Lack of CbrB in Pseudomonas putida affects not only amino acids metabolism but also different stress responses and biofilm development

Authors


E-mail icanper@upo.es; Tel. (+34) 9 5434 9052; Fax (+34) 9 5434 9376.

Summary

The CbrAB two-component system has been described in certain species of Pseudomonads as a global regulatory system required for the assimilation of several amino acids (e.g. histidine, proline or arginine) as carbon or carbon and nitrogen sources. In this work, we used global gene expression and phenotypic analyses to characterize the roles of the CbrAB system in Pseudomonas putida. Our results show that CbrB is involved in coordination with the nitrogen control system activator, NtrC, in the uptake and assimilation of several amino acids. In addition, CbrB affects other carbon utilization pathways and a number of apparently unrelated functions, such as chemotaxis, stress tolerance and biofilm development. Based on these new findings, we propose that CbrB is a high-ranked element in the regulatory hierarchy of P. putida that directly or indirectly controls a variety of metabolic and behavioural traits required for adaptation to changing environmental conditions.

Introduction

Bacteria can colonize very different niches because they can adapt to changing environmental conditions by altering their metabolism, physiology and behaviour. This capacity is the result of the coordinated action of regulatory systems that are triggered in response to environmental and/or physiological stimuli or signals to control the metabolic, physiological and behavioural traits that allow the cell to thrive in the new conditions. Likely due to their continuous exposure to changing environmental conditions, soil bacteria of the genus Pseudomonas possess an unusually high number of regulatory systems. Nearly 10% of the genes in the genome of the model strain Pseudomonas putida KT2440 encode products involved in signal transduction and gene regulation. This strain comprises a very high number of transcriptional regulatory proteins that confers the bacterium a very high degree of functional versatility. Among these proteins, we find 94 response regulator receiver determinants of two-component systems (Dos Santos et al., 2004), some of which have defined functions, whereas many others remain uncharacterized.

The response regulator, CbrB, in Pseudomonas aeruginosa is part of the CbrAB two-component system, which is required for growth on different nutrients, including several amino acids and other nitrogen-containing substrates as the sole carbon source (Nishijyo et al., 2001). CbrA is a sensor histidine kinase that contains transmembrane domains, and CbrB is an activator of σ54-dependent promoters belonging to the NtrC family. In silico analysis reveals the presence of a CbrB homologue in only a few groups of the γ-Proteobacteria other than the pseudomonads. The P. aeruginosa CbrAB system is directly required for expression of arginine and histidine catabolic genes. Notably, constitutive ntrB and ntrC alleles suppress the growth defect of a ΔcbrAB mutant on several amino acids (arginine, histidine, ornithine and proline), suggesting that some targets may be subjected to dual CbrAB/NtrBC control, depending on whether or not they are used as a carbon or nitrogen source (Li and Lu, 2007). Although the signal recognized by the CbrAB system remains unknown, suppression of the growth defect on proline as a carbon and nitrogen source by addition of a preferential carbon source (succinic acid) and restoration of the mutant phenotype by addition of a preferential nitrogen source (ammonium) suggests that, similar to NtrBC, CbrAB may respond to changes in the carbon/nitrogen balance (Nishijyo et al., 2001). Recently, a similar growth phenotype to that observed in P. aeruginosa has been described for cbrA and cbrB mutants in Pseudomonas fluorescens, and dual involvement of CbrB and NtrC in the regulation of the histidine catabolic hutU-G operon has been demonstrated in this organism (Zhang and Rainey, 2007; 2008a).

We recently described the Ntr regulon of the model soil bacterium P. putida KT2442 (Hervas et al., 2008; 2009).This organism also contains the cbrA and cbrB genes, which may share some targets with the Ntr system; however, their function has not been explored yet. In order to characterize the P. putida KT2442 Cbr regulon, we performed transcriptomic analysis of a cbrB mutant in comparison with the wild type KT2442 isogenic strain, as well as a double cbrB ntrC mutant in comparison with the single ntrC strain. Our analysis revealed that CbrB-dependent regulation of genes is related to carbon metabolism and amino acid uses but also to apparently unrelated functions, such as tolerance to metals, motility, chemotaxis and the transition from a planktonic lifestyle to the formation of surface-attached biofilm communities. Our data considerably expand the role of CbrB from a regulator of carbon metabolism to a general controller of multiple physiological and behavioural traits.

Results

The cbrB mutant is defective in the utilization of several amino acids as carbon or nitrogen sources

In order to test the ability of the of cbrB mutant to utilize different carbon and nitrogen sources, we initially compared the wild type P. putida KT2442 and cbrB mutant (MPO406) strains by Biolog phenotypic microarray analysis. This analysis showed that utilization of a number of amino acids and dipeptides as carbon or nitrogen sources may be impaired in the cbrB mutant (Table S1). The ability of KT2442, MPO406, an ntrC mutant (MPO201) and a double cbrB ntrC mutant (MPO400) to grow in minimal medium containing histidine, proline, arginine, ornithine, tyrosine or glutamic acid as the sole carbon source (ammonium was used as the nitrogen source), the sole nitrogen source (succinic acid was used as the carbon source), or the sole carbon and nitrogen source was further tested by batch culture in flasks.

Table 1 shows a summary of the growth data; for details of the growth curves, see Fig. S1. As shown in Table 1, the cbrB mutant was unable to use ornithine or tyrosine as a carbon source or as carbon and nitrogen sources and showed reduced growth relative to the wild type on the rest of the amino acids when tested as a carbon source or as a carbon and nitrogen source. In contrast, growth on these amino acids as the sole nitrogen source was similar to that of the wild type strain. The weak defect in proline utilization as the carbon or the carbon plus nitrogen source shown by the cbrB mutant was more evident in the double cbrB ntrC mutant, which suggests that NtrC can also activate proline utilization as the carbon source.

Table 1.  Growth properties of P. putida wild type, cbrB, ntrC and cbrB ntrC mutant strains on different C, N or C and N sources.
Source:HisProArgOrnTyrGlu
  1. +++: growth up to OD600 > 0.8 in 7 h; ++: growth up to OD600 > 0.8 in 10 h; +: growth up to OD600 > 0.8 in 15 h; +/−: growth up to OD600 > 0.8 in 21 h; −: no growth after 20 h.

Carbon
 KT2442++++++++++++
 MPO406 (cbrB)+++/−+
 MPO201 (ntrC)++++++++/−++++
 MPO400 (cbrBntrC)+/−+/−+/−+
Nitrogen
 KT2442++++++++++++++++
 MPO406 (cbrB)+++++++++++++/−+++
 MPO201 (ntrC)+++++++++++
 MPO400 (cbrBntrC)+++++++/−+
Carbon/nitrogen
 KT2442++++++++++++
 MPO406 (cbrB)+++++/−++
 MPO201 (ntrC)++++++++/−++++
 MPO400 (cbrBntrC)+/−+/−+/−+

The ntrC mutant showed impaired growth on ornithine, tyrosine and glutamate as nitrogen sources but grew as well as the wild type strain on histidine, proline or arginine as nitrogen sources. In contrast, growth of the ntrC mutant on amino acids, such as carbon or carbon and nitrogen sources, was generally comparable to that of the wild type. Use of proline as the nitrogen source was not affected in any of the ntrC or cbrB single mutants, but it was clearly impaired in the double cbrB ntrC mutant, thus suggesting that both NtrC and CbrB can activate the use of proline as the nitrogen source. A similar effect of the double mutation as compared with the single mutations was also observed when histidine was used as the nitrogen or the nitrogen and carbon source.

The cbrB mutant strain cannot use histidine as the sole carbon source, but it can use histidine when it is the only carbon and nitrogen source present in the medium. Apparently, NtrC activates utilization of histidine in this nitrogen-limiting condition; this use of histidine would not occur if ammonium were present as the nitrogen source (nitrogen excess condition). Interestingly, the cbrB mutant grows even better than the wild type when using His as the nitrogen and carbon source. This result may indicate some form of interference with activation at this particular promoter region by CbrB and NtrC when both are present. In support of this view, the ntrC mutant grows better than the wild type when using histidine as the C or the C+N source, which suggests that, in the absence of NtrC, CbrB can more efficiently activate histidine utilization.

Taken together, our results strongly suggest that CbrB is responsible for regulation of the uptake or catabolism of the amino acids tested when used as a carbon source, whereas NtrC appears to regulate the use of ornithine, tyrosine and glutamate as nitrogen sources. In addition, the severity of the phenotype of the double mutant relative to those of the single ntrC and cbrB mutants suggests that, as pointed out previously for P. aeruginosa (Li and Lu, 2007), NtrC and CbrB may have partially overlapping or redundant functions in P. putida.

Genome-wide analysis of the cbrB regulatory network

To further define the regulatory network of the CbrAB two-component system in P. putida, whole-genome transcriptional profiling of the wild type and cbrB strains was performed using DNA microarrays. Because our results suggest that CbrAB is involved in catabolism of several amino acids, and expression of amino acid utilization genes may require the presence of specific amino acids as inducers, we chose to grow the strains for transcriptomic analysis in medium containing a mixture of the 20 amino acids required for protein synthesis as the carbon and nitrogen sources. Comparison of the transcriptomic profiles from the wild type and the cbrB mutant revealed 91 differentially expressed genes whose transcription is influenced directly or indirectly by the CbrAB two-component system. The group of 51 genes that showed increased expression in the wild type strain relative to the cbrB mutant strain was designated as putatively CbrB-activated, whereas the group of 40 genes that showed increased expression in the cbrB mutant as compared with the wild type was designated as putatively CbrB-repressed (see Table S2).

Because our results suggest that the roles of the cbr and ntr systems may be redundant in P. putida, we aimed to discern the effect of CbrB on gene expression in the absence of a functional ntr system. To achieve this goal, we compared the transcriptomic profile of the ntrC mutant with that of the cbrB ntrC double mutant. In this case, we found 122 differentially expressed genes: 50 showed increased expression in the ntrC strain, and 72 showed increased expression in the cbrB ntrC strain (see Table S2).

Table 2 shows genes with assigned potential functions that were regulated by cbrB. A number of genes were not identified as regulated by cbrB in a wild type background (array 1), but they showed regulation by cbrB in an ntrC background (array 2). As evidenced by some growth phenotypes (Table 1), NtrC may at least partially compensate for the regulatory defect caused by the cbrB mutation at some operons. Conversely, we also found a number of genes that were regulated by cbrB in a wild type background but were not in the ntrC mutant background. The values for most of these genes in both arrays were very close to the cut-off and surpassed the limit in one array but not in the other.

Table 2.  Genes whose expression is altered by the cbrB mutation.
 Role/AnnotationArray 1Array 2
  1. Expression data obtained from the microarray analysis of the wild type versus cbrB mutant strains (Array1) and ntrC and ntrC cbrB mutant strains (Array 2), growing at mid-exponential phase in M9 minimal medium supplemented with the 20 amino acids (0.6 mM each). The plus sign indicates that the gene was CbrB-activated (downregulated in the mutant), and the negative sign indicates that the gene was CbrB-repressed (upregulated in the cbrB mutant strain). ‘√’ means the gene passed the cut-off, whereas ‘o’ means it did not.

 Amino acid uptake/metabolism  
PP1001 arcA-arginine deiminase
PP1002 arcD-arginine/ornithine antiporter
+PP0864-ornithine decarboxylase putativeo
+PP1059-amino acid permeaseo
+PP1068-amino acid ABC transporter ATP-binding proteino
+PP1069-amino acid ABC transporter permease proteino
+PP1070-amino acid ABC transporter permease proteino
+PP1138 braF -branched-chain amino acid ABC transporter ATP-bindin
+PP1141 braC -branched-chain amino acid ABC transporter periplasmico
+PP1297 aapJ-general amino acid ABC transporter periplasmic bindio
+PP1768 serC -3-phosphoserine aminotransferaseo
+PP1972 tyrB-1-aromatic-amino-acid aminotransferase
+PP2453 ansA-L-asparaginase IIo
PP3533-ornithine cyclodeaminaseo
PP4064 ivd-isovaleryl-CoA dehydrogenaseo
PP4065-3-methylcrotonyl-CoA carboxylase beta subunit putativeo
+PP4619-maleylacetoacetate isomerase putative
+PP5029 hutG -formiminoglutamase
+PP5030 hutI-imidazolonepropionaseo
+PP5032 hutH -histidine ammonia-lyaseo
+PP5033 hutU -urocanate hydrataseo
+PP5338 aspA-aspartate ammonia-lyase
 Carbon uptake/metabolism  
PP1022 zwf-1-glucose-6-phosphate 1-dehydrogenaseo
PP1023 pgl -6-phosphogluconolactonaseo
PP1024 eda -2-dehydro-3-deoxyphosphogluconate aldolase
PP1169 dctP-TRAP dicarboxylate transporter DctP subunito
PP1912 plsX-fatty acid/phospholipid synthesis protein PlsXo
PP2454 rbsB-ribose ABC transporter periplasmic ribose-binding pro
+PP3163 benC-benzoate dioxygenase ferredoxin reductase componento
+PP3365-acetolactate synthase catabolic putativeo
+PP4012-isocitrate dehydrogenaseo
PP4063-long-chain-fatty-acid-CoA ligase putativeo
PP4116 aceA -isocitrate lyaseo
+PP4403 bkdB -2-oxoisovalerate dehydrogenase lipoamide acyltransferaseo
+PP5346 oadA -oxaloacetate decarboxylase alpha subunito
PP5365-cyclopropane-fatty-acyl-phospholipid synthase putativeo
 Cell adhesion and motility/chemotaxis  
PP0320-methyl-accepting chemotaxis transducero
PP0806-lapF- surface adhesion protein putative
+PP1371-methyl-accepting chemotaxis transducero
PP3126-polysaccharide export proteino
PP3128- exopolysaccharide biosynthesis/transport protein, putative
PP3932-GGDEF domain proteino
PP4333-CheW domain proteino
+PP4377-flagellin FlaG putative
+PP4378 fliC -flagellin FliC
PP4383 flgI -flagellar P-ring protein precursor FlgIo
 Stress response  
+PP0915 sodB -superoxide dismutase (Fe)
+PP1209-cold-shock domain family proteino
+PP1360 groES -chaperonin 10 kDao
+PP1361 groEL -chaperonin 60 kDao
+PP2439 ahpC-alkyl hydroperoxide reductase C subunito
+PP4728 grpE -heat shock protein GrpE
+PP5000 hslV -heat shock protein HslVo
+PP5001 hslU -heat shock protein HslVU ATPase subunit HslUo
Transcription factors/sensors/regulators  
PP0191 pfrA anti-RNA polymerase sigma 70 factoro
PP0537 transcriptional regulator putativeo
PP0792 fruR transcriptional regulator (Cra)o
+PP1182-sensor histidine kinaseo
+PP2220-C4-type zinc finger protein DksA/TraR familyo
PP3179-transcriptional regulator LysR family
PP3412-DNA-binding response regulator LuxR familyo
PP3439-transcriptional regulator AraC familyo
PP3835-sensory box histidine kinaseo
PP4244 pfrI extracytoplasmic sigma factoro
PP4696 cbrB -transcriptional regulator
PP4781-sensor histidine kinaseo
PP5350-transcriptional regulator RpiR familyo
PP5375-transcriptional regulator LysR family
PP5383 copR -transcriptional activator CopRo
PP5384 copS -sensor protein CopSo
Others  
+PP0268 oprQ -outer membrane protein OprE3
+PP0883-porin putativeo
PP1083-bacterioferritin-associated ferredoxin putativeo
+PP1185 oprH-outer membrane protein H1o
+PP3155-outer membrane ferric siderophore receptor putativeo
PP3764-porin putative
PP3931-transporter sodium/sulfate symporter familyo
PP5207-ABC transporter ATP-binding protein/permease proteino
PP5208-ABC transporter permease proteino
PP5322-metal ion transporter putativeo

The regulated genes were grouped into six broad categories: amino acid uptake and/or metabolism, carbon uptake and/or metabolism, stress response, cell adhesion, motility and chemotaxis, sensor/regulator and a sixth group that included a number of genes that do not fit with the others but that we consider relevant (Table 2). The category of amino acid uptake and/or metabolism included several open reading frames related to transport of amino acids (PP1059, PP1068–70 and PP1138–41), as well as catabolic genes for histidine (hutGHU: PP5029, PP5032–33), tyrosine (tyrB1: PP1972; maleylacetoacetate isomerase: PP4619), arginine (arcA: PP1001), aspartate (aspA: PP5338), ornithine (ornithine cyclodeaminase: PP3533) and serine (serC: PP1768). All of the selected genes involved in amino acid uptake were downregulated in the cbrB mutant strain and, thus, are presumably CbrB-activated. The hutGHU, tyrB1, maleylacetoacetate isomerase, aspA and serC genes, which are involved in the catabolism of histidine, tyrosine, aspartic acid and serine, respectively, were also regulated in the same direction, whereas arcA and PP3533 appeared to be CbrB-repressed.

We also detected differential expression (in the cbrB mutant as compared with the wild type strain) of genes whose products are involved in carbon uptake or utilization of non-nitrogenated carbon substrates, such as dctP (dicarboxylate transport, PP1169) and rbsB (ribose transport, PP2454), or those involved in glucose utilization through the Entner-Doudoroff pathway (PP1022–PP1024). In relation to the tricarboxylic acid cycle, CbrB appears to favour the complete TCA cycle because it activates the gene encoding isocitrate dehydrogenase while repressing the gene encoding isocitrate lyase, which is the first enzyme of the glyoxylate shunt. In agreement with this expression pattern, we found that the cbrB mutant was unable to grow on citrate (see Fig. S1). In this category, we have also included a number of genes whose products are involved in membrane biogenesis (PP1912, PP4063 and PP5365), all of which appear to be repressed by CbrB.

Among the genes in the group of stress responses, we included those encoding several chaperonins, such as groEL-ES (PP1360–1361), cold-shock or heat-shock proteins, such as gprE (PP4728), and the hslVU operon, which encodes a heat-inducible general stress response proteasome that recycles misfolded proteins (PP5000 and PP5001) (Goldberg, 2003). Additionally, we included genes presumably involved in oxidative stress tolerance because they encode enzymes involved in detoxification, such as sodB (superoxide dismutase B, PP0915) and ahpC (alkyl hydroperoxyde reductase subunit C, PP2439). All genes in this category appeared to be activated by CbrB.

A number of genes related to bacterial cell adhesion, motility or chemotaxis were clustered in another category. This group included genes involved in polysaccharide biosynthesis and export (PP3126, PP3128), which contributes to cell-to-cell adherence, the surface adhesion protein LapF (PP0806) and a protein (PP3932) containing the GGDEF domain, which is known to be involved in adhesion of bacteria to solid substrates. Regarding the components of the flagella synthesis and biogenesis machinery, the genes coding for both the putative flagellin, FlaG (PP4377), and the structural subunit, FliC (PP4378), appear to be CbrB-activated, whereas PP4383, which encodes the flagellar P-ring precursor, FlgI, appears to be CbrB-repressed. Finally, three genes encoding putative chemotaxis proteins also have altered expression: PP0320 and PP1371, encoding methyl-accepting chemotaxis transducers, and PP4333, coding for CheW.

In the fifth group, we clustered a total of 16 genes encoding transcription factors, regulators or sensors, which suggests that CbrB may be indirectly regulating a number of different processes. Some genes appear activated, whereas others appear repressed by CbrB. Although potentially involved in gene regulation, the precise function of many of these genes is undefined. Nevertheless, these genes include: fruR (PP0792), which encodes the d-fructose-responsive global regulator that, in Escherichia coli (also called Cra), controls different pathways of carbon metabolism; the copR and copS genes, encoding a two-component regulatory system involved in copper homeostasis in Pseudomonas; and the genes encoding the transcription factors PfrI (an extracytoplasmic sigma factor similar to PvdS, PP4244) and PfrA (an anti-sigma 70, PP0191), both of which are required for biosynthesis of siderophores in fluorescent Pseudomonas. All of these regulatory genes with putative functions appear upregulated in the cbrB mutant strain. Among them is also cbrB (PP4696), which indicates that CbrB negatively autoregulates its own synthesis.

Finally, we delimited a sixth group containing genes encoding putative outer membrane proteins (oprQ: PP0268, oprH: PP1185, PP0883 and PP3764) or transport systems (PP3931, PP5207-08, PP5322) and a putative bacterioferritin-associated ferredoxin (PP1083), together with a number of genes encoding hypothetical proteins of unknown function (not included in Table 2 for simplicity).

In order to confirm some of the results from the microarrays, we performed expression analysis by quantitative RT-PCR using the same conditions as in the microarray assays. For this analysis, we chose tyrB1 (PP1972), encoding a tyrosine aminotransferase, hutU (PP5033), a urocanate hydratase for histidine utilization, lapF (PP0806) encoding the adhesion protein, LapF, and cbrB (PP4696), encoding the response regulator CbrB. RT-PCR was performed with total RNA from the wild type and cbrB strains grown to mid-exponential phase in minimal medium supplemented with a mixture of the 20 amino acids as the carbon and nitrogen source, in order to allow specific induction. The results are shown in Table 3.

Table 3.  Verification of selected DNA microarray results by quantitative RT-PCR.
 Ratio of expressiona
RT-PCRbMicroarrays
  • a. 

    Positive values for the ratio of expression in the arrays indicated that the gene was upregulated in the wild type strain, and negative values indicated that it was upregulated in the cbrB mutant strain. The same criteria were used for the quantitative RT-PCR data.

  • b. 

    The quantitative RT-PCR values are the means ± standard deviations of at least three independent experiments.

  • c. 

    The amplified cbrB region is located upstream of the transposon insertion in the cbrB mutant.

tyrB1 (PP1972)2.2 ± 0.91.9
hutU (PP5033)2.2 ± 1.12.1
lapF (PP0806)−4.5 ± 1.2−2.4
cbrB (PP4696)c−5.3 ± 1.8−1.9

The gene regulation, expressed as the ratio between expression in the wild type and the cbrB mutant strain, followed the same trend as the microarray data for the four genes analysed, and, in general, the fold change was higher than that obtained by the microarray analysis, thus confirming the previous results. The genes involved in the utilization of tyrosine and histidine were shown to be activated by CbrB. Low expression of these genes in the cbrB mutant is consistent with the growth defect previously observed in medium containing histidine or tyrosine as the sole carbon source. Quantitative RT-PCR analysis also confirmed that expression of lapF was upregulated in the cbrB mutant strain as compared with the wild type, indicating that it is repressed by CbrB. Finally, cbrB also had increased expression in the mutant strain, thus confirming that CbrB synthesis is negatively autoregulated.

The cbrB mutant shows altered metal ion homeostasis

The results of the transcriptomic analysis showed that the expression of some genes potentially involved in bacterial metal homeostasis was altered in the cbrB mutant. Some of these genes, such as copRS, pfrI and pfrA, and those coding for PP1083, PP3155 and PP5322 may be involved in bacterial interactions with metals in the environment.

To test directly whether the cbrB mutant displayed altered resistance to metal ions, serially diluted cultures of KT2442 and the cbrB mutant MPO406 were plated on LB-agar plates containing different concentrations of nickel, copper, iron, cobalt or zinc. The metal concentrations used on the plates were determined experimentally as those close to the concentrations preventing full growth of the wild type strain (data not shown). No difference whatsoever was found between the wild type strain and the cbrB mutant in their tolerance to nickel, iron or cobalt (not shown). Tolerances of both strains to copper and zinc, assessed as their survival on plates containing these metals, are shown in Fig. 1A and B. Viability of the wild type strain was reduced by the increasing metal concentrations (Fig. 1B). Although differences are small, these results indicate that the cbrB mutant is more tolerant than the wild type to copper and zinc, suggesting that some of the CbrB-repressed genes involved in metal metabolism may be involved in tolerance to these two metals.

Figure 1.

A. P. putida KT2442 and cbrB mutant strains' resistance to copper and zinc. Serial dilutions of each strain were plated on LB agar with no metal, 5 mM copper or 4 mM zinc added.
B. Survival was estimated as colony forming units observed for the wild type strain (black bars) and cbrB mutant (white bars) growing on LB agar with increasing concentrations of both metals.
C. Growth of the wild type strain (triangles) and the cbrB mutant (squares) in minimal M9 medium with the usual iron concentration (34 µM) (black) and in a similar medium with no iron added (white).

Iron scavenging ability was also tested because some of the genes with altered expression in the mutant may be involved in siderophore formation or uptake of the siderophore–iron complex. As shown in Fig. 1C, in the presence of iron, the cbrB mutant showed a longer lag phase in this minimal medium, but grew at the same rate as the wild type (doubling time of 60 and 67 min for the wild type and cbrB mutant respectively). Pseudomonas putida is a very efficient iron scavenger and can grow efficiently in minimal medium with no iron added (growth rate 70% of that with iron added), and reached the same cell density as the culture with iron added. However, the cbrB mutant was clearly more sensitive to iron limitation as it grew very slowly and reached half of the cell density of the culture with iron added (after 25 h, not shown). This result suggests that it has diminished iron scavenging capability.

The cbrB mutant has altered swimming motility

We showed above that several genes from the flagellar motility/chemotaxis cluster are differentially expressed in the presence of the cbrB mutation. To test whether these changes in expression lead to alterations in motility behaviour, we performed swimming motility assays using minimal medium soft agar plates, as described previously (Robleto et al., 2003). The results are shown in Fig. 2.

Figure 2.

Swimming motility assays for P. putida KT2442, cbrB and fleQ mutant strains in semisolid agar. The relative motility of the two mutant strains as compared with the wild type is shown (A) in M9 minimal medium supplemented with proline (left) or succinate plus ammonium (right). Plates were incubated at 30°C for 24 h before motility was scored. The results are based on three independent experiments. The picture (B) shows the result of one of the experiments.

The cbrB mutant strain showed about a 10-fold decrease as compared with wild type in swimming motility through M9 minimal medium soft agar, which was supplemented with 0.2% proline as the sole carbon and nitrogen source. Because the cbrB mutant exhibits slow growth in these conditions (Table 1, Fig. S1), a limited energy supply may be at least partly responsible for the observed defect. In order to avoid effects due to the growth rate, the same assay was also performed in soft agar medium supplemented with succinate and ammonium as a carbon and nitrogen source, respectively, which support growth of the cbrB strain at the same rate as the wild type (Fig. S1). Again, the cbrB mutant strain showed a clear, although somewhat smaller (2.5-fold), decrease in motility as compared with the wild type. The fleQ mutant used as a control was non-motile in both media.

Assayed this way, the altered motility phenotype of the mutant strain may be due to a defect in flagellar structure or function, impaired chemotaxis or both. In order to discriminate between these possibilities, we used microscopy to visualize the flagella of the cbrB mutant and wild type strains. Flagella staining and visualization revealed that P. putida KT2442 has several polar flagella, and the frequency of flagellated cells in the cbrB mutant is similar to that in the wild type (see Fig. S2). Confocal laser scanning microscopy and immunofluorescent staining of the flagellum with anti-flagellin antiserum failed to reveal significant differences in flagellar number or structure between the wild type and the cbrB mutant (Fig. S2). In addition, phase contrast microscopy of fresh cells grown in LB revealed that both strains were similarly motile (not shown), thus indicating that flagella are functional in the cbrB strain. We therefore propose that the decrease in motility detected in the soft agar plates when compared with the wild type may be due to a defect in chemotaxis towards non-exhausted medium.

The cbrB mutant exhibits an aggregative phenotype that results in increased biofilm and pellicle formation and flocculation

Pseudomonas putida forms biofilms efficiently on both biotic and abiotic surfaces (Espinosa-Urgel et al., 2000; Tolker-Nielsen and Molin, 2004; Gjermansen et al., 2006). The transcriptomic profile of the cbrB mutant shows significant changes in the expression levels of several genes encoding functions potentially involved in biofilm formation, such as signal transduction proteins containing GGDEF domains, exopolysaccharide biosynthesis or transport and the adhesin protein, LapF. To test whether the cbrB mutant showed differences in biofilm formation, its ability to colonize a plastic surface (polystyrene) was assessed by growing static cultures of the wild type and the cbrB, ntrC and cbrB ntrC mutants in LB and minimal broth in the wells of microtitre plates. The results are shown in Fig. 3A and B.

Figure 3.

Aggregative phenotypes of P. putida KT2442 and the cbrB, ntrC cbrB and fleQ mutants. Biofilm formation on LB medium (A) and M9 minimal medium supplemented with a mixture of the 20 amino acids (0.6 mM each) (B). Crystal violet rings were dissolved in ethanol, and A600 was read. Growth at 30°C (dark bars) was estimated by direct recording of A620 8.5 h after inoculation. Biofilm formation (clear bars) was quantitated as the capacity of the surface-attached biomass to retain crystal violet at the same time. Values are the means ± standard deviations of eight replicates in at least three independent experiments.
C. Overnight growth of P. putida KT2442 and cbrB mutant strains in LB medium at 30°C. The top arrow shows the biofilm at the liquid–air interphase, and the bottom arrow indicates the clumps formed by the cbrB mutant.
D. Flocculation test for the wild type strain (triangles) and the cbrB mutant (squares) based on sedimentation rates measured as the absorbance (at 600 nm) decay of the top of the liquid column of the culture over time.

The cbrB mutation resulted in a significant increase (threefold to fourfold) in the biomass attached to the wells of the polystyrene plates in both culture media. In contrast, the ntrC mutation did not affect biofilm formation in this system. A similar increase in planktonic growth was not observed because the cbrB and cbrB ntrC mutants displayed a somewhat lower OD620 than the wild type and ntrC strains, indicating that the increase in attachment is not the result of faster growth. As expected, a mini-Tn5 insertion mutant in the fleQ gene, used as a negative control, did not show any significant biofilm formation in this assay.

Biofilm formation in the liquid–air interphase was assessed by growing the wild type and cbrB strains in LB broth and visually assessing the accumulation of biomass on the surface by formation of a pellicle (Fig. 3C). A significant pellicle was observed in saturated cultures of the wild type strain, indicating that P. putida KT2442 can, to some extent, colonize the medium–air interphase. An equivalent culture of the cbrB mutant showed a much thicker pellicle, suggesting an increased ability to form a biofilm in this condition. Additionally, the cbrB mutant aggregated in liquid cultures and sedimented shortly after its removal from the shaker incubator, as shown in flocculation tests (Fig. 3D), indicating that cells in the culture medium clustered in large clumps. Taken together, our results indicate that the cbrB mutation provokes an aggregative phenotype that increases the formation of surface-attached communities by P. putida KT2442.

Discussion

Phenotypic microarrays, transcriptomic microarrays and growth tests of the wild type and the cbrB mutant strains clearly showed that the P. putida strain KT2442 response regulator CbrB is involved in utilization of a number of amino acids as carbon sources. These substrates can also be utilized as nitrogen sources, and the growth phenotype of an NtrC mutant (Table 1) indicated that their utilization as nitrogen sources is subjected to regulation by the general nitrogen control system, Ntr. This result suggests that some genes involved in these substrates' utilization may be subject to dual control by the Cbr and Ntr systems. This view is supported by the growth phenotype of the cbrB ntrC double mutant on some amino acids (e.g. proline) as compared with the single mutants. The more severe growth defect of the double mutant when using proline as the carbon or the nitrogen source suggests that CbrB and NtrC may at least partially substitute for each other in activating proline utilization. Conversely, it appears that NtrC may interfere with the function of CbrB and CbrB may interfere with the function of NtrC when activating the utilization of other amino acids, such as histidine. Despite particular differences in the way that the utilization of some amino acids is controlled among different species of Pseudomonas, our results are in full agreement with those previously reported for P. aeruginosa and P. fluorescens, which reveal the involvement of CbrB in amino acid catabolism (Nishijyo et al., 2001; Zhang and Rainey, 2007) and the relationship between the Cbr and the Ntr regulatory systems (Li and Lu, 2007; Zhang and Rainey, 2008a).

Global transcriptomic analysis also revealed altered expression of genes presumably affecting the Krebs/glyoxylate cycles, uptake of carbon sources and utilization of sugars through the major catabolic pathway in P. putida (Entner-Doudoroff), as glycolysis is incomplete. Interestingly, genes of the Entner-Doudoroff pathway are regulated by the carbon catabolite repressor Crc, which acts by binding to mRNA targets and repressing translation of the downstream gene (Collier et al., 1996; del Castillo and Ramos, 2007; Moreno and Rojo, 2008). A connection between the Cbr system and the catabolite repression exerted through Crc has been reported recently in P. aeruginosa; CbrB activates transcription of crcZ, which encodes a small RNA resembling the Crc target binding sites (Sonnleitner et al., 2009). However, this link may not be detected by transcriptomic analysis in many instances because Crc regulates mRNA translation. The fruR gene also showed altered expression in the cbrB mutant. Its orthologue in E. coli, encoding the catabolite repressor-activator (Cra) protein, controls carbon flux through different metabolic pathways (Ramseier, 1996; Saier, 1996). Cra is a dual regulator that represses transcription of genes encoding glycolytic and Entner-Doudoroff enzymes, whereas activating genes encode Krebs cycle, glyoxylate shunt and gluconeogenic enzymes. The Cra-activated genes show catabolite repression, whereas the Cra-repressed genes show high expression under the same conditions. Although the function of FruR in Pseudomonas has not been studied, its altered expression in the cbrB mutant suggests a connection between the Cbr regulation system and carbon catabolite regulation in P. putida. These expression data, together with the severe growth defect of the cbrB mutant when citrate is the only carbon and energy source (see Fig. S1), indicate that CbrB involvement in regulation of carbon catabolism may be extended from amino acids to other non-nitrogenated carbon sources and to central carbon catabolic pathways. A more extensive analysis of potential particular targets will allow further characterization of the link between CbrB and carbon source utilization.

Further characterization of the P. putida cbrB mutant phenotypes potentially related to genes whose expression pattern is altered indicated that CbrB may not be simply a metabolic regulator but may also regulate other functions unrelated to carbon or nitrogen metabolism. One of the functions affected in the cbrB mutant was tolerance to metal ions (Fig. 1). In relation to copper tolerance, the cbrB mutant showed increased expression of the genes encoding the CopRS two-component system, which is involved in copper homeostasis. Whether overexpression of this regulatory operon in the P. putida cbrB mutant would lead to increased copper tolerance is unclear because the molecular mechanisms of copper homeostasis are poorly understood. CopRS activates expression of the copABCD operon, which is involved in copper tolerance (Teitzel et al., 2006). On the other hand, P. putida insertion mutants in copA or copB showed very slightly increased tolerance (Adaikkalam and Swarup, 2005), which suggests that P. putida copABCD is not involved in this homeostasis. CopRS may be involved in copper tolerance through copCD only because, in P. fluorescens (which lacks CopAB), CopCD by themselves affect copper homeostasis (Zhang and Rainey, 2008b). Intriguingly, a P. fluorescens SBW25 copS deletion increases tolerance to copper (Zhang and Rainey, 2008b). Alternatively, CopRS could exert its effect by activating other uncharacterized operons. Obviously, further characterization is required to establish a direct link between CbrB and copper tolerance. The connection of zinc tolerance with CbrB is less obvious; we did not detect altered expression of any gene that might be involved. However, P. aeruginosa copRS induces both copper and zinc resistance (Caille et al., 2007). First, this finding connects the zinc-resistant phenotype with the altered expression of the regulatory operon. Second, it shows that copRS can regulate different metal resistance operons.

The growth phenotype of the cbrB mutant on medium lacking added iron indicated that the mutant had reduced scavenging ability as compared with the wild type strain. However, the apparent opposing effect of the cbrB mutation on expression of genes potentially involved in iron scavenging is intriguing. Consistently with its growth phenotype, we detected downregulation in the mutant of an outer membrane protein presumably involved in uptake of the iron–siderophore complex, thus suggesting that CbrB enhances iron scavenging capacity. However, we also detected upregulation of the pfrI and pfrA genes, which encode an extracytoplasmic sigma factor (PvdS in P. aeruginosa) and an anti-sigma 70, respectively, in the cbrB mutant strain. Although expression of siderophore biosynthetic genes is subject to negative control by the repressor Fur in the presence of iron, PfrI and PfrA are both required for siderophore biosynthesis in fluorescent Pseudomonas (Venturi et al., 1995; Leoni et al., 2000). This result suggests that siderophore production and, consequently, iron-scavenging capacity would be enhanced in the cbrB mutant. The iron scavenging capacity may be impaired in the mutant, despite the potentially enhanced siderophore production.

In addition, several members of the heat shock response system, including chaperonins (groES-groEL, grpE) and the HslVU protease, together with enzymes involved in the oxidative damage response system, were activated by CbrB, suggesting that the Cbr regulon is at least partly accountable for a general stress response that might be elicited by either nutrient limitation, oxidative damage or an as yet unknown stress signal. However, we have not identified any mutant phenotype related to these stress responses.

Motility is a very important feature in microbial physiology because it enables bacteria to drive themselves through the environment and promote optimal positioning of the cells. Motility yields the benefits of increased efficiency of nutrient acquisition, avoidance of toxic compounds or dispersal in the environment. The microarray data showed that the cbrB mutant has altered expression of some of the genes implicated in flagella synthesis and assembly and others encoding proteins putatively involved in chemotaxis (Table 2). In addition, lack of CbrB also resulted in clearly reduced bacterial motility in soft agar (Fig. 2), which supports the view that CbrB somehow controls motility. However, the cbrB mutant produces apparently functional flagella as abundantly as the wild type strain. Thus, the reduced motility of the cbrB mutant, which is assayed as the expansion capacity towards nutrient-rich zones, may indicate altered chemotactic responses that prevent the mutant from swimming efficiently towards microenvironments that provide optimal conditions for growth and survival, as has been shown for other chemotactic mutations (Kato et al., 1999; Ames et al., 2002; Barken et al., 2008). In support of this view, three genes potentially encoding chemotactic proteins showed altered expression in the cbrB mutant. Regardless of the underlying mechanism, it is clear that the cbrB mutation impaired the capacity of the bacteria to colonize more adequate niches.

A remarkable outcome of our analysis is the effect of the cbrB mutation on intercellular adhesion and colonization of solid surfaces. The cbrB mutant displayed an aggregative phenotype and develops a substantially thicker biofilm than the wild type strain both on solid–liquid and liquid–gas interphases. Various extracellular components, such as protein adhesins, proteinaceous appendages, exopolysaccharide and extracellular DNA, are implicated in the formation and maintenance of biofilms (Pamp et al., 2007). Consistently, transcriptomic analysis revealed upregulation in the cbrB mutant of several genes whose products are presumably involved in the biofilm formation process, such as those encoding exopolysaccharide export proteins, others with the GGDEF domain regulating expression of polysaccharides and proteins of the extracellular matrix (reviewed in Cotter and Stibitz, 2007). An additional example is LapF, a member of a family of ubiquitous large surface-attached proteins containing repetitive motifs; several of these are involved in cell–surface and cell–cell adhesion in Pseudomonads and in the adhesion of plants to solid surfaces (reviewed by Lasa and Penades, 2006). Biofilm formation is a developmental process in which large suites of genes are differentially expressed. In P. putida, this process is prevented under carbon-limiting conditions, and, in fact, preformed biofilm is dispersed as a response to nutrient starvation (Gjermansen et al., 2005; Klausen et al., 2006). Because CbrB would be mostly active under carbon-limiting conditions, our results suggest that CbrB regulates biofilm formation through repressing, either directly or indirectly, expression of genes required for the synthesis of multiple components involved in the cell surface architecture that promote biofilm formation. In the mutant, biofilm is formed much more efficiently.

The CbrAB regulatory system is generally associated with catabolism of nitrogen-containing compounds, such as amino acids and polyamines (Nishijyo et al., 2001). Although our results are consistent with this role of CbrB in P. putida, the identification of multiple other functions unrelated to carbon metabolism that are subject to CbrB-dependent regulation supports a broader cellular function for CbrB. Some of these functions may be related to the optimal adaptation to carbon limitation and could be regulated by the same environmental signal (i.e. carbon limitation) through CbrB. Under this condition, CbrB would extensively reprogramme the bacterial carbon metabolism, prevent establishment of the bacteria in this niche by preventing biofilm formation and promote chemotaxis to zones with higher nutrient concentration. Additionally, altered expression of other genes involved in other stress responses and additional phenotypes in the cbrB mutant indicate that the function of the cbr system may go even beyond the control of the different responses to nutrient limitation. We propose that CbrB is a highly ranked element in the regulatory hierarchy that controls multiple processes that may be important for optimal fitness of the bacteria in a changing environment. The large number of regulatory genes with altered expression in the cbrB mutant, which in turn could be involved in controlling particular sets of genes, supports this view.

Experimental procedures

Bacterial strains and growth conditions

All P. putida strains used were derived from the strain KT2442 (Franklin et al., 1981). Their phenotypes are shown in Table 4. The cbrB mutant strain contains a mini Tn5-luxAB-Km insertion at a position 669 bp downstream of the translational start of CbrB. Pseudomonas putida strains were grown at 30°C in LB medium or in M9 minimal medium (Mandelbaum et al., 1993), the latter containing sodium succinate (20 mM) as the carbon source and ammonium chloride (1 g l−1) as the nitrogen source. For growth tests, 25 ml cultures in 100 ml flasks were inoculated at an initial A600 of 0.05, incubated in M9 minimal medium containing proline, arginine, ornithine, glutamate, histidine (20 mM) or tyrosine (5 mM) as carbon, nitrogen or carbon and nitrogen sources and were monitored. Growth was followed by measuring turbidity at 600 nm (A600) every 2 h for up to 21 h in some cases. For expression analysis, inocula were incubated with shaking (180 r.p.m.) at 30°C until they reached the exponential phase (A600 = 0.6). The cells were collected and resuspended in 100 ml of fresh M9 minimal medium containing a mixture of the 20 amino acids as carbon and nitrogen sources at A600 = 0.15. This composition allowed specific induction of the pathways for the uptake or assimilation of any of the amino acids assayed. The cultures were then incubated for 2 h for specific induction, and cells were then collected by centrifugation, frozen in liquid nitrogen and stored at −80°C for RNA purification. Antibiotics were added, when necessary, at the following concentrations: 500 µg ml−1 carbenicilin, 25 µg ml−1 kanamycin, 5 µg ml−1 tetracycline and 20 µg ml−1 rifampicin. For solid media, Bacto-Agar (Difco, Detroit, MI) was added to a final concentration of 15 g l−1.

Table 4.  Bacterial strains used in this work.
NameRelevant characteristicsOrigin
  • a. 

    The MniTn5-luxAB-Km is inserted in codon 223 of cbrB in a way that luxAB orientation is opposite to that of cbrB.

Pseudomonas putida KT2442mt-2 hsdR1 RifrFranklin et al. (1981)
P. putida MPO201mt-2 hsdR1 RifrΔntrC::TcGarcia-Gonzalez et al. (2005)
P. putida MPO406amt-2 hsdR1 RifrcbrB:: MiniTn5-luxAB-KmJ.L. Ramos
P. putida MPO400mt-2 hsdR1 RifrΔntrC::Tc cbrB::MiniTn5-luxAB-KmThis work
P. putida fleQmt-2 hsdR1 RifrfleQ:: MiniTn5-KmA. Platero and F. Govantes (unpublished)

Phenotype microarray analysis

The phenotype microarray analysis was carried out by the Biolog commercial service (Biolog, Hayward, CA) using the PM Kit (Biolog Catalog # 12191), which contains 20 microplates containing the same medium but different carbon or nitrogen sources (Bochner et al., 2001). Phenotype microarray analysis use Biolog's redox chemistry, employing cell respiration as a universal reporter and the tetrazolium dye as redox indicator. The OmniLog instrument captures a digital image into kinetic graphs for 24 h. Differences in the plotted areas of redox activity monitored for 24 h between the wild type and cbrB mutant strains are quantified as scores.

Flocculation assay

An aliquot (600 µl) of overnight culture grown on LB medium was diluted and placed in a 1 ml spectrophotometric cuvette, and the absorbance was monitored over time. Sedimentation of the wild type and the cbrB mutant strains, estimated as the absorbance decay over time, is plotted in Fig. 3D.

RNA purification

Cell pellets were resuspended in 1 ml of TriPure Isolation Reagent (Tri Reagent LS, Molecular Research Center, Cincinnati, OH) and incubated at 60°C for 10 min for complete cell lysis. The solution was centrifuged at 13 000 r.p.m. for 10 min at 4°C, and the pellet was discarded. The samples were transferred to a 2 ml tube of Phase Lock Gel (Eppendorf) and 0.2 ml of chloroform was added to the supernatant, shaken, centrifuged and extracted with 1 vol. of phenol : chloroform : isoamylalcohol (25:24:1). After isopropanol precipitation, the pellet was resuspended in 150 µl of DEPC water. The samples were treated with DNase I (10 units, Roche) in the presence of 1 mM DTT, 10 mM MgCl2 and RNase inhibitor (20 units of Rnase Out, Invitrogen). After extracting the samples twice with equilibrated phenol : chloroform : isoamylalcohol (25:24:1) in a 2 ml tube of Phase Lock Gel and once more with acid phenol, the RNA was recovered by precipitation with sodium acetate and ethanol at −80°C. RNA integrity was checked by agarose gel electrophoresis, and the samples were purified using RNeasy columns (Qiagen, clean-up protocol).

P. putida genomic DNA microarrays, processing and data analysis

The genome-wide DNA microarray used in this work was described previously (Yuste et al., 2006). In brief, 20 µg of total RNA was transformed to cDNA with Superscript III reverse transcriptase using random hexamers as primers. After cDNA purification, each condition to be analysed was labelled with Cy3 or Cy5 fluorescent dyes (Amersham Biosciences). The microarray was scanned in an Axon GenePix 4000B microarray scanner with green and red lasers operating at 543 and 633 nm, respectively, to excite Cy3 and Cy5. Images were taken at 10 µm resolution and spot intensity was determined using the Genepix Pro 5.0 software. For each experiment, a minimum of three independent RNA extractions were made, and each extraction was analysed with at least two microarrays (three biological replicas, two technical replicas each). After background subtraction, signal intensities for each replica were normalized and statistically analysed using the Lowess intensity-dependent normalization method (Yang et al., 2002) included in the Almazen System software (BioAlma Bioinformatics S.L., Madrid). P-values were calculated with the Student's t-test algorithm based on the differences between log 2 ratio values for each replicate. Genes were considered differentially expressed when they fulfilled the following filter parameters: expression ratio of > 1.8 or < −1.8 and a P-value of < 0.05 (Benjamini et al., 2001).

Quantitative RT-PCR

Confirmation by quantitative RT-PCR of the expression of selected open reading frames was performed as described previously (Hervas et al., 2008) using 10 µg RNA for RT and random hexamers as primers. Target cDNAs (10 ng) from the experimental and reference samples were amplified in quadruplicate in separate PCR reactions using 0.3 µM of each primer. The PCR products were between 50 and 100 bp in length. A standard curve was made using serial dilutions from 25 to 0.0025 ng of P. putida KT2442 genomic DNA in order to quantify the relative abundance of transcripts in each sample. A melting curve analysis was performed for each amplification to identify non-specific amplification according the manufacturer's recommendations. The primer pairs used for each gene were: 5′-ACTTCCTCAAGCGCCTCTCG-3′ and 5′-CTAGGGTCGCTGATGGCAAC-3′ for tyrB1; 5′-GCAATCCGAAGTCGCCACT-3′ and 5′-GCGTTTAGATTGATATTGCCCC-3′ for lapF; 5′-ATCGAATGCCAGCAGAGCC-3′ and 5′-GACGTAGCGGGTTTCCAGG-3′ for hutU; and 5′-GCAGGACATGTACAGCAAGATCC-3′ and 5′-GCACATTGGAGTCGGTAGGC-3′ for cbrB.

Resistance to metal ions assays

Serial dilutions of mid-exponential cultures of P. putida KT2442 or its isogenic cbrB mutant strain at A600 0.5 were plated on LB plates supplemented with copper or zinc at the highest concentration of the metal that did not inhibit the growth of P. putida after overnight incubation at 30°C (i.e. 4 mM ZnCl2 and 5 mM CuCl2). Survival was estimated as the colony forming on the plates containing the metals, as compared with the viable counts on LB plates with no metal added. To quantify this tolerance, the viability of the wild type and cbrB mutant strains was estimated by plating serial dilutions of each strain in plates containing increasing concentrations of copper or zinc.

Iron scavenging capacity

The growth of the wild type and the cbrB mutant grown in a M9 minimal medium supplemented with succinate and ammonium with no iron added or containing 34 µM FeSO4 was monitored for 25 h at 30°C.

Motility assay

To study swimming motility, semisolid LB medium or M9 minimal medium containing succinate and ammonium (1 g l−1) or 20 mM proline (0.2% [wt/vol] Bacto Agar) were inoculated with a single fresh LB colony using a sterile wire. After incubation for 24 h at 30°C, the movement of the bacteria away from the inoculation point was determined relative to the movement of the P. putida wild type strain, as described previously (Robleto et al., 2003).

Flagella staining procedure

A single fresh LB colony of the P. putida KT2442 or cbrB mutant strain was carefully inoculated with a wire into a sterile water drop on a microscope slide and covered with a cover slip. Motility was monitored using an optical microscope prior to staining of the samples. The samples were incubated for 10 min at room temperature so that at least half of the cells could adhere to the surface, and two drops of Flagella Stain Reagent containing 0.6% crystal violet in ethanol, 2% tannic acid, 2.5% phenol and 5.7% aluminium potassium sulphate (BD BBL Diagnostic) were subsequently added at the edge of the cover slip, allowing its entrance to the sample by capillary action. After 15 min of incubation at room temperature, stained bacteria were visualized via phase contrast microscopy using a Leica DMI 4000B Inverted Microscope (Leica Microsystems GmbH). The same preparation was also visualized via confocal laser scanning microscopy using a Leica TCS SP2 Spectral confocal system (Leica Microsystems GmbH).

Flagella immunofluorescence visualization

One fresh LB colony of P. putida KT2442, MPO406 (cbrB), P. aeruginosa (PAO1) and the P. putida fleQ mutant strain was fixed in 40 µl of 4% formaldehyde in phosphate-buffered saline on a microscope slide for 30 min at room temperature. After washing with PBT (phosphate-buffered saline plus 0.1% Triton), the slide was blocked with 7% foetal calf serum in PBT for 45 min, and anti-FliC antibody (1:1000) was added and incubated for another 45 min. After three 10 min washes with PBT, the fluorescent Alexa 568-conjugated goat anti-rabbit IgG-conjugated antibody (Invitrogen) at a 1:300 dilution was incubated for 90 min for detection. The bacterial chromosome was stained with Hoechst reactive dye for 10 min prior to visualization using a Leica DMI 4000B fluorescence microscope.

Quantification of biofilm formation

A stationary-phase culture of P. putida growing in LB or M9 minimal medium supplemented with 20 mM succinate plus 1 g l−1 ammonium was rinsed with phosphate buffer and diluted 1:20 in either LB or M9 minimal medium containing a mixture of the 20 amino acids (0.6 mM each) as sole carbon and nitrogen sources. Biofilm formation was quantified as described (O'Toole and Kolter, 1998). Briefly, the supernatant was removed from the culture, and the remaining cells were stained with crystal violet dye. The surface-attached cells were quantified by solubilizing the dye in 95% ethanol, and the absorbance was determined at 620 nm.

Acknowledgements

We are grateful to all members of the laboratory for their insights and helpful suggestions, as well as Guadalupe Martín Cabello and Nuria Pérez Claros for technical help. The P. putida KT2442 cbrB mutant was kindly provided by Juan Luis Ramos, and P. aeruginosa Anti-FliC was kindly provided by Dr Reuben Rampal. The assistance with microscopy by Beatriz Mesa, Carlos Medina and Katherina García is much appreciated. This work was supported by Grants BIO2007-63754 and CSD2007-00005 from the Spanish Ministry of Science and Technology, Proyecto de Excelencia CVI-131 from the Andalusian government and by a fellowship from the Andalusian government to C.I.A.

Ancillary