A Physiological and Behavioral Mechanism for Leaf Herbivore-Induced Systemic Root Resistance

Abstract

Indirect plant-mediated interactions between herbivores are important drivers of community composition in terrestrial ecosystems. Among the most striking examples are the strong indirect interactions between spatially separated leaf- and root-feeding insects sharing a host plant. Although leaf feeders generally reduce the performance of root herbivores, little is known about the underlying systemic changes in root physiology and the associated behavioral responses of the root feeders. We investigated the consequences of maize (Zea mays) leaf infestation by Spodoptera littoralis caterpillars for the root-feeding larvae of the beetle Diabrotica virgifera virgifera, a major pest of maize. D. virgifera strongly avoided leaf-infested plants by recognizing systemic changes in soluble root components. The avoidance response occurred within 12 h and was induced by real and mimicked herbivory, but not wounding alone. Roots of leaf-infested plants showed altered patterns in soluble free and soluble conjugated phenolic acids. Biochemical inhibition and genetic manipulation of phenolic acid biosynthesis led to a complete disappearance of the avoidance response of D. virgifera. Furthermore, bioactivity-guided fractionation revealed a direct link between the avoidance response of D. virgifera and changes in soluble conjugated phenolic acids in the roots of leaf-attacked plants. Our study provides a physiological mechanism for a behavioral pattern that explains the negative effect of leaf attack on a root-feeding insect. Furthermore, it opens up the possibility to control D. virgifera in the field by genetically mimicking leaf herbivore-induced changes in root phenylpropanoid patterns.

Figure 1.

The root herbivore D. virgifera specifically avoids leaf-infested plants by recognizing systemic changes in soluble root components. A, Preference of D. virgifera larvae for roots of control versus leaf-infested plants in a two-arm belowground choice system (n = 15). B, Preference for roots of control versus leaf-infested plants with washed root systems in a petri dish setup (n = 19). C, Preference patterns after different durations of leaf infestation (n = 12). D, Preference for roots of damaged plants with and without defense elicitation by application of S. littoralis regurgitant (n = 12). E, Preference time course using a single leaf-induction event (n = 12–14). F, Preference for roots of control and leaf-infested plants with and without direct contact between the rhizosphere and phyllosphere (n = 12). G, Preference for root extracts of control and leaf-infested plants using agarose cubes (n = 15). Preference is expressed as percent choice corresponding to the proportions of independent replicates in which a given preference was observed (no choice: A–F, <10%; G, 21%). Asterisks indicate significant differences between treatments (*, P < 0.05; **, P < 0.01; and ***, P < 0.001).

Figure 2.

Leaf infestation alters soluble free and conjugated phenolic acids in the roots. Average concentrations of different phenolic acids in control roots (gray bars) and roots of leaf-infested plants (purple bars) are shown for crown and primary roots (±se). Shading indicates a significant overall treatment effect determined by ANOVA (P < 0.05). Asterisks indicate significant pairwise differences between treatments within root types (Holm-Sidak post hoc tests: *, P < 0.05; **, P < 0.01; and ***, P < 0.001). COMT, Caffeic acid O-methyl transferase; PAL, phenylalanine lyase; C, control; S.l., S. littoralis.

Figure 3.

Manipulating the biosynthesis of phenolic acids through cinnamate 4-hydroxylase (C4H) inhibition leads to the disappearance of the D. virgifera avoidance response toward leaf-infested plants. A, Preference for roots of buffer-treated and C4H-inhibited control and S. littoralis-infested plants (n = 12). B, Preference for roots of buffer-treated and C4H-inhibited control and artificially induced plants (n = 12). C, Preference for roots of buffer-treated, C4H-inhibited, and CA-complemented control and artificially induced plants. C4H was inhibited by application of the selective inhibitor piperonylic acid (PA; n = 23). Preference is expressed as percent choice corresponding to the proportions of independent replicates in which a given preference was observed (no choice, <10%). Asterisks indicate significant differences between treatments (*, P < 0.05; **, P < 0.01; and ***, P < 0.001).

Figure 4.

Genetic modification of the phenylpropanoid pathway leads to the disappearance of the D. virgifera avoidance response toward leaf-infested plants. A, Preference of D. virgifera for roots of control and leaf-infested wild-type (WT) plants, cinnamyl alcohol dehydrogenase (bm1) mutant plants, and caffeic acid O-methyl transferase (bm3) mutant plants (n = 12–13). B, Preference of D. virgifera for roots of control and artificially induced wild-type, bm1, and bm3 plants (n = 12). Preference is expressed as percent choice corresponding to the proportions of independent replicates in which a given preference was observed (no choice, <10%). Asterisks indicate significant differences between treatments (*, P < 0.05; **, P < 0.01; and ***, P < 0.001).

Figure 5.

Bioactivity-guided fractionation links changes in conjugated phenolic acids with D. virgifera preference patterns. A, Preference of D. virgifera for fractions of root extracts of control and leaf-infested plants (n = 15). The polarity gradient of the fractionation setup is shown in purple. B, Concentrations of free phenolic acids in root extracts of control and leaf-induced plants across the polarity gradient. Note that analyzed fractions cover the range of two fractions of experiment A. C, Concentrations of soluble, acid hydrolyzable phenolic acids. D, Concentrations of soluble, basic hydrolyzable phenolic acids. Shaded areas correspond to the bioactive fraction. S.l., S. littoralis; W., wounding; Reg., regurgitant.