Synergistic mechanism of Ag+–Zn2+ in anti-bacterial activity against Enterococcus faecalis and its application against dentin infection

Background Ag+ and Zn2+ have already been used in combinations to obtain both enhanced antibacterial effect and low cytotoxicity. Despite this, it is still unclear how the Zn2+ co-works with Ag+ in the synergistic antibacterial activity. The main purposes of this study were to investigate the co-work pattern and optimum ratio between Ag+ and Zn2+ in their synergistic antibacterial activity against E. faecalis, the possible mechanisms behind this synergy and the primary application of optimum Ag+–Zn2+ co-work pattern against the E. faecalis biofilm on dentin. A serial of Ag+–Zn2+ atomic combination ratios were tested on both planktonic and biofilm-resident E. faecalis on dentin, their antibacterial efficiency was calculated and optimum ratio determined. And the cytotoxicity of various Ag+–Zn2+ atomic ratios was tested on MC3T3-E1 Cells. The role of Zn2+ in Ag+–Zn2+co-work was evaluated using a Zn2+ pretreatment study and membrane potential—permeability measurement. Results The results showed that the synergistically promoted antibacterial effect of Ag+–Zn2+ combinations was Zn2+ amount-dependent with the 1:9 and 1:12 Ag+–Zn2+ atomic ratios showing the most powerful ability against both planktonic and biofilm-resident E. faecalis. This co-work could likely be attributed to the depolarization of E. faecalis cell membrane by the addition of Zn2+. The cytotoxicity of the Ag+–Zn2+ atomic ratios of 1:9 and 1:12 was much lower than 2% chlorhexidine. Conclusions The Ag+–Zn2+ atomic ratios of 1:9 and 1:12 demonstrated similar strong ability against E. faecalis biofilm on dentin but much lower cytotoxicity than 2% chlorhexidine. New medications containing optimum Ag+–Zn2+ atomic ratios higher than 1:6, such as 1:9 or 1:12, could be developed against E. faecalis infection in root canals of teeth or any other parts of human body.


Background
Infectious diseases pose a major threat upon the health and well being of humans [1,2]. The discovery and development of antibiotics in the last century has dramatically changed the outcome of treatment for infectious diseases. However, the development of bacterial resistance against even the most powerful antibiotics is rapidly becoming a major challenge for the application of antibiotics in combating infectious diseases [3][4][5]. Because of this concern, stringent standards for the prescription of antibiotics have been established to minimize the development of bacteria resistance, particularly in the hospital environment [5,6].
Antibacterial metal ions, such as silver ion (Ag + ) and zinc ion (Zn 2+ ), possess broad-spectrum antibacterial activities with no extensive development of bacterial resistance [7]. The silver ion has the most potent bactericidal activity among metal ions and has been used to treat bacterial infections for centuries [8][9][10]. However, the high cytotoxicity and discoloring potential of Ag + severely restricts its broad application in clinical medicine [11,12]. To overcome this shortcoming, many researchers attempted to use Ag + with other metal ions, in most cases the Zn 2+ , to achieve strong synergistic antibacterial effect and reduction in cytotoxicity [13,14]. This is because Zn 2+ is less toxic but also less bactericidal when compared with Ag + [10,15]. This effort of synergistic use of Ag + and Zn 2+ for infection control has been proven effective and efficient in biomedical, medical and dental studies [14,16,17]. Nevertheless, the mechanism by which Ag + and Zn 2+ interacts synergistically with each other is not clearly understood. Moreover, the optimum ratio of Ag + and Zn 2+ to achieve maximal synergistic antibacterial effect has not been established.
Enterococcus faecalis (E. faecalis) is a gram-positive facultative anaerobe that can cause refractory infections in the digestive system, urinary tract as well as root canals of human teeth [18][19][20]. Detection of E. faecalis in persistent infection of the root canals and around the root apex of teeth ranged from 24 to 77% [20], and is believed to be a major cause for the failure of root canal treatment [20,21]. Enterococcus faecalis develops strong resistance to extremely harsh environments including highly-alkaline pH, nutritional deficiency and many current clinical intra-canal medications, such as calcium hydroxide [22]. This resistance has made the control of E. faecalis infection a great challenge for dentists worldwide.
Silver ion and chlorhexidine have been reported to be very effective in the control of E. faecalis infection in root canals [23,24]. However, Ag + possesses high cytotoxicity and discoloring properties when it is in direct contact with the dentin of root canal walls and tissues surrounding the root apex [25]. Chlorhexidine is also cytotoxic to the apical tissues and produces toxic chemicals when it is used with sodium hypochlorite, a commonly used irrigant for canal debridement in root canal therapy [26,27]. Chemicals containing Zn 2+ have also been studied as an intracanal medication against E. faecalis, and were found to be more biocompatible but less effective in infection control when compared with medications containing Ag + [15,26]. Studies on the combined use of Ag + and Zn 2+ reported an improved antibacterial effect in both intracanal and dental implant surface infection control [28,29]. Nevertheless, the mechanism and optimum ratio of Ag + and Zn 2+ in their synergistic antibacterial activity are still elusive, especially in dealing with E. faecalis infection.
Membrane potential is a key feature of bacteria for maintaining membrane permeability, intracellular ionic balance, organelle function and quorum sensing, which are essential for the survival of bacterial communities [30][31][32]. Alteration of the bacteria membrane potential (i.e. hyper-polarization or de-polarization) is an important bactericidal mechanism for many antibacterial medications, including antibiotics [33]. Ion channels on the cell membrane regulate the trans-membrane potential, and the ionic strength in the extracellular matrix affects the activity of ion channels, resulting in the alteration of membrane potential [34]. It is not known whether alteration of membrane potential occurs in synergistic antibacterial activities involving Ag + -Zn 2+ . Accordingly, the objective of the present study was to investigate the optimum ratio of Ag + and Zn 2+ against E. faecalis biofilms on dentin and the potential mechanism behind this synergy. Silver ions, Zn 2+ and chlorhexidine were used independently as controls in the study.
After incubating at 4 °C for 24 h, 10 μL of the inoculum from each group was plated on brain heart infusion (BHI; Land Bridge Technology Co. Ltd., Beijing, China) broth agar (Biosharp, Hirono, Japan) and incubated anaerobically at 37 °C for 24 h. Bacteria inoculum mixed with only sterilized deionized water was included as blank control group for each test. All tests were performed in sextuplicate for each group and CFUs of E. faecalis were counted for group comparisons.

Dynamic growth-curves
A dynamic growth-curve method was used to observe the dynamic antimicrobial effect of Ag + , Zn 2+ or Ag + -Zn 2+ solutions, Briefly, for Ag + , a 4 mL suspension [ At 1, 2, 3, 4 and 5 h (2, 4, 6, 8 and 10 h for the Zn 2+ only group using 1 × 10 3 CFUs/mL E. faecalis) after incubation, 1 mL of the suspension was retrieved for optical density (OD) measurement at 600 nm using a spectrometer (UV-2401PC, Shimadzu Corp., Japan). Bacteria inoculum mixed with sterilized deionized water was used as the blank control group. For each group, the test was repeated six times. All solutions containing Ag + were prepared, stored and operated in the absence of light.

Determination of antibacterial efficiency of synergistic Ag + -Zn 2+ against Ag + only
The antibacterial efficiency of Ag + -Zn 2+ synergy was determined by calculating the bactericidal percentage against same amount of Ag + only, using the information derived from the aforementioned CFU counting and growth-curve tests. The bactericidal percentages against Ag + were calculated for each test using the following equation:

Minimum inhibitory concentration (MIC) and bactericidal concentration (MBC)
Minimum inhibitory concentration and MBC of Ag + , Zn 2+ or Ag + -Zn 2+ at different atomic ratios were determined by a serial microdilution assay. Briefly, for Ag + , 25.6 × 10 −2 mg/mL AgNO 3 solution was prepared. For Zn 2+ , 537.6 × 10 −2 mg/mL Zn(NO 3 ) 2 ·6H 2 O solution was prepared. For Ag + -Zn 2+ , Ag + -Zn 2+ solutions (25.6 × 10 −2 mg/mL AgNO 3 ) at different atomic ratios (1:1, 1:3, 1:6, 1:9 or 1:12) were prepared. From these starting solutions, a series of twofold dilutions were made by adding 50 μL suspension [1 × 10 5 CFUs/mL] of E. faecalis in double-concentrated BHI broth into each well of a 96-well plate with 50 μL Ag + , Zn 2+ or Ag + -Zn 2+ solutions at 37 °C for 24 h anaerobically. An optical density measurement was conducted by a micro-plate reader at 600 nm. Wells with the lowest Ag + , Zn 2+ or Ag + -Zn 2+ concentrations showing OD values most close to that of negative controls were determined to have MIC. MBC was determined by inoculating solutions from wells with Ag + , Zn 2+ or Ag + -Zn 2+ concentrations no less than MIC onto BHI agar plates and calculating the number of viable bacteria to be less than 4% of original bacteria. The test was repeated for three times.
The medium in each well was removed after 24 h. The cells were washed three times with α-MEM and cultured with 100 μL of α-MEM and 10 μL of CCK-8 solution at 37 °C for 4 h. The absorbance at 450 nm was measured using a micro-plate reader (Power Wave XS2, BioTek Instruments, VT, USA). Wells containing only α-MEM were used for background recording. Cells cultured in medium mixed with 10 μL of sterilized deionized water or 2% chlorhexidine were used as control groups.

Inhibition of E. faecalis biofilm formation on dentin by Ag + -Zn 2+
Dentin slices (4 mm wide × 4 mm long × 1 mm thick) were prepared from human extracted wisdom teeth under a protocol approved by the Ethics Committee of School and Hospital of Stomatology, Wuhan University. All dentin slices were cleaned using an ultrasonic bath in deionized water, 5.25% sodium hypochlorite and 17% ethylenediaminetetraacetic acid successively and finally in deionized water. The cleanliness of the dentin surfaces was checked by a field emission scanning electron microscope (FE-SEM; Ultraplus; Zeiss, Oberkochen, Germany). The dentin slices were autoclaved for 20 min at 121 °C in deionized water and incubated anaerobically in BHI broth for 24 h at 37 °C to ensure no bacterial contamination. Sterilized dentin slices were soaked into 3 mL of E. faecalis suspension [1 × 10 8 CFUs/mL] and incubated under anaerobic condition at 37 °C for 4 weeks. Fresh BHI broth was replaced every second day to remove dead cells and to ensure bacterial viability. After incubation, the slices were rinsed with sterile phosphate buffered saline (PBS) to remove floating bacteria and culture medium. Two randomly-selected dentin slices were observed by FE-SEM to verify the presence of E. faecalis biofilm on the dentin surfaces.
After inoculation of the dentin slices, 0.15 g of methylcellulose (Aladdin Industrial Corp., Shanghai, China) was mixed with 2 mL of Ag + -Zn 2+ solution (3.2 × 10 −2 mg/ mL AgNO 3 ) with atomic ratio of 1:1, 1:3, 1:9 or 1:12 (as described above), or 2% chlorhexidine. The latter was prepared by diluting 20% chlorhexidine (Adamas, Basel, Switzerland) with sterile deionized water. Addition of the methylcellulose enabled gels to be produced from those solutions. Dentin slices with E. faecalis biofilms were embedded in the gels and incubated anaerobically at 37 °C for 7 days in a 100% humidity environment. Gels prepared by mixing 2 mL of sterile PBS with 0.15 g methylcellulose were used as the negative control. Gels prepared by mixing 2 mL of 3.2 × 10 −2 mg/mL AgNO 3 or 67.2 × 10 −2 mg/mL Zn(NO 3 ) 2 ·6H 2 O with 0.15 g methylcellulose were used as positive control groups. One hundred dentin slices were divided randomly into 10 groups (N = 10). After incubation, all slices were gently washed in sterile PBS for five times to remove remnant gels. Four dentin slices of each group were randomly selected to be observed by FE-SEM (sputter-coated with conductive carbon or gold) and the other six slices from each group were immersed in 8 mL of fresh BHI broth at 37 °C. At 2, 4, 6, 8 and 10 h after incubation, 1 mL suspension was retrieved for OD measurement at 600 nm using a spectrometer (UV-2401PC).

Zn 2+ pretreatment
A Zn 2+ pretreatment study was designed to further understand the role of Zn 2+ in the synergistic antibacterial effect of Ag + -Zn 2+ . Briefly, a 10 mL suspension [1 × 10 8 CFUs/mL] of E. faecalis was centrifuged at 8000 rpm for 5 min. After centrifuging, the supernatant was discarded. To remove the remaining culture medium, the bacteria pellet was suspended in 10 mL of sterilized deionized water and re-centrifuged. The supernatant was then discarded. For Zn 2+ pretreatment, a 10 mL solution of 33.6 × 10 −2 mg/mL Zn(NO 3 ) 2 ·6H 2 O was added to the bacteria and incubated for 30 min at 4 °C. Sterilized deionized water (10 mL) was used as negative control. After 30 min, the Zn 2+ pretreated bacteria and negative control groups were centrifuged; the bacteria pellet obtained from each group was re-suspended in 10 mL of sterilized deionized water and centrifuged once again to remove the remaining Zn 2+ . The OD values of the suspensions were adjusted to 0.4. Then, 5 mL of Zn 2+ -pretreated bacteria suspension was mixed with 5 mL of 3.2 × 10 −2 mg/mL AgNO 3 solutions and 10 mL double-concentrated of BHI broth. For control groups, a 5 mL bacteria suspension from the negative control group was mixed with 5 mL of 3.2 × 10 −2 mg/mL AgNO 3 solution or sterilized deionized water, together with 10 mL of double-concentrated BHI broth. At 1, 2, 3 and 4 h after incubation at 37 °C, 1 mL of the suspension was retrieved for OD determination. Experiments were conducted in sextuplicate.
For each sample, fluorescence was acquired from 10,000 events by a flow cytometer (Beckman Coulter Inc., CA, USA). Signals were acquired with logarithmic amplification. DiOC 2 (3) was excited at 488 nm, and its green fluorescence was detected through a 525 ± 40 nm bandpass filter and red fluorescence was detected through a 690 ± 50 nm band-pass filter [32]. TO-PRO-3 was excited at 638 nm, and its red fluorescence was detected through a 660 ± 20 nm band-pass filter. Mean fluorescence intensity values of each sample were used for analysis. Because the fluorescence of a bacterial cell or clump is size-dependent, membrane potential was determined by the red/green ratio of DiOC 2 (3) and membrane permeability was evaluated by the values of TO-PRO-3 red/ DiOC 2 (3) green. For each group, the test was conducted in triplicate.

Statistical analysis
Statistical analysis was performed using one-way analysis of variance analysis with a post hoc Student-Newman-Keus test. When the normality or equal variance assumption of the data sets was violated, statistical analysis was performed using non-parametric Kruskal-Wallis analysis of variance with a post hoc Dunn test. For data in normal distribution, mean ± standard deviation (SD) was used, while for data violating normality, median ± quartile (i.e. P25-P75) was used. Statistical significance was preset at α = 0.05.

Synergistic pattern derived from CFUs counting
The CFUs formed on agar plates by E. faecalis incubated with Ag + were significantly less than the blank control except the 0.8 × 10 −2 mg/mL group (p < 0.05; Fig. 1). Unlike the bactericidal effect of Ag + , no significant difference was observed in the CFUs among the blank control and the various Zn 2+ group (p > 0.05; Fig. 2). There was no decline in CFUs with the increasing Zn 2+ concentrations. All the Ag + -Zn 2+ groups except the 1:1 group had significantly less CFUs when compared with the blank control and the Zn 2+ only group (p < 0.05; Fig. 3). The 1:9 and 1:12 Ag + -Zn 2+ groups with 1.6 × 10 −2 mg/mL AgNO 3 showed significantly better antibacterial effect than the Ag + only group containing 1.6 × 10 −2 mg/mL AgNO 3 (p < 0.05). The CFUs decreased extensively with increasing amounts of Zn 2+ in the Ag + -Zn 2+ mixtures.

Antibacterial efficiency of Ag + -Zn 2+ against Ag + only
The antibacterial efficiency of Ag + -Zn 2+ combinations calculated using the results derived from CFU counting and growth curve determination are shown in Tables 1 and 2, respectively. In Table 1, antibacterial efficiencies of groups with 1:6, 1:9 and 1:12 Ag + -Zn 2+ ratio were increased by 50.71, 62.28 and 60.05% respectively when compared with the use of only Ag + . In Table 2, despite the use of two bacteria concentrations (1 × 10 6 CFUs/ mL and 1 × 10 8 CFUs/mL), the antibacterial efficiencies of Ag + -Zn 2+ combinations exhibited the same trend as the results presented in Table 1. Antibacterial efficiency of the group with 1:12 Ag + -Zn 2+ ratio was increased by 66.27% (1 × 10 6 CFUs/mL) and 88.99% (1 × 10 8 CFUs/mL) when compared with the use of only Ag + . Antibacterial efficiency of the group with 1:9 Ag + -Zn 2+ ratio was found very close to that of 1:12 group.

MIC and MBC of Ag + -Zn 2+ combinations
Minimum inhibitory concentration and MBC results were shown in Table 3. MIC and MBC of Zn 2+ were undetectable for its limited antibacterial ability. MIC of Ag + or Ag + -Zn 2+ at different atomic ratios was 6.4 × 10 −2 mg/mL AgNO 3 . MBC of each group was 12.8 × 10 −2 mg/mL AgNO 3 except for the 1:12 group with the MBC being 6.4 × 10 −2 mg/mL AgNO 3 .
For CCK-8 tests of Ag + -Zn 2+ combinations, all ratio groups except the 1:12 group were not significantly different from the blank control (p > 0.05; Fig. 5c). Although the 2% chlorhexidine group, Ag + -Zn 2+ group with 1:12 ratio and Zn 2+ only groups showed significantly lower cell growth rate when compared with the blank control group (p < 0.05), 2% chlorhexidine solutions exhibited the most suppressive effect on cell growth (p < 0.05) in all these three groups.
Although the Ag + -Zn 2+ group with 1:12 ratio showed a slight suppressive effect on cell growth, its cytotoxicity was much lower than the 2% chlorhexidine group (p < 0.05; Fig. 5c).

Inhibition of E. faecalis biofilm grown on dentin using the optimum Ag + -Zn 2+ ratio
Based on the aforementioned findings, the two most potent Ag + -Zn 2+ ratios of 1:9 and 1:12 were used in this anti-biofilm test on dentin. The other two confirmed weaker ratios of 1:1 and 1:3 were also included for comparison. Optical density measurements after direct immersion of dentin slices into fresh BHI solution for 10 h indicated that there was significantly less biofilm accumulation on dentin slices treated with gels containing 2% chlorhexidine, Ag + only or Ag + -Zn 2+ combinations, when compared with the blank control group (p < 0.05; Fig. 6b). Dentin slices from the 2% chlorhexidine group and the Ag + -Zn 2+ group with 1:12 ratio had the least biofilm accumulation, especially when compared with the Ag + only group (p < 0.05; Fig. 6c). The Ag + -Zn 2+ group with 1:1 ratio showed statistically higher OD value than the 2% chlorhexidine group and the Ag + -Zn 2+ with 1:12 ratio (p < 0.05; Fig. 6c). Although no statistical difference was found among 1:3, 1:9, 1:12 and 2% chlorhexidine groups (p > 0.05; Fig. 6c), the OD values deceased from 1:3 to 1:12 group with 2% chlorhexidine and 1:12 ratio groups showing the least biofilm accumulation on dentin. Scanning electron microscopy confirmed the differences in biofilm accumulation on dentin slices from different groups, except for those treated with 2% chlorhexidine as 2% chlorhexidine killed and fixed the bacteria biofilm on the dentin surface (Fig. 7) [37].

Effect of Zn 2+ pretreatment on the bactericidal activity of Ag +
In Zn 2+ pretreatment study, optical density measurement at the 4th h showed that only the Zn 2+ pretreatment group had significantly lower OD values when compared with the control group (p < 0.05; Fig. 4f ). The OD value of Zn 2+ pretreatment group was also much lower than the Ag + only group despite no statistical difference was detected by the non-parametric statistics in this test.

The membrane potential and permeability of E. faecalis under different Zn 2+ conditions
In the membrane potential and permeability study, the membrane potential of E. faecalis exhibited a trend of CFUs/mL) pretreated with Zn 2+ only. (Asterisk: significant difference when compared with the blank control group (p < 0.05); ▲ significant difference when compared with the Zn 2+ only groups (p < 0.05); # significant difference when compared with the Ag + only groups (p < 0.05); Groups inside boxes in the corresponding chart had the same labels as the legend depolarization with increasing amounts of Zn 2+ (Figs. 8  and 9a). However, membrane permeability was not altered significantly (p > 0.05; Fig. 9b). The largest degree of depolarization of the membrane potential occurred when the concentrations of Zn(NO 3 ) 2 ·6H 2 O was more than 33.6 × 10 −2 mg/mL (p > 0.05 when compared with CCCP positive control group; Fig. 8).

Discussion
Silver ions has long since been used as a bactericidal metal element as its strong and broad-spectrum antibacterial ability. However, in this study, some E. faecalis were found still viable when the AgNO 3 concentration reached 51.2 × 10 −2 mg/mL. It might suggest that some strain of E. faecalis has developed the ability to resist the bactericidal effect of Ag + . On the other hand, the AgNO 3 concentration higher than 3.2 × 10 −2 mg/mL exhibited the cytotoxicity. Comparatively, the results from both CFUs counting and growth-curve methods indicated that the bactericidal ability of Zn 2+ is much weaker than Ag + , generally being inhibitive rather than bactericidal to E. faecalis. It has been reported that when the Zn 2+ was mixed with Ag + , the anti-bacterial effects would be significantly improved compared to either Ag + or Zn 2+ alone. However, the pattern and mechanism behind this synergistic activity is still unclear.
In this study, a series of atomic ratios between Ag + and Zn 2+ was produced by serial dilution, keeping the amount of Ag + constant while increasing the amount of Zn 2+ in the mixture. Presumably, this helps to evaluate the role contributed by Zn 2+ in promoting the antibacterial activity of the Ag + -Zn 2+ synergy. Thus, based on the measured MIC of silver nitrate, 1/2 MIC and 1/4 MIC of silver nitrate were selected as the concentration of Ag + in the Ag + -Zn 2+ combinations in order to kept Ag + at a relatively low level (1.6 × 10 −2 mg/mL AgNO 3 in which concentration of Ag + is 1.0 × 10 −2 mg/mL) to facilitate evaluation of its synergistic antibacterial activity with Zn 2+ . Results from the CFU counting method indicated that the synergistic antibacterial effect of Ag + -Zn 2+ combinations is significantly improved only when the Ag + -Zn 2+ ratio is higher than 1:6, i.e. 1:9 and 1:12 in this study. Interestingly, there was no much difference in the antibacterial effect between the 1:9 and 1:12 ratios, which might suggest that these two atomic ratios reach the peak of the synergistic antibacterial ability of Ag + -Zn 2+ combinations. Yang et al. reported a similar phenomenon when 1.7-7% Zn 2+ was added to 0.3% Ag + -doped, heatresistant honeycomb ceramic materials prepared from red mud industrial waste [13]. However, the combination ratios between the zinc and silver in that study were too   limited to identify the mechanism and optimum ratio for Ag + -Zn 2+ synergy against infection. Furthermore, the antibacterial effect of Ag + -Zn 2+ combinations in our study did not show a lower MIC than silver ions only (Table 3), and even the antibacterial effect of 1:12 group was found increased by only 88.99% (Table 2). Considering these findings, it might be more appropriate to use the word "antiseptic" than "antibacterial" in this manuscript. Despite these, these findings could be meaningful in designing new antiseptic or antibacterial materials containing both Ag + -Zn 2+ .
The cytotoxicity test revealed that the 1:9 Ag + -Zn 2+ ratios exhibited no suppressive effect on cell growth, while the 1:12 ration showed a slight suppressive effect. [Asterisk: significant difference when compared with the blank control group (p < 0.05); ▲ significant difference when compared with the Zn 2+ only group (p < 0.05); # significant difference when compared with the Ag + only group (p < 0.05); ■ significant difference when compared with 2% chlorhexidine (p < 0.05)] Considering the 1:9 ratio having the similar strong antibacterial activity as the 1:12 Ag + -Zn 2+ ratio, the 1:9 Ag + -Zn 2+ ratio might be more appropriate for in vivo applications against E. faecalis infection.
In the study against E. faecalis biofilm on dentin, 2% chlorhexidine gel was used as the positive control because it is an effective clinical antibacterial agent against E. faecalis in root canals [24]. SEM images showed that there Fig. 7 Representative canning electron microscopy images showing the inhibition of E. faecalis biofilms grown on dentin. a-c A biofilm treated with PBS gel for 7 days (a ×5000; b ×25,000; c ×50,000) (negative control). d-f A biofilm treated with 2% chlorhexidine gel for 7 days (d ×5000; e ×20,000; f ×35,000). g-i A biofilm treated with Ag + -Zn 2+ gel with 1:12 Ag + -Zn 2+ ratio (3.2 × 10 −2 mg/mL AgNO 3 ) for 7 days (g ×5000; h ×20,000; i ×35,000). j-l A biofilm treated with gel containing only Ag + (3.2 × 10 −2 mg/mL AgNO 3 ) for 7 days (j ×5000; k ×20,000; l ×50,000). m-o Biofilm treated with gel containing only Zn 2+ [67.2 × 10 −2 mg/mL Zn(NO 3 ) 2 ·6H 2 O] for 7 days (m ×5,000; n ×20,000; o ×25,000) seemed still many bacteria left on the dentin surface treated by 2% chlorhexidine, which is different from that of metal ions-treated dentins (Fig. 7). The reason for this difference is because 2% chlorhexidine could kill as well as fix the bacteria biofilm on the dentin surface [37]. The OD values of the Ag + -Zn 2+ group with 1:9 and 1:12 ratios were similar to that of 2% chlorhexidine, which might suggest that Ag + -Zn 2+ combination with a 1:9 or 1:12 atomic ratio was as efficacious as 2% chlorhexidine in eradicating E. faecalis biofilms on dentin. Considering that 2% chlorhexidine is significantly more cytotoxic than the Ag + -Zn 2+ combination with 1:9 or 1:12 atomic ratio, the latter may be used to produce new medication against E. faecalis biofilm infection on dentin. The bactericidal mechanism of chlorhexidine is different from that of metal ions. Chlorhexidine could kill the bacteria through binding to acidic groups on the surface of the bacteria [38,39], while metal ions normally enter the cell and combine with the enzyme or other proteins inside the cells, resulting in the cellular malfunctioning and death [40].
To better show and understand how the Ag + -Zn 2+ co-worked in antibacterial activities, a single E. faecalis bacteria strain was selected in this study to avoid interferences from different bacteria strains. The synergistic antibacterial effect of Ag + -Zn 2+ combinations could be different on other bacteria or on other materials surfaces with different surface topography [41]. Rathnayake et al. ever reported that multi-drug resistance in clinical isolates (71.2% of E. faecalis and 70.3% of E. faecium) was much higher than water isolates (only 5.7% E. faecium) [42]. Hence, further studies need to be conducted on clinical isolates and multi-species biofilms on different material surfaces, which may express resistance trait against Ag + -Zn 2+ effect.
The Zn 2+ pretreatment experiment was designed to clarify whether the Zn 2+ really improves the antibacterial activity of Ag + in the Ag + -Zn 2+ combination. In this experiment, a low Ag + concentration (0.8 × 10 −2 mg/ mL AgNO 3 ) was used to facilitate observation of the enhanced antibacterial effect produced by Zn 2+ pretreatment. The results suggested that Zn 2+ pretreatment significantly enhanced the sensitivity of E. faecalis to the bactericidal activity of Ag + . This phenomenon may partly explain why combining Ag + and Zn 2+ together produces a much stronger antibacterial effect than the use of Ag + alone. To further investigate how Zn 2+ makes E. faecalis more sensitive to Ag + , the membrane potential and permeability of E. faecalis was monitored by flow cytometry after the cells were exposed to different amounts of Zn 2+ . This is because membrane potential and permeability are closely related to the sensitivity of bacteria to ionic environment; ion homeostasis affects the proliferation, communication, metabolism and survival of bacteria [30,34]. Zinc ions have been reported to alter the mitochondria membrane potential of rat brain cells [43,44]. Nano-zinc oxide, for example, changes the permeability of the cell membrane [45,46]. The results of this study suggested that the Zn 2+ depolarizes the membrane potential but does not cause the bacterial cell membrane to rupture and leak. It is consistent with cytotoxicity testing, in which Zn 2+ showed an inhibitive instead of destructive effect against bacteria. The largest degree of depolarization happened when the concentrations of Zn(NO 3 ) 2 ·6H 2 O was more than 33.6 × 10 −2 mg/mL, which might probably explain why the Ag + -Zn 2+ combination was more potent when its atomic ratio was between 1:6 and 1:12.
Maintenance of membrane potential is dependent on normal functioning of the proton pumps and ion channels in the cell membrane [47,48]. Energy derived from adenosine triphosphate (ATP) is required for homeostasis of membrane potential, including ion transportation through the cell membrane and excretion of toxic heavy metal ions such as Ag + [49]. When the concentration of extracellular Zn 2+ increases as in the present study, the bacteria will transport some cations, including Na 2+ and Zn 2+ , into the cell to maintain their membrane potential. However, Zn 2+ is an inhibitor of ATP synthesis and oxidative phosphorylation in cells [50][51][52]. P-type ATPases are the essential enzymes involved in bacterial resistance to heavy metal ions [49,53,54]. Ion transportation through the cell membrane is hampered when ATP synthesis is inhibited by Zn 2+ . This may result in the accumulation of cations within the cell, including toxic Ag + . Intracellular cation accumulation, especially the heavy metal Ag + ions, could result in depolarization of membrane potential and ultimately destruction of cells. The potential pathways through which Zn 2+ influences Ag + transportation and accumulation requires investigation in future studies.

Conclusions
Based on the findings of the present study, it may be concluded that the synergistic antibacterial effect of Ag + and Zn 2+ is dependent upon the amount of Zn 2+ present in the medium. Atomic Ag + -Zn 2+ ratios higher than 1:6 (i.e. 1:9 and 1:12 in this study) appear to be optimum ratios against both planktonic E. faecalis as well as singespecies E. faecalis biofilms. This synergistic antibacterial effect may be attributed to the ability of Zn 2+ to depolarize the bacterial cell membrane. The 1:9 and 1:12 Ag + -Zn 2+ combinations are as efficacious as 2% chlorhexidine against E. faecalis biofilms that accumulate on dentin. New medications containing an optimal Ag + -Zn 2+ ratio may be developed against E. faecalis infection of tooth root canals as well as infections in other parts of the human body.