• Fusaric acid (FA) is a toxin produced by Fusarium species. Most studies on FA have reported toxic effects (for example, alteration of cell growth, mitochondrial activity and membrane permeability) at concentrations greater than 10−5m. FA participates in fungal pathogenicity by decreasing plant cell viability. However, FA is also produced by nonpathogenic Fusarii, potential biocontrol agents of vascular wilt fusaria. The aim of this study was to determine whether FA, at nontoxic concentrations, could induce plant defence responses.
• Nontoxic concentrations of FA were determined from cell-growth and O2-uptake measurements on suspensions of Arabidopsis thaliana cells. Ion flux variations were analysed from electrophysiological and pH measurements. H2O2 and cytosolic calcium were quantified by luminescence techniques.
• FA at nontoxic concentrations (i.e. below 10−6m) was able to induce the synthesis of phytoalexin, a classic delayed plant response to pathogen. FA could also induce rapid responses putatively involved in signal transduction, such as the production of reactive oxygen species, and an increase in cytosolic calcium and ion channel current modulations.
• FA can thus act as an elicitor at nanomolar concentrations.
In the present study, we focused on the effect of fusaric acid (FA), a nonspecific toxin produced by all Fusarium species (Bacon et al., 1996). FA elicits its action in a diversity of organisms, which are phylogenetically remote. In mammals, FA at concentrations higher than 10−5m induced effects on different organ systems, including cardiovascular and immune systems (Wang & Ng, 1999). In plants, most of the studies on FA reported toxic effects (e.g. alteration of membrane permeability, decrease of mitochondrial activity and O2 uptake, inhibition of ATP synthesis, inhibition of root growth) at concentrations higher than 10−5m (d’Alton & Etherton, 1984; Marréet al., 1993; Vurro & Ellis, 1997; Bouizgarne et al., 2004). Although its involvement in fungal pathogenicity has not been clearly established (Gapillout et al., 1996), higher concentrations of FA were detected in plant tissues infected with virulent strains than in those infected with avirulent strains (Harborne, 1989). FA could thus be involved in fungal pathogenicity by decreasing plant cell viability (Gapillout et al., 1996; Kuzniak, 2001). However, various nonpathogenic strains of F. oxysporum are considered as potential biocontrol agents of vascular wilt fusaria (Alabouvette et al., 2001; Cachinero et al., 2002). Under optimal conditions for toxin production, the culture of a nonpathogenic strain of F. oxysporum, used in biocontrol of the weed, Striga hermonthica, produced FA and dehydrofusaric acid, but no other toxins (Savard et al., 1997). As FA can induce typical rapid responses of plant to pathogen, such as production of reactive oxygen species (ROS) (Kuzniak et al., 1999; Kuzniak, 2001) and modifications of membrane potential in various cell types (d’Alton & Etherton, 1984; Marréet al., 1993; Bouizgarne et al., 2004), it is probably involved in the early steps of signal exchange between the fungus and the plant.
A. thaliana L. (ecotype Columbia) suspension cells were grown at 24 ± 2°C, under continuous white light (40 µE m−2 s−1), with rotation shaking in a 1-l round-bottom flask containing 350 ml of Gamborg culture medium (El-Maarouf et al., 2001; Reboutier et al., 2002; Brault et al., 2004). The pH of the culture medium was 5.8. Cells were subcultured weekly by a 10-fold dilution in fresh medium. Cell growth was estimated 24 h after the treatment by FA from the cell sediment obtained after 1 h with 50 ml of cell suspension.
For the measurement of O2 uptake, an oxygraph (Hansatech Instruments Ltd, Cambridge, UK), previously calibrated according to the manufacturer's instructions, was used. A 1.5-ml sample of treated cell suspension was transferred to the recording chamber and the oxygen consumption was measured for a time-period of 5 min. Values were converted to percentage of control, and the data represent the means of three replicates of two independent experiments. For real-time measurements, the O2 uptake of 1.5 ml of cells was determined over a 2-min time-period, then FA was added. Data represent the means of five replicates of two independent experiments.
Aequorin luminescence measurements
Cytoplasmic Ca2+ variations were recorded with an A. thaliana cell suspension expressing the apoaequorin gene (Brault et al., 2004). For calcium measurements, aequorin was reconstituted by overnight incubation of the cell suspension in Gamborg medium supplemented with 30 g l−1 sucrose and 2.5 µm native coelenterazine. Cell-culture aliquots (250 µl) were transferred carefully to a luminometer glass tube, and the luminescence counts were recorded continuously, at 0.2-s intervals, with a FB12-Berthold luminometer (Berthold Technologies, Bad Wildbad, Germany). Treatments were performed by pipette injection of 25 µl containing the effectors. At the end of each experiment, the residual aequorin was discharged by the addition of 10% (v/v) ethanol and 1 m CaCl2 (final concentration). The resulting luminescence was used to estimate the total amount of aequorin in each experiment. Calibration of the calcium measurement was performed by using the equation pCa = 0.332588(–log k) + 5.5593, where k is a rate constant equal to luminescence counts per second divided by the total remaining counts (Knight et al., 1996).
Measurement of H2O2
H2O2 release in the culture medium was quantified by measuring the chemiluminescence of luminol reacting with H2O2 (Jabs et al., 1997). Briefly, 1.5 ml of the cell suspension was inoculated with FA. Then, 200 µl of the culture medium was added to 600 µl of phosphate buffer (50 mm, pH 7.9) before the addition of 100 µl of 1.1 mm luminol and 100 µl of 14 mm K3[Fe(CN)6]. Chemiluminescence was monitored at 5-min intervals using an FB12-Berthold luminometer, with a signal integrating time of 0.2 s.
The suspension cells were impaled in the culture medium with borosilicate capillary glass (Clark GC 150F; Harvard Apparatus, Edenbridge, UK) micropipettes (resistance: 50 mΩ when filled with 600 mm KCl), as previously described (El-Maarouf et al., 2001; Reboutier et al., 2002; Brault et al., 2004). Individual cells were voltage-clamped using an Axoclamp 2B amplifier (Axon Instruments, Foster City, CA, USA) for discontinuous single electrode voltage clamp experiments (Finkel & Redman, 1984). Voltage and current were digitized with a personal computer fitted with a Digidata 1320A acquisition board (Axon Instruments). The electrometer was driven by pClamp software (pCLAMP8; Axon Instruments). The experiments were conducted on 4-d-old cultures [the main ions in the medium after 4 d of culture were 9 mm K+ and 11 mm No3−; Reboutier et al. (2002)]. Experiments were performed at 22 ± 2°C.
Extracellular pH measurements
Extracellular pH was measured directly in the medium of suspension cells (Brault et al., 2004). The experiments were run simultaneously in 6 × 10-ml flasks (control and tests) – each containing 1 g fresh weight (f. wt) of cells per 5 ml of suspension medium – with orbital shaking at 60 r.p.m. For each condition, the pH of the medium at the start of the experiment was between 5.8 and 6.0. Simultaneous changes in pH were followed by using the ELIT 808 ionometer with pH-sensitive combined electrodes functioning in parallel. The effectors were added when a stable pH was obtained and values were monitored for 45 min.
Camalexin determination was performed as described by Glazebrook & Ausubel (1994). For each sample, 20 ml of suspension cells was filtered, and 1 g of cells was heated in 350 µl of 80% (v/v) methanol for 20 min. The cells were removed and the methanol was evaporated under vacuum. The aqueous residue was extracted with two 50-µl aliquots of chloroform, which were combined and evaporated to dryness. The residue was dissolved in 100 µl of chloroform, applied to silica thin-layer chromatography plates and developed in ethyl acetate/hexane (9 : 1, v/v). Camalexin (Rf 0.81) was visualized by its blue fluorescence under a long-wave ultraviolet lamp (365 nm). The silica containing camalexin was scraped off the plate and camalexin was extracted into 1 ml of ethanol. The emission at 385 nm after excitation at 315 nm was measured using a Hitachi F2000 fluorimeter (Hitachi Ltd, Tokyo, Japan) and the camalexin concentration was calculated by comparison with a standard curve obtained by using purified camalexin kindly provided by Jane Glazebrook (Torrey Mesa Research Institute, San Diego, CA, USA).
Cell viability assay
Cell viability was assayed using the vital dye, neutral red, as described by Wendehenne et al. (2002). Cells (1 ml) were washed with 1 ml of a solution containing 175 mm mannitol, 0.5 mm CaCl2, 0.5 mm K2SO4 and 2 mm Hepes, pH 7.0, and incubated for 5 min in the same solution supplemented with neutral red to a final concentration of 0.01%. Cells that did not accumulate neutral red were considered to be dying. At least 500 cells were counted for each treatment. The experiment was repeated at least three times.
FA (5-butylpicolinic acid), acetic acid, luminol, pectolyase and 9-anthracen carboxylic acid (9-AC) were purchased from Sigma-Aldrich (Lyon, France). Uptima (Montluçon, France) supplied native coelenterazine. FA and acetic acid stock solutions were buffered at pH 5.8 with KOH.
Significant differences between treatments were determined by the Mann–Whitney test, and P-values of < 0.05 were considered significant.
Determination of toxic and nontoxic concentrations of FA
Numerous toxic effects of FA have been previously described on different animal and plant responses. FA notably inhibits cell growth and O2 uptake. We chose to determine toxic and nontoxic concentrations of FA on our model (suspension cells of A. thaliana) by scoring, in a dose-dependent manner, the cell suspension growth and the O2 uptake. The addition of FA to the medium resulted in a dose-dependent inhibition of the cell suspension growth after 24 h (Fig. 1a). The growth was drastically reduced for FA at concentrations of > 10−5m. At FA concentrations of < 10−6m, the growth was not significantly altered by comparison with the control. Similarly, the addition of FA to the medium resulted in a dose-dependent inhibition of cell-suspension O2 uptake (Fig. 1b). At FA concentrations of < 10−5m, the O2 uptake was not significantly altered compared with the control. Under the same experimental conditions, the O2 uptake was reduced for FA concentrations of > 10−5m. This decrease was rapid as the slope of O2 consumption changed a few seconds after the addition of 10−3m FA to the medium (data not shown).
From these data, we chose 10−7m FA as a nontoxic concentration and 10−4m or 10−3m FA as toxic concentrations. We then tested these FA concentrations on the early signalling events detectable in plant–pathogen interactions, namely ROS production, increase in internal Ca2+ and ion flux variations, and the later phytoalexin synthesis.
A rapid reaction of the Arabidopsis cell culture to FA was recorded as an increase in luminol-mediated chemiluminescence caused by H2O2 release into the culture medium. Within 10 min after the addition of 10−7m FA, the H2O2 concentration of the culture medium began to increase, reaching a maximum at c. 30 min (Fig. 2a). Thereafter, the H2O2 concentration decreased to a level similar to that of the control by c. 40 min. Treatment with the 10−4m toxic concentration of FA induced an oxidative burst of about the same intensity and kinetics as the nontoxic concentration (Fig. 2a). A hypo-osmotic chock used as a positive control largely increased H2O2 production (data not shown), as previously reported (Coelho et al., 2002), when nontreated cell suspensions showed no significant increase in the external H2O2 level during the experiments. Modulation of internal pH has been shown to modify ROS production (Suhita et al., 2004). To investigate the putative effect of cytosol acidification triggered by FA diffusion and acid load, we used acetic acid, another weak acid with a pKa similar to that of FA. As shown in Fig. 2(a), FA at 10−7 and 10−4m failed to increase the H2O2 level in the external medium.
Increase in cytosolic calcium
The addition of 10−7m FA induced a slight increase of the cytosolic calcium concentration ([Ca2+]cyt) within 30 s (Fig. 2b), thus representing one of the earliest detectable reactions of the cells. This increase remained stable for at least 20 min (data not shown). At 10−4m, FA also induced an increase in the [Ca2+]cyt, which displayed the same kinetic and amplitude as that of 10−7m FA (Fig. 2b). Addition of 25 µl of cell culture medium as a negative control failed to increase [Ca2+]cyt, whereas pectolyase, used as a positive control, induced a large and rapid increase of cytosolic Ca2+ (Fig. 2b), as previously reported (Carden & Felle, 2003).
Ion current variations
Under control conditions, successful impalements were routinely obtained in turgid cultured cells of A. thaliana. The value of the resting membrane potential (Vm) in the culture medium was −42 ± 3 mV (n = 88). By using electrophysiological and pharmacological analysis, we identified two main ion channel currents in the plasma membrane: a K+ outward rectifying current (KORC) and an anion current, which display the main hallmarks of slow anion channels (Reboutier et al., 2002).
A typical example of the Vm recording of a cultured cell treated with 10−7m FA is shown in Fig. 3(a). FA induced a transient hyperpolarization in less than 2 min. Then, the membrane depolarized slowly to the Vm value recorded before FA addition and even to slightly lower potential values (Fig. 3a,b). The same typical electrical behavior, hyperpolarization–depolarization, was observed by using FA at 10−3m, but the slight transient hyperpolarization was followed by a large irreversible depolarization (Fig. 3a,b). Acetic acid, used at the same concentrations as FA, did not mimick its effect on membrane potential (Fig. 3b), despite the putative regulation by internal pH variations of the main ion currents recorded in our model (Lacombe et al., 2000; Colcombet et al., 2005).
The anion current intensity was measured at different time-points during the Vm recording [i.e. (i) before FA addition, (ii) at optimal hyperpolarization (after c. 1 min), and (iii) at depolarization (Fig. 3a)]. Figure 3(c,d) shows the kinetics of these anion currents and the relative current voltage relationships that are consistent with the anion nature of the current. As indicated (Fig. 3c,d,e), 10−7m FA induced a transient decrease of anion current: the current intensity decreased during the hyperpolarization and then increased during the depolarization. With the toxic FA concentration (10−3m), the anion current variation evolved in the same ways as those induced by 10−7m FA, and the current intensities at maximal hyperpolarization and maximal depolarization were in the same range, regardless of the FA concentration (Fig. 3e). Acetic acid at 10−7 and 10−3m did not induce significant changes in anion currents (Fig. 3e). Acid load is therefore probably not responsible for the FA-induced modulation of anion currents.
As previously described, the KORC we recorded displayed the principal hallmarks of known outward K+ channels (Fig. 3f), a time-dependent kinetics and an activation threshold of the outward rectifying currents near (Schröeder, 1989). The KORCs were scored for the three specific time-points of the Vm recording [i.e. (i) before FA addition, (ii) at optimal hyperpolarization and (iii) at depolarization]. As indicated (Fig. 3f,g,h), the decrease of the KORC was not transient, with the intensity decreasing from the early hyperpolarization and continuing during the depolarization. With the toxic FA concentration (10−3m), the KORC variation evolved in the same ways as those induced by 10−7m FA, and the current decreases were in the same range for both FA concentrations (Fig. 3h). The use of acetic acid (at 10−7 and 10−3m) allowed us to check that the FA-induced decrease in KORC was not the result of acid load (Fig. 3h).
Culture medium alkalinization
We then analysed the putative role of plasma membrane H+-ATPases. FA was previously shown to alkalinize the external medium of different suspension cell cultures and to depolarize date palm root hairs by decreasing the H+-ATPase currents. In the present study, we observed a large depolarization of the cell in response to 10−4m FA, but not to 10−7m FA (Fig. 4); however, the anion and K+ current modulations were almost identical (Fig. 3h). We checked for the effect of FA on proton pumping – a hyperpolarizing mechanism – by following the external medium pH variations. External variations of pH were previously correlated to proton pumping on our model in the same conditions by using pharmacological agents efficient on H+-ATPases (Brault et al., 2004). Thus, the FA-induced alkalinization observed with 10−3m FA, but not with 10−7m FA (Fig. 4), suggested that the inhibition of the H+-ATPases could be responsible for the alkalinization and the large depolarization observed at FA toxic concentrations.
Induction of the phytoalexin, camalexin
We tested the induction of camalexin, a phytoalexin of A. thaliana, by FA. After 24 h, 10−7m FA induced the synthesis of camalexin (Fig. 5a). This synthesis was reduced for the 10−4 m toxic concentration, probably as a result of the increase in cell death observed at this concentration (Fig. 5b). Synthesis of phytoalexin from other species has been shown to be inhibited by anion channel blockers (Ebel et al., 1995; Jabs et al., 1997; Mithöfer et al., 2001). We showed that FA induced an increase in anion currents (Fig. 3c); thus we tested the effect of the anion channel blocker, 9-AC, on FA-induced camalexin synthesis. 9-AC was applied at 200 µm, a concentration sufficient to reduce significantly the anion current in our model. Figure 5(a) shows that 9-AC reduced camalexin synthesis induced by 10−7m FA, but this decrease was caused by the lethal effect of 9-AC (Fig. 5b) rather than by an effect on anion channel currents. Attempt to reduce 9-AC concentrations, or the use of other structurally unrelated anion channel blockers, were unsuccessful because of their lethal effect in our model, even at low concentrations (J. Briand, unpublished data).
The dose-dependent effect of FA recorded for A. thaliana cell suspension growth and O2 uptake enabled us to define toxic and nontoxic concentrations of FA. The growth and O2 uptake were drastically reduced for FA concentrations of > 10−5m, and the responses observed at FA concentrations of < 10−6m were not significantly altered compared with the control. These data are in accordance with (i) the increase of cell death recorded after 24 h at 10−4m FA, whereas 10−7m FA did not induce significant change (Fig. 5b) and (ii) with the results of other studies demonstrating the toxic effect of FA at concentrations of > 10−5m on the cell metabolism of various plant cell types (d’Alton & Etherton, 1984; Marréet al., 1993; Gapillout et al., 1996; Vurro & Ellis, 1997; Kuzniak, 2001). FA inhibits the oxidative phosphorylation through a direct action on the ATP-synthase/ATPase in rat liver mitochondria and maize root mitochondria, leading to the inhibition of oxygen uptake (Telles-Pupulin et al., 1998). In tomato mitochondria, 1 mm FA was shown to inhibit ATP synthesis (d’Alton & Etherton, 1984). Thus, the FA-induced decrease of available ATP could be responsible for the decrease of the plasma membrane H+-ATPase currents, as observed on date palm root hairs (Bouizgarne et al., 2004) and subsequently responsible for the large depolarizations observed on root hairs of different species (d’Alton & Etherton, 1984; Bouizgarne et al., 2004), on Egeria densa leaves (Marréet al., 1993) and on A. thaliana suspension cells at millimolar concentrations of FA (Fig. 3a). Furthermore, a reduced H+-ATPase activity is consistent with the alkalization of the external medium observed in response to FA concentrations of > 10−4m on the A. thaliana cell suspension (Fig. 4), on tomato cell suspensions (Gapillout et al., 1996; Kuzniak, 2001) and on leaf cells of E. densa (Marréet al., 1993). Thus, most of the described effects of FA at high concentrations could be explained by metabolism disruption, leading to cell death. However, FA could be considered as mildly toxic, as concentrations lower than 10−6m induced no change in O2 uptake, growth or death. From these data, we used 10−7m FA as a nontoxic concentration and 10−4m or 10−3m FA as toxic concentrations.
At nontoxic concentrations, FA induced a rapid and transient generation of H2O2. Although the amplitude of this increase was rather small under our experimental conditions, it revealed the rapid production of ROS, namely an oxidative burst, a characteristic event in the early phase of many plant–pathogen interactions (Baker & Orlandi, 1995; Jabs et al., 1997; Wojtaszek, 1997; Rajasekhar et al., 1999; Kawano, 2003). The same concentration of FA induced a rapid increase in internal Ca2+ variation, as frequently reported in plant–pathogen interactions (Atkinson et al., 1996; Levine et al., 1996; Lecourieux et al., 2002). This increase was low, but persistent in comparison with the effect of pectolyase used as a positive control. The pectolyase produced from Aspergillus japonicus is known for its important polygalacturonase activity (Mehri-Kamoun, 2001). F. oxysporum has been found to secrete various polygalacturonases, notably during the early stage of infection (Roncero et al., 2003). Thus, polygalacturonases and FA can participate in the sustained internal Ca2+ increase reported in response to pathogenic, but also nonpathogenic, strains of F. oxysporum (Olivain et al., 2003).
The nontoxic concentration of FA also modulated plasma membrane potential and ion currents. The early membrane hyperpolarization induced by FA involved a decrease of anion current; this current was then conversely regulated to allow membrane depolarization. This depolarization was accompanied by a decrease of KORC. These effects could not be attributed to acidification of cytosol triggered by diffusion and acid load of FA, as acetic acid did not modify ion-channel activities at the same concentrations. For toxic FA concentrations, the decrease of the H+-ATPase currents was added to the previous effects, explaining the largest depolarizations observed in this case.
In addition to these early physiological responses to nontoxic concentrations of FA, we observed the accumulation of a potentially antimicrobial phytoalexin, camalexin, a late defence response to pathogens and elicitors (Zhao et al., 2005). In different plant–pathogen interactions, the synthesis of phytoalexins has been reported to be reduced by anion channel blockers (Ebel et al., 1995; Jabs et al., 1997; Mithöfer et al., 2001). The FA-induced increase of anion current could not be correlated to phytoalexin synthesis owing to the toxic effect of the anion channel blocker, 9-AC. However, the different early responses observed at nontoxic FA concentrations, namely ROS production, [Ca2+]cyt increase and ion fluxes modulation, are frequently cross-linked in the plant response to pathogen and are also linked to phytoalexin synthesis (Price et al., 1994; Ebel et al., 1995; Ishihara et al., 1996; Jabs et al., 1997; Kawano & Muto, 2000; Able et al., 2001; Mithöfer et al., 2001). In conclusion, our data suggest that FA, in addition to its well-known toxic effects, and independently of any acid load, could act like an elicitor from nanomolar concentrations and elicit well-characterized plant cellular responses to pathogen attack.
We thank J. Glazebrook for providing purified camalexin, and R. Hasselberg and A. Dauphin for technical assistance. This work was supported by funds from the MENESR to EA 3514 and from the CIUFM to AI MA/01/23.