Amylase-Binding Protein B of Streptococcus gordonii Is an ...

INFECTION AND IMMUNITY, Oct. 2008, p. 4530–4537 0019-9567/08/$08.00⫹0 doi:10.1128/IAI.00186-08 Copyright © 2008, American Society for Microbiology. All Rights Reserved.

Vol. 76, No. 10

Amylase-Binding Protein B of Streptococcus gordonii Is an Extracellular Dipeptidyl-Peptidase䌤 Biswendu Chaudhuri,1 Susanna Paju,1† Elaine M. Haase,1 M. Margaret Vickerman,2 Jason M. Tanzer,3 and Frank A. Scannapieco1* Department of Oral Biology, School of Dental Medicine, State University of New York at Buffalo, Buffalo, New York1; Department of Periodontics and Endodontics, School of Dental Medicine, State University of New York at Buffalo, Buffalo, New York2; and Department of Oral Health and Diagnostic Sciences, School of Dental Medicine, University of Connecticut Health Center, Farmington, Connecticut3 Received 11 February 2008/Returned for modification 26 March 2008/Accepted 24 July 2008

The oral commensal bacterium Streptococcus gordonii interacts with salivary amylase via two amylasebinding proteins, AbpA and AbpB. Based on sequence analysis, the 20-kDa AbpA protein is unique to S. gordonii, whereas the 82-kDa AbpB protein appears to share sequence homology with other bacterial dipeptidases. The aim of this study was to verify the peptidase activity of AbpB and further explore its potential functions. The abpB gene was cloned, and histidine-tagged AbpB (His-AbpB) was expressed in Escherichia coli and purified. Its amylase-binding activity was verified in an amylase ligand binding assay, and its crossreactivity was verified with an anti-AbpB antibody. Both recombinant His-AbpB and partially purified native AbpB displayed dipeptidase activity and degraded human type VI collagen and fibrinogen, but not salivary amylase. Salivary amylase precipitates not only AbpA and AbpB but also glucosyltransferase G (Gtf-G) from S. gordonii supernatants. Since Streptococcus mutans also releases Gtf enzymes that could also be involved in multispecies plaque interactions, the effect of S. gordonii AbpB on S. mutans Gtf-B activity was also tested. Salivary amylase and/or His-AbpB caused a 1.4- to 2-fold increase of S. mutans Gtf-B sucrase activity and a 3to 6-fold increase in transferase activity. An enzyme-linked immunosorbent assay verified the interaction of His-AbpB and amylase with Gtf-B. In summary, AbpB demonstrates proteolytic activity and interacts with and modulates Gtf activity. These activities may help explain the crucial role AbpB appears to play in S. gordonii oral colonization.

deficient mutant strains of S. gordonii display diminished colonization on teeth of rats compared to wild-type strains, suggesting that this protein plays an important role in oral colonization by this bacterium (39). In the rat model, levels of colonization of S. gordonii on the teeth are highest in the presence of sucrose (37). This is due at least in part to the synthesis of glucans by glucosyltransferase G (Gtf-G), a major sucrose-metabolizing enzyme produced by this species, which facilitates biofilm formation (1, 21, 38, 40). Mutant strains deficient in abpB are not able to colonize the teeth of sucrose-fed rats as well as parental strains. In addition, abpB mutant strains fail to colonize when the animals are fed a sucrose-free, starch-rich diet (39). These data suggest that AbpB might play a role in colonization by interacting in some way with the glucans or Gtf of S. gordonii or by interacting with other salivary or bacterial proteins to influence oral biofilm formation. It has also been speculated that S. gordonii obtains nutrients in vivo from the metabolism of salivary components as well as ingested food in the oral cavity (5, 26). S. gordonii is well known to colonize endothelial surfaces, especially damaged heart valves, in cases of infective endocarditis (15). Extracellular proteases produced by oral streptococci have been shown to degrade host proteins including albumin (24), salivary proteins (4), casein (28), and gelatin and collagen (14). It has been suggested that such proteases enable nutrient acquisition during times of stress or slow growth, as is seen within the bacterial vegetation causing endocarditis or within biofilms (17). We report here that AbpB shows sequence similarity to members of the U34 family of dipeptidases, first isolated from Lactoba-

Saliva-bacterium interactions are important to the development and maintenance of the oral bacterial biofilms that are responsible for dental caries and periodontal diseases, two chronic diseases that result in worldwide morbidity and have considerable economic impact (7, 31, 32). Amylase, a major constituent of human saliva, binds specifically and with high affinity to a number of oral streptococcal species, including Streptococcus gordonii, but not to Streptococcus mutans or other mutans streptococci (6, 18, 33). Amylase-binding streptococci constitute a substantial proportion of the total cultivable flora on human teeth and appear to colonize only the mouths of mammals that secrete saliva with amylase activity (34). These findings suggest that the ability of S. gordonii to bind amylase is ecologically advantageous. To date, two S. gordonii amylase-binding proteins, AbpA and AbpB, have been identified (23, 29). AbpA (20 kDa) is transiently associated with the cell wall following its secretion and serves as the receptor for amylase binding to the bacterial cell surface (30). AbpB (82 kDa), also secreted by S. gordonii, does not mediate amylase binding to the bacterial cell surface, despite its ability to bind amylase when immobilized (23). However, in vivo, AbpB* Corresponding author. Mailing address: Department of Oral Biology, School of Dental Medicine, University at Buffalo, The State University of New York, Foster Hall, Buffalo, NY 14214. Phone: (716) 829-3373. Fax: (716) 829-3942. E-mail: [email protected]. † Present address: Institute of Dentistry, University of Helsinki, Helsinki, Finland. 䌤 Published ahead of print on 4 August 2008. 4530

STREPTOCOCCUS GORDONII AbpB IS A DIPEPTIDYL-PEPTIDASE

VOL. 76, 2008

cillus helveticus (8). This peptidase activity may enable S. gordonii to degrade host proteins such as collagen and fibrinogen. Furthermore, this protein may also interact with Gtf present in mixed-species dental plaque to modulate Gtf enzymatic activity. Such protein-protein interactions may play important roles in bacterial colonization of the oral cavity. MATERIALS AND METHODS Bacterial strains and culture conditions. Streptococcal strains were cultivated weekly from frozen stocks by plating on tryptic soy agar (Becton Dickinson) supplemented with 5% sheep blood and 0.5% yeast extract and incubated overnight in CO2-enriched air at 37°C. Escherichia coli strains were cultured in Luria-Bertani (LB) broth (Difco Laboratories, Detroit, MI) with constant shaking at 37°C and maintained on LB agar supplemented with ampicillin (50 ␮g/ml) for plasmid selection as needed. Streptococcal strains were routinely cultured in tryptic soy plus yeast extract broth. Preparation of streptococcal genomic DNA and PCR of abpB. Streptococcal genomic DNA was prepared from streptococci using a modification of a previously described method (36). Cells were cultured overnight in 5 ml of tryptic soy plus yeast extract broth containing 0.5% glycine, harvested by centrifugation, and resuspended in 500 ␮l of GTE (50 mM glucose, 25 mM Tris, 10 mM EDTA, pH 7) buffer containing 1 mg lysozyme and 50 U mutanolysin. To this, 100 ␮l of 20% sodium dodecyl sulfate (SDS) and 50 ␮l of 2% Sarkosyl were added. After incubation for 1 h at 37°C, 150 ␮l of 5 M sodium percholate was added and DNA was purified by phenol-chloroform extraction. PCR was performed in a 50-␮l reaction mixture containing 0.2 mM deoxynucleoside triphosphates, 0.4 mM of each primer constructed for amplification of the abpB based on the S. gordonii Challis CH1 sequence (abpBF4, 5⬘-GAAGCAATTGAGTTGCTTGC-3⬘; abpBR5, 5⬘-CAGACTTACGTCCAG CAGC-3⬘) (Invitrogen, Carlsbad, CA), 5 ␮l of 10⫻ PCR buffer, and 1 U Taq DNA polymerase (Promega, Madison, WI). PCR was performed for 30 cycles of 1 min at 94°C, 1 min at 55°C, and 2 min at 72°C. Amplified products were separated in a 1% agarose gel, stained with ethidium bromide, and visualized under UV. As a control for PCR, two universal 16S rRNA gene primers, SDBact033aA18 (5⬘-ACTCCTACGGGAGGCAGC-3⬘) and SDBact1525aS17 (5⬘-AAGGAGGTGATCCAGCC-3⬘), were used to confirm the presence of bacterial DNA in each sample (19). Expression of recombinant AbpB. The ligation-independent expression vector pET32Xa/LIC (Novagen, Madison, WI) was used to clone and express AbpB from S. gordonii. The abpB gene was amplified by PCR from S. gordonii Challis CH1 genomic DNA using forward primer OBC23 (5⬘-GGTATTGAGGGTCG CATGAAGAAGTTAACA-3⬘) and reverse primer OBC24 (5⬘-AGAGGAGAG TTAGAGCCCTCAACAGAAAAA-3⬘) (the 5⬘ ligation-independent cloning extensions for cloning are underlined). A 2,070-bp PCR product was purified from a 2% agarose gel and cloned into the pET32-Xa/LIC expression vector according to the manufacturer’s instructions to give p218-1. The recombinant plasmid was introduced by chemical transformation of NovaBlue single competent cells (Novagen) and selected on LB agar supplemented with ampicillin (100 ␮g/ml). Plasmids containing inserts of the correct size and orientation were purified using the Wizard Plus SV Minipreps DNA purification system (Promega, Madison, WI), and the cloned region was verified by nucleotide sequencing at the Roswell Park Cancer Institute DNA Sequencing Laboratory using an ABI Prism 3130XL genetic analyzer. Plasmid p218-1 was introduced into BL21(DE3)pLysS competent cells (Novagen), which were cultured in 50 ml LB broth supplemented with ampicillin (50 ␮g/ml) and 1% glucose until an optical density (OD) at 600 nm of 1.0. Recombinant AbpB was induced by the addition of 1 mM isopropyl-␤-D-thiogalactopyranoside (IPTG) for 2 h at 37°C with constant shaking. Cells were harvested by centrifugation at 4,000 ⫻ g for 15 min at 4°C, and the cell pellet was stored frozen at ⫺20°C. The induced E. coli cell pellet was suspended in 5 ml BugBuster master mix (Novagen). The inclusion bodies were collected by centrifugation and solubilized in binding buffer (500 mM NaCl, 6 M urea, 20 mM Tris-HCl, 5 mM imidazole, pH 7.9). For purification of histidine-tagged recombinant AbpB (His-AbpB), a His.Bind Quick 900 cartridge of Ni2⫹-charged His.bind cellulose (Novagen) was used. The solution of E. coli inclusion bodies containing His-AbpB was loaded onto the affinity cartridge, which was equilibrated with binding buffer. The cartridge was washed twice with binding buffer and once with wash buffer (500 mM NaCl, 20 mM Tris-HCl, 60 mM imidazole, pH 7.9). Finally, the His-AbpB was eluted in buffer (500 mM NaCl, 20 mM Tris-HCl, 1 M imidazole, pH 7.9). Eluted His-AbpB was dialyzed overnight against 50 mM Tris (pH 8) at 4°C and

4531

stored in aliquots at ⫺20°C. The identity of the purified protein was validated by peptide mass fingerprinting analysis of trypsin-digested His-AbpB by matrixassisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry, as previously described (3). Purification of native AbpB and His-AbpB and enzymatic assay. Molecular exclusion chromatography was initially used to purify native AbpB. Supernatant was collected from 1 liter of an S. gordonii Challis CH1 overnight culture by centrifugation at 6,000 ⫻ g for 10 min at room temperature, followed by filtration through a 0.22-␮m filter (polyethersulfone; Corning Inc., Corning, NY). The sterile filtrate was dialyzed extensively against distilled water and then lyophilized. The dried material was solubilized in 25 ml of sterile water to which 25 ml of 95% ethanol was added and kept at 4°C for 1 h. The resulting precipitate was collected by centrifugation at 1,000 ⫻ g for 10 min at 4°C, solubilized in 11 ml of distilled water, and dialyzed against 100 mM Tris-HCl–100 mM NaCl buffer, pH 8. The protein solution was then loaded on a 1.5- by 60-cm column of Bio-Gel P-60 (Bio-Rad, Hercules, CA). Each 5-ml fraction was evaluated for protein at an absorbance of 280 nm and for enzymatic activity using H-Gly-Pro-p-nitroanilide (H-Gly-Pro-pNa) as a chromogenic substrate in accordance with a previously described procedure (12). Briefly, 50 ␮l of substrate in 50 mM Tris (pH 7.8), 1 mM CaCl2, and 50 ␮l of either native AbpB or His-AbpB was added to wells of 96-well microtiter plates (Nunc, Rochester, NY), followed by incubation at 37°C. After 6 h the release of p-nitroaniline was measured at 405 nm using an AD 340 microplate reader (Beckman Coulter, Fullerton, CA). The following chromogenic substrates (Bachem Biosciences Inc., Torrance, CA) were tested at a final concentration of 1 mM: H-Pro-pNa, H-Ala-Pro-pNa, H-Gly-Pro-pNa, H-Arg-Pro-pNa, H-Gly-Arg-pNa, H-Ala-Ala-pNa, H-Lys-pNa, and H-LeuLeu-pNa. AbpB was also purified by immunoaffinity chromatography. Antisera against purified His-AbpB were produced commercially in New Zealand White rabbits using standard methods (GenScript Corp., Piscataway, NJ). Sera were obtained 10 days after the final booster injection, tested by immunoblotting, and stored in 1-ml aliquots at ⫺70°C. Preimmune sera, obtained from each rabbit before immunization, were tested to assure the absence of endogenous cross-reactive antibodies against streptococcal antigens. The final titer measured against the recombinant antigen was 1:128,000. An AbpB-specific affinity column was prepared using a commercial kit (Pierce, Rockford, IL) by following the manufacturer’s protocol. Briefly, immunoglobulin G (IgG) was isolated from rabbit serum by ammonium sulfate precipitation, desalted by dialysis, oxidized, and coupled at the Fc region to agarose beads. Similarly, a “sham” column was prepared using the globulin from the preimmune serum derived from the immunized rabbit. Culture supernatant (1 liter) from S. gordonii (strain CH1) incubated overnight in tryptic soy plus yeast extract medium was collected by centrifugation, passed through a 0.22-␮m filter, dialyzed exhaustively against distilled water at 4°C, and then lyophilized. The powder was suspended in 35 ml of distilled water, with a resulting protein concentration of 28.4 mg/ml as measured using the bicinchoninic acid assay (Pierce) with an albumin standard. Affinity chromatography was performed by incubating 32 ml of the suspension containing 909 mg of bacterial protein with 4 ml of the affinity matrix containing 5.3 mg bound IgG at 4°C with gentle tumbling on a rotator for 24 h. The mixture was then poured into a chromatography column, allowed to settle, and washed with phosphate-buffered saline until the effluent absorbance at 280 nm was less than 0.020. Bound material was eluted with 0.1 M glycine (pH 3) into tubes containing 50 ␮l 1 M Tris-HCl buffer (pH 9). The fractions were tested for protease activity as described previously (12); appropriate fractions were then pooled, dialyzed sequentially against phosphate-buffered saline and distilled water, and lyophilized. Identification of resulting products was determined by SDS-polyacrylamide gel electrophoresis (PAGE) and Western blotting using the anti-AbpB antiserum and the amylase-ligand blotting assay described below. Western blot and amylase ligand binding assays. To verify the identity of AbpB, Western blotting using the rabbit anti-AbpB antibodies was performed using standard methods (2). In addition, to determine the amylase-binding ability of His-AbpB, the amylase ligand binding assay was performed by electrotransfer of His-AbpB from SDS-PAGE gels to Immobilon-P membranes (Millipore, Bedford, MA) as previously described (13, 30). Briefly, after being blocked with 5% nonfat milk in Tris-buffered saline containing 0.1% Tween 20 (TBST), membranes were incubated with 1% (wt/vol) purified human salivary amylase (Sigma) in TBST for 30 min. After a washing, the blots were incubated with rabbit polyclonal anti-human amylase (Sigma) in TSBT, washed, and then incubated with goat anti-rabbit IgG conjugated with alkaline phosphatase (Promega). Finally, the blots were developed using the ProtoBlot Western blot alkaline phosphatase system (Promega). Antiamylase antibodies were unreactive against streptococcal proteins in the absence of amylase.

4532

CHAUDHURI ET AL.

Zymography and cleavage of host proteins by His-AbpB. Standardized preparations of native AbpB and His-AbpB were applied to a 12% SDS-PAGE gel containing 0.1% gelatin as a substrate for peptidase activity (17). Gels were then incubated at room temperature in 2.5% Triton X-100 for 30 min and subsequently at 37°C in a buffer containing 50 mM Tris-HCl (pH 7.6), 0.2 M NaCl, 5 mM CaCl2, and 0.02% Triton X-100 for 20 h. Clear zones of proteolysis were visualized following Coomassie blue staining. In addition, enzymatic cleavage of several native host proteins (␣-amylase, fibrinogen, and collagen VI [Sigma]) by purified His-AbpB was also evaluated. The target proteins (5 ␮g) were incubated at 37°C with 0.4 ␮g of His-AbpB for 20 h in a solution of 50 ␮l of 0.1 M Tris-HCl buffer (pH 7.5). These proteins, incubated under the same conditions without His-AbpB, served as negative controls. Enzymatic cleavage of the proteins was evaluated by SDS-PAGE as previously described (17). Gtf activity assay. Gtf activity was qualitatively determined in 12% polyacrylamide gels as previously described (42). Briefly, cell-free filtrates were resolved by SDS-PAGE, followed by an overnight incubation of gels in a solution containing Triton X-100 to renature the Gtf-G. Following this, 3% sucrose was added as a substrate, and the gels were incubated overnight. The synthesized glucan bands were visualized by staining with periodic acid and pararosaniline (Sigma). The sucrase and transferase activities of S. mutans Gtf-B were quantified as previously described (3, 9, 35). Briefly, 0.6 ␮g of purified Gtf-B (supplied by Anne Vacca-Smith, University of Rochester) was incubated in the presence or absence of various amounts of His-AbpB and/or amylase at room temperature for 5 min. To this, 50 ␮l of 0.2 M sodium phosphate buffer (pH 6.0) and 40 ␮l of 1 M sucrose were added to a final volume of 100 ␮l, and the reaction mixture was incubated overnight at 37°C. The amounts of glucose and fructose in the reaction mixture were measured by the use of an enzymatic redox reaction train using hexokinase, glucose-6-phosphate dehydrogenase, 6-phosphoglucose isomerase, and NADP⫹ (F-Kit; R-Biopharm, Mannheim, Germany). The amount of resulting fructose in the reaction mixture represents sucrase activity, and the difference between the amounts of free glucose and free fructose in the reaction mixture corresponds to the amount of glucosyl residues transferred to glucan, the transferase activity (35). Interaction of His-AbpB and amylase with Gtf-B of S. mutans. An enzymelinked immunosorbent assay (ELISA) was used to verify the interaction of His-AbpB and amylase with S. mutans Gtf-B. Twofold serial dilutions of purified amylase or His-AbpB were used to coat wells of a 96-well polystyrene microtiter plate (Nalge Nunc International, Rochester, NY). The plates were washed and blocked in accordance with the manufacturer’s instructions (Kirkegaard & Perry, Gaithersburg, MD). Following this, 1 ␮g of purified Gtf-B was added to each well and incubated at room temperature for 2 h. After a washing, the samples were probed with rabbit polyclonal anti-Gtf-B (kindly provided by Anne Vacca-Smith, University of Rochester) or rabbit anti-human amylase antibodies, as appropriate. The plates were then incubated, washed, and probed with the goat antirabbit secondary antibody provided in the ELISA kit. The plates were developed using the kit substrate according to the manufacturer’s instructions. Following color development, which reflected the amount of target protein bound to the coated antigen, the absorbance of the solution was read at 630 nm using a Beckman Coulter AD 340 microplate reader. Statistical analysis. Statistical analyses were conducted using SPSS software, version 11. Analyses of the enzyme activity were done using the two-sided Student t test.

RESULTS Prevalence of abpB among oral streptococcal species. Preliminary studies searched for the abpB gene by PCR analysis using abpB-specific primers and genomic DNA from 66 streptococcal strains. While abpB PCR amplicons were obtained from all 16 strains of S. gordonii tested, none of the other oral streptococci tested (12 strains of S. mitis, 3 strains of S. cristatus, 1 strain of S. parasanguinis, 2 strains of S. anginosus, 5 strains of S. mutans, 5 strains of S. sobrinus, 6 strains of S. salivarius, 3 strains of S. sanguinus, and 3 strains of S. oralis) showed a positive signal for the abpB gene (data not shown). This suggested that the primer target sequences are present only in S. gordonii or, alternatively, that the target sequences of abpB homologs may be present but not sufficiently conserved

INFECT. IMMUN.

at the annealing sites of the two oligomers chosen as PCR primers. Universal primers for the 16S rRNA gene yielded PCR amplicons from all strains tested, confirming the presence of bacterial template DNA in each PCR. Sequence analysis of AbpB. As previously reported, abpB contains an open reading frame of 1,959 bp that yields a 652amino-acid protein with a predicted molecular mass of 80 kDa, the first 24 amino acid residues of which comprise a hydrophobic signal peptide (23). In the recently completed S. gordonii genome sequence (GenBank accession number CP000725), abpB has been designated locus tag SGO_0162. A genomewide search for homologs and shared functional domains shows that strain Challis carries two additional genes that encode proteins that are about 50% and 34% identical to AbpB over the entire length of the protein at the amino acid level; the associated loci are SGO_0721 and SGO_0724, both encoding AbpB-like dipeptidase lipoproteins. Several other highly similar proteins were also revealed following BLAST searches of other streptococcal genomes, including those of S. suis (61% identical), S. sanguinis (45% identical), and S. pyogenes (45% identical). Similarities were also found with a peptidase from Lactobacillus delbrueckii subsp. lactis (30% identical). BLAST searching also revealed lower-similarity hits with proteins encoded by the S. pneumoniae and S. mutans genomes. These results suggest that AbpB-like proteins are conserved in several other related bacterial species and suggest that AbpB is a new member of the protease family U34 (peptidase C69). This search revealed over 40 U34 family homologs, among which are several reported eukaryotic and archeal proteins (data not shown). Proteins most similar to AbpB appear to be mainly from the Streptococcus and Lactobacillus species. The first member of the U34 family was isolated from Lactobacillus helveticus as a broad-specificity dipeptidase (dipeptidase A, encoded by pepDA) (8). Expression and purification of His-AbpB. A 2,070-bp DNA fragment containing the abpB open reading frame was cloned in frame into an IPTG-inducible pET32Xa/LIC expression vector (Fig. 1A). Although SDS-PAGE analysis detected little His-AbpB in the soluble cell lysate, the fusion protein was present in large amounts in solubilized inclusion bodies recovered from induced E. coli cells. His-AbpB was purified from the solubilized inclusion bodies using a nickel column. SDSPAGE gels of the resulting purified material stained with Coomassie blue showed a single band of ⬃100 kDa (Fig. 1B), consistent with the calculated mass of AbpB (82 kDa) plus the 20-kDa N-terminal His tag from the pET32Xa/LIC expression vector. The amylase ligand-binding assay confirmed that AbpB retained amylase-binding activity (Fig. 1B). Control experiments showed that the antiamylase antibodies did not crossreact with AbpB (data not shown). Peptide mass fingerprinting analysis was performed on trypsin-digested His-AbpB (data not shown). The molecular masses of the digested peptides obtained by MALDI-TOF mass spectrometry showed coverage of 25% and were matched with the theoretical peptide masses estimated using Peptidemass software (http://ca.expasy.org/tools/peptide-mass.html), indicating that the purified recombinant protein was homologous to the predicted AbpB of S. gordonii. Purification of native S. gordonii AbpB. In order to determine if AbpB is distinct from the extracellular x-prolyl dipep-

VOL. 76, 2008


FIG. 1. Cloning, purification, and characterization of AbpB. (A) Linear diagram of p218-1, expressing the His-AbpB fusion protein. T7-p, T7 promoter; T7-t, T7 terminator; lacO, lac operator; LIC site, ligation-independent cloning site; rbs, ribosome binding site; ORF, open reading frame. (B) SDS-PAGE of lysate fractions showing the expression of His-AbpB in E. coli harboring p218-1. Lane 1, molecular mass standard in kDa; lanes 2 and 3, Coomassie blue staining of soluble inclusion bodies and supernatant of IPTGinduced E. coli cells, respectively; lanes 4 and 5, Coomassie blue staining of 4 and 2 mg of purified protein, respectively; lanes 6 and 7, amylase ligand binding assay of 400 and 200 ng of His-AbpB, respectively.

tidyl-peptidase (⬃85 kDa) and arginine aminopeptidase (⬃70 kDa) of S. gordonii (11, 12), we partially purified native AbpB from S. gordonii Challis CH1 culture supernatants. With HGly-Pro-pNa as a substrate, fractions 3 and 4 obtained from BioGel P-60 gel filtration showed dipeptidase activity (Fig. 2A). SDS-PAGE analysis of pooled fractions 3 and 4 revealed a band of ⬃82 kDa (Fig. 2B). The amylase ligand binding assay (Fig. 2C), Western blotting using anti-AbpB antibody (Fig. 2D), and MALDI-TOF analysis confirmed that this band was AbpB of S. gordonii (data not shown). However, AbpB can be considered only partially purified since the pooled fraction also contained several other minor protein bands. Based upon its molecular size, we believe that AbpB is distinct from previously reported peptidases of S. gordonii. Attempts were also made to optimize the conditions for peptidase activity. The enzyme preparation was solubilized in 50 mM Tris-HCl buffer containing 1 mM CaCl2 at pH 7, 8, 9, and 10. Optimal activity was detected at pH 7, with lesser activity at pH 8 and no activity at pH 9 and 10. Protease activity of AbpB. Zymography using gelatin or casein has previously been used to demonstrate protease activities of proteins (17, 22). Gelatin zymography of His-AbpB, partially purified native AbpB, and immunoaffinity-purified AbpB revealed clear zones of proteolytic activity within the SDS-PAGE gelatin, suggesting a molecular mass of about 150 kDa (Fig. 2E). Since His-AbpB was present as a single band of 100 kDa by SDS-PAGE and native AbpB is known to be 82 kDa, it is likely that these proteins were retarded by the gelatin protein in the zymogram, as has been previously reported (17).

4533

Zymography also showed an additional clear zone at around the 50-kDa region, perhaps representing a proteolytic fragment of AbpB (Fig. 2E). In order to determine the enzyme specificity of AbpB, eight chromogenic endo- and aminopeptidase substrates were tested using purified His-AbpB and partially purified native AbpB. Of the eight substrates assessed, Ala-Pro, Gly-Pro, and Arg-Pro were most extensively hydrolyzed (Fig. 2F), suggesting that dipeptidase activity of AbpB was restricted to peptides containing a proline residue. To obtain further information on cleavage specificity, three mammalian proteins (human type VI collagen, amylase, and fibrinogen) were tested as substrates for AbpB peptidase activity. Of these three mammalian proteins, two are rich in proline residues (type VI collagen and fibrinogen) and one is poor in proline residues (amylase). While His-AbpB cleaved human type VI collagen and fibrinogen at 37°C, it did not cleave amylase under the same conditions (Fig. 2G). Immunoaffinity purification with anti-His-AbpB antibody bound to the column was also used to purify native AbpB from Streptococcus gordonii CH1 supernatant. The method yielded 0.023% purified AbpB (210 ␮g) from supernatant protein (909 mg). The affinity column effluent showed one weak band at 82 kDa and three other bands of approximately 52, 36, and 26 kDa (Fig. 3). Each of these bands was able to bind amylase (Fig. 3B, lane 2) and the anti-AbpB antibody (Fig. 3C, lane 2), indicating that they were likely related to the 82-kDa component. The 52-kDa band was sequenced but gave no N-terminal signal, suggesting that the N terminus of this protein was blocked. The 52-, 36-, and 26-kDa bands were not seen in concentrated supernatant either when blotted and probed with rabbit anti-AbpB antiserum or by the amylase-ligand binding assay, suggesting that they were in low abundance in the starting material. It is therefore likely that these lower-molecularweight bands are breakdown products from AbpB autodigestion or cross-reactive peptides encoded by other abpB-like genes that are present in the S. gordonii genome (for example an AbpB-like dipeptidase lipoprotein [accession number YP_001450022, locus tag SGO_0721]). The peptidase enzyme activity of the immunoaffinity-purified AbpB was assessed using H-Gly-Pro-pNa as the substrate. The starting S. gordonii Challis CH1 supernatant showed 3 OD units/␮g of protein, and the affinity-purified AbpB showed 628 OD units/␮g, representing just over a 200-fold enrichment in enzyme activity. Effects of amylase and His-AbpB on Gtf-B activity. Previous studies found that AbpB was precipitated from S. gordonii culture supernatants by the addition of purified salivary amylase (23) but not by the addition of bovine serum albumin (BSA). The resulting precipitate contained the streptococcal proteins AbpA, AbpB, and Gtf-G in addition to amylase. The sizes of the proteins, as determined by SDSPAGE, were identical to their predicted sizes and not degraded. Also, other studies demonstrated that AbpA from S. gordonii was able to interact with and modulate the enzyme activity of Gtf-B from S. mutans (3). The presence of amylase and/or AbpA increased both the sucrase and transferase component activities of S. mutans Gtf-B. The results suggested that an extracellular protein network of AbpAamylase-Gtf might influence the ecology of oral biofilms.

4534

CHAUDHURI ET AL.

INFECT. IMMUN.

FIG. 2. Purification of native AbpB, zymography, enzymatic assay, and cleavage of host proteins by AbpB. (A) Protein elution profile from Bio-Gel P-60 molecular exclusion column. Fractions of 5 ml were collected. Only fractions 3 and 4 displayed peptidase activity. (B) SDSPAGE analysis of pooled fractions 3 and 4 from Bio-Gel P-60 containing peptidase activity is shown with Coomassie blue staining. Lanes 1 and 2 contained 10 and 5 ␮g protein, respectively. (C) Amylase ligand binding assay of native, enriched AbpB (5 ␮g protein). (D) Western blot using anti-AbpB antibody with native, enriched AbpB (5 ␮g protein). (E) Gelatin zymography of native AbpB and His-AbpB. Lane 1, boiled AbpB (5 ␮g); lane 2, AbpB (5 ␮g); lane 3, boiled His-AbpB (4 ␮g); lane 4, His-AbpB (4 ␮g); lane 5, protein from E. coli cell extract (control); lane 6, recombinant amylase-binding protein A (rAbpA) as a control; lane 7, molecular mass standard in kDa. (F) Enzymatic assay of AbpB using chromogenic substrates. Release of p-nitroaniline, representing dipeptidase activity, was measured at 405 nm using a microplate reader. Dark bars, results using His-AbpB; white bars, results using native AbpB. (G) Proteolytic cleavage of host proteins by His-AbpB. Substrate proteins (5 ␮g each) were incubated with 400 ng of His-AbpB at 37°C for 20 h. Cleavage patterns were visualize by Coomassie staining. Lane 1, human collagen VI alone; lane 2, human collagen VI plus His-AbpB; lane 3, salivary amylase alone; lane 4, salivary amylase plus His-AbpB; lane 3, fibrinogen alone; lane 6, fibrinogen with His-AbpB. Arrowheads indicate protein bands in untreated samples degraded following incubation with AbpB.

FIG. 3. Immunoaffinity purification of AbpB. (A) Coomassie bluestained gel. Lane 1, culture supernatant starting material; lane 2, material eluted from the AbpB affinity column. (B) Amylase-ligand binding assay. Lane 1, culture supernatant starting material; lane 2, material eluted from the AbpB affinity column. (C) Western blot using anti-AbpB antibody. Lane 1, culture supernatant starting material; lane 2, material eluted from the AbpB affinity column.

Based on these results, we wished to determine if His-AbpB also interacted with Gtf-B. Therefore, the individual sucrase and transferase activities of purified S. mutans Gtf-B in the presence or absence of amylase and/or His-AbpB were quantified. Its sucrase activity was increased 1.4- to 2-fold (P ⱕ 0.01) and its transferase activity was elevated 3- to 6-fold (P ⱕ 0.001) in the presence of purified human amylase and/or His-AbpB (Fig. 4A and B). Simultaneous incubation of amylase and AbpB with Gtf-B did not cause additional or synergistic enhancement of Gtf activities. Under identical assay conditions, addition of BSA in control experiments had no effect on Gtf-B enzyme activity (data not shown). The physical interaction between His-AbpB and/or amylase and Gtf-B was verified by ELISA; the binding of His-AbpB and amylase to Gtf-B was in direct proportion to the amount of immobilized protein present (Fig. 5). Under identical assay conditions, immobilized BSA did not bind to Gtf-B, thus supporting the selectivity of Gtf-B binding with amylase or HisAbpB.


VOL. 76, 2008

4535

FIG. 4. Measurement of sucrase and transferase activities of Gtf-B. Purified Gtf-B enzyme was incubated at 37°C overnight with sucrose in the presence or absence of amylase and/or His-AbpB protein. The amounts of glucose and fructose present in the reaction mixtures were measured using the F-Kit. Transferase (A) and sucrase (B) activities of Gtf-B in the presence of amylase and/or His-AbpB protein are shown by the histograms.

DISCUSSION Previous studies have shown that S. gordonii produces at least two amylase-binding proteins, AbpA and AbpB (23, 30). Insertional inactivation of abpA completely eliminated amylase binding to the bacteria, demonstrating that AbpA is crucial for amylase binding to the cell surface (30). Interestingly, however, AbpAdeficient mutants were able to colonize the teeth of experimental rats, usually better than the wild type (39). To date, this protein appears to be unique, with no homologs produced by other bac-

FIG. 5. Binding of His-AbpB and amylase with Gtf-B. Four, 2, 1, 0.5, and 0.25 ␮g of either His-AbpB or amylase were used to coat wells of an ELISA plate. To each well, 0.5 ␮g of Gtf-B was added. Gray bars indicate binding of Gtf-B protein to immobilized His-AbpB; white bars indicate binding of Gtf-B to immobilized amylase.

terial species, as revealed by bioinformatic searches of the Protein Sequence Data Bank or other databases. In contrast, mutation of abpB does not affect the binding of amylase to the bacterial cell surface (23). However, AbpBdeficient mutants are defective in their ability to colonize rat teeth, suggesting this protein’s importance as a colonization determinant for S. gordonii (39). A search of the Protein Sequence Data Bank for sequences homologous with AbpB revealed significant sequence similarity with members of the U34 dipeptidase family of proteases (27). Previous sequence and structural analysis of U34 family members identified a conserved N-terminal cysteine that along with other residues might be important for enzymatic activity (27). The present paper supports the bioinformatic findings and provides functional evidence that AbpB is a protease; enzymatic assays confirmed that both recombinant and partially purified native AbpB display dipeptidase activity, and HisAbpB was able to degrade proline-rich proteins such as human type VI collagen and fibrinogen. AbpB was unable to degrade salivary amylase, likely due to the absence of proline-containing susceptible cleavage sites within the protein. The enzyme activity of the purified protein, however, was found to be modest. Several explanations can be offered for this observation, including use of unfavorable conditions during the enzyme assay, the absence of as yet unknown enzyme cofactors, structural modification of the protein as the result of purification, and/or autodegradation. AbpB appears distinct from previously described proteases of S. gordonii (11, 12, 17). The peptidase activity of AbpB may prove advantageous to the bacterium for nutrient acquisition or interaction with the tooth surface glycoprotein pellicle during oral colonization. Potential substrates for this enzyme could be the abundant proline-rich proteins in saliva. These proteins are known to play important roles in dental plaque

4536

CHAUDHURI ET AL.

formation and tooth remineralization and are found to be quite susceptible to proteolytic breakdown (10, 16). Of course, in light of the potential heterogeneity in function of the U34 family (27), AbpB may possess other heretofore-undescribed functions that influence the bacterium’s colonization potential. Sequence analysis also suggests that this protein may be a lipoprotein and therefore could be, in part, bacterial cell surface bound as well as released into the environment. S. gordonii produces a single Gtf (Gtf-G) (1), while Streptococcus mutans secretes three Gtfs (Gtf-B, Gtf-C, and Gtf-D) (20). Gtfs, long known as determinants of bacterial colonization of the teeth, synthesize extracellular glucans from sucrose (38). Their actions entail two sequential steps: the hydrolysis of sucrose into fructose and glucose and the transfer of glucosyl residues to form ␣-linked glucans (1). Gtfs consist of two relatively conserved structural domains, an N-terminal catalytic domain and a C-terminal glucan-binding domain (25). Although Gtfs of S. gordonii and S. mutans share conserved domains and are highly specific for sucrose, each Gtf exhibits distinct enzyme activities that alter the position of the resulting glucan’s glucosyl linkages and/or the acceptor molecule requirement for its catalytic activity (20). The interaction of amylase with Gtf enzymes on saliva-coated hydroxyapatite surfaces results in reduction in Gtf enzymatic activity and glucan formation (41), which in turn may modulate the colonization of oral streptococci. We have previously demonstrated that AbpB complexes with Gtf-G and amylase (3). S. gordonii exists within dental plaque and may influence cariogenic species such as S. mutans. We therefore assessed if S. gordonii AbpB could interact with and modify the enzymatic activity of S. mutans Gtf-B. The present study showed that soluble Gtf-B of S. mutans interacted with both soluble amylase and with His-AbpB, with resultant increased activities of its constituent catalytic functions. Indeed, His-AbpB alone was able to modulate Gtf-B activity. These results are consistent with our previous report that AbpB-deficient mutant strains of S. gordonii colonized rats at significantly lower levels than wild-type strains in the presence of a sucrose diet (39). Because Gtf requires sucrose as a substrate for extracellular glucan production, the in vitro observation that His-AbpB increases Gtf enzymatic activity may explain, in part, the rescue of colonization of abpB mutants (that fail to colonize rats eating a sucrose-free diet) if there is sucrose diet supplementation (39). It is also possible that AbpB could modify bacterial surface proteins such as Gtf-G (which has 2% Pro residues) through protease activity and that it also provides amidase activity (resulting in cell wall modification). Such modifications remain to be demonstrated. In summary, while immobilized AbpB interacts with amylase (as is seen, for example, in the amylase ligand binding assay), mutant strains of S. gordonii lacking AbpB retain amylase binding ability, suggesting that this protein is not involved in the direct binding of amylase to the cell surface. In vivo studies demonstrate that AbpB is important for S. gordonii colonization. In light of the demonstration of proteolytic activity it is possible that AbpB’s primary role in colonization may not be to bind amylase but rather to provide a crucial metabolic pathway for nutrient acquisition. Future studies will continue to dissect the role of amylase in S. gordonii colonization.

INFECT. IMMUN. ACKNOWLEDGMENTS This work was supported by USPHS grants DE07034 and DE09838 from the National Institute of Dental and Craniofacial Research. The technical assistance of Paul Bronson is gratefully acknowledged. REFERENCES 1. Banas, J. A., and M. M. Vickerman. 2003. Glucan-binding proteins of the oral streptococci. Crit. Rev. Oral Biol. Med. 14:89–99. 2. Burnette, W. N. 1981. “Western blotting”. Electrophoretic transfer of proteins from sodium dodecyl sulphate-polyacrylamide gels to unmodified nitrocellulose and radiographic detection with antibody and radioiodinated protein A. Anal. Biochem. 112:195–203. 3. Chaudhuri, B., J. Rojek, M. M. Vickerman, J. M. Tanzer, and F. A. Scannapieco. 2007. Interaction of salivary alpha-amylase and amylase-binding-protein A (AbpA) of Streptococcus gordonii with glucosyltransferase of S. gordonii and Streptococcus mutans. BMC Microbiol. 7:60. 4. Choih, S., Q. T. Smith, and C. F. Schachtele. 1979. Modification of human parotid saliva by oral Streptococcus sanguis. J. Dent. Res. 58:516–523. 5. de Jong, M. H., J. S. van der Hoeven, O. J. van, and J. H. Olijve. 1984. Growth of oral Streptococcus species and Actinomyces viscosus in human saliva. Appl. Environ. Microbiol. 47:901–904. 6. Douglas, C. W. 1983. The binding of human salivary alpha-amylase by oral strains of streptococcal bacteria. Arch. Oral Biol. 28:567–573. 7. Douglas, C. W. I. 1994. Bacterial-protein interactions in the oral cavity. Adv. Dent. Res. 8:254–262. 8. Dudley, E. G., A. C. Husgen, W. He, and J. L. Steele. 1996. Sequencing, distribution, and inactivation of the dipeptidase A gene (pepDA) from Lactobacillus helveticus CNRZ32. J. Bacteriol. 178:701–704. 9. Eto, A., T. C. Saido, K. Fukushima, S. Tomioka, S. Imai, T. Nisizawa, and N. Hanada. 1999. Inhibitory effect of a self-derived peptide on glucosyltransferase of Streptococcus mutans. Possible novel anticaries measures. J. Biol. Chem. 274:15797–15802. 10. Gibbons, R. J., D. I. Hay, and D. H. Schlesinger. 1991. Delineation of a segment of adsorbed salivary acidic proline-rich proteins which promotes adhesion of Streptococcus gordonii to apatitic surfaces. Infect. Immun. 59: 2948–2954. 11. Goldstein, J. M., A. Banbula, T. Kordula, J. A. Mayo, and J. Travis. 2001. Novel extracellular x-prolyl dipeptidyl-peptidase (DPP) from Streptococcus gordonii FSS2: an emerging subfamily of viridans streptococcal x-prolyl DPPs. Infect. Immun. 69:5494–5501. 12. Goldstein, J. M., D. Nelson, T. Kordula, J. A. Mayo, and J. Travis. 2002. Extracellular arginine aminopeptidase from Streptococcus gordonii FSS2. Infect. Immun. 70:836–843. 13. Gwynn, J. P., and C. W. Douglas. 1994. Comparison of amylase-binding proteins in oral streptococci. FEMS Microbiol. Lett. 124:373–379. 14. Harrington, D. J., and R. R. Russell. 1994. Identification and characterization of two extracellular proteases of Streptococcus mutans. FEMS Microbiol. Lett. 121:237–241. 15. Herzberg, M. C., M. W. Meyer, A. Kilic, and L. Tao. 1997. Host-pathogen interactions in bacterial endocarditis: streptococcal virulence in the host. Adv. Dent. Res. 11:69–74. 16. Jensen, J. L., M. S. Lamkin, and F. G. Oppenheim. 1992. Adsorption of human salivary proteins to hydroxyapatite: a comparison between whole saliva and glandular salivary secretions. J. Dent. Res. 71:1569–1576. 17. Juarez, Z. E., and M. W. Stinson. 1999. An extracellular protease of Streptococcus gordonii hydrolyzes type IV collagen and collagen analogues. Infect. Immun. 67:271–278. 18. Kilian, M., and B. Nyvad. 1990. Ability to bind salivary alpha-amylase discriminates certain viridans group streptococcal species. J. Clin. Microbiol. 28:2576–2577. 19. Kroes, I., P. W. Lepp, and D. A. Relman. 1999. Bacterial diversity within the human subgingival crevice. Proc. Natl. Acad. Sci. USA 96:14547–14552. 20. Kuramitsu, H. K. 2003. Molecular genetic analysis of the virulence of oral bacterial pathogens: an historical perspective. Crit. Rev. Oral Biol. Med. 14:331–344. 21. Kuramitsu, H. K. 1993. Virulence factors of mutans streptococci: role of molecular genetics. Crit. Rev. Oral Biol. Med. 4:159–176. 22. Lantz, M. S., and P. Ciborowski. 1994. Zymographic techniques for detection and characterization of microbial proteases. Methods Enzymol. 235: 563–594. 23. Li, L., J. M. Tanzer, and F. A. Scannapieco. 2002. Identification and analysis of the amylase-binding protein B (AbpB) and gene (abpB) from Streptococcus gordonii. FEMS Microbiol. Lett. 212:151–157. 24. Lo, C. S., and C. V. Hughes. 1996. Identification and characterization of a protease from Streptococcus oralis C104. Oral Microbiol. Immunol. 11:181– 187. 25. Mooser, G., and C. Wong. 1988. Isolation of a glucan-binding domain of glucosyltransferase (1,6-alpha-glucan synthase) from Streptococcus sobrinus. Infect. Immun. 56:880–884. 26. Palmer, R. J., Jr., K. Kazmerzak, M. C. Hansen, and P. E. Kolenbrander. 2001. Mutualism versus independence: strategies of mixed-species oral bio-

VOL. 76, 2008

27. 28. 29.

30.

31. 32. 33. 34. 35.


films in vitro using saliva as the sole nutrient source. Infect. Immun. 69:5794– 5804. Pei, J., and N. V. Grishin. 2003. Peptidase family U34 belongs to the superfamily of N-terminal nucleophile hydrolases. Protein Sci. 12:1131–1135. Rogers, A. H., P. S. Zilm, A. L. Pfenning, and N. J. Gully. 1990. Some aspects of protease production by a strain of Streptococcus sanguis. Oral Microbiol. Immunol. 5:72–76. Rogers, J. D., E. M. Haase, A. E. Brown, C. W. Douglas, J. P. Gwynn, and F. A. Scannapieco. 1998. Identification and analysis of a gene (abpA) encoding a major amylase-binding protein in Streptococcus gordonii. Microbiology 144(Pt. 5):1223–1233. Rogers, J. D., R. J. Palmer, Jr., P. E. Kolenbrander, and F. A. Scannapieco. 2001. Role of Streptococcus gordonii amylase-binding protein A in adhesion to hydroxyapatite, starch metabolism, and biofilm formation. Infect. Immun. 69:7046–7056. Rudney, J. D. 2000. Saliva and dental plaque. Adv. Dent. Res. 14:29–39. Scannapieco, F. A. 1994. Saliva-bacterium interactions in oral microbial ecology. Crit. Rev. Oral Biol. Med. 5:203–248. Scannapieco, F. A., E. J. Bergey, M. S. Reddy, and M. J. Levine. 1989. Characterization of salivary a-amylase binding to Streptococcus sanguis. Infect. Immun. 57:2853–2863. Scannapieco, F. A., L. Solomon, and R. O. Wadenya. 1994. Emergence in human dental plaque and host distribution of amylase-binding streptococci. J. Dent. Res. 73:1627–1635. Shiroza, T., N. Shinozaki-Kuwahara, M. Hayakawa, Y. Shibata, T. Hashizume, K. Fukushima, S. Udaka, and Y. Abiko. 2003. Production of a single-chain variable fraction capable of inhibiting the Streptococcus mutans glucosyltrans-

Editor: A. Camilli

36. 37.

38. 39.

40.

41.

42.

4537

ferase in Bacillus brevis: construction of a chimeric shuttle plasmid secreting its gene product. Biochim. Biophys. Acta 1626:57–64. Sulavik, M. C., G. Tardif, and D. B. Clewell. 1992. Identification of a gene, rgg, which regulates expression of glucosyltransferase and influences the Spp phenotype of Streptococcus gordonii Challis. J. Bacteriol. 174:3577–3586. Tanzer, J. M., L. K. Baranowski, J. D. Rogers, E. M. Haase, and F. A. Scannapieco. 2001. Oral colonization and cariogenicity of Streptococcus gordonii in specific pathogen-free TAN:SPFOM(OM)BR rats consuming starch or sucrose diets. Arch. Oral Biol. 46:323–333. Tanzer, J. M., M. L. Freedman, R. J. Fitzgerald, and R. H. Larson. 1974. Diminished virulence of glucan synthesis-defective mutants of Streptococcus mutans. Infect. Immun. 10:197–203. Tanzer, J. M., L. Grant, A. Thompson, L. Li, J. D. Rogers, E. M. Haase, and F. A. Scannapieco. 2003. Amylase-binding proteins A (AbpA) and B (AbpB) differentially affect colonization of rats’ teeth by Streptococcus gordonii. Microbiology 149:2653–2660. Tanzer, J. M., A. M. Thompson, L. P. Grant, M. M. Vickerman, and F. A. Scannapieco. 2008. Streptococcus gordonii’s sequenced strain CH1 glucosyltransferase determines persistent but not initial colonization of teeth of rats. Arch. Oral Biol. 53:133–140. Vacca-Smith, A. M., A. R. Venkitaraman, R. G. Quivey, Jr., and W. H. Bowen. 1996. Interactions of streptococcal glucosyltransferases with alphaamylase and starch on the surface of saliva-coated hydroxyapatite. Arch. Oral Biol. 41:291–298. Vickerman, M. M., and P. E. Minick. 2002. Genetic analysis of the rgg-gtfG junctional region and its role in Streptococcus gordonii glucosyltransferase activity. Infect. Immun. 70:1703–1714.