Phytohormone-based activity mapping of insect herbivore-produced elicitors

Abstract

In response to insect attack, many plants exhibit dynamic biochemical changes, resulting in the induced production of direct and indirect defenses. Elicitors present in herbivore oral secretions are believed to positively regulate many inducible plant defenses; however, little is known about the specificity of elicitor recognition in plants. To investigate the phylogenic distribution of elicitor activity, we tested representatives from three different elicitor classes on the time course of defense-related phytohormone production, including ethylene (E), jasmonic acid (JA), and salicylic acid, in a range of plant species spanning angiosperm diversity. All families examined responded to at least one elicitor class with significant increases in E and JA production within 1 to 2 h after treatment, yet elicitation activity among species was highly idiosyncratic. The fatty-acid amino acid conjugate volicitin exhibited the widest range of phytohormone and volatile inducing activity, which spanned maize (Zea mays), soybean (Glycine max), and eggplant (Solanum melongena). In contrast, the activity of inceptin-related peptides, originally described in cowpea (Vigna unguiculata), was limited even within the Fabaceae. Similarly, caeliferin A16:0, a disulfooxy fatty acid from grasshoppers, was the only elicitor with demonstrable activity in Arabidopsis thaliana. Although precise mechanisms remain unknown, the unpredictable nature of elicitor activity between plant species supports the existence of specific receptor-ligand interactions mediating recognition. Despite the lack of an ideal plant model for studying the action of numerous elicitors, E and JA exist as highly conserved and readily quantifiable markers for future discoveries in this field.

Plants defensively respond to insect herbivory both directly by inducing biochemical changes that impede pest growth and indirectly by promoting advantageous interactions with beneficial organisms through the release of volatile organic compounds (VOCs) (1–3). Specific biochemicals isolated from insect oral secretions (OS), saliva, and oviposition fluids may either amplify or suppress the expression of these herbivore-specific and wound-inducible plant responses (4–7). In addition to VOC production, insect-produced elicitors have broad activities on transcriptional regulation and non-volatile plant defenses including trypsin proteinase inhibitors and isoflavone phytoalexins (8–10). The past decade has included the discovery of multiple classes of insect-produced elicitors; however, general conclusions regarding modes of action cannot be drawn (5, 10–14).

Elicitor identification has relied largely upon the sequential chemical fractionation of insect fluids monitored by plant bioassays for induced physical or biochemical changes (4, 5). For example, a maize (Zea mays)-induced VOC production bioassay facilitated the purification of the fatty-acid amino acid conjugate (FAC) N-(17-hydroxylinolenoyl)-L-glutamine, termed volicitin, from beet armyworm (Spodoptera exigua) larvae OS (4). Related glutamine- and glutamate-containing FACs with elicitor activity have been subsequently described in numerous insects (15, 16). Unlike maize, lima bean (Phaseolus lunatus) and cotton (Gossypium hirsutum) do not exhibit rapid volicitin-induced VOCs, suggesting that FACs are not “general elicitors” of induced plant responses (17). To understand the lack of FAC activity in legumes, the phytohormone ethylene (E) was used as an inducible marker in cowpea (Vigna unguiculata) to guide the chemical fractionation of Spodoptera frugiperda larval OS and resulted in the isolation of chloroplastic ATP synthase γ-subunit-derived peptide elicitors termed inceptins (14, 18). More recently, a maize VOC bioassay was again used to monitor the chemical isolation of grasshopper (Schistocerca americana) OS elicitors and resulted in the identification of a complex mixture of disulfo-oxy fatty acids, termed caeliferins (11). Collectively, these results indicate that biochemical mechanisms mediating insect-herbivore recognition are diverse. Although information on (i) multiple classes of elicitors, (ii) specific model systems, and (iii) targeted biochemical plant responses now exists, to date there has been little progress in understanding patterns of herbivore-produced elicitor activity across plant families or conserved signaling mechanisms indicative of recognition.

The phytohormones most often associated with mediating plant responses to insects are jasmonic acid (JA), E, and salicylic acid (SA) (19–21). Jasmonates are key regulators of plant responses to damage, necrotrophic pathogens, and insect herbivory, whereas E modulates the magnitude of direct and indirect plant defenses produced in response to jasmonates (16, 22–28). SA is involved in systemic acquired resistance to biotrophic pathogens; however, its role in herbivore-induced responses is less clear (24, 29). Through signaling interactions with JA, SA has been shown to antagonize, not influence, or even synergize JA-mediated responses (21, 30, 31). In maize and tobacco (Nicotiana attenuata), both JA and E mediate FAC signaling, numerous transcriptional changes, and increased VOC emission (16, 20, 32, 33). Cowpea responds to inceptin through the induced production of JA, E, SA, and VOCs in damaged leaves; however, the phytohormone responses of plants to caeliferins have not been previously demonstrated (11, 14).

To create a broader view of elicitor action we examined the activity of FACs, inceptin, and caeliferin A16:0 in time course experiments for E, JA, SA, and subsequent VOC emission in cowpea (Fabaceae, Vigna unguiculata, cv California 5), soybean (Fabaceae, Glycine max, Williams 82), Arabidopsis thaliana (Brassicaceae, ecotype Landsberg erecta, Ler), eggplant (Solanaceae, Solanum melongena, Black beauty), and maize (Poaceae, Zea mays B73). Of the four plant families examined, each one exhibited induced production of E and JA in response to at least one elicitor class. We conclude that the elicitation of E and JA represent conserved responses that are highly idiosyncratic among species. This result supports receptor-mediated elicitor signaling and indicates that recognition systems for a single insect pest, within even related crops, cannot be inferred from plant phylogenies alone.

Results

Legumes Respond to Different Classes of Insect Produced Elicitors.

In cowpea, significant production of E, JA, and SA occurred within 2 h of treatment with inceptin but no other tested elicitor (Fig. 1 A, F, and K). In contrast, soybean produced no response to inceptin; however, volicitin strongly increased production of both E and JA within 2 h and this activity was significantly greater than that of N-linolenoyl-Gln (Fig. 1 B and G). Soybean SA levels did not significantly differ (Fig. 1L). To further examine the apparent selectivity of volicitin action in soybean, we tested three additional FACs, including N-linolenoyl-Glu, identified in Manduca sexta (15, 16). Soybean consistently exhibited significant induced E and JA responses to volicitin compared with all other treatments [supporting information (SI) Fig. S1]. These results demonstrate that related legumes exhibit specificity for different elicitors.

Fig. 1.

Activity of elicitor classes on an induced phytohormone production time course demonstrates widespread recognition. Plant treatments included undamaged controls (gray), damage plus buffer (black), volicitin (light blue), N-linolenoyl-Gln (dark blue), caeliferin A16:0 (green), and inceptin (red) with average (n = 4, ±SEM) (A-E) E (nl g⁻¹ hr⁻¹), (F-J) leaf JA (ng g⁻¹ FW) and (K-O) leaf SA (ng g⁻¹ FW) production in cowpea, soybean, A. thaliana Ler, eggplant, and maize B73, sequentially and respectively. At the time point of greatest mean change, different letters (a-c) represent significant differences with ANOVA P values <0.001 (Tukey test corrections for multiple comparisons; P < 0.05). Not statistically different (nsd) indicates ANOVA P values >0.05.

A. thaliana Selectively Responds to Caeliferin A16:0.

In the Brassicaceous model A. thaliana Ler, caeliferin A16:0 strongly induced both E and JA production within 2 h and a modest increase in SA after 4 h of application to damaged leaves (Fig. 1 C, H, and M). Neither FACs nor inceptin induced alterations in E, JA, or SA levels compared with damage alone (Fig. 1 C, H, and M). In addition to Ler, caeliferin A16:0 also induced E production in ecotype Columbia (Col-O) at 1 h (Fig. S2).

FAC Elicitors Are Differentially Active Within the Solanaceae.

As a member of the diverse genus Solanum (34), eggplant exhibited significant FAC-induced E and JA production within 2 h after damage but no response to inceptin or caeliferin A16:0 (Fig. 1 D and I). Eggplant SA levels did not significantly differ (Fig. 1N). In contrast to soybean, N-linolenoyl-Gln induced significantly more E in eggplant than volicitin at 2 h (Fig. 1 B and D); however, based on JA levels, these individual FAC activities were indistinguishable (Fig. 1I). Encouraged by the FAC-induced responses in eggplant, elicitation activity in tomato (Solanum lycopersicum) was examined. Damage significantly induced both E and JA production to a greater extent than in untreated controls within 2 h; however, none of the three elicitor classes increased E, JA, or SA, demonstrating elicitation differences in related species (Fig. S3).

Maize Displays Intraspecific Differences in Response to Elicitors.

FACs elicited rapid increases in E and JA production at 1 h in maize B73 (Fig. 1 E and J). Both inceptin and caeliferin A16:0 demonstrated weak and statistically insignificant activity on JA and E levels (Fig. 1 E and J). Maize B73 SA levels did not significantly differ (Fig. 1O). Surprisingly, caeliferin A16:0-induced VOC emission was also statistically intermediate between FAC and damage treatments (Fig. 2D). Caeliferin A16:0 has established VOC-inducing activity in “Delprim”; however, this maize hybrid is no longer commercially available (11). As an alternative, we examined the maize inbred line CML322, which has been crossed with B73 to generate a recombinant inbred population for genetic mapping (35). In CML322, both volicitin and caeliferin A16:0 resulted in significant increases in JA at 1 h and VOC emission at 5 h (Fig. S4). In contrast to volicitin, caeliferin A16:0 failed to induce E even in this otherwise responsive line (Fig. S4). To examine if FAC activity is idiosyncratic within a species, we tested volicitin-induced responses in five inbred lines. Consistent with conserved FAC recognition, all lines demonstrated significant 1.9- to 2.3-fold JA increases in response to volicitin; however, the subsequent induction of volatiles varied from 14.7-fold increases to no statistical difference (Fig. S5). Similar variation in the magnitude of Spodoptera OS-induced VOC emission in diverse maize cultivars and defense signaling in different tobacco populations has been previously described (36, 37).

Fig. 2.

Elicitor induced VOC emission. Average (n = 4, ±SEM) (A) cowpea, (B) soybean, (C) eggplant, and (D) maize B73 VOC emission (nl g⁻¹ 30 min⁻¹) 5 h following treatments including undamaged controls (gray), damage plus buffer (black), volicitin (light blue), N-linolenoyl-Gln (dark blue), caeliferin A16:0 (green), and inceptin (red). Using standard methods, induced VOC production in A. thaliana at 5 h was undetectable (data not shown). Different letters (a-d) represent significant differences. (All ANOVA P values <0.01 with Tukey test corrections for multiple comparisons; P < 0.05.)

Elicitor-Induced Phytohormone Production Is Associated with VOC Emission.

In cowpea, soybean, eggplant, and maize, elicitor classes that significantly promoted JA/E increases also resulted in induced VOC emission 5 h later (Fig. 2 A−D and Fig. S4). Curiously, in soybean, volicitin- and N-linolenoyl-Gln-induced VOC emission were not significantly different from each other (Fig. 2B). This result was unexpected as volicitin induced significantly greater levels of JA than N-linolenoyl-Gln (Fig. 1G). Although not statistically significant, on a mean basis, N-linolenoyl-Gln treatment resulted in a 2.1-fold increase in JA levels versus damage alone (Fig. 1G). Even small elicitor-induced changes in JA/E may be sufficient to promote VOC emission. Conversely, despite caeliferin A16:0 induction of both JA and E production in A. thaliana leaves, no evidence for induced VOC emission was detected at 5 h (data not shown).

Discussion

Empirical approaches are essential in determining how plants recognize insect herbivore attack. For example, cowpea and soybean both belong to the Fabaceae millettiod clade; however, based on E, JA, and VOC production, these crops recognize different elicitor classes present in larval Spodoptera pests (12, 14, 38). Curiously, lima bean and soybean have been used as interchangeable experimental systems to address legume recognition of Spodoptera feeding, elicitation, and FAC activity (39, 40). In soybean suspension cultures, volicitin proved less active than N-linolenoyl-Gln in triggering cytosolic Ca⁺² influx (39). Moreover, few differences could be found in the activity between D- and L-amino acid-derived FACs or even the surfactant SDS, leading to the conclusion that the effect was “linked to the overall physico-chemical properties of the amphiphilic compounds” (39). In contrast, L-glutamine-derived volicitin has been shown to be essential for induced maize VOC production and membrane binding activity (4, 41). Compared with leaf tissue, DNA microarray analyses in A. thaliana cell cultures demonstrate significantly greater numbers of constitutively down-regulated genes (42). To reconcile these findings, we hypothesize that select signaling pathways may be suppressed in cell cultures compared with intact plants. Clearly, plant species, cultivar, and growth conditions all need to be carefully examined before making even limited conclusions regarding elicitor recognition.

As evidence for recognition specificity in A. thaliana, herbivory by Pieris rapae larvae resulted in differential gene expression compared with mechanical damage alone (43). Consistent with elicitation, damaged leaves treated with P. rapae OS induced multiple defense-related transcripts; however, herbivory comparisons between the specialist P. rapae and the generalist Spodoptera littoralis resulted in surprisingly few differences. This effect was independent of dietary linolenic acid and led to the conclusion that FACs are not responsible for the induction of insect-inducible genes (44). We confirm this result by demonstrating that FACs are not active in A. thaliana when compared with the grasshopper elicitor caeliferin A16:0. Although little is known regarding A. thaliana-grasshopper interactions, herbivory on Brassica oleracea by the locust (Schistocerca gregaria) resulted in a 1,000-fold induction of lipoxygenase BoLOX transcript levels compared with a 100-fold damage-induced increase (45).

Consistent with established Solanaceous models, eggplant responds to FACs with significant increases in E, JA, and induced VOC emission (16, 20). In tobacco microarrays, FACs accounted for 55% of transcriptional regulation compared with crude OS of M. sexta, emphasizing both the importance of FACs and also other yet unidentified bioactive compounds (9). In contrast to eggplant, our search for phytohormone induction in tomato failed to detect rapid elicitation responses different from damage alone, thus emphasizing the inability to predict insect recognition mechanisms within plant genera. We hypothesize that damage-induced endogenous peptide elicitors, including systemin and hydroxyproline-rich systemins, may exist as the predominant means of herbivore-induced signal amplification in some members of this family (46, 47).

In the current study, elicitor-induced production of JA and E was closely associated with induced VOC emission. Indeed, multiple physiological and molecular approaches have demonstrated that both JA and E are important regulators of herbivore-induced VOCs (22, 25, 26). However, with the exception of E, we were unable to find evidence for induced VOC emission in A. thaliana treated with caeliferin A16:0. Although A. thaliana has been successfully used in numerous plant-insect and VOC biosynthesis studies, it is generally recognized that A. thaliana flowers and foliage emit exceedingly low VOC levels (48, 49). Given the slow kinetics (20–30 h) and low reported rates of VOC emission (3–5 ng g⁻¹ h⁻¹), it is possible that small effects of caeliferin A16:0 on induced VOC production may have been missed (48).

We draw three primary conclusions from this work. First, the activity of individual elicitors on closely related plant species is highly idiosyncratic; thus, no single model arthropod or plant system is ideal for studying the action of multiple elicitor classes (Fig. 3). Given the array of herbivore-plant interactions, we envision a large number of yet undiscovered elicitors and suppressors that will mediate plant responses (1). Moreover, the diversity of plant recognition systems and evolved manipulation by specific insect herbivores may eventually prove analogous to the complexity demonstrated in plant-pathogen interactions (50, 51). Second, JA and E exist as robust and highly conserved markers useful in the bioassay and discovery of additional elicitors. The induction time courses and specific chemistry of VOC and non-volatile defenses will vary among species; however, examples of rapid JA and E elicitation were found in all plant families examined (Figs. 1 and 3). As both mediators and markers for plant defense and disease susceptibility, phytohormones are firmly established in an array of plant-pathogen interactions and, increasingly, plant-insect interactions (13, 52). Additionally, changes in other phytohormone classes or suppression of basal levels may also serve as useful markers for suppressors of host responses. Third, FACs, and specifically volicitin, represent the most broadly active elicitors examined in these studies (Fig. 3). Potent activity in soybean, eggplant, and maize leads us to speculate that plant FAC recognition is an ancient evolutionarily conserved trait that can be reasonably expected in many additional plant families. Likewise, the inactivity of FACs in A. thaliana, cowpea, and tomato argues against relatively non-specific pore-formation mechanisms and supports plant recognition mediated by specific binding proteins (39, 41). ln contrast to FACs, inceptin represents a seemingly uncommon divergence from direct plant recognition of an insect-synthesized elicitor to an indirect recognition system based on insect catabolism (14, 18, 53).

Fig. 3.

Phytohormone responses to elicitors are highly conserved; however, activity is idiosyncratically distributed across angiosperm diversity. In order of activity frequency encountered, (i) FACs (blue) elicit JA and E in maize, eggplant, and soybean; (ii) caeliferin A16:0 (green) elicits JA and E in A. thaliana and JA in maize CML322; and (iii) inceptin (red) elicits JA, E, and SA in cowpea. In tomato, JA and E levels increased in response to damage alone with no additional elicitation detected.

In summary, we interpret the phylogenically idiosyncratic nature of elicitor action as additional indirect evidence for the existence of volicitin and inceptin receptors in maize and cowpea, respectively, and although little is known about caeliferins, modern tools for genetic mapping, microarray, and mutant analyses will expedite future discoveries (11, 18, 35, 44).

Materials and Methods

Plants.

Cowpea (Vigna unguiculata var. California Blackeye no. 5; The Wax Company), soybean (Glycine max var. Williams 82; ARS-GRIN accession PI518671), Arabidopis thaliana Columbia (Col-O) and Landsberg erecta (Ler), eggplant (Solanum melongena var. Black Beauty; W. Atlee Burpee), tomato (Solanum lycopersicum var. Castlemart II), and maize (Zea mays landrace inbreds B73, CML322, MS71, Ky21, NC358; source information at http://www.panzea.org/lit/germplasm.html) were germinated in MetroMix 200 (Sun Gro Horticulture Distribution) supplemented with 14-14-14 Osmocote (Miracle-Gro; Scotts). A. thaliana and maize B73 were grown under controlled conditions as reported elsewhere (54, 55). All other plants were maintained in a greenhouse with a 12-h photoperiod, minimum of 300 μmol m⁻² s⁻¹ of photosynthetically active radiation supplied by supplemental lighting, 70% relative humidity, and temperature cycle of 24 °C/28 °C (night/day).

Chemicals.

Volicitin, N-linolenoyl-Gln, caeliferin A16:0, and inceptin anticipated from cowpea/soybean [⁺ICDINGVCVDA⁻], A. thaliana [⁺ICDINGTCVDA⁻], eggplant [⁺ICDINGNCVDA⁻], and maize [⁺ICDVNGVCVDA⁻] were synthesized as previously described (11, 12, 14). Based on previous quantification of insect OS and optimized plant dose responses, we used 50 mM Na₂HPO₄ buffer (pH 8) stock solutions of FACs, caeliferin A16:0, and inceptin at pmol μl⁻¹concentrations of 100, 500, and 1, respectively.

Plant Bioassays.

Trials used 14- and 35-day-old plants for cowpea/maize and eggplant/tomato/soybean/A. thaliana, respectively. For all induction assays, the adaxial sides of new fully expanded leaves were scratched with a razor in three areas, removing 3%–5% of the waxy cuticle. The damage sites included the central leaf tip spanning both sides of the mid-rib and two mid-basal sections on opposite sides of the midrib. Elicitor test solutions of 5 μL or buffer alone were immediately applied and dispersed over the damage sites. Because of the small size, three leaves per plant were used for each A. thaliana treatment. Leaves remained on the intact plants for 1, 2, and 4 h before excision, sampling, and analyses. A separate set of plants was identically treated with leaves excised after 5 h to estimate short-term VOC emission.

Phytohormone and VOC Analyses.

For GC analysis of E, experimental leaves were excised throughout the designated time course and sealed in either 6- or 13-mL tubes for 60 min before head-space sampling as described (14). To quantify JA and SA, a 4-cm² section of leaf surrounding the treated site was weighed (50–100 mg), frozen in liquid N₂, processed, and analyzed by GC-MS as described elsewhere (55). Collection, quantification, and confirmation of leaf VOC emission followed established protocols (54). Five hours after treatments, individual leaves were excised from intact plants and placed in VOC collection chambers for 30 min. To normalize differences in leaf mass, VOC emission was calculated as ng g⁻¹ 30 min⁻¹. Specific VOCs included in estimates of total induced VOCs were as follows: cowpea, (Z)-3-hexenyl acetate, (E)-4,8-dimethyl-1,3,7-nonatriene (DMNT), methyl salicylate, and (E, E)-4,8,12-trimethyl-1,3,7,11-tridecatetraene (TMTT); soybean, (E)-β-ocimene and (E,E)-α-farnesene; eggplant, (Z)-3-hexenyl acetate, (E)-β-ocimene, linalool, DMNT, and (E,E)-α-farnesene; maize B73, (Z)-3-hexenyl acetate, (E)-β-ocimene, linalool, DMNT, indole, (E)-α-bergamotene, (E)-β-farnesene, and TMTT; maize CML322, (E)-β-caryophyllene, (E)-α-bergamotene, and (E)-β-farnesene.

Data Analysis.

At single time points containing the largest mean differences, ANOVAs were performed on amounts of E, JA, SA, and total induced VOCs. Significant treatment effects were investigated when the main effects of the ANOVAs were significant (P < 0.05). Where appropriate, Tukey tests were used to correct for multiple comparisons between control and treatment groups. Before statistical analysis, all data were subjected to square root transformation to compensate for elevated variation associated with larger mean values. The analysis was accomplished with JMP 4.0 statistical discovery software (SAS Institute). Construction of the simplified Angiosperm phylogenetic tree used the program Phylomatic (56). Specific families used in this work, including Fabaceae, Poaceae, Solanaceae, and Brassicaceae, were entered along with the Genus and species whereas all others were pruned from the tree.