Functional, genetic and chemical characterization of biosurfactants produced by plant growth-promoting Pseudomonas putida 267

Authors


  • All the authors have contributed equally to this work.

Jos M. Raaijmakers, Laboratory of Phytopathology, Wageningen University, 6709 PD Wageningen, The Netherlands. E-mail: jos.raaijmakers@wur.nl

Abstract

Aims:  Plant growth-promoting Pseudomonas putida strain 267, originally isolated from the rhizosphere of black pepper, produces biosurfactants that cause lysis of zoospores of the oomycete pathogen Phytophthora capsici. The biosurfactants were characterized, the biosynthesis gene(s) partially identified, and their role in control of Phytophthora damping-off of cucumber evaluated.

Methods and Results:  The biosurfactants were shown to lyse zoospores of Phy. capsici and inhibit growth of the fungal pathogens Botrytis cinerea and Rhizoctonia solani. In vitro assays further showed that the biosurfactants of strain 267 are essential in swarming motility and biofilm formation. In spite of the zoosporicidal activity, the biosurfactants did not play a significant role in control of Phytophthora damping-off of cucumber, since both wild type strain 267 and its biosurfactant-deficient mutant were equally effective, and addition of the biosurfactants did not provide control. Genetic characterization revealed that surfactant biosynthesis in strain 267 is governed by homologues of PsoA and PsoB, two nonribosomal peptide synthetases involved in production of the cyclic lipopeptides (CLPs) putisolvin I and II. The structural relatedness of the biosurfactants of strain 267 to putisolvins I and II was supported by LC-MS and MS-MS analyses.

Conclusions:  The biosurfactants produced by Ps. putida 267 were identified as putisolvin-like CLPs; they are essential in swarming motility and biofilm formation, and have zoosporicidal and antifungal activities. Strain 267 provides excellent biocontrol activity against Phytophthora damping-off of cucumber, but the lipopeptide surfactants are not involved in disease suppression.

Significance and Impact of the Study: Pseudomonas putida 267 suppresses Phy. capsici damping-off of cucumber and provides a potential supplementary strategy to control this economically important oomycete pathogen. The putisolvin-like biosurfactants exhibit zoosporicidal and antifungal activities, yet they do not contribute to biocontrol of Phy. capsici and colonization of cucumber roots by Ps. putida 267. These results suggest that Ps. putida 267 employs other, yet uncharacterized, mechanisms to suppress Phy. capsici.

Introduction

The genus Pseudomonas harbours plant and human pathogenic species, as well as species and strains that degrade xenobiotic compounds, promote plant growth, antagonize plant pathogenic fungi and oomycetes, or induce systemic resistance in plants (Ramos 2004). The interest in Pseudomonas species is due, in part, to their ability to produce a wide variety of antimicrobial metabolites, including enzymes, volatiles, cyclic lipopeptides (CLPs) and antibiotics (Raaijmakers et al. 2002, 2006; Haas and Défago 2005; Ongena and Jacques 2008). CLPs possess potent surfactant activity and are typically composed of a fatty acid tail connected to a cyclic peptide. CLPs are synthesized nonribosomally and exhibit considerable structural variation in both the fatty acid and the peptidic ring (Nybroe and Sørensen 2004; Fishbach and Walsh 2006; Raaijmakers et al. 2006). Derivatives of particular CLPs may be produced through incorporation of alternative amino acids into the peptide moiety (De Bruijn et al. 2008).

Numerous plant-associated Pseudomonads are known to produce CLPs with versatile functions (Nybroe and Sørensen 2004; Raaijmakers et al. 2006; Tran et al. 2007). For plant pathogenic Pseudomonads, CLPs are important virulence factors and for antagonistic Pseudomonas strains, they contribute to motility, biofilm formation, root colonization, antimicrobial activity and biocontrol of plant diseases (Nybroe and Sørensen 2004; Raaijmakers et al. 2006). The antimicrobial properties of CLPs can be attributed, in part, to the disruption of membrane integrity via pore formation (Nybroe and Sørensen 2004; Raaijmakers et al. 2006). For instance, the activity of CLP-producing Pseudomonads against oomycete plant pathogens has been partly ascribed to the disruption of the membranes of infectious zoospores (De Souza et al. 2003; Raaijmakers et al. 2006; De Bruijn et al. 2007; Tran et al. 2007). CLPs may also trigger a systemic defense response in plants, leading to enhanced resistance against fungal and oomycete pathogens (Ongena et al. 2007; Tran et al. 2007). CLPs may also adversely affect mycelium growth, causing hyphal swellings, increased branching and rosette formation, and reduce intracellular activity of fungal and oomycete plant pathogens (Thrane et al. 1999, 2000; Hansen et al. 2000; Tran 2007). Although many CLPs have been proposed to play an important role in the control of plant pathogens, a comparison between the biocontrol activity of the CLP-producing wild type strain and CLP-deficient mutants is lacking in most of these studies. Bais et al. (2004) showed that a mutant of Bacillus subtilis strain 6051, defective in surfactin production, was substantially less effective than the wild type strain in controlling root infection of Arabidopsis by P. syringae. Studies with P. fluorescens strain SS101 also showed reduced activity of the massetolide-deficient mutant against late blight of tomato (Tran et al. 2007). However, in biocontrol of Pythium root rot of wheat and apple seedlings, the massetolide-deficient mutant of strain SS101 was equally effective (Mazzola et al. 2007).

In a recent survey of the diversity and activity of biosurfactant-producing Pseudomonads in the rhizosphere of black pepper plants in Vietnam, approximately 300 biosurfactant-producing isolates were collected and characterized (Tran et al. 2008). Representative isolates were identified by 16S-rDNA sequencing as Ps. putida and were shown to produce the same biosurfactants based on LC-MS analysis. Based on the molecular masses, several of the biosurfactants were predicted to be chemically related to the CLPs putisolvin I and II (Tran et al. 2008). Putisolvins I and II were initially described for Ps. putida PCL1445, a strain that was selected for efficient root colonization and growth on naphthalene (Kuiper et al. 2004; Dubern et al. 2008). The Ps. putida strains isolated in Vietnam, including strain 267, were shown to provide significant control of Phytophthora capsici root rot and to promote shoot and root development of black pepper stem cuttings. In dual culture assays, the strains did not show any inhibitory effect on Phy. capsici mycelial growth. However, cell-free culture supernatants containing the biosurfactants were shown to lyse zoospores of Phy. capsici and Phy. infestans (Tran et al. 2008). Collectively these observations pointed to a possible role of the biosurfactants in controlling Phytophthora diseases on diverse crops.

In this study, the biosurfactants produced by Ps. putida strain 267 were further characterized by ESI-MS-MS analysis. To asses their functions and activity spectrum, biosurfactant-deficient mutants of strain 267 were generated by random plasposon mutagenesis and characterized genetically. In vitro assays and growth chamber bioassays were performed with wild type strain 267 and a biosurfactant-deficient mutant to determine the role of biosurfactant production in (i) activity against pathogenic fungi and oomycetes, (ii) biocontrol of Phytophthora damping-off of cucumber, and (iii) swarming motility and biofilm formation.

Materials and methods

Micro-organisms and growth conditions

Pseudomonas putida 267 (Tran et al. 2008), Ps. fluorescens SS101 (De Souza et al. 2003) and their respective mutants were routinely grown at 25°C on Pseudomonas Agar (PSA; Difco, France). For all strains, spontaneous rifampicin resistant derivatives were used. Mutant 10.24 is a transposon mutant of strain SS101 deficient in the production of the CLP massetolide A (De Souza et al. 2003; De Bruijn et al. 2008). Phy. capsici strain LT3239, originally isolated from pumpkin, was kindly provided by Dr. Kurt H. Lamour (University of Tennessee, Knoxville, TN, USA). It was routinely grown at 25°C on V8 juice medium (V8, N.V. Campell Foods Puurs, Belgium) amended with 3 g l−1 CaCO3 and 15 g l−1 agar. Phy. capsici zoospores were obtained from V8 juice agar plates (145-mm-diameter) fully overgrown by Phy. capsici during 3–4 days of incubation in the dark at 25°C. Plates were further incubated at room temperature (∼20°C) for 4–5 days to stimulate the production of sporangia. The plates were flooded with 20 ml of sterile distilled water and incubated at 4°C for 30 min and subsequently incubated at room temperature for 30 min to release the zoospores. Zoospore dilutions were made in sterile distilled water. The density of the zoospores was determined microscopically (100× magnification) and adjusted to a concentration of 104 zoospores ml−1.

Zoosporicidal and antifungal activity of Pseudomonas strains and biosurfactants

Bacterial cell suspensions (109 CFU ml−1) were prepared from colonies grown on PSA plates for 48 h at 25°C. A 10-μl-aliquot of the bacterial cell suspension was mixed on a glass slide with an Phy. capsici zoospores (104 ml−1) in a 1 : 1 ratio (v : v). Zoospore lysis was observed microscopically at 100× magnification for up to 2 min. Dual culture inhibition assays were performed by spot-inoculating Ps. putida 267 to the edge of an agar plate and incubation for 3 days at 25°C, followed by placing a fungal or oomycete agar plug (5 mm diameter) to the centre of the plate and incubation at diverse temperatures for up to 14 days, The oomycete and fungal plant pathogens tested in these dual culture assays were Phy. capsici LT3239 (grown at 25°C on V8 and Rye Sucrose Agar (Latijnhouwers et al. 2004)) Phy. infestans 88069 (grown at 18°C on RSA), Pythium aphanidermatum and P. ultimum (both at 25°C, 1/5 Potato Dextrose Agar, pH 7·0), Rhizoctonia solani AG2-2IIIB (25°C, 1/5 PDA pH 7·0), Botrytis cinerea B0510 (20°C, PDA), Sclerotium rolfsii QN002 (25°C, 1/5 PDA pH 7·0), Verticillium albo-atrum VA1 and V. dahliae JR2 (both at 22°C, PDA).

To determine the effect of purified biosurfactants on radial growth of the fungi and oomycetes, the biosurfactants were dissolved in water and added to agar medium (800 μl biosurfactant solution and 20 ml agar medium per plate) at different concentrations. The agar media used for the different fungi and oomycetes were as in the dual culture assays. A single plug of freshly grown mycelium was placed in the centre of the plate.

Purification and chemical analysis of biosurfactants

Biosurfactant extracts were prepared from cell cultures as described in detail by De Souza et al. (2003), De Bruijn et al. (2007) and Tran et al. (2008). To determine the critical micelle concentration (CMC), the extracts were dissolved in sterile demineralized water at different concentrations and surface tension measurements were performed at 25°C with a K6 tensiometer (Krüss GmbH, Hamburg, Germany). Extracts for reverse phase high performance liquid chromatography (RP-HPLC) were prepared in the same way, but were dissolved in a mixture of 15% HPLC-purified water, 45% acetonitrile, 40% methanol and 0·1% trifluoroacetic acid.

RP-HPLC analysis was performed as described previously (De Souza et al. 2003; De Bruijn et al. 2007, 2008; Tran et al. 2008). For determining the molecular masses of the individual RP-HPLC peaks, liquid chromatography-mass spectrometry (LC-MS) was performed as described previously (De Bruijn et al. 2007, 2008; Tran et al. 2008). For evaporative ionization tandem mass-spectrometry (ESI-MS-MS) analysis, 200 μg of the biosurfactant extract was dissolved in 6 ml dimethylsulfoxide. After membrane filtration, five injections of each 1150 μl were carried out on a Shimadzu autopreparative system with an Alltech end-capped 5 μm C18 column of 250 × 22 mm at 23 ml min−1 using an isocratic flow of 15% HPLC-purified water with 0·1% trifluoroacetic acid, 45% acetonitrile with 0·1% trifluoroacetic acid and 40% methanol. Fractions containing individual peaks were collected based on their UV signal and the eluent was removed using a rotary evaporator. All fractions were investigated by (+)-ESI-MS-MS using helium as a collision gas. Only a single parent ion was kept in resonance (isolation m/z 3) and all other ions were ejected from the trap without mass analysis. The ion was then agitated and allowed to fragment by collision-induced dissociation (CID).

Plasposon mutagenesis and gene sequencing

Random plasposon mutagenesis of Ps. putida strain 267 was performed by triparental mating with pTnMod-OTc as a donor (Dennis and Zylstra 1998) and pRK2013 as a helper (Figurski and Helsinki 1979). Mutants were selected on KB agar medium (King et al. 1954) without MgSO4, and supplemented with 50 μg ml−1 tetracycline and 100 μg ml−1 rifampicin. Mutants were transferred to microtitre plates containing 100 μl KB agar medium per well. After 2 days of growth at 25°C, colonies were dissolved in 100 μl of sterile distilled water and 10 μl droplets were tested in a drop-collapse assay (De Bruijn et al. 2007). Mutant cultures that did not show the typical drop-collapse as observed for wild type strain 267 were selected for further identification and sequencing of the regions flanking the plasposon insertion. Strain integrity of the mutants was confirmed by BOX-PCR as described by Tran et al. (2008).

Genomic DNA was isolated from 1·5 ml overnight cultures (Sambrook and Russell 2001) and digested with BamHI or PstI (Promega, Madison, WI, USA). Total digested genomic DNA was self-ligated and transformed by electroporation to Escherichia coli DH5α. Colonies were selected on Luria-Bertani (LB) plates containing 50 μg ml−1 tetracycline. Plasmid mini-preps were performed on 1·5 ml overnight cultures (Sambrook and Russell 2001), and the rescued plasposons were digested with BamHI or PstI to assess the insert sizes. The rescued plasposons were sequenced (BaseClear, Leiden, the Netherlands) using primers OTc-SEQ4 (5′-acggttcctggccttttgc-3′) and OTc-SEQ6 (5′-tgataaactaccgcattaaagc-3′). The obtained sequences were trimmed to end at the BamHI or PstI sites used in the plasposon rescue, and to remove plasposon and poor quality sequences. The flanking sequences were subsequently joined to obtain a single sequence (GenBank accession numbers EU886639EU886642 for mutants CP2, DP2, EP1 and HB1, respectively).

Swarming and biofilm assays

Swarming experiments were performed on soft agar plates (KB medium with 0·6% (w/v) agar). Bacterial cells grown for 24 h on PSA agar plates were dissolved in sterile distilled water to a final density of 109 CFU ml−1 (OD600 = 1), pelleted by centrifugation and washed once with sterile distilled water. Five microlitres of the cell suspension were placed in the centre of a soft agar plate and the ability of the bacterial colony to spread was evaluated after 24, 48 and 72 h of incubation at 25°C.

The biofilm assays were performed in flat-bottom 96-wells plates (Greiner-Bio One GmbH, Frickenhausen, Germany) according to the methods described by O’Toole et al. (1999) and De Bruijn et al. (2007). Wells were filled with 180 μl of SSM medium and 20 μl bacterial suspension (1 × 109 cells ml−1), and incubated for 24 h at 25°C. Biofilms were stained with crystal violet for visualization (De Bruijn et al. 2007).

Biocontrol activity of Ps. putida 267 against Phy. capsici on cucumber

To allow testing of biosurfactant-deficient mutants in disease assays, cucumber was chosen as a host for Phy. capsici. This pathosystem allows rapid and large-scale bioassays under controlled conditions. Among several tested Phy. capsici strains, strain LT3239 was the most virulent on cucumber. Cucumber seeds (cv. Chinese Slangen; Pieterpikzonen BV Holland, Heerenveen, the Netherlands) were surface sterilized and subsequently dried under continuous airflow prior to use. Trays with rockwool plugs (20-mm diameter and 25-mm height) were used for the bioassay. Bacterial inoculum was prepared by growing the Pseudomonas wild type and mutants on PSA plates at 25°C for 48 h. Bacterial cells were collected and washed in sterile distilled water and adjusted to a final density of 108 CFU ml−1 or 109 CFU ml−1. Four millilitres of the bacterial suspension or 4 ml of the biosurfactant solution (50 μg ml−1) were added to each rockwool plug. Next, one ml of zoospores of Phy. capsici LT3239 (104 ml−1) was added and one cucumber seed was sown in each rockwool plug and covered with a 1-cm layer of river sand. The rockwool trays were placed in boxes with a transparent cover and incubated in a climate chamber (25°C, 16 h photoperiod). Germination of the cucumber seeds and postemergence damping-off was scored 5 days after sowing. The disease incidence (%) was calculated by dividing the number of plants suffering from damping-off disease by the total number of seedlings. Each treatment had four replicates with 10 plants per replicate.

To study colonization of cucumber roots by bacterial strains, cucumber seeds were sown in a mixture of river sand and potting soil (4 : 1), amended with 106 cells ml−1. After 11 days of growth in a climate chamber (25°C, 16 h photoperiod), roots were harvested. Rhizosphere suspensions (roots with adhering soil) were prepared and serially diluted onto PSA supplemented with 40 μg·ml−1 ampicilin, 12·5 μg ml−1 chloramphenicol, 100 μg ml−1 Delvocid and 100 μg ml−1 rifampicin. Plates were incubated at 25°C for 3 days and colonies were counted.

Statistics

All experiments described were performed at least two times and representative results are shown. Disease incidence (%) was arcsin-transformed prior to statistical analysis. The area under the disease progression curve (AUDPC) was calculated using the trapezoidal integration method described by Landa et al. (2002). Statistical differences (P < 0·05) between treatments were analysed by anova followed by the Tukey test (SAS Institute, Inc., Cary, NC, USA). Normal distribution of the data and homogeneity of variances was tested prior to anova.

Results

Zoospore lysis, surface and antifungal activity of the Ps. putida 267 biosurfactants

Dose-response experiment performed with partially purified biosurfactants of Ps. putida 267 showed that lysis of Phy. capsici zoospores occurred within 90 s at concentrations of 25 μg ml−1 or higher (Fig. 1). Surface tension measurements showed a response curve typical for biosurfactants and indicate that the critical micelle concentration (CMC) of the biosurfactants from strain 267 is approximately 25 μg ml−1 (Fig. 1).

Figure 1.

 Dose–response relationship between the concentration of the biosurfactants of Pseudomonas putida 267 and the surface tension. The biosurfactants were dissolved in sterile distilled water at different concentrations. Mean values of three replicates are shown and error bars represent the standard errors. The vertical line indicates the minimum concentration at which lysis of zoospores of Phytophthora capsici occurs within 90 s after exposure.

Pseudomonas putida 267 did not inhibit hyphal growth of Phy. capsici in dual culture assays nor did strain 267 inhibit mycelial growth of the oomycete plant pathogens Phy. infestans, Pythium aphanidermatum and P. ultimum, and the fungal pathogens Rhizoctonia solani, Sclerotium rolfsii, Verticillium albo-atrum and V. dahliae. Only for the fungal pathogen Botrytis cinerea, a transient inhibition of mycelial growth was observed (data not shown). When B. cinerea was grown on plates containing the biosurfactants of Ps. putida 267 at concentrations ranging from 25 μg ml−1 to 200 μg ml−1, a significant reduction in radial mycelium growth was observed (Fig. 2a). Similar effects were found for R. solani (Fig. 2b). Growth of the other oomycete and fungal pathogens included in this study was not significantly inhibited on agar plates amended with the biosurfactants at concentrations up to 200 μg ml−1 (data not shown).

Figure 2.

 Inhibitory effect of the biosurfactants of Pseudomonas putida 267 on mycelium growth of the fungal plant pathogens Botrytis cinerea (panel a) and Rhizoctonia solani (panel b). Agar plates were amended with different concentrations of biosurfactants. Radial growth was recorded 1, 2, 3, 4 and 6 days after inoculation. Mean values of four replicates are shown and error bars represent the standard errors. (inline image, 0; inline image, 25; inline image,50; inline image, 100; inline image, 200 μg ml−1).

Generation of biosurfactant-deficient mutants of Ps. putida 267 and gene identification

A total of 1296 mutants of strain 267, obtained by random plasposon mutagenesis, were screened for loss of biosurfactant production in a drop-collapse assay. Four mutants were identified that had completely lost the ability to collapse a droplet of water. Plasposon rescue was performed on these four mutants and flanking sequences on both sides of the plasposon insertion were obtained. From the biosurfactant-deficient mutants CP2, DP2, EP1 and HB1, sequences of 231, 648, 939 and 933 bp, respectively, were obtained after editing. No overlap was found among these sequences. For the sequence obtained for mutant HB1, a 452-bp noncoding sequence upstream of the open reading frame did not reveal any homology to Genbank sequences. All other sequences were part of open reading frames. Their translated sequences yielded highest blastp hits (McGinnis and Madden 2004) with both characterized and uncharacterized nonribosomal peptide synthetases (NRPSs) involved in cyclic lipopeptide production by Pseudomonads. For each of the mutants, the highest identities found were to PsoA and PsoB, two NRPSs responsible for biosynthesis of the CLPs putisolvin I and II (Table 1).

Table 1.   Results of BlastP analyses of the translated open reading frames disrupted in four biosurfactant-deficient mutants of Pseudomonas putida 267
MutantAmino acidsBlastP hit*GenBank accessionE-valueIdentity (%)
  1. NRPS, nonribosomal peptide synthetase; Ps; Pseudomonas.

  2. *The highest three blastp hits are shown.

CP277Putisolvin synthetase PsoB (Ps. putida PCL1445)ABW173762e-1665
NRPS (Ps. entomophila L48)YP_6086052e-1259
NRPS (Ps. entomophila L48)YP_6086062e-1255
DP2206Putisolvin synthetase PsoB (Ps. putida PCL1445)ABW173766e-9280
Syringofactin synthetase SyfA (Ps. syringae pv. tomato DC3000)NP_7926331e-8070
NRPS (Ps. syringae pv. syringae B728a)YP_2356531e-8069
EP1312Putisolvin synthetase PsoB (Ps. putida PCL1445)ABW173763e-15487
NRPS (Ps. entomophila L48)YP_6086068e-14282
NRPS (Ps. entomophila L48)YP_6086056e-11773
HB1160Putisolvin synthetase PsoA (Ps. putida PCL1445)ABW173755e-5672
NRPS (Ps. entomophila L48)YP_6088772e-4359
Syringofactin synthetase SyfA (Ps. syringae pv. tomato DC3000)NP_7926333e-3751

Characteristics of biosurfactant-deficient mutant EP1

The biosurfactant-deficient mutant EP1, for which most sequence information was obtained, was further characterized phenotypically. Mutant EP1 is not able to reduce the surface tension of water (Table 2) and does not cause cessation of zoospore motility nor zoospore lysis (Table 2). In vitro assays further showed that wild type strain 267 was able to swarm on soft agar medium, whereas mutant EP1 had lost this ability completely (Fig. 3a, Table 2). Biofilm assays showed that mutant EP1 was capable of forming a substantial biofilm, whereas wild type strain 267 showed only marginal biofilm formation (Fig. 3b,c).

Table 2.   Phenotypic characteristics of Pseudomonas putida 267 and its biosurfactant-deficient mutant EP1
IsolateDrop collapse*Surface tension†Zoospore motility‡Zoospore lysis§ Swarming¶Biofilm formation**
  1. *5 μl droplets of bacterial cell suspensions (OD600 = 1) were tested in a drop–collapse assay on Parafilm; ‘+’ indicates a drop collapse.

  2. †Surface tension (mN m−1) of bacterial cell suspensions (OD600 = 1).

  3. ‡Zoospore motility was observed microscopically after addition of bacterial cell suspensions (OD600 = 1) to zoospores (104 zoospores ml−1) of Phy. capsici LT3239 in a 1 : 1 (v/v) ratio. A ‘+’ indicates cessation of zoospore motility.

  4. §Zoospore lysis was observed microscopically after bacterial cell suspensions (OD600 = 1) were mixed with zoospores (104 zoospores ml−1) of Phy. capsici LT3239 in a 1 : 1 (v/v) ratio. A ‘+’ indicates zoospore lysis.

  5. ¶Strains were tested for swarming by spotting 5 μl bacterial cell suspension (109 cells ml−1) on a soft agar (0·6% w/v) plate. A ‘+’ indicates the ability to swarm outwards.

  6. **Biofilm formation of the bacterial strains was tested in 96-well plates filled with 150 μl liquid KB medium per well. Biofilms were stained with crystal violet after 48 h of incubation.

  7. na, not applicable.

Water71·2nana
Ps. putida 267+31·5+++
EP174·0+
Figure 3.

 Swarming and biofilm characteristics of Pseudomonas putida 267 and its biosurfactant-deficient mutant EP1. (a) Swarming behaviour of Ps. putida 267 and EP1 on soft agar (0·6% w/v) plates. A single drop of a bacterial suspension was placed in the centre of a plate. Plates were incubated for 72 h. (b) Visual representation of biofilm formation by Ps. putida 267 and EP1. Cell cultures were grown in wells for 24 h and biofilms were stained with crystal violet. (c) Spectrophotometric quantification of the biofilm formed by Ps. putida 267 and EP1. The quantity of crystal violet stained biofilm was measured at 600 nm. Mean values of four replicates are given and error bars indicate the standard error. The asterisks indicates a statistically significant difference (P < 0·05).

LC-MS and MS-MS analysis of the Ps. putida 267 biosurfactants

RP-HPLC analysis of the partially purified biosurfactants of strain 267 showed at least five distinct peaks (Fig. 4). These five peaks are all missing in extracts obtained from the biosurfactant-deficient mutants of Ps. putida 267 (data not shown). The molecular masses of each of these five peaks was determined by LC-MS. Peak 1 consists of two different compounds with molecular masses of 1398 and 1420 with the same retention time (Fig. 4), which we were unable to separate (data not shown). The molecular masses of peaks 1 and 2 do not correspond to known CLPs. LC-MS further revealed that peaks 3, 4 and 5 (Fig. 4) have molecular masses of 1380 (peak 3) and 1394 (peaks 4 and 5), corresponding to the molecular masses of putisolvins I and II (Kuiper et al. 2004). Subsequent ESI-MS-MS analysis showed a fragmentation pattern for peak 3 identical to that of putisolvin I (Table 3). Putisolvin II differs from putisolvin I by substitution of a valine with a leucine/isoleucine residue, resulting in an increase of the molecular mass by 14 (Kuiper et al. 2004). The ESI-MS-MS profile of both peaks 4 and 5 showed some fragments with masses identical to those of putisolvin I, whereas the molecular masses of other fragments were 14 higher than that of putisolvin I (Table 3), consistent with the data of putisolvin II (Kuiper et al. 2004). Therefore, peaks 4 and 5 are most likely putisolvin II. The different retention times of peaks 4 and 5 (Fig. 4) most likely reflect an amino acid substitution that does not result in a shift in the molecular mass, such as a leucine to isoleucine substitution.

Figure 4.

 RP-HPLC chromatogram (206 nm) of the biosurfactant extract of Pseudomonas putida 267. The numbers above the peaks indicate the molecular masses as determined by LC-MS analysis.

Table 3.   Comparison of the masses of the MS-MS fragments of putisolvins I and II, and three putative biosurfactants produced by Pseudomonas putida 267
Putisolvin I and II†Peak 3‡Peaks 4 and 5‡
  1. †The molecular masses of the MS-MS fragments and the parent ion of putisolvin I are according to Kuiper et al. (2004). Fragments of putisolvin II (molecular mass of the parent ion is 1394) with an expected mass of +14 relative to putisolvin I are indicated with an asterisk (*).

  2. ‡For RP-HPLC peaks 3, 4 and 5 of Ps. putida 267 (Fig. 4), ESI-MS-MS analysis was performed. A ‘+’ indicates detection of a fragment with a mass identical to MS-MS fragments of putisolvin I (peak 3) or putisolvin II (peaks 4 and 5); nd: not detected.

1380* (parent ion)++
 341++
 454++
 567++
 686*++
 695++
 782++
 814*++
 863++
 881++
 927*++
 994++
1040*++
1081+nd
1169*++

Role of biosurfactants in the biocontrol activity of Ps. putida 267

Growth chamber assays were conducted to determine the activity of strain 267, mutant EP1, and its biosurfactants against Phy. capsici, the causal agent of pre- and postemergence damping-off of cucumber. Massetolide A-producing strain Ps. fluorescens SS101, massetolide A-deficient mutant 10.24, and purified massetolide A were included as references. The results showed that wild type strains 267 and SS101 significantly reduced pre- and postemergence damping-off of cucumber (Fig. 5, Table 4). The biosurfactant-deficient mutants EP1 and 10·24 were as effective as their parental strains, whereas addition of the biosurfactants of strain 267 or massetolide A did not provide any control of pre- and postemergence damping-off of cucumber (Fig. 5, Table 4). No significant differences in colonization of cucumber roots were observed between wild type strain 267 and mutant EP1 (strain 267: log CFU g−1 root = 7·62 ± 0·10; mutant EP1: log CFU g−1 root = 7·49 ± 0·12). These results indicate that, in spite of their zoosporicidal activities, these biosurfactants do not contribute to the high level of biocontrol activity of both strains 267 and SS101 against Phy. capsici on cucumber.

Figure 5.

 Pre- and postemergence damping-off of cucumber caused by Phytophthora capsici. (a) Pre-emergence damping-off of cucumber caused by Phytophthora capsici. ‘–’, no P. capsici zoospores were applied; ‘+’, cucumber seeds inoculated with zoospores of Phy. capsici LT3239. (b) Typical symptom of postemergence damping-off of cucumber caused by Phytophthora capsici. (c and d) Effect of application of Pseudomonas putida 267, its biosurfactants, or the biosurfactant-deficient mutant EP1 on pre-emergence (c) and postemergence (d) damping-off of cucumber, indicated as disease incidence (%). In the control treatment, only zoospores of Phy. capsici were applied. Applications of Ps. fluorescens SS101, massetolide A and the massetolide A-deficient mutant 10·24 were included as references. Pre- and postemergence damping-off was scored 5 days after sowing. Mean values of four replicates are shown and error bars represent the standard errors. Different letters indicate a statistically significant difference (P < 0·05).

Table 4.   Area under the disease progress curve (AUDPC) of postemergence damping-off of cucumber caused by Phytophthora capsici
Treatment*AUDPC†Tukey test‡
  1. *Cucumber seeds were sown in rockwool plugs amended with water (control), cell suspensions of Pseudomona putida 267, biosurfactant-deficient mutant EP1, Ps. fluorescens SS101, massetolide A-deficient mutant 10·24, biosurfactants of strain 267, or massetolide A. Rockwool plugs were inoculated with zoospores of Phy. capsici LT3239 and pre-emergence damping-off was assessed after 5 days of incubation.

  2. †Postemergence damping-off of cucumber seedlings was monitored after 5, 7, 10 and 13 days of plant growth. Based on these data, the area under the disease progress curve (AUDPC) was calculated for each of the treatments. Means of four replicates are shown.

  3. ‡Differences between treatments were analysed by the Tukey test and means with different letters are statistically significant different (P < 0·05).

Control700c
Ps. putida 267105a
Biosurfactant 267595bc
EP1175a
Ps. fluorescens SS101280ab
Massetolide A700c
10.24280ab

Discussion

Plant growth-promoting Pseudomonas putida strain 267, originally isolated from the rhizosphere of black pepper, produces biosurfactants that cause lysis of zoospores of the oomycete plant pathogen Phytophthora capsici. In the present study, several complementary approaches were adopted to (i) characterize the biosurfactants produced by strain 267, (ii) partially identify genes involved in their biosynthesis and (iii) evaluate the contribution of biosurfactant production in the control of Phytophthora damping-off disease of cucumber. Sequence analysis of four biosurfactant-deficient mutants of strain 267 indicated that the biosurfactants are produced via nonribosomal peptide synthesis and are most likely cyclic lipopeptide surfactants (CLPs). The highest identity of the disrupted genes in the mutants was with psoA and psoB, two genes involved in putisolvin biosynthesis in Ps. putida PCL1445 (Dubern et al. 2008). The structural relatedness of the biosurfactants of strain 267 with putisolvins I and II was further supported by LC-MS and ESI-MS-MS analyses. The other peaks observed in the chromatogram are most likely derivatives of putisolvins I and II, differing in amino acid composition or in their fatty acid tail. NMR and chiral-GC analyses will be required to confirm their exact structures. Although strains Ps. putida 267 and PCL1445 appear to produce the same biosurfactants, amino acid sequence variation between their respective synthetases was observed with identities ranging from 65% and 87%. This may explain, in part, the differences in the number of biosurfactants produced by the two strains: PCL1445 produces putisolvin I and II, whereas strain 267 produces putisolvin I, two variants of putisolvin II and at least three additional derivatives. Isolation and sequence analysis of the complete CLP biosynthesis cluster of strain 267 will be necessary to give insight into the level of similarity with the pso biosynthesis cluster.

Ron and Rosenberg (2001) proposed several natural roles for CLPs and other biosurfactants, including a function in antimicrobial activity, regulation of attachment and detachment to and from surfaces, and motility. CLPs may facilitate attachment to or detachment from surfaces depending on the surface properties of the CLP-producing strain and the CLPs that are produced (Neu 1996; De Bruijn et al. 2008). For Ps. putida strain 267, the biosurfactants were shown to inhibit biofilm formation and to be essential for surface motility. These results are consistent with the data obtained for putisolvin-producing strain PCL1445 (Kuiper et al. 2004). Addition of purified putisolvin I to the growth medium prior to incubation reduced biofilm formation by strain PCL1445 in a concentration-dependent manner (Kuiper et al. 2004). Similarly, an arthrofactin-deficient mutant of Pseudomonas sp. MIS38 formed unstable, but thicker biofilms than the wild type (Roongsawang et al. 2003). In contrast, viscosin and massetolide A are essential in biofilm formation by Ps. fluorescens SBW25 and SS101, respectively (De Bruijn et al. 2007, 2008). How CLPs influence biofilm formation is still unclear, but their effect on cell surface hydrophobicity may play an important role in this. Hydrophobic interactions and surface-active compounds have been widely suggested to play a role in the adherence of cells to surfaces (Neu 1996). Given the diversity in structures and hydrophobicities of various CLPs produced by Pseudomonas strains, we postulate that depending on the cell surface of the producing strain as well as the structure and hydrophobicity of the CLP produced, their role in biofilm formation may differ accordingly.

With respect to the role of biosurfactants in root colonization, Bais et al. (2004) postulated that CLP production may enable bacteria to efficiently colonize plant roots thereby providing protection to their host. Indeed, massetolide A was found to positively contribute to colonization of tomato roots by Ps. fluorescens SS101 and to protection against late blight disease (Tran et al. 2007). Similarly, the CLP amphisin was shown to contribute to colonization of sugar beet seeds and roots by Ps. fluorescens DSS73 (Nielsen et al. 2005). In this study, however, no difference in colonization abilities of strain 267 and its biosurfactant-deficient mutant EP1 were observed, suggesting that, at least for cucumber plants, putisolvin-like surfactants do not contribute to the rhizosphere competence of Ps. putida 267.

In vitro assays showed that the biosurfactants produced by strain 267 did not exhibit activity against mycelial growth of several fungal and oomycete plant pathogens. Growth of R. solani was not inhibited by strain 267, yet is was inhibited to some extent when the purified biosurfactants were added to the growth medium. The CLPs tensin and viscosinamide were also shown to adversely affect mycelial growth of R. solani (Thrane et al. 1999; Nielsen et al. 2000). In contrast, mycelium growth of B. cinerea was significantly inhibited by strain 267 and its purified biosurfactants. The different effects of the CLPs of strain 267 on different pathogens may reflect differences in membrane composition between these pathogens or the ability of these pathogens to resist or inactivate the CLPs. Although strain 267 did not have a direct growth-inhibitory effect on most of the pathogens in vitro, it may well exhibit biocontrol activity of these pathogens when applied to the host plants, as was shown for control of Phy. capsici on black pepper (Tran et al. 2008). The oomycete Phy. capsici Leonian infects a variety of solanaceous and cucurbitaceous hosts, including cucumber, tomato, eggplant, pumpkin, squash, and melon (Erwin and Ribeiro 1996;Hausbeck and Lamour 2004). Zoospores are important propagules in the infection process and a potential target to control Phy. capsici. The results of the present study show that, in spite of their zoosporicidal activity, the biosurfactants produced by Ps. putida 267 do not contribute to biocontrol of Phytophthora damping-off of cucumber. Similar results were obtained for massetolide A of Ps. fluorescens strain SS101. This is in contrast to results obtained in previous studies where massetolide A was shown to be involved in biological control of late blight of tomato by Ps. fluorescens SS101 (Tran et al. 2007), and where application of the biosurfactant rhamnolipid B suppressed diseases caused by Phy. capsici and Colletotrichum orbiculare on pepper and cucumber (Kim et al. 2000). In the present study, the amount of biosurfactants of strain 267 applied to the rockwool may not have been sufficient to lyse all of the zoospores of Phy. capsici, or zoospores may have encysted rapidly after introduction rendering them insensitive to lysis by the biosurfactants. Nevertheless, wild type strains 267 and SS101 and their surfactant-deficient mutants did provide substantial level of disease control, indicating that mechanisms other than biosurfactant production are involved in disease suppression. Similar observations were made for biocontrol of Pythium root rot of apple and wheat by Ps. fluorescens SS101 (Mazzola et al. 2007). To control Phy. capsici, crop rotation in conjunction with cultural and chemical control strategies are currently recommended in practice. However, these strategies have not yet provided significant economic control for diverse crops (Lamour and Hausbeck 2003; Drenth and Guest 2004; Hausbeck and Lamour 2004). In a previous study (Tran et al. 2008) and in the present study, Phy. capsici disease control was achieved on both black pepper and cucumber by application of antagonistic Pseudomonas strains 267 or SS101. Although the CLPs do not contribute to the control of Phytophthora damping-off of cucumber, the high levels of disease control achieved by strains 267 and SS101 may provide an attractive supplementary strategy to control this economically important oomycete pathogen.

Acknowledgements

This research was sponsored, in part, by a Bsik-subsidy from the Dutch Ministry of Economic Affairs (Ecogenomics; M.K.) and by the Vietnamese Ministry of Education and Training (MOET) through project 322 (H.T.). We thank Dr Kurt H. Lamour from University of Tennessee (Knoxville, TN, USA) for the Phytophthora capsici strain and Dr Teris van Beek, Pieter de Waard, Frank Claassen and Irene de Bruijn from Wageningen University (the Netherlands) for the LC-MS and ESI-MS-MS analyses.

Ancillary

Advertisement