INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Despite significant gains in the control and treatment of dental caries, this disease remains widespread and costly to treat. Caries formation results from the acidification of dental plaque driven by microbial metabolism of dietary carbohydrate. The pH of dental plaque can fall to values as low as 4.0 (35), causing dissolution of the tooth enamel (37). In addition, frequent consumption of fermentable carbohydrate leads to extended periods of plaque acidification, which encourages the emergence of an acid-tolerant, cariogenic plaque microflora enriched in organisms such as mutans streptococci and lactobacilli (6). The enrichment of dental plaque with cariogenic microorganisms is generally accompanied by the loss of less acid-tolerant, less cariogenic organisms, which are abundant in so-called healthy dental plaque. These organisms include the sanguis group streptococci and Actinomyces naeslundii (4), which, interestingly, are also capable of neutralizing the acids produced by glycolysis through the generation of ammonia from salivary substrates.

Caries development is usually a prolonged process involving cycles of demineralization and remineralization. Periods of plaque acidification and tooth demineralization are normally followed by phases of alkalinization with a return to more neutral plaque pH values (35), which promotes remineralization at the tooth surface (22, 23). It is when the phases of demineralization dominate that a carious lesion develops. Whereas the processes by which dental plaque becomes acidified have been intensively studied, the alkalinization phase and pH homeostasis during fasting periods are rather poorly understood. A number of mechanisms are thought to contribute to the alkalinization of dental plaque, including clearance of acids and sugars by saliva, buffering by salivary and bacterial components, and production of alkali by plaque bacteria, which occurs primarily through the metabolism of urea to ammonia by microbial urease activity.

Urea is secreted continuously in the range of 3 to 10 mM in saliva and crevicular fluids of healthy individuals (15) and is rapidly hydrolyzed by the urease enzymes of oral microflora. Existing data indirectly support a major role for ureolysis in plaque pH homeostasis. Elevated salivary urea and ammonia concentrations are correlated with marked reductions in the extent and duration of plaque acidification following a carbohydrate challenge (20). Urea hydrolysis can neutralize plaque acids (33) and may positively influence plaque ecology by preventing the pH from falling to levels that select for the outgrowth of aciduric, cariogenic microorganisms (6, 8, 9). In addition, ammonia released by ureolysis can promote remineralization of the tooth enamel (22, 26, 27). Clinical studies indicate that caries-resistant patients have elevated resting plaque pH values and that these values are not lowered to the same extent as those in caries-susceptible individuals following a carbohydrate intake (1). Margolis et al. (22) have confirmed that caries-resistant subjects have more alkaline resting plaque pH values than do caries-susceptible individuals and have correlated this, in part, with increased ammonium concentrations in plaque. Studies of patients with chronic renal failure have also shown that these patients, who have salivary urea concentrations often greater than 50 mM, have alkaline plaque pH levels and a very low incidence of dental caries, despite ingestion of a diet that is dominated by carbohydrates (29; E. P. Syrrakou, M.S. thesis, University of Rochester, Rochester, N.Y.).

Recently, attention has focused on the concept that the development of cariogenic plaque may result not only from the extensive acidification of dental plaque but also from a diminution in the alkali-generating capacity of oral biofilms colonizing a carious lesion (9). Loss of ammonia-producing bacteria from the complex populations on the tooth surface would reduce the capacity of plaque to neutralize acids and slow the return of plaque pH to more neutral values. This concept is consistent with two observations: (i) the resting plaque pH in healthy individuals is higher than the pH of the saliva which bathes the plaque, and (ii) the depth and duration of plaque acidification is greater in caries-prone subjects. Also reinforcing this concept is the demonstration that individuals with low salivary ureolytic capacity have a markedly diminished capacity to blunt glycolytic acidification (34). More recently, recombinant Streptococcus mutans strains carrying plasmid-borne urease genes were used in vitro to demonstrate directly that the levels of urease commonly found in healthy plaque are sufficient to offset a pH drop by using physiologically relevant levels of urea, even in the presence of a 10-fold molar excess of glucose (12). Importantly, the ureolytic capacity of dental plaque within a carious lesion and the prevalence of alkali-generating bacteria following sustained acidification of dental plaque remain unexplored.

Unfortunately, testing of hypotheses related to the base-producing capacity of biofilms and oral health in well-controlled clinical studies has yet to be undertaken. A variety of accepted animal caries models are available for testing the effects of various carbohydrates and therapeutics on caries formation, with the rat model appearing to be the most appropriate and widely accepted. However, these caries models are not readily adaptable to the study of alkali generation because of a lack of a suitable test organism and differences in the endogenous flora of rats and humans with regard to urease activity (see below). In clinical studies, essentially all of the data relating ureolysis to plaque pH homeostasis and oral health are restricted to total plaque from healthy individuals (5, 19, 30, 33). To address some of these deficiencies, recombinant ureolytic strains of the cariogenic plaque bacterium S. mutans which could be implanted in the plaque of experimental animals were constructed to test the hypothesis that enhancing the alkali-generating capacity of plaque could reduce the incidence of caries formation.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and growth conditions. The strains and plasmids used in this study and their relevant characteristics are described in Table 1. Escherichia coli was grown and maintained in L broth (31), and S. mutans strains were grown and maintained in brain heart infusion (BHI) medium in the presence of the appropriate antibiotics when necessary. S. mutans UA159 St^R was a streptomycin-resistant spontaneous mutant selected by plating S. mutans UA159 (UAB159) on BHI agar plates (1.5% Bacto agar) supplemented with 500 µg of streptomycin per ml. All growth media were obtained from Difco (Detroit, Mich.) and all chemicals were obtained from Sigma (St. Louis, Mo.) unless otherwise noted.

DNA manipulations. DNA isolation and recombinant DNA manipulations were carried out using established procedures (31). Genetically competent S. mutans (28) was prepared by growing S. mutans UA159 in BHI broth supplemented with 10% heat-inactivated horse serum (Life Technologies, Inc., Gaithersburg, Md.) at 37°C in a 5% CO₂ aerobic atmosphere. Competent S. mutans was transformed with plasmid DNA isolated from E. coli. Transformants were selected and maintained on BHI agar plates containing spectinomycin (250 µg/ml) or tetracycline (5 µg/ml) as needed. Integration of the antibiotic resistance and urease genes into the S. mutans chromosome was confirmed by Southern blotting (31).

Determination of enzyme activity. Urease activity was assayed in intact cells as previously described (10). To examine the nickel dependence of urease biogenesis, recombinant S. mutans strains were grown to an optical density at 600 nm of ca. 0.6 in BHI supplemented with NiCl₂ at concentrations of 0, 25, 50, 75 or 100 µM. The cells were concentrated by centrifugation, washed, resuspended in 10 mM sodium phosphate (pH 7.0), and assayed for urease activity. Activity was expressed in units (U), defined as the amount of enzyme required to hydrolyze 1 µmol of urea per min, and was normalized to cell dry weight. To examine urease activity expressed in the plaques of experimental animals, the mandibles and the maxillary palate of each animal were individually placed in sterile sodium phosphate buffer on ice and sonicated for 30 s at 10-s intervals. The suspensions from the left and right mandibles from all animals in each group were pooled, as were the suspensions from the maxillary palate from all animals in each group. The samples were centrifuged at 2,900 × g for 20 min at 4°C, resuspended in 4 ml of sodium phosphate buffer, and stored on ice. The amount of ammonia released was measured after a 6.5-h incubation in the presence of urea, with a no-urea-added control to detect any ammonia liberated from other sources. Activity was normalized to protein content as determined by the method of Bradford (7).

In vitro pH drop experiments. Recombinant S. mutans was grown in BHI broth supplemented with 0 or 25 µM NiCl₂ to an optical density at 600 nm of ca. 0.6. The cells were concentrated by centrifugation, washed, and resuspended in 1/10 the initial volume in ice-cold pH drop buffer (10 mM KCl, 1 mM MgCl₂). The glycolytic capacities of the cells were determined as described by Belli and Marquis (3). The ureolytic capacities of the strains and the influence of urea metabolism on environmental acidification as a result of glycolysis were determined by initiating the pH drop and simultaneously adding 56 mM glucose and 5 or 25 mM urea (12). The pH of the cell suspensions was monitored continuously for 1 h using a KCl glass electrode connected to either a Corning 320 or a Beckman 10 meter.

Animal studies. Eighteen litters of female Sprague-Dawley rat pups with their dams were obtained from Charles River Breeding Laboratories (Wilmington, Mass.). Prior to infection, the dams were screened for mutans streptococci by plating from oral swabs onto mitis salivarius (MS) agar on MS agar supplemented with bacitracin (1 µg/ml) (MSB agar). No animals were found that harbored detectable levels of mutans streptococci. Animals were also screened for and found to be free of sialoacroadenitis virus by immunologic screening, since this virus can affect salivary gland function, which in turn could affect the development of caries. At 19 days of age, the pups were weaned, grouped into nine groups of 16 animals, and caged in plastic cages. The rats were then infected on two consecutive days by oral swabbing with cultures of the various strains of S. mutans (Table 2) that had been grown to mid-exponential phase in low-molecular-weight medium (39). One group of animals was infected with S. mutans UA159 St^R, four groups (groups 2 to 5) were infected with the ureolytic strain S. mutans ACUS6, and four groups (groups 6 to 9) were infected with the nonureolytic control strain, S. mutans ACUS8. Two days after inoculation, successful infection was confirmed by plating oral swabs from each animal on MSB agar and on MSB agar containing appropriate antibiotics.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of recombinant, ureolytic S. mutans. The strategy used for constructing recombinant, ureolytic S. mutans is depicted in Fig. 1. Briefly, a spectinomycin resistance gene (38) was cloned 5', and in the opposite orientation, to the urease (ure) gene cluster on pMC12. Plasmid pMC12 (10) harbors the intact ureABC genes, encoding the enzyme subunits, and the ureEFGD genes, encoding the accessory proteins required for activation of the apoenzyme from S. salivarius 57.I (Fig. 1A). Recombinant bacteria harboring pMC12 express urease, but the amount of active enzyme produced is dependent on the concentration of NiCl₂ in the growth medium (10). The urease and spectinomycin genes were cloned as a cassette into the S. mutans lac sequences on the integration vector pVA2216 (18), with concomitant replacement of the tetracycline resistance gene. The resulting plasmid, pACUS2 (Fig. 1B), was linearized by restriction digestion at a unique PvuI site within the vector and used to transform competent S. mutans UA159. Allelic exchange between homologous plasmid and chromosomal sequences resulted in the establishment of the spectinomycin/urease cassette in the UA159 lac locus (Fig. 1C). Spectinomycin-resistant transformants were identified by selection on BHI agar supplemented with spectinomycin and were transferred to modified urea agar plates, which contained a pH indicator dye, to screen for the expression of urease activity. Colonies which produced a pink color, indicating the production of alkali from urea, were isolated and found to express urease activity when grown in the presence of 50 µM NiCl₂. One of the positive clones was selected and named ACUS4.

View larger version (26K): [in this window] [in a new window]   FIG. 1.   Construction of recombinant, ureolytic S. mutans. The ureABCEFGD genes and a partial ureI open reading frame from S. salivarius were cloned, adjacent to a Spr gene, within S. mutans lac sequences on the integration vector, pVA2216 (18). Introduction of the recombinant vector, pACUS2, into S. mutans allowed integration of the genes within the chromosomal lac locus by allelic exchange. The ure genes in the resulting strain, S. mutans ACUS4, were transcribed from a putative weak cognate promoter, PureA. To construct a strain expressing higher levels of urease activity, a second integration vector, pAC21, was constructed that harbored tet flanked by Spr and ure sequences. The introduction of pAC21 into ACUS4 allowed homologous recombination such that tet was integrated 5' to ure in ACUS4. As a result, the ure operon in this strain, S. mutans ACUS6, was transcribed from a strong promoter within tet, Ptet, at a constitutive level. Two control strains were also constructed using a similar strategy (data not shown). S. mutans ACUS5 and ACUS8 were inactivated at the lac locus by integration of the Spr gene alone or together with tet, respectively. Arrows beneath the gene clusters indicate transcriptional direction. Rows A to E are as described in the text.

1

1

Characterization of urease expression by recombinant, ureolytic S. mutans. S. mutans ACUS4 and S. mutans ACUS6 expressed urease activity, whose level was dependent on the concentration of exogenous NiCl₂ present in the growth medium (Fig. 2). As is commonly observed when urease genes are expressed in nonureolytic hosts (24), addition of NiCl₂ to the growth medium was needed to observe significant urease activity, because nonureolytic strains lack the high-affinity transport proteins that are needed to scavenge sufficient Ni²⁺ from the environment. When no additional NiCl₂ was added to the growth medium, ACUS4 and ACUS6 were poorly ureolytic, expressing only 2 × 10⁴ U and 5 × 10⁴ U mg of cell dry weight¹, respectively. However, when grown in the presence of 25 µM NiCl₂, ACUS4 expressed approximately 0.025 U mg of cell dry weight¹. When grown in 25 µM NiCl₂, ACUS6 expressed greater than threefold more activity than ACUS4 did (approximately 0.08 U mg of cell dry weight¹). Increasing the concentration of nickel to 50 µM slightly increased the level of urease activity expressed by ACUS6. S. mutans ACUS5 and ACUS8 were nonureolytic regardless of the concentration of exogenous NiCl₂ added to the growth medium (data not shown).

View larger version (24K): [in this window] [in a new window]   FIG. 2.   Nickel-dependent urease activity in recombinant, ureolytic S. mutans. S. mutans ACUS4 and S. mutans ACUS6 were grown to an optical density of 0.6 in BHI supplemented with nickel chloride and assayed for urease activity. Activity was expressed in units (micromoles of urea hydrolyzed per minute per milligram of cell dry weight. The graph is representative of results obtained with at least four separately grown bacterial cultures of each strain. In control experiments, no urease activity was detected for the parent strain, S. mutans UA159, or the recombinant strains, S. mutans ACUS5 and ACUS8, regardless of the concentration of nickel chloride (data not shown).

Relationship of urease activity to environmental pH-modulating capacity in vitro. The ability to modulate the level of activity of ureolytic S. mutans isolates by nickel supplementation and promoter strength allowed for exploration of the relationship between the level of urease activity and the capacity of the strains to retard glycolytic pH reduction. The organisms were grown in the absence of NiCl₂ and allowed to metabolize a solution of excess glucose. All strains reduced the suspension pH to a final value of approximately 3.5 at a comparable rate (Fig. 3; some data not shown). Growth in the presence of 25 µM NiCl₂ did not reduce the capacity of these strains to metabolize glucose and acidify the environment (Fig. 3). When grown in the presence of NiCl₂ and allowed to metabolize glucose and 25 mM urea, S. mutans ACUS4 reduced the extent of the pH fall by approximately 0.5 pH unit (Fig. 3A). The moderation of environmental acidification, as a result of urea metabolism, was more pronounced in the more strongly ureolytic strain, S. mutans ACUS6 (Fig. 3B). Specifically, when grown in the same concentration of NiCl₂ (25 µM) and expressing threefold more urease activity than ACUS4 (Fig. 2), cell suspensions of ACUS6 metabolizing glucose and as little as 5 mM urea had a final pH value almost 1.5 pH units higher than did those metabolizing glucose alone. Increasing the concentration of urea to 25 mM resulted in a final cell suspension pH of 6.5 at the end of the experiment (1 h), which was 3.0 pH units higher than that achieved following the metabolism of glucose alone. While it was apparent that the concentration of urea influenced the pH-moderating capacity of the ureolytic strains, the absolute level of urease present was an essential determinant in the depth and duration of the acidification. For example, when ACUS6 was grown in the absence of NiCl₂ and allowed to metabolize glucose and 25 mM urea, the final suspension pH was just 0.2 pH unit higher than when it metabolized glucose alone (Fig. 3B). Also, the nonureolytic control strains, S. mutans ACUS5 and ACUS8, had an identical pH drop profile to the wild-type strain, S. mutans UA159 (data not shown). No moderation of pH in the presence of urea was observed using the nonureolytic strains.

View larger version (14K): [in this window] [in a new window]   FIG. 3.   Moderation of glycolytic acidification as a function of urease activity and urea availability. ACUS6 and ACUS8 were grown in BHI in the absence (open symbols) or presence (solid symbols) of 25 µM nickel chloride to an optical density of ca. 0.6. Glucose was added to the cell suspension to initiate glycolysis together with 0 mM (, ), 5 mM (, ), or 25 mM (, ) urea. Data points were collected every 30 s over a 1-h period. Data points for every second for 15 min, every minute for 5 min, and every 5 min for 40 min are presented. The urease activity expressed by these strains was 0.27 × 103 and 23.0 × 103 U mg of cell dry weight1 for ACUS4 and 0.5 × 103 and 82.5 × 103 U mg of cell dry weight1 for ACUS6 in the absence and presence of nickel chloride, respectively.

Animal studies. A small pilot study demonstrated that all the strains would implant and that the level of urease could be modulated by provision of nickel in the drinking water (data not shown). Then, using ACUS6 and the otherwise isogenic, nonureolytic strain ACUS8, a full-scale experiment was undertaken to determine if increases in the levels of urease enzyme in vivo could moderate caries formation. As outlined in Table 2, rats were infected with either a streptomycin-resistant derivative of the wild-type strain, S. mutans UA159 St^R, the ureolytic recombinant S. mutans ACUS6, or the nonureolytic recombinant strain S. mutans ACUS8. All animals were provided with Diet 2000 containing 56% sucrose and one of the following: (i) sweetened (5% sucrose) drinking water, (ii) sweetened water supplemented with 50 µM NiCl₂, (iii) sweetened water supplemented with 50 mM urea, or (iv) sweetened water supplemented with 50 µM NiCl₂ and 50 mM urea. No differences were seen in weight gains of the various groups of animals during the course of the study.

View larger version (19K): [in this window] [in a new window]   FIG. 4.   Microbiological assessment of the microflora on the rat dentition. The right mandible from each animal was sonicated in 5 ml of sterile sodium phosphate buffer (pH 7.0). Diluted (1:100) and undiluted aliquots were plated on MSB to determine total mutans counts, on MSB containing streptomycin and spectinomycin (St/Sp) to determine total recombinant populations, and on SB agar to determine total cultivable flora (TCF). The graph represents the mean bacterial counts for all animals in each group. Error bars represent the standard error of the mean. Statistical analysis was carried out using the Tukey-Kramer HSD test.

View larger version (20K): [in this window] [in a new window]   FIG. 5.   Urease specific activity expressed by S. mutans ACUS6 and ACUS8 in vivo. The left and right mandibles (Ma) from all rats in each group were sonicated in sterile sodium phosphate (pH 7.0). The sonicate was pooled and assayed for urease activity as described above. Similarly, the pooled sonicate from the maxillae (Mx) of all rats in each group was assayed for urease activity. Activity is expressed as units per milligram of protein. Group 1 was infected with the wild-type strain. Groups 2 to 5 were infected with the ureolytic recombinant S. mutans ACUS6, and groups 6 to 9 were infected with the nonureolytic control strain S. mutans ACUS8. The supplements to the drinking water are indicated above the bars in the graph. Groups, 1, 2, and 6 received 5% sucrose-sweetened water (), the same water supplemented with 50 µM NiCl2 was provided to groups 3 and 7, the same water supplemented with 50 mM urea was provided to groups 4 and 8, and the same water supplemented with both 50 µM NiCl2 and 50 mM urea was provided to groups 5 and 9.

View larger version (27K): [in this window] [in a new window]   FIG. 6.   Mean caries scores from animals infected with the ureolytic recombinant S. mutans ACUS6. All four jaw quadrants from each animal were scored for smooth-surface (A) and sulcal (B) caries by the method of Keyes (17). The graph represents the mean caries score and standard error of the mean from all animals in each group as a percentage of total enamel (E) scores for ACUS6 infected animals fed a cariogenic diet, where 100% = 14 in panel A and 100% = 35 in panel B. All animals were fed a cariogenic diet with sweetened drinking water. Nickel chloride and urea were added to the drinking water as indicated underneath the bar chart. Statistical comparisons were made using the Tukey-Kramer HSD test (P = 0.05). Symbols above the bars indicate groups that are statistically different from control groups in caries scores. Only differences in E scores are indicated. E, total caries; Ds, slight dentinal involvement; Dm, moderate dentinal involvement; and Dx, extensive dentinal involvement.

View larger version (26K): [in this window] [in a new window]   FIG. 7.   Mean caries scores from animals infected with the nonureolytic recombinant S. mutans ACUS8. All four jaw quadrants from each animal were scored for smooth-surface (A) and sulcal (B) caries. The graph represents the mean caries score and standard error of the mean from all animals in each group as a percentage of total (E) scores for ACUS8-infected animals fed a cariogenic diet, where 100% = 12 in panel A and 100% = 37 in panel B. All animals were fed a cariogenic diet with sweetened drinking water. Nickel chloride and urea were added to the drinking water as indicated underneath the bar chart. Statistical comparisons were made using the Tukey-Kramer HSD test (P  0.05). The symbols used are identical to those in Fig. 6.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Previously, a recombinant strain harboring the S. salivarius 57.I ure cluster, S. mutans ACO4, had been constructed and used in vitro to test hypotheses relating pH-moderating potential to ureolytic capacity (12). Those studies established the feasibility of expressing recombinant ure genes in cariogenic plaque bacteria and demonstrated in principle that some minimum level of alkali-generating capacity was necessary to significantly blunt the pH fall due to glycolysis. In ACO4, the ure cluster was plasmid borne and the segregational stability was dependent on maintenance of antibiotic pressure. Once the urease genes were integrated into the chromosome in ACUS4 and ACUS6, they were stably maintained and expressed in the absence of antibiotic selection, without detectable deleterious effects on cell growth or glycolytic capacity. Consequently, as demonstrated in this communication, strains with the ACUS designation were suitable for exploration of the role of ammonia generation in vivo.

The requirement of the ureolytic S. mutans strains for nickel to produce an active urease, as well as the differential level of ure transcription in ACUS4 and ACUS6, turned out to be a major advantage in this study. Simply by modulating the concentration of available NiCl₂ in the growth medium, populations of cells were generated that expressed low, intermediate, or high levels of urease activity. In vitro analysis confirmed that increasing the ureolytic capacity of plaque bacteria increased their capacity to moderate acidification by glycolysis and reinforced that some critical level of urease activity was necessary to prevent the fall of the pH to values at which caries occurs. Clearly, ACUS6 produced a sufficient quantity of urease to have a considerable impact on the environmental pH whereas AUCS4 did not. Recombinant strains that could produce a functional urease in the absence of exogenous nickel would be useful in future in vitro and in vivo studies, and efforts to isolate the genes for the S. salivarius nickel uptake system are under way.

The decrease in caries in animals receiving drinking water supplemented with NiCl₂ is not entirely surprising. Nickel is a trace element in the biosphere which may be essential for animal nutrition. However, high concentrations of nickel can be highly toxic and carcinogenic, presumably through inhibition of enzyme activities and through DNA damage induced by lipid peroxidation, respectively (32). Notably, the levels of nickel in the drinking water were below that which we have determined to be toxic for S. mutans or to have significant impact on the growth of oral streptococci. Almost assuredly, then, the effects of nickel on caries seen in this study were not exerted through enhancement of urease in naturally ureolytic organisms, since the addition of nickel alone to drinking water did not enhance urease activity in controls. Instead, the decreases in caries formation were probably elicited through the toxicity of nickel to the bacteria and the inhibition of enzymes.

The decision to carry out the rat caries experiments with ACUS6 as the test strain was based on its pH-moderating capacity in the pH drop experiments. Specifically, the level of urease activity achieved by ACUS6 was 0.2 µmol of NH₄ produced min¹ mg of protein¹. In this case, the rate of ammonia production from urea was close to the glycolytic rate reported for S. mutans (0.08 to 0.25 µmol of lactate produced mg of cell dry weight¹) (3), so that with sufficient concentrations of urea, ACUS6 would be predicted to produced enough ammonia to neutralize the organic acids produced by glycolysis. Also, the maximum ureolytic rate demonstrated by ACUS6 grown in the presence of nickel was comparable to values predicted by Dibdin and Dawes (13) to be sufficient to moderate plaque acidification, preventing enamel demineralization.

Based on the in vitro studies, ACUS6 should have been only weakly ureolytic in animals which did not receive supplemental NiCl₂. Consistent with this prediction was the finding that urease levels in ACUS6-infected rats were comparable to those measured in animals infected with wild-type or the nonureolytic control strain ACUS8 when nickel was absent. In contrast, the provision of nickel and urea to animals harboring ACUS6 increased urease activity significantly and resulted in statistically meaningful reductions in both smooth-surface and sulcal caries. The additional decrease in caries was not observed in ACUS8-infected animals. Thus, the presence of ACUS6 and the production of urease by this strain was the key factor in the inhibition of caries. The most probable mechanism by which caries inhibition occurred was the direct effect of ammonia production on plaque pH. However, another possible effect of the presence of ammonia-producing streptococci may have been to influence the overall ecology of plaque, such that changes in plaque structure or composition also contributed to the decline in caries. Regardless, it is clear that the presence of recombinant, ammonia-producing organisms can moderate the initiation and progression of dental caries in experimental animals.

S. mutans UA159 St^R was included in this study as a positive control for the elicitation of dental caries since this strain implants and causes smooth-surface and sulcal caries in rodents (39). Although the values were not statistically different, S. mutans UA159 St^R appeared to be modestly less cariogenic on smooth surfaces than the ACUS strains when animals received the control diet only. As indicated, UA159 St^R is a spontaneously arising streptomycin-resistant strain of S. mutans. Decreases in the virulence of streptomycin-resistant strains of S. mutans have been reported previously (2). Thus, the most appropriate control for comparison of the cariogenicity of ACUS6 is ACUS8, since these strains harbor the same antibiotic resistance genes, lack lactose catabolism capacity, and differ only in their possession of urease genes.

It appears as though the limiting factor in ammonia generation from urea in the healthy human may be urea, not a lack of enzymatic activity. Specifically, urea is not found in any appreciable quantity in dental plaque (5), probably because it is rapidly hydrolyzed by the relatively high levels of urease present in total samples of plaque and saliva (around 1 U mg [wet weight]¹ in healthy individuals) (34). However, it has also been shown (34) that individuals who produce low levels of urease have a poor capacity to offset glycolytic acidification when provided with urea. It is also important to consider that caries is a site-specific disease and that loss of alkali-generating potential at a carious site could have devastating consequences without having any detectable impact on the total oral ureolytic potential. In the rat caries model, ambient levels of urea provided in saliva were insufficient to elicit dramatic changes in caries, even in the presence of organisms producing an active urease. However, addition to the drinking water, which is ingested only periodically, of as little as 50 mM urea was sufficient to inhibit smooth-surface and sulcal caries, provided that animals were infected with recombinant streptococci producing an active urease. It is also essential to note that inhibition of caries by the urease-producing streptococci occurred in the face of an overwhelming cariogenic challenge. Therefore, it can be concluded that relatively small differences in urea concentrations and in the amount of urease enzyme may dramatically affect caries initiation and progression.

The finding that increasing base production in dental plaque reduces not only smooth-surface caries but also sulcal scores is important. Sulcal surfaces are areas where the diet and acidic breakdown products are likely to be retained, more so than on the smooth surfaces of the teeth; therefore, the sulci are generally subjected to a more severe acid attack and are more prone to caries development. Clinically, sulcal and interproximal caries are far more common than smooth surface caries and fluoride is far less effective against these types of caries. It therefore appears that in contrast to some other caries prevention strategies, enhancement of ammonia generation in plaque may be an effective way to combat the more common types of carious lesions.

Bacterial strains with diminished virulence have been used in replacement strategies to control dental caries. Early studies with variants of S. salivarius showed that less cariogenic bacteria could reduce caries by preempting colonization by more virulent S. mutans strains (36). Implantable strains of S. mutans that are deficient in lactate dehydrogenase (16) and express a heterologous alcohol dehydrogenase gene also appear to be potentially effective for controlling caries by replacement therapy. Other approaches have explored expressing glucanase enzymes in commensal oral streptococci to degrade the adhesive plaque glucans produced by mutans streptococci (21). Use of alkali-generating bacteria should now also be considered potentially useful for replacement therapy. A strain of S. mutans was selected for the present study because of the compatibility with the rat caries model. Future studies could be done with noncariogenic or less cariogenic resident plaque streptococci, since most oral streptococci can be genetically manipulated. In fact, we have constructed ureolytic strains of S. gordonii, showing that the use of other bacterial hosts that colonize the teeth is certainly feasible. One clear advantage that ammonia-producing plaque bacteria have over other approaches is that they may favorably modify the supragingival plaque ecology by fostering an environment that inhibits the emergence of a cariogenic flora, rather than targeting a specific pathogenic agent or perturbing the normal metabolic or physiologic pathways, which could compromise competitive fitness.

A key consideration in the utility of replacement strains is their ability to compete with endogenous strains and the fact that ablation of endogenous activities or introduction of foreign genes can compromise the fitness of a bacterium. Arguably, the introduction of genes which produce ammonia from urea into oral streptococci may instead create strains which are better able to compete than the parent. First, urea can diffuse through the membrane, so that no energy is required to transport this compound. Once cleaved, the ammonia can neutralize the cytoplasm, raising the intracellular pH and creating a diminished requirement for the organisms to spend ATP to extrude protons. Recently, we have also found that the ureolytic oral bacteria S. salivarius (Y. M. Chen, C. A. Weaver, and R. A. Burne, submitted for publication) and A. naeslundii (25) can efficiently use the ammonia from urea as a source of nitrogen. Thus, recombinant ureolytic bacteria may gain a competitive advantage from a bioenergetic standpoint because they can access a source of nitrogen unavailable to other oral bacteria.

The economic importance of dental caries has resulted in a large research effort focusing on preventing caries formation. Fluoridation of water and dental hygiene products are effective but not completely so. Other approaches to control caries include immunization and use of antimicrobial agents. This communication provides compelling evidence to support the idea that increasing the alkali-generating capacity of plaque, perhaps by using recombinant ureolytic bacteria, may help to protect against caries formation. Modulation of the alkali-generating potential of dental plaque may also have great potential because it does not target particular etiologic agents and instead may work by fostering an ecologically healthy oral environment, which naturally controls the emergence of pathogenic microorganisms.