Vol. 172, No. 10JOURNAL OF BACTERIOLOGY, Oct. 1990, p. 5714-57230021-9193/90/105714-10$02.00/0Copyright C) 1990, American Society for Microbiology

Regulation of Fructose Metabolism and Polymer Synthesis byFusobacterium nucleatum ATCC 10953

STANLEY A. ROBRISH* AND JOHN THOMPSON

Laboratory of Microbial Ecology, National Institute ofDental Research, Bethesda, Maryland 20892

Received 5 February 1990/Accepted 11 July 1990

Energy for the anaerobic growth of Fusobacterium nucleatum ATCC 10953 can be derived from thefermentation of sugar (fructose) or amino acid (glutamate). During growth on fructose, the cells formed largeintracellular granules which after extraction yielded glucose by either acid or enzymatic hydrolysis. Theendogenous polymer was subsequently metabolized, and after overnight incubation of the cells in buffer, theglucan granules were no longer detectable by electron microscopy. Anaerobically, washed cells grownpreviously on fructose fermented this sugar to a mixture of lactic, acetic, and butyric acids, and littleintracellular glucan was formed. Aerobically, the cells slowly metabolized fructose to acetate. Provision ofglutamic acid as an additional energy (ATP) source elicited rapid synthesis of polymer by glycolyzing cells.Intracellular granules were not present in glutamate-grown cells, and under anaerobic conditions, the restingcells failed to metabolize [14C] fructose. However, the addition of glutamic acid to the suspension resulted inthe rapid accumulation of sugar by the cells. Approximately 15% of the '4C-labeled material was extractablewith boiling water, and by 31P nuclear magnetic resonance spectroscopy, this phosphorylated derivative wasidentified as [14C]fructose-l-phosphate. The nonextractable material represented [14C]glucan polymer. Fruc-tose-l-phosphate kinase activity in fructose-grown cells was fivefold greater than that in glutamate-grown cells.We suggest that the activity of fructose-l-phosphate kinase and the availability of ATP regulate the flow offructose into either the glycolytic or polymer-synthesizing pathway in F. nucleatum.

Fusobacterium nucleatum is a gram-negative, spindle-shaped organism (5, 15, 17, 43) which by association (11, 16,27, 37) is believed to play a role in the etiology of periodon-titis and gingivitis in humans (21, 25, 26, 38). The bacteriumrelies primarily on the catabolism of amino acids (e.g.,glutamate, histidine, lysine [2-4, 6, 10, 19, 23]) for energyproduction, and the end products of these fermentations areusually acetic, propionic, and butyric acids (4, 6, 17-19, 23,44). Penetration of periodontal tissue by the cytotoxic or-ganic acids (36), together with the amino acid-rich environ-ment of the gingival crevice (24, 35), facilitate the invasionand provide an ecological niche conducive to establishmentof the oral pathogen (11, 16, 27, 38).Bergey's Manual of Systematic Bacteriology (18) and

other sources (15, 17, 23) describe the fusobacteria asweakly saccharolytic. The early studies of Jackins andBarker (19), Loesche and Gibbons (23), and Coles (7), aswell as more recent reports (9, 31), attest to the limitedglycolytic activity of F. nucleatum. It is perhaps because ofthese observations that regulation of sugar transport andmetabolism by F. nucleatum has received little attention.However, in the past 3 years (31-33), we have obtainedpartial explanations for the weakly saccharolytic capacity ofthis organism by the discoveries (i) that transport andphosphorylation of sugars (e.g., glucose and galactose) de-pend on ATP generated via amino acid metabolism, (ii) thatintracellular sugar phosphates are rapidly converted to poly-mer, and (iii) that concomitant fermentation of amino acidscan prevent the degradation of endogenous polymer. Itseemed reasonable to assume that the transport and dissim-ilation of fructose would likewise depend on provision of afermentable amino acid, but this proves not to be the case.

This communication represents the first detailed descrip-

* Corresponding author.

tion of fructose transport and metabolism by F. nucleatum.In addition, our observations provide further insight into theunusual interrelationships which exist between amino acidand sugar fermentations in this organism (9, 31-33). Finally,we show how the cofermentation of an amino acid (gluta-mate) can regulate the entry of fructose into either thecatabolic or polymer-synthesizing pathway of the cell.

MATERIALS AND METHODSOrganism and culture maintenance. F. nucleatum ATCC

10953 was obtained from the American Type Culture Col-lection (Rockville, Md.). The organism was maintained bysemimonthly transfer in modified thioglycolate medium(Fisher Scientific Co., Springfield, N.J.) supplemented with20% (wt/vol) horse meat infusion and solid CaCO3. Cultureswere grown at 37°C under anaerobic conditions (GasPak;BBL Microbiology Systems, Cockeysville, Md.) and werelater stored at 4°C.Growth of cells. F. nucleatum was grown in a medium of

the following composition (grams per liter): Myosate (BBL),6; Neopeptone (Difco Laboratories, Detroit, Mich.), 20;NaCl, 2; Na2HPO4, 0.4; and Na2CO3, 0.8. A salts-vitaminmixture was added separately from a concentrated stocksolution (stored at 4°C) to give the following final concentra-tions (milligrams per liter): FeSO4 * 7H20, 4; MnCl2 4H20,1.9; MgSO4 7H20, 490; dithiothreitol, 100; biotin, 0.015;thiamine, 2.5; riboflavin, 2.5; pyridoxine, 5; pantothenicacid, 2.5; nicotinic acid, 5; folic acid, 0.025; andp-aminoben-zoic acid, 0.5. Fructose or ammonium glutamate was pro-vided as an energy source at a final concentration of S g/liter.

Preparation of cells. A 10-ml culture of F. nucleatumgrown previously overnight at 37°C in an anaerobic jar wastransferred to 200 ml of medium supplemented with salts-vitamin mixture and an appropriate energy source (gluta-mate or fructose). The culture was grown under anaerobicconditions until the late log or early stationary phase (ap-

5714

on Novem

ber 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

FRUCTOSE METABOLISM BY FUSOBACTERIUM NUCLEATUM 5715

proximately 22 h) and then was quickly transferred to acapped 250-ml polycarbonate bottle. The headspace wasflushed with an anaerobic gas mixture (5% C02, 5% H2, 90%N2), and the cells were collected by centrifugation at 5,000 xg for 20 min at 4°C. The supernatant fluid was discarded, andthe cell pellet was suspended as quickly as possible in 35 mlof anaerobically prepared buffer (referred to as Coles buffer[7]) containing 50 mM potassium phosphate (pH 7), 0.7 mMMgCl2, 0.1 mM AMP, and 0.1 mM NAD+. After beingflushed with anaerobic gas mixture, the tube was capped andthe cells were collected by centrifugation. The cell pellet wassuspended in 5 ml of Coles buffer to yield a homogeneousthick cell suspension, which was then maintained underanaerobic conditions at 0°C until required.

Preparation of cell extracts. Cells were harvested from 1liter of stationary-phase culture, washed, and suspended to 5ml with 50 mM potassium phosphate (pH 7) buffer containing0.7 mM MgCl2. The cells were disrupted by sonic oscillationtreatment (at 0°C, under anaerobic gas) with the microtip ofa Branson sonifier operating at 70% of maximum power.After a total of 3 min of sonication (with intermittent cooling),the suspension was centrifuged at 27,000 x g for 30 min at4°C. The supernatant fluid was removed and used for enzy-matic analyses.Assays for sugar phosphorylation (kinase activities) were

performed at room temperature and contained the followingin a final volume of 200 ,ul: 100 mM potassium phosphatebuffer (pH 7); 5 mM ATP, phosphoenolpyruvate (PEP), oracetylphosphate; 5 mM MgCl2; 2.5 mM [14C]sugar (specificactivity, 0.2 ,Ci/pRmol); and cell extract (1 to 2 mg ofprotein). At 0-, 5-, 10-, 20-, and 30-min intervals, 25-pIsamples were transferred to Whatman DE-81 (2.5-cm-diam-eter) filter disks. The disks were placed first in water toremove free sugar, then in acetone and air dried. Radiola-beled sugar phosphate retained on the filters was determinedby liquid scintillation spectrometry.Sugar transport studies. The procedure used to monitor

[14C]sugar uptake by resting (washed) cells of F. nucleatumhas been described previously (31). In brief, 4.8-ml volumesof Coles buffer containing appropriate additions (e.g., so-dium glutamate, 8 mM; [14C]fructose, 0.5 mM; specificactivity, 0.2 ,uCi/,pmol) were transferred to screw-cap tubes(16 by 125 mm) filled with anaerobic gas and capped withbutyl rubber septa (Bellco Glass, Inc., Vineland, N.J.).Transport experiments (at 37°C) were initiated by the addi-tion of 0.2 ml of anaerobic cell suspension (equivalent to 1.5to 2.5 mg [dry weight] of cells) to the transport assay. Atintervals, 0.5 ml of suspension was withdrawn with a tuber-culin syringe (previously flushed with anaerobic gas), andcells were collected by vacuum filtration through membranefilters (diameter, 25 mm; pore size, 0.45 pum; type HA;Millipore Corp., Bedford, Mass.). Cells retained on the filterwere rinsed with 4 ml of Coles buffer and dried, andcell-associated radioactivity was determined by liquid scin-tillation spectrometry (31).

Fructose utilization by resting cells. Resting (washed) cellswere suspended (to the equivalent of 12 to 15 mg of totalprotein) in 5 ml of Coles buffer containing 2 mM fructoseand, when required, 20 mM sodium glutamate. At intervals,0.5-ml samples were removed from each system with ananaerobic syringe and cells were removed by filtrationthrough Millex-GS (0.22-pum pore size; Millipore Corp.) filterunits. Filtrates were collected, and fructose was determinedby the Glucose Analysis Kit 115 (Sigma Chemical Co., St.Louis, Mo.) to which was added phosphoglucoseisomerase(EC 5.3.1.9). In some experiments, detection or confirma-

tion of the presence of fructose was provided by the resor-cinol procedure of Roe (34).

Extraction and identification of glycolytic intermediates.Washed cells of F. nucleatum grown previously on eitherfructose or glutamate were suspended (to the equivalent of20 mg of total protein) in 5 ml of Coles buffer containing 20mM sodium glutamate. After 10 min of incubation at 37°C,0.5 mM ['4C]fructose (specific activity, 2 ,uCi/,umol) wasadded to the suspensions and incubation was continued for afurther 10 min. Cells were collected by membrane filtration(47-mm-diameter membranes, 0.45-p.m pore size) and rinsedbriefly with distilled water at 0°C, and filters plus adheringcells were transferred to 8 ml of boiling water for 6 min. Thesuspensions were clarified by centrifugation, and the super-natant fluids were removed and lyophilized. The residueswere reconstituted with 0.5 ml of water, and the solutionswere passed through PD-10 gel filtration columns. Fractionscontaining low-molecular-weight compounds were pooled,frozen, and lyophilized. Each residue was reconstituted with50 p.l of water, and 10 p.1 of solution (containing approxi-mately 60,000 cpm) was applied to thin-layer sheets ofpolyethyleneimine (PEI)-cellulose. Radiolabeled glycolyticintermediates were then separated and identified by autora-diography as described previously (41, 42).

Extraction and'composition of intracellular granules. Fruc-tose-grown cells (37 mg [dry weight]) were suspended in 5 mlof Coles buffer containing 25 mM sodium glutamate and 1mM [14C]fructose (specific activity, 0.5 p.CiIp.mol). After 20min of incubation at 37°C (to allow for polymer synthesis),the cells were collected by membrane filtration and thenextracted in 8 ml of boiling water for 10 min. The suspensionwas centrifuged (10,000 x g for 15 min), supernatant fluidwas removed, and the cell pellet was taken up in 2 ml ofwater. Radiolabeled cell-associated polymer was extractedby a modification of the Somogyi procedure (39), in whichthe 2-ml cell suspension was mixed with 4 ml of 30% (wt/vol)KOH and heated in a sealed tube for 3 h at 100°C. Aftercooling, the solution was clarified by centrifugation andsupernatant fluid was removed and mixed with 2 volumes of95% (vol/vol) ethanol and left to stand overnight at 4°C. Thewhite precipitate collected by centrifugation was washedtwice by resuspension and centrifugation from 5 ml of 95%(vol/vol) ethanol. The residue of 14C-labeled polymer wasdried in vacuo and subsequently dissolved in 500 ,ul of water(total of -7 x 105 cpm). In the acid hydrolysis procedure,200 p.1 of polymer solution was diluted with 200 ,u1 of 2 N HCiand the mixture was heated in a sealed tube for 5 h at 100°C.The hydrolysate was lyophilized. In the enzymatic hydroly-sis procedure, 200 pul of polymer solution was mixed with 500pul of 50 mM acetate buffer (pH 4.6) containing 20 mg(approximately 48 U) of amyloglucosidase and the solutionwas left at room temperature for 4 h before lyophilization.The two residues obtained by acid and enzymatic hydrolysiswere each reconstituted in 1 ml of water. The samples wereapplied to and eluted from PD-10 columns with distilledwater, and fractions containing low-molecular-weight com-pounds were collected, pooled, and lyophilized. Each resi-due was dissolved in 300 p.l of water, and the two sampleswere desalted by passage through small, mixed-bed ion-exchange columns (0.5 ml each of Dowex-1-formate andDowex 50 H+ form resins). The water eluents (total, 5 ml)were lyophilized, and each residue (approximately 3 x 105cpm) was redissolved in 100 p.1 of water. For chromato-graphic analysis, 3 pu1 of each hydrolysate (10,000 cpm) wasapplied to 3MM chromatography paper (Whatman, Inc.,Clifton, N.J.) together with 7 p.1 (approximately 28,000 cpm)

VOL. 172, 1990

on Novem

ber 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

5716 ROBRISH AND THOMPSON

of a mixture of [14C]glucose and ["4C]fructose (5 mM eachsugar; specific activity, 0.2 ,uCi/,mol) as standards. Thesolvent used for descending chromatography (40 h) con-tained butyl acetate-ethanol-pyridine-H20 (8:2:2:1, vol/vol)and sugars were visualized by autoradiography.

Extraction of fructose derivative. Cells of F. nucleatum10953 were grown to the stationary phase in 2.4 liters ofmedium supplemented with 0.5% (wt/vol) ammonium gluta-mate as the energy source. The culture was divided, and theglutamate-grown cells were harvested and washed (anaero-bically) with 50 mM potassium phosphate (pH 7) buffercontaining 1 mM MgCl2. Each pellet was suspended in 20 mlof buffer, and sodium glutamate was added to each anaerobicsuspension to a final concentration of 200 mM. The suspen-sions were incubated for 10 min at 37°C, and then 1 ml ofanaerobic buffer was added to one suspension (control) and1 ml of 100 mM fructose solution was added to the other.Twenty minutes later, the suspensions were chilled rapidlyin ice-water and the cells were collected by centrifugation.The pellets were washed with 20 ml of water (0°C), and thenthe cells were extracted with 10 ml of boiling water for 10min. After clarification by centrifugation (10,000 x g for 30min), the two supernatant solutions were removed, frozen,and lyophilized. The residues were dissolved in 1 ml of 1 Nacetic acid, and precipitated (proteinaceous) material wasremoved by centrifugation. The supernatant solutions werepassed through PD-10 columns, and fractions containinglow-molecular-weight compounds were pooled, frozen, andlyophilized.31P-NMR spectroscopy. Prior to 31P-nuclear magnetic res-

onance (NMR) analysis, residues were reconstituted with0.4 ml of D20 (>99.8 atom %; Merck Sharpe & Dohme,Montreal, Quebec, Canada) and 0.1 ml of 1 M HEPES(N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid) buff-er (pH 8) containing 25 mM EDTA was added to eachsolution. 31P-NMR spectra were recorded on a JEOL GSX-500 spectrometer by using a (90° pulse-free induction decayacquisition-pulse delay)n sequence, with continuous broad-band proton decoupling. Spectral acquisition parametersincluded 11-pus 900 pulse, 15-kHz sweep width, 32,768 datapoints, 1.1-s data acquisition, 6.0-s pulse delay. Seventy-twofree induction decay signals were averaged for each spec-trum. Before Fourier transformation, the free inductiondecay signal was zero filled and then exponentially broad-ened to result in an additional 1.5-Hz line broadening in thefrequency domain spectrum.

Electron microscopy. Cells of F. nucleatum were collectedby centrifugation from 200 ml of medium containing eitherfructose or glutamate as the energy source. The cells werewashed twice by suspension and centrifugation from phos-phate-buffered saline and then were fixed for 4 h in 0.1 Mcacodylate buffer (pH 7.4) containing 2% (vol/vol) each offormaldehyde and glutaraldehyde. Fixed cells were rinsed incacodylate buffer and then were postfixed in cacodylatebuffer containing 2% osmium tetroxide and 1.5% (wt/vol)potassium ferrocyanide (20). The samples were dehydratedthrough a graded series of increasing concentrations ofethanol. The preparations were embedded in Spurr resin(40), and thin sections were cut with a diamond knife. Afterbeing mounted on copper grids, the sections were stainedlightly with Reynolds lead citrate (30) and examined underan electron microscope (110-CX; JEOL U.S.A. Inc., Pea-body, Mass.).

Analytical procedures. Cell proteins were determined by acombination of the NaOH solubilization and Lowry proce-dures described by Herbert et al. (14). L-Glutamic acid was

assayed with a glutamate dehydrogenase kit (BoehringerMannheim Biochemicals, Indianapolis, Ind.). Fructose-i-phosphate (F1P) kinase activity was determined by assay IIdescribed by Gottschalk and von Hugo (13):

FlP + ATP Mg2+, FlP kinase (in extract} FDP + ADP (1)ADP + PEP pyruvate kinase ATP + pyruvate (2)

Pyruvate + NADH + H+ lactate dehydrogenase lactate + NAD+ (3)

where FDP is fructose-1,6-diphosphate. The spectrophoto-metric determination of FlP kinase activity by these sequen-tial reactions relies on the formation of ADP in reaction 1.Cell extracts of F. nucleatum exhibit ATPase activity, andADP is also formed via hydrolysis of ATP. The contributionof and correction for ATPase activity were determined byusing a complete assay lacking FlP.

Analysis of fermentation products. Fermentation productswere quantitatively identified on an Aminex HPX-87H (300by 7.8 mm) organic acid separation column (Bio-Rad Labo-ratories, Richmond, Calif.). For each analysis, 5 pAl of astandard solution (containing 20 ,ug each of unlabeled glu-cose, fructose, and lactic, formic, acetic, and butyric acids)was mixed with 100 RI of appropriate filtrate. Then 20 RI ofthe mixture was injected onto the column, and 0.01 N H2SO4was used as the carrier solvent. Standard compounds weredetected and elution profiles were recorded with an LKBUvicord S detector (LKB, Bromma, Sweden) fitted with a206-nm interference filter. Column effluents were collected(200-pul fractions), and radioactivity present in 100 ,ul of eachfraction was determined by liquid scintillation spectrometry.The 14C-labeled fermentation products were identified bycomparison (and coelution) with absorbance peaks of theunlabeled internal standards.

Reagents. Radiolabeled sugars were purchased from Du-pont, NEN Research Products (Boston, Mass.). a-Amylo-glucosidase [1,4-a-D-glucan glucohydrolase, EC 3.2.1.3] andalkaline phosphatase (calf intestinal, EC 3.1.3.1) were ob-tained from Boehringer Mannheim Biochemicals. Precoatedplastic-backed layers of polyethyleneimine cellulose (Baker-flex) were purchased from J. T. Baker Chemical Co. (Phil-lipsburg, N.J.). The PD-10 gel filtration columns were ob-tained from LKB-Pharmacia Inc. (Piscataway, N.J.).

RESULTS

Growth studies. Experiments presented in Fig. 1 demon-strate the capacity ofF. nucleatum to utilize either glutamate(Fig. 1A) or fructose (Fig. 1B) as the energy source forgrowth. In each case, the increase in cell mass was concom-itant with the utilization of the appropriate substrate. Thebackground growth found in the absence of either fructose orglutamic acid (Fig. 1B) can be attributed to low concentra-tions of fermentable amino acids in the complex growthmedium (e.g., glutamate, lysine, serine, and threonine [2, 3,10, 23]).Polymer formation. Electron microscopic examination of

fructose-grown cells revealed large, densely-packed intracel-lular granules (Fig. 2A). These inclusions were not formedby the organism during growth on glutamic acid (Fig. 2C).The latter finding, and the fact that the granules becameradiolabeled when the cells were grown on [14C]fructose,suggested that they were formed from either fructose per seor a fructose derivative. Radiolabeled granules were ex-tracted by NaOH treatment and ethanol precipitation, andthe purified material was acid hydrolyzed. Paper chromatog-raphy revealed [14C]glucose as the sole radiolabeled sugar in

J. BACTERIOL.

on Novem

ber 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

FRUCTOSE METABOLISM BY FUSOBACTERIUM NUCLEATUM 5717

-fi1000+.-

a)

cn 400

0

0

-C300

30

25I

E10 ff15 w

10 E0

20Eco

10o

. 5 EOL

0 2 4 6 8 10 12 "24Time (hours)

FIG. 1. Growth of F. nucleatum in medium supplemented withammonium glutamate (A) or fructose (B) as the energy source.

Growth of the organism in the absence of glutamate or fructose isindicated (L) in the lower part of panel B.

the hydrolysate (Fig. 3, lane 2). The same result was

obtained after incubation of the granules with amyloglucosi-dase (Fig. 3, lane 3). The intracellular granules (whichapparently consist of a glucan [glycogen?] polymer) were

subsequently metabolized, and after overnight incubation ofthe cells in buffer, the granules were no longer detectable(Fig. 2B). Enzymatic and gas chromatographic analysesshowed that acetic, butyric, and D-lactic acids are theproducts of polymer fermentation (unpublished data).

Fructose uptake by resting cells. Fructose-grown organ-

isms, when suspended in Coles buffer containing ["'C]-fructose, failed to accumulate significant levels of the radio-labeled sugar under aerobic or anaerobic conditions (Fig.4A). However, the addition of Na+-glutamate to the anaer-

obic suspension elicited rapid uptake of sugar by the cells.Of the total radioactivity accumulated by the organisms,approximately 10 to 15% was extractable with boiling waterand was found to consist primarily of radiolabeled glycolyticintermediates, free sugar, and fermentation products (Fig.SA). The remainder (-85 to 90%) was recovered as 14C-labeled polymer. Previously, we have shown that the fer-mentation of glutamate by F. nucleatum is inhibited byexposure of the cells to air (31). The presence of the aminoacid caused little stimulation of [14C]fructose uptake by cellsmaintained under aerobic conditions (Fig. 4A). The consti-

tutivity of the fructose transport system was demonstratedby the finding of glutamate-stimulated transport of["'Cifructose by glutamate-grown cells (Fig. 4B). As withfructose-grown cells, >85% of the accumulated radioactivitywas recoverable as "'C-labeled polymer. However, theglutamate-grown cells were found to contain unusually highlevels of a derivative which, by its migration position onPEI-thin-layer chromatography (TLC), was tentatively iden-tified as a hexose monophosphate (Fig. SB, arrow).

Product of fructose transport by F. nucleatum. The putativesugar phosphate (Fig. SB, arrow) was eluted from the PEI-TLC and purified by ion-exchange chromatography. Afterincubation of the sample with alkaline phosphatase, all theradioactivity was recovered and identified by paper chroma-tography as [14C]fructose (Fig. SC). The identity of thefructose phosphate(s) (fructose-6-phosphate [F6P], FlP, or amixture of the two) was established by comparative 31P-NMR spectroscopy. Extracts were prepared from cellsmetabolizing (i) glutamate alone (Fig. 6A) or (ii) glutamateplus fructose (Fig. 6B). In the presence of fructose, twoadditional phosphorylated compounds were formed (peaks 1and 2) with resonances centered at 1.85 and 1.90 ppm,respectively. Addition of authentic F6P to this extract gen-erated a new peak in the spectrum at 1.39 ppm (Fig. 6C).However, the addition of FlP resulted in the enhancement ofthe signal at 1.90 ppm (i.e., peak 2), and this spectralco-incidence established that the [14C]fructose derivativeformed by F. nucleatum (Fig. SB, arrow) was exclusively the1-phosphate isomer. The phosphorylated compound respon-sible for the resonance at 1.85 ppm (peak 1) has not yet beenidentified, but our TLC results suggest that it is not radiola-beled and probably is not directly derived from [14C]fruc-tose.

Fructose metabolism by resting cells. Previous studies (Fig.4A) showed that in the absence of glutamic acid, little["'Cifructose was accumulated by fructose-grown cells un-der either aerobic or anaerobic conditions. Since the cellsappeared unable to transport the sugar, it was thus perplex-ing to find that under the anaerobic conditions used for thetransport experiments, the cells (i) depleted fructose fromthe medium (10 nmol of sugar consumed per mg of proteinper min [Fig. 7A]), and (ii) metabolized [U-14C]fructose (Fig.8) to a mixture of labeled fermentation products comprisingD-lactate, butyrate, and acetate. The paradox was resolvedby experiments presented in Fig. 9, in which resting cellswere incubated with [14C]fructose and rapidly extracted withboiling water. Although the cell extract contained relativelylow levels of radioactivity, several glycolytic intermediates(including FDP, PEP, and 2- and 3-phosphoglyceric acids),hexose monophosphate(s), and lactic acid were detected byPEI-TLC (Fig. 9A). The low level of cell-associated radio-activity was consistent with both the earlier transport data(Fig. 4A) and the fermentation of intracellular fructose (Fig.7A and 8). The metabolism of [14C]fructose was confirmedby the dramatic (and immediate) increase in the intracellularconcentrations of FDP and hexose monophosphate upon theaddition of iodoacetate (a glycolytic inhibitor [41]) to the cellsuspension (Fig. 9B). The inclusion of glutamate in themedium (Fig. 4A) stimulated the synthesis of intracellularpolymer, and the overall rate of fructose utilization by thecells increased by almost 100% (-20 nmol of fructoseconsumed per mg of protein per min [Fig. 7A]). In sharpcontrast to fructose-grown cells, the utilization of fructoseby glutamate-grown organisms was minimal (Fig. 7B). Al-though the presence of glutamate stimulated the rate offructose disappearance from the medium (Fig. 7B), this

VOL. 172, 1990

on Novem

ber 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

5718 ROBRISH AND THOMPSON

.<*: .,

B

as

FIG. 2. Thin-section electron micrographs of fructose-grown cells of F. nucleatum (A) and the same cells after overnight incubation inbuffer (B). Cells grown previously on glutamate are shown in panel C. Note the absence of the granular inclusions in panels B and C.Magnification, x32,500. Bar, 0.5 ,um.

effect was largely due to the deposition of intracellularpolymer rather than to an increase in the rate of sugarfermentation.Mechanism(s) of fructose phosphorylation. The formation

of FlP by cells of F. nucleatum (Fig. SB) suggests theoperation of a PEP-dependent fructose:phosphotransferasesystem (PTS) (28) in this organism. However, all attempts todemonstrate fructose-PTS activity in permeabilized cells orvia the in vitro reconstitution of membrane and cytoplasmiccomponents have been unsuccessful (data not shown). The

possibility of kinase-mediated phosphorylation of fructoseinvolving ATP, acetyl phosphate, or PEP as the phosphoryldonor has also been examined, again without success. Thus,under conditions in which the ATP-dependent phosphoryla-tion of glucose and galactose was readily demonstrable(Table 1), phosphorylation of fructose was not detected.That FlP is an intermediate in the metabolism of fructose isindicated by the presence of ATP-dependent FiP kinaseactivity in cell extracts of F. nucleatum. Significantly, theactivity of ATP-dependent FiP kinase in fructose-grown

- .. n ..--,-- - .:

Wm . -mmoio"

J. BACTERIOL.

4t

.1 ^!F4i 11%411-xi",-ftl..

on Novem

ber 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

FRUCTOSE METABOLISM BY FUSOBACTERIUM NUCLEATUM 5719

A. B.

SF. 2.-

lactate-

Hexose.phosphate

3-PGA-- * _ . :PEP .

P~~EPg

Origin -

11 2 3 4

FIG. 3. Identification by paper chromatography of the radiola-beled sugar produced by acid hydrolysis (lane 2) or by a-amyloglu-cosidase treatment (lane 3) of the ["4C]fructose-derived polymer inF. nucleatum. Lanes 1 and 4 contain radiolabeled fructose (Fru.)and glucose (Glu.) standards.

cells (-79 nmol of FlP converted to FDP per mg of proteinper min) was approximately fivefold greater than that deter-mined in glutamate-grown cells (-17 nmol of FlP convertedto FDP per mg of protein per min).

DISCUSSIONThe probable pathways for metabolism of fructose and

glutamic acid by F. nucleatum ATCC 10953 are presented in

1 2 3

FIG. 5. Identification by PEI-TLC of intracellular metabolitesformed from ["4C]fructose by cells of F. nucleatum grown previ-ously on fructose (A) or glutamate (B). Washed cells were incubatedwith ["4C]fructose, and intracellular metabolites were obtained byboiling water extraction (see Materials and Methods for details).SF.1 and SF.2 refer to solvent fronts 1 (H20) and 2 (0.5 M LiCI-2 Nformic acid, 1:1), respectively. 3-PGA, 3-Phosphoglyceric acid.Arrow in panel B indicates position of "4C-labeled sugar phosphateformed by glutamate-grown cells. This derivative when treated withalkaline phosphatase yielded ["4C]fructose as shown in the paper

chromatogram (C, lane 1). The migration positions of the standards['4C]glucose (Glu.) and ["4C]fructose (Fru.) are shown in lanes 2 and3, respectively, of panel C.

Fig. 10. This illustration provides the basis for discussion ofour results and depicts the interrelationships between sugarand amino acid fermentations in the oral anaerobe. Thesalient results from our investigation are the findings (i) thatF. nucleatum has the capacity to utilize fructose as an

xl( X10U

025

0 10 20

U_ ~~~~~~~~~8-15-

C-)~~~ ~ ~ ~ ~ ~ ~~~~

~0

0 10 20 30 40 0 10 20 30 40

Time (minutes)FIG. 4. Accumulation of [14C]fructose (and a fructose-derived product[s]) by washed cells of F. nucleatum grown previously on fructose

(A) or glutamate (B) as the energy source. Experimental conditions are indicated by the following symbols: O, aerobic; *, aerobic plusglutamate; O, anaerobic; *, anaerobic plus glutamate. Sampling and assay procedures are described in Materials and Methods.

C.

-Fru.

r' IFZ -Glu.

3:u

-Glu.

0

LL4

i

VOL. 172, 1990

- Fru.

on Novem

ber 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

5720 ROBRISH AND THOMPSON

A.

B.

1.v

C.

,-I I .-2

FIG. 6. 31P-NMR spectra of extracts prelnucleatum during metabolism of glutamatefructose (B). Panel C shows the spectrum obtzof 5 mM final concentrations of FlP and F6P sshown in B. The chemical shift scale is in paiPi resonance set at zero.

33 energy source for growth, (ii) that the transport and metab-olism of fructose by resting (i.e., nongrowing) cells is notdependent on provision of a fermentable amino acid asenergy source, and (iii) that excess ATP (generated viaamino acid fermentation) promotes the diversion of fructosefrom the glycolytic pathway toward the polymer-synthesiz-ing pathway of F. nucleatum (Fig. 10). Studies with restingcells, grown previously on either fructose or glutamic acid,were instrumental to elucidation of the mechanism(s) offructose dissimilation by F. nucleatum. Under anaerobicconditions, washed (i.e., resting) cells grown previously onfructose rapidly fermented the sugar to a mixture of D-lac-tate, butyrate, and acetate. The capacity of the cells to

33 transport fructose is in marked contrast to results from ourearlier studies (31) which revealed an obligate requirementfor a fermentable amino acid (e.g., glutamate, lysine, orhistidine) for the transport of glucose and galactose by F.nucleatum. Since the fermentation of fructose occurs via theEmbden-Meyerhof (glycolytic) pathway, it is not clear whyglucose and galactose are not also rapidly metabolized by F.nucleatum. The presence of intracellular FlP in the cellssuggests that this is the first derivative of fructose metabo-lism and that after phosphorylation via an ATP-dependentFlP kinase, fructose enters the glycolytic pathway at thelevel of FDP (Fig. 10). Glucose and galactose are phosphor-ylated by ATP-dependent kinases to yield glucose-6-phos-phate and galactose-l-phosphate, respectively, and thesecompounds would be expected to enter the glycolytic path-way at the level of glucose-6-phosphate. The capacity of F.nucleatum to rapidly ferment a particular sugar may dependon the point at which the various sugar phosphates enter theglycolytic pathway.

Although glutamate-grown cells failed to ferment fructose,1 PPM the resting cells accumulated high levels of FlP (when

pared from cells of F. glutamate was supplied as an energy source [Fig. 4B]), and(A) or glutamate plus the fructose monophosphate was rapidly transformed intoained after the addition polymer. Surprisingly, the polymer granules formed by;tandards to the extract growing and resting cells were composed of glucose ratherrts per million with the than fructose. a-Amyloglucosidase cleaves a(1--4)- and

a(1-*6)-glucosidic linkages, and the polymer is evidently aglucan and most probably glycogen. Experiments conductedwith glutamate-grown cells revealed the constitutivity of the

2.0A

E 1.5C X

a)~~~~~lE\ X

001.0

._-oLI0.5_-*

-0 10 20 30 40 0 10 20 30 40Time (minutes)

FIG. 7. Fructose utilization by washed cells of F. nucleatum grown previously on fructose (A) or glutamate (B). The cells were suspendedunder anaerobic conditions in buffered medium containing 2 mM fructose and either no further additions (A) or 20 mM sodium glutamate (A).

J. BACTERIOL.

1.xI

on Novem

ber 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

FRUCTOSE METABOLISM

40,000

30,000

20,000 F

10,000 ItLI0t 10 20 30 40 50 60

Fraction Number (200 j1l/fraction)

BY FUSOBACTERIUM NUCLEATUM 5721

70 80

FIG. 8. Products of the anaerobic fermentation of fructose by fructose-grown cells of F. nucleatum. Washed cells were suspended in Colesbuffer (1.6 mg of protein per ml) containing 2 mM [U-'4C]fructose (specific activity, 2 p,Ci/,umol). Filtrates were collected immediately afteraddition of the cells (zero time, Control) and after 90 min of incubation. High-performance liquid chromatographic analyses revealed only[14C]fructose (peak 1) in the control filtrate (0). After 90 min of incubation (shaded areas, *), labeled sugar was barely detectable andradioactivity was recovered as the following fermentation products: peak 2, D-lactate (-60%); peak 3, acetate (-7%); and peak 4, butyrate(-20%). Approximately 15% of the initial radioactivity was not recovered in our analyses, and this deficiency is attributed to the release of[14C]C02 which would accompany the formation of acetic and butyric acids (12). (In this experiment, the rate of fructose utilization by cellsunder anaerobic conditions was 12.7 nmol of sugar per mg of protein per min. Under aerobic conditions, the rate of sugar fermentationdecreased by 60% [5 nmol of fructose used per mg of protein per min], and only [14C]acetate was detected in the filtrates).

A. B.

SF. 2.?lactate-

Hexose

phosphate

3-PGA-

FDp

Origin -- "

0

LL-

FIG. 9. Qualitative analysis and identification (by PEI-TLC) ofintracellular metabolites of cells of F. nucleatum. Cell extracts wereprepared from organisms during metabolism of ['4C]fructose (A) andafter addition of 5 mM iodoacetic acid (a glycolytic inhibitor) to theglycolyzing cell suspension (B). 3-PGA, 3-Phosphoglycerate.

enzymes involved in polymer synthesis and of those systemsrequired for the transport and phosphorylation of fructose.However, the nature of the latter system(s), and the identityof the phosphoryl donor, have thus far proved elusive. Manybacteria translocate and simultaneously phosphorylate fruc-tose via the fructose:PEP-PTS (28), and the product of thisgroup translocation reaction is invariably FlP. Formation ofthis derivative by F. nucleatum suggests operation of thefructose-PTS, but our attempts to detect such activity inpermeabilized cells or via subcellular complementation havefailed. Fox et al. (8) suggest that acetylphosphate serves inlieu of PEP as the phosphoryl donor for PTS-mediated sugarphosphorylation in Escherichia coli. Since acetylphosphateis generated during glutamate fermentation by F. nucleatum,we tried this high-energy compound as a potential donor forthe fructose-PTS, but again without success (data not report-ed).

TABLE 1. In vitro phosphorylation of sugarsa by a cell extractbof F. nucleatum ATCC 10953

Sugar Resultc with the following phosphoryl donor:substrate None ATP PEP Acetylphosphate

Glucose 0.76 15.70 1.64 0.55Galactose 0.33 11.58 0.30 0.77Fructose NDd ND ND ND

a Assay conditions were as described in Materials and Methods.b The cell extract was prepared from glutamate-grown cells.c Results are expressed as nanomoles of [14C]sugar phosphorylated per

milligram of protein per minute.d ND, Not detectable.

VOL. 172, 1990

c

0

C.)

a-

w-

C.,a.

._

0

VF

ni

-1-icCL)

0U)

'f-

1. f.' .;:.;f -4 .: r; 1. ..

on Novem

ber 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

5722 ROBRISH AND THOMPSON

FRUCTOSE

I ATP F1PI 1

ADPlFDP

(Acetyl-P FP

PEP) TP? ADP I

-4'>2.ATPY3-PGA

I tP EP

PYRUVATECo2

Acetate

Butyrate X 0d2o

NH3, CO2acetic acidbutyric acid

glutamate

glutamate

a)

co

0)-

pF6P > tQ

G6P I

Acetyl-P

G1P ATP

4..4

ADP-glucose

GLUCANpathway of

polymer synthesis

LACTATEglycolyticpathway

FIG. 10. Diagrammatic representation of the interrelationshipsbetween fructose metabolism, polymer synthesis, and glutamatefermentation by F. nucleatum ATCC 10953. Numbers refer to theenzymes: 1, FlP kinase; 2, glyceraldehyde-3-phosphate dehydroge-nase (inhibited by iodoacetate [IAA]); 3, fructose-1,6-diphos-phatase; and 4, ADP-glucose pyrophosphorylase. ElI refers toenzyme IIlf' of the putative fructose-PTS, and gltp indicates a

glutamate-specific permease or active transport system. 3-PGA,3-Phosphoglycerate; TP, triosephosphates; G6P, glucose-6-phos-phate; Acetyl-P, acetylphosphate; GlP, glucose-i-phosphate.

Although it is generally regarded as an anaerobe, studiesby Loesche (22) showed growth of F. nucleatum in oxygentensions approaching 6%; however, the cells died quicklywhen exposed to air. Our attempts to detect fructose phos-phorylation were not performed under strict anaerobic con-ditions, and oxygen sensitivity of the putative fructose-PTSmust be considered. In this context, we note that the rate offructose metabolism by aerobic cells is only 40% of the ratedetermined under anaerobic conditions (Fig. 8, legend). It isconceivable that fructose accumulation occurs via activetransport and that the intracellular (free) sugar is phosphor-ylated by an ATP-dependent kinase to yield FlP, as de-scribed recently in Halobacterium vallismortis (1). How-ever, we were unable to detect FlP (or F6P) synthesis in

cell-free preparations using ATP, PEP, and acetylphosphateas potential phosphoryl donors (Table 1).An important contribution from our studies is the finding

that the fate of intracellular FlP is determined by theavailability and cofermentation of glutamic acid. Whereas inthe absence of glutamate, fructose was metabolized almostexclusively via the glycolytic pathway, in the presence of theamino acid, intracellular FlP was also channelled into poly-mer synthesis (Fig. 10). The biochemical basis for thisdiversion has yet to be established, but it is significant thatthe rate of glutamate fermentation (and ATP generation) byF. nucleatum is very high (-0.12 ,umol/mg [dry weight] ofcells per min [31]). If ATP production exceeds cellulardemand, then the surplus may be consumed in the reaction:ATP + a-glucose-i-phosphate -> ADP-glucose + PPi. Anincrease in the activity of ADP-glucose pyrophosphorylase(ATP:a-glucose-1-phosphate adenyl transferase, [29]) wouldthus pull the branchpoint metabolite (FDP) away from theglycolytic pathway and into the polymer-synthesizing path-way.

Fructose-grown cells rapidly ferment fructose (Fig. 7A),whereas fermentation of the sugar by glutamate-grown or-ganisms is barely detectable (Fig. 7B). The inability of thelatter cells to ferment fructose may, in part, be attributableto the low level of FlP kinase activity in these cells (approx-imately 20% of that determined in fructose-grown cells). If,because of the reduced activity of FlP kinase, the rate ofFDP formation becomes comparable to the rate at whichFDP is dephosphorylated to F6P (Fig. 10, step 3), thencontinued transport and metabolism of fructose will not bepossible. In summary, we believe that FlP kinase levels andATP availability are the two factors which regulate the entryof fructose into either the catabolic (glycolytic) or anabolic(polymer-synthesizing) pathway in F. nucleatum.

Previously (32), we showed that concomitant fermentationof glutamate by F. nucleatum prevented the degradation ofintracellular polymer derived from glucose or galactose. Inthe present study, by comparative electron microscopy (Fig.2A and B), we provided evidence for the converse, i.e., thatfructose-derived granules are rapidly metabolized by thecells in the absence of glutamate (starvation overnight). Theresults of this investigation and of previous studies (31-33)demonstrate the capacity of F. nucleatum (i) to form thesame intracellular polymer from glucose, galactose, andfructose under conditions of amino acid excess and (ii) toferment this sugar reserve under conditions of amino aciddeprivation. The biochemical synergism between sugar andamino acid fermentations may contribute to the survival ofF. nucleatum in the feast-or-famine environment of the oralcavity and to the persistence of this organism in periodontaldisease.

ACKNOWLEDGMENTS

We thank J. London and P. E. Kolenbrander for constructivecriticisms of the manuscript, and we express appreciation to S. E.Mergenhagen for his interest and support of our program. We thankC. Oliver, J. Waters, and 0. Ambrose for preparation of the electronmicrographs and acknowledge the help of W. M. Egan with high-resolution NMR spectroscopy. We thank Irma Gomez for technicalassistance and Charlette Cureton for preparation of the manuscript.

LITERATURE CITED1. Altekar, W., and V. Rangaswamy. 1990. Indication of a modifiedEMP pathway for fructose breakdown in a halophilic archae-bacterium. FEMS Microbiol. Lett. 69:139-144.

2. Bakken, V., B. T. Hogh, and H. B. Jensen. 1989. Utilization of

J. BACTERIOL.

I>-1.1=

-t--

on Novem

ber 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from

FRUCTOSE METABOLISM BY FUSOBACTERIUM NUCLEATUM 5723

amino acids and peptides by Fusobacterium nucleatum. Scand.J. Dent. Res. 97:43-53.

3. Barker, H. A. 1981. Amino acid degradation by anaerobicbacteria. Annu. Rev. Biochem. 50:23-40.

4. Barker, H. A., J. M. Kahn, and L. Hedrick. 1982. Pathway oflysine degradation in Fusobacterium nucleatum. J. Bacteriol.152:201-207.

5. Bennett, K. W., and B. I. Duerden. 1985. Identification offusobacteria in a routine diagnostic laboratory. J. Appl. Bacte-riol. 59:171-181.

6. Buckel, W., and H. A. Barker. 1974. Two pathways of glutamatefermentation by anaerobic bacteria. J. Bacteriol. 117:1248-1260.

7. Coles, R. S., Jr. 1977. Glucose utilization by resting cells ofFusobacterium polymorphum. Arch. Oral Biol. 22:87-90.

8. Fox, D. K., N. D. Meadow, and S. Roseman. 1986. Phosphatetransfer between acetate kinase and enzyme I of the bacterialphosphotransferase system. J. Biol. Chem. 261:13498-13503.

9. Gharbia, S. E., and H. N. Shah. 1988. Glucose utilization andgrowth response to protein hydrolysates by Fusobacteriumspecies. Curr. Microbiol. 17:229-234.

10. Gharbia, S. E., and H. N. Shah. 1989. The uptake of amino acidsfrom a chemically defined medium by Fusobacterium species.Curr. Microbiol. 18:189-193.

11. Gibbons, R. J., S. S. Socransky, S. Sawyer, B. Kapsimalis, andJ. B. MacDonald. 1963. The microbiota of the gingival crevicearea of man. II. Predominant cultivable organisms. Arch. OralBiol. 8:319-329.

12. Gottschalk, G. 1986. Bacterial fermentations, p. 208-282. In G.Gottschalk (ed.), Bacterial metabolism, 2nd ed. Springer-Ver-lag, New York.

13. Gottschalk, G., and H. von Hugo. 1982. Phosphofructokinasefrom Clostridium pasteurianum. Methods Enzymol. 90:82-87.

14. Herbert, D., P. J. Phipps, and R. E. Strange. 1971. Chemicalanalysis of microbial cells. Methods Microbiol. 5B:209-344. .

15. Hofstad, T. 1981. The genus Fusobacterium, p. 1464-1474. InM. P. Starr, H. Stolp, H. G. Truper, A. Balows, and H. G.Schlegel (ed.), The prokaryotes. A handbook on habitats,isolation, and identification of bacteria. Springer-Verlag, NewYork.

16. Hofstad, T. 1984. Pathogenicity of anaerobic gram-negativerods: possible mechanisms. Rev. Infect. Dis. 6:189-199.

17. Holdeman, L. V., E. P. Cato, and W. E. C. Moore. (ed.). 1977.Anaerobe laboratory manual, 4th ed. Virginia Polytechnic In-stitute and State University, Blacksburg.

18. Holdeman, L. V., R. W. Kelley, and W. E. C. Moore. 1984.Anaerobic gram-negative straight, curved and helical rods, p.602-662. In N. R. Krieg and J. G. Holt (ed.), Bergey's manualof systematic bacteriology, vol. 1. The Williams & Wilkins Co.,Baltimore.

19. Jackins, H. C., and H. A. Barker. 1951. Fermentative processesof the fusiform bacteria. J. Bacteriol. 61:101-114.

20. Karnovsky, M. J. 1971. Use of ferrocyanide-reduced osmiumtetroxide in electron microscopy. J. Cell Biol. 51:146a.

21. Loe, H., E. Theilade, and S. B. Jensen. 1965. Experimentalgingivitis in man. J. Periodontol. 36:177-187.

22. Loesche, W. J. 1969. Oxygen sensitivity of various anaerobicbacteria. Appl. Microbiol. 18:723-727.

23. Loesche, W. J., and R. J. Gibbons. 1968. Amino acid fermenta-tion by Fusobacterium nucleatum. Arch. Oral Biol. 13:191-201.

24. Maskinen, K. K., H. Hurttia, P.-L. Makinen, and K. Paunio.1984. Effects on oral health of mouthrinses containing xylitol,sodium cyclamate and sucrose sweeteners in the absence of oralhygiene. II. Relative composition of free amino acids in humancrevicular fluid. Proc. Finn. Dent. Soc. 80:13-19.

25. Moore, W. E. C., L. V. Holdeman, R. M. Smibert, I. J. Good,J. A. Burmeister, K. G. Palcanis, and R. R. Ranney. 1982.Bacteriology of experimental gingivitis in young adult humans.Infect. Immun. 38:651-667.

26. Moore, W. E. C., L. V. Holdeman, R. M. Smibert, D. E. Hash,J. A. Burmeister, and R. R. Ranney. 1982. Bacteriology ofsevere periodontitis in young adult humans. Infect. Immun.38:1137-1148.

27. Moore, W. E. C., R. R. Ranney, and L. V. Holdeman. 1982.Subgingival microflora in periodontal disease: cultural studies,p. 13-26. In R. J. Genco and S. E. Mergenhagen (ed.), Host-parasite interactions in periodontal diseases. American Societyfor Microbiology, Washington, D.C.

28. Postma, P. W., and J. W. Lengeler. 1985. Phosphoenolpyruvate:carbohydrate phosphotransferase system of bacteria. Micro-biol. Rev. 49:232-269.

29. Preiss, J. 1984. Bacterial glycogen synthesis and its regulation.Annu. Rev. Microbiol. 38:419-458.

30. Reynolds, E. S. 1963. The use of lead citrate at high pH as anelectron opaque stain in electron microscopy. J. Cell Biol.17:208-212.

31. Robrish, S. A., C. Oliver, and J. Thompson. 1987. Aminoacid-dependent transport of sugars by Fusobacterium nuclea-tum ATCC 10953. J. Bacteriol. 169:3891-3897.

32. Robrish, S. A., and J. Thompson. 1988. Suppression of polyglu-cose degradation in Fusobacterium nucleatum ATCC 10953 byamino acids. FEMS Microbiol. Lett. 55:29-34.

33. Robrish, S. A., and J. Thompson. 1989. Na+ requirement forglutamate-dependent sugar transport by Fusobacterium nucle-atum ATCC 10953. Curr. Microbiol. 19:329-334.

34. Roe, J. H. 1934. A colorimetric method for the determination offructose in blood and urine. J. Biol. Chem. 107:15-22.

35. Singer, D. L., and I. Kleinberg. 1983. The free amino acids inhuman dental plaque. Arch. Oral Biol. 28:873-878.

36. Singer, R. E., and B. A. Buckner. 1981. Butyrate and propion-ate: important components of toxic dental plaque extracts.Infect. Immun. 32:458-463.

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

38. Socransky, S. S., A. C. R. Tanner, A. D. Haffajee, J. D. Hillman,and J. M. Goodson. 1982. Present status of studies on themicrobial etiology of periodontal diseases, p. 1-12. In R. J.Genco and S. E. Mergenhagen (ed.), Host-parasite interactionsin periodontal diseases. American Society for Microbiology,Washington, D.C.

39. Somogyi, M. 1957. Preparation of glycogen, nitrogen- and phos-phorus-free, from liver. Methods Enzymol. 3:3-4.

40. Spurr, A. R. 1969. A low viscosity epoxy resin embeddingmedium for electron microscopy. J. Ultrastruct. Res. 26:31-43.

41. Thompson. J. 1978. In vivo regulation of glycolysis and charac-terization of sugar: phosphotransferase systems in Streptococ-cus lactis. J. Bacteriol. 136:465-476.

42. Thompson, J., and D. A. Torchia. 1984. Use of 31P nuclearmagnetic resonance spectroscopy and 14C fluorography in stud-ies of glycolysis and regulation of pyruvate kinase in Strepto-coccus lactis. J. Bacteriol. 158:791-800.

43. Walker, C. B., D. Ratliff, D. Muller, R. Mandell, and S. S.Socransky. 1979. Medium for selective isolation of Fusobacter-ium nucleatum from human periodontal pockets. J. Clin. Micro-biol. 10:844 849.

44. Werner, H., F. Neuhaus, and H. Hussels. 1971. A biochemicalstudy of fusiform anaerobes. Med. Microbiol. Immunol. 157:10-16.

VOL. 172, 1990

on Novem

ber 3, 2020 by guesthttp://jb.asm

.org/D

ownloaded from