UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Development of an enzyme linked immunosorbent assay for dehydroabietic acid in pulp mill effluents Li, Kai 1996

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1996-147797.pdf [ 10.26MB ]
Metadata
JSON: 831-1.0087843.json
JSON-LD: 831-1.0087843-ld.json
RDF/XML (Pretty): 831-1.0087843-rdf.xml
RDF/JSON: 831-1.0087843-rdf.json
Turtle: 831-1.0087843-turtle.txt
N-Triples: 831-1.0087843-rdf-ntriples.txt
Original Record: 831-1.0087843-source.json
Full Text
831-1.0087843-fulltext.txt
Citation
831-1.0087843.ris

Full Text

D E V E L O P M E N T O F A N E N Z Y M E L I N K E D I M M U N O S O R B E N T A S S A Y F O R D E H Y D R O A B I E T I C A C I D I N P U L P M I L L E F F L U E N T S by KAI LI B. Sc. Jiangsu Normal University, 1982 M . Sc. Suzhou University, 1985 M . Sc. University of British Columbia, 1992 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Wood Science) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH C O L U M B I A June 1996 © Kai L i , 1996 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada Da.e r?fj. 7 DE-6 (2/88) I A b s t r a c t A n enzyme-linked immunosorbent assay (ELISA) was developed for the quantification of resin acids in softwood pulp mill effluents. Various antigens were synthesized and polyclonal antibodies were successfully raised by immunizing rabbits with D l i A M - S U C - K L H or D H A -K L H , which represent antigens with and without a spacer arm between the hapten and the carrier protein. Different ELISA formats were evaluated for their assay sensitivities and the direct ELISA was applied to both chemi-thermomechanical pulp (CTMP) mill effluents and bleached kraft mill effluents (BKME), as well as fish bile for the quantification of resin acids. A GC protocol involving extraction, derivatization and GC quantification was developed as a standard method and used to compare the results with ELISA. The ELISA formats played an important role in the assay sensitivities and the 50% inhibition concentration ( I50) for D H A ranged from 12.3 to 113.2 ug/L. Indirect ELISA with the biotin-streptavidin system showed the highest assay sensitivity. Direct ELISA had an I50 of 49.7 ug/L, with a detection limit of 4.5 ug/L. It was considered the method of choice due to its simple assay protocol and the reasonably good sensitivity that was obtained. However, when we compared two polyclonal antibodies derived from the antigen with and without a spacer arm, no significant differences were observed in either assay sensitivity or cross-reactivity. Both polyclonal antibodies showed high cross-reactivities with the abietic type resin acids. Thus the polyclonal antibodies appeared to be group specific and could be used to quantify total abietic type resin acids. Various assay parameters, such as the concentrations of coating antigen and antibodies used, the incubation temperature and time, pH, ionic strength and metallic ions, were each evaluated for their influence on the assay sensitivity. The direct ELISA method was used to quantify D H A equivalents in both B K M E and CTMP effluents. For CTMP effluents, which usually contain high levels of resin acids together with a high chemical oxygen demand (COD) and a high biochemical oxygen demand (BOD), a large dilution factor was required to bring the i i resin acid concentration into the quantification window. The D H A equivalents determined by ELISA compared favourably with total abietic type resin acids determined by GC, and also correlated well with total resin acids. While dilution could be used to reduce the background to a negligible level for CTMP samples containing high resin acid concentrations, this method was not suitable for samples containing low resin acid concentrations such as B K M E . A significant background effect was encountered and the background interference could not be eliminated by dilution, as this brought the resin acid concentration to below the detection limit of the ELISA. Various techniques, such as filtration, extraction and fractionation were used to try to eliminate this background effect. However, all these methods were unsuccessful. By using a bio-treated B K M E as the diluent and creating a standard curve using the same bio-treated B K M E we were able to eliminate the background interference. Although the assay sensitivity decreased, the D H A equivalents determined in the B K M E by the ELISA method compared favourably with the amount of total abietic type and the total resin acids as determined by GC. The direct ELISA was also used for the quantification of resin acids in fish bile. Good recoveries were obtained with a D H A spiked control bile sample. Direct ELISA of a limited number of bile samples from fish exposed to pulp mill effluents indicated that this immunochemical method did not require sample hydrolysis and much less sample was needed when compared to the GC method (1-5 pi for ELISA vs 100-500 pi for GC). This work represents the first attempt at developing an antibody-based immunochemical method for the detection and quantification of resin acids in softwood pulp mill effluents. Compared with conventional instrumental methods such as GC, H P L C and GC-MS, the ELISA method was able to overcome several drawbacks of these alternative methods such as, extensive sample pre-treatment, expensive instrumentation and low sample output. Although not intended to replace the conventional analysis, ELISA will be particularly useful in the screening of large number of samples which can produce results within a short period of time. i i i Table of Contents Abstract i i List of Tables •. vii i List of Figures.... x List of Abbreviations xi i Acknowledgement xv Note ...xvi C H A P T E R 1 INTRODUCTION 1 1.1. Resin Acids 3 1.1.1.. Occurrence and toxicity of resin acids 3 1.1.2. Current methods of resin acid analysis and their drawbacks 10 1.1.2.1. Isolation 10 1.1.2.2. Separation and derivatization 11 1.1.2.3. Identification and quantification 12 1.1.2.4. Drawbacks of conventional analytical methods 13 1.2. Immunoassay 15 1.2.1. Introduction to immunoassay 15 1.2.1.1. Antibody and its structure 15 1.2.1.2. Immunoassays and their advantages 19 1.2.2. Preparation of antigens 24 1.2.2.1. Choice of haptens 24 1.2.2.2. Choice of proteins 25 1.2.2.3. Choice of spacer arms 26 1.2.2.4. General methods for the preparation of antigens 28 1.2.3. Production of polyclonal antibodies 31 1.2.4. The enzyme-linked immunosorbent assay (ELISA) ....34 1.2.4.1. Assay protocols 35 1.2.4.2. Cross-reactivity 38 1.2.4.3. Matrix effects and assay validation...., 39 1.3. Proposed Work and Objectives 41 C H A P T E R 2 D E V E L O P M E N T OF A GC PROTOCOL FOR RESIN ACID A N A L Y S I S 44 2.1. Introduction 44 2.2. Materials and Methods 46 2.2.1. Extractions of resin acids from pulp mill effluents 47 2.2.2. Sample clean-up using Solid-Phase Extraction (SPE) 47 2.2.3. Derivatization of resin acids 47 2.2.4. GC and GC-MS Analysis of resin acids 48 2.3. Results and Discussion 48 2.3.1. Comparison of extraction solvents 48 2.3.2. Solid-phase extraction prior to GC and GC-MS analyses..... 49 2.3.3. Effects of pH on extraction efficiency 53 iv 2.4. Conclusions 60 C H A P T E R 3 SYNTHESIS OF ANTIGENS A N D D E V E L O P M E N T OF A N ELISA FOR D H A 61 3.1. Introduction ". .61 3.2. Materials and Methods 62 3.2.1. Synthesis of antigens 63 3.2.1.1. Preparation of D H A M - S U C 63 • 3.2.1.2. Preparation of D H A M - S U C - K L H and D H A M - S U C - B S A . . . . 6 4 3.2.1.3. Preparation of D H A M - S U C - H R P 64 3.2.1.4. Preparation of D H A - K L H and D H A - B S A 65 3.2.1.5. Succinylation of B S A and K L H ..65 3.2.1.6. Preparation of POD-SUC-BSA and P O D - S U C - K L H . . . . 66 3.2.1.7. Preparation of DHA-OXI-SUC-BSA and D H A - O X I - S U C - K L H 66 3.2.1.8. Preparation of D H A - K L H - P I M and D H A - B S A - P L M 68 3.2.2. Immunization protocols 68 3.2.3. Detection of polyclonal antibodies against D H A 69 3.2.4. Indirect ELISA of D H A using the biotin-streptavidin system 70 3.3. Results and Discussion 70 3.3.1. Synthesis of antigens 70 3.3.1.1. Antigens with spacer arm: D H A M - S U C - K L H and D H A M - S U C - B S A 71 3.3.1.2. Antigens without spacer arm: D H A - K L H and D H A - B S A 74 3.3.1.3. Enzyme conjugate: D H A M - S U C - H R P .....75 3.3.1.4. Other antigens 75 3.3.2. Preliminary immunization results 79 3.3.2.1. Production of polyclonal antibodies using D H A M - S U C - K L H as antigen 79 3.3.2.2. Evaluation of the polyclonal antibodies 79 3.3.2.3. Cross-reactivity 82 3.3.2.4. Recovery of D H A by ELISA in Spiked Water and Effluent.. 84 3.3.2.5. Production and evaluation of other polyclonal antibodies -.85 3.4. Conclusions 87 C H A P T E R 4 E V A L U A T I O N OF TWO P O L Y C L O N A L ANTIBODIES AGAINST D H A A N D OPTIMIZATION OF ELISA 88 4.1. Introduction 88 4.2. Materials and Methods 89 4.2.1 ELISA protocols using microtitre plates 90 4.2.1.1. Conventional indirect ELISA 90 4.2.1.2. Indirect ELISA with biotin-streptavidin system 91 4.2.1.3. Direct ELISA :...91 4.2.2. Magnetic bead-based ELISA 92 4.2.2.1. Purification of polyclonal antibodies using the Protein A membrane 92 4.2.2.2. Coupling of antibodies to tosylactivated magnetic beads 93 v 4.2.2.3. Magnetic beads ELISA 93 4.3. Results and Discussion 94 4.3.1. Effects of the assay formats on the ELISA sensitivity 94 4.3.1.1. Conventional indirect ELISA 95 4.3.1.2. Indirect ELISA using the biotin-streptavidin system 96 4.3.1.3. Direct ELISA 96 4.3.1.4. ELISA using magnetic beads as support 100 4.3.2. Effect of a spacer arm on ELISA 103 4.3.2.1. Effects of a spacer arm on antibody production 103 4.3.2.2. Effects of spacer arm on the assay sensitivity 104 4.3.2.3. Effects of spacer arm on the specificity of the antibodies 107 4.3.2.4. Cross reactivity of antibodies to compounds other than resin acids 108 4.3.3. Evaluation of assay conditions in the direct ELISA 110 4.3.3.1. Coating temperature I l l 4.3.3.2. Incubation time for competition step 113 4.3.3.3. pH of the sample 114 4.3.3.4. Ionic strength of the sample 115 4.3.3.5. Effect of metallic ions 116 4.3.3.6. The effect of organic solvents on the ELISA 117 4.3.3.7. Effect of substrate on ELISA 118 4.3.4. Method evaluation 119 4.3.4.1. Within-assay and between-assay variation .119 4.3.4.2. Recovery of D H A from spiked matrices using direct ELISA. 122 4.3.4.3. False positive and false negative 124 4.3.4.4. Quantification of D H A in the presence of A B A or P IM 125 4.4. Conclusions 128 C H A P T E R 5 QUANTIFICATION OF RESIN ACIDS IN PULP M I L L EFFLUENTS USING ELISA. . . . 129 5.1 Intro duction 129 5.2 Material and methods 131 5.2.1. Competitive direct ELISA 132 5.2.2. GC analysis 132 5.3 Results and Discussion 133 5.3.1. ELISA for effluent samples from a CTMP mill 134 5.3.2. ELISA for B K M E samples 138 5.2.3. Effect of possible interfering material on the ELISA of B K M E 144 5.4. Conclusion 147 C H A P T E R 6 ELISA OF RESIN ACIDS IN FISH BILE 149 6.1. Introduction 149 6.2. Material and Methods 150 6.2.1. Direct ELISA 151 6.2.2. Hydrolysis of fish bile and GC analysis 152 6.3. Results and Discussion 152 6.3.1. Spiked bile 152 vi 6.3.2. Bile from fish exposed to effluent 154 6.3.3. Comparison between ELISA and GC 154 6.3.4. Effect of hydrolysis on ELISA 157 6.4. Conclusions 158 Chapter 7 S U M M A R Y A N D CONCLUSIONS 159 REFERENCES 164 vii List o f Tables Table 1.1 Range of resin acid concentrations (ug/L) found in untreated and biotreated whole mill effluents derived from various pulping processes 7 Table 1.2 Toxicity (96-h L C 5 0 in ug/L) of resin acids to rainbow trout) 9 Table 1.3 General procedures or the development of an immunoassay 23 Table 2.1 Description of pulp mill effluent sources 46 Table 2.2 Retention times (min) of resin acids on different GC columns 53 Table 2.3 Resin acid concentrations of untreated CTMP effluents determined by GC using different extraction solvents and pH 56 Table 2.4 Resin acid concentrations of B K M E sample (from aeration inlet) determined by GC using different extraction solvents and pH 57 Table 3.1 Hapten densities estimated in various hapten-protein conjugates 78 Table 3.2 Percentage cross-reactivity of the polyclonal antibodies from D H A M - S U C - K L H with different resin acids and the lower and upper detection limit (DL) 83 Table 3.3 Recovery of D H A from water and effluent spiked with D H A by the indirect ELISA with the biotin-streptavidin system 85 Table 4.1 Conditions used for different ELISA formats and the corresponding I 50 and detection limits obtained 97 Table 4.2 Comparison of results obtained in the indirect ELISA with biotin-streptavidin system using different antibodies and different coating antigens 105 Table 4.3 Comparison of sensitivity in the direct ELISA using polyclonal antibodies from D H A M - S U C - K L H and D H A - K L H 107 Table 4.4 Cross-reactivities determined by antibodies with and without spacer arm for some compounds with similar structures to resin acids 109 Table 4.5 Comparison of OD readings obtained from an inhibition study using polyclonal antibodies from D H A M - S U C - K L H and different concentration of D H A with the microtitre plates sealed and unsealed at coating stage in direct ELISA 112 Table 4.6 The effect of the incubation time of the competition step on the OD reading and the 50% inhibition concentrations in the direct ELISA using antibodies from D H A M - S U C - K L H 114 Table 4.7 Comparison of the final OD readings when OPD and TMB were used as HRP substrates in a direct ELISA using polyclonal antibodies from D H A M - S U C - K L H 119 Table 4.8 Analysis of variance (ANOVA) for within-assay and between-assay variation using standard in the direct ELISA performed over four consecutive days 120 Table 4.9 Comparison of the standard curves and I50 obtained over four consecutive days 121 Table 4.10 Recovery of D H A from spiked assay buffer using antibodies from D H A M - S U C - K L H in a direct ELISA 123 viii Table 4.11 Recovery of D H A from tap water spiked with D H A using antibodies from D H A M - S U C - K L H in a direct ELISA.. 123 Table 4.12 Recovery of D H A from biotreated B K M E spiked with D H A by direct ELISA... : 124 Table 4.13 The percentage of false positive and false negative responses determined by the direct ELISA 125 Table 4.14 Recovery of D H A equivalents in a D H A - P I M mixture as determined by the direct ELISA using antibodies from D H A M - S U C - K L H 127 Table 4.15 Recovery of D H A equivalents in a D H A - A B A mixture as determined by the direct ELISA using antibodies from D H A M - S U C - K L H 127 Table 5.1 The D H A equivalents determined by direct ELISA for a CTMP effluent sample diluted in assay buffer 134 Table 5.2 The D H A concentration of a CTMP effluent assayed over four consecutive days as determined by direct ELISA 135 Table 5.3 The determination of the individual resin acids (by GC) and the D H A equivalents (by direct ELISA) present in three different sites with the wastewater stream of a CTMP mill 136 Table 5.4 The direct ELISA results for a B K M E sample diluted in a biotreated B K M E free of resin acids 142 Table 5.5 The direct ELISA results for a high resin acid containing B K M E sample diluted in a biotreated B K M E free of resin acids 142 Table 5.6 The influence of humic acid on the direct ELISA determination of D H A present in assay buffer 146 Table 5.7 The influence of alkali lignin on the direct ELISA determination of D H A present in assay buffer 146 Table 6.1 Recovery of D H A equivalents by direct ELISA for spiked bile samples 153 Table 6.2 D H A equivalents determined by direct ELISA for a bile sample at different dilutions 154 Table 6.3 D H A equivalents determined by direct ELISA in fish bile and calculated bio concentration factor (BCF) 155 Table 6.4 Individual resin acid concentration found in a pooled bile sample by GC and the D H A equivalents by direct ELISA 156 ix List of Figures Figure 1.1 Common resin acids found in softwood pulp mill effluents 4 Figure 1.2 Schematic structure of antibody (IgG) 18 Figure 1.3 Structures of triazine haptens used for antigen preparations 27 Figure 1.4 Conjugation of carboxylic acid to protein using mixed anhydride method 29 Figure 1.5 Conjugation of carboxylic acid to protein using carbodiimide method 30 Figure 1.6 Synthesis of antigen from amines using succinic anhydride method 30 Figure 1.7 Synthesis of antigen from amines using glutaraldehyde method 31 Figure 1.8 Illustration of the immune response to dioxin hapten-protein conjugate 33 Figure 1.9 Major steps involved in an indirect (step 1-4) and direct ELISA (step 5-7) ... 37 Figure 1.10 A n overview of the proj ect 43 Figure 2.1 Gas chromatogram of extracts before and after sample clean-up using SPE ..51 Figure 2.2 Gas chromatogram of standards run on DJ3-5 and DB-17 columns 52 Figure 2.3 Recovery of resin acids for A), distilled water spiked with resin acids and extracted with ethyl acetate, B).distilled water spiked with resin acids and extracted with M T B E , C). biotreated B K M E spiked with resin acids and extracted with ethyl acetate 54 Figure 2.4 Recovery of resin acids in the presence of high molecular weight (FfMW) material and humic acid (HA) at different pHs 59 Figure 3.1 Schemes for an attempted synthesis of hapten 72 Figure 3.2 Schemes for an attempted synthesis of hapten 73 Figure 3.3 Schemes for the syntheses of antigens with and without spacer arm 74 Figure 3.4 Schemes for the syntheses of POD-SUC-BSA and P O D - S U C - K L H 76 Figure 3.5 Schemes for the syntheses of DHA-OXI-SUC-BSA and D H A - O X I - S U C - K L H 77 Figure 3.6 Determination of titre for four different rabbits immunized with D H A M - S U C - K L H 80 Figure 3.7 Checkboard titration for polyclonal antibodies from one rabbit immunized with D H A M - S U C - K L H 80 Figure 3.8 Inhibition curve for D H A using the indirect ELISA with biotin-streptavidin system 81 Figure 4.1 Inhibition curves obtained from different ELISA formats 99 Figure 4.2 Monitoring of IgG using U V and ELISA for fractions collected in the purification of polyclonal antibodies using M A C protein-A discs 101 Figure 4.3 Chemistry of Dynabead activation and binding of protein to beads 101 Figure 4.4 A typical standard curve of magnetic bead ELISA 102 Figure 4.5 Percentage of inhibition in the direct ELISA using antisera from D H A M - S U C - K L H and D H A - K L H , with 50.0 ug/L of inhibitors 108 Figure 4.6 The effect of sample pH on the final OD readings in the direct ELISA 115 Figure 4.7 The effect of ionic strength of the sample on the OD readings in the direct ELISA 116 Figure 4.8 The effect of metallic ions on the OD readings in the direct ELISA 117 x Figure 4.9 The effects of methanol on the direct ELISA using antibodies derived from D H A - K L H and D H A M - S U C - K L H 137 Figure 5.1 The correlation of D H A equivalents determined by direct ELISA and the total resin acid content determined by GC for various effluents collected at different sites in a CTMP mill 137 Figure 5.2 A typical D H A standard curve and the dilution curve for a B K M E sample ... 139 Figure 5.3 Correlation between D H A equivalents determined by direct ELISA and total resin acid content determined by GC for a range of B K M E samples collected from a mill 143 Figure 6.1 Steps involved in ELISA and GC analysis of resin acids in fish bile 158 xi List of Abbreviations 6 chemical shift ug microgram (0,1 microliter Ab antibody (antibodies) Ab-Ag complex of antibody -antigen interaction A B A abietic acid A g antigen A O X adsorbable organic halides B K M E bleached kraft mill effluent BOD biochemical oxygen demand B S A bovine serum albumin 1 COD chemical oxygen demand CR cross-reactivity CTMP chemi-thermomechanical pulping C V coefficient of variance d doublet dd doublet of doublet D C C dicyclohexylcarbodiimide D E C l-(3-dimethylaminopropyl)-3-ethyl-carbodiimide D H A dehydroabietic acid D H A - B S A conjugate from D H A and B S A D H A - K L H conjugate from D H A and K L H D H A M dehydroabietylamine D H A M - S U C - B S A conjugate from D H A M and B S A with a four-carbon spacer arm D H A M - S U C - H R P conjugate from D H A M and HRP with a four-carbon spacer arm D H A M - S U C - K L H conjugate from D H A M and K L H with a four-carbon spacer arm D M F dimethylformide D M S O dimethylsulfoxide EIA enzyme immunoassays ELISA enzyme linked immunosorbent assay EtOAc ethyl acetate Fab antibody fragment containing one of the antibody binding sites xii Fc region of antibody composed of the C-terminal domain of the heavy chain FLD flame ionization detector g gram GC gas chromatography GC-MS gas chromatography coupled with mass spectrometer Ff heavy chain of antibody H A humic acid H M W high molecular weight H P L C high performance liquid chromatography hr(s) hour(s) HRP horse radish peroxidase I 50 concentration required for a 50% reduction in inhibition IgA immunoglobulin A IgD immunoglobulin D IgE immunoglobulin E IgG immunoglobulin G IgM immunoglobulin M IR infrared spectrometry J coupling constant (in hz) K L H keyhole limpet hemocyanin L light chain of antibody L C 5 0 lethal concentration for 50%.population L M W low molecular weight M molar concentration M H S methyl heneicosanoate min minutes ml milliliter mmol millimole MS mass spectrometry M T B E methyl t-butyl ether m/z mass/ion ratio ng nanogram NHS N-hydroxysuccinimide nm wavelength in nanometer °C temperature in degree centigrade Xlll OD optical density O M P A o-methylpodocarpic acid PBS phosphate-buffered saline pH - log concentration of hydrogen ion, [H + ] P IM pimaric acid POD podocarpic acid R correlation coefficient RIA radioimmunoassays s.d. standard deviation T C A tricosanoic acid T M B tetramethylbenzidine TMP thermomechanical pulping TMS tetramethylsilane TSS total suspended solids xiv Acknowledgement I wish to express my sincere appreciation to Drs. John N . Saddler and Colette Breuil for their excellent guidance, valuable suggestions and constructive criticism, both during the progress of this research and in the preparation of this dissertation. I would also like to thank my other advisory committee members, Drs. Eric Hall, James P. Kutney, and Gerry Reimer, for giving me various comments and suggestions during the progress of this project. Thanks also go to Drs. Paul Bicho, Tao Chen, Ed de Jong, Alex Serreqi, Ken Wong, and Mr. Yong Gao for suggestions and helpful discussion on various research topics, and to Mr. Michael Chester, Miss Xiumei Feng and Mr. Kelvin Stark for their technical assistance. Special thanks to Drs. Paul Bicho, Carl Johansson and Alex Serreqi for proofreading. At the same time, I wish to thank all the members (both past and present) of the Chair of Forest Products Biotechnology for their help and friendship. Financial supports from Science Council of British Columbia (SCBC), National Science and Engineer Research Council (NSERC) and Harmac Pacific Inc. are greatly acknowledged. Finally, I am deeply indebted to my wife, Linda Y . P. Qin, and to my sons, Tony and Daniel, for their support, encouragement and co-operation. xv Note A l l names for resin acids used throughout this thesis are common names and the numbering system is derived from that for steroid as shown below for the abietane and pimarane skeletons (Rowe, 1969). The nomenclature and numbering system for resin acids used by Chemical Abstracts has changed repeatedly, seemingly with each collective index and index guide. In the literature, resin acids are usually named based on abietane, pimarane and isopimarane skeleton. For example, dehydroabietic acid (DHA) could be named as 8,ll,13-abietatrien-18-oic acid based on abietane. Currently (from 1973 to present), resin acids are indexed by Chemical Abstracts (CA) using phenanthrene as the parent compound for the tricyclic rings. Therefore, D H A is named as [1R-(1 a, 4ap, 10aa)]-l, 2, 3, 4, 4a, 9, 10, 10a-octahydro-l, 4a-dimethyl-7-(l-methylethyl)-l-phenanthrenecarboxylic acid. xvi The common names, C A names and registration numbers, and IUPAC names for eight common resin acids were listed below (Soltes and Zinkel, 1989): Common Name CA Registration No CA Name IUPAC Name Dehydroabietic acid 1740-19-8 1-phenanthrenecarboxylic acid 1, 2, 3, 4, 4a, 9, 10, lOa-octahydro-1, 4a-dimethyl-7-(l-methylethyl)-[ l R - ( l a , 4ap, 10aa)] 8,11,13-abietatrien-18-oic acid Abietic acid 514-10-3 1-phenanthrenecarboxylic acid 1, 2, 3, 4, 4a, 4b, 5, 6, 10, lOa-decahydro-1, 4a-dimethyl-7-(l-methylethyl)-[lR-(loc, 4aP, 4ba, 10aa)] 7,13-abietadien-18-oic acid Levopimaric acid 79-54-9 1-phenanthrenecarboxylic acid 1, 2, 3, 4, 4a, 4b, 5, 9, 10, lOa-decahydro-1, 4a-dimethyl-7-(l-methylethyl)-[ l R - ( l a , 4aP, 4ba, 10aa)] 8(14),12-abietadien-18-oic acid Neoabietic acid acid 471-77-2 1-phenanthrenecarboxylic acid 1, 2, 3, 4, 4a, 4b, 5, 6, 7, 9, 10, lOa-dodecahydro-1, 4a-dimethyl-7-(l-methylethylidene)-[lR-(loc, 4ap, 4ba, 10aa)] 8(14), 13(15)-abietadien-18-oic Palustric acid 1945-53-3 1-phenanthrenecarboxylic acid 1, 2, 3, 4, 4a, 5, 6, 9, 10, lOa-decahydro-1, 4a-dimethyl-7-(l-methylethyl)-[lR-(lot, 4ap, 10aa)] 8, 13-abietadien-18-oic acid Isopimaric acid 5835-26-7 1-phenanthrenecarboxylic acid, 7-ethyenyl-1, 2, 3, 4, 4a, 4b, 5, 6, 7, 8, 10, lOa-dodecahydro-1, 4a, 7-trimethyl-[ l R - ( l a , 4ap, 4ba, 7a 10aa)] 7, 15-isopimaradien-18-oic acid Sandaracopimaric acid 471-74-9 1-phenanthrenecarboxylic acid, 7-ethyenyl-1, 2, 3, 4, 4a, 4b, 5, 6, 7, 9, 10, lOa-dodecahydro-1, 4a, 7-trimethyl-[ l R - ( l a , 4ap, 4ba, 7a 10aa)] 8(14), 15-isopimaradienn-18-oic acid Pimaric acid 127-27-5 1-phenanthrenecarboxylic acid, 7-ethyenyl-1, 2, 3, 4, 4a, 4b, 5, 6, 7, 9, 10, lOa-dodecahydro-1, 4a, 7-trimethyl-[ l R - ( l a , 4ap, 4ba, 7p 10aa)] 8(14), 15-pimaradien-18-oic acid xvii CHAPTER 1 INTRODUCTION Canada is one of the largest pulp and paper producers in the world and the pulp and paper industry continues to play a significant role in the Canadian economy. For example, in 1994, 49 Canadian pulp mills represented 28% of the world's pulp production capacity and Canada ranked second in the world in terms of total wood pulp production (24.6 million tonnes). About 10.0 million tonnes, valued at $6.4 billion, were exported. These shipments alone accounted for 2.8% of Canada's total merchandise exports, and contributed one quarter of Canada's 1994 overall surplus of $24.6 billion (CPPA, 1995). British Columbia's contribution to Canada's total pulp and paper production in 1994 reached 7.6 million tonnes (31.1%) and 3.0 million tonnes (16.1%), respectively. These accounted for 16.8% of the total manufacturing shipments in British Columbia (CFI, 1995). Unfortunately, the pulp and paper industry has and continues to generate considerable amount of effluents due to the large amount of water used for chip washing, debarking, pulping, pulp washing, bleaching and papermaking operations; For example, in 1989, the total mill effluent discharged from all Canadian pulp mills averaged 104,000 m-Vd, which is approximately equal to the flow of the Columbia River in B.C. (CEPA, 1991). Another study indicated that approximately 2000 tonnes per day of biochemical oxygen demand (BOD) and 800 tonnes per day of total suspended solids (TSS) were released in 1987 (Halliburton and Chang, 1991). Therefore, the waste waters from the pulp and paper industry continue to be a highly profiled environmental issue. As a result, there is a high degree of public, regulatory and scientific concern over the discharge of pulp and paper mill effluents into the environment and many countries have legislated effluent discharge limits (CEPA, 1991). Tremendous efforts have been made in Canada to reduce effluent discharge and the Canadian pulp and paper industry wil l 1 have invested a total of $5.3 billion in pollution reduction measures between 1989 and 1996 to improve both air and water quality (Lachapelle, 1996). It is recognized that the wastewaters from the pulp and paper industry are highly heterogeneous and contain literally thousands of organic and inorganic substances derived from the wood furnish, process chemicals or formed during the pulping processes. Usually these liquid wastes are passed through some form of wastewater treatment facility before being discharged to the environment. However, even after wastewater treatment, the remaining liquid effluents can still contain a variety of undesirable substances which may cause adverse environmental effects when discharged. Therefore, these effluents are considered potential sources of both conventional and non-conventional pollutants. The so-called conventional pollutants include TSS, BOD and toxicity. These components usually have an immediate effect on the aquatic life in the receiving waters. To date, most abatement efforts have focused on the removal of solids and oxygen demand. Non-conventional pollutants are those trace amounts of organic compounds which are only present in the receiving waters in very low concentrations and usually have long term rather than immediate effects, such as bioaccumulation and sub-lethal toxicity in aquatic life (Cherwinsky and Murray, 1986). More than 250 toxic chemicals have been identified in pulp mill effluents (Suntio et al, 1988). Although these compounds can be usually classified into three main groups as neutrals, phenolics and acids, the chemical compositions of pulp mill effluents are not well characterized due to their complexity and the variability. These characterized compounds may represent only a very small fraction of the total chemicals present. It has been shown that a considerable amount of the toxic trace organic compounds found in pulp mill effluents are derived from the wood itself and are released during various pulping processes. Wood is composed of hundreds of complex molecules, which can be classified into four major groups of cellulose, hemicellulose, lignin and extractives (Sjostrom, 1993). While cellulose and hemicellulose are the primary source of wood fibre, it is usually the lignin and wood extractives that are modified during pulping and released into receiving waters and result 2 in fish toxicity. The extractives primarily consist of resin acids (from softwoods), fatty acid esters and unsaponifiables comprised of sterols, alcohols and hydrocarbons (Fengel and Wegener, 1984). It is now generally accepted that the acute toxicity of pulp mill effluents is largely attributed to resin and fatty acids derived from wood extractives, chlorinated phenols derived from lignin during pulp bleaching using chlorine or chlorine dioxide, and to lesser extent, a broad group of neutral compounds (McLeay & Assoc., 1987). 1.1. Resin Acids 1.1.1. Occurrence and toxicity of resin acids Resin acids are a class of diterpenoid carboxylic acids present mainly in the wood of conifer trees and are the major components of softwood extractives. They can be liberated from wood chips at different stages of the pulping process and are carried over into pulp mill effluents. The amount of resin acids in softwood pulp mill effluents can vary considerably depending on the wood furnish in the mill, the age of wood chips used, the mill process and the extent of biological treatment prior to effluent discharge (Taylor et al, 1988). In general, pulping of softwoods can release mg/L (mg/L) levels of the major resin acids in untreated whole mill effluents (Beak, 1987). A wide range of resin acid concentrations have been observed, ranging from undetectable to 10 mg/L for B K M E , while some C/TMP effluents can have levels as high as 50 mg/L (McLeay & Assoc., 1987). The resin acids commonly found in softwood pulp mill effluents are: dehydroabietic acid (DHA), abietic acid (ABA), neoabietic acid (NEO), pimaric acid (PIM), isopimaric acid (ISO), sandaracopimaric acid (SAND), levopimaric acid (LEV) and palustric acid (PAL) (Figure 1.1). They can be further classified into two major types. The pimaric type is characterized by the presence of both methyl and vinyl substituents at the C-13 position and the abietic type is 3 Neoabietic acid Levopimaric acid Palustric acid Sandaracopimaric acid Isopimaric acid Pimaric acid Figure 1.1 Common resin acids found in softwood pulp mill effluents 4 characterized by bearing only an isopropyl group at the C-13 position. Among these common resin acids, D H A is the most abundant and persistent, and can form chlorinated derivatives, such as monochloro- and dichlorodehydroabietic acids, by electrophilic substitution of chlorine on the aromatic ring during pulp bleaching using chlorine and chlorine dioxide. A l l of these resin acids have been found in the effluents of Canadian mills (McLeay & Associates, 1987). In addition there may be other, rarer compounds such as 7-oxodehydroabietic acid, 7, 15-isopimaric acid and 8,15-isopimaric acid, and hydroxydehydroabietic acids, which are only present in trace quantities (Taylor et al, 1988). Resin acids are the major components of wood extractives from softwood species. Since Canada is rich in softwood resources and its reserve of 17,800 million cubic metres represents over 15% of the world total non-tropical forest growing stock of softwoods (CFI, 1995), softwoods have become the major source of pulpwoods in Canada. Various resin acids have been found in Canadian pulpwoods (Swan, 1973). Some coniferous species, like the pines (Pinus) are generally rich in resin acids (1.5% based on oven dried wood). Although the presence of resin ducts is a normal feature of pines, spruces (Picea), larches (Larix) and Douglas fir (Pseudotsuga menziesii), this is not always indicative of high resin acid content as exemplified by the spruces which generally have a low resin acid content (0.12%). In other softwoods, like the true firs (Abies), hemlocks (Tsuga) and redwoods (Sequoia), resin ducts are normally absent although resin acids may be produced in response to injury. Finally, in certain conifers, such as western red cedar (Thujaplicata), resin acids are virtually absent. It should be realized that resin acid content is not only a function of inter species variations, it also differs in the distribution within the same tree. Usually heartwood is higher in resin acid content than sapwood (Hemingway and Hills, 1971). Highest resin acid content occurs in the bark itself as these compounds serve a defensive purpose against wood-boring insects and associated pathogenic microorganisms (Shrimpton, 1973). Due to the wide variation of resin acid content, the dominant species in a mill's furnish may not necessarily be the dominant source of resin acids in its effluent (Taylor et ai, 1988). There are several factors other than wood species 5 which affect the resin acid content, such as different geographical areas, different seasons for harvesting, the age and diameter of the tree, and seasoning of wood chips. For example, the resin acid content is known to decline as the duration of chip storage is increased due to resin acid auto-oxidation or biological oxidation (Rogers et al, 1971). Since resin acids are natural constituents of the wood furnish, they are released predominantly during bark removal and the pulping stage itself. Different resin acid concentrations have been found in the effluents from different pulping technologies. It has been shown that the concentration of resin acids in untreated pulp and paper wastewaters varies considerably and is affected by the type and the age of the wood feedstock and the mill operation conditions used (Taylor et al, 1988). Table 1.1 summarizes the concentrations of resin acids presented in the wastewaters before and after biological treatment from common pulping processes. In the kraft process (Mimms, 1993), which is the predominant pulping process used in Canada, wood chips are digested under pressure with a mixture of hot caustic soda and sodium sulfide. This cooking liquor solubilizes lignin and extractives, releasing free resin acids as their sodium salts, while leaving insoluble cellulose fibres as pulp. Most of these resin acid salts, along with some neutral materials are dissolved or suspended in the spent black liquor and are recovered as tall oil soap (Hough, 1985). However, the soap causes foam overflow occasionally, which can escape into the effluent stream and cause a "toxic" shock on the effluent treatment system. Levels of resin acids vary widely for both unbleached and bleached kraft whole mill effluents. Due to the very high level of dilution used in bleach plants, the concentration of resin acids is typically less in final effluents derived from bleached kraft pulping than those obtained from unbleached pulps. In mechanical pulping processes, the conversion of wood into fibres is achieved by mechanical grinding, with the help of thermal energy (thermomechanical pulping, TMP) or 6 Table 1.1 Range of resin acid concentrations (ug/L) found in untreated and biotreated whole mill effluents derived from various pulping processes (data in bracket are for biotreated effluents) (Taylor et al, 1988) Resin acid Pulping Process U K M E B K M E U S M E B S M E GDWD CTMP Abietic 30-9970 (<20-3630) <20-4800 (<10-1780) 520-4840 (437-500) <10-1000 (<10-100) 210-16000 (14-4200) 4210 (8860) D H A 990-5780 (<20-1930) 30-4580 (<1-2140) 700-4620 (247-1100) <20-8500 (10-700) 490-15100 (8-580) 5330 (15370) Isopimaric 70-4120 (<20-1420) <20-4800 (<10-930) 100-5070 (100-294) < 10-3 00 (<10-310) 150-9300 (12-7900) 2410 (6940) Levopimaric <10-2700 (< 10-30) <10-2400 (<1-1190) 100-510 (200) <10-100 (<10-60) 80-22000 (11-1800) Neoabietic <50-1200 <10-1000 (<1-150) 30-6800 (<1-3800) Palustric 90-100 (80) 300-7700 Pimaric 100-1830 (<20-890) <20-1010 (14-540) 490-1140 (20) <20-30 (<20) 20-6800 (<5-5700) 2230 (4350) Sandaracopimaric 650 (1830) U K M E : unbleached kraft mill effluent B K M E : bleached kraft mill effluent U S M E : unbleached sulphite mill effluent BSME: bleached sulphite mill effluent G D W D : groundwood, mechanical pulping (C)TMP: (chemi)thermomechanical pulping 7 both thermal energy and chemicals (chemi-thermomechanical pulping, CTMP). Despite very little delignification, wood extractives are still released by this process. In fact, CTMP effluents are highly concentrated wastes since little wash water is used in mechanical pulping and no chemical recovery process is applied. Therefore, any material dissolved out of the wood generally ends up in the effluents which, as a result, contain higher level of BOD and COD. Generally, these effluents are more toxic to fish and the microorganisms present in the waste water treatment systems than the effluents from pulp mills using the older mechanical pulping processes such as TMP. As both degradable and recalcitrant organic compounds are present in high concentration in the mechanical effluents, the resin acid content is also usually present at elevated levels, when compared to effluents from kraft and sulphite operations. The main issue regarding the amount of resin acids present in.pulp mill effluents is their toxicity to aquatic life. Since the 1930's, acute lethality tests have indicated that resin acids are responsible for a large proportion of the acute toxicity of pulp mill wastewaters to fish. Resin acids are still considered to be a primary source of toxicity to fish in Canadian softwood pulping effluents (McLeay & Assoc., 1987). It has been reported that resin acids may contribute to as much as 70% of the toxicity of whole pulp effluents from B K M E and up to 90% of the toxicity of whole pulp effluents from CTMP (Leach and Thakore, 1976, 1977). The lethal concentration (LC50) of individual resin acids to rainbow trout ranges from 0.2 to 1.7 mg/L at neutral pH (Table 1.2). This range is so narrow that the toxicity of pulp mill effluents containing resin acids wil l primarily depend upon the total concentration of resin acids rather than the toxicity of each individual form. Among the eight common resin acids, D H A is the most persistent and most abundant form found in most softwood pulp mill effluents. The mono and dichlorinated dehydroabietic acids, which are formed by electrophilic substitution of chlorine on the aromatic ring during pulp bleaching using chlorine, are substantially more toxic than the parent molecule (McLeay & Assoc., 1987). In addition to their acute toxicity, resin acids have also been shown to be responsible for sublethal effects, bioaccumulation and some genetic effects (Beak, 1987; Bonsor et al, 1988; McLeay and Assoc., 1987; Taylor et al, 1988). . A typical example of the 8 toxic effect of resin acids was the identification of D H A as the cause of a large scale fish-kill observed in the River Tajo in Spain during the Spring-Summer of 1991 (Munoz et al, 1994). Table 1.2 Toxicity (96-h LC50 in ug/L) of resin acids to rainbow trout (Taylor et al, 1988) Resin acids L C 5 0 Abietic acid 700-1500 Chlorodehydroabietic acid 600-900 Dehydroabietic acid 800-1740 Dichlorodehydroabietic acid 600-1200 Isopimaric acid 400-1000 Levopimaric acid 700-1000 Neoabietic acid 610-730 Palustric acid 500-600 Pimaric acid 700-1200 Sandaracopimaric acid 360 The acute toxicity of resin acids is strongly influenced by the pH of the receiving waters. For example, McLeay and Associates (1979) showed that resin acids are least toxic in alkali pH. In these conditions resin acids are ionized (98% at pH 9) and thus, are more polar, more water soluble and less soluble in lipids. D H A , for example, is 15 to 30 times more toxic at pH 6.5 than at pH 9.0. In 1991 the toxicity regulations of the Federal Fisheries Act were revised to state that mill effluents must have a 96h LC50 of no less than 100% if they are to be considered non-toxic. That is, more than 50% of the test fish must survive 96 hours in the undiluted (100%) mill effluent. The proposed limit of resin acid allowed in receiving water in the Province of Ontario 9 are currently set at 8 \xg/L at pH 7 and 14 u.g/L at pH 9 for D H A , while the total resin acid content is set at 25 and 62 ug/L at the respective pH (Taylor et al, 1988). Unlike the reduction in chlorinated phenolics that has been attained in the past ten years, toxicity due to resin acids will probably continue to be a potential problem despite changes in the pulping and bleaching technologies. This is primarily because resin acids are naturally derived rather than process derived. These compounds will continue to be discharged into receiving waters from pulp and paper industry until an economical viable closed cycle mill with no effluent discharge is developed. It is expected that, as problems associated with adsorbable organic halides (AOX) become less important, more attention will be paid to resin acids. 1.1.2. C u r r e n t methods o f res in a c i d ana lys is a n d t h e i r d r a w b a c k s During the past twenty years, various efforts have been made to develop accurate, sensitive and reliable analytical methods for resin acid analysis. However, as resin acids occur in a complex matrix, their detection and quantification are difficult to achieve. To date, a standardized method for the analysis of resin acids in pulp mill effluents still does not exist. Therefore, the analyst attempting to develop an appropriate analytical method must make a number of choices regarding the isolation, separation, purification and identification techniques that wil l be used. Major steps involved in resin acid analysis include isolation, derivatization and identification/quantification. 1.1.2.1. I s o l a t i o n Isolation of analytes from a sample matrix is of crucial importance in the entire analytical method since it defines the maximum recovery of the analytes of interest, as well as possible co-extractives that may interfere with the quantitation. Two procedures commonly used for the isolation of resin acids from pulp mill effluents are liquid-liquid extraction using various organic 10 solvents and solid phase extraction through adsorption of resin acids on macroreticular polymer. In the first method, the effluent is usually acidified and extracted with organic solvents, such as diethyl ether and methylene chloride (NCASI, 1986). Although solvent extraction is, in principle, fast and simple, direct solvent extraction of pulp mill effluent results in emulsions which increase preparation time and result in low recoveries. Very acidic conditions (pH 2-3) may be used to minimize emulsion. However, this can lead to other problems like the isomerization of resin acids, particularly levopimaric acid (Mahood and Rogers, 1975). A n alternative solvent extraction procedure using methyl tert-butyl ether (MTBE) under alkali conditions (pH 9) has been used to overcome many of the emulsion problems associated with acidic extraction (Voss and Rapsomatiotis, 1985). In solid phase extraction, resin acids are removed from the effluent under alkali conditions by adsorption onto a porous polymer called Amberlite XAD-2 . The resin acids which are adsorbed onto the polymer are then eluted with diethyl ether and methanol (Leach and Thakore, 1975). This method can overcome the limitations associated with solvent extraction, as there is no solvent to form emulsions and the pH is kept very alkaline. However, recoveries are generally low. Recently, commercially available solid phase extraction cartridges, packed with absorbents of different polarities, have made extraction and cleaning much easier (Analytichem International, 1990). These cartridges have been successfully used in the preconcentration of resin acids from an aqueous matrix (Richardson et al, 1992). This technique provides advantages such as ease of performance and lower solvent consumption. Extraction of resin acids by supercritical fluid has also been reported (Lee and Peart, 1992). 1.1.2.2. Separation and derivatization Following suitable isolation, resin acids may be separated and quantified by a number of different chromatographic techniques such as column, paper, liquid and gas-liquid chromatographies. Spectrometric techniques such as infrared, ultraviolet, nuclear magnetic 11 resonance and mass spectrometry may then be used for identification, once the resin acids have been separated (Browning, 1967). Currently, both gas chromatography (GC) and high performance liquid chromatography (HPLC) are widely used for identification purpose. However, derivatization is usually necessary for both these methods, as GC requires the volatility of the analyte and H P L C requires a chromphore present in the compound if U V detector is used. Derivatization of resin acids with carbonyldiimadazole (Kutney et al, 1981; NCASI , 1986) or coumarin (Richardson et al, 1992) has been used for H P L C analyses. In GC analysis using flame ionization detector (FED), derivatization with diazomethane. is the most common method used. However, diazomethane is toxic and explosive, and considered to be carcinogenic. Other derivatization procedures have been attempted, such as using triethyloxonium tetrafluoroborate to produce ethyl ester derivatives (NCASI, 1986). A n electron capture detector (ECD) can also be used in GC analysis when resin acids are derivatized as pentafluorobenzyl esters (Lee et al, 1990). It is apparent that the isolation and quantification of resin acids are still two imperfectly solved problems and that a sensitive, selective and inexpensive analytical method is highly desirable. 1.1.2.3. Identification and quantification Although GC has been applied to the analysis of resin acid methyl esters as early as 1959 (Hudy, 1959), this technique is still the method of choice currently used for separation and identification of particular resin acids. While early literature used packed columns with a variety of liquid phases (Zinkel and Han, 1986), quantification of individual resin acids can now be effectively achieved using high resolution capillary GC columns. For routine quantitative analysis of resin acids in pulp and paper mill effluents, gas chromatography using flame ionization detector (GC-FID) is the most commonly used method (Foster and Zinkel, 1982; Taylor et al. 1988). Gas chromatography coupled with mass spectrometer (GC-MS) provides a more confident identification, however, the instrumentation is far more expensive. H P L C can 12 also be used in the quantification after suitable derivatization (Kutney et al, 1981; NCASI , 1986; Richardson a/., 1992). Recently, an FfPLC based method has been reported to quantify D H A in pulp mill effluents without derivatization (Shepard et al, 1996). The method has shown good correlation between the D H A concentration and the amount of total resin acids present in mill effluents (r = 0.97). However, this still requires the use of expensive equipment and a dedicated technician. Also, this method lacks specificity as only D H A can give a response to the UV detector used. In addition to chromatographic methods, some colorimetric or spectrophotometric methods have also been developed for the quantification of resin acids. For example, Carpenter's method can be used to monitor the presence of specific resin acids containing conjugated diene structure, such as abietic acid, by reaction with 65% sulfuric acid and acetic anhydride to form a transient pink coloured compound which can be monitored at 518 nm (Carpenter, 1965). Recently, a rapid spectrophotometric procedure has also been developed for the determination of total resin and fatty acids in pulp and paper matrices (Sithole, 1993). 1.1.2.4. Drawbacks of conventional analytical methods While conventional analytical methods such as FfPLC, GC and GC-MS can generally provide reliable, accurate quantification of the analyte of interest, use of these methods in the measurement of environmental samples usually requires a considerable amount of sample work-up prior to final quantification. These procedures usually include 1). isolation of analytes from the sample matrix, 2). concentration of the analytes to achieve detection limits, 3). sample clean-up to reduce interference, 4). derivatization to meet the instrument requirements, 5). separation, 6). identification and 7). quantification (Keith et al, 1983). They also usually require expensive equipment, chemicals and trained personnel, and often result in low sample throughput. Some of analyses are extremely expensive due to the extensive workup procedures required. For example, a complete analysis of dioxins in pulp mill effluent was estimated to cost $l,100/sample while a 13 MESA (Municipal and Industrial Strategy for Abatement) group analysis package was estimated to cost $3300 for effluent and $4170 for sludge (Enviro-Test Laboratories, 1990). Resin acid analysis, as part of the above package, is estimated to be around $200/sample by GC and this wil l make daily analysis rather expensive. A further drawback is that the analytical results would not be available within a short period of time, due to the time required to transport the samples and the lengthy analytical procedure. Another consideration is that, since no standard method is available for resin acid analysis, results from various laboratories using different protocols for extraction, derivatization and chromatographic quantification may result in significant variations. A n inter-laboratory study involving eight laboratories from Canada, U S A and New Zealand indicated that variability ranged between 20-55% for the standard mixture and the variations among the real samples was even higher (Bicho et al, 1995a). A l l these factors limit the use of these conventional instrumental methods for the screening of numerous environmental samples. It is apparent that these current methods cannot easily cope with the steadily increasing demand for fast, accurate values neccessary to support regulatory compliance. Therefore, alternative analytical methods using fast and simple procedures, low cost and reasonable sensitivity are highly desirable. Immunochemical technologies, which have been routinely used in clinical laboratories, have proven to be highly accurate and reproducible analytical tools for medical diagnostics used in applications such as disease detection and drug monitoring. For example, immunoassays that can detect pregnancy are routinely used at home and in the doctor's office. Similar procedures have also been adopted in environmental analysis and are used as an alternative way of analyzing various low molecular weight compounds of environmental importance. In the past ten years, a considerable amount of progress has been made in the analysis of environmental pollutants, such as pesticides and their metabolites, industrial chemicals and microbial toxins (Vanderlaan et al., 1990; Sherry, 1992; Van Emon and Lopez-Avila, 1992; Van Emon and Gerlach, 1995a). These immunochemical methods require antibodies which specifically interact with the target analyte and the detection is based on the interaction between the target analyte 14 and the specific antibody. Although these procedures are not without some problems and limitations, such as cross-reactivity and the requirement of antibodies, immunoassays have proven to be sensitive methods and are especially suited to the analysis of environmental compounds that are difficult to detect by conventional methodologies. 1.2. I m m u n o a s s a y 1.2.1. I n t r o d u c t i o n to i m m u n o a s s a y Immunoassays are among the most useful immunochemical techniques for both research and clinical applications. This procedure is based on the interaction between an antibody and its antigen, which are two major components of the assay system. Antibodies can bind specifically to the antigen to form an antibody-antigen (Ab-Ag) complex, which can then be quantified using a variety of techniques (Chait and Ebersole, 1981; Harlow and Lane, 1988; Monroe, 1984). A wide range of methods can be used to detect and quantify either antigens or antibodies. With the appropriate assay formats, the detection and quantification can be remarkably quick and easy, yielding information that would be difficult to determine by other techniques. It should be recognized that although antibodies are produced by a biological process, immunoassays are nevertheless chemical analytical procedures and the reaction between antigen and antibody is also governed by Mass Action Law. 1.2.1.1. A n t i b o d y a n d its s t r u c t u r e A n antibody is produced by mammalian B-lymphocytes as part of the immune system's response to the invasion by a foreign substance (immunogen). By definition, immunogens are substances that, when administered to an animal in the appropriate manner, can elicit an immune response. There is a clear operational distinction between the immunogen and antigen. A n antigen is defined as any substance that can bind selectively to a specific antibody (Janeway and 15 Travers, 1994). Therefore, immunogens must be antigens. However some substances can be antigenic, but not immunogenic. Although they do not mean exactly the same thing, the terms antigen and immunogen are often used interchangeably. Therefore, in the subsequent text, only the term antigen is used, although it may refer to an immunogen. In general, only macromolecules which are foreign to the host animal and with some degree of chemical complexity have immunogenicity and can be used to immunize animals. The classes of molecules and structures most commonly used as antigens cover an enormous range of natural and synthetic materials such as proteins, nucleic acids, lipids, carbohydrates, viruses, bacteria, fungi and a wide variety of xenobiotics. An important requirement is that the molecules must be of fairly high molecular weight, usually greater than 10,000. As a general rule, the more complex the molecule, the more effective it wil l be as an antigen. Unfortunately, most of the. analytes in environmental analysis, such as synthetic pesticides and their metabolites, are low molecular weight compounds (<1000) that are not immunogenic (Erlanger, 1980). They can not by themselves stimulate antibody production. In order to make these low molecular weight compounds immunogenic, the small molecules (called haptens) must be covalently linked to a larger molecule (usually a protein, called carrier) that is immunogenic. The hapten-protein conjugate thus obtained can be used to immunize an animal. Since antigens are large molecules that can contain several antigenic sites or determinants, the antiserum generated in the animals immune system will contain a mixture of antibodies (called polyclonal antibodies). Usually, three distinct sets of antibodies are produced upon immunization with a hapten-carrier conjugate. One set comprises the hapten-specific antibodies that wil l react with the hapten on any carrier as well as free hapten. The second set of antibodies wi l l react with the unmodified carrier protein. Finally, some antibodies react only with the hapten-carrier conjugate used for immunization. Obviously, only those antibodies capable of recognizing the hapten (i.e., analyte) can be used in various immunochemical methods for the analysis of a specific analyte. 16 Chemically, antibodies are fairly large, multichain carbohydrate-containing proteins ranging in molecular weight from 150,000 to 900,000 Daltons and have isoelectric points ranging from 4.5 to 9.5. They are generally present at a concentration of 12-15 mg/ml in the blood serum, and nearly constitute one fifth of the serum protein content (Paraf and Peltre, 1991). There are several different classes of antibodies. Five primary classes of immunoglobulins have been identified in mammals and they are distinguished by the type of heavy chain found in the molecule. These are described as the IgG, IgM, IgA, IgD and IgE immunoglobulins. They differ in molecular weight, carbohydrate content and serum half-life, and also function somewhat differently in the clearance of foreign substances from the body. The most commonly used antibody in immunoassays is IgG, which represents 70% of the immunoglobulins present in animal serum. The molecule resembles the letter Y and.consists of four polypeptide chains, two identical heavy chains (H), of approximately 50-77 kd, and two identical light chains (L), of approximately 25 kd. These two peptide chains are held together and stabilized by intrachain and interchain disulfide bonds (Paraf and Peltre, 1991). Light chains are composed of two globular regions or domains, a constant ( C L ) and a variable ( V L ) one, while the heavy chains are composed of four domains, three constant (C j j l , Cjj2, and Cj-j3) and one variable (VJJ) . That is to say, each chain is composed of a variable region where the sequence of amino acids is unique for a given species of antibody and a constant region with a definite sequence of amino acids for a given heavy chain class or light chain type. The antigen binding sites are at the tip of each arm (Figure 1.2). Antibodies can be cleaved into small fragments. For example, papain splits the Ig molecule into three fragments of similar size: two Fab fragments and one Fc fragment (Figure 1.2). The Fc fragment is a dimer of the two last heavy chain domains and the Fab is a monomer composed of one light chain and the two upper heavy-chain domains. Digestion by pepsin produces one F(ab')2 fragment and numerous small peptides of the Fc portion. The resulting F(ab')2 fragment is composed of two disulfide-connected Fab units, plus the hinge region (Paraf and Peltre, 1991). As both Fab and F(ab')2 have variable domains, they maintain the specificity 17 for antigens,. They also offer several advantages over intact antibodies as reagents in an immunochemical technique, such as reduced non-specificity resulting from Fc interaction and higher sensitivity in antigen detection as a result of reduced steric hindrance from large protein epitopes. Fab H : V : Antigen-binding sites-carbohydrate < heavy chain variable region c H i c m CH2 C H 3 carbohydrate Fab papain L Fc L F.(ab')2 L : light chain C: constant region pepsin Figure 1.2 Schematic structure of antibody (IgG) As described previously, one of the characteristics of an antibody is its ability to bind specifically to an antigen to form an Ab-Ag complex. The Ab-Ag complex is linked entirely by noncovalent interactions, mainly by hydrogen bonds, electrostatic forces, Van der Waals forces and also hydrophobic bonds (Harlow and Lane, 1988). As no covalent bonds are involved, there is usually a thermodynamic reversibility to the antibody-antigen reaction. Similar to other 18 chemical equilibria, the strength of the interaction between Ab and Ag can be expressed using the affinity constant, Ka, which is obtained from the application of Mass Action Law for the following equilibrium. Ab + Ag <=> Ab-Ag Ka=[Ag-Ab]/[Ag][Ab] In practical terms, affinity describes the amount of antibody-antigen complex that will be found at equilibrium. Antibodies of high affinity for antigens have a K a from IO** to 10H liter/mole. Such antibodies wil l bind larger amounts of antigen in a shorter period of time than low-affinity antibodies. Generally, high-affinity antibodies perform better in immunochemical techniques. This is due to not only their higher capacity but also to the stability of the complex. For example, the half-time for dissociation of an antibody binding to a small protein antigen with high affinity is 30 minutes or more, while for a low-affinity antibody this time may be a few minutes or less (Harlow and Lane, 1988). Like all equilibrium reactions, the affinity constant for antibody-antigen interaction is affected by temperature, pH, and solvent. Changes in these factors may increase or decrease the number of antibody-antigen complexes found at equilibrium. These changes will affect the affinity constant, either driving the reaction toward complete binding or releasing bound antigen. It is known that antibodies are able to bind to not only their specific antigens through its Fab region, but also to some specific proteins such as protein A through its Fc region. They can also bind with non-specific supports such as plastic vessels and nitrocellulose filters. A l l these properties explain why antibodies can be a very useful, powerful and highly selective reagent in immunochemical techniques. 1.2.1.2. Immunoassays and their advantages Since immunologists demonstrated more than 35 years ago that antibodies could be generated and selectively bound to low molecular weight compounds (Farah et al, 1960), 19 biological and medical sciences have taken advantage of antibody specificity as a means to quantify hormones, pharmaceuticals and allergens successfully. This has also lead to a gradual acceptance of this technique in analytical chemistry, especially for use in environmental analysis. During the past fifteen years, antibodies have been raised against a variety of chemicals of environmental significance. These include pesticide residues and their metabolites, industrial chemicals and microbial toxins. Immunoassays have also been successfully applied in many fields where conventional chemical or instrumental methods are routinely used (Ramakrishna et al, 1990; Rittenburg et al, 1990; Sherry, 1992; Van Emon and Lopez-Avila, 1992; Van Emon and Gerlach, 1995a). Generally, immunoassays are classified according to the label, or signal generator, used in the detection system and they can be divided into radioimmunoassays (RIA), enzyme immunoassays (EIA) and fluorescent immunoassays (FIA) which use radioactive isotope, enzyme and fluorescent material as signal generator, respectively. Many others techniques have been developed for different assay purposes (Butt, 1984; Collins, 1985). While there are a variety of ways of performing an immunoassay, a competitive process is usually involved in the immunoassay of low molecular weight compounds. The unlabelled antigen (Ag) wil l compete with labelled antigen (Ag*) for binding sites on a limited quantity of antibody (Ab) as illustrated in the following equation (Berson and Yalow, 1968). Ag* + Ab + Ag <=> Ag*-Ab + Ag-Ab known limiting unknown amount amount amount where A g * is the antigen labelled with either radioisotope, enzyme or fluorescent material. Since the reaction follows the Law of Mass Action, the labelled antigen and the un-labelled antigen, analyte, wil l bind to the antibody in proportion to their relative concentrations in the reaction mixture. As a result, the amount of labelled antigen bound to the antibodies will 20 be inversely proportional to the amount of unlabelled antigen (analyte) in the sample. A standard curve, which is usually obtained by plotting the intensity of the signal versus amount of analyte, can be generated and used for the prediction of analyte concentration in the unknown samples (Van Emon et al, 1989). Earlier immunoassays used a radioisotope as the label. Since the introduction of a radioisotope to the hapten (analyte) is possible without structurally altering the antigen, the ideal competitive system with identical structure of labelled and free antigen can be realized. Although RIA is a sensitive and precise analytical method, it also has several disadvantages that create difficulties for its continued acceptance. These include difficulties in the preparation of the radiolabeled hapten or antibody, and the need for expensive equipment. Great care must also be taken when handling radioisotopes and their disposal is also a problem. RIAs are less desirable for environmental analyses because they are not field-portable and the half-life of commonly used 125j i s relatively short. Enzyme immunoassays (EIAs) are analogous to RIAs. However, instead of using a radioactive isotope as label, enzymes are used as a vehicle for signal production. The enzymes, which are usually horseradish peroxidase or alkaline phosphatase, can be used to label both antigen and antibodies, providing bases for many assay formats, Among the various EIA techniques, the enzyme-linked immunosorbent assay (ELISA) is one of the most widely used methods and various ELISA formats are available for detection and quantification of either antigens or antibodies (Voller, 1978; Voller et al, 1979). This procedure has been widely used in the immunoassay of toxic compounds in the environment (Sherry, 1992). Enzymes used as signal generators overcome all of the disadvantages of radioactive labels. For example, the use of EIA is not restricted to licensed laboratories; it can be performed with less expensive equipment; the enzyme-linked secondary antibodies from a number of species are commercially available; phase separation is easier; and, quantification only requires a colormetric recorder (plate reader). However, there is an additional need for the synthesis of enzyme-labeled hapten used in the direct ELISA and it is sometimes difficult to synthesize 21 enzyme-linked hapten with a particular hapten to enzyme ratio for maximal performance. Another disadvantage is that the enzyme and sample are incubated together, which may lead to problems, such as enzyme inactivation. Since immunoassays are based on the antibody-antigen interaction and the antibody only reacts with its specific antigen, or hapten (analyte), specific detection can be achieved. Due to the high specificity of the antibodies, they can be used to extract the analytes of interest from the sample matrix specifically, thus potentially improving the recovery of compound of interest. Less clean-up may be needed because of the specificity of the antibody-based determination, resulting in a considerable time saving. Chromatographic resolution may often be by-passed for the same reason and dozens of individual samples can be processed simultaneously with very simple equipment, increasing throughput per unit of time. Since an immunoassay is a specific, sensitive and quantitative method which is easy to perform, it is especially useful in the screening of large numbers of samples at the same time. By avoiding both false positive (leading to unnecessary repeat analysis by a more time-intensive method) and false negative (failure to detect a violative sample) results, a considerable amount of work can be saved. Another important point in favour of immunoassays is their cost-effectiveness. For example, the approximate cost per sample for the GC analysis of molinate, an active ingredient of a rice herbicide, was $50 for in-house work done by the researchers and $130-$200 for an outside contractor. This cost estimate was based on analyzing 10 samples per day with no replicate analysis. In contrast, the ELISA costs were estimated at $5-$6 per sample (in house work), based on 40 samples per day, three dilution per sample, and four replicate wells per dilution (Harrison et al, 1989). It has been reported that a fully automated ELISA system used to screen for plant pathogens is capable of processing 10000-15000 samples per day (Van Vuurde et al, 1988). It should be mentioned that some field-portable immunoassay kits are now commercially available. These assay kits enable rapid determination of some important target compounds at hazardous waste sites even by personnel unfamiliar with analytical chemistry methodologies (Kaufman and Clower, 1991; Van Emon and Gerlach, 1995a). 22 Immunoassays are now becoming very popular in the field of residual analysis (Van Emon and Gerlach, 1995a, 1995b; Kaufman and Clower, 1995; Linde and Goh, 1995). Immunochemical methods have also been integrated with other analytical techniques such as H P L C and L C (Lucas et al, 1995; Frutos and Regnier, 1993). The usefulness of immunoassays in environmental analysis is also apparent by the growing number of papers published during the last decade. It is also apparent that more and more companies are marketing immunoassay kits for environmental and food residue analysis (Van Emon and Gerlach, 1995a). As immunoassays employ antibodies as the major reagents, a considerable amount of time is required for the production of antibodies. General procedures for the development of an immunoassay are listed in Table 1.3 (Van Emon and Gerlack, 1995a). Table 1.3 General procedures for the development of an immunoassay Purpose Procedure Identify target analyte Refer to regulations for probable contamination on site. Develop selective antibodies Design haptens and synthesize antigens. Select immunization regime. Immunoassay format development Optimize reagent concentrations, incubation times, etc., and develop standard curve. Optimize analytical detection Check cross-reactivity, matrix effects, etc. Evaluate method usefulness Compare with traditional technology for appropriate set of samples Validate method Using data analysis procedures, quantify the method performance 23 1.2.2. Preparation of antigens Immunoassays require a specific antibody against the compound of interest. In order to produce such an antibody, a suitable antigen is indispensable. However, as mentioned previously, low molecular weight compounds, such as those toxic chemicals present in the environment, are not immunogenic and it is necessary to chemically link them to carrier proteins to form hapten-protein conjugates. The need for such a hapten-protein conjugate in the elicitation of an immune response is due to the fact that this complex has to be recognized by at least two types of cells. The hapten is recognized by B-lymphocytes and carrier protein by T-lymphocytes. This cooperative recognition is required for the majority of antigens. Therefore, from a chemical point of view, selection of an appropriate hapten and subsequent syntheses of hapten-protein conjugate (antigens) are the key steps in the development of an effective immunoassay. A number of factors need to be considered during the synthesis of antigens. These include the availability of the hapten, the choice of carrier molecule and spacer arm, the coupling method used, and the site used for coupling. A l l these factors wil l profoundly affect the production of useful antibodies and the ultimate sensitivity and specificity of the immunochemical assay using these antibodies. 1.2.2.1. Choice of haptens The practical criterion in hapten selection is the ease of synthesis. Structurally, the hapten must be closely related to the analyte so that the resulting antibodies will recognize the analyte later on. The functional groups used to conjugate with carrier proteins are usually -OH, -COOH, -SH, -CO, -CHO or - N H 2 . Fortunately, these groups are present in most natural products and synthetic chemicals. The 'perfect' hapten should contain as much of the complete structure of the target compound as possible, and plus a handle, or spacer arm, in some cases. This will facilitate 24 the production of the desired antibody since the chances are that the small hapten is masked by the bulky protein if the hapten-protein conjugate is reduced. If no satisfactory reactive groups are present in the hapten molecule, suitable functional groups, such as those listed above, must be introduced into the hapten molecule by chemical modification to facilitate the coupling. 1.2.2.2. Choice of proteins Various proteins have been used as carriers for coupling with a hapten. As the majority of the methods used for coupling involve a reaction between the activated carboxylic group in the hapten and the amine groups of the protein, the availability of a lysine group is important. The two frequently used proteins are bovine serum albumin (BSA) and keyhole limpet hemocyanin (KLH) (Harlow and Lane, 1988). It is not known to what extent the carrier influences the anti-hapten response. However, some proteins such as K L H are extremely immunogenic and may enhance the anti-hapten response as a result of the general stimulation of the immune system (Rittenberg and Amkraut, 1966). Due to its large size, K L H is more likely to precipitate during cross-linking, making handling K L H difficult. Serum albumin, on the other hand, has the advantage of being more soluble than K L H . Other proteins that can be used are ovalbumin, mouse serum album (MSA), or rabbit serum albumin (RSA). M S A or R S A may be used when the antibody response to the carrier molecule must be kept to a minimum. Regardless of the protein used, the same general functional groups are available for conjugation to the hapten. These include the amino groups of the N-terminal and lysine residues, the carboxylic groups of the C-terminal and aspartic and glutamic residues, the imidazo and phenolic functions of histidine and the tyrosine residues respectively, and sulfhydryl groups of cystein residues (Erlanger, 1980). 25 1.2.2.3. Choice of spacer arms It has long been known that antibody specificity is directed to sites on the hapten distal to the point of conjugation as predicted by Landsteiner (1945). This is because regions of the molecule used for coupling are most likely occluded due to their proximity to the polypeptide chain. Therefore, it is important to carefully examine the structural features of the molecule and make the desirable part of the molecule (determinant) unhindered or not masked by the larger protein structure. In order to achieve this, the coupling functionality is usually placed at the end of an alkyl chain to offset the hapten from the carrier protein. This spatial separation of the small hapten from the larger carrier provides a better exposure of the hapten for the immune response. The "visibility" of the hapten depends largely on the mode of conjugation. A small molecule that is directly covalently linked to some sites on a protein may suffer from a considerable amount of masking in the region of the hapten nearest to the site of linkage. Antisera raised against such a molecule would suffer from a lack of specificity for those determinant groups that have been masked. In order to raise antibodies against the hapten, it is essential that most of the hapten molecule be visible and accessible to generate an immune response (Erlanger, 1980; Tijssen, 1985). A common practice is to insert a spacer arm between the hapten and the protein. There are four major factors to be considered in spacer arm selection. (1) the position of spacer arm; (2) the length of the spacer arm; (3) the polarity of the spacer arm and (4) the functional group variations for ease of attachment to protein carriers and enhanced hapten density (Jung et al, 1989). The structure and length of spacer arm should be carefully chosen to reduce spacer recognition, while retaining specificity for the target molecule. If possible, functional groups in the spacer arm should be avoided to minimize spacer recognition. Generally, alkyl spacers are preferable to heteroatom spacers. The point of attachment between the hapten and the carrier can greatly influence the selectivity of the resulting antibodies. If, for example, it is intended to assay for a compound when an hydroxylated product also occurs in the tissue, this group should then 26 be distal to the point of conjugation. Conversely, if it is desirable to assay both the compound and its hydroxylated product, conjugating through the hydroxyl group wil l destroy the difference, generating an antiserum capable of recognizing both the molecular species to the same degree. Thus, by the careful choice of different linkage moieties it is possible to generate antiserum of different selectivities when used in the immunoassays (Robins et al, 1985). The importance of the spacer arm and the position of linkage can be shown in the following example in the production of antibodies specific to triazine herbicides (Figure 1.3). Carboxylic acid derivatives of atrazine (R = Et) and simazine (R = i-Pr) at two positions were synthesized and conjugated to carrier proteins. It was found that chloro and alkylamino groups were each important determinants of specificity. Replacement of the chlorine of atrazine with mercaptopropanoic acid produced haptens which resulted in class-specific antibodies. Haptens with hexanoic spacers at the 4- or 6-(alkylamino) positions have produced more compound-specific antibodies (Jung et al, 1989). NH(CH2)nCOOH Atrazine R = Et n = 1, 2, 3, 4 or 5 Simazine R = i-Pr Figure 1.3 Structures of triazine haptens used for antigen preparations (Jung et al, 1989) 27 1.2.2.4. General methods for the preparation of antigens Depending on the functional groups present in the hapten molecule and the stability and solubility properties of the hapten, a number of methods have been successfully used for conjugation between the hapten and the carrier protein. Most coupling methods rely on the presence of free amino (-NH2), sulfhydryl (-SH), phenolic (-OH), or carboxylic (-COOH) groups on the hapten molecule or protein. In practice, coupling of a carboxylic acid or an alcohol to protein is the most common approach. Usually a two-stage synthesis is applied. In the first step, a bifunctional reagent is used to introduce a spacer arm. In the second step, a derivative containing a free carboxylic acid generated in the first step is activated and coupled to the protein. Usually the hapten can be activated in a dry, water-miscible organic solvent such as dimethylformide (DMF) and dioxane. Aliquots of this active hapten are added to three or four different proteins in at least two hapten protein ratios. The products, with different hapten densities, can be used to immunize animals and to coat the microtitre plate in the immunoassay. Although not strictly necessary, it is a common practice to dialyse the reaction mixture or pass the crude conjugate through a column of Sephadex G-25 prior to final lyophilisation and storage. This insures that unreacted hapten, which is usually used in excess and still left after dialysis, has been totally removed from the final product. It is also important to ascertain the molar ratio of the hapten to protein in the conjugate. Both very low and very high densities of hapten should be avoided for optimal antisera quality. There are about 60 lysine terminals in B S A although not all of these are available for condensation with the hapten. For BSA, ratios of between 10:1 and 25:1 (hapten:protein) should be attained. While it is frequently possible to obtain an estimate of the molar ratio by spectroscopy, usually by U V or fluorescence, in some instances it may be necessary to disrupt the conjugate chemically and assay the free hapten by chromatographic techniques. Although tedious, this ensures that later work, which usually involves several months labour, is not wasted by injecting an inferior antigen, which may result in a very poor response. 28 Two widely used methods for coupling are the mixed anhydride and carbodiimide methods. The mixed anhydride method was originally developed for peptide synthesis (Vaughan and Osato, 1951, 1952). This direct coupling procedure is now routinely used for conjugation of carboxyl-containing haptens to proteins (Erlanger, 1980). In this method, the carboxylic group is converted to a mixed anhydride by reaction with t-butyl chloroformate in the presence of tributylamine. The mixed anhydride, which is much more active than the acid itself and does not need to be isolated from the reaction mixture, then reacts with the amino groups in the protein to form an amide (Figure 1.4). O ° C l - C - O - B u t ° ° H 2 N-protein ° R _ C — O H - R - C - O — C—O-Bu* R - C - N H — protein t B u 3N Figure 1.4 Conjugation o f carboxylic acid to protein using mixed anhydride method The carbodiimide method is another common technique for direct coupling of carboxylic acid to protein (Kurzer and Douraghi-Zadeh, 1967). One advantage of this method is that the reaction can be carried out in either aqueous media or anhydrous conditions. In this method (Figure 1.5), the amide between the haptenic acid and the amino group in the protein is formed in the presence of a carbodiimide which acts as dehydrating agent. The two most commonly used carbodiimides in the synthesis of hapten-protein conjugates are l-(3-dimethylaminopropyl)-3-ethyl-carbodiimide (DEC) and dicyclohexylcarbodiimide (DCC), the former is water-soluble and the later can be used in organic solvents such as D M F or DMSO when the hapten is not water-soluble. Carbodiimides react with the side chains of aspartic and glutamic acid as well as the carboxyl terminal groups. 29 o II H 2 N — p r o t e i n O II O || R - C — O H R - C - N H — p r o t e i n R - H N — C — N H - R ' R - N = C = N - R Figure 1.5 Conjugation of carboxylic acid to protein using carbodiimide method Compounds with hydroxyl groups usually cannot be conjugated to a protein directly. However, these compounds can react with succinic anhydride to form hemisuccinate. The carboxylic group in the hemisuccinate can then be conjugated to protein through the mixed anhydride method or carbodiimide method as described above. Other chemicals, such as glutaric anhydride, sebacoyl dichloride and the trans- 1,4-cyclohexan-dicarbonyldichloride, can also be used for the same purpose, resulting in the introduction of spacer arms with different lengths. There are several choices for a hapten with an amine group. In a similar procedure to compounds with hydroxyl groups, both succinic anhydride and glutaric anhydride can be used to introduce a spacer arm, then either the mixed anhydride method or carbodiimide method can be applied to link the hapten to protein (Figure 1.6). O =o R N H 2 — - R - N H - C O C H 2 C H 2 C O O H O II CI - C — O - B u * H 2 N - — prote in - R - N H — C O C H 2 C H 2 C O — p r o t e i n Figure 1.6 Synthesis of antigen from amines using succinic anhydride method 30 Direct conjugation of a compound containing an amine group to protein is also possible and can be achieved through the Mannich reaction (Teal et al, 1977). A n amine group containing compound can also react with acrylic acid to form a propionic acid derivative and then conjugated with protein through the carboxylic group (Weiler et al, 1981). Glutaraldehyde is another bifunctional coupling agent that links two compounds primarily through their amino groups (Figure 1.7). Glutaraldehyde cross-linking is one of the most stable linkages used. However, the glutaraldehyde bridge will often form a portion of an epitope recognized by the immunized animal. R - N H 2 + O H C — C H 2 C H 2 C H 2 — C H O + H2N—protein \ R-N=CHCH 2 CH 2 CH 2 CH=N—protein Figure 1.7 Synthesis of antigen from amines using glutaraldehyde method 1.2.3. Production of polyclonal antibodies Conventional antibody production requires several steps including injection of the immunogen into animals, withdrawal of blood to test the antibody levels, and finally exsanguination of the animals for collection of antisera. As mentioned previously, antibodies are produced in response to an antigen by a very complex mechanism. Upon injection of an antigen into an experimental animal, the lymphocytes in the animal wil l be stimulated and turned into proliferating B-lymphocytes. These B-lymphocytes subsequently mature into antibody-secreting plasma cells. Successful antibody production strongly depends on several factors such as the 31 animals' immune system, the characteristics of the immunogen, and the immunization protocols (Harlow and Lane, 1988). Conventional immunization leads to the production of polyclonal antibodies, which are a heterogeneous antibody population directed to one antigen (Erlanger, 1980). This is because numerous B-lymphocyte clones secrete various antibodies against different epitopes on the antigen molecule. Among the wide range of antibodies produced, only some of them are capable of recognizing the small hapten and binding to it to differing degrees. It is these kind of antibodies that are useful in the development of immunoassays for haptens. Figure 1.8 shows this diversity in the production of antibodies against dioxin (Vanderlaan et al, 1986). When the animal is immunized with a dioxin-protein conjugate, three kinds of antibodies could be produced by the lymphocytes, for example, cell A, B and C. Each lymphocyte responds to a particular determinant on the antigen molecule and produces identical antibodies with a unique combining site. The serum will contain all of these antibodies, and hence, is termed polyclonal antiserum. Generally, most of the antibodies produced will be similar to those represented by antibodies A and B, which react with protein alone or with some complex of the protein, hapten and linkage groups. Only a few of the antibodies, represented by antibody C, are capable of binding a specific hapten with high affinity, without cross-reacting with the protein or the linker. It should be noted that there will be many other hapten-binding antibodies that are produced by different lymphocytes. These antibodies differ from each other in the fine details of their respective combining sites. 32 Figure 1.8 Illustration of the immune response to dioxin hapten-protein conjugate (Modified from Vanderlaan et al, 1986) It is important to note that once produced, the antibodies of interest wil l recognize the hapten no matter if it is conjugated to the protein or not. The antibodies can be easily isolated from blood after the red and white cells have been removed and the antiserum thus obtained can usually be used directly. However, since polyclonal antibodies are heterogeneous with respect to binding affinity and analyte recognition, it is also a common practice to partially purify the antibody from antiserum prior to use, in order to remove serum protein, interfering antibodies or to concentrate specific ones. To achieve this objective, a variety of techniques can be used such as affinity chromatography, ammonium sulfate fractionation and ion-exchange chromatography 33 (Dunbar and Schwoebel, 1990). These steps can be highly efficient but are nearly always accompanied by substantial losses of the desired antibody. It has been shown that individual serum collections from the same animal species and from different animals will show considerable variations in titer and specificity. It is rarely possible.to reproduce antisera with the exactly the same titer and specificity, even in genetically identical animals. Accordingly, the specificity of polyclonal antisera will vary over time and between animals. The production of polyclonal antibodies is also restricted by several other factors, such as the quantity of serum that can be obtained at one time, the level of sustained response to repeated immunization, and the life span of the animal. 1.2.4. The enzyme-linked immunosorbent assay (ELISA) The enzyme-linked immunosorbent assay (ELISA) has been the most widely used EIA method due to its simple detection system, inexpensive instrumentation and versatility (Voller et al, 1979). It is considered to be the method of choice for immunoassays of low molecular weight chemicals of environmental importance (Morgan et al, 1983). Different ELISA formats have been developed for various purposes and all formats share three components: specific antibody, conjugated hapten and free hapten (target analyte). While most of the formats can be applied to the detection and quantification of protein antigens, some of them, such as double antibody sandwich ELISA, can not be applied to the hapten quantification due to the small size of the hapten molecule. Hapten detection is only possible in a competitive assay format measuring the inhibition of binding of a labelled species in the presence of hapten, or the analyte. The basis for measurement is the competition between the free hapten and the conjugated hapten for binding to the limited amount of specific antibody. For a successful ELISA, a number of inter-related parameters need to be varied to define a satisfactory system. It should be realized that the influence of pH, ionic strength, temperature, organic solvents and other non-specific parameters varies from assay to assay. It is difficult to predict the optimum 34 conditions for the hapten-antibody interaction since the forces holding these complexes together are quite heterogeneous. Use of suitable enzyme and substrate in ELISA is also crucial. A variety of enzymes and substrates are commercially available. Alkaline phosphatase (AP) and horse radish peroxidase (HRP) are the two commonly used enzymes because of their low costs, availability in pure form, stability in a variety of coupling procedures, suitable specific activities and the ease of their colorimetric assays. Different substrates are also available for a given enzyme, such as ortho-phenyldiamine (OPD), 2, 2'-azino-bis (3-ethylbenzthiazoline)-6-sulphonic acid (ABSA), and tetramethylbenzyline (TMB) for HRP, and slightly different sensitivity may be encountered. 1.2.4.1. Assay protocols It is evident that in order to maximize precision, assay protocols should be designed in such a way that a minimum number of pipetting steps and minimum sample manipulation are involved. Depending on whether the antibody or antigen is immobilized on the plastic surface, various assay formats can be designed. The direct and indirect competitive ELISAs are two basic formats which have been widely used (Figure 1.9). Other formats, such as labelled avidin-biotin (LAB) assay, bridged avidin-biotin (BRAB) assay and the sandwich assay are all based on these two basic formats (Harlow and Lane, 1988). In the indirect ELISA, the antigen is immobilized onto the surface of microtitre plate first. After the unbound sites on the microtitre plate are blocked, the solution of diluted primary antibodies and the sample are added. During this stage, the coating antigen and the free antigen (analyte) in the sample will compete for a limited amount of antibodies, to form A b - A g c o a t m g or A b - A g f r e e . The latter will be washed away. In order to quantify the amount of Ab-A.§coating> an enzyme-linked secondary antibody is further added to bind with the primary antibodies. Since the secondary antibody has been labelled with enzyme, upon addition of the substrate, colour will develop. The intensity of the colour can be measured by a 35 spectrophotometer (ELISA plate reader) at suitable wavelength after the reaction is stopped by sulfuric acid. Obviously, the colour intensity will be inversely proportional to the concentration of analyte in the sample. The more the analyte in the sample, the less the amount of antibodies bound to coating antigen, and the lower the colour intensity. A successful ELISA should be developed with characteristics such as low detection limit, high sensitivity, large quantitative window and low spectrum of antibody recognition or cross-reactivity. Before these characters can be evaluated, it is important to determine the amount of coating antigen and the dilution of primary antibodies which should be used in the ELISA procedures. This can usually be done by a two dimensional checkerboard titration, with different concentrations of coating antigen in Step 1 and different dilutions of primary antibodies in Step 2, without the presence of standard or analyte (Figure 1.9). Usually, a stock solution of standard in organic solvents, such as methanol, D M F or DMSO, is prepared. Poor solubility of the standard in the assay buffer may sometimes be encountered. In this case, organic solvents can be used to increase the solubility of the compounds in the buffer. Antibodies can often tolerate high concentrations of organic co-solvents such as ethanol, dimethylsulfoxide, acetonitrile, tetrahydrofuran, dioxane, methanol and propylene glycol. These solvents can enhance the solubility of the analyte, remove the analyte from surfaces, disrupt lipid micelles in the sample matrix, and increase ease of sample handling. The properties of antibodies vary widely and a concentration of organic solvent that has no effect on one immunoassay can dramatically decrease or even increase the sensitivity of other assays (Gee, 1988). In order to increase the assay sensitivity, a biotin-avidin system is frequently used. In this assay format, the incubation step involving the use of enzyme conjugated secondary antibodies (step 3) was replaced by two consecutive incubation steps: incubation with a biotin-conjugated secondary antibody followed by another incubation with an avidin-enzyme conjugate. The advantage of using a biotin-avidin system is that the avidin may be labelled with enzyme, fluorochromes, radiolabels, ferritin, etc. so that the same biotinylated antibodies may be used 36 o / f E f A A s — - p s—«• p A p A ' i O E O _ E 6 O S — » • P S—*- P polystyrene surface X Enzyme-labeled antibody Q Hapten-carrier Q Substrate for protein conjugate O enzyme P Coloured product Analyte Antibody A Enzyme-linked hapten Figure 1.9 Major steps involved in an indirect (step 1-4) and direct ELISA (step 5-7) 37 with the different avidin conjugates in various immunoassay formats, without the need for preparing separate labelled antibodies. The signal amplification was achieved by the extremely high affinity between biotin and avidin. Streptavidin from Streptomyces avidinii can also be used for the same purpose. In direct ELISAs, the primary antibodies are immobilized on the microtitre plate and a hapten-enzyme conjugate is required in the incubation step. Similarly, the free hapten (analyte) and hapten-enzyme conjugate will compete for a limited amount of antibodies. Again, the intensity of the colour is inversely proportional to the concentration of analyte in the sample. Compared with indirect ELISA, less steps are needed for the assay. However, the hapten-enzyme conjugate must be synthesized chemically. 1.2.4.2. Cross-reactivity Since the antibodies are produced against the chemical entity (epitope) in the antigen molecule, it is a common phenomenon that antibodies recognizing a specific epitope may also show cross-reactivity with other structurally related compounds. This is especially true for polyclonal antibodies since they contain various antibodies capable of recognizing different epitopes on the target molecule. Therefore, cross-reactivity becomes one of the special concerns for immunoassays. This is because closely related compounds, for instance, residual pesticide and its metabolites, may also be present in the sample matrix and the antibody wil l recognize both the analyte and those related structures, thus over-estimating the analyte of interest or causing false positive results. Generally, interferences may occur due to the presence of these compounds in the sample matrix. However, they can usually be avoided by selecting a suitable sample preparation procedure. It should also be mentioned that, although cross-reactivity is usually considered a disadvantage, sometimes, cross-reactivity may provide a method to quantify a group of chemicals (Hock, 1993). Multianalytes can also be quantified by using several antibodies 38 showing cross-reactivity. The major disadvantage of an immunoassay, on the other hand, is that an assay needs to be developed for each compound of interest. Such development involves at least several months work in synthesis of antigens and production of antibodies and requires specific facilities, particularly for the experimental animals used to raise the initial antibodies. 1.2.4.3. Matrix effects and assay validation When a standard curve is obtained, the assay of real samples can be performed. It is common practice in immunoassay to run several dilutions of the sample to determine sample matrix effects, which are caused by substances other than the analyte present in the sample or sample extract. In the ideal situation, linear results over a series of dilutions for sample assayed should be achieved for a positive sample, spiked negative sample and a standard in assay buffer. In the historical use of clinical immunoassays, samples are simple, well defined and with consistent matrices such as urine, blood and serum. However, the immunoassay developed for environmental contaminants must deal with sample matrices which are more complex, less well defined and more variable, such as soil, sediments and sludge. Therefore, components other than analytes in the sample matrix are very likely to interfere with the assay. Generally, the sensitivity of the antibody can be exploited by diluting the sample to minimize interferences and still achieve the desired detection limit (Gee et al, 1994). In order to determine sample matrix effects, several dilutions of the sample should be made and assayed. If the resulting sample dilution curve is parallel to the standard curve, sample matrix effects can be considered minimal. In such a case, sample clean-up may be omitted and analysis can be carried out directly. For example, quantitative analysis of many pesticides, such as molinate, parathion, chlorsulfuron, paraquat and maleic hydrazide, in simple matrices, such as ground and drinking water, juice and urine, have been carried out with little or no sample clean-up and matrix interference has not been a problem (Harrison et al, 1988). However, for more complex sample matrices, such as food products and soils, pretreatment such as extraction is sometimes required 39 prior to the immunoassay. For example, various extraction methods have been compared for the immunoassay of PCBs (Johnson and Van Emon, 1994). Sometimes it has been neccessary to generate the standard curves in a negative sample matrix rather than the assay buffer, although this usually entails some loss of assay sensitivity. This method has also been used in some cases to reduce the background effect. For example, pooled human control urine was used to generate the standard curve for the ELISA of alachlor residue in monkey urine (Feng et al, 1994). Similarly, drug-free equine urine was used to generate a standard curve for the detection of buprenorphine in post-race horse urine (Debrabandere et al, 1993). It has also been reported that, when ELISA was used to quantify carbaryl, which is an insecticide often found in water, soil extract, urine and honey, standard curves generated in the presence of matrices were parallel to that in buffer. However, different sensitivities were obtained (Marco et al, 1993a). Therefore, it seems that a general approach of resolving the matrix problem would be to run the standard curve in the presence of components that mimic the behaviour of the matrix (Marco et al, 1993b). It should be realized that sample workup necessary for immunoassay wil l vary with each combination of analyte and antibody, each of which may have a different tolerance for the matrix and other factors. Since quantitative analysis is a statistical rather than an absolute phenomenon, analytical recovery of spiked samples should show little variation, and inter- and intra-assay variations should be checked. Both random and non-random errors should be as small as possible. Random error determines the precision which is defined as the concordance between a series of measurements of the same sample. Non-random error determines the accuracy of the analysis which is the degree of approximation of the "true" value by the analysis. While random error is largely determined by factors which also operate in other analytical systems (e.g. operator skill, precision of pipetting devices, etc.), attention should be given to non-random errors which affect accuracy as these factors tend to be more assay specific and may not be obvious. The following points are some of the major factors which contribute to non-random error in immunoassays (Weiler, 1990): 40 1. Presence of cross-reactants 2. Presence of factors destabilizing or desorbing antibodies (e.g. surfactants, phenols) 3. Presence of factors influencing the antigen-antibody interaction 4. Antigen binding by other assay components (surface of vessels, carrier proteins such as albumins) 5. Presence of factors influencing the activity of the marker enzymes (e.g. inhibitors) in ELISA, sample autofluorescence (in FIA), radioactivity quenchers (in RIA) Therefore, the matrix effect is one of the most serious and most common sources of error in an immunoassay. A very important goal of validation is to determine if there is a matrix effect and then to eliminate or account for this interference. Whenever possible, immunoassay results should be correlated with those obtained by independent analytical methods, such as H P L C and GC. Among the many comparison parameters, the most obvious ones are cost, analytical speed and time, ease of performance, accuracy, precision, reproducibility, and detection limit using statistical support. If the immunoassay is used in the screening of samples, positive samples should be independently confirmed if a high level of confidence is to be maintained. Confirmation of all samples may be unnecessary in some monitoring applications if the assay has been carefully validated. 1.3. Proposed Work and Objectives As both the concern about toxic chemicals in the environment and the cost of the analyses for these environmental pollutants becomes greater, the development of inexpensive and accurate analytical methods will become increasingly desirable. Resin acids are among the major compounds that have been shown to be responsible for the fish toxicity of some pulp mill effluents. Although attempts have been made to develop effective methods for the detection and 41 quantification of resin acids (Taylor et al, 1988; Zinkel and Han, 1986), a standard method for resin acid analysis is still not available. For example, different laboratories using different analytical protocols have generated results with large variations (Bicho et al, 1995a). Due to the drawbacks of the currently used methods for detecting and quantifying resin acids, we have tried to develop an alternative method which could overcome some of the disadvantages of traditional instrument-based analytical methods. A n immunochemical method has been considered desirable in the quantification of toxic chemicals present in the environment due to its rapidity, ease of use, low cost and the ability to process large numbers of samples. Therefore, the overall objective of this project was to raise polyclonal antibodies specific to resin acids and to use these antibodies as immunological probes to quantify the resin acid content in various pulp mill effluents and process waters. An overview of the major work that was performed in this project is described in Figure 1.10. Different antigens were synthesized by coupling haptens to carrier proteins and subsequently used for the immunization of rabbits. The polyclonal antibodies obtained were then evaluated for their ability to recognize resin acids. The successful antibodies were selected to develop an enzyme-linked immunosorbent assay (ELISA). The method was subsequently validated by comparison with GC and the developed ELISA was applied to various samples, such as effluent samples from CTMP and B K M E and samples from fish bile. 42 Synthesis of antigens (With and without spacer arms) Polyclonal antibodies (Specific or generic antibodies) E L I S A (Direct and Indirect ELISA) Assay validation (ELISA vs GC) J Application to real samples (Effluents from C T M P and and B K M E , fish bile...) Figure 1.10 A n overview of the proj ect 43 CHAPTER 2 DEVELOPMENT OF A GC PROTOCOL FOR RESIN ACID ANALYSIS 2.1; Introduction Prior to embarking on a program of developing an immunoassay for resin acids we first had to establish a "standard" method to which it would be compared. Although various H P L C , GC and GC-MS methods have been developed to quantify resin acids (Judd et al, 1995; Lee and Peart, 1992; Morales et al, 1992; Stuthridge et al, 1992; Tavendale et al, 1995; Taylor et al, 1988; Voss and Rapsomatiotis, 1985), the first step in analysis by all of these procedures is the extraction of the resin acids from the pulp mill effluents. This can be done by either liquid-liquid extraction (NCASI, 1986) or solid phase extraction (Leach and Thakore, 1973; Richardson et al, 1992). Once isolated, resin acids are usually derivatized and then quantified by either H P L C (Richardson et al, 1992) or GC (Foster and Zinkel, 1982; Zinkel and Engler, 1977). Capillary GC is generally the method of choice due to its high sensitivity and resolution. Depending on the derivatization method used, detection can be achieved by either flame ionization (GC-FID), electron capture (GC-ECD) or mass spectrometry (GC-MS) (Lee et al, 1990; NCASI , 1986; Voss and Rapsomatiotis, 1985). This latter method can provide more sensitive detection as well as structural confirmation of the eight major resin acid structures. Liquid-liquid extraction is the most commonly used technique for the isolation of organic compounds from aqueous medium. However, it can be complicated by several factors such as emulsion formation, which retards the speed of extraction and recoveries, and co-extraction of hydrophobic material, which wil l interfere with subsequent chromatographic separation. Pulp mill effluents are complex matrices where small amounts of lipophilic resin acids co-exist with abundant lignin residues, humic acids and many other materials (Suntio et al, 1988). These materials probably interact with resin acids through various processes such as covalent 44 interaction, charge-transfer, van der Waals forces and hydrogen bonding (Morales et al, 1992). In addition, some resin acids are reported to be unstable and isomerization may occur, especially at low pH. Different pHs, ranging from 2 to 12, have been used in the liquid-liquid extraction of resin acids from pulp mill effluents ( Morales et al, 1992; NCASI, 1972, 1975, 1978 and 1986; Turner and Wallin, 1985; Voss and Rapsomatiotis, 1985; Richardson and Bloom, 1982; Wilkins and Panadam, 1987). Some of these researchers have advocated extraction procedures that use acidic media to ensure a minimal dissociation of resin acids. Alternatively, other workers have recommended procedures that use alkaline media to minimize isomerization. In addition, various organic solvents, such as diethyl ether, methylene chloride, chloroform and M T B E have also been evaluated (Voss and Rapsomatiotis, 1985). However, it has been found that the same extraction protocol may lead to different extraction efficiencies for different types of effluents. For example, Voss and Rapsomatiotis (1985) reported that highest extraction efficiencies were obtained using methyl t-butyl ether (MTBE) at pH 9 for bleached kraft mill effluent (BKME), while Orsa and Holmbom (1994) reported that similar extraction efficiencies were observed for effluents from thermomechanical pulp (TMP) over a pH range of 3.5 to 8. Alternatively, NCASI (1986) recommended extraction procedures using ethyl ether at a low pH. Due to the complex nature of the effluent composition, it is apparent that there is currently no universal extraction protocol suitable for all pulp mill effluents. A wide choice of isolation, derivatization and chromatographic options have resulted in a number of different approaches to overcome the problems associated with resin acid analysis. Although many papers have been published in this area, there has been no detailed attempt to evaluate the pH effect on the extraction efficiency of resin acids from various pulp mill effluents. In this initial work, we have investigated the effects of pH and solvents on the extraction of resin acids from various pulp mill process wastewater streams. A GC protocol including extraction, derivatization and sample clean-up using solid-phase extraction was established. We then carried out a preliminary 45 assessment on the possible influence of some pulp mill components such as high molecular weight material and humic material on the extraction of resin acids. 2.2. Materials and Methods Resin acid standards were obtained from Helix Biotech Corporation (Richmond, B.C., Canada). N-methyl-N-nitroso-p-toluene sulfonamide (Diazald®), tricosanoic acid (TCA) and humic acid (sodium salt) were purchased from Aldrich Chemical Company, Inc. (Milwaukee, Wis). Ethyl acetate (EtOAc) and M T B E were from Fisher (HPLC grade). Methyl heneicosanoate (MHS), O-methylpodocarpic acid (OMPA) and were obtained from Sigma Chemical Co. (St. Louis, MO). Bond Elut/Vac Elut system and the NH2 type solid-phase extraction cartridge (Bond Elut®) were obtained from Analytichem International (Harbor City, CA). A l l effluent samples (biotreated B K M E from M i l l A , untreated CTMP effluent from M i l l B and untreated B K M E from M i l l C) were obtained from three softwood pulp mills located in British Columbia, Canada (Table 2.1). Table 2.1 Description of pulp mill effluent sources M i l l A M i l l B MhTC Furnish Softwood spruce : pine : fir (35:55:10) spruce : pine : fir : cedar (30:32:18:6) Pulping method Kraft TMP/CTMP Kraft Production (adt/d) 1120 950 920 Bleach sequence D/C-EO/DED Peroxide D-EP/DED Effluent treatment Activated sludge Anaerobic/aerobic stabilizing basin Aerated lagoon 46 2.2.1 . E x t r a c t i o n s o f res in ac ids f r o m p u l p m i l l e f f luents Upon receipt, effluent samples were stored at 4°C with pH unadjusted and they were usually analyzed within two weeks. Effluent samples (200 ml for treated B K M E , 100 ml for untreated B K M E or 20 ml for untreated CTMP effluent) were spiked with surrogate (OMPA, 50.0 pi, 1.0 mg/ml in methanol), adjusted to a desired pH with either I M HC1 or I M NaOH, and extracted twice with an equal volume of either ethyl acetate or M T B E in a separation funnel. The organic phases from each extraction were combined and dried over anhydrous MgSC>4. The solvent was then removed under reduced pressure. A l l samples were extracted in duplicate and the data presented are mean values. 2.2.2. S a m p l e c l e a n - u p u s i n g S o l i d - P h a s e E x t r a c t i o n ( S P E ) A n NH2 type solid-phase extraction cartridge (Bond Elut®) was conditioned with hexane (2x2 mL) with a 5 kPa vacuum applied. The crude extracts were dissolved in 0.5 mL of chloroform and loaded onto the column. The neutral co-extracted organic compounds were eluted with 16 mL of solvent (hexane : diethyl ether, 1:8, v/v). The fatty and resin acids were eluted with 6 mL of acidic ether (acetic acid : diethyl ether, 3:97, v/v). The eluent solvent was removed by evaporation under a nitrogen stream (Chen et al, 1994). 2.2.3. D e r i v a t i z a t i o n o f res in ac ids The resin acid fraction was dissolved in 1.0 ml of diethyl ether and spiked with an internal standard (50 pL of MHS in methanol, 1.0 mg/ml) and a methylation standard (50 pL of T C A in methanol, 1.0 mg/ml). The mixture was derivatized with diazomethane, which was generated in-situ by reaction of N-methyl-N-nitroso-p-toluene sulfonamide (Diazald®) with alcoholic potassium hydroxide. It was delivered to the sample vial by a stream of nitrogen gas 47 until a persistent yellow colour was obtained. The ether was evaporated under nitrogen and the methyl esters of fatty and resin acids were redissolved in 1.0 mL of methanol. 2.2.4. GC and GC-MS Analysis of resin acids Individual resin acids were identified and quantified on a Hewlett Packard HP 5890 series II gas chromatograph (GC) equipped with an HP 7673 auto injector, flame ionization detectors and a dual column system (DB-5 and DB-17 fused silica capillary columns, 0.25-mm ID, 30-m long, 0.25 um thickness, from J & W Scientific). Helium and nitrogen were used as the carrier and make-up gases respectively. Injector and detector temperatures were 260°C and 290°C, respectively. The oven temperature was programmed at 60°C for 2 min, with an increase of temperature at 35°C/min to 170°C, then 0.6°C/min to 200°C and finally 35°C/min to 280°C. It was kept at this temperature for 10 minutes. The instrument detection limit was 1.0 ng. Analysis of standards indicated that the coefficient of variation was always below 8% and usually around 1-2%. 2.3. Results and Discussion 2.3.1. Comparison of extraction solvents From our initial investigations it was apparent that effective extraction of resin acids from pulp mill effluents was of primary importance in the accurate quantification of these compounds. In the past, various organic solvents with different polarities have been used, including diethyl ether, methylene chloride, chloroform and M T B E (Voss and Rapsomatiotis, 1985). Solvent mixtures, such as petroleum ether-acetone-methanol (McMahon, 1980), hexane-acetone-methanol (Wearing et al, 1984), and MTBE/dichloromethane (Morales et al, 1992) have also been suggested to either improve extraction efficiency or minimize emulsion formation. In theory, the efficiency of liquid-liquid extraction will be determined by the 48 solubilities of compounds of interests in the two immiscible solvents. As the solubility of resin acids in common organic solvents is much higher than in water, liquid-liquid extraction should be very effective. However, it is possible that differences in solvent polarities may lead to the co-extraction of material other than the compounds of interests. For example, it has been reported that, compared with diethyl ether, a petroleum ether/acetone/methanol (PAM) system can effectively extract resin and fatty acids from unbleached kraft mill process streams while most of the lignin material remains in the aqueous solution (McMahon, 1980). From the different solvents assessed in our initial evaluation of extraction efficiency, ethyl acetate and M T B E were selected for further detailed evaluation. Although ethyl acetate is a universal solvent, which is widely used in organic laboratories and successfully used in the extraction of pesticide residues from water (Chau and Afghan, 1982), it has not been evaluated in detail for the extraction of resin acids from pulp mill effluents. While several publications have favoured the use of M T B E (Voss and Rapsomatiotis, 1985), we found that ethyl acetate was as efficient as M T B E at recovering all of the resin acids (Figure 2.3, Table 2.3 and 2.4). As resin esters are unstable in both hydrocarbon solvents (Nestler and Zinkel, 1963) and halogenated solvents (Mayr et ai, 1982), we found that derivatized resin acid samples should be re-constituted in either ethyl acetate, M T B E or methanol, as soon as possible prior to analyses by GC. Samples re-constituted in this way were stable for at least three weeks, provided they were kept refrigerated. 2.3.2. Solid-Phase Extraction prior to G C and G C - M S analyses Since pulp mill effluents contain various organic components, solvent extraction usually extracts all of the lipophilic material which is present. Unfortunately, this co-extracted material may interfere with subsequent GC analysis. When large amounts of lipophilic compounds other than resin and fatty acids were co-extracted, it was shown that sample clean-up was required prior to GC analysis to avoid peak overlapping (Wilkins and Panadam 1987). Although sample 49 clean-up using silica gel chromatography after extraction of resin acids from pulp mill effluents has been successfully used (NCASI 1975), we adopted a solid-phase extraction (SPE) protocol (Chen et al. 1994) which was quicker and just as efficient. When a N H 2 type SPE cartridge was used, the carboxylic group of resin and fatty acids reacted with the amino groups of the solid phase'matrix to form an ammonium carboxylate ionic pair. The resin and fatty acids could be subsequently eluted by acidic ether. The resin and fatty acids could be recovered quantitatively and the background noise in the chromatogram was greatly reduced (Figure 2.1). Most of the past work on the identification and quantification of resin acids in pulp mill effluents has utilized GC with flame ionization detector (FID) (Taylor et al., 1988). However, FID is a non-specific detector and other compounds may co-elute with individual resin acids and be erroneously identified as resin acids. A dual column system involving two capillary columns connected to the same injection port was used in this study and a calibration curve for resin acid standards was established using the data from the DB-5 column, with the DB-17 serving as the reference column. In. our temperature program, levopimaric and palustric acids could not be resolved on the DB-5 column (Figure 2.2). However, their presence could be confirmed on the DB-17 column. As the polarity of the stationary phases of the DB-5 (with 5% phenyl-methylpolysiloxane) and DB-17 columns (with 50% phenyl-methylpolysiloxane) differed, a peak derived from a non-resin acid component that co-eluted with a resin acid on one column would not likely co-elute with the resin acid on the other column. Thus the use of both the DB-5 and DB-17 in tandem provided good confirmation of individual resin acid standards (Figure 2.2). Although good resolution could be achieved with the dual column GC system, GC-MS was routinely used to confirm the identities of the resin acids detected by GC analysis. Both the retention times of resin acids on the dual GC column (DB-5 and DB-17) and GC-MS system, and the ions monitored in GC-MS were determined (Table 2.2). Analysis of an extracted CTMP effluent by GC-MS confirmed that our GC protocol, using the dual column system, provided selective quantification of the resin acids. 50 A 6 0 0 0 5 0 0 0 4 0 0 0 3 0 0 0 6 0 0 0 5 0 0 0 4 0 0 0 H B 3 0 0 0 10 1 0 2 0 3 0 4 0 5 0 6 0 Min < u oo i—i Q u < o g 2 0 3 0 4 0 5 0 6 0 Min Figure 2.1 Gas chromatogram of extracts before and after sample clean-up using SPE 51 7000 r 6000 5000 4000 3000 30 8000 7000 6000 5000 30 Q < .CO CL O o < < Q Q — « < J i . < O Q_ 6 X CO < CQ < o UJ o \ u I 40 50 60 Retention Time on DB-5 (min) Q < CO O CO • L CO X o > U J I < < < x Q < < o o o UJ, o X o < X Q CN o _L 40 50 60 Retention Time on DB-17 (min) Figure 2.2 Gas chromatogram of standards run on DB-5 and DB-17 columns 52 Table 2.2 Retention times (min) of resin acids On different G C columns D B - 5 (min) DB-17 (min) G C - M S (min) Monitored ions (m/z) P I M 34.44 43.09 52.22 121, 180, 257, 316 S A N 35.52 44.75 54.07 121, 301, 316 ISO 39.07 50.26 56.19 241, 257, 316 P A L / L E V 40.14 52.55 56.38 241, 301, 316 D H A 42.56 57^02 57.18 239, 299, 314 A B A 45.92 57.02 58.00 241,256,316 N E O 50.94 58.00 58.53 121, 135, 316 M H S (IS) 52.38 48.63 58.50 143, 297, 340 2.3.3. Effects of p H on extraction efficiency A s mentioned previously, a wide range of pHs have been suggested by various researchers as means of enhancing extraction efficiency. We therefore decided to further investigate the effects of p H on the extraction efficiency of effluent samples from different mills using various pulping and effluent treatment technologies. Three p H values (2, 7 and 9) were compared and samples included distilled water which was spiked with resin acids, spiked biotreated B K M E , untreated C T M P effluent and untreated B K M E . A s a first step in the evaluation, distilled water was spiked with resin acids (250 pg/L of D H A , 250 ug/L of A B A and 300 ug/L of ISO in methanol) to a total concentration of 800 ug/L and then extracted with either ethyl acetate or M T B E at different pHs. G C analysis (in triplicates) indicated that lower pHs favored the extraction of resin acids in both cases (Figure 2.3). Since resin acids are weak organic acids, it was expected that a low p H would facilitate the extraction of these compounds from the aqueous media. A s the dissociation constants (pKa) for resin acids are around 7, at a higher p H , most of the resin acids should be ionized. .For 53 example, more than 95% of the D H A was ionized at pH 8 (Taylor et al. 1988). Our results confirmed that the unionized resin acids were easily extracted. This agreed with early reports which showed that the extraction efficiency of resin acids was inversely proportional to the pH of the aqueous medium. For example, NCASI (1986) found that higher recoveries were achieved at low pH when diethyl ether was used as the extraction solvent. Voss and Rapsomatiotis (1985) found a similar trend for spiked samples using M T B E as the solvent. We next wanted to evaluate the reproducibility of this method using real effluents. Initially, a series of B K M E grab samples were obtained from M i l l A at the outfall of the secondary biological treatment facility. These samples were chosen as earlier work had indicated that they contained low BOD and COD values and low concentrations of resin acids. No resin acids were detected when 200 ml of effluent was extracted and analyzed by GC. When this effluent sample was then spiked with resin acids (DHA, A B A and ISO in methanol) and kept at 4°C overnight, a similar resin acid profile to that observed with the spiked water was detected, with low pH favouring extraction (Figure 2.3). Recovery % 100.0 75.0 4 50.0 25.0 4-Figure 2.3 Recovery of resin acids for A), distilled water spiked with resin acids and extracted with ethyl acetate, B).distilled water spiked with resin acids and extracted with M T B E , C). biotreated B K M E spiked with resin acids and extracted with ethyl acetate, D). biotreated B K M E spiked with resin acids and extracted with M T B E 54 Although this initial work indicated that low p H favoured the extraction of resin acids from the spiked samples, it was possible that, in effluents containing higher concentrations of organic material ( C O D / B O D ) , the resin acids would not be readily extracted from this "matrix" of materials. A s effluents from C T M P mills using softwood furnish are known to contain substantially higher C O D and B O D values as wel l as higher resin acid content, an effluent from M i l l B was used as a representative sample. The samples were first centrifuged to remove fines and other solids prior to extraction. Due to the substantially higher resin acid content, only 20 m l of effluent was required for extraction. It was found that, regardless of the solvent used, emulsions occurred when effluents were extracted under alkaline conditions, especially when the ratio of solvent to effluent volume was low. This was probably due to the high resin and fatty acid concentration, with these acids behaving as surfactants at high p H and thereby facilitating emulsion formation. However, this problem was largely resolved by using large volumes of the organic solvent (equal or greater than the effluent volume). When the resin acid content of the untreated C T M P effluent was determined at different pHs and solvents (Table 2.3), the resin acids were again shown to be extracted more effectively at p H 2 rather than at p H 9, regardless of the solvent used. This is similar to the result reported by Ekman and Holmbom (1989) who also adopted low p H in their extraction protocol for resin acids in T M P effluents, when using diethyl ether as the extraction solvent. A s discussed earlier, in their neutral form, resin acids can be expected to be extracted more efficiently than their ionized forms and this may explain why extraction at low p H provided high recoveries. Previous studies by N C A S I (1986) indicated that the optimum resin acid extraction occurred after three extractions at p H 5 followed by two extractions at p H 2. This sequence was carried out to try to minimize the isomerization and emulsion problems. Turner and W a l l i n (1985) also used low p H (2.0) to extract resin acids in B K M E using methylene chloride. Our results using spiked water, spiked B K M E and untreated C T M P effluents 55 confirmed these earlier recommendations. However, these results differed from those of Voss and Rapsomatiotis (1985) who found that a recovery of 95% or greater was achieved in the extraction of B K M E under alkaline conditions (pH 8 to 10) using M T B E , while at p H 2 their recoveries were less than 70%. The difference between Voss's work and our work might be explained by the source of the effluents used. They used a B K M E which contained high concentrations of lignin and other high molecular weight material.-The lower recovery at the acidic p H could be due to some type of associations between resin acids and dissolved lignins or other high molecular weight materials present in the effluent. Recently, Orsa and Holmbom (1994) reported that, when M T B E was used to extract resin acids from T M P effluents, a comparable recovery "was achieved over a p H range of 3.5 to 8. W e obtained similar results with both M T B E and ethyl acetate, probably reflecting the similar nature of the effluent from T M P and C T M P mills. Table 2.3 Resin acid concentrations of untreated C T M P effluents determined by G C using different extraction solvents and p H E t O A c M T B E pH2 pH7 pH9 pH2 pH7 pH9 P L M 2.63 1.81 2.16 2.15 2.22 1.41 S A N D 0 0 0 0 0 0 ISO 4.48 4.05 3.86 4.09 2.95 3.09 P A L / L E V O 5.26 4.89 4.55 5.17 4.05 4.10 D H A 9.23 8.70 7.63 8.46 6.81 6.65 A B A 14.02 13.34 12.95 13.25 11.57 10.87 N E O 3.51 3.58 3.12 3.44 3.00 2.86 Total (mg/L) 39.13 36.37 34.21 36.56 30.60 28.98 56 To further assess the possible p H effect on the extraction of resin acids from B K M E , 20 litres of B K M E was collected from the inlet of the aeration basin of M i l l C. Resin acids were extracted using E t O A c and M T B E at different pHs and G C results indicated that, in this case, better resin acid recovery did occur when the extraction was performed at p H 9 (Table 2.4). In addition, it was found that less coloured material was extracted at p H 9 rather than p H 2. It is possible that the residual lignins and phenolics present in B K M E are ionized at high p H and are less extractable. It was also observed that a brown material, which was possibly lignins and other high molecular weight materials, accumulated at the liquid-liquid interface during extraction at low p H . Table 2.4 Resin acid concentrations of B K M E sample (from aeration inlet) determined by G C using different extraction solvents and p H E t O A c M T B E pH2 pH7 pH9 pH2 pH7 pH9 P I M 129 131 167 153 138 213 S A N D 135 102 151 215 160 186 ISO 171 159 176 109 153 175 P A L 72 96 148 80 122 135 D H A 268 298 333 283 307 337 A B A 166 179 219 103 147 244 N E O 65 47 75 64 55 51 Total (ug/L) 1006 1012 1269 1007 1082 1341 The difference in the p H used for extraction of resin acids from various types of effluent samples prompted us to investigate the effect of some probable effluent components on extraction efficiency. A s mentioned previously, pulp m i l l effluents are very complex and the presence of humic substances, carbohydrates and other dissolved organic materials can 57 complicate the extraction of resin acids. It is recognized that the presence of these types of materials can change the physio-chemical state of resin acids in aqueous systems by binding or association with them and that, generally, the solubility of hydrophobic organic compounds in water can be enhanced by such interactions (Maguire 1994). It is also recognized that these interactions w i l l vary depending on the nature of the dissolved organic materials that are present (Wang etal. 1995). Earlier work has shown that in bleached kraft m i l l effluents, the high molecular weight ( H M W ) material represented a significant fraction of the adsorbable organic halide and chemical oxygen demand (Morck et al. 1991). Bullock et al. (1995) in our research group had fractionated and characterized B K M E effluent to determine the relative effects that the H M W and low molecular weight ( L M W ) material had on the microbes present in waste treatment plants. A s it was possible that the H M W fraction might also interfere with resin acid analyses, we added different amounts of H M W material isolated from a B K M E by membrane filtration with 10,000 molecular weight cutoff, to distilled water spiked with D H A (350 ug/L) and ISO (450 ug/L). A s the H M W fraction had been stored as a concentrated solution (16x) it was diluted to a concentration similar to that of the original B K M E (6% v/v). The samples were extracted using ethyl acetate and analyzed by G C in triplicate. It was apparent that the addition of H M W material at p H 2 inhibited the extraction of resin acids and the recoveries decreased from 92% to 85%) when concentration of H M W material increased from 0 to 6% (Figure 2.4). The recovery was even lower at p H 9 when 6% H M W material was present and only 75% resin acids were recovered. However, this result is better than 65% recovery obtained when resin acids were extracted at p H 9 without the presence of H M W material (Figure 2.3, A ) . Therefore, there could be some kind of interaction among the solvent, D H A and H M W material, which might affect the extraction of resin acids. Although the H M W material exerted some effects on resin acid extraction, we thought it more probable that humic type material, which can be detected in most effluents, was more likely to interfere with resin acid extraction. Humic materials generally result from the 58 decomposition of organic matter and consist of a mixture of complex macromolecules having polymeric phenolic structures and they have been shown to be associated with hydrophobic low molecular weight compounds (Carter and Suffet, 1982; Kulovaara et al, 1987). W e added commercially available humic acid (sodium salt) to distilled water and spiked this suspension with D H A and ISO prior to extraction at different pHs using ethyl acetate (Figure 2.4). The values indicated that the interactions between the humic acid and resin acids were p H dependent, with the extraction efficiency of resin acids reduced at low p H (89%) and enhanced at high p H (97%). It is difficult to explain why the presence of humic acids influenced the extraction trend. It could be that at higher p H , both the resin and humic acids are ionized, and interactions which occur at low p H may not occur at high p H . However, it was clear that the presence of both the H M W material and humic acids affected the extraction efficiency of resin acids. It is apparent that the nature and components of the pulp mi l l effluent w i l l greatly influence the conditions that should be used to ensure optimum recovery of resin acids. Recovery % 100.0 -i 1 75.0 4-pHZ pH2 pH2 pH7 pH9 pH2 pH2 pH7 pH9 with with with with with with with with with 0% 3% 6% 6% 6% 0 ppm 150 ppm 150 ppm 150 ppm HMW HMW HMW HMW HMW HA HA HA HA Figure 2.4 Recovery of resin acids in the presence of high molecular weight ( H M W ) material and humic acid ( H A ) at different pHs (data are the average of three determinations) 59 2.4. Conclusions Liquid- l iquid extraction followed by G C - F J D analysis has been routinely used in the quantification of resin acids from pulp m i l l effluents. While a wide range of p H and solvents have been suggested for the solvent extraction of resin acids, it is apparent that no universal extraction protocol is suitable for all types of effluents. The optimum p H for liquid-liquid extraction should be considered carefully and it may vary depending on the source of effluent. The efficiency of l iquid-liquid extraction is dependent on several factors, such as p H , solvents and the nature of the pulp m i l l effluents. Our G C protocol which involved liquid-liquid extraction, S P E sample clean-up and the use of a dual column system provided a reliable quantitation of resin acids from various sample matrices. While both ethyl acetate and M T B E proved to be good solvents for the extraction of resin acids from pulp m i l l effluents, p H seemed to play a more important role in the efficiency of liquid-liquid extraction. Spiked water, spiked biotreated B K M E and untreated C T M P effluents were extracted more effectively at low p H . The improved extraction efficiency for untreated B K M E at p H 9 could be due to the presence of high concentrations of humic-like material. It is apparent that high p H could be used for the extraction of resin acids from B K M E containing high concentration of lignin-like materials, while a low p H was more suitable for chemi-thermo-mechanical pulp m i l l effluents. It is apparent that solvents and pHs used in the liquid-liquid extraction are two major factors affecting the the recovery of resin acids. The G C protocol developed in this study w i l l be applied to effluent samples from different sources. In order to make it a "standard" procedure which can be used to compare results with E L I S A , an extraction procedure using ethyl acetate was adopted and the pHs used were 2-3 for C T M P effluents and 8-9 for B K M E . It should be recognized that effluent composition can affect the extraction efficiency and the p H used for extraction should be checked for samples from different sources in order to achieve the highest extraction efficiency. 60 C H A P T E R 3 S Y N T H E S I S O F A N T I G E N S A N D D E V E L O P M E N T O F A N E L I S A F O R D H A 3.1. Introduction The use of an immunoassay for the detection and quantification of a chemical compound requires an antibody which w i l l react specifically against the compound of interest. A s described in the first chapter, antibodies are produced by an animal's immune system when a foreign substance is encountered. Unfortunately, small molecules with a molecular weight less than 1000, such as resin acids, are not immunogenic and do not stimulate antibody production (Erlanger, 1980). However, antibodies against these low molecular weight compounds (haptens) can be raised i f the animal was immunized with a hapten-protein conjugate. These conjugates are usually prepared by covalently linking a hapten to a carrier protein (Harlow and Lane, 1988; Weiler, 1990). Therefore, the synthesis of antigen or hapten-protein conjugates becomes the first step in the development of an immunoassay for a low molecular weight compound. It is also a critical step since the quality of the antibodies, such as affinity and specificity, w i l l be largely dependent on the antigen used for immunization which w i l l , in turn, determine the quality of the immunochemical application of the antibodies. The basic requirement for a hapten is that it must structurally mimic the analyte of interest. Resin acids possess a unique tricyclic skeleton, however, the structural differences between individual resin acids are quite small. They can be further divided into abietic type and pimaric type. The former has an isopropyl group on the G-13 position and the latter has a methyl and a v iny l group attached on the C-13 position. A s D H A is the most abundant and persistent resin acids found in softwood pulp m i l l effluents, it was selected as the target compound for E L I S A development. Generally, several factors have to be considered during the preparation of an antigen. These include, the appropriate position of the spacer arm, the type of spacer arm used 61 and the nature of the functional group used for coupling. The present study deals with the synthesis of various antigens, including those with and without a spacer arm, and those with a protein attached at different positions of the hapten molecule. These hapten-protein conjugates were then used to immunize rabbits. The polyclonal antibodies were characterized by the indirect E L I S A . 3.2. Materials and Methods Dehydroabietic acid ( D H A ) was obtained from I C N Biochemicals (Cleveland, Ohio) as a technical grade product and purified to 99.95% purity (Halbrook and Lawrence, 1966). Dehydroabietylamine ( D H A M ) was obtained from Sigma (St. Louis, M O ) as a technical grade product and purified to 99.0% purity (Gottstein and Cheney, 1965). Succinic anhydride, podocarpic acid (POD), N-hydroxysuccinimide (NHS), l-(3-dimethylaminopropyl)-3-ethyl-carbodiimde ( D E C ) , thionyl chloride were purchased from Sigma (St. Louis, M O ) . Tributylamine, tert-butyl chloroformate, hydroxylamine hydrochloride, N,N-dimethylformade ( D M F ) and dioxane were purchased from Aldr ich (Milwaukee, WI). The infrared spectra were recorded on a Perkin-Elmer 1600 Series F T I R spectrometer. The * H N M R spectra were recorded on a Bruker WH-400 and the chemical shift values were reported in mg/L, relative to tetramethylsilane (TMS) . Mass spectra were recorded on an A E I - M S - 9 0 2 (low resolution) and a Kratos-MS-50 (high resolution) spectrometer in the Department of Chemistry, University of British Columbia. Gas chromatography (GC) was performed on Hewlett Packard H P 5890 series II gas chromatograph equipped with an H P 7673 auto injector and flame ionization detector (FLD). G C - M S was performed on a Carloerba K R A T O S M S 8 0 R F A in the Department of Chemistry, University of British Columbia. A l l resin acid standards used in E L I S A were obtained from Hel ix Biotech Corporation (Richmond, B C , Canada). Carbonate-bicarbonate buffer capsules, phosphate-citrate buffer with sodium perborate capsules, o-phenylenediamine (OPD) and Freund's complete and incomplete 62 adjuvants were purchased from Sigma (St. Louis, M O ) . Biotinylated donkey anti-rabbit IgG was obtained from Amersham International pic, (Amersham, U K ) . Horseradish peroxidase (HRP), streptavidin-linked, was obtained from Southern Biotechnology Associates, Inc. (Birmingham, A L ) . B A C T O dehydrated skim milk was purchased from Difco Laboratories (Detroit, MI ) . De-ionized water was purified by the N A N O pure ultrapure water system (Barnstead-Themolyne). Phosphate-buffered saline (PBS, 0.15M) was composed of N a H 2 P 0 4 , N a 2 H P 0 4 , M g C l 2 and N a C l . The E L I S A plate was the 96-well IMMULON® 4 (flat bottom, Catalog N o . 01-010-3855) from Dynatech Laboratories Inc. (Chantilly, V A ) and the optical density was read on a T H E R M O m a x T M microtitre plate reader (Molecular Devices Corp., Menlo Park, C A ) . 3.2.1. Synthesis of antigens The antigens were prepared based on the literature procedures with necessary modifications discussed below (Erlanger, 1980; Robins, 1986). 3.2.1.1. Preparation of D H A M - S U C Dehydroabietylamine (285 mg, 1.0 mmol) and succinic anhydride (120 mg, 1.2 mmol) were dissolved in dry pyridine (10 ml) and stirred at room temperature for 24 hours. Water (20 ml) was added to the reaction mixture and extracted with ethyl acetate (50 m l x 2). The organic layer was washed with water and dried over magnesium sulphate. The solvent was removed by rotary evaporation and the crude product was purified on a silica gel column using ethyl acetate and hexane (3:2 with 0.1% acetic acid) as eluent. Succinylated dehydroabietylamine ( D H A M -S U C ) was obtained as an off-white powder (343 mg, 89%) and characterized by IR, * H N M R and M S as follows: IR ( C H C 1 3 ) v m a x : 1713, 1697 c m " 1 ( R - N H C O C H 2 C H 2 C O O H ) lU N M R ( C D C I 3 , 400 M H z ) 5: 7.16 (d, J=8Hz, 1H, H-14), 6.98 (dd, J=8 and 2 H z , 1H, H-12), 6.90 (d, J=2Hz, 1H, H - l l ) , 6.0 (t, J=2Hz, 1H, N - H ) , 3.20 (t, J=7Hz, 2 H , C O C H 2 C H 2 C O O H ) , 63 2.80 (t, J=7Hz, 2 H , C O C H 2 C H 2 C O O H ) , 1.22 (s, 3 H , H-20), 1.21 (s, 6H, H-16/17), 1.20 (s, 3 H , H-19), 0.90 (s, 2 H , H-18). E I - M S m/z: 131(100%), 173 (41.2%), 259 (18.6%), 285(3.1%), 370 (3.4%), 385 (10.9%). H i g h resolution molecular weight determination calculated for C24H35NO3: 358.2617, found: 358.2610. 3.2.1.2. Preparation of D H A M - S U C - K L H and D H A M - S U C - B S A D H A M - S U C (38.5 mg, 0.10 mmol) was dissolved in dry dioxane (10 ml). To this solution, tributyl amine (28.5 u l , 0.12 mmol) and tert-butyl chloroformate (16.4 p i , 0.12 mmol) were added and stirred at room temperature. After one hour, the resulting anhydride was added dropwise to a solution of B S A (660 mg, 0.01 mmol) in Tris buffer (20 ml , p H 8.8) and stirred at 4°C overnight. The resulting hapten-protein conjugate solution was dialysed against de-ionized water for 24 hrs with 4 changes of water (2 L ) . The dialysed solution was lyophilized and D H A M - S U C - B S A was obtained as a white foam (516 mg). D H A M - S U C - K L H was prepared in the same way as D H A M - S U C - B S A , except that K L H (60 mg, 2x10"^ mmol) was used instead of B S A . The hapten densities were estimated by titration of the available amino groups with trinitrobenzenesulfonic acid (Habeeb, 1966). 3.2.1.3. Preparation of D H A M - S U C - H R P Hapten-enzyme conjugate, D H A M - S U C - H R P , was prepared from D H A M - S U C and H R P by the carbodiimide method following the procedures of Schneider and Hammock (1992) with modifications. D H A M - S U C (1.2 mg, 3 pmol) was dissolved in dry D M F (1.0 ml) and to this solution, N H S (1.7 mg, 15 pmol) and D E C (5.9 mg, 30 pmol) were added and the solution was stirred at room temperature. After three hours, the resulting reaction mixture was added to a solution of H R P (6 mg, 0.14 pmol) in carbonate-bicarbonate buffer (6 m l , 0 .25M, p H 9.6) and stirred at room temperature for two hours and then at 4°C overnight. The resulting hapten-64 enzyme conjugate solution was desalted with a PD-10 column using the supplier's recommended procedure. The enzyme concentration, as determined by the B C A method (Sigma Total Protein Procedure TPRO-562), was 0.56 mg/ml. The final product, D H A M - S U C - H R P , was stored at -20°C. 3.2.1.4. Preparation of D H A - K L H and D H A - B S A D H A - B S A and D H A - K L H were synthesized by converting D H A into the corresponding acid chloride, followed by reaction with protein. The acid chloride was prepared according to the literature with some modifications (Zeiss and Martin, 1953). Briefly, D H A (30 mg, 0.10 mmol) was dissolved in dry benzene (15 ml). To this solution, thionyl chloride (0.23 ml) and two drops of D M F were added with stirring. The mixture was kept at 40°C for 30 minutes, then refluxed for 30 minutes. The solvents were removed under reduced pressure to provide a colourless oi l (31 mg). IR ( C H C 1 3 ) v m a x : 1790cm" 1 . The crude acid chloride was dissolved in anhydrous dioxane (0.5 ml) and B S A (220 mg in 20 m l of P B S , 3.3x10"-* mmol) was added while stirring. The reaction was maintained at room temperature overnight and the resulting solution was dialysed against deionized water for 24 hrs with 4 changes of water (2 L ) . The dialysed solution was lyophilized and D H A - B S A was obtained as a white foam (189 mg). D H A - K L H was prepared in a similar manner, except that K L H (60 mg, 2x10"^ mmol) was used instead of B S A . 3.2.1.5. Succinylation of BSA and K L H B S A and K L H was succinylated according to a literature procedure (Hung et al, 1980). Brief ly, succinic anhydride (420 mg, 4.1 mmol) was added in small portions to an aqueous solution of B S A (132 mg, 2x10"3 mmol, in 10 m l of distilled water) while the p H was maintained at 7 with I N N a O H . The solution was incubated at room temperature for 30 min.. The p H was then adjusted to 2.5 with I N HC1, and the solution was dialysed against deionized 65 water for 24 hrs with 4 changes of water (2 L ) . The dialysed solution was lyophilized and succinylated B S A was obtained as a white foam (110 mg). K L H was succinylated in a similar way except that 50.0 mg of K L H was used. 3.2.1.6. Preparation of POD-SUC-BSA and POD-STJC-KLH The succinylated B S A (35.0 mg) was dissolved in dry D M F (5.0 ml). To the solution was then added podocarpic acid (11.0 mg, 0.04 mmol) and D E C (29 mg, 0.15 mmol). The reaction mixture was stirred at room temperature overnight. The product was purified by dialysis against deionized water for 24 hrs with 4 changes of water (2 L ) . The dialysed solution was lyophilized and P O D - S U C - B S A was obtained as a yellowish foam (36.6 mg). P O D - S U C - K L H was prepared in a similar way using 8.0 mg of P O D and 20.0 mg of succinylated K L H , and the product was a yellowish foam (22.4 mg). 3.2.1.7. Preparation of D H A - O X I - S U C - B S A and D H A - O X I - S U C - K L H D H A was oxidized to 7 - O X O - D H A and then converted to oxime based on the procedure of Pratt (1951) with some modifications. Briefly, D H A (0.5 g, 1.67 mmol) was dissolved in an aqueous potassium hydroxide solution (0.1 g in 10 m l water). A solution of potassium permanganate (0.7 g, 4.43 mmol, in 15 m l water) was added dropwise with stirring at approximately the rate at which it was reduced. The reaction mixture was then stirred at room temperature for 3 hrs and any excess of potassium permanganate was destroyed by addition of sodium sulfite (Na2S03). The reaction mixture was acidified with 1.0N HC1 and extracted with ethyl acetate (50 m l x2). The combined organic extracts were washed with water (20 ml) and brine (20 ml), dried, filtered and rotary evaporated. The crude product was obtained as a semi-oi l (0.4 g), which was purified by column chromatography on silica gel using hexane:ethyl acetate (6:1, with 0.1% H A c ) as eluent and 7 - O X O - D H A was obtained as major product (white foam, 284 mg). U V V a x : 254nm (s 9270), 300nm (e 2150). lH N M R ( C D C 1 3 , 400 M H z ) 5: 66 7.85 (d, J=2.1, 1H, H-14), 7.38 (dd, J=8.2 and 2.1 H z , 1H, H-12), 7.27 (d, J=8.2 H z , 1H, H - l l ) , 2.92 (m, 1H, H-15), 1.35 (s, 3 H , H-19), 1.24 (s, 3H, H-20), 1.22 (d, J= l . l H z , 3 H , H-16), 1.21 (d, J = l . l H z , 3 H , H-17). E I - M S m/z: 253 (100%), 299 (30.5%), 314 (50.9%). H i g h resolution molecular weight determination calculated for C20H26O3: 314.1882, found: 314.1880. 7 - O X O - D H A (50.0 mg, 0.14 mmol), hydroxyamine hydrochloride (50.0 mg) and pyridine (0.5 ml) were dissolved in ethanol (10 ml) and the solution was refluxed for two hours. The reaction mixture was diluted with water (10 ml) and extracted with diethyl ether (20 m l x2). The ether extract was washed with water (20 m l x2), dried over anhydrous M g S 0 4 . After filtration, removal of the solvent under reduced pressure afford a crude product (colourless o i l , 40 mg). Purification by column chromatography on silica gel using hexane:ethyl acetate (6:1, with 0.1% H A c ) as eluent afforded pure oxime as white foam (31 mg), which decomposed at about 250°C. * H N M R ( C D C 1 3 , 400 M H z ) 5: 8.62 (s, broad, 1H, N - O H ) , 7.90 (d, J=2.1, 1H, H -14), 7.41 (dd, J=8.2 and 2.1 H z , 1H, H-12), 7.30 (d, J=8.2 H z , 1H, H - l l ) , 2.94 (m, 1H, H-15), 1.35 (s, 3 H , H-19), 1.26 (s, 3 H , H-20), 1.24 (d, J = U H z , 6H, H-16 and H-17). E I - M S m/z: 79 (46.2%), 199 (11.5%), 211 (34.6%), 250 (58.3%), 253 (100%), 268 (63.8%), 299 (41.8), 314 (66.7%o), 329 ( M + , 51.5%). H i g h resolution molecular weight determination calculated for C20H27NO3: 329.1991, found: 329.1984. The succinylated B S A (35.0 mg) was dissolved in dry D M F (5.0 ml). To the solution was then added oxime (10.0 mg, 0.030 mmol) and D E C (29 mg, 0.15 mmol). The reaction mixture was stirred at room temperature overnight. The product was purified by dialysis against deionized water for 24 hrs with 4 changes of water (2 L ) . The dialysed solution was lyophilized and D H A - O X I - S U C - B S A was obtained as a white foam (37.5 mg). D H A - O X I - S U C - K L H was prepared in a similar way using 8.0 mg of the oxime and 15.0 mg of succinylated K L H , and the product was a white foam (18 mg). 67 3.2.1.8. Preparation of D H A - K L H - P T M and DHA-BSA-PEVI B S A and K L H were modified by attachment of both D H A and P L M simultaneously. A mixture of D H A (15.0 mg, 0.05 mmol) and P L M (15.0 mg, 0.05 mmol) was dissolved in dry benzene (15 ml). To this solution, thionyl chloride (0.23 ml) and two drops of D M F were added with stirring. The mixture was kept at 40°C for 30 minutes, then refluxed for 30 minutes. The solvents were removed under reduced pressure and B S A (150 mg in 20 m l of PBS) was added while stirring. The reaction was maintained at room temperature overnight and the resulting solution was dialysed against deionized water for 24 hrs with 4 changes of water (2 L ) . The dialysed solution was lyophilized and D H A - B S A - P I M was obtained as a white foam (108 mg). D H A - K L H - P L M was prepared in a similar manner, except that K L H (60 mg, 2x10~5 mmol) was used instead of B S A . 3.2.2. Immunization protocols Each of four N e w Zealand white rabbits was injected subcutaneously with 2.0 m l of D H A M - S U C - K L H (1.75 mg dissolved in 1.0 m l of P B S with 0.25 m l of D M S O and emulsified with 1.25 m l of Freund's complete adjuvant). A t days 21, 38 and 50, about 10 m l of blood were withdrawn from each rabbit to monitor antibody production. A t each of these times, 1.0 m l of D H A M - S U C - K L H (1.0 mg dissolved in 0.5 m l of P B S with 0.25 m l of D M S O and emulsified in 0.75 m l of Freund's incomplete adjuvant) was injected intramuscularly. The antisera titres were examined using indirect E L I S A in which D H A M - S U C - B S A was used as the coating antigen. Seven days after the last injection, the rabbits were bled and the antisera were collected by centrifuge and removal of cell material (Harlow, 1988), and were stored at -80°C. Two N e w Zealand white rabbits were used in the immunization with each of the fol lowing antigens, D H A - K L H , P O D - S U C - B S A , D H A - O X I - S U C - B S A and D H A - K L H - P L M , in a similar way. Boosting injections were carried out approximately 3, 5 and 7 weeks after the 68 initial immunization. Seven days after the last boosting, rabbits were bled and the antisera were collected and stored at -80°C. 3.2.3. Detection of polyclonal antibodies against D H A The indirect E L I S A with the biotin-streptavidin system was used to detect the polyclonal antibodies present in the serum from rabbits immunized with antigens. The checkerboard titration was adopted and the major procedures were as follows. The plate was washed five times using P B S between each step to remove the unbound material and the plate was covered by a plastic f i l m to avoid evaporation during incubation. Step 1. The microtitre plate was coated with antigen at different concentrations (2 to 125 ng/100 ul/well) in carbonate-bicarbonate buffer (pH 9.6) and dried at 37°C overnight. Step 2. The excess binding sites on the microtitre plate were blocked using P B S with 2% milk (200 u,l/well) at 37°C for 1 hr. Step 3. Primary antibodies (diluted in 0.1% milk in PBS , 1:2000 to 1:128000) were added to the plate (100 ul/well) and incubated at 37°C for 2 hrs. Step 4. The biotinylated donkey anti rabbit IgG (diluted in 0.1% milk in P B S , 1:12000) was added to the plate (100 ul/well) and incubated at 37°C for 1 hr. Step 5. Streptavidin-HRP (diluted in 0.1% milk in P B S , 1:12000) was added to the plate (100 ul/weli) and incubated at 37°C for 1 hr. Step 6. O P D (1.0 mg/ml in citrate-phosphate buffer with sodium perborate) was added to the plate (100 ul/well), the reaction was stopped with sulphuric acid (2.5 M , 50 ul/well) after 5 minutes and O D at 490 nm was measured. 3.2.4. Indirect E L I S A of D H A using the biotin-streptavidin system The indirect E L I S A with biotin-streptavidin system was also used to assess the specificity of the antibodies to free D H A and other resin acids. The procedures are similar to 69 those described previously, except that DFIA or other resin acids were used as inhibitors in step 3. Step 1. The microtitre plate was coated with D H A M - S U C - B S A (25 ng/100 ul/well) in carbonate-bicarbonate buffer (pH 9.6) at 37°C overnight. Step 2. The excess binding sites on the microtitre plate were blocked using P B S with 2% milk (200 ^I/well) at 37°C for 1 hr. Step 3. Appropriate volumes of D H A or other resin acid standards (prepared in methanol as stock solution, about 1.0 mg/ml) were evaporated in glass test tubes and mixed with the primary antibodies (1:30000, diluted in 0.1% milk in PBS) . The mixture was added to the plate (100 ul/well) and incubated at 37°C for 2 hrs. Step 4. The biotinylated donkey anti rabbit IgG (diluted in 0.1% milk in P B S , 1:12000) was added to the plate (100 ul/well) and incubated at 37°C for 1 hr. Step 5. Streptavidin-HRP (diluted in 0.1% milk in P B S , 1:12000) was added to the plate (100 ul/well) and incubated at 37°C for 1 hr. Step 6. O P D (1.0 mg/ml in citrate-phosphate buffer with sodium perborate) was added to the plate (100 ul/well), the reaction was stopped with sulphuric acid (2.5 M , 50 ul/well) after 5 minutes and O D at 490 nm was measured. 3.3. Results and Discussion 3.3.1. Synthesis of antigens Both dehydroabietic acid ( D H A ) and dehydroabietylamine ( D H A M ) were obtained from commercial sources as technical products. D H A was purified to 99.9% purity by several steps 70 including isolation of its salt with 2-aminoethanol, acidification, recrystallization (Halbrook and Lawrence, 1966) and finally reversed phase preparative H P L C purification. D H A M was purified to 99.0% purity by the isolation of its acetic acid salt, recrystallization, alkalization and extraction (Gottstein and Cheney, 1965), and finally silica gel chromatography. 3.3.1.1. Antigens with spacer arm: D H A M - S U C - K L H and D H A M - S U C - B S A In order to develop an immunochemical method for the analysis of resin acids, it was important to select a suitable hapten which could be used to raise antibodies specific to resin acids. From a chemical point of view, pimaric acid should be the best choice for the hapten since it is structurally distinct from the other common resin acids. It is the only compound among the eight most common resin acids that possesses a methyl and a vinyl group, and it has an S configuration at the C-13 position. This structural feature is important from an immunological point of view, since the more unique the antigen structure is, the higher the possibility of producing more specific antibodies. However, we decided to choose D H A as our target compound, mainly because it is the most abundant and persistent resin acid in softwood pulp m i l l effluents and has been considered to be more representative of the total resin acids (Taylor et al, 1987). Initially, we attempted to introduce a spacer arm by reacting D H A with &-aminocaproic acid, a spacer molecule with six carbon atoms. W e then hoped to conjugate the derivatized D H A with B S A and K L H via the carboxylic group, using either the mixed anhydride method or the carbodiimide method. However, we were unable to put a spacer arm onto the D H A molecule. When dicyclohexylcarbodiimide (DCC) was used as dehydrating agent, the reaction stopped at the first stage and a D H A - D C C adduct with a molecular weight of 506 was formed. N o expected product was isolated (Figure 3.1). Use of water soluble l-(3-dimethylaminopropyl)-3-ethyl-carbodiimde (DEC) also failed and a D H A - D E C adduct with a molecular weight of 455 was isolated as the only product. When the mixed anhydride method was used, D H A could be easily converted to the mixed anhydride. However, further reaction 71 with co-aminocaproic acid led to formation of a by-product with a molecular weight of 231 and D H A was regenerated. These results indicated that the steric hindrance of the carboxylic group could be a major factor affecting its reactivity. H9C4OCOHN(CH2)5COOH MW=231 Figure 3.1 Schemes for an attempted synthesis of hapten 72 Another attempt was then made to convert D H A into an N-methyl amine via a series of wel l established procedures (Zeiss and Martin, 1953) (Figure 3.2). However, the last step was again unsuccessful. It is most likely that steric hindrance played an important role because the amine is a secondary amine with the nitrogen atom attached to a tertiary carbon atom and therefore, it is very sterically hindered. Although we were unable to achieve our goal from this synthetic pathway, the scheme did suggest that the acid chloride could be a suitable intermediate to start the reaction with. For example, it can be converted to dehydroabietylamine ( D H A M ) , by reaction with ammonia followed by reduction, and the D H A M should then easily react with succinic anhydride since it is a primary amine with the nitrogen atom attached to a primary carbon atom. Thus the steric hindrance should be much less. The acid chloride can also react with proteins directly to form hapten-protein conjugates. These two ideas resulted in the successful syntheses of antigens with and without spacer arm as described in Figure 3.3. toluene heating Figure 3.2 Schemes for an attempted synthesis of hapten 73 Fortunately, dehydroabietylamine ( D H A M ) is commercially available as a technical grade product. After purification via its acetic acid salt (Gottstein and Cheney, 1965), a spacer arm was introduced by the reaction of D H A M with succinic anhydride. The hapten was then conjugated to the proteins using the conventional mixed anhydride method (Figure 3.3). The hapten densities (hapten/protein ratio), determined by the titration of the free amino groups (Habeeb, 1966), were found to be approximately 9 for D H A M - S U C - B S A and 20 for D H A M -S U C - K L H (Table 3.1). D H A M D H A M - S U C D H A M - S U C - B S A D H A M - S U C - K L H D H A M - S U C - H R P DHACOC1 D H A - B S A D H A - K L H Figure 3.3 Schemes for the syntheses of antigens with and without spacer arm 3.3.1.2. Antigens without spacer arm: D H A - K L H and D H A - B S A After successful preparation of D H A M - S U C - K L H and D H A M - S U C - B S A , our attempts were again focused on the direct coupling of D H A to a carrier protein. A s discussed previously, the carboxylic group in D H A and other resin acids is bound to a tertiary carbon atom and the high steric hindrance retards its reactivity. For example, esterification of resin acids is less 74 efficient and requires large excess of alcohol and high temperature (Soltes and Zinkel , 1989). Therefore, an excess of thionyl chloride was used to convert D H A into the corresponding acid chloride, which is more reactive than the mixed anhydride. The acid chloride was formed almost quantitatively and IR showed a strong absorption at 1790 c m - 1 , which is from the carbonyl group in the acid chloride. Further reactions with B S A and K L H led to antigens without a spacer arm, D H A - B S A and D H A - K L H . The hapten density was 6 and 23 for D H A - B S A and D H A -K L H , respectively, as estimated by titration of the free amino groups (Habeeb, 1966). 3.3.1.3. Enzyme conjugate: D H A M - S U C - H R P Various attempts were made in order to synthesize the hapten-enzyme conjugate, a major reagent required by the direct E L I S A format. The reaction of D H A M - S U C with H R P using the same conditions for the synthesis of D H A M - S U C - B S A and D H A M - S U C - K L H was unsuccessful, probably because the H R P has a high isoelectric point (pi 8.8) and a low lysine content (6/mole) (Ngo, 1991). Therefore, D H A M - S U C was activated with both N -hydroxysuccinimide (NHS) and l-(3-dimethylaminopropyl)-3-ethylcarbodiimide hydrochloride ( D E C ) , then reacted with H R P in carbonate-bicarbonate buffer at p H 9.6. The product mixture was de-salted using a PD-10 column to give the enzyme conjugate, D H A M - S U C - H R P . The protein concentration was estimated by the B C A method (Sigma Total Protein Procedure T P R O -562) and was found to be approximately 0.56 mg/ml. 3.3.1.4. Other antigens Attempts were also made to attach the protein at different positions on the hapten molecules. Podocarpic acid (POD) was selected due to its similar structure to D H A . It was hoped that, by linking the carrier protein at a position far from the carboxylic group, the antibodies raised would recognize the entity of the molecule near the carboxylic group, thus providing generic antibodies to the resin acids. In order to attach the carrier protein to the 75 phenolic hydroxyl group in P O D , both B S A and K L H were first modified by reaction with succinic anhydride to form succinylated B S A and K L H (Hung et al, 1980). They were subsequently reacted with P O D using the carbodiimide method to give compounds P O D - S U C -B S A and P O D - S U C - K L H (Figure 3.4). Hapten density was 8 and 12, respectively (Table 3.1) OH HOOCCH 2 CH 2 CONHprotein COOH POD O—COCH 2 CH 2 C01VH-protein COOH POD-SUC-BSA P O D - S U C - K L H Figure 3.4 Schemes for the syntheses of P O D - S U C - B S A and P O D - S U C - K L H Another attempt was made to attach the protein at the C-7 position in D H A . This was achieved by oxidizing D H A to 7 - O X O - D H A , converting it to oxime (Pratt, 1951) and subsequently conjugating with succinylated B S A and K L H . D H A was easily oxidized to 7-O X O - D H A using potassium permanganate under alkaline conditions. G C analysis indicated that two products were formed during the oxidation and the structures were identified as 7 - O X O -D H A (major) and 7 - O X O - l 3 - h y d r o x y - D H A (minor) (Rogers and Mahood, 1981). The products were wel l resolved in a G C and the relative retention times were 1.00, 1.27 and 1.37 min for D H A (unreacted, 26%), 7 - O X O - D H A (65%) and 7 -OXO-13-hydroxy-DHA (6%), respectively. The major product, 7 - O X O - D H A , was purified by silica gel chromatography. The structure was confirmed by both proton and C-13 N M R . A H N M R showed that the chemical shift for H-14 has been moved from 6.88 mg/L in D H A to 7.85 mg/L and a new ^ C signal at a chemical shift of 198.8 mg/L was found for C-7, instead of 30.0 mg/L in D H A . There was also a bathochromic 76 shift in the U V spectra (300 nm compared with 275 nm in D H A ) . A l l this information implied that a conjugated aromatic ketone system had been formed due to oxidation at the C-7 position. Both the low and high resolution M S further confirmed the structure! The reaction of 7 - O X O -D H A with hydoxylamine provided the oxime, structure of which was confirmed by N M R and M S . The 1 3 C chemical shift for C-7 was shifted from 198.8 mg/L in 7 - O X O - D H A to 155.5 mg/L, due to structural modification at the C-7 position. A broad signal at 10.0 mg/L, caused by = N - O H , was found in its proton N M R . A molecular ion peak at w/z=329 and m/z=329.1984 in low and high resolution E I - M S further confirmed the structure. The oxime was then conjugated to succinylated B S A or K L H using the carbodiimide method to gave the final product D H A -O X I - S U C - B S A and D H A - O X I - S U C - B S A (Figure 3.5). The estimated hapten density was 3 and 7 for D H A - O X I - S U C - B S A and D H A - O X I - S U C - B S A , respectively (Table 3.1), indicating the conjugation was not effective. DHA 7-OXO-DHA Oxime H O O C C H 2 C H 2 C O —NH—protein DHA-OXI-SUC-BSA DHA-OXI-SUC-KLH Figure 3.5 Schemes for the syntheses of D H A - O X I - S U C - B S A and D H A - O X I - S U C - K L H 77 In another attempt to raise antibodies which were capable of recognizing both abietic type and pimaric type resin acids, it was decided to link both D H A and P L M , representatives from the abietic type and pimaric type resin acids, covalently to the same carrier protein and use this conjugate as the antigen to immunize rabbits. Therefore, a mixture of D H A and P L M in equal amounts was converted to the corresponding acid chlorides, as described previously, and conjugated to K L H and B S A directly. The hapten density was 18 and 10 for D H A - K L H - P L M and D H A - B S A - P L M , respectively, as estimated by titration of the free amino groups (Habeeb, 1966) (Table 3.1). Table 3.1 Hapten densities estimated in various hapten-protein conjugates Hapten-protein conjugate Hapten density (hapten/protein) D H A M - S U C - K L H * 19 D H A M - S U C - B S A 8 D H A M - S U C - H R P not determined D H A - K L H * 23 D H A - B S A 6 P O D - S U C - B S A * 8 P O D - S U C - K L H 12 D H A - O X I - S U C - B S A * 3 D H A - O X I - S U C - K L H 7 D H A - K L H - P L M * 18 D H A - B S A - P L M 10 * Antigens used for immunization. 78 3.3.2. P r e l i m i n a r y i m m u n i z a t i o n results The above ten antigens that were synthesized (Table 3.1) represented hapten-protein conjugates of different hapten structures and carrier proteins, with and without a spacer arm. Five of them were subsequently used to immunize the rabbits. It was expected that antibodies with different specificities would be derived from these antigens. 3.3.2.1. P r o d u c t i o n o f p o l y c l o n a l an t ibod ies us ing D H A M - S U C - K L H as an t igen Four rabbits were immunized with antigen bearing a four carbon spacer arm, D H A M -S U C - K L H . During the various stages of immunization, antibody production was monitored using indirect E L I S A . Antibodies were detected three weeks after the initial immunization. After three boosting injections, the titre had reached a 1:30,000 level and the rabbits were exsanguinated. Very high titers were observed for the antisera from each of the four different rabbits (Figure 3.6). Not surprisingly, it was found that there was a slight difference in the immune responses from the four rabbits when assessed by their abilities to recognize D H A . The antiserum from rabbit 3 were used in the subsequent studies- as it showed the best response in inhibition study. Figure 3.7 shows a three-dimensional titration plot of the antiserum from rabbit 3 with different concentrations of coating antigen ( D H A M - S U C - B S A ) . A n optical density reading of 1.0 at 490 nm was obtained with a coating antigen concentration of 25 ng/100 ul/well and an antiserum dilution of 1:30,000, i f the enzyme reaction was stopped after 5 minutes. 3.3.2.2. E v a l u a t i o n o f the p o l y c l o n a l an t ibod ies A s one of the main objectives of this work was to develop an immunoassay for resin acids, it was important to determine i f the polyclonal antibodies could recognize free D H A and resin acids. Initially, the antibodies were evaluated in a competitive indirect E L I S A . The inhibition curve with free D H A concentrations ranging from 250.0 to 1.0 ug/L is shown in Figure 3.8. The percentage of inhibition was calculated using equation 3.1. 79 ODat 490 nm 1:2000 1:4000 1:8000 1:16000 1:32000 1:64000 1:128000 Antibody dilution Figure 3.6 Determination of titre for four different rabbits immunized with D H A M - S U C - K L H 1:640oo Coating antigen Serum dilution Figure 3.7 Checkerboard titration for polyclonal antibodies from rabbit #3 immunized with D H A M - S U C - K L H 80 % Inhibition = 100 x A o - A A o [Eq.3.1] where A o is the absorbance with no D H A present, and A is the absorbance with D H A present. The detection limit was arbitrarily defined as the amount of analyte which yielded a 20% inhibition. Linear regression analysis of the data from 1.0 to 250.0 ug/L gave the equation; % Inhibition ( D H A ) = 11.68 + 29.36 L o g [ D H A ] (R = 0.993) From this equation, the 50% inhibition and 20% inhibition (the detection limit) points were calculated as 20.2 pg/L and 1,9 pg/L, respectively. It was clear that the antiserum raised against the antigen with a spacer arm, D H A M - S U C -K L H , recognized both the free D H A and the hapten present on the carrier protein ( D H A M - S U C -B S A ) used in the immobilized phase of the assay. Inhibit ion% 10 D H A concentration (ppb) 100 Figure 3.8 Inhibition curve for D H A using the indirect E L I S A with biotin-streptavidin system 81 3.3.2.3. Cross-reactivity To characterize the specificity of the polyclonal D H A antibodies, the seven major resin acids which are usually detected in the effluents of softwood species after they have been pulped, were assayed in a competitive inhibition E L I S A . The structures of the two main types of resin acids are described in Figure 1.1. Abietic types have an isopropyl side chain at position 13, pimaric types have both methyl and vinyl substituent at position 13. Two chlorinated resin acids, mono- and dichlorodehydroabietic acid, were also tested. These modified compounds are produced during chlorine bleaching, and are frequently isolated from effluents. The percentage cross reactivity (%CR) is defined as the ratio of the amount of D H A to the amount of the cross-reactant required to reduce binding by 50% as indicated in equation 3.2. % C R = 100 x ( I 5 0 f o r D H A / I 5 0 for cross-reactant) t E q 3 - 2 1 To compare the cross-reactivities of the nine resin acids indicated above, we determined the amount of each resin acid required to achieve 50% inhibition (IC50) and calculated the cross-reactivity of each resin acid as the ratio of the IC50 of D H A to that of the cross-reactant. The upper and lower limits of detection were arbitrarily defined as the amount of compound required to achieve 80% and 20% inhibition, respectively (Midgley et al, 1969). The lower detection limit, as determined above, was comparable to an alternative approach where the lower detection limit was defined as the amount of compound required to achieve an absorbance (A) which is separated from the control absorbance (Ao) by three standard deviations (Fleeker, 1987). The data in Table 3.2 summarize the cross-reactivity of the antiserum and indicate the upper and lower detection limits obtained when the antiserum was used in an indirect E L I S A using the biotin-streptavidin system. The antiserum was highly cross-reactive with the mono-and dichlorodehydroabietic acids, two derivatives of D H A . It also showed a high degree of cross-reactivity with the other abietic type resin acids, abietic and palustric acids. Each of these four analogues of D H A has an isopropyl group at the C-13 position.. Alternatively, the antiserum 82 showed lower or negligible cross reactivity with pimaric type resin acids, which have both methyl and vinyl groups at the C-13 position. Table 3.2 Percentage cross-reactivity of the polyclonal antibodies from D H A M - S U C - K L H with different resin acids and the lower and upper detection limit (DL) Resin A c i d I C 5 0 (Ug/L)* Cross reactivity (%) Upper D L (Ug/L)* Lower D L (Ug/L)* D H A 20.2 ± 3 . 5 100.0 212.3 ± 10.4 1 . 9 ± 0 . 8 C 1 D H A 22.8 ± 3 . 7 88.6 168.7 ± 11.0 3.1 ± 0 . 8 A B A 27.0 ± 5 . 0 74.8 254.8 ± 18.7 2.9 ± 0 . 9 C 1 2 D H A 28.6 ± 5 . 3 70.6 270.5 ± 1 9 . 8 3.0 ± 0 . 9 P A L 44.7 ± 9 . 0 45.2 515.7 + 41.1 3.9 ± 1.3 P L M 301.7 ± 4 2 . 3 6.7 1680.6 ± 9 4 . 0 54.2 ± 12.3 ISO 342.3 ± 5 8 . 2 •5.9 2732.0 ± 1 8 4 . 9 42.9 ± 11.8 S A N D 384.4 ± 7 3 . 2 5.3 3930.8 ± 3 0 1 . 9 3 7 . 6 ± 11.6 N E O 669.3 ± 1 1 2 . 2 3.0 5189.5 ± 3 4 6 . 4 86.3 ± 2 3 . 3 A l l the IC50, upper and lower detection limits are presented as average values from three determinations with standard deviation. Therefore, the polyclonal antibodies obtained from rabbits immunized with D H A M -S U C - K L H showed very high sensitivity to D H A . They could recognize D H A and other resin acids, especially for abietic type resin acids. Although these polyclonal antibodies might not be used to quantify the total resin acids present in an undefined mixture, they could probably be used to detect the generic class of abietic resin acids. To test this hypothesis, we used D H A and pimaric acid (PLM) as representatives of the abietic and pimaric type resin acids and prepared a 83 series of mixed analytes as potential inhibitors. When D H A and P I M were mixed in ratios from 3:1 to 1:3,.up to a total concentration of 1000 ug/L, it was found that the less cross-reactive component, pimaric acid, had a negligible effect on the quantitative detection of D H A . It was apparent that D H A could be quantified in the presence of P I M over a wide range of concentrations. This could be a general trend for abietic type and pimaric type resin acids, and should imply that the polyclonal antibodies can be used to quantify abietic type resin acids. 3.3.2.4. Recovery of D H A by E L I S A in Spiked Water and Effluent A s the initial work indicated that the polyclonal antibodies were capable of recognizing D H A , the recoveries of D H A in spiked tap water and biotreated B K M E were next evaluated by the indirect E L I S A with the biotin-streptavidin system. Prior to spiking with D H A , the absence of resin acids in the biotreated B K M E sample was confirmed by G C analysis. Both tap water and the effluent samples were spiked with various amounts of D H A , ranging from 1.0 to 500.0 ug/L, by the addition of suitable amounts of a D H A stock solution in methanol and kept at 4 °C overnight. These samples were then analyzed by the indirect E L I S A with slightly modified procedures. A 200 pi aliquot of the D H A spiked solution was mixed with 800 u l of the antiserum, which was diluted 1:24,000 using 0.1% milk in PBS . This maintained the final serum dilution at 1:30,000. The possible interference from the effluent matrix was also kept to a minimum since the sample was diluted five times. Subsequently, 100 ul of this solution was then transferred to each wel l in the microtitre plate and the remaining steps were carried out as previously described. The recoveries were satisfactory and ranged from 82% to 120%) (Table 3.3). Higher recoveries, with a larger percentage error, were found at low D H A levels. This was probably due to the proximity to the detection-limit of the assay. 84 Table 3.3 Recovery of D H A from water and effluent spiked with D H A by the indirect E L I S A with the biotin-streptavidin system Matrix D H A spiked (Ug/L) Recovered D H A (Ug/L)' Recovery (%) Water 500.0 486.2 + 35.2 93.6 ± 7 . 2 50.0 41.1 ± 7 . 4 82.2 ± 18.0 20.0 18.8 ± 1 . 8 94.0 ± 9.6 1.0 1.1 ± 0 . 3 110.0 + 27.3 Effluent 500.0 503.1 ± 4 1 . 0 100.6 ± 8 . 1 50.0 52.8 ± 6 . 5 105.6 + 12.3 20.0 16.9 ± 2 . 4 84.5 + 14.2 1.0 1.2 ± 0 . 3 120.0 + 25.0 *Data are the average from three determinations, each with eight replicate wells 3.3.2.5. P r o d u c t i o n a n d e v a l u a t i o n o f o the r p o l y c l o n a l an t ibod ies The antigen without the spacer arm ( D H A - K L H ) was also used to immunize rabbits using similar immunization protocols ( L i et al, 1994). After three booster injections following initial immunization, polyclonal antibodies with high titre were obtained. Preliminary evaluation using the indirect E L I S A with the biotin-streptavidin system indicated that the antibodies could be diluted 1:25,000, as determined by checkerboard titration. Detailed characterization of the polyclonal antibodies w i l l be discussed in the next chapter. When rabbits were immunized with the P O D - S U C - B S A , a high titre was obtained (at least 1:20,000) when P O D - S U C - K L H was used as the coating antigen. However, the polyclonal antibodies were incapable of recognizing free D H A up to concentration of 40 mg/L, although P O D was recognized at I50 around 1.9 mg/L. This could be due to the fact that the P O D and D H A had opposing stereochemistry at the C-4 position. Since the polyclonal antibodies were derived from P O D - S U C - B S A , with the protein attached to a position far from the carboxylic 85 group, the carboxylic part of the P O D could be the potential epitope for binding of the antibody. It is a common phenomenon that antibodies can distinguish different stereoisomers. For example, in the development of an immunoassay for abscisic acid, which is a plant growth inhibitor with four possible isomers, the use of an antigen synthesized from a racemic starting material elicited an antiserum which reacted predominantly with the unnatural (-)-isomer, i f the conjugation was achieved by reaction of carboxylic group with carrier protein. However, coupling through the carbonyl group via hydrazone intermediates elicited the production of antisera which reacted preferentially with (+)-isomer (Weiler, 1990). O n the other hand, antiserum from a rabbit immunized with D H A - O X I - S U C - B S A showed a relatively low titre and it could only be used in a 1:5000 dilution, as determined by checkerboard titration. Again, the polyclonal antibodies were incapable of recognizing free D H A when the indirect E L I S A with the biotin-streptavidin system was used. It is not known why the antibodies from D H A - O X I - S U C - B S A could not recognize free D H A or the other resin acids. It was probably due to the low hapten density of D H A - O X I - S U C - B S A as indicated in Table 3.1, and therefore, not effective in the elicitation of production of hapten specific antibodies. The response to the E L I S A was probably due to "spacer arm recognition" caused by antibodies which were active against the carrier protein or the bridge between the hapten and the carrier. Since the polyclonal antibodies from D H A M - S U C - K L H can recognize D H A - O X I - S U C - B S A when it was used as the coating antigen, it is evident that the D H A - O X I - S U C - B S A had the proper hapten attached to the B S A . Another possibility is that, when this antigen was injected into the rabbits, it was metabolized or decomposed. N o further experiments were performed using these antibodies. Polyclonal antibodies were also raised using rabbits immunized with a mixed antigen of K L H with both D H A and P I M attached to it ( D H A - K L H - P E V I ) . These polyclonal antibodies were capable of recognizing both D H A and P I M with high sensitivity when assayed individually using D H A - B S A - P I M as the coating antigen. However, the technical problems associated with the use of these antibodies for the quantification of both D H A and P I M simultaneously were not 86 resolved within the time frame of this work, primarily because of the lack of a suitable standard curve. More work needed to be done to decide how to best use these antibodies in the quantification of resin acids. 3.4. Conclusions Selection of haptens and synthesis of antigens are two of the most important steps in the development of an immunoassay for low molecular weight compounds. A total of ten hapten-protein conjugates and one hapten-enzyme tracer were synthesized, with hapten densities ranging from 3 to 23. Due to the steric hindrance, the carboxylic group in D H A is less reactive and the direct conjugation of D H A to protein was only achieved by the acid chloride method. Five hapten-protein conjugates were subsequently used to immunize rabbits for the production of polyclonal antibodies. However, useful antibodies were only obtained from rabbits immunized with either D H A M - S U C - K L H or D H A - K L H . The use of limited numbers of antigens in the immunization of rabbits did not lead to production of polyclonal antibodies with expected specificity, and more work could be carried out to design more suitable haptens. Obviously, antibodies capable of recognizing all of the resin acids could be useful since it is the total resin acid concentration in a particular effluent sample that primarily determines its toxicity. Alternatively, it would also be useful to raise antibodies which are highly specific to a particular resin acid since this would allow the monitoring of specific compounds in the sample. Therefore, we could produce either generic antibodies which cross-react equally wel l with each resin acid and could thus be used to detect the total resin acid content, or a more specific antibody, which could be used to quantify a specific resin acid. These options could be pursued by the careful design of haptens, the synthesis of antigens, and use of these new complexes in the production of both polyclonal and monoclonal antibodies. 87 CHAPTER 4 EVALUATION OF TWO POLYCLONAL ANTIBODIES AGAINST DHA AND OPTIMIZATION OF ELISA 4.1. Introduction The performance of an immunoassay is mainly determined by the quality of the antibodies. Usually antibodies with high affinity constants are more desirable for immunoassays than those of lower affinity. However, the affinity constant, as wel l as the association and the dissociation constants of the antibody-antigen complexes are fixed once antibodies are obtained. Fortunately, immunoassays designed for the quantification of low molecular weight compounds, such as E L I S A s , are based on a competitive process and antibodies are usually used in a limited amount. Therefore, the absolute amount of material required for 50% inhibition (I50) is not only a function of the affinity constant, but also a function of the assay format and the size of mass action reaction (Vanderlaan et al, 1990). The lower the amount of hapten-protein conjugate bound to the solid phase and the lower the amount of antibody used, the less competitor needed for inhibition and thus the more sensitive the assay. A successful E L I S A format should be designed in such a way that, at a particular concentration of inhibitor (analyte), maximal inhibition should be obtained. This can be achieved by optimization of assay conditions such as the concentrations of the coating antigen and the dilutions of the primary antibodies. When several antigens and antibodies are available, it is also possible to design an E L I S A with suitable sensitivity by using different combinations of antibodies and antigens. Under a particular assay format, maximum sensitivity can usually be achieved by decreasing the concentration of coating antigen and using antibodies with higher affinities to the free hapten. Assay formats and enzyme amplification are other optimization parameters which could be manipulated to improve the detection limit, assay time and other performance parameters (Johannsson et al, 1986). 88 r The influence of p H , ionic strength, temperature, organic solvents and other nonspecific parameters varies from assay to assay. It is generally difficult to predict the optimum conditions for the antibody-antigen interactions since the forces holding these complexes together are quite heterogeneous. Most of the optimum conditions reported in the literature such as buffer p H , and temperature are generally close to normal physiological conditions. In the work described in this chapter, the E L I S A conditions were optimized by investigation of such parameters as coating temperature, incubation time, p H , ionic strength, metallic ions, organic solvent and substrates for enzyme. The effect of the assay formats on the assay sensitivity were also evaluated. These included traditional indirect E L I S A , indirect E L I S A with biotin-streptavidin system and direct E L I S A . A format based on the magnetic beads was also evaluated. The effect of the spacer arm on the sensitivity of the assay was also evaluated by using different combinations of coating antigens and antibodies. Finally the method was assessed by determining the within-assay and between-assay variation and the percentage of false positive and false negative results. Recoveries were checked for different sample matrices spiked with D H A using direct E L I S A . These included assay buffer, tap water and a biotreated B K M E sample. 4.2. Materials and Methods A l l resin acid standards were obtained from Hel ix Biotech Corporation (Richmond, B C , Canada). Carbonate-bicarbonate buffer capsules, phosphate-citrate buffer with sodium perborate capsules, o-phenylenediamine (OPD), tetramethylbenzidine ( T M B ) and Freund's complete and incomplete adjuvants were purchased from Sigma (St. Louis, M O ) . Biotinylated donkey anti-rabbit IgG and goat anti-rabbit IgG, linked with horseradish peroxidase (HRP), were obtained from Amersham International pic, (Amersham, U K ) . Horseradish peroxidase (HRP), streptavidin-linked, was obtained from Southern Biotechnology Associates, Inc. (Birmingham, A L ) . B A C T O dehydrated skim milk was purchased from Difco Laboratories (Detroit, MI ) . De-89 ionized water was purified by the N A N O p u r e ultrapure water system (Barnstead-Themolyne). Phosphate-buffered saline (PBS, 0.15M) was composed of N a t f r P C ^ , N a 2 H P 0 4 , M g C l 2 and N a C l . Amine-terminated supermagnetic beads and Dynal Magnetic Particle Concentrator were purchased from Dynal . The E L I S A plate was the 96-well EVTMULON® 4 (flat bottom, Catalog N o . 01-010-3855) from Dynatech Laboratories Inc. (Chantilly, V A ) and the optical density was read on a T H E R M O m a x T M m i c r o t i t r e plate reader (Molecular Devices Corp., Menlo Park, C A ) . 4.2.1 E L I S A protocols using microtitre plates The major steps of E L I S A were listed below using D H A M - S U C - B S A as the coating antigen and the polyclonal antibodies from D H A M - S U C - K L H as an example. The microtitre plate was washed five times with P B S buffer, between each step, to separate the bound and unbound material on the microtitre plate. The plate was covered by a plastic f i lm to avoid evaporation during incubation. 4.2.1.1. Conventional indirect E L I S A The conventional indirect E L I S A was carried out using the following procedure: Step 1. The microtitre plate was coated with D H A M - S U C - B S A (100 ng/100 ul/well) in carbonate-bicarbonate buffer (pH 9.6) and dried at 37°C overnight. Step 2. Excess binding sites on the microtitre plate were blocked using 2% milk in P B S buffer (200 ul/well) at 37°C for 1 hr. Step 3. Appropriate volume of D H A stock solution in methanol was evaporated in a glass test tube and mixed with polyclonal antibodies from D H A M - S U C - K L H (1:8000, diluted in 0.1% milk in PBS) . The mixture was added to the plate (100 ul/well) and incubated at 37°C for 2 hrs. Step 4. HRP- l inked goat anti rabbit IgG (1:3000 diluted in P B S with 0.1% milk) was added to the plate (100 ul/well) and incubated at 37°C for 1 hr. 90 Step 5. O P D (1.0 mg/ml in citrate-phosphate buffer with sodium perborate, 100 ul/well) was added, the reaction was stopped with sulphuric acid (2.5 M , 50 ul/well) after 15 minutes and O D at 490nm was measured. 4.2.1.2. Indirect E L I S A with biotin-streptavidin system Indirect E L I S A with biotin-streptavidin system was carried out using the following procedure: Step 1. The microtitre plate was coated with D H A M - S U C - B S A (10 ng/100 ul/well) in carbonate-bicarbonate buffer (pH 9.6) and dried at 37°C overnight. Step 2. Excess binding sites on the microtitre plate were blocked using 2% milk in P B S (200 ul/well) at 37°C for 1 hr. Step 3. Appropriate volume of D H A stock solution in methanol was evaporated in a glass test tube and mixed with polyclonal antibodies from D H A M - S U C - K L H (1:40000, diluted in 0.1% milk in PBS) . The mixture was added to the plate (100 pl/well) and incubated at 3 7 ° C f o r 2 h r s . Step 4. Biotinylated donkey anti rabbit IgG (1:12000) was added to the plate (100 ul/well) and incubated at 37°C for 1 hr. Step 5. Streptavidin-HRP (1:12000) was added to the plate (100 pl/well) and incubated at 37°C for 1 hr. Step 6. O P D (1.0 mg/ml in citrate-phosphate buffer with sodium perborate, 100 pl/well) was added, the reaction with stopped using sulphuric acid (2.5 M , 50 pl/well) after 15 minutes and O D at 490 nm was measured. 4.2.1.3. Direct E L I S A Direct E L I S A was carried out using the following procedure: Step 1. The microtitre plate was coated with diluted primary antibodies (1:8000, 100 pl/well) in 91 carbonate-bicarbonate buffer (pH 9.6) and dried at 37°C overnight. Step 2. Excess binding sites on the microtitre plate were blocked using 2% milk in P B S (200 |il/well) at 37°C for 1 hr. Step 3. Appropriate volume of D H A stock solution in methanol was evaporated in a glass test tube and mixed with D H A M - S U C - H R P (1:16000). The mixture was added to the plate (100 ul/well) and incubated at 37°C for 1.0 hrs. Step 4. O P D (1.0 mg/ml in citrate-phosphate buffer with sodium perborate, 100 ul/well) was added, the reaction was stopped with sulfuric acid (2.5 M , 50 ul/well) after 15 minutes and O D at 490 nm was measured. 4.2.2. M a g n e t i c b e a d - b a s e d ELISA 4.2.2.1. P u r i f i c a t i o n o f p o l y c l o n a l an t ibod ies u s i n g the P r o t e i n A m e m b r a n e Antiserum from a rabbit immunized with D H A M - S U C - K L H was purified using protein-A affinity chromatography according to the procedure provided, with slight modifications (Amicon Inc., 1991). 1. Twenty M A C protein A Discs (47 mm) were inserted into the disc holder according to the instructions provided and flushed first with P B S (60 ml), then 0.15 M glycine (pH 2.3, 60 ml), and re-equilibrated with 0.2 M Tris buffer (pH 8.0, 100 ml). 2. Antiserum (10 ml) was filtered through a 0.2 um filter and passed through the M A C system using a peristaltic pump, back and forth several times, at a f low rate of 50 ml/min. 3. Excess sample was removed by washing with 0.2 M Tris buffer until O D 280 nm returns to baseline. 4. Antibodies were eluted by quickly passing 80 m l of glycine buffer (pH 2.3) through the system and fractions (3 m l each) were collected into test tubes containing 125 u l of I M Tris buffer (pH 8.0) until O D 280 reaches baseline. The IgG in each fraction was 92 monitored by measuring the O D at 280 nm and assay using E L I S A . IgG fractions were collected and concentrated using ultrafiltration. 5. The M A C system was regenerated by flushing the discs with P B S until p H reaches neutral, further flushing with 4% acetic acid (60 ml) and washing with P B S until p H back to neutral. The system was finally re-equilibrated in P B S containing 0.02% sodium azide for storage. 4.2.2.2. Coupling of antibodies to tosylactivated magnetic beads The purified polyclonal antibodies were coupled to tosylactivated Dynabeads M-280 according to the procedures provided (Dynal A . S., 1994). The following steps were followed: 1. Magnetic M-280 tosylactivated Dynabeads (10 mg/ml) were concentrated by removing the aliquot and resuspending in 0.05 M borate solution (pH 9.5) to a final concentration of 20 mg/ml. 2. Purified antibodies (IgG) were dissolved in 0.05 M borate solution to a concentration of 0.4 mg/ml, mixed with the concentrated M-280 tosylactivated Dynabeads (1.0 ml), and incubated 24 hrs at room temperature with slow tilting and rotation. 4. Dynabeads were collected by applying a magnetic field using Dynal Magnetic Particle Concentrator, the supernatant was discarded and the beads were washed three times with P B S buffer, 10 minutes each time, by slow tilting and rotation, and finally washed overnight at 4°C. 5. The IgG coated magnetic beads were collected and resuspend in P B S (2.0 ml) to a final concentration of 10 mg/ml. 4.2.2.3. Magnetic beads ELISA The direct E L I S A was performed using antibody coated magnetic beads as the solid support according to the following procedures: 93 1. The antibody-coated magnetic beads (10 mg/ml) were mixed wel l by vortexing, 50 p i of the suspension was withdrawn and resuspend in P B S with 0.1% milk (5.0 ml). 2. Equal volume (500 u l each) of sample in PBS , enzyme tracer (1:20,000 dilution in P B S with 0. P/o milk) and antibody-coated beads were mixed in a test tube (7x96mm), and incubated 30 minutes at 37°C. 3. The beads were separated by applying a magnetic field using Dynal Magnetic Particle Concentrator and decanting the aliquot. The beads were washed with P B S buffer three times. 4. O P D (1.0 m l , 1.0 mg/ml in citrate-phosphate buffer with sodium perborate ) was added to each tube and the reaction was stopped after 15 minutes using 2 M sulphuric acid (0.5 ml). 5. The aliquot'was transferred to a 96-well plastic plate and the O D was measured at 490 nm. 4.3. Results and Discussion 4.3.1. Effects of the assay formats on the ELISA sensitivity It is wel l known that the quality of an immunoassay is primarily determined by the antibodies and the assay sensitivity is limited by the affinity of the antibody for the analyte. However, various assay formats may affect the performance of antibodies due to different binding patterns, thus leading to different selectivities and specificities. Therefore, the absolute amount of material required for 50% inhibition w i l l be a function of the assay format, the size of mass action reaction and the affinity of the antibody for the antigen (Vanderlaan et al, 1990). Generally, the lower the amount of hapten-protein conjugate bound to the solid phase and the lower the amount of antibody used, the less competitor needed for inhibition and the more sensitive the assay. For example, when comparisons were made between direct and indirect E L I S A methods for abscisic acid, a plant growth inhibitor, the indirect E L I S A showed a much higher sensitivity (Belefant and Fong, 1989). In order to see i f different assay formats could improve the sensitivity of the E L I S A for D H A , three common E L I S A formats were evaluated 94 using antibodies from D H A M - S U C - K L H . These were conventional indirect E L I S A , indirect E L I S A with a biotin-streptavidin system and direct E L I S A . A n E L I S A procedure based on the magnetic beads was also evaluated. In our preliminary investigation, the enzyme reaction was stopped 5 minutes after addition of substrate and the O D was then measured. A s the intensity of the colour depends on several factors, such as the concentration of the coating antigen, the dilution of the primary antibody and.the time allowed for the enzyme catalysed reaction, it is important to select suitable conditions from these parameters to achieve an optimal E L I S A performance. The target endpoint, for example, an O D close 1.0, can be achieved by using lower amounts of coating antigen or antibodies i f the enzyme reaction time is extended. This w i l l not only save on the amount of valuable immunoreagents used, but also increase the assay sensitivity since it is a function of the size of the mass action reaction (Vanderlann et al, 1990). Therefore, it was decided to extend the reaction time from 5 minutes used previously to 15 minutes. Times ranging from several minutes to one hour have been reported in the literature and usually 15-30 min is often sufficient (Robins, 1986). 4.3.1.1. Conventional indirect E L I S A Usually the conventional indirect E L I S A uses an enzyme-linked secondary antibody as the signal generator. A two dimensional checkerboard titration was used to determine the suitable concentrations of coating antigen and the dilution of the primary antibodies. W e used an O D reading close to 1.0 (at 490 nm), after 15 minutes of enzymatic reaction, as an arbitrary end point. Since the E L I S A was designed to quantify D H A , primary antibodies should be used in a limited amount and the amount of coating antigen should also be low in order to achieve an effective competition. Therefore, a concentration of coating antigen of 100 ng/100 ul/well and a primary antibody dilution of 1:8000 was chosen. 95 Using these conditions, the competitive indirect E L I S A was performed using D H A as the inhibitor. The standard inhibition curve was determined and the 50% inhibition concentrations (I50) and the detection limits, defined as the amount of analyte required for a 20% inhibition, were calculated from the standard curve and listed in Table 4.1. For the conventional indirect E L I S A , it was found that the I50 and the detection limit were 113.2 and 11.7 pg/L, respectively. 4.3.1.2. Indirect E L I S A using the biotin-streptavidin system The indirect E L I S A with the biotin-streptavidin system was the format used in our preliminary investigation. This format required a biotinylated secondary antibody and an enzyme-linked streptavidin. The assay sensitivity usually improved significantly primarily because the signal had been amplified in this biotin-streptavidin system. A drawback of this format was that one more incubation step was required. Previously, the indirect E L I S A with the biotin-streptavidin system was carried out using 25.0 ng/lOOpl/well of D H A M - S U C - B S A as coating antigen and a 1:30,000 dilution of primary antibodies, while the enzymatic reaction was allowed to proceed for' 5 minutes before it was stopped ( L i et al, 1994). A s expected, when the reaction time was extended from 5 minutes to 15 minutes, a two-dimentional checkerboard titration indicated that the concentrations of coating antigen and antibodies could be further lowered. Only 10.0 ng/lOOpl/well of coating antigen and as high a dilution as 1:40,000 of primary antibodies could be used (Table 4.1). B y using lower concentrations of coating antigen and primary antibodies in conjunction with an extended enzyme-catalyzed reaction time, the assay reagents were saved and, more importantly, the assay sensitivity was improved (I50 of 12.3 pg/L vs 20.2 pg/L reported previously) ( L i et al, 1994). 96 4.3.1.3. Direct ELISA The direct E L I S A requires that antibodies are immobilized on the microtitre plate. Wi th this assay format, a similar two dimensional checkerboard titration could be used to determine the suitable concentrations of antibodies coated on the plate and the dilutions of enzyme tracer used in the incubation. Based on the same endpoint criteria, the polyclonal antibodies from D H A M - S U C - K L H could be diluted to 1:10,000 and the enzyme conjugate ( D H A M - S U C - H R P ) could be used at 1:16000. Under these conditions, an I50 of 49.7 ug/L was obtained (Table 4.1). This result is better than the previously reported I50 value of 261.6 ug/L which was obtained by util izing a higher concentration of coating antibody (1:8000) and enzyme tracer (1:10000), and a short enzyme reaction time (5 minutes) ( L i et al, 1996a). Obviously, the lower the amount of antibodies immobilized on the microtitre plate, the less the amount of enzyme tracer used, the lower the amount of D H A needed for competition. Only one step is required i f the microtitre plate was pre-coated and blocked. The disadvantage is that a hapten-enzyme conjugate is required for this format and care must be taken in its synthesis to avoid enzyme denaturation or deactivation. Table 4.1 Conditions used for different E L I S A formats and the corresponding I 50 and detection limits obtained Assay Coating material Antiserum 150 Detection format dilution (Ug/L)* l imit (ug/L)* Conventional D H A M - S U C - B S A 1:8,000 113.2 + 22.5 11.7 + 2.6 indirect E L I S A (lOOng/100 ul/well) Indirect E L I S A with D H A M - S U C - B S A 1:40,000 12.3 ± 2 . 1 1.9 + 0.3 biotin-streptavidin (lOng/100 ul/well) Polyclonal antibodies D H A M - S U C - H R P 49.7 + 7.9 4.5 + 1.3 Direct E L I S A (1:10,000, 100 ul/well) 1:16,000 * average of three determinations, each with eight replicate wells 97 Standard curves for the three E L I S A formats described previously are shown in Figure 4.1. It was obvious that the biotin-streptavidin system was the most sensitive of the three E L I S A formats. For both the direct and indirect E L I S A , the assay sensitivity was improved significantly by using less coating material and an extended enzyme-catalyzed reaction. The different sensitivities observed from the direct and indirect E L I S A (I50 of 49.7 pg/L vs 113.2 pg/L) may be due to the different reaction kinetics. In the indirect E L I S A , the coating antigen was immobilized on the microtitre plate and the competition occurs between the immobilized antigen and the free hapten in solution for limited amount of antibodies. However, in the case of the direct E L I S A , primary antibodies are immobilized on the microtitre plate and the competition occurs between the free hapten and the hapten-enzyme tracer, which are both in solution. Thus, this process may proceed more effectively. Although the direct E L I S A does not seem to be as sensitive as the indirect E L I S A with the biotin-streptavidin system, it requires the least steps and the shortest assay time of all of the three formats evaluated. These two characteristics are important for environmental analysis since large number of samples are usually analyzed and automation requires analytical procedures to be as simple as possible. Although the conventional indirect E L I S A showed the highest I50, the linear range was wider than obtained with the other two formats. This format could be used in cases where low sensitivity is required. Therefore, it is evident that assay formats play a significant role in the determination of assay sensitivity. The direct E L I S A showed promising results and had some advantages such as, simple assay format and short assay time. This format is widely used in environmental assays and many field-portable immunoassay kits are based on this format (Van Emon and Gerlack, 1995a). Although the direct E L I S A was not as sensitive as the indirect E L I S A with the biotin-streptavidin system in this study, alternative strategies could have been used to increase the assay sensitivity of the direct E L I S A . For example, the use of IgG or protein A as pre-coating to ensure the correct orientation of antibodies have been reported (Schneider and Hammock, 1992) and the use of a fluorogenic substrate instead of a chromogenic one may also increase the sensitivity of the E L I S A . 98 Inhibition% Figure 4.1 Standard curves obtained from different ELISA formats A: indirect ELISA B: indirect ELISA with the biotin-streptavidin system C: direct ELISA (Error bars indicate the standard deviation, n=8). 99 4.3.1.4. ELISA using magnetic beads as support Recently, magnetic beads have been successfully applied as a solid support in E L I S A s used for the detection of toxic chemicals in water and soil ( L a w r u k ^ al, 1995). Magnetic bead E L I S A s employ a direct format and require beads coated with antibodies. The linkage between the antibodies and the magnetic beads can be either through physical adsorption or covalent bonding. Since antiserum contains polyclonal antibodies together with other serum proteins, it is necessary to first isolate the IgG fraction and then to subsequently link the IgG to the magnetic beads. The antiserum from rabbits immunized with D H A M - S U C - K L H was first purified using protein A membrane affinity chromatography. This is a fast system which permits purification of IgG in a single step (Malakian et al, 1993). Due to the specific binding of protein A to IgG and the special design of M A C Protein-A Discs for fast f low (50 ml/min), efficient purification was achieved in a short period of time. Both U V and E L I S A were used to monitor the fractions of eluents from the M a c discs (Figure 4.2). After combining all of the IgG fractions, the polyclonal antibodies were concentrated using ultrafiltration. A concentration of 1.48 mg/ml was achieved, as determined by U V at 280 nm. After purification, the polyclonal antibodies, mainly IgG, were coupled to the activated magnetic beads, which are termed tosylactivated Dynabeads (Figure 4.3). The surface of the beads have been activated by p-toluenesulfonyl chloride treatment and therefore, ligand containing amino groups w i l l react with beads to form a covalent bond between bead and ligand. The reaction was performed in borate buffer. Since the antibodies are directly attached to beads which act as the solid support, the coating and blocking steps normally required in plate E L I S A can be omitted. Similar to the direct E L I S A , magnetic bead E L I S A requires only one incubation step involving sample solution, enzyme tracer and antibody-coated beads. The I50 and the detection limit for D H A were 139.1 100 OD@280nm for protein assay OD@490nm in ELISA 1.0 H 0.8 0.6 H 0.4 0.2 0.0 — • — OD@490nm — ° — OD@280nm 1.2 1.0 0.8 0.6 0.4 H 0.2 0.0 Fraction number Figure 4.2 Monitoring of IgG using U V and E L I S A for fractions collected in t purification of polyclonal antibodies using M A C protein-A discs . WM>—OH + BEAD)—O-S—(Cj)—Cli 3 (Tosylated bead) ft ^ NH2—protein O M D ) — N H — p r o t e i n + H 0 _ | — ( ^ ^ ) ~ C H 3 Figure 4.3 Chemistry of Dynabead activation and binding of protein to beads 101 pg/L and 3.5 ug/L, respectively (Figure 4.4). Although the I50 of the magnetic bead based E L I S A was higher than that in the plate E L I S A (49.7 pg/L), the linear range of the standard curve was wider. The higher I50 could be due to the high concentrations of antibodies coated on the beads. This was indicated by the low concentration of enzyme tracer (1:20,000) used in this format in order to achieve a final O D reading of 1.0 after 15 minutes of enzymatic reaction. Another advantage of using the magnetic bead E L I S A was that an incubation time as short as 30 minutes was required and the assay could be finished in 45 minutes. Inhibition% D H A concentration (ppb) Figure 4.4 A typical standard curve of magnetic bead E L I S A Therefore, there is a clear indication that various assay sensitivities can be achieved by using different assay formats. This may provide an easy way to quantify a broad range of resin acid concentrations from different sources. Among the four E L I S A formats evaluated, the indirect E L I S A with the'biotin-streptavidin system gave the best results in terms of I50 and 102 detection limit. The direct E L I S A was less sensitive than the indirect E L I S A with the biotin-streptavidin system. However, it required less assay steps and was faster. The magnetic bead E L I S A required the shortest assay time and showed a broader linear range for quantification. However, its I50 was higher than that of the direct E L I S A . It should be realized that, under a particular assay format, different sensitivities can also be achieved when using different amounts of coating antigens or different dilutions of antibodies. 4.3.2. Effect of a spacer arm on E L I S A A s discussed previously, a spacer arm in the immunizing antigen can be important in the elicitation of antibody production (Harrison et al, 1990). Wi th two sets of polyclonal antibodies derived from antigens with and without a spacer arm, both direct and indirect E L I S A were used to evaluate the effect of a spacer arm on the assay sensitivity and cross-reactivity. 4.3.2.1. Effects of a spacer arm on antibody production It has been generally recognized that the determinant groups of the hapten molecule in a hapten-protein conjugate must not be masked i f effective antibodies against a particular hapten are to be raised. A n y masking caused by the carrier's spatial conformation may lead to a deleterious effect on the production of antibodies against the hapten. The visibility of the hapten depends largely on the mode of conjugation. A small molecule that is directly linked to some site on a protein may suffer considerable masking in the region of the hapten nearest the site of linkage. The antisera raised would suffer a lack of specificity for those determinant groups that have been masked. The most obvious means of avoiding such a problem would be to physically extend the hapten out into space via a spacer arm. The effect of a spacer arm is best exemplified in the development of an R I A for diphenylhydantoin (Paxton et al, 1976). Two conjugates were used as antigens, one containing an acetyl group between the hapten and carrier protein and the 103 other a pentanoyl. Both were attached to the same site on the ethyleneurea nucleus. It was found that the antigen with the shorter spacer arm failed to elicit antibodies that could recognize free diphenylhydantoin, while the one with the longer spacer arm produced antibodies of high titre and specificity against the free hapten. However, our immunization results using antigens with and without a spacer arm indicated that polyclonal antibodies with high titre could be raised in both cases. The polyclonal antibodies could be used at more than 1:40,000 and 1:25,000 dilution, respectively, when either D H A M - S U C - B S A or D H A - B S A was used as a coating antigen in the indirect E L I S A with a biotin-streptavidin system (Table 4.2). Therefore, it was probable that in both cases, the hapten in the immunizing antigens was not masked by the carrier protein. This result implied that, in our case, the spacer arm was not an important factor in the elicitation of antibody production. This could be because the carboxylic group of the hapten was used to react with a lysine terminal of the carrier protein, which should be on the surface of the protein and may act as a spacer arm itself. 4.3.2.2. Effects of spacer arm on assay sensitivity A s discussed previously, a hapten should not be masked in the immunizing antigen so that antibodies against that hapten can be raised. It is also true that the hapten in the coating antigen (for indirect E L I S A ) or in the enzyme tracer (for direct E L I S A ) should not be masked, otherwise the antibodies w i l l be unable to bind to the hapten. Although polyclonal antibodies have been raised from immunizing antigens with and without a spacer arm, their reactivities to resin acids may be different. Wi th these two sets of polyclonal antibodies obtained from rabbits immunized with D H A M - S U C - K L H and D H A - K L H in hand, together with two coating antigens D H A M - S U C - B S A and D H A - B S A , it was possible to use different combinations of antibodies and coating antigens in E L I S A to see i f the assay sensitivities could be improved. Therefore, a checkerboard titration was again used to determine the suitable amount of coating antigen and 104 antibody dilution which could be used. When an indirect E L I S A with the biotin-streptavidin system was used, it was found that the amount of antigen ( D H A - B S A ) coated on the microtitre plate should be around 2.5 times that of D H A M - S U C - B S A , to reach a similar O D reading (Table 4.2). Since the hapten densities for these two antigens are similar (8 in D H A M - S U C - B S A and 6 in D H A - B S A , see Table 3.1, Chapter 3), it is probable that some of the hapten molecule in the D H A - B S A conjugate was somehow masked. The antiserum dilution, the 50% inhibition concentration (I50) and the detection limit (I2Q) are also listed in Table 4.2. Table 4.2 Comparison of results obtained in the indirect E L I S A with biotin-streptavidin system using different antibodies and different coating antigens Entry Coating antigen Antiserum 50% inhibition Detection limit (amount) dilution l 5 0 ( p g / L ) c I2o(Hg/L)c 1 D H A M - S U C - B S A (lOng/100 pl/well) 1:40,000 a 12.3+2.7 1.8 + 0.3 2 D H A - B S A (25 ng/100 pl/well) 1:40,000 a 35.8 + 4.5 3.8 + 0.6 3 D H A M - S U C - B S A (10 ng/100 pl/well) 1:25,000 ° 15.6 + 2.4 1.6 + 0.4 4 D H A - B S A (25 ng/100 pl/well) 1:25,000 b 25.0 + 3.1 2.2 + 0.4 a) Antiserum obtained from rabbit immunized with D H A M - S U C - K L H b) Antiserum obtained from rabbit immunized with D H A - K L H c) Average of three determinations. The 8-point standard curves were generated with six replicate wells for each point 105 The data in Table 4.2 showed that, among four possible combinations of coating antigens and antibodies, the I50 for D H A as determined by indirect E L I S A using the biotin-streptavidin system varied from 12.3 to 35.8 ug/L. This indicated that both coating antigens and antibodies had some effect on the assay sensitivity, although these effects seemed not as big as the assay formats did. The direct E L I S A was also used to evaluate the effect of the spacer arm on the assay sensitivity. Since only D H A M - S U C - H R P was available, a comparison was made between the two polyclonal antibodies from D H A M - S U C - K L H and D H A - K L H . The assay showed very similar sensitivities and the I50 was 49.7 and 52.4 pg/L, respectively (Table 4.3). Therefore, it was probable that the spacer arm did not play a significant role in the elicitation of antibody production and the characteristics of the antibodies towards D H A were very similar, based on the results from both indirect and direct E L I S A . Table 4.3 Comparison of sensitivity in the direct E L I S A using polyclonal antibodies from D H A M - S U C - K L H and D H A - K L H Coating A b obtained from Enzyme dilution ( D H A M - S U C - H R P ) 50% inhibition* 150 ( 1 ^ ) Detection l imit* 120 Og/L) D H A M - S U C - K L H (1:10,000, 100 p:l/Well) 1:16,000 49.7 ± 7 . 9 4.5 ± 1 . 3 D H A - K L H (1:8,000, 100 ul/well) 1:16,000 52.4 ± 7 . 4 4.8 ± 1 . 9 * Average of three determinations. The 8-point standard curves were generated with six replicate wells for each point 106 A m o n g the several assay formats tested using two polyclonal antibodies, the direct E L I S A was considered to be the method of choice for the quantification of D H A since it required the fewest assay steps and the sensitivity was reasonably good. Therefore, all subsequent work was done using the direct E L I S A and the antibodies from rabbits immunized with D H A M S U C - K L H , unless it is otherwise specified. 4.3.2.3. Effects of spacer arm on the specificity of the antibodies Previously we had shown that polyclonal antibodies from D H A M - S U C - K L H cross-reacted with other resin acids because of their structural similarity ( L i et al, 1994). To compare the cross-reactivity of two polyclonal antibodies, eight common resin acids and two chlorinated D H A s were used as inhibitors in the inhibition study. To simplify the procedure, the concentration of the inhibitor was kept the same (50.0 pg/L) and the direct E L I S A was employed due to its simplicity. N o significant difference was found between the two polyclonal antibodies in terms of their inhibition patterns (Figure 4.5). This again confirmed that the spacer arm did not appear to play a significant role in the production of polyclonal antibodies and that the antibodies raised were of a similar character. This could be because the haptens ( D H A and D H A M ) used in the synthesis of antigens were similar in structure and the position of the linkage was the same. Both polyclonal antibodies showed high cross-reactivities to abietic type resin acids. Cross-reactivities with pimaric type resin acids were considerably lower. 107 Inhibition% DHA C12DHA C1DHA ABA LEV PAL NEO PIM ISO SAND Figure 4.5 Percentage of inhibition in the direct E L I S A using antisera from D H A M - S U C - K L H and D H A - K L H , with 50.0 ug/L of inhibitors 4.3.2.4. C r o s s r e a c t i v i t y o f an t i bod ies to c o m p o u n d s o the r t h a n res in ac ids In addition to the common resin acids, some compounds with structures similar to resin acids were also used as inhibitors for the evaluation of two polyclonal antibodies. Direct E L I S A was again used. A l l of the compounds were tested at the 1500 p,g/L level, except for podocarpic acid and 4-isobenzoic acid which were tested at the 1000 pg/L level. It was found that podocarpic acid did not show any significant cross-reactivity with the two polyclonal antibodies, possibly because it lacked the isopropyl moiety in the structure, which is considered to be the epitope for antibody binding (Table 4.4). Some phytosterols such as P-sitosterol and stigmastanol, which are usually present in the wood furnish and can also be found in pulp m i l l effluents, were evaluated as potential cross-reactants. Cholesterol and androstenedione were also 108 Table 4.4 Cross-reactivities determined by antibodies with and without spacer arm for some compounds with similar structures to resin acids Chemical Name Chemical Structure A b ( D H A M - S U C - K L H ) A b ( D H A - K L H ) I50 (pmol/L) C R (%) I50 (pmol/L) C R (%) D H A COOH 0.16 100% 0.17 100% Podocarpic acid OH •"' COOH >3.65 <4.4% >3.65 <4.6% p-sitosterol H O ' ^ - ^ \ >3.61 <4.4% >3.61 <4.7% stigmastanol H O ^ - i >3.60 <4.4% >3.60 <4.7% Cholesterol >3.88 <4.1% >3.88 <4.4% Androstenedione >5.24 <3.1% >5.24 <3.2% Retene >6.41 <2.5% >6.41 <2.7% 4-isopropyl-benzoic acid HOOC >6.41 <2.5% >6.41 <2.7% 109 tested for the same purpose. However, we found that none of these compounds cross-reacted with the antibodies to any great extent. This result was expected since the carbon skeletons of these compounds were different from those of the various resin acids. Retene, one of the compounds isolated as a biotransformation product from pulp mi l l effluent (Judd et al, 1995), also showed little cross-reactivity. This could be due to its flat conformation, which is quite different from resin acids despite sharing a similar three-ring carbon skeleton. It is interesting to note that 4-isopropylbenzoic acid did not cross-react with the polyclonal antibodies. This implied that the epitope recognized by the antibodies is probably a moiety bigger than an isopropyl group. Although both antibodies showed little cross-reactivity to the compounds listed in Table 4.4, it is highly likely that some resin acid-related terpenoid neutrals, such as dehydroabietyl alcohol, aldehyde and other resin acid degradation intermediates, may show considerable cross-reactivity with polyclonal antibodies. Therefore, it is possible that the presence of these compounds may affect the E L I S A of resin acids although their concentrations in the pulp m i l l effluents are low. Evaluation of possible interference by these compounds should be done, once purified and characterized compounds are readily available. 4.3.3. Evaluation of assay conditions in the direct E L I S A While a two dimensional checkerboard titration can be used to determine the suitable concentration of coating antigen and dilutions of the antibodies:, which are the two important parameters determining the assay sensitivity, other parameters such as coating temperature, incubation times, p H and ionic strength of the samples may also affect the sensitivity, accuracy and precision of E L I S A . Therefore, all of these conditions should be optimized or compromised in such a way that antibodies are used economically and the required sensitivity is achieved. Direct E L I S A was again used due to the simplicity of this assay format. The antibodies used were derived from either the D H A M - S U C - K L H or D H A - K L H antigens. 110 4.3.3.1. Coating temperature Depending on the assay format, either the antigen or antibody can be immobilized on the microtitre plate. Immobilization should be carried out in such a way that the reactivity of the immunological components is strongly retained. Several factors such as the reactant concentration, coating time, temperature and p H need to be considered and all of these conditions should be strictly adhered to i f reproducible results are to be obtained. The antigen was always coated on the microtitre plate in the indirect E L I S A using carbonate-bicarbonate buffer (pH 9.6) at 37°C overnight. However, different temperatures were investigated for the application of the coating antibody on the microtitre plate in the direct E L I S A . Most of the papers in the literature have adopted the same coating conditions, which involve incubation of the antibody in carbonate-bicarbonate buffer (pH 9.6) at 4°C overnight. Incubation at 37°C for 4 hrs has also been used in some cases. It was believed that these conditions could avoid or minimize the possibility of deactivation or denaturation of the antibodies. In order to evaluate the effect of temperature on coating efficiency and the subsequent E L I S A result, three different coating temperatures of 37°C, room temperature (23 °C) and 4°C were employed during the immobilization of the antibodies (derived from rabbits immunized with D H A M - S U C - K L H ) . It was found that the O D readings were temperature dependent and they increased as the temperature increased. Incubation at 37°C overnight was more effective than incubation at 4°C overnight or at room temperature overnight. Final O D readings of 0.920, 0.653 and 0.354 were achieved for plates coated at 37°C, room temperature (23°C) and 4°C, respectively. Obviously, in the.case of 37°C incubation, more antibodies were immobilized on the microtitre plate. It seems that antibodies were stable and could tolerate 37°C overnight (-17 hrs). It should be noted that the coating buffer was evaporated to dryness after incubation at 37°C overnight. However, in the other two cases, the coating buffer had not been evaporated completely and some antibodies were still in solution and were subsequently washed away. This resulted in less 111 antibodies coated on the plate and may explain the low O D values observed. Therefore, in further work we placed a plastic sealer on the microtitre plate during incubation at 37°C to try to avoid evaporation. It was found that the O D reading for the unsealed plate was almost doubled compared to the values obtained with a sealed plate. This indicated that better immobilization was achieved when the plate was not sealed. A large variation in O D was also found with the sealed plate (Table 4.5). The O D reading was even lower than that obtained with the microtitre plate coated at 4°C. The reason was not clear. It is possible that the adsorption process is not favoured in solution at 37°C and that dissociation occurs at the same time. Table 4.5 Comparison of O D readings obtained from an inhibition study using polyclonal antibodies from D H A M - S U C - K L H and different concentration of D H A with the microtitre plates sealed and unsealed at coating stage in direct E L I S A D H A (Ug/L) Unsealed Sealed Mean O D * s.d. C V (%) Mean O D * s.d. C V (%) 125.0 0.306 0.020 6.6 0.040 0.022 53.7 62.5 0.374 0.017 4.5 0.109 0.035 31.8 31.3 0.505 0.025 5.0 0.147 0.028 19.0 15.6 0.610 0.045 7.4 0.168 0.061 36.0 7.8 0.718 0.055 7.7 . 0.195 0.071 36.5 0 0.920 0.050 5.5 0.248 0.086 34.7 * average of eight replicate wells. Since the adsorption is a random process, antibodies may be attached to the plate with different orientations, thus reducing the chances for the formation of antibody-antigen complex since the antigen binding sites (Fab) may have been blocked. It has been suggested that protein A or anti-rabbit IgG can be used as a precoating agent in the direct E L I S A . This would ensure that the primary antibodies bind to these materials exclusively using the Fc portion, thus a better 112 orientation of the Fab portion could be achieved. However, in our case, precoating with protein A or goat anti rabbit IgG did not show any advantages (data not shown). Direct immobilization of antibodies to microtitre plates without a protein A or an IgG precoating are popular. For example, excellent results were achieved for soil and water samples in the quantification of atrazine using an E L I S A with direct immobilization of antibodies (Wittmann and Hock, 1989; 1990). Due to the good immobilization observed, all subsequent coatings with polyclonal antibodies were directly performed on the microtitre plate at 37°C overnight without using a plastic sealer. 4.3.3.2. Incubation time for competition step Although the total time required by an assay is largely determined by the time needed for each incubation step, the final colour intensity is usually proportional to the amount of enzyme tracer bound to the microtitre plate. If the incubation time is too short, the final colour development step may not lead to a measurable signal due to inadequate amounts of enzyme tracer. However, the E L I S A is a competitive process and the antibody-antigen interaction itself is an equilibrium reaction, therefore, the time allowed for such competition w i l l not only affect the assay time, it w i l l also affect the assay sensitivity. Initially, the time allowed for the competition reaction was set for 2.0 hrs, as described in most of the assay protocols found in literature. To evaluate the effect of the incubation time on the assay sensitivity, the direct E L I S A using polyclonal antibodies from D H A M - S U C - K L H was used and incubation times from 45 to 120 minutes were compared. It was found that longer incubation times favoured colour development and higher O D readings were obtained (Table 4.6). The O D readings for samples containing no D H A (zero sample) increased from 0.59 to 1.34 when the incubation time was extended from 45 to 120 minutes. Obviously, in the latter case, more enzyme tracer had been bound to the antibodies immobilized on the microtitre plate. However, the I50 determination showed that the shorter incubation time resulted in better assay sensitivity. This is probably 113 because antibodies usually have a higher affinity for the hapten-protein conjugate, such as haptenated enzyme tracer, than for the free hapten. Thus, when a longer incubation time is involved, more enzyme tracer w i l l bind to the antibody and the free hapten must be present in a higher concentration to compete effectively with the enzyme tracer. It is also possible that the antibody-hapten complex could dissociate faster than the antibody-enzyme complex, thus leading to unsuccessful competition. Although 45 min of incubation time gave the highest inhibition, the final O D reading for the zero sample was relatively low (0.6). W e prefered to aim for a final O D reading of between 0.8-1.2. Therefore, a 60 min incubation was adopted and used in all subsequent direct E L I S A s . Table 4.6 The effect of the incubation time of the competition step on the O D reading and the 50% inhibition concentrations in the direct E L I S A using antibodies from D F L A M - S U C - K L H Incubation time (min) 45 60 90 120 O D at 490 nm for zero sample * 0.59 0.98 1.21 1.34 150 (Ug/L) 48.6 52.4 67.8 75.6 * average of eight replicate wells, CV<8% 4.3.3.3. p H of the sample Although the p H of the assay buffer (PBS) was always 7.4, the p H of the actual sample may differ. In the direct E L I S A , since the sample is mixed with the enzyme tracer diluted in assay buffer, this may change the p H of the assay system and affect the performance of the antibody and the antibody-hapten complex. Therefore, the p H of the P B S was adjusted to 5.5, 6.5, 7.4, 8.5, 9.5 and 10.5 using diluted HC1 or N a O H , and used as blank samples in the direct E L I S A . It was found that the final O D values were constant between p H 6.5 to 9.5. However, at pHs below 6.5 and above 9.5, the O D readings were considerably lower (Figure 4.6). In these 114 cases, the buffering capacity of P B S may have been exceeded. Since the hapten-enzyme conjugate and sample were incubated together, the final p H of the system was important as the enzyme could be deactivated or denatured at extreme pHs. The influence of p H on the E L I S A also depends on the antibodies that are used. It has been shown that, although some antibodies are sensitive to p H , other antibodies work equally wel l over a wide p H range (Kamata et al, 1996). OD@490nm Figure 4.6 The effect of sample p H on the final O D readings in the direct E L I S A using polyclonal antibodies from D H A M - S U C - K L H 4.3.3.4. Ionic strength of the sample In addition to the p H , the ionic strength of the sample is an another factor which may affect the assay. Although the assay buffer (PBS) was used at a concentration of 0.15 M , a range of 0.05 to 1.5 M was examined in the direct E L I S A . It was found that P B S over a concentration range of 0.05 to 0.4 M did not affect the assay significantly (Figure 4.7). However, higher concentrations did affect the O D reading. It seems that the functions of either the antibody or enzyme could have been influenced by this high ionic strength. Similar to p H , the influence of ionic strength on the E L I S A depends on the antibodies used. Some antibodies are very 115 susceptible to changing ionic strength while others may tolerate a wide range of ionic strengths (Kamatae/a/ . , 1996). OD@490nm 0.05 0.10 0.15 0.20 0.40 0.80 1.00 1.50M Figure 4.7 The effect of ionic strength of the sample on the O D readings in the direct E L I S A using polyclonal antibodies from D H A M - S U C - K L H 4.3.3.5. Effect of metallic ions Various metallic ions may be present in different sample matrices. For example, ions< derived from both alkaline earth metal and heavy metals are present in pulp m i l l effluents (Springer, 1993). These ions may interfere with the assay. To evaluate the effects of metallic ions on the E L I S A , some examples of common metallic ions found in C T M P effluents, such as N a + , K + , C a 2 4 " , M g 2 + , Z n 2 + a n d M n 2 + , were added to de-ionized water and assayed by the direct E L I S A . It was found that, at the tested concentrations (200 m M for N a + and 50 m M for other ions), none of the six ions affected the assay. The percentage of residual O D was higher than 85% (Figure 4.8). The recovery of samples spiked with 50.0 ug/L of D H A in the presence of the various individual ions was greater than 91% (individual recovery was not shown), indicating that these ions did not affect either the sensitivity or colour development of the assay. 116 OD@490nm Figure 4.8 The effect of metallic ions on the O D readings in the direct E L I S A using polyclonal antibodies from D H A M - S U C - K L H 4 . 3 . 3 . 6 . The effect of organic solvents on the E L I S A Due to the low detection limit of the E L I S A , no organic solvents were required to help facilitate the solubilization of either the standard or samples in the assay buffer when the E L I S A was performed. Usually, the D H A standard stock solution was prepared in methanol and an appropriate amount of D H A was taken from this stock solution, evaporated to dryness using a stream of nitrogen and then re-dissolved in assay buffer for use in E L I S A . To see i f we could omit this evaporation step, the effect of methanol on the performance of antibodies was evaluated by both the indirect and direct E L I S A s . It was found that the antibodies derived from D H A M - S U C - K L H could tolerate up to 10% methanol, while the antibodies derived from D H A -K L H could tolerate only around 2% methanol when the direct E L I S A format was used (Figure 4.9). The indirect E L I S A was also used to evaluate the effect of methanol and it was found that both the antibodies from D H A M - S U C - K L H and D H A - K L H could tolerate methanol concentrations of up to 5%. It seems that the effect of organic solvents on the assay varied from antibody to antibody. Antibodies capable of tolerating up to 30% methanol have previously been reported (Kramer et al, 1994). Therefore, the D H A stock solution in methanol could be used directly in the E L I S A without evaporation since the methanol content in the sample was kept wel l below 0.5% (v/v). 117 Residual activity % 120-. : m-\—i—i——i—i—i—i—i—i—i—i—i—i—i 0 0.1 0.2 0.5 1 2 5 10 20 30 40 50 Methanol in sample % (v/v) Figure 4.9 The effects of methanol on the direct E L I S A using antibodies derived from D H A - K L H (•) and D H A M - S U C - K L H (•) 4.3.3.7. Effect of substrate on E L I S A The performance of two chromogenic substrates for horseradish peroxidase, o-phenylenediamine (OPD) and 3, 3', 5, 5'-tetramethylbenzidine ( T M B ) , in E L I S A were evaluated. O P D was converted to an orange product after the reaction was stopped by sulphuric acid and the O D was monitored at 490 nm. T M B was converted to a blue product and O D was measured at 605 nm. The blue product turned yellow after acidification and the O D could be monitored at 450 nm. A t the same substrate concentration (1.0 mg/ml) and reaction time (15 min), the O P D resulted in a higher O D reading than the T M B when antibodies from D H A M - S U C - K L H were used in the direct E L I S A (Table 4.7). However, no significant difference was observed in the percentage of inhibition and I50 (45.8 ug/L for O P D vs. 43.4 ug/L for T M B ) . This result was similar to those obtained by others evaluating the substrates (Al-Kaiss i and Mostratos, 1983). Both O P D and T M B are commercially available as tablets and are easy to use. However, an 118 organic solvent such as D M S O is needed to dissolve T M B due to its low solubility in citrate-phosphate buffer. Therefore, O P D was used as the enzyme substrate in our subsequent studies. Table 4.7 Comparison of the final O D readings when O P D and T M B were used as H R P substrates in a direct E L I S A using polyclonal antibodies from D H A M - S U C - K L H D H A (Ug/L) O D 4 9 0 for O P D * Inhibition (%) O D 6 0 5 for T M B * Inhibition (%) 125.0 0.334 67.2 0.277 68.5 62.5 0.458 55.1 0.383 56.4 31.3 0.587 42.4 0.506 42.4 15.6 0.680 33.3 0.590 32.9 7.8 0.817 19.8 0.684 22.2 0 1.019 (150 = 45.8 ug/L) 0.879 ( I 5 0 = 43.4 ug/L) * average of eight replicate wells, with CV<10%. 4.3.4. M e t h o d e v a l u a t i o n 4.3.4.1. W i t h i n - a s s a y a n d between -assay v a r i a t i o n In E L I S A s , wel l to wel l variations (within-assay), and plate to plate variations (between-assay) are frequently observed. The so-called within-assay variation may reflect several factors involved in the assay, such as variation in the wells of the plate (usually within a C V of 4% as claimed by the supplier), the variation in pipetting and the uneven mixing of the assay reagents. The analytical variability within the assay can usually be controlled below 10% (coefficient of variance, C V ) for O D readings i f the assay is performed with care. Large variations are easily noticed as most E L I S A software w i l l generate good statistical data. However, i f a comparison is made between assays which are performed on different plates at different times, the total variation w i l l be larger. In this case, the total variation observed can be affected by both within-119 assay and between-assay variations. The latter variation usually contributes more than the former since more factors tend to be involved in between-assay variations, such as plate to plate and batch to batch variation in wells, and the day to day variations in assay conditions. It has also been observed that, while O D readings may fluctuate, the percentage inhibition is usually more stable. Therefore, a comparison of variation was performed by determining the percentage of inhibition instead of the O D readings. In order to determine the total variation, including within-assay and between-assay variations, the direct E L I S A s were performed using antibodies from D H A M - S U C - K L H over four consecutive days using identical protocols. D H A at concentrations ranging from 7.8 to 125.0 pg/L was used as the inhibitor. The analysis of variance ( A N O V A ) indicated that between-assay variation was a more important factor than within-assay variation (Bookbinder and Panosian, 1986; Rosner, 1986). The total variation was less than 11% (Table 4.8). Generally, the immunoassay precision ranges from 5-20% (coefficients of variation for replicate analysis run in different assays). However, the C V can be much higher, particularly when antigen levels were determined using the lower end of standard curves (Weiler, 1990). Table 4.8 Analysis of variance ( A N O V A ) for within-assay and between-assay variation using standard in the direct E L I S A performed over four consecutive days Point 1 125.0 pg/L Point 2 62.5 pg/L Point 3 31.3 pg/L Point 4 15.6 pg/L Point 5 7.8 pg/L Replica 8 8 8 8 8 Plates 4 4 4 4 4 Data point 32 32 32 32 32 %inhibit ion (mean) 60.3 52.6 44.9 32.7 25.0 C V (within) 3.2% 6.1% 3.6% 4.2% 6.4% C V (between) 8.0% 7.5% 6.4% 8.2% 8.8% C V (total) 8.6% 9.7% 7.3% 9.2% 10.9% 120 The variation in I50 values obtained from different determinations is also an indicator of assay precision. The I50 values were calculated from four standard curves and listed in Table 4.9. Although both the intercepts and the slopes of the standard curves from different days varied, the I50 varied only slightly and the C V was 6.5%. However, a total of 16 I50 values determined in a period of one month indicated that inter-assay variation could reach 17.5% (data not shown), which is considerably greater than the typical within-assay variation (usually less than 10%)). Therefore, a standard curve was created every time an assay was performed. This practice has also been recommended in E P A s " A User's Guide to Environmental Immunochemical Analysis" (Gee et al, 1994). If a given plate is subject to differences in manipulation time, temperature of incubation or other factors which may affect the equilibrium, the samples on that plate can still be compared to a calibration curve subjected to those same variables. Table 4.9 Comparison of the standard curves and I50 obtained over four consecutive days Plate N o . Standard curve R I 5 0 (ug/L) 1 Y % = -2.93 + 30.03 L o g X 0.991 57.9 2 Y % =-1.16 +29.55 L o g X 0.990 53.9 3 Y % = 3.85 + 26.03 L o g X 0.991 59.3 4 Y % = 2.41 +27.82 L o g X 0.993 51.4 55.6 ± 3 . 6 ( C V = 6.5%) It should be noted that under manual operation, a high C V in a typical E L I S A is not unusual. Provided the coefficient of variation values are not greater than 15%, 20% and 50% for within-assay, between-assay and between-laboratory assays, respectively, the accuracy of the assay is considered acceptable (Heitzman, 1988). 121 4.3.4.2. Recovery of D H A from spiked matrices using direct E L I S A In order to evaluate the E L I S A , various matrices were spiked with different amounts of D H A and assayed by the direct E L I S A using antibodies from D H A M - S U C - K L H . The matrices included assay buffer, tap water and a biotreated bleached kraft m i l l effluent ( B K M E ) which had been previously shown that resin acids were undetectable by traditional G C analysis. The E L I S A procedures were slightly modified so that equal volumes of the enzyme tracer in assay buffer (1:8000 in 0.1% milk in PBS) and the sample solution ( D H A in PBS) were mixed in a test tube prior to being added to the microtitre plate and incubated. This maintained the final dilution of enzyme tracer to 1:16000, as used previously. The first sample candidate was P B S , which was the buffer system used in previous E L I S A development work. P B S was spiked with 10.0 to 150.0 pg/L of D H A and left at 4°C overnight. When assayed using the direct E L I S A , recoveries ranging from 86 to 102% with low percentage of variation were obtained (Table 4.10). This result was expected since this is the simplest case and the chemical components in the .sample and in the standard were identical. Therefore, no matrix effect should present. The next matrix tested was tap water. Similarly, 10.0 to 150.0 pg/L of D H A was spiked in tap water and kept at 4 °C overnight. When assayed by the direct E L I S A using antibodies from D H A M - S U C - K L H , recoveries ranged from 87% to 98%, indicating that no matrix effect was encountered (Table 4.11). This was expected as tap water is generally clean and should not contain any interfering materials. A biotreated bleached kraft m i l l effluent ( B K M E ) sample, which contained no resin acids, within the G C detection limit of less tahn 0.5 pg/L, was next spiked with D H A and then analyzed by the direct E L I S A . It was found that recoveries ranged from 103 to 122% (Table 4.12). However, an average value of 2.2 pg/L of D H A was found in the control. If this was considered as a background value and subtracted from the recovered D H A , the results are more 122 realistic. They ranged from 78 to 105%. Poor recovery was observed at low spiked concentrations, possibly due to being close to the detection limit of the assay. Table 4.10 Recovery of DFIA from spiked assay buffer using antibodies from D F I A M - S U C -K L H in a direct E L I S A D H A spiked in Recovered Recovery sample(ug/L) (Ug/L)* (%) 150.0 141.3 ± 9 . 7 94.2 ± 6 . 5 100.0 93.7 ± 8 . 8 93.7 ± 8 . 8 75.0 74.7 ± 6 . 2 99.6 ± 8 . 3 50.0 5 1 . 0 ± 4 . 2 102.0 ± 8 . 4 25.0 22.9 ± 2 . 1 91.6 ± 8 . 4 10.0 8.6 ± 0 . 9 86.0 ± 9 . 0 * Average of three determinations, each with eight replicate wells. Table 4.11 Recovery of D H A from tap water spiked with D H A using antibodies from D H A M -S U C - K L H in a direct E L I S A D H A spiked in D H A recovered Recovery sample (ug/L) (Ug/L)* (%) 150.0 146.2 + 15.2 97.5 ± 10.1 75.0 81.1 ± 7 . 4 108.1 ± 9 . 9 50.0 48.1 ± 5 . 8 96.2 ± 11.6 25.0 23.9 ± 3 . 3 95.6 ± 13.2 10.0 8.7 ± 1.6 87.0 ± 16.0 * Average of three determinations, each with eight replicate wells. 123 Table 4.12 Recovery of D H A from biotreated B K M E spiked with D H A by direct E L I S A D H A spiked in sample (pg/L) Recovered D H A (Ug/L)* Recovery (%) Adjusted recovery (%) 160.0 170.6 ± 19.2 106.6 105.3 80.0 85.4 ± 11.4 106.7 104.0 40.0 41.9 + 6.9 104.7 99.3 20.0 22.1+3.3 110.3 99.5 10.0 10.3 ± 1.7 103.0 81.0 5.0 6.1+0.9 122.0 78.0 0 2.2 + 0.4 * Average of three determinations, each with eight replicate wells. 4.3.4.3. F a l s e pos i t i ve a n d fa lse negat ive The percentage of false positives and false negatives is an indicator of the reliability of a particular assay method. The definition of a false negative is a negative response for a sample that contains up to two times the stated detection level of the target analytes while the false positive is a positive response for a sample that contains analytes at one-half of the detection level (Lesnik, 1994). During the screening for the presence of toxic chemicals in a large number of environmental samples, a false negative may lead to dangerous exposure of toxic material to the environment. On the other hand, a false positive may lead to an unnecessary investment in time and money for unnecessary actions. In order to determine the percentage of false positive and false negative responses, 32 parallel samples were prepared using assay buffer and spiked with 10 pg/L 124 of D H A (two times the detection limit) and another 32 samples with 2.5 p,g/L of D H A (1/2 the detection limit). Direct E L I S A indicated that among the 32 samples (triplicate for each sample) spiked with 2.5 ug/L of D H A , 2 samples showed D H A concentration above 5.0 ug/L, a 6.3% false positive. N o false negatives were found for the 32 samples spiked with 10.0 ug/L of D H A (determined value less than 5.0 ug/L) (Table 4.13). Table 4.13 The percentage of false positive and false negative response determined by the direct E L I S A Total samples spiked with 2.5 ug/L of D H A 32 N o . of samples with determined concentration above 5.0 p,g/L 2 False positive 6.3% Total samples spiked with 10.0 ug/L of D H A 32 N o . of samples with determined concentration below 5.0 u.g/L 0 False negative 0% When the bio-treated B K M E was spiked with similar concentrations of D H A and assayed by the direct E L I S A , it was found that, 18 of the 32 samples spiked with 2.5 ug/L of D H A showed D H A concentration higher than 5.0 ug/L, causing 56.3% false positives. However, i f a nominal 2.2 ug/L of background reading was deducted, only 3 samples showed false positive values (9.4%). If the standard curve was generated using the same bio-treated B K M E , false positives were decreased to 6.3%. N o false negatives were determined. 4.3.4.4. Quantification of D H A in the presence of A B A or PIM A s described previously, the polyclonal antibodies against D H A cross-reacted with other resin acids, particularly the abietic type resin acids. Therefore, the ability of the assay to quantify 125 D H A in the presence of other resin acids was evaluated. D H A and P L M were selected as representatives of the abietic and pimaric type resin acids. They were mixed in proportions ranging from 3:1 to 1:3, with a total concentration from 7.8 to 125.0 pg/L. These samples were quantified using the direct E L I S A and the antibodies from the rabbit immunized with D H A M -S U C - K L H . The D H A equivalents determined for each sample are listed in Table 4.14. It was found that P L M with concentrations of three times more than D H A did not affect the quantification of D H A significantly as recoveries ranged from 71-121%, with an average recovery of 96 ± 18%. Therefore, it did not appear that the presence of P L M in the sample affected the quantification of D H A significantly. D H A and A B A are structurally similar abietic type resin acids. When D H A and A B A were mixed and assayed by the direct E L I S A , it was found that all recoveries for samples with a D H A : A B A ratio of 3:1 were above 90% (at an average of 93%) and the average recovery for samples with a D H A : A B A ratio of 1:1 was 86%. However, the recoveries for samples with a D H A : A B A ratio of 1:3 were lower, ranging from 58 to 76% (at an average of 68%) (Table 4.15). It should be realized that the standard curve was created using D H A , not A B A and this may result in the low recovery of A B A presented in the mixture. A s indicated in the literature, D H A is usually the predominant resin acid in most pulp m i l l waste water streams and the ratio of abietic acid to D H A is unlikely to be greater than 3:1. Therefore, it is possible that the responses from D H A and A B A are additive and the E L I S A can be used to quantify D H A and A B A together. A s the A B A and P L M are representatives of the abietic and pimaric type resin acids, it is probable that the contributions of the pimaric type resin acids are negligible and that the D H A equivalents determined by E L I S A using polyclonal antibodies from D H A M - S U C - K L H provide an approximation of the total amount of abietic type resin acids. 126 Table 4.14 Recovery of D H A equivalents in a D H A - P L M mixture as determined by the direct E L I S A using antibodies from D H A M - S U C - K L H D H A : P L M ratio Total amount (pg/L) 125.0 62.5 31.2 15.6 7.8 75:25 D H A concentration (ug/L) 93.8 46.9 23.4 11.7 5.9 D H A eq determined (ug/L) 108.0 59.4 24.1 12.7 4.2 Recovery (%) 115 127 103 109 71 50:50 D H A concentration (pg/L) 62.5 31.2 15.6 7.8 3.9 D H A eq determined (pg/L) 58.1 31.5 16.2 5.9 2.8 Recovery (%) 93 101 104 76 72 25:75 D H A concentration (pg/L) 31.2 15.6 7.8 3.9 2.0 D H A eq determined (pg/L) 35.2 16.9 6.2 2.8 1.8 Recovery (%) 113 108 79 72 90 Table 4.15 Recovery of D H A equivalents in a D H A - A B A mixture as determined by the direct E L I S A using antibodies from D H A M - S U C - K L H D H A A B A ratio Total amount (pg/L) 125.0 62.5 31.2 15.6 7.8 75:25 D H A eq determined (pg/L) 121.9 59.1 28.3 14.2 7.0 Recovery (%) 98 95 91 91 90 50:50 D H A eq determined (pg/L) 118.9 55.4 26.1 13.8 5.9 Recovery (%) 95 89 84 88 76 25:75 D H A eq determined (pg/L) 95.2 46.9 21.3 10.0 4.5 Recovery (%) 76 75 68 64 58 127 4.4. Conclusions Polyclonal antibodies have been raised against D H A , one of the major resin acids found in pulp m i l l effluents: Various E L I S A formats were compared. Immunizing antigens, with or without a spacer arm, did not show any significant difference in the elicitation of antibody production, while the polyclonal antibodies obtained reacted with D H A and other resin acids in a a similar way, in terms of I50 and cross-reactivity. However, different assay sensitivity could be achieved by using different assay formats. The indirect E L I S A with the biotin-streptavidin system was the most sensitive format producing an I50 of 12.3 pg/L for antibodies from D H A M - S U C - K L H . It was considerably better than the traditional indirect E L I S A which provided an I50 of 113.2 ug/L. The direct E L I S A had an I50 of 49.7 ug/L when the same antibodies were used. Although less sensitive than the indirect format with the biotin-streptavidin system, the direct E L I S A required less assay steps and was easy to perform. Although the magnetic bead E L I S A required the fewest assay steps, it was less sensitive (I50 of 139.1 pg/L). A m o n g the several parameters evaluated, the p H , ionic strength of the samples, and the presence of metallic ions had a relatively small influence on the assay. The polyclonal antibodies from D H A M - S U C - K L H and D H A - K L H could tolerate a 10% and 2% concentration of methanol, respectively, as determined by the direct E L I S A . When various matrices, such as assay buffer, tap water and biotreated B K M E , were spiked with D H A and assayed using the direct E L I S A , satisfactory recoveries were obtained. Within-assay and between-assay comparisons revealed that between assay variation was usually larger than within-assay variation. The cross-reactivity of the polyclonal antibodies from D H A M - S U C - K L H to abietic type resin acids may provide a possibility to quantify this group of resin acids. 128 CHAPTER 5 QUANTIFICATION OF RESIN ACIDS IN PULP MILL EFFLUENTS USING ELISA 5.1 Introduction Resin acids are considered to be the primary source of fish toxicity for effluents from pulp and paper mills using softwoods as the furnish (Taylor et al, 1988). This continues to be an issue of concern as resin acids are wood derived and w i l l continue to be released into pulp m i l l effluents, no matter which pulping process is used (Mcleay & Assoc., 1987). Among the major pulping processes currently used in Canada, mechanical pulping tends to produce effluents with a very high resin acid content, primarily because little wash water is used. Resin acid concentrations as high as 1000 mg/L have been detected in chemi-thermomechanical pulp ( C T M P ) m i l l effluents (Cornacchio and H a l l , 1988; L i u et al, 1991; Bicho et al, 1995a). Although the levels of resin acids present in bleached kraft m i l l effluents ( B K M E ) can vary significantly, they are generally lower than the levels found in mechanical pulp m i l l effluents since large amounts of water are normally used for pulp bleaching and washing (Taylor et al, 1988). Resin acids commonly found in Canadian pulp m i l l effluents are abietic, dehydroabietic, pimaric, isopimaric, levopimaric, neoabietic, palustric and sandaracopimaric acids (Taylor et al, 1988). Quantification of resin acids in pulp m i l l effluents is usually achieved by conventional analytical methods such as H P L C , G C or G C - M S (Foster and Zinkel , 1982; N C A S I , 1975, 1986; Voss and Rapsomatiotis 1985; Lee et al, 1990; Richardson et al, 1992, Shepard et al, 1996). However, due to the lengthy procedures involving extraction, derivatization and chromatographic separation and quantification, these methods are generally not suitable for routine analysis of large numbers of samples in a short period of time. In addition, the variations in the effluent composition make the multi-step analytical protocols difficult to perform. A 129 round-robin interlaboratory comparison involving eight laboratories from Canada, the United States and N e w Zealand indicated that interlaboratory variations were significant. Variations as high as 55% were found for the resin acid standards while the variations in the samples derived from the real samples were considerably higher (Bicho et al, 1995b). A s described in the previous part of this thesis, we have investigated several options such as p H and solvents used for sample extraction and the solid phase extraction used for sample clean-up ( L i et al, 1996b). The method developed was used as a standard method to compare the E L I S A with. Immunoassays are widely used in the detection and quantification of toxic chemicals, such as pesticide residues and industrial pollutants (Vanderlaan et al, 1990; Schneider and Hammock, 1992; Van-Emon and Lopez-Avi la , 1992; Hock, 1993; V a n Emon and Gerlach, 1995a; Kaufman and Clower, 1995; Linde and Goh, 1995). Since immunoassays are based on the interaction between an antibody and antigen (Voller et al, 1979), they are target specific and the antibodies are capable of recognizing analytes from a complex sample matrix, thus minimizing the work needed for sample pretreatment. A n enzyme-linked immunosorbent assay ( E L I S A ) based on polyclonal antibodies has been developed for dehydroabietic acid ( D H A ) , one of the major resin acids found in pulp m i l l effluents ( L i et al, 1994). This method has been validated by analysis of various sample matrices spiked with D H A and good linearity was obtained for different systems, including assay buffer, tap water and bio-treated B K M E ( L i et al, 1994). However, the assay has not yet been applied to representative pulp m i l l effluent samples. Previous results indicated that the polyclonal antibodies that we raised against D H A also cross-reacted with abietic type resin acids ( L i et al, 1994). Although this suggested that these polyclonal antibodies could therefore be used to quantify the total abietic type resin acids, it was possible that the presence of both degradation products derived from resin acids and various other inorganic and organic compounds in the untreated effluent matrix could interfere with the E L I S A quantification. It is wel l known that pulp m i l l effluents are complex in their chemical composition. More than 250 compounds have been identified and, together with other materials, these components 130 are collectively characterized as biochemical oxygen demand (BOD), chemical oxygen demand (COD) , and total suspended solids (TSS) (Suntio et al, 1988). The final objective of this project was to develop an immunochemical method which could be used to quantify resin acids in pulp m i l l effluents. In this part of the thesis, we describe an E L I S A protocol employing polyclonal antibodies against D H A as a group specific probe. The direct E L I S A was applied to various B K M E and C T M P effluents to determine i f the effluents could be assayed directly or i f some form of extraction or removal of interference was required before good linearity could be achieved. 5.2 Material and Methods A l l resin acid standards were obtained from Hel ix Biotech Corporation (Richmond, B C , Canada). Carbonate-bicarbonate buffer capsules, phosphate-citrate buffer with sodium perborate capsules, o-phenylenediamine (OPD) methyl heneicosanoate (MHS) , O-methylpodocarpic acid ( O M P A ) were purchased from Sigma (St. Louis, M O ) . N-methyl-N-nitroso-p-toluene sulfonamide (Diazald ), tricosanoic acid ( T C A ) , humic acid (sodium salt) and alkali l ignin (#37 095-9) were purchased from the Aldr ich Chemical Company, Inc. (Milwaukee, Wis) . Ethyl acetate (EtOAc) was purchased from Fisher ( H P L C grade). B A C T O dehydrated skim milk was purchased from Difco Laboratories (Detroit, MI) . The E L I S A plate was IMMULON® 4 (flat bottom, Catalog N o . 01-010-3855) from Dynatech Laboratories Inc. (Chantilly, V A ) and optical density was read on a T H E R M O m a x ^ M microplate reader (Molecular Devices Corp., Menlo Park, C A ) . Individual resin acids were identified and quantified on a Hewlett Packard H P 5890 series II gas chromatograph equipped with an H P 7673 auto injector, flame ionization detectors and a dual column system (DB-5 and DB-17 fused silica capillary column, 0.25-mm LD, 30-m long, 0.25 um thickness, from J & W Scientific). Both treated and untreated C T M P m i l l effluents and B K M E were obtained from three mills located in British Columbia, Canada, which used softwoods as their furnish (Table 2.1). They were collected over a period of more than six months and stored at -20°C before use. The polyclonal antibodies were raised from rabbits 131 immunized with D H A M - S U C - K L H , a conjugate between dehydroabietylamine and keyhole limpet hemocyanin ( K L H ) ( L i et al, 1994). The enzyme tracer was D H A M - S U C - H R P , which was prepared by conjugation of D H A M - S U C to horseradish peroxidase (HRP). 5.2.1. Competitive direct E L I S A The D H A equivalents in pulp m i l l effluents were determined using the following procedures. Brief ly, the microtitre plate was coated with polyclonal antibodies (1:8000, 100 p 1/well) i n carbonate-bicarbonate buffer (pH 9.6) and dried at 37°C overnight. The plate was washed four times using P B S buffer to remove the unbound material and the remaining binding sites on the plate were blocked with 2% milk in P B S (200 pl/well) at 37°C for 1 hr. Once the blocking agent was removed and the plate washed, the sample solution (or standard solution) in P B S buffer and the enzyme tracer (1:8000 in P B S with 0.1% milk) were mixed in equal volume and added to the microtitre plate (100 pl/well). The plate was incubated at 37°C for 1.0 hrs. After washing, the enzyme substrate O P D (1.0 mg/ml in citrate-phosphate buffer, 100 pl/well) was added and the reaction was stopped after 15 min using sulfuric acid (2.5 M , 50 pl/well). The colour intensity was measured at 490 nm using a plate reader. 5.2.2. G C analysis Individual resin acids were quantified using the following procedures. Briefly, untreated C T M P effluent (20.0 ml) or untreated B K M E effluent (100.0 ml) was spiked with surrogate (O-M P C A , 50.0 p i , 1.0 mg/ml in methanol), adjusted to p H 3 with 1 M HC1, and extracted twice with an equal volume of ethyl acetate. The organic phases from each extraction were combined and dried over anhydrous M g S 0 4 . The solvent was then removed under reduced pressure. The extract was dissolved in 1.0 m l of diethyl ether and spiked with an internal standard (50 p L of M H S in methanol, 1.0 mg/ml) and a methylation standard (50 p L of T C A in methanol, 1.0 mg/ml). The mixture was then derivatized with diazomethane, which was generated in situ by 132 the reaction of N-methyl-N-nitroso-p-toluene sulfonamide (Diazald) with alcoholic potassium hydroxide and delivered to the sample vial by a stream of nitrogen gas, until a persistent yellow colour was obtained. The ether was evaporated under nitrogen and the methyl esters of fatty and resin acids were redissolved in 1.0 m L of methanol. Individual resin acids were identified and quantified on a Hewlett Packard H P 5890 series II gas chromatograph (GC) equipped with an H P 7673 auto injector, flame ionization detectors and a dual column system (DB-5 and DB-17 fused silica capillary column, 0.25-mm LD, 30-m long, 0.25 um thickness, from J ' & W Scientific). Hel ium and nitrogen were used as the carrier and make-up gases, respectively. Injector and detector temperatures were 260°C and 290°C, respectively. The oven temperature was programmed at 60°C for 2 min, with an increase of temperature at 35°C/min to 170°C, then 0.6°C/min to 200°C and finally 35°C/min to 280°C when it was kept at this temperature for 10 minutes. 5.3 Results and Discussion Previously we had used an indirect E L I S A format with biotin-streptavidin system ( L i et al, 1994), in the work reported in this part of the thesis, a direct format was adopted due to its simplicity and short assay time. A typical standard curve with D H A concentrations ranging from 2.0 to 250.0 ug/L gave a regression equation of % Inhibition = 1.07 + 28.85 L o g [ D H A ] , with a coefficient of correlation of R=0.998. From this regression equation, the concentrations required for a 50% inhibition (I50) and detection limit, defined as the amount of D H A which yielded a 20% inhibition (Midgley et al, 1969), were calculated as 49.7 ug/L and 4.5 ug/L, respectively. A s there is day to day variation due to the small differences in timing, temperature or reagent age, a standard curve was included with each assay to ensure an accurate result. Prior to applying the direct E L I S A to effluent samples, negative sample matrices, such as assay buffer, tap water and biotreated B K M E , were spiked with D H A with concentrations 133 ranging from 5.0 to 320.0 pg/L and assayed by the direct E L I S A . Similar to the previous results obtained from the indirect E L I S A ( L i et al, 1994), recoveries were usually in the range of 85 to 110% (data not shown). 5.3.1. E L I S A for effluent samples from a C T M P mill Immunoassays developed for toxic chemicals in aqueous samples, such as drinking water and ground water, do not typically encounter interference by contaminating materials or matrix effects (Schneider and Hammock, 1992, Hottenstein et al, 1995). A s C T M P effluents contain high levels of C O D and B O D , and therefore, probably contain high concentrations of many chemical compounds, resin acid concentrations are also usually at elevated levels (Cornacchio and H a l l , 1988; L i u et al, 1991). A s a result, effluent samples must first be diluted in order to bring the resin acid concentrations into the quantification window. Due to the high dilution required, none of the other material in the effluent exerted any adverse effect on the quantification of resin acids and good linearity was maintained over a wide range of dilutions (Table 5.1). Table 5.1 The D H A equivalents determined by direct E L I S A for a C T M P effluent sample diluted in assay buffer Dilution 1:40 1:80 1:160 1:320 1:640 D H A determined (pg/L)* 133.80 ± 12.24 59.73 ± 6 . 0 8 33.12 ± 3.15 16.74 ± 1.49 8.14 ± 0 . 7 8 D H A in original (mg/L) 5.35 4.78 5.30 5.36 5.21 A v e = 5.20 mg/L s = 0.24 C V % = 4.6 * Mean value was calculated on eight replicate wells. 134 To test the precision of the assay, the same sample was assayed on four consecutive days using a single dilution (1:80). Between-assay and within-assay variation was determined by analysis of variance ( A N O V A ) (Bookbinder and Panosian,. 1986). It was found that between-assay variation was bigger than within-assay variation and the total C V was 11.78% (Table 5.2). Table 5.2 The D H A concentration of a C T M P effluent assayed over four consecutive days as determined by direct E L I S A Day 1 Day 2 Day 3 Day 4 Replica wel l 48 48 48 48 D H A determined (mg/L) 5.05 ±0.46 4.77 ±0.31 5.23 ±0.20 6.04 ±0.17 Variations Grand mean = 5.27 mg/L CV%(within) = 5.81 C V % (between) = 10.25 CV%(total) = 11.78 The precision obtained with this initial work encouraged us to further assess the accuracy of the assay. W e next assayed the resin acid content of three different wastewaters from various stages of a typical C T M P process by both the direct E L I S A and G C (Table 5.3). These samples represented waste streams containing low, medium and high concentrations of resin acids, from the aerobic stabilization basin, preacidification tank and preheater site, respectively. Although the E L I S A could not provide direct information about the concentration of individual resin acids, the D H A equivalents, as determined by E L I S A , compared favourably with the total concentration of abietic type resin acids determined by G C . These results indicated that the polyclonal antibodies could be effectively used as group specific antibodies to quantify abietic type resin acids. H i g h precison was obtained as the G C data were the average of triplicates ( C V from 3-14%o) and the E L I S A data were the average from three determinations (three dilutions for each samples and four replicate wells for each dilution, C V from 8-19%). 135 Table 5.3 The determination of the individual resin acids (by G C ) and the D H A equivalents (by direct E L I S A ) present in three different sites of the wastewater streams, from a C T M P m i l l Aerobic stabilization basin Preacidification tank Preheater site P L M 900 1400 24800 S A N 100 400 5300 ISO 1800 2800 34800 P A L / L E V 400 3800 33200 D H A 3600 10800 104700 A B A 2700. 5200 93100 N E O 300 1400 10100 7 - O X O - D H A 1100 300 9500 Total abietic type resin acids by G C (pg/L)* 8,100 21,500 250,600 Total resin acids by G C (pg/L)* 10,800 26,200 315,400 D H A eq. by E L I S A ( p g / L ) b 7,049 21,465 251,231 G C % ( D H A eq/Total Abietic) 87.0% 99.8% 100.2% a. average from three determinations ( C V from 3-14%) b. average from three determinations (each determination was composed of three dilutions for each samples and four replicate wells for each dilution, C V from 8-19%) Since it is generally the total concentration of resin acids that w i l l determine effluent toxicity, it would be highly desirable to have an indication of the total resin acid content of the effluent rather than just the abietic type resin acid content. In order to obtain this information, D H A equivalents as determined by E L I S A were correlated with the total resin acid content as determined by G C . When all the available data from E L I S A and G C determinations were 136 plotted, good agreement was obtained, with a coefficient of correlation of 0.993, which was significant at P < 0.001 (Figure 5.1). D H A equivalents determined by E L I S A (ppm) 350 J E L I S A -0.74 + 0.75 * G C (R = 0.993) 300 4 250 4 200 4 1504 1004 50 4 0 0 50 100 150 200 250 300 350 Total resin acid concentration determined by G C (ppm) Figure 5.1 The correlation of D H A equivalents determined by direct E L I S A and the total resin acid content determined by G C for various effluents collected at different sites in a C T M P m i l l The information obtained from this correlation allowed us to estimate the total resin acid content based on the E L I S A response. Using this regression equation and the assumption that the E L I S A quantified just the abietic type resin acids, the abietic type resin acids w i l l account for 66.6 to 74.8% of the total resin acids over a concentration range of 5 to 350 mg/L. This result is similar to both data in the literature ( M I S A , 1993) and the values and ratios that we have observed in our own work. For example, analysis of the resin acid concentrations from a bleached C T M P m i l l wastewater indicated that abietic type resin acids counted for 75.3% of the total resin acids (McCarthy et al, 1990). Usually, abietic type resin acids predominate in most pulp m i l l effluents. A report from the Ontario Municipal/Industrial Strategy for Abatement 137 ( M I S A ) programme involving 29 pulp mills indicated that abietic type resin acids account for 69.5% of the total resin acids ( M I S A , 1993). Although the E L I S A values were lower than the total resin acid values obtained by the G C , this was not surprising as the polyclonal antibodies primarily detected abietic type resin acids, which have a similar carbon skeleton to D H A . However, this good correlation still allowed us to estimate the total resin acid content based on the E L I S A response. 5.3.2. E L I S A for B K M E samples Unlike the effluents from a C T M P m i l l , B K M E usually contains lower levels of C O D and B O D due to the large volumes of water used in both the bleaching and pulp washing stages. The concentration of resin acids in the final effluents are also generally lower than those from C T M P mills. However, B K M E s are complex in their chemical composition and contain many chemical compounds derived from both wood furnish and introduced during pulping and bleaching (Suntio et al, 1988). When a B K M E sample was diluted using the assay buffer and analyzed by the direct E L I S A , a non-linear decrease of D H A equivalents was observed. It was apparent that the sample dilution curve was not parallel to the standard curve (Figure 5.2). It was also found that the patterns of the dilution curves for different B K M E samples were not the same, indicating that B K M E components varied between samples. For an immunoassay to be effective, the sample dilution curve should be parallel to the standard curve (Jung et al, 1989). Obviously, when this non-parallelism occurs, the reference standard curve cannot be used to predict the concentration of D H A equivalence in the original B K M E (Peterman, 1991). For example, the predicted resin acid concentrations ( D H A equivalents) for the particular B K M E sample used in Figure 5.2 were 261.8, 358.8, 276.3, 158.0 and 172.2 pg/L, respectively, for samples diluted from 1 to 16 times. However, as a result of the large variations in the data, the resin acid content in the sample could not be determined. This non-parallel phenomenon indicated that some components in the B K M E sample were affecting 138 Inhibition% Inhibition% Figure 5.2 A typical D H A standard curve and the dilution curve for a B K M E sample the assay. Although many factors may contribute to this phenomenon, it was most likely that this was due to background interference or a matrix effect, as the background was not maintained at the same level when the sample was diluted. Due to the complex and unknown nature of the effluent components, it was difficult to identify what kind of compounds were causing the problem. Different strategies were attempted to try resolve this problem. First we tried Mil l ipore membrane filtration (0.45 pm pore size) as this step was generally used for clean up of sample prior to H P L C analysis. Although this improved the assay by reducing the variations of predicted values for sample diluted, G C analysis indicated that there was a decrease in resin acid concentration before and after filtration. It was likely that a high portion of resin acids were retained on the membrane due to the hydrophobic nature of these compounds. W e next evaluated the possible effects of metallic ions on the E L I S A by testing the various metallic ions that have been found in effluents. Neither N a + (up to 200 m M ) , i 139 K + , C a 2 + , Mg2+, Zn^+or M n 2 + (all up to 50 m M ) had any adverse effect when added to the assay buffer, indicating that these ions were not the source of the interference. Extraction of analytes from aqueous solutions is a.common method that is routinely used in conventional instrumental analyses. This method has also been used as a simple sample pre-treatment, either for concentration of analytes or removal of interference prior to E L I S A (Fleeker, 1987; A g a et ai, 1994). When we tried to extract resin acids from effluent samples using ethyl acetate, although the performance of E L I S A was improved, the non-linearity was not completely resolved. Presumably, interference due to inorganic components was removed, implying that the organic components that were co-extracted with resin acids were the main cause of the observed non-linearity. Fractionation of the organic extracts using a NH/? tyPe s ° l id phase extraction cartridge (Chen et al, 1994) followed by E L I S A determination of these fractions indicated that the resin/fatty acids fraction accounted for 87% of the total D H A equivalents that were detected. A 9% and 4% contribution could be attributed to the two less polar fractions. However, the exact chemical nature of these fractions was not determined. These fractions may contain phytosterols and some other fatty alcohols, which showed negligible cross-reactivity to the antibodies used in the assay. However, abietyl alcohol, aldehyde and possible resin acid degradation products may also be in this fraction. Current work is trying to confirm the presence of these compounds in these fractions. It is desirable to avoid the extra steps of sample pre-treatment such as extraction by either liquid-liquid or solid phase extraction, as this may lead to incomplete recovery of the analyte of interest. Although dilution is the method of choice for quantifying liquid samples, sample dilution and the other methods discussed previously failed to eliminate or reduce the background interference encountered with B K M E samples. Since the background levels varied for the different diluted samples, we need to find a way of maintaining the same background levels of interfering compounds for each sample dilution. This approach has previously been used in some cases to reduce the background effect in an E L I S A . For example, pooled human control urine was used as the diluent to generate the standard curve for the E L I S A of alachlor residue in 140 monkey urine (Feng et al, 1994). It has also been reported that, when E L I S A was used to quantify carbaryl, an insecticide, in water, soil extract, urine and honey, standard curves generated in the presence of matrices were parallel to that in buffer, although different sensitivities were obtained (Marco et al, 1993a). Therefore, it seems that a general approach to solve matrix problems would be to run the standard curve in the presence of other factors that mimic the behavior of the matrix (Marco et al, 1993b). W e tried this approach by using a biotreated B K M E effluent, with undetectable amount of resin acids as determined by G C , to dilute each of the samples. When this procedure was used to generate the standard curve and to dilute the real effluents, good parallelism between the standard curve and the sample dilution curve was maintained (Table 5.4). However, the assay sensitivity slightly decreased (I50 increased to 77.9 ug/L). A n untreated B K M E sample with high resin acid contents could also be analyzed this way (Table 5.5). It should be noted that the matrix interferences are analyte dependent and techniques developed for eliminating the background effect for one analyte may not be applicable to another analyte in the same sample. For instance, matrix interferences are still a significant problem in the immunoassay of urinary Cortisol since many steroidal as well as non-steroidal substances are present in urine. Neither solvent extraction or an acid or alkaline wash of the urine extract appeared sufficient to remove the interference completely (Nahoul et al, 1992). To test the validity of using the biotreated effluent to dilute the sample for the E L I S A , the percentage of false positives and false negatives (Lesnik, 1994) was determined using a negative B K M E spiked with D H A . In the screening of toxic chemicals in a large number of environmental samples, false negative data erroneously indicate a clean sample or site and lead to dangerous exposure of toxic material to the environment. Alternatively, a false positive may lead to a considerable investment in time and money to take an unnecessary action. When 32 parallel samples were prepared by spiking a negative B K M E sample with 10.0 ug/L of D H A (two times the detection limit), and then assayed by the direct E L I S A , it was found that no sample showed D H A concentration below 5.0 ug/L (average of triplicate data for each sample). 141 This indicated that there were no false-negatives. However, when we assayed another 32 parallel samples prepared by spiking 2.5 pg/L of D H A (1/2 of the detection limit) in negative B K M E samples by the direct E L I S A , 2 samples showed D H A concentration above 5.0 pg/L, indicating a 6.3% false positive. However, i f the standard curve was generated using assay buffer, not bio-treated B K M E , 18 of the 32 parallel samples spiked with 2.5 pg/L of D H A showed D H A concentration higher than 5.0 pg/L, causing 56.3% false positives. This seems reasonable since a nominal 2.2 pg/L background reading was found for bio-treated B K M E under such a system. Table 5.4 The direct E L I S A results for a B K M E sample diluted in a biotreated B K M E with undetectable amount of resin acids Dilution factor 1 2 4 8 16 D H A eq. determined (pg/L)* 150.4 ± 15.9 83.5 + 6.2 39.2 + 3.5 22.3+2.2 10.5 ± 1.6 D H A eq. corrected for dilution (pg/L) 150.4 ± 15.9 167.0+12.4 156.8+ 14.0 178.4 + 17.6 168.0 + 19.6 A v e = 164.1 pg/L s = 10.8 pg/L C V = 6.6% * Mean value was calculated on eight replicate wells. Table 5.5 The direct E L I S A results for a high resin acid containing B K M E sample diluted in a biotreated B K M E with undetectable amount of resin acids Dilution factor 10 20 40 80 D H A eq. determined (l-ig/L)* 122.8 + 13.7 53.3+6.2 25.2 + 2.8 13.2 + 1.5 D H A eq. corrected for dilution (pg/L) 1228.0 + 137.0 1066.0 + 124.0 1008.0+ 112.0 1056.0 + 140.0 A v e = 1089.5 pg/L s = 95.7 pg/L C V = 8.8% Mean value was calculated on eight replicate wells. 142 A limited number of B K M E samples with low, medium and high resin acid content were analyzed by both the direct E L I S A and the G C methods. Correlation between D H A equivalents as determined by E L I S A and the total resin acids as determined by G C resulted in a linear regression curve with a coefficient of correlation of 0.994, significant at P < 0.001 (Figure 5.3). D H A equivalents determined by ELISA (ppb) 1400 1200 1000 800 600 400 -200 -0 0 200 400 600 800 1000 1200 1400 Total resin acid concentration determined by G C (ppb) Figure 5.3 Correlation between D H A equivalents determined by direct E L I S A and total resin acid content determined by G C for a range of B K M E samples collected from a m i l l Based on this regression equation, the abietic type resin acids in the B K M E (determined as D H A equivalents by E L I S A ) should represent 64.2 to 79.8% of the total resin acids from 100 to 1400 ug/L. This is a similar result to that found with a New Zealand bleached kraft pulp and paper m i l l wastewater which contained an average of 65.6 to 82.3% of abietic type resin acids as determined from 7 sites in the m i l l (Zender et al, 1994). 143 While most immunoassays are designed for a single analyte, it has been reported that, with careful hapten design, antibodies can be developed which can detect a particular class of compounds (Hock, 1993). Our results support this statement as it was apparent that the polyclonal antibodies could be used to quantify abietic type resin acids in both B K M E and C T M P m i l l effluents. Although identification of individual compounds must depend on other methods, such as G C and G C - M S , quantification and screening of large number of samples could be efficiently carried out by an E L I S A . 5.2.3. Effect of possible interfering material on the E L I S A of B K M E Although the background effect could be eliminated by the use of a biotreated B K M E as the diluent for sample preparation and standard curve generation, we continue to try to identify the materials responsible for the interfering background effect. It is probable that the presence of resin acid degradation products or some other structurally similar compounds could be the source of much of the interference. However, retene (a biotransformation product) and several major phytosterols such as B-sitosterol, stigmastanol and campesterol did not show immunoreactivity to the polyclonal antibodies. Although interference did not appear to be derived from these materials, it is possible that dissolved organic matter, such as humic material, could bind lipophilic organic compounds and thus affect their bioavailability, biodegradability and toxicity. This phenomenon has also been found in the bleached kraft m i l l effluents containing degraded lignin and chlorolignin (Kukkonen, 1992). Due to the similar chemical characteristics of l ignin and humic material (both are high molecular weight polymeric phenolics), we decided to evaluate the interference of humic material on the E L I S A to see i f it could influence the interaction of the resin acids with the antibodies. When the assay buffer was spiked with different levels of humic acid, no significant changes were observed in the direct E L I S A containing humic acid concentrations of up to 125.0 mg/L. This seemed to indicate that humic acid did not affect the binding of hapten-enzyme conjugate to the polyclonal antibodies. 144 However, when D H A was added at 104.0 ug/L, different percentages of inhibition were observed at different humic acid concentrations. The humic acid appeared to interact with D H A by limiting the amount of D H A that was available for competition (Table 5.6). It seems that the interference was dependent upon the concentration of humic acid as interference increased as the concentrations of humic material increased. The alkali l ignin was also evaluated for their effects on the E L I S A . Since the alkali l ignin was dissolved in D M S O as a highly concentrated stock solution (about 85 mg/ml), the amount of D M S O in the assay system was wel l below 0.1% and had no effect on the antibodies. It was apparent that the E L I S A responses were reduced significantly as the lignin concentration increased. This indicated that there were some kinds of interactions between lignin and the antibodies or enzyme tracer. When D H A was used at a similar level (104.0 ug/L), the percentage of inhibition also decreased as the lignin concentration increased. This indicated that the lignin further interacted with the D H A and inhibited the competition between the D H A and the antibodies (Table 5.7). Since the alkali lignin used was a polymeric kraft l ignin isolated from a softwood pulp m i l l , it should be quite similar to the lignin material present in the B K M E used in this study. Compared to the humic material, the phenolic type lignin material is more likely to interfere with the E L I S A since a much lower concentration was required to obtain a similar inhibitory effect. Although the interfering compounds were not identified, the presence of l ignin-related polymeric phenolic compounds was probably one of the main sources of interference. Future work should be carried out to identify the interfering components present in pulp m i l l effluents. Recently, an H P L C based method of quantifying D H A has shown an excellent relationships between total resin acid content as determined by the G C method and D H A concentration as determined by the H P L C method (r = 0.97) (Shepard et al, 1996). Although this method requires no extraction and provides values in a matter of hours, it still requires the use of expensive equipment and a dedicated technician. Wi th further work, a simple assay kit should allow the quicker, routine analysis of many m i l l streams using an E L I S A based method. 145 Table 5.6 The influence of humic acid on the direct E L I S A determination of D H A present in assay buffer Humic acid (mg/L) Without D H A (OD @ 490 nm) With D H A (104.0 pg/L) (OD @ 490 nm) Inhibition % 250.0 0.806 ± 0 . 0 2 9 0.485 ± 0 . 0 6 9 39.8 125.0 1.101 ± 0 . 0 8 3 0.644 ± 0 . 0 5 8 41.5 62.5 1.034 ± 0 . 0 7 5 0.572 ± 0 . 0 4 6 44.9 31.2 1.090 ± 0 . 0 7 7 0.436 ± 0 . 0 2 3 60.0 15.6 1.059 ± 0 . 0 8 4 0.410 ± 0 . 0 1 2 61.3 7.8 1.050 ± 0 . 0 7 2 0.398 ±0.020 62.1 0 1.050 ± 0 . 0 4 4 0.359 ± 0 . 0 0 8 65.8 Table 5.7 The influence of alkali lignin on the direct E L I S A determination of D H A present in assay buffer L i g n i n (mg/L) Without D H A (OD @ 490 nm) With D H A (104.0 pg/L) (OD @ 490 nm) Inhibition % 21.25 0.284 ± 0.024 0.229 ± 0.049 19.4 10.63 0.467 ± 0 . 0 1 1 0.276 ± 0 . 0 5 8 40.9 5.31 0.607 ± 0 . 0 6 0 0.307 ± 0.040 49.4 2.66 0.731 ± 0 . 0 8 1 0.337 ± 0 . 0 1 8 53.9 1.33 0.826 ± 0.043 0.349 ± 0 . 0 2 8 57.8 0.66 1.014 ± 0 . 0 8 7 0.358 ± 0 . 0 1 4 64.7 0 1.117 + 0.070 0.360 ± 0 . 0 3 0 67.8 146 5.4. Conclusion A direct E L I S A based on polyclonal antibodies has been successfully applied to the quantification of abietic type resin acids in B K M E and C T M P effluents. When the direct E L I S A was used to quantify resin acids present in effluents from C T M P containing softwoods as the fibre furnish we obtained a good correlation between E L I S A and G C . The D H A equivalents as determined by the E L I S A gave an indication of the total abietic type resin acids and the total resin acids. Analysis of effluent samples from different processing sites within the same m i l l such as the aerobic stabilization basin, the preacidification tank and the preheater site indicated that the E L I S A could be used to quantify resin acids present in various other waste streams in the m i l l . However, it is possible that the ratio of resin acids may differ when different wood furnishes or when different pulping processes are used. A review of the literature indicated that there were very few circumstances where the amount of total pimaric type resin acids was greater than total abietic type resin acids. This implied that an E L I S A based on D H A equivalents should be applicable to most m i l l wastewater streams. Although the proportion of resin acids and other organic/inorganic materials in the C T M P wastewater streams varied considerably, the E L I S A still showed a good correlation in all of these cases. However, we still do not know i f the E L I S A could be used effectively in circumstance such as when the hardwoods are used as the fibre furnishes or when variables such as different bleaching steps are incorporated into the process. Factors such as the presence of residual amounts of different bleaching agents such as sulfite and hydrogen peroxide and use of B C T M P effluents or C T M P effluents derived from hardwood furnish should be pursued in the near future. Although the background interference could be eliminated by simple dilution when C T M P effluents were used, significant background interference was encountered in the E L I S A of B K M E samples. There was a problem of non-linearity when the sample was diluted using assay buffer and assayed directly. Although we initially tried to remove this interference by various strategies such as filtration and extraction, this did not dramatically improve the assay. 147 Compounds with stnictures similar to resin acids such as phytosterols and retene did not show any significant cross-reactivities to the antibodies. The presence of lignin-related polymeric phenolic compounds was probably one of the main sources of interference. One successful strategy to reduce the background interference was to use a bio-treated B K M E as the diluent in order to maintain the background at a similar level. However, as the composition of bio-treated B K M E w i l l vary between mills, this might limit the use of such an approach on resolving this interference in the E L I S A . Further work is needed for identifying interfering materials present in the B K M E . This method can be used for routine analysis of resin acids in pulp m i l l effluents. L i k e the H P L C method (Shepard et al, 1996), E L I S A only provides information about the total abietic type resin acids. However, identification of individual resin acids must rely on G C or G C / M S . Since the D H A equivalents determined by E L I S A represent total abietic type resin acids, which predominate in resin acids found in pulp m i l l effluents, it could be useful in the prediction of total resin acid concentration in the effluents and even the effluent toxicity. The correlation curve between D H A equivalents determined by E L I S A and the total resin acids determined by G C could be different from m i l l to m i l l due to different wood furnish used, however, such a correlation could be useful, as for a particular m i l l , the furnish used should be stable at least in a certain peroid of time. 148 CHAPTER 6 ELISA OF RESIN ACIDS IN FISH BILE 6.1. Introduction Resin acids are considered the primary source of fish toxicity (Taylor et al, 1988) and although biological treatment results in the removal of the acute toxicity, resin acids and other persistent bioaccumulative constituents are sometimes present in the treated effluents and may cause a variety of subacute effects on the aquatic environment (McLeay, 1987). When fish are exposed to such contaminated waters, hydrophobic resin acids are likely to partition -from the water into the fatty tissues of the fish by uptake across the g i l l surfaces. After entering the vascular system, these lipid-soluble compounds bind nonspecifically to plasma lipoprotein and are then distributed throughout the body by the blood (Kruzynski, 1979). Resin acids in the bloodstream are trapped and enzymatically transformed in the liver. Within the liver, the acids are conjugated with glucuronic acid to form polar, water soluble complexes. These conjugates are then transferred to the bile and eliminated by excretion into the gut (Kruzynski, 1979). It has been found that resin acids can accumulate in fish bile. For example, a bioconcentration factor of 996 in bile was reported for fish exposed to D H A at 0.65 mg/L after 120 hours (Kruzynski, 1979). Therefore, the measurement of the resin acid content of fish bile could provide an alternative method for monitoring the resin acid load in the receiving waters where fish have been exposed (Oikari and Holmbom, 1981). Previously a strong correlation between bile concentration and water column concentration has been demonstrated for various xenobiotics detected in fish (Oikari, 1986). Therefore, the analysis of fish bile could provide a sensitive tool for the assessment of effluent composition and constituent distribution. Currently, the analysis of resin acids in fish bile is mainly based on chromatographic techniques such as G C and H P L C (Morales et al, 1992). These techniques have recently been combined with mass spectrometry such as G C - M S to confirm the presence of individual resin acids. Since more than 99% of the resin acids are in the conjugated forms with glucuronic acid 149 in fish bile (Oikari et al, 1984), the conventional instrumental analytical methods usually require a series of sample pre-treatments prior to G C , H P L C or G C - M S analysis (Morales et al., 1992, Oikari et al, 1984). These include hydrolysis of the bile sample to break down the conjugate (3-glucuronide ester) by either enzymatic (using 3-glucuronidase) or chemical methods (using acidic or alkaline conditions), extraction of free resin acids and derivatization. The whole procedure is time consuming and requires relatively large sample volumes (100- 500 pi). It should also be noted that the hydrolysis may not quantitatively release all of the bound constituents and may also destroy some compounds of importance. Therefore, an analytical procedure requiring small sample size and less pre-treatment would be highly desirable. Recently, immunoassays have been widely used in the detection and quantification of toxic chemicals, such as pesticide residues and industrial pollutants (Hock, 1993; Kaufman and Clower, 1995; Linde and Goh, 1995; Schneider and Hammock, 1992; Vanderlaan et al, 1990; Van-Emon and Lopez-Avi la , 1992; V a n Emon and Gerlach, 1995 a). Since immunoassays are based on the interaction between an antibody and antigen, they are target specific and the antibodies are capable of recognizing analytes from a complex sample matrix, thus minimizing the work required for sample pretreatment. In most cases, sample treatment could be by-passed or reduced to a minimum. Recently enzyme-linked immunosorbent assays ( E L I S A ) based on the polyclonal antibodies were developed for dehydroabietic acid ( D H A ) , one of the major resin acids found in pulp m i l l effluents ( L i et al, 1994; 1996a). In this part of the thesis, we describe the use of the direct E L I S A to quantify resin acids in file bile. The E L I S A results were compared with those obtained by conventional G C method. 6.2. Material and Methods A l l resin acid standards were obtained from Hel ix Biotech Corporation (Richmond, B C , Canada). Carbonate-bicarbonate buffer capsules, phosphate-citrate buffer with sodium perborate capsules and o-phenylenediamine (OPD), and methyl heneicosanoate ( M H S ) were purchased 150 from Sigma (St. Louis, M O ) . N-methyl-N-nitroso-p-toluene sulfonamide (Diazald) and tricosanoic acid ( T C A ) were purchased from Aldr ich Chemical Company, Inc. (Milwaukee, Wis) . The 96-well E L I S A plates were IMMULON® 4 (flat bottom, Catalog N o . 01-010-3855) from Dynatech Laboratories Inc. (Chantilly, V A ) and optical density was read on a T H E R M O m a x T M m i c r o t i t r e plate reader (Molecular Devices Corp., Menlo Park, C A ) . The polyclonal antibodies were raised from rabbits immunized with D H A M - S U C - K L H , a conjugate between dehydroabietylamine and keyhole limpet hemocyanin ( K L H ) ( L i et al, 1994) and the enzyme tracer was D H A M - S U C - H R P , prepared by conjugation of D H A M - S U C to horseradish peroxidase. The fish bile samples were supplied by West Vancouver Laboratory, Fisheries and Oceans, Canada. Individual resin acids were identified and quantified on a Hewlett Packard H P 5890 series II gas chromatograph (GC) equipped with an H P 7673 auto injector, flame ionization detectors and a dual column system (DB-5 and DB-17 fused silica capillary column, 0.25-mm ID, 30-m long, 0.25 pm thickness, from J & W Scientific). 6.2.1. Direct ELISA The D H A equivalents in fish bile samples were determined using the following procedures. Brief ly, the microtitre plate was coated with the polyclonal antibodies (1:8000, 100 ul/well) in carbonate-bicarbonate buffer (pH 9.6) and dried at 37°C overnight. The plate was washed four times using P B S to remove the unbound material and the remaining binding sites on the plate were blocked with 2% milk in P B S (200 ul/well) at 37°C for 1 hr. Once the blocking agent was removed and the plate washed, bile sample or standard were dissolved in P B S buffer, mixed with the enzyme tracer (1:8000 in P B S with 0.1% milk) in equal volume and added to the microtitre plate (100 ul/well). The plate was incubated at 37°C for 1.0 hr. After washing, the enzyme substrate O P D (1.0 mg/ml in citrate-phosphate buffer, 100 ul/well) was added and the reaction was stopped after 15 min using sulfuric acid (2.5 M , 50 ul/well). The colour intensity was measured at 490 nm using a plate reader. 151 6.2.2. Hydrolysis of fish bile and G C analysis Fish bile (200 ul) was added to 10 m l of 0.5 M K O H (90% in ethanol) and the mixture was refluxed at 70°C for 3 hrs. After hydrolysis, the reaction mixture was worked up by addition of water (10 ml), p H adjustment to 3 with diluted HC1 and extraction with ethyl acetate (20 m l x2). The organic layers were combined, dried over anhydrous M g S 0 4 and evaporated under pressure. The extract was dissolved in 1.0 m l of diethyl ether and spiked with an internal standard (50 u L of M H S in methanol, 1.0 mg/ml) and a methylation standard (50 p L of T C A in methanol, 1.0 mg/ml). The mixture was derivatized with diazomethane, which was generated in situ by reaction of N-methyl-N-nitroso-p-toluene sulfonamide (Diazald®) with alcoholic potassium hydroxide and delivered to the sample vial by a stream of nitrogen gas, until a persistent yellow colour was obtained. The ether was evaporated under nitrogen and the methyl esters of fatty and resin acids were redissolved in 1.0 m L of methanol. Individual resin acids were identified and quantified on a Hewlett Packard H P 5890 series II gas chromatograph (GC) equipped with an H P 7673 auto injector, flame ionization detectors and a dual column system (DB-5 and DB-17 fused silica capillary column, 0.25-mm ID, 30-m long, 0.25 pm thickness, from J & W Scientific). Hel ium and nitrogen were used as the carrier and make-up gases, respectively. Injector and detector temperatures were 260°C and 290°C, respectively. The oven temperature was programmed at 60°C for 2 min, with an increase of temperature at 35°C/min to 170°C, then 0.6°C/min to 200°C and finally 35°C/min to 280°C when it was kept at this temperature for 10 minutes. 6.3. Results and Discussion 6.3.1. Spiked bile To evaluate the matrix effect using E L I S A in the quantification of resin acids in fish bile, a control bile sample from fish which had not been exposed to pulp m i l l effluents was spiked 152 with D H A to a final concentration of 100 mg/L and assayed by the direct E L I S A . It was found that recoveries ranged from 96 to 123% at different dilution levels (Table 6.1). However, the E L I S A of the unspiked control bile under the same conditions indicated that the background interference was significant, i f the sample was not diluted to some degree. This was probably the main reason for the high recovery that was observed. It was found that a dilution of at least 1:500 was needed in order to bring the background to a negligible level. However, it should be noted that the dilution factor which can be used to eliminate background effect depends on the composition of the control bile. The adjusted recoveries, which were calculated by deducting the background reading, are also listed in Table 6.1. L o w recoveries were obtained at low spiking levels, possibly due to the detection limit of the assay. Table 6.1 Recovery of D H A equivalents by direct E L I S A for spiked bile samples D H A in spiked bile (pg/L) Sample dilution D H A recovered from spiked (Ug/L)* Background in unspiked (Ug/L) Recovery (%) Adjusted recovery (%) 400.0 1:250 435.5 ± 3 7 . 1 19.7 108.9 104.0 300.0 1:333 288.4 ± 3 1 . 0 8.3 96.1 93.4 200.0 1:500 234.3 ± 2 1 . 8 5.1 117.2 114.6 100.0 1:1000 113.6 ± 9.9 4.6 113.6 109.0 50.0 1:2000 60.2 ± 4 . 1 4.5 120.4 111.4 25.0 1:4000 30.7 ± 2 . 6 4.9 122.8 103.2 12.5 1:8000 13.1 ± 1.0 3.4 104.8 77.6 6.25 1:16000 6.9 ± 1.0 2.6 110.4 68.8 *Data are the average of eight wells. 153 6.3.2. B i l e f r o m fish exposed to e f f luent When the direct E L I S A was applied to a real bile sample, the dilution curve was parallel to the standard curve and good linearity was maintained over a wide range of dilutions. This seemed to indicate that little or no background interference was encountered. A n average concentration of 259.5 mg/L D H A equivalents was predicted (eight dilutions, 4 wells for each dilution) (Table 6.2). Table 6.2 D H A equivalents determined by direct E L I S A for a bile sample at different dilutions Dilut ion Mean (mg/L)* Std. Dev. (mg/L) C V % 1:100 278 9 3 1:200 287 43 15 1:400 320 19 6 1:800 270 23 8 1:1600 244 18 8 1:3200 242 70 28 1:6400 212 54 25 1:12800 222 42 19 grand mean = 259.5 ± 49.0 mg/L, C V = 18.9% * Data are the average of four wells. 6.3.3. C o m p a r i s o n between E L I S A a n d G C Since most of the resin acids in the fish bile were conjugated to glucuronic acid through their carboxylic group, they are present as glucuronides (Oikari et al, 1984), G C or L C protocols require a hydrolysis step prior to chromatographic analysis. However, no sample 154 pretreatment was employed in E L I S A procedures. This is because resin acids were conjugated to glucuronic acid in a similar manner to that used for the preparation of the antigens for immunization. A s the site for conjugation was far from the epitope used for antibody binding, the performance of the antibody would not be affected regardless of whether or not the compound is conjugated. This is the great advantage of the E L I S A over the G C . Another advantage in this particular case was that, due to the high resin acid concentration in the bile sample, less than 5 p i of bile was required and various dilutions could be made for the E L I S A . A series of bile samples, obtained from Chinook salmon exposed to pulp m i l l effluents (with 0.8 mg/L of resin acids) for different lengths of time, were assayed by the direct E L I S A and D H A equivalents ranging from 5.5 to 426.1 mg/L were obtained (Table 6.3). A s only a limited amount of fish bile was available, G C analysis of individual bile samples was not possible. However, the bioconcentration factors (BCF) calculated using the D H A equivalents as determined by the E L I S A varied from 6.9 to 532.6, which was in the normal range. A B C F value as high as 996 has been reported (Kruzynski, 1979). Table 6.3 D H A equivalents determined by direct E L I S A in fish bile and calculated bioconcentration factor (BCF) B i le sample D H A equivalents (mg/L) B C F A 159.7 199.6 B 19.9 24.9 C 27.9 34.9 D 426.1 532.6 E 5.5 6.9 F 15.7 19.6 155 In order to compare the E L I S A results with G C , another set of bile samples was pooled and assayed by the direct E L I S A . A 200.0 u l aliquot of this pooled sample was also hydrolysed under alkaline conditions (Oikari et al, 1984) and individual resin acids were quantified by G C . A value of 798.9 mg/L of total resin acids and 775.0 mg/L of abietic type resin acids were detected by G C . The direct E L I S A detected a D H A equivalent value of 826.5 mg/L (Table 6.4). This slightly higher E L I S A value could be explained by the incomplete characterization of some of the G C peaks. A l l the resin acid peaks listed in Table 6.4 were confirmed by G C - M S . Table 6.4 Individual resin acid concentration found in a pooled bile sample by G C and the D H A equivalents by direct E L I S A . Resin acid Concentration (mg/L) S A N 6.4 ISO 17.5 P A L / L E V O 3.4 D H A 743.4 A B A 2.0 N E O 26.2 Total Abietic Type (mg/L) 775.0 Total resin acids (mg/L) 798.9 D H A equivalents by E L I S A (mg/L)* 826.5 +75.9 * Mean value was calculated on three dilutions, each with eight replicate wells. 6.3.4. Effect of hydrolysis on ELISA A s described in the previous section, the E L I S A was performed directly on the bile sample without prior hydrolysis. To evaluate the necessity of the hydrolysis step prior to the 156 E L I S A , the same pooled bile sample was also assayed with and without prior hydrolysis. For the hydrolysed sample, resin acids were extracted with ethyl acetate and a quantity equivalent to 1.0 p i of original bile was evaporated and re-dissolved in P B S buffer at a 1:5000 dilution. For the un-hydrolyzed bile sample, a 1.0 p i aliquot of the original bile was also dissolved in P B S buffer in 1:5000 dilution. The direct E L I S A of these two samples showed that 786.0 ± 52.5 and 831.0 ± 73.0 mg/L of D H A equivalents were obtained for the hydrolysed and un-hydrolyzed samples, respectively. Although the un-hydrolyzed bile showed the higher value, a t-test indicated that there was no significant difference at p<0.05 level. This result demonstrated again that the polyclonal antibodies recognized a molecular entity other than the carboxylic group and that the modification of the carboxylic group, such as by the formation of a glucuronide ester, did not alter the reactivity of the antibodies. Therefore, it is probable that hydrolysis is not necessary for the quantification of resin acids in fish bile using the direct E L I S A . This feature, together with the small sample volume required, constitute the major advantages of the E L I S A over the traditional G C method (Figure 6.1). It should also be recognized that various samples can be processed simultaneously using the E L I S A , leading to a further reduction in analytical time. Since only limited amounts of bile samples were available to us, a corresponding G C analysis could not be performed in each case. However, our preliminary results look promising. This preliminary work indicated that we could quantify the resin acid content of fish bile using a minute sample volume that is inadequate for G C or H P L C analysis. The E L I S A method w i l l be further validated by analyzing further fish bile samples and comparing these results with G C determined values. 157 Bi le (100-500 ul) Hydrolysis (3-4 hrs) Dilution in buffer Extraction & concentration (0.5 hr) E L I S A (2.5 hrs) Derivatization (0.5 hr) Figure 6.1 Steps involved in E L I S A and G C analysis of resin acids in fish bile 6.4. Conclusions Fish bile could provide a sensitive tool for monitoring the concentration of toxic chemicals in receiving waters. This preliminary work demonstrated that the E L I S A could be used as an alternative method of quantifying the resin acid content of the bile from fish exposed to pulp m i l l effluents. Good recoveries were obtained for a control bile spiked with D H A . The D H A equivalents determined by the E L I S A compared favourably with the total abietic type resin acids obtained by G C determination. Since sample hydrolysis was not required by E L I S A , the quantification procedures were greatly simplified compared to those required for G C . 158 CHAPTER 7 SUMMARY AND CONCLUSIONS A literature search of several databases such as Biological Abstracts on C D - R O M , P O L T O X I (Pollution and Toxicology) on CD-ROM, OSH-ROM (Occupational Safety and Health on CD-ROM), M E D L I N E and A Q U A R E F (Canadian water resources and environment related) seemed to indicate that, as yet, no immunochemical method other than the work reported here has been developed for the detection and quantification of resin acids in pulp mill effluents. This thesis describes the development of an enzyme-linked immunosorbent assay (ELISA) for the detection of resin acids in the wastewater streams of pulp mills which used softwoods as the furnish. D H A was selected as the target compound since it is the most abundant and persistent resin acid found in most softwood pulp mill effluents. Both D H A and D H A M , as well as some other related compounds, were used as haptens and various antigens were synthesized by coupling them to carrier proteins, B S A and K L H , with and without insertion of a spacer arm. Among the five immunizing antigens synthesized, polyclonal antibodies were successfully raised by immunization of rabbits with D H A M - S U C - K L H and D H A - K L H , conjugates representing antigens with and without a spacer arm. Assay formats played a significant role in determining the assay sensitivities. Various assay formats, including the conventional indirect ELISA, indirect ELISA with the biotin-streptavidin system, direct ELISA and magnetic bead ELISA, were evaluated for their effects on the assay sensitivities. Among the three formats evaluated using microtitre plates, the I50 ranged from 12.3 ug/L to 113.2 pg/L. The indirect ELISA using the biotin-streptavidin system was the most sensitive format. The direct ELISA had an I50 of 49.7 pg/L and a detection limit of 4.5 p g/L. It was considered to be the method of choice since fewer assay steps were required and the sensitivity was reasonably good. Use of magnetic beads as a solid support further reduced the analytical time. However, the sensitivity was low (I50 of 139.1 pg/L). Assay sensitivity was also 159 evaluated using different antibodies derived from antigens with and without a spacer arm. These different antibodies showed similar sensitivities and cross-reactivities towards resin acids, indicating that the spacer arm did not play a significant role in enhancing the production of antibodies. Since both polyclonal antibodies derived from antigens with and without a spacer arm showed high cross-reactivity to abietic type resin acids, the antibodies could be group specific and adequate for measuring not only DFIA, but also the total abietic type resin acids which are the predominant resin acids found in most effluents when softwoods are used as the furnish. The immunochemical methods were applied to both B K M E and C T M P effluents, which are the wastewater streams from the two major pulping processes used in Canada. For softwood derived C T M P effluents, which usually contain high levels of resin acids, no interference by other effluent components was apparent, probably because a high dilution factor was applied and the dilution eliminated any possible background/matrix effects. The direct E L I S A lead to accurate quantification. The E L I S A results compared favourably with the G C results and a good linear correlation was found for the D H A equivalents, as determined by the E L I S A , and the total abietic type resin acids, as determined by G C . It also correlated wel l with the total resin acid concentration as determined by G C . However, it is wel l known that the chemical constitutes in the C T M P effluents are highly variable and various inorganic/organic materials may interfere with the E L I S A i f they are present in high enough concentrations. In addition, variations in the process conditions and chemicals used during the mechanical pulping and bleaching processes may also affect the assay. It is also possible that, when a hardwood furnish such as aspen is used, the low concentration of resin acids would result in lower dilution requirements. Thus future work should evaluate the effectiveness of the E L I S A under conditions where there are variations in both the feedstocks and process conditions used in mechanical pulp mills. When various B K M E samples were assayed, significant background interference was encountered. In this case, a high dilution could not be used as the low concentration of resin acids in the effluent generally resulted in values which were below the assay detection limit. A 160 non-linear decrease in concentration was predicted for effluent samples serially diluted in assay buffer, due to the non-parallelism observed between the standard curve and the sample dilution curve. Several strategies such as filtration, extraction and fractionation did not eliminate this background effect. Evaluation of the cross-reactivities for some compounds structurally similar to resin acids, such as retene and phytosterols, excluded them from interfering material. The presence of lignin-related phenolic compounds could be probably one of the main sources of interference encountered in the E L I S A of B K M E samples. The interference by the contaminating materials could largely resolved by the use of a bio-treated B K M E effluent which was free of resin acids, as the diluent for both B K M E sample and the medium for the standard curve generation. This greatly improved the assay performance and linearity could be maintained over a wide range of dilutions. Although the sensitivity was decreased (I50 increased from 49.7 to 77.9 pg/L), the E L I S A results compared favourably with the G C results. However, the strategies of using bio-treated B K M E as the diluent has some potential problems as the composition of the bio-treated B K M E w i l l probably differ between mills. It is probably that a bio-treated effluent from one m i l l could not be used for other m i l l effluents as the wood furnish and pulping conditions w i l l be different. Even within the same m i l l , care should be taken when using bio-treated B K M E as the diluent since the composition could change and influence the linearity and precision of the assay. Thus future work should evaluate the compatibility of using a bio-treated effluent from one m i l l when assaying effluents from other mills , which w i l l probably use different wood furnishes and bleaching technologies. Another obvious objective would be to characterize the material interfering with the E L I S A . The direct E L I S A was also used to quantify resin acids present in fish bile. Good recoveries were achieved with control bile sample spiked with D H A . The direct E L I S A of a limited number of bile samples from fish exposed to pulp m i l l effluents indicated that this immunochemical method did not require sample hydrolysis and much less sample was need (1-5 p i for E L I S A vs 100 - 500 pi for G C ) . The E L I S A could be an ideal alternative for the currently 161 used G C method as rapid resin acid determination could be performed on numerous individual fish. Compared with conventional instrumental methods such as G C , H P L C and G C - M S , which usually require extensive sample pre-treatment such as extraction, derivatization and chromatographic separation/detection, and which suffer from such disadvantages as high analytical cost and low sample output, the developed immunochemical method required less sample pre-treatment and was relatively easy to perform. Multiple samples could be processed simultaneously. However, two of the drawbacks of the E L I S A are, the considerable amount of time that has to be initially invested in the production of antibodies and the lack of information that is provided on the concentrations of individual resin acids. Whi le the results of this initial work are promising, it is recognized that the effluent samples that were tested were derived from a limited number of pulp mills. Therefore, it is important that future work should evaluate the E L I S A method for detecting resin acids in various m i l l streams such as effluent samples from diverse sources using various pulping methods and different wood furnishes. Although some good basic information about the action of polyclonal antibodies against resin acids has been obtained, several questions have yet to be resolved. For example, how can we develop more generic antibodies which can be used to quantify all resin acids? While polyclonal antibodies recognised mainly abietic type resin acids, is it possible to raise antibodies specific to pimaric type resin acids? H o w can we develop a novel assay format to use these polyclonal antibodies to quantify total resin acids i f both group specific antibodies were available? Is it possible to quantify individual resin acids in samples containing resin acid mixtures? It is probable that use of monoclonal antibodies would improve both the supply and quality of the antibodies. One of our continuing research goals is to isolate possible biotransformation intermediates or products, to help elucidate the resin acid biotransformation pathway. A s it has been demonstrated that antibodies against D H A cross-reacted with other resin acids, it is quite possible that the antibodies w i l l also cross-react with these intermediates i f the 162 epitopes are still present in these molecules. It should be possible to use methods such as affinity chromatography to isolate and identify some of resin acid degradation products. Although the microtitre plate based E L I S A s can be sensitive enough to detect and quantify compounds of interest in the laboratory, they require precision pipetting, incubation steps of one hour or more and the equipment needed to separate the free and the bound assay components. Therefore, the development of an E L I S A kit suitable for on-site detection would be more useful. Wi th the basic information obtained from this study, it should be possible to develop a dip-stick type kit which could be used to monitor resin acids directly at m i l l sites. 163 REFERENCES A g a , D . S., Thurman, E . M . and Pomes, M . L . (1994), Determination of alachlor and its sulfonic acid metabolite in water by solid-phsae extraction and enzyme-linked immunosorbent assay, Anal. Chem., 66, 1495-1499, A l - K a i s s i , E . and Mostratos, A . (1983), Assesment of substrate for horseradish peroxidase in enzyme immunoassays, J. Immunol. Meth., 58, 127-132. Albro , P. W . , Luster, M . I., Chae, K . , Chaudhary, S. K . , Clark, G , Lawson, L . D . , Corbett, J. T. and M c K i n n e y , J. D . (1979), A radioimmunoassay for chlorinated dibenzo-p-dioxins, Toxicol. Appl. Pharmacol, 50, 137-146. A m i c o n Inc. (1991), M A C Discs-Protein A , Amicon Publication N o . 1-383A. Analytichem International (1990), Sorbent Extraction Technology Handbook (ed by K . C. V a n Home), Analytichem International Inc. Beak Consultants L td . (1987), "Non-conventional Contaminants in Pulp and Paper M i l l Effluents", Prepared for Ontario Forest Industries Association. Belefant, H . , Fong, F. (1989), Abscisic acid E L I S A : organic acid interference. Plant Physiol, 91, 1467-1470. Berson, S. A . and Yalow, R. S. (1968), General principle of radioimmunoassay, Clin. Chim. Acta, 22,51-69. Bicho, P. A . , Strub. R. and Saddler, J. N . (1995a), Resin acid analysis laboratory round robin, final report, (unpublished). Bicho, P. A . Martin, V . and Saddler, J. N . (1995b), Growth, induction and substrate specificity of dehydroabietic acid-degeading bacteria isolated from a kraft m i l l effluent enrichment, Applied Environ. Microbiol, 61, 3245-3250. Bicho, P. A . , Moosbrugger, R , MacLean B . , Stark, K . and Saddler, J. N . (1995c), Sources of resin acids in the Quesnel River pulp m i l l (unpublished). Bonsor, N . , M c C u b i b , N . and Sprague, J. B . (1988), "Kraft m i l l effluent in Ontario". 260pp. Ontario Ministry of the Environment, Municipal-Industrial Strategy for Abatement ( M I S A ) , Technical Advirory Committee, Pulp and Paper Sector of M I S A , Queens's Printer for Ontario. Bookbinder, M . J. and Panosian, K . J. (1986), Correct and incorrect estimation of with-day and between-day variations, Clin. Chem., 32, 1734-1737. 164 Browning, B . L . (1967), Examination of the resin acid fraction, in "Method of Wood Chemistry", (ed. by B . L . Browning), Interscience Publishers, New York, p. 161-188. Bullock, C , Bicho, P. and Saddler, J. N . (1995). The influence of the high molecular weight fraction of a bleached kraft m i l l effluent on the microbial activity of activated sludge, Wat. Res., (in press). Butt, W . R. (1984), "Practical Immunoassay, The State of The Art " , Marcel Dekker, Inc., N e w York. Calverlry, R. A . , Jackman, R. and Pembroke-Hatterslry, J. J. (1987), Solide-phase sample preparation techniques for immunoassay, in "Immunoassays for Veterinary and Food Analysis-1" (ed. by B . A . Morris , M . N . Cli f ford and R. Jackman), Elsevier Applied Science, London and N e w York , p.93-108. Carpenter, W . L . (1965), Colorimetric measurement of resin acid soaps in pulp and paper m i l l wastes, TAPPI, 48(11), 669-671. Carter, K . R. (1992), On-site screening speeds sample analysis, Pollution Engineering, March 15, 1992, p.36. Carter, C. W . and Suffer, I. H . (1982), Binding of D D T to dissolved humic materials, Environ. Sci. Technol, 16 (11), 735-740. C E P A (1990), Canadian Environmental Protect Act, "Priority Substances Lis t Assessment Report N o 1, Polychlorinated Bibenzodioxins and polychlorinated dibenzofurans", Environmental Protection, Conservation and Protection, Environment Canada, Minister of Supply and Services Canada, Ottawa, Canada. C E P A (1991), Canadian Environmental Protect Act , "Priority Substances List Assessment Report N o 2, Effluents from Pulp M i l l s Using Bleaching", Environmental Protection, Conservation and Protection, Environment Canada, Minister of Supply and Services Canada, Ottawa, Canada. C F I (1995), Council of Forest Industries, "British Columbia Forestry Industry Fact Book", August, 1995. C F I (1995), Council of Forest Industries, "British Columbia Forest Industry Statistical Tables", A p r i l , 1995. Chait, E . M . and Ebersole, R. C. (1981), Clinical Analysis, A perspective on chromatographic and immunoassay technology, Anal. Chem., 53, 682A-692A. 165 Chau, A . S. Y . and Afghan, B . K . (1982), "Analysis of Pesticides in Water", C R C Press, Boca Raton, F L , V o l . I, II and III. Chen, T., Breuil , C , Carriere, S. and Hatton, J. V . (1994), Solid-phase extraction can rapidly separate l ipid classes from acetone extracts of wood and pulp, TAP PI Journal, 11 (3), 235-240. Cherwinsky, C. and Murray, D . (1986), "Preliminary investigation of trace contaminants in pulp and paper m i l l effluents", Ontario Ministry of the Environment, Water Resources Branch, Queen's Printer for Ontaruo, Toronto. Collins, W . P. (1985), "Alternative Immunoassays", John Wiley & Sons L t d . , N e w York. Cornacchio, L . A . and H a l l , E . R. (1988), Characteristics of C T M P wastewaters, W T C - B I O - 0 7 , Environment Canada, p. 1-14. C P P A (1995), "Reference Tables", Canadian Pulp and Paper Association. Debrabandere, L . , V a n Boven, M . , Laruelle, L . and Daenens, P. (1993), Routine detection of buprenorphine in horse urine, Anal. Chim. Acta, 275, 295-303. Dunbar, B . S. and Schwoebel, E . D . (1990), Preparation of Polyclonal Antibodies, in "Methods in Enzymology", V o l 182, "Guide to Protein Purification", (ed. by M . P. Deutscher), Academic Press, Inc., N e w York. Dynal A . S. (1994), Dynalbeads M-280 Tosylactivated, Product Information. Ekins, R. P. (1980), More sensitive immunoassays, Nature, 284, 14-15. Ekman, R. and Holmbom, B . (1989), Analysis by chromatography of the wood extractives in pulp and water samples from mechanical pulping of spruce, Nordic Pulp and Paper Res. J., 1, 16-24. Enviro-Test Laboratories (1990), "Enviro-Test Laboratories Fee Schedule". E P S (1971/72), "Pulp and Paper Effluent Regulations", Codes and Protocols Report 1, 7pp, Environment Canada, Environ. Prot. Service, Ottawa, Ontario. E P S (1978), "Guidelines for the Pulp and Paper Effluent Regulations", Codes and Protocols Report 2, Environment Canada, Environmental Protection Service, Ottawa, Ontario. E P S (1982), "Status Report on Abatement of Water Pollution from the Canadian Pulp and Paper Industry", Environment Canada, Environmental Protection Service, R e p t E P S 1/PF/l . 166 Erlanger, B . F. (1980), The preparation of antigenic hapten-carrier conjugate: a survey, in "Methods in Enzymology", V o l 70, "Immunochemical Techniques, Part A " (ed. by H . V . Vunakis and J. J. Langone), p.85. Farah, F. S., Kern, M . and Eisen, H . N.(1960), The preparation and some properties of purified antibodies specific for the 2,4-dinitrophenyl group, J. Exp. Med., 112, 1195-1210. Feng, P. C. C , Sharp, C. R. and Horton, S. R. (1994), Quantitation of alachlor residues in monkey urine by E L I S A , J. Agric. Food Chem., 42, 316-319. Fengel, D . and Wegener, G . (1984), "Wood: Chemistry, Ultra-Structure, Reactions", Walter de Gruyter & Co. , Berl in, p. 182-221, 258-262, 555-560. Fleeker, J. (1987), Two enzyme immunoassays to screen for 2,4-dichlorophenoxyacetic acid in water, J. Assoc. Off. Anal. Chem., 70 (5), 874-878. Foster, D . O. and Zinkel , D . F. (1982), Qualitative and quantitative analysis of diterpene resin acids by glass capillary gas-liquid chromatography, J. Chromatog., 248, 89-98. Frutos, M . E . and Regnier, F. E . (1993), Tandem chromatographic-immunological analysis, Anal. Chem., 65 (1), 17A-25A. Furniss, B . S., Hannaford, A . J., Rogers, V . , Smith, P. W . G . and Tatchell. A . R. (1978),. in "Vogel's Textbook of Practical Organic Chemistry", 4th ed., Longman, New York, pp. 129. Gafre, G . and Milstein, C. (1981), Preparation of monoclonal antibodies, strategies and procedures, in "Methods in Enzymology, V o l 73, Immunochemical techniques, Part B " p.3 (ed. by Vunakis, H . V . and Langone, J. J . ) Gee, S. J., Hammock, B . D . and Van Emon, J. M . (1994), " A User's Guide to Environmental Immunochemical Analysis", U S Environmental Protection Agency, Washington, D . C , 1994, EPA/540/R-94/509. Gee, S. J., Miyamoto, T., Goodrow, M . H . , Buster, D . and Hammock, B . D . (1988), Development of enzyme-linked immunosorbent assay for the analysis of the thiocarbate herbicide molinate, J. Agric. Food Chem., 36, 863-870. Gottstein, W . J. and Cheney, L . C. (1965), Dehydroabietylamine, A new resolving agent, J. Org. Chem., 30,' 2072-2073. Habeeb, A . F. S. F. (1966), Determination of free amino groups in proteins by trinitrobenzenesulfonic acid, Anal. Biochem., 14, 328-336. Halbrook, N . J. and Lawrence, R. V . (1966), The isolation of dehydroabietic acid from disproportionated rosin, J. Org. Chem., 31 (12), 4246-4247. 167 H a l l , E . A . H . (1991), "Biosensors", Prentice H a l l , Englewood Cliffs , N e w Jersey. H a l l , J. C , Deschamps, R. and McDermott, M . R. (1990), Immunoassays to detect and quantitate herbicides in the environment, Weed Technol., 4, 226-234. Halliburton, D . and Chung, A . (1991), Mornitoring and Regulation of Bleached Kraft M i l l Effluents: The Canadian Experience, Proceedings from International Conference on Blleached Kraft Pulp M i l l s , February 1991, Melbourne, Australia, p.397-408. Hammock, B . D . , Gee, S. J., Cheung, P. Y . K . , Miyamoto, T., Goodrow, M . H . , V a n Emon, J. M . and Seiber, J. N . (1986), in "Pesticide Science and Biotechnology", (ed. by R. Greenhalgh and T. R. Robert), Blackwell Scientific Publications, Oxford, 1987, p.309-316. Harlow, E . and Lane, D . (1988), "Antibodies, a laboratory manual", Cold Spring Harbor Laboratory. Harrison, R. O. , Braun, A . L . , Gee, S. J., O'Brian, D . J. and Hammock, B . D . (1989), Evaluation of an E L I S A for the direct analysis of molinate in rice field water, Food Agric. Immunol, 1, 37-51. Harrison, R. O. , Gee, S. J. and Hammock, B . D . (1988), Immunochemical methods of pesticide residue analysis, in "Biotechnology for Crop Protection" (ed. by P. A . Hedin, J. J. Menn and R. M . Hollingworth), A C S Symposium Series 379, Chapter 24, pp 316-330. Harrison, R. O. , Goodrow, M . H . , Gee, S. J. and Hammock, B . D . (1990), Hapten synthesis for pesticide immunoassay development, in "Immunoassays for Trace Chemical Analysis, Mornitoring Toxic Chemicals in Humans, Food, and the Environment", A C S Symposium Series 451, American Chemical Society, Washington, D C , p. 14-27. Heitzman, R. J. (1988), Immunoassay techniques for measuring veterinary drug residues in farm animals, meat and meat products, in "Immunoassays for Veterinary and Food Analysis-I" (ed. by B . A . Morris , M . N . Clif ford and R. Jackman), Elsevier Applied Science, London and N e w York, p.73-115. Hemingway, R. W . and Hi l l i s , W . E . (1971), Changes in fats and resin acids of Pinus radiata associated with heartwood formation. APPITA, 24 (6), 439-443. Herbert W . J. (1978), in "Handbook of experimental immunology", (ed. by Weir D . M . , 3rd edn.), Blackwel l Scientific Publications, Oxford, p. A.4 .1 . Hock, B . (1993), Enzyme immunoassays for pesticide analysis, Acta Hydrochim. Hydrobiol, 21 (2), 71-83. 168 Hottenstein, C. S., Jourdan, S. W . , Hayes, M . C , Rubio, F. M . , Herzog, D . P. and Lawruk, T. S. (1995), Determination of pentachlorophenol in water and soil by a magnetic particle-based enzyme immunoassay, Environ. Sci. Technol., 29, 2754-2758. Hough, G . (1985), "Chemical Recovery in the Alkaline Pulping Precesses", T A P P I Press, Atlanta, G A . Huber, S. J. and Hock, B.(1985), A solid-phase enzyme immunoassay for quantitative determination of the herbicide terbutryn, J. Plant Dis. Protect., 95, 147-156. Hudy, J. A . (1959), Resin acids-gas chromatography of their methyl esters, Anal. Chem., 31, 1754-1756. Hung, D . T., Benner, S. A . and Wil l iams, C. M . (1980), Preparation of an ecdysone immunogen for radioimmunoassay work, J. Bio. Chem., 255 (13), 6047-6048. Hunter, W . M . and Corrie, J. E . T. (1983), "Immunoassays for Clinical Chemistry", 2nd. edn, Churchil l Livingstone, Edinburgh. Jamieson S. (1991), Y o u don't have to treat your effluent nicely, you just have to treat it wel l , Pulp & Paper Canada, 92, 21-27. Janeway, C. A . Jr. and Travers, P. (1994), "Immunobiology: The Immune System in Health and Disease", London, Current Biology Limited and New York, Garland Publishing Inc. 1994. Johannsson, A . , E l l i s , D . H . , Bates, D . L . , Plumb, A . M . and Stanley, C. J. (1986), Enzyme amplification for immunoassay detection limit of one hundredth of an attomole, J. Immunol. M?//* . , 87, 7-11. Johnson, J. C. and V a n Emon, J. M . (1994), Development and evaluation of a quantitative enzyme linked immunosorbent assay for polychlorinated biphenyls, U S Environmental Protection Agency, Washington, D . C , 1994, EPA/600/R-94/112. Jones, S. A . , Halliburton, D . , Carey, J. H . , Carlisle, D . B . , Colodey, A . G . , Meyer, A . , Lockhart, W . L . , and Rogers, I. H . , (1991). Priority substances list assessment report N o . 2: effluents from pulp mills using bleaching. Canadian Environmental Protection Act , Environmental Protection, Conservation and Protection, Environment Canada, Minister of Supply and Services Canada, Ottawa, Canada. Judd, M . C , Stuthridge, T. R , Tavendale, M . H . , McFarlane, P. N . , Mackie, K . L . , Buckland, S. J., Randall, C. J., Hickey, C. W . , Rope, D . S., Anderson, S. M . and Steward, D . (1995), Bleached kraft pulp m i l l sourced organic chemicals in sediments from N e w Zealand rivers. Chemosphere, 3 0 (9), 1751-1765. 169 Jung, F. , Gee, S. J., Harrison, R. O., Goodrow, M . H . , Karu, A . E . , Braun, A . L . , L i , Q. X.and Hammock, B . D . (1989), Use of immunochemical techniques for the analysis of pesticide residues. Pestic. Sci., 26, 303-317. Kamata, N . , Enomoto, A . , Ishida, S., Nakamura, K . , Kurisaki , J. and Kaminogawa S. (1996), Comparison of p H and ionic strength dependence of interactions between monoclonal antibodies and bovine P-lactoglobulin, Biosci. Biotech. Biochem., 60 (1), 25-29. Kaufman, B . M . and Clower, M . Jr. (1991), Immunoassay of pesticides, J. Assoc. Off. Anal. Chem., 74, 239-247. Kaufman, B . M . and Clower, M . Jr.(1995), Immunoassay of pesticides: an update, J. Assoc. Off. Anal. Chem., 78 (4), 1079-1090. Keith, L . H . , Crummett, W . , Deegan, J., Libby, R. A . , Taylor, J. K . and Wentler, G . (1983), Principles of environmental analysis, Anal. Chem., 55, 2210-2218. K e l l y , M . M . , Zahnow, E . W . , Petersen, W . C , Toy, S. T. (1985), Chlorsulfuron determination in soil extracts by enzyme immunoassay, J. Agric. Food Chem. 33, 962-965. Kramer, P. M . , Marco, M . P. and Hammock, B . D . (1994), Development of a selective enzyme-linked immunosoebent assay for 1-naphthol-the major metabolite of carbaryl (1-naphthyl N -methylcarbamate). J. Agric. Food Chem., 42, 934-943. Kruzynski , G . M . (1979), Some effects of dehydroabietic acid on hydromineral balance and other physiological parameters in juvenile sockeye salmon Onchorhynchus nerka. Ph. D . dissertation, University of British Columbia, Vancouver, B . C , Canada. Kukkonen, J. (1992), Effects of lignin and chlorolignin in pulp m i l l effluents on the binding and bioavailability of hydrophobic organic pollutants, Wat. Res., 26 (11), 1523-1532. Kulovaara, M . , Kronberg, L . and Pensar G . (1987), Recoveries of some chlorophenolics and resin acids from humic water, Sci. Tot. Environ., 62, 291-296. Kurstak, E . (1986), "Enzyme Immunodiagnosis", Academic Press, N e w York. Kurzer, F. and Douraghi-Zadeh, K . (1967), Advances in the chemistry of carbodiimides, Chem. Rev., 67, 107-137. Kutney, J. P., Berset, J. D . , Hewitt, G . M.and Singh, M . (1988), Biotransformation of dehydroabietic, abietic and isopimaric acids by Mortierella isabellina immobilized in polyurethane foam, Appl. Environ. Microbio., 54(4), 1015-1022. Kutney, J. P. , Singh, M . , Hewitt, G , Salisbury, P. H . , Worth, B . R , Servizi, J. A . , Martens D . W . and Gordon R. W . (1981), Studies related to biological detoxification of kraft pulp m i l l 170 effluent I. The biodegradation of dehydroabietic acid with Mortierella isabellina, Can. J. Chem., 59, 2334-2341. Lachapelle, L . (1996), What we must do to meet global challenges, Papermaker, January, 1996, p.25. Landsteiner, K . (1945), "The Specificity of Serological Reactions", Harvard University Press, Cambridge. Lawruk, T. S., Gueco, A . M . , Jourdan, S. W . , Scutellaro, A . M , Fleeker, J. R , Herzog, D . P. and Rubio, F. M . (1995), Determination of chlorothalonil in water and agricultural products by a magnetic particle-based enzyme immunoassay, J. Agric. Food Chem., 43, 1413-1419. Leach, J. M . and Thakore, A . N . (1973), Identification of the constituents of kraft pulping effluent that are toxic to juvenile coho salmon (Onchorynchus kisutch), J. Fish. Res. Brd. Can., 30, 479-484. Leach, J. M . and Thakore, A . N . (1976), Toxic constituents in mechanical pulping effluents, TAPPI, 59, 129-132. Leach, J. M . and Thakore, A . N . (1977), Compounds toxic to fish in pulp m i l l waste streams, Progress in Water Technology, 9 (4), 787-798. Lee, H . B . , Peart, T. E . , Carron, J. (1990), Gas chromatographic and mass spectrometric determination of some resin and fatty acids in pulp mi l l effluents as their pentafluorobenzyl ester derivatives, J. Chromatog., 498, 367-379. Lee, H . B . and Peart, T. E . (1992), Super critical carbon dioxide extraction of resin and fatty acids from sediments at pulp m i l l sites. J. Chromatog., 594, 309-315. Lesnik, B . (1994), Immunoassay methods: development and implementation programme at the U S E P A , Food Agric. Immunol., 6, 251-259. L i , K . , Chester, M . , Kutney, J. P., Saddler, J. N . and Breuil , C. (1994), Production of polyclonal antibodies for the detection of dehydroabietic acid in pulp m i l l effluents, Anal. Lett, 27 (9), 1671-1688. L i , K . , Chen, T., Bicho, P., Breuil , C. and Saddler, J. N . (1996a), A comparison of gas chromatographic and immunochemical method for quantifying resin acids, in "Environmental Fate and Effect of Pulp and Paper M i l l Effluents" (ed. by M . R. Servos, K . R. Munkittrick, J. H . Carey and G . J. V a n der Kraak), St. Lucie Press, Florida, p. 119-127. L i , K . , Chen, T., Bicho, P. , Breuil , C. and Saddler, J. N . (1996b), Factors affecting gas chromatographic analysis of resin acids present in pulp m i l l effluents, Toxicol. Environ. Chem. (in press). 171 Linde, C. D . and Goh, K . S. (1995), Immunoassays (ELISAs) for pesticide residues in environmental samples, Pesticide Outlook, August, p. 18-23. L i u , H . W . , L o , S. N . and Lavallee, H . C. (1991), Characteristics of pollutants of source and biological treatment of a C T M P effluent, APPITA, 44, 133-138. L o , S. N . , Lavallee, H . C , Rowbottom, R. S., Meunier, M . A . and Zaloum, R. (1994), Activated sludge treatment of T M P mi l l effluents, TAPPI Journal, 77 (11), 167-178. Lucas, A . D . , Gee, S. J., Hammock, B . D . and Seiber, J. N . (1995), Integration of immunochemical methods with other analytical techniques for pesticide residue determination, J. Assoc. Off. Anal. Chem., 78 (3), 585-591. Luster, M . I., Albro , P. W . , Chae, K . , Lawson, L . D . , Corbett, J. T. and McKinney , J. D . (1980), Radioimmunoassay for quantitation of 2,3,7,8-tetrachlorodenzofurans, Anal. Chem., 52, 1497-1500. MacLean, B . , Vegt, A . and Driel , E . (1990), Fu l l scale anaerobic/aerobic treatment of T M P / B C T M P effluent at Quesnel River Pulp, in Proceedings of T A P P I Environmental Conference, Seattle, W A . A p r i l 9-11. p.647-661. MacLeay, D . J. (1991), Canadian Approaches and Test Methods for Determining and Monitoring the Aquatic Toxic Effects of B K M E , Proceedings from International Conference on Bleached Kraft Pulp M i l l s , February 1991, Melbourne, Australia, p.281-311. Maguire, R. J. (1994), Dissolved organic matter can reduce the efficiency of extraction of lipophilic chemicals from water, Preprint Extended Abstract, Divison of Environmental Chemistry, Americam Chemical Society, Washington, D C , August 21-25, 1994. Mahood, W . H . and Rogers, I. H . (1975), Separation of resin acids from fatty acids in relation to environmental studies, J. Chromatog., 109, 281-286. Malakian, A . , Golebiowska, M . and Bellefeuille, J. (1993), American Laboratory, Feb. 40R-40V. Marco, M . P. , Gee, S. J., Cheng, H . M . , Liang, Z . Y . and Hammock, B . D . (1993a), development of an enzyme-linked immunosorbent assay for carbayl, J. Agric. Food Chem., 41, 423-430. Marco, M . P., Hammock, B . D . and Kurth, M . J. (1993b), hapten design and development of an E L I S A for the detection of the mercapturic acid conjugates of naphthalene, J. Org. Chem., 58, 7548-7556. 172 Mayr, M . , Lorbeer, E . and Kratzl , K . (1982), Gas chromatographic separation of diterpene acids on glass capilary columns of different polarity, J. Amer. Oil Chem. Soc., 59 (1), 52-57. McCarthy, P. J., Kennedy, K . J. and Droste, R. L . (1990), Role of resin acids in the anaerobic toxicity of chemithermomechanical pulp wastewater, Wat. Res., 24 (11), 1401-1405. McFarlane, V . A . and Clark, J. U . (1989), Environmental occurrence, abundance, and potential toxicity of poly chlorinated biphenyl cogeners: considerations for congener-specific analysis, Environ. Health Perspect, 81, 225-239. McFarlane, P. N . , Stuthridge, T. R. and Clark, T. A . (1991), Fundamental aspects of the biological treatment of bleached kraft m i l l wastewaters, Proceedings from International Conference on Bleached Kraft Pulp M i l l s , February 1991, Melbourne, Australia, p.333-361. M c L e a y and Associates Ltd . (1987), "Aquatic Toxicity of Pulp and Paper M i l l Effluents: A Review". Prepared for Environment Canada, Fisheries and Oceans Canada, Canadian Pulp and Paper Assoc. and Ontario Ministry of the Environment, Report E P S 4 / P F / l . M c M a h o n , D . H . (1980), Analysis of low levels of fatty and resin acids in kraft m i l l process streams. TAPPI, 63(9), 101-103. Midgley, A . R , Niswender, G . D . and Rebar, R. W . (1969), Principles for the assessment of reliability of radioimmunoassay methods (precision, accuracy, sensitivity, specificity), Acta Endocrinol. Suppl., 142, 163-179. M i m m s , A . (1993), "Kraft Pulping, A Compilation of Notes", T A P P I Press, Atlanta, G A . M I S A (1993), "Draft Development Document for the Pulp and Paper Scetor Effluent Limits Regulation", I S B N 0-7778-0837-4, Environment Ontario, Queen's Printer for Ontario. Monroe, D . (1984), Enzyme Immunoassay, Anal. Chem., 56, 920A-931A. Morales, A . , Birkhole, D . A . and Hrudey, S. E . (1992), Analysis of pulp m i l l effluent contaminants in water, sediment, and fish bile —fatty and resin acids, Wat. Environ. Res., 64 (5), 660-668. Morck, R , Reimann, A . and Dahlman O. (1991), Chracterization of high molecular weight organic materials in modern softwood and hardwood bleached kraft m i l l effluents, in Environmental Fate and Effects of Bleached Kraft M i l l Effluents, Proceedings of SEP A , 155-163. Morgan, M . R. A , McNerney, R , Matthew, J. A . , Coxon, D . T. and Chan, H . W . S. (1983), A n enzyme-linked immunosorbent assay for total glycoalkaloids in potato tubers, J. Sci. Food Agric, 34, 593-598. 173 Munoz, M . J., Castano, A . , Blazquez, J., Vega, M . , Carbonell, G . , Ortiz, J. A . , Carballo, M . and Tarazona, J. V . (1994), Toxicity identification evaluation for the investigation of fish ki l ls : a case study, Chemosphere, 29(1), 55-61. Nahoul, K . , Patricot, M - C , Moatti, J-P. and Revol, A . (1992), Determination of urinary Cortisol with three commercial immunoassays, J. Steroid Biochem. Molec. Biol, 43 (6), 573-580. N C A S I (1972), "Evaluation of analytical procedures for the analysis of selected organic compounds in kraft effluents", Natl . Counc. Pap. Ind. A i r Stream Improv. Tech. B u l l . N o . 258. N C A S I (1975), "Improved procedures for the gas chromatographic analysis of resin and fatty acids in kraft m i l l effluents", Natl . Counc. Pap. Ind. A i r Stream Improv. Tech. B u l l . N o . 281. N C A S I (1978), " A n examination of porous polymer resins for analytical use and removal of compounds in kraft effluents responsible short term bioassay response", Natl . Counc. Pap. Ind. A i r Stream Improv. Tech. B u l l . N o . 302. N C A S I (1986), "Procedures for the analysis of resin and fatty acids in pulp m i l l effluents", Natl . Counc. Pap. Ind. A i r Stream Improv. Tech. B u l l . N o . 501. Ngo , T. T. (1991), Enzyme systems and enzyme conjugates for solid-phase E L I S A , in "Immunochemistry of Solid-Phase Immunoassay" (ed. by J. E . Butler), C R C Press, p.86-102. Nestler, F. H . M . and Zinkel , D . F. (1963), Separation of the methyl esters of resin acids by gas liquid chromatography, Anal Chem., 35, 1747-1749. Nisonoff, A . (1982), "Introduction to Molecular Immunology", Sinauer and Associates, Inc. Sunderland, Massachusetts. Oikari , A . and Holmbom, B . (1981), Analysis of trout bile can be used for monitoring resin acids load in receiving waters. S I T R A Workshop on the effects of pulp m i l l bleaching plant effluents on the receiving waters: Espoo/Otaniemi. Oikari , A . , Anas, E . , Kruzynski , G . and Holmbom, B . (1984), Free and conjugated resin acids in the bile of rainbow trout, Salmo gairdneri. Bull Environ. Comtam. Toxicol, 33, 233-240. Oikari , A . O. J. (1986), Metabolites of xenobiotics in the bile of fish in waterways polluted by pulp m i l l effluents, Bull. Environ. Contam. Toxicol, 36, 429-436. Orsa, F. and Holmbom, B . (1994), A convenient method for the determination of wood extractives in papermaking process waters and effleunts,. Journal of Pulp and Paper Science, 20 (12), J361-366. Paraf, A and Peltre, G . (1991), "Immunoassays in Food and Agriculture", Kluwer Academic, Boston. 174 Paxton, J. W . , Rowel l , F. J. and Ratcliff, J. E . (1976), Production and characterization of antisera to diphenylhydantoin suitable for radioimmunoassay, J. Immunol. Meth., 10, 317-327. Paul, W . E . (1984), "Fundamental Immunology", Raven Press, New York. Peterman, J. H . (1991), Immunochemical considerations in the analysis of data from non-competitive solid-phase immunoassay, in "Immunochemistry of Solid-Phase Immunoassay" (ed. by J. E . Butler), C R C Press, p.47-65. Peterman, J. H . and Butler, J. E . (1989), Application of theoretical considerations to the analysis of E L I S A data, Bio. Tech., 7, 608. . Pratt, Y . (1951), Derivatives of dehydroabietic acid, J. Amer. Chem. Soc, 73, 3803-3807. Priha, M . H . and Talka, E . T. (1986), Biological activity of bleached kraft m i l l effluent ( B K M E ) fractions and precess streams. Pulp Pap. Can., 87, 143-147 (T447). Ramakrishna, N . , Lacey, J., Candlish, A . A . G . , Smith, J. E . and Goodbrand, I. A . (1990), Monoclonal antibody-based enzyme linked immunosorbent assay of alatoxin B l , T-2 toxin, and Ochratoxin A in barley, J. Assoc. Off. Anal. Chem., 73(1), 71-76. Rappe, C. (1984), analysis of polychlorinated dioxins and furans, Environ. Sci. Technol., 18, 78A-90A. Richardson, D . E . , Bremmer, J. B . and O'Grady, B . V . (1992), Quantitative analysis of total resin acids by high performance liquid chromatography of their coumarine ester derivatives. J. Chromatog., 595, 155-162. Richardson, D . E . and Bloom, H . (1982), Analysis of resin acids in untreated and biologically treated thermomechanical pulp effluent. APPITA, 35(6), 477-482. Richardson, D . E . , Bremmer, J. B . and O'Grady, B . V . (1992), Quantitative analysis of total resin acids by high performance liquid chromatography of their coumarine ester derivatives. J. Chromatog., 595, 155-162. Rittenberg, M . B . and Amkraut, A . A . (1966), Immunogenicity of trinitrophenyl-hemocyanin: Production of primary and secondary anti-hapten precipitins. J. Immunol., 97, 421-430. Rittenburg, J. H . , Grouthaus, G . D . , Fitzpatrick, D . A . and Lankow, R. K . (1990), Rapid on-site immunoassay systems, in "Immunoassays for Trace Chemical Analysis", A C S Symposium Series 451, American Chemical Society, Washington, D C , p.28-39. 175 Robins, R. J., Morgan, M . R. A . , Rhodes, M . J. C. and Furze, J. M . (1985), Cross-reactions in immunoassays for small molecules, in "Immunoassays in Food Analysis" (ed. by B . A . Morris and M . N . Clifford), Elsevier Applied Science, London, p. 197-211. Robins, R. J. (1986), The measurement of low-molecular-weight, non-immunogenic compounds by immunoassay, in "Immunology in Plant Sciences" (ed. by H . F. Linskens and J. F. Jackson), Springer-Verlag, p.86-141. Rogers, I. H . (1973), Isolation and chemical identification of toxic components of kraft m i l l waters. Pulp Paper Canada, 74 (9), 111-116. Rogers, I. H . , Harris, A . G . and Rozon, L . R. (1971), The effect of outside storage on the extractives of white spruce and lodgepine. Pulp Paper Mag. Can. 72(6), T214-220. Rogers, I. H . and Mahood, H . W . (1981), Oxidized resin acids in douglas fir wood extractives, in "Advances in the Identification and Analysis of Organic Pollutants in Water" (ed. by L . H . Keith), A n n Arbor Science Publishers, Inc., Michigan, V o l 2, pi097-11,13. Rosner, B . (1986), "Fundamentals of Biostatistics" (2nd. edn.), Duxbury Press, Boston. Rowe, J. W . (1969), "Common and Systematic Nomenclature of Cycl ic Diterpenes", Forest Products Laboratory, Forest Service, U S D A , Madison, Wisconsin. Schneider, P. and Hammock, B . (1992), Influence of the E L I S A format and the hapten-enzyme conjugate on the sensitivity of an immunoassay for s-triazine herbicides using monoclonal antibodies, J. Agric. Food Chem., 40, 525-530. Shepard, D . B . , Sigfusson, D . A . and Sim S. Y . (1996). Effluent management: Determination of resin acids i n pulp m i l l effluent. Preprints of Papers to be presented at the 1996 Spring Conference, Canadian Pulp & Paper Association, Session 5, Paper 3, M a y 15-18, 1996, Jasper, Alberta, Canada. Sherry, J. P. (1992), Environmental chemistry: the immunoassay option, CRC Crit. Rev. Anal. Chem., 23 (4), 217-300. Shrimpton, D . M . (1973), Extractives associated eith wound response of lodgepole pine attacked by the mountain pine bettle and associated microorganisms, Can. J. Bot., 51, 527-533. Sithole, B . B . (1993), A rapid spectrophotometric procedure for the determination of total resin and fatty acids in pulp and paper matrices, TAPPIJournal, 76 (10), 123-127. Sjostrom, E . (1993), "Wood Chemistry Fundamentals and Applications", (2nd. edn.), Academic Press Inc., N e w York. 176 Smith, D . S., A l - H a k i e m , M . H . H . and London, J. (1981), A review of fluoroimmunoassay and immunofluorometric assay. Ann. Clin. Biochem., 18, 253-274. Soltes, E . J. and Zinkel , D . F. (1989), Chemistry of rosin, in "Naval Stores, Production-Chemistry-Utilization" (ed. by D . F. Zinkel and J. Russel), Pulp Chemicals Association, N e w York , p.261-345. Sprague, J. B . (1991), Environmentally desirable approaches for regulating effluents from pulp mills , Wat. Sci. Tech., 24, 361-371. Springer, A . M . (1993), Overview of water pollution and their impact: pulp and paper industry, in "Industrial Environmental Control Pulp and Paper Industry" (ed. by A . M . Springer, 2nd. ed.), T A P P I Press. Stanker, L , H . , Watkins, B . , Roger, N . and Vanderlann, M , (1987), Monoclonal antibodies for dioxin: antibody chracterization and assay development, Toxicology, 45, 229-243. Stuthridge, T. R , Tavendale, M . H . , Judd, M . C , Mackie, K . L . and McFarlane, P. N . (1992), Determination of bleached kraft m i l l effluent organic constituents in recipent media. T A P P I Environmental Conference Proceedings, p.219-231. Suntio, L . R , Shiu, W . Y . and Mackay D . (1988), A review of the nature and properties of chemicals present in pulp m i l l effluents, Chemosphere, 17 (7), 1249-1290. Swan, E . P. (1973), Resin acids and fatty acids of Canadian pulpwood-a review of the literature, Department of Environment, Canadian Forestry Service, Information Report V P - X - 1 1 5 , 21pp, Western Forest Products Laboratory, Vancouver, B . C . Tavendale, M . H . , Wil lk ins A . L . , Langdon A . G . , M a c K i e K . L . , Stuthridge T. S. and McFarlane, P. N . (1995), Analytical methodology for the determination of freely available bleached kraft m i l l effluent-derived organic constituents in reciepent sediments, Environ. Sci. Technol. 29, 1407-1414. Taylor, B . R , Yeager, K . L . , Abernethy, S. G . and Westlake, G . F. (1988), "Scientific Criteria Document for Development of Provincial Water Quality Objectives and Guidelines: Resin Acids" , 50pp, Ontario Ministry of the Environment, Water Resources Branch, Toronto, Ontario. I S B N 0-7729-4347-8, Queen's Printer for Ontario. Teal, J. D . , Clough, J. M . and Mark, V . (1977), Radioimmunoassay of vinblastine and vincristine, Br. J. Clin. Pharmacol., 4, 169-172. Tijssen, P. (1985), "Practice and Theory of Enzyme Immunoassays", Elservier, Amsterdam. Turner, S. A . and Wal l in , B . K . (1985), The analysis of toxic and nonconventional pollutants by F S O T G C / M S . TAPPI, 66 (8), 108-113. 177 Turoski, V . E . , Kuehnl, M . E . and Vine, B . F. (1981), Determination of resin and fatty acids in paper m i l l effluents by G C / M S , TAPPI, 64 (5), 117-121. Unkulvasapaul, M . (1984), Resin acids compounds in a kraft pulp m i l l and paper m i l l effluent and their distribution in bottom sediments near the outfall, Ph. D . dissertation, School of Marine Science, College of W i l l i a m and Mary in Virginia, Virginia. V a n Emon, J. M . and Gerlach, C. L . (1995a), A status report on field-portable immunoassay, Environ. Sci. Technol, 29 (7), 312A-317A. V a n Emon, J. M . and Gerlach, C. L . (1995b), The right environment for the immunoassay, CHEMTECH, November, p.51 -54. V a n Emon, J. M . , Seiber, J. N . and Hammock, B . D . (1987), Application of an enzyme-linked immunosorbent assay to determine paraquat im milk, beef and potatos, Bull. Environ. Contam. Toxicol, 39, 490-497. V a n Emon, J. M . , Seiber, J. N . , Hammock, B . D . (1989), Immunoassay Techniques for Pesticide Analysis, in " Analytical Methods for Pesticides and Plant Growth Regulators", (ed. by J. Sherma), Academic Press, Inc., p.217-263. V a n Emon, J. M . Hammock and B . , Seiber, J. N . (1986), Enzyme-Linked Immunosorbent Assay for paraquat and its application to exposure analysis, Anal. Chem., 58, 1866-1873. V a n Emon, J. M . and Lopez-Avi la , V . (1992), Immunochemical methods for environmental analysis, Anal Chem., 64, 79A-88A. V a n Vuurde, J. W . L . , Maat, D . Z . and Fanken, F. (1988), in "Biotechnology for Crop Protection", (ed. by P. Hedin, J. J. Menn and R. M . Hollingworth), A C S Symposium 379, American Chemical Society, Washington, D C , p.338-350. Vanderlaan, M . , Watkis, B E . and Stanker, L . (1988), Environmental monitoring by immunoassay, Environ. Sci. Technol, 22, 247-254. Vanderlaan, M . , Stanker, L . and Watkis, B . E . (1990), Immunochemical techniques in trace residue analysis, in "Immunoassays for Trace Chemical Analysis, Mornitoring Toxic Chemicals in Humans, Food, and the Environment", A C S Symposium Series 451, American Chemical Society, Washington, D C , p. 1-13. Vanderlaan, M . , V a n Emon, J., Watkis, B . and Stanker, L . (1986), Monoclonal antibodies in the detection of trace chemicals, in "Pesticide Science and Technology" (ed. by R. Greenhalgh and T. R. Robert), Blackwell Sci. Pub., Oxford, p.597-602. 178 Vaughan, J. R. Jr. and Osato, R. L . (1951), Preparation of peptides using mixed carboxylic acid anhydride, J. Amer. Chem. Soc., 73, 5553-5555. Vaughan, J. R. Jr and Osato, R. L . (1952), The preparation of peptides using mixed carbonic-carboxylic acid anhydride, J. Amer. Chem. Soc, 74, 676-678. Victor, P. A . et al. (1981), Environmental Protection Regulation: Water Pollution and the Pulp and Paper Industry, Technical report N o . 14, Economic Council of Canada, M I C 82-00206. Voller , A . (1978), The enzyme-linked immunosorbent assay (ELISA) , Ric. Clin. Lab., 8, 289-298. Voller , A . , B idwel l , D . E . and Bartlett, A . (1979), "The Enzyme Linked Immunosorbent Assay ( E L I S A ) , A guide with abstracts of microtitre plate applications", I S B N 0.906036.01.1 Voss, R. H . (1987), Trace organic contaminants in pulp and paper m i l l effluents and their environmental effects. P A P R I C A N : Miscellaneous Report M R 112, 21pp. Voss, R. H . and Rapsomatiotis, A . (1985), A n improved solvent extraction based procedure for the gas chromatographic analysis of resin and fatty acids in pulp m i l l effluents, J. Chromatog., 346, 205-214. Walden, C. C. and Howard, T. E . (1981), Toxicity of pulp and paper m i l l effluent - a review. Pulp Pap. Can., 82 (4), 115-121 (T143). Wang, C. X . , Yediler, A . , Peng, A . and Kettrup, A . (1995), Photodegradation of phenantherene in the presence of humic substances and hydrogen peroxide, Chemosphere, 30 (3), 501-510. Wearing, J. T., Ouchi, M . D . , Mortimer, R. D . , Kovacs, T. G . and Wong, A . (1984), Factors controlling resin and fatty acid dissolution in sulfite cooking of black spruce. J. Pulp Pap. Sci., 10(6), J178-181. Weiler, E . W . (1990), Immunoassay for the quantitation of low molecular weight compounds in plants, in "Pyrethroid Residues, Immunoassays for L o w Molecular Weight Compounds", (ed. by W . Blab and E . W . Weiler), Springer-Verlag, p. 147-220. Weiler, E . W . , Stockigt, J. and Zenk, M . H . (1981), Radioimmunoassays for the quantitative determination of scopolamine, Phytochemistry (Oxf), 20, 2009-2016. Wie , S. I. and Hammock, B . D . (1982), Development of enzyme-linked immunosorbent assay for resodue analysis of diflurobrnzuron and Bar Sir 8514, J. Agric. Food Chem., 30, 949-957. Wie , S. I. Hammock, B . D . (1984), Comparison of coating and immunizing antigen structure on the sensitivity and specificity for benzoylphenylurea insecticides, J. Agric. Food Chem., 32, 1294-1301. 179 Wilkins , A . L . and Panadam, S. (1987), Extractable organic substances from the discharges of a N e w Zealand pulp and paper m i l l . APPITA, 40 (3), 208-212. Wil lard , H . K . (1983), Toxicity reduction of pulp and paper m i l l wastewater, Contract Rep. 68-03-3028, W A 20. U . S . Environmental Protection Agency. Cincinnati, O H . Winter, G . and Milstein, C. (1989), Man-made antibodies, Nature, 349, 293-299. Wittman, C. and Hock, B . (1989), Improved enzyme immunoassay for the analysis of s-triazine in water samples, Food Agric. Immunol., 1, 211-224. Wittman, C. and Hock, B . (1990), Evaluation and performance characteristics of a novel E L I S A for the quantitative analysis of atrazine in water, plants and soil, Food Agric. Immunol., 2, 65-74. Zeiss, H . H . and Martin, W . B . Jr. (1953), Synthesis and stereochemistry of the 3-keto-A4-steroidal system from diterpenic acids, J. Amer. Chem. Soc, 75, 5935-5940. Zender, J. A . , Stuthridge, T. R., Langdon, A . G . , Wilkins , A . L . , Mackie, K . L . and McFarlane, P. N . (1994), Removal and transformation of resin acids during secondary treatment at a N e w Zealand bleached kraft pulp and paper m i l l , Wat. Sci. Tech., 29, 105-121. Zinkel , D . F. and Engler, C. C. (1977), Gas-liquid chromatography of resin acid esters. J. Chromatog., 136, 245-252. Zinkel , D . F. and Han, J. (1986), G L C determination of the resin acid composition in rosins and oleoresins: state of the art, Naval Stores Rev., 96(2), 14-16. 180 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0087843/manifest

Comment

Related Items