Vol. 56, No. 5INFECTION AND IMMUNITY, May 1988, p. 1314-13190019-9567/88/051314-06$02.00/0Copyright C 1988, American Society for Microbiology

Identification of a Galactose-Binding Lectin onFusobacterium nucleaturm FN-2

PATRICIA A. MURRAY,* DAVID G. KERN, AND JAMES R. WINKLERDepartment of Stomatology, Division of Periodontology, HSW 661, University of California San Francisco,

San Francisco, California 94143-0515

Received 17 August 1987/Accepted 3 February 1988

A previous study has suggested that Fusobacterium nucleatum FN-2 contains a galactose-binding protein(lectin) on the cell surface (P. A. Murray, V. Matarese, C. I. Hoover, and J. R. Winkler, FEMS Microbiol.Lett. 40:123-127, 1987). In the present study, the molecular specificity and size of this lectin were investigatedby several techniques. Whole-cell affinity chromatography with asialofetuin covalently coupled to Sepharose6MB demonstrated that 81% of 3H-labeled F. nucleatum were specifically eluted by 0.5 M galactose. Specificbinding was calcium dependent and did not occur in the presence of calcium chelators. Binding was inhibitedby preincubation with galactose. Agglutination of human parotid saliva by F. nucleatum was also inhibited bygalactose and its structural analogs. Inhibition by lactose was 2 times that of galactose, inhibition byp-aminophenyl galactosides was 4 times that of galactose, and inhibition by asialoglycopeptides was 100 timesthat of galactose. Similar inhibition results were obtained for hemagglutination of neuraminidase-treatederythrocytes. These findings suggest that the binding specificity of F. nucleatum FN-2 is more complex thansimply the recognition of the monosaccharide galactose. This is consistent with the concept that lectinsconsidered identical in terms of monosaccharide specificity can recognize fine differences in more complexstructures. To identify the specific bacterial component(s) involved in galactose recognition, proteins of F.nucleatum FN-2 were separated on a 4 to 11% gradient sodium dodecyl sulfate slab gel, transferred tonitrocellulose paper to renature bacterial binding sites, and then incubated with '251-labeled asialofetuin.Autoradiographs of the nitrocellulose revealed a band at a range of Mr 300,000 to 330,000 which was notpresent when the blots were preincubated with galactose. These data support the concept that F. nuckeatumFN-2 possesses a lectin that recognizes galactose and galactose-containing substrates.

The adherence of bacteria to host tissues is a specificinteraction mediated by bacterial surface molecules (adhe-sins) which combine with complementary structures on thehost cell surface (receptors) (2, 25). Several oral bacteriahave been shown to contain surface adhesins, or bindingproteins, and a lectinlike mechanism of adherence has beenproposed on the basis of inhibition of attachment by specificsugars (8-10, 21, 22, 28). In gram-negative bacteria, theseadhesins (lectins) include fimbriae and certain outer mem-brane proteins (13). Although it is recognized that surfacecomponents of oral bacteria participate in adherence andclearance phenomena, their properties remain largely ob-scure. The further identification and characterization ofthese binding proteins, as well as their receptor sites, willprovide a more complete understanding of their role inhost-parasite interactions.

Considerable data indicate that the oligosaccharide moi-eties of salivary glycoproteins interact specifically withbacterial ligands and subsequently play a major role inadherence and clearance mechanisms in the oral cavity (3,17, 18, 20-22, 29). Galactose (Gal)-binding lectins have beenidentified on several species of oral bacteria (4, 6, 10, 14, 22,23). Yet, although these bacteria possess ligands that inter-act with ,-galactosides, they differ in their properties ofadherence and colonization of oral tissues (11). It appearsthat lectins considered identical in terms of monosaccharidespecificity are able to recognize subtle structural differences(5). This is exemplified in the studies of Murray et al. (20,21), who identified a sialic acid-binding lectin on Streptococ-cus sanguis and Streptococcus mitis, with the greatest

* Corresponding author.

specificity towards a NeuAc(a2,3)Gal(,1,4)GalNAc se-quence (NeuAc, N-acetylneuraminic acid; GalNAc, N-acetyl-galactosamine). Furthermore, Firon et al. (8, 9) elegantlydemonstrated that bacterial lectins in the form of type Ifimbriae on the various species of enterobacteria, typicallyclassified as mannose specific, exhibited fine differences inactual sugar specificity.

In earlier studies, we described one such Gal-bindinglectin on Fusobacterium nucleatum FN-2 (22). Very little isknown about the detailed carbohydrate specificity of thereceptor to which this adhesin binds (i.e., the recognitionsite). The purpose of this investigation was to analyze thesugar specificity patterns of F. nucleatum by using a whole-cell affinity chromatography technique, hemagglutinationassays, and salivary agglutination assays. In addition, wehave identified the bacterial lectin on the basis of size bysodium dodecyl sulfate-polyacrylamide gel electrophoresis(SDS-PAGE) and Western blotting.

MATERIALS AND METHODSMaterials. Sialyllactose was purchased from Boehringer

Mannheimn Biochemicals. Fetuin, enzymes, sugars, NonidetP-40 (NP-40), and most reagents were obtained from SigmaChemical Co. Electrophoresis reagents were from Bio-RadLaboratories. Tryptic soy broth and brucella agar were fromDifco Laboratories. CNBr-Sepharose 6MB was purchasedfrom Pharmacia Fine Chemicals. Permlastic impression ma-terial was from Kerr. [2,8-3H]adenine and ['25I]sodium io-dide were obtained from New England Nuclear Corp. Mi-crotiter plates were purchased from Linbro, X-ray film wasfrom Eastman Kodak Co., and Quanta III intensifyingscreens were from Du Pont Co. The nitrocellulose mem-

1314

on February 9, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

GALACTOSE-BINDING LECTIN ON F. NUCLEATUM FN-2

brane (0.2 ,um pore size) was purchased from Schleicher &Schuell, Inc.

Bacteria and culture conditions. F. nucleatum FN-2 wasthe kind gift of S. Syed, University of Michigan, Ann Arbor.This strain was stored frozen in tryptic soy broth containing20% glycerol, 0.25% yeast extract, 2.5 ,ug of hemin per ml,2.5 ,ug of menadione per ml, and 0.01% dithiothreitol. Astock culture was maintained anaerobically (85% N2-10%H2-5% CO2) on laked rabbit blood brucella agar with weeklytransfers to new plates. Cells were grown anaerobically at37°C to late log phase of growth in tryptic soy brothcontaining 0.25% yeast extract, 2.5 ,ug of hemin per ml, 2.5,ug of menadione per ml, and 0.01% dithiothreitol. Radiola-beled bacteria were prepared by growing cells in the pres-ence of [3H]adenine (10 ,XCi/ml). The typical specific activityachieved was 3,000 to 5,000 cpm/106 cells. Cells wereharvested by centrifugation (8,000 x g for 15 min at 4°C) andwashed either in PBS (10 mM sodium phosphate, 154 mMNaCl [pH 7.2], and 0.02% sodium azide) or in buffered KCI(5 mM KC1 containing 2 mM sodium phosphate and 1 mMCaCl2 [pH 6.0]).

Whole-cell affinity chromatography. Asialofetuin was co-valently coupled to CNBr-activated Sepharose 6MB using 6mg of ligand per ml of gel as previously described (22). Thederivatized macrobeads (1.5 to 2 ml) were equilibrated inPBS in columns (0.5 by 5 cm) containing an 80-,um-pore-sizenylon mesh filter, which allows bacterial cells to passthrough but retains matrix beads. Radiolabeled bacteria (109cells) were suspended in 0.2% bovine serum albumin (BSA)in PBS (pH 7.2), carefully layered onto the affinity matrix,and incubated for 1 h at 4°C. In some experiments, to test theeffect of calcium on bacterial binding, calcium or calciumchelators were added to the incubation buffer. After incuba-tion, the columns were adjusted to a flow rate of 1 ml/minand the unbound fraction was recovered by washing with 10volumes (2 ml each) of PBS containing 0.2% BSA. Bacteriaspecifically bound were eluted with 8 volumes of 0.5 M Galin PBS containing 0.2% BSA (pH 7.2). The nonspecificallybound fraction was eluted with 10 volumes of 0.1% NP-40 inPBS (pH 7.2). The counts per minute recovered in thevarious fractions was determined by scintillation spectros-copy. The percentage recovery was determined by dividingthe counts per minute bound by the counts per minuteadded; the percentage specifically bound was calculated bydividing the counts per minute specifically eluted by the totalcounts per minute recovered.

Saliva collection. Whole paraffin-stimulated saliva fromfour donors was collected on ice into a protease inhibitorcocktail (5 mM benzamidine, 10 mM EDTA, 10 p,M pepsta-tin, 2 mM phenylmethylsulfonyl fluoride [pH 7.0]). Thesaliva was then pooled and centrifuged at 11,000 x g for 30min at 4°C.Human parotid saliva was obtained by means of a modi-

fied Carlson-Crittenden apparatus. Human submandibular-sublingual saliva was collected by means of a mouthpiececustom-made from Permlastic impression material. Salivaryflow was stimulated by 2% citric acid applied to the lateralborders of the tongue at 30-s intervals. Samples of both typesof saliva were collected on ice into the protease inhibitorcocktail and were centrifuged at 11,000 x g for 30 min at 4°C.

Salivary agglutination and agglutination inhibition. Visualassays for salivary agglutination were performed in 96-wellround-bottom microtiter plates at room temperature. Fresh-ly collected saliva samples (whole, parotid, or submandibu-lar-sublingual) were serially diluted in PBS. Next, 25 ,ul ofone type of saliva was added to an equivalent volume of

bacterial suspension (108 to 109 cells per ml) in a checker-board fashion in microtiter plates. Plates were allowed to sitfor 1 h at room temperature, after which salivary agglutina-tion was scored from 0 to 4+, with 4+ being 100% aggluti-nation of bacteria and saliva. Agglutination scores of 4+were assigned when there was an opaque, scalloped blanketof cells on the bottom of the well, whereas a score of 0 wasassigned when there was a round, tight pellet of cells. Scoresbetween these two numbers were based on the appearanceof the cells relative to the two extremes.

Inhibition studies were conducted by incubating 25 pul ofserially diluted inhibitors (2 to 100 mM in PBS with pHadjusted to 7.2) with 25 p.l of F. nucleatum FN-2 (finalconcentration, 5 x 108 cells per ml) for 1 h at roomtemperature with gentle agitation. Saliva (50 p.l) was addedto each well, and the plates were again gently agitated for 1h. Plates were scored as above. The MIC was determined tobe the concentration of inhibitor required for 50% inhibitionof agglutination.

Hemagglutination and hemagglutination inhibition. Rabbiterythrocytes (RBC) were treated with neuraminidase (typeVI, Clostridium perfringens) in sodium acetate buffer (2 mMsodium acetate, 154 mM NaCl [pH 5.8]) for 1 h at 37°C toremove the terminal sialic acid residues and expose thepenultimate Gal. Cells were then washed three times in PBSand either used immediately or stored in an equal volume ofAlsever solution (University of California San FranciscoCell Culture Facility) at 4°C for no longer than 1 week.Removal of sialic acid was tested by hemagglutination oftreated and untreated cells with Limulin lectin at a concen-tration of 2 p.g/ml (1).Washed asialo-RBC were diluted with PBS at a 4%

(vol/vol) suspension. Visual assays for bacteria-mediatedhemagglutination were performed in 96-well round-bottommicrotiter plates at room temperature. Twofold serial dilu-tions of asialo-RBC (25 pI) were added to 25 p.1 of seriallydiluted bacterial suspensions in a checkerboard fashion.After the plates were allowed to sit for 1 h, hemagglutinationwas scored from 0 to 4+, with 4+ being 100% hemaggluti-nation of asialo-RBC and bacteria (6, 7, 21). A score of 4+was given to those wells in which the cells formed ascalloped blanket of cells at the bottom of the well. Thosescored as 0 exhibited a tight, round pellet of cells at thebottom of the well.For hemagglutination inhibition studies, 25 pI of serially

diluted inhibitors (2 to 100 mM in PBS with pH adjusted to7.2) were added to 25 pI of bacteria (final concentration, 5 x108 cells per ml) and were incubated for 1 h with gentleagitation. Next, 50 pI of asialo-RBC were added, and theplates were again gently agitated for 1 h. Plates were scored,and the MIC was defined as the concentration needed toinhibit hemagglutination by 50%.

Binding of '25I-labeled asialofetuin to SDS-PAGE replicas ofF. nucleatum on nitrocellulose. F. nucleatum FN-2 (2 x 108cells) were solubilized in 20 p,l of 0.125 M Tris hydrochloridecontaining 6 M urea, 4% SDS, 20% glycerol, 10% P-mercap-toethanol, and 0.01% bromophenol blue (pH 6.8) (loadingbuffer). The mixture was boiled for 10 min, and then proteinswere separated by 4 to 11% gradient SDS-PAGE (16, 24).The molecular weight standards used were as follows: Mr29,000, carbonic anhydrase; Mr 45,000, ovalbumin; Mr66,000, bovine plasma albumin; Mr 97,400, phosphorylase b;Mr 116,000, ,-galactosidase; Mr 205,000, myosin.The resulting gels were cut into two pieces. One half was

stained with 0.5% Coomassie blue in 45% methanol-10%acetic acid. The other half was transferred to nitrocellulose

1315VOL. 56, 1988

on February 9, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

1316 MURRAY ET AL.

by means of a semidry electroblotter (Satorius) for 1 h at 0.8mA/cm2 of gel (15, 30). The nitrocellulose blots were rinsedonce with TBS (0.154 M NaCl-0.02 M Tris hydrochloride[pH 8.0]) and then were soaked with TBS containing 5%BSA-0.1% Tween 20 for 1 h at 37°C to block nonspecificattachment of ligand to the nitrocellulose blots. After block-ing, the nitrocellulose blots were rinsed a total of five timesfor 6 min, each time with TBS containing 0.05% Tween 20.Asialofetuin was radiolabeled with [1251]sodium iodide (spe-cific activity, 17.4 Ci/mg) by the chloramine-T method (12).125I-asialofetuin was dissolved in TBS containing 10 mMCaCl2 (1.3 x 106 cpm); blots were incubated in this mixturewith gentle agitation for 1 h at 37°C and then were incubatedovernight at 4°C. In some cases, to evaluate the specificity ofbinding, the incubations were done in the presence of100-fold excess asialofetuin or 0.5 M Gal. The blots werethen washed five times for 6 min each time with 0.05%Tween 20 in TBS containing 10 mM CaCl2, dried in air, andexposed to X-ray film with an intensifying screen at -70°Cfor 24 h.

KEi'NTT T'QI.EnUL I a

Affinity chromatography of F. nucleatunSepharose 6MB. F. nucleatum FN-2 label(nine was applied to a Sepharose 6MB cc

a

E

0

b

0 1 0 20 30Fraction volume (ml)

30-

E 20-0

t pF0

0.5M Galf95

0 10 20 30Fraction volume (ml

FIG. 1. Elution profile of binding of F. niSepharose 6MB covalently coupled to asialofettrose 6MB alone (b). F. nucleatum (109 cells)column at room temperature and incubated for:herent fraction (17%) was eluted off the columrepresented by the first peak on the graph. Thesents bacteria specifically bound and eluted by 0.

remaining two peaks are nonspecifically bouneremoved by 0.1% NP-40 (NP40) (2%). (b) The n(running through the column with PBS represeadded (control). In this case, no bacteria were eluThe arrows indicate when the eluent was added

30-C

.C

*

0.5

Galactose concentration (mM)FIG. 2. Effect of Gal concentration on the binding of F. nuclea-

tum FN-2 to asialofetuin-Sepharose 6MB at room temperature.Bacteria were preincubated and eluted with various concentrationsof Gal. The values shown on the graph represent the percentage ofinhibition in comparison with 0.5 M Gal, which inhibited attachmentby 100%. The concentration of Gal required to inhibit specificbinding by 50% was 20 mM.

bound to asialofetuin (Fig. la) and, as a control, to aa on asialofetuin- Sepharose 6MB column alone (Fig. lb). Of the cells added toed with [3H]ade- the test column, 81% were specifically eluted from asialofe-)lumn covalently tuin with 0.5 M Gal (Fig. la). The unbound fraction (eluted

with PBS) represented 17%, whereas the nonspecificallybound fraction (eluted with NP-40) represented the remain-ing 2% recovered. In the control columns containing noligand (Fig. lb), 95% of the added cells remained unbound,and no bacteria were eluted with 0.5 M Gal.To investigate the potential role of calcium in mediating

the attachment of F. nucleatum FN-2 to asialofetuin, weexamined the effects of adding divalent cation chelators toour affinity chromatography system. When either 1 mM

NP40 EDTA or 1 mM ethylene glycol-bis(,-aminoethyl ether)-N,N,N',N'-tetraacetic acid (EGTA) was added to the 0.2%BSA-PBS buffer before the addition of cells to the affinity

___________^column, the percentages of cells requiring elution by Gal(specific binding) were significantly reduced, i.e., by 91 and

4 0 5 0 95%, respectively (P < 0.001, Student's t test). On the otherhand, when 10 mM Ca2+ was also added to the buffercontaining 1 mM EDTA or 1 mM EGTA, the percentages ofcells requiring Gal for elution returned to near control levels(93%). This reversal of inhibition could not be duplicated bythe addition of 1 mM Mg2+ or 1 mM Mn2+.

Effects of Gal concentration on binding of F. nucleatum toasialofetuin-Sepharose 6MB. F. nucleatum FN-2 was prein-cubated and eluted from affinity columns of asialofetuin-Sepharose 6MB with several different concentrations of Gal

NP40 (Fig. 2). As the concentration of Gal used to preincubate and'1' elute the bacteria was increased, the numbers of bacteria

that bound to the column decreased. A concentration of4 0 5 0 approximately 20 mM Gal was needed for 50% inhibition of

I) binding of the bacteria to the affinity column (Fig. 2).ucleatum FN-2 to Saliva-mediated agglutination of F. nucleatum. Serial dilu-in (a) or to Sepha- tions of saliva were added to microtiter plates containing F.were added to the nucleatum FN-2 in a checkerboard fashion to compare the1 h. (a) The nonad- agglutination titers for whole, ductal parotid, and ductalIn with PBS and is submandibular-sublingual saliva (Table 1). For all cell con-major peak repre- centrations of F. nucleatum FN-2 tested, parotid saliva

d fractions of cells proved to be the best agglutinator. At the optimal bacterial)nadherent fraction cell concentrations (5 x 108 cells per ml), parotid salivamnted 95% of cells agglutinated F. nucleatum FN-2 at titers 8-fold higher thanited with 0.5 M Gal. submandibular-sublingual saliva and 16-fold higher than1. whole saliva. These findings confirm that there are compo-

INFECT. IMMUN.

on February 9, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

GALACTOSE-BINDING LECTIN ON F. NUCLEATUM FN-2

TABLE 1. Agglutinating activity of human salivas withF. nucleatum FN-2

Titer'F. nucleatum

(no. of cells/ml) Whole Parotid Submandibular-saliva saliva sublingual saliva

1 X 108 64 1,024 1285 x 108 128 2,048 2561 x 109 128 1,024 128

a Titers represent the reciprocal dilution giving 50% agglutination. Theprotein concentrations were 0.70, 0.67, and 0.60 mg/ml for whole, parotid, andsubmandibular-sublingual saliva, respectively.

nents in saliva capable of interacting with F. nucleatumFN-2 and that parotid saliva contains the greatest concen-tration of specific molecules. Consequently, parotid salivawas chosen for use in the experiments studying the inhibitionof saliva-mediated agglutination of F. nucleatum FN-2.

Inhibition of human parotid saliva-mediated agglutinationof F. nucleatum. To examine the specificity of the adhesin(s)of F. nucleatum FN-2 in the agglutination by human parotidsaliva, we used an inhibition assay. Various sugars andglycopeptides were tested to determine their effectiveness ininhibiting agglutination (Table 2). Of the 29 inhibitors tested,5 were glycopeptides and 23 were sugars. The hierarchy ofinhibition was as follows: asialoglycopeptides > p-amino-phenyl galactopyranosides, sialyllactose, and N-acetyl-D-galactosamine > D-lactose and phenyl-p-D-thiogalactoside> NeuAc > D-galactose > fucOse > p-nitrophenyl-P-D-galactopyranoside, p-nitrophenyl-p-D-lactopyranoside, andphenyl-p-D-galactoside (Table 2). The best inhibitors wereasialoglycopeptides, showing 100-fold potency over Gal.However, there was no distinction seen between inhibitionby N-linked asialofetuin, 0-linked asialofetuin, and asialo-transferrin glycopeptides. N-Acetyl-D-galactosamine and p-aminophenyl galactopyranosides were second in the relative

TABLE 2. Inhibition of F. nucleatum-mediated salivaryagglutination and hemagglutination by specific

sugars and glycopeptidesa

MICbInhibitor

Parotid salivac RBC

Asialofetuin glycopeptides (N-linked) 0.2 (100.0) 0.3Asialofetuin glycopeptides (0-linked) 0.2 (100.0) 0.3Asialotransferrin glycopeptides 0.2 (100.0) 0.3p-Aminophenyl-a-galactopyranoside 5 (4.0) 10p-Aminophenyl-3-galactopyranoside 5 (4.0) 10Sialyllactose 5 (4.0) 5N-Acetyl-D-galactosamine 5 (4.0) 10D-Lactose 10 (2.0) 20Phenyl-,B-D-thiogalactoside 10 (2.0) 20NeuAc 15 (1.3) 20D-Galactose 20 (1.0) 30Fucose 30 (0.7) >1,000p-Nitrophenyl-o-D-galactopyranoside 50 (0.4) 100p-Nitrophenyl-f-D-lactopyranoside 50 (0.4) 100Phenyl-p-D-galactoside 50 (0.4) 100

a The following sugars were tested and found to have MICs greater than1,000: glucuronic acid, galacturonic acid, fetuin glycopeptides, transferringlycopeptides, D-glucose, D-sucrose, fucosylamine, fucose-l-phosphate,mannose, D-xylose, D-turanose, p-nitrophenyl-3-D-glucopyranoside, p-nitro-phenyl-,3-D-mannopyranoside, and N-acetyl-D-glucosamine.

b Concentration of inhibitor (mM) required to inhibit agglutination orhemagglutination by 50%.

c Numbers in parentheses indicate the inhibition potency relative to Gal asa standard of reference on a molar basis.

potency of inhibition (4 times that of Gal), with both the a-and r-derivatives of p-aminophenyl galactopyranosidesdemonstrating equal inhibition. When the substitutions onGal were examined and compared, the p-aminophenyl galac-topyranosides were 10-fold better inhibitors than the p-nitrophenyl-3-D-galactopyranosides. In addition, inhibitionby N-acetyl-D-galactosamine was fourfold that of Gal, andinhibition by lactose and phenyl-p-D-thiogalactoside wastwofold that of Gal.

F. nuckatum-mediated hemagglutination and its inhibition.Previous studies have demonstrated that many F. nucleatumstrains have the ability to hemagglutinate rabbit asialo-RBC(7). We conducted similar studies confirming this point (datanot shown). Therefore, we developed a hemagglutinationinhibition assay to investigate the specificity of the adhe-sin(s) of F. nucleatum FN-2. It appears that asialo-RBC aresuitable receptor analogs for F. nucleatum FN-2 (7, 19) andwe wanted to compare the sensitivity and specificity ofhemagglutination versus salivary agglutination. Asialo-RBCwere added to microtiter plates containing F. nucleatumFN-2 and the same serially diluted sugars as in the salivaryagglutination assay. The results of these experiments arealso shown in Table 2. Although the same hierarchy ofpotency of inhibitors was noted, hemagglutination proved tobe less sensitive than salivary agglutination. The one excep-tion was fucose, which was a good inhibitor of salivaryagglutination but had no effect on hemagglutination.

Binding of '25I-labeled asialofetuin to F. nucleatum electro-blotted onto nitrocellulose membrane. To identify the bacte-rial component(s) involved in recognition of galactosyl resi-dues, we separated proteins of F. nucleatum FN-2 bySDS-PAGE, electroblotted them to nitrocellulose, and thenincubated the blots with 1251I-asialofetuin in the presence andabsence of 0.5 M Gal. Figure 3 shows the stained gel (lane 1)and developed autoradiographs (lanes 2 and 3). The darkband in lane 2, which showed binding of 125I-asialofetuin tostrain FN-2, had an electrophoretic mobility similar to thatof the three high-molecular-weight bands at approximately300,000, 310,000, and 330,000 seen on the stained gel ofstrain FN-2 (lane 1). When the FN-2 nitrocellulose replicawas preincubated in 0.5 M Gal (lane 3), the band was notseen. Similar findings were obtained when the blots werepreincubated with a 100-fold excess of asialofetuin (data notshown). Results demonstrate that the binding of 1251I-asialo-fetuin to the 300,000 to 330,000 bands was inhibited by thepresence of excess Gal, suggesting that this high-molecular-weight bacterial complex might represent the Gal-bindingadhesin(s) of F. nucleatum FN-2.

DISCUSSIONDental diseases are initiated by the colonization of the

tooth surface by pathogenic bacteria (11). The study offactors influencing the attachment and distribution of organ-isms in the oral cavity is therefore of considerable impor-tance. In vivo the adherence of bacteria to the tooth surfaceinvariably occurs via the salivary pellicle. Yet, to combatadherence, saliva interacts directly with bacteria, facilitatingclearance by mastication, movement of the tongue andcheeks, and swallowing (11). It is now recognized that theoligosaccharide moieties of salivary glycoconjugates play apivotal role in oral clearance and adherence mechanisms (18,29). In many instances, interactions are mediated by surfacestructures present on the bacteria, involving the formation ofspecific protein-carbohydrate complexes (3-5, 17, 20-23,28). The molecular mechanisms by which these interactionstake place have been the subject of our investigations.

1317VOL. 56, 1988

on February 9, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

1318 MURRAY ET AL.

- - 205k

^ - 116k

4 < 97k

i.t

.. ..

1 2 3

. <66k

. 4̂5k

29k

FIG. 3. Binding of 125I-labeled asialofetuin to SDS-PAGE repli-cas of F. nucleatum FN-2. A 4 to 11% gradient SDS-PAGE of F.nucleatum FN-2 was electroblotted to nitrocellulose, incubated with1251-asialofetuin in the presence and absence of Gal, and autoradio-graphed. Lane 1, Gel of strain FN-2, stained with Coomassie blue;lane 2, autoradiograph of strain FN-2 after incubation with 1251I_asialofetuin; and lane 3, autoradiograph of strain FN-2 preincubatedwith 0.5 M Gal before incubation with 1251I-asialofetuin. The posi-tions of the molecular weight markers (see Materials and Methods)are shown to the right.

We report here that F. nucleatum FN-2 possesses asurface-associated lectin with specificities for galactosylresidues. Galactosyl residues are commonly found as termi-nal sugars on salivary glycoconjugates and appear to befunctional receptor sites for bacterial Gal-binding lectins.We used an RBC model system of adherence to characterizethe lectin activity of F. nucleatum because RBC possessreceptors that are similar, if not identical, to the receptors onthe normal target cells (6-9, 14, 19). However, the apparentsensitivity of the salivary agglutination assay was 1.5 to 2times greater than that of the hemagglutination assay (Table2). When select sugars were used as inhibitors, a similarpattern was seen, except for fucose, which inhibited salivaryagglutination but not hemagglutination. The reasons for thisremain unclear, and the role of fucose as an inhibitor iscurrently being investigated in our laboratory. The results ofinhibition studies suggest that the lectin responsible forbinding F. nucleatum to human parotid saliva is complex innature and is capable of discerning subtle differences instructure. Alternatively, it is possible that multiple lectinsrecognizing different sugar determinants are present on thisorganism.

It is of note that in addition to the Gal-containing deriva-tives, fucose, NeuAc, and sialyllactose [NeuAc(2,3) and(2,6)Gal(P1,4)Glc (Glc, glucose)] also demonstrated inhibi-tion (30, 15, and 5 mM, respectively). However, the sialogly-copeptides of fetuin and transferrin did not show inhibitionat any concentration. Consequently, the conflicting data aredifficult to interpret and the role of NeuAc as a receptorrecognition site remains unclear. One speculation is that theinternal Gal residues of more complex structures may some-times be recognized. Another possibility is that F. nuclea-

tum has two separate adhesins, one for NeuAc and one forGal. Further studies are required to address these possibili-ties. The fact that galacturonic acid and glucuronic acid didnot inhibit agglutination of parotid saliva demonstrates thatthe NeuAc effect is not due to charge. It is tempting tospeculate that NeuAc, fucose, and Gal, in combinations oftwo or three, may have a synergistic effect in inhibitingFN-2-mediated agglutination of parotid saliva.

It must be considered that the presence of specific anti-bodies to F. nucleatum may contribute to the saliva-inducedagglutination. However, previous studies suggested that ahigh-molecular-weight mucinous glycoprotein was responsi-ble for the aggregating activity for oral F. nucleatum strains,and the contribution from specific antibodies was minimaland not inhibitable by Gal (7). Studies by Smoot and Falkler(27) demonstrated that absorption of immunoglobulin M andimmunoglobulin A from saliva did not remove salivaryagglutinating activity. Furthermore, we have shown thatbinding of 3H-F. nucleatum to nitrocellulose replicas ofelectrophoretically separated salivary proteins is only to theproline-rich glycoprotein of human parotid saliva (26).To identify the bacterial surface component involved in

the interaction with galactosyl residues, we separated pro-teins of F. nucleatum FN-2 by SDS-PAGE, transblotted thegels, and incubated the resultant blots with 1251I-asialofetuin.This procedure apparently allowed the renaturation of thelectinlike components that presumably mediate adherence toGal residues. Results suggest that this high-molecular-weightbacterial complex (300,000 to 330,000) represents the Gal-binding adhesin(s) of F. nucleatum FN-2 (Fig. 3).The biological role of this lectin and its interaction in vivo

with salivary glycoconjugates is a matter of speculation.Galactosyl residues are common constituents of salivaryglycoconjugates and may serve as possible receptor sites forF. nucleatum. Falkler et al. (6) indicated that mucinousglycoproteins may play a role in aggregating F. nucleatum inthe oral cavity. Other investigators have suggested that theserum glycoproteins containing Gal may bind to the surfaceof F. nucleatum and that this binding may influence coloni-zation in the gingival sulcus (27). Studies in our laboratoryhave demonstrated that F. nucleatum adheres to saliva-coated cementum in high numbers, and this attachment isnot inhibitable by Gal or Gal-containing glycopeptides (D. G.Kern, J. R. Winkler, and P. A. Murray, J. Dent. Res. 66(Special Issue):1155A, 1987 [manuscript in preparation]). Onthe other hand, F. nucleatum FN-2 specifically binds to aproline-rich glycoprotein in human parotid saliva via galac-tosyl residues (26). Collectively, these data implicate a rolefor this Gal-binding lectin of F. nucleatum in bacterialmediated clearance and suggest that adherence and clear-ance of F. nucleatum in the oral cavity occur by differentmechanisms.

ACKNOWLEDGMENTS

We thank Evangeline Leash for her editorial work on this paperand Valerie Matarese for her technical assistance.

This research was supported by Public Health Service grants DE07244 and DE 06945 from the National Institute of Dental Research.

LITERATURE CITED1. Barondes, S. H., and T. P. Nowak. 1978. Limulus polyphemus

agglutinin (Limulin). Methods Enzymol. 50:302-305.2. Beachey, E. H. 1981. Bacterial adherence: adhesin-receptor

interactions mediating the attachment of bacteria to mucosalsurfaces. J. Infect. Dis. 143:325-345.

3. Bergey, E. J., M. J. Levine, M. S. Molakala, S. D. Bradway, and

INFECT. IMMUN.

lo. 1- t'-T

-4 OF c)il 9

on February 9, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

GALACTOSE-BINDING LECTIN ON F. NUCLEATUM FN-2

I. Al-Hashipi. 1986. Use of photoaffinity cross-linking agentN-hydroxysuccinimidyl-4-azidosalicylic acid to characterizesalivary-glycoprotein-bacterial interactions. Biochem. J. 234:43-48,

4. Cisar, J. O., A. L. Sandberg, and S. E. Mergenhagen. 1984. Thefunction and distribution of different fimbriae on strains ofActinomyces viscosus and Actinomyces naeslundii. J. Dent.Res, 63:393-396.

5. Debray, H., D. Decout, G. Strecker, G. Spik, and J. Montreuil.1981. Specificity of 12 lectins towards oligosaccharides andglycopeptides related to N-glycosylproteins. Eur. J. Biochem.117:41-55.

6. Falkler, W. A., J. R. Mongiello, and B. W. Burger. 1979.Haemagglutination inhibition and aggregation of Fusobacteriumnucleatum by human salivary mucinous glycoproteins. Arch.Oral Biol. 24:483-489.

7. Falkler, W. A., Jr., and C. E. Hawley. 1977. Hemagglutinatingactivity of Fusobacterium nucleatum. Infect. Immun. 15:230-238.

8. Firon, N., I. Ofek, and N. Sharon. 1983. Carbohydrate speci-ficity of the surface lectiiis of Escherichia coli, Klebsiellapneumoniae, and Salmonella typhimurium. Carbohydr. Res.120:235-249.

9. Firon, N., 1. Ofek, and N. Sharon. 1984. Carbohydrate-bindingsites of the mannose-specific fimbrial lectins of enterobacteria.Infect. Immun. 43:1088-1090.

10. Gibbons, R. J., and J. V. Qureshi. 1978. Selective binding ofblood group-reactive salivary mucins by Streptococcus mutansand other oral organisms. Infect. Immun. 22:665-671.

11. Gibbons, R. J., and J. van Houte. 1980. Bacterial adherence andformation of dental plaque, p. 61-104. In E. H. Beachey (ed.),Receptors and recognition, series B, vol. 6. Bacterial adher-ence. Chapman & Hall, New York.

12. Greenwood, F. C., W. M. Hunter, and J. S. Glover. 1963. Thepreparation of 1311-labeled human growth hormone of highspecific radioactivity. Biochem. J. 89:114-119.

13. Jones, G. W., and R. E. Isaacson. 1983. Proteinaceous bacterialadhesins and their receptors. Crit. Rev. Microbiol. 10:229-260.

14. Kondo, W., M. Nakao, S. Shibata, S. Shizukuishi, R. Nakamura,and A. Tsunemitsu. 1983. Hemagglutinating activity of Leptotri-chia buccalis cells and their adherence to saliva-coaited enamelpowder. J. Periodontol. 54:163-172.

15. Kyhse-Andersen, J. 1984. Electroblotting of multiple gels: asimple apparatus without buffer tank for rapid transfer ofproteins from polyacrylamide to nitrocellulose. J. Biochem.Biophys. Methods 10:203-209.

16. Laemmli, U. K. 1970. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature (London)227:680-685.

17. Leffler, H., and S. H. Barondes. 1986. Specificity of binding ofthree soluble rat lung lectins to substituted and unsubstituted

mammalian 1-galactosides. J. Biol. Chem. 261:10119-10126.18. Levine, M. J., L. A. Tabak, M. Reddy, and I. D. Mandel. 1985.

Nature of salivary pellicles in microbial adherence: role ofsalivary mucins, p. 125-130. In S. E. Mergenhagen and B.Rosan (ed.), Molecular basis of oral microbial adhesion. Amer-ican Society for Microbiology, Washington, D.C.

19. Mongiello, J. R., and W. A. Falkler. 1979. Sugar inhibition oforal Fusobacterium nucleatum haemagglutination and cell bind-ing. Arch. Oral Biol. 24:539-545.

20. Murray, P. A., M. J. Levine, M. S. Reddy, L. A. Tabak, andE. J. Bergey. 1986. Preparation of a sialic-acid binding proteinfrom Streptococcus mitis KS32AR. Infect. Immun. 53:359-365.

21. Murray, P. A., M. J. Levine, L. A. Tabak, and M. S. Reddy.1982. Specificity of salivary-bacterial interactions. II. Evidencefor a lectin on Streptococcus sanguis with specificity for aNeuAca2,3 GalI1,3GalNac. Biochem. Biophys. Res. Commun.106:390-396.

22. Murray, P. A., V. Matarese, C. I. Hoover, and J. R. Winkler.1987. The identification of oral microbial lectins by cell affinitychromatography. FEMS Microbiol. Lett. 40:123-127.

23. Nagata, K., M. Nakao, S. Shibata, S. Shizukuishi, R. Nakamura,and A. Tsunemitsu. 1983. Purification and characterization ofgalactosephilic component present on the cell-surfaces of Strep-tococcus sanguis ATCC 10557. J. Periodontol. 54:163-172.

24. Nelville, D. M. 1971. Molecular weight determination of protein-dodecyl sulfate complexes by gel electrophoresis in a discontin-uous buffer system. J. Biol. Chem. 246:6328-6334.

25. Ofek, I., and A. Perry. 1985. Molecular basis for adherence totissues, p. 7-13. In S. E. Mergenhagen and B. Rosan (ed.),Molecular basis of oral microbial adhesion. American Societyfor Microbiology, Washington, D.C.

26. Prakobphol, A., P. A. Murray, and S. J. Fisher. 1987. Bacterialadherence on replicas of sodium dodecyl sulfate-polyacrylamidegels. Anal. Biochem. 164:5-11.

27. Smoot, C. N., and W. A. Falkler. 1981. Attachment of serumnon-antibody glycoproteins to the Fusobacterium nucleatumfound in periodontal disease. Arch. Oral Biol. 26:859-864.

28. Svanborg Eden, C., B. Andersson, H. Leffler, G. Magnusson, andG. Noori. 1985. Specificity for glycolipid receptors of adheringbacteria: experimental differences between uropathogenic Esch-erichia coli and Streptococcus pneumoniae, p. 14-17. In S. E.Mergenhagen and B. Rosan (ed.), Molecular basis of oralmicrobial adhesion. American Society for Microbiology, Wash-ington, D.C.

29. Tabak, L. A., M. J. Levine, I. D. Mandel, and S. A. Ellison.1982. Role of salivary mucins in the protection of the oralcavity. J. Oral Pathol. 11:11-17.

30. Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretictransfer of proteins from polyacrylamide gels to nitrocellulosesheets: procedure and some applications. Proc. Nati. Acad. Sci.USA 76:4350-4354.

VOL. 56, 1988 1319

on February 9, 2021 by guest

http://iai.asm.org/

Dow

nloaded from