ANALYTICAL SCIENCES JULY 2015, VOL. 31 591

Introduction

Organophosphate pesticides (OPs) are highly toxic because they inactivate acetylcholinesterase (AChE), an enzyme required for nerve function.1 They form covalent bonds to serine residues that are located in the active site of AChE, impairing the enzyme’s function. Subsequent build-up of acetylcholine blocks cholinergic nerve impulses, leading to paralysis, suffocation, and ultimately to death.2 Thiocholine (TCh) is one of the reaction products of AChE with acetylthiocholine (ATCh) as a substrate. Therefore, TCh has been used as an indicator to measure the activity of AChE for the detection of OPs, carbamic pesticide, and nerve agents.3,4 Currently, spectrophotometric detection using Ellman’s method is often carried out to measure TCh, which is based on the detection of the produced TCh.5 The detection is based on a reaction that uses TCh and 5,5′-dithiobis-2-nitrobenzoic acid (DTNB) to produce 2-nitro-5-thiobenzoic acid (TNB) with yellow color.6 Compared with spectrophotometric detection, electrochemical methods have advantages such as high sensitivity, rapid response, and good stability when incorporated in miniaturized devices.7–9 Moreover, the measurement is simple and can be carried out by untrained personnel.

For analyses that use expensive reagents such as enzymes, the amount of the reagent should be minimized. For this purpose, microfluidics, particularly plug-based microfluidics, are effective. The plug is a fragment of a solution formed in a microfabricated flow channel. In ordinary microfluidics based

on the use of continuous flows, most of the solution is wasted and only a part is used depending on the application. On the other hand, in plug-based microfluidics, the total volume of solutions can be minimized to the nL or pL order. In most of the plug-based microfluidic systems, immiscible oils were used to separate aqueous solutions. However, a serious problem in applying this platform to biochemical analyses is that the sensing area is contaminated with the oil. To address this, we have developed plug-based microfluidic systems in which the plugs are separated by air, thus resolving the problem of contamination.

In detecting the analyte electrochemically, amperometry is often used. However, in a plug of a very small volume, the redox compound detected on the working electrode depletes rapidly in the plug. As a result, the output current decreases rapidly as the volume of the plug decreases, making the measurement difficult. On the other hand, coulometry facilitates the measurement because the output charge, which is the integral of the current, increases as time elapses.

We have previously reported a microdevice for OPs based on two-step bienzyme reactions of AChE and choline oxidase and coulometric detection of one of the final reaction products, H2O2.10 The enzyme plug and substrate plug were formed in the flow channel by rhombuses structure and the final product was determined at the electrode area. In processing plugs of aqueous solutions, the interior of the flow channels must be sufficiently hydrophobic. Otherwise, plugs collapse while they are moved. Here, a problem is that the surface state changes by the adsorption of proteins. The influence of using the two enzymes was not negligible and residue of the enzymes often remained on the wall of the channel. In addition, the use of the two enzymes increases cost.10 To solve the problems, a coulometric

2015 © The Japan Society for Analytical Chemistry

† To whom correspondence should be addressed.E-mail: hsuzuki@ims.tsukuba.ac.jp

Microfluidic Device for Coulometric Detection of Organophosphate Pesticides

Jin WANG,* Takaaki SATAKE,* and Hiroaki SUZUKI**†

* Graduate School of Life and Environmental Sciences, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305–8572, Japan

** Graduate School of Pure and Applied Sciences, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305–8573, Japan

A microdevice for coulometric detection of organophosphate pesticides (OPs) was fabricated based on the measurement of the inhibition of an enzyme, acetylcholinesterase (AChE), by OPs. Thiocholine (TCh) produced in the enzymatic reaction of AChE with acetylthiocholine (ATCh) as a substrate was oxidized on a microelectrode array formed in a main flow channel. Volumes of plugs of necessary solutions were measured using a structure consisting of a row of rhombuses formed in an auxiliary flow channel. The plugs were merged and solution components were mixed at a T-junction formed with the main and auxiliary flow channels. A linear relationship was confirmed between the generated charge and the logarithm of the OP (malathion) concentration in a concentration range between 10–6 and 10–3 M with a correlation coefficient of 0.951. The lower limit of detection was 412 nM.

Keywords Coulometry, organophosphate pesticide, acetylcholinesterase, acetylthiocholine

(Received January 14, 2015; Accepted May 8, 2015; Published July 10, 2015)

592 ANALYTICAL SCIENCES JULY 2015, VOL. 31

device using a mono-enzyme system based on the detection of TCh was fabricated.

Experimental

Reagents and chemicalsReagents and materials used for fabrication and characterization

of the device were obtained from the following commercial sources: glass wafers (#7740, 3 i.d., 500 μm thick) from Corning Japan (Tokyo, Japan); a thick-film photoresist, SU-8 25, from Microchem (Newton, MA); prepolymer solution of polydimethylsiloxane (PDMS), KE-1300T, from Shin-Estu Chemical (Tokyo, Japan); AChE (EC 3.1.1.7, 827 units/mg, from Electrophorus electricus; and chloride salt of ATCh from Sigma-Aldrich (St. Louis, MO). A 50 mM phosphate buffer solution (PBS, pH 7.4) containing 0.1 M KCl was used to dissolve AChE. Commercial OP formulation (malathion) was purchased from Sumitomo Chemical Garden Products (Tsukuba, Japan). OP standard solutions were prepared with distilled-deionized water. For the analysis on the device, a solution containing ATCh and a solution containing OP and the enzyme were prepared. All reagents were of analytical grade.

Device fabricationThe microdevice consisted of a glass substrate with a three-

electrode system to measure charge generated accompanying the oxidation of TCh and a PDMS substrate with a flow channel structure (Fig. 1). The fabrication process was described in our previous work.8.10 The three-electrode system consisted of a platinum working electrode, an Ag/AgCl reference electrode, and a platinum auxiliary electrode. The working electrode consisted of a microelectrode array (40 thin strips of 600 μm × 400 μm) with a 130-μm inter-electrode distance (edge to edge). Two pinholes (30 μm in diameter) were formed on the silver layer of the reference electrode. AgCl was grown from the pinholes into the silver layer by applying a current of 50 nA for  20 min in a 0.1 M KCl solution. This structure realized excellent durability for reliable coulometric measurement.11 The flow channel structure including the main and auxiliary flow channels was formed with PDMS by replica molding using a template formed with the thick-film photoresist (SU-8). A T-junction formed with the main and auxiliary flow channels was used to form and manipulate solution plugs. The auxiliary flow channel consisted of an array of rhombuses. The shape of the rhombuses was designed so that solutions occupy each rhombus as a unit depending on the applied pressure by taking

advantage of changes in surface tension in the structure.8 Through holes were formed at the end of the main and auxiliary flow channels by punching, and silicone tubes (inner diameter: 500 μm) were connected there to introduce solutions and apply pressure.

Principles of coulometric determination of OPsThe enzymatic and electrochemical reactions used for the

detection of OP are as follows:

Acetylthiocholine + H2O ←→AChE Thiocholine + Acetic acid

Thiocholine ←→Platniumelectrode Thiocholine (ox) + 2H+ + 2e–

OPs are well-known inhibitors of AChE. In this study, ATCh was used as the substrate to measure a change in the AChE activity before and after the incubation with the OP. The degree of inhibition was calculated using the following equation:

Inhibition(%) = −−

×q qq q1 2

1 0100, (1)

where q0 is the background charge in the absence of the OP and the enzyme, q1 is the charge when the enzyme is not inhibited, and q2 is the charge when the enzyme is inhibited by the OP. The reduced activity of AChE leads to the decrease in the production of TCh, which then results in the decrease in current or charge generated on the working electrode. Therefore, the decrease in the measured charge can be related to the inhibitory effect of the OP or its concentration. In this study, malathion was used as a model OP.

A critical point in improving the detection limit of coulometric detection is how to increase the ratio of the Faradaic charge with respect to the background charge. Unlike amperometry, the background charge increases as time elapses because it is obtained as the integral of the generated current. To minimize the background and increase the contribution from the Faradaic charge, microelectrodes are effective. The small electrode area minimizes the background and the unique diffusion profile around microelectrodes enhances the Faradaic current. In addition, the sensitivity can also be improved by collecting as many molecules as possible. Although this can be achieved by simply making the time for measurement longer, another option has to do with how the molecules in the flow channel structure are collected. To solve these problems, we used an array of strips of microelectrodes that expanded in the flow channel.8

Fig. 1　The coulometric microdevice for pesticide determination. (A) Exploded view of the device. (B) Top view of the device with the electrodes and the flow channel structure. (C) Picture of the device.

ANALYTICAL SCIENCES JULY 2015, VOL. 31 593

Measurement procedureA solution containing ATCh and a solution containing

malathion and the enzyme were introduced into the flow channel and were processed in the form of liquid plugs. The volumes of the solutions were measured using the rhombus structure in the auxiliary flow channel, and the solutions were mixed by an operation using the T-junction (Fig. 2). In measuring the volume of a solution, a solution plug was introduced into the auxiliary flow channel by applying a negative pressure to the flow channel. The rhombus structure was designed so that the solution moved spontaneously once it passed the widest portion and stopped at the narrowest portion by capillary action. Two rhombuses (150 nL) were used to form each plug and a volume that corresponded to four rhombuses was equal to the volume of the sensing region (300 nL).8 Coulometry was conducted using an electrochemical workstation (Autolab PGSTAT13, Eco Chemie, Utrecht, The Netherlands) connected to a laptop computer. The potential applied to the working electrode was +0.65 V with respect to the on-chip Ag/AgCl reference electrode.7 After each measurement, the electrodes and the flow channel were washed with PBS.

For the coulometry, TCh was produced by the enzymatic reaction and was accumulated for 30 min in the plug of the merged solutions. The coulometry was started by introducing the substrate (ATCh) and enzyme (AChE) solutions into the flow channel structure and merging the plugs of the solutions. The signal produced accompanying the enzymatic reaction (q1) was recorded. After flushing all used solutions in the flow channel structures and washing them with PBS, a malathion standard solution containing AChE was introduced into the flow channel along with the substrate solution. After merging the plugs of the solutions, the change generated by the inhibited AChE was measured (q2). The concentration of ATCh was fixed at 2 mM.

Results and Discussion

Response of the device to ATChAs a first step, the response of the device to ATCh was checked

without the enzyme and malathion. Figure 3 shows the dependence of the generated charge on the concentration of ATCh. The curve gradually saturated with the increase in the ATCh concentration. The apparent Michaelis–Menten constant, Km

app, for ATCh determined by the electrochemical Eadie–Hofstee form of the Michaelis–Menten equation was 0.66 mM.1 The relative standard deviation (RSD) was within 3.2% for values obtained in triplicate measurements, demonstrating reproducible measurements using the device.

Fig. 2　On-chip processing of solution plugs. (A) A substrate solution and a mixture of the enzyme and malathion were introduced into the main flow channel. (B – D) The volume of each solution was measured. (E) The two plugs were merged in the main flow channel. (F) The new solution mixture was transported to the sensing region for detection.

Fig. 3　Dependence of the generated charge on the concentration of ATCh. The symbols correspond to three measurements.

594 ANALYTICAL SCIENCES JULY 2015, VOL. 31

Temperature dependenceIn order to choose the best condition for ATCh detection, the

influence of temperature on the enzyme activity was investigated. Responses were measured at 25, 37, and 50°C. The room temperature was set at 25°C for the measurement at 25°C. For experiments at 37 and 50°C, the chips were placed on a hot plate with the temperature set, respectively. As shown in Table 1, the enzyme activity decreased monotonically with the increase in temperature, and the maximum response was obtained at 25°C among the three temperatures. Although deactivation of the enzyme is anticipated at higher temperatures, the measured charge was substantial even at 50°C. Although higher activity may be observed at temperatures between 25 and 37°C, formation of air bubbles dissolved in the main flow channel became more significant when the temperature was elevated. No such troubles were experienced at 25°C. Therefore, all the measurements were carried out at 25°C.

Detection of malathion using the deviceDetermination of malathion was carried out based on the

enzymatic and electrochemical reactions mentioned earlier. The inhibition of AChE by OPs is irreversible and the existence of the OPs reduces the amount of TCh produced by the enzymatic reaction. Figure 4 shows calibration plots that illustrate the relation between the inhibition of AChE activity and malathion concentration. In the semi-log plot, the graph shows linearity in a concentration range between 10–6 and 10–3 M (R2 = 0.951). Gogol et al. and Arduini et al. designed screen-printed electrodes coated biosensors with 3.5 × 10–7 M for trichlorfon, 1.5 × 10–7 M for coumaphos5 and 5 × 10–7 M6 detection limit, respectively; Liu et al. used a carbon nanotube modified glassy

carbon electrode for TCh detection with a detection limit of 3 × 10–7 M for TCh.7 The lower limit of detection we obtained was 412 nM, which is comparable with those of previous reports  based on the same TCh detection. The maximum residue  levels (MRLs) represent the upper legal limit of the concentration of pesticide residues in food. Table 2 summarizes the MRLs of malathion for various foods in the U.S.A, Canada, and Mexico.12 Meanwhile, according to the Japan Food Chemical Research Foundation, in the case of malathion, it is 0.5, 1, 2, and 0.1 ppm, 0.01 ppm for tea, papaya, cauliflower, brown rice, and tomato juice, respectively. Although the MRL is different among vegetables, the following reasons can be considered. First, since the shape and/or structure of the edible portion of vegetables are different, the amount of pesticide

Table 1　Charge measured accompanying the enzymatic reactions at different temperatures

Temperature/°C Charge/μC

253750

7.63 ± 0.317.05 ± 0.123.50 ± 0.66

ATCh concentration: 2 mM. Each charge represents the average of three measurements. Standard deviations are also shown.

Table 2　Summary of current international tolerances and maximum residue limits (ppm)

Commodity U.S. Canada Mexico

Apple TBD 2.0 8

Avocado 0.2 8.0

Barley, grain, postharvest 8 8.0 (raw cereals)

Bean Dry, 2.0Succulent, 2.0

2.0 8

Blackberry 6 8.0

Blueberry 8 8.0

Carrot, roots 1 0.5 8

Cherry 3.0 6.0

Corn, grain, postharvest 8.0 8.0

Corn, sweet, kernel plus cod with husks removed

0.1

Cucumber 0.2 3.0 8

Eggplant 2.0 0.5 8

Garlic 1.0 0.5 8

Grape 4.0 8.0 8

Melon 1.0 8.0 8

Nut, macadamia 0.2

Papaya 1 8.0 1

Peach 6.0 6.0 8

Pear 3.0 2.0 8

Pepper 0.5 0.5 8

Potato 0.1 0.5 8

Pumpkin 1.0 3.0 8

Radish 0.5 0.5 8

Rice, grain, postharvest 30 8.0 (raw cereals)

8

Shallot, bulb 6.0

Squash, summer and winter

Summer, 0.2Winter, 1.0

3.03.0

8 (zuchinni)

Sweet potato, roots 0.1

Tomato 2.0 3.0 8

Turnip (including tops) Tops, 4.0Roots, 0.5

0.5

TBD = To be determined.

Fig. 4　Dependence of AChE inhibition on the concentration of malathion. Each point in the graph represents the average and standard deviations of three measurements.

ANALYTICAL SCIENCES JULY 2015, VOL. 31 595

rinsed off by rain is also different. As a result of this, residual pesticides are different. Second, vegetables are categorized into root and leafy vegetables, and the portion of vegetables exposed to pesticide are different between these two groups. Residual pesticides of root vegetables are very low compared with leafy vegetables. Third, resistibility to diseases and insects depends on the vegetable. This indicates that more pesticides are needed for low resistive vegetables. Fourth, growth periods and harvest periods vary depending on the vegetables. So the periods of applying pesticides are also different. Fifth, the balance between supply and demand in the market may make it difficult to set extremely low detection values for vegetables whose sufficient supply is difficult. The lower limit of detection achieved using our device is sufficient for tea, papaya, cauliflower and many other types of food. However, it is not sufficient for brown rice and tomato juice in Japan, but sufficient enough for all those in the U.S., Canada and Mexico according to the database shown in Table 2.

To lower the detection limit further, a longer incubation time and use of other enzymes with a smaller Km value will be effective to improve the detection limit.5,13 In this experiment, the incubation time for the mixture of malathion and AChE was 7 min. Therefore, there is still room to improve the performance if the requirement for the length of the incubation time is not so severe. Nevertheless, the device can reduce the consumption of very expensive reagents and can simplify the processing of solutions when performing the detection. In this sense, the device can be useful in realizing portable user-friendly instruments for on-site analyses of OPs.

Conclusions

A coulometric microdevice for OP determination was fabricated based on the inhibition of AChE. A mono-enzyme system and simplified processing of solutions in this microfluidic device

can realize low cost user-friendly analysis. A linear relationship (R2 = 0.951) was observed in the OP concentration range between 10–6 and 10–3 M with a 412 nM detection limit for malathion. The sufficient sensitivity and detection limit obtained demonstrates the excellent electroanalytical performance of this microdevice.

References

1. K. A. Joshi, J. Tang, R. Haddon, J. Wang, W. Chen, and A. Mulchandani, Electroanalysis, 2005, 17, 54.

2. F. M. Raushel, Nature, 2011, 469, 310. 3. S. Andreescu, L. Barthelmebs, and J. L. Marty, Anal. Chim.

Acta, 2002, 464, 171. 4. A. Crew, D. Lonsdale, N. Byrd, R. Pittson, and J. P. Hart,

Biosens. Bioelectron., 2011, 26, 2847. 5. F. Arduini, A. Cassisi, A. Amine, F. Ricci, D. Moscone, and

G. Palleschi, J. Electroanal. Chem., 2009, 626, 66. 6. G. Liu, S. L. Riechers, M. C. Mellen, and Y. Lin,

Electrochem. Commun., 2005, 7, 1163. 7. E. V. Gogol, G. A. Evtugyn, J. L. Marty, H. C. Budnikov,

and V. G. Winter, Talanta, 2000, 53, 379. 8. F. Sassa, H. Laghzali, J. Fukuda, and H. Suzuki, Anal.

Chem., 2010, 82, 8725. 9. S. Andresscu, T. Noguer, V. Magearu, and J. L. Marty,

Talanta, 2002, 57, 169. 10. J. Wang, H. Suzuki, and T. Satake, Sens. Actuators, B,

2014, 204, 297. 11. H. Suzuki and T. Taura, J. Electrochem. Soc., 2001, 148,

468. 12. R. P. Keigwin Jr., Reregistration Eligibility Decision (RED)

for Malathion, US EPA Archive document, 67-81 (2009). 13. P. Mulchandani, A. Mulchandani, I. Kaneva, and W. Chen,

Biosens. Bioelectron., 1999, 14, 77.