Vol. 58, No. 10
Protective Immunization against Experimental Bacteroides
(Porphyromonas) gingivalis Infection
PRISCILLA B. CHEN,'* LYNDA BETH DAVERN,1 ROBERT SCHIFFERLE,"2 AND JOSEPH J. ZAMBON" 2
Departments of Oral Biology' and Periodontology,2 School ofDental Medicine, State University ofNew York at Buffalo,
Buffalo, New York 14214
Received 27 February 1990/Accepted 13 July 1990
The effects of immunization in modulating the pathogenesis of Bacteroides (Porphyromonas) gingivalis
infection in a murine model system were examined. BALB/c mice were immunized by intraperitoneal injection
with B. gingivalis ATCC 53977 (one injection per week for 3 weeks), or with a lithium diiodosalicylate (LIS)
extract (one injection per week for 3 weeks), or with lipopolysaccharide (LPS; one intravenous or intraperi-
toneal injection) from this same strain. Two weeks after the final immunization, the mice were challenged by
subcutaneous injection of B. gingivalis ATCC 53977. Mice immunized with bacteria had no secondary lesions
and no septicemia, whereas mice immunized with LIS extract had few secondary lesions and no septicemia.
Mice immunized with LPS and nonimmunized mice demonstrated secondary abdominal lesions and septicemia
after challenge. Bacterial cells and LIS extract, but not LPS, induced serum antibody and antigen reactive
transfer, and in vitro lymphoproliferative responses. The present study suggests that immunization with a LIS
extract or whole cells may induce a protective response against experimental B. gingivalis infection.
as measured by enzyme-linked immunosorbent assay, immunoblot, Western immunoblot
Bacteroides (Porphyromonas) gingivalis is a gram-nega-
tive, anaerobic rod that has been associated with adult
periodontitis and abscesses of oral origin. This species
demonstrates significant heterogeneity as regards virulence
in animal models. Studies in our laboratory (25) and other
laboratories (8, 11, 30, 32) have demonstrated that B. gingi-
valis strains can be classified as either invasive or noninva-
sive depending on whether the strain migrates to a site
distant from the injection site (invasive) or remains localized
to the injection site (noninvasive). Immunization can alter
the course of experimental B. gingivalis infections in a
mouse model. Previously, we have shown that immunization
with invasive B. gingivalis ATCC 53977 (A7A1-28) resulted
in localization of a subsequent challenge infection with the
homologous strain (4). However, mice immunized with a
noninvasive strain of B. gingivalis or with Bacteroides
intermedius or with Ringer solution (4) developed spreading
infections after challenge with strain ATCC 53977. Similarly,
Okuda and colleagues (26) found that there were slight
reductions in the number ofB. gingivalis 381 recovered from
hamsters immunized with B. gingivalis 381 whole cells (a
noninvasive strain) or its hemagglutinin when compared with
sham-immunized controls. The present study sought to
determine ifbacterial extracts such as a membrane extract or
lipopolysaccharide (LPS) could confer protective immunity
against an experimental invasive B. gingivalis infection.
MATERIALS AND METHODS
Bacteria. B. gingivalis ATCC 53977 (A7A1-28) was iso-
lated in 1985 at the State University of New York at Buffalo
from a deep periodontal lesion in a 37-year-old male patient
with non-insulin-dependent diabetes mellitus. In mice, this
B. gingivalis isolate is invasive and produces ulcerated
lesions distant from the injection site and septicemia and,
often, death (25). The pathogenic potential of this strain is
similar to that reported for B. gingivalis W50 and W83 (30,
32). B. gingivalis ATCC 53977 was transferred from tryptic
soy agar (Difco Laboratories, Detroit, Mich.) supplemented
with 5% sheep blood (Crane Laboratories, Inc., Syracuse,
N.Y.), 5jigof hemin (Sigma Chemical Co., St. Louis, Mo.)
per ml, 1.0jigof menadione (Sigma) per ml, and 1% yeast
extract (BBL Microbiology Systems, Cockeysville, Md.),
henceforth called enriched tryptic soy agar (ETSA), to
trypticase soy broth containing hemin (5 ,g/ml) and mena-
dione (0.5 ,ug/ml). The microorganism was cultured in an
anaerobic chamber containing an atmosphere of 85% N2-
10% H2-5% CO2 (Forma Scientific, Marietta, Ohio) at 37°C
to late logarithmic-early stationary phase. Bacterial purity
was confirmed by phase-contrast microscopy, light micro-
scopic examination ofGram stains, and monitoring ofcolony
morphology after anaerobic culture. Cells were harvested by
centrifugation, washed three times in phosphate-buffered
saline (pH 7.2) and stored at -70°C until required.
Antigen preparation and characterization. Two different
preparations were prepared from whole cells ofB. gingivalis
ATCC 53977 for use in this study.
(i) LIS extract. Bacterial cells were extracted with lithium
diiodosalicylate (LIS; Eastman Kodak Co., Rochester,
N.Y.) by stirring 2 x 1010 bacteria per ml in 0.3 M LIS (16)
dissolved in Tris buffer (0.05 M, pH 7.2) for 1 h at room
temperature and then overnight at 4°C. The extracted cells
were removed by centrifugation at 12,000 x g. The super-
natant was dialyzed against distilled water until the optical
density at 323 nm (OD323) was <0.05 and was then concen-
trated by ultrafiltration and stored in aliquots at -70°C until
required. Protein content was determined by the method of
Lowry et al. (14) by using bovine serum albumin as a
standard. The LIS extract (1 mg of protein per ml) was
hydrolyzed at 100°C for 6 h with 2 N HCl for hexosamine
analysis and with 6 N HCl for 24 h for amino acid analysis
with a Beckman 6300 amino acid analyzer. Neutral sugar
content of the LIS extract was determined by the method of
Dubois et al. (5) by using glucose as a standard.
(ii) LPS. LPS was prepared by extracting bacterial cells
with hot phenol by using a modification of the method of
INFECTION AND IMMUNITY, OCt. 1990, p. 3394-3400
Copyright X 1990, American Society for Microbiology
EXPERIMENTAL B. GINGIVALIS INFECTION 3395
Millar et al. (19). Briefly, bacterial cells (10 to 15%, wt/vol)
were suspended in sodium phosphate buffer (0.05 M, pH 7.4)
containing NaCl (0.15 M) and disodium EDTA (1 mM) and
broken by means of glass beads with a Braun homogenizer
(five 1-min bursts) at 4°C. The glass beads were removed by
centrifugation at 200 x g. The bacterial cell suspension was
centrifuged at 12,000 x g, and the resulting supernatant was
ultracentrifuged at 80,000 x g. The pellet formed by ultra-
centrifugation was lyophilized. The pellet was then sus-
pended in distilled water (10 mg/ml) and mixed with phenol
(1:1) preheated to 68°C. The phenol-water mixture was
reheated to 68°C with stirring. The aqueous phase was
removed and the phenol phase was reextracted at 68°C with
an equal volume of water. The two aqueous phases were
combined, dialyzed against distilled water at 4°C, and then
lyophilized. Protein content was determined by the method
of Lowry et al. (14) by using bovine serum albumin as a
The two B. gingivalis antigens were compared by sodium
dodecyl sulfate-polyacrylamide gel electrophoresis by the
method of Laemmli (13) with a Hoeffer SE600 Mighty Tall
slab gel unit. The LIS extract and LPS were electrophoresed
on a sodium dodecyl sulfate-12% polyacrylamide resolving
gel and a 3% polyacrylamide stacking gel and stained with
silver (9) or with Coomassie brilliant blue R (Sigma). After
sodium dodecyl sulfate-polyacrylamide gel electrophoresis,
the antigens were also compared by Western immunoblot
analysis with a dry transfer system (Polyblot transfer appa-
ratus, Hayward, Calif.) for transfer to nitrocellulose. The
nitrocellulose was blocked with bovine serum albumin over-
night and then incubated with normal mouse serum (1/500) or
immunized mouse serum (1/500) for 3 h at room temperature.
The second antibody was goat anti-mouse immunoglobulin
G (IgG) or IgM conjugated to horseradish peroxidase (1/
3000; Bio-Rad Laboratories, Richmond, Calif.), and the
substrate was 4-chloro-1-naphthol (Bio-Rad).
Infection model. Female BALB/c mice, 6 to 8 weeks of age
(West Seneca Labs, West Seneca, N.Y.), were immunized
by intraperitoneal injection of 0.1 ml of B. gingivalis cells
(109 per mouse), LIS extract or LPS (100 ,ug/ml of pyrogen-
free saline). The LIS extract and B. gingivalis cells were
administered once a week for 3 weeks, whereas LPS was
administered once. Serum samples from each mouse were
obtained before and after immunization.
All mice were challenged 2 weeks after the final immuni-
zation by subcutaneous injection of 0.1 ml of B. gingivalis
ATCC 53977 bacteria (5 x 1010 per ml) into each oftwo sites
about 1 cm lateral from the midline on the dorsal surface.
The mice were then examined daily for the following: (i) the
size and presence of lesions at the injection site; (ii) pres-
ence, location, and appearance ofsecondary lesions; and (iii)
health status. Five mice from each group were sacrificed on
days 1, 2, 5, and 15 after challenge, and samples, including
blood from the subclavian artery, abscess fluid, and serous
exudate from any secondary lesion, were collected for
microbiological examination. Portions of each sample were
streaked onto ETSA and anaerobically cultured at 37°C for 7
days. The ETSA plates were then examined for growth ofB.
gingivalis, and the number of colonies was recorded, as was
the colony morphology and pigmentation. Bacterial cells
from the colonies were gram stained and examined by light
Immediately prior to challenge, spleens were obtained
from five normal and five immunized mice from each group
and tested as previously described (3). Briefly, single-cell
suspensions were prepared from each spleen and adjusted to
TABLE 1. Amino acid composition of LIS extract
No. of residues/
Aspartic acid .................
aResults are from an amino acid analyzer after hydrolysis in 6 N HCI at
105°C for 28 h.
2.5 x 106 cells per ml ofRPMI 1640 (GIBCO, Grand Island,
N.Y.) supplemented with antibiotics and 10% heat-inacti-
vated fetal calf serum (GIBCO). Cells were dispensed into
individual wells (200
plate, and the appropriate antigen or mitogen was added to
triplicate wells in 20-,ul volumes. The cell cultures were
incubated at 37°C in 5% C02-95% air for 72 h. Six hours
prior to harvest, [3H]thymidine (specific activity, 2.0 Ci/mM,
1 jxCi per culture) was added to each well. Cells were
collected by using a cell harvester, and the degree of
radiolabel incorporation was determined by liquid scintilla-
tion counting. Results were analyzed statistically by analysis
of variance and the Fisher least significant difference proce-
dure (20). Results from a representative experiment for five
individual spleens are presented as mean stimulation index
(E/C signifies mean counts per minute of stimulated cultures/
mean counts per minute of unstimulated cultures) ± stan-
Serum samples were tested for antibody to B. gingivalis
by immunoblotting on nitrocellulose with a Bio-Dot micro-
filtration apparatus (Bio-Rad) as previously described (18) by
using 4-chloro-1-naphthol (Bio-Rad) as the substrate or by
enzyme-linked immunosorbent assay (ELISA) as previously
described (2) with p-nitrophenylphosphate (Sigma) as the
substrate. Mouse sera were diluted 1/500. Goat anti-mouse
IgG conjugated to horseradish peroxidase (1/3000) for immu-
nodot or alkaline phosphatase (1/1000) for ELISA was the
second antibody. Color intensity was read at OD405 with an
ELISA reader (Dynatech).
.lI per well) of a 96-well microtiter
Analysis of LPS and LIS for protein by the method of
Lowry et al. (14) revealed less than 1% protein in LPS (2.5
p.g/ml) and 16 to 20 mg of protein per ml in the LIS extract.
The LIS extract was diluted to 1 mg of protein per ml for
neutral sugar, amino sugar, and amino acid analysis. The
results obtained for the LIS extract revealed neutral sugars
(8.73 ,ug/mg of protein), amino sugars (glucosamine, 11.4
,ug/mg of protein; and galactosamine, 11.0jig/mgofprotein),
and amino acids (Table 1). Staining with Coomassie brilliant
blue R revealed a banding pattern for LIS but not LPS (Fig.
VOL. 58, 1990
3396 CHEN ET AL.
X 3:~~4 3w
FIG. 1. Sodium dodecyl sulfate-polyacrylamide (12.5%) gel elec-
trophoresis of LIS extract and LPS from B. gingivalis ATCC 53977.
The LIS extract and LPS were stained with Coomassie brilliant blue
R250 (A) and silver nitrate (B). Lanes: 1 and 5, prestained molecular
mass standards in kilodaltons; 2 and 6, Salmonella typhimurium
LPS (5 jig, dry weight); 3 and 7, B. gingivalis ATCC 53977 LPS (5
jig, dry weight); 4, B. gingivalis ATCC 53977 LIS extract (20 jig of
protein); 8, B. gingivalis ATCC 53977 LIS extract (10 jig of protein).
1A), with major protein bands at 40 and 57 kilodaltons (kDa).
By contrast, staining with silver nitrate (9) revealed a more
intense ladderlike pattern associated with the B. gingivalis
LPS than with the LIS extract (Fig. 1B). Less than 1%
nucleic acids were present in the LPS, as determined by
absorption at 260 nm.
By 18 to 24 h after challenge with B. gingivalis ATCC
53977, all control mice and mice immunized with LPS had
on the abdomen (Table 2). These
spreading infections were hemorrhagic, ulcerative lesions
that covered the lower abdomen and often extended poste-
riorly to the base of the tail. Ulcerated lesions that rapidly
dried were observed at the injection site. These mice exhib-
ited moderate to severe cachexia, with ruffled hair, hunched
bodies, weight loss (Fig. 2), and an odor associated with B.
gingivalis infections. Microbiological sampling 24 h after
challenge demonstrated B. gingivalis in the blood, at the
injection site, and in secondary lesions on the abdomen
Mice immunized with B. gingivalis ATCC 53977 whole
bacterial cells usually demonstrated abscesses at the injec-
tion site after challenge but did not show secondary lesions
on the abdomen. Similarly, microbiological samples were
positive for B. gingivalis ATCC 53977 at the injection site
but were negative in samples from the abdomen or blood
(Table 2). These mice exhibited normal behavior, without
ruffled hair or hunched bodies, and lost less weight than the
other groups of mice (Fig. 2). Thus, immunization with B.
gingivalis ATCC 53977 prevented septicemia and generally
localized the challenge infection to the injection site.
Following challenge with B. gingivalis ATCC 53977, all
mice immunized with the LIS extract developed abscesses at
the injection site and 20% developed secondary infections.
B. gingivalis could be cultured from both the injection site
and the abdomen 2 days postchallenge but could not be
isolated from blood (Table 2). These mice did not have
ruffled hair and did not lose as much weight as the control or
LPS-immunized mice (Fig. 2), and their secondary lesions
were small and nonhemorrhagic and were often found
around the tail rather than on the abdomen.
Spleens were obtained prior to challenge from immunized
and control mice to evaluate in vitro lymphoproliferative
responses to bacterial antigens, phytohemagglutinin (Bur-
roughs Wellcome) and Escherichia coli 0111:B4 (Difco)
LPS. Spleen cells from all groups responded in vitro to
phytohemagglutinin and E. coli LPS (Fig. 3). Enhanced in
vitro lymphoproliferative responses to B. gingivalis ATCC
53977 bacteria and LIS extract were observed in spleen cells
from mice immunized with B. gingivalis ATCC 53977 and
contrast, spleen cells from mice immunized with LPS exhib-
ited lower in vitro lymphoproliferative responses to all B.
gingivalis antigens tested when compared with spleen cells
from mice immunized with B. gingivalis ATCC 53977 or
from LIS-immunized mice. A representative experiment is
presented in Fig. 3.
Before challenge, mice immunized with bacteria or LIS
extract demonstrated antibodies in serum to B. gingivalis
and LIS, but normal mice or mice immunized with LPS did
not (Table 3). LIS-immunized mice had low levels of anti-
body in serum to LPS (OD405, 0.12 to 0.2), whereas serum
samples from normal mice or mice immunized with B.
gingivalis ATCC 53977 cells or LPS did not contain antibod-
ies to LPS (Table 3). Different immunization protocols were
employed for immunization with B. gingivalis LPS (from one
to multiple immunizations, intravenously or intraperitoneal-
ly), but no IgM or IgG serum antibodies to LPS could be
as compared with normal (P < 0.05). In
TABLE 2. Clinical courses of B. gingivalis infection in mice after challengea
Site 2 (%)
B. gingivalis ATCC 53977, live (109/mouse, 3
B. gingivalis ATCC 53977-LIS (10 ,ug/mouse,
B. gingivalis ATCC 53977-LPS (10 ,ug/mouse)
Ringer solutiond (3 injections)
Abscess ± ulcerated
Abscess t ulcerated 20 (by day 4)
aChallenged with5 x 109 B. gingivalis ATCC 53977 administered by subcutaneous injection. For each immunizing agent, 25 mice were challenged. Site 1,
Injection site; Site 2, usually the abdomen.
bLesions occurring at site 2 were hemorrhagic and ulcerative.
cResults of microbiological sampling 2 days postchallenge. -, Negative; +1, <10 colonies; +2, 20 to 100 colonies; +3, >100 colonies, still countable; +4,
confluent growth, colonies were too numerous to count.
dPrereduced, anaerobically sterilized.
EXPERIMENTAL B. GINGIVALIS INFECTION 3397
FIG. 2. Weight loss (% of initial weight) following challenge with B. gingivalis ATCC 53977 (5 x 1010 cells per ml, 0.2 ml per mouse). Mice
were immunized with B. gingivalis ATCC 53977 bacteria (O----O), LIS extract (OL----OI), or LPS (A----A) or were not immunized (@-).
detected in this group of mice by either ELISA or immuno-
blot prior to challenge, whether test sera were diluted at
1/100 or 1/500. By 12 to 15 days postchallenge, however,
serum samples from all remaining mice contained antibodies
to B. gingivalis cells and LIS extract (Table 3). Serum
samples from mice immunized with B. gingivalis cells or LIS
extract reacted with bacterial cells, LIS extract, and LPS by
day 15 after challenge. Serum samples from mice immunized
with LPS reacted with B. gingivalis cells and LIS extract by
day 15 after challenge but did not react to LPS. Serum
samples from nonimmunized mice reacted with B. gingivalis
cells and LIS extract by day 15 after challenge. Analysis of
antibody reactivity by Western transfer demonstrated anti-
gen reactivity of serum samples from mice immunized with
B. gingivalis ATCC 53977 to the LIS extract but not LPS
(Fig. 4). Serum samples from nonimmunized mice did not
react to LIS extract or LPS preparation prior to challenge
infection (Fig. 4).
Studies of B. gingivalis infections in a murine model from
this laboratory (25) and other laboratories (8, 11, 30, 32) have
demonstrated that B. gingivalis strains demonstrate differing
abilities to invade tissue. Thus, B. gingivalis isolates can be
classified as invasive ifthey migrate to a site distant from the
injection site or noninvasive if they remain localized to the
injection site with subsequent abscess formation. We have
previously shown that immunization with an invasive B.
gingivalis isolate results in localization of a challenge infec-
tion with the homologous strain while immunization with a
noninvasive strain does not (4). In the study reported here,
active immunization with B. gingivalis ATCC 53977 also
resulted in localization of a challenge infection with the
homologous invasive strain. Immunization with B. gingivalis
AJW4, an invasive strain, also resulted in localization of a
challenge infection with B. gingivalis ATCC 53977 (data not
shown). Immunization with the LIS extract from B. gingi-
valis ATCC 53977 or AJW4 resulted in reduced severity of a
B. gingivalis ATCC 53977 challenge infection in BALB/c
mice and no septicemia. The B. gingivalis ATCC 53977 LIS
extract induced both humoral immune responses, as as-
sessed by Western transfer and ELISA, and in vitro lym-
TABLE 3. Reactivity of sera from immunized mice to B.
gingivalis ATCC 53977
Immunization agent and time of challenge
ELISA reactivity' (OD405)
B. gingivalis ATCC 53977 cells
B. gingivalis ATCC 53977-LIS extract
B. gingivalis ATCC 53977-LPS
aResults are presented as meanOD405 of duplicate wells of immunized or
control mouse serum samples (standard deviations are within 10%o of the
mean) minus mean OD405 of normal mouse serum plus antigen. The mean
OD4o5 + standard deviations were as follows: normal mouse serum without
antigen, 0.026 + 0.001; normal mouse serum plus B. gingivalis ATCC 53977
cells, 0.080 + 0.012; normal mouse serum plus LIS, 0.064 ± 0.003; normal
mouse serum plus LPS, 0.055 + 0.008. B. gingivalis ATCC 53977 cells, OD580,
0.3; LIS extract, 100 ,ug/ml; LPS, 5 ,ug/ml.
VOL. 58, 1990
CHEN ET AL.
FIG. 3. Lymphoproliferative response of mouse spleen cells to B. gingivalis antigens. Spleen cells (5 x 105 per well) from normal mice
(_)or mice immunized with B. gingivalis ATCC 53977 cells ( M), LIS extract ( M
for 72 h at 370C in 5% C02-95% air with phytohemagglutinin (0.2 ,ug per well), E. coli 0111:B4 LPS (2 ,ug per well), B. gingivalis ATCC 53977
bacteria (1/100 dilution of packed cells), B. gingivalis LIS extract (LIS, 2 ,ug per well), or B. gingivalis LPS (2 ,ug per well). Six hours prior
to harvest, [3H]thymidine (1 ,Ci per well) was added. Degree of [3H]thymidine incorporated was determined by liquid scintillation counting.
Results are expressed as the mean ratio + the standard deviation of mean counts per minute of cells cultured with stimulant to cells cultured
with media (E/C) for five mice. *, Significantly different from normal at P < 0.05.
), or B. gingivalis LPS (
) were cultured intriplicate
phoproliferative responses as demonstrated by tritiated thy-
midine incorporation. The LIS extract also induced
nonspecific lymphoproliferative response in spleen cells
from nonimmunized mice, suggesting a mitogenic compo-
nent associated with this extract. The lymphoproliferative
response of normal spleen cells to the LIS extract was less
than half that of mice immunized with B. gingivalis ATCC
53977 cells, suggesting that a specific lymphocyte response
was also involved.
LIS has been used to extract glycoproteins from erythro-
cytes (16) and a sialic acid-binding protein from Streptococ-
cus mitis (24). The results obtained in this study suggest that
LIS-extractable membrane proteins can induce a partially
protective immune response to an invasive B. gingivalis
infection. It remains to be determined if the membrane-
extractable proteins in the LIS extract are outer membrane
proteins. The results of this study, therefore, are consistent
with a report by Okuda et al. (26), who showed a slight
decrease in the number of B. gingivalis 381 on ligated
hamster first molars after immunization with whole cells or
extracted hemagglutinin. This LIS extract used in this study
contained both protein and LPS; however, the LPS content
was less than that in the LPS preparation. The abilities of
LIS-immunized mice to produce antibody to LPS suggest
that the B. gingivalis LPS used in this study needs to be
associated with a protein in order to be immunogenic. The
inability of serum samples from mice immunized with B.
gingivalis ATCC 53977 bacteria to respond to LPS until 12 to
15 days after challenge suggests that the LPS was not
exposed sufficiently on the surface to induce an immune
response, possibly because of encapsulation.
Active immunization of mice with B. gingivalis LPS did
not result in readily detectable IgM or IgG antibody levels to
LPS, nor did it reduce the severity ofB. gingivalis infection.
However, serum samples from LPS-immunized mice re-
acted with the LIS extract by day 15 postchallenge. These
results suggest that structural differences of LPS presented
as a component of a LIS-extractable membrane protein
versus phenol-extracted LPS are important in its role as an
antigen or that an antigen-specific unresponsive state oc-
curred. Spleen cell lymphoproliferative responses to B.
gingivalis antigens from LPS-immunized mice were similar
to or lower than those of nonimmunized mice. After chal-
lenge, these mice lost as much or more weight than controls
and had ruffled hair and hunched bodies. Extensive hemor-
rhagic lesions were observed on the abdomen, around the
tail, and on the hind legs ofthese mice. These results suggest
that LPS did not induce a protective immune response but
may have induced an antigen-specific unresponsive state. A
possible explanation for these results may be found in the
work of Elkins et al. (7) who showed that a single, subim-
munogenic injection of LPS induced the development of
antigen-specific immunological unresponsiveness. However,
the differences could also be due to structural differences in
LPS molecules. It remains to be determined if the immuni-
zation regimen utilized in this study results in a similar
tolerance or if the structure of the B. gingivalis LPS is
responsible. Alternatively, the extensive hemorrhagic le-
sions observed in LPS-immunized mice may be due to
production oftumor necrosis factor, since Bramanti et al. (1)
have reported on the induction of tumor necrosis factor by
B. gingivalis LPS.
Murine models of bacterial infection have shown that
active immunization with bacterial cells can prevent or
EXPERIMENTAL B. GINGIVALIS INFECTION
FIG. 4. Western blot analysis of LIS extract and LPS from B.
gingivalis ATCC 53977. Lanes: 1 and 4: molecular weight standard
in kilodaltons; 2 and 5, LPS (5 ,ug, dry weight); 3 and 6, LIS (20 Fg
ofprotein). Lanes 4, 5, and 6 were probed with normal mouse serum
and lanes 1, 2, and 3 were probed with sera from mice immunized
with B. gingivalis ATCC 53977.
decrease the incidence of a subsequent challenge infection
(6, 22). Efforts have focused on the isolation, purification,
and characterization of protective bacterial antigens. One
group, the outer membrane antigens, including porins, LPS,
and lipoproteins have been studied for their roles in induc-
tion of a protective immune response. The role of outer
membrane proteins in induction of a protective immune
response has been extensively studied in Salmonella typhi-
murium (29), Salmonella typhi (10), Brucella abortus (21),
Pseudomonas aeruginosa (17), Pasteurella multocida 3:A
(15), and Haemophilus influenzae type b (12) infections in
various animal models, in which immunization with the
outer membrane proteins is reported to confer protection
against infection. Antibodies to certain of these extractable
membrane proteins, including those from Neisseria menin-
gitidis (27), H. influenzae type b (23), and E. coli (31) are also
protective against lethal infection. Previous studies of the
outer membrane proteins ofB. gingivalis by sodium dodecyl
sulfate-polyacrylamide gel electrophoresis (33) revealed a
complex pattern with major bands at 40, 57, and 66 kDa and
a number of minor bands smaller than 40 kDa. A 74-kDa
band was present at 37°C but was absent at 100°C. Yoshi-
mura et al. (34) have reported that a 75-kDa protein from B.
gingivalis 381 was recognized by serum samples from peri-
odontitis patients and immunized rabbits. In this study, a 57-
and a 40-kDa band appeared to be the major protein bands,
as identified by staining a LIS-extractable membrane prep-
aration with Coomassie brilliant blue. Serum samples from
immunized mice recognized bands at approximately 57, 43,
40, 37, 30, 28, and 15 kDa. The protective potential of each
of these antigenic components remains unclear.
Other bacterial cell surface components can also be pro-
tective. For example, immunization with the capsular
polysaccharide from Bacteroides fragilis can prevent intra-
abdominal abscess formation in mice (28). It remains to be
determined if B. gingivalis capsular polysaccharide can be
used to immunize and prevent or localize an invasive B.
gingivalis infection. This study has shown that active immu-
nization with a LIS extract results in decreased severity of a
B. gingivalis ATCC 53977 challenge infection. By compari-
son, immunization with LPS does not afford protection from
a challenge infection with B. gingivalis ATCC 53977. The
specific role of both cell-mediated and humoral immunity in
modulating the pathogenesis of this infection is currently
The study was supported by Public Health Service grants
DE06880 and DE08240 from the National Institute of Dental Re-
We are grateful to Suzanne Michalek and Russell Nisengard for
their discussions about the manuscript, Robert Dunford for assis-
tance with statistical analyses, and Paul Dressel for photography.
ADDENDUM IN PROOF
Bacteroides gingivalis is now officially Porphyromonas
1. Bramanti, T. E., G. G. Wong, S. T. Weintraub, and S. C. Holt.
1989. Chemical characterization of biologic properties of lipo-
polysaccharide from Bacteriodes gingivalis strains W50, W83
and ATCC 33277. Oral Microbiol. Immunol. 4:183-192.
2. Chen, P., V. Bochaki, H. S. Reynolds, J. Beanan, D. N. Tatakis,
J. J. Zambon, and R. J. Genco. 1986. The use of monoclonal
antibodies to detect Bacteroides gingivalis in biological sam-
ples. Infect. Immun. 54:798-803.
3. Chen, P., and N. Doroszczak. 1982. Differences in lymphopro-
liferative responses to the bacterium Actinomyces viscosus in
various mammalian species. Arch. Oral Biol. 27:319-324.
4. Chen, P. B., M. E. Neiders, S. J. Millar, H. S. Reynolds, and
J. J. Zambon. 1987. Effect of immunization on experimental
Bacteroides gingivalis infection in a murine model. Infect.
5. Dubois, M., K. A. Gilles, J. K. Hamilton, P. A. Rebers, and F.
Smith. 1956. Colorimetric method for determination of sugars
and related substances. Anal. Biochem. 28:350-356.
6. Eisenstein, T. K., L. M. Miller, and B. M. Sultzer. 1984.
Immunity to infection with Salmonella typhimurium. Mouse-
strain difference in vaccine- and serum-mediated protection. J.
Infect. Dis. 150:425-435.
7. Elkins, K. L., P. W. Stashak, and P. J. Baker. 1987. Prior
exposure to subimmunogenic amounts of some bacterial lipo-
polysaccharides induces specific immunological unresponsive-
ness. Infect. Immun. 55:3085-3092.
8. Grenier, D., and D. Mayrand. 1987. Selected characteristics of
pathogenic and nonpathogenic strains ofBacteroides gingivalis.
J. Clin. Microbiol. 25:738-740.
9. Hitchcock, P. J., and T. M. Brown. 1983. Morphological heter-
ogeneity among Salmonella lipopolysaccharide chemotypes in
silver-stained polyacrylamide gels. J. Bacteriol. 154:269-277.
10. Isibasi, A., V. Ortiz, M. Varga, J. Paniagua, C. Gonzalez, J.
Moreno, and J. Kumate. 1988. Protection against Salmonella
typhi infection in mice after immunization with outer membrane
proteins isolated from Salmonella typhi 9,12,d,Vi. Infect. Im-
11. Kastelein, P., T. J. M. van Steenbergen, T. J. M. Bras, and J.
DeGraaff. 1981. An experimentally induced phlegmonous ab-
scess by a strain of Bacteroides gingivalis in guinea pigs and
mice. Antonie van Leeuwenhoek 47:1-9.
12. Kimura, A., P. A. Gulig, G. McCracken, Jr., T. A. Loftus, and
E. J. Hansen. 1985. A minor high-molecular-weight outer mem-
brane protein of Haemophilus influenzae type b is a protective
antigen. Infect. Immun. 47:253-259.
13. Laemmli, U. K. 1970. Cleavage of structural proteins during the
assembly of the head of bacteriophage T4. Nature (London)
VOL. 58, 1990
CHEN ET AL.
14. Lowry, 0. H., N. J. Rosebrough, A. L. Farr, and R. J. Randall.
1951. Protein measurement with the Folin phenol reagent. J.
Biol. Chem. 193:265-275.
15. Lu, Y. S., S. J. Afendis, and S. P. Pakes. 1988. Identification of
immunogenic outer membrane proteins of Pasteurella multo-
cida 3:A in rabbits. Infect. Immun. 56:1532-1537.
16. Marchesi, V. T., and E. P. Andrews. 1971. Glycoproteins:
isolation from cell membranes with lithium diiodosalicylate.
17. Matthews-Greet, J. M., and H. E. Gilleland. 1987. Outer mem-
brane protein F (porin) preparation ofPseudomonas aeruginosa
as a protective vaccine against heterologous immunotype
strains in a burned mouse model. J. Infect. Dis. 155:1282-1291.
18. Millar, S. J., P. B. Chen, and E. Hausmann. 1987. Monoclonal
antibody for identification ofBacteroides gingivalis lipopolysac-
charide. J. Clin. Microbiol. 25:2437-2439.
19. Miliar, S. J., E. G. Goldstein, M. J. Levine, and E. Hausmann.
1986. Modulation of bone metabolism by two chemically dis-
tinct lipopolysaccharide fractions from Bacteroides gingivalis.
Infect. Immun. 51:302-306.
20. Miller, R. G., Jr. 1981. Simultaneous statistical inference, 2nd
ed., p. 90-94. Springer-Verlag, New York.
21. Montarz, A. J., and A. J. Winter. 1986. Comparison of living
and nonliving vaccines for Brucella abortus BALB/c mice.
Infect. Immun. 53:245-251.
22. Mukkur, T. K. S., G. H. McDowell, B. A. D. Stoker, and A. K.
Lasceiles. 1987. Protection against experimental salmonellosis in
mice and sheep by immunization with aromatic-dependent Sal-
monella typhimurium. J. Med. Microbiol. 24:11-19.
23. Munson, R. S., Jr., J. L. Shenep, S. J. Barenkamp, and D. M.
Granoff. 1983. Purification and comparison of outer membrane
protein P2 from Haemophilus influenzae type b isolates. J. Clin.
24. Murray, P. A., M. J. Levine, M. S. Reddy, L. A. Tabak, and
E. J. Bergey. 1986. Preparation of a sialic acid-binding protein
from Streptococcus mitis KS32AR. Infect. Immun. 53:359-365.
25. Neiders, M. E., P. B. Chen, H. Reynolds, J. J. Zambon, M.
Shlossman, and R. J. Genco. 1989. Heterogeneity of virulence
among strains of Bacteroides gingivalis. J. Periodontal Res.
26. Okuda, K., T. Kato, Y. Naito, I. Takazoe, Y. Kikuchi, T.
Nakamura, T. Koyoshige, and S. Sasaki. 1988. Protective ef-
ficacy of active and passive immunizations against experimental
infection with Bacteroides gingivalis in ligated hamsters. J.
Dent. Res. 67:807-811.
27. Saukkonen, K., H. Abdillahi, T. Poolman, and M. Leinonen.
1987. Protective efficacy ofmonoclonal antibodies to class 1 and
class 3 outer membrane proteins of Neisseria meningitidis
B:15:PI.16 in infant rat infection model: new prospects for
vaccine development. Microb. Pathog. 3:261-267.
28. Shapiro, M. F., A. B. Onderdonk, D. L. Kasper, and R. W.
Finberg. 1982. Cellular immunity to Bacteroides fragilis capsu-
lar polysaccharide. J. Exp. Med. 155:1188-1197.
29. Udhayakumar, V., and V. R. Muthukkaruppan. 1987. Protective
immunity induced by outer membrane proteins of Salmonella
typhimurium in mice. Infect. Immun. 55:816-821.
30. van Steenbergen, T. J. M., F. G. A. Delemarre, F. Namavar, and
J. DeGraaff. 1987. Differences in virulence within the species
Bacteroides gingivalis. Antonie van Leeuwenhoek 53:233-244.
31. van Steenbergen, T. J. M., P. Kastelein, J. J. A. Touw, and J.
DeGraaff.1982. Virulence of black-pigmented Bacteroides
strains from periodontal pockets and other sites in experimen-
tally induced skin lesions in mice. J. Periodontal Res. 17:41-49.
32. Vuopio-Varkila, J., M. Karvonean, and H. Saxen. 1988. Protec-
tive capacity of outer-membrane components of Escherichia
coli in a systemic mouse peritonitis model. J. Med. Microbiol.
33. Williams, G. D., and S. C. Holt. 1985. Characteristics of the
outer membrane of selected oral Bacteroides species. Can. J.
34. Yoshimura, F., K.-I. Watanable, T. Takasawa, M. Kawanami,
and H. Kato. 1989. Purification and properties ofa 75-kilodalton
major protein, and immunodominant surface antigen, from the
oral anaerobe Bacteroides gingivalis. Infect. Immun. 57:3646-