Jasmonic acid and its precursor 12-oxophytodienoic acid control different aspects of constitutive and induced herbivore defenses in tomato

Abstract

The jasmonate family of growth regulators includes the isoleucine (Ile) conjugate of jasmonic acid (JA-Ile) and its biosynthetic precursor 12-oxophytodienoic acid (OPDA) as signaling molecules. To assess the relative contribution of JA/JA-Ile and OPDA to insect resistance in tomato (Solanum lycopersicum), we silenced the expression of OPDA reductase3 (OPR3) by RNA interference (RNAi). Consistent with a block in the biosynthetic pathway downstream of OPDA, OPR3-RNAi plants contained wild-type levels of OPDA but failed to accumulate JA or JA-Ile after wounding. JA/JA-Ile deficiency in OPR3-RNAi plants resulted in reduced trichome formation and impaired monoterpene and sesquiterpene production. The loss of these JA/JA-Ile -dependent defense traits rendered them more attractive to the specialist herbivore Manduca sexta with respect to feeding and oviposition. Oviposition preference resulted from reduced levels of repellant monoterpenes and sesquiterpenes. Feeding preference, on the other hand, was caused by increased production of cis-3-hexenal acting as a feeding stimulant for M. sexta larvae in OPR3-RNAi plants. Despite impaired constitutive defenses and increased palatability of OPR3-RNAi leaves, larval development was indistinguishable on OPR3-RNAi and wild-type plants, and was much delayed compared with development on the jasmonic acid-insensitive1 (jai1) mutant. Apparently, signaling through JAI1, the tomato ortholog of the ubiquitin ligase CORONATINE INSENSITIVE1 in Arabidopsis (Arabidopsis thaliana), is required for defense, whereas the conversion of OPDA to JA/JA-Ile is not. Comparing the signaling activities of OPDA and JA/JA-Ile, we found that OPDA can substitute for JA/JA-Ile in the local induction of defense gene expression, but the production of JA/JA-Ile is required for a systemic response.

Figure 1.

Jasmonate levels in wild-type and OPR3-RNAi plants. A, Jasmonates in wild-type (WT, dark blue) and OPR3-RNAi leaf tissue (RNAi, green) 40 min after wounding (W, filled bars) compared with unwounded controls (C, hatched bars). JA, JA-Ile, and 12-OH-JA were quantified by liquid chromatography-tandem mass spectrometry after solid-phase extraction of methanolic extracts. Jasmonate levels are given in nanomoles per gram of fresh weight (FW) as the mean ± sd of six biological replicates for wild-type plants. For OPR3-RNAi plants, three biological replicates were performed on each of three independent transgenic lines. B, OPDA, JA, and 12-OH-JA were quantified by GC-MS in wild-type and OPR3-RNAi plants 2 h after wounding (color scheme as in A). The experiment involved eight and four biological replicates on independent RNAi lines for wounded plants and unwounded controls, respectively. Asterisks indicate significant differences between OPR3-RNAi and wild-type plants (*P < 0.05, **P < 0.01, and ***P < 0.001).

Figure 2.

Bioassays assessing the relative contribution of OPDA and JA/JA-Ile to insect resistance. A, Dual-choice tests revealing a preference of M. sexta larvae for OPR3-RNAi and jai1-1 mutants over wild-type plants. Leaf discs from either OPR3-RNAi (three independent transgenic lines) or jai1-1 mutant plants and the corresponding wild-type controls (WT₁, cv UC82B; and WT₂, cv Castlemart) were offered to third-instar M. sexta larvae. The leaf area consumed within 4 h of feeding is indicated in percentages ± sd (n = 33, P < 0.001). B, No-choice tests assessing palatability of OPR3-RNAi and jai1-1 mutants compared with the wild type. A single forth-instar M. sexta larva was offered 500 mg of leaf discs from either OPR3-RNAi plants, jai1-1 mutants, or the respective controls. The leaf mass that was consumed within 30 min is indicated in percentages ± sd (n = 100 for OPR3-RNAi [four independent transgenic lines] and cv UC82B; n = 55 for jai1-1 and cv Castlemart). In A and B, the wild type is shown in dark blue, OPR3-RNAi in green, and jai1-1 in yellow. Both experiments were performed with leaf material from healthy plants (hatched bars) and from plants that were wounded 24 h before the experiment (filled bars). Asterisks indicate significant differences (Wilcoxon rank sum test; **P < 0.01 and ***P < 0.001). n.s., Not significant.

Figure 3.

M. sexta oviposition preference for OPR3-RNAi plants as a result of impaired terpene production. To assess oviposition preference, tomato wild-type and OPR3-RNAi plants were offered to a pair of mating M. sexta in a two-channel olfactometer. A choice for the wild type (WT) is indicated by blue bars, a choice for OPR3-RNAi in green, and no choice in gray. When no further odorants were added, the preference for RNAi plants was statistically significant (n = 35, χ² = 48.8, ***P = 0.0001). In a wild-type/wild-type comparison, the addition of cis-3-hexenal on one side of the olfactometer had no influence on oviposition preference (n = 16, χ² = 0.87, P = 0.64). By contrast, the addition of a mixture of terpenes reflecting the terpene content of wild-type trichomes in the OPR3-RNAi/OPR3-RNAi comparison had a significant deterrent effect (n = 21, χ² = 22.6, ***P < 0.0001). n.s., Not significant.

Figure 4.

Effect of OPR3 silencing on nutritional quality, trichome density, and volatile production. A, Nutritional quality. Total leaf nitrogen, bound carbon, and carbohydrate content (complete [total] sugars, reducing sugars, and starch) were analyzed in the wild type (black bars; n = 4 for total carbon and nitrogen, n = 15 for carbohydrates), and in OPR3-RNAi plants (gray bars; n = 12 for total carbon and nitrogen, n = 15 for carbohydrates) in milligrams per gram of fresh weight (FW). B, Trichome density. Trichome density is given as the number of type VI trichomes per square centimeter of leaf area for the wild type (black bar; n = 26) and three independent OPR3-RNAi lines (gray bars; n = 15, 15, and 5 for lines J55 [left], P3 [center], and J18 [right bar], respectively). C, Trichome volatiles. GC-MS analyses identified significant differences for 10 compounds identified as cis-3-hexenal, five monoterpenes (α-pinene, 2-carene, limonene, and α- and β-phellandrene), three sesquiterpenes (α-humulene, δ-elemene, and β-caryophyllene), and one unknown (mass spectrum in Supplemental Fig. S4) that were quantified in nanograms per microliter of trichome extract (1 µL corresponding to 2 mg of leaf material) in two technical replicates performed on each of two independent RNAi lines. Data in A to C show the mean ± sd for wild type in black and OPR3-RNAi in gray. The identification of δ-elemene may be an artifact because it may have formed from germacrene C in the injector during gas chromatography (Quintana et al., 2003). Asterisks in A and B indicate significant differences (*P < 0.05 and ***P < 0.001). FW, fresh weight.

Figure 5.

cis-3-Hexenal acts as a feeding stimulant for M. sexta larvae. A, Dual-choice tests were performed with an artificial diet to which extracts from wild-type (WT; black bars) or OPR3-RNAi trichomes (gray bars) were added (WT versus RNAi, left). To account for the reduced trichome density of OPR3-RNAi plants, the experiment was repeated with the amount of wild-type extract reduced to one-third (1/3 WT versus RNAi, center). In the last comparison, wild-type extract was complemented by the addition of cis-3-hexenal to match the concentration observed in OPR3-RNAi trichomes (WT+hex versus RNAi, right). B, Dual-choice tests were performed with an artificial diet to which synthetic compounds (a terpene blend or cis-3-hexenal) were added in concentrations reflecting the composition of the wild type (black bars) or OPR3-RNAi trichomes (gray bars). The consumed diet in A and B is given in percentages as the mean of 33 experiments for each genotype and treatment ± se. Asterisks indicate statistically significant differences (Wilcoxon rank sum test; ***P < 0.001). n.s., Not significant.

Figure 6.

Although the development of M. sexta larvae is indistinguishable on OPR3-RNAi and wild-type (WT) plants, it is much faster on jai1 mutants. A, Eighty 3-d-old larvae were placed on each of the two genotypes to be compared (wild-type cv UC82B, dark blue; and OPR3-RNAi, green). Larvae were weighed collectively at age 3 d and individually at age 11, 13, 15, 16, and 17 d. On the last day, the larvae were photographed (bottom) and the experiment was terminated because they were about to enter the wandering stage for pupation. B, In the same way, the development of 75 larvae was compared on the wild type (cv Castlemart, blue) and the jai1-1 mutant (yellow). The weight of the larvae was determined at age 7, 11, 12, 13, and 14 d. This experiment had to be terminated 3 d earlier, because larvae feeding on jai1-1 had already entered the wandering stage at age 14 d. Data in A and B show the mean weight of the larvae ± sd. Asterisks indicate significant differences (Student’s t test; ***P < 0.001).

Figure 7.

OPDA is sufficient for local defense gene induction but not for systemic wound signaling. The expression of the PI-II wound-response marker gene was analyzed by RT-PCR 8 h after wounding (+) in both the wounded leaf (loc) as well as systemic unwounded leaves (sys) of wild-type (WT) and OPR3-RNAi (RNAi) plants. The corresponding leaves of unwounded plants (−) were analyzed as controls. The same experiment was also performed with grafted plants. The different graft combinations are indicated as fractions, with the genotype of the scion above, and that of the root stock below the fraction line. One experiment (−,+) is shown for the control grafts (WT/WT and RNAi/RNAi), whereas one control and three wounded plants with independent RNAi lines as root stock or scion were analyzed in the case of the informative grafts (RNAi/WT and WT/RNAi; −,+++). RT-PCR amplification of EF-1α is shown as a control for RNA integrity and cDNA synthesis. PCR products were analyzed on 1.5% agarose gels stained with ethidium bromide.

Figure 8.

Conversion to JA is not required for wound-response gene induction by OPDA. The expression of the PI-II wound-response marker gene was analyzed in wild-type (WT) and OPR3-RNAi plants treated with linolenic acid (LA; 300 nmol/leaf), OPDA (6 nmol/leaf), JA (6 nmol/leaf), or the buffer control (C; 1% Tween 20 in 15 mm potassium phosphate buffer, pH 7.5). The compounds were applied as 5-µL droplets onto the surface of two leaves of 4-week-old plants. After 6 h, total RNA was isolated and used for RT-PCR expression analysis of PI-II and EF-1α, as a control for RNA integrity and cDNA synthesis. PCR products were analyzed on a 1.5% agarose gel stained with ethidium bromide.