Real-time monitoring of the pH microenvironment at the interface of multispecies biofilm and dental composites

Abstract

Bacterial-mediated local pH change plays an important role in altering the integrity of resin dental composite materials in a dynamic environment such as the oral cavity. To address this, we developed a 300-μm-diameter, flexible, solid-state potentiometric pH microsensor capable of detecting and quantifying the local pH microenvironment at the interface of multispecies biofilm and dental resin in real time over 10 days. We used fluorinated poly(3,4-ethylenedioxythiophene) as the back contact in our newly developed pH sensor, along with a PVC-based ion-selective membrane and PTFE-AF coating. The high temporal resolution pH data demonstrated pH changes from 7 to 6 and 7 to 5.8 for the first 2 days and then fluctuated between 6.5 to 6 and 6 to 5.5 for the remaining 8 days with the resin composite or glass slide substrate respectively. We could observe the fluctuations in pH mediated by lactic acid production within the biofilm and the reestablishment of pH back to 7. However, acid production started to overwhelm buffering capacity with the continuous feed of sucrose cycles and reduced the local pH nearer to 5.5. No such changes or fluctuations were observed above the biofilm, as the pH remained at 7.0 ± 0.2 for 10 days. The localized real-time monitoring of the pH within the biofilm showed that the pH shift underneath the biofilm could lead to damage to the underlying material and their interface but cannot be sensed external to the biofilm.

Graphical Abstract

1. Introduction

Nutrition and diet play significant etiological roles in many dental diseases, including dental caries, enamel developmental defects, and dental erosion [1–5]. Tooth demineralization is primarily mediated by bacterial metabolic activity, especially the production of a wide variety of organic acids as a result of fermentation of simple or complex carbohydrates [6]. Among the complex sugars, sucrose produces a sticky exopolysaccharide matrix that aids bacteria in adhering to teeth or dental materials and forming a three-dimensional (3D) structure, thus creating a unique microenvironment gradient inside the biofilm and on the outside surface matrix [7–10]. This further promotes the development of a separate microenvironment at the interface of the biofilm and the restored tooth, which has an important effect on the integrity of the bonded interface with dental composites.

The most commonly used dental restorative materials are polymer or methacrylate based [11]. The acid produced by bacteria has the potential to destabilize the ester bonds within the cross-linked polymer network, leading to marginal degradation and potentially failure of the overall restoration [12,13]. It has also been reported that the composite surface showed significant surface roughness at pH 6.0, which might facilitate the bacterial attachment and initiate the process of composites degradations [14,15]. In addition, when the bacteria-mediated local pH drops below 5.5, the teeth demineralization starts. Therefore, the combined effect of adjacent tooth demineralization along with the onset of composites degradation will exacerbate the problem that might lead to dental filling failure due to bacterial growth. Hence, understanding the interfacial chemical microenvironment between biofilm and dental composites is crucial for developing better dental restorative materials and testing the longevity of existing dental composites.

The major challenge lies in detecting and quantifying the chemical microenvironments in real time at this hard-to-reach interface [16]. Confocal laser scanning fluorescence microscopy (CLSM) has gained popularity in the measurement of pH inside the biofilm structure [6,17,18]. Typically, biofilm is grown in the presence of dextran modified with fluorescent dye to enable visualization of the fluorescent-labeled polysaccharide matrix through CLSM imaging. Koo et al. [17] reported that the pH at the bottom of the biofilm grown on hydroxyapatite disks shifted from 6 to 5.5 even in the presence of buffered brain heart infusion (BHI) growth medium. Koo et al. also recorded time-dependent pH images over a period of 2 h to avoid bleaching of the fluorescent dyes. Real-time biofilm monitoring is not possible because of the dye characteristics. In addition, the fluorescent-based pH calibration curve obtained in bulk solution does not always reflect the complex biofilm matrix in which the pH measurements are made. Thicker biofilm poses an additional challenge in imaging at the base of the biofilm, as it becomes increasingly difficult for the excitation beam to penetrate it [17]. Hardie et al. [19] revealed the presence of dynamic pH gradients during biofilm formation of Pseudomonas aeruginosa and Streptococcus mutans through fluorescent nanosensors. The experiment was stopped after 22 h, however, because of the limitation of the growth chamber conditions. Nanosensor variability also presents a limitation in its interaction with bacteria in specific scenarios such as complex multispecies biofilms. Potentiometry, in contrast, has inherent advantages of miniaturization and is suitable for long-term real-time monitoring in complex environments such as biofilms [20–22]. With the high selectivity of ionophore-based pH sensors, data acquisition would be more specific in such a complex matrix environment. The major challenges of using a pH sensor to monitor biofilm growth at the interface of interest are its stability over a long period (days), anti-biofouling characteristics, size, and flexibility.

To address these challenges, we developed a miniaturized 300-μm solid-state pH sensor with fluorinated poly(3,4-ethylenedioxythiophene) (PEDOT-F) as the back contact material, allowing the sensor to be placed at the hard-to-reach interface of dental composites and biofilm without disrupting the growing biofilm. In this study, our objective was to monitor pH during multispecies biofilm growth in real time to gather quantitative data on the pH change over the resin during a prolonged period, and without affecting the structure or function of the biofilm.

2. Materials and methods

Materials.

Hydrogen ionophore-I, tetradodecylammonium tetrakis(4-chlorophenyl)borate (ETH 500), and high molecular weight poly(vinyl chloride) (PVC) were purchased from Sigma Aldrich. Carboxymethylcellulose sodium salt, tetrabutylammonium hexafluorophosphate (TBAPF₆), and perfluoromethylcyclohexane were purchased from Alfa Aesar. Potassium tetrakis (4-chlorophenyl) borate (KTCIPB) and bis(2-ethylhexyl) sebacate (DOS) were purchased from TCI America. Tetrahydrofuran, dichloromethane, yeast extract, and poly[4,5-difluoro-2,2-bis(trifluoromethyl)-1,3-dioxoleco-tetrafluoroethylene] (PTFE-AF) were purchased from Sigma-Aldrich; trypticase peptone from BD Bioscience (Franklin Lakes, NJ, USA); protease peptone from Hardy Diagnostics (Santa Maria, CA, USA); and 125-μm Teflon-coated platinum/iridium wire from World Precision Instruments.

Instrumentation.

Electrochemical measurements were made with a six-channel high impedance potentiometer (Lawson Labs, Malvern, PA, USA) and a bipotentiostat (CH Instruments, Austin, TX, USA).

Synthesis of fluorinated EDOT (EDOT-F).

The EDOT-F was synthesized in a four-step sequence starting from the commercially available perfluorooctyl iodide and allylic alcohol. First, azobisisobutyronitrile-mediated perfluoroalkylation followed by base-mediated iodide elimination afforded the epoxide intermediate, which was subsequently hydrolyzed to give the diol intermediate containing the hydrophobic perfluoroalkyl side chain. Finally, this diol was coupled with dimethoxythiophene to complete the synthesis of the EDOT-F product. The details of the synthesis procedure and the corresponding NMR spectra (Figs. S1 and S2) can be found in Supporting Information.

Fabrication of solid-state ion-selective electrode.

Preparation of back contact.

A 125-μm-diameter Teflon-coated platinum wire was used as a working electrode. First, the 0.5-mm Teflon coating was stripped off on one end for use as the sensor and 1 cm was exposed on the other end for the electrical connection. The electrode was washed with ethanol and dried before electropolymerization of EDOT-F as a back contact material. An Ag wire and stainless-steel coil were used as a reference and a counter electrode, respectively. Electropolymerization of PEDOT-F was performed by cyclic voltammetry, using a bipotentiostat (CH Instruments, USA) and multiple cycling in the potential range of 0 to +1.6 V at a scan rate of 50 mV/s. The freshly prepared solution of 20 mM of EDOT-F containing 0.1 M of TBAPF₆ in acetonitrile was used for electropolymerization.

Ion-selective membrane coating.

After preparing the PEDOT-F back contact, the electrode was dipped four times in the ionophore cocktail. The final membrane composition was as follows: 4 wt% H⁺ ionophore, 2 wt% KTCIPB, 63 wt% DOS, and 31 wt% PVC. For the control electrode, the final membrane composition was 6 wt% ETH500, 63 wt% DOS, and 31 wt% PVC. All electrodes were dried at room temperature for 1 min and further coated with 4 μL of PTFE-AF dissolved in perfluoromethylcyclohexane three times. Sensors were later conditioned in artificial saliva at pH 4.5 for 12 h. The final diameter of the sensors or ion-selective electrodes was approximately 300 μm. The detailed design of the pH solid state sensor is shown in Fig. 1B.

Fig. 1.

Schematic diagram of (A) incubator setup with sensors, (B) 300-μm flexible ion-selective electrode design, and (C) flexible pH sensors on top of the substrate embedded in the biofilm. ficial saliva

Electrochemical characterization.

The pH sensors were calibrated in three different testing solutions, including artificial saliva, universal buffer, and basal medium mucin (BMM) (pH 7 to 4). The response times were measured both in artificial saliva and BMM solutions. The artificial saliva is composed of CaCl₂, MgCl₂, KCl, KH₂PO₄ and HEPES buffer, and BMM solution is composed of protease peptone, trypticase peptone, yeast extract, KCl and carboxymethyl-cellulose [23]. The reversibility was determined in BMM with additions of 1 M HCl and NaOH. Selectivity tests for the pH sensors were performed in the universal buffer composed of H₃BO₃, H₃PO₄, CH₃COOH, and NaOH with fixed interference method.

Dental plaque-derived multispecies biofilm grown on glass or resin substrates.

Dental plaque-derived multispecies biofilms were grown on glass and resin substrates inside a homebuilt specialized incubator system. Details of the biofilm growth chamber are shown in Fig. 1A. The resin composites were prepared by mixing of 50 wt% triethyleneglycoldimethacrylate (TEGDMA) and 50 wt% bis-phenol-A-glycidyl methacrylate (bis-GMA) in a centrifugal mixing system for 1 min at 2300 rpm (DAC 150, Flackteck, Landrum, South Carolina). Then, resin mixture also contained 0.4 wt% camphorquinone (photo initiator), 0.8 wt% ethyl-p-dimethylaminobenzoate (co-initiator), and 0.05 wt% butylated hydroxytoluene (inhibitor). The resin was loaded into a disk-shaped mold between glass slides and irradiated for 40s per side [23]. The cured resin was used for the biofilm incubator experiment. Each resin disk was covered with Kapton tape, with a small 3 × 3 mm area left exposed for growing biofilm (Fig. 1C). The disks were then fixed inside a 35-mm Petri dish with double-sided tape. Two pH sensors and one control were placed carefully over the substrate. In addition, we add one control and one pH wire sensor vertically and approximately 1-cm above the substrate in media solution inside the same petri dish. The Petri dish was placed in the incubator to grow the biofilm at 37 °C in the presence of flowing 5% CO₂ (Fig. 1A). BMM media was then flowed through the dish at 3.6 mL/h by using a syringe pump. After potentials recorded by the sensors are stable, a 1-mL solution of dental plaque was suspended in BMM media adjusted to an O.D. of 0.5 in the presence of 150 mM sucrose, and this was added to the Petri dish via a sterile injection port. The plaque had been collected from the molar teeth of a participant in accordance with strict university institutional review board protocols for human subjects (IBC 3117). The bacteria were inoculated inside the dish for 2 h before restarting the BMM solution flow at a rate of 3.6 mL/h. CO₂ flow was maintained at 5% throughout the experiment to ensure a constant positive pressure environment inside the box. Bacteria were fed 150 mM sucrose every 8 h pH sensor and control sensor measurements were recorded continuously with a multichannel high impedance potentiometer for 10 days.

Biofilm volume determination by SECM and CLSM.

Biofilm morphology and volume measurements were performed as previously explained in detail [24,25]. In short, the dental plaque biofilm grown on the resin or glass slide substrate was carefully fixed on a 35-mm Petri dish and placed on the SECM stage. Artificial saliva of pH 7.2 containing 1 mM ferrocene methanol was added to the dish at room temperature. A Pt ultra-microelectrode (diameter 10 μm) was used as the SECM probe. An Ag/AgCl wire was used as the reference electrode and a stainless-steel wire as an auxiliary electrode. The average biofilm volume was calculated by measuring a series of SECM 3D images at different locations on the biofilm surface (3 × 3 mm).

For CLSM experiments, the specimen with biofilm was taken out of the incubator box and incubated in the dark for 30 min with SYTO9/propidium iodide dye inside a biosafety cabinet. The biofilm sample was then placed on a confocal microscope platform for further imaging. A 488-nm laser source was used for both fluorescence dyes, and the emission was collected with a 480- to 520-nm and a 610- to 650-nm filter to visualize the biofilm height. COMSTAT2 software was used to calculate the volume of biofilm.

3. Results and discussion

Electrochemical characterization of pH sensors.

The sensors were calibrated in a universal buffer solution in the pH range of 4–11 and showed a near-Nernstian slope of 57.5 ± 0.5 mV/pH (n = 5) (as representative in Fig. 2A). Later, the sensors were also calibrated in artificial saliva solution and the four-point calibrations from pH 7 to 4 showed a Nernstian slope of 59.7 ± 2 mV/pH (n = 5) (as representative in Fig. 2A). We used BMM for all experiments to mimic human saliva in terms of viscosity and high protein content. The pH calibration with BMM demonstrated a near-Nernstian slope of 60.1 ± 2 mV/pH (n = 5), as shown representatively in Fig. 2A. The sensors showed a response time of 0.5 ± 0.1 s in artificial saliva (n = 5) and 8 ± 3 s in BMM solutions (n = 5) (as representative in Fig. 2B). The slower response time of pH sensors in BMM solution was due to its high viscosity and nonspecific adsorption of proteins present in BMM solution.

Fig. 2.

Electrochemical characterization of pH sensors. (A) Calibration curve of solid-state pH sensor in universal buffer (slope = 57.5 ± 0.5 mV/pH), arti (slope = 59.7 ± 2 mV/pH), and basal medium mucin (BMM) (slope = 60.1 ± 2 mV/pH). (B) Potentiometric response time of the sensor in artificial saliva (0.5 s) and BMM (8 ± 3 s) solutions. (C) Reversibility of the pH sensor in BMM from pH 7 to 6 and 5, and then back to 7. (D) Selectivity coefficient of pH sensors against major interfering ions such as Ca²⁺, Na⁺, and K⁺.

In our current biofilm experimental setup, reversibility of the sensors is an important parameter because of the unpredictable nature of bacterial metabolic activity and the corresponding change in local pH. The pH sensors showed excellent reversibility between pH 5 and 7, as shown in Fig. 2C. The pH values were validated by using a commercial glass pH electrode.

For long-term sensor experiments, it is essential to estimate sensor drift under the experimental conditions; therefore, we performed a series of control experiments. To characterize the drifts, we set up a series of control experiments in BMM media (pH 7) without any biofilm, using the same experimental setup as shown in Fig. 1A. While the control experiment without bacteria ran for 10 days, there is initial protein absorption from BMM solution on the ion-selective membrane that shifted the potential toward being more negative for the first 2 days [22,26]. In all cases, pH sensors and control sensors followed the same drift trends (Fig. S3A). The actual drift of pH sensors was estimated to be 1.2 mV/day after subtracting the background potential drifts. The pH readings obtained from our control experiment would have changed by 0.2 pH units over 10 days because of the drift of the pH sensor. With 3% error from pH readings and stable measurements for the control experiment, the pH sensors would be credible for the measurement at the interface of the bioactive material and biofilm. This is supported by the pH data for the BMM solution in control experiments and bulk BMM solution in biofilm experiments over 10 days showed an average value of 7.05 ± 0.18 (Fig. 3C, Fig. 4C, and Fig. S3B). We observed bacteria also adsorb and grow on the wire sensors but in significantly lower volume (10 μm or less instead of 100–300 μm thickness on the substrate). Henceforth, we are not expecting any significant change in local pH underneath the biofilm as the biofilm is continuously being flushed with a new media solution of pH 7.0.

Fig. 3.

Real-time monitoring of pH at the interface of multispecies biofilm and resin composite substrates at the left (A) and right (B) spots. The pH in bulk solution far away from the biofilm (C). Fluorescence image of multispecies biofilm growth for 10 days with sucrose feeding every 8 h and placement of pH sensors on the left (D) and right (E) location of the resin substrate. (For other trials see Table T1).

Fig. 4.

Real-time monitoring of pH at the interface of multispecies biofilm and glass substrates at the left (A) and right (B) spots. The pH in bulk solution is recorded vs time (C). Biofilm morphological images at the left (D) and right (E) spots were obtained by using SECM. (For other trials see Table T1).

The sensors were also tested for selectivity against major interfering ions (such as Na+, K⁺, and Ca²⁺) by the fixed interference method [25,27,28]. The sensors showed high selectivity against interfering ions, with a selectivity coefficient of −10.3 log units against Na⁺, −10.6 log units against K⁺, and −10.8 log units against Ca²⁺ (Fig. 2D).

Quantification of pH change at the bottom of the dental plaque-derived multispecies biofilm.

Local pH is important to the integrity of the substrate, especially biofilm-mediated corrosion in dental settings. Hence, we used our newly developed pH sensors to determine the change in local pH at the bottom of the biofilm in real time as the biofilm was developing on resin surfaces. In this study, we primarily tested the bacterial-mediated change in local pH at the interface of resin, a widely used dental filling biomaterial. A glass substrate was used as a control in this study to compare the relative biofilm growth and the corresponding change in local pH with the resin.

For the resin substrate (Fig. 3), we observed that the pH at the interface of resin and multispecies biofilm changed from an initial value of 7.0 to 6.0 over 10 days. At day 1, the bacteria were still adsorbing on the resin substrate and forming microcolonies, and thus the pH remained the same as the solution pH of 7.0. We observed that the bacteria took from day 1–3 to start forming a 3D matrix or biofilm. After day 4, the interfacial pH started fluctuating between 7.0 and 6.0 as the bacteria started forming biofilm and producing a significant amount of lactic acid by sucrose metabolism and were thus able to titrate the buffer locally (Fig. S4). However, the biofilm used in this study was dental plaque-derived multispecies biofilm, which means that the distribution of the bacterial species also had an important role in maintaining a healthy chemical microenvironment, or local pH in this case. In a multispecies microbial ecosystem, among other bacteria, the biofilm consists of lactic acid-producing species (Streptococcus mutans and Lactobacillus) and hydrogen peroxide-producing species (S. gordonii and S. sanguinis). It is well-known that prolonged sucrose exposure causes the biofilm to shift toward more pathogenic bacterial species or to the more abundant presence of lactic acid-producing bacteria, but it is unclear how the distribution shift of these bacterial species causes local pH change inside the biofilm. In this study, we observed interesting phenomena such as how the local pH shifted to the more acidic side (pH below 5.5) and how the bacterial biofilm as a whole attempted to reestablish a pH close to 7.0. A pH of 5.5 is critical in dental biofilm because tooth demineralization occurs below this pH. From the pattern of pH data (Fig. 3A and B), it is clear that the biofilm was adjusting the pH to a neutral level (i.e., the pH fluctuated from 7.0 to 6.0 and then from 6.5 to 6.0, and so on) and with each cycle, the pH slowly shifted toward the more acidic side. This pattern is contrary to the common belief that the pH changes linearly from 7.0 to 5.5 or below in multispecies biofilm. With sucrose feeding every 8 h favors lactic acid production and aids in titrating the local buffering capacity; thus, the biofilm as a whole cannot sustain a considered healthy pH range of 6.0–7.0. However, one should also observe that the fluctuation in pH occurred over 24 h or more, suggesting that the sensor captured the actual natural bacterial metabolic activities rather than noise. Fig. 3A and B shows different trends in local pH over the same 10-day period, as is evident from the different biofilm growth at two different spots. Since the fluorescence-based biofilm volume is collected at the end of 10 days, we can only compare the final pH 6.5 and 6.1 and biofilm volume per surface area 0.11 and 0.12 mm³/mm² at the two different spots, respectively. However, these fluorescence-based data do not give any indication of how the biofilms developed at the two different spots on the same substrate. The difference in trends in local pH data between the left and the right spots (Fig. 3A and B) might represent different biofilm growth rates or different biofilm species distribution (i.e., the right spot might have had more acid-producing species than the left side did). Spatial bacterial genomics is beyond the scope of the current study.

Similar differential pH trends were observed at two different spots in the glass substrate experiments (Fig. 4A and B). The biofilm on the left spot on the glass substrate showed a much slower change in local pH than did the biofilm on the right spot, the latter of which showed a sharp consistent drop in local pH at 5.5. However, the final pH in both cases (left and right) was observed to be close to 5. The biofilm grown on the glass substrate for 10 days also had a significantly higher volume than that on the resin substrate as bacterial adhesion is significantly favorable on glass substrate. We used SECM to measure the volume of the biofilm grown on the glass substrate because it was too thick for CLSM, as the laser cannot penetrate beyond a 200-μm depth. The biofilm volume per surface area at two locations on glass substrate are 0.35 and 0.36 mm³/mm², respectively. As observed in Figs. 3 and 4, local pH changes are higher on glass than on resin substrate. However, if we normalize the pH change with respect to biofilm volume as shown in Table T1, the local pH changes due to bacterial biofilm grown on glass and resin substrates are the same. This further confirmed that the biofilm growth rate on different substrates contributes to the chemical microenvironment at the biofilm-substrate interface and thus contributes to different final pH within the same period.

For both glass and resin substrate experiments, a separate pH sensor was also added to the bulk solution to determine whether the change in pH was local or whether it also affected the bulk solution outside the biofilm. We observed that the bulk pH remained constant over 10 days, confirming that the pH change was purely local near the biofilm and acid production within the biofilm was not able to titrate the bulk solution (Fig. 3C). This finding also demonstrates that our pH sensors were recording the correct pH in a complex matrix solution. The unchanged bulk pH of 7.0 also validates the observation that caries is local, i.e., cariogenic bacteria on one area of a tooth does not cause local pH changes at other sites without biofilms.

4. Conclusion

This high temporal resolution pH study demonstrates that the local pH produced by bacterial metabolic activity is different from the pH of the bulk solution. Such an observation was made possible with our newly developed fluorine-coated miniaturized pH sensor. The present study also shows that there is no significant difference in change in local pH at the interface of multispecies oral biofilm and biomaterial or glass substrate. However, the bacterial growth rate or biofilm volume might be significantly different on the different substrates such as glass and resin, leading to different final pH within the same period. Thus, this unique sensor opens up the opportunity to study other microbiome and biomaterials interfaces by using miniaturized solid-state pH sensors.

HIGHLIGHTS.

• Flexible, solid-state potentiometric pH microsensor of 300 μm was developed.

• The newly developed pH sensors showed a slope of 59.7 ± 2 mV/pH and stability for ten days in the complex biofilm matrix.

• The sensors can detect and quantify the pH microenvironment at the interface of oral biofilm and dental resin in real-time.

• The local pH at the interface of biofilm-resin showed an acidic shift while the bulk pH in the solution remained constant at neutral.

• The pH shift at their interface could lead to damage to the underlying material but cannot be sensed external to the biofilm.