TGA transcription factors and jasmonate-independent COI1 signalling regulate specific plant responses to reactive oxylipins.

Jasmonates and phytoprostanes are oxylipins that regulate stress responses and diverse physiological and developmental processes. 12-Oxo-phytodienoic acid (OPDA) and phytoprostanes are structurally related electrophilic cyclopentenones, which activate similar gene expression profiles that are for the most part different from the action of the cyclopentanone jasmonic acid (JA) and its biologically active amino acid conjugates. Whereas JA-isoleucine signals through binding to COI1, the bZIP transcription factors TGA2, TGA5, and TGA6 are involved in regulation of gene expression in response to phytoprostanes. Here root growth inhibition and target gene expression were compared after treatment with JA, OPDA, or phytoprostanes in mutants of the COI1/MYC2 pathway and in different TGA factor mutants. Inhibition of root growth by phytoprostanes was dependent on COI1 but independent of jasmonate biosynthesis. In contrast, phytoprostane-responsive gene expression was strongly dependent on TGA2, TGA5, and TGA6, but not dependent on COI1, MYC2, TGA1, and TGA4. Different mutant and overexpressing lines were used to determine individual contributions of TGA factors to cyclopentenone-responsive gene expression. Whereas OPDA-induced expression of the cytochrome P450 gene CYP81D11 was primarily regulated by TGA2 and TGA5, the glutathione S-transferase gene GST25 and the OPDA reductase gene OPR1 were regulated by TGA5 and TGA6, but less so by TGA2. These results support the model that phytoprostanes and OPDA regulate differently (i) growth responses, which are COI1 dependent but jasmonate independent; and (ii) lipid stress responses, which are strongly dependent on TGA2, TGA5, and TGA6. Identification of molecular components in cyclopentenone signalling provides an insight into novel oxylipin signal transduction pathways.

Introduction

Oxygenation of polyunsaturated fatty acids leads to the production of oxylipins, such as jasmonates and phytoprostanes, via enzymatic or non-enzymatic pathways (Mueller, 2004; Wasternack, 2007). Exogenous application of jasmonic acid (JA) inhibits mitosis, root growth, and seed germination (Swiatek ). Endogenous jasmonate biosynthesis is required for development of fertile flowers (Sanders ). Jasmonates also control abiotic and biotic stress responses with a concomitant induction of a variety of genes related to JA biosynthesis and defence (Devoto ). Biological activities have also been reported for 12-oxo-phytodienoic acid (OPDA), which is a precursor of JA biosynthesis. OPDA inhibits root growth and mitosis similarly to JA but induces a different set of genes (Taki ; Mueller ). Endogenous OPDA was recently shown to impede seed germination independent of JA biosynthesis and signalling (Dave ; Dave and Graham, 2012). Mutants with defects in oxylipin biosynthesis, signalling, and transport were used to establish the biological functions of both compounds (McConn and Browse, 1996; McConn ; Stintzi and Browse, 2000; Malek ; Park ; Mene-Saffrane ; Dave ; Stotz ). Such studies demonstrated that jasmonates protect plants against chewing insects (Howe ; McConn ; Pieterse ) and modulate host–pathogen interactions (Ton ; Laurie-Berry ; Pieterse ). OPDA was shown to protect specifically against necrotrophic pathogens and not by virtue of its being a JA precursor (Raacke ; Stotz ).

Phytoprostanes are non-enzymatically formed compounds with structural similarity to OPDA (Mueller, 2004). Similarly to JA and OPDA, these compounds inhibit root growth and mitosis and induce the production of secondary metabolites (Mueller ). The set of genes which is induced by phytoprostanes shows a strong overlap with the OPDA-responsive genes and only a small overlap with JA-induced genes. This can be explained by the presence of an α,β-unsaturated carbonyl group in OPDA and phytoprostanes, which are electrophilic cyclopentenones. In contrast, JA is a non-electrophilic and chemically unreactive cyclopentanone. The α,β-unsaturated carbonyl group is the reason for the higher chemical reactivity, which was suggested to be crucial for the biological activity (Farmer and Davoine, 2007).

Recently, substantial progress has been made towards understanding the signal transduction pathway mediating the response to jasmonates. JA–isoleucine (JA-Ile), the biologically active form of JA, is bound to the F-box protein COI1 in the presence of JASMONATE ZIM-domain (JAZ) protein family members (Chini ; Thines ; Sheard ). JAZ proteins act as negative regulators of jasmonate-responsive gene expression. Binding of JA-Ile leads to the degradation of JAZ proteins, resulting in the release of transcription factors such as MYC2, which promote the expression of jasmonate-responsive genes (Chini ). MYC2 was identified via positional cloning of a jasmonate-insensitive jin1 mutant allele (Berger ); JIN1 encodes the basic helix–loop–helix transcription factor MYC2 (Lorenzo ).

In contrast to the jasmonate signal transduction pathway, only little is known about the mechanism that mediates the effects of OPDA and phytoprostanes. Putative binding sites for TGA transcription factors are over-represented in promoters of phytoprostane-responsive genes, and specifically the TGA2, TGA5, and TGA6 factors were shown to regulate gene expression in response to cyclopentenone oxylipins (Mueller ). Induction of 30% and 60% of the genes in response to OPDA and the phytoprostane PPA1, respectively, did not occur in the tga2 tga5 tga6 mutant, which is defective in expression of all three TGA factor genes. However, the participation of other TGA factors in responses to these cyclopentenones has not been tested.

The primary aim of this study was to uncover signalling pathways that mediate the effects of reactive oxylipins on plant growth and stress responses, the jasmonate receptor COI1 and TGA transcription factors being of particular interest. With respect to stress responses, specific contributions of individual TGA factors to OPDA-dependent gene expression were determined using the cytochrome P450 gene CYP81D11, the regulation of which was further characterized recently (Köster ), the glutathione S-transferase gene GST25, and the OPDA reductase gene OPR1.

Materials and methods

Plant material and growth conditions

The jin1 and coi1-16 mutants together with their Arabidopsis thaliana (L.) Heynh. background Col-gl were those originally reported (Berger ; Ellis and Turner, 2002; Nickstadt ). The dde2-2 mutant in the background of ecotype Col-0 was previously published (Malek ). The tga6, tga2 tga5, and tga2 tga5 tga6 mutants as well as the tga1 tga4 double mutant were those originally described (Zhang ; Kesarwani ). All transgenic lines overexpressing TGA2, TGA5, or TGA6 were received from Professor Christiane Gatz. In addition to the previously published lines TGA2.1, TGA2.2, TGA5.1, TGA5.2, and TGA6.2 (Zander ), novel TGA5 and TGA6 lines were tested. All tga mutant and TGA-overexpressing lines were generated in the background of ecotype Col-0.

Seedlings were grown in liquid MS (Murashige and Skoog) medium containing 1% or 2% sucrose or on MS agar plates as previously described (Mueller ). Seedlings were grown with a 9h light/15h dark cycle at 22 °C under fluorescent light (150 µmol m–2 s–1).

Chemical treatments

Seedlings grown in liquid MS medium or on MS agar plates were treated with OPDA synthesized by enzymatic conversion of linolenic acid using linseed acetone powder (Parchmann ), JA (Sigma-Aldrich, St Louis, MO, USA), the phytoprostane PPA1 (Thoma ), or the prostaglandin PGA1 (Cayman Chemical, Ann Arbor, MI, USA).

Quantitative real-time PCR analysis

Total RNA was extracted from liquid-grown seedlings using the E.Z.N.A. plant RNA kit (Omega Bio-Tek, Norcross, GA, USA). Potential DNA contamination was removed using on-column digestion with DNase I. Following quantification using an ND-1000 UV-Vis Spectrophotometer (NanoDrop, Wilmington, DE, USA), 1 µg of total RNA was used for cDNA synthesis using M-MLV RNase H minus reverse transcriptase (Promega, Madison, WI, USA). Real-time PCR was performed using a QPCR SYBR Green Mix (Thermo Scientific, Lafayette, CO, USA). Primers are listed in Supplementary Table S1 available at JXB online, except for OPR1 and Act2/8, which have already been published (Mueller ; Ellinger ). Reactions were performed on a Mastercycler Realplex (Eppendorf, Wesseling-Berzdorf, Germany) or on a CFX96 Real-Time PCR Detection System (BioRad, Hercules, CA, USA) with 40 cycles of denaturation for 15 s at 95 °C, annealing for 20 s at 55 °C, and extension for 20 s at 72 °C. This program was followed by a melting curve analysis. Purified real-time PCR products were used for calibration using the relative standard curve method (Appplied Biosystems, Carlsbad, CA, USA). Three biological replicates were used for each data point.

Statistical analysis

Analysis of variance (ANOVA) was used for statistical analysis of root growth measurements. Levene’s test was used to determine homogeneity of variances. Data were transformed to achieve homogeneous variances. Alternatively, data were analysed using non-parametric statistics. Two-tailed tests were used with α < 0.05. The Relative Expression Software Tool V2.0.13 (Qiagen, Hilden, Germany) was used to determine the significance of pairwise comparisons of quantitative PCR data.

Results

Inhibition of root growth by phytoprostanes is dependent on COI1 but independent of jasmonate biosynthesis

An effect shared by jasmonates and phytoprostanes is the inhibition of root growth, which was previously measured in wild-type A. thaliana seedlings after treatment with OPDA or PPA1 (Mueller ). COI1 is known to mediate inhibition of root growth in response to exogenous JA or JA methyl ester. To test whether inhibition of root growth in response to phytoprostanes is also COI1 dependent, the response of the coi1 mutant was analysed. The root length of coi1 seedlings on medium containing 25 µM JA, OPDA, or PPA1 was similar to that of the control grown on MS medium without the addition of oxylipins (Fig. 1A). This demonstrates that inhibition of root growth by OPDA or phytoprostanes is dependent on COI1. In addition, this result shows that growth inhibition is not based on a toxic effect of cyclopentenones but on signalling processes.

Fig. 1.

Inhibition of root growth by oxylipins in different mutants. Seedlings of coi1-16 (A), tga2 tga5 tga6 (B), and tga1 tga4 (C) were grown together with their corresponding wild types on vertically oriented MS agar plates containing phytoprostane A1 (PPA1), 12-oxo phytodienoic acid (OPDA), or jasmonic acid (JA) in a final concentration of 25 µM, or the solvent <2% methanol (control or Cont.). Root lengths were measured after 8 d of growth. Shown are means of 20 seedlings ±95% confidence intervals. Letters indicate significant differences among means. Independent experiments (six for tga2 tga5 tga6, four for coi1 and tga1 tga4) were performed with similar results.

It is not clear whether OPDA exerts the observed effect directly or indirectly via JA biosynthesis because the coi1 mutant can convert OPDA to JA. So far, COI1 has only been shown to bind amino acid conjugates of JA and coronatine (Thines ; Katsir ). This raises the question of whether JA-Ile mediates the effect of PPA1. To investigate the possibility that an accumulation of JA-Ile upon PPA1 treatment is responsible for the inhibition of root growth, the dde2 mutant was tested. This mutant contains a knockout allele of the allene oxide synthase (AOS) gene (Malek ). As a result, the dde2 mutant no longer produces OPDA, JA, and JA-Ile (Köster ). Inhibition of root growth in the dde2 mutant in response to phytoprostane treatment was similar to the root growth inhibition observed in the wild type (Table 1). This clearly shows that the inhibitory effect of phytoprostanes on root growth is not mediated through OPDA or JA-Ile. These data also demonstrate that COI1 plays an important role in mediating root growth-inhibitory effects of oxylipins other than jasmonates.

Table 1.

Oxylipin-mediated root growth inhibition in the allene oxide synthase mutant dde2 and wild-type (Col-0) A. thaliana.

	Col-0			dde2
	Control	25 µM JA	25 µM PPA1	Control	25 µM JA	25 µM PPA1
Length (mm)	21.9±1.8	6.7±1.8	10.2±1.7	24.4±1.9	7.2±1.7	12.2±1.8
% Length	100	31	47	100	30	50

Sterilized seeds of Col-0 and dde2-2 were grown on vertically oriented square Petri dishes containing MS medium supplemented with 2% (w/v) sucrose and oxylipins in a final concentration of 25 µM. Control treatments contained the solvent methanol (<2%). Root length was determined after 7 d. Shown are means ±95% confidence intervals of 14–16 seedlings. Mann–Whitney U-tests revealed no significant effect of genotypes on treatment (P ≤ 0.129).

As mentioned above, induction of gene expression in response to cyclopentenones is impaired in the tga2 tga5 tga6 mutant. It was therefore investigated whether this mutant is also insensitive to oxylipin-triggered inhibition of root growth. On control medium without oxylipins, roots of the tga2 tga5 tga6 mutant were considerably shorter (54%) than wild-type roots (F 1,132=230.6, P < 0.001). Oxylipins strongly inhibited root growth. Root growth of the tga2 tga5 tga6 mutant was more sensitive to the presence of PPA1 (F 1,198=42.4, P < 0.001) and JA (F 1,208=5.3, P = 0.023) than wild-type roots (Fig. 1B). The difference in genotype-dependent inhibition of root growth by OPDA was not significantly different. Root lengths of the triple mutant were reduced to 15, 21, and 26% relative to the lengths on control medium in the presence of PPA1, OPDA, and JA, respectively; corresponding relative root lengths in the wild type were 56, 27, and 35%. These data illustrate that the transcription factors TGA2, TGA5, and TGA6 are not required for root growth inhibition in response to oxylipins. Instead, the tga2 tga5 tga6 mutant was particularly hypersensitive to PPA1.

Root growth was also analysed in tga1 tga4, a double mutant defective in expression of TGA1 and TGA4, which represents a different class of TGA factors. In contrast to the tga2 tga5 tga6 mutant, growth phenotypes of the tga1 tga4 mutant were identical to those of the wild type on control medium and on medium containing JA, OPDA, and PPA1 (Fig. 1C). This shows that TGA1 and TGA4 are not involved in regulating root growth in response to oxylipins.

Regulation of phytoprostane-responsive genes is dependent on class II TGA factors but not on COI1 and MYC2

The results on COI1-dependent inhibition of root growth by phytoprostanes prompted the investigation of whether induction of phytoprostane-responsive genes is dependent on COI1. A limited analysis of this latter oxylipin response was previously documented in coi1 mutant and wild-type plants using northern hybridization with two probes, one for the cytochrome P450 gene CYP81D11, which responds to diverse stimuli (Mueller ; Matthes ; Köster ), and the other one for the OPDA reductase genes OPR1/2, which are phytoprostane responsive but also up-regulated after OPDA and JA treatment (Mueller ). To challenge these previous findings, a more comprehensive analysis was performed using an independent method. Quantitative real-time PCR analysis of the above-mentioned genes as well as the glutathione S-transferase genes GST6 and GST25, which are related to detoxification, and the TolB-like gene was performed; all three genes are phytoprostane responsive; GST6 and TolB-like genes also show some up-regulation after OPDA treatment (Mueller ). To discriminate the effects of different classes of oxylipins, the MYC2 transcription factor mutant jin1 and expression of the vegetative storage protein gene VSP1, which is not responsive to phytoprostanes but shows COI1-dependent induction after JA treatment, were tested.

Relative to the wild type, induction of all tested phytoprostane-responsive genes by PPA1 or OPDA was not reduced in the jin1 and coi1 mutants (Fig. 2). The trend of the previously reported reduced induction of CYP81D11 in the coi1 mutant by reactive oxylipins (Mueller ) was confirmed; methodological differences are probably responsible for quantitative differences between northern hybridization and quantitative real-time PCR because CYP81D11 belongs to a gene family with 15 members (Bak ). Up-regulation of VSP1 and CYP81D11 after JA treatment was clearly reduced in both mutants. Reduction of VSP1 induction was stronger in the coi1 mutant than in the jin1 mutant, which is in agreement with published data (Benedetti ; Berger ). The jin1 mutant has a small effect on VSP1 expression because MYC2 acts in concert with MYC3 and MYC4 to regulate the expression of VSP1 (Fernandez-Calvo ). Together, these data show that, in contrast to inhibition of root growth, induction of the tested phytoprostane-responsive genes is not dependent on COI1.

Fig. 2.

Expression of oxylipin-responsive genes in the wild type and in mutants of the jasmonate pathway, coi1 (left column) and jin1 (right column). Seedlings were grown for 10 d in MS medium containing 2% sucrose under short-day conditions. The medium was exchanged for 75 µM phytoprostane A1 (PPA1), 75 µM 12-oxo phytodienoic acid (OPDA), 75 µM jasmonic acid (JA), or the solvent 0.5% methanol (control). After a treatment for 4h, RNA was extracted, converted into cDNA, and amplified using quantitative reverse transcription PCR. Expression of the cytochrome P450 gene CYP81D11, the glutathione S-transferase genes GST6 and GST25, the OPDA reductase gene OPR1, the TolB-like gene, and the gene encoding vegetative storage protein1, VSP1, is shown. Expression was normalized to the actin gene Act2/8, which was used as a constitutively expressed internal control. Expression of the wild-type control treatment was set to 1 and all other data were expressed relative to it. Presented are means and standard deviations of three independent experiments with different biological replicates.

It was previously shown by microarray and northern analysis that induction of CYP81D11 and OPR1/2 genes by oxylipins is reduced in the tga2 tga5 tga6 mutant (Mueller ). To compare the response of the triple mutant to exogenous JA and reactive oxylipins, target gene expression was analysed by quantitative reverse transcription PCR. To determine whether class II TGA factors specifically regulate oxylipin-induced gene expression, the class I TGA factor mutant tga1 tga4 was tested.

The tga2 tga5 tga6 mutant exhibited lower induction of CYP81D11, GST25, OPR1, and TolB-like by PPA1 and OPDA in comparison with the wild type. Expression of GST6 showed a tendency to lower induction than in the wild type, especially after treatment with OPDA (Fig. 3). These results are consistent with published data on CYP81D11, OPR1, TolB-like, and GST6 expression (Mueller ). In addition, the induction of all tested genes by JA was lower relative to the wild type. This result confirms the previous conception that, besides their involvement in responses to OPDA and phytoprostanes, TGA2, TGA5, and TGA6 mediate responses to exogenous JA (Mueller ; Köster ). In contrast to the triple mutant, induction of all tested genes was not reduced in the tga1 tga4 mutant. This suggests that TGA1 and TGA4 are not necessary for oxylipin responses.

Fig. 3.

Expression of oxylipin-responsive genes in the wild type and in tga2 tga5 tga6 (left column) and tga1 tga4 mutants (right column). Seedlings were grown for 10 d in MS medium containing 2% sucrose under short-day conditions. The medium was exchanged for 75 µM phytoprostane A1 (PPA1), 75 µM 12-oxo phytodienoic acid (OPDA), 75 µM jasmonic acid (JA), or the solvent 0.5% methanol (control). After a treatment for 4h, RNA was extracted, converted into cDNA, and amplified using quantitative reverse transcription PCR. Expression of the cytochrome P450 gene CYP81D11, the glutathione S-transferase genes GST6 and GST25, the OPDA reductase gene OPR1, and the TolB-like gene is shown. Expression was normalized to the actin gene Act2/8, which was used as a constitutively expressed internal control. Expression of the wild-type control treatment was set to 1 and all other data were expressed relative to it. Presented are means and standard deviations of three independent experiments with different biological replicates.

Differential regulation of phytoprostane-responsive genes in tga6, tga2 tga5, and tga2 tga5 tga6 mutants

To test the individual contributions of TGA2, TGA5, and TGA6 to cyclopentenone-regulated CYP81D11, OPR1, and GST25 expression, tga6, tga2 tga5, and tga2 tga5 tga6 mutants were used. In addition to OPDA, A. thaliana seedlings grown in MS medium were challenged with PGA1, a commercially available and structurally related cyclopentenone, which was previously shown to bind covalently to AtGST6 (Dueckershoff ).

CYP81D11 was induced 60- to 70-fold after treatment of wild-type seedlings for 4h with OPDA or PGA1 (Fig. 4). CYP81D11 reached >70% of the wild-type induction level in the tga6 mutant irrespective of the stimulus, suggesting that the absence of TGA6 does not have a significant effect on cyclopentenone-induced expression of this gene. Basal CYP81D11 levels did not differ between the tga6 mutant and the wild type, but basal expression levels were reduced >4-fold in the tga2 tga5 and tga2 tga5 tga6 mutants. Both OPDA- and PGA1-stimulated expression of CYP81D11 was significantly reduced in the tga2 tga5 double mutant, reaching <20% of induced wild-type levels. A further reduction in oxylipin-induced CYP81D11 expression occurred in the tga2 tga5 tga6 mutant, reaching <3% of wild-type expression, which was not significantly different from uninduced wild-type levels. TGA6 therefore exerts a significant effect on CYP81D11 expression in the absence but not in the presence of TGA2 and TGA5.

Fig. 4.

Expression of oxylipin-responsive genes in the wild type and tga mutants. Seedlings were grown for 10 d in MS medium containing 1% sucrose under short-day conditions. The medium was exchanged for 75 µM 12-oxo phytodienoic acid (OPDA), 75 µM prostaglandin A1 (PGA1), or the solvent 0.5% methanol (control). After a treatment for 4h, RNA was extracted, converted into cDNA, and amplified using quantitative reverse transcription PCR. Expression of the cytochrome P450 gene CYP81D11, the OPDA reductase gene OPR1, and the glutathione S-transferase gene GST25 is shown. Expression was normalized to the actin gene Act2/8, which was used as a constitutively expressed internal control. Expression of the wild-type control treatment was set to 1 and all other data were expressed relative to it. Means and standard errors of three biological replicates are shown. Significant differences among means indicated by letters were determined using the Relative Expression Software Tool V2.0.13 (Qiagen, Hilden, Germany).

OPR1 expression increased 10- and 21-fold after treatment of wild-type seedlings with OPDA and PGA1, respectively (Fig. 4). Basal OPR1 levels did not vary much between mutant and wild-type seedlings. In the tga6 mutant, expression of OPR1 reached only 46% and 26% of wild-type levels after induction with OPDA and PGA1, respectively. The response to PGA1 was significantly reduced, indicating that TGA6 plays an essential role in OPR1 induction. Up-regulation of OPR1 by OPDA reached 26% of wild-type levels in the tga2 tga5 mutant. Induction of OPR1 by PGA1 was significantly less in the tga2 tga5 mutant, reaching only 10% of wild-type levels. OPDA- and PGA1-responsive expression of OPR1 was further decreased in the tga2 tga5 tga6 mutant.

GST25 was induced 16- and 5-fold after treatment of wild-type plants with OPDA and PGA1, respectively (Fig. 4). GST25 expression reached 57% and 45% of wild-type levels in the tga6 mutant after induction with OPDA and PGA1, respectively. Cyclopentenone-induced GST25 expression levels were very similar in the tga6 and tga2 tga5 mutant, suggesting that induced GST25 expression is regulated similarly by TGA2 and TGA5 and by TGA6. The induction level in the tga2 tga5 tga6 mutant was <3% relative to the wild type and did not differ from uninduced wild-type levels. Quantitative differences in GST25 or OPR1 induction levels among experiments (as compared with Figs 2 and 3) are probably attributable to subtle changes in plant growth conditions.

Separate effects of three TGA factors on OPDA-induced gene expression

To examine further the contribution of individual TGA factors to OPDA-induced gene expression, TGA2-, TGA5-, or TGA6-overexpressing A. thaliana lines (Zander ) were used. TGA protein expression was readily detected in crude extracts from overexpressing plants (Supplementary Fig. S1 at JXB online). TGA protein expression varied among overexpressing lines but did not substantially alter the induction of target gene expression (Supplementary Figs S2, S3).

OPDA treatment of wild-type seedlings increased CYP81D11 expression 93-fold (Fig. 5). This level of induction was consistent across experiments in the wild-type background Col-0 (Figs 3, 4), but induction of CYP81D11 appeared to be quantitatively lower in the genotype Col-gl (Fig. 2). No induction of CYP81D11 by OPDA was observed in the tga2 tga5 tga6 mutant, which served as the genetic background for all three lines overexpressing TGA factors. CYP81D11 expression was significantly increased after OPDA treatment of TGA2.1- and TGA5.1-overexpressing lines by 46% and 23% of wild-type levels, respectively. However, OPDA induction of CYP81D11 was not significant in the TGA6.3-overexpressing line, reaching only 12% of wild-type levels. These results support the tga mutant data (Fig. 4) and demonstrate that TGA6 is not sufficient for induced CYP81D11 expression.

Fig. 5.

Expression of oxylipin-responsive genes in wild-type, tga2 tga5 tga6 mutant, and TGA-overexpressing plants. TGA overexpression occurred in the background of the tga2 tga5 tga6 mutant. Seedlings were grown for 10 d in MS medium containing 1% sucrose under short-day conditions. The medium was exchanged for 75 µM 12-oxo phytodienoic acid (OPDA) or the solvent 0.5% methanol (control). After a treatment for 4h, RNA was extracted, converted into cDNA, and amplified using quantitative reverse transcription PCR. Expression of the cytochrome P450 gene CYP81D11, the OPDA reductase gene OPR1, and the glutathione S-transferase gene GST25 is shown. Expression was normalized to the actin gene Act2/8, which was used as a constitutively expressed internal control. Expression of the wild type control treatment was set to 1 and all other data were expressed relative to it. Means and standard errors of three biological replicates are shown. Significant differences among means indicated by letters were determined using the Relative Expression Software Tool V2.0.13 (Qiagen, Hilden, Germany).

Effects of TGA2.1, TGA5.1 and TGA6.3 overexpression on OPDA-induced expression of OPR1 and GST25 were similar and were distinct from those of CYP81D11. Overexpression of each of the three transcription factors overcame the lack of OPR1 and GST25 induction after OPDA treatment in the tga2 tga5 tga6 mutant. Although TGA2 made a significant contribution to OPDA-induced expression of OPR1 and GST25, the effects of TGA5 and TGA6 were quantitatively larger.

Based on data from both mutant and transgenic seedlings, the response of CYP81D11 to OPDA is regulated directly or indirectly by TGA2 and TGA5. In contrast, TGA5 and TGA6 make a quantitatively larger contribution to OPDA-induced expression of OPR1 and GST25 than TGA2. These data suggest that at least two classes of OPDA-regulated genes exist.

Discussion

COI1 mediates root growth inhibition in response to phytoprostanes independent of jasmonates

Whereas root growth was not inhibited by JA, OPDA, or PPA1 in the coi1 mutant (Fig. 1A), the AOS mutant dde2 was fully sensitive to phytoprostane treatment (Table 1). This finding illustrates that root growth in this JA- and OPDA-deficient mutant is dependent on COI1 and that COI1 mediates jasmonate-independent responses to an electrophilic oxylipin. While similar JA-Ile-independent COI1-mediated responses were previously documented (Ribot ; Adams and Turner, 2010; Stotz ; Köster ; Ralhan ), the underlying mechanism has not been resolved. Based on these published results, apparently two jasmonate-independent COI1 pathways exist. Unlike the opr3 mutant, aos and coi1 mutants are impaired in defence responses against the necrotrophic ascomycete Sclerotinia sclerotiorum (Stotz ) and during wound-induced expression of AtPHO1;H10 (Ribot ), suggesting that OPDA mediates JA-Ile-independent COI1 responses. On the other hand, ethylene-dependent inhibition of root growth (Adams and Turner, 2010), susceptibility to Verticillium longisporum (Ralhan ), and induction of CYP81D11 in response to xenobiotics (Köster ) are altered in the coi1 but not in the aos mutant, suggesting that in this case COI1 exerts its effects independently of OPDA. Elegant grafting experiments showed that susceptibility to V. longisporum is dependent on a COI1-specific recognition event in the root (Ralhan ), suggesting that this organ may also play a role in mediating oxylipin responses. In analogy, we now show that the phytoprostane PPA1 signals through COI1 independently of OPDA and JA biosynthesis.

COI1 interacts with JAZ1, JAZ3, JAZ6, JAZ9, and JAZ10 in a JA-Ile- and coronatine-dependent manner (Melotto ; Chung and Howe, 2009; Sheard ). Although OPDA does not facilitate interactions of COI1 with JAZ1, JAZ3, and JAZ9 (Melotto ; Chung and Howe, 2009), the possibility cannot be excluded that cyclopentenones may promote interactions between COI1 and other JAZ proteins. JA-Ile induces 10 of the 12 JAZ family members as part of a negative feedback loop (Chini ). Analysis of transcript profiling in response to the phytoprostane PPA1 (Mueller ) did not indicate regulation of JAZ genes by this compound. Alternatively, binding of phytoprostanes to COI1 may facilitate interactions with other proteins that are not related to JAZ proteins but nevertheless act as co-receptors of COI1.

TGA factors 2, 5, and 6 activate oxylipin-responsive gene expression but impede inhibition of root growth by oxylipins

The TGA factors 2, 5, and 6 were shown to act as redundant members of the class II TGA factors during the establishment of systemic acquired resistance, which is regulated by the salicylic acid (SA) pathway (Zhang ). In addition, these transcription factors are involved in regulating gene expression in response to the jasmonate/ethylene pathway (Zander ). This pathway is important for resistance to necrotrophic pathogens, and the tga2 tga5 tga6 mutant is more susceptible to Botrytis cinerea than wild-type plants (Zander ). A possible explanation for this hypersusceptibility is perhaps reduced jasmonate/ethylene signalling and a strongly reduced expression of genes related to detoxification (Mueller ), leading to a reduced and slower metabolism of phytoprostanes and other toxic compounds. This is supported by results showing that in the tga2 tga5 tga6 mutant, cell death is elevated after treatment with tert-butyl hydroperoxide (Supplementary Fig. S4 at JXB online) and that sensitivity to xenobiotics is increased relative to the wild type (Fode ). Collectively, these data suggest that these three TGA factors play an important role in detoxification responses of plants.

The fact that the tga2 tga5 tga6 mutant still responded to oxylipins with a reduction in root growth (Fig. 1) suggests that this response is not dependent on these transcription factors. Although the growth of the triple mutant was reduced on MS agar medium relative to the wild type, inhibition of root growth by PPA1 was quantitatively larger in the tga2 tga5 tga6 mutant than in the wild type. The hypersensitivity of the triple mutant to a phytoprostane seems to support the proposed antagonism between these three TGA factors and MYC2 affecting ORA59 expression and jasmonate/ethylene-related gene expression (Zander ).

TGA-specific regulation of phytoprostane-responsive target genes

The putative detoxification genes CYP81D11, OPR1, and GST25 responded differently to TGA2, TGA5, and TGA6. CYP81D11 differed from GST25 and OPR1 in the level of induction by cyclopentenones but also in the specificity of induction by different TGA factors. Cyclopentenone-induced expression of CYP81D11 was more strongly regulated by TGA2 and TGA5 than by TGA6 (Figs 4, 5). At the most, overexpression of TGA factors resulted in an OPDA induction of ~50% relative to wild-type levels (Fig. 5). Thus, overexpression of single TGA factors results in partial induction of CYP81D11 expression, raising the possibility that TGA factors may become limiting due to the heterodimerization requirements of these transcription factors. In contrast, overexpression of TGA5 or TGA6 in the background of the tga2 tga5 tga6 mutant resulted in wild-type levels of GST25 and OPR1 expression after OPDA treatment (Fig. 5), suggesting that individual TGA factors can be sufficient for the induction of these genes. These results show that control of gene expression by TGA factors varies among target genes. In contrast to the results presented here, SA-induced expression of PR1 is blocked in the tga2 tga5 tga6 mutant, but wild-type induction levels are reached in tga6 and tga2 tga5 mutants, which demonstrates transcription factor redundancy with respect to PR1 expression (Zhang ). On the other hand, expression of PDF1.2 after induction with methyl jasmonate and 1-aminocyclopropane-1-carboxylic acid is similar in wild-type and tga6 mutant plants, whereas stimulus-induced expression is equally low in tga2 tga5 and tga2 tga5 tga6 mutants (Zander ). Thus, expression of PDF1.2 under these conditions is strictly dependent on TGA2 and TGA5. However, TGA factors indirectly regulate PDF1.2 expression (Zander ).

Unlike GST25, which is exclusively regulated by TGA2, TGA5, and TGA6, CYP81D11 was recently shown to be co-regulated by these TGA factors and COI1 (Köster ). Sequence analysis of the OPR1 promoter provides no evidence for the presence of a MYC2-responsive G-box, also suggesting a fundamental difference in regulation of CYP81D11 versus GST25 and OPR1 genes.

Contrast of the responses to COI1 or TGA2, TGA5, and TGA6

COI1 as well as TGA2, TGA5, and TGA6 induce related but distinct defence responses. For instance, susceptibilities of both coi1 and tga2 tga5 tga6 mutants to B. cinerea are elevated relative to the wild type (Thomma ; Zander ). Likewise, induction of PDF1.2 expression after B. cinerea inoculation is severely reduced in both types of mutants (Guo and Stotz, 2007; Zander ). However, coi1 and tga2 tga5 tga6 mutants differ in cis-jasmone-responsive gene expression patterns (Matthes ), demonstrating clear differences in these signal transduction pathways. This is not surprising because class II TGA factors were shown to activate indirectly the jasmonate/ethylene pathway that is controlled by COI1 (Zander ). Given that COI1 also fulfils distinct roles in regulation of responses to JA and to pathogens via combinatorial jasmonate/ethylene signalling, differences in observed physiological (Fig. 1) and defence responses (Figs 2, 3) can be reconciled.

Whereas PPA1 activates the expression of stress and detoxification genes, this compound down-regulates the expression of genes that contribute to cell growth and division (Mueller ), which may explain the fact that roots respond to phytoprostanes with growth inhibition (Fig. 1). Moreover, root growth inhibition in response to phytoprostanes is lessened by TGA2, TGA5, and TGA6, possibly because these proteins may influence the repression of gene expression associated with growth and division. In contrast, COI1 exerts a negative effect on root growth in response to cyclopentenones, although this receptor is only known to bind JA-Ile and coronatine.

Collectively, these data strongly suggest the existence of two phytoprostane signalling pathways (Fig. 6). One pathway regulates the expression of detoxification genes and is influenced positively by both COI1 and class II TGA factors. The second pathway inhibits root growth, which is mediated by COI1 but negatively influenced by the TGA factors. This proposed model can be reconciled with a previously published model on the antagonism between class II TGA factors and MYC2 (Zander ).

Fig. 6.

Tentative model explaining the observed effects of cyclopentenone oxylipins on root growth and expression of detoxification genes. Roles of the jasmonate receptor COI1 and TGA transcription factors in mediating oxylipin signalling are highlighted. Individual contributions of TGA factors were determined for induction of detoxification genes but not for root growth inhibition. The CYP81D11 promoter is primarily regulated by TGA2 and TGA5. GST25 and OPR1 promoters are primarily regulated by TGA5 and TGA6, although TGA2 does also contribute to the expression of these genes. MYC2 was previously shown to activate CYP81D11 expression (Köster ) presumably by binding to a G-box in the promoter sequence.

Supplementary data

Supplementary data are available at JXB online.

Figure S1. Relative expression of TGA factors in overexpressing A. thaliana lines.

Figure S2. Oxylipin-responsive gene expression in wild-type, tga2 tga5 tga6 mutant, and independent TGA-overexpressing plants.

Figure S3. Effect of different levels of TGA6 protein expression on plant growth and oxylipin-responsive gene expression.

Fig. S4. Cell death in tga2 tga5 tga6 mutant and Col-0 (wild-type) seedlings.

Table S1. Primers and probes used for quantitative RT–PCR.