Autoinduction of 2,4-Diacetylphloroglucinol Biosynthesis in the Biocontrol Agent Pseudomonas fluorescensCHA0 and Repression by the Bacterial Metabolites Salicylate and Pyoluteorin

Isolation and characterization of the 2,4-DAPG-negative mutant CHA630.Following Tn5 mutagenesis of P. fluorescens strain CHA0, one out of 1,500 mutants tested was unable to produce detectable amounts of 2,4-DAPG and MAPG in YME and OSG media (Table 2). The mutant strain, CHA630, was compared with the wild-type strain CHA0 for HCN, extracellular protease, salicylate, pyochelin, and fluorescent pigment production, TSO activity, prototrophy, and growth characteristics in different media, and no difference was found (data not shown). However, the mutant produced small amounts of pyoluteorin in OSG medium, whereas strain CHA0 grown to the same cell density did not (Table 2). Southern hybridization showed that the mutant CHA630 contained a single Tn5 insertion (data not shown). The Tn5 insertion was located to position 3617 in an open reading frame (ORF) termedphlH (see below and Fig. 1).

Cloning of the wild-type genes surrounding the Tn5insertion in strain CHA630.The genomic region adjacent to Tn5 in strain CHA630 was obtained by using the two-step cloning-out procedure described by Schnider et al. (47). The suicide plasmid pME3049, which carries a mercury resistance and the kanamycin resistance determinant of Tn5, was mobilized into strain CHA630. The integration of pME3049 into the chromosome occurred at a frequency of 10⁻⁶ to 10⁻⁷ per donor. Genomic DNA of a recombinant with a chromosomally integrated plasmid was digested with BamHI, ligated, and used to transformE. coli DH5α. The resulting plasmid pME6250B (Fig. 1) carries a 1.7-kb genomic DNA insert. The chromosome of strain CHA630::pME3049 was also digested with SacI. Self-ligation and transformation of E. coli led to plasmid pME6250S (Fig. 1) carrying an extended insert of 11.1 kb downstream of the Tn5 insertion site. For the isolation of the wild-type genes, plasmid pME6250B was transferred into the wild-type strain CHA0. The chromosomal DNA of CHA0::pME6250B was digested withHindIII and ligated. Plasmid pME6251H (Fig. 1), containing a 9.8-kb genomic fragment, was selected.

Sequence analysis of the phlA, -F, -G, and -H genes.The nucleotide sequence of 4,010 bp (expanded region shown in Fig. 1) revealed two ORFs designated phlA and phlF, by analogy with homologous genes of P. fluorescens Q2-87 (2), as well as two new ORFs named phlG and phlH (Fig.1). This sequence contained 57.9% G+C. The deduced product (362 amino acids, 38.5 kDa) of the phlA gene of strain CHA0 showed 83.3% identity with its homolog PhlA, the first gene product of the 2,4-DAPG biosynthetic operon in strain Q2-87 (2). The intergenic region upstream of phlA, encompassing 462 nucleotides, is relatively AT rich (56.0%) and contains two inverted repeats (Fig. 1). The divergently oriented phlF gene appears to start at a GTG codon, and its product (200 amino acids, 22.9 kDa) is 84.5% identical over the entire sequence with PhlF, a repressor protein involved in the regulation of 2,4-DAPG biosynthesis, ofP. fluorescens Q2-87 (2) and F113 (GenBank accession no. AF129856 ). An identical, putative helix-turn-helix motif (³⁴GYX₃SIX₂VX₅ASXPXIYXWWXNKX₂L⁶⁴), which is typical of DNA-binding regulatory proteins, exists near the PhlF N terminus of strains CHA0, Q2-87, and F113. The average amino acid change per codon score (8) was 0.76; scores of >0.8 are considered to be highly indicative of a helix-turn-helix motif (8). Downstream of phlF there is an inverted repeat that could function as a transcription terminator (ΔG, −21.9 kcal mol⁻¹) (Fig. 1). The adjacent ORF, tentatively named phlG, could encode a protein, PhlG (320 amino acids, 36.3 kDa), which shows no similarities to protein sequences with known functions deposited in the SwissProt and the EMBL GenBank. The deduced product PhlH (223 amino acids, 24.3 kDa) of the adjacent ORF shows similarities to regulatory proteins, e.g., IfeR of Agrobacterium tumefaciens (AF039653; 34% identity), a putative transcriptional regulator of Streptomyces coelicolor (AL035569; 29% identity), AcrR, a potential AcrAB operon repressor of E. coli (P34000; 21% identity), and members of the TetR/AcrR family of a series of other microorganisms (data not shown). The similarities in the deduced amino acid sequences of these proteins are concentrated within approximately 100 residues of the N termini; beyond this region, the proteins bear little resemblance to one another (data not shown). The ORF phlH is followed by an inverted repeat (ΔG, −17.0 kcal mol⁻¹) (Fig. 1) likely to be a transcription terminator.

Construction of phlA and phlF mutants.A phlA in-frame deletion mutation was created in the chromosome of strain CHA0 (Fig. 1) as described in Materials and Methods. In contrast to the wild type, the phlA mutant CHA631 obtained did not produce measurable amounts of 2,4-DAPG and MAPG in OSG medium (Table 2). 2,4-DAPG and MAPG production in strain CHA631 could be restored by complementation with pME6261 (Table 2), a plasmid which carries phlA under its own promoter (Fig. 1), whereas introduction of the cloning vector pME6030 alone had no effect. These results are in accordance with the findings of Bangera and Thomashow (2) and confirm the requirement of the biosynthetic genephlA for 2,4-DAPG production.

A chromosomal phlF mutant, CHA638, was constructed by insertion of the transcription termination element Ω-Km. Compared with the wild-type strain CHA0, the mutant CHA638 overproduced 2,4-DAPG about fourfold when grown in OSG medium at 24°C (Table 2) or 30°C (data not shown) to early stationary phase at an OD at 600 nm of 3.5. For restoration of the disrupted gene, phlF⁺ was inserted as a single copy via a Tn7-based system into the chromosome of strain CHA638. In the resulting strain CHA638.phlF⁺ the formation of 2,4-DAPG and MAPG was repressed to levels that were lower than those in the wild type, for reasons that are not understood (Table 2). Nevertheless, these results provide experimental evidence for PhlF acting as a repressor of 2,4-DAPG biosynthesis.

The phlA mutant CHA631, but not the wild-type strain CHA0 and the phlF mutant CHA638, excreted the antibiotic pyoluteorin (about 4 μM) when grown in OSG medium (Table 2). In the iron-poor GCM medium, the wild-type and both mutant strains produced the same amounts of the siderophores salicylate (about 60 μM) and pyochelin (about 40 μM). In addition, the mutants were indistinguishable from the wild type in terms of HCN production, fluorescence, production of extracellular enzymes, and growth in different media (data not shown).

Growth-dependent phlA expression and 2,4-DAPG production, and bacterial degradation of 2,4-DAPG to MAPG.Expression of a phlA′-′lacZ translational fusion carried by pME6259 (Table 1; Fig. 1) was monitored in strain CHA0 growing in OSG medium (Fig. 2). Biosynthesis of 2,4-DAPG and MAPG was determined in parallel. Expression of phlAoccurred from the mid-exponential to the early stationary growth phase (Fig. 2). Thereafter, β-galactosidase activity slowly declined, as expected for the long half-life of the enzyme. The kinetics of 2,4-DAPG and MAPG production paralleled expression of the phlA′-′lacZreporter construct in strain CHA0 (Fig. 2). 2,4-DAPG and MAPG synthesis started at mid-exponential growth phase, and the compounds were accumulated until the beginning of stationary growth phase, with 2,4-DAPG reaching a maximum concentration of about 100 μM (Fig. 2). Thereafter, 2,4-DAPG and MAPG concentrations in the medium steadily decreased, due to degradation by the bacterium. This was confirmed by the observation that 100 μM synthetic 2,4-DAPG added to OSG medium remained stable for at least 1 week in the absence of bacteria but was no longer detectable after a 30-h incubation in the presence of the 2,4-DAPG- and MAPG-negative mutants CHA631 (Fig.3) or CHA630 (data not shown). Interestingly, degradation of 2,4-DAPG by strain CHA631 was succeeded by a temporary accumulation of MAPG in the growth medium (Fig. 3), suggesting that MAPG can be a breakdown product of 2,4-DAPG.

• Open in new tab
• Download powerpoint

Fig. 2.

Production of 2,4-DAPG (■) and MAPG (□) and expression of a phlA′-′lacZ translational fusion in a growing culture of P. fluorescens CHA0. Strain CHA0 carrying the phlA reporter construct on plasmid pME6259 was cultivated in OSG medium at 30°C. At different ODs at 600 nm (▵), antibiotic production and specific β-galactosidase activities (○) were determined as described in Materials and Methods. Means ± standard deviations from three experiments are shown. Some of the error bars are too small to be shown.

• Open in new tab
• Download powerpoint

Fig. 3.

Degradation of 2,4-DAPG to MAPG by a growing culture of the 2,4-DAPG- and MAPG-negative phlA mutant CHA631. Before bacterial inoculation, 100 μM 2,4-DAPG was added to OSG medium. After different incubation periods at 24°C, concentrations of 2,4-DAPG in absence of bacteria (○) and of 2,4-DAPG (●) and MAPG (▾) in the presence of CHA631 were determined. Means ± standard deviations from four experiments are shown.

Expression of phlA is induced by 2,4-DAPG.To identify signals inducing phlA expression, late-exponential-phase, cell-free culture supernatants from strain CHA0 were extracted with dichloromethane. The extract enhanced expression of the phlA′-′lacZ reporter 15- to 20-fold in a CHA0 background growing in NYB, a medium which does not support 2,4-DAPG and MAPG production. The inducing signal was purified by chromatography on silica gel and by preparative HPLC. A single active peak was isolated with a retention time (19.5 min) and a UV spectrum identical to that of authentic 2,4-DAPG (25). Addition of 300 μM synthetic 2,4-DAPG to an NYB culture induced phlA′-′lacZ expression in CHA0 to a similar level as did the crude dichloromethane extract (data not shown), thus confirming the identification of the signal.

The effect of added 2,4-DAPG on phlA′-′lacZ expression was further analyzed in OSG, a medium which supports production of up to 100 μM 2,4-DAPG by strain CHA0 (Fig. 2). Addition of 100 μM synthetic 2,4-DAPG to OSG medium inoculated with CHA0/pME6259 advanced the expression of phlA′-′lacZ to the early exponential growth phase, and phlA′-′lacZ expression during exponential growth was induced sixfold by 2,4-DAPG (Fig.4; Table3). However, phlA′-′lacZexpression at the end of the exponential growth phase was the same with or without addition of 2,4-DAPG (Fig. 4). A fivefold-higher concentration of 2,4-DAPG did not further promotephlA′-′lacZ expression (Table 3). Remarkably, 100 or 500 μM MAPG also induced phlA′-′lacZ expression, although to a significantly lesser extent than did 2,4-DAPG (Table 3). In NYB amended with 2,4-DAPG, the maximal level of phlA′-′lacZ expression in CHA0 was similar to that observed in OSG medium (data not shown), suggesting that exogenously added 2,4-DAPG is an effective signal. To test whether induction of phlA expression by 2,4-DAPG was specific, we also assayed derivatives of strain CHA0 carrying a chromosomal lacZ fusion either in the HCN biosynthetic genehcnA (strain CHA207) or fused to an artificial constitutive promoter (strain CHA901). No differences in β-galactosidase specific activities could be detected throughout exponential and stationary growth of these strains in OSG medium amended or not with 100 μM 2,4-DAPG (data not shown). In conclusion, it appears that 2,4-DAPG induction is specific for phlA expression.

• Open in new tab
• Download powerpoint

Fig. 4.

Induction by 2,4-DAPG and repression by salicylate, pyoluteorin, and fusaric acid of phlA′-′lacZ expression in growing cultures of P. fluorescens CHA0 harboring pME6259. Specific β-galactosidase activities were determined for strain CHA0/pME6259 cultivated in OSG medium supplemented with 100 μM 2,4-DAPG (●), 100 μM pyoluteorin (□), 500 μM salicylate (◊), 500 μM fusaric acid (▿), or no effector (○). ODs at 600 nm (▵) are shown for CHA0/pME6259 grown in OSG without effector added. Addition of effectors to the medium had no effect on cell growth (data not shown). Means ± standard deviations from at least nine experiments are shown. Some of the error bars are too small to be shown.

Salicylate and pyoluteorin, extracellular metabolites of P. fluorescens CHA0, repress phlA expression and 2,4-DAPG production.To evaluate the specificity of induction by 2,4-DAPG further, we tested other extracellular metabolites of P. fluorescens CHA0, i.e., salicylate and pyoluteorin, for their effect on the expression of phlA′-′lacZ. Addition of 500 μM salicylate or 100 μM pyoluteorin to OSG medium repressedphlA expression throughout exponential and stationary growth of CHA0/pME6259 up to 10-fold, without affecting bacterial growth (Fig.4; Table 3). Addition of salicylate also nearly completely abolished 2,4-DAPG and MAPG production by CHA0/pME6259 at any growth stage in OSG medium (data not shown). Salicylate added at a 10-fold-lower concentration had the same repressive effect on phlA′-′lacZexpression (Table 3) and on the production of the two metabolites (data not shown). In a control experiment, 500 μM salicylate had no effect on the expression of chromosomal lacZ fusions inhcnA or in the protease biosynthetic gene aprA in the CHA0 derivatives CHA207 and CHA805 grown under the same experimental conditions. Likewise, salicylate did not affect expression of an hcnA′-′lacZ fusion carried by plasmid pME3219 (Table1) in a CHA0 background (data not shown). As additional controls, benzoate and acetophenone, two compounds having some structural resemblance to 2,4-DAPG and MAPG, were shown to have no significant effect on the expression of the phlA′-′lacZ reporter (Table3). Taken together, these results suggest that phlbiosynthetic genes as well as 2,4-DAPG and MAPG production in P. fluorescens CHA0 can be repressed by the bacterium's own extracellular metabolites salicylate and pyoluteorin.

Autoinduction of phlA by 2,4-DAPG requires the repressor PhlF.To study the regulation of the phlbiosynthetic genes, the phlA′-′lacZ reporter construct pME6259 was transferred to several mutants of P. fluorescensCHA0, which were grown in OSG medium. In the 2,4-DAPG- and MAPG-negative mutant CHA631 (ΔphlA),phlA′-′lacZ expression was about 10-fold lower than in wild-type strain CHA0 in the exponential growth phase (Fig. 4 and5A). Addition of 100 μM 2,4-DAPG to the medium led to induction of phlA expression in CHA631 in the exponential growth phase, although the final level was lower than that observed for the wild type (Fig. 5A). In the 2,4-DAPG- and MAPG-negative phlH mutant CHA630, the same partial restoration of phlA expression by added 2,4-DAPG could be observed (data not shown), suggesting that additional factors may be required for wild-type-level expression of phlA.

• Open in new tab
• Download powerpoint

Fig. 5.

Effect of 2,4-DAPG on the expression of aphlA′-′lacZ translational fusion carried by pME6259 in theP. fluorescens phlA mutant CHA631 (A), in thephlF mutant CHA638 (B), and in the gacA mutant CHA89 (C). Specific β-galactosidase activities were determined for strains grown in OSG medium at 30°C without (○) or with (●) addition of 100 μM 2,4-DAPG ODs at 600 nm for cultures grown in the absence (▵) and presence (▴) of 2,4-DAPG are shown. Means ± standard deviations from at least three experiments are shown. Some of the error bars are too small to be shown.

In the phlF mutant CHA638, phlA′-′lacZ was expressed earlier than in the wild type (Fig. 5B), confirming the role of PhlF as a repressor and explaining the fourfold increased 2,4-DAPG production in strain CHA638 (Table 2). Expression ofphlA′-′lacZ in the phlF mutant was not influenced by the addition of 2,4-DAPG (Fig. 5B). Intriguingly, phlAexpression in the phlF mutant CHA638 was not repressed by the addition of 500 μM salicylate (Table 3), and production of 2,4-DAPG and MAPG was diminished only by 20 to 25% at any growth stage (data not shown). Growth of strain CHA638 was not affected by 2,4-DAPG or salicylate. In the complemented mutant CHA638.phlF⁺, induction by 2,4-DAPG and repression by salicylate of phlA′-′lacZ expression was similar to that observed in the wild-type strain CHA0, although for unknown reasons the reporter construct was expressed at a threefold-lower level (Table 3).

To test whether phlA gene expression is regulated at the level of transcription, a transcriptional phlA-lacZ fusion carried by pME6710 (Table 1) was introduced into the wild-type strain CHA0, the phlA mutant CHA631, and the phlF mutant CHA638. As in the case of the translational phlA′-′lacZfusion (Fig. 4, 5A, and 5B; Table 3), the expression of the transcriptional phlA-lacZ fusion was fourfold lower in thephlA mutant (1,440 ± 163 Miller units) and sixfold higher in the phlF mutant (34,810 ± 1,240 Miller units) than in wild-type strain CHA0 (5,690 ± 140 Miller units) during the late exponential growth phase (OD at 600 nm of 3.0). These results confirm that phlA expression is regulated positively by 2,4-DAPG and negatively by PhlF at the transcriptional level.

PhlF-mediated regulation was confirmed in a heterologous host, E. coli DH5α, expressing the phlA′-′lacZ fusion on pME6259, with or without the phlF gene on pME6824 (Table 1). In strain DH5α carrying the phlA′-′lacZ reporter alone, addition of 100 μM 2,4-DAPG or 500 μM salicylate had no effect onphlA gene expression (Table 3). When the strain carried also the phlF gene, the expression of the phlAreporter was repressed about 10-fold (Table 3), confirming the role of PhlF as a phl repressor. The strong repression was probably due to the high copy number of plasmid pME6824. Expression ofphlA was significantly induced (1.6-fold) by 2,4-DAPG and repressed (1.6-fold) by salicylate when phlF was present in DH5α (Table 3). Poor uptake of the effectors by E. colimight be the reason for the low induction and repression factors.

P. fluorescens CHA0 requires the global response regulator GacA for 2,4-DAPG production (Table 2) (30). Accordingly, the expression of the phlA′-′lacZ fusion on pME6259 was strongly lowered in the gacA mutant CHA89 and was only marginally induced by 2,4-DAPG (Fig. 5C).

Fusaric acid, a fungal metabolite, represses phlA gene expression and 2,4-DAPG production.Fusaric acid, a mycotoxin produced by Fusarium spp., represses 2,4-DAPG production in strain CHA0 in vitro and on tomato roots (14). The effect of added fusaric acid (500 μM) on phlA′-′lacZ expression in strain CHA0/pME6259 was similar to that observed for salicylate and pyoluteorin. Fusaric acid repressed the expression of phlAby a factor of up to 10, depending on the growth stage (Fig. 4; Table3). Likewise, biosynthesis of 2,4-DAPG and MAPG was diminished to levels close to the detection limit when the strain was grown in presence of 50 or 500 μM fusaric acid (data not shown). Fusaric acid appears to act specifically on phl biosynthetic genes, since CHA0 derivatives carrying hcnA′-′lacZ fusions in the chromosome (CHA207) or on a plasmid (pME3219) were not affected by the compound (data not shown). The phlF mutant CHA638 was largely insensitive to fusaric acid with respect to phlAgene expression (Table 3) and 2,4-DAPG production (data not shown), and the fungal metabolite did not affect bacterial growth.

In a plate assay, F. oxysporum f. sp.radicis-lycopersici interfered with 2,4-DAPG production inP. fluorescens CHA0 but not in a phlF mutant. This was shown by growing bacterial strains in the presence or absence of Fusarium on malt agar plates, a medium which promotes 2,4-DAPG production by P. fluorescens CHA0 (25). 2,4-DAPG production by P. fluorescens strains was monitored as clear inhibition zones in the growth of B. subtilis, used as 2,4-DAPG-sensitive indicator (25). In the absence ofFusarium, inhibition zones obtained with the phlFmutant CHA638 were 18% ± 4% larger than those produced by the wild-type strain CHA0, thus illustrating derepression of 2,4-DAPG production in the phlF mutant. When the fungal pathogen was present on the same plate, inhibition zones produced by strain CHA0 were significantly smaller (17% ± 6%), whereas no reduction of inhibitory activity could be observed for the phlF mutant. Controls showed that the fungus had produced 4.3 μM fusaric acid on this medium. In summary, fusaric acid represses phlAexpression via PhlF, and this fungal interference can be demonstrated in vivo.

Autoinduction of 2,4-DAPG biosynthesis.This study shows, for the first time, that expression of phlA, the first gene of the 2,4-DAPG biosynthetic operon (2), is specifically autoinduced by 2,4-DAPG (Fig. 4; Table 3) and to a lesser extent by MAPG (Table 3). This positive autoregulation was evident in aphlA mutant, CHA631, in which exogenously added 2,4-DAPG compensated for the lack of 2,4-DAPG and MAPG production and partially restored phlA expression (Fig. 5A). The inducing effect of exogenous 2,4-DAPG was most pronounced in a medium (NYB) which does not promote the production of 2,4-DAPG by strain CHA0. Other examples of autoinduction in bacteria are known. The best-documented mechanism is the autoinduction of the biosynthesis of N-acyl-homoserine lactones, i.e., diffusible molecules that mediate cell-to-cell communication, commonly known as quorum sensing, in many gram-negative bacteria (19). Positive autoregulation of biosynthetic genes has also been described for the siderophores pyochelin in P. aeruginosa (42) and yersiniabactin in Yersinia enterolitica (35). Furthermore, indole-3-acetic acid was found to upregulate the expression of the ipdC gene encoding a key enzyme of its own biosynthetic pathway inAzospirillum brasilense (53). Finally, early work by Gutterson (21) proposed positive feedback regulation as a mechanism controlling the biosynthesis of oomycin A, an antifungal compound produced by P. fluorescens Hv37a.

Our work confirms the proposed role of PhlF as a pathway-specific repressor (2) acting at the transcriptional level and establishes PhlF as a mediator of autoinduction by 2,4-DAPG. Two observations support our findings. First, a phlF mutant, CHA638, was insensitive to 2,4-DAPG addition and kinetics ofphlA expression corresponded to those in the wild-type CHA0 with 2,4-DAPG added (Fig. 4 and 5B). Second, phlF was required for 2,4-DAPG-induced phlA expression in the heterologous host E. coli (Table 3). PhlF may bind to the promoter(s) of phlA, thereby repressing the expression of the 2,4-DAPG biosynthetic operon. 2,4-DAPG might bind to PhlF and thereby prevent the interaction of the repressor protein with the promoter, leading to increased expression of the 2,4-DAPG biosynthetic genes. This autoinduction circuit can be boosted by the global response regulator GacA (5, 30), as shown by the strong GacA dependence of phlA expression (Fig. 5C).

Repression of 2,4-DAPG biosynthesis.The 2,4-DAPG autoinduction circuit in P. fluorescens CHA0 was shown to be blocked by salicylate or pyoluteorin (Fig. 4; Table 3), both extracellular metabolites produced by this strain and numerous other pseudomonads (24, 55). Moreover, fusaric acid, a pathogenicity factor of F. oxysporum, also strongly repressed phlA expression (Fig. 4; Table 3), confirming, at a molecular level, earlier observations by Duffy and Défago (14) that the fungal metabolite abolishes 2,4-DAPG production by strain CHA0 in vitro and in the rhizosphere of tomato plants. PhlF seems to have a mediator role, since the presence of an intact phlF gene was required for repression ofphlA expression by salicylate, pyoluteorin, or fusaric acid (Table 3). One possible explanation could be that the repressing compounds might directly or indirectly compete with 2,4-DAPG for binding sites on PhlF.

Based on our observations, it is tempting to speculate that relative concentrations of 2,4-DAPG, salicylate, pyoluteorin, and perhaps other extracellular metabolites might be sensed by P. fluorescensand maintained at homeostatically balanced levels. For instance, the data in Table 2 indicate that 2,4-DAPG levels are inversely correlated with pyoluteorin concentration and pyoluteorin-negative mutants of strain CHA0 overproduce 2,4-DAPG and MAPG (31). Positive feedback regulation and repression by extracellular metabolites might allow the bacterium to adapt to and fine-tune levels of these extracellular metabolite in response to environmental changes and to avoid escalation of the autoinduction loop.

Kinetics of 2,4-DAPG biosynthesis.In P. fluorescens CHA0, 2,4-DAPG and MAPG were accumulated until the beginning of stationary growth phase (Fig. 2), as is typical for secondary metabolites. Surprisingly, thereafter the concentrations of the two metabolites in the growth medium steadily decreased. Initially 2,4-DAPG appears to be degraded to MAPG, which could then be further metabolized to compounds that have not been identified. At a temperature (i.e., 18°C) that is closer to that found in the natural habitat, accumulation and degradation rates of 2,4-DAPG were slowed down and the period of maximal concentrations was doubled compared to that at 30°C (data not shown). Why should maximal concentrations of a major biocontrol compound be available to the bacterium only during a limited period? As long as the physiological functions of 2,4-DAPG remain unknown, it is difficult to provide a good answer. Maybe 2,4-DAPG-mediated self-defense against competitors and predators is more effective when it does not operate permanently in the biocontrol bacterium. Our findings have two important implications. First, they illustrate that kinetic studies, rather than point measurements, can be important in studies on the regulation of antibiotic biosynthesis inP. fluorescens. Second, degradation of 2,4-DAPG to MAPG by strain CHA631 adds a further level of complexity to positive autoregulation and kinetics of 2,4-DAPG biosynthesis. Previously, MAPG has been proposed to be a direct precursor of 2,4-DAPG. A genomic region encoding an acetyltransferase activity capable of converting MAPG to 2,4-DAPG has been described for P. fluorescensstrain F113 (13, 48). In P. fluorescens strain Q2-87, the products of the 2,4-DAPG biosynthetic genesphlACB are necessary for the conversion of MAPG to 2,4-DAPG (2).

Implications for biocontrol.From an ecological point of view, rapid accumulation of the biocontrol compound 2,4-DAPG via positive autoregulation may be advantageous since it may allow the bacterium to respond promptly to competition with other microorganisms, including plant pathogens, in the rhizosphere. At a population level, the capacity of P. fluorescens to perceive exogenous 2,4-DAPG as a signal inducing 2,4-DAPG biosynthesis potentially implies a novel way of communication within or between populations of 2,4-DAPG producers. In this way, the 2,4-DAPG pool within a homogenous or mixed P. fluorescens population could rapidly be boosted to levels that are relevant to pathogen control (38). The importance of extracellular signal molecules for the communication within and between populations of fluorescent pseudomonads has recently been demonstrated for N-acyl-homoserine lactone-mediated expression of phenazine antibiotic genes in the rhizosphere of wheat (36). The potential for negative cross talk is illustrated by the fact that the fungal pathogen F. oxysporum was shown to break the autoregulatory mechanism (Fig. 4; Table 3) (14).