Puri¢cation and characterization of a surface-binding protein fromLactobacillus fermentum RC-14 that inhibits adhesion of

Enterococcus faecalis 1131

Christine Heinemann a;2, Johan E.T. van Hylckama Vlieg b, Dick B. Janssen b,Henk J. Busscher c;1, Henny C. van der Mei c, Gregor Reid a;b;2;*

a Department of Microbiology and Immunology, University of Western Ontario, Health Sciences Centre, London, Ont., Canada, N6A 5C1b Department of Biochemistry, University of Groningen, Nijenborgh 4, 9747 AG Groningen, The Netherlands

c Department of Biomedical Engineering, University of Groningen, Bloemsingel 10, 9712 KZ Groningen, The Netherlands

Received 4 April 2000; received in revised form 28 June 2000; accepted 6 July 2000

Abstract

Lactobacilli have been shown to be important in the maintenance of the healthy urogenital flora. One strain, Lactobacillus fermentum RC-14, releases surface-active components which can inhibit adhesion of uropathogenic bacteria. Using a quantitative method for determininginhibition of adhesion, a protein with high anti-adhesive properties against Enterococcus faecalis 1131 was purified. The N-terminalsequence of the 29-kDa protein was identical to that of a collagen-binding protein from Lactobacillus reuteri NCIB 11951, and exhibitedclose homology with a basic surface protein from L. fermentum BR11. The results suggest that this anti-adhesive cell surface protein ofLactobacillus could protect against uropathogens by preventing their adhesion. ß 2000 Published by Elsevier Science B.V. on behalf ofthe Federation of European Microbiological Societies.

Keywords: Lactobacillus ; Enterococcus ; Bacterial adhesion; Collagen

1. Introduction

Lactobacilli have been implicated for many years ascontributing to the prevention of intestinal and urinarytract infections [1,2]. Several mechanisms have been pro-posed to be involved including competitive exclusion anddisplacement of uropathogens [3], production of hydrogenperoxide, lactic acid and growth inhibitors [4,5], and therelease of biosurfactants [6]. The surface-active compo-nents of Lactobacillus fermentum RC-14 are believed tobe powerful defense agents because they inhibit the adhe-sion of a wide variety of uropathogens [7]. The amount ofmaterial released by one adhering Lactobacillus su¤ces tocoat a substratum surface area several times its geometri-cal area [8], rendering unfavorable conditions for uropath-

ogens to adhere. Velraeds et al. [6] demonstrated that of15 Lactobacillus strains tested, all released surface-activecomponents when grown to their mid-exponential and sta-tionary growth phases. The isolated components couldreduce the surface tension of water and were found tobe primarily proteinaceous in nature, based on infraredabsorption and X-ray photoelectron spectroscopy [9].The aim of the work described in this paper was to isolateand identify key components of the biosurfactant mixturereleased by L. fermentum RC-14 and determine their abil-ity to interfere with Enterococcus faecalis 1131 adhesion.

2. Materials and methods

2.1. Strains and culture conditions

Lactobacillus fermentum RC-14, an isolate from the uro-genital tract of a healthy woman [10], was stored in De-Man-Rogosa-Sharpe (MRS) broth (Merck, Germany)containing 7% (v/v) dimethyl sulfoxide at 320³C. Foreach experiment, the bacteria were thawed, streaked on

0378-1097 / 00 / $20.00 ß 2000 Published by Elsevier Science B.V. on behalf of the Federation of European Microbiological Societies.PII: S 0 3 7 8 - 1 0 9 7 ( 0 0 ) 0 0 3 3 1 - 1

* Corresponding author. Tel. : +1 (519) 646-6000, ext. 65256;Fax: +1 (519) 646-6110; E-mail : gregor@julian.uwo.ca

1 Also a¤liated with University of Western Ontario.2 Also a¤liated with Lawson Research Institute, 268 Grosvenor Street,

London, Ont. N6A 4V2, Canada.

FEMSLE 9555 10-8-00

FEMS Microbiology Letters 190 (2000) 177^180

www.fems-microbiology.org

MRS agar plates, and incubated at 37³C in an atmospherecontaining 5% CO2 for further culturing.

The uropathogen Enterococcus faecalis 1131 was storedin brain heart infusion (BHI) broth (Oxoid, UK) contain-ing 7% (v/v) dimethyl sulfoxide at 320³C. From frozenstocks, blood agar plates were inoculated and incubatedaerobically at 37³C to obtain cultures. Precultures wereprepared by inoculating 5 ml of BHI broth and subse-quent aerobic overnight incubation at 37³C.

2.2. Biosurfactant isolation

The production of the biosurfactant was based on themethod by Velraeds et al. [6]. A starter culture of L. fer-mentum RC-14 was grown at 37³C overnight in 30 mlMRS medium in the presence of 5% CO2. An amountof 1200 ml of MRS broth was inoculated with the over-night culture and incubated for 18 h. The bacteria wereharvested by centrifugation (10 000Ug, 5 min, 10³C),washed twice in demineralized water and resuspended in200 ml of PBS (10 mM KH2PO4/K2HPO4, 150 mM NaCl,pH 7.0). To obtain biosurfactant release, the bacteria werekept at room temperature for 2 h with gentle stirring.After centrifugation (10 000Ug, 10 min, 10³C) the super-natant containing the biosurfactant was ¢ltered (0.22-Wm,Millipore) and dialyzed against demineralized water at 4³C(Spectrapor membrane tube, MWCO, 60 00^8 000, Spec-trum Medical Industries, California). Finally, the dialysatewas freeze-dried and stored at 320³C.

2.3. Anti-adhesion assay

The anti-adhesive activity against E. faecalis 1131 wasquanti¢ed using a previously reported adhesion assay [11].Brie£y, the wells of tissue culture treated, polystyrene mi-crotiter plates (Corning Glassworks, Corning, NY) weretreated with 200 Wl of the solution to be tested for anti-adhesive activity. The plate was incubated for 18 h at 4³Cand subsequently washed twice with PBS. Control wellscontained bu¡er only. E. faecalis 1131 was grown at 37³Covernight under aerobic conditions. An amount of 200 Wlof a washed suspension of E. faecalis 1131 (3U108 bacte-ria ml31 in PBS) was added and incubated in the wells for4 h at 4³C. Unattached organisms were removed by wash-ing the wells three times with PBS. The adherent bacteriawere stained with crystal violet and optical density read-ings taken at 595 nm. The change in adherence was usedto calculate the percentage inhibition of adhesion. Fromvarious dilutions of the crude biosurfactant mixture and acalibration curve, the amount of material that caused a50% reduction of adhesion was calculated. This was de-¢ned as one inhibitory unit.

2.4. Size exclusion chromatography

For puri¢cation, the crude biosurfactant was dissolved

in PBS (3 mg ml31) and loaded onto a Sephacryl S-200 gel¢ltration column (700 ml, Pharmacia Biotech) previouslyequilibrated in PBS (pH 7.0). Elution was carried out withPBS (£ow rate, 30 ml h31 ; fractions, 10 ml). The proteincontent of the eluted fractions was determined with Coo-massie brilliant blue using bovine serum albumin as thestandard [12]. All fractions obtained from the various pu-ri¢cation steps were also tested for anti-adhesive activityusing the assay described above.

2.5. Ion exchange chromatography

For further puri¢cation, fractions 44^49 were pooledand dialyzed overnight against TMA bu¡er (10 mMTris^HCl, l mM L-mercaptoethanol, 3 mM NaN3, pH8.0) at 4³C. The dialysate was subsequently applied to aResource Q anion-exchange column (6 ml, Pharmacia Bio-tech) connected to a LCC500 type FPLC system (Pharma-cia Biotech). Bound protein was eluted with a 140-ml lin-ear gradient from 0 to 0.45 M NaCl in TMA bu¡er (£owrate, 5 ml min31 ; fractions, 2 ml).

2.6. SDS^PAGE and electroblotting

Sodium dodecyl sulfate polyacrylamide electrophoresis(SDS^PAGE) was performed as previously described [13].The gels were either stained with Coomassie brilliant blueR-250 or the protein was transferred to a PolyvinylideneDi£uoride (PVDF) membrane (Mini ProBlot1, AppliedBiosystems) by electroblotting. Brie£y, the unstained gelwas soaked for 5 min in transfer bu¡er (25 mM Tris^HCl/192 mM glycine bu¡er, pH 8.3 containing 15% methanol)before blotting. A PVDF membrane was pre-wetted with100% methanol and transferred to the blotting bu¡er.After electroblotting (200 mA, 2 h), the ¢lter was washedin distilled water and stained with Coomassie brilliant blueR-250. The blot was partially destained with 50% metha-nol.

2.7. N-terminal sequencing

Following electroblotting, N-terminal sequencing wasperformed with an automated sequencer (Model 494 Pro-cise/Model 477A, Applied Biosystems) by Eurosequence(Groningen, The Netherlands).

3. Results and discussion

An anti-adhesion component, which appeared to be a29-kDa protein, was puri¢ed from the crude biosurfactantmixture in two steps. The ¢rst step involved size exclusionchromatography and most of the anti-adhesive activity co-eluted with the bulk of the protein, except for some latereluting fractions which had the highest speci¢c activity(Fig. 1). Starting with 45 mg of crude material, containing

FEMSLE 9555 10-8-00

C. Heinemann et al. / FEMS Microbiology Letters 190 (2000) 177^180178

694 U mg protein31, 4.1 mg protein was recovered with aspeci¢c activity of 1290 U mg protein31. Further puri¢ca-tion using ion exchange chromatography yielded a singlepeak of activity with a speci¢c activity of 1716 U mgprotein31, although the yield was relatively low (1.8%),probably due to absorption losses (results not shown).SDS^PAGE correlated the activity to a protein with anapparent molecular mass of 29 kDa (Fig. 2). The N-ter-minal sequence of the puri¢ed protein was found to beASSAVNSELVHKGELTIGLE. Database searches re-vealed 100% identity of the obtained sequence to a colla-gen-binding protein from L. reuteri NCIB 11951 [14], and90% similarity to a basic surface protein from L. fermen-tum BR11 [15]. The collagen-binding protein of L. reuteriNCIB 11951 has been puri¢ed by a¤nity chromatographyon collagen-Sepharose by Aleljung et al. [16]. Its biologicalfunction is still unknown, but it is thought to function asan adhesin [14]. The N-terminal sequence of the anti-ad-hesion protein was also similar to the sequence of a basicsurface protein from L. fermentum BR11. This 32-kDaprotein, which is also thought to function as an adhesin,was puri¢ed by removing it from the bacterial cell surfaceby extraction with 5 M LiCl [15]. The authors suggest thatthe protein is anchored to the cell surface by electrostaticinteractions with acidic groups since it could easily beremoved from the cell surface with acidic bu¡ers. [15] Inrecent years, many microorganisms have been shown tomediate their adhesion to host tissues via binding to extra-cellular matrix proteins, such as ¢bronectin, laminin andcollagen [17]. Several studies have shown that lactobacillialso have the ability to bind to collagen and ¢bronectin[18,19] which suggests that binding to extracellular matrixproteins might be an important feature in their ability tocolonize the intestinal and urogenital tract. Interestingly,we isolated the same protein as Roos et al. [14], but basedon a di¡erent activity, namely the ability to inhibit theadhesion of uropathogenic enterococci. These results sug-

gest that binding of lactobacilli to extracellular matrixproteins is not only involved in the adhesion of these bac-teria to host tissues, but that release of these proteins alsoplays an important role in the ability of lactobacilli toprotect the host against invading pathogens.

Acknowledgements

This work was supported, in part, by the Natural Sci-ences and Engineering Research Council of Canada.

References

[1] Reid, G., Bruce, A.W., McGroarty, J.A., Cheng, K.-J. and Coster-ton, J.W. (1990) Is there a role for lactobacilli in prevention of uro-genital and intestinal tract infections? Clin. Microbiol. Rev. 3, 335^344.

[2] Blomberg, L., Hendriksson, A. and Conway, P.L. (1993) Inhibitionof adhesion of Escherichia coli K88 to ileal mucus by Lactobacillusspp.. Appl. Env. Microbiol. 59, 34^39.

[3] Millsap, K., Reid, G., van der Mei, H.C. and Busscher, H.J. (1994)Adhesion and displacement of Enterococcus faecalis by Lactobacillusand Streptococcus sp. from hydrophobic and hydrophilic substrata asstudied in a parallel plate £ow chamber. Appl. Env. Microbiol. 60,1867^1874.

[4] Klaenhammer, T.R. (1988) Bacteriocins of lactic acid bacteria. Bio-chemie 70, 337^349.

[5] McGroarty, J.A. and Reid, G. (1988) Detection of a Lactobacillussubstance that inhibits Escherichia coli. Can. J. Microbiol. 34, 974^978.

[6] Velraeds, M.M.C., van der Mei, H.C., Reid, G. and Busscher, H.J.(1996) Inhibition of initial adhesion of uropathogenic Enterococcusfaecalis by biosurfactants from Lactobacillus isolates. Appl. Env. Mi-crobiol. 62, 1958^1963.

Fig. 2. SDS^PAGE of puri¢ed anti-adhesion protein present in a biosur-factant mixture of L. fermentum RC-14. Lane 1, size markers; lane 2,crude biosurfactant mixture; lane 3, fractions 44^49 from gel ¢ltration;lane 4, puri¢ed anti-adhesion protein.

Fig. 1. Sephacryl S-200 elution pro¢le of components from L. fermen-tum RC-14 with anti adhesion-activity. Both activity and speci¢c activityof eluting fractions are showing. a, activity; b, speci¢c activity againstE. faecalis 1131 adhesion; dashed line, protein concentration.

FEMSLE 9555 10-8-00

C. Heinemann et al. / FEMS Microbiology Letters 190 (2000) 177^180 179

[7] Velraeds, M.M.C., van de Belt-Gritter, B., van der Mei, H.C., Reid,G. and Busscher, H.J. (1998) Interference in initial adhesion of uro-pathogenic bacteria and yeasts to silicone rubber by a Lactobacillusacidophilus biosurfactant. J. Med. Microbiol. 47, 1^5.

[8] Busscher, H.J. and van der Mei, H.C. (1997) Physico-chemical inter-actions in initial microbial adhesion and relevance for bio¢lm forma-tion. Adv. Dent. Res. 11, 24^32.

[9] Velraeds, M.M.C., van der Mei, H.C., Reid, G. and Busscher, H.J.(1996) Physicochemical and biochemical characterization of biosur-factants released by Lactobacillus strains. Col. Surf. B: Biointerfaces8, 51^61.

[10] Reid, G., Cuperus, P.L., Bruce, A.W., Tomeczek, L., van der Mei,H.C., Khoury, A.H. and Busscher, H.J. (1992) Comparison of con-tact angles and adhesion to hexadecane of urogenital, dairy andpoultry lactobacilli : e¡ect of serial culture passages. Appl. Env. Mi-crobiol. 58, 1549^1553.

[11] Reid, G., Heinemann, C., Velraeds, M.M.C., van der Mei, H.C. andBusscher, H.J. (1999) Biosurfactants produced by Lactobacillus.Methods Enzymol. 310, 426^432.

[12] Bradford, M.M. (1976) A rapid and sensitive method for the quanti-tation of microgram quantities of protein utilizing the principle ofprotein-dye binding. Anal. Biochem. 72, 248^254.

[13] Laemmli, U.K. and Favre, M. (1973) Maturation of the head of

bacteriophage T4. I. DNA packaging events. J. Mol. Biol. 80, 575^599.

[14] Roos, S., Aleljung, P., Robert, N., Lee, B., Wadstrom, T., Lindberg,M. and Jonsson, H. (1996) A collagen binding protein from Lacto-bacillus reuteri is part of an ABC transportersystem? FEMS Micro-biol. Lett. 144, 33^38.

[15] Turner, M.S., Timms, P., Hafner, L.M. and Gi¡ard, P.M. (1997)Identi¢cation and characterization of a basic cell surface-located pro-tein from Lactobacillus fermentum BR11. J. Bacteriol. 179, 3310^3316.

[16] Aleljung, P., Shen, W., Rozalska, B., Hellman, U., Ljungh, A.and Wadstrom, T. (1994) Puri¢cation of collagen-binding proteinsfrom Lactobacillus reuteri NCIB 11951. Curr. Microbiol. 28, 231^236.

[17] Patti, J.M. and Hook, M. (1994) Microbial adhesins recognizingextracellular matrix macromolecules. Curr. Opin. Cell Biol. 6, 752^758.

[18] Nagy, E., Froman, G. and Mardh, P.-A. (1992) Fibronectin bindingof Lactobacillus species isolated from women with and without bac-terial vaginosis. J. Med. Microbiol. 37, 38^42.

[19] Aleljung, P., Paulsson, M., Emody, L., Andersson, M., Naidu, A.S.and Wadstrom, T. (1991) Collagen binding by lactobacilli. Curr.Microbiol. 23, 33^38.

FEMSLE 9555 10-8-00

C. Heinemann et al. / FEMS Microbiology Letters 190 (2000) 177^180180