A rapid wound signal activates the systemic synthesis of bioactive jasmonates in Arabidopsis


  • Abraham J.K. Koo,

    1. Department of Energy Plant Research Laboratory, Michigan State University, East Lansing, Michigan 48824 USA
    Search for more papers by this author
  • Xiaoli Gao,

    1. Genetics Graduate Program, Michigan State University, East Lansing, Michigan 48824, USA
    Search for more papers by this author
  • A. Daniel Jones,

    1. Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing, Michigan 48824, USA
    Search for more papers by this author
  • Gregg A. Howe

    Corresponding author
    1. Department of Energy Plant Research Laboratory, Michigan State University, East Lansing, Michigan 48824 USA
    2. Department of Biochemistry and Molecular Biology, Michigan State University, East Lansing, Michigan 48824, USA
    Search for more papers by this author

*(fax: +1 517 353 9168; e-mail howeg@msu.edu).


Jasmonic acid (JA) and its biologically active derivatives (bioactive JAs) perform a critical role in regulating plant responses to wound stress. The perception of bioactive JAs by the F-box protein COI1 triggers the SCFCOI1/ubiquitin-dependent degradation of JASMONATE ZIM-DOMAIN (JAZ) proteins that repress the expression of JA-response genes. JA is required for many wound-inducible systemic defense responses, but little is known about the role of the hormone in long-distance signal relay between damaged and undamaged leaves. Here, we show that the wounding of Arabidopsis thaliana leaves results in the rapid (<5 min) accumulation of jasmonoyl-l-isoleucine (JA-Ile), the bioactive form of JA, in leaves distal to the wound site. The rapid systemic increase in JA-Ile preceded the onset of early transcriptional responses, and was associated with JAZ degradation. Wound-induced systemic production of JA-Ile required the JA biosynthetic enzyme 12-oxo-phytodienoic acid (OPDA) reductase 3 (OPR3) in undamaged responding leaves, but not in wounded leaves, and was largely dependent on the JA-conjugating enzyme JAR1. Interestingly, the wound-induced synthesis of JA/JA-Ile in systemic leaves was correlated with a rapid decline in OPDA levels. These results are consistent with a model in which a rapidly transmitted wound signal triggers the systemic synthesis of JA, which, upon conversion to JA-Ile, activates the expression of early response genes by the SCFCOI1/JAZ pathway.


Jasmonic acid (JA) and its receptor-active derivatives (bioactive JAs) regulate a wide range of biological processes, including plant defense (Browse and Howe, 2008; Glazebrook, 2005; Howe and Jander, 2008), secondary metabolism (Gundlach et al., 1992; Pauwels et al., 2009), growth control (Balbi and Devoto, 2008; Yan et al., 2007; Zhang and Turner, 2008), trichome initiation (Li et al., 2004; Yoshida et al., 2009) and reproductive development (Browse, 2005). JAs exert their effects by orchestrating large-scale changes in gene expression (Pauwels et al., 2009; Browse, 2009). Current models indicate that in cells containing low JA levels, JASMONATE ZIM-domain (JAZ) proteins repress JA signaling through the direct interaction with the transcription factor MYC2 (Chico et al., 2008; Chini et al., 2007; Katsir et al., 2008a; Thines et al., 2007). The transcriptional activation of JA-response genes is initiated by developmental and environmental cues that activate de novo synthesis of JA (Schaller et al., 2005; Wasternack, 2007). Increased JA levels stimulate JAZ binding to CORONATINE INSENSITIVE 1 (COI1), which is the F-box protein component of the E3 ubiquitin ligase, SCFCOI1 (Balbi and Devoto, 2008; Thines et al., 2007). The hormone-induced COI1–JAZ interaction triggers JAZ degradation via the ubiquitin/26S proteasome pathway, thereby releasing MYC2 from repression. The inhibition of JA signal output by the endogenous JAZ10.4 repressor requires protein–protein interaction through the ZIM domain, suggesting that JAZ homo- or heterodimerization is an important part of the mechanism for the repression of MYC2, and possibly other transcription factors (Chung and Howe, 2009).

We previously demonstrated that the JA-amino acid conjugate jasmonoyl-l-isoleucine (JA-Ile), but not JA, methyl-JA (MeJA) or the JA precursor 12-oxo-phytodienoic acid (OPDA), directly promotes the COI1–JAZ interaction, in the absence of any other plant proteins (Katsir et al., 2008b; Thines et al., 2007; Melotto et al., 2008). This finding extends previous work showing that the conjugation of JA to Ile by the enzyme JAR1 is required for many JA responses (Kang et al., 2006; Staswick, 2008; Staswick and Tiryaki, 2004). Structurally related JA-amino acid conjugates, including JA-Val and JA-Leu, also stimulate the binding of tomato COI1 to JAZ1/3 to varying degrees in vitro (Katsir et al., 2008b). The bacterial toxin coronatine (COR) is a close structural mimic of one [namely(+)-(3R,7S)-JA-Ile] of the two native stereoisomers of JA-Ile (Creelman and Mullet, 1997; Staswick, 2008; Wasternack, 2007). The ability of COR to bind COI1–JAZ complexes with high affinity (Kd ∼ 20 nm), in a manner that depends on COI1, indicates that COI1 (or a COI1–JAZ complex) is a component of the COR receptor, and is most likely to be the receptor for bioactive isomers of JA-Ile as well (Katsir et al., 2008b).

Leaf wounding is well characterized as a stimulus for the rapid activation of JA synthesis (Blechert et al., 1995; Chung et al., 2008; Glauser et al., 2008; McCloud and Baldwin, 1997; Strassner et al., 2002; Weber et al., 1997). Wound-induced accumulation of JA is controlled at the level of substrate availability; all JA biosynthetic enzymes are present in resting (e.g. unwounded) cells (Stenzel et al., 2003; Wasternack, 2007). Lipases that release JA precursors, such as linolenic acid, or possibly OPDA, from galactolipids in the plastid can generate substrates for JA synthesis (Buseman et al., 2006; Hyun et al., 2008; Ishiguro et al., 2001; Kourtchenko et al., 2007). Although there is evidence to indicate that these lipases are regulated by developmental signals (Ito et al., 2007), the mechanism by which lipases, or other rate-limiting enzymes in the JA biosynthetic pathway, are controlled in response to wounding is not known.

Many wound-induced plant defensive traits are expressed in leaves distal to the site of injury (Green and Ryan, 1972). These long-distance responses provide protection against future attacks, and involve the reprogramming of gene expression in undamaged systemic leaves. Considerable research effort has focused on the identification of systemic wound signals, and the mechanisms by which they are generated, transported and perceived in distal leaves. Among the signals implicated in wound-induced systemic responses are oligosaccharides, reactive oxygen species, systemin, green leaf volatiles, hydraulic signals, electrical signals and various plant hormones (Heil and Ton, 2008; Leon et al., 2001; Malone, 1993; Wasternack et al., 2006; Zimmermann et al., 2009). A key role for JA in systemic wound signaling has been established through the use of mutants that are defective in JA synthesis or perception (Leon et al., 1998; Rojo et al., 1998, 1999; Schilmiller and Howe, 2005; Titarenko et al., 1997; Wang et al., 2008). Grafting experiments in tomato showed that the wound-induced systemic expression of defensive proteinase inhibitors depends on JA synthesis in rootstock tissues, but not in distal leaves of the scion, and that this response also depends on the ability of distal leaves to perceive a JA signal (Li et al., 2005, 2002). A requirement for JA action in distal undamaged leaves is consistent with recent work showing that wounding or simulated herbivory results in the systemic accumulation of JA (Glauser et al., 2008) and JA-Ile (Suza and Staswick, 2008; Wang et al., 2008), and that wounding causes a systemic turnover of JAZ proteins (Zhang and Turner, 2008). Some systemic wound responses are controlled by signaling pathways that operate independently of JA (Howe, 2004; Titarenko et al., 1997).

Although it is well established that JA plays an important role in promoting systemic wound responses, little is known about how the JA pathway components are spatially organized and temporally engaged between damaged and undamaged leaves. Specific questions that remain to be answered include: (i) does the wound-induced accumulation of JAs in distal undamaged leaves correlate with wound responses in these tissues?; (ii) which JA derivatives activate systemic gene expression?; (iii) does the wound-induced systemic accumulation of JAs result from de novo synthesis in undamaged leaves, or from transport between damaged to undamaged leaves?; and (iv) what is the nature of the systemic wound signal? Here, we describe the results of a series of experiments indicating that a rapidly transmitted signal triggers the systemic synthesis of JA-Ile, which in turn activates target gene expression via the SCFCOI1/JAZ pathway.


Wound-induced systemic gene expression correlates with increased levels of JA-Ile and JAZ degradation

We used rosette leaves on adult plants as an experimental system in which to study the role of JA-Ile in systemic wound responses in Arabidopsis. Developmentally older leaves at the base of the rosette were mechanically wounded with a hemostat, and then used as a source of tissue for analysis of the local response. Younger undamaged leaves on the same rosette were used for the analysis of the systemic response. To establish the general timing of wound responses in this system, we used RNA blot analysis to measure the wound-induced expression of the JAZ5, JAZ7, OPR3 and OPCL1 genes, which are the primary (i.e. early) response genes in the JA signaling pathway (Chini et al., 2007; Chung et al., 2008). These transcripts accumulated locally and systemically within 15 min of wounding, attained peak levels 30–60 min after wounding, and declined at later time points (Figure 1a). The temporal expression pattern of JR2 (also known as CORI3/CYSTINE LYASE), which is a secondary (i.e. late) response gene (Rojo et al., 1998), was delayed in both local and systemic leaves in comparison with early response genes. In agreement with previous studies (Rojo et al., 1999), the wound-induced systemic expression of JR2 occurred prior to its expression in the damaged leaf. These results show that primary and secondary JA-response genes are expressed at temporally distinct stages (i.e. early and late, respectively) of the systemic wound response, and that the systemic signal involved in activating early response genes is perceived and processed in distal, undamaged leaves, within 15 min of wounding.

Figure 1.

 Wound-induced systemic activation of jasmonic acid (JA) signaling in Arabidopsis.
(a–c) Developmentally older leaves on 30-day-old wild-type (WS) plants were wounded with a hemostat. Damaged (Local) and undamaged distal (Systemic) leaves were harvested at the indicated time after wounding (TAW), and were analyzed for gene expression by RNA blotting (a) or JA-Ile accumulation by LC MS/MS (b, c). Each data point in (b) and (c) represents the mean ± SD of three biological replicates.
(d) Wound-induced degradation of JAZ1-GUS. Four rosette leaves on 18-day-old soil-grown 35S:JAZ1-GUS plants were wounded across the mid-rib with a hemostat. Wounded and control (0 min) plants were stained for GUS activity at the indicated times after wounding. Scale bar: 10 mm.

We used ultraperformance liquid chromatography–tandem mass spectrometry (UPLC MS/MS) to determine whether the JA-Ile levels (measured as JA-Ile plus JA-Leu) correlate with wound-induced changes in gene expression. The JA-Ile content in wounded leaves increased within 5 min of tissue damage, and attained peak levels ∼60 min after wounding (Figure 1b), in agreement with previous studies (Suza and Staswick, 2008; Chung et al., 2008; Glauser et al., 2008). Interestingly, JA-Ile levels in local damaged leaves remained relatively high (i.e. >1000 pmol g−1 fresh weight, FW) 3–6 h after wounding, when early gene expression had returned to the resting level. We also observed a rapid burst of JA-Ile in leaves distal to the wound site (Figure 1c). The JA-Ile content in systemic leaves increased about eightfold within 5 min of wounding, and peaked 15–30 min after wounding; the peak level of systemic JA-Ile was typically 10–20% of that in damaged leaves. The wound-induced systemic increase in JA-Ile preceded the induced expression of early response genes (e.g. OPR3). The rapid disappearance of early gene transcripts at later time points (e.g. 180 min) was correlated with a decline in JA-Ile content to basal levels. The systemic level of JR2 mRNA at this time point, however, remained high. This observation is consistent with the fact the JR2 is a secondary JA-response gene (Rojo et al., 1998), the JA-induced expression of which presumably depends on new synthesis of a transcription factor that is not directly repressed by JAZ proteins.

We employed a 35S:JAZ1-GUS reporter line (Thines et al., 2007) to investigate the effect of mechanical wounding on the turnover of JAZ1 in systemic leaves. In the absence of wounding, strong GUS expression was observed in the petiole and base of younger leaves, as well as in the apical meristem (Figure 1d, 0 min). The mechanical wounding of developmentally older leaves resulted in a time-dependent decrease in GUS staining in the undamaged younger leaves 15–30 min after wounding. Thus, the timing of wound-induced JAZ degradation correlates with the systemic increase in JA-Ile content and activation of early gene expression.

The jar1 mutant is deficient in the wound-induced systemic JA-Ile production and expression of response genes

We next evaluated the systemic wound response in a null mutant (designated jar1-12; Figure 2c and S1a) that harbors a T-DNA insertion in the gene that encodes the JA-conjugating enzyme, JAR1. In both jar1-12 and wild-type Columbia-0 (Col-0) plants, wounding caused a rapid increase (∼10-fold at the 15-min time point) in systemic JA levels, which then declined during the remainder of the time course (Figure 2a). This finding confirms that JA production is rapidly activated in leaves distal to the wound site (Glauser et al., 2008), and further shows that JAR1 activity is not required for this response. In contrast, jar1-12 reduced both the local (Figure S1b) and systemic JA-Ile content to about 20% of wild-type levels (Figure 2b). Although the JA-Ile deficiency resulting from the loss of JAR1 activity in damaged jar1 leaves did not affect the wound-induced gene expression (Chung et al., 2008; Suza and Staswick, 2008; Figure S1a), the JA-Ile deficiency in systemic jar1-12 leaves resulted in a reduced expression of all response genes tested (Figure 2c). These findings indicate that the wound-induced systemic production of JA-Ile is largely dependent on JAR1 activity, and that JAR1 plays a role in the systemic expression of response genes.

Figure 2.

 Systemic wound responses in the jar1 mutant.
(a, b) Time course of the wound-induced systemic accumulation of jasmonic acid (JA) (a) and JA-Ile (b) in wild-type (•, Col-0) and jar1-12 (○). Wounding was performed as described in the legend for Figure 1. Each data point represents the mean ± SD of three biological replicates.
(c) Systemic leaves from the same set of plants used in (a) and (b) were harvested at the indicated time after wounding (TAW), and were used for RNA gel blot analyses of various JA-response genes, or ACT8 as a loading control.

JA-Ile is a predominant JA-amino acid conjugate in systemic leaves

We used UPLC MS/MS to determine the extent to which JA-amino acid conjugates other than JA-Ile accumulate during the wound response. This analysis was performed with both Wassilewskija (WS) and Col-0 ecotypes, using a representative time point for the local (60 min) and systemic (15 min) responses. Among the JA conjugates measured were JA-Ile/Leu, JA-Val, JA-Gln, JA-Thr, JA-Phe, JA-Ala, JA-Met and JA-Trp, as well as the jasmonoyl conjugate of 1-aminocyclopropane-1-carboxylate, JA-ACC. With the exception of JA-Met, JA-Trp, and JA-ACC, which were not detected in the leaves of either ecotype, the level of all compounds increased in damaged leaves (Table 1). JA-Ile/Leu accounted for 90 and 93% of the wound-induced pool of the JA conjugates measured in damaged leaves of Col-0 and WS, respectively. We observed an increase in JA-Ile/Leu and JA-Val in leaves distal to the wound site. The systemic level of JA-Ile/Leu, however, was at least 20-fold higher than that of JA-Val in both Col-0 and WS. Control LC MS/MS experiments were performed to determine the relative proportion of JA-Ile and JA-Leu in the single LC peak corresponding to JA-Ile/Leu. The results showed that, depending on the ecotype, JA-Ile accounts for 85–95% of the JA-Ile/JA-Leu content in wounded leaves (Table 1). We conclude that JA-Ile is the most abundant of the known bioactive JAs to accumulate during the local and systemic wound response of Arabidopsis.

Table 1.   Quantification of jasmonic acid (JA)-amino acid conjugates in wounded Arabidopsis leaves
 WS ecotypeCol-0 ecotype
  1. aLeaves were wounded three times with a hemostat. Local wounded leaves (local) and systemic undamaged leaves (systemic) were harvested at 60 and 15 min, respectively, after wounding.

  2. bJA-Ile and JA-Leu co-eluted as a single peak using standard LC conditions. Additional control experiments (see Experimental procedures) showed that JA-Ile accounts for 90–95% of the JA-Ile/JA-Leu content in damaged Col-0 leaves harvested 1 h after wounding, and for 85–90% of the JA-Ile/JA-Leu content in damaged WS leaves harvested 1 h after wounding.

  3. cND, not detected.

  4. Values are expressed as pmol g−1 fresh weight leaf tissue, and indicate the mean of three biological replicates ± SD.

JA-Ile/Leub7 ± 42195 ± 621164 ± 7712 ± 41746 ± 221113 ± 52
JA-ValNDc71 ± 115 ± 6ND61 ± 46 ± 3
JA-GlnND18 ± 7NDND82 ± 11ND
JA-Thr2 ± 118 ± 54 ± 13 ± 446 ± 86 ± 5
JA-PheND17 ± 9NDND26 ± 6ND
JA-AlaND23 ± 5NDND18 ± 7ND

JA-Ile is synthesized de novo in systemic leaves

The wound-induced systemic accumulation of JA-Ile could result from the transport of JA-Ile from wounded to unwounded leaves or from the de novo synthesis of JA-Ile in undamaged leaves, in response to a long-distance signal. To distinguish between these possibilities, we developed a transgenic system in which the conversion of OPDA to JA can be spatially manipulated with a dexamethasone (DEX)-inducible promoter (Aoyama and Chua, 1997). A chimeric gene (Pdex:GFP-OPR3) encoding a green fluorescent protein (GFP)-OPR3 fusion under the control of a DEX-inducible promoter was introduced into the JA-deficient opr3 mutant. In the absence of DEX treatment, GFP-OPR3 mRNA did not accumulate in wounded or unwounded leaves of the transgenic line (Figure S2). The application of DEX to a subset of rosette leaves on Pdex:GFP-OPR3 plants resulted in the strong expression of GFP-OPR3 within 6 h of treatment. Control experiments showed that DEX application did not induce GFP-OPR3 expression in leaves distal to the site of application (Figure S2). The GFP-OPR3 fusion protein also accumulated in a time-dependent manner in leaves treated with DEX, but not in untreated leaves of the same plant (Figure 3a). The Agrobacterium tumefaciens-mediated transient co-expression of GFP-OPR3, and a peroxisomal marker protein (YFP-PTS1), in Nicotiana tabacum leaves showed that GFP-OPR3, which contains a PTS1 signal at the C terminus, was correctly targeted to peroxisomes (Figure 3b).

Figure 3.

 Functional complementation of opr3 with a dexamethasone (DEX)-inducible GFP-OPR3 transgene.
(a) Protein-blot analysis of GFP-OPR3 levels in DEX-treated Pdex:GFP-OPR3 plants. Protein extracts were prepared from Pdex:GFP-OPR3 leaves that were treated (L) with DEX for the indicated time. As controls, proteins extracts were also prepared from leaves distal to the site of DEX treatment (S), and from wild-type (WS) and opr3 leaves that were wounded 6 h before tissue harvest. Immunoblots were probed with anti-OPR3 (top panel) and anti-GFP (middle panel) antibodies, or were stained with Coomassie Blue (lower panel). Molecular mass standards (kDa) are indicated on the left. Arrows show the position of GFP-OPR3 and native OPR3 proteins. The anti-GFP antibody cross-reacts with an unidentified 54-kDa protein that most likely corresponds to the large subunit of Rubisco.
(b) GFP-OPR3 is targeted to peroxisomes. Confocal images of Nicotiana tabacum leaves co-infiltrated with Agrobacterium strains expressing Pdex:GFP-OPR3 and 35S:YFP-PTS1 are shown. Co-infiltrated leaves were treated with DEX 16 h prior to imaging. The merged image (yellow) shows the co-localization of GFP (green) and YFP (red) fluorescence. Scale bar: 10 μm.
(c) The partial complementation of wound-induced JA production in the opr3 mutant by the DEX-induced expression of GFP-OPR3. Wild-type (ecotype WS), opr3 and Pdex:GFP-OPR3 leaves were either wounded (+) or left undamaged (−), and were collected 2 h later for the quantification of JA levels. Two sets of Pdex:GFP-OPR3 plants were treated (+) with DEX 16 h prior to wounding. Each data point represents the mean ± SD of four biological replicates.

To determine whether the DEX-induced expression of GFP-OPR3 complements the opr3-mediated block in JA synthesis, we measured JA levels in Pdex:GFP-OPR3 leaves that were either treated or not treated with DEX prior to mechanical wounding. In the absence of DEX treatment, the level of JA in wounded Pdex:GFP-OPR3 plants was identical to that observed in wounded opr3 plants, which was ∼5% of wild-type (WS) levels (Figure 3c). Treatment of Pdex:GFP-OPR3 leaves with DEX prior to mechanical wounding increased the JA level above that in mock-treated plants; using our standard conditions for DEX treatment, the quantity of JA produced in DEX-treated, wounded Pdex:GFP-OPR3 leaves was approximately 40% of that in wild-type plants. These results indicate that the DEX-induced expression of GFP-OPR3 partially complements the JA deficiency caused by opr3. The level of wound-induced JA in a Pdex:OPR3 transgenic line expressing OPR3 alone (i.e. not fused to GFP) was similar to that of Pdex:GFP-OPR3 lines, suggesting that the partial complementation of the opr3 phenotype does not result from the reduced activity of the GFP-OPR3 fusion protein (Figure S3).

Having established an experimental approach to spatially manipulate the conversion of OPDA to JA, we used this system to determine whether the wound-induced systemic increase in JA-Ile depends on OPR3 function in the damaged or undamaged distal leaves. As illustrated in Figure 4(a), Pdex:GFP-OPR3 plants were subjected to four different treatments (I–IV) prior to the measurement of JA-Ile in the damaged (Figure 4b; local) and undamaged (Figure 4c; systemic) leaves of wounded plants. Damaged and undamaged leaves of control plants not treated with DEX (type-I treatment) did not accumulate JA-Ile to levels greater than that in the opr3 mutant. The pre-treatment of damaged leaves with DEX (type II) resulted in JA-Ile production (198 pmol g−1 FW) in the damaged leaves, and only a trace quantity (∼3 pmol g−1 FW) of JA-Ile in undamaged systemic leaves. In contrast, the pretreatment of distal leaves with DEX (type III) prior to wounding resulted in the accumulation of JA-Ile in the DEX-treated (systemic) leaves, but not in wounded leaves that were not treated with DEX. As expected, the application of DEX to both the damaged and undamaged leaves (type IV) resulted in wound-induced JA-Ile production in both sets of leaves. These findings indicate that the wound-stimulated production of JA-Ile in leaves distal to the wound site results mainly from de novo synthesis in undamaged leaves, rather than from the transport of JA/JA-Ile from the wound site.

Figure 4.

 Wound-induced systemic accumulation of JA-Ile requires OPR3 function in systemic leaves.
(a) Schematic illustration of four different treatments (types I–IV). The open circles denote leaves that were pre-treated with dexamethasone (DEX). See text for description. For all treatments, the older leaves on Pdex:GFP-OPR3 plants were mechanically wounded with a hemostat. Damaged (loc) and distal undamaged (sys) leaves were harvested 30 min after wounding for the analysis of JA-Ile levels.
(b, c) JA-Ile levels in local (b) and systemic (c) leaves harvested 30 min after wounding. For the unwounded controls (open bars), leaves that were pre-treated, or not, with DEX were harvested from a separate set of plants that were not subjected to wounding. Each data point shows the mean ± SD of three biological replicates.

Rapid exit of the systemic signal from the wounded leaf

The aforesaid results indicated that a rapid wound signal triggers JA-Ile synthesis in undamaged distal leaves. To estimate the time required for the exit of this signal from the wounded leaf, we damaged rosette leaves with a hemostat, and then removed them (by excising the petiole) at various times after wounding (Figure 5a). The JA-Ile content in systemic undamaged leaves was then measured at a single time point (15 min) after the initial wound event (Figure 5d). Control experiments showed that leaf excision alone was sufficient to trigger JA-Ile production (Figure 5b) and the robust expression of response genes (Figure 5c) in leaves distal to the wound site. Because the level of excision-induced JA-Ile accumulation (<50 pmol g−1 FW at 15 min post wounding) was well below the level of wound-induced systemic JA-Ile accumulation (typically >125 pmol g−1 FW), we used the excision-only control value (Figure 5d, dashed line) as a baseline to assess the time required for the passage of the systemic signal from hemostat-damaged leaves. The excision of damaged leaves 1 or 1.5 min after wounding resulted in the systemic accumulation of JA-Ile to levels comparable with that in the excision-only control (C2, Figure 5d). However, the excision of damaged leaves 2 min after wounding resulted in systemic JA-Ile levels (127 ± 21 pmol g−1 FW) that were comparable with that observed in the wound-only control (C3, Figure 5d). Similar results were obtained when wounded leaves were excised 5 or 10 min after wounding. These results demonstrate that the excision of undamaged leaves alone is sufficient for the rapid activation of JA-Ile synthesis and expression of response genes, and suggest that the maximal systemic production of JA-Ile involves a signal that exits the wounded leaf within 2 min of tissue damage.

Figure 5.

 Rapid exit of the systemic signal from wounded leaves.
(a) Diagram of experimental set-up.
(b, c) Effect of petiole excision on systemic JA-Ile accumulation (b) and marker gene expression (c). Leaves (not damaged with a hemostat) were excised at the petiole, and systemic leaves were collected at the indicated time points.
(d) Older leaves on wild-type (Col-0) plants were wounded and then excised from the plant at the indicated time points after wounding. Fifteen minutes after the initial hemostat-inflicted wound, systemic undamaged leaves were harvested for the analysis of JA-Ile content. Control treatments (C1–C3) were as follows: C1, basal level of JA-Ile in untreated (not crushed and not excised) plants; C2, level of JA-Ile in systemic leaves from plants in which local leaves were excised (but not damaged with a hemostat) 15 min prior to harvesting the systemic leaves for JA-Ile measurements; C3, level of JA-Ile in systemic leaves from plants in which the local leaves were wounded with a hemostat (but not excised) 15 min prior to harvesting the systemic leaves for JA-Ile measurements. Each data point represents the mean ± SD of four or five biological replicates.

Activation of the JA receptor does not induce JA-Ile synthesis in the absence of wounding

To test the idea that the wound-induced production of endogenous JA-Ile stimulates its own synthesis during the wound response, we measured JA-Ile accumulation in leaves treated with COR, a potent agonist of the JA receptor (Katsir et al., 2008b; Melotto et al., 2008). Application of COR to rosette leaves resulted in the accumulation of JAZ7 and OPR3 mRNAs within 1 h of treatment (Figure 6a). The level of COR-induced expression of these genes was similar to that observed 15 min after mechanical wounding. Whereas high levels of JA-Ile accumulated in wounded leaves, JA-Ile was not produced in leaves treated with COR (Figure 6b). Consistent with current views on how JA synthesis is regulated (Wasternack, 2007), we conclude that receptor activation and signal output, including the expression of JA biosynthesis genes, is not sufficient to stimulate the production of endogenous JA-Ile in the absence of wounding.

Figure 6.

 Activation of the jasmonate receptor does not induce the de novo synthesis of JA-Ile.
Wild-type (Col-0) plants were either wounded or treated with coronatine (COR) or mock treated (Mock). Leaves were harvested 15 min after wounding or 1 h after COR treatment, and were used for RNA blot analysis (a) or JA-Ile measurements (b). A separate set of plants remained untreated (None). Each data point represents the mean ± SD of three biological replicates.

Wounding causes a rapid depletion of OPDA in systemic leaves

To investigate the mechanism by which wounding activates systemic synthesis of JA-Ile, we used LC MS/MS to determine the time course of accumulation of the JA precursor, OPDA. OPDA levels in damaged wild-type (WS) leaves tripled within 5 min of wounding, and stabilized at a level eight times higher than pre-wound levels at 60 min (Figure 7a). However, in undamaged systemic leaves the free OPDA levels dropped by about 70% within 5 min of wounding (Figure 7c). The rapid systemic decline in free OPDA levels in response to wounding was also observed in the Col-0 ecotype (Figure S4). Previous studies have suggested that OPDA-containing galactolipids (referred to as Arabidopsides) may provide a reservoir of OPDA for JA synthesis in response to wounding (Stelmach et al., 2001; Buseman et al., 2006; Mosblech et al., 2009). We therefore measured the effect of wounding on local and systemic levels of five chemically distinct Arabidopside lipids (A–E). In damaged leaves, Arabidopside levels tripled within 5 min of wounding (Figure 7b), as previously reported (Buseman et al., 2006; Kourtchenko et al., 2007). In systemic undamaged leaves, however, Arabidopside levels were not affected by wounding (Figure 7d).

Figure 7.

 Time course of accumulation of free and lipid-esterified 12-oxo-phytodienoic acid (OPDA) in wounded leaves.
(a–d) Wound-induced accumulation of OPDA (a, c) and OPDA-containing galactolipids (arabidopsides A–E) (b, d) in damaged local (a, b) and undamaged systemic (c, d) leaves. Wild-type (WS) plants were wounded as described in the legend to Figure 1. Each data point represents the mean ± SD of three biological replicates.


Current models of JA signaling indicate that COI1–JAZ interactions promoted by JA-Ile lead to JAZ degradation and the subsequent activation of transcription factors (e.g. MYC2) that control JA-response genes. The fact that JA, MeJA and OPDA fail to stimulate the binding of COI1 to any JAZ protein characterized to date supports the hypothesis that JA-Ile is the major bioactive signal for JA responses (Browse, 2009; Howe and Jander, 2008; Katsir et al., 2008a; Staswick, 2008). Our results strengthen this view by demonstrating that JA-Ile is a key signal for wound-induced systemic gene expression in Arabidopsis. We show that mechanical tissue damage causes a ∼10-fold increase in systemic JA-Ile levels within 5 min of wounding, and up to a 30-fold increase at the peak of JA-Ile accumulation, thereby confirming and extending recent studies (Glauser et al., 2008; Suza and Staswick, 2008; Wang et al., 2008). We also show that the wound-induced systemic increase in JA-Ile precedes the expression of primary response genes, and that gene expression is spatially and temporally correlated with the wound-induced turnover of a JAZ1-GUS reporter protein. Consistent with this observation, Zhang and Turner (2008) reported that the systemic turnover of JAZ1-GUS occurs within 60 min of wounding.

A direct role for JA-Ile in wound-induced systemic gene expression is supported by an analysis of the jar1-12 mutant that produces normal JA levels, but only about 20% of wild-type JA-Ile levels, in damaged and undamaged leaves. Although the expression level of wound-response genes in damaged jar1 leaves was similar to that in the wild type (Chung et al., 2008; Suza and Staswick, 2008; this study), the reduced systemic levels of JA-Ile in jar1-12 plants were correlated with the decreased systemic expression of JA-response genes. We suggest that the level of bioactive JA-Ile in wounded jar1 leaves exceeds the level needed to saturate the transcriptional response, whereas the systemic level of JA-Ile in jar1 is sufficient to produce a response that, although robust, is not saturated. An alternative possibility is that another wound-induced JA signal activates systemic gene expression in jar1 plants (Chung et al., 2008; Katsir et al., 2008b; Suza and Staswick, 2008; Wang et al., 2008). The idea that relatively low levels (∼50 pmol g−1 FW) of JA-Ile can activate robust changes in gene expression is supported by our finding that the excision of leaves from unwounded wild-type plants results in an accumulation of low JA-Ile levels, and strong signal output in distal leaves. The collective data support a model in which wound-induced systemic increases in bioactive JA-Ile promote the SCFCOI1-dependent degradation of JAZs, thereby allowing downstream transcription factors to engage early-response genes.

The steep rise and fall of systemic JA-Ile levels in response to wounding was mirrored by corresponding changes in gene expression. The transient nature of this response implies the existence of an efficient mechanism to metabolize JA-Ile, which may involve the hydroxylation of JA or JA-Ile (Glauser et al., 2008; Miersch et al., 2008). The amplitude and duration of wound-induced JA-Ile accumulation in undamaged systemic leaves differed from that in wounded leaves in two important ways: (i) peak levels of the hormone in damaged leaves were much higher (∼10-fold) than in undamaged leaves; and (ii) high JA-Ile levels (>1000 pmol g−1 FW) in damaged leaves persisted until at least 6 h after wounding. Despite the elevated JA-Ile levels during later stages of the local response, it is noteworthy that early gene (e.g. JAZ7) expression rapidly declined at these time points, suggesting that damaged leaves may become desensitized to JA-Ile. Wound-induced production of dimeric JAZ repressors that are stabilized against hormone-induced degradation may be involved in this form of negative feedback control (Chung and Howe, 2009).

We developed the Pdex:GFP-OPR3 transgenic line as a tool to spatially manipulate wound-induced JA/JA-Ile synthesis in local damaged and distal undamaged leaves. Wounding did not elicit systemic JA/JA-Ile accumulation in Pdex:GFP-OPR3 plants that had the ability to produce JA/JA-Ile in damaged leaves only. Conversely, wounding did induce the systemic production of JA-Ile in Pdex:GFP-OPR3 plants that were complemented for OPR3 function in systemic undamaged leaves, but not in local damaged leaves. We conclude that the wound-induced systemic increases in JA/JA-Ile result from the de novo synthesis of these compounds in systemic leaves, rather than from transport from the site of wounding. Metabolic labeling experiments have also provided evidence that JA-Ile is synthesized de novo in leaves distal to the wound site (Wang et al., 2008). Because application of DEX to systemic leaves provided only partial complementation of the JA deficiency in Pdex:GFP-OPR3 plants, we cannot exclude the possibility that JA or its precursors produced in other parts of the plant (e.g. in petioles) are transported to the distal leaves during the wild-type wound response. The partial complementation of the JA deficiency in Pdex:GFP-OPR3 plants may reflect a failure of DEX to restore OPR3 activity to wild-type levels in all cells of the DEX-treated leaf.

Our results imply the existence of a systemic wound signal that activates JA synthesis in distal leaves, the production of which in damaged leaves does not require OPR3. Based on the results of time-course and petiole-excision experiments, we estimate that the speed of signal movement is less than 2 cm min−1, in agreement with recent work by Glauser et al. (2008). The rapid nature of this response suggests the involvement of a physical signal generated in response to leaf injury (Malone, 1993; Zimmermann et al., 2009). It has been shown that wound-induced changes in hydraulic pressure in tomato cause a basipetal mass flow capable of delivering chemical elicitors to distal leaves via the xylem, in a time frame that can account for the rapid (e.g. within 20 min of wounding) activation of gene expression (Malone, 1993; Malone and Alarcon, 1995).

A common mechanistic feature of the JA-dependent systemic wound response in Arabidopsis, tomato, and tobacco is the COI1-mediated perception of a bioactive JA signal in remote leaves (Li et al., 2002; Wang et al., 2008; this study). Grafting (Li et al., 2005, 2002) and biochemical studies (Hause et al., 2003) performed with tomato have provided strong evidence that a JA signal acts cell non-autonomously to promote wound-induced systemic expression of defensive proteinase inhibitors (Schilmiller and Howe, 2005; Wasternack et al., 2006). In other experimental systems, however, JA/JA-Ile is synthesized de novo in systemic tissues after the arrival of an unknown mobile signal (Wang et al., 2008; this study). Our work with Arabidopsis implies the existence of a JA-independent mobile signal. These and other recent findings (Wu et al., 2007) support the idea that wound-induced systemic responses may involve multiple long-distance signaling systems that operate at different temporal and spatial scales. It is conceivable that the rapid systemic responses we describe here play a role in priming the undamaged tissues for enhanced defense to biotic stress, similar to the priming of rhizobacteria-induced systemic resistance in Arabidopsis (Pozo et al., 2008). The role of wound-activated systemic JA/JA-Ile synthesis in induced resistance to insect herbivores, and the extent to which this rapid signaling pathway is conserved in different plant species, requires further investigation.

How is de novo JA-Ile synthesis activated in leaves distal to the wound site? The speed of the response implies a mechanism involving the rapid mobilization of a JA-Ile precursor or the post-translational activation of a pre-existing biosynthetic enzyme. The observation that systemic increases in JA-Ile are accompanied by increased JA accumulation (Figure 2a) supports the idea that JA pools are limiting for the synthesis of JA-Ile (Suza and Staswick, 2008). Thus, a pathway enzyme preceding JAR1 is likely to be the regulated step in systemic JA-Ile production. Our results showing a temporal link between the rise in systemic JA-Ile and the decline in free OPDA suggests that OPDA metabolism may be a control point for JA/JA-Ile production in distal leaves. In several independent experiments, the decline in free OPDA levels during the first 5 min of wounding ranged from 1–4 nmol OPDA g−1 FW, which is sufficient to account for the wound-induced systemic increase in JA-Ile. A precursor–product relationship between the rapidly metabolized pool of free OPDA and increased JA/JA-Ile remains to be demonstrated. It will be interesting to determine whether the rapid wound-induced depletion of systemic OPDA pools we observed in Arabidopsis occurs in other plant species.

Previous studies in tomato and tobacco have shown that leaf wounding produces a rapid xylem-borne signal that activates mitogen-activated protein kinases (MPKs), and, importantly, that MPKs play a role in the wound induction of JA synthesis (Kandoth et al., 2007; Seo et al., 1995, 1999; Stratmann and Ryan, 1997; Wu et al., 2007). It has also been shown that reversible protein phosphorylation is involved in the wound-induced changes in gene expression in Arabidopsis (Rojo et al., 1998). Elegant studies showing that specific MPK isoforms in Arabidopsis stimulate ethylene production by phosphorylation of ACC synthase (Liu and Zhang, 2004) raise the hypothesis that MPKs regulate the conversion of OPDA to JA in systemic leaves. Transport of OPDA into peroxisomes via the ATP-binding cassette transporter COMATOSE (Theodoulou et al., 2005) represents a potential point of post-translational regulation of OPDA metabolism. The peroxisomal flavoprotein OPR3, as well downstream enzymes in the β-oxidation pathway (Schaller et al., 2005; Browse, 2009), are also potential control points for the wound-induced post-translational regulation of JA synthesis. The X-ray crystal structure of OPR3 revealed that the enzyme forms a self-inhibited dimer, suggesting that OPR3 activity in vivo may be regulated by reversible phosphorylation (Breithaupt et al., 2006). Although the peroxisomal location of OPR3 indicates that it is not likely to be a substrate for cytosolic MPKs, the enzyme may be indirectly regulated by an MPK pathway, or perhaps by protein kinases that reside in peroxisomes (Ma and Reumann, 2008). The systemic wound response of Arabidopsis should provide a useful model system to understand the early signaling events involved in the activation of JA synthesis in response to mechanical stress and other inductive cues.

Experimental procedures

Plant material and growth conditions

Arabidopsis thaliana ecotypes WS and Col-0 were used as indicated as the wild type, and were grown as previously described (Koo et al., 2006). A T-DNA insertion allele (SALK_075487; designated jar1-12) of JAR1 was obtained from the Arabidopsis Biological Resource Center (ABRC, http://www.biosci.ohio-state.edu/pcmb/Facilities/abrc/abrchome.htm) (Alonso et al., 2003). The opr3 mutant and 35S:JAZ1-GUS transgenic line were described previously (Stintzi and Browse, 2000; Thines et al., 2007).


The (±)-JA (J2500), COR (C8115) and dexamethasone (D1756) were purchased from Sigma-Aldrich (http://www.sigmaaldrich.com). Chemically synthesized JA-amino acid conjugates used as standards for LC MS were provided by Dr Paul Staswick (Staswick and Tiryaki, 2004). [13C6]JA-Ile was synthesized as described by Chung et al. (2008). [2H5]OPDA (Delker et al., 2007) was a gift from Dr Claus Wasternack (Leibniz-Institute of Plant Biochemistry).

Plant treatments

Between three and five fully expanded, developmentally older rosette leaves on 30-day-old plants were wounded by crushing the leaf (two or three times) across the mid-rib with a hemostat. At various times after wounding, damaged (local) leaves and undamaged younger (systemic) leaves (four or five per plant) were harvested, frozen in liquid nitrogen and then stored at −80°C until they were used in RNA or JA extraction. The time elapsed between excising and emerging leaves into liquid N2 was <1 min. Solutions containing COR (10 μl per leaf of a 2.5 mm solution in 70% ethanol) and DEX (10 μl per leaf of a 30-μm solution in water) were applied as droplets to the adaxial leaf surface. DEX was applied 16 h prior to wounding.

Plasmid construction and transgenic plants

The Pdex:GFP-OPR3 binary vector was constructed as follows. The full-length OPR3 cDNA was PCR amplified from a cDNA template with the primer pair 5′-GGGCTCGAGATGACGGCGGCACAAGGGAA-3′ (the XhoI site is set in bold) and 5′-CCCACTAGTTCAGAGGCGGGAAAAAGGAG-3′ (the SpeI site is set in bold). The resulting PCR fragment was cloned into the XhoI and SpeI sites of the glucocorticoid-inducible vector system (Aoyama and Chua, 1997) to generate Pdex:OPR3. The open reading frame encoding GFP was amplified from the EGFP vector (Clontech, http://www.clontech.com) with primers 5′-AAACTCGAGATGGTGAGCAAGGGCGAGGA-3′ (the XhoI site is set in bold) and 5′-AAGCTTCTCGAGCCCGGGGAATTCGG-3′ (the XhoI site is set in bold), and then ligated into the XhoI site of Pdex:OPR3. Bacterial colonies containing plasmids (Pdex:GFP-OPR3) in which the GFP fragment was inserted in the correct orientation were identified and verified by DNA sequencing.

Pdex:GFP-OPR3 was transformed into the Arabidopsis opr3 mutant as previously described (Koo et al., 2006). opr3 plants used for transformation were sprayed with a solution of 100 μm MeJA every other day, beginning 3 days before floral dipping until the plants were dried for seed collection. Transformed Pdex:GFP-OPR3 seedlings (T1 generation) grown on soil for ∼10 days were selected for BASTA resistance by spraying a solution containing 0.5 μm glufosinate-ammonium (Finale; AgrEvo Environmental Health, NJ, USA) and 0.00025% Silwet L-77 (Lehle Seeds, http://www.arabidopsis.com) twice a week (for 2 weeks).

Subcellular localization was performed by the Agrobacterium-mediated transient co-expression of GFP-OPR3 and YFP-PTS1 (Orth et al., 2007) in N. tabacum leaves, as described by Chung and Howe (2009). N. tabacum leaves co-infiltrated with Agrobacterium strains harboring Pdex:GFP-OPR3 and 35S:YFP-PTS1 constructs were re-infiltrated, 48 h after the Agrobacterium infiltration, and 16 h prior to confocal imaging, with a solution containing 30 μm DEX. Confocal microscopy was performed with an Olympus FluoViewTM FV1000 microscope (http://www.olympus-global.com). Spectral images were recorded with a 488-nm argon laser line for excitation. The fluorescence emission range was recorded from 495 to 607 nm in 7-nm increments, using a 7-nm-wide slit. GFP and YFP fluorescence signals were separated with the FluoView linear unmixing algorithm provided by the manufacturer. N. tabacum leaves infiltrated with the same two constructs, but not treated with DEX, were used as a control to assess the potential interference between GFP and YFP signals (data not shown).

RNA and protein analysis

RNA extraction was performed with the Trizol reagent (Invitrogen, http://www.invitrogen.com) according to the manufacturer’s instructions. RNA blot analysis was performed as described previously (Li et al., 2002). Gene-specific probes were prepared by PCR amplification of the corresponding cDNA clones (Chung et al., 2008; Koo et al., 2006). A sequence-verified fragment (∼500 bp) of the JR2 (At4g23600) cDNA was amplified by RT-PCR using the primer set of 5′-GCCCCGGGGGCGCGCCACCTTGAGGTCCGCCACTAT-3′ and 5′-CGCCTAGGATTTAAATGCTCGGTCCATAAGAAGGTG-3′. The GFP probe was prepared by PCR amplification of the GFP gene in plasmid EGFP-Peroxi (Clontech). The ACT8 probe was used as a loading control.

Total protein was extracted from rosette leaves frozen in liquid N2 (∼100 mg FW) and was analyzed by protein-blot analysis, as previously described (Schilmiller et al., 2007). Anti-OPR3 and anti-GFP (Molecular Probes, http://www.invitrogen.com/site/us/en/home/brands/Molecular-Probes.html) primary antibodies were used at a 1:5000 and 1:2000 dilution, respectively. A peroxidase-conjugated anti-rabbit secondary antibody (Sigma-Aldrich) was used at a 1:20 000 dilution. Protein–antibody complexes were visualized by incubation with the SuperSignal West Pico Chemiluminescent substrate (Pierce, http://www.piercenet.com) according to the manufacturer’s instructions.

Quantification of metabolites by mass spectrometry

Measurement of JA derivatives and OPDA was performed by LC MS (Chung et al., 2008). Dihydro-JA (dh-JA), [13C6]JA-Ile and [2H5]OPDA were added as internal standards for the quantification of JA, JA-amino acid conjugates and OPDA, respectively. The transitions from deprotonated molecules to characteristic product ions were monitored in electroyspray negative mode for JA (m/z, 209 → 59), dh-JA (211 → 59), JA-Ala (280 → 88), JA-ACC (292 → 100), JA-Thr (310 → 266), JA-Gln (337 → 130), JA-Val (308 → 116), JA-Leu/Ile (322 → 130), JA-Met (340 → 148), JA-Phe (356 → 164), JA-Trp (395 → 203), [13C6]JA-Ile (328 → 136), OPDA (291 → 165) and [2H5]OPDA (296 → 170). Peak areas were integrated, and the analytes were quantified based on standard curves generated by comparing analyte responses to the corresponding internal standard.

Because this method does not distinguish JA-Ile from JA-Leu (i.e. both compounds elute as a single LC peak), values reported for JA-Ile represent the sum of JA-Ile plus JA-Leu. A separate set of control experiments was performed to determine the relative proportion of JA-Leu and JA-Ile in leaf extracts. This was accomplished by selective fragmentation of JA-Leu and JA-Ile, using a method similar to that described for the selective fragmentation of amino acids (Gu et al., 2007). Transitions of m/z 324 → 69 and 324 → 30 were used for JA-Ile and JA-Leu in electrospray positive mode to generate the calibration curves. Details of this methodology will be presented elsewhere. Results from these experiments showed that the quantity of JA-Leu in wounded leaves from 30-day-old Col-0 plants accounted for 5–10% of the JA-Ile + JA-Leu content, and, in the case of the WS ecotype, 10–15% of the JA-Ile + JA-Leu content. This finding is consistent with previous reports (Suza and Staswick, 2008). For the experiments described in Figure 3, JA levels were quantified by GC MS, as previously described (Koo et al., 2006).

Arabidopsides were extracted by heating leaf tissue in 3 ml of a solution of isopropanol containing 0.01% butylated hydroxytoluene (BHT) for 15 min at 90°C. Extraction of the lipid fraction was performed as described by Welti et al. (2002). The organic extract was dried under nitrogen gas and dissolved in a small volume (0.5–1 ml) of chloroform. An aliquot (80 μl) of this sample was prepared for LC by the addition of 25 μl each of a 100-μm solution of hydrogenated monogalactosyldiacylglycerol (MGDG) and digalactosyldiacylglycerol (DGDG) as internal standards (Matreya LLC, http://www.matreya.com), and the addition of 270 μl of a solution containing 15 mm ammonium acetate in methanol. The mixture was centrifuged at 16 000 g for 20 min at 4°C, and 200 μl of the resulting supernatant was transferred to sample vials for LC/time-of-flight (TOF) MS analysis on a Waters LCT Premier instrument (http://www.waters.com). Arabidopsides were separated on an Ascentis Express C18 column (2.7-μm particle size, 2.1 × 50 mm; Supelco, now part of Sigma-Aldrich) using a gradient of 10 mm ammonium formate (solvent A), methanol (solvent B) and isopropanol (solvent C). The gradient consisted of a linear increase from 60% A/40% B to 100% B in 4 min, a ramp to 100% C in 5 min and maintenance of 100% C for 2 min. The gradient was then adjusted to 60% A/40% B in 1 min. The column was re-equilibrated for 3 min prior to running the next sample. The flow rate was 0.4 ml min−1. The mass spectrometer was operated at a resolution of 6000 (full width at half maximum, FWHM) at m/z = 195, in negative ion mode, with a capillary voltage of 3000 V and a sample cone voltage of 20 V. The source and desolvation temperatures were maintained at 100 and 300°C, respectively, with desolvation and cone gas flow rates operated at 300 and 20 L h−1, respectively. Accurate mass LC MS data were collected over the range of m/z values from 50 to 1400 in centroid mode, and with a 0.2-s acquisition time per spectrum. Spectra with confirming fragment ions were generated using the quasi-simultaneous multiplexing of the aperture-1 potential. Details of this methodology will be presented in a separate paper. Data acquisition and processing were performed using Masslynx 4.1 (Waters).


We are grateful to Paul Staswick (University of Nebraska), Claus Wasternack (Leibniz-Institute of Plant Biochemistry) and Yuichi Kobayashi (Tokyo Institute of Technology) for providing jasmonate standards, and Florian Schaller (Ruhr-Universität Bochum) for the gift of the OPR3 antibody. We also acknowledge John Browse (Washington State University) and Rob Larkin (Michigan State University) for providing seed stocks, Jeff Dangl (University of North Carolina) for providing the DEX-inducible vector, and the ABRC for the distribution of cDNA clones and sequence-indexed T-DNA lines generated at the Salk Institute Genomic Analysis Laboratory. We thank Melinda Frame at the Center for Advanced Microscopy (Michigan State University) for assistance with confocal microscopy. We thank Hoo Sun Chung, Eliana Gonzales-Vigil, and Darya Howell for their helpful assistance during the course of the research. This work was supported by the National Institutes of Health (grant R01GM57795), the Office of Basic Energy Sciences at the US Department of Energy (grant DE–FG02–91ER20021), the National Science Foundation (grant DBI-0619489) and the MSU Mass Spectrometry Facility.