Streptococcus mutans: a new Gram-positive paradigm?

History of S. mutans research

In 1924, J. Clarke isolated an organism from carious lesions and called it S. mutans, because he thought the oval-shaped cells observed were mutant forms of streptococci (Clarke, 1924). However, it was only in the late 1950s that S. mutans received greater attention from the scientific community and, by the mid 1960s, it was recognized as a major aetiological agent in dental caries (Loesche, 1986). In the subsequent two decades, researchers began to uncover the pathophysiology of S. mutans. During this period, the first tools for studying S. mutans, both in vitro and in vivo, were developed. As a result of the efforts of these pioneer researchers, the major virulence traits of S. mutans were established: (i) the ability to produce large quantities of organic acids (acidogenicity) from metabolized carbohydrates; (ii) the ability to survive at low pH (aciduricity); and (iii) the ability to synthesize extracellular glucan-homopolymers from sucrose, which play a critical role in initial attachment, colonization and accumulation of biofilms on tooth surfaces (Banas & Vickerman, 2003; Bowen & Koo, 2011; Burne, 1998; Loesche, 1986). With the advances in molecular genetic techniques in the 1980s and 90s, scientists began to more rapidly understand how metabolic pathways enabled S. mutans to evolve into a specialized dental pathogen. Finally, with the release of the first complete S. mutans genome in 2002 (Ajdić et al., 2002), the scientific community took full advantage of the technologies that emerged in the genomic era, applying functional genomic, transcriptomic and proteomic approaches to better dissect the physiology, genetics and virulence mechanisms of S. mutans.

S. mutans as a model organism of pathogenic Gram-positive bacteria

Contrary to B. subtilis, which does not infect humans and has a free-living lifestyle, S. mutans is an obligate human pathogen. Thus, the knowledge gained from studies with S. mutans may be highly relevant to other pathogenic non-sporulating Gram-positive bacteria. With a relatively small genome of about 2 Mb with over 1900 genes (Ajdić et al., 2002), the S. mutans UA159 genome is much closer in size to those of closely related streptococcal and staphylococcal species (ranging from about 1.8 Mb in Streptococcus pyogenes to about 2.8 Mb in Staphylococcus aureus) (Table 1). More recently, next-generation sequence technology was used to obtain genome sequences of 57 clinical isolates of S. mutans (Cornejo et al., 2012). The S. mutans core genome was estimated to contain close to 1500 genes, with each genome containing, on average, 1636 genes. With a genome about half the size of E. coli (~4.7 Mb) or B. subtilis (~4.2 Mb) and with a GC content less than 40 %, it is, therefore, not surprising that a greater overlap in the mechanisms of gene regulation and metabolic pathways occurs between S. mutans and related Gram-positive pathogens, which also have compact genomes, low GC content and, often, a host-associated lifestyle.

Like E. coli and B. subtilis, S. mutans is highly amenable to genetic manipulation, is naturally competent, and is relatively safe to manipulate in the laboratory setting. Today, a large number of fast and reliable genetic tools are available to study the biology of S. mutans, including PCR-ligation mutagenesis (Lau et al., 2002), markerless in-frame deletion systems (Atlagic et al., 2006; Banerjee & Biswas, 2008; Merritt et al., 2007), gene knockdown strategies (Lemos et al., 2007b; Wang & Kuramitsu, 2005), integration vectors for genetic complementation or expression of foreign genes (Jung et al., 2009; Wen & Burne, 2001; Zeng & Burne, 2009), and a variety of E. coli–streptococcal shuttle vectors and reporter systems for gene expression analysis. In addition to sophisticated genetic tools, there are also many simple and efficient in vitro models to study S. mutans biofilms, which include batch and continuous-feeding models, as well as flow systems attempting to mimic salivary flow (Lemos et al., 2010). Notably, most of these systems are suitable for the development of more complex multispecies biofilms (Bradshaw & Marsh, 1999; McDermid et al., 1986; Shu et al., 2003). These model systems can be easily adapted for studies with non-oral and non-streptococcal species. Some of the most recent studies using S. mutans biofilm models are briefly discussed below.

Despite the advantages of having powerful analytical tools to study the genetics and biology of pathogenic bacteria, the translation of in vitro findings to animal models is critical for a complete understanding of bacterial pathogenesis. In this regard, there are well-established in vivo models for S. mutans, including rat (Burne et al., 1996; Koo et al., 2005; Yamashita et al., 1993) and transgenic mouse caries models (Burne et al., 1996; Catalán et al., 2011; Culp et al., 2005; Yamashita et al., 1993), as well as rabbit and mouse models of infective endocarditis (Bahn et al., 1978; Paik et al., 2003). In fact, it was the pioneering work conducted in the 1950s and 60s using hamsters and gnotobiotic rats that established S. mutans as the principal aetiological agent of dental caries (Fitzgerald & Keyes, 1960; Orland et al., 1955; Zinner et al., 1965). More recently, several studies have demonstrated the usefulness of the infection of the larvae of Galleria mellonella to study the pathogenic potential of S. mutans strains in systemic infections (Abranches et al., 2011; Gonzalez et al., 2012; Kajfasz et al., 2010). Despite its obvious limitations, this increasingly popular model may be viewed as a powerful and inexpensive screening tool for studying bacterial fitness and virulence, in particular for traits that are associated with resistance to host-derived stress factors during systemic diseases.

The oral biofilm: a battlefield of microbial interactions

Microbial biofilm communities are found on mucosal surfaces, including the gut and in the vaginal, oral and nasal cavities. It is widely recognized that the different interactions, synergistic or antagonistic, between members of these complex communities can shift the bacterial population from health to disease. This is particularly true for the oral cavity, where S. mutans, among other less well-characterized lactic acid-producing bacteria such as lactobacilli and bifidobacteria, tends to dominate the oral flora as caries progresses (Becker et al., 2002; Gross et al., 2010; Tanner et al., 2011a, b). As proposed by Marsh (1994) in the ecological plaque hypothesis, dental caries is a disease that results from a shift in the proportions of the resident flora as a consequence of changes in the environment (e.g. high sugar concentrations/low pH).

The easy access to oral samples from human volunteers and the ability to identify the major contributors in health and disease allowed researchers to probe some key bacterial interactions that disturb the homeostasis of oral biofilms. For example, Streptococcus gordonii and Streptococcus sanguinis, early colonizers of the tooth surface normally associated with health, were shown to antagonize the growth of S. mutans by producing H₂O₂ (Kreth et al., 2008). Likewise, Streptococcus oligofermentans, whose abundance appears to be inversely related to that of S. mutans and the presence of carious lesions, was shown to produce inhibitory quantities of H₂O₂ using lactic acid as a substrate (Tong et al., 2007). In this particular case, S. oligofermentans gains a competitive advantage over S. mutans by directly inhibiting its growth (via H₂O₂ production) and, at the same time, by facilitating the survival of an acid-sensitive flora. Conversely, S. mutans can produce bacteriocins (known as a mutacins) that are inhibitory against early colonizers (Merritt & Qi, 2012). By analysing the transcriptome of S. mutans in mixed-species cultures containing S. mutans, S. gordonii and the early colonizer Veillonella parvula, Liu and colleagues suggested that part of the deleterious effects of the H₂O₂ producer S. gordonii against S. mutans was mitigated by the presence of V. parvula (Liu et al., 2011). Many other studies, which were not restricted to S. mutans, have shown how microbial cell–cell interactions influence the oral microbiota composition and homeostasis (Kuramitsu et al., 2007; Whitmore & Lamont, 2011). Therefore, due to the relatively well-defined microbial composition of healthy and diseased sites and easy access for sampling, the oral biofilm microflora have served, and will likely continue to serve, as an excellent model system for studying bacterial interactions.

S. mutans as a model biofilm organism

It is well established that S. mutans has a biofilm-dependent lifestyle (Bowen & Koo, 2011; Burne, 1998; Marsh, 1994). In the oral cavity, a highly diverse microbial community is constantly interacting with a proteinaceous film (containing salivary proteins and bacterial exoproteins) present on the tooth surface, known as the pellicle. Specific groups of organisms (mostly streptococci and Actinomyces spp.) can adhere to the pellicle in low numbers, and subsequently these organisms co-adhere with other oral microbial species (Nobbs et al., 2009). S. mutans expresses adhesin-like surface-associated proteins (e.g. AgI/II family) that are capable of binding to receptors in the pellicle (Gibbons, 1989). Recently, the adhesin P1 (AgI/II) was shown to be an amyloid-forming protein that contributes to biofilm development by S. mutans (Oli et al., 2012). S. mutans also interacts with specific salivary proteins, such as common salivary protein-1 (CSP-1), which, in turn, helps the bacterium to bind saliva-coated apatitic surfaces (Ambatipudi et al., 2010). Thus, S. mutans can be present in this initial colonizing community on the tooth pellicle, albeit not as one of the most abundant organisms. However, when dietary sugars, particularly sucrose, become available, S. mutans can become the numerically dominant organism, thereby orchestrating the formation of cariogenic biofilms.

S. mutans-derived extracellular glucosyltransferases (Gtfs) are constituents of the pellicle and are capable of synthesizing glucans in situ from sucrose (Schilling & Bowen, 1992). The surface-formed glucans provide additional bacterial binding sites that favour local colonization by S. mutans, as it binds avidly and in large numbers to these polymers (Schilling & Bowen, 1992) through several envelope-associated glucan-binding proteins (Banas & Vickerman, 2003). Furthermore, in situ-formed glucans can mask host-derived bacterial binding sites in the pellicle (Schilling & Bowen, 1992), which may negatively impact adherence of other commensal species that bind more favourably to the pellicle, such as Streptococcus oralis, S. sanguinis and S. gordonii (Jenkinson, 2011; Nobbs et al., 2009). Gtfs also bind to many oral microbes, even those that do not naturally express Gtfs, thereby converting them into de facto glucan producers (Gregoire et al., 2011; McCabe & Donkersloot, 1977; Vacca-Smith & Bowen, 1998). The exopolysaccharides (EPS), mostly in the form of glucans, formed on the pellicle and on microbial surfaces promote local accumulation of microbial cells while forming a diffusion-limiting polymeric matrix that protects embedded bacteria (Koo et al., 2010; Xiao et al., 2012). In parallel, sucrose and other sugars are rapidly fermented by S. mutans within the EPS-rich matrix, creating highly acidic environments.

As discussed above, complex interspecies competition and synergism take place between S. mutans and other species during the process of biofilm development. The low-pH environment further induces EPS production by S. mutans (Li & Burne, 2001), and under these conditions the organism prospers, leading to biofilm accretion and further acidification. As the environmental stress increases (Lemos & Burne, 2008), the microbial diversity is dramatically reduced in favour of a highly aciduric and acidogenic flora (Gross et al., 2012, 2010; Palmer, 2010).

Recently, attention has been devoted to understanding how S. mutans takes advantage of these environmental cues to optimize its metabolism in the presence of other oral bacteria within a dynamically changing biofilm milieu. A combination of in situ EPS imaging and 3D pH mapping approaches with high-throughput quantitative proteomics (MudPIT) has been used to dissect the mechanisms by which S. mutans assembles the matrix and survives within intact mixed-species biofilms. The sequential assembly of a structurally complex and spatially heterogeneous 3D matrix helps to create a highly compartmentalized architecture within intact biofilms (Fig. 2). In turn, highly localized and structured acidic microniches are dynamically formed throughout intact biofilms. The spatiotemporal assembly of a 3D biofilm matrix and the regulation of metabolic and stress survival pathways by S. mutans are highly intertwined, and also are dependent on the presence of other oral species (Klein et al., 2012; Xiao et al., 2012). More specifically, co-culture with the early oral colonizers S. oralis and Actinomyces naeslundii induces the expression of proteins associated with EPS synthesis (GtfB and GtfC) and remodelling (DexA), of the glucan-binding GbpB protein and of acid stress-related proteins such as AtpD, FabM and GroES (see below for more details) (Xiao et al., 2012). While some of these interactions are unique to the oral environment, the knowledge generated from these studies has relevance beyond the mouth, as extracellular matrix and mixed-species communities are common features of many biofilms found in the body and in the environment.

Clearly, S. mutans represents a model that may be useful to further elucidate matrix assembly principles, responses to environmental cues and polymicrobial interactions. These studies can advance current understanding about how bacterial interactions with the matrix environment modulate the dynamics of biofilm development and how these interactions govern the shifts in microbial composition and creation of biological niches within mixed-species biofilms. Furthermore, the novel approaches to study S. mutans biofilms may have broader applicability. These models could be adapted for studies of other biofilm systems in which polymeric matrices are actively formed during biofilm development and varied microenvironments are created, such as those associated with cystic fibrosis and medical device-related infections.

S. mutans is the bacterial paradigm of lactic acid bacteria (LAB) stress responses

When compared to other species of LAB, S. mutans is, by far, the best-studied organism in terms of the genetics and physiology of stress responses. This is not entirely surprising, given that dental caries is directly linked to the ability of certain oral bacteria such as S. mutans to thrive under highly stressful conditions. Studies on S. mutans stress responses have served to reveal critical differences in the way that host-associated bacteria cope with environmental stresses when compared with bacteria that have both a free-living and host-associated lifestyle. In many cases, the stress regulon was shown to control a broader set of biological functions, including energy metabolism, biofilm formation and genetic competence (Lemos et al., 2005; Lemos & Burne, 2008; Matsui & Cvitkovitch, 2010). In the following two sections we will describe the key mechanisms of stress tolerance utilized by S. mutans and highlight studies with broader implications for bacterial stress responses that have served to challenge long-established dogmas.

Key contributions linked to studies on acid stress responses by S. mutans

The ability of S. mutans to cause disease depends on its abilities to irreversibly bind to teeth via the formation of extracellular glucan polymers, and to survive in the acidic environment created by its own metabolism of the sugars present in the human diet. Of note, elegant studies from the Burne group on carbon metabolism and carbon catabolism repression revealed a highly complex, hierarchical and efficient control of energy generation by S. mutans, and also indicated that carbon metabolism is intertwined with the expression of virulence-related traits (Abranches et al., 2003, 2008; Zeng & Burne, 2008, 2010; Zeng et al., 2006).

The pressure of life at low pH values has led to the evolution of interconnected stress response pathways that allow its persistence in the oral cavity. The acid stress response not only facilitates the survival of S. mutans, but also enables the organism to outcompete less aciduric species of oral bacteria. Results from experiments conducted with cultures of mixed oral bacterial species have shown that S. mutans dominates these microbial communities during low-pH growth, whereas less aciduric species, such as S. sanguinis, predominate during growth at neutral pH values (Bradshaw & Marsh, 1998; McDermid et al., 1986). These observations ignited a search for the mechanisms that control acid stress responses in S. mutans, and also generated considerable interest in understanding the interrelationships of the oral microbial flora, which has led to recent metagenomic sequencing projects.

The association between the aciduricity of S. mutans and its ability to outcompete other oral bacteria in mixed cultures can be explained by how the organism maintains intracellular pH homeostasis. S. mutans produces elevated levels of a membrane-bound F₁F₀-ATPase during growth in acidic conditions (Belli & Marquis, 1991). Moreover, counter to the well-known function of the membrane-bound F₁F₀-ATPase in the formation of ATP in E. coli and in mitochondria, the principal function of the S. mutans F-ATPase is to operate in reverse of the established paradigm, by pumping protons out of cells at the cost of ATP hydrolysis. This maintains the intracellular pH close to neutrality, which protects acid-sensitive enzymes in the glycolytic pathway (Bender et al., 1986). It was also shown that S. mutans produces higher levels of F-ATPase, with greater enzymic turnover at low pH, than other oral bacteria, with the exception of the highly aciduric lactobacilli (Bender et al., 1986; Sturr & Marquis, 1992). Thus, S. mutans not only produces prodigious amounts of organic acids, which act to inhibit the growth of acid-sensitive organisms and contribute to enamel demineralization, but also expresses comparatively high levels of F-ATPase to maintain intracellular pH homeostasis and preserve its own metabolic activity (Sutton & Marquis, 1987).

The observations concerning the acid-resistant F-ATPase prompted a subsequent investigation to determine whether changes in the membrane might contribute to F-ATPase activity. This study showed that membrane composition undergoes rapid remodelling during growth at low pH (Quivey et al., 2000). In fact, there is a near reversal of the proportions of saturated and unsaturated membrane fatty acids, with unsaturated fatty acids occupying nearly 60 % of the membrane composition during acidic growth (Fozo & Quivey, 2004; Quivey et al., 2000). Around this time, the fabM gene, encoding an enzyme with trans-2-cis-3-decenoyl isomerase activity, was identified and characterized in Streptococcus pneumoniae (Marrakchi et al., 2002) and shown to be responsible for the formation of unsaturated fatty acids. Parallel studies, conducted with S. mutans, showed that following mutation of the fabM gene, the organism could no longer produce unsaturated fatty acids (Fozo & Quivey, 2004). Moreover, loss of the unsaturated fatty acids resulted in an extremely acid-sensitive phenotype that could be rescued by the inclusion of long-chain, unsaturated fatty acids in the growth medium (Fozo & Quivey, 2004). In addition, inactivation of fabM resulted in a substantial reduction in the ability of S. mutans to cause caries in rats (Fozo et al., 2007). This was a departure from the E. coli paradigm of fatty acid synthesis, in which unsaturated fatty acids arise by the combined action of two enzymes, FabA, a dehydratase that forms unsaturated fatty acids from C₁₀ : hydroxy-fatty acid intermediates and FabB, the condensing enzyme that is capable of elongating unsaturated fatty acids (Marrakchi et al., 2002).

Shifting long-standing paradigms in protein translocation and the stringent response

The signal recognition particle (SRP) is conserved in all three domains of life and is involved in the translocation of membrane proteins (Keenan et al., 2001). The bacterial SRP translocation pathway was previously studied in E. coli and B. subtilis and in both cases was found to be essential for cell viability (Keenan et al., 2001; Phillips & Silhavy, 1992). However, proteins of the SRP pathways were unexpectedly shown to be dispensable in S. mutans (Hasona et al., 2005). Further studies revealed that overlapping and compensatory pathways exist in S. mutans and allow the organism to adapt and maintain membrane function in the absence of the SRP (Funes et al., 2009). Specifically, a gene duplication event enabled the specialization of a protein from the YidC/Oxa1/Alb3 family, involved in the insertion of proteins into membranes of bacteria, mitochondria and chloroplasts, to overlap functionally with the SRP (Funes et al., 2009). Although the SRP mutants showed impaired growth under a variety of stress conditions, this finding suggests that protein biogenesis and translocation pathways in S. mutans, and likely in other closely related organisms, may have fundamentally important differences when compared to the more widely studied organisms.

Another investigation that challenged a long-standing bacterial dogma came during the initial characterization of the S. mutans stringent response. The stringent response is a highly conserved global stress response mechanism that coordinates broad metabolic alterations necessary for adaptation under nutrient-limiting conditions (Potrykus & Cashel, 2008). This response is mediated by the accumulation of two modified guanine nucleotides, pyrophosphorylated GDP or GTP, collectively known as (p)ppGpp. For several decades, a single bifunctional enzyme known as Rsh (or RelA) was considered to be the sole enzyme controlling (p)ppGpp production and degradation in Gram-positive bacteria. However, the serendipitous finding that the S. mutans ΔrelA mutant was still able to produce (p)ppGpp (Lemos et al., 2004) led to the discovery of two small monofunctional (p)ppGpp-synthetase enzymes, named RelP and RelQ, in S. mutans (Lemos et al., 2007a). blast search analysis indicated that at least one of these enzymes was present in virtually every Gram-positive organism (Lemos et al., 2007a), and subsequent studies confirmed the functionality of these enzymes in other Gram-positive bacteria such as B. subtilis, S. pneumoniae and Enterococcus faecalis (Abranches et al., 2009; Battesti & Bouveret, 2009; Nanamiya et al., 2008). The identification of these small (p)ppGpp synthetase enzymes in Gram-positive bacteria reinvigorated the scientific interest in this fundamental stress response pathway. As a result, recent bioinformatic and genetic analysis indicated that single-domain (p)ppGpp-synthesizing and -hydrolysing enzymes can be found, alone or in combination, in a number of organisms, including archaea, bacteria, plants and animals (Atkinson et al., 2011; Das et al., 2009; Sun et al., 2010).

S. mutans competence: lessons for making genetic transformation possible for all streptococci?

Natural transformation (competence) is a phenomenon observed in many bacterial species and has been linked to increased genetic diversity, increased survival, biofilm formation, bacteriocin production and general stress responses (Kreth et al., 2005; Martin et al., 2006; Perry et al., 2009; Suntharalingam & Cvitkovitch, 2005). The molecular mechanisms underlying competence development in S. pneumoniae have been thoroughly dissected (Claverys & Havarstein, 2002). In this organism, as well as in all other competent streptococci, an alternative sigma factor (ComX) functions as a master regulator of competence. ComX is activated by the product of comC, a secreted competence-stimulating peptide (CSP), in a density-dependent (quorum-sensing) manner. Although S. pneumoniae laid the foundation for understanding competence in streptococci, work with S. mutans indicated that specific pathways associated with competence might vary among the different streptococci (Martin et al., 2006; Suntharalingam & Cvitkovitch, 2005). Interestingly, genomic analysis revealed that a complete ComX regulatory circuit and a number of the late competence genes are present in streptococcal species deemed ‘non-transformable’, including members of the pyogenic group (e.g. S. pyogenes and Streptococcus agalactiae). This observation raises the possibility that all streptococci may, under appropriate conditions, be capable of developing competence. However, there was no evidence that these bacteria encode a ComC-like peptide. More recently, a novel double-tryptophan peptide known as sigX-inducing peptide (XIP, encoded by comS) and an Rgg-type regulator (designated comR) responsible for inducing comX expression were identified in members of the salivarius group (e.g. Streptococcus salivarius, Streptococcus thermophilus), which led to the development of protocols for in vitro transformation of S. salivarius and S. thermophilus (Fontaine et al., 2010). Remarkably, ‘non-transformable’ members of the pyogenic and bovis groups of streptococci also possess a regulatory system homologous to ComR/ComS (Mashburn-Warren et al., 2010).

In contrast to S. pneumoniae, which relies on CSP to stimulate competence, S. mutans appears to be unique, as it carries both CSP- and XIP-dependent pathways. Despite an overlap with competence-related gene expression, the CSP regulatory circuit was later found to be dispensable for S. mutans competence and was implicated in non-lantibiotic bacteriocin production (Ahn et al., 2006; Kreth et al., 2005; Mashburn-Warren et al., 2010). Using a microfluidic, single-cell method, Son et al. (2012) provided novel clues on how S. mutans responds to extracellular signals to activate ComX. More specifically, the authors showed that in a chemically defined medium, exogenous CSP does not induce comX, whereas exogenous XIP uniformly activates comX expression. On the other hand, in complex medium, XIP does not induce comX, while CSP elicits a bimodal comX response from the population (Son et al., 2012). These recent findings indicate that our understanding of the mechanisms associated with bacterial competence is far from complete. The ease of genetic manipulation and the presence of the ComR/ComS regulatory pathway place S. mutans in a unique position to further advance our understanding of competence in streptococci and, possibly, help to overcome the barriers that have hampered in vitro genetic transformation of pyogenic streptococci and other streptococcal species.

Concluding remarks

In this article, we have presented evidence that the development of novel technologies and the rapid advances in dissecting the genetics and physiology of S. mutans have resulted in the emergence of this bacterium as a new Gram-positive model organism. The availability of multiple model systems to investigate the role of specific mutations in the pathophysiology of S. mutans, and the ability to recover living bacteria from each of the established models, strengthen the case for the use of this organism as a comparative tool for other Gram-positive pathogens. Because of the enormous amount of knowledge gathered from E. coli and B. subtilis, their contributions to present and future discoveries will continue to inform the development of additional bacterial systems. However, with the annotation of hundreds of complete bacterial genomes, functional genomics and other ‘omic’ technologies, it is now much easier to study other bacterial species. Thus, the progress made in recent decades places S. mutans in an interesting position to further advance basic microbiology research, especially in low-GC Gram-positive organisms.