INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Glucosyltransferases (GTFs) are extracellular enzymes which, through synthesizing glucans from sucrose, are involved in the attachment and accumulation of Streptococcus mutans on the tooth surface. S. mutans possesses three distinct genes encoding three different GTFs that produce water-soluble or -insoluble glucans (1, 7, 24, 31). Production of glucans, especially water-insoluble ones, is crucial for the development of smooth-surface carious lesions in animal experiments comparing GTF-deficient isogenic mutants of S. mutans to those of parental organisms (21, 33). GTF has two functional domains, i.e., an N-terminal catalytic sucrose-binding domain involved in sucrose hydrolysis and a C-terminal glucan-binding domain involved in the binding of the synthesized glucan polymer and presumably chain extension of the growing glucan polymers (11, 19, 20, 32).

An important application of studying the molecular pathogenesis of dental caries and identifying virulence factors of S. mutans is to develop a mucosal subunit vaccine which would inhibit these factors by inducing substantial levels of secretory immunoglobulin A (IgA) antibodies in saliva. Antibodies to GTF, for example, would be expected to inhibit glucan synthesis and consequently reduce the caries activity by preventing S. mutans accumulation on the tooth surface. The feasibility of this objective has been demonstrated by oral immunization experiments with GTF in animal models of dental caries (26). For vaccine development, the use of recombinant polypeptides representing protective epitopes within the GTF molecule should maximize the specificity and effectiveness of the induced salivary IgA response. We have previously subcloned the putative catalytic and glucan-binding regions (CAT and GLU, respectively) of S. mutans GTF-I and demonstrated that antibodies to either CAT (representing amino acid residues 253 to 628) or GLU (representing amino acid residues 1183 to 1473) could inhibit glucan synthesis by GTF, although anti-GLU antibodies were much more effective than anti-CAT antibodies (75% versus 22% inhibition, respectively) (10). These findings support earlier observations that antibodies to peptides corresponding to sequences within the CAT or GLU caused a moderate inhibition of GTF activity (3, 5, 16, 27, 28). Furthermore, subcutaneous immunizations in the salivary gland vicinity of rats with synthetic peptides representing components of the glucan-binding or catalytic region of GTF moderately reduced smooth-surface caries (30). Although antibody responses to selected peptides are expected to have even greater specificity against the catalytic or glucan-binding functions of GTF than antibodies to the whole putative domain, antibody responses to the latter should theoretically be less genetically restricted in human vaccinees.

We have previously found that intranasal (i.n.) immunization of mice with GLU alone, isolated from inclusion bodies found in the cytoplasmic fraction, could induce a moderate salivary IgA response (10). Aiming at increasing the solubility of GLU as well as enhancing its mucosal immunogenicity, we genetically linked GLU with thioredoxin. When fused to the N terminus of the polypeptide of interest, thioredoxin from Escherichia coli has previously been shown to enhance the solubility of polypeptides expressed recombinantly, and the linked proteins have been shown to preserve their biological activity (8, 17, 22). Furthermore, bacterial thioredoxin has been shown to enhance proliferation of murine T cells due to a thiol-related reducing capacity (2). It might be speculated that antigens presented as thioredoxin chimeras could induce T-cell proliferation and augment a specific antibody response.

In this study, we present the construction, expression, and purification of a thioredoxin-GLU chimeric protein (Thio-GLU). The immunogenic properties of this polypeptide were compared to those of the GLU polypeptide alone by immunization of mice via the i.n. route. Moreover, both constructs were evaluated for their ability to induce a protective salivary IgA response against S. mutans in a mouse infection model.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Genetic construction. A DNA fragment encoding the glucan-binding domain of gtfB was previously cloned into the expression vector pET20b(+) Novagen, Madison, Wis.), and the construct was named pET20b(+)-GLU (10). The DNA fragment encoding GLU was separated from vector sequences by restriction enzyme digestions with NcoI and XhoI, followed by gel electrophoresis and purification with the QIAEX gel extraction kit (Qiagen, Chatsworth, Calif.). The fragment was subcloned into the pET32b(+) expression vector in frame with the 3' end of a gene sequence encoding a 17.1-kDa thioredoxin fragment containing His-tag and S-tag sequences for ease of purification (Novagen), and the construct was named pET32b(+)-GLU (Fig. 1A). Both constructs [pET20b(+)-GLU and pET32b(+)-GLU] were electroporated into E. coli BL21(DE3), containing a genomic source of T7 RNA polymerase under lacUV5 promoter control, and transformed colonies were selected on Luria-Bertani agar plates (1% tryptone, 0.5% yeast extract, 1% NaCl, 0.1% dextrose, 1.8% agar) containing 50 µg of carbenicillin per ml as the selection for the plasmids. The presence of plasmids with sizes of 4.5 kb [pET20b(+)-GLU] or 6.8 kb [pET32b(+)-GLU] was confirmed by gel electrophoresis of plasmid preparations made by using the Wizard Miniprep DNA Purification System (Promega, Madison, Wis.).

View larger version (26K): [in this window] [in a new window]   FIG. 1.   (A) Maps of the plasmids used for electroporation of E. coli BL21(DE3) expressing the recombinant proteins GLU and Thio-GLU. (B) Coomassie blue stain of sodium dodecyl sulfate-polyacrylamide gel electrophoresis (3% polyacrylamide stacking gel and 12% polyacrylamide separation gel) of the recombinant Thio-GLU (lane 1) and GLU (lane 2) proteins purified by nickel column affinity chromatography. Molecular mass markers in kilodaltons are indicated on the left.

Protein expression and purification. An overnight culture of E. coli BL21(DE3) containing either pET20b(+)-GLU or pET32b(+)-GLU was diluted 1:15 and grown to mid-log phase at 30°C (approximately 2 h). The culture was induced by 1 mM IPTG (isopropyl--D-thiogalactopyranoside) and grown for an additional 3-h period. Soluble proteins were recovered by resuspending pelleted cells in binding buffer (0.5 M NaCl, 20 mM Tris-HCl [pH 7.9], 5 mM imidazole) and stored for 1 h at 70°C. Following thawing, the cell-lysate was sonicated twice for 10 s. Inclusion bodies containing insoluble GLU or Thio-GLU were recovered by centrifugation and solubilized in 6 M guanidine-HCl-0.1 M NaH₂PO₄-1 mM Tris-HCl (pH 8.0) by stirring at room temperature for 4 h. The lysate was again sonicated twice for 10 s before it was clarified by centrifugation and loaded on a precharged and equilibrated His-Bind Resin column (Novagen) (14). Briefly, the cell lysate was incubated with the His-Bind Resin column at 4°C overnight, and unbound protein was passed through the column by gravity. The column-bound protein was washed with 5 column volumes of 8 M urea-0.1 M NaH₂PO₄-1 mM Tris-HCl (pH 8.0), followed by 3 column volumes of 8 M urea-0.1 M NaH₂PO₄-1 mM Tris-HCl (pH 6.3). The protein was refolded by gradually lowering the initial urea content in 1 M steps of the refolding buffer (8 M urea, 0.5 M NaCl, 10 mM Tris-HCl, 20% glycerol [pH 7.4]) and then eluted by 0.25 M imidazole in refolding buffer (without urea). Finally, the recovered proteins, GLU or Thio-GLU, were dialyzed against 50 mM Tris-HCl (pH 7.9)-0.5 M NaCl-10% glycerol before being used for immunization.

Immunization protocol. Groups of six BALB/c mice, 8 to 10 weeks of age, were used for i.n. immunization with either 50 µg of purified GLU, 75.6 µg of Thio-GLU (corresponding to 25.6 µg of thioredoxin and 50 µg of GLU), or buffer (50 mM Tris-HCl [pH 7.9], 0.5 M NaCl, 10% glycerol). Each dose volume did not exceed 20 µl and was applied slowly to both nares. The immunizations were given on days 1, 14, and 28, and saliva, blood, and vaginal wash samples were collected prior to immunization and on days 13, 27, 41, 83, and 97 (Fig. 2), as previously described (10). Briefly, saliva samples were collected after stimulation of the salivary flow by intraperitoneal injection of 5 µg of carbachol (Sigma Chemical Co., St. Louis, Mo.), and the samples were stored at 70°C and centrifuged prior to enzyme-linked immunosorbent assay. The serum was obtained after centrifugation of blood samples collected by heparinized capillary pipettes from the retroorbital plexus. The vaginal wash samples were collected by flushing the vagina twice with 50 µl of phosphate-buffered saline. The animal experiments were performed according to National Institutes of Health guidelines and protocols approved by the University of Alabama at Birmingham Institutional Animal Care and Use Committee.

View larger version (7K): [in this window] [in a new window]   FIG. 2.   Diagram showing the immunization, infection, and sampling protocols used to study the protective effect of salivary antibodies against the glucan-binding region of S. mutans GTF in an experimental mouse model. Immunization was by the i.n. route with 50 µg of GLU or 75.6 µg of Thio-GLU. , antibiotic treatment for 5 consecutive days; , oral infection for 5 consecutive days with 2 × 109 S. mutans PC3379 cells; , collection of blood, saliva, and vaginal wash samples; , oral swab of infected mice; , mice sacrificed and mandibular molars scored for caries activity.

Quantitation of antibody responses. The levels of specific antibodies in samples were determined by enzyme-linked immunosorbent assay on Maxisorp microtiter plates (Nunc, Roskilde, Denmark) coated with 1 µg of GLU per ml, as previously described (10). Plates were blocked for 2 h at room temperature with 5% fetal calf serum in 0.01 M phosphate buffer (pH 7.2) containing 0.5 M NaCl and 0.15% Tween 20. The total concentration of IgA in saliva and vaginal washes was determined on microtiter plates coated with antibodies to mouse IgA. Peroxidase-labeled anti-mouse IgA or IgG antibodies were used as detection reagents, followed by o-phenylenediamine substrate with H₂O₂. The detecting and coating antibodies used in this study were purchased from Southern Biotechnology Associates, Inc., Birmingham, Ala. The levels of anti-GLU antibody and total IgA levels in the samples were calculated by interpolation on standard curves generated with a mouse Ig reference serum (ICN Biomedicals, Aurora, Ohio). Data were logarithmically transformed and statistical analysis (one-way analysis of variance in conjunction with the Tukey multiple-comparisons test) was performed by using the InStat program (Graphpad Software, San Diego, Calif.).

Infection protocol. Prior to infection, the immunized mice were treated for 4 consecutive days (starting at day 44) with 4,000 U of penicillin per ml in the drinking water and for 5 consecutive days with 4 mg of tetracycline and erythromycin per g of diet to temporarily suppress the oral flora to facilitate S. mutans infection. The diet at this time point, i.e., experimental day 44, was switched to a cariogenic powdered diet 301 containing 1% sucrose (18). One day after termination of the antibiotic treatment, the mice were checked for the suppression of the normal oral flora by swabbing two animals from each group with type 4 Calgiswabs (Spectrum, Houston, Tex.). The swabs were subsequently solubilized in 500 µl of saline by vortexing, and 100 µl was plated on 5% blood agar plates. The mice were orally infected from days 51 to 55 with 2 × 10⁹ CFU of tetracycline- and erythromycin-resistant S. mutans PC3379 by applying swabs presoaked with the bacterial solution. The bacteria had been grown overnight anaerobically in Todd-Hewitt broth containing 10 µg of tetracycline and erythromycin per ml. S. mutans PC3379 (provided by P. J. Crowley and A. S. Bleiweis, Gainesville, Fla.) is a spaP-complemented strain constructed from the spaP mutant strain PC3370 (4) and shown to be virulent in a gnotobiotic rat model (unpublished findings).

S. mutans colonization and caries assessment. S. mutans infection levels were assessed by swabbing all animals at weekly intervals for 5 weeks. The swabs were solubilized in 500 µl of saline by vortexing, and 10-fold dilutions were plated on Mitis-Salivarius agar plates with or without (to assess total oral streptococci levels) 10 µg of tetracycline and erythromycin per ml. The plates were incubated at 37°C under anaerobic conditions for 48 h, and colonies were counted by using a Bausch and Lomb microscope and a Spiral System magnifier. At day 120, the experiment was terminated, mice were sacrificed, and the mouse mandibles were removed, cleaned, and stained with murexide. The teeth were then sectioned, and the caries level was determined by the Keyes method (13) as previously described (12).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Recombinant protein expression. We have previously found that GLU is expressed in the insoluble phase in E. coli (10). The fusion of the thioredoxin subunit to GLU increased its solubility, although the chimera was still predominantly expressed in inclusion bodies. GLU and Thio-GLU polypeptides were therefore purified from the insoluble phase by means of His-tag affinity chromatography on a nickel-charged column, with a yield of approximately 12 and 15 mg of protein per liter of culture, respectively. The purity of the eluate was confirmed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Fig. 1B), and the polypeptides were used in mouse immunization experiments.

Immune responses. Evaluation of serum antibody responses in mice after i.n. immunization with GLU or Thio-GLU showed significantly enhanced levels of IgG anti-GLU starting on day 27 (P < 0.01) and continuing throughout the experiment when compared to sham-immunized mice (Fig. 3A). The specific serum IgA anti-GLU antibody response in GLU- or Thio-GLU-immunized mice reached statistical significance compared to that of sham-immunized animals at day 41 (P < 0.01) and persisted at even higher levels throughout the experiment (Fig. 3B).

View larger version (16K): [in this window] [in a new window]   FIG. 3.   Serum IgG (A) and IgA (B) anti-GLU responses. Results are expressed as the geometric means of the specific Ig responses multiplied or divided by the standard deviations of six mice given GLU, Thio-GLU, or buffer only by the i.n. route. The asterisks indicate antibody responses significantly different from these of sham-immunized animals (P < 0.01). The arrows indicate days of immunization, and the bars indicate 5 consecutive days of S. mutans infection.

View larger version (19K): [in this window] [in a new window]   FIG. 4.   Salivary IgA (A) and vaginal IgA (B) anti-GLU responses. Results are expressed as the geometric means multiplied or divided by the standard deviations of six mice given GLU, Thio-GLU, or buffer only by the i.n. route. The asterisks indicate antibody responses significantly different from those of sham-immunized animals (P < 0.01). The arrows indicate days of immunization, and the bars indicate 5 consecutive days of S. mutans infection.

Inhibition of S. mutans colonization. Following challenge with S. mutans PC3379 on days 51 to 55, control animals (given buffer only by the i.n. route) maintained high levels of S. mutans (~60% of total oral streptococci) throughout the study (Fig. 5). In contrast, GLU- and Thio-GLU-immunized mice showed significant reductions in S. mutans levels over time. The Thio-GLU-immunized group exhibited a pronounced 86% decrease (P < 0.05) at week 3, whereas the GLU-immunized group showed a maximum of a 52% decrease at week 5 (P < 0.05). Moreover, Thio-GLU-immunized animals displayed significantly lower levels of S. mutans (P < 0.05) than GLU-immunized animals at week 3 but not at other time points.

Caries protection. GLU- or Thio-GLU-immunized mice displayed significantly fewer buccal enamel (Fig. 6A) or dentinal (Fig. 6B) lesions (P < 0.01) than sham-immunized animals. The immunized groups had, in general, similar caries activity on proximal and sulcal surfaces when compared to unimmunized animals. The pronounced protective effect of anti-GLU immunizations on buccal lesions is consistent with the role of glucans in enhancing S. mutans colonization on the smooth surfaces of the teeth.

View larger version (13K): [in this window] [in a new window]   FIG. 5.   The percentage of S. mutans PC3379 per total oral streptococci in the oral cavity of immunized or control mice which were challenged with 2 × 109 CFU of the PC3379 strain for 5 consecutive days. The level of infection is presented as the geometric means of six mice.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

In the present study, we examined the ability of Thio-GLU or the GLU polypeptide alone to stimulate immunity against S. mutans-induced dental caries. Mice immunized with GLU or Thio-GLU by the i.n. route exhibited significantly elevated levels of specific salivary IgA antibodies, in addition to serum IgA, IgG, and vaginal IgA antibodies, in comparison to sham-immunized animals. Moreover, GLU- and Thio-GLU-immunized animals showed significant inhibition of S. mutans colonization and buccal enamel and dentinal lesion development, in contrast to control animals, which maintained high levels of S. mutans throughout the study and displayed significantly more buccal lesions.

The finding that GLU in the absence of any immunostimulating agent was capable of stimulating protective immunity via a mucosal route was somewhat surprising since protein antigens alone usually are poor mucosal immunogens. For example, the saliva-binding region of S. mutans surface antigen I/II was found to require fusion of cholera toxin (CT) A2/B subunits (CTA2/B) (6) or coadministration with monophosphoryl lipid A (18) to induce substantial mucosal or serum responses. In contrast, coadministration of GLU with CT or monophosphoryl lipid A did not further enhance antibody responses to GLU (our unpublished observations). It is possible that the GLU preparation might be in particulate form, due to a propensity of GLU to self-aggregate, which may have increased its ability to be taken up into mucosal inductive sites and subsequently by antigen-presenting cells. It is noteworthy that in this study the specific serum IgA response approached 10 µg/ml. In our experience with other protein antigens, this level of response requires the use of a potent adjuvant such as CT. Similarly, the fusion of thioredoxin to GLU did not significantly influence anti-GLU antibody responses, although salivary IgA responses were slightly elevated in the Thio-GLU-immunized mice compared to those of mice immunized with GLU alone. This finding might account for the relatively earlier and more intensive clearance of S. mutans in mice immunized with Thio-GLU, since the amount of antibody needed for protection has not been established. However, the lack of consistent statistical significance between the two experimental groups does not allow for claiming a conclusive immunoenhancing role for thioredoxin in the chimeric protein.

The difference in infection levels between GLU- and Thio-GLU-immunized animals was not completely reflected in the dental caries scores, since animals were kept on the cariogenic diet for an additional 5 weeks after the last assessment of the infection level to allow for substantial development of tooth decay. We therefore do not know whether differences in S. mutans colonization levels between the two groups at the 5-week time point (Fig. 5) are representative of later time points. There was an increase in S. mutans levels in Thio-GLU-immunized animals from week 3 to week 5 which may be related to a decrease in salivary IgA antibody levels. Likewise, levels of specific salivary antibody responses in GLU-immunized animals continued to increase throughout the time course of this study, as reflected in the decrease of S. mutans infection levels during the same time period.

View larger version (23K): [in this window] [in a new window]   FIG. 6.   Caries activity on buccal, proximal, and sulcal molar surfaces involving enamel (A) and slight dentinal (B) lesions. Values are the mean caries scores ± the standard deviations of six mice. The asterisks show caries levels which are significantly different (P < 0.01) from those of infected and unimmunized control animals.

The demonstration that caries protection was more pronounced on buccal than on sulcal or proximal surfaces was not surprising based on previous findings of others (23, 25). In studies by Taubman et al. (30) it was shown that serum antibodies to a 22-mer GLU synthetic peptide inhibited water-soluble glucan synthesis by S. mutans and Streptococcus sobrinus GTFs, whereas water-insoluble glucan synthesis only by S. sobrinus GTF was inhibited. The evidence that smooth-surface caries in animals immunized with the 22-mer GLU synthetic peptide was reduced to a greater extent after infection with S. sobrinus than S. mutans supports the importance of inhibiting water-insoluble glucan synthesis for protection. The water-insoluble glucan synthesis is believed to be especially important for S. mutans colonization of smooth surfaces where tenacious attachment is imperative for successfully resisting the flow of saliva and other ordinary mechanical forces (9, 15), whereas it may not be as important for colonization of retentive areas, such as proximal or sulcal areas, where streptococci can be passively trapped. Through their ability to inhibit GTF, salivary IgA anti-GTF antibodies are therefore thought to have a greater influence on smooth-surface caries.

Taken together, our data strongly suggest that specific salivary antibodies against the glucan-binding region of GTF prevent colonization of S. mutans in the oral cavity of mice fed a cariogenic diet. Since complete protection against S. mutans infection was not seen, it could be advantageous to combine different streptococcal antigens and thereby block different stages of the bacterial attachment and accumulation process. One apparent approach would be to combine the saliva-binding region of Ag I/II, which is involved in the initial attachment of S. mutans to the pellicle on the tooth surface, and GLU of GTF, which is involved in the production of glucans responsible for the subsequent accumulation process of S. mutans; such initiatives are underway (29, 34). Further studies in humans will also evaluate the properties of GLU as a mucosal vaccine against S. mutans-induced dental caries.