ANTIFUNGAL ACTIVITY OFSOLANUMFRUIT GLYCOALKALOIDS: IMPLICATIONS FOR FRUGIVORY AND SEED DISPERSAL

Abstract

We tested two sets of hypotheses about the role of secondary metabolites (glycoalkaloids) in ripe Solanum (Solanaceae) fruit. The Directed Toxicity Hypothesis states that these compounds are most toxic to frugivores that do not disperse seeds (e.g., fungi), whereas its alternative, the General Toxicity Hypothesis, states that they are generally toxic to all frugivores. The Removal Rate Hypothesis states that highly rewarding fruits, which are removed rapidly by dispersers, require only low levels of glycoalkaloids for protection against insects and microbes. The alternative Nutrient–Toxin Titration hypothesis states that highly rewarding fruits may be able to maintain higher levels of glycoalkaloids because high nutrient content may offset the negative effects of glycoalkaloids on dispersers. Our previous work with vertebrate frugivores supported the General Toxicity and Removal Rate hypotheses. We used artificial fruit pulp media to measure how the mycelial growth rate of nine fungal species was affected by two glycoalkaloids, α-solasonine and α-solamargine, both common in ripe Solanum fruits. In one set of fungal growth trials, we mimicked two species of fruit, S. americanum with low levels of glycoalkaloid in ripe fruit and S. carolinense with high levels, to determine the relative effects of nutrient and toxin content. In another set of trials, we examined the relative and interactive effects of the two glycoalkaloids. Finally, we tested whether or not increases in nutrient content could offset the negative effects of glycoalkaloids on fungal growth. All fungi tested were inhibited by the glycoalkaloids, with lower levels of antifungal activity for agars corresponding to (low-toxin) S. americanum fruits than to (high-toxin) S. carolinense fruits. Because fungi varied in their degree of inhibition by glycoalkaloids, we found only limited support for the Directed Toxicity Hypothesis. The degree of inhibition for fungi was, on average, similar in magnitude to the negative effects upon vertebrate feeding rates, providing further support for the General Toxicity Hypothesis. Additionally, we found no evidence for differences in antifungal effects between the two glycoalkaloids and no evidence of synergism between them. Nutrient effects were generally small, with little evidence that nutrient variation could compensate for the antifungal activity produced by levels of glycoalkaloids found in ripe S. carolinense fruits. Fruits that are presumably more profitable (S. americanum) are preferred by vertebrate frugivores and contain much lower levels of glycoalkaloid. These results support the Removal Rate Hypothesis. We conclude that glycoalkaloids of Solanum fruits may act as effective antifungal defenses. However, because they are also toxic to seed-dispersing vertebrates, their effects upon fruit–disperser interactions are probably general and important.