Magnesium-Dependent Promotion of H₂O₂ Production Increases Ecological Competitiveness of Oral Commensal Streptococci

Abstract

The pyruvate oxidase (SpxB)–dependent production of H₂O₂ is widely distributed among oral commensal streptococci. Several studies confirmed the ability of H₂O₂ to antagonize susceptible oral bacterial species, including caries-associated Streptococcus mutans as well as several periodontal pathobionts. Here we report a potential mechanism to bolster oral commensal streptococcal H₂O₂ production by magnesium (Mg²⁺) supplementation. Magnesium is a cofactor for SpxB catalytic activity, and supplementation increases the production of H₂O₂ in vitro. We demonstrate that Mg²⁺ affects spxB transcription and SpxB abundance in Streptococcus sanguinis and Streptococcus gordonii. The competitiveness of low-passage commensal streptococcal clinical isolates is positively influenced in antagonism assays against S. mutans. In growth conditions normally selective for S. mutans, Mg²⁺ supplementation is able to increase the abundance of S. sanguinis in dual-species biofilms. Using an in vivo biophotonic imaging platform, we further demonstrate that dietary Mg²⁺ supplementation significantly improves S. gordonii oral colonization in mice. In summary, our results support a role for Mg²⁺ supplementation as a potential prebiotic to promote establishment of oral health–associated commensal streptococci.

Introduction

Oral biofilms are functionally and structurally organized polymicrobial communities embedded in an extracellular matrix on the tooth surface, the oral mucosa, and dental restorations (Marsh and Zaura 2017; Bowen et al. 2018). Microbes within the biofilm have evolved synergistic and antagonistic mechanisms to compete for available ecological niches. Important for overall oral health, the polymicrobial community maintains an ecological equilibrium protecting the host from diseases such as dental caries and periodontitis as well as invading pathogens (Takahashi and Nyvad 2011). However, host behaviors and environmental perturbations can promote the overgrowth and virulence of oral pathobionts that would otherwise remain harmless. If left unchecked, this can trigger oral dysbiosis and pathology to the host (Takahashi and Nyvad 2011; Bowen et al. 2018; Zhou et al. 2018).

Commensal Streptococcus sanguinis and Streptococcus gordonii are pioneer colonizers of the oral biofilm. Several studies observed that S. sanguinis and S. gordonii are associated with oral health, and their abundance declines in individuals with caries and periodontal disease (Zeng et al. 2012; Marsh et al. 2015; Mira et al. 2017). Characteristic of both species is their pyruvate oxidase (SpxB)–dependent production of H₂O₂. Under aerobic conditions, the catalytic activity of SpxB generates H₂O₂, CO₂, and acetyl phosphate. Acetyl phosphate is further metabolized to acetate, generating adenosine triphosphate (ATP) (Kreth et al. 2008). The H₂O₂ produced by the numerous commensal SpxB-encoding commensal streptococci such as S. sanguinis and S. gordonii is one of the main inhibitors antagonizing the growth of cariogenic competitors such as Streptococcus mutans (Kreth et al. 2008; Huang et al. 2018). Furthermore, H₂O₂ is a determinant in the interactions between S. sanguinis/S. gordonii and Aggregatibacter actinomycetemcomitans, Porphyromonas gingivalis, and other bacteria, modifying the spatial-temporal arrangement of those species in the oral biofilm (Jakubovics et al. 2008; Stacy et al. 2014; Herrero et al. 2016; Zhou et al. 2017). Recent studies have demonstrated that H₂O₂ plays an important role in cell-cell aggregation and biofilm formation through the H₂O₂-mediated release of DNA (Kreth et al. 2009; Itzek et al. 2011). Extracellular DNA (eDNA) is not only a major constituent of the biofilm extracellular polymeric substances (EPS) (Whitchurch et al. 2002) but also serves as a source of DNA for horizontal gene transfer (Itzek et al. 2011; Schlafer et al. 2017). These observations demonstrate that H₂O₂ produced by S. sanguinis andS. gordonii plays a crucial role in early oral biofilm development and likely modulates the species composition in the oral cavity (Redanz, Cheng, et al. 2018).

Due to our focus upon studies of molecular commensalism (Kreth et al. 2017) as well as the role of H₂O₂ in oral microbial ecology (Redanz, Cheng, et al. 2018), we were interested to identify potential approaches to bolster H₂O₂ production in the oral cavity as a mechanism to strengthen commensal oral streptococci. Here, we provide evidence that magnesium supplementation could be one such approach and expand on a previous finding that magnesium promotes peroxidogenesis (Barnard and Stinson 1999).

Materials and Methods

Detailed descriptions of the materials and methods can be found in the Appendix.

Bacterial Strains and Culture Conditions

Strains are listed in Appendix Table 1. Bacteria were grown in brain heart infusion medium, supplemented with antibiotics when required, at 37°C and 5% CO₂.

DNA Manipulations

Standard nucleic acid recombinant protocols were used.

Construction of Mutacin V (SMU_1914c) Luciferase Reporter Strain

The Renilla luciferase reporter was constructed using an overlapping extension polymerase chain reaction (PCR) approach using an erythromycin resistance cassette as a selectable marker.

Measurement of H₂O₂ Production

H₂O₂ was determined with indicator Prussian blue plates or enzymatically using o-dianisidine and horseradish peroxidase.

RNA Isolation, Complementary DNA Synthesis, and Quantitative PCR

RNA was isolated using TRIzol and complementary DNA (cDNA) synthesized using SuperScript II RT (Invitrogen), and quantitative PCR was performed using SYBR green and 2^–ΔΔCT analysis.

Western Blot

Western blots were performed using SpxB-FLAG engineered proteins, anti-FLAG M2 antibodies, and enhanced chemiluminescence detection.

Luciferase Assay

Firefly and Renilla luciferase activities were determined using a GloMax 265 Discover plate reader (Promega).

Inhibition Assays

A deferred antagonism assay was used to determine interspecies inhibition.

Biofilm Analysis

Biofilms were stained with 0.1% safranin and absorption at 492 nm measured using a GloMax 265 Discover plate reader.

Dual-Species Biofilm Competition

S. sanguinis and S. mutans coculture biofilms were grown in polystyrene 24-well plates. Cell number was determined by colony-forming unit (CFU) counting and H₂O₂ production determined using the Amplex Red Assay.

Bioluminescent In Vivo Bacterial Imaging of Mouse Colonization

BALB/CByJ mice were used and inoculated with bioluminescent S. gordonii. Mice were imaged periodically using an IVIS Spectrum biophotonic imaging system.

Statistical Analysis

Software packages from SPSS and ImerTest were used for statistical analysis.

Results

Magnesium Promotes H₂O₂ Production in S. sanguinis and S. gordonii

Magnesium is a cofactor for SpxB catalytic activity and has previously been shown to stimulate peroxidogenesis (Barnard and Stinson 1999; Tittmann et al. 2005). To determine whether magnesium influences H₂O₂ production in S. sanguinis andS. gordonii, a colorimetric H₂O₂ assay based on Prussian blue agar plates was used (Saito et al. 2007). Prussian blue agar plates are based on BHI, which contains Mg²⁺ with reported concentrations between 60 and 350 µM (Jen and Shelef 1986; Damo et al. 2013), which is in the range of salivary concentrations from 246 to 452 µM (Monaci et al. 2002). Since increased peroxidogenesis has been reported before for S. gordonii, we used a range of Mg²⁺ concentrations (0 to 10 mM Mg²⁺) that were close to the previously reported 1 mM Mg²⁺ (Barnard and Stinson 1999). The amount of H₂O₂ produced by S. sanguinis and S. gordonii in the presence of Mg²⁺ was noticeably increased with 5 mM Mg²⁺ supplementation, as evident from the increased production of colored precipitate surrounding the respective producers (Fig. 1A, B). To exclude the possibility that Mg²⁺ itself enhances H₂O₂ detection on the colorimetric plates, a 30% commercial H₂O₂ solution was diluted into 3% and 0.3%, and then 15 µL of each was spotted onto the Prussian blue agar plates supplemented with different concentrations of Mg²⁺ and incubated aerobically at 37°C for 16 h. The dimensions of the blue circles on the indicator plates exhibited no statistically significant differences among the groups with or without Mg²⁺ (Appendix Fig. 1B). In contrast, the H₂O₂-reactive chemical sodium pyruvate greatly reduced Prussian blue detection of 3% and 0.3% H₂O₂ when supplemented at 1-, 5-, and 10-mM concentrations (Appendix Fig. 1C). This is consistent with the H₂O₂ scavenging activity of pyruvate (Constantopoulos and Barranger 1984). Moreover, the H₂O₂ concentration of the supernatant was determined in liquid cultures grown aerobically and normalized to OD_600nm. The results also showed that H₂O₂ production of S. sanguinis and S. gordonii increased significantly in the presence of Mg²⁺ (Fig. 1C). Furthermore, Western blot analysis was performed using isogenic strains of S. sanguinis and S. gordonii producing FLAG-tagged SpxB. As predicted, noticeably more SpxB was detected for S. sanguinis and S. gordonii in the presence of Mg²⁺ compared to the control group without Mg²⁺ (Fig. 1D).

Figure 1.

Addition of Mg²⁺ increases H₂O₂ productions in Streptococcus sanguinis and Streptococcus gordonii. (A) Representative images of H₂O₂ production of S. sanguinis and S. gordonii on Prussian blue agar plates with or without supplemented Mg²⁺. The arrow indicates the distance measured. (B) Quantitative analysis of H₂O₂ production on indicator plates; production was determined by measuring the distance between the border of the colony to the border of the halo formed by the Prussian blue precipitate, as illustrated by the arrow in Figure 1A. (C) Quantitative analysis of H₂O₂ production in liquid cultures; production was determined by the peroxidase method and normalized to optical density (OD_600nm). Presented is the average of 3 biological replicates. Significant differences are indicated by *P < 0.05. (D) Top, quantification of SpxB protein (72 kDa) abundance of S. sanguinis and S. gordonii in the presence of different concentrations of Mg²⁺ (the respective sodium dodecyl sulfate–polyacrylamide gel electrophoresis gel with Coomassie blue staining, shown as a loading control, can be found in Appendix Fig. 1A); Western blots are representative of 3 biological replicates. Relative Western blot band intensity was analyzed by ImageJ (National Institutes of Health) and was defined as the ratio of the Mg²⁺ supplementation group to those without Mg²⁺; the intensity of the unsupplemented group was arbitrarily defined as 1. (E) Relative spxB gene expression of S. sanguinis and S. gordonii in the presence of different concentrations of Mg²⁺; the expression was determined by quantitative reverse transcription polymerase chain reaction. (F) Specific luciferase activity of S. sanguinis and S. gordonii in the presence of different concentrations of Mg²⁺; the luciferase gene was inserted after the spxB promoter. Presented is the average of 3 biological replicates. Significant differences are indicated by *P < 0.05.

While SpxB protein abundance and H₂O₂ production demonstrated a dose-dependent correlation in our assays, the expression of spxB as determined by quantitative reverse transcription PCR (RT-qPCR) and luciferase reporter assays did not show the exact same pattern (Fig. 1E, F). Addition of 1 mM Mg²⁺ did significantly increase spxB expression in both assays for S. sanguinis and S. gordonii when compared to the control. However, addition of 5 or 10 mM Mg²⁺ did not further increase spxB expression in a dose-dependent manner when compared to the addition of 1 mM Mg²⁺. A similar result was also observed using a transcription fusion spxB-luciferase reporter of S. gordonii (Fig. 1F). The reason for the discrepancy is currently unknown but discussed below. The data strongly support that magnesium is able to promote S. sanguinis and S. gordonii H₂O₂ production.

Magnesium-Dependent Increased H₂O₂ Production Facilitates the Antagonistic Effect of S. sanguinis and S. gordonii against S. mutans

The growth of S. mutans, S. sanguinis, and S. gordonii is unaffected by Mg²⁺ supplementation at 1, 5, and 10 mM (Appendix Fig. 2A–C). No differences were observed for single-species biofilm formation of S. mutans, S. sanguinis, and S. gordonii in the presence of 0, 1, 5, and 10 mM Mg²⁺ (Appendix Fig. 2D). The luciferase activity determined with the ldh-firefly luciferase reporter strains of S. mutans, S. sanguinis, and S. gordonii exhibited no statistically significant difference in the presence of 0, 1, 5, and 10 mM Mg²⁺, which further confirmed that Mg²⁺ at low concentrations did not influence bacterial growth (Appendix Fig. 2E). Moreover, we found that the tested concentrations of Mg²⁺ exhibited no effect upon mutacin V (nlmC) gene expression in S. mutans (Appendix Fig. 2E), indicating that Mg²⁺ supplementation does not affect its production.

Next, we tested whether increased Mg²⁺-dependent production of H₂O₂ influences the competitiveness of S. sanguinis and S. gordonii using a deferred antagonism assay. S. sanguinis andS. gordonii were inoculated on the BHI plates 16 h prior to the inoculation of S. mutans. Obviously, stronger antagonism of S. mutans was observed when S. sanguinis and S. gordonii were grown in the presence of Mg²⁺ (Fig. 2A, B), and this effect disappeared when 5 mg/mL catalase was added to the BHI plates (Fig. 2C). This suggests that magnesium supplementation increased the antagonistic abilities of S. sanguinis and S. gordonii due to their increased capacity to produce H₂O₂ and inhibit S. mutans.

Figure 2.

Effect of Mg²⁺ on competitive growth of Streptococcus sanguinis/Streptococcus gordonii against Streptococcus mutans. (A) Inhibition assays between S. sanguinis/S. gordonii and S. mutans in the presence of different concentrations of Mg²⁺. S. sanguinis and S. gordonii were inoculated first. After overnight incubation, S. mutans was inoculated adjacent to S. sanguinis/S. gordonii. (B) Quantitative analysis of growth inhibition on the plates; the inhibition was determined by measuring the distance of the inhibition zone at the intersection with the pioneer colony, as illustrated by the arrow in (A). Presented is the average of 3 biological replicates. Significant differences are indicated by *P < 0.05. (C) Inhibition assays between S. sanguinis/S. gordonii and S. mutans in the presence of different concentrations of Mg²⁺ and 5 mg/mL catalase.

Magnesium Promotes H₂O₂ Production in Other Oral Commensal Streptococci and Also Improves Their Antagonism of S. mutans

The spxB gene is highly conserved among H₂O₂-producing oral commensal streptococci (Redanz, Cheng, et al. 2018). Therefore, we were curious to compare magnesium-dependent H₂O₂ production using additional oral commensal streptococci. Low-passage oral clinical isolates of Streptococcus parasanguinis, Streptococcus oralis ssp. dentisani, Streptococcus mitis, Streptococcus oralis, S. gordonii, and an S. sanguinis endocarditis isolate were tested for H₂O₂ production using H₂O₂ indicator plates supplemented with different concentrations of Mg²⁺. All of the tested clinical isolates exhibited a Mg²⁺ concentration-dependent increase in H₂O₂ production on agar plates and in liquid cultures (Fig. 3A, B).

Figure 3.

H₂O₂ production and antagonism assays of Streptococcus sanguinis 133-79 and other clinical isolates (JM1, Streptococcus parasanguinis, Streptococcus oralis ssp. dentisani, Streptococcus mitis, Streptococcus oralis) in the presence of different concentrations of Mg²⁺. (A) Quantitative analysis of H₂O₂ production on indicator plates; production was determined by measuring the distance between the border of the colony to the border of the halo formed by the Prussian blue precipitate. (B) Quantitative analysis of H₂O₂ production in liquid cultures; production was determined by peroxidase method and normalized to optical density (OD_600nm). Presented is the average of 3 biological replicates. Significant differences are indicated by *P < 0.05. (C) Inhibition assays between the H₂O₂ production clinical isolates and Streptococcus mutans in the presence of different concentrations of Mg²⁺. (D) Quantitative analysis of growth inhibition on the plates; the inhibition was determined by measuring the distance of the inhibition zone at the intersection with the pioneer colony, as illustrated by the arrow in (A). Presented is the average of 3 biological replicates. Significant differences are indicated by *P < 0.05. **S. oralis ssp. dentisani.

Furthermore, the antagonistic abilities of all strains tested against S. mutans were improved when grown in the presence of added Mg²⁺ (Fig. 3C, D). As expected, the addition of catalase abolished the antagonistic capabilities (Appendix Fig. 3). This indicates that magnesium can promote H₂O₂ production in a wide variety of spxB-encoding oral commensal streptococci to improve their competitive abilities.

Magnesium Increases S. sanguinis Competitiveness in a Dual-Species Biofilm with S. mutans

In the oral cavity, S. sanguinis competes with other species during biofilm development on the tooth surface. To determine whether Mg²⁺ supplementation can increase S. sanguinis competitiveness during biofilm formation with S. mutans, we simulated an environment that is highly selective for S. mutans by growing the cells in BHI supplemented with 50 mM sucrose. As shown in Figure 4A, S. mutans was able to dominate biofilm development when no magnesium was added to the growth medium, with nearly undetectable levels of S. sanguinis in both the biofilm fraction and the planktonic phase of the dual-species biofilms. This dramatically changed with the addition of 5 mM Mg²⁺, as S. sanguinis abundance increased by >5 orders of magnitude in the biofilm and planktonic phase. This is in stark contrast to S. mutans, which was unaffected by Mg²⁺ supplementation (Fig. 4A and Appendix Fig. 4). The concentration of H₂O₂ in the supernatants of the biofilm cocultures was significantly increased (Fig. 4B), in agreement with establishment of S. sanguinis in greater abundance when MgCl₂ was supplied.

Figure 4.

Dual-species biofilm competition and bioluminescent Streptococcus gordonii mouse colonization. (A) Streptococcus sanguinis (SK36) and Streptococcus mutans (UA159) after 24 h of cocultivation ±5 mM Mg²⁺. Cells were grown in brain heart infusion supplemented with 50 mM sucrose. Colony-forming units (CFUs) were determined from planktonic and biofilm cells after sonication. Shown are the geometric means and 95% confidence intervals. Significance levels were determined by t test (unpaired, 2-tailed) after log₁₀ transformation. (B) Shown is the H₂O₂ concentration after 24 h of cocultivation ±5 mM Mg²⁺. Significance levels of data from 3 independently performed experiments were determined by t test (unpaired, 2-tailed). *P < 0.05, **P < 0.01, ***P < 0.001. (C) Boxplot of the average radiance (p/s/cm²/sr) values recorded over 14 d for control mice and mice fed with 10 mM Mg²⁺ ad libitum in the drinking water (containing 5% sucrose and 5% fructose for both groups). (D) Representative in vivo bioluminescent images of control mice and mice fed with 10 mM Mg²⁺ colonized by bioluminescent S. gordonii carrying a green Renilla reporter gene controlled by the constitutive ldh promoter. The pseudo-color scale shows relative photon flux as average radiance (p/s/cm²/sr).

Mg²⁺-Dependent Modeling of Bacterial Ecology Using a Biophotonic Mouse Model

The Mg²⁺-dependent increase in H₂O₂ production was able to bolster commensal streptococcal competitiveness in dual-species in vitro biofilms. Based on this observation, we hypothesized that dietary supplementation of Mg²⁺ would positively influence commensal streptococcal establishment in an ecologically relevant environment such as the mouse oral cavity. We previously developed a luciferase-based murine biophotonic imaging assay to measure oral bacterial colonization in situ (Merritt et al. 2016). The colonization was recorded over 14 d after a single inoculation of reporter bacteria. A significant increase in bioluminescent S. gordonii was observable for up to 12 d in mice supplied with 10 mM Mg²⁺ (Fig. 4C, D). Afterward, the abundance of reporter bacteria began to diminish, likely due to growth competition from the native mouse flora.

Discussion

The initiation and development of oral diseases are closely associated with the development of oral dysbiosis (Bowen et al. 2018). Ecological interactions within the biofilm as well as bacterial metabolites are crucial factors modulating oral biofilm composition. H₂O₂ generated by the majority of commensal oral streptococci seems to play a prominent role in shaping the microbial composition and spatial-temporal arrangement of the oral biofilm (Jakubovics et al. 2008; Kreth et al. 2008; Redanz, Cheng, et al. 2018). Oral biofilms are subject to frequent and dramatic changes in pH, glucose concentration, and other environmental factors, which can affect the production of H₂O₂ (Redanz, Cheng, et al. 2018; Redanz, Masilamani, et al. 2018). Our focus has been to characterize the molecular mechanisms involved in the control of H₂O₂ production to determine whether this ability could be bolstered therapeutically. For example, recent in vitro studies demonstrated that the addition of arginine increased H₂O₂ production in S. sanguinis and S. gordonii and improved their antagonistic abilities against S. mutans (He et al. 2016; Huang et al. 2018). In the current study, we examined the effect of Mg²⁺ on H₂O₂ production in oral commensal streptococci and explored its ecological significance by measuring H₂O₂-dependent antagonism of S. mutans. Other divalent metal ions such as Mn²⁺ and Ca²⁺ have been shown to either stimulate H₂O₂ production in S. gordonii (Ca²⁺) (Barnard and Stinson 1999) or function as important cofactor (Mn²⁺) in the activity of the Lactobacillus plantarum pyruvate oxidase (Sedewitz et al. 1984), which shares 49% identity on the amino acid level with SpxB from S. sanguinis (Redanz, Cheng, et al. 2018). However, our results with Mn²⁺ and Ca²⁺ did not show a uniform response in S. sanguinis and S. gordonii (Appendix Fig. 5). This is consistent with the observation that 10 mM Mn²⁺ had no effect on H₂O₂ production in S. gordonii (Barnard and Stinson 1999). Thus, we focused on the effect of Mg²⁺.

Both methods used to determine H₂O₂ production consistently showed that H₂O₂ production by S. sanguinis, S. gordonii, and other H₂O₂-producing low-passage clinical isolates was increased significantly in the presence of Mg²⁺. Gene expression and Western blot analysis showed that magnesium upregulates spxB gene expression and increases SpxB protein abundance, thus promoting bacterial H₂O₂ production. We did observe that the addition of different concentrations of Mg²⁺ had an indistinguishable effect on spxB expression for S. sanguinis and S. gordonii. This was also observed for S. gordonii spxB luciferase reporter gene activity, while the luciferase activity for S. sanguinis still seemed to follow in a dose-dependent manner, albeit with lower activity when compared to S. gordonii (Fig. 1E, F). Consistent with our results, Barnard and Stinson (1999) reported that H₂O₂ synthesis by S. gordonii decreased upon removal of Mg²⁺, whereas stimulation of peroxidogenesis occurred with 1 mM Mg²⁺ but was insensitive to changes in the Mg²⁺ concentration. In general, gene expression and protein abundance are correlated, but discrepancies have been reported in the literature (Vogel and Marcotte 2012; Liu et al. 2016). One possible explanation for the observed discrepancy between the dose-dependent increase of H₂O₂ production and the expression of spxB might be differential posttranscriptional regulation or messenger RNA (mRNA) stability in S. gordonii and S. sanguinis. Furthermore, it was shown that the correlation between mRNA and protein abundance decreases when the protein degrades more slowly (Raj et al. 2006). Although we do not have actual data for SpxB protein turnover rates in either species, the SpxB protein is still detectable under anaerobic conditions, while spxB expression ceases pointing to a longer half-life of SpxB (Zheng et al. 2011), suggesting that such a correlation might exist in S. sanguinis and S. gordonii. However, further investigations need to clarify the discrepancy observed here.

Mechanistically, the increased production of H₂O₂ might be attributable to the function of Mg²⁺ for the binding of the essential cofactor thiamine diphosphate, an important component in the catalytic reaction of SpxB (Redanz, Cheng, et al. 2018). However, this effect would not explain the increase in spxB expression. Therefore, we tested the hypothesis that overall transcription might be increased with Mg²⁺ supplementation, since Mg²⁺ is also important for the catalysis of the DNA-dependent RNA polymerase (Nudler 2009). However, the ldh (lactate dehydrogenase) luciferase reporter activity of S. mutans, S. sanguinis, and S. gordonii was unaffected by the tested Mg²⁺ concentrations.

Most important, we found that the magnesium-dependent increased H₂O₂ production augmented the antagonistic activity of S. sanguinis and S. gordonii against S. mutans. S. mutans is a vigorous producer of multiple mutacins, which are bacteriocins effective against a variety of other closely related species, including S. sanguinis and S. gordonii (Qi et al. 2001; Hale et al. 2005; Hossain and Biswas 2011). Therefore, we wanted to exclude the possibility that Mg²⁺ influences mutacin gene expression. The luciferase activity observed with the nlmC-Renilla luciferase reporter strain of S. mutans exhibited no statistically significant difference in the presence of Mg²⁺. This suggests that the increased antagonism of S. mutans was specifically due to increased H₂O₂ production by S. sanguinis and S. gordonii, rather than influencing mutacin V production by S. mutans. However, since S. mutans produces multiple mutacins, we cannot exclude additional effects of Mg²⁺ on other mutacin genes or S. mutans biology in general that might interfere with its fitness.

The effect of Mg²⁺ supplementation was also confirmed in a dual-species biofilm with S. mutans and S. sanguinis. While highly selective for S. mutans due to the addition of 50 mM sucrose, S. sanguinis was able to establish significant numbers in the biofilm when Mg²⁺ was supplied. This was also accompanied by a significant difference in the H₂O₂ concentration that was detected in the Mg²⁺ supplemented biofilm. Aside from the effect on H₂O₂ production itself, a metabolic advantage due to the increased activity of SpxB and the production of energetically favorable acetyl phosphate could account for the enhanced competitiveness. We previously demonstrated that spxB mutants of S. sanguinis and S. gordonii are severely impaired in their ability to establish dual-species biofilms with S. mutans (Kreth et al. 2008), further supporting the crucial role of H₂O₂ and SpxB for growth competition with S. mutans. However, since Mg²⁺ can have additional effects on metabolic or regulatory activities, we cannot exclude that other mechanisms are in part responsible for the competitiveness observed here. For example, Mg²⁺ is able to stabilize macromolecular complexes, including bacterial membranes, and serves as cofactor in a variety of enzymatic reactions (Groisman et al. 2013). Although we did not observe any significant differences in the growth of S. sanguinis, S. gordonii, and S. mutans in the presence of different Mg²⁺ concentrations (Appendix Fig. 2A–C), fitness effects might only be obvious under stress situations such as dual-species competitions in biofilms.

The mouse oral cavity provides an ecologically relevant environment to simulate multispecies competition, host innate immunity, and environmental challenges such as saliva flow and mastication. We followed S. gordonii colonization using a murine biophotonic imaging assay over 14 d, since we previously confirmed the timeframe for successful colonization of bioluminescent S. gordonii in a dual-species colonization experiment (Merritt et al. 2016). Mice that drank water supplemented with 10 mM Mg²⁺ exhibited significantly higher S. gordonii bioluminescence values, suggesting that Mg²⁺ did provide an advantage in colonization and/or persistence throughout most of the assay. Overall, our results agree with recent studies that indicate the potential utility of prebiotics to modify oral microbiome ecology in favor of commensal streptococci (Nascimento et al. 2014; Scoffield et al. 2019). In conclusion, the experimental data presented here support the use of low levels of magnesium as a prebiotic to increase commensal streptococcal hydrogen peroxide production. Our results are also consistent with earlier reports that similar concentrations of magnesium in mouthrinse had a reducing effect on plaque accumulation (Skjorland et al. 1978) and thus could be a valid option for use in humans in the future.

Author Contributions

X. Cheng, S. Redanz, P. Treerat, H. Qin, contributed to conception, design, and data analysis, drafted and critically revised the manuscript; D. Choi, contributed to conception, design, and data analysis, critically revised the manuscript; X. Zhou, X. Xu, J. Merritt, J. Kreth, contributed to conception and design, critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work.