Cross Talk between 2,4-Diacetylphloroglucinol-Producing Biocontrol Pseudomonads on Wheat Roots

Abstract

The performance of Pseudomonas biocontrol agents may be improved by applying mixtures of strains which are complementary in their capacity to suppress plant diseases. Here, we have chosen the combination of Pseudomonas fluorescens CHA0 with another well-characterized biocontrol agent, P. fluorescens Q2-87, as a model to study how these strains affect each other's expression of a biocontrol trait. In both strains, production of the antimicrobial compound 2,4-diacetylphloroglucinol (DAPG) is a crucial factor contributing to the suppression of root diseases. DAPG acts as a signaling compound inducing the expression of its own biosynthetic genes. Experimental setups were developed to investigate whether, when combining strains CHA0 and Q2-87, DAPG excreted by one strain may influence expression of DAPG-biosynthetic genes in the other strain in vitro and on the roots of wheat. DAPG production was monitored by observing the expression of lacZ fused to the biosynthetic gene phlA of the respective strain. Dual-culture assays in which the two strains were grown in liquid medium physically separated by a membrane revealed that Q2-87 but not its DAPG-negative mutant Q2-87::Tn5-1 strongly induced phlA expression in a ΔphlA mutant of strain CHA0. In the same way, phlA expression in a Q2-87 background was induced by DAPG produced by CHA0. When coinoculated onto the roots of wheat seedlings grown under gnotobiotic conditions, strains Q2-87 and CHA0, but not their respective DAPG-negative mutants, were able to enhance phlA expression in each other. In summary, we have established that two nonrelated pseudomonads may stimulate each other in the expression of an antimicrobial compound important for biocontrol. This interpopulation communication occurs in the rhizosphere, i.e., at the site of pathogen inhibition, and is mediated by the antimicrobial compound itself acting as a signal exchanged between the two pseudomonads.

Some root-associated fluorescent pseudomonads produce and excrete secondary metabolites which are inhibitory to plant-pathogenic rhizosphere inhabitants, including fungi, bacteria, and nematodes (14, 16, 40). Among these metabolites the polyketide compound 2,4-diacetylphloroglucinol (DAPG) has received particular attention because of its broad-spectrum antifungal, antibacterial, and antihelminthic activity (14, 16, 19, 40). Numerous studies have demonstrated the determinative role of DAPG production in the suppression of a variety of soilborne diseases by fluorescent pseudomonads (14, 16, 40). Evidence for a role of DAPG in disease suppression has mainly come from studies on DAPG-deficient mutants which gave reduced plant protection and from the demonstration of DAPG production in the rhizosphere (3, 19, 24, 28, 35). DAPG-producing pseudomonads are commonly found in the rhizosphere of important crops such as cucumber, maize, pea, tobacco, tomato, and wheat (18, 22, 29, 33, 36, 43), and they have been shown to be an important biological component of the natural suppressiveness of certain agricultural soils to take-all disease of wheat (7, 33) and black root of tobacco (36).

The DAPG-biosynthetic locus has been identified in Pseudomonas fluorescens strains Q2-87, CHA0, and F113. The locus comprises the DAPG-biosynthetic genes phlACBD, which are flanked upstream by the divergently transcribed phlF gene, encoding a transcriptional regulator, and downstream by the phlE gene, coding for a putative export protein (2, 6, 15, 38). Regulation of DAPG biosynthesis is complex, and some of the regulatory elements involved have been identified recently. Pathway-specific control is brought about by the TetR-like regulatory protein PhlF, which represses the expression of the DAPG-biosynthetic operon by binding to the phlA promoter region (1, 6, 14, 15, 38). DAPG itself acts as the derepressing signal by dissociating the repressor PhlF from the phlA promoter (1); as a consequence, DAPG autoinduces its own biosynthesis (1, 15, 38). An additional TetR-like regulator, PhlH, encoded by the phlH gene, located downstream of phlF, has been identified as a positive regulator of DAPG synthesis in strain CHA0, but its precise function remains to be determined (14, 38). Further elements involved in control of DAPG biosynthesis include several sigma factors and a regulatory cascade controlled by the two-component system GacS/GacA (14).

In the natural rhizosphere environment, numerous biotic and abiotic factors may influence production of DAPG. These include mineral and carbon sources as well as metabolites released by bacteria, fungi, and plants (11, 14). Recent findings suggest that in particular certain phenolic compounds may function as signals affecting antibiotic gene expression in biocontrol pseudomonads. For instance, while exogenous DAPG acts as a positive signal stimulating DAPG biosynthesis (see above), other bacterial metabolites such as salicylate and pyoluteorin strongly repress DAPG production in P. fluorescens CHA0 (38). Fusaric acid, a toxin and pathogenicity factor produced by the phytopathogenic fungus Fusarium oxysporum, is also a potent inhibitor of DAPG production in strain CHA0 (10, 38).

In the present study, we have chosen DAPG as an example to get more insight in the signaling events influencing the expression of antibiotic-biosynthetic genes in Pseudomonas populations in the rhizosphere. Since different genotypes of DAPG producers co-occur naturally in the rhizosphere (18, 21, 22, 29, 34, 36, 43), we have chosen to study a model interaction between the two well-characterized DAPG producers P. fluorescens CHA0 (38, 42) and P. fluorescens Q2-87 (2, 41), each strain belonging to a distinct genotypic group (18, 21, 43). In both strains, DAPG is a major determinant of the biocontrol activity against take-all disease of wheat (19, 41). Our aim was to investigate whether DAPG produced by one strain may be perceived as a positive signal stimulating phl gene expression in the other strain. For this purpose, we monitored the expression of phlA-lacZ reporter constructs in mixtures of strain CHA0 and Q2-87 in vitro and in the rhizosphere of wheat.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

The bacterial strains and plasmids used in this study are described in Table 1. P. fluorescens strains were routinely cultivated on King's medium B (KB) agar (20), in KB broth, on Luria-Bertani (LB) agar (37), or in LB broth at 27°C, unless otherwise specified. Escherichia coli was grown on LB agar or in LB broth at 37°C. When required, tetracycline hydrochloride (125 μg/ml) and kanamycin sulfate (25 μg/ml) were added to the growth media.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant genotype and/or phenotypea	Reference(s) or source
E. coli DH5α	Laboratory strain	37
P. fluorescens
CHA0	Wild type; DAPG+	42
CHA631	CHA0 ΔphlA; DAPG−	38
Q2-87	Wild type; DAPG+	41
Q2-87::Tn5-1	Q2-87 phlD::Tn5; DAPG− Kmr	2, 41
Plasmids
pME6016	Cloning vector for construction of transcriptional lacZ fusions; Tcr	38
pME607E	Reporter for DAPG-biosynthetic gene expression in P. fluorescens Q2-87; pME6016 with a 0.7-kb phlA upstream fragment of Q2-87 and a phlA-lacZ fusion; Tcr	This study
pME6259	Reporter for DAPG-biosynthetic gene expression in P. fluorescens CHA0, containing a 0.7-kb phlA upstream fragment of CHA0 and a phlA-lacZ fusion; Tcr	38

^aKm^r, kanamycin resistance; Tc^r, tetracycline resistance.

Construction of phlA-lacZ reporters.

Standard techniques were used for extraction of chromosomal and plasmid DNA, restriction enzyme digestions, PCR, ligations, electrophoresis, and transformations (37, 38, 39). For the construction of pME607E (Table 1), a 0.9-kb fragment containing the Q2-87 phlACBD promoter region (i.e., the intergenic region between phlA and phlF) (14) was generated by PCR amplification (3 min at 95°C; 25 cycles of 1 min at 95°C, 30 s at 58°C, and 1 min at 72°C; and 5 min at 72°C) with Promega Pfu DNA polymerase (Catalysis, Wallisellen, Switzerland), forward primer Q2-87phlF (5′-TCGTGAATTCGCAGCAGGAAATTGAGGTTC-3′), and reverse primer Q2-87phlA (5′-GTCTGAATTCGGGAGTCATATGGGTTGGTG-3′). Primers each contained an artificial restriction site (italics) for EcoRI and were developed with sequence data available from GenBank (accession no. U41818). The PCR product was purified with the QIAquick gel extraction kit (Qiagen AG, Basel, Switzerland), digested with EcoRI, and ligated into the EcoRI site of pME6016, a cloning vector for construction of transcriptional lacZ fusions which is derived from the rhizosphere-stable plasmid pME6010 (38). Absence of mutations due to PCR techniques was confirmed by sequencing after ligation. The plasmid was introduced into Q2-87 and its derivatives by electroporation. The stability of the construct was checked by repeated subculturing in LB broth without antibiotic selection. After 65 generations, retention of the plasmid was still 92% ± 6%. The lacZ reporter construct on pME6259 (Table 1) used to monitor phlA gene expression in P. fluorescens strain CHA0 and its derivatives has been previously described in detail (38).

Influence of DAPG on phlA expression in P. fluorescens CHA0 and Q2-87.

Strains CHA0 and CHA631, both carrying a phlA-lacZ fusion on plasmid pME6259, as well as strains Q2-87 and Q2-87::Tn5-1, carrying a phlA-lacZ fusion on plasmid pME607E, were grown in LB broth at 30°C for 4.5 h with shaking at 180 rpm. Cells from these cultures were harvested, washed twice with sterile distilled water, and adjusted to an optical density at 600 nm (OD₆₀₀) of 0.01. For inoculation, 20-μl aliquots of the adjusted cell suspensions were added to 20 ml of KB broth without selective antibiotics in 100-ml Erlenmeyer flasks. When required, synthetic DAPG (Toronto Research Chemicals Inc., North York, Ontario, Canada) dissolved in methanol (20 μl per flask) was added to the growth medium to give a final concentration of 100 μM. Controls received the same amount of solvent. Flasks were sealed with cellulose stoppers and incubated at 27°C with rotational shaking at 180 rpm. Bacterial growth (OD₆₀₀) and β-galactosidase specific activities (quantified by the method of Miller [37]) of at least three independent cultures per experiment were monitored over a period of 20 h. The experiment was repeated three times, and a representative experiment for each strain is shown in Fig. 1 and 2.

FIG. 1.

Induction by DAPG of phlA-lacZ expression in growing cultures of P. fluorescens CHA0 (○ and •) and its DAPG-negative mutant CHA631 (▵ and ▴), both harboring pME6259. β-Galactosidase activity and growth (OD₆₀₀) were determined for bacteria cultivated in KB broth at 27°C without (open symbols) or with (filled symbols) addition of 100 μM DAPG. Means ± standard deviations of three replicate cultures are presented. Some of the error bars are too small to be shown.

FIG. 2.

Induction by DAPG of phlA-lacZ expression in growing cultures of P. fluorescens Q2-87 (○ and •) and its DAPG-negative mutant Q2-87::Tn5-1 (▵ and ▴), both harboring pME607E. β-Galactosidase activity and growth (OD₆₀₀) were determined for bacteria cultivated in KB broth at 30°C without (open symbols) or with (filled symbols) addition of 100 μM DAPG. Means ± standard deviations of three replicate cultures are presented. Some of the error bars are too small to be shown.

Assay for induction of phlA expression in dual cultures.

The influence of DAPG produced by one Pseudomonas strain on phlA expression in another Pseudomonas strain was tested in a dual-culture assay allowing a free exchange of metabolites, with the bacterial strains growing in the same medium but physically separated by a dialysis membrane. Effector strains (i.e., the DAPG producers CHA0 and Q2-87 or their DAPG-negative mutants CHA631 and Q2-87::Tn5-1, respectively) were grown outside, and reporter strains carrying a phlA-lacZ construct (i.e., CHA631/pME6259 and Q2-87::Tn5-1/pME607E) were grown inside a dialysis tube as described below. Reporter gene expression in the DAPG-negative mutants CHA631 and Q2-87::Tn5-1 was monitored to exclude potential effects of the bacteria's own production of DAPG. Inocula were prepared from bacterial strains grown in LB broth for 6 h at 27°C with shaking at 160 rpm.

For the dual-culture assay, Servapor dialysis tubes made up of regenerated cellulose (pore diameter, 25 Å; molecular weight cutoff, 12,000 to 14,000; Serva, Heidelberg, Germany) were cut to a length of 20 cm and attached to plastic funnels having a lid. Each tube was placed into a 300-ml Erlenmeyer flask which was partially filled with 275 ml of KB broth, and one funnel was attached in the upper part of the flask. Then 30 ml of KB broth was added to each dialysis tube. The flasks were sealed with cotton stoppers and autoclaved at 121°C for 20 min. The medium outside the dialysis tubes was inoculated with 275 μl of the LB broth cultures of the effector strains adjusted to an OD₆₀₀ of 0.01 with sterile distilled water. The medium inside the tubes was inoculated with 25 μl of the LB broth cultures of the reporter strains adjusted to an OD₆₀₀ of 0.1. For experiments with strain Q2-87 and its derivative as effectors and CHA631/pME6259 as the reporter, effector strains were inoculated 8 h prior to the reporter strain. For experiments involving CHA0 and its derivative as effectors and Q2-87::Tn5-1/pME607 as the reporter, effector and reporter strains were inoculated at the same time. Flasks were incubated at 27°C with rotational shaking at 110 rpm. Samples inside and outside the dialysis tubes were taken under sterile conditions over a period of 60 to 80 h. Bacterial growth was determined by measuring OD₆₀₀, and β-galactosidase activities of phlA-lacZ on pME6259 or pME607E were quantified by the method of Miller (37). To quantify DAPG concentration in the medium, samples taken from outside the tubes were extracted with ethyl acetate and analyzed by high-pressure liquid chromatography as described elsewhere (19, 28). At the end of the experiment the medium inside the tubes was also assessed for DAPG content. To check for potential cross-contamination, dilutions of samples from inside and outside the dialysis tubes were plated on KB agar and KB agar containing tetracycline. Experiments were repeated twice with three replicates per treatment and experiment. One representative experiment for each reporter strain is presented in Fig. 3 and 4.

FIG. 3.

Induction of phlA-lacZ expression in P. fluorescens CHA631 (ΔphlA) carrying pME6259 by cocultivation with the DAPG producer P. fluorescens Q2-87. β-Galactosidase activity (A) and growth (OD₆₀₀) (B) were determined for CHA631/pME6259 grown in dual culture with strain Q2-87 (black circles) or its DAPG-negative derivative Q2-87::Tn5-1 (gray circles) as effector strains or in the absence of an effector strain (white circles). (C) Growth (OD₆₀₀) of the effector strains Q2-87 (▴) and Q2-87::Tn5-1 (▵), as well as DAPG production by strain Q2-87 (♦). Effector and reporter strains were cultivated physically separated by a dialysis membrane in KB broth at 27°C. Means ± standard deviations of three replicate cultures are presented.

FIG. 4.

Induction of phlA-lacZ expression in P. fluorescens Q2-87::Tn5-1 (phlD::Tn5) carrying pME607E by cocultivation with the DAPG producer P. fluorescens CHA0. β-Galactosidase activity (A) and growth (OD₆₀₀) (B) were determined for strain Q2-87::Tn5-1/pME607E grown in dual cultures with CHA0 (black circles) or its DAPG-negative derivative CHA631 (gray circles) as effector strains or in the absence of an effector strain (white circles). (C) Growth (OD₆₀₀) of the effector strains CHA0 (▴) and CHA631 (▵) and DAPG production by strain CHA0 (♦). Effector and reporter strains were cultivated physically separated by a dialysis membrane in KB broth at 27°C. Means ± standard deviations of three replicate cultures are presented.

Assay for induction of phlA expression on wheat roots.

To determine whether DAPG could be perceived as a signal also on roots, effector strains (i.e., DAPG producers CHA0 and Q2-87 and DAPG-negative mutants CHA631 and Q2-87::Tn5-1) and reporter strains carrying a phlA-lacZ construct (i.e., CHA0/pME6259, CHA631/pME6259, Q2-87/pME607E, and Q2-87::Tn5-1/pME607E) were tested in coinoculation experiments in the rhizosphere of wheat grown under gnotobiotic conditions. The gnotobiotic system has been described in detail elsewhere (19, 27). Briefly, 300 g of artificial soil was added to 1-liter Erlenmeyer flasks. The artificial soil consisted of pure vermiculite (expanded with 30% H₂O₂), quartz sand with different sizes of particles, and quartz powder mixed at a ratio of 10:70:20 (wt/wt/wt) and moistened with 10% (wt/wt) distilled water (17). Flasks were sealed with cotton wool stoppers and autoclaved at 121°C for 20 min. Bacterial effector and reporter strains were grown in LB broth for 8 h at 27°C with rotational shaking at 160 rpm. Cells were harvested by centrifugation, washed twice, and diluted with sterile 0.9% NaCl solution to the cell density required for seedling inoculation. Seeds of winter wheat (Triticum aestivum cv. Arina) were surface disinfected for 10 min in sterile distilled water heated to 50°C, followed by 30 min in 1% (vol/vol) sodium hypochlorite. Then they were thoroughly rinsed with sterile distilled water. Seeds were pregerminated for 2 days on 0.85% water agar at 24°C in the dark. For inoculation, seedlings were incubated for 1 h in bacterial suspensions containing 5 × 10⁷ CFU of a single strain (phlA-lacZ reporter) or a mixture of two strains (phlA-lacZ reporter plus effector strain)/ml at the following concentrations. For mixtures of CHA0/pME6259 or CHA631/pME6259 as the reporter strain with Q2-87 or Q2-87::Tn5-1 as the effector strain, the suspension contained 5 × 10⁷ reporter cells and 5 × 10⁸ effector cells per ml. For mixtures of Q2-87/pME607E or Q2-87::Tn5-1/pME607E as the reporter strain with CHA0 or CHA631 as the effector strain, the suspension contained 5 × 10⁷ reporter cells and 5 × 10⁷ effector cells per ml. Strain Q2-87 appeared to be a less competitive root colonizer than strain CHA0. When used as effectors, Q2-87 and its derivative therefore were applied at a 10-times-higher concentration to allow for induction of phlA expression. The β-galactosidase activity of the bacterial suspensions used for seedling inoculation never exceeded 0.8 U per 10⁷ CFU. Five seedlings were then transferred to each flask, covered with soil, and amended with 25 ml of sterile Knop's modified nutrient solution (17). Flasks were arranged in a completely randomized design in a growth chamber at 70% relative humidity and 18°C with light (160 μmol/s/m²) for 16 h, followed by an 8-h dark period at 13°C.

At 2 and 14 days after inoculation, plants were removed from the flasks, gently shaken to discard loosely adherent soil, and briefly washed with autoclaved water to remove tightly adhering soil. Roots from one flask were pooled and placed into a 50-ml Erlenmeyer flask containing 10 ml of sterile 0.9% NaCl solution. Flasks were rigorously shaken for 20 min on a rotary shaker at 400 rpm. Aliquots of 2 ml of the resulting suspensions were transferred to Eppendorf tubes and kept on ice until further use. β-Galactosidase activity was quantified according to the method of Miller (37) using 200 μl of the root suspensions. Units per 10⁷ CFU were calculated with the following formula: 1,000 × OD₄₂₀/(time in minutes × sample volume in milliliters) = units of β-galactosidase (27, 28). Control treatments of plants inoculated with strains which did not contain the reporter plasmids were used to subtract background activity from calculations. Bacterial root colonization was determined by plating serial dilutions of the root suspensions on KB agar for total bacterial root colonization and onto KB agar supplemented with tetracycline for root colonization of phlA reporter strains. Stability of plasmids pME6259 and pME607E was determined by replica plating of colonies grown on KB agar onto KB agar supplemented with tetracycline. After incubation for 48 h at 27°C, the percentage of tetracycline-resistant colonies, indicative of the presence of pME6259 or pME607E, was evaluated.

The experiment was repeated twice with CHA0 derivatives as reporters and three times with Q2-87 derivatives as reporters. Experiments consisted of at least four replicate flasks (each flask containing five plants) per treatment. Data of all trials together were first analyzed for trial-by-treatment interaction by analysis of variance using Systat, version 9.0 (Systat Inc., Evanston, Ill.). Data from different trials could not be pooled, and therefore individual trials are presented (see Tables 2 and 3). Means were separated with Fisher's protected least significant difference (LSD) test (LSD at P = 0.05).

TABLE 2.

Induction of phlA expression in P. fluorescens CHA0 and its DAPG-negative derivative CHA631 on wheat roots in the presence of the DAPG producer P. fluorescens Q2-87

Expt and strain(s)a	β-Galactosidase activity (U/107 CFU) and root colonization (107 CFU/g) after:b
	2 days			14 days
	β-Galactosidase activity	Root colonization		β-Galactosidase activity	Root colonization
		CHA0 derivatives	Q2-87 derivatives		CHA0 derivatives	Q2-87 derivatives
1
CHA0/pME6259	17.3†	6.0†	ND	11.3†	29.8†	ND
CHA0/pME6259 + Q2-87	29.3‡	2.5‡	8.1*†	12.5†	24.8†‡	23.7†
CHA0/pME6259 + Q2-87::Tn5-1	17.6†	3.1‡	12.4*†	9.2†	19.4‡	26.5†
CHA631/pME6259	7.6†	7.4†	ND	5.3†	18.0†	ND
CHA631/pME6259 + Q2-87	15.8‡	3.2‡	6.5†	6.3†	10.2‡	18.1†
CHA631/pME6259 + Q2-87::Tn5-1	8.5†	2.0‡	7.4†	6.5†	11.4‡	24.2†
2
CHA631/pME6259	3.8†	12.4†	ND	4.3†	29.6†	ND
CHA631/pME6259 + Q2-87	7.7‡	4.5‡	6.8†	3.9†	14.6‡	35.8†
CHA631/pME6259 + Q2-87::Tn5-1	3.3†	2.1‡	8.3*†	3.7†	10.7‡	42.3*†

^aStrains CHA0 (wild type, DAPG⁺) and CHA631 (ΔphlA, DAPG⁻) carrying the phlA-lacZ reporter construct on pME6259 were inoculated onto wheat seedlings, alone or in combination (at a 1:10 ratio) with either strain Q2-87 (wild type, DAPG⁺) or strain Q2-87::Tn5-1 (phlD::Tn5, DAPG). Plants were grown under gnotobiotic conditions. Experiments 1 and 2 are independent.

^bValues represent the means of four replicate flasks, each flask containing five plants. Means for the same set of three treatments (without effector, with Q2-87, with Q2-87::Tn5-1) within a column followed by a different symbol (dagger or double dagger) are significantly different (P ≤ 0.05) according to Fisher's protected LSD test. *, significantly different (P ≤ 0.05) from population sizes of coinoculated CHA0 derivatives. ND, not detected.

TABLE 3.

Induction of phlA expression in P. fluorescens strain Q2-87 and its DAPG-negative mutant Q2-87::Tn5-1 on wheat roots in the presence of DAPG-producing P. fluorescens CHA0

Expt and strain(s)a	β-Galactosidase activity (U/107 CFU) and root colonization (107 CFU/g) afterb:
	2 days			14 days
	β-Galactosidase activity	Root colonization		β-Galactosidase activity	Root colonization
		Q2-87 derivatives	CHA0 derivatives		Q2-87 derivatives	CHA0 derivatives
1
Q2-87/pME607E	24.7†	4.0†	ND	57.8†	12.5†	ND
Q2-87/pME607E + CHA0	48.9‡	0.3‡	8.8*†	46.9†	2.8‡	21.5*†
Q2-87/pME607E + CHA631	19.9†	0.7‡	13.6*†	44.0†	3.3‡	19.3*†
Q2-87::Tn5-1/pME607E	5.2†	5.8†	ND	10.1†	14.8†	ND
Q2-87::Tn5-1/pME607E + CHA0	15.4‡	0.5‡	9.0*†	13.5†	3.4‡	20.6*†
Q2-87::Tn5-1/pME607E + CHA631	7.9†	0.5‡	7.5*†	12.0†	2.0‡	13.1*†
2
Q2-87::Tn5-1/pME607E	3.9†	1.6†	ND	14.4†	3.0†	ND
Q2-87::Tn5-1/pME607E + CHA0	11.8‡	0.2‡	4.1*†	27.9‡	1.6†	23.3*†
Q2-87::Tn5-1/pME607E + CHA631	4.2†	0.2‡	7.3*†	10.3†	2.9†	13.7*†
3
Q2-87::Tn5-1/pME607E	8.9†	0.9†	ND	49.3†	1.9†	ND
Q2-87::Tn5-1/pME607E + CHA0	26.4‡	0.1‡	2.6*†	119.4‡	0.9†	4.6*†
Q2-87::Tn5-1/pME607E + CHA631	10.5†	0.1‡	4.6*‡	62.3†	1.4†	6.4*†

^aStrains Q2-87 (wild type, DAPG⁺) and Q2-87::Tn5-1 (phlD::Tn5 DAPG⁻) carrying the phlA-lacZ reporter construct on pME607E were inoculated onto wheat seedlings, alone or in combination (at a 1:1 ratio) with either strain CHA0 (wild type, DAPG⁺) or strain CHA631 (ΔphlA, DAPG⁻). Plants were grown under gnotobiotic conditions. Experiments 1, 2, and 3 are independent.

^bValues represent the means of four replicate flasks, each containing five plants. Means for the same set of three treatments (without effector, with CHA0, with CHA631) within a column followed by a different symbol (dagger or double dagger) are significantly different (P ≤ 0.05) according to Fisher's protected LSD test. *, significantly different (P ≤ 0.05) from population sizes of coinoculated Q2-87 derivatives. ND, not detected.

RESULTS

DAPG induces phlA expression in P. fluorescens strains CHA0 and Q2-87.

Expression of a phlA-lacZ reporter fusion carried by pME6259 in strain CHA0 growing in KB broth gradually increased to levels of about 1,500 Miller units at an OD₆₀₀ of 2.5 (Fig. 1). In the DAPG-negative mutant CHA631, phlA expression was lower than in the wild-type strain during the exponential growth phase but reached similar levels when cells had grown to an OD₆₀₀ higher than 1.8. The addition of 100 μM synthetic DAPG to KB broth advanced and enhanced (up to sixfold) phlA-lacZ expression in strain CHA0 (Fig. 1). In the DAPG-negative mutant, phlA expression was also induced by DAPG, although to a lesser extent than in the wild type. In strain Q2-87, expression of a phlA-lacZ fusion carried by pME607E increased during exponential growth and attained levels of 7,000 to 8,000 Miller units in the early stationary growth phase (Fig. 2). Expression of phlA was significantly lower in the DAPG-negative mutant Q2-87::Tn5-1. The addition of 100 μM DAPG to the growth medium induced phlA expression in both the wild type and the mutant (Fig. 2). Remarkably, β-galactosidase activity in Q2-87::Tn5-1 grown in the presence of DAPG was increased up to 10-fold during the exponential growth phase; however, final levels were similar to those for the wild type. Bacterial growth was not affected by the addition of DAPG (data not shown). Taken together, these results show that exogenous DAPG added at a physiological concentration (Fig. 3 and 4) (38) can significantly advance and enhance expression of the biosynthetic gene phlA in the two biocontrol pseudomonads CHA0 and Q2-87.

Induction of phlA expression in strains CHA631 and Q2-87::Tn5-1 by DAPG producers in dual-culture experiments.

It was then of interest to see whether DAPG excreted by one pseudomonad (termed the effector strain) could affect expression of DAPG-biosynthetic genes in another pseudomonad (termed the reporter strain). The two pseudomonads were physically separated by growing the reporter strain in KB broth inside a dialysis tube and the effector strain in KB broth surrounding the dialysis tube. DAPG-negative reporters were used to exclude potential effects of the strains' own DAPG production. When strain CHA631 carrying pME6259 was grown inside the tube, phlA-lacZ expression slowly increased during exponential growth and reached a maximum of 984 ± 72 Miller units at an OD₆₀₀ of 2.4 (Fig. 3A). In the presence of the DAPG-producing strain Q2-87 growing in the medium outside the dialysis tube, phlA expression in strain CHA631 was enhanced three- to eightfold throughout the experiment (Fig. 3A). In contrast, no such induction occurred when the reporter strain was exposed to the DAPG-negative derivative Q2-87::Tn5-1 (Fig. 3A). No β-galactosidase activity in the medium surrounding the dialysis tubes could be measured, indicating that no CHA631/pME6259 cells had passed through the dialysis membranes. The growth of the reporter strain was significantly reduced by the presence of either effector strain (Q2-87 or Q2-87::Tn5-1) in the medium surrounding the tube (Fig. 3B). When grown alone, the reporter CHA631/pME6259 inside the dialysis tube reached population densities similar to those of effector strains (Fig. 3B and C). It is possible that reporter strains growing inside the tubes under less-optimal conditions were more sensitive to nutrient competition from the effector strains. DAPG production by strain Q2-87 in the medium outside the dialysis tubes reached maximum levels of almost 50 μM (Fig. 3C). At the end of the experiment, the DAPG concentrations inside and outside the tubes were both about 35 μM, indicating that DAPG was partly degraded (38).

Similar results were obtained when Q2-87::Tn5-1 carrying a phlA-lacZ fusion on pME607E was the reporter strain inside the dialysis tube and CHA0 or CHA631 was the effector strain growing in the medium surrounding the tubes (Fig. 4). In the absence of effector strains, the maximal level of phlA expression in Q-87::Tn5-1 was 925 ± 151 Miller units when cells had grown to an OD₆₀₀ of 0.9, and β-galactosidase activity remained at this level until the end of the experiment (Fig. 4A). The growth of Q2-87::Tn5-1 inside the tube was strongly inhibited in the presence of strain CHA0 or its DAPG-negative derivative CHA631 grown outside the tube (Fig. 4B). Nevertheless, phlA expression in Q2-87::Tn5-1 was induced up to threefold when the strain was grown in the presence of DAPG-producing CHA0, whereas the DAPG-negative mutant CHA631 had no such effect (Fig. 4A). DAPG production by strain CHA0 in the surrounding medium reached a maximal level of 19.5 ± 4.0 μM after 64 h of incubation (Fig. 4C). The DAPG concentrations measured at the end of the experiment, after 74 h, were 12.5 ± 1.1 μM outside the dialysis tube and 11.9 ± 1.1 μM inside the dialysis tube.

Thus, the presence of a DAPG-producing, but not of a DAPG-negative, strain greatly increased the expression of a phlA-lacZ reporter in strains CHA631 and Q2-87::Tn5-1, indicating that DAPG released by one strain can be perceived as a signal inducing expression of DAPG-biosynthetic genes in the other strain.

Induction of phlA expression in CHA0 and Q2-87 and their derivatives on wheat roots by coinoculation with DAPG-producing strains.

To test whether DAPG-mediated cross talk between biocontrol strains of Pseudomonas occurs also in situ, the influence of the DAPG producer Q2-87 on phlA expression in wild-type CHA0 and its DAPG-negative mutant CHA631, both carrying the reporter plasmid pME6259, on wheat roots was monitored under gnotobiotic conditions. Two days after planting, expression of the phlA-lacZ reporter in strain CHA0 was about 17 U per 10⁷ CFU in the absence of effector strains (Table 2). Coinoculation with the DAPG-producing strain Q2-87 enhanced phlA expression in strain CHA0 by about 70%, whereas coinoculation with the DAPG-negative mutant Q2-87::Tn5-1 had no such effect (Table 2). Twelve days later, however, the inducing effect of Q2-87 on phlA expression in strain CHA0 had disappeared (Table 2). In the DAPG-negative derivative CHA631, phlA expression in general was lower than that in the wild-type CHA0 at both time points (Table 2). However, induction of phlA expression following coinoculation with Q2-87 was stronger in CHA631 than in wild-type CHA0 (Table 2). Reporter gene expression in CHA631 was not altered by the presence of the DAPG-negative mutant Q2-87::Tn5-1. The inducing effect of strain Q2-87 could no longer be observed 14 days after inoculation (Table 2).

When CHA0 or CHA631 carrying pME6259 was inoculated alone on wheat seedlings, population densities varied between 0.6 × 10⁸ and 1.2 × 10⁸ CFU per g of roots 2 days after planting and between 1.8 × 10⁸ and 3.0 × 10⁸ CFU per g of roots 14 days after planting (Table 2). Coinoculation with Q2-87 or Q2-87::Tn5-1 reduced root colonization levels of CHA0 derivatives two- to sixfold after 2 days and by about 20 to 60% after 14 days (Table 2). In coinoculation treatments, CHA0 and Q2-87 derivatives were inoculated at a ratio of 1:10 at the beginning of the experiment. The assessment of root colonization levels 14 days later revealed that in general there were no longer any significant differences between the population sizes of Q2-87 and CHA0 derivatives (Table 2). The retention levels of plasmid pME6259 in CHA0 and CHA631 cells were still 97.8% ± 0.8% and 98.1% ± 1.1%, respectively, at the end of the experiment.

In a second set of experiments, wild-type Q2-87 and its DAPG-negative derivative Q2-87::Tn5-1, both carrying a phlA-lacZ construct on plasmid pME607E, were used as reporter strains for DAPG produced by strain CHA0. Two days after planting the wheat seedlings, phlA expression in the absence of effector strains was severalfold higher in strain Q2-87 than in its derivative Q2-87::Tn5-1 (Table 3). When the reporter strains were coinoculated with the DAPG producer CHA0, phlA expression was enhanced twofold in the wild-type Q2-87 and threefold in the mutant Q2-87::Tn5-1 (Table 3). In contrast, phlA gene expression in the reporter strains was not influenced by the presence of strain CHA631, which is unable to produce DAPG (Table 3). In experiments 2 and 3, the inducing effect of CHA0 on phlA expression in Q2-87::Tn5-1 persisted until the end of the experiment (Table 3).

When inoculated alone, Q2-87 and Q2-87::Tn5-1 carrying pME607E had population densities that varied between 0.9 × 10⁷ and 5.8 × 10⁷ CFU per g of wheat roots after 2 days and that reached 1.9 × 10⁷ to 14.8 × 10⁷ CFU per g of roots after 14 days (Table 3). When these strains were coinoculated with strain CHA0 or CHA631, population densities of Q2-87 derivatives were markedly (6- to 13-fold) reduced (Table 3). After 14 days these differences had disappeared in two out of three experiments. At the beginning of the experiment, Q2-87 and CHA0 derivatives were coinoculated at a ratio of 1:1. The ratio for the population density of Q2-87 to that of CHA0 derivatives drastically changed to between 1:15 and 1:46 at 2 days after inoculation (Table 3). After 14 days, CFU counts of Q2-87 derivatives were still greatly outnumbered (by five to 15 times) by those of CHA0 derivatives, indicating that Q2-87 may be a less efficient root colonizer than CHA0. At the end of the experiment, 98.2% ± 0.9% and 98.5% ± 2.2% of Q2-87 and Q2-87::Tn5-1 cells, respectively, still carried plasmid pME607E.

In summary, these coinoculation experiments demonstrate that DAPG is exchanged as a signal between the biocontrol pseudomonads CHA0 and Q2-87, thereby activating expression of biosynthetic genes for this antifungal compound on wheat roots, i.e., at the site of pathogen suppression.

DISCUSSION

DAPG as a signal compound in the rhizosphere.

Numerous biotic and abiotic signals can influence the expression of antimicrobial metabolites in fluorescent pseudomonads colonizing the rhizosphere (11, 14, 28). In the present interpopulation signaling study, we provide evidence for the first time that DAPG produced in a mixed population of the two genotypically distinct pseudomonads CHA0 and Q2-87 can be perceived as a positive signal boosting the expression of DAPG-biosynthetic genes in the rhizosphere of wheat. Several lines of evidence support our finding. First, addition of exogenous DAPG at physiological levels to growing cultures of CHA0 and Q2-87 strongly enhanced expression of phlA in both strains (Fig. 1 and 2). This autoinduction of DAPG biosynthesis was most evident in DAPG-negative CHA0 and Q2-87 mutants, in which addition of DAPG at least partially restored phlA expression (Fig. 1 and 2). Together with the results from previous studies (1, 38), the present findings thus establish positive autoregulation as a conserved control mechanism of DAPG biosynthesis in P. fluorescens. Second, in coinoculation experiments the DAPG producer Q2-87 induced phlA gene expression in strain CHA0 or its phlA mutant CHA631 in liquid culture (Fig. 3) and on wheat roots (Table 2). Vice versa, strain CHA0 had the same effect on Q2-87 and its DAPG-negative derivative (Fig. 4; Table 3). Third, DAPG-negative mutants were unable to enhance expression of DAPG-biosynthetic genes in a coinoculated reporter strain (Fig. 3 and 4; Tables 2 and 3), suggesting that DAPG indeed was the signal mediating cross communication between the two pseudomonads. Fourth, strains CHA0 and Q2-87 were both capable of excreting DAPG in the medium used for the dual-culture assay (Fig. 3 and 4) and in the rhizosphere of wheat (3, 19).

In recent years, there has been increasing interest in gaining better insight into the mechanisms by which biocontrol pseudomonads communicate among each other and how they interact with the rhizosphere microbial community. Some of the molecules involved in rhizosphere signaling and regulation of antibiotic production by biocontrol pseudomonads have been identified. N-Acyl-homoserine lactones (AHLs) are currently the best-documented class of these signal molecules. They are used by many plant-associated bacteria for the control of gene expression and cell-cell communication in a cell density-dependent manner termed quorum sensing (32). In the biocontrol strains Pseudomonas aureofaciens 30-84 and Pseudomonas chlororaphis PCL1391 the production of phenazine antibiotics, an important trait for pathogen suppression, is under quorum sensing control (4, 32). Other rhizobacteria may positively or negatively affect AHL signaling in these biocontrol strains (25, 31). Here we present a novel example of positive interpopulation signaling between rhizosphere bacteria. We show that DAPG-producing biocontrol pseudomonads can interact synergistically by stimulating each other in the expression of their DAPG-biosynthetic genes. The signal molecule is DAPG itself. The situation is reminiscent of AHL signaling to a certain extent inasmuch as DAPG also induces its own biosynthesis and acts as a diffusible signal at intra- and interpopulation levels. Interestingly, DAPG appears to act also on the expression of other biocontrol traits since it strongly represses the expression of biosynthetic genes for pyoluteorin, another potent antifungal compound produced by some pseudomonads (14). Other examples of signaling in the rhizosphere are rare. An exciting example of cross talk in the rhizosphere is provided by fusaric acid, a pathogenicity factor of F. oxysporum f. sp. radicis-lycopersici that represses phlA expression and DAPG production in P. fluorescens CHA0 and thereby abolishes the bacterium's capacity to suppress tomato root and crown rot caused by the pathogen (10, 38). Nonpathogenic isolates of Fusarium producing fusaric acid were also found to repress DAPG gene expression in the rhizosphere of wheat (27). Recent evidence suggests that besides DAPG and fusaric acid a number of other phenolic metabolites (e.g., pyoluteorin and salicylate) of microorganisms and plants may affect production of antimicrobial metabolites in P. fluorescens (14, 38).

Differences in the responsiveness of CHA0 and Q2-87 to signaling.

The two DAPG-producing biocontrol pseudomonads used in this study differ in their geographical and agricultural origins as well as in a number of ecological, physiological, and genetic characteristics. Strain Q2-87 was originally isolated from the rhizosphere of the monocot wheat grown in a U.S. take-all-suppressive soil (41), whereas strain CHA0 was isolated from roots of the dicot tobacco grown in a Swiss soil suppressive to black root rot (17, 42). In contrast to Q2-87, CHA0 produces the antimicrobial compound pyoluteorin in addition to DAPG (18). Based on amplified rDNA restriction analysis (ARDRA) and other molecular fingerprinting methods applied to a worldwide collection of DAPG producers (18, 21, 22, 43), strains CHA0 and Q2-87 belong to two clearly distinct genotypes. Strain CHA0 groups with members of the ARDRA1 group, which includes pseudomonads producing pyoluteorin in addition to DAPG, whereas strain Q2-87 clusters in the ARDRA2 group, comprising DAPG-producing pseudomonads that do not produce pyoluteorin (18, 43). Considering these differences between CHA0 and Q2-87, it seems not surprising that the two strains differ also in their responses to the signal molecule DAPG in the rhizosphere. This is illustrated by our observation that the enhancing effect of CHA0 on phlA expression in Q2-87::Tn5-1 on wheat roots was stronger and longer lasting (Table 3) than the effect of Q2-87 on expression of the corresponding gene in CHA631 (Table 2). A possible explanation for this effect is that other factors present in the rhizosphere might negatively affect expression of DAPG-biosynthetic genes in CHA0 but not in Q2-87. There is evidence for differential responses of the two strains to certain environmental signals, with strain CHA0 and other members of the ARDRA1 group being significantly more sensitive to repression of DAPG gene expression by phenolic compounds, such as salicylate and fusaric acid, than strain Q2-87 and other members of the ARDRA2 group of DAPG producers (8, 26; C. Keel, unpublished data).

There is a further striking difference in the behavior of the two biocontrol pseudomonads. Q2-87 seems to be significantly less competitive than CHA0 in the wheat rhizosphere since Q2-87 cells were greatly outnumbered by CHA0 cells at 2 and 14 days after seedling inoculation with a 1:1 mixture of the two strains (Table 3). This may be partly explained by the fact that Q2-87 and its derivatives, when inoculated alone, generally reached lower population densities on wheat roots than CHA0 and its derivatives (Tables 2 and 3), indicating that Q2-87 might be a weaker root colonizer in the plant assay used in this study. Earlier studies indicated that strain Q2-87 appears to be a less competitive root colonizer than strains of DAPG producers with other genotypes (21, 22, 34). It is also possible that the growth of Q2-87 was inhibited by certain metabolites released by strain CHA0. Previously, strain Q2-87 was reported to be inhibited by several other fluorescent pseudomonads in a plate assay in vitro (30). Another possibility is that strain Q2-87 is less efficient or even impaired in using certain substrates provided by the wheat root exudates (34). Interestingly, strains CHA0 and Q2-87 differ considerably in their profiles for utilization of 128 carbon sources (43). Effective adaptation to the utilization of root exudate components, such as carboxylic acids, sugars, and certain amino acids, is considered to be a prerequisite for rhizosphere competence (23).

Implications for biocontrol.

One approach to overcome inconsistent biocontrol activity is the application of combinations of different biocontrol agents. For plant-beneficial pseudomonads, strain mixtures and combinations with other bacteria or fungi often provided more-effective disease control than the application of an individual biocontrol pseudomonad alone (5, 9, 12, 30). One approach to obtain a successful microbial biocontrol consortium is to apply mixtures of biocontrol agents which display different disease-suppressive mechanisms that are complementary to each other. De Boer and coworkers (5) combined Pseudomonas strains effective in siderophore-mediated competition for iron and induction of systemic plant resistance to improve control of Fusarium sp.-mediated wilt of radish. Dunne et al. (13) applied a mixture of the DAPG producer P. fluorescens F113 and a proteolytic rhizobacterium to enhance suppression of Pythium sp.-mediated damping-off of sugar beet.

Another approach may be to combine strains which stimulate each other in the expression of relevant biocontrol genes. Here, we demonstrate that a combination of the two plant-protecting strains CHA0 and Q2-87 had a synergistic effect, resulting in enhanced expression of the genes for DAPG biosynthesis, a crucial trait in disease suppression. Such synergism may be relevant for biocontrol, since in this way the DAPG pool within a mixed Pseudomonas population in the rhizosphere may rapidly be boosted to levels that are relevant to pathogen control (35). Since both CHA0 and Q2-87 use DAPG to effectively suppress take-all of wheat (19, 41), it will be interesting to see whether a combination of the two strains results in improved DAPG production and control of the disease. To this end, our preliminary results from experiments carried out in a gnotobiotic take-all assay (19) indicated that the strain mixture was only slightly (but not significantly) superior to the individual strains in suppressing disease and that these effects were not always consistent (data not shown). It is possible that the use of a different plant-pathogen assay system or combinations of other DAPG-producing pseudomonads may have resulted in better biocontrol. In future approaches the natural diversity of different genotypes of DAPG-producing pseudomonads (18, 21, 34, 43) may be exploited to design strain combinations that result not only in enhanced and consistent DAPG production in various soil environments but also in improved growth, activity, and competitiveness in the rhizosphere. In addition, more attention should be paid to effects of the host plant and the soil environment, which both can strongly influence DAPG expression in pseudomonads (11, 28).