Reexamination In Vitro and In Situ of an Antibacterially Modified Experimental Dental Resin Composite with Molecular Methods: A Pilot Study
Purpose. To introduce additional methods to detect and to quantify single pathogens in the complex biofilm formation on an antibacterial dental material. Materials and Methods. A conventional (ST) and an antibacterial dental composite (B) were manufactured. In vitro: specimens were incubated with a mixture of early colonizers. Bacterial adhesion was analyzed by TaqMan PCR after 8/24 h. In situ: TaqMan PCR and 16S rRNA Next Generation Sequencing (NGS) were performed. Results. In vitro: after 8 h incubation, B was covered by 58.6% of the bacterial amount that was attached to ST. After 24 h, the amount of attached bacteria to ST remained constant on ST only slightly lower on B. In situ: after 8 h the amount of adhering A. viscosus and S. mitis was prominent on ST and reduced on B. NGS revealed that S. sanguinis, S. parasanguinis, and Gemella sanguinis were the mainly attached species with S. sanguinis dominant on ST and S. parasanguinis and G. sanguinis dominant on B. Conclusions. Initial biofilm formation was altered by B. A shift between actinomycetes and streptococci was observed in situ. TaqMan PCR and 16S rRNA NGS revealed comparable results in situ and demonstrated the usefulness of NGS to characterize complex bacterial communities.
Pellicle and bacterial adherence is considered to play an important role in early plaque formation [1–3]. It is of interest to avoid or at least to diminish bacterial adherence not only on hard tooth tissues but also on dental restorative materials [4–7]. Hence many attempts have been made to modify dental resin composites to that effect [1, 8–13]. To test the expected antibacterial effects mainly confocal laser scanning microscopy [2, 13–15] or fluorescence microscopy [1–3, 8, 16] and predominantly bacteria as S. gordonii , S. mutans [13, 18–21], S. mitis [1, 8, 18, 22, 23], S. oralis [1, 8, 17, 18, 20], S. sanguinis [1, 8, 17, 24], A. naeslundii [1, 2, 8, 15], A. odontolyticus [21, 25], and A. viscosus [8, 26–28] were used.
This work focuses on the reexamination of the bacterial attachment to an experimental dental resin composite, called material B . Material B is modified with 4% (m/m) of the antibacterial monomer (5-chloro-2-2,4-dichlorophenoxy)phenyl methacrylate (called Methacryl-Irga in the following); its antibacterial properties were examined in monoclonal in vitro settings in comparison with a standard, called ST, with fluorescence microscopy . (5-chloro-2-2,4-dichlorophenoxy)phenol is known under the trademarks Irgasan or Triclosan, respectively, and is a well-proven broad spectrum antimicrobial agent. It inhibits the enoyl-acyl-carrier protein reductase component of type II fatty acid synthase in bacteria, the mammalian fatty acid synthase, and provides anticariogenic activity [29, 30]. Irgasan/Triclosan was used to synthesize Methycryl-Irga . The aforesaid investigation reports good antibacterial effectiveness for material B for the streptococci S. oralis and S. sanguinis and S. mitis and for the actinomycetes A. naeslundii and A. viscosus .
The goal of the present investigation was to analyze bacterial mixtures instead of a single isolated species of the aforesaid bacteria in their in vitro and in situ attachment to ST and the antibacterially modified material B by real-time PCR (TaqMan-PCR) and Next Generation Sequencing (NGS). Since the published TaqMan-PCRs to detect A. naeslundii [31, 32], A. viscosus, or S. mitis  were not found to be sufficient specific and sensitive when reevaluated they were newly designed. Although the present work must be considered as a pilot study it is thought to be valuable because additional methods are introduced to detect and to quantify simultaneously single pathogens in the complex situation of biofilms. It was hypothesized that there is no difference in total bacteria load between the materials ST and B.
2. Material and Methods
For detailed information about manufacturing of the test materials, refer to the literature ; the respective raw materials are listed in Table 1. The standard ST represents a conventional dental resin composite. Material B differs from ST by replacing 4% (m/m) of the resin matrix with (5-chloro-2-2,4-dichlorophenoxy)phenyl methacrylate (see Table 2). discs (thickness: mm; diameter: mm) were made and cured 40 s each side and then stored for 14 days in water at 37°C, disinfected, and wet-polished (sterile water) on one side with fine and superfine polishing-discs according to . Twenty-four discs were used for the in vitro and four for the in situ adhesion assay.
2.1.1. Bacterial Plate Cultures
To test the analytical specificities and the sensitivities of primers and probes bacteria (Table 3) were grown on COS blood agar plates (bioMerieux GmbH, Nürtingen, Germany) and incubated for 48 h at 37°C: actinomycetes under anaerobic conditions and streptococci under 5% (v/v) carbon dioxide. The same culture conditions were utilised for counting the CFUs of each strain.
2.1.2. Bacteria Liquid Cultures for In Vitro Adhesion Assay
Bacteria were grown overnight at 37°C in 2.98% (w/v) thioglycollate (THIO-) broth (Merck GmbH, Darmstadt, Germany), streptococci by shaking under atmospheric air, and actinomycetes under anaerobic conditions. Dilutions of 1 : 6 to 1 : 10 were cultivated for 2–4 h up to a turbidity of 1 McFarland, which corresponds to an optical density of 0.2–0.3 OD at 600 nm. Cells were harvested by centrifugation (5 min at 4,000 ×g) and suspended in the same volume of PBS, and a mixture of bacterial suspensions in equal parts was made. Total genomic DNA of 350 μL bacterial mixture was prepared for subsequent quantification of each species by TaqMan-PCR.
2.2. In Vitro Adhesion Assays
Human saliva was collected from ten healthy volunteers, prepared, and sterilized according to Rüttermann et al. . Approval from the ethics committee was obtained (Ethics Committee of the Medical Faculty of Heinrich-Heine-University, Düsseldorf, Germany, internal study number: 2912). The participants gave their consent verbally, as they all were part of the academic faculty staff of the department. Documentation of the oral consent was done by making a list of prospective saliva donors. All donors who subsequently gave their oral consent were checked-off on this list. The oral consent was explicitly judged as adequate, and the need for written informed consent from the participants was waived by the ethics committee. For each test material (ST, B) and incubation period (8 h/24 h), six discs (24 discs in total) were placed with the polished side up in a 24-well plate and preincubated with 250 μL sterile saliva (1 : 2 in PBS) for 2 h at 37°C. Next, the disks were gently rinsed two times with PBS (37°C) and 350 μL of the bacterial mixture was added to each pellicle coated disk and incubated at 37°C under 5% CO2 (v/v) for 8 h or 24 h. PBS was used as negative control. After 8 h or 24 h of incubation the supernatant was removed and the disks were gently washed with 300 μL PBS (37°C) and transferred to a clean well of a 24-well plate before preparation of total genomic DNA.
2.3. In Situ Adhesion Assays
The Ethics Committee of the Medical Faculty of Heinrich-Heine University Düsseldorf, Germany, stated that there is no need for further vote as the two participants of the in situ experiment were two of the authors themselves (UP and SR), who gave their written consent. Their saliva samples and mucosal swabs were taken for analysis. One disk of each material was separated into four parts to obtain suitable pieces to be inserted in individual oral splints on the upper jaw at each molar on the buccal side [33, 34]. The splints were steadily worn for 8 hours overnight, removed, and gently washed with 300 μL PBS (37°C) and all quarters of a disk were transferred to one clean well on a 24-well plate prior to preparation of total genomic DNA.
2.4. Preparation of Total Genomic DNA
Total genomic DNA was isolated with the Gram-positive protocol of the DNA Blood and Tissue Kit (Qiagen GmbH, Hilden, Germany). (a) Cells of the bacterial suspensions were collected by centrifugation (10 min at 13,000 ×g), resuspended in 300 μL enzymatic lysis buffer, and incubated for 30 min at 37°C. (b) Bacterial-coated disks were incubated with 300 μL lysis buffer and the 8 h specimens after overnight storage at −20°C were incubated for at least 30 min at 37°C.
30 μL Proteinase K (>600 mU/mL) and 300 μL Buffer AL (Qiagen GmbH) were added to each sample and incubated at 56°C for 1 h. After addition of 300 μL ethanol mixtures were transferred to DNeasy Mini spin columns (Qiagen GmbH). Total genomic DNA was eluted in 100 μL and stored at −20°C until use.
2.5. Design of Primers and Probes
Multiple sequence alignments were calculated with the MegAlign algorithm of the DNASTAR sequence analysis software (DNASTAR Inc., WI, Madison, USA). Gene sequences were retrieved from the nucleotide database of the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/nucleotide/). The final design of the specific primers and probes (Table 4) (MWG-Biotech GmbH, Ebersberg, Germany; metabion GmbH, Planegg-Martinsried, Germany; Eurogentec, Seraing, Belgium) was either calculated with the Primer Express Software (Applied Biosystems Inc., Foster City, CA, USA) or taken from the literature .
In house TaqMan-PCRs were carried out in a total volume of 25 μL consisting of 12.5 μL Eurogentec PCR MasterMix NO ROX (Eurogentec GmbH Reference: RT-QP2X-03NR, containing buffer, dNTPs with dUTP, HotGoldStar DNA polymerase, 5 mM MgCl2, Uracil-N-glycosylase, and stabilizers), 300 nM of each forward and reverse primer, 200 nM labelled probe, and 2.5 μL of template DNA. Quantification standards were amplicon carrying plasmids in three different concentrations (105, 103, and 102 copies/μL) (Table 4). Each sample was analyzed twice. Thermal cycling, fluorescent data collection, and analysis were done with the iCycler (BioRad Laboratories, Munich, Germany). The cycling program was as follows: 1 cycle at 50°C for 10 min and 1 cycle at 95°C for 10 min followed by 45 cycles at 95°C for 15 s and 60°C for 1 min.
2.7. PCR Amplicon Library Preparation
PCR amplification of the 16S rDNA hypervariable region V6 was done with a pool of five degenerated forward and four degenerated reverse primers expected to target the known groups of bacteria according to Huber et al. . The V6 region was bidirectional sequenced with two different primer sets: (a) 5′-ends of the forward primers fused with the A-adaptor and the reverse primers 5′-ends fused with truncated P1 (trP1)-adaptors (Life Technologies GmbH, Darmstadt, Germany, User Guide #4468326C, Ion Amplicon Library Preparation, Fusion Method) and (b) trP1 fused to the 5′-ends of the forward primers and A to the 5′-ends of the reverse primers.
Per 25 μL PCR reaction the amplicon library preparation consisted of 1 μL DNA (1% (v/v) total DNA-preparation/sample, volunteer 1 only), 1 U Platinum Taq DNA Polymerase High Fidelity, 2 mM MgSO4, 0.2 mM dNTPs, and 0.4 μM of each forward and reverse primer. The cycling was as follows: 94°C/3 min, 36 cycles: 94°C/15 s, 58°C/30 s, 68°C/1 min, and finally 68°C/30 min and hold at 4°C. The PCR-quality was checked with gel electrophoresis on 2% (w/v) agarose gel in 1x Tris Borat EDTA buffer. Amplicon library was purified (QIAquick PCR Purification Kit, Qiagen GmbH) and quantified (Qubit Quant-iT HS DNA Assay, Invitrogen GmbH, Darmstadt, Germany) and finally quality and size of amplicons were determined with the Agilent 2100 Bioanalyzer DNA Lab Chip 1000 (Agilent Technologies GmbH, Böblingen, Germany). The concentration of the amplicon library for clonal amplification was adjusted to 6 pmol/L.
2.8. Emulsion PCR and Sequencing
The emulsion PCR for enrichment of template positive IonSphere particles was done with the Ion OneTouch 200 Template Kit (Life Technologies GmbH) and the quality of the IonSphere particles was checked with the Qubit 2.0 Fluorometer (Invitrogen GmbH) and the IonSphere Quality Control Kit (Life Technologies GmbH). The amplicon libraries were sequenced twice to increase sequencing depth with the Ion Torrent Personal Genome Machine (PGM), the Ion Torrent 314 Chips, and the Ion PGM 200 Sequencing Kit (Life Technologies GmbH) according to the respective protocol.
2.9. NGS (Next Generation Sequencing) Analysis
2.9.1. Sequencing Read Preprocessing
Raw sequencing reads were filtered for low quality and PCR primer sequences applying three consecutive quality filter steps to the two individual sequence data sets per in situ sample: (a) reads <10% of bases quality score ≤25 were discarded (FASTQ quality filter/FASTX-Toolkit, http://cancan.cshl.edu/labmembers/gordon/fastx_toolkit), (b) PCR primer sequences were removed by semiglobal alignment with Cutadapt (http://journal.embnet.org/index.php/embnetjournal/article/view/200), and (c) in Cutadapt the minimum length of the trimmed reads was set to 50 bases.
2.9.2. Diversity and Taxonomic Analyses
Preprocessed high quality sequence reads were by BLAST (Basic Local Alignment Search Tool) applying the Human Oral Microbiome Database (HOMD), version 13.2, 16S rRNA RefSeq, according to Chen et al. . BLAST analyses were calculated with the CLC Genomics Workbench Software Version 6.0.5 (Qiagen GmbH) applying CLC standard BLAST parameters. Query sequences matching more than one target sequence with equal BLAST scores were randomly assigned.
Matching reads per taxon were normalized for the total number of query sequences for each sample to compare the relative bacterial composition of the two in situ samples. The cut-off for comparative analysis of relative abundances was set to 0.1% after normalization to focus on a more robust subset of the identified taxa.
2.10.1. In Vitro Adhesion Assay
Each material, ST and B, was analyzed sixfold at 8 h and 24 h, respectively. In TaqMan-PCR, DNA samples were analyzed in duplicate. Descriptive statistics were accomplished; that is, means and standard deviations were calculated.
2.10.2. NGS Analysis
Technical reproducibility of the Ion Torrent sequencing reaction was controlled using linear regression coefficient analysis.
Species-specific regions for the bacteria tested and the number of copies per genome are shown in Table 4. Due to three (A. naeslundii) and four (S. mitis) 16S rDNA copies per genome, these TaqMan-PCRs had higher sensitivities than the other ones which targeted single copy genes. For A. viscosus, S. oralis, S. sanguinis, and S. gordonii, the glucosyltransferase (GTF) gene (existent in most streptococci ) was used for targeting S. gordonii (gtfG), S. oralis (gtfR), and S. sanguinis (gtfP) and the sialidase gene (nanH) for targeting A. viscosus. The efficiency of the TaqMan-PCRs ranged from 94.5% (S. sanguinis) to 99.9% (A. odontolyticus) and the coefficient of determination ranged from 0.990 to 0.999 (ideal value = 1.000). Cross reactions with cognate species amounted to less than 0.001%.
3.2. In Vitro Adhesion Assay
The quantification of the different species in the inoculum with TaqMan-PCRs proved the bacterial concentration of a McFarland 1 suspension as species-specific but not always of the same value (Figure 1). Streptococci were overrepresented in the inoculum in contrast to the actinomycetes. After 8 h of incubation the antibacterial material B was covered with 58.6% of the bacterial amount of the standard material ST. After 24 h the amount of attached bacteria remained constant on material ST (99.8%) and increased on material B to 92.6% of the bacterial amount on ST. The bacterial composition on both materials was quite similar. More than 90% of the attached bacteria accounted for S. sanguinis, S. oralis, and S. gordonii with 36% S. oralis after 8 h and 50% after 24 h of incubation. A. viscosus (3%), A. naeslundii (<1%), and S. mitis (1%) were underrepresented on both materials ST and B after 8 h which further declined after 24 h (Figure 1 and Table 5).
3.3. In Situ Adhesion Assay
The bacterial loads in saliva of volunteers 1 and 2 were similarly concentrated ( and , resp.), dominated by A. odontolyticus in volunteer 1 (Figure 2(a1)) and A. odontolyticus and S. mitis in volunteer 2 (Figure 2(a2)). The diversity of bacteria and their quantities varied in the mucosal swabs volunteer-specifically (Figures 2(b1) and 2(b2)). Volunteer 1 showed less bacterial load than volunteer 2. A. viscosus was missing and the streptococci dominated. Volunteer 2 showed 1,000-fold increased actinomycetes and streptococci with a dominance of streptococci.
After 8 h of in situ carriage, A. viscosus and S. mitis dominated on ST with respect to both volunteers (Figures 2(c1) and 2(c2)). The load of these species was heavily reduced on material B of volunteer 1, but on material B of volunteer 2 the reduction was only due to the disappearance of A. viscosus (Figure 2(c2)). Volunteer 1 showed 10-fold increased bacterial load on the test materials compared to volunteer 2.
3.4. Next Generation Sequencing (NGS)
Reproducibility of the 16S amplicon sequencing reactions was confirmed by comparing two duplicated data sets from each of the materials ST and B which showed very good correlations of 99.97% and 99.51%, respectively. Therefore, after preprocessing the two filtered sequence data sets per in situ sample were combined for subsequent analyses to increase sequencing depth. Consequently, after 8 h in situ the specimens of volunteer 1 showed 504,710 for ST and 415,258 for material B high quality sequence reads, probably indicating a lower bacterial load of 82% on material B.
BLAST analyses of the test materials’ sequences with HOMD 16S rRNA database yielded five taxa on the phylum and 11 taxa on the genus level for both of the materials (Figure 3(a)). Firmicutes were the dominant phylum, Streptococcus and Gemella were the dominant genera, and bacteroidetes were almost absent. Material B had higher amounts of actinobacteria especially of Rothia and Actinomyces than ST. Both materials showed 247 taxa species-wise with two sequence reads for the respective rRNA target sequence, and the main species were S. sanguinis, S. parasanguinis, and G. sanguinis. S. sanguinis (approx. 37% of the microbiota) and S. parasanguinis (approx. 25% of the microbiota) dominated on material ST, whereas G. sanguinis (approx. 25% of the microbiota) dominated on material B (Figure 3(b)). R. mucilaginosa and A. odontolyticus showed seven- and fourfold higher amounts on material B, whereas S. sanguinis load was twofold higher on ST (Figure 3(c)).
This pilot study reevaluates the antibacterial effect of the experimental dental resin composite material B in comparison with a standard dental resin composite ST  with more sophisticated methods. The test materials (raw materials, manufacturing/synthesis, formulation, properties, antibacterial effects, and preparation of specimens and saliva) and the rationale for their choice as well as for the selection of the bacteria (A. viscosus, A. naeslundii, S. mitis, S. oralis, and S. sanguinis) have already been thoroughly described by Rüttermann et al. [1, 8]. The results of the aforesaid literature are strongly limited because the attachment of only single isolated bacteria was investigated on the test materials with fluorescence microscopy . Biofilm formation is very complex and does not include only the few bacteria individually tested. Much more bacteria live in the oral cavity and coexist with saliva and protein components. Therefore, it is certainly of scientific value to investigate the test materials’ antibacterial effects on complex bacteria mixture in vitro and preferably in situ. This justifies the present pilot study which introduced TaqMan-PCRs and NGS to observe simultaneously distinct pathogens in complex bacterial mixtures. Though more bacteria were identified by TaqMan-PCRs and especially by NGS, only the formerly investigated species A. viscosus, A. naeslundii, S. mitis, S. oralis, and S. sanguinis  are discussed here.
Although TaqMan-PCRs for A. naeslundii [31, 32], A. viscosus, and S. mitis had been published, these PCRs were reevaluated and found to be less specific and sensitive. Therefore, the TaqMan-PCRs were newly designed for all investigated species.
The results of the in vitro adhesion assay, which were quantified by TaqMan-PCR (Figure 1), showed slightly fewer bacteria after 24 h on material B than on ST. The greatest difference was observed after 8 h of incubation, as the amount of attached bacteria on material B counted only 58.6% of the bacterial load on standard material ST. This contradicted our formerly published observation, found by fluorescence microscopy, that there was no difference of the total bacterial load between the incubation periods . This might be due to the different incubation methods (PBS instead of medium) or to an altered adherence mechanism of the individual bacterial species in a polymicrobial mixture. Considering the differences between the test materials the present investigation and the former analysis of Rüttermann et al.  coincide that there are definitely fewer bacteria, considered overall as well as species-wise, on material B than on ST, especially in the first hours of biofilm formation (8 h). In contrast to the other streptococci, the adherence capacity of S. mitis was reduced on both materials tested (Figure 1). This finding is also in good accordance with formerly published results of our research group . Although attachment of A. viscosus and A. naeslundii was reduced to less than 5% of the total bacterial load on both test materials, still fewer actinomycetes loads were detected on material B than on ST. Concluding, the distinct actinomycetes and streptococci species of the polybacterial mixture showed a similar attachment-behaviour like the monobacterial cultures . In using Taqman PCR as a molecular approach, quantification of single species in complex bacterial mixtures is now possible, which was impractical by fluorescence microscopy. However, in contrast to the fluorescence staining method, in which the ratio between vital and nonvital cells can be estimated, amplification of bacterial DNA by TaqMan-PCR does not distinguish between DNA of a vital or nonvital cell. By implementation of propidium monoazide technique as a discriminator  in future, it should be possible to differentiate between vital and nonvital cells even in Taqman-PCR.
One should keep in mind that we used a static biofilm model to simulate the oral situation. We decided to do so, because the advantages of a dynamic model are not crucial for the short observation periods for the analysis of the initial biofilm (8/24 h).
The in situ TaqMan-PCRs (Figure 2) did not coincide in all aspects with its in vitro version after 8 h of incorporation. The mucosal swabs showed substantial differences in the bacterial composition and load between the volunteers (Figures 2(b1) and 2(b2)), which was also the case with the test materials (Figures 2(c1) and 2(c2)). The fact that some of the six species tested were not found in the saliva or rarely detected in the mucosal swabs but showed dominant adherence to the test materials illustrated that the interactions among the magnitude of oral species and between the species and the formulation of the test material play an important role in the resulting bacterial adhesion. Future studies should try to clarify these phenomena. Attempts that were made in biofilm research to simulate the oral cavity [39–41] revealed very low discriminatory power of single species, indicating that the oral cavity is certainly the best biofilm reactor at present.
Thus, Next Generation Sequencing has the opportunity to characterise the complexity of the oral microbiome in saliva and mucosal swab and attached to the test material and, depending on the discriminatory power of the target region, to enable a relative quantification of the bacterial families, genera, or species. The target region of the NGS analysis used here (Figure 3) was already described in the literature  to elucidate shifts in bacterial communities in subgingival plaque. The comparison of NGS and TaqMan-PCRs showed higher PCR-derived genome equivalents than NGS sequencing reads, but both confirmed lower loads of A. viscosus and A. naeslundii and higher loads of A. odontolyticus on material B than on ST (Figure 4). It was difficult to identify and to compare the quantities of the streptococci due to the low discriminatory power of the short V6 region in streptococci (e.g., S. oralis, S. mitis, S. sanguinis, S. pyogenes, and S. gordonii). TaqMan-PCRs proved lower loads of these streptococci to material B (Figure 4(b)) than NGS analysis. Identical NGS reads complicated the differentiation between S. oralis, S. mitis, or S. sanguinis (Figure 4(a): Ssan/Sora) and between S. gordonii and S. pyogenes (Figure 4(a): Sgor/Spyo). These results indicated that longer stretches of the hypervariable 16S rDNA regions are required for microbiota analyses down to species-level.
The following can be concluded:(i) In vitro and in situ TaqMan-PCRs and NGS analyses coincided with the literature  that fewer bacteria adhered on material B than on ST.(ii)The literature’s  species-wise testing of antibacterial effects gave only tentative results.(iii)The literature  and the in vitro TaqMan-PCRs did not coincide with the in situ TaqMan-PCRs and the NGS in the differentiation of the adhered bacteria.(iv)Sound results can only be obtained by in situ/in vivo testing.(v)The following limitations need to be considered:(a)TaqMan-PCR should be upgraded to differentiate between vital and nonvital cells.(b)Target region of NGS analysis needs to be optimized to better differentiate between the (early oral colonizing) bacteria.Although the statistical power of this study is too low to give detailed results, the data reveal that TaqMan-PCR and NGS analysis are highly valuable molecular methods to investigate biofilms on restorative materials in the oral cavity. Sound results can only be achieved by clinical studies and a statistically sufficient number of volunteers which will follow this pilot study.
The authors declare that they have no competing interests.
The authors thank Dana Belick (Institute of Medical Microbiology and Hospital Hygiene, University of Düsseldorf) for her excellent technical assistance.
F. Aykent, I. Yondem, A. G. Ozyesil, S. K. Gunal, M. C. Avunduk, and S. Ozkan, “Effect of different finishing techniques for restorative materials on surface roughness and bacterial adhesion,” The Journal of Prosthetic Dentistry, vol. 103, no. 4, pp. 221–227, 2010.View at: Publisher Site | Google Scholar
A. Al-Ahmad, A. Wunder, T. M. Auschill et al., “The in vivo dynamics of Streptococcus spp., Actinomyces naeslundii, Fusobacterium nucleatum and Veillonella spp. in dental plaque biofilm as analysed by five-colour multiplex fluorescence in situ hybridization,” Journal of Medical Microbiology, vol. 56, no. 5, pp. 681–687, 2007.View at: Publisher Site | Google Scholar
L. Mei, H. J. Busscher, H. C. van der Mei, Y. Chen, J. de Vries, and Y. Ren, “Oral bacterial adhesion forces to biomaterial surfaces constituting the bracket-adhesive-enamel junction in orthodontic treatment,” European Journal of Oral Sciences, vol. 117, no. 4, pp. 419–426, 2009.View at: Publisher Site | Google Scholar
R. Z. Thomas, H. C. Van Der Mei, M. H. Van Der Veen, J. J. De Soet, and M. C. Huysmans, “Bacterial composition and red fluorescence of plaque in relation to primary and secondary caries next to composite: an in situ study,” Oral Microbiology and Immunology, vol. 23, no. 1, pp. 7–13, 2008.View at: Publisher Site | Google Scholar
H. Slutzky, E. I. Weiss, I. Lewinstein, S. Slutzky, and S. Matalon, “Surface antibacterial properties of resin and resin-modified dental cements,” Quintessence International, vol. 38, no. 1, pp. 55–61, 2007.View at: Google Scholar
J. C. Meiers and G. A. Miller, “Antibacterial activity of dentin bonding systems, resin-modified glass ionomers, and polyacid-modified composite resins,” Operative Dentistry, vol. 21, no. 6, pp. 257–264, 1996.View at: Google Scholar
T. M. Auschill, E. Hellwig, A. Sculean, N. Hein, and N. B. Arweiler, “Impact of the intraoral location on the rate of biofilm growth,” Clinical Oral Investigations, vol. 8, no. 2, pp. 97–101, 2004.View at: Google Scholar