Chemical Structures of Organophosphate Pesticides Evaluated in a Cross-Reactivity Study

DOCTOR OF PHILOSOPHY

Development of bioanalytical approaches as model tools for the rapid detection of pesticides

Bin Salom, Abdullah

Award date: 2018

Awarding institution: Queen's University Belfast

Link to publication

Terms of use All those accessing thesis content in Queen’s University Belfast Research Portal are subject to the following terms and conditions of use

• Copyright is subject to the Copyright, Designs and Patent Act 1988, or as modified by any successor legislation • Copyright and moral rights for thesis content are retained by the author and/or other copyright owners • A copy of a thesis may be downloaded for personal non-commercial research/study without the need for permission or charge • Distribution or reproduction of thesis content in any format is not permitted without the permission of the copyright holder • When citing this work, full bibliographic details should be supplied, including the author, title, awarding institution and date of thesis

Take down policy A thesis can be removed from the Research Portal if there has been a breach of copyright, or a similarly robust reason. If you believe this document breaches copyright, or there is sufficient cause to take down, please contact us, citing details. Email: [email protected]

Supplementary materials Where possible, we endeavour to provide supplementary materials to theses. This may include video, audio and other types of files. We endeavour to capture all content and upload as part of the Pure record for each thesis. Note, it may not be possible in all instances to convert analogue formats to usable digital formats for some supplementary materials. We exercise best efforts on our behalf and, in such instances, encourage the individual to consult the physical thesis for further information.

Download date: 27. Sep. 2021

Development of bioanalytical approaches as model tools for the rapid detection of pesticides

Abdullah Bin Salom

This thesis is submitted to the Queen’s University Belfast for the degree of Doctor of Philosophy

Institute for Global Food Security School of Biological Sciences Faculty of Medicine, Health and Life Sciences Queen’s University Belfast

October 2017

Thesis Acknowledgement

I am grateful to Allah Almighty for the good health and well-being that were necessary to complete this long journey.

I gratefully acknowledge my employer, the Saudi Food and Drug Authority, for sponsoring my PhD studies as well as understanding the challenges that I have faced during all these years.

I would like to express my sincere gratitude to my supervisor, Dr. Katrina Campbell, for her continuous support of my PhD study and related research, and for her patience, motivation, and immense knowledge. Her guidance has helped my research and the writing of this thesis.

In addition, my sincere thanks also goes to my ex-supervisors, Dr. Iva Chianella and Professor Mohammed Zourob, for their encouragement, insightful comments and continued support, even after I moved from Cranfield University.

I am also grateful to Sara, a post-doctoral research fellow, and Tom, a technician both of whom worked within the Institute for Global Food Security. I am extremely thankful for their shared expertise and valuable guidance. I am also grateful to all the laboratory staff for their assistance. I would also like to thank my fellow doctoral students for their feedback, cooperation and of course friendship.

I am also grateful to my son, Basil, for his patience and emotional support during my PhD studies.

Summary

Pesticides are chemicals used internationally in the agricultural sector to repel and kill pests. However, misuse or misapplication of pesticides can cause detrimental effects for humans and animals, including livestock and non-target organisms. In addition, many studies have linked pesticides with certain adverse symptoms and death. As a result, regulatory authorities around the world monitor pesticide levels in food supplies and update maximum residue levels as needed. As part of addressing the concerns of pesticide presence in food stocks, rapid, sensitive and cheap methods for sample preparation and early detection systems need to be developed to detect pesticide residues in food for on- site analysis before it reaches consumers.

The introduction to Chapter 1 discusses pesticide classifications and their use in food production, regulations governing pesticides, their use and current methods used for pesticide analysis. Also, molecularly imprinted polymers (MIPs) and biosensors related to organophosphate pesticides (OPPs) were discussed. It addresses MIP production, synthesis and different materials involved in these processes that occur in the literature, including templates, functional monomers, cross-linkers, initiators and solvents. In addition, this review discusses various analytical applications, such as solid-phase extraction, chromatographic columns, biosensors and catalysis as well as the use as micro- and nanoparticles. Furthermore, it outlines current pesticide detection methods employed in environmental and food safety applications. It discusses the materials and biosensors used for target enrichment and pre-concentration processes, recognition elements such as enzymes, antibody whole-cells, MIPs and aptamers, and other techniques that use free recognition elements and nanoparticles to construct biosensors for OPPs detection.

Chapter 2 discusses the development of MIPs, of which the first attempt used precipitation polymerisation before thermos-polymerisation methods to evaluate four different porogens (acetonitrile, dichloromethane, chloroform and toluene) and their impact on MIP affinity and selectivity. Initially, chlorpyrifos (CPF) and chlorpropham (CIPC) were selected as templates and the synthesised MIPs were evaluated for solid phase extraction (SPE) application in both organic and aqueous solutions as loading

iii

carriers. However, only the MIP prepared using acetonitrile as the porogen for CPF in aqueous solution as the carrier showed promising results that allowed for a full optimisation for future work. During this study polymer capacity, washing solutions and selectivity were also investigated.

Chapter 3 focuses on the development of a detection system based on an immune-optical planar waveguide biosensor for the rapid selective individual screening of two organophosphate pesticides, chlorpyrifos (CPF) and parathion (PT). The assay conditions were investigated and optimised for CPF detection using MBio assay buffer in a single assay. A cross-reactivity study with pesticides of similar structure was performed in buffer, which showed high selectivity towards CPF. To demonstrate a real application, the biosensor assay was evaluated for the determination of CPF in strawberries at the maximum residue level of 200µg/kg using a simple extraction and dilution approach. For the detection of PT, a broad specificity antibody was employed. The method was first investigated and optimised in MBio assay buffer for sensitivity and selectivity of the assay and then evaluated for the detection of other pesticides of similar chemical structure. The results indicated that the biosensor recognized a number of pesticides. The biosensor assay was again evaluated for the determination of parathion in strawberries. Due to the lower maximum residue level required of 50µg/kg and to improve the sensitivity of the assay, an additional sample preparation step in drying the strawberry extracts and reconstituting them in buffer was necessary.

Chapter 4 advances the work discussed in Chapters 3, and describes the development of a proof of concept approach to create a multiplex microarray to detect two agricultural pesticides, CPF and PT, simultaneously in one assay using a multiplex MBio planar waveguide biosensor. However, due to the challenges and complexities of the antibodies employed the results showed that the multiplex assay was not accurate for the detection of both pesticides due to the cross-reactivity of the antibody used to detect PT in this study. Nonetheless, the results of the MBio planar waveguide biosensor used in this research have demonstrated that rapid individual assays have been developed for CPF and PT. Due to the portability of this technology, these results are promising for field analysis with improvements in sample preparation and sensitivity being required. In addition, for both pesticides, the multiplex assay remains a possible concept for

transferring this technology to the field of pesticide analysis for multiple pesticide detection and recognition. However, instead of purchasing commercial antibodies as required due to time constraints in this project, the design of highly specific antibodies for individual pesticides would be preferential.

TABLE OF CONTENTS

Thesis Acknowledgement ...... ii Summary ...... iii TABLE OF CONTENTS ...... vii LIST OF FIGURES ...... xiii LIST OF TABLES ...... xviii LIST OF ABBREVIATIONS ...... xx Chapter 1: Literature review ...... 1 1.1 Abstract ...... 2 1.2 An Introduction to Pesticides ...... 4 1.3 Chemical classification of pesticides ...... 5 1.3.1 Organochlorines ...... 5 1.3.2 Carbamates ...... 5 1.3.3 Pyrethrins and pyrethroids ...... 6 1.3.4 Organophosphates ...... 6 1.4 Pesticide use in food production ...... 9 1.4.1 Pesticide use in agriculture ...... 9 1.4.2 Pesticide use in aquaculture ...... 10 1.5 Health implications ...... 11 1.5 Legislation governing pesticide use and harmonisation ...... 11 1.6 Current methods of pesticide analysis ...... 13 1.6.1 Analytical Methods of Analysis ...... 13 1.6.1.1 High Performance Liquid Chromatography (HPLC) ...... 14 1.6.1.2 Liquid chromatography/mass spectrometry (LC/MS) ...... 15 1.6.1.3 Gas chromatography/mass spectrometry (GC/MS) ...... 16 1.6.2 Bioanalytical Techniques ...... 17 1.6.2.1 Immunological assays ...... 18 1.6.2.2 Enzyme-linked immunosorbent assay (ELISA)...... 20 1.6.2.3 Lateral-flow (LF) ...... 20 1.6.2.4 Biosensors ...... 21 Bio-recognition receptors ...... 24 Antibody based biosensors ...... 25 Aptamers-based biosensors ...... 27 Enzyme-based biosensor methods ...... 28 Whole-cell based biosensors ...... 31 Molecularly imprinted polymer-based biosensors ...... 32 Nanomaterials ...... 36 1.6.2.5 Microarray technology to nanosensors for food analysis ...... 37 1.7 Sample preparation ...... 40 1.7.1 Liquid-liquid extraction (LLE) ...... 40

vii

1.7.2 Solid phase extraction (SPE) ...... 41 1.7.3 Other techniques ...... 41 1.7.4 Molecularly imprinted polymers ...... 42 1.7.4.1 Polymerization procedures ...... 46 Bulk polymerisation ...... 48 Precipitation polymerisation ...... 48 1.7.4.2 Templates ...... 49 1.7.4.3 Functional monomers ...... 50 1.7.4.4 Cross-linkers ...... 53 1.7.4.5 Solvents (porogens) ...... 53 1.7.4.6 Initiator ...... 54 1.7.4.7 Rational synthesis of Molcular imprinting ...... 55 1.7.4.7 Application of MIPs for sample preparation ...... 56 1.7.4.7.1 Sorbents for solid-phase extraction...... 56 1.8 Target pesticides ...... 58 1.8.1 Chlorpyrifos ...... 58 1.8.1.1 Toxicity ...... 60 1.8.1.2 Exposure ...... 60 1.8.1.3 Regulations ...... 62 1.8.2 Parathion ...... 63 1.8.2.1 Toxicity ...... 64 1.8.2.2 Exposure ...... 65 1.8.2.3 Physical and chemical characteristics...... 65 1.8.2.4 Regulation ...... 66 1.8.2.5 Analysis ...... 67 1.8.3 Chlorpropham ...... 67 1.9 Aims and objectives ...... 69 1.9.1 Aims ...... 69 Chapter 2: Development and evaluation of molecularly imprinted polymers to concentrate CPF and CIPC in water samples ...... 71 2.1 Abstract ...... 72 2.2 Introduction ...... 73 2.2.1 MIP applications for sample preparation ...... 73 2.2.2 General extraction procedure ...... 73 2.2.3 The chapter aim ...... 76 2.3 Materials and methods ...... 77 2.3.1 Reagents and Chemicals ...... 77 2.3.2 HPLC analysis ...... 77 2.3.3 Preparation of stock and working solutions ...... 78 2.3.4 Preparation of MIPs by precipitation polymerisation (first attempt) .... 78 2.3.5 Preparation of MIPs by bulk polymerisation ...... 79

viii

2.3.5.1 Molecularly imprinted polymer–solid-phase extraction procedure (loading CIPC and CPF in organic solvent) ...... 80 2.3.5.2 A rapid optimisation of packing materials and loading volumes ..... 81 2.3.5.3 Polymer mass capacity ...... 81 2.3.5.4 Optimisation of washing protocols ...... 82 2.3.5.5 MISPE extraction procedure in aqueous solution ...... 82 2.3.5.6 Selectivity study ...... 83 2.4 Results and discussion ...... 83 2.4.1 Precipitation polymerisation ...... 84 2.4.1.1 Initial synthesis by precipitation polymerisation ...... 84 2.4.1.2 Evaluation of different factors ...... 84 2.4.2 Bulk polymer preparation ...... 86 2.4.3 The development of standard calibration curves by using HPLC ...... 87 2.4.4 Evaluation of loading solvents for CIPC and CPF ...... 89 2.4.4.1 CIPC evaluation ...... 89 2.4.4.2 CPF evaluation ...... 90 2.4.4.3 Loading CPF and CIPC in aqueous solution ...... 91 2.4.5 The MIP packing volume ...... 91 2.4.6 Polymer mass capacity ...... 92 2.4.7 Washing optimisation ...... 92 2.4.8 Adsorption experiment ...... 94 2.4.8.1 Evaluation of different CIPC-polymers ...... 94 2.4.8.2 Evaluation of different CPF-polymers ...... 96 2.4.9 Selectivity study ...... 98 2.5 Conclusions...... 102 Chapter 3: Detection of organophosphate pesticides using a planar waveguide immunosensor ...... 103 3.1 Abstract ...... 104 3.2 Introduction ...... 106 3.2.1 The chapter aim ...... 109 3.3 Materials and methods ...... 110 3.3.1 Reagents and chemicals ...... 110 3.3.2 LC-MS/MS determination of CPF and PT in strawberry samples ...... 110 3.3.3 Immunological Reagents ...... 111 3.3.4 Characterisation of immunological reagents ...... 112 3.3.4.1 UV-VIS Spectroscopy ...... 112 3.3.4.2 Characterisation of immunological reagents by ELISA ...... 112 Determination of the optimal coating antigen concentration and antibody titre...... 112 Evaluation of antibodies sensitivity by ELISA ...... 113 3.3.5 Microarray construction ...... 114

3.3.6 General MBio array protocol ...... 116 3.3.7 Optimisation of coating antigen concentrations and antibody titre ..... 117 3.3.8 Calibration curves in buffer ...... 118 3.3.9 Specificity of antibodies in buffer ...... 119 3.3.10 Sample preparation development and evaluation ...... 119 3.3.10.1 Chlorpyrifos extraction procedure ...... 120 Evaluation of 100% and 20% (MeOH) as extraction procedures at a 1:8 extract dilution ...... 120 Evaluation of 100% and 20% (MeOH) as extraction procedures at a 1:4 extract dilution ...... 120 Re-evaluation of the extraction procedure by MeOH and a 1:8 extract dilution at higher CPF concentration ...... 121 Re-evaluation of the extraction procedure by MeOH methanol and a 1:4 extract dilution ...... 121 Examination of matrix effects and assay recovery ...... 122 Planar Waveguide Assay for CPF ...... 123 3.3.10.2 Parathion extraction procedure ...... 123 Evaluation of 100% (MeOH) as extraction procedure and a 1:4 extract dilution...... 124 Evaluation of 100% (MeOH) as extraction procedure and a 1:8 extract dilution...... 124 The extract drying and reconstitution step ...... 124 Examination of matrix effects and assay recovery ...... 125 Planar waveguide assay for PT ...... 126 3.3.11 Validation study ...... 126 3.3.11.1 Limits of detection (LOD) and limits of quantification (LOQ) .... 127 3.3.11.2 Intra and inter -day reproducibility ...... 127 3.3.11.3 Recovery ...... 128 3.4 Results and discussion ...... 128 3.4.1 Sample screening ...... 129 3.4.2 Initial characterisation ...... 132 3.4.2.1 UV-VIS Spectrophotometer...... 132 3.4.2.2 Chequerboard ELISA for determination of optimal ratios ...... 134 Chlorpyrifos-coating antigen and anti-CPF antibody dilution ...... 134 Parathion - coating antigen and ant- PT-mAb 1E2 antibody dilution ...... 136 3.4.3 MBio assay protocol ...... 138 3.4.3.1 MBio assay optimisation ...... 138 Chlorpyrifos ...... 139 Parathion ...... 140

3.4.4 Development of calibration curves using the MBio Biosensor ...... 142 3.4.5 Specificity of antibodies in the MBio assay buffer ...... 144 3.4.6 Matrix evaluation ...... 149 3.4.6.1 Development of a sample preparation procedure for chlorpyrifos extraction ...... 150 3.4.6.2 Development of a sample preparation procedure for parathion extraction ...... 155 3.4.6.3 Drying and reconstitution procedure ...... 158 3.4.6.4 Development of a calibration curve for strawberry samples ...... 159 3.4.7 Validation study ...... 162 3.4.7.1 Chlorpyrifos ...... 162 Limits of detection and quantification ...... 162 Intra-and-inter batch reproducibility ...... 163 Recovery ...... 163 Comparison of the developed immunoassay method ...... 163 3.4.7.2 Parathion ...... 165 Limits of detection and quantification ...... 166 Intra-day study ...... 166 Recovery ...... 166 Comparison of the developed immunoassay method ...... 167 3.5 Conclusion ...... 169 Chapter 4: Challenges in the development of a multiplexing array to detect organophosphate pesticides ...... 171 4.1 Abstract ...... 172 4.2 Introduction ...... 173 4.2.1 The chapter aim ...... 175 4.3 Materials and Methods ...... 176 4.3.1 Reagents and chemicals ...... 176 4.3.2 Microarray construction ...... 176 4.3.3 Immunoassay procedure ...... 176 4.3.4 Evaluation of a multiplexed assay for the simultaneous detection of CPF and PT in buffer ...... 177 4.3.4.1 CPF-pAb specificity to both coating antigens (CPF-OVA /H11- OVA) ...... 177 4.3.4.2 PT-mAb 1E2 specificity to both coating antigens (CPF- OVA/H11-OVA) ...... 177 4.3.5 Development of multiplexed calibration curves in buffer ...... 177 4.3.6 Real sample analysis ...... 178 4.4 Results and discussion ...... 179 4.4.1 MBio assay protocol ...... 179 4.4.2 Specificity evaluation ...... 179

4.4.2.1 Evaluation of pesticide conjugates versus CPF-pAb ...... 179 4.4.2.2 Evaluation of pesticide conjugates versus PT-mAb 1E2 ...... 180 4.4.3 Evaluation of a multiplexed assay for OPPs detection in buffer ...... 182 4.4.3.1 Evaluation of a multiplexed assay for OPPs detection in a real sample ...... 185 4.5 Conclusion ...... 187 Chapter 5: Conclusions ...... 188 5.1 Conclusions...... 189 5.2 Future work and perspectives ...... 193 Chapter 6: References ...... 197 Appendices: ...... 243

xii

LIST OF FIGURES

Figure 1-1: General chemical structure of OPPs ...... 7

Figure 1-2: Schematic presentation of a biosensor ...... 22 Figure 1-3: Graph shows the number of published papers for biosensors and pesticides in general and OPP detection over the last 33 years (from Web of Science at 28/08/2017)...... 23 Figure 1-4: The number of published papers for molecularly imprinted polymer present in the literature with different search terms in the last 33 years (from Web of Science at 21/09/2017)...... 43 Figure 1-5: Schematic diagram of molecularly imprinted polymer synthesis to produce template-monomers network by non-covalent interaction...... 47

Figure 1-6: Chemical structure of AIBN ...... 54

Figure 1-7: Radical polymerisation process ...... 55

Figure 1-8: Structure of Chlorpropham ...... 68

Figure 2-1: Principles of solid phase extraction (Lucci, et al., 2012)...... 74 Figure 2-2: General procedure for the synthesis of the bulk imprinted polymers (Haupt, 2008) ...... 80 Figure 2-3: An image of particle size of approximately 1 µm produced by precipitation polymerisation and taken by a light microscope, 40X magnification...... 84 Figure 2-4: Some of the polymers produced by protocols of different precipitation polymerisations. Relatively solid (A), nano-polymer (B), milky-like (C) and gel- like (D)...... 85 Figure 2-5: The bulk polymer after 24h polymerisation (A) and after grinding and sieving (B)...... 87 Figure 2-6: A calibration curve of CIPC in ACN at the concentration range of 0– 1000 ng/mL...... 88 Figure 2-7: A calibration curve of CPF in ACN at the concentration range of 10– 1000 ng/mL...... 88 Figure 2-8: CIPC washed amount (%) with different mixtures of washing solutions (5ml). Comparison between washed amounts of CIPC from columns loaded with 30 ml of 500ng/ml...... 92 Figure 2-9: CPF washed amounts (%) with different mixtures of washing solutions (5ml). Comparison of washed amounts of CPF from column loaded with 30 ml of 500ng/ml...... 93

xiii

Figure 2-10: Washed amount of CIPC after loading 30 ml (500 ng/mL CIPC) prepared in water onto MIPs and NIPs. 5 mLs ACN:water (20:80,v:v) was used for washing, and four MIPs were synthesised in different porogens...... 95 Figure 2-11: Recovery of CIPC in an elution fraction after loading 30 mL (500 ng/mL) in an aqueous solution and washing. Elution step: 5 mLs, MeOH: acetic acid (90:10;v:v) was used for elution. Four MIPs were synthesised in different porogens...... 96 Figure 2-12: Washed amounts of CPF after loading 30 ml (500 ng/mL CPF) prepared in water for MIPs and NIPs. 5 mLs methanol-water (50:50, v:v) was used for washing. Four MIPs were synthesised in different porogens; all were dried and reconstituted in 1 mL of ACN...... 96 Figure 2-13: Recovery of CPF from an elution fraction after loading 30 mL (500 ng/mL) in an aqueous solution and washing. 5 mLs, MeOH: acetic acid (90:10;v:v) was used for elution. Four MIPs were synthesised in different porogens; all were dried and reconstituted in 1 mL of ACN...... 98 Figure 2-14: Chemical structures of compounds used to test their affinities towards CPF-polymers...... 99 Figure 2-15: Percentage of CPF and other compounds washed off from the MIP with 5 mLs (MeOH;H2O; 50:50, v:v)...... 100 Figure 2-16: Percentage of CPF and other compounds recovered after elution with 5 mLs MeOH: acetic acid (90:10; v:v)...... 101 Figure 3-1: MBio SnapEsi reader and a rack of assay cartridges. The reader is connected to a laptop computer...... 107 Figure 3-2: Schematic representation of the planar waveguide used in the MBio biosensor...... 108

Figure 3-3: The Scienion sciFLEXARRAYER S5 nanoplotter...... 115 Figure 3-4: MBio components and an assembly tool. Cartridge components include (A) vacuum tool, (B) upper, (C) lid, (D) wick pad, (E) waveguide, (F) gaskets and (G) assembly press...... 116 Figure 3-5: A representative schematic image showing a 2×22 array format and the array layout of different conjugates on each waveguide slide for optimsation...... 117 Figure 3-6: Representative schematic for an array layout used for printing and devloping a calibration curve in buffer at 100 µg/mL of coating antigen. .... 118

Figure 3-7: Chromatogram of CPF detection using UPLC-MS/MS in methanol. 129

Figure 3-8: Chromatogram of PT detection using UPLC-MS/MS in methanol. .. 130

Figure 3-9: CPF calibration curve for strawberry extracts (0-200 ng/mL)...... 130

xiv

Figure 3-10: PT Calibration curve for strawberry extracts (0-200 ng/mL)...... 131 Figure 3-11: Screening of strawberry blank extract showing no distinctive peaks for A) CPF at 4.67 min and B) PT at 4.04 min...... 132 Figure 3-12: Comparison of ultraviolet absorption spectrums of CPF standard, CPF-OVA and OVA from 220 nm to 400 nm...... 133 Figure 3-13: Comparison of ultraviolet absorption spectrums of PT standard, H11- OVA and OVA showing the wavelength from 220 nm to 400 nm...... 134 Figure 3-14: Standard calibration curves for CPF illustrating percent normalised response values versus CPF concentrations in icELISA format for CPF.pAb. Each value represents the mean of two replicates in buffer solutions. Plates were coated with 100 µL of 0.25 µg/100µL of CPF-OVA...... 136 Figure 3-15: Calibration curves of PT illustrating percent normalised responses versus PT concentrations in icELISA format for PT-mAb 1E2. Each value represents the mean of two replicates in buffer solutions. Plates were coated with 100 µL of 0.25 µg/100µL of H11-OVA...... 138 Figure 3-16: Results of different concentrations of coating antigens against an individual mixture of CPF standard and CPF-pAb dilutions, (A) 1/100 and (B) 1/250...... 140 Figure 3-17: Results from different conjugate concentrations against PT standard concentrations of 0, 100 and 1000 ng/mL. (A) Antibody dilution at 1/250, and (B) antibody dilution at 1/500...... 141 Figure 3-18: Chlorpyrifos calibration curve in MBio assay buffer showing fluorescence intensity versus CPF concentration using CPF-pAb (1/100) and CPF-OVA (100 µg/mL)...... 143 Figure 3-19: Parathion calibration curve in MBio assay buffer showing fluorescence intensity versus PT concentration using PT m-Ab IE (1/250) and HII-OVA (100 µg/mL)...... 143 Figure 3-20: Chemical structures of organophosphate pesticides evaluated in a cross-reactivity study...... 145 Figure 3-21: Cross-reactivity of CPF-pAb with structurally related OP insecticides. Figure A shows analytes with no cross reactivity at 10000 ng/mL and figure B shows those with some degree of cross-reactivity. CPF is presented in both figures...... 146 Figure 3-22: Cross-reactivity of PT-mAb 1E2 with other related OPPs shown as standard curves using MBio assay buffer. (A) shows different cross-reactivity and (B) shows no cross-reactivity at 10000 ng/mL, the highest concentration used in the experiment...... 148

Figure 3-23: Evaluation of strawberry matrix effects with CPF calibration curves in buffer, before and after extraction at 1:8 (dilution factor of 40), with CPF.pAb at 1/100...... 153 Figure 3-24: Evaluation of strawberry matrix effects by developing CPF calibration curves in buffer before and after extraction at a 1:4 dilution (dilution factor of 20), with CPF.pAb at 1/100...... 155 Figure 3-25: Evaluation of strawberry matrix effects by comparing three calibration curves (buffer, extracted curves, spiked before and after) at a 1:4 extract dilution...... 156 Figure 3-26: Evaluation of strawberry matrix effects by comparing extracted calibration curves (spiked before and after) at the extract dilution of 1:8 with the calibration curve in a MBio assay buffer. The three curves represent the concentration ranges as ng/g (A), ng/mL (B)...... 157 Figure 3-27: Evaluation of two volumes of a mBio assay buffer (200 µL and 500 µL) used to reconstitute dried strawberry extract and their comparison with buffer. 50 ng/g and an equivalent concentration were used to spike the strawberry homogenate before and after extraction...... 159 Figure 3-28: Three curves indicating the results of spiked (tissue and extract) and MBio assay buffers in a 10-minute assay. The curves represent the concentration ranges as ng/g (A) and ng/mL (B) after evaporation of 1 mL and reconstitution in 500 µL of a MBio assay buffer...... 161 Figure 3-29: Four populations of strawberry samples (n=20), fortified at different CPF concentrations...... 162 Figure 3-30: Four populations of strawberry samples (n=10) fortified at different PT concentrations...... 166 Figure 4-1: Theory and principles of the planar waveguide approach used in the MBio biosensor (Lochhead et al., 2011)...... 174 Figure 4-2: Representative array layout showing the arrangement of different spots (volume;25 nL), controls (fluorescently-labeled BSA-647), and H11-OVA and CPF-OVA (coating antigens)...... 176 Figure 4-3: CPF-pAb against both CPF-OVA and H11-OVA using CPF standards in MBio assay buffer...... 180 Figure 4-4: PT-mAb 1E2 against both H11-OVA and CPF-OVA using PT standard in MBio assay buffer. Fluorescent responses (A) and normalised data (B). . 181 Figure 4-5: Multiplexed calibration curves for both Abs with CPF-OVA and H11- OVA. Fluorescent responses against CPF (A) and normalised responses data (B)...... 183

xvi

Figure 4-6: Multiplexed calibration curves developed to detect CPF and P using both Abs T. Fluorescent responses against PT (A) and normalised responses (B).184 Figure 4-7: Multiplexed calibration curves for CPF and PT in a real sample. CPF concentrations (A) and PT concentrations (B)...... 186

xvii

LIST OF TABLES

Table 1-1: The main OPPs subclasses and their chemical structures (Chambers and Levi, 1992)...... 8

Table 1-2: Selected analytical techniques for pesticide analysis ...... 14

Table 1-3: Different bio-recognition elements for pesticide detections ...... 24

Table 1-4: Enzyme-based methods used in pesticide sensing techniques ...... 30 Table 1-5: Platforms developed to detect OPPs using whole cells as recognition elements ...... 31 Table 1-6: Platforms developed and coupled with MIPs as recognition elements for OPP detection ...... 33 Table 1-7: Various nanomaterials and composites were used in biosensor applications to detect OPP residues...... 37

Table 1-8: Chemical structures of OPPs used as templates for MIP production ... 44 Table 1-9: General advantages and disadvantages of non-covalent and covalent imprinting (Mirsky and Yatsimirsky, 2010) ...... 46

Table 1-10: MIP templates and their applications ...... 50

Table 1-11: Templates and polymerisation reagents that used to develop MIPs. .. 52

Table 1-12: Common solvents that used for imprinting in non-covalent approach54 Table 1-13: Physical and chemical properties of CPF, Modified from (NLM, 2017a) ...... 61 Table 1-14: Physical and chemical properties of PT, Modified from (NLM, 2017b) ...... 66

Table 2-1: Different compositions of synthesised MIPs ...... 79 Table 2-2: Different precipitation polymerisation conditions reported in the literature...... 86 Table 2-3: Percentages of retained analytes (CIPC) after loading in four different organic solvents at (100 ng/ml) in triplicate ...... 89 Table 2-4: Percentages of retained analytes (CPF) after loading in four different organic solutions at (100 ng/ml) in triplicate...... 90

Table 3-1: ACQUITY UPLC mobile phase gradient ...... 111

Table 3-2: OD responses of checkerboard ELISA titration ...... 135 Table 3-3: Average OD responses at different CPF-pAb dilutions against CPF (0- 1000) ...... 135 xviii

Table 3-4: OD responses for checkerboard ELISA titration ...... 137 Table 3-5: Average OD responses at different PT-mAb dilutions against PT (0-1000) ...... 137 Table 3-6: Cross reactivity profile of CPF-related compounds for CPF-pAb in a MBio assay buffer using a MBio reader ...... 147 Table 3-7: Cross-reactivity profiles of PT-related compounds with PT-mAb 1E2 in MBio assay buffer using an MBio reader...... 149 Table 3-8: Evaluation of two extraction procedures of CPF from strawberry sample spiked before and after extraction for 1:8 extract dilution at 5 ng/mL. Average response (n=6)...... 151 Table 3-9: Evaluation of two extraction procedures of CPF from spiked strawberry samples before and after extraction for 1:4 extract dilution at 10ng/mL. Average response (n=6)...... 151 Table 3-10: Evaluation of two extract dilutions at 1:8 and 1:4 of spiked strawberry samples before and after extraction; 5 ng/mL and 50 ng/mL or equivalent were used for fortification. Average response (n=6) ...... 152 Table 3-11: The percentage of the florescence responses generated from both extracted curves relative to the MBio assay buffer...... 154 Table 3-12: Percentage of florescent responses generated from both curves in relation to a MBio assay buffer...... 160 Table 3-13: Measured concertation, standard deviation, coefficients of variation and recovery values for samples spiked with CPF at three levels. Values were averaged (N=20) ...... 163 Table 3-14: Measured concertation, standard deviation, coefficients of variation and recoveries for samples spiked with PT at three levels. Values were averaged (n=10) ...... 166 Table 4-1: Comparison of different responses generated with Ab mixed in buffer equivalent to the zero standard in buffer ...... 185 Table 5-1 Summary of advantages and disadvantages of MIPs for analytical applications (Mahony et al., 2005)...... 190

xix

LIST OF ABBREVIATIONS

3-CA 3-Chloroanline 4-CA 4-Chloroaniline AA Acrylic acid Ab Antibody ABIN Azobisisobutyronitrile AChE Acetylcholinesterase ACN, Acetonitrile Ag Antigen ATCh Acetylthiocholine BChE Butyrylcholinesterase BSA Bovine serum albumin,

CHCl3 Chloroform CIPC Chlorpropham CL Chemiluminescence COOH Carboxyl group CPF Chlorpyrifos CR Cross-reactivity CV Cyclic voltammetry CV Coefficient of variation

DDH2O Double-distilled water EC European Commission EFSA European Food Safety Authority EGDMA Ethylene glycol dimethacrylate ELISA Enzyme-linked immunosorbent assays EU European Union FAO Food and Agriculture Organisation of the United Nations FM Functional monomer G Gram GaR Goat Anti-Rabbit GaM Goat Anti-Mouse GC Gas chromatography GCE Glassy carbon electrode GC–MS/MS Gas chromatography–tandem mass spectrometry GHS Classification and labelling of chemicals GCE Glassy carbon electrode

H2O Water HPLC High performance liquid chromatography HRP Horseradish peroxidase IARC International Agency for Research on Cancer

IC20 20% inhibition concentration

IC50 50% inhibition concentration

IC80 80% inhibition concentration Kg Kilogram L Litre LC Liquid chromatography LC/MC Liquid chromatography/mass spectrometry LF Lateral-flow LLE Liquid-liquid extraction LOD Limit of detection LOQ Limit of quantification MAA Methacrylic acid Mg Milligram µg Microgram MIPs Molecularly imprinted polymers MISPE Molecularly imprinted polymers solid phase extraction µL Microliter µm Micrometer MMA Methyl methacrylate MMIPs Magnetic molecularly imprinted polymers PM Parathion-methyl MRL Maximum residue limits MRMs Multi-residue based methods MS Mass spectrometry MS/MS Tandem mass spectrometry MWCNTs Multi-walled carbon nanotubes

N2 Nitrogen gas NaCl Sodium Chloride NAFTA The North American Free Trade Agreement NIPs Non-imprinted polymers Nl Nanolitre nM Nanomolar NMOF Nanometal organic framework film NPIC National Pesticide Information Center NRDC Natural Resources Defense Council OECD The Organisation for Economic Co-operation and Development OP Organophosphate OPP Organophosphate pesticide OPAA Organophosphorus acid hydrolase OPH Organophosphorus hydrolase OVA Ovalbumin

xxi

PANNA Pesticide Action Network of North America PBS Phosphate buffered saline PH Parathion hydrolase PIC Prior Informed Consent Pg Picogram PNP P-nitrophenol PPM Part per million PPT Parts per trillion PT Parathion PT-OVA Parathion conjugate QDs Quantum dots RASFF Rapid Alert System for Food and Feed RUP Restricted Use Pesticide SAM A self-assembled monolayer SD Standard deviation SERS Surface-enhanced Raman spectroscopy SPE Solid phase extraction SPR Surface plasmon resonance SWV SWV square-wave voltammetry TCh Thiocoline TIR Total internal reflection TQ-MS Triple quadrupole mass spectrometer TRIM Triethylene glycol dimethacrylate TWG Technical Working Group UHPLC-MS/MS Ultra-high performance liquid chromatography tandem mass spectrometry USEPA United States Environmental Protection Agency UV-Vis Ultraviolet-visible spectrophotometry, VX Nerve agent v/v Volume per volume WHO The World Health Organization WP WP wetable powder XL XL cross-linker

xxii

Chapter 1

Chapter 1: Literature review

Chapter 1

1.1 Abstract

Currently, a large number of various types of pesticides have been developed for use in the agricultural industry to manage pests and weeds. Some of these pesticides disperse or degrade quickly in ecosystems. However, in other cases, humans may become directly exposed to traces of pesticide residues due to crop application, by consumption of food or through the food chain in animal and animal feeds, all of which can pose risks to human health (Gilbert-López et al., 2009). As a result, the use of certain pesticides is banned in food production and other maximum residue limits (MRLs) have been updated. Therefore, feed and food products are regularly monitored for pesticides to avoid any possible threats to animal or human health.

Owing to various concerns regarding food security and trade barriers between countries, more sensitive and reliable methods of analysis to detect pesticide residues are important (LeDoux, 2011). Furthermore, the pesticide regulation, enforcement, and remediation processes are major parts of monitoring the environment for the presence of compounds harmful to public health and ecosystems (Kim et al., 2017). Generally speaking, pesticides are environmental contaminants commonly used throughout the globe, and millions of tonnes per year are applied to improve agricultural crop yields (Donkor et al., 2016).

Organophosphate pesticides (OPPs) are a large class of low molecular weight compounds with various chemical and physical properties. Various OPPs, such as insecticides, nematicides, fungicides and herbicides, are commonly used in the agriculture and aquaculture sector to enhance crop production. OPPs are designed primarily to kill pests but they can also cause severe neurotoxic effects in mammals and wildlife with their residues remaining after pesticide applications for extended periods in agricultural fields, aquaculture ponds and ecosystems. As these pesticides are generally regarded as highly toxic, there are increasing concerns about the risks associated with these pesticides for ecosystems, animal and human health due to their overuse, misuse and frequent applications in food production settings.

Recently, the uncertainty associated with these risks has driven the development to either improve analytical methods or provide alternative solutions to measure OPPs

Chapter 1

contamination in feed and food samples. For example, the technology of molecular imprinted polymers is a process to create artificial polymeric receptors with a predetermined selectivity and specificity for a given analyte. Due to the reported robust polymeric matrices mimicking that of natural receptors or antibodies with high selectivity, MIPs have been explored in particular as solid phase extraction sorbents for sample preparation and biosensor recognition elements for OPP and other targets. However, while some studies have demonstrated the efficiency of using MIPs for various targets, particularly in separation processes, this approach still suffers from challenges such as slow mass transfer, template leaching and non-specific binding during the analysis process.

Traditionally high performance liquid chromatography (HPLC) and gas chromatography (GC) coupled to mass spectrometry have been used to detect and analyse OPP residues. However, these methods have their limitations, including complexity and expense, time requirements for analysis and restrictions to use in the laboratory. Therefore, there is an increasing importance to monitor these potentially neurotoxic OPPs with a high degree of sensitivity and selectivity at sites of contamination. For simplicity and portability of biosensor methods, different biosensing techniques have been investigated for the analysis of environmental and agriculture samples for the presence of OPPs. To date, researchers have coupled various recognition elements such as enzyme- and antibody- based applications and others, with various detection methods, such as electrochemical, optical, and piezoelectric approaches, in addition to advancement in nanomaterials to construct a broad range of biosensors. No individual approach meets all the requirements for the end user in real life scenarios. Therefore, a combination of approaches should be considered.

Chapter 1

1.2 An Introduction to Pesticides

Pesticide is defined by Food and Agriculture Organization of the United Nations (FAO) as “any substance intended for preventing, destroying, attracting, repelling, or controlling any pest including unwanted species of plants or animals during the production, storage, transport, distribution and processing of food, agricultural commodities, or animal feeds or which may be administered to animals for the control of ectoparasites. The term includes substances intended for use as a plant growth regulator, defoliant, desiccant, fruit thinning agent, or sprouting inhibitor and substances applied to crops either before or after harvest to protect the commodity from deterioration during storage and transport” (FAO, 2001). Different pesticide classifications exist, as they are classified based on their intended purpose, pesticide mode of action, targeted pest species and chemical structures. In addition, a different classification within the class exists, such as the spectrum of activity, mode of formulation and toxicity or risk potential (Zacharia, 2011). Pesticide use has become an integral aspect of agricultural, aquaculture and environmental practices, and are increasingly applied in ever-greater quantities for agricultural purposes. Over the last decades, the widespread use and benefits of pesticides for the commercial agriculture sector have markedly enhanced its productivity and prevented losses by eliminating detrimental pests. However, excessive and improper application of pesticides in agricultural fields, aquaculture and ecosystems has increased the levels of pesticide contamination in food supplies, which has necessitated the detection and monitoring of these pesticides in a broad range of feed and food products (Yan et al., 2014). In addition, while pesticides increase productivity, they also pose many challenges, such as interference with and disruption of biological systems in many living organisms.

In fact, many pesticides have been banned in several countries due to their detrimental effects on humans, other mammals, insects and ecosystems in general. Of particular concern are OPPs; a large chemical class of compounds mainly used as insecticides, nematicides, fungicides and herbicides. Many of these pesticides are also considered neurotoxic for higher living organisms and harmful to ecosystems due to their ability to interfere and inhibit cholinesterase enzymes in humans and animals (Ma and Chen, 2014). In fact, some OPPs, such as chlorpyrifos, have even been banned from indoor use (Kitada

Chapter 1

et al., 2008) and officials in many countries have regulated and continue to review the application of other OPP pesticides (Valdés-Ramírez et al., 2009).

1.3 Chemical classification of pesticides

Chemical classifications are important for researchers who study pesticides and the environment. Researchers generally classify pesticides into four categories: organochlorines, organophosphates, carbamates and, pyrethrins and pyrethroids (Salghi et al., 2011).

1.3.1 Organochlorines

The first chemical class of pesticide produced was Organochlorine compounds whose structures contain carbon, hydrogen and chlorine. They are synthetic organic materials and include pesticides such as dichlorodiphenyltrichloroethane (DDT), cyclodienes (aldrin, and dieldrin) and hexachlorocyclohexane (lindane). These pesticides contaminated the environment, and were found to be in different locations because of pesticide drift and prolonged persistence after the introduction of the pesticide through application in agriculture (Abdollahi et al., 2004). Under national and international regulations, most or all organochlorine pesticides were banned for agricultural use due to their adverse health effects to human, animals and their long-term persistence in the environment (VoPham et al., 2017; Pedro et al., 2017). Although this class was banned, some of its pesticides are still being used and detected in food, in developing countries (Jin et al., 2017).

1.3.2 Carbamates

Carbamate pesticides are synthetic and organic chemicals originated from carbamic acid, and they are grouped mainly as N-methylcarbamate insecticides and N-allylcarbamate herbicides, or fungicide (Nantia et al., 2017; Ameno, 2005). Such pesticides include carbaryl, methiocarb and thiodicarb, and they target insects through inhibition of acetylcholinesterase enzyme and cause acetylcholine to accumulate. As a result, toxicity symptoms are seen after exposure that could happen by any route though mainly through inhalation or ingestion (Fishel, 2005).

Chapter 1

1.3.3 Pyrethrins and pyrethroids

This class is another agriculture pesticide group and initially originated from the natural compound before organic synthetic insecticides were produced to improve stability (Shafer et al., 2005). They are commonly utilised in agriculture for crops protection and commodity storage. They are neurological active substances and kill target pests by interfering with their sodium channels which are an important component of membrane proteins, leading to paralysis and death. They have a lower toxicity to mammals compared to other groups and are less persistent in the environment. However, they bioaccumulate in the fatty tissue of animals which may cause possible health implications to humans on consumption of meat products of affected animals (LeDoux, 2011).

1.3.4 Organophosphates

OPPs were developed to replace the organochlorines and are used primarily to control insects due to their insecticidal properties that were discovered in substances related to sarin, soman and tabun, also known as nerve gases. Compared to the class of organochlorine insecticides, OPPs exhibit higher toxicity to both insects and vertebrates but degrade quicker in the environment. These two features, in higher control and less persistence, have given OPPs a competitive edge in the agricultural-based practice. Since the development of the OPP class in the last century, over 1500 OP compounds with specific features have been produced and applied as pesticides in ecosystems with various formulations. However, this success has encouraged OPP use in household and domestic gardens, as well as in the environment, which has led to considerable growth in their applications. Furthermore, the toxicity mode of action of organophosphates is similar to carbamate and relies on disrupting the acetylcholinesterase that regulates acetylcholine, in the nervous system, by deactivation (Jamal et al., 2002). Therefore, these compounds can be highly potent chemicals to humans. For example, parathion-methyl (PM) has been deemed one of the most highly toxic OPPs due to the p-nitrophenyl functional group present in its chemical structure, which degrades to yield methyl paraoxon that is even more toxic (Fang et al., 2015). As a result, many regulatory bodies worldwide have regulated OPP residues.

Chapter 1

Chemical structures and classification of OPPs

OPPs are artificial and organic compounds, and include esters, amides and derivatives of thiol, such as phosphoric, phosphonic, phosphorothioic, or phosphonothioic acids. The general chemical structure of organophosphates includes a central phosphorus atom with a double bond attached to either a sulphur or an oxygen atom, as shown in Figure 1-1. Only P=O compounds, known as oxons, are the metabolites of the oxidative desulfuration of parent P=S esters, known as thions. The X group is attached to a phosphor atom and varies in chemical structure. The R groups are either directly or indirectly attached to the phosphorus atom via an oxygen or nitrogen atom, and can be chemically different.

R1 R2 R1 R2

X X

Oxons Thions

Figure 1-1: General chemical structure of OPPs

The OPPs have twelve subclasses with names based on their chemical structures.

Different groups such as alkoxy, aryloxy and thioalkoxy commonly represent R1 and R2 while X represents a leaving group. For example, if the R group is connected directly to a phosphorus atom, they are called phosphinates; to an oxygen atom, phosphates; and to a nitrogen atom, phosphoramidates. Also, if the R1 group is attached to a phosphorus atom and the R2 group to an oxygen atom, the result is designated a phosphonate (Samuels and Obare, 2011). Table 1-1 shows OPP subclasses and their different chemical structures.

Chapter 1

Table 1-1: The main OPPs subclasses and their chemical structures (Chambers and Levi, 1992). Phosphorus group Structure Common name Phosphate chlorfenvinphos, crotoxyphos, dichlorvos dicrotophos, heptenphos, mevinphos, monocrotophos, naled, phosphamidon, TEPP, tetrachlor vinphos, triazophos

O -alkyl amiton, demeton-S-methyl, omethoate, phosphorothioate oxydemeton-methyl, phoxim, vamidothion

azothoate, bromophos, bromophos-ethyl, chlorpyriphos, chlorpyriphos-methyl, coumaphos, diazinon, dichlofenthion, fenchlorphos, fenitrothion, fenthion, parathion, parathion methyl, pyrazophos, pyrimiphos-ethyl, pyrimiphos-methyl, sulfotep, temephos, thionazin Phosphorodithioate amidithion, azinophos-ethyl, azinophos-methyl, dimethoate, dioxathion, disulfoton, ethion, formothion, malathion, mecarbam, menazon, meth-idathion, morphothion, phenthoate, phorate, phosalone, phosmet, prothoate, thiometon

S- alkyl profenofos, trifenofos phosphorothioate

S- alkyl prothiofos, sulprofos phosphorodithioate

Phosphoramidate cruformate, fenamiphos, fosthietan

Phosphorotriamidate triamiphos

Phosphorothioamidate methamidophos

Chapter 1

Phosphorus group Structure Common name isofenphos

Phosphonate butonate, trichlorfon

Phosphorothionate EPN, trichlornat, leptophos, cyanofenphos

1.4 Pesticide use in food production

1.4.1 Pesticide use in agriculture

The application of pesticides is an integral part of agricultural productivity, as their use helps to control diseases and avoid food losses, both in the field and during storage periods. As stated above, one system of classification is that based on target pest species or their intended use, and these include herbicides, insecticides, fungicides, rodenticides, avicides, nematocides and so on (Salas et al., 2000). In addition, pesticides are used in different agricultural settings to protect crops, forests and wetlands during pre- or post- harvest phases (Cortina-Puig et al., 2010). As organic compounds, pesticides are biologically active, and thus inevitably either directly or indirectly find their way into mammalian tissues via the food chain or through drinking water (Kim et al., 2017). Some pesticides have been determined to be toxic to human health due to their ability to interfere with metabolic activity (Cortina-Puig et al., 2010). Recently, environmental and food safety concerns have been raised that are associated with the presence of pesticide residues as a result of the misuse of pesticide applications. For example, improper pesticide selection, excessive use of pesticides in agricultural applications, or that pre- harvest interval periods are not followed after application, all of which can allow a great amount of residues into food designated for human consumption (Yan et al., 2014; Chen

Chapter 1

et al., 2012). As a result, regulations have been enacted across the globe, and specific limits on allowable pesticide levels present in water and crops have been adopted. Also, for some pesticides, monitoring and reviewing methods have been established to ensure human, animal and environmental safety and reduce possible risks (Handford et al., 2015; Cortina-Puig et al., 2010). However, more robust pesticide detection systems are required to support monitoring of pesticide residues levels in ecosystems, foodstuffs and soils (Yan et al. 2014).

1.4.2 Pesticide use in aquaculture

Aquaculture is an important sector and defined by EU as “the rearing or cultivation of aquatic organisms using techniques designed to increase the production of those organisms beyond the natural capacity of the environment” (Council directive, 2006). This includes the farming of fish, shellfish and others under controlled systems in the environment of either seawater or fresh water. For example, seawater aquaculture is in cages or in the water column, with organisms such as oysters, clams, mussels, shrimp and salmon. Whereas freshwater aquaculture occurs in ponds or recirculating tanks, with organisms such as catfish, trout, tilapia and bass (NOAA Fisheries, 2017). The production quantity and quality of the aquaculture production is continuously increasing. The amount of fish aquaculture produced internationally soared from around 13 to 67 million tons from 1990 to 2012 (FAO, 2014). Pesticide contamination of aquaculture produce can occur in two ways: either through the intentional use to kill unwanted organisms or unintentionally through the use of contaminated ingredients of the provided feeds which include residues of pesticides and veterinary drug, persistent organic pollutants and others (Tacon and Metian, 2008). Additionally, marine environments are placed for receiving agricultural waste leading to aquatic pollution (Kaoud, 2015). Accumulation may occur in target bodies used in food production, which in turn may cause a critical environmental concern for the end consumer. Pesticides are used to control or reduce the increase in growth of aquatic organisms and include organophosphate, carbamate and pyrethroid pesticides. For example, carbaryl pesticide is used to kill ghost shrimp and mud shrimp (Weston, 2000), trichlorfon is used to treat parasites in baths (Guimarães et al., 2007) and cypermethrin is used to treat sea lice in baths (Willis et al., 2004).

Chapter 1

1.5 Health implications

The implications of pesticide use on human health are considered as potentially serious public health issues around the world. Our body receives pesticides directly by exposure, from the environment or as residues from the food (Tegtmeier and Duffy, 2004). The adverse health effects that are triggered by pesticides can happen in two ways either in a few hours or days as acute toxicity or in the long term may take years as chronic toxicity, which is more serious and can cause severe damage (Devi, 2009). Although all necessary measurements have been taken to regulate and restrict pesticide use by individuals and workers, there are several illnesses that have been linked to pesticides from the chronic exposure. It was reported that a number of neoplasms include breast cancer, prostate cancer, lung cancer, brain cancer, colorectal cancer, testicular cancer and pancreatic cancer (Mostafalou and Abdollahi, 2013). Another health problem is related to birth defects and developmental toxicity where pesticide exposure was connected to antepartum, postpartum and congenital disorders (Weselak et al., 2007). Pesticide exposure was linked to reproductive disorders in both males and females, with a number of pesticides that interfere with endocrine functionality intoxicating the reproductive system (Saadi and Abdollahi, 2012; Tiemann, 2008). Despite the fact that the cause of Parkinson disease is not well-understood yet, there has been a strong correlation between a number of reported cases and individuals who were exposed to pesticides (Van et al., 2012). Another study by Hayden found an important correlation between workers exposed to OPPs and symptoms of Alzheimer’s disease at a later stage in life (Hayden et al., 2010). In the literature, there are other health effects reported such as Amyotrophic lateral sclerosis, diabetes, cardiovascular disease, chronic nephropathies and chronic respiratory disease (Mostafalou and Abdollahi, 2013).

1.5 Legislation governing pesticide use and harmonisation

The harmonization of pesticide regulations and standards has been proposed and in some cases made available by international organisations to reduce the risk and trade barriers globally. The process underwent several stages, for example, the Code of Conduct on the Distribution and Use of Pesticides was adopted in 1985, and reviewed in 2014 by the FAO. This code contains voluntary standards which are highly important when a country

Chapter 1

has inappropriate or limited domestic pesticide regulations (Handford et al., 2015). A number of treaties and international documents are available for any nation to embrace for pesticide management such as the Basel Convention on the Control of Transboundary Movements of Hazardous Wastes and Their Disposals, the Rotterdam Convention on the Prior Informed Consent (PIC) Procedure for Certain Hazardous Chemicals and Pesticides in International Trade, and the Stockholm Convention on Persistent Organic Pollutants.

The World Trade Organisation (WTO) provides trade regulations of goods movements between its members. The WTO issues different agreements such the Sanitary and Phytosanitary Agreement, which encourage acquiring international standards, guidelines or recommendations such as Codex standards on pesticide residues. Codex provides non- compulsory MRLs. However, it is valuable especially for nations where a national pesticide system is weak or not developed (WTO, 2017). The Organisation for Economic Co-operation and Development (OECD) has 30 countries that is responsible for different tasks and allows for greater harmonization between the members. The OECD pesticide program enables the sharing of the pesticide registration system and the cooperation for risk assessment. They work to harmonize the methods for testing to promote consistency between governments and to minimize the risk associated with registration, handling and pesticide use (OECD, 2017). The Globally Harmonised System of Classification and Labelling of Chemicals (GHS) was established to help classify chemicals based on the hazard levels, which allows for good hazard communication in terms of labelling and safety data sheets. This system makes necessary information obtainable for international trade, and harmonised guidelines for environmental and human safety (UNECE, 2017). Moreover, the North American Free Trade Agreement (NAFTA) was created between Canada, the US, and Mexico to harmonize the requirement of pesticide regulations. For decreasing the cost and boosting the efficiency of pesticide control, a Technical Working Group (TWG) on Pesticides was developed (USEPA, 2016). In Europe for example maximum acceptable thresholds of 0.1 µg/L and 0.5 µg/L, were introduced as limits for single OPP compounds and collective pesticides respectively for drinking water (Kuster et al., 2009). These thresholds apply to the majority of OPPs, which pose greater challenges for detection.

Chapter 1

1.6 Current methods of pesticide analysis

Chemical contamination can occur in food supplies and ecosystems in trace amounts, and the detection of pesticide residues relies on factors such as sensitive analytical or bioanalytical instrumentation in combination with rapid, simple reliable sample preparation; and clean-up techniques (Kang et al., 2012). Various analytical methods are currently available to detect and analyse OPPs that involve liquid or gas chromatography that are coupled with different detectors such as mass spectrometry (HPLC-MS, GC-MS). Apart from conventional analytical methods, advances in the development of other analytical techniques to facilitate and improve detection sensitivities for OPPs include the use of biosensors. There is ongoing research being conducted in order to reduce the cost of pesticide monitoring processes and to promote more rapid and portable detection capabilities. In fact, a number of studies have been published in the last two decades that have reported advances in producing relatively rapid, simple and reliable biosensor devices that can easily monitor food safety and environmental quality (Verma and Bhardwaj, 2015). Different bioanalytical approaches in the use of immunoassays or biosensors such as electrochemical and optical biosensors have been used for OPP detection (Zhang et al., 2014). To date, researchers have developed many techniques and methods for rapid pesticide detection, and the literature contains studies on pesticides in general and OPPs in particular that show developments in the analysis of food and environmental samples (Liu et al., 2008; Mulchandani, 2011; Van and Pletschke, 2011). However, many of these recent approaches have not yet been deemed viable in the market in order to justify their widespread use.

1.6.1 Analytical Methods of Analysis

Monitoring of pesticides residues in food and feed to not exceed MRL is mandatory by law in order to provide safe food and protect human health. Pesticide analysis by conventional analytical techniques for identification and quantification involve HPLC, GC, or coupled with mass spectrometry due to its efficiency regarding the selectivity, sensitivity and high capacity (Villaverde et al., 2016; Cardeal et al., 2011). Table 1-2 represents different analytical techniques applied for pesticide analysis. The majority of the current analytical techniques available for OPP detection are laboratory-based

Chapter 1

illustrating pesticide detection in specific commodities with varying limits of quantification (LOQs), limits of detection (LODs) and times for analysis.

Table 1-2: Selected analytical techniques for pesticide analysis. Limit of Limit of Total Parameter Matrix Quantification Detection Run Reference (LOQ) (LOD) Time Esteve-Turrillas et al. ELISA Strawberry - 20 - 30 ng/L 2 days (2015) Immunoassay Red and ELISA 0.01 ng/mL 0.005 ng/mL 2 days Τsialla et al. (2015) white wine Fab Vegtable - - - Rao et al. (2016) Fluorescent PBS bufer - 1.0 ng/mL 15 m Zou et al. (2010) sensor Milk Electrochemical - 15 ng/L 15 m Campanella et al. (2011) vegtalbe Optical MILK - 5.0 × 10−8 M - Tomassetti et al. (2015) Immunosensor Optical wine - 0.6 ng/mL 10 m Koukouvinos et al. (2017) Electrochemical potato - 1.4 μg/kg 50 m Valera et al. (2014) Electrochemical Vegetable - 0.014 ng/mL - Cao et al. (2015) Electrochemical water - 0.2 ng/L - Tran et al. (2013) Electrochemical Rice - 0.1 ng/mL - Deep et al. (2015) Electrochemical Vegetable - 52 pg/L - Mehta et al. (2016) 0.0017-0.2667 GC-MS grapes - 35 m Lian et al., ( 2010) Gas mg/kg chromatography GC-EI-HRMS Tomato ≤1 μg/kg ≤ 1 μg/kg 10 m Mol et al. (2016) GC-MS-MS Sea food 0.037-0.153 ng/g 0.011-0.046 ng/g 22 Li et al. (2017a) HPLC-MS Olive oil - 0.001-0.500 mg/L 32 Barrek et al. (2003) Cappage 0.001- HPLC-MS/MS 0.01 mg/kg 13 m Mol et al. (2003) and grap 0.004 mg/kg. Liver and Hamamoto et al., LC-MC/MS 0.9 to 15.2 ng/g 0.3 to 4 ng/g 31 m musel (2017) red and LC-ESI- white 1 μg/L - - Berset et al. (2017) MS/MS wines LC-ESI- Tea 0.01 -0.05 mg/kg - - Jiao et al. (2016) MS/MS LC-ESI- produce 0.01 mg/kg - 33 m Mol, et al. (2007) MS/MS LC-ESI- 2 μg/kg 0.7 μg/kg 15 m Zheng et al. (2017) MS/MS Liquid LC-ESI- Water and 1 ng/L and - 36 Pastor-Belda et al. (2017) chromatography MS/MS fruit 5 ng/kg 0.004-0.012 LC-MS/MS honeybee 0.028-0.094 7 m Amulen et al. (2017) µg/kg UPLC-MS/MS Water 0.025 µg/L - 10 m Marín et al. (2009) UHPLC- Fruit and - ˂ 50 μg/kg 20 m Wang et al. (2016a) QTOF–MS vegetable UHPLC/TOF- vegetable 0.8 to 11.8 μg/kg 0.3 to 3.8 μg/kg 5 m Sivaperumal et al. (2015) MS and fruit UPLC-TOF- Nurmi and Pellinen, Wastewater 28-49 ng/L 15-26 ng/L 16 m MS (2011). 43 and 113.2 HPLC-HRMS urine - - Cortéjade et al. (2016) ng/mL UHPLC- Whole ˂3.57 µg/L ˂9.97 µg/L 8 m Bordin et al. (2017). MS/MS wheat

1.6.1.1 High Performance Liquid Chromatography (HPLC)

HPLC is one of the popular chromatographic techniques used for pesticides analysis and has been applied for relatively polar and non-volatile substances. Generally speaking, it separates chemically different analytes present in mixtures between two mobile phases and stationary phases. Normally, a small volume of a sample extract is injected into a

Chapter 1

chromatographic column and the sample is moved by the flow of mobile phase and each analyte present in the sample retains to the stationary material slightly at different times, a different speed, causing the separation to occur. The analyte nature and the conditions of the two phases will determine the degree of retention time for each analyte. Ultraviolet– visible (UV/VIS) detectors such as diode array (DAD) or photodiode array (PDA) are the most widely used for pesticide analysis using HPLC (Tuzimski and Sherma, 2015). However, HPLC applications are perhaps suitable for pesticide analysis for a sample that is not complicated and owns insignificant matrix effects otherwise the sample might need "Quick, Easy, Cheap, Effective, Rugged, and Safe" (QuEChER) extraction methods associated with a clean-up step to achieve the required sensitivity (Rejczak and Tuzimski, 2017). Moreover, HPLC is still used to quantify pesticides even with more advanced techniques being introduced. For example, multi-residues including fourteen fungicides, were identified in grape samples (Otero et al., 2003), three different insecticides (Al- Rimawi, 2014) and seven different herbicides and the main metabolites related to s- triazine (Teju et al., 2017) were identified in water samples.

1.6.1.2 Liquid chromatography/mass spectrometry (LC/MS)

LC-MS is a commonly used technique for the analysis of a large number of trace levels of pesticide residues. LC-MS has two components, LC for separation and MS for analytical detection. It is increasingly used to detect, identify, and quantify multi-pesticide classes for the reason that it requires minimal work for sample preparation in a complex matrix sample and provides the level of required sensitivity, selectivity and specificity for regulatory bodies (Farré et al., 2014; De et al., 2014; Hu et al., 2014). LC-MS is the most advanced analytical technique used for more polar pesticides that are thermolabile and not volatile. The system provides details about the structure of the analysed analyte. It also does not necessitate purifying the sample, and it allows for a simultaneous detection of analytes that have a significantly wide range of polarities. Furthermore, in a classical LC-MS system, initially, a small volume of the sample extract passes through a chromatographic column, LC system, to separate different analytes. Once the sample fractions are eluted from the column in different times, it then undergoes the ionization stage before being subjected to a mass analyser and detection. LC-MS has become a standard method for pesticide detection in many laboratories (Fan et al., 2014; Masiá et

Chapter 1

al., 2014). Different pesticides in edible samples have been detected by LC-MS, for instance, four fungicides and its major metabolites (Yoshioka et al., 2004) were identified in citrus fruits, and propamocarb (Hiemstra and de Kok, 2002) in vegetable samples.

1.6.1.3 Gas chromatography/mass spectrometry (GC/MS)

GC/MS is another analytical technique that comprises of two sections, GC and MS and working by the same principle as LC/MS. However, it is frequently used for pesticide analysis where pesticides are known to be volatile or semi-volatile and thermally stable. The system enables pesticides to be separated, identified and quantified with high sensitivity in one analytical run. A small volume of a sample extract is introduced into GC where it vaporizes under high temperature and moves along a chromatographic column through a gas carrier as a mobile phase. The analytical separation provided by GC uses the analytes chemical characteristics in the sample and affinity to the stationary phase. At a different time, the analytes leave the column and are subjected to a mass spectrometer where substances are ionised to fragments and the chemical structures identified based on their masses which have a specific spectrum for determination. Furthermore, several methods were developed to quantify pesticides in several matrices, for example, a few acaricides (Rial-Otero et al., 2007) and eleven organochlorine pesticide residues (Tahboub et al., 2006) were detected in honey, twenty-three insecticides and fungicides in vegetable samples (González-Rodríguez et al., 2008) and triazines (Sanchez-Ortega et al., 2009) in underground water samples.

In addition to basic HPLC, LC-MS and GC-MS, advanced techniques such MS/MS for pesticides residues were introduced and applied for quantification and confirmation especially for multi-class pesticide residues at ultra-trace levels in vegetables and fruits. These techniques include gas chromatography–tandem mass spectrometry (GC– MS/MS), liquid chromatography–tandem mass spectrometry (LC–MS/MS) (Sharma, et al., 2010) and ultra-high performance liquid chromatography tandem mass spectrometry (UHPLC-MS/MS) which are able to provide superior capabilities. The sample preparation for the systems needs optimisation to cover a wide range of pesticides with high selectivity and sensitivity. Two mass analysers (MS/MS) are involved to produce chemical structures of fragmented ions in two steps. First, initial ion fragments are

Chapter 1

produced by first MS, and then the selected ions of a specific mass-to-charge ratio are fragmented again before the new fragments are separated and identified where then the detector can use the new fragments to determine information about their chemical structure.

For multi-pesticide residue methods, liquid chromatography–tandem quadrupole mass spectrometry (LC–MS/MS) with electrospray ionization was used to detect a hundred and seventy-one pesticides in many samples of plant origin (Hiemstra and De 2007), two hundred and ninety-nine pesticides in groundwater samples (Sandstrom et al., 2016). Liquid chromatography–time-of-flight mass spectrometry (LC–TOF-MS) was used to quantify fifteen pesticides in fruit and vegetable samples (Ferrer et al., 2005) and in water to identify a hundred and one pesticides and their degradation (Ferrer and Thurman, 2007). Ultra-high performance liquid chromatography and tandem mass spectrometry (UPLC-MS/MS) was used to detect different pesticide residues in fruits (Shi et al., 2016). In addition, gas chromatography–tandem quadrupole mass spectrometry (GC-MS/MS) was also used to detect many pesticide residues in honey samples (García et al., 2017; Saitta et al., 2017) and to identify two hundred and thirteen pesticides residues in several samples of vegetable oils (Vázquez et al., 2016).

Although these methods are reliable, offer the use of multi-residue based methods (MRMs) and can identify several pesticide residues in a single analysis, they are also generally tedious, laborious and expensive. In addition, these methods require well- equipped laboratories and trained technicians. Additionally, the required sample pre- treatments for these techniques is often deemed to be complicated. Therefore, they are not appropriate for field-testing and in situ monitoring of samples (Suri et al. 2009; Kaltsonoudis et al., 2003).

1.6.2 Bioanalytical Techniques

Bioanalytical chemistry is a subdiscipline of analytical chemistry that involves the separation, detection, identification and quantification of biological samples in different settings. It often involves the study of molecules such as proteins, peptides, DNA, toxins and drugs and more recently nanomaterials. In the context of rapid diagnostics bioanalytical methods have evolved from traditional immunological methods of ELISA

Chapter 1

to lateral flow devices (LF) to biosensors and nanoarray approaches that require three components of biorecognition elements, the antigen or target to be detected and either a labelled or label free detection system. Similary, advancement in biosensor technology have been illustrated with alternative biorecognition elements to antibodies to offer new detection approaches.

1.6.2.1 Immunological assays

Antibodies belong to a class of proteins known as immunoglobulins (Ig) that are generated through animal immunization, hybridoma technology, and/or recombinant techniques. Antibodies are an important part of immunoassay development, and most of their production methods require the immunization of an animal with an antigen. In this way, the immune system can identify a molecule or antigen (i.e. immunogenicity) that can then be associated specifically with an antibody (reactogenicity). Those molecules that feature both immunogenicity and reactogenicity are known as complete antigens, and those that only exhibit reactogenicity are known as incomplete antigens. A hapten is another name for an incomplete antigen, and includes a vast number of small molecules, such as pesticides, herbicides, drugs, explosives, polycyclic aromatic hydrocarbons and metal ions. However, haptens alone cannot make the immune system produce antibodies. Therefore, they need to be conjugated covalently to a larger carrier or molecule, usually a protein. Furthermore, natural antibodies, or immunoglobulins, are classified into different groups (e.g. IgG, IgM, IgA, and IgE) that supply animals with a defence mechanism (Ban and Blake, 2012). Antibodies contain recognition sites that allow for their high specificity and capacity to bind a foreign substance, which is an essential feature in biosensor fabrication (Hennion and Barcelo, 1998). They have been used not only for a rapid detection of analytes in water, body fluids, soil and food extracts but are also used in a sample preparation before detection. Due to affinity and specific characteristics that antibodies exhibit, the analytical competency of an immunochemical method can be selected. Also, depending on the properties of a given antibody, it can be used for a range of techniques. For instance, immunoassay methods are used for applications requiring sensitive detection, while for those analytes requiring clean-up and concentration, immunoaffinity chromatography and flow injection immunoassay systems can be used (Franek and Hruska, 2005).

Chapter 1

Immunoassays rely on a basic principle that involves a chemical reaction between an antigen and an antibody, which is governed by the Law of Mass Action. The high affinity of the antibody for its target (i.e. antigen) leads to a low detection limit and sensitive immunoassay (Hennion and Barcelo 1998). Moreover, immunoassays are rapid, specific, sensitive and relatively easy and inexpensive to conduct. Therefore, they have been developed either to detect individual compounds (Liu et al., 2017a; Dai et al., 2017; Moreno et al., 2001) or groups of similar pesticides (Xu et al., 2010; Alcocer et al., 2000; Xu et al., 2009). In addition, immunoassays that determine a group of pesticides supply either qualitative or semi-quantitative data for the entire group. In contrast to traditional methods, immunoassays have some advantages, they do not require sophisticated instruments or devices, and use only a very small amount of organic solvent. Furthermore, the capability of initial screening for environmental or food samples is another advantage. Lastly, this type of analysis is particularly useful when other methods for pesticide detection are tedious or expensive (Gabaldón et al., 1999).

Immunoassays yields relatively quick qualitative and semi-quantitative results in laboratory settings and its formats require three elements: generic or specific antibodies, a conjugated hapten, and a target analyte, for competition. The competitive assays that are used for small molecules can be divided into two categories; direct and indirect competitive immunoassays. For a direct competitive assay, a labelled-Ab is directly bound to the conjugated antigen. Whereas, in the indirect competitive assay, conjugated- antigen and a target analyte compete to bind the available unlabelled-Ab before a secondary-Ab is used to generate a signal for detection, and this type is mostly preferred for pesticide analysis. Moreover, the binding degree of conjugated antigen to Ab relies on a pesticide amount and this permits the identification of the target concentration indirectly. The concentration then is calculated based on developing a calibration curve from known target concentrations. (Goel, 2013; Fan and He, 2011; Toscano et al., 1998).

For non-competitive immunoassay, various technologies have been reported in the literature. The most commonly used technique to identify antigen concentration is the sandwich assay. However, the antigen has to be large enough to be successfully measured, and this approach uses both unlabelled and labelled antibodies. However, the analyte involved has to have more than one binding site (Hennion and Barcelo, 1998).

Chapter 1

A range of immunological methods exist based on the detectable labels available to measure biological binding. Some examples include radioimmunoassays (RIA), enzyme- linked immunosorbent assays (ELISA), fluorescence immunoassays (FIA), fluorescence and polarization immunoassays (FPIA), and others (Bock et al., 2007).

1.6.2.2 Enzyme-linked immunosorbent assay (ELISA)

ELISA is a well-known immunoassay technique and requires an antibody or antigen to be immobilised onto a surface utilized in a 96-well plate format. In addition, many pesticide detection methods have been developed for pesticide analysis, and substantial work has been carried out to develop relatively simple and rapid alternative methods for pesticide analysis. A competitive ELISA method can be used to identify a pesticide concentration in a sample in question by the Ab-Ag interaction. However, it sometimes encounters drawbacks related to complex matrix preparation, cross-reactivity and incomplete validation (Andreu and Picó, 2012). In this technique, the most widely used labels are alkaline phosphatase, galactosidase and horseradish peroxidase (HRP) (Fan and He, 2011). There are available ELISA methods developed to detect the main pesticides, including azoxystrobin in fruit juice (Parra et al., 2012), picoxystrobin in different food samples (Mercader et al., 2012), ethyl carbamate in Chinese vegetable (Luo et al., 2017) and dufulin residue in water and agricultural samples (Chen et al., 2017a).

1.6.2.3 Lateral-flow (LF)

LF, also known as an immunochromatographic technique, is a straightforward tool able to identify the presence or absence of a target substance, providing a rapid result avoiding expensive equipment. It is comprised of two elements; a separation part which is based on the sample movement and antibody-antigen interaction. It has been applied in different formats with test strips being the most widely implemented immunochromatographic technique (Dzantiev et al., 2014). Once the sample is introduced, it migrates along the porous carrier. Initially, it was used for diagnostic purposes such as the pregnancy test due to being a simple and user-friendly technique; it eases the rapid identification of many individual analytes inexpensively (Qian and Bau, 2003). In the past, the immunochromatographic technique was used to define a different analysis process whereby it separated and concentrated analytes in an immunochromatographic column

Chapter 1

(Weller, 2000). For the LF test strip, locations are determined for antibody or conjugate immobilisation where the binding takes place when the analytes are introduced in the sample. Several applications have been developed for many areas such as agriculture, environment, packaging, natural toxin, allergen and pathogen detection (Dzantiev et al., 2014).

In the field of food safety and quality, the occurrence of contaminants such as pesticides adulterate the commodities for international trade. Therefore, rapid identification techniques for early detection to monitor food contaminants are emerging. LF applications have been used to detect pesticides in food, water and environment using a test strip. It was used to detect pesticides in water samples, for instance, carbofuran and triazophos (Guo et al., 2009) and atrazine (Kaur et al., 2007), in vegetable samples such as organophosphate and carbamate (Han et al., 2012), carbaryl and endosulfan (Zhang et al., 2006), in fruit samples such as of thiabendazole (Goeselova et al., 2013), in beverage samples such as organophosphate pesticides (Hossain et al., 2009) and only buffer targeting endosulfan, cypermethrin, deltamethrin and fenvalerate (Kranthi et al., 2009) and 2, 4-dichlorophenoxyacetic acid (Weetall and Rogers, 2002).

1.6.2.4 Biosensors

In previous decades, the fabrication of biosensors for pesticide analysis received substantial consideration as a potential alternative. A biosensor possesses three integrated parts (Figure 1-2) 1) recognition elements, such as enzymes, DNA probes and antibodies attached to a transducer to identify the analyte; 2) a complementary binder and antigen; and 3) a transducer used to convert a reaction to a quantifiable signal using various sensing techniques. As a result of the reaction between the analyte and its biological recognition element and then convert the result into an electronic signal (Audrey et al., 2012). The extent of the resulting signal is proportional to the target chemical concentration bound to the bio-recognition element.

Chapter 1

Figure 1-2: Schematic presentation of a biosensor

Biosensors are ideal for environmental monitoring and detection, as they can be mobile, selective, sensitive and allow for rapid detection in real time. However, the majority of biosensors can only be used for an individual analyte, which is their main drawback, as they cannot analyze multiple pesticides simultaneously (Raz and Haasnoot, 2011). Immunosensors are a group of biosensors characterised by their highly selective affinity binding between antibodies and antigens on the surface of a transducer relative to competition from their specific target analytes. These immunosensors are specific to particular molecules and could be used to measure single or total toxicity (Jiang et al., 2008). A broad range of methods that have been used to develop various biosensor platforms designed either to detect OPPs or improve general detection sensitivity. To assess the literature, the search terms “biosensor” combined with “pesticide” and “biosensor” and “pesticide” combined with “organophosphate” were entered into the Web of Science search engine to identify the available research publications from 1985 to 2017. Figure 1-3 illustrates the number of publications published over this period with respect to biosensors for pesticide detection and then specifically for OPPs. While most of the biosensor detection methods that appear in the literature have not been marketed commercially, they have all been evaluated in laboratory settings.

Chapter 1

900

800

700

600

500

400

300

200

No. ofNo. PapersPublished 100

0 1985-1989 1990-1994 1995-1999 2000-2004 2005-2009 2010-2017 Year of Oublication

Biosensor & pesticides Biosensor, pesticides & organophosphate

Figure 1-3: Graph shows the number of published papers for biosensors and pesticides in general and OPP detection over the last 33 years (from Web of Science at 28/08/2017).

In the last few decades, a broad range of biosensor platforms have been developed to screen and quantify rapidly pesticide residues in various solutions and matrices. Biosensors are generally categorized into two types based either on the nature of their recognition elements (such as enzymes, antibodies, nucleic acids, whole-cells, molecularly imprinted polymers and aptamers) or on the transducers used (such as electrochemical, optical and piezoelectric).

In addition, biosensor assays offer many advantages over traditional methods, such as highly specific detection of targets in complicated mixtures, sensitivity in detecting low pollutant concentrations, user-friendly procedures, shorter sample preparation requirements, higher accessibility with the possibility to be performed in the field, quick response rates, lower costs, mobility and real-time monitoring capabilities. Indeed, many researchers consider the use of biosensors as an increasingly viable approach for detecting chemical and biological targets in food, environmental and clinical settings (Verma and Bhardwaj, 2015).

Chapter 1

Bio-recognition receptors Together with advancements in fields such as chemistry, biotechnology and nanotechnology that have improved biosensor performance and facilitated pesticide detection methods, ongoing research has also improved sensing elements. Many researchers have addressed improvements to bio-recognition elements in the design of different biosensors using materials in addition to enzymes, such as antibodies, whole- cells, recombinant DNA. Unlike enzymes, antibodies have biological and sensing characteristics that can recognize individuals or groups of pesticides based on their chemical structures, and establish a basis for analytical detection. However, some challenges exist with the antibodies, such as their relatively high cost from commercial sources, a lack of availability of suitable antibodies, general instability at extremes of pH and in a solvent or an extract and limited binding characteristics for sensitivity and specificity. Alternatively, MIPs, electroactive materials and nanoparticle composites have been used for OPP detection. Table 1-3 summaries the most common recognition elements that have been applied to date in OPP detections.

Table 1-3: Different bio-recognition elements for pesticide detections. Receptors Pesticide Transducer Limit of Sample Reference detection Parathion-methyl Optical 2.4 x 10-10 M Aqueous solution Arjmand et al. (2017) OPPs Electrochemical 4.5-14 ng/mL PBS buffer Guler et al. (2017) Ethyl paraoxon Electrochemiluminescence 0.3 pM PBS buffer Wang et al. (2017b) Chlorpyrifos Electrochemical 10-3 M PBS buffer Mogha et al. (2016) Chlorpyrifos Electrochemical 0.02 µg/L Water & vegetable Wang et al. (2016a). Parathion Optical – ES SERS 4×10-14 M Aqueous El Alami et al. (2016) Paraxon and Electrochemical – 0.7 & 3.9 nM Water Lang et al. (2016) dimethoate amperometric Malathion Electrochemical – 1 fm Spiked vegetable Kaur et al. (2016b) amperometric CPF-methyl Electrochemical- DPV 0.001 ppb Spiked vegetable Kaur et al. (2016a) Malathion Electrochemical – CV 4.08 nM Guler et al. (2016) Amplification Ethyl- paraxon Electrochemical - CV 0.33 nM Spiked vegetable Wang et al. (2016b) Dimethoate Electrochemical - CV 1×10-13 PBS buffer Zheng et al. (2016) OPPs ECL 0.3pM Spiked vegetable Wang et al. (2016c) Paraoxon and Colorimetric reaction or 4.6 μM and 200 nM Water Bui and Abbas. (2015) Enzyme derivatives UV–vis absorption Parathion Colorimetric probe 0.081 ng/mL Spiked water Bala et al. (2015) Paraoxon Electrochemical 0.5 µM Apple Zhang, Y. (2015) OPPs optical biosensor 0.1 ng/mL Buffer Kaur et al. (2015) Fenitrothion fluorescence detector 1.34 μg/mL Grape and orange Diaz et al. (2015) juices Sarin, VX and pH meter 100 pM, 100 pM and Buffer Yu et al. (2014) paraxoxon (potoluminescent quantum 8 nM dots) Malathion and Colorimetric microdevice 90 nM and 33 nM Buffer Wang et al. (2014) acephate Methamidophos, liquid crystal biosensor 10 ng/mL Buffer Ding et al. (2014) trichlorfon and paraoxon Paraoxon Colorimetric dipstick 10−7 mol/l (4pmol) Spiked water Pohanka, M. (2012) Paraoxon Fluorescence quenching 3.5 × 10−12 mol/L Buffer Wang, K. (2011) Paraoxon Spectrophotometry 32 to 48 ppb Spiked drinking Pohanka et al. (2010) water

Chapter 1

Receptors Pesticide Transducer Limit of Sample Reference detection Paraoxon, sarin and Colorimetric dipstick 5 × 10−8 M Water Pohanka et al. (2010) VX Paraxon and Fluorescence spectroscopy 2.5 μM and 10 μM Buffer Ramanathan and parathion Simonian. (2007) Chlorpyrifos pH indicator (Optical 0.04 mg/L Spiked water Kuswandi et al. (2008) fiber) Chlorphyrifos Optical 0.01 μM Buffer Chand and Gupta (2007)

Dichlorvos and pH fluorescence indicator Not reported Drinking water Vamvakaki and paraoxon Chaniotakis. (2007) Paraoxon Fluorescence-based 1 to 800 μM Aqueous solution Viveros et al.(2006) Fenitrothion Spectrofluorometer 5.5 to 19.5 μmol/L Buffer Díaz et al. (1997) Paraoxon pH meter Buffer Simonian et al. (1997) Parathion-methyl Electrochemical 3.2 nM Spiked fruit Govindasamy et al. amperometric (2017) Parathion-methyl & Electrochemical 12 nM & 10 nM Spiked vegetable Bao et al. (2016) Free parathion amperometric receptors MCP & DDVP Electrochemical – CV 0.036 & 0.012 nM Spiked Orange Sundarmurugasan et al. (2016) Fenitrothion CV 0.1 ng mL Pakohpi Wang et al. (2016d) Parathion-methyl CV 1.1×10-8 M Water Liu et al. (2016) Parathion Electrochemical- 46 pg/L PBS buffer Mehta et al. (2017) impendence OPP Electrochemical-impedric 0.1 ng/mL Rice Deep et al. (2015) Antibody Parathion Ameperometric 0.5×10-8 sunflower oil Martini et al. (2015) Parathion Electrochemical- 5 ng/L Spiked vegetable Mehta et al. (2016) impendence Parathion Luminescence 1 ppb to 1 ppm Water Kumar et al. (2016) Biomarkere Electrochemical 0.02 nM acetate buffer Du et al. (2011) Chlorpyrifos Optical 55 ng/L Drinking water Mauriz et al. (2006) Parathion-methyl Electrochemical- CV/DPV 3.16×10-10 Buffer He et al. (2017) Trizophos MIP-ECL 3.1×10-9 g/L Spiked vegetable Li et al. (2016). Methyl parathion SPR 10-13 mol/L Spiked in ACN Tan et al. (2015) Parathion QCM sensor 0.02 mM NaOH solution Özkütük et al. (2013) Methyl parathion CV 3.4 × 10−10 mol/L NaClO4 solution Li et al. (2012b) Molcurrly Omethoate Chemiluminescent 4.2 10-8 g/mL Pear, carrot, Ge et al. (2012) imprinted eggplant polymer Triazophos Chemiluminescent 2.5 nM Vegetable Xie et al. (2010) Paraoxon Voltammetric sensor 1.0×10−9 Water and Alizadeh. (2010) vegetable Parathion Cyclic voltammetry 0.003 mg/kg Vegetable Yang et al. (2009) Paraoxon Ellman assay In picomoles Water Marx and Zaltsman. (2003) Parathion-methyl Optical microplate 0.1-1 ppm Water Mishra et al. (2017a) Parathion-methyl Electrochemical 29 nM PBS buffer Bao et al. (2017) amperometric Nerve agent Microplate biosensor 4 µM Mice skin Mishra et al. (2017b) Electrochemical Parathion Colorimetric 10 µM DMSO Chong et al. (2016) Whole- Dimethoate Colorimetric/ chemo- 8.25±0.3 & 20±9.5 Buffer Dar et al. (2016) sensor probe ppm cells OPP Fluorescence 2 μM Phosphate buffer Liu et al. (2013a) spectrophotometer Paraoxon Optical fluorescence 5 μM Tap water Kim et al. (2013) Paraoxon and Amperometric 2.8 ppb and 5.3 ppb Citrate phosphate Lei et al. (2004b) parathion-methyl buffer P-nitrophenol Amperometric 5 nM (0.7 ppb) Citrate phosphate Lei et al. (2004a) buffer Malathion Electrochemical 0.001 ng/mL Spiked vegetable Prabhakar et al. (2016) Aptamer DPV/impendence Parathion-methyl electrochemical 1 × 10−12 M Water Viswanathan et al. DNA and chlorpyrifos (2009)

Antibody based biosensors

As pesticides are relatively small molecules with extremely low molecular weights, the ideal assay for the immunosensor pesticide detection is based on a competitive inhibition

Chapter 1

format whereby the analyte of interest (i.e. pesticide) competes with known levels of antigen for a limited number of antibody binding sites. As the pesticide concentration increases, antibodies are displaced from binding to the antigen, resulting in a decrease in a signal produced by a label on the primary or secondary antibody, whereas if antibodies are bound to the antigen then a detectable signal is measured through the labelled primary or secondary antibody. The response is indirectly proportional to the concentration of the pesticide present. Moreover, immunosensor methods have shown great potential to become a cost-effective approach for on-site monitoring of agricultural and environmental samples. Although the main drawback to develop immunosensor assays lies in producing antibodies with sufficient levels of specificity and high affinity for target molecules and reusability, immunosensors have emerged as a viable alternative approach for rapid screening and pesticide detection as a step before the confirmation using reference methods (Mauriz et al., 2006). For example,

Immunosensors have been used for a single or a group of OPP detection depending on the nature of the Ab specificity. The affinity interaction involved in the processes is based on immobilizing antibody or its antigen on a transducer as a recognition element to detect a target analyte. The resultant antigen-antibody complex during the assay can then be detected and quantified. In a different approach, a rapid electrochemical immunosensor for direct detection of OP-AChE. The group used zirconia nanoparticles (ZrO2 NPs) to modify a screen printed electrode using anti-human AChE monoclonal antibodies labeled with lead phosphate-apoferritin (LPA–anti-AChE). After the OP-AChE adducts were produced via incubation with paraoxon, CPF was captured by ZrO2NPs particles and labeled-antibody was used for detection that was achieved by a sandwich-like immunoreaction. This platform produced a linear response for concentrations ranging from 0.05 nM to 10 nM with a detection limit as low as 0.02 nM. Furthermore, this method overcame the challenges associated with Ab unavailability by selectivity introducing ZrO2NPs toward phosphorylated AChE and using an apoferritin label to amplify the resulting signal (Du et al., 2011). Mauriz et al. (2006) developed an anti- chlorpyrifos monoclonal antibody (Mab) to construct a portable surface plasmon resonance (SPR) platform to detect chlorpyrifos in water samples. The covalent immobilization was achieved on a gold-coated sensing surface, and the assay showed

Chapter 1

excellent sensitivity at a three-fold magnitude (55 ng/L) in buffer samples. After testing this method with real samples, the results demonstrated no matrix effects across these samples and the detection limits ranged between 45 to 64 ng/L.

In another study, Deep et al. (2015) produced a new platform to test PT using a nanometal organic framework film (NMOF) to provide greater and accurate readability for small molecules. Parathion-Ab was attached to carboxylic acid functional groups on the film, which was then coupled to electrochemical impedance transducer. It showed a detection limit of 0.1 ng/mL in rice samples. This hybrid process of a NMOF structure and Ab afforded the required level of specificity for parathion detection as well as sensor stability and sensitivity. A different work by Mehta et al. (2016) showed a very low LOD at 52 pg/L with the dynamic range of 0.1–1000 ng/L to detect PT in buffer by an electrochemical platform and the accuracy study showed an excellent recovery from spiked tomato and carrot samples. The same author improved the detection sensitivity to 46 pg/L with the dynamic range of 0.01–106 ng/L when they incorporated graphene quantum dot (GQD) on the sensing surface. The high sensitivity proposed in this method was due to the synergetic effect using a functionalized graphene quantum dot (GQD) (Mehta et al., 2017).

Aptamers-based biosensors Aptamers (i.e. DNA or RNA) are single-stranded nucleic acids that connect different bases to form new oligonucleotides for a wide range of target molecules. They can be produced in vitro using a selection process—called the systemic evolution of ligands by exponential enrichment—and can selectively bind their targets from proteins to small organic molecules (Chen et al., 2013). Aptamers are emerging as new artificial receptors used for analytical applications, and their features include high affinity and specificity, relative ease of production, high water solubility with chemical stability and ease of chemical modification. To date, analytical assays for pesticide detection by aptamers are not very well developed, and only one publication on this work was found in the literature. In this study, an aptamer was synthesized specifically for malathion, and was then immobilised on an electrode modified with a composite of chitosan/iron oxide and fluorine tin oxide for electrochemical measurements. The results of this work show a LOD of around 0.001 ng/mL with 15 minutes of analysis in buffer. In addition, vegetable

Chapter 1

samples fortified with malathion showed recoveries between 80 and 92% (Prabhakar et al., 2016).

Enzyme-based biosensor methods

The majority of the published literature on OPP detection using biosensors is enzyme- based. The enzymes used in these studies have been shown to be highly specific in detecting a number of chemicals, and their role in bioassays is to measure OPP residues by either inhibition or catalysis. Standard enzymes used for this purpose include AChE, butyrylcholinesterase (BChE), alkaline phosphatase (ALP) (acid or alkaline), tyrosinase, organophosphorus hydrolase (OPH) and aldehyde dehydrogenase and others that have been employed to measure total toxicity in samples.

First, the inhibition reaction measures indirectly the presence of inhibitors such as OPPs and other compounds decrease the enzymatic activity because of hindering the serine group at an active site of an enzyme, which leads to the generation of a phosphorylated serine through phosphorylation. Reaction measurements are based on monitoring the enzyme activity before and after pesticide exposure, and the activity is then calculated to yield an inhibition percentage related to the level of pesticide concentration. Two cholinesterase enzymes, AChE and BChE, are typically used in these reactions, and distinct substrates are employed. For example, AChE hydrolyzes acetyl esters, such as acetylcholine, whereas BChE hydrolyzes butyrylcholine. In addition, other enzymes such as tyrosinase (Sigolaeva et al., 2017; Mayorga-Martinez et al., 2014; Solná et al., 2005a), ALP (Mulyasuryani and Prasetyawan, 2015), peroxidase (Solná et al., 2005b) and methyl-parathion hydrolase (MPH) (Liu et al., 2013b) have also been used.

Secondly, the catalysis process whereby enzymes show specificity only to OPPs and catalyse the hydrolysis of acetylcholine to produce different products, protons, organophosphorus acid and alcohol. These products can then be measured directly by enzymes include OPH, phospho-triesterase, organophosphorus acid hydrolase (OPAA) and parathion hydrolase (PH), and these enzymes have broad specificity that are capable of hydrolyzing a number of OPPs that contains p-nitrophenyl OPs. During these reactions, enzymes undergo a hydrolysis process and cleave a number of OP subclasses that contain P–O, P–F, P–S and P-CN bonds with pesticides such as paraoxon, parathion, coumaphos,

Chapter 1

diazinon, chlorpyrifos, methyl-parathion, and nerve agent compounds, among others (Mulyasuryani et al., 2015; Liu et al., 2008; Walker et al., 2007).

The advantages of the hydrolysis process include the potential reusability of materials and continuous real-time monitoring. In addition, this approach can exploit enzymes in different ways and can be used to detect either single or multiple pesticides simultaneously. For example, single enzymes have been used to detect parathion-methyl (Arjmand, 2017), paraoxon (Gong et al., 2012), parathion (Pedrosa et al., 2007), and bienzymes (i.e. AChE and OPH) have been employed to detect paraoxon (Zhang et al., 2015a). Moreover, a byproduct is produced as a result of these enzymatic reactions. For example, inhibition processes or catalytic reactions between OPH and OPP who contains p-nitrophenyl substituents such as paraoxon, parathion, MP, coumaphos, diazinon, and chemical warfare agents (sarin and soman) produce p-nitrophenol (PNP), which then can be detected as well.

In addition, a colourimetric method was developed based on AChE and BChE inhibition utilizing indoxylacetate as a substrate for a paper-based sensor. The colour changes in spiked tap and rain water samples with paraoxon as an inhibitor showed a limit of detection at 10−7 mol/L with stability for a one-month period (Pohanka, 2012). In contrast, Jin et al. used a fluorogenic reagent in which a pH-sensitive fluorescence probe was developed as a pH indicator to detect OPP residues in a flow injection system. This procedure was carried out by mixing the enzyme solution contained acetylcholine with paraoxon as a substrate and the fluorescent dye intensity was measured based on the level of H+ ions released following the hydrolysis process, which provided a detection limit of 12 ng/mL in 10 minutes in a spiked water (Jin et al., 2004). In another study, a method was developed to detect paraoxon via enzyme inhibition and fluorescence quenching. Thiocholine (TCh) as a product of AChE hydrolysis able to generate a blue fluorescent compound after binding with a thiol-reactive material, which then undergoes a reduction process in the presence of paraoxon before a fluorescence quenching by a PNP compound occurred showing a final detection limit in a water sample at 3.5×10−12 mol/L (Wang et al., 2011b). Yu et al. (2014) used a catalytic enzymatic reaction (AChE) in which they exploited ATCh for coupling photoluminescent CdTe QDs. OPPs were detected through the hydrolysis of ATCh by AChE. The subsequent release of both H+ and a positively

Chapter 1

charged thiol group (-SH) from thiocholine showed improved sensitivity results with detection limits at 100 pM, 100 pM and 8 nM for sarin, VX and paraoxon respectively, with a dynamic range found to be between 8 nM and 8 µM in a buffer solution for paraoxon. Table 1-4 illustrates different enzyme-based methods that have been used to develop biosensing techniques.

Table 1-4: Enzyme-based methods used in pesticide sensing techniques. Pesticide Transducer Limit of detection Sample Reference Chlorpyrifos pH indicator (Optical fiber) 0.04 mg/L Spiked water Kuswandi et al. (2008)

Chlorpyrifos Surface plasmon resonance 0.01 μM Buffer Chand and Gupta. (2007) DFP and parathion Fluorescence spectroscopy 2.5 μM and 10 μM Buffer Ramanathan and Simonian. (2007) Diazinon and Electrochemical 6 × 10-12 M and Aqueous Sigolaeva et al. (2017) Chlorpyrifos 8 × 10-12 M Dichlorvos and pH sensitive fluorescence indicator Not reported Drinking water Vamvakaki and paraoxon Chaniotakis. (2007) Fenitrothion fluorescence detector 1.34 μg/mL Grape and Diaz et al. (2015) orange juices Fenitrothion Spectrofluorimetric methods, 5.5 to 19.5 μmol/L Buffer Díaz et al. (1997) fluorogenic substrates Malathion Electrochemical 4.08 nM Buffer Guler et al. (2016) Malathion and acephate Colorimetric microdevice 90 nM and 33 nM Buffer Wang et al. (2014b) Methamidophos, liquid crystal biosensor 10 ng/mL Buffer Ding et al. (2014) trichlorfon and paraoxon OPPs Eggshell membrane-based optical 0.1 ng/mL Buffer Kaur et al. (2015) biosensor OPP (sarin, VX and pH meter (potoluminescent 100 pM, 100 pM and Buffer Yu et al. (2014) paraxoxon) quantum dots) 8 nM Paraoxon Optical 4 × 10−14 M El Alami et al. (2016) Paraoxon Surface plasmon resonance 0.5 µM Apple Zhang et al. (2015a) Paraoxon Colorimetric reaction or UV–vis 4.6 μM Water Bui and Abbas. (2015) absorption Paraoxon Colorimetric dipstick 10−7 mol/l (4pmol) Spiked water Pohanka. (2012) Paraoxon Fluorescence quenching 3.5 × 10−12 mol/L Buffer Wang et al. (2011b) Paraoxon Spectrophotometry 32 to 48 ppb Spiked drinking Pohanka et al. (2010a) water Paraoxon, sarin and Colorimetric dipstick 5 × 10−8 M Water Pohanka et al. (2010b) VX Paraoxon Fluorescence-based biosensor 1 to 800 μM Aqueous Viveros et al. (2006) solution Paraoxon pH meter - Buffer Simonian et al. (1997) Parathion Colorimetric probe 0.081 ng/mL Spiked water Bala et al. (2015) Parathion-methyl Optical 0.24 nM Buffer Arjmand et al. (2017)

Enzyme-based biosensors can only detect the general toxicity of inhibitors and cannot provide information related to specific targets. In other words, while enzymes are very sensitive in detecting the presence of OPPs, they lack selectivity and do not provide high efficiency for generating qualitative or quantitative data for individual pesticides. As a result, antibodies as secondary recognition materials are generally incorporated into biosensor platforms after enzymes.

Chapter 1

Whole-cell based biosensors

Whole-cells are another recognition element that researchers have used to metabolize a substrate for detection. Whole-cell-based sensing has been employed to avoid the challenges presented by other conventional methods and attempted to construct easy to use, inexpensive and portable biosensors for onsite detection. Whole-cell based applications have been adopted in many fields, including food safety. In addition, they are good bio-recognition elements for pesticide detection because bacteria can degrade different pesticides. Furthermore, they have been investigated for detoxifying OPPs to develop other biosensors. Table 1-5 lists whole-cell biosensor platforms coupled with various materials.

Table 1-5: Platforms developed to detect OPPs using whole cells as recognition elements. Pesticide substrate Transducer Limit of detection Sample Reference P-nitrophenol Amperometric 5 nM (0.7 ppb) buffer Lei et al. (2004a) Paraoxon Optical fluorescence 5 μM tap water Kim et al. (2013) Paraoxon and parathion-methyl Amperometric 2.8 and 5.3 ppb buffer Lei et al. (2004b) Parathion Colorimetry 10 μM buffer Chong and Ching. (2016) Parathion-methyl Fluorescence 2 μM buffer Liu et al. (2013b) spectrophotometer Parathion-methyl Optical biosensor 0.1–1 ppm buffer Mishra et al. (2017a)

Lei et al. (2004a) evaluated the results of modifying the surface of a carbon paste electrode (CPE) with a nafion polymer for immobilization of Arthrobacter sp. JS443 to produce a whole-cell based amperometric biosensor. This biosensor was able to detect PNP, a byproduct of OPP hydrolysis. The results showed a detection limit of 5 nM (0.7 ppb) and the biosensor stability lasted for 3 days when stored in a citrate-phosphate buffer at 48ºC. The platform was applied successfully to measure PNP in spiked lake water samples. The merit presented here was its selectivity to PNP. Differently, the same group investigated a hybrid system for which two layers were prepared on CPE for sensing, whereby OPH for an initial hydrolysis process and Arthrobacter sp. strain JS443 for later PNP oxidation to target those OPPs such as paraoxon, parathion-methyl, parathion, fenitrothion, and EPN with PNP-substituents. These two layers were co-immobilized on electrodes for further detection by an amperometric platform. The pesticides were first hydrolyzed to produce PNP, and then underwent an oxidation process by bacteria to release carbon dioxide for quantification and showed detection limits of 2.8 ppb (10 nM) and 5.3 ppb (20 nM) in a buffer for paraoxon and PM, respectively. It had no interference with other

Chapter 1

phenolic compounds and was stable for over 40 repetitions. Also, the assay was tested in spiked river and lake water samples and showed no matrix effects (Lei et al., 2004b).

In another study, Liu et al. (2013b) constructed a biosensor to detect OPPs by combining two enzymes OPH and MPH. MPH was labelled with a green fluorescent protein (GFP) to produce a MPH/GFP fluorescent marker on the cell surface of Escherichia coli bacteria. Results showed that the cell surfaces exhibited broader target specificity concerning to OPPs to using OPH or MPH solely whereby when six different pesticides were prepared each at a 0.2 mM, a complete degradation was observed with a very sensitive florescence marker for pH changes in buffer samples. This method could potentially avoid the challenges associated with purification costs, and the enzyme stability.

Kim et al. (2013) developed a whole-cell array biosensor by coating the surface of a 96- well microplate with Escherichia coli cells, through mussel adhesive protein as a linker, capable of producing OPH. The results showed that the activity of paraoxon hydrolysis and fluorescence microscopy analysis enhanced cell-immobilizing efficiency and increased stability when compared to other procedures. It showed a detection limit of 5 µM with the biosensor stability of 28 days and could be reused 20 times. In a different approach, Mishra et al. (2017) used a biohybrid method to detect PM. Functionalized silica nanoparticles and Sphingomonas sp. cells were used to coat a 96 well microplate. After analysis by an optical biosensor, the sensitivity was improved, and the detection limit in buffer was in the range of 0.1–1 ppm.

Molecularly imprinted polymer-based biosensors

Along with using biological recognition elements, many researchers have investigated the use of artificial materials such as MIPs. They have been developed to resemble biological recognition elements (Ge et al., 2012). MIP-based biosensors are designed to detect compounds through selective recognition sites in various matrices (Yang et al., 2009). According to the literature, MIPs have been produced to detect PM (Tan et al., 2015; Li et al., 2012b), paraoxon (Özkütük et al., 2013; Alizadeh, 2010; Marx and Zaltsman, 2003), omethoate (Ge et al., 2012), triazophos (Xie et al., 2010) and parathion (Yang et al., 2009). In addition, many MIPs were used as recognition elements and had been

Chapter 1

coupled with various electrochemical, optical and piezoelectric platforms. Table 1-6 below summaries several applications for MIP-based biosensors that have been used for OPP detection.

Table 1-6: Platforms developed and coupled with MIPs as recognition elements for OPP detection. Pesticide Transducer Limit of detection Sample Reference Omethoate Chemiluminescent 4.2 108 g/mL Pear, carrot, eggplant Ge et al. (2012) Paraoxon Voltammetric sensor 1.0×10−9 Water and vegetable Alizadeh (2010) Parathion Cyclic voltammetry 0.003 mg/kg Spiked vegetable Yang et al. (2009) Paraoxon Ellman assay In picomoles Water Marx and Zaltsman. (2003) Parathion QCM sensor 0.02 mM NaOH solution Özkütük et al. (2013) Parathion-methyl Electrochemical 3.16 ×10 −10 mol/L Buffer He et al. (2017) Parathion-methyl SPR 10-13 mol/L Spiked in ACN Tan et al. (2015) Parathion-methyl CV 3.4 × 10−10 mol/L NaClO4 solution Li et al.(2012b) Triazophos electrochemiluminescence 3.1 × 10−9 g/L Vegetable Li et al. (2016) Triazophos Chemiluminescent 2.5 nM vegetable Xie et al. (2010)

MIPs and other materials as the interface of the sensor were introduced. He and coworkers (2017) developed MIPs with innovative techniques to detect PM. Porphyrin was attached to MIP microspheres for modification of a composite of carboxyl graphene and gold nanoparticles to improve and amplify the signal. The electrochemical analysis showed a LOD at 3.16 ×10 −10 mol/L with a dynamic range between 1.0 ×10 −6 mol/L and 8.0 ×10−9 mol/L (He et al., 2017). A different method to improve the sensitivity of triazophos detection was produced by the electro-chemiluminescence light measurement, a membrane was used for modification. MIP was polymerised on membrane at GCE which was previously modified with a composite of gold nanoparticle and carbon nanotube. The fabricated sensor was able to detect triazophos residues in vegetable samples with a LOD at 3.1 × 10−9 g/L with a range between 3.1 × 10−8 and 3.1 × 10−5 g/L and the recovery was performed and found to be in the range of 94-108% (Li et al., 2016). Li et al. developed a fast, selective and sensitive MIP-based electrochemical sensor by combining MIPs with capacitive transduction to improve capacitance sensitivity for a real time pesticide detection. An electro-polymerization method was used to produce a MIP film for the modification of a gold nanoparticle (AuNP) electrode for PM identification. The analysis was made by impedance spectroscopy (EIS) methods for distilled, tap, river, rainwater, and fruit samples. The results showed satisfactory recovery levels for all samples, and the sensor response for PM was linear, ranging from 7×10−8 mol/L to 1×10−6 mol/L with a detection limit of 3.4×10−10 mol/L (Li et al., 2012). Zhang et al. (2016) used

Chapter 1

porphyrins as a functional monomer to develop molecular imprinting membrane. They synthesised a coloured MIM using dimethyl methylphosphonate as a template, with two mixed monomers of both zinc porphyrin and MAA. Thermal polymerization was used on a glassy chip surface previously functionalized with EDGMA. The membrane sensor showed selectivity towards the target and the detection limit was found to be 0.1 µmol/L in water samples.

Tan et al. fabricated another sensitive MIP-based optical sensor, using SPR with a bare gold surface sensor chip. The sensor was applied to a PM solution in acetonitrile. The MIP film showed a linear response of between 10-13-10-10 mol/L with a detection limit as low as 10-13 mol/L. The film was evaluated with diuron, tetrachlorvinphose and fenitrothion, and demonstrated high selectivity and affinity towards parathion-methyl (Tan et al., 2015). Another group developed a MIP film and modified a surface of piezoelectric crystals to detect paraoxon. The film exhibited selectivity and sensitivity towards paraoxon with a linearity between 0.02 and 1µM and a detection limit of 0.02 mM in a solution. However, the sensor also showed selectivity for PT at high concentrations (Özkütük et al., 2013). Ge et al. (2012) coupled molecularly imprinted microspheres with chemiluminescence to target two classes, organophosphate and carbamate, using omethoate and carbofuran, respectively. They employed a multi-branch flow cell to conduct multi-residue analysis in which each branch was packed with a corresponding polymer, and a chemiluminescence transducer was connected to flow injectors in order to analyze the two pesticides. The chemiluminescence-molecular imprinting (CL–MI) sensor improved the sensitivity and selectivity. For omethoate, the linearity response ranged from 1.0×10-7 to 9.0×10-6 g/mL with a detection limit of 4.2×10- 8 g/mL. This method showed favourable applicability in detecting these pesticides in food samples, and the MIP was also able to separate and pre-concentrate samples.

Another group developed a MIP film and modified a surface of piezoelectric crystals to detect paraoxon. The film exhibited selectivity and sensitivity towards paraoxon with a linearity between 0.02 and 1µM and a detection limit of 0.02 mM in a solution. However, the sensor also showed selectivity for PT at high concentrations (Özkütük et al., 2013). Ge et al. (2012) coupled molecularly imprinted microspheres with chemiluminescence to target two classes, organophosphate and carbamate, using omethoate and carbofuran,

Chapter 1

respectively. They employed a multi-branch flow cell to conduct multi-residue analysis in which each branch was packed with a corresponding polymer, and a chemiluminescence transducer was connected to flow injectors in order to analyze the two pesticides. The chemiluminescence-molecular imprinting (CL–MI) sensor improved the sensitivity and selectivity. For omethoate, the linearity response ranged from 1.0×10- 7 to 9.0×10-6 g/mL with a detection limit of 4.2×10-8 g/mL. This method showed favourable applicability in detecting these pesticides in food samples, and the MIP was also able to separate and pre-concentrate samples.

Alizadeh et al. (2010) examined imprinting paraoxon (OP metabolite) for biosensor detection. The authors investigated three methods to integrate MIPs with an electrochemical transducer: 1) embedding the MIP polymer in carbon paste (MIP-CP); 2) coupling the MIP with a glassy carbon electrode surface using poly epychloro hydrine (MIP-PECH-GC); and 3) preparing a thin layer of MIP/graphite for coupling onto a GC electrode. The sensor assessment was based on its selectivity to paraoxon and the effect of the washing step on different electrode responses in a real sample. The MIP particles embedded into a carbon paste electrode demonstrated the most favourable results, in that a satisfactory limit of detection was achieved at 1.0×10−9 nmol/L with high selectivity and sensitivity for paraoxon in water and vegetable samples by a volumetric sensor. In a different study, researchers used a new MIP to develop another electrochemical sensor for PM detection. An electropolymerisation method was used to synthesise a thin film on glassy carbon electrode using a silicon precursor, tetraethyl orthosilicate, and vinyltriethoxysilane as a functional monomer. The PM-imprinted film was deposited on the surface of the electrode. MIP sensitivity was evaluated in a phosphate buffer and showed a detection limit of 8.9 × 10−9 mol/L (Tan et al., 2010). In addition, Yang et al. developed a MIP film. Polyethylenemine as a functional monomer was grafted on surfaces of silica gel particles (MIP-PEI/SiO2). The composite was mixed with a chitosan solution and then the mixture was used to modify the transducer surface of cyclic voltammetry and linear sweep voltammetry. The MIP was very selective to PT and exhibited linearity at a concentration of 0.015-15 mg/kg and a detection limit of 0.003 mg/kg in spiked vegetable samples (Yang et al., 2009). In a different approach, Wei et al. (2011) prepared an ultrathin MIP film by surface initiated radical polymerisation on a

Chapter 1

sensor chip for acephate detection by surface plasmon resonance. The biosensor film displayed good selectivity toward acephate when compared to other pesticides, including malathion, phoxim, chlorpyrifos, methamidophos, profenofos and trichlorfor. Additionally, the biosensor was evaluated using real samples and the detection limits of apple and cole, Chinese vegetable, were 1.14 ×10−13 M 4.29 ×10−14 M, respectively. Moreover, Li at al. (2016) adopted a surface–molecular approach to develop a MIP electrochemiluminescent sensor for triazophos. A composite of gold nanoparticle and carbon nanotube was employed to modify a glassy carbon electrode before using electropolymerization to imprint triazophos. The resultant MIP based sensor demonstrated high sensitivity and selectivity towards triazophos and showed a limit of detection of 3.1×10-9 g/L in vegetable samples.

Nanomaterials

The growth in nanomaterial applications has opened new areas of research in analytical detection techniques. In addition, various nanomaterials have been identified as possessing their inherent physical-chemical characteristics, such as size, shape and type. Besides their electrochemical and optical features, nanoparticles also afford higher surface to volume ratios that permit the attachment of greater numbers of biomolecules, which allows for more available positions that can react with a range of targets. The size and shape of these materials can be easily controlled, and a composite of several materials can be produced (Guler et al., 2017). Nanomaterials either a single or a composite are normally immobilised onto a sensing surface to produce an innovative interface. Some are very conductive and can accelerate the electron transfer during the analysis of electrochemical enzyme-based biosensors between the enzyme redox centre and the sensing surface (Liu et al., 2013c). As a result, these characteristics have shown nanomaterials to be favourable for the production of improved assays (Verma and Bhardwaj, 2015). In addition, the synergistic effects produced by combing different nanomaterials and sensors have been investigated in the last few years, and the design of nanomaterial-based biosensors have been researched intensively to study sensitivity and stability. Table 1-7 lists different nanomaterials and composites that have been used in biosensor applications to detect OPP residues.

Chapter 1

Table 1-7: Various nanomaterials and composites were used in biosensor applications to detect OPP residues. Detected pesticides Nano materials/ composite Biosensor Detection limit Sample Reference Trichlorfon and coumaphos polyaniline (PAn) make sure electrochemical 1×10-6 and buffer Ivanov et al. (2003) 9 ×10-8 mol/L Methylphosphonic acid Zirconium phosphate/ Optical > 5×10−5 M Ethanol Newman et al. (2007) (MPA)anddiethylchlorophos phosphonate modified AuNPs phate parathion and chlorpyrifos Single-walled carbon nanotubes electrochemical 1×10−12 M water Viswanathan et al. (2009) Parathion-methyl Poly(malachite green)/graphene electrochemical 2.0 nM Aqueous Xu et al. (2012c) nanosheets–nafion solution Parathion-methyl MWCNTs/poly(acrylamide) electrochemical 2.0× 10−9 mol/L water Zeng et al. (2012) (MWCNTs-PAAM) Parathion-methyl Bismuth oxyiodide, nanoflake Optical 0.04 ng/mL PBS Gong et al. (2012) arrays and titanium dioxide buffer (BiOINFs/TiO2) Parathion-methyl Polished aluminium oxide electrochemical 0.035 ng/mL buffer Wei et al. (2012) nanoparticles (α-Al2O3) Dursban Cd0.5Zn0.5S-rGO nanocomposite electrochemical 0.3 ng/mL buffer Liu et al. (2014) Malaoxon and paraoxon Garnished silver nanoparticles electrochemical 0.1 and 0.05 nM buffer Evtugyn et al. (2014). Malathion MWCNTs and nanoporous gold electrochemical 0.5 ng/mL buffer Ding et al. (2014) Parathion-methyl reduced graphene oxide electrochemical 0.06 ng mL buffer Liang et al. (2014) nanosheets (GNs) CdTe QDs. Parathion-methyl Graphenenano sheets-gadolinium electrochemical 1 nM buffer Li et al. (2014b) Prussian blue Parathion-methyl Nanoporous gold/MWCNTs electrochemical 4.8×10−8 g mL water Wan et al. (2014). Chlorpyrifos IrOx NPs and enzyme electrochemical 0.003 μM BPS Mayorga-Martinez et al. (2014) Monocrotophos Nanoporous polyaniline and electrochemical 0.1 ppb PBS Kaur & Srivastava. zeolite (PANI-Nano-ZSM-5) buffer (2015) Chlorpyrifos Ag/Cu–graphene electrochemical 4 × 10−12 M Water Sreedhar et al. and (2015) soil Parathion Nanocrystal metal organic electrochemical 0.1 ng/mL Rice Deep et al. (2015) frameworks and antibodies Diazinon a micro IrOx electrochemical 3 µM buffer Wang et al. (2015) Malathion conducting polymer (CP)- 1 fM buffer Kaur et al. (2016b) Poly(3,4-ethylenedioxythi ophene) (PEDOT/ MWCNT) Chlorpyrifos-methyl MWCNTs electrochemical 0.001 ppb buffer Kaur et al. (2016a) Parathion-methyl Titanium dioxide nanofibers/ electrochemical 12 nM and 10 nM buffer Bao et al. (2016) and parathion carboxylic acid functionalized multi-walled carbon nanotubes Chlorpyrifos Graphene oxide (RGO) and electrochemical 10− 13 buffer Mogha et al. (2016) zirconium Oxide (ZrO2/RGO) Chlorpyrifos Fe3O4 Nanoparticle and electrochemical 0.02 µg/L buffer Wang et al. (2016a) Graphene Dimethoate Graphene and cobalt (Co3O4) electrochemical 1.0 × 10−13 M buffer Zheng et al. (2016) nanoparticles Paraoxon and dimethoate Gold nanorods electrochemical 0.7 nM and Lang et al. (2016). 3.9 nM Monocrotophos and Zinc oxide electrochemical 0.036 and Orange Sundarmurugasan et Dichlorvos 0.012 nM al. (2016) Dimethoate Gold NPs Colorimetry 8.25± 0.3 - - Dar et al. (2016) 20 ± 9.5 ppm Malation, methidathion, and Graphene oxide (rGO-NH2) electrochemical 4.5, 9.5 and buffer Guler et al. (2017) chlorpyrifos ethyl nafion (NA) and Agnanoparticles 14 ng/mL

1.6.2.5 Microarray technology to nanosensors for food analysis

Microarray-based analytical systems have become an appealing replacement method due to their high throughput levels for a test, high sensitivity, enhanced reproducibility, small

Chapter 1

sample requirements, speed, and ease of automation (Zhang et al., 2012b). Microarray assays are based on specific biomolecular recognition on a certain well-defined heterogeneous substrate. Antibody molecules are the most commonly used materials to identify target analytes on microarrays, due to their specificity and sensitivity that can meet required low EU MRLs. In addition, in recent years a synthetic molecular recognition element has been produced as an affinity material for the analysis of chemical contaminants, such as molecularly imprinted polymers (Zhang et al., 2012b).

Microarray technology is an advanced system suitable for analysis of contaminants such as mycotoxins, biotoxins, pesticides and pharmaceutical residues (Spanjer et al., 2008; Basta et al., 2005). The principle of this technology is based on a particular biomolecular recognition pattern for a defined heterogeneous substrate. Microprinting or microspotting methods are used to immobilise a conjugate or an antibody on a selected solid support surface. Using this method, recognition and identification of the target compound in samples can be achieved semi-quantitatively or quantitatively. Moreover, this approach is based primarily on using antibodies as recognition elements to provide specificity and sensitivity; it is one of the most popular techniques used for pollutant analysis. This technique has several advantages, such as high throughputs, sensitivity; enhanced reproducibility; low sample consumption; reduced processing time, and ease of computerization (Zhang et al., 2012b; Raz and Haasnoot, 2011; Angenendt, 2005).

Screening multi-residues in a single assay affords many advantages, such as simplicity, speed, lower costs, use of fewer reagents and shorter time periods. Generally, there are two main categories of multiplexed assays: planar microarrays (i.e. planar waveguides) and bead‐based microarrays (i.e. suspension arrays). With respect to planar microarrays, most assays are conducted on a spatially resolved surface using several immunological reagents that are attached on planar support and are encoded using the coordinates of their positions. However, these approaches are restricted to a small number of pesticides for field analysis. For example, immunochromatographic test strips (immunostrips) have been used to identify both methyl-parathion and imidacloprid simultaneously using bifunctional antibody and pesticides-labeled enzymes (Shu et al., 2017); the analysis was performed on traditional Chinese medicine samples in 22 minutes. In another study, Xu et al. developed a strip-based immunoassay utilising specific nanocolloidal gold-labeled

Chapter 1

monoclonal antibodies to detect both imidacloprid and thiamethoxam in a semiquantitative analysis (Xu et al., 2012a). In addition, Rubtsova et al. developed a multi‐spot membrane strip to identify three pesticides (atrazine, simazine and terbuthylazine) in water samples using an electrochemiluminescence system (Rubtsova et al., 1998). For the bead‐based microarrays, Luminex xMAP is a leading platform based on flow cytometry. A few studies have reported multiplexing techniques for pesticide detection (Wang et al., 2014a; Guo et al., 2013; Biagini et al., 2004). These innovative platforms that rely on immunological assays have demonstrated the benefits of such methods. Nevertheless, most of these mentioned methods above were constrained by laboratory analysis. However, although the strip tests for pesticides are used for on-site analysis, it is still restricted by the number of targets that can be detected and mostly provide a qualitative test (Sajid et al., 2015). The MBio SnapEsi reader using a planar waveguide is advantageous over the test tripe and offers the portability for on-site analysis and the capability to build a full array of multiple pesticides from different classes (multiplexing). The MBio system were demonstrated for a singleplex method to detect paralytic shellfish toxins (Meneely et al., 2013) and multiplex methods to detect three antibiotics in milk (McGrath et al., 2015), five groups of harmful algal toxins in seawater samples (McNamee et al., 2014) and three different targets in a medical sample (Lochhead et al., 2011). Indeed, the MBio biosensor is an innovative technique showing a promise for portability, and both singleplexing and multiplexing methods.

1.6.2.6 Computational modelling Several software applications have been used to predict responses from available data provided by biosensors samples to classify different pesticides. Specifically, artificial neural networks (ANNs) and support vector machines (SVMs) have been used to determine the presence of OPPs. Based on given enough data, these computational simulations softwares can be trained to give accurate estimations and reduce false negative errors. SVMs have been used to improve the prediction of accurate classification of the results produced by multi-array biosensors for a nerve agent and other OPPs such as parathion, malathion, dichlorvos, trichlorfon, paraoxon, and diazinon in samples (Sadik et al., 2004). In addition, Istamboulie et al. (2009) used ANNs to improve the selectivity of those biosesnors previously developed to identify chlorpyrifos and

Chapter 1

chlorfenvinfos. ANNs were used to simulate responses of chemometric data that generated by amperometric measurements for these two pesticides utilising different enzymes as recognition elements. Interestingly, both software packages have shown improvements in the selectivity and accuracy when compared to control assays.

1.7 Sample preparation

Determining pesticide residue levels in fruits, vegetables and water requires analysis, and the analytical method depends on sample preparation and instrument selection. It is important to note that sample preparation is a key to the efficient and accurate analysis of pesticide residues and their corresponding trace compounds. Therefore, the aim of sample preparation is to isolate trace amounts of analytes from a complicated matrix and to remove as many interferences as possible from a sample matrix (Ferrer et al., 2005). It improves the detection accuracy and reduces components which cause the deterioration of the analytical machines through blocking or clogging of tubes reducing breakdowns which then boosts the laboratory productivity (Flanagan et al., 2008). Most typical sample preparation protocols for Gas chromatography (GC) and high-performance liquid chromatography (HPLC) coupled to mass spectrometry require the following general steps: homogenisation of a food sample, extraction of target pesticides with appropriate solvents, and a clean-up step by centrifugation or filtration to produce a sample extract. Also, a sorbent such as C18 may be required using a certain protocol to fractionate or free the retained targets. There are several sample clean-up protocols available for pesticides and they have their own pros and cons. The selected protocol will influence the analyte recovery and vary from one to another (Vas and Vekey, 2004).

1.7.1 Liquid-liquid extraction (LLE)

Traditionally, the LLE approach was the most popular extraction technique for pesticide residues and requires mixing mainly immiscible solvents with the sample to equilibrate the partition of the target pesticide and other interferences in the two phases (donor and acceptor) (Hanson, 2013). The technique is used for the standard sample preparation because it is simple, robust, efficient and relatively cheap in materials (Zhang et al., 2012a). Although it still has been used either separately (Tahboub et al., 2006, Wang, et al., 2017a) or in conjunction with clean-up techniques such as low-temperature

Chapter 1

purification (Sobhanzadeh et al., 2011), or dispersive solid phase extraction (Fang et al., 2017), the drawbacks of this technique are time-consuming, requiring a high amount of solvent and additional laboratory work (Zhang et al., 2012a).

1.7.2 Solid phase extraction (SPE)

The SPE is the most common technique applied for sample preparation, especially for pesticide residue analysis in fruit and vegetable (Sabik et al., 2000). A solvent extract is loaded through a chromatographic column contain sorbent which previously conditioned and activated, able to hold the targeted analytes and other coexisting substances. Followed by the wash of unwanted components and the elution of the analytes using suitable solvents to recover relatively clean targets (Simpson, 2000; Chen et al., 2010). Commercially, there are several SPE columns with different types of sorbents and bed dimensions such as C18. The extraction takes place by retaining a target analyte to an adsorbent material in a column before releasing it again in the elution (Chen et al., 2007). SPE has an uncomplicated experimental protocol, simple to operate, requires less solvent and also can be automated, which allows for SPE to be a replacement to LLE for sample preparation specifically clean up and pre-concentration. Although it is extremely efficient for pesticide analysis, it still has drawbacks. For example, it is not a straightforward procedure but needs some time to select adsorbents and appropriate solvents for washing and elution especially when many pesticides require analysis. Also, commercially available SPE columns are disposable, which add constraints when considering available budgets (Zhang et al., 2012a).

1.7.3 Other techniques

As each method has its shortcomings, many sample preparation methods for pesticides analysis in food have been introduced to overcome some of these disadvantages. The sample preparation techniques include Supercritical-ﬂuid extraction which is used for analytes present in a solid material to speed the extraction time and enables the solvent to penetrate into the matrix and then to extract the analyte (Martín et al., 2011; Rial-Otero et al., 2008). Pressurized-liquid extraction is used for solid and semi-sold samples under high pressure and temperature to accelerate the solvent extraction (Barriada-Pereira et al., 2007; Giergielewicz-Możajska et al., 2001). Similarly, microwave-assisted extraction is

Chapter 1

used for fast extraction and initially was for the extraction of organic contaminants before being effectively applied for pesticide residues (Beyer and Biziuk, 2010; Ji et al., 2007). Ultrasound-assisted extraction was developed for fast and efficient extraction where the generated effect by acoustic cavitations can ease the damage of the structures of the food cells and allow organic solvents to penetrate and release contaminants (Hemwimon et al., 2007). Gel permeation chromatography is a chromatographic technique, and the separation relies on the molecule size which determines the elution time of each analyte. Matrix solid-phase dispersion involves both extraction and clean-up steps in one single procedure (Rezaei and Hosseini, 2011; Ramos et al., 2008). Finally, solid phase microextraction requires the analytes to go through the partition equilibrium between the matrix and the stationary phase (Kimm et al., 2002).

1.7.4 Molecularly imprinted polymers

MIPs are synthetic materials that have been utilised as SPE sorbents, it is produced by a polymerisation technique through a combination of a monomer and a cross-linker with a target template molecule for imprinting combined with an appropriate organic solvent and initiator to start the polymerisation. Once the template is removed from the resultant solid matrix polymer, the polymer has cavities which are complementary in size, shape and chemical functionality to the template representing recognition sites. These sites can then bind the template or similar structures. The MIP provides specific recognition with great affinity to the imprinted template. Therefore, the MIP is able to rebind and remove the target biological and chemical compounds such as amino acids, protein, nucleotide, pollutants and pesticides from a sample extract or solution. These targets can be found in biological, environmental, drugs and food samples (Vasapollo et al., 2011; Alexander et al., 2006). The MIP is highly appealing for recognizing and enriching pesticide residues selectively (Lu et al., 2010; Hantash et al., 2006). MIPs have been synthesised in different formats for many pesticides (Pichon and Chapuis-Hugon, 2008), and permit the direct analysis of sample extracts. For example, a trace amount of diazinon was detected in a water sample using MISPE before HPLC analysis (Rahiminejad et al., 2009). Thiophosphate and diethyl dithiophosphate were detected in urine after applying MISPE before GC-MS analysis (Santos et al., 2012).

Chapter 1

As MIPs can resemble biological receptors such as antibodies, many fields have benefited from their development, including clinical analysis, medical diagnostics, and environmental monitoring. Furthermore, in the last two decades, molecularly imprinted materials have been used in a variety of applications, such as affinity chromatography, solid phase extraction, catalysis, sensory devices and capillary electrochromatography, all using various approaches for MIP synthesis and polymerisation (Haupt and Mosbach, 2000). To review the literature and understand trends in the use of MIPs for pesticide and organophosphate analysis, the search terms “molecularly imprinted polymer” combined with “pesticide” and “molecularly imprinted polymer” combined with “pesticide” and “organophosphate” were entered into the Web of Science online database to determine the number of available research publications from 1985 to 2017. Figure 1-4 illustrates the number of publications produced over this period that relate to MIPs and pesticides and then specifically for OPPs.

6000

5000

4000

3000

2000

No. of of No. PapersPublished 1000

0 1985-1989 1990-1994 1995-1999 2000-2004 2005-2009 2010-2017

MIP, pesticide and organophosphate MIP and pesticide MIP

Figure 1-4: The number of published papers for molecularly imprinted polymer present in the literature with different search terms in the last 33 years (from Web of Science at 21/09/2017). The earliest study of MIP technology to detect an organophosphate pesticide was published in 2001, when Jenkins et al. (2001) reported producing polymers as a recognition element that was used for a fiber optic probe to detect an organophosphate nerve agent (sarin and soman) in water samples. Since that time, over 40 publications have

Chapter 1

reported the use of MIPs for OP applications, either in terms of direct MIP development or for testing different parameters. Table 1-8 illustrates some of the OPP compounds that have been used as templates for MIP production.

Table 1-8: Chemical structures of OPPs used as templates for MIP production. Pesticide Structure Pesticide Structure Acephate Triazophos CH3

CH 3

Fenitrothion Paraoxon

Diazinon Parathion CH3

CH3

Diphenyl- phosphate Parathion- methyl

Monocrotophos Disulfoton

Chapter 1

Further, they are utilised in different applications including immunoassays, catalysis, affinity chromatography, sensory devices and SPE (Vasapollo et al., 2011; Mirsky et al., 2010). Furthermore, they have the capability of resisting aggressive conditions, such as high temperature, pressure, extreme pH, and organic solvents. Moreover, they are inexpensive to be synthesised and can be manufactured with relatively good reproducibility and stability. Nonetheless, this technology still has some challenges to be overcome such as template leaching, poor accessibility of the binding sites, low binding capacity and non-specific binding.

Non-imprinted polymers (NIPs) are prepared in the same way, but no template is present, which means they do not have template-specific cavities. However, they still possess the ability to adsorb template molecules and ions in a non-specific way (Pichon and Chapuis- Hugon, 2008). Furthermore, NIPs have shown potential for water and wastewater treatment and other contaminants, and have shown the ability to purify a large number of molecules simultaneously at a lower cost than MIPs, largely due to their non-specificity. However, one drawback of NIP’s concerns their competition with natural organic matter (Murray et al., 2011; Murray and Ormeci, 2012). In addition, MIPs are not easy to prepare in aqueous media, and they can adsorb equal or even greater amounts of template molecules from water in some cases (Deng et al., 2009; Meng et al., 2012). As a result, NIPs could perhaps be regarded as a purification process for small molecules in their own right (Murray and Ormeci, 2012). For example, when a NIP was compared with activated carbon to adsorb bisphenol A from both river and lake water, less interference was found (Xie et al., 2011).

Imprinting methods can be classified based on chemical interactions between the functional monomer and template, this include covalent imprinting, semi-covalent imprinting and non-covalent (Rosengren-Holmberg et al., 2009). Table 1-9 summarises the advantages and disadvantages of covalent and non-covalent imprinting methods. The terms covalent and non-covalent in this context refer to the type of interaction occurring during the polymerisation step and inside the binding sites between a template and functional monomers.

Chapter 1

Table 1-9: General advantages and disadvantages of non-covalent and covalent imprinting (Mirsky and Yatsimirsky, 2010). Covalent Non-covalent Choice of the template and functional Restricted to templates and Wide variety monomer functional monomers with appropriate functionalities. Preparation of the MIP Preparation of a covalently Easy and straightforward bound pre-polymerisation complex necessary. Polymerisation conditions Less critical; the pre- Critical; only noncovalent polymerisation complex is interaction govern the pre- more stable polymerisation complex Binding sites(affinity) Homogeneous Heterogeneous Selectivity Higher selectivity Lower selectivity Removal of the template after Cleavage of the monomer- Easy and mild condition polymerisation template linkage necessary Rebinding kinetics Slow fast

1.7.4.1 Polymerization procedures

MIP synthesis requires several reagents, and it is critical to choose the right ratio, combination of reagents and right polymerisation conditions in order to obtain an efficient MIP. This section will describe only the non-covalent approach, as this has been extensively used to produce MIP materials. It involves functional monomer, cross-linker, an imprinted molecule (template) and an organic solvent, so-called porogens. In this method, the nature of chemical interactions between templates and functional monomers are governed by forces such as electrostatic interaction include hydrogen bonding, (ion- pairing) and - stacking and/or weak van der Waals forces. In the beginning, functional monomers arrange around a template by functional groups in a self-assembly and the template–monomer network becomes cross-linked to each other after adding a cross- linker during polymerisation. A porogen and an initiator are added and then polymerisation takes place by either thermal or photochemical initiation to form an imprinted matrix. At the end of the process, the template needs removing by washing with suitable solvents capable of disrupting the bond. This can be carried out with uncomplicated methods using, for example, a soxhlet extractor or rinsing the material in a tube several times. Once the template is removed, there would be empty cavities left in the polymeric structure. As a result, the resulting polymer is a highly cross-linked, three- dimensional network polymer with imprinted cavities which are complementary in size, shape and chemical functionalities to the template whereby they are able to distinguish

Chapter 1

and rebind exclusively the target compound and other analogous that have similar structures (Mayes and Whitcombe 2005). Figure 1-5 shows a schematic diagram of the MIP preparation steps.

Figure 1-5: Schematic diagram of molecularly imprinted polymer synthesis to produce template-monomers network by non-covalent interaction.

To date, various methods have been proposed and employed to prepare MIPs. However, bulk polymerisation is a conventional method that is widely used due to its simplicity and low cost. Methods have been used to produce MIPs in spherical forms to detect OPPs included precipitation polymerisation (Xie et al., 2010), dispersion polymerisation (Wang et al., 2013), suspension polymerisation (Erdem et al., 2010), and surface initiation (Yang et al., 2009), emulsion, and two-step polymerisations etc. (Vasapollo et al., 2011; Pichon and Chapuis-Hugon, 2008), whereby greater control over particle sizes and porosity is possible.

Chapter 1

Bulk polymerisation This method requires only a single-step reaction which all reagents (i.e. template, monomer, cross-linker and initiator) are dissolved in a suitable porogen. The mixture is sometimes cooled, purged with nitrogen to remove oxygen from the media and then polymerised by applying heat above 60 °C or UV light at room temperature. Then, the resultant solid polymer is crushed, ground and sieved to produce particles of appropriate sizes. Usually, during the sieving, particles in a small range of sizes (from 5 to 150 µm) are collected depending on the desired application. However, this method has drawbacks, as it requires a long period of time, results in a significant amount of polymer loss and can also change the morphology of recognition sites. As a result, the resultant particles have irregular shapes that are not suitable for certain applications such as chromatography; however, they still can be used for SPE (Ye et al., 2002). Another issue that the MIP suffer from is a template leaching as a result of uncomplete extraction (Anfossi et al., 2009). Unfortunately, the template leaching can interfere with the results of binding assays or adsorption tests, especially when only trace analysis is required (Kang et al., 2012). Furthermore, the presence of non-specific binding sites that can appear in a MIP network can also be problematic and requires careful tuning of different parameters to improve selectivity (Yan et al., 2014). For this reason, various methods have been suggested to achieve more straightforward spherical MIP preparation to meet the requirements of different applications in which homogenous particles are preferred with uniform-shaped for both of micro- and nano- sizes (Ye and Mosbach, 2001).

Precipitation polymerisation Precipitation polymerisation is the second most commonly used technique for the preparation of MIP. The method requires reagents similar to those employed in bulk polymerisation; the only significant difference is that its reaction takes place in a highly diluted system by a porogen. This method produces mostly particles in nanometre sizes, which can be utilised in capillary electrochromatography (Say et al., 2003) and immunoassay (Ye et al., 1999). Furthermore, recently, a novel precipitation method for one-step preparation of monodisperse and spherical MIP beads has been developed to suit HPLC and SPE applications. Moreover, during the polymerisation process, particle precipitation occurs when polymer chains develop and achieve a critical mass that cannot

Chapter 1

be suspended. At this stage, particles precipitate and produce microspheres or aggregate particles with sizes ranging in the scale of nm to µm after optimisation (Geng et al., 2015). The advantage of this method is that it does not require surfactants or stabilizers, even though careful optimization of polymerisation conditions is required (Wang et al., 2003). Another matter concerns its template, which can affect particle sizes, and in some cases increase them. In a few cases, the template has no effect (Sambe et al., 2006; Wang et al., 2003). Also, there is no a standardised protocol for the precipitation polymerisation.

1.7.4.2 Templates

In the MIP synthesis, a selected template is used to interact with a functional monomer and form desired interactions. To some extent, the chemical properties of template can affect the imprinting stability of the monomer-template network, particularly, the uniformity and strength of resultant imprinted binding sites. For example, the nature and number of interaction sites between a monomer and a template and the configuration of the monomer-template structure can increase its strength, recognition capability and high selectivity (Tan, 2010; Sellergren, 1999). Moreover, it is essential for a template not to contain any functional groups, such as a thiol or a hydroquinone moiety, as these groups contain free electrons that can interact with free radicals and hinder the polymerisation process. In addition, any given template must be stable at high temperatures and resistant to the UV radiation exposure that are both commonly applied to initiate polymerisation (Cormack and Elorza, 2004). However, a MIP development is impractical for highly toxic compounds due to safety and contamination concerns and can present challenges in obtaining sufficient supplies (Baggiani et al., 2007). Furthermore, while the temperature normally used to begin the thermal polymerisation process is around 60°C or higher, the exothermic nature of polymerization can elevate this temperature. As a result, the reaction makes the initiation temperature not representative because the internal temperature can be much higher than the external temperature. Moreover, the pressure inside the polymer can also increase, which relates to temperature influences. These conditions can affect template imprinting and MIP performance. For example, capacity and separation factors were both evaluated at low and high polymerisation temperatures (80 °C and 25 °C) to produce two MIPs. The internal temperatures were elevated to 187 °C and 78 °C for the MIPs prepared at 80 °C and 25 °C, respectively. Obviously, an increase occurred in the

Chapter 1

two factor values as the polymerisation temperature decreased, which most likely can affect both templates and template-monomer complex stability (Piletska, et al., 2009). Different templates used for MIP synthesis and their application are presented in Table 1-10.

Table 1-10: MIP templates and their applications. Template Matrix LOD / Recovery (%) Application Reference Parathion methyl ACN 10-13 Optical sensor Tan et al. (2015) Parathion methyl Water 1 nM A potentiometric sensor Geng et al. (2015) Acephate Apple 1.14×10-13 Optical sensor Wei et al. (2011) Parathion methyl Solution 8.9×10-9 Electrochemical sensor Tan et al. (2010) Paraxon Water 1×10-9 Electrochemical sensor Alizadeh et al. (2010) cappage Parathion Vegetable 0.003 µg/mL Electrochemical sensor Yang et al. (2009) Parathion Buffer 0.005 µg/mL Electrochemical Marx et al. (2004) OPP Water ˂ 10 mg/L Optical sensor Jenkin et al. (2001) Trichlorfon and Monocrotophos Vegetable 0.011 & 0.015 mg/kg/ Solid phase extraction Li et al. (2017b) and water 86.5 to 91.7 Fenthion Olive oil 5 µg /L / 96 Solid phase extraction Bakas et al. (2014) Trichlorfon and monocrotophos Vegetable 4.2 µg &1.2 ng/g / Solid phase extraction Wang et al. (2014b) 88.5 - 94.2 4-(dimethoxyphosphorothioylamino)butanoic Fruit 0.3-1.6 µg/kg / 81- Solid phase dispersion Wang et al. (2013) acid) a 105 Trichlorfon and Monocrotophos fruits 0.011 & 0.015 mg/kg Solid phase extraction Xin, et al. (2013) / 86.5 to 91.7 Diethyl thiophosphate – DETP Urine 3 µg/L / - Solid phase extraction Santos et al. (2012) Tisulfoton Strawberry 0.02-0.05µg/g / 65.25 - Solid phase extraction Baldim et al. (2012) 87.70 Diazinon Water / 90 Solid phase extraction Rahiminejad et al. (2009) Trichlorfon Vegetable 92.5 µg/kg/ 80.2 - 95.8 Chromatographic column Zhao et al. (2014a) Diethyl(3methylureido)(phenyl) methylphosphonate - - Chromatographic column Kang et al. (2012) (DEP) b Triazophos vegetable 2.5 nM / 93 to 103 On-line enrichment Xie et al. (2010) Diazinon fruit 0.02 µg/L / 96.0–104.0 Dispersive Solid-Phase Bazmandegan-Shamili et Microextraction al. (2016) Diazinon - - Equilibrium binding Rahiminezhad et al. experiments (2010) Paraxon - - Catalysis Say et al. (2005) Paraxon - - Catalysis Erdem et al. (2010) a Generic template for (trichlorfon, malathion, acephate, methamidophos, omethoate, dimethoate, phosphamidon, monocrotophos, and parathion- methyl). b a dummy template. c acetylcholinesterase inhibitor (organophosphorus oxon metabolite).

1.7.4.3 Functional monomers

Functional monomers are responsible for the binding interactions in the imprinted sites by the functional groups. During the polymerisation, therefore, solider interactions between the template and the functional monomer are provided, leading to MIP with higher binding capacity and selectivity. For a general approach, templates containing basic groups, acid functional monomers are chosen and the other way around. Normally

Chapter 1

excessive number of moles of template is used to favour the formation of template, functional monomer assemblies. It is clearly very significant to match the functionality of the template with the functionality of the functional monomer in a complementary style in order to maximise complex formation and consequently the imprinting effect (Cormack & Elorza, 2004). Moreover, the selection of monomers is not easy to predict, and although many compounds are reported, only a small number is commonly used (Karim et al., 2005). Scientists by experience have developed notions to produce a combination of monomers and other regents which are expected to produce satisfactory results. The molar ratio of the template and the monomer is usually 1:4 and higher in non- covalent imprinting (Yan, 2004; Cormack & Elorza, 2004). Some frequently used monomers with different functionalities include acrylic acid (AA), methacrylic acid (MAA), vinylbenzoic acid, acrylamido-methylpropanesulphonic acid, vinylpyridine, vinylimidazole and acrylamide. MAA is the most commonly used functional monomer for MIP production and has been exploited for a wide spectrum of templates (Cameron et al., 2006). However, while MAA is widely used as an acidic functional monomer for many basic templates, other acidic monomers were also evaluated, as they exhibit various acidity levels showing higher efficiency for imprinting where they can establish ionic interactions with an imprinted template (Piletska et al., 2008). For example, MAA, AA and 2-acrylamido-2-methyl-1-propane sulfonic acid (AMPS) were evaluated, and AMPS (i.e. the most acidic monomer (pK, 1)) formed an ion-pair complex with templates in methanol. However, as weaker acidic monomers, MAA (pK, 54.65) and AA (pK, 54.2) were less effective, which was confirmed when the values of dissociation rates were calculated and showed the strongest complex results were achieved by the AMPS when compared with AA and MAA (Sergeyeva et al., 2001). In a different study, an alternative to MAA, 2-(trifluoromethyl) acrylic acid (TFMAA) with a strong acidity was reported for nicotine molcular imprinting, which was able to form stronger and firm complex showing higher affinity than MAA to nicotine (Matsui et al., 1997).

By far, MAA is the most commonly used functional monomer for MIP production and has been employed for a broad spectrum of templates (Alexander et al., 2006). The popularity of MAA lies in its capability to accept or donate a hydrogen bond. It has been reported that MAA is the most utilised functional monomer in the MIP synthesis for OPPs

Chapter 1

(Tan et al., 2015; Geng et al., 2015; Bakas et al., 2014; Rahiminejad et al., 2009) for parathion-methyl and diazinon. Furthermore, Bakas et al. (2014) used computer modelling to stimulate successful monomer-template interactions by screening 20 functional monomers before MAA was selected because of its highest binding energies among all monomers. Table 1-11 below lists monomers and other materials that have been used to develop MIPs for OPP detection.

Table 1-11: Templates and polymerisation reagents that used to develop MIPs. Template Solvent Functional Monomer Cross Linker Method References Acephate Acetonitrile MAA TRIM Surface-grafting by Wei et al. (2011) UV Diazinon Chloroform MAA EGDMA surface imprinting Bazmandegan-Shamili et al. (2016) Diazinon Chloroform MAA Polyethylene Facile and versatile Ma et al. (2012) alcohol bottom-up self- assembly strategy Diazinon acetonitrile MAA EGDMA Bulk Rahiminezhad et al. toluene 4vpy (2010) chloroform AA Disulfoton Toluene MAA EGDMA Bulk Baldim et al. (2012) DETP Acetonitrile 4-vinylpyridine EGDMA Bulk Santos et al. (2012) Diazinon Chloroform MAA EGDMA Bulk Rahiminejad et al. (2009) Diethyl(3-methylureido) Chloroform MAA EGDMA Precipitation Kang et al. (2012) (phenyl)methylphosphon AA ate Fenthion DMF AA EGDMA Bulk Bakas et al. (2014) Triazophos acetonitrile/ AA EGDMA Precipitation Xie et al. (2010) Toluene Trichlorfon and Chloroform MAA EGDMA, surface imprinting Li et al. (2017b) Monocrotophos Trichlorfon methanol/toluene MAA γ –MAPS Bulk Zhao et al. (2014b) Trichlorfon and Chloroform MAA EGDMA Bulk Wang, et al. (2014) Monocrotophos trichlorfon and Chloroform MAA EDGMA Bulk Xin et al. (2013) Monocrotophos Parathion-methyl Acetonitrile MAA EGDMA Bulk Tan et al. (2015) Parathion-methyl Methanol MAA Precipitation Geng et al. (2015) Parathion-methyl - vinyltriethoxysilane - electro-deposition Tan et al. (2010) by sol–gel Parathion polyethyleneimine EGDMA Surface imprinting Yang et al. (2009) Parathion Ethanol silanes containing NA Sol-gel process Marx et al. (2004) phenyl and primary amine groups Paraxon Toluene and MAH–Cu(II)a EGDMA Suspension Say et al. (2005) acetonitrile polymerization Paraxon Chloroform MAA EGDMA Bulk Alizadeh et al. (2010) Paraxon Acetonitrile MAH-M2+ EGDMA Suspension Erdem et al. (2010) polymerization 4(dimethoxyphosphoroth Acetonitrile AA EGDMA Bulk Wang et al. (2013) ioylamino ) butanoic acid 4 OPP - styrene divinyl Photopolymerisation Jenkin et al. (2001) benzene a methacryloyl–histidine–copper(II), FM functional monomer, XL, cross-linker b methacryloylhistidine-Co2+, -Ni2+, and -Zn2+ cacetylcholinesterase inhibitor (organophosphorus oxon metabolite)

Chapter 1

1.7.4.4 Cross-linkers

A cross-linker plays an important role and can shape the morphology of the polymer matrix and preserve its structural configuration after the template is removed. To ensure optimum efficiency and achieve structural specificity, a cross-linker should be applied in sufficient quantities, usually 70-90% of the resultant MIP (Mayes and Whitcombe, 2005). Although, a general formula at a molar ratio of 1:4:20 for the template: monomer: cross- linker respectively is used (Li et al., 2012a), the amount of the cross-linker used can be further optimized and produce MIPs with sufficient flexibility and speed the equilibrium process between the liberation and rebinding of the template (Jenkins et al., 2001). However, consideration for possible interactions between cross-linker and both of template and monomer, which might interrupt or block, need to be paid (Spivak, 2005). The cross-linkers include for example Ethylene glycol dimethylacrylate (EGDMA), divinylbenzene (DVB), trimethylolpropane trimethacrylate (TRIM).

Several cross-linkers appear in the literature, such as EGDMA, divinylbenzene (DVB), (TRIM) and polyethylene alcohol (PVA). EGDMA is the most commonly used cross- linker to develop MIP applications for OPPs (i.e. almost 70%). Other such as TRIM and DVB have also been employed in MIP production, cross-linking agents that used and found in the literature are listed in 1-12 above.

1.7.4.5 Solvents (porogens)

Organic solvents are employed in MIP synthesis procedures, and their role involves dissolving the template and other reagents during the pre-polymerisation process. In the polymerisation process, the solvent is responsible for both producing porous structures and enhancing the bonds between the template and the monomer(s). Solvents are responsible for the formation of pores and their distribution on the polymer surface. Moreover, solvents can influence the interaction between a template and the functional monomers, particularly in non-covalent methods. In addition, solvent polarity can decrease the interaction between monomers and a template and subsequently impact the formation of hydrogen bonds, thus reducing imprinting quality (Mirsky and Yatsimirsky, 2010). Table 1-12 illustrate some common solvents used for MIP production.

Chapter 1

Table 1-12: Common solvents that used for imprinting in non-covalent approach. Mw Boiling point Density Polarity Dielectric H-bond Solvent (g/mol) (°C) (g/mL) Index constant strength Acetonitrile 41.05 81.6 0.786 5.8 37.5 6.1 Chloroform 119.38 61.7 0.795 4.1 4.81 5.7 Dichloromethane 84.93 39.8 1.326 3.1 9.08 7.1 Tetrahudrofuran 72.11 66.0 0.886 4.0 7.6 8.0 Tolune 92.14 110.6 0.867 2.4 2.38 (25) 2.1 T=20°C

1.7.4.6 Initiator

The selection of initiators is based on the properties of a target molecule. Aazo-initiators are commonly used as free radical initiators, and azobisisobutyronitrile (AIBN) (Figure 1-6) is considered to be the most commonly used initiator used for polymerisation (Cameron et al., 2006). Because these initiators have electrical neutrality, they possess less selectivity, and more options are available in solvent selection, pH environment and polarities (Sellergren, 2000). Moreover, polymerisation at low temperatures results in better outcomes than a thermal approach achieved at elevated temperatures (Li et al., 2012a). For example, monomer-template network stability can be increased at low temperatures when interactions are based on electrostaticity (Lanza and Sellergren, 2004).

Figure 1-6: Chemical structure of AIBN

During the polymerisation process, an initiator forms a free radical species at the initiation step, and the free radicals then attacks the carbon double bonds (-C=C-) of the acrylate groups in the monomer and cross-linker. Moreover, during the propagation step, the reactive radicals again react with additional double bonds to form a new radical and grow to form a single bond with a longer chain. As the polymer grows and the monomer

Chapter 1

molecules are integrated to form the backbone, the solution becomes a gel and then a solid material (Figure 1-7). Finally, the polymerisation process is terminated when two growing oligomers chain react with each other (Tajau et al., 2014, February).

Figure 1-7: Radical polymerisation process

1.7.4.7 Rational synthesis of Molcular imprinting

The principle of imprinting suggests that an optimal ratio between monomer and template exists to provide a polymer with satisfactory selectivity. The reason for this relates to the ratio used during polymerisation, which governs the nature of the monomer/template assembly within the polymer and eventually the properties of the imprinted site. Some studies have suggested that even though higher ratios result in a smaller share of total amount that possesses selective binding sites, a higher percentage of these binding sites have high affinity. However, other studies have shown that when the concentration of a monomer is increased beyond a certain point and reaches saturation, polymer selectivity weakens and non-specific binding sites increase as a result of more monomer being incorporated into the polymer (Fish et al., 2005).

Chapter 1

Ideally, an optimal ratio is accomplished empirically by evaluating several polymers prepared with different formulations with increasing monomer amounts. The rationale behind this approach is believed to originate with the solution complex between functional monomers and templates. Based on the general formation mechanism of MIP binding sites, the formation of each individual binding site in a polymer is attributed to an individual template molecule surrounded by functional monomers in a pre-polymer complex. It has been assumed that each complex in the pre-polymer solution creates a binding site, and that increasing of the concentration of components or binding affinities of the complex in a pre-polymerization mixture could predict an increase in the pre- polymer complex. Correspondingly, an increase in the number of final binding sites occurs for the imprinted polymer, leading to increased binding or selectivity factors per gram of the polymer. The pre-polymer complex can be increased by either raising the amount of functional monomer or the amount of template, or both (Spivak, 2005). However, template concentrations can be significantly reduced in a pre-polymerisation mixture, showing no major losses in MIP quality or performance (Brüggemann et al., 2000). Moreover, the strength of the interactions formed in the monomer/template network determines the efficiency of its binding sites and their selectivity to a target molecule. In addition, it is worth mentioning that optimising the molar ratio of MIP components significantly affects polymer binding sites and selectivity.

1.7.4.7 Application of MIPs for sample preparation

1.7.4.7.1 Sorbents for solid-phase extraction

Many studies have reported the use of MISPE as selective sorbents. With respect to OPPs, some research has investigated and developed various polymers for use in MISPE applications, including sample such as water (Rahiminejad et al., 2009), strawberries (Baldim et al., 2012), fruit (Bazmandegan-Shamili et al. 2016; Wang et al., 2013), vegetables (Zhao et al., 2014) and olive oil (Bakas et al., 2014). Rahiminejad et al. (2009) synthesised a MIP using a thermos-polymerisation method for diazinon, which used to extract a trace amount of diazinon from a drinking water sample, and recovered greater than 90% of the compound of the spiked standard. In fact, the authors observed that diazinon recovery improved as they increased the pH levels in the sample. In another study, Xin et al. (2013) also applied an MISPE protocol, which was produced by thermo- 56

Chapter 1

polymerisation to detect two OPPs, trichlorfon and monocrotophos. The researchers took a different approach by utilising both analytes as mixed templates. Interestingly, the resultant MIP showed sufficient rebinding and adsorption capacities for these targets. The MISPE was coupled with GC to analyse pesticide residues in spiked rape samples and the detection limits for trichlorfon and monocrotophos were found to be 0.28 μg/L and 0.090 μg/L, respectively. To confirm their applicability in a different matrix, the same polymer was again successfully applied to quantify the same pesticide residues in leek samples.

Wang et al. (2014b) ran this experiment differently in order to produce MIPs and detect trichlorfon and monocrotophos pesticides. Instead of developing polymers using mixed templates, each MIP was prepared separately and the resultant MIPs then were mixed at a ratio of 20% and 80% respectively for preparing MISPE column. It was then coupled with a HPLC (MISPE-HPLC) for direct quantification of pesticide residues in spiked rape and cauliflower samples. The detection limits for trichlorfon and monocrotophos were found to be at 4.2 µg/g and 1.2 ng/g, respectively.

Bakas et al. (2014) prepared a MIP sorbent using a thermo-polymerisation method to extract fenthion from olive oil. In this study, the researchers modelled molecular recognition and used library screening to identify a suitable functional monomer. Acrylamide was identified, and dimethyl formamide as a porogen to produce the MIP. The MIP exhibited excellent selectivity towards fenthion when compared to other OPPs in olive oil samples, and a recovery of 96 % with a detection limit of 5 µg/L were reported. Baldim et al. (2012) synthesised another MIP to purify an extract of strawberry samples for five OPPs using disulfoton as a template. The thermos-polymerisation method coupled with GC was used to quantify: diazinon, disulfoton, parathion, chlorpyrifos and malathion, with detection limits that varied between 0.02 and 0.05 μg/g, with average recovery rates between 65.25 and 87.70%.

In different techniques, Bazmandegan-Shamili et al. (2016) adopted a one-step surface imprinting technique employing precipitation polymerisation to synthesise magnetic MIP (MMIP). The resultant MMIP was applied as dispersive solid-phase microextraction for purification and pre-concentration of diazinon from real samples. The magnetic cores were activated with Fe3O4 particles and used diazinon as a template, MAA and EGDMA

Chapter 1

as functional monomer and a cross-linker respectively. This technique was coupled with an HPLC analysis, and the average recovery percentages from real samples (tomato, cucumber, apple, and water) were found to be 96.0 –104.0 %, with a detection limit of 0.02 µg/L. In another study, Li et al. (2017b) synthesised novel MIPs, with mixed templates, utilising a metal-organic framework as a support material during the MIP preparation. The system was used for pre-concentration and detection of trichlorfon and monocrotophos in fruit samples and showed recovery rates ranging from 86.5 to 91.7 %, and detection limits of 0.011 and 0.015 mg/kg, for trichlorfon and monocrotophos respectively in apple and pear samples. In addition, the sensitivity of MISPE-HPLC was improved when compared with a traditional procedure. In addition, column chromatography is another application that has benefited from the MIPs. Although only one publication has appeared in the literature with the use for on-line enrichment. In this study, Xie et al. (2010) prepared a MIP-column for triazophos (TAP) using precipitation polymerisation to produce spherical particles. This method was integrated with a chemiluminescence (CL) assay system as a detection method. The MIP selectively designed to capture TAP improved significantly the detection. The researchers produced a sensitive method with a detection limit of 2.5 nM, well below the results of a traditional CL assay. As a result, the researchers suggested that this approach could be used for online enrichment and the detection of pesticide residues in vegetables.

1.8 Target pesticides

The target pesticides for this study were chlorpyrifos, parathion and chlorpropham.

1.8.1 Chlorpyrifos

Chlorpyrifos (CPF) [O,O-diethyl O-(3,5,6-trichloro-2-pyridyl) phosphorothioate] belongs to the family of organophosphate pesticides. CPF is an efficient broad-spectrum organophosphorus agricultural insecticide acaracide and miticide, having activity towards a wide range of insects and domestic pests. CPF works by inhibiting the enzymes of the nervous system related to cholinesterase activity in insects but also has the same effects in animal and human biological systems. Moreover, CPF has been used in agricultural, industrial, and residential settings which has led to its intentional and accidental entry into terrestrial and aquatic ecosystems. Subsequently, it poses a possible risk to animal and

Chapter 1

human health concerning chromosomal damage and bladder cancer (Ma and Chen, 2014). In addition, CPF uses have been banned in household settings in some countries such as the US (Kitada et al., 2008).

It is widely utilised globally to control insect and anthropoid agricultural pests found in soil, grazing lands, food and feed crops, to protect livestock (cattle ear tags) and to combat public health pests such as mosquitoes and fire ants (NPIC, 2017). For individual plant uses, it can be applied to alfalfa, tobacco, vegetables such as Brassicas (broccoli, cauliflower, and cabbage), vines, asparagus, citrus fruits; (Oranges, Lemons and Limes, Grapefruit, Nectarines), and tangerines, corn, cotton, grapes, mint, onion, peanuts, pome and stone fruits, cherries, peaches, nectarines, plums, prunes, and apples, strawberries, bananas, figs, soybeans, sugar beets, rice, sunflower, sweet corn, sweet potatoes, tree nuts; almonds, hazelnuts, pecans, and walnuts and wheat, cereal maize, sorghum and mushrooms (Gomez, 2009). Based on pesticide consumption patterns for targets, CPF is used on corn, (39%); alfalfa, (6%); cotton, (3%); sorghum, (1%); citrus and deciduous plants, fruits/nuts, (21%) and non-agricultural uses (31%) (Spectrum Laboratories, 2014). It is moderately persistent in nature and can be detected in soil samples after two months. Considerable amounts of CPF residues have been reported in apple, grape and orange fruit for various market-based surveys (Latif et al., 2011). In 2015, an EFSA report stated that an analysis of over 81,000 samples of food products in the European Union showed that almost 50% contained CPF residues, including strawberries in particular which exceeded MRL by 2.5% for the analysed samples (EFSA, 2015).

CPF accounts for a major group of insecticides utilised in agricultural and domestic settings. Global usage of CPF is greater than 50,000 kg/year and amounts to 5 million kg/year in Europe and the United States, respectively. Consequently, frequent incidences of elevated CPF residues in food samples and ecosystems have been reported (Alavanja et al., 2013). Owing to CPF usage in the agriculture sector (soil, water and food) and point-source discharges, its use is responsible for the toxicity in many aquatic ecosystems, and it has been detected in livestock via contaminated water (Rodríguez et al., 2013). As CPF is a common, persistent and toxic insecticide, the European Union strictly regulates its use and frequently monitors its maximum residue levels (MRLs), as stated in European Commission (EU) regulation 2016/60 (Brancato et al., 2017).

Chapter 1

1.8.1.1 Toxicity

CPF works by irreversibly inhibiting acetylcholinesterase, an enzyme produced by the nervous system, thereby interfering with signalling processes of the neurotransmitter acetylcholine to prevent its breakdown, thereby causing its accumulation at cholinergic synapses. This build-up of acetylcholine depolarizes postsynaptic cells and causes them to become resistant to subsequent acetylcholine release, causing, among other effects, neuromuscular paralysis and can eventually result in death. This effect does not only occur in insects, but also in animal and human biological systems (Ma and Chen, 2013). Moreover, CPF is widely used in the agricultural, industrial, and residential settings, which has led to its entry into aquatic ecosystems, particularly those that contain fish (Solomon et al., 2014). Subsequently, it poses a possible risk to mammalian health, as it can cause chromosomal damage and bladder cancer (Ma and Chen, 2013).

In addition, acute CPF toxicity may impact central nervous, cardiovascular, and respiratory systems and can irritate the skin and eyes. Acute exposure symptoms include numbness, tingling sensations, loss of coordination, headaches, dizziness, tremors, nausea, abdominal cramps, sweating, blurred vision, difficulty breathing, respiratory depression, and slowed heartbeat. Prolonged exposure to CPF may result in the delayed reported symptoms such as impaired memory and concentration, disorientation, severe depression, irritability, confusion, headaches, speech difficulties, delayed reaction times, nightmares, sleepwalking, drowsiness and insomnia. An influenza-like condition with headache, nausea, weakness, loss of appetite, and malaise has also been reported as a result of exposure to CPF (Hod, 2011).

The World Health Organisation (WHO) has designated CPF as a Class II pesticide, which includes those exhibiting moderate toxicity. However, based on the International Agency for Research on Cancer (IARC) report, there is no clear evidence that CPF can be classified as a carcinogen, and the US EPA placed it in a Group E, indicating there is no data supporting CPF carcinogenicity in humans (Dikshith, 2016).

1.8.1.2 Exposure

Many studies have indicated that exposure to CPF occurs through consumption of contaminated food, inhalation and dermal exposure during agricultural activities such as

Chapter 1

pesticide mixing and application. However, CPF absorption through the skin is not common, unless it is absorbed through cuts on the body, which may lead to systemic toxicity. With respect to human health, the effects of CPF exposure depend on quantity, period and exposure frequency as well as the health of an individual (Rathod and Garg, 2017). It has been reported that about 70% of CPF exposure occurs through oral consumption, while less than 3% is absorbed through the skin. CPF is readily bio- transformed in the liver and normally passed on within 24 hours, mainly through urination (Gilani et al., 2016).

Several studies have been conducted to investigate CPF exposure. For example, a US study found a high level of CPF in a group of children aged 6 and below who lived in a non-agricultural region; the primary cause was the consumption of produce containing CPF residues. Another cohort study evaluated the exposure to CPF in preschool age children by comparing consumption of conventional versus organic food over three days. This study concluded that children who ate conventional food showed significantly higher levels of CPF (Curl et al., 2003).

CPF is highly efficient on targets in three different ways: contact, ingestion, and through vapour actions. It is a colourless to white crystalline solid, is only slightly soluble in water, and is soluble in most organic solvents. The physical and chemical properties of CPF are presented in Table 1-13.

Table 1-13: Physical and chemical properties of CPF, Modified from (NLM, 2017a). Common name Chlorpyrifos IUPAC 0, 0-diethy 0-3, 5, 6- trichloro-2pyridyl phosphorus thioate Chemical formula C9H11Cl3NO3PS

Chemical structure

CH3

CAS number 2921-88-2 Molecular weight 350.62 g mol-1 Physical state and color Crystalline solid, colorless to white Melting point 42 °C Boling point 160°C Density 1.40 g/cm3 Water solubility 0.0014 g/L (1.4 mg/L) at 25 °C Half life 3.5 and 20 days (in water) Trade name Dursban, Lorsban

Chapter 1

The U.S. Environmental Protection Agency (EPA) has reported that more than 400 different commercially available products contain CPF as an active ingredient in a variety of formulations and delivery systems (USEPA, 2002). CPF is marketed under different formulations, with the most common products including emulsifiable concentrates (EC), granular (GR) wettable powders (WP), water-dispersible granules, micro-encapsulated suspensions, and gel-based products (Eaton et al., 2008). Farmers apply CPF via airborne or direct application to soil, fruits, and vegetables (Gilani et al., 2016).

1.8.1.3 Regulations

In the European Union, CPF is reviewed by the European Food Safety Authority (EFSA) according to EU MRLs. A 2015 EFSA ruling based on a very conservative risk assessment indicated that the existing MRLs for various commodities raised concerns for consumer health. Therefore, new MRLs were adopted in Annex II and Part B of Annex III of EC Regulation No 396/2005. In January 2016, the EC produced a new regulation, 2016/60, that concerning the reduction of MRLs for many raw foods; this regulation came into force on 10 August 2016 (Commission regulations, 2016). Moreover, because CPF has been shown to have detrimental effects on human health, especially for children, the US EPA banned its household usage in any form in 2001. In addition, the EPA’s Health Risk Assessment of CPF for Registration Review stated that estimated exposure of CPF residues from food commodities violates the safety limits determined under the US Food, Drug and Cosmetic Act (Ruiz, 2017; Kitada et al., 2008). However, in early 2017, the EPA rejected a petition filed by the Natural Resources Defense Council (NRDC) and the Pesticide Action Network of North America (PANNA) to ban all CPF uses on food commodities on the basis of insufficient evidence associating the pesticide with brain damage from prenatal exposure and toxic drift (Ruiz, 2017).

Globally, MRLs for many plants, including fruits and vegetables, are established by the Codex Alimentarius Commission, the EFSA, the US EPA and other authorities. For strawberries, the Codex Alimentarius and EFSA set MRL set levels of 0.3 mg/kg and 0.2 mg/kg for CPF, respectively (Brancato et al., 2017). In Europe, CPF and other pesticides are monitored using an early detection system called the European Rapid Alert System for Food and Feed (RASFF). According to the 2015 RASFF annual report, CPF was

Chapter 1

found to be one of the most detected pesticide residues reported in 2013, 2014 and 2015 (RASFF, 2017a). In addition, RASFF identified food products containing CPF, and 282 different notifications were issued, mainly for fruits and vegetable products. For example, peppers were the most reported plants, followed by leafy greens, mint and beans. Strawberries were also among other fruits for which CPF was detected (RASFF, 2017b). This suggests that site screening methods for CPF could help prevent contaminated products from entering the food supply chain.

1.8.2 Parathion

Parathion (PT) is another pesticide that is widely used in agriculture mainely as insecticdes. It has various applications that control insect populations for a number of crops. As other OPPs, it is an acetylcholinesterase inhibitor and is extremely toxic, as exposure to PT can result in respiratory pain and muscular difficulties. As a result of this extreme toxicity, it is banned from agricultural use in the US, the European Union and many other countries. However, PT is still produced in China and is also used in less- developed countries, and PT residues have been detected in different commodities (Funari et al., 2013).

PT (phosphorothioic acid, O,O-diethyl O-(4-nitrophenyl)) ester was introduced in 1947. It is a broad spectrum, non-systemic, highly effective organophosphate insecticide that is used as an acaricide with contact and fumigant actions that can produce adverse stomach and respiratory effects in humans. It is widely used in many countries to curb a wide range of pests, including sucking and chewing insects, aphids, mites, beetles, lepidoptera, leaf hoppers, leaf miners, and other insects normally located on cereal crops, fruits, vegetables, vines, forage crops, cotton and ornamentals. It also is used to treat foliage pre- harvest and controls several soil insects, such as wireworms, rootworms, symphilids and nematodes for beets and ornamental plants. Prior to its ban in the US in 1991, it was restricted to alfalfa, barley, rapeseed, corn, cotton, sorghum, soybean, sunﬂower and wheat crops. In 2000, it was also recommended for voluntary cancellation by its manufacturers (IARC, 2017a). Moreover, its usage patterns show that it was mainly used on wheat (49%), deciduous fruits and nuts (14%), vegetables (14%), sorghum (11%), cotton (4%), citrus (1%), and barley and oats (7%) (Spectrum Laboratories, 2016).

Chapter 1

Historically, available data on PT indicated that its estimated production in the US in the 1970s was about 6000 tonnes per year. Between 1980 and 1981, its production was 1.2 and 2000–5000 tonnes in India and Western Europe, respectively. In 2004, its production volume reached more than 1000 tonnes/year, which qualified it for the list of the highest chemical production volumes. In addition, PT is produced globally by a number of companies, mainly in China (IARC, 2017a).

Since PT production began in the middle of the last century, it was one of the major pesticides used internationally. Later, its use declined as it was banned in many countries due to its high mammalian toxicity and selective toxicity for certain enzymes. PT is very chemically stable, persists in the environment for prolonged periods and can ultimately enter the food production chain (Sai et al., 2017). This has elevated concern that PT is detrimental to both human health and aquatic ecosystems. Today, PT is banned from use in the EU, US, UK and many other countries. However, many countries, particularly those in the developing world, still use PT for agricultural purposes (Mehta et al., 2016). Therefore, it remains necessary to monitor PT residues in food and ecosystems.

1.8.2.1 Toxicity

PT is an acetylcholinesterase enzyme inhibitor and an extremely toxic pesticide. It has been included in the Acute Toxicity Category 1 due to its acute eye, skin, and inhalation effects (USEPA, 2000). It is a potent plasma, red blood cell and brain acetyl cholinesterase inhibitor and is also toxic by all routes of exposure. Cholinesterase inhibition occurs following acute, subchronic and chronic exposures to low doses of PT (USEPA, 2000), and can lead to symptoms such as convulsions, poor vision, tremor, dyspnea, lung oedema and respiratory effects. Due to its extreme toxicity, it was classified as a Restricted Use Pesticide (RUP) by the US Environmental Protection Agency (EPA) and can only be purchased and applied by certified applicators. Furthermore, PT exposure has been linked to an increased risk of cancer and malignant transformation of some cells by causing genomic instability. The nature of these symptoms relate to the degree of exposure (Sai et al., 2017). PT is classified in Group 2B by the International Agency for Research on Cancer (IARC) as a possible carcinogen and

Chapter 1

in class Ia by the World Health Organization (WHO) as extremely hazardous (IARC, 2017b; WHO, 2010).

1.8.2.2 Exposure

Before PT was banned in many countries, both workers and the general public were exposed to its toxic effects. Occupational exposure to pesticides mainly occurs through dermal contact, and several reports have indicated PT poisoning among workers. In addition, the presence of mist applications on agricultural sites increase the risk of inhalation by workers or those who live in a close proximity to these sites. Furthermore, the magnitude of exposure to PT varies greatly depending on application and environmental conditions. In contrast, people are exposed to PT mainly through the consumption of contaminated foodstuffs, drinking water and living near farms where the pesticide may be applied (USEPA, 2017a).

Furthermore, retrospectively, in the US in 1977 and 1982 surveys were conducted on patients hospitalised following pesticide poisoning, which found annual numbers of unintentional pesticide poisoning and occupational cases to be approximately 2380 and 814, respectively. Among all the different pesticides used by farmers and pesticide applicators, the OPP class accounted for about 43%. In addition, out of the 10 pesticides that were identified, PT was at the top of the list for the highest number of hospitalized pesticide poisoning cases (Krieger, 2010). In a different survey conducted between 1991and1992 in potato fields in Ecuador, 33 cases of pesticide poisoning were reported, with two cases reported specifically for PT exposure (Cole et al., 2000).

1.8.2.3 Physical and chemical characteristics

The physical and chemical characteristics of PT are presented in Table 1-14.

Chapter 1

Table 1-14: Physical and chemical properties of PT, Modified from (NLM, 2017b). Common name Parathion, Parathion-ethyl; Ethyl Parathion IUPAC O,O-diethyl O-(4-nitrophenyl) phosphorothioate Chemical formula C10H14NO5PS Chemical structure

CAS number 56-38-2 Molecular weight 291.261g mol-1 Physical state and color yellowish liquid Melting point 6.1°C Boling point 375°C Density 1.26 g/cm3 Water solubility 12.4 mg/l at 25°C Half life 1-10 days Trade name Aileron, Aphamite, Bladan, Folidol, Fosferno, Niram, Paraphos and Rhodiatox

PT used to be applied by repeated spraying by fans or aircrafts on a wide variety of orchards, rows, and field crops (DHHS, 2017). PT formulations were produced and contained dilute sprays prepared from 15% or 25% wettable powders; 2–8% emulsifiable concentrates; 0.5–2% dusts; 10% granules or 10% aerosol (IARC, 2017a).

1.8.2.4 Regulation

Due to the concerns associated with PT for both wildlife and public health, it has been either banned or phased out by many countries, including Australia, Belize, Bulgaria, Hungary, India, Ireland, Japan, Malaysia, New Zealand, Philippines, Portugal, Sweden, Turkey, United Kingdom, and the US (IARC, 2017a). In the European Union, PT was placed in the Council Directive 91/414/EEC of 15 July 1991 relating to the distribution of plant protection products on the market. Later, PT was removed from Annex I based on commission decision 2001/520/EC of 9 July 2001, and all authorised products that contained PT for plant protection were withdrawn by 2002. Previously, all PT formulations except capsule suspensions were included in the Prior Informed Consent (PIC) Procedure of the Rotterdam Convention on the International Trade of Hazardous

Chapter 1

Chemicals, where certain severely hazardous pesticides were listed in Annex III (Commission Decision, 2005; FAO/UNEP, 2005). The European Union has set the maximum limit of PT residues in fruits, vegetables and other food products between 20 and 50 μg/kg (Della Ventura et al., 2017). In the US, PT applications were restricted in 1991 to mitigate risk for workers; this restriction covered aerial application (i.e. crop dusting) of emulsifiable concentrates for the nine crops stated above, and the termination of all its uses occurred in 2003 (USEPA, 2000).

PT residues have still been detected by the European Rapid Alert System for Food and Feed (RASFF), which notifies the EU members of food and feed safety alerts. Between 2003 and 2014, alerts were raised for contaminated fruit and vegetable samples, including sweet peppers from the Netherlands, oranges from Italy and pears from the United Kingdom (RASFF, 2017c).

1.8.2.5 Analysis

Conventional detection methods are normally used for PT residue analysis, including liquid chromatography and gas chromatography methods coupled with various detectors. PT is detected in multi-residue approaches for its identification with other organophosphate pesticides using GC-MS and HPLC-MS/MS methods (Chen et al., 2017b). Moreover, other detection methods are also used, such as enzyme immunoassay and electrochemical and optical biosensors using various bio-recognition elements (Sai et al., 2017).

1.8.3 Chlorpropham

Chlorpropham (CIPC) (Isopropyl N-(3-chlorphenyl) carbamate) belongs to the carbamate pesticide family (Figure 1-8). It is a selective, systemic pesticide that is used as a herbicide, plant growth regulator and sprout inhibitor (Smith and Bucher, 2012). Moreover, CIPC is a well-known effective pesticide and is used to control weeds as a pre- emergence and post-emergence herbicide for some vegetables and flowers. It also inhibits potato sprouting in storage following harvest and loading (Cunnington, 2008). CPIC was originally registered to control weeds as a pre- and post-emergence herbicide for some vegetables, forbs and other non-food crops. In addition, it is used to inhibit potato sprouting during storage following a harvest (Cunnington, 2008). However, while CIPC

Chapter 1

has a low toxicity profile, concern about its use stems from its metabolite, 3-Chloroanline (3-CA), which possesses a similar structure to 4-Chloroaniline (4-CA) a well-known carcinogen (USEPA, 1996). Normally, after stored potatoes are sprayed with CIPC, residues remain in the form of CIPC and its main degradation compound (3-CA) in or on the tubers, mainly the skins. As a result, regulatory authorities have become increasingly concerned about the safety of CIPC use for potatoes. This concern is primarily based on the formation of 3-CA, which is regarded as a more toxic pollutant but also because it has structural similarities to 2-chloroaniline (2-CA) and 4-CA. These aniline derivatives are recognised as dangerous materials and possible carcinogens to humans (Paul et al., 2016). As a result, the issues surrounding CIPC have pressured regulatory bodies, such as European Food Safety Authority (EFSA), to minimize its maximum allowed residue levels in potato crops. Furthermore, these issues concern potato-producing countries. For example, in the US, maximum CIPC residue levels were reduced from 50 ppm to 30 ppm for fresh potatoes, and these levels are restricted between 5 to 10 ppm for European Union member states; other countries such as Canada and Australia have also conducted CIPC residue reassessments (Kleinkopf et al., 2003).

CH3

Figure 1-8: Structure of Chlorpropham

When applied for potato storage, CIPC residues and the main degradation compound, 3- CA can be detected in or on the skin of potato tubers. There are now increasing concerns from regulatory authorities on the use of this pesticide due to the formation of the 3-CA which is regarded as a more toxic pollutant but more importantly from the fact that it has similarities in structure to 2-CA and 4-CA. These aniline derivatives are recognised as hazardous compounds and possible carcinogens to humans (Paul et al., 2016).

Chapter 1

1.9 Aims and objectives

1.9.1 Aims

Currently, there are multiple approaches available for pesticide determination most of which are laboratory based either due to the high solvent sample extraction and concentration steps required or to the expensive non portable analytical equipment used for analysis. To date no single approach is utilized or recommended for pesticide analysis in the field due to the chemical diversities of the numerous target pesticide compounds causing complexities in rapid sample extraction, the high sensitivity that is required and the lack of user friendly cost effective tools available. The aim of this thesis was to examine different bioanalytical approaches as model tools for the rapid detection of pesticides.

The aim was to be achieved through three key objectives. The first objective was the synthesis and optimisation of MIPs to be utilized as SPE, affinity materials, to purify environmental samples for CPF and CIPC as a method for sample preparation and concentration for implementation into a rapid detection technique. The MIP synthesis involved bulk polymerisation whereby the assessment of various parameters was evaluated including different porogens, evaluation of the sample in organic and aqueous solution, the optimization for loading of the target and washing for recovery and the determination of the selectivity of the MIPs for the target compounds.

The second objective was to show the feasibility of a planar waveguide microarray biosensor system for individual pesticide analysis using outsourced reagents for CPF and PT. This involved the affinity characterisation of antibodies relative to the antigen, the assay construction in the optimised selection of concentrations of immunological reagents and the determination of the assay sensitivity and specificity for each target compound in buffer. To complete the process the development of a sample preparation method was required for the target compounds in strawberries as a model fruit suitable for the MBio planar waveguide detection system to illustrate the required specificity, sensitivity, time and cost-effectiveness to be fit for purpose relative to set MRLs.

Chapter 1

The third objective was to integrate and combine the individual assays to determine the feasibility for the simultaneous detection of multiple pesticides using the microarray technology illustrating the required specificity, sensitivity, time and cost-effectiveness to be fit for purpose.

Chapter 2

Chapter 2: Development and evaluation of molecularly imprinted polymers to concentrate CPF and CIPC in water samples

Chapter 2

2.1 Abstract

A comparative study of four different porogens and their effect on the affinity and selectivity of MIPs synthesised by thermo-polymerisation methods was conducted. CPF and CIPC were selected as the templates, MAA was chosen as the functional monomer, and acetonitrile, dichloromethane, chloroform and toluene were employed as the porogens for MIP synthesis. The subsequent evaluation of the results was conducted by using them as SPE sorbents with organic and aqueous solutions as analyte loading carriers. For each analysis, the analyte retention results of each experiment were evaluated using HPLC relative to a reference concentration (100 ng/mL). When the organic solvents were used as loading carriers, the MIPs exhibited weak retention characteristics for both CIPC and CPF. By contrast, the MIPs showed high affinities to these compounds when aqueous solutions were used as a loading medium. A rapid optimisation approach for washing protocols was used, which employed several mixtures containing organic solvents and water, and also different amounts of MIP polymer packed in a SPE tube was then carried out. Under the optimised conditions, the MIPs showed high binding capacities when tested in an aqueous solution for adsorption that were found to be higher than 16.7 mg/g polymer. Overall, the MIP prepared in acetonitrile for CPF showed the most promising results when tested in an aqueous solution. The selectivity of the CPF MIP prepared in ACN was evaluated against other pesticides, such as CIPC, TCP, 3-CA and IPC, and showed satisfactory selectivity characteristics compared to other compounds. Thus, the MIP for CPF prepared in ACN could be considered as a promising candidate for further study for application in real samples.

Chapter 2

2.2 Introduction

2.2.1 MIP applications for sample preparation

Although SPE is a robust method normally used in many applications for clean-up and pre-concentration of analytes, it is often not selective and can lead to interference in matrix components. Thus, specificity, selectivity and sensitivity, together with high extraction efficiency, can be obtained using sorbents based on MIPs (Núñez et al., 2012). Similar to the selectivity provided by immunoaffinity sorbents, MIP is used for sample matrix clean-up, enrichment and selective extraction of analytes in a complex matrix (Lucci et al., 2012).

2.2.2 General extraction procedure

MISPE is based in the same principles as traditional SPE methods, and is based on four main stages, as shown in Figure 2-1. The method works primarily by partitioning compounds between sample matrices and sorbents (Poole, 2003), and the process entails conditioning or equilibrium, loading, washing and elution. The ideal MISPE procedure might involve pH adjustment, dilution, and filtration. The first stage of the MISPE process entails conditioning or wetting by a solvent, which involves passing it into a SPE cartridge to activate functional groups on the polymer. The sorbent and sample matrix are then treated by matching polarity and using the same solvent. The second stage involves loading a sample of a known volume, depending on the aim of the experiment; this loading can be accomplished using different strategies. It includes binding and elution, interference removal and fractionation methods. The third stage involves washing, in which an appropriate solvent or mixture of solvents is used to remove unwanted interferences, leaving the analytes intact. The final stage is elution, in which the desirable analytes are removed from the sorbent using a suitable solvent; this normally involves a more powerful solvent than the one used in the washing stage, as it should be able to disrupt the primary and secondary retention interactions between the sorbent and the analytes (Liška, 2000).

Chapter 2

Figure 2-1: Principles of solid phase extraction (Lucci, et al., 2012).

MIPs have been synthesised by different methods using various pesticide templates and were applied as SPE for sample preparation and other such as affinity sensors. Many targets of OPPs have been used as templates to develop MIPs. Usually for pesticide analysis, sample pre-concentration and enrichment are critical in order for them to be detectable, particularly for trace analysis. In addition, various interference that may be present in samples can complicate the effective detection and quantification of target compounds. However, this challenge can be managed by applying separation techniques.

In the field of pesticide analysis, MIPs have been synthesised to detect various pesticide classes, including OPP and atrazine (Siemann et al., 1996) and carbamate (Mena et al., 2002). These MIPs have been utilised in SPE and sensing elements for affinity sensors. In addition, MISPE has been used to purify specific target analytes in complex matrices for identification and quantification purposes. For the carbamate class, MIP-related studies have been published for methyl-carbamate pesticides (carbaryl), pirimicarb and

Chapter 2

metocarb. Moreover, in an example of a separation application, Qian et al. (2010) used MIPs, prepared by bulk polymerisation to extract metolcarb traces from plant matrices (i.e. cabbage, cucumber and pear). Further, another MISPE protocol was developed using bulk polymerisation methods to extract pirimicarb from water samples (Mena et al., 2002). Finally, pirimicarb-MIP was prepared as bulk in a micropipette tip in order to extract it from tomato tissue (Zhou et al., 2010).

With respect to the OPPs class, work has been conducted to develop MIPs for a number of pesticides. In the following examples, MIPs were prepared to detect either individual or groups of pesticides in some cases, they were prepared for a pesticide and its metabolite (Bakas et al., 2013). Examples of OP-imprinted pesticides include fenitrothion (De Barros et al., 2010), acephate (Wei et al., 2011), chlorpyrifos (Liu et al., 2010), malathion (Chang et al., 2011), quinalphos (Sanagi et al., 2011), trichlorfon (Zhao et al., 2014), methamidophos (Xiao-hong et al., 2009), parathion-methyl (Yang et al., 2009), monocrotophos (Yan et al., 2007) and triazophos (Xie et al., 2010). In addition, MIPs have been synthesised to simultaneously detect the presence of some OPPs. For example, Zhu et al. (2005) developed a MIP that can bind monocrotophos, mevinphos, phosphamidon and omethoate that was based on the pesticides’ polarity in water and soil samples. In addition, Wang et al. (2013) developed a MIP to determine traces of nine organophosphorus pesticide residues (i.e. trichlorfon, malathion, acephate, methamidophos, omethoate, dimethoate, phosphamidon, monocrotophos, and methyl parathion) in fruit samples. Moreover, MIPs have been produced using bulk polymerisation method for malathion in order to purify four OPPs from strawberry samples with detection limits in the 0.02 to 0.05 μg/g range (Baldim et al., 2012). In addition, Liu et al. produced MIP microspheres (MIPMs) for CPF via emulsifier-free polymerisation through a single-step swelling method using polystyrene microspheres as seeds (Liu et al., 2010). The MIPMs were then tested in contaminated water samples for CPF removal. In another example, MIPs were developed to detect quinalphos using bulk polymerisation and were applied as a sample enrichment for a number of OPP residues, including CPF in fruit samples (i.e. apple and orange) (Sanagi et al., 2013).

MIPs are used as selective sorbents for molecularly imprinted solid phase extraction (MISPE). This is considered an important field in the area of analytical separation

Chapter 2

sciences among other applications (Lasáková and Jandera, 2009). Although SPE is a robust method normally used in many applications for clean-up and pre-concentration of analytes, it is often not selective and can lead to interference in matrix components. Thus, specificity, selectivity and sensitivity, together with high extraction efficiency, can be obtained using sorbents based on MIPs (Núñez et al., 2012). Similar to the selectivity provided by immunoaffinity sorbents, MIP is used for sample matrix clean-up, enrichment and selective extraction of analytes in a complex matrix (Lucci et al., 2012).

2.2.3 The chapter aim

The aim of this study was to synthesise MIPs specific to CPF and CIPC and develop novel solid phase extraction protocols. For this, four different solvents were tested as suitable porogens for the synthesis of MIPs, and several optimizations of the SPE protocols were conducted to identify the best MIP compositions and produce MIP sorbents for SPE suitable for a sample purification

Chapter 2

2.3 Materials and methods

2.3.1 Reagents and Chemicals

The organic standards chlorpyrifos (CPF), chlorpropham (CIPC), propham (IPC), 3- chloroaniline (3-CA) and 3,5,6-trichloro-2-pyridinol (TCP) were used to prepare stocks and \ standard solutions. In addition, MAA, TRIM, EGDMA, AIBN, silicon oil, acetonitrile (ACN), methanol (MeOH), toluene (TOL), dichloromethane (DCM), chloroform (CHL), methanol and acetic acid were purchased from Sigma-Aldrich (UK). Deionized water (DH2O) was prepared in the laboratory. All chemicals and solvents used were HPLC-analytical grade.

2.3.2 HPLC analysis

Analysis of CIPC was performed using an Agilent 1200 series HPLC binary pump system (Agilent, Berks., UK) equipped with an Agilent photodiode array with multiple wavelength detectors; CIPC was detected at 236 nm. A calibration standard curve for CIPC was prepared in acetonitrile with a concentration range of 10-1000 ng/mL, and 20μL was injected automatically into a Gemini C18 (150 x 4.6 mm diameter, 5 μm particle size) column (Phenomenex, UK). The mobile phases were (35%, 35%, 30%, v;v;v) for analytical grade acetonitrile, methanol and HPLC grade water, respectively (Camire et al., 1995). The flow rate of the mobile phase was 1 mL/min under isocratic conditions, the column temperature was set at 35°C and the temperature of the samples was held at 4°C using an Agilent G1330B cooled autosampler. The run time for the CIPC was 8 minutes, followed by 3 minutes of reconditioning. This analysis procedure was also used for quantification of IPC and 3-CA. For CPF, the same HPLC analytical conditions were used, and CPF was detected at 230 nm. A calibration standard curve of CPF was prepared in acetonitrile with a concentration range of 10-1000 ng/mL, and the mobile phases were (85%;15%, v;v) for analytical grade methanol and HPLC grade water, respectively (Lu et al., 2013). The flow rate of the mobile phase was 1 mL/min under isocratic conditions, the column temperature was set at 42°C and the temperature of the samples held at 4°C using an Agilent G1330B cooled autosampler. The run time for CPF was 11 minutes, followed by 3 minutes of reconditioning. For TCP, a mobile phase of

Chapter 2

acetonitrile/water (75:25, v/v) with 0.0025% formic acid was used, and the flow rate was 0.4 mL/min, which was adapted from a similar work (Williamson et al., 2006).

2.3.3 Preparation of stock and working solutions

Stock solutions of CPF and CIPC (1mg/mL) were prepared by dissolving dried materials in four different organic solvents (ACN, CHL, DCM and TOL). These working standard solutions were made by diluting the stock solution in ACN to produce different standard concentrations for calibration curves and preparing different organic solutions with the required concentration for the loading step.

2.3.4 Preparation of MIPs by precipitation polymerisation (first attempt)

An attempt to prepare CIPC-MIP and CPF-MIP by a precipitation polymerisation was conducted. For polymer preparation, 0.1 mmol of the templates, CPF and CIPC (35.06 and 21.1 mg) and 0.4 mmol of the monomer MAA (34.4 mg) were mixed with ACN, toluene or a combination of the two different volumes (between 5-11 mL) in a 20 ml round vial at room temperature. The cross-linker EGDMA (2 mmol, 396.6 mg) and different amounts of the initiator AIBN (3 to 97 mg) were then added to the mixture. After dissolving all the polymerisation components, the solutions were sonicated in a water bath and degassed using N2 for 5-10 minutes. The vials were then sealed with a screw cap and either immersed in an oil bath placed on a hot plate/magnetic stirrer (RO 15 power IKA®- Werke, GmbH & Co, (Staufen, Germany) or incubated in an oven. The temperature was gradually increased from room temperature to 60°C over a two-hour period and then left for 24 hours. Non-imprinted polymers (NIPs) were prepared with the same procedure but without the presence of the template. The polymers were characterised under a light microscope (Zeiss Axioskop).

Different parameters were evaluated for the polymer development at the micro scale. First, only the initiator quantities (i.e. 11.8, 12.1, 15, 20.1 and 25.2 mg) were varied and used for polymerisation with no changes to other reagents. Second, the polymerisation was performed either in an oil bath or incubator (Stuart S150 Orbital Incubator). The vials were rotated in the incubator at a 100 and 250 rpm range. Third, the quantities between the monomer and cross-linker (60.7:435.3) (60.9:455) (60.6:472.9) were also evaluated in mg. In addition, two different solvents (ACN and TOL) were investigated either

Chapter 2

separately or in a mixture of ACN-TOL (25:75; v:v). These two solvents were previously found to be ideal for the production of particles of micron size (Turiel et al., 2005). Another reported factor was the functional monomer and the ratio between both monomer and initiator as (34.6:11.8) (34.9:12.1) (45.1:15) (49:20) (55:25.2) and (51.2:12) in mg.

2.3.5 Preparation of MIPs by bulk polymerisation

Polymers were prepared via thermo-polymerisation for two different templates (CPF and CIPC) using four different porogens. The polymer compositions are reported in Table 2-1. MIPs M1 to M4 were prepared for CIPC and M5 to M8 were prepared for CPF with ACN, CHL, DCM and TOL employed for each template. For each MIP, a corresponding NIP was prepared using the same composition but with the template absent.

Table 2-1: Different compositions of synthesised MIPs. Polymer Template Functional Crosslinker Molar solvent Polymerisation monomer ratio* condition M1 CIPC MAA EDGMA 1:4:20 ACN Thermal 65 °C M2 CIPC MAA EDGMA 1:4:20 CHCL3 Thermal 65 °C M3 CIPC MAA EDGMA 1:4:20 DCM Thermal 65 °C M4 CIPC MAA EDGMA 1:4:20 TOL Thermal 65 °C M5 CPF MAA EDGMA 1:4:20 ACN Thermal 65 °C M6 CPF MAA EDGMA 1:4:20 CHCL3 Thermal 65 °C M7 CPF MAA EDGMA 1:4:20 DCM Thermal 65 °C M8 CPF MAA EDGMA 1:4:20 TOL Thermal 65 °C * Molar ratio of template = functional monomer: crosslinking monomer

The polymers were separately prepared for both targets as follows: CIPC (149.65 mg; 0.7 mmol) and CPF (245 mg; 0.7 mmol) as the template and MAA (240.9 mg; 2.8 mmol) were dissolved in 3.017 ml of each porogen (ACN, CHL, DCM and TOL). EDGMA (2775.8 mg; 14 mmol) and AIBN (30.1 mg) were then added to the solution. Next, a balance (Mettler Toledo, PB303) was used to weigh the reagents in a 15 ml glass vial and the pre-polymerisation mixture was then sonicated for 3-5 minutes to ensure complete dissolution. The mixture was then purged with N2 gas for 5-7 minutes to remove dissolved oxygen before it was sealed with a screw cap and parafilm. NIPs were prepared in the same manner but in the absence of the templates. Thermo-polymerisation was conducted in a Pyrex crystallising dish filled with oil that was heated on a hot plate at 65 °C for 24 hours. Then, the obtained monoliths were crushed and ground using a pestle and mortar (Fisher Scientific, UK) and 75 and 125 µm mesh size sieves (Retsch) were

Chapter 2

used to sieve the polymer with methanol to obtain particles. To remove the template, the resulting polymers were transferred into a paper filter, placed in a Soxhlet apparatus and extracted for 24 hours with a mixture of methanol and acetic acid (90:10;v:v) (250 mL). The same volume of methanol was also used to equilibrate the polymers before they were dried in a column oven (CTO-10AVP, Shimadzu) at 55 °C for 24 hours to remove any remaining solvent before starting the SPE experiments. Non-imprinted polymers were also produced following the same procedure without the presence of the templates. Figure 2-2 shows the preparation steps of the bulk imprinted polymers.

Figure 2-2: General procedure for the synthesis of the bulk imprinted polymers (Haupt, 2008).

2.3.5.1 Molecularly imprinted polymer–solid-phase extraction procedure (loading CIPC and CPF in organic solvent)

The experiments were separately conducted for both MIPs produced for CIPC and CPF and were evaluated across all four porogens by testing each polymer with the four solvents (i.e. ACN, CHL, DCM and TOL) as loading carriers for the analytes. The procedure was conducted as follows: a measured quantity (100 mg) of the MIPs was used to pack 12 SPE tubes (polypropylene) with PE frits of 20 μm porosity (1 mL). The SPE tubes were then placed in a Visiprep™ SPE vacuum manifold with 12 positions (Supelco, Sigma–Aldrich, UK). Next, different standard solutions of CIPC or CPF (100 ng/mL) were prepared in the four organic solvents. To equilibrate the polymers, the MISPE tubes

Chapter 2

were first conditioned by passing 5 mL of MeOH under vacuum to activate the polymers, followed by 3 mL of the solvent that matched the loading solution. The solid phase in the tubes was not allowed to dry at any time during the analysis. Then, 1 mL of each CIPC or CPF solution prepared in ACN, CHL, DCM and TOL was loaded into the MISPE tubes under vacuum, and a separate experiment for each target was conducted, with a flow rate of approximately 1 mL/min. Next, a volume (1 mL) of a similar solvent was used for washing. Finally, 1 mL of methanol-acetic acid (90:10; v/v) was used for elution. The experiments for each polymer and loading solution were performed in triplicate. The filtrates were then collected in chromatography vials for HPLC analysis, and after ACN was used as a loading solvent, the collected filtrates were injected directly into HPLC for analysis. The filtrates from CHL, DCM and TOL and elution were dried under a stream of nitrogen gas until dry and were reconstituted with 1 mL of ACN before conducting the analysis. Next, 100 ng/mL of CIPC solution and CPF solution were prepared in ACN and quantified to be used as a reference. The percentages of each analyte bound were then calculated by subtracting the analyte recovery values in loading, washing and elution from the reference solution.

2.3.5.2 A rapid optimisation of packing materials and loading volumes

Before the MIP protocol was carried out, a rapid optimisation of packing volumes and analyte concentrations was conducted. Different polymer amounts (30, 50 and 100 mg) of CIPC-MIPs and CPF-MIPs were evaluated. First, 100 mg of each MIP prepared in ACN was evaluated at a concentration of 100 ng/mL (CIPC and CPF). Then, the experiment was repeated to evaluate performance at smaller polymer amounts of 50 and 30 mg for each MIP, which were used to pack each tube independently. Each volume (50 and 30 mg) was separately evaluated at three concentrations (150, 300 and 500 ng/mL) of each analyte. The procedure was carried out by increasing the concentration and lowering the amount of each polymer in three replicates.

2.3.5.3 Polymer mass capacity

After the polymer volumes were selected, each CIPC-MIP and CPF-MIP (30 mg) was used to pack the MISPE tubes, and the binding capacity was determined by continuously loading the pesticide standard solutions (500 ng/mL) until the point of saturation. MeOH (3 mL) and H2O (3 mL) was used to condition and equilibrate the polymer, and 1 L (500 81

Chapter 2

ng/mL) of each pesticide solution (CPF or CIPC) was loaded onto the corresponding MISPE. Then, 5 mL of MeOH-acetic acid (90:10, v/v) was used to elute the analytes from the MIP. For the loading step, 1 L of pesticide standard solution was loaded in 10 mLs aliquots, which were collected separately for quantification. All filtrates in the loading, washing and elution steps were then collected and quantified using HPLC. The binding capacity of the polymers was reported as µg of pesticide bound/g polymer and calculated with the following formula:

Binding capacity = [(500 − X)/0.03g] where 500 (µg) is the total loaded amount of pesticide, X (µg) is the amount of pesticide- residue-found in the filtrates, and 0.03 (g) is the polymer mass of a MISPE. The value was then extrapolated for 1g of the MIP.

2.3.5.4 Optimisation of washing protocols

An attempt was made to develop washing protocols for both CIPC and CPF. Four different mixtures of organic solvent-water—ACN-H2O (10:90; v:v), (20:80;v:v), (30:70;v:v) and MeOH-H2O (50:50;v:v)—were tested for suitability as washing solvents. Following the conditioning and loading of each MISPE tube with 500 ng/mL of corresponding analyte prepared in aqueous solution (30 mL), 10 mLs of each washing solution was passed through each MISPE tube. Then, 5 mLs MeOH containing 10% acetic acid was used for the elution, which was adapted from a comparable study. A similar procedure was conducted for each NIP as a blank. For the washing solution, each mL of the filtrate was collected separately for analysis. The filtrates (1 mL) were then collected and dried (where appropriate), and the concentrations of CIPC or CPF were quantified using HPLC.

2.3.5.5 MISPE extraction procedure in aqueous solution

A measured quantity (30 mg) of each polymer (ACN-MIP, CHL-MIP, DCM-MIP and TOL-MIP) of CIPC and CPF was independently packed in a 1 mL SPE tube. Each MISPE was first conditioned using MeOH (3mLs) and H2O (3mLs) under vacuum, and water was used to equilibrate the polymers for loading in the aqueous solution and the solid phase was not allowed to dry at any time during the analysis. Standard concentrations of CIPC and CPF (500 ng/mL) were then prepared in distilled water (30 mLs) and loaded (5

Chapter 2

mLs aliquots) into each tube with a flow rate of approximately 1 mL/min. Next, a volume (5 mLs) of the washing protocol of ACN-H2O (20:80 v:v) for CIPC and MeOH-H2O (50:50 v:v) for CPF were used for the corresponding polymers. Finally, 10 mLs of an elution mixture of MeOH-acetic acid (90:10; v:v) was used to elute the analytes. For each polymer, a filtrate of 5 mLs of the loading solution and a filtrate of 1 mL of the washing and elution were collected in individual vials for HPLC analysis. All experiments were performed in triplicate and NIP was also prepared for each condition. The CPF was pre- concentrated six times whereby 30 mLs were used for CPF loading and 5 mLs of elution mixture were used for recovery.

2.3.5.6 Selectivity study

The selectivity study was conducted solely for CPF-MIP prepared in ACN. Analytical standards of other pesticides CIPC, IPC, 3-CA and TCP were prepared separately in aqueous solutions by first preparing stock solutions (1mg/mL) in ACN. They were then diluted in H2O to prepare the required concentrations (500 ng/mL). Ten tubes were prepared, five packed with MIP (30 mg) and five packed with NIP (30 mg). All tubes were conditioned with MeOH (3 mLs) and equilibrated with H2O (3 mLs). Next, 30 mLs of each aqueous solution of CPF, CIPC, IPC, 3-CA and TCP were loaded into the corresponding MISPE and NISPE. Then, five mLs of MeOH-H2O (50:50, v;v) were used for washing. Finally, five mLs of MeOH and acetic acid (90:10; v/v) were used for elution each analyte. The filtrates of all the steps were collected for quantification by HPLC in triplicate.

2.4 Results and discussion

Initially, attempts were made to develop polymers for solid phase extraction by precipitation polymerisation. However, after evaluating several procedures, we realized that they could not produce polymers that could be packed in a cartridge for SPE application. Alternatively, four porogens were evaluated through synthesis of different polymers by bulk method in order to select an appropriate solvent and develop a solid phase extraction procedure.

Chapter 2

2.4.1 Precipitation polymerisation

2.4.1.1 Initial synthesis by precipitation polymerisation

The initial synthesis employed a formula (1:4: 20; template: monomer: cross-linker) that is commonly used for bulk polymerisation with one difference which is using a diluted system (Lai et al., 2007). The results produced polymers with spherical and uniform particles of around 1 µm (Figure 2-3), which was confirmed by an initial characterisation using an optical microscope. Unfortunately, MIPs with such small particles were unsuitable for packing SPE tubes with frits having a porosity of 20 µm.

Figure 2-3: An image of particle size of approximately 1 µm produced by precipitation polymerisation and taken by a light microscope, 40X magnification.

2.4.1.2 Evaluation of different factors

A large number of polymers were produced at different conditions in an attempt to produce MIP particles larger than 20 µm. The types and quantities of functional monomers, cross-linkers, initiators, and porogens can affect particle size (Wang, et al., 2003). Therefore, the results of the first attempt in which the reagents were fixed and the initiator was varied showed no observed clear effects on particle size under a microscope; they were still of sub-micron size and thus too small to be packed into a SPE tube. All

Chapter 2

evaluated parameters resulting from changing the ratio between monomer and cross- linker reported that particle sizes can be increased (Shi et al., 2012) using two different solvents: ACN and TOL separately, or as ACN-TOL (25:75; v:v) which was reported to be ideal for the production of particles in micron size (Turiel et al., 2005) showed milky polymer, gel-like polymer and polymer in nano-scale. When the cross-linker was increased and all the other polymerisation components were kept fixed, the results produced a solid polymer. Furthermore, even the rotation of vials has been reported to affect polymer size (Wang et al., 2003). Figure 2-4 shows some of the polymers that were produced in various experiments.

Figure 2-4: Some of the polymers produced by protocols of different precipitation polymerisations. Relatively solid (A), nano-polymer (B), milky-like (C) and gel-like (D).

While some studies have reported producing MIPs by precipitation polymerisation for SPE applications, the polymerisation conditions used in this research could not produce MIPs with appropriate particle sizes. The most successful polymer resulting from these experiments exhibited particle sizes of 1 µm, which comported with the results of other studies that produced similar or smaller particle sizes (Lai et al., 2007). Moreover, many protocols in the literature have reported the successful synthesis of polymers at micron sizes, but the optimisation conditions to obtain particles above 10 µm does not seem straightforward. Further, the protocols available in the literature are typically not

Chapter 2

standardised, and several parameters are required, such as the volume of a monomer and a cross-linker, a solvent type and volume, an initiator concentration (Li et al., 2003) and rotation speeds (Wang et al., 2003). All of these should be taken into account and optimised to successfully obtain microparticles of desired sizes. Table 2-2 lists some examples of polymerisation conditions.

Table 2-2: Different precipitation polymerisation conditions reported in the literature. Template Functional Crosslinker AIBN Solvents size Reference mmol monomer µm 0.2 0.8 4 0.02 25 mL CAN - Fan et al. (2014) 1.5 4 28.8 1.9 128 mL ACN:T 3.5 Wang et al. (2003) 0.4 4.9 5.6-9.6 2.5 40 mL CAN 0.4-1 Zhang et al. (2012c) 0.17 0.68 3.40 0.25 12.5 mL ACN:T 3-5 Turiel et al. (2005) 1 4 20 0.2 30 mL T 80-150 Mao et al. (2013) 0.67 2.56 13.88 0.24 (40) 60 mL T - Dai et al. (2011) 0.3 1.5 2.4 10 mg 60 ml ACN:T - Yin et al. (2011) 0.23 12 16 0.30 (50) 85:25 4 Sadeghi et al. (2012) 0.25 1 5 10 mg 50 ml T 2 Yan et al. (2007)

The synthesis of the polymers by precipitation polymerisation was deemed to be not feasible and beyond the scope of this study. Alternatively, bulk polymerisation was used to produce MIPs as a preliminary study to select an appropriate solvent and develop a solid phase procedure, as historically a large number of MISPEs have been synthesised and reported in the literature using this method.

2.4.2 Bulk polymer preparation

For this study, a non-covalent method was selected due to its wide applicability and flexibility in providing relatively weak interactions such as hydrogen bonds and electrostatic and ionic interactions between monomers and a template. Thus, the bulk format was deemed the most straightforward, simple and suitable for SPE application. Many MIPs have been prepared for OP and carbamate pesticide classes by bulk polymerisation (Baldim et al 2012; Sanagi et al., 2013; Pereira and Rath, 2009; Qian., 2010; Baggiani., 2005 and Denderz et al., 2012). The interaction between MAA and the corresponding targets takes place via the oxygen atoms present in the structure of both compounds (CIPC and CPF) (Yao, 2008; Sanagi et al., 2011). In addition, the solvent’s physical and chemical properties govern the assembly forces between the templates and the functional monomers used. As a result, this process can weaken effective imprinting in non-covalent methods, and polymer morphology and porosity can also be affected by

Chapter 2

the solvent type. Therefore, we decided to evaluate ACN, DCM, CHL and TOL in this study by developing different MIPs for each pesticide. Figure 2-5 shows the polymers before and after grinding.

Figure 2-5: The bulk polymer after 24h polymerisation (A) and after grinding and sieving (B).

2.4.3 The development of standard calibration curves by using HPLC

The HPLC sensitivity for the identification of CIPC and CPF was then evaluated and showed satisfactory linearity for the calibration curves within the selected ranges. The coefficients of correlation (R2) were 0.99 and 1 for CIPC and CPF, respectively, (Figure 2-6 and Figure 2-7). However, HPLC sensitivity was not sufficient to accurately detect both CIPC and CPF at low concentrations.

Chapter 2

80 R² = 0.9969

40 Peak Peak area 30

0 0 100 200 300 400 500 600 700 800 900 1000

CIPC concentration (ng/mL)

Figure 2-6: A calibration curve of CIPC in ACN at the concentration range of 0–1000 ng/mL. 40 R² = 1 35

20 Peak Peak area 15

0 0 100 200 300 400 500 600 700 800 900 1000 CPF concentration (ng/mL)

Figure 2-7: A calibration curve of CPF in ACN at the concentration range of 10–1000 ng/mL.

Chapter 2

2.4.4 Evaluation of loading solvents for CIPC and CPF

Based on HPLC sensitivity, the concentration of 100 ng/mL used for loading was found to be optimum and fell within the legal limit range for most fruit and vegetable samples. It has been shown that using the same solvent for both polymerisation and loading can establish memory effects for analyte adsorption, maximise hydrogen bond interactions and impart affinity and selectivity to resultant MIPs (Yan, 2004). The result of the initial evaluation using the organic solvents ACN, DCM, CHL and TOL as carriers for loading each analyte in a corresponding polymer showed weak binding for both analytes to the MIPs and NIPs.

2.4.4.1 CIPC evaluation

The processed results of the HPLC analysis are presented in Table 2-3, which lists the adsorption percentages of CIPC for different MIPs and corresponding NIPs that were produced with different porogens. In addition, the table reports a broad range of CIPC affinity to the polymers ranging from 2-74% for MIPs and from 7-88% for NIPs.

Table 2-3: Percentages of retained analytes (CIPC) after loading in four different organic solvents at (100 ng/ml) in triplicate. CIPC - Synthesised polymer (CIPC-MIP) ACN- ACN- CHL- CHL- DCM- DCM- TOL- TOL- Carriers SD SD SD SD SD SD SD SD MIP NIP MIP NIP MIP NIP MIP NIP

ACN 44.98 0.61 35.05 2.68 18.30 1.97 23.35 2.38 73.14 8.06 41.89 1.99 66.08 8.42 68.78 4.67

CHL 42.14 0.00 34.69 0.00 10.10 1.45 26.19 5.55 67.82 6.91 65.82 17.04 73.72 2.81 88.11 10.90

DCM 35.40 1.63 40.37 2.22 2.21 1.45 7.57 4.34 61.17 6.41 52.52 23.37 44.05 9.94 61.13 9.57

TOL 43.56 2.68 8.09 1.84 12.31 2.50 20.19 2.89 56.51 5.02 27.92 7.19 46.30 15.34 35.96 17.96

In general, the MIPs and NIPs prepared in TOL and DCM showed the highest affinity for most loading carriers. In addition, TOL-MIP and DCM-MIP exhibited the highest affinity (74%) when ACN and CHL were used as carrier solutions. Overall, CIPC was adsorbed poorly in CHL-MIP, and exhibited the lowest adsorption percentages around (2-26 %) for all carrier solutions. In addition, the results of the organic solvent applications showed no major differences for all MIPs and NIPs, except for ACN-MIP. In this case, when TOL was the carrier, the difference was about 38%. In some cases, the NIPs showed higher adsorption levels than the MIPs, such as when TOL-MIP was used with ACN, CHL and DCM as carriers.

Chapter 2

Although binding occurred between different MIPs and CIPC, CIPC was not completely bound. In addition, MIP affinity results for CIPC did not show a significant difference when compared to the corresponding NIPs. The explanation for this behaviour could be attributed to the polarities of these solvents, as low-polar solvents promote H-bond interactions for analytes, which were most likely stronger than the H-bonds for the interactions with MIPs and NIPs.

2.4.4.2 CPF evaluation

When the evaluation was carried out for CPF polymers, the results indicated that the calculated percentages of CPF attained by all MIPs and corresponding NIPs were even less satisfactory than those when CIPC accounted for less than 50% of loaded CPF in the different organic solutions. Table 2-4 lists the CPF percentages after the CPF analytes were quantified. The CPF binding percentages ranged from about 24% to almost 39% for MIPs and 33% to 57% for NIPs.

Table 2-4: Percentages of retained analytes (CPF) after loading in four different organic solutions at (100 ng/ml) in triplicate. CPF - Synthesised polymer (CIPC-MIP) ACN- ACN- CHL- CHL- DCM- DCM- TOL- TOL- Carrier SD SD SD SD SD SD SD SD MIP NIP MIP NIP MIP NIP MIP NIP

ACN 22.13 0.52 21.12 1.96 19.71 1.48 36.84 3.93 35.06 3.74 57.07 5.09 38.74 4.33 21.73 7.13

CHL 20.88 0.59 33.23 17.97 32.56 1.48 23.13 5.14 29.35 2.45 39.14 2.45 19.84 2.83 28.34 2.83

DCM 19.14 0.57 18.75 1.23 22.28 1.48 30.85 6.80 24.46 7.34 42.40 3.74 18.90 1.64 34.96 1.64

TOL 23.99 1.24 21.56 1.69 26.56 6.47 34.27 3.93 37.51 3.74 48.92 2.45 28.34 2.83 49.13 12.78

The results clearly show that CPF adsorption for different polymers was very weak, and there were no significant differences between MIPs and NIPs. For example, the CHL- MIP and DCM-MIP showed lower binding levels than their corresponding NIPs for most loading carriers except ACN. Even other polymers with different carriers showed only small levels of CPF binding and no noticeable differences between MIPs and NIPs; this could be attributed to their relatively high affinity to the organic media rather than to either the MIPs or NIPs. This was most likely a result of a poor medium environment that suppressed the hydrogen bond interactions between the analyte and MAA that caused these analytes to remain in the hydrophobic environment. Further, one explanation for these results might be that the hydrophobic targets preferred to establish interactions with

Chapter 2

the aprotic and low-polar organic solvents rather than with the more hydrophilic polymer matrix (Beltran et al., 2010).

In fact, employing a porogen as a medium for analyte rebinding can have beneficial results and show strong and selective binding due to memory effects as well as MIP swelling properties, which are assumed to vary with different solvents (Yan, 2004; Meier, et al., 2012). However, this is not always the case, as the opposite was also observed in a different study (Denderz et al., 2012). As a result, the MIP affinity study was repeated using an aqueous solution as a medium for loading.

2.4.4.3 Loading CPF and CIPC in aqueous solution

Aqueous solutions were used as carriers to increase adsorption towards different MIPs. It is known that changing a medium from an organic solvent to an aqueous solution will alter the chemistry of the interactions between templates and polymers. In such environments, binding is mainly due to nonspecific interactions through electrostatic and hydrophobic interactions as well as through hydrogen bonds (albeit to a lesser degree) (He et al., 2007). Under these conditions, the washing step becomes crucial in the selectivity of MIPs, as it can disrupt the weakest nonspecific interactions (Siemann et al., 1996). Therefore, a rapid optimisation procedure for the packing volume, analyte concentrations, and washing solvents was conducted.

2.4.5 The MIP packing volume

The result of the first evaluation of MIP-CIPC and MIP-CPF using 100 mg and an analyte concentration of 100 ng/mL showed that the sorbents exhibited a high binding capacity after loading CIPC and CPF solutions (500 mLs). Thus, the packed polymer amounts were reduced to 50 mg and 30 mg and the analyte concentrations were varied to include 150, 300 and 500 ng/ml. After each polymer was loaded with a volume of 500 mLs with a corresponding solution at each concentration, there were no signs of these analytes in the filtrates and only negative solutions. Therefore, 30 mg of the polymer and 500 ng/ml of the analyte concentrations were selected due to the high binding capacity demonstrated by the MIPs. The polymer quantity in this work was in agreement with Xie et al. (2010).

Chapter 2

2.4.6 Polymer mass capacity

Breakthrough is defined as the volume of a loading solution at which a total analyte adsorption can no longer be attained due to the saturation of a polymer (Chapuis et al., 2004). For this study, the results indicated that whilst a large volume of standard solution 1 L (500 ng/mL) was loaded into MISPE tubes, traces in the filtrate were still not observed after HPLC analysis. Based on this result, the binding capacity for the MIPs was calculated as higher than 16.7 mg/g polymer for both CIPC and CPF.

2.4.7 Washing optimisation

The MIP prepared in ACN was the only one that consistently exhibited stronger binding characteristics than its corresponding NIP. As a result, the washing step was first optimised using ACN-MIP and ACN-NIP prepared for both analytes. CIPC and CPF in aqueous solution were completely adsorbed after loading and confirmed by HPLC analysis. The results of the quantification of CIPC in the washing filtrate are presented in Figure 2-8.

80 MIP-washing

NIP-washing 60

Washed amount (%)

(30:70; ACN:H2O)(20:80; ACN:H2O)(10:90; ACN:H2O) (50:50; MeOH:H2O)

Figure 2-8: CIPC washed amount (%) with different mixtures of washing solutions (5ml). Comparison between washed amounts of CIPC from columns loaded with 30 ml of 500ng/ml.

Chapter 2

The results showed that when the polarity of the mixture was decreased, a higher amount of analyte was washed. At 10% ACN, the washing solution was not able to disrupt the interaction between CIPC and the polymer, and when the percentage was increased to 20%, about 5% and 10% were removed from both MIP and NIP, respectively, and a large amount of the target analyte (between 65% and 75%) was removed from both MIP and NIP respectively at 30% ACN. When 50% of the mixture was MeOH, about 55% and 60% of CIPC were washed from the MIP and NIP, respectively. For all tested polymers, the results showed that the washed amounts of CIPC from the MIPs was slightly lower than those of the NIPs, which suggests that there were stronger interactions present between the MIPs and CIPC than the corresponding NIPs. However, these differences were not substantial. As a result, a ACN-H2O (20:80; v:v) mixture was selected. With respect to the CPF analyte, the results of HPLC analysis are presented in Figure 2-9. In either case, the mixtures containing ACN were not capable of disrupting the nonspecific binding for MIPs or NIPs. This could be explained by the low polarity of the mixture, which was unable to break the interactions established between the analyte and its respective MIP.

MIP-washing 3 NIP-washing

Washed amount (%)

(30:70; ACN:H2O)(20:80; ACN:H2O)(10:90; ACN:H2O) (50:50; MeOH:H2O)

Figure 2-9: CPF washed amounts (%) with different mixtures of washing solutions (5ml). Comparison of washed amounts of CPF from column loaded with 30 ml of 500ng/ml.

Chapter 2

As the interaction between CPF and MAA is based on hydrogen bonds, it is possible for this bond to be disrupted by less polar solvents (He et al., 2007). However, ACN is slightly more polar than MeOH (polarity index 5.8 for ACN and 5.1 for MeOH) (Hemwimol et al., 2006). Therefore, MeOH (50%) was able to disrupt the non-specific interactions and remove a greater amount of CPF from the NIP than the MIP; this result appeared in each mL analysed by HPLC. After the evaluation of two volumes (5 and 10 mLs) of the solution used for washing, the 5 mLs amount was selected because most of the target analyte was removed using the first 5 mLs. The reason that the NIP released a higher amount of CPF than the MIP was perhaps due to the presence of recognition sites. In addition, 5 mLs of MeOH containing 10% acetic acid was used for elution, which was based on a comparable study (Siemann et al., 1996 and Mayes and Whitcombe, 2005).

2.4.8 Adsorption experiment

2.4.8.1 Evaluation of different CIPC-polymers

All different MIPs prepared for CIPC were tested with ACN-H2O (20:80; v:v) for washing and acetic acid-MeOH (10:90; v:v) for elution. Figure 2-10 shows the results of the quantified CIPC by HPLC from all filtrates of the washing step. CIPC was completely retained by different polymers at the loading step. In addition, contradictory results were observed, whereby higher CIPC amounts were washed from the MIP than the NIP for all polymers except the one prepared in ACN, for which the NIP was slightly higher than the MIP; and this was the most acceptable result among all experiments. This could be attributed to the washing protocol, which required further optimisation.

Chapter 2

20 MIP NIP 15

Washed amount (%)

Tol ACN CHL DCM Figure 2-10: Washed amount of CIPC after loading 30 ml (500 ng/mL CIPC) prepared in water onto MIPs and NIPs. 5 mLs ACN:water (20:80,v:v) was used for washing, and four MIPs were synthesised in different porogens.

Then, CIPC was eluted from the polymers with 5 mLs of the elution mixture. As expected, the amount eluted from the NIPs was higher than or similar to that of the MIPs, except the polymer prepared in ACN, which could indicate the presence of recognition sites. Figure 2-11 shows the eluted CIPC values from different MIPs and NIPs after HPLC quantification. The CIPC amount recovered was pre-concentrated six times, 30 mLs were used to load CIPC into the tubes and 5 mLs of elution mixture was used for recovering it. While a promising ACN-MIP affinity for CIPC was demonstrated, the difference remained insignificant. However, we concluded that the CIPC polymer experiment requires additional optimisation to improve the washing protocol that would account for other factors, such as testing additional washing mixtures, flow rates, and pH values.

Chapter 2

100

80 MIP NIP

Recovery (%) 20

Tol ACN CHL DCM Figure 2-11: Recovery of CIPC in an elution fraction after loading 30 mL (500 ng/mL) in an aqueous solution and washing. Elution step: 5 mLs, MeOH: acetic acid (90:10;v:v) was used for elution. Four MIPs were synthesised in different porogens.

2.4.8.2 Evaluation of different CPF-polymers

The results of this experiment indicated that no CPF was detected in the filtrates and thus was completely retained by all polymers due to hydrophobic and electrostatic interactions

(Mayes and Whitcombe, 2005). The optimised washing protocol MeOH-H2O (50:50; v:v) was then used, and the percentage of CPF was calculated in the filtrates (Figure 2-12).

10.0 MIP NIP 7.5

5.0

2.5

Washed amount (%)

0.0

Tol ACN CHL DCM

Figure 2-12: Washed amounts of CPF after loading 30 ml (500 ng/mL CPF) prepared in water for MIPs and NIPs. 5 mLs methanol-water (50:50, v:v) was used for washing. Four MIPs were synthesised in different porogens; all were dried and reconstituted in 1 mL of ACN.

Chapter 2

The results show that all three different polymers prepared in ACN, CHL and TOL released a greater amount of CPF from NIPs than MIPs. As expected, this suggests higher affinity and the likelihood of recognition sites (Schirmer and Meisel, 2006). Figure 2-13 presents the results of CPF quantification in the elution filtrates, and shows that CPF amounts recovered from the MIPs prepared in ACN, CHL and DCM were higher than their corresponding NIPs. However, only the difference between the MIP and NIP prepared in ACN was high enough. This difference was considered the best of all especially the same behaviour of the polymers was noticed with each single experiment where the MIP attained higher CPF than the NIP. In fact, for the other polymers, the differences fell inside the standard deviation of the measurements, and thus they are considered inappropriate. By contrast, the NIP prepared in TOL seemed to release a slightly higher amount of CPF in the elution step. Furthermore, one result for all polymers was that CPF was not fully recovered in the 5 mL elution mixture, even if the amount lost at the washing step was considered. Further, a subsequent 5 mL of elution mixture was passed through the tubes, but the amount of CPF released was below the HPLC LOD value. Nevertheless, the MIP prepared in ACN possessed the highest affinity and showed the greatest amount of CPF recovered at the elution step (around 75%), representing the best performing MIP.

Chapter 2

100 MIP NIP 80

Recovery (%) 20

Tol ACN CHL DCM

Figure 2-13: Recovery of CPF from an elution fraction after loading 30 mL (500 ng/mL) in an aqueous solution and washing. 5 mLs, MeOH: acetic acid (90:10;v:v) was used for elution. Four MIPs were synthesised in different porogens; all were dried and reconstituted in 1 mL of ACN.

We concluded that the MIP prepared in ACN using a standard MIP synthesis protocol was the most promising material. While other studies have reported the use of DCM, toluene and chloroform as porogens to produce MIPs for other pesticides (Xin et al., 2013; Baldim et al., 2012; Schirmer and Meisel, 2006; Zhu et al., 2005), this finding comports with those of other studies for which MIPs were produced for OPPs using ACN as a porogen (Wang et al., 2013; Santos et al., 2012); Wei et al., 2011). Furthermore, only the polymer prepared in ACN for CPF was considered for further characterisation.

2.4.9 Selectivity study

As the polymer performance could not be evaluated solely by investigating binding and capacity, its target selectivity to other related compounds was also necessary. The MIP selectivity study towards CIPC, IPC, 3-CA and TCP which chemical structures are shown in Figure 2-14; the pesticide bindings were evaluated using the CPF protocol.

Chapter 2

TCP IPC

3-CA

Figure 2-14: Chemical structures of compounds used to test their affinities towards CPF- polymers.

Not surprisingly, the results of the quantified analytes in the filtrates at the loading and washing steps showed that all compounds were completely adsorbed. In addition, between 55-85% of these compounds were removed at the washing step (Figure 2-15). Over 90% of CPF was still selectively adsorbed. The results also indicated that the MIP exhibited low selectivity for these compounds when compared to CPF, which possibly had higher molecular recognition characteristics.

Chapter 2

100

Washed amount (%) 20

IPC CPF CIPC 3-CA TCP

Pesticide compunds Figure 2-15: Percentage of CPF and other compounds washed off from the MIP with 5 mLs (MeOH;H2O; 50:50, v:v).

At this washing step, most of the retained compounds were eluted, and about 80% of CPF was significantly recovered (Figure 2-16). A possible explanation for this result was that the interactions occurred mainly via non-specific binding between the MIPs and the tested analytes. This could also be attributed to different log Pow (partition coefficient in n- octanol/water) values, which describe the lipophilic or hydrophobic properties of a compound. The log Pow values of CPF, CIPC, TCP, IPC, and 3-CA are 5.27, 3.51, 2.63,

2.6 and 1.72, respectively. CPF exhibited a higher log Pow value than other compounds, which suggests that the strong adsorption to the polymer as high log Pow would promote hydrophobic interaction. While TCP and IPC have almost similar log P values, they were expected to behave in a similar fashion, instead they had different amounts washed from the polymers, the elusion resulted in almost the same volumes. The resultant amount of TCP in the washing step was underestimated, which is most likely a result of an experimental error. Further, 3-CA is a highest polar compound due to its amino group

(NH2), which can acquire a positive charge. Therefore, since 3-CA has pka value of 3.52 (T3DB, 2016), which is responsible for acidity levels, perhaps it is not totally neutral and you might have a small degree of positive charges that makes it interacting with the negative charges in the polymer (due to the carboxylic groups (–COO-) of MAA. As a result, a small amount remained after the washing step and it was eluted during the

100

Chapter 2

recovery. This result indicated the contribution of hydrophobic effects. Therefore, selectivity could have been influenced by the hydrophobic properties of each compound as a result of different log Pow values under the applied conditions.

100

Recovery (%) 20

IPC CPF CIPC 3-CA TCP Pesticide compounds

Figure 2-16: Percentage of CPF and other compounds recovered after elution with 5 mLs MeOH: acetic acid (90:10; v:v).

This result demonstrates that under the tested conditions, MIP prepared in ACN was sufficiently selective for detecting CPF in aqueous solution among other compounds that might present in real samples. The main implicit mechanism for the template selectivity of this MIP could be based on a shape recognition of the binding sites present in the polymeric structure (De Barros et al., 2010).

101

Chapter 2

2.5 Conclusions

MISPE sorbents for selective binding of both pesticides CIPC and CPF were synthesised using a thermo-polymerisation method with different porogens, including acetonitrile, chloroform, dichloromethane and toluene. These organic solvents were also used as loading carriers for CIPC and CPF. The results showed that they were inappropriate for this application and that the polymers exhibited weak retention. An aqueous solution was also used as a loading carrier for CIPC and CPF, and both pesticides showed high affinities to the polymers. Moreover, CPF-MIP showed a high binding capacity (greater than 16.7 mg/g polymer) when tested in an aqueous medium.

A rapid optimisation work for washing and packing different sorbent volumes in a SPE tube was conducted. Several mixtures containing organic solvents and water were used in the washing process and the optimal conditions were identified. After evaluating the MISPE for CIPC with an aqueous solution for analyte loading, the results showed no significant difference between MIP and NIP. In fact, the NIP exhibited slightly higher affinity than the MIP in most polymers, except when the MIP was prepared in ACN, which exhibited a negligible difference. Due to time constraints, the experiments for CIPC were discontinued. By contrast, the MIPs prepared in ACN, CHL and TOL for CPF showed higher affinities than the NIPs. Among all, ACN-MIP showed the highest affinity to CPF when tested in aqueous solution. Moreover, when MIP selectivity towards CPF was assessed, the MIP was relatively selective compared with other pesticides.

Although the MIP that was produced for CPF could be suitable to extract CPF from real samples. Further optimisation of other factors such as pH, loading volume, washing and elution mixtures would be preferable to improve the SPE protocol.

102

Chapter 3

Chapter 3: Detection of organophosphate pesticides using a planar waveguide immunosensor

103

Chapter 3

3.1 Abstract

Organophosphate pesticides are extremely toxic compounds that act as acetylcholinesterase inhibitors in animals and humans. Due to the nature of their application they may contaminate agri-food products. Conventional methods for pesticide detection are laboratory-based. However, rapid detection methods for on-site analysis using biosensors are highly appealing due to the low cost, rapidity and achievable sensitivity. The aim of this study was to illustrate the use of an immunosensor for rapid pesticide detection. A fluorescence planar waveguide immunosensor was employed as a rapid detection platform using indirect competitive inhibition assays and a fluorescent- labelled secondary antibody, with strawberries as a model food product, and parathion and chlorpyrifos as target pesticides. Chlorpyrifos detection was achieved by immunosensor with high specificity and an IC50 of 185.3 ng/g and a dynamic range (IC20-

IC80) of 11.4-1629.6 ng/g fitting to the maximum residue level with the LOD and LOQ of 28.2 ng/g and 79.1 ng/g respectively. The average % recovery obtained for three different concentrations of chlorpyrifos was 109.5 ± 5.6 with a %CV <17%. Parathion detection was achieved with an IC50 of 44.5 ng/g and a dynamic range of 10.1-147.4 ng/g though a broader specificity was observed in the detection of other organophosphate compounds. The LOD and LOQ of the assay were found to be 6.5 ng/g and 15.4 ng/g, respectively. The average % recovery showed 110.9 ± 8.3 with a %CV <19% at three different concentrations. Both independent assays are promising as screening tools that could be applied to additional food products prior to confirmatory analysis. Though these methods would be more suitable to use in follow-up monitoring for both quality and safety control. However, with specially designed immunoreagents the benefits of multiplexing analysis by immunosensor could be realised.

Both assays are promising screening tools that could be applied to additional food products, particularly for those exhibiting CPF regulations prior to confirmatory analysis. As a result, these could be suitable tools for follow-up monitoring for both targeted quality and safety control. While sensitivity enhancements may be required for other food products with lower maximum residue values, these assays offered improved sensitivity compared with currently available and published detection methods for on-site analysis.

104

Chapter 3

However, the PT assay was not limited to specific detection purposes, but alternatively could be utilised as a class selective platform.

105

Chapter 3

3.2 Introduction

In recent years, concerns about the risks associated with pesticide use in agriculture have increased, and a rise in adverse health effects due to exposure to pesticide residues in foodstuffs has been confirmed by scientific research. The OP class of pesticides is one of the most widely used classes employed for pest control (Xu, et al., 2011). In recent years, CPF was banned from domestic uses in the USA, and the use of MRLs for many fruit and vegetable has been reduced in the EU (Commission Regulations, 2016). PT is a toxic organophosphate pesticide that has been banned in the EU and many other countries. However, it still can be traded illegally and can still be found in many food products. Moreover, these pesticides can enter the food chain through contaminated food, water or animal feed pathways, and as acetylcholinesterase, inhibitors can cause acute and chronic toxicities for humans as well as insects.

For pesticide analysis, conventional detection methods are normally used, including GC- MS and HPLC-MS/MS (Chen et al., 2017a). Whilst these analytical techniques are available, they require expertise and skilled staff. Similarly, immunological methods such as ELISA offer rapid and sensitive screening tools, but are presently limited by portability constraints. However, for site monitoring, sensitive, inexpensive, reliable, and effective affinity-based immunoassays have recently emerged. For example, the SnapEsi reader from MBio Diagnostics Inc. (Figure 3-1) is based on a new technique of planar waveguide technology, and uses a fluorescent label for detection that holds great potential for portability and multiplexing.

106

Chapter 3

Figure 3-1: MBio SnapEsi reader and a rack of assay cartridges. The reader is connected to a laptop computer.

This biosensor system, which uses a disposable cartridge and a fluorescence SnapEsi reader, employs a combination of two techniques: conventional fluorescence assays, such as fluorescence immunoassays, and a planar waveguide fluorescence illumination system (Mukundan et al., 2009). To date, a few studies have used this technique to deliver a proof of concept (Lochhead et al., 2011; Devlin et al., 2013;McNamee et al., 2014; McGrath et al., 2015). For example, Meneely et al. developed a rapid method for marine purposes that investigated the efficacy of using planar microarrays to detect paralytic shellfish toxins. The measurement system could perform assay with a run time of 15 minutes, which is feasible for on-site use (Meneely et al., 2013). The platform was used to detect toxins with a single-use cartridge that can be used for either portable singleplex or multiplex approaches. Thus technique directs a laser beam by a planar waveguide composed of a high refractive index material (i.e. glass or plastic) surrounded by a lower refractive index material (i.e. specimen or fluid). Figure 3-2 depicts a schematic representation of this technique for analysis by a biosensor.

107

Chapter 3

Figure 3-2: Schematic representation of the planar waveguide used in the MBio biosensor.

Briefly stated, the detection system measures far-red wavelengths generated by a laser source that are propagated by total internal reflection (TIR) through the surface of a planar waveguide at the interface between the waveguide high refractive index and an adjacent field. An evanescent field is created in the area between a sample medium and the waveguide surface, which penetrates the sample medium. Evanescent arrays that escape from the waveguide are used to excite and illuminate fluorescent dyes (fluorophores) present near the surface, which are used for labelling antibodies (McGrath et al., 2015; McNamee et al., 2014). In addition, nano-spotting technologies can be used to precisely immobilise a specific volume of immunochemical reagents on the solid surface, whereas fluorescently labelled antibodies can be used as binding materials within the liquid medium. The SnapEsi reader can then be used to detect and quantify levels of fluorescent signals for each array after a binding event. Ideally, antibody binding to the target conjugates immobilised on the surface is optimized and a high fluorescent signal is generated when no target analytes are present in the sample. When target analytes are present in the sample, fewer antibodies bind to the target conjugate, thereby diminishing the fluorescent signal (McGrath et al., 2015).

108

Chapter 3

3.2.1 The chapter aim

The aim of this study was to utilise the MBio planar waveguide biosensor to develop rapid singleplex immunoassays to detect individually CPF and PT organophosphate residues in strawberry samples to serve as a model system that requires minimal sample preparation.

109

Chapter 3

3.3 Materials and methods

3.3.1 Reagents and chemicals

For this study, Tween 20, bovine serum albumin (BSA), ovalbumin (OVA), dried skimmed milk (marvel), casein, dibasic sodium phosphate, monobasic sodium phosphate, sodium chloride, sulfuric acid, phosphate buffered saline (PBS) tablets, sodium bicarbonate, sodium carbonate, and Alexa Fluor™ 647 conjugates were purchased from Sigma-Aldrich (Dorset, U.K). In addition, pesticide standards for chlorpyrifos, coumaphos, parathion, phoxim, triazophos, dichlofenthion, azinphos-ethyl, phosalone, parathion-methyl, cyanophos and disulfoton were purchased from Sigma-Aldrich (Dorset, UK). First, to prepare working standards for the immunoassay, all standards were prepared as stock solutions in HPLC grade methanol at a concentration of 1mg/mL before making serial dilutions in an MBio assay buffer. Anti-chlorpyrifos polyclonal antibodies and chlorpyrifos–OVA protein conjugate were purchased from Agrisera (Vännäs, Sweden). In addition, Professor Zhen-Lin Xu of South China Agricultural University, China provided anti-organophosphate monoclonal antibody 1E2 and H11-ovalbumin (H11-OVA) conjugate as a gift (Xu et al., 2012b). Finally, horseradish peroxidase (HRP)- conjugated (Goat-anti-rabbit HRP) was purchased from DAKO (Produktionsvej, Denmark), tetramethylbenzidine substrate solution (TMB/E) was purchased from Millipore (Hertfordshire, UK), and Alexa Fluor 647 goat anti-rabbit IgG (GaR) antibodies and Alexa Fluor-647 goat anti-mouse IgG (GaM) antibodies were purchased from Invitrogen Ltd. (Paisley, Scotland).

3.3.2 LC-MS/MS determination of CPF and PT in strawberry samples

The samples were screened for CPF residues using an Acquity ultra-performance liquid chromatography (UPLC) i-Class coupled to a Xevo triple quadrupole mass spectrometer (TQ-MS) (Waters, Manchester, UK) using multiple reaction monitoring (MRM). Prior to sample screening, the UPLC-MS/MS were calibrated using chlorpyrifos and parathion standards prepared in methanol solutions ranging in concentrations from 0 to 1000 ng/mL. Chromatography was performed with an ACQUITY UPLC BEH C18 column (50 mm x 2.1 mm i.d., 1.7 µm particle size, 130 Å pore size) (Waters, UK). The two mobile phases consisted of water containing 0.1% formic acid (v/v) and methanol with 0.1% formic acid

110

Chapter 3

(v/v). The flow rate was set at 0.40 mL/min with the methanol concentration held at 10 % for 0.25 min, followed by an increase to 95% over 4.5 min and maintained at 95% for 0.25 min before returning to 10% concentration for a 1 min re-equilibration before the next injection, with the injection volume set at 5 μL (Table 3-1). Detection and quantification were achieved using targeted analysis via MRM involving the fragmentation of specific precursor ions (parent) using argon as the collision gas to at least 2 product ions (daughters). The system was operated in electrospray positive mode (ESI+) according to the following parameters: capillary voltage 1kV, source temperature 120°C, desolvation temperature 400°C, desolvation gas flow 750 L/h. The strawberry fruits were then washed, cut and homogenised, and an aliquot was used to prepare six samples fortified with 50 µL of CPF standards (0, 2.5, 5, 50, 200 and 1000 ng/g) and six samples fortified with 50 µL of PT standards (0, 2.5, 5, 50, 500 and 1000 ng/g).

To extract the samples, 1 g aliquot of the homogenised strawberry mixture was transferred into a plastic tube (15 mL) and fortified by the standards. Then, 4.95 mL of methanol was added and the samples were vortexed and placed in a roller mixer (Dynalon, SRT9/120V/60 9-Roller Analog Tube Mixer with Fixed Speed, 120V - 60Hz [IKA]) for 30 min before centrifugation (Thermo Legend T Plus Centrifuge, Thermo Scientific) at 4600 rpm for 10 min. Then, 1 mL of each extract containing 0, 0.5, 1, 10, 40 and 200 ng/mL for chlorpyrifos, and 0, 0.5, 1, 10, 100 and 200 ng/mL for parathion, was decanted and transferred into HPLC vials for a subsequent screening by the UPLC-MS/MS.

Table 3-1: ACQUITY UPLC mobile phase gradient. Time(min) Flow rate Solvent A Solvent B Curve Initial 0.4 90.0 10.0 0 0.25 0.4 90.0 10.0 6 4.75 0.4 5.0 95.0 6 5.00 0.4 5.0 95.0 6 5.10 0.4 90.0 10.0 2 6.00 0.4 90.0 10.0 6

3.3.3 Immunological Reagents

The immunological reagents CPF-pAb (3 mg/mL) and its conjugate CPF-OVA (3 mg/mL) were sourced commercially from Agrisera; the synthesis protocols are described elsewhere (Manclus et al., 1994). CPF- pAb was raised in a rabbit using CPF-BSA as the

111

Chapter 3

immunogen, and PT-mAb 1E2 (39.87mg/mL) and its conjugate H11-OVA (0.5 mg/mL) were received from a friend. PT-mAb was raised in mice using H11-BSA (immunogen) (Xu et al., 2010).

3.3.4 Characterisation of immunological reagents

3.3.4.1 UV-VIS Spectroscopy

A NanoDrop 8000 microvolume UV-VIS spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA) was used to characterise the reagents. The pesticide standards (CPF and PT) (1mg/mL), CPF-OVA (3 mg/mL), H11-OVA (0.5 mg/mL) and OVA (1mg/mL) were scanned over a range of 220-750 nm to determine if conjugations to OVA could be confirmed. A full scanning procedure was followed according to the company protocol. Initially, the device was washed by adding of 5µl of double-distilled water

(DDH2O) to the pedestal and was analysed. The pedestal was then wiped clean and the device was calibrated by applying to the pedestal 5µl of phosphate buffer as a blank and analysing. Each sample listed above was then analysed, and the same washing and blank calibration steps were followed after each sample analysis. Moreover, the concentration of CPF-pAb and PT-mAb 1E2 at A280nm were also determined by the same protocol to confirm the concentrations.

3.3.4.2 Characterisation of immunological reagents by ELISA

Determination of the optimal coating antigen concentration and antibody titre.

Reagents of CPF and PT were characterised using an indirect competitive inhibition ELISA assay format with a chequerboard design of coating antigen concentrations and antibody titres. Coating antigens, CPF-OVA and H11-OVA were prepared in carbonate/bicarbonate coating buffer (0.1M pH9.6). Nunc MaxiSorb ELISA 96-well Microtitre plates (Falcon 353070, Becton Dickinson, USA) were coated with 100 µl/well of 0.25, 0.5 and 1 µg/100 µL of CPF-OVA coating antigen, respectively, and for parathion at 0.25 µg/100 µL of H11-OVA coating antigen. Additional wells were coated with equivalent concentrations of the proteins BSA and OVA, respectively, as positive and negative controls. The plates were incubated overnight at 4 ºC. The coating buffer was then discarded, 200 µl/well of blocking buffer (2% marvel in deionised water) was added to each well, and the plates were incubated at 37ºC for 1 hour in a shaker. The blocking

112

Chapter 3

solution was then discarded for subsequent steps. Each CPF-pAb or PT-mAb 1E2 (1/1k, 1/2k, 1/4k, 1/8k, 1/16k and 1/32k) was serially diluted in a phosphate buffer (0.1M pH 7.2), and each dilution was added to each appropriate well of the corresponding plate that was coated with a different concentration of CPF-OVA, H11-OVA, BSA and OVA proteins. Phosphate buffer (0.1M pH7.2) was used as the negative standard, and CPF and PT standards at a concentration of 100 ng/mL prepared in a phosphate buffer (0.1M pH 7.2) were used as the positive standard.

Immediately following the addition of the antibody dilution (50 µL/well), 50 µL of each standard was added to the appropriate positive and negative designated wells for each plate. The plates were incubated overnight in the dark at 4°C and then washed four times with a wash buffer (0.0125% Tween 20 in 0.15M NaCl). A secondary antibody—goat anti-rabbit-horseradish peroxidase-labelled for CPF and goat anti-mouse-horseradish peroxidase-labelled for PT previously diluted to 1/2000 in phosphate buffer (0.1M pH7.2)—was added to each well of the corresponding plates, and the plates were incubated for 55 minutes at 37°C. The secondary antibody was then discarded and the plates were washed eight times with wash buffer. Then, TMB/E (100 µL/well) was added to each well and the reaction was stopped after five minutes using 25 µL/well of 2.5M sulphuric acid. The optical density (OD) for each well was immediately read at 450 nm using a TECAN Safire2 plate reader.

Evaluation of antibodies sensitivity by ELISA

Nunc MaxiSorb ELISA 96-well Microtitre plates (Falcon 353070, Becton Dickinson, USA) were coated with each coating antigen (CPF-OVA and H11-OVA) at a concentration of 0.25 µg/100 µL and were incubated overnight at 4°C. The solution was then discarded and the plate was blocked with 2% Marvel for 1 hour at 37°C. Dilutions at 1/250, 1/500 and 1/1K of each antibody (CPF-pAb and PT-mAb 1E2), (50 µL/well) at dilutions of 1/250, 1/500 and 1/1K were then added to the appropriate wells of the corresponding plate to assess the sensitivity of the antibodies at different dilutions. Then, 50 µl of CPF and PT standards at concentrations ranging from 0-1000 ng/mL that were prepared by serial dilution using phosphate buffer (0.1M pH7.2) were added to the appropriate wells of the plates before overnight incubation in the dark at 4°C. The plate was then washed four times with a wash buffer (0.0125% Tween 20 in 0.15M NaCl).

113

Chapter 3

Secondary antibodies—goat anti-rabbit or goat anti-mouse-horseradish peroxidase- labelled antibody previously diluted to 1/2000 in phosphate buffer (0.1M pH7.2)—were added to the appropriate plates, which were then incubated for 55 minutes at 37°C. The secondary antibodies were then discarded and the plates washed eight times with a wash buffer. Finally, TMB/E (100 µL/well) was added to each well and the reaction was stopped after five minutes using 25 µL/well of 2.5M sulphuric acid. The optical density for each well was read immediately at 450 nm using a TECAN Safire2 plate reader.

3.3.5 Microarray construction

Initially, each coating antigen was prepared for the first dilutions at concentrations of 20, 50, 100 and 200 µg/mL, respectively, by mixing a required volume of CPF-OVA or H11- OVA in PBS (0.1M;7.2 pH) to produce 200 µg/mL. Then, serial dilutions were prepared for the remaining concentrations, followed by the final dilution in filtered printing buffer (100 mM sodium phosphate, 50 mM sodium chloride, 100 μg/ml BSA, 0.005 % Tween- 20, pH 8.0) at a ratio of 1:1 (first dilution: printing buffer) to produce concentrations at 10, 25, 50, and 100 µg/mL, respectively. Then, BSA-647 (2 μg/ml) was prepared by the same procedure, whereby 4 µg/mL and 2 µg/mL were prepared for the first dilution in PBS and final dilution in a filtered printing buffer in the ratio of 1:1. A lower concentration of printing buffer was prepared by mixing 50% filtered printing buffer with 50% degassed water. The Scienion sciFLEXARRAYER S5 (Scienion AG, Berlin, Germany) (Figure 3-3) nanoplotter was then primed, programmed and utilised to spot 25 nL of BSA-647 and each coating antigen concentration onto chemically activated planar waveguides (MBio Diagnostics Inc., Boulder, Colorado, U.S.A) in a 2×22 array format.

114

Chapter 3

Figure 3-3: The Scienion sciFLEXARRAYER S5 nanoplotter.

After the nanoplotter was primed and programmed, the BSA-647 solution was first spotted at each of the four corners of the array in order to assist automatic array alignment by the SnapEsi reader (MBio Diagnostics Inc., Boulder, Colorado, U.S.A.). Then, the selected conjugate concentrations were spotted according to each designed plan and a printing buffer was used to spot the spacing locations within the arrays. All spots were printed at room temperature with 65% humidity. The printed waveguides were then left in the spotter for 1 hour to allow for the humidity to decrease before transferring them to a suitably controlled chamber at 29 °C and 30% humidity for overnight incubation. The waveguides were then placed in a plastic rack, secured with rubbers and shaken vigorously for 30 seconds in a protein-based blocking solution (casein, 5g; Tween-20, 0.5g; PBS×10, 100 mL and water, 900 mL), left to soak for 5 min and rinsed in deionised water. Afterward, the waveguides were dried by spinning in a Thermo Legend T Plus Centrifuge (Thermo Scientific) at 1000 rpm for 5 min at room temperature. All cartridge parts presented in Figure 3-4 were assembled by first placing the moulded upper housing on a specifically designed tool and then attaching the gaskets with double-sided adhesive to the printed waveguides using vacuum assistance to form a 5mm-wide fluidic channel. Then, the waveguides were inserted into the cartridge housing using 50-mm-long, 5-mm-

115

Chapter 3

wide and 0.145-mm-high flow channel dimensions before wicks, an absorbent pad, and lids were attached to complete the cartridge assembly. The cartridge inlet was designed to receive and reserve assay fluids flowing under gravity-assisted capillary action via the flow channel with the absorbent pad trapping any waste during the assay. A combination of capillary action and hydrostatic pressure allowed for consistent and negative fluid flow in the cartridges.

Figure 3-4: MBio components and an assembly tool. Cartridge components include (A) vacuum tool, (B) upper, (C) lid, (D) wick pad, (E) waveguide, (F) gaskets and (G) assembly press.

3.3.6 General MBio array protocol After the cartridge parts were assembled, they were placed on the MBio rack and reagents were prepared and mixed just prior to analysis. Then, 100 µL of each prepared standard concentrations/sample (CPF or PT) was independently mixed with 100 µL of each antibody (CPF-pAb or PT- mAb 1E2) dilution at 1:1 ratio to provide target pesticide-

116

Chapter 3

antibody mixtures. After that, 150 µL of each mixture was injected through a cartridge inlet and incubated for 5 min. Then, the same volume (150 µL) of each secondary antibody (i.e. goat-anti Rabbit-Alexa-647 or goat-anti mouse-Alexa-647) at a concentration of 10 µg/mL (1/200) prepared in MBio Assay Buffer (PBS, 1% BSA, 0.05% Tween20) was added and incubated for an additional 5 min. Each cartridge was then immediately read using the MBio SnapEsi Reader. All responses were determined before the data were exported into Microsoft Excel for processing.

3.3.7 Optimisation of coating antigen concentrations and antibody titre

For CPF, BSA-647 was evaluated for non-specific binding towards CPF-pAb before optimisation, whereby a concentration of 2µg/mL was spotted on the waveguides in replicates of ten by the procedure stated above. Then, three different solutions—a MBio assay buffer, a CPF-pAb, and a non-specific antibody raised to BSA prepared in a MBio assay buffer at 1/250 dilution—were applied to develop the assay using a MBio SnapEsi Reader in replicates. In addition, a chequerboard assay was evaluated using the planar waveguide platform in order to optimise and select the concentrations and titres of the immunological reagents applied. The coating antigen was spotted on the waveguides as described above. First, BSA-647 at 2 µg/mL was spotted as a control at all four corners of the planar waveguides. Then, 25 nL of 4 different concentrations (10, 25, 50, and 100 μg/mL) of each CPF-OVA and H11-OVA was spotted separately on each waveguide in replicates of four (Figure 3-5) and the remaining spacing was spotted with a printing buffer.

Figure 3-5: A representative schematic image showing a 2×22 array format and the array layout of different conjugates on each waveguide slide for optimsation.

117

Chapter 3

The cartridges were tested with two different antibody dilutions (1/1K and 1/5K) that were mixed independently with negative and positive CPF standards of 0 and 10 ng/mL. Then, the spotting procedure was repeated with the same four concentrations of CPF- OVA and was followed by testing higher antibody concentrations of 1/100 and 1/250 dilutions. Negative and positive CPF working standard solutions of 0, 100 and 1000 ng/mL were also used. With respect to PT, three antibody dilutions (1/100, 1/250 and 1/500) were mixed independently using negative and positive PT standards of 0, 100 and1000 ng/mL, respectively.

3.3.8 Calibration curves in buffer

After the assay protocol was optimised, a concentration of 100 µg/mL was selected for use with both CPF-OVA and H11-OVA, a dilution of CPF-pAb (1/100) and PT-mAb 1E2 (1/250) was selected, and calibration curves were developed. Coating antigens of CPF- OVA and H11-OVA were spotted in two different experiments in replicates of eight on each planar waveguide (Figure 3-6). Standard solutions of each pesticide were prepared in a MBio assay buffer at a 0-1000 ng /mL range (0, 0.5, 1, 5, 10, 25, 50, 100 and 1000 ng /mL) for CPF and 0-100 ng/mL range (0, 0.25, 0.625, 1.25, 2.5, 5, 12.5, 25, 50 and 100) for PT and were then mixed with a corresponding antibody dilution (CPF-pAb [1/100] or PT-mAb 1E2 dilution [1/250]) according to the developed MBio array protocol. For each curve, midpoint concentrations (IC50) and dynamic ranges (IC20 - IC80) were determined using BIA evaluation software 4.1.

Figure 3-6: Representative schematic for an array layout used for printing and devloping a calibration curve in buffer at 100 µg/mL of coating antigen.

118

Chapter 3

3.3.9 Specificity of antibodies in buffer Cross-reactivity (CR) was investigated for each assay (CPF-pAb and PT-mAb 1E2) of the ten structurally related OPP analytes. For each assay, chlorpyrifos, coumaphos, parathion, phoxim, triazophos, dichlofenthion, azinphos-ethyl, phosalone, parathion- methyl, cyanophos and disulfoton were tested at a concentration range of 0-10000 ng/mL (0, 5, 10, 50, 100, 1000 and 10000 ng/mL) following preparation in a MBio assay buffer (PBS, pH 7.2, 1% BSA, 0.05% w/v Tween 20). Under optimised conditions, each standard was mixed with each antibody dilution at a 1:1 ratio and was indivdidually analysed according to the developed MBio array protocols. The responses were used to calculate the midpoint concentrations (IC50) and the dynamic ranges (IC20 - IC80) of each pesticide calibration curve. Cross-reactivity was calculated as a percentage relative to the pesticide (CPF and PT midpoint) as follows:

CR (%) = (IC50 of analyte/IC50 of analogue) × 100

3.3.10 Sample preparation development and evaluation

Two batches of strawberry fruits sourced from different countries were purchased from local stores in Belfast and then chopped and homogenised in a blender (Oster, model 6805). Several aliquots of the homogenate were then secured in individual plastic bags and frozen at -20ºC. The general extraction procedure was adapted from Otieno et al. (2013) with minor modification: a 1g aliquot of the homogenised strawberry material was thawed and weighed in a plastic tube (15 mL). Then, 5 mL of methanol was added to unspiked samples and 4.95 mL to spiked samples, and the samples were vortexed and placed in a roller mixer (Dynalon, SRT9/120V/60 9-Roller Analog Tube Mixer with Fixed Speed, 120V-60Hz (IKA)) for 30 min before centrifugation (Thermo Legend T Plus Centrifuge, Thermo Scientific) at 4600 rpm for 10 min. Finally, the required volume of each extract was decanted for subsequent analysis. A different procedure for sample preparation was developed to extract each target; pure methanol was used for CPF and a similar extraction protocol with an extra dry step to achieve the required sensitivity for PT as discussed individually below for each target.

119

Chapter 3

3.3.10.1 Chlorpyrifos extraction procedure

Initially, smaller-scale experiments for two extraction procedures were assessed using 1) 100% methanol and 2) 20% methanol in MBio assay buffer in order to extract CPF from the strawberry samples before repeating the extraction process using only methanol. It was also necessary to assess appropriate dilutions of the strawberry extract at two dilution ratios (1:8 and 1:4) to evaluate the matrix effect of strawberry extract.

In the experiment, each extraction procedure used three samples—spiked tissue, spiked extract and a blank extract—and each group of three samples was spiked and similarly diluted but differently extracted.

Evaluation of 100% and 20% (MeOH) as extraction procedures at a 1:8 extract dilution

Six strawberry samples were prepared according to the extraction procedure stated above. Three samples were extracted by 100% and another three samples by 20% MeOH, which applied for each two samples individually. Two samples were spiked with a volume of 50µL of a CPF solution (4000 ng/mL) representing 200 ng/g, and each extract was diluted at a ratio of 1:8 (20µL:140µL; extract: MBio assay buffer). The remaining four samples were extracted in the same manner, and two samples were diluted as (125µL:875µL; extract: MBio assay buffer) before a volume (950 µL) of the dilution was spiked with 50µL of a CPF solution (100 ng/mL) to produce 5 ng/mL. The last two negative extract samples were diluted (20µL:140µL; extract: MBio assay buffer) and used as blank samples. All prepared samples were mixed individually at a 1:1 ratio using CPF-pAb dilution (1/100) for analysis.

Evaluation of 100% and 20% (MeOH) as extraction procedures at a 1:4 extract dilution Six strawberry samples were prepared. The first two samples were spiked with 50µL of a CPF solution (4000 ng/mL) representing 200 ng/g, and each extract was diluted at a ratio of 1:4 (40µL:120µL; extract: MBio assay buffer). For the subsequent spike, two extracts from the remaining four negative samples were diluted (250µL:750µL; extract: MBio assay buffer) and then 950µL of each dilution was spiked with 50µL of a CPF solution (200 ng/mL) equivalent to 10 ng/mL. The remaining two negative extracts were

120

Chapter 3

used as blank samples. The pairs of samples that were similarly prepared represent a different extraction procedure for each sample. In addition, CPF concentrations at 0, 5 and 10 ng/mL were prepared in MBio assay buffer for comparison. All prepared samples were mixed individually at a 1:1 ratio with CPF-pAb dilution (1/100) for analysis. However, the same extraction procedure and dilutions were repeated using only methanol as an extraction solvent, and two CPF concentrations (5 ng/mL and 50 ng/mL) were evaluated.

Re-evaluation of the extraction procedure by MeOH and a 1:8 extract dilution at higher CPF concentration

Six samples were prepared. First, two homogenate samples were spiked with 50µL (4000 ng/mL and 40,000 ng/mL) representing 200 ng/g and 2000 ng/g, respectively. After extraction in methanol, each extract was then diluted at a ratio of 1:8 (20µL:140µL; extract: MBio assay buffer). For the subsequent sample spike for spike after, two negative extracts were diluted (125µL:875µL; extract: MBio assay buffer) and then 950µL of each dilution was spiked with 50µL (100 ng/mL and 1000 ng/mL) equivalent to 5 ng/mL and 50 ng/mL. For the remaining two negative samples, each extract was diluted (20µL:140µL; extract: MBio assay buffer) and used as a blank. Similar CPF concentrations at 0, 5 and 50 ng/mL were prepared in a MBio assay buffer. All prepared samples were mixed individually at a 1:1 ratio with CPF-pAb dilution (1/100) for analysis.

Re-evaluation of the extraction procedure by MeOH methanol and a 1:4 extract dilution

Similarly, six samples were prepared, and two homogenate samples were spiked with 50µL (2000 and 20,000 ng/mL) representing 100 ng/g and 1000 ng/g, respectively. Following extraction in methanol, each extract was diluted at a ratio of 1:4 (40µL:120µL; extract: MBio assay buffer). For the remaining four samples, two negative extracts were diluted (250µL:750µL; extract: MBio assay buffer) and 950µL of each dilution was spiked with 50µL (100 ng/mL and 1000 ng/mL) to produce 5 ng/mL and 50 ng/mL. The remaining negative extracts were diluted (250µL:750µL; extract: MBio assay buffer) and used as blank samples. Similar concentrations of CPF 0, 5 and 50 ng/mL were prepared

121

Chapter 3

in MBio assay buffer, and all prepared samples were mixed individually at a 1:1 ratio with CPF-pAb (1/100) for analysis.

Examination of matrix effects and assay recovery

In order to examine the effects of the strawberry matrix on the assay and recovery of CPF, calibration curves were generated using MBio assay buffer and for extracted strawberry, spiked tissue and strawberry extracts at a concentration range of 0-2000 ng/g or equivalent. To generate calibration curves, 50 µL of required CPF standard solutions dissolved in methanol were used to fortify the sample aliquots before and after extraction. The developed curves for the spikes before and after extraction were then examined relative to the buffer calibration curves. The comparison of the samples spiked after extraction compared to those of the buffer curves revealed any matrix effects, and the comparison of the samples spiked before and after extraction will illustrate assay recovery. Fourteen homogenised strawberry samples (1g) free of CPF residues were prepared, and seven samples were spiked with 0, 50, 100, 200,400, 1000 and 2000, ng/g) for tissue spike. Following the extraction by methanol, the strawberry extracts of each sample spiked before (spiked tissues) were then diluted at a ratio of 1:8 (20µL:140µL; extract: MBio assay buffer). For the spike after, each blank extract of the remaining seven samples was diluted (125µL:875µL; extract: MBio assay buffer) and then 950 µL of each dilution was spiked with 50 µL of equivalent concentrations (0, 1.25, 2.5, 5, 10, 25 and 50 ng/mL) for the spike extract. In addition, another curve was prepared in MBio assay buffer at an equivalent concentration. A CPF-pAb dilution of 1/100 was prepared and mixed with each standard/sample at a 1:1 ratio, and 150 µL of each sample/antibody mixture was used to develop the assay. However, assay sensitivity was reduced during this validation experiment, which was required to modify the dilution step. As a result, a lower strawberry extract dilution of 1:4 was used at a concentration range of 0-2000 ng/g. For this experiment, 16 samples of homogenised strawberry (1g) were prepared, eight of which were spiked beforehand with 0, 25, 50, 100, 200, 400, 1000 and 2000, ng/g. Following extraction in methanol, each extract was then diluted at a ratio of 1:4 (300µL:900µL; extract: MBio assay buffer). However, for the extract spike after, each negative extract was first diluted by the same ratio and then 950µL of each dilution was spiked with a 50 µL of CPF equivalent concentration (0, 1.25, 2.5, 5, 10, 20, 50 and 100

122

Chapter 3

ng/mL). Another curve was prepared in a MBio assay buffer using equivalent concentrations. Finally, the same MBio array protocol of mixing CPF-pAb dilution of 1/100 for each sample (1:1) and 150 µL of each sample/antibody mixture was used to develop the assay.

Planar Waveguide Assay for CPF

Planar waveguides were prepared with 25 nL of BSA-647 (2 μg/ml) at all four corners, and CPF-OVA (100 μg/mL) in replicates of eight; the cartridges were built according to the previously described procedure. After the CPF-pAb and sample mixtures were incubated for 5 min, a GaR secondary antibody was added, and the cartridges were again incubated at room temperature for another 5 min. Fluorescence response was obtained by reading the cartridge on a MBio SnapEsi reader and recording the data on a laptop. Then, the data was exported to excel and the mean values and standard deviations (±SD) of the results were calculated. The midpoint concentration (IC50) and the dynamic range (IC20 -

IC80) of each calibration curve (buffer, spiked tissue and spiked extract) were determined.

3.3.10.2 Parathion extraction procedure

To examine the effects of the strawberry matrix on assay performance and PT recovery, calibration curves were prepared in a MBio assay buffer for extracted strawberry, spiked tissue and extracts at a concentration range of 0-500 ng/g or equivalent. To generate calibration curves, 50 µL of a required PT standard solution in methanol was used to fortify each sample aliquot prior to and after extraction. The fortified samples were then examined relative to the buffer calibration curves. The comparison of the samples fortified after extraction to the buffer curves illustrated the matrix effects, and the comparison of the samples fortified before and after extraction showed the recovery of the assay. Seven homogenised strawberry samples (1g) free of PT residues were then spiked with 0, 10, 25, 50, 100, 200 and 500 ng/g for spiking tissues and for equivalent concentrations (0, 0.5, 1.25, 2.5, 5, 10 and 25 ng/mL) for spiking extracts after extraction, respectively. To simplify this process, methanol was used to extract PT directly from the samples, and the strawberry extracts were first evaluated at two dilutions (1:4 and 1:8) to be comparable to the chlorpyrifos assay.

123

Chapter 3

Evaluation of 100% (MeOH) as extraction procedure and a 1:4 extract dilution

For the extract dilution at the 1:4 ratio, seven homogenate samples were fortified at 0, 10, 25, 50, 100, 200 and 500 ng/g and an additional seven blank samples were prepared. Following extraction, the required volumes were decanted and all positive extracts were first diluted to achieve a 1:4 ratio (40 µL:120µL; extract: MBio assay buffer). Then, only the negative extracts for the curve spiked tissue were diluted (250 µL:750µL; extract: MBio assay buffer) and 950 µL of each diluted sample was fortified with 50 µL of the equivalent concentrations (0, 0.5, 1.25, 2.5, 5, 10 and 25 ng/mL). Further, an additional curve was prepared in a MBio assay buffer at equivalent concentrations. Finally, a PT- mAb 1E2 dilution of 1/250 was prepared and mixed with each sample at a 1:1 ratio. 150 µL of each sample/antibody mixture was used to develop the assay, and another curve was prepared in a MBio assay buffer with similar concentrations.

Evaluation of 100% (MeOH) as extraction procedure and a 1:8 extract dilution

The conditions were adjusted and the experiment was repeated in order to evaluate a higher extract dilution. The same procedure described above was used, with the only difference being an extract dilution at a 1:8 ratio and a calibrant increased from 7 to 10 points to cover the concentration range 0-4000 ng/g for spiking tissues before extraction and the equivalent of 0 -100 ng/mL for spiking extracts after extraction. Twenty samples were prepared, the first ten of which were fortified with 50 µL of 0, 10, 25, 50, 100, 200, 500, 1000, 2000 and 4000 ng/g. The remaining ten samples used for spiking extracts after extraction and the same extraction procedure was followed. Each fortified extract was diluted at a 1:8 ratio (extract: MBio assay buffer; 20µL:140µL), the negative extracts were diluted first at 1:8 (extract: MBio assay buffer; 125 µL:875µL), and 950 µL of each diluted extract was spiked with 50 µL of the equivalent concentrations (0, 0.25, 0.625, 1.25, 2.5, 5, 12.5, 25, 50 and 100 ng/mL). Then, PT-mAb 1E2 dilutions of 1/250 were prepared and mixed with each sample at a 1:1 ratio. Finally, 150 µL of each sample/antibody mixture was used to develop the assay, and an additional curve at equivalent concentrations (0-100 ng/mL) was prepared in a MBio assay buffer.

The extract drying and reconstitution step

The sample preparation procedure was again altered, and an additional drying step was added to concentrate the sample to improve assay sensitivity. The initial evaluation was 124

Chapter 3

conducted by preparing two batches of strawberry homogenate, and each batch contained three samples (blank, spiked tissue (before extraction) and spiked extract (after extraction)). The extraction procedure was conducted as follows. Six homogenate samples were prepared, two of which were fortified with 50 µL of PT (50 ng/g) for the spike before, two of which using an equivalent concentration (10 ng/mL) for the spike after (dry extract), and the remaining two were prepared without PT to serve as blanks. Following the extraction of all samples, 1 mL of each strawberry extract was decanted, transferred into a 5 mL glass tube, placed in a TurboVap evaporator (Caliper Life Sciences, Hopkinton, MA, USA) and dried at 45 °C for 90 minutes under a nitrogen stream. Two volumes (200 µL and 500 µL) of MBio assay buffer were used to reconstitute the dry residues for the first and second batches, respectively. An additional two samples of PT concentration (10 ng/mL) were prepared in two volumes using a MBio assay buffer (200 and 500 µL) as a blank in order to mimic the experimental conditions. Finally, a PT- mAb 1E2 dilution (1/250) was prepared and mixed with each sample at a 1:1 ratio, and 150 µL of each sample/antibody mixture was used to develop the assay.

Examination of matrix effects and assay recovery

A final optimised volume (500 µL) of MBio assay buffer was used to develop full calibration curves in a MBio assay buffer for extracted strawberry, spiked tissue and extract samples. Calibration curves were required to again examine the effects of the strawberry matrix on assay performance, and fortified sample curves before and after extraction were examined relative to the buffer calibration curves. Sixteen homogenised strawberry samples (1g) were prepared, eight of which were spiked with 50 µL of PT (0, 5, 10, 25, 50, 100, 200 and 400, ng/g) for spikes before, and the remaining eight equivalent concentrations (0, 1, 2, 5, 10, 20, 40 and 80 ng/500 µL) were used to spike the negative dry residue samples for spikes after. Following extraction, drying and reconstitution in 500 µL, a PT-mAb 1E2 dilution (1/250) was prepared and mixed with each sample at a 1:1 ratio. Finally, 150 µL of each sample/antibody mixture was used to develop the assay, and another equivalent concentration of 0-80 ng/500 µL was prepared to develop a curve in an MBio assay buffer.

125

Chapter 3

Planar waveguide assay for PT

Planar waveguides were prepared and 25 nL of labelled-BSA-647 (2 μg/ml) at all four corners, H11-OVA (100 μg/mL) in replicates of eight were spotted and cartridges were built as described previously. After the PT-mAb 1E2 dilution (1/250) was prepared and mixed with each sample (1:1), 150 µL of each sample/antibody mixture was used to develop the assay. After a five-minute incubation period, a GaM secondary antibody was added and the cartridge was incubated at room temperature again for an additional five minutes. Fluorescence responses were obtained by reading the cartridge and recording the data using MBio SnapEsi reader and a laptop, and mean values and standard deviations (±SD) of the results were calculated. In addition, the midpoint concentrations

(IC50) and dynamic ranges (IC20 - IC80) of each calibration curve (buffer, spiked tissue and spiked extract) were determined.

3.3.11 Validation study

For both target detections, the optimised immunoassay methods were validated based on the guidelines of 657/2002/EC (Commission Decision, 2002) in relation to a semi- quantitative method. For each experiment, calibration curves were prepared for strawberry samples at 0-2000 ng/g (0, 50, 100, 200,400, 1000 and 2000 ng/g) for CPF and 0-400 ng/g for PT (0, 5, 10, 25, 50, 100, 200 and 400 ng/g) and in buffer with equivalent concentrations. One standard curve was prepared in a MBio assay buffer as a reference, and the remaining two curves were produced by fortifying strawberry samples prior to (tissue) and after extraction (extract), and dry residues in the case of PT.

In addition, strawberry samples were fortified with three different concentrations (0, 100, 200, and 300 ng/g) for CPF and 0, 25, 50 and 75 ng/g for PT (n=10). The samples were prepared for calibration curves to include 40 samples for different concentrations. Following the extraction steps of adding methanol, mixing, and centrifugation, each sample (1g) was fortified according to the previously discussed procedure. The samples were then extracted according to the developed procedures for each target, at 1:4 ratio extract dilution for CPF, and drying the extract and reconstitution in 500 µL MBio assay buffer for PT. After extraction, each MBio array protocol was used accordingly by mixing

126

Chapter 3

individually standards/samples with antibody dilutions at a 1:1 ratio for subsequent MBio reader analysis.

For CPF, sensitivity, dynamic range, recovery, repeatability (intra-day) and reproducibility (inter-day) analyses and limits of detection and quantification were conducted; all of these aspects except reproducibility were conducted for PT. In addition, standard deviation (SD) and coefficient of variation (CV%) values were determined for both targets at each concentration. However, due to limited availability of reagents, only two experiments were analysed over two days for CPF and one experiment was carried out for PT.

3.3.11.1 Limits of detection (LOD) and limits of quantification (LOQ)

LOD and LOQ were determined for each pesticide by calculating the concentrations of 20 blank samples (over two experiments) for CPF and 10 blank samples for PT. The LOD and LOQ were calculated after all responses were interpolated from the extracted curves (spiked tissue) and then each average concentration was reported using the following equations:

LOD = average concentration + (3×Standard deviation)

LOQ = average concentration + (10× Standard deviation)

3.3.11.2 Intra and inter -day reproducibility

Intra and inter-day precision were determined by analysing the strawberry extract samples that were fortified with three different concentrations (100, 200, and 300 ng/ for CPF and 25, 50 and 75 ng/g for PT). The samples were fortified with a concentration exceeding the LOD at 200 ng/g, which is equivalent to the MRL stipulated in EC legislation for strawberries and other crops. The fortification of CF at 100 ng/g and 300 ng/g were lower and higher than CPF regulatory limits by 50%. Again, a fortification of strawberry samples with a concentration above the LOD at 50 ng/g was conducted, which is equivalent to the MRL stipulated by EC regulations for strawberries and other crops. The fortification of PT at 25 ng/g and 75 ng/g were lower and upper than PT regulatory limit by 50%. The resulting variance among replicates assayed at several concentrations within the standard curve over one day shows the level of intra-day precision, while the variance

127

Chapter 3

among different assays (batches) conducted on different days illustrates inter-day precision. Calculated concentrations were interpolated from the known curve results of the spiked tissue samples that were developed on the same day for each experiment.

3.3.11.3 Recovery

CPF and PT were spiked in strawberry samples (n=20) at concentrations of 100, 200 and 300 ng/g (n=10) and 25, 50 and 75 ng/g, respectively. The samples were then extracted following each extraction procedure and were analysed using an MBio reader. Recovery values were identified by interpolating the response of each sample from the extracted calibration curve (spiked before) and compared with the actual fortification levels, expressed as a percentage according to the following equation:

Re%= (Fc/Ic)*100 in which Re represents recovery percentage, Ic and Fc represent initial concentration and final concentration respectively.

3.4 Results and discussion

This work illustrates the potential of using MBio biosensor technology to detect the harmful insecticides CPF and PT. The quality and safety of agricultural products have become increasingly important as consumer concerns about the contamination of food products with pesticides used in the food chain have grown. For monitoring and control purposes, a rapid, sensitive and portable system that can be utilised in the field to prevent and mitigate pesticide contamination is highly appealing. A competitive inhibition assay has been shown previously to be effective in identifying small molecular weight molecules (i.e. less than 900 dalton) (McNamee et al., 2014) using the MBio SnapEsi Reader, which was deemed suitable in feasibility studies that detect the low molecular weights of pesticides such as CPF and PT. For this procedure, pesticide and corresponding coating antigens compete for available sites in antibodies (PT.mAb 1E2 or CPF.pAb), and antibody inhibition increases as the standard concentration increases in a medium. Thus, pesticide concentrations can be indirectly quantified using fluorescently-labelled secondary antibodies. Further, if fewer antibodies are bound to a surface coating antigen,

128

Chapter 3

the fluorescent dye generates a weaker signal as a result of the presence of higher concentrations of pesticides.

3.4.1 Sample screening

The UPLC-MS/MS was tuned to detect both targets. Figure 3-7 and Figure 3-8 show the chromatogram results of CPF and PT detection, respectively, and the run times using a methanol solution. Retention times of 4.67 and 4.04 min, respectively, were observed and MS transitions of 350.00 > 198.00 and 352 > 200 for CPF and 292.10 > 236.05 and 292.10 > 264.10 for PT were used for quantification. Figure 3-9 and Figure 3-10 show CPF and PT calibration curve results for preparation in the strawberry matrix with an R2 value of 0.99.

161026_AS_007 3: MRM of 6 Channels ES+ TIC (CPF) 3.99e6

0 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 161026_AS_007 2: MRM of 4 Channels ES+ Figure 3-7: Chromatogram of CPF detection using UPLC-MS/MS in methanol4.04 . TIC (Parathion) 2.00e6

0 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 161026_AS_007 1: MRM of 2 Channels ES+ TIC (Glyphosate) 6.66e5 129

0.41

0 Time 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 161026_AS_007 3: MRM of 6 Channels ES+ TIC (CPF) 3.99e6

Chapter 3

0 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 161026_AS_007 2: MRM of 4 Channels ES+ 4.04 TIC (Parathion) 2.00e6

0 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 161026_AS_007 1: MRM of 2 Channels ES+ Figure 3-8: Chromatogram of PT detection using UPLC-MS/MS in methanol.TIC (Glyphosate) 6.66e5

0.41

Compound name: CPF Correlation coefficient: r = 0.999154, r^2 = 0.998309 Calibration curve: 420.06 * x + 280.401 Response type: External Std, Area Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None

80000

70000 0 Time 0.5060000 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 50000

40000 Response 30000

20000

10000

-0 ng/mL -0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 Figure 3-9: CPF calibration curve for strawberry extracts (0-200 ng/mL).

130

Chapter 3

Compound name: Parathion Correlation coefficient: r = 0.999933, r^2 = 0.999865 Calibration curve: 125.009 * x + 26.2868 Response type: External Std, Area Curve type: Linear, Origin: Exclude, Weighting: 1/x, Axis trans: None

25000

20000

15000

Response 10000

5000

-0 ng/mL -0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180 190 200 Figure 3-10: PT Calibration curve for strawberry extracts (0-200 ng/mL).

The analytical results of both strawberry batches generated by UPLC-MS/MS under current conditions confirmed that the strawberry samples were free from CPF and PT residues. The results also showed that the standard concentration of the lowest spiked extract—2.5 ng/g, equivalent to 0.5 ng/mL—had two peaks as qualifier and quantifier, while the blank strawberry extract had no detectable residues of CPF or PT, as both a qualifier and a quantifier were not detected. The mass chromatograms are depicted in Figure 3-11 show no distinctive peak for CPF or PT, but rather only baseline noise.

131

Chapter 3

Figure 3-11: Screening of strawberry blank extract showing no distinctive peaks for A) CPF at 4.67 min and B) PT at 4.04 min.

3.4.2 Initial characterisation

Many factors can affect immunoassay sensitivity. However, the two main issues are poor antibody-antigen interactions and increases in non-specific binding. As a result, after the immunological reagents were obtained, the coating antigens (CPF-OVA and H11-OVA) were characterised using UV-Vis Nanodrop spectroscopy against the pesticide standards and OVA in order to confirm the conjugation. ELISA was also used to evaluate sensitivity and determine the quality of the reagents.

3.4.2.1 UV-VIS Spectrophotometer

Figure 3-12 shows the absorption spectrums resulting from UV-VIS spectroscopy trials for CPF, CPF-OVA, PT, H11-OVA and the protein control (OVA) in the range of 220– 750 nm, and Figure 3.13 shows the results for PT. The specific absorbance contributions of CPF were detected at two ultraviolet functional groups at 229 nm and 289 nm, and the contributions of CPF-OVA was observed at 223 nm and 265 nm. With respect to PT, the

132

Chapter 3

specific absorbance contributions were detected at 223 and 283 nm, and for H11-OVA, the contributions were observed at 226 and 268 nm. However, the results for the carrier protein (OVA) showed either a small or no specific absorbance, and 226 nm was detected as the highest absorbance value. In both experiments, an absorbance change between the conjugates (CPF-OVA and PT-OVA) and OVA was evident and the shift in wavelengths determined the course of the conjugation.

4 CPF (1 mg/mL) 3 CPF.OVA (3 mg/mL) OVA (1 mg/mL) 2

Absorbance (AU) 1

0 200 250 300 350 400

Wavelength (nm)

Figure 3-12: Comparison of ultraviolet absorption spectrums of CPF standard, CPF-OVA and OVA from 220 nm to 400 nm.

133

Chapter 3

8 7 6 PT (1mg/mL) 5 H11-OVA (0.5mg/mL) 4 OVA (1mg/mL) 3

Absorbance (AU) 2 1

0 200 250 300 350 400

wavelength (nm)

Figure 3-13: Comparison of ultraviolet absorption spectrums of PT standard, H11-OVA and OVA showing the wavelength from 220 nm to 400 nm.

In addition, protein concentrations of CPF-pAb and PT-mAb 1E2 were measured at 280 nm and found to be 3mg/mL and 39.87mg/mL, respectively. Finally, the measurement procedure was relatively rapid and convenient and required no additional reagents, incubation time, or standard curves.

3.4.2.2 Chequerboard ELISA for determination of optimal ratios

Chlorpyrifos-coating antigen and anti-CPF antibody dilution

Chequerboard titration is a technique that is used to assess an antibody titre and identify the range and optimum antibody/antigen dilutions or concentrations for ELISA development. Table 3-2 presents the results of the chequerboard analysis using CPF-OVA coating antigen concentrations (0.25, 0.5 and 1 µg/100 µL) against serial CPF-pAb dilutions with negative and positive CPF concentrations at 100 ng/mL. These results show average OD responses of approximately 0.5 for negative samples at all coating concentrations. Therefore, 1/1K or lower (antibody dilution) and 0.25 µg/100µL (coating antigen) were selected to determine antibody sensitivity. In addition, the protein controls showed high signals for BSA, indicating binding of CPF-pAb, as BSA was the carrier protein for the immunogen and a zero response for OVA which was used to couple CPF pesticide to produce the CPF-OVA (conjugate), which suggests that no binding occurred between CPF-pAb and OVA.

134

Chapter 3

Table 3-2: OD responses of checkerboard ELISA titration. Coating CPF Concentrations 1/1k SD 1/2k SD 1/4k SD 1/8k SD 1/16k SD 1/32k SD 100 (µg/100µL) ng/mL 0.44 0.19 0.27 0.01 0.21 0.00 0.17 0.04 0.13 0.00 0.13 0.00 neg 1 0.30 0.08 0.19 0.00 0.19 0.00 0.15 0.04 0.11 0.00 0.10 0.01 pos 0.47 0.02 0.26 0.02 0.19 0.00 0.15 0.02 0.11 0.00 0.09 0.01 neg 0.5 0.27 0.00 0.18 0.00 0.16 0.01 0.11 0.02 0.09 0.00 0.09 0.02 pos 0.42 0.07 0.20 0.00 0.16 0.00 0.11 0.03 0.08 0.01 0.10 0.03 neg 0.25 0.21 0.02 0.15 0.01 0.15 0.02 0.12 0.01 0.08 0.00 0.07 0.01 pos 2.62 0.22 1.03 0.05 0.30 0.01 0.09 0.01 0.09 0.01 0.05 0.01 neg Protein 2.84 0.10 1.20 0.12 0.34 0.02 0.10 0.00 0.07 0.01 0.06 0.00 pos BSA 1 BSA 0.5 BSA 0.25 OVA 1 OVA 0.5 OVA 0.25

After the selection of the optimum conjugate concentration (0.25 µg/100µL) and three antibody dilutions (1/250, 1/500 and 1/1K) were determined the sensitivity evaluation results for the indirect competitive assay against CPF concentrations (0-1000 ng/mL) were calculated. Table 3-3 presents these results including average OD and SD.

Table 3-3: Average OD responses at different CPF-pAb dilutions against CPF (0-1000). CPF (ng/mL) CPF-pAb dilution SD CPF-pAb dilution SD CPF-pAb dilution SD (1/250) (1/500) (1/1000) 0 1.16 0.06 0.48 0.05 0.27 0.01 1 1.18 0.02 0.48 0.01 0.28 0.01 5 1.16 0.05 0.44 0.02 0.22 0.02 10 1.15 0.04 0.40 0.06 0.16 0.02 50 0.75 0.09 0.24 0.01 0.12 0.01 100 0.57 0.04 0.20 0.01 0.10 0.00 500 0.32 0.03 0.16 0.01 0.07 0.01 1000 0.23 0.02 0.10 0.01 0.06 0.00

IC50 (ng/mL) 64.17 32.92 13.69 Dynamic range 25.1-174.9 9.1-136.8 3.6-63 (C20-C80) (ng/mL)

The results show that only a 1/250 dilution at the selected CPF concentration range was found to be optimum, which was the highest OD response range from 1.16 to 0.23 arbitrary units (AU). However, dilution at 1/500 and 1/1K produced very weak responses that fell outside the optimum range. In addition, Figure 3-14 shows the results of the percentage normalised responses versus the CPF concentrations for each calibration curve. A comparison of these curves shows that sigmoidal inhibition curves were produced varying the assay sensitivity of three CPF-pAb dilutions. It was anticipated that the most sensitive icELISA immunoassay would be the highest antibody dilution of 1/1K, but the signal proved insufficient. Each calibration curve illustrates the concentration

135

Chapter 3

range of 0-1000 ng/mL, which allowed for the assay sensitivity (IC50) for CPF to be calculated. The midpoint concentration that allows for 50% inhibition were 64.2, 32.9 and 13.7 ng/mL at the dilutions of 1/250, 1/500 and 1/1K respectively. For this study, the theoretical LODs were calculated as a 20% inhibitory concentration (IC20) and were found to be 25.1, 9.1 and 3.6 for 1/250, 1/500 and 1/1K dilutions, respectively. Nevertheless, the reagents were shown to be functional for biosensor development for the antibody/antigen interactions measured.

120

100 CPF.pAb dilution (1/250)

80 CPF.pAb dilution (1/500)

CPF.pAb dilution (1/1000) 60

Normlised data (%) 20

0 0.1 1 10 100 1000 10000

CPF Concentration (ng/mL)

Figure 3-14: Standard calibration curves for CPF illustrating percent normalised response values versus CPF concentrations in icELISA format for CPF.pAb. Each value represents the mean of two replicates in buffer solutions. Plates were coated with 100 µL of 0.25 µg/100µL of CPF-OVA.

Parathion - coating antigen and ant- PT-mAb 1E2 antibody dilution Table 3-4 presents the results obtained by checkerboard titration when a concentration of H11-OVA (0.25 µg/100µL) coating antigen was used against serial PT-mAb 1E2 dilutions 1/250, 1/500, 1/1K, 1/2K, 1/4K and 1/8K with negative and positive PT concentrations at 100 ng/mL. After the plates were measured, the average OD response results showed a high signal at the lowest antibody dilution (1/250), and the responses gradually decreased until the highest dilution for negative samples. However, the protein

136

Chapter 3

control results showed high signals for the control protein, BSA, indicating binding of PT-mAb 1E2 which was used as a carrier protein, immunogen, for animal immunisation and a zero response for OVA which was used to couple PT and produced the H11-OVA (conjugate), meaning there was no binding occurring between PT-mAb 1E2 and OVA.

Table 3-4: OD responses for checkerboard ELISA titration. Coating PT 100 1/250 SD 1/500 SD 1/1k SD 1/2k SD 1/4k SD 1/8k SD concentrations ng/mL 2.62 0.22 2.69 0.40 2.51 0.25 1.71 0.08 0.63 0.12 0.32 0.25 neg 0.25 0.35 0.05 0.20 0.01 0.12 0.00 0.08 0.01 0.08 0.00 0.09 0.07 pos Protein 1.09 0.05 0.17 0.02 0.08 0.01 0.07 0.01 0.08 0.01 0.09 0.09 neg 1.19 0.15 0.23 0.00 0.10 0.00 0.09 0.01 0.08 0.00 0.08 0.08 pos BSA 1 BSA 0.5 BSA 0.25 OVA 1 OVA0.5 OVA 0.25

As a result, antibody dilutions of 1/2K, 1/4K, 1/8K and 1/32K and 2.5 µg/mL (conjugate coating) were selected to develop the inhibition curves for the following experiment to determine Ab sensitivity. Table 3-5 presents the inhibition curve results and the average

OD responses and SD for each dilution, and shows IC50 values with different sensitivities and the calculated values of the dynamic ranges.

Table 3-5: Average OD responses at different PT-mAb dilutions against PT (0-1000). PT.mAb PT.mAb PT.mAb PT.mAb PT.mAb PT (ng/mL) dilution SD dilution SD dilution SD dilution SD dilution SD (1/2K) (1/4K) (1/8K) (1/16K) (1/32K) 0 2.52 0.19 2.01 0.20 1.54 0.11 0.93 0.02 0.51 0.00 1 2.59 0.30 1.85 0.05 1.28 0.02 0.56 0.00 0.29 0.00 5 1.45 0.01 0.82 0.10 0.46 0.01 0.21 0.00 0.12 0.00 10 0.93 0.08 0.46 0.00 0.28 0.00 0.13 0.00 0.08 0.00 50 0.28 0.02 0.16 0.01 0.11 0.00 0.07 0.00 0.05 0.00 100 0.24 0.05 0.11 0.01 0.07 0.00 0.06 0.00 0.05 0.00 500 0.07 0.00 0.05 0.00 0.05 0.00 0.05 0.00 0.05 0.00 1000 0.06 0.01 0.05 0.00 0.06 0.00 0.05 0.00 0.05 0.00 IC50 (ng/mL) 5.74 3.78 2.64 1.31 1.09 Dynamic range IC20-IC80 2.69-12.56 1.65-8.79 1.09-6.61 0.29-4.19 0.18-3.93 (ng/mL)

In addition, the results of the percent-normalised responses versus PT concentrations for each dilution are shown in Figure 3-15. These results show that all PT-mAb 1E2 dilutions produced sigmoidal curves. Each calibration curve illustrates the concentration range

between 0-1000 ng/mL, allowing for PT assay sensitivity (IC50) to be calculated for actual concentrations (ng/mL). The signal produced by each dilution confirmed antibody binding, and signal inhibition increased as the PT-mAb 1E2 dilutions increased. The

137

Chapter 3

reagents were confirmed to be effective for biosensor development for the Ab/Ag interactions measured.

120 PT.mAb 1E2 (1/2k) 100 PT.mAb 1E2 (1/4k) PT.mAb 1E2 (1/8k) 80 PT.mAb 1E2 (1/16k) PT.mAb 1E2 (1/32K) 60

Normlised data (%) 20

0 0.1 1 10 100 1000 10000

Parathion Concentration (ng/mL)

Figure 3-15: Calibration curves of PT illustrating percent normalised responses versus PT concentrations in icELISA format for PT-mAb 1E2. Each value represents the mean of two replicates in buffer solutions. Plates were coated with 100 µL of 0.25 µg/100µL of H11-OVA.

3.4.3 MBio assay protocol

The current study illustrates the first time an attempt to use MBio technology to detect the OPPs CPF and PT by immunoassay using CPF-pAb and PT-mAb 1E2.

3.4.3.1 MBio assay optimisation

It was necessary to confirm the absence of non-specific binding by different solutions against the fluorescently-labelled dye (BSA-647) that could have influenced the immunoassay results. The results of testing the CPF-pAb (raised using BSA as a carrier protein) and the MBio assay buffer showed responses that were similar and that fell within the same range for all solutions. These results confirmed that no non-specific binding occurred. The CPF-OVA conjugate concentrations were immobilized onto the planar waveguides by the spotter and were optimised to generate fluorescent responses from the blank (negative samples) between 150 – 300 AU.

138

Chapter 3

Chlorpyrifos For the initial titration, antibody dilutions of 1/1K and 1/5K were used against the concentration of 0 and 10 ng/mL, but the responses generated were weak and fell below 25 AU. Thus, the experiment was repeated with lower CPF-pAb dilutions (1/100 and 1/250) against higher CPF concentrations (0, 100 and 1000 ng/mL). This resulted in a significant increase in the responses of both dilutions (252.5 and 143.3 AU at 1/100 and 1/250 CPF-pAb dilution, respectively) for the blanks (negative samples). All responses were plotted against CPF concentrations for both CPF-pAb dilutions (Figure 3.16) and show that the responses from the negative sample increased as conjugate concentrations (CPF-OVA) increased.

300 CPF.OVA (10 µg/mL) 250 CPF.OVA (25 µg/mL)

200 CPF.OVA (50 µg/mL) CPF.OVA (100 µg/mL) 150

100

Fluorescence intensity 50

0 0 100 1000 Chlorpyrifos concentrations (ng/mL)

139

Chapter 3

150 CPF.OVA (10µg/mL) CPF.OVA (25 µg/mL) CPF.OVA (50 µg/mL) 100 CPF.OVA (100 µg/mL)

Fluorescence intensity

0 0 100 1000 Chlorpyrifos concentrations (ng/mL)

Figure 3-16: Results of different concentrations of coating antigens against an individual mixture of CPF standard and CPF-pAb dilutions, (A) 1/100 and (B) 1/250.

Clearly, Figure 5-16 shows that all conjugates at both CPF-pAb dilutions exhibited responses of less than 150 AU. The only exception appeared at the conjugate concentration of 100 µg/mL in which CPF-pAb dilution (1/100) was used and the responses fell within a range of 150 to 300 AU (252.5 AU). As a result, these optimised immunoreagent parameters were used to develop a calibration curve and other studies in buffer and strawberry matrix.

Parathion

Figure 3-17 shows the results of initial antibody titrations using antibody dilutions of 1/250 and 1/500 against PT standards 0, 100 and 1000 ng/mL (Section 3.3.7).

140

Chapter 3

200 H11-OVA (10 µg/ml)

H11-OVA (25 µg/ml) 150 H11-OVA (50 µg/ml) H11-OVA (100 µg/ml) 100

Fluorescence intensity

0 100 1000 Parathion concentrations (ng/mL)

150 H11-OVA (10 µg/ml) H11-OVA (25 µg/ml)

100 H11-OVA (50 µg/ml) H11-OVA (100 µg/ml)

Fluorescence intensity

0 100 1000 Parathion concentration (ng/mL)

Figure 3-17: Results from different conjugate concentrations against PT standard concentrations of 0, 100 and 1000 ng/mL. (A) Antibody dilution at 1/250, and (B) antibody dilution at 1/500.

The above figures clearly show that almost all conjugates at both antibody dilutions exhibited responses of 110 AU and lower (ranging between 9 and 110 AU), except at the conjugate concentration of 100 µg/mL, for which an antibody dilution of 1/250 was used

141

Chapter 3

(Figure 5.17, A). The average response reading of the four replicates was 172 AU for the blank sample, which fell within the range of 150-300 AU. Therefore, these optimised immunoreagent conditions (1/250 for Antibody dilution and 100 μg/mL as a conjugate concentration) were used to develop a calibration curve and to carry out additional studies on buffer and fruit matrix to evaluate PT detection.

3.4.4 Development of calibration curves using the MBio Biosensor

Ideally, in developing methods to detect contaminants, a calibration curve is produced in a buffer as an initial evaluation step prior to development in a matrix. The reason for this is to minimize any confounding effects that may present with a real sample relative to buffer in a similar condition, such as interference from cross-reactants, non-specific binding and matrix effects, sample pH or ionic strengths, or other substances that can interact with the target analyte or antibody detection. Therefore, for this study the curve was initially developed in an MBio assay buffer (Section 3.3.8) by preparing nine calibrants within the range and ten calibrants to show a sigmodal curve. (Figure 3-18 and Figure 3-19), respectively, for CPF and PT results). Furthermore, the sensitivity of this method was calculated as a midpoint (IC50), and was found to be 6 ng/mL. The assay dynamic range (IC20 - IC80) was determined as 0.9 to 46 ng/mL, respectively, with a standard deviation (n=6) for each calibrant of between 1.9 – 13 AU for CPF. For PT, the midpoint (IC50) and the assay dynamic range (IC20 - IC80) were determined to be 4.7 ng/mL and to range from 1.3 to 25.4 ng/mL, respectively, with a standard deviation (n=8) for each calibrant of between 2.3 – 26.9 AU.

142

Chapter 3

240

200

160

120

Fluorescence intensity 40

0 0.1 1 10 100 1000 10000

Chlorpyrifos concentration (ng/mL)

Figure 3-18: Chlorpyrifos calibration curve in MBio assay buffer showing fluorescence intensity versus CPF concentration using CPF-pAb (1/100) and CPF-OVA (100 µg/mL).

250

200

150

100

Fluorescence intensity

0 0.1 1 10 100 1000

Parathion concentration (ng/mL)

Figure 3-19: Parathion calibration curve in MBio assay buffer showing fluorescence intensity versus PT concentration using PT m-Ab IE (1/250) and HII-OVA (100 µg/mL).

143

Chapter 3

3.4.5 Specificity of antibodies in the MBio assay buffer

Evaluation of specificity is essential for diagnostic methods, as it can help determine if an antibody is able to identify a specific compound that exhibit high affinity and recognise other compounds from a related family that can avert false/positive results. Figure 3-20 presents the cross-reactivity studies conducted for the two antibodies—CPF.pAb and PT- mAb 1E2—toward other OPPs and their chemical structures. These compounds may be found in fruit and vegetable samples, and as a result could be co-extracted from a real sample matrix. Therefore, it was important to determine the potential cross-reactivity of these compounds as sources of interference. In terms of CPF, based on the measurable responses from individual pesticides, the analysis showed that CPF-pAb exhibited either a small percentage (<3%) or no cross-reactivity for these tested structurally related analytes. Figure 3-21 presents the curves generated for the different assays developed for each OPP during this study.

144

Chapter 3

CL CL

OCH3 OC2H5 O2N O CL O OCH3 OC2H5

Chlorpyrifos CH3 Parathion-methyl CH3

CH3

CH3 Parathion Coumaphos CL

C H O CL O 2 5 OC2H5

OC2H5 C2H5

Dichlofthion Trizophos O

OCH3 S H5C2 O C2H4 S C2H5 N s

N OC H OCH 2 5 N 3 Disulfoton Azinphos-methyl

N NC OC2H5 O O OC2H5 OCH3

OCH3 NC

Phoxim Cyanophos

CH S 2 OC2H5

OC2H5

Phosalone Figure 3-20: Chemical structures of organophosphate pesticides evaluated in a cross- reactivity study.

145

Chapter 3

225 CPF 200 Parathion-methyl

175 Disulfoton 150 Triazophos

125 Cyanophos 100 Parathion 75

Fluorescence intensity Dichlofenthion 50 Phoxim 25 0 Coumaphos 0.1 1 10 100 1000 10000

Pesticide Concentrations (ng/mL)

150 125

100 CPF Azinphous 75 Phosalone 50 Fluorescence intensity 25

0 0.1 1 10 100 1000 10000

Pesticide concentrations (ng/mL)

Figure 3-21: Cross-reactivity of CPF-pAb with structurally related OP insecticides. Figure A shows analytes with no cross reactivity at 10000 ng/mL and figure B shows those with some degree of cross-reactivity. CPF is presented in both figures.

146

Chapter 3

The CPF-pAb results showed either weak or no cross-reactivity (specificity) for all of the studied OPPs. CR percentages were also calculated as 100%, 2.3% and 0.04 % for CPF, phosalone and azinphos, respectively, showing very low levels (< 3%) of cross-reactivity. However, no cross-reactivity observed with the other analysed OPPs for concentrations up to 10000 ng/mL, the highest standard used in the study. Table 3-6 represents all OPP profiles, inhibition levels (IC80, IC50 and IC20) and cross-reactivity percentages.

Table 3-6: Cross reactivity profile of CPF-related compounds for CPF-pAb in a MBio assay buffer using a MBio reader. OP Pesticides IC50 (ng/mL) CR (%) IC20 (ng/mL) IC80 (ng/mL) Chlorpyrifos 5.4 100 0.7008 42.3 Parathion > 10000 None > 10000 > 10000 Parathion-methyl > 10000 None > 10000 > 10000 Azinphos 17192 0.04 1004 294485 Dichlofenthion > 10000 None > 10000 > 10000 Disulfoton > 10000 None > 10000 > 10000 Triazophos > 10000 None > 10000 > 10000 Phoxim > 10000 None > 10000 > 10000 Phosalone 249 2.3 70.9 874 Cyanophos > 10000 None > 10000 > 10000 Coumaphos > 10000 None > 10000 > 10000

With respect to PT, the analysis showed that PT-mAb 1E2 exhibited different cross- reactivities ranging from 0.47 to 315.14% for some evaluated structurally-related compounds. However, PT-mAb 1E2 showed no cross reactivity with phosalone, parathion-methyl and cyanophos at 10000 ng/mL. Florescent responses of each compound were plotted against corresponding pesticide standards and presented in Figure 3-22.

147

Chapter 3

200

Parathion 150 Chlorpyrifos Phoxim

100 Trizophos Azinphos-methyl Disulfoton

Fluorescence intensity 50 Coumaphos Dichlofenthion Parathion-methyl 0 0.1 1 10 100 1000 10000 Pesticide concentrations (ng/mL)

200

Parathion 150 Phosalone

Cyanophos 100

Fluorescence intensity

0 0.1 1 10 100 1000 10000

Pesticide concentrations (ng/mL)

Figure 3-22: Cross-reactivity of PT-mAb 1E2 with other related OPPs shown as standard curves using MBio assay buffer. (A) shows different cross-reactivity and (B) shows no cross- reactivity at 10000 ng/mL, the highest concentration used in the experiment.

148

Chapter 3

Most of the studied compounds cross-reacted differently, and the cross-reactivities were calculated as 100%, 315.15%, 21.89%, 12.95%, 2.37%, 0.95%, 0.89%, 0.49 and 0.47 for parathion, coumaphos, triazophos, dichlofenthion, phoxim, parathion-methyl, disulfoton, CPF and azinphos, respectively. The results showed high-affinity binding to coumaphos, even higher than parathion, while the remaining compounds exhibited cross-reactivity results of less than 22%. Further, cyanophos and phosalone showed no inhibition at the highest concentration used (10000 ng/mL) indicating there was no cross-reactivity. In addition, when possible, for each OPP the midpoint concentrations (IC50), dynamic range

(IC20-IC80) and theoretical lower and upper detection limits were calculated for each assay curve. Also, the cross-reactivity results were calculated as a percentage relative to a PT standard. Table 3-7 lists the different inhibition levels (IC80, IC50 and IC20) and percentages of cross-reactivity for all compounds used in this study, as determined by the MBio reader.

Table 3-7: Cross-reactivity profiles of PT-related compounds with PT-mAb 1E2 in MBio assay buffer using an MBio reader. OP Pesticides IC50 (ng/mL) CR (%) IC20 (ng/mL) IC80 (ng/mL) Parathion 9.26 100 2.88 29.74 Chlorpyrifos 1888 0.49 399 8933 Parathion-methyl 974 0.95 773 1247 Azinphos 1960 0.47 525.11 7375 Dichlofenthion 71.51 12.95 96.71 52.88 Disulfoton 1041 0.89 214.6 5051 Triazophos 42.35 21.87 5.157 347.80 Phoxim 391.60 2.37 32.63 4699 Phosalone > 10000 None > 10000 > 10000 Cyanophos > 10000 None > 10000 > 10000 Coumaphos 2.94 315.14 1.34 6.45

3.4.6 Matrix evaluation

It is important to highlight the goal of this work, which was the development of a desirable procedure with as few steps as possible to serve as a rapid diagnostic method. Further, food samples require preparation, such as homogenisation and blending, to ensure uniform samples before extraction. As a result, traditional extraction procedures are performed with suitable solvents to separate targets from the other matrix components. Some common techniques used to reduce matrix effects include either a clean-up step, extract dilutions and concentrating extracts to dryness, which can reduce matrix effects

149

Chapter 3

and decrease the concentration of any potential interfering materials before assay development (Melo et al., 2012; Hennion and Barcelo, 1998). Further, the merit of any analytical procedure concerns the absence of any interferences that are normally present in real solid samples or that arise from reagents, as interferences can lead to inaccurate measurements by increasing or decreasing a response (false positive or false negatives).

Strawberries are globally one of the most popular fruits, and are typically consumed around the world either whole or juiced, or in processed products such as jam or incorporated in baby food, which could be a significant source of contamination. Therefore, this assay was developed to determine CPF and PT levels and to evaluate its performance with strawberry samples as a model for analysis.

3.4.6.1 Development of a sample preparation procedure for chlorpyrifos extraction

In developing an immunoassay method for CPF detection in strawberry samples, the analysis was kept as simple as possible to ensure the suitability of the extraction procedure and an appropriate extract dilution and to achieve a successful measurement with the MBio platform. For pesticide extraction, different organic solvents are commonly used. such as methanol, acetonitrile, acetone and others, particularly for pesticides with lower solubility rates in water and solid matrix samples. Nevertheless, such organic solvents can cause strong matrix effects in an immunoassay performance, and different sample dilutions are normally required to be used 20-100 times to reduce these effects (Xu et al., 2012a). For example, acetone has been used as an extraction solvent for CPF from different vegetable matrices (Gabaldon et al., 2007). By contrast, Garcés-García et al. examined different organic solvents—acetonitrile, dimethylsulfoxide ethyl-lactate, methanol and 2-propanol and mixtures of acetone/acetonitrile and hexane/acetonitrile— to extract the same compound, and found that methanol achieved the best extraction results for CPF (Garcés-García et al., 2006). As CPF is a hydrophobic compound and thus has limited solubility in water, methanol was considered for the extraction procedure for which two dilutions of strawberry extract during sample preparation were evaluated.

In general, after evaluating the two extract dilutions at 1:8 and 1:4 (as described in Section 3.3.10.1), the results showed no significant difference between the two extraction methods (100% and 20% methanol in buffer) at the selected CPF concentration (5

150

Chapter 3

ng/mL). When the samples at 1:8 dilution were analysed, the results showed an extremely small difference in the responses of both spiked samples prior to the extraction from both extraction methods. Similar results were observed for the spiked sample after extraction by both methods (Table 3-8). However, the response of the blank sample derived from the extraction solution of 100% methanol was reduced by about 60% compared to the corresponding sample of 20% methanol, which was most likely an experimental error.

Table 3-8: Evaluation of two extraction procedures of CPF from strawberry sample spiked before and after extraction for 1:8 extract dilution at 5 ng/mL. Average response (n=6). 100% methanol 20% methanol Sample Response MBio assay buffer 140 140 Sample blank extract 107 182.5 Spike before extraction 124.8 128 Spike after extraction 189.5 178

With respect to the 1:4 dilution, the results of the blank (negative samples) derived from both extraction methods showed similarities. However, the responses derived from the spiked samples (tissue and extract) with 100% methanol extraction were suppressed compared to 20% methanol in buffer (Table 3-9). As a result, there was no marked difference between both extraction procedures, and the results of these experiments was not conclusive, as both extraction methods did not allow the researcher to determine the appropriate dilution.

Table 3-9: Evaluation of two extraction procedures of CPF from spiked strawberry samples before and after extraction for 1:4 extract dilution at 10ng/mL. Average response (n=6). 100% methanol 20% methanol Sample Response MBio assay buffer 190 190 Sample blank extracted 210.6 211 Spike before extraction 117.8 216.3 Spike after extraction 153.7 199

However, the results of the second experiment with concentrations at 5 ng/mL and 50 ng/mL and using only methanol as the extraction solvent showed improvement (Table 3-10). The responses of the blank samples diluted at 1:8 and 1:4 were 211 and 237.7 AU respectively, and both were amplified relative to a buffer sample due to matrix effects. The comparison between the two extract dilutions showed that the response at 1:8 dilution was lower than that of the 1:4 dilution by only 11.2%. In relation to the spiked

151

Chapter 3

samples (tissue and extract) at the two concentrations, the results showed the same behaviour. The spikes before and after extraction were in close proximity at each dilution, and as expected, the matrix effects for the 1:8 dilution were less influenced by the matrix compared to those for the 1:4 dilution. This result was clearly due to the fact that a lower volume of the sample was used for the dilution. Therefore, the 1:8 dilution was selected to develop the immunoassays in the following experiments.

Table 3-10: Evaluation of two extract dilutions at 1:8 and 1:4 of spiked strawberry samples before and after extraction; 5 ng/mL and 50 ng/mL or equivalent were used for fortification. Average response (n=6). Response 1:8 dilution 1:4 dilution Sample CPF 5 ng/mL CPF 50 ng/mL CPF 5 ng/mL CPF 50 ng/mL MBio assay buffer 160.7 160.7 160.7 160.7 Blank sample extracted 211 211 237.7 237.7 Spike before extraction 144.1 64.7 164.2 76.9 Spike after extraction 141.7 64.7 148.2 65.2

The response amplifications of the diluted extract relative to the MBio assay buffer could be attributed to a number of factors. For example, interferences such as organic solvents, colour or polyphenols compounds and co-extracted particulates are absent in buffer (Dankwardt, 2001). Moreover, the technique of evanescent wave fluorescence is claimed to be suitable for samples with high matrix effects, because it allows for a real-time analyses with relative resistance to matrix effects, as measurements are made at the waveguide surface where no excessive fluorophores, particulates, and interfering compounds exist, unlike bulk fluid analysis. In addition, with this technique AlexaFluor- 647 dyes are excited at longer wavelengths than the endogenous fluorophores that naturally occur in samples (Ligler et al., 2003).

Development of a calibration curve in strawberry matrix

After the extraction method was optimised and the 1:8 extract dilution was selected, these conditions were applied to develop the calibration curve in a real sample (Figure 3-23). Although the extracted curves (tissue and extract) were diluted before developing the assay, both curves were amplified and shifted to the left by almost 25% compared with the blank MBio assay buffer for blank samples, as a result of the matrix effects. However, the matrix effects were decreased at positive points and the difference became smaller at

152

Chapter 3

higher CPF concentrations, which indicated that the immunoassay was minimally influenced at these concentrations.

250

200 Spiking in MBio assay buffer 150 Spiking before extraction Spiking after extraction

100

Fluorescence intensity 50

0 1 10 100 1000 10000

Chlorpyrifos concentrations (ng/g)

Figure 3-23: Evaluation of strawberry matrix effects with CPF calibration curves in buffer, before and after extraction at 1:8 (dilution factor of 40), with CPF.pAb at 1/100.

In addition, the inhibition curve generated from the spiked extracts showed slightly suppressed responses of 5% and 7% at 25 ng/mL and 50 ng/mL, respectively. However, these differences were negligible and could be attributed to the variation in the responses generated by the MBio reader. Table 3-11 lists the percentages of different responses generated at all points for both extracted curves relative to the curve produced for the MBio assay buffer. Generally, the results showed some matrix effects were observed at both extracted curves compared with the curve in buffer. The inhibition response in buffer differed from that of a real sample, and antibodies can behave differently at each concentration. As a solid matrix was used, an assay can also be affected by the pH, the ionic strength and interferences from other constituents present in a sample (Hennion and Barcelo 1998).

153

Chapter 3

Table 3-11: The percentage of the florescence responses generated from both extracted curves relative to the MBio assay buffer. CPF concentration (ng/mL) Spike before (%) Spike after (%) 0 131.9 131.9 1.25 119.3 103.5 2.5 109.0 114.8 5 125.2 111.7 10 109.4 111.9 25 132.8 95.1 50 95.1 92.7

The averages (n=8) at each point were calculated, and the midpoints (IC50) were 5.1, 4.2 and 3.2 ng/ml for buffer, spike before and spike after extraction, respectively. This equated to 163.4 and 128.3 ng/g for spiked tissue (before extraction) and spiked extract (after extraction), respectively, which suggests a satisfactory level of sensitivity of under the legal threshold for CPF presence in most fruit (200 ng/g). The average standard deviation of 8 spots for the three calibration curves was found to range between 2 and 21

AU. Unfortunately, it was noted that after several experiments, the IC50 values at 163.38 ng/g shifted and became less sensitive over time, resulting in responses above the legal limits, which affected the sensitivity of the method and the validation study. This phenomenon could be attributed to the deterioration of CPF-pAb. Therefore, the dilution step was adjusted and re-evaluated using a lower dilution ratio of 1:4 instead of 1:8, as this a higher dilution was seen to reduce assay sensitivity (Wang et al., 2005). Three curves were developed according to the procedure stated in the methods section (Section

3.3.10.1 and Figure 3-24). The results of the calculated IC50 showed marked improvement, showing assay sensitivities below the legal limit for CPF of 6.9, 7.3 and 7.1 ng/mL for buffer, spiked tissues and spiked extracts, respectively, with equivalent concentrations of 146.9 and 142.4 ng/g for spiked tissues and spiked extracts, respectively. While the inhibition of the extracted curves shifted to the left by 36.5% due to the matrix effects relative to the calibration curve in buffer, the remaining points were less affected. Subsequently, these conditions were used for the validation study.

154

Chapter 3

150

100 Spiking MBio assay buffer

Spiking before extraction Spiking after extraction 50

Fluorescence intensity

0 1 10 100 1000 10000

Chlorpyrifos concentrations (ng/g)

Figure 3-24: Evaluation of strawberry matrix effects by developing CPF calibration curves in buffer before and after extraction at a 1:4 dilution (dilution factor of 20), with CPF.pAb at 1/100.

3.4.6.2 Development of a sample preparation procedure for parathion extraction

To develop the immunoassay method for PT analysis, several attempts were made to simplify and speed up this method and to ensure its suitability for measurement of PT using the MBio platform.

For initial matrix evaluation, methanol was used as the extraction solvent, as it showed success as solvent for CPF using the same matrix, as discussed in the previous section. In addition, Zeng et al. (2007) used methanol in a previous study to extract PT from different plant samples. The adapted extraction procedure was found to be simpler and shorter than other extraction methods for PT extraction from a solid matrix reported in the literature (e.g. Deep et al., 2015; Funari et al., 2015; Yang et al., 2009; Zhang et al., 2009; Li et al., 2006) which required hours to days for extraction and involved hazardous solvents and a clean-up step.

During the extraction procedure, steps were taken to improve the immunoassay and its required sensitivity. Two different dilution extract ratios (1:4 and 1:8) with an overall dilution factor of 1:20 and 1:40, respectively, were initially applied in order to allow for

155

Chapter 3

direct comparison with the CPF method described in Section 3.3.10.2. However, sensitivity of the broad specificity PT.mAb 1E2 relative to the MRL in strawberries was not adequate at these dilution ratios. Therefore, an extra drying step and reconstitution in a buffer was required to enhance the assay sensitivity. Efforts were taken to avoid extra steps, such as using SPE for clean-up, in order to keep the procedure feasible for use as a rapid diagnostic method.

The results of the initial extract dilution (1:4 ratio) for the sample spiked before the extraction (tissues) and after the extraction (extract) are presented in Figure 3-25. It was observed that the strawberry matrix caused strong effects in both extracted curves as compared to the buffer. The responses of the blank samples of the extracted curves were amplified and shifted by about 60 % from the corresponding sample in buffer. Similar effects were observed for most of the remaining calibrants.

400

300 MBio assay buffer 200 Spiking before extraction

Spiking after extraction

100

Fluorescence intensity

0 0.1 1 10 100 Parathion concentrations (ng/mL)

Figure 3-25: Evaluation of strawberry matrix effects by comparing three calibration curves (buffer, extracted curves, spiked before and after) at a 1:4 extract dilution.

At 1:4 extract dilution ratio, the matrix effects were substantive. Therefore, a wider range of working concentrations with a higher extract dilution at a 1:8 ratio were considered to achieve a sigmodal curve and reduce matrix effects. To this end, the procedure was adjusted. Figure 3-26 shows the resulting extracted curves and the curve in buffer in ng/g

156

Chapter 3

(A) and ng/mL (B). However, the immunoassay method at this dilution ratio compromised its sensitivity.

350

300

250 MBio assy buffer Spike before extraction 200 Spiking after extraction 150

100

Fluorescence intensity

0 0.1 1 10 100 1000 10000

Parathion concentrations (ng/g)

350

300

250 MBio assay buffer 200 Spiking before extraction Spiking after extraction 150

100

Fluorescence intensity 50

0 0.1 1 10 100 1000

Parathion concentration (ng/mL)

Figure 3-26: Evaluation of strawberry matrix effects by comparing extracted calibration curves (spiked before and after) at the extract dilution of 1:8 with the calibration curve in a MBio assay buffer. The three curves represent the concentration ranges as ng/g (A), ng/mL (B).

157

Chapter 3

Clearly, Figure 5.26 shows an improvement and a reduction in the matrix effects. At a 1:8 extract dilution ratio, the results of extracted curves produced (to a large degree) sigmoidal curves for both curves (spiked tissues and extracts). Further, the responses for each blank sample of the extracted curves were amplified by about 50-60 % relative to the curve in buffer. However, while the matrix effects were reduced, the sensitivity (IC50) was calculated at 371.7 ng/g and 678.9 ng/g for spiked tissues and spiked extracts, respectively, which suggests that the assay conditions used for CPF detection were not sensitive enough to detect PT and meet the 50 ng/g limit. As a result, this method was not fit for this purpose, and thus it was necessary to alter the sample preparation procedure by adding an extra step to dry the extracts and reconstitute the dry residue in buffer. In others studies, similar attempts to improve sensitivity and reduce matrix effects were conducted for extracting PT from solid samples (Zhang et al., 2009; Li et al., 2006).

3.4.6.3 Drying and reconstitution procedure As both the 1:4 and 1:8 extract dilutions were insufficient in reducing matrix effects and improving the sensitivity of the method, a different approach was taken by incorporating a drying step into the sample preparation procedure. This additional step was undertaken using fenitrothion and triazophos analysis in order to improve the sensitivity of reported immunoassay methods (Du et al., 2015; Cho et al., 2004). Initially, it was necessary to conduct a rapid test to evaluate the appropriateness of two volumes of MBio assay buffer that were used to reconstitute the dried strawberry residues. However, PT solubility varies at 200 and 500 µL, and the presence of other constituents can affect PT-mAb 1E2 binding behaviour. Therefore, 50 ng/g, which represents the MRL of PT in many fruits, was selected to spike the strawberry homogenate. After following this revised extraction procedure, an average of 8 responses was calculated for each sample (Figure 3-27). The results showed that testing of three blank samples to include buffer and dried residues in 200 µL and 500 µL volumes produced responses of 185.8, 79.0 and 173.7 AU, respectively. Figure 3-27 shows that the responses generated from the samples that were reconstituted in 500 µL were comparable to those in buffer, which were negligibly lower by only 6.5%.

158

Chapter 3

200

150 Control (blank) Spike before 100 spike after

Fluorescence intensity

MBio assay buffer MBio assay buffer MBio assay buffer ReconstitutedReconstituted in 200µL in 500µLReconstitutedReconstituted in 200µL in 500µLReconstitutedReconstituted in 200µL in 500µL

Figure 3-27: Evaluation of two volumes of a mBio assay buffer (200 µL and 500 µL) used to reconstitute dried strawberry extract and their comparison with buffer. 50 ng/g and an equivalent concentration were used to spike the strawberry homogenate before and after extraction.

However, with respect to reconstituting the extract in 200 µL, responses were significantly suppressed and dropped by about 58% relative to the blank buffer sample from just above 180 to around 80 AU. Similarly, the same pattern for the spiked sample (tissues and extracts) was observed. However, the suppression response for the 200 µL volume could be attributed to a slow flow rate, as the sample was more concentrated (i.e. had a higher viscosity); the antibody may also have been inhibited. Based on the sample preparation protocol discussed above, the 500 µL volume was selected to reconstitute the dried residues in order to again develop a full calibration curve for a real sample.

3.4.6.4 Development of a calibration curve for strawberry samples

The optimised conditions and the developed procedure for sample preparation were used to develop three full calibration curves. The range of PT concentrations was modified slightly to become 0-400 ng/g, equivalent to 0-160 ng/mL to suit the new conditions as described in Section 3.3.10.2. When the responses were generated for the extracted curves (before and after extraction), they were compared as a % relative to the calibration curve

159

Chapter 3

in MBio assay buffer. The results are summarised in Table 3-12 shows the inhibition percentage varied across the calibrants in both curves. In addition, the % responses indicated that only the first and last two calibrants were suppressed for the spike before whereas for the spike after all calibrants were suppressed except at the highest concentration 160 ng/mL as a result of the matrix effects. Generally, the result indicated that the spike after samples was mostly suppressed by about 25% at the point of 40 ng/mL whereas the spike before sample was mostly amplified by 48% at the point of 20 ng/mL. This 25% would be indicative of the matrix effects whereas the differential between before and after would be indicative of the recovery.

Table 3-12: Percentage of florescent responses generated from both curves in relation to a MBio assay buffer. PT concentrations (ng/mL) Spike before (%) Spike after (%) 0.01 84.50 84.50 2 91.14 77.76 4 107.02 77.78 10 126.00 88.33 20 148.32 78.83 40 136.29 74.06 80 98.93 83.56 160 75.39 111.56

The curves are presented in Figure 3-28 and based on their dynamic range, the calculated sensitivity for all curves showed a minor difference in IC50. The spiked dried residue (spike after) exhibited a higher sensitivity than the buffer (5.5 ng/mL and 7.5 ng/mL, respectively). However, the IC50 values calculated from the spiked tissues exhibited lower sensitivity at 21.8 ng/mL, which could be a result of matrix effects. The IC50 values were also calculated as ng/g, and results showed 54.6 and 13.8 ng/g for spiked tissues and extracts (before and after, respectively). These results suggest that assay sensitivity was just above the action limit (50 ng/g). However, this experiment could not be optimised further at this stage due to time constraints and a limited quantity of the reagents.

160

Chapter 3

175

150 Mbio assay buffer

125 Spiking before extraction

100 Spiking after extraction

Fluorescence intensity 25

0 0.1 1 10 100 1000

Parathion concentration (ng/g)

175 150 125 Mbio assay buffer Spiking before extraction 100 Spiking after extraction 75

Fluorescence intensity 25

0 0.1 1 10 100 1000

Parathion concentration (ng/mL)

Figure 3-28: Three curves indicating the results of spiked (tissue and extract) and MBio assay buffers in a 10-minute assay. The curves represent the concentration ranges as ng/g (A) and ng/mL (B) after evaporation of 1 mL and reconstitution in 500 µL of a MBio assay buffer.

161

Chapter 3

3.4.7 Validation study

3.4.7.1 Chlorpyrifos

The average IC50 calculated over two days was 185.3 ng/g with a dynamic assay range of 11.4-1629.6 ng/g. However, the assay sensitivity was reduced by about 52.3 % in day two, which could be explained by the deterioration of CPF.pAb. Furthermore, across all 20 samples, both the blank samples and the lowest fortified concentration used in the experiment (100 ng/g) showed no overlap. Figure 3-29 presents the results of the blank and positive strawberry samples at three different concentrations.

400

300 0 ng/g 100 ng/g 200 ng/g 200 300 ng/g

100

CPF concentrationCPF (ng/g)

0 0 5 10 15 20 Sample population

Figure 3-29: Four populations of strawberry samples (n=20), fortified at different CPF concentrations.

Limits of detection and quantification The LOD and LOQ values for CPF were 28.2 ng/g and 79.1 ng/g, respectively. However, some of the spots in the same planar waveguide of the negative samples showed a minor reduction in response, which could be attributed to low level of interference from other sample components that gave some positive concentrations.

162

Chapter 3

Intra-and-inter batch reproducibility The intra- and inter-day precision study was performed using three different concentrations. The analysis was carried out using 10 replicates for intra-day precision; the same experiment was conducted on a different day to determine inter-day precision. Moreover, CPF showed good precision for both intra-day and inter-day analyses with acceptable CV%. Table 3-13 shows the results of these experiments.

Table 3-13: Measured concertation, standard deviation, coefficients of variation and recovery values for samples spiked with CPF at three levels. Values were averaged (N=20). Spiking level Measured concentration CV (%) Mean (ng/mL) (ng/g) ± S.D. recovery (%) 100 114.99 ± 15.29 13.4 114.4 Day 1 200 220.75 ± 23.3 10.6 110.4 300 295.7 ± 22.3 7.5 98.6 100 113.63 ± 19.24 16.9 113.6 Day 2 200 225.94 ± 21.95 9.7 112.97 300 318.80 ± 27.08 8.5 106.3 Overall 100 113.9 ± 16.9 14.9 113.9 (2 days) 200 223.0 ± 22.2 9.9 111.7 300 307.00 ± 26.8 8.7 102.4

Recovery The recovery results of three concentrations that were spiked in strawberry sample tissues and analysed by the MBio reader showed that the assay method recovered a high percentage of CPF. Table 3-13 shows the recovery results. The average recovery rates of CPF was 109.5% ± 5.6, ranging from 98.6% to 114.4 %; and the average CV% was 11.2%, ranging from 7.5% to 16.9%, indicating acceptable overall average recovery values. Furthermore, the results confirmed that the assay was sufficiently sensitive to meet the international thresholds of the Codex Alimentarius and the EU MRL (200 ng/g and 300 ng/g, respectively) for CPF detection. The method’s precision was based on repeatability results that produced CV% of less than 17%, and the mean percentage recoveries also suggested satisfactory accuracy of the method.

Comparison of the developed immunoassay method

The results of this study were compared with those of previous studies. Many analytical methods have been reported in the literature that evaluated the detection of CPF in food and environmental samples. Generally speaking, competitive immunoassays for semi-

163

Chapter 3

quantitative identification of CPF in strawberry samples compared well with biosensor and conventional analytical techniques. The method developed in this study demonstrated excellent sensitivity and comparable or even superior results when compared to similar immunoassay methods developed for rapid and on-site analysis. A direct comparison can be made to the work of Kim et al., (2011) who developed an immunochromatographic assay test strip to detect CPF in vegetable samples and found comparable or lower limits of detection. While the assay detection duration for test strip’s study was fewer than 10 minutes, sample preparation required 24 hours, whereas our method required relatively minimal work and less than 1 hour for sample preparation (Kim et al., 2011).

In a different immunoassay approach, Liu et al. used ELISA to detect CPF in environmental samples and reported a detection limit of 30 ng/g (Liu et al., 2011). Similarly, Soler et al. showed comparable sensitivity results with a theoretical limit of detection of 1-1.75 ng/mL that was calculated in a buffer (Soler et al., 2008). In addition, Chen et al. employed a fluorescence-linked immunosorbent assay (FLISA) for CPF detection and reported detection limits of 8.4 ng/mL (Chen et al., 2010a) and 3.8 ng/mL (Chen et al., 2010b). While FLISA exhibited a shorter assay time than ELISA, both methods showed comparable assay sensitivities. However, the main disadvantage to these studies is that they are not rapid and are thus inappropriate for field analysis.

In addition, several biosensors have been used to detect CPF in laboratory settings and various limits of detection have been reported. For example, electrochemical biosensor- based enzyme studies reported values of 0.05 ng/mL (Chen et al., 2015) and 0.986 ng/mL (Hunde et al, 2017). However, despite the fact that these methods are sensitive, they lack selectivity, and also CPF was identified in liquid samples as a general inhibitor. Further comparison to other methods included electro- and optical-immunosensors that produced several detection limits at the scale of pg/mL. For example, electro-immunosensor studies reported limits of 0.01 ng/mL (Wang et al., 2017c); 0.0063 ng/mL (Sun et al., 2015); 0.014 ng/mL (Cao et al., 2015); and 0.046 ng/mL (Sun, et al., 2012). While these methods were more sensitive than our method, they were also lab-based and thus could not be performed in the field. In addition, the very low detection limits reported in these studies can be explained by their use of highly advanced nanomaterials as a part of their recognition element.

164

Chapter 3

With respect to the optical and portable SPR, results showed detection limits ranging from 0.045 to 0.064 ng/mL (Mauriz et al., 2006). In comparison to our method, while SPR is a portable tool with superior sensitivity for CPF detection, the authors used an anti- chlorpyrifos monoclonal antibody that showed higher specificity compared to the polyclonal antibody used in our study. In addition, they detected CPF in a water sample using a complicated extraction procedure requiring a relatively high volume (500 mLs) of the sample, and the sample had to be passed through SPE columns before drying for 20 minutes. The signal also had to be amplified for 20 minutes to complete the analysis, whereas in our study only minimal sample preparation was required and the analysis took only 10 minutes.

Finally, traditional methods such as chromatography have been used to detect CPF and have produced different detection limits. For example, Hazer et al. reported 62 ng/mL for HPLC (Hazer et al., 2017), Lazic et al. reported 4 ng/mL for GC (Lazic et al., 2011), and Tian et al. reported 1-10 ng/g (Tian et al., 2016) for HPLC-MS/MS. However, most of these methods exhibited lower detection limits and required a lengthy extraction process and some type of sample clean-up procedure. Further, while some of these methods were more sensitive, the method developed in this study is appropriate for such applications in terms of screening and follow-up detection.

3.4.7.2 Parathion

The results showed that the IC50 for PT detection was found to be 44.5 ng/g, which provided sufficent sensitivity and a dynamic range of 10.1- 147.4 ng/g. Furthermore, both the blank samples and the lowest fortified concentrations used in this experiment (25 ng/g) showed no overlap across all 10 samples. Figure 3-30 presents the results of the blank and positive strawberry samples fortified at three different concentrations.

165

Chapter 3

100

60 0 ng/g 25 ng/g 40 50 ng/g 75 ng/g 20

PT ConcenttionPT (ng/g)

0 0 2 4 6 8 10 Sample population

Figure 3-30: Four populations of strawberry samples (n=10) fortified at different PT concentrations.

Limits of detection and quantification

The LOD and LOQ values for PT were found to be 6.5 ng/g and of 15.4 ng/g for the assay, respectively. The values were lower for PT and showed greater sensitivity compared to the assay for CPF.

Intra-day study The results of the intra-day repeatability study showed satisfactory levels of precision, with acceptable CV% values ranging between 10.2 and 18.8 and (Table 3-14).

Table 3-14: Measured concertation, standard deviation, coefficients of variation and recoveries for samples spiked with PT at three levels. Values were averaged (n=10). Spiking level Measured concentration CV (%) Mean (ng/g) (ng/g) ± S.D. recovery (%) 25 29.32 ± 5.50 18.8 117.17 Day 1 50 57.03 ± 5.64 9.9 114.06 75 75.76 ± 7.47 10.2 101.01

Recovery Study results showed that the method recovered a high percentage of PT with acceptable mean recovery values, as indicated in the tables above. The results of an overall average

166

Chapter 3

recovery analysis were found to be 110.9 ± 8.3. This value was over 100%, which could be a result of matrix effects, variability in the analysis or the presence of other inhibitors. And while the validation was only performed over a single day, it showed a promising assay approach to detecting PT in strawberry samples that would meet the international standards introduced by Codex Alimentarius and the EU MRL for PT detection in most fruit and vegetable samples at (50 ng/g). In addition, the CV% values were calculated at less than 19%, which falls within reasonable limits for each concentration of fortification.

Comparison of the developed immunoassay method

In terms of PT, the overall performance of the developed assay was compared to published work, including immunoassays, bioanalytical techniques such as immunosensors and conventional methods. It worth mentioning that most recent methods for PT were conducted using buffer or water samples, and LODs were determined accordingly. Furthermore, only two methods were developed for rapid detection used in on-site analysis; the majority were developed for methods that required a laboratory setting. Comparisons were also made to other rapid-detection methods. For example, Gerent and Spinelli (2016) used an electrochemical biosensor for PT detection in skim milk samples and produced a LOD of 55.7 nmol/L. However, while this method was sensitive, the researchers used different recognition elements, such as bismuth-film electrodes, which showed a high specificity for PT (Gerent and Spinelli, 2016). In another method, a LOD at 0.081 ng/mL was reported for a rapid detection method that demonstrated high sensitivity. However, this method was based on acetylthiocholine, which is enzyme-based technique, and is not selective to PT. The platform also required cysteine capped gold nanoparticles to enhance detection (Bala et al., 2015).

With respect to available immunoassays, several ELISAs were developed for PT detection. However, most of these methods were used to detect PT in aqueous solutions, such as 25 ng /mL in buffer (Zhang et al., 2015); 2 ng/mL in buffer (Liu et al., 2010); and 0.08 ng/mL in PBS (Gui et al., 2009). For other samples (cucumber, rice and corn) and water, LODs were 10 ng/g and 5 ng/mL respectively (Liu et al., 2009). The LOD of 10 ng/g was in agreement with our method developed for PT in strawberry samples. Other immunosensors, mainly those using electrochemical platforms, have also been utilised for PT quantification. For example, detection limits that have been reported include 0.046 167

Chapter 3

pg/mL in a PBS buffer (Mehta et al., 2017); 0.025 ng/L in a PBS buffer (Mehta et al., 2016) and 0.1 ng/mL for a rice sample (Deep et al., 2015). The high sensitivity level that was produced for rice samples in the latter study could be attributed to the use of a hybrid surface of both a nanometal organic framework material and an anti-parathion antibody for sensing PT. Further, a quartz crystal microbalance was also used with a reported LOD of 60 ng/mL in water (Funari et al., 2013).

In addition, our results were compared to those using biosensors. For example, Yang et al. (2009) used an electrochemical biosensor to detect PT in vegetable samples (cucumber and cabbage) and reported a LOD of 3 ng/g. These results can be attributed to the high affinity of a MIP as a sensing surface used for detection. Nevertheless, their result was comparable to our work. Further comparisons were also made to conventional methods. For example, a LOD of 4.7 ng/mL for cow milk samples and 1.5 ng/g for leek samples were reported for analysis using a GC-MS method (Rodrigues et al., 2011) and GC- MS/MS (Qu et al., 2010), respectively. These detection limits are comparable and in agreement with the present study. And while these methods can achieve multi-reside analysis, which is excellent for pesticide monitoring, the obvious advantage of our method is its relative simplicity, lower costs, rapid analysis and promising portability. Furthermore, sensitivity levels were higher when comparisons were made with more advance techniques using LC–MS/MS, with a LOD reported in water samples of 0.3 ng/mL (Yang and Lee, 2010).

Unfortunately, the analytical detection of a target in different matrices cannot be directly compared in all cases, where some detection limits are reported in different units as ng/mL. Thus, when compared with our results, some of the LODs reported above for liquid samples may produce less accurate comparisons.

168

Chapter 3

3.5 Conclusion

This method developed is a simple, rapid, sensitive and reproducible method to detect CPF and PT in strawberry samples in a 10-minute assay based on a portable optical planar waveguide approach. After evaluating different conjugate concentrations (CPF: CPF- OVA and PT: H11-OVA) and Abs dilutions (CPF.pAb and PT-mAb 1E2) for both targets, coating antigens were immobilised on waveguides before titrating with antibody dilutions to determine appropriate concentrations and the optimal antibody dilutions.

Under optimised conditions, the developed assay showed selectivity towards CPF when ten structurally-related compounds were tested in buffer. Also, minimal sample preparation for CPF extraction from the strawberry sample was required and was achieved in less than one hour. Assay parameters were validated and they demonstrated repeatability and precision, with a LOD of 28.2 ng/g, which meets the safety limits of various regulatory authorities. The developed assay also showed satisfactory sensitivity when compared to available and published detection methods for on-site analysis. However, we were challenged by a shortage of immunoreagents, which limited our validation work. Moreover, this promising assay method could be applied as a screening tool for CPF detection in agricultural applications, including those for fruits and vegetables, and for environmental samples. In addition, this assay could be used for follow-up monitoring purposes for both quality and safety control. Future work will focus on completing the validation study, and perhaps using enhancement techniques of combining the MIP technology to further improve assay sensitivity.

With respect to PT detection, the study results describe a relatively simple assay method, and showed that under optimised conditions PT-mAb had cross-reactivity with different degrees to a number of OPPs in which coumaphos showed a very high affinity at 315.5%. The results for the remaining nine pesticides ranged from zero to 21.9% in buffer. The method also showed promising results after assay parameters were validated and

demonstrated satisfactory repeatability in a single run, with IC50 and LOD found to be 44.5 ng/g and 6.5 ng/g, which meets the MRLs set by various regulatory authorities. Most of the published methods that reported LODs in environmental samples were conducted mainly in buffer and water samples, which made it difficult to make direct comparison. Also, for this study sample preparation required an additional drying step in order to

169

Chapter 3

achieve satisfactory sensitivity that could be applied in real sample applications. Therefore, further validation would be necessary. As PT-mAb 1E2 exhibited broad specificities, it would be worthwhile to develop an extraction procedure that encompasses the additional pesticides that showed high cross-reactivity with PT, as well as evaluating them in a single assay. In addition, a sample preparation could be investigated to improve simplicity of use and portability.

170

Chapter 4

Chapter 4: Challenges in the development of a multiplexing array to detect organophosphate pesticides

171

Chapter 4

4.1 Abstract

Multiplexing microarrays to develop rapid and sensitive methods for pesticide detection are in their infancy. The concept of microarray diagnostics could be transferred and applied to a broad range of applications in environmental and food safety monitoring. Here, is discussed an attempt to develop a multiplex microarray for the detection of two harmful pesticides, CPF and PT, found in agricultural products and water. This study was a preliminary investigation and was conducted as a proof of concept. Two antibodies, CPF-pAb, which is specific for CPF, and PT-mAb 1E2, which has broad specificity for organophosphate pesticides, were used to determine the feasibility of a multiplexed assay that followed the evaluation of single assays. Both CPF-OVA and H11-OVA were printed onto planar waveguide slides, and the molecular interaction of each antibody was evaluated both individually and in combination. For the multiplexed assays, the results of calibration curves developed both in buffer and for real samples indicated a specificity of CPF-OVA to CPF, whereas H11-OVA showed affinity to both CPF and PT. Due to the broad specificity of PT-mAb 1E2 to a number of organophosphate pesticides, the applicability of this multiplex assay for specific pesticide identification was limited, and it was determined that more specific antibodies are required. Interference of reagents between different assays highlights one of the challenges of antibody microarray construction on a single channel device. Further studies would need to be performed that apply this concept but use specific antibodies to individual pesticides to develop a highly sensitive and specific multiplex assay for pesticide identification using this planar waveguide biosensor. This application could be used for future pesticide detection methods that include CPF-Ab and other specific organophosphate pesticides after optimisation in a single assay.

172

Chapter 4

4.2 Introduction

The increasing occurrence of potentially harmful pollutants and contaminants in food supplies and ecosystems is a serious issue. Many contaminant classes such as pesticides, toxins, and antibiotic residues can pose threats to human health through their uptake in food supplies and in ecosystems. Several pesticides in many formulations have been widely used to protect edible plants from insects and to improve crop yields. However, the increasing presence of a number of pesticides detected in food supplies is of extreme global concern. To conduct the required measurements to detect these harmful contaminants, significant analytical monitoring methods are used in order to ensure public safety.

To this end, using multi-residue screening methods for discrete samples is vital to safeguarding public health. These methods are advantageous, because they are rapid, sensitive and relatively simple approaches to multi-residue analysis for monitoring purposes. To date, the identification of several pesticide residues have been based mainly on traditional methods (Section 1.6.1). Further, most on-site immunoassays used to detect pesticides have usually been singleplex assays that are geared for either single compound or class-selective pesticides.

For field analysis, immunochromatographic test strips (immunostrips) have been widely used. For example, one method was developed to simultaneously identify the presence of both methyl-parathion and imidacloprid using bifunctional antibody and pesticide-labeled enzymes (Shu et al., 2017); this analysis was performed on traditional Chinese medicine samples and took only 22 minutes. In another study, Xu et al. developed a strip-based immunoassay method utilising specific nanocolloidal gold-labeled monoclonal antibodies in a semiquantitative analysis to detect imidacloprid and thiamethoxam (Xu et al., 2012a). In addition, Rubtsova et al. developed a multi‐spot membrane strip to identify three pesticides (atrazine, simazine and terbuthylazine) in water samples using an electrochemiluminescence system (Rubtsova et al., 1998). However, while strip tests for pesticides are used for on-site analysis, they remain restricted by the number of targets that can be detected and largely only provide a qualitative test (Sajid et al., 2015).

173

Chapter 4

One approach to transferring testing sites from laboratories to on-site multiplexing involves using the SnapEsi reader from MBio Diagnostics Inc. (see the discussion in Chapter 3), a tool that is capable of analysing multiple targets in individual samples. In this method, a waveguide surface is coated with different coating antigens and mixed reagents are applied for assay development (Figure 4-1). This approach is discussed in Chapter 3.

Figure 4-1: Theory and principles of the planar waveguide approach used in the MBio biosensor (Lochhead et al., 2011).

To date, a few studies have used this technique to deliver proofs of concept. In the medical diagnostic field, Lochhead et al. developed a relatively simple method to quantify Treponema bacteria and the HIV-1 and hepatitis C viruses (Lochhead et al., 2011). In addition, Devlin et al. developed a method to detect microcystins in water using a portable biosensor. This group developed two assays, one for free microcystins and another for intracellular microcystins that can be completed in 15 and less than 30 minutes, respectively (Devlin et al., 2013). Another approach was also used to detect five groups of harmful algal toxins in seawater samples in approximately 15 minutes (McNamee et al., 2014). Further, in a study targeting three antibiotics in milk samples, McGrath et al.

174

Chapter 4

developed a rapid multiplexed assay for field detection to detect chloramphenicol in raw dairy milk, which required only a five-minute assay duration (McGrath et al., 2015).

4.2.1 The chapter aim

The aim of this chapter was to construct a multiplex assay using the optical waveguide array biosensor for the simultaneous detection of CPF and PT to illustrate a model system applicable to multiple pesticide analysis. The first objective was to combine the two individual assays for CPF and PT into one assay utilising the same reagents and show as a proof of principle in a buffer system. The second objective was to combine the sample preparation to develop a single sample preparation protocol for the analysis of real samples using strawberries as the model food.

175

Chapter 4

4.3 Materials and Methods

4.3.1 Reagents and chemicals

All materials are the same as those described in Section 3.3.1.

4.3.2 Microarray construction

For each assay, optimised concentrations of CPF-OVA and H11-OVA conjugates (100 μg/mL) were prepared, and were used to detect the corresponding organophosphate pesticides in a single assay. First, 200 µg/mL of each conjugate was prepared in phosphate buffer saline (PBS) (0.1M; 7.2 pH), and 100 µg/mL of each conjugate was then prepared in filtered concentrated printing buffer (100 mM sodium phosphate, 50 mM sodium chloride, 100 μg/ml BSA, 0.005% Tween-20, pH 8.0) (1:1 ratio; initial dilutions: printing buffer). BSA-647 (647- dye-labelled; 2 μg/mL) was prepared using the same procedure at 4 µg/mL and 2 µg/mL for initial and final dilutions, respectively. A lower concentration of printing buffer was prepared by mixing 50% of filtered concentrated buffer with 50% degassed water. Section 3.3.6 describes all the steps, including microarray printing, waveguide blocking, cartridge building and assay protocols.

4.3.3 Immunoassay procedure

First, eighteen waveguides were placed in a spotter. Then, BSA-647 at 2 μg/mL, CPF- OVA and H11-OVA conjugates at 100 µg/mL were immobilised on each waveguide in replicates of 8, and printing buffer was used to print spacing locations, as shown in Figure 4-2.

Figure 4-2: Representative array layout showing the arrangement of different spots (volume;25 nL), controls (fluorescently-labeled BSA-647), and H11-OVA and CPF-OVA (coating antigens).

176

Chapter 4

4.3.4 Evaluation of a multiplexed assay for the simultaneous detection of CPF and PT in buffer

To evaluate the reagents, eighteen waveguides were divided into three groups. The evaluation procedure included: 1) CPF-pAb specificity to both conjugates (CPF- OVA/H11-OVA); 2) PT-mAb 1E2 specificity to both coatings (CPF/H11-OVA); and 3) development of multiplex calibration curves in buffer and for a real sample.

4.3.4.1 CPF-pAb specificity to both coating antigens (CPF-OVA /H11-OVA)

A developed assay for a singleplex for CPF-pAb was utilised to evaluate the first six waveguides in buffer using six working standard concentrations of CPF (0, 1, 5, 10, 25, and 100 ng /mL) prepared in MBio assay buffer. First, to investigate the binding specificity to both coatings, 100µL of each standard was mixed with CPF-pAb at 1/100 dilution (1:1 ratio). Then, 150 µL of the mixture was used to load each cartridge and was then incubated for 5 minutes. Thereafter, a secondary antibody (goat-anti-rabbit-alexa 647) at a dilution of 1/200 was added to each cartridge and incubated for an additional 5 minutes.

4.3.4.2 PT-mAb 1E2 specificity to both coating antigens (CPF-OVA/H11-OVA)

The singleplex assay developed for PT-mAb 1E2 was again used to evaluate the second six waveguides in buffer. Six working standard concentrations of PT at 0, 1, 2.5, 5, 10 and 100 ng/mL were first prepared in MBio assay buffer. Then, to investigate the binding specificity to both conjugates, 100µL of each standard was mixed with PT-mAb 1E2 (1/250 dilution; 1:1 ratio). Finally, 150 µL of the mixture was pipetted in each cartridge and was incubated for 5 minutes, and the secondary antibody (goat-anti-mouse-alexa- 647) at a dilution of 1/200 was added through the cartridge inlet and was incubated for an additional 5 minutes.

4.3.5 Development of multiplexed calibration curves in buffer

For the remaining six waveguides, a mixture of both standards was prepared as follows: 0, 2, 10, 20, 50 and 200 ng /mL for CPF and 0, 2, 5, 10, 20 and 200 ng/mL for PT. Higher concentrations were prepared (concentrations × 2) before mixing both standards together (1:1 ratio) to achieve the required concentrations at each calibrant. Similarly, higher antibody dilutions of CPF-pAb and PT-mAb 1E2 were prepared at 1/50 and 1/125,

177

Chapter 4

respectively, before preparing a mixture of both antibodies at a 1:1 ratio to achieve the required dilutions (1/100 and 1/250) respectively. Then, each mixture of both standards and antibody dilutions were mixed again at a 1:1 ratio. To prepare a single solution, both secondary antibodies (goat anti-rabbit alexa-647 and goat anti-mouse alexa-647) were prepared directly at a dilution 1/200 in 1 mL by adding 5µL of each secondary antibody into 990 µL (MBio assay buffer). Then, 150 µL of each mixture (standards/antibodies) was pipetted in each cartridge inlet and was incubated for 5 minutes. Finally, the mixed secondary antibody (goat anti-mouse alexa-647) was added and incubated for an additional 5 minutes.

4.3.6 Real sample analysis

Sample preparation

The final optimised extraction procedure for PT was applied in order to extract both analytes from the spiked strawberry samples. First, aliquots (1 g) of the homogenised strawberry sample (n=12) were transferred to plastic tubes (15 mL). Each sample was then spiked with 25µL of CPF concentrations (0, 1, 12.5, 25, 50, 100, 200, 500, 1000, 2000, 4000 and 8000 ng/g) that were prepared in methanol and 25μL of PT concentrations (0, 0.1, 1.25, 2.5, 5, 10, 20, 50, 100, 200, 400 and 800 ng/g) that were prepared as mixed standards in a MBio assay buffer; 4.95 mL of methanol was then added. The samples were then vortexed, placed in a tube roller machine for 30 minutes, and centrifuged at 4600 rpm for 10 min. Then, 1 mL of each extract was decanted and transferred to a 5 mL glass tube for drying under a stream of nitrogen gas at 45ºC for 90 minutes. For the following step, the dry residues were reconstituted in 500 µL of MBio assay buffer.

Evaluation of a multiplexed assay for the simultaneous detection of CPF and PT in a real sample

Similar to the multiplexed calibration curve in buffer, waveguides, cartridges and antibody solutions were prepared and twelve strawberry samples were spiked per the sample preparation process outlined above. After extraction, each dry residue was reconstituted with 500 µL of MBio assay buffer, and a concentrated dilution (dilution × 2) of CPF-pAb and PT-mAb 1E2 were prepared at 1/50 and 1/125, respectively, before mixing both antibodies at 1:1 to achieve the required dilutions (1/100 and 1/250 dilution,

178

Chapter 4

respectively). Then, each sample mixture/antibody dilution was again mixed at a 1:1 ratio. Both secondary antibodies (goat anti-rabbit alexa-647 and goat anti-mouse alexa-647) were prepared directly at a dilution 1/200 in 1 mL by mixing 5µL of each one with 990 µL (MBio assay buffer). Finally, 150 µL of each mixture sample/antibody) was pipetted in each cartridge inlet and was incubated for 5 minutes before 150 μL of the mixed secondary antibodies solution was added and the mixture was incubated for an additional 5 minutes.

4.4 Results and discussion

This work describes attempt of a small-scale study conducted as a proof of concept to develop a multiplexed indirect competitive immunoassay for the simultaneous detection of CPF and PT.

4.4.1 MBio assay protocol

Based on the work described in Chapter 3, which provided the optimal concentrations of pesticide conjugates and antibody dilutions, the spotted reagents allowed us to evaluate the binding between antibodies and conjugates that were immobilized onto the waveguides. Previously, CPF-pAb and PT-mAb 1E2 used with the blank samples were optimised in buffer to provide responses of between 150 and 300 at 252.5 AU and 172 AU, respectively. Theareafter, protocols and procedures to spot and optimise microarrays for CPF and PT were explored (Chapter 3) in order to provide the most favourable concentrations of 100 μg/mL for CPF-OVA and H11-OVA. Although both conjugates had similar concentrations, the antibody dilutions were different (1/100 for CPF-pAb and 1/250 for PT-mAb 1E2), PT-mAb 1E2 was more concentrated than CPF-pAb, and was required to be diluted to show the inhibition.

4.4.2 Specificity evaluation

4.4.2.1 Evaluation of pesticide conjugates versus CPF-pAb

After the measurements were completed for the first set of six cartridges and responses were generated from both conjugates at CPF concentrations 0-100 ng/mL using the MBio reader and a secondary antibody (GaR) to detect binding, the results were plotted against CPF concentrations (Figure 4-3). The curves illustrated that CPF-pAb showed specificity

179

Chapter 4

towards CPF-OVA but had no affinity toward H11-OVA at all concentrations. Also, higher CPF concentrations than 100 ng/mL were required to achieve inhibition of the assay.

250

200

CPF-OVA 150 PT-OVA

100

Fluorescence intensity

0 0.1 1 10 100 Chlorpyrifos concentrations (ng/mL)

Figure 4-3: CPF-pAb against both CPF-OVA and H11-OVA using CPF standards in MBio assay buffer.

4.4.2.2 Evaluation of pesticide conjugates versus PT-mAb 1E2

The responses were generated from both coating antigens at PT concentrations 0-100 ng /mL (Section 4.3.4.2); the results are shown in Figure 4.4. The results showed that PT- mAb 1E2 recognised both conjugates (CPF-OVA and H11-OVA) with different affinities, and produced a response of 101 AU for H11-OVA and 180 AU for CPF-OVA. Surprisingly, the response of CPF-OVA showed higher responses towards PT-mAb 1E2.

180

Chapter 4

200

150 CPF-OVA PT-OVA 100

Fluorescence intensity

0 0.1 1 10 100

Parathion concentrations (ng/mL)

120 100

60 CPF-OVA

40 H11-OVA

Normlised signal 20

0 0.1 1 10 100

Parathion concentrations (ng/ml)

Figure 4-4: PT-mAb 1E2 against both H11-OVA and CPF-OVA using PT standard in MBio assay buffer. Fluorescent responses (A) and normalised data (B).

181

Chapter 4

Evidently, PT-mAb 1E2 recognised both coating antigens with higher affinity to CPF- OVA. The bind between PT-mAb and PT-OVA was estimated to be about 45% weaker than the bind between PT-mAb and CPF-OVA. Moreover, both conjugates showed dynamic ranges: the calculated IC20, IC50 and IC80 values for H11-OVA were found to be 2.5, 6 and 14.8 ng/mL, respectively, and for CPF-OVA, 2.6, 6.2 and 15.7, respectively.

4.4.3 Evaluation of a multiplexed assay for OPPs detection in buffer

The last six cartridges were used to evaluate multiplex immunoassay performance for the detection of both analytes. For a competitive immunoassay using both antibodies, analyte standards and secondary antibodies (GaR and GaM), the results were plotted separately, as different concentration ranges were used for CPF and PT. Figure 4-5 shows the results of CPF-pAb inhibition in a multiplexing assay.

400

300 CPF-OVA H11-OVA 200

100

Fluorescence intensity

0 0.1 1 10 100

Chlorpyrifos concentration (ng/mL)

182

Chapter 4

120

100

80 CPF-OVA 60 H11-OVA

Normlised signal (%) 20

0 0.1 1 10 100

Chlorpyrifos concentrations (ng/mL)

Figure 4-5: Multiplexed calibration curves for both Abs with CPF-OVA and H11-OVA. Fluorescent responses against CPF (A) and normalised responses data (B).

Figures 4-3 and 4-4 (A) above clearly show relatively close responses for 200 and 180 AU, respectively, under different antibodies for the blank sample of CPF-OVA. However, a considerable increase in the response shifted from around 200 to 385 AU (Figure 4-5 (A)) when the response from the multiplex assay was compared with the result for a single assay using only CPF-pAb and GaR. This could be explained by the fact that the higher concentrations of both CPF-pAb and PT-mAb 1E2 recognised the coating antigen CPF- OVA in the assay. This suggests that they were available for binding, and lead to their detection by both labelled-secondary antibodies, which caused the amplification of the response. Further, inhibition levels of IC20, IC50 and IC80 were calculated for CPF concentrations as 1.5, 10.1 and 40.3 ng/mL, respectively.

183

Chapter 4

400

300 CPF-OVA H11-OVA 200

100

Fluorescence intensity

0 0.1 1 10 100

Paration concentrations (ng/mL)

120

100 CPF-OVA 80 H11-OVA 60

Normlised data (%) 20

0 0.1 1 10 100

Paration concentrations (ng/mL)

Figure 4-6: Multiplexed calibration curves developed to detect CPF and P using both Abs T. Fluorescent responses against PT (A) and normalised responses (B).

The responses generated from the same multiplexed assay were used again to separately plot the calibration curve against PT concentrations (Figure 4-6). The response of PT- OVA for the blank sample showed no response against CPF-pAb (Figure 4-3). However, when the comparison was made between the results for only PT-reagents, a slight increase

184

Chapter 4

was evident from around 100.3 to 121.3 AU (Figure 4-4 A and 4-5 A). This could be attributed to the fact that PT-mAb 1E2, which has a broad specificity, was used for PT detection and that it partially recognised PT-OVA as well as other OPPs. For this assay, the inhibition levels of IC20, IC50 and IC80 for PT were calculated as 2.9, 7.6 and 19.2 ng/mL, respectively. Table 4-1 outlines the responses generated at different conditions.

Table 4-1: Comparison of different responses generated with Ab mixed in buffer equivalent to the zero standard in buffer. Reagents CPF-OVA H11-OVA CPF + CPF-pAb 200 17 PT + PT-mAb 180 100.3 Mixed reagents and standards 385 121.3

4.4.3.1 Evaluation of a multiplexed assay for OPPs detection in a real sample Following the development of multiplexed calibration curves to analyse CPF and PT in spiked strawberry samples before extraction (see Section 4.3.6), the responses were separately plotted against each pesticide (Figure 4-7).

600

400 CPF-OVA H11-OVA

200

Fluorescence intensity

0 0.1 1 10 100 1000 10000

Chlorpyrifos concentrations (ng/g)

185

Chapter 4

600 500

400 CPF-OVA H11-OVA 300

200

Fluorescence intensity 100

0 0.1 1 10 100 1000

Parathion concentrations (ng/g)

Figure 4-7: Multiplexed calibration curves for CPF and PT in a real sample. CPF concentrations (A) and PT concentrations (B).

Figure 4-7 shows that matrix effects slightly suppressed the CPF curve and amplified the PT curve responses for the blank samples. The CPF curve was suppressed by about 8.6% from 385 AU in buffer to 351.9 AU in the matrix, and the PT response was amplified by about 431% from 121 AU in buffer to 521.5 AU in matrix. In addition, the results showed that the food matrix contained different constituents that can be extracted with the target pesticide. Put together, these factors can either negatively or positively influence assay performance (Pugazhenthi and Ayyadurai, 2013). Another issue encountered was that of cross-reactivity from the broad specificity antibody (PT-mAb 1E2), which made it difficult to precisely quantify the pesticides. Under these conditions, the inhibition levels of IC50 and IC80 for PT-mAb 1E2 were found to be 884.5 and 2408 ng/mL, respectively, for CPF-pAb and 160 and 252.7 ng/mL, respectively for PT-mAb 1E2. However, the IC20 could not be calculated due to the reduced assay sensitivity, which requires further sample preparation to improve the method. Thus, the current multiplex microarray was not able to detect both pesticides due to the cross-reactivity and interference caused by the reagents observed in the experiment.

186

Chapter 4

4.5 Conclusion

Two antibodies, CPF-pAb, which is specific for chlorpyrifos, and PT-mAb 1E2, which has shown a broad specificity for organophosphate pesticides, which based on the development and evaluation of single assays, were evaluated for the feasibility of a multiplexed assay. Here, the suitability of transferring the concept of multiplexing using the MBio platform to detect CPF and PT simultaneously was evaluated. Following the printing of CPF-OVA and H11-OVA onto planar waveguides, the molecular interactions of each antibody were evaluated both individually and in combination. The resulting calibration curves developed both in buffer and for a real sample indicated that there was cross-reactivity, and PT-mAb- 1E2 showed affinity to both pesticides (CPF and PT). However, CPF-pAb was only specific for CPF. This illustrates the compatibility-related challenges of reagents in constructing a single and specific assay, the sensitivities required relative to each antibody used and maximum desired residue levels to be detected. However, because PT-mAb 1E2 showed specificity to a number of organophosphate pesticides, the multiplex assay for specific pesticide identification was not feasible under these conditions but could be considered further as a qualitative tool. Nonetheless, it is worth considering this work as the basis for further study using CPF-pAb and other specific antibodies to develop highly sensitive singleplex assays before moving to a multiplex assay combinations for pesticide determination using a single channel planar waveguide biosensor.

187

Chapter 5

Chapter 5: Conclusions

188

Chapter 5

5.1 Conclusions

Currently, many pesticides have been produced for the agriculture sector to control many pests and weeds. Over time, the organophosphate and carbamate pesticide classes have replaced the organochlorine class, which has been banned from use. While many pesticides disperse or degrade quickly in ecosystems, in other cases, humans may become directly exposed to traces of pesticide residue through crop applications and consumption of contaminated food throughout the food chain which can pose risks to human health. Currently, traditional analytical methods are still used for routine pesticide analysis in water and food samples. As a result, due to various concerns regarding food security and trade barriers between countries, more rapid, sensitive and reliable analytical methods to detect pesticide residues are necessary.

The overall aim of this research was to develop a rapid diagnostics system for pesticide analysis utilising a combination of bioanalytical tools. The first element was the evaluation of molecularly imprinted polymer sorbents for their equivalency to SPE for sample preparation. The concept was to employ such techniques to be utilised as effective sample preparation and clean up tools as the front line component of food analysis implementing rapid, sensitive and cost-effective (bio)analytical methods for OPPs. The second component was the construction and design of a biosensor approach as the rapid sensitive detection system for the identification of individual pesticides or multiple pesticides simultaneously.

The study objectives and achievements are as follows:

Objective 1: Development of molecularly imprinted solid-phase extraction

In order to purify pesticides and quantify them reliably in trace amounts from a complex sample, a clean-up step is required. As a primary study for porogen selection, an evaluation of different molecularly imprinted polymers was first conducted with four different porogens (ACN, DCM, CHL and TOL) using bulk methods for imprinting two different templates (CIPC and CPF) (see Chapter 3). Only a polymer developed using ACN as a porogen with an aqueous solution as a carrier for CPF showed complete retention and thus promising results. In addition, a mass capacity study of the CPF- polymer was higher than 16.3µg/g for the polymer. The developed polymer showed some

189

Chapter 5

selectivity to CPF and could be used for sample purification. In addition, the best combination of porogens and polymerisation reagents can be used for a further study applying the method on a real sample. Both MIPs and antibodies have been used to purify many samples for pesticide separation as well as incorporation into an affinity biosensor. Table 5-1 lists advantages and disadvantages of MIPs compared to antibodies.

Table 5-1 Summary of advantages and disadvantages of MIPs for analytical applications (Mahony et al., 2005). Advantages Disadvantages Cost-effective alternative to biomolecule- Lower catalytic capabilities than biological based recognition counterparts Ease of preparation; enhanced thermal and Binding site heterogeneity providing a distribution of chemical stability vs. antibodies binding site affinities Can be prepared in different formats Template bleeding requires suitable template (bead/block/thin film) following need of the analogue for the imprinting step and affects application quantitative applications Can be stored for years without loss of affinity Grinding and sieving of bulk polymer for SPE/LC for target analyte applications is labour-intensive and inefficient in material yield. Techniques for bead formation may be necessary to obtain more homogeneous MIP particles within a given size range for better chromatographic performance

Both MIPS and antibodies have their specific advantages and disadvantages but no system offers perfection as yet or is without its difficulties. For future research both MIPS and antibodies may be considered further as affinity clean up tools for sample preparation that could be applied in tandem for biosensor analysis.

Objective 2: Detection of organophosphate pesticide using a planar waveguide immunosensor

The residues of CPF and PT, toxic compounds, can be found in many agriculture and environmental products, and samples are typically analysed in a laboratory setting for detection by conventional methods, such as HPLC and GC. However, it is impossible to conduct such analysis on-site. This research has demonstrated the potential of rapid and sensitive methods for a competitive inhibition assay in 10 minutes after extraction in strawberry samples as the model food. The assay was developed using a fluorescence planar waveguide immunosensor, a technique that is potentially portable and can be used

190

Chapter 5

for the semi-quantitative screening in situ. The extraction of CPF from spiked strawberry samples was achieved in less than 1 hour and 1:4 extract dilution in MBio assay buffer was appropriately selected for application with the MBio platform. Sample preparation for PT was optimised using strawberry samples but whereby a concentration step was required to achieve a satisfactory sensitivity threshold relative to the legislative level. Unfortunately, this concentration step involving the evaporation of solvent and reconstitution in a lower volume of buffer is less favourable for an application in rapid screening and semi-quantitative analysis in the field. Furthermore, analytical results for CPF showed LOD and LOQ of 28.2 ng/g and 79.1 ng/g, respectively, and a dynamic range of 11.4-1629.6 ng/g with an average IC50 of 185.3 ng/g. The study results also demonstrated average recoveries 109.5 ±5.6, with a coefficient variation of less than 17%. The PT quantification results showed 6.5 ng/g and 15.4 ng/g for LOD and LOQ, respectively, and a dynamic range of 10.1-147.4 ng/g with IC50 of 44.5 ng/g. The initial validation process showed suitable repeatability and reproducibility using the MBio platform, which has the potential to be applied for food products in field screening tests prior to confirmatory analysis using conventional physicochemical methods. In addition, this method could be used for follow-up monitoring for both targeted quality and safety controls. However, to address the unfavourable side of a broad-specificity antibody, the method may be adjusted in the future to provide a qualitative result for key organophosphates for which the antibody can detect. PT may not be specifically measured but a total measure of pesticide contamination for a group of pesticides could be determined.

Objective 3: Challenges in the development of a multiplexing array to detect organophosphate pesticides

Currently, many different methods for pesticide detection exist that are able to identify targets at and often below tolerance levels, including single residue methods (SRMs) and multiresidue methods (MRMs). Although these methods are reliable and used for regulatory purposes, their drawbacks include long analysis durations, the need for trained personnel, relatively high costs and the need to be performed in a laboratory. Alternatively, the multiplex MBio planar waveguide biosensor assay can provide a semi- quantitative method capable of quantifying targets present in a sample. An attempt to

191

Chapter 5

multiplex two pesticides (CPF and PT) in a single assay was conducted using one sample preparation procedure for both analytes in buffer and real samples. The results showed a higher CPF signal in buffer perhaps as a result of the broad specificity antibody used for PT. However, the signals were amplified dramatically when the evaluation was performed using a real sample due to matrix effects. Nevertheless, this attempt showed the promise of the multiplex applicability for pesticide analysis. To be able to perform effective multiplex analysis, antibodies specific to individual pesticides would be the optimal approach. Though, these are currently unavailable commercially and would need to be considered in relation to the carrier protein used in production and the biosensor array to be constructed to minimise interference. If utilising a broad specificity antibody it would be more suitable if the cross-reactivity was 100% for all pesticides and whereby the total pesticide levels could be determined for the known group and a qualitative response provided. Both recognition elements, MIP and antibody have been used to purify many samples for pesticides separation as well as incorporated for affinity biosensor.

Both MIPs and antibodies have been used as affinity capture tools for separation columns and diagnostic platforms. For most commercial applications, antibodies prevail with a wide share of the market. They are undeniably extremely specific and selective for several kinds of chemicals and can be generated on a large scale. The use of antibodies in diagnostic assays is widely recognised and they are considered the gold standard in relation to sensitivity and affinity. Regrettably, they also show some drawbacks and cannot be always the best option to use because of instability out of their aqueous environment or under harsh conditions such as organic solvents, extreme pH or temperature. Moreover, the production process can be expensive and time consuming to ensure an antibody of a certain specificity or selectivity is produced. The production of antibodies for small molecules it is not straightforward whereby chemical coupling between haptens and a suitable carrier protein to generate an immune response is required; the use of animals is a necessity with occasionally a lengthy immunisation plan and their purification and characterisation requires additional expertise. For MIPs to compete with antibodies, they need to acquire the features of antibodies such as solubility, specificity and affinity while preserving the advantages of MIPs such as low cost, a short development time and high stability, high tolerance for mechanical stress, high pressures,

192

Chapter 5

high temperatures, wide pH ranges, a variety of solvents. In addition, MIPs do not require the use of animals, and can be prepared for a wide variety of targets. However, the well- known disadvantages such as the necessity for template molecules via MIP production, the cross-reactivity often correlated with MIPs, issues in template removal where it is not always easy to completely remove the template leading to potentially leaching during the analysis, showing inaccurate detections explains their limited applications. Furthermore, the gap between pure research and application needs to be reduced or eliminated, such as scaling-up and reproducibility in order to fulfil real implementations. MIPs integration into sensors or transformation of the template binding into an electrical signal can also be laborious. Although MIPs have been produced and used in sensors, as recognition elements mimicking antibodies and catalysts, MIPs are most frequently applied as affinity adsorbents for sample pre-concentration and separation in HPLC and SPE systems.

5.2 Future work and perspectives

Commercially, while agricultural practices are increasingly dependent on pesticides for plant production safety and quality, these practices have resulted in contamination and samples that can exceed tolerance levels, particularly in developing countries whose monitoring systems are not well-established. It is therefore necessary to develop techniques for sample purification as well as portable methods for rapid detection for on- site analysis. Immunosensors are well-known for providing sensitive, simple and relatively inexpensive methods for single or multi-residue assays and portable systems that can become a means to ensure human health and animal feed. The current research showed a number of findings that deserve further investigation. First, with respect to developing molecularly imprinted solid-phase extraction, it is worth testing acetonitrile as a porogen to synthesise magnetic MIPs. Other aspects relating to the type of functional monomer in terms of a level of acidity as well as the ratio of a template to the functional monomer are worthy of investigation. In addition, NIPs where non-specific binding sites are present on the backbone of the polymer, have been used as sorbent materials for sample purification and worth testing the NIP further especially when working with relatively a high concentration in a scale of nanogram. Furthermore, the pH of a solution, loading volume, washing and elution mixtures play a role in improving the MISPE protocol. The relative number of MIP applications for sample preparation have increased

193

Chapter 5

and have been used for sample purification and pre-concentration as clean-up techniques for food samples. In addition to the several applications that have been developed, including SPE, magnetic MIP and others, new arrays of MIP forms have been produced, including hybrid composites between MIPs and other materials. As a result, the number of commercial MIP applications will most likely increase over time. Though also worthy of note are the potential of antibody based immunoaffinity columns as future applications in sample clean up and concentration.

Second, the assay method developed for CPF detection in strawberry samples using an MBio platform showed promising results and should be further validated using different matrices. Third, the assay evaluated for PT detection in a strawberry samples using PT.mAb 1E2 should include the remaining pesticides and use the percentage inhibition rate. This process could make the method applicable for a qualitative screening assay. Finally, little research has been conducted for on-site multi-pesticide detection by biosensor using MBio platforms for multiplexing. The current work demonstrated the feasibility of transferring the technique into the field of pesticide analysis for on-site rapid detection. Therefore, other specific antibodies that are produced for different organophosphate pesticides as well as CPF should be used to develop a pioneer portable assay for multi-pesticide detection in a real agriculture sample. The immunoassay can be investigated further to improve the sensitivity of the detection limit by looking at different physicochemical factors influencing assay performance, such as pH, ionic strength, blocking solution, and diluting solution, were optimized.

Multiplexing is a new trend in rapid analysis for the detection of several targets in a single assay. In the last 10 years, new devices, such as MBio and Luminex Xmap, have been introduced for multiple analyte detection with the capability to multiplex different targets in one assay. However, to date few studies exist that investigate pesticide immunosensor multiplexed detection. However, these platforms are promising and further research in the pesticide field is expected in the near future.

Biosensor applications showed capabilities for monitoring environmental samples, and have used arrays of recognition elements to efficiently construct analytical platforms that can detect targets in samples of varying complexity. However, while considerable work

194

Chapter 5

has been done on enzymes and antibodies to improve biosensor performance—for example, using genetically engineered microorganisms and antibody fragments— molecularly imprinted polymers, aptamers and nanotechnology are new sensing materials that have been used to develop new biosensors and improve their detection sensitivities. These promising new elements have a wide range of properties. Generally speaking, the environmental disciplines have not fully explored the capabilities of biosensors, especially compared to their applications in the field of public health, for which a major share of commercially available biosensors are used for medical applications. However, challenges remain for these techniques in terms of constructing more reliable pesticide detection tools that match the capabilities for use in the field compared to those of multi- analytes analysis now offered by mass spectrometry in the laboratory setting.

A broad range of biosensors—such as enzyme-, antibody-, whole-cell and MIP-based biosensors and nanosensors—has been developed to detect OPP residues with varying detection limits. In addition, biosensors have been used to monitor OPP residue levels in buffer, water, vegetable, fruit, juice and soil samples and have exhibited a range of detection limits, storage stabilities, reproducibility and sensitivity.

Clearly, the use of biosensors has proved to be a viable assay method for OPP analysis in a range of environmental samples. In addition, a number of enzymes, mainly AChE and antibody-based biosensors, have been produced all over the world. Enzyme-based biosensors are sensitive to the presence of targets and are able to detect OPP traces well below the maximum residue level for samples. Indeed, biosensor applications have become increasingly promising and their resultant signals could be greatly improved through different approaches. However, while these biosensor applications have their advantages, at present, they are used only for screening due to broad enzyme specificity. Thus, it would be necessary to develop additional biosensors for future applications especially for those based on microarray techniques due to several advantages such as sensitivity; enhanced reproducibility; low sample consumption; reduced processing time, and ease of computerization.

Additional studies have reported biosensor detection of OPPs using materials such as antibodies, whole-cells, DNA, and artificial materials such as MIP and aptamer. Using antibody-based biosensors is a well-established technique for pesticide detection and 195

Chapter 5

specific to targets that use Ab and Ag for OPP detection. In addition, emerging techniques such as molecular imprinting polymers are gaining increased attention due to their benefits. These polymers can be produced to detect different targets, including pesticides, and exhibit favourable shelf-life stability and improved detection limits. However, more research is required to improve and standardise polymerisation methods and increase the number of detected pesticides in a given sample. Moreover, biosensor-based nanomaterials have been increasingly used in biosensor platforms, as they offer innovative properties in many fields, and have improved detection limits. However, most of these developed platforms were applied in buffer assays and thus should be fully tested with environmental samples. Therefore, additional research needs to be done to develop more robust biosensor platforms to allow for the lowest possible detection limits for OPP residues in real samples, as well as improving their stability, repeatability and bio- recognition elements.

196

Chapter 6

Chapter 6: References

Abdollahi, M., Ranjbar, A., Shadnia, S., Nikfar, S., & Rezaie, A. (2004). Pesticides and oxidative stress: A review. Medical Science Monitor: International Medical Journal of Experimental and Clinical Research, 10(6), RA141-7. doi:4163 [pii]

Alavanja, M. C., Ross, M. K., & Bonner, M. R. (2013). Increased cancer burden among pesticide applicators and others due to pesticide exposure. CA: A Cancer Journal for Clinicians, 63(2), 120-142.

Alcocer, M., Dillon, P., Manning, B., Doyen, C., Lee, H., Daly, S., et al. (2000). Use of phosphonic acid as a generic hapten in the production of broad specificity anti- organophosphate pesticide antibody. Journal of Agricultural and Food Chemistry, 48(6), 2228-2233.

Alexander, C., Andersson, H. S., Andersson, L. I., Ansell, R. J., Kirsch, N., Nicholls, I. A., et al. (2006). Molecular imprinting science and technology: A survey of the literature for the years up to and including 2003. Journal of Molecular Recognition, 19(2), 106-180.

Alizadeh, T. (2010). Comparison of different methodologies for integration of molecularly imprinted polymer and electrochemical transducer in order to develop a paraoxon voltammetric sensor. Thin Solid Films, 518(21), 6099-6106.

Al-Rimawi, F. (2014). a hplc-uv method for deteermination of three pesticides in water. International Journal of Advances in Chemistry, 2(2), 1-8.

Ameno, K. (2005). Carbamate pesticides. Drugs and poisons in humans (pp. 559-570) Springer.

Amulen, D. R., Spanoghe, P., Houbraken, M., Tamale, A., de Graaf, D. C., Cross, P., et al. (2017). Environmental contaminants of honeybee products in uganda detected using LC-MS/MS and GC-ECD. PloS One, 12(6), e0178546.

Andreu, V., & Picó, Y. (2012). Determination of currently used pesticides in biota. Analytical and Bioanalytical Chemistry, 404(9), 2659-2681.

Anfossi, L., Baggiani, C., Baravalle, P., Giovannoli, C., Guzzella, L., & Pozzoni, F. (2009). Molecular recognition of the fungicide carbendazim by a molecular imprinted polymer obtained through a mimic template approach. Analytical Letters, 42(5), 807-820.

197

Chapter 6

Angenendt, P. (2005). Progress in protein and antibody microarray technology. Drug Discovery Today, 10(7), 503-511.

Arjmand, M., Saghafifar, H., Alijanianzadeh, M., & Soltanolkotabi, M. (2017). A sensitive tapered-fiber optic biosensor for the label-free detection of organophosphate pesticides. Sensors and Actuators B: Chemical, 249, 523-532.

Audrey, S., Beatriz, P., & Jean-Louis, M. (2012). Biosensors for pesticide detection: New trends. American Journal of Analytical Chemistry, 2012

Baggiani, C., Anfossi, L., Baravalle, P., Giovannoli, C., & Giraudi, G. (2007). Molecular recognition of polycyclic aromatic hydrocarbons by pyrene-imprinted microspheres. Analytical and Bioanalytical Chemistry, 389(2), 413-422.

Baggiani, C., Baravalle, P., Anfossi, L., & Tozzi, C. (2005). Comparison of pyrimethanil- imprinted beads and bulk polymer as stationary phase by non-linear chromatography. Analytica Chimica Acta, 542(1), 125-134.

Bakas, I., Oujji, N. B., Istamboulié, G., Piletsky, S., Piletska, E., Ait-Addi, E., et al. (2014). Molecularly imprinted polymer cartridges coupled to high performance liquid chromatography (HPLC-UV) for simple and rapid analysis of fenthion in olive oil. Talanta, 125, 313-318.

Bakas, I., Oujji, N. B., Moczko, E., Istamboulie, G., Piletsky, S., Piletska, E., et al. (2013). Computational and experimental investigation of molecular imprinted polymers for selective extraction of dimethoate and its metabolite omethoate from olive oil. Journal of Chromatography A, 1274, 13-18.

Bala, R., Sharma, R. K., & Wangoo, N. (2015). Highly sensitive colorimetric detection of ethyl parathion using gold nanoprobes. Sensors and Actuators B: Chemical, 210, 425-430.

Baldim, I. M., de Oliveira Souza, Marília Cristina, da Cunha Souza, João Carlos Jacinto, Figueiredo, E. C., & Martins, I. (2012). Application of the molecularly imprinted solid-phase extraction to the organophosphate residues determination in strawberries. Analytical and Bioanalytical Chemistry, 404(6-7), 1959-1966.

Ban, B., & Blake, D. A. (2012). Recombinant Antibodies and Non-Antibody Scaffolds for Immunoassays. In Advances in Immunoassay Technology. InTech.

Bao, J., Hou, C., Dong, Q., Ma, X., Chen, J., Huo, D., et al. (2016). ELP-OPH/BSA/TiO 2 nanofibers/c-MWCNTs based biosensor for sensitive and selective determination

198

Chapter 6

of p-nitrophenyl substituted organophosphate pesticides in aqueous system. Biosensors and Bioelectronics, 85, 935-942.

Bao, J., Hou, C., Huo, D., Dong, Q., Ma, X., Sun, X., et al. (2017). Sensitive and selective electrochemical biosensor based on ELP-OPH/BSA/TiO2NFs/AuNPs for determination of organophosphate pesticides with p-nitrophenyl substituent. Journal of the Electrochemical Society, 164(2), G17-G22.

Barrek, S., Paisse, O., & Grenier-Loustalot, M. (2003). Determination of residual pesticides in olive oil by GC–MS and HPLC–MS after extraction by size-exclusion chromatography. Analytical and Bioanalytical Chemistry, 376(3), 355-359.

Barriada-Pereira, M., González-Castro, M., Muniategui-Lorenzo, S., López-Mahía, P., Prada-Rodríguez, D., & Fernández-Fernández, E. (2007). Comparison of pressurized liquid extraction and microwave assisted extraction for the determination of organochlorine pesticides in vegetables. Talanta, 71(3), 1345-1351.

Basta, N., Ryan, J., & Chaney, R. (2005). Trace element chemistry in residual-treated soil. Journal of Environmental Quality, 34(1), 49-63.

Bazmandegan-Shamili, A., Dadfarnia, S., Shabani, A. M. H., Saeidi, M., & Moghadam, M. R. (2016). High-performance liquid chromatographic determination of diazinon after its magnetic dispersive solid-phase microextraction using magnetic molecularly imprinted polymer. Food Analytical Methods, 9(9), 2621-2630.

Beltran, A., Borrull, F., Marcé, R., & Cormack, P. (2010). Molecularly-imprinted polymers: Useful sorbents for selective extractions. TrAC Trends in Analytical Chemistry, 29(11), 1363-1375.

Berset, J., Mermer, S., Robel, A., Walton, V., Chien, M., & Field, J. (2017). Direct residue analysis of systemic insecticides and some of their relevant metabolites in wines by liquid chromatography–mass spectrometry. Journal of Chromatography A, 1506, 45- 54.

Beyer, A., & Biziuk, M. (2010). Comparison of efficiency of different sorbents used during clean-up of extracts for determination of polychlorinated biphenyls and pesticide residues in low-fat food. Food Research International, 43(3), 831-837.

Biagini, R., Smith, J., Sammons, D., MacKenzie, B., Striley, C., Robertson, S., et al. (2004). Development of a sensitivity enhanced multiplexed fluorescence covalent microbead immunosorbent assay (FCMIA) for the measurement of glyphosate,

199

Chapter 6

atrazine and metolachlor mercapturate in water and urine. Analytical and Bioanalytical Chemistry, 379(3), 368-374.

Bock, S. A., Munoz-Furlong, A., & Sampson, H. A. (2007). Further fatalities caused by anaphylactic reactions to food, 2001-2006. The Journal of Allergy and Clinical Immunology, 119(4), 1016-1018. doi:S0091-6749(06)03814-0 [pii]

Bordin, A. B., Minetto, L., do Nascimento Filho, I., Beal, L. L., & Moura, S. (2017). Determination of pesticide residues in whole wheat flour using modified QuEChERS and LC–MS/MS. Food Analytical Methods, 10(1), 1-9.

Brancato, A., Brocca, D., De Lentdecker, C., Erdos, Z., Ferreira, L., Greco, L., et al. (2017). Review of the existing maximum residue levels for chlorpyrifos according to article 12 of regulation (EC) no 396/2005. EFSA Journal, 15(3).

Brüggemann, O., Haupt, K., Ye, L., Yilmaz, E., & Mosbach, K. (2000). New configurations and applications of molecularly imprinted polymers. Journal of Chromatography A, 889(1), 15-24.

Bui, M. N., & Abbas, A. (2015). Simple and rapid colorimetric detection of p-nitrophenyl substituent organophosphorous nerve agents. Sensors and Actuators B: Chemical, 207, 370-374.

Cameron, L., Lalich, N., Bauer, S., Booker, V., Bogue, H., Samuels, S., et al. (2006). Occupational health survey of farm workers by camp health aides. Journal of Agricultural Safety and Health, 12(2), 139-153.

Camire, M. E., Bushway, R. J., Zhao, J., Perkins, B., & Paradis, L. R. (1995). Fate of thiabendazole and chlorpropham residues in extruded potato peels. Journal of Agricultural and Food Chemistry, 43(2), 495-497.

Campanella, L., Eremin, S., Lelo, D., Martini, E., & Tomassetti, M. (2011). Reliable new immunosensor for atrazine pesticide analysis. Sensors and Actuators B: Chemical, 156(1), 50-62.

Cao, Y., Sun, X., Guo, Y., Zhao, W., & Wang, X. (2015). An electrochemical immunosensor based on interdigitated array microelectrode for the detection of chlorpyrifos. Bioprocess and Biosystems Engineering, 38(2), 307-313.

Cardeal, Z. L., Souza, A. G., & Amorim, L. C. (2011). Analytical methods for performing pesticide degradation studies in environmental samples INTECH Open Access Publisher.

200

Chapter 6

Chambers, J. E., & Levi, P. E. (1992). Organophosphates: Chemistry, fate, and effects.

Chand, S., & Gupta, B. (2007). Surface plasmon resonance based fiber-optic sensor for the detection of pesticide. Sensors and Actuators B: Chemical, 123(2), 661-666.

Chang, M., Shu-ling, Y., Hong, Z., Xiu-zhen, G., Zi-yu, C., Lin, S., et al. (2011). Preparation of malathion imprinted polymer and its application for solid-phase extraction. Chinese Journal of Analysis Laboratory, 7, 016.

Chapuis, F., Pichon, V., Lanza, F., Sellergren, B., & Hennion, M. (2004). Retention mechanism of analytes in the solid-phase extraction process using molecularly imprinted polymers: Application to the extraction of triazines from complex matrices. Journal of Chromatography B, 804(1), 93-101.

Chen, D., Jiao, Y., Jia, H., Guo, Y., Sun, X., Wang, X., et al. (2015). Acetylcholinesterase biosensor for chlorpyrifos detection based on multi-walled carbon nanotubes-SnO2- chitosan nanocomposite modified screen-printed electrode. Int.J.Electrochem.Sci, 10, 10491-10501.

Chen, G., Jin, M., Du, P., Zhang, C., Cui, X., Zhang, Y., et al. (2017b). A sensitive chemiluminescence enzyme immunoassay based on molecularly imprinted polymers solid-phase extraction of parathion. Analytical Biochemistry, 530, 87-93.

Chen, J., Duan, C., & Guan, Y. (2010a). Sorptive extraction techniques in sample preparation for organophosphorus pesticides in complex matrices. Journal of Chromatography B, 878(17), 1216-1225.

Chen, L., & Li, B. (2012). Application of magnetic molecularly imprinted polymers in analytical chemistry. Analytical Methods, 4(9), 2613-2621.

Chen, L., Ding, L., Jin, H., Song, D., Zhang, H., Li, J., et al. (2007). The determination of organochlorine pesticides based on dynamic microwave-assisted extraction coupled with on-line solid-phase extraction of high-performance liquid chromatography. Analytica Chimica Acta, 589(2), 239-246.

Chen, X., Huang, Y., Duan, N., Wu, S., Ma, X., Xia, Y., et al. (2013). Selection and identification of ssDNA aptamers recognizing zearalenone. Analytical and Bioanalytical Chemistry, 405(20), 6573-6581.

Chen, Y. P., Ning, B., Liu, N., Feng, Y., Liu, Z., Liu, X., et al. (2010b). A rapid and sensitive fluoroimmunoassay based on quantum dot for the detection of chlorpyrifos residue in drinking water. Journal of Environmental Science and Health Part B, 45(6), 508-515.

201

Chapter 6

Chen, Y., Li, J., Lu, P., Hu, D., Xue, W., & Ding, X. (2017a). Development of a polyclonal antibody-based indirect competitive enzyme-linked immunosorbent assay to detect dufulin residue in water, soil and agricultural samples. Food and Agricultural Immunology, 1-12.

Chen, Y., Ren, H. l., Liu, N., Sai, N., Liu, X., Liu, Z., et al. (2010a). A fluoroimmunoassay based on quantum dot− streptavidin conjugate for the detection of chlorpyrifos. Journal of Agricultural and Food Chemistry, 58(16), 8895-8903.

Cho, Y. A., Seok, J., Lee, H., Kim, Y. J., Park, Y. C., & Lee, Y. T. (2004). Synthesis of haptens of organophosphorus pesticides and development of immunoassays for fenitrothion. Analytica Chimica Acta, 522(2), 215-222.

Chong, H., & Ching, C. B. (2016). Development of colorimetric-based whole-cell biosensor for organophosphorus compounds by engineering transcription regulator DmpR. ACS Synthetic Biology, 5(11), 1290-1298.

Cole, D. C., Carpio, F., & León, N. (2000). Economic burden of illness from pesticide poisonings in highland ecuador. Revista Panamericana De Salud Pública, 8(3), 196- 201.

Commission Decision. (2002). No. 2002/657/EC implementing council directive 96/23/EC concerning the performance of analytical methods and the interpretation of results. Official Journal of the European Communities L221:8–34, L221, 8-.

Commission Decision. (2005). Commission decision of 18 novmber 2005 adopting community import decisions for certain chemicals pursuant to regulation (EC) no 304/2003 of the european parliament and of the council and amending decision 2000/657/EC. Official Journal of the European Union, (L 304/46)

Commission regulation. (2016). Commission regulation (eu) 2016/60, january amending annexes ii and iii to regulation (ec) no 396/2005 of the european parliament and of the council as regards maximum residue levels for chlorpyrifos in or on certain products. Official Journal of the European Union, (L 14/1.)

Cormack, P. A., & Elorza, A. Z. (2004). Molecularly imprinted polymers: Synthesis and characterisation. Journal of Chromatography B, 804(1), 173-182.

Cortéjade, A., Kiss, A., Cren, C., Vulliet, E., & Buleté, A. (2016). Development of an analytical method for the targeted screening and multi-residue quantification of environmental contaminants in urine by liquid chromatography coupled to high resolution mass spectrometry for evaluation of human exposures. Talanta, 146, 694- 706.

202

Chapter 6

Cortina-Puig. M. Istamboulie., G. Noguer. T. and Marty. J. L. (01, January, 2010). Analysis of pesticide mixtures using intelligent biosensors In Vernon S. Somerset (Ed.), Intelligent and biosensors (pp. 205-216) InTech.

Council Directive. (2006). Council directive 2006/88/EC of 24 october 2006 on animal health requirements for aquaculture animals and products thereof, and on the prevention and control of certain diseases in aquatic animals. Off J Eur Comm L, 328, 14-56.

Cunnington, A. (2008). Developments in potato storage in Great Britain. Potato Research, 51(3-4), 403.

Curl, C. L., Fenske, R. A., & Elgethun, K. (2003). Organophosphorus pesticide exposure of urban and suburban preschool children with organic and conventional diets. Environmental Health Perspectives, 111(3), 377-382.

Dai, Y., Wang, T., Hu, X., Liu, S., Zhang, M., & Wang, C. (2017). Highly sensitive microcantilever-based immunosensor for the detection of carbofuran in soil and vegetable samples. Food Chemistry, 229, 432-438.

Dankwardt, A. (2001). Immunochemical assays in pesticide analysis. Encyclopedia of Analytical Chemistry.

Dar, A. I., Walia, S., & Acharya, A. (2016). Citric acid-coated gold nanoparticles for visual colorimetric recognition of pesticide dimethoate. Journal of Nanoparticle Research, 8(18), 1-8.

De Barros, L. A., Martins, I., & Rath, S. (2010). A selective molecularly imprinted polymer-solid phase extraction for the determination of fenitrothion in tomatoes. Analytical and Bioanalytical Chemistry, 397(3), 1355-1361.

De Oliveira Silva, R., de Castro, R. C., Milhome, M. A. L., & do Nascimento, R. F. (2014). Liquid chromatography–electrospray ionization–tandem mass spectrometry method for determination of twenty multi-class pesticide residues in cashew. LWT- Food Science and Technology, 59(1), 21-25.

Deep, A., Bhardwaj, S. K., Paul, A., Kim, K., & Kumar, P. (2015). Surface assembly of nano-metal organic framework on amine functionalized indium tin oxide substrate for impedimetric sensing of parathion. Biosensors and Bioelectronics, 65, 226-231.

Della Ventura, B., Iannaccone, M., Funari, R., Ciamarra, M. P., Altucci, C., Capparelli, R., et al. (2017). Effective antibodies immobilization and functionalized

203

Chapter 6

nanoparticles in a quartz-crystal microbalance-based immunosensor for the detection of parathion. PloS One, 12(2), e0171754.

Denderz, N., Lehotay, J., Čižmárik, J., Cibulková, Z., & Šimon, P. (2012). Thermodynamic study of molecularly imprinted polymer used as the stationary phase in high performance liquid chromatography. Journal of Chromatography A, 1235, 77-83.

Devi, P. I. (2009). Health risk perceptions, awareness and handling behaviour of pesticides by farm Workers1. Age (Years), 45(23), 70.

Devlin, S., Meneely, J. P., Greer, B., Greef, C., Lochhead, M. J., & Elliott, C. T. (2013). Next generation planar waveguide detection of microcystins in freshwater and cyanobacterial extracts, utilising a novel lysis method for portable sample preparation and analysis. Analytica Chimica Acta, 769, 108-113.

DHHS, US Department of Health and Human Services. (2017). Agency for toxic substances and disease registry.TOXICOLOGICAL PROFILE FOR PARATHION. Retrieved 03/02, 2017, from https://www.atsdr.cdc.gov/toxprofiles/tp205.pdf

Díaz, A. N., Sánchez, F. G., Bracho, V., Lovillo, J., & Aguilar, A. (1997). Enzymatic determination of fenitrothion by cholinesterase and acetylcholinesterase on fluorogenic substrates. Fresenius' Journal of Analytical Chemistry, 357(7), 958-961.

Diaz, A., Sanchez, F., Aguilar, A., de Vicente, A., & Bautista, A. (2015). Fast stopped- flow enzymatic sensing of fenitrothion in grapes and orange juice. Journal of Food Composition and Analysis, 42, 187-192.

Dikshith, T. S. S. (2016). Handbook of chemicals and safety CRC Press.

Ding, J., Zhang, H., Jia, F., Qin, W., & Du, D. (2014). Assembly of carbon nanotubes on a nanoporous gold electrode for acetylcholinesterase biosensor design. Sensors and Actuators B: Chemical, 199, 284-290.

Donkor, A., Osei-Fosu, P., Dubey, B., Kingsford-Adaboh, R., Ziwu, C., & Asante, I. (2016). Pesticide residues in fruits and vegetables in ghana: A review. Environmental Science and Pollution Research, 23(19), 18966-18987.

Du, D., Chen, A., Xie, Y., Zhang, A., & Lin, Y. (2011). Nanoparticle-based immunosensor with apoferritin templated metallic phosphate label for quantification of phosphorylated acetylcholinesterase. Biosensors and Bioelectronics, 26(9), 3857- 3863.

204

Chapter 6

Du, P., Jin, M., Yang, L., Du, X., Chen, G., Zhang, C., et al. (2015). A rapid immunomagnetic-bead-based immunoassay for triazophos analysis. RSC Advances, 5(99), 81046-81051.

Dzantiev, B. B., Byzova, N. A., Urusov, A. E., & Zherdev, A. V. (2014). Immunochromatographic methods in food analysis. TrAC Trends in Analytical Chemistry, 55, 81-93.

Eaton, D. L., Daroff, R. B., Autrup, H., Bridges, J., Buffler, P., Costa, L. G., et al. (2008). Review of the toxicology of chlorpyrifos with an emphasis on human exposure and neurodevelopment. Critical Reviews in Toxicology, 38(sup2), 1-125.

EFSA, the European Food Safety Authority. (2013). The 2010 european union report on pesticide residues in food. EFSA Journal, 11(3), 3130.

EFSA, the European Food Safety Authority. (2015). The 2013 european union report on pesticide residues in food. EFSA Journal, 13(3:4038,), 169.

El Alami, A., Lagarde, F., Tamer, U., Baitoul, M., & Daniel, P. (2016). Enhanced raman spectroscopy coupled to chemometrics for identification and quantification of acetylcholinesterase inhibitors. Vibrational Spectroscopy, 87, 27-33.

Erdem, M., Say, R., Ersöz, A., Denizli, A., & Türk, H. (2010). Biomimicking, metal- chelating and surface-imprinted polymers for the degradation of pesticides. Reactive and Functional Polymers, 70(4), 238-243.

Esteve-Turrillas, F. A., Abad-Somovilla, A., Quiñones-Reyes, G., Agulló, C., Mercader, J. V., & Abad-Fuentes, A. (2015). Monoclonal antibody-based immunoassays for cyprodinil residue analysis in QuEChERS-based fruit extracts. Food Chemistry, 187, 530-536.

Evtugyn, G. A., Shamagsumova, R. V., Padnya, P. V., Stoikov, I. I., & Antipin, I. S. (2014). Cholinesterase sensor based on glassy carbon electrode modified with ag nanoparticles decorated with macrocyclic ligands. Talanta, 127, 9-17.

Fan, M. T, and J, He. (2011). Pesticide immunoassay. In S. Margarita (Ed.), Pesticides - strategies for pesticides analysis (2nd ed., pp. 293) Croatia : InTech.

Fan, S., Zhao, P., Yu, C., Pan, C., & Li, X. (2014). Simultaneous determination of 36 pesticide residues in spinach and cauliflower by LC-MS/MS using multi-walled carbon nanotubes-based dispersive solid-phase clean-up. Food Additives & Contaminants: Part A, 31(1), 73-82.

205

Chapter 6

Fang, W., Zhang, X., Chen, Y., Wan, L., Huang, W., Shen, A., et al. (2015). Portable SERS-enabled micropipettes for microarea sampling and reliably quantitative detection of surface organic residues. Analytical Chemistry, 87(18), 9217-9224.

Fang, Y., Tian, W., Pei, F., Li, P., Shao, X., Fan, Y., et al. (2017). Simultaneous determination of pesticide residues and antioxidants in blended oil using a liquid- liquid extraction combined with dispersive solid phase extraction method. Food Chemistry, 229, 347-353.

FAO and WHO, Food and Agricultural Organisation of the United Nations and World Health Organisation. (2014). International code of conduct on pesticide management. Retrieved June/06, 2017, from http://www.fao.org/fileadmin/templates/agphome/documents/Pests_Pesticides/Cod e/

FAO, Food and Agriculture Organization of the United Nations. (2001). Definitions for the purposes of the codex alimentarius. Retrieved June/16, 2017, from http://www.fao.org/docrep/005/y2200e/y2200e07.htm.

FAO, Food and Agriculture Organization of the United Nations. (2014).

FAO/UNEP, Joint the United Nations Environment Programme and the Food and Agriculture Organisation of the United Nations. (2005). Decision guidance document: Parathion. secretariat for the rotterdam convention on the prior informed consent procedure for certain hazardous chemicals and pesticides in international trade., 2017, from http://www.pic.int/Portals/5/DGDs/DGD_Parathion_EN.pdf

Farré, M., Picó, Y., & Barceló, D. (2014). Application of ultra-high pressure liquid chromatography linear ion-trap orbitrap to qualitative and quantitative assessment of pesticide residues. Journal of Chromatography A, 1328, 66-79.

Ferrer, I., Garcia-Reyes, J. F., Mezcua, M., Thurman, E. M., & Fernández-Alba, A. R. (2005). Multi-residue pesticide analysis in fruits and vegetables by liquid chromatography–time-of-flight mass spectrometry. Journal of Chromatography A, 1082(1), 81-90.

Fish, W. P., Ferreira, J., Sheardy, R. D., Snow, N. H., & O'Brien, T. P. (2005). Rational Design of an Imprinted Polymer: Maximizing Selectivity by Optimizing the Monomer–Template Ratio for a Cinchonidine MIP, Prior to Polymerization, Using Microcalorimetry. Journal of liquid chromatography & related technologies, 28(1), 1-15.

206

Chapter 6

Fishel, F. M. (2005). Pesticide toxicity profile: Carbamate pesticides. Pub no.PI- 51.Gainesville, FL: Pesticide Information Office, Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of Florida,

Flanagan, R. J., Taylor, A. A., Watson, I. D., & Whelpton, R. (2008). Fundamentals of analytical toxicology John Wiley & Sons.

Franek, M., & Hruska, K. (2005). Antibody based methods for environmental and food analysis: A review. Veterinarni Medicina, 50(1), 1-10.

Funari, R., Della Ventura, B., Carrieri, R., Morra, L., Lahoz, E., Gesuele, F., et al. (2015). Detection of parathion and patulin by quartz-crystal microbalance functionalized by the photonics immobilization technique. Biosensors and Bioelectronics, 67, 224-229.

Funari, R., Della Ventura, B., Schiavo, L., Esposito, R., Altucci, C., & Velotta, R. (2013). Detection of parathion pesticide by quartz crystal microbalance functionalized with UV-activated antibodies. Analytical Chemistry, 85(13), 6392-6397.

Gabaldón, J. A., Maquieira, A., & Puchades, R. (1999). Current trends in immunoassay- based kits for pesticide analysis. Critical Reviews in Food Science and Nutrition, 39(6), 519-538.

Gabaldon, J., Maquieira, A., & Puchades, R. (2007). Development of a simple extraction procedure for chlorpyrifos determination in food samples by immunoassay. Talanta, 71(3), 1001-1010.

Garcés-García, M., Brun, E. M., Puchades, R., & Maquieira, Á. (2006). Immunochemical determination of four organophosphorus insecticide residues in olive oil using a rapid extraction process. Analytica Chimica Acta, 556(2), 347-354.

García, M. G., Duque, S. U., Fernández, A. L., Sosa, A., & Fernández-Alba, A. (2017). Multiresidue method for trace pesticide analysis in honeybee wax comb by GC-QqQ- MS. Talanta, 163, 54-64.

Ge, S., Zhao, P., Yan, M., Zang, D., & Yu, J. (2012). Multi-branch chemiluminescence– molecular imprinting sensor for sequential determination of carbofuran and omethoate in foodstuff. Analytical Methods, 4(10), 3150-3156.

Geng, F., Ding, J., Jia, C., Ding, B., & Qin, W. (2015). A potentiometric biosensing system based on an isolated degrading bacterium klebsiella sp. MP-6 for the determination of methyl parathion. RSC Advances, 5(43), 34475-34480.

207

Chapter 6

Gerent, G. G., & Spinelli, A. (2016). Environmentally-friendly in situ plated bismuth- film electrode for the quantification of the endocrine disruptor parathion in skimmed milk. Journal of Hazardous Materials, 308, 157-163.

Giergielewicz-Możajska, H., Dąbrowski, Ł., & Namieśnik, J. (2001). Accelerated solvent extraction (ASE) in the analysis of environmental solid samples—some aspects of theory and practice. Critical Reviews in Analytical Chemistry, 31(3), 149-165.

Gilani, R. A., Rafique, M., Rehman, A., Munis, M. F. H., & Chaudhary, H. J. (2016). Biodegradation of chlorpyrifos by bacterial genus pseudomonas. Journal of Basic Microbiology, 56(2), 105-119.

Gilbert-López, B., García-Reyes, J. F., & Molina-Díaz, A. (2009). Sample treatment and determination of pesticide residues in fatty vegetable matrices: A review. Talanta, 79(2), 109-128.

Goel, P. (2013). Immunodiagnosis of pesticides: A review. African Journal of Biotechnology, 12(52), 7158-7167.

Goeselova, S., Holubová, B., Blazkova, M., & Fukal, L. (2013). Colloidal gold based lateral flow assay for the one-step rapid detection of thiabendazole. Chemicke Listy, 107(11), 875-879.

Gomez, L. E. (2009). Use and benefits of chlorpyrifos in US agriculture. Dow AgroSciences LLC,

Gong, J., Wang, X., Li, X., & Wang, K. (2012). Highly sensitive visible light activated photoelectrochemical biosensing of organophosphate pesticide using biofunctional crossed bismuth oxyiodide flake arrays. Biosensors and Bioelectronics, 38(1), 43- 49.

González-Rodríguez, R. M., Rial-Otero, R., Cancho-Grande, B., & Simal-Gándara, J. (2008). Determination of 23 pesticide residues in leafy vegetables using gas chromatography–ion trap mass spectrometry and analyte protectants. Journal of Chromatography A, 1196, 100-109.

Govindasamy, M., Chen, S., Mani, V., Akilarasan, M., Kogularasu, S., & Subramani, B. (2017). Nanocomposites composed of layered molybdenum disulfide and graphene for highly sensitive amperometric determination of methyl parathion. Microchimica Acta, 184(3), 725-733.

208

Chapter 6

Gui, W., Liu, Y., Wang, C., Liang, X., & Zhu, G. (2009). Development of a direct competitive enzyme-linked immunosorbent assay for parathion residue in food samples. Analytical Biochemistry, 393(1), 88-94.

Guimarães, A., De Assis, H. S., & Boeger, W. (2007). The effect of trichlorfon on acetylcholinesterase activity and histopathology of cultivated fish oreochromis niloticus. Ecotoxicology and Environmental Safety, 68(1), 57-62.

Guler, M., Turkoglu, V., & Basi, Z. (2017). Determination of malation, methidathion, and chlorpyrifos ethyl pesticides using acetylcholinesterase biosensor based on Nafion/Ag@ rGO-NH 2 nanocomposites. Electrochimica Acta, 240, 129-135.

Guler, M., Turkoglu, V., & Kivrak, A. (2016). Electrochemical detection of malathion pesticide using acetylcholinesterase biosensor based on glassy carbon electrode modified with conducting polymer film. Environmental Science and Pollution Research International, 23(12), 12343.

Guo, Y., Liu, S., Gui, W., & Zhu, G. (2009). Gold immunochromatographic assay for simultaneous detection of carbofuran and triazophos in water samples. Analytical Biochemistry, 389(1), 32-39.

Guo, Y., Tian, J., Liang, C., Zhu, G., & Gui, W. (2013). Multiplex bead-array competitive immunoassay for simultaneous detection of three pesticides in vegetables. Microchimica Acta, 180(5-6), 387-395.

Hamamoto, K., Iwatsuki, K., Akama, R., & Koike, R. (2017). Rapid multiresidue determination of pesticides in livestock muscle and liver tissue via modified QuEChERS sample preparation and LC-MS/MS. Food Additives & Contaminants: Part A, 1-10.

Han, Z., Chi, C., Bai, B., Liu, G., Rao, Q., Peng, S., et al. (2012). Chromogenic platform based on recombinant drosophila melanogaster acetylcholinesterase for visible unidirectional assay of organophosphate and carbamate insecticide residues. Analytica Chimica Acta, 720, 126-133.

Handford, C. E., Elliott, C. T., & Campbell, K. (2015). A review of the global pesticide legislation and the scale of challenge in reaching the global harmonization of food safety standards. Integrated Environmental Assessment and Management, 11(4), 525-536.

Hanson, C. (2013). Recent advances in liquid-liquid extraction Elsevier.

209

Chapter 6

Hantash, J., Bartlett, A., Oldfield, P., Dénès, G., O’Rielly, R., Roudiere, D., et al. (2006). Use of an on-line imprinted polymer pre-column, for the liquid chromatographic-UV absorbance determination of carbaryl and its metabolite in complex matrices. Journal of Chromatography A, 1125(1), 104-111.

Haupt, K. (2008). Synthesis of molecularly imprinted polymers by bulk polymerization. Retrieved 10/01, 2017, from http://karstenolaf.free.fr/Imprinting.html

Haupt, K., & Mosbach, K. (2000). Molecularly imprinted polymers and their use in biomimetic sensors. Chemical Reviews, 100(7), 2495-2504.

Hayden, K. M., Norton, M. C., Darcey, D., Ostbye, T., Zandi, P. P., Breitner, J. C., et al. (2010). Occupational exposure to pesticides increases the risk of incident AD: The cache county study. Neurology, 74(19), 1524-1530. doi:10.1212/WNL.0b013e3181dd4423 [doi]

Hazer, O., Akkbik, M., Demir, D., & Turhan, Y. (2017). Determination of carbendazim and chlorpyrifos in selected fruits and vegetables samples using QuEChERS-HPLC- FD. Eurasian Journal of Analytical Chemistry, 12(2)

He, B., Mao, Y., Zhang, Y., Yin, W., Hou, C., Huo, D., et al. (2017). A porphyrin molecularly imprinted biomimetic electrochemical sensor based on gold nanoparticles and carboxyl graphene composite for the highly efficient detection of methyl parathion. Nano, 12(04), 1750046.

He, C., Long, Y., Pan, J., Li, K., & Liu, F. (2007). Application of molecularly imprinted polymers to solid-phase extraction of analytes from real samples. Journal of Biochemical and Biophysical Methods, 70(2), 133-150.

Hemwimol, S., Pavasant, P., & Shotipruk, A. (2006). Ultrasound-assisted extraction of anthraquinones from roots of morinda citrifolia. Ultrasonics Sonochemistry, 13(6), 543-548.

Hemwimon, S., Pavasant, P., & Shotipruk, A. (2007). Microwave-assisted extraction of antioxidative anthraquinones from roots of morinda citrifolia. Separation and Purification Technology, 54(1), 44-50.

Hennion, M., & Barcelo, D. (1998). Strengths and limitations of immunoassays for effective and efficient use for pesticide analysis in water samples: A review. Analytica Chimica Acta, 362(1), 3-34.

210

Chapter 6

Hiemstra, M., & de Kok, A. (2002). Determination of propamocarb in vegetables using polymer-based high-performance liquid chromatography coupled with electrospray mass spectrometry. Journal of Chromatography A, 972(2), 231-239.

Hiemstra, M., & De Kok, A. (2007). Comprehensive multi-residue method for the target analysis of pesticides in crops using liquid chromatography–tandem mass spectrometry. Journal of Chromatography A, 1154(1), 3-25.

Hod, R. (2011). Chlorpyrifos blood level and exposure symptoms among paddy farmers in sabak bernam, malaysia. International Journal of Public Health Research, 1(1), 1- 6.

Hossain, S. Z., Luckham, R. E., McFadden, M. J., & Brennan, J. D. (2009). Reagentless bidirectional lateral flow bioactive paper sensors for detection of pesticides in beverage and food samples. Analytical Chemistry, 81(21), 9055-9064.

Hu, F., He, L., Yang, J., Bian, K., Wang, Z., Yang, H., et al. (2014). Determination of 26 veterinary antibiotics residues in water matrices by lyophilization in combination with LC–MS/MS. Journal of Chromatography B, 949, 79-86.

Hunde, T., Berhe, M., Tadese, A., Tirfu, M., Woldu, A., Menasbo, B., et al. (2017). Nano fe 3 O 4–graphite paste modified electrochemical sensor for phosphatic pesticide- chlorpyrifos. Momona Ethiopian Journal of Science, 9(1), 76-89.

IARC, International Agency for Research on Cancer. (2017a). IARC monographs on the evaluation of carcinogenic risks to humans. THE MONOGRAPHS. some organophosphate insecticides and herbicides. lyon – france. Volume 112 (2017). Retrieved 05/01, 2017, from http://monographs.iarc.fr/ENG/Monographs/vol112/ index.php

IARC, International Agency for Research on Cancer. (2017b). Agents classified by the IARC monographs, volumes 1–119. Retrieved 05/17, 2017, from http://monographs.iarc.fr/ENG/Classification/

Istamboulie, G., Cortina-Puig, M., Marty, J., & Noguer, T. (2009). The use of artificial neural networks for the selective detection of two organophosphate insecticides: Chlorpyrifos and chlorfenvinfos. Talanta, 79(2), 507-511.

Ivanov, A., Evtugyn, G., Lukachova, L. V., Karyakina, E. E., Budnikov, H., Kiseleva, S., et al. (2003). New polyaniline-based potentiometric biosensor for pesticides detection. IEEE Sensors Journal, 3(3), 333-340.

211

Chapter 6

Jamal, G. A., Hansen, S., Pilkington, A., Buchanan, D., Gillham, R. A., Abdel-Azis, M., et al. (2002). A clinical neurological, neurophysiological, and neuropsychological study of sheep farmers and dippers exposed to organophosphate pesticides. Occupational and Environmental Medicine, 59(7), 434-441.

Jenkins, A. L., Yin, R., & Jensen, J. L. (2001). Molecularly imprinted polymer sensors for pesticide and insecticide detection in water. Analyst, 126(6), 798-802.

Ji, J., Deng, C., Zhang, H., Wu, Y., & Zhang, X. (2007). Microwave-assisted steam distillation for the determination of organochlorine pesticides and pyrethroids in chinese teas. Talanta, 71(3), 1068-1074.

Jiang, X., Li, D., Xu, X., Ying, Y., Li, Y., Ye, Z., et al. (2008). Immunosensors for detection of pesticide residues. Biosensors and Bioelectronics, 23(11), 1577-1587.

Jiao, W., Xiao, Y., Qian, X., Tong, M., Hu, Y., Hou, R., et al. (2016). Optimized combination of dilution and refined QuEChERS to overcome matrix effects of six types of tea for determination eight neonicotinoid insecticides by ultra performance liquid chromatography–electrospray tandem mass spectrometry. Food Chemistry, 210, 26-34.

Jin, M., Fu, J., Xue, B., Zhou, S., Zhang, L., & Li, A. (2017). Distribution and enantiomeric profiles of organochlorine pesticides in surface sediments from the bering sea, chukchi sea and adjacent arctic areas. Environmental Pollution,

Jin, S., Xu, Z., Chen, J., Liang, X., Wu, Y., & Qian, X. (2004). Determination of organophosphate and carbamate pesticides based on enzyme inhibition using a pH- sensitive fluorescence probe. Analytica Chimica Acta, 523(1), 117-123.

Kaltsonoudis, C., Lamari, F., Prousalis, K., Karamanos, N., & Tsegenidis, T. (2003). Analysis of carbendazim and thiabendazole in lemons by CE-DAD. Chromatographia, 57(3), 181-184.

Kang, S., Xu, Y., Zhou, L., & Pan, C. (2012). Preparation of molecularly imprinted polymers: Diethyl (3‐methylureido)(phenyl) methylphosphonate as a dummy template for the recognition of its organophosphate pesticide analogs. Journal of Applied Polymer Science, 124(5), 3737-3743.

Kaoud Hussein, A. (2015). Article review: Heavy metals and pesticides in aquaculture: Health problems. European Journal of Academic Essays, 2(9), 15-22.

212

Chapter 6

Karim, K., Breton, F., Rouillon, R., Piletska, E. V., Guerreiro, A., Chianella, I., et al. (2005). How to find effective functional monomers for effective molecularly imprinted polymers? Advanced Drug Delivery Reviews, 57(12), 1795-1808.

Kaur, B., & Srivastava, R. (2015). A polyaniline–zeolite nanocomposite material based acetylcholinesterase biosensor for the sensitive detection of acetylcholine and organophosphates. New Journal of Chemistry, 39(9), 6899-6906.

Kaur, J., Singh, K. V., Boro, R., Thampi, K., Raje, M., Varshney, G. C., et al. (2007). Immunochromatographic dipstick assay format using gold nanoparticles labeled protein− hapten conjugate for the detection of atrazine. Environmental Science & Technology, 41(14), 5028-5036.

Kaur, N., Thakur, H., & Prabhakar, N. (2016a). Conducting polymer and multi-walled carbon nanotubes nanocomposites based amperometric biosensor for detection of organophosphate. Journal of Electroanalytical Chemistry, 775, 121-128.

Kaur, N., Thakur, H., Kumar, R., & Prabhakar, N. (2016b). An electrochemical sensor modified with poly (3, 4-ethylenedioxythiophene)-wrapped multi-walled carbon nanotubes for enzyme inhibition-based determination of organophosphates. Microchim Acta, 183(7), 2307-2315.

Kaur, N., Thakur, H., Pathak, S., & Prabhakar, N. (2015). Acetylcholinesterase immobilised eggshell membrane-based optical biosensor for organophosphate detection. International Journal of Environmental Analytical Chemistry, 95(12), 1134-1147.

Kim, C. S., Choi, B., Seo, J. H., Lim, G., & Cha, H. J. (2013). Mussel adhesive protein- based whole cell array biosensor for detection of organophosphorus compounds. Biosensors and Bioelectronics, 41, 199-204.

Kim, K., Kabir, E., & Jahan, S. A. (2017). Exposure to pesticides and the associated human health effects. Science of the Total Environment, 575, 525-535.

Kim, Y. A., Lee, E., Kim, K., Lee, Y. T., Hammock, B. D., & Lee, H. (2011). Competitive immunochromatographic assay for the detection of the organophosphorus pesticide chlorpyrifos. Analytica Chimica Acta, 693(1), 106-113.

Kimm, G. L., Hook, G. L., & Smith, P. A. (2002). Application of headspace solid-phase microextraction and gas chromatography–mass spectrometry for detection of the chemical warfare agent bis (2-chloroethyl) sulfide in soil. Journal of Chromatography A, 971(1), 185-191.

213

Chapter 6

Kitada, Y., Kawahata, H., Suzuki, A., & Oomori, T. (2008). Distribution of pesticides and bisphenol A in sediments collected from rivers adjacent to coral reefs. Chemosphere, 71(11), 2082-2090.

Kleinkopf, G. E., Oberg, N. A., & Olsen, N. L. (2003). Sprout inhibition in storage: Current status, new chemistries and natural compounds. American Journal of Potato Research, 80(5), 317.

Koukouvinos, G., Τsialla, Z., Petrou, P. S., Misiakos, K., Goustouridis, D., Moreno, A. U., et al. (2017). Fast simultaneous detection of three pesticides by a white light reflectance spectroscopy sensing platform. Sensors and Actuators B: Chemical, 238, 1214-1223.

Kranthi, K., Davis, M., Mayee, C., Russell, D., Shukla, R., Satija, U., et al. (2009). Development of a colloidal-gold based lateral-flow immunoassay kit for ‘quality- control’assessment of pyrethroid and endosulfan formulations in a novel single strip format. Crop Protection, 28(5), 428-434.

Krieger, R. (2010). Hayes' handbook of pesticide toxicology Academic press.

Kumar, P., Kim, K., Bansal, V., Paul, A. K., & Deep, A. (2016). Practical utilization of nanocrystal metal organic framework biosensor for parathion specific recognition. Microchemical Journal, 128, 102-107.

Kuster, M., de Alda, M. L., & Barceló, D. (2009). Liquid chromatography–tandem mass spectrometric analysis and regulatory issues of polar pesticides in natural and treated waters. Journal of Chromatography A, 1216(3), 520-529.

Kuswandi, B., Fikriyah, C. I., & Gani, A. A. (2008). An optical fiber biosensor for chlorpyrifos using a single sol–gel film containing acetylcholinesterase and bromothymol blue. Talanta, 74(4), 613-618.

Lai, J., Yang, M., Niessner, R., & Knopp, D. (2007). Molecularly imprinted microspheres and nanospheres for di (2-ethylhexyl) phthalate prepared by precipitation polymerization. Analytical and Bioanalytical Chemistry, 389(2), 405-412.

Lang, Q., Han, L., Hou, C., Wang, F., & Liu, A. (2016). A sensitive acetylcholinesterase biosensor based on gold nanorods modified electrode for detection of organophosphate pesticide. Talanta, 156, 34-41.

Lanza, F., & Sellergren, B. (2004). Molecularly imprinted polymers via High‐Throughput and combinatorial techniques. Macromolecular Rapid Communications, 25(1), 59- 68.

214

Chapter 6

Lasáková, M., & Jandera, P. (2009). Molecularly imprinted polymers and their application in solid phase extraction. Journal of Separation Science, 32(5‐6), 799- 812.

Latif, Y., Sherazi, S. T. H., & Bhanger, M. I. (2011). Monitoring of pesticide residues in commonly used fruits in hyderabad region, pakistan. American Journal of Analytical Chemistry, 2(08), 46.

Lazic, S., Sunjka, D., Grahovac, N., Vukovic, S., & Jaksic, S. (2011). Determination of chlorpyrifos in water used for agricultural production. Poljoprivreda i Sumarstvo, 57(4), 17.

LeDoux, M. (2011). Analytical methods applied to the determination of pesticide residues in foods of animal origin. A review of the past two decades. Journal of Chromatography A, 1218(8), 1021-1036.

Lei, Y., Mulchandani, P., Chen, W., Wang, J., & Mulchandani, A. (2004a). Arthrobacter sp. JS443‐Based whole cell amperometric biosensor for p‐Nitrophenol. Electroanalysis, 16(24), 2030-2034.

Lei, Y., Mulchandani, P., Chen, W., Wang, J., & Mulchandani, A. (2004b). Whole cell– enzyme hybrid amperometric biosensor for direct determination of organophosphorous nerve agents with p‐nitrophenyl substituent. Biotechnology and Bioengineering, 85(7), 706-713.

Li, C., Wang, C., Wang, C., & Hu, S. (2006). Development of a parathion sensor based on molecularly imprinted nano-TiO 2 self-assembled film electrode. Sensors and Actuators B: Chemical, 117(1), 166-171.

Li, D., Zhang, X., Kong, F., Qiao, X., & Xu, Z. (2017b). Molecularly imprinted solid- phase extraction coupled with high-performance liquid chromatography for the determination of trace trichlorfon and monocrotophos residues in fruits. Food Analytical Methods, 10(5), 1284-1292.

Li, H., Wang, Z., Wu, B., Liu, X., Xue, Z., & Lu, X. (2012b). Rapid and sensitive detection of methyl-parathion pesticide with an electropolymerized, molecularly imprinted polymer capacitive sensor. Electrochimica Acta, 62, 319-326.

Li, H., Xie, T., Shi, D., Jin, J., & Xie, C. (2016). Enhanced electrochemiluminescence of luminol at the gold nanoparticle/carbon nanotube/electropolymerised molecular imprinting composite membrane interface for selective recognition of triazophos. International Journal of Environmental Analytical Chemistry, 96(13), 1300-1311.

215

Chapter 6

Li, P., Rong, F., & Yuan, C. (2003). Morphologies and binding characteristics of molecularly imprinted polymers prepared by precipitation polymerization. Polymer International, 52(12), 1799-1806.

Li, S., Ge, Y., Piletsky, S. A., & Lunec, J. (2012a). Molecularly imprinted sensors: Overview and applications Elsevier.

Li, W., Liu, D., Li, J., Gao, J., Zhang, C., Wang, P., et al. (2017a). Matrix solid-phase dispersion combined with GC–MS/MS for the determination of organochlorine pesticides and polychlorinated biphenyls in marketed seafood. Chromatographia, 80(5), 813-824.

Li, Y., Xu, M., Li, P., Dong, J., & Ai, S. (2014b). Nonenzymatic sensing of methyl parathion based on graphene/gadolinium prussian blue analogue nanocomposite modified glassy carbon electrode. Analytical Methods, 6(7), 2157-2162.

Lian, Y., Pang, G., Shu, H., Fan, C., Liu, Y., Feng, J., et al. (2010). Simultaneous determination of 346 multiresidue pesticides in grapes by PSA-MSPD and GC-MS- SIM. Journal of Agricultural and Food Chemistry, 58(17), 9428-9453.

Liang, H., Song, D., & Gong, J. (2014). Signal-on electrochemiluminescence of biofunctional CdTe quantum dots for biosensing of organophosphate pesticides. Biosensors and Bioelectronics, 53, 363-369.

Ligler, F. S., Taitt, C. R., Shriver-Lake, L. C., Sapsford, K. E., Shubin, Y., & Golden, J. P. (2003). Array biosensor for detection of toxins. Analytical and Bioanalytical Chemistry, 377(3), 469-477. Liška, I. (2000). Fifty years of solid-phase extraction in water analysis–historical development and overview. Journal of Chromatography A, 885(1), 3-16.

Liu, B., Zhou, P., Liu, X., Sun, X., Li, H., & Lin, M. (2013c). Detection of pesticides in fruits by surface-enhanced raman spectroscopy coupled with gold nanostructures. Food and Bioprocess Technology, 1-9.

Liu, F., Du, Y., Cheng, Y., Yin, W., Hou, C., Huo, D., et al. (2016). A selective and sensitive sensor based on highly dispersed cobalt porphyrin-Co3O4-graphene oxide nanocomposites for the detection of methyl parathion. Journal of Solid State Electrochemistry, 3(20), 599-607.

Liu, J., Yang, M., Huai, L. F., & Chen, Y. T. (2010). Removal of chlorpyrifos from contaminated water using molecularly imprinted polymeric microspheres. Paper

216

Chapter 6

presented at the Bioinformatics and Biomedical Engineering (iCBBE), 2010 4th International Conference on, pp. 1-4.

Liu, L., Suryoprabowo, S., Zheng, Q., Song, S., & Kuang, H. (2017). Rapid detection of aldicarb in cucumber with an immunochromatographic test strip. Food and Agricultural Immunology, 28(3), 427-438.

Liu, Q., Cai, J., Huan, J., Dong, X., Wang, C., Qiu, B., et al. (2014). A visible light photoelectrochemical biosensor coupling enzyme-inhibition for organophosphates monitoring based on a dual-functional cd 0.5 zn 0.5 S-reduced graphene oxide nanocomposite. Analyst, 139(5), 1121-1126.

Liu, R., Yang, C., Xu, Y., Xu, P., Jiang, H., & Qiao, C. (2013b). Development of a whole- cell biocatalyst/biosensor by display of multiple heterologous proteins on the escherichia coli cell surface for the detoxification and detection of organophosphates. Journal of Agricultural and Food Chemistry, 61(32), 7810-7816.

Liu, S., Yuan, L., Yue, X., Zheng, Z., & Tang, Z. (2008). Recent advances in nanosensors for organophosphate pesticide detection. Advanced Powder Technology, 19(5), 419- 441.

Liu, S., Zheng, Z., & Li, X. (2013a). Advances in pesticide biosensors: Current status, challenges, and future perspectives. Analytical and Bioanalytical Chemistry, 405(1), 63-90.

Liu, Y. H., Chen, J., Guo, Y. R., Wang, C. M., Liang, X., & Zhu, G. N. (2011). A sensitive monoclonal antibody-based enzyme-linked immunosorbent assay for chlorpyrifos residue determination in chinese agricultural samples. Journal of Environmental Science and Health Part B, 46(4), 313-320.

Liu, Y. H., Wang, C. M., Guo, Y. R., Liang, X., Gui, W., & Zhu, G. N. (2010). Monoclonal antibody produced by heterologous indirect enzyme-linked immunosorbent assay and its application for parathion residue determination in agricultural and environmental samples. Journal of Environmental Science and Health Part B, 45(6), 547-556.

Liu, Y., Wang, C., Gui, W., Bi, J., Jin, M., & Zhu, G. (2009). Development of a sensitive competitive indirect ELISA for parathion residue in agricultural and environmental samples. Ecotoxicology and Environmental Safety, 72(6), 1673-1679.

Lochhead, M. J., Todorof, K., Delaney, M., Ives, J. T., Greef, C., Moll, K., et al. (2011). Rapid multiplexed immunoassay for simultaneous serodiagnosis of HIV-1 and

217

Chapter 6

coinfections. Journal of Clinical Microbiology, 49(10), 3584-3590. doi:10.1128/JCM.00970-11 [doi]. Lu, P., Li, Q., Liu, H., Feng, Z., Yan, X., Hong, Q., et al. (2013). Biodegradation of chlorpyrifos and 3, 5, 6-trichloro-2-pyridinol by cupriavidus sp. DT-1. Bioresource Technology, 127, 337-342.

Lu, Q., Chen, X., Nie, L., Luo, J., Jiang, H., Chen, L., et al. (2010). Tuning of the vinyl groups’ spacing at surface of modified silica in preparation of high density imprinted layer-coated silica nanoparticles: A dispersive solid-phase extraction materials for chlorpyrifos. Talanta, 81(3), 959-966.

Lucci, P., Pacetti, D., Núñez, O., & Frega, N. G. (2012). Current trends in sample treatment techniques for environmental and food analysis. Chromatography-the most versatile method of chemical analysis () InTech.

Luo, L., Lei, H., Yang, J., Liu, G., Sun, Y., Bai, W., et al. (2017). Development of an indirect ELISA for the determination of ethyl carbamate in chinese rice wine. Analytica Chimica Acta, 950, 162-169.

Ma, G., & Chen, L. (2014). Determination of chlorpyrifos in rice based on magnetic molecularly imprinted polymers coupled with high-performance liquid chromatography. Food Analytical Methods, 7(2), 377-388.

Ma, Y., Li, H., & Wang, L. (2012). Magnetic-luminescent bifunctional nanosensors. Journal of Materials Chemistry, 22(36), 18761-18767.

Mahony, J., Nolan, K., Smyth, M. and Mizaikoff, B. (2005), "Molecularly imprinted polymers—potential and challenges in analytical chemistry", Analytica Chimica Acta, vol. 534, no. 1, pp. 31-39.

Manclus, J. J., Primo, J., & Montoya, A. (1994). Development of a chlorpyrifos immunoassay using antibodies obtained from a simple hapten design. Journal of Agricultural and Food Chemistry, 42(6), 1257-1260. Mao, B., Jia, Z., Mao, X., Wang, K., & Ji, S. (2013). Determination of triazine residues in herbal plants using molecularly imprinted polymer extraction followed by UPLC- MS/MS. Analytical Methods, 5(14), 3582-3590.

Marín, J. M., Gracia-Lor, E., Sancho, J. V., López, F. J., & Hernández, F. (2009). Application of ultra-high-pressure liquid chromatography–tandem mass spectrometry to the determination of multi-class pesticides in environmental and wastewater samples: Study of matrix effects. Journal of Chromatography A, 1216(9), 1410-1420.

218

Chapter 6

Martín, L., Julio, L. F., Burillo, J., Sanz, J., Mainar, A. M., & González-Coloma, A. (2011). Comparative chemistry and insect antifeedant action of traditional (clevenger and soxhlet) and supercritical extracts (CO 2) of two cultivated wormwood (artemisia absinthium L.) populations. Industrial Crops and Products, 34(3), 1615- 1621.

Martini, E., Tomassetti, M., & Campanella, L. (2015). Determination of traces of several pesticides in sunflower oil using organic phase immuno electrodes (OPIEs). Talanta, 132, 503-512.

Marx, S., & Zaltsman, A. (2003). Molecular imprinting of sol gel polymers for the detection of paraoxon in water. International Journal of Environmental & Analytical Chemistry, 83(7-8), 671-680.

Masiá, A., Blasco, C., & Picó, Y. (2014). Last trends in pesticide residue determination by liquid chromatography–mass spectrometry. Trends in Environmental Analytical Chemistry, 2, 11-24.

Mauriz, E., Calle, A., Lechuga, L., Quintana, J., Montoya, A., & Manclus, J. (2006). Real- time detection of chlorpyrifos at part per trillion levels in ground, surface and drinking water samples by a portable surface plasmon resonance immunosensor. Analytica Chimica Acta, 561(1), 40-47.

Mayes, A., & Whitcombe, M. (2005). Synthetic strategies for the generation of molecularly imprinted organic polymers. Advanced Drug Delivery Reviews, 57(12), 1742-1778.

Mayorga-Martinez, C. C., Pino, F., Kurbanoglu, S., Rivas, L., Ozkan, S. A., & Merkoçi, A. (2014). Iridium oxide nanoparticle induced dual catalytic/inhibition based detection of phenol and pesticide compounds. Journal of Materials Chemistry B, 2(16), 2233-2239.

McGrath, T. F., McClintock, L., Dunn, J. S., Husar, G. M., Lochhead, M. J., Sarver, R. W., et al. (2015). Development of a rapid multiplexed assay for the direct screening of antimicrobial residues in raw milk. Analytical and Bioanalytical Chemistry, 407(15), 4459-4472.

McNamee, S. E., Elliott, C. T., Greer, B., Lochhead, M., & Campbell, K. (2014). Development of a planar waveguide microarray for the monitoring and early detection of five harmful algal toxins in water and cultures. Environmental Science & Technology, 48(22), 13340-13349.

219

Chapter 6

Mehta, J., Bhardwaj, N., Bhardwaj, S. K., Tuteja, S. K., Vinayak, P., Paul, A., et al. (2017). Graphene quantum dot modified screen printed immunosensor for the determination of parathion. Analytical Biochemistry, 523, 1-9.

Mehta, J., Vinayak, P., Tuteja, S. K., Chhabra, V. A., Bhardwaj, N., Paul, A., et al. (2016). Graphene modified screen printed immunosensor for highly sensitive detection of parathion. Biosensors and Bioelectronics, 83, 339-346.

Meier, F., Elbert, S. M., & Mizaikoff, B. (2012). A novel approach for the direct determination of residual template molecules in molecularly imprinted polymer matrices. Analytical Methods, 4(9), 2755-2758.

Melo, A., Aguiar, A., Mansilha, C., Pinho, O., & Ferreira, I. M. (2012). Optimisation of a solid-phase microextraction/HPLC/Diode array method for multiple pesticide screening in lettuce. Food Chemistry, 130(4), 1090-1097.

Mena, M., Martınez-Ruiz, P., Reviejo, A., & Pingarron, J. (2002). Molecularly imprinted polymers for on-line preconcentration by solid phase extraction of pirimicarb in water samples. Analytica Chimica Acta, 451(2), 297-304.

Meneely, J. P., Campbell, K., Greef, C., Lochhead, M. J., & Elliott, C. T. (2013). Development and validation of an ultrasensitive fluorescence planar waveguide biosensor for the detection of paralytic shellfish toxins in marine algae. Biosensors and Bioelectronics, 41, 691-697.

Mercader, J. V., Parra, J., Esteve-Turrillas, F. A., Agulló, C., Abad-Somovilla, A., & Abad-Fuentes, A. (2012). Development of monoclonal antibody-based competitive immunoassays for the detection of picoxystrobin in cereal and oilseed flours. Food Control, 26(1), 162-168.

Mirsky, V. M., & Yatsimirsky, A. (2010). Artificial receptors for chemical sensors John Wiley & Sons.

Mishra, A., Kumar, J., & Melo, J. S. (2017a). An optical microplate biosensor for the detection of methyl parathion pesticide using a biohybrid of sphingomonas sp. cells- silica nanoparticles. Biosensors and Bioelectronics, 87, 332-338.

Mishra, R. K., Mohan, A. V., Soto, F., Chrostowski, R., & Wang, J. (2017b). A microneedle biosensor for minimally-invasive transdermal detection of nerve agents. Analyst, 142(6), 918-924.

220

Chapter 6

Mogha, N. K., Sahu, V., Sharma, M., Sharma, R. K., & Masram, D. T. (2016). Biocompatible ZrO 2-reduced graphene oxide immobilized AChE biosensor for chlorpyrifos detection. Materials & Design, 111, 312-320.

Mol, H. G., Rooseboom, A., van Dam, R., Roding, M., Arondeus, K., & Sunarto, S. (2007). Modification and re-validation of the ethyl acetate-based multi-residue method for pesticides in produce. Analytical and Bioanalytical Chemistry, 389(6), 1715-1754.

Mol, H. G., Tienstra, M., & Zomer, P. (2016). Evaluation of gas chromatography– electron ionization–full scan high resolution orbitrap mass spectrometry for pesticide residue analysis. Analytica Chimica Acta, 935, 161-172.

Mol, H. G., van Dam, R. C., & Steijger, O. M. (2003). Determination of polar organophosphorus pesticides in vegetables and fruits using liquid chromatography with tandem mass spectrometry: Selection of extraction solvent. Journal of Chromatography A, 1015(1), 119-127.

Moreno, M. J., Abad, A., & Montoya, A. (2001). Production of monoclonal antibodies to the N-methylcarbamate pesticide propoxur. Journal of Agricultural and Food Chemistry, 49(1), 72-78.

Mostafalou, S., & Abdollahi, M. (2013). Pesticides and human chronic diseases: Evidences, mechanisms, and perspectives. Toxicology and Applied Pharmacology, 268(2), 157-177.

Mukundan, H., Anderson, A. S., Grace, W. K., Grace, K. M., Hartman, N., Martinez, J. S., et al. (2009). Waveguide-based biosensors for pathogen detection. Sensors, 9(7), 5783-5809.

Mulchandani, A. (2011). Microbial biosensors for organophosphate pesticides. Applied Biochemistry and Biotechnology, 165(2), 687-699.

Mulyasuryani, A., & Prasetyawan, S. (2015). Organophosphate hydrolase in conductometric biosensor for the detection of organophosphate pesticides. Analytical Chemistry Insights, 10, 23-27. doi:10.4137/ACI.S30656 [doi].

Nantia, E. A., Moreno-González, D., Manfo, F. P., Gámiz-Gracia, L., & García-Campaña, A. M. (2017). QuEChERS-based method for the determination of carbamate residues in aromatic herbs by UHPLC-MS/MS. Food Chemistry, 216, 334-341.

221

Chapter 6

Newman, J. D., Roberts, J. M., & Blanchard, G. (2007). Optical organophosphate/phosphonate sensor based upon gold nanoparticle functionalized quartz. Analytica Chimica Acta, 602(1), 101-107.

NLM, National library of medicine. National center for biotechnology information. (2017a). Pubchem compound database. compound summery for CID7230. Retrieved 03/11, 2017, from https://pubchem.ncbi.nlm.nih.gov/compound/chlorpyrifos#section=Top

NLM, National library of medicine. National center for biotechnology information. (2017b). Pubchem compound database.compund summery for CID 991. Retrieved 06/13, 2017, from https://pubchem.ncbi.nlm.nih.gov/compound/parathion#section =Top

NOAA fisheries, National Oceanic and Atmospheric Administration. What is aquaculture? Retrieved June/2, 2017, from http://www.nmfs.noaa.gov/aquaculture/what_is_aquaculture.html

NPIC, National pesticide information center. (2017). Chlorpyrifos, technical fact sheet. Retrieved 05/17, 2017, from http://npic.orst.edu/factsheets/archive/chlorptech.html Núñez, O., Gallart-Ayala, H., Martins, C. P., & Lucci, P. (2012). New trends in fast liquid chromatography for food and environmental analysis. Journal of Chromatography A, 1228, 298-323.

Nurmi, J., & Pellinen, J. (2011). Multiresidue method for the analysis of emerging contaminants in wastewater by ultra performance liquid chromatography–time-of- flight mass spectrometry. Journal of Chromatography A, 1218(38), 6712-6719.

OECD, Organisation for Economic Co-operation and Development. (2017). Pesticide programme;agricultural pesticides and biocides. Retrieved June/03, 2017, from http://www.oecd.org/env/ehs/pesticides-biocides/

Office of pesticide programs, reregistration eligibility decision for chlorpyrifos. Retrieved 10/0, 2017, from https://archive.epa.gov/pesticides/reregistration/web/html/status .html

Otero, R. R., Grande, B. C., & Gándara, J. S. (2003). Multiresidue method for fourteen fungicides in white grapes by liquid–liquid and solid-phase extraction followed by liquid chromatography–diode array detection. Journal of Chromatography A, 992(1), 121-131.

222

Chapter 6

Otieno, P., Owuor, P., Lalah, J., Pfister, G., & Schramm, K. (2013). Comparative evaluation of ELISA kit and HPLC DAD for the determination of chlorpyrifos ethyl residues in water and sediments. Talanta, 117, 250-257.

Özkütük, E. B., Diltemiz, S. E., Özalp, E., Gedikbey, T., & Ersöz, A. (2013). Paraoxon imprinted biopolymer based QCM sensor. Materials Chemistry and Physics, 139(1), 107-112.

Parra, J., Mercader, J. V., Agulló, C., Abad-Somovilla, A., & Abad-Fuentes, A. (2012). Generation of anti-azoxystrobin monoclonal antibodies from regioisomeric haptens functionalized at selected sites and development of indirect competitive immunoassays. Analytica Chimica Acta, 715, 105-112.

Pastor-Belda, M., Garrido, I., Campillo, N., Viñas, P., Hellín, P., Flores, P., et al. (2017). Combination of solvent extractants for dispersive liquid-liquid microextraction of fungicides from water and fruit samples by liquid chromatography with tandem mass spectrometry. Food Chemistry, 233, 69-76.

Paul, V., Ezekiel, R., & Pandey, R. (2016). Sprout suppression on potato: Need to look beyond CIPC for more effective and safer alternatives. Journal of Food Science and Technology, 53(1), 1-18.

Pedrosa, V. A., Caetano, J., Machado, S. A., Freire, R. S., & Bertotti, M. (2007). Acetylcholinesterase immobilization on 3‐Mercaptopropionic acid self assembled monolayer for determination of pesticides. Electroanalysis, 19(13), 1415-1420.

Pereira, L. A., & Rath, S. (2009). Molecularly imprinted solid-phase extraction for the determination of fenitrothion in tomatoes. Analytical and Bioanalytical Chemistry, 393(3), 1063-1072.

Pichon, V., & Chapuis-Hugon, F. (2008). Role of molecularly imprinted polymers for selective determination of environmental pollutants—a review. Analytica Chimica Acta, 622(1), 48-61.

Piletska, E. V., Guerreiro, A. R., Whitcombe, M. J., & Piletsky, S. A. (2009). Influence of the polymerization conditions on the performance of molecularly imprinted polymers. Macromolecules, 42(14), 4921-4928.

Piletska, E. V., Guerreiro, A. R., Romero-Guerra, M., Chianella, I., Turner, A. P., & Piletsky, S. A. (2008). Design of molecular imprinted polymers compatible with aqueous environment. Analytica chimica acta, 607(1), 54-60.

223

Chapter 6

Pohanka, M. (2012). Acetylcholinesterase based dipsticks with indoxylacetate as a substrate for assay of organophosphates and carbamates. Analytical Letters, 45(4), 367-374.

Pohanka, M., Karasova, J. Z., Kuca, K., & Pikula, J. (2010a). Multichannel spectrophotometry for analysis of organophosphate paraoxon in beverages. Turkish Journal of Chemistry, 34(1), 91-98.

Pohanka, M., Karasova, J. Z., Kuca, K., Pikula, J., Holas, O., Korabecny, J., et al. (2010b). Colorimetric dipstick for assay of organophosphate pesticides and nerve agents represented by paraoxon, sarin and VX. Talanta, 81(1), 621-624.

Poole, C. F. (2003). New trends in solid-phase extraction. TrAC Trends in Analytical Chemistry, 22(6), 362-373.

Prabhakar, N., Thakur, H., Bharti, A., & Kaur, N. (2016). Chitosan-iron oxide nanocomposite based electrochemical aptasensor for determination of malathion. Analytica Chimica Acta, 939, 108-116.

Pugazhenthi, T.R., and Ayyadurai, K. (2013). Immunoassay analysis – a tool for detection of pesticide residues in dairy products. Indian Journal of Fundamental and Applied Life Sciences, 3(1), 48.

Qian, K., Fang, G., He, J., Pan, M., & Wang, S. (2010). Preparation and application of a molecularly imprinted polymer for the determination of trace metolcarb in food matrices by high performance liquid chromatography. Journal of Separation Science, 33(14), 2079-2085.

Qian, S., & Bau, H. H. (2003). Analysis of lateral flow biodetectors: Competitive. Analysis, 9, 26-2003.

Qu, L., Zhang, H., Zhu, J., Yang, G., & Aboul-Enein, H. Y. (2010). Rapid determination of organophosphorous pesticides in leeks by gas chromatography–triple quadrupole mass spectrometry. Food Chemistry, 122(1), 327-332.

Rahiminejad, M., Shahtaheri, S., Ganjali, M., Forushani, A., & Golbabaei, F. (2009). Molecularly imprinted solid phase extraction for trace analysis of diazinon in drinking water. Journal of Environmental Health Science & Engineering, 6(2), 97- 106.

Rahiminezhad, M., Shahtaheri, S., Ganjali, M., Koohpaei, A., Rahimi Forushani, A., & Golbabaei, F. (2010). An experimental investigation of the molecularly imprinted

224

Chapter 6

polymers as tailor-made sorbents of diazinon. Journal of Analytical Chemistry, 65(7), 694-698.

Ramanathan, M., & Simonian, A. L. (2007). Array biosensor based on enzyme kinetics monitoring by fluorescence spectroscopy: Application for neurotoxins detection. Biosensors and Bioelectronics, 22(12), 3001-3007.

Ramos, J., Rial-Otero, R., Ramos, L., & Capelo, J. (2008). Ultrasonic-assisted matrix solid-phase dispersion as an improved methodology for the determination of pesticides in fruits. Journal of Chromatography A, 1212(1), 145-149.

Rao, M., Li, Y., Dong, J., Wu, W., Xu, Z., Sun, Y., et al. (2016). Production and characterization of a single-chain fab fragment for the detection of O, O-diethyl organophosphorus pesticides. Analytical Methods, 8(15), 3140-3147.

RASFF, the Food and Feed Safety Alerts system. (2017c). European commission RASFF portal Retrieved 06/01, 2017, from https://webgate.ec.europa.eu/rasffwindow/portal /?event=SearchResultPrint&StartRow=0

RASFF, The rapid alert system for feed and food. (2017a). The rapid alert system for feed and food 2015 annual report, RASFF annual report, european commission, health and food safety, 2016. Retrieved 04/15, 2017, from https://ec.europa.eu/food/safety/rasff/reports_publications_en

RASFF, The Rapid Alert System for Food and Feed. (2017b). The rapid alert system for food and Feed 2014 annual report. Retrieved 06/19, 2017, from https://webgate.ec.europa.eu/rasff-window/portal/?event=SearchByKeyword &NewSearch=1&Keywords=chlorpyrifos

Rathod, A. L., & Garg, R. (2017). Chlorpyrifos poisoning and its implications in human fatal cases: A forensic perspective with reference to indian scenario. Journal of Forensic and Legal Medicine, 47, 29-34.

Raz, S. R., & Haasnoot, W. (2011). Multiplex bioanalytical methods for food and environmental monitoring. TrAC Trends in Analytical Chemistry, 30(9), 1526-1537.

Rejczak, T., & Tuzimski, T. (2017). QuEChERS-based extraction with dispersive solid phase extraction clean-up using PSA and ZrO 2-based sorbents for determination of pesticides in bovine milk samples by HPLC-DAD. Food Chemistry, 217, 225-233.

Rezaei, F., & Hosseini, M. M. (2011). New method based on combining ultrasonic assisted miniaturized matrix solid-phase dispersion and homogeneous liquid–liquid extraction for the determination of some organochlorinated pesticides in fish. Analytica Chimica Acta, 702(2), 274-279. 225

Chapter 6

Rial-Otero, R., Gaspar, E., Moura, I., & Capelo, J. (2007). Gas chromatography mass spectrometry determination of acaricides from honey after a new fast ultrasonic- based solid phase micro-extraction sample treatment. Talanta, 71(5), 1906-1914.

Rodrigues, F. d. M., Mesquita, P. R., de Oliveira, L. S., De Oliveira, F. S., Menezes Filho, A., Pereira, Pedro A de P, et al. (2011). Development of a headspace solid-phase microextraction/gas chromatography–mass spectrometry method for determination of organophosphorus pesticide residues in cow milk. Microchemical Journal, 98(1), 56-61.

Rodríguez, D. C., Carvajal, S., & Peñuela, G. (2013). Effect of chlorpyrifos on the inhibition of the enzyme acetylcholinesterase by cross-linking in water-supply samples and milk from dairy cattle. Talanta, 111, 1-7. Rosengren-Holmberg, J. P., Karlsson, J. G., Svenson, J., Andersson, H. S., & Nicholls, I. A. (2009). Synthesis and ligand recognition of paracetamol selective polymers: Semi-covalent versus non-covalent molecular imprinting. Organic & Biomolecular Chemistry, 7(15), 3148-3155.

Royo, S., Martínez-Máñez, R., Sancenón, F., Costero, A. M., Parra, M., & Gil, S. (2007). Chromogenic and fluorogenic reagents for chemical warfare nerve agents' detection. Chemical communications, (46), 4839-4847.

Rubtsova, M. Y., Samsonova, J., Egorov, A., & Schmid, R. (1998). Simultaneous determination of several pesticides with chemiluminescent immunoassay on a multi‐ spot membrane strip. Food and Agricultural Immunology, 10(3), 223-235.

Ruiz, L. M. (2017). Exploring the harmful health effects of chlorpyrifos on children: An argument for policy reform.

Saadi, H. S., & Abdollahi, M. (2012). Is there a Link between Human Infertilities and Exposure to Pesticides?,

Sabik, H., Jeannot, R., & Rondeau, B. (2000). Multiresidue methods using solid-phase extraction techniques for monitoring priority pesticides, including triazines and degradation products, in ground and surface waters. Journal of Chromatography A, 885(1), 217-236.

Sadik, O., Land, W. H., Wanekaya, A. K., Uematsu, M., Embrechts, M. J., Wong, L., et al. (2004). Detection and classification of organophosphate nerve agent simulants using support vector machines with multiarray sensors. Journal of Chemical Information and Computer Sciences, 44(2), 499-507.

226

Chapter 6

Sai, N., Sun, W., Wu, Y., Sun, Z., & Huang, G. (2017). A sensitive immunoassay for parathion based on covalent linkage between small molecules hapten microtiter plates surface. Journal of the Iranian Chemical Society, 14(1), 257-268.

Saitta, M., Di Bella, G., Fede, M. R., Lo Turco, V., Potortì, A. G., Rando, R., et al. (2017). Gas chromatography-tandem mass spectrometry multi-residual analysis of contaminants in italian honey samples. Food Additives & Contaminants: Part A, 1- 9.

Sajid, M., Kawde, A., & Daud, M. (2015). Designs, formats and applications of lateral flow assay: A literature review. Journal of Saudi Chemical Society, 19(6), 689-705.

Salas, B. V., Duran, E. I. G., & Wiener, M. S. (2000). Impact of pesticides use on human health in mexico: A review. Reviews on Environmental Health, 15(4), 399-412.

Salghi, R., Errami, M., Hammouti, B., & Bazzi, L. (2011). Pesticides in the modern world-trends in pesticides analysis. In Tech, Rijeka.p, 71

Sambe, H., Hoshina, K., Moaddel, R., Wainer, I. W., & Haginaka, J. (2006). Uniformly- sized, molecularly imprinted polymers for nicotine by precipitation polymerization. Journal of Chromatography A, 1134(1), 88-94.

Samuels, T. A., & Obare, S. O. (2011). Advances in analytical methods for organophosphorus pesticide detection. In Pesticides in the Modern World-Trends in Pesticides Analysis. InTech.

Sanagi, M. M., Salleh, S., Ibrahim, W. A. W., & Naim, A. A. (2011). Determination of organophosphorus pesticides using molecularly imprinted polymer solid phase extraction. Malaysian Journal of Analytical Sciences, 15(2), 175-183.

Sanagi, M. M., Salleh, S., Ibrahim, W. A. W., Naim, A. A., Hermawan, D., Miskam, M., et al. (2013). Molecularly imprinted polymer solid-phase extraction for the analysis of organophosphorus pesticides in fruit samples. Journal of Food Composition and Analysis, 32(2), 155-161.

Sanchez-Ortega, A., Unceta, N., Gómez-Caballero, A., Sampedro, M., Akesolo, U., Goicolea, M., et al. (2009). Sensitive determination of triazines in underground waters using stir bar sorptive extraction directly coupled to automated thermal desorption and gas chromatography–mass spectrometry. Analytica Chimica Acta, 641(1), 110-116.

Sandstrom, M. W., Kanagy, L. K., Anderson, C. A., & Kanagy, C. J. (2016). Determination of pesticides and pesticide degradates in filtered water by direct

227

Chapter 6

Aqueous-Injection Liquid Chromatography-Tandem Mass Spectrometry (No. 5- B11). US Geological Survey.

Santos, M. G., Vitor, R. V., Andrade, F. L., Martins, I., & Figueiredo, E. C. (2012). Molecularly imprinted solid phase extraction of urinary diethyl thiophosphate and diethyl dithiophosphate and their analysis by gas chromatography–mass spectrometry. Journal of Chromatography B, 909, 70-76.

Say, R., Erdem, M., Ersöz, A., Türk, H., & Denizli, A. (2005). Biomimetic catalysis of an organophosphate by molecularly surface imprinted polymers. Applied Catalysis A: General, 286(2), 221-225.

Say, R., Ersöz, A., & Denizli, A. (2003). Selective separation of uranium containing glutamic acid molecular-imprinted polymeric microbeads. Separation Science and Technology, 38(14), 3431-3447.

Schirmer, C., & Meisel, H. (2006). Synthesis of a molecularly imprinted polymer for the selective solid-phase extraction of chloramphenicol from honey. Journal of Chromatography A, 1132(1), 325-328.

Sellergren, B. (1999). Polymer-and template-related factors influencing the efficiency in molecularly imprinted solid-phase extractions. TrAC Trends in Analytical Chemistry, 18(3), 164-174.

Shafer, T. J., Meyer, D. A., & Crofton, K. M. (2005). Developmental neurotoxicity of pyrethroid insecticides: Critical review and future research needs. Environmental Health Perspectives, 123-136.

Sharma, D., Nagpal, A., Pakade, Y. B., & Katnoria, J. K. (2010). Analytical methods for estimation of organophosphorus pesticide residues in fruits and vegetables: A review. Talanta, 82(4), 1077-1089.

Shi, X., Liu, J., Sun, A., Li, D., & Chen, J. (2012). Group-selective enrichment and determination of pyrethroid insecticides in aquaculture seawater via molecularly imprinted solid phase extraction coupled with gas chromatography-electron capture detection. Journal of Chromatography A, 1227, 60-66.

Shu, Q., Wang, L., Ouyang, H., Wang, W., Liu, F., & Fu, Z. (2017). Multiplexed immunochromatographic test strip for time-resolved chemiluminescent detection of pesticide residues using a bifunctional antibody. Biosensors and Bioelectronics, 87, 908-914.

228

Chapter 6

Siemann, M., Andersson, L. I., & Mosbach, K. (1996). Selective recognition of the herbicide atrazine by noncovalent molecularly imprinted polymers. Journal of Agricultural and Food Chemistry, 44(1), 141-145.

Sigolaeva, L. V., Gladyr, S. Y., Mergel, O., Gelissen, A. P., Noyong, M., Simon, U., Pergushov, D.V., Kurochkin, I.N., Plamper, F.A. and Richtering, W. (2017). Easy- preparable butyrylcholinesterase/microgel construct for facilitated organophosphate biosensing. Analytical Chemistry, 89(11), 6091-6098.

Simonian, A., Rainina, E., & Wild, J. (1997). A new approach for discriminative detection of organophosphate neurotoxins in the presence of other cholinesterase inhibitors. Analytical Letters, 30(14), 2453-2468.

Simpson, N. J. (2000). Solid-phase extraction: Principles, techniques, and applications CRC press.

Sivaperumal, P., Anand, P., & Riddhi, L. (2015). Rapid determination of pesticide residues in fruits and vegetables, using ultra-high-performance liquid chromatography/time-of-flight mass spectrometry. Food Chemistry, 168, 356-365.

Smith, M. J., & Bucher, G. (2012). Tools to study the degradation and loss of the N- phenyl carbamate chlorpropham—A comprehensive review. Environment International, 49, 38-50.

Sobhanzadeh, E., Bakar, N. K. A., Abas, M. R. B., & Nemati, K. (2011). Low temperature followed by matrix solid-phase dispersion-sonication procedure for the determination of multiclass pesticides in palm oil using LC-TOF-MS. Journal of Hazardous Materials, 186(2), 1308-1313.

Soler, C., Girotti, S., Ghini, S., Fini, F., Montoya, A., Manclús, J. J., et al. (2008). Analysis of chlorpyrifos in water, fruit juice, and honeybee extract by chemiluminescent ELISA. Analytical Letters, 41(14), 2539-2553.

Solna, R., Dock, E., Christenson, A., Winther-Nielsen, M., Carlsson, C., Emnéus, J., et al. (2005a). Amperometric screen-printed biosensor arrays with co-immobilised oxidoreductases and cholinesterases. Analytica Chimica Acta, 528(1), 9-19.

Solna, R., Sapelnikova, S., Skládal, P., Winther-Nielsen, M., Carlsson, C., Emnéus, J., et al. (2005b). Multienzyme electrochemical array sensor for determination of phenols and pesticides. Talanta, 65(2), 349-357.

229

Chapter 6

Solomon, K. R., Williams, W. M., Mackay, D., Purdy, J., Giddings, J. M., & Giesy, J. P. (2014). Properties and uses of chlorpyrifos in the united states. Ecological risk assessment for chlorpyrifos in terrestrial and aquatic systems in the united states (pp. 13-34) Springer.

Spanjer, M. C., Rensen, P. M., & Scholten, J. M. (2008). LC–MS/MS multi-method for mycotoxins after single extraction, with validation data for peanut, pistachio, wheat, maize, cornflakes, raisins and figs. Food Additives and Contaminants, 25(4), 472- 489.

Spectrum Laboratories. (2014). Chemical fact sheet on chlorpyrifos- a newsletter. Retrieved 04/10, 2017, from http://www.speclab.com/compound/c2921882.htm

Spectrum Laboratories. (2016). Chemical fact sheet, parathion. Retrieved 08/06, 2017, from http://www.speclab.com/compound/c56382.htm

Spivak, D. A. (2005). Optimization, evaluation, and characterization of molecularly imprinted polymers. Advanced Drug Delivery Reviews, 57(12), 1779-1794.

Sreedhar, N., Kumar, M. S., & Krishnaveni, K. (2015). Sensitive determination of chlorpyrifos using Ag/Cu alloy nanoparticles and graphene composite paste electrode. Sensors and Actuators B: Chemical, 210, 475-482.

Sun, X., Cao, Y., Gong, Z., Wang, X., Zhang, Y., & Gao, J. (2012). An amperometric immunosensor based on multi-walled carbon nanotubes-thionine-chitosan nanocomposite film for chlorpyrifos detection. Sensors, 12(12), 17247-17261.

Sun, Z., Wang, W., Wen, H., Gan, C., Lei, H., & Liu, Y. (2015). Sensitive electrochemical immunoassay for chlorpyrifos by using flake-like fe 3 O 4 modified carbon nanotubes as the enhanced multienzyme label. Analytica Chimica Acta, 899, 91-99.

Sundarmurugasan, R., Gumpu, M. B., Ramachandra, B. L., Nesakumar, N., Sethuraman, S., Krishnan, U. M., et al. (2016). Simultaneous detection of monocrotophos and dichlorvos in orange samples using acetylcholinesterase–zinc oxide modified platinum electrode with linear regression calibration. Sensors and Actuators B: Chemical, 230, 306-313.

Suri, C. R., Boro, R., Nangia, Y., Gandhi, S., Sharma, P., Wangoo, N., et al. (2009). Immunoanalytical techniques for analyzing pesticides in the environment. TrAC Trends in Analytical Chemistry, 28(1), 29-39.

T3DB. (2015). The toxin and toxin target database. Retrieved 11/01, 2016, from http://www.t3db.ca

230

Chapter 6

Tacon, A. G., & Metian, M. (2008). Aquaculture feed and food safety. Annals of the New York Academy of Sciences, 1140(1), 50-59.

Tahboub, Y. R., Zaater, M. F., & Barri, T. A. (2006). Simultaneous identification and quantitation of selected organochlorine pesticide residues in honey by full-scan gas chromatography–mass spectrometry. Analytica Chimica Acta, 558(1), 62-68.

Tan, X., Li, B., Liew, K., & Li, C. (2010). Electrochemical fabrication of molecularly imprinted porous silicate film electrode for fast and selective response of methyl parathion. Biosensors and Bioelectronics, 26(2), 868-871.

Tan, Y., Ahmad, I., & Wei, T. (2015). Detection of parathion methyl using a surface plasmon resonance sensor combined with molecularly imprinted films. Chinese Chemical Letters, 26(6), 797-800.

Tegtmeier, E. M., & Duffy, M. D. (2004). External costs of agricultural production in the united states. International Journal of Agricultural Sustainability, 2(1), 1-20.

Teju, E., Tadesse, B., & Megersa, N. (2017). Salting-out assisted liquid-liquid extraction for the preconcentration and quantitative determination of 8 herbicide residues simultaneously in different water samples with high performance liquid chromatography. Separation Science and Technology, (just-accepted)

The state of world fisheries and aquaculture opportunities and challenges., rome. Retrieved June/04, 2017, from http://www.fao.org/3/a-i3720e.pdf

Tian, F., Liu, X., Xu, J., Dong, F., Zheng, Y., Hu, M., et al. (2016). Simultaneous determination of phoxim, chlorpyrifos, and pyridaben residues in edible mushrooms by high-performance liquid chromatography coupled to tandem mass spectrometry. Food Analytical Methods, 9(10), 2917-2924.

Tiemann, U. (2008). In vivo and in vitro effects of the organochlorine pesticides DDT, TCPM, methoxychlor, and lindane on the female reproductive tract of mammals: A review. Reproductive Toxicology, 25(3), 316-326.

Tomassetti, M., Martini, E., Campanella, L., Favero, G., Sanzó, G., & Mazzei, F. (2015). A new surface plasmon resonance immunosensor for triazine pesticide determination in bovine milk: A comparison with conventional amperometric and screen-printed immunodevices. Sensors, 15(5), 10255-10270.

Toscano, I. A. S., Nunes, G. S., & Barceló, D. (1998). Immunoassays for pesticide analysis in environmental and food matrices. Food Technology and Biotechnology, 245-255.

231

Chapter 6

Tran, H., Reisberg, S., Piro, B., Nguyen, T., & Pham, M. (2013). Label‐Free electrochemical immunoaffinity sensor based on impedimetric method for pesticide detection. Electroanalysis, 25(3), 664-670.

Τsialla, Z., Ucles-Moreno, A., Petrou, P., Fernandez-Alba, A. R., & Κakabakos, S. E. (2015). Development of an indirect enzyme immunoassay for the determination of thiabendazole in white and red wines. International Journal of Environmental Analytical Chemistry, 95(13), 1299-1309.

Turiel, E., Tadeo, J., Cormack, P., & Martin-Esteban, A. (2005). HPLC imprinted- stationary phase prepared by precipitation polymerisation for the determination of thiabendazole in fruit. Analyst, 130(12), 1601-1607.

Tuzimski, T., & Sherma, J. (2015). High performance liquid chromatography in pesticide residue analysis Crc Press.

UNECE, United Nations Economic Commission for Europe. (2017). Globally harmonized system of classification and labelling of chemicals (GHS). Retrieved 04/19, 2017, from https://www.unece.org/trans/danger/publi/ghs/ghs_welcome_e.html

USEPA United States Environmental Protection Agency, & US EPA. NAFTA TWG five-year strategy 2016 - 2021. Retrieved June 05, 2017, from https://www.epa.gov/pesticides/nafta-twg-five-year-strategy-2016-2021.

USEPA, United States Environmental Protection Agency. (2000). RED facts: Ethyl parathion. prevention, pesticides and toxic substances. EPA-738-F00-009. september 2000. washington (DC): United states environmental protection agency. Retrieved 07/05, 2017, from https://archive.epa.gov/pesticides/reregistration/web/ pdf/0155fct.pdf

USEPA, United States Environmental Protection Agency. (2017a). Parathion. Hazard summary. Retrieved 05/04, 2017, from https://www.epa.gov/sites/production/files /2016-09/documents/parathion.pdf

USEPA, US Environmental Protection Agency. (1996). Reregistration eligibility decision for chlorpropham.prevention, pesticides EPA 738-R-96-023 environmental protection and toxic substances october 1996. Retrieved 03/21, 2017, from https://archive.epa.gov/pesticides/reregistration/web/html/index-65.html

USEPA. (2002). Interim reregistration eligibility decision for chlorpyrifos.united states environmental protection agency.

232

Chapter 6

Valdés-Ramírez, G., Gutiérrez, M., Del Valle, M., Ramírez-Silva, M., Fournier, D., & Marty, J. (2009). Automated resolution of dichlorvos and methylparaoxon pesticide mixtures employing a flow injection system with an inhibition electronic tongue. Biosensors and Bioelectronics, 24(5), 1103-1108.

Valera, E., García-Febrero, R., Pividori, I., Sánchez-Baeza, F., & Marco, M. (2014). Coulombimetric immunosensor for paraquat based on electrochemical nanoprobes. Sensors and Actuators B: Chemical, 194, 353-360.

Vamvakaki, V., & Chaniotakis, N. A. (2007). Pesticide detection with a liposome-based nano-biosensor. Biosensors and Bioelectronics, 22(12), 2848-2853.

Van Dyk, J. S., & Pletschke, B. (2011). Review on the use of enzymes for the detection of organochlorine, organophosphate and carbamate pesticides in the environment. Chemosphere, 82(3), 291-307.

Van Maele-Fabry, G., Hoet, P., Vilain, F., & Lison, D. (2012). Occupational exposure to pesticides and parkinson's disease: A systematic review and meta-analysis of cohort studies. Environment International, 46, 30-43.

Vas, G., & Vekey, K. (2004). Solid‐phase microextraction: A powerful sample preparation tool prior to mass spectrometric analysis. Journal of Mass Spectrometry, 39(3), 233-254.

Vasapollo, G., Sole, R. D., Mergola, L., Lazzoi, M. R., Scardino, A., Scorrano, S., et al. (2011). Molecularly imprinted polymers: Present and future prospective. International Journal of Molecular Sciences, 12(9), 5908-5945.

Vázquez, P. P., Hakme, E., Uclés, S., Cutillas, V., Galera, M. M., Mughari, A., et al. (2016). Large multiresidue analysis of pesticides in edible vegetable oils by using efficient solid-phase extraction sorbents based on quick, easy, cheap, effective, rugged and safe methodology followed by gas chromatography–tandem mass spectrometry. Journal of Chromatography A, 1463, 20-31.

Verma, N., & Bhardwaj, A. (2015). Biosensor technology for pesticides—a review. Applied Biochemistry and Biotechnology, 175(6), 3093-3119.

Villaverde, J. J., Sevilla-Morán, B., López-Goti, C., Alonso-Prados, J. L., & Sandín- España, P. (2016). Trends in analysis of pesticide residues to fulfil the european regulation (EC) no. 1107/2009. TrAC Trends in Analytical Chemistry, 80, 568-580.

233

Chapter 6

Viswanathan, S., Radecka, H., & Radecki, J. (2009). Electrochemical biosensor for pesticides based on acetylcholinesterase immobilized on polyaniline deposited on vertically assembled carbon nanotubes wrapped with ssDNA. Biosensors and Bioelectronics, 24(9), 2772-2777.

Viveros, L., Paliwal, S., McCrae, D., Wild, J., & Simonian, A. (2006). A fluorescence- based biosensor for the detection of organophosphate pesticides and chemical warfare agents. Sensors and Actuators B: Chemical, 115(1), 150-157.

VoPham, T., Bertrand, K. A., Hart, J. E., Laden, F., Brooks, M. M., Yuan, J., et al. (2017). Pesticide exposure and liver cancer: A review. Cancer Causes & Control, 1-14.

Walker, J. P., Kimble, K. W., & Asher, S. A. (2007). Photonic crystal sensor for organophosphate nerve agents utilizing the organophosphorus hydrolase enzyme. Analytical and Bioanalytical Chemistry, 389(7-8), 2115-2124.

Wan, J., Yin, G., Xing, L., & Yang, X. (2014). Simple method for preparing methyl parathion sensor based on nanoporous gold/MWCNTs electrodes. International Journal of Environmental Analytical Chemistry, 94(2), 183-193.

Wang, B., Wang, H., Zhong, X., Chai, Y., Chen, S., & Yuan, R. (2016c). A highly sensitive electrochemiluminescence biosensor for the detection of organophosphate pesticides based on cyclodextrin functionalized graphitic carbon nitride and enzyme inhibition. Chemical Communications, 52(28), 5049-5052.

Wang, B., Ye, C., Zhong, X., Chai, Y., Chen, S., & Yuan, R. (2016b). Electrochemical biosensor for organophosphate pesticides and Huperzine‐A detection based on pd wormlike Nanochains/Graphitic carbon nitride nanocomposites and acetylcholinesterase. Electroanalysis, 28(2), 304-311.

Wang, B., Zhong, X., Chai, Y., & Yuan, R. (2017a). Ultrasensitive electrochemiluminescence biosensor for organophosphate pesticides detection based on carboxylated graphitic carbon nitride-poly (ethylenimine) and acetylcholinesterase. Electrochimica Acta, 224, 194-200.

Wang, C., Qu, L., Liu, X., Zhao, C., Zhao, F., Huang, F., et al. (2017b). Determination of a metabolite of nifursol in foodstuffs of animal origin by liquid–liquid extraction and liquid chromatography with tandem mass spectrometry. Journal of Separation Science, 40(3), 671-676.

Wang, H., Zhao, G., Chen, D., Wang, Z., & Liu, G. (2016a). A sensitive acetylcholinesterase biosensor based on screen printed electrode modified with

234

Chapter 6

Fe3O4 nanoparticle and graphene for chlorpyrifos determination. International Journal of Electrochemical Science, 11(12), 10906-10918.

Wang, J., Cormack, P. A., Sherrington, D. C., & Khoshdel, E. (2003). Monodisperse, molecularly imprinted polymer microspheres prepared by precipitation polymerization for affinity separation applications. Angewandte Chemie International Edition, 42(43), 5336-5338.

Wang, J., Suzuki, H., & Satake, T. (2014). Coulometric microdevice for organophosphate pesticide detection. Sensors and Actuators B: Chemical, 204, 297-301.

Wang, J., Yokokawa, M., Satake, T., & Suzuki, H. (2015). A micro IrO x potentiometric sensor for direct determination of organophosphate pesticides. Sensors and Actuators B: Chemical, 220, 859-863.

Wang, K., Wang, L., Jiang, W., & Hu, J. (2011). A sensitive enzymatic method for paraoxon detection based on enzyme inhibition and fluorescence quenching. Talanta, 84(2), 400-405.

Wang, L., Dong, J., Wang, Y., Cheng, Q., Yang, M., Cai, J., et al. (2016d). Novel signal- amplified fenitrothion electrochemical assay, based on glassy carbon electrode modified with dispersed graphene oxide. Scientific Reports, 6, 23409. doi:10.1038/srep23409 [doi]

Wang, S., Yu, C., & Wang, J. (2005). Enzyme immunoassay for the determination of carbaryl residues in agricultural products. Food Additives and Contaminants, 22(8), 735-742.

Wang, W., Han, Z., Liang, P., Guo, D., Xiang, Y., Tian, M., et al. (2017c). Co 3 O 4/PAn MAGNETIC NANOPARTICLE-MODIFIED ELECTROCHEMICAL IMMUNOSENSOR FOR CHLORPYRIFOS. Digest Journal of Nanomaterials & Biostructures (DJNB), 12(1)

Wang, X., Mu, Z., Shangguan, F., Liu, R., Pu, Y., & Yin, L. (2014a). Rapid and sensitive suspension array for multiplex detection of organophosphorus pesticides and carbamate pesticides based on silica–hydrogel hybrid microbeads. Journal of Hazardous Materials, 273, 287-292.

Wang, X., Qiao, X., Ma, Y., Zhao, T., & Xu, Z. (2013). Simultaneous determination of nine trace organophosphorous pesticide residues in fruit samples using molecularly imprinted matrix solid-phase dispersion followed by gas chromatography. Journal of Agricultural and Food Chemistry, 61(16), 3821-3827.

235

Chapter 6

Wang, X., Tang, Q., Wang, Q., Qiao, X., & Xu, Z. (2014b). Study of a molecularly imprinted solid‐phase extraction coupled with high‐performance liquid chromatography for simultaneous determination of trace trichlorfon and monocrotophos residues in vegetables. Journal of the Science of Food and Agriculture, 94(7), 1409-1415.

Wang, Z., Cao, Y., Ge, N., Liu, X., Chang, Q., Fan, C., et al. (2016a). Wide-scope screening of pesticides in fruits and vegetables using information-dependent acquisition employing UHPLC-QTOF-MS and automated MS/MS library searching. Analytical and Bioanalytical Chemistry, 408(27), 7795-7810.

Weetall, H. H., & Rogers, K. R. (2002). A simple assay for 2, 4-dichlorophenoxyacetic acid using coated test-strips. Analytical Letters, 35(8), 1341-1348.

Wei, C., Zhou, H., & Zhou, J. (2012). Ultrasensitively sensing acephate using molecular imprinting techniques on a surface plasmon resonance sensor. Talanta, 83(5), 1422- 1427.

Weller, M. G. (2000). Immunochromatographic techniques–a critical review. Fresenius' Journal of Analytical Chemistry, 366(6-7), 635-645.

Weselak, M., Arbuckle, T., & Foster, W. (2007). Pesticide exposures and developmental outcomes: The epidemiological evidence. Journal of Toxicology and Environmental Health, Part B, 10(1-2), 41-80.

Weston, D. P. (2000). Ecological effects of the use of chemicals in aquaculture. Paper presented at the Use of Chemicals in Aquaculture in Asia: Proceedings of the Meeting on the use of Chemicals in Aquaculture in Asia 20-22 may 1996, Tigbauan, Iloilo, Philippines, pp. 23-30.

WTO. (2017). Agreement of the application of sanitary and phytosanitary measures (SPS agreement), article 3, harmonisation. Retrieved June / 1, 2017, from https://www.wto.org/english/tratop_e/sps_e/spsagr_e.htm

Williamson, L. N., Terry, A. V., & Bartlett, M. G. (2006). Determination of chlorpyrifos and its metabolites in rat brain tissue using coupled‐column liquid chromatography/electrospray ionization tandem mass spectrometry. Rapid Communications in Mass Spectrometry, 20(18), 2689-2695.

Willis, K. J., & Ling, N. (2004). Toxicity of the aquaculture pesticide cypermethrin to planktonic marine copepods. Aquaculture Research, 35(3), 263-270.

236

Chapter 6

World Health Organization. (2010). The WHO recommended classification of pesticides by hazard and guidelines to classification 2009. Retrieved 6/24, 2017, from http://www.who.int/ ipcs/publications/pesticides_hazard/en/

Xiao-hong, L., Yu-hua, C., & E-nuo, G. (2009). Molecularly imprinted polymer microspheres for solid-phase extraction of methamidophos residue in milk. Chinese Journal of Analysis Laboratory, 6, 001.

Xie, C., Zhou, H., Gao, S., & Li, H. (2010). Molecular imprinting method for on-line enrichment and chemiluminescent detection of the organophosphate pesticide triazophos. Microchimica Acta, 171(3-4), 355-362.

Xin, J., Qiao, X., Xu, Z., & Zhou, J. (2013). Molecularly imprinted polymer as sorbent for solid-phase extraction coupling to gas chromatography for the simultaneous determination of trichlorfon and monocrotophos residues in vegetables. Food Analytical Methods, 6(1), 274-281.

Xu, M., Zhu, J., Su, H., Dong, J., Ai, S., & Li, R. (2012c). Electrochemical determination of methyl parathion using poly (malachite green)/graphene nanosheets–nafion composite film-modified glassy carbon electrode. Journal of Applied Electrochemistry, 42(7), 509-516.

Xu, T., Xu, Q. G., Li, H., Wang, J., Li, Q. X., Shelver, W. L., et al. (2012a). Strip-based immunoassay for the simultaneous detection of the neonicotinoid insecticides imidacloprid and thiamethoxam in agricultural products. Talanta, 101, 85-90.

Xu, Z., Deng, H., Deng, X., Yang, J., Jiang, Y., Zeng, D., et al. (2012b). Monitoring of organophosphorus pesticides in vegetables using monoclonal antibody-based direct competitive ELISA followed by HPLC–MS/MS. Food Chemistry, 131(4), 1569- 1576.

Xu, Z., Shen, Y., Zheng, W., Beier, R. C., Xie, G., Dong, J., et al. (2010). Broad- specificity immunoassay for O, O-diethyl organophosphorus pesticides: Application of molecular modeling to improve assay sensitivity and study antibody recognition. Analytical Chemistry, 82(22), 9314-9321.

Xu, Z., Xie, G., Li, Y., Wang, B., Beier, R. C., Lei, H., et al. (2009). Production and characterization of a broad-specificity polyclonal antibody for O, O-diethyl organophosphorus pesticides and a quantitative structure–activity relationship study of antibody recognition. Analytica Chimica Acta, 647(1), 90-96.

Yan, M. (2004). Molecularly imprinted materials: Science and technology CRC press.

237

Chapter 6

Yan, S., Fang, Y., Yao, W., & Gao, Z. (2007). Characterization and quality assessment of binding properties of the monocrotophos molecularly imprinted microspheres prepared by precipitation polymerization in toluene. Polymer Engineering & Science, 47(9), 1302-1308.

Yan, X., Li, H., Yan, Y., & Su, X. (2014). Developments in pesticide analysis by multianalyte immunoassays: A review. Analytical Methods, 6(11), 3543-3554.

Yang, Q., Sun, Q., Zhou, T., Shi, G., & Jin, L. (2009). Determination of parathion in vegetables by electrochemical sensor based on molecularly imprinted polyethyleneimine/silica gel films. Journal of Agricultural and Food Chemistry, 57(15), 6558-6563.

Yang, T., & Lee, M. (2010). Electrically assisted solid-phase microextraction combined with liquid chromatography–mass spectrometry for determination of parathion in water. Talanta, 82(2), 766-770.

Yao, W., Fang, Y., Gao, G. L., & Cheng, Y. (2008). Adsorption of carbaryl using molecularly imprinted microspheres prepared by precipitation polymerization. Polymers for Advanced Technologies, 19(7), 812-816.

Ye, L., & Mosbach, K. (2001). Molecularly imprinted microspheres as antibody binding mimics. Reactive and Functional Polymers, 48(1), 149-157.

Ye, L., Cormack, P. A., & Mosbach, K. (1999). Molecularly imprinted monodisperse microspheres for competitive radioassay. Analytical Communications, 36(2), 35-38.

Ye, L., Mosbach, K., Spivak, D. A., & Sibrian-Vazquez, M. (2002). Molecularly imprinted materials: Towards the next generation. MRS Online Proceedings Library Archive, 723.

Yoshimatsu, K., Reimhult, K., Krozer, A., Mosbach, K., Sode, K., & Ye, L. (2007). Uniform molecularly imprinted microspheres and nanoparticles prepared by precipitation polymerization: The control of particle size suitable for different analytical applications. Analytica Chimica Acta, 584(1), 112-121.

Yoshioka, N., Akiyama, Y., & Teranishi, K. (2004). Rapid simultaneous determination of o-phenylphenol, diphenyl, thiabendazole, imazalil and its major metabolite in citrus fruits by liquid chromatography-mass spectrometry using atmospheric pressure photoionization. Journal of Chromatography A, 1022(1), 145-150.

238

Chapter 6

Yu, T., Ying, T., Song, Y., Li, Y., Wu, F., Dong, X., et al. (2014). A highly sensitive sensing system based on photoluminescent quantum dots for highly toxic organophosphorus compounds. Rsc Advances, 4(16), 8321-8327.

Zacharia, J. T. (2011). Identity, physical and chemical properties of pesticides. Pesticides in the Modern World-Trends in Pesticides Analysis, 10-14.

Zeng, K., Yang, T., Zhong, P., Zhou, S., Qu, L., He, J., et al. (2007). Development of an indirect competitive immunoassay for parathion in vegetables. Food Chemistry, 102(4), 1076-1082.

Zeng, Y., Yu, D., Yu, Y., Zhou, T., & Shi, G. (2012). Differential pulse voltammetric determination of methyl parathion based on multiwalled carbon nanotubes–poly (acrylamide) nanocomposite film modified electrode. Journal of Hazardous Materials, 217, 315-322.

Zhang, C., Zhang, Y., & Wang, S. (2006). Development of multianalyte flow-through and lateral-flow assays using gold particles and horseradish peroxidase as tracers for the rapid determination of carbaryl and endosulfan in agricultural products. Journal of Agricultural and Food Chemistry, 54(7), 2502-2507.

Zhang, L., Liu, S., Cui, X., Pan, C., Zhang, A., & Chen, F. (2012a). A review of sample preparation methods for the pesticide residue analysis in foods. Open Chemistry, 10(3), 900-925.

Zhang, L., Wang, Z., Wen, Y., Shi, J., & Wang, J. (2015b). Simultaneous detection of parathion and imidacloprid using broad-specificity polyclonal antibody in enzyme- linked immunosorbent assay. Analytical Methods, 7(1), 205-210.

Zhang, W., Asiri, A. M., Liu, D., Du, D., & Lin, Y. (2014). Nanomaterial-based biosensors for environmental and biological monitoring of organophosphorus pesticides and nerve agents. TrAC Trends in Analytical Chemistry, 54, 1-10.

Zhang, Y., Arugula, M. A., Wales, M., Wild, J., & Simonian, A. L. (2015a). A novel layer-by-layer assembled multi-enzyme/CNT biosensor for discriminative detection between organophosphorus and non-organophosphrus pesticides. Biosensors and Bioelectronics, 67, 287-295.

Zhang, Y., Kang, T., Wan, Y., & Chen, S. (2009). Gold nanoparticles-carbon nanotubes modified sensor for electrochemical determination of organophosphate pesticides. Microchimica Acta, 165(3-4), 307-311.

239

Chapter 6

Zhang, Y., Qian, L., Yin, W., He, B., Liu, F., Hou, C., et al. (2016). A dual read-out molecularly imprinted composite membrane sensor based on zinc porphyrin for the detection of dimethyl methylphosphonate. Chemical Research in Chinese Universities, 32(5), 725-730.

Zhang, Z., Cheng, Z., Zhang, C., Wang, H., & Li, J. (2012c). Precipitation polymerization of molecularly imprinted polymers for recognition of melamine molecule. Journal of Applied Polymer Science, 123(2), 962-967.

Zhang, Z., Li, P., Hu, X., Zhang, Q., Ding, X., & Zhang, W. (2012b). Microarray technology for major chemical contaminants analysis in food: Current status and prospects. Sensors, 12(7), 9234-9252.

Zhao, T., Gao, H., Wang, X., Zhang, L., Qiao, X., & Xu, Z. (2014a). Study on a molecularly imprinted solid-phase extraction coupled to capillary electrophoresis method for the determination of trace trichlorfon in vegetables. Food Analytical Methods, 7(6), 1159-1165.

Zhao, T., Wang, Q., Li, J., Qiao, X., & Xu, Z. (2014b). Study on an electrochromatography method based on organic–inorganic hybrid molecularly imprinted monolith for determination of trace trichlorfon in vegetables. Journal of the Science of Food and Agriculture, 94(10), 1974-1980.

Zheng, W., Park, J., Zhang, D., El-Aty, A. A., Kim, S., Cho, S., et al. (2017). Determination of fenobucarb residues in animal and aquatic food products using liquid chromatography-tandem mass spectrometry coupled with a QuEChERS extraction method. Journal of Chromatography B, 1058, 1-7.

Zheng, Y., Liu, Z., Zhan, H., Li, J., & Zhang, C. (2016). Studies on electrochemical organophosphate pesticide (OP) biosensor design based on ionic liquid functionalized graphene and a co 3 O 4 nanoparticle modified electrode. Analytical Methods, 8(26), 5288-5295.

Zhou, J., Ma, C., Zhou, S., Ma, P., Chen, F., Qi, Y., et al. (2010). Preparation, evaluation and application of molecularly imprinted solid-phase microextraction monolith for selective extraction of pirimicarb in tomato and pear. Journal of Chromatography A, 1217(48), 7478-7483.

Zhu, X., Yang, J., Su, Q., Cai, J., & Gao, Y. (2005). Selective solid-phase extraction using molecularly imprinted polymer for the analysis of polar organophosphorus pesticides in water and soil samples. Journal of Chromatography A, 1092(2), 161-169.

240

Chapter 6

Zou, Z., Du, D., Wang, J., Smith, J. N., Timchalk, C., Li, Y., et al. (2010). Quantum dot- based immunochromatographic fluorescent biosensor for biomonitoring trichloropyridinol, a biomarker of exposure to chlorpyrifos. Analytical Chemistry, 82(12), 5125-5133.

241

Appendices:

243

Appendix A: Publications

244