A facile microfluidic paper-based analytical device for acetylcholinesterase inhibition assay for rapid detection of pesticide residues in food -

A facile microfluidic paper-based analytical device for acetylcholinesterase inhibition assay for rapid detection of pesticide residues in food

A novel method has been developed for an improved acetylcholinesterase (AChE) inhibition assay via organic solvent extraction combined spontaneous in situ solvent evaporation on microfluidic paper-based analytical devices. Enzyme pre-immobilization procedure was spared and AChE was added to the system after sampling step until a complete in-situ solvent evaporation process was performed on chip. IC₅₀ levels of the six investigated organophosphate and carbamate pesticides indicated a completely eliminated influence of solvents on AChE behavior with the new method. Most importantly, analytical performances were significantly improved in food sample measurements. Reduction in matrix effect was observed when acetonitrile was adopted for lettuce sample pretreatment instead of water. Studies on different pesticides suggested a remarkably decreased discrimination effect on recoveries from sample pretreatment with the new developed method. The recovery level for phoxim spiked head lettuce samples reached (107.5 ± 14.2) %, in comparison with that of (18.6 ± 1.4) % from water-based extraction. Spiked water and apple juice samples with carbaryl concentration of as low as 0.02 mg/L were also successfully recognized with the present method by visual detection.

Method and materials

AChE was purchased from Sigma-Aldrich (St. Louis, USA). Bovine serum albumin (BSA) was obtained from BioRoYee (Beijing, China). PBS and Tris powders were purchased from Amresco (Cleveland, USA). Pesticide standards including phoxim, chlorpyrifos, carbaryl, triazophos, carbofuran and methamidophos were purchased from Shanghai Pesticide Research Institute (Shanghai, China). HPLC-grade acetonitrile, methanol, acetone, hexane and ethyl acetate were purchased from Merck (Darmstadt, Germany). Whatman™ 1# filter paper was used to fabricate paper chips.

Chip fabrication

Chips employed in current work were fabricated by wax printing method. Briefly, hydrophobic wax patterns were designed by Corel Draw software and printed onto the Whatman™ 1# filter paper using a wax printer (ColorQube 8580, Xerox). The paper was then heated with an oven at 115 °C for 10 min. During heating, the printed wax patterns became melted and penetrated into the paper substrate to form hydrophobic channel walls. The fabricated chip was of 35 mm × 35 mm, consisting of a flower-shaped hydrophilic pattern with one common reagent reservoir located at the chip center, and eight sensing zones. Each sensing zone was with the diameter of 5 mm and was connected to the common reagent reservoir through a hydrophilic channel (4 mm × 1.5 mm).

The effects of different solvents on the activity of AChE were investigated in solution and on chip, respectively. The tested solvents included water, acetonitrile, ethyl acetate, acetone, hexane and methanol. In bulk solution experiment, 70 mL of AChE solution (100 U/mL) and the investigated solvent of the same volume were mixed in a tube and incubated for 10 min before 70 mL of IPA solution (3 mM) was added into the mixture. The color development was recorded after 7 min. A paper-based microfluidic chip was employed for the study of solvent influence. According to primary tests, an individual sensing zone of the chip could be properly filled by 1.10 mL of liquid, so the volume of reagent pipetted onto each sensing zone was fixed at 1.10 mL. Briefly, different solvents were pipetted into the sensing zones (one solvent per zone with 1.10 mL of each solvent) and were let evaporate for 5 min before AChE solution (100 U/mL, 1.10 mL for each sensing zone) was added. The whole chip was then incubated in atmosphere for 10 min before 1.10 mL of IPA solution (3 mM) was added to each sensing zone. Finally, the color message was recorded by a desktop scanner 7 min after the addition of IPA solution and the color intensity (CI) value of each zone was measured by Photoshop software.

Measurement conditions of the on-chip AChE inhibition assay including reagent volumes, pH, IPA and enzyme concentrations were tested and optimized before pesticide measurement. According to the results of the optimization experiment, reactions were carried out at pH 8.0, and concentration of IPA and AChE were set at 3 mM and 100 U/mL for all assays. Standard solutions of six OP and CM pesticides including chlorpyrifos, phoxim, carbaryl, triazophos, carbofuran and methamidophos were prepared in acetonitrile and measured with current method. Pesticide solutions at different concentrations were first dropped onto the sensing zones of each chip with 1.10 mL per zone. The whole chip was placed in air for 5 min to allow the complete evaporation of acetonitrile from chip surface before 1.10 mL of AChE solution was subsequently added in each sensing zone. After 10 min incubation, 25 mL of IPA (pH 8.0) was introduced to the common reagent reservoir at chip center so the capillary action force could drive the solution to fill all sensing zones simultaneously. The chip was then placed into a laboratory-made humidity chamber for color development, and signal collection was performed 7 min later with a scanner. CI of each sensing zone was measured with Photoshop. The obtained values were finally fitted with Hill equation in Origin software to construct the calibration curve of each pesticide. IC50 values were calculated accordingly.

Measurement of OP and CM pesticides in spiked samples with different extraction solvents Pesticides spiked in head lettuce samples at the concentrations close to each relevant IC50 were measured and recovery data were compared between different groups categorized by extraction solvents. Pesticides involved in this section included phoxim, carbaryl, carbofuran and methamidophos. And the spiked concentrations of the four pesticides were 0.8, 0.4, 0.01 and 2 mg/kg, respectively. The four different pesticides were added separately into prechopped head lettuce samples to yield spiked samples. One portion (3.0 g) of each spiked sample was extracted with 3 mL of acetonitrile while another portion (3.0 g) of the same sample was extracted with 3 mL of ddH₂O. The tubes containing sample and corresponding solvent were then vortexed for 5 min and centrifuged at 4500 rpm for 10 min to obtain the supernatants for detection. 0.75 g of solid NaCl was added into the acetonitrile tube before centrifugation to assist phase separation. Non-spiked samples were prepared by the same protocol and the obtained supernatants were used to prepare matrix-matched standards. CI values of samples were interpolated in the calibration curve obtained from the matrix-matched standards on the same chip.

Water and apple juice samples spiked with 0.02 mg/L of carbaryl were used to test the possibility of rapid visual screening of samples with extremely low pesticide residual levels by the new developed method. Prior to the detection, pH of the apple juice samples was adjusted to 7.0 by 1 M NaOH. 1 mL of ethyl acetate was added into 20 mL of the spiked sample for extraction. After the addition of 4 g NaCl, the mixture was vortexed and centrifuged to achieve supernatant for assessment. Samples and blank controls were tested with the new method and visual detection was achieved by comparing the final color intensity.

Results

The influences of different solvents on pesticide extraction

Six different OP and CM pesticides which presented different water solubility and were frequently detected in routine assessment of vegetable and tea samples were chosen as model pesticides for extraction tests. Samples spiked with the pesticide mixtures were extracted with different constituted solvents. Recoveries of different pesticides under each conditionwere measured with the aid of UPLC-MS/MS analysis. Non-polar solvent like hexane was also tested for the pesticide extraction. Acetonitrile presented the best extraction efficiency among the four solvents/mixtures in both wet (cabbage) and dry (tea) sample cases. The recoveries of all six pesticides ranged between (83.59 ± 2.90) % ~ (98.30 ± 0.92) % for cabbage and 74.84 ± 1.80%, 106.39 ± 0.36% for tea samples when extracted with acetonitrile. In contrast, water which was widely employed in sample extraction for conventional AChE inhibition assays provided quite deficient extractions. And the achieved extraction recoveries depended highly on the water-solubility of the target pesticides under this situation. This dependence resulted in serious differentiation effect among pesticides. For instance, using pure water as the extraction solvent, 77.22% of methamidophos spiked in cabbage was detected in the extract while only 14.81% of chlorpyrifos was measured under the same condition. The differentiation effect became even worse for tea samples of which the water content was lower. Study result indicated that pesticides with low water solubility tended to retain in sample matrices when the sample was treated with water or mixture containing low ratio of organic solvent and thus presented inferior extraction recoveries. For instance, the recoveries of chlorpyrifos and phoxim extracted with water or mixed solvents were all below 20% in both cabbage and tea samples. Meanwhile, it was observed in the test that other than pesticide’s physico-chemical properties, water content in sample matrix also exerted influence on extraction outcome. As a result, the four pesticides with low water-solubility showed extremely low recoveries when they were extracted by water dominated solvent mixtures from tea samples. Chlorpyrifos and phoxim spiked in tea samples were involved as model pesticides based on the result of the above tests. The new results confirmed that compared to pure water group, low concentration (5% v/v) of methanol did not make observable difference in extraction for both of the pesticides. And even the 50% (v/ v) methanol-water mixture provided quite limited improvement. The recovery data of chlorpyrifos were (7.0 ± 2.9) %, (5.7 ± 0.7) %, (15.7 ± 0.9) % and (90.3 ± 1.2) % for water, 5% methanol-water, 50% methanol-water and 100% methanol, respectively. According to these experiments, it is reasonable to assume that pesticide extraction with low content of organic solvents is unreliable, and that high concentration (or pure) organic solvent based sample pretreatment is still preferred and necessary for acceptable sample pretreatment.

The influences of different solvents on the AChE activity was then assessed in bulk solution as well as on paper-chips in current work. In solution condition, AChE was first incubated with water and 5 routine organic solvents (including acetonitrile, acetone, ethyl acetate, methanol and hexane), respectively. IPA was subsequently dropped in to probe AChE activity after the incubation. The mixture rapidly turned dark blue in water tube upon the addition of IPA solution. In sharp contrast, there was almost no color alteration in the tube containing acetonitrile, methanol, acetone or ethyl acetate even after hours, revealing negligible hydrolysis of IPA and thus a complete loss of AChE activity in these solvents. Replacement of H₂O by organic solvent molecules changed the secondary structure of the protein and led to the denaturation of the enzyme. In the case of non-polar solvent like hexane, it was much more difficult for the molecule to penetrate into the aqueous phase, where the enzyme was located in. Most importantly, hexane is not a hydrogen bond donor or acceptor. Hence, it could not occupy the active sites on enzyme from H₂O, and AChE retained the activity thereby. When the experiments were performed on commercial rapid detection cards, the phenomenon was basically consistent with the result from bulk solution group. In order to compare the data from different experimental groups through the whole study, each CI value obtained was standardized by dividing the CI value from the control device (incubated with water) of the same test. Compared to the control group incubated with H₂O, the colors from organic solvent groups were significantly lighter, suggesting inhibited AChE activity in these solvents. Serendipitously, some difference in phenomenon between the experiments on chip and the tests in tube was observed in the above tests. Utilizing non pre-immobilized enzyme reagent and spontaneous solvent evaporation on chip, developed a novel protocol for the enzyme inhibition assay based pesticide measurements on chip. Pesticide dissolved in organic solvent was loaded onto the chip before any other reagents were added. The organic solvent evaporated rapidly due to the large specific surface area of mPADs. AChE introduction was set at 5 min after sampling in current study for the sake of operation consistency. By this means, solvent’s contact with AChE could be fundamentally eliminated.

The six OP and CM compounds, including chlorpyrifos, phoxim, carbaryl, triazophos, carbofuran and methamidophos, were prepared at a series of concentrations in acetonitrile and measured under the optimized condition. According to the corresponding calibration curves, the LODs of the six pesticides for our method were calculated to be 0.77, 0.39, 0.25, 1.29, 0.006 and 1.39 mg/L, respectively. Pesticide standards were also dissolved and measured in acetone, ethyl acetate and methanol. The data were consistent with those from acetonitrile solution. All these results suggested a completely eliminated influence of solvents on AChE behavior with the new protocol and thus in theory, pesticides dissolved or extracted by any routine solvents could be directly measured.

In the conventional AChE inhibition assay strategies, using bulk solution reaction or rapid detection cards with preimmobilized AChE, the employment of organic solvent was problematic. In current work, we eliminated the organic solvent effect on AChE activity through performing solvent evaporation on chip before the addition of AChE solution. Different pesticides (including phoxim, carbaryl, carbofuran and methamidophos) spiked in head lettuce samples were extracted and measured, respectively. And the spiked concentration of each pesticide was set around the individual LOD obtained with standard solutions in order to validate the sensitivity and practicality of current method in real world sample assessment.

Acetonitrile was selected as the extraction solvent in this section as it was prevalently used in routine sample pretreatment. When the spiked samples were extracted by water, recoveries were found to vary regularly according to the pesticides’ physicochemical property. The water-miscible pesticides were extracted more effectively by aqueous solution, while those with poor water solubility could hardly be accumulated in aqueous phase. As a result, significant discrimination effect emerged among the tested pesticides, with recoveries ranging from 18.6% to 73.2% methamidophos. By resolving the inhibition problem of organic solvents on enzymes, experiments based on organic solvent extraction provided significantly improved results. The recovery of phoxim reached 107.5% when acetonitrile was used as the extracting solvent, compared to 18.6% in water-extract based measurement. Recovery data for the rest three pesticides ranged from 75% to 112%. The well-improved recoveries and significantly decreased discrimination effect with organic solvent extraction guaranteed the accuracy of present method in real sample measurement. The matrix effect was frequently observed in water extracts measurement in this study. Interestingly, similar matrix effect did not appear in acetonitrile extracts. It was assumed that these endogenous AChE inhibitors were prone to be distributed to aqueous phase against organic solvent phase during acetonitrile extraction process.

Different pesticides could be detected with similar recovery levels by current method, which is fairly important in real sample screening for decreasing false negative results. Meanwhile, the compatibility of enzyme reaction and different solvents in current method allows more sensitive assessment for some special circumstances. For instance, the pesticide concentration in contaminated surface water is normally much lower than the corresponding IC50. Water and apple juice samples spiked with 0.02 mg/L of carbaryl were used to test the possibility for rapid screening of the type of samples mentioned above. It was worth mentioning that the spiked concentration was about 12.5 times lower than its IC50 (0.253 mg/L for carbaryl) obtained with standard solutions. A water-immiscible polar solvent ethyl acetate was employed to perform carbaryl extraction from spiked water and apple juice samples. The samples were first mixed thoroughly with ethyl acetate at 1:20 (solvent: sample, v: v) ratio and then the obtained supernatant was pipetted directly onto chip and measured. Both of the spiked samples were successfully identified. The differences could also be easily identified by the naked eye. The whole process required no additional facilities except for an ordinary pipette and the whole measurement was accomplished within 30 min. Considering that the original IC50 of carbaryl obtained with standard solutions in current work was around 0.25 mg/L, this result suggested a pesticide enrichment of more than 10 times was achieved during the organic solvent extraction, and hence the detection sensitivity was remarkably increased compared to the conventional AChE assay protocol without solvent extraction. Pre-enrichment of pesticide with such method required the employment of a water-immiscible solvent and correspondingly, it was expected to perform well only with pesticides of low watersolubility. Measurement of high polar pesticides with good water solubility such as methamidophos was expected to be difficult. However, this method could still be useful when the list of potential contaminants was known beforehand. Meanwhile, it was also noticed that slight signal suppression occurred in the measurement of apple juice with the pesticide pre-enrichment protocol. In order to realize more accurate quantitative assessment of such samples, sample pre-cleaning might still be preferred when pre-enrichment of potential target compounds was adopted.

In conclusion, a novel method for AChE inhibition assay based pesticide measurement was developed in current work, highlighting the use of organic solvents for sample pretreatment and in situ spontaneous evaporation of solvents on mPADs. Compared to conventional biosensor strategies, AChE was introduced onto chip after sampling step and enzyme pre-immobilization process was spared in current method. Ideal extraction and measurement of different spiked pesticides were achieved with the present strategy. The decrease of recovery discrimination among pesticides was significant with current protocol which imaginably would contribute to reducing false negative results and improving method reliability in real world food sample analysis, especially for the measurement of pesticides with poor water-solubility. The combination of organic solvent extraction strategy and mPAD technology provided a rapid, accurate and reliable protocol for pesticide residues analysis. For accurate quantitative measurement of samples rich in endogenous active compounds, further integration of more functional elements like pre-cleaning unit might be necessary on the platform. Compared to the conventional rapid detection dipstick or card protocol for pesticide detection, there is additional liquid manipulation steps needed in practice, mostly from the on-line enzyme introduction process. A reliable pesticide extraction combined AChE inhibition assay protocol was achieved in this paper which was critical for real world sample assessments. This is the first report on direct sampling of organic extracts for AChE inhibition assay and it might provide a new perspective for other enzyme activity based detection or enzyme reactions.

Reference:

Jin, L., Hao, Z., Zheng, Q., Chen, H., Zhu, L., Wang, C., Liu, X. and Lu, C., 2020. A facile microfluidic paper-based analytical device for acetylcholinesterase inhibition assay utilizing organic solvent extraction in rapid detection of pesticide residues in food. Analytica Chimica Acta, 1100, pp.215-224.