Peptide-modified Microelectrode-based Potentiometric Device for 2,4,6-trinitrotoluene Molecule Detection

1Research and Development Center for Five-Sense Devices, Kyushu University, 744 Motooka, Nishiku, Fukuoka 819-0395, Japan 2Department of Information Science and Electrical Engineering, Kyushu University, 744 Motooka, Nishiku, Fukuoka 819-0395, Japan 3Department of Chemical Science and Engineering, Tokyo Institute of Technology, 2-12-1 O-Okayama, Meguro-ku, Tokyo 152-8552, Japan 4Institute for Advanced Study, Kyushu University, 744 Motooka, Nishiku, Fukuoka 819-0395, Japan


Introduction
Electrical sensing has been widely used for the detection of heavy metal ions, proteins, DNA, and some hazardous molecules. (1)(2)(3)(4)(5)(6) In particular, the development of electrical sensing devices for detecting hazardous explosive molecules has been rapid since such molecules have become a major threat to global security and also an important environmental issue because of the contamination of agricultural land and products and the threat to normal life. (7) To date, great progress has been made, and devices such as quartz crystal microbalance (QCM) and techniques including square wave voltammetry (SWV), electrochemical impedance spectroscopy (EIS), and colorimetric method have been used for the detection of nitroaromatic compounds in both liquid and vapor phases. (8)(9)(10)(11)(12) In particular, field-effect transistor (FET)based biosensors have been considered in the detection of nitrates and some small molecules because of their easy integration, low cost, and high throughput with the target receptor functionalized on the semiconducting layers. (2,4,5) FET-based potentiometric biosensing detects the concentration of biomolecules captured on a semiconducting gate electrode, where the charged biomolecules change the carrier concentration. (2,13) An extended gate-FET has been successfully fabricated for the detection of protein biomolecules (e.g., lysozyme and thrombin) and the influenza virus, realizing a platform for a simple and rapid diagnostic system by measuring the change in potential. (14)(15)(16)(17) Therefore, potentiometric biosensing methods are promising for various applications in the near future.
In this study, a newly fabricated 2,4,6-trinitrotoluene (TNT) recognition peptide TNTHCDR3-modified microelectrode-based potentiometric device was used for the rapid, specific, and real-time detection of TNT molecules. The TNT recognition peptide was synthesized in accordance with our previous work. (18,19) The mechanism for TNT explosive detection is based on the change in potential, which is mainly caused by the charge transfer between the positively charged group in TNT molecules and the negatively charged TNTHCDR3 peptide during the binding interaction. The sensitivity, selectivity, and stability of the peptide-based potentiometric device were investigated to evaluate its performance.

Experimental Procedure
The details of the microelectrode-based potentiometric device fabrication procedure were described in a previous work. (16) Briefly, 10 disc-shaped gold microelectrodes (500 μm in diameter) were connected to an electrometer/high-resistance meter (Keithley, Cleveland, OH) to monitor the change in voltage. The measurement system is illustrated in Figs. 1(a) and 1(b). The flow system measurement procedure is described as follows. The potentiometric sensor chip was firstly anchored to the sensing unit and placed in an incubator with the temperature fixed at 25 ℃ to prevent interference from the environment. Then, the explosive sample (250 µL for each measurement) was injected into the flow system through the sample injector and pumped to pass it through the microelectrode area (sensing unit) where the explosive recognition peptide was covalently immobilized on the surface. The system running buffer, 0.01 M phosphatebuffered saline (PBS, 0.138 M NaCI, 0.0027 M KCI, pH 7.4) containing 0.05%Tween 20 (T) and 5% DMF (N,N-dimethylformamide) (PBST in 5% DMF), with a nitrogen flow was chosen to stabilize the flow system (flow rate: 500 µL/min) to prevent environmental interference. When the explosive sample arrived at the microelectrode area, the change in electrical potential at the microelectrode was monitored using our laboratory-built data analysis system. After the reaction, the sample was pushed out by running buffer to a waste well. Each measurement cycle (reaction time of 400 s) started when the explosive samples began to react with the peptidebased gold surface and ended when the electrical potential returned to the baseline.
The preparation and characterization of the sensor chip surface are shown in Fig. 1(c) and described as follows. The PEG6-COOH aromatic dialkanethiol selected as a self-assembled monolayer (SAM) reacted with the gold surface through the formation of a strong Au-S bond similarly in our previous studies. (20,21) Wettability measurements showed that the water contact angle changed (61.0 ± 1.0° to 35.2 ± 0.9°) as a result of SAM treatment. The SAM on each microelectrode was formed by immersing the potentiometric device chip in 1 mM PEG6-COOH aromatic dialkanethiol overnight. The surface was rinsed with ethanol and Milli-Q water at least three times. Then, a mixture solution of 1-ethyl-3-(3-dimenthylaminopropyl) carbodiimide (EDC) and N-hydroxy succinimide (NHS) was added to activate and convert the carboxyl terminal groups to NHS esters. In accordance with a previous study, the 15-mer amino acid sequence (ARGYSSFIYWFFDFC) obtained from the complementarity-determining region heavy chain (HCDR) in the anti-TNT monoclonal antibody was determined as the TNT recognition peptide (named TNTHCDR3), whereas the TNTLCDR3 (sequence: CLQHYSAPYTC) peptide derived from the light chain of the complementarity-determining region (LCDR) in the anti-TNT monoclonal antibody was demonstrated to show no appreciable binding to the TNT explosive. (18) 1000 ppm TNT recognition peptide was immobilized through the amine coupling method to obtain the highest surface density for TNT binding, which was followed by blocking with ethanolamine solution. The surface morphology with/without the anchored peptide was evaluated by atomic force microscopy (AFM) as shown in Fig. 1(d). A control experiment was performed by immobilizing the control peptide TNTLCDR3 on the gold surface by the same method. 1.9, 15.7, 31.4, 62.7, 125.4, and 250.8 ppm TNT solutions were prepared in PBST-5% DMF buffer solution, each of which was injected into the flow system three times to perform three cycles of measurements. The change in potential was recorded when TNT molecules were captured by the recognition peptide in real time. The sensor surface was regenerated to the baseline potential by running buffer (PBST-5% DMF) after each measurement.

Results and Discussion
The results in Figs. 2(a) and 2(b) show a strong linear relationship between the change in potential and the concentration (1.9 to 62.7 ppm with correlation coefficient R 2 = 0.9900) of TNT samples. The sensor started to saturate at a concentration higher than 62.7 ppm according to the curve. The limit of detection (LOD) was calculated to be 600 ppb (3σ in background solution). Table 1 shows the results of various studies of using different peptides as TNT receptors for sensitive and selective detection. A higher sensitivity was obtained in the present work than in our previous studies using the same peptide-based surface plasmon resonance (SPR) sensor as well as similar sensitivity to other reported methods. The results in Fig. 3 show the changes in the potentials of a TNTLCDR3-functionalized potentiometric chip and a TNTHCDR3-functionalized chip for TNT solutions with concentrations of 250.8 and 501.5 ppm. The TNTLCDR3-anchored potentiometric device exhibited a negligible response (equal to the background response) for both TNT concentrations, whereas the TNTHCDR3 peptideanchored chip exhibited a significantly greater response than the control chip. Therefore, the interaction mechanism between the TNT molecule and the TNTHCDR3 recognition peptide  was investigated. In previous studies, a computational model predicted that the TNT molecule binds to the peptide named antennal specific protein-1 (P1ASP1) from the honeybee through a hydrogen bond with tryptophan (Trp 17 ) amino acid (W) and to a single-walled carbon nanotube (SWCNT) through π-π interaction on the surface. The sequence Trp-Phe-Val-Ile (WFVI) in P1ASP1 at the C-terminal plays an important role in the reaction with the TNT molecule. (10) The highly stable polymer structures designed to recognize TNT exhibited the shape and electrostatic complementarity between the negatively charged region of triphenylene and the positively charged region of TNT, where electrostatic potentials were calculated. (22) As a commonly accepted mechanism, the aromatic rings of TNT are electron-deficient, which enables strong interactions with the electron-rich amino groups of 3-aminopropyltriethoxysilane (APTES). This would result in the formation of a Meisenheimer complex because of the charge transfer from amino groups to aromatic rings. (23) The TNT-binding peptide core amino sequence WHW was synthesized through the phage display method. (24,25) In this study, the net charge of the TNT recognition peptide TNTHCDR3 (ARGYSSFIYWFFDFC) was negative because of the isoelectric point (pI = 5.7) in the buffer solution (pH 7.4). On the basis of the above mechanism, the change in potential was mainly caused by the charge transfer between the positively charged group in the TNT molecule and the negatively charged peptide TNTHCDR3 when the TNT molecule and peptide interacted via the π-π interaction (WFY) and a hydrogen bond, which is illustrated in Fig. 1(e). RDX, 2, 6-DNT, DNP-glycine, and NB-gly-gly have similar molecular structures to the TNT molecule. These four TNT analogues were used to investigate the selectivity of the rationally designed peptide. The TNTHCDR3 peptide-modified sensor exhibited a significant response to TNT and a background-level response to the four TNT analogues (Fig. 4), indicating the specific interaction between TNT and the peptide and hardly any nonspecific binding to the four TNT analogues. The stability of the peptide-based potentiometric device response was also investigated. Figure 5 shows the binding response for the same sensor chip over one week. The response to 125.4 ppm TNT explosive was about 96.0% of the original value after 4 days and 77.9% after one week, indicating that the retention of the binding ability to TNT explosive was acceptable. The repeatability of the response of the peptide-based potentiometric device was calculated with a standard deviation of 0.85 (n = 5) at a TNT concentration of 31.4 ppm, which showed good repeatability of the sensor response.

Conclusion
We have demonstrated a TNT recognition peptide-functionalized potentiometric device for sensitive and highly selective TNT molecule detection in an aqueous solution. The LOD is expected to be improved to meet the need for the detection of ultralow-concentration traces of TNT in the future.