Coevolutionary transitions from antagonism to mutualism explained by the Co-Opted Antagonist Hypothesis

Abstract

There is now good evidence that many mutualisms evolved from antagonism; why or how, however, remains unclear. We advance the Co-Opted Antagonist (COA) Hypothesis as a general mechanism explaining evolutionary transitions from antagonism to mutualism. COA involves an eco-coevolutionary process whereby natural selection favors co-option of an antagonist to perform a beneficial function and the interacting species coevolve a suite of phenotypic traits that drive the interaction from antagonism to mutualism. To evaluate the COA hypothesis, we present a generalized eco-coevolutionary framework of evolutionary transitions from antagonism to mutualism and develop a data-based, fully ecologically-parameterized model of a small community in which a lepidopteran insect pollinates some of its larval host plant species. More generally, our theory helps to reconcile several major challenges concerning the mechanisms of mutualism evolution, such as how mutualisms evolve without extremely tight host fidelity (vertical transmission) and how ecological context influences evolutionary outcomes, and vice-versa.

Fig. 1. Study system involving the hawkmoth Manduca sexta and its associated plants in southern Arizona, USA.

M. sexta nectar-feeds (orange arrows) and oviposits (green arrows) at D. wrightii and D. discolor. The outcome of these interactions could be affected by M. sexta interactions with the alternative nectar source, A. palmeri, or the alternative larval host plant, P. parviflora. Paintings by Julie Johnson (Life Science Studios).

Fig. 2. Coevolution from antagonism to mutualism in one-plant species communities.

Panels (a, b) show interaction outcomes as a function of attraction (v_i) and defense (h_max − h_i), where h_max is the maximum herbivory rate (Methods). Parameter space regions depict ecological outcomes and arrows highlight coevolutionary effects. The ancestral insect persists as a pure antagonist above the dashed gray line (Eq. 1) and is extinct below. Co-option of the antagonist as a pollinator and the evolution of pollination benefits (b_i > 0) expands the green mutualistic region by moving the interaction breakdown boundary (Eq. 3; solid black line) away from the interaction transition boundary (Eq. 2; dotted black line), as depicted by the green arrows (see Supplementary Movies 1 and 2). The insect persists as a net antagonist despite being co-opted as a pollinator within the gray regions and goes extinct within the white regions. Simultaneous to the evolution of pollination benefits, coevolution of attraction and defense drives the transition to net mutualism, as depicted by the orange arrows pointing from the ancestral coESSs (white points) to the new coESSs (black points). Empirical estimates of the coESSs are not included here (as in Fig. 3) due to data limitations. Panels (c, d) plot the equilibrium densities of each plant species (solid green lines) and insect larvae per plant of each Datura species (dashed green lines) over evolutionary time, τ. Panels (e, f) plot the coevolutionary dynamics of pollination benefits (b_i; black lines), attraction (v_i; blue lines), and defense (h_i; purple lines).

Fig. 3. Coevolution from antagonism to mutualism in the two-plant species community.

Panels (a, b) show interaction outcomes as a function of attraction (v_i) and defense (h_max – h_i), where h_max is the maximum herbivory rate. The ancestral insect persists as a pure antagonist above the dashed gray line in panel a and is extinct below. There is no dashed gray line in panel b because the ancestral insect can persist on the ancestral D. wrightii alone. The insect persists as a net antagonist within the gray regions and is extinct within the white regions. Plant evolution of pollination benefits (b_i > 0) expands the green mutualistic regions by separating the interaction breakdown boundary (solid black line) and the interaction transition boundary (dotted black line), as depicted by green arrows. Simultaneously, coevolution of attraction and defense drives the transition to net mutualism, as depicted by orange arrows from ancestral coESS (white points) to the new coESS (black points). Blue points give empirical estimates of the coESS (panel (a): h_w = 1 ± 0.4; v_w = 4.3 ± 0.6; panel (b): h_d = 2 ± 0.8; v_d = 2.3 ± 0.4), where the crossbars show variation in leaf consumption and the standard error of floral visitation (n = 89 plants for D. wrightii; n = 33 plants for D. discolor) (Methods). While the plant species coevolve in the model (see Supplementary Movie 3), panels (a, b) are plotted with the other plant species held at its final coESS for clarity. Panels (c, d) plot the equilibrium densities of each Datura species (solid green lines) and insect larvae per plant of each Datura species (dashed green lines) over evolutionary time, τ. Panels (e, f) plot the coevolutionary dynamics of pollination benefits (b_i; black lines), attraction (v_i; blue lines) and defense (h_i; purple lines).

Fig. 4. Coevolutionary transitions in more generalized trophic interactions.

Interaction outcomes in the presence of an alternative larval host plant (panels a, b) or an alternative nectar source (panels c, d) are plotted as a function of attraction (v_i) and defense (h_max – h_i), where h_max is the maximum herbivory rate. The ancestral insect persists as a pure antagonist above the dashed gray lines and is extinct below. The insect persists as a net antagonist within the gray regions and is extinct within the white regions. Plant evolution of pollination benefits (b_i > 0) expands the green mutualistic regions by moving the interaction breakdown boundary (solid black line) away from the interaction transition boundary (dotted black line), as depicted by green arrows (see Supplementary Movies 4 and 5). Simultaneously, coevolution of attraction and defense drives the transition to net mutualism, as depicted by orange arrows from ancestral coESSs (white points) to new coESSs (black points). Empirical estimates of the coESSs are not included due to data limitations. Panels (e–h) plot the equilibrium densities of each Datura species (solid green lines) and insect larvae per plant of each Datura species (dashed green lines) over evolutionary time, τ. Panels (i-l) plot the coevolutionary dynamics of pollination benefits (b_i; black lines), attraction (v_i; blue lines), and defense (h_i; purple lines).

Fig. 5. Varying the costs associated with plant and insect traits reveals coevolutionary outcomes.

Insect costs associated with herbivory (c^H_I,i) and visitation (c^V_I,i) are varied in panels (a, b) and plant costs associated with defense (c^H_P,i) and attraction (c^V_P,i) are varied in panels (c, d) (Methods). Green regions indicate evolutionary transitions from antagonism to net mutualism. Gray regions indicate that the interaction is net antagonistic despite evolutionary co-option of the antagonist. White regions indicate evolutionary purging of the antagonist. Black points give evolutionary parameter values used in the model (Methods).