|Home | About | Journals | Submit | Contact Us | Français|
Streptococcus gordonii and Streptococcus mutans avidly colonize teeth. S. gordonii glucosyltransferase (GtfG) and amylase-binding proteins (AbpA/AbpB), and S. mutans glucosyltransferase (GtfB), affect their respective oral colonization abilities. We investigated their interrelationships and caries association in a rat model of human caries, examining the sequence of colonization and non- vs. high-sucrose diets, the latter being associated with aggressive decay in humans and rats. Virulence-characterized wild-types of both species and well-defined mutants of S. gordonii with interrupted abpA and gtfG genes were studied. While both S. gordonii and S. mutans were abundant colonizers of rat’s teeth in the presence of either diet, if inoculated singly, S. mutans always out-competed S. gordonii on the teeth, independent of diet, strain of S. mutans, simultaneous or sequential inoculation, or presence/absence of mutations of S. gordonii’s abpA and gtfG genes known to negatively or positively affect its colonization and to interact in vitro with S. mutans GtfB. S. mutans out-competed S. gordonii in in vivo plaque biofilm. Caries induction reflected S. mutans or S. gordonii colonization abundance: the former highly cariogenic, the latter not. S. gordonii does not appear to be a good candidate for replacement therapy. These results are consistent with human data.
The mutans streptococci are strongly cariogenic, while other tooth-colonizing oral streptococci are thought to be either non-cariogenic or minimally so (Hamada and Slade, 1980; Loesche, 1986; Tanzer et al., 2001a). Thus, there is interest in “bacterial interference therapy” to pre-empt establishment or replace mutans streptococci from tooth surfaces by non-cariogenic tooth-colonizing bacteria. This concept is not new (Aly et al., 1982; Sprunt and Leidy, 1982; Tanzer et al., 1985a,b; Hillman et al., 2000). The non-mutans species—formerly classified as Streptococcus sanguis, which now includes the species Streptococcus sanguinis, Streptococcus oralis, and Streptococcus gordonii—are found in high numbers on supragingival tooth surfaces. S. sanguis had been suggested as a replacement organism, since it has been negatively associated with caries (Krasse et al., 1968; de Stoppelaar et al., 1969; Loesche and Syed, 1973; Loesche et al., 1975; Swenson et al., 1976). While recent in vitro studies suggest that S. gordonii should be capable of such replacement (Kreth et al., 2008; Liu et al., 2011), other in vivo microbiological assessments of carious human teeth continue to suggest otherwise (Gross et al., 2010). We know of no human populations free of S. gordonii and mutans streptococci that could serve as “clean” study subjects. We therefore tested, in rats initially free of these bacteria, the hypotheses that S. gordonii can colonize and persist on the dentition despite simultaneous, previous, or subsequent inoculation by S. mutans strains, and vice versa. We also tested the dependency of these relationships upon dietary sucrose. We also tested abilities of mutants of S. gordonii deficient in genes that encode colonization determinants to affect this relationship.
Bacterial strains, plasmids, and primers used in this study are presented in Appendix 2.
Details of the rationales and construction of S. gordonii mutant strains of Challis CH1 are presented in Appendix 2. Briefly, a mutant of the wild-type S. gordonii Challis stain was constructed having a kanamycin-resistant phenotype to facilitate recovery and quantification during animal experiments. A chromosomal integration site was selected empirically, based upon the gene order of convergent open reading frames at loci SGO_2076 encoding a putative phage integrase and SGO_2077 encoding a hypothetical protein with no identified conserved domains and no known function. Thus, S. gordonii strain KS1 was generated by replacement of SGO_2076 and a portion of the downstream intergenic region (bp 2140291-2141556; GenBank CP000725) with an aphIII determinant encoding kanamycin resistance.
Strain KS1gtfG was constructed with the 2.5-kb dsDNA Xho1/Sst1 fragment of pKS1 carrying the aphIII determinant between SGO2076 flanking regions. In this case, the linear fragment was used to transform the CH1 derivative strain CH1ΔgtfG in which gtfG encoding the single S. gordonii glucosyltransferase enzyme had been deleted (Banas et al., 2007).
abpA-deficient mutants in KS1 or KS1gtfG were constructed by disruption of abpA by an integrated tetM gene that encodes tetracycline resistance (EMBL accession no. X56353). After sequence verification, selected transformants were designated strain KS1abpA, in which abpA was disrupted, and strain KS1gtfG/abpA, a double-mutant with both a gtfG deletion and abpA disruption, respectively.
All mutant strains were compared with S. gordonii parental strain Challis for similar colony morphology on blood agar (BA) and Mitis-Salivarius Agar (MSA), Gram stain, and biochemical/metabolic profiles. No differences were noted in growth rates between the parental and mutant strains. To compare amylase-binding activity between wild-type and mutant strains, we used a Western blot amylase ligand-binding assay as previously described (Chaudhuri et al., 2007, 2008). Gtf activity was qualitatively determined in 12% polyacrylamide gels as previously described (Vickerman and Minick, 2002).
Experiments with rats were approved by the IACUC, University of Connecticut Health Center. Eight in vivo experiments were performed. For details, see Appendix 2. Their designs are depicted in Figs. 1A and and1B.1B. For each experiment, 21-day-old weanlings were randomly assigned from multiple litters born on the same day. In the simultaneous inoculation protocol (A) (Tanzer et al., 1985a), one day after provision of test diet, rats were either inoculated (Experimental Day 0) simultaneously with equal numbers of stipulated S. gordonii strain, its isogenic derivative, and an S. mutans strain; inoculated with one or the other of them; or remained un-inoculated. In the sequential inoculation protocol (B), by contrast, one day after provision of test diet, weanlings were either inoculated (Experimental Day 0) with one of the test strains or remained un-inoculated. Ten days later, inoculation of the potentially competing species was given both to a previously inoculated group and to a group previously un-inoculated. Some groups remained un-inoculated. In this way, sequentially doubly- and singly-inoculated animals, as well as un-inoculated ones, could be studied.
Cross-contamination among rat groups, reversion of mutants to parental phenotypes, indigenous or unintended extraneous infection, or adverse health effects did not occur. Single pure-culture inoculants were always recovered in high numbers from swabbed and sonicated teeth. However, for doubly inoculated animals, whether simultaneously or sequentially inoculated by S. gordonii and S. mutans strains, recoveries were always dominated by S. mutans on sonicated molars, independent of sequence of inoculation and sucrose- or sucrose-free diet. Caries scores reflected S. mutans predominance.
We previously studied mutants of S. gordonii that did not express abpA or gtfG derived from a spontaneous streptomycin-resistant mutant of S. gordonii Challis, strain Challis-S; these strains were used in colonization and caries studies in the rat (Tanzer et al., 2001b, 2003, 2008). Now we examined their abilities to compete with S. mutans strains for colonization of rats’ teeth. When the genome sequence of Challis CH1 was determined, we constructed strain KS1 with a chromosomally integrated kanamycin resistance gene for selection in animal model studies to verify and extend our initial experiments. Use of strains derived from strain KS1 minimized the possibility that previous observations were influenced by pleiotropic effects of the undefined mutation(s) that conferred streptomycin resistance in strain Challis-S (Appendix 2).
Figs. 2A--2C2C show colonization by S. gordonii strains Challis-S, its abpA-defective mutant Challis-ST, and S. mutans 10449S, inoculated in equal doses into sucrose-eating rats, either singly or simultaneously (Fig. 1A), and respective caries scores. Molars became heavily colonized (Figs. 2A, ,2B)2B) by S. mutans or S. gordonii strains if inoculated singly, and the Challis-ST mutant tended to establish in slightly higher numbers (p = 0.116) than its parent, as previously reported (Tanzer et al., 2003). But when S. mutans 10449S was simultaneously inoculated with either S. gordonii strain, it greatly outnumbered them in absolute (Fig. 2A) counts recovered (p < 0.001). As a percentage of recoverable flora (Fig. 2B), both of these species dominated the flora, but when inoculated simultaneously, S. mutans dominated, while S. gordonii virtually disappeared (p < 0.001). Caries scores (Fig. 2C) reflected colonization status: high S. mutans, with high scores (all comparisons, p < 0.001), while high S. gordonii only slightly augmented scores of un-inoculated rats attributable to their indigenous oral flora (p < 0.02). There was no evidence that S. gordonii co-colonization inhibited S. mutans 10449S cariogenicity.
We then hypothesized that prior colonization of teeth by S. gordonii would prevent their displacement by subsequent S. mutans inoculation, and vice versa. Use of the sequential inoculation protocol (Fig. 1B) and the same inoculants (Figs. 2D--2F)2F) showed that both S. mutans and S. gordonii were very numerous on molars if they were inoculated singly. They were less numerous if either was inoculated at day 10, but at this time the abpA mutant was established in higher numbers than its progenitor (all, p < 0.001). By contrast, when initial S. gordonii inoculation was followed at day 10 by 10449S inoculation, the 10449S displaced it, on absolute (Fig. 2D) and percentage (Fig. 2E) bases (p < 0.001). If rats had been first inoculated by S. mutans 10449S, and 10 days later by either S. gordonii strain, S. mutans was not displaced (p < 0.001), and S. gordonii was established poorly. The percentage of total recoverable flora represented by the sum of S. mutans and S. gordonii was very high and virtually the same for each inoculated group. S. gordonii strains caused minor augmentation of caries scores (Fig. 2F) (p < 0.05), while S. mutans caused great augmentation (p < 0.001). Inoculation by S. mutans at day 10 resulted in slightly lower caries scores than inoculation at day 0 (p < 0.001), probably reflecting reduced S. mutans challenge-time. There was neither inhibition nor augmentation of cariogenicity of S. mutans by prior colonization by S. gordonii.
The cariogenic and colonization superiority of S. mutans was hypothesized to depend on its well-known production of a variety of extracellular glucans uniquely from sucrose, its potent acidogenesis from sucrose, and other less-defined sucrose-supported functions. To test this, we repeated the simultaneous inoculation protocol using a sucrose-free diet, substituting powdered cornstarch (Appendix 2) for powdered sucrose. For this experiment, we used parental strain S. gordonii KS1 and its isogenic abpA-deficient mutant (Appendix 2). While KS1 colonized better (p < 0.001) than KS1abpA and S. mutans 10449S on both absolute and relative count bases (Figs. 3A, ,3B),3B), all inoculants were recovered in high numbers and percentages from teeth. However, neither S. gordonii strain displaced 10449S or became dominant in its presence (p < 0.001), although KS1abpA colonized better than its KS1 parent (p < 0.001). S. mutans 10449S was recovered in higher numbers and percentages from rats also inoculated by KS1abpA rather than KS1 (p < 0.001), suggesting an interactive effect associated with loss of S. gordonii AbpA expression. Importantly, the competitive advantage of S. mutans appeared independent of sucrose.
The caries scores (Fig. 3C) were very low, as expected in rats eating a high-cornstarch/sucrose-free diet (Tanzer et al., 2003). While total lesion scores were statistically higher for the S. mutans-inoculated groups than for either un-inoculated or S. gordonii-inoculated groups (p < 0.05), they were low by comparison with those observed in experiments with a high-sucrose diet.
While results indicating superiority of S. mutans in competition with S. gordonii seemed convincing, they could have been influenced by complex interactive effects of AbpA with the Gtfs of both/either S. gordonii and/or S. mutans, viz. with GtfG and GtfB, or by pleiotropic effects resulting from undefined antibiotic resistance mutations. It was also possible that S. mutans strain 10449S might have unusual potency in competition with S. gordonii. Therefore, the simultaneous inoculation protocol using gtfG- and abpA-mutants of S. gordonii strain KS1 (Appendix 2), their progenitors, and a different S. mutans strain (LT11) was performed in rats fed the high-sucrose diet. Again, single inoculants became very numerous on teeth, but if any S. gordonii strain had been simultaneously inoculated with S. mutans LT11, only the latter was found in very high numbers and percentages on molars (Figs. 3D, ,3E)3E) (all, p < 0.001). Also, KS1gtfG colonized less well than its progenitor (absolute, p = 0.027; percent, p < 0.001). The percentage of total recoverable flora constituted by LT11 plus KS1gtfGabpA double-mutant was lower than for all other doubly inoculated rat groups (p < 0.001), suggesting complex in vivo interactive effects between AbpA and GtfG of S. gordonii and/or GtfB of S. mutans, as previously reported in vitro (Chaudhuri et al., 2007). Nonetheless, in doubly inoculated rats, S. mutans LT11 vastly predominated on the surfaces of the teeth. The previously observed superior competitiveness of S. mutans strain 10449S, thus, did not appear either unique or attributable to pleiotropic effects.
The caries scores (Fig. 3F) reflected the S. mutans LT11 colonization of the rats (p < 0.001); S. gordonii strains, again, caused little or no augmentation of caries scores (p > 0.05). Double inoculation by S. mutans LT11 plus KS1 or LT11 plus KS1abpA resulted in increased total caries score (both, p < 0.005), but double inoculation by LT11 plus either KS1gtfG or KS1gtfGabpA did not.
Finally, to confirm that the advantage of S. mutans over S. gordonii was not sucrose-dependent, using KS1 and its site-specific constructs, KS1gtfG and KS1gtfGabpA, we carried out the simultaneous inoculation protocol using the cornstarch, sugar-free diet in which extracellular glucan synthesis was impossible. All singly inoculated strains heavily colonized the molar teeth. When any of the KS1 strains were simultaneously inoculated with S. mutans 10449S, the latter dominated the numbers and percentages on teeth (all comparisons, p < 0.002) (Figs. 4A, ,4B).4B). KS1 and its gtfG and abpA deletion mutants colonized comparably (p > 0.05). However, the double-mutant KS1gtfGabpA colonized less well than KS1 (p = 0.027) on an absolute count basis (Fig. 4A). The percentage of total recoverable flora represented by the sum of S. mutans and S. gordonii ranged from 70 to 100%. Only the combined inoculum of S. mutans 10449S + S. gordonii KS1gtfG evidenced any decreased ability to colonize (Fig. 4B) (p < 0.001). Remarkably, the S. mutans strain 10449S out-competed all S. gordonii site-specific constructs in rats fed the sugar-free, high-cornstarch diet. The caries scores were minimal, as expected for this sucrose-free diet (data not shown).
We addressed the relationships in vivo of 2 plaque biofilm streptococci, S. mutans and S. gordonii, as a function of the presence/absence of dietary sucrose, and explored possible pre-emption or displacement of S. mutans on teeth through bacterial interference. A well-defined rat dental caries model was used, because much literature shows its relevance to predicting behavior of oral flora on human teeth and predicting results of caries-preventive intercessions (Tanzer, 1986), and because healthy mutans-free and S. gordonii-free rats maintained in our barrier-breeding facility provided “clean” control un-colonized rats, harboring normal rodent gut flora. We know of no human populations free of S. gordonii or other amylase-binding oral streptococci and mutans streptococci that could serve as study subjects, were one to contemplate such studies.
This report describes studies that revealed consistent patterns: (A) No strain of S. gordonii (either wild-type or mutants deficient in gtfG, abpA, or gtfGabpA) could compete with S. mutans for colonization in the plaque biofilm; (B) failure to compete is evident when these bacteria are inoculated in very large, equal numbers, either simultaneously or sequentially, after a 10-day period allowed for their establishment; (C) superiority of S. mutans’ colonization ability is not unique to strain 10449S; it is also apparent for S. mutans strain LT11; and (D) the superiority of S. mutans colonization is independent of whether rats are fed a high-sucrose diet or an otherwise identical sucrose-free, high-starch diet. The effect, therefore, cannot be attributed to the production of extracellular glucans (Tanzer et al., 1974; Kuramitsu, 1993; LeBlanc, 1994) by S. mutans or to other sucrose-dependent phenomena. It must have another, as-yet-undefined, basis(es).
In the presence of a sucrose diet, S. mutans colonization induces high caries lesion scores for all categories of surfaces of the crowns of teeth. Induction of lesions by S. gordonii wt and various mutants, when inoculated singly, is modest, and sometimes lesion scores are not significantly greater than those for un-inoculated control rats. In the presence of a starch diet, caries lesion induction associated with inoculation by either S. mutans or S. gordonii strains is low: for S. mutans, statistically significant, but biologically unimpressive; for S. gordonii, virtually none, consistent with previous observations (Tanzer, 1979; Tanzer et al., 2001a, 2003).
A recent study using an in vitro biofilm model indicated that hydrogen peroxide produced by S. gordonii inhibited S. mutans growth, suggesting that it should inhibit colonization of in vivo dental plaque biofilm by S. mutans (Kreth et al., 2008). However, the present in vivo plaque biofilm studies in rats with normal salivary secretion showed no inhibition of S. mutans by S. gordonii, consistent with previous literature on human dental plaque (Krasse et al., 1968; de Stoppelaar et al., 1969; Folke et al., 1972; Loesche and Syed, 1973; Loesche et al., 1975; Staat et al., 1975; van Houte and Duchin, 1975; Swenson et al., 1976; Emilson and Westergren, 1979; Tanzer et al., 2001a). Furthermore, previous reports of high peroxidase levels in saliva of rats and humans (Nickerson et al., 1957; Morioka et al., 1969; Ihalin et al., 2006) also suggest the impotence of S. gordonii-produced peroxide to inhibit S. mutans. Recent observations on advanced caries lesions in young human teeth, using bacterial 16S sequence analysis methods, are consistent with the present data, and indicate that S. gordonii diminishes greatly in caries-associated plaque biofilm, while S. mutans persists (Gross et al., 2010).
In summary, S. mutans out-competes S. gordonii on teeth, independent of host sucrose diet, sequence of inoculation, strain of S. mutans, and the presence or absence of mutations of abpA and gtfG genes of S. gordonii implicated in its colonization and interactions with S. mutans glucans. The mechanisms that explain these observations are yet unknown. S. gordonii does not appear to be a good candidate for replacement therapy against caries, even though, by itself, it is an excellent colonizer of the teeth.
The authors thank Jessica Kilham and Jeff Eckleberry for excellent library and graphics assistance, respectively.
This work was supported by grants DE-09838 and DE-07034 from the US National Institutes of Health.
The author(s) declare no potential conflicts of interest with respect to the authorship and/or publication of this article.
A supplemental appendix to this article is published electronically only at http://jdr.sagepub.com/supplemental.