Correspondence: Gunnel Svensäter, Department of Oral Biology, Faculty of Odontology, Malmö University, Malmö SE-20506, Sweden. Tel.: +46 40 665 8495; fax: +46 40 929 359; e-mail: email@example.com
Pseudomonas aeruginosa and Staphylococcus epidermidis are common opportunistic pathogens associated with medical device-related biofilm infections. 16S rRNA-FISH and confocal laser scanning microscopy were used to study these two bacteria in dual-species biofilms. Two of the four S. epidermidis strains used were shown to form biofilms more avidly on polymer surfaces than the other two strains. In dual-species biofilms, the presence of P. aeruginosa reduced biofilm formation by S. epidermidis, although different clinical isolates differed in their susceptibility to this effect. The most resistant isolate coexisted with P. aeruginosa for up to 18 h and was also resistant to the effects of the culture supernatant from P. aeruginosa biofilms, which caused dispersal from established biofilms of other S. epidermidis strains. Thus, different strains of S. epidermidis differed in their capacity to withstand the action of P. aeruginosa, with some being better equipped than others to coexist in biofilms with P. aeruginosa. Our data suggest that where S. epidermidis and P. aeruginosa are present on abiotic surfaces such as medical devices, S. epidermidis biofilm formation can be inhibited by P. aeruginosa through two mechanisms: disruption by extracellular products, possibly polysaccharides, and, in the later stages, by cell lysis.
Microbial biofilms can generally be defined as sessile communities of cells that are irreversibly attached either to a substratum or to each other, and that are embedded in a matrix of polymeric substances (Donlan & Costerton, 2002). Biofilms isolated from various environments differ in thickness (from single cells to multilayers) and species diversity as well as the composition and amount of extracellular polymeric substances present. For instance, densely packed dental plaque biofilms harbour a diverse consortium of microorganisms, with more than 500 known species (Marsh & Percival, 2006). A similar wide microbial diversity is also seen in the gastrointestinal tract (Hentschel et al., 2003), while medical device-related biofilms are likely to contain much fewer species (Tunney et al., 1998; Viale & Stefani, 2006; Stickler, 2008). Thus, because multispecies communities represent the normal situation, interactions between different bacterial species will be central to the development of naturally occurring biofilms.
On contact with surfaces, bacteria undergo a range of phenotypic changes, for instance, genes involved in alginate synthesis in Pseudomonas aeruginosa are upregulated within 1 h after surface attachment (Davies & Geesey, 1995). In 8-h biofilms of P. aeruginosa, the pattern of protein expression has been shown to be different from that of planktonic cells and at least half of the detectable proteome was differentially expressed in 6-day-old biofilms (Sauer et al., 2002). Bacterial responses to the numerous microniches that develop in thicker biofilms, with differences in pH, nutrient levels and oxygen tension, further contribute to the phenotypic and genotypic diversity (Ehrlich et al., 2005; Stewart & Franklin, 2008). An important medical consequence of the phenotypic changes associated with microbial biofilm growth is the high degree of resistance to stress (Mathee et al., 1999) and antimicrobial agents (Donlan & Costerton, 2002) as well as increased virulence (Wang et al., 2008), making biofilm-induced diseases difficult to treat.
Microbial biofilms are estimated to be involved in the majority of human infections, especially those of a chronic nature including urinary tract infections, cystic fibrosis, infective endocarditis and infections associated with medical devices, including central venous and other in-dwelling catheters (Fux et al., 2005). Many of these infections are the result of opportunistic colonization of new sites by the commensal flora. Staphylococcus epidermidis, normally found on the skin, as well as P. aeruginosa, can often be isolated from chronic wound- (Gjødsbøl et al., 2006) and catheter-associated infections (Finkelstein et al., 2002; Lyytikäinen et al., 2002). Staphylococcus epidermidis strains isolated from different types of infections and geographical regions show a high degree of genetic diversity (Miragaia et al., 2007), some of which is associated with differences in virulence and antibiotic resistance (Ziebuhr et al., 2006).
As emphasized by Marsh (2005), the different species present in a natural biofilm are not just passive neighbours, but rather are involved in a range of physical, metabolic and molecular interactions with each other. In the case of P. aeruginosa and S. epidermidis, several studies have shown that substances from P. aeruginosa, such as pyocyanin (Baron & Rowe, 1981), LasA (Barequet et al., 2009), quinolines (Machan et al., 1992) and rhamnolipids (Haba et al., 2003), have an antistaphylococcal activity. Recently, P. aeruginosa supernatant has been shown to be capable of disrupting established S. epidermidis biofilms through an extracellular polysaccharide-dependent process (Qin et al., 2009).
The aim of this work was to study S. epidermidis and P. aeruginosa during the formation of dual-species biofilms. To investigate this in relation to the clinical situation, we have used fresh isolates of S. epidermidis from skin and a peritoneal catheter and compared them with laboratory strains.
Materials and methods
Bacterial strains and culture
Pseudomonas aeruginosa [National Collection of Type Cultures (NCTC) 6750] and S. epidermidis strains [American Type Culture Collection (ATCC) 49461, Culture Collection, University of Göteborg (CCUG) 44858, fresh isolates Mia (isolated from skin of healthy person) and C103 (isolated from peritoneal dialysis catheter)] were grown on blood agar in 5% carbon dioxide (CO2) at 37 °C. Colonies were transferred to liquid Todd–Hewitt (TH) growth medium and incubated in 5% CO2 at 37 °C overnight. Samples were then transferred to fresh TH and incubated in 5% CO2 at 37 °C until the mid-exponential growth phase, corresponding to OD600 nm≈0.5, was reached.
Adherence and biofilm formation
Both ibiTreat μ-Slide eight-well cultures plates and the ibiTreat μ-Slide VI flow-cell system for Live Cell Analysis (Integrated BioDiagnostics, Munich, Germany) were used in this study. These have a physically treated cycloolefin-(co)polymer surface with a surface energy of around 72 mN m−1. Mid-exponential growth-phase cells of P. aeruginosa or S. epidermidis were harvested by centrifugation (4000 g, 15 min at 4 °C), washed in TH and adjusted to OD600 nm=0.5 corresponding to 1 × 109 cells mL−1 for P. aeruginosa and 4 × 108 cells mL−1 for S. epidermidis. For monoculture biofilm experiments of P. aeruginosa or S. epidermidis, cell suspensions containing 4 × 108 CFU mL−1 were used. For the dual-species biofilm experiments, a suspension containing 2 × 108 CFU mL−1 of P. aeruginosa and 2 × 108 CFU mL−1S. epidermidis was made. The flow cells were then inoculated with 100 μL of the cell suspension and incubated under static conditions for up to 24 h in a humid atmosphere of 5% CO2 at 37 °C before being subjected to BacLight LIVE/DEAD staining according to the manufacturer's instructions, a crystal violet biofilm assay or 16S rRNA-FISH, followed by confocal laser scanning microscopy (CLSM). The crystal violet biofilm assay, based on the method described by O'Toole et al. (1999), was performed on 4-h biofilm cells cultured in ibiTreat μ-Slide eight-well culture plates. Briefly, planktonic cells were removed by aspiration and biofilms were washed with water before incubation with 0.1% aqueous crystal violet for 10 min at room temperature. After removal of excess crystal violet by aspiration, wells were washed with water and air-dried. The retained crystal violet was extracted using 96% ethanol and absorbance was measured at 590 nm.
Effect of P. aeruginosa (NCTC 6750) on established biofilms of S. epidermidis
Staphylococcus epidermidis strains Mia and C103 were allowed to form biofilms for 9 h as described above. Pseudomonas aeruginosa NCTC 6750 was then added and the flow cells were maintained for 12, 18 or 24 h in 5% CO2 at 37 °C. The biofilms were studied using 16S rRNA-FISH, followed by CLSM.
Effect of culture supernatants from P. aeruginosa on established biofilms and biofilm formation of S. epidermidis
Flow cells were inoculated with a mid-exponential growth-phase suspension of P. aeruginosa NCTC 6750 (109 CFU mL−1) and maintained in TH medium, 5% CO2 at 37 °C, for 24 h. The supernatant was then harvested, filtered (0.20 μm) and stored at −20 °C until use. Staphylococcus epidermidis biofilms were then either exposed to the supernatant continuously during growth for 9 or 20 h or for 1 h following 9 or 20 h of growth. The biofilms were then subjected to 16S rRNA-FISH and examined using CLSM. Because the P. aeruginosa supernatant was prepared in TH medium, 9- and 20-h biofilms of S. epidermidis were also exposed to fresh TH (which had not been in contact with P. aeruginosa) for 1 h as a control. To control for the possible growth effect of introducing this new TH medium, biofilms of S. epidermidis were also subjected to 16S rRNA-FISH directly after the 9 or 20 h of growth.
For 16S rRNA-FISH, supernatants were removed from the flow cells and the biofilms were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) [0.1 M NaCl, 7 mM K2HPO4, 2.5 mM KH2PO4, pH 7.3] 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 of Tris-HCl, pH 7.5, and 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 ultrapure 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 sulphate (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. Pseudomonas aeruginosa was identified using the PsaerA probe (5′–3′ sequence GGTA ACCGTCCCCCTTGC) (Hogardt et al., 2000), while the STA3 probe (5′–3′ sequence GCACATCAGCGTCAGT) (Tavares et al., 2008) was used to identify S. epidermidis. The probes were fluorescently labelled with ATTO-565 or ATTO-488. 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 ultrapure water. For 16S rRNA-FISH on bacteria in suspension, a centrifugation step (6000 g, 5 min, 4 °C) was included after each 16S rRNA-FISH step to allow the recovery of the cells.
Analysis of cell viability
To investigate the viability of the cells, dual-species biofilms (after 9 and 20 h) were stained using the BacLight LIVE/DEAD staining kit according to the manufacturer's instructions. The spent culture medium from these biofilms, and from the established S. epidermidis biofilms exposed to P. aeruginosa for 12, 18 and 24 h, was also harvested and centrifuged (6000 g, 5 min, 4 °C) to recover the detached cells. These were then either stained using the BacLight LIVE/DEAD staining kit or subjected to 16S rRNA-FISH. Aliquots of the spent medium were also cultured on blood and 110 agar plates to investigate the proportions of P. aeruginosa and S. epidermidis, respectively.
CLSM, and image and statistical analyses
An Eclipse TE2000 inverted confocal laser scanning microscope (Nikon Corporation, Tokyo, Japan) was used to observe the flow cells and 20 randomly selected areas (0.05 mm2) of each sample, covering a total substratum area of 1 mm2, were photographed. Green fluorescence was provided by an Ar laser (488 nm laser excitation) and red fluorescence was given by a G-HeNe laser (543 nm laser excitation). Image analysis (surface coverage) was carried out using the function ‘Cell Counting-Batch’ in the software package bioimage_l (Chavez de Paz, 2009). The acquired data are reported as the mean and SD of at least three experiments. Differences between the mean surface coverage in the presence or absence of P. aeruginosa cells or culture supernatant were compared using a two-tailed t-test, and P values <0.05 were considered significant.
Biofilm formation by the test strains
Biofilm formation by individual S. epidermidis strains (ATCC 49461, CCUG 44858, Mia and C103) in mini flow cells was followed over 20 h using 16S rRNA-FISH staining (Fig. 1). For all strains, the maximum surface coverage was reached after 4 h, although the total surface coverage achieved differed between them. While ATCC 49461 and CCUG 44858 reached levels of 10.8 ± 9.7% and 16.9 ± 21.6%, Mia and C103 showed surface coverages of 55.2 ± 9.3% and 81.0 ± 3.3%, respectively. These results were confirmed using a crystal violet biofilm assay that showed strains Mia and C103 to be more avid biofilm formers than ATCC 49461 and CCUG 44858 (data not shown). Thus, it is clear that the strains of S. epidermidis used here vary in their capacity for biofilm formation. Biofilms were allowed to develop over the following 16 h, and although the levels declined somewhat, the pattern remained the same, with C103 and Mia showing a much higher surface coverage (54.4 ± 9.7% for C103, 52.3 ± 22.6% for Mia) than CCUG 44858 and ATCC 49461 (0.5 ± 0.1% for CCUG 44858, 2.4 ± 0.9% for ATCC 49461). The laboratory strain of P. aeruginosa (NCTC 6750) reached a maximum surface coverage of 14.6 ± 11.2%, and although this figure was comparable to that for the laboratory strains of S. epidermidis, it was reached after 1 h rather than the 4 h seen for the S. epidermidis strains (data not shown).
Competition between S. epidermidis and P. aeruginosa (NCTC 6750) during biofilm formation
When the strains of S. epidermidis, which were poor biofilm formers in their own right (CCUG 44858 and ATCC 49461), were investigated for their biofilm-forming capacity in the presence of P. aeruginosa, virtually no S. epidermidis cells could be detected after 9 h (data not shown). The strains that were more proficient biofilm formers (C103 and Mia) (see Fig. 1) were therefore investigated for their biofilm formation in the presence of P. aeruginosa.
After 9 h in a mono-species culture (control), the Mia strain showed a surface coverage of 54 ± 9% (Fig. 2a), while the C103 strain showed a coverage of 45 ± 13% (Fig. 2e). When the same numbers of S. epidermidis cells as used above were coinoculated with equal numbers of P. aeruginosa cells and followed for 9 h, the Mia strain showed a threefold decrease in surface coverage (18 ± 4%) over the control (Fig. 2b). The viability of the attached cells was the same (around 98%) in both the dual- (Fig. 2d) and the mono-species biofilms (Fig. 2c) suggesting that S. epidermidis Mia was removed, but not killed by P. aeruginosa. In the C103 strain, the effect of P. aeruginosa was much more pronounced, with a ninefold decrease in surface coverage (5 ± 5%) over 9 h in dual-species biofilms with P. aeruginosa (Fig. 2f) as compared with the single-species biofilm level (45 ± 13%) (Fig. 2e). The viability of the C103 strain was again similar (c. 98%) in both the mono- (Fig. 2g) and the dual-species (Fig. 2h) biofilms, again indicating that P. aeruginosa does not kill S. epidermidis in the biofilms. Thus, the strains Mia and C103 appear to differ in their susceptibility to the effect of P. aeruginosa on biofilm formation.
To further investigate the behaviour of S. epidermidis Mia and P. aeruginosa during biofilm formation, the two species were inoculated into flow cells in equal proportions and the total surface coverage as well as the relative proportions were followed over 24 h (Fig. 3). The proportion of S. epidermidis cells found in the biofilms after 6 h was less than that in the original inoculum, confirming the inhibitory effect of P. aeruginosa on S. epidermidis adhesion. The total surface coverage increased up to 18 h (54.7 ± 11.1%) and both species contributed to this increase. This suggests that those S. epidermidis cells that did attach were able to coexist with P. aeruginosa and grow for up to 18 h. After this time point, the relative numbers of S. epidermidis cells in the biofilms declined sharply, while the level of surface coverage by P. aeruginosa remained almost the same. The total viability of the cells released from the biofilms (found in the culture medium) as shown by BacLight LIVE/DEAD staining had decreased (67.8 ± 18.0% LIVE bacteria) and 16S rRNA-FISH studies showed the cells to be mostly P. aeruginosa. Culturing on blood and 110 agar showed that the total number of S. epidermidis cells in the system was reduced by a factor of 100, compared with 9 h, suggesting that S. epidermidis cells in the biofilm had undergone lysis, leading to further detachment between 18 and 24 h. The decrease in S. epidermidis surface coverage in the presence of P. aeruginosa was seen, although the effect was not so pronounced, even when 10 000 times fewer P. aeruginosa cells were present (data not shown).
Effect of P. aeruginosa on established S. epidermidis biofilms
Pseudomonas aeruginosa was added to 9 h biofilms of S. epidermidis Mia and C103, and the relative levels of both species in the biofilms were investigated after 12, 18 and 24 h using 16S rRNA-FISH (Fig. 4). After 12 h, the level of both S. epidermidis strains had decreased to 80% of the original level, indicating that P. aeruginosa cells had begun to establish themselves in the biofilms. Over the following 12 h, the levels of P. aeruginosa cells in the biofilms increased successively until S. epidermidis constituted <10% of the total biofilm surface coverage. These data suggest that P. aeruginosa cells are capable of exerting their effect of removing S. epidermidis cells even when these are present in established biofilms.
BacLight LIVE/DEAD staining as well as culturing on blood agar showed 90% of the cells in the biofilm culture fluid to be viable after 12 h. Culturing on 110 agar confirmed that these cells included S. epidermidis cells released from the biofilms. Thus, P. aeruginosa cells are capable of removing S. epidermidis cells from established biofilms without causing cell lysis.
Effect of P. aeruginosa biofilm supernatant on biofilms of S. epidermidis
To investigate how P. aeruginosa may be exerting its effect on S. epidermidis biofilms, staphylococcal cell suspensions were mixed with the culture supernatant of P. aeruginosa and biofilm formation was followed for 9 or 20 h. As can be seen in Table 1, the surface coverage of the S. epidermidis strains 49461 and C103 after 9 h decreased relative to the control during growth in the presence of the supernatant (P<0.001). To further study this effect, established 9- or 20-h biofilms of S. epidermidis were exposed to P. aeruginosa supernatant for 1 h and the cells were visualized using 16S rRNA-FISH as above. Control S. epidermidis biofilms were exposed to TH medium for 1 h. As shown in Fig. 5, after 9 h, there was no influence on the surface coverage of the Mia strain as compared with the control, suggesting that established biofilms of this strain could not be dispersed by the P. aeruginosa biofilm supernatant. In this respect, the Mia strain differed from the others tested (C103, ATCC 49461 and CCUG 44858), where clear reductions in surface coverage were seen in response to the supernatant from biofilm cells of P. aeruginosa. In the more established (20 h) biofilms (Fig. 5), a significant reduction in coverage with the Mia strain in response to the P. aeruginosa biofilm supernatant was seen (P<0.05).
Table 1. Biofilm formation of Staphylococcus epidermidis strains in the absence (control) and presence (+sup) of Pseudomonas aeruginosa supernatant after 9 and 20 h, presented as surface coverage (%)
Values significantly different from controls, P<0.001.
Biofilm cells were visualized using the 16S rRNA FISH probes PsaerA (P. aeruginosa) and STA3 (S. epidermidis) and surface coverage was calculated after CLSM studies. All values are means ± SD of triplicate experiments.
Staphylococcus epidermidis is an important opportunistic pathogen that ranks among the major causative agents of infections associated with in-dwelling catheters, implanted heart valves and orthopaedic devices (Rogers et al., 2009). Biofilm formation on polymer surfaces is thus an essential virulence factor for this bacterium. Staphylococcus epidermidis has been reported to interact specifically with host proteins such as fibrinogen (Hartford et al., 2001) and type I collagen (Arrecubieta et al., 2007) as well as attaching to biomaterials such as polystyrene via, for example, the fimbriae-like polymers SSP-1 and SSP-2 (Timmerman et al., 1991; Veenstra et al., 1996). However, the adhesion of S. epidermidis to abiotic surfaces has been recognized to be governed largely by surface hydrophobicity (Otto, 2009), which is mediated by, for instance, the bifunctional autolysin/adhesin AtlE (Heilmann et al., 1997). In this study, we have shown a clear difference in the adherence capacity of different strains of S. epidermidis, with those isolated from skin and the peritoneal dialysis catheter showing an avid adhesion and biofilm formation on abiotic polymer surfaces. It is thus likely that the lower degree of adhesion seen for the other strains of S. epidermidis in this work is due to differences in the expression of surface proteins. This emphasizes the dangers of drawing generalized conclusions regarding the adhesion of different species from studies of individual strains.
Several substances from P. aeruginosa including LasA (Barequet et al., 2009), the quorum-sensing molecule N-(3-oxodoceanoyl) homoserine lactone (Kaufmann et al., 2005), quinolines (Machan et al., 1992), pyocyanin (Baron & Rowe, 1981) and rhamnolipids (Haba et al., 2003) have been reported to have antistaphylococcal activity. From our study, it is evident that the presence of P. aeruginosa in dual-species inocula to a certain degree inhibited biofilm formation by S. epidermidis. For the S. epidermidis strain C103, the coverage was reduced by ninefold relative to the mono-species biofilms. A (3–4-fold) reduction in surface coverage was also seen when the supernatant from P. aeruginosa cultures was present during biofilm formation by the strains of S. epidermidis (Table 1).
In a recent study, it was also found that established S. epidermidis biofilms could be disrupted by extracellular products, mainly polysaccharides, from the PAO1 laboratory strain of P. aeruginosa (Qin et al., 2009). In static-chamber cultivation systems, 24-h, multilayered biofilms were reduced to single layers when exposed to the supernatant from planktonic cultures of P. aeruginosa for 2 h. These findings are consistent with our results for S. epidermidis strains ATTC 49461, CCUG 44858 and C103, where established biofilms were disrupted by the addition of P. aeruginosa cells or a 1-h exposure to the culture supernatant from P. aeruginosa. In the work by Qin et al. (2009), most of the detached S. epidermidis cells were shown to be viable and this was also the case in our study, suggesting that the same mechanisms were in operation here. Qin and colleagues present a possible explanation for the detachment by showing that while quorum-sensing-controlled factors were not involved, supernatants from P. aeruginosa pelA and pslF-negative mutants as well as the pelApslBCD double-negative mutant, which are deficient in polysaccharide synthesis, had a reduced ability to disrupt S. epidermidis biofilms.
Interestingly, the inhibition of S. epidermidis biofilm formation was much less pronounced for one of the strains (Mia), suggesting that this was, to some extent, resistant to the effects of P. aeruginosa and that different strains of staphylococci may differ in their capacity to resist dispersal by P. aeruginosa polysaccharide in vivo. Within dual-species biofilms of S. epidermidis (Mia) and P. aeruginosa, the two species were mostly found adjacent to each other rather than as separate populations, suggesting that they can coexist within the same biofilms for at least 18 h (Fig. 3). After 24 h, however, even the resistant S. epidermidis strain disappeared from the dual-species biofilms – most likely due to lysis by P. aeruginosa.
The ability of S. epidermidis and P. aeruginosa to coexist in the same environment is supported by clinical experience, where they can be isolated from the same types of infections (Finkelstein et al., 2002; Lyytikäinen et al., 2002). The mechanisms governing interactions between these two species in different clinical situations are currently unknown, but our data suggest that S. epidermidis biofilm formation can be inhibited by P. aeruginosa through two mechanisms: in the early stages through disruption by extracellular products, possibly polysaccharides, and, in the later stages, by cell lysis.
We thank Ulrika Troedsson, Madeleine Blomqvist and Agnethe Henriksson for their excellent technical assistance. This study was supported by the Knowledge Foundation and the Crafoord Foundation (J.R.D.), Sweden.