochemicals (direct defence), causing potential incompatibility between host plant resistance and biological control. On the other hand, herbivores feeding on plants with high levels of allelochemicals may develop more slowly and have an impaired immune function, leading to longer exposure of vulnerable stages to parasitism and predation and causing potential synergism between host plant resistance and biological control. Studies of interactions between direct and indirect defences are almost exclusively based on studies with crop species. It has frequently been observed that domesticated plants have lower levels of defensive chemicals than their wild relatives.

In this study I used a natural plant-herbivore-parasitoid complex. The main focus in this thesis is the effect of direct defence chemicals in the plant on higher trophic levels in the system, to detect if there is potential for a conflict between direct chemical defence and indirect defence of the plant.

 

Natural enemies of herbivores are a potential source of indirect defences for a plant. Plants can attract these enemies by providing shelter (domatia), food or signalling them with volatiles. The efficiency of these indirect defences depends on the effect of the natural enemy on the herbivore. Parasitoids often have an optimal host instar in which to parasitize their hosts. In this optimal instar they develop faster and/or grow bigger and have higher survival. In chapter 2 I examined this optimal host size for a generalist koinobiont tissue-feeding larval endoparasitoid, Hyposoter didymator, and two of its natural hosts, Spodoptera exigua and Chrysodeixis chalcites. Koinobiont parasitoids attack hosts that continue feeding and growing during parasitism. In contrast with hemolymph-feeding koinobionts, tissue-feeding koinobionts face not only a minimum host size for successful development, but also a maximum host size, since consumption of the entire host is often necessary for successful egression. I hypothesized that the range of host instars suitable for successful parasitism by H. didymator would be much more restricted in the large host C. chalcites than in the smaller S. exigua. In contrast with our predictions, C. chalcites was qualitatively superior to S. exigua in terms of the survival of parasitized hosts, the number of parasitoids able to complete development and adult parasitoid size. However, in both hosts the proportion of mature parasitoid larvae that successfully developed into adults was low at the largest host sizes. Our results suggest that qualitative, as well as quantitative factors are important in the success of tissue-feeding parasitoids.

One of the direct defence mechanisms of plants against herbivores is the production of allelochemicals. However, the effects of these defence compounds is not necessarily restricted to herbivores but can extend to higher trophic levels in the food chain, including the predators and parasitoids of herbivores. In chapter 3 I examined the effects of two defence chemicals of Plantago lanceolata, the iridoid glycosides (IGs) aucubin and catalpol, on the performance of two generalist and two specialist herbivores and their endoparasitoids. Furthermore, I studied the sequestration of these chemical compounds in the herbivore-parasitoid complex of the specialist herbivore Melitaea cinxia. In general, the performance of generalist herbivores was negatively correlated with the levels of IGs but effects on the performance of their parasitoids were less apparent. Moreover, because herbivores developed more slowly on high IG plants, instars vulnerable to parasitism suffered an increased period of exposure to the parasitoids. On the other hand, effects on specialist herbivores differed between M. cinxia and the other specialist herbivore, Junonia coenia. The development of J. coenia was slower when feeding on plants containing high IG levels, whereas the pattern was reversed in M. cinxia. Similarly, development of Cotesia melitaearum, a gregarious endoparasitoid of M. cinxia caterpillars, benefited when it developed in larvae reared on P. lanceolata genotypes with high levels of IGs. Iridoid glycosides were detected in all tissues of the specialist herbivore M. cinxia, in its endoparasitoids and in two of their hyperparasitoids. In pupae and adults, the fraction of catalpol, the more toxic of the two IGs, increased with trophic level.

Another characteristic of IGs is that they can be oviposition stimulants for specialist herbivores and feeding stimulants for their larvae, especially when their performance is better on plants that contain IGs. In chapter 4 I studied the effect of IGs and aspects of plant size (mainly the number of leaves) of the host plant P. lanceolata, on the oviposition behaviour of its specialist herbivore M. cinxia. A previous study of the same species showed that oviposition was associated with high levels of aucubin in plants in the field, but it did not distinguish whether the higher levels of aucubin were the cause (active choice) or consequence (induction) of oviposition. I conducted a set of dual- and multiple-choice experiments between plants with different levels of IGs, in cages and in the field. In the cages I found a positive correlation between the pre-oviposition level of aucubin and the number of ovipositions, indicating an active oviposition decision for plant with higher aucubin level, rather than plant induction following oviposition. The results also suggest a threshold concentration below which females do not distinguish among levels of IGs. In contrast to the cage experiment, in the field the size of the plant appeared to be a more important stimulus than the IG concentrations, with bigger plants receiving more ovipositions than the smaller ones, regardless of their secondary chemistry. Therefore, the predominant cues used for oviposition may be dependent on environmental conditions.

That not only plant chemicals play a role in oviposition choice, is also clear from chapter 5, in which I looked at the oviposition preference, habitat use and food plant suitability of another specialist on P. lanceolata, M.athalia. In a big cage experiment I studied the oviposition choice of this butterfly. For the oviposition experiment I used eight different plant species, all containing IGs. The plant species the females preferred for oviposition were Veronica chamaedrys, V. spicata and P. lanceolata. All of these plants grow in open meadows, which is where I also found the adults flying most frequent in the field. The difference in host plant and habitat use between Åland, Finland (where the field observations where done) and other regions, may reflect local adaptation to land use practices and geology which maintain clusters of small open meadows. Despite the fact that the presence of IGs is an important trait distinguishing host from non-host species used by M. athalia, oviposition preference within the group of (potential) host species and among individual plants within host species was largely independent of IG concentration. Although the adult butterflies chose specific plant species for oviposition, the immediate surrounding of these host species was more important than the IG concentrations of these plants. Plants in plots surrounded by bare ground received significantly more egg batches than plants in plots surrounded by vegetation. The larvae of M. athalia did not profit from the oviposition choice of their mother. They performed equally well on all the 13 plant species used for the performance experiment, except for V. officinalis.

Many natural plant populations exhibit significant genetic variation in their levels of chemical defence against herbivores and pathogens. In P. lanceolata I also found variation in their levels of IGs. One of the factors that could contribute to the maintenance of this variation is the presence of fitness costs of chemical defence. In chapter 6 I examined if there were fitness costs of having higher levels of IGs. This would imply that in the absence of natural enemies, the production, transport, storage, self-detoxification, activation and/or turnover of secondary plant compounds results in lower plant fitness as resources used for these processes cannot be used for growth, survival or reproduction. In this chapter I describe a regrowth experiment to investigate whether there are trade-offs between resistance and one specific aspect of tolerance, the ability to regrow after defoliation. I let plants with different levels of IGs grow under two nutrient conditions, poor and rich. After 8 weeks I clipped the aboveground biomass, and let the plants regrow for five weeks. The questions I asked were: do high-IG plants (1) suffer allocation costs in terms of shoot and root growth, (2) have reduced regrowth ability (tolerance) after defoliation and (3) are such costs more pronounced under nutrient stress? I found that the total biomass produced by high-IG plants was not lower than that of low-IG plants. However, high-IG plants produced fewer inflorescences (a reproductive cost) and allocated less biomass to roots than low-IG plants. After regrowth, root mass of high-IG plants grown under nutrient-poor conditions was significantly lower than that of low-IG plants. I speculate therefore that if there would be repeated defoliation, high-IG plants would eventually fail to maintain shoot regrowth capacity and that trade-offs between resistance (having high IG levels) and tolerance in this system may not show up until repeated defoliation events occur.

In chapter 7 I discuss the results of the effects of IGs on herbivores and their parasitoids and whether these observed patterns differ among generalist and specialist herbivore-parasitoid combinations. I conclude that how direct and indirect defence mechanisms interact depends on the combination of species involved. A possible conflict between these two defence strategies may arise if the plants that are most attractive to natural enemies also possess strong direct chemical defences that exhibit clear negative effects on the performance of predators and parasitoids, or when there are metabolic trade-offs between these two kinds of defences. Future studies with natural plant species should examine wether these conflicts really exist in nature