Background

Verticillium species are wide-spread soil-borne fungi which cause vascular diseases in many plant species and are responsible for devastating diseases for plants that can thwart agricultural production. The vascular wilt fungus Verticillium dahliae (Vd), for instance, infects more than 200 plant species, among them agriculturally and horticulturally important crops and ornamental plants [1]-[3]. It is estimated that Vd infections are responsible for several billions of dollars of annual crop losses worldwide. Vd has a broad host range and infects plants from temperate to subtropical climates [1]. Because of their complex life style of the Verticillium species, their control by classical pesticides or fungicides is difficult; therefore, the isolation of Verticillium-resistant cultivars is an important task for the breeders (cf. [4],[5]).

Genetic resistance against Verticillium wilt diseases has been reported for several plant species [1],[2]. The Ve gene provides resistance against race 1 isolates of Vd in tomato [6],[7] and the tomato gene is also functional after expression in Arabidopsis [8]. Many studies have used Arabidopsis for the isolation of Vd-resistant germplasm [9],[10] or the identification of novel resistance traits following mutagenesis [2],[10]-[14]. Furthermore, quite recently, a large number of proteins and metabolites from different organisms as well as phytohormones have been described to be involved in establishing partial resistance against Verticillium wilt [15]-[22].

Like other Verticillium species, Vd can overwinter as mycelium in host plants or soil. The fungus can also form seed-like structures called microsclerotia, long-lived survival structures of clusters of melanized cells with thick walls, which survive in the soil without a host plant or in association with plant material for up to 20 years [23],[24]. The microsclerotia germinate in response to stimuli from root exudates [25]. The hyphae penetrate and grow inter- and intracellularly through the root cortex toward the central cylinder of the root [26],[27]. They enter the xylem cells of the root, from where they colonize the xylem of the hypocotyl and leaves. Ultimately, the water transport is disrupted which results in the wilt phenotype [1]-[3]. Verticillium species are considered as hemibiotroph: a biotrophic phase within root xylem without a visible disease phenotype is followed by a necrotrophic phase in the aerial parts of the plant.

The spread of the pathogen occurs primarily by root infections from the soil. Therefore rhizosphere bacterial strains such as Pseudomonas putida B E2, Pseudomonas chlororaphis K15 or Serratia plymuthica R12 [28] or bacterial isolates [29] have been shown to function as efficient biocontrol agents against Vd spread. The microbial bioagents induce antibiosis, parasitism, competition and secretion of enzymes such as glucose oxidase, chitinase and glucanase which results in the induction of disease resistance in the hosts [12],[30].

To our knowledge, there is no report on endophytic fungi which can be used as biocontrol agent against Vd in Arabidopsis. Piriformospora indica (Pi), a cultivable basidiomycete of Sebacinales, colonizes the roots of many plant species including Arabidopsis [31],[32]. Like other members of Sebacinales, Pi is found worldwide in association with roots [33] and stimulates growth, biomass and seed production of the hosts [31],[34]-[36]. The fungus promotes nitrate and phosphate uptake and metabolism [35],[37]. Pi also confers resistance against abiotic [38],[39] and biotic stress [40].

Here, we demonstrate that Pi is an efficient biocontrol agent that protects Arabidopsis from Vd infection in vitro and in vivo by inhibiting growth of Vd in roots. Furthermore, we give evidence that a Vd-exudate compound induces cytoplasmic Ca2+ ([Ca2+]cyt) elevation and the Vd-disease development is dependent on the ethylene-activated transcription factor EIN3.

Results

Pi inhibits growth of Vd on PDA agar plates

Pi and Vd were co-cultivated as described in Methods on a PDA agar plate for 3 weeks. Figure 1(A and B) demonstrates that Pi strongly inhibits growth of Vd hyphae. The Vd colony in the dual culture is significantly smaller than the Vd colony growing without Pi. Furthermore, the number of microsclerotia produced by Vd in the dual culture is less than the number of microsclerotia produced by Vd growing alone. No obvious inhibition zone can be detected. In contrast, growth of Pi is barely affected by the presence of Vd. This prompted us to test the role of Pi in protecting Arabidopsis plants against Vd infection.

Figure 1
figure 1

Pi inhibits growth of Vd on agar plates. (A) Typical plates from 3 independent experiments are shown. (B) Quantification of the colony. The diameter of the Pi and Vd mycelia on the agar plate is given in cm. Bars represent SDs. Asterisks indicate significant differences, as determined by ANOVA (*** P ≤ 0.001).

Full size image

Arabidopsis seedlings pretreated with Pi are protected against Vd infection

To investigate whether Pi can protect Arabidopsis for Vd infection, we exposed the seedlings first to Pi prior to Vd infection. Seedlings not exposed to any of the two fungi or to one of the two fungi alone served as controls (cf. Methods). The performance of the seedlings was measured after 10, 14 and 21 days, by visible inspection and measuring the fresh weights. After 10 days of co-cultivation, seedlings treated with Vd or Pi alone showed ~30% increase in the biomass compared to the untreated control seedlings. A comparable increase in the biomass was observed when the seedlings were first exposed to Pi and then to Vd or vice versa (Figure 2A). This slight increase in the biomass suggests that both fungi initially form a beneficial interaction with the seedlings, and is consistent with the idea that this phase represents a biotrophic interaction of Vd with Arabidopsis roots. On the 14th day, seedlings infected by Vd alone or first with Vd followed by Pi (1V2P) showed obviously the disease symptoms. The leaves of these seedlings became paler and the roots browner compared to the seedlings exposed to Pi or 1P2V treatments, although no significant differences in the biomass were observed for the different fungal treatments, except for Pi treatment (Figure 2A). In contrast, on the 21st day, seedlings exposed to Vd alone or exposed to Vd prior to exposure to Pi (1V2P) were severely damaged. Their fresh weights were reduced or no longer measurable. Pi treatment alone resulted in a ~30% increase in the fresh weight (Figure 2A). Interestingly, seedlings which were pretreated with Pi and then exposed to Vd (1P2V) had the same fresh weights as untreated control seedlings, although the visible inspection showed some photo-bleaching (Figure 2B). This clearly demonstrates that Pi protects Arabidopsis seedlings against Vd-induced wilt. Therefore, this experimental set-up was used to study the protective function of Pi in greater details.

Figure 2
figure 2

Pi protects Arabidopsis seedlings from Vd infection. (A) Fresh weights of seedlings after 10, 14 and 21 days of co-cultivation or mock-treatments on Petri dishes. The seedlings were exposed to either Pi or Vd alone or in combination as described in the Methods and Additional file 1: Figure S1. C: seedlings treated without fungi; Vd: seedlings treated with Vd; Pi: seedlings treated with Pi; 1P2V: seedlings first treated with Pi for 4 days followed by Vd; 1V2P: vice-versa as 1P2V. n.d: no detectable (seedlings were dead, no fresh weight could be determined). The data are based on 3 independent experiments with 16 seedlings each. Bars represent SDs. Asterisks indicate significant differences, as determined by ANOVA (* P ≤ 0.05; ** P ≤ 0.01; *** P ≤ 0.001). (B) The phenotype of typical seedlings on 21st day. (C) PDI for seedlings exposed to Vd. For treatments, cf. Methods and Additional file 1: Figure S1. The data are based on 3 independent experiments with 16 seedlings each. Bars represent SDs.

Full size image

The results were confirmed by calculating the Percentage Disease Index (PDI) for those seedlings treated with Vd. After 10 days of co-cultivation, the PDI for Vd and 1V2P seedlings was ~20%, and after 14 days 40-50%. After 21 days, the PDI was almost 100%. In contrast, seedlings pretreated with Pi prior to exposure to Vd (1P2V) showed a slow increase in the PDI, which reached ~30% after 21 days (Figure 2C).

Furthermore, the amount of total chlorophyll (Chl) is a sensitive marker for the fitness of a plant. On the 4th day, the shoots of Vd- and Pi- treated plants contained slightly higher Chl levels than control seedlings (Figure 3). On the 10th day, the Chl content of Vd- treated seedlings is comparable to that of control seedlings not exposed to the pathogen. Furthermore, while 1P2V seedlings had the same amount of Chl as Pi seedlings, the Chl content in 1V2P seedlings was significantly reduced (Figure 3). Comparable results were obtained for the 14th day, except that the Chl content for 1P2V seedlings was reduced compared to Pi seedlings (Figure 3). On the 21st day, Pi seedlings had the highest Chl content, 1P2V seedlings had the same amount of Chl as control seedlings not exposed to a fungus, while the Chl levels in the Vd and 1V2P plants were strongly decreased (Figure 3). This confirms the protective function of Pi against Vd infection in Arabidopsis leaves.

Figure 3
figure 3

Total chlorophyll content (mg/g fresh weight) in shoots. The data were obtained 4, 10, 14 and 21 days after the fungal treatments (cf. Methods, Additional file 1: Figure S1 and legend to Figure 2A). The data are based on 3 independent experiments with 16 seedlings each. Bars represent SDs. Asterisks indicate significant differences to the untreated control, as determined by Student's t-test (* P ≤ 0.05; ** P ≤ 0.01; *** P ≤ 0.001).

Full size image

Pathogenesis and application of pathogen-associated molecular patterns induce stomata closure [41]. In control plants not exposed to any fungus, between 5 and 12% of the stomata were closed. Three days after exposure of the roots to Vd, ~25% of the stomata were closed (Figure 4A), and this increased to ~30% until the 7th day. The 1V2P treatment showed ~25% stomata closure at the 7th day, and this value is comparable to that for seedlings treated with Vd alone. In contrast, exposure of the roots to Pi or first to Pi followed by Vd did not result in stomata closure and these values are comparable to those of the untreated controls (Figure 4B). This indicates that Pi prevents Vd-induced stomata closure. These results demonstrate that stomatal closure correlates nicely with the amount of total chlorophyll.

Figure 4
figure 4

Stomata closure rate in leaves after 3 (A) and 7 (B) days. The data are based on 3 independent experiments with 16 seedlings each. Bars represent SDs. Asterisks indicate significant differences to the untreated control, as determined by Student's t-test (* P ≤ 0.05; ** P ≤ 0.01; *** P ≤ 0.001).

Full size image

Pi represses Vd-induced genes in shoots

Vd induces defense gene expression in shoots. After 1 d, the mRNA levels for PR1 and PR2 representing SA-inducible genes and PDF1.2 for the JA/ET pathway, ERF1 and VSP2 for ET pathway were upregulated in the leaves of Vd-exposed seedlings. Except for PR2, none of the other genes responded to Pi exposure (Figure 5). After 14 d, Vd-exposed seedlings showed an even stronger upregulation of the defense genes in the leaves (Figure 5). Pretreatment of the seedlings with Pi prior to Vd infection resulted in the repression of defense gene expression compared to seedlings which were not pretreated with Pi. This provides additional evidence for the protective function of Pi against Vd infection. Furthermore, plant glutamate receptor-like (GLR) genes, GLR2.4, GLR2.5 and GLR3.3 code for putative Ca2+ transporters and are involved in defense responses [42]-[44]. We observed that GLR2.4 (but not GLR2.5 and GLR3.3) was upregulated in the leaves of Vd-exposed seedlings and repressed in the leaves of seedlings which were pretreated with Pi prior to Vd exposure (Figure 5 and Additional file 1: Figure S2). RabGAP22 is required for defense to V. longisporum and contributes to stomata immunity [22]. For Vd, RabGAP11 is upregulated after exposure to Vd and significantly repressed in seedlings which were pretreated with Pi (Figure 5).

Figure 5
figure 5

Induction of defense genes in the shoots of Arabidopsis seedlings 1 and 14 days after the fungal treatments, relative to the untreated control. The data represents fold induction (mRNA level +fungal treatments/mRNA level -fungal treatments; fold of control is set as 1.0). For experimental details, cf. Methods, Additional file 1: Figure S1 and legend to Figure 2A. The data are based on 5 independent experiments with 16 seedlings each. Bars represent SDs. Asterisks indicate significant differences, as determined by Student's t-test (* P ≤ 0.05; **P ≤ 0.01; *** P ≤ 0.001).

Full size image

Pi strongly represses Vd-induced phytohormone accumulation in shoots

The phytohormones JA, JA-Ile, OPDA, SA, ABA and ET are crucial for the activation of defense responses. Figure 6 demonstrates that these phytohormones accumulated after Vd infection in the shoots of Arabidopsis seedlings. The phytohormone levels were also high in the 1V2P samples, while in all other cases [Control (C), Pi, 1P2V], they showed significantly lower levels. Thus, Vd-induced phytohormone accumulation is repressed if the roots are colonized by Pi prior to their exposure to Vd. Interestingly, application of Pi to roots which were already exposed to Vd did not repress the accumulation of the phytohormones in the shoots.

Figure 6
figure 6

Phytohormone levels in the shoots 21 days after the different fungal treatments. For experimental details, cf. Methods, Additional file 1: Figure S1, and legend to Figure 2A. The data are based on 3 independent experiments with 12 seedlings each. Bars represent SDs. Asterisks indicate significant differences, as determined by Student's t-test (* P ≤ 0.05; ** P ≤ 0.01; *** P ≤ 0.001).

Full size image

Pi inhibits Vd propagation and microsclerotia formation

Quantification of the amount of Vd DNA demonstrated that Vd and 1V2P seedlings contain twice as much pathogen DNA than 1P2V seedlings in both roots (Figure 7A and D) and shoots (Figure 7B and E). Interestingly, the amount of Pi DNA in the roots is identical in all Pi-treated samples and not affected by a pretreatment with Vd (Figure 7C and F). Furthermore, microscopic analysis demonstrated that the number of microsclerotia was strongly reduced in root tissue pretreated with Pi (Figure 8). This demonstrates that Pi inhibits Vd propagation and microsclerotia formation in the roots, while Vd does not affect the propagation of Pi in Arabidopsis roots.

Figure 7
figure 7

The amount of fungal DNA in the roots and shoots of Arabidopsis seedlings exposed to the 5 treatments (cf. legend to Figure2A). For experimental details, cf. Methods and Additional file 1: Figure S1. The measurements were performed for the 14th(A, B, C) and 21st (D, E, F) day. The data are based on 3 independent experiments with 12 seedlings each. Bars represent SDs. Asterisks indicate significant differences compared to Vd (A, B, D, E) or to Pi (C and F), as determined by Student's t-test (* P ≤ 0.05; ** P ≤ 0.01; *** P ≤ 0.001).

Full size image
Figure 8
figure 8

Pi inhibits the formation of Vd micosclerotia in roots, irrespective of whether the roots were first exposed to Pi ( 1P2V ) or first to Vd ( 1V2P ). The analysis was performed 21 days after infection. Left: microscopy of root sections with microslerotia (black spots). Right: Quantification of the number of microsclerotia. The data are based on 3 independent experiments with 12 seedlings each. Bars represent SDs. Asterisks indicate significant differences to Vd, as determined by Student's t-test (* P ≤ 0.05; ** P ≤ 0.01; *** P ≤ 0.001).

Full size image

Long-term experiments confirmed the results obtained for seedlings

In order to study long term interaction, the seedlings were grown according to the 5 regimes on Petri dishes for 10 days before transferred to sterile vermiculite for additional 14 days. All (C) seedlings and those exposed to Pi (Pi) were alive. Exposure of Pi-pretreated plants to Vd resulted in ~20% loss of the plants. However 80% of the plants, which were either exposed to Vd alone or first to Vd followed by Pi, died (Figure 9A). Furthermore, we measured the fresh weights of the seedlings which survived the treatments. Plants exposed to Pi alone showed a ~30% increase in the fresh weight. The fresh weights of 1P2V plants were comparable to those not exposed to any fungus. Vd- and 1V2P-treated seedlings showed significantly decreased fresh weights compared to all other treatments (Figure 9B). Finally, the Vd DNA amount in both shoots and roots was lower in 1P2V-treated plants compared to those treated with Vd alone or first with Vd followed by Pi (1V2P) (Figure 9C). Comparable to the results obtained with seedlings in Petri dishes (Figure 7), the Pi DNA content was the same in all Pi-treated roots (Figure 9C). This confirms that Pi inhibits Vd growth, but not vice versa.

Figure 9
figure 9

Confirmation of the results for adult plants, grown in sterile vermiculite. After exposure of the seedlings to the 5 treatments in Petri dishes for 10 days (cf. legend to Figure 2A), they were transferred to Magenta boxes with sterile vermiculite for 14 days. (A) Number of survived plants. (B) Fresh weight of plants. (C) Fungal DNA content in roots and shoots. The data are based on 3 independent experiments with 16 seedlings each. Bars represent SDs. Asterisks indicate significant differences to Vd, as determined by Student's t-test (* P ≤ 0.05; ** P ≤ 0.01; *** P ≤ 0.001).

Full size image

EIN3 is required for full susceptibility of Arabidopsis to Vd

The strong upregulation of the phytohormone levels in the leaves of seedlings grown in the presence of Vd was further investigated for ET. Pantelides et al. [11] have shown that ET perception via ETR1 is required for Vd infection in Arabidopsis. We observed a strong requirement of EIN3 for Vd-induced disease development in Arabidopsis leaves. ein3 seedlings which were exposed to Vd alone or were first treated with Vd before application of Pi perform better than wild-type seedlings (Figure 10A, B and Additional file 1: Figure S3). Interestingly, the ET level in ein3 seedlings is much higher than in wild-type seedlings, even in the absence of Vd. Exposure of the seedlings to Vd stimulate ET accumulation even further (Figure 10C and Additional file 1: Figure S4). This suggests that ein3 seedlings try to compensate the lack of EIN3-induced genes by further stimulating ET biosynthesis, in particular after Vd infection. Taken together, these data demonstrate that EIN3-induced genes are required for pathogenicity of Vd.

Figure 10
figure 10

EIN3 is required for full susceptibility of Arabidopsis to Vd . (A) The representative picture (3 independent experiments with 32 plants each) was taken after 21 days inoculation with Vd. (B) Number of survived seedlings. (C) Ethylene levels in WT and ein3 seedlings after exposure to Vd. Bars represent SDs. Asterisks indicate significant differences, as determined by Student's t-test (* P ≤ 0.05; ** P ≤ 0.01; *** P ≤ 0.001).

Full size image

Vd induces [Ca2+]cyt elevation in WT roots, but not in roots of a Ca2+ response mutant

Pathogen-associated molecular pattern-triggered immunity is often initiated by [Ca2+]cyt elevation, which can be induced by exudated compounds from pathogenic fungi [cf. [45] and ref. therein]. Since the putative plasma membrane-localized Ca2+-transporter gene GLR2.4 was upregulated by Vd, we tested whether exudated compounds from Vd can induce [Ca2+]cyt elevation in roots. An exudate preparation from the mycelium was applied to the roots of transgenic pMAQ2 Arabidopsis lines expressing the Ca2+-sensor apoaequorin. Under resting conditions, 21 d-old pMAQ2 lines gave [Ca2+]cyt values of 70 ± 0.6 nM (n = 16). A rapid and transient increase in the [Ca2+]cyt concentration is observed 40 sec after the application of Vd preparation (Figure 11A). Discharge at the end of the experiment demonstrates that less than 5% of the reconstituted aequorin was consumed after the stimuli, which ensures that the amount of aequorin in the sample is not limiting for the Ca2+ signal (data not shown). The [Ca2+]cyt reached a peak of ~ 400 nM after 90 to 120 sec (Figure 11A). Subsequently the Ca2+ levels steadily decreased. No [Ca2+]cyt elevation is observed with the PBS buffer treatment (Figure 11A). The magnitude of the [Ca2+]cyt response is dose-dependent (data not shown). Furthermore, an Arabidopsis cytoplasmic calcium elevation mutant1 (cycam1) which does not show [Ca2+]cyt elevation in response to exudate preparation from various pathogenic fungi [45] also failed to induce [Ca2+]cyt elevation in response to the Vd preparation (Figure 11B). This indicates that cycam1 is impaired in the response to exudate preparations from various pathogens. Furthermore, we crossed the apoaeqorin gene into the glr2.4, glr2.5 and glr3.3 knock-out background. Figure 11B demonstrates that the Vd exudate preparation induced [Ca2+]cyt elevation in the knock-out backgrounds, indicating that these putative plasma membrane-localized transporters do not participate in the Ca2+ uptake from the extracellular space, although the gene GLR2.4 was upregulated in Vd-infected seedlings (Figure 5).

Figure 11
figure 11

Vd exudate preparation induces [Ca2+] cyt elevation in A. thaliana seedlings expressing cytosolic aequorin. (A) Roots of 21-day old pMAQ2 in Col-0 seedlings were dissected and incubated overnight in 7.5 μM coelenterezine. The roots were challenged with 50 μl of the Vd preparations. [Ca2+]cyt level was calculated from the relative light unit (RLU) at 5's integration time for 10 min. The arrow indicates the time (60 s) of addition of the stimuli/PBS buffer. For all experiments, 10 mM phosphate buffer (PBS, pH 7.0) was used as control and gave background readings. All curves and values represent average of five independent experiments with eight replications in each experiment. (B) Vd exudate preparation does not induce [Ca2+]cyt elevation in the cycam1 mutant, but induces [Ca2+]cyt elevation in pMAQ2 lines in the glr2.4, glr2.5 and glr3.3 background.

Full size image

To investigate whether [Ca2+]cyt elevation is required for disease development, cycam1 was infected with Vd and the development of the mutant seedlings was compared to that of the WT seedlings. No obvious difference of the disease symptoms in the aerial parts could be detected, which suggests that [Ca2+]cyt elevation is not essential for Vd propagation (Additional file 1: Figure S6).

Discussion

Our data demonstrate that Pi is a very efficient biocontrol agent for Vd wilt in Arabidopsis. Pi restricts Vd growth both on agar plates (Figure 1) and in Arabidopsis roots, in particular when they were first colonized by Pi prior to infection with Vd (Figure 7). Molecular and biochemical analyses demonstrate that the reduced growth rate of Vd in Pi-pretreated Arabidopsis roots retards defense gene expression (Figure 5), the accumulation of defense-related phytohormones (Figure 6) and stomata closure (Figure 4). The performance of the seedlings is significantly better (Figure 2) and this also continues after shifting the seedlings to vermiculite for a longer period of time (Figure 9). Pi not only inhibits growth of Vd mycelia in Arabidopsis roots, but also prevents the spread of the pathogen to the aerial parts of the plant (Figure 7). Furthermore, microsclerotia formation is strongly reduced (Figure 8). Previously, several soil-borne bacteria have been identified as biocontrol agents for Verticillium wilt [29],[46]-[48]. Vd can induce antimicrobial metabolites such as rutin in potato [49] or pathogenesis-related proteins in Arabidopsis [12] which participates in pathogen resistance. Prieto et al. [50] demonstrated that root hair colonization plays an important role in Pseudomonas spp.-mediated biocontrol activity against Verticillium wilt in olive roots. Furthermore, the Bacillus subtilis strain NCD-2 functions as a biocontrol agent against cotton Verticillium wilt, and the cotton PhoR/PhoP, two component regulatory systems, were involved in the biocontrol capability of the bacterium [51]. Also quorum sensing is involved in the biocontrol activity of Serratia plymuthica against Vd[52]. Moderate drought influences the effect of arbuscular mycorrhizal fungi as biocontrol agents against Verticillium-induced wilt in pepper [53]. It appears that quite different mechanisms control the fungal spread, probably because of the complicated lifestyle of the pathogen which allows microbial interference at different levels and in different plant tissues.

An increasing number of genes were recently identified to be involved in establishing partial resistance to Verticillium wilts (cf. Background). Pathogen attack including root colonization by Vd is associated with stomata closure as one of the first line of plant defense (Figure 4). RabGAP22 is required for defense against V. longisporum and contributes to stomatal immunity [20]. RabGAP11 gene is upregulated by Vd and repressed by Pi (Figure 5). Finally, defensins play a role in the plant defense against Vd[19].

Control of microsclerotia formation is crucial for preventing Verticillium spread in nature and agriculture. Our data demonstrate that Pi is quite efficient in restricting microsclerotia formation in Arabidopsis roots (Figure 8), presumable because the pathogen cannot grow fast enough in the presence of Pi. Microsclerotia formation is also suppressed by Verticillium itself, i.e. by the fungal transcription activator of adhesion Vta2, and fungi impaired in Vta2 are unable to colonize plants and induces disease symptoms [21]. Taken together, Pi restricts Vd growth as well as hyphal and microslerotia propagation, which - in turn - causes that the plant defense processes get activated at a lower level compared to Vd treatments which might depend on Pi-plant-Vd interaction-pattern and the attack strategy of Vd. This is not only important for better performance of individual plants, but has also severe long-term consequences for the control of the Vd spread via microsclerotia in ecosystems and agricultural areas.

GRL homologs are associated with Ca2+ influx through the plasma membrane. Figure 5 demonstrates that the mRNA level for GLR2.4 is upregulated in the leaves of Vd-infected Arabidopsis seedlings and these responses are restricted by a pretreatment of the seedlings with Pi. GLR3.3 is involved in plant defense and resistance to Hyaloperonospora arabidopsidis[44]. The protein also mediates glutathione-triggered [Ca2+]cyt transients, transcriptional changes, and innate immunity responses in Arabidopsis [54]. GLR2.5 is upregulated in Arabidopsis cell cultures upon wounding [43] and GLR2.4 is induced by nematodes in Arabidopsis roots [42]. GLR2.4, also called AUGMIN subunit 8, is a microtubule plus-end binding protein that promotes microtubule reorientation in hypocotyls [55],[56]. Microtubules and microtubule orientation are important for plant defense and immunity [56],[57] and also involved in Vd-Arabidopsis interaction. Hu et al. [18] demonstrated that histone H2B monoubiquitination is involved in regulating the dynamics of microtubules during the defense response to Vd toxins in Arabidopsis. Yuan et al. [58] showed that Vd toxins disrupted microfilaments and microtubules in Arabidopsis suspension-cultured cells. Figure 11A shows that exudate compounds from Vd induces [Ca2+]cyt elevation in Arabidopsis roots. In order to test whether the [Ca2+]cyt elevation is mediated by one of the three GLRs, we generated transgenic glr3.3, glr2.5 and glr2.4 knock-out lines in the apoaequorin background and found that the [Ca2+]cyt response is not controlled by the three GLRs (Figure 11B), although the mRNA level of GLR2.4 is upregulated upon Vd infection (Figure 5). This suggests that GLRs have different functions in the Vd-Arabidopsis interaction. However, an ethylmethansulfonate-induced Arabidopsis mutant named cycam1 which is unable to induce [Ca2+]cyt elevation in response to exudate preparations from Alternaria brassicae, Rhizoctonia solani, Phytophthora parasitica var. nicotianae and Agrobacterium tumefaciens[45] did not respond to the Vd exudate preparation (Figure 11B). This demonstrates that at least one of the Vd-induced signaling events leading the opening of Ca2+ channels or the channels themselves are identical to those responding to exudate preparations from other pathogens [45]. However, the reduced Ca2+ response in the cycam1 mutant does not affect the disease development. It remains to be determined which is the active compound inducing the [Ca2+]cyt response in Arabidopsis roots, and what is the mutated gene in the cycam1 mutant.

Several exudated compounds have been postulated to induce pathogenicity in plants. Klosterman et al. [3] proposed that based on the sequence information of Verticillium species, pathogenicity may be caused by a cocktail of different compounds and elicitors with different functions in the complex pathogenicity procedure. A Verticillium crude toxin preparation has been often used, although the exact composition of this preparation and the role of the individual compounds are not clear. For instance, recently Yao et al. [59] have demonstrated that the Vd toxin preparation stimulates nitric oxide production in Arabidopsis which serves as a signaling intermediate downstream of H2O2 to modulate dynamic microtubule cytoskeleton. This may link the Vd toxin function again to GLR2.4, who's mRNA level is upregulated after Vd infection (Figure 5). Wang et al. [60] reported on the purification and characterization of a novel hypersensitive-like response-inducible protein elicitor named PevD1 from Vd that induces resistance responses in tobacco. The relationship of the bioactive compound that induces the [Ca2+]cyt response to the toxins which induce disease responses needs to be investigated.

Interestingly, we did not observe a linear relationship between the propagation of Vd in the seedlings and the accumulation of defense-related phytohormone levels. For instance, the phytohormone levels were always high when the seedlings were exposed to Vd, irrespective of whether they were exposed to Vd alone, pretreated with Pi or first with Vd followed by Pi (Figure 6), although, growth of Vd was strongly reduced by the Pi pretreatment (Figure 1). This suggests that even low infection rates of Vd are sufficient to stimulate the accumulation of the defense hormones. This might be a precaution, although propagation of Vd is inhibited when the roots were pretreated with Pi.

Various reports showed the involvement of plant hormones in the control of Verticillium growth in Arabidopsis. Stabilization of cytokinin levels enhances Arabidopsis resistance against V. longisporum[17]. The fungus also requires JA-dependent COI1 function in roots to elicit disease symptoms in Arabidopsis shoots [15]. Ethylene perception via the receptor ETR1 is required for Vd infection in Arabidopsis [11]. Enhanced resistance of etr1-1 plants, but not of SA-, JA- or other ET-deficient mutants against Vd infection indicate a crucial role of ETR1 in defense against this pathogen. We observed a particularly striking resistance of the Arabidopsis ein3 mutant against Vd infection in vivo and in vitro (Additional file 1: Figure S5). This is consistent with the reports by Pantelides et al. [11] for etr1, although they did not observe a significant role of EIN3 in their studies. Our data demonstrate that EIN3 plays an important role in pathogenicity and will provide an important tool to identify EIN3-regulated genes which are required for Vd disease development. Furthermore, the ET level in the ein3 mutant exposed to Vd is much higher compared to Vd-exposed WT seedlings (Figure 10C). This suggests a feedback loop by which the lack of EIN3-induced defense responses in the ein3 mutant results in an additional stimulation of ET synthesis.

Conclusions

In summary, our data demonstrate that Pi is a very efficient biocontrol agent for Vd. This is mainly caused by the restriction of Vd growth in the presence of Pi. There appears to be additional mechanisms which prevent pathogenicity of Vd in the presence of Pi. For instance, the phytohormone levels accumulate to comparable levels in Vd and 1P2V seedlings, although Vd propagation is restricted in the presence of Pi (Figure 1). Since Pi pretreatment severely reduces defense gene expression in spite of a comparable phytohormone level in these tissues, additional signals from Pi must participate in the regulation of the immune response against Vd.

Methods

Growth conditions of seedlings and fungi

A. thaliana wild-type (ecotype Columbia-0) seeds, seeds of the glr2.4, glr2.5, glr3.3 and ein3 mutants as well as of cycam1 mutant [45] were surface-sterilized and placed on Petri dishes with MS media [61]. After cold treatment at 4°C for 48 h, plates were incubated for 11 days at 22°C under long day conditions (16 h light/8 h dark; 80 μmol m-2 sec-1). Pi was grown for 3-4 weeks on KM medium as described previously [62]. For detailed information see Section A and B in Johnson et al. [63]. Vd (FSU-343, Jena Microbial Resource Center, Germany) was grown for 2-3 weeks on Potato Dextrose Agar (PDA) medium [64].

Co-cultivation assays

For co-cultivation assays 13 day-old A. thaliana seedlings of equal size were used. Co-cultivation of A. thaliana and the fungi Pi and/or Vd was performed under in vitro culture conditions on a nylon membrane on PNM media as described by Johnson et al. ([63], Section C1 - Method 1) with a few modifications. Vd was grown for 12 days and Pi for 10 days on the membrane on top of PNM medium in Petri dishes. 13-day old Arabidopsis seedlings were then transferred to the Pi or Vd plates, or mock-treated (no fungal mycelium; C). For the shifting experiments, the seedlings were transferred to plates with the other fungus after 4 days (from Vd to Pi or vice-versa). Including the (C), five different treatments were compared: (1) Arabidopsis seedlings grown without Pi or Vd (C); (2) without Pi and with Vd (Vd); (3) with Pi and without Vd (Pi); (4) with Pi for 4 days before transfer to Vd plates (1P2V) and (5) with Vd for 4 days before transfer to Pi plates (1V2P). The seedlings were harvested between 1 and 21 days after exposure to the first fungus (or mock-treatment) for further analysis. A time scheme is shown in Additional file 1: Figure S1. The light intensity (80 μmol m-2s-1) was checked weekly. Shoots and roots were harvested separately for DNA and RNA analyses.

Long term co-cultivation in sterile vermiculite

30 g vermiculite was placed into one Magenta box (Sigma-Aldrich, Germany) and autoclaved at 121°C for 30 min. After the addition of 40 ml of sterile liquid PNM medium, Arabidopsis seedlings grown in Petri dishes for 10 days were transferred to the sterile vermiculite boxes (1 plant per box). For each treatment, 16 seedlings were analyzed. After 10 days, the number of survived plants, their biomass and fungal DNA content were determined.

Gene expression analysis

RNA was isolated from shoots and reverse-transcribed for Real-time quantitative PCR analysis, using an iCycler iQ Real-time PCR detection system and iCycler software version 2.2 (Bio-Rad). Total RNA was isolated from 5 independent biological experiments of Arabidopsis shoots. cDNA was synthesized using the Omniscript cDNA synthesis kit (QIAGEN) using 1 μg RNA. For the amplification of the RT-PCR products, iQ SYBR Green Supermix (Bio-Rad) was used according to the manufacturer's protocol in a final volume of 20 μl. The iCycler was programmed to 95°C 3 min, 40 × (95°C 30 sec, 57°C 15 sec, 72°C 30 sec), 72°C 10 min, followed by a melting curve program 55°C to 95°C in increasing steps of 0.5°C. All reactions were performed in triplicate. The mRNA levels for each cDNA probe were normalized with respect to the glycerin-aldehyde-3-phosphate dehydrogenase (GAPDH) mRNA level. The primer pairs are given in Additional file 1: Table S1.

Quantification of fungal DNAs by PCR

Genomic DNA extraction was conducted with DNeasy Plant Mini Kit. 12.5 ng DNA was taken for PCR template. The reactions were performed with gene-specific primers, as given in Additional file 1: Table S1. For details see Camehl et al. [65].

Dual culture of Pi and Vd

Dual culture of Pi and Vd on agar plates was performed as described by Johnson et al. [66]. A Pi plug with 5 mM diameter was placed at one end of a PDA plate and a Vd plug of the same size at the other end of the plate. The plates were incubated at 22-24°C in dark and 75% relative humidity. Photos were taken after 3 weeks of co-cultivation.

Percentage disease index (PDI) calculation

Disease index was calculated with the following formula:

PDI= n 1 x 1 + n 2 x 2 + n 3 x 3 + n 4 x 4 + n 5 x 5 Total number of leaves observed × maximum grade ×100

n1-5 = number of affected leaves of the respective disease.

Severity grade (0-5), x1-5 = disease severity grade based on the percentage of affected leaf area. 1, 1% ≤ × ≤ 10%; 2, 10% < × ≤ 20%; 3, 20% < × ≤ 30%; 4, 30% <× ≤ 40%; 5, × > 40%; ×100: calculated in percentage scale. Disease severity was estimated on the basis of affected leaf area. 1-5 disease severity grades were described by Naik and Lakkund [67],[68].

Quantification of jasmonic acid (JA), JA-isoleucine (JA-Ile), abscisic acid (ABA), salicylic acid (SA), oxophytodinoic acid (OPDA) and ethylene (ET)

Independent samples of 250 mg shoot material were collected from each treatment. Phytohormone extractions (JA, JA-Ile, ABA, SA and OPDA) were performed by adding 1 ml ethyl-acetate containing 60 ng of D2-JA and 40 ng of D6-ABA, D4-SA and JA-13C6-Ile (OPDA has the same internal standerd as JA) to 100 mg ground tissues. All samples were then vortexed for 10 min and centrifuged at 13,000 rpm for 20 min at 4°C. The supernatants were collected and evaporated to dryness at 30°C using a vacuum concentrator. Residues were resuspended in 500 μl MeOH:H2O (70:30, v/v) and centrifuged at 13,000 rpm for 10 min. The supernatants were collected and measured with the API 3200 LC-MS/MS system (Applied Biosystems, Framingham, USA) as previously described [69].

For ET measurements, 100 mg shoot material from each treatment was collected into 4 ml vials (Roth, Germany). After 4 h ET accumulation, the measurement was performed with the ETD-300 ethylene detector (Sensor Sense B.V., Nijmegen, The Netherlands).

Chlorophyll content was determined according to Yang et al. [70] and based on g fresh weight.

Quantification of microsclerotia

Roots of Arabidopsis seedlings from the 3 treatments with Vd were harvested after 3 weeks of co-cultivation in Petri dishes and transferred to a microscopic glass slide with 80 μl lactic acid/glycerol/H2O (1:1:1). The number of the microsclerotia formed in the roots was calculated averagely per root visually under the light microscope (magnification: x200). The experiment was performed 3 times independently and for each treatment the roots of 12 seedlings were analysed.

Cytoplasmic Ca2+ ([Ca2+]cyt) measurement

Aequorin based luminescence measurements were performed using 21-day old individual wild-type (WT) plants and mutants grown in Hoagland medium [71]. WT aequorin (pMAQ2) plants served as control [72]. Mutants (glr2.4, glr2.5 and glr3.3) were crossed back to wild-type expressing aequorin. After 2 generation selection based on [Ca2+]cyt responses and RT-PCR of T-DNA insertion examination, the homozygote seeds were used for the described experiments. Primers used for homozygosity tests are given in Additional file 1: Table S1. For [Ca2+]cyt measurements, approximately 70% of the roots per seedling was dissected and incubated overnight in 150 μl of 7.5 μM coelentrazine (P.J.K. GmbH, Germany) in the dark at 20°C in a 96 well plate (Thermo Fischer Scientific, Finland, cat. no. 9502887). Bioluminescence counts from roots were recorded as relative light units (RLU) with a microplate luminometer (Luminoskan Ascent, version 2.4, Thermo Electro Corporation, Finland).

Preparation of exudates from mycelia of Vd

A 5 mM Vd fungal plug was inoculated in Czapek's medium as described in Zhen et al. [73] and grown for 3 weeks. Then, the fungal culture was filtered through double layers of filter paper and the filtrate was centrifuged at 10,000 g for 30 min to remove the spores. The supernatant was dialyzed with a dialysis membrane (MWCO) (Spectra/Por® Float-A-lyzer®) in 10 mM phosphate buffer pH 7.0 at 4°C for 24 h. The dialyzed solution was frozen and lyophilized. The powder was dissolved in distilled water and the solution was filtered through a 0.45 μm pore size Millipore filter (Roth, Germany). The resulting filtrate was used as exudate for further experiments.

Statistics

All statistical analyses were performed using Excel or SPSS 17.0 (SPSS Inc., Chicago, IL, USA) for ANOVA.

Availability of supporting data

All the supporting data are included as additional file.

Additional file