The Global Carbon Metabolism Regulator Crc Is a Component of a Signal Transduction Pathway Required for Biofilm Development by Pseudomonas aeruginosa

Bacterial strains, media, and chemicals.

All of the bacterial strains and plasmids used in this study are shown in Table 1. P. aeruginosa PA14 was grown on rich (Luria Bertani [LB]) medium or minimal medium (as indicated in each experiment) at 37°C. Unless otherwise stated, the minimal medium used was minimal M63 salts (28) supplemented with glucose (0.2%), MgSO₄ (1 mM), and Casamino Acids (CAA; 0.5%). Antibiotics were added at the following concentrations: tetracycline, 150 μg/ml; ampicillin, 750 μg/ml; carbenicillin, 1 mg/ml. All enzymes for DNA manipulation were purchased from New England Biolabs (Beverly, Mass.). All plasmids were constructed in E. coli JM109 by using standard protocols (1) and transferred to P. aeruginosa by electroporation (3).

Enzyme assays.

To generate cell material for enzyme assays, bacterial cultures were grown in 250 ml of minimal medium (18) with 20 mM mannitol as the sole carbon and energy source for at least 9 generations in log phase to fully induce the hex regulon (including glucose 6-phosphate dehydrogenase, mannitol dehydrogenase, and NAD-specific glyceraldehyde 3-phosphate dehydrogenase). An aliquot of cells was transferred to minimal medium containing 40 mM succinate and 20 mM mannitol, and growth was continued for 2 to 2.5 generations. Bacteria were chilled, centrifuged, and washed with 1/10 volume of cold minimal medium, and the cell pellets were frozen.

Extracts were prepared from frozen cell pellets by suspending the cells in 2 to 3 volumes of Tris-Cl (50 mM; pH 8)-dithiothreitol (1 mM). The suspended cells were broken by passage through a French press at 16,000 lb/in² and clarified by centrifugation at 170,000 × g for 30 min (50,000 rpm in a Sorvall T-1270 rotor). Mannitol dehydrogenase was assayed on freshly prepared extracts, while the other assays were performed on thawed extracts. Glucose 6-phosphate dehydrogenase and mannitol dehydrogenase were assayed as previously described (14, 34). The NAD- and NADP-dependent glyceraldehyde 3-phosphate dehydrogenase assays were similar to those of Tiwari and Campbell (35). Briefly, 1-ml reaction mixtures containing NAD or NADP at 0.5 mM; Tris-Cl at 100 mM (pH 8), sodium arsenate at 20 mM, l-cysteine at 3.3 mM, and sodium fluoride at 20 mM were incubated at 22°C. Reactions were started with the addition of glyceraldehyde 3-phosphate to 2 mM. For the determination of specific activities, extracts were assayed for protein by the method of Lowry et al. (17) with bovine serum albumin as the standard.

Twitching motility assays.

Twitching motility was assessed by stabbing cells into a very thin (2-mm) LB agar plate (1.5% agar) with a toothpick and incubating the plate for 24 to 48 h at 37°C. A haze of growth at the agar-plate interface and the characteristic flat, spreading colony morphology indicate strains that are proficient for twitching motility (21, 23).

Biofilm formation assay. (i) Screen for mutants defective in biofilm formation.

The assay used to screen for biofilm formation-defective mutants is based on the ability of bacteria to initiate biofilm formation on polyvinyl chloride (PVC) plastic. The initiation of biofilm formation was assayed as previously described (25) by determining the ability of cells to adhere to the wells of 96-well microtiter dishes made of PVC (Falcon 3911 Microtest III Flexible Assay Plate; Becton Dickinson Labware, Oxnard, Calif.) using a modification of a previously reported protocol (11). The medium (100 μl/well) was inoculated with either (i) cells patched to LB agar plates using a multiprong device or (ii) a 1:100 dilution from an overnight LB culture. After inoculation, plates were incubated at 37°C for the indicated time and then 25 μl of a 1% crystal violet solution was added to each well (this dye stains the cells but not the PVC). The plates were incubated at room temperature for ∼15 min, rinsed four times thoroughly and vigorously with water, blotted on paper towels, and scored for biofilm formation (see Fig. 1).

(ii) Quantitation of biofilm formation.

Biofilm formation was quantified as previously described (25). Briefly, the crystal violet was solubilized in 95% ethanol and the A₅₇₀ was determined.

(iii) Microscopy.

Visualization of bacterial cells attached to PVC was performed by phase-contrast microscopy (400× magnification) using a Nikon Diaphot 200 inverted microscope (Nikon Corp., Tokyo, Japan). The images were captured with a black-and-white CCD72 camera integrated with a Power Macintosh 8600/300 computer with video capability. The images were processed with Scion Image software, a modification of NIH Image (National Institutes of Health, Bethesda, Md.) by the Scion Corporation (Frederick, Md.).

Molecular techniques.

The DNA sequence flanking the transposon mutant insertion site was determined by using arbitrary primed PCR as previously reported (27). All other protocols were performed as previously reported (1, 25).

Plasmid construction.

Plasmid pSMC31 (crc⁺ Cb^r) carries a wild-type copy of the crc locus derived from P. aeruginosa PAO1. This plasmid was built by first amplifying crc⁺ from pPZ352 (19) by PCR using primers which generated a BamHI site at the 5′ end and an EcoRI site at the 3′ end of the crc gene. The PCR product was digested with EcoRI and BamHI and then ligated to pUCP18, which had also previously been digested with EcoRI and BamHI.

EM.

LB agar plates without NaCl were prepared and inoculated as described for the twitching motility assays above. Approximately 10 μl of distilled water (dH₂O) was dropped on the edge of the colony, the bacteria were allowed to become suspended for ∼30 s, and then the copper-coated grid was floated on the cell-containing dH₂O drop for ∼30 s, rinsed for 5 s with dH₂O, stained for 45 s with phosphotungstate (2%), rinsed for 5 s with dH₂O, and then gently blotted and air dried (D. Wozniak, personal communication). For each strain, at least 100 cells were examined. The pili are fragile structures; they are apparently easily lost during the preparation of electron microscopy (EM) grids. Samples were analyzed with a JEOL 1200 EX transmission electron microscope at 80 kV.

β-Galactosidase assays.

β-Galactosidase assays were performed as previously described without modification (22). The data presented are averages of two cultures assayed in duplicate (a total of four data points).

Isolation of the crc mutant.

We previously performed a screen for mutants of P. aeruginosa PA14 defective in biofilm formation and identified mutations in genes required for flagellum-mediated swimming and type IV pilus-mediated twitching motility (24). In this search for biofilm formation mutants, we also obtained a sad-24 mutant strain with an insertion in the crc gene. The biofilm formation phenotype of this mutant is shown in Fig. 1A.

Comparison of the DNA sequence flanking the sad-24::Tn5 insertion to DNA sequences in the GenBank database revealed 100% identity over a 34-amino-acid region to the previously described crc locus of P. aeruginosa PAO1 (19), and this allele was therefore designated crc-24. The location of the insertion is near the 3′ end of the crc gene, and the mutated gene is predicted to code for a protein whose C-terminal 38 amino acids are deleted and replaced with 9 amino acids derived from a transposon sequence. We observed other phenotypes in the strain carrying the crc-24::Tn5 mutation, including overproduction of a pigment (presumably pyocyanin [4]) and a decrease in type IV pilus-mediated twitching motility.

It is important to note that the growth rate of the crc mutant is indistinguishable from that of the wild type in glucose-containing minimal medium supplemented with CAA (the same medium and growth conditions used in the screen to isolate this mutant; data not shown). Therefore, the defect in the crc mutant appears to be specific to biofilm formation and not a general defect in growth. Finally, in addition to its defect in biofilm formation on PVC plastic, this strain is also defective for biofilm formation on polystyrene, polycarbonate, and polypropylene (data not shown).

Complementation of the crc mutation.

In order to confirm that the phenotypes of the crc-24 mutant strain were indeed caused by the mutation in the crc gene, we performed complementation analysis. The crc gene of P. aeruginosa PAO1 had been cloned previously (19), and using the cloned gene, we built a plasmid construct which carries only crc⁺ and no other complete open reading frame (pSMC31; see Methods and Materials for details of the plasmid construction). Figure 1B shows the results of the complementation assay performed on these strains. The plot in Fig. 1B shows that based on quantitation of biofilm formation using the crystal violet assay, plasmid pSMC31 (crc⁺) fully rescued the biofilm formation defect of the crc-24 mutant strain to levels equivalent to those of the wild type (carrying either plasmid). As expected, the mutant carrying the vector control remained defective in biofilm formation.

The result above was confirmed and extended by directly visualizing biofilm formation by these strains using phase-contrast microscopy (Fig. 1C). After incubation for 8 h, the wild type (carrying either plasmid) had formed a dense monolayer of cells punctuated with microcolonies on the surface of the PVC plastic. Examination of the crc-24 strain carrying pUCP18 (vector-only control) revealed a dispersed monolayer of cells markedly different from that of the wild-type strain. However, providing the crc-24 mutant with a wild-type copy of crc on pSMC31(crc⁺) resulted in a biofilm formation phenotype indistinguishable from that of the wild-type strain.

In addition to assessing biofilm formation, we also investigated whether pSMC31 (crc⁺) could complement the defect in twitching motility of the crc-24 mutant (Fig. 2). The hazy zone of growth around the wild-type colonies is the “twitching zone”; that is, the area containing cells which have used type IV pilus-mediated twitching motility to move away from the point of inoculation. The wild-type strain (carrying either plasmid) and the crc-24 mutant carrying pSMC31 (crc⁺) both formed an obvious twitching zone with an average size of 9.5 mm (measured as the distance from the colony edge to the edge of the twitching zone). In contrast, the crc-24 mutant carrying pUCP18 (vector control) produces a twitching zone about sixfold smaller (1.5 mm) than that produced by the wild-type strain. Furthermore, the crc-24 mutant carrying pSMC31 (crc⁺) no longer overproduced pyocyanin (data not shown). Based on these results, we conclude that the crc-24 mutation caused the observed phenotypes.

The biofilm formation phenotype of a crc mutant is identical to that of strains carrying mutations in genes required for type IV pilus biogenesis.

The crc-24::Tn5 mutant of P. aeruginosa PA14 is defective for twitching motility. This observation suggested that crc is required for the synthesis or function of type IV pili. As assessed by phase-contrast microscopy, the crc-24 mutant showed a pattern of dispersed cells on the PVC plastic surface after incubation for up to 8 h (Fig. 1C), a phenotype identical to that observed previously for mutants lacking type IV pili (24). Therefore, the crc mutant has a biofilm formation defect indistinguishable from that of a strain which lacks type IV pili.

The crc-24::Tn5 mutant produces type IV pili.

Three lines of evidence demonstrated that although the crc-24 mutant is defective in twitching motility, this strain still has at least some functional type IV pili. The first line of evidence is that the crc-24 mutant retains residual twitching motility, as shown in Fig. 2. Second, the crc-24 mutant remains sensitive to infection by bacteriophage SN-T (15), which utilizes type IV pili as a receptor (26). The abilities of phage SN-T to form plaques on the wild-type strain and the crc-24 mutant are indistinguishable (data not shown). It is interesting that although the crc-24 mutant has decreased twitching motility, it is still sensitive to phage SN-T infection. This observation may reflect the sensitivity of the phage infection assays, that is, the possibility that only a single phage or a few phage are sufficient for infection and lysis. The wild-type strain may be bound by many more phage particles than are required for cell lysis. No plaques were formed on other type IV pilus mutants (such as pilB and pilC mutants), even using a phage lysate at a titer of ∼3 × 10⁷ PFU/ml. Finally, we directly visualized type IV pili by using EM. Type IV pili are polar structures found at the same end of the cell as the flagellum. As shown in Fig. 3, in the wild-type strain the type IV pili are readily visible as multiple, polarly localized rods which are somewhat thinner than the single polar flagellum. These pili are also visible in the crc mutant, consistent with the results of the phage susceptibility tests and twitching motility assays. However, in the pilB strain, which is not susceptible to lysis by phage SN-T and retains no detectable twitching motility, no pili were visible.

A crc null mutant of P. aeruginosa PAO1 is also defective for biofilm formation.

As mentioned above, mutations in crc were originally isolated in P. aeruginosa PAO1 based on their defects in the regulation of carbon metabolism. However, these mutations were not tested for defects in biofilm formation. The constructed crc::Cb^r mutation in P. aeruginosa PAO1, like the crc::Tn5 mutation in P. aeruginosa PA14, results in a strain defective in biofilm formation (data not shown). We have also shown that a strain carrying the crc::Cb^r allele is also defective for twitching motility and remains sensitive to lysis by type IV pilus-specific bacteriophage SN-T. Prince and coworkers had also indicated that a crc mutant of P. aeruginosa PAO1 is defective in twitching motility (4). These data further confirm a role for crc in biofilm formation and demonstrate that the defect in biofilm formation is not specific to P. aeruginosa PA14.

Effects of crc-24 on carbon metabolism in P. aeruginosa PA14.

The crc locus was originally identified in P. aeruginosa PAO1 during a screen for genes that participate in catabolite repression control of enzymes involved in mannitol utilization (37). In order to determine if a P. aeruginosa PA14 crc mutant is also defective in the regulation of carbon metabolism (in addition to the other observed phenotypes), we assessed the crc-24 mutant isolated in the biofilm mutant screen for its effects on catabolite repression in P. aeruginosa PA14. These data are summarized in Table 2. In wild-type P. aeruginosa PAO1, the levels of enzyme activities required for glucose and mannitol utilization are repressed in the presence of succinate (29, 34). Similarly, levels of representative enzymes involved in mannitol and glucose catabolism in P. aeruginosa PA14 (mannitol dehydrogenase, glucose 6-phosphate dehydrogenase, and NAD-specific glyceraldehyde 3-phosphate dehydrogenase) are up to eightfold higher when cells are grown in the absence versus the presence of succinate. The level of these enzymes in P. aeruginosa PA14 carrying the crc-24 allele is no longer as strongly repressed (or not repressed at all) when cells are grown in the presence of succinate. This is the phenotype that originally defined the crc locus (37). The NADP-specific glyceraldehyde 3-phosphate dehydrogenase (as opposed to the NAD-specific enzyme) is not regulated by Crc, and the enzyme level shows a less-than-twofold change when cells are grown in the presence versus the absence of succinate.

Strains which express amidase are sensitive to the amide analog fluoroacetamide (FAA), which is converted to the toxic metabolite fluoroacetate by this enzyme (5). Expression of amidase is normally repressed when cells are grown on succinate, and this repression is dependent on Crc. Strains without functional Crc express amidase even in the presence of succinate and die when FAA is included in the medium. Thus, FAA sensitivity acts as an indirect assay of Crc function. Figure 4 shows the growth of the wild type, the crc-24 mutant, and the crc-24 mutant carrying pSMC31 (crc⁺) on minimal medium containing succinate with (A) or without (B) FAA. All of the strains grew equally well on minimal medium containing succinate without FAA (Fig. 4B). In contrast, the crc-24 mutant showed no growth on the FAA-containing medium (Fig. 4A). The wild type and the crc-24 mutant carrying the wild-type copy of crc on a plasmid grew well on this medium. Therefore, although the strain carrying the crc-24 mutation was isolated in a screen for mutants unable to form a biofilm, this strain is also defective in the Crc-mediated, succinate-dependent catabolite repression of enzymes required for carbohydrate catabolism.

Mutations in crc decrease the expression of pilA, the pilin structural gene.

Phibbs and coworkers reported that P. aeruginosa PAO1 carrying mutations in crc had increased levels of enzymes and transporters involved in carbohydrate metabolism even under conditions in which expression of these functions should have been repressed (i.e., in the presence of TCA intermediates such as succinate) (20, 37). However, the mechanism by which Crc could alter gene expression has not yet been determined.

We have obtained evidence that Crc can function to regulate gene expression at the level of transcription. Specifically, strains with mutations in crc have markedly lower levels of expression of pilA, which codes for the major pilin subunit, and pilB, which codes for an accessory factor required for pilus assembly. These genes are positioned adjacent to each other on the chromosome but are transcribed divergently (21, 23). We introduced plasmid-borne pilA-lacZ (pMSZ25) and pilB-lacZ (pPB18) transcriptional fusions (along with a vector control, pDN19) into wild-type P. aeruginosa PAO1 and its crc::Cb^r derivative. The results of β-galactosidase assays performed on these strains are summarized in Fig. 5. Expression of the pilA-lacZ transcriptional fusion activity was 132-fold higher in the wild type than in the crc mutant strain (crc⁺, 3,965 miller units [MU]; crc mutant, 30 MU). Similarly, the expression of the pilB-lacZ transcriptional fusion in the wild type was 641 MU, compared to 98 MU from this same fusion in the crc::Cb^r mutant (a decrease of 6.5-fold). As expected, the vector control showed only minimal activity in both genetic backgrounds.