action, and is characterized by great diversity— especially in those plant taxa on which phytoche-mists have focused their attention (Jones and Firn 1991). In fact, Harborne (1993) considers the three primary areas of biological diversity to be angiosperms, insects, and secondary compound chemistry. Chemical defences have been classified according to various plant-herbivore theories (discussed by Speight et al. 1999). In broad terms, a distinction has been made between toxins (produced in small quantities by rare or ephemeral plants) and digestibility-reducing allelochemicals ('quantitative' or deterrent defences, produced in large quantities by long-lived or apparent plants). Mattson (1980) pointed out that toxic compounds are associated with nitrogen-rich plants such as legumes, and many of the toxins are nitrogen-based (e.g. alkaloids, cyanogenic glycosides, non-protein amino acids), although others (cardiac glycosides) are not (Harborne 1993). Moreover, proteinase inhibitors are nitrogen-based but not toxic. Digestibility-reducing allelochemicals, on the other hand, tend to occur in plants adapted to low nitrogen environments (such as Eucalyptus) and are carbon-based (tannins, terpenoids). According to the carbon-nutrient balance hypothesis of Bryant et al. (1983), the production of defences is costly to plants and involves a trade-off between growth and defence. Nitrogen-based defences increase in concentration when nitrogen availability to plants increases, while concentrations of carbon-based defences decrease (Kyto et al. 1996). More specifically, Haukioja et al. (1998) have proposed that protein synthesis, because of the requirement for phenylalanine, competes with synthesis of condensed tannins. Conflicting results are common in tests of the carbon-nutrient balance hypothesis (Hamilton et al. 2001) and other hypotheses generated to explain patterns in plant defences: see Cipollini et al. (2002) for an example involving measurement of many different variables in a thorough study of leaf chemistry and herbivory. The distinction between toxic and deterrent plant allelochemicals can be misleading: a recent review uses the term 'antinutrients' for natural products which reduce nutrient availability to insects (Felton and Gatehouse 1996). Reduction of nutrient availability may involve chemical modification of the nutrient, formation of complexes with it, or interference with its digestion or absorption, and the effect of antinutrients can be overcome by providing the insects with supplemental nutrients.
Feeding experiments involving these aromatic compounds require careful design and have shown some complex effects. For example, Raubenheimer and Simpson (1990) found no interactive effects of tannic acid and protein-carbohydrate ratios in Locusta, which is a grass-feeding oligophage, and Schistocerca, which is polyphagous and receives greater exposure to tannins in its diet. In the short term, tannic acid had a phagostimulatory effect on Schistocerca, and this must be distinguished from compensatory consumption of an inferior diet. More recent use of the geometric approach has shown that tannic acid is more effective as a feeding deterrent in Locusta than as a post-ingestive toxin, but its effect is markedly influenced by the proportions of protein and carbohydrate in the food (Behmer et al. 2002). Tannic acid has a stronger deterrent effect when foods contain a large excess of carbohydrate relative to protein, whereas locusts are willing to consume relatively large amounts of tannic acid included in protein-rich diets. These authors suggest that this is another factor favouring carbon-based defences in resource-poor habitats.
Eucalyptus species contain very high concentrations of phenols and essential oils (mixtures of terpenoids), yet these do not seem to affect feeding by chrysomelid beetles, which are far more constrained by the low nitrogen content of Eucalyptus foliage (Fox and Macauley 1977; Morrow and Fox 1980). This explains the apparent paradox of heavy grazing in spite of these quantitative chemical defences. Phenolics have varied effects, both inhibitory and stimulatory, on the performance of insects (Bernays 1981) and negative effects on feeding and growth may involve a variety of mechanisms, including oxidative stress. Felton and Gatehouse (1996) exclude tannins from their review of plant antinutrients.
Caterpillars maintain strongly alkaline midguts, and Berenbaum's (1980) survey of published gut pH values for 60 species showed that those feeding on woody plant foliage have higher gut pH than those feeding on herbaceous plants. Condensed tannins are characteristic of trees, and high pH may be advantageous in reducing the stability of tannin-protein complexes. The association between phylogeny, diet, and midgut pH was considered further by Clark (1999). Exopterygote insects have near-neutral midguts, and physicochemical conditions in grasshopper guts are apparently not influenced by patterns of host plant use (Appel and Joern 1998), although Frazier et al. (2000) suggest that some grasshopper diets pose significant acid-base challenges. In an excellent review, Appel (1994) has stressed the importance of the gut lumen of insect herbivores as the site of interaction of nutrients (often refractory), insect and plant enzymes, allelochemicals, and pathogens. These interactions are affected by widely differing pH, redox conditions, and antioxidant activities. Although alkaline pH weakens protein-tannin binding, this mode of action is now considered less likely than oxidation of phenols to reactive quinones (Appel 1994). Oxidation of allelochem-icals to toxic metabolites is, in fact, favoured by high pH. However, low oxygen levels in the gut lumens of herbivores reduce the rates of oxidation of allelochemicals (Johnson and Rabosky 2000). Ascorbate, an essential nutrient for many insects, is an antioxidant that maintains phenols in a reduced state in the gut lumen, minimizing their negative effects, and the recycling of ingested ascorbate may be the biochemical basis of differing tolerances to tannins among caterpillar species (Barbehenn et al. 2001).
Research on plant allelochemicals has shifted primarily to antinutrient proteins, because of their enormous potential in plant biotechnology (Lawrence and Koundal 2002). Development of transgenic crops expressing genes for insect resistance was first based on expression of Bt toxins, but transgenic plants equipped with genes for protei-nase inhibitors and lectins are providing interesting opportunities for collaboration between chemists, physiologists, and applied ecologists. Proteinase inhibitors are inducible plant defences that are synthesized in leaves as a direct response to feeding, not only at the site of attack, but also throughout the plant, although the response declines with plant age. They are also constitutively produced in seeds and storage organs of many staple crops (Jongsma and Bolter 1997). The signalling cascade that is initiated by feeding damage and leads to proteinase inhibitor gene expression is described by Koiwa et al. (1997). Proteinase inhibitors work by binding directly to the active sites of the enzymes to form complexes, mimicking the normal substrates but effectively blocking the active sites. Digestion of plant protein is inhibited and the insects are effectively starved of amino acids and prone to amino acid deficiencies.
Soybean trypsin inhibitor was the first proteinase inhibitor shown to be toxic to insects, and the trypsin inhibitors are particularly well known, partly because trypsin is commonly used in screening procedures for proteinase inhibitors (Lawrence and Koundal 2002). Based on primary sequence data, there are at least eight families of serine proteinase inhibitors in plants (Koiwa et al. 1997). Cysteine proteinase inhibitors (cystatins) are best studied in rice and are effective against some Coleoptera, whereas serine proteinase inhibitors are most effective against Lepidoptera. Effects on performance are commonly evaluated in insects feeding either on artificial diets containing the proteins or on the transgenic crops (this provides an opportunity to evaluate the effect of plant allelochemicals in natural diets, by comparing herbivore performance with that on unmodified crops). Jongsma and Bolter (1997) present data from numerous studies investigating the effects of proteinase inhibitors on various fitness parameters of insects. Frequently, the results of feeding experiments have been disappointing in comparison to those from in vitro experiments with gut extracts and proteinase inhibitors. Moreover, inhibitory effects may be surprisingly poor in transgenic plants, as shown by tomato moth larvae, Lacanobia oleracea (Noctuidae) subjected to a soybean inhibitor in artificial diets and in transgenic tomato plants (Fig. 2.13) (Gatehouse et al. 1999). There are many factors that may be responsible for such discrepancies, such as expression levels in the plant tissue, inhibitor-enzyme affinity, diet quality,
Was this article helpful?