Fusarium oxysporum is a ubiquitous vascular wilt plant pathogen causing severe yield losses in a wide range of economically important crops. In this study, the interaction between Fusarium oxysporum f. sp. raphani and Arabidopsis thaliana plants impaired in the salicylate (SA), jasmonate (JA) and ethylene (ET) defence signalling pathways was investigated to better understand the nature of this plant–microbe interaction. The in planta bioassays revealed a key role for the ETR1 receptor as the etr1-1 mutant plants exhibited statistically less Fusarium wilt symptoms compared to the other mutant and Col-0 plants. Quantitative polymerase chain reaction (qPCR) analysis associated the decrease in symptom severity shown in etr1-1 plants with reduced vascular growth of the pathogen, suggesting the activation of defence mechanisms in etr1-1 plants against F. oxysporum. Furthermore, the early activation and increased accumulation of the SA-responsive PR1, PR2 and PR5 genes in the etr1-1 plants, in contrast to the Col-0 plants that showed higher transcript levels of the JA/ET-responsive PR3, PR4 and PDF1.2 genes after F. oxysporum inoculation, can lead to speculation that F. oxysporum hijacks ETR1-mediated ethylene signalling to promote disease development in plants.
Fusarium oxysporum is a ubiquitous soilborne pathogen that causes vascular wilt in more than 100 species of plants and is amongst the top 10 most important scientific/economic plant pathogenic fungi around the world, according to a recent survey (Dean et al., 2012). Disease symptoms include vascular browning, leaf epinasty, stunting, progressive wilting, defoliation and plant death (Agrios, 2005). Affected plants are mostly from the tropical and subtropical areas, probably because wilt symptoms are more pronounced at elevated temperatures (Di Pietro et al., 2003; Berrocal-Lobo & Molina, 2008). Because F. oxysporum grows better in warmer conditions, there is growing concern about the possibility that global warming might positively influence its incidence (Berrocal-Lobo & Molina, 2008). The control of F. oxysporum is especially difficult because the fungus survives for several years in the soil as resting structures, chlamydospores. In addition, resistance appears to be genetically complex and thus is a difficult trait to confer by breeding (Berrocal-Lobo & Molina, 2008). Therefore, it is quite important to investigate the molecular and genetic bases of plant innate immunity against one of the most devastating plant pathogens worldwide.
Upon pathogen recognition by plants, several signal transduction pathways are activated, influencing the disease outcome. The role of the signalling pathways mediated by salicylate (SA), jasmonate (JA) and ethylene (ET) in the innate immune response of Arabidopsis thaliana is well established (Glazebrook, 2005). These pathways can act either synergistically or antagonistically. It is known that SA and JA are mutually inhibitory for the expression of many genes (Berrocal-Lobo & Molina, 2008), while induced expression of some genes requires both ET and JA or only one of these signals. In addition, there are also cases of negative interaction between ET and JA signalling (Glazebrook, 2005).
Studies of the Arabidopsis–F. oxysporum interaction indicate that both SA- and JA-dependent defence signalling pathways are important for an effective defence response against F. oxysporum (Edgar et al., 2006). Increased resistance to F. oxysporum in transgenic or mutant plants has been associated with increased JA-responsive gene expression, such as the overexpression of defence genes PDF1.2 and CHIB, or transcription factors ERF1 and ERF2 that activate JA-responsive defence gene expression (Anderson et al., 2004; Berrocal-Lobo & Molina, 2004; Lorenzo et al., 2004; McGrath et al., 2005; Dombrecht et al., 2007). On the other hand, despite compromised JA-dependent defence responses, the JA perception mutant coronatine insensitive 1 (coi1) exhibits a high level of resistance to wilt disease caused by F. oxysporum (Thatcher et al., 2009). This response was found to be independent of SA-dependent defences, as coi1/NahG plants showed similar disease resistance to coi1 plants. However, other studies support a role for SA in defence against F. oxysporum because exogenously applied SA provides increased resistance to the pathogen (Edgar et al., 2006), and plants defective in SA accumulation/biosynthesis (such as NahG or eds5 plants) show increased susceptibility to F. oxysporum (Berrocal-Lobo & Molina, 2004, 2008; Diener & Ausubel, 2005; Dombrecht et al., 2007).
In view of the above, the objectives of this study were: (i) to clarify the role of SA, JA and ET in the defence of Arabidopsis against F. oxysporum f. sp. raphani using mutants in the corresponding signalling pathways, (ii) to determine the correlation between disease severity and degree of fungal proliferation in the vascular tissues by real-time qPCR, and (iii) to monitor the expression level of several defence genes involved in the SA and ET/JA response pathways in resistant and susceptible F. oxysporum f. sp. raphani–A. thaliana mutant interactions.
Materials and methods
Origin of seeds
Seeds of A. thaliana ecotype Columbia (Col-0) and the transgenic line NahG were provided by Syngenta; Col-0 mutants etr1-1, ein2-1, ein3-1, ein4, ein5-1, jar1-1, npr1-1, pad3-1 and pad4-1 were obtained from the Nottingham Arabidopsis Stock Centre, and eds5/sid1 and sid2 (Nawrath & Métraux, 1999) were provided by J. P. Métraux (University of Fribourg, Switzerland). All seeds were stored at 4°C. Seeds were sown directly into 6-cm-diameter pots, each containing approximately 200 cm3 soil (Potground; Klasmann). The pots were placed at 25°C with a 14 h photoperiod at 50–60% relative humidity in a controlled environment growth chamber. The plants were watered and fed with a nutrient solution (XL 60, Hortifeeds) as needed.
Fungal strains and inoculum preparation
The F. oxysporum f. sp. raphani WCS600 isolate (provided by C. M. J. Pieterse; Utrecht University, The Netherlands), with known pathogenicity against A. thaliana plants (Leeman et al., 1995), was used in the experiments. The fungal isolate was cryopreserved by freezing a suspension of conidia in 25% aqueous glycerol at –80°C. Before being used, fungus was transferred to potato dextrose agar (Merck) at 28°C for 3 days. For the bioassays, a suspension of 107 conidia mL−1 in sterile distilled water (SDW) was prepared from a culture grown for 3 days at 28°C in sucrose sodium nitrate liquid medium (Sinha & Wood, 1968).
Fusarium oxysporum f. sp. raphani WCS600–Arabidopsis bioassays
Fusarium oxysporum was applied as a soil drench inoculum because it is a soilborne pathogen that senses the presence of the plant through elaborate signalling mechanisms that lead to host recognition, root penetration, breakdown of host defences, proliferation within the host tissue and establishment of disease (Di Pietro et al., 2003). For this purpose, 3-week-old plants were inoculated with F. oxysporum by soil drenching with 10 mL of 107 conidia mL−1. Control plants were mock-inoculated with 10 mL SDW. On the day of inoculation, the soil was semidry in order to absorb the drenched inoculum.
Disease severity at each observation was calculated from the number of leaves that showed fusarium wilt symptoms (Anderson et al., 2004) as a percentage of the total number of leaves of each plant, and was recorded periodically for 30 days post-inoculation (dpi). Disease ratings were plotted over time to generate disease progression curves. The area under the disease progress curve (AUDPC) was calculated by the trapezoidal integration method (Campbell & Madden, 1990). Disease severity was expressed as a percentage of the maximum possible AUDPC for the whole period of the experiment and is referred to as relative AUDPC. The experiment was repeated three times with 15 replicates per experiment.
DNA extraction and qPCR fungal quantification
Ten plants from each treatment (Col-0 and etr1-1 plants infested with F. oxysporum) and replication (three replications, a total of 30 plants per treatment) were harvested for real-time quantitative PCR (qPCR) analysis at 6-day intervals from 6 dpi to 18 dpi. For each sampled plant, the above-ground parts (leaves and stem tissues) were cut at soil level, rinsed with SDW and ground to a fine powder using an autoclaved mortar and pestle in the presence of liquid nitrogen. Total DNA was isolated according to Dellaporta et al. (1983) and was quantified by spectrophotometry and agarose gel electrophoresis. qPCR assays for the quantification of F. oxysporum were conducted using the primer pair PG5-F and PG5-R (Table 1), designed for the endopolygalacturonase PG5 gene (AB256876) of F. oxysporum. QuantiFast SYBR Green PCR (QIAGEN) master mix was used for the amplification reactions. qPCR was performed in a Stratagene Mx3005P thermocycler. The results were analysed with MxPro qPCR software. For sample equilibration, the A. thaliana α2-tubulin gene (M84696) was targeted using the primer pair TUBα-F/TUBα-R (Table 1; Pantelides et al., 2010). PCR cycling started with an initial step of denaturation at 95°C for 5 min, followed by 40 cycles of 95°C for 10 s and 60°C for 30 s. The primer specificity and formation of primer dimers were monitored by dissociation curve analysis. All qPCRs were performed in duplicate. The absence of non-specific products and primer dimers was confirmed by the analysis of melting curves. To quantify the DNA levels of F. oxysporum, the PCR product of the primer pair PG5-F/PG5-R was cloned into pGEM (Promega) and seven 10-fold dilutions of the plasmid, ranging from 3 × 109 to 3 × 103 copies, were used to generate a standard curve (r2 = 0·99).
RNA isolation and qPCR determination of transcript levels
In many studies, resistance against F. oxysporum appears to correlate with reduced vascular colonization and overexpression of defence genes in the stem rather than root tissues (Edgar et al., 2006; Thatcher et al., 2009; Trusov et al., 2009). Therefore, in the present study, the above-ground parts (leaves and stem tissues) from the pathogen- and mock-inoculated Col-0 and etr1-1 plants were collected for RNA analysis. Samples were collected from 10 randomly selected plants per treatment (Col-0 and etr1-1 plants either inoculated with F. oxysporum or mock-inoculated) and replication (three replications, a total of 30 plants per treatment) and were immediately frozen in liquid nitrogen and stored at –80°C. For each sample, total RNA was isolated according to Brusslan & Tobin (1992). The RNA samples were treated with DNase I (Invitrogen) to eliminate traces of contaminating genomic DNA. The RNA concentration was measured on a NanoDrop ND-1000 spectrophotometer (Saveen Werner). First-strand cDNA was synthesized using SuperScript II (Invitrogen) following the manufacturer's procedure. Primers used for qPCR are listed in Table 1. The PCR efficiency for each amplicon was calculated by employing the linear regression method on log (fluorescence) per cycle number data, using Lin-RegPCR software (Ramakers et al., 2003). The qPCRs were performed in duplicate. The absence of non-specific products and primer dimers was confirmed by the analysis of melting curves. The A. thaliana a2-tubulin gene (M84696) was used as an internal standard to normalize small differences in cDNA template amounts. For data analysis, average threshold cycle (Ct) values were calculated for each gene of interest on the basis of three independent biological samples.
Data on relative AUDPC were transformed with the √ (x +1) transformation before analysis of variance (anova) was applied. When a significant (P ≤0·05) F-test was obtained for treatments, data were subjected to means separation by Tukey's multiple range test. Data on relative F. oxysporum DNA quantification and relative gene expression were analysed by carrying out a two-sample t-test (P ≤0·05) at each sampling date.
Impaired perception of ET via ETR1 reduces fusarium wilt symptoms
The role of SA, JA and ET signalling in defence against F. oxysporum was assessed in A. thaliana mutants impaired in these pathways by recording the fusarium wilt symptoms.
The first disease symptoms appeared in the form of wilting in all genotypes at 5 dpi. Disease severity progressed rapidly in all genotypes except etr1-1, which showed less prominent symptoms and slower disease development (Figs 1a & 2). Foliar symptoms were recorded for the next 25 days, until 30 dpi. On that day, disease severity in etr1-1 plants was 16%; in the other genotypes, the disease severity ranged from 40 to 72% (Fig. 1a). Consequently, the relative AUDPC analysis revealed that etr1-1 plants exhibited statistically less fusarium wilt symptoms compared to the other mutant and Col-0 plants (Fig. 1b). It is worth mentioning that the relative AUDPC value of the etr1-1 plants is six times less than that of the Col-0 plants. Therefore, the pathogenicity tests revealed that the ETR1 receptor has a key role in plant defence against F. oxysporum, whereas the other SA, JA and ET mutants did not affect the disease resistance of Arabidopsis plants (Fig. 1).
qPCR Fusarium oxysporum quantification reveals reduced colonization in etr1-1 plants
To determine whether the reduced disease severity of etr1-1 mutants is associated with a reduction in the amount of the pathogen in the vascular tissues of the plants, Col-0 and etr1-1 plants were inoculated with F. oxysporum, and the level of vascular fungal colonization was assessed by real-time qPCR.
qPCR analysis showed that, by 6 dpi, F. oxysporum was present in the vascular tissues of both Col-0 and etr1-1 plants (Fig. 3); however, the amount of F. oxysporum in Col-0 plants was seven times greater than in etr1-1 mutants. Six days later at 12 dpi, the amount of F. oxysporum in the Col-0 and etr1-1 plants had declined 1·6- and 1·3-fold, respectively. At 12 dpi, the relative amount of pathogen DNA was six times higher in Col-0 than etr1-1 plants (Fig. 3). At 18 dpi, there was no difference in the amount of the pathogen in both genotypes as a result of further decline of the relative DNA level of the pathogen in the Col-0 plants and a steady relative DNA level of the fungus in the etr1-1 plants (Fig. 3).
Impaired perception of ET via ETR1 causes transcriptional changes in response to Fusarium oxysporum infection
The data from the pathogenicity and fungal quantification experiments indicated that etr1-1 plants showed increased resistance to F. oxysporum. The expression levels of genes involved in the SA-, ET- and JA-associated defence mechanisms were therefore monitored in etr1-1 and Col-0 plants to elucidate potential components of basal resistance. Overexpression of the transcript levels of all examined genes (PR1, PR2, CHI/PR3, PR4, PR5, PDF1.2, AtMYC2, ERF2 and NPR1) was observed in both genotypes in response to F. oxysporum inoculation, indicating the activation of plant defence mechanisms upon fungal attack (Fig. 4a). It is notable that the SA-related defence genes PR1, PR2, PR5 and NPR1 were expressed at a higher level in etr1-1 than Col-0 plants (Fig. 4b), while the reciprocal observation was made for the ET/JA-linked defence genes CHI/PR3, PR4 and PDF1.2 that showed higher transcript levels in Col-0 than etr1-1 plants (Fig. 4c). In the etr1-1 plants, the transcript levels of all the examined genes, except PR1 and PR5, increased with time (Fig. 4a). The PR1 gene was overexpressed at 2 dpi and then its expression level declined, while the expression level of PR5 remained constant throughout the experimental time period (Fig. 4a). Interestingly, in the etr1-1 plants PDF1.2, which is part of the ET/JA-triggered defence mechanism, was underexpressed at 2 dpi and was overexpressed at the later stages of the infection process (Fig. 4a). The estimation of the etr1-1/Col-0 ratio of the transcript levels of the examined genes revealed that the highest ratio in the expression levels of the PR1, PR2 and PR5 genes was attained at 2 dpi, early in the infection process (Fig. 4b). On the other hand, the relative expression levels (etr1-1/Col-0 ratio) over time of the transcription factors NPR1 and AtMYC2 did not show as high fluctuations as PR1, PR2 and PR5 (Fig. 4b).
In the Col-0 plants, the transcript levels of all the examined genes increased with time, with the PR1 gene showing the lowest levels of overexpression (Fig. 4a). The estimation of the Col-0/etr1-1 ratio of the transcript levels of the examined genes showed that the most prominent difference between the two genotypes was in the transcript levels of PDF1.2 and ERF2 that exhibited the highest ratio at 2 dpi and then declined (Fig. 4c). In contrast, the highest Col-0/etr1-1 ratio of the expression levels of PR3 and PR4 was attained at 6 dpi, remaining at the same level at 12 dpi (Fig. 4c).
The wilt-inducing strains of F. oxysporum are responsible for severe damage on many economically important plant species (Fravel et al., 2003). The most cost-effective, environmentally safe method of control is the use of resistant cultivars. In the absence of resistant cultivars, broad-spectrum biocides, particularly methyl bromide, are used to fumigate soil before planting, with the drawback that such treatments are environmentally damaging (Fravel et al., 2003). Therefore, it is of great importance to develop other means of control based on the natural potential of plants to resist pathogens. To this end, a better understanding of the plant defence responses to pathogens such as F. oxysporum strains is required.
In the present study, the resistance of A. thaliana to F. oxysporum f. sp. raphani was determined in mutants of the SA, JA or ET defence signalling pathways. The AUDPC analysis showed that plants mutated in the ETR1 receptor were six times less infected than the wildtype Col-0 plants, whereas other ET mutants (ein2, ein5, ein4) along with SA (sid2, eds5, NahG) and JA (jar1-1) mutants were as susceptible as the wildtype plants. This result is in agreement with the findings of Geraats et al. (2003) who also observed a smaller percentage of diseased etr1-1 mutants compared to wildtype plants upon F. oxysporum infection. Likewise, Lund et al. (1998) have observed that the ET-insensitive never-ripe mutant of tomato displays increased tolerance to the vascular wilt pathogen F. oxysporum f. sp. lycopersici. It has also been shown that etr1-1 mutant Arabidopsis plants are resistant to the vascular pathogens Verticillium dahliae and Clavibacter michiganensis subsp. michiganensis (Balaji et al., 2008; Pantelides et al., 2010).
It is well established that the ET and JA signalling pathways synergistically regulate defence against necrotrophic plant pathogens (Glazebrook, 2005). Even if F. oxysporum, V. dahliae and C. michiganensis subsp. michiganensis are considered as necrotrophs, the etr1-1 mutation resulted in increased resistance instead of susceptibility. An explanation for this discrepancy may lie in the fact that these are vascular wilt pathogens which are present in the xylem vessels where they interfere with the host's translocation of water and nutrients, only invading and destroying surrounding parenchyma cells at later stages. In contrast, necrotrophs generally invade and kill cells of the epidermis and cortex (Geraats et al., 2003; Agrios, 2005). Therefore, vascular wilt pathogens might be better understood by considering them as hemibiotrophs.
Interestingly, the other ET mutants, ein4, ein2-1, ein3-1 and ein5-1, were not resistant to F. oxysporum. This can be explained by the role and position of these genes in the ET signalling pathway. The ET receptors are subdivided into two subfamilies: ETR1 is a member of the subfamily 1 and EIN4 is a member of the subfamily 2 (Guo & Ecker, 2004), which might explain the different responses of etr1-1 and ein4 mutants to F. oxysporum. Furthermore, the specific expression and localization of these ET receptors is not known in detail in relation to the position of the fungus. The other genes, EIN2-1, EIN3-1 and EIN5-1, are located downstream of ET receptors and act as positive regulators of ET responses (Guo & Ecker, 2004).
The qPCR monitoring of the F. oxysporum infection process in the Col-0 and etr1-1 plants revealed that the retarded wilt symptoms in the etr1-1 plants reflect reduced pathogen amounts in the vascular tissues compared to the Col-0 plants. This observation implies an early activation of plant defence mechanisms in etr1-1 compared to Col-0 plants. In contrast, Thatcher et al. (2009) observed that there was no difference in the degree of F. oxysporum colonization of the wild type and resistant coi1 plants until later stages of infection, when host necrosis was well developed. Therefore, it was suggested that the biotrophic phase of the hemibiotrophic infection process of F. oxysporum was not affected in coi1, but that the necrotrophic phase, when host cell death and lesion development occurs, does require a functional COI1 gene in the host (Thatcher et al., 2009). It is clear that etr1-1 and coi1 interfere with F. oxysporum in a different manner: coi1 plants appear to be tolerant to the pathogen at the early stages of infection, whereas a functional ETR1 receptor is needed by the fungus to overcome the defence mechanisms that seem to be deployed in resistant etr1-1 plants, according to the present data.
In addition, a lack of positive correlation between pathogen growth within plant tissues of Col-0 plants and symptom severity was observed (Figs 1a & 3). This discrepancy has been noted in several plant–pathogen interactions (Lund et al., 1998; Cecchini et al., 2002; Veronese et al., 2003). A possible explanation of this interesting decoupling could be that wilting is most probably caused by a combination of pathogen activities, such as toxin production and host defence responses (Agrios, 2005), that cause changes in normal plant growth and development, such as accelerated flowering, senescence and programmed cell death.
Following challenge by F. oxysporum, the wildtype Col-0 plants and the etr1-1 mutant plants exhibited elevated transcript levels of all the examined genes, compared to the mock-inoculated plants, except for PDF1.2 in etr1-1 plants at 2 dpi. The JA-responsive defence gene PDF1.2 has been long associated with plant resistance against F. oxysporum (Anderson et al., 2004; McGrath et al., 2005; Edgar et al., 2006); however, in the present study the PDF1.2 gene was underexpressed in the resistant etr1-1 plants at 2 dpi, compared to the mock-inoculated etr1-1, and it was also less expressed in the inoculated etr1-1 than in Col-0 at all time points. Likewise, Thatcher et al. (2009), recorded a down-regulation of the PDF1.2 gene in coi1 plants resistant to F. oxysporum compared to the wildtype Col-0 plants, at 4 dpi. In the present study, it was observed that etr1-1 plants had elevated expression of the SA-mediated defence genes PR1, PR2 and PR5, and attenuated expression of the ET/JA-associated genes PR3, PR4 and PDF1.2 compared to the Col-0 plants upon F. oxysporum infection.
In parallel, the ERF2 transcription factor that positively regulates the expression of the JA-responsive genes PDF1.2, PR3 and PR4 (Brown et al., 2003) was expressed at higher levels in Col-0 than in etr1-1 plants. On the other hand, the AtMYC2 gene has been identified as a negative regulator of its own transcription and of JA-dependent plant defences (Anderson et al., 2004; Lorenzo et al., 2004; Dombrecht et al., 2007). In agreement with these observations, etr1-1 plants that exhibited a higher expression of AtMYC2 showed a decreased expression of PR3, PR4 and PDF1.2 (Fig. 4b,c). Together, these results indicate that JA-dependent genes such as PDF1.2 are unlikely to contribute to the resistance observed in etr1-1 plants.
Previous research has shown that F. oxysporum is sensitive to SA-dependent defences (Edgar et al., 2006; Trusov et al., 2009). Indeed, exogenous SA application activated PR1 and PR2 genes in the leaves of Col-0 plants, but not in roots, and provided increased F. oxysporum resistance (Edgar et al., 2006). Furthermore, the PR-regulator gene NPR1 was overexpressed in etr1-1 compared to Col-0 plants, in agreement with the elevated levels of PR1, PR2 and PR5 genes in etr1-1. It has been suggested that NPR1 is part of the crosstalk control between the SA and JA signalling pathways (Spoel et al., 2003). The higher transcript levels of NPR1 in etr1-1 compared to wildtype plants may support the idea that crosstalk between the SA and JA pathways via NPR1 plays an important role in fine-tuning the plant defence response to Fusarium infection, as has been described for Pseudomonas syringae pv. tomato (Pst) infection (Zhao et al., 2003). In tomato, SA-mediated defence responses are more effective against Pst DC3000 infection than JA-mediated responses and Zhao et al. (2003) have reported that Pst DC3000 uses the type-III secretion system and the phytotoxin coronatine to activate the JA signalling pathway and to repress pathogenesis - related proteins (PRs) to promote disease.
Pathogen recognition is the first and crucial step in plant defence (Glazebrook, 2005). In this respect, induction of the ‘wrong’ defence mechanism often results in higher pathogen proliferation and in the subsequent death of the plant (Anderson et al., 2004; Takahashi et al., 2004; Asselbergh et al., 2008). The data from the present study pinpoint the key role of the ethylene receptor ETR1 in F. oxysporum interaction with plants. It is tempting to speculate that F. oxysporum hijacks ETR1-mediated ET signalling to promote disease development in plants, as it is also known that F. oxysporum produces ET in vitro (Tzeng & DeVay, 1984), by avoiding the early and high accumulation of the potentially effective SA-dependent defence genes.