Results
The analysis of variance for larval weight evolution during 26 days showed that all sources of variation [random (first-order autoregressive covariance structure of within-subject correlation among weights recorded on the same lava; σ2 = 0.7539 ± 0.0163) and fixed effects] were significant (p < 0.05) except treatments, but the interaction treatment x genotype was significant suggesting that maize genotypes would differentially respond to feeding treatments. At the end of the bioassay, larvae reared on stems from control susceptible inbred plants weighed less than larvae fed on the resistant ones (Fig. 1). Resistant plants pre-conditioned by 48 hours of feeding by MCB larvae decreased significantly larval weight compared to its control, while average weight of larvae fed with susceptible plants pre-conditioned by 48 hours feeding did not differ from its corresponding control larval weight (Fig. 1 and 2). On the other hand, susceptible plants exposed to feeding by larvae for a longer period (nine days) had opposite effects on larval weight to resistant plants and long-term attack to the resistant genotype caused stronger effect on the weight of larvae than short-term attack. In addition, induced responses by long-term attack made that the mortality of larvae fed on the resistant plants significantly higher than that of larvae fed on the susceptible ones although no differences for mortality between both inbreds were detected at control conditions (Table 1).
Attending to the metabolomics approach, 4362 different features were detected among the different samples (Supplementary table S1). 194 and 192 ions were selected as important features (based on VIP-score values of the PLS-DA analysis) in the short and long-term responses, respectively, of the susceptible inbred line; 108 of them being important in both responses (Fig. 3). On the other hand, 101 ions out of the 188 and 227 detected as relevant in the short and long-term responses of the resistant inbred, respectively, were common in both responses. Although 194 and 188 ions had important contributions to the 48h responses to S. nonagrioides in the susceptible and resistant inbreds, respectively, only 57 of them contributed to both responses. Ninety ions were selected as important features based on VIP-score values of the PLS-DA analysis for the responses to 9-day stem borer attack in both inbreds. Finally, 30 ions were involved in the four responses. Identification and characterization of the metabolites from which the ions detected could be derived have been done with the ions with relevant effects on the long term response, either detected across genotypes or genotype-specific (Table S2).