•Synthetic chemical elicitors of plant defense have been touted as a powerful means for sustainable crop protection. Yet, they have never been successfully applied to control insect pests in the field.
•We developed a high-throughput chemical genetics screening system based on a herbivore-induced linalool synthase promoter fused to a β-glucuronidase (GUS) reporter construct to test synthetic compounds for their potential to induce rice defenses.
•We identified 2,4-dichlorophenoxyacetic acid (2,4-D), an auxin homolog and widely used herbicide in monocotyledonous crops, as a potent elicitor of rice defenses. Low doses of 2,4-D induced a strong defensive reaction upstream of the jasmonic acid and ethylene pathways, resulting in a marked increase in trypsin proteinase inhibitor activity and volatile production. Induced plants were more resistant to the striped stem borer Chilo suppressalis, but became highly attractive to the brown planthopper Nilaparvata lugens and its main egg parasitoid Anagrus nilaparvatae. In a field experiment, 2,4-D application turned rice plants into living traps for N. lugens by attracting parasitoids.
•Our findings demonstrate the potential of auxin homologs as defensive signals and show the potential of the herbicide to turn rice into a selective catch crop for an economically important pest.
Chemical genetics offers an efficient and powerful approach to discover synthetic elicitors, and is widely applied in human drug therapy and drug development (Walsh, 2007; Gendron et al., 2008; Done et al., 2010). Given the labor-intensive and time-consuming assays that are needed to directly screen for inducers of herbivore resistance, chemical genetics may represent a useful alternative in this context.
Rice, one of the most important food crops in the world, suffers heavily from insect pests (Cheng & He, 1996), making it an important target for novel pest control strategies. Previous studies with rice have shown that herbivore attack induces a variety of plant hormones, including JA, SA and ET, which subsequently regulate defensive responses, including the release of volatile organic compounds (VOCs) (Lou et al., 2005a,b, 2006; Lu et al., 2006; Zhou et al., 2009). Some of these herbivore-induced VOCs, for example linalool, are strongly attractive to Anagrus nilaparvatae, an egg parasitoid of the rice brown planthopper (BPH) Nilaparvata lugens (Lou, 1999), one of the most important rice pests in Asia (Heong & Hardy, 2009). Exogenous application of JA can partially mimic the defensive reaction of rice (Lou et al., 2005a; Zhou et al., 2009) and can enhance the parasitism of N. lugens eggs by A. nilaparvatae in the glasshouse and the field (Lou et al., 2005b). However, as JA is expensive to produce and may have negative effects on plant growth (Parthier, 1991), its applicability as a plant strengthener remains limited.
We developed a specific high-throughput screening system to discover novel synthetic elicitors with potential for application. Using a 1.55-kb promoter region of the S-linalool synthase gene (OsLIS; Yuan et al., 2008), fused to a β-glucuronidase (GUS) reporter gene, we screened over 100 chemicals for their potential to induce the expression of the reporter gene. With this approach, we identified 2,4-dichlorophenoxyacetic acid (2,4-D) as a defense elicitor. By combining chemical and transcriptional analysis, reverse genetics and behavioral assays in the field, we unravel the mechanism by which 2,4-D acts as a defense elicitor and discuss how it may be used to control N. lugens in the field.
Materials and Methods
For the experiments, the following rice (Oryza sativa L.) genotypes were used: Xiushui 11 wild-type (WT), as-acs (Lu et al., 2011) and as-lox (Zhou et al., 2009). The LISp::GUS reporter lines were created as described below. For all lines, pregerminated seeds were cultured in plastic bottles (diameter, 8 cm; height, 10 cm) in a glasshouse (28 ± 2°C, 14 h light : 10 h dark). Ten-day-old seedlings were transferred to 20-l communal hydroponic boxes with rice nutrient solution (Yoshida et al., 1976). After 30–35 d, seedlings were transferred to individual 500-ml hydroponic plastic pots, each pot with one or two plants. Plants were used for experiments 4–5 d after transplanting.
Colonies of the rice striped stem borer (SSB) Chilo suppressalis (Walker) and Nilaparvata lugens (Stål) (BPH) were originally obtained from rice fields in Hangzhou, China, and maintained on Xiushui 11 rice seedlings in a controlled climate room at 26 ± 2°C, 12 h light phase and 80% relative humidity. Anagrus nilaparvatae Pang et Wang colonies were also obtained from rice fields in Hangzhou, China, and maintained with BPH eggs on Xiushui 11 rice plants.
Generation and characterization of the LISp::GUS reporter lines
To obtain the promoter region of the linalool synthase OsLIS (TIGR ID: Os02g02930), a 1.55-kb genomic DNA fragment upstream of the gene was amplified by PCR using the following primer pair: F1551, 5′-CGAAGCTTCGTGTTCATGTACCCTTTT-3′; R1551, 5′-AGGTCGACATATCAAGCTCAAGCGAGT-3′ (the bold letters represent digest sites of enzymes Hind III (forward primer) and Sal I (reverse primer); the italic letters represent the primers corresponding to the sequence of OsLIS). The PCR product (1551 bp) was inserted into the pCAMBIA-1391 vector to fuse it with GUS, yielding a transformation vector. The vector was inserted into the rice variety Xiushui 11 using Agrobacterium tumefaciens-mediated transformation, as described by Hiei et al. (1994). Homozygous T2 plants were selected using GUS staining or hygromycin resistance screening (Zhou et al., 2009). The number of insertions was determined by Southern hybridization of genomic DNA using a PCR fragment of the GUS gene as a probe. For this, genomic DNA was extracted from rice leaves using a cetyl trimethyl ammonium bromide procedure (Ausubel et al., 1987). DNA (20 μg) was digested using various restriction enzymes, size fractionated in a 0.8% w/v agarose gel, transferred to a nylon membrane (Hybond N+, Millipore) and cross-linked with UV. Southern hybridization was achieved using the DIG High Prime DNA Labeling and Detection Starter Kit II (Roche), according to the manufacturer’s procedure. The DNA template for the GUS probe was obtained by PCR using primers GUS-F (5′-GCAACTGGACAAGGCACT-3′) and GUS-R (5′-GCGTCGCAGAACATTACA-3′). Two T2 homozygous lines (L15-15, L15-38), each containing a single insertion of LISp::GUS (Supporting Information Fig. S1a), were then selected and used for subsequent experiments.
Quantitative real-time PCR (QRT-PCR)
All QRT-PCR experiments were carried out using five independent biological replicates. Total RNA from each sample was isolated using the SV Total RNA Isolation System (Promega), following the manufacturer’s instructions. One microgram of each total RNA sample was reverse transcribed using the PrimeScript™ RT-PCR Kit (TaKaRa, Otsu, Japan). The primers and probes used for mRNA detection of target genes by QRT-PCR are shown in Table S1. The QRT-PCR assay was performed on a CFX96™ Real-Time system (Bio-Rad) using a Premix Ex Taq™ Kit (TaKaRa). A rice actin gene OsACT (TIGR ID: Os03g50885) was used as an internal standard to normalize cDNA concentrations.
Quantitative GUS activity assay
GUS activities in transgenic plants were analyzed according to the method described in Jefferson et al. (1987) with some modifications. For quantitative analysis, plant tissues were homogenized in a GUS assay buffer (50 mM potassium phosphate, 10 mM EDTA, 0.1% Triton X-100, 0.1% Sarcosyl), and an aliquot of the supernatant was incubated at 37°C for 30 min after 4-methylumbelliferyl-β-d-glucuronide (4-MUG) was added as the substrate. The amount of 4-methylumbelliferone (4-MU) formed by the GUS reaction was determined using a DTX880 Multimode Detector (Beckman, Fullerton, CA, USA). Protein concentrations were determined using the method described by Bradford (1976). Each treatment at each time interval was replicated five times.
JA treatment Two methods were used to induce plants with JA. First, plants were individually sprayed with 2 ml of JA (100 μg ml−1) in 50 mM sodium phosphate buffer (titrated with 1 M citric acid to pH 8, with 0.01% Tween-20). Control plants were sprayed with 2 ml of the buffer (BUF group). Second, plant roots were treated with JA, which was added to the nutrient solution at a concentration of 5 mg l−1, for 24 h. Nonmanipulated plants were used as controls (C).
2,4-D treatment To expose plants to 2,4-D, individual rice seedlings were grown in a nutrient solution (pH 4.8), and 2,4-D (minimum purity > 99%, Sigma), at concentrations ranging from 0.5 to 50 mg l−1, was added (see Figs 1–8). Nonmanipulated plants grown in nutrient solution without 2,4-D were used as controls (C).
SSB treatment To expose plants to SSB, they were individually infested with a third-instar larva of SSB that had been starved for 2 h before the experiments. Control plants (C) did not receive any caterpillars.
BPH treatment Individual plants were infested with 15 gravid BPH females that were confined in a glass cylinder (diameter, 4 cm; height, 8 cm; with 48 small holes (diameter, 0.8 mm)), and the top of the cylinder was covered with a piece of sponge. One empty cylinder was attached to control plants (noninfested).
Behavior and performance experiments
SSB performance measurement To measure the effect of 2,4-D on SSB, 5-d-old, preweighed larvae were placed on stems of plants whose roots were either treated or not with 2,4-D. One larva per plant was used, and 30 replicates were carried out per treatment. Larval mass was measured 6 d after the start of the experiment. The relative rate of increase of larval mass was calculated according to the following equation: (larval mass at the end – larval mass at the beginning)/larval mass at the beginning × 100%.
BPH preference measurement To determine the colonization and oviposition preferences of BPH, two plants (one control plant vs one plant whose roots were treated with 2,4-D) were confined within a glass cylinder, into which 15 gravid adult BPH females were introduced. The number of BPH on each plant was counted 1, 2, 4, 8, 24 and 48 h after the release of BPH, and, after 72 h, BPH were removed and the eggs on each plant were counted. The experiment was repeated six to eight times.
Responses of A. nilaparvatae females to rice volatiles were measured in a Y-tube olfactometer, which was similar to that described in Lou et al. (2005a). Plants were randomly assigned to the following treatments: control (C); BPH; 2,4-D; and BPH + 2,4-D. The behavioral response of parasitoids exposed to the following pairs of odor sources was then determined: control plants (C) vs plants whose roots were treated with 2,4-D at a dose of 2 mg l−1 for 48 h (2,4-D); plants infested by BPH for 48 h (BPH) vs plants infested by BPH and treated with 2,4-D at a dose of 2 mg l−1 for 48 h (BPH + 2,4-D). For each treatment, five replicates were carried out, and the odor sources were replaced by a new set of plants after testing 10 wasps. Thus, for each odor source combination, a total of 50 females was tested. All bioassays were conducted between 09:00 and 17:00 h, and the temperature in the room was maintained at 25–28°C.
JA and SA analysis
Plant hormones were quantified following four different treatments: control (C); 2,4-D; SSB; and 2,4-D + SSB. For the 2,4-D treatment, plant stems were harvested 1 d after the roots had been treated with 2,4-D at a concentration of 2 mg l−1. For the 2,4-D + SSB treatment, stems were harvested after treatment with 2,4-D at a concentration of 2 mg l−1 for 1 d followed by SSB infestation for 3 h. Each treatment at each time interval was replicated five times. JA and SA were analyzed by GC-MS using labeled internal standards, as described by Lou & Baldwin (2003).
To measure ET, individual control and 2,4-D-treated plants (see 2,4-D treatment in plant treatments described earlier) were placed into sealed glass cylinders (height, 50 cm; internal diameter, 4 cm). ET production was determined by taking 5 ml of headspace using a syringe from the cylinder 0, 12, 24 and 48 h after the start of the treatment. Each treatment at each time interval was replicated five times. The ET samples were analyzed by GC as described by Lu et al. (2006).
Collection, isolation and identification of volatile compounds
The collection, isolation and identification of rice volatiles were carried out using the same method as described in Lou et al. (2005a). Volatiles from plants subjected to four different treatments were collected: control (C); 2,4-D; BPH; and BPH + 2,4-D. The treatments were similar as those for the olfactometer bioassays (described above). Each treatment was replicated six times. The detected compounds were expressed as the percentage of peak areas relative to the internal standard (IS) per 8 h of trapping.
Trypsin proteinase inhibitor (TrypPI) analysis
The activity of TrypPI was measured in plants subjected to the following treatments: control (C); 2,4-D; SSB; and SSB + 2,4-D. For the SSB + 2,4-D treatment, plants were treated with 2,4-D (2 mg l−1) together with SSB third-instar larva infestation (one larva per plant). Stems (0.2–0.3 g per sample) were harvested 3 d after the start of the treatment. The TrypPI activity was measured using a radial diffusion assay as described by van Dam et al. (2001). Each treatment was replicated five times.
To evaluate the role of 2,4-D in plant–insect interactions in the field, experiments were carried out in September 2010 in a rice field in Yangzhou, China. The experimental field consisted of 12 plots (3 m × 3 m), with each plot surrounded by a buffer zone of 1 m planted with rice (Fig. S2). The 12 plots were randomly assigned to four treatments: control (C); 2,4-D (2 mg l−1; 2MG); 2,4-D (5 mg l−1; 5MG); and buffer (BUF). Each treatment consisted of three independent replicates. Plots were sprayed with 500 ml of 2,4-D in 50 mM sodium phosphate buffer (titrated with 1 M citric acid to pH 8, including 0.01% Tween; 2MG and 5MG), the buffer only (BUF) or kept nonmanipulated (C). The number of BPH nymphs, female and male adults and predatory spiders that were naturally present in the different plots were recorded 0, 6 and 10 d after the treatments. The numbers of parasitized and nonparasitized BPH eggs were counted at 0 and 10 d after the treatments. To do so, 20 plants from each block were cut off at the base and dissected under a microscope using the method described by Xiang et al. (2008). Based on these data, the relative changes in abundance of parasitized and nonparasitized BPH eggs, nymphs, adults and predatory spiders were calculated for the buffer and 2,4-D-treated plots relative to the control plots.
Differences in the levels of TrypPI, JA and SA, SSB performance and transcript levels of OsLIS in different lines were analyzed by one-way ANOVAs. In the case of significant treatment effects (P <0.05), Duncan’s multiple range tests were used to test for significant differences between the groups. Differences in experiments involving two treatments were determined by Student’s t-tests. All tests were carried out with Statistica (Statistica, SAS Institute Inc., Cary, NC, USA).
A high-throughput system for chemical elicitors of herbivore defenses
To screen new synthetic defense elicitors, we cloned a 1.55-kb promoter sequence of OsLIS (Fig. S3), fused it to a GUS reporter gene (Fig. 1a) and created two independent single-insertion reporter lines (LISp::GUS, L15-15 and L15-38, Fig. S2a). There was no obvious difference in growth between WT plants and transgenic lines (Figs 1c, S1b–d). To determine whether the LISp::GUS lines can be used to screen defense elicitors, we tested JA, which is known to efficiently induce rice plant defense responses and linalool release. As expected, the expression levels of OsLIS were increased significantly 12 and 24 h after JA treatment (Fig. 1d). Consistent with this finding, JA enhanced significantly GUS activity in the leaves (Fig. 1e). Because of the low constitutive GUS activities in the roots, JA treatment at a concentration of 5 mg l−1 for 24 h resulted in an even more pronounced response (Fig. 1b,f). These findings show that the two LISp::GUS lines can be used for high-throughput screening for chemical elicitors, and that the analysis of root GUS activity may be the most sensitive approach.
A chemical screen reveals a candidate elicitor of 2,4-D
To screen candidate elicitors, 30-d-old seedlings of the two LISp::GUS lines were incubated for 48 h with different compounds at concentrations between 1 and 50 mg l−1, followed by quantification of GUS activities in stems and roots. We found that one of these compounds, 2,4-D, a chemical analog of indole-3-acetic acid (IAA), increased significantly GUS activities in roots and stems of the two lines, especially in roots: the GUS activities in roots of 2,4-D-treated plants were two- to three-fold higher than in control plants (Fig. 1g). Quantitative dose–activity analysis revealed that 2,4-D is active at concentrations as low as 1 mg l−1, and its maximal effective concentrations were 5 mg l−1 for roots and 20 mg l−1 for stems (Fig. 1g). The application of higher concentrations of 2,4-D resulted in a decline in GUS activity in roots and stems (Fig. 1g). At a concentration of 2 mg l−1, no negative effects were observed within 20 d.
2,4-D enhances the transcript levels of three mitogen-activated protein kinase (MAPK) genes and OsWRKY53
MAPK cascades (Zhang & Klessig, 2001; Rodriguez et al., 2010) and WRKY transcription factors (Asai et al., 2002; Qiu et al., 2008) play an essential role in the mediation of plant immune responses (Wu et al., 2007; Skibbe et al., 2008). Thus, we investigated whether treatment with 2,4-D alters the expression of three MAPK genes (OsMPK3 and OsMPK6, the homologs of NaWIPK and NaSIPK, and OsMEK3, the homolog of AtMEK3) and one WRKY gene (OsWRKY53, the homolog of NaWRKY6) in plants. Transcript levels of the three MAPK genes and OsWRKY53 increased 4–8 h after treatment with 2 mg l−1 of 2,4-D, and peaked at 8–12 h (Fig. 2a–d), showing that 2,4-D regulates the transcription of MAPKs and WRKYs.
2,4-D induces JA and ET biosynthesis, but represses SA
The JA, SA and ET signaling pathways play pivotal roles in herbivore-induced signaling (Kessler et al., 2004; Bostock, 2005; Mewis et al., 2005). Therefore, we determined whether 2,4-D elicits the biosynthesis of these signaling molecules. As shown in Fig. 3(a), JA levels in stems of 2,4-D-treated plants (2 mg l−1 for 24 h) were 1.51-fold higher than those in control plants. 2,4-D treatment also increased SSB-elicited JA levels: JA levels in plants treated with 2,4-D for 1 d, followed by 3 h of SSB infestation, were 1.3-fold higher than in SSB-infested plants without 2,4-D (Fig. 3a).
A rice 13-lipoxygenase gene, OsHI-LOX, has recently been reported to be involved in herbivore-induced JA biosynthesis (Zhou et al., 2009). We thus examined the changes in transcript levels of OsHI-LOX and another JA biosynthesis-related gene, OsAOS1, a putative allene oxide synthase (AOS), in plants after treatment with 2,4-D at a concentration of 2 mg l−1. Consistent with the measured JA levels, 2,4-D treatment resulted in mRNA accumulation of the two genes, starting 4 h after treatment (left panels of Fig. 3b,c). Moreover, SSB-elicited transcript levels of the two genes were enhanced when plants had been treated with 2,4-D (right panels of Fig. 3b,c).
2,4-D treatment also elevated ET production in rice plants (Fig. 4a). The levels of ET released from plants treated with 2 mg l−1 of 2,4-D for 12, 24 and 48 h were 1.64-, 2.12- and 2.57-fold higher than those from control plants (Fig. 4a). ACC synthase (ACS), which converts S-adenosyl-l-methionine (S-AdoMet) to 1-aminocyclopropane-1-carboxylic acid (ACC), is a key enzyme for ET biosynthesis (Lin et al., 2009). In rice, OsACS2 was found to mediate herbivore-induced ET production (Lu et al., 2011). We therefore measured the changes in transcript levels of OsACS2 in plants treated with 2,4-D. In accordance with the measured ET production, the expression levels of OsACS2 were enhanced significantly after 2,4-D treatment (left panel of Fig. 4c). Treatment with 2,4-D for 24 h also enhanced SSB-elicited transcript levels of OsACS2 (right panel of Fig. 4c).
SA levels, especially in stems of SSB-elicited plants, decreased when their roots were treated with 2 mg l−1 of 2,4-D (Fig. 4b). SA levels in plants treated with 2,4-D for 24 h, followed by 3 h of SSB infestation, were only 64.5% of those in plants that were infested by SSB, but not treated with 2,4-D (Fig. 4b). NPR1 has been reported to function downstream of SA (Pieterse & Van Loon, 2004). Here, we found that 2,4-D treatment increased the mRNA levels of OsNPR1 in both SSB-infested and noninfested plants (Fig. 4d).
2,4-D enhances TrypPI levels and induces VOCs
To test the effect of 2,4-D on effective rice defenses, we measured the activity of TrypPIs, which are important defense proteins in rice against chewing herbivores such as SSB (Zhou et al., 2009), as well as plant volatiles that can attract natural enemies of BPH (Lou et al., 2005b). We found that levels of TrypPIs in rice stems were enhanced strongly after root treatment with 2,4-D at a concentration of 2 mg l−1 for 3 d (Fig. 5a). 2,4-D treatment also increased SSB-elicited TrypPI levels (Fig. 5a).
Treatment with 2,4-D also increased dramatically the release of rice volatiles: the total amount of volatiles was 138.7% that of control plants (Fig. 6, Table 1). Moreover, levels of nine individual compounds (2-heptanol, α-pinene, limonene, linalool oxide, linalool, methyl salicylate, unknown 3, farnesene and zingiberene) were significantly higher in 2,4-D-treated plants than in control plants. Treatment with 2,4-D also resulted in the release of more volatiles when plants were infested by BPH. The total amount of volatiles emitted from BPH-infested plants whose roots were treated with 2 mg l−1 of 2,4-D was 1.22-fold higher than that from plants that were infested by BPH alone. The emission of 12 chemicals was increased significantly in BPH-infested plants treated with 2,4-D (Fig. 6, Table 1).
Table 1. Comparison of volatile compounds (mean ± SE, n =6) emitted from plants treated with 2 mg l−1 of 2,4-dichlorophenoxyacetic acid (2,4-D), infested with rice brown planthopper (Nilaparvata lugens, BPH), infested with BPH and treated with 2,4-D (BPH + 2,4-D) and controls (C)
BPH + 2,4-D
Letters in the same row indicate significant differences between treatments (P <0.05, Duncan’s multiple-range test).
2.81 ± 0.44c
3.38 ± 0.40bc
4.46 ± 0.18b
6.20 ± 0.47a
2.06 ± 0.17c
2.96 ± 0.16b
4.48 ± 0.36a
5.19 ± 0.31a
1.08 ± 0.25b
1.76 ± 0.22ab
1.81 ± 0.26ab
3.66 ± 0.19a
2.30 ± 0.12c
3.14 ± 0.23b
4.54 ± 0.26a
4.88 ± 0.17a
3.32 ± 0.51b
4.10 ± 0.52ab
4.84 ± 0.91ab
5.48 ± 0.60a
5.27 ± 0.43c
9.12 ± 0.55b
10.38 ± 0.75b
16.17 ± 1.14a
1.67 ± 0.15c
2.52 ± 0.15b
5.30 ± 0.38a
5.17 ± 0.26a
5.21 ± 0.54c
9.72 ± 0.72b
10.29 ± 1.48b
13.74 ± 0.92a
3.58 ± 0.42b
4.00 ± 0.14b
7.46 ± 0.24a
6.64 ± 0.63a
2.79 ± .0.37b
4.05 ± 0.30a
4.62 ± 0.22a
4.26 ± 0.24a
2.13 ± 0.37b
3.76 ± 0.35a
4.73 ± 0.36a
4.29 ± 0.41a
0.84 ± 0.08b
1.07 ± 0.05b
2.91 ± 0.31a
3.83 ± 0.26a
1.12 ± 0.16c
1.70 ± 0.16bc
2.31 ± 0.31b
2.97 ± 0.23a
1.21 ± 0.18a
1.35 ± 0.20a
1.18 ± 0.17a
1.46 ± 0.22a
3.87 ± 0.25b
4.57 ± 0.24b
4.79 ± 0.36b
5.84 ± 0.36a
2.17 ± 0.27c
2.88 ± 0.26bc
3.26 ± 0.23b
5.08 ± 0.39a
2.29 ± 0.16c
3.20 ± 0.38bc
3.76 ± 0.35b
5.65 ± 0.63a
1.66 ± 0.11c
2.50 ± 0.13b
2.38 ± 0.22b
3.32 ± 0.27a
1.35 ± 0.21c
1.53 ± 0.18bc
1.97 ± 0.20b
2.96 ± 0.38a
1.49 ± 0.20b
1.70 ± 0.12b
3.04 ± 0.39a
3.72 ± 0.18a
5.36 ± 0.35c
7.73 ± 0.79b
8.89 ± 0.61b
10.87 ± 0.53a
2.22 ± 0.18c
3.18 ± 0.30bc
4.02 ± 0.68b
6.02 ± 0.65a
2.38 ± 0.30c
3.72 ± 0.36bc
4.32 ± 0.82b
7.20 ± 0.65a
1.64 ± 0.27c
1.81 ± 0.18bc
2.67 ± 0.49ab
3.30 ± 0.21 a
1.01 ± 0.15b
1.09 ± 0.13b
1.58 ± 0.17a
1.64 ± 0.23a
2.71 ± 0.29a
2.63 ± 0.38a
2.97 ± 0.24a
2.55 ± 0.38a
63.54 ± 3.09d
89.17 ± 3.12c
112.96 ± 5.39b
142.09 ± 5.93a
2,4-D functions upstream of JA and ET pathways
Our results show that 2,4-D treatment elicits the biosynthesis of JA and ET (Figs 3a, 4a). To confirm that 2,4-D acts upstream of JA and ET biosynthesis, we further investigated the changes in the levels of OsLIS transcripts and TrypPIs, two markers that can be induced by 2,4-D (Figs 1, 5), in transgenic plants with impaired JA and ET signaling. We used an antisense OsHI-LOX line (as-lox), which produces only 50% of the JA levels relative to WT plants when infested by SSB (Zhou et al., 2009), and an antisense-OsACS2 line (as-acs), which produces 40% of ET levels of WT plants on infestation with SSB (Lu et al., 2011). As found above (Figs 1, 5), 2,4-D induced OsLIS transcripts and TrypPI activity in WT plants. However, this induction was reduced in as-lox and as-acs lines (Fig. 7). This suggests that 2,4-D functions upstream of JA and ET.
2,4-D induces direct and indirect resistance in the laboratory and the field
Based on the results above, we hypothesized that 2,4-D should protect plants against herbivores. In the laboratory, SSB caterpillars gained less mass on 2,4-D-treated rice plants than on untreated control plants (Fig. 5b). BPH female adults, however, were found more frequently on plants whose roots were treated with 2 mg l−1 of 2,4-D than on control plants (Fig. 5c). Similarly, BPH female adults laid significantly more eggs on 2,4-D-treated plants than on control plants (Fig. 5c). In accordance with the volatile measurements, females of the BPH parasitoid A. nilaparvatae were more strongly attracted by plants treated with 2,4-D than control plants, and preferred BPH-infested plants treated with 2,4-D over BPH-infested plants (Fig. 5d).
In the field, higher densities of female and male adults, as well as eggs of BPH, were observed on plants sprayed with 2,4-D than on plants sprayed with the buffer (Fig. 8a,b,d). Interestingly, however, the number of nymphs did not differ between the treatments (Fig. 8c). 2,4-D application strongly enhanced the parasitism of N. lugens eggs by the egg parasitoid A. nilaparvatae (Fig. 8e). When 2 and 5 mg l−1 of 2,4-D were sprayed on rice plants, the parasitism of N. lugens eggs by A. nilaparvatae on the plants was 2.37- and 1.81-fold higher than that on plants sprayed with the buffer (Fig. 8e). The density of spiders was not increased by 2,4-D treatment (Fig. S4). No negative effects of 2,4-D on plants were observed within 30 d in the field.
2,4-D as a defense elicitor
Synthetic chemical elicitors can be used to control pests by eliciting plant defense, but are also powerful tools to elucidate the mechanisms behind plant defense responses (Kawasumi & Nghiem, 2007). Therefore, it is of importance to construct robust high-throughput systems to find new chemical elicitors. Here, based on our previous finding that linalool in rice plants can be induced by herbivore infestation and JA treatment, we established such a system. The 1.55-kb promoter region upstream of OsLIS fused to a GUS reporter gene is an efficient reporter construct to measure the induction of plant defense (Fig. 1e,f). The discovery that 2,4-D is a strong elicitor of rice defenses shows how useful this system can be to find novel bioactive compounds (Figs 2–8).
Even at low doses, 2,4-D induced MAPK genes and WRKY transcription factors (Fig. 2), both of which play essential roles in early plant defense signaling (Wu et al., 2007; Qiu et al., 2008; Skibbe et al., 2008; Rodriguez et al., 2010). Furthermore, phytohormone biosynthetic genes were activated and JA and ET levels increased in 2,4-D-treated plants (Figs 3, 4), resulting in an increased activity of TrypPIs (Fig. 5) and elevated emission of plant volatiles (Fig. 6, Table 1). The fact that these defense responses are attenuated in transgenic plants impaired in JA and ET signaling (Fig. 7) indicates that 2,4-D acts upstream of JA and ET. Our observation that 2,4-D mimics processes that are induced by herbivory and JA application (Lou et al., 2005a,b; Zhou et al., 2009) underscores its proposed mode of action: Plants most probably recognize 2,4-D and quickly activate MAPK cascades and/or WRKYs. The activation of MAPKs and WRKYs enhances the transcript levels of other transcription factors, including other WRKYs, and defense signal biosynthesis-related genes, such as OsHI-LOX, OsAOS1 and OsACS2, which, taken together, increase the concentrations of JA and ET. The resulting signaling profile then shapes the rice defense response. 2,4-D, as an auxin homolog, seems to have similar effects in rice as IAA in Arabidopsis (Woltering et al., 2005; Chandler, 2009; Kazan & Manners, 2009), suggesting that 2,4-D binds to auxin receptors in the plant, thereby triggering an immune response. Clearly, the auxin pathway deserves further attention as a potential modulator of plant defenses (Kazan & Manners, 2009).
Interestingly, 2,4-D treatment enhanced the expression levels of OsMPK3 and OsMPK6, the homologs of NaWIPK and NaSIPK, which have been reported to positively regulate the biosynthesis of JA and SA in Nicotiana attenuata (Wu et al., 2007), but reduced the levels of SA (Fig. 4). A similar phenomenon was also found in cultivated tobacco plants, in which silencing of WIPK resulted in higher levels of SA (Seo et al., 1995). This suggests that the function of OsMPK3 and OsMPK6 differs between plant species. In rice, an antagonistic role between the JA and SA signaling pathways has been reported (Lee et al., 2004). Thus, one possible reason for the observed phenomenon is a suppressive effect of the JA pathway (activated by OsMPK3 and OsMPK6) on the production of SA. Additional reasons may be a direct inhibitory role of OsMPK3, OsMPK6 or other factors that are elicited by 2,4-D on SA biosynthesis. These hypotheses need to be investigated further.
In Arabidopsis, NPR1 is a major regulator of pathogen-induced SA-mediated systemic acquired resistance (SAR), and is involved in cross-talk between SA- and JA-dependent defense pathways (Spoel et al., 2003; Loake & Grant, 2007). We found here that 2,4-D enhanced the expression levels of OsNPR1, but suppressed the production of SA (Fig. 4). Recently, Knoth et al. (2009) have found that 3,5-dichloroanthranilic acid and 2,6-dichloroisonicotinic acid, both of which have largely similar structures to 2,4-D, elicit SA-independent but NPR1-dependent defense responses in Arabidopsis. This suggests that these chemicals can bypass SA to activate NPR1 directly or indirectly.
2,4-D: herbicide, susceptibility factor or plant strengthener?
2,4-D is the most widely used herbicide in the world (Stebbins-Boaz et al., 2004; Teiheira et al., 2007). It is mainly used to control broad-leaf weeds in monocotyledonous crops, including rice (Lyon et al., 1993; Wright et al., 2010). Our study shows that the application of 2,4-D may influence crop resistance much more profoundly than commonly assumed. An early study reported that maize plants treated with 2,4-D rendered corn plants more susceptible to the European corn borer Ostrinia nubilalis and corn leaf aphids (Oka & Pimentel, 1976). Similarly, rice was found to be more susceptible to SSB on 2,4-D treatment (Ishii & Hirano, 1963). However, these effects are likely to result from increased nitrogen concentrations in treated plants (Ishii & Hirano, 1963; Norris & Kogan, 2000) and are not necessarily related to volatile signals. In contrast with these earlier studies, the current work shows that a rice genotype with a high capacity for induced defenses resists SSB attack more strongly after treatment with 2,4-D, a result that is likely to stem from the increased TrypPI activity. Furthermore, because of the increase in volatile emissions, 2,4-D-treated rice became more attractive for BPH and its egg parasitoid A. nilaparvatae. Herbicides have been reported to influence infestation rates of arthropods in crop plants (Norris & Kogan, 2000), but few have previously been shown to act as defense elicitors. 2,4-D has been reported to induce the production of glucosinolates in Arabidopsis (Dalgaard et al., 2003) and the expression of glutathione-S-transferase genes (Wagner et al., 2002; Fode et al., 2008). The new evidence for the activity of 2,4-D as a strong elicitor of immune responses of rice may thus have far-reaching consequences for integrated pest management strategies.
The use of 2,4-D in pest control
From the above results, it becomes evident that 2,4-D can profoundly influence rice defenses and, as a consequence, alter the interaction of the plant with other organisms, up to the third trophic level. This was confirmed in field experiments, in which 2,4-D-treated plants attracted significantly larger numbers of BPH adults, which led to a significantly higher egg load on treated plants. Parasitism by A. nilaparvatae was also increased on the same plants, which may explain the small numbers of nymphs on the plants. This effect of 2,4-D application on two trophic levels has interesting control potential, as it can concentrate pest insects and natural enemies on the same plants, and thereby may be used to redistribute and reduce pest damage in neighboring plants. This is the principal behind the push–pull strategy in pest control, whereby some plants are highly attractive to pests, whereas others are not, or are even repellent (Cook et al., 2007). In the most successful example of a push–pull approach, natural enemies of the pest are attracted to the same plants that repel the pest (Khan et al., 2010). This is different for 2,4-D application, as it results in attraction of the pest and its natural enemy. If correctly exploited, this may have its advantages: planting rice in a conventional manner and applying 2,4-D to the edges of a field would create a trapping zone that reduces pest damage in the inner parts of the field. This would also lead to larger numbers of egg parasitoids in these zones, keeping pest reproduction to a minimum and creating a source of parasitoids that can attack the next generations of the pest, also in the untreated parts of the fields. If needed, insecticide treatments could be limited to the edges of the plots after parasitoid emergence. Contrary to common push–pull approaches (Cook et al., 2007), such a strategy could be realized with minimal additional costs to the farmer and no loss of cultivated area. As 2,4-D at low concentrations has no negative effects on rice development (Liu, 1963), the current study may thus provide the basis for the development of a novel and simple pest control strategy in rice. Further studies should confirm the effects of 2,4-D in a large-area field, and determine the best application method, including concentration, application time and application area. Furthermore, it should be verified to what extent the strategy can reduce the vectoring of viruses by BPH. As a result of its simple implementation, a 2,4-D-based control technique for BPH could be adopted quickly in rice-growing regions around the world.
The study was jointly sponsored by the National Basic Research Program of China (2010CB126200), the Innovation Research Team Program of the National Natural Science Foundation of China (31021003), the Natural Science Foundation of Zhejiang Province (D3080282) and the earmarked fund for Modern Agro-industry Technology Research System.