Beyond the oral microbiome


  • Howard F. Jenkinson

    Corresponding author
    1. School of Oral and Dental Sciences, University of Bristol, Lower Maudlin Street, Bristol BS1 2LY, UK.
      E-mail; Tel. (+44) 117 342 4424; Fax (+44) 117 342 4313.
    Search for more papers by this author

  • Allegories are from the song ‘Hey, hey, my, my (into the black)’ composed and sung by Neil Young (1979).

E-mail; Tel. (+44) 117 342 4424; Fax (+44) 117 342 4313.


The human oral microbiome currently comprises 600–700 taxa, but estimates suggest that overall species numbers may turn out to be higher (∼1200). Within the oral cavity, groups of microbial species become arranged into surface-localized communities that vary considerably in composition according to sites of establishment. Factors such as nutrient availability, pH, toxic metabolites, shear forces and host conditions contribute to modelling the structure and activities of these oral microbial communities. With development of more rapid and accurate molecular techniques it has become possible to begin to characterize the genome contents of individual communities. However, understanding the phenotypic interactions between cultivable microorganisms within communities is essential in order to complement the genomic data. This will then enable construction of microbial community interactomes, incorporating genomic and spatial information with functional knowledge of physical and metabolic interplays between the microorganisms. Enlightenment of the changes in genome composition and phenotypic interactions as functions of niche, time and intrusions will help towards developing better means of manipulating communities for host benefit.


Out of the blue and into the black

The oldest recognized microbial ecosystem is within the human oral cavity. Some of the animalcules seen by van Leeuwenhoek in the white stuff between his teeth, using his home-made microscopes, are clearly visible today in micrographs of dental plaque. Most recognizable would be pairs or chains of spherical streptococci, unusual rod-shaped forms of Actinomyces and Fusobacterium, and the twisting spiralling forms of Spirochetes. After this first look, nearly 360 years passed before it was established that dental caries (tooth decay) was caused by Streptococcus mutans (Fitzgerald and Keyes, 1960). Then, in successively shorter time intervals, a range of anaerobic bacteria such as Bacteroides (now Porphyromonas) gingivalis were cultivated in the 1970s from periodontal disease lesions (Socransky, 1970), and complex mixtures of anaerobic bacteria including Porphyromonas, Peptostreptococcus (now Parvimonas), Prevotella, etc. were cultivated from infected root canals and pulps (Sundqvist, 1994). Recently, it has been possible to identify, by molecular techniques, potentially all of the bacterial species present in the human oral cavity (Dewhirst et al., 2010). There has been an exponential evolution in technologies, e.g. anaerobic culture, PCR, proteomics, pyrosequencing, which has driven this progress.

It may be noticed from the timeline that oral microbiology began with some vision, progressed with better understanding of disease aetiology, and has now come to an unbelievable climax, knowledge of possibly all the microorganisms present within the oral ecosystem. But how do we begin to build upon this knowledge? Can we satisfactorily deduce how communities are formed, when more than 50% of microorganisms cannot yet be cultivated? Is there a core oral microbiome as may be indicated for the gut (Turnbaugh et al., 2009)? The oral cavity environment is arguably more heterogeneous than the gut with regard to morphology, tissue types, pH and flow rates. Which of the microorganisms are most ecologically relevant in respect of oral health or disease? Have the real pathogens been studied? Should we be working with monocultures or with microbial communities? Are oral disease models satisfactory? Have we come out of the blue and into the black?

Molecular taxonomy and diagnosis

Making use of the conserved sequences within 16S rRNA, in conjunction with analysis of non-conserved loops, was a clever development in progressing ideas that species might be delineated rather simply thoughsequence variations of small stretches of rDNA. At the time of writing there are 1 613 000 16S rRNAs (, partial or complete, in the sequence databases. The sequences for oral microorganisms have been collated into the HOMD database (Chen et al., 2010), in which the rRNA dataset has 335 830 rRNA sequences downloaded from Ribosomal Database Project-II database. In early work, primers corresponding to relatively conserved regions at each end of the 16S rRNA gene, so-called universal primers, were utilized to PCR-amplify sequences from total DNA extracted from pools of cultivated oral microorganisms. In this way approximate proportional representations of sequences within a population were made. However, only those rRNA sequences from genes with sequences that exactly matched the primers were generated.

An extension to these methods was introduced with reverse-capture checkerboard hybridization technique, which utilized immobilized 16S rRNA sequences from cultivated or non-cultivated oral bacteria, to hybridize with labelled 16S rRNA or DNA sequences prepared from oral microorganisms (Paster and Dewhirst, 2009). Technically this proved laboursome, and was superseded by oligonucleotide array technology, wherein 16S rRNA-based oligonucleotides were printed onto glass slides, and hybridized with fluorescently labelled rDNA sequences obtained via PCR from clinical samples. The current next-generation sequencing technologies, such as 454 pyosequencing, eliminate the need for 16S rRNA primers and cloning, and generate reads of up to half a million sequences from a single sample. Such have been the technological advances in probing for microbial diversity within the oral ecosystem. But it is important to stress that bacteria are still currently named on the basis of phenotype. Consequently, there are thousands of 16s rRNA sequences within the databanks belonging to unnamed microorganisms. This is a problem and, together with the millions of nucleic acid and protein sequences now binned into GenBank, we have a situation in which data collection has, for the time being, outpaced cultivable and analytical capabilities.

The oral microbiome project has shown that hundreds more microbes exist within oral cavity niches than have ever been identified by conventional laboratory cultivation techniques. However, notwithstanding the fact that not all bacteria present in a sample are cultivable, the microorganisms isolated are alive. Molecular identification, on the other hand, gains information on the genomes present, from microorganisms dead or alive. Because dental plaque develops temporally, and communities show both compositional ripples and ecological catastrophes, single sample molecular assessments provide just a metagenomic snapshot. An important future development in this area will be to perform longitudinal studies of microbial community development at specific oral cavity sites. In such studies, the appearance of new DNA species over time should represent the acquisition of those microorganisms into the communities.

Oral cavity microbiome

The king is gone but he's not forgotten

The more recent explosion in technical sophistication and high-speed sequencing should not belittle the work of dedicated microbiologists over 50 years in characterizing new species of cultivable oral bacteria, and correlating disease conditions with individual or groups of microorganisms. Although the number of oral microbial species is estimated to be approximately 700 (there are 624 taxa in HOMD) (Dewhirst et al., 2010), there remains debate as to what constitutes a normal oral species. If a cut-off is applied, such that the likelihood that a species will be found in 99.5% of genomes sequenced, then the potential species count rises to about 1200. Because the oral cavity is the portal through which most microorganisms enter the body, it is also possible to identify genomes associated with food, e.g. Rhizobium, drink, e.g. Saccharomyces carlsbergensis, and air, e.g. Legionella.

Back to the present, the main questions now raised by the oral microbiome data relate to fathoming microbial functions in the contexts of health or disease conditions (Xie et al., 2010). In a recent study of oral bacterial communities in healthy subjects it was found that each individual's mouth houses a unique collection of bacterial species. However, about 15 bacterial genera were conserved among the individuals (Bik et al., 2010) showing that communities tend to be more similar when classified at the level of genus, and raising the possibility of defining a core microbiome (Zaura et al., 2009). On top of this, every site in a mouth has a non-random subset of bacteria, the make-up of which varies according to whether the site is basically healthy or diseased (Jenkinson and Lamont, 2005). This is very apparent from studies comparing the microbiota of teeth from children with or without dental caries (Gross et al., 2010; Kanasi et al., 2010); subgingival microbial communities of adults presenting with periodontitis compared with healthy periodontal status (Colombo et al., 2009); the oral microbiotas of HIV-infected subjects (Aas et al., 2007); and the root canal microbiotas from subjects with different types of endodontic infections (Siqueira and Rôças, 2009). Typically, microbiotas found at diseased sites are more structurally complex than those at corresponding healthy sites, suggesting that it might be easier to define microbial community activities at these latter sites.

Historically, the quest has been to identify cultivable microorganisms that were regularly associated with different oral diseases. Studies of caries-free versus caries-positive subjects, and data from in vivo experimental models, have suggested that mutans streptococci, e.g. S. mutans, Streptococcus sobrinus, were major causative agents of dental caries (Takahashi and Nyvad, 2011). Other studies have shown that three species of bacteria, so-called the red group, were consistently associated with adult periodontal disease (Socransky et al., 1998). However, molecular methodology has come up with some interesting discoveries that challenge dogma (Table 1). For example, the anaerobic Gram-positive bacillus Scardovia wiggsiae has been designated a new caries pathogen (Tanner et al., 2011). This organism was found in plaque from carious lesions in children in the presence or absence of S. mutans. Future studies on S. wiggsiae are anticipated to determine pathogenic potential in experimental in vivo models, characterize virulence factors and assess usefulness as a risk indicator for dental caries.

Table 1.  Changing concepts in oral disease aetiologies.
DiseaseTraditional microbial pathogen aetiologyNew concept pathogenic communities
  1. Additional references: Kazor and colleagues (2003); Faveri and colleagues (2008); Haffajee and colleagues (2008); Riggio and colleagues (2008).

Dental cariesStreptococcus mutans, S. sobrinus, S. downei, Lactobacillus acidophilus, L. casei, L. fermentum, L. rhamnosus, Actinomyces naeslundii, A. odontolyticusBifidobacterium dentium, S. mutans, Scardovia wiggsiae, B. longum, B. adolescentis, Prevotella spp., Selenomonas spp., Lactobacillus spp.
Root cariesActinomyces gerencseriae, A. israelii, Streptococcus spp., Lactobacillus spp., A. naeslundiiActinomyces spp., Lactobacillus spp., Prevotella denticola, Pseudoramibacter spp., Enterococcus faecalis
PulpitisE. faecalis, Porphyromonas endodontalis, A. odontolyticus, Parvimonas micra (Peptostreptococcus micros), Prevotella intermediaBifidobacterium spp., S. intermedius, Lactobacillus spp., A. israelii, Treponema denticola, P. micra, Prevotella baroniae, Dialister invisus, Olsenella uli, Prevotella spp.
GingivitisFusobacterium spp., Actinomyces spp., P. micra, P. gingivalisEubacterium nodatum, Eikenella corrodens, Fusobacterium nucleatum, P. micra
Aggressive periodontitisAggregatibacter actinomycetemcomitans, Capnocytophaga spp.Selenomonas spp., Prevotella spp., A. actinomycetemcomitans, Filofactor alocis, Tannerella forsythia, T. denticola, P. gingivalis
Chronic periodontitisP. gingivalis, T. forsythia, T. denticola, Fusobacterium spp.P. micra, Campylobacter gracilis, E. nodatum, Eubacterium saphenum, T. forsythia, P. gingivalis, Prevotella spp., Treponema spp., Selenomonas noxia E. corrodens
Halitosis (oral malodour)Prevotella spp., P. gingivalis, Actinomyces spp.Atopobium parvulum, Dialister phylotype, Eubacterium sulci, TM7 phylum, Solobacterium moorei, Prevotella spp.

Another potential cariogenic pathogen recently identified is Bifidobacterium dentium (Ventura et al., 2009). This is closely related to gut commensal bifidobacteria, but has acquired genes for survival in dental plaque at low pH, and does not colonize the edentulous mouth (Mantzourani et al., 2010). These various discoveries widen viewpoint of the causative agents of dental caries past the mutans streptococci. In a similar way, molecular methods have identified Dialister invisus, Olsenella uli and Synergistes spp. as prevalent in persistent root canal infections (Siqueira and Rôças, 2005), questioning a previously held belief that Enterococcus faecalis often played a significant role in the disease process (Siqueira and Rôças, 2009). Another potential paradigm shift results from the revelation that periodontitis-associated microbial communities are hugely more complex than previously believed, with possibly new pathogens, e.g. Eubacterium saphenum being implicated in periodontitis (Abiko et al., 2010) in addition to P. gingivalis and Tannerella forsythia (Table 1). Of note is that genomic analyses have revealed the presence of multiple phylotypes of Treponema within periodontal pockets (Moter et al., 2006) and it is suggested that members of this genus are almost always found within oral soft tissue lesions (Table 1).

Recent molecular genomic studies have thus confirmed some of the previous correlations made between the presence of bacterial species and specific disease conditions, but have identified some new and potentially significant associations. However, the discussion up until now has exclusively involved bacteria. The oral cavity is also home to more than 100 species of fungi (Ghannoum et al., 2010), and untold numbers of protozoan species. While fungi seem often to be omitted from microbiomics, yeasts such as Candida albicans clearly contribute to a wide range of oral infections, e.g. denture stomatitis, periodontitis, and interact specifically with some of the oral bacteria (Bamford et al., 2009). Protozoa, as well as undiscovered bacteriophage, would be continuously depleting the bacterial communities, while the bacteria themselves indulge in molecular war games between each other. Now that components of oral microbial communities have been identified, the objectives are to better understand the architectural, metabolic and genetic activities of these communities. For this, the new molecular taxonomic information has to be integrated with more conventional genetic, biochemical and physiological approaches, and it will then become evident how these communities function. The physical and chemical interactions that occur between the microorganisms determine how the community grows and survives, and influences the niche. The interactions can be studied through the employment of metabolomics, which characterize the metabolic product profiles of the community (Takahashi et al., 2010), proteomics (Rudney et al., 2010) and transcriptomics. Ultimately, it will be possible to incorporate all of this information into generating microbial community interactomes, essentially describing the functional consequences of the microbiome.

The meaning of life

The oral bacterial microbiome data have revealed numerous microorganisms that were never known to be present, and so laboratory cultivation and characterization of individual microorganisms is more urgent than ever before. Thousands of new genes have suddenly become available for investigation. Some of these could encode exciting new potential: for example, pathways for production of novel antibiotics, undiscovered toxins or new antibiotic resistance determinants. However, even for bacteria whose genomic sequences have been available for several years, there is still the problem that up to 50% of the assigned open reading frames encode polypeptides of unknown functions. Knockouts of these open reading frames may not have readily identifiable phenotypes, or alternatively can be lethal. Where phenotypes are difficult to pinpoint, transcriptional microarrays or metabolite profiles could become useful, under strictly controlled growth conditions, for comparison of mutant and wild-type activities. Phenotypes may also not become apparent until the environmental growth conditions are changed. For example, some genes are only switched on at specific pH, or only expressed when cells undergo surface growth in biofilms.

Identifying those genes that are crucial for generation and stability of oral microbial community interactomes is an important future objective. This will require development of new technologies to analyse, predict and model genetic and genomic interactions between microorganisms within communities, in vitro and in vivo. It also depends upon success with expanding the microbial living cultivar, and applying genetic and phenotypic analyses to newly cultivated bacteria. One objective for diagnostic and clinical impact would be to predict that a particular microbial community might evolve to become pathogenic. Another might be that community microbiomes or interactomes may be diagnostic in identifying an unknown underlying host condition, such that the subject may then be treated accordingly.

Community development

There's more to the picture than meets the eye

There are several general observations that can be made, now that the oral microbiome has been revealed, and many of the oral bacterial genomes sequenced. Possibly no-one seriously believed that there would be such immense biodiversity, and the extent of horizontal gene transfer and gene duplications seems to be enormous. Horizontal gene transfer of virulence determinants is associated with phage transduction, and some genes, e.g. merA (mercury resistance) have transferred between Gram-positive and Gram-negative bacteria (Roberts and Mullany, 2006). Duplicated sequence scaffolds have also generated families of proteins, such as fibronectin-binding proteins in Treponema denticola (Bamford et al., 2010) and cell wall-associated proteins in streptococci (Nobbs et al., 2009). However, a major concept that has emerged is that oral microorganisms do not live in isolation; they live in communities within which there must be multiple growth dependencies, synergies and antagonisms. In addition, there may be obligate dependencies between microorganisms for growth and survival. As has been pointed out before, one of the possible reasons why individual bacteria might be non-cultivable is because they depend upon another organism or group of organisms to grow. If it is possible to determine the chemical bases of such metabolic dependences, it would be gradually more possible to cultivate the currently uncultivable. Even then, some of the essential growth conditions might involve host factors that at present cannot be satisfactorily reproduced in the laboratory setting.

Dental plaque comprises a mixture of microorganisms, microbial polymers, salivary glycoproteins and other host-derived molecules. Plaque formation begins with the adherence of mainly Gram-positive bacteria to the salivary pellicle, which is a thin layer of salivary constituents that becomes rapidly adsorbed onto a clean enamel surface. Subsequent growth and multiplication leads to the formation of an early biofilm, sometimes referred to as a linking film. This provides a new surface for other microorganisms to colonize, with intermicrobial adherence (coaggregation) generating diversity within the developing biofilm community. At the gingival margins, where tooth meets gum, the communities consist of Gram-positive and Gram-negative bacteria. However, at subgingival sites the communities often consist of mainly Gram-negative obligately anaerobic bacteria. Among these are species that have been historically associated with periodontal disease conditions. While there is temporal maturation of the oral microbiota in situ, there is also increasing prevalence of periodontal pathogens with age in children (Papaioannou et al., 2009). It is thought that, ultimately, age-associated defects in innate immune functions contribute to infection-driven periodontal disease in the elderly (Hajishengallis, 2010).

Primary colonizers

For many years it has been acknowledged that the first bacteria to colonize a clean tooth surface in the human mouth are mainly streptococci (Nyvad and Kilian, 1990). However, the prediction would now be that a much wider range of organisms are able to first colonize the enamel pellicle, based upon recent data from in vivo studies. Fluorescent molecular probe and deep sequencing technologies have yet to be applied to investigate the microbiomes associated with temporal deposition onto enamel. The microbial communities associated with plaque formed on tooth surfaces from clinically healthy subjects (Aas et al., 2005; Diaz et al., 2006) contained Gemella (Streptococcus-like), Granulicatella, Abiotrophia, Actinomyces and Veillonella (Fig. 1), as well as Selenomonas and Capnocytophaga. Interestingly, Veillonella parvula is the first ranked taxon in the oral microbiome making up 6.6% of total clones observed (Dewhirst et al., 2010), but many clones have resisted cultivation. The genus Streptococcus is a very successful colonizer; not only are more than 25 species of viridans streptococci now defined, the pathogenic streptococci such as Streptococcus pyogenes, Streptococcus agalactiae and Streptococcus dysgalactiae are all able to colonize the human nasopharynx (Nobbs et al., 2009).

Figure 1.

Primary (early) colonizing bacteria of salivary pellicle-coated oral cavity surfaces, e.g. tooth enamel. Cultivable streptococci that are initial colonies of salivary pellicle include S. oralis, S. mitis, S. sanguinis, S. gordonii, S. parasanguinis and S. cristatus (showing characteristic bundles of cell-surface fibrils). The first three species tend to be of highest incidence. Veillonella parvula (also var. atypica), which utilize lactate, are almost always found in association with streptococci. Abiotrophia spp., e.g. defectiva, Gemella spp., e.g. haemolysans, Rothia dentocariosa, Atopobium spp., e.g. parvulum, and Granulicatella spp., e.g. adaicens, were known as nutritionally variant streptococci and many species are uncultivable. Gemella are aerobic diplococci with surface fibrillar coating, while Atopobium and Granulicatella are closely related strict anaerobes. Rothia (Gram-positive oval-shaped cells) and Neisseria (Gram-negative cocci or short rods) are both prominent components in early (∼8 h) communities. The cells that are touching each other have been shown to physically interact (coaggregate). Transition to a disease community, leading to enamel decay (caries), occurs with acquisition of Scardovia and Bifidobacterium species, lactobacilli, S. mutans and Capnocytophaga spp., with corresponding decreases in the proportions of most of the primary colonizing streptococci. Data collected principally from Aas and colleagues (2005) and Diaz and colleagues (2006).

The streptococci possess adherence and metabolic capacities that enable them to colonize a wide repertoire of oral cavity sites, such as epithelium, salivary pellicle-coated tooth surfaces, denture and implant surfaces, tonsils and tongue (Nobbs et al., 2009). Different species of streptococci have predilections for different surfaces, for example Streptococcus salivarius is almost uniquely associated with colonization of the tongue (Aas et al., 2005). It is not really clear how such site preferences are determined, but theories include recognition by bacterial cells of tissue-specific receptors and ability to outcompete other organisms through metabolic nuances or production of bacteriocin killing factors. Some of the interactions that occur between early colonizing bacteria have been mapped out, and are described below and shown in Fig. 1. The molecular basis of the associations between the newly cultivated and previously cultivated bacteria has not yet been worked out.

Primary interactome

The streptococci utilize an array of surface proteins, many of the genes for which are horizontally dispersed across the species, in order to adhere to glycoproteins present within the oral environment (Nobbs et al., 2009). Actinomyces naeslundii (oris) cells are reported to utilize type 1 fimbriae to adhere to statherin and proline-rich proteins present within salivary pellicle formed on enamel surfaces (Li et al., 2001). Actinomyces are also able to interact via their type II fimbriae with the linear phosphopolysaccharides expressed on the surfaces of Streptococcus oralis, Streptococcus mitis and Streptococcus sanguinis cells containing GalNAcβ1-3Gal (or Galβ1-3GalNAc) linkages (Yang et al., 2009). The physical interactions between Actionmyces and Streptococcus are probably associated in many instances with metabolic interactions that favour both partners (Palmer et al., 2001), but which are currently undefined. It is well documented that Veillonella species interact metabolically with lactic acid-producing bacteria such as Streptococcus spp. The veillonellae utilize lactate generated by the streptococci as a carbon and energy source (Bradshaw and Marsh, 1998), potentially promoting streptococcal growth through lifting of end-product inhibition of glycolysis. Veillonella species are almost always found along with streptococci in molecular studies of oral microbial populations.

Secondary interactome

The development of secondary interactomes is probably less clearly understood overall because of the larger variety of species involved. The maturing communities become constructed upon the physical associations of secondary microorganisms with antecedent (primary) bacteria. The secondary communities differ more widely, depending upon the site within the oral cavity at which they develop, and the host conditions present. For example, a rather defined secondary colonization event occurs when cells of the yeast C. albicans bind to streptococci and other early colonizers, e.g. Actinomyces that have deposited upon denture acrylic, to produce biofilms that potentially exacerbate denture stomatitis (Bamford et al., 2009). An intermediary secondary interactome event, involving physical interactions, chemical signalling and transcriptional responses, occurs between Streptococcus species and P. gingivalis. The P. gingivalis major fimbriae interact with glyceraldehyde-3-phosphate dehydrogenase present on the streptococcal cell surface (Nagata et al., 2009), and the minor fimbriae interact with adhesin protein SspB (Daep et al., 2011), while streptococci express autoinducer-2 to facilitate dual species community development (McNab et al., 2003). Because there is no competition for sugars between these two bacterial species (P. gingivalis are asaccharolytic), they coexist in harmony. The streptococci presumably deplete the environmental oxidants for P. gingivalis, while extensive proteolytic activities of the latter probably provide streptococci with peptides.

In the development of caries (tooth decay) it is envisaged that, as a result of changing to a more carbohydrate (sugar)-based diet, more frequent feeding, and consumption of acidic drinks, conditions become established that favour a more acidogenic microbiota (Bradshaw and Marsh, 1998). In severe early childhood caries, Smutans, Swiggsiae and Fusobacterium nucleatum showed the highest associations (Tanner et al., 2011). The occurrence of S. sobrinus was infrequent. Development of severe caries seems to be associated with a decline in proportions of S. mitis, Streptococcus gordonii, Streptococcus cristatus and Capnocytophaga (Gross et al., 2010). In addition, bifidobacteria may play a role in progression of caries in children and in adults (Mantzourani et al., 2009). However, Veillonella and Actinomyces species are almost always present in higher proportions in carious subjects, showing that there is species overlap between disease-associated and non-disease-associated communities.

Many of the concepts of oral microbial community interactions have been developed by Kolenbrander and co-workers (Kolenbrander et al., 2010). These concepts have, in the main, been depicted in a visually memorable manner, and have emphasized the multiple microbial cell–cell interactions (coaggregations) that have been shown to occur in laboratory studies. However, it is important to realize that not all coaggregating bacteria are found at one site, and that out of the 700 or more oral bacterial taxa, only 10% or less of these are actually found at a single site. The secondary interactomes may be very different, depending upon whether the communities are above or below the gum line, superficial or invasive, acute or chronic, periodontic or endodontic, and if the host is immune-proficient or immunocompromised, male or female, child or adult; the various permutations and comparisons are immense. Given that also many of the secondary colonizers are presently uncultivable, diagrammatic depictions of interactions occurring between bacteria in a typical secondary community are not yet possible. However, systems level analysis of communities by molecular imaging techniques (Valm et al., 2011) suggests that we will soon be able to map microbial communities with some precision. Once techniques such as these can be utilized in longitudinal studies of population development at different oral cavity sites, it will be possible to define temporal and positional knowledge of microbial community patterns.

Dependence and defence

Synergistic interactions

The various oral microbial communities become established through making best use of the nutrients available and surviving a wide range of environmental challenges. The utilization by one organism of the metabolic waste products of another, such as described above by Veillonella and Streptococcus, represents one way in which metabolic traits may be dovetailed for co-survival. In a similar fashion, metabolic cross-feeding of Aggregatibacter actinomycetemcomitans by streptococci producing L-lactate appears to ensure persistence of A. actinomycetemcomitans in a polymicrobial infection model, and was shown to enhance pathogenesis (Ramsey et al., 2011). Other synergistic interactions occurring between primary and secondary colonizers in biofilm formation are increasingly being identified (Periasamy and Kolenbrander, 2009), but the metabolic mechanisms operating are generally not yet known. The advantages that community living brings, aside from simple cross-feeding by partner organisms, is that metabolic interactions can be intercalated such that organisms may collectively utilize a common substrate. For example, glycosylated proteins such as mucins provide a major source of sugars and amino acids for microorganisms in the oral cavity and nasopharynx (King, 2010). Groups of organisms may more efficiently utilize a complex polysaccharide through combined activities of specific hydrolases produced by the different species. A second advantage for a microbial community biofilm is that the interior cells are protected from attack by protozoa and macrophages, and are in many instances shown to be less susceptible to antibiotics (Harriott and Noverr, 2010).

Antagonistic interactions

It is important to recognize that there is continuous indigenous control of communities through competitive or inhibitory interactions occurring between microorganisms. Secreted molecules such as hydrogen peroxide, lactic acid and bacteriocins deter neighbouring bacteria, and S. mutans produces all of these compounds (Kreth et al., 2008). Six lantibiotic-type mutacins (bacteriocins post-translationally modified to contain lanthionine) have been identified, with mutacin I and mutacin II being similar to nisin (Lactococcus lactis) and salivaricin (S. salivarius) respectively (see Hossain and Biswas, 2011). At least two non-lantibiotic mutacins are also produced by Smutans (mutacins IV and V) and these have wide-ranging activities against almost all species of streptococci (Hossain and Biswas, 2011). The production of mutacins might therefore assist Smutans, which is not considered to be a primary colonizer, to take control of a developing cariogenic community as conditions become more favourable. Presumably the bacteria often found in association with S. mutans, e.g. bifidobacteria, veillonellae (see Fig. 3) are less susceptible to mutacins. Interestingly, control of mutacin production is closely coordinated with the development of competence for DNA uptake (Okinaga et al., 2010) and is influenced by expression of luxS gene producing AI-2 (Merritt et al., 2005).

Figure 3.

Oral microbial interactome based upon in vitro studies of coaggregation (grey connecting lines) with metabolic cooperation (green lines), or metabolic inhibition (red dashed lines). Key to lettering: An, Actinomyces naeslundii (oris); Aa, Aggregatibacter actinomycetemcomitans; Ba, Bifidobacterium adolescentis; Ca, Candida albicans; Co, Capnocytophaga ochracea; Fn, Fusobacterium nucleatum; Np, Neisseria pharyngis; Pn, Prevotella nigrescens; Pg, Porphyromonas gingivalis; Sf, Selenomonas flueggei; Sg, Streptococcus gordonii; Sm, Streptococcus mutans; So, Streptococcus oralis; Td, Treponema denticola; Tf, Tannerella forsythia; Vp, Veillonella parvula (atypica).

Streptococcus salivarius, a main colonizer of the tongue dorsam, produces at least six lantibiotic peptides (salivaricins: SalA, SalA2, SalB, Sal9, G32, streptin) that kill a range of streptococci (Wescombe et al., 2011). On agar plates, SalA lantibiotic secreted by S. salivarius K12 strongly inhibits growth of most S. pyogenes strains (Upton et al., 2001). Nevertheless, S. salivarius K12 and S. pyogenes MGAS6180 seem to be able to coexist within biofilms. S. salivarius and S. pyogenes alone both form confluent biofilms on saliva-coated surfaces. In dual species biofilms, S. pyogenes cells make up a significant proportion of the biovolume and tend to form societies within a denser mass of S. salivarius cells (Fig. 2). S. pyogenes cells within biofilms must be somehow protected from the killing effects of S. salivarius bacteriocins, perhaps by generation of less-sensitive persister cell groups in response to sensing SalA peptide (Upton et al., 2001). These results demonstrate that inhibitory activities of bacteriocins as defined in vitro may not necessarily be manifested within biofilms in vivo.

Figure 2.

Confocal laser scanning microscopy images showing coexistence of S. pyogenes (Group A Streptococcus) with bacteriocin-producing Streptococcus salivarius K12 in 24 h biofilms formed upon salivary pellicle. A. S. pyogenes MGAS6180 alone forms confluent and compact biofilms of depth ∼15 µm. B. In dual species biofilms with S. salivarius, the density of S. pyogenes cells (green) is much reduced, with pillar-like structures formed of depth up to ∼50 µm. C. S. salivarius K12 forms a confluent biofilm in dual species, but this is less compact (YZ stack) than the S. pyogenes monospecies biofilm. D. Merged images of (B) and (C) showing integration of S. pyogenes and S. salivarius (YZ stack) with formation of societies of S. pyogenes cells that access the surface of the mature biofilm (XY stack) to ∼50 µm. Green (FITC), S. pyogenes; red (propidium iodide), S. salivarius. The plot shows relative biovolumes of S. pyogenes and S. salivarius in dual species biofilms (D), indicating that S. pyogenes makes up about 40% of the biovolume. Images and biofilm data provided by Chris Wright.

Bacteriocin production is not restricted to the oral Gram-positive bacteria. Prevotella nigrescens produces a bactericidal protein active against P. gingivalis, T. forsythia and Actinomyces species (Kaewsrichan et al., 2004) suggesting that subgingival biofilm communities are also subject to intrinsic interspecies growth regulation. Bacteriocins are also reported to be produced by F. nucleatum (Ribeiro-Ribas et al., 2009) and by Eikenella corrodens (Apolônio et al., 2008), while genes encoding potential bacteriocins have been identified in T. denticola (Bamford et al., 2010). It is thought that by harnessing the structural features determining the functions of some of these molecules it might be possible to generate new antimicrobial agents active against periodontal pathogens.

Future perspectives

Rust never sleeps

With the new microbiome data and more recent laboratory studies showing specific interactions between oral microorganisms, we can begin to further develop spatial and functional maps of oral microbial communities. Although there are no data yet for multiple individual communities, it is possible to generate simple matrices showing how oral microorganisms interact, not just physically but also metabolically. Some of the known interactions occurring between oral microorganisms have been integrated into a simplified interactome diagram (Fig. 3). For some of these interactions there are no functional explanations, aside from retention within the community (Nagaoka et al., 2008), but it is probable that the driving forces for coaggregation are of benefit to survival. Interactions that are known to be metabolically beneficial to the partners are also shown in Fig. 3. Where metabolic interactions between bacteria are known to be antagonistic, the bacteria usually do not have the ability to coaggregate (Fig. 3). These observations provide some assuagement for the notion that oral bacteria are able to coaggregate as this function has been selected for being beneficial to the development of the community.

The final frontier for the oral microbial ecologist, the oral microbiome, has been realized and is a feat of technology, managed by talented evolutionary microbiologists and taxonomists. Metagenomic sampling of individual sites within the oral cavity shows that there are probably hundreds of different micro-communities (Bik et al., 2010), each of them defining a niche. New research approaches will then be needed to reconstruct these microbial communities in vitro, with models that might incorporate differentiating epithelia, fibroblasts, salivary or plasma molecules and innate immune components. Reductionist approaches will continue to be necessary in order to define mechanisms underlying microbial community maturation. With genetic techniques now being developed or refined for Prevotella, Fusobacterium, Treponema, etc. it will be possible to evaluate the roles of specific bacterial genes in a wide range of polymicrobial interactions. Furthermore, as culturing requirements for individual microorganisms have become defined, bacteria such as S. wiggsiae, Filofactor alocis and D. invisus will be characterized for production of colonization and virulence factors. Novel and sophisticated techniques will need to be developed in order to analyse the composition and properties of many different types of community. This will advance microbiome data past cataloguing phyla, and on to quantifying individual species and utilizing co-isolation data to define metabolic synergies or dependencies. In addition, the genetic and metabolic profiles of oral microbial populations may be utilized as surrogate markers for potential health concerns, such as nutritional deficiency or immune dysfunction, or increased risk of coagulopathy, diabetes, or birth complications. All of these studies will be bioinformatically intensive, and will depend upon the expertise of molecular microbiologists, protein biochemists, microbial physiologists and geneticists to gain biological sensitivity.

The notion that microbial communities, as opposed to individual microorganisms, are disease determinants has significant impact upon clinical practice and disease control. The use of broad spectrum antibiotics that, with the intention of removing a pathogen, also deplete the host microbiota, is recognized now as a medical risk. In a simple scenario, if a pathogenic microorganism is dependent upon other bacteria in order to survive, removal of partner bacteria could provide a means to eradicate the pathogen. These kinds of approaches can only be properly planned if the interactions between the microorganisms that go to make up communities are fully understood. The ability to more specifically manipulate the composition of microbial communities is a challenge for the future.


The author thanks Rich Lamont, Steve Kerrigan, Aras Kadioglu and Paul Kolenbrander for sharing with him their thoughts, ideas and reagents. Also major thanks to members of the Oral Microbiology laboratory, past and present, for their dedication to the cause of microbiology research. Special thanks to Lindsay Dutton and Chris Wright for the provision of colour figures, and to Ang Nobbs for helping with the manuscript. I should apologize to some authors because, owing to space limitation, it has not been possible to cite all relevant primary publications. Research of the Author is supported by NIH (NIDCR) (R01DE016690-06) and the Wellcome Trust (#081855 and #084979).