Laboratory of Phytopathology, Wageningen University, Wageningen, The Netherlands
Microbial Ecology Department, Netherlands Institute of Ecology (NIOO-KNAW), Wageningen, The Netherlands
Correspondence: Jos M. Raaijmakers, Microbial Ecology Department, Netherlands Institute of Ecology (NIOO-KNAW), 6708 PB Wageningen, The Netherlands. Tel.: +31 317473497; fax: +31 317473675; e-mail: email@example.com
Pseudomonas fluorescens SS101 produces the cyclic lipopeptide massetolide with diverse functions in antimicrobial activity, motility, and biofilm formation. To understand how massetolide biosynthesis is genetically regulated in SS101, c. 8000 random plasposon mutants were screened for reduced or loss of massetolide production. Of a total of 58 putative mutants, 45 had a mutation in one of the three massetolide biosynthesis genes massA, massB, or massC. For five mutants, the insertions were located in the known regulatory genes gacS, gacA, and clpP. For the remaining eight mutants, insertions were located in clpA, encoding the ClpP chaperone, in phgdh, encoding D-3-phosphoglycerate dehydrogenase, in the heat shock protein-encoding dnaK, or in the transmembrane regulatory gene prtR. Genetic, chemical, and phenotypic analyses showed that phgdh, dnaK, and prtR are indeed involved in the regulation of massetolide biosynthesis, most likely by transcriptional repression of the LuxR-type regulator genes massAR and massBCR. In addition to their role in massetolide biosynthesis, dnaK and prtR were found to affect siderophore and extracellular protease(s) production, respectively. The identification of new regulatory genes substantially extended insights into the signal transduction pathways of lipopeptide biosynthesis in P. fluorescens and into regulation of other traits that may contribute to its life-style in the rhizosphere.
Lipopeptides (LP) are produced by diverse bacterial genera and are composed of a lipid tail linked to a short linear or cyclic oligopeptide (Raaijmakers et al., 2010). LP are surface-active compounds, exhibit broad-spectrum antibiotic activities, and have diverse natural functions in motility, biofilm formation, and virulence (Raaijmakers et al., 2006; Raaijmakers & Mazzola, 2012). The LP massetolide A was identified in Pseudomonas fluorescens SS101, a biocontrol strain isolated from the wheat rhizosphere (de Souza et al., 2003; de Bruijn et al., 2008), and has potent surfactant and broad-spectrum antimicrobial activities (Gerard et al., 1997; van de Mortel et al., 2009). LP biosynthesis is governed by large multimodular nonribosomal peptide synthetases (NRPS) and is well studied in Pseudomonas and Bacillus (Finking & Marahiel, 2004; Raaijmakers et al., 2006). In contrast, relatively little is known about the regulatory networks, and the signal transduction pathways involved in LP biosynthesis. Among the global regulatory systems, two-component regulators play an important role in the regulation of LP biosynthesis (Raaijmakers et al., 2010). For example, the GacS/GacA two-component system in Pseudomonas functions as a master switch, and a mutation in either one of the two genes results in loss of LP production (Kitten et al., 1998; Koch et al., 2002; Dubern et al., 2005; de Bruijn et al., 2007). Also in Bacillus, LP biosynthesis is regulated by a two-component system as was shown for ComA/ComP in surfactin biosynthesis (Sullivan, 1998). Also quorum sensing plays a role in the regulation of LP biosynthesis in some species and strains. For example, in Bacillus, the cell-density-dependent pheromone ComX and the phosphatase RapC are involved in surfactin biosynthesis (Duitman et al., 2007). For Pseudomonas fluorescens 5064 and Pseudomonas putida PCL1445, N-acyl homoserine lactones (N-AHLs) regulate viscosin and putisolvin biosynthesis, respectively (Cui et al., 2005; Dubern et al., 2006). However, for various other Pseudomonas strains belonging to the same species, including P. fluorescens strain SS101, subject of this study, no evidence was found for a role of N-AHL-mediated regulation of LP biosynthesis (Dumenyo et al., 1998; Kinscherf & Willis, 1999; Andersen et al., 2003; de Bruijn et al., 2007, 2008). This indicates that cell-density-dependent regulation of LP biosynthesis can differ between species and between strains of the same species. In addition to these two global regulatory systems, LuxR-type transcriptional regulators that flank the LP biosynthesis genes have been shown to regulate syringomycin, syringopeptin, syringofactin, putisolvin, entolysin, viscosin, and massetolide biosynthesis (Lu et al., 2002; Wang et al., 2006; Berti et al., 2007; Dubern et al., 2008; de Bruijn & Raaijmakers, 2009b; Vallet-Gely et al., 2010). In P. putida, also DnaK was identified as an additional regulator of putisolvin biosynthesis (Dubern et al., 2005).
The study presented here focuses on identification of regulatory genes of massetolide biosynthesis in the beneficial rhizobacterium P. fluorescens SS101. Massetolides consist of a 9-amino acid cyclic peptide moiety linked to a 3-hydroxydecanoic acid tail and belong to the viscosin LP group (Raaijmakers et al., 2006). Massetolide biosynthesis is governed by three NRPS genes, designated massA, massB, and massC, flanked by two LuxR-type regulatory genes massAR and massBCR (de Bruijn et al., 2008; de Bruijn & Raaijmakers, 2009b). We previously identified the two-component system GacS/GacA and the serine protease ClpP as regulators of massetolide biosynthesis in strain SS101 (de Bruijn & Raaijmakers, 2009a). At the transcriptional level, ClpP-mediated regulation of massetolide biosynthesis appears to operate independently from the regulation by GacS/GacA (de Bruijn & Raaijmakers, 2009a). Based on these previous findings, a tentative regulation model was proposed where ClpP regulates, alone or together with a yet unknown chaperone other than ClpX, massetolide biosynthesis via degradation of putative transcriptional repressors of massAR, and/or via modulation of the citric acid cycle and amino acid metabolism (de Bruijn & Raaijmakers, 2009b). The aims of this study were to (1) perform a genome-wide search for new regulatory genes of massetolide biosynthesis in P. fluorescens SS101, (2) determine the role of these genes in the regulation of massetolide production, and (3) investigate the putative role of these regulatory genes in other phenotypic traits of P. fluorescens SS101. To this end, we screened two independent random plasposon mutant libraries (c. 8000 mutants total) for a reduced or loss of massetolide production. Thirteen putative regulatory mutants were found. Genetic, phenotypic, chemical, and transcriptional analyses were performed to elucidate the functions of three regulatory genes in massetolide biosynthesis and in other phenotypic traits, including swarming motility, siderophore production, and extracellular protease activity.
Materials and methods
For a detailed description of the Materials and Methods used in this study, we refer to the Data S1 (Supporting Information).
Screening for massetolide-deficient mutants of P. fluorescens SS101
Two independent libraries of 520 and 7500 random TnMod plasposon mutants of strain SS101 were screened in a drop collapse assay (Fig. 1a) for reduced or loss of massetolide production. The drop collapse assay is a highly reliable proxy (de Bruijn et al., 2008; de Bruijn & Raaijmakers, 2009a) for massetolide production in P. fluorescens SS101. A total of 58 putative massetolide-deficient mutants were found. The regions flanking the plasposon insertion were sequenced for all 58 mutants. In 45 mutants, the insertion was located in massA, massB, or massC. The insertions in the other 13 mutants were located in three genes described previously for their role in massetolide biosynthesis (de Bruijn & Raaijmakers, 2009a) and in four putative new regulatory genes. The three known regulatory genes were the caseinolytic protease gene clpP (n = 1), the sensor kinase gene gacS (n = 3), and its cognate response regulator gene gacA (n = 1). The four putative new regulatory genes were clpA (n = 4; PflSS101_3193), dnaK (n = 2; PflSS101_4633), prtR (n = 1; PflSS101_3280), and phgdh (n = 1; PflSS101_5176). The clpA gene encodes the chaperone of the ClpP serine protease and most likely regulates massetolide biosynthesis via ClpP. The regulatory role of ClpA and its interplay with ClpP was not further investigated here. Instead, a more detailed functional analysis was conducted for phgdh, dnaK, and prtR.
To confirm the role of these three genes in the regulation of massetolide biosynthesis, RP-HPLC analysis showed that these three mutants were indeed all deficient in the production of massetolide A and its derivatives (Fig. 1b). Complementation of the mutants with each of the corresponding target genes cloned into the stable vector pME6031 restored massetolide production, whereas no massetolides were detected in the empty-vector control (Fig. 1a and b). Massetolide biosynthesis is known to be essential for swarming motility of strain SS101 (de Bruijn et al., 2008). All three mutants lost their ability to swarm on soft agar (0.6% w/v; Fig. 1c). Swarming motility was restored by complementation with the corresponding target gene, albeit with a slightly different swarming pattern (Fig. 1c), which may be due to effects of the copy number of the plasmid used for complementation. Collectively, these results indicate that all three genes are indeed required for massetolide biosynthesis in SS101.
Characterization of regulatory mutant ∆phgdh
D-3 phosphoglycerate dehydrogenase (phgdh) is known to be involved in the biosynthesis of the amino acid L-serine. It converts 3-phosphoglycerate into 3-phosphohydroxypyruvate which in turn is converted into 3-phosphoserine by 3-phosphoserine aminotransferase (PSAT). Finally, 3-phosphoserine is converted into L-serine by phosphoserine phosphatase (PSP; van der Crabben et al., 2013). In P. fluorescens SS101, the phgdh gene is 1230 bp, and blastx analysis showed 89–99% identity to homologs in other Pseudomonas genomes (Fig. 2a). A mutation in phgdh significantly reduced the expression of massA, massB, and massC and of the LuxR-type transcriptional regulators massAR and massBCR (Fig. 2b). Growth of the phgdh mutant was adversely affected in the stationary phase, and this deficiency was largely restored by genetic complementation (Fig. 2c). Based on the drop collapse assay, which is a highly reliable proxy for massetolide production in P. fluorescens SS101, we observed a restoration of massetolide production to wild-type level with the addition of increasing concentrations of L-serine to the growth medium (Fig. 2d). Collectively, these results indicate that phgdh regulates massetolide biosynthesis via modulation of L-serine biosynthesis.
Characterization of regulatory mutant ∆dnaK
The dnaK gene codes for a molecular chaperone of the Hsp70 protein family and was shown previously to regulate, together with its adjacent genes dnaJ and grpE, the biosynthesis of the LP putisolvin in P. putida PCL1445 (Dubern et al., 2005). In P. fluorescens SS101, the complete dnaK gene comprised 1923 bp and blastx analysis showed 86–97% identity to dnaK in other Pseudomonas genomes (Fig. 3a). Like in P. putida PCL1445 and other Pseudomonas species, dnaK is flanked in strain SS101 by the chaperone-encoding gene dnaJ and the heat shock protein-encoding gene grpE (Fig. 3a). qRT-PCR analyses showed that the transcript levels of massA, massB, massC, and the two regulatory genes massAR and massBCR were significantly decreased in the dnaK mutant (Fig. 3b). The growth rate of the dnaK mutant in KB broth was reduced relative to that of wild-type SS101, particularly in the lag phase; the mutant also reached relatively lower cell densities in the stationary phase (Fig. 3c).
In P. putida PCL1445, putisolvin production decreased with increasing temperature (Dubern et al., 2005). Here, we also observed that, based on tensiometric analysis of the cell-free culture filtrates, massetolide production decreased with increasing temperatures for both wild-type SS101 and the complemented dnaK mutant (Fig. S1, Supporting Information). Next to these phenotypes, we observed a significant difference in fluorescence between wild-type SS101 and the dnaK mutant when grown in King's medium B (KB) broth, with the wild type being more green fluorescent than the dnaK mutant. Subsequent phenotypic analysis on CAS-agar indicator plates confirmed that a mutation in dnaK adversely affects siderophore production (Fig. 3d). Spectrophotometric analysis (A400 nm) of cell-free culture filtrates of SS101 and of the dnaK mutant grown in KB broth confirmed the results of the CAS-agar plate assays, with a reduced siderophore production in the dnaK mutant (not shown). This alteration in siderophore production by the dnaK mutant was observed at four different incubation temperatures (i.e., 10, 18, 25, and 28 °C; not shown) and therefore seems not to be temperature dependent which is in contrast with the results shown for massetolide production. Collectively, these results showed that, along with massetolide production, a mutation in dnaK also affects siderophore production in SS101.
Characterization of regulatory mutant ∆prtR
PrtR was previously reported to be a novel antisigma factor and transmembrane activator which interacts with ECF (extra-cytoplasmic function) sigma factors of the σ70 family (Burger et al., 2000; Mascher, 2013). The neighboring prtI gene is an ECF sigma factor and usually encoded in an operon with prtR (Mascher, 2013). blastx analysis of the complete prtR gene (738 bp) of SS101 showed 43–87% identity to prtR homologs in P. putida, P. fluorescens, and P. protegens genomes (Fig. 4a). The prtI and prtR genes were not found in P.syringae or P.aeruginosa strains (Fig. 4a). In SS101, prtR is flanked by prtI and one transcription unit was predicted based on SoftBerry FGENESB analysis (Fig. 4a). qRT-PCR analyses showed that, similar to the results described above for the other two regulatory genes, transcript levels of the three massetolide biosynthesis genes and the two LuxR-type regulatory genes were significantly decreased in the prtR mutant (Fig. 4b). The prtR mutation did not affect the growth of P. fluorescens SS101 in the lag and exponential phase, but did so in the stationary phase (Fig. 4c). Complementation of the prtR mutation with the prtIR operon did not fully restore the growth, which may be due to effects of plasmid-mediated copy number of prtR, prtI, or both.
In P. fluorescens LS107d2, the prtIR genes were shown to be involved in temperature-dependent regulation of extracellular protease activity (Burger et al., 2000). In the study by Burger et al. (2000), the prtR and prtI mutants produced extracellular proteases at 23 °C but not at 29 °C. In SS101, extracellular protease activity was reduced in the prtR mutant (Fig. 4d) at all three temperatures (25 °C, 29 °C, and 30 °C) tested. Collectively, these results indicate that prtR regulates massetolide biosynthesis as well as extracellular protease production in SS101.
Massetolides are lipopeptide biosurfactants required for swarming motility, biofilm formation, and broad-spectrum antimicrobial activities (Gerard et al., 1997; de Souza et al., 2003; de Bruijn et al., 2008; van de Mortel et al., 2009). Here, we analyzed 58 massetolide-deficient mutants and discovered four new regulatory genes of massetolide biosynthesis, that is, clpA, phgdh, dnaK, and prtR. Consistent with the results of our previous study on the role of the GacS/A two-component system and the ClpP serine protease (de Bruijn & Raaijmakers, 2009a), mutations in phgdh, dnaK, and prtR adversely affected transcription of the three massABC biosynthesis genes, most likely through transcriptional repression of one or both of the LuxR-type regulatory genes massAR and massBCR (Fig. 5). Although we screened a large library of more than 8000 random mutants, we most likely did not cover the entire genome in strain SS101 to identify all regulatory genes of massetolide biosynthesis. Work by Liberati et al. (2006) on 34 176 random mutants of Pseudomonas aeruginosa PA14 (genome 6.53 Mb) showed that 75% of the predicted genes were mapped with on average 4.3 transposon insertions per gene. They also pointed out that there is bias in transposon insertion sites and that large genes tend to have a higher frequency of insertions than relatively small genes. This may explain that the majority (n = 45) of the massetolide mutants had the insertion in the large (6–13 kb) massABC genes and that no mutants were found in, for example, the small massAR (795 bp) and massBCR (672 bp) genes.
Based on our previous study (de Bruijn & Raaijmakers, 2009b), we proposed a model in which the serine protease ClpP regulates massetolide biosynthesis, alone or together with a chaperone other than ClpX, by degradation of putative transcriptional repressors of massAR or via modulation of the citric acid cycle and amino acid metabolism (Fig. 5). In that study, we also showed that addition of proline and glutamic acid to the growth medium can partially complement the deficiency in swarming motility of the clpP mutant (de Bruijn & Raaijmakers, 2009a). The results obtained here for phgdh, a key gene in L-serine biosynthesis, further extend the hypothesis that amino acid metabolism and in particular serine biosynthesis affects massetolide production. Given that serine makes up two of the nine amino acids in massetolide A, a possible scenario may be that a phgdh mutation depletes the cellular pool of serine thereby affecting massetolide biosynthesis/assembly (Fig. 5). More experiments will be required to support this hypothesis.
DnaK, together with the flanking genes dnaJ and grpE, was identified previously by Dubern et al. (2005) for its role in the regulation of putisolvin biosynthesis in P. putida. They postulated that DnaK, DnaJ, and GrpE may be required for proper folding or activity of other regulators of the putisolvin biosynthesis gene psoA or alternatively may be necessary for proper assembly of the putisolvin NRPSs (Dubern et al., 2005). In our study, we found that DnaK regulates massetolide as well as siderophore biosynthesis in P. fluorescens SS101. Based on these results, one might speculate that DnaK has a more global function in the regulation of NRPS genes, but more work needs to be done to support this suggestion. For P. putida, Dubern et al. (2005) also showed that dnaK expression was under the control of GacS/GacA. Preliminary results of whole-genome microarray analyses of the Gac mutant of strain SS101, however, suggest that phgdh, dnaK, prtR, and clpA expressions are not under the control of the GacS/GacA system. Hence, in the adapted regulation model (Fig. 5), the Gac-signal transduction route is kept separate from the other regulatory genes.
Bacteria possess different means to connect an extracellular input with an appropriate cellular response. Following one-component and two-component systems, extra-cytoplasmic function σ factors (ECFs) represent the third most abundant type of bacterial signal transduction. PrtR interacts with ECF sigma factors of the σ70 family and is required for aprX protease expression in P. fluorescens LS107d2 (Burger et al., 2000). In P. entomophila, prtR affects aprA protease production and contributes to pathogenicity (Liehl et al., 2006). PrtR was also found in P. fluorescens WH6 to regulate the biosynthesis of the germination-arrest factor (GAF), which is a predicted small peptide or amino acid analog with herbicidal activity (Kimbrel et al., 2010; Halgren et al., 2013). Here, we showed, for the first time, that prtR also regulates massetolide biosynthesis in P. fluorescens SS101.
In conclusion, the identification of at least three new regulatory genes substantially extended our insight into the regulatory network of lipopeptide biosynthesis in P. fluorescens SS101. Based on the results presented, we postulate that these three genes most likely regulate massetolide biosynthesis via one or both of the LuxR-type transcriptional regulators flanking the massA and massBC biosynthesis genes. Apart from their role in massetolide biosynthesis, we also showed that dnaK and prtR affect other traits in strain SS101 that may contribute to its lifestyle in the rhizosphere.
We are very grateful to the Graduate School of Experimental Plant Sciences (EPS) for financing this project. The authors of this manuscript have no conflict of interest to declare. This is publication No.5587 of the Netherlands Institute of Ecology (NIOO-KNAW).