Terpenoid antifeedants against insects: a behavioural and sensory study.

This thesis describes a study on the behavioural and sensory effects of terpenoid antifeedants on several insect species. The main aim was to elucidate the mechanisms of action of terpenoid antifeedants. From a fundamental point of view, this will yield insight in the role of these compounds in host plant selection by insects. From an applied perspective, knowledge of the level of effectiveness and of putative structure-activity relationships, provides a basis to assess the potential for using terpenoids as crop protection agents.The effects of specific molecular structures were studied comparatively in different unrelated insect species. The mechanism(s) of action of sesquiterpene drimanes were investigated by measuring sensory responses and observing their effects on behaviour. Furthermore, long-term and toxic effects were studied.Larvae of the large white butterfly Pieris brassicae L. (Lepidoptera: Pieridae) were used as model insect for studying the sensory perception of sesquiterpene drimane antifeedants by insects. By comparing sensory and behavioural responses to 15 drimane antifeedants, it was shown that the frequency of impulses from the deterrent cell in the medial sensillum styloconicum significantly correlates to the behavioural response in dual-choice tests. In addition, the drimanes caused inhibition of sensory responses to feeding stimulants. The results suggest that in this insect a 'labeled line' coding principle is used for the perception of antifeedants, combined with inhibition of receptor cells sensitive to feeding stimulants (chapter 2).The temporal aspects of the behavioural effects of five drimanes on P. brassicae larvae were studied with aid of detailed, 1 min interval behavioural observations in a no-choice test (chapter 3). It was found that two of the five drimanes (polygodial and warburganal) inhibited feeding only during the beginning (0-30 min) of the tests and that two other drimanes (confertifolin and 'compound 4') became effective 30-90 min after the onset of the test. The fifth compound (isodrimenin) inhibited feeding during the whole 3 h period. It is concluded that some drimanes ( e.g. confertifolin and compound 4) have postingestive, toxic effects resulting in inhibition of feeding. These and the results of chapter 2 indicate that analogous drimane antifeedants can inhibit feeding in P. brassicae through multiple mechanisms of action. Probably P. brassicae larvae in no-choice situations soon habituate to drimane antifeedants, unless the drimanes are toxic. The results indicate that, when developing a structure-activity relationship for a series of antifeedants, it is essential to distinguish the mode of action which underlies inhibition of feeding.The behaviour of larvae of the Colorado potato beetle, Leptinotarsa decemlineata Say (Coleoptera: Chrysomelidae) was studied when exposed to three drimanes in 3 h no-choice tests (chapter 4). In contrast to P. brassicae larvae L. decemlineata larvae were inhibited from feeding throughout the 3 h no-choice tests when treated with polygodial or warburganal, while to compound 4 habituation occurred after the first 15 min of the test. The results indicate that polygodial and warburganal are either strong antifeedants or possess toxic properties as well, preventing that habituation occurs. Topical application of polygodial and warburganal to the larval cuticle also inhibited feeding. This could indicate that toxic properties of these molecules might contribute to feeding inhibition, although the effects of topical application are not necessarily the same as the effects after ingestion of the compounds ( e.g. not in the case of P. brassicae ; chapter 3).The hypothesis that feeding deterrents in potential food plants are decisive in food plant selection by insects (chapter 1) was originally derived from experiments with the Colorado potato beetle. However, little was known about the sensory perception of such compounds by this beetle. In this study the role of an epipharyngeal sensillum in the perception of antifeedants was investigated (chapter 5). Electron microscopy revealed that the epipharyngeal sensillum is innervated by five neurons. Electrophysiological experiments showed that one of these cells responds to water, a second to sucrose and a third to two feeding deterrents ( i.e. compound 4 and sinigrin) that had been found effective in a behavioural test. Furthermore, the response of the sucrose sensitive cell was strongly inhibited by the drimane and only slightly by sinigrin. From a comparison of behavioural and sensory responses it is concluded that probably both the response of the epipharyngeal deterrent cell ànd peripheral interactions exerted by feeding deterrents on the sucrose sensitive cell in this sensillum determine the potency of feeding deterrents in L. decemlineata larvae. The results provide a physiological basis for the hypothesis that the presence of feeding deterrents in potential food plants is a decisive cue in food plant selection by L. decemlineata larvae.The sensitivities of nymphs of two aphid species, Myzus persicae Sulzer and Aphis gossypii Glover (Homoptera: Aphididae) to 11 drimane compounds, applied to the lower surface of artificial diet sachets, were compared in dual-choice tests. In general, A. gossypii nymphs were less sensitive to the drimanes than M. persicae nymphs. Warburganal and polygodial were highly active as deterrents and/or feeding inhibitors against both species. In dual-choice tests in which the upper parts of the nymphal antennae were ablated, it was found that both species do not detect the drimane polygodial anymore. It was concluded that nymphs of M. persicae and A. gossypii detect polygodial and probably the other drimanes tested with contact chemosensilla at the tips of their antennae. The ablation studies also showed that in both species no tarsal, labial or epipharyngeal sensilla are involved in detecting polygodial within 24 h. The results indicate that deterrents on the leaf surface might play a role in the selection of host plants by these and other aphid species.Chapter 7 reports on additional bioassays with drimanes on several insect species. In chapter 8, several synthetic analogues, derived from the C-9 side chains of the diterpenoids clerodin and ajugarin I, were tested for their effects on feeding of P. brassicae and L. decemlineata larvae and M. persicae nymphs. Several compounds showed moderate activity against P. brassicae larvae. Interestingly, the most effective compound most closely resembled the structure of the furopyran-fragment of the triterpenoid azadirachtin, which fragment had previously been shown to be a highly effective feeding deterrent against Spodoptera littoralis .In conclusion, the mode of perception as well as the structure-activity relationship of terpenoid antifeedants considerably differs between the insect species examined in this study (chapter 9). For practical implementation of terpenoid antifeedants in crop protection, this means that research should be focussed on selected pest insect species. Furthermore, the possibilities for combined use with other (biological) control agents should be investigated