Damage-inducible defenses in plants are controlled in part by jasmonates, fatty acid-derived regulators that start to accumulate within 30 s of wounding a leaf.
Using liquid chromatography–tandem mass spectrometry, we sought to identify the 13-lipoxygenases (13-LOXs) that initiate wound-induced jasmonate synthesis within a 190-s timeframe in Arabidopsis thaliana in 19 single, double, triple and quadruple mutant combinations derived from the four 13-LOX genes in this plant.
All four 13-LOXs were found to contribute to jasmonate synthesis in wounded leaves: among them LOX6 showed a unique behavior. The relative contribution of LOX6 to jasmonate synthesis increased with distance from a leaf tip wound, and LOX6 was the only 13-LOX necessary for the initiation of early jasmonate synthesis in leaves distal to the wounded leaf.
Herbivory assays that compared Spodoptera littoralis feeding on the lox2-1 lox3B lox4A lox6A quadruple mutant and the lox2-1 lox3B lox4A triple mutant revealed a role for LOX6 in defense of the shoot apical meristem. Consistent with this, we found that LOX6 promoter activity was strong in the apical region of rosettes. The LOX6 promoter was active in and near developing xylem cells and in expression domains we term subtrichomal mounds.
Potent cyclopentanone mediators of the wound response in plants are made when leaves are damaged. These compounds, jasmonates (Wasternack, 2007; Browse, 2009; Schaller & Stintzi, 2009), are derived from plastidial fatty acids, and include the powerful regulator jasmonoyl-isoleucine (JA-Ile) (Staswick & Tiryaki, 2004; Fonseca et al., 2009). Biologically active jasmonates reprogram the expression of defense genes that lead to reduced feeding in a wide variety of invertebrates (Howe & Jander, 2008) and in at least one vertebrate herbivore (Mafli et al., 2012). A first striking feature of jasmonate synthesis is that it is complex, requiring three cellular compartments, at least 10 intermediates, and several inter-organellar transport steps (Wasternack, 2007; Acosta & Farmer, 2010). The process begins in plastids with the dioxygenation of fatty acids in reactions catalyzed by 13-lipoxygenases (13-LOXs), which are conserved proteins found in plants (Andreou & Feussner, 2009), protists, fungi, some bacteria and numerous metazoans (Ivanov et al., 2010). In jasmonate synthesis, 13-LOXs oxygenate α-linolenic acid in a regio- and stereo-selective manner to produce 13(S)-hydroperoxylinolenic acid. Then, this 18-carbon precursor is dehydrated and cyclized to 12-oxo-phytodienoic acid (OPDA) in the plastid before export to peroxisomes. In the peroxisome, OPDA and its 16-carbon homolog dinorOPDA, both cyclopentenones, are reduced to cyclopentanones and thereafter shortened by successive β-oxidation steps leading to the 12-carbon prohormone jasmonic acid (JA). Finally, the process of biosynthesis of regulators like JA-Ile is terminated when JA is exported from the peroxisome for conjugation to L-isoleucine (Staswick & Tiryaki, 2004).
An additional notable feature of inducible jasmonate biosynthesis is that it is not confined to the site of tissue damage, although the nature of the signals that lead to early increases in the levels of jasmonates in distal tissues is still debated (Koo & Howe, 2009). However, in Arabidopsis thaliana, these jasmonate-inducing signals travel in the low centimeter per minute time range (Glauser et al., 2008; Koo et al., 2009) from a wounded leaf to distal leaves that share vascular connections (Glauser et al., 2009). Importantly, being among the first steps of jasmonate synthesis, 13-LOXs are therefore likely candidates as decoders of signals that originate from wounds. All four 13-LOXs encoded in the A. thaliana genome are capable of oxygenating α-linolenic acid in vitro (Bannenberg et al., 2009). However, the relative contribution of each of these LOXs (LOX2, LOX3, LOX4 and LOX6) to jasmonate synthesis in the wounded plant is unknown. To date only LOX2 in A. thaliana has a well-characterized role in wound-induced jasmonate synthesis. This isoform produces the bulk of the JA measurable in a leaf 1–4 h after mechanical damage (Bell et al., 1995; Schommer et al., 2008; Glauser et al., 2009; Seltmann et al., 2010). However, the lox2-1 mutant still displays JA and JA-Ile synthesis in the first 5 min after wounding (Glauser et al., 2009), meaning that other 13-LOXs must also be active in vivo. After wounding, the first measurable increases in JA concentrations were reported to occur c. 30–45 s (Glauser et al., 2009) but, because JA accumulation is exponential (Glauser et al., 2009), and also because of the fact that low basal concentrations of JA exist in healthy leaves, the exact initiation time of wound-induced JA synthesis is unknown. It is nevertheless clear that JA accumulation in distal leaves lags behind that in the wounded leaf, but is measurable within 2 min after wounding a leaf sharing vascular connections (Glauser et al., 2009). The accumulation of JA-Ile is also rapid, occurring within 2 min in the wounded leaf and slightly later in leaves distal to a wound (Koo et al., 2009).
Here, using all possible combinations of double, triple and quadruple A. thaliana 13-LOX mutants, we analyzed JA and JA-Ile concentrations in a 190-s timeframe after wounding. We identified the 13-LOX necessary for the initiation of JA synthesis in tissues distal to a wound and found that its impact increased with distance from the damage site. The results of this study show which enzyme contributes most to distal JA and JA-Ile synthesis soon after wounding, and they reveal when and where this occurs. We then designed experiments to investigate potentially cryptic biological functions of this gene in defense against a generalist herbivore. The results of this assay were consistent with the activity patterns of its LOX6 promoter.
Materials and Methods
Plant material, growth and wounding experiments
Arabidopsis thaliana (L.) Heynh. T-DNA insertion mutants were obtained from the European Arabidopsis Stock Center (NASC). At least two different alleles for LOX3 (At1g17420; lox3A = SALK_119404, lox3B = SALK_147830, lox3D = SALK_062064), LOX4 (At1g72520; lox4A = SALK_071732, lox4B = SALK_017873) and LOX6 (At1g67560) (lox6A = SALK_138907, lox6B = SALK_083650) were used (Caldelari et al., 2010). The lox2-1 (At3g45140) mutant was described in Glauser et al. (2009). Double, triple and quadruple mutants were obtained by crossing (sterility was rescued with methyl jasmonate in plants homozygous for lox3 and lox4 alleles (Caldelari et al., 2010)). Plants were genotyped according to Caldelari et al. (2010) and Glauser et al. (2009). Plants were grown at 21°C, 10 h light d−1 (100 μmol s−1 m−2) photoperiod and 70% humidity, except for the assessment of lateral branching, when plants were grown in 16 h light: 8 h dark. Leaves from 5- to 6-wk-old plants were numbered from oldest to youngest. All experiments for jasmonate measurements were performed between 13:00 and 16:00 h. For wounding, 50% of leaf 8 was crushed with a metal forceps. Approximately 10 s was required to wound one leaf in this way. In accordance with previous experiments (Glauser et al., 2009) timing was initiated after the last second of mechanical injury infliction. Leaves were snap-frozen in liquid N2 and stored at −80°C before extractions.
JA and JA-Ile quantification
Frozen tissues (500 mg, from 6-wk-old plants) were ground in a ball mill extractor with 50 μl of 18O2JA (Mueller et al., 2006) and JA-13C6Ile (Kramell et al., 1997) internal standards (40 ng ml−1) before extraction with isopropanol. Highly apolar compounds such as chlorophyll were removed on a C18 solid-phase extraction cartridge using methanol : water (85 : 15, v/v) for elution. The eluate was concentrated and dissolved in 100 μl methanol : water (85 : 15, v/v). Jasmonates were detected by tandem mass spectrometry in the multiple reaction monitoring (MRM) mode on a Quattro micro™ API mass spectrometer (Waters, Milford, MA, USA) with an electrospray ionization (ESI) interface coupled with Agilent LC system (Hewlett Packard, Ramsey, NJ, USA). The ESI conditions were as follows: capillary voltage 3300 V; cone voltage 24 V; extractor 3 V; RF Lens 0 V; source temperature 120°C; desolvation temperature 350°C; cone gas flow 900 l h−1 and desolvation gas flow 27 l h−1. Detection was performed in the negative ion mode. The MRM transitions are: JA, 209.1 > 58.7; 18O2JA, 213.1 > 62.8; JA-Ile, 322.2 > 130.0; JA-13C6Ile, 328.2 > 136.0 (parent > daughter). The separation was carried out on a Kinetex 2.6-μm C18 100-Å column (100 × 3.0 mm) (Phenomenex, Torrance, CA, USA). A gradient was performed at a flow rate of 0.4 ml min−1 with the following solvent system: A = 0.1% formic acid/water, B = 0.1% formic acid/acetonitrile; 5% B for 3 min, 5–75% B for 11 min, 75–95% B for 2 min, 95% B for 2 min and 95–5% B for 2 min. The limit of quantification (LOQ = 3× limit of detection) was 4.2 pmol g−1 fresh weight (FW) for JA and 1.5 pmol g−1 FW for JA-Ile. Data below LOQ were not considered as informative.
Total RNA from 5-wk-old plants (purified according to Onate-Sanchez & Vicente-Carbajosa, 2008) was retro-transcribed to cDNA with the M-MLV reverse transcriptase RNAse H(-) (Promega, Madison, WI, USA) according to the manufacturer's instructions. Five microliters of 20× diluted cDNA was used for qPCR reactions. The PCR reaction mixture contained 0.2 mM deoxynucleotide triphosphates, 2.5 mM MgCl2, 0.5× SYBR Green I (Invitrogen, Paisley, UK), 30 nM 6-carboxy-X-rhodamine, 0.5 units of GoTaq polymerase (Promega), and 0.25 μM of each primer in a total volume of 20 μl. JASMONATE-ZIM DOMAIN 10 (JAZ10; At5g13220) data were standardized to the UBC21 ubiquitin conjugase reference gene (At5g25760). The primers and PCR program used were as in Gfeller et al. (2011).
LOX6Pro:GUS plant transformation
The LOX6 promoter (At1g67560, amplified using 5′-CGG GGT ACC GGT TGT TGA AAT TTC TGA TGC T-3′ and 5′-TTC CCC CCC GGG TTT TGT TTG GAG TTT GGC AGT-3′ primers) was cloned using XmaI and KpnI (New England Biolabs, Ipswich, MA, USA) into the pUC57-L4-KpnI/XmaI-R1 plasmid producing a pEN-L4-LOX6Pro-R1 as pENTRY clone.
The pUC57-L4-KpnI/XmaI-R1 plasmid was generated by Joop Vermeer (University of Lausanne) by introducing L4-KpnI/XmaI-R1 att recombination and restriction sites into pUC57 (Invitrogen). pEN-L1-GUS-L2 entry clones were obtained with Gateway technology (Invitrogen) according to the manufacturer's instructions with β-glucuronidase (GUS) cDNA (amplified from pEN-L2-S*-L3 (Ghent University, Ghent, Belgium) using 5′-GGG GAC AAG TTT GTA CAA AAA AGC AGG CTC AAT GTT ACG TCC TGT AGA AAC C-3′ and 5′-GGG GAC CAC TTT GTA CAA GAA AGC TGG GTA TCA TTG TTT GCC TCC CTG CTG C-3′ primers) and pDONR™/ZEO (Invitrogen). Adaptors in primer sequences are underlined. The final LOX6Pro:GUS constructs were generated using a double Gateway reaction into pEDO097pFR7m24GW. pEDO097pFR7m24GW was generated by inserting the fluorescence-accumulating seed technology (FAST) cassette (Shimada et al., 2010) into pH7m24GW (Invitrogen) by Ester M. N. Dohmann (University of Lausanne). Wild-type (WT) Columbia (Col-0) plants were transformed using Agrobacterium tumefaciens cells as described previously (Berberich et al., 2008). Transformed seeds expressing red fluorescence protein (RFP) were selected by fluorescence microscopy. The T1 generation was used for experiments. A total of 45 rosettes from 5-wk-old plants were GUS-stained for examination. Of these, 39 showed the staining pattern we report, and six plants showed no staining.
GUS staining and light microscopy
Staining and embedding were adapted from Scheres et al. (1994). Five-week-old rosettes were prefixed in acetone : water (90 : 10, v/v) for 1 h on ice, washed twice in 50 mM sodium phosphate buffer (pH 7.2) and vacuum-infiltrated for 10 min, then left overnight at 37°C in the dark in staining solution (10 mM Na2EDTA; 50 mM sodium phosphate buffer, pH 7.2; 1 mM K4Fe(CN)6; 1 mM K3Fe(CN)6; 0.1% (v/v) Triton X-100; 0.5 mg ml−1 X-Gluc). Rosettes were then either transferred into ethanol : water (70 : 30, v/v) for photography with Keyence Digital microscope VHX-500F (Mechelen, Belgium) or fixed (glutaraldehyde/formaldehyde/50 mM sodium phosphate (pH 7.2) buffer: 2 : 5 : 43, v/v/v) for resin embedding. For embedding, stained rosettes were dehydrated in an ethanol gradient (10%, 30%, 50%, 70%, 90% and twice absolute) for 30 min each. Embedding in Technovit 7100 resin (Haslab GmbH, Ostermundigen, Switzerland) was carried out according to the manufacturer's instructions. Sections (4 μm) were made with a Leica RM2255 microtome using disposable Leica TC-65 blades. Sections were photographed with a Leica DM5500 microscope.
Insect feeding trials
Separate plants were grown in individual pots. Eleven pots of 4-wk-old plants of each genotype were isolated in plexiglass boxes. Newly hatched Spodoptera littoralis (Boisduval; Noctuidae; Lepidoptera) caterpillars (four larvae per pot) were placed on each plant. The weight of larvae from one box per number of recovered larvae was considered as one replicate. For each box, larvae were harvested after 8, 10 or 14 d. For insect bioassays on wounded plants, three equidistant (120°) fully expanded leaves per plant were wounded over 50% of the leaf apex with forceps 30 min before adding the freshly hatched larvae. For assays with WT, lox2-1 lox3B lox4A triple and lox2-1 lox3B lox4A lox6A quadruple mutant plants, the plants were kept after the insect feeding trial and transferred to a 16-h light photoperiod for lateral branch development.
Analysis of early JA and JA-Ile accumulation in 13-LOX single, double, triple and quadruple mutants
We used an experimental design in which a single leaf, leaf 8, was wounded. Jasmonate and transcript levels in this leaf as well as in distal leaf 13 on the same plant were then measured at precise time-points. The two leaves in question were chosen because they are known to share a connected vasculature: they are part of the same parastichy (Dengler, 2006). Previous research has shown that when leaf 8 is wounded, JA accumulates in this leaf and also in leaf 13, but JA accumulates to a lesser extent in leaf 9 which is not directly coupled through the vasculature to leaf 8 (Glauser et al., 2009). In the present study we used high-performance liquid chromatography–tandem mass spectrometry in the multiple reaction monitoring (MRM) mode and cognate isotope-labeled internal standards to work in the low pmol range for both JA and JA-Ile quantification. The basal concentrations of JA observed in unwounded leaves of the WT (Col-0) (9 ± 3 pmol g−1 FW in leaf 8 and 6 ± 1 pmol g−1 FW in leaf 13) were similar to those reported previously (Glauser et al., 2009). Forty seconds after completion of wounding (a process that took < 10 s) we detected statistically robust increases in JA concentrations in the wounded WT leaf 8 (Fig. 1). Significant JA accumulation began at c. 90 s in leaf 13. The calculated speed of the signal(s) that triggers JA accumulation in the distal leaf 13 was c. 7.5 cm min−1 (Supporting Information Fig. S1), slightly faster than in a previous report, where a different experimental design was used (Glauser et al., 2009). Likewise, JA-Ile concentrations increased 90 s after wounding leaf 8 and at 190 s in distal leaf 13. Therefore, it takes c. 50 s to convert JA to JA-Ile in the wounded leaf, while it appears to take longer than 100 s in a distal connected leaf.
Jasmonate levels were analyzed systematically in T-DNA insertion mutants as well as in the lox2-1 null mutant (Glauser et al., 2009). The results showed that two to three different alleles of each of LOXs 3, 4 and 6 behaved similarly in terms of their effects on jasmonate synthesis (Figs 1, S2) and that concentrations of wound-induced JA in lox2-1 (Fig. 1) were similar to those reported previously for this mutant and for plants in which LOX2 expression was down-regulated by co-suppression (Glauser et al., 2009; Seltmann et al., 2010). These detailed analyses revealed that lox6A showed a delay of c. 50 s in wound-stimulated JA accumulation in leaf 8; lox2-1 showed less delay in JA accumulation. By contrast, JA-Ile accumulation was not impeded in lox2-1, but was slower in lox6A (Fig. 1). A similar observation was made for the connected leaf 13 where only the lox6A mutant displayed a delay in distal JA and JA-Ile accumulation. However, at 190 s post-wounding, the JA-Ile/JA ratio was similar in WT and lox6A.
Using the lox2-1 mutant and the lox3B, lox4A and lox6A alleles we constructed double, triple and quadruple mutants. None of these mutants had a markedly different phenotype to the WT except that, wherever the homozygous lox3B and lox4A alleles occurred, plants were male sterile. These various genotypes were then used to investigate LOX roles in jasmonate synthesis. Consistent with the literature, we found that the redundancy of LOX3 and LOX4 described for male fertility (Caldelari et al., 2010) was also apparent in early wound events. At the 190-s time-point the combined lox3B lox4A mutations reduced JA concentrations by c. 20% in the wounded leaf 8 but not in the distal leaf 13 (Fig. 2). As opposed to the modest effect of the lox3B lox4A double mutant on wound-inducible jasmonate concentrations, the lox2-1 lox6A double mutant had the strongest effect of all double mutant combinations on both JA and JA-Ile accumulation within the wounded leaf. As seen in Fig. 2, all double and triple mutants down-regulated for either LOX6 or LOX2 displayed a delay in JA accumulation in the wounded leaf 8. Furthermore, double and triple mutants harboring the lox6A allele displayed a large delay in distal JA and JA-Ile accumulation up to the first 190 s following the wounding. Conversely, the triple mutant lox2-1 lox3B lox4A, where LOX6 is the only functional 13-LOX, accumulated as much JA and JA-Ile in leaf 13 190 s after the wounding leaf 8 as did the WT.
Examination of inducible jasmonate concentrations was then extended to the four 13-LOX triple mutants. The analysis of these mutants further confirmed that the reduction of LOX6 expression greatly reduced wound-inducible JA and JA-Ile concentrations and was essential for early distal jasmonate synthesis. In all three triple mutant combinations that contained the lox6A allele, reducing LOX6 expression resulted in reduced JA and JA-Ile concentrations both in the wounded leaf 8 and in unwounded leaf 13. Finally, neither JA nor JA-Ile increases were verifiable or quantifiable in the lox quadruple mutant at any time-point as values were well below the limit of quantification (LOQ). In summary, for the time course we used, LOX6 was the major quantitative contributor to JA and JA-Ile accumulation in the wounded leaf as well as in the distal unwounded leaf. While LOX2 was necessary for WT JA production levels in the wounded leaf 8 it made no detectable contribution to JA induction in leaf 13 during the time course we used.
Rapid jasmonate accumulation in wounded and unwounded sectors of leaf 8
Our previous results (Figs 1, 2) suggested the hypothesis that LOX6 might contribute mostly to early wound-inducible jasmonate synthesis at sites distal to the damaged area. To investigate this we first looked more closely for evidence of LOX6 activity in the two tissue environments generated in wounded leaves. We separated the crushed part from the healthy part of wounded leaf 8 and repeated the quantification analysis with the WT and with lox6A. By doing this we found that JA accumulated faster in the crushed part of the wounded leaf than in the uncrushed part in the WT (Fig. 3). In these plants only c. 25 s was required for the first observable increase in JA concentration in the crushed sector. A further 15 s was required for JA build-up to begin in the uncrushed part of the wounded WT leaf. During these analyses we repeatedly noticed slightly higher basal JA concentrations in the tips of unwounded WT leaves than in the basal half (Fig. 3). Nevertheless, 90 s after wounding, JA concentrations were similar in the two leaf regions. There was a minor impact on JA accumulation in lox6A compared with the WT in the crushed sector of the wounded leaf but at 90 s post-wounding the ratios of JA and JA-Ile remained similar for the two genotypes (JA/JA-Ile = 14.85 ± 7 for WT and 15.7 ± 1.6 for lox6A). By contrast, the uncrushed area of leaf 8 in the lox6A mutant was unable to accumulate JA or JA-Ile up to the 90-s time-point. This result supports a role for LOX6 in the rapid accumulation of JA-Ile in regions distal to a wound.
Local and distal JAZ10 expression analysis
The JAZ10 gene is a well-described early marker for jasmonate signalling in wounded leaves (Yan et al., 2007; Acosta & Farmer, 2010). We investigated whether mutations affecting JA and JA-Ile accumulation could impact JAZ10 transcript levels. For this, the design of the previous experiment (Fig. 3) was retained, separating the crushed part of the leaf from the healthy tissue proximal to the wound. In agreement with the results on JA and JA-Ile concentrations we found that the absence of functional LOX6 had (1) no measurable effect on JAZ10 expression within the wound itself, (2) only a weak but significant effect on JAZ10 transcript levels in the unwounded portion of leaf 8, and (3) a strong effect on JAZ10 expression in distal leaf 13 (Fig. 4). At the 60-min time-point (insets in Fig. 4) the levels of JAZ10 transcript were similar in lox6A and the WT in the uncrushed part of leaf 8. By contrast, we did not detect increased JAZ10 levels in leaf 13 of the lox6A mutant even 60 min after wounding leaf 8. To extend these observations we tested the effect of a second lox allele, lox6B, on wound-induced JAZ10 expression in the unwounded part of wounded leaf 8. As for the lox6A mutant, the lox6B mutant reduced JAZ10 expression relative to the same tissue zone in the WT (Fig. S3).
Defense assays with a lepidopteran herbivore
A first defense assay using Spodoptera littoralis larvae was then set up to test insect performance on the WT, lox6A, and the lox2-1 lox3B lox 4A lox 6A quadruple mutant lacking all four functional 13-LOX genes (Fig. 5a). In this assay, larvae gained weight rapidly on the quadruple mutant and had consumed most of the plants by day 8 (inset in Fig. 5a). By contrast, we did not observe altered weight gain on lox6A relative to the WT either at this time or 6 d later. As LOX6 was necessary for early-phase jasmonate synthesis distal to a wound, we established a second bioassay in which both WT and lox6A plants were wounded before releasing neonate larvae. Unwounded plants served as controls. After 14 d, caterpillars that had fed on the unwounded WT showed an increased weight gain relative to the pre-wounded WT (P = 0.002, t-test; Fig. 5b), but we did not observe such evident differences in caterpillar weight on unwounded lox6A and pre-wounded lox6A.
A further bioassay was then set up. This next assay used the unwounded WT and quadruple mutant controls, but this time, instead of using lox6A, we used its mirror mutant lox2-1 lox3B lox4A. At the conclusion of the experiment (after caterpillars had been allowed to feed for 10 d) the impact of the different plant genotypes on caterpillar weight was assessed. We found that insects grew to similar weights on lox2-1 lox3B lox4A and lox2-1 lox3B lox4A lox6A and weighed less after feeding on the WT (Fig. 5c). These initial results, therefore, did not reveal any additional roles of LOX6 in defense until we noted that feeding patterns were strikingly different on the triple and quadruple mutants. In the case of the quadruple mutant the insects migrated to and fed on the center of the rosette, consuming apical tissues including the youngest leaves. Instead, on the triple mutant, the insects fed more on expanded leaves and avoided the young leaves in the shoot apex region. The final outcome of this experiment on the morphology of the plants was then assessed by removing the insects and allowing all three plant genotypes to proceed to flowering. When this was done the quadruple mutant showed more leaves and more lateral branches than both the WT and the lox2-1 lox3B lox4A triple mutant (respectively, P = 7.2E-7 and P = 1.0E-5, t-test; Fig. 5d). The undamaged WT and triple and quadruple mutants all produced single apical infloresence shoots (Fig. S4). The observations on the effects of insect feeding on post-damage branching correlated with the fact the larvae did not feed heavily on the expanded leaves of the quadruple mutant and instead focussed on younger leaves (Fig. 5e–j). That is, the insects preferentially consumed material from the apical region of the quadruple mutant whereas they fed in the opposite manner on the WT and on the triple mutant, preferring the older leaves (Fig. 5k–m).
Tissue localization of LOX6 promoter expression
By constructing and expressing a LOX6pro:GUS fusion gene we observed that LOX6 promoter activity was highest in both midrib and minor veins in young leaves (leaves 10 to > 20) close to the shoot apex (Fig. 6a), and consistent with results from herbivory assays. Interestingly, staining resulting from GUS activity was not observed in the midrib of older leaves (leaves 6–9), and was almost completely lost for the oldest leaves (leaves 5–1) where staining was observed mainly at the peripheral area of the lamina. In order to identify the cell types in which the LOX6 promoter was most active, we examined sections of a young unexpanded leaf near the apex. The first section shown (Fig. 6b) was cut transversally to the midrib. Here, the highest LOX6 promoter activity was observed in the xylem (Fig. 6b). The second section (Fig. 6c) was cut longitudinally along the midrib. This section shows a stained metaxylem cell with annular secondary wall thickenings and, next to it, an unstained protoxylem vessel cell with spiral secondary cell wall thickenings. This configuration of GUS-stained metaxylem cells adposed against GUS-negative protoxylem cells was highly conserved. GUS staining was not restricted exclusively to xylem cells and we also observed staining in parenchyma cells subtending trichomes. These subtrichomal mounds of LOX6 expression (Fig. S5) were close to stained xylem cells.
The 13(S)-hydroperoxidation of triunsaturated fatty acids is a crucial (Staswick & Tiryaki, 2004; Fonseca et al., 2009) and potentially regulatory step in jasmonate synthesis, although it is not known whether or not the fatty acids that are oxygenated by jasmonate-producing 13-LOXs are free or esterified in vivo. Increases in concentrations of free α-linolenic acid have been observed in wounded tomato (Solanum lycopersicum) leaves (Conconi et al., 1996) and this may require the action of acyl hydrolases. Consistent with this, such enzymes are known to play a role in the developmentally programmed synthesis of the jasmonates necessary for correct flower maturation in A. thaliana (Ishiguro et al., 2001). Also, concerning wounding, an acyl hydrolase that participates in damage-induced jasmonate synthesis has been reported from the leaves of a wild tobacco (Nicotiana attenuata) (Kallenbach et al., 2010). However, this contrasts to the situation for damage-induced jasmonate synthesis in A. thaliana leaves where, to date, such enzymes have not been identified (Ellinger et al., 2010). Finally, arabidopsides, secondary metabolites that contain OPDA and dinorOPDA, are made directly on galactolipids without prior de-esterification of the substrate fatty acids (Nilsson et al., 2012). Whether or not their fatty acid substrate is bound or free, 13-LOXs catalyze the first stereoselective step in jasmonate synthesis.
Several features that we had observed previously in leaves 8 and 13 in the wounded WT (Glauser et al., 2009) were reproduced in the present study. These included exponential increases in JA accumulation and an estimated wound signal velocity from leaf 8 to leaf 13 of c. 7.5 cm min−1. In the present experiments, the first signs of wound-induced JA accumulation in the damaged sector of the WT leaves were observed at 25 s after wounding and this lagged by a further 15 s in the unwounded half of the wounded leaf. JA-Ile accumulated to lower concentrations than did JA, but by 100 s increases in both JA and JA-Ile concentrations were detectable in both damaged and undamaged sectors of the wounded WT leaf. By 190 s after wounding, we observed a concentration of 150 pmol g−1 FW JA and 10 pmol g−1 FW JA-Ile in the wounded leaf. These concentrations are low compared with the maximum concentrations reported previously for later time-points. For example, at 90 min after wounding the concentrations of JA can reach 12 nmol g−1 FW (Reymond et al., 2004) and those of JA-Ile can reach 1.8 nmol g−1 FW (Koo et al., 2009). Nevertheless, by employing MRM we were able to work quantitatively at low jasmonate concentrations.
LOX6 activity is responsible for most of the rapidly synthesized JA and JA-Ile distal to a wound
The contribution of each LOX to JA and JA-Ile synthesis depends on the time of leaf harvest after wounding, but LOX6 stands out as being the 13-LOX having the highest impact on the earliest phase of JA and JA-Ile accumulation. Behind LOX6 and LOX2 in relative importance in rapid jasmonate production at this time were LOX3 and LOX4, and the lox3B lox4A double mutant reduced JA production in the wounded leaf. Analysis of each of the triple mutants carrying these two alleles did not allow us to distinguish the relative contributions of each of LOX3 and LOX4 to JA synthesis but further served to confirm the importance of LOX6 and LOX2 in early jasmonate production. The triple mutants are potentially useful as they isolate each individual LOX gene to function alone. For example, an observation of interest came from the lox3B lox4A lox6A triple mutant that retains the WT LOX2 gene. As seen in Fig. 2, this triple mutant produced JA but not JA-Ile in leaf 8 in response to wounding. It is known that LOX2 ablation in the lox2-1 mutant did not affect JAZ10 expression and, when compared with the WT, did not significantly alter concentrations of JA-Ile measured in the 5–90-min interval after wounding (Glauser et al., 2009). This and the current study suggest that signaling through JA-Ile perception was unnecessary for the regulatory events that led to the rapid LOX6-promoted JA accumulation we observed in response to a wound.
LOX6 has roles in the defense of young leaves
To date only one 13-LOX gene, LOX2, has been shown to have a role in defense in A. thaliana. The lox2-1 mutation caused a > 99% reduction in the concentrations of OPDA and dinorOPDA-containing arabidopsides in wounded leaves (Glauser et al., 2009). Similar very strong reductions in the concentrations of these putative defense compounds were also observed in plants in which LOX2 expression was suppressed through RNA interference (Seltmann et al., 2010). Consistent with a role in defense for arabidopsides, the lox2-1 mutation facilitated weight gain in the generalist herbivore Spodoptera littoralis (Glauser et al., 2009). Interestingly, and unlike the effect on the growth of S. littoralis, the lox2-1 mutation did not affect wound-inducible JAZ10 expression (Glauser et al., 2009). By contrast, the lox6A mutant reduced both the synthesis of jasmonates and JAZ10 expression upon wounding, and this effect was distance-dependent. The strongest reduction in JAZ10 transcript induction in lox6A relative to the WT was seen in the unwounded leaf 13. This observation became important when we conducted a standard herbivory assay by monitoring S. littoralis weight gain on the WT and on the lox6A mutant. We found no difference in larval development on the two plant genotypes in this assay, whereas the quadruple mutant was extremely sensitive to S. littoralis. However, the fact that we had observed an increasingly dominant role of LOX6 on jasmonate production distal to a wound led us to develop a different bioassay protocol in which three equidistant leaves on rosettes of each genotype were wounded before allowing neonate insects to feed. When this was done, an increased weight gain in insects was seen on the unwounded WT relative to that on the pre-wounded WT. By contrast, we did not see increased weight gain for S. littoralis that had fed on unwounded lox6A relative to pre-wounded lox6A. The severe mechanical wounding we used strongly activates jasmonate synthesis (Figs 1, 2). On the other hand, it is known that lepidopteran larvae do not strongly activate JA synthesis in the first 9 h of feeding (Reymond et al., 2004). The results support a defense role for LOX6 in responding to long-distance signals through the synthesis of jasmonates that trigger defense gene expression.
In order to further investigate the biological roles of LOX6 we used the lox6 mirror mutant lox2-1 lox3B lox4A in which three 13-LOXs (LOX2, LOX3 and LOX4) were genetically ablated, isolating the LOX6 gene to function alone. In contrast to the quadruple mutant in which young leaves were attacked aggressively by the herbivores, these small leaves near the apex of the lox2-1 lox3B lox4A triple mutant were left undamaged. The assays revealed that LOX6 has a function in defending the youngest leaves at the growing center of the plant, but our observations extended beyond this. When left to produce floral shoots, the lox2-1 lox3B lox4A triple mutant and the quadruple mutant showed very different phenotypes, with the latter displaying strong lateral branching. Removal of apical meristemic tissue by herbivores or by cutting is well known to enhance lateral meristem activity (Trumble et al., 1993) and this is also observed after cutting primary infloresence shoots (Stirnberg et al., 2002). This was the case in the S. littoralis-damaged quadruple mutants where new lateral inflorescence branches compensated for the loss of the shoot apex. Defending the tissues surrounding the shoot apical meristem is likely to be a key biological role of 13-LOXs and this was demonstrated in our experiments for LOX6. Considering the results of the first bioassay with WT and lox6A we do not rule out roles in defense of the shoot apex for the other 13-LOXs.
LOX6 as a potential regulator of early jasmonate production in and near xylem cells
In a few cases the cell types in which promoters of jasmonate synthesis genes have been studied and some of their gene products have also been localized to leaf cells or tissues. Previous studies have localized allene oxide cyclase (AOC) protein to vascular parenchyma and bundle sheath cells in tomato, whereas LOX and allene oxide synthase (AOS) were expressed strongly in many leaf cell types in this plant (Hause et al., 2000; Stenzel et al., 2003a). Upon wounding of tomato leaves, jasmonates accumulate strongly in the epidermis, in the mesophyll, and in parenchyma cells associated with the xylem and phloem (Mielke et al., 2011). In A. thaliana, unwounded leaves contain LOX, AOS and AOC proteins (Stenzel et al., 2003b). The AOS promoter was active in vascular bundles in the leaves of A. thaliana (Kubigsteltig et al., 1999) and the promoters of both LOX3 and LOX4 were expressed in or near the vascular system of A. thaliana cotyledons (Vellosillo et al., 2007). Concerning jasmonates themselves, the concentrations of JA and JA-Ile that we worked with were too low to be localized directly with current technology. In A. thaliana the bulk of the JA detected 90 min after leaf wounding was found in the lamina with significantly less JA in the region of the midvein (Glauser et al., 2008). Returning to LOX6, the promoter is reported to have low activity in A. thaliana seedlings (Vellosillo et al., 2007). However, our observations of LOX6 promoter activity in rosettes shed new light on spatial aspects of early wound-induced jasmonate synthesis. In the present study we found that LOX6 promoter activity was high in and near xylem cells in young tissues and this might explain why expanded leaves (in which many xylem cells will have fully differentiated) have lower levels of promoter activity. It is probable that much of the jasmonate that is produced in the first 3.5 min after wounding is made in living xylem cells or in their progenitors, before cell differentiation into tracheary elements. Tracheary element differentiation is itself frequently wound-inducible (Fukuda, 1996) and, interestingly, the JAZ10 promoter is known to be active in xylem cells in A. thaliana (Sehr et al., 2010). We frequently observed GUS-positive metaxylem-like cells adjacent to fully differentiated GUS-negative protoxylem (Fig. 6c). This appears to be a major niche of LOX6 promoter activity with a secondary niche being subtrichomal mounds of cells between the xylem and trichomes (Fig. S5). We cannot at this point comment on the functional significance of this potentially novel expression domain.
It is likely that there are many layers of regulation of jasmonate biosynthesis (Schaller & Stintzi, 2009). Several plant and animal lipoxygenases related to LOX6 have been shown to be activated by Ca2+ (Tatulian et al., 1998; Oldham et al., 2005). LOX6 itself might therefore be regulated by divalent metal ions (e.g. Ca2+ or Mg2+). Either LOX6 first builds OPDA pools that are converted to JA and JA-Ile upon wounding (Koo et al., 2009), or the enzyme catalyzes de novo wound-stimulated jasmonate from fatty acids (Glauser et al., 2009). In either case, we show that the activity of this lipoxygenase is necessary for decoding long-distance wound signals in leaves into early, defense-associated jasmonate production. In summary, our data show in new detail where and when the first jasmonates are made in wounded A. thaliana leaves. All four 13-LOXs contribute to wound-stimulated jasmonate production in A. thaliana leaves and, at the cellular level, metaxylem appears to be a site of early jasmonate production after wounding.
We thank A. Chételat (University of Lausanne) for technical help, L. Dubugnon (University of Lausanne) for help in producing several mutants, F. Schweizer (University of Lausanne) for insect larvae, E. M. N. Dohmann, J. Vermeer, D. Gasperini, I. Acosta, K. Neiminen, M. Yamazaki, D. Roppolo, S. Stolz, and S. Mousavi (University of Lausanne), G. Glauser (University of Neuchatel), and P. Eugster and G. Marti (University of Geneva), for plasmids and/or technical advice. I. Acosta, D. Gasperini and S. Wege (University of Lausanne) provided helpful comments on the manuscript. This work was supported by Swiss National Science Foundation Grants 3100A0-122441 and 31003A-138235 (to E.E.F.) and 205320-135190 (to J.L.W. and E.E.F.).