Real-Time TaqMan PCR for Quantifying Oral Bacteria during Biofilm Formation

Real-time PCR assays are powerful, convenient tools that are capable of quantification and identification in a single assay. The key to the success of the TaqMan method is designing specific primers and probes based on the nucleotide sequence of the target organism. Genes encoding species-specific proteins or 16S ribosome RNA genes are generally used as the targets of real-time PCR (7, 13). However, it is difficult to isolate specific nucleotide sequences from the genes of organisms for which species-specific proteins have not been characterized or that distinguish the 16S rRNA genes of closely related species. Human oral dental plaque is characterized by a number of closely related species.

This study used genomic subtractive hybridization to prepare species-specific PCR primers and TaqMan probes for early colonizer microorganisms in dental biofilms: Streptococcus gordonii, Streptococcus mitis, Actinomyces naeslundii, and Actinomyces viscosus. The technique was initially developed to isolate differences in cDNA pools (6), but it has also successfully identified genomic differences between closely related strains (1, 12). This report describes both the identification of species-specific sequences for two Streptococcus and two Actinomyces species using subtractive hybridization and the quantification of these four early oral bacteria colonizers and two cariogenic bacteria—-Streptococcus mutans and Streptococcus sobrinus—-during dental biofilm formation, using a TaqMan PCR assay.

The bacterial strains used in this study are listed in Table 1. S. gordonii, S. mitis, A. naeslundii, and A. viscosus were grown anaerobically (85% N₂, 10% H₂, 5% CO₂) at 37°C in Todd-Hewitt broth (Difco Laboratories, Detroit, Mich.). Chromosomal DNA was purified using a Puregene DNA isolation kit (Gentra Systems, Minneapolis, Minn.) according to the manufacturer's instructions.

Genomic subtractive hybridization was performed as previously described (10). The chromosomal DNA of Streptococcus oralis ATCC 10557 was used as driver DNA for S. gordonii and S. mitis, and the chromosomal DNAs of A. naeslundii and A. viscosus were each used as driver DNA for the other. The oligonucleotide adapters used in this study are listed in Table 2. The second-round PCR products were digested with Sau3AI, cloned into BamHI-digested pBluescript II SK⁺ (Stratagene, La Jolla, Calif.), and then used to transform Escherichia coli DH5α (Takara Bio Company, Shiga, Japan). Eight colonies were selected randomly from each cDNA bank containing 200 colonies, and the nucleotide sequences were determined using an ABI PRISM 310 genetic analyzer (Applied Biosystems, Foster City, Calif.). A total of 32 fragments ranged from 150 to 550 bp and had different nucleotide sequences.

A nucleotide database search (blastn) (2) revealed that 21 insertions have no significant sequence homology to any known DNA sequences in GenBank, and the nucleotide sequences of seven fragments were detected in the driver bacterial DNA. The other three insertions are homologous to 16S rRNAs of oral bacteria, and the eight insertions share homology with the genes encoding unknown proteins. Table 3 shows the characteristics of the Sau3AI fragments selected to design the primers and probes.

A protein database search (blastp) (2) revealed that the protein-encoding region in the fragment isolated from S. gordonii shared 52% identity with the putative type I site-specific DNase in Streptococcus pyogenes M1, which is a strict human pathogen that causes a variety of diseases, including pharyngitis, scarlet fever, impetigo, and erysipelas (3). The protein encoded by the S. mitis fragment shared 38% identity with a glycine-rich protein of the plant Arabidopsis thaliana (4). The product of the A. naeslundii fragment shared 33% identity with an unknown protein of Mesorhizobium loti, which is a symbiotic bacterium that associates with plants (5). The product of the A. viscosus fragment shared 35% identity with an unknown protein of the rice species Oryza sativa (11). The blastx similarities of the products showed expected values of >0.1, except for the S. gordonii sequence (E = 7e − 37). Although we did not investigate this further, the strain-specific products may partly explain the characteristics of these organisms.

The S. gordonii-, S. mitis-, A. naeslundii-, and A. viscosus-specific primers and probes were designed from the Sau3AI DNA fragments by genomic subtractive hybridization using the software Primer Express 1.5 (Applied Biosystems) (Table 2). The specificities of the DNA sequences between the primers were initially confirmed by blastn and then further confirmed by conventional PCR with the oral bacterial DNA listed in Table 1. We confirmed that PCRs with the primers were positive for all intended target species and that no cross-reactivity with other nontarget species was observed.

Using these primers and probes, we developed a TaqMan PCR assay to quantify these bacteria. Amplification and detection were performed using the ABI PRISM 7700 sequence detection system (Applied Biosystems), as previously described (13). Standard curves for each organism were plotted for each primer-probe set using the Ct (the cycle number at which the threshold fluorescence was reached) values obtained by amplifying successive 10-fold dilutions of a known concentration of DNA. The DNA concentrations were 1.1 × 10⁹ CFU/ml for S. gordonii, 0.9 × 10⁸ CFU/ml for S. mitis, 3.3 × 10⁸ CFU/ml for A. naeslundii, and 1.0 × 10⁸ CFU/ml for A. viscosus. The numbers of CFU were determined by plating culture dilutions on suitable agar plates. The assay was capable of detecting bacterial DNA linearly for dilutions from 10⁻³ to 10⁻⁸ for every organism. Using this approach, correlations between Ct and CFU were observed (Fig. 1). Detection and quantification were linear over the following ranges for the different species: 2.2 × 10¹ to 2.2 × 10⁶ cells for S. gordonii, 1.8 × 10⁰ to 1.8 × 10⁵ cells for S. mitis, 6.6 × 10⁰ to 6.6 × 10⁵ cells for A. naeslundii, and 2.0 × 10⁰ to 2.0 × 10⁵ cells for A. viscosus. The presence of PCR inhibitors in dental plaque was assessed using the fluorescence levels for serial dilutions of each lysed bacterium. In this study, lysates with or without 10 μg (wet weight) of added dental plaque that did not contain any target bacteria showed no inhibition (data not shown). S. mutans and S. sobrinus were quantified as previously described (13).

Using this real-time PCR assay, we examined the numbers of four early colonizer organisms and two cariogenic bacteria in dental plaque from five individuals (Table 4). Dental plaque specimens were collected from the buccal side of the upper first molar. We suspended 100 mg (wet weight) of plaque in 1.0 ml of phosphate-buffered saline (0.12 M NaCl, 0.01 M Na₂HPO₄, 5 mM KH₂PO₄ [pH 7.5]), placed it in sterile screw-cap tubes with 1.0 g of 0.1-mm-diameter glass beads, and beat it for 20 min at regular intervals in a Mini-BeadBeater 8 (BioSpec Products, Inc., Bartlesville, Okla.) at 4°C. The supernatant without beads was used as the template for the real-time PCR assay. The cell numbers per milligram (wet weight) of plaque ranged from 0 to 1.25 × 10⁴ for S. mutans, 2.02 × 10⁰ to 1.08 × 10⁵ for S. sobrinus, 0 to 3.56 × 10² for S. gordonii, 1.47 × 10³ to 3.05 × 10⁵ for S. mitis, 0 to 5.27 × 10⁵ for A. naeslundii, and 0 to 1.45 × 10⁰ for A. viscosus. Large numbers of S. mitis were detected in all samples, while S. gordonii and A. viscosus were minor species. Although more investigations are required, our results are in accord with previous findings (8, 9).

Our study revealed that the TaqMan assay is accurate and useful for quantifying early colonizer organisms and cariogenic bacteria in dental biofilms. Furthermore, genomic subtractive hybridization facilitated the isolation of specific DNA sequences for the target organism versus related organisms. Real-time PCR in combination with genomic subtractive hybridization may be useful for studying microbial ecosystems involved in the formation and etiology of biofilms.

• Open in new tab
• Download powerpoint

FIG. 1.

Standard curves generated from known numbers of S. gordonii DL1 (A), S. mitis 903 (B), A. naeslundii ATCC 51655 (C), and A. viscosus ATCC 43146 (D). Linearity was shown for these organisms. The respective correlation coefficients were 0.998, 0.998, 1.00, and 1.00.

• Copyright © 2004 American Society for Microbiology