Regulation of Jasmonate-mediated Plant Responses in Arabidopsis

Abstract

Jasmonates (JAs) are signalling molecules that play a key role in the regulation of metabolic processes, reproduction, and defence against pathogens and insects. JAs regulate responses that are both local and systemic, and which are affected by outputs from signalling pathways regulated by ethylene, salicylic acid and auxin. This is a review of recent advances in our understanding of the regulation of JA perception in Arabidopsis thaliana, the different signalling functions of biologically active JAs, the post-translational control of JA responses leading to substantial transcriptional reprogramming, and the influence of other signalling pathways of systemic JA responses.

JASMONATES AS SIGNALLING MOLECULES

Regulation of changes in the relative levels of biologically active signalling molecules named jasmonates (JAs) contribute to the control of metabolic, developmental and defensive processes in plants (Weber et al., 1997). Jasmonic acid (JA) is a terminal product of the octadecanoid pathway and several intermediates in the pathway for JA biosynthesis are biologically active, as are some derivatives of JA (Fig. 1 and Turner et al., 2002). These different signalling molecules affect a variety of plant processes (Creelman and Mullet, 1997), including fruit ripening, production of viable pollen, root growth, tendril coiling, response to wounding and abiotic stress, and defences against insects and pathogens. Figure 1 illustrates a model for the biosynthesis of JAs following wounding or pest attack and during pollen development.

Fig. 1. Model for the octadecanoid biosynthetic pathway following wounding or pest attack and in pollen development. The activation of a phospholipase (PLD or DAD1) may result from the elicitation of a membrane receptor. Abbreviations for enzyme names are in bold and underlined: AOC, allene oxide cyclase; AOS, allene oxide synthase; DAD1, defective anther dehiscence1; JMT, S -adenosyl-l-methionine:jasmonic acid carboxyl methyltransferase; LOX, lipoxygenase; OPR3, OPDA reductase3; PLD, phospholipase. Abbreviations for names of intermediates are in bold: 13-HPOT, 13-hydroperoxylinolenic acid; OPC 8 : 0, 3-oxo-2(2′pentenyl)-cyclopentane-1-octanoic acid; OPDA, 12-oxo-phytodienoic acid.

Our understanding of jasmonate signalling is complicated by the presence of multiple acyclic or cyclic oxidation products derived from the catabolism of fatty acids—oxylipins—that regulate many defence and developmental pathways in plants. The activity of JA, its precursor 12-oxo-phytodienoic acid (OPDA) and other oxylipins (Krumm et al., 1995; Bate and Rothstein, 1998) as signals for defence suggests that host responses to attackers may be regulated by a complex mix of signals, which has been termed the oxylipin signature (Weber et al., 1997). Hause et al. (2000) showed that distinctive oxylipin profiles are produced by different external stimuli and by developmental cues. Overexpression of a JA methyltransferase gene increases resistance to Botrytis cinerea, suggesting that MeJA induces pathogen defence responses (Seo et al., 2001); the methylation of JA to its methylester (MeJA) is catalysed by an S-adenosyl-L-methionine : JA carboxyl methyltransferase (JMT; Fig. 1). An important advance in our understanding of jasmonate-signalled responses has been made recently by Stintzi et al. (2002), who used a biochemical genetic approach to show that OPDA (a precursor of JA, see above and Fig. 1) is a physiological signal for defence, which induces broad-spectrum resistance in the absence of JA.

ROLES OF JASMONATES AND REGULATION OF JASMONATE BIOSYNTHESIS

Our understanding of the perception of stresses and developmental cues is limited. Responses mediated by JAs can be triggered by a series of diverse abiotic stresses and elicitor molecules (Doares et al., 1995; Kramell et al., 1995; Parchmann et al., 1997; Leon et al., 2001; Turner et al., 2002). Presumably, either different stimuli interact with a common receptor regulating JA biosynthesis or, more likely, interact with different receptors which regulate signalling pathways that converge on the pathway for JA biosynthesis.

In arabidopsis, JAs inhibit root elongation (Staswick et al., 1992) and are required for pollen development, anther dehiscence (Feys et al., 1994; McConn and Browse, 1996; Sanders et al., 2000; Stintzi and Browse, 2000) and defence against insects (McConn et al., 1997) and necrotrophic pathogens (Thomma et al., 1999). JA is also required for protection from ozone damage (Overmyer et al., 2000; Rao et al., 2000) and is the primary signal in protective alkaloid production in Eschscholtzia californica cell cultures (Byun, 2000).

The production of JAs ultimately leads to the induction of many genes, including those for vegetative storage proteins (VSPs; Benedetti et al., 1995), a thionin (Thi2.1; Epple et al., 1995; Vignutelli et al., 1998) and a plant defensin (PDF1.2; Penninckx et al., 1998). JAs also induce transcription of genes that regulate JA synthesis (Fig. 1), including DAD1, LOX2, AOS, OPR3 and JMT (Heitz et al., 1997; Laudert and Weiler 1998; Mussig et al., 2000; Ishiguro et al., 2001; Seo et al., 2001). Microarray analysis confirmed that five out of 41 genes responding to JA are required for JA biosynthesis, substantiating the extent of a positive feedback regulatory system for JA biosynthesis (Sasaki et al., 2001).

The arabidopsis mutant constitutive expression of vegetative storage protein (cev1) has constitutive production of JA and ethylene, constitutive expression of PDF1.2, Thi2.1 and the chitinase CHI (Ellis and Turner, 2001) and has enhanced defences against the pathogens Erysiphe cichoracearum and Pseudomonas syringae and improved resistance against the aphid Myzus persicae (Ellis et al., 2002b). Ellis et al. (2002a) have shown that cev1 acts at an early step in the stress perception/transduction pathway, and induces JA and ethylene synthesis. The cev1 mutant phenotype is partially suppressed in the coronatine insensitive 1 (coi1) and in the ethylene resistant 1 (etr1) mutant backgrounds, and the triple mutant, cev1;coi1;etr1 is wild type except for slightly shorter roots (Ellis et al., 2002a). cev1 has been mapped by positional cloning and encodes the cellulose synthase CeSA3 (Ellis et al., 2002a). This suggests that the cell wall might mediate JA- and ethylene-dependent stress and defence responses.

Studies on JA biosynthetic mutants have shown that JAs have critical roles in pollen maturation and dehiscence and wound-induced defence against biotic attacks. Allene oxide synthase (AOS), a cytochrome P450 enzyme (CYP74A), catalyses dehydration of 13-(S)-hydroperoxylinolenic acid to 12,13-epoxy-linolenic acid (allene oxide), the first committed step in JA synthesis (Fig. 1). A knock-out mutant defective in CYP74A was isolated (Park et al., 2002), which contained reduced amounts of JA and was male sterile. Male sterility could be rescued by exogenous application of methyl jasmonate or by complementation with constitutive expression of the wild-type AOS gene—a gene that regulates JA synthesis (see above). Plants in which AOS was overexpressed had enhanced wound-induced induction of Arabidopsis thaliana vegetative storage protein 2 (AtVSP2), demonstrating the role of AOS as a modulator of the wound signal.

MUTANTS IDENTIFIED IN THE JA SIGNAL TRANSDUCTION PATHWAY

Mutant screens for insensitivity to coronatine (a structural analogue of methyl jasmonate) and to methyl jasmonate itself (Staswick et al., 1992; Feys et al., 1994), and a screen for mutants that do not express a luciferase reporter for the VSP promoter in the presence of JA (Ellis and Turner, 2001), have been conducted to saturation. These screens should have recovered mutants in receptors for coronatine or JA that would be insensitive to these compounds. However, the screens identified only alleles of the genes coronatine insensitive (coi1; Xie et al., 1998), jasmonate insensitive (jin1 and 4; Berger et al., 1996) and jasmonate resistant (jar1; Staswick et al., 1998). COI1 and JAR1 have been isolated but neither defines an obvious receptor for JA. It is assumed, therefore, that there is redundancy amongst the JA receptors. COI1 encodes an F-box protein related to TIR1, a component of an ubiquitin-like E3 complex that is involved in plant auxin response (Ruegger et al., 1998; Xie et al., 1998). Analysis of jar1-1 has shown that this locus is involved in protection against a variety of stresses that plants encounter, such as resistance to the opportunistic soil fungus Pythium irregulare (Staswick et al., 1998), systemic resistance against various other pathogens (van Loon et al., 1998; Clarke et al., 2000) and limiting damage from ozone exposure (Overmyer et al., 2000; Rao et al., 2000). jar1-1 plants are fertile, indicating that JAR1 is not required for all jasmonate responses. Recently, positional cloning indicated that JAR1 belongs to a multigene family that includes the auxin-induced soybean GH3 (Abel and Theologis, 1996; Staswick et al., 2002). Fold prediction modelling and an in vitro biochemical assay revealed that JAR1 is structurally related to the firefly luciferase superfamily of adenylate-forming enzymes. Surprisingly, therefore, JAR1 apparently modifies JA and the JA-insensitive jar1 phenotype, indicating that adenylation of JA is required for some but not all JA responses. Curiously, the suppressor of constitutively photomorphogenic 1 (cop1), fin219 (Hsieh et al., 2000), maps to the JAR1 locus, but displays no increase to resistance to MeJA.

Lipoxygenases (LOXs) catalyse the oxygenation of fatty acids to their hydroperoxy derivatives (Fig. 1). Jensen et al. (2002) used the luciferase reporter of the JA-responsive LOX2 (Bell et al., 1995) promoter to screen for mutants with aberrant expression of luciferase activity. Three recessive mutants that underexpress the reporter, designated jue1, 2 and 3, as well as two recessive mutants that overexpress the reporter, designated joe1 and 2, were isolated. Genetic analysis indicated that reporter overexpression in the joe mutants requires COI1, suggesting that they act prior to COI1 to regulate LOX2 expression. joe1 responded to MeJA with increased anthocyanin accumulation, while joe2 responded with decreased root growth inhibition. In addition, wild-type induction of the reporter and endogenous LOX2 expression by the serine-threonine protein kinase inhibitor staurosporine was deficient in joe2. This indicates that the joe2 mutation may lead to inactivation of a kinase or its substrate, while joe1 may act prior to the phosphorylation event in a JA signal pathway.

AN E3 UBIQUITIN LIGASE REGULATES JA RESPONSES IN ARABIDOPSIS

The coi1 mutants are unresponsive to growth inhibition by MeJA, are male sterile, fail to express JA-regulated genes that code for vegetative storage proteins, VSPs (Benedetti et al., 1995), thionin, Thi2.1, and the plant defensin PDF1.2, and are susceptible to insect herbivory and to pathogens (McConn et al., 1997; Thomma et al., 1998). coi1-16 is a recently cloned allele that was isolated for failure to activate the vegetative storage protein 1 (VSP1) promoter (Ellis and Turner, 2002). Fertility of coi1-16 is temperature-sensitive. Further alleles of coi1 have also been isolated in screens for susceptibility to bacterial disease (Kloek et al., 2001). COI1 gene encodes a 66 kD protein containing an N-terminal F-box motif, and a leucine rich repeat (LRR) domain (Xie et al., 1998).

F-box proteins are components of SCF (SKP1, CDC53p/CUL1 F-box protein) complexes (Bai et al., 1996), where they function as specific receptors targeting proteins to ubiquitin-mediated proteolysis (Glickman and Ciechanover, 2002). Gagne et al. (2002) identified 694 potential F-box genes in A. thaliana, making this gene superfamily one of the largest currently known in plants, although a role in the SCF complex has been demonstrated for only a handful of these genes (Gray et al., 1999; Samach et al., 1999; Dieterle et al., 2001; Woo et al., 2001; Devoto et al., 2002).

One focus of present research is to identify the target proteins for the ubiquitin (Ub)/proteasome-mediated degradation. Increasing evidence suggests that this proteolytic pathway plays an integral role in plant development, responsiveness to hormones, light, sucrose and defence by selectively removing abnormal polypeptides and short-lived regulatory proteins (Ciechanover, 2000; Ellis et al., 2002c). Immunoprecipitates of epitope-tagged COI1 from transgenic arabidopsis plants co-precipitate with components of the SCF complex, SKP1 proteins, cullin and Rbx1, confirming that COI1 forms an SCF^COI1 complex in vivo (Devoto et al., 2002; Xu et al., 2002). COI1 is therefore expected to form a functional E3-type ubiquitin ligase in plants. A key question is what is the substrate COI1 recruits for ubiquitination? Devoto et al. (2002) used a yeast two-hybrid screen with COI1 as bait, and co-immunoprecipitation with epitope-tagged COI1, to demonstrate that the histone deacetylase RPD3b (reduced potassium dependency; Murfett et al., 2001) and the small subunit of RUBISCO bind to COI1 in vivo. These two proteins might be potential substrates for COI1-mediated ubiquitination and possibly mediate JA response. Such roles remain to be elucidated. Significantly, in yeast, RPD3b binds to COI1, but not to the gene product of coi1-16 (Ellis and Turner, 2002) in which the mutation is a L245F in a leucine-rich repeat of the COI1 protein.

Histone deacetylase functions to maintain the balance between acetylation and deacetylation of histones and forms an important mechanism in the regulation of gene transcription in eukaryotes (Hassig et al., 1997; Pazin et al., 1998; Lusser et al., 2001). Histone deacetylation is believed to decrease accessibility of chromatin to the transcription machinery (Fig. 2A). Mammalian homologues of yeast proteins known to interact with each other and to be involved in a ubiquitin signalling pathway, p97/VCP/Cdc48p and a phospholipase A2-activating protein, have been shown to associate with a cytoplasmic murine histone deacetylase 6 (mHDAC6) (Seigneurin-Berny et al., 2001), establishing a link between protein acetylation and protein ubiquitination. According to one model, a target regulator (R) that might be a histone deacetylase could form a hormone-sensitive multimeric complex which might contribute to suppress the expression of JA response genes, via an SCF^COI1 complex. This would require a JA-induced signal to modify the regulator, possibly by phosphorylation, which might be recruited by COI1. Subsequent degradation of the target regulator would permit induction of JA responsive genes (Fig. 2B). It has been observed that SCF E3 ubiquitin ligases can modify membrane transport, subcellular localization, transcription and protein kinase activity via monoubiquitylation (Pickart et al., 2001). Accordingly, the SCF^COI1 complex might modify the activity of a target regulator or have a dual role in its activation and successive degradation via ubiquitination (Fig. 2B).

Fig. 2. A, Histone deacetylases form complexes with other proteins. Model of a histone deacetylase complex in plants. The hypothetical complex contains a histone deacetylase (RPD3, Murfett et al., 2001) and retinoblastoma (Rb)-associated proteins, which have been identified in plants (Lusser et al., 2001). A multisubunit complex (MSC) is expected to mediate binding of the histone deacetylase to the DNA. Recruitment of the histone deacetylase complex to a particular region of the genome will deacetylate histones. This may allow positively charged lysine residues to allow the formation of heterochromatin. * Acetylated histones. B, COI1 signalling as a model for ubiquitin-mediated protein degradation in arabidopsis. COI1, Skp1, AtCUL1 (Cullin) and AtRbx1 (Rbx1), form an SCF^COI1 ubiquitin ligase complex (depicted here are the only components identified so far). Here a signal activates synthesis of JA and phosphorylation (P) of a target acting as negative regulator (R) of jasmonate responsive genes, which now binds COI1. The ubiquitinated protein is destroyed in the proteasome. Alternatively, the SCFCOI1 ubiquitin ligase complex might activate the regulator via monoubiquitination. K, Kinase; U, ubiquitin.

The plausibility of RUBISCO as a target for COI1 is that, in arabidopsis, senescence is associated with increased JA levels in leaves (He et al., 2002), reduced expression of the small subunit of RUBISCO, and enhanced chlorophyll loss (Parthier, 1990).

The SCF^COI1 shares component proteins with the SCF^TIR1 complex, which regulates auxin responses. Possibly, cross-talk between the JA and the auxin signal pathways is mediated by interaction between these two protein complexes.

SYSTEMIN AND JASMONATE

JA signalling has been extensively studied in arabidopsis and tomato, where the two proposed pathways appear to differ significantly. For example, arabidopsis mutants defective in JA synthesis or perception are deficient in defence responses and are male sterile (Feys et al., 1994; McConn and Browse, 1996; Vijayan et al., 1998), whereas tomato mutants apparently defective in JA synthesis or perception have deficient defences, but are male fertile (Howe et al., 1996). Similarly, the systemic induction of JA responses in tomato occurs through the well-characterized systemin signal pathway (Constabel et al., 1995), but in arabidopsis there is no evidence for an equivalent pathway, even though systemic signalling can be demonstrated (Kubigsteltig et al., 1999).

Systemin, an 18-amino-acid polypeptide, was identified as the primary signal for the systemic activation of defence genes in leaves of wounded tomato plants (Pearce, 1991). Systemin causes a cascade of intracellular signalling events leading to the release of linolenic acid from membranes, and its conversion to oxylipin molecules that signal the expression of defence genes (Ryan, 2000; Howe and Schilmiller, 2002). Systemin induces the production of H₂O₂, and the subsequent synthesis of JA and induction of defence gene expression (Orozco-Cardenas et al., 2001). Scheer and Ryan (2002) have recently identified the systemin receptor SR160 from Lycopersicon peruvianum as a member of the LRR receptor kinase family, with high amino acid identity and domain similarities to the BRI1 receptor kinase from arabidopsis. The cell-type specific expression pattern of genes encoding prosystemin and some JA biosynthetic enzymes has previously suggested that wound-induced release of systemin into the vascular system activates JA biosynthesis in surrounding vascular tissues in which JA biosynthetic enzymes are located (Ryan, 2000). A role for JAs in intercellular signalling is supported by the fact that application of JA/MeJA to one leaf induces wound-inducible proteinase inhibitors (PI expression) in distal untreated leaves (Farmer and Ryan, 1992). Li et al. (2002) used classical grafting techniques to examine long-distance wound signalling in mutants that are deficient either in JA biosynthesis or in JA perception. Their findings question the role of systemin in systemic wound signalling in tomato plants and, interestingly, ascribe a central role to JAs in this response. In agreement with the above are the studies performed by Strassner et al. (2002). Heterologous expression of three tomato isoforms of 12-oxophytodienoate reductase (OPR) has recently limited the role for JA biosynthesis in the activation of wound response gene expression in systemic as compared with wounded tissues. Local but not systemic increase of OPDA and JA after wounding supports these observations. Moreover in contrast to previous assumptions (Schaller, 2001), the octadecanoid pathway appears to be confined to plastids and peroxisomes and does not involve the cytosolic compartment.

JA responses and long-distance JA-dependent signalling in arabidopsis and tomato may therefore turn out to be more similarly regulated than previously suspected, neither involving systemin. Arabidopsis is, therefore, a suitable model to study JA signalling in plants.

OUTPUT CROSS-TALK

It has become clear that there is a significant amount of ‘cross-talk’ between the different hormone-dependent signalling pathways. For example, the auxin-resistant mutant axr1 also exhibits resistance to exogenous ethylene and JA (Lincoln et al., 1990; Tiryaki and Staswick, 2002). The allele axr1-24 has decreased sensitivity to other inhibitors of root growth such as the ethylene precursor 1-aminocyclopropane-1-carboxylic acid, 6-benzylamino- purine, epi-brassinolide and abscisic acid. AXR1 is necessary for resistance to Pythium irregulare in arabidopsis and the effect of JAR1 and AXR1 is additive showing involvement in the same response pathway. Surprisingly, JA responsive genes LOX2, AOS and AtVSP were also induced by IAA in axr1-24.

Previous studies have identified numerous genes that act in concert with AXR1 in a ubiquitin-like proteasome pathway that mediates auxin signalling (Ruegger et al., 1998; Yeh et al., 2000). AXR1 and ECR1 form a heterodimeric enzyme that activates the ubiquitin-like RUB protein (del Pozo et al., 2002). As already described above, jasmonate and auxin might use a similar signalling mechanism. However, coi1 is not altered in its response to auxin (Feys et al., 1994), suggesting that these are separate signalling pathways. The isolation and characterization of the axr1-24 allele supports the hypothesis that JA and auxin might act through a common signalling intermediate that also affects response to other plant hormones. Further evidence that AXR1 is of general importance for different pathways that are controlled by E3-mediated protein degradation has been provided recently by Schwechheimer et al. (2002). JA root growth inhibition and VSP transcript induction are impaired in axr1-3 mutants. Most interestingly, AXR1 also participates in the repression of photomorphogenesis in the dark, a process that requires the activity of a non-SCF-type E3 consisting of the RING finger protein COP1.

The cev1 mutant has been used to investigate cross-talk between the JA, ethylene and salicylic acid (SA) signal pathways (Ellis et al., 2002a). Treatment of cev1 with SA suppresses expression of PDF1.2 and enhances expression of PR1, though less so than in wild-type plants. coi1 mutants, which are deficient in JA perception/response, have slight but significant PR1 expression, indicating that a COI1-dependent signal normally suppresses PR1 in untreated plants. The double mutant cev1;coi1 expresses neither PDF1.2 nor Thi2.1, confirming that expression of these genes requires the JA perception-response pathway regulated by COI1. The mutant ethylene resistant (etr1) was used to make the double mutant cev1;etr1, in which PDF1.2 expression was absent, confirming a requirement for an ethylene signal for PDF1.2 transcription (Ellis and Turner, 2001). Interestingly, Thi2.1 is constitutively expressed in this double mutant indicating that ethylene signalling suppresses the transcription of Thi2.1. These results emphasize the positive and negative cross-talk between the JA, SA and ethylene signal pathways.

A convergence point between ethylene and JA pathways is represented by transcriptional activation of ethylene-response-factor1 (ERF1), encoding a transcription factor, which regulates the expression of pathogen response genes that prevent disease progression (Lorenzo et al., 2003). The authors have shown also that overexpression of ERF1 can rescue the defence-response defects of coi1 and ein2 (ethylene-insensitive2) by restoring PR gene expression, representing a likely downstream component of both ethylene and jasmonate signalling pathways. Supporting evidence is provided by transcriptome analysis of 35S:ERF1 plants, which reveal an overlap between ERF1- and ethylene/jasmonate up- and downregulated genes.

Although JA and SA signal cascades activate different sets of plant defence genes (Thomma et al., 1998) or even act antagonistically (Felton et al., 1999), there is substantial communication between the pathways (Stout et al., 1999; Moran et al., 2001). Plants use insect-derived signals to regulate their defence pathways (Moran et al., 2001). The first example of use by insects of plant signal molecules has been provided recently. In response to plant defences, herbivores increase their detoxifying arsenal, including cytochrome P450 (Li et al., 2000) leading to the concept of ‘signal-eavesdropping’ on plant defence signals.

SUMMARIZING COMMENTS

Many of the known plant signal pathways are defined by the signalling molecules they produce and respond to, including JA, SA, auxin and ethylene. Recent evidence indicates that these signal pathways are not linear, but are integrated through a network of cross-talking connections that appear to co-ordinate responses. A hypothetical model of interaction is shown in Fig. 3. A current challenge is to define the connections, which should begin to reveal how plant responses to biotic and abiotic stresses are integrated.

Fig. 3. Model of the interaction between jasmonate (JA), ethylene, auxin and salicylic acid (SA) signalling pathways. JA, via COI1, and ethylene, via ETR1 and EIN2, act synergistically and in an ERF1-dependent manner to induce the expression of PDF 1.2. cev1 has constitutive JA and ethylene signalling. A wound signal might induce the production of JA and this will stimulate the expression of JA responsive genes, like Thi 2.1 and VSP. Jasmonate and auxin might use a similar signalling mechanism which involves AXR1. COI1-dependent PR1 repression represents antagonism between JA and SA signalling. Arrows and bars indicate positive and negative interaction, respectively. The interaction between signalling pathways may vary when other output responses are considered.

ACKNOWLEDGEMENTS

The authors are grateful to Roberto Solano and Oscar Lorenzo who made available unpublished information to assist the preparation of this review. Many thanks also to Sarah B. Nettleship for critical comments. Research work in J.G.T.’s laboratory is supported by grants from the Biotechnology and Biological Sciences Research Council.

Received: 10 February 2003; Returned for revision: 9 April 2003; Accepted: 28 May 2003 Published electronically: 18 July 2003