Vol. 57, No. 11

Cloning and Expression in Escherichia coli of the Genes of theArginine Deiminase System of Streptococcus sanguis NCTC 10904

ROBERT A. BURNE,* DAWN T. PARSONS, AND ROBERT E. MARQUISDepartments of Microbiology and Immunology and ofDental Research, University ofRochester School of

Medicine and Dentistry, Rochester, New York 14642

Received 30 May 1989/Accepted 11 August 1989

The common oral bacterium Streptococccus sanguis can degrade arginine via the arginine deiminase (AD)system. The three enzymes of this system, AD, ornithine carbamyltransferase (OTC), and carbamate kinase(CK), catalyze the breakdown of arginine to ornithine, C02, and two molecules of ammonia, with theproduction of ATP from ADP. The genes of the AD system, which are subject to complex regulation in the oralstreptococci, have been isolated in bacteriophage lambda by screening for AD activity. The AD gene,

designated arcA, was expressed from recombinant bacteriophage or in cells harboring plasmid subclones fromthis phage at a level up to 1,000-fold lower than the level in fully derepressed S. sanguis but apparently underthe control of its own promoter. By subcloning in Escherichia coli mutants defective in anabolic OTC (argFargL) and CK (carB), it was demonstrated that the genes for S. sanguis OTC and CK were located adjacent tothe AD gene. The levels of expression of the OTC and CK genes (arcB and arcC, respectively) were also very

low in E. coli, although arcC expression was not as poor as arcA and arcB expression when compared with thelevels found in S. sanguis. Also, arcB and arcC were unable to complement the defects in their anaboliccounterparts. Introduction of the entire AD system or subclones which encoded only the AD gene into E. coliharboring defects in arginine and pyrimidine biosynthesis resulted in a 10- to 15-fold decrease in the level ofAD activity, suggesting that arginine or its metabolites may regulate AD expression. Transposon mutagenesiswas utilized to construct defined mutants of S. sanguis with mutations in the AD gene cluster. AD gene

expression in these mutants indicated that the expression of the AD genes in this organism is stronglyinterrelated. The isolation and partial characterization of the arc genes represents the first step in the geneticmanipulation of the AD system in the oral streptococci for analysis of the regulation of AD, analysis of the roleof the system in plaque ecology, and utilization of the system to modulate the cariogenicity of dental plaque.

A variety of bacteria, including members of the genera

Streptococcus, Bacillus, Pseudomonas, Clostridium, Lacto-bacillus, Treponema, and Mycoplasma, are able to utilizearginine, catabolically, via the arginine deiminase (AD)system (1, 13). Many of these organisms can utilize arginineas a sole source of energy for growth (1, 13, 42, 47). The ADpathway involves three main enzymes. AD catalyzes thehydrolysis of arginine to citrulline and ammonia. Then,catabolic ornithine carbamyltransferase (ornithine transcar-bamylase; OTC) catalyzes the reaction of citrulline withinorganic phosphate to produce ornithine and the high-energy compound carbamyl phosphate. The final step in thepathway is catalyzed by catabolic carbamate kinase (CK),which can transfer the phosphate group of carbamyl phos-phate to ADP to produce ATP, with the concomitant pro-duction of ammonia and CO2. Anabolic variants of OTC andCK are widely distributed among procaryotes and eucary-otes, in which they function in arginine and pyrimidinebiosyntheses (13, 18).The AD system plays important roles in cell physiology

and is subject to fairly complex genetic and biochemicalregulation. For example, in Pseudomonas aeruginosa, ca-tabolism of arginine facilitates fermentative growth on argi-nine, with ATP being derived from the reaction catalyzed byCK (1, 13, 47). Genetic regulation of the AD system byoxygen, arginine, glucose, and cellular energy levels hasbeen observed in a variety of organisms, such as P. aerug-inosa, Streptococcus faecalis, Bacillus licheniformis, andsome lactic streptococci (1, 13, 35, 39, 43, 47). The genes for

* Corresponding author.

the P. aeruginosa AD system have been isolated and shownto be tightly linked in operon form (30). Additionally, the ADpathway enzymes isolated from these and other organismsdemonstrate various complex patterns of allosteric regula-tion (1).The AD system of Streptococcus sanguis, a common oral

bacterium isolated in high numbers from the tooth surfaceand saliva of humans (20), is tightly regulated at the geneticlevel. The genes of the system are repressed in the presenceof glucose and are inducible by arginine (17). S. sanguis alsoproduces an arginine-ornithine antiporter (38) which is sim-ilar to that described for Streptococcus lactis (15) and whichexchanges the substrate of the pathway, arginine, for one ofthe end products, ornithine, with no energy expenditure.This catabolic arginine transport system is coordinatelyregulated with the AD system (38). Arginine-specific ami-nopeptidases which may function to liberate arginine fromsalivary and bacterial proteins in plaque (11, 41) have beenidentified in the oral streptococci (22). Recently, such a

peptidase was isolated from S. sanguis and partially charac-terized in this laboratory. S. sanguis is inherently less acidtolerant than other dental plaque organisms, such as Strep-tococcus mutans and Lactobacillus casei (4), and can beprotected against lethal acidification by catabolism of argin-ine via the AD pathway (8, 34). Protection probably occurs

through the production of ammonia and the associated rise inthe environmental pH (34). This protection may be critical tothe survival of bacteria such as S. sanguis in dental plaque,in which pH values can drop below 4.0 and in which cyclesof acidification-alkalinization occur frequently.Our principal interest in the AD system of S. sanguis

3540

INFECTION AND IMMUNITY, Nov. 1989, p. 3540-35480019-9567/89/113540-09$02.00/0Copyright © 1989, American Society for Microbiology

on May 22, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

S. SANGUIS AD GENES 3541

stems from an intention to manipulate the AD systemgenetically in plaque bacteria so that the resulting organismswill be able to colonize the teeth and modulate the carioge-nicity of dental plaque by catabolizing arginine. For exam-ple, with the isolation of the genes from S. sanguis, it shouldbe possible to introduce the AD system into the cariogenicbacterium S. mutans (20, 29), which lacks the AD system.These recombinant organisms should have a selective ad-vantage because of their ability to produce one molecule ofATP per molecule of arginine catabolized. As an alternativestrategy, the AD genes could be altered in vitro to be morehighly expressed or nonrepressible by glucose and thenreturned to S. sanguis. The ammonia-producing streptococcinot only should be less cariogenic themselves but also couldreduce the overall cariogenic potential of dental plaque. Thegeneral approach of using replacement therapy in the pre-vention of dental caries has been explored by others withStreptococcus salivarius (45) and lactate dehydrogenase-deficient S. mutans (21) and has the advantage that theseorganism may be effective when cariostatic agents, such asfluoride, are only moderately effective, for example, in pitand fissure caries or caries in interproximal surfaces (20, 29).The AD system appears to be a major contributor to the

rise in plaque pH that normally occurs after glycolyticacidification. In fact, the pH of plaque normally rises abovethat of the saliva which bathes it (2, 25). The abilities of S.sanguis and other plaque bacteria to degrade salivary pro-teins and peptides were clearly demonstrated (11, 25, 41),and the arginine liberated from salivary proteins was de-graded, primarily via the AD pathway (25). Studies havedemonstrated that arginine and arginine-containing com-pounds are far more effective than other amino acids orpeptides at inhibiting the pH drop following a sugar chal-lenge and in eliciting a subsequent pH rise (25, 49, 50).Sialin, an arginine-containing tetrapeptide found in saliva,was particularly effective (25), even at concentrations as lowas 40 ,uM. Fractionation of salivary proteins indicates thatalthough the amount of arginine in free or bound form insaliva is low, there is a high proportion of arginine in thelow-molecular-mass fractions (<10 kilodaltons) which pre-sumably can readily diffuse into plaque and be metabolizedby S. sanguis (11, 25). Well-characterized peptides such asthe histatins, which have a low molecular mass (3 to 5kilodaltons), represent a major proportion of total salivarysecretions. They bind with high affinity to the tooth surfaceand are composed of approximately 10 to 12.5 mol% arginine(37). Thus, ammonia production from arginine by oral strep-tococci, such as S. sanguis and Streptococcus mitior, couldplay a direct, ameliorative role in the initiation and progres-sion of dental caries.

Studies focusing on differences in the plaques of caries-resistant and caries-prone subjects (2, 11, 25) indicated thatthere were few differences in the abilities of plaques fromthese individuals to effect a pH drop when challenged with afermentable carbohydrate but that there were significantdifferences in the environmental pH rise following acidifica-tion (11, 25). More recently (A. Naini and I. D. Mandel, J.Dent. Res. 68:318, 1989), significant differences were foundin the abilities of plaques from caries-resistant and caries-prone individuals to produce ammonia and effect a pH risewhen given arginine in vitro. These studies suggest that oneof the key factors in inhibiting dental caries may be afunction of the capacity for dental plaque to produce ammo-nia from arginine in salivary peptides.

Despite its probable importance in oral ecology and dis-ease processes, little is known of the genetic regulation,

TABLE 1. Bacterial strains, bacteriophages, and plasmids

Bacterial strain Genotype or relevant phenotypeand reference or source

StrainsE. coliJM83 ................. ara lac pro rpsL thi 480d lacZ AM15

(48)DH5S ................. [F- endAl hsdRJ7 (r- m+)] supE44

thi-J recAl gyrA96 relAl A(argF-lacZYA)U169 +80d lacZ AM15(Bethesda Research Laboratories)

N134 .................. A(gpt-lac)5 relAl spoTI thi-l argI68;OTC deficient (N. Glansdorf; CGSCa)

JEF8 ................. thr-31 AcarB8 relA metBI; CK deficient(N. Glansdorf; CGSC)

LE392 ................. F- hsdR514 (rK MK+) supE44 supF58.A(lacIZY)6 ga/K2 gaIT22 metBItrpR55 (46)

NM538 ................. Permissive host for X-EMBL3nonrecombinants (Promega)

NM539 ................. Restrictive host for X-EMBL3nonrecombinants (Promega)

730B ................. F- supE44 thi-J leuB6 lacYl tonA121;mini-Mu Mu dE strain (26)

M8820 Mu cts ..........F- araD139 A(ara-/eu)7697 A(proAB-argF-/acIPOZYA)XIII rpsL; Mu cts(9) transductant recipient

S. sanguis NCTC 10904 ... AD positive

Bacteriophagesx-EMBL3 ................. Lambda cloning host (Promega)x-AD1 ................. AD-producing recombinant phage (this

study)

Plasmid vectorspUC19 ................. 48pMK4................. 44pVA891 ................. 31

a CGSC, B. Bachmann, E. coli Genetic Stock Center, Yale University,New Haven, Conn.

biochemistry, and physiology of the AD system in oralstreptococci. As described in this report, the genes encodingthe enzymes of the AD system have been isolated bymolecular cloning, and defined mutants of the AD systemhave been constructed. The cloning, organization, and pe-culiarities of expression in Escherichia coli of the genes forthe S. sanguis AD pathway are described, and the potentialutility of the manipulation of the AD system for cariesprevention is discussed.

MATERIALS AND METHODS

Bacterial strains, bacteriophage, plasmids, and growth me-dium. The bacterial strains and plasmids used in this studyare listed in Table 1. E. coli strains were maintained on L(36) or MacConkey medium supplemented, when necessary,with 50 ,ug of ampicillin, 50 ,ug of kanamycin, or 10 ,ug ofchloramphenicol per ml. Liquid cultures were grown in Lbroth or in TB medium, a rich medium consisting of 12 g oftryptone, 24 g of yeast extract, and 4 ml of glycerol in 1 literof 0.1 M potassium phosphate buffer (pH 7.0). Ampicillin orchloramphenicol was added to 50 or 10 ,ug/ml, respectively.Manipulations of bacteriophage lambda were done as de-scribed by Maniatis et al. (33). For identification of AD-

VOL. 57, 1989

on May 22, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

3542 BURNE ET AL.

producing clones in the lambda library, phage were plaquedon the appropriate recipient on YT (33) medium supple-mented with 1.0% arginine, 0.2% glucose, and 0.1% 2,3,5-triphenyltetrazolium chloride. The plates were incubatedanaerobically for 48 h in a GasPak jar (BBL MicrobiologySystems, Cockeysville, Md.). S. sanguis strains were main-tained on brain heart infusion (BHI) agar and grown in BHIbroth supplemented, when necessary, with 5 ,ug of erythro-mycin per ml. For assessment of AD system enzyme levels,S. sanguis was grown under derepressing conditions in BHIbroth plus 0.8% galactose and 1.0% arginine. All mediumcomponents were from Difco Laboratories, Detroit, Mich.,and supplements were purchased from Sigma Chemical Co.,St. Louis, Mo.DNA manipulations and genetic techniques. Chromosomal

DNA was prepared from S. sanguis NCTC 10904 as previ-ously described (6, 10). Partial Sau3AI fragments of S.sanguis DNA, enriched for 12- to 17-kilobase-pair (kbp)fragments by sucrose gradient centrifugation, were clonedinto the BamHI sites of the left and right arms of X-EMBL3,the DNA was packaged in vitro with a kit (Promega Biotec,Madison, Wis.), and appropriate E. coli strains were infectedas outlined by the supplier.CsCl-banded plasmid DNA was prepared as previously

described (19). Recombinant plasmids were screened in E.coli DH5a by the protocol of Holmes and Quigley (23).Plasmid DNA prepared by this method was digested withrestriction enzymes without further purification. Mappingand subcloning experiments were done with restriction en-zymes and T4 DNA ligase (Bethesda Research Laboratories,Inc., Gaithersburg, Md.) under conditions recommended bythe supplier. E. coli was transformed by the CaCl2 method(32), and transformants of S. sanguis were obtained essen-tially as described by Lawson and Gooder (27).

Transpositional mutagenesis of cloned genes was accom-plished as described by Kuramitsu (26). Mu dE is a mini-Muderivative (9) of bacteriophage Mu which contains a Kmrdeterminant for use in E. coli and an Emr determinant thatfunctions in the oral streptococci. Briefly, plasmids (encod-ing Cm9 harboring known restriction fragments were intro-duced by transformation into E. coli 730B, which harbors thethermally inducible Mu dE. Cultures were grown to themid-exponential phase at 30°C, and phage induction andconcomitant transposition were achieved by shifting thetemperature to 42°C for 5 min, followed by incubation at37°C for 2 h. The resulting phage lysate was treated withCHC13 and used to infect the recipient, E. coli M8820 (9).Transductants harboring plasmids which contained transpo-son insertions were selected on medium supplemented withkanamycin and chloramphenicol. Plasmid DNA was pre-pared from the transductants by the rapid boiling method(23), and the insertion site of the transposon was mappedwith BamHI or HindIII. Plasmids thus prepared were linear-ized with an appropriate restriction enzyme and used totransform competent S. sanguis NCTC 10904, and transfor-mants were selected on BHI plates with 10 ,ug of erythro-mycin per ml.Enzyme preparation and assays. Phage lysates for assay of

AD system enzymes were prepared by infecting E. coliLE392 with X EMBL3 or recombinant phage in the mid-exponential phase of growth at a multiplicity of infection of5. For measurement of enzyme levels in E. coli harboringrecombinant plasmids, cells were sonicated and centrifugedat 37,000 x g for 30 min at 4°C (6), and the supernatantfraction was used for assays. AD system enzyme levels in S.sanguis were measured in permeabilized cells (8). Hexanoic

acid was added to preparations at a final concentration of 10mM. Samples were used directly or were stored at -20°Cwith no loss of activity for >1 month.Enzyme assays were performed by standard methods. For

the AD reaction, mixtures (500 ,ul) contained 25 mM argin-ine, 10 mM hexanoic acid, 50 mM Tris-maleate buffer (pH6.0), and the protein or cell sample. Reactions were termi-nated by the addition of an equal volume of 10% trichloro-acetic acid solution, and samples were centrifuged at 12,700x g for 1 min. The citrulline produced was measured in areaction mixture containing 500 ,ul of trichloroacetic acid-precipitated sample, 700 pl of phosphoric acid-sulfuric acid(3:1), and 250 [lI of diacetyl monoxime (3, 17). This reactionmixture was boiled for 15 to 30 min. Citrulline standardswere treated in the same manner. OTC was assayed in thedirection of citrulline production (35, 43). CK was assayedby measuring the production ofATP as previously described(43), with the following modification. After incubation, 1volume of 10% trichloroacetic acid was added, and thereaction mixture was centrifuged at 12,700 x g for 1 min.ATP was measured in a reaction mixture containing 200 mMTris hydrochloride (pH 7.4), 0.37 mM NADP, 30 mMglucose, 8 mM MgCl2, 15 U of glucose-6-phosphate dehy-drogenase per ml, and 6 U of hexokinase per ml. ATPstandards were treated identically to the samples. ATPmeasurements were linear from 5 to 150 nmol. All incuba-tions were done at 37°C. Protein concentrations were deter-mined with a protein assay kit from Bio-Rad Laboratories,Richmond, Calif., as outlined by the supplier. Dry weightswere determined with cells which were washed twice indistilled H20 and dried to a constant weight in tared alumi-num pans.

RESULTS

Cloning and expression of the AD gene in E. coli. Thegenetic library of S. sanguis NCTC 10904 constructed asdescribed in Materials and Methods was amplified in E. coliNM539, and lysates were used to infect E. coli LE392 forplating on L medium supplemented with arginine, glucose,and 2,3,5-triphenyltetrazolium chloride. The screen forclones containing the AD gene(s) relied on the production ofa functional AD enzyme. Under anaerobic conditions, E.coli fermented glucose to produce acid and the 2,3,5-triphe-nyltetrazolium chloride in the medium remained colorlessand soluble. When the AD enzyme produced by a recombi-nant phage catalyzed the hydrolysis of arginine to produceammonia, the environment became more alkaline, so thatthe 2,3,5-triphenyltetrazolium chloride formed a red, insol-uble precipitate. Therefore, phage plaques possessing afunctional AD gene appeared red. In this manner, fourclones were isolated and subsequently shown to encode afunctional AD gene product.

Following plaque purification, recombinant and controlphage were used to infect LE392. The resulting lysates wereassayed for AD activity. One recombinant phage, X-AD1,was selected for further study because cells infected with itproduced the highest levels of the AD enzyme. Cells whichwere infected with the X-AD1 recombinant phage produced afunctional AD enzyme, while there was no detectable ADactivity in cells infected with the vector, X-EMBL3 (Table 2)E. coli produces repressible, anabolic OTC and CK enzymeswhich are involved in arginine and pyrimidine biosyntheses(13, 18). Cells infected with X-EMBL3 produced OTC andCK at levels of 0.250 and 37.32 U/mg of protein, respec-tively. Cells infected with X-AD1 produced higher levels of

INFECT. IMMUN.

on May 22, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

S. SANGUIS AD GENES 3543

TABLE 2. Expression levels of AD system enzymesin phage lysatesa

Level of:Bacteriophage

ADb OTCc CKd

X-EMBL3 0.000 0.250 37.32X-AD1 0.130 0.441 86.04

a Bacteriophage lysates were prepared by infecting mid-exponential-phasecultures of E. coli LE392 with the appropriate bacteriophage at a multiplicityof infection of 5. Hexanoic acid was added to the resulting lysates to 10 mM.Lysates were assayed directly, without storage.

b AD levels in lysates are expressed in units per milligram of protein, whereone unit is the amount of enzyme required to liberate one micromole ofcitrulline per hour.

c OTC activity was assessed in terms of citrulline synthesis. One unit isdefined as the amount of enzyme required to synthesize one micromole ofcitrulline per hour from ornithine and carbamyl phosphate.

d CK activity was measured in bacteriophage lysates by measuring ATPsynthesis from ADP and carbamyl phosphate. One unit is defined as theamount of enzyme required to produce one micromole of ATP per hour.

both OTC and CK, suggesting that the genes for the threeenzymes of the AD system had been isolated on a singlefragment. However, some compensatory increase in endog-enous OTC and CK due to the depletion of arginine by ADcould not be excluded.

Characterization of expression of the AD genes. Restrictionmapping indicated that X-AD1 contained a 13-kbp insert,which was subsequently transferred to pUC19 as a singleSall fragment with the Sall sites present in the X-EMBL3cloning vector. The resulting construction (Fig. 1) wasdesignated pMAD1. pMAD1 was used to transform E. coliJM83, and transformants were used to prepare 37,000 x gsoluble fractions. The AD gene was found to be expressedby E. coli (Table 3). Similarly, the OTC and CK genes wereexpressed in E. coli JM83(pMAD1) at levels significantlyhigher than in the control [E. coli JM83(pUC19)]. The level

' 5 HEBg H

FIG. 1. Recombinant plasmid pMAD1 containing a 13-kbp frag-ment of S. sanguis NCTC 10904 DNA inserted into the unique Sallsite of pUC19. The thick line represents pUC19 DNA, and the thinline represents S. sanguis DNA. A, AvaI; B, BamHI; Bg, BgIIl; E,EcoRI; H, HinclI; Hd, HindIII; P, PstI; S, SaIl.

TABLE 3. Expression levels of AD system enzymesin 37,000 x g soluble fractions of E. colia

Level of:E. coli strain

AD OTC CK

Wild typeJM83(pUC19) 0.000 0.024 5.46JM83(pMAD1) 0.123 0.104 21.90

argF and argIN134(pUC19) 0.000 0.000 1.32N134(pMAD1) 0.010 0.011 1.44

carBJEF8(pUC19) 0.000 0.012 0.00JEF8(pMAD1) 0.008 0.014 1.44a 37,000 x g soluble fractions were prepared from stationary-phase cultures

of the appropriate bacterial strain by lysozyme treatment, sonication, andcentrifugation at 37,000 x g for 30 min. The resulting supernatant fraction wasused for enzyme assays. Enzyme levels are expressed as in Table 2.

of expression of the AD gene in E. coli was found to be up to1,000-fold lower than that in fully derepressed S. sanguiswhen permeabilized cells were used for assays (data notshown). The low levels of gene expression in E. coli weresurprising because E. coli has been found to express strep-tococcal genes on multicopy plasmids particularly well (5, 7,14).The E. coli OTC and CK genes should have been re-

pressed by arginine and/or uracil in the rich medium (18).The observed increase in the levels of OTC and CK in strainsharboring pMAD1 (Table 3) could possibly have been due toelevated levels of the endogenous E. coli enzymes. There-fore, pMAD1 was introduced by transformation into E. colistrains with arginine and pyrimidine biosynthesis defects. Amutant lacking OTC activity because of deletions in the argFand argl genes (E. coli N134) displayed low but measurableOTC activity when harboring pMAD1 (Table 3). Likewise,E. coli JEF8, containing a deletion in the gene encoding thecatalytic subunit of CK, carB, demonstrated consistentlymeasurable CK activity when transformed with pMAD1.The CK gene product was not produced at levels as low asthose of the products of the AD and OTC genes whencompared with the levels found in S. sanguis. The data areconsistent with reports indicating that the CK gene may besubject to additional levels of control in the oral streptococci(17, 24) and, therefore, independently regulated. The dataclearly indicated that the genes for the three AD enzymeswere located on pMAD1 and were clustered in the S. sanguischromosome, perhaps similarly to the arc gene organizationin P. aeruginosa (30).As previously mentioned, the poor expression of the AD

gene in E. coli may have been due to insufficient free argininein the cell to induce or derepress the gene. Recent data fromour laboratory suggested that the levels of arginine in richmedium required to induce the system fully may be ratherhigh (>10 mM). The results of the experiments with theargFI and carB mutants (Table 3) suggested but did notprove a lack of induction. The results did indicate that theintroduction of pMAD1 into E. coli strains with deletions inboth the argF and argI genes or in the carB gene resulted ina marked decrease in the expression of the AD gene. Levelsof AD enzyme in these strains were around 10- to 15-foldlower than those found in an E. coli arginine prototroph(JM83). Of interest, the presence of the argF mutation in E.coli DH5a was associated with a three- to fivefold decrease

VOL. 57, 1989

on May 22, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

3544 BURNE ET AL.

ai 10X 0,

o1 o c0 10 20

Time (h)FIG. 2. Growth (0) versus production of the AD enzyme (0). E.

coli DH5ao harboring pMAD1 was diluted 1:100 from a 24-h culturein L broth supplemented with ampicillin into fresh TB medium plusampicillin. Cells were removed at various times, 37,000 x g solublefractions were prepared, and AD enzyme activity was assessed asdescribed in Materials and Methods. O.D.6., Optical density at 600nm.

in the level ofAD gene expression. Also, although the levelsof OTC and CK were clearly different (Table 3) betweenarginine prototrophs and auxotrophs, definite conclusionscould not be drawn, since the relative contributions of theendogenous E. coli enzymes in JM83 to the observed OTCand CK levels were not known. Nevertheless, the levels ofthe AD gene product found in E. coli JM83, when comparedwith those in N134 and JEF8, were sufficiently different notto be accounted for by gene dosage.The times in the culture cycle of E. coli harboring pMAD1

at which AD activity was detected indicated that the AD

gene was subject to temporal regulation (Fig. 2). Optimalexpression did not occur until very late in the stationaryphase. Expression during the exponential phase of growthwas extremely low. The experimental data shown in Fig. 2were obtained with E. coli DH5a, but regardless of the E.coli host chosen, expression of the genes for the AD systemwas very low or undetectable until late in the stationaryphase. AD production in S. sanguis also occurs rather late inthe growth cycle, even without catabolite repression.

Subcloning and mapping of the AD, OTC, and CK genes.To map the location of the AD gene on pMAD1, wegenerated a bank of subclones (Fig. 3) with the plasmidvector pMK4 (44). These constructions were introduced intoE. coli DH5a. A single 2.1-kbp HinclI fragment (pMAD-S47)was both necessary and sufficient for AD activity. The3.0-kbp BamHI fragment present in pMAD-S18, which con-tains this HincIl fragment, also could serve to transform E.coli to the AD-positive phenotype. Subcloning of the ADgene did not result in enhanced expression indicative of arelief of repression, although E. coli cells harboring pMAD-S18 or pMAD-S47 did express slightly higher levels of ADactivity than did cells carrying pMAD1. Slightly higherexpression was observed for the subclones carried in otherE. coli strains as well. This increase in AD activity likely wasattributable to size differences and copy numbers of thevarious subclones.Mapping of the OTC and CK genes was done by examin-

ing the expression levels of various subclones in the appro-priate mutant strains. The 3.0-kbp BamHI fragment couldrender E. coli N134 OTC positive. The adjacent 3.7-kbpHinclI fragment was also demonstrated to be both necessaryand sufficient for CK gene expression, albeit at a very lowlevel. The levels of OTC and CK enzyme activity in thesestrains were comparable to those found in the appropriate

ENZYME ACTIVI1TUnits/ mg Protein

AD OTC CK

SH H HBH HB AAHdI I 11I 11 111

Ap Plac Bg P >

B C C BH CHBpMAD-SI8 I I III

Cm Ap PlacH C C H CH

pMAD-S471 ICm Ap Plac

H

pMAD-S42

H P HdAAII I II

-~~~~~L-Cmr Ap lPlac

S 0.036 0.011

0.103 0.024

0.115 ND

BHI I ND ND

FIG. 3. AD system enzyme levels from subclones from pMAD1. Plasmids were introduced into E. coli DH5a for measuring AD activity,N134 for measuring OTC activity, and JEF8 for measuring CK activity. The thick lines represent vector sequences, and the thin lines

represent S. sanguis DNA. Restriction enzyme sites are abbreviated as in Fig. 1. ClaI sites (C) in pMAD-S18 and pMAD-S47 were used to

determine the orientation of the insert. The arrows above each insert refer to the orientation of that fragment with respect to the orientationshown for pMAD1. The arrow and designation Plac denote the location and direction of transcription of the lac promoter. Cell extracts, assays,and unit definitions are as outlined in Table 2. ND, Not detected.

spMADl L 1.44

ND

ND

1.44

PLASNM RESTRICnON MAP

. X I. . . . . .

INFECT. IMMUN.

on May 22, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

S. SANGUIS AD GENES 3545

11 18 2 5

BPH1 TRB A A Hd P S

FIG. 4. Mu dE insertions in the 3.0-kbp BamHI and overlapping3.7-kbp HincII fragments shown to encode AD, OTC, and CKactivities (Fig. 3). Insertions were constructed as described in thetext. Insertions 2, 11, and 18 map to the 2.1-kbp HincIl fragment,which expresses AD activity. Insertion 5 maps to the 3.7-kbp HinclIfragment, which expresses CK activity in E. coli JEF8.

parent strain carrying pMAD1 (Table 2), indicating that, aswith AD gene, subcloning did not result in the alleviation ofpoor expression. Notably, the addition of arginine and uracilto the culture medium did not increase AD system geneexpression from pMAD1 or its subclones.The expression of the genes of the AD system from

pMAD1 or any of its subclones was not influenced byinduction of the lac promoter by isopropyl-p-D-thiogalacto-pyranoside (data not shown), suggesting that the AD, OTC,and CK genes were transcribed from their cognate strepto-coccal promoters. The AD and OTC genes could be tran-scribed either from separate promoters or as a polycistronicmessage, but the CK gene must have its own promoter, sinceit is unlikely that the construction pMAD-S42 could containeither the AD or the OTC promoter in the proper orientation.

Neither pMAD1 nor its subclones could complement theCK or OTC defects of N134 or JEF8, as measured by aconversion to prototrophy for arginine and/or uracil. Thelack of complementation likely reflects the opposite regula-tory influences on the OTC and CK genes of E. coli and S.sanguis; levels of expression are increased by depletion ofarginine or pyrimidines for E. coli (13, 18) and excessarginine and depletion of glucose for S. sanguis (17). Also,the streptococcal enzymes, like other catabolic OTC and CKproteins (1), may not function well in the anabolic direction.

Insertional inactivation with Mu dE. Bacteriophage Mu dE(26) was used to mutagenize the 3.0-kbp BamHI fragmentencoding the AD gene. This phage is a construction of MudII 4041 (9) with a Kmr marker for use in E. coli and an Emrmarker from Tn9J7 for use in the oral streptococci. Inser-tions were mapped with HindIII and BamHI. These con-structions were linearized with EcoRI and introduced bytransformation into S. sanguis NCTC 10904. The insertiondesignated 10904-mu2 resulted in the loss of AD activity,with a concomitant decrease in OTC function (Fig. 4 andTable 4). 10904-mull and 10904-mul8 resulted in a severe

TABLE 4. Expression levels of AD system enzymesa inS. sanguis NCTC 10904 and S. sanguis treated with Mu dEb

Level of:S. sanguis strain

AD OTC CK

10904 5.82 10.38 1.7710904-mu2 0.00 1.14 3.3910904-mu5 0.06 2.46 0.00a AD system enzyme levels were measured in permeabilized cells from

stationary-phase cultures. Activity is defined as units per milligram of cell dryweight, where one unit is the amount of enzyme required to produce onemicromole of product per hour.

b Cells were grown under derepressing conditions in BHI broth supple-mented with 0.8% galactose and 25 mM arginine.

but not complete loss of AD activity, with decreased OTCactivity, suggesting that these insertions are not locatedwithin the AD structural gene but perhaps define a regula-tory region. These strains synthesized a functional CK geneproduct with CK activity 1.5- to 2-fold higher than that of theparent. Taken together, these data confirmed that subclonesof this region encode AD and OCT. Recently, one of usdemonstrated that the arcA and arcB genes described hereshow significant DNA sequence homology to the arcA geneof P. aeruginosa and the OTC genes of P. aeruginosa and E.coli (R. A. Burne, unpublished observations).Mutants lacking CK activity were obtained by introducing

insertions distal to the 2.05-kbp Hincll fragment encodingthe AD gene in the adjacent 3.7-kbp HinclI fragment (Fig. 2)into S. sanguis. The resulting Emr transformants (Fig. 4)(10904-mu5) were completely devoid ofCK activity (Table 4)and also produced abnormally low levels of AD and OTC.The data reinforced the observations that cells harboringpMAD1 and pMAD1-S42 express the S. sanguis CK geneand confirmed that the genes encoding the AD pathwayenzymes are tightly clustered on the S. sanguis chromo-some.

DISCUSSION

The genes encoding the three enzymes of the S. sanguisAD pathway have been isolated in bacteriophage lambda andsubsequently transferred to pUC-based plasmids. The genesfor the S. sanguis AD system, in keeping with the nomen-clature adopted for the Pseudomonas AD genes (30, 46),have been designated arcA (AD), arcB (catabolic OTC), andarcC (catabolic CK). Subcloning, mapping, and transposonmutagenesis studies have indicated that the genes for thispathway are tightly clustered on the S. sanguis chromo-some, as they are on the P. aeruginosa chromosome. Theresults of Southern blot experiments with S. sanguis DNAand AD gene fragments indicated that there were no rear-rangements of the AD genes during cloning (data notshown). Similar tight genetic linkages of coordinately regu-lated genes have been described previously in the oralstreptococci (6, 14, 16).The low-level expression of the AD genes in E. coli is of

particular interest, since it may provide insight into regula-tion in the streptococci. Previous reports describing thecloning and expression of genes from the oral streptococciindicated that the genes are highly transcribed and translatedby E. coli (5, 7, 14), presumably because streptococcalpromoters and ribosome-binding sites function well in E. coli(14). The genes for the AD system, particularly arcA andarcB, are poorly expressed in E. coli, and the gene productsare not detectable until very late in the stationary phase. Thesimplest interpretation is that the regulatory elements of theAD genes are not efficiently recognized by E. coli. There-fore, the genes are expressed at low levels, and the lateappearance of AD, OTC, and CK is due to the slowaccumulation of stable gene products. Another possibility isthat the genes are positively regulated in S. sanguis and thata regulatory element is missing from our clones. In fact,there is indirect evidence that such a regulator exists in P.aeruginosa (47) and S. faecalis (43). However, we haveobtained high-level expression of the arcA and arcB geneson plasmids in S. mutans (R. A. Burne, D. T. Parsons, andR. E. Marquis, J. Dent. Res. 68:1015, 1989). Possibly thereis something unique to the oral streptococci or gram-positiveorganisms for the activation of these genes. Low-levelexpression of the AD genes in OTC and CK mutants of E.

VOL. 57, 1989

on May 22, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

3546 BURNE ET AL.

coli would support the view that free arginine or a metaboliteof arginine may be involved in regulation of the system (1,13, 17, 35, 43). A decrease in cytoplasmic ATP levels couldalso regulate AD gene expression in E. coli, as previouslysuggested for other organisms (35, 43). On the basis of themapping data, it is unlikely that the 2.1-kbp HinclI fragmentencodes a repressor protein. Established models for tran-scriptional attenuation or autogenous regulation could ex-plain our observations. For example, late in the growthphase, levels of free arginine from proteolytic breakdownmay be sufficiently high to induce the AD genes. Alterna-tively, a decrease in ATP levels could modulate transcrip-tion. Consistent with either the attenuator or autogenousregulation model is the inability to remove low-level expres-sion by subcloning of the arcA gene. Yet another possibilityis that there is a posttranslational processing step whichoccurs in the oral streptococci but not in E. coli and which isrequired for the optimal expression or function of the ADsystem. Membrane-associated forms of AD have been foundin Mycoplasma hominis (28), and we have found that asignificant amount of S. sanguis AD activity is associatedwith the membrane fraction of S. sanguis. No evidence ofsuch an association with membranes of E. coli harboringpMAD1 or its derivatives has been detected. Clearly, be-cause of the complexity of the regulation of the AD system,studies done with DNA sequencing and gene fusions coupledto protein studies are needed to address poor expression inE. coli as well as AD gene regulation in S. sanguis.The effects of Mu dE insertions were consistent with the

mapping and biochemical data obtained in this study andconfirm that the genes expressed in E. coli are the S. sanguisAD genes. The insertional mutagenesis data suggest that, invivo, the genes could be transcribed as an operon, with thegene order being arcCAB, or that arcAB and arcC aredivergently transcribed. However, subcloning and expres-sion of the AD and OTC genes, apparently under the controlof their own promoter(s) and in the absence of the arcC geneand its promoter, argue against a three-gene operon model.Moreover, relatively little is known about the metabolitesand proteins which govern the expression of this gene clustersuch that down-regulation of the arc genes, as seen with10904 mu-5, may occur in response to accumulation ordepletion of a metabolite rather than to polar effects ordisruption of regions harboring divergent promoters. Addi-tionally, CK may be regulated by factors unrelated to the ADsystem since, in addition to ATP synthesis, carbamyl phos-phate can participate in the synthesis of intracellularpolysaccharides by serving as a phosphate donor to freeglucose (24). Data presented here and elsewhere (17, 24, 43)support the idea that arcC is probably subject to regulationeither coordinately or similarly to the arcA and arcB genesbut could respond to other influences independently of thearcA and arcB genes. Because of this complex regulationand because so little is known about mediators of ADexpression, a detailed study with insertional mutagenesiscould yield misleading results. Recently, the S. sanguis ADand OTC proteins were purified (R. A. Burne, unpublishedresults). These proteins are approximately 50 and 35 kilodal-tons, respectively, similar to other AD and OCT proteins (1,30). If the CK gene product is similar in molecular size to theP. aeruginosa or S. faecalis protein (35 to 39 kilodaltons) (1,30), then the entire AD gene cluster would be predicted tooccupy only 3 to 3.5 kbp. It is quite reasonable to coupleDNA sequencing with functional studies to dissect themolecular bases of differential AD gene expression. Cur-rently, the 3.0-kbp BamHI fragment is being sequenced.

The mini Mu-derived AD mutants of S. sanguis will beuseful for a number of other studies. Defined AD systemmutants can be used in in vitro studies to determine the roleof the AD system in the physiology and acid tolerance of S.sanguis. Studies with "mixed plaques" in continuous cul-tures (M. Bartels, M.S. thesis, University of Rochester,Rochester, N.Y., 1986) can also be used to assess the role ofthe system in the competitiveness of S. sanguis. The abilityof such mutants to cause caries can be assessed in the ratcaries model. All of these studies could logically be extendedto other oral streptococci, such as Streptococcus rattus andS. mitior, which possess the system.One of our objectives is to construct S. mutans strains

which have been genetically engineered to contain the ADsystem. These ammonia-producing streptococci could thenbe tested in animal models for their ability to (i) successfullyimplant; (ii) compete with plaque bacteria; and (iii) modulateplaque cariogenicity in such a way that they result in a netreduction in the amount of dental disease. Construction ofstrains of S. mutans with the AD system integrated into thechromosome is under way. The gtfA gene, cloned by anumber of laboratories (5, 6, 40) and of known nucleotidesequence (16), has been selected as the site for integration inthe initial test system, although alternative strategies can beused. Importantly, there should be nothing particularlydetrimental to S. mutans strains which possess the ADsystem, since S. rattus, the closest relative of S. mutansamong the oral streptococci (12), has the AD system and isfrequently isolated from the human mouth (20, 29). How-ever, optimization of these genetically engineered strains forimplantation would undoubtedly be necessary and could beaccomplished by passage on surfaces in continuous culturesor in gnotobiotic animals.Another strategy which is a logical spinoff of the use of the

AD system in S. mutans would be to introduce a bacterialurease into S. mutans. Urea is found in micromolar tomillimolar quantities in saliva and in very high quantities indialysis patients (38, 49, 50). High salivary urea levels havebeen associated with a lower caries incidence (38). Ureolyticorganisms have been isolated in plaque, and a correlationbetween their presence and an inhibition of plaque acidifi-cation has been described (25, 49, 50). It should be possibleto obtain a bacterial urease, perhaps from plaque bacteria,which could function when expressed in an organism such asS. mutans or S. sanguis. These strategies represent noveland potentially very powerful methodologies for the preven-tion of dental caries, and such methodologies should beapplicable to other oral diseases such as root surface cariesand periodontal disease.

ACKNOWLEDGMENTS

We thank Gary R. Bender for technical assistance in the initialisolation of recombinant bacteriophage and Larry Tabak, AidaCasiano-Col6n, and Kurt Schilling for helpful discussions.

This work was supported by Public Health Service grant P01-DE07003 from the National Institutes of Health for the RochesterCariology Center. R.A.B. was supported by the Cariology TrainingProgram of the National Institute of Dental Research (Public HealthService grant 1 T32 DE07165).

LITERATURE CITED1. Abdelal, A. T. 1979. Arginine catabolism by microorganisms.

Annu. Rev. Microbiol. 83:139-168.2. Abelson, D. C., and I. D. Mandel. 1981. The effect of saliva on

plaque pH in vivo. J. Dent. Res. 60:1634-1638.3. Archibald, R. M. 1944. Determination of citrulline and allantoin

INFECT. IMMUN.

on May 22, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

S. SANGUIS AD GENES 3547

and demonstration of citrulline in blood plasma. J. Biol. Chem.156:121-142.

4. Bender, G. R., S. Sutton, and R. E. Marquis. 1986. Acidtolerance, proton permeability, and membrane ATPases of oralstreptococci. Infect. Immun. 53:331-338.

5. Burne, R. A., B. Rubinfeld, W. H. Bowen, and R. E. Yasbin.1986. Cloning and expression of a Streptococcus mutans glu-cosyltransferase gene in Bacillus subtilis. Gene 47:201-209.

6. Burne, R. A., B. Rubinfeld, W. H. Bowen, and R. E. Yasbin.1986. Tight genetic linkage of a glucosyltransferase and dextra-nase of Streptococcus mutans GS-5. J. Dent. Res. 65:1392-1401.

7. Burne, R. A., K. Schilling, W. H. Bowen, and R. E. Yasbin.1987. Expression, purification, and characterization of an exo-,3-D-fructosidase of Streptococcus mutans. J. Bacteriol. 169:4507-4517.

8. Casiano-Colon, A., and R. E. Marquis. 1988. Role of the argininedeiminase system in protecting oral bacteria and an enzymaticbasis for acid tolerance. Infect. Immun. 54:1318-1324.

9. Castilho, B. A., P. Olfson, and M. J. Casadaban. 1984. Plasmidinsertion mutagenesis and lac gene fusion with Mini-Mu bacte-riophage transposons. J. Bacteriol. 158:488-495.

10. Chassy, B. M. 1976. A gentle method for the lysis of oralstreptococci. Biochem. Biophys. Res. Commun. 69:603-608.

11. Cowman, R. A., S. S. Baron, R. J. Fitzgerald, R. E. Stuchell,and I. D. Mandel. 1983. Comparative growth responses of oralstreptococci on mixed saliva or the separate submandibular andparotid secretions from caries-active and caries-free individuals.J. Dent. Res. 62:946-951.

12. Coykendall, A. L. 1987. Proposal to elevate the subspecies ofStreptococcus mutans to species status, based on their molec-ular composition. Int. J. Syst. Bacteriol. 27:26-30.

13. Cunin, R., N. Glansdorff, A. Pierard, and V. Stalon. 1986.Biosynthesis and metabolism of arginine in bacteria. Microbiol.Rev. 50:314-352.

14. Curtiss, R., III. 1985. Genetic analysis of Streptococcus mutansvirulence. Curr. Top. Microbiol. 118:253-277.

15. Driessen, A. J. M., B. Poolman, R. Kiewiet, and W. N. Konings.1987. Arginine transport in Streptococcus lactis is catalyzed bya cationic exchanger. Proc. Natl. Acad. Sci. USA 84:6093-6097.

16. Ferretti, J. J., T.-T. Huang, and R. R. B. Russell. 1988. Se-quence analysis of the glucosyltransferase A gene (gtfA) fromStreptococcus mutans Ingbritt. Infect. Immun. 56:1585-1588.

17. Ferro, K. J., G. R. Bender, and R. E. Marquis. 1983. Coordi-nately repressible arginine deiminase system in Streptococcussanguis. Curr. Microbiol. 9:145-150.

18. Glansdorff, N. 1987. Biosynthesis of arginine and polyamines, p.321-344. In J. L. Ingraham, K. B. Low, B. Magasanik, M.Schaechter, and H. E. Umbarger (ed.), Escherichia coli andSalmonella typhimurium: cellular and molecular biology. Amer-ican Society for Microbiology, Washington, D.C.

19. Godson, G. N., and D. Vapnek. 1973. A simple method ofpreparing large amounts of fX174 RFI supercoiled DNA.Biochim. Biophys. Acta 299:516-520.

20. Hamada, S., and H. D. Slade. 1980. Biology, immunology, andcariogenicity of Streptococcus mutans. Microbiol. Rev. 44:331-384.

21. Hillman, J. D. 1982. Analysis of lactate dehydrogenase mutantsof Streptococcus mutans and their potential use in replacementtherapy for caries prevention, p. 191-194. In D. Schlessinger(ed.), Microbiology-1982. American Society for Microbiology,Washington, D.C.

22. Hiraoka, B. Y., M. Mogi, K. Fukasawa, and M. Harada. 1986.Coordinate repression of arginine aminopeptidase and threeenzymes of the arginine deiminase pathway in Streptococcusmitis. Biochem. Int. 12:881-887.

23. Holmes, D. S., and M. Quigley. 1981. A rapid boiling method forthe preparation of bacterial plasmids. Anal. Biochem. 114:193-197.

24. Keevil, C. W., P. D. Marsh, and D. C. Ellwood. 1984. Regulationof glucose metabolism in oral streptococci through independentpathways of glucose 6-phosphate and glucose 1-phosphate. J.Bacteriol. 157:560-567.

25. Kleinberg, I., J. A. Kanapka, and D. Craw. 1976. Effect of salivaand salivary factors on the metabolism of the mixed oral flora, p.433-464. In H. M. Stiles, W. J. Loesche, and T. C. O'Brien(ed.), Proceedings: microbial aspects of dental caries (a specialsupplement to Microbiology Abstracts), vol. 3. InformationRetrieval, Inc., Washington D.C.

26. Kuramitsu, H. K. 1987. Utilization of a mini mu transposon toconstruct defined mutants in Streptococcus mutans. Mol. Mi-crobiol. 1:229-231.

27. Lawson, J. W., and H. Gooder. 1970. Growth and developmentof competence in group H streptococci. J. Bacteriol 102:820-825.

28. Lin, J.-S. L. 1986. Arginine deiminase of Mycoplasma hominis:cytoplasmic and membrane forms. J. Gen. Microbiol. 132:1467-1474.

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

30. Luthi, E., A. Mercenier, and D. Haas. 1986. The arcABC operonrequired for fermentative growth of Pseudomonas aeruginosaon arginine: TnS-751-assisted cloning and localization of struc-tural genes. J. Gen. Microbiol. 132:2667-2675.

31. Macrina, F. L., R. P. Evans, T. A. Tobian, D. L. Hartley, D. B.Clewell, and K. R. Jones. 1983. Novel shuttle plasmid vehiclesfor Escherichia-Streptococcus transgeneric cloning. Gene 25:145-150.

32. Mandel, M., and A. Higa. 1970. Calcium dependent bacterio-phage DNA infection. J. Mol. Biol. 53:159-162.

33. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

34. Marquis, R. E., G. R. Bender, D. R. Murray, and A. Wong.1987. Arginine deiminase system and bacterial adaptation toacid environments. Appl. Environ. Microbiol. 53:198-200.

35. Mercenier, A., J.-P. Simon, C. Vander Wauven, D. Haas, and V.Stalon. 1980. Regulation of enzyme synthesis in the argininedeiminase pathway of Pseudomonas aeruginosa. J. Bacteriol.144:159-163.

36. Miller, J. H. 1972. Experiments in molecular genetics. ColdSpring Harbor Laboratory, Cold Spring Harbor, N.Y.

37. Oppenheim, F. G., T. Xu, F. M. McMillian, S. M. Levitz, R. D.Diamond, G. D. Offner, and R. F. Troxier. 1988. Histatins, anovel family of histidine-rich proteins in human parotid saliva.J. Biol. Chem. 263:7472-7477.

38. Peterson, S., J. Woodhead, and J. Crall. 1985. Caries resistancein children with chronic renal failure: plaque pH, salivary pH,and salivary composition. Pediatr. Res. 19:796-799.

39. Poolman, B., A. M. Driessen, and W. N. Konings. 1987. Regu-lation of arginine-ornithine exchange and the arginine deiminasepathway in Streptococcus lactis. J. Bacteriol. 169:5597-5604.

40. Robeson, J. P., R. G. Barletta, and R. Curtiss III. 1983.Expression of a Streptococcus mutans glucosyltransferase genein Escherichia coli. J. Bacteriol. 153:211-221.

41. Rogers, A. H., P. S. Zilm, N. J. Gully, and A. L. Pfennig. 1988.Response of Streptococcus sanguis to arginine-containing pep-tides. Infect. Immun. 56:687-692.

42. Schimke, R. T., C. M. Berlin, E. W. Sweeney, and W. R.Carroll. 1986. The generation of energy by the arginine dihy-drolase pathway in Mycoplasma hominis 07. J. Biol. Chem.241:2228-2236.

43. Simon, J.-P., B. Wargnies, and V. Stalon. 1982. Control ofenzyme synthesis in the arginine deiminase pathway of Strep-tococcus faecalis. J. Bacteriol. 150:1085-1090.

44. Sullivan, M. A., R. E. Yasbin, and F. E. Young. 1984. Newshuttle vectors for Bacillus subtilis and Escherichia coli whichallow the rapid detection of inserted fragments. Gene 29:21-26.

45. Tanzer, J. M., A. B. Kurasz, and J. Clive. 1985. Inhibition ofecological emergence of mutans streptococci naturally transmit-ted between rats and consequent caries inhibition by Strepto-coccus salivarius TOVE-R infection. Infect. Immun. 49:76-83.

46. Tiemeyer, D. C., S. M. Tilghman, and P. Leder. 1977. Purifica-tion and cloning of a mouse ribosomal gene fragment in coliphage lambda. Gene 2:173-191.

47. Vander Wauven, C., A. Pierard, M. Kley-Raymann, and D.

VOL. 57, 1989

on May 22, 2021 by guest

http://iai.asm.org/

Dow

nloaded from

INFECT. IMMUN.

Haas. 1984. Pseudomonas aeruginosa mutants affected in an-

aerobic growth on arginine: evidence for a four-gene clusterencoding the arginine deiminase pathway. J. Bacteriol. 160:928-934.

48. Vieira, J., and J. Messing. 1982. The pUC plasmids, an

M13mp7-derived system for insertion mutagenesis and sequenc-ing with synthetic universal primers. Gene 19:259-268.

49. Wjeyeweera, R. L., and I. Kleinberg. 1989. Acid-base pHcurves in vitro with mixtures of pure cultures of human oralmicroorganisms. Arch. Oral Biol. 34:55-64.

50. WiJeyeweera, R. L., and I. Kleinberg. 1989. Arginolytic andureolytic activities of pure cultures of human oral bacteria andtheir effects on the pH response of salivary sediment and dentalplaque in vitro. Arch. Oral Biol. 34:43-53.

3548 BURNE ET AL.

on May 22, 2021 by guest

http://iai.asm.org/

Dow

nloaded from