Article available online at http://www.idealibrary.com on Microbial Pathogenesis 2001; 31: 261–270

PATHOGENESISMICROBIAL

doi:10.1006/mpat.2001.0469

Teichoic acid enhances adhesion ofStaphylococcus epidermidis to immobilizedfibronectinMuzaffar Hussain∗, Christine Heilmann, Georg Peters &Mathias Herrmann†

Institute of Medical Microbiology, University Hospital of Muenster, Germany

(Received June 11, 2001; accepted in revised form August 30, 2001)

Adhesion is a prerequisite for coagulase-negative staphylococci to cause invasive disease andmay be mediated by adhesive host molecules adsorbed on implanted polymers. In this study,we can confirm previous observations demonstrating binding of Staphylococcus epidermidis tofibronectin (FN) adsorbed polymer surfaces. So far, the nature of FN-recognizing adhesin(s) in S.epidermidis remains elusive. Since teichoic acids (TA) have been shown to exert binding functionsfor extracellular matrix molecules in several Gram-positive species, we have purified wall TA ofS. epidermidis laboratory strains KH11 and RP62A, as well as clinical isolate AB9. Using apolymethylmethacrylate (PMMA) coverslip adhesion assay, a microtitre plate assay and a particleagglutination assay, we found that purified TA significantly enhanced adhesion of S. epidermidisKH11 and RP62A to FN coated surfaces. Enhanced adhesion was dose-dependent and saturable.Preincubation, either of microorganisms or of FN coated surfaces, with TA promoted adhesion,while adhesion to TA-adsorbed PMMA was comparably low. This observation may suggest apotential role of cell wall carbohydrates as bridging molecules between microorganisms andimmobilized FN in early steps of S. epidermidis pathogenesis. 2001 Academic Press

Key words: Staphylococcus epidermidis, bacterial adhesion, fibronectin, teichnoic acid.

considered harmless commensals, are now re-Introductioncognized as the most important pathogens forinfections of implanted prosthetic material and

Staphylococcus epidermidis and other coagulase- catheter-associated blood stream infection. Thenegative staphylococcal species, previously events associated with initiation and course of

these infections have been intensely studied, anddistinctive mechanisms contributing to adhesion

† Present address: Institute of Medical Microbiology and of microorganisms to the uncoated polymer [1,Hygiene, University of the Saarland Medical School, Bldg. 2], proliferation and accumulation on the surface#43, Homburg 66421, Germany.resulting in biofilm formation [3–6], and per-∗Author for correspondence. E-mail: muzaffa@uni-muenster.

de sistence of the sessile bacterial population until

0882–4010/01/120261+10 $35.00/0 2001 Academic Press

262 M. Hussain et al.

removal of the polymer material [7, 8] have beenidentified.

Generally, it is thought that S. epidermidisprosthetic device infections occur throughcolonization of naked polymer surfaces bymicroorganisms resident in the surgical or thecatheter wound [9]. However, in addition toindwelling device infection, S. epidermidis arealso capable of causing invasive disease, suchas osteomyelitis [10], deep tissue infection [11]or native valve endocarditis [12]. Therefore, inthese entities as well as in certain types of poly-mer-associated infection (e.g., tunnel infectionsof implanted catheters or ventricular assist de-vice drive lines), S. epidermidis may bind toimmobilized host factors rather than to un-adsorbed plastic [13]. In vitro studies have dem-onstrated that S. epidermidis adheres to polymer-adsorbed fibronectin (FN), yet, in contrast to theFN-binding proteins in S. aureus [14], a FN-binding adhesin in S. epidermidis has not beenidentified. Teichoic acid (TA), an essential wallconstituent of staphylococci, has been implicatedin the binding of S. epidermidis binding fibrinclots [15] and in the adhesion of S. aureus toepithelial cells [16]. Therefore, the goal of thisstudy was to evaluate a potential role of S. 0.0

1.0 (b)

OD

460

nm

KH11 RP62A AD22 AB7 AB9

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0.0

0.7 (a)

Adh

esio

n (

%)

KH11 RP62A AD22 AB7 AB9

0.6

0.5

0.4

0.3

0.2

0.1

epidermidis TA as a candidate molecule me-Figure 1. Promotion of adhesion of five strains ofdiating binding of S. epidermidis to immobilizedS. epidermidis to FN-adsorbed surfaces. (a) PMMAFN.coverslips (b) microtitre plate wells. Binding to BSA(Φ) used as control and to FN (Ε). Experimentalconditions described in Materials and Methods.

Results

Interaction of S. epidermidis with soluble FN coated PMMA cover slips. Pre-adsorption ofFN to PMMA promoted adhesion of the five S.and immobilized FNepidermidis isolates tested in this study [Fig. 1(a)].(iv) Adhesion to FN coated microtitre plate.(i) Coagglutination test. S. epidermidis bound

soluble FN as demonstrated in a coagglutination Promotion in adhesion of these isolates to FN-adsorbed surfaces was also observed when usingreaction using formalin fixed and heat killed

Staphylococcus aureus Cowan 1 cells sensitized a polystyrene microtitre plate adhesion assay[Fig. 1(b)]. Adhesion of strain KH11 to FN coatedwith anti-FN-Abs. All five tested S. epidermidis

isolates revealed a positive agglutination re- polystyrene surfaces was found to be dose-dependent with concentrations of FN 5–50 �g/action, yet agglutination occurred to a different

extent with strains KH11 and AD22 yielding a ml (data not shown).(+++), strains RP62A and AB7 a (++) andstrain AB9 a (+) reaction. (ii) Particle ag-glutination test. Using FN coated latex particles Effect of treatment of bacteria with trypsin

and periodate oxidation on adhesionresults similar to those of coagglutination testresults were obtained. Two strains, KH11 andAD22, showed strong agglutination reaction This experiment was performed to characterize

the nature of the receptor for FN on staphylo-(+++), two strains, RP62A and AB7, ag-glutinated moderately (++) and a fifth strain, coccal cells surface. In separate experiments bac-

terial cell surface proteins were removed byAB9, agglutinated weakly (+). (iii) Adhesion to

Teichoic acid and S. epidermidis adhesion 263

Table 1. Adhesion of S. epidermidis to FN-adsorbed PMMA or FN-adsorbed polystyrene and effect of solubleTA

Radiometric adhesion assay Microtitre plate assay

Albumin FN FN plus TA Albumin FN FN plus TAPercent adhesion (mean, SD) OD492nm (mean, SD)

KH11 0.03 (0.01) 0.51 (0.05) 1.74 (0.11) 0.09 (0.02) 0.22 (0.06) 0.54 (0.08)RP62A 0.09 (0.07) 0.57 (0.06) 1.74 (0.12) 0.08 (0.50) 0.53 (0.04) 0.79 (0.09)AD22 0.04 (0.02) 0.65 (0.02) 1.59 (0.16) 0.04 (0.03) 0.57 (0.13) 0.93 (0.04)AB7 0.06 (0.05) 0.27 (0.03) 0.74 (0.10) 0.08 (0.03) 0.10 (0.01) 0.42 (0.03)AB9 0.09 (0.06) 0.39 (0.01) 0.45 (0.08) 0.05 (0.04) 0.32 (0.04) 0.56 (0.10)

trypsination and carbohydrates were removed p-anisidine or AgNO3/NaOH. In another chro-matogram, a glycerol–phosphate spot was dis-by periodate oxidation. The adhesion reaction

was found to be both trypsin and periodate- closed with ammonium molybdate+perchloricacid+HCl mixture. The chromatogram stainedsensitive, because pre-treatment of bacteria with

these reagents caused a reduction of KH11 ad- with ninhydrin showed only one spot close tothe alanine marker. Hydrolysates prepared fromhesion to FN-PMMA by 65 and 74%, and of

RP62A by 61 and 63%, respectively. The effect isolated TA from S. epidermidis RP62A, KH11and AB9 did exhibit these characteristics typicalof trypsin suggested a possible role of bacterial

surface proteins in the interaction with im- for staphylococcal wall TA [18, 19]. (ii) En-zymatic and chemical analysis. In TA from strainmobilized FN, adhesion assays were performed

using a mutant defective in expression of AtlE. RP62A, 165 �g of glucose and 40 �g of glu-cosamine were determined per mg of freeze-AtlE is an autolysin of strain S. epidermidis O-

47 with vitronectin (VN) binding activity in dried material. Taking molecular masses of poly-mer units of glucose and glucosamine as 162Western-ligand assays [1], and the AtlE analogue

from strain S. epidermidis strain AB9 recognized and 203, respectively, then these data suggestspresence of 1 �M of glucose and 0.2 �M of glu-both FN and VN [17]. Adhesion of the mutant

to FN-PMMA was slightly elevated (mean ad- cosamine in each mg of freeze-dried material.(iii) Reaction with lectins. TA isolated from threehesion, 0.86%) compared to the wild-type strain

(0.34%), but this was accompanied by slightly strains (RP62A, KH11 and AB9) showed strongprecipitation reactions with lectins concan-enhanced adhesion to BSA-PMMA (0.11 vs

0.06%). Thus, AtlE does not appear to mediate avalin-A and lectin from Triticum vulgaris (wheatgerm lectin) specific for glucose and glucosamineadhesion of S. epidermidis O-47 to FN-adsorbed

surfaces. respectively. The reaction with lectins suggestspresence of glucose and glucosamine as glucosysubstituents.

Purification and analysis of TA

(i) Thin layer chromatography. Cell walls isol- Role of TA in the FN–S. epidermidisinteractionated from the middle layer of the sucrose gra-

dient were hydrolysed in 6 M HCl for 18 h at105°C, and amino acids were separated by thin These findings prompted further evaluation of

the role of cell wall TA to the binding reaction.layer chromatography. Only alanine, glycine,glutamic acid and lysine were detected with Upon addition of TA purified by anion exchange

chromatography (final concentration, 16.5 �g/ninhydrin, a clear indication of a pure cell wallpreparation. TA extracted from purified cell ml), an increase in adhesion of all five tested

isolates, both in the radiometric adhesion assaywalls by either procedure was hydrolysed in3 M HCl for 3 h and analysed by thin layer as well as in the microtitre plate assay, were

observed with strain AB9 being least promotedchromatography. Spots having Rglucose valuesclose to those of glucose, glucosamine and (Table 1). Using the particle agglutination assay,

addition of TA purified from either RP62A, AB9glycerol–phosphate were identified with

264 M. Hussain et al.

0.0

3.0(b)

Adh

esio

n (

%)

KH110.0

2.0(a)

Adh

esio

n (

%)

RP62A

2.5

2.0

1.5

1.0

0.5

1.5

1.0

0.5

KH11

***

***

Figure 2. (a) Effect of different TA preparations on promotion of S. epidermidis adhesion to FN-adsorbedPMMA. Untreated (Φ) or FN-pre-adsorbed (25 �g in 1 ml PBS, 60 min, 37°C) (ΕΦ, Ε, and ΕΦ) PMMAcoverslips were incubated with 4×106 cfu/ml of [3H]-thymidine-labelled bacteria in 1 ml of Ca++/Mg++-PBS supplemented with 0.5% human serum albumin (60 min, 37°C, shaking water bath), then washed. Thenumber of adherent microorganisms was expressed as percentage of inoculum. In some experiments,suspensions containing radiolabelled bacteria and FN-PMMA were supplemented with TA purified fromthe DEAE cellulose column (Ε) or from the concanavalin-A column (ΕΦ) (16.5 �g/ml, respectively) preparedas described in Materials and Methods. Shown are means of three determinations±SD. Statistical analysiswas performed using one-way analysis of variance (ANOVA) and subsequent Student-Newman-Keuls testfor determination of the significance levels of differences between observations. ANOVA, P<0.001; ∗∗∗,P<0.01 of FN alone vs DEAE- or concavalin-A-eluate supplement. (b) Effect of two TA preparations onadhesion of S. epidermidis KH11 to FN-PMMA. TA was prepared either from S. epidermidis KH11 (Ε) orfrom RP62A (ΕΦ) and used to supplement the adhesion buffer as described in Fig. 1.

or KH11 to FN coated beads resulted in a mod- extent when compared with TA from strainKH11 [Fig. 2(b)]. Adhesion was found to be dose-erate agglutination (+/++), yet, when FN-

beads were first treated with purified TA and dependent with concentrations of >16.5 �g/ml,resulting in saturation of TA-mediated adhesionthen incubated with washed cells of strains

KH11, RP62A or AB9, an immediate and strong enhancement (Fig. 3). While these experimentsdemonstrated the adhesion-promoting effect of(+++) agglutination reaction was detected.

Control (BSA-adsorbed) beads revealed no TA in co-incubation with FN-PMMA andstaphylococci (Fig. 4), this effect was also to beagglutination whether incubated with TA alone

or with TA plus bacterial cells. Strains KH11 observed, albeit to a lesser extent, if PMMAcoverslips were preincubated with TA thenand RP62A were further studied. Lectin-affinity

chromatography did not enhance the biologic rinsed and tested for adhesion. Moreover, theextent of adhesion of bacteria pre-adsorbed withactivity of purified TA, since TA-enhanced ad-

hesion was similar using either purification pro- TA then washed was found to be almost identicalto adhesion to FN-PMMA in the presence of TA.cedure [Fig. 2(a)]. No discernible difference in

adhesion was observed when TA isolated upon In contrast, microorganisms adhered to PMMApre-adsorbed with TA only to a small extent.either TA extraction procedure (trichloroacetic

acid or lysostaphin–lysozyme extraction) wasused (not shown). For subsequent experiments, Interaction of TA with DIG-labelled ligandsTA was prepared from digestion of cell wallswith lysostaphin–lysozyme and purified by DIG labelled FN bound very well to TA and DIG

labelled FG bound to a lesser extent to immo-anion-exchange chromatography. The promo-tion of adhesion by TA was not isolate-de- bilized TA on a nitrocellulose membrane, while

DIG labelled VN did not show any binding topendent, i.e. TA purified from strain RP62Apromoted adhesion of strain KH11 to a similar TA (data not shown).

Teichoic acid and S. epidermidis adhesion 265

addition of oxidized TA are: strain KH11 OD492nm,0.24±0.05 (mean±SD, n=3), 0.25 (0.08); strainAB9 0.36 (0.07), 0.35 (0.05).

Effect of soluble FN on adhesion to TAcoated polystyrene surface

When a bacterial suspension containing 50 �g/ml FN was added to a microtitre plate wells pre-coated with TA, a slightly enhanced adhesion of

0.2

165

1.4

Teichoic acid added (µg/ml)

Adh

esio

n (

%)

1.65 16.5

0.4

0.6

0.8

1.0

1.2

strain KH11 and AB9 was observed: strain KH11,OD492nm, 0.09±0.03 (mean±SD, n=3) enhancedFigure 3. Adhesion promotion by TA as a functionto 0.14 (0.05) and strain AB9, 0.07 (0.04) enhancedof TA concentration. Various concentrations of TAto 0.12 (0.05). The possible explanation includes:were supplemented to the adhesion buffer as de-

scribed in Fig. 1, and adhesion of radiolabelled S. (i) a small amount of immobilized TA on aepidermidis KH11 was subsequently determined. polystyrene surface or (ii) the fact that functionalShown are means of triplicate determinations±SD. binding sites of TA may not be any more avail-

able for bridging between cell surface com-ponents and soluble FN after immobilization ofTA.

Discussion

In this study, we describe the interaction of S.epidermidis with FN and identify cell wall TA asan adhesive molecule recognizing adsorbed FN.After contact with blood, plasma proteins aresequentially adsorbed to polymeric surfaces,with high-abundant, low molecular-weight pro-teins, such as albumin, being rapidly and pref-

0.0

2.0

Adh

esio

n (

%)

RP62A

1.5

1.0

0.5

KH11

*** ***

erentially adsorbed but subsequently replacedFigure 4. Effect of coincubation vs preincubation ofby less abundant, more surface active proteinsTA either with FN-PMMA or with bacteria. Controlsuch as FN and fibrinogen [20]. These ob-assays either without FN (Φ) or with FN-PMMAservations are in good agreement with dataalone (ΕΦ) were performed as described in Fig. 1.obtained from ex vivo human [21] and canineAlternatively, TA (16.5 �g/ml) was either added dur-

ing the adhesion incubation (Ε), or either FN-PMMA [22] catheter materials witnessing deposition of(ΕΦ) or radiolabelled bacteria (P) were preincubated significant amounts of adhesive molecules onwith TA, then washed and incubated for bacterial the polymer substrates. Accordingly, variousadhesion. In some experiments, PMMA coverslips studies on the role of FN on S. epidermidis ad-without FN were incubated in the presence of TA hesion, albeit with contrasting results, have been(16.5 �g/ml) and radiolabelled bacteria (Q). Shown reported [21, 23–29], while adhesion to otherare means±SD of three determinations. ANOVA, immobilized matrix proteins was found to showP<0.001, ∗∗∗, P<0.01 of FN-PMMA alone vs co-

more interstrain variability or to be low [23].incubation or preincubation with TA.Experimental conditions contributing to dif-ferences in physicochemical binding forces, not-ably in surface hydrophobicity, may account atleast for a part of these differences [1, 26, 30].Effect of oxidized TA on adhesionTo model the physiologic conditions addressedin this study, we have performed our assays inThe oxidized TA failed to enhance adhesion of

strains KH11 and AB9 to polystyrene microtitre a physiologic, i.e. albumin-containing milieu,and by using three different types of assays weplate coated with FN. Values before and after

266 M. Hussain et al.

have confirmed that polymer-adsorbed FN does is also necessary to explore the potential of thesefindings for novel prophylactic or therapeuticindeed promote adhesion of S. epidermidis to the

substrate. strategies directed against S. epidermidis in-fections.Recently, we have reported on putative S.

epidermidis adhesins recognizing FN in Western-ligand blots [17]. While one of the releasedmolecules was ornithine carbamoyltransferase,

Materials and Methodsan intracellular enzyme not involved in adhesionto surfaces, here we present evidence that an-

Bacterial strains and culture mediaother S. epidermidis protein recognizing VN andFN in Western-ligand assays, the S. epidermidis

S. epidermidis isolates KH11 [36] and RP62Aautolysin AtlE, does not contribute to adhesion(ATCC 35984) were used. In addition, clinical S.to FN. Thus, a proteinaceous adhesin re-epidermidis blood stream isolates AD22, AB7 andcognizing FN remains unidentified in S. epi-AB9 were selected from the strain collection ofdermidis.our institute for further study. For determinationGram-positive cocci have been shown to in-of the role of AtlE on adhesion to FN, in someteract with whole cells or extracellular matrixexperiments wild-type strain O-47 and its AtlE-protein, notably FN, via TA and/or lipoteichoicdeficient mutant, mut1 [1], was also used. Iden-acid. In numerous studies {reviewed in [31]},tification of clinical isolates as S. epidermidis onstreptococcal lipoteichoic acid (LTA) has beena species level was performed using the ID 32suggested to bind to FN and these observationsbiochemical test kit (BioMerieux, Marcy-l’Etoile,have prompted a model of streptococcal at-France). Working cultures were maintained ontachment to extracellular matrix or eukaryoticblood agar plates (for strain mut1 supplementedcells involving a two-step mechanism includingwith erythromycin 10 �g/ml), and fresh over-LTA and proteinaceous receptors [32]. Thisnight cultures were grown on brain–heart broth/model has been challenged but provides a basisagar (Merck, Darmstadt, Germany) and Mullerfor further research in multiple-step adhesionHinton broth/agar (Mast, Merseyside, U.K.).mechanisms of Gram-positive cocci including

staphylococci [15, 16, 33]. In general, researchon the role of cell wall TA as adhesins is

Solid phase adhesion assayhampered by the fact that mutants of sta-phylococci deficient in TA are not available, since

For radiomateric analysis of S. epidermidis ad-wall TA appear to be essential for survival.hesion to coated solid surfaces, a previouslyOur evidence suggesting a role of TA as andescribed assay was used [37]. Briefly, solu-S. epidermidis adhesin for FN is based on thetion containing 25 �g/ml human plasma FNfollowing observations: (i) the demonstration of(Chemicon, Temecula, CA, U.S.A.) was alloweddose-dependency and saturability of TA-to adsorb to PMMA coverslips for 60 min atenhanced adhesion of S. epidermidis to adsorbed37°C. Thereafter, coverslips were washed withFN, (ii) the fact that S. epidermidis adheres onlyPBS and incubated in a shaking water bath withslightly to TA-pre-adsorbed PMMA in the ab-[3H] thymidine-labelled S. epidermidis cellssence of FN, and (iii) the effect of carbohydrate-[4×106 cfu in 1 ml PBS containing 0.5% bovinemodifying periodate on adhesion. As proteaseserum albumin (Sigma-Aldrich, Deisenhofen,treatment also reduced adhesion, both protein-Germany)] for 60 min at 37°C. After adhesion,and carbohydrate-type ligands may be involved,the PMMA coverslips were washed with PBShowever, it has been demonstrated for strep-three times, and adherent cpm counts were de-tococci that trypsin digestion of bacteria removestermined. In some experiments TA was addedmost of the TA from cell surface [34]. Our find-as g of glucose per ml of adhesion assay.ings suggest that TA forms a bridge between

the FN molecule and the S. epidermidis cell sur-face. This may occur via rebinding of TA throughits anionic backbone to cationic molecules on Microtitre plate adhesion assaycell surface [35], however, additional research,e.g. using mutants defective in TA-modifying Polystyrene plates (96 well) (Greiner, Frick-

enhausen, Germany) were coated with FNenzymes, is necessary to further determine themolecular mechanisms involved. Further work (50 �g/ml−1 in 50 mM sodium carbonate buffer,

Teichoic acid and S. epidermidis adhesion 267

pH 9.6) for 18 h at 4°C, then blocked with 3% Preparation of cell wallsBSA in TBS (25 mM Tris–HCl, 100 mM NaCl,pH 7.5) and washed. Bacteria were grown for For isolation of cell wall, bacteria were grown

in 800 ml TSB in two 1 l flasks, each inoculated18 h at 37°C, washed once with TBS, adjustedto an OD578nm 1.0 in TBS, and to each well 200 �l with 0.5 ml of an 18 h culture in the same

medium and incubated with shaking at 37°C forof bacterial suspension was added. Plates wereincubated at 37°C for 1 h, then wells were 18 h. Bacteria were pelleted by centrifugation

(10 000×g, 10 min, 4°C) and washed twice withwashed three times, stained with safranin for1 min, and the plate was read at 492 nm on a distilled water. Washed bacteria were broken by

the method of Huff et al. [41], shaking withTitertek-multiscan-8 (Flow laboratories, Bonn,Germany). To study the effect of soluble FN on glass beads in a Braun cell homogenizer (Braun,

Melsungen, Germany). For the separation ofadhesion to TA coated polystyrene surface, theabove detailed experiment was performed with walls by sucrose gradient centrifugation, the

method of Yoshida et al. [42] was used. Cellthe following two modifications: (i) the surfacewas coated with TA (16.5 �g/ml) and (ii) the walls from the middle layer of sucrose gradient

were resuspended in 50 mM Tris–HCl pH 7.8bacterial suspension contained 50 �g/ml FN.plus trypsin 250 �g/ml, and kept at 37°C withshaking for 4 h. The suspension was centrifuged(32 000×g, 30 min, 4°C), sedimented walls wereCoagglutination test washed with 1 M NaCl, then six times withdistilled water, and freeze-dried. For charac-The preparation of coagglutination reagent terization, freeze-dried cell walls (5 mg) in 6 M(SAC+anti-FN-Abs) involved fixing of protein HCl (1 ml) in a sealed glass tube were heated atA found on surface of S. aureus cowan1 (ATCC 105°C in an oven for 18 h. Acid hydrolysate was12598) cells with formaldehyde and then killing dried in a dessicator over silica gel in presencebacteria by heating [38, 39]. To the resulting of NaOH under vacuum. Acid from hydrolysateprotein A solid phase, anti-FN-Abs (Dako) were was completely removed by suspending thecoupled. Bacteria were grown in BHI for 18 h at dried hydrolysate in a few drops of water and37°C with shaking, washed twice with PBS and repeating the drying step at least three times.incubated at room temperature after suspension Amino acids were separated by thin layerin PBS containing 50 �g/ml FN. After 1 h, FN chromatography and detected by ninhydrintreated bacteria were washed twice with PBS spray.and resuspended in PBS to OD598nm of 1.0. For

the coagglutination test, 10 �l of FN exposedbacterial suspension was mixed with 20 �l of Isolation of TA from cell wallscoagglutination reagent on a microscope slide.After 3 min reactions were scored as negative TA was extracted from cell walls by two(−) or from weakly positive (+) to strongly methods: (i) digestion with lysozyme+positive (+++). No agglutination was found lysostaphin. Cell walls (1 mg/ml) were sus-when the PBS used to suspend bacteria and pended in 50 mM Tris–HCl, pH 8, containingcoagglutination reagent was tested as a control. 1.2 mM EDTA (Sigma) and 0.145 M NaCl (Merck,

Darmstadt, Germany). The suspension was firstdigested with a mixture of recombinant ly-sostaphin (Applied Micro Inc., New York,Particle agglutination assayU.S.A.) and lysozyme (Merck) for 3 h at 37°C,followed by 1 mg/ml trypsin (Sigma) digestionMethod of Naidu et al. [40] was used. Briefly,

200 �l of blue latex beads (Sigma) were pre- for 4 h at 37°C. After enzyme treatment, thesuspension was centrifuged (30 000×g, 4°C,adsorbed with FN (100 �g/ml) or with BSA for

control in sodium carbonate buffer (pH 9.5), then 30 min), then passed through a 0.22 �m mem-brane filter (Millipore Corporation, Bedford,washed. Bacteria were grown overnight in BHI,

washed, and resuspended in PBS. FN-latex U.S.A.) to remove cell wall fragments. Five vol-ume of ethanol was added to the filtrate. Afterbeads (20 �l) were mixed with 20 �l of bacterial

suspension, and after 3 min the agglutination overnight equilibration at 4°C, the liquid wasdiscarded and the precipitate containing TA wasreaction was scored as negative (−), or from

weak positive (+) to strong positive (+++). freeze-dried. (ii) Extraction with trichloroacetic

268 M. Hussain et al.

acid (TCA) by the method of Archibald et al. and McAllan [50] with glucosamine. HCl(Sigma) as standard was used.[43].

Reaction with lectinsAnion-exchange and affinitychromatography Two lectins, concanavalin-A (Sigma) and lectin

from Triticum vulgaris (wheat germ lectin;TA was fractionated by ion-exchange chromato- Sigma), were used. TA suspension (20 �l) wasgraphy on DEAE-cellulose and affinity chro- added to lectin solution (10 mg/ml, 20 �l) on amatography as described previously for the frac- glass slide. Results were recorded as negativetionation of extracellular products [19, 44]. (i) or positive (clumping) after 1 min.Briefly, the DEAE cellulose (Sigma) column(1.5×11 cm) was equilibrated with 0.1 M am-monium acetate (Merck) buffer (pH 5.5), and Interaction of TA with FN25 mg of crude TA in 2 ml buffer was loadedonto the column. The column was eluted step- TA was immobilized on nitrocellulose mem-wise with increasing concentrations of NaCl in brane (Schleicher & Scuell, Dasel, Germany) bybuffer. Fractions were analysed by a phenol/ allowing a 5 �l TA solution to penetrate thesulfuric acid assay [45] and an assay for total membrane and subsequently dried. The re-phosphorus [46]. Fractions making peaks were maining binding sites on the membrane werepooled, dialysed against distilled water for 24 h, blocked by 3% BSA (fraction V, Sigma), thenand freeze-dried. (ii) Affinity chromatography. separate membranes were probed with DIG-Material eluted with 0.5 M NaCl in 50 mM am- labelled FN, FG or VN for 1 h. The membranesmonium acetate from DEAE-cellulose was were washed three times with TBS-T and trans-loaded on a concanavalin A sepharose 4-B col- ferred to a solution of alkaline phosphatase-umn (Sigma) equilibrated with 0.1 M sodium conjugated anti-DIG-Abs (Roche, Mannheim,acetate (pH 6.0) containing 1 M NaCl, 1 mM of Germany) for 1 h. Spots binding FN, FG or VNeach CaCl2, MgCl2, MnCl2 and 0.02% NaN3. The were detected in an alkaline phosphatase colourcolumn was washed with 20 bed volumes of reaction (Bio-Rad) in accordance with the in-loading buffer, then bound material was eluted structions of the supplier.with five bed volumes of 0.1 M methyl �-D-glucopyranoside in loading buffer, and assayed

Removal of surface proteins andfor phosphorus. The eluted material was dia-carbohydrateslysed against distilled water then freeze-dried.

(i) Trypsin treatment. Bacteria were treated with25 �g/ml trypsin in PBS (pH 7.2) at 37°C forAnalysis of TA by thin layer1 h. Cells were washed twice with PBS andchromatographyresuspended in PBS for testing. (ii) Periodateoxidation. Bacterial cell surface carbohydratesTA was hydrolysed in 3 M HCl for 3 h at 100°Cwere oxidized by exposure in 0.01 M NaIO4in a boiling water bath. Acid from hydrolysatein 0.05 M sodium acetate, pH 4.2 for 30 min,was removed as detailed earlier for cell wallwashed twice with PBS and resuspended in PBShydrolysis. Prepared hydrolysate was separatedfor testing. (iii) Periodate oxidation of TA. Aon thin layer chromatography plates and de-100 �l of TA solution was oxidized by addingtected using standard methods [47, 48].100 �l of 0.02 M NaIO4 in 0.10 M sodium acetatepH 4.2. After 30 min, 10 �l of ethylene glycol wasadded, and the pH was adjusted to 7.0 with 1 M

Estimation of glucose and glucosamine in Tris–HCl pH 8.0.TA preparation

Glucose concentration was determined by a Statisticshighly specific hexokinase method of Carroll etal. [49] using glucose kit 115-A (Sigma). For All experiments were performed at least in

triplicate and repeated as indicated. Statisticalmeasurement of glucosamine, method of Levvy

Teichoic acid and S. epidermidis adhesion 269

11 Rupp ME. Coagulase-negative staphylococcal infec-analysis was performed with the Biostatisticstions: an update regarding recognition and manage-program [51] on a PC.ment. Curr Clin Top Infect Dis 1997; 17: 51–87.

12 Whitener C, Caputo GM, Weitekamp MR, KarchmerAW. Endocarditis due to coagulase-negative staphylo-cocci. Microbiologic, epidemiologic, and clinical con-siderations. Infect Dis Clin North Am 1993; 7: 81–96.Acknowledgement

13 Vaudaux PE, Yasuda H, Velazco MI, et al. Role ofhost and bacterial factors in modulating staphylococcaladhesion to implanted polymer surfaces. J Biomat ApplThe authors wish to acknowledge R. A. Proctor for1990; 5: 134–53.critical reading of this manuscript. This work has

14 Flock JI, Groman G, Jonsson K, et al. Cloning andbeen funded by the German Minister for Education expression of the gene for a fibronectin-binding proteinand Research (grant 01KI9750/9), by the Medical from Staphylococcus aureus. EMBO J 1987; 6: 2351–7.Faculty of the University of Muenster (grant # 15 Chugh TD, Burns GJ, Shuhaiber HJ, Bahr GM. Ad-HE119840), and by Deutsche Forschungsgemein- herence of Staphylococcus epidermidis to fibrin plateletschaft, Collaborative Research Center 492, project B9. clots in vitro mediated by lipoteichoic acid. Infect Immun

1990; 58: 315–9.16 Aly R, Levit S. Adherence of Staphylococcus aureus to

squamous epithelium: role of fibronectin and teichoicacid. Rev Infect Dis 1987; 9 (Suppl 4): S341–50.

References 17 Hussain M, Peters G, Chhatwal GS, Herrmann M. Alithium chloride-extracted, broad-spectrum-adhesive42-kilodalton protein of Staphylococcus epidermidis is

1 Heilmann C, Hussain M, Peters G, Gotz F. Evidence ornithine carbamoyltransferase. Infect Immun 1999; 67:for autolysin-mediated primary attachment of Staphy- 6688–90.lococcus epidermidis to a polystyrene surface. Mol Mi- 18 Hussain M, Wilcox MH, White PJ. The slime ofcrobiol 1997; 24: 1013–24. coagulase-negative staphylococci: Biochemistry and

2 Tojo M, Yamashita DA, Goldman DA, Pier GD. Isolation relation to adherence. FEMS Microbiol Rev 1993; 104:and characterization of a capsular polysaccharide ad- 191–208.hesin from Staphylococcus epidermidis. J Infect Dis 1988; 19 Hussain M, Hastings JG, White JG. Comparison of cell-157: 713–22. wall teichoic acid with high-molecular-weight extra-3 Heilmann C, Schweitzer O, Gerke C, Vanittanakom N, cellular slime material from Staphylococcus epidermidis.Mack D, Gotz F. Molecular basis of intercellular ad- J Med Microbiol 1992; 37: 368–75.hesion in the biofilm-forming Staphylococcus epidermidis. 20 Green RJ, Davies MC, Roberts CJ, Tendler SJB. Com-Mol Microbiol 1996; 20: 1083–91. petitive protein adsorption as observed by surface4 Mack D, Fischer W, Krokotsch A, et al. The intercellular plasmon resonance. Biomaterials 1998; 20: 385–91.adhesin involved in biofilm accumulation of Staphy-

21 Vaudaux PE, Pittet D, Haeberli A, et al. Host factorslococcus epidermidis is a linear �-1,6-linked glukosamino-selectively increase staphylococcal adherence on in-glykan: Purification and structural analysis. J Bacteriolserted catheters: a role for fibronectin and fibrinogen/1996; 178: 175–83.fibrin. J Infect Dis 1989; 160: 865–75.5 McKenney D, Hubner J, Muller E, Wang Y, Goldmann

22 Vaudaux PE, Francois P, Proctor RA, et al. Use of ad-DA, Pier GB. The ica locus of Staphylococcus epidermidishesion-defective mutants of Staphylococcus aureus toencodes production of the capsular polysaccharide/define the role of specific plasma proteins in promotingadhesin. Infect Immun 1998; 66: 4711–20.bacterial adhesion to canine arteriovenous shunts. Infect6 Hussain M, Herrmann M, von Eiff C, Perdreau-Immun 1995; 63: 585–90.Remington F, Peters G. A 140-kilodalton extracellular

23 Herrmann M, Vaudaux PE, Pittet D, et al. Fibronectin,protein is essential for the accumulation of Staphy-fibrinogen, and laminin act as mediators of adherencelococcus epidermidis strains on surfaces. Infect Immunof clinical staphylococcal isolates to foreign material. J1997; 65: 519–24.Infect Dis 1988; 158: 693–701.7 Younger JJ, Christensen GD, Bartley DL, Simmons JCH,

24 Baldassari L, Donelli G, Gelosia A, Simpson AW, Chris-Barrett FF. Coagulase-negative staphylococci isolatedtensen GD. Expression of slime interferes with in vitrofrom cerebrospinal fluid shunts: importance of slimedetection of host protein receptors of Staphylococcusproduction, species identification, and shunt removalepidermidis. Infect Immun 1997; 65: 1522–6.to clinical outcome. J Infect Dis 1987; 156: 548–54.

25 Rupp ME, Ulphani JS, Fey PD, Bartscht K, Mack D.8 Ziebuhr W, Heilmann C, Gotz F, et al. Detection ofCharacterization of the importance of polysaccharidethe intercellular adhesion gene cluster (ica) and phaseintercellular adhesin/hemagglutinin of Staphylococcusvariation in Staphylococcus epidermidis blood cultureepidermidis in the pathogenesis of biomaterial-basedstrains and mucosal isolates. Infect Immun 1997; 65:infection in a mouse foreign body infection model.890–6.Infect Immun 1999; 67: 2627–32.9 Rupp ME, Archer GL. Coagulase-negative staphylo-

26 Galliani S, Viot M, Cremieux A, Van der Auwera P.cocci: Pathogens associated with medical progress. ClinEarly adhesion of bactermic strains of StaphylococcusInfect Dis 1994; 19: 231–45.epidermidis to polystyrene: influence of hydrophobicity,10 Refsahl K, Andersen BM. Clinically significantslime production, plasma, albumin, fibrinogen, andcoagulase-negative staphylococci: identification and

resistance patterns. J Hosp Infect 1992; 22: 19–31. fibronectin. J Lab Clin Med 1994; 123: 685–92.

270 M. Hussain et al.

27 Dunne WM, Burd EM. Fibronectin and proteolytic frag- 39 Harlow E, Lane D. Antibodies. A Laboratory Manual. Coldments of fibronectin interfere with the adhesion of Spring Harbor, NY: Cold Spring Harbor Laboratory,Staphylococcus epidermidis to plastic. J Appl Bacteriol 1993; 1988: 620–1.74: 411–6. 40 Naidu AS, Paulsson M, Wadstrom T. Particle ag-

28 Muller E, Takeda S, Goldman DA, Pier GB. Blood glutination assays for rapid detection of fibronectin,proteins do not promote adherence of coagulase-neg- fibrinogen, and collagen receptors on Staphylococcusative staphylococci to biomaterials. Infect Immun 1991; aureus. J Clin Microbiol 1988; 26: 1549–54.59: 3323–6. 41 Huff E, Oxley H, Silverman CS. Density gradient pat-

29 Paulsson M, Kober M, Freij LC, Stollenwerk M, Wesslen terns of Staphylococcus areus cells and cell walls duringB, Ljungh A. Adhesion of staphylococci to chemically growth and mechanical disruption. J Bacteriol 1964; 88:modified and native polymers, and the influence of 1115–62.preadsorbed fibronectin, vitronectin, and fibrinogen. 42 Yoshida AC, Heden CG, Cedergren B, Edebo L. ABiomaterials 1993; 14: 845–53. method for the preparation of undigested bacterial cell

30 Brash JL. Hydrophobic polymer surface and their inter- walls. J Microbiol Technol Eng 1961; 3: 151–9.actions with blood. Ann NY Acad Sci 1977; 283: 356–71. 43 Archibald AR, Baddiley J, Shaukat GA. The glycerol

31 Hasty DL, Courtney HS. Group A streptococcal teichoic acid from walls of Staphylococcus epidermidis 12.adhesion. All of the theories are correct. Adv Exp Med Biochem J 1968; 110: 583–8.Biol 1996; 408: 81–94. 44 Hussain M, Hastings JGM, White PJ. Isolation and

32 Hasty DL, Ofek I, Courtney HS, Doyle RJ. Multiple composition of the extracellular slime made by co-adhesins of streptococci. Infect Immun 1992; 60: 2147–52. agulase-negative staphylococci in a chemically defined

33 Proctor RA. The staphylococcal fibronectin receptor: medium. J Infect Dis 1991; 163: 534–41.Evidence for its importance in invasive infections. Rev 45 Dubois M, Gilles KA, Hamilton JK, Rabers PA, SmithInfect Dis 1987; 9: S335–40. F. Colorimetric method for determination of sugars and34 Wagner B, Wagner M, Ryc M, Bicova R. Ultrastructural related substances. Anal Chem 1956; 28: 350–6.localization of lipoteichoic acid on group A streptococci. 46 Ames B. Assay of inorganic phosphorus, total phos-In: Kimura Y, Kotani S, Shiokwa S, Eds. Recent advances phorus and phosphates. Meth Enzymol 1966; 8: 115–8.in streptococci and streptococcal disease. Bracknell: Red-

47 Block RJ, Durrum EL, Zweig G (Eds). A manual of paperbooks Ltd., 1985: 192–4.chromatography and paper electrophoresis, 2nd edn. New35 Ofek I, Simpson WA, Beachey EH. Formation of mo-York: Academic Press, 1958: 171–214.lecular complexes between structurally defined M pro-

48 Dawson RMC, Elliot DC, Elliot EH, Jones KM (Eds).tein and acylated and deacylated lipoteichoic acid ofMethods for detection of biochemical compounds onStreptococcus pyogenes. J Bacteriol 1982; 149: 426–33.paper and thin layer chromatograms, with some notes36 Peters G, Locci R, Pulverer G. Adherence and growthon separation. Data for Biochemical Research. Oxford:of coagulase-negative staphylococci on surfaces of intra-Clarendon Press, 1969: 509–91.venous catheters. J Infect Dis 1982; 146: 479–82.

49 Carroll JJ, Smith N, Babson AL. A colorimetric serum37 Vaudaux PE, Waldvogel FA, Morgenthaler JJ,glucose determination using hexokinase and glucose-6-Nydegger UE. Adsorption of fibronectin onto poly-phosphate dehydrogenase. Biochem Med 1971; 4: 171–80.methylmethacrylate and promotion of Staphylococcus

50 Levvy GA, McAllan A. The N-acetylation and es-aureus adherence. Infect Immun 1984; 45: 768–74.timation of hexosamines. Biochem J 1959; 73: 127–32.38 Kessler SW. Use of protein-A bearing staphylococci for

51 Glantz SA. Biostatistics. New York: McGraw Hill, 2000:the immunoprecipitation and isolation of antigen fromcells. Meth Enzymol 1981; 73: 442–58. 473.