Terrestrial plants are subject to attacks by plant-feeding insects. The lipid-derived hormone jasmonate (JA), a collective term used to describe jasmonic acid and its precursors and derivatives, is a key phytohormone that orchestrates many of the defense responses against insects (Wasternack andHause, 2013, Howe et al., 2018). Rapid induction of the JA-dependent signaling pathway is critical for a timely response to fast moving aggressors like insects. Indeed, judging from the speed of JA-responsive marker gene expression, the JA signaling pathway is induced within several minutes of insect herbivory or mechanical tissue injury (Mousavi et al., 2013, Toyota et al., 2018).
The molecular details of transcriptional regulation in the JA signaling pathway have been revealed (Chini et al., 2007, Thines et al., 2007, Yan et al., 2007). The centerpiece for this mechanism is a nuclear residing co-receptor complex consisting of CORONATINE INSENSITIVE 1 (COI1) and a JASMONATE ZIM-domain (JAZ) protein (Xie et al., 1998).
COI1 is the F-box protein part of the E3 ubiquitin ligase complex, Skp1-Cul1-F-box protein (SCF COI1 ) and JAZs are transcriptional repressors of transcription factors (TFs) that control JAresponsive gene expression. The complex formation between COI1 and JAZ facilitated by JA results in the polyubiquitination and subsequent proteolytic degradation of JAZs which then leads to a transcriptional activation of JA-regulated genes. Since the physical interaction between COI1 and JAZ requires the bioactive form of JA, most prominently, (+)-7-iso-jasmonoyl-Lisoleucine (JA-Ile) (Fonseca et al., 2009), it implies that JA must first be present for this transcriptional system to work.
The core JA biosynthetic pathway begins in the plastid and proceeds through the peroxisome before finally being converted to JA-Ile in the cytosol (Vick and Zimmerman, 1983, Schaller and Stintzi, 2009, Koo, 2018, Wasternack and Feussner, 2018). The generally accepted first step of JA biosynthesis is the liberation of 18-carbon fatty acids (FAs) containing three double bonds (C18:3 9,12,15 ) called -linolenic acid (-LA) from phospholipids or galactolipids in the plastid membrane by phospholipase A-type 1 (PLA1) lipases (Conconi et al., 1996, Ryu, 2004, Bonaventure et al., 2011). DEFECTIVE IN ANTHER DEHISCENCE 1 (DAD1) (At2g44810; PLA-Iβ1) is the first established lipase to be involved in JA biosynthesis (Ishiguro et al., 2001). In addition, there are seven PLA1s that group closely with DAD1 in phylogenic trees that also have predicted plastid transit peptides named DAD1-like PLA1s (Rudus et al., 2014). Of these, DONGLE (DGL) was proposed to be the primary lipase involved in woundinduced JA biosynthesis in leaves (Hyun et al., 2008) but a subsequent study disputed the claim and instead identified PLA-Iγ1 as another contributor to JA biosynthesis in wounded leaves (Ellinger et al., 2010). Other recent studies have identified PLASTID LIPASE2 (PLIP2) and PLIP3 to be involved with ABA-induced JA biosynthesis (Wang et al., 2018).
Upon release from the membrane lipids by phospholipases, -LA is converted into cis-(+)-12-oxophytodienoic acid (OPDA) by 13-LYPOXYGENASE (LOX), ALLENE OXIDE SYNTHASE (AOS) and ALLENE OXIDE CYCLASE (AOC) in the plastid. OPDA is further metabolized in the peroxisome by a series of enzymes, including OPDA REDUCTASE 3 (OPR3), OPC-8:0 CoA LIGASE1 (OPCL1), ACYL COA OXIDASE1/5 (ACX1/5) and other oxidation cycle enzymes to produce jasmonic acid. Jasmonic acid is finally conjugated to an amino acid, most prominently, isoleucine by JASMONATE RESISTANT 1 (JAR1) in the cytosol (Staswick and Tiryaki, 2004).
Although the biosynthetic pathway is relatively well characterized, the regulatory aspects of the pathway and how JA biosynthesis is initiated upon wounding remain unclear (Koo and Howe, 2009, Bonaventure and Baldwin, 2010, Scholz et al., 2015, Mielke et al., 2021). The amount of JA in unwounded leaves can vary widely depending on developmental stage and environmental conditions but it is generally very low and only detectable by sensitive modern mass spectrometers (Creelman and Mullet, 1995, Schmelz et al., 2003, Glauser et al., 2009).
Wounding activates rapid de novo synthesis of JA within 2-5 min both locally and systemically (Chung et al., 2008, Glauser et al., 2008, Koo et al., 2009). The fast timing suggests that the biosynthetic capacity (e.g., enzymes) may be already present in untreated leaves before wounding (Maffei et al., 2007).
In this study, we first used a series of physiological, pharmacological, genetic, and kinetic analyses of gene expression and hormone profiling to demonstrate that the early spiking of JA upon wounding does not depend on transcriptional or translational induction of JA biosynthetic genes. We then confirm and add to earlier findings that JA biosynthesis is only limited by substrate availability. Next, by using a transgenic system, we demonstrate how a decoupling between responses to wounding and JA prevents perpetual synthesis of JA in wounded leaves, and identify DAD1-like PLA1 and JAR1 genes to the selective responders to wounding. We then use DAD1 as a wound-inducible model lipase to demonstrate that transient activation of DAD1 transcription can trigger JA biosynthesis to a small extent but that additional wound-activated post-transcriptional steps boost DAD1-mediated JA synthesis. We show that this boosting effect does not require direct tissue damage by showing that it occurs over long distance in undamaged leaves. Finally, we report the findings about DAD1 protein stability under normal and stress conditions, as an example of post-transcriptional mode of regulation.
When rosette leaves (leaf 3,4; local) of 24-d-old Arabidopsis were mechanically wounded using a pair of hemostats jasmonic acid and JA-Ile levels rose linearly for about 30 min (Supplemental Fig. S1A and B), consistent with what have been reported before (Chung et al., 2008). Clear increases can be detected within 5 min. The level of JA before 5 min can be inferred through extrapolating a straight line through the origin (0 min), indicating that JA is made well before 5 min. There was also a rapid synthesis of jasmonic acid and JA-Ile in the systemic undamaged leaves (leaf 6,7) of wounded plants within 5 min of local leaf wounding (Supplemental Fig. S1C and D), as reported before (Glauser et al., 2009, Koo et al., 2009). A time course gene expression analysis was carried out on tissue samples collected together with the above JA profiling samples to see how transcription of JA biosynthetic genes and other JA responsive genes respond to wounding compared with the speed of hormone accumulation. All early (OPR3, OPCL1, JAZ7) and late (JAR1) gene markers were induced by wounding (Supplemental Fig. S1E). The earliest significant increases were detected at around 10 min for OPCL1 and JAZ7, and 20 min for OPR3 (Fig. 1E), and 30 min for JAR1 (Supplemental Fig. S1E). Similar kinetic behavior of transcription was observed in the systemic leaves with increases of OPCL1sys detected the earliest among the four markers at 10 min (Supplemental Fig S1F). Thus, all marker genes were observed to lag behind the increases of JA in both the local and systemic tissues.
Next, we tested whether inhibition of gene expression could affect wound-induced JA biosynthesis. Plants were pretreated with 1 mM cordycepin, a potent transcriptional inhibitor (Sorenson et al., 2018) for 1 h. Such treatment resulted in complete inhibition of JAZ7 and OPR3 transcription by wounding, demonstrating the efficacy of the treatment (Fig. 1A). Hormone measurements in those plants showed that cordycepin treatment did not cause JA levels to change compared to the mock treatment (Fig. 1B). When both the mock and cordycepin pretreated plants were wounded (1 h), there were no measurable differences in jasmonic acid or JA-Ile levels between the two groups (Fig. 1B), showing that transcriptional inhibition of JA biosynthetic genes had minimal impacts on wound-induced JA accumulation.
Building onto a similar idea, we then tested whether inhibition of protein translation could have any impact on wound-induced JA levels. Plants were pre-incubated in buffers containing 0.2 mM cycloheximide (CHX) which is a potent translational inhibitor (Chung et al., 2008).
Presence of CHX eliminated synthesis of DAD1-Myc protein in a transgenic plant (to be described more in later sections), both short term (15, 30, 60 min) and long term (6 h), demonstrating the efficacy of the treatment (Fig. 1C). However, wounding in the presence or absence of CHX had no impact on JA levels (Fig. 1D). CHX by itself did not change JA levels compared to mock as well (Fig. 1D). These results show that blocking transcription or translation has no major effect on wound-induced JA accumulation and that enzymes needed for initial JA biosynthesis are already present in these tissues before wounding.
To test whether JA biosynthesis can be initiated by the induction of several JA biosynthetic genes, Arabidopsis leaves were sprayed with 5 coronatine (COR), a bacterial toxin and a potent mimic of JA-Ile which is known to induce JA responses (Katsir et al., 2008). COR induced most, if not all, JA biosynthetic genes, including LOX2, AOS, AOC1, OPR3, OPCL1, and ACX1 as shown by RNA-Seq experiment (Attaran et al., 2014) (Supplemental Fig. S2A,), as well as by time course qRT-PCR analysis of two marker genes, OPR3 and JAZ8 (Supplemental Fig. S2B). However, the same treatment did not cause endogenous jasmonic acid nor JA-Ile to increase during the course of 12 h treatment (Supplemental Fig. S2C), consistent with earlier studies probing similar questions using structural variants of JA or isotope-labeled JA precursors (Koch et al., 1999, Miersch and Wasternack, 2000, Pluskota et al., 2007, Scholz et al., 2015) and our earlier experiment using COR (Koo et al., 2009). Additionally, to see if pretreatment with COR has any impact on subsequent wound-induced JA biosynthesis, plants were first sprayed with COR for 1 h and then wounded for 30 min. Wounding increased jasmonic acid in both mock and COR-treated plants but there was no additional increase of jasmonic acid levels in the COR pretreated group (Supplemental Fig. S2C). For JA-Ile, there was even a strong reduction by COR pretreatment (Supplemental Fig. S2C), which may be attributed to the increased turnover (Caarls et al., 2017, Heitz et al., 2019, Poudel et al., 2019) since many genes involved in JA-Ile catabolism (e.g., CYP94C1, ILL6) are also induced by COR (Supplemental Fig. S2A).
If induction of biosynthetic gene expression doesn't trigger JA biosynthesis, what will trigger JA biosynthesis? Here, we revisited some of the earlier studies where JA responses were elicited by exogenous supply of JA precursors, such as -LA (Vick and Zimmerman, 1983, Farmer and Ryan, 1992, McConn and Browse, 1996, Christeller and Galis, 2014). In our experiment, we used intact Arabidopsis seedlings and a semi-in vitro system using isolated pea chloroplasts to study the kinetics of the -LA conversion to JA metabolites. When Arabidopsis seedlings (12-d-old) were incubated in a liquid media containing 50 or 100 M -LA, there was a dose-and time-dependent increase of jasmonic acid (Supplemental Fig. S3A and B). The timedependent was clear within 5 min of incubation which lasted until the end of the assay period of 1 h. We additionally treated fad3fad7fad8 mutant with exogenous -LA. The fad3fad7fad8 mutant is deficient in endogenous -LA and C16:3 FAs and consequently cannot produce JA even by wounding (Supplemental Fig. S3). However, when incubated with -LA, fad3fad7fad8 was able to accumulate jasmonic acid to levels equivalent to that produced by WT. This shows that even though fad3fad7fad8 lacks endogenous -LA or JA, it still possesses biosynthetic capacity to convert -LA to JA. The fact that both WT and fad3fad7fad8 produced similar levels of JA from exogenous -LA substrate shows that the bulk of the JA in the WT was also made from the exogenously supplied precursor (-LA) and not by secondary elicitation of more endogenous JA biosynthesis. Another important implication of these results is that the preexisting enzymes of JA biosynthesis in these seedlings are constitutively active and does not require additional activation steps to be able to catalyze the biosynthetic steps.
Next, intact chloroplasts isolated from pea (Pisum sativum) were tested for semi-in vitro synthesis of OPDA by exogenous -LA. Pea has been extensively used for isolating large quantities of intact chloroplasts (Koo et al., 2004). In the absence of exogenous -LA, there was minimal change of OPDA level over a course of 90 min incubation period. However, when supplied with -LA (100 M), a dramatic increase of OPDA (>1,200 pmol/mg chlorophyll (mgChl)) was observed within 5 min that saturated after 5 min (Supplemental Fig. S3E). This is a relatively large amount of OPDA compared to OPDA produced by wounded pea leaves (<50 pmol/mgChl) (Supplemental Fig. S4A). A trace amount of jasmonic acid (<5 pmol/mgChl) and no JA-Ile was detected in the incubation mixture as expected (Supplemental Fig. S4B). The precursor-product relationship was established by feeding these chloroplasts with a stableisotope labeled -LA ([ 13 C1]--LA) which was converted to [ 13 C1]-OPDA (Supplemental Fig. S3F) with no significant increase in unlabeled OPDA. These results demonstrate that similar to the whole seedlings, isolated chloroplasts which rely on all of their OPDA biosynthetic enzymes on the imports from the cytosol (since all are encoded by nuclear-genome) already possess the full biosynthetic capacity and that no other elicitation steps are needed to be able to convert exogenous -LA to OPDA.
Membrane lipid hydrolysis that generates free FAs including -LA is generally considered as the first step of JA biosynthesis. The substrate feeding experiments showing that the entire biosynthetic steps beyond the precursor generation step can run freely which strongly implies that such lipolysis step is the rate limiting step for the initiation of JA biosynthesis. Thus, we examined expression of seven plastid-localized DAD1-like PLA1s, including DAD1, DGL and PLA-11 that have previously been published for their role in JA biosynthesis (Yang et al., 2007, Rudus et al., 2014, Kelly and Feussner, 2016). Of these, five showed varying degree of transcript increases upon wounding (Fig. 2A). Induction was most prominent with DAD1 followed by DGL, PLA1-I⍺2, PLA1-Iβ2, and PLA1-I𝛾1, although statistical significance was weak (p>0.05) for all except DAD1. Even though absolute comparison of expression levels between genes are not accurate for this type of semi-quantitative qRT-PCR analysis, DAD1 consistently gave the highest relative expression level changes followed by DGL and PLA1-Iβ2. Although they are expressed in response to wounding, none of them were induced until 20 min after wounding which is significantly slower than the increase of JA, casting doubts that their transcriptional induction elicits JA biosynthesis. Expression of DAD1-like PLA1s in the systemic leaves of wounded plants also resembled their local expression pattern (Supplemental Fig. S5), showing that the systemic wound signal can trigger their gene expression although, again, their induction needed at least 20 min (Supplemental Fig. S1C and D).
We then looked at the expression of these seven DAD1-like PLA1s in the COR-treated RNAseq data (Fig. 2B) (Attaran et al., 2014). Interestingly, none of them were induced by COR. This is a clear deviation from the other JA biosynthetic genes that are induced by both wounding and COR (Supplemental Figs. S1 and S2). This pattern of expression was further verified with the DAD1 gene by qRT-PCR in a side-by-side wounding vs. jasmonic acid treatment comparisons (Fig. 2C). Expression of marker genes JAZ7 and OPR3 showed that the jasmonic acid treatment was effective. In addition to DAD1, JAR1, another key enzyme in the production of JA-Ile, was also found to be exclusively induced by wounding but not by COR or jasmonic acid (Fig. 2B and C) (Suza and Staswick, 2008).
The observation that expression of DAD1-like PLA1 genes can be turned on preferentially by wounding and not by JA/COR unlike other JA biosynthetic genes may be an important regulatory mechanism for preventing perpetual synthesis of JA. This is because if the entire JA pathway was under a positive feedback regulation, even a minor tissue damage or any other kind of stress or developmental program that causes JA to accumulate will be caught in a neverending loop of JA synthesis. To disprove such auto-amplification theory, we simulated the freerunning "lipase-JA-lipase-JA" circuit in planta. We generated a transgenic plant with a gene construct carrying the recombinant DAD1 gene controlled by a JA-inducible OPR3 promoter.
Function of DAD1 for wound-induced JA biosynthesis in leaves is not proven through loss-offunction studies most likely due to functional redundancies among plastidial PLA1s (Ellinger et al., 2010). However, DAD1 was chosen here based on its proven biochemical activity (Ishiguro et al., 2001) in JA-biosynthesis and it being the strongest wound-inducible PLA1 in leaves (Fig. 2). For the JA-inducible promoter, a 1.5-kb upstream region of OPR3 gene was used that had been shown to be effective in driving the expression of reporter genes in response to JA, wounding or insect herbivory (Body et al., 2019). The resulting transgenic Arabidopsis lines carrying OPR3promoter:DAD1 construct (OPR3pro:DAD1) were severely stunted and constitutively accumulated anthocyanin even when grown under standard growth conditions (Fig. 3A, B and D). These phenotypes were reminiscent of plants grown on JA-containing media.
Introduction of the OPR3pro:DAD1 construct in a JA-deficient aos mutant background suppressed the phenotypes and reverted back to the WT phenotype (Fig. 3A and D), showing that the stressed phenotypes of OPR3pro:DAD1 were due to the JA pathway. RT-PCR analysis of DAD1 transcripts showed that DAD1 was expressed at high levels even in mock-treated OPR3pro:DAD1 plants (Fig. 3C). Additive effects of exogenous JA on DAD1 expression in the OPR3pro:DAD1 plants were not obvious due to the already high basal levels. However, the induction was clearer in the OPR3pro:DAD1/aos plants where the basal level of DAD1 transcripts was low (Fig. 3C). Exogenous JA treatment resulted in a substantial increase of anthocyanin in OPR3pro:DAD1 compared to equally treated WT or OPR3pro:DAD1/aos (Fig. 3D). Hormone measurements showed that OPR3pro:DAD1 plants constitutively accumulated high levels of jasmonic acid (~800 pmol/gFW) compared to WT or OPR3pro:DAD1/aos (<10 pmol/gFW) (Fig. 3E). JA-Ile levels were also higher in the OPR3pro:DAD1 plants except that its relative content compared to jasmonic acid (~0.5%) was lower than those normally observed in wounded tissues (~10%) (Fig. 3E). This is likely contributed by low expression of JAR1 in OPR3pro:DAD1 plants even with increased jasmonic acid as shown earlier (Fig. 2C). These results illustrate the detrimental impacts of having DAD1 expression controlled by JA, and thereby explain why it is necessary to have its promoter not respond to JA.
We next wanted to know whether expression of DAD1 is sufficient to trigger JA biosynthesis or there are other layers of regulation besides gene expression. Although, the chronic JA phenotypes of OPR3pro:DAD1 plants imply that DAD1 expression without wounding can cause JA production, that could be due to OPR3 promoter activity in various cell types at diverse developmental stages beginning from the embryos. In addition, as mentioned before, what was puzzling was that wound induction of DAD1 expression (and other PLA1s) lags far behind JA biosynthesis (Fig. 2 and Supplemental Fig. S5) implying that gene expression is not the primary driver of initial JA burst. However, if preexisting DAD1 (and other PLA1) enzymes are responsible for JA biosynthesis upon wounding, then question rises as to why then they would not cause JA to accumulate in higher levels in resting plants. In order to further probe these questions, we generated a chemical-inducible transgenic system (Supplemental Fig. S6)
where DAD1 transcription can be induced by exogenous application of dexamethasone (dex) that does not occur in plants. Full-length DAD1 fused to a C-terminus Myc epitope tag was cloned into a dex-inducible vector (Aoyama and Chua, 1997), and the resulting construct was transformed into Arabidopsis (Pdex:DAD1-Myc). Out of the 16 T1 plants that survived antibiotic marker selection, six lines displayed significant induction of DAD1 transcript when their leaves were treated with 30 M of dex (Supplemental Fig. S6A). Those six lines also contained more JA (Supplemental Fig. S6B). A homozygous line was selected and used for more detailed analyses.
A time series experiment showed that the increase of DAD1 transcripts can be detected within 4 h of dex application and the levels continued to rise (8 h) (Fig. 4A). Immunodetection of DAD1-Myc using an antibody against the Myc epitope showed DAD1-Myc protein to be also induced by 4 h and continue to increase (8 h) (Fig. 4B). No DAD1-Myc protein was detected prior to the treatment with dex (0 h), showing absolute dependence of its expression on dex. Two of the three detected bands increased upon dex treatment over time and thus appeared to be specific to the DAD1-Myc. The upper weaker band (P) is likely to be the precursor form of DAD1-Myc before the cleavage of the chloroplast transit peptide and the stronger lower band (M) the mature form, judging from their sizes and preferential partitioning to the supernatant and the pellet fractions, respectively, upon centrifugation of chloroplasts (Supplemental Fig. S7A).
Apart from the two, one which appears to be a nonspecific band was detected in all samples. This band appears only when certain batches of commercial Myc antibodies are used (e.g., Fig. 4B, 6A, 7A and Supplemental Fig. S7A) and does not appear when other batches are used (Figs.
5C, 7B-C and 7F). Importantly, correlated with the increases of DAD1-Myc transcripts and proteins, there were increases of jasmonic acid and JA-Ile (Fig. 4C and Supplemental Fig. S7B).
This shows that expression of DAD1 can trigger JA production without wounding. However, the levels were quite low. Compared to ca. 7 nmol/gFW of jasmonic acid induced by wounding (2 h), dex triggered only ca. 0.5 nmol/gFW (Fig. 4C). JA-Ile to jasmonic acid ratio was also low (< 2%) when induced by dex compared to that by wounding (~ 10 %) (Fig. 4C and Supplemental Fig. S7B). This low level of JA-Ile was similar to what was observed in OPR3pro:DAD1 plants (Fig. 3E) and likely to be related to the unresponsiveness of JAR1 expression to JA (Fig. 2C and Supplemental Fig. S7C and D).
We then tested whether wounding has any additional effect on DAD1-induced JA biosynthesis. For this, Pdex:DAD1-Myc plants that had already been treated with dex were subsequently wounded and were compared with those that received only dex or only wounding treatments. Singular treatment with dex had very little effect on the increase of jasmonic acid (~ 50 pmol/gFW) (Fig. 4E) as seen before (Fig. 4C). Wounding raised the level to ~ 8 nmol/gFW which is equivalent to the level reached by wounded WT leaves (Fig. 4E). However, when both dex and wounding were applied together in the Pdex:DAD1-Myc plants, the jasmonic acid level rose to as much as 50 nmol/gFW, which is about 6-fold compared to wounding alone and 1000fold compared to dex treatment alone (Fig. 4E). DAD1 transcript levels remained similar between dex and dex+wound treated plants during these treatments (Fig. 4D). These results show that although DAD1 expression alone (by dex) can elicit some JA synthesis additional wounding significantly enhances this DAD1-mediated JA biosynthesis.
While carrying out these experiments, we also noticed that there were variations in the levels of JA induced by dex between experiments with the jasmonic acid levels ranging from ca. 50 pmol/gFW (Fig. 4E) to ca. 500 pmol/gFW (Fig. 4C). We eventually found that these variations were strongly influenced by the age of the plants (Fig. 5A). JA content was especially very low (< 2 pmol/gFW) in plants younger than 15-d-old but increased slowly until 25-d, and by 30-d, the levels jumped to 7 nmol/gFW. This was not due to variabilities in the inducibility of DAD1 transcripts nor DAD1 proteins by dex in these plants (Fig. 5B and C). It was also not due to the variability in their abilities to synthesize JA by wounding in these different-aged plants (Fig. 5D). The reason for this developmental variation is still unclear but the observations from earlier stage (<15-d) plants are especially telling because they represent cases where presence of abundant DAD1 proteins is failing to trigger JA biosynthesis, implying additional layer(s) of regulation for JA biosynthesis upon wounding. Wounding can unlock this limitation imposed upon DAD1 regardless of their ages, because the cotreatment with dex and wounding resulted in large boosts of JA synthesis in both stages (Fig. 4E and Fig. 5E).
The wound signal that amplifies JA production by DAD1-Myc can be transmitted systemically over a long distance This putative regulatory element that is activated by wounding to boost DAD1-Mycmediated JA synthesis may be physicochemical by nature, happening as a result of random cell breakage rather than by a controlled signaling mechanism. Examples of such uncontrolled events may include random mixing of DAD1-Myc enzymes with broken membrane debris out of cellular context. This possibility can be tested by looking at the systemic wound responses in undamaged intact tissues. If DAD1-Myc proteins that had been pre-induced (by dex) in these systemic leaves can be activated to boost JA biosynthesis as seen earlier by systemic wound signal coming from remote damaged leaves, this would discount the artifactual uncontrolled enzyme theory and favor a regulated posttranslational activation model. To pre-induce systemic DAD-Myc expression, the systemic leaves (leaf 6 and 7) were treated with dex for 6 h prior to any wounding (Fig. 6). Then, the untreated local leaves (leaf 3 and 4) were wounded. JA was measured in the systemic undamaged leaves 15 min after the local leaf wounding. As a control, the same experiment was carried out on Pdex:DAD1-Myc that had not been treated with any dex.
As expected, DAD1-Myc protein was induced by the dex treatment in the systemic leaves but not in the local untreated leaves (Fig. 6A). Wounding of the control plants (no dex) increased systemic jasmonic acid levels to ~70 pmol/gFW (Fig. 6B inset). The dex-alone in the systemic leaves induced jasmonic acid to ~7 nmol/gFW (UW+dex in Fig. 6B) as seen before (Fig. 5).
However, when the local leaves were wounded this systemic jasmonic acid level rose to ~35 nmol/gFW (W+dex in Fig. 6B), a 5-fold increase from dex alone and a 500-fold increase from wounding alone.
DAD1 protein is unstable and can be stabilized by wounding but degrades more quickly in the presence of -LA One of our hypotheses was that the molecular target for the signaling event that boosted DAD1-Myc-mediated JA biosynthesis may be the lipase itself. We began exploring this possibility by monitoring the dynamics of DAD1-Myc proteins levels over time (Fig. 7A). DAD1-Myc was first induced by treating Pdex:DAD1-Myc plants with dex for 12 h, and then 0.2 mM CHX was added to inhibit protein translation. Protein extracts at various time points were then probed with antibodies against Myc (for DAD1-Myc) and compared with other JAbiosynthetic marker proteins, LOX, AOC and JAR1 (Fig. 7A). Interestingly, in contrast to LOX, AOC, or JAR1 that remained largely unchanged over the monitored period of 3 h, DAD1-Myc levels were markedly reduced by 1 h indicative of faster protein turnover. Using similar assay, we then tested whether DAD1-Myc protein stability is affected by the presence of -LA which is the product of DAD1 catalysis and the primary precursor of JA biosynthesis (Fig. 7B). Inclusion of different concentrations of -LA in the incubation media was found to promote the degradation of DAD1-Myc protein, resulting in a clear reduction of signal by 40 min with 50 M -LA and 10-20 min with 100 M -LA. Encouraged by the results, we then tested whether wounding has any effect (Fig. 7C). In contrast to -LA, wounding delayed degradation of DAD1-Myc, maintaining signals after 90 min when most of them in the no wounding samples disappeared. Next, we created a catalytically inactive version of DAD1-Myc enzyme, DAD1 mut -Myc, by introducing Ala substitutions to the amino acid residues that constitute the catalytic triad (S295A, D352A, and H416A) (Ishiguro et al., 2001) which is critical for the catalytic activity of lipases (Fig. 7D). Consequently, stably transformed Arabidopsis lines expressing this mutated version of gene (Pdex:DAD1 mut -Myc) were not able to induce JA biosynthesis by dex nor could boost its accumulation by dex+wound treatment (Fig. 7E). Interestingly, the mutated DAD1 mut -Myc protein persisted longer than DAD1-Myc (Fig. 7F). Together, these results show that DAD1 protein stability may be one of the targets of regulation, perhaps contributing to elicitation of JA biosynthesis in wounded leaves.
Although JA biosynthetic pathway has been elucidated (Vick andZimmerman, 1983, Wasternack andHause, 2013) and transcription of biosynthetic genes has been extensively studied (Reymond et al., 2004, Devoto and Turner, 2005, Pauwels et al., 2009, Howe et al., 2018), how JA can be produced so quickly by wounding remains unclear. The fast synthesis of JA by wounding cannot be easily explained by a model that depends on transcriptional activation of JA biosynthetic enzymes. The biosynthetic gene transcripts lagged far behind JA accumulation both in the local and systemic leaves (Fig. 2, Supplemental Figs. S1 and S5).
Induction of JA biosynthetic gene expression by COR or JA could not elicit de novo synthesis of JA, consistent with earlier reports (Koch et al., 1999, Miersch and Wasternack, 2000, Pluskota et al., 2007, Koo et al., 2009, Scholz et al., 2015) nor did it result in more synthesis of JA when wounded subsequently (Supplemental Fig. S2). This is consistent with failed previous attempts to elevate basal JA levels by simply overexpressing a few biosynthetic enzymes in transgenic plants even though subsequent wounding of these plants may have raised JA levels over WT (Laudert et al., 2000, Bachmann et al., 2002, Stenzel et al., 2003, Sharma et al., 2006, Wasternack, 2007, Rudus et al., 2014) despite claims of success by some others (Harms et al., 1995). More direct evidence against the transcriptional activation model came from our inhibitor studies where blockage of either transcription or translation had no major impact on wound induced JA biosynthesis (Fig. 1).
In fact, the positive feedback model of JA synthesis by JA-responsive gene expression can be problematic. This was demonstrated by our transgenic system (OPR3pro:DAD1) that was designed to produce more JA by JA-responsive gene expression (Fig. 3). We found that there may be at least two steps of JA biosynthesis that prevent this run-on feedforward mechanism.
One is the lipid hydrolysis step at the beginning of JA biosynthesis and the second is the final conjugation step that joins nascent jasmonic acid with Ile. All seven plastidial DAD1-like PLA1s and JAR1 selectively responded to wounding but not to JA or COR (Fig. 2). In addition, on multiple occasions (Fig. 3E and Supplemental Fig. S7), JAR1 acted as a limiting factor for JA-Ile increases despite high jasmonic acid levels. For examples, JA-Ile levels were ~ 0.5 % of jasmonic acid levels both in OPR3pro:DAD1 and dex-induced Pdex:DAD1-Myc. These levels are much lower compared to the wound response where JA-Ile levels typically reach 5-20% of the jasmonic acid levels (Supplemental Fig. S1A and B) (Suza and Staswick, 2008) or even higher (> 20%) in the systemic leaves (Supplemental Fig. S1C and D). In this way, plants seem to have been selecting to exclude PLA1s and JAR1 from the positive feedforward mechanism which is prevalent among other JA metabolic and signaling genes.
With regards to the major limiting factor for JA biosynthesis, there is a wide consensus among scientists that it is the substrate or the precursor availability (Vick and Zimmerman, 1983, Farmer and Ryan, 1992, McConn and Browse, 1996, Wasternack, 2007, Christeller and Galis, 2014). Our substrate feeding assays using whole Arabidopsis seedling and isolated pea chloroplasts (Supplemental Fig. S3A-B) once again made it clear that exogenously supplied -LA can be converted to JA without wounding. It is important to point out the implications of this observation, that is, all enzymes in the pathway are present and active in leaves prior to wounding. This is not to say that any other additional regulatory step(s) may not exist, for example, to fine tune the catalytic potential of various enzymes in the pathway, but to emphasize the fact that the biosynthesis will run its full course without any need for an intervention as long as the substrate is available. The fact that fad3fad7fad8 could metabolize exogenous -LA to JA with the same rigor as the WT showed that the basal level of biosynthetic enzymes is maintained even in the absence of JA (Supplemental Fig. S3D). The fad3fad7fad8 results also discount the possibility of -LA acting as a signaling molecule to initiate JA biosynthesis (e.g., by triggering more lipid hydrolysis) rather than being a mere substrate for these feeding experiments. This is further supported by the stable isotope feeding experiment where [ 13 C]-labelled -LA was converted to [ 13 C]-OPDA without any increases in endogenous OPDA (Supplemental Fig. S3F).
The chloroplast feeding experiment also shows that the chloroplasts even when isolated from the cytosol have enough enzymes to sustain initial synthesis of OPDA. The reaction, however, plateaued within 5 min. It is not clear whether this is due to the complete exhaustion of -LA substrates or cessation of enzyme catalysis (either by turnover or inactivation). Immunoblots seem to indicate that at least LOX and AOC proteins seem to be present for hours (Fig. 7A) although this may be little different in isolated chloroplasts where fresh supply of enzymes from the cytosol is cut off. In addition, we routinely observed that the isolated chloroplasts are active hours after preparation. The substrate exhaustion theory can be explained by other metabolic sinks besides OPDA synthesis, such as incorporation into glycerol lipids (Koo et al., 2005).
Alternatively, we cannot rule out the possibility of feedback inhibition by OPDA. More biochemical assays are needed to explore this avenue of research.
Several lipases involved in JA biosynthesis have been described from multiple species (Ishiguro et al., 2001, Hyun et al., 2008, Kallenbach et al., 2010, Cai et al., 2014, Wang et al., 2018). However, those involved in wound-induced JA biosynthesis in Arabidopsis have remained elusive due to problems attributed to gene redundancy (Ellinger et al., 2010). Our gene expression study identified several of these plastidial DAD1-like PLA1s to be induced by wounding both in the local and systemic leaves. Of these, we chose DAD1 as a model lipase in our study because of its proven biochemical function in JA biosynthesis. Although contribution of DAD1 to wound-induced JA biosynthesis in leaves is not demonstrated through loss-offunction studies (Ellinger et al., 2010) and DAD1 is considered mainly as a flower lipase, its strong induction by wounding in leaves combined with its enzymatic activity makes it likely to contribute to JA biosynthesis in the wounded leaves. JA accumulation in both the OPR3pro:DAD1 and Pdex:DAD1-Myc plants support this prediction. However, there is more to it than its mere presence that elicits JA biosynthesis.
First, the fast speed of wound-elicited JA biosynthesis (Supplemental Fig. S1) would suggest action of pre-existing DAD1 (and other PLA1s). But the low basal levels of JA in untreated leaves would indicate their limited activity. Second, the chronically stressed phenotypes of OPR3pro:DAD1 plants would suggest its enzymatic action in the absence of wounding (Fig. 3). It must however be taken into account that OPR3 promoter driven DAD1 can potentially be expressed in all cell types throughout all developmental stages. In addition, since OPR3pro:DAD1 is under the positive feedforward regulation by JA, all these cells at all developmental stages are expected to contain higher levels of JA than those in the WT plants.
Nevertheless, it does show that when DAD1 is present (in high levels) it will cause JA to be made. Third, transient expression of DAD1-Myc in Pdex:DAD1-Myc plants induced JA accumulation (Fig. 4). However, the JA levels varied depending on the developmental stage of the leaves with the younger than 15-d old plants showing minimal accumulation. In fact, the JA levels stayed relatively low in most developmental stages until the plants reached 30 d (Fig. 5).
The fact that the induced DAD1-Myc protein levels remained similar throughout these developmental stages and that there was essentially no difference in their abilities to synthesize JA upon wounding clearly show that there is more to it than the mere presence of DAD1 proteins for the full elicitation of JA biosynthesis. This is consistent with previous reports showing minimal increases in JA when DAD1 or other PLA1s were expressed under cauliflower mosaic virus 35S promoter (Ellinger et al., 2010, Rudus et al., 2014). These studies were mostly focused on wound-induced JA levels but their unwounded data show little change from the controls. It may be worthwhile to note that sudden increase in JA at 30 d coincide with the vegetative-to-reproductive transition. The significance of this is currently unknown. Fourth, regardless of plant ages, wounding boosted the production of JA to several fold higher than that by DAD1-Myc expression alone (by dex) or by wounding alone (Figs. 4 and 5). This boosting effect cannot be explained by differences in DAD1-Myc expression levels since they were not different whether the plants were wounded or not whether the plants were young or old. It is also unlikely scenario that DAD1-Myc has somehow impacted other lipase activity. This boosting effect can be recapitulated in remote undamaged leaves indicating that the effect is based on a signaling event rather than by misregulation, for example, by random mixing of enzymes and substrates in broken cells. Thus, it seems like DAD1 when expressed can induce JA biosynthesis. However, the degree to which it can elicit JA is heavily influenced by the cell type and developmental stage contexts. Importantly, wound signal can lift sanctions imposed upon DAD1, allowing it to exert its greater potential. DAD1-Myc protein was found to be more labile than other biosynthetic enzymes such as LOX, AOC or JAR1 (Fig. 7). Fast turnover is a hallmark of tight regulation. Plants may have to maintain certain level of lipases all the time for rapid response upon attacks. For this, lipases may be expressed at low levels all the time. However, plants seem to degrade them constantly to avoid their overaccumulation. The benefit must be outweighing the high cost of maintaining this system. DAD1 stability was also affected by the presence of -LA and by wounding. Mutation in catalytic region could delay the degradation. While this does not explain, and may not even be directly linked to, the initiation of JA biosynthesis, it shows that the DAD1 stability is sensitive to factors occurring during wound-induced JA biosynthesis. More research is needed to understand the nature of this modulation of DAD1 protein stability by -LA and wounding, and its relationship to the regulation of JA biosynthesis.
Arabidopsis was grown under long day conditions (16 h light) with 100 -120 E m -2 s -1 light intensity in growth chambers kept at 22 °C. The wild-type (WT) used for all experiments was Arabidopsis thaliana ecotype Columbia-0 (Col-0). fad3fad7fad8 (McConn and Browse, 1996) was a gift from Dr. John Browse at Washington State University. Seedlings were either grown on solid Murashige and Skoog (MS) media (Caisson Laboratories) (0.7% w/v phytoblend agar, 0.7% w/v sucrose) or on soil. Pisum sativum var. little marvel (Green Seed Company, Springfield, MO) used for chloroplast isolation was surface sterilized by 50 % bleach and imbibed at 4 °C for two days prior to being sowed on a soil mixture of half perlite half soil. The peas were grown in short day conditions (10 h light / 14 h dark, 100 -120 E m -2 s -1 ) at 22 °C.
All tissues were harvested and flash frozen in liquid nitrogen and stored in -80 °C until use.
(±)-Jasmonic acid, methyl jasmonic acid (MeJA), coronatine (COR), -LA ((9Z,12Z,15Z)octadeca-9,12,15-trienoic acid)), [ 13 C1]--LA, dexamethasone (dex), cycloheximide (CHX), and cordycepin were purchased from MilliporeSigma (Burlington, MA). JA-Ile, [ 13 C6]-JA-Ile, OPDA, [ 2 H5]-OPDA, and dihydro-JA have been described previously (Koo et al., 2009). The primary antibody for JAR1 was raised in rabbits following the company's instruction (Cocalico Biologicals, Stevens, PA). The full-length JAR1 cloned into a pET28a vector with an N-term His tag (Westfall et al., 2012) was a gift from Dr. Joseph Jez from Washington University in St.
Louis. The construct was transformed into Rosetta 2 (DE3) cells and the recombinant protein was purified using Ni-nitrilotriacetic acid affinity chromatography (Qiagen, Hilden, Germany) followed by desalting and concentration via ultrafiltration (Amicon Ultrafilters, MilliporeSigma) before sending to the company. Polyclonal antibody against Myc tag (rabbit), plastidial LOX, and peroxidase-conjugated anti-rabbit secondary antibody were purchased from Abcam (Cambridge, UK), Agrisera (Sweden) and MilliporeSigma, respectively. Anti-AOC (Hause et al., 2000) was gift from Dr. Bettina Hause of Leibniz Institute of Plant Biochemistry, Halle (Saale), Germany.
Sequence information for DNA primers used in this study is in Supplementary Table S1.
The Pdex:DAD1-Myc plant binary vector construct was made by PCR-amplifying the full-length DAD1 open reading frame (ORF) fused to 4×Myc epitope tag using overlapping PCR technique using SpeI_DAD1_F and SpeI_FlagMyc_R primers and an equimolar mixture of DAD1 ORF and 4×Myc DNA fragments as templates. The two template fragments were each individually prepared by PCR using primer sets SpeI_DAD1_F and ov_DAD1-Myc_R, and ov_DAD1-Myc_F and SpeI_Myc_R, respectively. The resulting DAD1-Myc was cloned into the SpeI sites of the glucocorticoid-inducible vector system (Pdex) (Aoyama andChua, 1997, Koo et al., 2009) to generate Pdex:DAD1-Myc. The site-directed mutagenesis of Pdex:DAD1 mut -Myc was done using the Q5 Site-Directed Mutagenesis Kit (New England BioLabs, Ipswich, MA) following the manufacturers protocol. Mutagenic primers were designed using the NEBase Changer tool and can be found in Supplementary Table S1. Mutagenesis reactions were performed on the above generated DAD1-Myc ORF cloned in pGEM-Teasy vector (Promega, Madison, WI). The resulting construct with substitutes in the lipase consensus motif GSHLG to AAAAA was then subcloned into Pdex vector using the SpeI restriction site. The OPR3pro:DAD1 was constructed by first putting the DAD1 into a modified pBI121 binary vector (Schilmiller et al., 2007) using the SpeI site. A 1.5 kb promoter region of the OPR3 was amplified by PCR from the WT genomic DNA using ClaI_OPR3p_F and BamHI_OPR3p_R primers and the resulting PCR product was cloned in front of the DAD1 using ClaI and BamHI sites.
Above generated three plasmids were first transformed into the C58C1 strain of Agrobacterium tumefaciens, then into WT or aos backgrounds using the floral dip method (Clough and Bent, 1998). The flowers of aos were sprayed with 100 M MeJA solution once every day beginning from 3 days prior to and 5 days post the floral dipping. Seeds harvested from resulting plants (T1) were screened for resistance to either glufosinate-ammonium (10 g mL -1 ) for Pdex:DAD1-Myc and Pdex:DAD1 mut -Myc or kanamycin (50 g mL -1 ) for OPR3pro:DAD1-Myc.
Wounding was administered by crushing across the midrib two to three times using a pair of serrated-tip hemostats. Systemic wounding was performed similarly except that two adult leaves (leaf 3,4) were wounded per rosette and two younger unwounded leaves (leaf 6,7) were harvested. Solution containing dex (30 in 0.01% Triton X-100) was either sprayed (to saturation) or applied as tiny droplets (20-30 L per leaf) to the adaxial side of the leaf for the indicated durations (typically 6-12 h). -LA feeding of Arabidopsis seedlings was done by incubating the seedings in liquid MS media containing -LA. -LA was dried under N2 gas with few droplets of ammonium hydroxide and reconstituted in liquid MS media (< 0.01% DMSO) to their final concentrations. For CHX treatment, 12-14-d old seedlings were transferred from MS plate to liquid MS media containing 0.2 mM CHX and incubated for the indicated times.
Cordycepin treatment was carried out similarly in liquid MS media containing 1 mM of cordycepin 1 h before subsequent treatments. Wounding of CHX/cordycepin treated seedlings was done by indiscriminately but consistently crushing seedlings several times with hemostats while submerged in the media. COR (5 M) was sprayed evenly on the surface of fully expanded mature leaves for indicated times.
Intact chloroplasts were isolated from 10-12-d old pea seedlings using continuous Percoll (GE Healthcare Life Sciences, Chicago, IL) density gradient method (Perry et al., 1991).
Chlorophyll content was determined according to method by (Arnon, 1949). -LA feeding assay was carried out largely according to previously described methods (Koo et al., 2004). The reaction mixture consisted of 200 L incubation buffer containing 100 M -LA and 50 g chlorophyll-equivalent chloroplasts. The reaction was initiated by adding chloroplasts to the reaction mixture in ambient temperature (25 °C) under light (80 -100 E m -2 s -1 ) while constantly shaking on a benchtop orbital shaker. Reaction was stopped by adding equal volume (200 L) of 100% methanol containing internal standards for hormone analyses. Crude chloroplasts from Arabidopsis for immunoblot analysis were isolated based on procedures by (Salie et al., 2016) with modification. About 10 g of freshly harvested 20-25-d old leaf tissues were homogenized in ice-cold grinding buffer (50 mM HEPES-KOH pH 8.0, 330 mM sorbitol, 1.5 mM MnCl2, 2 mM MgCl2, 2 mM EDTA, 0.1% (w/v) BSA) using a polytron homogenizer (Kinematica, Switzerland). The homogenate was passed through two layers of Miracloth, prewet with homogenization buffer, and then centrifuged at 2,600 ×g at 4 °C for 20 min. The supernatant was kept as the fraction containing no chloroplasts. Chloroplasts in the pellet fraction were lysed for 30 min in ice-cold lysis buffer (50 mM HEPES-KOH pH 8.0, 10% (v/v) glycerol, 0.5% (v/v) Triton X-100). Lysates were then further homogenized in a Dounce homogenizer on ice and then centrifuged at 30,000 ×g for 20 min at 4 °C.
Proteins were typically extracted with 50 mM sodium phosphate buffer (pH 7.0) containing 10% glycerol, 250 mM NaCl, 0.1% SDS, 1% Triton, and protease inhibitor tablets (Thermo Scientific, Waltham, MA). Frozen tissue (50-100 mg) was ground into a fine powder using TissueLyserII (Qiagen). About three times the tissue weight volume of extraction buffer was added and upon brief vortex, samples were spun down for 1 min at 16,000 x g to remove debris.
Samples were incubated with sample buffer consisting of 2×Laemmli buffer and 6 M urea at 37 °C for 30 min before loading into 10% SDS-PAGE gel (20 g per lane). For immunoblot analysis, proteins were transferred to PVDF membrane and probed with primary antibodies against AOC, Myc and JAR1 at a 1:3000 dilution and LOX at a 1:15,000 dilution. An anti-rabbit secondary antibody (MilliporeSigma) conjugated to peroxidase was used at a 1:15,000 dilution.
Protein-antibody complexes were visualized with SuperSignal West Pico Chemiluminescent substrate (Thermo Scientific) and exposed on an X-ray film (Midwest Scientific, Valley Park, MO).
Total RNA was extracted from 50-100 mg of frozen tissues that were ground to a fine powder while frozen. Ten times tissue volume of TRIzol reagent (Thermo Fischer Scientific, Waltham, MA) was used for extraction followed by purification using the Direct-zol RNA MiniPrep Plus Kit (Zymo Research, Irvine, CA) following manufacturers' instructions. One g of RNA was reverse transcribed using the iScript Reverse Transcription Supermix (BioRad, Hercules, CA) following manufacturer's instructions. This was used as a template for semiquantitative PCR (qPCR) with the iTaq SYBR Green Supermix (BioRad) in a CFX96 Touch real-time PCR detection system (BioRad). ACT8 (At1g49240) was used as the internal reference gene. The oligonucleotide primers for JAZ7, JAZ8, OPR3, JAR1 and ACT8 have been described Arabidopsis seedlings (14-d) were incubated with mock (0.01% w/v ethanol in water) or 1 mM Cordy in liquid MS media for 1 h before wounding (1 h (A) or 0.5 h (B)). C-D, Protein immunoblot (C) of DAD1-Myc protein expression and JA levels (D) in Pdex:DAD1-Myc plants incubated in MS media with or without 0.2 mM cycloheximide (CHX). After 15 min of preincubation, DAD1-Myc expression was induced by adding 30 M dexamethasone (dex) and incubated for shown duration of time. DAD1-Myc in the protein extract was probed with an antibody against Myc epitope tag. For JA measurements, the CHX-treated and untreated (mock) plants were either not wounded (UW) or wounded for another 1 h. Bar graphs represent mean ± SD of three biological replicates. Letters in graphs indicate statistical significance (P < 0.05) as determined by pairwise t-tests. at 15-d stage. E, Jasmonic acid and JA-Ile in WT, OPR3pro:DAD1 and OPR3pro:DAD1/aos grown as in (A). Data represent mean ± SD of three biological replicates with statistical significance (pairwise t-tests, P < 0.05) denoted by letters above the bars. or both (dex+W). Ten L of 30 M dex was added as small droplets on the adaxial surface of the leaf and incubated for 8 h. Wounding was administered by crushing leaves twice across the midrib using a hemostat at 6 h post dex treatment, and tissue was harvested after 2 h of wounding. M dex was added as small droplets on the adaxial surface of the leaf and incubated for 8 h.
Wounding was administered by crushing leaves twice across the mid-rib using a hemostat at 6 h post dex treatment, and tissue was harvested after 2 h of wounding. Data represent mean ± SD of three biological replicates with statistical significance (pairwise t-tests, P < 0.05) denoted by letters above the bars. Samples marked with asterisk were treated with dex for entire duration (18 h) without CHX. 35 Fig. 2 36 Fig. 3 37 Fig. 4 39 Fig. 6 40 Fig. 7