•Jasmonic acid (JA) is a plant signalling compound that has been implicated in the regulation of mutualistic symbioses. In order to understand the spatial distribution of JA biosynthetic capacity in nodules of two actinorhizal species, Casaurina glauca and Datisca glomerata, and one legume, Medicago truncatula, we determined the localization of allene oxide cyclase (AOC) which catalyses a committed step in JA biosynthesis. In all nodule types analysed, AOC was detected exclusively in uninfected cells.
•The levels of JA were compared in the roots and nodules of the three plant species. The nodules and noninoculated roots of the two actinorhizal species, and the root systems of M. truncatula, noninoculated or nodulated with wild-type Sinorhizobium meliloti or with mutants unable to fix nitrogen, did not show significant differences in JA levels. However, JA levels in all plant organs examined increased significantly on mechanical disturbance.
•To study whether JA played a regulatory role in the nodules of M. truncatula, composite plants containing roots expressing an MtAOC1-sense or MtAOC1-RNAi construct were inoculated with S. meliloti. Neither an increase nor reduction in AOC levels resulted in altered nodule formation.
•These data suggest that jasmonates are not involved in the development and function of root nodules.
Jasmonic acid (JA) and its derivatives, such as its methyl ester (MeJA) and amino acid conjugates, are signalling compounds that play a role in vegetative and propagative plant development, the response to wounding and plant defence (Liechti & Farmer, 2002; Wasternack, 2007; Browse, 2009; Koo & Howe, 2009). JA biosynthesis represents one branch of the oxylipin pathway, which starts with the oxygenation of linoleic or α-linolenic acid by lipoxygenases (LOXs), leading to the formation of (9S)-hydroperoxy linoleic or α-linolenic acid (9-HPOD/9-HPOT) or (13S)-hydroperoxy linoleic or α-linolenic acid (13-HPOD/13-HPOT) (Andreou & Feussner 2009; Supporting Information Fig. S1). The LOX reaction products are substrates of at least six different enzyme families involved in pathways leading to the production of signalling compounds, such as jasmonates, leaf aldehydes or divinyl ethers acting as antimicrobial and antifungal compounds, and a plant-specific blend of volatiles (Mosblech et al., 2009). Only the α-linolenic acid-derived product 13-HPOT can be converted to jasmonates (Browse, 2009; Koo & Howe, 2009; Acosta & Farmer, 2010). During the first committed step of JA biosynthesis, allene oxide synthase (AOS) converts 13-HPOT to 12,13-epoxy α-linolenic acid, which is converted to cis-12-oxo-phytodienoic acid (OPDA) by an allene oxide cyclase (AOC). This first part of the JA biosynthetic pathway occurs in plastids, whereas the conversion of OPDA to JA occurs in peroxisomes (Browse, 2009). OPDA is transferred to the peroxisome, where it is converted to JA by the action of OPDA reductase 3 (OPR3) and three rounds of β-oxidation (Koo & Howe, 2009). Although the LOX reaction product, 13-HPOT, is the precursor of many different oxylipins, the products of AOS/AOC and OPR3 are committed to JA biosynthesis (Fig. S1).
JA may be involved in mycorrhiza-induced resistance by serving as a systemic defence signal to improve general fitness in arbuscular mycorrhizal symbioses (Pozo & Azcon-Aguilar, 2007). It also seems to be involved in the regulation of specific plant processes, such as isoflavonoid biosynthesis and the establishment of sink organs (Hause & Schaarschmidt, 2009). To date, not much is known about the role of JA during the course of symbiotic plant–bacteria interactions. The expression patterns of LOX genes have been analysed in some legume root nodule symbioses (Perlick et al., 1996; Porta et al., 1999). However, there are many members of the LOX gene family, and it is unclear whether LOX, the expression of which is induced in nodules, is actually involved in the biosynthesis of JA (Hause & Schaarschmidt, 2009). The expression patterns of genes encoding other enzymes of the oxylipin pathway have not been analysed in nodules. In addition, the exogenous application of JA/MeJA has been shown to affect the response of M. truncatula to rhizobial signal factors (Nod-factors; Miwa et al., 2006; Sun et al., 2006). Further-more, the application of MeJA to the shoots of the model legume Lotus japonicus affected nodulation (Nakagawa & Kawaguchi, 2006), and a hypernodulating mutant of soybean was shown to have increased JA content in leaves, leading to the hypothesis that JA might be involved in the mediation of the response to the shoot-derived signal in the autoregulation of nodulation (Seo et al., 2007; Kinkema & Gresshoff, 2008).
In order to analyse the role of JA/MeJA in nodule development, we set about comparing the JA biosynthetic capacity by determining the levels and cell-specific localization of AOC, the enzyme catalysing the second committed step in JA/MeJA biosynthesis. Furthermore, the levels of JA were analysed in the roots and nodules. To investigate whether alterations in the cellular JA biosynthetic capacity affect root nodule development, M. truncatula roots were transformed with an MtAOC1-sense or an MtAOC1-RNAi construct. The nodulation phenotype of the resulting composite plants containing roots with altered AOC levels was analysed via nodule number and morphology.
Materials and Methods
Plant and bacterial growth conditions
Casuarina glauca Sieber ex K. Sprengel and Datisca glomerata (C. Presl) Baillon were germinated on soil (Frühstorfer Erde Typ EP Nr. 340, Industrie Erdenwerk Archut, Lauterbach-Wallenrod, Germany) and transferred to a hydroponic system with one-quarter strength Hoagland’s medium (Hoagland & Arnon, 1938) after c. 2 months, as described previously (Pawlowski et al., 2003). Frankia strain Thr (Girgis & Schwencke, 1993) was used to infect C. glauca, whereas D. glomerata plants were infected with crushed nodules taken from soil-grown plants. Nodules were harvested 6–8 wk after infection. Medicago truncatula Gaertn. seeds were germinated on soil and, after c. 4 wk, plants were transferred to an aeroponic system with medium according to Lullien et al. (1987), and infected with Sinorhizobium meliloti 1021 grown in TY medium (Behringer, 1974), washed with and resuspended in double-distilled H2O. Nodules and roots were harvested 2–3 wk after infection.
Alternatively, M. truncatula seeds were scarified for 10 min in concentrated H2SO4, washed extensively in sterile double-distilled H2O and germinated on 1% water agar in Petri dishes, initially for 1 d at 24°C in the dark and then for 2 d at 16 h low light intensity and 8 h dark. Seedlings were transferred to pots containing expanded clay (Lecaton, 2–5 mm particle size; http://www.fiboexclay.de) wetted with one-quarter strength Hoagland’s medium. Plants were infected with wild-type S. meliloti 2011, a fixJ (GMI347; David et al., 1988) or a fixG (GMI394; Kahn et al., 1989) mutant strain, and grown under constant conditions (220 μmol m−2 s−1 for 16 h; 22°C). Plants were watered once a week with Hoagland’s solution and otherwise with double-distilled H2O. Root systems, including nodules, were harvested 7, 10 and 17 d after infection.
Root transformation of M. truncatula
For the knockdown of the expression of both MtAOCs, an inverted repeat construct was created in pRNAi (Limpens et al., 2004). The target region is located directly downstream of the plastid targeting signal, covering 379 bp of the MtAOC1-cDNA (AJ308489), and is homologous to that of MtAOC2 (AJ866733; Isayenkov et al., 2005). This region was PCR amplified using the following primers, RNAi_SwaI/BamHI 5′-ATACTAGTCCATGGTAGAGG-AAGTCCTGCATACC-3′ and RNAi_SpeI/NcoI 5′-ATG-GATCCATTTAAATAGCTGGCAAATCAGCAACAC-3′. Both PCR products were inserted into pRNAi according to Limpens et al. (2004). The resulting hairpin construct was cloned with KpnI-PacI restriction sites into the binary vector pRedRootII (kindly provided by R. Geurts, Wageningen, the Netherlands). The 35S::MtAOC1 cDNA expression cassette (Isayenkov et al., 2005) was also inserted in pRedRootII, allowing the selection of plants containing transgenic roots on kanamycin-containing medium. Empty pRedRootII served as a control. All binary vectors were introduced into Agrobacterium rhizogenes ArquaI (Quandt et al., 1993) by electrotransformation. Root transformation was performed as described previously (Boisson-Dernier et al., 2001; Mrosk et al., 2009). The cultivation of plants with transgenic roots and inoculation on pouch-paper were performed according to Journet et al. (2006). Cultivation and inoculation in expanded clay were performed as described for wild-type plants. Only nodules formed on transgenic roots selected by DsRED fluorescence using a fluorescence stereomicroscope (Leica MZ FLIII with DsRED1-Filter; http://www.leica.de) were harvested and analysed.
Isolation of RNA, cDNA synthesis and real-time PCR using Taq-Man probes and primers were performed as described previously (Isayenkov et al., 2005). All assays were performed on single transgenic root systems in three technical replicates each. ΔCt values were calculated by subtracting the Ct values of the MtAOC1 gene from the arithmetic mean of the Ct value of MtEF1-α. ΔΔCt values were calculated by subtracting the ΔCt values obtained for each sample from the ΔCt value of the plant material obtained from transformations with empty pRedRootII.
Protein isolation, protein gel electrophoresis and Western blot analysis
JA was extracted as described previously (Schmelz et al., 2004) with some modifications. Fifty milligrams of frozen plant material finely ground in liquid nitrogen were weighed and transferred to a 2 ml tube; 1 ml of diisopropylamine containing 100 ng of D6-JA (kindly provided by Otto Miersch, Halle/Saale, Germany) was added. The mixture was sonicated for 15 min and 1 ml of chloroform was added, followed by a second step of 15 min of sonification. For detection, JA was converted to its pentafluorobenzyl ester according to Mueller & Brodschelm (1994) by adding 17 mg of pentafluorobenzylbromide (Sigma) and incubating at 60°C for 1 h. After evaporation under a nitrogen stream, the residue was dissolved in 1 ml of diethyl ether and filtered through a filter paper. For complete recovery of the pentafluorobenzyl esters, the sample tube was washed with 1 ml of n-hexane which was also filtered and combined with the diethyl ether filtrate. The solution was evaporated under a stream of nitrogen.
The vapour phase extraction was carried out at 270°C for 5 min with argon as carrier gas and the substances absorbed by a SuperQ column (100 × 4.6 mm SDB-L Strata; pore size, 260 Å; Phenomenex, http://www.phenomenex.com). Subsequently, the substances were eluted from the SuperQ columns with 3 ml of n-hexane and 3 ml of ethyl acetate. The solution was concentrated using a rotating evaporator. The remainder was dissolved in 40 μl of dichloromethane and subjected to gas chromatography coupled with mass spectrometry.
The analysis was carried out using a ThermoFinnigan (http://www.thermo.com) Polaris Q mass-selective detector connected to a ThermoFinnigan Trace gas chromatograph equipped with a capillary Rtx-5MS column (15 m × 0.25 mm, 0.25 μm coating thickness; Restek, http://www.restek.com). Helium was used as carrier gas (1 ml min−1). A temperature gradient of 100°C for 1 min, 100–300°C at 8°C min−1 and 300°C for 5 min was used. The pentafluorobenzyl esters were detected by negative chemical ionization with ammonia as the ionization gas. For quantification, the ions 215 (D6-JA; Rf = 16.15, 16.51 min) and 209 (JA; Rf = 16.21, 16.56 min) were used.
Immunolocalization for light microscopy was performed as described previously by Stumpe et al. (2006). Briefly, roots and nodules were fixed using 3% (w/v) paraformaldehyde and 0.1% Tween-20 in phosphate-buffered saline (PBS) (0.14 M NaCl, 2.7 mM KCl, 6.5 mM Na2HPO4, 5 mM KH2PO4, pH 7.3) and dehydrated in a graded ethanol series. Optionally, nodules were then stained overnight at 4°C in 0.1% (w/v) toluidine blue dissolved in 95% ethanol. Embedding in Steedman’s wax or in PEG1500 was performed as described previously (Vitha et al., 1997; Isayenkov et al., 2005). Immunolabelling of sections, 4 or 10 μm thick, was carried out with rabbit anti-AOC antibody (diluted 1 : 500) followed by a goat anti-rabbit IgG antibody conjugated with AlexaFluor488 (Molecular Probes, http://www.invitrogen.com; diluted 1 : 250). Sections were then stained for 5 min in 0.01% (w/v) toluidine blue in PBS to suppress the autofluorescence of the cell walls. Optionally, sections were treated with 0.1 μg ml−1 4′,6-diamidino-2-phenylindole (DAPI). Sections were mounted using the ProLong® Gold Antifade Kit (Molecular Probes). Use of the respective pre-immune serum served as control and yielded no signal. Fluorescence was detected with an Olympus BX51 microscope (Olympus BioSystems GmbH, http://www.olympus.de) equipped with a SIS ColorView II digital camera and analySIS® FIVE image-analytical software (Soft Imaging System GmbH, http://www.soft-imaging.net), or a Zeiss AxioImager (Carl Zeiss AG, http://www.zeiss.de).
Immunogold electron microscopy
Immunolocalization on transmission electron microscopy was performed as described previously by Baluška et al. (2004). Sections were observed in a LEO EM 906E (Zeiss) or Hitachi-600 (Hitachi, http://www.hitachi-hitec.com) electron microscope at 75 kV.
Protein levels of AOC in roots and nodules of M. truncatula, C. glauca and D. glomerata
To assess protein levels per organ and to verify the applicability of the antibodies, total protein extracts from roots and nodules of M. truncatula, D. glomerata and C. glauca were used for Western blotting (Fig. 1). AOC protein of c. 26 kDa was detected in extracts from roots as well as nodules (Fig. 1b). No differences could be found between AOC protein levels in roots and nodules, except that AOC protein levels were lower in C. glauca than in the two other species.
Immunolocalization of AOC in nodules of the three symbiotic plants
Next, the distribution pattern of AOC was analysed in roots and nodules, assuming that spatial differences in protein distribution in both organs may be indicative of basal JA levels and JA biosynthetic capacity. AOC is a plastid-located enzyme (Stenzel et al., 2003) and has been shown to be restricted to the central cylinder in nonmycorrhizal roots of M. truncatula (Isayenkov et al., 2005). Immunolocalization of AOC in M. truncatula nodule sections showed that the protein was localized in the plastids of the uninfected cells of the nitrogen fixation zone, the proximal part of the prefixation zone and in plastids in the nodule cortex, including the vascular parenchyma (Fig. 2a–h). Hence, AOC was detected only in uninfected cell types. A more detailed study revealed that AOC occurred in different types of plastid, such as the amyloplasts of uninfected cells of the inner tissue and of the nodule cortex (Fig. 2c–f), and in the lens-shaped plastids that are found in uninfected cells of the younger part of the pre-fixation zone (Fig. 2g) and the vascular parenchyma (Fig. 2h). Fig. 2(h) shows the difference between the lens-shaped plastids of the stellar parenchyma and the round amyloplasts in cells of the nodule cortical parenchyma. To analyse whether the absence of AOC protein in infected cells is a result of nitrogen fixation, immunolocalization of AOC was performed on nodules induced by three different S. meliloti strains: wild-type S. meliloti 2011, a fixG::Tn5 mutant not affected in symbiotic nitrogen fixation and a fixJ::Tn5 strain inducing inefficient nodules. The AOC distribution pattern, however, was the same for both nitrogen-fixing and non-nitrogen-fixing nodules (Fig. S2). In all nodules analysed, AOC was detected in the plastids of cells of the nodule cortex and of uninfected cells of the inner tissue. The AOC amounts seemed to be higher in the non-nitrogen-fixing nodules than in the nitrogen-fixing ones, but this could be a result of the larger number of amyloplasts in noninfected cells of non-nitrogen-fixing than nitrogen-fixing nodules.
In the two actinorhizal species, green fluorescence was detected in the infected cells in which Frankia was beginning to form vesicles (Fig. 2g,h for C. glauca; data not shown for D. glomerata), but immunogold electron microscopy showed that this was caused by a cross-reaction of the anti-AOC antibody with a protein in Frankia hyphae (data not shown). In addition, the amyloplasts of the uninfected cells in the outer cortex in nodules of C. glauca contained AOC protein (Fig. 2g,h). Moreover, in D. glomerata nodules, a signal was observed in amyloplasts of uninfected cortical cells (Fig. 2i,j) and of young infected cells (Fig. 2k,l). This difference is probably because, in D. glomerata, amyloplasts disappear from cortical cells during the course of the infection process, whereas, in C. glauca, amyloplasts are only found in uninfected cortical cell types (M. Schubert et al., unpublished).
JA levels in roots and nodules of the three symbiotic systems
Our immunolocalization experiments revealed significant levels of AOC in uninfected cells adjacent to infected cells in all three types of nodule examined, suggesting JA biosynthetic activity in nodules. Therefore, JA levels were examined in roots and nodules of the three symbiotic systems.
The harvesting of nodules takes considerably longer than the harvesting of roots, and involves wounding and mechanical disturbance of the root system. It has been shown that wounding induces JA production (Koo & Howe, 2009) and mechanical disturbance causes an increase in JA levels in the roots and shoots of M. truncatula (Tretner et al., 2008). Given the touch-induced transcript accumulation of AOS in Arabidopsis shoots and wheat leaves (Mauch et al., 1997; Lee et al., 2005), and of AOC in M. truncatula shoots (Tretner et al., 2008), the induction of JA biosynthesis by mechanical disturbance is likely to be a general phenomenon. Therefore, controls had to be introduced to ensure that any differences observed between JA levels in roots and nodules were not caused by differences in harvesting procedures. Levels of JA were compared in roots and nodules harvested from nonpretreated plants or root systems vs material harvested from plants or root systems that had been mechanically disturbed (stroked and shaken, called ‘shaken’ in this article) for 30 min. In the case of C. glauca and D. glomerata, the root system was separated from the shoot before being subjected to mechanical disturbance. The results show that, in all three plant species, JA levels were increased by mechanical disturbance in roots and nodules (Fig. 3a). In M. truncatula, but not in the actinorhizal plants, JA biosynthesis in response to mechanical disturbance was consistently greater in roots than in nodules. Furthermore, no significant differences could be detected between AOC levels in control vs shaken organs (Fig. 1c). However, no significant differences were detected in JA levels between noninoculated roots and nodules in all three plants.
JA in nodulated vs noninoculated root systems of M. truncatula
As the mechanical treatment of roots might cause the biosynthesis of JA, we analysed JA levels in whole root systems of nodulated vs noninoculated plants. Medicago truncatula was selected for these experiments because of the availability of rhizobial mutants inducing non-nitrogen-fixing nodules in this plant. Nodules of roots infected with either the non-nitrogen-fixing fixJ mutant bacteria or wild-type rhizobia exhibited a similar pattern of distribution of wild AOC protein (Fig. S2). The root systems of plants nodulated with nitrogen-fixing wild-type S. meliloti 2011, or non-nitrogen-fixing fixJ::Tn5 mutant bacteria, were harvested at different time points after infection and their JA levels were determined (Fig. 3b). Noninoculated plants grown in nitrogen-free growth medium were used as controls. The data confirmed that there were no significant differences in JA levels between noninoculated vs efficiently nodulated root systems. Although a transient increase in JA levels after infection was observed in root systems inoculated with wild-type bacteria, the differences between JA levels among all root systems (noninoculated, nitrogen-fixing nodules, fixJ::Tn5-induced nodules) at different time points were not statistically significant (Fig. 3b). Nevertheless, as whole root systems were analysed, these results do not exclude a possible difference in JA levels in specific nodule cell types.
Nodulation of M. truncatula transgenic roots with altered expression levels of MtAOC1
The observed subcellular localization of AOC encouraged us to question whether cell-specific biosynthesis of JA played a role in nodule development. Therefore, analogous to the studies on the role of JA in arbuscular mycorrhization (Isayenkov et al., 2005), A. rhizogenes-mediated transformation was applied to create composite M. truncatula plants with transgenic roots expressing an MtAOC1-sense construct, an MtAOC1-RNAi construct or an empty pRedRootII vector, all under the control of the cauliflower mosaic virus (CaMV) 35S promoter. Transgenic roots also expressed the marker gene encoding DsRED to facilitate their identification using red fluorescence. The effect of the transgenes on MtAOC1 expression levels in transgenic root systems was confirmed by quantitative RT-PCR (Fig. 4a). Transcript levels were reduced by the MtAOC1-RNAi construct to c. 20% and increased by the MtAOC1-sense construct to 20–60-fold, both in comparison with the empty vector control. Moreover, the altered transcript levels resulted in changed protein levels as analysed by Western blot of transgenic roots and nodules (Fig. S3). In MtAOC1-RNAi roots, AOC protein was reduced compared with the empty vector control, whereas MtAOC1-sense-expressing roots exhibited elevated levels of AOC protein. In all roots analysed, the marker protein DsRED was clearly detectable at similar amounts. To check whether altered AOC protein levels resulted in altered JA levels, we used wounding as a stimulus, which is known to increase JA levels in M. truncatula roots (Tretner et al., 2008). Therefore, the roots of composite plants were wounded and the JA content was determined (Fig. 4b). Compared with the empty vector control, which showed a wound-induced six-fold increase in JA content, the MtAOC1-RNAi-induced reduction in AOC level abolished the increase in JA content. Conversely, higher AOC protein levels, resulting from the expression of the MtAOC1-sense construct, led to a slightly higher JA content in response to wounding. However, no effect of the transgenes on JA levels in noninoculated and nodulated roots was detected (Fig. 4c).
Nodules formed at transgenic roots were analysed using cytological methods to investigate alterations in their morphology (Fig. 5). Macroscopically, the nodules of all transgenic roots appeared to be normal (Fig. 5a). Longitudinal sections of transgenic nodules from all root types showed normal morphology compared with wild-type nodules of M. truncatula (Fig. 5b,c). The different zones (meristematic; infection; nitrogen fixation; senescing zone) were distinguishable and did not exhibit abnormalities. Moreover, immunolocalization revealed that the modulation of MtAOC1 expression levels resulted in the desired effect on MtAOC protein amounts; as shown in Fig. 5(d,e), the sense construct led to increased amounts of MtAOC1 in uninfected cells of the inner tissue and in the nodule cortex, whereas the activity of the RNAi construct abolished the formation of MtAOC protein.
To independently analyse the effect of the transgenes on nodule number, composite plants containing transgenic roots were either grown in expanded clay and infected by the application of S. meliloti to the stem base (Fig. 6a) or grown on pouch-paper (Journet et al., 2006) and infected by submersion in low-density suspensions of S. meliloti (Fig. 6b). Differences in nodule number could be detected between the independent experiments, but not between the three types of plant (MtAOC1-RNAi, empty vector, MtAOC1-sense) within each experiment. Moreover, the nodules of all transgenic roots seemed to be functional, as they were pink in colour (Fig. 5a) and none of the composite plants showed symptoms of nitrogen deficiency (data not shown). This suggests that the altered capacity of roots to produce JA has no influence on the number and functionality of nodules in M. truncatula.
The localization of AOC protein in the nodules of three different root nodule symbioses was analysed by immunolocalization, and the levels of JA in roots and nodules were determined. To determine whether a possible cell-specific production of JA, possibly undetectable by the extraction methods used, might play a role in nodulation, a transgenic approach was employed to modulate the JA biosynthetic capacity of roots and developing nodules.
AOC catalyses a committed step in JA biosynthesis. Hence, the cell-specific distribution of AOC in nodules should reflect the cell types in which JA can be synthesized. The correlation between the occurrence of AOC protein and JA biosynthesis capacity has been widely used to elucidate the role of JA in developmental processes (e.g. Hause et al., 2000; Cenzano et al., 2007). In this study, AOC protein was confined to plastids of uninfected cell types in the three types of root nodules examined. In the nodules of M. truncatula, AOC was found in the nodule cortex and vascular parenchyma, and in the uninfected cells of the central tissue. In C. glauca and D. glomerata, AOC protein was detected in amyloplasts of uninfected cortical cells. In all three species, the cell-specific occurrence of AOC was the same, leading to the conclusion that only noninfected cells of nodules have the capacity to synthesize JA.
This study shows the induction of JA biosynthesis within 30 min of mechanical disturbance in the roots and nodules of M. truncatula, C. glauca and D. glomerata. Touch-induced accumulation of AOS and AOC transcripts has also been shown in wheat, Arabidopsis and M. truncatula (Mauch et al., 1997; Lee et al., 2005; Tretner et al., 2008). In M. truncatula, increases in JA levels in response to mechanostimulation were equal to those in response to wounding. However, this rise in JA levels preceded the induction of AOC expression in M. truncatula, indicating that the transcriptional induction is a result of enhanced JA levels, and not vice versa (Tretner et al., 2008). Similar results have been found for other plant species, where increased AOC levels in transgenic plants led to increased JA levels only on wounding, but not in nonstressed plants (Wasternack, 2007). Consistent with this finding, our results showed that the rise in JA levels did not involve an increase in AOC protein in any of the species examined. That mechanical disturbance of root systems leads to the production of signal molecules is not surprising, as the development of the root system must be regulated in order to counteract wind stresses (Stokes et al., 1995). Moreover, the enhanced JA levels in mechanically stressed nodules indicate that nodules are capable of synthesizing JA.
As a result of the mechanical disturbance involved in nodule harvesting, a comparison between the JA contents of roots and nodules of M. truncatula, C. glauca and D. glomerata is technically challenging. JA biosynthetic capacity, as indicated by the levels of AOC, seemed to be similar in the roots and nodules of M. truncatula, C. glauca and D. glomerata, whereas JA biosynthetic capacity, as indicated by the accumulation of JA in response to mechanical disturbance, appeared to be different, as the JA levels in shaken M. truncatula nodules were much higher than those in the other two species. A comparison of the JA contents of whole root systems of M. truncatula over 17 d revealed that the JA contents of the root systems changed slightly during the course of root and nodule development, but remained at basal levels known also from nonstressed vegetative organs of other plants (Miersch et al., 2008). However, in our analyses, no differences between the JA levels in noninoculated and nodulated roots could be detected.
As a result of the extraction of whole root systems, a possible cell-specific rise in JA content could not be excluded. To conclusively determine whether JA affects the induction, development or function of nodules, composite M. truncatula plants were created by overexpression of MtAOC1 or by RNAi-mediated silencing of MtAOC1 expression in transgenic roots. The composite plants harbouring transgenic roots after A. rhizogenes-mediated transformation represent a valuable tool for the analysis of the biotic interactions of the root (Mrosk et al., 2009). The effects of the transgenes were confirmed at the transcript and protein levels. However, the altered MtAOC1 levels had no significant effect on the overall JA levels in roots. This paralleled the results presented by Isayenkov et al. (2005), who used MtAOC1 antisense constructs in transgenic root systems of M. truncatula and did not find an effect of the antisense construct on JA levels in nonmycorrhizal roots. They did, however, find an effect on mycorrhizal roots, where JA levels increased dramatically in control roots, but not in roots containing the MtAOC1 antisense construct. Similarly, in this study, wounding led to increased JA levels in control roots and in roots overexpressing MtAOC1, but did not cause a significant increase in the JA levels of roots expressing the MtAOC1-RNAi construct.
As nodulation has been shown not to affect JA levels in wild-type root systems, any observed effect of MtAOC1 overexpression or knockdown on nodulation would indicate that JA acts at the cellular level. However, this was not observed in our study. Therefore, although the repression of JA production in arbuscule-containing cells of M. truncatula roots reduced arbuscular mycorrhizal colonization, the alteration in JA production in uninfected cell types of roots and nodules had no effect on nodulation and nodule development.
This result was surprising, in view of the fact that exogenously applied JA/MeJA has been shown to affect Ca2+ spiking and gene induction in response to Nod-factors (Miwa et al., 2006; Sun et al., 2006), and has been implicated in the autoregulation of nodulation as a shoot-derived signal (Nakagawa & Kawaguchi, 2006; Seo et al., 2007). However, recent results suggest that JA cannot represent a shoot-derived signal (Magori & Kawaguchi, 2010), and the effects on Nod-factor signal transduction might also depend on OPDA-/JA-dependent induction of Ca2+ signals in response to wounding or herbivore attacks, as opposed to a direct role in nodule induction (Hause & Schaarschmidt, 2009).
JA has been suggested to be involved in sink signalling in arbuscular mycorrhiza (Isayenkov et al., 2005). Nodules of M. truncatula contain large amounts of starch, mostly in uninfected cells (Fig. 2), indicating that carbon sources from the phloem (sucrose) are available in surplus. Hence, it would be consistent that only uninfected cell types contain AOC and synthesize JA only under carbon-limiting conditions. If carbohydrates become limited, JA might be synthesized, leading to increased sink strength of nodules. Nodule sink strength is not dependent on bacterial nitrogen fixation; even bacteria-free nodules accumulate starch (Lin et al., 1988; Malek, 1988; Joshi et al., 1991; Blauenfeldt et al., 1994). Therefore, JA synthesis for sink signalling would be expected to be independent of nodule nitrogen fixation.
Our results reveal that the reduction of the JA biosynthetic capacity via RNAi-mediated knockdown of MtAOC1 expression in the root systems of M. truncatula does not affect nodule induction or development. This suggests that, even though JA synthesis in nodules is important in arbuscular mycorrhizal symbioses, it does not seem to play a role in root nodule development. In addition, our study shows that JA levels increase significantly on mechanical disturbance in all plant organs examined to an extent likely to affect transcription patterns. This has implications for the handling of plant material for transcript and protein analyses.
We thank Uwe Wedemeyer and Susanne Mester (Göttingen) for taking care of the plants, Ulrike Huth, Regina Franke (Halle/Saale) and Sabine Freitag (Göttingen) for dependable technical assistance, and N. Govindarajan (European Neuroscience Institute, Göttingen) for critical reading of the manuscript. A.Z. and K.P. acknowledge support by the Marie Curie Research Training Network INTEGRAL. This project was also supported by a grant from the German Research Council (DFG) to K.P. and a grant from the German Academic Exchange Service (DAAD) to K.D. K.D. acknowledges support by the Russian Fund for Basic Research (Grant no. 08-04-01710-a) and by the Russian Ministry of Education and Science (P289, P623, 02.740.11.0276).