Electron-Transfer Mediator for a NAD-Glucose Dehydrogenase-Based

Nov 7, 2013 - In the present study, a new electron-transfer mediator was designed and synthesized for the application selectively to the NAD–GDH sys...
0 downloads 26 Views 1MB Size
Download PDF
Article pubs.acs.org/ac

Electron-Transfer Mediator for a NAD-Glucose Dehydrogenase-Based Glucose Sensor Dong-Min Kim,† Min-yeong Kim,† Sanapalli S. Reddy,† Jaegeol Cho,‡ Chul-ho Cho,‡ Suntae Jung,‡ and Yoon-Bo Shim*,† †

Department of Chemistry and Institute of Biophysio Sensor Technology (IBST), Pusan National University, Busan 609-735, South Korea ‡ Sensing & Interaction Lab, DMC R&D Center, Samsung Electronics Company, Ltd., Gyeongi-do 443-742, South Korea ABSTRACT: A new electron-transfer mediator, 5-[2,5-di (thiophen-2-yl)1H-pyrrol-1-yl]-1,10-phenanthroline iron(III) chloride (FePhenTPy) oriented to the nicotinamide adenine dinucleotide-dependent−glucose dehydrogenase (NAD−GDH) system was synthesized through a Paal− Knorr condensation reaction. The structure of the mediator was confirmed by Fourier-transform infrared spectroscopy, proton and carbon nucler magnetic resonance spectroscopy, and mass spectroscopy, and its electrontransfer characteristic for a glucose sensor was investigated using voltammetry and impedance spectroscopy. A disposable amperometric glucose sensor with NAD−GDH was constructed with FePhenTPy as an electron-transfer mediator on a screen printed carbon electrode (SPCE) and its performance was evaluated, where the addition of reduces graphene oxide (RGO) to the mediator showed the enhanced sensor performance. The experimental parameters to affect the analytical performance and the stability of the proposed glucose sensor were optimized, and the sensor exhibited a dynamic range between 30 mg/dL and 600 mg/dL with the detection limit of 12.02 ± 0.6 mg/dL. In the real sample experiments, the interference effects by acetaminophen, ascorbic acid, dopamine, uric acid, caffeine, and other monosaccharides (fructose, lactose, mannose, and xylose) were completely avoided through coating the sensor surface with the Nafion film containing lead(IV) acetate. The reliability of proposed glucose sensor was evaluated by the determination of glucose in artificial blood and human whole blood samples.

D

Both glucose oxidase (GOx) and glucose dehydrogenase (GDH) are used for the preparation of enzyme-based glucose sensors. Of the two, GOx is the most commonly used enzyme in commercial glucose sensors. However, GOx-based glucose sensors are susceptible to the oxygen concentration in the measuring media. As an alternative to GOx, GDH has been used in glucose sensors. The GDH may contain one of three cofactors: pyrroloquinoline quinone (PQQ),6 nicotinamide adenine dinucleotide (NAD),7 and flavin adenine dinucleotide (FAD).8,9 Among these, PQQ-GDH has the disadvantage of low selectivity caused by the oxidation of a variety of saccharides such as mannose, maltose, lactose, etc.10 FAD− GDH-based sensors also have disadvantages, such as a complicated preparation process, high-cost, and difficult operation. Although the NAD−GDH type exhibits higher substrate selectivity and stability than PQQ−GDH, finding an adequate electron-transfer mediator for a NAD−GDH system is difficult. Thus, developing a new mediator for NAD−GDH to provide improved selectivity is crucial.

iabetes mellitus is a serious metabolic disease, and the number of diabetes patients has continuously increased despite advances in modern medical science and technology.1,2 To prevent and reduce complications from diseases associated with diabetes mellitus, close monitoring of blood glucose levels is required. There have been numerous studies regarding glucose sensors, but patients still demand stress-free use and more reliable sensors.3 Of the various glucose sensors, the performances of enzyme-based glucose sensors have been extensively studied since Clark and Lyons first introduced them in 1962. The historical development of biosensors can be categorized into three generations.4 The first generation biosensor is dependent on the use of the natural oxygen cosubstrate enzyme and the detection of hydrogen peroxide. In the second generation, the limitations of first generation biosensors were overcome by using artificial mediators. Finally, the third generation allowed for direct electrical communication between the enzyme and the electrode surface. Today, most commercially available glucose sensors are considered second generation.5 The second generation has been improved by enhanced electron transfer between the redox center of the enzyme and the electrode surface through an artificial mediator, which is an important advantage for the performance of the sensor. © XXXX American Chemical Society

Received: October 7, 2013 Accepted: November 7, 2013

A

dx.doi.org/10.1021/ac403217t | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

industrial, Korea). Cyclic voltammograms (CVs) and amperograms were recorded using a potentiostat/galvanostat, Kosentech model PT-1 (South Korea). Electrochemical impedance spectroscopy was performed using an EG&G Princeton Applied Research PARSTAT 2263 at a given potential from 100 kHz to 100 mHz at a sampling rate of five points-per-decade. Synthesis of Mediator. Synthesis of 1,4-di(2-thienyl)-1,4butanedione. The starting material, 1,4-di(2-thienyl)-1,4butanedione was synthesized according to a method described in the literature.22 A solution of thiophene (9.6 mL, 0.12 mol) and succinyl chloride (5.5 mL, 0.05 mol) in CH2Cl2 (100 mL) was added dropwise to a suspension of AlCl3 (16 g, 0.12 mol) in CH2Cl2 (15 mL). The mixture was stirred at 18−20 °C for 4 h and poured onto ice. Then, concentrated HCl (5 mL) was added and the resulting dark green organic phase was washed with concentrated NaHCO3 (25 mL) and dried over MgSO4. After the solvent was evaporated, the blue-green solids remained and were suspended in ethanol. The mixture was filtered and washed three times with 300 mL of ethanol. Column chromatography [SiO2, CH2Cl2:hexane (1:1)] and recrystallization from ethanol produced 9.98 g (39.81 mmol; 80%) of 1,4-di(2-thienyl)-1,4-butanedione as a white solid. FTIR (KBr): 3101−3083 cm−1, 2918 cm−1, 1651 cm−1. 1H NMR (300 MHz; CDCl3): 3.40 (s, 4H), 7.15 (t, 2H), 7.68 (d, 2H), 7.82 (d, 2H). 13C NMR (300 MHz; CDCl3): 33.17, 128.17, 132.14, 133.69, 143.76, 191.41 (ppm). 2. 5-(2,5-di(Thiophen-2-yl)-1H-pyrrol-1-yl)-1,10-phenanthroline. 1,10-Phenanthrolin-5-amine (590 mg; 3.03 mmol), 1,4-di(2-thienyl)butane-1,4-dione (765 mg; 3.05 mmol), and PTSA (74 mg; 0.43 mmol) were dissolved in dry toluene (200 mL) in a RB flask fitted with a Soxhlet extractor filled with 3 Å molecular sieves and a condenser. The solution was refluxed for 48 h, and the reaction was monitored by TLC using 100% CH2Cl2 as an eluent. After the allotted time, the mixture was cooled to room temperature and the toluene was removed in a rotary evaporator. The residue was redissolved in a minimum volume of CH2Cl2 and applied to a silica column. Elution with 100% CH2Cl2 afforded rapidly off the column to provide 1.05 g (85%) of the pure product, 5-(2,5-di(thiophen-2-yl)-1H-pyrrol1-yl)-1,10-phenanthroline. M.p.: 130−131 °C. 1H NMR (300 MHz; CDCl3): d 6.51 (m, 2H), 6.65 (m, 2H), 6.72 (s, 2H), 6.88 (dd, J = 5.1 Hz, 2H), 7.52 (m, 1H), 7.69 (m, 2H), 8.00 (s, 1H), 8.27 (d, J = 1.72 Hz, 1H), 9.17 (dd, J = 4.3 Hz, 1H), 9.28 (dd, J = 4.3 Hz, 1H). 13C NMR (75 MHz; CDCl3): 102.87, 112.90, 116.88, 118.95, 121.52, 122.24, 127.60, 128.00, 128.58, 130.41, 132.28, 133.59, 136.43, 139.90, 149.86, 154.69 (ppm). MS (EI): m/z 409 (M+). 3. 5-(2,5-di(Thiophen-2-yl)-1H-pyrrol-1-yl)-1,10-phenanthroline Iron(III) Chloride. A solution of FeCl3 (400 mg; 2.50 mmol) in absolute ethanol (20 mL) was added with stirring to a solution of 5-(2, 5-di (thiophen-2-yl)-1H-pyrrol-1-yl)-1,10phenanthroline (1.0 g; 2.40 mmol) in absolute ethanol (30 mL). A deep red precipitate was immediately formed, and the reaction mixture was stirred at room temperature for approximately 1 h. The supernatant was filtered, and the precipitate was washed with ethanol and dried under vacuum. Yield: 1.12 g (82.5%). M.p.: 124−125 °C. 1H NMR (300 MHz; CDCl3): d 6.53 (m, 2H), 6.66 (m, 2H), 6.75 (s, 2H), 6.86 (dd, J = 5.4 Hz, 2H), 7.54 (m, 1H), 7.68 (m, 2H), 8.03 (s, 1H), 8.25 (d, J = 1.78 Hz, 1H), 9.16 (dd, J = 4.6 Hz, 1H), 9.30 (dd, J = 4.3 Hz, 1H). 13C NMR (75 MHz; CDCl3): 103.14, 111.12, 117.61, 119.02, 121.98, 124.37, 127.14, 128.89, 129.12, 130.58,

Electron-transfer mediators include organics, inorganics, and metal−organics.11 Of the organic mediators, methylene blue,12 quinone, and quinone derivatives13,14 have been extensively studied for a long time. The main inorganic mediators are the hexacyano-complexes of iron15,16 and ruthenium.17 Fe(CN)64− is the most widely used complex in commercial glucose sensors. Additionally, ferrocene-derivatives18,19 and Os2+/3+-complexes20 are usually used as metal−organic mediators. A 1,10phenanthroline quinone mediator is used for the glucose sensor strip.21 Most mediators are used for FAD−GOx/GDH systems, and only few adequate mediators have been used for the NAD−GDH system. In the present study, a new electron-transfer mediator was designed and synthesized for the application selectively to the NAD−GDH system based on the redox property of an iron complex with a phenanthroline attached to a bisthiophene− pyrrole backbone. This system is expected to stably adsorb onto the carbon electrode substrate through the π−π interaction and to increase the conductivity of the sensor probe, resulting in an enhanced electron-transfer process to the NAD−GDH system. Fourier-transform infrared spectroscopy (FT-IR), proton and carbon nucler magnetic resonance (1H- and 13C NMR) spectroscopy, and mass spectroscopy were used to confirm the molecular structure of FePhenTPy. The glucose sensor was assembled on a screen-printed carbon electrode (SPCE) composed of FePhenTPy (with and without reduced graphene oxide, RGO) for the mediator, and NAD−GDH was used as the enzyme system. The performance of the NAD−GDHbased glucose sensor was investigated in terms of the pH, applied potential, temperature, humidity, and interfering molecules. Finally, the glucose sensor was evaluated by determining the glucose concentration in artificial blood and human whole blood samples.

■

EXPERIMENTAL SECTION Materials. Thiophene, aluminum chloride (AlCl3), succinyl chloride, 1,10-phenanthrolin-5-amine, and p-toluene sulfonic acid (PTSA) were purchased from Sigma Company (St. Louis, MO). Column chromatography was performed on the sillica gel 60 (70−230 mesh) to purify the synthetic products. Tetrabutylammonium perchlorate (TBAP, electrochemical grade) was purchased from Fluka. It was purified according to a general method and dried under vacuum at 1.33 × 10−3 Pa. NAD+-dependent glucose dehydrogenase was purchased from Toyobo Company. Dimethyl sulfoxide (DMSO, 99.9%, anhydrous, sealed under N2 gas), nicotinamide adenine dinucleotide oxidize form (NAD+), glucose, fructose, lactose, mannose, xylose, ascorbic acid, acetaminophen, lead(IV) acetate, and Nafion (5%) were purchased from Sigma Company. All of the aqueous solutions were prepared in doubly distilled water, which was obtained from a Milli-Q water-purifying system (18 MΩ cm). Instruments. A Bruker Advance 300 Spectrometer was used to record the 1H- and 13C NMR spectra in CDCl3. The chemical shifts are given in ppm downfield from tetramethylsilane (TMS). The FT-IR spectra were recorded using a JASCO FT-IR spectrometer. The electrochemical experiments were performed using an all-in-one screen-printed carbon electrode (SPCE). Modified carbon and carbon were used as the working, reference, and counter electrodes, respectively. Carbon and silver inks (Jujo Chemical) were used in the screen-printing process. The SPCEs were printed on the polystyrene-based film using a screen printer (BANDO B

dx.doi.org/10.1021/ac403217t | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Scheme 1. Synthetic Route of FePhenTPy

131.66, 133.14, 136.55, 139.14, 147.57, 155.23 (ppm). MS (EI): m/z 571 (M+). Fabrication of Glucose Sensor Probe. The fabrication process of the glucose sensor using FePhenTPy [with or without reduced graphene oxide (RGO)] and enzyme (NAD− GDH) involves three steps. To prepare the RGO, 50 mg of graphene oxide (GO) was added to a flask containing 50 mL of distilled water (1 mg/mL), and sonication was performed to uniformly disperse the GO. Then, 100 μL of hydrazine and 300 μL of ammonia were added with stirring and the flask was immersed into an oil bath at 100 °C for 24 h. The resulting black-colored RGO powder was filtered, washed with distilled water, and then dried at 60 °C in a vacuum oven.23 The RGO was dissolved in 25% ethanol to give a concentration of 0.2 mg/ mL. The resulting solution was sonicated for 1 h. After that, 2 mM FePhenTPy was dissolved in 0.2 mL DMSO and then added to 0.8 mL of the RGO solution. Three microliters of the FePhenTPy-containing RGO (FePhenTPy-RGO) was dropped onto the SPCE and dried at 35 °C for 5 min. Five microliters of 4 mg/mL NAD−GDH solution was dropped onto the FePhenTPy-RGO/SPCE and dried at 35 °C for 5 min. Finally, 1.0 μL of the protecting membrane (1% Nafion) was dropped onto the NAD−GDH/FePhenTPy−RGO/SPCE and dried at room temperature.

■

RESULTS AND DISCUSSION Synthesis and Electrochemical Behavior of the Mediator. The mediator (FePhenTPy) was synthesized from

Figure 2. (A) CVs recorded for (dotted line) ferricyanide-, (dashed line) FePhenTPy-, and (solid line) FePhenTPy-RGO-based NAD− GDH glucose sensors at 100 mg/dL glucose concentration. (B) Comparison of calibration plots for glucose sensing at different concentrations using ferricyanide- and FePhenTPy-RGO-based NAD−GDH glucose sensors.

1,4-di(2-thienyl)-1,4-butanedione, 1,10-phenanthrolin-5-amine, and FeCl3 and the ligand (PhenTPy) was synthesized through a Paal−Knorr pyrrole condensation reaction in the presence of a catalytic amount of PTSA in dry toluene, as shown in Scheme 1. The solid product FePhenTPy was obtained through the reaction between PhenTPy and FeCl3 in absolute ethanol. This product was characterized using spectroscopic methods (FT-IR, 1 H- and 13C NMR, and mass spectrometry), which agreed with the proposed structure. The voltammetric characterization for FePhenTPy was first performed in a DMSO solution containing 0.1 M TBAP. As shown in Figure 1, the CVs recorded from −0.2 V to +0.7 V at different scan rates between

Figure 1. CVs recorded for the FePhenTPy in the DMSO containing 0.1 M TBAP at various scan rates. Inset: The scan rate dependency of the redox current of FePhenTPy. C

dx.doi.org/10.1021/ac403217t | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

The mediation performance of FePhenTPy was examined for the glucose sensor, which was compared with that of a sensor using the ferricyanide ion as a mediator instead of the proposed FePhenTPy. Figure 2A shows the CVs recorded for glucose sensing using ferricyanide, FePhenTPy, and FePhenTPy-RGO as mediators from 0.0 to 0.8 V at a scan rate of 0.1 V/s in a 100 mg/dL glucose solution. The onset potential (Eonset), the anodic peak potential (Epa), and the anodic peak current (ΔIpa) of the ferricyanide-based glucose sensor are 0.38 V, 0.57 V, and 2.35 μA, respectively. In the case of the FePhenTPy-RGObased glucose sensor, Eonset and Epa are 0.26 and 0.43 V, showing a negative potential shift, whereas Eonset and Epa are 0.27 and 0.43 V, respectively, for the FePhenTPy-based sensor system, which are very similar to the potential of FePhenTPyRGO. The ΔIpa for the FePhenTPy-RGO (12.44 μA) mediator is comparatively higher than that of the FePhenTPy-based sensor system (9.9 μA), and it is 5.3 times higher than the response of the ferricyanide-mediated sensor. In addition, the electron-transfer property of FePhenTPy adsorbed SPCE was further investigated using a 0.1 M KCl containing 10 mM ferricyanide solution because the redox peaks of FePhenTPy adsorbed on the SPCE were not clearly observed in a 0.1 M PBS (pH 7.4). In accordance with the Nicholson method, the heterogeneous electron-transfer rate constant of ferricyanide ion on the FePhenTPy adsorbed SPCE, k0, is calculated as 1.68 × 10−3 cm/s, while the k0 of bare SPCE is calculated as 8.74 × 10−5 cm/s, which shows 19.2 times enhanced values compared with the bare SPCE, when the FePhenTPy was adsorbed on the SPCE. As shown in Figure 2B, the calibration curves for glucose detection were separately obtained using ferricyanide-, FePhenTPy-, and FePhenTPy-RGO-based NAD−GDH systems in the range of glucose concentration between 20 and 600 mg/dL. All sensor systems responded linearly to the glucose concentration. The slope of the calibration curve using the ferricyanide-based glucose sensor was 0.021, while the slope of the FePhenTPy-, and FePhenTPy-RGO-based glucose sensors were 0.064 and 0.067, respectively. The slopes of FePhenTPyand FePhenTPy-RGO-based systems were very similar, but the linear range of the FePhenTPy-RGO system was wider than that of the FePhenTPy system. The response of the FePhenTPy-RGO-based sensor was three times greater than that using ferricyanide. Hence, FePhenTPy-RGO was more effective as an electron-transfer mediator due to the stable adsorption of FePhenTPy onto the carbon electrode surface, and the elevated conductivity of the sensor surface revealed the enhanced electron-transfer process to a NAD−GDH system. Surface Characterization of the Sensor Probe Layers. Impedance spectroscopy was performed to evaluate the charge transfer resistance of each probe layer that affects the sensitivity of the sensors. Figure 3 shows the Nyqiust plots obtained for each modified electrode surface with and without RGO in a 0.1 M PBS (pH 7.4) solution. In the equivalent circuit, Rs represents the solution resistance, Rct (Rp1 + Rp2) represents the charge-transfer resistance, CPE is the constant-phase element, and W is the Warburg element. Values for the parameters of Rs, Rct, CPE, and W were obtained by fitting the experimental data to the equivalent circuit using Zview2 impedance software. The Rct values of bare and FePhenTPymodified SPCE were 1361.7 and 453.7 KΩ (Figure 3A), respectively, indicating that the FePhenTPy adsorption onto the probe made the resistance lower. In addition, the doping with RGO to the FePhenTPy caused the Rct value to reduce to 4.25 KΩ (Figure 3B), where it was decreased 106.8 times

Figure 3. Nyquist plots of impedance measurements for (A) bare SPCE (black), FePhenTPy (red), NAD−GDH/FePhenTPy (blue), and Nafion/NAD−GDH/FePhenTPy (green), (B) FePhenTPy-RGO (black), NAD−GDH/FePhenTPy-RGO (red), and Nafion/NAD− GDH/FePhenTPy-RGO (blue).

0.01 V/s and 0.25 V/s reveal the quasi-reversible redox peaks of FePhenTPy that is proportional to the root-mean-square values of the scan rates. With an increase in the scan rate, the separation between the oxidation and reduction peak potentials increases, and that confirms the quasi-reversible process of the redox couples. To estimate the electron-transfer rate of the FePhenTPy redox process, the Nicholson method24 was applied. Here, the rate constant k0 for quasi-reversible reactions is estimated by determining ΔEp versus v and using the relation between ΔEp and the parameter ψ, which is defined as ψ=

(DO/DR )α /2 k0 (πDofv )1/2

where ψ is the equilibrium parameter constant, DO and DR are the diffusion coefficients, v is the scan rate, α is the transfer coefficient, and f = F/RT, where F is the Faraday constant, R is the gas constant, and T is the absolute temperature. When α is in the range of 0.3−0.7, ΔEp is almost independent of α and becomes a function of ψ only. As shown in the inset of Figure 1, ΔEp, which is calculated as the midpoint between Epa (0.31 V) and Epc (0.212 V), is 0.098 at 0.05 V/s. Hence, ψ is 0.75 at ΔEp = 0.098 V. The DO and DR are determined to be 1.52 × 10−7 and 1.58 × 10−7 cm2/s, respectively using the Randles−Sevcik equation. The heterogeneous electron-transfer rate constant of 1.0 mM FePhenTPy in a DMSO solution, k0, is calculated as 7.3 × 10−4 cm/s. D

dx.doi.org/10.1021/ac403217t | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Figure 4. The effects of (A) medium pH, (B) the application potential on the oxidation peak current, (C) temperature, and (D) humidity on the sensor performance. Interference effects from (E) AA, AP, DA, UA, caffeine, and (F) other monosaccharides (fructose, lactose, mannose, and xylose).

compared to that of only FePhenTPy-modified SPCE. The final NAD−GDH immobilized sensor probe coated with Nafion shows an Rct value of 13.6 KΩ with RGO (Figure 3B) and 1089.7 KΩ without RGO (Figure 3A). This result indicates that the FePhenTPy and RGO cause the sensor probe to be more conductive and facilitate the electron-transfer process. Optimization of the Experimental Parameters That Affect the Sensor Performance. The effects of the medium pH, applied potential, temperature, and humidity on the sensor performance were examined using artificial blood samples with a glucose concentration of 159 mg/dL at measurement conditions of 15−37 °C and 20−80% relative humidity. As shown in Figure 4A, the maximum response occurs at the physiological pH of 7.4, which is the typical pH dependence of the response current for an enzyme-based sensor. As shown in Figure 4B, the applied potential changed from 0 to 0.7 V. In this case, the maximum response was obtained at 0.55 V and this potential was used for all subsequent experiments. As shown in Figure 4C, the detection sensitivity is steady between

20 and 30 °C; however, it increases above 30 °C and decreases below 20 °C. This behavior is due to the change in the enzyme activity at different temperatures. In the case of humidity (Figure 4D), the sensitivity of the sensor is steady between 20 and 50%; however, there is a small increase in the sensitivity at the humidities above 50%. In addition, the effect of storage time on the sensitivity of the sensors was examined by evaluating the stability of the sensors at 20 °C and 50% humidity as the normal condition. The sensitivities at 20 °C after 8 h are 94.9%, 99.9%, and 98.5% (compared with the initial measuring values under normal condition) at 30, 50, and 70% humidity, respectively. Interference Effects. There are various species that interfere with glucose detection in the blood sample, including ascorbic acid (AA, 10 mg/dL), acetaminophen (AP, 15 mg/ dL), dopamine (DA, 10 mg/dL), uric acid (UA 10 mg/dL), caffeine (10 mg/dL), and other monosaccharides [mannose (10 mg/dL), lactose (10 mg/dL), xylose (10 mg/dL), and fructose (30 mg/dL)]. In the cases of AA, AP, DA, and UA, the E

dx.doi.org/10.1021/ac403217t | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

Table 2. Comparison of the Results Obtained in the Determination of the Glucose Concentration of the Whole Blood Samples by the Proposed Sensor and the Comparative Sensor System glucose concentration (mg/dL) samples sample 1

sample 2

sample 3

Figure 5. The linear calibration plots for the glucose concentration in the artificial blood samples using the amperometric method. The applied potential was set at 0.55 V. Inset: amperometrc responses recorded for different glucose concentration.

sample 4

sample 5

oxidation potentials are similar to glucose; thus, serious interference is observed with the amperometric enzyme glucose sensors. Generally, using the Nafion membrane effectively removes the interference of AA because it is negatively charged and repels negatively charged organic species, such as AA. However, it is difficult to remove the interference caused by neutral or positively charged, very small organics, such as AP, DA, etc. These interfering species cannot be removed using only a Nafion membrane. AP interferes strongly after being dosed as a medicine. Thus, we tried to remove the interfering small neutral molecules, AP, by using a coat of 1 μL of Nafion (1%, 100 μL) containing lead(IV) acetate (1 mg) as a preoxidant (mediator/NAD−GDH/oxidant containing Nafion membrane). When the very small amount of lead(IV) acetate as an oxidizing agent was present in the Nafion layer on the sensor probe, interfering species are removed by the preoxidation reaction before they reach the electrode surface. So, the current signal from AA, AP, DA, UA, and caffeine were decreased by 100%. As a result, Figure 4E shows no interference from AA, AP, DA, UA, caffeine, or other monosaccharides because the GDH is very selective for glucose against other monosaccharides including fructose, lactose, mannose, and xylose. This result is agreed with the data in Figure 4F, where there is no interference from other monosaccharides. Amperometric Analysis of Glucose in Standard Artificial Blood and Real Blood Samples. Glucose analysis using the disposable sensor was performed with an artificial blood sample that had a 42% hematocrit. To prepare the standard artificial blood samples with different glucose concentrations, the hematocrit value was controlled to 10% in the final standard solutions (29, 73, 119, 159, 250, 348, and 434 mg/dL) by dilution with a glucose (600 mg/dL)-

sample 6

sample 7

proposed method

mean (SD)

comparative method

mean (SD)

103.7 106 118.4 99 98 105.1 109.7 112.7 115.9 105.9 93.7 95.2 77.4 87.9 87.1 111.8 118.9 121.9 123 105.6 121.5

109 (±7.9)

111 117 107 91 84 91 106 121 114 88 102 111 86 100 93 120 131 115 108 112 119

111.7 (±5.1)

100.7 (±3.8) 112.8 (±3.1) 98.3 (±6.7) 84.2 (±7.9) 117.5 (±5.2) 116.7 (±9.6)

88.7 (±4.0) 113.7 (±7.5) 100.3 (±11.6) 93 (±7.0) 122 (±8.2) 113 (±5.6)

containing saline solution. Chronoamperometry and cyclic voltammetry were applied to obtain the calibration plots for the standard glucose solution. CVs recorded for the sensor probe show an oxidation peak of the mediator at 0.51 V. On the basis of this value, a potential of 0.55 V was applied for 5 s in the chronoamperometric analysis. Figure 5 demonstrates that the chronoamperometric result shows a dynamic range of 29−434 mg/dL. In this case, the result yields a regression equation for ΔIP (μA) = (0.78 ± 0.05) + (0.039 ± 0.004) [C] (mg/dL) with a correlation coefficient of 0.996. The detection limit was determined to be 12.02 ± 0.6 mg/dL from blank noise signals (95% confidence level, k = 3, n = 5). The reproducibility of the data was examined using 10 different sensors that reveal a relative standard deviation of ±5% at the glucose concentration range from 29 to 434 mg/dL (Table 1). The reliability of the proposed glucose sensor was also evaluated by determining the glucose concentration in whole blood samples from seven healthy volunteers (n = 7) (Table 2). The values of the glucose concentration using the proposed glucose sensor were compared to values obtained using a commercial glucose monitoring meters (One Touch ultra, Lifescan Inc.). The result shows good agreement between the proposed sensor and the commercial glucose monitoring

Table 1. Statistical Values (Mean, Standard Deviation, Relative Standard Deviation) At Different Glucose Concentrations (n = 10) glucose concentration (mg/dL) mean standard deviation (SD) relative standard deviation (RSD)

29

73

119

159

250

348

434

1.898 0.0879 ±4.63%

3.561 0.0702 ±1.97%

5.838 0.2199 ±3.77%

7.625 0.1026 ±1.76%

10.762 0.2380 ±2.20%

14.001 0.3222 ±4.05%

16.902 0.4376 ±3.91%

F

dx.doi.org/10.1021/ac403217t | Anal. Chem. XXXX, XXX, XXX−XXX

Analytical Chemistry

Article

(16) Cecchet, F.; Marcaccio, M.; Margotti, M.; Paolucci, F.; Rapino, S.; Rudolf, P. J. Phys. Chem. B 2006, 110, 2241−2248. (17) Crumbliss, A. L.; Hill, H. A. O.; Page, D. J. J. Electroanal. Chem. Interfacial Electrochem. 1986, 206, 327−331. (18) Hendry, S. P.; Cardosi, M. F.; Turner, A. P. F.; Neuse, E. W. Anal. Chim. Acta 1993, 281, 453−459. (19) Tian, F.; Zhu, G. Anal. Chim. Acta 2002, 451, 251−258. (20) Zafar, M. N.; Wang, X.; Sygmund, C.; Ludwig, R.; Leech, D.; Gorton, L. Anal. Chem. 2012, 84, 334−341. (21) Forrow, N. J.; Sanghera, G. S.; Watkin, J. L.; Walters, S. Biosens. Bioelecton. 2005, 20, 1617−1625. (22) Hwang, J.; Son, J. I.; Shim, Y.-B. Sol. Energy Mater. Sol. Cells 2010, 94, 1286−1292. (23) Park, S.; An, J.; Pott, J. R.; Velamankanni, A.; Murali, S.; Ruoff, R. S. Carbon 2011, 49, 3019−3023. (24) Nicholson, R. S. Anal. Chem. 1965, 37, 1351−1355.

systems (listed in Table 2). The data were evaluated through the paired t test, where the calculated t value (0.16) was less than the critical t value (2.45) at a 95% confidence level (n = 7).

■

CONCLUSIONS A new electron-transfer mediator for the NAD−GDH system was successfully designed and synthesized for use in glucose sensors, with a heterogeneous aromatic bisthiophen−pyrrole backbone structure bearing an iron−phenanthroline complex. The mediator is easily adsorbed onto the carbon surface through π−π interactions and enhances the electron-transfer mediation from the NAD−GDH molecules to the electrode surface. The sensor probe based on the new mediator and the enzyme system demonstrated excellent performance for the detection of glucose at concentrations ranging from 30 to 600 mg/dL with a detection limit of 12.02 ± 0.6 mg/dL. In addition, using the lead(IV) acetate-containing Nafion film removed the interferences from acetaminophen, uric acid, ascorbic acid, and other small organics. The proposed glucose sensor was successfully applied to the analysis of the whole blood (finger prick blood) sample, in which the results demonstrated excellent performance compared with the commercial glucose monitoring meters (One Touch ultra, Lifescan Inc.).

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: +82-51-510-2244. Fax: 8251-514-2430. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS

■

REFERENCES

This work was supported by the Program through NRF grant funded by the MEST (Grant 2010-002-9128) and Healthcare & Sensing Lab and Advanced Device Team, DMC R&D center of Samsung Electronics Company, Ltd. (S. Korea, Suwon).

(1) Zimmet, P.; Alberti, K. G. M. M.; Shaw, J. Nature 2001, 414, 782−787. (2) Brownlee, M. Nature 2001, 414, 813−820. (3) Scognamiglio, V. Biosens. Bioelecton. 2013, 47, 12−25. (4) Wang, J. Chem. Rev. 2008, 108, 814−825. (5) Heller, A.; Feldman, B. Acc. Chem. Res. 2010, 43, 963−973. (6) Zayats, M.; Katz, E.; Baron, R.; Willner, I. J. Am. Chem. Soc. 2005, 127, 12400−12406. (7) Gorton, L.; Dominguez, E. Rev. Mol. Biotechnol. 2002, 82, 371− 392. (8) Tsujimura, S.; Kojima, S.; Kano, K.; Ikeda, T.; Sato, M.; Sanada, H.; Omura, H. Biosci. Biotechnol. Biochem. 2006, 70, 654−659. (9) Tamashita, Y.; Ferri, S.; Huynha, M. L.; Shimizua, H.; Yamaokaa, H.; Sode, K. Enzyme Microb. Technol. 2013, 52, 123−1128. (10) Frias, J. P.; Ellison, G. M.; Lim, C. G.; Montandon, C. M. Diabetes Care 2010, 33, 728−729. (11) Heller, A.; Feldman, B. Chem. Rev. 2008, 108, 2482−2505. (12) Willner, I.; Baron, R.; Willner, B. Biosens. Bioelecton. 2007, 22, 1841−1852. (13) Kaku, T.; Karan, H. I.; Okamoto, Y. Anal. Chem. 1994, 66, 1231−1235. (14) Battaglini, F.; Koutroumanis, M.; English, A. M.; Mikkelsen, S. R. Bioconjugate Chem. 1994, 5, 430−435. (15) Jaffari, S. A.; Turner, A. P. F. Biosens. Bioelecton. 1996, 12, 1−9. G

dx.doi.org/10.1021/ac403217t | Anal. Chem. XXXX, XXX, XXX−XXX