Marine and terrestrial herbivores display convergent chemical ecology despite 400 million years of independent evolution

Abstract

Chemical cues regulate key ecological interactions in marine and terrestrial ecosystems. They are particularly important in terrestrial plant-herbivore interactions, where they mediate both herbivore foraging and plant defense. Although well described for terrestrial interactions, the identity and ecological importance of herbivore foraging cues in marine ecosystems remain unknown. Here we show that the specialist gastropod Elysia tuca hunts its seaweed prey, Halimeda incrassata, by tracking 4-hydroxybenzoic acid to find vegetative prey and the defensive metabolite halimedatetraacetate to find reproductive prey. Foraging cues were predicted to be polar compounds but instead were nonpolar secondary metabolites similar to those used by specialist terrestrial insects. Tracking halimedatetraacetate enables Elysia to increase in abundance by 12- to 18-fold on reproductive Halimeda, despite reproduction in Halimeda being rare and lasting for only ∼36 h. Elysia swarm to reproductive Halimeda where they consume the alga's gametes, which are resource rich but are chemically defended from most consumers. Elysia sequester functional chloroplasts and halimedatetraacetate from Halimeda to become photosynthetic and chemically defended. Feeding by Elysia suppresses the growth of vegetative Halimeda by ∼50%. Halimeda responds by dropping branches occupied by Elysia, apparently to prevent fungal infection associated with Elysia feeding. Elysia is remarkably similar to some terrestrial insects, not only in its hunting strategy, but also its feeding method, defense tactics, and effects on prey behavior and performance. Such striking parallels indicate that specialist herbivores in marine and terrestrial systems can evolve convergent ecological strategies despite 400 million years of independent evolution in vastly different habitats.

Keywords: chemical cue; defense; eavesdropping; herbivory; prey tracking.

Fig. 1.

Elysia host preference. Number of trials in which an Elysia colonized one of 14 common seaweeds and seagrasses (n = 20) (A), three co-occurring seaweeds in the genus Halimeda (n = 20) (B and C), or a cotton ball laced with H. incrassata-conditioned seawater vs. seawater only (n = 40) (D), when offered in a still water arena (A, B, and D) or in the field (C). Choice was assessed after 2 h (A–C) or within a 5-min period (D). Results were analyzed by a Cochran’s Q (A–C) or Fisher’s exact (D) test. In A–C, different letters above bars indicate significant differences among seaweeds in terms of Elysia colonization frequency, as determined by Wilcoxon sign tests (corrected for multiple comparisons). AL, A. longicaulis; CC, Caulerpa cupressoides; CP, Caulerpa prolifera; CS, Caulerpa sertularioides; DC, Dictyosphaeria cavernosa; HI, H. incrassata; HM, H. monile; HO, H. opuntia; PC, Penicillus capitatus; PD, Penicillus dumetosus; RP, Rhipocephalus phoenix; SF, S. filiforme; TT, T. testudinum; US, Udotea sp.

Fig. 2.

Chemical attractants produced by H. incrassata. (A) Elysia hunts its seaweed prey by tracking H. incrassata cues. Shown are the structures of 4-HBA, which Elysia uses to locate vegetative H. incrassata, and HTA, which Elysia uses to locate reproductive H. incrassata during rare and ephemeral spawning events. Isolation of 4-HBA and HTA are reported in Figs. S1 and S2, respectively. (B and C) Mean (+ SE) number of Elysia/100 g seaweed that colonized H. incrassata in nature 24 h after seaweeds were enriched with 4-HBA or with a solvent control (n = 20) (B) or with extracts from vegetative or reproductive H. incrassata (n = 20) (C). Each result was analyzed by a Mann–Whitney test.

Fig. S1.

Isolation of attractant cues produced by H. incrassata. Number of trials in which an Elysia first colonized a cotton ball coated with a natural concentration of an extract fraction from H. incrassata (black bars) or a solvent control (gray bars) when the two were offered together in a still water arena (n = 15 trials per fraction). Choice was assessed within a 5-min period. Asterisks indicate a significant difference in the frequency with which a fraction and its control were colonized (*P < 0.05, **P < 0.01, ***P < 0.001) according to a Fisher’s exact test. Brackets indicate the path of bioassay-guided fractionation in successive panels. (A) liquid-liquid partition; (B) normal-phase silica chromatography; (C) reversed-phase C₁₈ silica chromatography; (D and E) preparative thin-layer silica chromatography. See SI Materials and Methods for separation techniques. Bioassays led to the isolation of the attractant 4-HBA (see Fig. 2A for structure).

Fig. S2.

Isolation of attractant cues produced by reproductive H. incrassata. Number of trials in which an Elysia first colonized a cotton ball coated with a natural concentration of an extract fraction from reproductive H. incrassata (black bars) or a solvent control (gray bars) when the two were offered together in a still water arena (n = 10 trials per fraction, except for B+C-3+4-D+E-4, where n = 20). Choice was assessed within a 5-min period. Statistical analysis was performed as in Fig. S1. Brackets indicate the path of bioassay-guided fractionation in successive panels. (A) liquid-liquid partition; (B) reversed-phase HP20ss chromatography; (C) normal-phase silica chromatography; (D) reversed-phase C₁₈ silica HPLC; (E and F) preparative thin-layer silica chromatography; (G) reversed-phase C₁₈ silica HPLC. See SI Materials and Methods for separation techniques. Bioassays led to the isolation of the attractant HTA (see Fig. 2A for structure).

Fig. S3.

Isolation of Elysia antipredator chemical defenses. Number of trials in which the wrasse T. bifasciatum rejected a pellet containing a natural concentration of an extract fraction from Elysia (black bars) following ingestion of a solvent control pellet (gray bar) (n = 20 trials per fraction). Statistical analysis was performed as in Fig. S1 (*P < 0.05, **P < 0.01, ***P < 0.001). Bioassays led to the isolation of the predator deterrent HTA (fraction B-3), which is produced by H. incrassata, used by Elysia as a cue to find reproductive seaweeds during their brief period of spawning, and sequestered by Elysia.

Fig. 3.

Effects of Elysia grazing on H. incrassata. Mean (+ SE) number of new segments produced (per eight marked branches) (A) and marked branches lost (per individual seaweed) (B) by H. incrassata after 3 d, when occupied by a natural density of Elysia (black bars, n = 29) or no Elysia (gray bars, n = 19). (C) Number of H. incrassata that lost branches that were occupied (black bar) or unoccupied (gray bar) by Elysia for 5 d (n = 15). Seaweeds that lost their empty enclosure also lost their occupied enclosure. (D) Experimental setup. Results were analyzed by a Welch’s t test (A), a Mann–Whitney test (B), or McNemar’s test (C).

Fig. 4.

Effect of fungal injection on H. incrassata tissue loss. Mean (+ SE) number of segments lost from H. incrassata above the site of inoculation when injected with a fungus isolated from the radula of Elysia (A) or with the common marine fungal pathogen L. thalassiae (B). Fungal effects (black bars) were assessed after 8 d, relative to media (gray bars) or needle-puncture (white bars) controls that were applied to other branches of the same seaweed (n = 20). Data from each experiment were analyzed by a Friedman’s test. Letters above bars indicate significant differences among treatments within an experiment, via Friedman’s post hoc tests.