Introduction

Biological control is an alternative strategy to reduce the use of chemicals thus contributing to the environment preservation. Biofungicides are produced by bacteria, fungi or yeasts. Antagonistic bacteria represent a realistic alternative to prevent plant diseases. They may act by several mechanisms including antagonism or antibiosis, competition, predation or hyperparasitism and stimulation of the natural defence of plants.

Several Pseudomonas strains have been studied for their antagonism and their ability to protect plants (Amein et al. 2008). They have several advantages namely their abundance in the rhizosphere, which represent 0.1–1% of rhizobacteria (Haas and Keel 2003), a great adaptation capacity to the nutrients present in their environment (Ghiglione et al. 2000) and the synthesis of several antagonistic molecules, such as 2,4-diacetylphloroglucinol, pyrrolnitrin, HCN, phenazine and pyoluteorin, which were the subject of detailed studies for their activities, production and action modes (Ayyadurai et al. 2006). Many Pseudomonas strains, including P. fluorescens Pf5 (Nowak-Thompson et al. 1994) and Pseudomonas sp. M18 (Hu et al. 2005) produce bioactive molecules.

Pyoluteorin is produced exclusively by certain Pseudomonas strains. It is associated with the control of Pythium which is responsible for root disease (Howell and Stipanovic 1980). This phenazine increases survival in soil environments and is essential for the biological control activity of certain Pseudomonas strains (Mavrodi et al. 2006). The production of these two molecules was coded by two clusters of genes which are rarely present in the same strain (Liu et al. 2006). In the present study, a strain of Pseudomonas fluorescens Pf1TZ was isolated. Its molecular characterization showed that it harbored genes from pyoluteorin and phenazine genes clusters. The Pf1TZ antifungal activities, in vitro and in planta, were investigated and HPLC-LC-MSn showed that the antagonism was due to several molecules, from which the most active one had a molecular mass of 504 Da.

Materials and methods

Isolation and identification of Pf1TZ

Pseudomonas fluorescens Pf1TZ was isolated from the rhizosphere of almond trees (Sfax, Tunisia) using the standard serial dilution method. Single fluorescent colonies were isolated and screened for the antifungal activity against Aspergillus niger using the co-culture method on PDA medium. Only one isolate was retained. This bacterium was identified using 16S rDNA sequence alignment. The PCR amplification of 16S rDNA gene was carried out using Taq DNA polymerase (Promega) and the universal primers, Rp1 and Rp2 (Supplementary Table 1), in minicycler (GeneAmp PCR system 2700, Applied Biosystems). Amplified PCR product was purified and sequenced with an automatic sequencer.

Search by PCR for genes involved in the Pseudomonas antifungal activity

Specific primers from sequence genes involved in the synthesis of antifungal molecules by Pseudomonas spp. were designed by alignment of the available sequences obtained from Genebank using the multiple sequence alignment software ClustalW. Three molecules were the subject of this study, namely, pyoluteorin (pltAf and pltAfr), pyrrolnitrin (prnDf and prnDfr) and phenazine (phzCf and phzCfr). Details of PCR amplifications are given in the Supplementary Table 1. PCR fragments were purified and sequenced with an automated DNA sequencer. Phylogenetic trees were constructed using Clustal X software.

Antifungal activity assays

Pf1TZ was tested for in vitro antagonism against several phytopathogenic fungi (Table 1) by standard co-culture technique on PDA. First, PDA plates were covered with 100 μl fungal spore suspension adjusted to 106 spores ml−1 using a Thomas counter. Next, 10 μl a cell suspension of Pf1TZ were removed from a 24 h culture in NBY (nutrient broth/yeast extract) and placed in the plate centre. Assay plates were incubated at 28°C for 3–5 days and diameters of growth inhibition zones were measured. To test the supernatant activity, wells were drilled in the PDA + rifampicin (30 μg ml−1), covered with spores and then 100 μl supernatant were added to the wells. For the quantification of the chromatographic fraction activities, 15 μl was deposited on sterile discs placed on PDA covered with spore suspension. All tests were repeated three times and data was subjected to analysis of variance (ANOVA) using EXCEL software. The obtained standard deviations were all less than 0.01.

Table 1 Evaluation of the antifungal activities of Pseudomonas fluorescens Pf1TZ
Full size table

Production, extraction and purification of Pf1TZ antifungal compounds

A single colony of the strain Pf1TZ was pre-cultured in NBY broth (0.3% meat extract; 0.5% peptone; 0.5% yeast extract) for 8–12 h at 27°C and 140 rpm, yielding approx. 109 c.f.u ml−1. Further culture was performed in NBY broth amended with a filter-sterilized ZnSO4·7H2O (1 mM) and autoclaved glycerol 1% (w/v). The culture was incubated for 48 h at 27°C with shaking at 140 rpm. It was then acidified to pH 2 with HCl and shaken with the same volume of ethyl acetate for 2 h, centrifuged for 10 min at 3000×g. Dry organic extract was determined. The residue was dissolved in 10 ml methanol for fractionation by flash chromatography (see Supplementary Table 2). Fractions were evaporated and weighed. Antifungal activities of all fractions were tested on A. niger. Active fractions were subsequently tested on the fungi presented in Table 1. The active molecules were purified using an analytic HPLC (Supplementary Table 3) and analyzed on an ion trap mass spectrometer fitted with an electrospray ionization (ESI) interface operating in both positive and negative ionization modes.

Application of Pf1TZ to the in vitro vine protection

Vitis viniferea L. cv. Chardonny was selected to test the endophytism and the plant protection capacity of Pf1TZ. Plantlets of Vitis viniferea were propagated by using nodal explants in a 25 mm diameter test tube containing 15 ml Martin Medium (Compant et al. 2005). The culture was in a growth chamber under white light with 16 h photoperiod at 26°C.

Endophytism test

To inoculate plants with Pf1TZ, 103 cells produced in NBY were washed and suspended in 200 μl PBS then deposited on the surface of the Martin Medium of the 45 days old plantlets (4–5 leaves). To test the endophytism of this strain, roots, stem and leaves of 1 week inoculated vitroplants were separated and washed with a serial-dilutions of bleach (from 10−2 to 10−5) and three times with sterile water. Each plantlet part was then crushed with a sterile pestle in PBS. Next, 100 μl were spread on the NBY agar plate and then incubated at 30°C for 48 h. One fluorescent colony was picked for identification using the specific primers (pltAf and pltAfr). In order to determine the population density of Pf1TZ in internal tissues of vitroplants, isolation and enumeration of bacterial cells were performed onto different vegetative parts of plants weekly during 3 months.

Confocal microscopy localization of Pf1TZ in plant tissues

From bacterized plantlets, portions of leaf, stem and root were cut and then bleached in 70% (v/v) ethanol for 30 min. These portions were finely cut and then observed with confocal microscopy (LSM 510, Zeiss). Non-inoculated plantlets were used as a control.

In vivo protection of vine plantlets

Every week and during 3 months, three plantlets inoculated with Pf1TZ were infected with 10 μl of spores suspension (106 spores ml−1) of B. cinerea that was poured on the surface of the upper leaf and then incubated for 10 days. As negative control, the same treatment was carried out on plantlets that were not inoculated by Pf1TZ.

Results

Identification and characterization of Pf1TZ

Pf1TZ was isolated from the rhizosphere. Antifungal activity was evaluated against A. niger. The isolate was identified by amplification, purification, sequencing and alignment of 1.5 kb PCR fragment of 16S rDNA. The Blastn alignment showed 98% sequence similarity with Pseudomonas fluorescens Pf-5 (AF394844). Based on this identification it is expected that the Pf1TZ strain contains the genes involved in the synthesis of pyoluteorin, phenazine or pyrrolnitrin. Using the specific primers, only amplifications of pyoluteorin and phenazine were obtained with an amplification of 0.64 kb fragment from pltA gene and 1.4 kb fragment from phzC gene as expected (Fig. 1). Sequencing analysis of these fragments by comparison with the gene bank helped to conclude that Pseudomonas fluorescens Pf1TZ had a/gene(s) of pyoluteorin and phenazine clusters. The phylogenetic trees constructed using the sequences of pltA and phzC and sequences available on the NCBI data base showed that these two DNA fragments were significantly different from their homologous in other strains, forming new branches (Fig. 2).

Fig. 1
figure 1

PCR amplification of gene fragments from clusters encoding for pyoluteorin and phenazine biosynthesis in Pseudomonas fluorescens Pf1TZ. a Lane 1 λ DNA-HincII digest. Lane 2 Amplification of 0.64 kb fragment DNA from pltA gene using pltAf and pltAfr primers. Lane 3 Negative control. b Lane 1 λ DNA-HindIII digest. Lane 2 Amplification of 1.4 kb fragment DNA from pltA gene using phzCf and phzCfr. Lane 3 Negative control

Full size image
Fig. 2
figure 2

Phylogenic trees of pltA (a) and phzC (b) DNA sequences of Pf1TZ and other Pseudomonas strains. ClustalX software (M18: Pseudomonas sp.; PACS171b: P. aeruginosa; LESB58: P. aeruginosa; Pf-5: P. fluorescens; PAO1: P. aeruginosa; NJ-10: P. fluorescens; Psd: P. putida; 383: Burkholderia sp.) was used

Full size image

Production and purification of Pf1TZ antifungal molecules

To improve the production of Pf1TZ antifungal molecules, several media were tested. Among these, the medium NBY + ZnSO4·7H2O (1 mM) + glycerol 1% (w/v) prompted the best antifungal activity of Pf1TZ. Antifungal compounds were totally extract from the liquid culture (1.6 g l−1) using ethyl acetate and were separated using flash chromatography. Three active fractions, Pf1TZ-17, Pf1TZ-23 and Pf1TZ-41, were eluted, respectively, at 80% cyclohexane + 20% acetonitrile, 55% cyclohexane + 45% acetonitrile and 75% methanol + 25% acetonitrile, with respective returns of 223, 142 and 602 mg. The LC-MSn analysis showed that these three fractions were different and the most active compound, Pf1TZ-23, has a molecular mass of 504 Da. Unexpectedly, molecular masses of 272 Da and 224 Da corresponding to pyoluteorin and phenazine, respectively, were not detected as main constituents in any active fractions.

Antifungal activity assays

The antifungal activities of Pf1TZ and its products were assessed by several methods (Table 1). The sensitivity of fungi is variable. A. alternata and B. cinerea were the most sensitive to both the crude fungicides and fractions obtained by flash chromatography particularly Pf1TZ-23. The results obtained with the culture supernatant and the organic extract could be compared because the tested quantity of the organic extract corresponded to 100 μl of the supernatant (equivalent to 8 × 108 c.f.u. ml−1). The activity of the organic extract was greater than the supernatant activity due to release of antifungal compounds from bacterial cells and their membranes by the acid hydrolysis before the liquid–liquid extraction. F. culmorum, F. graminearum and F. oxysporum were less sensitive to Pf1TZ fungicides than other tested fungi with a total insensitivity to the fraction Pf1TZ-17. For A. niger, A. alternata and F. oxysporum, a second activity zone appeared around the inhibition zone, characterized by a lower fungus growth density than the control and with a loss of characteristic pigmentation of these three fungi. Depigmentation was not observed when a culture of fungus was made on PDA medium from an inoculum taken from the second zone.

Application of Pf1TZ for the protection of vine plantlets

Pf1TZ endophytism

Six hours after the inoculation of Vitis viniferea plantlets, Pf1TZ colonized the root tissue, forming visible rings around the roots, which agreed with the work described by Chin-A-Woeng et al. (1997). This could be explained by the use of root exudates by the bacterium as carbon source. The isolation and characterization of the bacterium from leaves of vitroplant, after 1 week of its inoculation in the medium, indicated the endophytism of Pf1TZ. The monitoring of the population density of the bacterium in the inner tissues of plantlets showed that it colonized roots with an average density of 4 × 104 c.f.u. mg−1 of fresh roots. The bacterium is also present in stems and leaves with respective densities of 32 × 102 c.f.u. mg−1 and 152 c.f.u. mg−1. For further confirmation of this result, plant tissues were observed by confocal microscopy using to the auto-fluorescence of Pf1TZ. Figure 3 shows that inoculated plantlets roots had a very intense fluorescence particularly at the cell walls (Fig. 3a, b), but the fluorescence was more diffuse within the stem cells (Fig. 3c, d). In leaves, the fluorescence was less intense and presented both inside cells and in walls (Fig. 3e, f). These results confirmed the endophytic character of Pf1TZ strain and consolidated the enumeration of the bacterium inside plant.

Fig. 3
figure 3

Endophytism determination of P. fluorescens Pf1TZ by confocal microscopy using its auto-fluorescence and bacterium localization inside the inner tissues of Vitis viniferea plantlets. Roots (a), stems (c) and leaves (e) from a non-inoculated plantlet as a control. Roots (b), stems (d) and leaves (f) from 3 weeks inoculated plantlets with Pf1TZ cells. Bars 10 μm

Full size image

Plant protection

Protection monitoring of vine vitroplants during 3 months after inoculation of the bacterium shows a total protection against the pathogen B. cinerea, responsible for grey mould, if at least 3 weeks of bacterization was provided. This period was necessary for the bacteria to be present in suitable density to protect vitroplants from the pathogen (Fig. 4).

Fig. 4
figure 4

Application of Pf1TZ for plantlets protection from Botrytis cinerea grey mould disease. a Inoculated plantlets for 1 (b), 2 (c) and 3 weeks (d). All plantlets were infected by B. cinerea spores and incubated for 10 days. Bar 1 cm

Full size image

Discussion

Pseudomonas spp. are very abundant in soil and play a key role in the rhizosphere (Dowling and O’Gara 1994). This group is well known for its antifungal activity against several plant pathogenic fungi since it synthesises a variety of molecules via different pathways (Haas and Défago 2005). Here we describe the isolation and characterization of Pseudomonas fluorescens Pf1TZ harbouring antifungal activity against a variety of phytopathogenic fungi. To investigate this activity, we selected a number of fungi that are known to cause serious diseases on many different plants, namely, A. niger, A. alternaria and B. cinerea in addition to F. culmorum, F. graminearum and F. oxysporum which were reported to be relatively resistant to chemical and biological pesticides (Khan et al. 2006).

Fungicides of Pf1TZ were separated by flash chromatography and three active fractions were retained, namely, Pf1TZ-17, Pf1TZ-23 and Pf1TZ-41. Quantification of antifungal activities showed that Pf1TZ crude fungicide and its products are highly efficient against A. alternaria and B. cinerea and very efficient against A. niger. The activity of Pf1TZ against Fusarium was shown to be related only to Pf1TZ-23 and Pf1TZ-41 fractions, therefore, revealing the action specificity of Pf1TZ-17 compound.

The search for genes involved in antifungal activity evidenced presence of genes encoding pyoluteorin and phenazine. The co-existence of these two clusters is well known in the strain Pseudomonas sp. M18 (Hu et al. 2005) and was succinctly cited in a Pseudomonas fluorescens strain (Liu et al. 2006). Nevertheless, LC-MS analysis of Pf1TZ antifungal molecules showed that the strain acts by other molecules rather than pyoluteorin and phenazine suggesting that their clusters of genes are partial or truncated. The structure characterization of Pf1TZ active molecules by NMR is in progress. The synthesis of several antifungal molecules by Pf1TZ suggests its stratagem for dominating the host rhizosphere.

Pf1TZ activity was demonstrated in planta by its application for the protection of vine vitroplants against B. cinerea. The sensitivity of the latter to Pf1TZ fungicides prompted us to test its protective capacity in vivo. The verification of the Pf1TZ endophytism by confocal microscopy allowed us to infer that the bacteria preferentially colonized the cell membranes of roots. However, it was present at lower levels in stems and leaves. The bacteria enumeration in the plant tissues and the protection monitoring against B. cinerea had shown that total protection was reached after 3 weeks of inoculation, which was the time required for the bacteria to colonize the entire plant. Indeed, the growth of the bacterium in planta was ten times slower than in rich medium in vitro (Osburn et al. 1989). The effectiveness of Pf1TZ to protect plantlets against B. cinerea is closely linked to its antogonism and endophytism. It is not ruled out that the strain Pf1TZ also acted indirectly by inducing systemic resistance of the plant, this hypothesis is being studied. This dual activity further encourages field application of Pf1TZ as a biological control agent.

In conclusion, our data suggests that Pseudomonas fluorescens Pf1TZ is a candidate for biological control agent since it has shown antifungal activities in both in vitro and in planta conditions. Furthermore, the in vitro activities that were observed against several fungi, such as Alternaria, have encouraged us to include in our studies other plants that are widely affected by these fungi.