Climate change leads to phenology shifts of many species. However, not all species shift in parallel, which can desynchronize interspecific interactions. Within trophic cascades, herbivores can be top–down controlled by predators or bottom–up controlled by host plant quality and host symbionts, such as plant-associated micro-organisms. Synchronization of trophic levels is required to prevent insect herbivore (pest) outbreaks. In a common garden experiment, we simulated an earlier arrival time (~2 weeks) of the aphid Rhopalosiphum padi on its host grass Lolium perenne by enhancing the aphid abundance during the colonization period. L. perenne was either uninfected or infected with the endophytic fungus Epichloë festucae var. lolii. The plant symbiotic fungus produces insect deterring alkaloids within the host grass. Throughout the season, we tested the effects of enhanced aphid abundance in spring on aphid predators (top–down) and grass–endophyte (bottom–up) responses. Higher aphid population sizes earlier in the season lead to overall higher aphid abundances, as predator occurrence was independent of aphid abundances on the pots. Nonetheless, after predator occurrence, aphids were controlled within 2 weeks on all pots. Possible bottom–up control of aphids by increased endophyte concentrations occurred time delayed after high herbivore abundances. Endophyte-derived alkaloid concentrations were not significantly affected by enhanced aphid abundance but increased throughout the season. We conclude that phenology shifts in an herbivorous species can desynchronize predator–prey and plant–microorganism interactions and might enhance the probability of pest outbreaks with climate change.

Plants have developed a variety of defence strategies against herbivores. One possible strategy is the induced production of metabolites following herbivore attack. Plant-associated micro-organisms can be the source of such defensive compounds. For example, cool-season grasses can be associated with systemic endophytic fungi of the genus Epichloё, which produce herbivore-toxic alkaloids. In a controlled common garden approach, we tested the hypothesis that different types of herbivory induce endophyte growth and increase the endophyte-mediated production of three bioactive alkaloids which can deter or toxify herbivores. During 18 weeks, we analysed biweekly endophyte and alkaloid concentrations in the grass Lolium perenne infected with the endophytic fungus Epichloё festucae var. lolii. The experiment was conducted throughout the field season and compared three different herbivore treatments to the control treatment (herbivory exclosure). We showed that the concentration of the vertebrate toxic alkaloid lolitrem B increased following clipping (a simulation of grazing herbivores), while the insect deterring alkaloid peramine increased following locust herbivory (biting–chewing herbivores). The endophyte concentration increased slightly following clipping (P = 0·09). Sap sucking aphids altered neither endophyte nor alkaloid concentrations. Our study provides evidence for an herbivore-specific induction of endophyte-mediated responses following herbivore attack on its host grass. Our results suggest that the grass–endophyte symbiosis involves a close chemical crosstalk between the interacting partners.