Herbivore-induced volatiles induce the emission of ethylene in neighboring lima bean plants

Authors

  • Gen-ichiro Arimura,

    1. Bio-oriented Technology Research Advancement Institution, Tokyo 105–0001, Japan,
    2. Laboratory of Insect Physiology, Graduate School of Agriculture, Kyoto University, Kyoto 606–8502, Japan,
    Search for more papers by this author
  • Rika Ozawa,

    1. Bio-oriented Technology Research Advancement Institution, Tokyo 105–0001, Japan,
    2. Laboratory of Insect Physiology, Graduate School of Agriculture, Kyoto University, Kyoto 606–8502, Japan,
    Search for more papers by this author
  • Takaaki Nishioka,

    1. Laboratory of Insect Physiology, Graduate School of Agriculture, Kyoto University, Kyoto 606–8502, Japan,
    Search for more papers by this author
  • Wilhelm Boland,

    1. Max Planck Institute for Chemical Ecology, Carl-Zeiss-Promenade 10, D-07745 Jena, Germany,
    Search for more papers by this author
  • Thomas Koch,

    1. Max Planck Institute for Chemical Ecology, Carl-Zeiss-Promenade 10, D-07745 Jena, Germany,
    Search for more papers by this author
  • Frank Kühnemann,

    1. Institute for Applied Physics, University of Bonn, Wegelerstrasse 8, D-53115 Bonn, Germany, and
    Search for more papers by this author
  • Junji Takabayashi

    Corresponding author
    1. Center for Ecological Research, Kyoto University, Otsuka 509–3, Hirano, Kamitanakami, Otsu, 520–2113, Japan
      For correspondence (fax +81 77 549 8235; e-mail junji@ecology.kyoto-u.ac.jp
    Search for more papers by this author

For correspondence (fax +81 77 549 8235; e-mail junji@ecology.kyoto-u.ac.jp).

Summary

Herbivore attacks induce leaves to emit a specific blend of volatiles. Here we show that exposure to Tetranychus urticae-induced volatiles, as well as T. urticae infestation and artificial wounding, activates the transcription of the genes involved in the biosynthesis of ethylene [S-adenosylmethionine (SAM) synthetase and 1-aminocyclopropane-1-carboxylic acid oxidase] and a gene involved in the biosynthesis of polyamines from SAM (SAM decarboxylase) in lima bean leaves. Moreover, exposure of leaves to any one of the seven major chemical components of T. urticae-induced volatiles also induces expression of these genes. Furthermore, we found that, when lima bean plants were exposed to T. urticae-induced volatiles, they emitted ethylene. Lima bean plants infested by T. urticae and artificially wounded plants also emitted ethylene. Endogenous polyamine levels were not increased in the exposed leaves or the infested leaves, suggesting that polyamine production from SAM was only slightly promoted at the metabolic levels present in the leaves. We found that jasmonate (JA) accumulated in leaves exposed to T. urticae-induced volatiles, and that both JA and salicylate (SA) accumulated in leaves infested by T. urticae. These findings, as well as results of pharmacological analyses, suggest that, in leaves exposed to T. urticae-induced volatiles, ethylene biosynthesis might be regulated by pathways involving JA and the ethylene positive feedback loop. They also suggest that ethylene biosynthesis might be regulated by signaling pathways involving JA, SA and ethylene in T. urticae-infested leaves.

Introduction

Plants have developed a multitude of defense mechanisms against herbivore attacks (Karban and Baldwin, 1997). One class of defense mechanisms is referred to as induced indirect defense. An example of induced indirect defense is emission of specific blends of volatiles that attract the natural carnivorous enemies of herbivores (De Moraes et al., 1998; Dicke et al., 1990b, 1999; 1996; Turlings et al., 1990). A well-studied example of this type of induced indirect defense involves a tri trophic system consisting of lima bean plants (Phaseolus lunatus), herbivorous spider mites (Tetranychus urticae) and carnivorous mites (Phytoseiulus persimilis) that prey on the spider mites. Infestation by spider mites induces lima bean leaves to emit a blend of volatiles that attracts the predatory mites (Takabayashi and Dicke, 1996). This blend of volatiles also induces plants downwind from the infested plants to become more attractive to the predatory mites (Dicke et al., 1990a) and less susceptible to spider mites (Arimura et al., 2000a).

It is very interesting that the blend of volatiles released from T. urticae-infested lima bean leaves induces the expression of several genes in neighboring lima bean leaves. These genes have been identified as those encoding pathogenesis-related (PR) proteins, lipoxygenase (LOX), phenylalanine ammonia-lyase (PAL), and farnesyl pyrophosphate synthetase (FPS) (Arimura et al., 2000a). The primary components of T. urticae-induced volatiles are terpenoids, methyl salicylate (MeSA), and C6 volatiles (Dicke et al., 1999). Exposing lima bean leaves to vapors of any of the terpenoid or C6 components of T. urticae-induced volatiles has been shown to induce expression of genes for PR proteins, LOX, PAL and FPS (Arimura et al., 2000a; 2001). The level of expression of these genes has been mimicked by exposure to exogenous jasmonic acid (JA), which is known to mediate wound responses in plants (Farmer and Ryan, 1992; McConn et al., 1997).

We recently prepared cDNA microarrays from the cDNA library of lima bean leaves, and used them to detect other transcripts expressed in response to T. urticae infestation and T. urticae-induced volatiles (Arimura et al., 2000b). Among the transcripts detected, it was determined that approximately 90 genes were expressed at high levels (in addition to the genes for PR proteins, LOX, PAL and FPS). Two of the fragments detected were assumed to be from transcripts of the S-adenosylmethionine (SAM) synthetase (SAMS) gene and 1-aminocyclopropane-1-carboxylic acid (ACC) oxidase (ACO) gene. These two genes code for enzymes that catalyze the biosynthesis of ethylene (Adams and Yang, 1979), one of the signals that mediate direct responses to mechanical injury (O'Donnell et al., 1996), herbivore infestation (Kahl et al., 2000) and application of a purified fungus cellulase elicitor (Piel et al., 1997). The transcription of SAMS and ACO is also regulated by the plant's stage of development (Boerjan et al., 1994; Peleman et al., 1989), hormones (Lasserre et al., 1996; Gómez-Gómez and Carrasco, 1996) and/or stresses including wound stimuli (Espartero et al., 1994; Lasserre et al., 1996). This suggests that lima bean leaves might produce ethylene in response to T. urticae-induced volatiles as well as in response to T. urticae infestation. It has been suggested that ethylene and JA synergistically induce the transcription of defense genes in plants (O'Donnell et al., 1996; Penninckx et al., 1998).

In addition, using cDNA microarray analysis, we identified a SAM decarboxylase (SAMDC) gene whose expression was induced in lima bean leaves infested with T. urticae or exposed to T. urticae-induced volatiles (unpublished data). SAMDC is a key enzyme in the biosynthesis of the polyamines (PAs) spermine (spm) and spermidine (spd) from putrescine (Slocum, 1991). Spermine and spermidine are small basic molecules that are believed to promote plant growth and development (Walden et al., 1997). Several studies have shown that PA levels are raised in plants infected with pathogens and those treated with gibberellic acid (Evans and Malmberg, 1989; Yamakawa et al., 1998). SAM is a precursor for both ethylene biosynthesis and PA biosyntheses in plants. These pathways have been shown to be interdependent: it has been demonstrated that ethylene and PAs negatively regulate the synthesis of each other and compete for their common precursor, SAM (Evans and Malmberg, 1989). However, there are counterexamples in which ethylene and PAs are not mutually antagonistic (Evans and Malmberg, 1989).

In the present study, the expression of SAMS, ACO and SAMDC was induced in leaves (detached and intact) in response to infestation, artificial wounding, T. urticae-induced volatiles, or any one of the seven major chemical components of T. urticae-induced volatiles. We then investigated whether lima bean plants produce ethylene and PAs in response to T. urticae-induced volatiles as well as in response to T. urticae infestation. We believe that such induction of ethylene and/or PAs may lead to defensive actions against T. urticae (Arimura et al., 2000a), following the expression of defense genes such as those for PR proteins, because ethylene is considered one of the signaling factors for defenses against herbivory or physiological injury in plants (Kahl et al., 2000; O'Donnell et al., 1996). Based on these findings, as well as results of pharmacological analyses and quantitative analysis of JA and salicylate (SA), we propose a mechanism for the regulation of biosynthesis of ethylene and PAs in lima bean plants.

Results

cDNAs encoding proteins for the biosynthesis of ethylene and PAs

We completely sequenced three cDNA clones, encoding the enzymes SAMS (1659 nucleotides), ACO (1159 nucleotides), and SAMDC (1518 nucleotides), each with a poly(A) tail. A deduced coding region existed within each clone: SAMS, nucleotides 247–1425, 392 amino acids, 43.1 kDa; ACO, nucleotides 27–974, 315 amino acids, 36.2 kDa; SAMDC, nucleotides 283–1347, 354 amino acids, 39.2 kDa. The amino acid sequence of lima bean SAMS showed homology of 92, 59 and 60% with SAMS from A. thaliana, humans and yeast, respectively (Figure 1a). Also, the amino acid sequence of lima bean SAMS contains two conserved signature motifs, ADOMET SYNTHETASE 1 {G-A-G-D-Q-G-x(3)-G-[FYH]} (a.a. 120–130) and ADOMET SYNTHETASE 2 {G-[GA]-G-[ASC]-F-S-x-K-[DE]} (a.a. 267–275). Plant SAMS can be classified into two groups based on differences at specific positions (Schröder et al., 1997). According to this classification system, the lima bean SAMS identified in the present study is classified as type I. Lima bean ACO shows amino acid homology of 77, 78 and 68% with ACO from tobacco, tomato and rice (Figure 1b). Lima bean SAMDC shows amino acid homology of 70, 51 and 38% with SAMDC from tobacco, rice, and humans (Figure 1c). Also, lima bean SAMDC amino acid sequence contains a SAM decarboxylase signature motif, ADOMETDC {[SA]-[FY]-[LIV]-L-[STN]-E-S-S-[LIVMF]-F-[LIV]}, from a.a. 63–73.

Figure 1.

Comparison of the deduced amino acid sequences of lima bean SAMS, ACO and SAMDC with those of several other genes.

(a) The amino acid sequences of SAMS from lima bean, A. thaliana (accession No. M55077), human (D49357), and yeast (J03477) are shown.

(b) The amino acid sequences of ACO from lima bean, tobacco (X98493), tomato (AB013101), and rice (X85747) are shown.

(c) The amino acid sequences of SAMDC from lima bean, tobacco (U91924), rice (AJ251899), and human (M21154) are shown.

The alignments were created with the CLUSTALW program. The amino acid residues that are identical among the sequences from at least three species aligned here are indicated by white letters in the black box. Gaps in the sequences are indicated by dashes. The conserved motifs ADOMET SYNTHETASE 1 (solid line), ADOMET SYNTHETASE 2 (broken line), and ADOMETDC (dotted line) are lined up with the amino acid sequences of SAMS and SAMDC. The nucleotide sequence data reported in this paper will appear in the DDBJ/EMBL/GenBank databases under the accession numbers AB062358(SAMS), AB062359(ACO), and AB062360(SAMDC).

Expression analysis of genes involved in ethylene and PA biosyntheses

We used a simple bioassay system (upper part of Figure 2) in which lima bean leaves were exposed to T. urticae-induced volatiles. Using reverse transcription-polymerase chain reaction (RT–PCR) analysis with a relatively low number of cycles, we detected the expression of SAMS, ACO and SAMDC in uninfested neighboring leaves in response to 1-day exposure to the volatiles from T. urticae-infested leaves (Figure 2). The expression of ACO was detected at 1 day but not at 3 days, whereas SAMDC continued to be expressed in the exposed leaves for 3 days. One day of T. urticae infestation induced the expression of SAMS and SAMDC in the leaves, and ACO was expressed in the infested leaves after 3 days. In a control experiment, in which uninfested leaves or neighboring leaves were exposed to volatiles from uninfested leaves, we did not detect expression of SAMS, ACO or SAMDC at any time during the 3-day experimental period. Expression of basic PR-2 (β-1,3-glucanase gene), a defense gene, was induced in both infested and exposed leaves, a result that supports our earlier findings (Arimura et al., 2000a; Ozawa et al., 2000).

Figure 2.

Gene transcripts elicited by volatiles from lima bean leaves.

To test effects of T. urticae-induced volatiles, receiver leaves [Control(R) and T. urticae(R)] were exposed to volatiles from emitter leaves [uninfested (Control) and T. urticae-infested (T. urticae), respectively,] for 1 or 3 days. To test effects of wound-induced volatiles, receiver leaves [Wound (R)] were exposed to volatiles from artificially wounded leaves (Wound) for 3 or 24 h. Total RNA was isolated from the leaves, and the expression of SAMS, ACO, SAMDC and PR-2 was analyzed using RT–PCR. Each experiment was performed three times, with similar results each time.

Artificial wounds also induced the expression of SAMS, ACO, SAMDC and PR-2 in leaves after 3 h (Figure 2). However, wound-induced volatiles elicited the expression of only SAMDC and PR-2 in undamaged exposed leaves during a 24-h observation period. From this, we infer that de novo ethylene biosynthesis was only slightly activated by volatiles from artificially damaged leaves.

Differences in gene expression profiles between herbivore-induced volatiles and wound-induced volatiles appear to be due to differences in composition of the induced volatile blends (Arimura et al., 2000a).

In previous studies, we have used detached leaves. This could be seen as a weakness of our methodology. Therefore, in the present study, we analyzed expression of SAMS, ACO, SAMDC and PR-2 in leaves of intact lima bean plants exposed to volatiles from T. urticae-infested plants for 1 or 3 days in a container (Figure 3). In the leaves of receiver plants, we detected low levels of transcript accumulation for all four genes. In contrast, in a control experiment in which uninfested lima bean plants were exposed to volatiles from other uninfested plants, we did not detect expression of any of the four genes in the receiver leaves. The receiver leaves of potted plants showed lower gene transcript levels than the detached leaves. This could be due to the fact that the amount of volatiles produced by potted lima bean plants infested by T. urticae for 1 day was approximately 10 times less than the combined amount of the major components of the volatile blend produced by detached leaves infested by T. urticae for 1 day (Arimura et al., 2001).

Figure 3.

Gene transcripts elicited by volatiles from potted lima bean plants.

Uninfested intact plants were exposed to volatiles released from uninfested [Control (R)] or T. urticae-infested [T. urticae(R)] plants for 1 or 3 days, in a plastic container. The expression of genes in the leaves was analyzed as in Figure 2.

Volatile compounds responsible for the induction of gene transcripts

We exposed uninfested lima bean leaves to chemically synthesized vapors of four terpenoid compounds and analyzed gene expression in the leaves (left side of Figure 4). β-Ocimene induced the expression of SAMS, ACO, SAMDC and PR-2 after 24 h of exposure. (E)-4,8-Dimethyl-1,3,7-nonatriene (DMNT) also induced the expression of these genes, as indicated by the observed increases in their transcript levels at 3 and 24 h. (E,E)-4,8,12-Trimethyl-1,3,7,11-tridecatetraene (TMTT) induced the expression of these genes after 3 h, but its effect on the levels of SAMS, SAMDC and PR-2 had decreased after 24 h. Linalool induced a high level of SAMDC expression, as observed 3 and 24 h after exposure, and induced a low level of ACO expression after 3 h. No SAMS transcripts were detected at any time during the 24-h exposure to linalool.

Figure 4.

Gene transcripts elicited by synthetic vapors.

Two uninfested leaves were enclosed in a glass container together with a piece of cotton wool containing β-ocimene (E)-4,8-dimethyl-1,3,7-nonatriene (DMNT) (E,E)-4,8,12-trimethyl-1,3,7,11-tridecatetraene (TMTT), linalool (Z)-3-hexenol (E)-2-hexenal or (Z)-3-hexenyl acetate, dissolved in dichloromethane, for 3 or 24 h. The expression of genes in the leaves was analyzed as in Figure 2.

In addition, we analyzed the expression profiles of SAMS, ACO, SAMDC and PR-2 in response to C6 volatiles, which are released in response to infestation with T. urticae and/or artificial wounding (right side of Figure 4). Exposure of lima bean leaves to (Z)-3-hexenol, which was a major component of the C6 volatiles released by artificial wounding (but not T. urticae infestation) in our experiments, only induced the expression of PR-2. In contrast (E)-2-hexenal, which is a minor component of the volatile blend induced by T. urticae infestation, induced strong expression of SAMS, ACO and SAMDC, and weak expression of PR-2. (Z)-3-Hexenyl acetate, which is another major component of the volatile blend induced by T. urticae infestation, only induced SAMS expression during the 24-h exposure.

Time course of ethylene emission in response to herbivore-induced volatiles and herbivore infestation

Young lima bean plants exposed to volatiles from T. urticae-infested plants emitted ethylene at the level of parts per billion (p.p.b.) (Figure 5a). Ethylene emissions by the exposed plants reached a maximum 17.5–20 h after the start of the experiment, and then gradually decreased over the next 11 h. After that, emission continued at low levels with small unimodal patterns.

Figure 5.

Detection of ethylene released from lima bean plants.

(a) We measured ethylene released from lima bean plants exposed to volatiles from T. urticae-infested plants (three replicates). The experimental setup for this measurement is demonstrated at the top. The experiment took place under a 16-h-light/8-h-dark cycle, at 25°C. The black bars indicate darkness, and the white bars indicate light. Each experiment was performed three times, with similar results each time.

(b) We measured ethylene released from lima bean plants infested with T. urticae (four replicates), artificially wounded plants (three replicates), and undamaged plants that served as controls (three replicates).

Lima bean plants also emitted ethylene after artificial wounding or infestation with T. urticae (Figure 5b). The emissions increased rapidly 16–24 h after the onset of infestation, and then decreased over the next 8 h. This unimodal pattern of emission was repeated over the course of the following 3 days; the maximum amount of ethylene and the time at which the peak occurred differed somewhat among the cycles. We further observed that artificial wounding of lima bean plants (punching) elicited the release of ethylene from the plants. In contrast, unexposed lima bean plants that were not infested or wounded emitted very low levels of ethylene during the experimental period.

Endogenous PA levels in lima bean leaves

Our finding that expression of SAMS and SAMDC was induced in infested leaves as well as leaves exposed to T. urticae-induced volatiles (Figure 2) suggests that increased amounts of PAs (spm and spd) should have been found in those leaves, because SAMS and SAMDC are involved in the biosynthetic pathways of PAs. However, we found that the amounts of spm and spd in the leaf infested with T. urticae for 3 days and the leaf exposed to volatiles from the infested leaf for 3 days were not significantly higher than the amounts detected in an uninfested leaf (exposed leaf, P = 0.07 for spm and P = 0.53 for spd; infested leaf, P = 0.08 for spm and P = 0.67 for spd; t-test) (Table 1).

Table 1.  The endogenous levels of spm and spd in lima bean leaves
Duration
(Days)

Spm level (nmol per gram FW) aSpd level (nmol per gram FW) a
Control bT. urticae-induced
volatiles
T. urticae
infestation
Control bT. urticae-induced
volatiles
T. urticae
infestation
  • a

    Data represent the mean of four replications ± SE.

  • b

    Uninfested leaves kept in a glass container served as controls.

084.8 ± 5.0144.0 ± 5.0
140.0 ± 1.434.9 ± 1.7158.5 ± 6.4127.1 ± 5.3
370.6 ± 9.946.3 ± 2.446.9 ± 4.1142.6 ± 11.9130.8 ± 9.5133.0 ± 14.3

Signal transduction of ethylene and PA biosynthetic genes

Exogenous JA (0.1 mm) strongly elicited SAMS, ACO and SAMDC expression in lima bean leaves for 1 day (Figure 6a). Ethylene-releasing compound (ethephon; 10 mm) and MeSA vapor (30 p.p.b.) weakly induced the expression of SAMS and ACO after 1 day. In contrast, the expression of SAMDC was induced by spm (0.3 mm) and spd (0.3 mm), but not by ethephon or MeSA. The expression of PR-2 was induced by JA and ethephon.

Figure 6.

Signaling mechanisms of expression of ethylene and PA synthesis genes.

(a) Aqueous jasmonate (JA; 0.1 mm), ethephon (10 mm), spermine (spm; 0.3 mm), or spermidine (spd; 0.3 mm) was applied to the petioles of leaflets for 1 day. Methyl salicylate (MeSA) was applied to the leaves as a gas (approximately 30 p.p. b.), in a glass container, for 1 day. The expression of SAMS, ACO, SAMDC and PR-2 in the leaves was analyzed. Leaves treated with 10 mm MES (pH 5.5) served as controls.

(b) SHAM (10 mm), STS (0.1 mm), BAPTA (5 mm), or staurosporine (2 µm), was applied to the petioles of leaves. Three hours after each application, the leaves were exposed to volatiles from T. urticae-infested leaves for 1 day or infested with T. urticae for 3 days in a glass container (Figure 2), and the expression of genes in the leaves was analyzed. The controls were leaves that were only treated with water.

(c) Two detached leaves were exposed to T. urticae-induced volatiles for 1 day (gray bars) or 3 days (black bars). During the same time period, two other leaves were infested with T. urticae. Uninfested leaves kept in a glass container for 0 (white bars), 1 or 3 days served as controls. Endogenous JA and SA levels of the leaves were determined. Data represent results of three individual replicates (mean ± standard error).

To investigate the possible involvement of JA in the expression of SAMS, ACO, SAMDC and PR-2 in lima bean leaves exposed to volatiles from T. urticae-infested leaves, we analyzed the inhibitory effects of SHAM on transcription (Figure 6b). SHAM is an inhibitor of LOX, which is a key enzyme in JA biosynthesis via the octadecanoid pathway (Bell et al., 1995). Prior treatment of leaves with SHAM abolished SAMS, ACO, SAMDC and PR-2 expression in the receiver leaves after 1 day. STS, an inhibitor of the binding of ethylene to ethylene receptors, weakened the expression of SAMS, ACO and PR-2, but did not affect the level of SAMDC expression. In addition, we analyzed the involvement of Ca2+ influx into cells and protein phosphorylation in the level of gene expression in the receiver leaves. A chelator of extracellular Ca2+ (BAPTA) and an inhibitor of serine/threonine protein kinase (staurosporine) completely blocked expression of SAMS, ACO, SAMDC and PR-2 after 1 day.

In the leaves infested with T. urticae, prior treatment of leaves with SHAM abolished ACO and SAMDC expression and slightly weakened SAMS and PR-2 expression after 3 days (Figure 6b). STS weakened the expression of ACO, PR-2 and SAMDC in leaves, but did not affect the level of SAMS expression. BAPTA and staurosporine completely blocked the expression of all four genes after 3 days of infestation.

Accumulation of JA and SA in lima bean leaves

We measured the amount of endogenous JA and SA in leaves infested with T. urticae and leaves exposed to T. urticae-induced volatiles (Figure 6c). The amounts of JA and SA in the exposed leaves were 1.6 times greater than the amounts in the control leaves after 1 day of exposure (P = 0.01, t-test). However, after 3 days of exposure, the amounts of JA and SA in the receiver leaves were no longer elevated (P = 0.23). The amount of JA in the infested leaves was elevated after 1 day of infestation (1.9 times the amount measured in the control leaves, P < 0.01) and after 3 days of infestation (2.7 times, P = 0.02). Large amounts of free endogenous SA had accumulated in the T. urticae-infested leaves after 3 days of infestation (approximately 6 times the amount measured in the control, P = 0.02), whereas we observed no significant difference in SA levels between the receiver leaves and the controls for any duration of exposure (1 day, P = 0.78; 3 days, P = 0.32) (Figure 6c).

Discussion

Biosynthesis of ethylene and PAs

We demonstrated that ethylene emission from lima bean plants could be triggered not only by T. urticae infestation or artificial wounding but also by exposure to volatiles from infested lima bean plants (Figure 5). Previous studies have demonstrated that ethylene is emitted from plants that are herbivore-infested (Kahl et al., 2000), wounded (O'Donnell et al., 1996), or treated with a fungal elicitor (Piel et al., 1997). Our finding that ethylene was emitted from plants exposed to T. urticae-induced volatiles is novel. We also found that the levels of endogenous spm and spd were only slightly elevated in the infested leaves and neighboring leaves (Table 1). These results suggest that ethylene biosynthesis, one of the pathways that have SAM as a precursor, was selectively promoted in these leaves. Nevertheless, because the expression of SAMS and SAMDC was induced in leaves that were either infested or exposed to T. urticae-induced volatiles (Figure 2), we infer that the transcriptional activation of PA biosynthetic genes is not always indicated by levels of PA de novo biosynthesis. Alternatively, PAs might accumulate locally in intercellular spaces. It has been reported that the accumulation of spm in intercellular spaces is increased in tobacco leaves infected with tobacco mosaic virus (Yamakawa et al., 1998).

Furthermore, we discovered that, in infested lima bean leaves or those exposed to T. urticae-induced volatiles, the emission of ethylene is accompanied by the expression of SAMS and ACO (Figures 2, 3 and 5). In a previous study, we observed the induction of an ACC synthase (ACS) gene in lima bean plants in response to such stimuli (Arimura et al., 2000b). In several plant species, the conversion of SAM to ACC (and subsequent oxidation of ACC) is considered to be the rate-limiting step in ethylene synthesis (Lasserre et al; Vriezen et al., 1999; Yang and Hoffman, 1984). We therefore suggest that ethylene production in lima bean leaf tissues is limited by the de novo synthesis of ethylene biosynthesis enzymes.

The signaling pathway for the expression of genes

Figure 7 shows proposed signal cascades that we believe might be involved in ethylene biosynthesis in lima bean leaves in response to T. urticae infestation or volatiles emitted from nearby infested leaves. These cascades, involving a JA-dependent signal transduction pathway, were deduced from the following present findings: (1) both artificial wounding and exogenous application of JA induced the expression of ethylene synthesis genes (Figure 6a); (2) JA accumulated in leaves infested with T. urticae and leaves exposed to T. urticae-induced volatiles (Figure 6c); (3) SHAM, an inhibitor of the JA synthetic enzyme LOX, suppressed the expression of SAMS and ACO in the infested leaves and in the receiver leaves (Figure 6b). Moreover, the inhibitory effect of SHAM was abolished by simultaneous treatment of leaves with exogenous JA and SHAM (data not shown). In addition, SA might also be involved in ethylene synthesis in infested leaves, because SA induces the expression of SAMS and ACO (Figure 6a). Moreover, high levels of endogenous SA were found to accumulate in infested leaves (Figure 6c). It has been reported that, in tomato fruits and pear cell suspension cultures, SA inhibits ethylene biosynthesis by blocking ACS at the level of transcription and enzymatic activity, respectively (Leslie and Romani, 1986; Li et al., 1992). Therefore, it remains to be determined whether SA has a positive or negative effect on ethylene biosynthesis in infested leaves. We believe that SA is not involved in the signaling cascades in receiver leaves, because of the following findings: (1) SA only slightly accumulated in receiver leaves (Figure 6c); (2) our previous study showed that an SA-inducible gene encoding acidic PR-4 is not expressed in receiver leaves (Arimura et al., 2000a).

Figure 7.

A possible mechanism for the signal transduction pathway of ethylene biosynthesis in lima bean leaves in response to T. urticae infestation and T. urticae-induced volatiles.

Abbreviations: ACO, 1-aminocyclopropane-1-carboxylic acid oxidase; JA, jasmonate; Met, methionine; SA, salicylate; SAM, S-adenosylmethionine; SAMS; SAM synthetase

We also found that ethylene itself activates the expression of ethylene synthesis genes in both receiver leaves and infested leaves (Figure 6). This ethylene-regulated feedback has been observed in several plant species. Ethylene is responsible for the positive feedback regulation of ACS and ACO expression in tomato fruits (Rottmann et al., 1991), carnation petals (Woodson et al., 1992) and etiolated pea seedlings (Peck and Kende, 1995). It has also been implicated in the negative feedback regulation of ACS expression in deepwater rice internodes (Bleecker et al., 1987).

In addition, de novo ethylene synthesis following herbivore infestation or exposure to T. urticae-induced volatiles might be secondarily mediated by Ca2+ influx into cells and protein phosphorylation (Figure 6b). In contrast, Kwak and Lee (1997) have suggested that both Ca2+ influx and protein phosphorylation are required for ACO expression, but not SAMS expression, in peas in response to ethylene. The signal cascades of ethylene synthesis genes may therefore differ among plants.

We also found that the expression of SAMDC was induced by several stimuli, including infestation, wounding, wound-/herbivore-induced volatiles and application of JA and PAs (Figures 2 and 6a). In leaves infested by T. urticae or exposed to T. urticae-induced volatiles, Ca2+ influx and protein phosphorylation were required for SAMDC expression (Figure 6b). Also, the ethylene inhibitor STS suppressed SAMDC expression in infested leaves (Figure 6b). These results might be due to interactions between ethylene synthesis pathways and PA synthesis pathways (Evans and Malmberg, 1989).

Volatiles responsible for the induction of ethylene synthesis genes and PA synthesis genes

In a previous study, we showed that the expression of PR, LOX, PAL and FPS genes in lima bean leaves is induced by the terpenoid compounds DMNT, TMTT and β-ocimene (Arimura et al.. 2000a). The present results suggest that linalool, one of the T. urticae-induced terpenoids, induces the expression of ACO and SAMDC. (Z)-3-Hexenol is a major component of the phytochemical vapors released from artificially wounded lima bean leaves immediately after they are wounded (Arimura et al., 2000a). In the present study (Z)-3-hexenol induced expression of PR-2, but did not induce expression of any other genes that were analyzed (Figure 4). This strongly supports our finding that expression of ethylene synthesis genes was triggered by exposure to T. urticae-induced volatiles but not wound-induced volatiles (Figure 2). Thus, there is the possibility that multiple mechanisms are involved in the process by which reception of volatiles by receiver leaves leads to activation of a complex signal transduction mechanism dependent on JA and/or ethylene.

In addition, the expression of ethylene synthesis genes in the receiver leaves was not found to be associated with wound-induced volatiles (Figure 2), even though the wound-induced volatiles included p.p.b. levels of ethylene (Figure 5b). Thus, we conclude that higher concentrations of ethylene than those present in these wound-induced volatiles might be required for activation of ethylene synthesis.

Ethylene promotes the transcription of PR genes

We concluded that ethylene mediates the signal transduction of PR-2 transcription in both receiver and infested leaves, based on the following results: (1) the ethylene-releasing compound ethephon elicited PR-2 expression (Figure 6a); (2) the ethylene inhibitor STS blocked the induction of PR-2 expression in both receiver and infested leaves (Figure 6b). Recent studies have elucidated the mechanism governing regulation of transcription of PR genes, and have shown that a GCC box within a cis-acting element of the promoter region of PR genes is an important element of ethylene-regulated transcription (Ohme-Takagi and Shinshi, 1990; Raventós et al., 1995). We therefore infer that ethylene production has a positive effect on the expression of PR genes in lima bean leaves. We further conclude that ethylene and JA act synergistically to induce the expression of ethylene synthesis genes and PR genes in lima bean leaves. In tomato, tobacco and Arabidopsis plants, ethylene and JA are known to act together to regulate the expression of defense genes (encoding proteinase inhibitor, PRs and PDF1.2, respectively) during wound response or pathogen infection (O'Donnell et al., 1996; Penninckx et al., 1998; Xu et al., 1994).

In the present study, we demonstrated that multifunctional signaling cascades involving ethylene and JA are activated in herbivore-infested plants and plants exposed to herbivore-induced volatiles. We conclude that activation of these signaling cascades results in efficient induction of defense genes. More extensive pharmacological analyses and analyses using elaborate mutant and transgenic plants are needed to clarify such complex signaling cascades in plants.

Experimental procedures

Plants and mites

Lima bean plants (Phaseolus lunatus cv. Sieva) were reared in plastic pots (diameter, 12 cm; depth, 10 cm) in a climate-controlled greenhouse (25 ± 2°C). We used plants with two fully expanded primary leaves (2–3-weeks-old) for the experiments. The herbivorous mites (T. urticae) were obtained from a laboratory-maintained culture reared on kidney bean plants (P. vulgaris cv. Nagauzuramame) under the same conditions as those described above for the lima bean plants.

Bioassays

In all experiments using detached leaves, with the exception of the ethylene measurement, bioassays were performed according to the methods we described in a previous report (Arimura et al., 2000a) (see upper part of Figure 2).

For bioassays using potted lima bean plants, we placed approximately 100 T. urticae females on each of a plant's primary leaves. In each experiment, two infested potted plants (together in one plastic dish filled with water) were positioned on one side of a plastic container (18 l). We then placed two uninfested potted lima bean plants on the other side of the container. The container was then sealed, and was kept at 25°C, 50–70% R.H., 16L:8D (2150 lux, fluorescent lights). During the bioassay period, the container was opened once every day for 5 min for air exchange. To prevent invasion of the uninfested plants by T. urticae, we used wet cotton wool to partition the container into two areas: one for infested plants and one for uninfested plants. Prior to subsequent analyses, we checked the plants to ensure that the receiver leaves were not infested by T. urticae.

Each assay was repeated three times. The three different containers used for these procedures were washed with hot water and then ethanol, and dried at 65°C.

Because the whole plants were confined in the sealed container for 1–3 days, the available CO2 may have been depleted, and this could cause several types of stress. We checked for the effects of such stress on the plant's photosynthetic activity by monitoring the photosynthesis yield (arbitrary unit) in attached and detached leaves, using a photosynthesis yield analyzer (model: MINI-PAM, WALZ, Effeltrich, Germany). However, we saw no evidence to indicate that the photosynthetic activity of the plants was affected by stress (Arimura et al., 2001).

Chemical treatment

JA (0.1 mm; Wako, Kyoto, Japan), silver thiosulphate (STS, 0.1 mm; Palace Chemical Co., Ltd, Yokohama, Japan), and 1,2-bis-(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA, 5 mm; Sigma, St Louis, MO, USA) were applied to the petioles of leaflets in aqueous solutions. Salicylhydroxamic acid (10 mm; SHAM; Sigma), ethephon (10 mm; Wako), spm (0.3 mm; Wako), and spd (0.3 mm; Wako) were administered in 50 mm MES (pH 5.5). Staurosporine (2 µm; Wako) was applied as an aqueous solution with 0.1% DMSO. Leaves were exposed to methyl salicylate (Wako) as a gas (approximately 30 p.p.b.) in a glass bottle (2 l).

β-Ocimene, DMNT and TMTT were synthesized in the laboratory. Linalool and C6 volatiles were obtained from Tokyo Chemical Industry and Wako, respectively. Each C6 volatile was dissolved in pure dichloromethane (1 µg per µl solution), and 10 µg of each C6 volatile was then impregnated into a piece of cotton wool. The amount of each synthetic compound used in the bioassays was based on the amount of DMNT (a terpenoid volatile that is a major component of infestation-induced volatiles) emitted by lima bean leaves infested with T. urticae (approximately 0.35 µg h-1 leaf−1 after 23–24 h of infestation) (Arimura et al., 2001). After evaporation of the solvent, we enclosed each piece of cotton wool in a container (7 l) together with 2 leaves for 3 or 24 h, under the conditions described above (in the Bioassays section). All chemicals, except β-ocimene, were more than 98% pure. β-ocimene was a mixture of E (about 70%) and Z (about 30%) isomers. A piece of cotton wool containing only dichloromethane was used as the control.

Analysis of gene expression

Total RNA was extracted from two leaves by means of the acid guanidinium-phenol-chloroform method (Chomczynski and Sacchi, 1987). First-strand cDNA was synthesized from 1 µg of total RNA using AMV reverse transcriptase XL (Life Sciences Inc., Petersburg, FL, USA) and one of 3 gene-specific primers. These 3 primers (5'-ATACTCAACAGTGACTTGGG-3', 5'-CATCCTGGAAGAGAAGGATG-3' and 5'-TTGTAGTCCCACAGGTTTTG-3') corresponded to nucleotides 758–777, 575–594, and 522–541 of lima bean SAMS, ACO and SAMDC, respectively. The first-strand cDNA was amplified by adding Taq DNA polymerase (Takara, Otsu, Japan) and one of three additional primers. These three primers (5'-AGACATGCACCAAAACCAAC-3', 5'-TGACTAAAGAGCACTACCAG-3' and 5'-CTCCATCTTTCCAATTGCTC-3') corresponded to nucleotides 377–396, 181–200, and 205–224 of lima bean SAMS, ACO and SAMDC, respectively. The SAMS cDNA was amplified with 20 cycles of 95°C for 40 sec, 58°C for 90 sec, and 72°C for 60 sec. ACO and SAMDC cDNAs were amplified with 21 cycles of 95°C for 40 sec, 55°C for 90 sec and 72°C for 60 sec. These amplifications were performed at a sub-optimal dynamic range before leveling off. Synthesis of basic PR-2 (β-1,3-glucanase) cDNA was performed using the method of Arimura et al. (2000a), with the number of PCR cycles reduced to 21. PCR product and total RNA (5 µg) were separated by electrophoresis in 1.5% agarose gel and detected by staining with ethidium bromide. No PCR product was amplified in control amplifications that lacked RT. All RNA analyses were performed in triplicate, using a new set of plants for each iteration.

Measurement of ethylene, PAs, JA and SA

Ethylene production was measured non-invasively in real-time with a photoacoustic laser spectrometer containing a line-tunable infrared CO2 laser and a resonant photoacoustic cell (Beβler et al., 1998; Kahl et al., 2000). Five plantlets in vials were infested with approximately 600 T. urticae, placed in a glass cuvette (2.5 l), and exposed to a constant flow of catalytically cleaned air (passed through a platinum catalyst at 450°C to remove hydrocarbons). Artificially wounded plantlets were prepared by punching 20 holes (diameter, 7 mm) into each leaf. For measurement of ethylene production in the receiver plants, two additional cuvettes (1.6 l) were positioned downstream of the first cuvette (Figure 5a). One of these downstream cuvettes contained five receiver plantlets that were then exposed to T. urticae-induced volatiles, and the other was an empty reference cuvette of the same dimensions. The air from each cuvette was pumped through a liquid-N2 trap to remove CO2 and H2O, and then pumped into a photoacoustic detection cell that was used to measure the ethylene concentration every 3 min. The amount of ethylene produced by the receiver plants was determined by subtracting the amount detected in the empty cuvette.

For quantification of free PAs in lima bean leaves, fresh tissues were homogenized in 5 ml of perchloric acid and centrifuged twice at 14 000 g for 5 min. PAs in the supernatant were derivatized with benzoyl chloride, using diaminohexane as an internal standard, according to the method described by Flores and Galston (1982), with some modifications. Separation and quantification of PA derivatives were performed using an HPLC system (model LC-10AT, Shimadzu, Kyoto, Japan) equipped with a UV detector, under the following conditions: column, Inertsil ODS-3 (GL Sciences, Tokyo, Japan; i.d., 4.6 mm; length, 150 mm); column temperature, 40°C; mobile phase, 60% (v/v) methanol; flow rate, 0.8 ml min−1; and detection, 254 nm.

JA was extracted from the two leaves according to the method of Weber et al. (1997), with slight modifications. Dihydrojasmonic acid was used as an internal standard. The amounts of JA and dihydrojasmonic acid present were determined by GC–MS. GC was performed with a Hewlett-Packard 6890 with an HP-5MS capillary column: i.d., 0.25 mm; length, 30 m; film thickness, 0.25 µm; injection temperature, 250°C. MS was performed with a Hewlett-Packard 5973 mass selective detector, with ionization at 70eV. The temperature gradient used was as follows: 60°C for 1 min, 60–120°C at 20°C min−1, 120–180°C at 3°C min−1, and 180–300°C at 30°C min−1.

SA was extracted from one leaf and quantified according to the method of Malamy et al. (1992), with minor modifications. Separation and quantification were performed using a high-performance liquid chromatography system equipped with a spectrofluorescence detector (model RF-550 A, Shimadzu) under the following conditions: column, µBondasphere 300 5-µm C-18 (Waters, Milford, MA, USA; i.d., 3.9 mm × 150 mm); column temperature, 25°C; mobile phase, 23% (v/v) methanol in 20 mm sodium acetate (pH 5.0); flow rate, 1 ml min−1; excitation wavelength, 313 nm; and emission wavelength, 405 nm. All data were corrected to compensate for losses.

Acknowledgements

We would like to thank Dr Y. Ohashi and Dr S. Seo for their technical assistance in the measurement of SA, and Dr K. Matsuda for donating dihydrojasmonic acid, an internal standard for the measurement of JA. This research was supported by the Program for Promotion of Basic Research Activities for Innovative Biosciences (Bio-oriented Technology Research Advancement Institution) and CREST of JST (Japan Science and Technology Corporation).

Ancillary