Expression and Immunogenicity of Hemagglutinin A from Porphyromonas gingivalis in an Avirulent Salmonella enterica Serovar Typhimurium Vaccine Strain

Abstract

Porphyromonas gingivalis is a major etiologic agent of periodontitis, a chronic inflammatory disease that ultimately results in the loss of the supporting tissues of the teeth. Previous work has demonstrated the usefulness of avirulent Salmonella enterica serovar Typhimurium strains as antigen delivery systems for protective antigens of pathogens that colonize or cross mucosal surfaces. In this study, we constructed and characterized a recombinant S. enterica serovar Typhimurium avirulent vaccine strain which expresses hemagglutinin A and carries no antibiotic resistance markers. HagA, a major virulence-associated surface protein, is a potentially useful immunogen that contains an antigenic epitope which, in humans, elicits an immune response that is protective against subsequent colonization by P. gingivalis. The hagA gene, including its promoter, was cloned into a balanced-lethal Salmonella vector and transferred to the vaccine strain. Heterologous expression of HagA was demonstrated in both Escherichia coli JM109 and S. enterica serovar Typhimurium vaccine strain χ4072. The HagA epitope was present in its native configuration as determined by immunochemistry and immunoelectron microscopy. Purified recombinant HagA was recognized by sera from mice immunized with the S. enterica serovar Typhimurium vaccine strain. The HagA-specific antigen of the vaccine was also found to be recognized by serum from a periodontal patient. This vaccine strain, which expresses the functional hemagglutinin protein, induces a humoral immune response against HagA and may be useful for developing a protective vaccine against periodontal diseases associated with P. gingivalis.

Porphyromonas gingivalis is considered a major etiologic agent of adult and refractory periodontal disease. Hemagglutinins are bacterial surface proteins that often function as adhesins by which bacteria attach to host cells (8). Adherence to host cells is required for virulence of mucosal pathogens. Consequently, prevention of or interference with adherence of a particular bacterial pathogen by molecules such as antibodies to the adhesin prevent colonization and disease (33, 43). For example, multiple MAbs against the F41 adhesive fimbrial antigen of enterotoxigenic Escherichia coli (ETEC) protected animals against a challenge with F41-positive ETEC (56).

Multiple hemagglutinin genes have been cloned from P. gingivalis by functional screening (38, 50, 51). One of these, the gene coding for hemagglutinin A from P. gingivalis, has been isolated and shown to contain four large direct repeats (25). When a P. gingivalis expression library was screened for clones which bind human oral epithelial cells, all positive clones were found to have DNA homology to hemagglutinin A (16). Thus, an immune response to HagA or other hemagglutinins might prevent the colonization of P. gingivalis by inhibiting its adherence to oral tissues.

Vaccination against a disease may have both prophylactic and therapeutic value. Immunization with a vaccine containing killed P. gingivalis suppresses the progress of experimental periodontitis in Macaca fascicularis, suggesting that immunization against P. gingivalis may be an effective means of controlling the disease (57). Unfortunately, vaccines based on killed bacteria can cause toxic reactions (42). Subunit vaccines may reduce the problems associated with inactivated bacterial particles because of their defined chemical and physical properties. However, the production of adhesins for subunit vaccines is often difficult due to contamination with other virulence factors during the tedious process of purification. Other potential limitations include low levels of immunogenicity and failure to induce the desired type of immune response compared with natural infection (45).

Oral administration of vaccines induces a secretory immunoglobulin A (IgA) response upon absorption of the antigen by the gut-associated lymphoid tissue (GALT) (41). The most successful vaccines developed against intracellular bacteria have been based on replication-competent, avirulent or attenuated bacteria such as the BCG strain of Mycobacterium bovis (19) and most of all Salmonella enterica serovar Typhimurium (9). Salmonella is an effective antigen delivery system to the GALT, which initiates production of specific secretory immunoglobulin A for protection against pathogens that colonize or cross mucosal surfaces to initiate infection. This has been established as an effective means of stimulating significant levels of specific mucosal secretory immunoglobulin A directed against a variety of heterologous antigens and has also been shown to stimulate the production of serum antibodies and cell-mediated responses (7).

Salmonella vaccine strains expressing a streptococcal adhesin (24, 59), Listeria extracellular proteins (13, 28–30), a Leishmania surface glycoprotein (40), the Campylobacter immunoreactive transport protein (48), an Entamoeba protective antigen (60), the hepatitis B virus core antigen (54), the Bordetella pertussis filamentous hemagglutinin (23), and P. gingivalis hemagglutinin B (17) have been constructed. In this study, we sought to obtain and characterize an avirulent S. enterica serovar Typhimurium vaccine strain which expresses another potentially useful immunogen of P. gingivalis that may confer functional protection from periodontal tissue destruction induced by P. gingivalis. Heterologous expression of hemagglutinin A was obtained in both E. coli and the S. enterica serovar Typhimurium avirulent vaccine strain χ4072.

MATERIALS AND METHODS

Bacterial strains, plasmids, cell lines, and media.

S. enterica serovar Typhimurium χ4072 SR-11 (Δcya Δcrp Δasd) was used as the vaccine strain. The plasmid expression vector pYA292 is the vector component of this balanced-lethal system (21). A single copy of the asd gene per cell is sufficient for normal growth of S. enterica serovar Typhimurium Δasd, allowing the plasmid to be present in low copy numbers. E. coli χ6097 (F⁻ Δ[lac-pro]rpsL ΔasdA4[zhf-2::Tn10]thi φ80dlacZΔM15), also with asd deleted, was used as a cloning host for pYA292-based constructs (44). E. coli JM109 [recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 Δ(lac-proAB) (F′ traD36 proAB lacI^qZΔM15)] and E. coli DH5α (Life Technologies, Gaithersburg, Md.) were used for other routine cloning procedures. Plasmid pHagA2 contains 8,585 bp of the hagA gene (25) cloned into the XbaI-SacI sites of pBluescript II(+). Plasmid pYA292, E. coli χ6097, and S. enterica serovar Typhimurium χ4072 were kindly provided by Roy Curtiss III, Washington University, St. Louis, Mo.

Both S. enterica serovar Typhimurium and E. coli were cultured aerobically on Luria-Bertani medium (53) or on plates solidified with 1% agar, with the addition of dl-α,ɛ-diaminopimelic acid (50 μg/ml) for the plasmid-free Δasd strains. pBluescript transformants were grown in medium supplemented with 100 μg of ampicillin per ml.

Recombinant DNA manipulations.

Plasmids were isolated by the alkali lysis method on purification columns (Qiagen, Santa Clarita, Calif.). Recombinant DNA techniques (restriction endonuclease digestion, DNA fragment purification, alkaline phosphatase treatment, and ligation) were essentially as described previously (53) or as specified by the manufacturer. Restriction enzymes were obtained from Promega Corp. (Madison, Wis.) or New England Biolabs (Beverly, Mass.), calf alkaline phosphatase was obtained from Boehringer Mannheim (Indianapolis, Ind.), and T4 DNA ligase was obtained from Life Technologies. Oligonucleotide synthesis was performed by Genosys (The Woodlands, Tex.). DNA sequencing was performed at the University of Florida Interdisciplinary Center for Biotechnology Research Core laboratory using ABI 373 and 377 Perkin-Elmer/Applied Biosystems automated DNA sequencers. A Robotic Workstation (ABI Catalyst 800) and a Perkin-Elmer Cetus PEC 9600 thermocycler were used in fluorescent cycle sequencing reactions. After adapter was added, the ligation mix was heated at 45°C for 5 min before the ligase and the ligase buffer were added. The incubation proceeded overnight at 16°C. Chemically competent E. coli χ6097 was prepared by the method of Hanahan (26) in SOB medium supplemented with DAP. Electrocompetent S. enterica serovar Typhimurium χ4072 cells were prepared as previously described (5) in the presence of DAP.

PCR screening.

Transformed E. coli χ6097 colonies were screened by PCR (52) with four oligonucleotides to amplify an internal sequence of hagA and to confirm the presence of the hagA insert in pYA292, as follows: ST2/1, 5′-GCGGAATTCAGCTTCGATACGCAAACGCTTCCTAACG-3′ corresponding to nucleotides 1070 to 1097 of the hagA coding strand (25); ST2/2, 5′-CGATA ACTGCAGTATTACGCAGGCAAATCTACCGTACGCTCGATCC-3′ corresponding to nucleotides 4203 to 4231 of the hagA noncoding strand; PA2, 5′-GCGGATCCACCTTTTGAAAGTATTAAAGATTAATG-3′ complementary to bases 338 to 364; and TB4, 5′-GGCTCGTATAATGTGTGGA-3′ corresponding to nucleotides −57 to −39 upstream of the Met codon in pYA292 (21). In addition, to confirm the presence of the full-size hagA in the expression plasmid pNM1.1, PCR was performed with primers flanking the gene, as follows: upstream oligonucleotide 208, 5′-TTTCGCTCGCCGTCCTATTATC-3′ corresponding to nucleotides 387 to 408 of the coding strand, and downstream oligonucleotide 207-2, 5′-CGATCGGTTGGTAGAGCATAC-3′ complementary to nucleotides 8273 to 8293 of the noncoding strand (25).

Before the reactions were performed the suitability of the primers was verified with Oligo 4.0 software. The reactions were performed in a final volume of 50 μl containing 2.35 mM MgCl₂, 0.3 μM each primer, 0.4 mM each deoxynucleoside triphosphate (each), 1.25 U of Taq DNA polymerase (Promega), and 0.78 U of Pfu DNA polymerase (Stratagene, La Jolla, Calif.). Amplifications included an initial denaturation step of 1 min at 94°C followed by 30 cycles each of 94°C for 30 s, 55°C for 30 s, and 68°C for 6 min, with a final extension step of 10 min at 72°C, using a PTC-100 thermal cycler (MJ Research, Watertown, Mass.).

Immunological techniques.

Optimal dilutions of antibody, secondary antibody conjugate, and color substrate were selected by a series of dot blots and Western blots tested in multiwell incubation trays. Monoclonal antibody (MAb) 61BG1.3 (IgG1 isotype), (kindly provided by Rudolf Gmür, Institute of Oral Microbiology and General Immunology, Zürich, Switzerland [22]) was used to detect the expression of the target protein. Serum from subcutaneously challenged mice (see Fig. 3B) was obtained as previously described (36). Immunodot blots were used to detect HagA expression, as follows. Bacteria were collected, washed and resuspended in phosphate-buffered saline (PBS), and probe sonicated three times for 20 s on ice with a microsonicator (Kontes, Vineland, N.J.) in the presence of Complete proteinase inhibitor (Boehringer Mannheim). After centrifugation at 16,000 × g, the supernatant was collected for immunoanalysis. For selection of transformants expressing the target protein, colonies were lifted onto Nitro ME nitrocellulose filters (MSI, Westboro, Mass.) or Protran BA83 (Schleicher & Schuell, Keene, N.H.) and colony immunoscreening was performed (53).

FIG. 3.

Transmission electron micrographs of strains used in this study. (A) S. enterica serovar Typhimurium χ4072(pNM1.1) vaccine strain incubated with MAb 61BG1.3. (B) Same strain reacted with control unrelated antibody. (C) S. enterica serovar Typhimurium χ4072(pYA292) vector-only control reacted with MAb 61BG1.3. (D) P. gingivalis 381 reacted with MAb 61BG1.3 (positive control). (E) Same as panel D, with an unrelated antibody.

For Western immunoblots, 40-μl samples were mixed with 6× loading sample buffer (100mM Tris [pH 6.8], 5% [wt/vol] sodium dodecyl sulfate [SDS], 50% glycerol, 7.5% β-mercaptoethanol, 0.00125% bromphenol blue) and were loaded onto 10 to 20% gradient polyacrylamide gels after incubation in a boiling-water bath for 3 min. The gels were run in Tris-SDS buffer by the method of Laemmli (36a), and the proteins were reversibly visualized with a zinc staining kit (Bio-Rad, Hercules, Calif.). Broad-range molecular weight standards were used (Bio-Rad). After destaining, the proteins were transferred to a nitrocellulose membrane in a Trans Blot device (Bio-Rad) by standard procedures (3). The membranes were blocked (the blocking solution consisted of 5% Carnation dry nonfat milk and 0.02% sodium azide in Tris-buffered saline [TBS]) and reacted for 1.5 h with MAb 61BG1.3 diluted 1:20 or with serum from mice orally immunized by gastric intubation with recombinant Salmonella vaccine diluted 1:500 in the blocking solution (1% nonfat milk and 0.02% azide in TBS). The secondary antibody, alkaline phosphatase-conjugated goat anti-mouse IgG (Fisher), at a 1:500 dilution in the blocking solution, was applied for 1 h. Developing tablets (Sigma, St. Louis, Mo.) containing (after being dissolved) 0.01% nitroblue tetrazolium and 0.016% 5-bromo-4-chloro-3-indolyl phosphate (the color substrate) were used to develop the blots. Protein concentrations were determined with the bicinchoninic assay reagents (Sigma) as specified by the manufacturer. Proteins on blotted membranes were reversibly visualized with Ponceau S solution (Sigma).

Mouse immunization.

For oral immunization, a single colony of Salmonella vaccine strain was grown in Luria-Bertani broth at 37°C. Western analysis of the strain was done prior to immunization to confirm the presence of HagA in the cell lysate. After centrifugation at 5,000 × g, the bacterial pellet was resuspended in 0.1 M NaHCO₃ to yield ∼10¹⁰CFU/ml. Female BALB/c mice, 8 to 10 weeks old, were obtained from Charles River Laboratories, Inc. (Bar Harbor, Maine). The mice were intubated twice with 0.1 ml of Salmonella suspension at a 2-week interval. Serum was collected 9 weeks after the boost.

Human subject sera.

For immunoanalysis, serum from a clinical periodontal patient was obtained as described previously (39). The protocol for using adult human subjects was reviewed and approved by the University of Florida Institutional Review Committee. The patient was diagnosed with adult group 2 periodontitis (39). In this group, the alveolar bone loss is ≥8 mm at no more than one site and there are any number of sites with 4 to 5.9 mm of bone loss. The serum used contains anti-P. gingivalis antibodies at the following titers: anti-P. gingivalis 33277, 279 μg/ml; anti-P. gingivalis 381, 371 μg/ml; and anti-P. gingivalis W83, 103 μg/ml. For age-matched normal controls, the mean numbers are 20, 17, and 5 μg/ml, respectively.

IEM.

To detect a previously identified HagA-specific epitope (22) in the Salmonella vaccine strain, the vaccine cells were examined by immunoelectron microscopy (IEM) at the Electron Microscopy Core Laboratory of the Interdisciplinary Center for Biotechnology Research, University of Florida. Whole cells from the analyzed strains were collected from freshly grown liquid cultures. Minimum fixation was used to preserve the native conformation of the antigenic determinants. MAb 61BG1.3 was used as the primary antibody, and unrelated mouse MAb of matching isotype was used as a control. All samples were prelabeled and postlabeled after embedding and cutting of thin sections.

For prelabeling of bacteria with MAb 61BG1.3, the bacterial samples were pelleted by centrifugation, washed in PBS (pH 7.2), and fixed for 15 min in 4% paraformaldehyde. After fixation, the samples were washed twice in PBS. Each sample was split into two equal parts, and each part was incubated for 10 min with 1% bovine serum albumin. The samples were centrifuged, the supernatant was removed, and 0.25-ml volumes of hybridoma cultures containing either the mouse MAb 61BG1.3 or an unrelated IgG1 isotype-matched control mouse MAb were diluted 1:20 with PBS and added to the pellets. The pellets were resuspended and incubated overnight at 4°C.

For embedding of prelabelled bacteria, samples were pelleted and washed three times in PBS, dehydrated in an ethanol series to 100% ethanol, and then infiltrated and embedded in Unicryl.

For postlabeling of Unicryl sections of the bacteria samples with MAb 61BG1.3, thin sections of Unicryl-embedded samples on Formvar-coated nickel grids were incubated on 10-μl drops of 1% milk in high-salt Tween buffer (HST) (pH 7.2) for 10 min. The grids were blotted with filter paper and placed on either MAb 61BG1.3 or the control MAb which had been diluted 1:20 in HST. After the grids were incubated with the antibody overnight at 4°C in a moist environment, they were washed twice for 10 min each in HST buffer and once in PBS. They were then incubated for 1 h at room temperature on drops of anti-mouse IgG (Jackson ImmunoResearch Laboratories, Inc.) labeled with 18-nm-diameter gold particles, diluted 1:40 in PBS, and centrifuged before use. Finally the grids were washed three times for 10 min each in PBS, incubated on Trump fixative for 10 min, washed in distilled water, poststained with uranyl acetate and lead citrate, and examined under a Hitachi 7000 transmission electron microscope.

RESULTS

Construction of the HagA expression plasmid pNM1.1.

To construct a plasmid for heterologous expression of HagA, hagA (25), including its upstream and downstream regulatory regions, was obtained from pHagA2 by digestion with SacI and SalI and ligated into the SalI site of pYA292 by using a synthetic SalI-SacI adapter (Fig. 1). E. coli χ6097 cells transformed with this construct were grown on Luria-Bertani agar plates in the presence of 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) (40 μg/ml). White colonies were screened by PCR for the presence of a hagA insert as described in Materials and Methods. Plasmid pNM1, with the expected size, was isolated from a positive colony, and the presence of the hagA insert was confirmed by PCR with mixed pairs of primers (vector plus insert): ST2/1-ST2/2 and ST2/2-TB4 primer pairs were used for positive reactions, and the ST2/1-TB4 pair was used for the negative control (a graphic representation of the oligonucleotides is given in Fig. 1). The presence and orientation of the insert were also confirmed by sequencing reactions with TB4 and PA2 primers. Analysis of the expression plasmid pNM1 with primers flanking the full-size 8-kb insert demonstrated the presence of hagA (data not shown).

FIG. 1.

Construction of pNM1 (see Materials and Methods and Results). The hagA coding sequence including flanking regions was cloned in pYA292 by using a SalI-SacI adapter. Striped line, hagA open reading frame; dotted boxes, hagA flanking sequences; arrows, primers used for PCR analysis. Not to scale.

Construction of the recombinant S. enterica serovar Typhimurium χ4072 vaccine strain.

pNM1 was reisolated and transferred to S. enterica serovar Typhimurium χ4072 by electroporation. To screen S. enterica serovar Typhimurium χ4072 for transformants which expressed HagA, colony immunoscreening on 30 transformants was performed. Both E. coli χ6097 and S. enterica serovar Typhimurium χ4072 were transformed with the pNM1 expression construct. Screening of 30 S. enterica serovar Typhimurium colonies by enzyme-linked immunodetection of HagA with the 61BG1.3 monoclonal antibody resulted in the identification of six HagA-positive S. enterica serovar Typhimurium transformants. A plate with E. coli χ6097 transformants was used as a positive control. This immunodetection assay revealed that only 20% of Salmonella transformants expressed HagA compared to 100% of the control E. coli χ6097 transformants. Five of the S. enterica serovar Typhimurium HagA-positive transformants were grown, and plasmid DNA preparations (pNM1.1 to pNM1.5) were made and analyzed by PCR with internal and mixed pairs of primers, as described for the initial screening for pNM1. The size of each of these five plasmids was found to be equal to that of the E. coli-derived original. One of them (pNM1.1) was chosen for further study. Plasmid DNA was isolated from pNM1.1, and the insert was sequenced with TB4 and PA2 primers. The results confirmed the presence in S. enterica serovar Typhimurium of sequences identical to those of the plasmid (pNM1) isolated from E. coli.

Immunochemical analyses.

SDS-polyacrylamide gel electrophoresis and Western blot analyses were performed with MAb 61BG1.3 to detect the expression of HagA in the S. enterica serovar Typhimurium χ4072 vaccine strain (Fig. 2A, lane 5) compared to its expression in this strain containing vector only (lane 6) and in E. coli (lanes 1 and 3) compared to E. coli containing the vectors only (lanes 2 and 4). The HagA epitope, recognized by MAb 61BG1.3 (22), was detected in the Salmonella vaccine strain and in both recombinant E. coli strains used, JM109 (lane 1) and χ6097, the cloning host for Asd⁺ plasmid pYA292-based constructs (lane 3). The MAb was not reactive with the control Salmonella strain containing the vector only (lane 6). Interestingly, the level of expression from the same expression plasmid, pNM1.1, appeared to be lower in E. coli than in S. enterica serovar Typhimurium (lanes 3 and 5).

FIG. 2.

(A) Western immunoblot analysis of soluble cellular fractions of S. enterica serovar Typhimurium and E. coli with HagA-specific MAb 61BG1.3. Lanes: 1, E. coli(pHagA1); 2, E. coli(pUC18); 3, E. coli(pNM1); 4, E. coli(pYA292); 5, S. enterica serovar Typhimurium(pNM1.1); 6, S. enterica serovar Typhimurium(pYA292); 7, P. gingivalis 381. (B) Western immunoblot of recombinant HagA with HagA-specific MAb 61BG1.3 as the primary antibody (lane 1) and sera from Salmonella-immunized mice (lane 2). (C) Western blotting with serum from a periodontitis patient. Lanes 1: S. enterica serovar Typhimurium(pNM1.1); 2, P. gingivalis 381; 3, S. enterica serovar Typhimurium(pYA292). Molecular masses are given in kilodaltons.

To test the immunogenicity of HagA expressed by the vaccine strain, immune mouse serum was assayed for antibody to HagA by Western immunoblotting of a previously cloned 100-kDa recombinant HagA fragment (E. Kozarov and A. Progulske-Fox, unpublished data). The presence of specific anti-HagA antibodies in the serum was clearly demonstrated in Fig. 2B, lane 2. A 1-μg sample of protein was loaded in each of lanes 1 and 2. As a control, a similar blot was reacted with the specific anti-HagA MAb (lane 1).

To determine if the vaccine-derived HagA can be recognized by human antibodies, serum from a periodontitis patient was also analyzed by Western blotting. As shown in Fig. 2C, a high-molecular-weight band is visible in the vaccine strain preparation (lane 1) as compared to the Salmonella vector control (lane 3). Reactive antigen bands are also evident with the positive control, lysed P. gingivalis cells (lane 2). The presence of a unique protein species from the hemagglutinin-expressing vaccine strain that reacts with the serum from the periodontitis patient (lane 1) and is absent from the control (lane 3) is clearly evident.

Immunoelectron microscopy.

To determine the cellular location of the expressed HagA antigen, immunoelectron microscopy was performed on the S. enterica serovar Typhimurium χ4072(pNM1.1) vaccine strain (Fig. 3A), S. enterica serovar Typhimurium χ4072 (Fig. 3C, vector-only control), and P. gingivalis 381 (Fig. 3D, positive control). These results demonstrate the expression of the antigen in the vaccine strain. Figures 3D and E are positive and negative controls respectively.

DISCUSSION

Periodontitis in humans is thought to be caused by a group of predominantly gram-negative anaerobic bacteria, among which P. gingivalis is prominent. Considerable scrutiny is required to select useful immunogens that can elicit functional protection against periodontal tissue destruction induced by oral microorganisms that already colonize or infect the host (31). Immunization with a vaccine containing killed whole cells of P. gingivalis suppresses the progress of experimental periodontitis in M. fascicularis (57). However, a vaccine composed only of specific protective antigens is most desirable.

Hemagglutinin A is the largest member of a family of P. gingivalis proteins, including hemagglutinins A and D, whose genes were isolated via functional screening for hemagglutinating activity (25, 50). They have extensive homology to each other and to other abundant P. gingivalis proteins including protease PrtP (4), PrtH (20), protease RGP-1 (47), protease PrtR (35), argingipain (46), and Arg1 (11). One of the four ∼450-amino-acid (aa) repeats making up more than half of the HagA polypeptide is the common shared motif. The MAb used here for detection of expression of HagA, 61BG1.3 (22), provides passive protection against recolonization of P. gingivalis in humans (6) and recognizes an epitope present in HagA and in the proteins to which HagA has homology. With five copies of the epitope, HagA itself is a multivalent carrier. The 61BG1.3 epitope may be a component of a binding domain common to multiple gene products of this organism and may thus represent a functionally important target of the specific immune response of the host to P. gingivalis (11). The existence of multiple gene products containing a common epitope has previously been reported for Moraxella catarrhalis. The high-molecular-weight UspA protein of M. catarrhalis is present on the surface of all M. catarrhalis disease isolates examined to date and contains the epitope for a MAb (MAb 17C7) which enhances the pulmonary clearance of this organism in a mouse model system (27). Recently, the presence of a second M. catarrhalis gene, uspA2, which also encodes the MAb 17C7-reactive epitope, has also been reported (1). Interestingly, both UspA1 and UspA2 proteins closely resemble adhesins produced by other bacterial pathogens. With P. gingivalis, the hemagglutinating adhesin HagA (25) shares a 25-aa residue protective epitope found in the arginine-specific protease (11) and the lysine-specific protease (51) of this organism. Thus, construction of a vehicle for delivering the HagA-encoded antigens may be an efficient way of eliciting an immune response capable of preventing colonization of P. gingivalis in humans.

The apparent processing of the HagA polypeptide (Fig. 2A, lane 5) may be because many P. gingivalis gene products are posttranslationally processed to contribute to the formation of multimeric surface protein-adhesin complexes (37). It is established that various cell surface and secretory proteins are processed in P. gingivalis (34).

The epitopes recognized by sera from periodontitis patients have been previously reported to fall within the beta subunit, a hemagglutinin and/or adhesin component of the arginine-specific proteases of P. gingivalis (ArgI, ArgIA, and ArgIB) (11). The antibody response in animals to a protease carrying both catalytic and hemagglutinating domains is confined only to the adhesive part of the protein, suggesting that the catalytic portion is not exposed (J. Travis, personal communication). Thus, a HagA vaccine which includes an epitope common to a family of hemagglutinins in addition to proteases may be an effective immunogen against a variety of virulence factors.

The goal of the present effort was to determine if HagA can be expressed in an immunogenic form in a Salmonella-based live vaccine strain. In this study we show that hemagglutinin A, a 2,628-aa P. gingivalis protein which agglutinates erythrocytes and is implicated in the virulence of the bacterium, can be expressed in an attenuated Salmonella vaccine strain. Serum from mice immunized with the vaccine strain react with purified HagA. In addition, we have demonstrated that the HagA antigen of the vaccine strain is recognized by antibodies in the serum of a periodontal patient.

The live vaccine strain, S. enterica serovar Typhimurium χ4072, is both avirulent and immunogenic but retains its ability to colonize the GALT (12). This vaccine strain has been used previously to express another hemagglutinin from P. gingivalis (18). The nonfused filamentous hemagglutinin of Bordetella pertussis, an important adhesin in the early interactions between the bacterium and host cells, has also been efficiently expressed in S. enterica serovar Typhimurium (23). Plasmid pYA292 has also been used to express streptococcal surface antigens (15), Entamoeba hystolytica antigens (10), and hepatitis B virus antigens (54).

The presence of the hagA gene-associated protein in Salmonella is demonstrated by immunoanalysis. This suggests that although no E. coli-like ribosome binding sequence is present in the 5′-untranslated region of hagA, the E. coli and Salmonella transcription and translation machinery still functions to express this gene. Multiple protein bands are recognized in both E. coli and P. gingivalis in these blots, which suggests that the 2,628-aa target protein is being processed by proteases.

In gram-negative bacteria, many periplasmic and outer membrane preproteins have a signal sequence at the N terminus which is cleaved during translocation of the protein through the cytoplasmic membrane. By using the PSORT algorithm (58), the structure of the N-terminal region of HagA is predicted to be typical of a prokaryotic signal peptide, initiating inner membrane transfer of the precursor. It consists of positive N-terminal charges, a central hydrophobic region, and an Ala signal peptidase cleavage site (Fig. 4). Cleavage at the predicted cleavage sites would result in an outer membrane-embedded or secreted protein.

FIG. 4.

N-terminal amino acids of HagA. +, basic-type residue; H, hydrophobic-type residue; P, polar uncharged-type residue.

In addition to being surface exposed, HagA is probably released from the cells since hemagglutination activity in the culture medium has been reported for P. gingivalis 381 (32). In our study, Western immunoblotting of spent culture medium from different P. gingivalis strains demonstrated an abundance of protein species recognized by the protective antibody (data not shown). These proteins may either be independently released or constitute a component of blebs, i.e., 100-nm membrane vesicles released by P. gingivalis. Proteins secreted by Mycobacterium, another mucosal pathogen, have also been suggested to be major immune targets (2). Secreted proteins are preferentially recognized by T cells before somatic proteins (14), and it has been shown that secreted or surface-localized antigens in Salmonella display superior efficacy over that of somatic display (29, 55). Accordingly, by using IEM we demonstrated surface expression of the target protein in Salmonella. An immune response to HagA may be efficient against different strains of the pathogen. In addition, MAb 1A1, which recognizes the same epitope as MAb 61BG1.3, strongly inhibits the agglutination of human erythrocytes by P. gingivalis culture supernatant (11). These findings suggests that Salmonella expressing HagA would be a good choice as a live-vaccine candidate.

In conclusion, heterologous expression of hemagglutinin A, a major virulence-associated surface protein of P. gingivalis, was demonstrated in an avirulent vaccine strain of S. enterica serovar Typhimurium. A balanced-lethal non-antibiotic-resistant expression vector for the Salmonella host system was used for the expression. Successful delivery of the target protein via the mucosal immune system was demonstrated by the presence of antibodies in the sera of mice which had received the vaccine strain by gastric intubation. A well-characterized epitope of the HagA protein (22) was shown to be present in its native configuration by immunochemistry and IEM analysis of the vaccine strain. The presence of hemagglutinin A on the Salmonella surface was demonstrated by IEM. In addition, the HagA antigen of the vaccine strain was recognized by antibodies present in the serum of a human periodontitis patient. Finally, by using purified recombinant HagA, the presence of specific anti-hemagglutinin A antibodies in the serum of orally immunized mice was established. Therefore, testing of this vaccine construct for the elicitation of a protective immune response is continuing.