INFECTION AND IMMUNITY, Mar. 1985, p. 638-6470019-9567/85/030638-10$02.00/0Copyright ©3 1985, American Society for Microbiology

Biochemical and Immunobiological Properties of Lipopolysaccharide(LPS) from Bacteroides gingivalis and Comparison with LPS from

Escherichia coliTOSHIHIKO KOGA, TATSUJI NISHIHARA, TAKU FUJIWARA, TOSIKI NISIZAWA, NOBUO OKAHASHI,

TOSHIHIDE NOGUCHI, AND SHIGEYUKI HAMADA*

Department of Dental Research, The National Institute of Health, Kamiosaki, Shinagawa-ku, Tokyo 141, Japan

Received 28 August 1984/Accepted 28 November 1984

Lipopolysaccharides (LPSs) were isolated from Bacteroides gingivalis and Escherichia coli by the phenol-water and butanol-water procedures. The phenol-water-extracted LPS from B. gingivalis 381 was composed of46% carbohydrate, 23% hexosamine, 18% fatty acid, and 5% protein. The major component sugars of thispreparation were glucose, glucosamine, rhamnose, galactose, galactosamine, and mannose, and theirmolecular ratio was 1:0.9:0.7:0.6:0.6:0.4, respectively. Neither heptose nor 2-keto-3-deoxyoctonate was

detected. the butanol-water-extracted LPS from this strain was composed of 76% glucose, 7% fatty acid, and13% protein, and it was associated with a number of polypeptides (13 to 56 kilodaltons). The main fatty acidof both LPS preparations was palmitic acid. It was found that biological activities of LPS from B. gingivaliswere comparable to those of LPS from E. coli in terms of activation of the clotting enzyme of Limulusamebocyte lysate, mitogenicity, polyclonal B cell activation, and stimulation of interleukin 1 production inBALB/c mice. Furthermore, LPS-nonresponsive C3H/HeJ spleen cells were found to yield good mitogenicresponses to both phenol-water-extracted LPS and butanol-water-extracted LPS from B. gingivalis or

butanol-water-extracted LPS from E. coli. On the other hand, spleen cells from LPS-responsive C3H/HeN miceresponded well to all these LPS preparations.

Lipopolysaccharide (LPS), localized in the cell wall ofgram-negative bacteria, is known to have a broad spectrumof biochemical and immunochemical activities (12, 23). TheLPS molecule from bacteria of the family Enterobacteria-ceae consists of a polysaccharide region covalently bound toa lipid region, termed lipid A (5). The lipid A contains acentral backbone of the phosphorylated (1- 6) linkedglucosamine disaccharide. The polysaccharide region can bedivided into a side chain (O antigen) and a core region. Thecore polysaccharide of most LPS preparations consists ofglucose, galactose, 2-acetamido-2-deoxyglucose, and hep-tose, which are linked to 2-keto-3-deoxyoctonate (KDO) oflipid A (5).

It has been reported that the composition of LPS fromBacteroides species is unique; it lacks KDO, heptose, and3-hydroxy myristic acid, which are unique to enterobacterial

LPS, and its biological potency is considerably lower thanthat of enterobacterial LPS (15, 18, 24). Recently, Joiner etal. (13) reported that LPS from Bacteroides fiagilis is a

potent mitogen for spleen cells from both LPS responder(C57B/1OScN) and nonresponder (C57BL/1OScCR andC3H/HeJ) mice. On the contrary, Wannemuehler et al. (34)have shown that LPS from B. fragilis induces mitogenicresponses in spleen cells from LPS responder (C3H/HeN)but not LPS nonresponder (C3H/HeJ) mice.

In this study we isolated LPS from Bacteroides gingihalis,one of the major pathogenic bacteria of periodontal diseases(7, 8, 29, 30), by the phenol-water (35) and butanol-water(22) extraction procedures. Furthermore, we comparedchemical, biological, and immunochemical properties ofLPS preparations from B. gingivalis and Escherichia coli.

* Corresponding author.

MATERIALS AND METHODS

Bacterial strains and growth conditions. B. gingivalis 381(originally isolated by S. Sockransky, Forsyth Dental Cen-ter, Boston, Mass.) and E. coli strain K235 (19) were ob-tained from Y. Yamamoto, Sunstar Co., Osaka, Japan, andJ. R. McGhee, University of Alabama, Birmingham, Ala.,respectively. B. gingivalis was grown in Trypticase soy broth(BBL Microbiology Systems, Cockeysville, Md.) supple-mented with yeast extract (15 g/liter), hemin (5 mg/liter), andmenadione (1 mg/liter) at 37°C for 4 days in a 5% CO2atmosphere. E. coli was grown aerobically in Trypticase soybroth at 37°C for 24 h. Both cells were harvested by centrif-ugation, washed three times with distilled water, and lyoph-ilized.

Mice. BALB/c, BALB/c (ni/lni), and C3H/HeJ mice were

obtained from the Department of Veterinary Science, Na-tional Institute of Health, Tokyo, Japan. C3H/HeN micewere purchased from CLEA Japan, Ltd., Tokyo. Youngmale mice (5 to 7 weeks) were used throughout this study.

Extraction of LPS. Lyophilized cells (10 g) were sus-

pended in 350 ml of pyrogen-free water and 350 ml of 90%phenol (35). The mixture was stirred vigorously at 65°C for20 min and then centrifuged at 7,000 x g for 20 min. Theaqueous phase was removed, and the phenol phase andinsoluble precipitate were reextracted with 350 ml of water.The combined solution of aqueous phase was dialyzedextensively against distilled water, lyophilized, and was

termed crude phenol-water-extracted LPS (PW-LPS).Lyophilized cells (10 g) were suspended in 240 ml of 0.85%

NaCI. An equal volume of water-saturated butanol was thenadded (23). The reaction mixture was stirred for 15 min at4°C and centrifuged at 7,000 x g for 30 min at 4°C. Theaqueous phase was removed, and the butanol phase and

638

Vol. 47, No. 3

on March 2, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

IMMUNOBIOLOGY OF B. GINGIVALIS LPS 639

precipitate were reextracted twice with 120 ml of 0.85%NaCI. The combined extracts of the aqueous phase weredialyzed against distilled water and lyophilized. This wastermed crude butanol-water-extracted LPS (BW-LPS).

Purification of LPS. Crude LPS (1 g) was suspended in 100ml of pyrogen-free water and centrifuged at 100,000 x g for3 h. The pellet was suspended in 20 ml of 10 mM Tris buffer(pH 7.4) containing 0.1 mM ZnC12 and 400 p.g of nuclease P1from Penicillium citrinum (Yamasa Shoyu Co., Ltd., Choshi,Japan). The reaction mixture was incubated in 37°C for 16 hand then dialyzed extensively against distilled water. Thedialyzed solution was centrifuged. The pellet was washedwith pyrogen-free water by centrifugation and lyophilized.Enzyme treatment of LPS. PW-LPS from B. gingivalis (1

mg/ml) was treated with equal amounts of trypsin (Boehrin-ger Mannheim GmbH, Mannheim, Federal Republic ofGermany) or pronase E (Kaken Kagaku Ltd., Tokyo, Japan)for 18 h at 37°C at their optimal pHs. The enzyme activitywas terminated by heating the reaction mixture at 100°C for5 min, followed by dialysis against distilled water, and thenlyophilized. The lyophilized LPS was used for the mitogenassay (see below).

Chemical analyses. Hexose, deoxyhexose, hexosamine,protein, and phosphorus were quantitated as previouslydescribed (26). Heptose, KDO, and fatty acid esters werequantitated by the colorimetric methods of Wright andRebers (36), Karkhanis et al. (14), and Itaya and Ui (9),respectively.

For sugar and fatty acid analyses, LPS (2 mg) washydrolyzed for 6 h at 100°C in 2 ml of 2 N HCI. Thehydrolysate was suspended in 10 ml of a solvent consistingof chloroform, methanol, and water (4:10:5[vol/vol]). Themixture was filtered through a filter paper to remove coarsedebris. After chloroform (10 ml) and water (10 ml) wereadded to the filtrate, the mixture was centrifuged at 5,000 xg for 20 min. Each of the aqueous and chloroform phaseswas dried by evaporation and used for the sugar and fattyacid analyses.

Neutral and amino sugars were identified as trimethylsilylderivatives (26) and alditol acetate derivatives (25, 27). Thetrimethylsilyl derivatives were analyzed by a gas-liquidchromatograph (Shimadzu Works, Ltd., Kyoto, Japan) fittedwith a glass column (3 mm by 2.6 m) of 10% silicone GESE-30 on Chromosorb W (60 to 80 mesh; Wako PureChemical Industries, Ltd., Osaka, Japan). On the otherhand, the alditol acetate derivatives were chromatographedon 3% ECNSS-M on Gas-Chrom Q (100 to 200 mesh; Wako)in a glass column (3 mm by 2.1 m).

Fatty acids were identified as their methyl esters (4). Thedried chloroform phase was treated with 4 ml of 5% HCI inmethanol at 105°C for 3 h in sealed tubes and dried under astream of N2 gas. After the addition of 2% potassiumbicarbonate solution, the esters were extracted into hexaneand quantitated in a gas chromatograph, fitted with glasscolumns (3 mm by 2.1 m) of 25% diethyleneglycolsuccinateon Chromosorb W (60 to 80 mesh; Wako) and 1.5% siliconeGE SE-30 on Chromosorb W (60 to 80 mesh; Wako) at1700C.For amino acid analysis, LPS preparations (1 mg each)

were hydrolyzed in 6 N HCI, in vacuo, at 110°C for 24 h, andthen analyzed by a Hitachi 835-50 amino acid analyzer(Hitachi Ltd., Tokyo, Japan) (26).

SDS-polyacrylamide gel electrophoresis. Sodium dodecylsulfate (SDS)-polyacrylamide slab gels were prepared asdescribed by Laemmli (16). LPS preparations (150 jig each)were heated at 100°C for 5 min in 10 mM Tris-hydrochloride

buffer (pH 6.8) containing 2.5% SDS, 5% 2-mercaptoeth-anol, and 20% glycerol. Electrophoresis was performed at aconstant voltage of 150 V for 5 h with 12.5% resolving and4% stacking gels. Proteins were stained with Coomassiebrilliant blue R-250. A low-molecular-weight electrophoresiscalibration kit (Pharmacia Fine Chemicals, Uplsala, Swe-den) was used for molecular weight standards.Limulus amebocyte lysate clotting activity. Colorimetric

blood endotoxin determination reagent (Pyrodik; SeikagakuKogyo Co. Ltd., Tokyo, Japan) was used for the measure-ment of Lirnulius amebocyte lysate clotting activity. Thereagent consists of Limulus amebocyte lysate and a syn-thetic chromogenic substrate (protecting group-leucine-glycine-arginine-p-nitroanilin) (10). The reagent (0.1 ml) and0.1 ml (10 pg) of LPS to be tested were incubated for 30 minat 37°C. After the reaction was terminated by adding 0.6 Nacetic acid (1 ml), the A405 was measured. LPS from E. coli0127:B8 (Sigma Chemical Co., St. Louis, Mo.) was used asa positive control.

Mitogenic activity. Murine spleen cells from BALB/c,BALB/c (nu/nu), C3H/HeJ, and C3H/HeN mice and thymo-cytes from BALB/c mice were suspended in RPMI 1640medium (GIBCO Laboratories, Grand Island, N.Y.) supple-mented with penicillin (100 U/ml), streptomycin (100 .ag/ml),and N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid(HEPES) buffer (pH 7.2, 15 mM), and adjusted to appropri-ate cell concentrations before culture. Lymphoid cells werecultured in triplicate in a 96-well microculture plate in a totalvolume of 0.2 ml of the medium, containing various amountsof LPS or concanavalin A (ConA; Sigma), at 37°C in ahumidified 5% CO2 atmosphere. Cultures were pulsed with0.25 p,Ci of [3H]thymidine ([3H]Td1; 24 Ci/mmol; Amer-sham Corp., Buckinghamshire, United Kingdom) during thefinal 6 h of a 48-h incubation. After the labeling period, cellswere harvested onto glass fiber filters by using a multiple-cellharvester. These filters were dried, placed in vials withscintillation fluid, and counted on an Aloka scintillationcounter.The effect of polymyxin B on nlitogenic responses of

C3H/HeJ and C3H/HeN spleen cells was examined as fol-lows. Spleen cells (7 x 105) were cultured 37°C for 48 h in 0.2ml of the RPMI 1640 medium, containing various amounts ofpolymyxin B sulfate (104U/mg; Taito Pfizer Co., Ltd., To-kyo, Japan) and 5 ,ug of stimulant (PW-LPS or ConA).Cultures were pulsed for the final 6 h of incubation with[3H]TdR (0.25 ,uCi), and the amount of [3H]TdR uptake wasdetermined as described above.Immune responses in mice. In vitro polyclonal responses to

LPS of murine spleen cells and adjuvant activities of LPSwere examined as follows. Spleen cells from BALB/c micewere suspended in RPMI 1640 medium supplemented withpenicillin (100 U/ml), streptomycin (100 [Lg/ml), 2-mercapto-ethanol (0.05 mM), and 20% fetal calf serum (FCS; ArmourPharmaceutical Co., Kankakee, Ill.). Cells (5 x 106) wereprecultured for 0 to 24 h in a 24-well microculture plate in atotal volume of 0.6 ml of the medium with or without sheeperythrocytes (SRBC; 2 x 106), and then LPS preparations(10 pg each) were added. After incubation for 4 days at 37°Cin an atmosphere of 5% CO2 in air, nonadherent cells wereremoved from each culture well, washed once with incom-plete minimal essential medium (GIBCO), and resuspendedto a defined volume with minimal essential medium for theplaque-forming cell (PFC) assay.

For estimation of in vivo polyclonal responses to LPS,BALB/c mice were injected intravenously with LPS (100p.g). After 4 days, the spleen of the mice was removed.

VOL. 47, 1985

on March 2, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

640 KOGA ET AL.

TABLE 1. Chemical composition of LPSs from B. gingivalis 381and E. coli K235

% (Dry wt) LPS from:Componenta B. gingivalis 381 E. coli K235

PW-LPS BW-LPS PW-LPS BW-LPS

Neutral sugarRhamnose 9.9 0.8 0.9 0.7Fucose b _ 12.8 4.8Mannose 5.6 - 4.1 1.8Galactose 9.7 8.7 3.7Glucose 15.0 76.1 13.0 4.8Heptose 7.2 2.7KDO - 5.9 2.1Unknown Ac 0.1Unknown Bc 5.5 0.2

Amino sugarGlucosamine 13.8 0.4 18.9 9.6Galactosamine 8.7 - -

Fatty acid 18.3 6.8 25.6 22.0

Protein 5.2 12.8 0.6 45.4

Phosphorus 1.5 0.2 2.5 1.1a Neither ribose nor deoxyribose was detected by gas-liquid chromatogra-

phy.b-, Not detected.c Calculated by assuming that the response adjustment factor of unknown

sugars was the same as that of glucose with gas-liquid chromatography.

Spleen cells were washed and resuspended in minimalessential medium for PFC assay.PFC assay. Nonadherent cells remnoved from in vitro

cultures or spleen cells from immunized mice were assayedin triplicate for direct anti-SRBC PFC responses with thehemolytic plaque technique of Cunningham and Szenberg(1).

Induction of IL-1 secretion. Induction of interleukin 1(IL-1) secretion was done as follows. Macrophages ofBALB/c mice were induced by intraperitoneal injection of1.5 ml of thiogylcolate medium (Difco Laboratories, Detroit,Mich.). Three days later, peritoneal exudate cells werecollected by peritoneal lavage with 5 ml of Hanks balancedsalt solution (GIBCO) containing 1% FCS, washed, andresuspended in RPMI 1640 medium supplemented withpenicillin (100 U/ml), streptomycin (100 p.g/ml), 25 mM

TABLE 2. Fatty acid composition of LPSs from B. gingivalis 381and E. coli K235% (Dry wt) total fatty acids from:

Fatty B. gingivalis 381 E. coli K235acid

PW-LPS BW-LPS PW-LPS BW-LPS

C12:0 -a 5.1 4.0C14:0 4.5 8.3 8.3C14:OH 69.7 47.0C15:iso 3.4 3.9C160 41.1 65.2 10.7 25.5C6l --11.9 3.0C17:iso 15.5 11.2C17:0 2.8 4.0C18:0 37.3 15.2 1.9 2.6C18:1 2.4 5.6a-, Not detected.

HEPES buffer (pH 7.2), and 5% FCS. The peritoneal cells(2.4 x 106) were plated in 35-mm petri dishes at 37°C in anatmosphere of 5% CO2 in air for 2 h and washed three timeswith Hanks balanced salt solution containing 1% FCS. Theadherent cells were incubated in RPMI 1640 medium with orwithout stimulant for 24 h. Supernatants were collected andcentrifuged for 10 min at 1,000 x g.

IL-1 activity in culture supernatants was determined asdescribed by Vacheron et al. (33). In brief, thymocytes fromBALB/c mice were suspended to a density of 7 x 106 cellsper ml in RPMI 1640 medium containing antibiotics andHEPES. The cell suspension (0.1 ml) was seeded into a96-well microculture plate, and an equal volume of serialdilutions of macrophage culture supernatants was added.Cultivation was carried out with or without the addition of asubmitogenic concentration of ConA (1 ,ug/ml) and incu-bated in an atmosphere of 5% CO2 in air for 48 h. They werepulsed for the final 6 h with 0.25 ,uCi of [3H]TdR.

RESULTSChemical composition of LPSs from B. gingivalis and E.

coli. Purified PW-LPS from B. gingivalis 381 was composedof 15.0% glucose, 13.8% glucosamine, 9.9% rhamnose, 9.7%galactose, 8.7% galactosamine, 5.6% mannose, 5.5% un-known sugar, 18.3% fatty acid, 5.2% protein, and 1.5%phosphorus (Table 1). No heptose, KDO, ribose, or deoxy-ribose was detected. The molar ratio of glucose, glucos-amine, rhamnose, galactose, galactosamine, and mannosewas 1.0:0.9:0.7:0.6:0.6:0.4. respectively. On the other hand,BW-LPS from this strain was composed of 76.1% glucose,6.8% fatty acid, and 12.8% protein. Amino sugar and phos-phorus contents of the BW-LPS were lower than those of thePW-LPS. The major fatty acids of LPS from B. gingivaliswere C16:0 (palmitic), C18:0 (stearic), and C17:iso (Table 2).The amino acid composition of LPS preparations is shown inTable 3. The major amino substance of PW-LPS from B.gingivalis was ethanolamine, whereas in BW-LPS from thisstrain, aspartic acid, threonine, serine, glutamic acid, glycine,alanine, and leucine were predominat.

TABLE 3. Amino acid composition of LPSs from B. gingivalis381 and E. coli K235a

Composition (nmol/mg) in:

Amino acid B. gingivalis 381 E. coli K235PW-LPS BW-LPS PW-LPS BW-LPS

Aspartic acid 29 62 2 294Threonine 30 50 1 192Serine 51 39 3 135Glutamic acid 35 48 3 254Glycine 22 48 45 280Alanine 39 49 b 246Valine - 38 169Methionine 3 1 1Isoleucine 10 29 - 106Leucine 18 43 1 186Tyrosine - 28Phenylalanine 10 25 2 95Lysine 12 34Histidine 3 9 - 41Arginine 8 21Ethanolamine 390 10 82 51NH3 328 116 343 479

a LPS preparations (1 mg each) were hydrolyzed in 6 N HC1 at 100°C for 14h in a vacuum-sealed hydrolysis tube.b , Not detected.

INFECT. IMMUN.

on March 2, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

IMMUNOBIOLOGY OF B. GINGIVALIS LPS 641

MW 1 2 3 4 5(x103)94_-d67-..-

67 ~w __43-op

20-*m

FIG. 1. SDS-polyacrylamide gel electrophoresis of LPS prepa-rations from B. gingivalis 381 and E. coli K235. Electrophoresis was

performed with 12.5% acrylamide. containing 150 ,ug of each LPS.The gel was stained with Coomassie brilliant blue. Lane 1, Molec-ular weight markers; lane 2, PW-LPS from E. coli; lane 3, BW-LPSfrom E. coli; lane 4, PW-LPS from B. gingivalis; lane 5, BW-LPSfrom B. gingivalis.

The component sugars of LPS from E. coli K235 were

glucosamine, glucose, fucose, galactose, heptose, KDO,mannose, and rhamnose (Table 1). Neither ribose nor deoxy-ribose was detected. The BW-LPS contained 45% protein,whereas the PW-LPS contained little protein. The mainconstituent fatty acids of these preparations were C14:OH(P-hydroxy myristic) and C16:0 (palmitic) acids (Table 2). Theamino acid composition of E. coli LPS differed from that ofB. gingivalis LPS (Table 3). The ethanolamine content ofPW-LPS from E. coli was considerably lower than that of B.gingivalis PW-LPS.

SDS-polyacrylamide gel electrophoresis. Coomassie bluestaining of SDS-polyacrylamide slab gels (Fig. 1) revealedmany protein bands in BW-LPSs from B. gingivalis and E.coli but no detectable bands in PW-LPSs from both species.Proteins in BW-LPS from E. coli appeared as a majorpeptide band at 13 kilodaltons with minor peptides of 37, 28,14, and 12 kilodaltons (Fig. 1, lane 3). On the other hand,BW-LPS from B. gingiivalis associated with major polypep-tides of 56, 44, 30, 28, 19, and 13 kilodaltons (Fig. 1, lane 5).

Activation of clotting enzyme in Limulus amebocyte lysateby LPS. Limulus amebocyte lysate clotting activity of LPSpreparations was measured by using a colorimetric endo-toxin determination reagent. PW-LPS preparations from E.coli 0127:B8 and B. gingivalis strongly activated the clottingenzyme (Table 4). However, both LPS preparations from E.coli K235 had a little weaker activity in comparison withthese PW-LPS preparations. BW-LPS from B. gingivalis didnot significantly activate the enzyme at the concentrationemployed.

Mitogenic activity of LPS. The mitogenic activities of LPSpreparations from B. gingivalis and E. coli were determinedin spleen cells of BALB/c, BALB/c (nu/nu), C3H/HeN, andC3H/HeJ mice and thymocytes of BALB/c mice.

All LPS preparations examined were found to be stronglymitogenic for BALB/c, BALB/c (nu/nu), and C3H/HeNmurine spleen cells at concentrations between 5 and 50jig/ml (Fig. 2 through 5), whereas no LPS preparations weremitogenic for thymocytes of BALB/c mice (Fig. 2). Theabove results indicate that these LPS preparations are B cellmitogens in murine spleen cells.

BW-LPS from E. coli K235 was mitogenic for C3H/HeJspleen cells, but PW-LPS from this strain was found to benonmitogenic for the spleen cells at all doses (5 to 50 p.g/ml)and cell densities (8 x 104 to 7 x 105 cells per well) tested(Fig. 4 and 5). On the other hand, all LPS preparations fromB. gingivalis, whatever methods for their extraction, in-duced significant mitogenic responses in C3H/HeJ spleencells when 7 x 105 cells per well were employed (Fig. 4).Even at lower cell densities, PW-LPS from B. gingivalisinduced good mitogenic responses in C3H/HeJ spleen cells,although these responses were a little lower than those inC3H/HeN cell cultures (Fig. 5).

Contamination of mitogenically active proteins in PW-LPS from B. gingivalis could be attributed to the mitogeni-city for C3H/HeJ spleen cells. Therefore, the effects oftrypsin and pronase E on PW-LPS from B. gingivalis wereexamined. The mitogenicity of the PW-LPS, as measured inC3H/HeJ spleen cells, was considerably resistant to theproteolytic effects of these enzymes (Fig. 6). These resultsindicate that the mitogenicity of PW-LPS from B. gingivalisfor C3H/HeJ lymphocytes is not attributable to contamina-tion of such proteins.The effects of polymyxin B on mitogenic responses of

C3H/HeN and C3H/HeJ spleen cells were examined. Addi-tion of polymyxin B to cultures significantly inhibited themitogenicity of PW-LPS from E. coli for C3H/HeN spleencells, but had virtually no effect on PW-LPS from B.gingi'alis and ConA (Fig. 7A). All of these mitogens werescarcely affected by the addition of polymyxin B to C3H/HeJspleen cell cultures (Fig. 7B).

In vitro and in vivo immune response to SRBC. The effectsof LPS on the in vitro immune response of BALB/c mousespleen cells to SRBC were studied by the hemolytic plaqueassay. All LPS preparations induced in vitro direct (immuno-globulin M) PFC responses to SRBC (Table 5), indicatingthat they are polyclonal B cell activators. The numbers ofPFCs induced by PW-LPSs from both species were signifi-cantly higher than those induced by BW-LPS preparations.For determining the optimal condition for adjuvant assay,BALB/c spleen cells (2 x 106 cells per well) and BW-LPS (10p.g) from B. gingivalis were added. The peak of the responsespecific for SRBC was obtained at 6 h of the preculture timewith SRBC (Fig. 8). Therefore, the spleen cells were cul-tured with SRBC and 6 h later with LPS preparations fordetermining adjuvant activity. As shown in Table 5, PW-LPS and BW-LPS from B. gingivalis and PW-LPS from E.coli strongly enhanced responses to SRBC. PW-LPS andBW-LPS from B. gingivalis are much higher in activity thanthose from E. coli. The adjuvant effect of BW-LPS from E.coli was weak.

TABLE 4. Activation of clotting enzyme of Limilius amebocytelysate by LPSs from B. gingivalis 381 and E. cOli K235 and

0127:B8

LPS (10 pg Activity ofOrganism tube) clotting enzyme

None 0.033B. gingivalis 381 PW-LPS 1.090

BW-LPS 0.045E. coli K235 PW-LPS 0.233

BW-LPS 0.666E. coli 0127:B8 PW-LPS 1.077

" Colorimetric endotoxin determination reagent (0.1 ml) and 0.1 ml of LPSpreparation (0.1 ng/ml) were incubated for 30 min at 37°C. After the reactionwas terminated by adding 1 ml of 0.6 N acetic acid, the A40, was measured.

VOL. 47, 1985

on March 2, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

642 KOGA ET AL.

A. Spleen cells B.Thymocytes

None F=~~~~~~.r

Con AB.gingivalis 381

PW-LPS

BW-LPSEcoli K 235 -

PW-LPS

BW-LPS

0 10 20 30 50e H) TdR Uptake

(kcpm)

I

0 10 20HJTdR Uptake(kcpm)

FIG. 2. Mitogenic responses of BALB/c mouse spleen cells (A) and thymocytes (B) to LPSs from B. gingivalis 381 and E. coli K235.Spleen cells or thymocytes (7 x 105) were cultured for 48 h with 1 p.g (O), 5 ,ug (LI), or 10 ,ug (O) of stimulant per well or without stimulant(U) in 0.2 ml of RPMI 1640 medium. Data are expressed as the mean ± standard error (SE) of triplicate cultures.

BALB/c mice were injected intravenously with LPS (100,ug), and 4 days later, at the peak response, the number ofPFCs to SRBC was assayed by using the spleen of eachmouse. All LPS preparations caused a significant increase inthe number of PFCs to SRBC (data not shown).

Induction of IL-1 secretion by LPS. IL-1 activity wasassessed by measuring the incorporation of [3H]TdR bycultured murine thymocytes. PW-LPSs from B. gingivalis

and E. coli induced mouse peritoneal cells to produce IL-1activity. The macrophage culture supernatants were mito-genic and enhanced the stimulation ofBALB/c mouse thymo-cytes in the presence of a submitogenic concentration ofConA (1 pug/ml) (Fig. 9 and Table 6). This enhancement was

not due to the combined action of ConA and PW-LPS addedto the peritoneal cells-that is, PW-LPSs at 0.5 to 50 ,I.g/mlwere not mitogenic for thymocytes in the presence of 1 .g

None e

ConA

B.girngivalis 381

PW-LPS ,

BW-LPS -I

E coli K235PW-LPS

BW-LPS .ZZ ; ;; ;

0 10 20 30 40 50(3 H)TdR Uptake ( kcpm)

FIG. 3. Mitogenic responses of BALB/c (nu/lni) mouse spleen cells to LPSs from B. gingiv'alis 381 and E. coli K235. Spleen cells (7 x 105)were cultured for 48 h with 1 ,ug (O), 5 jig (L), or 10 ,ug (LI) of stimulant per well or without stimulant (U) in 0.2 ml of RPMI 1640 medium.Data are expressed as the mean ± SE of triplicate cultures.

I:::::::: ::::---

b . - . Ar r

77...H

I

INFECT. IMMUN.

c

on March 2, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

IMMUNOBIOLOGY OF B. GINGIVALIS LPS 643

A. C3H/HeN

None IP

Co n A .-" //////

a.gingivalis381 ......-.-.

PW-LPS

BW-LPSSE.coli K235 s....

PW- LPS

BW-LPS.

6 20 40 60 80(3H)TdR Uptake

( kcpm )

B. C3H/HeJ

r.r:..:*; .* -.-: -.-:;--:: .:-: . .::::::.. .j

t:;:.-..- . .}l

6 20 40 60 so i6o(3H) TdR Uptake

(kcpm)

FIG. 4. Mitogenic responses of C3H/HeN (A) and C3H/HeJ (B) mouse spleen cells to LPSs from B. gingivalis 381 and E. coli K235. Spleencells (7 x 105) were cultured for 48 h with 1 ,ug (O), 5 jig (O), or 10 ,ug (O) of stimulant per well or without stimulant (U) in 0.2 ml of RPMI1640 medium. Data are expressed as the mean ± SE of triplicate cultures.

of ConA per ml (Table 7). In addition, no IL-1 productionwas observed when BW-LPSs from B. gingivalis and E. coliwere used at 10 and 100 ,uglml.

DISCUSSIONMany investigators have used C3H/HeJ mice as an exper-

imental model for analyses of LPS responses (20). Thismouse strain is markedly resistant to lethal effects of LPS(31). It has been considered that this unresponsiveness ofC3H/HeJ mice to LPS results from a generalized inability ofC3H/HeJ lymphoid cells (e.g., B lymphocytes and macro-phages) and nonlymphoid cells (e.g., embryonic fibroblasts)to respond to LPS (20). However, it was observed byMorrison et al. (21) and Sultzer and Goodman (32) thatLPS-protein complexes are mitogenic for C3H/HeJ B lym-phocytes. This protein component (lipid A-associated pro-tein) consists of a number of distinct components withmolecular weights from ca. 8,000 to 40,000 which have beenfound on the outer cell membrane (20). In this study, weshowed that BW-LPSs from E. coli and B. gingivalis, whichassociated with a number of polypeptides, were mitogenicfor C3H/HeJ mouse spleen cells (Fig. 1 and 4), whereasreference PW-LPS from E. coli was not mitogenic for thespleen cells, as was reported by others (20, 21). Of interestwas that protein-free PW-LPS from B. gingivalis was alsomitogenic for C3H/HeJ spleen cells (Fig. 1, 4, and 5).However, the presence of trace amounts of mitogenicallyactive protein in the PW-LPS cannot be excluded. There-fore, we treated the PW-LPS exhaustively with trypsin andpronase. However, the mitogenicity of the PW-LPS was notdiminished by the treatment with these proteolytic enzymes(Fig. 6).

Recently, Joiner et al. (13) reported that PW-LPS from B.fragilis also was a potent mitogen for spleen cells fromboth LPS responder (C57BL/1OScN) and nonresponder(C57BL/1OScCR and C3H/HeJ) mice. More recently, Wan-nemuehler et al. (34) have reported that PW-LPS from B.fragilis induced mitogenic responses in both C3H/HeN andC3H/HeJ spleen cell cultures when 8 x 105 cells per well

were employed, but a typical responsive-nonresponsive pat-tern develops when lower numbers of spleen cells (8 x 104 to5 x 105 cells per well) are tested with this LPS preparation.On the basis of these results, they concluded that themitogenic response pattern of Bacteroides lipid A is similarto that reported for either E. coli lipid A or Salmonella lipidA. However, PW-LPS from B. gingivalis induced significantmitogenic responses in C3H/HeJ spleen cells even at low celldensities (8 x 104 to 5 x 105 cells per well) (Fig. 5).Moreover, polymyxin B, known to inhibit the mitogenicityof lipid A from E. coli (11), had virtually no effect on themitogenicity of PW-LPS from B. gingivalis for C3H/HeNspleen cells (Fig. 7). In addition, we were unable to demon-strate KDO and heptose, major components of lipid A in E.coli LPS, and no lipid A from B. gingivalis PW-LPS withmild acid treatment was obtained (unpublished data). Theseresults indicate that PW-LPS from B. gingivalis may containa mitogenically active substance different from lipid A of theLPS of members of the Enterobacteriaceae.Both LPS preparations from B. gingivalis were found to

be mitogenic for BALB/c and BALB/c (nulnu) spleen cellsbut not for BALB/c thymocytes (Fig. 2 and 3). These resultsindicate that LPS from B. gingivalis is a B cell mitogen. Inaddition, these LPS preparations induced in vitro and in vivodirect PFC responses to SRBC (Table 5). It has beenreported that established lesions of chronic inflammatoryperiodontal disease in humans are primarily infiltrated withB cell lesions (17, 28). Therefore, polyclonal activation of Bcells induced by B. gingivalis LPS may participate in thedevelopment of periodontal disease.The chemical compositions of LPS preparations from B.

gingivalis 381 differed considerably with different extractionmethods. The main sugars of PW-LPS from this strain wereglucosamine, glucose, rhamnose, galactose, galactosamine,and mannose. It should be noted again that this LPS lackedheptose and KDO (Table 1). These findings are consistentwith those reported by Mansheim et al. (18) and Nair et al.(24). On the other hand, the most prominent sugar ofBW-LPS from B. gingivalis 381 was glucose, and trace

;m';7;n_

VOL. 47, 1985

on March 2, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

644 KOGA ET AL.

40

0 5 10

(C3H /HeN)B. B.gingivalis

PW-LPS

( C3H/HeJ)E B.gingivalis

PW-LPS

0 5 10

Stimulant (jig /well)

FIG. 5. Mitogenic responses of C3H/HeN and C3H/HeJ mouse spleen cells to ConA and LPSs from B. gingivalis 381 and E. coli K235.Various numbers of spleen cells from C3H/HeN (A, B, and C) and C3H/HeJ (D, E, and F) mice were cultured for 48 h with the indicated dosesof ConA (A and D), PW-LPS from B. gingivalis 381 (B and E), and PW-LPS from E. coli K235 (C and F). Symbols: 0, 7 x 105 cells per well;0, 5 x 105 cells per well; A, 2 x 105 cells per well; A, 8 x 104 cells per well.

0-c

L-

_I0 20 40 0 20 40

0 5 10PW-LPS(ug/well)

FIG. 6. Mitogenicity of enzyme-treated PW-LPS from B. ging-ivalis 381 for C3H/HeJ mouse spleen cells. Spleen cells (7 x 105)were cultured for 48 h with or without the indicated doses ofenzyme-treated PW-LPS from B. gingiv'alis 381 in 0.2 ml of RPMI1640 medium. The results are expressed as arithmetric mean of thestimulation index (counts per minute for tested culture/counts perminute for culture with medium alone, 3,361 + 307 cpm). Theenzyme control possessed no significant mitogenic activity. Sym-bols: 0, no enzymes; 0, trypsin: A, pronase E.

Polymyxin B (U /well)FIG. 7. Effects of polymyxin B on mitogenic responses of

C3H/HeN (A) and C3H/HeJ (B) mouse spleen cells. Spleen cells (7x 1(O) were cultured for 48 h with 5 jig of ConA (@), PW-LPS fromB. gingi alis 381 (A), or PW-LPS from E. coli K235 (-) in 0.2 ml ofRPMI 1640 medium containing various amounts of polymyxin B.C3H/HeN and C3H/HeJ cultures to which no polymyxin B had beenadded (100% response) responded as follows: ConA, 72,832 + 8,217cpm and 93,014 ± 5.104 cpm, respectively; PW-LPS from B.gingivalis 381, 37,724 ± 2,055 cpm and 22,072 + 2,309 cpm,respectively: PW-LPS from E. coli K235. 44,536 ± 8,893 cpm and3,155 ± 80 cpm. respectively.

E

0-1-zX

40

20

FE coliPW-LPS

IRJ &. -1

x

cl-

C0._

n

INFECT. IMMUN.

2(

on March 2, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

IMMUNOBIOLOGY OF B. GINGIVALIS LPS 645

TABLE 5. Effect of LPS on in vitro immune responses to SRBC'No. of PFCs per 106 spleen cells

Stimulant (mean ± SE):(10 jig/well)

Without SRBC With SRBC

None 11 ± 4 152 ± 31

B. gingivalis 381PW-LPS 168 ± 9 823 ± 53BW-LPS 52 ± 12 644 ± 44

E. coli K235PW-LPS 152 ± 22 552 ± 48BW-LPS 59 ± 5 179 ± 14

a After spleen cells (5 x 106) from BALB/c mice were precultured for 6 hwith or without SRBC (2 x 106) in 0.6 ml of RPMI 1640 medium supplementedwith 20% FCS, LPS (10,Lg) was added. After 4 days of incubation, PFCresponses were determined. Data are expressed as the mean + SE of fourdifferent cultures.

amounts of rhamnose, glucosamine, and two unknown sug-ars were present (Table 1). This type of sugar composition issimilar to that of LPS obtained by the phenol-water proce-dure from Bacteroides oralis (2).Mansheim et al. (18) reported that LPS from B. gingivalis

(B. asaccharolyticus) contains no detectable ,3-hydroxy fattyacids. In contrast, Nair et al. (24) have found an abundanceof hydroxy fatty acids. We could not detect hydroxy fattyacids in our LPS preparations (Table 2). The cause of thisdiscrepancy is unknown, but it may be ascribed in part todifferences in the extraction method and the gas-liquidchromatographic procedures for fatty acid analyses. How-ever, the main fatty acid of LPS from B. gingivalis reportedby both investigators was palmitic acid, which was inagreement with our findings (Table 2).

McIntire et al. (19) isolated LPS from E. coli K235 by thephenol-water method and demonstrated the presence ofglucosamine, galactosamine, glucose, galactose, rhamnose,

tn 600

400

to(l

200

a-L

(nI

F-SRBC

tBW-LPS

i-None0 3 6

Of

24Preculture Time (h)

FIG. 8. Effect of preculture time with SRBC on in vitro adjuvantactivity of BW-LPS from B. gingiv'alis 381. Spleen cells (5 x 106)from BALB/c mice were precultured for 0 to 24 h with SRBC (2 x106) in 0.6 ml of RPMI 1640 medium supplemented with 20% FCS,and then BW-LPS (10 ,ug) from B. gingivalis was added. After 4days, PFC responses were determined. Data are expressed as themean + SE of four different cultures.

E

3-

-aCK

pI

1

, -i i -

0 1 A

* .1:32 1:16 1:8 1:4

Supernatant Dilution ofMacrophage Culture

FIG. 9. IL-1 activity in various dilutions of culture supernatantsof mouse peritoneal macrophages that had been stimulated by LPSs(100 ,ug/ml) from B. gingivalis 381 and E. coli K235. BALB/c mousethymocytes (7 x 105 cells; 0.1 ml) were cultured for 48 h in theabsence ----) or presence (-) of ConA (0.2 ,ug) with 0.1 ml ofvarious dilutions of supernatants of macrophage cultures stimulatedwith PW-LPS from B. gingivalis (0) or PW-LPS from E. coli (A).Data are expressed as the mean ± SE of triplicate cultures.

heptose, and KDO as sugar components of this preparationby paper chromatography and colorimetric methods. Incontrast to these results, we have shown the presence offucose and mannose, but were unable to detect galactos-amine by gas-liquid chromatography and amino acid analy-sis. The major fatty acid components of the PW-LPS prep-aration from E. coli K235 were found to be ,B-hydroxymyristic (18% of the total dry weight of LPS), palmitic (3%),myristic (2%), and lauric (1%) acids (Table 2). These resultsare generally consistent with those reported by McIntire etal. (19), except for myristic acid. They failed to detectmyristic acid in their E. coli LPS preparation.

IL-1 is known to have many activities, such as enhance-ment of immune responses, stimulation of thymocyte prolif-

TABLE 6. Induction of IL-1 secretion by LPS in BALB/c miceStimulant added to Concn (Lg/ml) [3HJTdR uptake"

macrophages (cpm)

None 129 ± 20

B. gingivalis 381PW-LPS 100 1,093 ± 55

10 368 97BW-LPS 100 106 ± 22

10 144 62

E. coli K235PW-LPS 100 2,058 ± 29

10 1.067 ± 57BW-LPS 100 207 ± 25

10 296 ± 87a Thymocytes from BALB/c mice were suspended to a density of 7 x 106

cells per ml in RPMI 1640 medium. The cell suspension (0.1 ml) and an equalvolume of a 1:4 dilution of supernatants of macrophage cultures that had beenstimulated with LPS were seeded into a 96-well microculture plate. Cultiva-tion was carried in the presence of ConA (0.2 jig). Data are expressed as themean t SE of three different cultures.

VOL. 47, 1985

2

6-

on March 2, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

646 KOGA ET AL.

TABLE 7. Combined effects of ConA and LPS in mitogenicresponses of BALB/c mouse thymocytes

Concn [3HJTdR uptake' (cpm):Stimulant (L/elStimulant( ell) Without ConA With ConA

None 198 ± 34 NDbConA 1 8,215 ± 238 ND

0.5 1,103 ± 261 ND0.2 250 + 33 ND

B. gingivalis 381PW-LPS 10 292 ± 59 216 ± 15

5 338 132 177 ± 231 288 39 213 ± 490.1 304 ± 19 256 ± 10

E. coli K235PW-LPS 10 291 ± 39 323 ± 121

5 306 ± 100 175 ± 251 288 ± 33 228 ± 330.1 356 ± 101 213 ± 19

a BALB/c thymocytes (7 x 105 cells) were incubated in a total volume of 0.2ml of RPMI 1640 medium with LPS (0 to 10 pug) in the absence or presence ofConA (0.2 ktg). Data are expressed as the mean ± SE of three differentcultures.

b ND, Not determined.

eration, activation of B cells, stimulation of proteinase andprostaglandin production by connective tissue, and stimula-tion of bone resorption (3, 6). It was found in this study thatPW-LPS from B. gingivalis stimulated IL-1 production bymacrophages to activate BALB/c thymocytes (Fig. 9 andTable 6). These findings suggest that otle of mediators ofalveolar bone loss in human periodontitis could be IL-1stimulated by B. gingivalis LPS.LPS from B. gingivalis has been reported to display

significantly less biological potency than that from Salmo-nella species and E. coli in terms of Limrlulis lysate gelation,chicken embryo lethality, dermal Shwartzman reaction andpyrogenicity in rabbits, and mitogenicity for mouse spleencells (18, 24). In the present study, we isolated LPSs from B.gingivalis and E. coli under the same conditions and com-pared their biological properties with each other. It wasrevealed that biological activities of LPS from B. gingivaliswere comparable to those of E. coli LPS in terms ofactivation of clotting enzymes of Limulus amebocyte lysate,mitogenicity, polyclonal B cell activation, adjuvanticity, andstimulation of IL-1 production. In particular, LPS from B.gingivalis was much higher in adjuvant activity than LPSfrom E. coli (Table 5).

In conclusion, it is apparent now that LPS preparationsobtained with different extraction methods from B. ging-ivalis, a suspected periodontopathic bacterium, possessmarked immunobiological potencies on lymphocytes andmacrophages.

ACKNOWLEDGMENTSWe thank E. Jirillo, University of Bari, Bari, Italy, for his critical

review of the manuscript.

LITERATURE CITED

1. Cunningham, A. J., and A. Szenberg. 1968. Further improve-ments in the plaque technique for detecting single antibody-form-ing cells. Immunology 14:599-600.

2. Dahlen, G., H. Nygren, and H.-A. Hansson. 1978. Immunoelec-tron microscopic localization of lipopolysaccharides in the cellwall of Bacteroides oralis and Fusobacterium nucleatuim. In-fect. Immun. 19:265-271.

3. Dinarello, C. A. 1984. Interleukin-1. Rev. Infect. Dis. 6:51-95.

4. Flesher, A. R., and R. A. Insel. 1978. Characterization oflipopolysaccharide of Haemophililuis inflienzae. J. Infect. Dis.138:719-730.

5. Galanos, C., 0. Luderitz, E. T. Rietschel, and 0. Westphal.1977. Newer aspects of the chemistry and biology of bacteriallipolysaccharides, with special reference to their lipid A com-ponent, p. 239-335. In T. W. Goodwin (ed.), Internationalreview of biochemistry, vol. 14. University Park Press, Balti-more.

6. Gowen, M., D. D. Wood, E. J. Ihrie, M. K. B. McGuire, andR. G. G. Russel. 1983. An interleukin 1 like factor stimulatesbone resorption in vitro. Nature (London) 306:378-380.

7. Heijl, L., J. Wehnstrom, J. Lindhe, and S. S. Socransky. 1980.Periodontal disease in gnotobiotic rats. J. Periodontal Res.15:405-419.

8. Irving, J. T., M. G. Newman, S. S. Socransky, and J. D. Heeley.1975. Histological changes in experimental periodontal diseasein rats monoinfected with a gram-negative organism. Arch. OralBiol. 20:219-220.

9. Itaya, K., and M. Ui. 1965. Colorimetric determination of freefatty acids in biological fluids. J. Lipid Res. 6:16-20.

10. Iwanaga, S., T. Morita, T. Harada, S. Nakamura, M. Niwa, K.Takada, T. Kimura, and S. Sakakibara. 1978. Chromogenicsubstrates for horseshoe crab clotting enzyme. Its applicationfor the assay of bacterial endotoxins. Haemostasis 7:183-188.

11. Jacobs, D. M., and D. C. Morrison. 1977. Inhibition of themitogenic response to lipopolysaccharide (LPS) in mouse spleencells by polymyxin B. J. Immunol. 118:21-27.

12. Jirillo, E., S. M. Michalek, and J. R. McGhee. 1984. Lipopoly-saccharide regulation of the immune response: LPS effects onlymphoreticular cells and mucosal immune responses. EOS Riv.Immunol. Immunofarm. 4:21-31.

13. Joiner, K. A., K. P. W. J. McAdam, and D. L. Kasper. 1982.Lipopolysaccharides from Bacteroides fragilis are mitogenic forspleen cells from endotoxin responder and nonresponder mice.Infect. Immun. 36:1139-1145.

14. Karkhanis, Y. 0., J. Y. Zeltner, J. J. Jackson, and D. J. Carlo.1978. A new and improved microassay to determine 2-keto-3-deoxyoctonate in lipopolysaccharide of gram-negative bacteria.Anal. Biochem. 85:595-601.

15. Kasper, D. L., A. A. Lindberg, A. Weintraub, A. B. Onderdonk,and J. Lonngren. 1984. Capsular polysaccharides and lipopoly-saccharides from two strains of Bacteroides fragilis. Rev.Infect. Dis. 6:S25-S29.

16. Laemmli, U. K. 1970. Cleavage of structural proteins during theassembly of the head of bacteriophage T4. Nature (London)227:680-685.

17. Mackler, B. F., T. C. Waldrop, P. Schur, P. B. Robertson, andB. M. Levy. 1978. IgG subclasses in human periodontal disease.I. Distribution and incidence of IgG subclass bearing lympho-cytes and plasma cells. J. Periodontal Res. 13:109-119.

18. Mansheim, B. J., A. B. Onderdonk, and D. L. Kasper. 1978.Immunochemical and biologic studies of the lipopolysaccharideof Bacteroides melaninogenicius subspecies asaccharolyticus. J.Immunol. 120:72-78.

19. McIntire, F. C., H. W. Sievert, G. H. Barlow, R. A. Finley, andA. Y. Lee. 1967. Chemical, physical, and biological properties ofa lipopolysaccharide from Escherichia coli K-235. Biochemistry6:2363-2372.

20. Morrison, D. C. 1983. Bacterial endotoxins and pathogenesis.Rev. Infect. Dis. 5:S733-S747.

21. Morrison, D. C., S. J. Betz, and D. M. Jacobs. 1976. Isolation ofa lipid A bound polypeptide responsible for "LPS-initiated"mitogenesis of C3H/HeJ spleen cells. J. Exp. Med. 144:840-846.

22. Morrison, D. C., and L. Leive. 1975. Fractions of lipopolysac-charide from Escherichia coli O111:B4 prepared by two extrac-tion procedures. J. Biol. Chem. 250:2911-2919.

23. Morrison, D. C., and R. J. Ulevitch. 1978. The effects ofbacterial endotoxins on host mediation systems. Am. J. Pathol.93:527-617.

24. Nair, B. C., W. R. Mayberry, R. Dziak, P. P. Chen, M. J.Levine, and E. Hausmann. 1983. Biological effects of a purifiedlipopolysaccharide from Bacteroides gingi'alis. J. Periodontal

INFECT. IMMUN.

on March 2, 2020 by guest

http://iai.asm.org/

Dow

nloaded from

IMMUNOBIOLOGY OF B. GINGIVALIS LPS 647

Res. 18:4(-49.25. Niedermeier, W. 1971. Gas chromatography of neutral and

amino sugars in glycoproteins. Anal. Biochem. 40:465-475.26. Okahashi, N., T. Koga, H. Akada, and S. Hamada. 1983.

Purification and immunochemical characterization of Strepto-coccuis sangiis serotype I carbohydrate antigen. Infect. Immun.39:552-558.

27. Sawardeker, J. S., J. H. Sloneker, and A. Jeanes. 1965. Quanti-tative determination of monosaccharides as their alditol acetatesby gas liquid chromatography. Anal. Chem. 37:1602-1604.

28. Seymour, G. J., and J. S. Greenspan. 1979. The phenotypiccharacterization of lymphocyte subpopulations in establishedhuman periodontal disease. J. Periodontal Res. 14:39-46.

29. Slots, J. 1977. The predominant cultivable microflora of ad-vanced periodontitis. Scand. J. Dent. Res. 85:114-121.

30. Slots, J. 1982. Importance of black-pigmented Bacteroides inhuman periodontal disease. p. 27-45. In R. J. Genco and S. E.Mergenhagen (ed.), Host-parasite interaction in periodontaldiseases. American Society for Microbiology. Washington, D.C.

31. Sultzer, B. M. 1968. Genetic control of leukocyte responses toendotoxin. Nature (London) 219:1253-1254.

32. Sultzer, B. M., and G. W. Goodman. 1976. Endotoxin protein: aB-cell mitogen and polyclonal activator of C3H/HeJ lympho-cytes. J. Exp. Med. 144:821-827.

33. Vacheron, F., M. Guenounou, and C. Nauciel. 1983. Induction ofinterleukin 1 secretion by adjuvant-active peptidoglycans. In-fect. Immun. 42:1049-1054.

34. Wannemuehler, M. J., S. M. Michalek, E. Jirillo, S. I. Willi-amson, H. Hirasawa, and J. R. McGhee. 1984. LPS regulation ofthe immune response: Bacteroides endotoxin induces mitoge-nic, polyclonal, and antibody responses in classical LPS respon-sive but not C3H/HeJ mice. J. Immunol. 133:299-305.

35. Westphal, O., and K. Jann. 1965. Bacterial lipopolysaccharides.Extraction with phenol-water and further applications of theprocedure. Methods Carbohydr. Che,m. 5:83-91.

36. Wright, B. G., and P. A. Rebers. 1972. Procedure for determin-ing heptose and hexose in lipopolysaccharides-modification ofthe cysteine-sulfuric acid method. Anal. Biochem. 49:307-319.

VOL. 47. 1985

on March 2, 2020 by guest

http://iai.asm.org/

Dow

nloaded from