Diversity of structure and antiherbivore activity in condensed tannins

INTRODUCTION

Tannins were seldom mentioned in the ecological literature prior to Feeny's (1968, 1969, 1970) work with oak and winter moth caterpillars, but achieved prominence when it was hypothesized that they possess general chemical properties that confer defense against herbivores (Feeny 1976, Rhoades and Cates 1976). The widespread occurrence and high concentrations of tannins in many plant species are now frequently interpreted as the result of selective pressures exerted by herbivores (Coley 1983, 1986, Begon et al. 1990: 111-113, Ricklefs 1990:134-135, Herms and Mattson 1992, Skogsmyr and Fagerstrom 1992).

Despite this, the literature is replete with conflicting data regarding herbivore-tannin interactions. Recent reviews have challenged generalizations (Scriber and Ayres 1988, Bernays et al. 1989, Schultz 1989, Clausen et al. 1992, Feeny 1992). Some research supports the hypothesis that tannins act as an antiherbivore defense by binding with dietary protein and digestive enzymes to limit assimilation in herbivores (e.g., Mitaru et al. 1984, Robbins et al. 1987a, b, 1991, Cooper et al. 1988, Hagerman and Robbins 1993). Other studies suggest that toxicity, rather than digestive inhibition, is the mode of action (Lindroth and Batzli 1984, Steinly and Berenbaum 1985, Blytt et al. 1988, Thomas et al. 1988, Karowe 1989, Mole et al. 1990, Voltura and Wunder 1994). Finally, there are frequent reports that tannins have no detectable effect on herbivores (Bernays et al. 1980, Klock and Chan 1982, Manuwoto and Scriber 1986, Martin et al. 1987, Smith et al. 1992, McArthur and Sanson 1993a), or that they even function as nutritive substrates supporting growth (Bernays and Woodhead 1982).

Research on tannins has been plagued by methodological difficulties. Many studies have not used purified tannins, and extraction procedures have varied among those that have (Hagerman and Butler 1980, 1989, Mole and Waterman 1987a, b, Hagerman 1988, Cork and Krockenberger 1991). Assessments of antiherbivore activity have involved correlative studies, manipulations of whole-plant physiology, amended natural diets, and artificial diets (Schultz 1989). In some cases, different techniques seem to lead to different conclusions (Berenbaum 1986). However, methodological explanations seem insufficient to explain the range of results. Instead, there may be great natural variation in tannin-herbivore interactions, due to structure specificity in the effects of tannins (Zucker 1983, Clausen et al. 1990, Van Altena and Steinberg 1992, Boettcher and Targett 1993), and to physiological variation in the responses of herbivores to tannins (Robbins et al. 1991, Hagerman and Robbins 1993).

Research on tannins has also been obfuscated with the use of lax definitions and characterization of tannins. For instance, tannins are usually defined as substances that (1) have high molecular mass, (2) are polyphenolic, (3) are water soluble, and (4) possess the ability to precipitate proteins. This working definition is useful in describing some chemical and physical characteristics of tannins, but is nonetheless vague. For this reason, many studies have attempted to further classify tannins into hydrolyzable vs. condensed tannins. Hydrolyzable tannins are esters of sugars (usually glucose) with gallic acid, whereas condensed tannins are polymers of flavanols (e.g., Fig. 2). Unfortunately, these subsequent classifications are still vague and often misleading. For instance, it is commonly held that the nonhydrolyzable tannins (i.e., condensed tannins) are not cleaved in mildly acidic or basic media because of their carbon-carbon linkages, and, hence, cannot be assimilated by herbivores. However, condensed tannins are readily cleaved via basic and acidic catalyses into products that could easily pass through the gut wall. Because the products of condensed tannin cleavage will depend on the condensed tannin structure, it seems reasonable that condensed tannin structure would have significant biological consequences that are difficult to generalize.

We conducted experiments to address the following hypotheses: ([H.sub.1]) there is extensive variability, even among closely related plants, in the chemical structure of their condensed tannins; ([H.sub.2]) condensed tannins with different structure differ in antiherbivore activity; ([H.sub.3]) herbivores differ in their overall sensitivity to condensed tannins; ([H.sub.4]) there are interactions between tannin structure and herbivore physiology, so that the same condensed tannin can have different effects on different herbivores; ([H.sub.5]) condensed tannins afford protection for plants against many potential herbivores, but herbivores may evolve adaptations to circumvent the tannin defenses of their host.

To test these hypotheses, we purified and characterized the structure of condensed tannins of 16 woody plant species from seven genera and six families ([H.sub.1]), and assayed the effects of these tannins on the growth and development of six insect species from four genera and two orders ([H.sub.2-5]). We also conducted one experiment using a phenolic glycoside, to compare its activity with that of the condensed tannins. Because our objective was to test broad patterns of variation in insect-tannin interactions, our sample of tannins included as many different plant species as we could manage; we deliberately included some plants that are naturally encountered by our assay insects, and some that are so phylogenetically distant from natural hosts that we could assume no evolutionary history of interactions. Similarly, our sample of insects included as many different species of tree-feeding folivores as we could work with in our laboratory in Alaska.