Structural and Antigenic Characteristics of Streptococcus sobrinus Glucan Binding Proteins

Abstract

Three purified glucan binding proteins (GBP-2, GBP-3, and GBP-5) from Streptococcus sobrinus 6715 were compared structurally by mass spectroscopy of tryptic fragments and antigenically by Western blot analysis with rat antisera to each GBP or to peptides containing putative glucan binding epitopes of mutans streptococcal glucosyltransferases. Structural and antigenic analyses indicated that GBP-3 and GBP-5 are very similar but that both are essentially unrelated to GBP-2. None of these S. sobrinus GBPs appeared to have a strong antigenic relationship with GBPs from Streptococcus mutans. Thus, S. sobrinus GBP-2 and GBP-3 appear to be distinct proteins with potentially different functions. S. sobrinus GBP-5 may be a proteolytic fragment of GBP-3, or, alternatively, the genes coding for these proteins may be closely related.

The ability of mutans streptococci to bind to glucans synthesized in situ by glucosyltranferases (GTFs) is presumed to be an important feature in the development of dental plaques containing these streptococci (7). Binding may be mediated by cell wall-associated glucan binding proteins (GBPs). Many proteins with glucan binding properties have been identified in Streptococcus mutans and Streptococcus sobrinus. Each GBP has the ability to bind α-1,6-glucan, although other glucan linkages may impart higher binding constants. Three S. sobrinus GBPs have been described. Wu-Yuan and Gill (19) reported that an 87-kDa GBP from S. sobrinus B13 was the major GBP recovered by affinity chromatography. They observed that this S. sobrinus protein had a weak antigenic relationship with an S. mutans GBP. Further, they found an aggregation-deficient mutant which had lost the ability to secrete this protein under the growth conditions utilized, thus implicating the 87-kDa GBP in aggregation phenomena. Ma and coworkers (8) have described an S. sobrinus GBP, identified in strain 6715, which had an apparent molecular mass of approximately 60 kDa. They also identified a mutant S. sobrinus strain which lacked the 60-kDa GBP and was unable to aggregate in the presence of α-1,6-linked glucan, although this strain contained the 87-kDa protein. Landale and McCabe (6) identified a much smaller GBP (a homodimer of 7.5 kDa) from S. sobrinus 6715. This GBP, designated GBP-1, was shown to bind soluble glucan synthesized by S. sobrinus GTF via a site that accommodated eight glucose residues. However, none of these S. sobrinus GBPs have been cloned or sequenced.

S. mutans also secretes several, apparently distinct, GBPs. GBP-A, purified by Russell and coworkers (10) and cloned and sequenced by Banas and coworkers (2), has an apparent molecular mass of 74 kDa by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). GBP-A appears to predominate among the GBPs secreted by S. mutans (13) and has approximately 50% sequence homology with the putative glucan binding regions of GTFs of mutans streptococci (2, 17). A second, somewhat faster-migrating S. mutans protein, GBP-B (59 kDa), has little antigenic relationship with GBP-A by enzyme-linked immunosorbent assay or in Western blots and appears to induce a significant salivary immune response in humans (13). Recently, a third S. mutans GBP (GBP-C), with an apparent molecular mass of 64 kDa by SDS-PAGE, was identified (11). This GBP was detected only when S. mutans cultures were stressed during growth (11). Unlike GBP-A, GBP-C has no sequence similarity to the glucan binding domains of GTFs.

In the course of studying the role of GBPs in the molecular pathogenesis of mutans streptococci, we detected at least three GBPs in S. sobrinus 6715, only one of which (an 87-kDa GBP) had been previously purified (19). The purpose of the present study was to investigate the structural and antigenic relationships of several of these S. sobrinus GBPs (designated GBP-2, GBP-3, and GBP-5 [Table 1]) by mass spectroscopy and Western blot analysis.

TABLE 1.

GBPs from S. sobrinus and S. mutans

Protein	Apparent molecular mass (kDa) by:
	Previously reported SDS-PAGE (reference)	Data from this study
		SDS-PAGE	Mass spectroscopy
S. sobrinusa
GBP-2	87 (19)	82	71
GBP-3	—b	68	52
GBP-4	60 (8)	58	NDc
GBP-5	—	49	41
S. mutans
GBP-A	74 (10)	72	59
GBP-B	59 (13)	58	42
GBP-C	64 (11)	ND	ND

^a The numbering for S. sobrinus GBPs in this table begins at GBP-2 because Landale and McCabe (6) previously designated a small GBP (a homodimer of 7.5 kDa) as GBP-1. The selection of SDS-polyacrylamide gel concentration precluded detection of this smaller GBP in this study. 

^b —, not previously reported. 

^c ND, not done. 

Mutans streptococci used for GBP preparation were grown anaerobically (10% CO₂, 90% N₂) overnight in chemically defined media, as previously described (13). The final cultivation in sealed 16-liter vessels was accomplished in 15 h at 37°C. After neutralization, culture supernatants were obtained by centrifugation, filtered on Durapore GVWP filters (Millipore Corp., Bedford, Mass.), and then incubated for 4 h with Sephadex G100 (S. sobrinus strains) or Sephadex G150 (S. mutans strains). GBPs were eluted from the Sephadex with 3 M guanidine-HCl and concentrated by ultrafiltration, followed by gel filtration on Superose 6 (fast-protein liquid chromatograph; Pharmacia) and ion-exchange chromatography on Mono-Q (Pharmacia). Conditions for affinity chromatography on Sephadex and gel filtration on Superose of GBPs of both mutans streptococcal strains were identical to those previously described for an S. mutans SJ GBP-A and GBP-B preparation (13). The S. sobrinus 6715 GBPs in the GBP-containing Superose pool were separated by ion-exchange chromatography (Mono-Q HR 5/5 column in 0.02 M bis-Tris, 6 M urea, HCl [pH 6.5]) after dialysis against the Tris-urea-HCl buffer. GBPs were eluted from the column with a gradient formed with 0 to 1 M NaCl (elution rate of 1 ml/min with a slope of 0.6 mM NaCl/ml). GBPs were further enriched by rerunning on Mono-Q at one-half of the slope of the initial gradient. The relative concentrations of GBPs eluting in the first Mono-Q run were estimated by planimetry.

Antisera to separated GBPs were prepared by subcutaneous injection of Sprague-Dawley rats with 5 μg of protein in complete Freund adjuvant on day 0 and incomplete Freund adjuvant on day 14. Sera, prepared from blood collected 28 days after the second injection, were stored at −70°C until use in Western blot analysis.

SDS-PAGE of proteins was performed for 1 h at 17 mA/gel on 7% polyacrylamide gels containing 0.01% SDS, with a 4% stacking gel in an air-cooled slab-gel apparatus (Mighty Small; Hoefer Scientific Instruments, San Francisco, Calif.), as previously described (13). For Western blot analysis, SDS-PAGE-separated proteins were electrophoretically transferred to nitrocellulose for 1 h at 200 mA. After blocking, the blotted proteins were incubated with antisera (1:200 final dilution) from rats that had been immunized with GBPs or with the synthetic peptide construct GLU or DREP (14). The GLU peptide sequence, TGAQTIKGQKLYFKANGQQVKG, has significant homology with repeating sequences both in S. mutans GBP-A and in the C-terminal third of all mutans streptococcal GTFs (1–5, 9, 12). The DREP peptide sequence, VTDRYGRISYYDGNSGDQIRN, has significant homology with repeating sequences in the C-terminal third of S. mutans GTF-I (12) (Table 2). Although both peptides may reflect putative glucan binding domain structures (18), they differ in that the GLU domain aromatic pair is YF and the bracketing glycines are eight residues apart, while in the DREP domain the aromatic pair is YY and the glycines are nine residues apart. Furthermore, GLU- and DREP-associated domains in GTFs have far higher homologies within than between the respective domains. After incubation with rat anti-GBP or peptide reagents, membranes were developed for rat immunoglobulin G (IgG) antibody with 1:100 dilutions of biotinylated affinity-purified goat anti-rat γ chain antibody (Zymed). Bands were visualized with streptavidin-horseradish peroxidase (Zymed), followed by addition of a solution of 0.05% 4-chloro-1-naphthol, 16.7% methanol, and 0.015% hydrogen peroxide.

TABLE 2.

Sequence homologies of representative glucan binding domains in S. downei, S. sobrinus, and S. mutans GTFs, S. mutans GBP-A, and the synthetic peptides GLU and DREP

Source (reference)	Amino acid position	Representative sequence	% Homology with GLU peptide	Amino acid position	Representative sequence	% Homology with DREP peptide
S. downei GTF-I (9)	1303	TGAQTIKGQKLYFRANGQQVKG	100	1400	VTGNDGKLRYYDANSGDQAFN	57
S. sobrinus GTF-I (1)	1298	TGAQTIRGQKLYFKANGQQYEG	86	1436	VTGTDGKVRYYDANSGDGAFN	48
S. sobrinus GTF-S (3)	1234	TGEQTIDGQKVFFQDNGVQVKG	82	1117	VWYDGKKAYYYDDNGRTWTNK	29
S. mutans GTF-I (12)	1289	TGARTINGQLLYFRANGVQVKG	77	1324	VTDRYGRISYYDGNSGDQIRN	100
S. mutans GTF-SI (12)	1264	TGTVTFNGQRLYFKPNGVQAKG	73	1287	IRDANGYLRYYDPNSGNEVRN	48
S. mutans GTF-S (4)	1267	VGVQTINGKTYYFGQDGKQIKG	50	1314	ATTDSQNNQYYFGSDGVAVTG	24
S. mutans GBP-A (2)	274	IGWRTIGLL.YYFDTNGVQVEG	50	388	DIAERDGKVYYLDEDSGQVVK	19
Synthetic peptides
GLU		TGAQTIKGQKLYFKANGQQVKG
DREP					VTDRYGRISYYDGNSGDQIRN

Molecular masses estimated from Coomassie blue-stained SDS-polyacrylamide gels were calculated with 44- to 205-kDa nonbiotinylated or biotinylated protein standards (Sigma Chemical Co., and Bio-Rad Laboratories, respectively) by using the Southern hyperbola equation (15). The molecular masses of the standards were plotted on the x axis against their mobilities (R_f) on the y axis. The estimated molecular masses of the unknowns were obtained by least-squares fit with Sigma Plot 4.0 for Windows 95 (SPSS Inc., Chicago, Ill.). A standardized nomenclature for S. sobrinus GBPs is proposed in Table 1 (e.g., GBP-2) and used herein.

Mass spectroscopy of intact S. sobrinus 6715 GBPs and peptides, prepared by overnight tryptic digestion (0.02 μg/μl; Promega), was conducted at the Molecular Biology Core Facility of Dana-Farber Cancer Institute with a PerSeptive Biosystems Voyager-DE-STR mass spectrometer. Analyses were performed by the matrix-assisted laser desorption ionization time of flight method. Protein samples to be digested were first enriched by the methods described above, followed by SDS-PAGE of each GBP in a separate lane. Gel slices containing GBPs were then removed and processed as described by William et al. (16) to provide tryptic fragments for spectroscopy. Tryptic digestion and mass spectroscopy of bovine serum albumin was also performed, for control purposes.

Several prominent GBPs (GBP-2, 82 kDa; GBP-3, 68 kDa; GBP-5, 49 kDa) can be routinely observed in SDS-PAGE of GBP mixtures from S. sobrinus 6715 or DS (Fig. 1, lanes 5 and 6) after elution from Sephadex G150. These S. sobrinus 6715 GBPs were separated by a combination of gel filtration in 6 M guanidine-HCl and ion-exchange chromatography on Mono-Q (Fig. 2). S. sobrinus 6715 GBP-2, previously separated by similar techniques from S. sobrinus B13 and identified as an 87-kDa protein by Wu-Yuan and Gill (19), eluted at a salt concentration of 0.18 M and was the most prominent GBP in the Mono-Q elution profile (Fig. 2). A second prominent GBP, GBP-3, eluted at a salt concentration of 0.15 M and migrated to a position of 68 kDa in SDS-PAGE (Fig. 3). This GBP-3 had not previously been described. A third GBP (GBP-5), also not previously described, eluted at a salt concentration of 0.11 M and migrated to a position of 49 kDa in SDS-PAGE (Fig. 3). A component whose migration characteristics (58 to 60 kDa) corresponded to previously described S. sobrinus GBP-4 (8) was detected in S. sobrinus 6715 but not in strain DS (Fig. 1). This GBP could not be purified from other GBP components.

FIG. 1.

Silver-stained SDS–7% polyacrylamide gel of guanidine eluates and purified GBPs from mutans streptococci. Lanes 1 and 2 contain enriched preparations of S. mutans GBP-A (0.5 μg [lane 1]) and GBP-B (0.4 μg [lane 2]). Lanes 3 to 6 contain GBP mixtures from S. mutans Ingbritt, S. mutans SJ, S. sobrinus 6715, and S. sobrinus DS, respectively. These GBP mixtures were obtained by 6 M guanidine elution of GBPs from Sephadex after incubation with culture supernatant, as described in the text. The relative migration positions of S. sobrinus GBP-2 to GBP-5 and GTF are indicated by the labeled arrows on the right side of the gel. The molecular masses (in kilodaltons) of coelectrophoresed standards are indicated on the left side of the gel.

FIG. 2.

Purification of S. sobrinus 6715 GBPs. (A) A₂₈₀ elution profile of GBPs eluted from Sephadex G100 by 3 M guanidine-HCl and separated with Superose 6 in 6 M guanidine-HCl. GTF enzymatic activity (release of total reducing sugars) is illustrated by the open circles, which indicate A₅₄₀ after the Somogyi assay. (B) Elution profile of GBPs in the shaded Superose 6 peak (arrow [A]) after ion-exchange chromatography in a Mono-Q 5/5 column. The NaCl gradient used to elute the GBPs is indicated by the dotted line connecting the open circles. The molecular mass characteristics of GBP-2, GBP-3, and GBP-5 are given in Table 1.

FIG. 3.

Coomassie blue-stained SDS–7% polyacrylamide gel of chromatographically purified GBPs from S. sobrinus 6715. In lane 1 is the GBP pool (2 μg) from Superose 6 gel filtration (Figure 2A). In lanes 2, 3, and 4 are the GBP-5-, GBP-3-, and GBP-2-containing pools, respectively, from Mono-Q ion-exchange chromatography (Figure 2B). Lanes 2, 3, and 4 contain 0.1 μg (each) of protein. The migration positions of the purified S. sobrinus GBPs are indicated by the labeled arrows on the right side of the gel. No other components were detected in the GBP-2 (lane 4) and GBP-5 (lane 2) pools, although a trace of GBP-2 can be detected in the GBP-3 pool (lane 3). The estimated molecular masses (in kilodaltons) of coelectrophoresed standards are indicated on the left side of the gel.

The physical relationships among GBP-2, GBP-3, and GBP-5 from S. sobrinus 6715 were examined. Gel slices of chromatographically enriched GBPs, after SDS-PAGE (Fig. 3), were digested by trypsin, followed by mass spectroscopy, to determine the molecular masses of the released peptide fragments. This analysis revealed that approximately 75% of the released fragments of GBP-3 were shared by GBP-5, indicating that much of the primary sequences of these GBPs were identical (Table 3). The tryptic fragment profile of GBP-2, on the other hand, bore little resemblance to either GBP-3 or GBP-5, since only one GBP-2 peptide had a mass similar to that of the other GBPs (GBP-3). Thus, the distinctiveness of GBP-3 tryptic peptides from those of GBP-2 indicated that S. sobrinus 6715 GBP-3 is a unique protein and not simply a proteolytic fragment of the larger GBP-2.

TABLE 3.

Tryptic peptides in GTF, GBP-2, GBP-3, and GBP-5

Protein	No. of tryptic peptidesa
	GTF	GBP-2	GBP-3	GBP-5
GTF	20	1	0	0
GBP-2		14	0	1
GBP-3			19	14
GBP-5				20

^a Total number of tryptic peptides in each protein is shown in boldface type; numbers of peptides with identical molecular masses shared by different proteins are shown in standard type. 

Antigenic relationships among S. sobrinus GBPs were examined in Western blotted S. sobrinus 6715 GBP mixtures eluted from Sephadex (Fig. 4). Despite the presence of several GBPs and GTFs in this mixture (Fig. 1), the antiserum to GBP-2 reacted only with GBP-2. Antisera to GBP-3 and GBP-5 showed little or no reactivity with GBP-2 but showed strong and nearly indistinguishable reactions with GBP-3 and GBP-5 in the unseparated S. sobrinus GBP guanidine eluate. The minor antibody reaction seen with GBP-2 in the anti-GBP-3 serum (Fig. 4, lane 8) is likely to be the result of a response induced to the trace of GBP-2 in the GBP-3 pool (Fig. 3). Similarly, GBP-2, but not GBP-3 or GBP-5, reacted with IgG antibody to the GLU synthetic peptide, whose sequence contained a putative glucan binding domain (Fig. 5). Antisera to the DREP synthetic peptide did not react with these S. sobrinus GBPs or GTFs (not shown), presumably because the DREP peptide construct has greater homology with S. mutans than with S. sobrinus GTF sequences. None of the rat antisera induced by injection with enriched preparations of S. sobrinus 6715 GBP-2, GBP-3, or GBP-5 reacted with glucan binding components in unseparated GBP guanidine eluates of S. mutans (Fig. 4).

FIG. 4.

Western blot of S. mutans SJ GBPs (left) and S. sobrinus 6715 GBPs (right) eluted from Sephadex G100 after electrophoresis on an SDS–7% polyacrylamide gel. Blots were developed for IgG in sera (1:200) from rats injected with the enriched GBP pools shown in Fig. 3. Antisera to GBP-5, GBP-3, or GBP-2 were added to lanes 2 to 4 and 7 to 9, respectively. An antiserum to a sham-immunized rat was added to lanes 1 and 6. The molecular masses of protein standards (lane 5) are indicated (in kilodaltons) on the left of the blot.

FIG. 5.

Western blot of transblotted SDS–7% polyacrylamide gels developed for IgG in serum (1:50) from a rat injected with the GLU peptide construct. Lanes 1 and 2 contain previously electrophoresed S. sobrinus 6715 and S. sobrinus DS guanidine eluates, respectively. Lane 3 contains previously electrophoresed enriched S. sobrinus GBP-2 (0.6 μg). Lane 4 contains molecular standards whose estimated masses (in kilodaltons) are indicated to the right of the blot. The stained S. sobrinus components migrating between the 205- and 116-kDa positions reflect the migration of GTF (see Fig. 1). The stained component migrating to approximately 100 kDa is likely to be degraded GTF.

The relationship between the S. sobrinus 6715 GBPs and GTF, which also has glucan binding activity, is unclear. None of the GBPs appear to have significant structural relationships with S. sobrinus GTF, since only one tryptic peptide of similar mass (Table 3) was shared between GTF and any GBP (GBP-2). This observation would also indicate that these three GBPs were not created by proteolytic digestion of GTF. Interestingly, however, antisera to GBP-3 and to GBP-5, as well as anti-GLU peptide antibody, appeared to bind GTF in Western blotting (Fig. 4 and 5, respectively). On the other hand, the anti-GBP-2 reagent showed no reaction with S. sobrinus GTF despite its relatively high concentration in this preparation (Fig. 1). Molecular biological approaches, now under way, should help to clarify these relationships.

Taken together, these results indicate that GBP-2 and the newly described GBP-3 and GBP-5 appear to be the most prominent GBPs secreted by S. sobrinus strains. Under our enrichment conditions, GBP-4, previously described by Ma et al. (8), was a minor component in the two S. sobrinus strains used in this study. The fastest migrating GBP, GBP-5, bears strong antigenic and structural resemblances to GBP-3 and thus may be a proteolytic derivative of the larger GBP, although several tryptic peptides were unique to GBP-5. None of these S. sobrinus GBPs appear to have a strong antigenic relationship with GBPs from S. mutans.

The functional roles of each of these constitutively secreted components are still unresolved. All have the ability to bind glucan. Mutant strains of S. sobrinus, altered in the ability to aggregate with glucan, have been shown to be deficient in GBP-2 (19) or GBP-4 (8). However, the nature and extent of these mutations were unclear. Identification and modification of individual genes responsible for the synthesis of each GBP may help to reveal their participation in mechanisms leading to the accumulation of mutans streptococci on dental surfaces.