Principles of microbial classification

As stated above, the purpose of classification schemes is to develop a logical arrangement of organisms based on their similarities and relationships. This requires the determination and comparison of as many characteristics as possible, although in identification schemes, only a few key discriminatory tests may be needed to distinguish between certain organisms. Early classification schemes relied heavily on morphological and simple physiological criteria such as the shape of the cell, and the pattern of fermentation of simple sugars (Table 3.2). In effect, these approaches analysed only a fraction of the components encoded by the genetic material of the cell (the genome). Chemotaxonomy, in which there is a broader analysis of more complex components of the cell (for example, the chemical composition of the cell wall or of membrane lipids, whole cell protein profiles, etc) led to major improvements in classification schemes. The structure of cells can also be compared using immunological techniques (serology), in which specific antibodies (polyclonal or monoclonal) are used to detect cell surface antigens.

Table 3.2 Some characteristics used in microbial classification and identification schemes

Contemporary classification schemes are based more on determining the genetic relatedness among strains. As the properties of an organism are dictated by what is encoded by its genome, the ultimate comparison is to determine the similarity in DNA base composition. The mole percentage of guanine (G) plus cytosine (C) in the total DNA can be determined. Organisms with markedly different G+C contents are unrelated, while organisms that have similar G+C values are closely related, although similarity in gross DNA composition is not unequivocal proof of close relatedness because the base pairs could be organised in a different sequence. In such a situation, genotypic similarity can be confirmed by determining the degree of relatedness (homology) between the DNA from two strains i.e. the abilities of heat-denatured, single strands of DNA from different strains to re-anneal with each other, or to a reference strain, during slow cooling (DNA–DNA hybridization). A high level of homology reflects an overall similarity in the nucleotide sequences from the DNA of the two strains being compared, and hence confirms the close taxonomic relationship of the strains.

Classification is now dominated by comparisons of the sequence of 16S ribosomal RNA genes (16S rRNA). Within the rRNA gene, some stretches of DNA sequence are conserved while other areas are highly variable and reflect evolutionary divergence. These genes are around 1500 nucleotides long, which is short enough for rapid sequencing in conventional laboratories (using automated DNA sequencing equipment) but long enough to provide valuable discriminatory information to demonstrate similarities and differences among strains. The conserved regions can be used as a template for the design of ‘universal’ PCR (polymerase chain reaction) oligonucleotide primers that can be used to amplify the rest of the 16S rRNA gene, which can then be sequenced to identify differences in the variable regions. The sequences can be compared with those from other microorganisms, and to sequences in nucleotide databases, so that the relationship of an isolate to known species can be determined and evolutionary (phylogenetic) trees can be developed. The technique is relatively rapid, and has facilitated the analysis of a far wider range of bacteria than was previously possible. Comparison of 16S rRNA gene sequences have revolutionized the field of microbial taxonomy, and has clarified the classification of many previously heterogeneous groups of oral bacteria, such as the streptococci (Table 3.3) and anaerobic Gram positive rods formerly grouped as Eubacterium species (Fig. 3.5). In addition to classifying unknown strains, this approach can also be used to identify isolates, and has many advantages over conventional cultural approaches (see later; Fig. 3.2).

Table 3.3 Species of oral streptococci isolated from humans

* mutans streptococci also include S. ferus (isolated from rats), S. macacae and S. downei (serotype h) (isolated from monkeys).

Fig. 3.2 The main stages in determining the microbial composition of the microflora of samples from the mouth using either culture or molecular approaches.

A consequence of classification is the proposal of new species. A species represents a collection of strains that share many features in common, and which differ considerably from other strains. Once a species has been recognised, then a type strain is nominated that has properties representative of the species. Type strains are held in national collections, such as the American Type Culture Collection, ATCC, or the National Collection of Type Cultures, NCTC, which is located in the United Kingdom.

A species may be divided into subspecies if minor but consistent phenotypic variations can be recognised. Likewise, groups of strains within a species can sometimes be distinguished by a special characteristic. For example, strains with a special biochemical or physiological property are termed biovars or biotypes, while strains with a distinctive antigenic composition are described as serovars or serotypes, and can be recognised using appropriate antibodies. Molecular approaches can also be adapted for subtyping strains within a species. Whole genomic DNA can be digested by different restriction enzymes (endonucleases), which cut the nucleic acid in specific places; these digests are then electrophoresed on an agarose gel to generate a chromosomal fingerprint. Different strains generate different patterns (restriction fragment length polymorphisms, RFLPs), although strains that appear to give similar patterns need to be compared after digestion with more than one enzyme. This approach can yield patterns too complex to analyse. In order to overcome this problem, the restriction fragments can be blotted onto nitrocellulose or nylon membranes, and hybridized with a suitably labelled probe to give a simpler profile. A ‘universal probe’ can be used, based on part of the rRNA gene sequence, and this can yield different patterns within a species (ribotyping). Ribotyping has allowed insights to be made into the transmission of strains among family members, especially from mothers to their babies, and in demonstrating the turnover of specific ribotypes (and specific clones) at a site (Ch. 4).

Principles of conventional microbial identification

Once organisms have been correctly classified using rigorous techniques, then more simple identification schemes can be devised in which only limited numbers of key discriminatory properties are compared (Fig. 3.1, Table 3.2). The first stage might involve the reaction of an organism with the Gram stain, and the determination of cellular morphology. Bacteria are then described as being, for example, Gram positive cocci or Gram negative rods, etc. Depending on the outcome of that division, simple physiological tests may be performed, such as the determination of sugar fermentation patterns, the profiles of acidic fermentation products following glucose metabolism, or selected enzyme activities. The rapid detection (ca. 4 hours) of constitutively-expressed enzymes by concentrated suspensions of bacteria has simplified the identification of some groups of bacteria. Ecologically-relevant substrates that detect enzymes such as glycosidases that cleave sugar residues from salivary mucins are now more commonly used to differentiate groups of bacteria that had previously been difficult to separate, for example, some oral streptococci. Some of these tests have been incorporated in kits and sold commercially, together with computerized databases, to facilitate identification.

Monoclonal antibodies and nucleic acid (oligonucleotide) probes have been developed for the rapid identification of some species, but primarily those associated with disease. Such antibodies and probes can be labelled with a signalling group to aid in detection; examples include fluorescent dyes (Fig. 3.3), radiolabels, or enzymes such as horseradish peroxidase. These techniques have the advantage that organisms can be detected directly in plaque or a clinical sample without the need for lengthy culturing, although a potential drawback is that they can detect dead as well as viable cells.

Fig. 3.3 Examples of currently unculturable bacteria in sub-gingival plaque visualized using fluorescent in situ hybridization (FISH) techniques. (A) 16S rRNA oligonucleotide probes were used to detect bacteria (green) and Treponema spp (red).

Courtesy of Dr Annette Moter and produced with permission by Springer-Verlag GmbH & Co. (Norris SJ, Paster BJ, Moter A and Gobel UB, The genus Treponema, in The Prokaryotes, Third Edition; Springer, Berlin, 2007). (B) 16S rRNA oligonucleotide probes were used to detect bacteria (green) and members of the TM7 phylum (blue). Courtesy of Dr Cleber Ouverney and the American Society for Microbiology (see Ouverney et al Appl Environ Microbiol 2003 69:6294–6298).

Conventional microbial identification schemes can only be used when organisms have been isolated and grown in pure culture. Inevitably, the procedure for achieving pure culture (sample dispersion and dilution, growth on selective and non-selective agar plates, incubation conditions, etc) leads to the introduction of bias towards those microorganisms that grow quickly and easily under laboratory conditions (Fig. 3.2). Alternative, culture-independent, procedures have evolved out of modern taxonomic approaches, and these are giving a more accurate picture of the diversity (richness) of the microflora from a wide range of habitats.

The impact of molecular microbial ecology

As stated above, the genetic relatedness of microorganisms is now primarily determined by comparisons of 16S ribosomal RNA (rRNA) gene sequences. The biggest impact of this approach has been in the analysis of diverse communities of microorganisms from a number of habitats (molecular microbial ecology), including the mouth. Comparisons of the number of cells in samples that can be observed by microscopy versus those that can be cultured in the laboratory, even when using the most advanced techniques, has shown that only a proportion of the microflora at a site can be grown. The culturable fraction can vary from less than 1% of the total cell count in some marine habitats to about 50% of the oral microflora. These organisms that cannot currently be grown are termed ‘unculturable’; the reasons for this may be ignorance of an essential nutrient or other growth requirement, or because they evolved to grow as part of a community of cells, rather than as an isolated pure culture. Molecular ecological approaches have enabled the construction of phylogenetic trees that include currently unculturable organisms, so that in many cases, a genus and species name can still be assigned. Oligonucleotide probes can also be synthesized so that the presence of these organisms can be determined relatively simply in clinical samples using rapid tests, such as the polymerase chain reaction (PCR) or by in situ hybridization, usually with a fluorescent label (Fig. 3.3). An important benefit of these PCR-based molecular approaches is their potential to detect organisms that are present in low numbers. Although the properties of these uncultivable organisms cannot be determined using conventional tests (such as sugar fermentation patterns or their antibiotic sensitivity profile), databases of genes exist that can be interrogated to search for homologous sequences with known functions. This can provide insights into potentially important properties of these unculturable organisms, such as their cell wall structure, their virulence attributes, and the metabolic pathways they might use.

The existence of these unculturable bacteria cannot be ignored, since their presence at a site could be highly significant; for example, the aetiological agent of syphilis is a spirochaete, Treponema pallidum, that still cannot be grown in the laboratory. Two large families of novel unculturable bacteria have been identified in the mouth (Fig. 3.3), and these are commonly detected in deep periodontal pockets; further details will be presented later in this Chapter. In addition, some genera contain examples of both culturable and unculturable species; for example, there are around 50 Treponema species that can be seen by microscopy and detected by molecular approaches but which cannot be grown (Fig. 3.3A). Similarly, the molecular analysis of the microflora associated with dento-alveolar abscesses and endodontic infections has consistently identified novel groups of bacteria that were not recognised, or were grossly under-estimated, in parallel cultural studies (Fig. 3.2).

Molecular approaches have also been developed to compare the diversity of oral microbial communities from different sites in health and disease. These approaches include microbial community profiling using denaturing gradient gel electrophoresis (DGGE). Total genomic DNA is extracted from clinical samples, amplified by PCR using universal primers for bacterial 16S rRNA genes, and products resolved on polyacrylamide gels with a denaturing gradient. DGGE profiles can be analysed using appropriate software, and novel or discriminatory bands can be excised from the gel, cloned, and sequenced, enabling a presumptive identification to be made. Alternatively, techniques based on checkerboard DNA–DNA hybridization using labelled whole genomic probes and nylon membranes can be used to simultaneously screen multiple clinical samples for around 40 different pre-selected microbial species.

These culture-independent approaches are radically changing our perception of the diversity of the resident oral microflora in health and disease. Work is underway to identify the signature DNA sequences of all of the microorganisms in the mouth (culturable and unculturable). In the future, DNA microarray chips will be available, with the ability to detect each component or specific subsets of the resident oral microflora, enabling clinicians to rapidly screen for the presence or absence of over 700 organisms, thereby leading to the promise of improved diagnostic and treatment opportunities.

Difficulties arising from recent advances in microbial classification

Although recent advances have led to improvements in the classification of oral bacteria, such improvements have also generated a number of difficulties when interpreting or comparing early data when a previous (and sometimes flawed) nomenclature was in use. The classification of many groups of oral bacteria has been revolutionized in a relatively short time period, with many new genera and species described. A species highlighted in an early study may now have been reclassified and hence renamed, and so new and old terminologies may coexist in the scientific literature. For example, Streptococcus sanguis has been reported in the literature for many decades but, as of 1989, its description became more limited and organisms that were previously included within this species are now known to be sufficiently different as to warrant a distinct species epithet, e.g. S. gordonii. Consequently, some strains reported in earlier studies as S. sanguis may not have the same properties as strains more recently identified as S. sanguis sensu stricto. Furthermore, the Latin name originally given to some of the oral streptococci has recently been shown to be erroneous, and so S. sanguis is now termed S. sanguinis. For similar reasons, S. parasanguis, S. rattus, S. cricetus and S. crista have been renamed as S. parasanguinis, S. ratti, S. criceti and S. cristatus, respectively. Thus, great care has to be taken when interpreting older (and not so old) scientific literature.

Microbial taxonomy is a dynamic area with existing species being reclassified due to the application of more stringent tests, together with the recognition of genuinely newly-discovered species from sites such as periodontal pockets and infected root canals. The emphasis paid to the classification and identification of the oral microflora is necessary because without valid subdivision and accurate identification of isolates, the specific association of species with particular diseases (microbial aetiology) cannot be determined. Likewise, it has to be accepted that further changes in microbial classification schemes will occur in the future and new genera and species will be identified. The properties of the main groups of microorganism found in the mouth will now be described.

Gram positive cocci