Live and let die: hydrogen peroxide production by the commensal flora and its role in maintaining a symbiotic microbiome

Summary

The majority of commensal oral streptococci are able to generate hydrogen peroxide (H₂O₂) during aerobic growth, which can diffuse through the cell membrane and inhibit competing species in close proximity. Competing H₂O₂ production is mainly dependent upon the pyruvate oxidase, SpxB and to a lesser extent the lactate oxidase LctO, both of which are important for energy generation in aerobic environments. Several studies point to a broad impact of H₂O₂ production in the oral environment, including a potential role in biofilm homeostasis, signaling, and interspecies interactions. Here, we summarize the current research regarding oral streptococcal H₂O₂ generation, resistance mechanisms, and the ecological impact of H₂O₂ production. We also discuss the potential therapeutic utility of H₂O₂ for the prevention/treatment of dysbiotic diseases as well as its potential role as a biomarker of oral health.

Introduction

The human host is closely associated with complex microbial communities consisting primarily of bacteria, but also of viruses, fungi, archaea, and certain protozoa¹. The colonization of oral mucosal and hard tissue surfaces by microorganisms is accompanied by the vigorous intra- and interspecies exchange of information among diverse microbial communities as well as cross-kingdom communication between the microbes and host ^2,3. The ideal outcome of microbial colonization is to create a host-protective environment characterized by a consortium of commensal and mutualistic flora species. The protection afforded by the resident commensal flora can be viewed both temporally as well as spatially: temporally, since the initial colonization after birth can contribute to a lifetime of proper host immune system function ⁴ and spatially, since community interactions amongst the flora as well as flora-host interactions both yield mostly local effects ⁵. To receive these protective benefits from the flora, it is critical to maintain microbial symbiosis ⁶. There are various recognized mechanisms that promote a symbiotic relationship with the flora, including the production of hydrogen peroxide (H₂O₂) by the commensal oral streptococci.

The oral cavity has several surfaces covered with microbes dwelling in biofilms and the composition of the biofilm community is determined by the respective environment ^7–9. For example, the process of dental biofilm formation has been the center of several in vivo oral microbial ecology studies ^10–13. Distinctive patterns of colonization sequences have been described, which can develop into a mature biofilm community with enormous complexity and structure. Initial colonization of the saliva bathed tooth surface in the first 4 to 8 hours is dominated by oral streptococci and Corynebacterium ssp. as recently shown ^12,14. This dominance extended up to 16 hours into biofilm development and Streptococcus accounted for about 20% of the taxa ¹⁴. Incidentally, oral streptococci are also the main source for ecologically relevant H₂O₂ production ¹⁵. The growth of initial colonizers during biofilm development in the first 16 hours seemed to be exponential, suggesting ideal conditions for facultative anaerobic and aerotolerant species. Further biofilm development favors the integration of Gram-negative anaerobic species including Fusobacterium, Prevotella and Porphyromonas, but growth of the biofilm population seemed to slow down after 16 hours ¹⁴.

Initial biofilm development requires the formation of saliva-derived macromolecular complexes on the tooth surface leading to the formation of the acquired enamel pellicle. This process takes place within seconds after a clean enamel surface is exposed to saliva ¹⁶. Mechanistically, the abundance of streptococci as initial colonizers can be explained by surface adhesins specifically recognizing salivary proteins that are part of the acquired enamel pellicle. Integration of other species, including later colonizers, occurs via specific cell-surface receptors also displayed by streptococci ¹⁷, thus streptococci together with a few other species are able to determine the spatial and temporal development of oral biofilms and are important for biofilm homeostasis. Several synergistic and antagonistic mechanisms have been described that contribute to oral biofilm dynamics ¹⁸, and H₂O₂ production seems to play a significant role ¹⁵.

In the time since our previous review about the role of H₂O₂ in oral biofilm ecology ¹⁵, there have been important advances in our understanding of oral microbial ecology and species composition in health and disease. One of the key outcomes of those studies is that polymicrobial diseases such as caries and periodontal disease are the ecological consequences of dysbiosis. For example, caries is triggered by overgrowth of a subset of the oral flora that are inherently aciduric and acidogenic ¹⁹. We have previously suggested that the conserved ability of commensal oral streptococci to produce H₂O₂ is important for oral health ²⁰. Here we review the latest evidence detailing our current knowledge of the biology of streptococcal H₂O₂ production as well as the role of H₂O₂ as key component of oral ecology.

Relevant Chemistry of hydrogen peroxide and reactive oxygen species (ROS) production

i). Oxygen redox pathways generate H₂O₂ as an intermediate

The generation of H₂O₂ during cellular redox reactions occurs due to the sequential univalent reduction of molecular oxygen ²¹. H₂O₂ is the second intermediate formed by the addition of one electron and two protons to the highly unstable superoxide anion. H₂O₂ is relatively stable and unreactive in an abiotic environment ^22,23. Pure solutions of macromolecules like nucleic acids, proteins (metal-free), lipids, and polysaccharides are typically resistant to H₂O₂ oxidation. However, this is not the case in a biological environment, which is replete with transition metals such as Fe(ii) or Cu²⁺ that act as reducing agents for H₂O₂ ^22,24. The further reduction of H₂O₂ results in the addition of an electron and a proton to generate H₂O and the extremely reactive and short-lived hydroxyl radical, which is the major reactive oxygen species (ROS) reacting with biomolecules ²¹. The half-life of the hydroxyl radical in a biological system is only about 1 ns. Its reaction with biomolecules is non-selective and is only limited by diffusion ²⁵.

ii). H₂O₂ involvement in the production of hydroxyl radicals via Fenton chemistry

In a cellular context, the generation of hydroxyl radicals is primarily catalyzed by Fenton chemistry via the transition metal ion pool inside the cell, especially soluble Fe(II) ²⁶. Fe(II) donates one electron to one H₂O₂ molecule to generate a hydroxyl radical, a hydroxide ion, and oxidized Fe(III). The hydroxyl radical will subsequently react rapidly with any organic cellular compound in the immediate vicinity ²⁷

iii). Cytotoxicity of H₂O₂

Much of the cytotoxicity of H₂O₂ is the result of oxidative DNA damage induced by hydroxyl radicals generated through Fenton chemistry ^27,28. Hydroxyl radicals can trigger direct strand breakage through the oxidation of the ribose moieties of the DNA backbone ^27,29,30. In addition, the nucleobases can be oxidized, with guanine being one of the major targets of oxidation ³¹. Interestingly, two kinetically distinguishable modes of killing by H₂O₂ seem to exist as shown with the model organism Escherichia coli ^32,33. Cells treated with H₂O₂ concentrations lower than 3 mM were more susceptible compared to H₂O₂ concentrations between 5 to 20 mM. Above 20 mM H₂O₂, the survival rate is inversely proportional to the H₂O₂ concentration, as would be expected. The dual range of H₂O₂ susceptibility is attributed to separate killing mechanisms. Lower concentrations of H₂O₂ will mostly damage DNA due to Fenton reaction hydroxyl radicals inducing a lethal mutation rate ^27,34. The ability of Fe to bind to DNA is a key feature of this mechanism ³⁵. The second mode of killing at higher H₂O₂ concentrations seems to be the result of damage to other target(s) ^33,34, including proteins and lipids. However, the biologically relevant H₂O₂ concentration bacteria normally encounter is usually in the nM to lower mM range ^36–39. Hydrogen peroxide concentrations up to 7 mM have been reported for streptococci ³⁷, although those measurements were mostly performed in batch cultures that may not be reflective of the normal H₂O₂ production capacity of streptococcal biofilms. Regardless, it seems unlikely that E. coli would encounter >20 mM H₂O₂ during growth in bacterial communities. A similar dual susceptibility to H₂O₂ has also been reported for Streptococcus thermophilus ⁴⁰, which is intriguing since both organisms utilize distinct mechanisms to detoxify H₂O_2. E. coli directly degrades H₂O₂ via the catalase enzyme ⁴¹, whereas streptococci typically lack catalase and rely on other mechanisms ^42,43. It is worth noting that S. mutans is equally susceptible to killing when treated with 1 to 5 mM H₂O₂ ⁴⁴. Thus, the dual susceptibility to H₂O₂ killing might not be generalizable for all species and/or growth conditions.

H₂O₂ mediated damage can also occur to enzymes with iron-sulfur clusters abolishing enzyme activity, as well as enzymes that require a single Fe²⁺ ion as cofactor for enzymatic activity ³¹. Other potential targets are lipids in the cell membrane of bacterial cells. However, this has not been studied with oral biofilm bacteria and the lack of the main target for lipid peroxidation, polyunsaturated fatty acid chains in most bacterial membranes, makes it unlikely that this mode of killing action occurs ^27,38.

Sources of microbial hydrogen peroxide production in the oral biofilm

In the oral biofilm, naturally produced H₂O₂ crosses the bacterial membrane of producer cells and is released into the environment where it can influence neighboring cells and ultimately oral biofilm ecology. Our group and others have demonstrated that abundant H₂O₂ production in the pioneer colonizer S. sanguinis is critically dependent upon the activity of the pyruvate oxidase, SpxB. SpxB knock-out mutants do not produce sufficient extracellular H₂O₂ to inhibit competing H₂O₂-susceptible species like Streptococcus mutans ⁴⁵. SpxB-dependent H₂O₂ production has also been shown experimentally for S. gordonii ⁴⁵ and Streptococcus parasanguinis ⁴⁶ as well as Streptococcus infantis, Streptococcus oralis, Streptococcus cristatus and Streptococcus mitis as shown on Prussian Blue H₂O₂ indicator plates ^47,48 (Fig. 1, unpublished).

Fig. 1:

Comparison of H₂O₂ production by S. oralis ssp. tigurinus J22, S. infantis LZ2, S. mitis 12261, and S. cristatus wild type strains and their respective isogenic spxB mutants. H₂O₂ production from all strains was assessed on H₂O₂ indicator plates after overnight incubation at 37ºC aerobically. H₂O₂ production leads to the precipitation of Prussian blue during aerobic growth, which forms a blue halo around the producer ^47,48.

i). Pyruvate oxidase, SpxB

The initial characterization of the pyruvate oxidase from S. sanguinis was reported by Jan Carlsson and colleagues. Permeabilized cells exhibited pyruvate oxidase activity dependent upon thiamine pyrophosphate (TPP), flavin adenine dinucleotide (FAD), Mg²⁺, and orthophosphate ⁴⁹. Catalytic activity was also significantly increased in cells grown aerobically ⁴⁹, while the reaction itself is oxygen-dependent ⁵⁰. SpxB catalyzes the oxidative decarboxylation of pyruvate to the high energy metabolite acetyl-phosphate in addition to CO₂, H₂O₂, and ATP. Thus, the reaction provides an energetic growth advantage for the producer ¹⁵. Our group confirmed the presence of SpxB under anaerobic conditions, suggesting that the enzyme is present in cells even when H₂O₂ production is not detectable ⁵¹. Consistent with the results from Carlsson, we observed S. sanguinis spxB expression to increase about 25-fold in aerobically grown cultures with the highest spxB expression occurring in late log phase ⁵¹. Even so, S. sanguinis still produces small amounts of SpxB under anaerobic conditions when pyruvate is primarily metabolized via the pyruvate formate lyase ⁴⁹. The reason for this is not entirely clear, but it could potentially provide a growth advantage once oxygen becomes available. Alternatively, SpxB might have additional moonlighting functions that have yet to be discovered.

The biochemical mechanism of SpxB catalysis has not been elucidated in oral streptococci, but this has been investigated using the Lactobacillus plantarum pyruvate oxidase (LpPox) ortholog. Several characteristics between LpPox and S. sanguinis SpxB are conserved and they share 49% identity (Table 1) ⁵². Molecular modeling using the automated protein structure homology-modelling server SWISS-MODEL (https://swissmodel.expasy.org) ⁵³ predicted a high conservation of the S. sanguinis SpxB protein structure compared to the known structures of the pyruvate oxidases from L. plantarum ⁵⁴ and Aerococcus viridans ⁵⁵ (Fig. 2).

Table 1:

Comparison of SpxB among different species

Species	Strain	number of amino acids	% identity	% positives
Streptococcus sanguinis	SK36	591	100	100
Streptococcus gordonii	CH1	591	98	99
Streptococcus mitis	B6	591	98	98
Streptococcus peroris	ATCC 700780	591	98	98
Streptococcus infantis	ATCC 700779	591	98	98
Streptococcus oralis	Uo5	591	98	98
Streptococcus pseudopneumoniae	IS7493	591	98	98
Streptococcus pneumoniae	R6	591	98	98
Streptococcus parasanguinis	CC87K	591	98	98
Streptococcus vestibularis	ATCC 49124	591	98	99
Streptococcus salivarius	HSISS4	591	97	98
Streptococcus cristatus	AS1.3089	591	98	99
Streptococcus anginosus	C1051	591	98	99
Streptococcus australis	ATCC 700641	591	96	98
Streptococcus himalayensis	HTS2	593	89	94
Streptococcus acidominimus	NCTC11291	593	85	92
		583 (pseudogene)*	70	83
Aerococcus viridans	CCUG4311	592	69	82
Lactobacillus plantarum	ZJ316	603 (PoxB) **	49	66

% identity: identity on the amino acid level to the query sequence of spxB from S. sanguinis.

% positives: amino acids with a conservative exchange were physiochemical properties are preserved when compared to the query sequence of spxB from S. sanguinis.

^*The second pyruvate oxidase gene in S. acidominimus possesses an internal stop codon.

^**L. plantarum encodes several pyruvate oxidases, shown here is the relevant pyruvate oxidase used for molecular modeling.

Fig. 2:

Protein structure homology modeling of SpxB from S. sanguinis using the automated protein structure homology-modelling server SWISS-MODEL. The reference proteins used were from Aerococcus viridans and Lactobacillus plantarum. The quaternary structure quality estimate (QSQE) based upon the modeling with Aerococcus viridans SpxB is 0.78. This protein has 69% sequence identity to the S. sanguinis SpxB. The QSQE score ranges from 0 to 1 with higher numbers indicating higher reliability.

From the LpPox crystal structure, several aspects of the catalytic process have been revealed, including the roles of the enzyme-bound co-factors TPP and FAD ^56,57. The reaction occurs in several steps beginning with a nucleophilic attack on the bound pyruvate molecule by TPP leading to acetyl-TPP and FADH₂. In a subsequent oxygen-dependent mechanism, FADH₂ is oxidized to FAD generating H₂O₂, the acetyl-TPP reacts with inorganic phosphate to create acetyl-phosphate, and TPP is regenerated ^57,58 (Fig. 3). A similar catalytic mechanism would be expected for S. sanguinis SpxB. However, LpPox can substitute any of the divalent metal ions Mg²⁺, Mn²⁺, and Ca²⁺ to fix the di-phosphate moiety of TPP, while the S. sanguinis SpxB seems to favor Mg^{2+ 52}.

Fig. 3:

Generalized overview of pyruvate oxidase mechanism, Fenton reaction, and DNA damage. The pyruvate oxidase catalyzes the oxidative decarboxylation of pyruvate and requires the cofactors TPP and FAD. During the multistep reaction, two reducing equivalents are transferred from TPP to FAD, yielding 2-acetyl-thiamine pyrophosphate and FADH₂. FADH₂ is re-oxidized by O₂ which generates H₂O₂ and FAD. Ac-TPP is then cleaved phosphorolytically to yield acetyl phosphate and TPP ¹⁴⁶. Fe²⁺ is able to non-specifically bind to DNA and react with the generated H₂O₂ in the Fenton reaction to cause DNA damage in the form of DNA strand breakage (as shown here) or base alterations in the DNA.

i.i). Distribution and chromosomal location of spxB in oral streptococci

SpxB-dependent H₂O₂ generation is conserved among most of the oral streptococci. Using the amino acid sequence of SpxB from S. sanguinis SK36 as a reference, we performed a BlastP search of the entire Human Oral Microbiome Database (http://www.homd.org/) ⁵⁹. A surprisingly high conservation of the amino acid composition (≥95 identity/≥98% similarity) and protein length (591 amino acids) was observed among oral streptococci (Table 1).

Among streptococci, spxB is typically expressed as a single gene controlled by its own promoter (Fig. 4). In several species, the immediate upstream open reading frame encodes a P-type ATPase copper transporter. Copper is a trace metal that can participate in the Fenton reaction and is also able to disrupt the iron sulfur clusters of various proteins, several of which participate in a variety of cellular processes influenced by H₂O₂. Accordingly, exposure of Streptococcus pyogenes to H₂O₂ leads to the induction of genes with predicted roles in iron sulfur cluster assembly ⁶⁰. Therefore, a coordinated regulation of copper homeostasis and hydrogen peroxide production could be one mechanism employed as a strategy to mitigate the potential toxicity triggered by iron-sulfur cluster protein oxidation.

Fig. 4:

Chromosomal context of spxB in multiple oral streptococci. The sequences of the given species and strains as well as the displayed gene names and functions were obtained from NCBI. Species were sorted with respect to their analogy in gene assortment. Related or functionally connected genes are shown in matching colors. In case the gene function was not already given in the NCBI database (“conserved hypothetical protein” or “hypothetical protein”), related or functionally connected genes were identified by BLAST sequence comparison. Figures were manually generated using Microsoft Powerpoint. Catabolite responsive elements (cre boxes) are indicated (♦) and were predicted as previously described ¹²¹. The respective cre boxes of S. sanguinis and S. gordonii were experimental validated ¹²³.

i.ii). Evidence of in vivo expression

In an effort to demonstrate the in vivo relevance of pyruvate oxidase, spxB gene expression was quantified in clinical plaque samples from different individuals ⁶¹. Universal spxB primers were developed for streptococci and could readily amplify spxB from isolates identified as H₂O₂ producers on Prussian Blue agar plates ^47,48. Conversely, isolates exhibiting no H₂O₂ production also failed to yield spxB PCR amplicons ⁶¹. The expression characteristics of spxB was also determined in these clinical plaque samples. While spxB transcript levels were quite variable between the 9 subjects, the day to day variability was quite low and seemed relatively constant over time for one selected subject ⁶¹. The ability to detect spxB expression in oral plaque samples points to its active role in oral streptococcal physiology. Moreover, spxB is expected to play a particularly important role in early biofilm development, since SpxB-positive streptococci comprise significant fractions of newly formed oral biofilms ⁶² and these biofilms also contain significantly higher oxygen concentrations compared to older, mature biofilms ⁶³.

ii). Other sources of streptococcal hydrogen peroxide production

Another commonly encoded H₂O₂-producing enzyme of oral streptococci is the FMN (Flavin mononucleotide)-dependent lactate oxidase (LctO; sometimes referred to as Lox) ^64,15. The metabolic function of LctO is the aerobic oxidation of L-lactate to generate pyruvate. Since pyruvate is also the substrate for SpxB, both LctO and SpxB are linked in the central metabolism of oral streptococci during aerobic growth. The catalytic mechanisms of SpxB and LctO are different, since no TPP is involved in LctO oxidation of L-lactate. However, for both enzymes the cofactor FMN is initially reduced to FMNH₂ and then subsequently oxidized to regenerate FMN and produce H₂O₂ ⁶⁵.

The LctO-dependent inhibitory ability of Streptococcus oligofermentans toward S. mutans has been studied in detail ^66–68. (Note: Streptococcus oligofermentans has been recently re-classified as Streptococcus cristatus ⁶⁹, but for the reminder of the review we adhere to its old designation). Overall, the H₂O₂-dependent inhibition of S. mutans by S. oligofermentans seems to be a coordinated event between SpxB and LctO, with SpxB producing the majority of H₂O₂ during early and logarithmic growth and LctO dominating during stationary phase. This might be partially due to the fact that the lactate concentration is initially limited, thus precluding LctO from producing much H₂O₂. However, lactate will accumulate during growth and can later be used by LctO to produce pyruvate and H₂O₂. Interestingly, inactivation of SpxB in S. oligofermentans has a negative effect on the activity of LctO, further suggesting these two enzymes are coordinately controlled. Furthermore, in the context of a multispecies biofilm, LctO activity would presumably play an important ecological role in the presence of copious lactate producers like S. mutans. It has been suggested that this lactate-dependent oxidase activity ensures successful competition with S. mutans and may partially explain the high prevalence of S. oligofermentans in certain caries-free individuals ^68,70. Other than S. oligofermentans, the role of LctO has only been studied in S. pyogenes ⁷¹ and S. pneumoniae ⁷². Therefore, it remains to be determined whether other oral streptococci can exploit LctO activity to inhibit cariogenic species.

In addition to SpxB and LctO, S. oligofermentans may utilize another oxidase to produce competitive, S. mutans-inhibiting quantities of H₂O₂ ^73–75. This enzyme was initially classified as an L-amino acid oxidase (LAAO) because of its ability to use peptone and amino acids such as L-aspartic acid, L-tryptophan, L-lysine, L-isoleucine, L-arginine, L-asparagine and L-glutamine to produce significant amounts of H₂O₂. Tong and colleagues suggested LAAO activity could be important under specific growth conditions ⁷⁴. For example, when saliva is used as sole nutrient source, polypeptides might be catabolized into free amino acids that could serve as LAAO substrates to yield α-keto acids, ammonia, and H₂O₂. However, the activity of the enzyme was later revised and aminoacetone was reported to be the preferred substrate for purified LAAO ⁷⁵. This was independently confirmed and the structure of the enzyme solved. Thus, the enzyme was renamed to aminoacetone oxidase or AAO. Its activity previously measured using amino acids substrates is therefore likely the result of either promiscuous catalysis on non-preferred substrates or an artifact of the in vitro reaction conditions ⁷³. Nonetheless, AAO seems to be a unique enzyme in oral streptococci and there is strong evidence to suggest that S. oligofermentans acquired this gene via horizontal gene transfer ⁷⁶. Although LctO-mediated inhibition of S. mutans is ten-fold stronger as compared to AAO ⁷⁴, it might still provide a selective advantage for S. oligofermentans under certain ecological conditions. Unlike SpxB and LctO, AAO also has the added advantage of producing ammonia as a byproduct ⁷³, which, like the arginine deiminases of other commensal oral streptococci, might similarly provide protection against plaque acidification ⁷⁷.

Other fundamental redox reactions in the cell can also yield modest amounts of H₂O₂. For example, the regeneration of NAD⁺ via NADH oxidase can yield H₂O₂ ⁷⁸. In general, NADH oxidases reduce molecular oxygen to either H₂O₂ (via Nox-1) or H₂O (via Nox-2) using NADH as a substrate ⁷⁸. Among streptococci, differences exist on the distribution of the nox-1/2 genes. S. mutans encodes nox-1 and nox-2 ⁷⁹, while S. sanguinis only encodes nox-2 ⁷⁸. The H₂O₂ production capacity of nox-1 is apparently limited, since S. mutans does not produce inhibiting amounts of H₂O₂ ⁴⁵_. In a recent study, Ge et al. demonstrated that a nox-2 knock-out mutant in S. sanguinis loses its ability to compete with S. mutans due to a decrease in extracellular H₂O₂ production ⁷⁸. Interestingly, this effect on H₂O₂ production was suggested to be indirect since the intracellular concentration of H₂O₂ was actually slightly increased and no deleterious impacts were observed upon SpxB activity. The authors also reported that the nox-2 deletion affects membrane fluidity, which has led to the speculation that the observed lack of S. mutans inhibition might result from a decrease in H₂O₂ diffusion across the membrane ⁷⁸.

Self-compatibility and the response to hydrogen peroxide stress

When comparing the aerobic growth of H₂O₂-producing streptococci in the presence and absence of externally added catalase, it is evident that H₂O₂ production yields self-toxicity ⁸⁰. Typically, streptococci do not encode catalase, suggesting that the benefit from excreting relatively large amounts of H₂O₂ outweighs the potential negative effects its production might pose on the producer itself. The absence of catalase or other commonly employed H₂O₂-detoxifying enzymes in oral streptococci also suggests that other important, perhaps uncharacterized mechanisms are involved in protection from H₂O₂ toxicity. Certainly, it is also possible that in the oral cavity H₂O₂ toxicity might not be as pronounced as under laboratory conditions. Saliva has an inherent capacity to deplete H₂O₂ via the salivary lactoperoxidase system, which catalyzes the oxidation of thiocyanate by H₂O₂ to produce the antimicrobial compound hypothiocyanite ^81–83. It is currently unclear to what extent, if any, that lactoperoxidase might influence the local H₂O₂ concentrations in oral biofilms. Furthermore, saliva flow aides in the dilution of H₂O₂ ⁸⁴. However, saliva flow can be limited, for example, in biofilms or interproximal spaces. The actual concentration of H₂O₂ in saliva or at the biofilm-tooth interface has not been determined in vivo. The detection is hindered by several problems, including the presence of the salivary lactoperoxidase system which can eliminate H₂O₂ before its reliable detection. The lactoperoxidase substrates (thiocyanate) and products (hypothiocyanite) can be measured in saliva however, and this allows salivary H₂O₂ concentrations to be extrapolated to approximately 10 µM ⁸⁵. We currently lack the sensor technologies required to directly measure H₂O₂ at the biofilm interface of patients, but one would expect this value to be significantly higher than the 10 µM value estimated for saliva. Recent real-time measurements of the H₂O₂ concentration 100 μm above an in vitro polymicrobial biofilm found up to 1.4 mM of H₂O₂ can be generated ⁸⁶. This suggests that during early biofilm development, in which there is an abundance of SpxB positive streptococci, considerable amounts of H₂O₂ can be produced that presumably requires efficient cellular defense mechanisms to avoid H₂O₂-dependent toxicity.

A key bacterial mechanism to prevent H₂O₂ toxicity is through the avoidance of the Fenton reaction. We have previously demonstrated in S. sanguinis and S. gordonii that mutants lacking Dps, a ferritin-like protein involved in iron sequestration from DNA, are about 10⁴-fold more susceptible to exogenous H₂O₂ compared to their respective wild types ⁴³. A similar hypersensitivity was also reported for S. mutans as well as other streptococci, suggesting that Dps is a broadly conserved and critical protein for the defense against H₂O₂ ^87,88. In S. mutans, superoxide dismutase (Sod) mutants exhibit a similar hypersensitivity as Dps mutants ⁸⁹, while the same Sod mutations in S. sanguinis and S. gordonii only result in about two to three-fold increases in H₂O₂ sensitivity ⁴³. TrxB mutants of S. sanguinis and S. gordonii also exhibit dramatically different H₂O₂ sensitivity phenotypes. TrxB is a thioredoxin reductase that catalyzes the reduction of thioredoxin and plays important role in the formation of disulfide bonds in oxidized proteins. An S. sanguinis TrxB mutant is severely impaired in its H₂O₂ resistance, whereas the S. gordonii TrxB mutant exhibits no obvious survival defects after H₂O₂ challenge ⁴³.

Studies of the S. pneumoniae pyruvate oxidase SpxB demonstrated an unexpected direct role for SpxB in the protection against H₂O₂ toxicity. S. pneumoniae SpxB deletion mutants are 10² to 10³-fold more susceptible to exogenously added H₂O₂ ⁹⁰. The source of this change in H₂O₂ susceptibility was determined to be dependent upon post translational modifications of SpxB itself, in which sulfenylation of SpxB occurs during endogenous H₂O₂ production as part of a reactive oxygen species adaptation mechanism ⁹¹. Although not entirely understood, the sulfenylation of SpxB is postulated to serve as an “H₂O₂ sink” that scavenges intracellular H₂O₂ ⁹¹. Thus, S. pneumoniae has apparently evolved an intriguing mechanism to use SpxB to mitigate the oxidative stress that is created through its own catalytic activity. However, this mechanism does not appear to be a conserved characteristic of SpxB orthologs, as spxB mutations do not impact H₂O₂ sensitivity for either S. sanguinis or S. gordonii ⁴³. Based upon our current knowledge of oxidative stress tolerance, it would appear that the basic machinery employed by oral streptococci are part of a common toolbox. Yet, the observed differences among streptococci for proteins like SpxB, Sod, and TrxB suggest that there are significant aspects of H₂O₂ resistance that have yet to be discovered in these species. This is particularly evident from a recent publication illustrating a wide variety of oxidative stress tolerance phenotypes among different strains of S. mutans that all exhibit similar expression patterns for known components of the ROS response machinery ⁹². For further information about streptococcal ROS response mechanisms, see ⁹³.

Hydrogen peroxide production and biofilm ecology

Our view of oral disease development has evolved in the last two decades due to advances in our understanding of the ecology within oral biofilms ⁹⁴. It is now widely accepted that caries and periodontal disease are byproducts of oral dysbiosis ⁹⁵. However, due to a historical focus upon pathogenic bacteria, a major aspect of oral health has been relatively understudied: the protective abilities of the commensal flora. Currently, there is a paucity of studies revealing the biological mechanisms that support symbiosis, a research area that we refer to as molecular commensalism ²⁰. A better understanding of such mechanisms could provide valuable new strategies to prevent dysbiotic diseases. In fact, there is already evidence that such an approach is feasible. By promoting the growth of alkali-producing oral streptococci via arginine supplementation, it is possible to achieve a net increase in dental plaque pH and lower caries scores ^77,96. H₂O₂ production is another aspect that could be exploited, as H₂O₂ is specifically associated with the commensal flora. Cariogenic species like S. mutans and most periodontopathogens are all exquisitely sensitive to H₂O₂ toxicity ^15,97.

i). H₂O₂ and interspecies competition in the oral biofilm

To better understand how H₂O₂ production is influenced by the immediate biofilm environment, scanning electrochemical microscopy (SECM) has been used to monitor H₂O₂ production in real-time at the surface of an S. gordonii-S. mutans dual species biofilm ⁹⁸. In combination with a pH probe, Joshi and colleagues identified the Achilles heel of S. gordonii H₂O₂ production. In an in vitro dual species biofilm assay grown in artificial saliva, S. gordonii is able to produce a considerable amount of H₂O₂ at a neutral pH. In fact, S. gordonii can slow the process of biofilm acidification by increasing spxB expression and H₂O₂ production once the environmental pH starts to fall due to sugar fermentation. This adjustment in H₂O₂ production provides a temporary check on S. mutans growth. However, once the environmental pH falls below the physiological buffering capacity of saliva, both spxB expression and H₂O₂ production diminish, which allows S. mutans to eventually dominate ⁹⁸. This clearly shows that a prolonged reduction of the salivary pH not only promotes the growth of aciduric species, but also impairs the ability of some commensal species to effectively antagonize H₂O₂-sensitive pathobionts like S. mutans ⁹⁸.

The effect of the environmental pH on the H₂O₂ production capability of oral streptococci seems to be a key issue in their competitive abilities. Contrary to the observation with S. gordonii, S. sanguinis exhibited no change in the production of H₂O₂ when growing at pH6, demonstrating that the response to a lower pH is not a universal decrease in H₂O₂ production ⁹⁹. Although the observed phenotypic response showed consistency among different strains, the authors would like to emphasize that the ability of streptococci to adjust to environmental perturbations needs to be assessed on the species and sometimes strain level and should not be generalized ^92,99. Interestingly, the observed reduction of spxB expression and H₂O₂ production in S. gordonii leads to a redirection of the metabolic flux at the pyruvate node ⁹⁹. Since less pyruvate is metabolized through SpxB catalytic activity, an increase in lactic acid production was observed, which is also reflected in an increased expression of the lactate dehydrogenase gene ldh. The increase in lactic acid production itself is an adaptation mechanism improving the survival of S. gordonii at lower pH ⁹⁹. S. sanguinis in general produces less H₂O₂ than S. gordonii and is more susceptible towards lower pH. Furthermore, the reduced H₂O₂ production at lower pH observed in S. gordonii still exceeds S. sanguinis production ability ⁹⁹. Perhaps, the production of H₂O₂ serves a slightly different ecological purpose since S. sanguinis is one of the most-abundant early colonizers ²⁰ and biofilm mediated protection mechanisms may protect the S. sanguinis population against low pH ¹⁰⁰. At the same time, higher abundance means less energy needs to be channeled towards H₂O₂ production in general, while S. gordonii as a species with lower abundance needs to be more aggressive to ensure competitiveness.

The impact of environmental pH upon competitive H₂O₂ production has also been examined in other species ^99,101,102. S. oligofermentans, a potent H₂O₂ producer, is able to inhibit the growth of S. mutans at pH6 in deferred antagonism assays, which is quite surprising, since S. oligofermentans growth at pH6 is compromised to a much greater degree compared to S. mutans. Even with its impaired growth at pH6, S. oligofermentans still produces up to 1.5-fold more H₂O₂ compared to pH7 ¹⁰¹. At a pH of 5.5, S. oligofermentans growth defects are too severe to remain competitive with S. mutans, despite its aggressive H₂O₂ production. The inhibitory spectrum of commensal oral streptococcal H₂O₂-production was also confirmed for periodontopathogens like Aggregatibacter actinomycetemcomitans, Prevotella intermedia, and Porphyromonas gingivalis ^97,103, suggesting that H₂O₂ production might similarly influence the composition of subgingival biofilms.

ii). Tipping the balance – influencing the production of H₂O₂

A key outcome for studies of molecular commensalism would be to develop new approaches to sustain a symbiotic biofilm community, even under conditions that otherwise would lead to dysbiosis. Conceivably, this could be achieved either by increasing the number of health-associated bacteria, such as SpxB-encoding streptococci, and/or by increasing the H₂O₂ output of producer species. Clinically, it has been shown that S. sanguinis isolates from caries-free individuals have increased H₂O₂ production relative to high caries individuals, which further supports the health-protective role of H₂O₂ ¹⁰⁴. In addition, our previous studies of an S. gordonii CcpA mutant have demonstrated the conceptual utility of exogenously manipulating H₂O₂ production to achieve a desired ecological outcome. CcpA is a transcriptional regulator of spxB and its deletion increases spxB expression and H₂O₂ production ¹⁰⁵. Consequently, CcpA mutants also exhibit an enhanced ability to antagonize the growth of S. mutans. Similarly, in vitro studies of a 3-species S. gordonii, S. mutans, and A. naeslundii biofilm model have demonstrated that L-arginine supplementation can prevent S. mutans dominance and favor S. gordonii growth even in the presence of sucrose ¹⁰⁶, a carbohydrate that strongly promotes the dysbiotic overgrowth of S. mutans ¹⁰⁷. In this system, L-arginine was able to increase the pH of the biofilm, as expected, but it had the added benefit of increasing the production of H₂O₂ by S. gordonii as well. Conversely, in a 14-species biofilm model containing both H₂O₂-producing streptococci and periodontopathogens, the addition of clinically relevant concentrations of the H₂O₂-inactivating enzymes myeloperoxidase, lactoperoxidase, and erythrocyte catalase strongly favored the growth of the periodontopathogens at the expense of the commensals ¹⁰⁸. Overall, there is ample clinical and experimental evidence establishing a clear inverse correlation between the abundance of H₂O₂ produced in an oral biofilm community and the prevalence of pathobionts.

When comparing the oral biofilms of different people, it is evident that considerable variability exists not only at the species level, but even among different strains of the same species ¹⁰⁹. This has important ecological implications for one’s individual likelihood of developing oral dysbiosis. For example, differences in the H₂O₂ susceptibility of an individual’s pathobionts can influence how much H₂O₂ production is required to sustain symbiosis. This is evident from a recent study of S. mutans clinical isolates, which demonstrated how variability among the oxidative stress tolerance phenotypes of these strains has a dramatic influence upon their competitiveness in a multispecies biofilm setting ⁹². S. mutans clinical isolates were mixed with S. oralis and A. naeslundii and grown as biofilms for up to 91 hours. One particular S. mutans strain (Smu81) that was isolated from subgingival plaque ¹¹⁰ exhibited poor competitive abilities after 67 and 91 hours of incubation in the presence of sucrose, while all other S. mutans strains eventually became the dominant biofilm species under the same growth conditions ⁹². In a deferred antagonism assay, S. oralis was able to severely inhibit the growth of Smu81, but this effect was completely abolished in the presence of catalase ⁹². It is still a mystery why different strains of S. mutans can exhibit such a wide range of oxidative stress tolerance.

Similarly, clinical isolates of H₂O₂-producing streptococci also exhibit a range of H₂O₂-production phenotypes. Thus far, we have identified three distinct groups of H₂O₂-producers: group I – high-level H₂O₂ production, carbon catabolite repression (CCR) insensitive; group II – low-level H₂O₂ production, CCR insensitive; and group III – CCR-sensitive H₂O₂ production (Fig. 5). Importantly, like S. mutans oxidative stress tolerance, the H₂O₂-production phenotypes of these isolates were largely determined at the strain level ¹²¹. Further details regarding the role of CCR-dependent regulation of H₂O₂-production will be described later.

Fig. 5:

Representative species for the three identified H₂O₂-production groups and their ability to inhibit S. mutans ± glucose.

As mentioned previously, environmental pH is another factor that can have a profound impact upon H₂O₂ production. In a recent study by Huang and colleagues, alkali-generating arginolytic species were isolated from subjects with and without caries ¹¹¹. Overall, 6 different Streptococcus species were detected with S. sanguinis being predominant (67.9%) followed by S. gordonii (8.9%), S. intermedius (8.9%), Streptococcus cristatus (8.9%), S. australis (3.6%), and S. parasanguinis (1.8%) ¹¹¹. Each of these species also encodes spxB (see Table 1) and further analysis determined H₂O₂ production to be their principal mechanism of S. mutans antagonism ¹¹². While H₂O₂ production is negatively impacted by acidic growth conditions for some streptococci ⁹⁹, the arginine deiminase system is highly acid tolerant ^113,114 and could serve to bolster SpxB activity to prolong competitive H₂O₂ production.

iii). Beyond competition – H₂O₂ signaling, biofilm formation, and community development

Another emerging aspect of H₂O₂ production is its influence on bacterial synergistic interactions, potentially functioning as a signal molecule. The role of streptococcal H₂O₂ in interspecies interactions with A. actinomycetemcomitans has been the subject of several excellent publications ^115–117. Initial investigations with S. gordonii and A. actinomycetemcomitans revealed that the latter is able to utilize the lactic acid produced by S. gordonii as a carbon source, thus providing metabolic complementation and potentially promoting virulence ¹¹⁶. The dilemma for A. actinomycetemcomitans is its susceptibility to the H₂O₂ produced by S. gordonii. To minimize its H₂O₂ exposure, A. actinomycetemcomitans increases the production of its catalase gene katA as well as a novel gene dspB ¹¹⁵. DspB is an extracellular enzyme that promotes the dispersal of A. actinomycetemcomitans biofilms by hydrolyzing the major component of its extracellular matrix, poly-N-acetylglucosamine. This allows A. actinomycetemcomitans to keep a “safe” distance from S. gordonii in mixed species murine abscesses. Normally, A. actinomycetemcomitans and S. gordonii both form small aggregates of about 250–1000 cells in experimental abscesses, while those aggregates increased about two-fold in an A. actinomycetemcomitans DspB mutant ¹¹⁵. Moreover, the DspB mutant aggregates were either contiguous to or within 4 µm of those of S. gordonii, while the wild type maintained a distance between 4 and 13 µm from the S. gordonii aggregates. Presumably, this distance provides the optimal balance between the benefits provided by carbon source acquisition and the limitations of its catalase activity ¹¹⁵. In a separate mixed species study of A. actinomycetemcomitans and S. parasanguinis, it was demonstrated that a close proximity between both species can be maintained by triggering a decrease in spxB expression ¹¹⁸. In fact, dual species biofilms exhibited a significant increase in bio-volume and biofilm thickness as compared to mono-species biofilms, while both species remained intimately associated. Interestingly, A. actinomycetemcomitans catalase did not play an important role in this assay, presumably because the down regulation of spxB limited H₂O₂ production to a level at which katA expression is unnecessary.

A similar phenomenon has been reported for mixed species biofilms of S. gordonii and A. naeslundii ¹¹⁹. Like A. actinomycetemcomitans, A. naeslundii is very susceptible to inhibition by H₂O₂ and encodes catalase. Interestingly, in co-cultures with A. naeslundii, S. gordonii exhibits reduced oxidative damage to its surface proteins, which is likely a benefit provided in trans from A. naeslundii catalase ¹¹⁹. Thus, certain combinations of species might prove beneficial to H₂O₂ producers by minimizing the self-toxicity associated with their H₂O₂ production. Taken together, multiple lines of evidence support the role of H₂O₂ in shaping the ecology within the polymicrobial dental biofilm through both pathobiont inhibition and interspecies interactions. Further knowledge of the biogeographical landscape of the numerous polymicrobial physical ¹² and chemical interactions ¹²⁰ will be required to fully understand the dynamics of these interactions and how they may influence the balance between symbiosis and dysbiosis.

Genetic regulation of streptococcal spxB expression

The expression of spxB is under the control of several transcriptional regulators that are conserved in oral streptococci and respond to either metabolic or environmental stimuli. Some species- and strain-specific variations exist in their mode of control and it is currently unknown how these different control elements are coordinated in a hierarchy of transcriptional regulation. In addition, a specific regulatory mechanism has yet to be established for the decreased spxB expression observed during the switch from aerobic to anaerobic growth conditions. Likewise, it is currently unclear whether the previously described pH dependence of H₂O₂ production is the result of transcriptional regulation, post-transcriptional regulation, or an indirect consequence of a metabolic adjustment.

i). CcpA

Oral streptococci have access to and metabolize a large variety of carbohydrates. Glucose is the most important and preferred carbon source, but its concentration in the oral cavity fluctuates as a consequence of the host diet. In the presence of glucose, most streptococci will suppress the expression of genes involved in the utilization of other carbon sources, a process referred to as carbon catabolite repression (CCR) ¹²¹. We previously identified distinct spxB transcriptional responses to glucose from S. sanguinis and S. gordonii. S. gordonii follows the classic carbon catabolite repression response and decreases spxB expression in the presence of glucose, while spxB expression remains unchanged in S. sanguinis, although its expression is consistently lower than in S. gordonii overall ^105,122,123. Given this difference in regulation, it was surprising to find a consensus sequence for the catabolite control protein A (CcpA) (cre) in both species. CcpA is the main global regulator for carbohydrate metabolism in Gram-positives ¹²¹. Interestingly, S. pneumoniae spxB expression is also controlled by CcpA, but is not affected by glucose similar to S. sanguinis ¹²⁴. The regulatory role of CcpA has been confirmed by mutagenesis, and as expected, spxB expression is increased in the CcpA mutants of both S. sanguinis and S. gordonii, suggesting that S. sanguinis CcpA-dependent control of spxB expression is glucose independent ¹²³. To further explore the regulatory role of CcpA in both species, promoter binding studies were performed and confirmed to depend upon the two cre sites for both species ¹²³. Despite this, it is still not clear why the two species differ in their CCR response for spxB, but we have some evidence indicating that a regulatory mechanism upstream of CcpA is most likely responsible. Furthermore, we observed similar carbohydrate-dependent and carbohydrate-independent spxB control mechanisms from a variety of clinical isolates of different oral streptococci, which indicates that the distinctions between S. gordonii and S. sanguinis are commonly found in other species as well ¹²³.The ecological implications of this differential regulation of spxB are still unclear and this is a subject of active investigation.

ii). SpxR

One of the first spxB transcriptional regulators identified in S. pneumoniae is SpxR. Mutational studies confirmed the role of SpxR as a transcription activator of spxB gene expression. SpxR mutants exhibit decreased H₂O₂ production and are severely impaired for virulence in a murine infection model ¹²⁴. The regulatory role of SpxR was also demonstrated in S. sanguinis ¹²⁵. The domain architecture of SpxR contains a helix-turn-helix motif for DNA binding as well as CBS and HotDog domains. CBS and HotDog domains seem to be ancient and ubiquitous motifs with potential sensory functions monitoring the cellular energy status by binding to AMP, ATP, or S-adenosyl methionine ¹²⁶. Thus, it has been suggested that SpxR connects spxB expression to the energy and metabolic status of the cell ^124,125, but this prediction has yet to be experimentally confirmed and the signal sensed by SpxR remains unknown.

iii). SpxA1 and SpxA2

SpxA1 and SpxA2 are both transcriptional regulators implicated in the control of spxB expression. Both regulators appear conserved among streptococci and fulfill several important functions in oxidative stress tolerance as well as biofilm and competence development. A detailed analysis in S. sanguinis demonstrated that a deletion of spxA1 decreased the production of H₂O₂ threefold along with a similar reduction in spxB expression. In the spxA2 deletion background, H₂O₂ production was not found to be significantly affected ¹²⁷, but this result has been recently challenged ¹²⁸. Interestingly, SpxA1/2 exhibit significant homology, including a Cys-X-X-Cys motif that has previously been shown to sense the intracellular redox state ¹²⁹. Further studies are required to better understand the mechanistic role of both transcriptional regulators for spxB regulation.

iv). Two-Component Systems (TCS) VicRK and SptRS

Environmental information is commonly transmitted via TCS signaling in bacteria ^130,131. Two S. sanguinis TCS have been previously demonstrated to influence spxB gene expression ^132,133. The initial discovery that the TCS VicRK positively regulates spxB expression as well as biofilm formation and the release of extracellular DNA demonstrates that H₂O₂ production is involved in important cellular developmental processes ¹³³. This is consistent with our previous studies of the connection between spxB expression, H₂O₂ production, and the induction of extracellular DNA release ¹³⁴. More recently, the S. sanguinis SptSR TCS has also been demonstrated to control spxB expression ¹³². In contrast to the VicRK TCS, a deletion of either sptS or sptR increases the production of eDNA and H₂O₂. While VicR was shown to bind the spxB promoter ¹³³, this was not possible for SptR, although a putative consensus binding site was identified in the spxB promoter ¹³². Therefore, it is currently unknown whether the regulatory effect of the SptRS TCS is direct or indirect. The latter is a distinct possibility, since the sptR and sptS mutations trigger significant transcriptional changes in the regulatory genes vicR and spxR ¹³².

Streptococcal H₂O₂ production at extra-oral locations

The reservoir of oral streptococci is not confined to the oral cavity, as S. dentisani, S. tigurinus, S. oralis, S. oligofermentans, S. mitis, S. infantis, and S. gordonii have all been isolated from a various areas of nasopharynx from healthy adults ¹³⁵. In addition, it has been suggested that streptococci are generally protective for paranasal sinuses to certain diseases such as chronic rhinosinusitis ¹³⁶. Although there is currently no experimental evidence of H₂O₂ production as a key facet of this protection, the availability of oxygen in the nasopharynx would certainly allow for its production.

Oral streptococci are also commonly associated with the lung microbiome, and their presence in patients with either chronic obstructive pulmonary disease (COPD) or cystic fibrosis (CF) has been shown ^137,138. For COPD patients, the streptococcal community in the lungs was shown to originate primarily from the oral cavity through aspiration based upon comparisons of the overall similarity of the bronchial, peripheral lung, nasal, and oral microbiomes ¹³⁸. This suggests that aspiration and colonization by oral streptococci is likely a common occurrence. A recent clinical study found that a pediatric control cohort had a higher percentage of Streptococcus in their broncho-alveolar lavage fluid compared to an age-matched cohort of pediatric CF patients, while the study failed to detect any Streptococcus in adult CF patients ¹³⁷. Over time, the non-traditional CF taxa like the streptococci seem to be replaced by the traditional CF taxa including Pseudomonas, Staphylococcus, and Stenotrophomonas ¹³⁷. This observation may indicate that the non-traditional taxa including Streptococcus can colonize better and initially provide some level of protection. It has also been hypothesized that oral streptococci may play an important role in increasing the diversity of the cystic fibrosis microbiome, thus promoting patient stability ¹³⁹. Mechanistically, this is consistent with the findings by Scoffield and Wu, which demonstrated that oral streptococci successfully interfere with Pseudomonas aeruginosa through H₂O₂ production ⁴⁶. Surprisingly, multiple streptococci including S. parasanguinis, S. sanguinis and S. gordonii all failed to inhibit the growth of P. aeruginosa in antagonism assays when using regular growth medium, but exhibited inhibition when nitrite was supplemented. This effect was also SpxB-dependent. In a Drosophila melanogaster infection model, the authors demonstrated that H₂O₂-producing oral streptococci are similarly protective against P. aeruginosa in the presence of nitrite ⁴⁶. Nitrite concentrations seem to be elevated in the sputum of CF patients ¹⁴⁰ and can be converted to peroxynitrite in the presence of H₂O₂. It has been suggested that peroxynitrite is potentially internalized through a permease to exert its antimicrobial effect upon P. aeruginosa ⁴⁶. It is also worth noting that multiple CF studies have detected oral streptococci from the Streptococcus anginosus group (formerly Milleri group) as being strongly associated with CF pathogenesis, namely S. intermedius and S. constellatus ^141,142. Coincidently, after searching the annotated genomes available in the Human Oral microbiome Database (HOMD), both S. intermedius and S. constellatus were not found to encode spxB and are thus not expected to be significant H₂O₂ producers.

In conclusion, oral streptococci are present in multiple sites of the human host. Most likely, their dissemination originates from the oral cavity. In future studies, it would be interesting to determine whether commensal oral streptococci exert similar health protective benefits outside of the oral cavity via their H₂O₂ production.

Outlook

Hydrogen peroxide production plays an important role in the metabolism of oral streptococci and effects the ecology of dental biofilms. Experimental evidence over the last decade supports a critical role for H₂O₂ in maintaining oral symbiotic homeostasis. With our deeper understanding of the ecological aspects of oral diseases, it has been suggested that future treatments are likely to remodel dysbiotic communities back to a state of symbiosis with the host ^19,20,143. This concept is a promising alternative to conventional therapies, but its translation is only likely to occur with a more thorough understanding of molecular commensalism ²⁰. The manipulation of H₂O₂ production among the commensal flora is certainly one conceivable strategy to prevent and/or reverse dysbiosis.

Furthermore, a recent systematic review about the diagnostic accuracy of current Caries Risk Assessment (CRA) methods concluded “the analyzed methods may lead to patients with increased risk not being identified” ¹⁴⁴. Clinical CRA guidelines approved by the American Dental Association fail to consider the polymicrobial nature of caries and lack any molecular markers sampled directly at the tooth surface where caries occurs ¹⁴⁵. The identification of such microbial biomarkers to monitor changes at the tooth surface before the clinical manifestation of caries is detectable would greatly improve early interventions. Currently, the caries disease process cannot be accurately determined in the absence of observable demineralization. Therefore, predictive preclinical biomarkers are clearly desirable and much needed. H₂O₂ production might prove to be a useful biomarker of the underlying ecology within the oral cavity.