The Development of Immunotechniques for Environmental Monitoring of Pesticides

THE DEVELOPMENT OF IMMUNOTECHNIQUES FOR ENVIRONMENTAL MONITORING OF PESTICIDES

By MOHAMAD FAWAZ KATMEH BSc, MSc [Surrey)

A thesis submitted to the University of Surrey for the degree of Doctor of Philosophy

November 1994

University of Surrey Robens Institute of Health and Saftey Guildford, Surrey GU2 5XH ProQuest Number: 27600339

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a com plete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest

ProQuest 27600339

Published by ProQuest LLO (2019). Copyright of the Dissertation is held by the Author.

ProQuest LLO. 789 East Eisenhower Parkway P.Q. Box 1346 Ann Arbor, Ml 48106- 1346 S u m m a r y

This research project has investigated the feasibility of using immunotechniques for the immuno-detection and monitoring of pesticides in different matrices, mainly water residues. The maximum admissible concentration (MAC) for a single pesticide (0.1/^g/L), which was set by the European community, has been successfully detected by the use of immunoassay methods. Competitive enzyme-linked immunosorbent assays (ELISA) suitable for the determination of the urea herbicides isoproturon, chlortoluron and 2,4-D in different types of water and biological fluids have been developed. The production of highly specific antibodies utilised in this work was achieved by covalently coupling thyroglobuhn protein with the compound or a synthesised hapten. As a result, the cross -reactivities of the antisera with a range of related and unrelated pesticides was shown to be negligible. Furthermore, the limits of detection of the isoproturon and chlortoluron ELISA methods were 0.03 and 0.015 ^Mg/L respectively, well below the MAC for pesticides in drinking water, whilst the ELISA assay limit of detection for 2,4-D was 50 y«g/L which would be sensitive enough for monitoring 2,4-D in air for occupational exposure studies. In addition, reproducible and quantitative recoveries of isoproturon, chlortoluron and 2,4-D firom water, obtained from various sources, and biological fluids were possible without any sample preparation requirements. The feasibility of using a luminescent assay with a photographic end point has shown considerable potential by providing a rapid, simple and portable means of monitoring multiple water samples for the presence of pesticides. The assay was able to identify samples containing a pesticide or a class group of pesticides at or above the MAC level. Consequently, this would greatly facihtate the increased monitoring of water supphes in other areas of environmental analysis. A further study was carried out to demonstrate that screening pesticides qualitatively with non-instrumental immunoaffinity columns based on the principles of affinity chromatography and enzyme immunoassay was possible. Summary

The method produced quahtative results in 20 min by determining samples containing certain concentrations of pesticides and above as positive samples (Le. 0.12 iMg/L or above for isoproturon and 0.11 /Wg/L or above for chlortoluron). Samples with a concentration of pesticides below these levels would be considered as negative (no pesticides present at these limits of detection). The advantages of this procedure are that it requires no sample dilutions, separation steps, or precise timing (except for reading colour). Moreover, the technique characteristics made it ideally suited for monitoring pesticides on-site in a different range of matrices which could be performed by non -skilled personnel thereby reducing the workload involved in meeting EC regulations. The data reported in this thesis demonstrate the potential use of immunoassays in environmental analysis potentially offering a substantial improvement to the monitoring programmes of pesticides by providing tests which are sensitive, rapid, inexpensive and simple to perform. This should encourage further work on the development of immunoassays for other pesticides and for toxic or hazardous chemicals in general.

n TO my wife Claire, my son Rateb and my parents

m Acknowledgements

I would like to thank the Health and Safety Executive for providing the funding for this project. In the completion of this thesis I would like to express my special thanks to my supervisors Dr Wynne Aheme and Dr Derek Stevenson for their guidance and advice in conducting this task. My thanks must be extended to Dr George Frost, who sadly passed away recently, for his invaluable experience and co operation regarding the chemical synthesis of the haptens. I take this opportunity to extend my condolences to his family. I would also like to thank Dr Jack Firth for his invaluable advice on this work. My grateful thanks go to all members of ClifMar Associates Limited, in particular Sean Lovegrove, for their endless help. My gratitude must also be extended to Dr Piotr KwasowskL and Dr Miguel Godfrey for their technical advice and help during the chromatographic study development. I would like to thank those who have worked in 26 AY 19 School of Biological Sciences over the past three years for providing a friendly atmosphere in which to work. I would like to thank Dr Shelagh Hampton and Mr Brian Morris for their support and encouragement throughout the course of this work. Finally, the unceasing and selfless enthusiasm and help of my dear wife and parents have been a constant source of encouragement throughout this study.

IV C o n te n ts

Sum m ary...... i Acknowledgements...... iv C o n te n ts ...... v F ig u res...... xii T ab les...... xviii Abbreviations...... xxii

CHAPTER ONE INTRODUCTION...... 1 1.1 GENERAL INTRODUCTION...... 2 1.2 HISTORY ...... 2 1.3 PESTICIDES, CHEMICALS AND THE ENVIRONMENT ...... 3 1.3.1 Water contamination by pesticides...... 3 1.3.1.1 Process of dissipation ...... 3 1.3.2 Soil contamination by pesticides ...... 8 1.4 PESTICIDES: THE ENVIRONMENTAL LEGACY...... 9 1.5 TOXICITY OF PESTICIDES ...... 11 1.5.1 Short term health effects ...... 11 1.5.2 Long term health effects ...... 12 1.6 PESTICIDES CLASSIFICATION...... 14 1.6.1 Herbicides ...... 14 1.6.2 Fungicides ...... 16 1.6.3 Insecticides ...... 16 1.7 MODE OF ACTION OF HERBICIDES...... 16 1.7.1 Inhibition of respiration ...... 16 1.7.2 Inhibition of photosynthesis ...... 16 1.7.3 Interference with protein synthesis ...... 17 1.7.4 Interference with lipid synthesis ...... 17 1.8 DETERMINATION OF HERBICIDE RESIDUES...... 17 1.8.1 Standard methods ...... 17 1.8.2 Imm unoassays ...... 20

V 1.8.2.1 Immunological principles ...... 20 1.8.2.2 Antisera ...... 22 1.8.2.3 Labels (Tracers) ...... 23 a) Radioim m unoassay ...... 23 b) Enzyme im m unoassay ...... 25 c) Enhanced luminescence in ELISA ...... 29 1.8.2.4 New approaches for trace immuno-analysis ...... 30 1.9 APPLICATION OF IMMUNOASSAY TO PESTICIDES ANALYSIS...... 30 1.9.1 Food analysis ...... 31 1.9.2 Water analysis ...... 31 1.9.3 Soil and plant analysis ...... 31 1.9.4 Human exposure studies ...... 32 1.10 AIMS OF PRESENT STUDY...... 32

CHAPTER TWO MATERIALS AND METHODS...... 34 2.1 REAGENTS ...... 35 2.1.1 General reagents ...... 35 2.1.2 Herbicide compounds ...... 36 2.1.3 Radiolabels ...... 36 2.1.4 Organic solvents ...... 36 2.1.5 Disposable plastics ...... 37 2.1.6 Manual liquid handling ...... 37 2.1.7 Dialysis ...... 37 2.1.8 Glassware ...... 37 2.2 EQUIPMENT ...... 38 2.2.1 General equipment ...... 38 2.2.2 Specialised equipment ...... 38 2.3 MATERIALS ...... 38 2.3.1 General buffer solutions ...... 38 2.3.2 Radiolabels ...... 40 2.3.3 Dextran-coated charcoal (DCC, 2.5%) ...... 40 2.3.4 Preparation of pesticide stock standard solutions ...... 40 2.3.5 Water sam ples ...... 40 2.4 METHODS...... 41

VI 2.4.1 Antiserum production ...... 41 2.4.1.1 Preparation of immunogen ...... 42 2.4.1 2 Antisera ...... 43 2.4.2 RIA procedures ...... 43 2.4.3 DEAE purification of antisera ...... 45 2.4.4 HRPO-hapten conjugate ...... 45 2.4.5 Competitive ELISA ...... 46 2.4.6 Competitive ECLIA ...... 46 2.4.7 Competitive Immunoaffinity column assay ...... 47

CHAPTER THREE DEVELOPMENT OF ISOPROTURON AND CHLORTOLURON ELISA METHODS ...... 48 3.1 INTRODUCTION TO UREA HERBICIDES...... 49 3.2 HAPTENS...... 50 3.2.1 Chemical synthesis of isoproturon hapten "N-[4-(l- methylethyl)phenyl]-N-methyl-N-carboxypropyl- urea" ...... 50 3.2.2 Chemical synthesis of chlortoluron hapten [N-(3-chloro-4- methylphenyl)-N-methyl-N-carboxypro^l urea] ...... 52 3.3 IMMUNOGENS...... 53 3.4 IMMUNISATION PROCEDURE...... 54 3.5 ANTISERA SCREENING PROCEDURE...... 54 3.6 PURIFICATION OF ANTISERA...... 55 3.7 HRPO-HAPTEN CONJUGATE...... 55 3.8 OPTIMISATION OF ASSAY CONDITIONS...... 56 3.8.1 Isoproturon ELISA ...... 61 3.8.1.1 Evaluation of the purified antiserum (titre and displacement) ...... 61 3.8.1.2 Assay protocol for standard curve ...... 62 3.8.1.3 Assay standard curve optimisation ...... 62 3.8.1.3.1 Microtitre plates ...... 62 3.8.1.3.3 Coating antibody ...... 64 3.8.1.3.2 Enzyme label ...... 64 3.8.1.3.4 Incubation tim e ...... 64 3.8.1.4 Assay validation ...... 67 3.8.1.4.1 Cross-reactivity ...... 67

vii 3.8.1.4.2 Recovery of isoproturon from different water samples ...... 70 3.8.1.4.3 Matrix effects ...... 70 3.8.1.4.4 Competitive isoproturon ELISA final protocol ...... 74 3.8.1.4.5 Recovery of isoproturon from different water samples using the final protocol ...... 75 3.8.1.4.6 Assay sensitivity and reproducibility ...... 77 3.8.1.4.7 Stability and preparation of standards ...... 78 3.8.1.4.8 Measuring isoproturon in biological fluids using ELISA ...... 81 3.8.1.4.8.1 Detection of isoproturon (or metabolites) in mice ...... 85 3.8.2 Chlortoluron ELISA ...... 86 3.8.2.1 Evaluation of the purified antiserum (titre and displacement) ...... 86 3.8.2 2 Assay standard curve optimisation ...... 88 3.8.2.2.1 Coating antibody ...... 88 3.8.2.2.2 Enzyme label ...... 88 S.8.2.3 Assay validation ...... 92 3.8.2.3.1 Cross-reactivity ...... 92 3.8.2.3.2 Competitive chlortoluron ELISA final protocol ...... 93 3.8.2.3.3 Recovery of chlortoluron from different water samples using the final protocol ...... 93 3.8.2.3.4 Assay sensitivity and reproducibility ...... 93 3.8.2.3.5 Stability and preparation of standards ...... 95 3.8.2.3.6 Measuring chlortoluron in biological fluids using ELISA ...... 95 3.9 DISCUSSION...... 101 3.9.1 Antisera ...... 102 3.9.2 Assay optimisation ...... 103 3.9.3 Assay validation ...... 104

mil CHAPTER FOUR DEVELOPMENT OF 2,4-D AND MCPA ELISA METHODS...... 109 4.1 INTRODUCTION...... 110 4.2 2,4-D AND MCPA IMMUNOGENS...... I l l 4-3 IMMUNISATION PROCEDURE...... 112 4.4 ANTISERA SCREENING PROCEDURE...... 112 4.5 PURIFICATION OF ANTISERA...... 112 4.6 HRPO-HAPTEN CONJUGATE...... 115 4.7 2,4-D ELISA...... 115 4.7.1 Antiserum dilution and displacement curves ...... 115 4.7.2 Assay protocol for 2,4-D standard curve ...... 117 4.7.3 Standard curve optimisation ...... 117 4.7.3.1 Coating antibody ...... 117 4.7.3.2 2,4-D peroxidase label ...... 117 4.7.4 Competitive 2,4-D ELISA final protocol ...... 120 4.7.4 Assay validation ...... 120 4.7.4.1 Cross-reactivity ...... 120 4.7.4.2 2,4-D assay variation ...... 121 4.7.4.3 Stability of 2,4-D standard ...... 124 4.7.4.4 Recovery of 2,4-D from various w ater ...... 124 4.7.4.5 Recovery of 2,4-D from biological fluids ...... 126 4.8 MCPA ELISA...... 129 4.8.1 Antiserum and displacement curves ...... 129 4.8.2 Alternative MCPA im m unogen ...... 134 4.8.2.1 Chemical synthesis of MCPA h a p te n ...... 134 4.8.2.2 MCPA hapten-thyroglobulin conjugate ...... 134 4.8.2.3 Production and screening of the MCPA antisera ...... 135 4.9 DISCUSSION...... 136 4.9.1 2,4-D ...... 136 4.9.2 MCPA ...... 137

CHAPTER FIVE DEVELOPMENT OF ISOPROTURON AND CHLORTOLURON ECUA METHODS...... 139 5.1 INTRODUCTION...... 140 5.2 ECLLA FOR ISOPROTURON AND CHLORTOLURON...... 144

IX 5.2.1 Assay sensitivity, reproducibility and recovery of spiked samples ...... 147 5.3 PHOTOGRAPHIC ENHANCED CHEMLUMINESCENT ENDPOINT ASSAY FOR CHLORTOLURON...... 152 5.3.1 Types of photographic film ...... 152 5.3.2 Camera luminometer ...... 153 5.3.3 Photographic assay technique ...... 153 5.3.4 Photographic assay validation...... 156 5.3.4.1 Comparison between standards and spiked samples 156 5.3.4.2 Accuracy of chlortoluron photographic endpoint assay...... 159 5.3.4.3 Correlation between chlortoluron ELISA, ECLIA and photographic endpoint techniques ...... 159 5.4 DISCUSSION...... 161

CHAPTER SIX DEVELOPMENT OF IMMUNOAFFINITY CHROMATOGRAPHY COLUMNS FOR ISOPROTURON AND CHLORTOLURON...... 164 6.1 INTRODUCTION...... 165 6.1.1 History...... 165 6.1.2 The theory of immunoaffinity chromatography ...... 165 6.1.3 Antibody-antigen interaction ...... 166 6.2 ISOPROTURON IMMUNOAFFINITY COLUMNS...... 168 6.2.1 Immobilisation of antibodies and column preparation ...... 168 6.2.1.1 Affinity support matrices ...... 168 6.2.1.2 Antibody immobilisation ...... 169 6.2.2 Optimisation of the elution conditions using ^^C isoproturon label ...... 170 6.2.3 Theoretical binding capacity of the columns ...... 172 6.2.4 Isoproturon standard curve using isoproturon label ...... 173 6.2 4.1 A direct competitive assay ...... 180 6.2.4.2 An indirect competitive assay ...... 180 6.2.5 Optimisation of the conditions using HRPO isoproturon label 182 6.2.5.1 Choice of elution buffer ...... 182 6.2.5.2 Standard curve using isoproturon peroxidase enzyme label ...... 184

X G.2.5.3 Choice of elution buffer volume ...... 187 6.2.5.4 Choice of isoproturon enzyme peroxidase concentration ...... 187 G.2.5.5 Optimisation of the analyte sample mixture volume applied onto the column ...... 191 6.2.6 Validation of the isoproturon standard curve using immunoaffinity columns ...... 191 6.2.6.1 Indirect competitive isoproturon assay final protocol 191 G.2.6.2 Assay sensitivity and reproducibility "Inter-column variation" ...... 193 G.2.6.3 Stability and preparation of the analyte mixture ...... 193 G.2.6.4 Colour development within the columns ...... 194 6.2.6.5 Correlation studies between isoproturon ELISA and immunoaffinity column technique ...... 197 6.3 CHLORTOLURON IMMUNOAFFINITY COLUMNS...... 199 6.3.1 Antibody immobilisation ...... 199 6.3.2 Assessm ent of theoretical binding capacity of the colum ns...... 199 6.3.3 Optimisation of the chlortoluron enzyme peroxidase concentration ...... 201 6.3.4 Accuracy of chlortoluron immunoaffinity column system ...... 205 6.3.5 Correlation studies between chlortoluron ELISA and Immunoaffinity column technique ...... 207 6.4 DISCUSSION...... 208

CHAPTER SEVEN GENERAL DISCUSSION AND CONCLUSION...... 214 7.1 GENERAL DISCUSSION...... 215 7.2 FUTURE WORK...... 221 7.3 CONCLUDING REMARKS...... 222 REFERENCES...... 224 APPENDIX...... 238

XI F ig u r e s

Paae No Figure 1-1 Six departments involved in pesticides approval for use. 9 Figure 1-2 The principles of competitive immunoassays. 21 Figure 1-3 A schematic diagram showing the stages in a coated 27 antibody ELISA. Figure 1-4 A schematic diagram showing the stages in a coated 28 antigen ELISA.

Figure 2-1 N-hydroxysuccinimide "active ester” conjugation 42 method.

Figure 3-1 Urea herbicides (ureides). 49 Figure 3-2 Chemical synthesis of the isoproturon hapten. 52 Figure 3-3 Chemical synthesis of the chlortoluron hapten. 53 Figure 3-4 Isoproturon and chlortoluron protein conjugates. 54 Figure 3-5 Immunisation charts for (a) isoproturon and (b) 57 chlortoluron. Figure 3-6 Evaluation of isoproturon and chlortoluron antisera 58 using RIA (I'^C labels). Figure 3-7 Purification of isoproturon 4676 (IV:B) and 59 chlortoluron 1826 (V:C) bleeds on DEAE eluted with 0.0IM phosphate buffer pH 6.4. Figure 3-8 Comparison between isoproturon standard curves 60 using recently and previously prepared isoproturon peroxidase label. Figure 3-9 Isoproturon antiserum dilution and displacement 63 curves using ELISA. Figure 3-10 Optimisation of the isoproturon coating antibody 65 dilutions usm g ELISA.

J C ll Figures

Paae No

Figure 3-11 Optimisation of the isoproturon HRPO label for molar 66 ratio and dilution using ELISA. Figure 3-12 The effect of various incubation conditions on 68 isoproturon standard curves (n=4). Figure 3-13 The effect of different matrices used as dilution 72 buffer on the Bo isoproturon standard curve. Figure 3-14 The effect of adding normal serum to the isoproturon 73 HRPO label solution on the Bo standard curve. Figure 3-15 hitra-plate variation of isoproturon standard 79 curve with precision profile using ELISA. Figure 3-16 Isoproturon standard curves constructed by 80 colleagues with varied experience in immunoassay. Figure 3-17 Intra-plate variation of isoproturon standard 84 curves constructed in plasma and urine with precision profiles using ELISA. Figure 3-18 The elimination of "immunoreactive" isoproturon from 87 mice injected with isoproturon and measured by ELISA. Figure 3-19 Chlortoluron antiserum dilution and displacement 89 curves using ELISA. Figure 3-20 Optimisation of the chlortoluron coating antibody 90 dilutions using ELISA. Figure 3-21 Optimisation of the chlortoluron HRPO label for 91 molar ratio and dilution using ELISA. Figure 3-22 Intra-assay variation of chlortoluron standard curve 96 with a precision profile using ELISA. Figure 3-23 precision of chlortoluron standard curve with 98 chlortoluron standard solutions at pH 5 and stored at 4°Cover a period of one month. Figure 3-24 Chlortoluron standard curves constructed hi plasma 100 and urine with precision profiles.

X l l l Figures

Paae No

Figure 4-1 2,4-D and MCPA chemical structures. 110 Figure 4-2 2,4-D and MCPA protein conjugates. I l l Figure 4-3 The immunisation charts for (a) 2,4-D and (b) MCPA 113 immunisation charts. Figure 4-4 Purification profile of (a) 2,4-D and (b) MCPA antibodies. 114 Figure 4-5 Antiserum dilution and displacement curves of 2,4-D 116 using ELISA. Figure 4-6 Optimisation of the 2,4-D coating antibody dilution 118 using ELISA. Figure 4-7 Optimisation of the 2,4-D HRPO label for molar ratio 119 and dilution. Figure 4-8 Intra-plate variation of 2,4-D standard curve with 123 precision profile. Figure4-9 2,4-D standard curves made up in plasma and urine 128 with precision profiles. Figure 4-10 MCPA antiserum dilution and displacement curves 130 using ELISA. Figure 4-11 Displacement of MCPA antiserum dilution curves with 131 increasing concentrations of MCPA standard. Figure 4-12 Displacement of MCPA antiserum dilution curves 132 withlOmg/L MCPA, immunogen and thyroglobulin. Figure 4-13 Displacement of MCPA antiserum dilution curves with 133 lOmg/L MCPA, immunogen and thyroglobulin using a non-equilibrium system. Figure 4-14 The proposed chemical structure of the synthesised 134 MCPA hapten. Figure 4-15 Imm unisation chart for the MCPA immunogen 135 prepared with adipimidate using ELISA.

XIV Figures

Page No Figure 5-1 Chemilmniiiescent reaction of luminol. 141 Figure 5-2 Signal characteristics of enhanced luminescence 141 (adapted from Amersham International brochure). Figure 5-3 Proposed reactions sequence in the peroxidase- 142 catalysed chemiluminescent oxidation of luminol (adapted from Misra and Squatrito, 1982). Figure 5-4 Illustrating the reaction sequences for enhanced 143 chemiluminescent light production. Figure 5-5 Isoproturon and chlortoluron standard curves 146 obtained by ECLIA. Figure 5-6 Intra-assay variation of ECLIA isoproturon standard 150 curve. The mean and SD of 6 wells for each standard on one plate. Figure 5-7 hitra-assay variation of ECLIA chlortoluron standard 151 curve. The mean and SD of 6 wells for each standard on one plate. Figure 5-8 Dynatech camera luminometer. 154 Figure 5-9 Optimisation of chlortoluron antibody and enzyme 155 label dilutions using the camera luminometer. Figure 5-10 Visual comparison between different concentrations of 157 chlortoluron standards using the photographic technique. Figure 5-11 Chlortoluron assay, using camera luminometer, 158 containing chlortoluron standards and spiked river water samples.

Figure 6-1 Diagrammatic representation of the mechanism 166 involved in immunoaffinity chromatography. Figure 6-2 The intermolecular attractive forces binding antigen 167 to antibody (adapted from Roitt et al, 1985).

XV Figures

Paae No Figure 6-3 Purification profile of the isoproturon solution 174 using an immunoaffinity column with 0.3% HCl, pH 2, as the elution buffer. Figure 6-4 The effect of adding different percentages of ethanol to 175 the 0.3% HCl elution buffer, pH 2, on the elution profile of isoproturon label using an immunoaffinity column. Figure 6-5 Purification profile (n=5) of the ^"^C isoproturon label 176 using a mixture of 0.3% HCl, pH 2, and 35% ethanol as an elution buffer. Figure 6-6 The breakthrough curves for isoproturon 177 immunoaffinity columns packed with various concentrations of antiserum. Figure 6-7 The theoretical binding capacities of isoproturon 178 immunoaffinity columns, ng/ 1.2g wet weight solid phase, packed with different concentrations of antiserum . Figure 6-8 The theoretical binding capacity of isoproturon 179 immunoaffinity columns containing various amounts of the immobilised solid phase. Figure 6-9 The mean of six isoproturon standard curves 181 performed by an indirect competitive assay on immunoaffinity columns using isoproturon radiolabel. Figure 6-10 The effect of various elution buffers on the peroxidase 183 protein activity. Figure 6-11 Presentation of a competitive enzyme immunoassay 185 system using immunoaffinity columns. Figure 6-12 Inter-column variation of isoproturon immunoaffinity 186 column for six standard curves.

XVI Figures

Paae No Figure 6-13 Optimisation of the elution buffer (mixture of PBS 188 buffer and 35% ethanol) volume required to recover the retained isoproturon enzyme label solution. Figure 6-14 Optimisation of the isoproturon enzyme label dilutions 189 to enhance the sensitivity of the isoproturon standard curve using immunoaffinity columns. Figure 6-15 The effect of adding 2.5% normal serum to the 190 isoproturon label solution on the sensitivity of isoproturon standard curve using immunoaffinity columns. Figure 6-16 Optimisation of the analyte mixture volume required 192 to saturate the 50mg isoproturon SPE column. Figure 6-17 The effect of reducing the pH of the analytes mixture 194 (pH 5) on the isoproturon standard curve limit of detection. Figure 6-18 The breakthrough curves for chlortoluron 200 immunoaffinity columns packed with various concentrations of antiserum. Figure 6-19 The theoretical binding capacities of chlortoluron 202 immunoaffinity columns packed with different concentrations of antiserum. Figure 6-20 The theoretical binding capacity of chlortoluron 203 immunoaffinity columns containing various amounts of the immobilised solid phase. Figure 6-21 The effect of different dilutions of the chlortoluron 204 enzyme label upon the sensitivity of the standard curve using chlortoluron immunoaffinity colu m n s. Figure 6-22 Inter-column variation of chlortoluron standard 206 curves using chlortoluron immunoaffinity column.

XVll T a b le s

Paae No Table 1-1 Pesticide breaches (Lees and McVeigh, 1988) in 8 England during the period July 1985-June 1987. Table 1-2 Summary of studies investigating possible links 13 between cancer and pesticide use among farmers. Table 1-3 Classification of herbicides (adapted from Has sail, 15 1990). Table 1-4 Labels used in immunoassay. 25

Table 2-1 Protocol for antiserum assessment. 44

Table 3-1 Specificity of isoproturon antiserum towards a 69 selection of herbicides. Table 3-2 The recovery of isoproturon from different spiked water 70 samples by using ELISA. Table 3-3 Improvement of isoproturon recovery from tap water 74 samples by the addition of normal serum to the enzyme label solution. Table 3-4 Recovery of isoproturon from different water sources 76 by using ELISA (final protocol). Table 3-5 Intra-plate variation for isoproturon standard curves 77 using ELISA. Table 3-6 Inter-plate variation for isoproturon standard curves. 78 Table 3-7 The stability of standard solutions in different 81 storage and pH conditions. Table 3-8 Intra and inter-assay variation of isoproturon 82 standard curves in plasma and the recovery of isoproturon from spiked plasma samples.

x v i i t Tables

Paae No Table 3-9 Intra and inter-assay variation of isoproturon 83 standard curves in urine and the recovery of isoproturon in spiked urine samples. Table 3-10 Amount of "immunoreactive" isoproturon measured by 86 ELISA. Table 3-11 Specificity of the chlortoluron antiserum towards a 92 selection of herbicides. Table 3-12 Percentage recovery of chlortoluron from different 94 water sources assayed using the chlortoluron ELISA (final protocol). Table 3-13 Intra-plate variation between chlortoluron standard 97 curves using ELISA. Table 3-14 Inter-plate variation of chlortoluron standard 97 curves using ELISA. Table 3-15 Intra and inter-assay variation of chlortoluron 99 standard curves in plasma and the recovery of chlortoluron from spiked plasma samples. Table 3-16 Intra and inter-assay variation of chlortoluron 101 standard curves in urine and the recovery of chlortoluron from spiked urine samples.

Table 4-1 Specificity of 2,4-D antiserum towards a selection of 121 herbicides. Table 4-2 Within-plate variation for 2,4-D standard curves 122 using ELISA. Table 4-3 Between-plate variation for 2,4-D standard curves 122 using ELISA. Table 4-4 Stability of 2,4-D solutions in tap water stored at 4PC 124 over a period of one month.

XIX Tables

Pooe No Table 4-5 Recovery of 2,4-D from different water samples 125 using ELISA. Table 4-6 Intra and inter-assay variation of 2,4-D standard 126 curves in plasma and the recovery of 2,4-D from spiked plasma samples. Table 4-7 Intra and inter-assay variation of 2,4-D standard 127 curves in urine and the recovery of 2,4-D in spiked urine samples.

Table 5-1 Within-plate variation of isoproturon and 145 chlortoluron standard curves using ECLIA obtained at various incubation times of the Amerlite reagent. Table 5-2 Intra and inter-assay variation of isoproturon 148 standard curves using ECLIA and the recovery of isoproturon from spiked tap water samples over a period of two weeks. Table 5-3 Intra and inter-assay variation of chlortoluron 149 standard curves using ECLIA and recovery of chlortoluron from spiked tap water samples over a period of two weeks. Table 5-4 Accuracy of chlortoluron photographic endpoint 159 immunoassay using a Dynatech camera luminometer. Table 5-5 Comparison of results for chlortoluron concentration 160 obtained by ELISA, ECLIA and photographic endpoint techniques.

Table 6-1 Properties required by the affinity support matrix 168 (Weetall, 1973). Table 6-2 Immunoaffinity chromatographic matrices. 169 Table 6-3 Inter-column variation for isoproturon standard curve 193 using immunoaffinity columns.

XX T ables

Paae No Table 6-4 Accuracy of isoproturon immunoaffinity columns. 197 Table 6-5 Comparison of results for isoproturon concentration 198 obtained by ELISA and Immunoaffinity colum n methods. Table 6-6 hiter-column variation for chlortoluron standard 205 curve using immunoaffinity columns. Table 6-7 Accuracy of chlortoluron immunoaffinity columns. 207 Table 6-8 Comparison of chlortoluron concentration results 208 obtained by ELISA and immunoaffinity column methods.

XXI Abbreviations

2,4-D 2-4-dichlorophenoxy acetic acid Ab Antibody AChE Acetylcholinesterase Ag Antigen Ag* Antigen Label AMPPD Disodium 3-(4-methoxyspiro[ 1,2-dioxentane-3,2'-tricyclo]-4- yDphenyl phosphate A/S Antiserum

BCG Bacillus Calmette-Guerin Bo Binding (in absence of unlabeUed analyte)

Cone Concentration CV Coefficient of Variation

DCC Dextran-Coated Charcoal DEAE Diethylaminoethyl Dist Distilled DMA Dimethyl Adipimidate DMF Dimethyl Formamide

E Enzyme ECLIA Enhanced Chemiluminescent Immunoassay EIA Enzyme Immunoassay ELISA Enzyme-Ltoked Imm unosorbent Assay

GC Gas Chromatography

H Harvesting HPLC High Performance Liquid Chromatography hr Hour HRPO Horse-radish Peroxidase

IgG Immunoglobulin G IP Intra-Peritoneal

xxii Abbreviations

M Mouse MAC Maximum Admissible Concentration MCPA 2-metbyl-4-chlorophenoxy acetic acid m in Minute

N/A Not Applicable NC Negative Control NSB Non Specific Binding

OD Optical Density OPD O-phenylenediamine

PBS Phosphate Buffered Saline PBSG Phosphate Buffered Saline & Gelatin PBST Phosphate Buffered Saline & 0.05% v/v "Tween 20 PS Porous Silica

RIA Radioimmunoassay

SD Standard Deviation S t Standard STS Soft Tissue Sarcoma SPE Solid Phase Extraction

T Total Count TLC Thin Layer Chromatography TMB 3,3 ,5,5' tetramethylbenzidine

X X V ll CHAPTER ONE INTRODUCTION

Chapter 1 1.1 GENERAL INTRODUCTION Pesticides are a group of chemical compounds which have contributed to an overall improvement in food production. The effectiveness and potential economic benefits of pesticides have led to their widespread use in controlling agricultural pests and disease vectors. When man first started to deliberately grow plants for food, he soon learned that production was greatest when crop species were grown alone, free from competition. Thus, monoculture was developed and the concept of weeds as unwanted plants was bom. As weed and pest infestations began to seriously limit the production of crops, methods were devised to combat them. The success of the operation depended upon three measures: a) The rotation of crops. b) The removal of weed seed. c) The disturbance of soil by cultivation. Until the advent of chemical pesticides these were central features of the art of weed control.

1.2 HISTORY The early part of the 20th century saw the introduction of the first forms of biological control. Cages of poorly fed hens were dragged out into pest-infected fields and set free in the most affected areas. The object was simply that the hungry bird would effectively control the pest insect. Organic solvents such as chloroform and carbon tetrachloride were also used at this time for the control of soil-based pests (Thomson and Abbott, 1966). The first selective herbicide in British agriculture was copper sulphate, first used in1898, to control broad-leaved weeds in cereal crops (Hassall, 1990). Other mineral salts (ferrous sulphate, sodium chlorate) were also used and later sulphuric acid. During the late 1930s the search for chemicals which would regulate the growth of plapts was undertaken. This led to the development of organic herbicides i.e. MCPA (2-methyl-4-chlorophenoxy acetic acid) and 2,4-D (2,4-dichlorophenoxy acetic acid) (Lockhart et al, 1982). Experimental work also started in Britain in 1941 on the development of DNOC (di-nitro-o-cresylate) which had been used in France since 1933. Throughout the 1950s and 1960s both the range of pesticides and their use e^qpanded steadily. The decade of the

Chapter 1 2 1970s brought increased awareness of the growing danger posed by undesirable plants (Hassall, 1990).

1.3 PESTICIDES. CHEMICALS AND THE ENVIRONMENT Many thousands of chemicals are produced in everyday life both industrially and domestically, and are often present at extremely low concentrations. Most known substances have no known effect on health and the environment. Others are moderately hazardous, a number are dangerous and a few so potentially harmful that they require highly controlled handling and storage. The role of pesticides in agriculture has been to protect crops and to increase production in order to feed the burgeoning world population. The pesticide stoiy is slowly changing from one of an environmental benefit to one of long term environmental concern. Contamination of the environment by pesticides can occur in a number of different ways e.g. air, soil, food or water. Water pollution by pesticides is contributing on a larger scale to environmental contamination than any other means (Jones, 1992). Pollution of the water supply system by pesticides can occur in various ways. These include: careless over spraying of water courses, through drains or standing water, run-off from sprayed or treated areas, careless disposal of containers, the illegal disposal of waste pesticides in farm soakaways, washing down of contaminated equipment spills and leaching from soil in treated areas (The BMA, 1992). Factory discharges are normally kept to a minim um and can be carefully controlled and accounted for.

1 .3 .1 Water contamination by pesticides 1.3.1.1 Process of dissipation Although pesticides have proved very useful, their steadily increasing use has led to concerns over their appearance in drinking water. Pesticides enter water supplies by a process of dissipation from the soil. As some pesticides are persistent in soil, their duration can be determined and so are an important factor in the movement of residues into water supplies. Pesticides may be lost from the soil either by physical removal of the unchanged molecule or by degradation (Walker et al, 1982).

Chapter 1 Physical removal The ways in which physical removal may occur are by volatihsation, leaching and uptake by plants. D Adsorption This is a process in which interaction occurs between the adsorbent surface and the molecules or ions of the adsorbate causing the adsorbate to be attracted to the surface and reducing its concentration in solution. An important factor to determine the extent of adsorption of pesticides by the soil is the amount or type of the colloidal constituents (organic matter and clay present). Many clay particles resemble the negative radical of a weak acid (e.g. COO" in acetic acid), and thus the negative clay particle attracts to its surface positive ions such as hydrogen and calcium. II) Volatilisation In theory even such involatile materials as the ureas could be lost significantly in this way but in practice movement of the pesticide into the soil and adsorption reduce such losses. Volatilisation may also be reduced by using granular rather than liquid formulations. Vapour losses after spraying a wet soil are greater than firom a dry soil because of the competition between water and herbicide for adsorption sites, so that as the water content increases there is a reduction in the amount of pesticide which is unable to volatilise due to its adsorption. Evaporation usually increases with increasing temperature. However, if the rise in temperature produces a significant loss of soil water, more adsorption sites will be made available to the pesticide so that it is possible, under some circumstances, to demonstrate a decrease in the vapour loss of a pesticide with increasing temperature. m) Leaching Leaching is the downward movement of a substance by water through the soil. The movement of a pesticide by leaching may determine its effectiveness as a pesticide. Leaching is very important because it is the process whereby water supplies are contaminated. Rain leaches the pesticide into the upper soil layer. Weed seeds germinating in the presence of the herbicide are killed. Large seeded crops such as com, cotton and peanuts planted below the area of high pesticide concentration may not be injured. The extent to which a pesticide is leached is determined principally by: a) The amount of rain passing downward through the soil.

Chapter 1 4 b) The solubility of the pesticide. c) The adsorptive relationship between the pesticide and the soil. In general, those pesticides that leach are rapidly degraded and those that are slowly degraded do not leach. However, the greatest losses from soil to drainage water occurs after heavy rains. TV) Uptake bv plants Some proportion of pesticide will be removed from the soil by the roots of both tolerant and susceptible plants. Little is known of the extent of this process although there are a few examples, such as the case of maize which is able to absorb and metabolise considerable quantities of simazine (Walker et al, 1982).

Degradation Decomposition is the major route for pesticide disappearance from soil. Besides, degradation products are usually less toxic than the parent compound but they may have greater persistence and other kinds of biological activity. Degradation is possible in three ways, photochemically, chemically or enzymatically, whereby different steps may be performed by different mechanisms. I) Photochemical decomposition Photodecomposition has been reported for many pesticides (Lockhart et al, 1982). This process beguas when the pesticide molecule absorbs light energy; this causes excitation of the electrons and may result in breakage or formation of chemical bonds. A number of pesticides have been shown to be decomposed when exposed to ultraviolet light, although usually the m axim um rate of breakdown occurs at wavelengths rather shorter than those which reach the soil from the sun. TTl Chemical decomposition Many pesticides contain chemical groupings which are susceptible to hydrolysis but in aqueous solution at pH values simila r to those in soils, such hydrolyses would be slow. In soil, however, many types of surface are present and there are many different species in solution so catalysis may well occur. Furthermore, oxidation/reduction reactions are an important feature of soil chemistry. Oxidation by atmospheric oxygen dissolved in soil water is a probable mode of degradation of some pesticides, and reduction is also possible under anaerobic conditions.

Chapter 1 in) Microbial decomposition Microbial breakdown of pesticides appears to fall into one of two categories. In the first, degradation proceeds at a steady rate, approximately proportional to the concentration of chemical in the soil, with no lag phase. An explanation of this behaviour is that microbial enzymes degrading natural substrates in the soil carry out similar transformations on structurally related molecules, a process known as co-metabolism. Simazine, diuron, triazines and other ureas fall into this group (such compounds are usually of long persistence). In the second category, slow degradation by the non-specific mechanism is followed by a period of rapid breakdown. The e^qplanation of this is that microbial enzymes adapt to the new substrate during the initial lag phase. Such behaviour has been found for many herbicides of short persistence in soil and has been extensively studied for the phenoxyalkanoic acid herbicides.

1.3.1.2 The Extent of Water Contamination The actual scope of water contamination by pesticides (other toxic components of pesticide formulations, breakdown products or toxic substances formed during water treatment) is very difficult to quantify because of a wide range of influential parameters (i.e. sampling frequency, site of selection, sampling technique, availability of analytical techniques, nature of pesticide). However, it is evident that this is a serious and widespread global problem (Lees and McVeigh, 1988). Relatively little is known about the range and concentration of pesticides present in drinking water at very low levels that would be associated with long term exposure. Besides current analytical techniques are often not sensitive enough to measure such low levels. The worst case of contamination is probably that of ground water since most compounds when they have reached the ground water are unlikely to degrade significantly and will therefore, without treatment, be present continuously in drinking water. Ground water provides ~ 30% of drinking water in England and Wales and up to 70% of supplies in Eastern and Southern England. It also contributes to the "base flow" of rivers which are the source of ~ 70% of England and Wales drinking water (Lees and McVeigh, 1988). The persistence and mobility of pesticide compounds are amplified in ground water systems according to the British Geological Survey 1987 (The BMA, 1992).

Chapter 1 There are a range of water treatments that can be applied to remove pesticides from drinking water. These include: a) Simple filtration which generally removes only a small proportion of any pesticide. b) Coagulation procedures which remove a higher proportion of certain pesticides (organophosphorus compounds). c) Activated carbon and granular activated carbon (GAC) which has been reported to remove 99% of certain pesticides under favourable conditions. However, the majority of water supplies in the UK rely mainly on simple filtration and coagulation. Activated carbon is used in some cases but not as a routine treatment of drinking water. This perhaps potentiates the extent of water contamination. Between June 1985-July 1987 a survey of pesticide pollution in drinking water in England and Wales was carried out by Friends of the Earth (Lees and McVeigh, 1988). It reported a number of breaches of the European Community (EC) Drinking Water Directive; incidences where the maximum admissible concentration (MAC) for single pesticide of 0.1/^g/L was exceeded. The pesticide contamination of various water sources (ground, surface, reservoir and a blend of six of the present existing water regions) in UK were reported upon. Table 1-1 shows 16 pesticides detected in water supply areas/sources above the MAC for a single pesticide 0.1/ig/L. The incident levels of several pesticides, notably atrazine and simazine, have posed not only a problem in the UK but also in other countries in Western Europe including France, West Germany, Norway and Italy (Lees and McVeigh, 1988). The potential health hazards of pesticides whether due to chronic (long term) or acute exposures are not well documented especially in many cases of pesticides introduced several years ago and also there are no safety records available. Therefore, the incidence of water contamination levels by pesticides is of growing concern. The inadequate safety tests and monitoring of water need to be improved.

Chapter 1 Table 1-1 Pesticide breaches (Lees and McVeigh, 1988) in England during the period June 1985-July 1987.

Pesticide No records Overall mean Overall range fig/L fig/L

Atrazine 166 0.36 0.11-4.5 Simazine 96 0.34 0.11-1.97 Mecoprop 47 0.35 0.12-1.01 MCPA 11 0.37 0.16-0.84 MCPB 9 0.43 0.12-0.41 Dimethoate 6 0.24 0.12-0.6 2,4-D 5 0.36 0.11-0.56 2,4-5-T 5 0.39 0.11-0.5 Dicamba 3 0.21 0.12-0.3 2,3,6-TBA 4 0.73 0.12-0.28 Chlortoluron 2 0.48 0.18-1.4 Propazine 9 0.22 0.44-0.58

1 .3 .2 Soil contamination by pesticides There is also a risk to human health from chemicals ingested in food grown on contaminated soil and to a lesser degree through accidental ingestion of the soil itself and indirect absorption following contact between soil and the skin. Those chemicals also adversely affect natural habitats. It is more likely that some pesticides will be eaten with the plant than absorbed from the drinking water. There are three main pathways by which a pesticide in the soil can enter a plant (Topp, 1986). These are: a) Root uptake and subsequent upward movement in the plant. b) Uptake of vapour from the surrounding air. c) Uptake by external contamination of shoots by soil and dust, followed by retention in the plant cuticle or penetration through it. The dichotomy between the benefits of pesticides and the possible risks to human health from their presence in soil serves to remind us of the difficult balance

Chapter 1 8 between risk and benefit which is central to any discussion of chemicals in the environment.

1.4 PESTICIDES; THE ENVIRONMENTAL LEGACY The laws covering the manufacture and use of pesticides can be confusing to both those required to comply with them and those seeking to use them to improve standards. In the UK the responsibility for administering and enforcing the legislation is split between two different Government agencies: the Health and Safety Executive (HSE) and the Ministry of Agriculture Fisheries and Food (MAFF). Figure 1-1 illustrates the six Government departments which are involved in the process of approving pesticides for use.

COMPANY > \ p a t a Scientific Sub -committee

V Evaluation MAFF Pesticide Safety Division Advisory Committee (Health and Safety Executive » on Pesticides for wood preservatives) D ecision A dvice <------Ministry of Agriculture Food and Fisheries Department of Health

Department of Environment

Department of Employment

Scottish Office

Welsh Office

Figure 1-1 Six departments involved in pesticides approval for use.

The Health and Safety at Work etc. Act 1974 is the main piece of health and safety legislation in the UK. Specific regulations under the 1974 Act are relevant to pesticide health and safety (The BMA, 1992). These are as follows: a) Control of substances hazardous to health regulations (COSHH) 1988.

Chapter 1 b) Health and safety-control of industrial major accident hazard regulation 1984. c) Health and safety-classification, packaging and labelling of dangerous substances regulations 1984. d) Health and safety-road traffic (carriage of dangerous substances in packages etc.) regulations 1986. In the UK, two voluntary schemes were used until 1986. The toxicity and possible hazards of pesticides were controlled by Pesticides Safety Precautions Scheme (PSPS) whereas the effectiveness and appropriate uses of products were the concern of the Agricultural Chemicals Approval Scheme (ACAS). These were replaced by the Control of Pesticides Regulations 1986. The major provision is that no pesticide may be used commercially unless it has been given provisional or full approval on the grounds of safety and efficacy by Ministers. The legal limit for individual pesticides in drinking water in Europe is the EC Maximum Admissible Concentration (MAC) of 0.1/tg/L (ppb) and total pesticide concentrations should not exceed 0.5/^g/L. For some pesticides these concentrations are many times less than was allowed under previous legislation. For example, the previous UK Government guideline concentration for atrazine was 30/ig/L. The MAC is not necessarily a toxicity indicator but a limit set to protect water supplies from pollution. The World Health Organisation (WHO) adopts a different approach. It considers the toxicology of individual substances and recommends a guideline value for each substance based on the assumption of lifelong consumption. The British Government has asked the European Commission to review the pesticides standards with the aim of securing different standards for mdividual pesticides based on toxicology. For the present, the EC standards are embodied in the regulations which water companies are required to meet (Jones, 1992). Guideline concentrations for pesticides in drinking water have been issued by the UK Government. The values quoted vaiy between chemicals and attempt to take account of known toxicity data. For example, guideline concentrations for atrazine, simazine, chlortoluron and MCPA are 30, 30, 8 and 10/^g/L respectively. The World Health Organisation has made an independent assessment of some of the pesticides currently in use and has set guideline

Chapter 1 10 concentrations for drinking water of 2, 17 and 0.5^g/L for atrazine, simazine and MCPA respectively. Several pesticides have been detected at concentrations above the MAC in water sources in the UK as well as elsewhere in Europe. According to the survey carried out by Lees and McVeigh (1988) many incidents of pesticide contamination of various water sources in six out of ten UK water regions were reported. The survey did not quote the number of analyses which took place which were below the MAC. Nevertheless, the apparently high incidence of water contamination by pesticides is of great concern and improved monitoring programmes to assess water quality are required to meet EC regulations and to allay public concern.

1.5 TOXICITY OF PESTICIDES Poisoning throughout the world by pesticides are thought to exceed one million annually (Friends of the Earth, 1993). A report in 1986 from the World Health Organisation estimated that the average mortality world wide may be -2800 death per annum and those working in agriculture run a higher risk than almost any other group in society (The BMA, 1992). The adverse effects of a particular pesticide wiU depend upon its intrinsic biological properties (toxicity), the route of exposure (inhalation, ingestion or skin exposure), level and duration of exposure, any co-exposure to other chemicals included in its formulation and the individual sensitivities of those exposed to it. There are two types of exposure which need to be considered: short term exposure (acute) and long term exposure (chronic).

1 .5 .1 Short term health effects Acute reactions usually occur while the chemical is being used or shortly afterwards. Most acute reactions last only a short time and the majority of victims recover completely without long-term comphcations. One of the most detailed analyses of the extent of acute pesticide poisoning has been carried out by the Regional Poisoning Treatment Centre at the Royal Infirmary in Edinburgh (Proudfoot, 1988). All admissions following acute exposures to pesticides over a 6 year period (1981-1986) were identified. In the 6 years under review there were over 9000 acute poisoning admissions to the centre, fifty seven of which followed alleged exposure to pesticides (0.6% of the total); thirty eigjit of the a dm issio n s

Chapter 1 11 were due to parasuicide (suicide attempts or self harm). The majority of patients had no symptoms or relatively minor and short Uved ones, such as nausea, vomiting, abdominal pain, diarrhoea, coughing and breathlessness.

1 .5 .2 Long term hecdth ejffects Given today's extensive use of pesticides, both for agricultural and non- agricultural purposes, it is almost impossible for any member of the population to avoid daily exposure to very low levels of several different pesticides in food and water. Consequently, there is concern about possible adverse effects on human health arising from continual long term low level exposure; that is the potential for chronic toxicity. Chronic effects are only likely to become apparent after prolonged exposure to a chemical. In the case of cancer this may be a period of several decades which leads to difficulties in identifying the causative agent. Pesticides that produce a "unique" disease in man are most likely to be identified, particularly if they produce a similar effect in animal species. However, if a pesticide increased the incidence of a very common human disease, it might be almost impossible to identify the cause (The BMA, 1992). Three major areas have been subjects of study into the possible effects of long term pesticide exposure: cancer, reproductive hazards and neurological disorders. The carcinogenicity of pesticides have been consistently investigated by the studies of farming occupations. These have been well summarised in probably the most extensive review of chronic health effects of pesticides by Sharp and colleagues (The BMA, 1992) and are reproduced in Table 1-2. The soil fumigant 1,3-dichloropropene has been associated with several case reports of leukaemia (Markovitz and Crosby, 1984) while exposure to triazine (Donna, 1989), chlordane (Wang and MacMohn, 1979), heptachlor (MacMahon, 1988) and other organochlorine (Morgan, 1980) pesticides do not appear to be carcinogenic in man. Exposure to phenoxy herbicides (2,4-D, 2,4,5- T, MCPA and contaminants) and chlorophenols constitute the majority of studies of defined exposure to either applicators or farmers. Results show an association between exposure to these compounds and Hodgkin's disease (Riihimaki, 1982), leukaemia and soft tissue sarcomas (Hoar-Zahm, 1988). Skin effects are an occupational hazard for agricultural workers (Mathias and Morrison, 1988) and many specific compounds have been implicated; for example: anilazine (Schuman

Chapter 1 12 and Dodson, 1985) and deltameüirin (Wang, 1988). In addition, skin diseases appear to be the main cause of adverse health effects in appUcators of agrochemical.

Table 1-2 Summary of studies investigating possible links between cancer and pesticide use among farmers.

Region Outcome Type of activity

California Leukaemia Fanning Hodgkin's lymphoma Farming Multiple myeloma Farming W ashington Leukaemia Poultry Multiple myeloma Farming South-east USA Multiple myeloma Poultry Uterine cervix Farming Nebraska Leukaemia Com, insecticide, poultry Texas Leukaemia Farming Iowa Leukaemia Farming Lymphoma Farming Multiple myeloma Farming Prostatic cancer Farming Stomach cancer Farming Wisconsin Lymphosarcoma None noted Reticulum cell sarcoma Cattle, dairy, small grains Britain Soft-tissue sarcoma Farming New Zealand Lymphoma Farming Multiple myeloma Farming

Studies of the effects of pesticides on fertility and fetal development are in principle much simpler to conduct than investigations into cancer, because the at-risk population is more readily identified and the period of exposure is much shorter. Dibromochloroproprane (DBCP) is the pesticide most clearly implicated

Chapter 1 13 in the impairment of fertility. Studies by Whorton and colleagues (Whorton et al, 1977) showed that almost half the workers had lower than normal sperm counts compared with less than 10% in non-exposed workers. Subsequent studies have confirmed that the degree and length of exposure to DBCP is directly related to reduced sperm count (The BMA, 1992). Organochlorine pesticides such as DDT pass through the placenta with an average level in the newborn blood reaching around a third of that in maternal blood. The problem is most evident in the third world where exposure to DDT continues. Studies in Brazil (Procianoy, 1981) and India (Saxena, 1981) have shown DDT levels significantly higher in the cord blood of pre-term than in-term infants. Finally, the nervous system has been recognised as a target organ for pesticide toxicity for several decades. Organophosphates may cause two delayed effects from acute and chronic exposure: delayed polyneuropathy and neurobehavioural effects (The BMA, 1992). Moreover, organophosphates depress cholinesterase activity in a dose-dependent relationship (Ames, 1989) but do not appear to cause chronic neurotoxicity (Misra, 1988). However, the effects of pesticides on the nervous system can be difficult to detect (Savage, 1988).

1.6 PESTICIDES CLASSIFICATION Pesticides are most commonly chemicals and their classification is based on the organisms that are attacked i.e. chemicals that kill weeds are called herbicides, chemicals that prevent the growth of fungal diseases are known as fungicides and chemicals that kill insects are called insecticides.

1.6.1 Herbicides Active ingredients with herbicidal properties belong to many chemical families and have been formulated in numerous ways. Thus, in a list of products approved in the UK, the entries for herbicides occupy more space than those for insecticides and fungicides combined (The BMA, 1992). It is therefore necessary to divide herbicides into manageable sub-groups and one way to do so is shown in Table 1-3.

Chapter 1 14 Table 1-3 ClEissification of herbicides (adapted from Hassall, 1990).

Group 1 Applied to foliage

TypeA Kill all foliage unless directionally sprayed TypeB Kill broad-leaf weeds in cereals, grass TypeC KiU grasses (sometimes in a broad-leaf crop) TypeD KiU grasses in cereals TypeE Kill broad-leaf weeds in various dicotyledonous crops

Group 2 Foliar and soil action on young weeds

TypeA Inhibitors of photosynthesis TypeB Herbicides that affect cell division Type C Substances that disrupt membrane structure or function

Group 3 Soil-acting (often soil incorporated)

TYPEA Substances that disrupt fatty acid metabolism TypeB Herbicides that affect meristematic growth

The initial division relates to overall agricultural function (foliar or soil application, control of seeding or of established weeds). Further subdivision utilises probable mode of action or (for leaf-applied substances) types of plants that are selectively killed or selectively protected. Group 1 foliar herbicides are usually applied to established weeds, sometimes to clear ground for crops or more often to kill established weeds in a standing crop. Group 2 herbicides usually have a quite different function, since they are mainly used against weeds that have just emerged or are e^qiected to emerge shortly after application. The stage of growth of the crop at the time of application may vary; it may be an established crop, like trees in an orchard. The soil-acting herbicides of group 3 have yet other functions, for example they may be applied long before a crop is sown or planted in order to eradicate subterranean growth of troublesome

Chapter 1 15 perennial weeds, a task that is often difficult to achieve in the presence of an annual crop (Hassall, 1990).

1.6.2 Fungicides Many well known fungicides are compounds of sulphur, copper or mercury. Fungicides such as calcium hydroxide, mercuric chloride or mercurous chloride are also effective fungicides, but are poisonous to man. Synthetic organic compounds such as dithiocarbamates have also been developed.

1.6.3 Insecticides Some insecticides kill by stomach actions, i.e. they poison the insect when eaten (arsenic, fluoride). Other insecticides are used in the gaseous form, they are called fumigants. Organic compounds derived from plants include the nicotine alkaloids, rotenone and pyrethum; the latter being the one still used to an extent (Lockhart, 1982). Newer synthetic insecticides i.e. DDT and malathion have replaced the older types of insecticides, although DDT has been banned firom use in the developed countries since 1974 (The BMA, 1992).

1.7 MODE OF ACTION OF HERBICIDES The ultimate aim of a herbicide application is to kill unwanted plants and this may be achieved by inhibition or disruption of a variety of processes:

1.7.1 Inhibition of respiration Herbicidal action on respiratoiy metabolism may involve interference with the degradation of glucose to pyruvate (glycolysis), the oxidation of various organic acids (Kreb s cycle), the electron transport system or the mechanism by which oxidation is coupled to the phosphorylation of ADP to form ATP in the mitochondria (Sagar et al, 1982). Inhibitors of oxidative phosphorylation include the carbamates and phenylureas.

1.7.2 Inhibition of photosynthesis Many of the herbicides which affect ATP formation also affect the light reactions of photosynthesis. Recent studies have allowed the classification of herbicidal inhibitors of the photochemically-induced reactions into the classes (Moreland, 1980): electron inhibitors (e.g. triazines), uncouplers (e.g. perfluidone), energy-transfer inhibitors (e.g. 1,2,3-thiadiazolyl-phenyl-ureas).

Chapter 1 16 inhibitory uncouplers (e.g. dinitrophenols) and electron acceptors (e.g. certain bipyridyliums). It was reported that the natural photosynthetic phosphorylation as a cyclic process produces ATP at the expense of light. The urea herbicides block the normal pathway by which electrons in chlorophyll are replenished; the chloroplasts are oxidised and they become bleached and the leaf is said to become chlorotic whereby it soon dies (Ashton and Crafts, 1973).

1 .7 .3 Intetference with protein synthesis Various stages in protein synthesis are also sites of action for several groups of herbicides including the benzoic acids, dinitroanihnes, phenoxy-acids and triazines. The action of dinitroanihnes on protein synthesis, however, may be indirect, resulting from interference with the action, development and transport of hormones. Again, many of the herbicides which affect protein synthesis also inhibit ATP synthesis by oxidative or photophosphorylation. Consequently, it may be difficult to distinguish direct and indirect effects on protein synthesis (Sagar et al, 1982).

1 .7 .4 Interference with lipid synthesis Certain herbicides are known to interfere with the synthesis of hpids which are important constituents of cell and organelle membranes. Among the herbicides which inhibit lipid synthesis are dichlobenH and pentachlorophenol. Low concentrations of certain growth-regulator herbicides may stimulate hpid formation (i.e. 2,4-D) (Sagar et al, 1982).

1.8 DETERMINATION OF HERBICIDE RESIDUES 1 .8 .1 Stcuidard methods Until the 1950s bioassays, non specific methods (such as total chlorine) and colorimetric assays were used for pesticide analysis (Middelem, 1971). The limit of detection of these assays in water were in the region of 0.5mg/L and although some colorimetric assays have been employed more recently, they have been largely replaced by more sophisticated techniques (Handa, 1988). Thin layer chromatography (TLC) was useful as a qualitative screening method and was the first multiresidue assay system for the detection of pesticides of similar structures and their metabolites. Despite the current re surgence of interest in TLC, relatively few papers were published concerning pesticide analysis by TLC (Sherma, 1986). However, analyses are now almost

Chapter 1 17 exclusively performed by gas chromatography (GC) and high performance liquid chromatography (HPLC). Current methodologies for the measurement of pesticides have recently been reviewed (Sherma, 1989). Pesticide analysis by GC has employed various types of columns and detectors. Packed column phases have been described for the separation of all classes of pesticides and metabolites and the electron capture, and flame detectors provide for their sensitive and selective detection and quantification (Lee and Chan, 1983). Organochlorine compounds for example are frequently analysed using electron capture detectors (Neicheva, 1988) which are also suitable for multiresidue assays (Lee and Chan, 1983). Electrolyte conductivity detectors were used in the analysis of triazines demonstrating a detection limit of 50-100 /(g/L (Purkayastha et al, 1973). GC/mass spectrometry has been employed to detect numerous pesticides such as atrazine and diazinon in water and soil (Lopez-Avüa et al, 1985). Despite detection limits of 0.1-1.0 /ig/L being achieved, stable-labelled isotopes were required to spike each sample prior to extraction (water samples were extracted with methylene chloride; soü samples were extracted with acetone/hexane). As a result, GC has some disadvantages, primarily its use is restricted to volatile compounds only, and it requires time- consuming sample preparation steps. The limitation of GC (volatility requirements) has been overcome by the use of HPLC. HPLC has been successfully used on pesticides for a number of years (Pommery et al, 1990). Fluorometric, UV and electrochemical detectors have been used for quantitation of pesticides. The choice of stationary phase depends on the class of compound being measured and include reversed-phase systems and bonded phases (CN and NH 2) (Lawrence and Turton, 1978) (Osselton and Snelling, 1986). Many HPLC methods in the literature have been reported for measuring pesticide compounds, for example the determination of phenylurea pesticides by HPLC with UV and photoconductivity detectors, showing a detection limit of 500 /ig/L, (Walters et al, 1984). Also, determination of phenylurea herbicides by HPLC with electrochemical detection showed a detection limit of 50/ig/L (Chiavari and Bergamini, 1985). Some researchers have determined pesticide levels in drinking water with a limit of detection varied between 0.5-0.1 fig/L. However, samples required solid phase extraction prior to HPLC analysis (Balinova, 1993). Although, HPLC is a powerful technique which avoids some of

Chapter 1 18 the disadvantages encountered with GC, it still has some major limitations which include the lack of sensitive and specific detectors and difficulties in the direct linking with mass spectrometry. Ultraviolet spectrophotometry (UV) has been utilised to a much lesser extent than other techniques, such as GC and HPLC, because of its lower sensitivity and selectivity. However, it has been claimed (Singh, 1989) that it is a very precise and reproducible method for the quantification of dinitroaniline herbicides (Traore and Aaron, 1989). In spite of its moderate sensitivity (limits of detection ranging firom 1 to 7 mg/L), this technique presents the advantage of being inexpensive and easily interfaceable with HPLC and TLC. Moreover, fluorescence spectrometry has been combined with TLC for the qualitative or quantitative analysis of organophosphates and carbamates pesticides and with HPLC for the analysis of herbicides and fungicides (Sherma, 1989, 1991, 1993). Fluorometric HPLC (or TLC) detectors are extremely sensitive (ng/ml range) and selective, but the number of fluorescent pesticides is relatively small. Chromatographic techniques for pesticide determination are not ideally suitable for large scale monitoring programmes. Although, solid phase extractions are now being increasingly used, large sample volumes and extensive extraction and clean up procedures are generally necessary prior to analysis Increasing the cost and time involved for each assay (Junk and Richard, 1988). An example for the application of solid phase extraction is the organochlorine pesticide analysis of water which were carried out at the Robens Institute (Prapamontol, 1991). The method required an initial rinse of the Cg-cartridges in methanol using a vacuum followed by the addition of 100ml samples of water onto the small cartridges which pulled through at 20-25ml per minute. Then, the cartridges were briefly rinsed with water, dried and eluted with ethyl acetate for direct injection onto the GC whereby a detection limit in the range of 0.1 /fg/L was achieved (Stevenson, 1991). Furthermore, the equipment required for chromatographic analyses is relatively expensive and therefore often restricted to specialised laboratories. These disadvantages can be offset by the fact that several compounds can often be analysed at one time. Also, the use of mass- spectrometry allows clear identification of analytes in a complex matrix. The fact that many pesticides inhibit enzymes has led to the introduction of biological techniques for the detection and estimation of these various

Chapter 1 19 compounds. For example, organophosphorous pesticides have the ability to inhibit acetylcholinesterase (AChE) as illustrated in the reaction below: AChE acetylcholine —> choline + acetic acid Therefore, the immobilisation of a cholinesterase on a pH electrode gives rise to a biosensor sensitive to pesticide residues since these compounds inhibit the activity of the enzyme on its substrate. This technology may allow continuous or semi-continuous monitoring of water supplies. Pesticide residue analysis has been a fundamental part of safe and effective product development, but there is now additional pressure to provide more information about the fate of pesticides in the environment. Frequent monitoring for the presence of pesticides in food, drinking water etc. is required in order to meet the needs of present legislation relating to the use and levels of pesticides. The use of chromatographic techniques for pesticide residue analysis are not entirely suited to the task of rapid and frequent monitoring of a large number of samples (screening purpose). Thus, less time-consuming, cheaper methods of analyses are required to complement existing techniques. Immunoassays have the potential to improve monitoring programmes by providing assays which are inexpensive, rapid and simple to perform.

1 .8 .2 Immunoassays 1.8.2.1 Immunological principles Immunochemical techniques based on the use of specific antibodies provide a convenient, cost-effective option for the purpose of monitoring pesticides in different matrices. In time, semi-quantitative extra-laboratoiy tests utilising solid phase antibodies should become a reality for a number of pesticides allowing the rapid identification of samples breaching MAC levels for instance. Immunoassays were first described some 30 years ago (Yalow and Berson, I960) (Ekins, 1960) and were rapidly exploited by clinical chemists for the sensitive and specific analysis of hormones, proteins, peptides and drugs. Further developments have resulted in the commercial availability of diagnostic reagents and kits for a wide range of analytes as well as tests suitable for use by semi trained and untrained personnel, e.g. for confirming pregnancy and predicting ovulation. Other scientific fields, e.g. forensic science, food science and

Chapter 1 20 microbiology, have now also widely adopted immunoassays and their potential in the field of environmental analysis is rapidly gaining recognition. Immunoassays are based on the principle of competitive binding or saturation analysis (Walker, 1977) (Ekins, 1991). The analyte (Ag) in samples or in standards competes with a constant amount of a labelled form of the analyte (Ag*) commonly called the label or tracer, for a limited number of antibody binding sites (Ab) specific for the analyte as illustrated in Figure 1-2. At the end of a period of incubation. I.e. when equilibrium is reached (since the reaction obeys the Laws of Mass Action), the antibody bound fraction and the unbound or "free" fraction are separated from each other in a procedure known as "phase separation". The amount of the label associated with either the bound or free fraction is then measured.

V' V' BOUND FREE

Figure 1-2 The principles of competitive immunoassays.

The distribution of the labelled antigen between the bound and free fraction can then be related to the amount of antigen present. Thus, standard curves can be constructed and unknown concentrations of antigen determined. This type of immunoassay is known as a competitive or limited reagent assay in that there is a limited amount of antibody present. Non-competitive or reagent

Chapter 1 21 excess assays (immunometric assays) are also commonly used (Jackson and Ekins, 1986) although these are generally less suitable for small molecular weight compounds such as pesticides. Two key reagents are therefore required to set up an immunoassay: an antiserum with the desired specificity towards the analyte and an immunoreactive labelled form of that analyte. A suitable phase separation technique must also be found.

1.8.2.2 Antisera The main requirement for the development of an immunoassay is that a specific antiserum to the analyte is available. However, if the antiserum is not commercially available, antisera can be produced in a number of laboratory animal species using a suitable immunisation schedule. Antisera to small molecules (haptens*. MW < 5000 daltons) such as pesticides, which are not normally immunogenic, are produced in response to the injection of an immunogenic conjugate. This is prepared by covalently linking the hapten to a carrier protein such as bovine serum albumin, ovalbumin or thyroglobulin. Conjugation usually takes place through the lysine or carboj^lic acid residues of the protein. The chemical reaction may be achieved by a number of methods, notably the mixed anhydride reaction (Erlanger et al, 1957), carboditmide condensation reactions (Kurzer and Douraghi-Zadeh, 1967) or by the use of hydro^grsuccinimide esters (Anderson et al, 1964). The use of other bifunctional reagents, such as glu tar aldehyde (Reichlin et al, 1968) and N-(m- maleimidobenzoyloxy)-succinimide (MBS) (Kitigawa et al, 1981), have also been described. The type of conjugation reaction used depends mainly upon the functional groups available on the hapten. Some haptens contain reactive groups such as carboxyl, amino or hydroxyl groups through which direct conjugation can take place, but often a more appropriate analogue or derivative of the compound has to be used. The specificity of a resulting antiserum will greatly depend upon the point of attachment of the protein to the hapten, since antibodies to haptens are directed primarily to determinants furthest away from the site of conjugation (Landsteiner, 1945). Thus, the choice of conjugate requires careful consideration. Following the production of a hapten-protein conjugate purification by dialysis or gel filtration may be necessary to remove any unreacted hapten. The

Chapter 1 22 conjugate is then prepared for immunisation by émulsification with an adjuvant containing inactivated bacteria (Bacillus Calm ette-G uerin [BCG]). W hen the animal receives this mixture as the primary immunisation, this adjuvant serves to sensitise the immune system and to ensure that the immunogen is released slowly. BCG is absent from any subsequent booster immunisation which are administered at appropriate intervals. However, once an immune response has been identified, it appears to be advantageous to leave the animal for a period of 5-6 months to allow the antibody titre to fall down before any boosting takes place (Aheme and Marks, 1979). Each bleed obtained from the animal requires extensive characterisation by measurement of antibody titre, avidity and specificity. Titre is an indication of the concentration of antibodies present and is defined as the dilution of the antiserum that achieves 50% of the maximum binding of a fixed amount of the labelled antigen. Furthermore, the avidity of the antibody is the energy with which it binds to its specific antigen and it is equivalent to the association constant (K) as shown in Figure 1-2. Also it is important to determine the specificity of the antiserum i.e. the extent to which the binding is not influenced by compounds of similar structure. Therefore, antibodies for use in immunoassays should possesshigh quality and show the desired specificity. Antisera which are produced by active immunisation are polyclonal in nature and a mixture of antibodies with heterogeneous specificity and avidity. In contrast, monoclonal antibodies produced by hybridoma techniques are homogeneous as well as being available in large and theoretically unlimited amounts (Edwards, 1985). Monoclonal antibodies are of enormous value in the analysis of complex systems and as diagnostic tools, but their use in analytical immunological assays is still controversial (Kinders and Hass, 1990).

1.8.2.3 Labels (Tracers) a) Radioimmunoassay Traditionally and until recently, the most commonly used type of immunoassay has been radioimmunoassy (RIA) where the antigen is labelled with a radioactive label. The sensitivity of a RIA depends to a large extent on the type and amount of label used (Ekins, 1985). Small molecules such as drugs and steroid hormones can often be radiolabelled to high specific activity using ^H or and are in many ways ideal labels. The incorporation of tritium into the

Chapter 1 23 molecule causes minimal structural change and therefore little change in immunoreactivity. For some antigens it has not been possible to obtain high specific activity labels using 8-emitting isotopes and radioiodinated antigens have been used instead.

Counting techniques for qj-q considerably easier and cheaper than scintillation counting but there are major disadvantages associated with the use of Because of the short half-life of (60 days) labelled antigens have to be prepared frequently. Radioiodination techniques involve the use of relatively la r^ amounts of radioiodine and are, therefore, a potential health hazard. Although, the use of radioisotopes provide high specific activity labels and offer high sensitivity to the assay procedures, the short shelf life of reagents, expensive counting equipment, waste disposal problems and potential health risks caused a drawback for the use of the technique. Also the incorporation of a large atom into the antigen can markedly alter the immunoreactivity of small molecules such as drugs, which can not be radioiodinated directly and require to be 'tagged' with a tyrosine or histidine containing moiety in which case the immunoreactivity will be altered even further (Nars and Hunter, 1973) (Robinson et al, 1975). Over the last fifteen years considerable effort has been made to find suitable alternatives to the use of radiolabels in immunoassays as illustrated in Table 1-4. Alternative labels should be stable and the endpoint easily quantitated on readily available equipment.

Chapter 1 24 Table 1-4 Labels used in immunoassay.

Type qflabel Examples

Radioisotopes 14c, 3H, 1251 75se, 57Co Enzymes 8 -D-Galactosidase Alkaline phosphatase HRPO Fluorescence Fluorescein Rhodamine UmbeUiferones Lanthanide Chelates Luminescence Luciferase / Luciferins Peroxidase Luminol (and derivatives) Aciidinium esters Miscellaneous Proteins Viruses Co-enzymes

One important aspect of the use of non-isotopic labels is that homogenous (non separation) assays can be developed, and opportunity for the development of single, fast and portable kits for "positive or negative" answers become possible.

b) Enzyme immunoassay Presently, enzyme immunoassays (EIA) are the main alternative to RIA. Enzyme labels are stable and the endpoint, for instance a change in colour, can be monitored on equipment which is available in most laboratories. The sensitivity of EIA can now be as high as many directly comparable RIA techniques and both enzyme labelled antigens and antibodies have been extensively used in simple emyme linked immunosorbent assay (ELISA) (Khanna, 1991). These are usually carried out in a 96-well microtitre plate format and qualitative, semiqualitative or quantitative assays are in every day use in a

Chapter 1 25 variety of applications. There are many different forms of ELISA (Kemeney and Challcombe, 1988) and two formats of ELISA are generally suitable for the analysis of pesticides as shown in Figures 1-3 and 1-4. In both assays one of the components is adsorbed hydrophobically to polystyrene plastic surfaces. This procedure is known as "coating" and is carried out h y simply filling the wells with a solution of the coating material for a period of time. Uncoated material is removed by washing the plate. Coating of proteins is commonly carried out in alkaline buffers at approximately pH 9.6. On the other hand, for some analytes it may be necessary to reduce non-specific binding to the plastic by blocking the uncoated sites with an inert protein such as gelatin, album in or casein. Once the coating has taken place the next stages in the assay can be carried out with further washing steps included to remove any unbound reagents. The times and incubation periods required at each stage have to be optimised for each assay. Typically, incubation times are 1-2 hours at 37°C although overnight incubations at 4°C are also favoured especially for the coating step. At the end of the assay, the enzyme is quantitated by the addition of an appropriate chromagen /substrate which produces a coloured product. The amount of product formed is measured using a spectrophotometric plate reader. For some applications of ELISA a visual estimate of the colour produced is all that is required. Several enzymes have been employed in ELISA, b u t the m ost widely used are horseradish peroxidase (HRPO), alkaline phosphatase and 6-galactosidase. Enzyme conjugates for use as labels in ELISA can be prepared by a number of chemical coupling reactions (Blake and Gould, 1984). The reactions should be carefully chosen to retain both immunoreactivity and enzymatic activity. The chromagen / substrate used at the end of an ELISA will depend on the enzyme used to prepare the labelled component of the assay. Several chromagens are available for HRPO producing different coloured endpoints suitable for visual examination. Many of them have unfbrtuhately been shown to be carcinogenic or mutagenic in test systems. o-Phenylenediamine (OPD) previously one of the commonest substrates, producing a yellow colour which changes into orange when the reaction is stopped by the addition of acid, has been largely replaced by tetramethyl benzidine (TMB) since it is thought to be less hazardous. TMB produces a pale blue colour changing to yellow on the addition of acid. On the

Chapter 1 26 1- Coat solid phase with antibody

Wash

I IMI.IIII 1 IIII 1:11111 Xmurn mi iiiiiii X .nti im , ninm

2- Add no analyte (Bo) Add an a ly te (standard or sample)

uniJiiiifiiiiiiMiiiiULiiiiiiiiinnmiiiiii ......

3- Add enzyme label ^ ^ ^

I III! TfIII! ifii n il HI Iiii'i/ii Tr III! III! Ylilt rtii n il

4- Incubate and wash

5- Add substrate

Colour or Light

Figure 1-3 Schematic diagram showing the stages in a coated antibody ELISA or ECLIA.

Chapter 1 27 1 Coated with antigen {conjugate)

Wash riut jiinnm nin mmii III! iiiHmnmiii i nTiuiiiiii iiii in iiuiiii iiii nil miiiinm

2' Add the related Ab Add the related Ab Bo (no analyte) B (standard or sample)

nil III IIUIIII IIII TTT1 nTTT'imjiu nil III iiti iiiriiii IIII lui iiimn

3' Add enzyme labelled second antibody

I nil iiiTini nil iiin u n n ini nii jlu iin iin I nil nniiii nn in nuiniini nii mi nnim

4- Incubate and wash

5 Add substrate

Colour or Light

Figure 1-4 Schematic diagram showing the stages in a coated antigen ELISA or ECLIA.

Chapter 1 28 other hand, alkaline phosphatase obtained from calf Intestine is more expensive than HRPO, but its conjugates are very stable and its colorimetric and fluorimetrlc substrates are described to be non-hazardous. Para-nitrophenyl phosphate is one of the most widely used chromagen/substrates for alkaline phosphatase and methyl umbelliferryl phosphate is an alkaline phosphatase fluorescent substrate which produces a fluorescent endpoint (Burd et al, 1977).

c) Enhanced luminescence in ELISA The high sensitivity at which chemiluminescent reactions can be measured has provided the opportunity of developing other types of ELISA. However, chemiluminescent signals are usually very short-lived which presents some practical difficulties in measurement. Enhanced chemiluminescent reactions in which an intense, prolonged light signal is produced have now been effectively utilised in ELISA methodology with both HRPO and alkaline phosphatase labels. In the HRPO chemiluminescent reaction luminol is oxidised by hydrogen peroxide in the presence of an enhancer, P-Iodophenol to produce a signal which reaches a peak at 2 minutes and is stable for at least 20 minutes. The light emitted has a maximum absorbance between 400 and 450nm, suitable for most commercially available luminometers (Thorpe et al, 1985a). HRPO enhanced chemiluminescence has been applied to a wide range of assays both as research tools and for routine use. In addition, several luminescent microplate readers are now available commercially and the glow reaction can also be used to measure immunoassay end-points in a semi-quantitative manner by recording the signal onto high speed Polaroid film using a portable camera luminometer (Thorpe et al, 1985b). Enhanced luminescent endpoints have also been developed for alkaline phosphatase reactions (Thorpe et al, 1989). The reaction is based on a 1,2- dioxetane derivative (AMPPD) which has been stabilised with an adamantyl group and also contains a phosphate group. When the phosphate group is enzymatically cleaved by alkaline phosphatase, AMPPD becomes destabilised and decomposes with emission of light at a rate of which is dependent on the enzyme concentration (Bronstein et al, 1989).

Chapter 1 29 1.8.2.4 New approaches for trace immuno-analysis The conventional enzyme immunoassays have been successfully used in the detection of a range of analytes present in trace amounts. But as quantification is a function of enzyme activity, the method is sensitive to sample matrix effects, enzyme instability, temperature and incubation timing (Zuk et al, 1985). Recent advances in ELISA technology has led to the development of assays which can be used with the minimum of equipment and performed rapidly. Enzyme assays can be developed to give a colour change at a specific concentration. Also it has been possible to combine all of the reagents on paper strips often called dipsticks (Buck et al, 1986). They are of great use in areas such as emergency rooms, doctor's surgery or for "in the field measurement" (Litman et al, 1980).

1.9 APPLICATION OF MMPNOASSAY TO PESTICIDES ANALYSIS In spite of the advantages of immunoassays, it is only relatively recently that the potential of immunoassay techniques has been realised for the analysis of pesticides in the environment. This is probably because there is not sufficient knowledge about the advantages of the assays or the expertise and equipment are not available to the relevant laboratories. Moreover, the limited availability of specific antibodies to pesticides and other related compounds prevents the widescale use of immunoassays for environmental analysis. The description of antibody production to DDT and malathion was reported in 1970 (Centeno et al, 1970). A review on immunochemical anafysis of pesticides was published the following year (Ercegovich, 1971) which illustrated the advantages of immunological methods for pesticide analysis as supplemental methods for screening and confirmatory assays. RIAs were among the first immunoassay described for pesticides (Vallejo et al, 1982) (Knopp et al, 1985) but the ELISA format is now widely employed. Different forms of ELISA have been described for pesticides but the most applied enzyme system is HRPO with a colorimetric end-point. In general, the coated antibody ELISA format has been widely used although some assays utilise an antigen coated solid phase format (Niewola et al, 1986) (Hall et al, 1992).

Chapter 1 30 1 .9 .1 Food analysis At present, increased monitoring of food products suspected of being contaminated by pesticides is required by the food industry (Kaufman and Glower, 1991). Ercegovich developed an RIA m ethod for th e insecticide parathion which was capable of detecting lOywg/kg in lettuce extracts without cleanup (Ercegovich et al, 1981). The herbicide diclofopmethyl was determined in milk, wheat, soyabeans and sugarbeets by enzyme immunoassay using horseradish peroxidase coupled to diclofop as a label and compared with a similar method using diclofop coupled to fluorescein amine (Schwalbe et al, 1984). Immunoassays have been also used in food industry as ELISA methods for microbiological identification, the measurement of bacterial toxins and anabolic steroids have been developed (Morris and Clifford, 1985). An ELISA for methyl 2- benzimidazole carbamate, the degradation product of benomyl, has been reported (Bushway et al, 1990) in which a portable photometer is used, allowing analysis away from the normal anafytical laboratory.

1 .9 .2 Water analysis The legal limit for individual pesticides in drinking water set by European countries (MAC) is of OAjug/L (ppb) and total pesticide concentrations should not exceed 0.5/^g/ /L. The MAC is not necessarily a toxicity indicator but a limit set to protect water supplies firom pollution and thus frequent monitoring is required. For example, ELISA techniques for measuring atrazine and paraquat have been developed in our laboratory which have limits of detection of 0.03 and 0.0 Ifig/L respectively (Aheme, 1991). Immunoassays can have an important role to play in the monitoring of water for the presence of pesticide micro-contamination. The assays can be formatted to give qualitative or quantitative results for identifying those samples which require further analysis by more expensive techniques, thus reducing the time and cost needed for water sample analysis.

1 .9 .3 Soil and plant analysis Many of the assays developed for water analysis have also been used for the measurement of pesticides in soil and plants. Such assays are required to assess the persistence and effectiveness of pesticide residues. It is often essential to determine the presence of any previously applied chemicals before the next

Chapter 1 31 crop is planted. Soil and plant samples unlike water wül require some pre-assay treatment. Bushway et al (1988) reported an assay for measuring atrazine in soil which requires shaking or sonicating the soil sample for a short time in either water or acetonitrile. A more extensive soil extraction procedure was used for the analysis of paraquat in soil (Niewola et al, 1986) involving reflux of the sample in 6M acid followed by neutralisation of the extract.

1 .9 .4 Human exposure studies There is increasing evidence that occupational and domestic exposure to pesticides may have short or long term health effects. Biological monitoring is increasingly being used to detect whether workers have been C3qx>sed to and have absorbed the pesticide (Woollen, 1993) (Chester, 1993). This can be ascertained by measuring the concentrations of the pesticide or of its metabolites in biological media. Also in the case of accidental or deliberate ingestion of pesticides it may be necessary to monitor plasma levels to ensure that the correct treatment is given. Rapid, sensitive and simple procedures would facilitate analytical determinations. Immunochemical techniques satisfy these criteria and have become common analytical methods in clinical laboratory practice. Immunoassays can play a major role in occupational exposure monitoring (VaijEmon et al, 1986) and in the measurement of pesticides in plasma and urine. In our laboratory, enhanced luminescence immunoassays for detecting atrazine and paraquat in plasma and urine have been developed (Aheme et al, 1991). The assays are rapid, require no sample preparation and are sensitive to less th an l/ig/L. Moreover, a radioimmunoassay for biological monitoring studies of 2,4-dichloropheno3yacetic acid have been reported on occupationally exposed sprayers (Knopp and Glass, 1991). However, applications in the biological monitoring of occupationally or environmentally e^qxDsed persons are still in the early stages (Hammock et al, 1987) (Jung et al, 1989).

1.10 AIMS OF PRESENT STUDY The first aim of this project is to study the feasibility of developing immunoassays which have the potential to provide rapid and mexpensive tests for the detection and measurement of toxic substances mainly pesticides i.e. chlortoluron, isoproturon, 2,4-D and MCPA in a variety of sample matrices (water samples and biological fluids). As the main advantage of immunoassay

Chapter 1 32 over more conventional techniques is one of sensitivity, the assays were developed with the view of specifically measuring pesticides at or below the EC MAC of 0.1/fg/L. The second aim was to immobüise the available antisera on solid phases and to investigate their potential for the development of portable test kits which can be used in extra-laboratoiy locations by relativefy unskilled personnel.

Chapter 1 33 CHAPTER TWO MATERIALS AND METHODS

Chapter 2 34 2.1 REAGENTS 2 .1 .1 General reagents i) BDH Chemicals Ltd., Poole, Dorset, UK. Disodium hydrogen orthophosphate AnalaR grade. Potassium dihydrogen orthophosphate AnalaR grade. Sodium dihydrogen phosphate AnalaR grade. Hydrogen peroxide solution 30%. Acetic acid. Lactose. Glycine. Sodium chloride.

ii) Sigma Chemical Co Ltd., Poole, Dorset, UK. Charcoal, activated, untreated powder 250-350 mesh. Polyoxyethylene sorbitan monolaurate "Tween 20". Sodium barbitone. 3,3,5,5 Tetra methyl benzidine (TMB). Citric acid. N-hydro5y succinimide. Dicyclohejyl carbodiimide. Sodium azide. Thyroglobulin. Sodium cyanoborohydride 90-95%. Thimerosal.

iii) Amersham International PLC., Buckinghamshire, UK. Amerlite signal enhancement rea^nts.

iv) Pharmacia LTD., Milton Keynes, Bucks, UK. Dextran T-70. Optiphase 'safe' scintillation fluid. Diethylaminoethyl DEAE sephadex disc cartridge.

Chapter 2 35 v) ClifBftar A sso cia tes L im ited ., University of Surrey, Guildford, UK. Aldehyde activated porous silica. Porous silica (Matrex™, particle size 90-130 /^m, pore size lOOOA®)

vi) Evans., Leatherhead, Surrey, UK. BCG vaccine (Bacille Calmette-Guerin).

2 .1 .2 Herbicide compounds Grey Hound., Birkenhead, Merseyside, UK. Isoproturon Chlortoluron Chlorbromuron Chlormethiuron Chloroxuron Diuron Fluometuron Linuron Metobromuron Metoxuron Neburon MCPA (2-methyi-4-chloropheno5y acetic acid). 2,4-D (2,4-dichlorophenoxy acetic acid).

2 .1 .3 Radiülfjbéls Sigma Chemical Co. Ltd., Poole, Dorset, UK. 2,4-dichlorophenoxy acetic acid carbo3y-14C: 314.8MBq/mmol.

Ciba Giegy., Cambridge, UK. I'^C chlortoluron [phenyl-(u)-1 ] - chlortoluron: 637.5MBq/mmol. I'^C isoproturon (phenyl-(u)- I^c] - isoproturon: 268MBq/mmol.

2 .1 .4 Orgcmic solvents BDH Chemicals Ltd., Poole, Dorset, UK.

Chapter 2 36 Methanol AnalaR grade. Absolute alcohol 100% AnalaR grade. Dimethyl formamide AnalaR grade.

2.1.5 Disposable plastics Columns were purchased from Lab M (Bury, Lancs, UK). The colorimetric 96 well microtitre plates used in the ELISA techniques were Nunc-Immuno plates Maxisorp F96 (certified grade 1) supplied h y Gibco Europe (Uxbridge, UK). The round bottom, flexible polyvinyl chloride, microtitration plates used for the camera luminometer and the chemiluminescent microtitration plates, solid flat bottomed white Microlite ™2, were purchased from Dynatech Labs Ltd (Billingshurst, West Sussex, UK). Scintillation vials (GK 16 B scintminivial tubes & caps; No: 10200) were purchased from Rocket of London Ltd (Tyne & Wear, UK). LP3 and LP4 tubes were purchased from Luckhams Ltd (Sussex, UK). AH pipette tips were bought from Alpha Laboratories (Eastleigh, Hampshire, UK).

2.1.6 Manucd liquid handling a) AH single pipetting was performed using pipettes manufactured by GHson (Luton, Beds, UK) and by Labsystem UK Ltd (Basingstoke, Hants, UK). b) A Zippette 10 ty Jencons sci Ltd (Leighton Buzzard, UK) was used for repeat dispensing of scintiHation fluid.

2 .1 .7 Didlysis i) For smaH (<10ml) volumes, ceHulose membrane tubing (Sigma, D-9277) 10mm wide was used to perform the dialysis. Prior to use the appropriate length of tubing was boHed for an hour in water. Samples were transferred and sealed into doubly knotted tubing and incubated in three changes of 1 Htre volumes of the chosen buffer for 48 hours at 4*C stirring constantly. ii) Volumes between 10-1000ml were dialysed in 22mm wide Visking tubing suppHed Ty MediceH Int (London, UK).

2.1.8 Glassware AH glassware used was sent on a daffy basis to an in-house washing facffity and rinsed with deionised water prior to use. Usually aH smaH volumes of glassware were new and used once only.

Chapter 2 37 2.2 EQUIPMENT 2.2.1 General equipment Small quantities of reagents (l-200mg) were weighed out into vessels on a Mettler H 18 balance, accurate to plus or minus 0.1 mg. Larger quantities of reagent were weighed out on a Mettle PI200 N top-pan balance, accurate to plus or minus lOmg. Both these instruments were supplied by GallenKamp (Technico House, Christopher Street, London, UK). Measurements of buffer pH were made using a Kent pH meter model 7020, fitted with a bil 9142 sealed (dry) electrode. The pH meter was calibrated before each evaluation by using pH buffer solutions made up on a weekly basis using pH buffer capsules (pH's 4, 7 and 9) manufactured Ty Russel and supplied ly the Sigma Chem Co. General assessment of pH were performed using narrow range (0.0-6.0, 4.5-10 and 7.0-14) pH test strips firom the Sigma Chem Co.

2 .2 .2 Specialised equipment a) Microtitre plate reader Labsystems M ultiskan BICHROMATIC supplied ly Labsystem (Basingstoke, Hampshire, UK). b) The Enfer 1000 chemiluminescent plate reader and Enfer 1000 four place microtitration plate shaking incubator were supplied by Enfer Scientific (UK) Ltd (Derby, UK). c) Microtitre plate camera luminometer-Dynatech laboratories Ltd (Billingshurst, West Sussex, UK). d) WALLAC 1410 liquid scintillation counter supplied ly Pharmacia Ltd (Milton Keynes, Bucks, UK).

2 .3 BIATERIALS Double glass distilled water was used for the preparation of assay solutions.

2.3.1 General buffer solutions

Citric Acid Buffer p H 5 7. lOg di-sodium hydrogen orthophosphate anhydrous 5.22g citric acid Dissolved in 1 litre of distilled water.

Chapter 2 38 Phosphate Buffered Saline (PBS) p H 7.4 8.00g sodium chloride 0.20g potassium chloride 0.20g potassium di-hydrogen orthophosphate 2.9 g di-sodium hydrogen orthophosphate anhydrous Dissolved in 1 litre of distilled water

Phosphate Buffered Saline with Tween IPBST) 0.50ml Tween 20 Dissolved in 1 litre of PBS pH 7.4.

Phosphate Buffered Saline with Gelatin fPBSG) p H 7.4 I) Phosphate Buffer pH 7.4 200ml of O.IM potassium di-hydrogen phosphate (containing 0.014% thimerosal) was mixed with 800ml of O.IM di-sodium hydrogen phosphate (containing 0.014% thimerosal). n) Following mixing of 12g sodium chloride cind 2.0g gelatin per litre distilled water, phosphate buffer as at (a) was added and the resultant PBSG buffer stored at 4°C until use.

Sodium Barbitone Buffer p H 9.6 14.40g barbitone sodium Dissolved in 1 litre distilled water.

Phosphate Buffer (0. IM) d H 6.8 6.96g di-sodium hydrogen orthophosphate anhydrous 7.69g sodium di-hydrogen orthophosphate anhydrous Dissolved in 1 litre of distilled water.

Phosphate Buffer (0.0 IM) t>H 6.4 0.38g di-sodium hydrogen orthophosphate anhydrous 1.15g sodium di-hydrogen orthophosphate anhydrous Dissolved in 1 litre of distilled water.

Chapter 2 39 Acetic acid flM) d H 2.7 57.29ml glacial acetic acid made up in 1 litre of distilled water.

Tetramethyl Benzidine TMB (chromagen/substrate) 100/d TMB stock (20mg/ ml in dimethyl formamide DMF)was added to 20ml citrate buffer and 10/d H 2O2 (30% v/v) immediate^ before use.

2.3.2 Radiolabels Substock solutions: -isoproturon 1/100 dilution -chlortoluron 1/100 dilution 14c -2,4-D 1/100 dilution All substock solutions were made up with methanol.

2.3.3 Dextran-coated charcoal (DCC, 2.5%) l.Og of Dextan T-70 and lOg charcoal were mixed in 400ml 0.05M phosphate buffer (pH 7.4) by stirring overnight at 4°C. The mixture was then centrifuged (2500rpm, 4°C, 30 min) to remove charcoal fines. The de-fined pellet was resuspended in a 400ml of buffer and the suspension, stored at 4 ^ for up to three months, was always thoroughly stirred before use.

2.3.4 Prepcaration of pesticide stock standard solutions Each pesticide (isoproturon, chlortoluron, 2,4-D and MCPA) was prepared (1 mg/ml) separately in methanol. Each compound was accurately weighed (5mg) on a five place balance and made up to a volume (5ml) in a 5ml glass bottle. These stock standard solutions were aliquoted into 5 x 1ml portions each and stored at 4°C in the refrigerator.

2.3.5 Water scanples a) Tap water was collected in a 5 litre glass conical flask firom a tap in 26 AY 19 general laboratory sink, SOBS, University of Surrey. The tap was turned on for 30 minutes to drain the pipe before water was collected. b) River Ouse water was collected (in 1990) firom different sites of the river (100ml each). Some of each river water sample was filtered using Whatman filter paper 1. c) University lake water was collected in the morning (2 Litre).

Chapter 2 40 d) Guildford canal water was collected in the afternoon (2 Litre).

2.4 METHODS 2 .4 .1 Antiserum production The most important single factor in establishing an immunoassay is the production of a suitable antiserum. This depends predominantly on the production of antibodies that possess a high avidity and selectivity for the material to be assayed (Kemeny and Chantier, 1989). The immunogenicity of a compound is related to its molecular weight and the rigidity of the compound's structure. Purity of antigens is also an important consideration. Although, it was reported that impure substances could improve antisera production compared to pure preparations, with the impurities acting as adjuvants (Hum and Landon, 1971), the specificity of the antisera could be affected significantly ly the impurity of the antigen. The use of an adjuvant, such as mineral oil, was shown to increase the chance of an antibody response. The adjuvant was mixed with the immunogen to form a stable emulsion. It is believed to work by releasing the immunogen over a period of time, thereby preventing its rapid uptake by the circulation avoiding excessive degradation. The adjuvant also facilitates phagocytosis of the immunogen, an essential step of antibody production, in addition to causing the formation of local granulomatous lesions which can act as a focus of antibody production (Morris, 1985). In theory, antisera can be obtained firom any animal able to initiate an immune response. The choice depends on the nature of the immunogen, and the use to which the antiserum is to be put. Rats can be used for screening some preparations for immunogenicity prior to injecting into larger species (Robinson et al, 1975). Although antisera to small molecular weight compounds have been widely produced in rabbits, the relatively small volume of antisera produced can be limiting, and thus for this reason sheep and goats are increasingly used. Also, antisera raised against smaH molecules tend to be of low titre and large volumes of serum firom sheep and goats are preferable (Aheme and Marks, 1979). It is favourable to immunise a large number of animals to increase the chance of attaining a suitable antisera, although in reality the number of animals is often limited l y space and cost.

Chapter 2 41 The route of immunogen administration is also important. Few critical studies have been made of the effect of this, but Hum and Landon (1971) reported sites in decreasing order of effectiveness: lymph nodes, intra-articular, intra-dermal, intra-muscular, intra-peritoneal and intra-venous. The timing of bleeds and boosting injections is a matter for conjecture, although Aheme and Marks (1979) favoured a rest period of several months (4-6) between boosts once a response had been established.

2.4.1.1 Preparation of immunogen An N-hydroxy succinimide active ester method (Anderson, 1963) was used for conjugating haptens synthesised for isoproturon, chlortoluron, 2,4-D and MCPA to thyroglobulin. This method links carboxyl groups on haptens to amino residues on the carrier protein (thyroglobulin). Fig 2-1 illustrates the reaction sequences for hapten-carrier protein conjugation.

R-BRIDGB-COOH R-BRIDGE-CO PROTEIN + I + OH ® NHg I I ^1^

HjjO

[ActiveEster ) ( N-Hydrfijqfsmxiniinide ) ^ ^-N=C=N- ___

( Dicyclohexglcarbodiimide )

R-BRUXxE-CO-NH-PROTEIN

[Sitbstituted. Urea )

R= Hapten

Figure 2-1 N-hydroxysuccinimide "active ester" conjugation method.

Chapter 2 42 Method Protocol lOO/iM N-hydroxy succinimide was dissolved in 0.25ml DMF 300/

2.4.1 2 Antisera Two mature Suffolk sheep for each immunogen (hapten-carrier protein conjugate) were used for immunisation using conventional procedures. 5.0mg thyroglobulin conjugate and 0.1ml Bacüle Calmette-Guerin (BCG) vaccine in 1ml sterile water was emulsified with 2ml of non ulcerative |^eunds adjuvant (NUFA) and injected intramuscularly into the hind legs and back. The animals were boosted approximately two months later with half the primary dose containing no BCG. Boosting doses were given at appropriate intervals (every 4-5 months) vhich resulted in a significant immune response. Blood was collected firom the jugular vein ten days after each immunisation and at appropriate intervals. After allowing the blood to clot, the serum was separated and stored at 4°C with the addition of sodium azide (0.1% final concentration).

2 .4 .2 RIA procedures The protocol used for the construction of antiserum dilution curves (and displacement curves) from which the antiserum titre can be calculated is shown in Table 2-1:

Chapter 2 43 Table 2 -1 Protocol for antiserum assessment.

Total NSB Antiserum Displacement Binding tube Tubes # (fd) (fd) (fd)

Diluent (PBSG) 600 500 400 300 Radiolabel 100 100 100 100 Antiserum 100 100 Standard (analyte) _ 100 DCC 100 100 100 Supernatant 500 500 500 500 Scintillation fluid 4000 4000 4000 4000

The assay was set up in duplicate utilising disposable LP3 plastic tubes. Assay buffer (PBSG) was used to dilute the antisera with a range of 1/20 to 1/320. Similarly, radiolabel substock solutions were diluted as appropriate equivalent to radioactivity count of 1000 CPM in total tubes. Tubes containing diluent, radiolabel and antiserum were vortex-mixed and incubated on ice for one hour. The antibody-bound analyte was then separated from the free fraction by incubation with ice-cold DCC for ten minutes on ice followed by centrifugation (2500ipm for ten minutes at 4°C). Then SOOjul of the resulting supernatant (bound fraction) were transferred into polyethylene scintillation vials. The vials were mixed and the radioactivity in the vials was measured by counting each sample for four minutes in a liquid scintillation counter. Duplicate results were measured and calculated as a percentage of the total counts (T) following subtraction of counts due to non specific binding (NSB). Percentage bound B/T% was calculated for each point. Antiserum titre was determined by evaluating the dilution of antiserum that bound 50% of the available bound radiolabel. Moreover, antiserum avidity, the energy of binding of the antibody for its antigen, was indirectly determined by inspection of the relevant antiserum dilution and displacement curves, the greater the displacement the greater the avidity.

Chapter 2 44 2 .4 .3 DEAE purification o f cm tisera The diethylaminoethyl (DEAE) Sephadex disc (an anion exchanger) was used to separate immunoglobulins (lÿl) from other serum proteins. The principle of this separation was to allow the IgG immunoglobulins, which are positively charged at pH 6.4, to pass through the anion exchanger while retaining negatively charged proteins such as albumin.

Protocol i) Neat antiserum (2ml) was diluted 1:10 in 0.0IM phosphate buffer pH 6.4. ii) DEAE Sephadex disc was washed with 75ml of 0.0 IM phosphate buffer. iii) The diluted antiserum was applied onto the disc followed by enough 0.0 IM phosphate buffer to collect 10 x 5ml fractions. iv) The optical density (OD) of each fraction was m easured at 280nm. v) The most concentrated fractions were pooled and the OD was measured again at 280nm. Then the OD reading was divided by 1.4 to convert the optical density into concentration (A 1% 280 nm= 14; Little and Donahue, 1967).

V i) The purified immunoglobulin was stored at 4PC with the addition of thimerosal solution (0.1% final concentration). vii) The disc was cleaned with 75ml of IM acetic acid and regenerated with 75ml of 0. IM phosphate buffer pH 6.8 followed by 75ml of 0.0IM phosphate buffer pH 6.4 ready for purification of the next antiserum. viii) Finally, the disc was stored at 4PC after 75ml of 0.0 IM phosphate buffer pH 6.4 containing 0.1% thimerosal was run through.

2 .4 .4 HRPO-hapten conjugate Haptens were also conjugated to HRPO using the N-hydroxy succinimide active ester method (see 2.4.1.1) at different nominal molar ratios (HRPO:hapten) of 1:40, 1:20, 1:10 and 1:5. The enzyme conjugates were dialysed against phosphate buffered saline, PBS pH 7.4, (3 x l L) over 48 hours. The concentration of the proteins (HRPO) in the retentâtes were measured by a spectrophotometer at 405nm and the solutions were stored at 4 ^ in the presence of 0.1% thimerosal solution.

Chapter 2 45 2 .4 .5 Competitive ELISA Final Assay Protocol: 96 well polystyrene microtitre plates were utilised as the solid phase for the antibody ELISA. 200/d aliquots of the purified sheep anti-hapten serum, diluted in sodium barbitone buffer pH 9.6, were dispensed into each well (excluding the outer ring of weUs to minimise the differential surface effect of the plate edges caused during the manufacturing process of the plate). The antibody was incubated overnight at 4°C in a humidified air tight container. Any uncoated material was removed ty emptying the contents of the coated plate and washing it three times with phosphate buffered saline containing 0.05% Tween 20 (PBST). 100/fl of a series of herbicide standards, diluted in tap water (1(X), 10, 1, 0.1, 0.01 and 0 ng/ml) were added to individual wells in duplicate. In addition, 100/d of enzyme labelled herbicide, diluted in PBS buffer pH 7.4 and containing 2.5% normal serum, were added to all wells. The plate was incubated in the shaker incubator (Enfer) for one hour at room temperature. Excess unbound labelled herbicide or herbicide standard was removed ty emptying the contents of the plate and washing it three times with normal tap water (tap sink). 200/d aliquots of peroxidase chromagen / substrate (TMB) were added to each well of the plate to measure the amount of enzyme labelled herbicide bound to the immobilised antibody and subsequently incubated in the shaker for 30 minutes at room temperature. Enzyme activity was stopped by the addition of 50/d/well IM HCl. Finally, the amount of colour developed on the plate was read using a Labsystems Multiskan plate reader at 450nm and the standard curves were evaluated offline using a 4 parameter logistic plot (Ria Calc, Wallac Ltd., Milton Keynes). Figure 1-3 demonstrates a schematic representation of this reaction sequence. During the development and vahdation of ELISA the procedures and conditions may differ from this scheme and are described in the appropriate section.

2 .4 .6 Competitive ECLIA Final Assay Protocol: The purified sheep anti-herbicide was adsorbed to the solid phase as specified in 2.4.5 using 96 well flat bottom Microhte plates for antisera adsorption.

Chapter 2 46 A series of herbicide standards diluted in tap water were added to the plate followed by the addition of 100/^1 of enzyme labelled herbicide, diluted in PBS buffer containing 2.5% normal serum, to all wells. The plate was placed into a humidified container and incubated in a shaker (as mentioned in 2.4.5) for one hour at room temperature. Excess unbound labelled herbicide and herbicide standard was removed by emptying the contents of the plate and washing three times with normal tap water. 200/d aliquots of Amerlite signal solution (see materials) were added to each well of the plate and then incubated for ten minutes at room temperature. Finally, the light signal was measured using the microtitre plate luminometer (Enfer). Figure 1-3 also illustrates the competitive coating antibody ECLIA reactions.

2 .4 .7 Competitioe Immunocffhiity column assay Final assay protocol: 2ml of the contaminated sample were mixed with 5/d of IM HCl to bring the pH to 5 followed by the addition of the enym e labelled herbicide solution. After mixing, the above mixture was transferred onto a 1ml immunoaffinity column packed with silica bound antibodies. The column was washed with 3ml of PBS buffer pH 7.4 to remove the unbound labelled and unlabelled herbicide. Finally, 100/^1 of TMB chromagen-substrate were added onto the column to allow colour development. The colour obtained was compared with the control zero standard's colour (mixture of various matrices previously shown to be non- immunoreactive) run in a similar manner to the contaminated samples. The presentation of the competitive enzyme immunoassay using immunoaffinity columns is illustrated in Figure 6-11.

Chapter 2 47 CHAPTER THREE DEVELOPMENT OF ISOPROTURON AND CHLORTOLURON ELISA METHODS.

Chapter 3 48 3.1 INTRODUCTION TO UREA HERBICIDES Urea is the amide of carbdmic acid. In most members of the ureide (i.e. substituted urea) family of herbicides, urea is trisubstituted in the way indicated in Figure 3-1. One of the amino groups carries either two methyl groups or one methyl and one methojy group. The other amino group is substituted with a benzene ring which, in most cases, contains halogen atoms. Ureides are solids with a low vapour pressure (10"®-10"® mmHg) at room temperature and possess aqueous solubilities ranging from 4 to 700ppm (Hassall, 1990).

CH(CH3)2 CHg

V V NH NH NHa •u 4.

CO .1 ,1 1 /\ /\ NÜ 2 CHg OHs CH3 CHg

UREA ISOPROTURON CHLORTOLURON

Figure 3-1 Urea herbicides (ureides).

Ureides are adsorbed on soil organic matter and, to a lesser extent, on clay particles. Adsorption on soil constituents not only determines persistence, leaching characteristics and availability for root uptake but also influences the rate of degradation by micro-organisms. The principal mode of action of urea herbicides is to disrupt the light reaction of photosynthesis, although there is evidence that the action of ureides is not «cclusivety on photosynthetic tissues (seedlings may die before emerging from the soil and larger plants die even if kept in the dark) (Hassall, 1990). Acute toxicity of ureides to mammals is usually low (Brown, 1980). On the other hand, low concentrations of some ureides have been known to inhibit testicular DNA synthesis in mice (Seiler, 1979) which in turn can reveal potential

Chapter 3 49 carcinogenicity. Sarkar and Gupta (1993) have reported neurotoxicity effects in mice by a single oral administration of isoproturon. Isoproturon and chlortoluron are phenylurea herbicides which have been detected in drinking water supplies in the UK (Lees and McVeigh, 1988). The presence of urea herbicides in water supplies is an emerging problem of concern to the water industry in UK today. Indeed, results available from the Department of Environment in 1988 indicate that isoproturon and chlortoluron as well as certain commonly used triazine and phenoxy alkanoic acid herbicides were present at levels exceeding the MAC (Roberts, 1991). Such information indicates that there is a need by the water industry in the UK (and elsewhere) to monitor water supplies for herbicides. Besides, due to the lack of publications in the literature about convenient methods for measuring isoproturon and chlortoluron these two herbicides have been chosen for further studies. Isoproturon, also known as 3-(4-isopropylphenyl)-l,dimethylurea, is commonly used in the UK to control broad leaf weeds and grasses particularly in the cultivation of cereals, whereas chlortoluron is used extensively in the control of the four types of wild oats and black grass in winter barley. The traditional and approved techniques for measuring isoproturon and chlortoluron, i.e. chromatographic methods, involve extensive sample preparation, using solvent or solid phase extraction procedures, which adds to assay costs and reduces the number of samples that can be handled at one time (Sidwell and Ruzicka, 1976). Besides, they reveal limits of detection in the /wg range which are not adequate enough to detect the MAC of a single pesticide in drinking water (Walters et al, 1984). Therefore, as an alternative, quantitative enzyme-linked immunosorbent assays (ELISA) have been developed for the detection and measurement of isoproturon and chlortoluron offering highly sensitive, simple and rapid techniques of analysis.

3.2 HAPTENS After several meetings and discussion, the chemical synthesis of all haptens were prepared by Dr. G. Frost at Robens Institute.

3 .2 .1 Chemical synthesis o f isoproturon hcq^ten '*N-f4r(l-methylethyl)phenyy- N-methÿl-N~carboxypropyl~ urea” The first stage of the synthesis, involving the formation of isopropylphenyl isocyanate, was carried out with triphosgene (bis(trichloromethyllcarbonate) as a

Chapter 3 50 convenient and less hazardous replacement for the monomer. Triphosgene (0.85g, 2.86mmol) was weighed into a two-neck flat-bottomed flask (150ml) fitted with a reflux condenser and an addition funnel mounted on a suitable hot plate stirrer. Dry toluene (6ml) was added and the solution stirred using a magnetic follower. A solution of 4-isopropyianiline (I.14g, 8.4mmol) in dry toluene (7ml) was added dropwise and the funnel was washed out with dry toluene (5ml). The mixture was refluxed gently and all solids slowly dissolved with the evolution of HCl gas (CaCl2 guard tube). Heating was continued for two hours and the flask and contents were left overnight at room temperature. For the second stage, the intermediate isocyanate solution was evaporated to remove toluene and the viscous residue was added dropwise to a stirred solution of N-methylaminobutyrlc acid hydrochloride (I.3g, 8.5mmol) and sodium hydroxide (0.74g, 18.5mmol) in water (10ml) in a 50ml conical flask. The solution was stirred for two hours at room temperature at which time the pH was 11. A fine precipitate was formed and a small amount of this was filtered off and washed with water (7ml).The filtrate was acidified with 2M HCl to pH 3 and the resulting precipitate was filtered on a Whatman No 50 paper (Buchner funnel) and washed with water until the filtrate was pH 5. The product was dried in a vacuum over CaCl 2 to 725mg. Examination by TLC (Ethyl acetate : Hexane, 4:1) showed one major spot only. The proposed structure of the hapten is shown in Figure 3-2.

Chapter 3 51 (CHgla-CH- ^ ^'N H g + 1 /3 (CClgOjg CO

4-(l-methylethyl}phen.ylamine CO TYiphosgene piis(tiHchloromethy1)carbon.eit^

(CH3)2-c h - ^ y-N C O

4 -fl-methylethyl)pkenylisocyanate (U)

N-methylaminobiityric acid CH3-NH(CH2)3-C00H

/CH 3

(CH3)2-CH- ^ /) -NH-CO-N^ (CH2)3-C00H

IT-i4-(l-methylethy1)phenyll-N-methyl-N-carboxypropyl urea (IH)

Figure 3-2 Chemical synthesis of the isoproturon hapten.

3.2.2 Chemical synthesis of chlortoluron hapten [N-(3-chlora4- methylphenyl)-N-methyl-N-€xu'boxypropgl ureaj The chlortoluron hapten was synthesised under similar conditions to isoproturon, but with 3-chloro-4-methylaniline replacing isopropylaniline at the first stage. The compound also showed one major spot only by TLC. Figure 3-3 illustrates the chemical reactions for synthesising the chlortoluron hapten.

Chapter 3 52 1 /3 (CCIgOlg CO

3-chloro-4-methylaniline CO Triphosgene [bis(trichloromethyl)carbonatej

CH3 - -NOG

3-chloro-4-methylphenyUsocycmate (H)

N-methylcuninobutyric acid CH3-NH(CH2)3-C00H

/CH 3

CH3 - -NH-CO-N \ (CH2)3-C00H

N'-3-chloro-4-methyIphenyl-N-methyl-N-carboxypropyl (HÇurea

Figure 3-3 Chemical synthesis of the chlortoluron hapten.

3 .3 IMMUNOGENS The isoproturon and chlortoluron haptens were conjugated to bovine thyroglobulin using N-hydroxy succinimlde active ester (Anderson, 1963) as mentioned in 2.4.1.1. This method links carboxyl groups on haptens to amino residues on the carrier protein. Figure 3-4 illustrates the presumed structures of the isoproturon and chlortoluron immunc^ns.

Chapter 3 53 ^CÜ3 (CHgjgCH- -NH-CO-N \ (CHgjg-CO NH thyroglobuliii

ISOPROTURON IMMUNOGEN

-NH-CO-N

(CH^)3'CO-NH-thyroglobulin

CHLORTOLURON IMMUNOGEN

Figure 3-4 Isoproturon and chlortoluron protein conjugates.

3.4 IMMUNISATION PROCEDURE Two mature Suffolk sheep were utilised for each Immunogen (isoproturon and chlortoluron immunogens) to raise antisera using conventional procedures (as described in 2.4.1.2). Blood was collected from the jugular vein seven days after the primary dose, then this was repeated twice per month from then on. One of the sheep immunised with the isoproturon immunogen died at an early stage (four months after the primary dose). Five booster doses were given to the remaining isoproturon sheep whereas seven booster doses were given to chlortoluron sheep at appropriate intervals which resulted in a significant immune response. The serum was separated after allowing the blood to clot and stored at 4°C with the addition of 0.1% sodium azide. Experience in this laboratory has shown that antisera stored at 4°C in the presence of sodium azide are very stable up to 15 years (Middleton et al, 1988).

3.5 ANTISERA SCREENING PROCEDURE A radioimmunoassay (RIA) procedure using isoproturon and chlortoluron radiolabels and dextran coated charcoal was utilised to screen the antisera for the presence of specific antibodies. The assay procedure was carried out as described in 2.4.2 to measure the titre of the bleeds. The results of the

Chapter 3 54 immune responses are shown in Figure 3-5. Titre is the dilution of the antiserum which binds 50% the immunoreactive radio-label as shown in Figure 3-5. The two sheep immunised with the immunogen isoproturon (4651 and 4676) responded positively by producing specific antibodies to isoproturon. Although sheep 4651 died, sheep 4676 was boosted five times and showed a good binding capacity towards isoproturon. As a result, the fourth boost, second bleed (IV:B) has displayed significant displacement with the addition 0.5/

3.6 PURIFICATION OF ANTISERA Following initial screening for the presence of antibodies in an RIA using labels (isoproturon and chlortoluron), the chosen antisera were partially purified on DEAE cellulose ion exchan^ discs (LKB). 2ml firom each serum were diluted to 20ml with 0.0IM phosphate buffer pH 6.4 and applied to the disc. Immunoglobulins (I^s), \dnch are positively charged at pH 6.4, pass the anion exchanger which retains negatively chared proteins as described in 2.4.3. 2ml volume per fraction were collected and the immunoglobulins were measured on a spectrometer at 280mn as shown in Figure 3-7. The most concentrated fractions were pooled and measured at 280nm. The concentration of immunoglobulin in the partially purified isoproturon antibodies was found to be 1.29mg/ml whereas the chlortoluron immunoglobulins concentration was 1.39mg/ml. The purified antisera were then stored at 4°C with 0.1% thimerosal solution added as preservative. Purified antisera are stable for at least a year but each new batch must be checked for the exact dilutions to be used in the assay.

3.7 HRPO-HAPTEN CONJUGATE The isoproturon and chlortoluron haptens were conjugated to HRPO by the N-hydroxy succinimide active ester technique (2.4.1.1). Various nominal molar ratios between HRPO protein and isoproturon or chlortoluron hapten

Chapter 3 55 (HRPO : hapten) were used: 1:40 (lOmg HRPO: 150/d active ester), 1:20 (lOmg HRPO : 75/d active ester), 1:10 (lOmg HRPO : 38/d active ester) and 1:5 (lOmg HRPO : 19/d active ester). The mixtures were left rolling overnight at room temperature. Next day enzyme conjugates were dialysed for 48 hours against PBS buffer pH 7.4 and the protein concentrations of the peroxidase in the retentâtes were measured using a spectrophotometer at 405 nm. The final concentration of peroxidase in the isoproturon conjugate solution was 2.3mg/ml vdiereas the final concentration of peroxidase in the chlortoluron conjugate solution was 2.1mg/ml. The enzyme solutions were then stored at 4 ^ after the addition of 0.1% thimerosal solution. The concentrated enzyme labels appeared to be very stable, and freshly diluted enzyme label were made for each daily assay. An isoproturon HRPO enzyme label prepared in March 1991 was used until July 1992. At this time, although the label was still usable in the assay, the colour produced ty the HRPO had decreased by approximately 39% and a new label was made. Figure 3-8 shows a comparison between the isoproturon standard curves obtained with the new and old isoproturon HRPO labels. The shape of the curves were not significantly different but the total amount of colour intensity (OD) was much lower in the older preparation and this may have an adverse effect on the performance of the assay.

3.8 OPTIMISATION OF ASSAY CONDITIONS In immunoassays, reagents must be titred to find out which conditions will produce the best assay. The standard curve should have sufficient sensitivity to detect the analyte at the chosen level for the particular application. Ideally, it should cover the range of concentrations likely to be encountered, give good recoveries and have a low inter (between) and intra (within) assay variation. To achieve this, the antibody and label concentrations as well as the incubation times and temperatures are altered and tested in various combinations. Additionally, the assay diluent may require alteration (i.e. specific diluents are required to prepare the standard curve for each sample).

Chapter 3 56 (a) Titre 500 450 400 350 300 Sheep 4676 250 boost boost boost boost boost Sheep 4651 200 orims 150 100 50

0 10 20 30 40 50 60 70 80 90 100 Week No

(b) Titre 1000 boost boost boost 900

800 prims 700 600 Sheep 1826 500 Sheep 1825 400 300 200 100

0 10 20 30 40 50 60 70 80 90 100 Week No

Figure 3-5 Immunisation charts for (a) isoproturon, bleed (IV:B) and (b) chlortoluron, ® ; bleed (V:C).

Chapter 3 57 B/T% 100 90 80 70 60 A/S dilution curve 50 0.5|jg/L isoproturon 40 30 20 10

10 100 1000 Isoproturon antiserum dilutions A/s 4676 (Bleed IV:B)

B/T% 100 90 80 70 60 A/S dilution curve 50 0,5m g/L chlortoluron 40 30 20 10

100 100010 Chlortoluron antiserum dilutions A/s 1826 (Bleed V;C) Figure 3-6 Evaluation of isoproturon and chlortoluron antisera using RIA (14c labels).

Chapter 3 58 (a) OD 280nm

2.5

1.5 -

0.5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Fractions (2ml each)

(b) OD 280nm 3 r

2.5

1.5

0.5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Fractions (2ml each)

Figure 3-7 Purification of (a) isoproturon 4676 (IV:B) and (b) chlortoluron 1826 (V:C) bleeds on DEAE eluted with 0.0 IM phosphate buffer pH 6.4. (§): Pooled fractions.

Chapter 3 59 B/Bo% 100 90 80 70 60 ISO-HRPO 3/91 50 ISO-HRPO 7/92 40 30 20 Bo-1.2

Bo-1.9

0.001 0.01 0.1 1 10 100 1000 Isoproturon pg/L

Figure 3-8 Comparison between Isoproturon standard curves using recently and previously prepared isoproturon horseradish peroxidase label (ISO-HRPO).

Chapter 3 60 3 .8 .1 Isoproturon ELISA 3.8.1.1 E)valiiation of the purified antiserum (litre and displacement) The direct competitive ELISA method was adopted for the measurement of isoproturon vdiereby the purified antibody was coated onto a microtitre plate. Standard solutions of isoproturon, samples and isoproturon enzyme label are added together causing competition between isoproturon and enzyme label for the antibody binding sites. This type of ELISA method offers simplicity by minimising the number of steps in the assay, and above all reducing the effect of non-specific binding which could be produced by the bridge between the hapten and its coating carrier protein conjugate resulting in reduced sensitivity. Titre is defined as the dilution of the antiserum which binds to 50% of the immunoreactive protein of the added eu 2yme label. It was determined by means of an antiserum dilution curve which was carried out as follows : 200^1/well of five serial dilutions of the partially purified antiserum (1/500, 1/1000, 1/2000, 1/4000 and 1/8000) diluted in sodium barbitone buffer pH 9.6 were added to the wells of a microtitre plate in duplicate and incubated in a moist chamber overnight at 4^. Similarly, six dilutions of the pre-immune serum were applied to the plate to determine the binding specificity of the purified isoproturon antibody. Next day, the plate was washed three times with 0 .15M PBS buffer pH 7.4 which contains 0.05% "Tween 20" (PBST) to remove any excess of uncoated antibodies. 100^1/well PBS buffer were added onto all antiserum dilution curve wells followed by the addition of another 100/^1/well isoproturon enzyme label (molar ratio 1:40) diluted with PBS buffer at 1/1000 dilution (2.3 jug/ml peroxidase) to bring the total volume of each well to 200^1. The plate was then incubated in the Enfer shaker for two hours at 37°C. Finally, the plate was washed three times with tap water and 200^1/well chromagen-substrate solution were applied to all wells and left for 30 minutes shaking at 37°C for colour development. The reaction was stopped by adding 50/fi/well IM HCl and read at 450nm using the Labsystem microtitre plate reader. The results were presented by plotting the optical density of the enzyme label bound by antibodies against the dilution of antiserum as demonstrated in Figure 3-9. The term avidity, on the other hand, is used to denote the potential antigen binding displayed by an antibody. An antibody with high avidity will display a tendency to bind a lot of antigen and, conversely, an antibody with low

Chapter 3 61 avidity will have a tendency to bind few antigen molecules (Edwards, 1985). Performing a displacement curve is an Indirect way to measure the antiserum avidity. Therefore, a displacement curve was produced together with the above antiserum dilution curve on the same microtitre plate. The displacement curve was carried out as for the antiserum dilution curve, but lOO/fl/well isoproturon standard at 0.5/

3.8.1.2 Assay protocol for standard curve The inner 60 wells of a microtitre plate (as described in 2.4.5) were coated with 200jWl/well partially purified isoproturon antibodies diluted with 0.07M sodium barbitone buffer pH 9.6 and incubated in a moist chamber overnight at 4°C. The plate was then washed three times in PBST buffer. 100/d/well serial dilutions of isoproturon standards (100, 10, 1, 0.1, 0.01 and 0//g/L), diluted with distilled water, were added to the plate followed with the addition of 100/d/well isoproturon enzyme label diluted with PBS buffer. The plate was then incubated and shaken for one hour at 37°C. After incubation, the plate was washed three times with tap water, and 200/d/well TMB chromagen-substrate were added into all wells and left shaking for 30 minutes at 37®C. Finally, the reaction was stopped by the addition of 50/d/well IM HCl and the colour endpoint was read at 450nm as before. The standard curves were evaluated offline using a 4 parameter logistic plot (RiaCalc, Wallac Ltd).

3.8.1.3 Assay standard curve optimisation 3.8.1.3.1 Microtitre plates A number of microtitre plates were tested during the assay optimisation procedures. These included flexible fiat bottomed PVC, flat bottomed irradiated polystyrene Immuno-2-microtitre plates and polystyrene Maxisorb microtitre plates. The non-specific binding was determined by coating the wells with 0.07M sodium barbitone buffer pH 9.6 containing no antibody preparation. The assay was then performed as described above and the optical density in the wells was taken as the non-specific binding. Although all above microtitre plates displayed minimum non-specific binding, irradiated Immuno-2-microtitre plates (Dynatech)

Chapter 3 62 OD 450nm 2.8

2.4

1.6 A/S dilution curve —I— Pre-immune serum 1.2 —^ 0.5|jg/L Isoproturon

0.8

0.4

100 1000 10000 Isoproturon antiserum dilutions

Figure 3-9 Isoproturon antiserum dilution and displacement curves (antiserum 4676, bleed 1V:B) using ELISA.

Chapter 3 63 were chosen because of their high binding capacity (OD=2.1).

3.8.1.3.3 Coating antibody Four different dilutions of purified antibody solution (1/1000, 1/2000, 1/4000 and 1/8000) made up in sodium barbitone buffer pH 9.6 were added to a microtitre plate to construct four standard curves. Isoproturon enzyme label at 1/1000 dilution (1:40 molar ratio) was utilised with all standard curves. Figure 3-10 illustrates that 1/2000 dilution of the isoproturon coating antibody yielded the steepest standard curve with a high Bo reading.

3.8.1.3.2 Enzvme label The optimum dilution of isoproturon enzyme label was chosen applying serial dilutions (1/1000, 1/2000, 1/4000, 1/8000 and 1/16000) of enzyme label from each molar ratio (1:5, 1:10, 1:20 and 1:40 ”HRPO:hapten"), diluted with PBS buffer, onto standard curves constructed on one microtitre plate using 1/2000 dilution optimised above for coating antibody dilution. Figure 3-11 illustrates the effect of various dilutions of enzyme label with different molar ratios (HRPO:hapten) on the standard curves. Although a standard curve was achieved at each dilution and for each molar ratio, 1/4000 dilution (1:20 molar ratio) displayed a high optical density with the steepest sloped curve.

3.8.1.3.4 Incubation time Firstly, four isoproturon standard curves were constructed on four microtitre plates using the optimum dilutions of coating antibody and enzyme label. The plates were incubated for different initial coating times as follows: Plate 1 one hour incubation at 37°C. Plate 2 two hours incubation at 37°C. Plate 3 three hours incubation at 37°C. Plate 4 overnight (16 hours) incubation at 4®C. The incubation for the competing reaction between the labelled and unlabelled isoproturon was carried out for one hour at 4°C. Secondly, three isoproturon standard curves constructed on three plates were coated overnight with antibodies. Then the plates were incubated differently for the competing reaction between labelled and unlabelled isoproturon as follows:

Chapter 3 64 OD 450nm 2.8

2.4

— 1/1000 1.6 -4- 1/2000 1.2 1/4000 1/8000 0.8

0.4

0.001 0.01 0.1 1 10 100 1000 Isoproturon pg/L

Figure 3-10 Optimisation the isoproturon coating eintibocfy dilutions using ELISA. Legend shows dilutions of coating antibody.

Chapter 3 65 OD 450nm OD 450nm

2.5- 2.5 — 1/1000 1/1000 -4- 1/2000 —t— 1/2000 1.5: 1/4000 1.5 1/4000 -G- 1/8000 -B - 1/8000 1/16000 1/16000

0.5 0.5

0.001 0.01 0.1 1 10 1000100 0.001 0.01 0.1 10 100 10001 Isoproturon pg/L Isoproturon pg/L Molar ratio=l:40 Molar rati0=1:20

OD 450nm OD 450nm 1.8

1.6 1.5 1/1000 1/1000 -I- 1/2000 -4 - 1/2000 1/4000 0.9; 1/4000 0.8 -e- 1/8000 -B- 1/8000 1/16000 0.6 -44- 1/16000

0.4 0.3

0.001 0.01 0.11 10 100 1000 0.001 0.01 0.1 1 10 1000100 Isoproturon pg/L Isoproturon pg/L Molar rati0=1:10 Molar rati0=1:5

Figure 3-11 Optimisation of the isoproturon HRPO label for molar ratio and dilution using ELISA. Legends show dilutions of isoproturon HRPO label.

Chapter 3 66 Plate 1 one hour incubation at 37°C. Plate 2 two hours Incubation at 37°C. Plate 3 three hours incubation at 37°C. Thirdly, two isoproturon standard curves constructed on two plates were coated overnight at 4 ^ . Next day, the competing reaction was incubated for one hour at 37°C, although one plate was shaken on the Enfer shaker and the other plate was left without shaking. Figure 3-12 demonstrates the effect of different incubation conditions on the isoproturon standard curve. As a result, the incubation of the coating antibody overnight at 4 ^ accompanied by one hour incubation for the competing reaction at 37°C whilst the plate was shaking were the chosen conditions as the standard curve possessed the steepest slope under these conditions.

3.8.1.4 Assay validation 3.8.1.4.1 Cross-reactivitv Antiserum specificity results firom the action of a population of individual antibody molecules directed against different determinants on the antigen. When some of the determinants of the anti^n are shared by another compound then a proportion of the antibodies will also react with the second compound causing cross-reaction. Although cross-reacting molecules often bear similar chemical structures, there is evidence that antibodies recognise the overall configurations of antigens. It is possible that antibodies recognise particular three dimensional electron cloud shapes rather than (or as well as) specific chemical structures (Roitt and et al, 1986). Additionally, there is evidence that the specificity of an antiserum changes according to the conditions of the assay such as incubation times, pH, temperature and ionic strength of the buffer (Pudek and et al, 1985). This study was carried out to illustrate the specificity of the isoproturon antiserum as it is necessary to determine whether or not the antiserum is completely specific to its antigen, and if it is not, to what extent it cross-reacts with related structures. The specificity of the isoproturon antiserum was assessed by deriving cross-reactivity curves for different types of herbicide compounds. A series of standard curves were set up vdiereby one consisted of decreasing concentrations of isoproturon (100, 10, 1, 0.1, 0.01 and 0 fig/l.) whilst the others were composed of greater concentrations of compounds (KXkX), KXX), 1(X), 10, 1 and 0/^g/L)

Chapter 3 67 OD 450nm

1.61

1.2 —— 1 hr Incnbatloa 37C 2 hr Incubation 37C 0.8 — 3 hr Incubation 37C —B- Overnight(16 hr)4C

0.4

0.001 0.01 0.1 1 10 100 1000 Isoproturon pg/L Incubation time for coating antibody

OD 4S0nm

1.6

1.2 1 hr Incubation 37C 2 hr Incubation 37C 0.8 3 h r Incubation 37C

0.4

0.001 0.01 0.1 1 10 100 1000 Isoproturon pg/L Incubation time for competing reaction between labelled and unlabelled isoproturon

OD 450nm

1.6 1.2 Shaking 0.8 Non shaking

0.4 -

0.001 0.01 0.1 1 10 100 1000 Isoproturon pg/L The effect of shaking the plate during the competing stage on the standard curve

Figure 3-12 The effect of various incubation conditions on isoproturon standard curves (n=4).

Chapter 3 68 which might possibly cross-react. The same amounts of isoproturon label and antiserum were added to the plates. Percentage cross-reactivity was calculated from the weights of isoproturon and cross-reactants required to reduce the binding at zero standard concentration by 50%. In general, an antiserum which cross-reacts 100% with another compound would measure that compound equally well as the specific antigen. Cross-reactivity of < 1% implies a particular compound hardly combines with the antibody over the range studied and would not be likely to interfere in the assay for the specific antigen. Table 3-1 shows the degree of cross-reactivity of the isoproturon antibody with various herbicide compounds. Although not all urea herbicides have been tested for cross-reactivity, the antibody demonstrates remarkably high specificity towards isoproturon. It was not surprising that cross- reactivity with herbicides from different classes (i.e.atrazine) was minimal.

Table 3-1 Specificity of isoproturon antiserum towards a selection of herbicides.

C om pound Cross-reaction(%)

Isoproturon 100 Chlor toluron 1.187 Metoxuron 0.1 Chlorbromuron 0.085 Chlorsulfiiron <0.001 Metobromuron < 0.001 Monolinuron < 0.001 2,4-D <0.001 MCPA < 0.001 Atrazine <0.001 Simazine <0.001 Mecoprop < 0.001 Propyzamide < 0.001 Paraquat dichloride < 0.001 MCPB <0-001

Chapter 3 69 3.8.1.4.2 Recovery of isoproturon froTn different water samples The accuracy of measurement can be assessed by measuring the recovery of anatyte in authentic samples spiked with a known amount of anafyte. Water samples from different water sources (river water, lake water and tap water) were spiked with different concentrations of isoproturon stock solution (Ig/L) as shown in Table 3-2. All samples were measured against an isoproturon standard curve which was made up with distilled water. Table 3-2 demonstrates the inaccuracy of measurement in the water samples obtained by using ELISA. In addition, measurement of non-spiked samples proved to be non-immunoreactive.

T able 3 -2 The recovery of isoproturon from different spiked water samples using ELISA.

Isoproturon A m ount CV% Recovery a d d e d fig/L determined pg/L (n=6) W

University of Surrey lake 0.2 0.389 9.4 195 0.8 1.452 12.7 182 River Ouse 0.2 0.390 10.6 195 0.8 1.154 8.8 144 Tap water 0.2 0.289 9.3 145 0.8 1.079 6.9 135

3.8.1.4.3 Matrix effects It is possible that the actual ionic strength of the water samples shows individual variation and exerts an inconsistent effect on the binding of the antibody to the antigen which was different from the distilled water used to dilute the standards. Also the inaccurate recovery observed in Table 3-2 could be due to the presence of organic matter in the river samples which interferes with the antigen-antibody binding reaction. The effect of different matrices was

Chapter 3 70 investigated by comparing standard curves prepared with different solutions (tap water which had previously been shown to have no immunoreactivity, distilled water and PBS buffer of various ionic strengths). Although Figure 3-13 shows no significant matrix effect between the three sets of standards plotted as B/Bo, the observed difference in the Bo for the various matrices could have counted for the inaccurate recoveries. Tap water was chosen for diluting standards as its matrix most closefy resembles that of driiaking water supplies in general. Moreover, since accurate recovery of pesticides firom biological fluids has previously been observed (Hardcastle et al, 1990), the effect of adding normal sheep serum to the enzyme- labelled isoproturon solution was investigated. This was carried out by constructing four isoproturon standard curves, made up with tap water, on one plate. Each curve was obtained by adding the enzyme conjugate with different amounts of normal sheep serum (1%, 2%, 2.5%, and 3%). At the same time, four tap water samples were spiked with isoproturon standard solution at different concentrations to assess the recovery of isoproturon against one of the above standard curves. As shown in Table 3-3 the addition of 2.5% serum to the label solution of the assay caused significant improvement in the recovery of isoproturon firom the spiked samples accompanied with a slight increase in the Bo of the standard curve. Figure 3-14. Therefore, the final protocol adopted the use of tap water for diluting the standards and the addition of 2.5% normal sheep serum to the enzyme label solution. The water samples run in the assays were "authentic" and contained variable amounts of non homogenous suspended solids which may bind isoproturon to varying degrees and cause variation. Small sample volumes were used in the assay and so the effect of the suspension may have been considerable. This was investigated by running water samples in an assay, using the final optimised protocol, before and after centrifugation to settle the solids. No significant difference to the overall results was obtained.

Chapter 3 71 B/Bo% 100 90

60 Bo-1.48 Tap water 50 Dist. water 40 Bo-1.73 PBS buffer 30 20 10

0.001 0.01 0.1 1 10 100 1000 Isoproturon pg/L

Figure 3-13 The effect of different matrices used as dilution buffer on the Bo isoproturon standard curve.

Chapter 3 72 B/Bo% 1001 90

70 Bo-1.4 1% serum 60 2% serum Bo-1.7 2.5% serui 40 Bo-1.7 3% serum 30 20 10

0.001 0.01 1 10 100 10000.1 Isoproturon pg/L

Figure 3-14 The effect of adding normal serum to the isoproturon HRPO label solution on the Bo standard curve.

Chapter 3 73 T able 3 -3 Improvement of isoproturon recovery from tap water samples by the addition of normal serum to the enzyme label solution.

Isoproturon 1% seru m 2% seru m 2.5% seru m 3% seru m A d d e d pg/L Recovery% Recovery% Recovery% Reœvery% (n=6) (n=6) (n=6) (n=6)

0.15 147 131 109 110 0.5 130 135 113 105 0.7 151 128 92 113 1 119 123 97 108

3.8.1.4.4 Competitive isoproturon ELISA final protocol Using the information gained from the preceding experiments the following protocol was finally adopted: A microtitre plate was coated with 200;d/well of partially purified isoproturon antibody (0.65 ^g/ml) (sheep 4676 IV:B) diluted in 0.07M sodium barbitone buffer pH 9.6 at 1 /2000 dilution. The plate was incubated overnight at 4°C in a moist chamber and washed three times with PBST on the next day. Standard solutions (lOO/d per well) of isoproturon in the range 0-100 pg/h, made up in tap water, were applied to the plate followed by the addition of 100/^1/weU HRPO -labelled isoproturon at 1 /4000 dilution (0.575 p g /v o l peroxidase), diluted in PBS buffer which contains 2.5% normal serum sheep, to bring the total volume to 200/il/well. The plate was incubated in a shaker incubator (Enfer) for one hour at 37°C. After incubation, the plate was washed four times with tap water and 200/d per well of chromagen-substrate TMB were placed in all the wells. Finally, after a further 30 minute incubation the reaction was stopped by the addition of IM HCl (50/d/well) and the plate was read using the Labsystems Multiskan plate reader at 450 nm and the standard curves were evaluated off-line using a four- parameter logistic plot (RiaCalc, Wallac, Milton Keynes, UK).

Chapter 3 74 3.8.1.4.5 Recovery of isoproturon from different water samples using the final protocol Different water sample sources (River Ouse water. University of Surrey campus lake water. University of Surrey Manor Farm site pond water, Surrey research park site lake water. River Wey Guildford and tap water) were spiked with different concentrations of isoproturon stock solution and the percentage recovery of isoproturon was determined using the ELISA protocol. Table 3-4 demonstrates an improvement in the percentage recovery of isoproturon (81- 110%) from various water sources making a reproducible and quantitative recovery possible.

Chapter 3 75 Table 3 -4 Recovery of isoproturon from different water sources by using ELISA (final protocol).

Isoproturon A m ount cv% Recovery a d d e d pglL determined pg/L (ru6) W

Tap water (day 1) 0.2 0.184 2.1 92 0.8 0.651 4.2 81 Tap water (day 2) 0.2 0.162 9.1 81 0.8 0.736 7.2 92 Tap water (day 3) 0.2 0.187 5.9 94 0.8 0.814 8.7 102 River Ouse 0.2 0.162 8.4 81 0.8 0.845 7.7 106 River 0.2 0.169 5.1 85 0.8 0.771 6.9 96 University lake 0.2 0.182 7.5 91 0.8 0.714 4.9 89 University Manor Farm 0.2 0.213 4.8 107 0.8 0.841 6.2 105 Research Park lake 0.2 0.170 3.6 85 0.8 0.871 5.1 109

Chapter 3 76 3.8.1.4.6 Assay sensitivitv and reproducibility a) Intra-plate variation The intra-assay variation was determined for the eight standard isoproturon solutions (100, 10, 1, 0.5, 0.1, 0.05, 0.01 and 0 pg/L) by constructing six standard curves on one microtitre plate using the final protocol. The mean and range of values obtained (expressed as a percentage of the maximum binding B/Bo) for each isoproturon standard are shown in Table 3-5. The detection limit of the assay, defined as the concentration equivalent to a three standard deviation (SD) fall from binding at Bo (the binding measured in the absence of isoproturon), was 0.03pg/L, well below the EC MAC for pesticides in drinking water. Figure 3-15 shows the intra-assay variation between isoproturon standard curves which ranged from 2.6-5.6% across the curves.

Table 3 -5 Intra-plate variation for isoproturon standard curves using ELISA.

Isoproturon standard M ean(n=6) SD CV% pgIL BfBo%

100 5 ±0.13 2.6 10 10 ±0.37 3.7 1 30 ±1.7 5.6 0.5 45 ±2.4 5.3 0.1 67 ±3.01 4.5 0.05 79 ±3.9 4.9 0.01 90 ±4.8 5.3 0 (Bo=1.57) 100 ±3.9 3.9

The intra-assay variation for the standard curves were also measured by two colleagues (one was skilled in immunoassay techniques whereas the other was a novice with immunoassay) following the steps described in the final protocol. The results obtained from the skilled colleague show high reproducibility (CV%) ranging from 2.3-6.7%, whereas the data produced by the

Chapter 3 77 non skilled colleague showed CV% between 6.2-8.1% which is still acceptable. Figure 3-16.

b) Inter-assau variation The between-plate variation was determined for eight standard isoproturon solutions by constructing six standard curves, each in duplicate, separated onto six microtitre plates. All plates were treated in the same way but executed on different days. Table 3-6 demonstrates the variation of isoproturon standard curves performed on different microtitre plates which ranged from 7.2-15.1%.

Table 3-6 Inter-plate variation for isoproturon standard curves.

Isoproturon standard M ea n (n -6 ) SD CV% pg/L B/Bo%

100 13 ±1.2 9.2 10 33 ±2.4 7.2 1 53 ±5.7 10.7 0.5 61 ±5.9 9.7 0.1 70 ±8.2 11.7 0.05 88 ±10.9 12.4 0.01 96 ±14.3 14.9 0 (Bo=1.32) 100 ±15.1 15.1

3.8.1.4.7 Stabilitv and preparation of standards To further minimise day to day variation, individual standard solutions were frozen over a period of one month to eliminate variation due to daily preparation of standards. Each individual standard (100, 10, 1, 0.1, 0.01 and 0 pg/h) was prepared in tap water, aliquoted into tubes in volumes sufficient for one assay and frozen. They were thawed at room temperature immediately before use in the assay. This was compared with standard solutions made up freshly every day hi tap water and standard solutions made up with tap water before and after adjusting the pH of the water to pH 5 (since isoproturon is more stable in

Chapter 3 78 B/Bo% CV% 50 100 45 90 40 Bo-1.57 35 70 30 60 50 25 40 20 30 15 20 10

0.001 0.01 1 100.1 100 1000 Isoproturon pg/L

Figure 3-15 Intra-plate variation of isoproturon standard curve with precision profile using ELISA. ( I ), standard deviation; (-X-), mean of six standards; (-A-), coefficient of variation.

Chapter 3 79 (a)

B/Bo% CV% 110 50 100 45 90 40 80 Bo-1.4 70 30 60 50 20 40 30 20

0.001 0.01 0.1 1 10 100 1000 Isoproturon pg/L

(b)

B/Bo% CV% IIO jt 50 100 45 9 0 - 40 80 Bo-1.6 70 30 60 25 50 20 40 30

0.001 0.010.11 10 100 1000 Isoproturon pg/L Figure 3-16 Isoproturon standard curves constructed by colleagues with varied experience in immunoassay [(a) skilled colleague, (b) non skilled colleague]. ( 1 ), standard deviation; (-X-), mean of six standards; (-A-), coefficient of variation.

Chapter 3 80 acidic media (Merck Index, 1989), stored at 4 ^ to be used randomfy during the one month study period. Table 3-7 shows the statistical evaluation of the assays using these different prepared standard solutions. The CV% of standard solutions made up with tap water pH 5 revealed high reproducibility with CV% ranging from 2 to 5%. Furthermore, standards made up daily with normal tap water (pH 7.8) also showed good stability with a variation range of 3 to 6.9%. On the other hand, the use of frozen standard solutions caused a significant increase in the CV% between standard curves (11-15%) over the investigation period. Therefore, standard solutions for all assays were made up with tap water pH 5 in large volumes and stored at 4°C for future use.

T ab le 3 7 The stability of standard solutions in different storage and pH conditions.

Standard Frozen Tap w a ter Tap w a ter Tap w a te r pH 7.8 St pH 7.8 St pH 7.8 St pH 5 St (1 month) (daily made) (1 month 4^C) (1 month 4^0 mean CV% mean CV% mean CV% mean CV% B/Bo% (n=6) B/Bo% (n=6) B/Bo% (n=6) B/Bo% (n=6)

100 18 11 11 4.2 16 7.9 9 3.6 10 42 13 29 3.3 39 9.7 25 4.4 1 66 15 41 6.2 61 10.1 38 4.1 0.1 82 13 65 6.9 79 11.4 68 2.7 0.01 97 10 81 5.1 92 14 86 4.9 0 100 9 100 6.8 100 10.4 100 5

3.8.1.4.8 Measuring isoproturon in biological fluids using ELISA The widespread use of isoproturon may cause accidental or deliberate ingestion or gradual increase in body load with chronic use. An immunoassay is a convenient diagnostic and monitoring aid in the case of poisoning or to detect occupational exposure levels. Also the amount of plasma required by immunoassay (100/d) is possible to collect e.g. from finger pricks rather than

Chapter 3 81 venipuncture. Six isoproturon standard solutions (100, 10, 1, 0.1, 0.01 and Qfig/h) made up in normal human plasma and stored at 4°C were used to assess the intra and inter-assay variation over a period of 10 days. In addition, plasma samples spiked with different concentrations of isoproturon sub-stock solution (0.1, 0.5 and 1 /fg/L) were measured with each assay standard curve. Table 3-8 demonstrates the intra and inter assay variation of isoproturon in plasma across the standard curve and the percentage recovery of isoproturon from plasma spiked samples over a period of 10 days. Furthermore, Figure 3-17 shows significant increase in the optical density of the zero (presence of label only) accompanied with a slight improvement of the sensitivity. This would allow further reduction in the amount of enzyme label applied onto the standard curve and thus enhance the sensitivity of the assay even further.

Table 3 -8 Intra and inter-assay variation of isoproturon standard curves in plasma and the recovery of isoproturon from spiked plasma samples.

Isoproturon Within-assay variation Between-assay variation S ta n d a rd fig/L Mean SD CV% Mean SD CV% B/Bo% (n=6) B/Bo% (n=6)

100 6 ±0.3 5 9 ±0.9 10 10 14 ±0.8 5.7 19 ±2.4 12.6 1 30 ±2.2 5.6 40 ±5.1 12.8 0.1 57 ±3.5 6.5 61 ±7.3 11.9 0.01 79 ±4.8 6.1 81 ±8.7 10.7 0.(Bo=2.1) 100 ±6.3 6.3 100 ±9.8 9.8

Isoproturon a d d e d Amount determined CV% Recovery ygIL Mean(yg/L) SD (n=5) W

Plasma 0.1 0.124 ±0.013 10.5 124 0.5 0.433 ±0.036 8.3 86.6 1 1.034 ±0.075 7.3 103.4

Chapter 3 82 Similarly, urine was used as a matrix for standard curves to demonstrate within and between-assay variation over a similar period (10 days) and to monitor the recovery of isoproturon from spiked urine samples as shown in Table 3-9. However, the use of urine has shown to decrease the optical density of the zero standard curve and the sensitivity of the assay significantly. Figure 3-17.

Table 3 -9 Intra and inter-assay variation of isoproturon standard curves in urine and the recovery of isoproturon in spiked urine samples.

Isoproturon Within-assay variation Between-assay variation S ta n d a rd fig/L M ean SD CV% M ean SD CV% B/Bo% (n=6) BIBo% (n=6)

100 14 ±1.2 8.6 15 ±1.7 11.3 10 33±2.2 6.7 35 ±4.2 12 1 52 ±3.9 7.5 51 ±7.5 14.7 0.1 75 ±3.7 4.9 77 ±6.3 8.2 0.01 94 ±3.2 3.4 93 ±10.4 11.1 0 (Bo=1.02) 100 ±5.9 5.9 100 ±15.8 15.8

Isoproturon added Amount determined CV% Recovery jugiL M ean (fig/L) SD (n=5) (%)

Urine 0.1 0.119 ±0.013 10.9 119 0.5 0.451 ±0.044 9.8 90.2 1 0.976 ±0.075 7.7 97.6

Chapter 3 83 (a)

B/Bo% CV% 110

100 Bo-2.1 45 90 40 80 35 70 30 60 25 50 20 40 30 20

0.001 0.01 0.1 10 100 10001 Isoproturon pg/L (b)

B/Bo% CV% 110 50 100 45 90 40 80 Bo-1.02 35 70 30 60 25 50 20 40 30 15 20 10 10

0.001 0.010.1 10 1001 1000 Isoproturon p g/L Figure 3-17 Intra-plate variation of isoproturon standard curves constructed in (a) plasma and (b) urine with precision profiles. ( I ), standard deviation; (-X-), mean of six standards; (-A-), coefficient of variation.

Chapter 3 84 3.8.1.4.8.1 Detection of isoproturon (or metabolites) in mice The potential of the isoproturon ELISA method for monitoring isoproturon in biologiCcd fluids was further investigated by injecting 22 mice (DBA2 males) simultaneously with lOmg/Kg (0.2mg/mouse) isoproturon standard solution made up with distilled water from the stock standard (lOmg/ml with methanol) using the intra-peritoneal (IP) route for injection. The protocol was carried out by harvesting two mice at time intervals during a 48 hour period. The blood (1- 1.5ml) was collected in 2ml Eppendorf tubes (each mouse was giving 1-1.5ml of blood) and left for 2 hours to clot. The clotted blood samples were then centrifuged at 2000 rpm and the serum supernatants were transferred onto 1ml Eppendorf tubes and stored at -40°C until the next stage of assessment by ELISA. Isoproturon standard solutions were diluted in adult male mouse serum to construct a standard curve while serum samples collected as described were diluted with adult male mouse serum (1/10, 1 /100, 1 / 16(X) and 1 /3200) to allow the detection of any high concentrations of isoproturon present in early time periods. The final protocol of the isoproturon standard curve, described in 3.8.1.4.4, was adopted and the concentrations of the unknown serum samples were obtained from the standard curve automatically using RiaCalc and corrected for the dilutions used. Table 3-10 shows the amount of "immunoreactive" isoproturon (or metabolites) detected by the ELISA system. Significant elimination of isoproturon (or metabolites) from the mice was observed eight hours after injection of isoproturon standard. Figure 3-18. Therefore, the isoproturon ELISA method has shown potential for monitoring isoproturon (or its related metabolites) in different biological fluids. Isoproturon (or its metabolites) could be detected in serum for a period of hours after exposure which would depend on the dose ingested and could be useful for occupational exposure monitoring.

Chapter 3 85 Table 3 10 Amount of "immunoreactive" isoproturon measured ty ELfôA.

Harvesting Time Isoproturon/Metabolites Dilution Minute Amount Detectedfig/L (n=2) Required

NO 0 None 5 1760 1/3200 10 866 1/3200 15 181 1/100 30 102 1/100 60 15.3 1/100 120 4.4 1/10 240 1.3 1/10 480 0.14 1/10 1440 0.01 None 2880 0 None

3 .8 .2 Chlortoluron ELISA 3.8.2.1 Evaluation of the purified antiserum (titre and displacement) A direct competitive ELISA method as described in 3.8.1 was also adopted for another urea herbicide, chlortoluron. Antiserum dilution and displacement curves were constructed on one microtitre plate as follows: 200/d/well of six serial dilutions of chlortoluron partially purified antiserum (1/500, 1/1000, 1/2000, 1/4000, 1/8000 and 1/16000) diluted in sodium barbitone buffer pH 9.6 were added into a microtitre plate in duplicate and incubated in a moist chamber overnight at 4°C. At the same time, similar dilutions to the above were utilised on the same plate using the pre-immune serum for determining the specific binding of the chlortoluron antiserum. Next day, the plate was washed three times with PBST to remove any excess of uncoated antibodies. For all antiserum dilution curve wells, 100/d/well PBS buffer were applied into the plate followed with the addition of another 100/^1/well chlortoluron enzyme label (molar ratio 1:20) diluted with PBS buffer at 1/1000 dilution (3.1/4g/ml peroxidase) and containing 2.5% normal human

Chapter 3 86 Isoproturon pg/L 10000

1000

100 =

0.1

0.01

0.001 5 15 240 480 1440 2880 Time (minute)

Figure 3-18 The elimination of "immunoreactive" isoproturon from mice injected with isoproturon and measured ELISA.

Chapter 3 87 serum to bring the total volume of each weU to 200/fl. On the other hand, for all displacement curve wells 100^1/well of 0.5/(g/L chlortoluron standard solution were applied into the m icrotitre plate followed with the addition of lOO/zl/weU of the same dilution of the chlortoluron enzyme label used above. The plate was then incubated in the Enfer shaker for one hour at 37°C. Finally, the plate was washed three times with tap water and 200/d/well TMB chromagen-substrate solution were added to plate and left shaking for 30 minutes at 37°C. The colour developed was halted by adding 50/d/well IM HCl and read at 450 nm using Labsystem microtitre plate reader. Figure 3-19 shows high binding of the antiserum to HRPO and significant displacement by 0.5 /(g/L chlortoluron standard.

3.8 2.2 Assay standard curve optimisation The assay standard curve procedure was carried out as described in 3.8.1.2 while all assay conditions optimised for isoproturon were used for the chlortoluron assay.

3.8.2.2.1 Coating antibody Four different dilutions of chlortoluron purified antibody solution (1/100, 1/2000, 1/4000 and 1/8000) made up with sodium barbitone buffer pH 9.6 were applied on a microtitre plate to perform four standard curves. A 1 /1000 dilution of chlortoluron enzyme label (1:20 molar ratio) was added to all standard curves. Figure 3-20 shows that the standard curve for 1/8000 dilution of the coated antibody possessed the steepest curve.

3.8.2.2.2 Enzyme label Several dilutions of chlortoluron enzyme label (1/10(X), 1 /2000, 1/4(X)0, 1/8000 and 1/16000) from each molar ratio prepared previously (1:5, 1:10, 1:20 and 1:40 "HRPO:hapten") were used in the assay for standard curves constructed on one plate which was coated with the optimum dilution of the coating antibody 1/8000 (3.8.2.2.1). The addition of 1/16000 dilution of the chlortoluron enzyme label (1:20 molar ratio) into the chlortoluron assay produced the standard curve with the steepest slope as demonstrated in Figure 3-21.

Chapter 3 88 OD 450nm 2.4

1.6 — A/S dilution curve 1.2 —I— Pre-immune serum —^ 0.5pg/L chlortoluron 0.8

0.4

100 1000 10000 Chlortoluron antiserum dilutions

Figure 3-19 Chlortoluron antiserum dilution and displacement curves (antiserum 1826, bleed V:C) using ELISA.

Chapter 3 89 OD 450nm 2.8

2.4

1/1000 1.6 1/2000

1.2 ^ 1/4000 -a - 1/8000 0.8

0.4

0.001 0.01 0.1 1 10 100 1000 Chlortoluron pg/L

Figure 3-20 Optimisation of the chlortoluron coating antibody dilutions using ELISA. Legend shows dilutions of chlortoluron coating antibody.

Chapter 3 90 OD 450nm 2.8 OD 450nni

2.8 2.4 2.4

1/1000 — 1/1000 —t— 1/2000 1.6 - 4 - 1/2000 1.6 1/4000 1/4000

-B - 1/8000 - s - 1/8000

- X - 1/16000 1/16000 0.8 0.8 -

0.4 0.4 -

0.001 0.01 0.1 1 10 100 1000 0.001 0.01 0.1 1 10 100 1000 Chlortoluron (jg/L Chlortoluron pg/L Molar rat]o=l:40 Molar ratlo=l:20

OD 450nm OD 450nm

1.6 1.6 -

1/1000 1/1000 —t— 1/2000 — I— 1/2000 1/4000 1/4000 0.8 -B- 1/8000 0.8 1/8000 -)<- 1/16000 1/16000

0.4 0.4

0.001 0.01 1 10 1000.1 1000 0.001 0.01 0.11 10 100 1000 Chlortoluron pg/L Chlortoluron pg/L Molar ratio=l:10 Molar ratio=l:S

Figure 3-21 Optimisation of the chlortoluron HRPO label for molar ratio and dilution using ELISA. Legends show dilutions of chlortoluron HRPO label.

Chapter 3 91 3.8 2.3 Assay validation 3.8.2.3.1 Cross-reactivitv The specificity of chlortoluron antiserum was assessed by deriving cross reactivity curves for different pesticide compounds. The procedure was followed as explained in 3.8.1.4.1. Table 3-11 demonstrates the degree of cross-reactivity of chlortoluron antibody with various pesticides which reveals different percentages of cross-reactivity with various phenylurea herbicides. As expected, structurally unrelated compounds i.e. triazines did not show any sign of cross-reactivity.

Table 3-11 Specificity of the chlortoluron antiserum towards a selection of herbicides.

Compound Cross-reajctton(%)

Chlortoluron 100 Chlorbromuron 71 Isoproturon 47 Metoxuron 8.8 Metsulfuron 3.45 Chlorsulfuron 1.3 Metamitron < 0.001 2,4-D <0.001 MCPA < 0.001 Atrazine <0.001 Simazine < 0.001 Mecoprop <0.001 Propyzamide < 0.001 Paraquat dichloride <0.001 Terbutiyn < 0.001 MCPB <0.001

Chapter 3 92 3.5.2.3.2 Competitive chlortoluron ELISA final protocol Using the information gained from the preceding experiments the following protocol was finally adopted as follows: A microtitre plate was coated with 200/d per well (0.17 //g/ml) of partially purified (by ion exchange) chlortoluron antibody (sheep 1826 V:C) diluted in 0.07M sodium barbitone buffer pH 9.6 (1/8000 dilution). The plate was incubated overnight at 4PC in a moist chamber and washed three times with PBST on the next day. Standard solutions (100/^1 per well) of chlortoluron in the range 0-100 jWg/L, made up in tap water, were applied to the plate followed by the addition of 100//1 per well HRPO-labelled chlortoluron at 1/16000 dilution (0.13 fig/ml peroxidase), diluted in PBS buffer containing 2.5% normal serum sheep, to bring the total volume to 200/^1/well. The plate was incubated in a shaker incubator (Enfer) for one hour at 37°C. After incubation, the plate was washed four times with tap water and 200/d per well of TMB chromagen-substrate was placed in all the wells. Finally, after a further 30 minute incubation in the shaker the reaction was stopped by the addition of IM HCl (50/d per well) and the plate was read using the Labsystems Multiskan plate reader at 450 nm. Standard curves were evaluated off-line using a four-parameter logistic plot (RiaCalc, Wallac, Milton Keynes, UK).

3.8.2.3.3 Recovery of chlortoluron from different water samples using the final protocol A range of water samples obtained from various sources (3.8.1.4.5) were spiked with different concentrations of chlortoluron stock solution and measured by chlortoluron ELISA method to detect the percentage recovery of the spiked chlortoluron in each sample. Table 3-12 shows the recovery percentage of chlortoluron from different water matrices ranging between 78-127%.

3.8.2.3.4 Assay sensitivitv and reproducibility a) Intra-assaii variation The intra-plate variation between chlortoluron standard curves (100, 10, 1, 0.5, 0.1, 0.05, 0.01 and 0 fig/L) performed on one microtitre plate was determined as described in the final protocol procedure. Table 3-13 demonstrates the intra-assay variation between chlortoluron standard curves which ranged from 2.7 - 6.5% across the curves. Also, the detection limit of of the chlortoluron

Chapter 3 93 Table 3 12 Percentage recovery of chlortoluron from different water sources assayed using the chlortoluron ELISA (final protocol).

Chlortoluron A m ount CV% Recovery a d d e d fig/L determined fig/L (n=6) (%)

Tap water (day 1) 0.2 0.162 4,3 81 0.8 0.691 5.7 86 Tap w ater (day 2) 0.2 0.209 11.2 105 0.8 0.703 8.2 88 Tap water (day 3) 0.2 0.156 7.8 78 0.8 0.810 6.7 101 River Ouse 0.2 0.223 9.3 112 0.8 1.012 6.4 127 River Wey 0.2 0.194 7.9 97 0.8 0.878 5.3 110 University lake 0.2 0.207 4.8 104 0.8 0.874 5.1 109 University Manor Farm 0.2 0.223 6.5 112 0.8 0.821 8.4 103 Research Park lake 0.2 0.177 4.4 89 0.8 0.787 7.3 98

Chapter 3 94 assay, Figure 3-22, was 0.015 /^g/L well below the EC MAC for pesticides in drinking water.

b) Inter-assau variation Six chlortoluron standard curves (100, 10, 1, 0.5, 0.1, 0.05, 0.01 and 0 fig/L) were constructed in duplicate during a period of 10 days to assess the variation of the curve between assays. Table 3-14 illustrates the inter-plate variation of the standard curves which ranged from 6.9-12.4%.

3.5.2.3.5 Stabilitv and preparation of standards Chlortoluron standard solutions made up with tap water, pH 5, and stored at 4°C revealed high stability within a period of one month. This was detected by constructing six isoproturon standard curves randomly during this period. The limit of detection of the mean standard curve was 0.02 fig/L and the precision of the assay ranged between 3.2-7.1% (Figure 3-23).

3.5.2.3.6 Measuring chlortoluron in biological fluids using ELISA Six chlortoluron standard solutions (100, 10, 1, 0.1, 0.01 and 0 fig/L) were made up in normal human plasma and urine and stored at 4®C for further investigation. The determination of the intra and inter-variation for each biological fluid was carried out exactly as described in 3.8.1.4.8 and illustrated in Tables 3-15, 3-16 with all percentage recovery of the spiked samples. Furthermore, the use of plasma in the standard curve assay increased the optical density of the zero significantly but caused no improvement to the sensitivity of the assay as shown in Figure 3-24.

Chapter 3 95 B/Bo% CV% 50 ICO 45 90 40 80 35 70 30 60 25 50 40 20 30 15 20 10

0.001 0.01 0.1 1 10 100 1000 Chlortoluron pg/L

Figure 3-22 Intra-assay variation of chlortoluron standard curve with a precision profile using ELISA. ( 1 ), standard deviation; (-X-), mean of six standards; (-A-), coefiOcient of variation.

Chapter 3 96 Table 3-13 Intra-plate variation between chlortoluron standard curves using ELISA.

Chlortoluron standard M ea n (n -6 ) SD CV% fig/L B/Bo%

100 5 ±0.2 5 10 6 ±0.16 2.7 1 11 ±0.7 6.4 0.5 15 ±0.8 5.3 0.1 19 ±1.04 5.5 0.05 41 ±1.6 3.9 0.01 63 ±4.1 6.5 0(Bo=1.91) 100 ±4.9 4.9

Table 3-14 Inter-plate variation of chlortoluron standard curves using ELISA.

Chlortoluron standard M ean(n=6) SD CV% fig/L B/Bo%

100 7 ±0.8 11.4 10 9 ±1.09 12.1 1 12 ±0.96 8 0.5 16 ±1.1 6.9 0.1 21 ±1.8 8.6 0.05 47 ±5.1 10.8 0.01 66 ±8.2 12.4 0 (Bo=1.59) 100 ±9.3 9.3

Chapter 3 97 B/Bo% CV% 110 50 100 45 Bo-1.41 90 40 35 70 30 60 25 20 15 30 20 10

0.001 0.01 0.1 1 10 100 1000 Chlortoluron pg/L

Figure 3>23 Precision of chlortoluron standard curve with chlortoluron standard solutions at pH 5, stored at 4°C, over a period of one month. ( I ), standard deviation; (-X -), mean of six standards; (-A-), coefficient of variation.

Chapter 3 98 Table 3-15 Intra and inter-assay variation of chlortoluron standard curves in plasma and the recovery of chlortoluron from spiked plasma samples.

Chlortoluron Within-assay variation Between-assay variation S ta n d a rd fig/L M ean SD CV% M ean SD CV% B/Bo% (n=6) B/Bo% (n=6)

100 4 ±0.17 4.3 6 ±0.5 8.3 10 8 ±0.54 6.8 10 ±1.1 11 1 13 ±0.72 5.5 20 ±2.5 12.5 0.1 20 ±0.98 4.9 28 ±4.1 14.6 0.01 59 ±2.6 4.4 64 ±7.3 11.4 0 (Bo=1.78) 100 ±5.1 5.1 100 ±8.9 8.9

Chlortoluron added Amount determined CV% R ecovery fig/L M ean (fig IL) SD (n=5) W

Plasma 0.1 0.113 ±0.01 8.8 113 0.5 0.537 ±0.06 11.2 107 1 0.945 ±0.08 8.5 95

Chapter 3 99 (a) B/Bo% CV% 110 1 50

100 ) 45 Bo-1.91 90 40 80 70 30 60 25 50 20 40

20 - 10 10^^

0.001 0.01 0.1 1 10 100 1000 Chlortoluron pg/L

(b)

B/Bo% CV% 110 50 100 45 Bo-0.89 90 40 80 70 30 60 25 50 20 40 30 20 10

0.001 0.01 0.1 1 10 100 1000 Chlortoluron pg/L

Figure 3-24 Chlortoluron standard curves constructed in (a) plasma and (b) urine with precision profiles. ( I ), standard deviation; (-X-), mean of six standards; (-A-), coefficient of variation.

Chapter 3 100 Table 3-16 Intra and inter-assay variation of chloitoluron standard curves in urme and the recovery of chlortoluron from spiked urine samples.

Chlortoluron Within-assay variation Between-assay variation Standard pug/L Mean SD CV% Mean SD CV% B/Bo% (n=6) B/Bo% (n=6)

100 5 ±0.13 2.6 4 ±0.42 10.5 10 11 ±0.6 5.5 12 ±1.6 13.3 I 15 ±1 6.6 18 ±2.6 14.4 0.1 24 ±1.1 4.6 30 ±3.9 13 0.01 66 ±3.3 5 62 ±5.8 9.4 0 (Bo=0.89)l 00 ±4.9 4.9 100 ±11.1 11.1

Chlortoluron added Amount determined CV% Reœvery jug/L M ean (fig/L) SD (n=5) W

Urine 0.1 0.87 ±0.097 11.1 87 0.5 0.521 ±0.05 9.6 104 1 0.934 ±0.1 10.7 93

3.9 DISCUSSION The ELISA techniques developed and described here for isoproturon and chlortoluron, proved to be rapid, sensitive and specific, fulfilling the needs of present legislation relating to the use and levels of pesticides in the environment. Isoproturon and chlortoluron which are both commonly used as herbicides were chosen because of their widespread use in the environment to control grasses and so far no simple and sensitive methods are available for detecting these phenylurea herbicides at low concentrations. Although Li'egeois and his colleagues have successfully developed indirect competitive ELISA methods for measuring isoproturon in drinking water and soiL the assays were not sensitive enough {Ifig/L) to detect the MAC level of the herbicide in drinking water

Chapter 3 101 (Li'egeois and et al, 1991) and the limit of detection of isoproturon in soil was only between 20-250yMg/L (Li'egeois and et al, 1992). The ELISA systems described in this chapter were based on a direct competition between the enzyme labelled and unlabelled related herbicides offering a speedy, simple and highly sensitive system for monitoring pesticides in the environment.

3 .9 .1 A n tisera The two phenylurea herbicides (isoproturon and chlortoluron) are of small molecular weight (between 207-213 daltons), rendering them non- immunogenic compounds. As such, they require covalent binding to large protein molecules in order for them to be recognised as antigenic by the animals immune system. However, the chemical structures of isoproturon and chlortoluron do not possess any suitable functional group (i.e. OH, COOH, NH 2) for conjugation proteins, and therefore, the addition of one of these functional groups was necessary to enable binding with carrier proteins. The high specificity of isoproturon and chlortoluron antibodies was achieved by covalently coupling thyroglobulin with carboigrl groups synthesised on the secondary amino groups on the haptens. It is indicated ly Landsteiner's Principle that antibody specificity is directed primarily towards that portion of the hapten furthest removed firom the functional group used to link it to the carrier protein (Landsteiner, 1945). The two immunogens were injected into four mature Suffolk sheep which were boosted five times at regular intervals to enhance the immune responses. All animals produced a significant immune response following the primary injection except sheep 1826, immunised with the chlortoluron-thyroglobulin immunogen, which gave a low immune response until the fifth boost when the titre and the avidity improved dramatically. No improvement in the immune response sheep 4676 (isoproturon) was observed after the fourth boost. Antisera were screened using an RIA technique using radio labels because of the commercial availability of the labels. Negative results would be regarded as negative immune response rather than a label preparation problem (i.e. if ELISA is used there is always uncertainty that the enzyme label is immunogenic). Bleeds IV:B (fourth boost, second bleed) firom sheep 4676 and V:C (fifth boost, third bleed) sheep 1826 displayed acceptable titres and significant displacements with 0.5 fig/L of the relevant herbicide. Thus, they were used for

Chapter 3 102 assay developments after being partially purified on an ion exchange disc as explained in 3.9.2.

3 .9 .2 Assay optimisation In the development of an ELISA, optimisation of the coating antibody dilution, and enzyme label concentration were essential. Much time was taken in assay optimisation and validation in an attempt to attain the most sensitive and robust assay. The choice of microtitre plate depended on personal preference and cost. PVC plates were reported to possess a great capacity for protein (De Savigny and Voiler, 1980) (Kemeny and Challacombe, 1988), although they were not reported to bind more than the irradiated polystyrene plates now available (Urbanek and et al, 1985). It was observed in this study that the irradiated polystyrene Immuno-2 -microtitre plates generally possessed high binding capacity which supported the finding of Urbanek (1985). Thus, polystyrene Immuno-2-microtitre plates were used throughout the assay developments. Furthermore, the hnmuno- 2-microtitre plate compared with Maxisorb showed marginally improved standard curves as determined ty assay sensitivity. The high sensitivity of the isoproturon standard curve was observed when the coating antibody was incubated overnight at 4°C (16 hrs). IgG immunoglobulins show increased adsorption stability on polys^ene surface at low temperature (Green, 1991). Also, incubating the competing reaction between the related label and unlabelled herbicide and shaking at 37°C for one hour produced high optical density with high sensitivity (i.e. steepest slope) for the standard curve. One hour incubation at 37°C for the competing step w^as enough to reach the equilibrium between the labelled and unlabelled isoproturon for the antibody binding sites whereas increasing the incubation time proved to show low optical density which was probably due to the dissociation of the bound fraction from the antibodies. The performance of an immunoassay is directly dependent on two key reagents, the antiserum with the desired specificity towards the analyte and the immimoreactive labelled form of that analyte (Aheme, 1984). The most important requirement for a coating antibody is that once bound to the solid phase, it should have a high binding capacity for the relevant antigen (Voiler and et al, 1978). The use of whole sheep antiserum for coating resulted in low binding of

Chapter 3 103 both the isoproturon and enzyme conjugates. This was probably due to competitive binding and the possible interference of many other proteins present in serum (Kemeny and Chantier, 1989). Consequently, ion exchanged purified antisera were utilised for all assay developments. As with any assay, optimisation of the capture (coating) antibody concentration was essential for a sensitive and reproducible assay. Low concentrations of herbicides could not be detected if the coating antibocfy concentration was too high or too low (Engvall and Perlmann, 1972). The mechanism of protein adsorption to plastic surface is not fully understood, but is known to be reversible. It therefore followed that excess capture antibody adsorption would increase protein loss and reduce assay sensitivity (Engvall and et al, 1971). The optimum antibody concentrations for isoproturon and chlortoluron were 0.65 yUg/ml (1/2000 dilution) and 0.174 ^g/ml ( 1 /8000 dilution) respectively determined by checkerboard assays. The other essential reagent for immunoassay is the label. The N- hydrojtysuccinimide active ester method of enzyme conjugation produced satisfactory labels for isoproturon and chlortoluron which retained similar enzyme activity for at least six months. Besides, the active ester method provides a convenient way to monitor the success of the activation by producing clear crystals which are easy to observe. Several molar ratios (HRPO:hapten) were carried out to determine the optimum ratio between the hapten and the enzyme and the optimum enzyme label dilutions for isoproturon and chlortoluron were 1/4000 (0.58 )Mg/ml peroxidase) molar ratio 1:20 and 1/16000 (0.13 ^wg/ml peroxidase) molar ratio 1:20 respectively determined Isy checkerboard assays.

3 .9 .3 A ssa y vaU dation Validation of an assay was essential to determine how robust the assay was under routine use. Each antiserum was characterised for specificity towards closely related herbicides using the method described by Dr Aheme (Aheme and Marks, 1979). Although not all urea herbicides have been determined for cross reactivity, the isoproturon antibody illustrates remarkable specificity by showing no cross-reactivity with a lar^ range of pesticides even with the closely related urea herbicide chlortoluron. On the other hand, the chlortoluron antibody displays significant percentage cross-reactivity with various phenylurea herbicides (i.e. chlorbromuron, isoproturon, Metoxuron). This lack of specificity towards some phenylureas could turn into an advantage, in contrast to the

Chapter 3 104 clinical application of some Immunoassays vhere absolute specificity for a single compound is required, by utilising the antibody for screening purposes for phenylurea herbicides in general. This is an advantage of immunoassay using non-specific antisera in forensic toxicology, e.g. where a negative result excludes a range of pesticides from finther investigation while a positive result indicates the class of compound present and gives an approximate estimate of the amount, and thus simplifying the choice of method for confirming the immunoassay result (Smith, 1985). Consequently, samples contaminated with pesticides could be screened using the chlortoluron antibody for the presence of phenylurea herbicides. The ELISA assays for isoproturon and chlortoluron were reliable and robust from day to day. Standards prepared in tap water, reduced to pH 5, were shown to be particularly stable over a one month period. It could be that a reduction in the pH causes the protonation of the amino group attached to the aromatic ring resulting in increased solubility of the above two herbicides in aqueous solutions, and hence substantially improving the recovery. The intra assay relative coefficient of variation (CV) ranged from 2.6-5.6% and 2.7-6.5% for isoproturon and chlortoluron respectively across the standard curves compared with the iater-assay CV which ranged from 7.2-15.1% isoproturon and 6.9-12.5% chlortoluron. Furthermore, the detection limits of the assays, defined as the concentration equivalent to a three standard deviation (SD) fall from binding at Bo (the binding measure in the absence of the standard), were 0.03 /fg/L and 0.015 ^g/L for isoproturon and chlortoluron respectively, well below the EC MAC for a single pesticide in drinking water. Water samples from a variety of sources are likely to exhibit a wide range of pH and ionic strength, and often contain suspended solids. Initial experiments showed that differences in matrix effects between standards and samples resulted in non-quantitative and variable recovery of isoproturon and chlortoluron from authentic water samples. Previous ejqieiience has shown improved recovery of atrazine and paraquat from biological samples (Hardcastle, 1990), and the addition of normal serum (i.e. human or sheep) at a final concentration of 2.5% to the enzyme conjugate solution was found to overcome the matrix effects seen with ground water samples. Thus, quantitative recovery of isoproturon and chlortoluron using ELISA techniques was then achieved without extensive

Chapter 3 105 sample preparation, i.e. utilising solvent or solid phase extraction procedures, as is required by the GC and HPLC methods. The extensive use of isoproturon and chlortoluron in agriculture may cause progressive micro-exposure of workers to low levels of herbicides which could occur by absorption through skin, inhalation and accidental ingestion. Immunoassays offer a convenient diagnostic aid in case of poisoning and provide an opportunity to cany out occupational monitoring of workers. The ELISAs developed for water analysis were re-optimised for the analysis of isoproturon and chlortoluron in plasma and urine by diluting the standards with the appropriate analyte free biological fluid. The intra-plate variation ranged from 5-6.6% and 4.3-6.8% for isoproturon and chlortoluron respectively in plasma and 3.4-8.6% and 2.6-6.6% in urine. Reproducible and quantitative recoveries of isoproturon and chlortoluron firom plasma and urine were possible. Plasma samples spiked with isoproturon and chlortoluron standard solutions at different concentrations showed recoveries ranging firom 85 to 124%, and urine samples spiked with isoproturon and chlortoluron solutions showed recoveries ranging firom 87 to 119%. Moreover, although the sensitivities of the plasma standard curves were slightly improved (0.025 fig/'L and 0.01 ^g/L for isoproturon and chlortoluron respectively), the optical densities of the zero (presence of label only) obtained in the plasma standard curves were significantly mcreased. This could be due to the co-operative binding of other proteins present in the plasma which increase the colorimetric end point of the assay. As a result, further reduction in the amount of enzyme label used in the assa}^ can be achieved, thus improving the assay sensitivity even further. However, it was observed that the optical densities obtained by the urine standard curves were significantly lower contributing to a decrease in the sensitivities of the curves. One possible eiqplanation could be that in urine the presence of a wide range of compounds accompanied with low protein concentrations (i.e. salts, urea) minimise the binding conditions of the antibody towards its analyte resulting in a reduction in the colorimetric end point of the assay. Finally, the isoproturon ELISA method was chosen to demonstrate the potential of ELISA technique for monitoring pesticides in biological fluids. A preliminary study was undertaken to demonstrate the time course required for detecting the presence of "immunoreactive" isoproturon in the blood of animals

Chapter 3 106 injected with isoproturon standard. 0.2 mg/mouse (lOmg/Kg) isoproturon standard solution, equal to the LD 0.25 (Merck Index, 1991), was injected into 22 DBA2 male mice intra-peritoneahy (IP) in order to monitor the herbicide in the circulation over a 48 hour period. This low level of isoproturon was chosen to prevent the administration of a toxic dose to the mice, and also to prove that the assay was highly sensitive. The procedure was accomplished by harvesting two mice at regular intervals. Two mice acted as negative controls. The collected serum samples were assessed for the presence of isoproturon or its metabolites against the standard curve, made up with normal mouse serum, using the isoproturon ELISA method. The elimination rate of isoproturon from the circulation was determined (@ tj/ 2= 3.96 min and 8 ti/ 2= 2 hr) using PC Nonlin software and assuming a two compartmental model (correlation coeffLcient= 0.99). The area under the curve, when plasma isoproturon concentration was plotted against time was also calculated at 27609 fig /L .m in using the same two compartmental model. Although little information is available on the metabolism of isoproturon, it has been reported that isoproturon is metabolised to 4-isopropylanilme in mice (Merck Index, 1991). This could result with a cross-reaction between the isoproturon antibody and the isoproturon metabolites as the 4-isopropyl group is contained in both isoproturon and 4-isopropylaniline chemical structures. Practicalty, since not enough information is available in the literature about the metabolism of isoproturon in animals or human, the immune response detected by isoproturon antibody using ELISA might not be specific to isoproturon compound, and the antibody might have cross-reacted with different isoproturon metabolites present in plasma. However, this cross-reaction would be an advantage vdien there is a need for a screening assay. Further knovdedge about the pharmacokinetic studies of isoproturon is required to understand the metabolic rate of the compound, and hence the use of mass-spectrometry is essential to identify the metabofites. Also, antibody specificity should be measured towards the identified metabolites, once isolated, by performing cross reactivity tests. The results showed significant decrease in the concentration of immunoreactive isoproturon after one hour of injection, whereas complete elimination of the compound firom the circulation was observed after eig|ht hours.

Chapter 3 107 In summaiy, ELISA methods for isoproturon and chlortoluron have been developed with limits of detection well below the MAC for single pesticides in drinking water supplies. Furthermore, measurement of samples from various matrices (i.e. water, biological fluid) was performed by a direct application of samples into the assay without prior sample extraction. Also, it has been shown that this methodology may be a suitable technique for occupational exposure studies. Consequently, the high limit of detection produced and the lengthy sample preparation required by the leading conventional techniques (i.e. GC and HPLC) can be substituted by the use of im m unoassays.

Chapter 3 108 CHAPTER FOUR DEVELOPMENT OF 2,4-D AND MCPA ELISA METHODS

Chapter 4 109 4.1 INTRODUCTION 2-4-dichlorophenoxy acetic acid (2,4-D) and 2-met±tyl-4-chlorophenoxy acetic acid (MCPA) are phenoxy acetic acid derivatives used for killing all types of weed. They control broad leafed weeds in grassland and cereal crops vdiich play a crucial role in the diet of man and farm animals. Phenoxyacetic acid herbicides were introduced as a result of work carried out both in the USA and in Britain around 1944, although the substance now known as 2,4-D was used earlier as a plant growth regulator (HassaÜ, 1990). All members of the group have a chlorine atom attached to C-4 of the phenojQr acetate benzene ring and either a chlorine atom or a methyl group on C- 2. Sometimes an additional chlorine atom is present on C-5. Figure 4-1 illustrates the chemical structures of 2,4-D and MCPA.

CH3

-O-CHg-COOH O O H g COOH

2 ,4 -D MCPA

Figure 4-1 2,4-D and MCPA chemical structures.

Formulations of MCPA or 2,4-D are usually either as esters emulsified in oil or as water-soluble amines or other salts. Amine salts are usually very soluble in water and such formulations are convenient for low-volume application. Ester formulations are insoluble in water and more toxic to weeds than are the ionised forms. They are , however, much less selective. MCPA is extensively used in Europe but 2,4-D is often preferred in the USA. The difference probably reflects the relative costs of raw materials at a regional level. The sodium salt of MCPA is more soluble in water than is that of 2,4-D (27% as against 4.5%) and may cause less jet blockage when the volume of application is low. MCPA-Na is more selective than 2,4-D-Na but it is also less effective against many (but not all) weed species (Hassall, 1990). Current concerns about potential health hazards connected with pesticide use have focused on 2,4-D as a suspected cancer-causing agent (Hoar et al.

Chapter 4 110 1986). Also, cases of soft tissue sarcoma (STS) were observed among persons potentially exposed to MCPA (Lynge, 1993). Thus, the widespread use of these herbicides and associated health concerns have made monitoring environmental and biological samples for the presence of 2,4-D and MCPA desirable. A number of chromatographic methods have been developed for the measurement of 2,4-D and MCPA (Geerdink et al, 1991) (Coquart and Hennion, 1993). The methods, however, require extensive sample preparation including derivatization before samples can be analysed. On the other hand, although immunoassays for the detection of 2,4-D and MCPA in water, plasma and urine have been published (Knopp et al, 1985) (Fleeker, 1987) (Hall et al, 1989), aU assays developed were not sensitive enough to detect the MAC for a single pesticide in drinking water. Therefore, the need for simple, low cost and sensitive assays for monitoring 2,4-D and MCPA in various matrices has been addressed ty the production of antisera to the above herbicides. The development of direct competitive ELISA methods for 2,4-D and MCPA have been described in this chapter.

4.2 2,4-D AND MCPA nmUNOGENS The 2,4-D and MCPA haptens were linked to bovine thyroglobulin by N- hydro^y succinimide active ester method (Anderson, 1963) as described in 2.4.1.1. The method involves the conjugation of the carboxyl groups already existing on both haptens to amino residues on thyroglobulin protein. Figure 4-2 illustrates the proposed structures of the 2,4-D and MCPA immunogens.

Cl-(4 -O-CHg-CO-NH-thyro^obuIin

2,4-D IMMUNOGEN

CI-(4 ^-O-CHg-CO-NH-thyroglobiilm

MCPA IMMUNOGEN

Figure 4-2 2,4-D and MCPA protein conjugates.

Chapter 4 111 4-3 nmUNlSATION PROCEDURE 5 mg from each immunogen was then injected intra-muscidarly into two mature Suffolk sheep (1837, 1838 for 2,4-D, 1835 and 1836 for MCPA) and the animals were boosted again with half the primary dose three months later. Seven boosting doses were then given to the sheep at appropriate intervals which resulted in moderate immune responses. Blood was collected from the jugular vein of the sheep and serum was separated after allowing the blood to clot and stored at 4°C with the addition of 0.1% sodium azide.

4.4 ANTISERA SCREENING PROCEDURE l"^C 2,4-D label was used in a RIA for screening 2,4-D and MCPA antisera. Due to the commercial unavailability of ^^C MCPA label, it was hoped that there would be sufficient cross reactivity of any MCPA antibodies with 2,4-D to detect an immune response. The results of the immune response for 2,4-D were plotted as titre and B/T% for MCPA (low response) against num ber of weeks since initial Immunisation, Figure 4-3. Although 2,4-D sheep 1838 had higher concentrations of antibody than sheep 1837, both animals did not display any displacement with the addition of 0.5/fg/L 2,4-D standard. Bleed 1838 (IV:B) was chosen for immunoglobulin purification because of its high titre. On the other hand, the two MCPA sheep 1835, 1836 showed a very low immune response throughout the seven booster doses, which could be related to the use of 2,4-D label for screening, and no displacement was obtained with 0.5/

4.5 PURIFICATION OF ANTISERA 2mls of the favoured antisera bleeds, from above, were partially purified on DEAE cellulose ion exchange discs following the method described in 2.4.3. Figure 4-4 shows the absorbance of the fractions from the DEAE discs. 2,4-D fractions, 5 to 7, were pooled, and the concentration of the immunoglobulins was calculated from the OD at 280nm to be 1.16mg/ml (Little and Donahue, 1967), whereas MCPA fractions, 3 to 4, were pooled and the concentrations of the immunoglobulins was 0.97 mg/ml. The purified antisera were then stored at 4°C with 0.1% thimerosal solution added for preservation.

Chapter 4 112 Titre (a) 100 90 boost boost boost boost boost boost 80 70

prime Sheep 1838 50 Sheep 1837 40 30 20

0 20 40 60 80 100 120 140 Week No

B/T% (b) 140 boost boost boost boost boost boost boost 130 120 prime 110 100 90

Sheep 1835 70 - 60 - Sheep 1836 50 - 40 - 30 - 20 - 10 -

0 20 40 60 80 100 120 Week No

Figure 4-3 The immunisation charts for (a) 2,4-D, Bleed (IV:B) and (b) MCPA. ® : Bleed (V:C).

Chapter 4 113 (a)

OD 280nm 3

2.5

1.5

0.5

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Fractions (2ml each)

(b)

OD 280nm 2.5 r

1.5 -

0.5

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Fractions (2ml each)

Figure 4-4 Purification profile of (a) 2,4-D and (b) MCPA antibodies. Pooled fi'actions.

Chapter 4 114 4.6 HRPO-HAPTEN CONJUGATE 2,4-D and MCPA haptens were also conjugated to HRPO by the N-hydroxy succinimide active ester method at different nominal ratios, HRPO : hapten, of 1:5, 1:10, 1:20 and 1:40 as described in 3.7. The final concentrations of peroxidase were I.97mg/ml and 1.83mg/ml for 2-4D and MCPA respectively. The enzyme conjugates were then stored at 4°C after the addition of 0.1% thimerosal solution.

4.7 2.4 D ELISA A competitive ELISA method for 2,4-D was initially adopted due to its simplicity and the advantage that several different reagent concentrations and combinations could be investigated at the same time.

4 .7 .1 Antiserum dilution and displacement curves The titre of the purified 2,4-D antiserum was assessed by the antiserum dilution curve, whereas avidity of the antibody was assessed indirectly by the displacement curve. The procedures were carried out as follows: A polystyrene microtitre plate was coated with serial dilutions of 2,4-D antibody (1/500, 1 /1000, 1 /2000, 1 /4000 and 1 /8000) diluted in sodium barbitone buffer pH 9.6, and incubated in a moist chamber overnight at 4°c ^ temperature. Next day, the plate was washed with PBST buffer. Two types of solution were made up in PBS buffer: (A) and (B). Solution (A) was 100 fig/L 2,4-D standard and solution (B) was 1/1000 dilution of 2,4-D enzyme label (1:20 molar ratio). To produce the antiserum dilution curve, lOOjWl/well firom solution (B) followed with 100/d/well firom PBS buffer were added to the plate whereas for the displacement curve 100/d firom each solution (A) and (B) were added to each well. The plate was then incubated in an Enfer shaker for one hour at 37°C, as optimised in 3.8.1.3.4. Finally, the plate was washed with tap water three times and 2(X)/d of TMB substrate-chromagen were applied to all wells. After a further 30 minutes incubation in the Enfer shaker at 37°C, the colour development was stopped by the addition of IM Hydrochloric acid (HCl), 50//l/weU. The plate was read using a Labsystem microtitre plate reader at 450nm. Figure 4-5 shows high optical density readings for all antiserum dilutions and significant displacement with the addition of 100 /fg/L 2,4-D

Chapter 4 115 OD 450nm 2.4

1.6 A/S dilution curve 1.2 —I— Pre-immune serum lOOpg/L 2,4-D 0.8

0.4

100 1000 10000 2,4-D antiserum dilutions

Figure 4-5 Antiserum dilution and displacement curves of 2,4-D (antiserum 1838, bleed IV: B) using ELISA.

Chapter 4 116 standard. Thus, further optimisation of the enzyme label dilutions and coating antibodies were required.

4 .7 .2 Assay protocoljor 2,4-D sta n d a r d curve The inner 60 wells of a microtitre plate were coated with 200/^1/well partially purified 2,4-D antibodies diluted with 0.07M sodium barbitone buffer pH 9.6 and incubated in a moist chamber overnight at 4^0. The plate was then washed three times in PBST buffer. 100/d/well serial 2,4-D standards (10000, 10(X), ICX), 10, 1 and 0 /^g/L) diluted with distilled water were added to the plate followed with the addition of 100/d/well 2,4-D enzyme label diluted with PBS buffer. The plate was then incubated with shaking for one hour at 37°C. After incubation, the plate was washed three times with tap water, and 200/^1/well TMB chromagen-substrate were added into all wells and left for 30 minutes at 37^ shaking. Finally, the reaction was stopped by the addition of 50/^1/well IM HCl and the colour endpoint was read as before.

4 .7 .3 Standard curve optimiscdion 4.7.3.1 Coating antibody Several dilutions of the purified 2,4-D antibody (1/1000, 1/2000, 1/4000 and 1/8000), made up with sodium barbitone buffer pH 9.6, were utilised to construct four 2,4-D standard curves in duplicate. The dilution of the 2,4-D enzyme label was fixed at I/IOCX) (1:20 molar ratio) and applied to all standard curves. Figure 4-6 shows 1/1000 dilution (1.16/fg/ml) of antibody gave the best standard curve as it possessed the steepest slope with high optical density readings.

4 7.3.2 2,4-D peroxidase label Serial dilutions (1/1000, 1/2000, 1/4000, 1/8000 and 1/16000) of 2,4-D HRPO label from each molar ratio (1:5, 1:10, 1:2b and 1:40 "HRPO:hapten") ,diluted with PBS buffer pH 7.4, were applied onto standard curves, constructed on one plate, using the above optimum antibody dilution (1/1000). Figure 4-7 shows the effect of different label dilutions at different molar ratios. The dilution of 1/2000 (0.985 figfrol peroxidase) with a molar ratio of 1:40 showed the steepest slope curve with high OD readings.

Chapter 4 117 OD 450nm 2.4

1.6 1/1000 1/2000 1/4000 -B- 1/8000 0.81

0.4

1 100.1 100 1000 10000 100000 2,4-D ijg/L

Figure 4-6 Optimisation of the 2,4-D coating antibody dilution using ELISA. Legend shows dilutions of antibody used.

Chapter 4 118 OD 450nm OD 450nm 2.4 2.4

1.6 — 1/1000 1.6 — 1/1000 -t- 1/2000 - + - 1/2000 1/4000 1/4000 -s- 1/8000 -S- 1/8000 0.8 1/16000 0.8 1/16000

0.4 0.4:

0.001 0.01 0.11 10 100 1000 0.001 0.01 0.1 1 10 100 1000 2,4-D pg/L 2,4-D pg/L Molar ratio=l:40 Molar ratio=l:20

OD 450nm OD 450nm 1.6

— 1/1000 0.8 1/1000 — 1/2000 -+- 1/2000 0.8 1/4000 1/4000 -e - 1/8000 - e - 1/8000 -X- 1/16000 0.4 -X- 1/16000 0.4;

0.001 0.01 0.1 100 1000110 0.001 0.01 0.11 10100 1000 2,4-D pg/L 2,4-D pg/L Molar ratlo=ldO Molar ratio=l:5

Figure 4-7 Optimisation of the 2,4-D HRPO label for molar ratio and dilution. Legends show dilutions of 2,4-D HRPO label used.

Chapter 4 119 4 .7 .4 Competitwe 2,4-D ELJSAJînal protocol Using the information gained from the previous chapter the following protocol was finally adopted: A microtitre plate was coated with 2 0 0 /û per well (1.16 fig/ml) of partially purified (by ion exchange) 2,4-D antibody 1838 (IV:B) diluted in 0.07M sodium barbitone buffer pH 9.6 (1/1000 dilution). The plate was incubated overnight at 4 ^ in a moist chamber and washed three times with PBST buffer on the next day. Standard solutions (100/d per well) of 2,4-D in the range 0-10000 fig/L, m ade u p in tap water, were applied to the plate followed by the addition of 100/d/well HRPO-labelled 2,4-D at 1/2000 dilution (0.985 fig/ml peroxidase, molar ratio 1:40), diluted in PBS buffer which contains 2.5% normal serum sheep, to bring the total volume to 200/fl per well. The plate was incubated in a shaker incubator (Enfer) for one hour at 37®C. After incubation, the plate was washed four times with tap water and 2 0 0 fil p ^ well of chromagen-substrate were placed in all the wells. Finally, after a further 30 minute incubation the reaction was stopped by the addition of IM HCl (50/

4 .7 .4 Assrty vaiidatùm 4.7.4.1 Cross-reactivity The selectivity of the 2,4-D antiserum was assessed by deriving cross reactivity curves for different types of herbicides. In this procedure, a serial of standard curves were constructed, one consisting of increasing concentrations of 2.4-D (0, 1, 10, 100, 1000 and 10000 fig/h), and the others of the same concentrations of herbicide compounds which might cross-react. Then, 1 /2000 dilution of 2,4-D HRPO label solution, containing 2.5% normal serum, was added to all the curves. Table 4-1 demonstrates the degree of cross-reactivity of 2.4-D antibody with different herbicides which revealed high degree of specificity towards 2,4-D.

Chapter 4 120 T able 4 -1 Specificity of 2,4-D antiserum towards a selection of herbicides.

C om pound Cross-reaction(%)

2,4-D 100 MCPB 9.39 MCPA 7.198 Dichlorprop 2.442 Mecoprop 0.811 2,4,5-T 0.35 Fenoprop 0.275

Isoproturon < 0 .0 0 1

Chlortoluron < 0 .0 0 1

Metoxuron < 0 .0 0 1

Chlorbromuron < 0 .0 0 1

Chlorsulfuron < 0 .0 0 1

Metsulfiiron < 0 .0 0 1

Metamitron < 0 .0 0 1

Atrazine < 0 .0 0 1

Simazine < 0 .0 0 1

Propyzamide < 0 .0 0 1

Paraquat dichloride < 0 .0 0 1 Terbutiyn < 0 .0 0 1

4.7 4.2 2»4-D assay variation a) Intra-assay variation Six 2,4-D standard curves with eight standard 2,4-D solutions (0, 1, 10, 50, 100, 500, 1000 and 10000 fig/L), diluted with tap water, were constructed on a microtitre plate. Table 4-2 illustrates the mean of the six standard curves and their variation within the plate. Also, Figure 4-8 shows the mean standard curve which has a detection limit of 50 fig/h, defined as the concentration equivalent to a three standard deviation (SD) fall firom binding at Bo (the binding measured in the absence of 2,4-D), and CV% less than 5.3%.

Chapter 4 1 2 1 T able 4 -2 Within-plate variation for 2,4-D standard curves using ELISA.

2,4-D sta n d a rd M ean(n=6) SD CV% fig/L B/Bo%

100 0 0 9 ±0.4 4.4 1000 24 ±1.3 4.6 500 48 ±2.5 5.2 100 65 ±2.9 4.5 50 77 ±3.9 5.1 10 89 ±4.3 4.8

1 97 ±4.8 4.9

0 (Bo=1.33) 100 ±3.9 3.9

b) Inter-assay variation Six 2,4-D standard curves, each consisting of eight standards in duplicate, were performed on six different days to determine the CV% from day to day assays. Table 4-3 shows a range of variation between 4.1-7.3%.

Table 4-3 Between-plate variation for 2,4-D standard curves using ELISA.

2,4-D sta n d a rd M ean(n=6) SD CV% yg/L B/Bo%

1 0 0 0 0 10 ±0.57 5.7 1000 26 ±1.9 7.3 500 45 ±3 6.7

100 6 8 ±4.8 7.1 50 74 ±3.3 4.5

10 89 ±5.4 6 .1

1 101 ±6 .1 6 0 (Bo=1.48) 100 ±6.9 6.9

Chapter 4 122 B/Bo% CV% 50 100 45 90 BO-1.3 40 35 70 30 60 25 50 40 20 30 15 20 10

0.1 1 10 100 1000 10000 100000 2,4-D pg/L

Figure 4-8 Intra-plate variation of 2,4-D standard curve with precision profile.( I ), standard deviation; (-X -), mean of six standards; (-A-), coejfficient of variation.

Chapter 4 123 4.7.4 3 Stability of 2,4-D standard Ten 2,4-D standard curves constructed from six standard solutions, already made up with tap water and stored at 4°C over a period of one month, were compared to determine the stability of 2,4-D standard in solution during this period. Table 4-4 demonstrates remarkable precision of 2,4-D solution with CV% less than 8.2%.

Table 4 -4 Stability of 2,4-D solutions in tap water stored at 4°C over a period of one m onth.

2,4-D s ta n d a rd M ean(n= 10) SD CV% fig/L B/Bo%

1 0 000 8 ±0.55 6.9 1000 27 ±2 .1 7.8 100 63 ±4.5 7.1

10 83 ±5 6

1 94 ±7.7 8 .2

0 (Bo=1.29) 100 ±7.4 7.4

4.7.4.4 Recovery of 2,4-D firom various water Different samples of water were spiked with two different concentrations of 2,4-D standard (200 and 800 fig/h) and the recovery of the spiked samples was assessed using the 2,4-D ELISA method. Table 4-5 shows a complete (80-111%) recovery of 2,4-D from these samples indicating the possibility of monitoring this herbicide in water at these conditions.

Chapter 4 124 T able 4 -5 Recovery of 2,4-D from different water samples using ELISA.

2,4-D A m ount CV% R ecovery A d d e d fig/L Determined fig/L (n=6) (%)

Tap water (day 1)

2 0 0 191 6.3 96 800 834 5.6 104 Tap water (day 2) 2 0 0 178 7.1 89 800 721 4.2 90 Tap water (day 3) 2 0 0 165 6.9 83

800 807 5.7 101 River Ouse

2 0 0 182 5.1 91 800 781 6.7 98 River Wey 2 0 0 159 7.3 80 800 732 8 .1 92 University lake 2 0 0 21 6.9 105 800 846 5.7 106 University Manor Farm

2 0 0 2 2 2 4.2 111 800 789 6 .8 99 Research Park lake 2 0 0 216 6 .6 108 800 742 5.9 93

Chapter 4 125 4 7.4.5 Recovery of 2,4-D firom biological ftoids Samples of normal plasma and urine were spiked with 2,4-D standard solution at three different concentrations (1000, 500, and 100 jKg/L) and measured by the 2,4-D ELISA method. The standard solutions were made up with plasma for detecting the spiked plasma samples and urine for detecting spiked urine samples. Tables 4-6 and 4-7 demonstrate the within and between-plate variation of 2,4-D standard curves in plasma and urine respectively and include the recovery percentage of 2,4-D from the spiked plasma and urine samples. Figure 4-9 displays the intra-plate variation between 2,4-D standard curves made up with plasm a and urine.

Table 4 -6 Intra and inter-assay variation of 2,4-D standard curves in plasma and the recovery of 2,4-D from spiked plasma samples.

2,4-D Wittiin-assay variation Between-assay variation S ta n d a rd fig/L M ean SD CV% M ean SD CV% BlBo% (n=6) B/Bo% (n=6)

1 0000 7 ±0.31 4.4 10 ±0 .8 8 8 .8 1000 21 ±0.98 4.7 28 ±2.7 9.6 100 59 ±3.1 5.2 65 ±4.6 7.1 10 77 ±4.6 6 80 ±8.3 10.4 1 90 ±6.2 6.9 92 ± 8 .6 9.3

0 (Bo=1.7) 100 ±5.5 5.5 1 0 0 ±8 .1 8 .1

2,4-D a d d e d Amount determined CV% Recovery mg/L M ean (mg/L) SD (n=6) W

Plasma

0 .1 0 .1 1 2 ±0.0096 8 .6 112 0.5 0.537 ±0.048 8.9 107 1 1.089 ±0.069 6.3 109

Chapter 4 126 T ab le 4 -7 Intra and inter-assay variation of 2,4-D standard curves in urine and the recovery of 2,4-D in spiked urine samples.

2,4-D Within-assay variation Between-assay variation S ta n d a rd figlL Mean SD CV% Mean SD CV% B / Bo% (n=6) B/Bo% (n=6)

1 0 0 0 0 13 ±0.9 6.9 12 ±1.1 9.2

1 0 0 0 31 ±2.7 8.7 34 ±2 .8 8 .2 100 6 8 ±4.1 6 65 ±6.9 10.6 10 8 8 ±5 5.7 91 ± 8 .6 9.5 1 102 ±7.3 7.2 99 ±10.5 10.6

0 (Bo=0.9) 1 0 0 ±5.4 5.4 1 0 0 ±8.9 8.9

2,4-D a d d e d Amount determined CV% Recovery mg/L M ean (mg/L) SD (n=6) W

Urine

0 .1 0.086 ±0.0078 9.1 8 6 0.5 0.532 ±0.044 8.3 106 1 0.912 ±0.059 6.5 91.2

Chapter 4 127 (a)

B/Bo% CV% 110 50 100 45 90 40 80 35 70 30 60 25 50 20 40 30 15 20 10 1 0 -

0.1 1 10 100 1000 10000 100000 2,4-D ng/L

(b)

B/Bo% CV% 110 50 45 90 40 Bo-0.9 80 35 70 30 60 25 50 20 40 30 20 10

0.11 10 100 1000 10000 100000 2,4-D Mg/L

Figure 4-9 2,4-D standard curves made up in (a) plasma and (b) urine with precision profiles. ( I ), standard deviation; (-X -), mean of six standards; (-A-), coefficient of variation.

Chapter 4 128 4.8 MCPA ELISA 4 .8 .1 Antiserum and. displacement curves Antiserum dilution and displacement curves were performed to assess the titre and indirectly the avidity of the chosen MCPA antibody. The procedure was carried out as described in 4.7.1 with antibody dilutions of 1/250, 1/500, 1/1000, 1/2000 and 1/4000 and the MCPA HRPO-label dilution was 1/500 (molar ratio 1:40). Initial attempts to displace the MCPA label with lOOfig/L MCPA standard, made up with tap water, however, proved to be unsuccessful as shown in Figure 4-10. Therefore, using the same dilutions of antiserum, displacement was attempted with higher concentrations of MCPA (250, 500, 1000 and 10000/ig/L). No evidence of significant displacement was seen at these increased concentrations as demonstrated in Figure 4-11. In order to assess the specificity of the MCPA antilxxfy to its anatyte, the thyroglobulin-MCPA conjugate used to immunise the sheep, thyroglobulin and MCPA standard were used in an attempt to displace the label at a concentration of lOOOO/fg/L made up with tap water. Only the immunogen demonstrated a significant capacity to displace the label firom the antibody binding sites. Figure 4-12. Non-displacement with MCPA may possibly be due to a strong binding between the enzyme label and the MCPA antibody, and it was decided to attempt a non-equilibrium system whereby the MCPA standard would be given a chance to bind to the coating antibody before the label was added. Thus, 10000//g/L MCPA standard was added to the assay and incubated for one hour before the MCPA enzyme conjugate was added to the assay at a dilution of 1 /500 (molar ratio 1:40). This was followed by further incubation for half an hour. It was hoped that reducing the chance of the label binding to the antibody by prior ejqjosure to the standard would enhance the antibody binding towards MCPA. The immunogen and thyroglobulin were also included in this assay at 10000//g/L concentration. Figure 4-13 shows virtually no displacement with lOOOO^g/L thyroglobulin and MCPA standard whereas significant displacement was obtained hy lOOOOfig/L concentration of immunogen. The result suggests a recognition of the antibody towards the linkage between the MCPA and thyroglobulin which may have masked any significant reduction of antibody specificity towards the MCPA molecule. In many cases it has been shown that the separation of the hapten and the protein by a carbon

Chapter 4 129 OD 450nm

1.6

1.2 A/S dilution curve — Pre-immune serum 0.8 ^ lOOpg/L MCPA

0.4

100 1000 10000 MCPA antiserum dilutions

Figure 4-10 MCPA antiserum dilution and displacement curves using (antiserum 1835, bleed V: C) ELISA.

Chapter 4 130 OD 450nm 2

1.6

1.2 ~*~ A/S dilation curve 2S0yg/L MCPA -B- 500pg/L MCPA -X- lOOOpg/L MCPA 0.8 - 0 - lOOOOpg/L MCPA

0.4

0 200 2000 MCPA antiserum dilutions

Figure 4-11 Displacement of MCPA antiserum dilution curves with increasing concentrations of MCPA standard.

Chapter 4 131 OD 450nm 2

1.6

A/s dilution curve 1.2 lOmg/L Immunogen lOmg/L thyroglobulin 0.8 ^ lOmg/L MCPA

0.4

0 >- 200 2000 MCPA antiserum dilutions

Figure 4-12 Displacement of MCPA antiserum dilution curves withlOmg/L MCPA, immunogen and thyroglobulin.

Chapter 4 132 OD 450nm

1.6

A/S dilution curve 1.2 4K- lOmg/L Immunogen lOmg/L thyroglobulin 0.8 -X- lOmg/L MCPA

0.4

200 2000 MCPA antiserum dilutions

Figure 4-13 Displacement of MCPA antiserum dilution curves with lOmg/L MCPA, immunogen and thyroglobulin using a non-equilibrium system.

Chapter 4 133 bridge increases the Immunogenicity (Robinson et al, 1975). Presumably, the insertion of a spacer group allows the hapten to be more easily recognised by the circulating lymphocytes. Consequently, another MCPA immunogen was prepared with a longer spacer group between the hapten and the carrier protein.

4 .8 .2 AUemative MCPA immunogen After several meetings and discussion, the synthesis of the new hapten of MCPA was made by Dr. Frost at Robens Institute.

4.8.2.1 Chemical synthesis of MCPA hapten 20g (O.IM) MCPA was mixed w ith 14ml (0.13M) diethylenetriam ine in a 150ml flat-bottomed flask set up on a stirrer-hot plate with distillation head, condenser and receiver. The internal temperature was raised slowly and at 140- 145°C the slow distillation of a cloudy fluid began. After collecting the distillate, the hapten was readily soluble in 5% acetic acid forming a clear solution and was then subjected to examination by TLC where one major spot was visualised. The proposed chemical structure is illustrated in Figure 4-14.

^ C H 3

C l - / V O-CH2 -CO-NH-CH2 -CH2 -NH-CH2 -CH 2 -NH2

4-dtiloio-2-methyIphenoxyaeetamidoethyliminoethylamme

Figure 4-14 The proposed chemical structure of the synthesised MCPA hapten.

4.8 2.2 MCPA hsqpten-thyroglobiilin conjugate Bifunctional imidoesters are extremely useful as protein cross-linking agents between two amino groups, especially when the compound to be conjugated is insoluble in aqueous solution, (Wold, 1967). The procedure was carried out as described (Al-Bassam et al, 1978): 7.46mg MCPA hapten and 6.38 mg dimethyl adipimidate (DMA) were dissolved in 500/^1 dry methanol containing 5% N-ethylmorpholine at 2QPC, After 30 minutes,

6 8 mg thyroglubulin (1:200, thyroglobulin:hapten) dissolved in 5ml of bicarbonate

Chapter 4 134 B/T% 140 130 120 boost 110 boost 100 90 Srime Sheep 9179 "29"

60 Sheep 9179 "88"

40 30 20

0 10 20 30 40 50 60 70 80 Week No

Figure 4-15 Immunisatioii chart for the MCPA immunc^en prepared with adipimidate using ELISA.

buffer pH 9.9 was added to the above mixture and left for 90 minutes. The reaction was then stopped by the addition of 5ml of tris buffer pH 7.5 and the mixture was dialysed against 3 x XL bicarbonate buffer pH 9.6 over 48 hours.

4.8.2.3 Production and screening of the MCPA antisera The above MCPA immunogen was used for the im m unisation of two sheep (9179"29" and 9179"®®") as previously described. Although three booster doses were given to the animals over a period of one year, a significant immune response was not observed after screening the antisera with an ELISA using MCPA peroxidase enzyme label (prepared from the new hapten). Figure 4-15 aXx)ve. The low immune response obtained by ELISA was reconfirmed with a RIA using 14(2 2,4-D label, and as a result, the work on the MCPA was suspended.

Chapter 4 135 4.9 DISCUSSION 4.9.1 2,4-D A specific and easy to perform ELISA method for the phenoxyacetic acid herbicide, 2,4-D, has been described in this chapter. Although the assay did not possess sufficient sensitivity to meet the requirements of the EC MAC (0.1/ig/L) for a single pesticide in drinking water, the assay would be sensitive and simple enough for monitoring 2,4-D in air (occupational ejqxjsure) where the 2,4-D level of contamination could reach between 500^g-50mg/L (Deutsche Forschungsgemeinschaft, 1989), (Grover et all, 1985). The radioimmunoassay for detecting 2,4-D in exposed workers m agriculture and forestry published by Knopp offered a valuable technique for 2,4-D biological monitoring (Knopp and Glass, 1991). However, the potential hazards and disadvantages which are contributed by the use of radiolabels make the assay undesirable for routine laboratory work. The high specificity of the 2,4-D antibodies was attained by a direct covalent conjugation of thyroglobulin protein to a carbojg^hc acid functional group already existing on the hapten. Thus, the conjugation took place with the parent compound and not with a structural analogue. The cross-reactivity of the antiserum with a wide range of related and unrelated pesticides was negligible. Even the most closely similar structure herbicide MCPA showed virtually negligble degree of cross-reactivity (< 7.2%). The ELISA assay for 2,4-D showed a remarkable reproducibility for day to day variation and ty utilisiag standards made up in tap water a limit of detection of 50^g/L was achieved. Furthermore, 2,4-D is an acidic compound which exercises high stability in aqueous solutions. This was proved by obtaining low CV% (<8.2%) for standard curves constructed over a period of at least one month which makes the adjustment of the water pH unnecessary unlike the isoproturon and chlortoluron herbicides (3.8.1.4.7). Reproducible and quantitative recovery of 2,4-D firom different water sources and biologcal fluid were possible. Various water samples spiked with 2,4-D solution showed recoveries ranging firom 80 to 111%, and biological fluid

samples spiked with 2,4-D solution showed recoveries ranging firom 8 6 to 112%. Although the number of publications for 2,4-D immunoassays reported similar sensitivities to the assay developed hi this chapter, the availability of such specific antibodies has given us the opportunity for a wider use of the

Chapter 4 136 antisemm (i.e. preparation of affinity chromatography columns to be used as on/off line sample extraction technique prior to sample analysis by the traditional chromatographic methods).

4 .9 .2 MCPA The initial MCPA antiserum produced has proved to be unsuitable for the development of an ELISA method since displacement only occurred with the thyroglobulin-MCPA immunogen which confirms the non-specificity of the antibodies. The results suggest that the antibodies had low avidity for the herbicide MCPA. The antibodies were also able to recognise the MCPA peroxidase label since without this binding, no colour would have been produced. Moreover, using the immunogen as a displacing compound demonstrated high antibody affinity towards the immunogen, while thyroglobulin displayed no immunoreactivity. The low antibody avidity could be related to the existence of a reactive carboxyl group on the MCPA chemical structure which was utilised for direct coupling with thyrogobulin via the active ester method. Therefore, no bridge between the carrier protein and the hapten was required. It is possible that the small hapten molecule would be masked in the large surface of the protein molecule. Hence, the polyclonal antibodies produced may contain a large population of antibodies which recognise the MCPA coupled to the carrier protein rather than MCPA itself. This also explains the recognition of the MCPA antibodies to the enzyme label, as MCPA was conjugated to peroxidase protein using the same chemical reaction. The antibodies reveal preferential binding towards the HRPO-enyme conjugate label rather than MCPA standard itself. However, althougi the 2,4-D and MCPA immuno^ns possess similar chemical structures and they were both prepared by the same conjugation method, the sheep injected with the 2,4-D immunogen produced an adequate immune response to 2,4-D itself. This could possibly be related to individual variation in response of animals towards immunogens (e.g. the two sheep immunised by the same chlortoluron immunogen responded in a totally different manner from each other). Consequently, since the insertion of a spacer goup between the hapten and the carrier protein enhances the immunogenicity of the conjugate (Robinson et al, 1975), an attempt to synthesise a new hapten for MCPA with a longer

Chapter 4 137 bridge has been carried out. The immunogen was constructed so that the primary aliphatic amino group on the spacer arm (diethylene triamine) of the synthesised MCPA hapten was linked to the amino group of thyrogobulin protein molecule using dimethyl adipimidate (DMA). The sera collected from the sheep after three booster doses over a period of one year showed no immune response. One explanation might be due to significant alterations to the electronic configuration of the hapten molecule during the chemical synthesis procedure which caused minimum recognition of MCPA and thus, no immune response was observed. On the other hand, a study of different methods for the conjugation of a morphine molecule to a carrier protein (Robinson et al, 1975) resulted in no immune response when DMA was utilised as a bifunctional reagent. This could be explained as the side chain, DMA, linking the hapten to the carrier protein lies in a semicircular shape, thus projecting the hapten towards the carrier making the small molecule unavailable for recognition by the circulating lymphocytes (personal communication, G. Court to B. A. Morris). Therefore, the feasibility of the above hypothesis could explain the lack of immune response obtained ty the new MCPA antibodies as the long chain between the hapten and the carrier protein was significantly lengthened by the use of DMA (6 carbon chain), which might ultimately mask the hapten. Alternatively, an addition of a primary amino goup to the MCPA carboxyl goup could be an advantage since the bifunctional reagent, i.e. glutaraldehyde, can be used to couple the protein and hapten through their primary amino groups, under aqueous conditions, offering a bridge of five carbon between the carrier and hapten that may increase the immunogenicity of the conjugate. However, the work on MCPA was suspended due to the shortage of time available. In summary, an ELISA method for 2,4-D herbicide was developed with a detection limit of 50 ^g/L. This limit is suitable for 2,4-D air monitoring without prior sample preparation. Although the development of ELISA for MCPA ended without success, it furthered our knowledge considerably about the preparation of antigenic hapten-carrier conjugates.

Chapter 4 138 CHAPTER FIVE DEVELOPMENT OF ISOPROTURON AND CHLORTOLURON ECLIA METHODS

Chapter 5 139 5.1 INTRODUCTION Several types of non-isotopic labels (e.g.; enzymes, fluorescence and luminescence) have been successfully utilised in immunoassays as alternatives to radioactivity. Their characteristics have provided the means of developing rapid, cost-effective assays appropriate to laboratories with various requirements. In addition, simple, accurate immunoassays and immuno-devices have been devised for "on site or in the field" testing and monitoring. However, the disadvantage associated with colorimetric and fiuorimetric determinations of enzyme activity is that the procedure must be accurately timed. Because of the steady increase in signal as the reaction progesses, the signal development time (usually 15 to 30 minutes) must be accurately controlled. The reaction needs to be stopped by the addition of a reagent at a specific time. Figure 5-2a. Chemiluminescent labels are suited for use in immunoassays as they are stable and can be detected with high sensitivity over a wide linear range with a theoretical backgound of zero (Krlcka and Thorpe, 1983). The phenomenon of chemiluminescence involves the generation of excited states by a chemical reaction which upon relaxation to the ground state result in photonic emission (Weeks and Woodhead, 1988). Several types of chemiluminescent compounds are available for use as labels in immunoassays. Luminol and isolumiaol have been widely utilised as luminescent labels but have the disadvantage of low quantum yield, especially when conjugated to antigens and antibodies. The fight production is catalysed by haem-containing proteins in the presence of hydrogen peroxide in a highly alkaline solution as illustrated in Figure 5-1. Furthermore, to maximise the fight output a period of incubation in strongly alkaline conditions prior to fight measurement is required (Barnard et al, 1985). The acridinium esters were introduced to overcome these problems as much milder oxidising conditions are required for light production (Weeks et al, 1983). Besides, the quantum yield of light is not reduced h y conjugation because the reactive species is dissociated firom the rest of the molecule before the emission of photons. Both of these chemiluminescent endpoints produce a flash of fight necessitating accurate measurement times and equipment which enables the addition of reagents directly into the vial in the measuring position.

Chapter 5 140 ELjOg / OH

C atalyst ^No

Figure 5-1 Chemiluminesœnt reaction of luminol.

Horseradish peroxidase has been extensively used as a stable enzyme label. In 1983 it was reported that luciferin could increase the production of light ly several-fold from the HRPO-catatysed oxidation of luminol (Whitehead et al, 1983). Later other chemical reagents, notably p-iodophenol, were shown to be more effective enhancing agents (Thorpe et al, 1985). Amersham International has developed this new approach to the use of chemiluminescence as the detection system in a reaction based on horseradish peroxidase- chemiluminescence coming from the oxidation of luminol. The discovery of a series of special enhancers has allowed optimisation to give a prolonged, high intensity light signal. Also, the light signal produced can be measured and re measured without critical timing Figure 5-2b.

Standard-ELISA Enhanced chemiluminescent-ELISA

signal

10 20 30 0.2 10 15 20 30 Time (minutes) Time (minutes) (a) (b)

Figure 5-2 Signal characteristics of enhanced luminescence (adapted from Amersham International brochure).

Chapter 5 141 A generally accepted reaction scheme for enhanced luminescence (Misra and Squatrito, 1982) involves the oxidation of luminol by a complex between hydrogen peroxide and peroxidase to produce a luminol radical. Luminol radicals then undergo further reactions resulting in the formation of an endoperoxide which then degrades to yield nitrogen and an electronically excited 3- aminophthalate emitting light at a peak wavelength of approximately 425 nm on the return to its ground state. A simplified reaction scheme of peroxidase- catalysed chemiluminescence is shown in Figure 5-3.

HgOg / Peroxidase Lumincd i y Luminol radical Oxidation & Luminol endoperoxideII Aminophthalate Degradation +

L ight Back to the ground state * at ~ 425 nm 3-Aminophthalate + N2 (Electronically excited)

Figure 5-3 Proposed reactions sequence in the peroxidase-catalysed chemiluminescent oxidation of luminol (adapted from Misra and Squatrito, 1982).

The addition of specific enhancers has allowed prolonged and high intensity light signals to be obtained as illustrated in Figure 5-4. Although the precise mechanism of enhanced luminescence is not fully understood, it is believed that the enhancer itself acts as an electron donor enabling the reaction to cycle more rapidty, and thus overcoming the rate-limiting step in the reaction of luminol and H 2O2 (Cunningham et al, 1994) The characteristics of enhanced luminescence, i.e. an increased, prolonged and relatively stable light output (2 0 minutes), pro\4de a means of increasing assay sensitivity and of simplifying the endpoint measurement of luminescent immunoassays. The initiating reagent can be added outside the measuring instrument and samples can be re-read if necessary. In addition, established

Chapter 5 142 immunoassays which utihse HRPO labels can be readily converted into luminescent assays with the potential of increasing sensitivity. A further advantage of enhanced luminescent endpoints in inmumoassays is that they can be used in systems suitable for alternative site (i.e. extra-laboratory) use. The glow of light produced in enhanced luminescent systems is sufficiently intense that photographic film can be exposed to the signal to give semi-quantitative or qualitative results with a permanent visual record (Ireland and Samuel, 1989).

glow ing Light Peroxidase

Antibody Enhancer 2 0

Jlashing Light NH OH OH

NH2 o O

Luminol Aminophthalic acid

Figure 5-4 Illustrating the reaction sequences for enhanced chemiluminescent light production.

Other chemiluminescent systems are now also available. Chemiluminescent dioxetane-based substrates for alkaline phosphatase-labelled immuno-reagents have the ability to produce enhanced and prolonged light signals (Bronstein et al, 1989).

Chapter 5 143 5.2 ECLIA FOR ISOPROTURON AND CHLORTOLURON Although the isoproturon and chlortoluron ELISA methods described earlier offer a limit of detection weU below the MAC level for a single pesticide in water, the development of ECLIA was carried out a) to compare sensitivities (between ELISA and ECLIA) and b) to devise a portable system which can be used "in the field" by non-skilled personnel, ECLIAs for both isoproturon and chlortoluron herbicides, with a limit of detection below the MAC (0.1 fig/L), have been developed in this chapter. The final protocols for the isoproturon and chlortoluron ELISAs were adopted exactly as described previously in 3.8.1.4.4 and 3.S.2.3.2 respectively except that the amount of HRPO bound to the chemiluminescent microtitration plates (solid flat bottomed white Microlite ^^2) was quantified by the addition of an enhancement reagent (AmerUte; Amersham International) instead of the chromagen-substrate (TMB). The light emitted was measured within 2-20 minutes using the Enfer chemiluminescent plate reader. The coating antibody and the enzyme labelled dilutions for both isoproturon and chlortoluron were adopted from the final ELISA protocols. In addition, the light signal produced by the luminescent endpoint system was also optimised over a period of 2 to 2 0 minutes for the optimum reproducibility. The protocol was carried out as follows:

2 0 0 /d/well of isoproturon and chlortoluron coating antibodies at 1 / 2 0 0 0 isoproturon and 1/8000 chlortoluron dilutions were added into separate chemiluminescent microtitre plates, after being diluted in 0.07M sodium barbitone buffer pH 9.6. The plates were incubated overnight at 4 ^ in a moist chamber. On the next day, the plates were washed three times with PBST buffer pH 7.4 and standard solutions (100/^1/well) of isoproturon and chlortoluron in the range of 0-100/zg/L, made up in tap water, were applied to the plates. This was followed the addition of 100/d/well isoproturon and chlortoluron HRPO- labels at 1 /4000 isoproturon and 1 /16000 chlortoluron dilutions, diluted in PBS buffer which contained 2.5% normal serum sheep, into their related plates to bring the total volume to 200/^1/well. The plates were then incubated in a shaker incubator (Enfer) for one hour at 37°C. After incubation, they were washed four times with tap water and 200/d/weH Amerlite luminescent reagent were placed in all the wells. Finally, the light signal produced in the wells was measured at 2, 5, 10, 15 and 20 minutes from the addition of Amerlite using the Enfer

Chapter 5 144 chemüiiminescent plate reader. As a result, 10 minutes incubation of Amerlite reagent with chlortoluron bound label was chosen as the readings demonstrated low coefficients of variance at this time. Table 5-1. The standard curve was plotted using percentage label bound (B/Bo) against the concentration of standard, and concentrations of herbicide in unknown samples were obtained from this standard curve. Figure 5-5 illustrate the light signals measured for the isoproturon and chlortoluron standard curves.

Table 5-1 Within-plate variation of isoproturon and chlortoluron standard curves using ECLIA obtained at various incubation times of the Amerlite reagent.

Isoproturon CV%

2m in 5m in lO m in 15m in 2 0m in

100 10.1 9.5 4.1 6.3 7.9

10 13.3 7.8 7.3 12.2 10.3

1 9.7 10.1 6 .6 11.3 9.5 0 .1 14.5 9.9 5.2 8.9 8.4

0 .0 1 11.9 11 7.7 1 0 .6 8.7 0 12.1 10.4 6.9 9.9 12.1

Chlortoluron CV% ( pg/L 2m in 5m in lO m in 15m in 20 m in

100 14.3 9.9 4.2 6 .1 4.5 10 15.1 7.4 6.9 7.5 6 .1 1 10.9 8.7 4.3 6.3 7.8

0 .1 11.2 10.9 5.5 8 .1 8.4

0 .0 1 12.7 8.4 7.1 6 .6 8 .2 0 10.3 9.3 6.4 8.9 9.1

Chapter 5 145 B/Bo% 100 Bo-5115 90 80 70 60 50 40 30

20 -

0.001 0.01 0.1 1 10 100 1000 Isoproturon pg/L

B/Bo% 100 90 80 70 60 50 40 30

10 -

0.010.001 0.1 1 10 100 1000 Chlortoluron Mg/L

Figure 5-5 Isoproturon and chlortoluron standard curves obtained by ECLIA.

Chapter 5 146 5 .2 .1 Ass€My sensitürity, reprodxicibüUÿ cmd recovery o f spiked soMiiples a) Intra-assay variation The isoproturon and chlortoluron enhanced chemiluminescent intra-assay variation were assessed for eight standard solutions (100, 10, 1, 0.5, 0.1, 0.05, 0.01 and 0 Mg/Lh made up in tap water and stored at 4°C, by performing six standard curves (n= 6 ) for each herbicide on separate chemiluminescent microtitre plates. Table 5-2 and Table 5-3 show the mean (expressed as a percentage of the maximum binding B/Bo), SD and CV values obtained from each isoproturon and chlortoluron standard curves respectively. In addition. Figures 5-6 and 5-7 illustrate the standard curves obtained for isoproturon and chlortoluron with the intra-assay coefficient of variation indicated for each point. Furthermore, the limits of detection for the isoproturon and chlortoluron assays were 0.025/xg/L and 0.01/

Chapter 5 147 Table 5-2 Intra and inter-assay variation of isoproturon standard curves using ECLIA and the recovery of isoproturon from spiked tap water samples over a period of two weeks.

Isoproturon Wiffiin-assay variation Between-assay variation S ta n d a rd pg/L Mean SD CV% Mean SD CV% B/Bo% (n=6) B/Bo% (n=6)

100 3 ±0 .1 3.3 8 ±0.92 11.5

10 8 ±0.5 6.3 13 ±1.4 1 0 .8

1 23 ± 1.2 5.2 30 ±2 .8 9.3 0.5 40 ±3.0 7.5 44 ±6 .0 13.6 0 .1 64 ±5.0 7.8 71 ± 10.1 14.2

0.05 76 ±4.5 5.9 82 ±9.5 1 1 .6

0 .0 1 8 8 ±6 .1 6.9 91 ± 8 .8 9.7

0 (Bo=4520) 1 0 0 ±7.1 7.1 1 0 0 ±14.0 14.0

Isoproturon added Amount determined CV% Recovery fig/L M ean (fig/L) SD (n=6)

Tap water Q.l 0.078 ±0.009 11 78 0 .2 0.235 ±0.023 9 118 0.5 0.528 ±0.073 13 106

Chapter 5 148 T able 5 -3 Intra and inter-assay variation of chlortoluron standard curves using ECLIA and recovery of chlortoluron firom spiked tap water samples over a period of two weeks.

Chlortoluron Within-assay variation Between-assay variation Standard puglh M ean: SD CV% M ean SD CV% B/Bo% (n=6) BfBo% (n=6)

100 2 ±0 .1 1 5.5 3 ±0.26 8.7 10 5 ±0.23 4.6 4 ±0.37 9.25

1 9 ±0.51 5.7 11 ± 1.2 11

0.5 13 ±0 .8 6 .2 15 ±1.9 12.7

O.I 2 2 ±1.5 6 .8 28 ±2.5 8.9

0.05 63 ±3.3 5.2 70 ±7.8 1 1 .1

0 .0 1 77 ±5.1 6 .6 81 ± 8 .6 1 0 .6

0 (Bo=3811) 1 0 0 ±7.4 7.4 100 ±13.2 13.2

Chlortoluron added Amount determined CV% Recovery figfL M ean (fig/L) SD (n=6) (%)

Tap water 0 .1 0.081 ±0.008 9 81 0 .2 0.187 ±0.027 14 94

0.5 0.54 ±0.061 11 108

Chapter 5 149 B/Bo% CV%

Bo*4520

0.001 0.01 0.1 1 10 Isoproturon pg/L

Figure 5-6 Intra-assay variation of ECLIA isoproturon standard curve. ( I ), standard deviation; (-X-), mean of six standards; (-A-), coefficient of variation.

Chapter 5 150 B/Bo% CV% 110 50

100 B o-3881 45 90 40 80 35 70 30 60 25 50 - 20 40 15 30 20 10 10

0.001 0.01 0.1 1 10 100 1000 Chlortoluron pg/L

Figure 5-7 Intra-assay variation of ECLIA chlortoluron standard curve. ( I ), standard deviation; (-X -), mean of six standards; (-A-), coefficient of variation.

Chapter 5 151 5.3 PHOTOGRAPHIC ENHANCED CHEBflLUMINESCENT ENDPOINT ASSAY FOR CHLORTOLURON In chemiluminescent assays light emission is usually measured with a photomultiplier tube or a silicon photodiode. A less expensive and simpler alternative technique for measuring light is the use of photographic film. The combination of chemiluminescent and photographic film offers a number of advantages: a) Multiple assays can be performed simultaneously using low cost portable instrumentation which requires no power source. b) Ideally suited for "in field" screening procedures. c) A perm anent visual record of results is obtained. d) Rapid assays are possible, particularly if an instant development type of film is used. The major limitations of using photographic film as a detector are that it is less sensitive than a photomultiplier tube and the photographic results are semiquantitative rather than quantitative. However, a further degree of quantitation can be achieved using, for example, a densitometer. The feasibility of using the enhanced chemiluminescent immunoasseys for screening water samples has been assessed and adapted onto the Dynatech Microlite camera luminometer to provide a semi-quantitative assay based on a photographic record of the luminescent endpoint. The chlortoluron enhanced chemiluminescent assay was chosen in this case as an example, although the isoproturon assay could equally well have been used.

5 .3 .1 Types of photographicfUm Early work was performed with conventional black and white film, but more recently instant film has been preferred because of the convenience of automatic processing (personal communication with Polaroid pic). Four types of instant film have been used, Polaroid Land Types 47 (ASA 3000), 57 (ASA 3000), 410 (ASA 10,000) and 612 (ASA 20,000) black and white print film. The sensitivity of a film is represented by a speed number, based on either the ASA or DIN film speed system. The higher the number the more sensitive the film. Polaroid Type 612 has a speed of ASA 20,000 and is the most sensitive instant film generally available. Ideally, a film intended for use as a detector of chemiluminescence should combine high sensitivity with a low contrast as this would both detect

Chapter 5 152 very low light levels and give a graded response in image density across a wide range of exposure (Kricka and Thorpe, 1986).

5 .3 .2 CcanercL lu m in am eter The camera luminometer is a hgjit-tight box located on top of a film holder as illustrated in Figure 5-8. A sliding shutter is interposed between the film holder and the body of the instrument. In the closed position the shutter supports the reaction microtitre plate holder, whereas in the open position the holder drops onto the film. The holder is made in. the form of an array of holes to separate the light emission firom individual weUs preventing "cross talk" firom well to well (Dynatech camera luminometer manual book).

5 .3 .3 Photographic assay technique The concentrations of chlortoluron coating antiserum and its HRPO label were optimised to meet the requirement of a short (20 minutes) assay incubation time. Shorter incubation times required higher antibody and label concentrations (Hardcastle et al, 1989). Therefore, two dilutions of chlortoluron antibody (1/2000 and 1/4000) and three dilutions of enzyme label (1/4000, 1/8000 and 1 /16000) were used for optimisation and as a result, a concentration of 0.68 fig/L chlortoluron coating antibody (1/2000 dilution) and a dilution of 1/4000 (0.52 ^g/ml peroxidase) chlortoluron enzyme label proved to be the optimum conditions for an assay time of approximately 20 minutes. Figure 5-9. In addition, at a coating concentration of 1/4000, no photographic exposure was observed (i.e. binding of HRPO-conjugate was too low). The protocol was carried out as follows: Flexible round-bottomed PVC microtitre plates (Dynatech) were coated with 200/d/well of partially purified chlortoluron antiserum (sheep 1826, B.vc ), diluted in 0.07M sodium barbitone buffer pH 9.6, at 1 /2000 dilution for 16 hours at 4% . On the next day, the plate was washed three times with PBST buffer pH 7.4 using a m anual washing bottle. lOO/zl/weU of various standard chlortoluron concentrations (0.3, 0.1, 0.05 and 0 fig/L), made up in tap water, were dispensed mto the wells followed by the addition of 100/d/well chlortoluron HRPO label at a dilution of 1/4000 into all wells. The plate was then incubated for 20 minutes at room temperature in an enclosed box. After washing the plate three times with tap water, 200/d/well of enhanced luminescent reagent (Amerlite, Amersham

Chapter 5 153 i à

DYNATECH I MICF=iC3L.ITE

1339 - B@ a

P R O C E S & S a O M - ”

CAMERA LL/M/fsjOMETER

Figure 5-8 Dynatech camera luminometer (a, Polaroid film back; b, shutter; c, mask; d, microtitre plate; e, lid and timer).

Chapter 5 154 10 98765432

The entire plate was coated with 1 /4000 antibody dilution. (a)

The entire plate was coated with 1/2000 antibody dilution. (b) Label dilution 1 /4000 in rows 2, 3 and 4. Label dilution 1 /8000 in rows 5, 6 and 7. Label dilution 1 /16000 in rows 8, 9 and 10.

Figure 5-9 Optimisation of chlortoluron antibody and enzyme label dilutions using the camera luminometer.

Chapter 5 155 International pic) was added to each well. 10 minutes later, the glowing plate was positioned in the mask and placed in the Dynatech camera luminometer which had been loaded with Polaroid Type 612 film. The lid was replaced and the shutter withdrawn allowing the mask to drop onto the film for a period of 40 seconds and the shutter was then closed. Finally, the instant film was pulled from the film holder and after 40 seconds (automatic processing) the film backing was removed to reveal the developed photograph. The degree of ejqxjsure was proportional to the intensity of light emission. Moreover, the intensity of light was inversely related to the concentration of chlortoluron standard solution. Figure 5-10.

5 .3 .4 Photographic assay validation 5.3.4.1 Comparison between Standards and spiked samples A round bottomed PVC microtitre plate was used to visually assess eight different concentrations of chlortoluron spiked river water samples (0.01, 0.03, 0.06, 0.09, 0.11, 0.15, 0.2, and 0.3 fig/L) dispensed in triplicate. In addition, three chlortoluron standard solutions (0.1, 0.05and 0 fig/L), made up in tap water, were added onto the plate to make the visual comparison with the spiked samples possible. Figure 5-11 illustrates the visual differences between sample concentrations and standard solutions. The darker the exposure, the higher the chlortoluron concentration in the well.

Chapter 5 156 CC H M L c c C 10 9 8 7 4 3 2 •• e##B • • # # # c •• e##D

C = Negative control 0 jug/L L = 0.05 /ig/L chlortoluron M = 0.1 jttg/L chlortoluron H = 0.3 /xg/L chlortoluron

Coating antibody dilution 1 /2000 Chlortoluron label dilution 1/4000 Film exposed for 40 seconds

Figure 5-10 Visual comparison between different concentrations of chlortoluron standards using the photcgraphic technique.

Chapter 5 157 The keyfor interpretation of the above photograph:

1 0 9 8 7 6 5 4 3 2 Standard/^g/L 0.1 0.05 0 0.1 0.05 0 0.1 0.05 0 B Water sample yg/L 0.2 0.2 0.2 0.06 0.06 0.06 0.15 0.15 0.15 0 Standard fig/L 0.1 0.05 0 0.1 0.05 0 0.1 0.05 0 D Water sample fig/L 0.3 0.3 0.3 0.11 0.11 0.11 0.03 0.03 0.03 E Standard fig/L 0.1 0.05 0 0.1 0.05 0 0.1 0.05 0 F Sample&Standard 0.09 0.09 0.09 0.09 O.lSt 0.01 0.01 0.01 0.01 G Blank "no HRPO" & 8 8 8 8 8 8 8 8 H

Figure 5-11 Chlortoluron assay, using camera luminometer, containing chlortoluron standards and spiked nver water samples.

Chapter 5 158 5.3.4 2 Accuracy’ of chlortoluron photogrsqihic endpoint assay Nine different concentrations of chlortoluron spiked river water samples (0.03, 0.06, 0.09, 0.1, 0.12, 0.15, 0.2, 0.3 and 0.4 /zg/L), run in triplicate, were applied onto five PVC microtitre plates to be screened by photographic endpoint assays. The photc^aphs obtained were examined ty ten independent observers at the department to score the wells as being fighter or darker than the concentration of 0.1 fig/L standard wells. 0.05 and 0 figlL standard wells were also included on the plates for reference. Although the predictions of sample concentrations were satisfactory, it was observed that the accuracy of the assays was poorer in the range 0.1-0.12 fig/L. Table 5-4 demonstrates the percentage of correct answers obtained the ten observers for the estimation of chlortoluron concentration.

Table 5 -4 Accuracy of chlortoluron photographic endpoint immunoassay using a Dynatech camera luminometer.

Chlortoluron concentration % C orrect a n sw e r fig/L Identifying above or below 0.1yg/L

0.03 90 0.06 100 0.09 90 0.1 70 0.12 80 0.15 80 0.2 100 0.3 100 0.4 100

5.3.4 3 Correlation between chlortoluron ELISA, ECLIA and photographic endpoint techniques Fourteen river water samples were spiked with chlortoluron standard ty an independent colleague at concentrations which were undisclosed. The samples

Chapter 5 159 were then determined quantitative^ using chlortoluron ECLIA and ELISA and qualitatively by photographic endpoint immunoassay. Also, twelve independent observers were involved in the estimation of chlortoluron concentrations from the spiked samples using the developed photographs. Table 5-5 compares the quantitative and qualitative results obtained hy the above three techniques.

Table 5-5 Comparison of results for chlortoluron concentration obtained by ELISA, ECLIA and photographic endpoint techniques.

Sample Amount spiked Amount measured Photographic assay No figlL ELISA ECLIA < or > O.lfiglL jugIL /ng/L

1 0 0 0 100% <0.1 2 0.03 0.054 0.047 100% <0.1 3 0.05 0.059 0.061 100% <0.1 4 0.06 0.053 0.065 100% <0.1 5 0.07 0.062 0.08 30% >0.1; 70% <0.1 6 0.08 0.068 0.058 40% >0.1; 60% <0.1 7 0.1 0.093 0.111 50% >0.1; 50% <0.1 8 0.12 0.112 0.108 80% > 0.1; 20% <0.1 9 0.13 0.139 0.147 100% >0.1 10 0.16 0.152 0.178 100% >0.1 11 0.18 0.196 0.164 100% >0.1 12 0.2 0.193 0.212 100% >0.1 13 0.25 0.265 0.274 100% >0.1 14 0.3 0.322 0.276 100% >0.1

It has been observed. Table 5-5, that there was a degree of diffîctilty in estimating samples of chlortoluron concentrations in the range 0.07 to 0.12 fig/L; however. there was still a high percentage of correct results. On the other hand, for concentrations outside this range the results were 100% accurate. Furthermore, th ere was significant correlation between the ELISA and ECLIA results (r=

Chapter 5 160 0.9718, Y= (0.9)X + 0.02) and P value < 0.0001 which is considered highly significant.

5.4 DISCUSSION Enhanced luminescent end point assays offer great sensitivity and relatively stable light output which give the system an advantage as the signal can be measured and re-measured without critical timing. In addition, established immunoassays which utilise HRPO labels can be readily converted into luminescent assays, by merely adding the relevant substrate (luminescent reagent) with the potential of increasing sensitivity. Although the isoproturon and chlortoluron ELISA methods, developed in chapter three, offer high sensitivity which can easily detect the MAC of a single pesticide in water, the desire for the development of an "in field" technique using a camera luminometer made the utifisation of enhanced chemiluminescent endpoint assays an attractive possibility. Furthermore, the photographic endpoint assay enables the provision of results in the form of photographic records where semi-quantitative and qualitative measurements are appropriate (Kricka and Thorpe, 1986). The principle of enhanced chemiluminescent immunoassays is based on the detection of light signal produced by the oxidation of luminol in the presence of a specific enhancer. The signal yielded is prolonged and of high intensity which minimises strict measurement timing. Amerlite luminescent reagents (Amersham International pic) consisting of enzyme substrate, activator and enhancer incorporated into tablet form were used in the development procedures. Using the same conditions as determined in the final ELISA protocols of isoproturon and chlortoluron, ECLIA systems for both herbicides were established. The standard curves of the two herbicides ranged firom 100-0 pgfL with mean CV across the standards of 3.3-7.1% isoproturon and 4.6-7.4% chlortoluron (n=6). The theoretical sensitivities of the two curves (3 SD inhibition of the binding at zero analyte concentration) were 0.025 and 0.01 /fg/L isoproturon and chlortoluron respectively (compare to 0.03 ^g/L isoproturon and 0.015 ^g/L chlortoluron obtained by ELISA) well below the MAC of single pesticide in water. However, there is scope for even more sensitivity with the ECLIAs by reducing the amount of enzyme label applied. Moreover, the recoveries of both herbicides spiked into tap water samples at different concentrations

Chapter 5 161 were 78-118% isoproturon and 81-108% chlortoluron which were within the acceptable range. Although a quantitative enhanced luminescent technique proved to be effective, easy and rapid to use offering high sensitivity and acceptable reproducibility and recovery, the system requires detectors which need power source facilities and therefore is not suitable for monitoring pesticides at extra laboratory locations. Photographic endpoint immunoassays for triazine and paraquat herbicides have been reported (Hardcastle et al, 1989) indicating the general applicability of the technique for environmental pollutants. The feasibility of using a camera luminometer and simple procedures applicable to "field" conditions for the detection of chlortoluron was investigated. The assay was based on the use of a portable Dynatech camera luminometer which has great potential as an extra-laboratory monitoring technique. Using the reagents developed for the chlortoluron quantitative ECLIA, a photographic endpoint assay has been developed to facilitate a qualitative detection of chlortoluron in a large number of samples within a 30 minutes period. This was achieved by using triple the concentrations of both chlortoluron antibody and HRPO-label, stated in the ECLIA protocol, in order to reduce the incubation time to 20 minutes without substantial loss of assay sensitivity. Although in this developmental phase plates were coated overnight prior to use, for routine use, plates (or strips) could either be stored in the presence of coating solution at 4°C (Hardcastle et al, 1989) or could be dried and packed for long term storage as for other commercially available enzyme immunoassay test kits (e.g. EnviroGard Test Kits for Pesticide Detection). A number of spiked chlortoluron river water samples and chlortoluron standards were assayed using the camera luminometer, and individuals were asked to mark the results as above, equal to or below the MAC by comparing the brightness of the sample well on the photograph with the 0.1 /^g/L standard wells. A chlortoluron solution of 0.05 ^g/L was clearly observed to be below the MAC but observers had more difficulty in distinguishing the 0.12 yg/L standard from the MAC level. Furthermore, river water samples spiked with chlortoluron standard solution at levels below the MAC (as confirmed in the ELISA and ECLIA) were m ostly considered to be negative w ith respect to th e 0.1 fig/L standard. The mean percentage correct answer was 88%. On the other hand.

Chapter 5 162 samples containing 0.13 /«g/L or greater of chlortoluron were identified correctly (100%) as being darker than the MAC (0.1 figlL standard). The method developed facihtates the screening out of samples with very low chlortoluron concentrations and identifies those with high concentrations which breach the MAC. Therefore, it reduces the number of samples requiring chromatographic confirmation, thereby reducing costs of analysis. It may also be a useful device for monitoring operating procedures during water processing at water intake points or for monitoring accidental spfils of potentially toxic compounds. The photographic endpoint assay has considerable potential by offering a rapid, simple, cost effective and portable means of monitoring multiple water samples for the presence of pesticides. Moreover, the ability to identify samples which are contaminated by a pesticide or a class group of pesticides at or above the MAC level may greatly facilitate the increased monitoring of water supplies required by EC regulations. Further applications of an enhanced chemiluminescent immunoassay using a camera luminometer have already been developed for the detection of antibody to hepatitis B virus surface antigen (Ireland and Samuel, 1989) and glucose (Carter and Whitehead, 1982) in human serum. The assays have proved to be both specific and sensitive and provide a permanent photographic record. Subsequently, the photographic end point system illustrates the possibilities of using such a technique in various matrices without requiring expensive automated equipment or an external electricity supply and can therefore be used where only minimal laboratory facilities are available.

Chapter 5 163 CHAPTER SIX DEVELOPMENT OF IMMUNOAFFINITY CHROMATOGRAPHY COLUMNS FOR ISOPROTURON AND CHLORTOLURON

Chapter 6 164 6.1 INTRODUCTION 6.1.1 History Affinity chromatography is a unique separation technique in which the material to be separated is selectively adsorbed onto an immobilised secondary molecule. In this way, specific molecules can be isolated and purified firom complex mixtures of biological materials on a functional rather than a physiochemical basis (Absolom, 1981). A refinement of the basic affinity technique was the introduction of immunoadsorption or immunoaffinity chromatography, which added the selectivity and specificity of immunological reactions to the separation procedure. Thus, in a more specialised form of affinity chromatography an immobilised antibody or antigen is used as the ligand (the immobilised molecule), and the selective separation comes through immunological reactions (Fuchs and Sela, 1979). The first application of immunoaffiniiy chromatography was described by Campbell and his colleagues (Campbell et al, 1951). They used diazotized aminobenzylceUulose to immobihse an antigen for the isolation of specific antibodies directed against the immobilised antigen. Since this early start many different types of immunoaffinity matrices have been described and reported in the literature (Sfiman and Katchalski, 1966) (Wllchek et al, 1984) (Walters, 1985).

6 .1 .2 The theory of iMnnumoctffmity chromatography Immunoaffinity chromatography depends on both the strength and selectivity of an immobilised antibody (used as the ligand) for the isolation of the desired material, through a specific antibody-antigen reaction. The general approach of this technique is to immobilise a selected antibody by chemically bonding it to an inert support matrix (Phillips, 1989). This antibody-coated matrix is then packed into a column and a solution containing the specific antigen is passed over the immobilised antibody. During this process, the binding sites of the antibody come in contact with its specific antigen and capture it. Since the antibody is chemically bound to the inert packing matrix, the captured antigen is retained while the unreactive materials pass through the column in the mobile phase. The captured antigen is released and recovered by changing the nature of the mobile phase in such a way that dissociation of the immobilised antibody-antigen complex is achieved. Once the complex is fully dissociated, the

Chapter 6 165 antigen is released into the mobile phase and passes through the column where it is collected and measured as illustrated in Figure 6-1.

Dissociation agent C □

Figure 6-1 Diagrammatic representation of the mechanism involved in immunoaffinity chromatography. (A) The material to be purified (triangular shape) selecfivefy attaches to the immobilised ligand (antibody) whereas the non-immunoreactive material (square) passes through the column. (B) The formation of the antibody-antigen complex. (C) The introduction of a dissociation agent which breaks the bonds of the complex and releases the required purified material

The general reactions governing immunoaffinity chromatography are summarised in the following two equations:

Immobilised antibody + A n tig en t ^ Immobilised complex

Immobilised complex + Elution agent t ^ Free antigen

6.1.3 Antibody-antigen interaction The antigen-antibody interaction takes place by the formation of multiple non-covalent bonds between the antigen and the amino acids of the antibody binding sites. These non-covalent bonds are created by the effect of four weak bonding forces (hydrogen bonds, electrostatic attraction. Van der Waals forces and hydrophobic bonds). Although the attractive forces involved in these bonds are individually weak by comparison with covalent bonds, the multiplicity of the bonds leads to a considerable binding energy (Roitt et al, 1985). Furthermore, the non-covalent bonds are critically dependent on the distance between the

Chapter 6 166 interacting groups. Thus, the interacting groups must be close in molecular terms before these forces become significant. As a result, the antigenic determinant and antigen combining sites must have complementary structures to enable them to combme with sufficient binding energy. Figure 6-2. Moreover, although antigen-antibody reactions can show a high level of specificity (Roitt, 1991), a degree of antibody cross-reactivity may be observed towards compounds possessing structures similar to the original antigen.

an tibody a n tig en

hydrogen H bonding H + dectxostatic

Van der Waals > <

hydrophobic iiiiiiiiiiiiii _ water excluded % iiiiiiiiiiiiiiiiiiiiiiiiiiiir

Figure 6-2 The intermolecular attractive forces binding antigen to antibody (adapted from Roitt et al, 1985).

Simple devices for performing immunological assays in non-laboratory settings (e.g. home, bedside, doctor's office and out-patient clinic) are a recent innovation (Valkirs and Barton, 1985), (Osikowicz et al, 1990). There are a number of reasons for performing extra-laboratory assays and these include: a medical emergency, doctor-patient convenience and monitoring or screening (i.e. to assist rapid screening procedures for large numbers of contaminated samples). Immunological tests for detecting pregnancy (hCG in urine) and ovulation (FSH in urine) were the first tests to become widely available, and today many examples of these test devices are commercially available (Kricka, 1993). Their success has fuelled the development of a wide range of other tests and test devices. An enzyme immunochromatography method for measuring theophylline

Chapter 6 167 has been published (Zuk et al, 1985), the method providing a novel test strip immunoassay for quantifying the drug in biological fluid, and it is also well suited for on-site testing in non-laboratory environments. Moreover, a semi quantitative method for measuring the immunoreactivity of triazine was reported using an enhanced luminescent immunoassay with a photographic end point

(Hardcastle et al, 1989). The object of this study W cis to develop immunoaffinity chromatography columns for the detection of pesticides in water residues using isoproturon and chlortoluron as model compounds, and also to determine their feasibihty as simple qualitative systems suitable for "in the field" water monitoring.

6.2 ISOPROTURON IMMUNOAFFINITY COLUMNS 6 .2 .1 Immobilisation of antibodies and column preparation 6.2.1.1 AfEimty support matrices A good support matrix suitable for immunoaffinity chromatography must possess several specialised properties such as a large surface area which is mechanically stable, and possess suitable side-chains which can easily be modified for ligand attachment. The structure must also be chemically inert to ensure that non-specific adsorption does not take place during the affinity separation (Weetall, 1973). Table 6-1 summarises the requirements for an ideal affinity support matrix.

Table 6 -1 Properties required by the affinity support matrix (Weetall, 1973).

Properties o f affinity matrices

1. Mechanically stable 2. Chemically stable, especially under elution conditions 3. Possess a surface chemistry which is easily derivatised 4. Possess a surface which provides easy ligand access 5. Possess a surface which is inert and does not adsorb materials non- specifically 6. Give good flow-rate characteristics when packed in a column

Chapter 6 168 Currently, although there are many commercially available matrices to choose from (Table 6-2), porous silica (PS) matrix was chosen for this work since it is rigid and mechanically stable. It also has a high particle density allowing rapid settled bed formation. Furthermore, different iomc strengths or pH have minimal effects on the bed volume of the silica-based matrix in comparison to the organic matrices such as Sepharose or Sephadex. Porous silica matrices have the advantage of being exceptionally cheap (Godfrey, 1993).

Table 6-2 Immunoaffinity chromatographic matrices.

Matrix Characteristics Example

Dextran Chemically stable Sephadex (Pharmacia Ltd) Agarose Cheap Sepharose (Pharmacia Ltd) Cellulose Solvent resistant ExceUulose (Pierce Europe) Polyacrylamide pH stable (1-10) Bio-Gel P. (Bio-Rad Labs) Porous silica Rigid, h ig i flow rates Matrex (Amicon) Glass Rigid, high pore stability Controlled pore glass CPG

6.2.1.2 Antibody immolnlisation AH polyclonal antibodies usually need some form of clean-up procedure (i.e. ion exchange chromatography on DEAE or salt precipitation with ammonium sulphate) before they can be used as an immunoaffinity ligand (Phillips, 1989). However, solid phases made by immobilising whole antiserum have longer shelf-lives than those prepared using purified IgGs. This could possibly be due to the protein presence in antisera providing mechanical and/or physicochemical support to immobilised antibodies ((jkxifrey, 1993). Therefore, non-purified antiserum of isoproturon 4676 (Bleed IV:B) was utilised for the preparation of affinity columns. The binding of isoproturon antibodies, via their amino groups, to aldehyde activated porous silica (supplied by ClifMar Associates Limited) formed Schfff s base bonds. These bonds were found to be unstable in aqueous conditions and slowly hydrolysed allowing the antibodies to leach from solid phases. Therefore,

Chapter 6 169 sodium cyanoborohydride was used to stabilise the Schiffs base bonds whilst leaving the aldehyde groups and the ligand (antibody) on the solid phase undiminished (Godfrey, 1993). The amount of antiserum to be immobilised was estimated empirically to cover a wide range of antibody concentrations (2.5, 5, 10, 20, 40, 80, 160, 320, 640 and 1280 /d neat antiserum) vhile the protocol was carried out as follows: Ten polypropylene disposable separation columns (11.0 x 1.0 cm; Lab M No. D823, UK), each provided with a polyethylene matrix support frit, were each packed with 0.5 g aldehyde activated silica. Each column was washed with 50ml PBS buffer to remove any remaining trace of g^utaraldehyde on the solid phase. Next, 5ml of PBS buffer was dispensed into each column followed ly the addition of the appropriate amount of the neat isoproturon antiserum. The columns were then closed from both sides and left rolling on a rotamixer for two hours at room temperature. Once again each column was washed with 10ml PBS buffer, and 5ml of lOmM sodium cyanoborohydride (NaBHgCN) pH 6, made up with IM glycine buffer, was carefully added to individual columns which were then rotated overnight at room temperature. Next day, each column was washed with 10ml of 0-3% HCl pH 2 followed with 20ml PBS buffer. Finally, the columns were stored at 4°C after the addition of 5ml 0.14^M lactose solution, containing 0.1% thimerosal, into each column.

6 .2 .2 Optimisation of the elution mnditions using isoproturon label The dissociation of an antibody-antigen complex can be achieved by a variety of techniques such as the introduction of excessive hydrogen or hydroxyl ions by changing the pH of the elution buffer (Phillips, 1985). In a similar manner, the ionic strengths of the elution buffer could be altered by the addition of chaotropic salts such as 3M sodium thiocyanate and 4M sodium chloride (Phillips, 1988). Furthermore, an antibody-antigen complex can be split by the addition of any solution which contains polarity-reducing agents (i.e. ethylene glycol, methanol and ethanol). These agents act by reducing the polarity of the solution surrounding the antibody-antigen complex and thus neutralising the hydrophobic forces responsible for the attraction (Chase, 1983). Initially, 0.3% hydrochloric acid (HCl) was utilised as an elution buffer (advised by ClifMar Associates Limited) to determine its potential for eluting the retained fractions on the column. One of the columns prepared previously.

Chapter 6 170 containing 80/d isoproturon antiserum (chosen randomly), was used for this optimisation study. The procedure was executed as follows: 10ml of isoproturon label (163 6q/ml), made up with PBS buffer, were added onto the column followed by the addition of 20ml PBS buffer to remove any unbound radiolabel from the column. 40ml of 0.3% HCl pH 2 were then applied onto the column to elute the retained labelled fractions. The eluate from the entire purification process was collected (10ml/fraction) under gravity flow. Finally, 0.5ml from each collected fraction were transferred into scintillation vials, 4ml of OptiPhase added and the radioactivity in each vial counted. The results in Figure 6-3 show that only 26% radioactivity of the retained label on the solid phase was recovered by the use of 40ml 0.3% HCl (-33 column volumes). It has been reported that hydrophobic bonds may contribute up to half the total strength of the antigen-antibody bond (Roitt et al, 1985). Thus, one of the polarity reducing agents, ethanol, was investigated ly adding various percentage volumes of it to the elution buffer (0.3% HCl pH 2). The optimisation procedure for the percentage of ethanol required to be added to the elution buffer was as follows: 10ml of the above isoproturon label solution were added into the immunoaffinity column followed by the addition of 20ml of PBS as a washing buffer. 10ml from each elution buffer, containing different percentages of ethanol (0, 1%, 5%, 10%, 15%, 20%, 25%, 30%, 35% ) were applied onto the column. Meanwhile, fractions of 10ml each from the wash and elution were collected and measured for their ^"^C radioactivity as described before. Figure 6-4 shows a significant improvement in the elution profile achieved by the addition of 35% ethanol to 0.3% HCl elution buffer pH 2. The purification process was repeated five times using 35% ethanol to confirm the reproducibility of the elution profile and to assess the dénaturation effect of the elution buffer on the immobilised antibody. Consequently, 98% of the retained isoproturon label was recovered by the addition of eight column volumes (10ml) of the elution buffer, v/v 0.3% HCl pH 2 containing 35% ethanol. Figure 6-5, and thus it was used as an elution buffer during the development process.

Chapter 6 171 6 .2 .3 Theoreticcd binding capacUg o j the colu m n s The theoretical binding capacity of an immunoaffinity column (qm) is defined as the maximum amount of the antigen that can be bound by ligand (immobilised antibodies), qm can be calculated from the breakthrough curve, defined as the variation of the ligate concentration in the bed outlet plotted as C/Co whereby C is the ligate concentration in the outlet of the column and C is the ligate concentration in the inlet (Chase, 1984). The performance of the adsorption stage of a packed affinity process can best be determined h y jro n tcd a n a lysis, in which a fixed concentration of ligate (Cq) is continuously applied to the immunoaffinity column whilst the variation in ligate concentrations (C) is monitored. In this study, the procedure for isoproturon was carried out as follows: 30ml isoproturon (Ing/ml), in tap water, was spiked with 125/d i^^C isoproturon (163 6q/ml) and applied into each isoproturon immunoaffinity column prepared earlier, containing 0, 2.5, 5, 10, 20, 40, 80, 160, 320, 640 and 1280 pel isoproturon antiserum, while eluate fractions of 1ml were collected firom the columns under gravity flow and the amount of radioactivity in each fraction determined- Finally, each column was washed with 10ml of PBS buffer followed by the addition of 5ml of elution buffer. Figure 6-6 was constructed ly plotting C/Co (expressed as a percentage) against the number of eluate fractions. It shows that the immobilisation of 20/d of isoproturon antiserum resulted in the steepest slope of the breakthrough curve indicating the lowest possible concentration of antibody able to be immobilised onto the support matrix whilst retaining a significant binding capacity. For each column the volume of analyte solution (V) required to saturate 50% of the column binding capacity was determined by drawing a line parallel to the abscissa from the point which represents 50% of the maximum value of the curve's C/Cq and dropping a vertical to the abscissa. Thus, qn, was calculated from the following equation:

9 ^ = v x a where the theoretical binding capacity. V = volume of analyte solution. Co = the concentration of the ligate solution applied to the column.

Chapter 6 172 Figure 6-7 shows a plot of for columns against the amount of ligand (isoproturon antiserum) used for immobilisation. The lower the binding capacity of the immunoaffinity column the lower the limit of detection produced. Hence, the addition of 20/d isoproturon antiserum to a 0.5g diy activated porous silica showed the lowest q^ value (8.55 ng/1.2g wet weight solid phase). On the other hand, the immobilisation of less than 20/d antiserum showed an increase in the non-specific binding as compared to the binding obtained by zero antiserum column (containing pre-immune serum). This might be due to insufficient antibodies being immobilised in the correct orientation on the support matrix (possibly by attachment via amino groups in the hypervariable region of the antibody), and hence 20pi\ isoproturon antiserum solid phase was chosen for further investigation. In addition, columns packed with activated and inactivated silica were utilised to eliminate the possibility of any non-specific binding which could have emerged by the support matrix. The detection of pesticides in water at a level of O.l/tg/L requires further reduction in the theoretical binding capacity of the column. This was met by paclring various amounts of the 20/d antiserum solid phase (50, 100, 150, 300 and 600 mg wet weight solid phase) into smaller plastic disposable solid phase extraction (SPE) columns (IX 0.2 cm) already provided with firits. The theoretical binding capacity for each column was then measured as described previously. Figure 6-8 shows that the lowest q%n value can be achieved by packing 50mg wet weight solid phase into an immunoaffinity column (1.2 ng/50 mg wet weight solid phase). Consequently, the amount of 50 mg wet weight solid phase was selected for further development.

6 .2 .4 IsopTOtUTon standard curve using isoproturon label Isoproturon standard curves were obtained on immunoaffinity columns using two types of assay format, direct (the reading is directly proportional to the analyte concentration which is displacing the analyte label) and indirect (the reading is indirectly proportional to the concentration of the analyte which is competing with the analyte label in the mixture) competitive assays. Columns were chosen randomly for standard solution application.

Chapter 6 173 CPM (Thousands) 100-

Total EL W1 W2 E l E2 E3 14C isoproturon (lOml/fraction)

Figure 6-3 Purification profile of the ^ isoproturon solution using an immunoaffinity column with 0.3% HCl, pH 2, as the elution buffer. EL, excess label not bound to the column; W, trapped label recovered by washing; E, eluting specific bound anafyte.

Chapter 6 174 CPM (Thousands) 100 1 90 - 80- 70- 60- 50- 40 - 30-

2 0 -

1 0 -

Total EL W1 W2 E0% El% E5% E10% E15% E20% E25% E30% E35% 14C isoproturon (lOml/fraction)

Figure 6-4 The efifect of adding different percentages of ethanol to the 0.3% HCl elution buffer, pH 2, on the elution profile of isoproturon label using an immunoaffinity column. EL, excess label not bound to the column; W, trapped label recovered by washing; E, eluting specific bound analyte.

Chapter 6 175 CPM (Thousands) 80 75 - 70 - 65 - 60 - 55 - 50 - 45 - 40 - 35 - 30 - 25 -

20 - 15 -

10 - 5 - 0 ^ Total EL W1 W2 El E2 E3 E4 14C iscproturon (lOml/fraction)

Figure 6-5 Purification profile (n=5) of the isoproturon label using a mixture of 0.3% HCl, pH 2, and 35% etheinol as an elution buffer, EL, excess label not bound to the column; W, trapped label recovered by washing; E, eluting specific bound analyte.

Chapter 6 176 C/Co% 120 110 100 90 80 “ 0 pi Ab

- 4- 2.5pl Ab 70 5 pi Ab 60 - B - 10 pi Ab 50 20 pi Ab 40 40 pi Ab 30 20 10

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 Eluate fractions (Iml/fraction)

C/Co% 120 110 100 90

—^ 80 pi Ab

70 - 4- 160 pi Ab 60 -4f- 320 pi Ab 50 ~ B - 640 pi Ab 40 1280 pi Ab 30 20 10

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 Eluate fractions (Iml/fraction)

Figure 6-6 The breakthrough curves for Isoproturon immunoafOnity columns packed with various concentrations of antiserum. Legends show volumes of antibody bound to 1.2g solid phase per column.

Chapter 6 177 qm 40

15 -

0 2.5 5 10 20 40 80 160 320 640 Antiserum immobilised (pi)

Figure 6-7 The theoretical binding capacities of isoproturon immunoaffinity columns, ng/1.2g wet weight solid phase, packed with different concentrations of antiserum.

Chapter 6 178 qm 10

50 100 150 300 600 Wet weight solid phase (mg/column)

Figure 6-8 The theoretical binding capacity of isoproturon nnmunoaJBBnity columns containing various amounts of the immobilised solid phase.

Chapter 6 179 6.2 4.1 A direct competitive assay The assay was carried out as follows: 1ml isoproturon label solution (1.26 G6q/g), in PBS buffer, was applied into each of the six 50mg SPE columns. Each column was then washed with 3ml PBS buffer followed by the addition of 1ml isoproturon standard solutions (0, 0.01, 0.1, 1, 10 and 100 ^g/L) made up with tap water. Finally, 1ml elution buffer was added into every column to recover any retained ligate, and columns were regenerated by passing through 3ml PBS buffer. The eluate for aU purification stages was collected at 1ml per fraction under gravity flow and the radioactivity measured. It was observed that the addition of isoproturon did not displace more than 10% label applied into the column at all concentrations tested. This could possibly be due to the fact that the equilibrium between the two competing components (the retained radiolabel and the standard) and the antibodies was not achieved, as the passage of standard solution through the column took place in a short period of time (lOmin).

6.2 4.2 An indirect competitive assay The procedure was performed as follows: Six 0.5ml isoproturon label solution (1.26 G6q/g), in PBS bufier, were mixed with 0.5ml of a standard isoproturon solution (0, 0.01, 0.1, 1, 10 and lOO^g/L) which was made up with tap water. The six mixtures were applied onto six 50mg SPE columns independently. The columns were then washed with 3ml PBS buffer each followed by the addition of 1ml elution buffer into individual columns. Similarly, all columns were regenerated by passing through 3ml PBS buffer through each column. Radioactivities in the collected fractions were determined as before. Figure 6-9 shows the mean of six isoproturon standard curves with a limit of detection of 1.4 /^g/L (calculated by three standard deviation fall from binding at Bo). Although the columns were chosen randomly, every time the standard curve was conducted variation between-columns was less than 9.2%. The interpretation of the results could be that both species of isoproturon have equal chance of binding to the immobilised antibodies during the passage period of the mixture through the column. Hence, the format of an indirect competitive assay was adopted for further system development.

Chapter 6 180 B/Bo% CV% 120 r 50 110.: 45 100)^ 40 9 0 - 35 80 - 70 - 30 60 - 25 50 - 20 40 - 15 30 - 10 20^;.

0.001 0.01 0.1 1 10 100 1000 Isoproturon pg/L

Figure 6-9 The mean of six isoproturon standard curves performed by an indirect competitive assay on immunoaffinity columns using isoproturon radiolabel. { I ), standard deviation; (-X -), mean of six standards; (-A-), coefficient of variation.

Chapter 6 181 6.2.5 Optimisation of the conditions using HRPO isoproturon label The novelty of this work is the development of immunoaffinity columns for monitoring pesticides "in the field" with the ability of visually distinguishing samples with a concentration of pesticides above the MAC compared to a negative control. This was achieved by performing a colorimetric end point assay on the columns, and thus the use of an enzyme label was essential for the development of colour.

6.2.5.1 Choice of elution buffer It was thought that the low pH used in the previous elution buffer (mixture of 0.3 HCl and 35% ethanol, pH 2) was likely to cause some dénaturation of the HRPO enzyme protein. Also, the presence of 35% ethanol in the above buffer might have a negative effect on the enzyme activity. Hence, other elution buffers which possess the ability to reduce the hydrophobic interaction between an Ab-Ag complex and maintain integrity of the elution were investigated, and the procedure was as follows: Seven dilutions of pure bovine peroxidase enzyme (7.8, 15.65, 31.25, 62.5, 125, 250 and 500 ng/L), in PBS buffer, were used along with the following elution buffers: 0.3% HCl pH 2, 0.3%HC1 and 35% ethanol pH 2, ethanol, PBS buffer pH 7.4 with 35% ethanol, 3M sodium thiocyanate, 5M sodium chloride, PBS buffer pH 7.4 with 50% ethylene glycol. 100/d/well of each elution buffer were added to the seven duplicate dilutions of peroxidase (100/^1/well) bringing the total of each well to 200/d/well. The plates were incubated for five minutes at room temperature followed by the addition of 50/d/well TMB into all wells and the plates were then left incubating for another five minutes. Finally, the reaction was stopped ly the addition of 50/d/well IM HCl and the colour was read at 450 nm using the Labsystems Multiskan plate reader. Figure 6-10 demonstrates a significant denaturing effect when a low pH solution or 100% ethanol solvent were incubated with the enzyme, hi comparison to the other elution buffers used, a negligible denaturing effect was observed on the enzyme activity by incubating a mixture of PBS buffer pH 7.4 with 35% ethanol. Consequently, the mixture (PBS buffer pH 7.4 with 35% ethanol) was then investigated for its potential as an elution buffer.

Chapter 6 182 OD (450 nm) 2.6 2.4 2.2 0.3% HCl 1.8 - f - 0.3% HC1&35% ethanol 1.6 ethanol 1.4 -B- PBS & 35% ethanol 1.2 —X— 3M Na thiocyanate - e - 5M NaCl 0.8 PBS & 50% glycol 0.6 - g - PBS 0.4 0.2

1 10 100 1000 Peroxidase enzyme ng/L

Figure 6-10 The effect of various elution buffers on the peroxidase protein activity. Legend shows the different solutions used for elution.

Chapter 6 183 6.2 5.2 Standard curve using isoproturon peroxidase enagmie label Columns were selected randomly for standard solution application. The procedure was carried out as described in 6.2.4.2, using the isoproturon peroxidase label, as follows: Six 0.5ml aliquots of HRPO isoproturon label solution, in PBS buffer diluted at 1/4000 (0.575/(g/ml peroxidase), were mixed with 0.5ml of a standard isoproturon solution (0, 0.01, 0.1 1, 10 and 100 ^g/L) made up in tap water and incubated at 4°C. The six mixtures were applied into six 50mg SPE columns individually. The columns were then washed with 3ml PBS buffer each followed by the addition of 1ml elution buffer (mixture of PBS buffer pH 7.4 with 35% ethanol). All columns were regenerated by passing 3ml PBS buffer through each column. The eluate was collected at 1ml per fraction under gravity flow. Figure 6- 11 illustrates the stages involved in a competitive enzyme immunoassay system using immunoaffinity columns. Finally, 100^1 of each collected fraction were dispensed into the wells of microtitre plates ia duplicate followed ly the addition of 100/m1 TMB chromagen-substrate. After incubating the plates for 5 min at room temperature, the reaction was stopped the addition of 50/fl/weU IM HCl and the colour was read at 450 nm. The standard curve procedure was repeated six times to estimate the variation between-columns while columns were chosen randomly each time. Figure 6-12 demonstrates the mean of six standard curves obtained using six immunoaffinity columns, each containing 50mg solid phase, with a detection limit of l^fg/L (calculated by 3 SD fall from binding at Bo). The inter-column variation of the standard curves was less than 9.3 % even though columns were selected randomly for the application of the standard solution. Meanwhile, a minimal background reading was observed at the washing step using PBS buffer, indicating a minimum leakage of the captured analyte through the columns. Furthermore, the mixture of PBS buffer pH 7.4 with 35% ethanol showed remarkable abihty to elute the bound analytes (the labelled and unlabelled isoproturon) from the columns without any significant denaturing effects on the HRPO protein.

Chapter 6 184 1 ImmunoafSnity column

2> Add mixture of labelled 3 Wash with PBS buffer to and unlabelled analytes remove unbound analytes > k(AN k«'^n

> <1

> % 4- Elute the retained analytes 5 Measure the eluted enzyme activity

K<>Nk<< M % ‘ 'V i

Figure 6-11 Presentation of a competitive enzyme immunoassay system using immunoaffinity columns. The triangular symbols represent analytes and the triangle attached to the encircled E represents labelled analytes.

Chapter 6 185 B/Bo% CV% 110 50 100 45 90 40 35 70 30 60 25 20 40 15 30 20 10

0.001 0.01 0.1 1 10 100 1000 Isoproturon pg/L

Figure 6-12 Inter-column variation of Isoproturon immunoaflSuaity column for six standard curves. { I ), standard deviation; (-X-), mean of six standards; (-A-), coefficient of variation; (-X-). non-specific binding background during the washing process.

Chapter 6 186 6.2.5 3 Choice of elution buffer volume The minimum volume of the elution buffer required to elute all the retained analytes from a 50mg isoproturon SPE column was optimised as follows: Four 1ml aliquots of isoproturon enzyme label (1/4000 dilution in PBS buffer) were applied into four 50mg SPE columns. The columns were each washed with 3ml PBS buffer followed by the addition of the elution buffer at different volumes (1, 0.75, 0.5 and 0.25 ml) into each column. Finally, the enzyme activity in the eluate fractions was determined as described previously. The addition of 0.5, 0.75 and 1 ml elution buffer resulted in similar analyte recoveries (94%) as shown in Figure 6-13. Furthermore, using 0.5ml elution buffer facilitates a reduction in the required assay time, and minimises any possible long term dénaturation effect on the immobilised antibodies, and thus it was chosen for further development. On the other hand, the addition of 0.25ml elution buffer showed a recovery of only 52%.

6.2.5 4 Choice of isoproturon enzyme peroxidase concentration The optimisation procedure was carried out as described in G.2.5.2 using four dilutions of the isoproturon enzyme label (1/4000, 1/8000, 1/16000 and 1/32000), in PBS buffer, to conduct four standard curves. 0.5ml elution buffer was utilised for each column to elute all the retained analyte. Subsequently, the eluate was collected at Iml/fraction and measured for its enzyme activity as outlined previously. As a result, the addition of 0.144 and 0.072/fg/ml (1/16000 and 1/32000 dilutions) isoproturon enzyme label showed a significant improvement in the sensitivity of the standard curve in comparison to the other enzyme dilutions used as it possessed the steepest slope with high OD reading. Figure 6-14. However, the OD of the Bo was markedly reduced when 1/32000 label dilution was used, and thus, 1 /16000 dilution of isoproturon label solution (0.144jUg/ml peroxidase) was selected for further research. Furthermore, the addition of 2.5% normal serum to the enzyme label solution as explained in 3.8.1.4.3 was investigated, and the assay was repeated five times to confirm the reproducibility of the results. Figure 6-15 demonstrates a slight increase in the sensitivity of the standard curve, 0.48 pg/L, when 2.5% normal serum was applied to the label solution, and hence the addition of 2.5% serum was adopted for ffirther development.

Chapter 6 187 OD (450 nm)

I H Im l ■ 0.75ml iiiil 0.5ml M 0.25ml

Total W l W2 W3 E Elution profile of enzyme isoproturon

Figure 6-13 Optimisation of the elution buffer (mixture of PBS buffer and 35% ethanol) volume required to recover the retained Isoproturon enzyme label solution. EL, excess label not bound to the column; W, trapped label recovered by washing; E, eluting specific bound analyte. Legend shows volumes used for elution.

Chapter 6 188 B/Bo% llOr 100 90

— 1/4000 label 60 1/8000 label

Bo-1.21 ^ 1/16000 label 40 -a- 1/32000 label 30 B o-0.32 20

0.001 0.01 0.1 1 10 100 1000 Isoproturon pg/L

Figure 6-14 Optimisation of the isoproturon enzyme label dilutions to enhance the sensitivity of the isoproturon standard curve using immunoaffinity columns. Legend shows dilutions of HRPO-conjugate.

Chapter 6 189 B/Bo% CV% 110-r 50 100 45 40 80 B o-0.75 35 70 30 60 25 50 20 40 15 30 10

0.001 0.01 0.1 1 10 100 1000 Isoproturon pg/L

Figure 6-15 The effect of adding 2.5% normal serum to the isoproturon label solution on the sensitivity of isoproturon standeird curve using immunoaffinity columns. ( I ), standard deviation; {-X -), mean of five standards; (-A-), coefficient of variation.

Chapter 6 190 6.2.5 5 Optimisation of the analyte sample mixture volume applied onto the colum n The optimisation process for the amount of sample mixture (HRPO label with isoproturon standard) required to saturate 50mg isoproturon SPE column under gravity flow was executed using a comparison between five standard curves in the range 0-100 /^g/L (six standards, in tap water). Each curve was constructed consecutively adding different volumes of mixture (0.5, 1, 2, 2.5 and 3 ml) to each SPE column. The assay was carried out as described in 6.2.5.2 using the optimum dilution of the enzyme labelled solution (1 /16000, containing 2.5% normal serum) and also 0.5ml elution buffer volume. The addition of 2ml mixture solution showed a marked improvement in the sensitivity of the standard curve as shown in Figure 6-16. The results could be justified as the duration of passage through the column is increased when a large volume of analytes solution is added, thus increasing the possibility of the isoproturon standard and label to bind to the immobilised antibodies and reach an equilibrium. On the other hand, the addition of 2.5 and 3ml of mixture did not show any further improvement in the sensitivity of the standard curve which might suggest that the column had already been saturated with analytes by a lower mixture solution volume (i.e. 2ml).

6 .2 .6 Validation of the isoproturon standard curve tising immunoaffinity colum ns 6.2.6.1 Indirect competitive isoproturon assay final protocol Using the information gained firom the preceding experiments the following protocol was finally adopted: Six 1ml aliquots of HRPO isoproturon label solution (0.144 ^g/m l peroxidase) in PBS buffer diluted at 1/16000 and containing 2.5% normal serum, were mixed with 1ml of a standard isoproturon solution (0, 0.01, 0.1, 1, 10 and 100 fig/D prepared in tap water. The six 2ml mixtures were applied onto six 50mg SPE columns and each mixture was collected in one fraction / colum n- The columns were each washed with 3ml PBS buffer followed by the addition of 0.5ml elution buffer (mixture of PBS buffer pH 7.4 with 35% ethanol). One fraction /column of the elution buffer was collected and the enzyme activity in each fraction was detected as outlined before. Finally, all columns were regenerated by passing 3ml PBS buffer through each column under gravity flow (0.5 m l/min).

Chapter 6 191 B/Bo% 110 r 100 90 80 70 — 3ml H — 2.5m l 60 2m 1 50 Bo-0.52 -S ~ 1ml 40 0.5ml Bo-0.71 30 Bo-0.78 20

0.001 0.01 0.11 10 100 1000 Isoproturon pg/L

Figure 6-16 Optimisation of the analyte mixture volume required to saturate the 50mg isoproturon SPE column. Legend shows volume of original anafyte mixture used.

Chapter 6 192 6.2 6.2 Assay sensitivity and reproducibility '^Inter-column vcwicUion" Between-columii variation was determined for the six standard isoproturon solutions (0, 0.01, 0.1, 1, 10 and 100 fig/L) by constructing 10 standard curves using the final protocol. The mean, expressed as a percentage of the maximum binding B/Bo, SD and CV for each isoproturon standard are shown in Table 6-3, The limit of detection of the standard curve, defined as the concentration equivalent to a three standard deviation(s) fall firom binding at Bo (containing no isoproturon standard), was 0.15 fig/L. Moreover, the inter-column variation between isoproturon standard curves ranged firom 5.8-1 o.1% across the curve.

T a b le 6 -3 Inter-column variation for standard curve using isoproturon immunoaffinity columns.

Isoproturon standard M ean(n= 10) SD CV% fig/L BfBo%

100 4 0.23 5.8 10 23 2.1 9.1 1 47 3.5 7.4 0.1 86 8.7 lo.i 0.01 97 9.5 9.8 0 (Bo=0.75) 100 9.3 9.3

6 2.6.3 Stability and preparation of the analyte mixture The analyte mixtures for six isoproturon standard solutions (0, 0.01, 0.1, 1, 10 and 100 f^g/L) were prepared ly mixing 25ml of each standard solution, in tap water, with 25ml isoproturon label solution (0.144 fig/L peroxidase) containing 2.5% normal serum. Since better stability of isoproturon was obtained with low pH as explained in 3.8.1.4.7, the mixtures were then adjusted to pH 5 by adding 5/d IM HCl into 2ml mixture solution and appUed into six isoproturon SPE columns utilising the final protocol to construct a standard curve. Ten standard curves were consequently constructed over a period of one

Chapter 6 193 month to assess day to day variation. This was compared with standard curves in which no pH adjustment took place. Figure 6-17 shows substantial improvement to the limit of detection of the standard curve (0.025 /^g/L) vhen pH 5 was used for the analyte mixture solutions. This might be due to the abihty of the immobilised antibodies, capable of binding with high affinity to the isoproturon standard and label, to disregard the non-specific components present in the mixture solution in conditions of low pH (pH 5). This reduces the binding stabüity of the non-specific compounds thereby favouring dissociation, whilst offering httle resistance to association with the isoproturon label and standard enhancing the sensitivity of the standard curve. Furthermore, the variation between columns across the standard curves, obtained by changing the pH of each sample to 5, was less than 8.7%.

6.2 6.4 Colour developmcait within the columns The objective of this work was to develop immunoaffinity columns capable of visually detecting low levels of pesticides in different matrices. This qualitative detection would be based on colour differences between the positive sample (sample with a concentration of pesticide above 0.1 fig/h) and the zero standard sample. In order to achieve this, the chromagen-substrate (TMB) was appfied directly into the columns after the analyte mixture had been added and the colum ns washed. It was thought necessary to reduce the effects on the zero standard sample, to be compared with unknown samples of various matrices. This was performed by mixing an equal volume firom each matrix (River Ouse wrater. River Wey Guildford, University of Surrey campus lake water, tap water, human plasma and human urine) which was then stored at 4°C. The colour development protocol within the columns was carried out as follows: Three 1ml isoproturon standard solutions (0.075, 0.15 and 0.3 /fg/L), in tap water, were mixed with three 1ml isoproturon label solutions at 1/16000 dilution (0.144 fig/ml peroxidase) containing 2.5% normal serum bringing the total volume of each mixture to 2ml. The pH was adjusted as described 6.2.6.3. Each mixture solution plus 2ml zero standard were applied into four 50mg isoproturon SPE column followed ly the addition of 3ml PBS wrashing buffer into each of the columns. Finally, 100/^1/column TMB solution were added facüitating colour production as shown in the photo below.

Chapter 6 194 B/Bo% pH 7.8 CV% llOr 50 Bo‘0.76 100 45 90- 40

70 30 60 25 50 40 30 - 20^^ 10 10 -

0.001 0.01 0.1 10 1001 1000 Isoproturon pg/L

B/Bo% pH 5 CV% 50 45 Bo-0.65 90 - 40 80 35 70 30 60 25 20 40 30 15 10 10

0.001 0.01 0.1 1 10 100 1000 Isoproturon p g/L

Figure 6-17 The effect of reducing the pH of the analytes mixture (pH 5) on the isoproturon standard curve limit of detection. ( I ), standard deviation; (-X-), mean of ten standards; (-A-), coefficient of variation.

Chapter 6 195 0.075 0 / 3 / L

The Photo Illustrates the colour differences between the positive samples containing analyte and zero standard sample using immunoaffinity columns.

Chapter 6 196 The colour produced was inversely proportional to the amount of isoproturon present in the sample. As a result, the mixtures with a standard solution of 0.15, and 0.3//g/L were distinguished visually from the zero standard sample during the first three minutes of TMB addition whilst the contrast between the unknown and the zero samples disappeared after this period. The above procedure was repeated six times using different standard solutions (0.08, 0.1, 0.11, 0.12, 0.13, 0.14, 0.15 and 0.2 /^g/L) and the colour produced was examined by five independent observers to score columns which contain sample with any concentration of isoproturon. Table 6-4 shows that samples with an isoproturon concentration of 0.12 /

Table 6 -4 Accuracy of isoproturon immunoaffinity columns.

Isoproturon concentration (n=6) % Correct answer jug/L P ositive or N egative

0.08 100% N egative 0.1 100% N egative 0.11 100% N egative 0.12 100% Positive 0.13 100% Positive 0.14 1(X)% Positive 0.15 100% Positive 0.2 100% Positive

6 2.6.5 Correlation studies between isoproturon ELISA and immunoaffinity column technique Sixteen samples of various matrices (River Ouse water. River Wey Guildford, University of Surrey campus lake water, Surrey research park site lake water, distilled water, tap water, human plasma and human urine) were spiked with isoproturon standard, by a colleague in the laboratory, at different

Chapter 6 197 concentrations which were undisclosed. The samples were then determmed quantitatively using isoproturon ELISA and qualitatively by the isoproturon immunoaffinity column technique. Moreover, eight independent observers were involved in the qualitative determination of the isoproturon concentration of the spiked samples as bemg positive (isoproturon concentration 0.12 /^g/L or above) or negative (isoproturon concentration below 0.12 fig/h) samples using the standard zero sample as a comparison. Table 6-5 compares the results achieved by the above two systems.

Table 6-5 Comparison of results for isoproturon concentration obtained by ELISA and immunoaffinity column methods.

Sam ple Amount spiked Amount measured Immunoaffinity ELISA colum n No ng/L ng/L Positive or Negative

Tap water 1 0 0 100% N egative 2 0.125 0.139 100% Positive Distilled water 3 0.15 0.161 100% Positive 4 0.2 0.213 100% Positive River Ouse 5 0.03 0.018 100% N egative 6 0.15 0.172 100% Positive River Wey 7 0.11 0.127 100% N egative 8 0,18 0.201 100% Positive Research Park lake 9 0.09 0.077 100% N egative 10 0.17 0.159 100% Positive University lake 11 0.05 0.069 100% N egative 12 0.08 0.101 100% N egative Human plasma 13 0.125 0.157 100% Positive 14 0.25 0.273 100% Positive Human urine 15 0.16 0.144 100% Positive 16 0.32 0.298 100% Positive

Chapter 6 198 Samples spiked with isoproturon above 0.12 /ig/L concentration were scored as being positive samples showing a 100% accuracy. Table 6-5. On the other hand, samples spiked with lower isoproturon concentration than 0.12 ^wg/L were cited as being negative samples with again accuracy of 100%. Furthermore, the results obtained by the ELISA method confirmed those determined by the immunoafBnity column technique.

6.3 CHLORTOLURON IMMUNOAFFINITY COLUMNS 6 .3 .1 Antibody immobilisation The amount of chlortoluron antiserum to be immobilised on the aldehyde activated silica was estimated empirically to cover a wide range of antibody concentrations (0, 2.5, 5, 10, 20, 40, 80, 160, 320, 640 and 1280 y l n eat antiserum) while the protocol was carried out exactly as described in 6.2.1.2. The immunoaffinity columns were then stored at 4®C after the addition of 5ml 0.14/^M lactose solution, containing 0.1% thimerosal, into each column.

6 .3 .2 Assessment of theoretical binding capacity of the columns The performance of the packed chlortoluron ImmunoafOnity columns was determined hyfrontal analysis as follows: 30ml chlortoluron (Ing/ml), in tap water, were spiked with QOfil chlortoluron substock solution (146 6q/ml) and applied into each chlortoluron immunoaffinity column prepared earlier. The eluate from the purification procedure was collected, at 1ml/fraction, under gravity flow and the fractions were then measured for their radioactivity as described in 6.2.3. Finally, columns were regenerated and equilibrated by the addition of 5ml/column elution buffer and 10ml/column PBS washing buffer. The breakthrough curves were plotted as C/Co (expressed as a percentage). Figure 6-18, which shows that the immobilisation of 10^1 of chlortoluron antiserum resulted in a steep breakthrough curve. Furthermore, a low amount of ligate solution (5ml) was required to saturate the 10/<1 chlortoluron antiserum immunoaffinity colum n indicating a low binding capacity.

Chapter 6 199 C/Co% 120 110 100 90 80 0 (j 1 Ab 70 -H— 2.5pl Ab 5 jjl Ab 60 -B - 10 pi Ab 50 20 pi Ab 40 40 pi Ab 30 20 10

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 Eluate fractions (Iml/fraction)

C/Co% 120 110 100 90

— 80 pi Ab 70 - f - 160 pi Ab 60 320 pi Ab 50 -G - 640 pi Ab 40 1280 pi Ab 30 20 10

1 3 5 7 9 11 13 15 17 19 21 23 25 27 29 Eluate fractions (Iml/fraction)

Figure 6-18 The breakthrough curves for chlortoluron immunoaffinity columns packed with various concentrations of antiserum. Legends show volumes of antibody bound to 1.2g solid phase per column.

Chapter 6 2 0 0 The theoretical binding capacity for each column was measured as outlined in 6,2.3. Figure 6-19 shows that the immobilisation of 10/il antiserum gave the lowest q^i value (7.2 ng/1.2 wet weight solid phase). On the other hand, the immobilisation of 2.5 and Sfil antiserum per column showed an increase in the non-specific binding as compared with the binding obtained by the zero antiserum column. Thus, 10/d chlortoluron antiserum solid phase was chosen for further investigation. Various amounts of the solid phase, containing 10/d antiserum, (50, 100, 150, 300, and 600 mg wet weight solid phase) were packed into SPE columns to achieve further reduction in the column binding capacity and the theoretical binding capacity for each column was then measured as described in 6.2.3. mg Figure 6-20 shows the lowest qm value (0.94 ng/50 wet weight solid phase), and therefore it was selected for further development.

6 .3 .3 Optimisation of the chlortoluron enzyme peroxitUise concentration Columns were selected randomly for standard solution application and the procedure was executed by the use of the information gained from the isoproturon development assay as follows: Six 1ml HRPO chlortoluron label solution at various dilutions (1/8000, 1/ 16(K)0, 1/32000 and 1/64000), in PBS buffer, were mixed with 1ml of a standard chlortoluron solution (0, 0.01, 0.1, 1, 10 and 100 fig/L) prepared in tap water. The six 2ml standard mixtures of each label dilution were adjusted to pH 5 and then applied into six 50mg SPE columns separately. The columns were each washed with 3ml PBS washing buffer followed by the addition of 0.5ml elution buffer which was collected at one fraction per column. Finally, the enzyme activity in the collected fractions was determined as outlined previously. Figure 6-21 shows a substantial improvement in the sensitivity of the standard curve when chlortoluron enzyme solution at 1/32000 dilution (6.5 ng/ml peroxidase) was applied (as it possessed the steepest slope with high OD reading). However, the use of 1/64000 dilution caused a significant reduction in the OD readings of the standard curve. Therefore, 1/32000 dilution of the chlortoluron enzyme solution was used to study the inter-column variation by constructing ten standard curves over a period of one month. Table 6-6 demonstrates the variation between the ten curves which ranged from 6.2-9.7% and Figure 6-22 shows the mean of ten standard curves with a detection limit of 0.01/ig/L.

Chapter 6 201 qm 40

2.5 10 20 40 160 320 640 Antiserum immobilised (pi)

Figure 6-19 The theoretical binding capacities of chlortoluron immunoaffinity columns packed with different concentrations of antiserum.

Chapter 6 2 0 2 qm 10

50 100 150 300 600 Wet weight solid phase (mg/column)

Figure 6-20 The theoretical binding capacity of chlortoluron immunoaffinity columns containing various amounts of the immobilised solid phase.

Chapter 6 203 B/Bo% 110 100 90 80 70 — 1/8000 label 60 1/16000 label 50 1/32000 label 40 -B- 1/64000 label 30 B o-0.92 20

0.01 0.10.001 1 10 100 1000 Chlortoluron pg/L

Figure 6-21 The effect of different dilutions of the chlortoluron enzyme label upon the sensitivity of the standard curve using chlortoluron immunoaffinily columns. Legend shows the dilutions of HRPO-conjugate used.

Chapter 6 204 T ab le 6 -6 Inter-column variation for standard curve using chlortoluron immunoaffinity columns.

Chlortoluron standard M ean(n=10) SD CV% fig/L B/Bo%

100 1 0.094 9.4 10 5 0.31 6.2 1 21 1.9 9 0.1 48 4.6 9.6 0.01 79 5.3 6.7 0 (Bo=0.55) 100 9.7 9.7

6 .3 .4 Accuracy of chlortoluron immunocfffinity column system Eight chlortoluron standard solutions (0.09, 0.1, 0.11, 0.12, 0.13, 0.14, 0.15 and 0.2 fig/L), made up with tap water, plus the zero standard sample prepared previously (6.2.6.4) were mixed with chlortoluron enzyme solution at a 1 /32000 dilution as described in 6.2.6.4. The results were repeated six times and estimated by five independent colleagues, as illustrated in Table 6-7, samples with concentrations below 0.11 fig/L being scored as negative (no chlortoluron present), whilst samples with a chlortoluron concentration of 0.11 jug/L and above were scored as being positive (contain chlortoluron).

Chapter 6 205 B/Bo% CV% 50 100 45 9 0 - 40 35 70 30 60 25 50 20 40 30 15 20 10

0.001 0.01 0.1 1 10 100 1000 Chlortoluron pg/L

Figure 6-22 Inter-column variation of chlortoluron standard curves using chlortoluron immunoaffinity columns. ( I ), standard deviation; (-X-), mean of ten standards; (-A-), coefficient of variation.

Chapter 6 206 T able 6 -7 Accuracy of chlortoluron immunoafiinity columns.

Chlortoluron concentration (n=6) % Correct answer fig/L Positive or Negative

0.09 100% N egative 0.1 100% N egative 0.11 100% Positive 0.12 100% Positive 0.13 100% Positive 0.14 100% Positive 0.15 100% Positive 0.2 100% Positive

6 .3 .5 Correlation studies between chlortoluron ELISA and inununoc^^inity column technique Sixteen samples of various matrices (River Ouse water. River Wey Guildford, University of Surrey campus lake water, Surrey research park site lake water, distilled water, tap water, human plasma and human urine) were spiked with chlortoluron standard, by a colleague in the laboratory, at different concentrations which were undisclosed as carried out in 6.2.6.5. The samples were then determined quantitatively and qualitatively using the chlortoluron ELISA and immunoaffinity column technique respectively. Moreover, eight independent observers were involved in the qualitative determination of the chlortoluron concentration of the spiked samples as being positive (chlortoluron concentration 0.11 fig/L or above) or negative (chlortoluron concentration below 0.11 jicg/L) samples using the zero standard sample for a comparison. Table 6-8 compares the results achieved by the above two systems. Samples spiked with chlortoluron concentration of 0.11 fig/L or above were scored as being positive samples with 100% correct results accuracy .whilst samples spiked with chlortoluron concentration below 0.11 fig/L were cited as being negative samples showing again a 100% accuracy of correct results.

Chapter 6 207 Furthermore, the results obtained by the ELISA method confirmed those determined by the immunoaffinity column technique.

T a b le 6-8 Comparison of chlortoluron concentration results obtained by ELISA and immunoaffinity column methods.

Sample Amount spiked Amount measured Immunoc^mity ELISA column No jug/L fig/L Positive or Negative

Tap water 1 0 0 l6o% Negative 2 0.12 0.135 100% Positive Distilled water 3 0.11 0.121 100% Positive 4 0.18 0.169 100% Positive River Ouse 5 0.05 0.022 100% Negative 6 0.09 0.067 100% Negative River Wey 7 0.12 0.133 100% Positive 8 0.2 0.211 100% Positive Research Park lake9 0.1 0.085 100% Negative 10 0.15 0.167 100% Positive University lake 11 0.08 0.098 100% Negative 12 0.02 0.01 100% Negative Hum an plasm a 13 0.15 0.139 100% Positive 14 0.2 0.22 100% Positive Human urine 15 0.1 0.089 100% Negative 16 0.3 0.288 100% Positive

6.4 DISCUSSION Affinity chromatography is a unique separation procedure in which the molecule to be separated is selectively adsorbed by an immobilised ligand. A refinement of this technique has been the development of immunoaffinity chromatography, which adds the selectivity and specificity of immunological reactions to the separation process. In this work, a feasibility study for the rapid

Chapter 6 208 monitoring of pesticides for "in the field" tests using immunoaffinity columns with a colorimetric end point assay was carried out in which antibodies were utilised as the ligands, and the affinity separation results were dependent upon the antibody immunological activity. Two phenyl urea herbicides isoproturon and chlortoluron were chosen as model compounds for the study. The immunoaffinity columns were prepared by the use of aminopropyl porous silica already aldehyde activated. The amino groups on the antibodies were then bound to the aldehyde activated matrices forming Schiffs base bonds. These bonds were reduced to a stable covalent state by the addition of sodium cyanoborohydride reducing the risk of antibody leach from the solid phases. This approach was chosen as the binding of ligands to matrices via amine bonds has previously been reported as the most stable means of their immobihsation (Godfrey, 1993) (Godfrey et al, 1993). The amount of antibody immobilised onto the support matrices was apphed empirically at various concentrations, and the performance of the adsorption stage was determined hyjtontcd analysis. The shape of the adsorptive breakthrough curves obtained by the addition of 20/d isoproturon and 10/d chlortoluron antisera to each 1.2g wet weight silica showed steep curves fridicating the lowest possible concentration of antibodies able to be immobilised on the support matrix whilst retaining significant binding capacity. Also, the immobilisation of 20 and 10 /fl isoproturon and chlortoluron respectively onto each 1.2 g wet weight silica gave lower theoretical binding capacity (qm), 8.55 ng/1.2 g wet weight solid phase isoproturon and 7.2 ng/1.2 wet weight solid phase chlortoluron, than the other amount of antibodies added resulting in columns with a low limit of detection as required to detect the MAC level (0.1 /ig/L). In addition, it was observed that reducing the amount of the packed solid phases to 50mg per column, for both isoproturon and chlortoluron, caused a decrease in the theoretical binding capacity which in turn improved the columns sensitivity even further. The system was initially developed using a label of the related analyte (isoproturon or chlortoluron) to minimise any possibilities of an enzyme label conjugation procedure failure. Furthermore, as it was reported that the hydrophobic bonds may contribute up to half the total strength of the antigen- antibody bond (Roitt et al, 1985), the conditions of the elution buffer were

Chapter 6 209 optimised by the addition of a polarity reducing agent (ethanol) to the buffer. As a result, the mixture of 35% ethanol with HCl pH 2 showed a recovery of more than 90% of the retained antigen using ansilyte label as an indicator. However, no apparent recovery of the retained antigen was obtained when peroxidase en 2yme label antigen was used. This was thought to be due to the denaturating effect of the low pH of the mixture (pH 2). Thus, the addition of 35% ethanol to a PBS buffer pH 7 showed the optimum elution condition to recover more than 95% of the specifically bound antigen to the solid phase (isoproturon or chlortoluron) without any significant effect on the peroxidase enzyme protehi label. The sensitivity of the standard curves for the two phenyl urea herbicides was improved by the addition of 2.5% normal serum to the label solution followed by reducing the concentration of the analyte enzyme label in the mixture (HRPO label/standard) applied into the column. Moreover, increasing the volu m e of the label-standard mixture applied into each column contributed to a further improvement to the sensitivity of the standard curve as it increases the possibility of the standard and label to bind to the immobilised antibodies and reach an equilibrium. Besides, lowering the pH of the label-standard mixture to pH 5 during the inter-column variation study to stabilise the standard over a long period revealed a substantial improvement in the sensitivity of the standard curves. This was reported previously by Godfrey (1992). Low pH facilitates the ability of the immobilised antibodies to reduce the stability of the non-specific compounds binding thereby favouring dissociation whilst offering little resistance to association with the label and standard producing a highly sensitive standard curve. The final limits of detection obtained for both isoproturon and chlortoluron standard curves were 0.02/fg/L and 0.01/ig/L respectively with an inter-column variation of less than 10%, although the columns were selected randomly for each standard application. The addition of the chromagen-substrate TMB directly into the columns was carried out for the development of a visual qualitative system capable of detecting breaches in the MAC level of pesticides in different matrices. The qualitative detection was based on colour differences between samples containing concentrations of pesticides, 0.12/(g/L isoproturon and 0.11/^g/L chlortoluron or above, and the zero standard sample. Furthermore, the colour produced by the

Chapter 6 210 addition of TMB was inversely proportional to the amount of pesticide present in the sample. Meanwhile, the zero standard was prepared by mixing equal volumes from a wide range of different matrices (stored at 4°C) and then used for visual colour comparison with unknown contaminated samples of various matrices. For a duration of 3 min, from the addition of TMB, the colours produced in the sample mixture column and that of the zero standard could be differentiated by eye. The difference in colour production is due to the low concentration of enzyme label retained on the column containing the analyte sample, as the presence of anatyte competes with the enzyme label reducing its retention on the column. In comparison a larger amount of enzyme label is retained on the column with the zero standard. The higher amount of enzyme retained from the zero standard sample resulted in a shorter production time of a visible signal than was the case for samples containing analyte. Samples of different matrices spiked with isoproturon at a concentration below 0.12 /fg/L were hardly differentiated visually from the zero standard sample, whilst samples spiked with isoproturon at a concentration of 0.12 ng/l> and above were visually distinguished from the zero standard sample. The estimation of the results was made by eight independent observers. In addition, eightindependent observers were not able to distinguish samples applied into immunoaffinity columns with chlortoluron concentration below 0.11 fig/L from the zero standard, whereas concentrations of 0.11 /fg/L chlortoluron and above were easily distinguished from the zero. The qualitative results obtained by the immunoaffinity columns technique showed an accuracy of 100% correct results at below or above 0.12/fg/L and 0.11 figfl> for isoproturon and chlortoluron respectively which were then confirmed by those measured by ELISA method. The application of the spiked urine samples into the columns caused a significant decrease in the column's flow rate. Thus, it was necessary to regenerate the columns using lar^r volumes of elution and washing buffer. The reduced flow rate was not observed after the application of spiked plasma samples. This result might be due to the chaotropic action of urine (due to the presence of urea) which may have created a certain degree of dénaturation of the immobilised antibodies resulting in their agglutination, and thus causing a significant decrease in the flow rate of the columns.

Chapter 6 211 This chapter has demonstrated the feasibility of screening pesticides quahtatively with non-instrumental immunoaffinity columns based on the principles of affinity chromatography coupled to enzyme immunoassay. The method gives a qualitative result in 20 min without the requirement of sample dilutions, or separation steps. The characteristics of the technique make it ideally suited for monitoring pesticides on-site in a different range of matrices which could be carried out by a non-skilled person thereto reducing the workload involved in meeting EC regulations. Obviously it would be necessary to make further refinements to the system, especially regarding the colour signal development, to enable it to be used commercially. One alternative could be the optimisation of the chromagen- substrate (TMB) concentration which might result in an increase in the duration of the colour signal difference between the zero and mixture analyte sample. An attempt was carried out by adding lower concentrations of TMB onto the columns, however the intensity of the colour signal obtained was low in all columns making the visual colour comparison unsuccessful. Nevertheless, the desired result may be achieved by using a HRPO precipitating substrate i.e. DAB (3,3-diaminobenzidine tetrahydrochlorlde) vdierel^ electrons are transferred ty HRPO from the DAB to peroxide yielding an insoluble brown-coloured end product (PIERCE, 1994). This brown insoluble marker may contribute to a clearer colour signal difference between the zero and mixture analyte sample. An approach for obtaining quantitative results could possibly be achieved by packing the solid phase into capiHaiy columns. The end of the columns would then be inserted into the mixture anafyte solution, containing a limited amount of label mixed with the analyte sample, and the liquid components would be allowed to migrate up the length of the columns ly capillary action. Thus, as the labelled and unlabelled analyte migrate past the immobilised antibody, they would specifically immuno-bind to the solid phase in the columns. Subsequently, the end of the columns would be inserted again in a developer solution containing TMB for colour signal. Using this approach, the quantification of results would be based on the distribution of the enzyme label thereby measuring the height of the colour migration which in turn could be compared to a conversion table, based on calibrator analysis, to translate the height into concentration units. Nevertheless, even without such refinements, the system

Chapter 6 212 already achieved in this study forms the prototype for exciting developments in the field of environmental monitoring.

Chapter 6 213 CHAPTER SEVEN GENERAL DISCUSSION AND CONCLUSION

Chapter 7 214 7.1 GENERAL DISCUSSION The large number of chemicals used and produced in our everyday lives (i.e. for agriculture, food production, food processing, transport emissions and manufacturing) has led to the need for improved methods to monitor the environment. Environmental analysis is needed to assess the distribution of chemical micropollution, monitoring the extent of pollution over a time period, and as an aid to prevent the spreading of further contamination. Furthermore, increased monitoring of the environment is also required to comply with various national and international regulations aimed to anticipate hazards caused by toxic chemicals and to assess the attendant risks in order to protect human health. The work described in this thesis has concentrated on one area of concern, although the principles could be applied to other areas of environmental analysis. Pesticides are a group of chemical compounds which have contributed tremendously to an overall improvement in agriculture (food production). However, the presence of pesticide residues in water, soil and food is an escalating problem that has aroused public concern over potential health hazards. As a result, the legislation related to pesticides has become more and more stringent over the years, requiring larger numbers of more sensitive tests to be performed on various types of environmental samples. In 1985, the EEC issued a Drinking Water Directive, no. 80/778/EEC, (Bates, 1989) setting standards for the quality of water for human consumption, irrespective of the source. In this Directive, pesticides are considered as a single group, with a Maximum Admissible Concentration (MAC) set at 0.1 /tg/L for any individual substance in the group, and 0.5 /^g/L for total pesticides and related products. It is also important to detect pesticide concentrations for economic reasons. For instance, if a maize crop has been triazine protected, the residual triazine concentration in the soil should be monitored prior to using the same field to grow another vegetable which might be affected by the presence of triazine. Pesticide residue analysis has conventionally been achieved using traditional and approved techniques such as GC , HPLC and mass spectrometry. Although these methods are highly accurate and generally achieve adequate limits of detection, the demands of sample preparation procedures are often time consuming which result in a reduced sample handling capacity. Also they require sophisticated and

Chapter 7 215 expensive equipment, handled by skilled technicians, performing a tedious analytical operation. There is, therefore, a need for simple, fast and inexpensive analyses which could complement the existing techniques allowing technicians to easily and rapidly perform large numbers of assays in the laboratory as a means of screening. Ultimately, the ability to detect pesticide residues at extra laboratory sites for either occupational health reasons or environmental reasons would be an advantage. This thesis describes feasibility studies to show the potential use of immunoassays in improving monitoring programmes of pesticides by providing tests which are sensitive, rapid, cost effective and simple to perform. Moreover, the model compounds studied in this project (isoproturon, chlortoluron, 2,4-D and MCPA) were chosen because of their extensive use in UK agriculture and elsewhere. Besides, health concerns raised by the widespread use of these compounds have made monitoring environmental and biological samples for their presence desirable. The two competitive ELISA methods of isoproturon and chlortoluron developed here were optimised for high sensitivity in order to meet the requirement of the MAC level of 0.1/fg/L for individual pesticides in drinking water. The steps involved in the performance of the assays were simple and required inexpensive equipment. Also, sample pretreatment was not necessary and only small sample volumes were required for each assay resulting in a highly desirable technique for environmental analysis. High sensitivity is not always required by other applications such as air monitoring. The 2,4-D ELISA method developed during this study has a limit of detection suitable for monitoring the work place. Regrettably, the production of a suitable MCPA antibody for ELISA was not successful which could possibly be due to the use of an unsuitable method for preparing the immunogen. However, the ELISA assays for isoproturon, chlortoluron and 2,4-D were reliable and robust from day to day and by using standards prepared in tap water (previously shown to have no effect on the antibody-antigen interaction) limits of detection of 0.03, 0.015 and 50 yMg/L respectively were achieved. The isoproturon and 2,4-D antisera have dernonstrated remarkably high specificity towards their analytes showing a negligible degree of cross-reactivity with various herbicide compounds. Additionally, the chlortoluron antiserum showed a significant degree of cross reactivity with the sub-class of phenylurea herbicides. This could be an

Chapter 7 216 important feature for the chlortoluron assay in screening work whereby a negative result excludes a range of suspected analytes from the same group of compounds from further investigation. A positive result, however, indicates the presence of a class of herbicides and gives an estimate of the amount, thus identifying the method required for confirming the ELISA result. On the other hand, although a number of chromatographic methods have been published in the literature for the determination of sub-classes of pesticides in different matrices, all methods possessed limits of detection in the range of 1-100 /ig/L and they required samplespreconcentration procedure prior to analyte analysis (Walters et al, 1984) (DiCorcia and Marchetti, 1991). Quantitative recoveries of isoproturon, chlortoluron and 2,4-D herbicides were obtained from spiked samples in various matrices (i.e. different water sources, biological fluids) demonstrating the accuracy of the methods and their suitability for environmental application. Also the short duration of the assay (2- 4 hrs) and the ability to analyse relatively large sample numbers simultaneously have increased the potential use of ELISA for pesticides in environmental monitoring. Although the assays used here were shown to be robust, widescale use will rely on commercialisation and appropriate manufacture of kits. The isoproturon and chlortoluron antisera have already been incorporated into ELISA kits. The EnviroGard pesticide detection kits of the two herbicides are now available commercially along with a wide range of other pesticide compounds. The kits offer quantitative laboratory tests for the detection of the related pesticide in water and other aqueous solutions. The ELISA assays are also versatile in that both colour and luminescent end-points could be used providing similar performances. Using the enhanced chemiluminescence immunoassay, semi-quantitative results of chlortoluron levels (i.e. samples identified as greater or less than the MAC) were obtained by visual inspection of a photographic end-point produced from the glowing ELISA plate in a portable camera luminometer. Such assays would be useful for semi- quantitative results in the laboratory or in extra-laboratory sites. The phenylurea herbicide chlortoluron was chosen in this study as a model example whilst any other pesticide compounds could equally be detected and measured provided a suitable antiserum towards the pesticide was available.

Chapter 7 217 The major advantage offered by photographic film as a detection system in luminescence is that of simplicity, a permanent visual record being obtained without the use of complex equipment. Obviously this advantage remained relatively unattractive whilst photographic emulsions required conventional "wet" development, but the availability of the Polaroid and related systems of "instant" development provided a simple inexpensive alternative. An arrangement which made best use of such a detection system was the combination of solid-phase reagents (antibodies immobilised onto microtitre plates, permitting a simultaneous monitoring of as many as 60 wells of the microtitre plate) with a chemiluminescent reaction and photographic detection. Luminescent reactions have several advantages over conventional colorimetric or spectrophotometric reactions since they are rapid, extremely sensitive, and they may be coupled with a wide range of analyses, (Carter and Whitehead, 1982). However, the chemiluminescent end-point immunoassays developed for isoproturon and chlortoluron showed similar sensitivities to those produced by ELISA methods exhibiting similar low intra-plate variation across the standards. The chlortoluron ECLIA, based on the oxidation of lumtnol in the presence of hydrogen peroxide in a reaction catalysed by HRPO, provided a sensitive "threshold" type test for the presence or absence of chlortoluron. The visual estimate of chlortoluron levels from the photograph was accurate and easily distinguished. The response was such that the film was completely exposed by the glowing Bo wells on the plate and the degree of exposure was indirectly proportional to the amount of chlortoluron present in the samples. As a result, samples with chlortoluron concentrations exceeding 0.1/fg/L showed a low degree of exposure in comparison to the zero standard and they were easily identified. Confirmation of chlortoluron concentrations in selected immunopositive samples could be subsequently obtained by more conventional techniques (i.e. chromatographic methods or ELISA). This photographic screening procedure, using an enhanced chemiluminescent immunoassay, offered semi-quantitative values for chlortoluron immunoreactivity within a short duration (30 min provided plates already coated with antibody were available). The assay required simple equipment (i.e. wash and drop bottles) and could be carried out by semi-skilled persons in extra-laboratoiy locations. Additionally, providing all the necessary

Chapter 7 218 controls and reference standards are in place at least 20 samples can be assayed in duplicate on each plate. This technique could have potential in the monitoring of pesticides in other types of environmental samples. These include the screening of soil, food and could also be used for human exposure screening studies (Ireland and Samuel, 1989). Although some sample treatment may be required for these types of matrices, the extraction steps are often much less complicated than those required for conventional analysis. For example, in studies of a kit for detection of triazine herbicides in food, a typical residue extraction solvent (acetonitrile) was used and then diluted with water to levels tolerated by the im m unoassay (Glower, 1991). As a large quantity of antisera was readily available in stock, it provided the impetus for investigating the possible development of portable test kits. Feasibility studies of antibodies bound to solid phases (porous silica) and packed into disposable columns (immunoaffinily columns) were carried out to illustrate the possibility of providing extra-laboratory tests for monitoring pesticides. This enables the detection at very low concentrations, in a wide range of environmental samples by unskilled personnel. Again isoproturon and chlortoluron were chosen in this study as examples whilst other pesticides could have equally been used provided a suitable antiserum existed. After optimising the assay conditions, with regard to the amount of antibodies required to be bound to the porous silica, label dilutions and elution buffer, the systems demonstrated sensitivities of 0.025 and 0.01 pg/L for isoproturon and chlortoluron respectively. The addition of the chromagen- substrate TMB into the related immunoaffinity columns, after the application of sample-label solution mixtures, facilitated the development of a portable assay capable of screening samples for the presence of isoproturon and chlortoluron in various matrices. Consequently, samples with contamination levels of 0.12 fig/h isoproturon and 0.11 pg/L chlortoluron and above were visually identified as positive samples in comparison to the zero standard sample. The progression of colour produced hy samples containing the above levels of herbicides were slower than samples containing concentrations lower or no herbicides. This is due to the amount of label retained on the columns being low (i.e. analyte/analyte label competition) in those containing the contaminated samples. Furthermore, the ability of estimating the two herbicides in various water sources and biological

Chapter 7 219 fluids has added another important feature to the technique as sample preparation was no longer required. Consequently, the differences in colour intensity between the positive and negative samples provided an attractive method for screening pesticides at low levels hi different types of environmental samples without sample treatment which may thereby reduce the workload involved in achieving EEC regulations. The duration of the sample screening procedure on the column was less than 20 minutes making the technique highly desirable for a quick estimate in the field. Also, each immunoaffinity column was able to screen 30 samples before a degree of loss in the column's binding capacity was observed. Hence, the reusability of the columns a large number of times would bring down the test cost substantially which makes it even more attractive. Although immunochromatography devices have been used in the past (Kricka, 1993) for different classes of analytes (i.e. hormones, tumour markers and drugs), this is the first time the technique has been applied to environmental monitoring. Further work is required for the technique to be applied for general use, however, the assay does represent a rapid, simple, cost effective and portable method for screening different types of environmental samples for the presence of pesticides. The feasibility studies carried out in this thesis illustrate the potential use of immunoassay techniques to improve the monitoring program of pesticides in the environment. Moreover, the assays could be applied for analysing a wider range of toxic compounds provided suitable antisera are available. Antibodies against a number of fungal metabolites (e.g. aflatoxin and mycotoxin) have been raised and ELISA methods for the determination of these toxic compounds have been successfully developed (Chu, 1991) (Park et al, 1991). Furthermore, occupational exposure to abietic acid, a marker of exposure to solder fume, has been recognised as a cause of occupational asthma, and thus was a major concern to the HSE (Cullen et al, 1992). The existing HPLC methods for determining abietic acid are unsuited for rapid screening because they involve expensive and sophisticated equipment, and extensive sample preparation limiting the number of assays that can be performed. An attempt to produce antiserum towards abietic acid was executed which resulted in a low immune response. Time limitations did not allow further studies on the preparation procedure of the immunogen.

Chapter 7 220 Immunoaffinily columns, where antibodies are bound to solid phases such as silica, can be utilised as a specific clean up step in a multistage sample preparation procedure i.e. the analysis of residues in a complex matrix. In addition, the possible combination between immunoaffinity columns and chromatographic techniques (the on-line combination of immunoaffinity column with HPLC and the off-line combination with GC-MS) offers a desirable feature for pesticide analysis combining the advantages of the chromatography automation and the high selectivity of the antibodies. Preliminary studies of sample preparation techniques using immunoaffinity columns has been carried out at the Robens Institute for a number of pesticides (isoproturon, chlortoluron and 2,4-D) using the antisera described in this thesis. The approach produces high selectivity, simplicity and a rapid extraction procedure whereby final determ ination is made by HPLC. Although the immunoassay techniques provide the advantages of speed, simplicity and selectivity for the detection and measurement of pesticides in the environment, it is only relatively recently that the potential of the techniques has been widely considered. This was probably due to the shortage of suitable antibodies, nevertheless these are now being produced for a range of different compounds. Also the reluctance of analysts to accept immunological assays is gradually being overcome. Recently, more than half of the total clinical immunoassay market was aimed at low molecular-weight analytes, amounting to an estimated $580 million in sales in the United States by 1990 (Vanderlaan et al, 1991). The techniques require minimal or no sample treatment, thus the sample workload can be increased in comparison to the existing conventional methods. Besides, for many environmental monitoring programs and also for operating procedures positive or negative results are adequate, especially in the absence of skilled personnel, and this would be easily achieved by the use of coloured or luminescent visual end-point immunoassays. Indeed simple devices for performing immunological assays in non-laboratory settings are a recent innovation (Osikowicz et al, 1990) and kits for quantitative measurement of various pesticides are now commercially available.

7 .2 FUTURE WORK A possibility of improving the sensitivity of the 2,4-D assay could be by the use of ECLIA method since its Bo reading is usually much higher than that

Chapter 7 221 by ELISA facilitating further reductions in the concentrations of the coating antibody and enzyme label. Hence, a calibration curve with high sensitivity could be achieved. Furthermore, an addition of a primary amino group to the MCPA carbo^gl group could create a hapten soluble enough in aqueous solution to be used with the biftmctional reagent, i.e. glutaraldehyde. This enables the coupling of the protein and hapten through their primary amino groups offering a bridge of five carbons between them which may increase the immunogenicity of the conjugate. Further refinements to the immuno-affinity columns are required, especially regarding the colour signal development, to enable the system to be used commercially. One alternative could be the use of an insoluble chromagen, DAB (3,3'-diaminobenzidine tetrahydrochloride), whereby an insoluble brown- coloured end product would result from electron transfer, by HRPO, from DAB to the peroxide. This brown insoluble marker may produce the clearer visual difference required between the zero and the mixture analyte samples. Moreover, an approach for achieving quantitative results by immunoaffinity columns could possibly be obtained by packing the solid phase into capillary columns. The quantification of results by this approach would be based on the distribution of the enzyme label through the column thereby measuring the height of the colour migration which in turn could be compared to a conversion table to translate the height into concentration units. A class specific immunoassay could be developed as the specificity of an antibody to a small molecule, i.e. pesticides, can be influenced to a large degree by the design of the immunogen used to induce antibody formation. Thus, an attempt to raise antibodies towards the N-N dimethyl urea moiety of the phenylurea herbicides could be highly desirable for identifying the presence of the phenylurea compounds in complex matrices.

7.3 CONCLUDING REMARKS The potential use of immunoassays for pesticide residue analysis should complement, not replace, the traditional methodologies. The type of assay used will depend upon the concentration and matrix of the analyte and the specific analytical requirements. In summary, the work described in this thesis has demonstrated the analytical versatility of antibodies and thus exemplified how

Chapter 7 222 further work will ensure an important place for immunoassays in environmental monitoring of pesticides as weU as other toxic compounds.

Chapter 7 223 REFERENCES

References 224 Absolom, D. R., Affinity chromatography. Separation and Purification Methods, 1981, 10 ,239-286.

Aheme, G. W., and Marks, V., Radioimmunoassay techniques. Techniques in Metabolic Research. 1979, B 218, 1-19.

Aherne, G. W., Enhanced luminescent immunoassays for environmental monitoring. Analytical Proceedings, 1990, 27, 100-101.

Aherne, G. W., Hardcastle, P., Saleem, P., and England, N., Enhanced chemiluminescent immunoassays for environmental monitoring. In: Bioluminescence and Chemiluminescence Current Status. Stanley, P. E., eind Kricka, L. J., Eds., John Wiley and Sons, Chichester, 199 1, 91-98.

Aheme, G. W., Immunoassay and its applications to water analysis. Trends in Analytical Chemistry, 1984,3 ,215-217.

Aherne, G. W., Immunochemical analysis of pesticides in the Aquatic Environment: s-triazine herbicides. In: Chemistry. Agriculture and the Environment. Richardson, M. L., Ed., The Royal Society of Chemistry, Cambridge, 1991, 46-59.

Al-Bassam, M. N., O'Sullivan, M. J., Gnemmi, E., Bridges, J. W., and Marks, V., Double-antibody enzyme immunoassay for nortriptyline. Clinical Chemistry, 1978, 24 ,1590-1594.

Ames, R. G., Cholinesterase activity depression among California agricultural pesticide applicators. American Journal of Industrial Medicine, 1989, 15, 143-150.

Anderson, G. W., Zimmerman, J. E., and Callahan, F. M., The use of esters of N- hydroxysuccinimide in peptide synthesis. Journal oj the American Chemical Society, 1964,86, 1839-1842.

Ashton, F. M., and Crafts, A. S., Mode of action of herbicides. In: Mode o f Action of Herbicides. Ashton, F. M., and Crafts, A. S., Eds., John Wiley and Sons, New York, 1973, 100-109.

Balinova, A., Solid-phase extraction followed by high-performance liquid chromatographic analysis for monitoring herbicides in drinking water. Journal o f Chromatography, 1993, 643 1, , 203-207.

Barnard, G. J. R., Kim, J. B., and WiUiams, J. L., Chemiluminescence immunoassays and immunochemiluminometric assays. In: Alternative Immunoassays. Collins, W. P., Ed., John Wiley & Sons, Chichester, 1985, 123- 152.

Bates, J. A. R., Pesticides and international organisations. In: World Directory of Pesticide Control Organisations. Royal Society of Chemistry, 1989, ISBN 0-85186- 723-5, 1-15.

Blake, C., and Gould, B. J., Use of enzymes in immunoassay techniques. Analyst, 1984, 109, 533-547.

References 225 BMA., Pesticides, chemicals and the environment. The BMA (British Medical Association) guide to Pesticides., Chemicals and Health. Morgan, D. R., 2nd ed., Edward Arnold, London, 1992, 25-54.

Bronstein, I., Voyta, J. C., Thorpe, G. H. G., Kricka, L. J., and Armstrong, G., Chemiluminescent assay of alkaline phosphatase applied in an ultrasensitive enzyme immunoassay of thyrotropin. Clinical Chemistry, 1989, 35 , 1441-1446.

Brown, V. K., Test animals. In: Acute Toxicity in Theory and Practice. Bridges, J. W., and Grasso, P., Eds., John Wiley and Sons, Chichester, 1980, 35-41.

Buck, D., On site theophylline analysis using enzyme immunochromatography. American Clinical Proceedings Review, 1986, 5, 36-41.

Burd, J. F., Wong, R. C., Feeney, J. E., Carrico, R. J., and Boguslaski, R. C., Homogeneous reactant-labeUed fluorescent immunoassay for therapeutic drugs exemplified by getamicin determination in human serum. Clinical Chemistry, 1977,23, 1402-1408.

Bushway, R. J., and Savage, S. A., Determination of methyl 2- benzimidazolecarbamate in finit juices by immunoassay. Food Chemistry, 1990, 35, 51-58.

Bushway, R. J., Perkins, B., Savage, S. A., Lekousi, S. J., and Ferguson, B. S., Determination of atrazine residues in water and soil by enzyme immunoassay. Bulletin of Environmental Contamination and Toxicology, 1988, 40, 647-654.

Campbell, D. H. E., Luescher, E., and Lerman, L. S., Isolation of antibody by means of a cellulose-protein antigen. Proceedings of the National Academy of Sciences of the United States of America, 195 1 , 37, 575-582.

Carter, T. J. N., and Whitehead, T. P., Investigation of a novel sohd-phase chemiluminescent analytical system, incorporating photographic detection, for the measurement of glucose. Talanta, 1982, 29, 529-531.

Centeno, E. R., Johnson, W. J., and Sehon, A. H., Antibodies to two common pesticides, DDT and Malathion. International Archives of Allergy, 1970, 37, 1-13.

Chase, H. A., Affinity separations utilising immobilised monoclonal antibodies-a new tool for the biochemical engineer. Chemical Engineering Science, 1983, 39 7 ,1099-1125.

Chase, H. A., Prediction of the performance of preparative affinity chromatography. Journal of chromatography, 1984, 297, 179-202. Chester, G., Evaluation of agricultural worker exposure to, and absorption of, pesticides. Annals of Occupational Hygiene, 1998,37 ,509-523.

References 226 Chiavari, G., and Bergamini, C., Determination of phenylurea herbicides by high- performance liquid chromatography with electrochemical detection. Journal o f Chromatography, 1985, 346,369-375.

Chu, F. S., Current immunochemical methods for mycotoxin analysis. In: Immunoassays for Trace Chemical Analysis. Vanderlaan, M., Stanker, L. H., Watkins, B. E., and Roberts, D. W., Eds., American Chemical Society, Washington, DC, 199 1, 140-157.

Clower, M., Pesticide residues in food, US food and drug administration's program for immunoassay. In: Immunoassays for Trace Chemical Analysis. Vanderlaan, M., Stanker, L. H., Watkins, B. E., and Roberts, D. W., Eds., American Chemical Society, Washington, DC, 199 1, 49-58.

Coquart, V., and Hennion, M. C., Determination of phenoxyacid herbicides in drinking water at the ppt level by liquid chromatography and on-line selective preconcentration. The Science of the Total Environment, 1993, 132 , 349-360.

Cullen, R. T., Cherrie, B., and Soutar, C. A., Immune responses to colophony, an agent causing occupational asthma. Thorax, 1992,47/12, 1050-1055.

Cunningham, M., Simmonds, A., Harvey, B., and Harris, M., Non-radioactive methods for the detection of nucleic acids. Comments, Amersham Life Science, 1994,21 ,1-8.

De Savlgny, D., and Voiler, A., The communication of ELISA data from laboratory to clinician. Journal of Immunoassay, 1980, 1 , 105-128.

DFG Deutsche Forschungsgemeinschaft., Maximum concentrations at the workplace and biological tolerance values for working materials 1989. Report No. XXV Commission for the Investigation of Health Hazards of Chemical Compounds in the work Area. VCH Verlagsgesellschqft, 1989.

DiCorcia, A., and Marchetti, M., Rapid and sensitive determination of phenylurea herbicides in water in the presence of their anilines by extraction with a carbopack cartridge followed by liquid chromatography. J o u rn a l o f Chromatography, 1991,541,365-373.

Donna, A., Triazine herbicides and ovarian epithelial neoplasm. Scandinavian Joumalof Work, Environment and Health, 1989, 15,47-53.

Edwards, R., Antibodies: polyclonal and monoclonal. In: Immunoassay an Intoduction. Edwards, R., Ed., 1st éd., WiUiam Heinemann Medical Books, 1985, 55-73.

Ekins, R. P., Current concepts and future developments. In: Alternative Immunoassays. Collins, W. P., Ed., John Wiley and Sons, Chichester, 1985, 219- 239.

Ekins, R. P., The estimation of thyroxine in human plasma by an electrophoretic technique. Clinical Chemistry Acta, i960, 5, 453-459.

References 227 Ekins, R., Immunoassay design and optimisation. In: Principles and Practice of Immunoassay. Price, C. P and Newman, D. J., Eds., Macmillan Publishers Ltd, 1991, 96-154.

Engvall, E., and Perhnann, P., Enzyme-linked immunosorbent assay, ELISA: in Quantification of specific antibodies by enzyme-labelled anti-immunoglobulin in antigen-coated tubes. Journal of Immunology. 1972, 109 , 129-135.

Engvall, E., Jonsson, K., and Perhnann, P., Enzyme linked immunosorbent assay. II Quantitation assay of protein antigen, immunoglobulin G, by means of enzyme-labelled antigen and antibody-coated tubes. Biochimica Et Biophysica Acta. 1971, 251, 427-434.

Ercegovich, C. D., Anatysis of pesticide residues: immunological techniques. In: Pesticides Identification at the Residue Level. Advances in Chemistry Series, No. 104. Gould, R. F., Ed., American Chemical Society, Washington D. C, 1971, 162- 177.

Ercegovich, C. D., Vallejo, R. P., Gettig, R. R., Woods, L., Bogus, E. R., and Mumma, R. O., Development of a radioimmunoassay for parathion. Journal o f Agriculture and Food Chemistry, 198 1, 29, 559-563.

Erlanger, B. F., Borek, F., Beiser, S. M., and Lieberman, S., Steroid-protein conjugates: praparation and characterisation of conjugates of bovine serum albumin with testosterone and cortisone. Journal of Biological Chemistry, 1957, 228, 713-727.

Fleeker, J. R., Two Enzyme immunoassays to screen for 2,4- Dichloropheno 2Qracetic acid in water. Journal of the Association of Official Analytical Chemistry. 1987, 70, 874-878.

Friends of Earth., Publications catalogue Autumn, 1993.

Fuchs, S., and Sela, M., Immunoadsorbents, In: Handbook of Experimental Immunology. Weir, D. M., Ed., Blackwell Scientific Publications Oxford, England, 3rd ed., 1978, 1, 10.1-10.6.

Geerdink, R. B., Graumans, A. M., and Viveen, J., Determination of pheno^gracid herbicides in water. Journal of Chromatography. 199 1, 547 ,478-483.

Godfrey, M. A. J., Assessment of silica-based solid phases in bioseparations. PhD Thesis, University of Surrey (UK), 1993.

Godfrey, M. A. J., Kwasowski, P., Clift, R., and Marks V., A sensitive enzyme- linked immunosorbent assay (ELISA) for the detection of staphylococcal protein A (SpA) present as a tracer contaminant of murine immunoglobulins purified on immobilised protein A. Journal of Immuaolopical Methods, 1992, 149,21-27.

Godfrey, M. A. J., Kwasowski, P., Chft, R., and Marks V., Assessment of the suitability of commercially available SpA affinity solid phases for the purification of murine monoclonal antibodies at process scale. Journal of Immunological Methods, 1993, 160, 97-105.

References 228 Green, D. J., Optimisation of coating parameters in microplate immunoassays. An industry sponsored workshop at the national AACC meeting in Washington, D.C., Labsystems, 199 1

Grover, R., Cessna, A. J., and Kerr, L. A., Procedure for the determination of 2,4- D and dicamba in inhalation. Dermal, hand-wash, and urine samples from spray applicators. Journal of Environmental Science and Health, 1985, B 20 , 113-128.

Hall, J. C., Deschamps, R. A., and Krieg, K. K., Immunoassays for the detection of 2,4-D and picloram in river water and urine. Journal of Agriculture and Food Chemistry, 1989, 37, 981-984.

Hall, J. C., Wilson, L. K., and Chapman, R. A., An immunoassay for metolachlor detection in river water and soil. Journal of Environmental Science and Health, 1992, 27 ,523-544.

Hammock, B. D., Gee, S. J., Cheung, P. Y. K., Miyamoto, T., Goodrow, M. J., Van Emon, J., and Seiber, J. N., Utihty of immunoassays in pesticide trace analysis. In: Pesticide Science and Biotechnology. Greenhalgh, R., Roberts, T. R., Eds., Blackwell Scientific. Oxford, 1987, 309-316.

Handa, S. K., Spectrophotometric method for the microdetermination of bendiocarb standard residues in water. Journal of the Association of Official Analytical Chemists, 1988, 71, 51-52.

Hardcastle, A., Aheme, W., Thorpe, G.H. G., Photographic screening procedure for triazine herbicides in water using an enhanced luminescent immunoassay. In: proceeding of the lAWPRC conference. Wheeler, D., Richardson, M. L., and Bridges, J., Eds., Watershed 89, 1989, 2, 293-296.

Hardcastle, A., Saleem, P., and Aheme, G. W., Immunoassays of atrazine and paraquat in biological fiuids using enhanced luminescence. In: proceedings of ACB National Meeting. Martin, S., and Halloran, S. P., Eds., Association of Chemical Biochemists, Brighton, 1990, B 22, 65-66.

Hassall, K. A., Herbicides apphed to foliage and that are mainly soil-acting against seedling weeds. In: The Biochemistry and Uses of Pesticides. Hassall, K. A., Ed., 2nd éd., Macmillan Press Ltd, 1990, 383-472.

Hoar, S. K., Blair, A., Holmes, F. F., Boysen, C. D., Robel, R. J., Hoover, R., Fraumeni, J. F., Agricultural herbicide use and risk of lymphoma and soft-tissue sarcoma. Joumal-American Medical Association. 1986,256, 1141-1147.

Hoar-Zahm, S. K., A case referent study of soft tissue sarcoma and Hodgkin's disease. Scandinavian Journal of Work, Environment and Health, 1988, 14, 224- 230.

Hum, B. A. L., Landon, J., Antisera for radioimmunoassay. In: Radioimmunoassay Methods. Kirkham, K. E., and Hunter. W. M., Eds., Churchill Livingstone, Edinburgh and London, 1971, 121-142.

References 229 Ireland, D., and Samuel, D„ Enhanced chemilumlnescence ELISA for the detection of antibody to hepatitis B virus surface antigen. Journal of Bioluminescence and Chemiluminescence, 1989,4, 159-163.

Jackson, T. M., and Ekins, R. P., Theoretical limitations on immunoassay sensitivity. Current practice and potential advantages of fluorescent Eu3+chelates as non-radio-isotopic tracers. Journal of Immunological Methods, 1986,86, 13-20.

Jones, J. G., Pesticides: the environmental legacy. The Royal Society of Chemistry, A Synposium of the Analytical Division on Fungicides, Herbicides and Insecticides, 1992, 15-28.

Jung, R., Gee, S. J., Harrison, R. O., Goodrow, M. H., Karu, A. E., Braun, A. L., and Hammock, B. D., Use of immunochemical techniques for the analysis of pesticides. Pesticide Science, 1989, 26, 303-317.

Junk, G. A., and Richard, J. J,, Organics in water: solid phase extraction on a small scale. Analytical Chemistry, 1988, 60, 451-454. Kaufman, B. M., and Clower, M., Immunoassay of pesticides. Journal of the Association of official Analytical Chemists, 1991,74: ,239-247.

Kemeny, D. M., Challacombe, S. J., Microtitre plates and other solid phase supports. In: ELISA and Other Solid Phase Immunoassays. Kemeny, D. M., and Challacombe, S. J., Eds., John Wiley and Sons, Chichester, 1988, 31-55.

Kemeny, D. M., Chantier, S., An Introduction to ELISA. In: ELISA and Other Solid Phase Immunoassays. Kemeny, D. M., and Challacombe, S. J., Eds., John Wiley and Sons, Chichester, 1988, 1-31.

Khanna, P., Homogenous enzyme immunoassay. In: Principles and Practice of immunoassay. Price, C. P and Newman, D. J., Eds., Macmillan Publishers Ltd, 1991, 326-365

Kinders, R. J., and Hass, G. M., Interference in immunoassays by human anti mouse antibodies. European Journal of Cancer, 1990,26 , 647-648.

Kitigawa, T., Shimozono, T., Aikawa,T., Yoshida, T., and Nishimura, H., Preparation and characterisation of hetro-bifunctional cross-linking reagents for protein modifications. Chemical and Pharmceutical Bulletin, 1981,29, 1130-1135.

Knopp, D., Glass, S., Biological monitoring of 2,4-dichlorophenoxyacetic acid- exposed workers in agriculture and forestry. International Archives of Occupational and Environmental Health, 199 1 , 63, 329-333.

Knopp, D., Nuhn, P., and Dobberkau, H. J., Radioimmunoassay for 2,4- dichloropheno]gacetic acid. Archives o f Toxicology, 1985, 58, 27-32.

Kricka, L. J., and Thorpe, G. H. G., Chemiluminescent and bioluminescent methods in analytical chemistry. Analyst, 1983, 108, 1274-1296.

Kricka, L. J., and Thorpe, G. H. G., Photographic detection of chemiluininescent and bioluminescent reactions. Methods inEnzymology , 1986, 133, 404-420.

References 230 Kricka, L. J., Extra-laboratoiy immunoassays: technology for rapid analyses of tests ranging from drugs to infectious diseases. Clinical Biochemistry, 1993, 26, 11-13.

Kurzer, F., and Douraghi-Zadah, K., Advances in the chemistry of carbodiimides. Chemical Reviews, 1967,67 , 107-152.

Landsteiner, K., Artificial conjugated antigens: serological reactions with simple chemical compounds. In: The Specificity of Serological Reactions. Landsteiner, K., Ed., Dover Publications INC, NewYork, 1962, 156-209.

Lawrence, J. F., and Turton, D., High performance hquid chromatographic data for 166 pesticides. Journal of Chromatography, 1978, 159, 207-226.

Lee, H., and Chan, A. S. Y., Determination of trifiuranin, diallate, triallate, atrazine, barban and benzoyl propethyl in natural waters at parts per trillion levels. Journal of the Association of official Analytical Chemists, 1983, 66, 651-658.

Lees, A., and McVeigh, K., An investigation of pesticide pollution in drinking water in England and Wales. Friends of the Earth, Report on breaches of the EC Drinking Water Directive, London, 1988, ISBN 0 905966 65 1.

Li'egeois, E., Dehon, Y., de Brabant, B., Copin, A., and Portetelle, D., Isoproturon as a model to illustrate the ELISA method for the detection and quatitation of pesticides in environment. In: Proceedings of an International Symposium Environmental Biotechnology, Ostende, 199 1 , 69-72.

Li'egeois, E., Dehon, Y., de Brabant, B., Perry, P., Portetelle, D., and Copin, A , ELISA test, a new method to detect and quantify isoproturon in sod. The Science of the Total Environment, 1992, 123 , 17-28.

Litman, D. J., Handon, T. M., and UUman, E. F., Enzyme channelling immunoassay: a new homogeneous enzyme immunoassay technique. Analytical Biochemistry, 1980, 106,223-229.

Little, J. R., and Donahue, H., Spectral properties of proteins and small molecules of immunological interest. In: Methods in Immunology and Immunochemistry. Williams, C. A and Chase, M. A., Eds., Academic Press, New York, 1968, 2, 343-363.

Lockhart, J. A. R., Holmes, J. C., and Mackay, D. B., The evolution of weed control in British agriculture. In: Weed Control Handbook: principles. Roberts, H. A., Ed., 7th ed.. Black Well Scientific Publication, Oxford, 1982,37-63.

Lopez-Avlla, V., Hirata, P., Kraska, S., Flanagan, M., and Talor, J. H., Determination of atrazine, lindane, pentachlorophenol and diazinon in water and soil by isotope dilution GC-MS. Analytical Chemistry, 1985, 57, 2797-2801.

Lynge, E., Cancer in phenoxy herbicide manufacturing workers in Danmark, 1947-1987 an update. Cancer Causes Control. 1993, 4 , 261-272.

References 231 Markovitz, A., and Crosby, W. H., Chemical carcinogenesis: a soil fumigant, 1,3- dichloropropene, as possible cause of hematologic malignancies. Archives of International Medicine, 1984, 144,1409-1411.

Mathias, C. G. T., and Morrison, j. H., Occupational skin diseases. United S ta tes. Archives o f Dermatology (Chicago), 1988, 124, 1519-1524.

Mcmahon, B., A second follow-up mortility in a cohort of pesticide apphcators. Journal of Occupational Medicine, 1988,30,429-432.

Merck Index. An encyclopedia of chemicals, drugs, and biologicals. Budavari, S., Ed., 11th ed. Merck and CO., INC, USA, 1989.

Middelem, C., and Van, H., Assay procedures for pesticide residues. In: Pesticides in the Environment, Stevens, R. W., Ed., Marcel Dekker INC, NewYork, 1971, 1 , 309-315.

Middleton, B. A., Morgan, L. M., Aheme, G. W., and Marks, V., The effect of storage temperature and physical state on the performance of antisera in radioimmunoassay after long-term storage. Annals of Clinical Biochemistry, 1988, 25, 89-95.

Misra, H. P., and Squatrito, P. M., The role of superoxide anion in peroxidase- catalyzed chemilumin escence of luminol. Archives of Biochemistry and Biophysics, 1982, 215, 59-65.

Misra, U. K., A study of nerve conduction velocity late responses and neuromuscular synapse function ia organo phosphate workers in india. Archives of Toxicology, 1988, 61,496-500.

Moreland, D. E., Mode of action of herbicide. In: Pesticide Chemistry in the 20th Century. Plimmer, J. R., Ed., American Chemical Society, Washington, D. C., 1977, 56-75.

Morgan, E . p.. Morbidity and mortality in workers occupationally exposed to pesticides. Archives of Environmental Contamination and Toxicology, 1980, 9, 349- 382.

Morris, B. A., Principles of immunoassay. In: Immunoassays in Food Analysis. Morris, B. A., and Clifford, M. N., Eds., Elsevier Apphed Science Pubhshers, London, 1985,21-52.

Nars, P. W., and Hunter, W. M., A method for labelling oestradiol-176 with radioiodine for radioimmunoassays. In: Proceedings of the Society for Endocrinology. Joumalof Endocrinology, 1973, 57, xlvii-xlviii.

Neicheva, A., Kovacheva, E., and Marudov, G., Determination of organophosphorus pesticides in apples and water by gas-hquid chromatography with electron-capture detection. Journal of Chromatography, 1988, 437, 249-253.

Niewola, Z., Benner, J. P., and Swaine, H., Determination of paraquat residues in soil by an enzyme linked-immunosorbent assay. Analyst, 1986, 3, 399-403.

References 232 Osikowicz, G-, Beggs, M., Brookhart, P., Caplan, D., Ching, S., Eck, P., Gordon, J., Richerson, R., Sampedro, S., Stimpson, D., and Walsworth, F., One-step chromatographic immunoassay for qualitative determination of choriogonadotropin in urine. Clinical Chemistry, 1990, 36, 1586-1586.

Osselton, M. D., and Snelling, R. D., Chromatographic identification of pesticides. Journal of Chromatography,1986, 368, 265-271.

Park, D. L., Njapau, H., Rua, S. M., and Jorgensen, K. V., Field evaluation of immunoassays for aflatoxin contamination in agricultural commodities. In: Immunoassays for Trace Chemical Analysis. Vanderlaan, M., Stanker, L. H., Watkins, B. E., and Roberts, D. W., Eds., American Chemical Society, Washington, DC, 1991,162-169.

Phillips, T. M., and Frantz, S. C., Isolation of specific tymphocyte receptors by high-performance immunoaffinity chromatography. Journal of chromatography, 1988, 444, 13-20.

Phillips, T. M., High-performance immunoaffinity chromatography. In: Advances in Chromatography Biotechnological Applications and Methods. Giddings, J. C., Grushka, E., and Brown, P. R., Eds., Marcel Dekker INC, New York, 1989, 29, 134-170.

Phillips, T- M., Queen, W. D., More, N. S., and Thompson, A. M., Protein A- coated glass beads: a universal support media for high performance immunoaffinity chromatography. Journal of chromatography, 1985,327,213-219.

PIERCE. Enzymatic labelling and detection. Catalogue and Handbook. 1994,209- 230.

Pommeiy, J., Mathieu, D., and Lhermitte, M., High performance hquid chromatographic determination of atrazine in human plasma. Journal o f Chromatography, 1990,526,569-574.

Prapamontol, T., The development, validation and application of chromatographic methods to stucfy organochlorine pesticides in milk and water. PhD Thesis, University of Surrey (UK), 1991.

Procianoy, R. S., and Schvartsman, S., Blood pesticide contamination in mothers and their new bom infants, relation to prematurity. Acta Paediatrica Scandinavica, 1981, 70, 925-928.

Proudfoot, A. T., Poisoning treatment centre admissions foUowing acute incidents involving pesticides. Human Toxicology, 1988, 7, 255-258.

Pudek, M. R., Seccombe, D. W., Jackson, B. E., and Humphries, K., Effect of assay conditions on cross reactivity of digoxin-like immunoreactive substance(s) with radioimmunoassay kits. Clinical Chemistry. 1985,31 , 1806-1810.

Purkayastha, R., and Cochrane, W. P., Comparison of electron capture and electrolyte conductivity detectors for the determination of triazine herbicides. Journal of Agriculture and Food Chemistry, 1973, 21, 93-98.

References 233 Reichlln, M., Schnure, J. J., and Vance, V. K., Induction of antibodies to porcine ACTH in rabbits with nonsteroidogenic polymers of BSA and ACTH. Proceedings oj the Society for Experimental Biology and Medicine, 1968, 128,347-350.

Riihimaki, V., Mortality of 2,4-dichlorophenoxyacetic acid and 2,4,5- trichlorophenoxyacetic acid herbicide applicators in Finland. Scandinavian Journal of Work Environment and Health. 1982, 8, 37-42.

Roberts, T. R., Pesticides in water-human health, agricultural and environmental aspects, hi: Chemistry , Agriculture and the Environment. Richardson, M. L., Ed., The Royal Society of Chemistry, England, 199 1, 429-443.

Robinson, J. D., Aheme, G. W., Teale, J. D., Morris, B. A.,and Marks, V., Problems encountered in the use of radioiodination tags in radioimmunoassays for dmgs. In: Radioimmunoassay in Clinical Biochemistry, Pasternak, C. A., Ed., Heyden, London, 1975, 113-123.

Robinson, J. D., Morris, B. A., Piall, E. M., Aheme, G. W., and Marks, V., The effect of bridge length and degree of derivatisation on the production of antibodies to morphine, use of rats in the screening of dmg-protein conjugates for immunoreactivity. In: Radioimmunoassay in Clinical Biochemistry, Pasternak, C. A., Ed., H ^den, London, 1975, 101-112.

Roitt, I. V., The recognition of antigen 1- primary interaction. Essential Immunology. Roitt, T. V., Ed., 7th ed., Blackwell Scientific Publications, Oxford, 1991, 65-84.

Sagar, G. R., Caseley, J. C., Kirkwood, R. C., and Parker, C., Herbicides in plants. In: Weed Control Handbook: Principles. Roberts, H. A., Ed., Blackwell Scientific Publications, Oxford, 1982, 64-86.

Sarkar, S. N., and Gupta, P. K., Neurotoxicity of isoproturon, a substituted phenylurea herbicide, in mice. Journal of the Indian Experimental Biology, 1993, 31 ,977-981.

Savage, E. P., Chronic neurological sequelae of acute organophosphate pesticide poisoning. Archives of Environmental Health, 1988, 43, 39-45.

Saxena, M. C., Organochlorine pesticides in specimens fi'om women undergoing spontaneous abortion, premature or full term dehvery. Journal of Analytical Toxicology, 1981, 5, 6-9.

Schuman, S. H., and Dodson, R. L., An outbreak of contact dermltitis in farm workers. Journal of American Academy of dermatology, 1985, 13, 220-230.

Schwalbe, M., Dorn, E., and Beyerman, K., Enzyme immunoassay and fluoroimmunoassay for the herbicide diclofop-methyl. Journal of Agriculture and Food Chemistry, 1984, 32, 734-741.

Seiler, J. P., Herbicidal phenylalkylureas as possible mutagens, II chemical basis of mutagenic activity. Pesticide Biochemistry and Physiology, 1979, 12, 183-190.

References 234 Sherma, J., Determination of trizine and. chlorophenoxyacid herbicides in natural water samples by phase extractio and quantitative TLC. Journal of Liquid Chromatography, 1986, 9, 3433-3438.

Sherma, J., Pesticides. Analytical Chemistry, 1989, 61, 153 R-165 R.

Sherma, J., Pesticides. Analytical Chemistry, 1991, 63, 118 R-130 R.

Sherma, J., Pesticides. Analytical Chemistry, 1993, 65, 40 R-54 R.

SidweH, J. A., and Ruzicka, H. A., The determination of substituted phenylurea herbicides and their impurities in technical and formulated products by use of liquid chromatography. Anofyst, 1976, 101, 111-121.

Silman, I. H., and Katchalski, E., Water insoluble derivatives of enzymes, antigens and antibodies. Annual Review of Biochemistry, 1966, 35, 873-901.

Singh, R., Simultaneous determination of fenamiphos, its sulphoxide and sulphone in water by high performance hquid chromatography. Analyst, 1989, 114,425-427.

Smith, R. N., Radioimmunoassay in forensic science. In: Drug Determination In Therapeutic And Forensic Contexts. Reid, E., and Wilson, I. D., Eds., Plenum, 1985, 14, 261-268.

Steward, M., Antigen recognition. In: Immunology. Roitt, I., Brostoff, J., and Male, D., Eds., 2nd ed., Gower Medical Pubhshing, London, 1989, 7.1-7.10.

Thomson, J., and Abbott, D. C., Pesticide residues, history, alternatives and analysis. The Royal Institute of Chemistry, Lecture Series, 1966, 3, 1-14.

Thorpe, G. H. G., Bronstein, I., Kricka, L. J., Edwards, B., and Voyta, J. C., Chemiluminescent enzyme immunoassay of @ fetoprotein based on adamantyl dioxetane phenylphosphate substrate. Clinical Chemistry, 1989, 35,2319-2321.

Thorpe, G. H. G., Kricka, L. J., Moseley, S. B., and Whitehead, T. P., Phenols as enhancers of the horseradish peroxidase-luminol-hydrogen peroxide reaction: Apphcation in luminescence-monitored enzyme immunoassays. Clinical Chemistry, 1985, 31 , 1335-1341.

Thorpe, G. H. G., Moseley, S. B., Kricka, L. J., Stott, R. A., and Whitehead, T. P., Enhanced luminescence determination of horseradish peroxidase conjugates, apphcation of benzothiazole derivatives as enhancers in luminescence assays on microtitre plates. Analytical Chemica Acta, 1985, 170, 101-107.

Topp, E., Factors Affecting the Uptake of ^^C-labeUed Organic Chemicals by Plant and by Soil. Ecotoxicology and Environmental Safety. 1986, 11, 219-228.

Traore, S., and Aaron, J. J., Analysis of trifluralin and other dinltroaniline herbicide residues by zero-order and derivative ultraviolet spectrophotometry. Analyst, 1989, 114, 609-613.

References 235 Urbanek, R., Kemeny, D. M., and Samuel, D., Use of enzyme-linked Immunosorbent assay for measurement of allergen specific antibodies. Journal o f Immunological Methods, 1985, 79, 123-131.

Valkirs, G. E., and Barton, R., Immunoconcentration^^ -a new format for solid phase im m unoassays. Clinical Chemistry, 1985,31,1427-1431.

Vallejo, R. P., Bogus, E. R., and Mumma, R. O., Effects of hapten structure and bridging groups on antisera specificity in parathion immunoassay design. Journal of Agricultural and Food Chemistry, 1982,30,572-580.

Van Emon, J., Hammock, B., and Seiber, J. N., Enzyme-linked immunosorbent assay for paraquat and its application to exposure analysis. Analytical Chemistry, 1986,58, 1868-1873.

Vanderlaan, M. Stanker, L., and Watkins, BV, Immunochemical techniques in trace residue analysis. In: Immunoassays for Trace Chemical Analysis. Vanderlaan, M., Stanker, L. H., Watkins, B. E., and Roberts, D. W., Eds., American Chemical Society, Washington, DC, 1991, 2-13.

Voiler, A., Bartlett, A., and Bidwell, D. E., Enzyme immunoassays with special refemce to ELISA techniques. Journal of Clinical Pathology. 1978, 31, 507-520.

Walker, A., Briggs, G. G., Greaves, M. P., Hance, R. J., and Thompson, A. R., Herbicides in soil. In: Weed Control Handbook: Principles. Roberts, H. A., Ed., Blackwell Scientific Publications, Oxford, 1982, 87-105.

Walker, W. H. C., An approach to immunoassay. Clinical Chemistry, 1977, 23, 384-402.

Walters, R. R., Diol-bonded silica. In: Affinity Chromatography a Practical Approach. Dean, P. D. G., Johnson, W. S., Middle, F. A., Eds., IRL Press Ltd, England, 1985,25-26.

Walters, S. M., Westerby, B. C., and Gilvydis, D. M-, Determination of phenylurea pesticides h y high-performance liquid chromatography with UV and photoconductivity detectors in series. Joumalof Chromatography, 1984, 317, 533- 544.

Wang, H. H., and MacMahon, B., Mortality of pesticides applicators. Journal o f Occupational Medicine, 1979, 21, 741-744.

Wang, S., Health survey among farmers exposed to deltamethrin in the cotton fields. Ecotoxicology and Environmental Scifety, 1988, 15, 1-6.

Weeks, I., and Woodhead, J. S., Chemilumlnescence immunoassay. In: ELISA and Other Solid Phase Immunoassays. Kemeny, D. M., and Challacombe, S. J., Eds., John Wiley & Sons Ltd, 1988, 265-277.

Weeks, I., Beheshti, I., McCapra, F., Campbell, A. K., and Woodhead, J. S., Acridinium esters as high-specific-activity labels in immunoassay. Clinical Chemistry, 1983,29 , 1474-1479.

References 236 Weetall, H.H., Affinity chromatography. Separation and Purification Methods, 1973,2 ,199-229.

Whitehead, T. P., Thorpe, G. H. G., Carter, T. J. N., Groucutt, C., and Kricka, L. J., Enhanced luminescence procedure for sensitive determination of peroxidase- laballed conjugates in immunoassay. Nature, 1983, 304, 158-159.

Whorton, D., Infertility in male pesticide workers. Lancet, 1977, 2, 1259-1261. Wilchek, M., Miron, T., and Kohn, J., Affinity chromatography. M ethods in Enzymology, 1984, 104, 211-248.

Wold, F., Bifuctional reagents. Methods inEnzymology. 1967, 11, 617-640.

Woollen, B. H., Biological monitoring for pesticide absorption. A nnals o f Occupational Hygiene, 1993,37 , 525-540.

Yalow, R. S., and Berson, S. A., Assay of plasma insuhn in human subjects by immunological meyhods. Nature, 1959, 184, 1648-1650.

Zuk, R. F., Ginsberg, V. K., Houts, T., and Rabbie, J., Enzyme immunochromatography-quantitative immunoassay requiring no instrumentation. Clinical Chemistry, 1985,31 ,1144-1150.

Stevenson, D., The low concentrations of hazardous compounds present in the environment or the human body mean that some form of sample concentration and extraction is almost inevitable. Laboratory Equipment Digest, 1991, September, 17-17. ______

References 237