- © 1998 American Society of Plant Physiologists
Abstract
Activation of the plant defensin gene PDF1.2 in Arabidopsis by pathogens has been shown previously to be blocked in the ethylene response mutant ein2-1 and the jasmonate response mutant coi1-1. In this work, we have further investigated the interactions between the ethylene and jasmonate signal pathways for the induction of this defense response. Inoculation of wild-type Arabidopsis plants with the fungus Alternaria brassicicola led to a marked increase in production of jasmonic acid, and this response was not blocked in the ein2-1 mutant. Likewise, A. brassicicola infection caused stimulated emission of ethylene both in wild-type plants and in coi1-1 mutants. However, treatment of either ein2-1 or coi1-1 mutants with methyl jasmonate or ethylene did not induce PDF1.2, as it did in wild-type plants. We conclude from these experiments that both the ethylene and jasmonate signaling pathways need to be triggered concomitantly, and not sequentially, to activate PDF1.2 upon pathogen infection. In support of this idea, we observed a marked synergy between ethylene and methyl jasmonate for the induction of PDF1.2 in plants grown under sterile conditions. In contrast to the clear interdependence of the ethylene and jasmonate pathways for pathogen-induced activation of PDF1.2, functional ethylene and jasmonate signaling pathways are not required for growth responses induced by jasmonate and ethylene, respectively.
INTRODUCTION
Higher plants induce various defense responses when they are attacked by microbial pathogens, such as fungi, bacteria, or viruses. These defense responses include suicide of the attacked host cell (the so-called hypersensitive response); the production of antimicrobial secondary metabolites (called phytoalexins); the production of pathogenesis-related (PR) proteins, of which many exert antimicrobial properties; and the production and oxidative cross-linking of cell wall polymers. The efficacy of these defense responses often determines whether plants are susceptible to infection by a pathogen.
Elicitors secreted by or released from microbial invaders are the primary signal for induction of plant defense responses (Ebel and Cosio, 1994). Each pathogen produces a particular mixture of elicitors, which are sometimes accompanied by suppressors, and these molecules interact with receptors on the host cells that further translate the primary signal into particular events in the plasma membrane, the cytosol, and/or the nucleus (Shirasu et al., 1996). Induction of some defense genes requires the generation of secondary endogenous signal molecules (stress hormones) by the challenged cells in the infection site. The secondary signal molecules in turn set in motion signal transduction cascades in receiving cells, eventually leading to activation of pathogen-responsive genes. Secondary signal molecules thus serve to amplify and spread the response of the host after initial recognition of the pathogen. Several secondary signal molecules whose synthesis is increased in response to elicitor recognition and that are involved in the activation of defense genes have been identified. These include H2O2 or other active oxygen species (reviewed in Lamb and Dixon, 1997), salicylic acid (SA; reviewed in Dürner et al., 1997), ethylene (reviewed in Boller, 1991), jasmonic acid (JA; reviewed in Creelman and Mullet, 1997), and abscisic acid (reviewed in Zeevaart, 1988), but it is likely that still others remain to be discovered.
Some of these molecules, such as H2O2, have a high turnover rate and can only activate genes in a limited area immediately surrounding the infection site (Levine et al., 1994). Others, such as SA, are transported over a longer distance and can be spread via the vascular system to organs that are distant from the infection site (Shulaev et al., 1995). Defense genes responding to signal molecules with limited mobility are said to be locally induced, whereas those responding to mobile signals are said to be systemically induced. The response of a pathogen-responsive gene to a particular signal molecule(s) depends primarily on the presence and relative position of binding sites for transcription factors in its promoter (Somssich, 1994).
It is well documented that particular sets of pathogen-responsive antimicrobial protein genes require SA accumulation for their induction by pathogens. Transgenic tobacco plants expressing a bacterial gene (nahG from Pseudomonas putida) for an SA-converting enzyme were unable to accumulate SA when infected, and they failed to induce genes for extracellular PR-1, PR-2, and PR-3 proteins (Friedrich et al., 1995). These plants also had reduced susceptibility to tobacco mosaic virus, Pseudomonas syringae pv tabaci, Phytophthora parasitica, and Cercospora nicotianae (Gaffney et al., 1993; Delaney et al., 1994). Similarly, nahG-expressing Arabidopsis plants infected with pathogens had drastically reduced levels of SA and transcripts for extracellular PR-1 proteins compared with infected control plants (Lawton et al., 1995). These plants were more susceptible to infection by P. syringae pv tomato and Peronospora parasitica (Delaney et al., 1994). The genes for extracellular PR proteins in tobacco and Arabidopsis are strongly induced when SA is applied externally to leaves (Brederode et al., 1991; Ward et al., 1991; Uknes et al., 1992).
Evidence is now accumulating that not all pathogen-responsive antimicrobial protein genes are dependent on SA for their induction. For instance, the induction of tobacco genes for vacuolar PR-2 and PR-3 proteins after infection by Erwinia carotovora was similar in nahG-expressing plants and wild-type plants (Vidal et al., 1997). Pathogen-responsive tobacco genes for vacuolar PR-2, PR-3, and PR-5 proteins are known to be induced much more efficiently by the external application of ethylene than SA (Brederode et al., 1991; Xu et al., 1994; Beffa et al., 1995), and elicitor-induced induction of a vacuolar PR-5 gene in tobacco could be blocked by norbornadiene, which is a blocker of ethylene action (Chang et al., 1995). In Arabidopsis, we have recently identified a gene (PDF1.2) encoding an antifungal peptide belonging to the family of plant defensins (Penninckx et al., 1996; see Broekaert et al., 1995, for a review on plant defensins). This gene is induced by exogenous treatment with either ethylene or methyl jasmonate (MeJA) but not by SA, whereas the opposite was observed for induction of the Arabidopsis PR-1 gene. Inoculation of either wild-type or nahG-expressing Arabidopsis plants with the fungus Alternaria brassicicola caused systemic accumulation of PDF1.2, indicating that pathogen-induced accumulation of this protein is SA independent. On the other hand, both an ethylene-insensitive mutant, called ein2-1, and a jasmonate-insensitive mutant, coi1-1, were unable to accumulate substantial amounts of PDF1.2 in inoculated leaves as well as in noninoculated leaves of inoculated plants (Penninckx et al., 1996).
Our finding that both a component of the ethylene signal transduction pathway (EIN2) and one of the jasmonate signal transduction pathway (COI1) are involved in pathogen-induced activation of the Arabidopsis PDF1.2 gene raises questions about the possible interaction between the ethylene and jasmonate response pathways in this defense response. In this study, we present evidence that concomitant triggering of the ethylene and jasmonate pathways is required for PDF1.2 induction to occur.
RESULTS
ETR1 Is Required for Pathogen-Induced Activation of PDF1.2
It was previously observed that the activation of PDF1.2 in response to challenge with A. brassicicola is abolished in the ethylene-insensitive mutant ein2-1 but only partially in another ethylene-insensitive mutant, etr1-3 (Penninckx et al., 1996). ETR1 encodes an ethylene receptor, whereas EIN2 codes for a downstream ethylene signal transduction component (Kieber, 1997). The weak influence of the etr1-3 mutation on PDF1.2 expression casts some doubts on the involvement of the entire ethylene signal transduction chain and thus on that of ethylene itself.
To clarify this point, Arabidopsis plants carrying the etr1-1 mutation, conferring stronger ethylene insensitivity than the allelic etr1-3 mutation, were tested for their ability to induce PDF1.2 when infected with A. brassicicola. Pathogen-induced expression of PDF1.2 in etr1-1 was apparently blocked at the transcriptional level, because no plant defensin mRNA could be detected in leaves of pathogen-treated plants (Figure 1A). Consistent with this observation, the levels of plant defensins in pathogen-treated and nontreated systemic leaves of etr1-1 mutants after inoculation with A. brassicicola were found to be at least 12-fold lower than the levels in corresponding leaf samples of pathogen-treated wild-type plants (Figure 1B). In a similar test, it was found that pathogen-induced accumulation of plant defensins in etr1-3 mutants was only weakly affected compared with that of wild-type plants (results not shown). This result is consistent with our previous findings (Penninckx et al., 1996). These results point to the requirement of a functional ethylene receptor and thus most likely of ethylene production itself in pathogen-induced activation of PDF1.2.
Ethylene and JA Activate the PDF1.2 Gene via Parallel Signaling Pathways
The observation that pathogen-induced expression of PDF1.2 is blocked both by mutants affected in the ethylene response (ein2-1 and etr1-1) as well as by a mutant affected in its response to jasmonate (coi1-1) implies that ethylene and JA somehow interact with each other to affect expression of this gene (Penninckx et al., 1996). Conceptually, three different models for the interaction between ethylene and JA can be conceived (Figure 2). The first model implies that pathogen recognition leads to increased ethylene production, which in turn results in stimulated JA production and subsequent PDF1.2 activation. The second model is identical to the first, except that the hierarchy between the ethylene and JA signals is switched. In the third model, we propose that ethylene and JA do not act in a sequential manner but rather via parallel pathways, both of which need to be activated for induction of the PDF1.2 gene upon pathogen recognition.
Induction of Plant Defensins in Arabidopsis Wild-Type Plants (Col-0) and in the Ethylene Receptor Mutant etr1-1.
(A) RNA gel blots probed with either PDF1.2 or Tubulin β-1 (Tub) riboprobes. The samples represent 4 μg of total RNA.
(B) Plant defensin (PDF) contents as determined by ELISA, using an antigen affinity-purified anti–Rs-AFP1 (Raphanus sativus–antifungal protein 1) antiserum. Values are means (±se) of three independent determinations.
Arabidopsis plants were inoculated with A. brassicicola (A.bras.) by applying 5-μL drops of a spore suspension (5 × 105 spores per mL) on four lower rosette leaves (five drops per leaf). Control plants were treated identically with 5-μL drops of water (H2O). Pathogen-treated leaves (1°) and nontreated leaves of the same plants (2°) were collected 2 days after inoculation for RNA extractions and 3 days after inoculation for protein extractions. Total RNA and proteins were extracted as described in Methods. The experiment was repeated twice with similar results.
It was previously shown that inoculation of Arabidopsis wild-type plants (ecotype Columbia [Col-0]) with A. brassicicola results in a remarkable increase in the levels of JA in the leaves (Penninckx et al., 1996). The first model predicts that such a pathogen-triggered rise in JA level would not occur in the ethylene-insensitive mutant ein2-1, whereas according to models 2 and 3, the ein2-1 mutation would not affect pathogen-stimulated JA production. To test these divergent predictions, ein2-1 and wild-type plants (Col-0) were inoculated with the pathogen A. brassicicola. After inoculation, JA levels were monitored at different time points. As shown in Figure 3, JA levels in fungus-inoculated wild-type plants started to increase 24 hr after inoculation. This increase was more dramatic 48 hr after inoculation, and JA levels peaked ~72 hr after inoculation. In the ein2-1 mutant, pathogen-stimulated production was clearly not abolished but was even more pronounced relative to that in wild-type plants. When wild-type plants were exposed to 50 ppm ethylene for different periods (0.5, 2, 4, 8, 12, 24, or 48 hr), PDF1.2 transcript levels increased from 8 hr after treatment, whereas no increase in endogenous JA above background levels in air-treated plants occurred until 24 hr after treatment, reaching 10-fold higher levels at 48 hr (results not shown). Hence, all these results are in conflict with the predictions made based on model 1.
To further discriminate between models 2 and 3, ethylene production by wild-type plants and coi1-1 mutants was measured in response to inoculation with A. brassicicola. Model 2 predicts that the coi1-1 mutant would be blocked in its ability to stimulate ethylene production when attacked by a pathogen, whereas model 3 implies that the ethylene response in the coi1-1 mutants would not be reduced versus that of wild-type plants. Inoculation of wild-type plants with A. brassicicola resulted in ethylene production levels that were approximately three times higher than those in mock-inoculated plants 36 hr after inoculation (Figure 4). An increase in ethylene production levels was also observed in the coi1-1 mutant plants 36 hr after inoculation with A. brassicicola, reaching levels that were approximately fivefold higher compared with levels in mock-inoculated coi1-1 mutants. Because pathogen-stimulated ethylene production clearly was not abolished in the coi1-1 mutants, we conclude that model 2 is not valid. Also in conflict with model 2 is our observation that application of drops containing 50 μM MeJA on leaves of Arabidopsis plants does not result in increased ethylene emission relative to control plants whether measured 0.5, 2, 4, 8, 12, 24, or 48 hr after treatment, whereas PDF1.2 transcript levels were already increased 4 hr after treatment (results not shown).
An alternative way to verify the validity of the three different models is to check whether the PDF1.2 gene can be induced in MeJA-treated ein2-1 mutants or in ethylene-treated coi1-1 mutants. Treatment of wild-type plants with 50 μM MeJA resulted in a marked elevation of plant defensins and PDF1.2 transcripts relative to those in untreated wild-type plants. In contrast, the content of plant defensins and their transcripts in ein2-1 mutants was not elevated relative to those in untreated plants (Figures 5A and 5B). This observation is in conflict with model 1, which predicts that jasmonate-induced PDF1.2 accumulation would not be affected by the ein2-1 mutation. Exposure of wild-type plants to ethylene boosted protein and mRNA levels of plant defensins compared with air-exposed plants. However, no accumulation of plant defensins could be observed in ethylene-treated coi1-1 mutants (Figures 5A and 5B), which contradicts the predictions of model 2. Hence, model 3, proposing parallel ethylene and jasmonate signaling pathways, is the only model that is in agreement with all observations.
Three Alternative Models for the Interaction between Ethylene and Jasmonate Signals during Activation of the PDF1.2 Gene in Pathogen-Challenged Arabidopsis Plants.
Model 1 implies that pathogen infection initially stimulates production of ethylene, which subsequently stimulates production of jasmonates, which in turn activates PDF1.2. Model 2 implies that pathogen infection initially leads to enhanced production of jasmonates, which subsequently triggers elevated production of ethylene, which in turn controls PDF1.2 expression. Model 3 predicts that pathogen infection results in simultaneous production of ethylene and jasmonates, which are both required for induction of PDF1.2. The positions of the gene products EIN2 and COI in the different pathways are indicated by short arrows.
Synergistic Induction of PDF1.2 by Ethylene and Jasmonate
Our finding that JA and ethylene act via parallel pathways to activate the PDF1.2 gene suggests that combinations of JA and ethylene may have a synergistic effect on the activation of this gene. It was observed that treatment of Arabidopsis plants, grown under sterile conditions in agar, with either ethylene or MeJA resulted in a much weaker induction of PDF1.2 compared with the same treatments applied to plants grown on nonsterile potting soil (Figures 6A and 6B). For instance, treatment of sterile plants with 5 ppm of ethylene did not cause a detectable induction of PDF1.2, whereas induction was clearly observed in nonsterile plants. However, combined application of 5 ppm of ethylene and 0.05 μM MeJA boosted the level of PDF1.2 transcripts in sterile plants to approximately the same level as that in sterile plants treated with 50 μM MeJA alone. When applied in combination to sterile plants, 5 ppm of ethylene and 5 μM MeJA caused a rise in PDF1.2 transcripts that was significantly higher relative to separate treatments with 50 ppm of ethylene or 50 μM MeJA and that was of the same order of magnitude as that observed for nonsterile plants treated with either 50 ppm of ethylene or 50 μM MeJA alone (Figure 6A). Similar observations were made using ELISA assays to detect PDF1.2 (Figure 6B). Treatment of nonsterile plants with combinations of ethylene and MeJA only marginally increased PDF1.2 levels compared with treatments with either hormone applied separately (results not shown).
Activation of PDF1.2 by Superoxide Anion–Generating Paraquat Requires Functional Ethylene and Jasmonate Response Pathways
One of the most rapid events observed during plant–pathogen interactions is the production of reactive oxygen species, which are initially in the form of superoxide anion, by an NADPH oxidase complex (Lamb and Dixon, 1997). We have previously shown that treatment of Arabidopsis leaves with subherbicidal levels of the superoxide-generating compound paraquat can induce expression of PDF1.2, whereas no activation of the SA-dependent PR-1 gene was observed at the doses used (Penninckx et al., 1996). Therefore, we investigated whether signal components of the ethylene and JA response pathways are required for paraquat-induced activation of PDF1.2. As shown in Figures 5A and 5B, accumulation of plant defensins and their transcripts was strongly reduced in paraquat-treated ein2-1 mutants compared with similarly treated wild-type plants, whereas this accumulation was below the level of detection in paraquat-treated coi1-1 mutants. Hence, these results strongly suggest that reactive oxygen species act upstream of ethylene and JA.
EIN2 and COI1 Are Not Involved in Developmental Cues Influenced by Jasmonate and Ethylene, Respectively
In addition to being involved in defense responses, ethylene and JA are also implicated in developmental and growth processes of plants (Creelman and Mullet, 1997; Kende and Zeevaart, 1997). Growth of Arabidopsis seedlings in the presence of ethylene causes drastic inhibition of hypocotyl elongation, radial swelling of hypocotyls, and exaggeration of the curvature of the apical hook (Kieber, 1997), whereas MeJA causes drastic reduction of root elongation in Arabidopsis seedlings (Staswick et al., 1992). We investigated whether the ethylene-insensitive ein2-1 mutant was insensitive to MeJA-induced inhibition of root elongation, as is coi1-1. We also investigated whether the JA-insensitive coi1-1 mutant was insensitive to inhibition of hypocotyl elongation caused by ethylene, as is ein2-1. Root elongation of MeJA-treated ein2-1 seedlings and hypocotyl elongation of ethylene-treated coi1-1 seedlings were inhibited to the same extent as they were for similarly treated wild-type seedlings (Figure 7). Hence, EIN2 is implicated in growth processes influenced by ethylene but not in those influenced by JA, with the reverse being true for COI1.
Time Course of Endogenous JA Content in Arabidopsis Wild-Type Plants (Col-0) and Ethylene-Insensitive Mutants (ein2-1) When Inoculated with A. brassicicola (Closed Symbols) or Mock Inoculated with Water (Open Symbols).
JA levels, expressed as nanograms of free JA per gram fresh weight of tissue, were determined by HPLC with extracts from Arabidopsis leaves collected 0, 12, 24, 36, 48, 60, 72, 84, and 96 hr after inoculation with 5-μL drops (five drops per leaf) of either an A. brassicicola spore suspension at 5 × 105 spores per milliliter or distilled water. Each data point is the average of measurements on two separate sets of three plants each. The experiment was repeated once with similar results.
Time Course of Ethylene Production Levels in Arabidopsis Wild-Type Plants (Col-0) and Jasmonate-Insensitive Mutants (coi1-1) When Inoculated with A. brassicicola (Closed Symbols) or Mock Inoculated with Water (Open Symbols).
Ethylene production, expressed as nanomoles of ethylene per hour per gram fresh weight of tissue, was determined 0, 12, 24, 36, 48, 60, 72, 84, and 96 hr after inoculation with 5-μL drops (five drops per leaf) of either an A. brassicicola spore suspension at 5 × 105 spores per milliliter or distilled water. Each data point is the average (±sd) of measurements on two separate sets of two plants each. The experiment was repeated twice with similar results.
DISCUSSION
We demonstrated previously that pathogen-induced activation of the Arabidopsis PDF1.2 gene, which encodes an antifungal plant defensin, is abolished in a mutant carrying the ein2-1 mutation. EIN2 is a signal transduction component involved in the response to ethylene and acts downstream of the ethylene receptor, of which ETR1 forms a part (Kieber, 1997). Our finding that PDF1.2 induction is also abolished in the disfunctional etr1-1 allele implies that not only downstream response elements, such as EIN2, but also the ethylene receptor and hence most probably the entire signal transduction chain are involved in PDF1.2 activation. It should be noted that the impact of the etr1-3 mutation on pathogen-induced PDF1.2 activation is much weaker compared with that of the etr1-1 or ein2-1 mutations. This observation has some implications for the interpretation of the results of Bent et al. (1992), who showed that the development of chlorotic symptoms in Arabidopsis plants inoculated with either Pseudomonas or Xanthomonas pathogens was severely reduced in ein2-1 mutants compared with wild-type plants but not in etr1-3 (previously called ein1-1) mutants. Because the etr1-3 mutation has only a weak impact on the induction of pathogen-responsive genes, these experiments might be inconclusive regarding the involvement of ETR1 in the induced response to pathogens. Nevertheless, the experiments of Bent et al. (1992) have been translated by several authors into a model in which ETR1 is involved in the ethylene response but not in the pathogen response, whereas EIN2 is involved in both responses (Ecker, 1995; Kieber, 1997). Obviously, our data call this model into question.
Induction of Plant Defensins in Arabidopsis Wild-Type Plants (Col-0), Ethylene-Insensitive Mutants (ein2-1), and Jasmonate-Insensitive Mutants (coi1-1).
(A) RNA gel blots probed with either PDF1.2 or Tubulin β-1 (Tub) riboprobes. The samples represent 4 μg of total RNA. (B) Plant defensin (PDF) contents as determined by ELISA, using an antigen affinity-purified anti–Rs-AFP1 antiserum. Values are means (±se) of three independent determinations.
Arabidopsis leaves were inoculated with 5-μL drops (five drops per leaf) of water (con), paraquat (PQ; 25 μM), and MeJA (45 μM in 0.1% [v/v] ethanol). Ethylene treatment was performed by placing plants in an airtight chamber with an ethylene concentration of 50 ppm. All leaf samples were collected 48 hr after the initiation of treatment. The experiment was repeated two times with similar results.
The lack of pathogen-induced PDF1.2 activation in the jasmonate perception mutant coi1-1 implies that not only the ethylene response pathway but also a jasmonate-based signaling pathway is required for the induction of this gene. The fact that PDF1.2 induction could not be triggered by ethylene treatment of the coi1-1 mutant or by jasmonate treatment of the ein2-1 mutant clearly points to the conclusion that the ethylene and jasmonate signaling pathways are triggered independently and concomitantly and rules out the possibility of sequential activation of the jasmonate pathway by ethylene, or vice versa. We did not observe any elevated ethylene production after treatment of Arabidopsis plants with exogenous MeJA. This is in clear contrast with the situation in tomato leaves, where application of JA results in a rapid but transient increase in ethylene emission (O'Donnell et al., 1996). On the other hand, we observed that treatment of Arabidopsis plants with exogenous ethylene resulted in a rise in JA levels, as previously reported by Laudert and Weiler (1998) for treatment of Arabidopsis with the ethylene-releasing compound ethephon. Increased JA production triggered by ethylene treatment, however, was a rather late response and clearly occurred after PDF1.2 induction upon ethylene exposure, thus arguing against ethylene-stimulated JA production being the primary stimulus for PDF1.2 induction. Ethylene-stimulated JA production may nevertheless contribute to and enhance PDF1.2 activation in the late stages of infection by a pathogen. Although triggering of both the ethylene and jasmonate signaling pathways appears to be the main cause of pathogen-induced expression of PDF1.2, our experiments do not show at which point both pathways converge to affect expression of this gene. It is possible that transcription factors activated by the ethylene response pathway cooperate with transcription factors activated independently by the jasmonate response pathway to form a functional transcription initiation complex. Alternatively, signal transduction components activated upon jasmonate perception may alter the sensitivity of cells to the ethylene signal input or vice versa. It is known, for instance, that EREBP1, a gene encoding a transcription factor that is involved in the transcription of ethylene-activated genes, is induced upon treatment of tobacco plants with jasmonate (Horvath et al., 1998). A jasmonate-stimulated increase in EREBP1 levels might increase the responsiveness of the cell to the ethylene signal input.
Synergistic Induction of Plant Defensins by Ethylene and Jasmonate.
(A) RNA gel blots probed with either PDF1.2 or Tubulin β-1 (Tub) riboprobes. The samples represent 4 μg of total RNA.
(B) Plant defensin (PDF) contents as determined by ELISA, using an antigen affinity-purified anti–Rs-AFP1 antiserum. Values are means (±se) of three independent determinations.
Arabidopsis leaves from either agar-grown sterile plants or soilgrown nonsterile plants were inoculated with 5-μL drops (five drops per leaf) of MeJA (at the indicated concentration in 0.1% [v/v] ethanol) or 0.1% (v/v) ethanol (EtOH) either in the presence or the absence of ethylene (at the concentration indicated) in the atmosphere of an airtight chamber. All leaf samples were collected 48 hr after the initiation of treatment. The experiment was repeated three times with similar results.
The fact that treatment of plants with either ethylene or MeJA results in PDF1.2 induction (Penninckx et al., 1996) is in apparent conflict with our model proposing concomitant and cooperative ethylene and jasmonate signaling events. Indeed, this model (Figure 2) implies that only treatment with a combination of ethylene and jasmonate can effectively trigger PDF1.2 expression. Consistent with the model, however, we found that plants grown under sterile conditions and treated with either jasmonate or ethylene alone show very weak PDF1.2 induction, whereas application of both hormones simultaneously leads to strong synergistic PDF1.2 induction. We do not know why plants grown under nonsterile conditions behave differently in this respect. Obviously, the environmental conditions of plants grown on agar in a closed container and plants grown in potting soil differ in many parameters, including mineral nutrition, water availability, and gas exchange. Any of these parameters might influence the physiological response of plants to applied hormones. One other difference to be considered is the lack of microorganisms in the rizosphere of plants grown on agar under sterile conditions. It has previously been observed that the interaction between Arabidopsis roots and particular soil-borne Pseudomonas species leads to a systemic change in the resistance of the plants to pathogens such as Fusarium oxysporum f sp raphani (Pieterse et al., 1996). Such an interaction might alter the sensitivity of the plants to both jasmonate and ethylene, or alternatively, it might alter the endogenous levels of these hormones. In support of the latter idea, we found that plants grown in potting soil have basal ethylene emission levels that are at least fivefold higher relative to those of plants of the same age grown axenically (I.A.M.A. Penninckx, unpublished results). Higher basal ethylene production levels in soil-grown plants may act synergistically with externally applied MeJA to boost PDF1.2 levels.
Synergistic interactions between ethylene and MeJA have previously been observed for the induction of the gene encoding pathogen- and wound-responsive osmotin (vacuolar PR-5) in tobacco (Xu et al., 1994). The regulation of this gene may turn out to be very similar to that of PDF1.2 in Arabidopsis, although the latter is not induced by wounding (Penninckx et al., 1996). In tomato, induction of the wound-responsive proteinase inhibitor Pin2 gene occurs when plants are treated with JA but not with ethylene. However, treatment of plants with inhibitors of ethylene biosynthesis or ethylene perception abolished the responsiveness of this gene to both wounding and JA treatment (O'Donnell et al., 1996). Hence, the ethylene signal pathway appears to cooperate somehow with the JA response pathway to effect expression of a stress-related gene in tomato.
ein2-1 and coi1-1 Plants Are Not Affected in Developmental Cues Influenced by Jasmonate and Ethylene, Respectively.
Wild-type (Col-0), ein2-1, and coi1-1 plants were grown under sterile conditions with a 12-hr photoperiod on Murashige and Skoog (MS) medium in the absence or presence of 100 μM MeJA (top). The same ecotypes were grown under sterile conditions in darkness in airtight containers in the absence or presence of 50 ppm ethylene in the atmosphere (bottom).
Ethylene has also been shown to affect the sensitivity of Arabidopsis plants to SA. When Arabidopsis plants were treated with ethylene, lower concentrations of SA were required to induce the PR-1 gene (Lawton et al., 1994). However, in this case, the requirement for ethylene appears to be less strict because etr1-1 and ein2-1 mutants still induce the PR-1 gene when challenged by a pathogen or treated with SA (Lawton et al., 1994, 1995; Penninckx et al., 1996).
In this study, we present genetic evidence for the strict requirement of both the jasmonate and the ethylene signaling pathways for the induction of a pathogen-responsive gene. Interestingly, we found that the JA and ethylene response pathways do not affect ethylene- and JA-mediated developmental processes, respectively, such as hypocotyl and root elongation. Our observations in this respect are in line with those of Berger et al. (1996), who reported that Arabidopsis JA response mutants other than coi1-1 did not show an altered sensitivity to ethylene-induced hypocotyl growth reduction. Hence, ethylene and JA regulate at least three distinct responses in Arabidopsis, depending on whether they act alone or in combination (Figure 8). The effect of ethylene on hypocotyl elongation has been suggested to involve interaction with auxins, possibly based on inhibition of polar auxin transport by ethylene (Hobbie, 1998). Cooperative interactions between signaling response pathways may be seen as means developed by plant species to increase the number of distinct gene repertoires that can be controlled by a limited set of signaling molecules and hence to increase behavioral plasticity.
METHODS
Biological Material
The ethylene response mutants ein2-1 (Guzmán and Ecker, 1990) and etr1-1 (Bleecker et al., 1988; Chang et al., 1993) were obtained from the Arabidopsis Biological Resource Center (Columbus, OH; accession numbers CS3071 and CS237, respectively). The jasmonate response mutant coi1-1 (Feys et al., 1994) was obtained from J. Turner (University of East Anglia, Norwich, UK). Because this mutation is recessive and causes male sterility, we identified coi1 mutants in F2 plants grown from seed from selfed COI1/coi1 hemizygous plants. The F2 population was therefore subjected to different treatments as follows. Leaves from each individual plant were collected separately, and the plants were then grown until seed set. Individuals that did not form siliques were identified as having the coi1/coi1 genotype. All mutant and transgenic lines listed above were derived from the Arabidopsis thaliana Columbia (Col-0) ecotype. Growth and spore harvesting of the fungus Alternaria brassicicola (MUCL 20297; Mycothèque Université Catholique de Louvain, Louvain-la-Neuve, Belgium) were done as described previously (Broekaert et al., 1990).
Three Distinct Responses Controlled by Ethylene and Jasmonate in Arabidopsis.
Ethylene is involved in the control of hypocotyl elongation and curvature and possibly in interactions with other hormones such as auxins, but not with jasmonates. Ethylene and jasmonates interact to regulate the expression of particular defense-related genes such as PDF1.2 upon pathogen perception. Jasmonates are involved in the control of root elongation, possibly in interaction with other hormones excluding ethylene.
Plant Growth Conditions, Chemical Application, and Inoculation
Arabidopsis seeds were sown on flower potting compost containing a macronutrient supplement (Asef, Didam, The Netherlands) in Petri dishes. Seeds were vernalized for 2 days at 4°C after sowing. After 5 days of incubation in a growth chamber (20°C daytime temperature and 18°C nighttime temperature, with a 12-hr photoperiod at a photon flux density of 100 μE m−2 sec−1), seedlings were transferred to pots (5 × 4 × 4 cm) containing potting compost supplemented with macronutrients and grown under the same conditions as given above. Irrigation was with tap water via the trays carrying the pots. Surface-sterilized seeds were sown on sterile Murashige and Skoog medium (Sigma Chemical Co.), in some cases including methyl jasmonate (MeJA), as indicated in Results, in sterile translucent polypropylene pots. Seeds were vernalized for 2 days at 4°C after sowing and incubated under the same conditions as soil-grown plants. Plants were 4 weeks old when treated.
Paraquat (25 μM), MeJA (45 μM in 0.1% [v/v] ethanol), and 0.1% (v/v) ethanol were applied at the concentrations indicated as 5-μL droplets on leaves (five drops per leaf for soil-grown plants; one drop per leaf for sterile plants). The stock solution of MeJA was 45 mM in ethanol. Ethylene treatment was performed by placing pots in an airtight translucent chamber in which gaseous ethylene was injected via a silicon rubber septum. The ethylene concentration in the chamber was verified by gas chromatography. Control plants for the ethylene experiment were placed in an identical chamber without ethylene.
Inoculation with A. brassicicola was done by applying 5-μL drops of a spore suspension (density of 5 × 105 spores per mL in distilled water) on four lower rosette leaves (five drops per leaf). Control plants were treated identically with water droplets. The plants with drops of spore suspension or water were placed randomly (if different genotypes were treated simultaneously) in a propagator flat with a clear polystyrene lid and kept at high humidity for 2 days to stimulate infection by hyphal germlings. Thereafter, lids were removed, and the plants were incubated further until the harvesting of leaf material. The treated leaves (cotyledons and rosette leaf numbers 1 to 4) were collected separately from the nontreated leaves (rosette leaf numbers 5 to 8) of the same plants. The isolate of A. brassicicola and inoculation conditions used here caused limited brown necrotic lesions under the drops of spore suspension within 48 hr after inoculation, and these lesions failed to spread further.
RNA Gel Blot Analysis
RNA was extracted by the phenol–LiCl method according to Eggermont et al. (1996) from tissues frozen in liquid nitrogen and stored at −80°C. RNA gel blotting was performed as described previously (Penninckx et al., 1996). Blots were prehybridized, hybridized with digoxigenin-labeled antisense RNA probes (either derived from PDF1.2- or tubulin β-1–encoding expressed sequence tags), and developed by immunochemiluminescence, as described previously (Penninckx et al., 1996).
ELISA
Proteins were isolated from frozen leaf material, as described by Penninckx et al. (1996). Protein concentrations were determined in the crude extracts according to Bradford (1976) by using BSA as a standard. After heat treatment (10 min at 80°C) of the extract, the heat-stable soluble protein fraction was analyzed in a competition ELISA, as outlined previously (Penninckx et al., 1996).
Quantitative Analysis of Jasmonic Acid and Ethylene Production
The extraction and quantitative analysis of jasmonic acid (JA) were performed as previously described (Penninckx et al., 1996). For measurements of ethylene emission, 4-week-old soil-grown Arabidopsis plants were either inoculated with A. brassicicola as described above or mock inoculated with water. After 0, 12, 24, 36, 48, 60, 72, 84, and 96 hr of incubation, the plants were removed from their soil substrate, their roots were washed with tap water, and each plant was placed in a 20-mL glass bottle closed with a rubber cap. After 3 hr, a 1-mL gas sample was withdrawn from the headspace with a syringe and analyzed by gas chromatography on a Delsi Nermag 200 (Delsi, Paris, France) instrument equipped with a Poropack R (vinylpyrolidone) column and a flame ionization detector.
ACKNOWLEDGMENTS
This work was supported in part by a grant from the Fonds voor Wetenschappelijk Onderzoek Vlaanderen (FWO) and in part by a grant from the Swiss National Science Foundation for the jasmonic acid measurements. I.A.M.A.P. is the recipient of a predoctoral fellowship of the Vlaams Instituut voor Bevordering van het Wetenschappelijk-Technologisch Onderzoek in de Industrie (IWT). B.P.H.J.T. is a Research Assistant of the FWO. We thank Dr. John Turner for providing seed from Arabidopsis mutants and Dr. Maurice De Proft for advice on ethylene emission measurements.
- Received May 7, 1998.
- Accepted October 12, 1998.
- Published December 1, 1998.
REFERENCES
