Effects of clinical isolates of Pseudomonas aeruginosa on Staphylococcus epidermidis biofilm formation

Authors


  • Editor: Roger Bayston

Correspondence: Julia R. Davies, Department of Oral Biology, Faculty of Odontology, Malmö University, Malmö SE-20506, Sweden. Tel.: +46 40 665 8492; fax: +46 40 92 9359; e-mail: Julia.Davies@mah.se

Abstract

Pseudomonas aeruginosa is often found in chronic infections, including cystic fibrosis lung infections and those related to chronic wounds and venous ulcers. At the latter sites, P. aeruginosa can be isolated together with Staphylococcus epidermidis, and we have therefore explored the effect of clinical isolates and laboratory strains of P. aeruginosa strains on colonization by S. epidermidis in dual-species biofilms. Biofilm formation was assayed using 16S rRNA FISH and confocal laser scanning microscopy. Among the six P. aeruginosa strains tested, one particular strain, denoted 14:2, exerted a significant inhibitory effect, and even after 6 h, S. epidermidis levels in dual-species biofilms were reduced by >85% compared with those without P. aeruginosa. Interestingly, strain 14:2 was found to be negative for classical virulence determinants including pyocyanin, elastase and alkaline protease. Therefore, we suggest that less virulent phenotypes of P. aeruginosa, which may develop over time in chronic infections, could counteract colonization by S. epidermidis, ensuring persistence and dominance by P. aeruginosa in the host micro-habitat. Further studies are required to explain the inhibitory effect on S. epidermidis, although extracellular polysaccharides produced by P. aeruginosa might play a role in this phenomenon.

Introduction

Pseudomonas aeruginosa can be identified in a range of infections, particularly those with a tendency to become chronic, such as lung infections in patients with cystic fibrosis (Wagner & Iglewski, 2008), those related to venous ulcers (Dowd et al., 2008) and infections associated with in-dwelling medical devices (Finkelstein et al., 2002). The most well-documented virulence property of P. aeruginosa is its ability to produce and secrete elastase (Woods et al., 1982), alkaline protease (Howe & Iglewski, 1984), pyocyanin (Lau et al., 2004), rhamnolipids and a range of exotoxins (Smith & Iglewski, 2003). The expression of many of these factors is known to be differentially regulated through quorum-sensing systems in response to prevailing environmental conditions (Williams et al., 2000). Thus, progressive selection pressure during chronic infection may affect the expression of virulence factors and, indeed, less virulent phenotypes of P. aeruginosa do appear in cystic fibrosis patients with chronic lung infections (Luzar & Montie, 1985). In addition to the secretion of extracellular enzymes and toxins, persistence in the host has been linked to the ability of P. aeruginosa to adhere to and form biofilms on tissues and abiotic surfaces. Within these biofilms, communities of bacteria are embedded in a matrix of extracellular polymeric substances consisting of proteins, polysaccharides and nucleic acids largely derived from the bacteria themselves. In mucoid strains of P. aeruginosa, this matrix appears to be dominated by alginate. In nonmucoid strains, however, the matrix is considered to be composed of two recently described polysaccharides encoded by the psl and pel genes. These are Psl, a polymer rich in mannose and galactose residues, and Pel, a glucose-rich polymer (Ryder et al., 2007).

Natural biofilms are rarely mono-species communities, but are composed of several bacterial species. In chronic wounds and chronic venous ulcers as well as on in-dwelling catheters, P. aeruginosa can be found together with Staphylococcus epidermidis, a normal colonizer of the skin (Holley et al., 1992; Marra et al., 2005; Gjødsbøl et al., 2006). Our previous work has shown that one type strain of P. aeruginosa (NCTC 6750) present in a biofilm can exert an inhibitory effect on colonization by freshly isolated strains of S. epidermidis (Pihl et al., 2010). In another study by Qin et al. (2009), a similar effect was seen for the P. aeruginosa strain PAO1 and these authors have proposed that the effect is mediated by polysaccharide production via a quorum-sensing-independent mechanism. These observations prompted us to explore whether the inhibitory effect on S. epidermidis biofilm formation is unique to the type strains NCTC 6750 and PAO1 or is also present among clinical isolates of P. aeruginosa. In the present study, we confirm that the phenomenon is common to several freshly isolated P. aeruginosa strains and may thus be of importance in the progression of chronic infections where these two species are present. One of the P. aeruginosa strains had a greater capacity to prevent S. epidermidis colonization than the type strains studied previously and, interestingly, while this strain produced extracellular polysaccharide, it lacked the production of virulence factors such as elastase, pyocyanin and alkaline protease.

Materials and methods

Pseudomonas aeruginosa strains

Nonmucoid clinical isolates of P. aeruginosa (14:2, 23:1, 27:1 and 15159) were derived from patients with chronic venous ulcers (Schmidtchen et al., 2001, 2003). Patients had not been treated with antibiotics before isolation of the strains. In addition, two nonmucoid laboratory strains of P. aeruginosa, NCTC 6750 and PAO1 (ATCC BAA-47), were obtained from the National Collection of Type Cultures (NCTC) and American Type Culture Collection (ATCC).

Staphylococcus epidermidis strains

The staphylococcal strain Mia was isolated from the skin of a healthy person, while the others (C103, C116, C121, C164 and C191) were isolated from the external and luminal sides of the subcutaneous or the intraperitoneal part of dialysis catheters from five peritoneal dialysis patients. These patients were undergoing renal transplantation and showed no clinical signs of infection. The isolates were identified as Gram-positive cocci and showed growth as white colonies on staphylococcus-specific 110 agar (Chapman, 1949). All the strains were also catalase positive and oxidase negative (Barrow & Feltham, 1993), showing that they are staphylococci. However, they were also found to be negative in the Pastorex Staph Plus agglutination test (Bio-Rad) (Weist et al., 2006), indicating that they do not correspond to Staphylococcus aureus. Further identification was carried out using 16S rRNA gene sequencing. Strains were stored at −80 °C and not subcultured more than twice.

Preparation of mono- and dual-species biofilms

Bacteria were grown in Todd–Hewitt (TH) medium and incubated in 5% CO2 at 37 °C until the mid-exponential growth phase, corresponding to OD600 nm≈0.5, was reached. Cells were harvested by centrifugation (4000 g, 15 min at 4 °C), washed in TH and adjusted to OD600 nm=0.5 (corresponding to 109–1012 CFU mL−1 for P. aeruginosa and 108 CFU mL−1 for S. epidermidis).

Monoculture biofilms of the staphylococcal strains or P. aeruginosa were established in ibidi flow cells (μ-Slide VI for Live Cell Analysis, Integrated BioDiagnostics) by inoculating channels with a mid-exponential growth-phase cell suspension containing 2 × 108 CFU mL−1. The slides were maintained under static conditions for 6 h in 5% CO2 at 37 °C, and the biofilms were then subjected to 16S rRNA FISH and confocal laser scanning microscopy (CLSM). Each experiment was carried out in duplicate and two independent experiments were performed.

The staphylococcal strains identified as good biofilm formers in the monoculture studies (Mia, C103 and C121) were used in the dual-species experiments. They were mixed in equal proportions with the different P. aeruginosa strains, corresponding to 2 × 108 CFU mL−1 of each species. Biofilm formation was followed for 6 h under static conditions in 5% CO2 at 37 °C, and the biofilms were studied using 16S rRNA FISH and CLSM. Each experiment was carried out in duplicate and two independent experiments were performed.

16S rRNA FISH

Pseudomonas aeruginosa was identified using the PsaerA probe (5′–3′sequence GGTAACCGTCCCCCTTGC) (Hogardt et al., 2000) fluorescently labelled with ATTO-488 (green). Staphylococcus epidermidis was identified using the STA3 probe (5′–3′sequence GCACATCAGCGTCAGT) (Tavares et al., 2008) fluorescently labelled with ATTO-565 (red). For 16S rRNA FISH, supernatants were removed from the flow cells and the biofilms were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) overnight at 4 °C before being washed with cold sterile PBS. Bacterial biofilm cells were permeabilized using lysozyme (70 U mL−1) in 100 mM Tris-HCl, pH 7.5, 5 mM EDTA for 9 min at 37 °C and lysostaphin (0.1 mg mL−1) in 10 mM Tris-HCl, pH 7.5, for 5 min at 37 °C. The biofilms were then washed with ultra-pure water and dehydrated with 50%, 80% and 99% ethanol for 3 min, respectively, after which the flow cells were inoculated with 30 μL of hybridization buffer [0.9 M NaCl, 20 mM Tris-HCl buffer, pH 7.5, with 0.01% sodium dodecyl sulfate (SDS) and 25% formamide] containing 20 ng μL−1 of oligonucleotide probe PsaerA or 18 ng μL−1 of probe STA3 and incubated at 47 °C for 90 min in a humid chamber. In dual-species biofilms, a probe cocktail containing 20 ng μL−1 of oligonucleotide probe PsaerA and 18 ng μL−1 of probe STA3 in hybridization buffer was used. After hybridization, the slides were incubated with washing buffer (20 mM Tris-HCl buffer, pH 7.5, containing 5 mM EDTA, 0.01% SDS and 159 mM NaCl) for 15 min at 47 °C, and then rinsed with ultra-pure water.

CLSM and image analysis

An Eclipse TE2000 inverted confocal laser scanning microscope (Nikon Corporation, Tokyo, Japan) was used to observe the flow cells and 20 randomly selected areas of each sample, covering a total substratum area of 0.9 mm2, were photographed. Green fluorescence was provided by an Ar laser (488 nm laser excitation) and red fluorescence was provided by a G–HeNe laser (543 nm laser excitation). Image analysis (substratum coverage) was carried out using the function ‘Cell Counting-Batch’ in the software package bioimage_l (Chávez de Paz, 2009).

Effect of culture supernatants of different strains of P. aeruginosa on established biofilms of S. epidermidis

For the preparation of biofilm supernatants, mid-exponential growth-phase cultures (corresponding to 109 CFU mL−1) of the P. aeruginosa strains (NCTC 6750, PAO1, 14:2, 23:1, 27:1 and 15159) in TH medium were inoculated into tissue culture flasks and allowed to grow in biofilms under static conditions for 24 h (5% CO2, 37 °C). Culture supernatants were collected and subjected to centrifugation (10 min, 3000 g), sterile filtered (0.20 μm) and stored at −20 °C until use.

Six-hour S. epidermidis biofilms were exposed to P. aeruginosa biofilm supernatants for 1 h and then visualized using 16S rRNA FISH with the STA3 probe and examined using CSLM. The viability of the attached cells was investigated in parallel biofilm cultures using the BacLight LIVE/DEAD stain according to the manufacturer's instructions. To investigate the viability of dispersed cells of S. epidermidis, aliquots of the spent medium were cultured on 110 agar or stained using BacLight LIVE/DEAD staining. Two independent experiments were performed.

Homoserine lactones

The production of N-butanoyl-l-homoserine lactone (C4-HSL) was studied with a well-diffusion assay using the reporter strain Chromobacterium violaceum CV026 as described by Ravn et al. (2001). Culture supernatants from 24-h biofilms were extracted twice with equal volumes of ethyl acetate acidified with 0.5% formic acid. The combined extracts were then vacuum-dried and the residues were dissolved in 0.5 mL of ethyl acetate acidified with 0.5% formic acid and stored at −20 °C until use. Luria–Bertani (LB) agar seeded with C. violaceum CV026 (cultured overnight in LB broth supplemented with 20 μg mL−1 kanamycin, 28 °C) was poured onto prewarmed LB agar and allowed to solidify (10 μL C. violaceum culture mL−1 LB). Wells punched into the agar were filled with 50 μL of the solvent extracts and incubated for 24 h at 28 °C. Synthetic C4-HSL (Sigma) (1 mM) and TH medium were used as positive and negative controls, respectively. The presence of purple pigmentation around the wells indicated violacein production by C. violaceum CV026 in response to C4- to C8-HSL (McClean et al., 1997).

Pyocyanin production

Pyocyanin production was investigated by inoculating Pseudomonas medium A agar (Atlas & Parks, 1993) with the P. aeruginosa strains and incubating for 24 and 48 h in 5% CO2 at 37 °C. The production of the phenazine pigment pyocyanin was indicated by the presence of green colour around the CFU.

Zymography

Protease expression in biofilms of the different strains was determined by electrophoresis on Novex Zymogram gels (Invitrogen) according to the manufacturer's instructions. Briefly, equal volumes of biofilm culture supernatant were mixed with Novex Tris-Glycine SDS sample buffer and subjected to electrophoresis on 10% zymogram gelatin gels (nonreduced) for 2 h at 125 V. To remove SDS, gels were washed with renaturing buffer for 30 min at room temperature and incubation was then performed overnight at 37 °C on a shaking platform in developing buffer. Gels were stained with Coomassie blue G-250 in 20% ethanol for 3 h and destained in 25% ethanol. Protease-containing fractions were visualized as clear bands against a dark background. The total repertoire of extracellular proteins was also investigated by mixing biofilm culture supernatants with NuPAGE sample buffer (Invitrogen) and subjecting them to electrophoresis on 10% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) gels under reducing conditions for 1 h at 180 V. Gels were then stained with Coomassie blue according to the manufacturer's instructions. For the detection of P. aeruginosa elastase, proteins from the gels were electroblotted onto PVDF membranes (Immobilon-P, Millipore) at 50 V for 2 h at 4 °C. After blocking with 5% skim milk in Tris-buffered saline with 0.05% Tween-20, membranes were incubated first with a rabbit anti α-elastase antibody [a generous gift from Dr J. Fukushima; see also Schmidtchen et al. (2003)] diluted 1 : 750 and then an HRP-conjugated goat anti-rabbit Ig antibody diluted 1 : 2500. Antibody binding was visualized using the ECL Western blotting reagent (Pierce).

Extracellular polysaccharides

The production of extracellular polysaccharides by P. aeruginosa strains was studied using the lectins Hippeastrum hybrid agglutinin (HHA) and Marasmium oreades agglutinin (MOA) (recognizing galactose and mannose residues, respectively) (Ma et al., 2007). Twenty four-hour biofilms prepared as described below were washed twice in 100 μL PBS and then incubated with MOA or HHA [0.1 mg mL−1 in PBS (7 mM K2HPO4, 2.5 mM KH2PO4, pH 7.3, containing 0.1 M NaCl)] for 2 h at room temperature. Biofilms were washed four times (100 μL) with PBS before examination using CLSM.

Statistical analysis

Statistical analysis was performed using a one-way anova with a Bonferroni post-test to compare different strains.

Results

Biofilm formation by P. aeruginosa or S. epidermidis strains

Investigation of the different P. aeruginosa strains showed that they varied in their ability to form biofilms over 6 h in the flow cells. The clinical isolates (14:2, 23:1, 27:1 and 15159) and PAO1 showed a low degree of biofilm formation (1.5–5% surface coverage), while the type strain NCTC 6750 was a relatively good biofilm former (22% surface coverage) (Fig. 1a).

Figure 1.

 Biofilm formation by strains used in this study. (a) Six-hour biofilms of different strains of Pseudomonas aeruginosa (PAO1, NCTC 6750, 14:2, 23:1, 27:1 and 15159) in ibidi μ-Slide VI flow cells visualized using 16S rRNA FISH with the PsaerA probe. (b) Six-hour biofilms of strains of Staphylococcus epidermidis (Mia, C103, C121, C164), Staphylococcus lugdunensis (C116) and Staphylococcus warneri (C191) detected using 16S rRNA FISH and the STA3 probe.

Because we were interested in studying the effect of different P. aeruginosa strains on biofilm formation by S. epidermidis, the ability of a number of different, freshly isolated, S. epidermidis strains to form mono-species biofilms was also investigated. After 6 h of growth in flow cells, the clinical isolates of S. epidermidis showed substantial differences in biofilm-forming ability, with the surface coverage ranging from 0.4–0.2 mm2 for strains Mia, C103, C121 and C164, to 0.009 mm2 for strains C116 and C191 (Fig. 1b). Thus, the clinical isolates could be divided into two clear groups: good and poor biofilm formers. Subsequent 16S rRNA gene analysis later revealed the good biofilm formers to be strains of S. epidermidis, while the poor biofilm formers (C116 and C191) were identified as Staphylococcus lugdunensis and Staphylococcus warneri, respectively.

Effect of P. aeruginosa on biofilm formation by S. epidermidis

To study the effects of P. aeruginosa on the ability of S. epidermidis to form biofilms, equal numbers of S. epidermidis (strains C103 or C121) and P. aeruginosa cells (strains 14:2 or 15159) were inoculated into the flow cells and maintained for 6 h. Image analysis showed the level of surface coverage by the P. aeruginosa strains in the dual-species biofilms to be in the same range as that seen for the mono-species ones (Fig. 2g and h). The presence of P. aeruginosa strain 14:2 in the biofilms caused large reductions in colonization by S. epidermidis strains: 88% for strain C103 (Fig. 2b) and 86% for strain C121 (Fig. 2e) compared with their respective controls (Fig. 2a and d). However, the presence of the P. aeruginosa strain 15159 reduced biofilm-formation by the S. epidermidis strains C103 (Fig. 2c) and C121 (Fig. 2f) by only 34% and 38%, respectively, over the control (the equivalent mono-species levels) (Fig. 2a and d). Thus, although both the P. aeruginosa strains cause some degree of inhibition of biofilm formation by S. epidermidis, the effect is much greater for strain 14:2 than 15159.

Figure 2.

 Dual-species biofilms of Pseudomonas aeruginosa and Staphylococcus epidermidis after 6 h in ibidi μ-Slide VI flow cells. Control, mono-species biofilms of S. epidermidis strains C103 and C121 are shown in (a and d), respectively, while mono-species biofilms of P. aeruginosa strains 14:2 and 15159 are shown in (g and h), respectively. The effect of P. aeruginosa strain 14:2 (b, e) as well as the effect of P. aeruginosa strain 15159 (c, f) on biofilms of S. epidermidis strains C103 (b, c) and C121 (e, f) are also shown. Staphylococcus epidermidis was detected using the STA3 probe labelled with ATTO-565 (red) and P. aeruginosa with the PsaerA probe labelled with ATTO-488 (green), and both were visualized using CLSM.

The effects of all the different strains of P. aeruginosa (PAO1, NCTC 6750, 14:2, 23:1, 27:1 or 15159) on the ability of S. epidermidis (Mia, C103 or C121) to form biofilms were also studied as above. For the Mia strain, even after 6 h of co-culture in biofilms, the presence of all the P. aeruginosa strains reduced colonization compared with the control and the effect was significant (P<0.05) for strains PAO1 and 23:1 (Fig. 3). For S. epidermidis strains C103 and C121, a significant reduction in colonization (P<0.05) was seen when strain 14:2 was present in the dual-species biofilms. The S. epidermidis strain C121 appeared to be generally more resistant to the effect of P. aeruginosa than the other two (Fig. 3) and an increase in surface coverage was seen in the presence of NCTC 6750.

Figure 3.

 The effect of Pseudomonas aeruginosa on colonization by Staphylococcus epidermidis. Surface area covered with S. epidermidis (strains Mia, C103, C121) after 6 h in dual-species biofilms ibidi μ-Slide VI flow cells with P. aeruginosa strains (NCTC 6750, PAO1, 14:2, 23:1, 27:1, 15159), visualized using 16S rRNA FISH and CLSM. Control values represent the surface coverage of the S. epidermidis strains in 6-h biofilms in the absence of P. aeruginosa. *Values significantly different from control (P<0.05).

In summary, of the P. aeruginosa strains studied here, 14:2 had the greatest effect in inhibiting biofilm formation by S. epidermidis, giving rise to a 50% reduction for strain Mia and a >85% reduction for strains C103 and C121. Staphylococcus epidermidis strain C121 differed somewhat from the other two in that it was more resistant to P. aeruginosa.

Dispersal of established S. epidermidis biofilms by the supernatant from biofilms of P. aeruginosa

Established 6-h biofilms of the three S. epidermidis strains (Mia, C103 or C121) corresponding to a total area of 0.8 mm2 were exposed to biofilm supernatants from P. aeruginosa strains (PAO1, NCTC 6750, 14:2, 23:1, 27:1 or 15159) or TH medium (control) for 1 h. Cells remaining in the biofilms were then visualized using 16S rRNA FISH. The results for S. epidermidis strain C121 are shown in Fig. 4. Supernatants of all the P. aeruginosa strains had an effect on the S. epidermidis biofilms and the reduction in coverage was significant (P<0.001) for strains PAO1, 6750, 14:2, 23:1 and 27:1, but not for 15159. As for the dual-species biofilms shown in Fig. 3, a pronounced effect was seen for strain 14:2. Similar effects were seen with the P. aeruginosa supernatants for the other S. epidermidis strains (Mia and C103), although the effects were less pronounced (data not shown). To determine whether the dispersal effect on S. epidermidis biofilms was due to cell lysis, S. epidermidis cells remaining in the biofilms after exposure to the P. aeruginosa biofilm supernatants were examined with the BacLight LIVE/DEAD stain. For all the S. epidermidis strains (Mia, C103 and C121), over 90% of the cells were viable after treatment with each of the P. aeruginosa supernatants (data not shown). Similarly, the level of viability of the dispersed cells was over 90% as shown by staining or growth on 110 agar.

Figure 4.

 The effect of supernatants from Pseudomonas aeruginosa biofilms on established biofilms of Staphylococcus epidermidis strain C121. After 6 h of growth in ibidi μ-Slide VI flow-cells, biofilms of S. epidermidis strain C121 were inoculated with the supernatant from biofilms of the P. aeruginosa strains (NCTC 6750, PAO1, 14:2, 23:1, 27:1, 15159) for 1 h before being subjected to 16S rRNA FISH. Controls were incubated for 1 h with fresh TH medium before being subjected to 16S rRNA FISH. *Values significantly different from the control (P<0.001).

Characteristics of the P. aeruginosa strains

In order to investigate what might be responsible for the variable effect of the P. aeruginosa strains (PAO1, NCTC 6750, 14:2, 23:1, 27:1 and 15159), biofilm supernatants were investigated for the release of a number of known virulence factors. The type strain PAO1 and the clinical isolate 15159 were found to be positive for the production of the quorum-sensing signal C4-HSL, while all the other strains were negative (Table 1). All the P. aeruginosa strains were positive for pyocyanin production, except 14:2 and 27:1, which were negative in this assay (Table 1). These results indicate that the repertoire of extracellular products released from the cells varies according to the strain. The secretion of extracellular proteases from P. aeruginosa cells growing in biofilms was investigated with zymography of culture supernatants (Fig. 5a). This showed differences between the strains in their degree of gelatinase activity. The supernatants from the two laboratory strains: PAO1 and NCTC 6750 as well as the clinical isolate 15159 contained at least three major bands of proteolytic activity at >150, 70 and 50 kDa. The >150 kDa enzyme has been identified previously by immuno-blotting and N-terminal sequencing as a multimeric form of P. aeruginosa elastase (Schmidtchen et al., 2003). In the same study, P. aeruginosa alkaline protease was demonstrated to band at around 50 kDa. This 50 kDa band, but not the higher molecular weight fractions, was also present in supernatants from strains 23:1 and 27:1 while the culture supernatant from biofilms of strain 14:2 appeared to lack any proteolytic activity. SDS-PAGE of the same material under reducing conditions confirmed differences in the extracellular protein profiles between the strains (Fig. 5b). Two different protein banding patterns could be identified, with strains PAO1, NCTC 6750 and 15159 showing a similar pattern and 14:2, 23:1 and 27:1 strains sharing many common bands. A strongly stained band, present at around 36 kDa in supernatants from the latter group, was recognized by antibodies against P. aeruginosa elastase (Fig. 5c), and thus corresponds to monomers of the enzyme. In the zymogram gels, this material is present as multimers at Mw>150 kDa (see Fig. 5a). Thus, it appears that the six P. aeruginosa strains fall into three different phenotypic categories: PAO1, NCTC 6750 and 15159, which produce elastase and alkaline protease, 23:1 and 27:1, which appear to produce only alkaline protease, and strain 14:2, which lacks extracellular protease activity.

Table 1.   Production of the homoserine lactone (C4-HSL) by 24-h-old biofilms and pyocyanin after a 24-h growth on agar by various strains of Pseudomonas aeruginosa
Extracellular productsP. aeruginosa strains
PAO1NCTC 675014:223:127:115159
  1. Detailed conditions are given in Materials and methods.

C4-HSL++
Pyocyanin++++
Figure 5.

 Production of proteases by Pseudomonas aeruginosa strains during biofilm growth. Culture supernatants from biofilms of P. aeruginosa strains (PAO1, NCTC 6750, 14:2, 23:1, 27:1 and 15159) were subjected to (a) zymography on gelatine gels or (b) SDS-PAGE on 10% gels and stained with Coomassie blue G-250 as described in the Materials and methods. (c) Western blot of a 10% SDS-PAGE gel stained with a rabbit anti α-elastase antibody as described in the Materials and methods.

The production of mannose- and galactose-rich exopolymeric substances by P. aeruginosa cells during biofilm growth was studied using lectin staining with HHA and MOA (Fig. 6). The patterns of staining with the two lectins were very similar, and some mannose- and galactose-containing polysaccharides were seen for all strains. PAO1 showed the greatest level while strain 27:1 produced very low amounts. For the remaining strains, the amount of polysaccharides produced lay between these values.

Figure 6.

 Polysaccharide production by Pseudomonas aeruginosa growing in biofilms. Lectin staining of P. aeruginosa biofilms by strains PAO1, NCTC 6750, 14:2, 23:1, 27:1 and 15159. The strains were grown for 24 h in ibidi μ-Slide VI flow cells and stained with the mannose-specific [0.1 mg mL−1 HHA-fluorescein isothiocyanate (HHA-FITC)] or galactose-specific (0.1 mg mL−1 MOA-FITC) lectins for 2 h and visualized using CLSM.

Discussion

Biofilms are now recognized as the dominant mode of bacterial growth in vivo and the ability to form them can thus be regarded as a prerequisite for colonization (Costerton et al., 1999). While all the P. aeruginosa strains used here formed biofilms, the type strain NCTC 6750 was the most avid biofilm former (see Fig. 1a). However, even this strain has a low biofilm-forming capacity compared with the S. epidermidis isolates. When the two bacterial species (P. aeruginosa and S. epidermidis) were cultured in dual-species biofilms, the capacity of P. aeruginosa to form biofilms was unaffected by the presence of S. epidermidis (Fig. 2). On the contrary, colonization by S. epidermidis was generally reduced in the presence of the Pseudomonas strains (Figs 2 and 3) and the supernatant from P. aeruginosa biofilms had the capacity to disperse cells from preformed S. epidermidis biofilms (Fig. 4). This effect could not be attributed to lysis of S. epidermidis as both cells remaining in the biofilms and those that were detached in the presence of the supernatant were still viable. The S. epidermidis strains varied somewhat in their susceptibility to this effect and the reasons for this are unclear. However, a range of factors are known to be involved in biofilm formation by S. epidermidis, including surface adhesins and extracellular polysaccharides (Agarwal et al., 2010), and it is possible that the differential expression of surface components among strains may be causing the differences, where more resistant ones express lower levels of the target for the P. aeruginosa products.

Despite some variability in the capacity of P. aeruginosa strains to exert their effects, both cells and biofilm supernatants of strain 14:2 consistently exerted an inhibitory effect on all the S. epidermidis strains tested. Thus, it was of interest to compare the products released from strain 14:2 with those from the other P. aeruginosa strains. A high degree of phenotypic variation was found among type strains of P. aeruginosa and those isolated from chronic skin wounds with respect to the production of virulence determinants such as pyocyanin and extracellular protease. The six strains fell into three categories: the first included the two type strains as well as one of the clinical isolates (PAO1, NCTC 6750 and 15159), the second contained the clinical isolates 23:1 and 27:1 and, finally, strain 14:2 (also a clinical isolate) formed a group on its own. In the first group, all strains expressed pyocyanin, elastase and alkaline proteinase, and two of the three produced the quorum-sensing molecule C4-HSL, while the second group showed no expression of C4-HSL or elastase. Interestingly, strain 14:2 was negative for the expression of C4-HSL, pyocyanin and the proteases. A similar spread in the expression of virulence factors and quorum-sensing molecules among P. aeruginosa strains has been described by others, for instance, Luzar & Montie (1985) and Lee et al. (2005), who investigated chronically infected cystic fibrosis patients. Both studies showed not only variations between strains isolated from different patients but also changes associated with disease progression. Isolates from patients with more advanced disease showed lower pyocyanin and protease production, suggesting that the evolution of P. aeruginosa strains towards a less virulent phenotype may confer a survival advantage during chronic infection. Thus, in our study, the clinical isolate 14:2, which had the greatest inhibitory effect on biofilm formation by S. epidermidis and lacked the production of C4-HSL, pyocyanin and proteases, may represent a less virulent strain that has become adapted to enhance its persistence in the chronic sore environment (Lee et al., 2005).

In a recent study by Qin et al. (2009), extracellular products from P. aeruginosa were shown to disrupt S. epidermidis biofilms and it was suggested that extracellular polysaccharide could be responsible for the effect. Thus, the authors proposed that extracellular polysaccharides from P. aeruginosa may represent a novel target for the development of agents to control S. epidermidis biofilms at sites of infection. Mannose- and galactose-containing extracellular polysaccharides were detected in biofilms of all the strains of P. aeruginosa tested here, and thus the inhibition of S. epidermidis biofilm formation seen in our study may occur through a mechanism similar to that proposed by Qin and colleagues for biofilm dispersal. Expression of the two extracellular polysaccharides, Pel and Psl, is known to vary according to the strain and environmental conditions (Branda et al., 2005). Although 14:2 did not appear to produce higher levels of these polysaccharides than the other strains, which could account for its enhanced effect on S. epidermidis biofilms, it is possible that, for instance, differences in their relative expression may play an important role. In addition, coating with surface-active rhamnolipids from P. aeruginosa has been shown to inhibit adhesion and cause detachment of S. epidermidis from surfaces (Rodrigues et al., 2006), suggesting that such molecules may also represent candidates for mediating the effects seen in this study. Further studies are required to determine whether or not this is the case.

In conclusion, we have shown that strains of P. aeruginosa vary in their ability to affect biofilm formation by S. epidermidis and that the strain with the greatest effect appeared to lack the production of the classical virulence factors. In infections where both species are present, the outcome over time is likely to be highly influenced by the phenotype of the strains involved.

Acknowledgements

We thank Agnethe Henriksson, Ulrika Troedsson and Madeleine Blomqvist for excellent technical support. We wish to express our gratitude to Professor David Beighton, KCL Dental Institute, London, UK, for sequencing of staphylococcal strains. The reporter strain C. violaceum CV026 was a kind gift from Professor Peter Greenberg, University of Washington, USA. This study was financially supported by the Knowledge Foundation and the Crafoord Foundation, Sweden.

Ancillary