Farnesol inhibits development of caries by augmenting oxygen sensitivity and suppressing virulence-associated gene expression inStreptococcus mutans

Streptococcus mutans is a primary etiological agent of dental caries. Farnesol, as a potential antimicrobial agent, inhibits the development ofS. mutans biofilm. In this study, we hypothesized that farnesol inhibits caries development in vitro and interferes with biofilm formation by regulating virulence-associated gene expression. The inhibitory effects of farnesol to S. mutans biofilms on enamel surfaces were investigated by determining micro-hardness and calcium measurements. Additionally, the morphological changes ofS. mutans biofilms were compared using field emission scanning electron microscopy and confocal laser scanning microscopy, and the vitality and oxygen sensitivity ofS. mutans biofilms were compared using MTT assays. To investigate the molecular mechanisms of farnesol’s effects, expressions of possible target genesluxS, brpA, ffh, recA, nth, and smx were analyzed using reverse-transcription polymerase chain reaction (PCR) and quantitative PCR. Farnesol-treated groups exhibited significantly higher micro-hardness on the enamel surface and lower calcium concentration of the supernatants as compared to the-untreated control. Microscopy revealed that a thinner film with less extracellular matrix formed in the farnesol-treated groups. As compared to the-untreated control, farnesol inhibited biofilm formation by 26.4% with 500 µmol/L and by 37.1% with 1,000 µmol/L (P<0.05). Last, decreased transcription levels of luxS, brpA, ffh, recA, nth, and smx genes were expressed in farnesol-treated biofilms. In vitrofarnesol inhibits caries development and S. mutans biofilm formation. The regulation of luxS, brpA, ffh, recA, nth, and smx genes may contribute to the inhibitory effects of farnesol.


Introduction
Dental caries, a biofilm-related oral infectious disease, continues to be a significant public health problem in many parts of the world. Streptococcus mutans is an important element of the dental plaque biofilm, and plays a crucial role in the development of dental caries [1] . S. mutans in biofilms is physically and functionally distinct from planktonic cells and displays an enhanced resistance to antibiotics [2] . Within biofilms, S. mutans metabolizes many sugars and produces various organic acids which subsequently lowers the pH of the surrounding environment and eventually leads to the dissolution or de-mineralization of enamel over time [3] . Furthermore, S. mutans synthesizes extracellular polysaccharides (EPS) from dietary sucrose. EPS enhances the virulence of dental biofilms by promoting the attachment and accumulation of cariogenic microorganisms or by maintaining the bulk and structural stability of the biofilm. Therefore, developing antibacterial agents with the capacities to weaken acid generation and to decrease the EPS content of S. mutans biofilms will help to prevent dental caries.
The investigation of natural products and their derivatives has been one of the most attractive strategies for developing new medicines [4] . Farnesol, a sesquiterpene alcohol (3,7,11-trimethyl-2,6,10-dodecatrien-1-ol) and the first quorum-sensing molecule discovered in eukaryotes, can be found in some essential oils [5] , propolis (a resinous beehive product), and citrus fruits [6] . Studies have shown that this molecule has inhibitory effects against such microbes as Paracoccidioides brasiliensis [7] , Fusarium graminearum [8] , Staphylococcus epidermidis [9] , and Burkholderia pseudomallei [10] . Topical application of farnesol could interrupt the accumulation of biofilms, reduce the polysaccharide content of biofilms, and weaken the ability of S. mutans to generate and tolerate acid. To date, very few studies have evaluated caries development in vitro, oxygen sensitivity and morphological changes of S. mutans biofilms exposed to farnesol, and little is known about the associated molecular mechanisms [11] .
The luxS, brpA, ffh, recA, nth, and smx genes have been found to be related to S. mutans biofilm virulence. The luxS gene has a primary role in regulating autoinducer-2 production, which is an important quorum-sensing molecule in bacteria, and is able to elicit a specific response across species barriers [12] . The brpA gene encodes a predicted surface-associated protein associated with biofilm formation, autolysis, and cell division [13] . The ffh gene encodes a homolog of the 54 kDa subunit of a signal recognition particle involved in resistance to acid stress [14] . The recA gene influences glucosyltransferase (GTF) expression, and GTFs are essential for cellular adhesion [15] . The nth gene encodes a predicted EndoIII-related endonuclease involved in DNA replication, recombination, and repair [16] . The smx gene encodes a class II-like AP endonuclease that regulates the exonucleolytic activity and the acid-adaptive response of an organism [17] . In brief, luxS and brpA are involved in the regulation of S. mutans biofilm formation associated with EPS-produc-tion and responses to oxidative stress, while ffh, recA, nth, and smx play a role in the acid tolerance of S. mutans [12][13][14][15][16][17][18] .
In this study, we hypothesized that farnesol, a chemical compound with antimicrobial properties, may inhibit the development of caries, and interfere with the formation of S. mutans biofilms by regulating virulence-associated gene expression. To test this, the inhibitory effects of farnesol to S. mutans biofilms on enamel surfaces were investigated. Furthermore, the morphological changes and vitality of the biofilm were compared among farnesol-treated and untreated control groups. Additionally, the virulence-associated genes underlying the inhibitory effects of farnesol were assessed.

Materials and methods
Bacterial strains and growth conditions S. mutans UA159 was kindly provided by the Institute of Microbiology, Chinese Academy of Sciences. This strain is a proven virulent cariogenic dental pathogen with a known genomic sequence [19] .

Test compounds
Stock solutions (100 mmol/L) of farnesol (E,Efarnesol; Sigma) were dissolved in 100% (vol/vol) methanol and frozen at -70°C until use [20] . Farnesol was diluted to a working concentration in BHI broth for all experiments. Three concentrations, 250, 500, and 1,000 µmol/L, of farnesol were chosen for the MTT [3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide] (KeyGen Bio Co., Nanjing, China) assay of biofilm formation, based on a previous study [11] . The same series of concentrations were also used in determining micro-hardness of the enamel blocks and measuring calcium in the supernatants while 1,000 µmol/L farnesol was chosen for morphological obser-vations, the oxygen sensitivity assay and transcription level study.
The experiments were performed mainly in two groups: farnesol-treated and untreated control. The farnesol-treated group was S. mutans biofilm treated with farnesol diluted in BHI broth. The untreated control groups were S. mutans biofilm treated with an equal volume of deionized water, instead of farnesol, in BHI solution.

Biofilm preparation and farnesol treatments
Biofilms of S. mutans were formed on the surface of enamel blocks or flat-bottomed culture flasks in BHI broth containing 1% (w/v) sucrose [21] . Briefly, the standardized suspension was dispensed into flat-bottom plates, and the organisms were grown undisturbed for 24 hours to allow initial biofilm formation. After a 24hour-incubation of the suspension, non-adherent cells were removed by thoroughly washing the biofilms three times with PBS (pH 7). At this point, the culture medium was replenished, and equal volumes of farnesol or deionized water were added to the farnesol-treated and untreated control groups, respectively. For this treatment, the biofilms were exposed to farnesol or deionized water for 1 minute, rinsed three times in a sterile saline solution, and transferred to fresh culture medium. The medium was replaced twice daily, which means that biofilms in each group were treated a total of five times, at 24, 36, 48, 60, and 72 hours after the cells were added to the flat-bottom plates. All experiments were performed in triplicate on three separate occasions.
Cells from the same batch were used for the farnesoltreated and untreated control groups for the initial adherence step.

In vitro caries development assays
The study and use of patients' teeth for this study were approved by the Ethics Committee of Dentistry Hospital, Nanjing Medical University (Institutional Review Board) (IRB; approval number PJ2011-019-03). The isolated teeth without caries or dysplasia were collected from the patients whose teeth had to be extracted for orthodontic reasons. Before the surgical extraction of teeth, the participants were informed of the objective of this study and the uses of the teeth in vitro experiments. All study participants provided written informed consent. After extraction, the teeth were washed with normal saline and stored at 4°C for 1 week. Stereomicroscopic observation showed that the enamel on the teeth was intact without white spots. Enamel blocks (6 mm Â 6 mm Â 2 mm in length, width, and thickness) were cut from the surface coronal region of premolars perpendicular to the tubule direction using low-speed diamond slicers. Then, the enamel blocks were prepared and sterilized by gamma radiation as previously described [22] . Subsequently, the blocks were covered with nail varnish, with only a 4 mm Â 4 mm window exposed to bacterial culture. The enamel blocks were rehydrated in BHI overnight before each in vitro caries assay.
The 40 blocks were randomly and blindly divided equally into five groups: three farnesol-treated S. mutans biofilms (250, 500, and 1,000 µmol/L farnesol), one untreated control with S. mutans biofilm, and one untreated control without S. mutans.
The enamel blocks were placed into 12-well flatbottom tissue culture plates (Cellstar Greiner Bio-One, Frickenhausen, Germany), and 1,000 µL of standardized S. mutans suspension or BHI broth without S. mutans was dispensed onto them. Biofilm preparation and farnesol treatments were the same as described above. The biofilms were then treated with deionized water or different concentrations of farnesol twice daily, followed by medium replenishment, until the end of the experiment (at 72 hours). Meanwhile, the supernatants from each well were collected for calcium content measurements.
The enamel blocks were gently removed from the wells, washed twice with PBS, and then put into sterile 2 mL snap-top vials containing 0.5 mL of PBS. The vials were placed in an ultrasonic water bath and sonicated for 15 minutes to remove the biofilm. The blocks were then transferred to fresh 2 mL tubes, dried at 37°C, and stored at 4°C.
The micro-hardness of the enamel blocks was measured by the Micro-Hardness Tester (DHV-100, Shanghai Technology Co., Shanghai, China) before and after the experiments. The blocks were fixed on metal discs using a very thin layer of cyanoacrylate (MDS Adhesive QX-4; MDS Products Inc., Laguna Hills, CA) for micro-hardness analyses. Each enamel block was statically loaded with 100 g for 15 seconds with a Vickers diamond pyramid indenter. Then, the diagonal length of the formed indentation was measured, and the Vickers micro-hardness values (Hv) were converted to micro-hardness values [23] . Six indentations were measured in the window and the average of these measurements was used [24] . Each experiment was repeated three times.
The mineral loss of the enamel blocks was analyzed by comparing the calcium concentrations in the collected supernatants between the farnesol-treated and untreated control groups. The calcium ions atomized by flame were in ground state indicated by selective absorption on the spectrum. The calcium concentration was determined using an AA6300 atomic absorption spectrophotometer (Shimadzu, Japan) at a wavelength of 422.7 nm [25] . Concentrations were determined from a standard curve after calibrating the instrument with known concentration standards. Under the constant condition, the relationship between absorbance and concentration of calcium ion was applied to the Beer-Lambert law: A = KC (A is the measured absorbance, C is the analytic concentration, and K is a constant). The assays were performed in triplicate.

Comparison of biofilm organization by FE-SEM and CLSM
Biofilms in the farnesol-treated and untreated control groups were formed on the surface of glass-bottomed chamber slides (Nunc, Roskilde, Demark). Biofilm preparation and farnesol treatments were the same as described above. The medium was aspirated after a 72hour-incubation at 37°C, and biofilms were then washed three times with PBS (pH 7) and placed in 2.5% [vol/ vol] glutaraldehyde overnight at 4°C. All experiments were performed in triplicate on three separate occasions.
For confocal laser scanning microscopy (CLSM) observation, the samples were washed with distilled water twice, and stained with L-13152 LIVE/DEAD BacLight™ bacterial cell stain containing 6 µmol/L SYTO-9 dye and 30 µmol/L propidium iodide (Molecular Probes Inc., Eugene, OR). Bacterial cell stain (100 µL) was added into 1 mL medium for 15 minutes at 37°C in the dark. After staining, biofilms were examined using a Zeiss LSM710 confocal laser scanning microscope (Carl Zeiss, Inc., Oberkochen, Germany) with 495 nm argon ion lasers. Biofilms were observed using a 40 Â water-immersion objective lens (40 Â /1.2 w). The thickness of the biofilms was measured from the top to the bottom of the biofilm, and the maximum thickness of each region was calculated using the Z-stack of Zeiss ZEN software (Carl Zeiss, Inc., Oberkochen, Germany), which is a scanning image parameter. Each sample was scanned at intervals of 1 µm, and images were acquired randomly from at least five different regions of each surface analyzed. The regions were from the upper left, upper right, lower left, lower right, and middle fields.

Measurement of S. mutans biofilm formation
The wells of 96-well, flat-bottom microtiter plates (Cellstar Greiner Bio-One, Frickenhausen, Germany) were filled with 100 µL of standardized S. mutans suspension. Biofilm preparation and farnesol treatments were the same as described above. For the biofilm formation assay, the biofilms were treated with different concentrations of farnesol (0, 250, 500, or 1,000 µmol/ L) followed by the replenishment of their medium twice daily until the end of the experiment. Farnesol-untreated control biofilms were treated with deionized water instead of farnesol. After 72 hours, the medium was removed and the wells were carefully washed three times with PBS (pH 7).
The biofilm survival rate of the farnesol-treated and untreated control groups were compared using an MTT (KeyGen Bio Co., Nanjing, China) assay [27] . Briefly, 100 µL of MTT solution (2 mg/mL of MTT in sterile PBS) was placed in each well and the plates were incubated at 37°C for 3 hours in a dark place. The MTT solution was then gently removed, and intracellular formazan crystals were dissolved by adding 150 µL of dimethyl sulfoxide (Ameresco, OH, USA). Colorimetric analysis using a microtiter plate reader (Bio Tek Instruments Inc., VT, USA) at wavelength of 490 nm indicated the changes in metabolic activity of the biofilms.

Measurement of S. mutans biofilm oxygen sensitivity
The wells of 96-well, flat-bottom microtiter plates were filled with 100 µL of standardized S. mutans suspension. Biofilm preparation and farnesol treatments were the same as described above. Before the oxygen sensitivity assay, the biofilms were treated with 1,000 µmol/L farnesol twice daily. After the biofilm had been growing for 72 hours, hydrogen peroxide was added to the biofilms in a final concentration of 58.8 mmol/L (0.2%, v/v) for 30, 60, or 120 minutes. The biofilm survival rate of the farnesol-treated and untreated control groups were compared using an MTT assay, as described above [27] .

Measurement of the expression levels of genes potentially targeted by farnesol
Total RNA was isolated from the 72-hour-old biofilms from the farnesol-treated and untreated control groups using a modified one-step method [28] . After centrifugation, the cells were washed three times using PBS (pH 7) and re-suspended in 200 µL of TE buffer (10 mmol/L Tris HCL, 1 mmol/L EDTA; pH 8) with lysozyme (5 g/L) (Sigma, Germany) on ice for 30 minutes. Then, 20 µL 10% SDS was added to the samples and they were boiled for 1 minute followed by the addition of 100 µL of TRIzol reagent (Invitrogen) for 5 minutes. After the samples were placed on ice for 2 minutes, 100 µL of chloroform was added and the tubes were vortexed for 20 seconds before being centrifuged again. RNA was finally precipitated from the aqueous phase with isopropanol, and the resulting pellets were dried and suspended in 20 µL of diethylpyrocarbonatetreated water.
The RNA samples were treated with DNase (Takara), and reverse transcription was performed on 5 µg of total RNA using avian myeloblastosis virus reverse transcriptase XL (Takara) following the manufacturer's instructions. Primers ( Table 1) were all designed by Shanghai Generay Bio-Tech Co., Ltd., using 16S RNA as a reference. The synthetic cDNA described above was used for q-PCR analysis, which was performed on an ABI 7300 Fast real-time PCR machine (Applied Biosystems, Rotkreuz, Switzerland), using FastStart Universal SYBR green master mix (Roche, Applied Science, Penzberg, Germany). Amplification was achieved using the following cycle settings: 10 minutes at 95°C, followed by 40 cycles of 95°C for 15 seconds, 60°C for 60 seconds, and 95°C for 15 seconds. A melting curve was analyzed via amplification to ensure the absence of primer dimers. Expression of each gene was calculated using the 2 -ΔΔCt method [29] , with 16S rRNA used as a reference.

Statistical analysis
All quantitative experiments were done in triplicate for statistical analyses. The data were analyzed using analyses of variance (ANOVAs), and paired t-test was performed for intra-group comparisons. When significant differences were detected, pairwise comparisons among all the groups were calculated. Data were analyzed using the SPSS17.0 software (SPSS Inc., Chicago, IL, USA). Comparisons resulting in P values of less than 0.05 were considered to be statistically significant. In the q-PCR analysis, genes were considered to be differentially expressed if the values for 2 -ΔΔCt were upregulated at least 2.0-fold compared with the control group in three independent experiments.

Farnesol inhibits caries development in vitro
The hardness analysis results showed that farnesol inhibited enamel demineralization caused by S. mutans. Before the inoculation of S. mutans, there were no significant differences in micro-hardness among these enamel blocks in the five study groups (P > 0.01). After the inoculation, however, the micro-hardness of the enamel surface was decreased in the blocks covered with farnesol-untreated S. mutans biofilms (P < 0.01). Additionally, blocks from the three farnesol-treated S. mutans biofilm groups had significantly higher enamel surface micro-hardness than those from the-untreated control biofilms, respectively (P < 0.05). Furthermore, a significant difference was also observed in the enamel surface micro-hardness between the group treated with 1,000 µmol/L farnesol and the other two groups treated with the lower concentrations of farnesol, respectively (P < 0.05) ( Table 2).
As another measure of the effects of farnesol treatment 16S rRNA-R GTATCTAATCCTGTTCGCTACCCAC on caries development, the calcium levels in the supernatant from the culture media of the enamel blocks described above were examined. The calcium analysis results showed that the calcium concentrations in all the supernatants were elevated at 24, 48, and 72 hours. After 24 hours of culture in medium, the calcium concentrations of both the farnesol-treated and untreated control groups reached a steady-state ( Table 3). The calcium concentrations in the supernatants were lower in the group treated with 1,000 µmol/L farnesol than those in the groups treated with 250 or 500 µmol/L farnesol or in the untreated control group, respectively (P < 0.05).

Different morphological changes in S.mutans biofilm
The morphological changes of the farnesol-treated and-untreated control biofilms were characterized by microscopy ( Fig. 1). At lower magnification, FE-SEM showed that the biofilms formed in the farnesoluntreated control group were bulky and appeared to be structurally stable (Fig. 1A). On the other hand, the biofilms formed in the farnesol-treated group were less voluminous and their cells aggregated into clusters (Fig.  1B). At higher magnification, FE-SEM showed that there was less extracellular matrix in the farnesol-treated biofilms than in the-untreated control biofilms. Additionally, although the S. mutans still kept their primary shapes in the farnesol-treated biofilms, the proportion of long-chained shaped bacteria was decreased inside those biofilms as compared to that in the-untreated control biofilms (Fig. 1C-D).
The morphological changes in S. mutans biofilms between the farnesol-treated group and the untreated control groups were also analyzed using CLSM (Fig.  2). The biofilms in the farnesol-untreated control group were continuous and uniform. Inside the farnesoluntreated control biofilms, S. mutans appeared to be distributed evenly with small aggregates (Fig. 2A). After treatment with farnesol, biofilms appeared to be discontinuous and dispersed. The farnesol-treated S. mutans aggregated into large clumps (Fig. 2B). The thickness of biofilm in the farnesol-treated groups was thinner than that in the untreated control. These results suggest that farnesol contributes to the accumulation of S. mutans biofilms, while reduces the thickness of these biofilms.

Farnesol inhibits the formation of S. mutans biofilms
The effects of farnesol on biofilm formation were analyzed in our study using three different farnesol concentrations (Fig. 3). The S. mutans biofilm formation was significantly inhibited in the groups treated with 500 or 1,000 µmol/L farnesol as compared with that in the groups treated with only 250 µmol/L farnesol or in those that were untreated controls (P < 0.01). Additionally, compared with the values in the untreated control group, the OD 490 values of S. mutans biofilms were decreased by 26.4% and 37.1% in the groups treated with 500 and 1,000 µmol/L farnesol, respectively. The results suggest that the inhibitory effect of farnesol on S. mutans biofilm formation is related to its concentration.

Farnesol treatment increases oxygen sensitivity of S. mutans biofilms
After 30, 60, and 120 minutes of incubation in 0.2% hydrogen peroxide, the oxygen sensitivity of farnesoltreated and untreated control S. mutans biofilms (Fig. 4) was examined to determine whether this quality is affected by farnesol treatment. During 0 to 30 minutes of incubation in 0.2% hydrogen peroxide, the survival rate of S. mutans in the farnesol-treated group was lower than that in the-untreated control group (Fig. 4). After the biofilms were treated with 0.2% hydrogen peroxide, the survival rate of S. mutans in the farnesol-treated group remained stable for 30, 60, and 120 minutes. Additionally, in contrast to the biofilms exposed to farnesol, the survival rate of S. mutans in the untreated control group decreased significantly after 30 minutes of incubation in 0.2% hydrogen peroxide, and The micro-hardness of enamel surfaces covered BHI broth alone or with farnesol-treated or vehicle control S. mutans biofilms was evaluated. SMH, Enamel surface microhardness; SMH1, S. mutans + 250 mmol/L farnesol; SMH2, S. mutans + 500 mmol/L farnesol; SMH3, S. mutans + 1,000 mmol/L farnesol; SMH4, S. mutans + 0 mmol/L farnesol; SMH5, no S. mutans + 0 mmol/L farnesol. The variation between the micro-hardness before and after treatments was analyzed using ANOVAs and paired t-test was performed for intra-group comparisons: *P < 0.05, **P < 0.01. decreased even further after 60 minutes of incubation in 0.2% hydrogen peroxide. The OD 490 value of the biofilms in the untreated control group was approximately 1/3 lower after 60 min of incubation in 0.2% hydrogen peroxide than that in the initial biofilm (0 minute). In the last, the OD 490 value in the farnesoluntreated control group was close to that in the farnesoltreated group after 120 minutes of incubation in 0.2% hydrogen peroxide.

Gene expression in S. mutans biofilms
To determine if farnesol treatment affected the expression of S. mutans virulence-associated genes, the levels of luxS, brpA, ffh, recA, nth, and smx expression in farnesol-treated and -untreated control biofilms were investigated. The q-PCR results demonstrated that the transcription levels of luxS, brpA, ffh, recA, nth, and smx were all downregulated in the farnesol-treated biofilms, as compared with that in the untreated control biofilms (Fig. 5A). The q-PCR assay showed that the transcription levels of luxS, brpA, ffh, recA, nth, and smx were reduced approximately 1-fold in the farnesol-treated biofilms as compared to that in the untreated control biofilms (Fig. 5B, P < 0.05).

Discussion
Dental caries, which continues to affect the majority of the world's population, is attributed to a disturbance in the equilibrium between tooth substance and microbial deposit, and results in a loss of tooth mineral by the action of bacterial acidic metabolites. Formation of microbial biofilms on the tooth surfaces is the initial dynamic process of dental caries, thus making the control of dental biofilms essential in the prevention of carious lesions [30][31] .
Investigating natural products may provide innovative therapeutic agents to treat dental caries. Previous studies have reported that certain concentrations of farnesol affect the growth, metabolism, and polysaccharide contents of S. mutans biofilms [32] . However, to our knowledge, no data are available about the exact mechanism by which farnesol affects S. mutans biofilms [11] , nor have there been any study about the inhibitory effects of farnesol to S. mutans biofilms on enamel surface, and the morphological changes or oxygen sensitivity of S. mutans biofilms exposed to farnesol.
A few studies have shown that surface microhardness is one of the key physical properties of enamel that can access the effect of chemical agents on hard teeth [26,33] . The demineralization process caused by S. mutans biofilms decreases enamel surface micro-hardness, and eventually leads to caries lesion. Previous studies have confirmed that micro-hardness evaluation (micro indentation hardness testing) is sensitively for accessing the demineralization levels of an enamel surface [33] . In this study, we took the lower premolars from the similarly aged patients and chose enamel blocks with similar micro-hardness for comparison. Our results showed that the reduction of surface enamel micro-hardness by S. mutans was reduced by farnesol treatment, suggesting that farnesol prevent S. mutansrelated caries development and can be used as a potential agents against S. mutans [34] .
In this study, we also used FE-SEM and CLSM to investigate whether any morphological differences could be observed between the farnesol-treated anduntreated control biofilms. It is well known that the extracellular matrix, including the polysaccharide matrix, can play a critical role in affecting the virulence of plaque, and influence the physical and biochemical properties of biofilms [35] . We found that the farnesoltreated biofilms appeared to be looser, thinner, and have less extracellular matrix than the-untreated control biofilms. The FE-SEM and CLSM results from our study further confirmed that farnesol reduce the The supernatant of biofilms from the vehicle control and farnesol-treated groups were collected and their calcium concentrations were tested. The data were analyzed using a two-way ANOVA with interaction. Multiple comparisons within a certain group were performed using S-N-K method. *within a certain group, the 48 hours and 72 hours calcium concentrations were significantly different from the calcium concentration at 24 hours, P < 0.05; #within a certain group, the 24 hours and 72 hours calcium concentrations were significantly different from the calcium concentration at 48 hours, P < 0.05. The calcium concentrations in the supernatants were lower in the C3 than those in the C1, C2, or C4, respectively (P < 0.05). C, concentration of the supernatant; C1, S. mutans + 250 mmol/L farnesol; C2, S. mutans + 500 mmol/L farnesol; C3, S. mutans + 1,000 mmol/L farnesol; C4, S. mutans + 0 mmol/L farnesol; C5, no S. mutans + 0 mmol/L farnesol. pathogenicity of S. mutans by inhibiting biofilm formation and the associated extracellular matrix production.
Previous studies have reported that the decreasing of long-chained cells is the initial morphological change in the cell apoptosis process, and this alternation is also one of the acid tolerance mechanisms of S. mutans [36] .
Our microscopy results showed that there are fewer long-chained cells in the farnesol-treated biofilms than that in the-untreated control biofilms. The results suggest that the reduction of long-chained cells in the farnesol-treated biofilms could be a damage mechanism of farnesol against S. mutans biofilm.
In this study, we found that the transcription levels of luxS and brpA were downregulated in the farnesoltreated biofilms. The luxS and brpA genes are known for their roles in the regulation of biofilm formation and in the stress tolerance of S. mutans. Mutant strains lacking brpA have a limited ability to grow and accumulate on surfaces, while displaying increased sensitivity to acid [16,37] . Mutant strains lacking luxS are affected in nearly every aspect of their cellular physiology and  associated virulence, including carbohydrate and central metabolism, nucleic acid metabolism and DNA repair, amino acid and protein synthesis, stress responses, fatty acid synthesis, and other cellular processes [18] . Our findings suggest that farnesol inhibit the development of caries by downregulating the transcription levels of luxS and brpA of S. mutans.
It was reported that the biofilm formed by a luxS mutant strain adopted a notably more granular appearance, in contrast to the smooth and confluent layer formed by wild type S. mutans [38] . Accordingly, we found that when compared with the-untreated control biofilms, these farnesol-treated biofilms had downregulated levels of luxS compared to the-untreated control biofilms and appeared to be less voluminous, with their cells aggregated into clusters. The result suggested that farnesol change the smooth layer to granular appearance by regulating luxS expression.
The low-pH survival strategy of S. mutans is to induce a DNA repair system to protect or repair DNA damage caused by the harmful effects of intracellular acidification [36][37] . Some proteins such as RecA, Nth, Smx, and Ffh are involved in this process. RecA is thought to be a recombinase protein, Nth an endonuclease, and Smx an apurinic-apyrimidinic (AP) endonuclease. The S. mutans ffh mutant strains exhibit a defect in the H + /ATPase activity at acidic pH values, while H + /ATPase protects wild-type S. mutans against environmental acid stress by regulating pH homeostasis [14] . In our study, we found that the transcription Fig. 4 Effects of farnesol on oxygen sensitivity of S. mutans biofilms. Mature biofilms from the farnesol-untreated control group and the farnesol-treated group were collected before (0 minute) and after (30, 60, 120 minutes) treatment with 0.2% (v/v) hydrogen peroxide. Viable cells from the farnesol-treated biofilms (•) and untreated control biofilms (l) were assayed using an MTT assay at different phases (0, 30, 60, and 120 minutes). The data represent meanAEstandard error.

Fig. 5
Gene expression in farnesol-treated and untreated control S. mutans biofilms. The transcription levels of luxS, brpA, ffh, recA, nth, and smx were measured in the farnesol-treated and-untreated control groups. A: Expressions of luxS, brpA, ffh, recA, nth, and smx as determined by electrophoretic images of PCR products. C, farnesol-untreated control group; F, farnesol-treated group. B: Expressions of luxS, brpA, ffh, recA, nth, and smx as determined by q-PCR. *2-DDCt < 0.5 and P < 0.05. Fig. 3 Effects of farnesol on S. mutans biofilm formation. Cell densities of farnesol-treated and untreated control S. mutans biofilms were assayed by measuring the optical density at a wavelength of 490 nm (OD 490 ). The data were analyzed using a one-way ANOVA. The data represent meanAEstandard error. Multiple comparisons between experimental groups and the control group were performed using Dunnett method. **P < 0.01; ***P < 0.001. levels of ffh, recA, nth, and smx were all downregulated in the farnesol-treated biofilms compared with the levels in control biofilms, contributing to the inhibitory effect of farnesol against S. mutans biofilms.
Our results also found that farnesol inhibits biofilm formation by decreasing the synthesis of extracellular matrix and by decreasing the proportion of long chain cells, which result in a lower tolerance to acidic stress. These consequences may stem from the farnesolinduced downregulation of luxS, ffh, recA, nth, and smx, which are involved in the regulation of acid tolerance in S. mutans [16,18,[36][37] .
In conclusion, this study shows that farnesol inhibits caries development and S. mutans biofilm formation by regulating the expression of virulence-related genes. The results suggest that using farnesol as a rinsing agent may prevent the biofilm infection known to cause dental caries. However, further studies will be needed to test its clinical effectiveness.