INFEcTION AND IMMUNITY, Aug. 1992, p. 3175-31850019-9567/92/083175-11$02.00/0Copyright © 1992, American Society for Microbiology

Cloning of a Locus Involved in Streptococcus mutansIntracellular Polysaccharide Accumulation and Virulence

Testing of an Intracellular Polysaccharide-Deficient MutantGRACE SPATAFORA HARRIS,lt* SUZANNE M. MICHALEK,2 AND ROY CURTISS III'

Department of Biology, Washington University, St. Louis, Missouri 63130,1 and Department ofMicrobiology,University ofAlabama at Birmingham, Birmingham, Alabama 352942

Received 18 February 1992/Accepted 8 May 1992

The streptococcal transposon Tn916 (Tcr) was used to isolate mutants of Streptococcus mutans altered inglycogen accumulation to investigate whether glycogenlike intracellular polysaccharides (IPS) play animportant role in S. mutans-induced caries formation. S. mutans UA130 (serotype c) was transformed with theEscherichia coli plasmid pAM620 (Tn916), and the resultant transposon libraries were screened for glycogencontent by iodine staining. A transposon mutant, designated SMS201, demonstrated a glycogen-deficientphenotype on glucose-enriched medium. Quantitative electron microscopy confirmed that IPS concentrationswere significantly reduced in SMS201 relative to the wild-type progenitor strain, UA130. Importantly,reversion to wild type correlated at all times with loss of the transposon. Transposon excisants were used tofacilitate cloning of the streptococcal gene(s) involved in glycogen biosynthesis and storage. A 2.1-kbchromosomal determinant (glgR) which encodes a putative regulator of S. mutans glycogen accumulation wasisolated. A stable deletion mutation (AglgR) was subsequently generated in E. coli and introduced into S.mutans by allelic exchange. The resultant glycogen synthesis-deficient mutant, SMS203, demonstrated a

significantly reduced cariogenic potential (P < 0.01) on the buccal, sulcal, and proximal surfaces of teeth ingermfree rats, relative to animals challenged with the glycogen synthesis-proficient progenitor strain, UA130.These observations confirm previous reports (J. M. Tanzer, M. L. Freedman, F. N. Woodiel, R. L. Eifert, andL. A. Rinehimer, p. 597-616, in H. M. Stiles, W. J. Loesche, and T. L. O'Brien, ed., Proceedings inMicrobiology. Aspects of Dental Caries. Special Supplement to Microbiology Abstracts, vol. 3, 1976) whichimplicate IPS as significant contributors to the S. mutans cariogenic process.

Tooth decay is a chronic disease and perhaps the mostubiquitous bacterial infection in humans (23). Streptococcusmutans, the principal etiologic agent of human dental caries,is indigenous to the oral flora and often colonizes the oralcavity shortly after tooth eruption. Much of what we knowabout the damage caused to host tissues after colonizationwith this pathogen has derived from studies with germfreerats (26, 29), in which evidence suggests that dental caries isa multifactorial infectious process.Prominent among the proposed virulence traits for S.

mutans is the ability of the organism to metabolize ferment-able carbohydrates in the host diet and to produce acid in theoral cavity at concentrations sufficient to demineralize toothenamel (6, 7). Interestingly, S. mutans produces acid notonly from exogenous dietary carbohydrates, but also fromintracellular polysaccharides (IPS), glycogenlike storagepolymers with al,4 and al,6 linkages (15). In gram-negativemicroorganisms, ADP-glucose pyrophosphorylase (EC2.7.7.27), glycogen synthetase (EC 2.4.1.21), and branchingenzyme (EC 2.4.1.18) are all known to be involved in thebiosynthesis of IPS (2, 31). While the former enzymes havebeen identified in S. mutans, glycogen branching activity hasyet to be described in the streptococci. Nevertheless, glyco-gen accumulation in S. mutans is likely to follow a pathwaysimilar to that described in the members of the familyEnterobacteriaceae. Specifically, ADP-glucose is synthe-sized from ATP and glucose-i-phosphate by the ADP-

* Corresponding author.t Present address: Department of Microbiology, University of

Alabama at Birmingham, Birmingham, AL 35294.

glucose pyrophosphorylase, while glycogen synthetase cat-alyzes the formation of glycogen by using ADP-glucose asthe glucosyl donor. The net result of this glycogen accumu-lation in S. mutans may be to foster the progression of cariesdevelopment by prolonging the period of exposure of hosttissues to organic acid\s when dietary sugars have beendepleted from the oral cavity, such as at non-meal times.Tanzer and coworkers (38) have reported that nitrosogua-

nidine-generated mutants of S. mutans with defects relatedto IPS are less cariogenic in vivo than wild-type S. mutans.Conclusions drawn from these studies are tentative, how-ever, since the lesions generated in these mutants werenever genetically or biochemically defined. Lending furthersupport to an association of IPS with S. mutans-inducedcaries formation is the observation that streptococci isolatedfrom carious lesions are predominantly synthesizers of IPS,while those isolated from caries-inactive plaque are intra-cellular polysaccharide-negative variants (13).To define more precisely the role of intracellular polysac-

charides, presumably glycogen, in S. mutans virulence, andto facilitate cloning of the gene(s) involved in S. mutansglycogen accumulation, we used transposon mutagenesisand gene replacement strategies to construct mutants of S.mutans UA130, a cariogenic serotype c strain. In this report,we describe the construction and characterization of S.mutans glycogen-deficient mutants SMS201 and SMS203and examine their cariogenic potential in germfree rats. Ourfindings indicate that mutants defective in IPS accumulationare indeed less virulent than their IPS synthesis-proficientwild-type progenitors.

3175

Vol. 60, No. 8

on July 18, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

3176 HARRIS ET AL.

TABLE 1. Bacterial strains and plasmids

Genotype and phenotype' Source or reference

Eschenichia coli K-12DH5ct

NM539CGSC 4445 (X6153)CGSC 4447 (X6155)

Streptococcus mutansUA130SMS201SMS203

PlasmidspAM620pVA838pYA2329pYA3000pYA3001pYA3023pYA3034pYA3044pYA3045pYA3053pYA3076pYA3077

F- endAl hsdRl7(rK MK+) supE44 thi-I recAl gyrA96 relAlD(lacZYA-argF)U169+80d(lacZ)DM15

supF hsdR lacY (P2coxB)glgAglgC

S. mutans serotype c Glg+UA130::Tn916 Glg-UA130 AglgX Glg-

pVA891::pAD1EcoRI F'::Tn916 Emr TcrShuttle vector; Emr CmrpCP13 derived (9); Kmr SprpYA2329::18.5-kb PstI insert:Tn916 Kmr Spr TcrpYA2329::2.1-kb PstI insert Kmr SprpUC19 derived, 6.9-kb BglII insert; AprpVA838 derived, 8.3-kb XbaI insert; EmTpUC19 derived, 2.1-kb PstI insert; ApTpUC19 derived, 2.1-kb PstI insert; ApTpACYC184 derived, 16-kb SalI insert; CmrpYA3023 derived; aph AglgX Apr KmrpBR322 derived; aph AglgX Kmr Tcr

16

14B. BachmannB. Bachmann

P. CaufieldThis workThis work

3924This laboratoryThis laboratoryThis laboratoryThis workThis workThis workThis workThis workThis workThis work

a Em, erythromycin; Tc, tetracycline; Km, kanamycin; Sp, spectinomycin; Cm, chloramphenicol; Ap, ampicillin.

MATERIALS AND METHODS

Bacterial strains and plasmids. The bacterial strains andplasmids used in the present study are described in Table 1.

Culture conditions. Escherichia coli DH5a cells were

grown aerobically at 37°C in Lennox broth supplementedwith 0.1% dextrose. Todd-Hewitt (TH) broth (Difco Labo-ratories, Detroit, Mich.) containing 5% horse serum (GIBCOLaboratories, Grand Island, N.Y.) was used to propagatecultures of S. mutans routinely grown at 37°C in an anaero-

bic GasPak System (BBL Microbiology Systems, Cockeys-ville, Md.). Agar (Difco) was added to a final concentrationof 1.5% in the preparation of solid media. Colony morphol-ogies were examined periodically on mitis salivarius agar(Difco), and streptococci were subsequently checked forpurity upon microscopic examination.

Electroporation of E. coli was performed as describedpreviously (11). For S. mutans, overnight cultures of UA130were diluted 1:20 in fresh TH broth and grown as standingcultures at 37°C for 3.5 h. After incubation on ice for 15 min,cells were harvested by centrifugation and washed twice insuccession with 1 ml of cold 300 mM raffinose. Cells resus-

pended in 120 ,ul of cold 300 mM raffinose were divided into40-,u aliquots for subsequent electroporation.To assay for intracellular polysaccharide content, we grew

S. mutans anaerobically at 37°C for 5 days on Jordan'smedium (17) supplemented with 2% glucose with appropriateantibiotic selection.

Bacteriophage libraries of S. mutans genomic DNA were

stored as liquid lysates and used subsequently to infectgrowing cultures of E. coli NM539 at a multiplicity ofinfection of 0.01. After adsorption at 37°C, infected cellswere plated as overlays in 0.5% NZYM top agarose (34).Plaque formation proceeded at 37°C in a moisture chamberfor 12 to 16 h.For growth rate and acid production determinations, S.

mutans UA130 and SMS203 were grown as standing culturesin TH broth with or without 2% glucose. Samples were

withdrawn at various time points and measured for cell

density (optical density at 560 nm) and pH with an LKBUltrospec 4050 spectrophotometer and Orion digital pHmeter, respectively.

Cell preparations for biochemical and ultrastructural anal-yses involved inoculating 500 ml of TH broth containing 2%glucose with 5 ml of a UA130 or SMS203 overnight culture.Standing cultures grown at 37°C were maintained at pH 7.0for up to 12 h with 5 M NaOH. Cells were harvested bycentrifugation and used in the preparation of crude enzyme

extracts as described previously (2). For electron micros-copy, cells harvested at 5 (late logarithmic), 7 (early station-ary), and 12 (stationary) h were subsequently fixed in 3%glutaraldehyde as previously described (10).

Mutagenesis of S. mutans with Tn916. The E. coli plasmidpAM620 (36), which harbors the streptococcal transposonTn916, served as an efficient delivery vehicle for the gener-ation of transposon insertions into the S. mutans chromo-some. Specifically, 10 ,ug of pAM620 was used to transform40 ,ul of S. mutans UA130 by electroporation. The resultanttransposon mutants were selected on TH agar supplementedwith 5 ,ug of tetracycline per ml. Approximately 3,000colonies demonstrating resistance to tetracycline (Tcr) weresubsequently streaked onto Jordan's medium and analyzedfor IPS content.Assay for IPS content. The accumulation of IPS in S.

mutans was measured qualitatively by staining coloniesgrown on Jordan's medium with a 0.2% I2 in 2% KI solution(13). After incubation in the dark, glycogen-deficient colo-nies demonstrated a yellow phenotype, while glycogen-proficient colonies stained dark brown.

Preparation of crude enzyme supernatants. Crude proteinextracts were prepared from S. mutans cultures as describedpreviously (2), except that cells were disrupted in a homog-enizer (Braun, Melsungen, Germany) for 3 min and cooledwith intermittent bursts of liquid CO2. The disrupted cellsuspension was centrifuged at 37,000 x g for 20 min at 0°C,and the supernatant, termed the crude enzyme preparation,

Strain or plasmid

INFECT. IMMUN.

on July 18, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

IPS AND S. MUTANS VIRULENCE 3177

was standardized with a bicinchoninic acid protein assay

reagent (Pierce).Assay for ADP-glucose pyrophosphorylase and glycogen

synthetase activities in S. mutans. The reduction of nitrobluetetrazolium was measured at an optical density of 540 nm toassess S. mutans ADP-glucose pyrophosphorylase (glgC)and glycogen synthetase (glgA) activities. The assays were

performed in the dark and at room temperature as describedpreviously (12, 30), and optical densities were recorded withan enzyme-linked immunosorbent assay reader.

Isolation and purification of DNA. Chromosomal DNAfrom S. mutans was isolated by using a modification of themethod of Marmur (25). The DNA was purified by CsCl-ethidium bromide equilibrium density gradient centrifuga-tion and digested with restriction enzymes (Promega Biotec,Madison, Wis.) according to the recommendations of thesupplier. Plasmid DNA was extracted from E. coli by thealkaline lysis method of Birnboim and Doly (3).

Cloning and subcloning of the glycogen gene (glg) locus fromS. mutans. Chromosomal DNA isolated from glycogen-altered transposon mutants was digested with restrictionendonucleases, resolved on agarose gels, transferred tonitrocellulose membranes, and probed with radiolabeledpAM620 according to standard protocols (34). Restrictionfragments containing the transposon along with S. mutansflanking DNA were cloned into the cosmid vector pYA2329,and E. coli transductants (pYA3000) demonstrating transpo-son-encoded tetracycline resistance were isolated. Trans-ductants subsequently grown in the absence of tetracyclineselection gave rise to the transposon excisant, pYA3001.pYA3001 harbors a 2.1-kb PstI fragment which is putativelyinvolved in S. mutans glycogen accumulation.For subcloning, DNA fragments were isolated by electro-

elution (Elutrap) after electrophoresis in 0.7% agarose. Spe-cific fragments were subsequently ligated with pBR322,pUC19, pACYC184, or pVA838 DNA previously digestedwith appropriate restriction endonucleases. T4 DNA poly-merase and T4 DNA ligase used in subcloning experimentswere purchased from Promega Biotec, and the reactionconditions were as specified by the manufacturer.

Electroporation of bacterial strains. Plasmids purified byCsCl-ethidium bromide equilibrium density gradient centrif-ugation were introduced into E. coli by electroporation byusing a Gene Pulser apparatus (Bio-Rad Laboratories, Rich-mond, Calif.) as described previously (11). Transformantswere selected on L agar supplemented with ampicillin,chloramphenicol, kanamycin, or tetracycline at a final con-

centration of 100, 15, 50, or 15 ,ug/ml, respectively.For S. mutans UA130, electroporation conditions of 25

,uF, 1.25 kV, and 400 Ql were applied to a cell-DNA (40 1./10

,ug) mixture in a chilled 0.2-cm cuvette (Bio-Rad). After a

90-min recovery period in TH broth without antibioticselection, electroporants were plated onto TH agar contain-ing erythromycin, kanamycin, or tetracycline at a finalconcentration of 10, 250, or 5 ,ug/ml, respectively. Subinhib-itory concentrations of erythromycin (75 ng/ml) were incor-porated into solid media to induce expression of thepVA838-borne erythromycin determinant (24).Complementation studies. Subclones pYA3023, pYA3034,

pYA3044, pYA3045, and pYA3053 described in Table 1 wereused to transform E. coli or S. mutans glycogen-deficienthosts by electroporation as previously described. Transfor-mants selected on L agar or TH agar supplemented withappropriate antibiotics were subsequently transferred toJordan's medium and screened for glycogen content byiodine staining.

Preparation of radiolabeled probes. 32P-labeled probeswere prepared in vitro by nick translation based on theprocedure of Rigby et al. (33). Unincorporated label wasseparated from nick-translated DNA by chromatography onSephadex G-100 columns (Pharmacia).Genomic library construction. High-molecular-weight

DNA from S. mutans UA130 was partially digested withSau3A and size selected on 40% sucrose gradients as de-scribed previously (34). DNA in the size range of 10 to 25 kbwas ligated into BamHI-generated arms of lambda EMBL3(14). The ligation mixture was subsequently packaged invitro (Packagene; Promega Biotec) and used to infect E. coliNM539 (Table 1). A library representative of the S. mutansgenome was obtained as determined by the formula ofClarke and Carbon (8). This library was subsequently platedonto NZYM agar and screened by the method of Benton andDavis (1). The 2.1-kb glg fragment from S. mutans was usedas a probe for library screening.

Gel electrophoresis and hybridization methods. Electropho-resis was done in Tris-borate-EDTA or Tris-acetate-EDTAbuffer in horizontal submerged 0.7% agarose gels for 16 h at40 V. DNA was transferred to nitrocellulose membranes(Schleicher & Schuell, Inc., Keene, N.H.) by the method ofSouthern (37). Hybridizations with radiolabeled probes wereperformed in 50% formamide for 24 h at 42°C. Filters weresubsequently washed in 2x SSC (lx SSC is 0.15 M NaClplus 0.015 M sodium citrate, pH 7.0)-0.1% sodium dodecylsulfate (SDS) for 30 min with agitation, followed by twostanding washes at 500C in 0.5x SSC-0.1% SDS. Filterswere air dried and exposed to X-ray film (DuPont) for 24 h at-700C.

Electron microscopy. After fixation in glutaraldehyde, cellswere washed four times in 0.01 M phosphate (pH 6.2)containing 0.8 M KCl and 0.01 M Mg acetate and thensuspended for 30 min in the Veronal-acetate buffer of Kellen-berger et al. (18). Cells subsequently fixed for 2 to 4 h in 1%osmium tetroxide at room temperature were treated with0.5% uranyl acetate overnight and embedded in Polybed 8-12(Polysciences, Inc.). Sections were prepared on a SorvallMT-2 ultramicrotome with a diamond knife (DuPont) andplaced on nickel grids coated with Butvar-carbon for stain-ing. To enhance visualization of bacterial IPS, the sectionswere treated with 1% periodic acid for 30 min at roomtemperature and then with sodium chlorite as previouslydescribed (10), before staining with uranyl acetate and leadcitrate. Specimens were examined in a Hitachi H-600 elec-tron microscope at magnifications ranging from x40,000 tox 150,000.Virulence testing. The cariogenic potential of glycogen-

proficient and glycogen-deficient S. mutans was determinedin young gnotobiotic Fischer rats. Specifically, 19-day-oldweanling rats were challenged orally with approximately 108CFU of the appropriate test strain per ml. Animals weremaintained subsequently on a sterile caries-promoting dietcontaining 5% sucrose (26) provided either ad libitum (24h/day) or at restricted feeding times (6 h/day). Colonizationwas assessed 2 days postchallenge and then weekly for theduration of the experiment by collecting oral swab samplesand culturing them on mitis salivarius agar with or without250 ,ug of kanamycin per ml. Rats sacrificed at 45 days of agewere scored for caries (19), and plaque microbiology wasassessed on mitis salivarius agar with appropriate selection.Microorganisms were subsequently streaked onto Jordan'smedium and stained with iodine as described above to assessglycogen content.

Statistical analysis. All data presented are expressed as the

VOL. 60, 1992

on July 18, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

3178 HARRIS ET AL.

B

78 2 3 4 5 6 7 8----- _

__

FIG. 2. Assay for S. mutans IPS. S. mutans was grown anaero-

bically for 5 days on Jordan's medium and subsequently stained with0.2% I2-2% KI (13). (A) Colorless phenotype represents a deficiencyof IPS in the SMS201 transposon mutant. (B) Wild-type progenitorstrain UA130 is rich in IPS as reflected by a dark brown phenotype.

FIG. 1. Analysis of chromosomal DNA isolated from each ofeight randomly selected S. mutans Tcr transformants. (A) HindIII-digested chromosomal DNAs stained with ethidium bromide andresolved by electrophoresis in 0.7% agarose. (B) Southern blot ofchromosomal DNAs from panel A probed with radiolabeledpAM620 (Tn916). Represented in the first lane is DNA from theUA130 parent strain which lacks the transposon.

mean + the standard error of the mean. Statistical analysisof quantitative electron microscopy was performed by usingthe Student t test. Means and standard errors determined forcaries scores were evaluated by analysis of variance andmultiple mean comparisons by using the Duncan test. Dif-ferences were considered to be significant when P < 0.05was obtained.

RESULTSIsolation of S. mutans UA130 glycogen-deficient mutants by

Tn916 transposon mutagenesis. The streptococcal transposonTn916 was used to isolate mutants of S. mutans UA130altered in glycogen accumulation, as well as to facilitatecloning of the structural gene(s) involved in glycogen bio-synthesis and breakdown. Transposon libraries were gener-ated as described in Materials and Methods, and Tcr trans-formants were obtained for pAM620 with an averagefrequency of 10-6 transformants per recipient cell. Southernblot analysis of HindIII-digested chromosomal DNAs iso-lated from eight randomly selected Tcr transformants, usingradiolabeled pAM620 as a probe, revealed junction frag-ments of various sizes (Fig. 1); indeed, this suggests a

random insertion of Tn916 into the S. mutans genome.Moreover, since HindIII cuts only once within Tn916, thetwo hybridizing fragments resolved for each mutant indicatea single insertion event into the S. mutans chromosome.Nearly 3,000 tetracycline-resistant transposon mutants

were subsequently plated onto Jordan's medium containing2% glucose and screened for alterations in glycogen contentby iodine staining. A single mutant, SMS201, was deficientin its ability to accumulate IPS as indicated by a yellowphenotype upon staining with iodine (Fig. 2). The preciseexcision of Tn9O6 gave rise to glycogen-proficient revertants(as indicated by a brown phenotype) with an average fre-quency of 10-'. All revertants were correlated with the loss

of tetracycline resistance, thus suggesting that the glycogen-deficient phenotype of SMS201 occurred as a direct conse-quence of Tn916 insertion.

Cloning of locus involved in S. mutans glycogen accumula-tion. To isolate the region of Tn916 insertion, the transposon(16.4 kb) and sequences flanking it on the SMS201 chromo-some were cloned as an intact 18.5-kb PstI fragment onto thepCP13-derived cosmid vector, pYA2329 (Fig. 3). Resultantligation mixtures were packaged in vitro, and lysates wereused to transduce E. coli DH5a. Colonies demonstratingresistance to both kanamycin (Kmr) (cosmid encoded) andtetracycline (Tn916 encoded) harbored pYA3000 as con-firmed by minipreparation analysis. Transposon excisantswere obtained subsequently by growing individual culturesof recombinant E. coli (Kmr Tcr) in the absence of tetracy-cline selection. The resultant tetracycline-sensitive excisants(>90%), designated pYA3001, harbored a 2.1-kb PstI frag-ment which represents a chromosomal determinant (glgX)putatively involved in S. mutans glycogen accumulation.To facilitate subsequent genetic analyses, we cloned a

larger region of this locus from the S. mutans UA130chromosome. Specifically, the 2.1-kb PstI fragment frompYA3001 was nick translated and used as a probe to screena recombinant EMBL3 bacteriophage library of S. mutansgenomic DNA as described in Materials and Methods.Plaques which hybridized with this probe were purified asdescribed previously (1). Analysis of the purified phageDNAs, prepared from crude plate lysates with Lambdasorbphage adsorbent (Promega Biotec), resulted in the isolationof a 16-kb SalI fragment which has been subcloned to yieldpYA3053. A restriction map of pYA3053 is shown in Fig. 4.

Complementation of bacterial hosts deficient in glycogenaccumulation. The 16-kb SalI fragment illustrated in Fig. 4,and subclones thereof, were used to complement glycogendefects in E. coli or S. mutans hosts. Complementation toGlg+ was not observed in E. coli hosts X6153 and X6155,which bear mutations in the glgA and glgC genes, respec-tively; gigA and glgC code for enzymes which are requiredfor glycogen biosynthesis in gram-negative microorganisms(31). In contrast, complementation was observed in S.mutans, in which pYA3034, a pVA838-derived subclone(Fig. 4), successfully restored glycogen-accumulating abilityto the glycogen-deficient transposon mutant SMS201. South-ern blot analysis of chromosomal DNA isolated from the

A

INFEcr. IMMUN.

on July 18, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

IPS AND S. MUTANS VIRULENCE 3179

2L1

:1'TniLf

Tc SMS201

Sal Smia

aph

CIAP

T4 DNALigase

FIG. 3. Cloning of the locus involved in S. mutans glycogenaccumulation. Transposon Tn916 and sequences flanking it on theSMS201 chromosome were cloned onto the cosmid pYA2329, andthe resultant recombinant, pYA3000, was transduced into E. coliDH5a. In the absence of selective pressure, precise transposonexcisants were identified (pYA3001) which harbored a 2.1-kb chro-mosomal determinant (glgX) putatively involved in S. mutansglycogen accumulation. ClAP, calf intestinal alkaline phosphatase.

pYA3034-containing mutant confirmed that transposition ofTn916 to another locus on the S. mutans chromosome couldnot account for the observed reversion of SMS201 to glyco-gen proficiency (data not shown). The minimal subfragment

f/amp tetjl E pYA3076 (Pstl) 2 30

POR322Ellel ~~~~~960kb

436 k6r / amp

Pstil/Ec Pstl/EcoRI

T4 DNA ligase /5;l\ ~~~~~~~~Psti

(PStli 0.00

Saill

(Pstl) 230pYA3O771I.S0 kb

P6R322

-\_ SmalSina

FIG. 5. Disruption of the cloned glycogen gene locus from S.mutans. The streptococcal glycogen gene locus (glgX), cloned as a6.9-kb BglII fragment on plasmid pYA3023, was partially digestedwith PstI, and a 7.2-kb fragment lacking glgX was isolated byelectroelution as described in Materials and Methods. This purifiedfragment was treated with T4 DNA polymerase and subsequentlyblunt-end ligated to the SmaI-cut aph determinant from transposonTn1S45. The resultant recombinant, pYA3076, was digested withPstI and EcoRI and then ligated to PstI-EcoRI-cut pBR322 toeliminate ampicillin resistance. The final construct is designatedpYA3077. Brackets denote restriction sites not regenerated uponligation. Poll, polymerase I; ori, origin.

Sa Bg X PH E E H SP H P H 59 SP X E Sa..

I (pYA3034)Xa E E.HS m Bg SP X

ktb

I~~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~

FIG. 4. Restriction map of the 16-kb insert DNA from pYA3053which harbors a putative glycogen gene locus (glgX) from S.mutans. Also represented is pYA3034, a pVA838-derived subclonewhich restored glycogen-accumulating ability to the S. mutansglycogen-deficient transposon mutant, SMS201. Bg, BglII; E,EcoRI; H, HindIII; P, PstI; Sa, SalI; Sp, SphI; X, XbaI.

which complemented the defect in S. mutans was containedon an 8.3-kb XbaI fragment.

Construction of stable glycogen-deficient mutant by genereplacement. Gene replacement is a preferred moleculargenetic approach to mutant construction since it ensuresstrain stability. To generate a stable S. mutans mutantdeficient in glycogen accumulation for subsequent virulencetesting in vivo, we partially digested plasmid pYA3023 (Fig.5) with PstI. A resultant 7.2-kb PstI fragment lacking the2.1-kb glycogen gene locus was electroeluted from agarosegels, treated with T4 DNA polymerase, and blunt-end ligatedto a 2.3-kb SmaI fragment carrying the aph determinant fromtransposon TnJS45 (5) (aph encodes resistance to kanamycinwhich is expressible in both E. coli and S. mutans). Thisconstruction gave rise to plasmid pYA3076 (Fig. 5), whichwas to be used in subsequent transformations of S. mutans.To avoid introducing penicillin resistance into streptococciwhere it does not naturally occur, the 13-lactamase genecontained on plasmid pYA3076 was removed upon furtherdigestion with PstI and EcoRI. Subsequent ligation with aPstI-EcoRI digest of plasmid pBR322 gave rise to plasmidpYA3077. This plasmid was subsequently isolated from E.

-

VOL. 60, 1992

on July 18, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

3180 HARRIS ET AL.

A

p5 ": 59El IB 30! :stl [IPot (Bg 11.10WIHl p

- 'FeP a;>

X C. ,a 5 ;g |\ jt psS /\9a:_

I" El Br-_ 5

..~~~~~~~~~~~~~~~~a,3

III.

I

W! =si! T 1 p-1i1"I

UAI30Gig.

III 1

H11 dlill .AMHI.AI I I I

FIG. 6. Construction of S. mutans SMS203 and confirmation of allelic exchange by Southern blot analysis. (A) PstI-linearized pYA3077 wasintroduced into S. mutans UA130 by electroporation, and glycogen-deficient transformants demonstrating resistance to kanamycin (250 pLg/ml)were selected. The predicted allelic exchange is illustrated. (B) Radiolabeled probes designated I, II, and III were used in Southern hybridizationsof chromosomal DNA isolated from S. mutans SMS203 and UA130. (C) Southern blot analyses confirmed the predicted double-crossoverrecombinational event illustrated in panel A. Indeed, the resultant hybridization pattern shown here is consistent with the exchange of the defectiveglycogen locus on pYA3077 for the wild-type region on the S. mutans chromosome. Lanes: 1, UA130 and PstI; 2, SMS203 and PstI; 3, UA130 andBamHI; 4, SMS203 and BamHI; 5, UA130 and BgllI-XbaI; 6, SMS203 and BglII-XbaI; 7, UA130 and BamHI-BglII; 8, SMS203 and BamHI-BglII.

coli DH5a transformants demonstrating resistance to bothkanamycin and tetracycline, as described in Materials andMethods.

Inactivation of putative glycogen gene locus in S. mutansand confirmation of allelic exchange by Southern blot analysis.The pMo1 replicon of pYA3077 does not function in S.mutans; thus, the Kmr marker contained on the plasmid canbe rescued only by integration into the chromosome. Themechanism of integration involves recombination betweenthe streptococcal sequences which flank the deleted glyco-gen locus on plasmid pYA3077 and the homologous regionon the wild-type S. mutans chromosome (Fig. 6A). Specifi-cally, PstI-linearized pYA3077 was transformed into S.mutans UA130 by electroporation, and Kanr transformantswere analyzed for the expected double-recombination eventby Southern blot analysis. Indeed, the radiolabeled probesillustrated in Fig. 6B revealed hybridization patterns onSouthern blots (Fig. 6C) consistent with the exchange of thedefective glgX locus on pYA3077 for the wild-type region onthe S. mutans chromosome. That is, hybridization to aslightly larger DNA fragment was observed in a BglII-XbaIdigest of chromosomal DNA isolated from the mutant strain,SMS203, relative to the hybridization observed in the wild-type progenitor, UA130; this reflects the replacement of the2.1-kbglgX locus on the UA130 chromosome with the 2.3-kbaph determinant from transposon Tn1545. Importantly, thedeletion mutant, designated SMS203, was also glycogendeficient upon staining with iodine. Moreover, SMS203

remained kanamycin resistant and glycogen deficient formore than 100 generations when passaged in vitro in theabsence of antibiotic selective pressure. Hence, the stabilityof this mutant facilitated subsequent in vivo investigations ofintracellular polysaccharides and their role in the cariogenicprocess.

Biochemical analysis of enzyme activities in S. mutansUA130 and SMS203. Crude extracts prepared from S. mu-

tans UA130 and SMS203 were standardized for proteincontent and assayed for ADP-glucose pyrophosphorylaseand glycogen synthetase activities as described in Materialsand Methods. As shown in Fig. 7, both of these activitieswere significantly repressed in the glycogen-deficient mutantSMS203 relative to the activities observed in the wild-typeprogenitor strain, UA130. The coordinate repression ofthese enzymes in the glycogen-deficient mutant suggests thatglgX, a regulatory region which governs the expression ofboth glgC and gigA, has been disrupted. Preliminary nucle-otide sequence analysis of this regulatory region, now des-ignated glgR, revealed two open reading frames which readin the same direction, each preceded by a putative ribosome-binding site (data not shown).

Acidogenicity of S. mutans UA130 and SMS203 in vitro. Theacidogenicity of the IPS-deficient mutant, SMS203, ap-peared weaker than that of its wild-type progenitor, UA130,especially when grown in the absence of exogenous glucose(Fig. 8). Indeed, the contribution of IPS to S. mutans acidproduction is demonstrated by the drop in pH observed over

B. _2/-

j-.": ,.

L Ikb

n11

dph

SM5203G1g-

I

INFECT. IMMUN.

on July 18, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

IPS AND S. MUTANS VIRULENCE 3181

A.0.8 -

UA 300.7-/

0.6 -r

0.5- /

0 0.41t/8

0-4 t / < ~~~~~SMS2030.2-1- /

F / /4 * UAI30

ou' 4 ^) SMS203

0 10 20 30 40 50 60 70 80 90

Time (min)

el

0.2 [ A ~~~~~~~~UA130

,/0.15+ /

E t ~~~/

4 r /~~~~~~~~~~-[ t 1 t . tSMS203

0.05 4 UA130

SMS203

0 10 20 30 40 50 60 70 80 90

Time (min)

FIG. 7. Biochemical analysis ofglycogen accumulation in S. mutans. Crude protein extracts prepared from S. mutans UA130 and SMS203 werestandardized for protein content and subsequently assayed for ADP-glucose pyrophosphorylase (glgC) (A) or glycogen synthetase (gigl) (B)activities as described in Materials and Methods. Both enzyme activities are repressed in the SMS203 mutant strain relative to the UA130 wild typeas demonstrated by the ATP-linked (for glgC) or ADP-linked (for gIgA) reduction of nitroblue tetrazolium (see insets). OD, optical density.

time for cells starved of the exogenous carbohydrate. pHvalues are higher overall when exogenous glucose is limiting,however, since under such conditions, only intracellularreserves are available for acid production by S. mutans.

Quantitative ultrastructural analysis of S. mutans IPS.Since both UA130 and SMS203 demonstrated similar growthrates in TH broth supplemented with 2% glucose (Fig. 8),cells from each were harvested in parallel at the late-logarithmic (5-h), early-stationary (7-h), and stationary(12-h) phases for quantitative electron microscopy. Thinsections from each time point were pretreated with periodicacid and sodium chlorite to enhance the affinity of bacterialIPS for subsequent staining with uranyl and lead salts (10).Indeed, IPS first appeared as distinct electron-dense bodiesin the cytoplasm of the wild-type strain during the late-logarithmic phase (Fig. 9A). The accumulation of IPS in the

1.0

+ o

I

0.8

0.6

0.4

0.2

0.0

0 2 4 6 8 10 12 14

Time (hrs)

FIG. 8. Growth and acidogenicity of S. mutans UA130 (----)and SMS203 ( ). Cultures grown in TH broth supplemented with2% glucose were monitored for cell density (optical density at 560nm [0, 0]) and pH (A, A) at the time points designated. pHdeterminations were also performed for cultures grown in theabsence of glucose (l, *).

mutant was relatively delayed, however (Fig. 9B), becomingprominent in the cytoplasm for the first time during theearly-stationary and stationary phases of growth (Fig. 9Dand 9F). Moreover, IPS at the stationary phase were presentat significantly reduced concentrations (P < 0.05) in theSMS203 deletion mutant (10.13 + 0.9 granules per 2 cm2 ofcell section) relative to those concentrations observed in thewild-type progenitor strain, UA130 (21.75 + 2.1 granules per2 cm2 of cell section). The average number of granules perunit area was determined by examining approximately 40cell sections which demonstrated distinct trilaminar mem-brane profiles subsequent to staining with uranyl and leadsalts. Chemical quantitation of IPS (10) (data not shown) wasin good agreement with the cytochemical data presentedhere based on IPS granule count.

Virulence of glycogen-deficient mutants in gnotobiotic rats.The ability of S. mutans glycogen-deficient mutants to causecaries in a germfree rat model system was examined. Ratsinfected with the glycogen synthesis-proficient UA130strain, or with glycogen synthesis-deficient SMS203, weresacrificed 45 days postchallenge, and their teeth were scoredfor carious lesions by the method of Keyes (19). The resultsof these experiments are summarized in Table 2. In sum-mary, the mean caries scores for rats infected with theglycogen synthesis-deficient mutant were significantly lower(P < 0.01) than those scores for rats infected with theglycogen synthesis-proficient wild-type strain. This effectwas pronounced on all tooth surfaces examined (i.e., buccal,sulcal, and proximal) and was slightly more obvious inanimals maintained on 6-h-per-day food access protocols.Importantly, isolates from all animals were characterizedboth biochemically and genetically to verify their identitywith the original test strains. The glycogen synthesis-profi-cient phenotype of S. mutans UA130 and the glycogensynthesis-deficient phenotype of S. mutans SMS203 wereconsistent with those of the strains used to challenge the ratsinitially. In addition, Southern blot analysis of chromosomalDNA prepared from several of these isolates revealed hy-bridization patterns identical to those of the original infectingstrain (data not shown). In vivo analysis of SMS201 revealed

VOL. 60, 1992

on July 18, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

B

N1

.

FIG. 9. Electron microscopy of S. mutans glycogen accumulation. Thin sections of S. mutans glycogen-proficient UA130 (A, C, and E)and glycogen synthesis-deficient SMS203 (B, D, and F) were pretreated with periodic acid and sodium chlorite to enhance the affinity of IPSfor uranyl and lead salts. IPS, which appear as electron-dense granules, are significantly more abundant (P < 0.05) in the wild-type strain thanin the glycogen synthesis-deficient deletion mutant for all growth phases examined. (A and B) Late-logarithmic phase; (C and D)early-stationary phase; (E and F) stationary phase (1 cm = 167 nm).

3182

on July 18, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

IPS AND S. MUTANS VIRULENCE 3183

TABLE 2. Virulence of S. mutans UA130 and isogenic mutant SMS203 in monoinfected germfree rats (n = 6)

Caries score (mean ± SE)b

Strain Diet' Buccal Sulcal Proximal

E Ds Dm Dx E Ds Dm Dx E Ds Dm Dx

SMS203 R 9.2 ± 0.5C 7.0 ± 0.5C 4.5 ± 0.7C 2.5 ± 0.7C 14.7 ± 0.9C 11.2 ± 1.1C 2.8 0O.8C 1.5 ± 0.6C 1.5 ± 0.6C 0.2 ± 0.2C 0 0UA130 R 15.3 ± 0.9 12.2 ± 0.9 9.7 + 0.6 6.8 ± 0.5 20.0 ± 0.6 16.0 ± 0.9 8.0 ± 1.3 4.7 ± 1.1 6.0 ± 0.8 1.3 ± 0.6 0 0SMS203 AL 11.2 ± 0.4C 8.3 + 0.5C 6.3 ± 0.4C 3.8 ± 0.7C 17.7 ± 0.6C 12.7 ± 0.7C 3.8 ± 0.3c 2.0 ± 0.4C 3.3 ± 0.4C OC 0 0UA130 AL 16.0 ± 0.5 12.5 ± 0.4 9.7 ± 0.5 7.7 ± 0.4 20.8 ± 0.8 16.0 ± 0.7 9.7 ± 0.6 5.5 ± 0.3 5.7 ± 0.3 1.7 ± 0.3 0 0

a R, diet restricted to 6 h/day; AL, ad libitum.b Caries scores were determined by the method of Keyes (19). Abbreviations: E, enamel involvement; Ds, slight dentinal involvement; Dm, moderate dentinal

involvement; Dx, extensive dentinal involvement.c Significant difference (P < 0.01) between experimental strain score and UA130 score.

that the cariogenic potential of this glycogen synthesis-deficient transposon mutant was also significantly attenuated(P < 0.05) relative to its wild-type S. mutans progenitor,UA130 (data not shown).

DISCUSSION

Recent reports of S. mutans glucosyltransferase (27) andfructosyltransferase (27, 35) mutants confirm the importanceof both glucans and fructans to the cariogenic process. In thepresent study, we determined IPS in S. mutans to be yetanother significant virulence determinant in plaque-mediateddisease. The importance of IPS to S. mutans virulencesupports previous reports in the literature which describe anassociation of these glycogenlike storage polymers with S.mutans-induced caries formation (38). While these earlierstudies employed nitrosoguanidine-generated mutants of S.mutans for analysis in conventional rats, our strategy was toutilize gene disruption and allelic exchange to geneticallyengineer the IPS synthesis-deficient mutant, SMS203, forsubsequent virulence testing in germfree animals. Impor-tantly, no obvious changes in the phenotype of SMS203 werenoted other than the expected growth on kanamycin, thealteration in glycogen-accumulating ability upon iodinestaining, and the diminished acidogenic potential of thismutant in vitro. The growth rates of both the mutant andwild type were comparable in TH broth supplemented with2% glucose; this suggests that glucose transport was notaffected by the deletion in SMS203. The IPS synthesis-deficient strain was also able to adhere to glass at levelssimilar to those of the wild-type strain (data not shown). Thisis further supported by the efficiency of colonization ob-served in vivo; that is, the recovery of SMS203 from ratmolar surfaces was comparable to that of the wild-type strain(data not shown). Consequently, SMS203 represents a suit-able mutant for examining the specific contribution of IPS tooral disease.

In monoinfected germfree rats maintained on a caries-promoting diet (26), SMS203 proved to be significantly lesscariogenic (P < 0.01) than the IPS synthesis-proficientwild-type strain, UA130, on the buccal, sulcal, and proximalmolar surfaces of teeth. These differences were slightly morepronounced for animals maintained on the 6-h-per-day foodaccess protocol. Taken collectively, these observations sup-port the hypothesis that IPS provide S. mutans with anendogenous source of carbohydrate which can be metabo-lized when exogenous fermentable substrates have beendepleted from the oral cavity; as a result, a prolongedduration of acid production in the oral cavity and a concom-itant lower resting pH in the plaque microflora foster the

demineralization of tooth enamel and the onset of dentaldecay. Thus, we ascribe the failure of SMS203 to causedecay as proficiently as its wild-type progenitor to a signifi-cantly reduced availability of IPS in the former. Indeed, thedecreased acid production demonstrated in vitro by SMS203relative to the wild-type strain supports this hypothesis. Insummary, IPS are significant contributors to S. mutansvirulence.While the mechanisms which regulate the accumulation of

IPS in S. mutans remain totally unexplored, gene expressionin gram-positive bacteria, and specifically in the oral strep-tococci, is often thought to be influenced by mechanisms ofcatabolite repression (22). For instance, nucleotide se-quences in the promoter-operator region of the amyE gene inBacillus subtilis have been shown to be essential for thecatabolite repression of this gene (28). Moreover, expressionof the S. mutans fructanase gene has been shown to beinducible by exogenous fructose and repressible by glucose,with maximal expression at low growth rates (4). It may bethat the streptococcal glycogen genes are similarly regulatedby such substrate availability, catabolite repression, and/orbacterial growth rate. Indeed, the glycogenlike polysaccha-ride granules observed in thin sections of S. mutans UA130appear to accumulate during the late-logarithmic and early-stationary phases of growth, when nutrients have beendepleted from the environment and pH conditions havebecome unfavorable. This is consistent with glycogen accu-mulation in E. coli and Salmonella typhimunum, whichappears to be induced under similar conditions of nutrientdeprivation (32).Biochemical determinations from crude extracts of UA130

and SMS203 indicate that the mutation at the S. mutansglgRlocus represses the activities of ADP-glucose pyrophosphor-ylase (glgC) and glycogen synthetase (glgA), both enzymesinvolved in the synthesis of bacterial glycogen (2, 31).Importantly, glycogen accumulation via other metabolicpathways appears to be of little or no importance in S.mutans (2). The significantly reduced glgC and gigA activi-ties observed in the SMS203 mutant as a consequence ofglgR inactivation suggest that the glgR gene product acts asa positive regulator of S. mutans glycogen accumulation.Note, however, that while GlgR may enhance glycogen geneexpression in S. mutans, it is clearly not essential for thebiosynthesis of IPS.As determined by nucleotide sequence analysis, GlgR is

the putative gene product of an open reading frame whichshares no significant homology with other sequences in theGenBank data base. In contrast, a second open readingframe located immediately downstream shares some nucle-otide similarity with the glgB gene from Bacillus stearother-

VOL. 60, 1992

on July 18, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

3184 HARRIS ET AL.

mophilus (20). Importantly, glgB, which encodes glycogenbranching activity, has never been described in the strepto-cocci. Indeed, Southern blot analyses with glgB gene probesfrom Synechococcus sp. (21) and B. stearothermophilus (20)confirm the presence of a glgB homolog in S. mutans. Thecontinued sequence analysis of this glycogen gene locus in S.mutans should reveal more of the structural organization ofthese and possibly other IPS-related genes on the chromo-some. Moreover, the cloning and characterization of glgR,glgB, and other genes involved in S. mutans glycogenbiosynthesis could provide further insight into the molecularmechanisms by which this oral pathogen regulates glycogenaccumulation in the plaque environment. Overall, these andother studies should extend our currently limited knowledgeof the molecular genetic basis for gene regulation in gram-positive microorganisms.

ACKNOWLEDGMENTS

We appreciate the assistance of Michael Veith in the preparationof S. mutans thin sections for electron microscopic analyses. Wealso thank J. A. K. W. Kiel for providing the glgB probes. SusanHollingshead, Michael Hudson, John Mauer, and Steven Tinge areacknowledged for their critical reviews of the manuscript. We thankSteven Tinge for his assistance with figure preparation and CecilyHarmon and Gloria Richardson for their participation in animalexperimentation.

This work was supported by Public Health Service grants F32DE05999 to G.S.H., DE09081 and DE08182 to S.M.M., andDE06673 to R.C. III from the National Institute of Dental Research.

REFERENCES1. Benton, W. D., and R. W. Davis. 1977. Screening lambda gtll

recombinant clones by hybridization to single plaques in situ.Science 196:180-181.

2. Birkhed, D., and J. M. Tanzer. 1979. Glycogen synthesispathway in Streptococcus mutans strain NCTC 10449S and itsglycogen synthesis-defective mutant 805. Arch. Oral Biol. 24:67-73.

3. Birnboim, H. C., and J. Doly. 1979. A rapid alkaline extractionprocedure for screening recombinant plasmid DNA. NucleicAcids Res. 7:1513-1523.

4. Burne, R. A., K. Schilling, W. H. Bowen, and R. E. Yasbin.1987. Expression, purification, and characterization of an exo-P-D-fructosidase of Streptococcus mutans. J. Bacteriol. 169:4507-4517.

5. Caillaud, F., P. Trieu-Cuot, C. Carlier, and P. Courralin. 1987.Nucleotide sequence of the kanamycin-resistance determinantof the pneumococcal transposon TnIS45: evolutionary relation-ships and transcriptional analysis of aphA-3 genes. Mol. Gen.Genet. 207:509-513.

6. Charlton, G., R. J. Fitzgerald, and P. H. Keyes. 1971. Determi-nation of saliva and dental plaque pH in hamsters with glassmicroelectrodes. Arch. Oral Biol. 16:649-654.

7. Charlton, G., R. J. Fitzgerald, and P. H. Keyes. 1971. Hydrogenion activity in dental plaque of hamsters during metabolism ofsucrose, glucose and fructose. Arch. Oral Biol. 16:655-661.

8. Clarke, L., and J. Carbon. 1976. A colony bank containingsynthetic ColEl hybrid plasmids representative of the entire E.coli genome. Cell 9:91-99.

9. Darzins, A., and A. M. Chakrabarty. 1984. Cloning of genescontrolling alginate biosynthesis from a mucoid cystic fibrosisisolate of Pseudomonas aeruginosa. J. Bacteriol. 159:9-18.

10. DiPersio, J. R., S. J. Mattingly, M. L. Higgins, and G. D.Shockman. 1974. Measurement of intracellular iodophilicpolysaccharide in two cariogenic strains of Streptococcus mu-tans by cytochemical and chemical methods. Infect. Immun.10:597-604.

11. Dower, W. J., J. F. Miller, and C. W. Ragsdale. 1988. Highefficiency transformation of E. coli by high voltage electropora-tion. Nucleic Acids Res. 16:6127-6145.

12. Fox, J., S. Govons, and J. Preiss. 1972. Glycogen synthetasefrom Eschenichia coli B. Methods Enzymol. 28:539-545.

13. Freedman, M. L., J. M. Tanzer, and R. L. Eifert. 1976. Isolationand characterization of mutants of Streptococcus mutans withdefects related to intracellular polysaccharide, p. 583-596. InH. M. Stiles, W. J. Loesche, and T. L. O'Brien (ed.), Proceed-ings in microbiology. Aspects of dental caries. Special supple-ment to Microbiology Abstracts, vol. 3. Information Retrieval,Inc., London.

14. Frischauf, A. M., H. Lehrach, A. Poustka, and N. Murray. 1983.Lambda replacement vectors carrying polylinker sequences. J.Mol. Biol. 170:827-842.

15. Hamilton, I. R. 1976. Intracellular polysaccharide synthesis bycariogenic microorganisms, p. 683-701. In H. M. Stiles, W. J.Loesche, and T. L. O'Brien (ed.), Proceedings in microbiology.Aspects of dental caries. Special supplement to MicrobiologyAbstracts, vol. 3. Information Retrieval, Inc., London.

16. Hanahan, D. 1983. Studies on transformation of Eschenichia coliwith plasmids. J. Mol. Biol. 166:557-580.

17. Jordan, H. V., R. J. Fitzgerald, and A. E. Bowler. 1960.Inhibition of experimental caries by sodium metabisulfite and itseffect on the growth and metabolism of selected bacteria. J.Dent. Res. 39:116-123.

18. Kellenberger, E., A. Ryter, and J. Sechaud. 1958. Electronmicroscope study of DNA-containing plasms. II. Vegetativeand mature phage DNA as compared with normal bacterialnucleoids in different physiological states. J. Biophys. Biochem.Cytol. 4:671-676.

19. Keyes, P. 1958. Dental caries in the molar teeth of rats. lI. Amethod for diagnosing and scoring several types of lesionssimultaneously. J. Dent. Res. 37:1088-1099.

20. Kiel, J. A. K. W., J. M. Boels, G. Beldman, and G. Venema.1991. Molecular cloning and nucleotide sequence of the glyco-gen branching enzyme gene (glgB) from Bacillus stearothenno-philus and expression in Escherichia coli and Bacillus subtilis.Mol. Gen. Genet. 230:136-144.

21. Kiel, J. A. K. W., H. S. A. Elgersma, G. Beldman, J. P. M. J.Vossen, and G. Venema. 1989. Cloning and expression of thebranching enzyme (glgB) from the cyanobacterium Synecho-coccus sp. PCC7942 in Escherichia coli. Gene 78:9-17.

22. Lane, M. A., K. W. Bayles, and R. E. Yasbin. 1991. Identifica-tion and initial characterization of glucose-repressible promot-ers of S. mutans. Gene 100:225-229.

23. Loesche, W. J. 1986. Role of Streptococcus mutans in humandental decay. Microbiol. Rev. 50:353-380.

24. Macrina, F. L., J. A. Tobian, K. R. Jones, R. P. Evans, and D. B.Clewell. 1982. A cloning vector able to replicate in Escherichiacoli and Streptococcus sanguis. Gene 19:345-353.

25. Marmur, J. 1961. A procedure for the isolation of deoxyribo-nucleic acid from microorganisms. J. Mol. Biol. 3:208-218.

26. Michalek, S. M., J. R. McGhee, and J. M. Navia. 1975. Viru-lence of Streptococcus mutans: a sensitive method for evaluat-ing cariogenicity in young gnotobiotic rats. Infect. Immun.12:69-75.

27. Munro, C., S. M. Michalek, and F. L. Macrina. 1991. Carioge-nicity of Streptococcus mutans V403 glucosyltransferase andfructosyltransferase mutants constructed by allelic exchange.Infect. Immun. 59:2316-2323.

28. Nicholson, W. L., Y. K. Park, T. M. Henkin, M. Won, M. J.Weickert, J. A. Gaskell, and G. H. Chambliss. 1987. Cataboliterepression-resistant mutations of the Bacillus subtilis alpha-amylase promoter affect transcription levels and are in anoperator-like sequence. J. Mol. Biol. 198:609-618.

29. Orland, F. J., J. R. Blayney, R. W. Harrison, J. A. Reyniers,P. C. Trexler, M. Wagner, H. A. Gordan, and T. D. Luckey.1954. Use of the germfree animal technique in the study ofexperimental dental caries. I. Basic observation on rats rearedfree of all microorganisms. J. Dent. Res. 33:147-174.

30. Ozaki, H., and J. Preiss. 1972. ADP-glucose pyrophosphorylasefrom Escherichia coli B. Methods Enzymol. 28:406-413.

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

32. Preiss, J., and D. A. Walsh. 1981. The comparative biochemistry

INFECT. IMMUN.

on July 18, 2018 by guesthttp://iai.asm

.org/D

ownloaded from

IPS AND S. MUTANS VIRULENCE 3185

of glycogen and starch, p. 199-314. In V. Ginsburg, and P.Robbins (ed.), Biology of carbohydrates. John Wiley & Sons,Inc., New York.

33. Rigby, P. W. J., M. Dieckman, C. Rhodes, and P. Berg. 1977.Labeling deoxyribonucleic acid to high specific activity in vitroby nick translation with DNA polymerase I. J. Mol. Biol.113:237-251.

34. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecularcloning: a laboratory manual, 2nd ed. Cold Spring HarborLaboratory, Cold Spring Harbor, N.Y.

35. Schroeder, V. A., S. M. Michalek, and F. L. Macrina. 1989.Biochemical characterization and evaluation of virulence of afructosyltransferase-deficient mutant of Streptococcus mutansV403. Infect. Immun. 57:3560-3569.

36. Senghas, E., J. M. Jones, M. Yamamoto, C. Gawron-Burke, andD. B. Clewell. 1988. Genetic organization of the bacterial

conjugative transposon Tn916. J. Bacteriol. 170:245-249.37. Southern, E. M. 1975. Detection of specific sequences among

DNA fragments separated by gel electrophoresis. J. Mol. Biol.98:503-517.

38. Tanzer, J. M., M. L. Freedman, F. N. Woodiel, R. L. Eifert, andL. A. Rinehimer. 1976. Association of Streptococcus mutansvirulence with synthesis of intracellular polysaccharide, p.

597-616. In H. M. Stiles, W. J. Loesche, and T. L. O'Brien(ed.), Proceedings in microbiology. Aspects of dental caries.Special supplement to Microbiology Abstracts, vol. 3. Informa-tion Retrieval, Inc., London.

39. Yamamoto, M., J. M. Jones, E. Senghas, C. Gawron-Burke, andD. B. Clewell. 1987. Generation of Tn5 insertions in streptococ-cal conjugative transposon Tn916. Appl. Environ. Microbiol.53:1069-1072.

VOL. 60, 1992

on July 18, 2018 by guesthttp://iai.asm

.org/D

ownloaded from