S. mutans UA159 was acquired from ATCC (Cat. #700610), which had its genome sequenced [14] . This microorganism was grown in Brain Heart Infusion (BHI) (Difco, MD, USA) at 37˚C and atmosphere enriched with 10% CO₂. Chlorhexidine digluconate (20% aqueous solution, Cat. #C9394) was purchased from Sigma-Aldrich (St. Louis, MO, USA).

The MIC and MBC for planktonic cells were determined by overnight standard 96-well plate microdilution method and agar plating [15] . Two-fold dilutions of the drug were prepared in each test well containing Brain Heart Infusion broth (BHI), with CHX concentrations ranging from 0.56 µg/ml to 72 µg/ml. Aliquots of bacterial suspension (20 µL), adjusted spectrophotometrically to match the optical density of A_550nm = 0.5, were added to each 180 µL of drug-containing culture medium. Control wells included culture medium only, culture medium and CHX, and wells without CHX. Plates were closed and incubated for 18 h at 37˚C and atmosphere enriched with 10% CO₂. After incubation, MIC was determined visually and defined as the lowest drug concentration in which growth-turbidity was not observed. Assays were repeated three times in duplicate. Viable counting from control wells and from test wells was performed on BHI agar plates to determine MBC. MBC was defined as the lowest drug concentration in which no colony-forming units were observed on the plate.

Biofilm analyses were formed in 20 mm diameter, 15 mm deep polystyrene multidishes (six wells plates), by incubation of 20 µL of bacterial suspension (A_550nm = 0.5) and 5 mL BHI supplemented with 0.1% sucrose. After 18 h of growth, CHX was added directly to media-containing wells at final concentrations of 50 µg/ml, 100 µg/ml, 500 µg/ml, 800 µg/ml or 1000 µg/ml. Wells which did not receive CHX were set as control. Whole plate was closed and homogenized on an orbital mixer (80 rpm) for 60 s, and incubated for additional 5 min at 37˚C and atmosphere enriched with 10% CO₂. Afterwards, the medium was aspirated and biofilm was washed three times with 5 mL of 0.9% NaCl solution (saline) to remove non-adherent cells. Whole biofilm layer was dislodged with the aid of a cell disrupter (Lifter Cell/3008; Costar, Corning, NY) in presence of 2 mL saline. Part of dislodged biofilm suspension was subjected to expression analysis (below) and part (50 µL) subjected to dilution for viable counting onto BHI agar plates aiming at quantifying MBC in the biofilm (MBC-b). MBC was defined as the lowest drug concentration after exposure that prevented colony forming units growing on agar medium. Assays were performed in duplicate and repeated three times.

The expression levels of all genes were normalized by amplification of 16S rRNA of S. mutans as an internal standard. Overnight S. mutans broth culture was inoculated into 300 mL of pre-warmed BHI and let growth until reaches A_550nm = 0.3. Homogenized culture was then distributed into 20 mL aliquots in new tubes, and exposed to sub-lethal doses of CHX at 1.1 µg/ml, 2.2 µg/ml and 4.5 µg/ml each. Tubes which received the same volume of saline instead of CHX were set as control. All tubes were incubated for 2, 4 and 6 h at 37˚C, 10% CO₂. Aliquots of cultures (50 µL) at each time point were subjected to viable counting onto BHI agar plates. Suspensions were quickly cooled on ice and centrifuged at 5500 g at 4˚C for 4 min. Cells were washed in cold saline and stored with 220 µL of 10 mM Tris, 1 mM EDTA, pH 8.0 (TE) at −80˚C until use.

Biofilms exposed to CHX at 100 µg/ml, 500 µg/ml and 800 µg/ml were selected to analysis of gtf expression. Dislodged cell suspensions were centrifuged, washed in cold saline and stored with 220 µl of TE at −80˚C, until use.

Total S. mutans RNA from either planktonic cells or biofilms was extracted and purified using an RNeasy RNA isolation column (Qiagen-Sciences, MD). Briefly, frozen cells were mechanically disrupted with 0.16 g of 0.1 mm diameter Zirconium Beads (Biospec, OK, USA) on a Mini-bead beater (Biospec), at maximum power (3 cycles of 30 s with 60 s on ice). Further RNA purification, followed by digestion with on-column RNase-free DNase I, was performed as recommended by the manufacturer. To remove small traces of remaining DNA, purified RNA was treated with DNase I (Invitrogen, Life Technologies, Carlsbad, CA, USA) according to the manufacturer’s protocol. RNA sample concentrations were determined at A_260nm (NanoDrop, Thermo Scientific, USA) and integrity by 1.2% formaldehyde agarosegels stained with ethidium bromide [17] .

Reverse transcription of experimental samples together with negative controls was carried out with 1 µg RNA using Superscript III RT (Invitrogen, Life Technologies, Carlsbad, CA, USA) as previously described [16] . After reaction, water was added, so that the product concentration was relative to RNA 10 ng/μl. Controls for cDNA synthesis included samples without RNA template and samples without reverse transcriptase.

Quantification of cDNA was performed using the Step One Plus Real-Time PCR System (Applied Biosystems, Life Technologies, USA) and all primers were constructed by Invitrogen (Invitrogen Life Technologies, Carlsbad, CA, USA) using nucleotide sequence information provided [16] (Table 1). Each reaction mixture (25 µL) contained 1× SYBR Green PCR Master Mix (Applied Biosystems), 0.6 µM of each appropriate forward/ reverse primer and 3 µL of cDNA sample (30 ng relative RNA). Cycling conditions included initial denaturation at 95˚C for 10 min, followed by a 50-cycle amplification, consisting of denaturation at 95˚C for 15 s, annealing at 54˚C for 15 s and extension at 72˚C for 30 s each. Controls included reaction mixtures without cDNA template to rule out primer dimers formation or presence of contaminating DNA. A standard DNA amplification curve starting from 300 ng to 0.003 ng (10-fold dilutions) and a melting-point product curve were obtained for each primer set/run. Assays were performed in duplicate with three or more independent RNA samples.

Relative expression was calculated by Pfaffl mathematical model [17] . Briefly, cycle threshold values of expression levels (C_t) obtained from dilution curves were used to determinate quantitative PCR (qPCR) efficiency for each primer set. Each gtf gene of treated cells was normalized to the 16S rRNA gene (internal control) [16] . These values were then compared to those from the non-treated, 16S-normalized control, to determine the changes in gtf genes expression. Using C_t values, statically significance was obtained at p < 0.01 or p < 0.05, using ANOVA Dunnett Bilateral test, under BioEstat 5.0 free software.

CHX MIC and MBC for planktonic cells were 2.2 µg/ml and 18 µg/ml, respectively, while CHX MBC for biofilm (MBC-b) was 800 µg/ml (data not shown). Thus, sub-micvalues were established in concentrations lower than 2.2 µg/ml and 800 µg/ml for planktonic cells and biofilm, respectively. Planktonic cells exposed to 4.5 µg/ml CHX and 9 µg/ml CHX reduced the bacterial counts by nearly 6 and 20 fold, respectively. In biofilms, a high coefficient of variation (56.4%, p < 0.01) of CFU/ml counts was observed only in 500 µg/ml CHX. In 100 µg/ml and control groups, these coefficients were below 20% (p < 0.01). In 800 µg/ml, there was no cell growth.

The values of Ct 16S rRNA gene did not vary in the experiments with planktonic cells [mean ct 13.24 (±0.48)]. Regarding the CHX exposure, there was 7% variation between exposure and no exposure, whereas taking into account the control and minor concentrations (C1: 1.1 µg/ml and C2: 2.2 µg/ml), this variation was less than 5% (data not shown). Thus, the 16S rRNA gene can be successfully used as a reference gene on our conditions. Moreover, Ct values of 16S rRNA did not change in response to CHX in biofilm-growing cells [mean Ct 13.24 (±0.48)]. Ct values of 16S rRNA showed higher coefficient of variation (6.05%) in the group with the highest CHX concentration (C3: 800 µg/ml) than with other groups (C1: 100 µg/ml and C2: 500 µg/ml; <3.0%).

In planktonic cells, CHX increased gtfC (p < 0.05) and gtfD (p < 0.01) expression only in cell cultures with 6 h growth, at C3 (4.5 µg/ml) concentration, with 9- and 20-fold, respectively (Figure 1). In the biofilm environment, the expression of gtfB and gtfD was reduced in C2 (500 µg/ml) and C3 (800 µg/ml) concentrations, while there was decreased gtfC expression only for C2 concentration, when compared with the control (p < 0.01) (Figure 2).

SEM analyses of planktonic cells and biofilm exposed or not to CHX are shown in Figure 3 and Figure 4.

Several assays were performed for these morphological analyses and the results are representative of three independent experiments. Cell surfaces did not show any obvious change when cells were exposed to increasing CHX concentrations (1.1, 2.2, or 4.5 µg/ml) for 2 h (data not shown). However, as demonstrated by SEM (Fig- ure 3), after 4 or 6 h of exposure at less than-MBC (4.5 µg/ml), several S. mutans cells wilted or their membranes ruptured and, as a result, lost a substantial amount of cytoplasm material. SEM analyses of initial control biofilms and those treated with CHX are shown in Figure 4, where a decrease in cell chain length and matrix were found when biofilms were exposed to different CHX concentrations (1.1 µg/ml to 4.5 µg/ml), while 1200 µg/ml and 2000 µg/ml of CHX caused extensive precipitation of unknown material on the slide (data not shown).

The gene expression of planktonic cells and biofims were observed in the most appropriate experimental growth periods for planktonic cells (2, 4 and 6 h) and biofilms (18 h). Planktonic cells were exposed to the chlorhexidine from the early growth to analyze its growth kinetics. Biofilms were exposed to CHX for 5 min to mimic the oral cavity environment when exposed to CHX during fast mouthwash with the drug.

In this study, methods of RNA extraction from planktonic cells and biofilms were similar, which makes possible the direct comparison of these different conditions of the same bacterial microorganism. The polysaccharides elimination of the biofilm, as a prior step to extraction [21] [22] , was not necessary in our in vitro conditions due to the shorter time of biofilm formation and low concentration of sucrose.

The 16S rRNA gene has been successfully used as a reference gene in studies on S. mutans in biofilm and planktonic cells [16] [20] [23] -[26] . Our results indicated that the expression of 16S rRNA S. mutans remained stable when exposed to a CHX environment regardless of whether it is in planktonic cells or in biofilms. This suggests that the 16S rRNA gene can be used as a reference gene on analysis of expression in the two conditions described above.

The effect of antimicrobial agents on gene expressions that are heavily involved in biofilm formation is of particular interest to dentistry [24] [25] . Different patterns of gene expression between biofilms and planktonic cells were observed [20] . Thus, Li and Burne [27] found that the expression of gtf genes was enhanced in bacteria growing as biofilms compared with planktonic growth, implying thata cell-density-dependent intercellular communication system may play a role of regulating thegtfB/C genes expression [28] .

According to our findings, CHX inhibited the gtfs expression of S. mutans organized as biofilm (Figure 2) but increased the gtfs expression in planktonic cells (Figure 1). Koo et al. [25] demonstrated that apigenin significantly decreased the expression of gtfB and gtfC and increased the expression of gtfD in S. mutans growing in the planktonic state. A similar profile of gtf expression was obtained with biofilms.

Tam et al. [24] observed that iodine and povidone iodine (PI) in a tetraglycol-carrier cause enhancement of expression of gtfB in S. mutans in biofilms but not in planktonic bacteria. PI in water induced expression of gtfB and gtfC in planktonic bacteria. Thus, the vehicle type used together with the active compound or molecule can determine the increased or decreased expression of these genes.

It is probable that planktonic cells express more gtfs (Figure 1) as a defense mechanism against stress when exposed to CHX. Cells from biofilms are more protected, and they do not need to produce more gtfs. In biofilms, CHX inhibited the gtfs expression (Figure 2). This is a clear indication that CHX at sub-lethal concentrations may have an adverse effect on the adhesion process of these cells on the tooth surface, which is relevant for clinical purposes ofcontrolling oral biofilms’ development in vivo.

According to Wade [29] , high CHX concentrations, as used in clinical practice, kill the cells, and this is not interesting for the balance of microbiota in the biofilm. Successful antimicrobial agents are able to maintain the oral biofilm at levels compatible with oral health but without disrupting the natural and beneficial properties of the resident oral microflora [30] .

Thus, in biofilms, low CHX concentrations may have anticariogenic effect by inhibiting the expression of enzymes associated with virulence of S. mutans. Furthermore, CHX when used in concentrations of 0.12% and 0.2% for a long time has many side effects such as teeth pigmentation and taste changes [31] .

Radford et al. [32] demonstrated that biofilm formation in vivo can be inhibited using sub-MIC concentrations of CHX. It is possible that the reduction in the formation of the biofilm matrix exposed to low CHX concentrations (Figure 4) may be related to the affinity of CHX with proteins, as well as the ability to reduce the expression of some gtfs (Figure 2 and Figure 4).

According to Dunne [33] , CHX only kills outer layers of biofilm bacteria, leaving the healthy microorganisms in the inner layers and allowing the growth of these bacteria later when CHX is not present. It is also noteworthy to understand that possible conformational changes in biofilm structure (e.g., water channels) may contribute to the diffusion of CHX into deeper biofilm layers [12] . This variable might explain the unexpected difference in gtf expression in planktonic and biofilm forms because viable cells might be protected by the low diffusion rate of the agent in the biofilm matrix.

In this study, planktonic cells, when exposed to higher CHX concentrations, suffered membrane rupture (Figure 3). Tattawasart et al. [13] found similar results on the CHX effect on Pseudomonas stutzeri species sensitive to the drug. Vitkov et al. [34] , using an ex vivo model, showed that some bacterial cells from biofilm sustained ruptured membranes when exposed to CHX. This observation can be explained by the fact that CHX, which is positively charged, binds tightly to negatively charged bacteria membrane, causing its disruption [35] . Under such microbial stress, it is expected that enzyme expression can be altered as observed at concentration 4.5 µg/ml (Figure 1).