<em>Epichloë</em> endophytes impair performance and preference of <em>Rhopalosiphum padi</em> by disrupting aphid feeding behaviour in tall fescue

Mutualistic symbioses between plants and fungi are widespread in nature and are increasingly being adopted in agroecosystems to enhance economic and environmental sustainability. Temperate forage grasses often associate with Epichloë endophytic fungi that, by hindering herbivore performance, confer protection against insect damage, aphids in particular. However, despite the well-known impacts of endophytes on aphid preference and performance, there is scant information on how such effects are caused by alterations in feeding behaviour. The main objective here was to explore interactions between a grass plant, its fungal endophyte, and aphids, focusing on using the Electrical Penetration Graph technique to understand in detail the various aspects of aphid feeding behavior. Experiments were conducted in a greenhouse with tall fescue cultivars INIA Aurora and INIA Fortuna, grown with (+) or without (-) the AR584-Epichloë strain of Epichloë coenophiala, and exposed to bird-cherry oat aphids, Rhopalosiphum padi. Aphid performance was evaluated in antibiosis tests (no choice), at both population and individual levels, and preference in antixenosis (choice) tests. Aphid feeding behaviour was studied using the Electrical Penetration Graph technique in both seedlings and adult plants. Daily fecundity and aphid population growth were both lower on endophytic plants, indicating a preference for endophyte-free plants. Aphids exhibited impaired feeding behaviour on endophytic plants, with shorter phloem activities. Similar patterns were observed in seedlings. These results demonstrate that endophytes negatively affect aphid phloem sap ingestion, explaining the subsequent insecticidal and antifeedant effects, and that these can explain the ulterior insecticidal and antifeedant consequences. Methods All data from both the no-choice (antibiosis) and choice (antixenosis) experiments, including aphid population and individual responses, were analysed using mixed model ANOVA (PROC MIXED, SAS 2009). Normality assumptions were assessed using the Shapiro-Wilk test statistic (PROC UNIVARIATE, SAS 2009). The models examined the main effects of cultivar (four populations and five individuals), with individual seedlings or plants treated as random factors. Pairwise comparisons between cultivars were conducted using least-squares means (LS MEANS statement), with means separated using an adjusted Tukey method for multiple comparisons (α = 0.05). Mean differences were compared using the least-squares mean difference (LSD) from an adjusted Fisher’s protected LSD method for multiple comparisons (α = 0.05). Additionally, a Pearson rank correlation coefficient test was used to analyse the relationship between plant dry mass and plant height, specifically in plants inoculated with aphids or those that remained aphid-free during the aphid population experiment. Results from EPG measurements did not exhibit a normal distribution (Table 3, variables 1–8, 10–11, 13–22). Therefore, Mann–Whitney U rank sum tests were employed to assess differences in aphid feeding behaviour between plants with different endophytic status. Fisher’s exact test was used to evaluate disparities in the proportions of individuals engaging in specific types of activity between plants with different endophytic status (Table 3, variables 9, 12, 15, 23, and 24). Significance levels were adjusted for multiple testing using the false discovery rate (FDR) correction method at α = 0.05 (Benjamini and Hochberg 1995). A comprehensive description of the FDR procedure is provided in Benjamini et al. (2001). All statistical analyses for these experiments were conducted using InfoStat Professional v2011p software (Di Rienzo et al. 2013).