biofilm formation by staphylococcus epidermidis on peritoneal dialysis catheters and the effects of...

RESEARCH ARTICLE

Biofilm formation by Staphylococcus epidermidis on peritonealdialysis catheters and the effects of extracellular products fromPseudomonas aeruginosaMaria Pihl1, Anna Arvidsson2, Marie Skep€o3, Martin Nilsson4, Michael Givskov4,5, Tim Tolker-Nielsen4, GunnelSvens€ater1 & Julia R. Davies1

1 Department of Oral Biology, Faculty of Odontology, Malm€o University, Malm€o, Sweden

2 Department of Biomaterials, Sahlgrenska Academy, Gothenburg University, Gothenburg, Sweden

3 Department of Theoretical Chemistry, Lund University, Lund, Sweden

4 Department of International Health, Immunology and Microbiology, Faculty of Health Sciences, University of Copenhagen, Copenhagen, Denmark

5 Singapore Centre on Environmental Life Sciences Engineering, Nanyang Technological University, Nanyang, Singapore

This paper increases our understanding of the mechanisms by which bacteria colonise non-biological surfaces in thepresence of host proteins and in addition shows that there are bacterial molecules with the potential to be used as anti-adhesive or anti-bacterial molecules in biofilm-related infections.

Keywords

peritoneal dialysis catheter; flow-cell model;

bacterial adherence; biofilm; rhamnolipid.

Correspondence

Julia R. Davies, Department of Oral Biology,

Faculty of Odontology, Malm€o University,

Malm€o SE-20506, Sweden.

Tel.: +46 40 665 8492

fax: +46 40 929359

e-mail: [email protected]

Received 10 October 2012; revised 25

February 2013; accepted 25 February 2013.

Final version published online 3 April 2013.

doi:10.1111/2049-632X.12035

Editor: Ake Forsberg

Abstract

Biofilm formation by Staphylococcus epidermidis is a cause of infections related toperitoneal dialysis (PD). We have used a PD catheter flow-cell model incombination with confocal scanning laser microscopy and atomic force microscopyto study biofilm formation by S. epidermidis. Adherence to serum-coated catheterswas four times greater than to uncoated ones, suggesting that S. epidermidisbinds to serum proteins on the catheter surface. Pseudomonas aeruginosa biofilmsupernatant interfered with the formation of a serum protein coat thereby reducingthe capacity for biofilm formation in S. epidermidis. Supernatants from DpelA,DpslBCD and DrhlAB strains of P. aeruginosa showed no differences from thewild-type supernatant indicating that the effect on serum coat formation was notdue to rhamnolipids or the PelA and PslBCD polysaccharides. Supernatant fromP. aeruginosa also dispersed established S. epidermidis biofilms. Supernatantslacking PelA or PslBCD showed no differences from the wild type but that from aDrhlAB strain, showed reduced, but not abolished, capacity for dispersal. Thissuggests that rhamnolipids are involved but not wholly responsible for the effect.Thus, supernatants from P. aeruginosa contain promising substances for theprevention and treatment of biofilm infections, although further work is required toidentity more active components.

Introduction

Staphylococcus epidermidis is an integral part of the com-mensalmicrobial communities on human skin but also acts asan important opportunistic pathogen, particularly in relation toin-dwelling medical device-related infections (Otto, 2009). Inpatients undergoing peritoneal dialysis (PD), S. epidermidisand other coagulase-negative staphylococci (CoNS) are thecause of a significant number of peritonitis episodes (Finkel-stein et al., 2002). The most common routes of bacterialaccess to the peritoneal cavity are through touch contamina-tion during instillation of the dialysis fluid and from infectionsat the catheter exit site (Cameron, 1995). As well as being

life-threatening, repeated episodes of peritonitis can causescarring and thickening of the peritoneum, resulting ininadequate dialysis and technique failure (Rubin et al., 1991).As PD catheters are implanted into the peritoneal cavity,

they are immediately covered with a layer of proteinsderived from the peritoneal fluid. These can act as adher-ence sites for bacteria and catheters removed from patientswith clinical symptoms of infection have been shown to becovered with microbial biofilms (Dasgupta et al., 1987).Biofilms on PD catheters can be difficult or impossible toeradicate and often consist of mixed microbial populations,which include both Staphylococci and Gram-negative rods(Barraclough et al., 2010).

Pathogens and Disease (2013), 67, 192–198, © 2013 Federation of European Microbiological Societies. Published by Blackwell Publishing Ltd. All rights reserved192

Pathogens and Disease ISSN 2049-632X

The presence of Pseudomonas aeruginosa in a mixedspecies biofilm has been shown to exert an inhibitory effectupon S. epidermidis (Qin et al., 2009; Pihl et al., 2010a, b).Pseudomonas aeruginosa is known to produce a range ofextracellular products including: elastase, pyocyanin, alka-line protease and rhamnolipids whose expression may beregulated through quorum-sensing mechanisms. Rhamnoli-pids are known to have biosurfactant properties that cancause cell detachment and contribute to the maintenance ofchannels within P. aeruginosa biofilms (Davey et al., 2003;Boles et al., 2005; Pamp & Tolker-Nielsen, 2007). However,P. aeruginosa rhamnolipids have also been shown to causedisruption of biofilms comprised of other species, such asBordetella bronchiseptica (Irie et al., 2005) and Bacilluspumilis (Dusane et al., 2010). Recently, the exopolysaccha-rides, Pel and Psl have also been proposed as agentscapable of dispersing biofilms of S. epidermidis (Qin et al.,2009). The pel operon in P. aeruginosa is responsible forthe production of the glucose-rich polymer, Pel (Friedman &Kolter, 2004), while the psl locus encodes the galactose-and mannose-rich exopolysaccharide, Psl (Ma et al., 2007).Therefore, extracellular products from P. aeruginosa mayrepresent good candidates for use in treatment strategies toeliminate S. epidermidis biofilms on in-dwelling medicaldevices such as catheters.We have developed a model to study biofilm formation by

S. epidermidis on PD catheters in the presence of serumproteins and have also used this, as well as a flow-cellmodel, to investigate the effects of biofilm supernatant fromP. aeruginosa on S. epidermidis biofilm formation. Theresults obtained suggest that rhamnolipids as well as otherunidentified P. aeruginosa components can disperse estab-lished S. epidermidis biofilms and components other thanrhamnolipids possess the capacity to displace surface-associated serum proteins to which S. epidermidis can bind.

Materials and methods

Bacterial strains and culture

Staphylococcus epidermidis strain C121 was isolated from aPD catheter (Pihl et al., 2010a), whereas P. aeruginosastrains 14:2 and 15159 were from chronic venous ulcers(Schmidtchen et al., 2001, 2003). Bacteria were grownin Todd-Hewitt medium (TH) (Difco) (5% CO2, 37 °C)until the mid-exponential growth phase was reached(OD600 nm = 0.5). Cells were centrifuged (4000 g, 15 min,4 °C), washed and adjusted to OD600 nm = 0.5 (correspond-ing to 108 CFU mL�1 for S. epidermidis and 109 CFU mL�1

for P. aeruginosa).Mutants deficient in production of Pel polysaccharide

(DpelA), Psl polysaccharide (DpslBCD) or rhamnolipids(DrhlAB) were constructed in P. aeruginosa strain 15159using the knockout plasmids pMPELA (Starkey et al., 2009),pMPSL-KO1 (Kirisits et al., 2005) and a rhlAB knockoutplasmid described by Boles et al. (2005). The knockoutplasmids were transferred into P. aeruginosa by triparentalmating as described previously (Andersen et al., 1998)using the helper strain E. coli HB101/RK600 with selection

on Pseudomonas isolation agar plates supplemented with60 lg mL�1 gentamicin. Resolution of single cross-overevents was achieved by streaking on 5% sucrose plates viathe counter-selectable sacB marker on the knockout plas-mid.

Preparation of P. aeruginosa biofilm supernatants

Mid-exponential growth phase cells (OD600 nm= 0.5) ofP. aeruginosa strains 14:2 and 15159 as well as the DpelA,DpslBCD or DrhlAB mutants in strain 15159 were grown asbiofilms in TH in tissue culture flasks (24 h, 5% CO2, 37 °C).Culture supernatants were centrifuged (15 min, 4000 g),filtered (0.45 lm), sterile-filtered (0.20 lm) and stored at�20 °C.

Characterization of catheter surfaces with conditioningfilms

Catheter pieces were coated at 37 °C with 10% (v/v) TH,10% (v/v) horse serum – inactivated by heating to 54 °C for1 h (hih serum), 10% (w/v) bovine serum albumin (BSA) orsupernatant from P. aeruginosa strain 14:2 for 24 h andrinsed with PBS. An uncoated catheter served as a control.Atomic force microscopy (AFM), (Dimension 3000 SPMTM;Digital Instruments), was performed in TappingModeTM in airusing etched silicon probes (Digital Instruments, SantaBarbara) with cantilever lengths of 125 nm and resonancefrequencies of 260–300 kHz. Areas (10 9 10 lm2, n = 9)were measured at a scan rate of 1.0 Hz and the arithmeticaverage height deviation from a mean plane (Sa) wascalculated. Three catheter pieces with each coating andthree different areas per piece were analysed.Conditioning films of serum proteins were further analy-

sed using two-dimensional gel electrophoresis (2-DE).Proteins were desorbed from PD catheters coated overnightwith serum using 0.012% (v/v) Triton X-100 and 0.006%(v/v) Tween 80 (1 h, RT). The desorbate was subjected toisoelectric focusing on IEF strips, pH 4–7 followed by SDS-PAGE on 14% gels and protein spots identified usingLC-MS/MS as described previously (Davies et al., 2009).

Biofilm formation by S. epidermidis C121 in a catheterflow-cell model

Silicone PD catheters (GAMBRO PDCATH), with theperforated parts removed, were coated by circulating 10%(v/v) hih serum, 10% (w/v) BSA or supernatant fromP. aeruginosa strain 14:2 for 24 h, 0.7 mL min�1, 37 °C.Uncoated catheters served as a control. Mid-exponentialgrowth phase cells in TH were allowed to adhere (2 h,0.7 mL min�1, 37 °C) and fresh TH then circulated for 24 h.Traps were used to prevent bubble formation (FC34,BioSurface Technologies Corp.). The distribution of S. ep-idermidis was studied with 16S rRNA gene FISH using theSTA3 probe (5′-3′sequence GCACATCAGCGTCAGT)(Tavares et al., 2008) labelled with ATTO-565 (red) aspreviously described (Pihl et al., 2010a, b). Catheters werecleaved longitudinally, flattened and viewed with an Eclipse

Pathogens and Disease (2013), 67, 192–198, © 2013 Federation of European Microbiological Societies. Published by Blackwell Publishing Ltd. All rights reserved 193

M. Pihl et al. Staphylococcus epidermidis biofilms on PD catheters

TE2000 inverted confocal laser scanning microscope(CLSM) (Nikon Corporation, Tokyo, Japan). Three experi-ments were performed and three pieces per catheter werestudied.

Growth of S. epidermidis in the presence ofsupernatant from P. aeruginosa

To determine whether supernatant from P. aeruginosa hada bacteriostatic effect, S. epidermidis was grown insuspension culture in TH broth or TH broth containing0.1% supernatant from P. aeruginosa strains 15159 or14:2 for 9 h at 37 °C in 5% CO2. Growth was monitoredhourly by measuring absorbance at 600 nm, and experi-ments were carried out in duplicate using independentcultures.

Biofilm formation by S. epidermidis in the presence of acoating with supernatant from P. aeruginosa

To investigate whether supernatant from P. aeruginosacould be used as a coating to prevent adherence ofS. epidermidis to surfaces, ibidi l-Slide VI flow-cells(Integrated BioDiagnostics, Munich, Germany) werecoated with either 10% hih serum for 18 h at roomtemperature followed by supernatant from wild-type strain15159 for 2 h at 37 °C in 5% CO2 or supernatant fromwild-type strain 15159 for 18 h at room temperaturefollowed by 10% hih serum for 18 h at room temperature.After coating, the channels were rinsed with TH broth andmid-exponential growth phase cells of S. epidermidisintroduced and allowed to form biofilms for 24 h (37 °C,5% CO2). After washing with TH, biofilms were visualizedusing the LIVE/DEAD� BacLightTM stain.

Dispersal of established S. epidermidis biofilms byP. aeruginosa supernatants and purified rhamnolipids

To investigate the effects of extracellular products fromP. aeruginosa on established S. epidermidis biofilms, mid-exponential growth phase cells of S. epidermidis wereintroduced into uncoated ibiTreat l-slide VI flow-cells orflow-cells coated with 10% hih serum for 18 h and allowed toform biofilms for 24 h (37 °C, 5% CO2). The establishedbiofilms were then exposed to native biofilm supernatantsfrom the wild-type 15159 strain or biofilm supernatantsheated to 60 °C for 15 min. After washing, the biofilms werestained with LIVE/DEAD� BacLightTM stain and viewed usingCLSM. To investigate the role of Pel, Psl and rhamnolipids,S. epidermidis biofilms formed on 10% hih as above wereexposed to biofilm supernatants from the DpelA, DpslBCDor DrhlAB mutants for 1 h. After washing with TH, biofilmswere stained with LIVE/DEAD� BacLightTM stain and viewedusing CLSM. To test the effects of rhamnolipids specifically,biofilms prepared as above were exposed to a purifiedmixture of 50% w/v mono- and 50% w/v di-rhamnolipids(JBR515; Jeneil Biosurfactant, Saukville, WI) at final con-centrations of 1, 10 or 100 ng mL�1 in PBS for 1 h at 37 °Cand analysed as above.

Image analysis and statistics

For each experiment, where not otherwise stated, 20randomly selected areas (total 0.9 mm2) were photo-graphed and substratum coverage estimated using thefunction ‘Cell Counting-Batch’ in the bioImage_L softwarepackage (Ch�avez de Paz, 2009). In each case, threeindependent experiments were undertaken and statisticalanalysis carried out using a one-way ANOVA with a Bonferronipost test to compare different treatments.

Results

Coatings on PD catheters and their effects on biofilmformation by S. epidermidis

In this study, serum was used to model the protein filmderived from the peritoneal fluid, which forms on PDcatheters in vivo. To identify the serum proteins whichadhered to the catheter surface, the conditioning film wasdesorbed and subjected to 2-DE. This revealed more than100 proteins of which the most abundant seen using massspectrometry were albumin, apolipoprotein A-1 precursorand haptoglobin precursor (Fig. 1).To confirm that the different coatings actually formed

conditioning films, uncoated catheters or catheters coatedwith 10% serum, 10% BSA or supernatant from biofilms ofP. aeruginosa strain 14:2 were investigated using AFM. Ineach case, coating increased the surface roughnesscompared with the uncoated catheter [Sa = 18.92 � 3.50(SD) nm] (Fig. 2, upper panel). Serum and BSA showed

Fig. 1 2-DE of a serum-derived protein coat on PD catheters. After

coating for 24 h with heat-inactivated horse serum, proteins were

desorbed from PD catheters using a solution of 0.012% (v/v) Triton

X-100 and 0.006% (v/v) Tween 80. The desorbate was subjected to

isoelectric focussing (pH 4–7) in the first dimension and SDS-PAGE in

14% gels in the second dimension and the gels stained with silver.

Protein spots were identified by LC-MS/MS.


Staphylococcus epidermidis biofilms on PD catheters M. Pihl et al.

similar degrees of roughness (Sa = 24.77 � 5.84 andSa = 25.09 � 6.27 nm, respectively), while the cathetercoated with the biofilm supernatant from P. aeruginosahad the greatest surface roughness with Sa = 37.64 �8.46 nm, that is, twice that of the uncoated surface. Thesedata thus confirm that serum, BSA and P. aeruginosasupernatant all form conditioning films on the cathetersurfaces which increase the surface roughness.Biofilm formation by S. epidermidis C121 was then

investigated in the PD flow-cell model in the presence ofthe different conditioning films. After 24 h, the biofilmsurface coverage on catheters coated with 10% serumwas fourfold higher than that on uncoated ones demon-strating that the serum coating contains proteins whichmediate binding of S. epidermidis (Fig. 2, lower panel). On10% BSA, the surface coverage was significantly lower thanon the serum suggesting that, even though albumin was thedominant protein in the serum-derived conditioning film, asshown using 2-DE (Fig. 1), other serum proteins areinvolved in binding of S. epidermidis to the serum coat.When catheters were coated with P. aeruginosa superna-tant, the surface coverage by S. epidermidis was lower thanthat seen in the presence of serum and similar to that seenfor the uncoated and albumin-coated surfaces indicating

that S. epidermidis did not bind to surface-associatedcomponents of the supernatant.

Effect of supernatant from P. aeruginosa on planktonicgrowth of S. epidermidis

To determine whether the biofilm supernatant had a bacte-riostatic effect, S. epidermidis cells were grown in TH brothsupplemented with 0.1% supernatant from P. aeruginosastrains 15159 and 14:2. This revealed that 14:2 supernatanthad no inhibitory effect, while that from 15159 had a small,but distinct, inhibitory effect in the late exponential phase(Fig. 3).

Effect of supernatant from P. aeruginosa on serumconditioning films

In these studies, an ibidi flow-cell model, where the surfacecoverage in the presence of a serum coating was greaterthan on the PD catheters, was used. Initially, coverage onuncoated flow-cells was compared with that on serum-coated ones and, as for the PD catheters, coating withserum increased adherence of S. epidermidis (Fig. 4). Todetermine whether supernatant from P. aeruginosa hasthe potential to prevent binding of S. epidermidis in thepresence of serum, flow-cells were coated with supernatantfollowed by serum, after which S. epidermidis cells wereallowed to attach for 24 h. Under these conditions, the levelof adherence was reduced by 90 � 3% as compared withthat on serum alone. As the supernatants had no (strain14:2), or only a small (strain 15159) effect upon growth, itappears most likely that the supernatant inhibited binding ofthe serum proteins to which S. epidermidis adhered.To determine whether components in the supernatant

from P. aeruginosa also had the capacity to displace anexisting serum coat, flow-cells coated with serum wereexposed to supernatant from P. aeruginosa strain 15159

Uncoated:Sa = 18.92 ± 3.50 nm

Serum:Sa = 24.77 ± 5.84 nm

Albumin:Sa = 25.09 ± 6.27 nm

Pa 14 : 2:Sa = 37.64 ± 8.46 nm

(a) (b) (c) (d)

Fig. 2 Influence of coating on surface roughness and biofilm formation

by Staphylococcus epidermidis in PD catheters. Uncoated PD catheters

or catheters coated with 10% heat-inactivated horse serum, 10% BSA

or supernatant from Pseudomonas aeruginosa strain 14 : 2 biofilms

were subjected to atomic force microscopy in TappingModeTM in air

using etched silicon probes (upper panel). Mid-exponential growth

phase cells of S. epidermidis C121 in TH were allowed to adhere (2 h,

0.7 mL min�1, 37 °C) to uncoated PD catheters or catheters coated with

10% serum, 10% BSA or biofilm supernatant from P. aeruginosa strain

14 : 2. After 24 h in TH, adhered cells were stained using fluorescence

in situ hybridization and visualized with CSLM. The bars show mean

surface coverage lm�2 � SE of three independent experiments (lower

panel).

Fig. 3 Effect of Pseudomonas aeruginosa supernatants on planktonic

growth of Staphylococcus epidermidis. Streptococcus epidermidis

was grown in suspension culture in TH broth (○) or TH broth containing

0.1% supernatant from P. aeruginosa strains 15159 (▲) or 14:2 (■) for9 h at 37 °C in 5% CO2. Growth was monitored hourly by measuring

absorbance at 600 nm. The graph shows the mean � SE of duplicate

experiments.



after which S. epidermidis cells were allowed to attach for24 h. This procedure also reduced bacterial adherence by87 � 1% compared with the level seen on the serum-coatedsurface (Fig. 4), suggesting that components in the super-natant were capable of displacing serum proteins from thesurface. Similar reductions were seen for supernatantscollected from mutant strains of P. aeruginosa, which lackedproduction of rhamnolipids (85 � 1%), PelA (76 � 3%) orPsl (87 � 1%). Displacement of serum proteins from thesurface is thus not mediated by rhamnolipids or the Pel Aand PslBCD polysaccharides.

Dispersal of established S. epidermidis biofilms bysupernatant from P. aeruginosa

EstablishedbiofilmsofS. epidermidis formedonPDcatheterswere exposed to supernatant from two P. aeruginosa strains(14:2 or 15159) for 1 h. Exposure to each caused a markedreduction in coveragewith only 10 � 2%of theoriginal biofilmremaining (Fig. 5). More than 96 � 1% of the cells remainingwere viable as shown by LIVE/DEAD� BacLightTM stainingshowing they were not killed by the treatment. A similarreduction in surface coverage was seen even after heatinactivation of the supernatant from strain 15159, suggestingthat proteolysis is unlikely to underlie the dispersal effect.

To further investigate which components cause dispersalof S. epidermidis biofilms, biofilms in ibidi flow-cells wereexposed to supernatants from wild-type P. aeruginosa15159 and the DrhlAB, DpelA or DpslBCD strains (Fig. 6).Treatment with supernatants lacking pelA or pslBCD causedcomparable levels of dispersal to that seen for the super-natant from the wild-type strain, suggesting that neither pelAnor pslBCD was responsible for the effect. In contrast,DrhlAB supernatant resulted in significantly less dispersal

SerumUncoatedSerum +Pa 15159

Pa 15159 +Serum

Fig. 4 Effect of coating with Pseudomonas aeruginosa supernatants on

biofilm formation by Staphylococcus epidermidis. Uncoated ibiTreat

l-slide VI flow-cells or flow-cells coated with 10% serum (18 h, room

temperature), 10% serum (18 h, room temperature) followed by super-

natant from P. aeruginosa strain 15159 biofilms (2 h, 37 °C) or

supernatant from P. aeruginosa biofilms (18 h, room temperature)

followed by 10% serum (18 h, room temperature) were incubated with

S. epidermidis C121 for 24 h. Adhered bacteria were analysed using

LIVE/DEAD� BacLightTM staining (upper panel). The graph shows the

mean surface coverage lm�2 � SE of three independent experiments

(lower panel).

Fig. 5 Dispersal of established Staphylococcus epidermidis biofilms on

PD catheters by supernatants of Pseudomonas aeruginosa. Biofilms of

S. epidermidis C121 grown for 24 h and analysed using fluorescence

in situ hybridization (FISH) before ( ) or after (□) exposure to

supernatants from P. aeruginosa strains 14 : 2 and 15159 for 1 h.

The graphs show mean surface coverage lm�2 � SE of three

independent experiments.

Fig. 6 Dispersal of established Staphylococcus epidermidis biofilms by

supernatants from DrhlAB, DpelA and DpslBCD mutant strains of

Pseudomonas aeruginosa as well as P. aeruginosa rhamnolipids.

Biofilms of S. epidermidis C121 were established in ibiTreat l-slide VI

flow-cells for 24 h prior to exposure to supernatants from the wild-type

15159 strain or DrhlAB, DpelA and DpslBCD isogenic mutants. Adhered

cells were stained using LIVE/DEAD� BacLightTM and visualized with

CSLM. The graph shows the surface coverage lm�2 expressed as a

percentage of control (surface coverage on serum-coated surfaces)

� SE of three independent experiments. The insert shows the mean

surface coverage lm�2 expressed as a percentage of control (surface

coverage on serum-coated surfaces) � SE of three independent

experiments after treatment with different concentrations of rhamnoli-

pids.



than that seen with the wild-type supernatant, reducing thesurface coverage to 59 � 2% of the control level. Thisindicates that rhamnolipids play a role in dispersal ofestablished biofilms of S. epidermidis. S. epidermidis bio-films were then challenged with different concentrations of amixture of mono- and di-rhamnolipids (Fig. 6, insert). Thisshowed that rhamnolipids had a dose-dependent effect thusconfirming that they play a role in dispersal. However, thefact that the dispersal effect was not completely abolished inthe supernatant lacking rhamnolipid (Fig. 6) suggests thatother components in the supernatant also play a role in thisphenomenon.

Discussion

Biofilm-related infections pose a serious threat to patientswith indwelling medical devices and new strategies to treatsuch infections are urgently required (Donlan, 2001). Toshed light upon the formation of biofilms on PD catheters,we have investigated the adherence of S. epidermidis in aPD catheter flow-cell model. To mimic in vivo conditions,where catheters inserted into the peritoneal cavity arecovered with proteins from the peritoneal fluid, (Cameron,1995), catheters were coated with serum. Adherence ofS. epidermidis was fourfold higher on coated than onuncoated surfaces suggesting that the bacterial cells boundto serum proteins on the catheter surface. Although thedominant surface-associated serum protein identified using2-DE was albumin, the surface coverage on pure albuminwas lower than that on serum suggesting that S. epidermi-dis C121 interacted mainly with less abundant serumproteins. This conclusion is supported by studies showingthat S. epidermidis expresses adhesins which can bind, forinstance, vitronectin (Gotz, 2002) and fibronectin (Hussainet al., 2001; Williams et al., 2002). Another parameteraffecting bacterial coverage on surfaces is topography.AFM has been used to study protein adsorption onto smoothbiomaterial surfaces (Dekeyser et al., 2008), and here, thistechnique confirmed that serum, BSA and P. aeruginosasupernatant adhered to the catheter surface increasing thesurface roughness on the nanometre level from approxi-mately 19 nm for the uncoated surface to 25 nm for serumand BSA and 38 nm for P. aeruginosa supernatant. Ascoverage on the serum-coated surface was greater than onthe albumin-coated one, surface roughness at this levelappeared to be less important for adherence of S. epide-rmidis (spherical diameter of 1 lm approximately.) thanspecific interactions with serum proteins.To investigate whether P. aeruginosa supernatant could

reduce adherence of S. epidermidis to serum-coated sur-faces, flow-cells were coated with supernatant followed byserum, or serum followed by supernatant. Following boththese treatments, the surface coverage of S. epidermidisafter 24 h was significantly lower than that seen on serum-coated surfaces. Possible interpretations of these dataare that the supernatant reduced adherence either byinhibiting binding of or displacing serum proteins to whichS. epidermidis adheres. To identify the components medi-ating the effect, supernatants from mutant strains lacking

rhlA, pelA or pslBCD were tested in the same way.Treatment with these supernatants gave similar results tothat of the wild-type strain indicating that components otherthan rhamnolipids, pelA and pslBCD are responsible for theeffect.In previous investigations we, and others, have shown

that culture fluid from P. aeruginosa can disperse estab-lished biofilms of S. epidermidis (Qin et al., 2009; Pihl et al.,2010a, b). In this study, the same phenomenon wasobserved in the PD catheter model using supernatants fromtwo strains of P. aeruginosa (14:2 and 15159). A similareffect was seen in the ibidi flow-cell model on serum-coatedsurfaces where only 23 � 6% of the biofilm remainedfollowing exposure to supernatant from P. aeruginosa. Thesupernatant from the mutant strain lacking production ofrhamnolipids showed a significantly reduced dispersalcapacity compared with the wild-type strain suggesting thatrhamnolipids are involved in the phenomenon. This wasconfirmed using purified rhamnolipids which gave rise to adose-dependent dispersal of S. epidermidis biofilms.Rhamnolipids, which are glycolipids, are known to exhibitbiosurfactant properties (Nitschke et al., 2005), andP. aeruginosa rhamnolipids have previously been shownto play a role in mediating detachment of P. aeruginosa frombiofilms as well as to exert an anti-adhesive effect againstother species, such as Bordetella bronchiseptica (Irie et al.,2005) and Bacillus pumilus (Dusane et al., 2010).In this study, the dispersal effect was not completely

abolished in the DrhlAB mutant strain (40% of the biofilmwas still removed) suggesting that substances other thanrhamnolipids are also involved in this process. One putativecandidate is a small messenger fatty acid cis-2-decenoicacid produced by P. aeruginosa in biofilm cultures whichhas been shown to induce a dispersion response in biofilmsformed by, for example, Staphylococcus aerueus (Davies &Marques, 2009). In a recent study by Qin et al., 2009; it wasproposed that established biofilms of S. epidermidis couldbe dissipated by extracellular polysaccharides. In thepresent study, however, no difference was seen betweensupernatants from mutant strains lacking production of thepolysaccharides Pel and Psl and wild type suggesting thatthese substances did not mediate the effect. The differencesin results may be attributable to strain differences in thestudies as P. aeruginosa strains have been shown to differin their dispersal capacity and S. epidermidis strains alsodiffer in their susceptibility to P. aeruginosa supernatants(Pihl et al., 2010b).In conclusion, we have shown that treatment with biofilm

supernatant from clinical strains of P. aeruginosa reducesadherence of S. epidermidis by interfering with the forma-tion of a conditioning film of serum proteins to which thebacteria bind. In addition, the supernatants had thecapacity to remove established biofilms of S. epidermidis,an effect that was mediated, in part, by rhamnolipids. Theresults presented here indicate that supernatants fromP. aeruginosa contain promising substances for the pre-vention and treatment of infections related to peritonealdialysis, although further work is required to determine theiridentity.



Acknowledgements

We thank Madeleine Blomqvist and Agnethe Henriksson forexcellent technical assistance and gratefully acknowledgeThe Aberdeen Proteome Facility (jointly funded by theSHEFC, BBSRC and the University of Aberdeen) for proteinidentification. We also thank our colleagues at Gambro ABfor their intellectual input to this work. This study wassupported by the Knowledge Foundation, Sweden; HjalmarSvenssons Research Foundation, The Royal Society of Artsand Sciences, Gothenburg, Sweden and by the DanishCouncil for Independent Research.

References

Andersen JB, Sternberg C, Poulsen LK, Bjorn SP, Givskov M &

Molin S (1998) New unstable variants of green fluorescent protein

for studies of transient gene expression in bacteria. Appl Environ

Microbiol 64: 2240–2246.Barraclough K, Hawley CM, McDonald SP, Brown FG, Rosman JB,

Wiggins KJ, Bannister KM & Johnson DW (2010) Polymicrobial

peritonitis in peritoneal dialysis patients in Australia: predictors,

treatment and outcomes. Am J Kidney Dis 55: 121–131.Boles BR, Thoendel M & Singh PK (2005) Rhamnolipids mediate

detachment of Pseudomonas aeruginosa from biofilms. Mol

Microbiol 57: 1210–1223.Cameron JS (1995) Host defences in continuous ambulatory

peritoneal dialysis and the genesis of peritonitis. Pediatr Nephrol

9: 647–662.Ch�avez de Paz LE (2009) Image analysis software based on color

segmentation for characterization of viability and physiological

activity of biofilms. Appl Environ Microbiol 75: 1734–1739.Dasgupta MK, Bettcher KB, Ulan RA, Burns V, Lam K, Dossetor JB

& Costerton JW (1987) Relationship of adherent bacterial biofilms

to peritonitis in chronic ambulatory peritoneal dialysis. Perit Dial

Bull 7: 168–173.Davey ME, Caiazza NC & O’Toole GA (2003) Rhamnolipid

surfactant production affects biofilm architecture in Pseudomonas

aeruginosa PAO1. J Bacteriol 185: 1027–1036.Davies DG & Marques CN (2009) A fatty acid messenger

is responsible for inducing dispersion in microbial biofilms.

J Bacteriol 191: 1393–1403.Davies JR, Svensater G & Herzberg MC (2009) Identification of

novel LPXTG-linked surface proteins from Streptococcus

gordonii. Microbiology 155: 1977–1988.Dekeyser CM, Zuyderhoff E, Giuliano RE, Federoff HJ, Dupont-

Gillain C & Rouxhet PG (2008) A rough morphology of the

adsorbed fibronectin layer favors adhesion of neuronal cells.

J Biomed Mater Res A 87: 116–128.Donlan RM (2001) Biofilms and device-association infections.

Emerg Infect Dis 7: 277–281.Dusane DH, Nancharaiah YV, Zinjarde SS & Venugopalan VP

(2010) Rhamnolipid mediated disruption of marine Bacillus

pumilis biofilms. Colloids Surf B Biointerfaces 81: 242–248.

Finkelstein ES, Jekel J, Troidle L, Gorban-Brennan N, Finkelstein

FO & Bia FJ (2002) Patterns of infection in patients maintained on

long-term peritoneal dialysis therapy with multiple episodes of

peritonitis. Am J Kidney Dis 39: 1278–1286.Friedman L & Kolter R (2004) Genes involved in matrix formation in

Pseudomonas aeruginosa PA14 biofilms. Mol Microbiol 51: 675–690.

Gotz F (2002) Staphylococcus and biofilms.Mol Microbiol 43: 1367–1378.

Hussain M, Heilmann C, Peters G & Herrmann M (2001) Teichoic

acid enhances adhesion of Staphylococcus epidermidis to

immobilized fibronectin. Microb Pathog 31: 261–270.Irie Y, O’Toole GA & Yuk MH (2005) Pseudomonas aeruginosa

rhamnolipids disperse Bordetella bronchiseptica biofilms. FEMS

Microbiol Lett 250: 237–243.Kirisits MJ, Prost L, Starkey M & Parsek MR (2005) Charac-

terization of colony morphology variants isolated from

Pseudomonas aeruginosa biofilms. Appl Environ Microbiol 71:

4809–4821.Ma L, Lu H, Sprinkle A, Parsek MR & Wozniak DJ (2007)

Pseudomonas aeruginosa Psl is a galactose- and mannose-rich

exopolysaccharide. J Bacteriol 189: 8353–8356.Nitschke M, Costa SG & Contiero J (2005) Rhamnolipid surfactants:

an update on the general aspects of these remarkable biomol-

ecules. Biotechnol Prog 21: 1593–1600.Otto M (2009) Staphylococcus epidermidis-the ‘accidental’ patho-

gen. Nat Rev Microbiol 7: 555–567.Pamp SJ & Tolker-Nielsen T (2007) Multiple roles of biosurfactants

in structural biofilm development by Pseudomonas aeruginosa.

J Bacteriol 189: 2531–2539.Pihl M, Davies JR, Chavez de Paz LE & Svensater G (2010a)

Differential effects of Pseudomonas aeruginosa on biofilm

formation by different strains of Staphylococcus epidermidis.

FEMS Immunol Med Microbiol 59: 439–446.Pihl M, Chavez de Paz LE, Schmidtchen A, Svensater G & Davies

JR (2010b) Effects of clinical isolates of Pseudomonas aerugin-

osa on Staphylococcus epidermidis biofilm formation. FEMS

Immunol Med Microbiol 59: 504–512.Qin Z, Yang L, Qu D, Molin S & Tolker-Nielsen T (2009)

Pseudomonas aeruginosa extracellular products inhibit staphylo-

coccal growth, and disrupt established biofilms produced by

Staphylococcus epidermidis. Microbiology 155: 2148–2156.Rubin J, Herrera GA & Collins D (1991) An autopsy study of the

peritoneal cavity from patients on continuous ambulatory perito-

neal dialysis. Am J Kidney Dis 18: 97–102.Schmidtchen A, Wolff H & Hansson C (2001) Differential proteinase

expression by Pseudomonas aeruginosa derived from chronic leg

ulcers. Acta Derm Venereol 81: 406–409.Schmidtchen A, Holst E, Tapper H & Bjorck L (2003) Elastase-

producing Pseudomonas aeruginosa degrade plasma proteins

and extracellular products of human skin and fibroblasts, and

inhibit fibroblast growth. Microb Pathog 34: 47–55.Williams RJ, Henderson B, Sharp LJ & Nair SP (2002) Identification

of a fibronectin-binding protein from Staphylococcus epidermidis.

Infect Immun 70: 6805–6810.



biofilm formation by staphylococcus epidermidis on peritoneal dialysis catheters and the effects of...

Documents