3 RESULTS
3.1 Effects of multiple antagonists and attack order on plant gene transcripts
Transcription of plant genes associated with hormone biosynthesis across four pathways: (i) salicylic acid (ICS1 ), (ii) jasmonic acid (LOX2 ), (iii) abscisic acid (AO3) , and (iv) gibberellic acid (GA2ox ), were induced by PEMV when S. lineatus was not present (Fig. 2A-D, Table S1, Pillai = 0.942, P =0.002). However, there was no induction of any of these biosynthesis genes in response to PEMV when S. lineatus was present, indicating that S. lineatus inhibited plant defense against PEMV (Table S1, A x W interaction, Pillai = 1.509, P = 0.021, Fig. 2). In both the presence and absence of PEMV, S. lineatusinduced transcription of LOX2 , but S. lineatus did not directly modify the expression of ICS1, AO3 or GA2ox (Fig. 2B-D, Table S1). When PEMV-infectious A. pisum attacked following S. lineatus , there was greater induction of LOX2 compared to when S. lineatus attacked alone (Table S1, A x W interaction, Pillai = 1.509, P = 0.021, Fig. 2). In contrast to the antagonism exerted by S. lineatus on the response to PEMV, this represents enhancement of plant defense when PEMV infection followed attack byS. lineatus .
All three defense response transcripts (PR1, DDR230,PsLectin ) were induced by PEMV when S. lineatus was not present (Fig. 3); similarly, each transcript was induced by S. lineatus when PEMV was not present (Fig. 3, Table S1; A ✕ W interaction, F = 2.64, P = 0.111). When S. lineatusattacked second, the expression level of PR1 and Lectindid not change compared to when weevils were absent. The effects of PEMV on the transcripts was modified by the presence of S. lineatusand attack order. While DDR230 was induced by PEMV (Table S1,F = 47.181, P < 0.001), this effect diminished when S. lineatus was present after PEMV (Fig. 3B). Similarly, the effects of PEMV on PR1 were inhibited when S. lineatusattacked second (Fig. 3), whereas that the induction of lectin by PEMV was not altered by S. lineatus in either order (Fig. 3).
3.2 Effects of multiple antagonists and attack order on plant phytohormones
We observed variation in phytohormones in response to A. pisum(Table S2, Pillai = 0.95, P < 0.001) andS. lineatus (Table S2, Pillai = 1.195, P < 0.001). PEMV-infectious A. pisum strongly induced salicylic acid (Table S2, F = 254.2, P < 0.001), but this was inhibited when S. lineatus attacked after PEMV (Fig. 4A, Tukey HSD). PEMV did not affect jasmonic acid (Table S2, F = 0.97, P = 0.34), but the order of S. lineatus did (Table S2, F = 5.30, P = 0.018). Both S. lineatus (Table S2, F = 4.10, P = 0.037) and infectiousA. pisum induced abscisic acid (Table S2, F = 9.96, P = 0.006) and this effect was contingent on the attack order (Table S2, A ✕ W, F = 4.32, P = 0.032, Fig. 4, Tukey HSD). Jasmonic acid levels were suppressed by S. lineatus when attacking prior to non-infectious sham A. pisum , but not on plants already attacked by PEMV (Fig. 4B, Tukey HSD).
3.3 Effects of multiple antagonists and attack order on plant nutrients
Feeding by S. lineatus increased the total amino acid levels (GLM, χ2 = 9.19, P = 0.01, Fig 5), but PEMV-infectious A. pisum did not (GLM, χ2 = 0.044, P = 0.83), and this effect was not modified depending on attack order (GLM, A ✕ W interaction, χ2 = 0.24, P = 0.63). Non-metric multidimensional scaling (NMDS) analysis of amino acid composition also showed that changes to amino acid availability was most different among treatments for alanine, arginine, lysine, and glycine (Ordination plot, Fig S1).