A disposable acetylcholine esterase sensor for As(III) determination in groundwater matrix based on 4-acetoxyphenol hydrolysis

Abstract

There is a lack of field compatible analytical method for the speciation of As(III) to characterize groundwater pollution at anthropogenic sites. To address this issue, an inhibition-based acetylcholine esterase (AchE) sensor was developed to determine As(III) in groundwater. 4-Acetoxyphenol was employed to develop an amperometric assay for AchE activity. This assay was used to guide the fabrication of an AchE sensor with screen-printed carbon electrode. An As(III) determination protocol was developed based on the pseudo-irreversible inhibition mechanism. The analysis has a dynamic range of 2-500 μM (150 – 37,500 μg L⁻¹) for As(III). The sensor exhibited the same dynamic range and sensitivity in a synthetic groundwater matrix. The electrode was stable for at least 150 days at 22 ± 2 °C.

Graphical Abstract

1. Introduction

Arsenic contamination of groundwater has been found in many places under a wide range of hydrogeochemical conditions. The arsenic in groundwater is predominantly inorganic, with concentrations varying from < 0.5 to 5,000 μg L⁻¹.¹ Since the 1930’s, many arsenic poisoning incidents have been reported. Most of them were caused by ingestion of groundwater polluted by geogenic or anthropogenic arsenic.^2-4 Systematic study has demonstrated that arsenic is a carcinogen and a general toxin. Long-term ingestion of arsenic causes not only skin, bladder, or lung cancer; but also damage to the neural, cardiovascular, and reproductive systems. For vulnerable people, the harmful level can be as low as < 1 μgL⁻¹.⁵

The primary source of groundwater arsenic pollution can be geogenic or anthropogenic.^6,7 In anthropogenic pollution, arsenic is released from sites with a history of mining or smelting, coal burning, manufacture or application of arsenic agrochemicals, and wood preservation with chromated copper arsenate. Once arsenic is released from the primary sources into groundwater, it undergoes redistribution through redox reactions, sequestration by forming minerals in sediment, and desorption or dissolution from the sediments.

To manage anthropogenic pollution, the primary source needs to be characterized, contained, and treated.⁸ Runoff of arsenic from primary source causes acute impact on local lives, diffuse pollution downstream, and pollutant infiltration of groundwater. Examples include fish kills by unmanaged acid mine drainage,⁸ expansive arsenic pollution resulting from mine flood⁹ or uncontrolled discharge from arsenic chemical plant.¹⁰ For mine waste management, the contamination at the site needs to be characterized in terms of mineral source, contaminant abundance in the minerals, reactivity of the minerals, and main flow path of the contaminated water.¹¹ Anthropogenic site remediation can be time consuming and expensive. At Vinland Superfund site, the cleanup involved excavation and flushing of 682,558 tons of contaminated soil, pumping and treating two million gallons of groundwater per day. As of 2018, the project has cost $219.4 million since the start of cleanup in 1992.^12,13

Arsenic contamination was a concern in one third of the superfund sites in the US.¹⁴ A special type of pollution is caused by leachate from landfills, as it can mobilize arsenic in sediments by reductive dissolution.^15-17 The pollution needs to be managed and treated to prevent diffusion to surface water.^18,19 This effort entails monitoring and characterization of the flow path, plume distribution, and composition of groundwater.^10,17

Arsenic speciation is essential for the understanding of pollution source and pollutant transportation. Inorganic arsenic exists as As(III) and As(V) in As-O₂-H₂O system, with distribution governed by pH and redox potential. As(III) is the predominant species under reductive and acid-to-neutral condition.¹ In natural water, arsenic mobilization and re-distribution depends on the interconversion of As(III) and As(V). This interconversion is complex as it is directly coupled to the redox reactions involving iron, sulfur, and manganese at specific location.^10,20

As(III) was suggested as the dominant species in groundwater pollution caused by landfill leachate, but the instability of As(III) in samples casted doubts on the accuracy of earlier results in lab analysis. This is a significant concern because As(III) is highly different from As(V) in toxicity, bioavailability, mobility, and remediation need.^21,22 Inorganic arsenic specification in the lab usually involves a chromatographic separation and a spectrometric detection.^21,23,24 Although separation-spectrometry methods are highly sensitive and accurate, environmental sampling and sample management have been difficult. Samples from the field generally need to be cleaned, acidified, and stored under controlled condition, mainly to prevent As(III) oxidation.²⁵ However, general treatment is often inadequate because arsenic can be precipitated by adsorption to iron oxide minerals when sample is exposed to oxygen or formation of orpiment (AS2S3) when acidified.^26,27 Preservation methods need to be systematically developed for specific matrix to allow adequate time for lab analysis.^28,29

The sampling issues are commonly mitigated with in-field separation. Different species of arsenic are separated with solid phase extractions and sent for lab analysis.³⁰ In this procedure, water matrix may have strong impact on extraction capacity and the analysis throughput is still limited by sample management. ³¹

Electroanalysis based on different types of stripping voltammetry have been used in field speciation.^10,32 Although these methods are highly sensitive, they are still in development for routine environmental monitoring. Notable issues to be addressed include: robustness of electrodes, passivation of electrodes, interference caused by common elements such as Cu(II), requirement for HCl (2 M) as electrolyte solution, and cost of instrumentation, etc.^33-35

Many biomolecules can selectively bind to arsenic with different mechanisms for As(III) and As(V), therefore they may be used as bioreceptors for speciation.^36,37 These biomolecules include DNA fragments, aptamers, and enzymes or proteins. The specific interactions have been employed to develop many optical and electrochemical sensors. However, most of the reported biosensors are not ready for field use because they do not meet important practical qualifications. The fabrication needs to be simple yet tunable for mass production. The resulting sensors must be stable in storage, deployment, and operation. Finally, potential biosensors should maintain their selectivity and sensitivity in realistic water matrices.

To minimize the development risk, we picked an amperometric assay using acetylcholinesterase (AchE) as the bioreceptor. AchE is known to be inhibited by As(III).³⁸ The enzyme is not only extremely efficient, but also highly stable in solution. For field use, electrochemical transduction is preferred because it is highly sensitive and can be miniaturized.³⁹ We selected carbon screen-printed electrodes (SPE) as the base because of their low cost, flexibility, suitability for mass production, robustness when used in different aqueous matrices.^40,41 In addition, disposal of AchE sensor has less regulatory restrictions because the analysis does not involve hazardous chemicals or genetically modified organisms.^33,36,42

AchE has long been used to develop enzyme electrodes for pesticides, nerve agents, and heavy metals.^43-46 It offers distinctive advantages in catalytic efficiency and direct evidence of neurotoxicity. With a catalytic efficiency (k_cat/K_m) > ×10⁸ M⁻¹s⁻¹, the catalysis is essentially diffusion controlled.⁴⁷ The storage stability, ranging between 5-120 days, appears to be dependent on the fabrication methods.⁴⁴ For As(III) determination, there is a track record of improvements with the AchE biosensor and assay method. AchE was first immobilized on Pt electrode in 1964 for potentiometric assay of nerve agents.⁴¹ In 1998, AchE was immobilized on a graphite rotating disc electrode to determine As(III) via an amperometric assay.⁴⁸ A decade later, a disposable AchE sensor was fabricated on SPE with a carbon working electrode⁴⁹,⁵⁰. In the amperometric assays, the immobilized AchE hydrolyzes acetylthiocholine iodide (ATCI) to give thiocholine, which is detected by anodic oxidation.⁵¹ However, ATCI has been proven unsuitable for the transduction because the iodide can be oxidized at the same working potentials as thiocholine and might catalyze the oxidation of As(III).^52,53 In a later report, acetylthiocholine chloride (ATCCl) was employed as the substrate. To improve anodic oxidation, a platinum electrode was modified with a nanocomposite of graphene oxide and Ru(II)-tris(bipyridyl). The working potential was lowered to 0.214 V vs Ag/AgCl.⁵⁴

For free AchE from electric eel, the inhibition by As(III) was found to be pseudo-irreversible.³⁸ The association rate was 1.28 × 10² M⁻¹ min⁻¹, and the dissociation rate constant was 1.9 × 10⁻³ min⁻¹. Based on these values, we estimated that 1 μM of As(III) should cause 6.25% inhibition. Even though ligands such as pyridine-2-aldoxime methiodide (2-PAM) have been shown to promote the binding, it still requires at least 30 min to reach equilibrium when [As(III)] is at low μM concentrations. Surprisingly, this mechanism was disregarded in all published As(III) determinations using AchE biosensors. The assay protocols were all based on the assumption of rapid binding and dissociation to reach steady state. In one study using an amperometric assay, steady current was reached in 9 s after As(III) addition.⁵⁴ They showed no difference in response time between low and high concentration of As(III).^48,49,54 The immobilized AchE was reactivated in 15 s by buffer wash.⁴⁸ In addition, these sensors all exhibited very different sensitivity, with dynamic ranges from 0.2-20 nM ⁴⁸ to 28.7 −1600 μM.⁵⁰ Apparently, the inhibition of AchE might not play a significant role in the selectivity for As(III) in these previously reported AchE sensors.

In this AchE sensor development study, we propose to use 4-acetoxyphenol as the substrate based on the assumptions that it would be efficiently hydrolyzed by AchE but not oxidized at the working potential of the hydrolysis product hydroquinone (Fig. 1). We expected that the selectivity of Reaction 2 on carbon electrode would not require a redox mediator, thereby simplifying sensor fabrication.

Fig. 1.

Amperometric assay of AchE activity with 4-hydroxyphenol was the substrate

Although many immobilization methods have been developed, it is hard to predict their impacts on the kinetics of the immobilized enzymes.⁵⁵ Among the options, crosslinking with glutaraldehyde (GA) offers several important advantages.^56-58 GA is not only highly reactive but also flexible, because it can crosslink lysine residues as oligomers of different lengths. Bovine serum albumin (BSA) is lysine rich therefore is used as the carrier protein. The ratio of BSA with AchE can be adjusted to optimize enzyme performance including activity, sensitivity to As(III), and stability. The cross-linking is straightforward and easy to scale up.

To ensure AchE is acting as the bioreceptor for As(III), the kinetic mechanism of the inhibition has been evaluated with the electrodes and compared with the previous kinetic analysis of the free enzyme.³⁸ A protocol consistent with the mechanism has been developed for the assay of As(III). The assay has been tested in groundwater matrix and the storage stability of the sensor has been evaluated at ambient temperature for 5 months.

2. Experimental

2.1. Chemicals and Materials

Acetylcholinesterase (E.C. 3.1.1.7) Type VI-S from Electrophorus electricus (electric eel) with activity of 217 U mg⁻¹ protein, was purchased from Sigma-Aldrich. 4-Acetoxyphenol was purchased from Combi-Blocks (San Diego, CA, USA). All other chemicals were purchased from Sigma-Aldrich (reagent grade) and used as received.

Carbon Screen Printed Electrodes (SPE), including those with small round working electrodes (2 mm OD, RRPE1001C) and large rectangular working electrodes (4 × 5 mm, RRPE1002C), were purchased from Pine Research Instrumentation (Durham, NC, USA).

2.2. AchE immobilization on SPE

A mixture of BSA and AchE (containing 0.50 μg. or 0.11 U of the AchE, 8 μg of BSA in 10 μl of 0.06 M phosphate, pH 7.0) was spread on the working electrode of carbon SPE (RRPE1002C). The proteins were deposited on the electrode surface by air drying, and then crosslinked by adding 10 μl of 0.0021% of glutaraldehyde solution in water. The electrode was air dried at room temperature for 16 h to complete the crosslinking. To hydrate the sensor, it was stored in 0.1 M Tris-HCl, pH 7.0 for > 24 h at room temperature before use.

2.3. Electrochemical Measurements

Electrochemical experiment was carried out in the low volume SPE cell (Pine Research Instrumentation, Durham, NC, USA), and the condition was controlled by a DY2013 Potentiostat with DY2000 software (Digi-Ivy, Inc. Austin, TX, USA). The working potential was set with the Ag/AgCl electrode on the SPE as the reference. For cyclic voltammetry study, the SPE with a small working electrode (2 mm ID round, RRPE1001C) was used. For chronoamperometry study and AchE electrode preparation, we used the SPE with a larger working electrode (4 × 5 mm rectangular RRPE1002C).

All experiments were carried out at 22 ± 2 °C. The current in amperometric determination was recorded for 150s. The current typically stabilized in < 60s. The average of the readings between 120s and 150s was used as the steady-state current. The collected data was exported to Excel for data presentation and analysis.

2.4. Inhibition of immobilized AchE by Sodium Arsenite

The inhibition of the immobilized AchE by As(III) was investigated for protocol development. An AchE electrode was incubated in a NaAsO₂ solution for varying periods of time before removal to determine the residual activity. The inhibition was determined by the decrease of steady state current during amperometric measurement. The concentrations of NaAsO₂ were 5, 20 and 100 μM, and the incubation times were varied between 1 min and 60 min. The inhibition data was fit to a reversible pseudo-first order reaction model (Eq. 1) using a non-linear least-square method.⁵⁹

$$\begin{gathered} \mathsf{E}\underset{k_{-1}}{\overset{k_1[\mathsf{As}]}{\rightleftharpoons }}\mathsf{E}•\mathsf{As} \\ \text{ln}\frac{[E]-[E{]}_{\infty }}{[E{]}_0-[E{]}_{\infty }}=-(k_1[As]+k_{-1})t \end{gathered}$$ Eq. 1

2.5. [As(III)] determination protocol

The concentration of As³⁺ was determined by its correlation with the degree of inhibition of AchE. The AchE electrode was incubated with arsenite solution in 0.1 M Tris-HCl, pH 8.0 for 1 h at 22 ± 2 °C. The activity of the electrode was measured before and after the inhibition with an amperometric method based on the reaction sequence in Figure 1. The inhibition was calculated by Eq. 2, where I is inhibition, i_o is the steady state current in the absence of As³⁺, and i_i is the steady state current of the electrode after inhibition.

$$I=\frac{i_0-i_i}{i_0}$$ Eq. 2

Experimental planning and analysis were carried out with JMP software (SAS Institute, Cary, NC, USA)

2.6. Total arsenic determination with ICP-AES

Total arsenic (including As(III) and As(V)) was determined with an iCAP 6500 Inductively Coupled Spectrometer ( Thermo Scientific, Waltham, MA, USA ). The detection limit is 5 ppb.

3. Results and Discussion

3.6. Electrochemical characterization of the substrate and product

The electrochemical reactions were first evaluated with 5 mM solutions of 4-acetoxyphenol or hydroquinone (QH₂) in 0.1 M phosphate, pH 7.0. Cyclic voltammetry was performed with initial potential at −0.8 V, scan rate at 50 mV s⁻¹, and switching potential at 0.8 V.

Hydroquinone gave two well defined peaks at 195 and −275 mV in the voltammogram (Fig. 2). These peaks were reproducible in successive scans with the same electrode. The peak separation at 50 mV s⁻¹ scan rate was 470 mV, indicating the interfacial kinetics was slow. Further investigation showed that the peak potential was a function of scanning rate. The peak potential separation increased from 420 to 672 mV as the scan rate increased from 20 to 1000 mV s⁻¹. The peak current ratios were in the range between 0.85 and 1.09 (Supporting Information, Fig 1 and Table 1). The results show that the electrode reaction of hydroquinone is quasi-reversible on the carbon electrode: the reaction was chemically reversible but the kinetics was not. Despite the non-ideal behavior of Reaction 2, the anodic peak height in CV was linearly correlated to the concentration of reactant in the range from 0.05 to 10 mM when the scan rate was 1000 mV s⁻¹ (Supporting Information, Fig. 2 and 3).

Fig. 2.

Cyclic voltammograms of hydroquinone and 4-acetoxyphenol

The electrochemistry of quinone has been extensively studied and employed in many applications.⁶⁰ The interconversion between 1,4-benzoquinone (Q) and the corresponding hydroquinone (QH₂) is complex as the reaction involves two electron transfer reactions and two protonation steps. The mechanism is dependent on pH, working potential, and the reaction medium.⁶¹ A nine-membered square scheme was used to analyze the pathways for the intermediate interconversion via electron transfer and protonation.⁶¹ In phosphate buffer at pH 7.2, Reaction 2 gave two widely separated peaks (ΔE = 334 mV) in cyclic voltammetry with a glassy carbon electrode. The pathway for Reaction 2 was postulated to involve a HeHe mechanism, although these four reactions could involve concerted steps.

The proposed AchE substrate 4-acetoxyphenol did not undergo any redox reactions in CV. This is consistent to an earlier report, indicating the oxidation took place at 1.1 V (vs Ag/AgCl) with a glassy carbon electrode.⁶² The large difference in oxidation potentials between the substrate and product is highly desirable for selective determination of QH₂. It is feasible to develop an amperometric method based on Reaction 2 (Fig. 1) in a way similar to the reaction with 4-aminophenyl acetate as the substrate.^63,64 The oxidation of 4-aminophenol is more complicated as it involves two electron transfer, and two protonation reactions to give p-iminoquinone as the intermediate; which is subsequently hydrolyzed to give Q as the final product.⁶⁵ 4-Aminophenol acetate was oxidized when the potential was above 0.4 V (with a sodium chloride saturated calomel electrode) in 0.1 M phosphate at pH 7.9.⁶³ It also gave significant background in CV in 0.1 M phosphate, pH 7.5.⁶⁴ It was not widely adopted for AchE sensor because it can undergo spontaneous oxidation.⁴³ 4-Acetoxyphenol offers the advantages of a simpler reaction mechanism, a cleaner background, and improved stability.

Voltammetry study with 2 mM QH₂ shows that the oxidation started from 0 V. The steady state current increased linearly as the potential increased to 0.2 V, then leveled off at higher potential (Supporting Information, Fig. 4). Therefore, we set working potential at 0.35 V for maximum response. Amperometric assay with working potential at 0.35 V gave a linear correlation between steady state current and [QH₂] in the range between 0.05 to 10 mM. Regression analysis showed that the sensitivity was 8.76 μA mM⁻¹. For 4-acetoxyphenol, there was also a low yet appreciable current with a sensitivity of 0.150 μA mM⁻¹. Therefore, the sensitivity for the product was 57 folds higher than that for the substrate. Based on the result from CV study, 4-acetoxyphenol was unlikely to be oxidized at 0.35 V. However, it might undergo spontaneous hydrolysis to give QH₂, which was oxidized to give the background current. Another plausible cause of the background current is from impurity, as the substrate was only 96% pure. These concerns are addressable by a study with purified 4-acetoxyphenol.

3.7. Immobilization of AchE and hydrolysis of 4-acetoxyphenol

We observed the formation of QH₂ by HPLC (data not shown) when 4-acetoxyphenol was incubated with AchE in 0.1 M phosphate buffer, pH 7. The rate of QH2 formation depended on the amount of AchE and incubation time. Since no product was found in the control, we concluded that AchE could catalyze Reaction 1 (Fig. 1).

We used a two-stage protocol in which the protein solution was deposited and air-dried on the working electrode, followed by applying GA solution to crosslink the proteins. Immobilizing AchE on the working electrode significantly increased the steady state current in the solution of 4-acetoxyphenol, and the current is dependent on the load of AchE. Therefore, the transduction based on the hydrolysis of 4-acetoxyphenol is feasible for AchE activity assay.

In the initial optimization of AchE immobilization, the variables included the amount of AchE and BSA, the buffer, and the concentration of GA. The objective was to maximize the steady state current. The assay was carried out in 0.1 M phosphate, pH 7.0 because the reverse reaction in As(III) inhibition (Eq. 1) was slower in phosphate buffer.³⁸ The substrate concentration was 20 mM.

For the first step, we investigated the impact of AchE in the range of 0.5 – 10 μg, BSA in the range of 0 – 30 μg in 0.1 M phosphate with pH between 6.5 – 7.5 (Supporting Information, Table 2). For crosslinking step, we evaluated the GA concentration from 0.001% to 0.015%. Other than the pH value of phosphate buffer, all factors had strong impact on the activity of AchE electrode. The yield of the current ranged from about 7 – 50 μA μg⁻¹ of AchE in the scope of conditions. The concentration of GA was the most significant factor. For crosslinking of 5.3 μg AchE with 10 μg of BSA, increasing GA from 0.005 to 0.015% caused current decrease from 99.7 μA to 68 μA. BSA stabilized AchE in the immobilization. When GA was set at 0.01%, increasing BSA from 10 to 30 μg caused current increase from 88.6 to 105.3 μA. In general, the concentration of GA should be adjusted based on the total protein (Supporting Information, Table 2).

Higher concentration of GA also reduced the sensitivity to As(III) inhibition. For crosslinking of 5.3 μg AchE with 20 μg of BSA, increasing GA from 0.005 to 0.015% caused a reduction of maximum inhibition (measured with 1 mM sodium arsenite) from 69.3% to 65.7%. A similar trend was observed for other combinations of crosslinking conditions (Supporting Information, Table 2).

Importantly, the sensitivity to the inhibition had an inverse correlation with the amount of AchE in the range from 1 to 5 μg (Fig. 4). This trend was observed at both 0.1 mM and 0.5 mM of As(III). To finalize the optimization, we further reduced the amount of AchE to 0.5 μg and adjusted the amount of BSA to 8 μg. The protein mixture was immobilized by crosslinking with 10 μl of 0.0021% GA. The electrodes fabricated under this condition was more sensitive. The inhibition was about 55% with 100 μM As(III).

Fig. 4.

Dependence of inhibition on the amount of immobilized AchE. The electrode was prepared by crosslinking 5 μg (red circle) or 1 μg (blue square) of AchE with 10 μg of BSA by 10 μl of 0.005% GA (n = 3)

There was a clear dependence of steady state current on the substrate concentration with the AchE electrode (Fig. 5). Fitting the data to Michaelis-Menten equation gave a model with R² = 0.9931, indicating the immobilized AchE exhibited similar kinetics to that in the free form, in which the reaction was diffusion controlled. The V_max was 24.2 ±1.1 μA, corresponding to a current density of 121 + 5.25 μA cm⁻². The K_M was 5.92 ± 0.65 mM. When substrate concentration was set at 20 mM, the reaction rate should be 77% of the V_max.

Fig. 5.

Dependence of steady state current on substrate concentration

The immobilization protocol was efficient for the preparation of AchE electrodes. In an evaluation with 56 electrodes, the initial steady state current (A₀) ranged from 17.9-26.8 μA. The A₀ distribution appears to be random (Shapiro-Wilk test W = 0.9793, Prob < W 0.4455) (Fig. 6). The average A₀ was 21.7 ± 1.5 μA. One can routinely prepare 100 electrodes per day. It would be feasible to produce enough sensors to support field study at this productivity.

Fig, 6.

The distribution of A₀ of the AchE electrodes made by glutaraldehyde immobilization

In repeated use, these electrodes showed limited operational stability. The electrodes all underwent activity loss in 10 repeated uses. In each successive use, the steady state current was reduced by about 2 - 4% (Supporting Information, Fig. 5). If re-use is needed, all the measurements would need to have the same repeats. Since the current material cost is less than $3 per electrode, it is preferable to use these electrodes as one-use, disposable sensors.

3.3. Inhibition of immobilized AchE and measurement protocol

It was previously shown that Tris and high pH facilitated the binding of As(III) to free AchE,³⁸ therefore the inhibition study was carried out by incubating the electrode in arsenite solution in 0.1 M Tris-HCl, pH 8.0. The concentration of As(III) was 5, 20, and 100 μM, respectively; and samples were taken over the course of one hour.

The rate and extent of inhibition was found to be similar to the case with free AchE, which was dependent on both time and As(III) concentration (Fig. 7). Fitting the data to a model based on the reversible pseudo-first order mechanism (Eq. 1) gave a k₁ of 2.84 × 10³ M⁻¹ min⁻¹, and a k₋₁ of 1.75 × 10⁻¹ M⁻¹. Apparently, the association and dissociation for immobilized AchE were both faster than those for free AchE. The regression coefficient R² was 0.7475 for this fitting, suggesting that the kinetics might not be fully represented by the model. Nevertheless, it was clear that the protocol previously reported based on a steady state mechanism with rapid binding was unsuitable as the equilibrium was not reached instantaneously.^48-51,52,54 In addition, the mechanism showed that dissociation was still slow enough that the reaction could be treated as being irreversible in protocol development. Since the K_i was 6.16 × 10⁻⁵ M, the lower detection limit should be in low μM concentrations.

Fig. 7.

Inhibition of immobilized AchE by Arsenite (cycles, 5 μM, squares 20 μM, triangles 100 μM)

Based on the kinetic feature, our protocol involved a one-hour incubation of the electrode in arsenite solution in 0.1 M Tris-HCl, pH 8.0 followed by residual activity (i_i) determination. The inhibition was calculated by Eq. 2. The inhibition increased rapidly in the concentration range of 2-20 μM of As(III), then much slower as the concentrations were between 20 – 500 μM. (Fig. 8A). Over the wide dynamic range, the dependence of inhibition on [As(III)] could be represented with a logarithm relationship. The I_Max was typically 70-75%, much higher than those reported earlier. ^48,49

Fig. 8.

The correlation between I% and [As(III)] in Tris-HCl buffer

A focused evaluation at lower range of As(III) revealed detailed characteristics of the sensor (Fig. 8B). Corresponding to 1 to 20 μM As(III), the inhibition increased from 2.7 to 44.9 %. Least square regression gave a straight line with a R² of 0.93. The line had an intercept (I%) of 5.3 ± 1.68, and a slope (I% μM⁻¹) of 2.27 ± 0.18. Therefore, the limit of detection (LOD) was 3σ + intercept = 10.3% of uncorrected inhibition, corresponding to the concentration of As(III) at 2.2 μM. This result was consistent with our kinetic data. The estimated inhibition at 20 μM was 50.7 ± 5.7%.

The current AchE biosensor is useful for field testing of samples with [As(III)] > 2 μM (150 ppb). To expand its application, we expect to increase the sensitivity and precision through precise immobilization and bioreceptor improvement.^43,44,66,67

3.4. Speciation of As(III) and Correlation to ICP-AES measurement

The AchE sensor was evaluated for speciation with mixtures of 5 μM Na₂HAsO₄ · 7H₂O with 0 – 20 μM of NaAsO₂ in 0.1 M Tris-HCl, pH 8.0. The concentration of As (III) was determined by the AchE sensor, while the concentration of total arsenic was validated by ICP-AES.

The recovery of As (III) based on AchE-sensor ranged from 95-170% (Table 1). At concentrations of 2 and 4 μM, the test results were 60-70% higher than the real value. There are also large standard deviations associated to the results. This is probably caused largely by random error; because at low [As(III)] the equilibrium takes longer to reach and more sensitive to random interference. In addition, systematic error caused by linear fit may also be significant.³⁸ The accuracy of determination should be improved by spiking with known amount of As(III) to 8 μM – 20 μM, where the recoveries are between 95 – 109%.

Table 1.

Speciation composition and results by ICP-AES and AchE-Sensor

Test	As(V) (μM)	As(III) (μM)	ICP-AES (μM)	Total As Recovery (%)	AchE-Sensor (μM, n = 5)	Average As(III) Recovery (%)
1	5.0	0.0	5.20	104	−3.70 ± 3.44
2	5.0	2.0	7.19	102	3.40 ± 2.05	170
3	5.0	4.0	9.21	101	6.46 ± 2.30	161
4	5.0	8.0	13.17	101	8.73 ± 1.01	109
5	5.0	14.0	19.33	101	14.11 ± 1.15	100
6	5.0	20.0	25.32	101	19.00 + 1.15	95

The results from the AchE sensor and ICP-AES are highly correlated as shown by the cross-comparison (Fig. 9). Regression analysis gives a line with a slope of 0.992 ± 0.140, an intercept of – 5.13 ± 2.1(μM) and R² = 0.9262. Based on the equation, the x-intercept is 5.17 μM, confirming that the sensor is not sensitive to As(V).⁶⁸

Fig. 9.

Comparison of total As determined by ICP-AES and As(III) determined by AchE-sensor

3.5. Determination of As(III) in simulated groundwater

The impact of groundwater on the AchE sensor was tested using synthetic groundwater. The composition was created based on the water chemistry data collected in the Shepley’s Hill Landfill Superfund Site at Fort Devens, MA.^17,69 The concentration of a cation was the median of its concentrations in 41-43 samples. The concentration of an anion was the median of its concentrations in 36-43 samples (Supporting Information, Table 3). The groundwater was largely anoxic but not sulfidic, with a mean oxidation-reduction potential at −99.5 ± 6.77 mV. Nitrogen existed as ammonia. All iron was Fe(II). The Fe(II) in nitrogen sparged synthetic groundwater was stable for less than 6 hours, therefore the groundwater was prepared freshly. To promote the AchE inhibition by As(III), Tris salt was added to the groundwater sample at a concentration of 0.1 M, pH 8.

The standard curve with groundwater is similar to that with Tris-HCl (Fig. 10). Between 0-20 μM, the inhibition was 6- 49%. The detection limit is 8.6 %, corresponding to 1.1 μM (82.5 μg L⁻¹). The ratio of the regression slopes is 97.6%, suggesting a nearly full recovery in the groundwater matrix.

Fig. 10.

The correlation between I% and [As(III)] in synthetic groundwater

3.6. Storage Stability of the AchE electrode

We tested 6-day activity loss of AchE electrodes in buffers including phosphate or Tris-HCl with pH from 6 to 8, at temperatures from 4 – 25 °C. It was found the electrodes had essentially no activity loss in 0.1 M Tris-HCl, pH 7.0 under ambient temperature (22 ± 2 °C). Prolonged evaluation with AchE electrodes from the same batch showed that they were stable over 150 days (Fig. 11).

Fig. 11.

Storage stability of the AchE electrodes in 0.1 M Tris-HCl, pH 7.0 at 20-24 °C (n=5 for each time point)

Arsenic pollution from anthropogenic sources poses great environmental risk because As concentration is usually very high (> ppm level) at the origin.⁷ Groundwater contamination is the most important pathway for arsenic transportation from anthropogenic sources.⁷ The migration and sequestration can be highly complicated because the migration depends on the geochemistry, hydrology, and climate in local area ^1,11,70 A disposable field sensor for speciation is highly desirable to characterize the spatial and temporal variations of As(III). The speciation information would be useful to better characterize the arsenic plume in groundwater, differentiate the sources of arsenic in groundwater, understand the stability of arsenic associated with the sediment, and evaluate the risk of surface water contamination from groundwater discharge.

4. Conclusion

To address the unmet need for As speciation in field, a disposable AchE sensor has been developed based on the practical needs. A novel reaction sequence for signal transduction was employed to simplify sensor design. The amperometric assay was used to optimize sensor fabrication, evaluate the mechanism for AchE inhibition by As(III), and characterize sensor performance. This sensor has a linear range of 0-20 μM, with LOD of 1-2 μM As (III). It can determine As(III) in the mixture with As(V) as validated by ICP-AES. The sensor has been manually fabricated with a throughput of 100 day⁻¹ person⁻¹. It is stable for 150 days at room temperature. Therefore, this lab-made sensor can be used to characterize water pollution in anthropogenic sites.

An example of the potential application of the biosensor would be for the characterization of groundwater pollution in Shepley’s Hill Landfill Superfund Site. The overall As(III) recovery from simulated groundwater matrix is 97.6% in the range of 0-20 μM (1500 μg L⁻¹). The sensitivity is adequate to locate the hot spots of As(III) along with their spatial and temporal scales. In one survey, the concentrations of total As ranged from 177 – 995 μg L⁻¹ (2.4 – 13.3 μM) in 39 out of 43 samples.⁶⁹ (Supporting Information, Fig. 6). According to water chemistry and prior surveys, the As was expected to be mostly arsenite.⁷¹ Combined with geographic system information, the sensor can be used to generate maps of As(III) plume with 150 μg L⁻¹ boundary concentration.

Fig. 3.

Standard curves for amperometric assay at 0.35 V for QH₂ (blue) and 4-acetoxyphenol (red)