|Home | About | Journals | Submit | Contact Us | Français|
Real-time PCR analysis of the total bacterial load in advanced carious lesions has shown that the total load exceeds the number of cultivable bacteria. This suggests that an unresolved complexity exists in bacteria associated with advanced caries. In this report, the profile of the microflora of carious dentine was explored by using DNA extracted from 10 lesions selected on the basis of comparable total microbial load and on the relative abundance of Prevotella spp. Using universal primers for the 16S rRNA gene, PCR amplicons were cloned, and approximately 100 transformants were processed for each lesion. Phylogenetic analysis of 942 edited sequences demonstrated the presence of 75 species or phylotypes in the 10 carious lesions. Up to 31 taxa were represented in each sample. A diverse array of lactobacilli were found to comprise 50% of the species, with prevotellae also abundant, comprising 15% of the species. Other taxa present in a number of lesions or occurring with high abundance included Selenomonas spp., Dialister spp., Fusobacterium nucleatum, Eubacterium spp., members of the Lachnospiraceae family, Olsenella spp., Bifidobacterium spp., Propionibacterium sp., and Pseudoramibacter alactolyticus. The mechanisms by which such diverse patterns of bacteria extend carious lesions, including the aspect of infection of the vital dental pulp, remain unclear.
Dental caries is a progressive disease that expands from the initial focus of enamel degradation and the subsequent exposure of the underlying dentine to a point where microorganisms gain access to the tubular network of dentine that extends to the dental pulp (18). Depending on the response patterns of the odontoblasts that line the pulp chamber, bacterial invasion is sometimes arrested by a layer of so-called secondary or sclerotic dentine. Extension of the lesion can lead to necrosis of the pulp tissue by toxic action, immunopathological response, or frank invasion by bacteria. If left unchecked, the lesion can extend to involve the supporting bone and track to soft tissue, where cellulitis is a potential outcome (17).
Dissolution of the enamel matrix by organic acids, particularly lactic acid produced by mutans streptococci, is considered to be the primary event in caries development (8, 33). Recent analysis has, however, indicated the complexity of the bacterial biofilm associated with dental caries initiation in children (3). Diverse taxa, including a variety of streptococci, bifidobacteria, and organisms belonging to the genera Actinomyces, Selenomonas, Leptotrichia, Campylobacter, and Capnocytophaga were detected (3). In this context, there is a relative paucity of information concerning the groups of bacteria associated with the ensuing phase of lesion development, namely, invasion and destruction of the dentine matrix. Our previous analysis of the lesion of advanced caries by culture and real-time PCR (19) indicated the presence of many undetected or uncharacterized bacteria. It is evident that incomplete knowledge of the microbial populations that contribute to extension of the carious lesion enforces an empirical approach to therapy rather than specific antimicrobial therapy that might allow more conservative treatment options.
Since many organisms have yet to be isolated and cultured (26), PCR-based techniques are now widely used to identify organisms from an environment. These methods typically use broad-range or universal primers to amplify DNA that has been extracted directly from the environment. This is done based on the premise that DNA is efficiently extracted from most organisms present in the environment and that each species of DNA is amplified proportionately by PCR and cloned without bias (16). The sequence of choice for environmental surveys is the 16S rRNA gene due to its recognized species-specific conservation (35). In the present study, we used population analysis based on 16S rRNA gene(s) to gain insight into the diversity and abundance of the major groups of bacteria within the matrix of 10 advanced caries lesions selected on the basis of comparable total microbial load and high or low abundance of Prevotella spp. (19).
The source of material for analysis was carious dentine obtained in a manner approved by the Human Ethics Committee of Central Sydney Area Health Service (reference no. 6/96) (19, 23). Adult patients elected to have extractions for unrestored anterior, premolar, and molar teeth that presented with large coronal dentine caries lesions that, by macroscopic examination, had not penetrated to the underlying vital pulp tissue. Selected teeth had adjacent periodontal probing depths of <4 mm. After the removal of superficial plaque and debris overlying the lesion, the carious zone of decalcified and partially decalcified dentine was excavated, weighed, and resuspended in reduced transport fluid (31) to a concentration of 10 mg (wet weight) of dentine per ml prior to processing and extraction of bacterial DNA by using the ATL buffer reagent (Qiagen, Clifton Hill, Victoria, Australia) as previously described (19, 23). This method of DNA extraction releases DNA from gram-negative bacteria, as well as from many gram-positive bacteria, including lactobacilli when cultured under anaerobic conditions (4, 19).
Ten dentinal caries samples were chosen for comprehensive bacterial analyses from a pool of 65 (19) on the basis of comparable total microbial load (range, 1.4 × 108 to 1.1 × 109 bacteria per mg [wet weight] of dentine) and variable characteristics with respect to high or low abundance of Prevotella spp. as previously determined by real-time PCR enumeration (19). When we chose the 10 samples, patient age and sex were cross matched within the limitations of the sample pool, and a mixed presentation of pulpal histopathology (19) was chosen to maximize the extent of the species and/or phylotypes detected (Table (Table11).
Using the universal primers previously described by Nadkarni et al. (23), each of the 10 samples were subjected to PCR amplification with a FTS-320 thermal cycler (Corbett Research, Sydney, Australia). PCR was performed in triplicate in a total volume of 100 μl by using the AmpliTaq Gold PCR core reagent kit (Applied Biosystems) containing 300 nM concentrations of each of the universal forward and reverse primers. An initial denaturation step of 95°C for 10 min was followed by 25 cycles of 95°C for 15 s and 60°C for 1 min, with a final resting temperature of 4°C. The triplicate PCR samples were combined and purified by using the Wizard PCR Preps DNA purification system (Promega, Sydney, New South Wales, Australia) before being cloned into linearized plasmids by using the pGEM-T Easy vector system (Promega) according to the manufacturer's instructions. Ligated plasmids were transformed into competent Escherichia coli XL1-Blue cells as described previously (4). Between 105 and 110 white colonies were randomly selected from each of the 10 samples and inoculated into 5 ml of Luria-Bertani broth supplemented with 100 μg of ampicillin ml−1 and incubated at 37°C with shaking for 18 h.
A 2-ml aliquot of each of the overnight cultures was used to extract plasmid DNA by using the Wizard Plus MiniPreps DNA purification system (Promega) according to the manufacturer's instructions. The purified plasmids were separated through a 1% agarose gel in 40 mM Tris-acetate buffer (pH 8.0) containing 1 mM EDTA (TAE buffer), stained with 500 ng of ethidium bromide ml−1, and visualized under short-wavelength UV light. Chimeric plasmids containing cloned 16S rRNA gene inserts were selected for sequencing.
Purified plasmids were sequenced at Westmead DNA Sequencing Facility, Westmead Hospital, Wentworthville, New South Wales, Australia, by using the universal forward primer (23). Approximately 410 to 430 bp of each 16S rRNA gene sequence were obtained by this approach.
Chimeric 16S rRNA sequences were identified by the program Chimera_Check, accessed through the Ribosomal Database Project (5), and discarded. All other sequences were compared to 16S rRNA gene in the Ribosomal Database Project (5) and in GenBank. Sequences sharing >99% nucleotide identity to a known species or to an uncultured bacterium were grouped, and a representative sequence was selected for further analysis. The representative sequences were aligned by using the program CLUSTAL W (32) and a distance matrix was calculated by using Dnadist with the Jukes-Cantor model (13). A phylogenetic tree was constructed by the neighbor-joining method of Saitou and Nei (28) using the program Neighbor. Phylogenetic data were subjected to bootstrap analysis of 1,000 replicates by using Seqboot and Consense accessed through the Australian National Genomic Information Service (http://www.angis.org.au).
Ninety-eight percent of the 942 amplicons of the 16S rRNA from 10 different carious lesions displayed a high level of sequence identity (>99%) to either the 16S rRNA sequences of known species or to those of uncultured bacteria lodged in GenBank. The exceptions were three sequences most closely related to Lactobacillus pontis, nine to Clostridium leptum, three to the Lachnospiraceae oral clone MCE9-31, one to Prevotella ruminicola, and one to Denitrobacterium detoxificans. The region of the 16S rRNA genes that were sequenced did not allow differentiation of the closely related oral streptococcal species, Streptococcus sanguinis, Streptococcus oralis, and Streptococcus mitis, which collectively represented 10 amplicons. Similarly, the lack of sequence divergence in the 16S rRNA amplicons did not allow Lactobacillus casei and Lactobacillus rhamnosus or Lactobacillus panis and Lactobacillus reuteri to be differentiated with confidence. Overall, 75 different taxa were identified from the 942 16S rRNA amplicons. In each of the contributing samples, the number of taxa ranged from 7 in sample 3 to 31 in sample 7. In lesions where lactobacilli were the predominant bacterial flora, an average of nine different species or phylotypes were found to predominate in the lesions.
To provide an overview of the diversity of species associated with advanced caries, a phylogenetic tree was constructed from 75 16S rRNA sequences representing the different species and phylotypes in the ten carious lesions (Fig. (Fig.1).1). Of the 75 different bacterial lineages identified, the major families were the Lactobacteriaceae and the Prevotellaceae, with 15 and 13 distinct phylogenetic lineages, respectively. Other diverse families included the Acidaminococcaceae (seven lineages), Eubacteriaceae (six lineages), and Lachnospiraceae (six lineages). Less diverse were the Streptococcaceae, Coriobacteriaceae, and Bifidobacteriaceae, each comprising three distinct lineages. The remaining species were found to represent various divergent bacterial lineages which, with the exception of Pseudoramibacter alactolyticus, Fusobacterium nucleatum, and Propionibacterium spp., were generally found to comprise minor proportions in the 10 carious lesions.
Species belonging to the family Lactobacillacae were abundant and represented 50% of total amplicons (Table (Table2).2). Since this family was dominant in four of the eight lesions in which they were detected, comprising 97, 85, 98, and 95% of the bacteria in samples 1, 2, 3, and 4, respectively, they were designated high-Lactobacillus samples (Table (Table22).
The family Prevotellaceae were also diverse and abundant, representing 15% of total species in the 10 carious dentine samples. The Prevotellaceae were detected in six lesions. They were dominant in samples 8 and 9, comprising 39 and 46%, respectively, of the species present (Table (Table2).2). These carious dentine samples were referred to as high-Prevotella and were found to be populated by a range of other species or phylotypes unlike the high-Lactobacillus samples. In particular, Olsenella spp. and Pseudoramibacter alactolyticus were found in significant proportions in these lesions. For instance, in sample 8, Olsenella spp. comprised 23% of the amplicons, whereas in sample 9, Olsenella spp. and Pseudoramibacter alactolyticus comprised 14 and 26% of amplicons, respectively.
In dentine samples 5, 6, and 7, lower proportions of lactobacilli and prevotellae were identified, with lactobacilli accounting for 67, 11, and 11% of the amplicons and prevotellae accounting for 24, 24 and 20%, respectively (Table (Table2).2). These carious dentine samples were referred to as mid-Lactobacillus/Prevotella. Although sample 5 was composed primarily of lactobacilli and prevotellae, samples 6 and 7 contained relatively high proportions of species belonging to the Lachnospiraceae, with 16 and 13% of the species, respectively, belonging to this family. In addition, F. nucleatum was the predominant species in sample 6, representing 37% of the amplicons.
Since no lactobacilli were detected in the carious dentine sample 10 and only two Prevotella spp. were found, this sample was designated low-Lactobacillus/Prevotella (Table (Table2).2). In this sample, Pseudoramibacter alactolyticus and Propionibacterium spp. were predominant, with 36 and 30%, respectively, of the species belonging to these two genera.
Of the 15 major lineages in the family Lactobacillaceae, nine species/phylotypes were identified with higher frequency (Fig. (Fig.1).1). These were generally identified in the high-Lactobacillus samples. Of note was the observation that Lactobacillus casei and/or Lactobacillus rhamnosus were identified in samples where significant proportions of Prevotella were also found. Among the Prevotella, 5 of the 13 species or phylotypes were identified with higher frequency. In particular, the uncharacterized Prevotella spp., corresponding to oral phylotype IDR-CEC-0032 (GenBank accession number AY550995), were found to comprise the highest proportion of these Prevotella.
Phylogenetic analysis based on 16S rRNA gene(s) has demonstrated the extreme complexity of microbial populations in a variety of habitats within the oral cavity, with the current estimate of the total oral flora to be ca. 700 species (14). This complexity is particularly evident for the gingival crevice (26) and dorsum of the tongue (14) but also for the dental plaque biofilm associated with the initiation of the carious enamel lesion (3). It is believed that only 50% of bacteria in the oral cavity are cultivable (30), requiring a molecular approach to elucidate the diversity of microflora associated with disease progression.
The present molecular study was undertaken to define the profile of bacteria associated with carious dentine after 25 rounds of PCR, with 30 cycles being the most commonly used in other studies (16, 20, 26). Only the most prevalent bacteria were detected by the methods used since after 25 rounds of PCR the chance of detecting a given bacterium with 95% confidence requires the bacterium to be present as ≥3% of the population when 100 clones are sequenced. This implies that less-abundant species or phylotypes were not disclosed. Although the method used to extract and protect DNA from degradation readily lyses gram-negative bacteria, as well as many gram-positive bacteria such as lactobacilli growing under anaerobic conditions, the method does not allow for the quantitative extraction of DNA from all gram-positive bacteria whether or not they are growing anaerobically (4, 23). Thus, streptococci, including the mutans streptococci which are associated with the initiation of dental caries (8, 33), would be underestimated by this method. However, in the 10 samples analyzed, the relative levels of cultivable streptococci were known in relation to the colony counts for Prevotellaceae and lactobacilli (and other species ). Based on the chance of finding streptococci in the 10 samples analyzed, the data for 3 of the samples were at odds with the cultivable data. These were samples 1 and 6, in which more streptococci would be expected to be detected, and sample 4, in which no streptococci would be expected after 25 rounds of PCR amplification despite streptococci being known to be present in the sample. Interestingly, a recent study of five carious dentine samples reported that the numbers of cariogenic Streptococcus mutans were very low in three of the five samples analyzed (20). The same study also reported very low levels of anaerobic gram-negative bacteria, suggesting a bias toward extraction of DNA from gram-positive bacteria consistent with the DNA extraction protocol used (6).
Although there is a risk of overinterpretation of the dynamics of the microbial populations during progression of the carious lesion within the dentine matrix, the results of the current study of 10 selected carious lesions, chosen on the basis of comparable total microbial load and variable Prevotella load (19), show a high degree of variability. Up to 31 different taxa were detected in each of the 10 carious dentine samples, similar to the diversity reported to be associated with endodontic infections (21) and to the diversity in carious dentine seen by using a combination of culture-based and cloning techniques (20), but greater than the level of diversity determined in previous carious dentine studies with only culture-based methods (1, 7, 11).
Lesions 1, 2, 3, and 4 were dominated by lactobacilli in various combinations, a finding compatible with real-time PCR analysis of these samples that indicated the paucity of Fusobacterium spp. and Micromonas (formally Peptostreptococcus) micros, Porphyromonas endodontalis, and Porphyromonas gingivalis (19). Rather, high levels of Lactobacillus delbrueckii, Lactobacillus ultunensis, Lactobacillus crispatus, and Lactobacillus salivarius (Table (Table2)2) were associated with these lesions in agreement with the finding by Munson et al. (20) that taxa of the genus Lactobacillus dominate carious dentine. It is possible that successful early colonization of the dentinal matrix by lactobacilli is favored by abundant dietary carbohydrate that, once fermented, creates an acidic environment which excludes subsequent colonization by other species. Exclusion may also depend on the fermentation by-products of some of these lactobacilli, since those of Lactobacillus delbrueckii, Lactobacillus ultunensis, and Lactobacillus crispatus (a group of organisms more frequently associated with other regions of the gastrointestinal tract  are not well characterized. It is known, however, that the proportions of lactic, acetic, and propionic acids vary considerably between lesions (9). Lactate-dependent Veillonella spp. were infrequently detected, and even though other members of the family Acidaminococcaceae, such as Selenomonas spp., as well as species such as Pseudoramibacter alactolyticus, are capable of metabolizing lactate (12), none of these bacteria were associated with high-Lactobacillus lesions. This suggests that lactate is not the major fermentation by-product in these lesions. However, these lactate-utilizing bacteria were found in other lesions, particularly those containing Prevotella (Table (Table2).2). Interestingly, Pseudoramibacter alactolyticus, which can also ferment both carbohydrates and lactate, is found in 56% of endodontic infections (29) and was present at high levels in samples 9 and 10 where no lactobacilli were detected.
In contrast to the high-Lactobacillus lesions, five of the six carious dentine samples with a low abundance of lactobacilli were associated with high levels of Prevotellaceae, as well as members of either the family Lachnospiraceae, or Olsenella spp. (Table (Table2).2). Numerically abundant Prevotella organisms were mainly allocated to novel, uncultured forms of unknown biology with a major group represented by phylotype IDR-CEC-0032, which is most closely related to a Prevotella-like rumen bacterium (22).
Invasion of bacteria into the dentinal tubules implies that properties such as adhesion and motility are involved. Members of the families Acidaminococcaceae, such as Selenomonas spp., and the Lachnospiraceae possess flagellae and hence are motile, although the large size of some selenomonads (2 by 10 μm) could limit their penetration into tubules which at the dentine-enamel junction are ca. 1 to 2 μm in diameter (15). On the other hand, nonmotile proteolytic bacteria such as Porphyromonas gingivalis have been shown to invade dentinal tubules in vitro by coaggregation to chains of Streptococcus gordonii since these chains extend into the tubules as the Streptococcus organisms multiply (18). Our data suggest, however, that neither the Streptococcaceae nor the porphyromonads are found in high numbers in samples containing high proportions of Lactobacillus and/or Prevotella spp. (Table (Table2).2). This finding supports our previous analysis of 65 carious dentine lesions wherein the abundance of highly proteolytic organisms, such as Porphyromonas gingivalis and Porphyromonas endodontalis (27), was found to be uniformly low (19) despite the expectation that the exposed decalcified and denatured collagenous framework of the dentinal tubules would provide a rich nutrient source (2). This implies that an essential nutrient, such as hemin (25, 36), may be limiting or that some other unspecified environmental factor restricts the dominance of the porphyromonads. However, other bacteria, including Micromonas micros, certain Prevotella spp. and the Propionibacterium sp. oral strain FMA5, which accounted for 30% of the bacterial population in sample 10, have been shown to degrade proteins such as gelatin and albumin (10, 24, 34). Such bacteria may therefore provide peptides and amino acids required by other bacteria within the polymicrobial population. Interestingly, Propionibacterium sp. oral strain FMA5 was one of only three taxa found in all five carious dentine samples recently analyzed by Munson et al. (20).
Since no particular consistent pattern of dominance was observed in the 10 selected carious dentine samples, it is probable that dominance by individual taxa was achieved after early invasion of the exposed dentine matrix. This would suggest that the composition of the dominant groups could depend on early events, including invasion by species present in supragingival plaque on the tooth surface and/or random contamination of the open lesion from salivary and dietary sources. Rates of lesion progression would then depend on the nature and abundance of acidogenic and aciduric organisms and, at least initially, on the continued availability of dietary fermentable carbohydrates. Migration of proteolytic bacteria to the proximity of the pulp to exploit additional nutrient resources could drive the final phase of the carious lesion and result in pulpal death and translocation of the bacteria into the pulp chamber. Accordingly, the composition and pathogenic potential of bacterial populations subsequently associated with endodontic infections might be determined by relatively early events occurring in the establishment of the carious dentine lesion.