Cross-talk between wound signalling pathways determines local versus systemic gene expression in Arabidopsis thaliana

Authors

  • Enrique Rojo,

    1. Departamento de Genética Molecular de Plantas, Centro Nacional de Biotecnología CSIC, Campus Cantoblanco UAM, Cta. Colmenar Viejo km. 15,500, 28049 Madrid, Spain
    Search for more papers by this author
  • José León,

    1. Departamento de Genética Molecular de Plantas, Centro Nacional de Biotecnología CSIC, Campus Cantoblanco UAM, Cta. Colmenar Viejo km. 15,500, 28049 Madrid, Spain
    Search for more papers by this author
  • José J. Sánchez-Serrano

    Corresponding author
      For correspondence (fax +34 91 5854506; e-mail jjss@cnb.uam.es).
    Search for more papers by this author

For correspondence (fax +34 91 5854506; e-mail jjss@cnb.uam.es).

Summary

Plants react to mechanical damage by activating a set of genes, the products of which are thought to serve defensive functions. In solanaceous plants, cell wall-derived oligosaccharides and the plant hormones jasmonic acid and ethylene participate in the signalling network for wound-induced expression of proteinase inhibitors and other defence-related genes, both in the locally damaged and in the systemic non-damaged leaves. Here we show that in Arabidopsis thaliana, these signalling components interact in novel ways to activate distinct responses. In damaged tissues, oligosaccharides induce the expression of a specific set of wound-responsive genes while repressing jasmonic acid-responsive genes that are activated in the systemic tissues. The oligosaccharide-mediated repression of the jasmonic acid-dependent signalling pathway is exerted through the production and perception of ethylene in the locally damaged tissue. This cross-talk between separate wound signalling pathways thus allows the set up of different responses in the damaged and the systemic tissues of plants reacting to injury.

Introduction

Plants react to mechanical injury, produced by abiotic or biotic agents, by activating a set of responses that include, in most cases, the transcriptional activation of wound-responsive (WR) genes. While some of them may have a defensive role against organisms that are feeding on the plant or entering through wounds, the function of the others may be related more to wound healing (Bowles 1990). Several components involved in the transduction of wound signals have been identified in the tomato. Among them, oligosaccharides, the peptide systemin, the plant hormones jasmonic acid (JA), abscisic acid (ABA) and ethylene, and electrical signals, have all been shown to support a complex signalling network leading to activation of proteinase inhibitors and other WR genes (Bishop et al. 1984; Doares et al. 1995; O’Donnell et al. 1996; Pearce et al. 1991; Peña-Cortés et al. 1995; Weiss & Bevan 1991; Wildon et al. 1992). The results obtained with this set of genes suggested that a unified wound signalling pathway, operating through JA, activates the expression of the same genes in the locally wounded leaves and in the systemic non-damaged ones (Bergey et al. 1996; Peña-Cortés et al. 1995). In brief, this model suggests that both oligosaccharides (thought to be local signals) and systemin (the putative signal for systemic WR gene activation) activate the octadecanoid pathway for JA synthesis, triggering WR gene expression in the foliage (Bergey et al. 1996; Doares et al. 1995; Howe et al. 1996), and ethylene amplifies the JA signal leading to maximal levels of response (O’Donnell et al. 1996). This single signalling pathway, however, does not account for the differences observed in the pattern of proteins accumulating in local and systemic tissues of wounded tomato plants (Dalkin & Bowles 1989; Lightner et al. 1993). Recently, it has been suggested that wound signalling pathways independent of JA may exist in tomato (O’Donnell et al. 1998). The presence of separate wound signalling pathways, dependent and independent of JA, has been firmly established in Arabidopsis thaliana (León et al. 1998; McConn et al. 1997; Nishiuchi et al. 1997; Rojo et al. 1998; Titarenko et al. 1997). We have previously reported the characterization of wound-inducible genes in A. thaliana exhibiting distinct patterns of expression upon wounding, regarding both time–course and spatial distribution of transcript accumulation (Titarenko et al. 1997). Some of these newly characterized genes are predominantly active in damaged leaves rather than in the systemic ones. In addition, they are not responsive to JA treatment and, moreover, are also induced by mechanical damage in the JA-insensitive coi1 mutant (Feys et al. 1994), indicating that their expression upon wounding is fully independent of JA perception. However, the signal(s) activating this JA-independent transduction pathway was not identified. Here we report that in A. thaliana, JA and oligosaccharides antagonistically trigger wound signalling pathways that activate, through cross-talk, different sets of genes in wounded and systemic tissues.

Results and Discussion

Oligosaccharides activate a JA-independent wound signalling pathway

Previously we have reported the cloning and characterization of wound-inducible genes in A. thaliana. A subset of these genes [JR1, JR2 and vegetative storage protein (VSP)] was also JA-inducible. However, two genes [choline kinase (CK) and WR3] were induced by a separate wound signalling pathway, in a manner independent of JA synthesis and perception (León et al. 1998; Rojo et al. 1998; Titarenko et al. 1997). To elucidate the inducing factor(s) for the JA-independent wound signalling pathways, we analysed whether oligosaccharides could activate these wound-responsive genes in intact Arabidopsis plants, as these molecules have been shown to participate in wound signalling in other plant species. In particular, the ability of oligogalacturonides (OGA) to induce wound-responsive gene expression has been assessed as they may, in vivo, be released upon wounding by mechanical or enzymatic disruption of pectic cell wall components (Benhamou et al. 1990). A wound-inducible polygalacturonase activity recently described in tomato (Bergey et al. 1999) may be involved in the production of endogenous oligogalacturonide elicitors upon wounding. In addition, fungal-derived chitosan oligosaccharides have been shown to mimic the wound response, activating protein-ase inhibitor genes in tomato through the same signalling pathway as OGA (Doares et al. 1995; Howe et al. 1996). As shown in Fig. 1, treatment with chitosan specifically activated the expression of genes inducible through the JA-independent pathway, with a time–course very similar to that observed upon wounding, suggesting that these molecules act early in wound signalling. Treatments with plant cell wall-derived OGA or with the proteinase inhibitor-inducing factor (PIIF) (Ryan 1992) gave essentially identical results (Fig. 1). The activation of CK and WR3 genes indicated that oligosaccharides are readily taken up by liquid-grown Arabidopsis plants. In this system, OGA are likely to have direct access to all damaged leaves in contact with the culture medium (Baydoun & Fry 1985). The transient activation observed even when the signal was constitutively present (i.e. plants treated with oligosaccharides) indicated that a process of desensitization of the signalling cascade was taking place, as reported for perception of chitin fragments in tomato cells (Felix et al. 1998). Moreover, pretreatment with protein kinase inhibitors blocked, at similar inhibitory concentrations, the induction of CK and WR3 by wounding and oligosaccharide treatment (data not shown), suggesting that CK and WR3 activation by both of these stimuli is transduced through a common signalling pathway that involves protein phosphorylation (Farmer et al. 1989). Taken together, these results indicate that oligosaccharides are indeed a primary signal involved specifically in the activation of a JA-independent wound-induction pathway.

Figure 1.

Effect of wounding and oligosaccharide treatment on gene expression.

Wild-type plants grown in liquid culture were wounded (W) or treated with chitosan (CHI), the proteinase inhibitor inducing factor (PIIF) or oligogalacturonides (OGA), and samples were collected at the times indicated (in hours). Northern blots were hybridized to 32P radiolabelled probes from the genes indicated on the right.

Two separate signalling pathways are involved in wound-induced gene expression

Tomato plants respond to treatment with OGA or chitosan with de novo synthesis of JA, which is required for oligosaccharide-induced accumulation of proteinase inhibitor I and II transcripts (Doares et al. 1995; Howe et al. 1996; O’Donnell et al. 1996). In contrast, in Arabidopsis plantlets grown in liquid medium, treatment with chitosan did not significantly alter jasmonate levels in non-wounded plants [14.31 ± 2.84 nmoles g–1 fresh weight (FW) in control plants, compared with 12.41 ± 2.75 nmoles g–1 FW after 1 h treatment with 250 μg ml–1 chitosan]. Moreover, chitosan treatment blocked the twofold increase in jasmonate content observed in non-treated plants 1 h after wounding (28.10 ± 4.87 nmoles g–1 FW in wounded non-treated plants and 15.78 ± 2.27 nmoles g–1 FW in wounded plants treated with 250 μg ml–1 chitosan). As treatment of Arabidopsis plantlets with OGA or chitosan does not activate the expression of genes specific for the JA-dependent pathway (Fig. 1), wound-induced JA production and subsequent JR gene expression in Arabidopsis is likely to be independent of oligosaccharide release at wound sites, and may be triggered by an as yet unidentified analogue of the tomato systemin (Ryan & Pearce 1998). On the other hand, in the Arabidopsis JA-insensitive coi1 mutant (Feys et al. 1994), CK and WR3 genes are induced to wild-type levels by wounding or treatment with oligosaccharides (Fig. 2, and data not shown), indicating that their activation is fully independent of JA perception. Therefore, wound signals are transduced in Arabidopsis through at least two alternative pathways, termed JA- and oligosaccharide-dependent (OSD) pathways hereafter.

Figure 2.

JA-independent induction of WR gene expression.

JA-insensitive coi1 plants grown in liquid culture were wounded (W) or treated with chitosan (CHI), and samples were collected at the times indicated (in hours). Northern blots were hybridized to 32P radiolabelled probes from the genes indicated on the right. A control hybridization to RNA from liquid-grown wild type plants 1.5 h after wounding is also shown (WT).

The OSD signalling pathway represses JA-dependent gene expression in locally damaged tissues

A possible significance of maintaining separate wound signalling pathways could be the activation of distinct responses in local and systemic parts of wounded plants (Titarenko et al. 1997). At early times after wounding, genes specifically induced through the OSD pathway (CK and WR3) attained maximal levels of activation in damaged leaf tissue, whilst transcripts of genes induced through the JA-dependent pathway (JR1, JR2 and VSP) accumulated maximally in systemic tissues (Fig. 3). Although JR transcripts also accumulated in the damaged leaves at later times after wounding (but only in the case of JR2 attaining similar levels to those observed in systemic tissues), CK and WR3 transcripts were barely detectable in systemic leaves. The spatial pattern of expression upon wounding of oligosaccharide-responsive genes thus correlates with the distribution expected for oligosaccharides released upon mechanical damage, which are likely to migrate to only relatively short distances from the wound site (Baydoun & Fry 1985). However, signal strength for the JA-dependent pathway should also be maximal near the wound sites, but none the less JR genes are induced in these tissues only at later times after wounding and, in most cases, to a lesser extent than in systemic tissues. The observed OSD repression of wound-induced JA synthesis may be in part responsible for the reduced expression of JR genes in damaged tissues. To gain a further insight into the mechanisms responsible for this pattern of expression, we first analysed whether the signal molecule, JA, was present in injured leaves of soil-grown plants. As shown in Fig. 4(a), during the first 2 h after wounding there was a fivefold increase in jasmonate content in damaged leaves, where expression of JR genes was largely blocked, and less than twofold in systemic leaves, where JR genes were induced to maximal levels. Thus, the OGA released upon mechanical wounding were not able to block the wound-induced increase in JA levels observed in damaged leaves. OSD repression of JA synthesis is likely to be restricted to the immediate vicinity of the wound site, where a higher concentration of OGA is to be expected. The evidence suggests that, in damaged leaves, wound-induced JA-dependent gene activation has to be downregulated downstream of JA synthesis as well. In these tissues, the onset of the wound-induced accumulation of JR genes and the desensitization of the OSD pathway are synchronized (compare, for instance, the pattern of CK and JR2 expression in damaged leaves shown in Fig. 3). As every stimulus activating the OSD pathway, such as treatments with protein phosphatase inhibitors, mobilizers of Ca2+ stores, antagonists of CaM, and protein synthesis inhibitors, all blocked JA-induced JR gene expression (León et al. 1998; Rojo et al. 1998), the release of oligosaccharides at the wound site was probably responsible for the repression of JR gene expression observed in locally wounded tissues. To test this possibility, plants were treated with JA in the presence or absence of oligosaccharides. As shown in Fig. 4(b), expression of CK and WR3 was induced in the oligosaccharide-treated plants as expected, and this induction was not affected by JA. Interestingly, treatments with either chitosan or the endogenous elicitors OGA or PIIF concomitantly blocked the JA induction of JR genes (VSP, JR1 and JR2). Most probably, this inhibition was not caused by quenching of JA by the oligosaccharides, as it could not be overcome by increasing concentrations of JA up to 1 mm (data not shown). Thus, all the evidence supports the idea that activation of the OSD pathway blocks the JA-responsive pathway.

Figure 3.

Local and systemic induction of WR gene expression.

Soil-grown wild-type plants were wounded, and samples from the damaged leaves (W) or the non-damaged systemic ones (WS) were collected at the times indicated (in hours) after wounding. Northern blots were hybridized as described in Figs 1 and 2.

Figure 4.

Oligosaccharides block JA-induced expression of JR genes at a step downstream of JA synthesis.

(a) Leaves of soil-grown plants were wounded and, at the indicated times (in hours) after wounding, jasmonate content was measured in damaged (W) and systemic (WS) leaves. Values presented are the mean from three independent experiments (bars represent the standard deviation).

(b) Liquid-grown plants were treated with JA (JA) or an equal amount of solvent (C), or treated with either chitosan, proteinase inhibitor-inducing factor or oligogalacturonides for 30 min, and subsequently with JA (JA + CHI; JA + PIIF; JA + OGA, respectively). Samples were collected 1 h after JA application. Northern blots were conducted as in Figs 1 and 2.

Ethylene production and perception is required for oligosaccharide-dependent repression of JA-induced gene expression

It was therefore of interest to elucidate how oligosaccharides repressed the JA pathway downstream of JA synthesis. It has been shown previously that, in tomato plants, wounding or treatment with OGA or JA induced the synthesis of ethylene, and that de novo synthesis of both ethylene and JA was required for wound-induced expression of the proteinase inhibitor II gene (O’Donnell et al. 1996). Arabidopsis also responded to wounding or chitosan treatment with a transient increase in ethylene production relative to untreated plants (Fig. 5a). However, in sharp contrast to tomato, JA treatment did not result in any enhanced ethylene synthesis over untreated plants. These results suggest that, in Arabidopsis, ethylene production in response to mechanical damage is triggered via the OSD pathway. We examined the possible role of ethylene in the regulation of gene expression through this pathway, taking advantage of the ethylene-insensitive mutants (ein and etr mutants) available in Arabidopsis (Johnson & Ecker 1998). The similar levels of chitosan-induced CK and WR3 transcripts in wild-type plants and the ethylene-insensitive mutants (Fig. 5b) suggested that OSD expression of these genes is independent of ethylene perception. In contrast, JA-induced accumulation of JR1, JR2 and VSP transcripts in the ethylene-insensitive mutants was several-fold higher than in wild-type plants, in which treatments with at least 10-fold higher JA concentrations were required for attaining the levels of JR transcript accumulation observed in the etr1 mutant (Fig. 6). Treatment with 1-aminocyclopropane-1-carboxylic acid, a precursor of ethylene synthesis, had no effect on the endogenous JA level (13.45 ± 2.68 nmoles g–1 FW compared with 14.31 ± 2.84 nmoles g–1 FW in non-treated plants) but inhibited, as ethylene itself, JA-induced expression of JR1, JR2 and VSP in wild-type plants, whereas treatment with norbornadiene, a competitive inhibitor of ethylene, activated on its own the expression of these JR genes, albeit to a lesser extent than JA (data not shown). This negative effect of ethylene on JA-induced JR gene expression is in contrast with the synergistic effect that both hormones play on the expression of the defensin PDF1.2 gene (Penninckx et al. 1998). The delay in the time–course of PFD1.2 induction upon fungal attack (maximal accumulation at 48 h after fungal infection) compared with the wound-induced expression of the JR genes (reaching their maximal levels in the systemic tissues after 2 h) may explain this discrepancy, and suggests that cross-talk between these signalling pathways may have both positive and negative effects on gene expression. Indeed, one of the genes we identified (JR3) appears to be regulated by positive interactions of both ethylene and JA signals (Rojo et al. 1998; E. Rojo and J.J. Sánchez-Serrano, unpublished results) in a manner that may resemble PDF1.2 induction. Clearly, a more detailed study of the regulation of expression of these genes is required to understand the molecular mechanisms underlying the interaction of these wound signalling pathways.

Figure 5.

Ethylene involvement in the regulation of WR gene expression.

(a) Ethylene production in response to wounding (W) and JA (JA) and/or chitosan (CHI) was determined in liquid-grown plants. Values are presented as fold increase over the ethylene concentration in their respective-non-treated plants, and are the mean from three independent experiments with three samples per treatment (bars represent the standard deviation).

(b) Liquid-grown wild-type Columbia (Col) or the ethylene-insensitive etr1-3 (etr1), ein2-5 (ein2), or ein3-1 (ein3) plants were treated with chitosan, and harvested for RNA extraction at the times indicated above the lanes. Northern blots were hybridized to the probes indicated on the right.

Figure 6.

Ethylene perception negatively affects JA-induced gene expression.

Liquid-grown wild-type Columbia (Col) or the ethylene-insensitive etr1-3 (etr1), ein2-5 (ein2), or ein3-1 (ein3) plants were treated with JA, at the concentrations indicated above the lanes, and total RNA was isolated after 6 h and analysed by Northern blotting.

Rather than activating local wound-induced gene expression, the activation of ethylene production signalled through the OSD pathway may therefore be involved in blocking the JA-dependent pathway in damaged tissues. Interestingly, JR genes accumulated to higher levels in the locally damaged leaves of the etr1 mutant than in those of wild-type plants (Fig. 7a), as would be expected if ethylene was involved in vivo in JR gene repression in those tissues. Similar results were obtained with the Arabidopsis ein2 and ein3 mutants (data not shown). Thus, all mutations tested which alter sensitivity to ethylene, from etr1 (which disturbs a receptor; Hua & Meyerowitz 1998) to ein3 (in a gene encoding a transcription factor; Solano et al. 1998), affect JA-induced JR gene expression. This suggests that ethylene action may be exerted way down the transduction pathway, perhaps at the level of JR gene transcription. In the etr1 mutant, JR1 and VSP genes were induced by JA to similar levels in the presence or absence of chitosan (Fig. 7b). The use of the weak allele etr1-3 (Penninckx et al. 1998) may explain the somewhat lower JR2 transcript levels observed in mutant plants treated with JA and chitosan. Consistent with this hypothesis, treatment of etr1 plants with l-α-[2-aminoethoxyvinyl]-glycine, an inhibitor of ethylene synthesis (Yang & Hoffman 1984), fully alleviated chitosan-mediated blocking of JA-induced JR2 expression (data not shown).

Figure 7.

Ethylene synthesis and perception are required for repression of JR genes in damaged tissues.

(a) Soil-grown Columbia (Col) or etr1-3 (etr1) plants were wounded, and accumulation of JR transcripts was monitored by Northern blot analysis, both in the damaged (W) and in the systemic leaves (WS) 2 h after wounding, and in leaves from unwounded plants (C).

(b) Liquid-grown Columbia (Col) or etr1-3 (etr1) plants were treated (+) or not (–) with JA and/or chitosan as indicated above the lanes. Samples were collected 2 h later for Northern blot analysis.

Taken together, our results suggest that oligosaccharide-dependent, wound-induced ethylene production and subsequent perception are required for down-regulating JR gene expression in damaged tissues.

Conclusion

We have integrated these experimental results into a model for wound signalling in Arabidopsis, shown in Fig. 8. In this model, mechanical damage triggers several signalling cascades that regulate the expression of distinct sets of genes in wounded and systemic tissues of the plant. The principal features of the model are (i) oligosaccharides (OGA) released from plant cell walls upon wounding (Benhamou et al. 1990; Bergey et al. 1999) block JA synthesis, and activate ethylene production and a set of JA-independent genes (CK and WR3) in the vicinity of the wound site; (ii) the ethylene generated suppresses JA-induced gene expression in the damaged leaves downstream of JA biosynthesis. The relatively short-distance movement of OGA (Baydoun & Fry 1985) and ethylene diffusion may prevent suppression of JA-induced gene expression in systemic leaves. (iii) JR gene (JR1, JR2 and VSP) induction in these distant tissues may involve a systemic signal produced or released in the damaged tissues in a JA-dependent manner, as indicated in previous works (Farmer et al. 1992; Royo et al. 1996). Although the signals involved are conserved, this regulatory system for wound responses is thus different from the well-characterized response in solanaceous plants, where JA and oligosaccharide signals act in sequence and ethylene potentiates their effects (Doares et al. 1995; O’Donnell et al. 1996). The ability of plants to co-ordinately mount local and systemic responses to mechanical damage by the activation and suppression of different sets of genes in these tissues may have important applications for future work aimed at engineering plants with improved responses to biotic and abiotic stresses.

Figure 8.

A model for wound signal transduction in Arabidopsis thaliana.

The putative role of endogenous oligosaccharides (OGA), jasmonic acid (JA) and ethylene (C2H4) in gene activation in the damaged (wounded leaf) and non-damaged (systemic leaf) leaves of wounded Arabidopsis plants is depicted. The components of the ethylene and JA transduction pathways (ETR1, EIN2, EIN3, and COI1) are located at positions likely for their action. Arrows indicate positive action; blunted lines indicate negative regulation. The set of genes active in the respective tissues is encircled.

Experimental procedures

Plant material and treatments

Arabidopsis thaliana ecotype Landsberg erecta plants, grown either in soil or in liquid culture, were used in most experiments. Where indicated, ethylene-insensitive plants containing either the dominant allele etr1-3 (etr1; formerly ein1) or the recessive ein2-5 (ein2) or ein3-1 (ein3) (Guzman & Ecker 1990<!hide{Q1} Au: Guzman & Ecker 1990 has been changed to Guzman & Ecker 1990 so that this citation matches the list>) were used and compared with their corresponding wild-type background, ecotype Columbia. Plants were grown in liquid culture as described elsewhere (Rojo et al. 1998). Prior to treatments, the liquid medium was removed from the wells and replaced with 1 ml of fresh medium. Unless otherwise indicated, plantlets were either wounded as described elsewhere (Rojo et al. 1998) or treated with either 250 μg ml–1 chitosan, previously hydrolysed with nitrous acid (Hadwiger & Beckman 1980), OGA (degree of polymerization 7–14; from a 10 mg ml–1 aqueous stock solution, a kind gift of J.C. Cabrera, Instituto Nacional de Ciencias Agrícolas de Cuba), PIIF (average degree of polymerization 20; from a 10 mg ml–1 aqueous stock solution, a kind gift of C.A. Ryan, Institute of Biological Chemistry, Washington State University) or 50 μm JA (mixed isomers; Apex Organics, UK; from a 100 mm stock solution in N,N-dimethylformamide). l-α-[2-aminoethoxyvinyl]-glycine (Sigma, USA) was used at 5 μm final concentration (from a 2 mm stock solution in water). Plants were grown in soil in the greenhouse at 20°C under a 16 h day/8 h night light regime until the stem elongated to approximately 6 cm. For wounding experiments, half of the rosette leaves of the plant were damaged by thoroughly crushing with a pair of fine tweezers (approximately 50% of the surface was damaged). Systemic leaves were non-damaged rosette leaves of wounded plants.

RNA analysis

Total RNA isolation and Northern blotting techniques with 32P radiolabelled cDNA probes were performed as described elsewhere (Rojo et al. 1998). Equal RNA loading was verified by initial visualization of the ethidium bromide-stained ribosomal RNA in the agarose gel, and hybridization to a 32P-labelled 18S RNA probe after last stripping of the blot.

Experiments were performed independently at least five times, yielding highly reproducible results. Single, representative experiments are shown in the figures.

Measurements of jasmonate content in plant tissues

Jasmonates were extracted overnight in 80% methanol, at 4°C with stirring. Extracts were cleared by passing through C-18 Sep-Pak cartridges (Waters Corp., USA). Eluates were vacuum concentrated to aqueous phase and 1/10 volume of 2 m sodium acetate, pH 4.0, was added, and extracted twice with 1 volume of dichloromethane. The organic phase was separated by centrifugation, transferred to a glass vial and vacuum dried. Samples were methylated with freshly prepared diazomethane and the jasmonate content determined by competitive ELISA assay using multiwell, protein A-coated, polystyrene plates containing immobilized jasmonate antibodies (Royo et al. 1999). Jasmonate concentrations were determined by competition with JA coupled to alkaline phosphatase, using p-nitrophenylphosphate (Sigma, USA) as reaction substrate, and quantified using a calibration standard curve of methyl jasmonate (10–7500 pmoles). Values were corrected for the estimated recovery of the experiment (55–85%), calculated by spiking samples with known amounts of JA and methyl jasmonate.

Measurements of ethylene production

Ethylene accumulation was measured in a Shimatzu G8a gas chromatograph as described elsewhere (Ruíz-Argüeso et al. 1978). Forty seeds of A. thaliana ecotype Landsberg erecta were sown in 5-ml Erlenmeyer flasks containing 2 ml of liquid MS medium supplemented with 0.5% sucrose, and grown for 16 days (800 μl of fresh medium was added 7 and 14 days after sowing). Prior to treatments, culture medium was removed and replaced with 1 ml of fresh medium. Plants were either wounded or treated with JA 50 μm and/or 250 μg ml–1 chitosan. Approximately 0.5 ml of the head space volume was removed at the indicated times for ethylene determination.

Acknowledgements

We wish to thank Professor C.A. Ryan and J.C. Cabrera for the kind gifts of PIIF and OGA, respectively, Professor T. Ruíz-Argüeso for help in ethylene determinations, and Drs J. Salinas, R. Solano and G. Griffiths for many helpful suggestions and comments on the manuscript. Tomás Cascón provided excellent technical assistance. We also gratefully acknowledge the photographic work by Inés Poveda and Angel Sanz. Seeds from the A. thaliana etr1-3, ein2-5 and ein3-1 mutants were obtained from the Arabidopsis Biological Resource Center at the Ohio State University. Seeds from the A. thaliana coi1 mutant were kindly provided by Dr J.G. Turner (University of East Anglia, Norwich, UK). Financial support was provided by the Programa de Biotecnología from the Spanish Comisión Interministerial de Ciencia y Tecnología (grant BIO96–0532-CO2–01 to J.J.S-S). J.L. was supported by a postdoctoral contract from the Spanish Ministerio de Educación y Ciencia.

Ancillary