Glucose Microsensor with Covalently Immobilized Glucose Oxidase for Probing Bacterial Glucose Uptake by Scanning Electrochemical Microscopy

Abstract

We have developed a new dual-tip glucose sensing scanning electrochemical microcopy (SECM) probe by covalently immobilizing the glucose oxidase (GOD) enzyme onto an ultra-micro electrode (UME) to measure the local glucose consumption of Streptococcus mutans (S. mutans) biofilms. GOD was immobilized on a novel enzyme immobilization matrix of functionalized multiwalled carbon nanotubes (f-MWCNTs) and 1-butyl-4-methylpyridinium hexafluorophosphate (ionic liquid/IL) packed into the etched Pt UME. The highly selective GOD-based SECM tip showed a high current density of 94.44 (±18.55) μA·mM⁻¹·cm⁻² from 0.1 mM to 1.0 mM at 37 °C as a result of the synergetic effects of f-MWCNTs and ionic liquid. The detection limit of the new 25 μm diameter glucose sensor is 10.0 μM with a linear range up to 4.0 mM. The sensor was successfully used to quantify the rate of glucose consumption of S. mutans biofilms in the presence of sucrose. S mutans catabolizes both glucose and sucrose, producing lactic acid, reducing the local pH, and causing dental caries. With sucrose, S. mutans produces exopolysaccharides to enhance bacterial adhesion on the tooth surface; subsequent lactic acid production reduces the local pH, resulting dental caries. Because of the high selectivity of the sensor, we were able to quantify glucose consumption in the presence of sucrose. S. mutans preferentially consumed sucrose in a mixed diet of both sucrose and glucose. Furthermore, using this unique fast-response (~ 2 s) glucose sensor, we were for the first time able to map the distribution of the glucose consumption profile in the local environment of S. mutans biofilm. These findings provide insight into how the fast-growing S. mutans creates nutrient-depleted regions that affect the survival and metabolic behavior of other bacterial species within oral biofilm.

GRAPHICAL ABSTRACT

INTRODUCTION

Human oral biofilm is a well-studied bacterial community in which the metabolites and metabolic byproducts of the inhabitant species greatly dictate the survival and pathogenesis of bacteria in their host sites. S. mutans is a facultative anaerobic bacterium found in oral biofilm and is known to be the main contributor to the formation of dental caries which is the process of demineralization of tooth enamel due to bacterial activity leading to dental cavity formation¹. This bacterium catabolizes the sugars in an individual’s diet to produce lactic acid as the final product of the glycolytic pathway, at the same time decreasing the pH of the local bacterial environment below 5.5. Tooth enamel demineralizes as a result, leading to dental carries. Acidic local pH from S. mutans activity also disrupts healthy bacterial hemostasis in the oral cavity.

Genomic studies have shown that S. mutans is capable of catabolizing a large range of carbohydrates, including sucrose, glucose, fructose, galactose, mannose, and starch². Sucrose is known as the most cariogenic among all fermentable carbohydrates because, in addition to its fermentation for energy production, it is also the precursor for the formation of exopolysaccharide rich S. mutans biofilms, a critical factor in their virulence behavior, in contrast to glucose, which produces the least exopolysaccharides³. A regular diet composed of multiple sugars gives S. mutans the opportunity to uptake sucrose and other simple sugars such as glucose, although no quantitative evidence is available on the sugar uptake behavior of S. mutans when exposed to such a diet. It is important to study real-time fluctuations of local glucose concentration in the presence and absence of sucrose because these fluctuations are crucial indicators of the metabolic performance of S. mutans and its virulence behavior when exposed to a mixed diet.

Bacterial metabolic activity results in the depletion of nutrients and the accumulation of metabolic by-products. These chemical changes are always localized to an environment that extends only up to a few hundred micrometers from the biofilm surface. Genomic studies⁴, competition assays on solid or liquid media^5,6, and spectroscopic techniques⁷ have been used for decades to explore the genetic regulation and metabolic pathways of oral bacteria. These techniques have limited ability, however, to provide quantitative information on real-time chemical changes in the local microbial environment. SECM has been extensively used to study biological systems because of its ability to reach the local environment and to quantify the molecules of interest associated with bacterial⁸ or cellular^9,10 metabolic activity with high spatial resolution. The nondestructive nature of this technique makes it an ideal method for real-time studies involving live systems. SECM is a scanning probe technique in which an ultra-micro electrode (UME) (radius ≤25 μm) serves as the scanning probe¹¹. Many of the previously reported SECM studies on bacteria were performed with standard platinum or gold UMEs^12,13. When specific analytes in the bacterial environment need to be monitored, however, specific micron-sized sensors are required. Therefore, highly sensitive and selective glucose microsensors need to be designed to measure the local glucose concentration around S. mutans in the presence of other sugars.

The enzyme glucose oxidase (GOD) is a promising sensing mechanism in the design of glucose microsensors¹⁴. High selectivity of the enzyme leaves fewer possibilities for interference in a complex biological system. Although numerous studies have been published on glucose sensors that use GOD¹⁵, only a few of them have been SECM compatible. Kueng et al. reported their work on the fabrication of ultra-micro glucose sensors for SECM with a dual-tip probe composed of two 5 μm Pt UMEs¹⁶. They co-immobilized GOD and hexokinase within an electropolymerized polymer layer on one Pt UME. They used the unmodified second Pt UME to fix the probe-substrate distance by approach curves and used the glucose sensor to image ATP transport by SECM through a polycarbonate membrane layer. Their electropolymerized enzyme layer was not confined to the electrode surface but extended to the glass surface, resulting in a glucose sensor with a larger geometric surface area than that of the UME. The same group integrated the glucose sensor in an atomic force microscopy tip by localized entrapment of GOD within a xylylene polymer layer^17,18. The combined probe has shown a low sensitivity of 2.283 pA·mM⁻¹, which limits its use in probing very low glucose concentrations. Ciobanu et al. compared two different approaches for fabricating glucose UMEs¹⁹. In the first approach, they co-immobilized the enzyme with electropolymerized 2-aminophenol and in the second, they cast the sensor by touching the Pt UME tip onto a droplet of solution containing the enzyme glutaraldehyde and bovine serum albumin. As reported, neither sensor allowed them to obtain approach curves for fixing the probe-to-substrate distance. Therefore, to position the electrode at a known distance above the substrate, they used an inverted microscope along with the SECM setup. Later, Soldà et al. also reported a detailed comparison study on ultra-micro glucose sensors fabricated with electropolymerized polymers and with cross-linking agents, bovine serum albumin, and glutaraldehyde²⁰. They showed that glucose microsensors that were fabricated via cross-linking had better sensitivity than did electropolymerized glucose sensors. Nonetheless, this method had technical difficulties in that the enzymes could be selectively immobilized only on the electrode surface when ultra-microsensors were fabricated.

The ability to obtain approach curves to fix the sensor-to-substrate distance while achieving higher sensitivity with a small surface area is the foremost challenge of designing SECM compatible enzymatic glucose sensors. Immobilization of enzymes with cross-linking agents involves technical difficulties in selectively modifying only the electrode surface of UME, whereas immobilization in electropolymerized polymer layers results in sensors with a larger effective geometric area than that of the original UME. In addition, the enzymes immobilized through these methods do not exhibit the same activity as free enzymes, because of the non-optimum orientations and hindrance of active sites caused by the immobilization matrix. Direct electrophoretic deposition of enzyme molecules has been used as an enzyme-immobilizing strategy for UME preparation²⁰, as this method is more effective in limiting enzyme immobilization to the electrode surface. Nonetheless, this approach is challenging because of the tendency of the deposited enzymes to leach out from the electrode surface during extended SECM experiments. This alters sensor performance during usage. Additional protective coatings such as Nafion can be applied to reduce enzyme leach, but they lower the sensitivity of the sensor. The requirement to make the measurements in a bacterial system in the presence of sugars such as sucrose adds another layer of complication to the sensor design. Although GOD is highly selective for glucose, metabolites or biomolecules originating from the bacteria and sucrose can interfere with the enzyme-immobilizing matrix, affecting sensor performance. To the best of our knowledge, no publications to date have reported using a glucose microsensor as an SECM tip capable of performing an approach curve and measuring the local glucose concentration above live biofilm.

Hence, we have developed a dual-tip SECM probe with a new highly sensitive glucose UME by covalently immobilizing GOD onto a 25 μm Pt UME modified with an enzyme immobilization matrix. This new sensor fabrication method preserves the geometric surface area of the UME, retaining maximum enzyme activity similar to that of the free enzyme, and is capable of performing a feedback approach curve. The new SECM probe was used to quantify the glucose consumption of S. mutans biofilms. Since the cariogenic behavior of S. mutans is directly related to their carbohydrate catabolism, it is important to study their carbohydrate uptake behavior. To study the preferential uptake of sugars by S. mutans, as in a normal diet composed of both simple and complex sugars, we quantified the bacterial glucose uptake in the presence of sucrose. The minimum interference of sucrose in the glucose sensor response allowed us to use the newly developed dual SECM probe for quantification of the rate of bacterial glucose consumption in the presence of sucrose in real time.

EXPERIMENTAL

For details about chemicals, instrumentation, fabrication of dual SECM probe with a pH sensor and the preparation of S. mutans biofilms, please see the relevant sections in the supporting information.

Fabrication and characterization of the dual SECM probe with an ultra-micro glucose sensor.

The f-MWCNT required for Pt UME modification was prepared according to a previously reported method⁸. Briefly, 30.0 mg of MWCNT was refluxed with 60.0 mL of concentrated HNO₃ acid at 80 °C for 72 h. The remaining acid was removed by filtration and the f-MWCNT was washed thoroughly with DI water to remove all acids before being vacuum dried for 12 h at 40 °C. Carboxylic functionalization of MWCNT was quantified by acid-base titration.

A dual-tip SECM probe that contains two ultra-micro Pt electrodes was fabricated. One Pt electrode surface was then selectively modified into a glucose microsensor. To fabricate the dual-tip SECM probe, we pulled an 8.0 cm long theta capillary (1.50 mm OD and 1.02 mm ID) with the capillary puller (Model P-30, Sutter Instrument Co., CA, USA, pulling parameters; Heat=700, Pull=650). The pulled end was sealed and a 1.0 cm long by 25 μm diameter Pt wire (Goodfellow Cambridge Ltd., UK.) was inserted into each barrel of the capillary and sealed under vacuum. Copper wires were inserted from the unsealed end and an electrical connection between the Pt wire and the copper wire was secured with silver epoxy (Epoxy technology, IL, USA.). The fabricated probe was polished to expose the Pt surface and further smoothed by sequential polishing on a polishing pad with 1.0 μm, 0.3 μm, and 0.05 μm alumina powder (BUEHLER, IL, USA). The polished electrode was tested by obtaining cyclic voltammetry in a solution of 1.0 mM ferrocene methanol solution in 100.0 mM KCl. To make the enzymatic sensor, we electrochemically etched one Pt electrode surface by applying a 5.0 V potential across a graphite rod in a solution of saturated CaCl_2:HCI:H₂O (60:4:36 v/v%) to obtain a 5.0 μm deep cavity^21,22. This cavity was thoroughly cleaned by sonicating, first in deionized water and then in 70% ethanol in a sonic bath, to remove any residual etching solution. The cavity was then dried under a stream of nitrogen gas. f-MWCNT:IL (10:90 w/w%) was thoroughly hand mixed in a ceramic dish to make an even paste. A portion of this freshly prepared paste was spread onto a glass slide and packed into the electrode cavity by gently tapping the electrode onto the paste. The packed electrode was then immersed in 10.0 mM phosphate buffer solution (PB) (pH = 6.0) containing GOD, 30.0 mM N-hydroxysuccinamide (NHS), and 15.0 mM 1-Ethyl-3-(3-dimethylaminopropyl) carbodiimide HCl (EDC) for 12 h at 4 °C for covalent coupling of enzymes onto the electrode surface. After 12 h, non-covalently attached enzymes were removed by gently washing the electrode with PB (pH=6.0) and stored in PB until use. Glucose sensor characterization was performed by the standard addition method.

Probing the bacterial glucose uptake with SECM.

Bacterial biofilms were prepared by immobilizing S. mutans with alginate hydrogel within a cavity of a polydimethylsiloxane (PDMS) mold made inside a glass-bottom Petri dish (figure 1 A).

Figure 1:

(A) Schematic representation showing the SECM probe positioned above the S. mutans -alginate biofilm inside the PDMS mold. (B) The SECM setup to perform the biofilm-related experiments.

For SECM experiments, the Petri dish with the PDMS mold containing the alginate bacteria gel was mounted inside the heating mantle on the SECM stage (figure 1 B). The CaCl₂ solution was replaced with artificial saliva (AS) solution containing 1.0 mM ferrocyanide. The probe-substrate distance was fixed by performing a negative feedback approach curve on the substrate with the Pt electrode of the SECM probe at room temperature (figure S-1). The ferrocyanide solution was then replaced with fresh AS solution after washing the biofilm with AS solution a few times to remove residual ferrocyanide. The temperature of the solution was first raised to 37 °C and then +0.5 V (vs Ag/AgCl in 1.0 M KCl) was applied to the enzymatic sensor to quantify the local glucose concentration consumed by bacteria. During the X and Z scans experiments, the tip was moved at a rate of 0.5 μm/s to preserve the steady state glucose concentration profile.

RESULTS AND DISCUSSION

Electrochemical characterization of the dual SECM probe and glucose microsensor.

The new SECM probe with ultra-micro glucose sensor was fabricated by selectively modifying one of the Pt UMEs of the dual tip. An optical microscopic picture of this probe with the Pt UME and glucose microsensor is shown in figure 2 A. After their fabrication, the new dual-tip probes were characterized by obtaining the negative feedback approach curves in 1.0 mM ferrocenemethanol solution. The unmodified Pt UME of the dual tip was used to obtain the feedback approach curve. Figure 2 B shows the experimental approach curve fitted with the theoretical approach curve.

Figure 2:

(A) Optical microscopic image of the new dual-tip SECM probe showing the Pt UME and the glucose sensor. (B) Negative feedback approach curve obtained with the new dual-tip SECM probe in 1.0 mM ferrocenemethanol solution fitted with the theoretical approach curve of the electrode with the same dimensions. (C) Schematics of the new SECM probe with the unmodified Pt UME and the glucose UME. (D) Schematic representation (not drawn to scale) of the covalently attached glucose oxidase enzyme on to the f-MWCNT exposed to the surface of the IL-f-MWCNT matrix packed into the etched cavity on the Pt UME.

The enzyme immobilization method plays a crucial role in the glucose sensor’s performance. It is essential to use an immobilizing method that does not significantly change the geometric surface area of the electrode, allows enzymes to adhere to it strongly and to use an immobilizing matrix that can load a high number of enzymes to yield a higher sensitivity. To address these requirements, we designed the new ultra-microsensor by covalently immobilizing the enzyme molecules onto the electrode surface through an immobilization matrix. One of the electrodes of the dual SECM probe was electrochemically etched and the resulting cavity was packed with a new enzyme immobilization matrix composed of IL and f-MWCNT. GOD was then covalently immobilized onto the UME. This glucose sensor fabrication approach preserves the geometric area of the electrode surface, protects enzyme activity, and provides a highly conductive enzyme immobilization matrix that preserves the electrical properties of the sensor. The schematics of the developed SECM probe with the glucose microsensor and an illustration of the enzyme immobilization approach are shown in figure 2 C and D.

Fabricated glucose sensors were calibrated by the standard addition method at +0.5 V vs the Ag/AgCl (1.0 M KCl) reference electrode in AS (pH=7.2) at room temperature (23 °C) and at 37 °C. As shown in equations 1 and 2, hydrogen peroxide (H₂O₂) is produced during the enzymatic oxidation of glucose and is electrochemically detected at the electrode surface. The resulting current from the electrochemical oxidation of hydrogen peroxide is directly proportional to the oxidized enzyme substrate concentration.

$$D-\mathit{\text{glucose}}+GOD\left(FAD\right)\to D-\mathit{\text{glucono}}-\delta -\mathit{\text{lactone}}+GOD\left(FAD{H}_2\right)$$ (1)

$$GOD\left(FAD{H}_2\right)+O_2\to GOD\left(FAD\right)+H_2{O}_2$$ (2)

The increase in current per added glucose concentration was normalized for the geometric surface area of the electrode. Figure 3 C and D show the calibration plots for the sensor at both temperatures. The lower detection limit of the new sensor was recorded as 10 μM. At 23 °C in the linear range from 100.0 μM to 3.0 mM, the sensitivity was 53.45 (±6.20) μA·mM⁻¹·cm⁻². As shown in Figure 3 C, at 37 °C, the current density plot shows two distinct linear ranges. From 100.0 μM to 1.0 mM, the sensor exhibits a sensitivity of 94.44 (±18.55) μA·mM⁻¹·cm⁻² and from 1.0 mM to 4.0 mM, a sensitivity of 60.0 (±2.86) μA·mM⁻¹·cm⁻². The increase in current density at higher temperatures is due to the increase in the enzyme turnover rate at 37 °C and a further decrease in charge transfer resistance in IL at higher temperatures⁸. When a steady-state current corresponding to glucose oxidation is obtained, the enzyme turnover rate is directly proportional to the current. Since enzymes are biological catalysts, they should follow Michaelis-Menten kinetics: two linear ranges in the calibration plot are due to the saturation of enzyme active sites with the substrate at higher substrate concentrations. To further confirm this, we obtained the Lineweaver-Burk plot by graphing 1/[Glucose] vs 1/Current. The plot showed a linear relationship (figure S-2) throughout the added glucose concentration range, up to 8.0 mM. The observed high sensitivity of the glucose microsensor confined to a 25 μm UME could be a contribution of both the immobilization matrix properties and the enzyme immobilizing strategy. The new matrix composed of f-MWCNT and IL is efficient in both electron transfer and peroxide oxidation, which gives a higher current signal for glucose oxidation. In addition, the calculated Michaelis-Menten constant for the immobilized GOD is 29.0 mM, which is similar to the values reported for free enzymes²⁰. This confirms that the activity of the immobilized enzyme corresponds to that of the free enzyme, demonstrating the excellence of the enzyme immobilizing method for use in sensor fabrication.

Figure 3:

(A) Optimization of the composition of the f-MWCNT-IL paste for immobilizing GOD at 23 °C in AS with pH=7.2 at +0.5 V vs Ag/AgCl reference electrode (B) Cyclic voltammograms obtained for Pt UME, f-MWCNT-IL-packed UME, and GOD-immobilized sensor in standard glucose solutions at 23 °C in AS with pH=7.2. (C) Calibration plots of the optimized glucose microsensor at 23 °C and at 37 °C in AS solutions with pH=7.2 at 0.5 V vs Ag/AgCl reference electrode. (D) Calibration of the glucose microsensor at lower concentrations at 23 °C and at 37 °C in AS solutions with pH=7.2 at 0.5 V vs Ag/AgCl reference electrode.

The new glucose microsensor quantifies glucose concentrations indirectly from the current produced by enzyme-generated hydrogen peroxide oxidation. Hence, the enzyme immobilization matrix should be efficient in hydrogen peroxide oxidation. MWCNT composites have been widely used for electrochemical hydrogen peroxide detection because of their excellent catalytic properties for peroxide oxidation and high electrical conductivity^8,23,24. MWCNTs were used as the GOD immobilizing matrix because their biocompatibility and electrochemical properties in biosensor design has already been proven^25,26. Commercially available pristine MWCNTs were chemically treated with nitric acid to introduce carboxylic functional groups that can immobilize enzymes via carbodiimide coupling. During functionalization, an average of 7.0 w/w% (n=2) carboxylic functional groups were obtained, allowing the immobilization of a significant number of enzyme molecules. Ionic liquid containing carbon composites have been successfully tested for electrochemical sensor fabrication because of their high conductivity and broad potential window 8,27. We mixed f-MWCNT with IL (1-butyl-4-methylpyridinium hexafluorophosphate) in a ratio of 90:10 (w/w%) into a paste and packed it into the cavity on the Pt UME surface. GOD was then covalently immobilized onto the carboxylic functional groups of the f-MWCMT that were exposed on the electrode surface. Hydrogen peroxide produced by GOD in the presence of glucose is oxidized and the resulting electrons are efficiently transferred through f-MWCNT and IL to the Pt electrode in the UME. The presence of π electrons in pyridine group-bearing IL (1-butyl-4-methylpyridinium hexafluorophosphate) further lowers the charge transfer resistance²⁷ and increases the sensitivity of the electrodes. This is an advantage when designing an UME, since it allows detection of very low concentrations of analytes. Since IL is the sole contributor to the binding strength of the matrix, it is important to maintain a high weight percentage of IL. Poor binding strength will cause leaching of the matrix components during experiments, affecting the sensitivity of the sensor. Therefore, when determining the optimum composition, we considered the pastes that contain more than 50% (w/w) IL. Figure 3 A shows the comparison of glucose sensor sensitivity upon changing the IL composition from 50% to 90% in the paste. Increasing the carboxylic functional groups by increasing the f-MWCNT percentage should theoretically increase the amount of immobilized enzyme molecules, thereby increasing the sensor sensitivity; however, according to the data shown in figure 3 A, IL had the greatest effect on increasing the sensitivity. During acid treatment, in addition to functionalization, more defective sites are also introduced to the MWCNT, decreasing their electrical conductivity²⁸. A lower f-MWCNT percentage and a higher IL percentage thus produces a highly conductive paste, resulting in a higher sensitivity of the sensor. Therefore, a 10% f-MWCNT-containing paste was selected as the optimum level for sensor fabrication. The amount of GOD immobilized onto the sensor is controlled by changing the enzyme activity of the coupling solution. During the optimization of paste composition, coupling solutions with 600 U/mL GOD activity were used and this was kept constant during further sensor fabrications.

Analytical performance of the glucose microsensor.

GOD from Aspergillus sp. is highly selective for ꞵ-D-glucose and this makes it ideal for use in the preparation of glucose sensors to quantify glucose in complex biological systems¹⁴. However, molecules such as sugars and proteins have the tendency to adsorb onto the electrode surface and can affect sensor performance. We thus used the new glucose sensor to quantify glucose consumption in sucrose containing AS (defined in the experiment section) solutions and to study the effect of sucrose and other possible interferences (figure S-3) in glucose sensor performance. The sequential addition of glucose followed by the sequential addition of sucrose up to 3.0 mM is shown in figure 4 A. No change in the glucose signal was observed with the addition of sucrose to the solution. Moreover, we calibrated the glucose sensor in the presence of 3.0 mM sucrose, the same sucrose concentration that we used for the biofilm experiments. Current density plots with and without sucrose are shown in figure 4 B; glucose sensor sensitivity did not change with the addition of sucrose.

Figure 4:

(A) Interference test: the glucose response of the new sensor with sequential addition of glucose followed by sucrose in AS with pH=7.2 at 37 °C. (B) Calibration plots obtained for a glucose microsensor in the presence of and absence of sucrose in AS with pH=7.2 at 37 °C. (C) Response time of the glucose microsensor. (D) Calibration plots of the glucose microsensor obtained before and after the SECM experiment in AS with pH=7.2.

Since the glucose biosensors have mainly been designed for use in SECM studies, the response time of the sensors should be taken into consideration. Figure 4 C shows the response time of a single glucose sensor with the addition of a known concentration of glucose in the bulk solution. The new sensor showed a fast response time of 1.91 (±0.33) s upon addition of glucose, proving its ability to be incorporated as an SECM probe. It is important that sensor response be preserved across SECM experiments with bacteria, as these experiments extend over a longer time. Therefore, sensors were calibrated before and after the SECM experiments to ensure that sensor performance remained unchanged; the change in current density before and after the SECM experiments was 10.05 (±3.97) %. The sensors retained about 90% of their initial sensitivity during the SECM experiments, proving their excellent stability during such studies.

Studying the glucose uptake behavior of S. mutans bacterial biofilms.

We immobilized the bacteria within a biocompatible alginate hydrogel matrix to make a simulated biofilm, as it was difficult to maintain a constant bacterial population with naturally grown S. mutans biofilm. Our laboratory has previously used alginate to prepare simulated biofilms with oral bacterial species, including S. mutans, and we have shown that bacteria remain metabolically active in alginate hydrogel²⁹.

Since glucose sensor performance is directly affected by environmental factors, specific attention was paid to other chemical changes in the local bacterial environment during carbohydrate catabolic processing. S. mutans is known to produce lactic acid during carbohydrate catabolism, decreasing the local pH and affecting sensor response. We used an iridium oxide deposited pH microsensor to evaluate the pH change in the local bacterial environment on addition of 1.0 mM glucose and 3.0 mM sucrose. The pH change caused by the bacteria depends on the initial pH and the buffering capacity of the AS. The average stimulated saliva pH of an adult has been reported to be 7.2³⁰ and, for the biofilm experiments, we used AS with pH 7.2 containing 30.0 mM HEPES as the principal buffering component. No significant change in pH above the biofilms was observed on addition of both sugars (figure S-4). The presence of other possible interferences such as depleted dissolved oxygen concentration and hydrogen peroxide in the local bacterial environment was also studied prior to glucose consumption experiments (figure S- 5). Their contribution to glucose sensor performance was minimal and assumed to be negligible.

We also investigated bacterial glucose consumption without sucrose. Figure 5 A shows the steady-state glucose concentration (0.242 mM) 50 μm above the biofilm after addition of 1.0 mM glucose. In the control experiment, where alginate gel was prepared without bacteria, no change in glucose concentration was observed. Therefore, the observed decreased glucose concentration is due to bacterial consumption. The average remaining glucose concentration observed 50 μm above the biofilm was 0.194 (± 0.057) mM, an 80.6% reduction from the bulk concentration. The rise in glucose concentration in the control alginate substrate indicates the time required to attain a homogenous glucose concentration across the Petri dish.

Figure 5:

(A) Steady-state glucose concentration observed 50 μm above the biofilm after adding 1.0 mM glucose concentration to the media at 37 °C in AS with pH=7.2. (B) Steady-state glucose concentration observed 50 μm above the biofilm after adding 1.0 mM glucose + 3.0 mM sucrose to the media at 37 °C in AS with pH=7.2. (C) Z-direction glucose concentration profile from 50 μm (z=0) to 1000 um distance above the biofilm surface with 1.0 mM glucose in AS with pH=7.2 at 37 °C. (D) X-direction glucose concentration 100 μm above the biofilm with 1.0 mM glucose in AS with pH=7.2 at 37 °C.

From bacterial consumption results, glucose depletion profiles were created in the local biofilm environment. Since many oral bacterial species other than S. mutans also use glucose, these glucose-depleted regions affect the metabolic behavior of other slow-growing, glucose- consuming bacterial species. Therefore, the 3D distribution of the glucose consumption profiles plays a key role in biofilm structure. To analyze the dimensions of the glucose consumption profile, we performed z and x scans under the same experimental conditions in the presence of 1.0 mM glucose 100 μm above the simulated biofilm. According to z scan data shown in figure 5 C, the profile extends up to about 600 μm from the biofilm surface. In figure 5 D, x scan data are given for three x-direction scans performed 100 μm above the biofilm surface after adding 1.0 mM glucose. Scan a (figure 5 D) was performed above the biofilm where the average diameter of the biofilm is about 2500 μm and the scans are 200 μm apart from each other. According to the observed glucose concentration change in x scans, the glucose consumption profile extends up to about >600 μm from the edge of the S. mutans-alginate biofilm along the x direction. The decreased remaining glucose concentration observed form scans b and c (figure 5 D) is due to the variability of the biofilm sample. The >600 μm glucose depletion zone surrounding the biofilm indicates the spatial distribution of the glucose profile, which in turn signifies the availability of nutrients for the adjacent bacterial species (figure 6).

Figure 6:

Schematics showing the glucose-depleted region created in the S. mutans local bacterial environment.

To investigate the bacterial glucose uptake behavior in the presence of sucrose, we performed the same experiment by introducing both 1.0 mM glucose and 3.0 mM sucrose at the same time (figure 5 B). The observed average glucose concentration 50 μm above the biofilm was 0.925 (± 0.072) mM. This is a 73% decrease in glucose uptake compared with experiments in which only glucose was introduced to the biofilms. S. mutans can excrete exoenzymes that catabolize sucrose into fructan and glucose (figure S-6). To check whether the higher remaining glucose concentration observed was due to extracellular sucrose catabolism, we measured the glucose sensor response by adding only 3.0 mM sucrose (figure S-7). No change was observed in glucose concentration 50 μm above the biofilm, indicating that no glucose contribution to the glucose signal originates from extracellular sucrose catabolism. This shows that, in a regular mixed diet, when simple glucose sugar is available, bacteria still prefer sucrose uptake. Extracellular sucrose catabolism by this species produces glucose and fructose, increasing their local concentrations (figure S-6). The uptake of both glucose and fructose mainly occurs through the phosphoenolpyruvate sugar phosphotransferase (PTS) system². With an abundance of sugars to uptake in the proximity of the bacterial membrane, there will be less uptake of added glucose and this could be why we observe less bacterial glucose uptake when sucrose is present.

CONCLUSION

A dual SECM probe with a new ultra-micro glucose sensor was fabricated by covalently immobilizing GOD onto an f-MWCNT-IL composite-modified 25 μm Pt UME. The enzyme immobilization matrix that we developed has the unique advantages of high electrical conductivity and efficient catalytic activity for peroxide oxidation. Moreover, covalent immobilization of the enzymes allowed us to preserve maximum enzyme activity close to that of the free enzyme. These features enabled us to achieve higher sensitivity for glucose detection while miniaturizing the size of the electrode. With this new enzyme electrode fabrication approach, the geometric surface area of the UME is preserved. This allowed us to use the UMEs in SECM scanning modes at close proximity to the electrode surface without damaging the sensor. The sensor also has a fast response time of ~ 2 s. Therefore, it is ideal for use in real-time measurements of bacterial biofilms or any type of living cells with SECM. In addition, sensor response was not affected by sucrose, allowing us to use it to probe S. mutans glucose uptake in the presence of sucrose. From this work, for the first time, glucose consumption of S. mutans was quantitatively measured in real time, along with the local distribution of the glucose consumption profile in the local bacterial environment. The glucose consumption data collected in the presence of sucrose reveals the preferential sucrose consumption behavior of S. mutans bacteria in a mixed diet composed of sucrose and simple sugars such as glucose. This information will open new pathways for molecular studies of this sugar uptake behavior in pathogenic bacteria. Detailed studies related to the effect of local sugar concentrations on S. mutans PTS activity and the regulation of sucrose catabolizing exoenzyme production are currently being undertaken in the lab.