Label-Free Electrochemical Detection of S. mutans Exploiting Commercially Fabricated Printed Circuit Board Sensing Electrodes

This paper reports for the first time printed-circuit-board (PCB)-based label-free electrochemical detection of bacteria. The demonstrated immunosensor was implemented on a PCB sensing platform which was designed and fabricated in a standard PCB manufacturing facility. Bacteria were directly captured on the PCB sensing surface using a specific, pre-immobilized antibody. Electrochemical impedance spectra (EIS) were recorded and used to extract the charge transfer resistance (Rct) value for the different bacteria concentrations under investigation. As a proof-of-concept, Streptococcus mutans (S. mutans) bacteria were quantified in a phosphate buffered saline (PBS) buffer, achieving a limit of detection of 103 CFU/mL. Therefore, the proposed biosensor is an attractive candidate for the development of a simple and robust point-of-care diagnostic platform for bacteria identification, exhibiting good sensitivity, high selectivity, and excellent reproducibility.


Introduction
Bacteria comprise a ubiquitous type of microorganism, involved in numerous clinical, environmental, and industrial phenomena; indicative examples include infectious diseases, food and water safety, and insulin synthesis [1]. In all of the applications of this vast spectrum, there is the need to identify and quantify the presence of the bacteria with high sensitivity and specificity, ideally in real-time. The detection (i.e., quantification range, limit of detection) and usability (i.e., time to result, cost) requirements for each case vary, but nonetheless underpin in a similar manner all of them. For this reason, a multitude of research groups around the world have been proposing various bacterial detection technologies as alternatives to laborious standard protocols such as bacterial culture. For example, matrix-assisted laser desorption/ionization-time-of-flight MALDI-TOF mass spectrometry of culture supernatants, nucleic acid based methods (i.e., fluorescence in situ hybridization, DNA microarrays, real-time polymerase chain reaction (PCR). In all of the aforementioned cases though, the assay completion is laborious, requiring several sample preparation steps in order to achieve efficient quantification. Optical quantification methods, such as surface plasmon resonance imaging (SPRi) have also been reported, showing promising limits of detection in liquid culture media, but has also yet to be demonstrated in bacterial multiplication-free systems [2]. samples which could be detected down to 10 3 CFU/mL. Several groups have demonstrated S. mutans detection using different biosensing methods [14][15][16][17][18][19][20]. While capable of rapid measurements, these sensors require complicated detection protocols which may not be ideally suited for point-of-care diagnosis. Here we report for the first time a PCB implemented label-free electrochemical S. mutans detection platform using a very simple detection protocol. Streptococcus mutans was detected on the PCB sensing surface using a specific antibody which was pre-immobilized on the PCB. Electrochemical impedance spectrometry (EIS) was used to record the charge transfer resistance (R ct ) value with different S. mutans concentrations.

Preparation of Bacterial Cultures
Brain heart infusion (BHI) broth was prepared by dissolving 3.7 g of the BHI base in 100 mL of distilled water. BHI plates were prepared by dissolving 5.2 g of the BHI agar in 100 mL distilled water. Luria broth was prepared by dissolving 2 g of Luria broth base in 100 mL of distilled water. Luria agar (LA) plates were prepared by 3.5 g of the LA powder to 100 mL of distilled water.
To prepare Streptococcus mutans cultures, 10 µL of the revived MTCC890 strain were first plated on BHI agar plates and allowed to grow overnight at 37 • C. From these plates, overnight liquid cultures were set up in fresh BHI broth. It was further diluted 400 times and incubated at 37 • C and 237 rpm for 48 h. 3 mL of the liquid culture was centrifuged at 15,000 rpm for 10 min to obtain a bacterial pellet. The pellet was washed thrice with warm 1X PBS (pH 7.4). The final washed pellet was diluted in 2 mL of 0.01 mol/L PBS (pH 7.4) and used for ELISA experiments. For the electrochemical detection, 6 mL of the primary culture was centrifuged and washed three times as previously described. The washed pellet was then diluted with 1 mL of 0.01 mol/L PBS (pH 7.4). We then serially diluted this bacterial suspension to prepare different concentrations of bacteria.
To determine the specificity of our sensor to S. mutans, we used E. coli (MG1655) as a negative control. We plated 10 µL of the revived E. coli on LA plates and incubated them overnight at 37 • C. From these plates, an overnight culture was set up. From the overnight culture, a concentration of 10 10 CFU/mL was prepared in 2 mL of 0.01 mol/L PBS (pH 7.4). A high concentration of E. coli was chosen to test the binding specificity of the prepared electrodes.

Thiolation of Anti-Streptococcus mutans Primary Antibody
The specificity of the anti-Streptococcus mutans polyclonal antibody was determined by indirect ELISA, as described in the supplementary material. The primary antibody was thiolated (-SH) as previously reported with minor modifications [21,22]. Briefly, 1 mL of 100 µg/mL anti-Streptococcus mutans IgG was incubated in a solution of Traut's reagent in PBS containing 2 mM EDTA for 1 h at room temperature with gentle agitation. A 10-fold molar excess of Traut's reagent per mol antibody was used to ensure full thiolation to the lysine side chains of IgG. Excess (unconjugated) Traut's reagent was removed by centrifugation for 30 min at 10,000 rpm. Thiolated anti-Streptococcus mutans IgG was dissolved in 1 mL PBS (pH 7.4) and used immediately for sensor immobilization.

Design of Printed Circuit Board Sensing Electrodes for S. Mutans Detection
The sensing electrodes were designed in PCB CAD software (version 17.1.9, Altium ® ) and commercially fabricated in a standard PCB manufacturing facility (Lyncolec Ltd, Poole, UK). The copper electrodes were electroplated with a hard-gold finish in order to exploit the pore-free deposition and low contact resistance achieved by this technique [23]. The gold-plated electrodes were exploited as working, counter, and reference electrodes and connected to a pocketstat (Ivium, Netherland) to record the signals.

Sensor Fabrication
The gold-plated PCB electrodes were cleaned prior to anti-Streptococcus mutans IgG immobilization by ultrasonication in acetone, ethanol, and water, respectively for 15 min followed by 30 min ultrasonication in a solution containing 5:1:1 water, ammonium hydroxide (20%), and hydrogen peroxide (30%) [24,25]. Immobilization of the anti-Streptococcus mutans IgG to the working electrode was carried out by incubating 10 µg/mL of thiolated anti-Streptococcus mutans IgG solution for 1 h at room temperature, followed by thorough rinsing with PBS, and drying with purified N 2 gas. A volume of 5 µL of the thiolated antibody was employed for spotting it on the surface via manual drop casting on the working electrode surface, assuring all liquid was in contact with the surfaces of the other two electrodes. To minimize nonspecific binding and enhance the stability of the immobilized antibody, 1% bovine serum albumin (BSA) dissolved in PBS solution was incubated on the antibody-immobilized electrode for 30 min at room temperature, followed by rinsing twice with PBS and drying with purified N 2 gas. The prepared electrodes were stored at 4 • C [26][27][28][29].

Experimental Setup and Electrochemical Measurements
Cultures of S. mutans bacteria were serially diluted in PBS to obtain different concentrations, which were subsequently used for EIS measurements without any further processing. 50 µL of bacteria spiked in PBS was dispensed onto the sensor and incubated for 30 min at 4 • C, followed by thorough rinsing with PBS and drying with purified N 2 gas. EIS measurements were performed immediately after bacteria incubation using a Helios electrochemical potentiostat (Ivium pocketstat, Netherlands), connecting the three PCB sensing electrodes through wires and a commercial Peripheral Component Interconnect (PCI) express connector. A micromachined Teflon tape was adhered on the board around the sensing area, in order to separate one electrochemical cell on the PCB from another, and locally confine the reagents. The impedance spectra were recorded using gold as counter and pseudo-reference electrodes at an open circuit potential (OCP) in the frequency range of 100 kHz to 0.5 Hz, with a 60 mV amplitude in 50 µL of 100 mM PBS (pH 7.4) containing 2 mM of the [Fe(CN) 6 ] 3−/4− redox couple [30]. All electrochemical data were obtained at room temperature (25 • C). The cyclic voltammograms were performed in a three-electrode configuration with cycling the potential between −0.3 V and 0.4 V (scan rate: 0.05 V·s −1 ). Figure 1 schematically illustrates the label-free S. mutans detection protocol on the PCB sensor surfaces. The thiolated capture anti-Streptococcus mutans IgG was immobilized effortlessly on the PCB surface because of the strong thiol-Au interaction. Non-specific binding on the sensor surface was minimized by BSA, allowing the functionalized sensor to selectively bind with S. mutans. The electrochemical impedance spectra were recorded at each step. The complete set up of the electrochemical detection of bacteria using pocketstat is shown in Figure 2. Figure 3 shows typical EIS spectra of bare PCB, antibody-immobilized PCB, and antibody-and BSA-modified PCB surfaces, respectively, in 0.1 M PBS containing 2 mM K 4 Fe(CN) 6 and 2 mM K 3 Fe(CN) 6 . As seen in Figure 3, the charge transfer resistance (R ct ) dramatically increases from 6.26 kΩ to 27.19 kΩ, confirming that the antibody was immobilized on the sensor surface. The R ct value was further increased to 36.49 kΩ when BSA was incubated on the antibody-modified surface, confirming that BSA successfully blocked the unoccupied surface. The R ct values for the ferri/ferrocyanide are increased with successive protein layer formation on the sensing surface because the protein coverage at the PCB Au surface hindered the rate of electron transfer to the Au, causing the observed increase in the R ct .

Description of Label-Free S. mutans Detection Scheme
Micromachines 2018, 9, x FOR PEER REVIEW 5 of 9 respectively, in 0.1 M PBS containing 2 mM K4Fe(CN)6 and 2 mM K3Fe(CN)6. As seen in Figure 3, the charge transfer resistance (Rct) dramatically increases from 6.26 kΩ to 27.19 kΩ, confirming that the antibody was immobilized on the sensor surface. The Rct value was further increased to 36.49 kΩ when BSA was incubated on the antibody-modified surface, confirming that BSA successfully blocked the unoccupied surface. The Rct values for the ferri/ferrocyanide are increased with successive protein layer formation on the sensing surface because the protein coverage at the PCB Au surface hindered the rate of electron transfer to the Au, causing the observed increase in the Rct.   respectively, in 0.1 M PBS containing 2 mM K4Fe(CN)6 and 2 mM K3Fe(CN)6. As seen in Figure 3, the charge transfer resistance (Rct) dramatically increases from 6.26 kΩ to 27.19 kΩ, confirming that the antibody was immobilized on the sensor surface. The Rct value was further increased to 36.49 kΩ when BSA was incubated on the antibody-modified surface, confirming that BSA successfully blocked the unoccupied surface. The Rct values for the ferri/ferrocyanide are increased with successive protein layer formation on the sensing surface because the protein coverage at the PCB Au surface hindered the rate of electron transfer to the Au, causing the observed increase in the Rct.

Preparation of S. mutans Cultures for Electrochemical Detection
In order to determine the concentration of bacteria in our electrochemical test samples, the primary culture was serially diluted by a factor of 10 and the colonies were counted. Ten microliters of each solution were plated on solid agar plates (TYCSB) and incubated for 48 h to allow the growth of bacteria colonies. From counting the number of colonies, the concentration of the primary culture was found to be approximately 1.68 × 10 9 col/mL.

Quantification of S. mutans Detection via Impedimetric Measurement
In order to validate the antibody binding to S. mutans, an indirect ELISA experiment was performed as described in the supplementary materials ( Figures S1 and S2). The ELISA data confirmed that the primary antibody binds to our strain of S. mutans successfully. Hence, we proceeded with the functionalization of our sensor via immobilization of the antibodies on the working electrode surface, and recording of the respective EIS spectra.
The Rct is the key parameter associated with the binding of the target to the capture antibody on the sensor surface, and hence, to the concentration of the S. mutans in the sample solutions. Rct can be conveniently extracted from the EIS spectrum by either direct analysis of the spectrum or by fitting the spectrum to the Randles equivalent circuit (Figure 4a (inset)). The Randles equivalent circuit is composed of solution resistance (RS) in series with the parallel integration of the double-layer capacitance (Cdl) and the charge transfer resistance (Rct) [31]. The Warburg impedance is known to be represented as the straight line with a 45° phase angle and is closely associated with the diffusion of the redox species in solution; if the Warburg impedance element is absent and the Rct values of the sensor surfaces are significantly large, this implies that a significantly large amount of a charge transfer impeding material is found on the modified sensor. The Rct of redox species is modulated by their inherent electron transfer rate and the presence of any charge transfer-impeding material on the electrode surface that the redox species must penetrate to reach the electrode surface, thus increasing the magnitude of the Rct. Impedimetric measurements were carried out after 60 µL of the redox couple was injected. Figure 4a shows representative Rct values obtained from the PCB-implemented sensor surfaces treated with different concentration of S. mutans bacteria between 10 3 CFU/mL and 10 10 CFU/mL. Figure 4b presents the corresponding calibration plot of S. mutans detection, demonstrating the anticipated increase of Rct with increasing S. mutans concentration. This assay exhibits a lower limit of detection (LOD) of 10 3 CFU/mL. Three repetitions were done on the same PCB sensing surface. We observed a nonlinear behavior in the calibration curve. The signal was slowly increased

Preparation of S. mutans Cultures for Electrochemical Detection
In order to determine the concentration of bacteria in our electrochemical test samples, the primary culture was serially diluted by a factor of 10 and the colonies were counted. Ten microliters of each solution were plated on solid agar plates (TYCSB) and incubated for 48 h to allow the growth of bacteria colonies. From counting the number of colonies, the concentration of the primary culture was found to be approximately 1.68 × 10 9 col/mL.

Quantification of S. mutans Detection via Impedimetric Measurement
In order to validate the antibody binding to S. mutans, an indirect ELISA experiment was performed as described in the supplementary materials ( Figures S1 and S2). The ELISA data confirmed that the primary antibody binds to our strain of S. mutans successfully. Hence, we proceeded with the functionalization of our sensor via immobilization of the antibodies on the working electrode surface, and recording of the respective EIS spectra.
The R ct is the key parameter associated with the binding of the target to the capture antibody on the sensor surface, and hence, to the concentration of the S. mutans in the sample solutions. R ct can be conveniently extracted from the EIS spectrum by either direct analysis of the spectrum or by fitting the spectrum to the Randles equivalent circuit (Figure 4a (inset)). The Randles equivalent circuit is composed of solution resistance (R S ) in series with the parallel integration of the double-layer capacitance (C dl ) and the charge transfer resistance (R ct ) [31]. The Warburg impedance is known to be represented as the straight line with a 45 • phase angle and is closely associated with the diffusion of the redox species in solution; if the Warburg impedance element is absent and the R ct values of the sensor surfaces are significantly large, this implies that a significantly large amount of a charge transfer impeding material is found on the modified sensor. The R ct of redox species is modulated by their inherent electron transfer rate and the presence of any charge transfer-impeding material on the electrode surface that the redox species must penetrate to reach the electrode surface, thus increasing the magnitude of the R ct . Impedimetric measurements were carried out after 60 µL of the redox couple was injected. Figure 4a shows representative R ct values obtained from the PCB-implemented sensor surfaces treated with different concentration of S. mutans bacteria between 10 3 CFU/mL and 10 10 CFU/mL. Figure 4b presents the corresponding calibration plot of S. mutans detection, demonstrating the anticipated increase of R ct with increasing S. mutans concentration. This assay exhibits a lower limit of detection (LOD) of 10 3 CFU/mL. Three repetitions were done on the same PCB sensing surface. We observed a nonlinear behavior in the calibration curve. The signal was slowly increased with the lower S. mutans concentration, and the increment was much faster when the concentration of S. mutans was near 10 8 CFU/mL. The protein layer coverage on the PCB Au surface is the key factor for the R ct value change during the impedance measurement. We assume that the sudden jump behavior of the R ct from 10 8 to 10 9 was because of more potential protein layer formation on the Au surface that effectively hindered the electron transfer rate to the Au. Finally, we observed a saturation of the impedance signal when the concentration of S. mutans reached near 10 10 CFU/mL. with the lower S. mutans concentration, and the increment was much faster when the concentration of S. mutans was near 10 8 CFU/mL. The protein layer coverage on the PCB Au surface is the key factor for the Rct value change during the impedance measurement. We assume that the sudden jump behavior of the Rct from 10 8 to 10 9 was because of more potential protein layer formation on the Au surface that effectively hindered the electron transfer rate to the Au. Finally, we observed a saturation of the impedance signal when the concentration of S. mutans reached near 10 10 CFU/mL.

Specificity Study of the Developed Sensor
The specificity of this sensor was studied by performing measurements on PBS buffer samples spiked with E. coli, which is a Gram-negative, facultative anaerobic, rod-shaped, coliform bacterium of the genus Escherichia that is commonly found in the lower intestine of warm-blooded organisms. As shown in Figure 5, the Rct value of 10 10 CFU/mL E. coli modified sensor surface is similar to that generated from the non-spiked sample, which was used as a blank control. In contrast, the Rct value from the sample containing 10 6 CFU/mL S. mutans is significantly higher than that of E. coli, suggesting that our assay is highly specific to S. mutans.

Conclusions
In this paper we report the first PCB implemented label-free electrochemical bacterial biosensor. The thiolated capture antibody was successfully immobilized on a PCB sensing surface and we demonstrated that we can detect S. mutans selectively at a LOD of 10 3 CFU/mL. This work suggests

Specificity Study of the Developed Sensor
The specificity of this sensor was studied by performing measurements on PBS buffer samples spiked with E. coli, which is a Gram-negative, facultative anaerobic, rod-shaped, coliform bacterium of the genus Escherichia that is commonly found in the lower intestine of warm-blooded organisms. As shown in Figure 5, the R ct value of 10 10 CFU/mL E. coli modified sensor surface is similar to that generated from the non-spiked sample, which was used as a blank control. In contrast, the R ct value from the sample containing 10 6 CFU/mL S. mutans is significantly higher than that of E. coli, suggesting that our assay is highly specific to S. mutans. with the lower S. mutans concentration, and the increment was much faster when the concentration of S. mutans was near 10 8 CFU/mL. The protein layer coverage on the PCB Au surface is the key factor for the Rct value change during the impedance measurement. We assume that the sudden jump behavior of the Rct from 10 8 to 10 9 was because of more potential protein layer formation on the Au surface that effectively hindered the electron transfer rate to the Au. Finally, we observed a saturation of the impedance signal when the concentration of S. mutans reached near 10 10 CFU/mL.

Specificity Study of the Developed Sensor
The specificity of this sensor was studied by performing measurements on PBS buffer samples spiked with E. coli, which is a Gram-negative, facultative anaerobic, rod-shaped, coliform bacterium of the genus Escherichia that is commonly found in the lower intestine of warm-blooded organisms. As shown in Figure 5, the Rct value of 10 10 CFU/mL E. coli modified sensor surface is similar to that generated from the non-spiked sample, which was used as a blank control. In contrast, the Rct value from the sample containing 10 6 CFU/mL S. mutans is significantly higher than that of E. coli, suggesting that our assay is highly specific to S. mutans.

Conclusions
In this paper we report the first PCB implemented label-free electrochemical bacterial biosensor. The thiolated capture antibody was successfully immobilized on a PCB sensing surface and we demonstrated that we can detect S. mutans selectively at a LOD of 10 3 CFU/mL. This work suggests

Conclusions
In this paper we report the first PCB implemented label-free electrochemical bacterial biosensor. The thiolated capture antibody was successfully immobilized on a PCB sensing surface and we demonstrated that we can detect S. mutans selectively at a LOD of 10 3 CFU/mL. This work suggests that it may be possible to exploit PCB biosensors to potentially profile multiple bacteria in completely untreated samples, such as saliva. The proposed PCB which is designed and fabricated in a standard PCB manufacturing facility shows promising bacteria detection results which could be particularly useful for point-of-care and field uses in developing countries.