Systemic resistance and lipoxygenase-related defence response induced in tomato by Pseudomonas putida strain BTP1
© Akram et al. 2008
Received: 22 January 2008
Accepted: 10 November 2008
Published: 10 November 2008
Skip to main content
© Akram et al. 2008
Received: 22 January 2008
Accepted: 10 November 2008
Published: 10 November 2008
Previous studies showed the ability of Pseudomonas putida strain BTP1 to promote induced systemic resistance (ISR) in different host plants. Since ISR is long-lasting and not conducive for development of resistance of the targeted pathogen, this phenomenon can take part of disease control strategies. However, in spite of the numerous examples of ISR induced by PGPR in plants, only a few biochemical studies have associated the protective effect with specific host metabolic changes.
In this study, we showed the protective effect of this bacterium in tomato against Botrytis cinerea. Following treatment by P. putida BTP1, analyses of acid-hydrolyzed leaf extracts showed an accumulation of antifungal material after pathogen infection. The fungitoxic compounds thus mainly accumulate as conjugates from which active aglycones may be liberated through the activity of hydrolytic enzymes. These results suggest that strain BTP1 can elicit systemic phytoalexin accumulation in tomato as one defence mechanism. On another hand, we have shown that key enzymes of the lipoxygenase pathway are stimulated in plants treated with the bacteria as compared with control plants. Interestingly, this stimulation is observed only after pathogen challenge in agreement with the priming concept almost invariably associated with the ISR phenomenon.
Through the demonstration of phytoalexin accumulation and LOX pathway stimulation in tomato, this work provides new insights into the diversity of defence mechanisms that are inducible by non-pathogenic bacteria in the context of ISR.
All plants have active defense mechanisms against pathogen attacks. If defense mechanisms are triggered by a stimulus prior to infection by a virulent plant pathogen, disease symptoms can be reduced. Some plant growth-promoting rhizobacteria (PGPR) are able to reduce disease through the stimulation of inducible plant defense mechanisms that render the host plant more resistant to further pathogen ingress. Since this induced systemic resistance (ISR)  is long-lasting and not conducive for development of resistance in the targeted pathogen, this phenomenon can be the basis of new plant disease control strategies both for greenhouse cultures and under field conditions, particularly in integrated pest management strategies [2–4]. ISR is phenotypically similar to the well-studied systemic acquired resistance (SAR) activated after a first infection by an incompatible necrotising pathogen . However, the signal transduction pathway and the molecular basis underlying ISR differ in many aspects from the pathogen-induced SAR. In the last case, several well-characterized defense reactions such as hypersensitive reaction (HR) , oxidative burst , reinforcement of cell wall structures through lignification or callose deposition [8–10], accumulation of antimicrobial phytoalexins [10–13] and induction of defense-related proteins with antifungal properties [14, 15] have been extensively reported in many plant species. By contrast, protective mechanisms involved in ISR are just beginning to be elucidated. For instance, the reinforcement of cell wall structures through lignification or callose deposition [16, 17], the accumulation of antimicrobial phytoalexins [18–20] and the induction of defense-related proteins with antifungal properties [21–23] have been reported following interactions with a pathogen. The activation of systemic resistance by nonpathogenic rhizobacteria has been also associated with the induction of lipoxygenase (LOX) activity in bean and tomato [24–27].
Plant LOX may be involved in growth and developmental control processes, through the biosynthesis of regulatory molecules and volatile compounds involved in insect attraction, but also in defense responses to pathogen, wounding and stress [28–31]. LOX catalyzes the incorporation of molecular oxygen in polyunsaturated fatty acids to yield the corresponding fatty acid hydroperoxides. These compounds are substrates for other enzymes such as (i) peroxygenase (POX) leading to the conversion into fungitoxic epoxy- and hydroxy-derivatives; (ii) allene oxide synthase (AOS) leading to the production of jasmonates known to be involved in signaling events and regulation of plant defense genes expression; (iii) hydroperoxide lyase (HPL) forming short chain aldehydes that are believed to behave as "volatile phytoalexins", and (iv) divinyl ether synthase (DES) which has been detected only in the Solanaceae [28, 31].
A non-pathogenic Pseudomonas putida strain (BTP1) isolated in the laboratory was shown to enhance the level of resistance in cucumber and bean against the pathogens Pythium aphanidermatum and Botrytis cinerea respectively. These studies revealed that the disease protective effect was associated with a systemic increase of antifungal phytoalexins in cucumber tissues  and the stimulation of the LOX pathway in bean . In this work, we first demonstrate the ISR-related protective effect triggered by P. putida BTP1 in tomato but also wanted to further characterize the plant defense mechanisms that could contribute to this enhanced level of resistance. On the basis of previous results obtained with this strain, we have more specifically investigated the accumulation of antifungal compounds and the possible LOX pathway induction in infected leaves with regard to disease symptom reduction.
Pseudomonas putida strain BTP1, isolated from barley roots, was originally selected for its specific features regarding pyoverdine-mediated iron transport [32, 33]. It was maintained and prepared for use in the ISR assays as previously described . The fungal pathogen Botrytis cinerea used for tomato infection was grown as described [32, 34].
Tomato seeds (Solanum lycopersicum L. cv Merveille des Marchés) were soaked, before sowing, for 10 min in BTP1 cell suspension at a concentration of 108 CFU/ml in 0.01 M MgSO4 or in 0.01 M MgSO4 without bacteria in the case of control plants. Then seeds were sown in 10 cm-pots containing sterilized potting soil (Brill Substrate GmbH, KG, Germany) previously mixed with bacterial inoculum to a final concentration of 3.107 CFU/g or with an equal volume of sterile water for untreated control plants. Tomato plants were germinated at 26 ± 2°C in the greenhouse with a 16 h photoperiod. Fifteen days after sowing, 20 ml of a bacterial suspension at 108 CFU/ml was added to the roots of BTP1-treated plants in order to ensure a high level of colonization by the strain. Five-week-old tomato plants were used for infection with B. cinerea on excised third leaves [35, 36]. The spore suspension was prepared by growing the fungus for three weeks on an oat-based medium (oatmeal 25 g l-1; agar 12 g l-1) at room temperature in the dark. During the last week, plates were placed under 16h-photoperiod UV illumination to induce sporulation. The spore suspension was prepared by harvesting spores in sterile peptone water containing 0.01% Tween 80. After removing mycelial debris by filtration through several layers of cheese cloth, the suspension was centrifuged for 5 min at 5000 g and the spores were resuspended in 0.01 M glucose, and 6.7 mM KH2PO4 to a final concentration of 105 spores/ml.
Detached tomato leaves were infected with B. cinerea by depositing ten 4-μl drops of the pathogen spore suspension. Disease incidence was expressed in terms of the percentage of B. cinerea lesions that clearly grew out of the inoculum drop zone to produce spreading lesions. It was recorded 24, 48 and 72 h after infection by using 20 excised third leaves per treatment and 45 leaves per treatment were used for final scoring at 96 h post-infection.
After grinding frozen tissue samples in liquid nitrogen, RNA was extracted and purified by a phenol/SDS method . In all cases, 20 μg total RNA was loaded per lane on formaldehyde agarose gels and blotted onto Hybond N+ membranes (Amersham, Little Chalfont, UK). Equal lane loading was checked by visualising ethidium bromide-stained ribosomal RNA after agarose gel electrophoresis. The Northern blots were hybridised with DNA probes labelled by random priming in the presence of [α-32P]dATP, according to the procedure recommended by the manufacturer (Random Primers DNA Labelling System; Invitrogen, Carlsbad, Calif.). The probe used was a 705 bp cDNA including the coding sequence of the PR-1a tomato gene . After hybridization, the blots were washed and then exposed to X-ray film (Fujifilm, Japan) for at least 24 h.
Methanolic extracts of ground leaf tissues ground in liquid nitrogen were prepared from two-g samples of frozen material collected at the specified time points .
Free and glycosidic phenolic compounds as well as aglycones released after acid hydrolysis were extracted from leaves first in 80% methanol (10 ml/g of material fresh weight) for 18 h in the dark at room temperature and then for an additional 4 h in pure methanol under the same conditions. Methanolic extracts were then collected by filtration on a Whatman no. 1 filter paper disc and concentrated by rotoevaporation to a final volume of 80 mL (aqueous fraction). Concentrates were then partitioned three times against the same volume of hexane to remove waxes and pigments and then against diethyl ether in the same conditions. Free phenolics migrated in this last organic phase, which was then evaporated to dryness and redissolved in methanol for further analyses. Aglycones were recovered from the aqueous fraction after acid-hydrolysis (HCl 4N, 100°C for 90 min) and subsequent extraction against diethyl ether and concentration to the same final volumes.
The fungitoxic activity of the various extracts was tested against Cladosporium cucumerinum after thin layer chromatography on silica gel plates (TLC). 35 μl of hydrolysed leaf extracts in diethyl ether corresponding to 35 mg leaf fresh weight were applied on TLC plate and developed with a mixture of dichloromethane: hexane: methanol (6:4:1, v/v/v). A conidial suspension of C. cucumerinum was then sprayed on the dried plate and fungitoxic zones (white spots) were observed after incubation for 48 h.
Acid-hydrolyzed extracts from third leaves were analyzed by reverse phase high performance liquid chromatography (HP 1100 series system, Agilent Technologies) on a LiChrospher 100 RP C-18 column (250 by 4.6 mm, 5-μm packing; Merck, Darmstadt, Germany). The apparatus was coupled with a photodiode array detector allowing on-line visualization of the absorbance spectra of eluted molecules and peak purity verification. Thirty-μl volumes were injected and eluted with a gradient of acetonitrile in H2O/acetic acid 0.1%, as follows (time in min/percentage acetonitrile/flow rate in ml/min: 0/5/0.25, 2/5/2.5, 2.5/5/1, 5/5/1, 15/30/1, 25/40/1, 40/45/1, 50/60/1, 60/65/1, 65/95/1, 75/95/1, 76/5/1, 83/95/1. Data were analyzed using the Chemstation Software (Agilent Technologies). For each treatment, HPLC runs were repeated at least twice and results presented are representative of all tested aliquots from all experiments.
Spores were prepared as described for the ISR assays but were finally resuspended in sterile distilled water to the desired final concentration of 106 spores/ml. This suspension was used immediately. Antifungal activities of the various tomato leaf extracts were evaluated in microculture assays by growing the fungus in sterile 96-well microplates in a final volume of 125 μl containing 100 μl of clarified V8 juice medium (5% V8 juice in distilled water, centrifuged at 3,000 g for 5 min). Each well was inoculated with 15 μl of the B. cinerea spore suspension and treated with 10 μl of methanol (controls) or methanolic solutions of the various compounds or extracts to be tested. Preliminary assays showed that methanol at this final concentration reduced B. cinerea growth by about 25% compared to water treatment but fungal development was absolutely not impaired and we consider these conditions acceptable since significant increases in optical density were recorded during the 48 h-time period of the test. Fungal growth in the presence of the various compounds/extracts was monitored by measuring the optical density (OD) at 620 nm with a microplate reader (Beckman Coulter AD340) at 0 h and after 24 h and 48 h of incubation. In every experiment, eight replicates (wells) in the same plate were used for each treatment and the experiments were repeated three times with newly prepared B. cinerea spore suspensions. Growth inhibitions were calculated on the basis of final ODs within experiments.
The antifungal compound used in these tests was purified by repeated HPLC injections of 20-μl aliquots of the crude aglycone extract in the same conditions as described above. Samples were monitored spectrophotometrically at 214 nm and 2-ml samples were collected automatically over the entire run time. Fractions containing the peak of interest were pooled and evaporated in a Speed-Vac concentrator. The pure compound was resolubilized in pure methanol in order to obtain a final concentration similar to the one in the crude extract. Fractions collected all over the runs except the one containing the antifungal compound were pooled and the resulting solution was lyophilized and the residue resolubilized in methanol so as to test the remaining constituents at the same concentration as the elicitor.
The mass spectrometry analysis was performed with a 9.4 tesla Apex-Qe FTICR mass spectrometer (Bruker Daltonics, Billerica, MA) in positive ion mode. The sample solution (10 μl of the sample, 490 μl of MeOH and 500 μl of water) was infused via an external Apollo electrospray ion source at a flow rate of 120 μl/h with the assistance of N2 nebulizing gas. The off axis sprayer was grounded, the end-plate was set to -4 kV and the inlet capillary was set to -4.5 kV for the generation of positively charged ions. N2 heated drying gas (250°C) was applied to assist desolvation of ESI droplets. For CID (Collision Induced Dissociation) experiments, the precursor ion (m/z = 288.25) was selected and accelerated at the entrance of the collision-cell. To produce this acceleration, the Col-Cell Trap voltage was increased from -3.5 to -12.5 V. The collision gas was argon at a pressure of ~10-3 mbar.
The preparation of crude extracts and the determination of enzyme activities were performed by following the methods detailed in Ongena et al. (2004) . LOX activity and LHP activity were determined spectrophotometrically by monitoring hydroperoxide increase or decomposition at 234 nm.
The enhanced state of resistance induced by some non-pathogenic Pseudomonas strains was occasionally related to the SAR response. In order to determine if P. putida BTP1 induces a SAR-type response in tomato the expression level of the PR1a gene, considered as good molecular marker for SAR, was examined. This was carried out by hybridisation of a labelled specific cDNA probe on Northern blots. Before infection both P. putida BTP1-treated plants and control plants exhibited the same low level of PR1a gene expression (data not shown). This limited accumulation of PR1a transcripts could be due to wounding during leaf excision. After infection by B. cinerea both types of plants showed very high levels of PR1a transcripts (data not shown).
Peak area corresponding to this compound (Fig. 3B) showed increased size in all plants after infection, suggesting an accumulation of this molecule. While peak areas were similar before infection in control and P. putida BTP1-treated plant extracts, we clearly observed a more important and rapid accumulation during the first 16 hours after pathogen challenge in bacterized plants as compared with control plants.
The antifungal compound eluted at 47 minutes was purified by HPLC and its molecular weight was determined by electrospray mass spectrometry at a value of m/z 288.23279 Da and 310.235194 Da for the protonated molecular ion [M+H]+ and [M+Na]+ respectively) (Fig. 3C). This corresponds to an exact mass of 287.253278 Da for the molecule, which could have the brut formula C16H33NO3.
LHP activities were also measured in extracts prepared from leaves collected in three separate experiments before and after pathogen challenge. Prior to pathogen challenge, LHP activity in plants colonized by P. putida BTP1 was similar to that observed in non-inoculated plants. However, a higher LHP activity was observed in P. putida BTP1-treated plants in comparison to control plants two and four days after infection. After four days, bacterized plants showed 1.7-fold higher LHP activities than control plants (Fig. 5C). Time-course stimulation of both LOX and LHP activities are associated with the reduction of disease symptoms (Fig. 5B).
As a plant pathogen can induce SAR, non-pathogenic rhizobacteria can also stimulate a phenotypically similar induced systemic resistance (ISR) response in the host plant. Nevertheless molecular events underlying ISR are less well understood than in the case of SAR. In this context the work presented here was initiated with the aim to correlate metabolic changes with the observed B. cinerea disease reduction in tomato plants treated at the root level with P. putida BTP1.
The signal transduction pathway leading to ISR usually requires responsiveness to jasmonate and ethylene [41–43]. Thus, in most cases, ISR differs from SAR. However a SAR-type response involving free SA accumulation and/or PR protein expression was also occasionally reported in the systemic resistance induced by some PGPR [23, 44]. As a first step in the characterization of specific host metabolic changes associated with the protective effect, northern analyses showed that the resistance induced by P. putida BTP1 in tomato is not accompanied by the expression of the PR1a gene considered as good marker of SAR . The P. putida BTP1-induced resistance observed here in tomato is therefore most probably not related to SAR. This statement could moreover be generalized to other plant species since the P. putida BTP1-triggered ISR response in bean is SA-independent  and cucumber plants treated with the bacterium did not show any induction of PR-8 gene expression .
We investigated the possible induction of ISR-related defence mechanisms tightly involved in restriction of pathogen ingress. Our results first showed a clear accumulation of antifungal products in leaves of P. putida BTP1-treated plants compared to controls (Fig. 2A). These compounds only accumulate after pathogen challenge and are probably de novo synthesized in plant cells as conjugates since their activity can be visualized in biotests only after acid hydrolysis. However, the time-course accumulation of this material inhibitory to B. cinerea is associated with the reduction of disease symptoms (Fig. 3) suggesting an involvement of the molecule in direct pathogen inhibition. In general, the relationship between accumulation of conjugates and resistance of some plants has already been well established in other pathosystems and is argued to be a necessary component of the chemical response that could be involved in pathogen restriction [19, 46–48]. It is postulated that fungitoxic aglycones are gradually liberated from inactive conjugates by cleavage of the sugar bond by hydrolytic enzymes released by the pathogen itself during tissue ingress . Results from our previous work on ISR in cucumber treated with P. putida BTP1 and challenged with Pythium aphanidermatum also illustrated the accumulation of antifungal phenolic compounds in a conjugated form. These antifungal compounds mainly accumulated in cucumber roots and leaves as a systemic response [19, 20]. Intriguingly, cucumber and bean plants systemically protected with P. putida BTP1 did not accumulate any antifungal product after challenge with Colletotrichum lagenarium or with B. cinerea respectively [26, 50], suggesting that phytoalexin accumulation is not always associated with P. putida BTP1-induced ISR.
Despite the information provided by mass spectrometry, the molecule could not be identified in this work. Several lines of evidence suggest that this compound does not correspond to a phenylpropanoid-derived phenolic. Firstly its molecular weight, UV-visible spectral properties and HPLC retention time, are clearly different from those of various representative standard compounds under the same conditions (data not shown). Secondly this metabolic pathway is not stimulated as suggested by the absence of differential expression of the PAL gene (unpublished results). For the same reasons, it also does not correspond to the major tomato phytoalexins such as rhishitin, lubimin and solavetivone [13, 51, 52]. This hydrophobic molecule accumulates concomitantly with the stimulation of the LOX enzyme but nothing indicates that it could therefore derive from the oxylipin pathway. Both its spectral properties and molecular weight do not correspond to those of LOX-derived products that may be toxic for B. cinerea such as the 9- or 13-hydroperoxides, their corresponding hydroperoxy-octadecadi(tri)enoic acids and the phytodienoic acid or, to a lower extend, the hydroxy-octadecadienoic and colnelenic acids . Further NMR and 2D-MS analyses are underway and will certainly provide further insights about the chemical nature of the molecule. In this work, significantly enhanced levels of LOX activities were observed in P. putida BTP1-elicited tomato plants compared to controls (Fig. 5A). Higher levels of the global activity of all hydroperoxide-degrading enzymes were also maintained over four days after pathogen challenge. It is thus obvious that the entire metabolic route leading to oxylipins was induced in resistant tomato plants inoculated with P. putida BTP1. Further oxylipin profiling in these induced plants is been performed in order to identify bioactive LOX-derived products that accumulate concomitantly with enzyme stimulation.
This LOX pathway can be activated under different abiotic stress conditions, in response to treatment with chemicals or biotics elicitors but also following interactions with pathogens . This metabolic route leads to the synthesis of various compounds displaying antimicrobial  or signaling activities or both [43, 55]. It has been recently well illustrated by the study of Prost and collaborators showing the inhibitory activity of a number of oxylipins toward various bacterial, oomycete and fungal plant pathogens . Included in the tested organisms, B. cinerea is highly sensitive to some of these LOX-derived products, especially polyunsaturated fatty acid hydroperoxides and their reduced forms. The strong LOX stimulation observed in BTP1-induced tomato plants could result in a accumulation of such hydroperoxides that may constitute a first chemical barrier to B. cinerea penetration. Oxylipins may thus directly or indirectly contribute to the restriction of pathogen ingress in local or systemic defense reactions such as SAR. LOX induction associated with PGPR-induced systemic resistance was occasionally reported [24, 25] but to our knowledge, evidence for a stimulation of the complete LOX pathway correlating with disease reduction was only reported in bean protected by P. putida BTP1 against B. cinerea infection. LOX and hydroperoxide lyase (HL) were indeed significantly stimulated during the first 4 days after pathogen challenge of bacterized plants. In parallel a more rapid consumption of these enzyme's respective substrates and significantly higher concentrations of the fungitoxic final product Z-3-hexenal were reported . As the quantities effectively produced within the leaf tissues might be in the range that could locally restrict hyphal penetration of B. cinerea, we have postulated that these molecules could play a role as "volatile" phytoalexins. In the present work on tomato, headspace-gas chromatography analyses of fresh leaf materials did not show any accumulation of these aldehydes or other volatile products in plants bacterized with P. putida BTP1 (data not shown). This suggests that a given rhizobacterium can induce the LOX pathway in various plants but with different outcomes regarding the type of oxylipin that will ultimately accumulate in infected tissues and putatively restrict pathogen ingress. It also suggests that this species specificity may lie in the hydroperoxide degradation to different end-products in function of the relative activities of peroxidase, divinyl ether synthase, allene oxide synthase, hydroperoxide lyase and lipoxygenase.
According to the priming concept, our work on tomato treated with BTP1 suggests a pathogen-dependant systemic activation of the defense reaction. Whether they are related or not, both accumulation of antifungal compounds and stimulation of the first part of the oxylipin pathway only occur when bacterized plants were challenged with B. cinerea (Fig. 3 and 5). The results thus show that pathogen perception is necessary for LOX activation in tomato but it is significantly enhanced upon pre-treatment of the plant at the root level with strain BTP1. In light of recent researches, such a pathogen-dependant enhanced expression of host defense mechanisms appears to be a common feature of ISR induced by beneficial rhizobacteria (and fungi) in plants. PGPR-inoculated roots do react locally to colonization by the bacteria and are primed but defense responses are not activated directly in the whole plant and are only observed in leaf tissues upon pathogen perception .
In spite of the numerous examples of ISR induced by PGPR in plants, only a few biochemical studies have associated the protective effect with specific host metabolic changes [4, 19, 53, 54]. Through the demonstration of LOX pathway stimulation in tomato, this work provides new insights into the diversity of defence mechanisms that are inducible by non-pathogenic bacteria in the context of ISR. LOX induction occurs in tomato and bean but not in cucumber indicating that such a stimulation of the oxylipin pathway is not strictly related to the ISR induced by P. putida BTP1. Molecular events associated with this phenomenon seem to depend on the plant, the inducing microorganism and even the pathogen. Further studies on a broader set of pathosystems are thus required to evaluate the specificity of the involvement of some specific defence mechanisms more accurately.
The first author would like to thank the Atomic Energy Commission of Syria (AECS) for financial support. This work received financial support from the Program F.R.F.C. n° 2.4624.06 (National Funds for Scientific Research, FRS-F.N.R.S., Belgium). M. Ongena is research associate at the FRS-F.N.R.S.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.