Epichloë endophytes alter inducible indirect defences in host grasses

Abstract

Epichloë endophytes are common symbionts living asymptomatically in pooid grasses and may provide chemical defences against herbivorous insects. While the mechanisms underlying these fungal defences have been well studied, it remains unknown whether endophyte presence affects the host's own defences. We addressed this issue by examining variation in the impact of Epichloë on constitutive and herbivore-induced emissions of volatile organic compounds (VOC), a well-known indirect plant defence, between two grass species, Schedonorus phoenix (ex. Festuca arundinacea; tall fescue) and Festuca pratensis (meadow fescue). We found that feeding by a generalist aphid species, Rhopalosiphum padi, induced VOC emissions by uninfected plants of both grass species but to varying extents, while mechanical wounding failed to do so in both species after one day of damage. Interestingly, regardless of damage treatment, Epichloë uncinata-infected F. pratensis emitted significantly lower quantities of VOCs than their uninfected counterparts. In contrast, Epichloë coenophiala-infected S. phoenix did not differ from their uninfected counterparts in constitutive VOC emissions but tended to increase VOC emissions under intense aphid feeding. A multivariate analysis showed that endophyte status imposed stronger differences in VOC profiles of F. pratensis than damage treatment, while the reverse was true for S. phoenix. Additionally, both endophytes inhibited R. padi population growth as measured by aphid dry biomass, with the inhibition appearing greater in E. uncinata-infected F. pratensis. Our results suggest, not only that Epichloë endophytes may play important roles in mediating host VOC responses to herbivory, but also that the magnitude and direction of such responses may vary with the identity of the Epichloë-grass symbiosis. Whether Epichloë-mediated host VOC responses will eventually translate into effects on higher trophic levels merits future investigation.

Figure 1. Total VOC emissions (mean ±1SE) from naturally endophyte free (E-; circles) vs infected (E+; triangles) tall fescue (a) and meadow fescue (b) in response to aphid or mechanical damage.

C: untreated control; W: mechanical wounding; A: aphid infestation. For comparison, manipulatively endophyte free tall fescue (ME-; squares) was included. All damage treatments were initiated at the same time, then VOC collections conducted at 1 day after mechanical wounding, or at 6 and 12 days after aphid addition. Statistical details are shown in tables 1 and 2.

Figure 2. Individual VOCs of tall fescue whose release changed significantly in response to endophyte infection and 12 days of aphid feeding.

E-: naturally endophyte free; E+: naturally endophyte infected; ME-: manipulatively endophyte free; C: untreated control; A: aphid damage. Statistical details are shown table S1.

Figure 3. Individual VOCs of meadow fescue which changed significantly in response to endophyte infection and 12 days of aphid feeding.

E-: naturally endophyte free; E+: naturally endophyte infected; C: untreated control; A: aphid damage. Statistical details are shown table S5.

Figure 4. PLS-DA plots based on comparisons among VOC blends emitted by differently treated plants at 12 days after aphid addition.

E-: naturally endophyte free; E+: naturally endophyte infected; ME-: manipulatively endophyte free; C: control; A: aphid feeding. For tall fescue (upper panel), a clear separation was seen between control and infested plants, which clustered mainly on the right and left side of the plot, respectively, whereas plants of different endophyte status within each damage treatment largely overlapped. For meadow fescue (lower panel), E- and E+ plants grouped separately while overlapping somewhat, with the strongest separation occurring between infested E- plants and any other treatment. Statistical details concerning compounds responsible for the clustering are given in tables S1 and S5.