In Vivo Colonization with Candidate Oral Probiotics Attenuates Colonization and 1 Virulence of Streptococcus mutans

24 A collection of 113 Streptococcus strains from supragingival dental plaque of caries-free 25 individuals were recently tested in vitro for direct antagonism of the dental caries pathogen 26 Streptococcus mutans , and for their capacity for arginine catabolism via the arginine deiminase 27 system (ADS). To advance their evaluation as potential probiotics, twelve strains of commensal 28 oral streptococci with various antagonistic and ADS potentials were assessed in a mouse model 29 for oral (i.e., oral mucosal pellicles and saliva) and dental colonization under four diets (healthy 30 or high-sucrose, with or without prebiotic arginine). Colonization by autochthonous bacteria was 31 also monitored. One strain failed to colonize, whereas oral colonization by the other eleven 32 strains varied by 3 log units. Dental colonization was high for five strains regardless of diet, six 33 strains increased colonization with at least one high-sucrose diet, and added dietary arginine 34 decreased dental colonization of two strains. Streptococcus sp. A12 (high in vitro ADS activity 35 and antagonism) and two engineered mutants lacking the ADS (∆ arcADS ) or pyruvate oxidase- 36 mediated H 2 O 2 production (∆ spxB ) were tested for competition against S. mutans UA159. A12 37 wild type and ∆ arcADS colonized only transiently, whereas ∆ spxB persisted, but without altering 38 oral or dental colonization by S. mutans . In testing four additional candidates, S. sanguinis 39 BCC23 markedly attenuated S. mutans’ oral and dental colonization, enhanced colonization of 40 autochthonous bacteria, and decreased severity of smooth surface caries under highly cariogenic 41 conditions. Results demonstrate the utility of the mouse model to evaluate potential probiotics, 42 revealing little correlation between in vitro antagonism and competitiveness against S. mutans in 43 vivo . Our results demonstrate in vivo testing of potential oral probiotics can be accomplished 47 and can yield information to facilitate the ultimate design and optimization of novel anti-caries 48 probiotics. We show human oral commensals associated with dental health are an important 49 source of potential probiotics that may be used to colonize patients under dietary conditions of 50 highly varying cariogenicity. Assessment of competitiveness against dental caries pathogen 51 Streptococcus mutans and impact on caries identified strains or genetic elements for further 52 study. Results also uncovered strains that enhanced oral and dental colonization by 53 autochthonous bacteria when challenged with S. mutans , suggesting cooperative interactions for 54 future elucidation. Distinguishing a rare strain that effectively compete with S. mutans under 55 conditions that promote caries further validates our systematic approach to more critically 56 evaluate probiotics for use in humans.


59
Dental caries remains a highly prevalent disease and global health problem. Caries 60 results from repetitive and/or prolonged demineralization of tooth enamel driven by exposure to 61 low pH from organic acids produced by acidogenic oral bacteria during fermentation of dietary 62 carbohydrates (1). Counterbalancing acid production and demineralization are alkalinization of 63 oral biofilms by certain commensal bacteria, which in addition to salivary buffering and delivery 64 of supersaturated calcium and phosphate ions, promotes enamel remineralization (1). 6 of host factors, such as specific salivary constituents, on colonization by commensals and the 115 induction of caries by oral pathogens (43-45). 116 To identify and test putative probiotic strains in the prevention and treatment of caries we 117 have taken a systematic approach, targeting commensal streptococci which represent an 118 abundant genera found in healthy dental plaque (28). First, we recently isolated 113 119 Streptococcus strains representing ten species from supragingival dental plaque of individuals 120 free of clinical lifelong dental caries. Each strain was tested in vitro for two specific phenotypes: vivo. Also, estimates of recovered total bacteria were determined using a novel qPCR assay 153 targeting conserved regions of the ubiquitous single copy gene, rpsL (30S ribosomal protein S12) 154 (48), rather than by CFU on blood agar plates (44, 45); the latter may overlook bacteria rendered 155 non-viable during molar sonication. Subtraction of recovered genomes of inoculated strains 156 from total recovered genomes thus estimates the population of recovered murine autochthonous 8 their drinking water, ad libitum (high-sucrose diet). We reasoned that a probiotic must 161 effectively colonize the oral soft and hard tissues irrespective of the cariogenicity of an 162 individual's diet, as diets will vary among humans, and from day-to-day for a given individual. 163 Two other diets were created by addition of 1.5% arginine to each primary diet to determine 164 whether arginine provided as a prebiotic influence's colonization by a candidate probiotic strain. 165 An increase in colonization with added dietary arginine would suggest that some minimum level 166 of ADS activity may be required for a strain to more effectively counteract acids produced by 167 members of the autochthonous bacterial population, and may therefore require arginine as a 168 prebiotic to be competitive against S. mutans. Conversely, decreased colonization with added 169 arginine indicates that simultaneous use of arginine as a prebiotic may be contraindicated. Diets 170 were based on the nutritionally balanced diet, AINS-93G (50), rather than the commonly used 171 cariogenic diet, Diet 2000, which is nutritionally deficient in vitamins and minerals that likely 172 influence its cariogenic properties (51). Constituents of each diet are given in Table 1. extents. Oral colonization, as assessed from oral swabs, ranged from 10 2 genomes for S. mitis 178 BCC15 to 10 5 genomes for BCC32 and A12. There was only a single example in which oral 179 colonization was significantly increased by added arginine (i.e., S. gordonii BCC32), whereas S. 180 mitis BCC08 displayed decreased colonization. Though, in both cases, these differences were 181 observed only with the high-sucrose diets and were inconsistent among swabs at experimental 182 days 10 and 20. Interestingly, S. gordonii BCC32 has very high ADS activity compared to 9 nearly undetectable ADS activity in S. mitis BCC08 (Fig. 1S). However, oral colonization by all 184 other strains with either similar or higher ADS activity were not impacted by arginine.

185
Collective results suggest oral colonization is mostly independent of added dietary arginine and a 186 strain's ADS activity.

187
The only example where increased sucrose showed a trend towards impacting oral 188 colonization was for A12, where recoveries from oral swabs were consistently at least 5-fold 189 higher in the high-sucrose diet versus the average diet at day 10 and day 20. In only a few cases 190 did oral colonization under the same diet change significantly from experimental day 10 to day 191 20, suggesting oral colonization mostly achieved steady-state levels by day 10.

192
Dental colonization: Five of the twelve strains exhibited relatively high levels of dental 193 colonization with all four diets, ranging from about 10 5 to slightly above 10 6 genomes (i.e., S. swabs were in nearly all cases consistent among the four diets and between experimental day 10 10 and day 20, suggesting resident bacteria are at steady-state levels by experimental day 10, 207 regardless of diet. With respect to colonization of molars, mouse commensals also displayed few 208 and relatively minor differences between diets, especially when comparing a high-sucrose diet to 209 its respective average diet. A notable exception were mice inoculated with S. intermedius A3 in 210 which recoveries were extremely high from mice fed the average and high-sucrose diets, but 211 were dramatically reduced by addition of arginine to each diet, mirroring in large part 212 colonization by S. intermedius A3. Furthermore, mouse commensals recovered from dental 213 biofilms of mice inoculated with S. cristatus A52, which failed to colonize, were consistent 214 across diets (10 6 to 2 x 10 6 genomes) and not significantly higher than that seen with mice 215 challenged with the other eleven colonizing strains of human commensals (10 5 to 10 6 genomes).

216
These results suggest the total population of resident dental bacteria were only moderately 217 impacted when mice were infected with a human commensal, regardless of diet. In all cases,

223
The novel strain Streptococcus sp. A12, a relative of Streptococcus australis based on 224 comprehensive phylogenomic analyses (9), has been studied in vitro due to its high levels of 225 antagonism against S. mutans and ADS activity (Fig. S1) (5, 7, 9, 52). One mechanism used by 226 A12 to inhibit growth of S. mutans, in vitro, is H 2 O 2 production through pyruvate oxidase, 227 encoded by spxB (9). Also, production of ammonia from arginine via the ADS is predicted to 228 help counteract oral biofilm acidification by S. mutans to promote pH homeostasis, in vivo, thus 11 creating an environment less favorable for the emergence of aciduric organisms. We therefore 230 set out to determine whether A12 interferes with colonization by S. mutans or colonization of 231 autochthonous bacteria, in vivo, and if either outcome was affected by elimination of the ADS or 232 pyruvate oxidase, using mutant strains ∆arcADS, lacking ADS activity or ∆spxB, respectively.

233
Colonization: Before testing A12 and its mutant strains, we first determined whether 234 each mutant colonizes the mouse oral cavity and dentition. Ultimately, we wanted to examine 235 A12 and its mutants under strongly cariogenic conditions to construct a rigorous test of each 236 strain's colonization and potential competitiveness against S. mutans, which would provide 237 insights into the relative contribution of antagonism and pH moderation to competitive fitness. 238 We therefore examined colonization with mice fed the high-sucrose diet with 1.5% arginine and 239 4% sucrose water. In subsequent competition experiments, added arginine was then posited to 240 further amplify any deficits in competition associated with deletion of the ADS. In these initial 241 colonization experiments, we used the same timeline for colonization as in the previous set of 242 experiments with the twelve strains of human commensals ( Fig. 2A). As shown in Fig. 2B, oral 243 colonization by the ADS mutant was not significantly different than A12 WT. However, dental 244 colonization by the ADS-deficient mutant was approximately 80% less then A12 WT, but still 245 sufficient for further testing. In contrast, both oral and dental colonization of A12 ∆spxB were 246 similar to A12 WT. Oral and dental colonization of mouse commensals were not significantly 247 different between the WT and each mutant group. The mean increase in body weights between 248 groups were also not significantly different (Fig. S2B).

249
Colonization by S. mutans: In testing human commensals in competition experiments, 250 our strategy was to first establish colonization by the commensal, followed by oral inoculations 251 with S. mutans UA159 and monitoring of colonization. We reasoned that first establishing the 12 human commensal likely reflects future clinical applications of probiotics in which patients at 253 high caries risk would first undergo comprehensive removal of supragingival dental biofilms, 254 followed by administration of a probiotic immediately thereafter, and subsequent self-255 administration of probiotic at periodic intervals after brushing of teeth and/or use of an oral 256 antiseptic. Because our initial oral colonization results indicated human oral commensals 257 appeared to reach a steady-state by experimental day 10 (swab 1) or earlier, we scheduled 258 subsequent inoculations with S. mutans to start at experimental day 7. A concern with this 259 strategy was that mice would not be inoculated with S. mutans until 6.5 to 7 weeks of age. It is 260 established that colonization of rodents by S. mutans and subsequent development of dental 261 caries is greatest when inoculated before weaning at age 21 days, when tooth eruption is in its 262 early stages, and that colonization and caries then declines markedly with age (43, 53). In 263 contrast, BALB/c mice greater than 8 weeks of age were able to be colonized with S. mutans 264 after antibiotic suppression of the oral microbiota (41, 42) and to induce measurable caries (54).

265
It was thus necessary to first establish in our model how well S. mutans colonized and induced 266 caries, but to also determine whether colonization and induction of caries by S. mutans decreased 267 when inoculations were initiated at experimental day 7, compared to day 0. We used the high-   conditions. Importantly, the incidence and severities of smooth surface and sulcal caries were 281 highly similar after inoculating mice with S. mutans UA159 on experimental day 7 compared to 282 day 0 ( sucrose with 5% sucrose water (43-45). There also were no differences in the mean increase in 286 body weights between the two groups during the experiment (Fig. S2C). starting S. mutans inoculations, but with more frequent oral swabbing to monitor oral 294 colonization. As explained above, mice were fed the high-sucrose diet with 1.5% added arginine 295 and 4% sucrose water. A group in which initial inoculations were without added bacteria (mock 296 inoculations) was included as a control group for S. mutans alone. As shown in Fig. 4B, oral and 297 dental colonization by S. mutans in the mock group were each robust, whereas oral colonization 14 by murine commensals were lower but persistent. Similar to results of the validation experiment 299 (Fig. 3B, Inoculated Days 7-11 group), dental colonization by mouse commensals was about 2 300 logs greater than its oral colonization and comparable to that of S. mutans. There were distinct 301 differences in colonization between the three strains of A12. First, A12 WT was undetectable in 302 swabs at experimental day 27 and barely detectable a day later in dental biofilms. A12 ∆arcADS 303 appeared even less competitive against S. mutans, as oral colonization was undetectable a week 304 earlier than the WT and undetectable in molar biofilms. In stark contrast to A12 WT, oral 305 colonization by S. A12 ∆spxB was unexpectantly persistent, with recoveries comparable to its 306 recovery from dental biofilms. Oral colonization by murine autochthonous bacteria in this group 307 was also persistent and similar to the mock group, while recoveries of mouse commensals in 308 both the A12 WT and A12 ∆arcADS groups were erratic at times, but nonetheless comparable or 309 slightly higher than in the mock group by experimental day 27, respectively. Importantly, none 310 of the A12 strains had an impact on dental colonization by S. mutans. Furthermore, mean 311 increase in body weights between the four groups were not significantly different (Fig. S2D). inclusion of a mock group were the same as with the A12 strains. As shown in Fig. 5B, oral and 322 dental colonization by S. mutans in the mock group were at high levels compared to murine 323 autochthonous bacteria, although oral colonization of murine commensals was persistent.

324
Oral colonization by S. gordonii BCC32 progressively decreased to low levels, as did 325 murine autochthonous bacteria. Noteworthy, though, was the approximate 10-fold increase in 326 oral colonization of mouse commensals following initial inoculations with S. gordonii BCC32, 327 as well as the more than 3-fold decrease in S. mutans compared to the mock group at 328 experimental day 27. A moderate level of S. gordonii BCC32 was recovered from dental 329 biofilms. Compared to the mock group the recovery from dental biofilms of S. mutans was 330 unaltered, whereas autochthonous bacteria were enhanced. Importantly, the very low recovery 331 (about 500 genomes) of mouse commensals from swabs on day 27 (swab 5) compared to nearly 332 10 6 genomes recovered from mandibular molars on day 28, demonstrates oral swabs capture 333 bacteria primarily from non-dental biofilms, most likely from saliva, epithelial biofilms (i.e., 334 mucosal pellicles) and papillary groves of the tongue.  The most striking results were seen with S. sanguinis BCC23. Its oral colonization was 348 stable, then increased during the final week. Its presence was associated with significant and 349 consistently higher levels of murine autochthonous bacteria when compared to the mock group.

350
Conversely, the levels of oral colonization of S. mutans were depressed initially, but increased 351 during the final week in conjunction with S. sanguinis BCC23 and autochthonous bacteria. More 352 importantly, dental colonization of S. mutans was nearly 4-fold lower than in the mock group, 353 and also markedly lower than autochthonous bacteria. Dental colonization by S. sanguinis 354 BCC23 was greater than 10 5 genomes, equivalent to S. mutans. Furthermore, mouse 355 commensals in dental biofilms were significantly greater than in the mock group. There were 356 also no differences in the mean increase in body weight between each of the five groups during 357 the experiment (Fig. S2E).

384
Oral and dental colonization in relation to dietary sucrose: In the context of dental 385 caries, dietary carbohydrates are a critical determinant of the oral microbiome and oral health, 386 due in large part to the ability of S. mutans to rapidly utilize sucrose to produce a structural 387 matrix of insoluble glucans that can greatly enhance the cariogenic potential of oral biofilms (1).

388
The high cariogenicity of the high-sucrose diet used in this study was confirmed, as smooth 389 surface and sulcal caries of mice inoculated with S. mutans alone were comparable to prior experiments that incorporated Diet 2000, containing 56% sucrose and 5% sucrose water (43). In 391 contrast, the average diet contained nearly 70% less sucrose, and mice were supplied with sterile 392 drinking water without added sucrose. Omission of sucrose from the drinking water further 393 reduces the cariogenicity of a diet, likely by decreasing the frequency of exposure of S. mutans to 394 sucrose. For example, in a pilot study using our earlier mouse caries model with mice fed Diet 395 2000, the incidences of total smooth surface caries, total sulcal caries and recovery of S. mutans 396 from molar sonicates were 84%, 62% and 49% less, respectively, in mice provided sterile water, 397 compared to mice provided with 5% sucrose water (Culp, unpublished observations). Therefore, 398 in light of the differences in cariogenicity between the average and high-sucrose diets, the ability 399 of the great majority of candidate probiotics to colonize the oral cavity and dental biofilms at 400 relatively moderate to high levels, regardless of the level of dietary sucrose, is considered a 401 reflection of their adaptation to dental biofilms of caries-free individuals and is a highly desirable 402 attribute, as diet will likely vary in cariogenicity among patients taking probiotics. In addition, 403 colonization of non-dental sites in the oral cavity such as oral epithelium and within papillary 404 groves of the tongue potentially creates a reservoir for persistent recolonization of dental 405 biofilms. Of particular note was the inexplicable extreme increase in molar colonization by S. 406 intermedius A3 with the high-sucrose diet compared to the average diet. Additional 407 investigations are required explain this phenotype.  and 2-fold greater, respectively, than that for S. sanguinis BCC23. S. mitis BCA12 and S. 499 sanguinis BCA8 have only slightly lower levels than S. sanguinis BCC23 (see Fig. S3).

500
Furthermore, the failure of strains other than S. sanguinis BCC23 to affect dental colonization by 501 S. mutans is not related to their ability to colonize dental biofilms, as each strain colonized molar 502 teeth at levels higher or slightly lower than S. sanguinis BCC23 when mice were fed the same 503 diet (Fig. S3). These combined results strongly suggest antagonism against S. mutans, in vitro, 23 which is highly dependent on growth conditions, does not correlate with competitiveness in vivo, 505 at least under the conditions referenced herein, and further warrant that caution should be 506 exercised when extrapolating in vitro phenotypes to the in vivo environment. In vivo testing of 507 competitiveness of a candidate probiotic thus represents an important discriminating assessment 508 to identify strains for further study, especially in light of the phenotypic heterogeneity displayed 509 by oral streptococci is not species specific, nor always reflected by genotype (18). Initial  Collective results demonstrate human oral commensals strongly associated with dental 518 health are generally well adapted to colonize both the soft and hard tissues of mice under highly 519 cariogenic and healthier dietary conditions, and identifies a highly attractive probiotic candidate, 520 S. sanguinis BCC23. Health-associated dental isolates from humans thus represents a source of 521 putative probiotic strains with the potential to colonize dental and oral biofilms of patients, and persistent colonization of epithelial and/or dental biofilms. The model is further amenable to 528 explore the effectiveness of the dose and frequency of administration of a probiotic or prebiotic. 529 We therefore demonstrate the utility of in vivo assessments to more stringently evaluate the oral 530 fitness of candidate strains to help facilitate the rational design and optimization of novel 531 probiotic strategies to target microbial ecology in protection of supra-gingival dental surfaces.   Table S3.     Fig. 4B (1) and Fig. 5B (2) that included colonization by S. mutans and mouse commensals.

911
Results are mean ± SE (A, n = 10 mice per group; B, n =    Colonization by S. mutans and mouse oral commensals from mandibular molars and from oral swabs at the indicated number of days relative to the first inoculation of S. mutans UA159, as determined by qPCR. Mice were fed the high-sucrose diet with 4% sucrose water. Results are from a single experiment and expressed as mean ± SE of recovered genomes (n = 20 per group). *, p ≤ 0.05 versus the previous swab or an earlier swab as indicated by its number (e.g., 1 for Swab 1). #, p ≤ 0.05 versus S. mutans swab day 48 of Day 7 group, and mouse commensals in all swabs of Day 7 group and all swabs but day 29 of Day 0 group. Statistical comparisons by one-way ANOVA with Tukey's multiple comparisons test. and mouse oral commensals (closed squares, dotted lines) from oral swabs 1-5 taken at the times indicated in A and from sonicates of mandibular molars (M). Mice were fed the high-sucrose diet plus 1.5% arginine with 4% sucrose water. Results are mean ± SE (n = 14 mice per group) of recovered genomes estimated by qPCR. *, p ≤ 0.05 versus the previous swab or an earlier swab as indicated by its number (e.g., 2 for Swab 2). M, p ≤ 0.05 versus the same point in the mock group. W, p ≤ 0.05 versus S. A12 wild type group. Statistical comparisons by one-way ANOVA with Tukey's multiple comparisons test. from oral swabs 1-5 taken at the times indicated in A and from sonicates of mandibular molars (M). Mice were fed the high-sucrose diet plus 1.5% arginine with 4% sucrose water. Results are mean ± SE (n = 14 mice per group) of recovered genomes estimated by qPCR. *, p ≤ 0.05 versus the previous swab or an earlier swab as indicated by its number (e.g., 2 for Swab 2). M, p ≤ 0.05 versus the same point in the mock group. Statistical comparisons by one-way ANOVA with Tukey's multiple comparisons test. Comparisons between the two mock groups in competition experiments of Fig. 4B (1) and Fig. 5B (2) that included colonization by S. mutans and mouse commensals. Results are mean ± SE (A, n = 10 mice per group; B, n = 14) of recovered genomes estimated by qPCR. Statistical comparisons between swabs at a given point are by one-way ANOVA with Sidak's's multiple comparisons test. Statistical comparisons between recoveries from mandibles are by the two-tailed unpaired t test. *, p ≤ 0.05 versus same point in the alternate experiment.