Steroids 69 (2004) 245–253

A novel enzyme-linked immunosorbent assay for ethynylestradiolusing a long-chain biotinylated EE2 derivative

Christian Schneidera, Heinz F. Schölera, Rudolf J. Schneiderb,∗a Institute of Environmental Geochemistry, University of Heidelberg, Im Neuenheimer Feld 236, D-69120 Heidelberg, Germany

b Institute of Plant Nutrition, University of Bonn, Karlrobert-Kreiten-Str. 13, D-53115 Bonn, Germany

Received 7 January 2003; received in revised form 29 December 2003; accepted 26 January 2004

Available online 20 May 2004

Abstract

Ethynylestradiol (EE2) is one of the most potent endocrine disrupting compounds capable to induce estrogenic effects even at tracelevel concentrations in the aquatic environment. Methods for detecting EE2 in such concentrations are generally based on GC or HPLCcoupled to at least one mass spectrometer. Another approach are immunoassays and sensor systems but for most designs, derivatives ofEE2 are required (e.g. for coupling to carrier proteins, enzyme or fluorescent labels, etc.). Here we present the straightforward synthesis andcomplete characterization of a new long-chain biotinylated EE2 derivative. The new EE2 derivative is used as tracer in a direct competitiveenzyme-linked immunosorbent assay (ELISA) for the determination of EE2. With pure water, the limit of detection (LOD, signal-to-noiseratio, S/N= 3) and the test midpoint were found to be 14 and 136 ng l−1, respectively. Cross reactivity (CR) was tested for 10 endogenoussteroids and the BSA-conjugate used for immunization, as well as a synthetic precursor of the conjugate. Among the naturally occuringcompounds, CR was determined to be maximum for metabolites of EE2 conjugated at ring-position 3 (17% and 37% for 3-glucuronideand 3-sulphate, respectively). Assay stability was tested against humic substances and organic solvents. Increasing amounts of organicsolvents in the sample caused a clear decrease in sensitivity, presence of humic substances lead to an overestimation of EE2.© 2004 Elsevier Inc. All rights reserved.

Keywords: Ethynylestradiol; EE2; Biotinylation; Steroid; Biotin; Synthesis; ELISA

1. Introduction

The 1990s have witnessed a great expansion of interestin the implications of environmental exposure of wildlifeand humans to potential endocrine disruptors[1]. Gutendorfand Westendorf[2] identified more than 50 compounds ex-hibiting endocrine disrupting effects, with the majority ofthese compounds being of estrogenic nature. Estradiol andits metabolites estrone and estriol have been detected in Eu-rope [3–6], North America[6] and Asia[7] in municipaland industrial effluents. The pharmaceutical estrogen 17�-ethynylestradiol (EE2) frequently prescribed as an oral con-traceptive has been detected at trace-level concentrations inthe low ng l−1-range in sewage effluents in Italy, Germany,England, The Netherlands and the U.S.[3,4,6,8–13]. In sur-face waters EE2 was only detected in a few samples. Somestudies reported values up to 4.3 ng l−1 [3]. However, mostof the samples had levels below 1 ppt[3,8,12]. Even though

∗ Corresponding author. Tel.:+49-228-732856; fax:+49-228-732489.E-mail address: schneider@uni-bonn.de (R.J. Schneider).

the detected EE2 concentrations are very low there mightstill be an impact on the aquatic ecosystem, as has been indi-cated by laboratory studies using vitellogenin as biomarkerfor endocrine disruption in rainbow trout[14].

For quantifications in the low ng l−1-range analystsrely on highly sophisticated methods such as GC–MS,HPLC–MS or even tandem MS techniques. As an alterna-tive, sensor techniques and immunochemical assays suchas enzyme-linked immunosorbent assay (ELISA) can beused. In the past, several studies have been published us-ing immunoassays for measuring EE2 in different matrices[15–19]. Available test kits for EE2 are optimized for bio-logical specimens such as blood and urine. Applicability toenvironmental samples was neither considered nor assessedfor the majority of these kits. Here we present the facilesynthesis and complete characterization of a new long-chainbiotinylated EE2 derivative (Fig. 1). We have employedan ε-aminocaproic acid spacer connecting the biotin moi-ety to an EE2-6-carboxymethyloxime with the intentionof maximizing the biotin–avidin interaction[20,21] in theapplication of the derivative in a new ELISA. The ELISA

0039-128X/$ – see front matter © 2004 Elsevier Inc. All rights reserved.doi:10.1016/j.steroids.2004.01.003

246 C. Schneider et al. / Steroids 69 (2004) 245–253

HO

HO

NO

O

HNNH

OHN

O

S

HN

NH

O

Fig. 1. Z-form of 3,17�-dihydroxy-1,3,5(10)-17�-ethynylestratrien-6-one-carboxymethyloxime-hydrazido-(ε-biotinyl)-aminocaproamide.

based on the long-chain biotinylated ethynylestradiol isoptimized for environmental analysis and shows excellentsensitivity and selectivitiy.

2. Experimental

2.1. Reagents and materials

All reagents were of analytical or biochemical grade andwere used as received. 1,3,5(10)Estratrien-17�-ethynyl-3,17-diol-6-one-6-carboxymethyloxime and the corre-sponding BSA conjugate were purchased from Steraloids(London, GB), the biotin-LC-hydrazide was obtainedfrom Molecular Bioscience (Boulder, CO, USA), Ace-tonitrile was purchased from LGC Promochem (Wesel,Germany), tetramethylbenzidine (TMB) and TweenTM 20were purchased from Serva (Heidelberg, Germany), humicacid (HA) was from Carl Roth (Karlsruhe, Germany) thestreptavidin-HRP conjugate (S5512) and all other reagentswere supplied by Sigma–Aldrich (Taufkirchen, Germany).Ultrapure water was obtained by running demineralizedwater through a Milli-Q water purification system fromMillipore (Eschborn, Germany). 0.2�m syringe filterswere otained from Gelman (Ann Arbor, MI, USA), TLCplates (AlugramTM) were from Macherey–Nagel (Düren,Germany). The microtiter plates used were 96 flat-bottomwells with high binding capacity (MaxisorpTM) from Nunc(Roskilde, Denmark). For the immunoassay the followingbuffers were used: PBS 1 (10 mM PBS pH 7.6, containing150 mM NaCl and 2.85% TweenTM 20), PBS 2 (8 mM PBSpH 7.6, containing 145 mM NaCl), PBS 3 (8 mM PBS pH7.6, containing 145 mM NaCl and 0.5% TweenTM 20), coat-ing buffer (50 mM sodium carbonate buffer, pH 9.6) andsubstrate buffer (200 mM citric acid monopotassium saltbuffer containing 3 mM perhydrol and 0.01% sorbic acidpotassium salt, pH 3.8). A stabilized TMB-solution (41 mMTMB, 8 mM tetrabutylammoniumborhydrid in dimethylac-etamide) was prepared according to Frey et al.[22]. Thefinal substrate solution consists of 540�l TMB-solution in21.5 ml substrate buffer and was freshly prepared for eachrun.

2.2. HPLC

Preparative high-performance liquid chromatography wascarried out on a Perkin-Elmer system (advanced LC sampleprocesssor ISS 200, binary LC pump 250, LC 235 diodearray detector) using a Prodigy ODS column (preparative;10�m, 250 mm× 21.2 mm; Phenomenex, Aschaffenburg,Germany). Separations were run at 20◦C with a flow rate of7 ml min−1 with the following acetonitril/water (adjusted topH 9 with ammoniumhydroxide) gradient: starting with 25%acetonitril for 5 min, increasing linearly to 40% acetonitrilwithin 8 min, continuing with 60% acetonitril, increasingto 100% within 12 min. UV detection was performed at awavelength of 260 nm.

2.3. NMR spectroscopy

NMR spectra were recorded on a Bruker AM 300 (300and 75 MHz, for1H and 13C, respectively) at room tem-perature.1H and13C NMR chemical shifts are reported inparts per million downfield from tetramethylsilane and refer-enced to residual solvent resonances (1H NMR: 3.31 ppm forCHD2OD in CD3OD; 13C NMR, 49.15 ppm for CD3OD).1H NMR signals are given followed by multiplicity, cou-pling constantsJ in hertz, and integration in parentheses.For complex coupling patterns, the first coupling constantlisted corresponds to the first splitting listed.

2.4. Mass spectrometry

Exact masses were determined by FAB MS analysis run-ning in positive ion mode using 3-nitrobenzylalcohol as ma-trix on a Concept 1H instrument from Kratos.

Synthesis of 3,17�-dihydroxy-1,3,5(10)-17�-ethynyles-tratrien-6-one-carboxymethyloxime-hydrazido-(ε-biotinyl)-aminocaproamide

6.8�l of isobutylchloroformiate was added at−15◦Cto a solution of 20 mg 3,17�-dihydroxy-1,3,5(10)-17�-ethynylestratrien-6-one-carboxymethyloxime and 8.3�l ofN-methylmorpholin in 1.04 ml dry DMF. After stirringfor 3 min the activated steroid was added to a suspensionof 19.4 mg of the biotin-LC-hydrazide in 4.16 ml of dryDMF and stirred overnight at 4◦C. Reaction progress wasmonitored by TLC on precoated silica gel plates (thickness0.25 mm), spots being detected under a UV lamp. Withcompletion of the reaction the cloudy suspension of thebiotin-LC-hydrazide changed to a clear solution. For purifi-cation the solution was directly injected into the preparativeHPLC system. The combined fractions of the product wereevaporated to dryness under reduced pressure. Solid phasematerial stemming from the preparative column precipitatedwithin the product and was removed by membrane filtra-tion using methanol as solvent and a 0.2�m hydrophilicpolypropylene membrane. Again methanol was evaporatedand the product dried under vacuum.

C. Schneider et al. / Steroids 69 (2004) 245–253 247

Yield: 52%. 1H NMR, (CD3OD):� 7.31 (d,J = 2.8 Hz,1H, 4-H), 7.13 (d,J = 8.5 Hz, 1H, 1-H), 6.74 (dd,J = 8.5;2.8 Hz, 1H, 2-H), 4.64 (s, 2H, oxime), 4.43 (pseudo-dd,J = 7.8; 5.1 Hz, 1H, bio), 4.24 (dd,J = 7.9; 4.5 1H,biotin), 3.18 (m, 1H, SC-H), 3.13 (m, 2H, NCH2), 2.87(dd, J = 12.7; 5.1 Hz 1H, SCHa), 2.82 (s, 1H, 17-alkyne),2.65 (d,J = 12.7, 1H, SCHb), 2.3–2.1 (m, 6H), 2.0–1.9(m, 3H), 1.8–1.3 (m, 20 H), 0.81 (s, 3H, 18-methyl);13CNMR, (CD3OD): � 176.1, 173.7, 170.4, 169.9, 166.2,165.0, 157.4, 135.1, 132.1, 126.9, 118.7, 111.7, 80.3, 73.8,63.5, 61.7, 57.1, 51.2, 42.7, 41.1, 40.3, 40.0, 39.3, 37.0,36.9, 35.2, 33.7, 31.8, 30.9, 30.1, 29.8, 29.6, 27.5, 27.0,26.5, 23.8, 22.2, 13.3 ppm; MS (FAB): [M+ Na]+ 759.3;[M] + 737.4; [M-urea]+ 679.4; [M-(methyl, ethynyl, bothhydroxyfunctions]+ 663.5; [M-(methyl, ethynyl, one hy-droxyfunction, urea]+ 621.3; [M-(methyl, ethynyl, bothhydroxyfunctions, urea fragment, steroid ring d]+ 563.3;[ethynylestradiol-6-oxime+ Na]+ 329.1; [ethynylestradiol-6-oxime]+ 307.1; [biotin+ Na]+ 234.0; [biotin]+ 312.0.HPLC (preparative): Rt 11–12 min.

2.5. Bio-LC-EE2 ELISA

2.5.1. Antisera productionTwo New Zealand white rabbits were immunized in co-

operation with Eurogentec (Herstal, Belgium) using a BSAconjugate of 1,3,5(10)estratrien-17�-ethynyl-3,17-diol-6-one-6-carboxymethyloxime. Primary immunization wasperformed intradermally at 10 multiple sites by injectingthe immunogen emulsified with Freund’s complete adju-vant. Booster injections were administered after 2, 4 and 8weeks after priming using Freund’s incomplete adjuvant.Final bleedings were taken after 3 months. The serum fromthe rabbit which showed highest titer and sensitivity in theassay was used in this study.

2.5.2. Optimized ELISA procedureDirect competitive ELISA format was adapted for the

analysis of EE2. ELISA was performed as follows. Mi-crotiter plates were coated with the polyclonal antibody (di-lution 1:50 000, 200�l per well) in coating buffer. The plateswere covered with parafilm to prevent evaporation. Afterovernight incubation at 20◦C, the plates were washed threetimes with PBS 1 using an automatic plate washer (Flexi-wash, Asys Hitech, Eugendorf, Austria). For construction ofthe calibration curve an EE2 stock solution was prepared us-ing ethylacetate and then further diluted with ultrapure wa-ter to obtain standard solutions covering the concentrationrange between 0.001 and 1000�g l−1. After the three-cyclewashing step the plates had EE2 standards added (100�lper well) and were shaken at room temperature for 30 min.This was followed by addition of the synthesized Bio-LC-EE2 in PBS 2 (0.025 pmol ml−1, 100�l per well); the plateswere shaken at room temperature for another 30 min, fol-lowed by a second three-cycle washing step. To the plateswas then added streptavidin-HRP conjugate solution in PBS

3 (50 ng ml−1, 200�l per well) and shaking continued atroom temperature for 30 min, followed by a third three-cyclewashing step. Finally, substrate solution was added (200�lper well) and incubated for 30 min. The enzyme reactionwas stopped by addition of sulfuric acid (50�l per well,1 M). Absorbance was measured at 450 nm and referencedto 650 nm with a plate reader (VMax, Molecular Devices,Munich, Germany). All determinations were made at leastin triplicate on different plates. The mean values were fit-ted to a four-parametric logistic equation (4PL)[23], opticaldensities were normalized to parameter A of the 4PL andplotted against log C.

2.5.3. Cross reactivity (CR) determinationThe relative sensitivity of the immunoassay towards the

compounds listed inTable 1was determined by assaying adilution series of each substance in water. Cross reactivityis calculated as the ratio of molar concentrations at the in-flection points (midpoints) of the corresponding calibrationcurves and expressed in percentage relative to the midpointfor EE2.

2.5.4. Effects of organic solvents and humic acid on theELISA

Matrix effects of organic solvents and a model humic acidas interfering agents were studied using serial dilutions ofEE2 in solutions of the respective substances. Three organicsolvents (acetone, acetonitrile, methanol) were tested fortheir compatibility with the ELISA; concentrations coveredthe range from 5 to 40% solvent in EE2 standard. Commer-cial humic acid was tested in concentrations ranging from0.2 to 10 mg l−1.

3. Results and discussion

3.1. Synthesis of 3,17β-dihydroxy-1,3,5(10)-17α-ethynylestratrien-6-one-carboxymethyloxime-hydrazido-(ε-biotinyl)-aminocaproamide

We present a facile synthesis and the complete charac-terization for a new biotinylated ethynylestradiol derivative.The synthesis relies on activation of the steroid for 3 minat −15◦C using isobutylchloroformiate as leaving group inDMF as solvent[24]. Modifications of this procedure pub-lished for the biotinylation of estradiol[25] applied a harshermethod for activation substituting DMF with dioxane andstirring for 30 min at 8◦C. However, promising results inthe biotinylation of estradiol using the more gentle activa-tion at−15◦C (data not shown) led us to the procedure forethynylestradiol presented here enabling a very smooth ki-netic control in the activation of the steroid.

As a thorough purification of the product is an indispens-able prerequisite for the application of this new ethynylestra-diol derivative in immunochemistry, we have developed anew preparative HPLC method for its purification. Problems

248 C. Schneider et al. / Steroids 69 (2004) 245–253

Table 1Cross reactivity of selected estrogens with ethynylestradiol antiserum

Compound Cross reactivity (%)a Standard deviation Structure

Ethynylestradiol (EE2)b 100 –

17�-Estradiol (E2)b 0.2 0.03

Estronec <0.1 –

Estriolc <0.1 –

E2-3-glucuronide sodium saltb 0.4 0.05

E2-3-sulphate sodium saltb <0.1 –

E2-17-sulphate sodium saltc <0.1 –

E2-17-glucuronide sodium saltb <0.1 –

C. Schneider et al. / Steroids 69 (2004) 245–253 249

Table 1 (Continued )

Compound Cross reactivity (%)a Standard deviation Structure

EE2-3-glucuronide sodium saltc 17 0.11

EE2-3-sulphate sodium saltc 37 4.3

EE2-17-glucuronide sodium saltc <0.1 –

EE2-6-carboxy-methyloximec 99 6.7

EE2-BSA (immunogen)c 435,000 69,000

All values are determined in triplicate on different microtiter plates.a Cross reactivity is calculated as a ratio of molar concentrations at the inflection points of the corresponding calibration curves and expressed in

percentage based on 100% response of EE2.b Compounds were supplied by Sigma.c Compounds were supplied by Steraloids.

arose as solid phase material precipitated in the productwhen evaporating the combined fractions to dryness. Thisdrawback was overcome by using a simple membrane fil-tration to remove the solid phase. Purity of the isolatedproduct could be confirmed by NMR. The biotinylatedethynylestradiol was obtained in 52% yield due to thementioned difficulties in the process of purification.

3.2. Characteristics of the Bio-LC-EE2 ELISA

3.2.1. Optimization of assay conditionsUnlike noncompetitive ELISAs, in which an excess of

reagents is used to increase detectability, limiting concen-

trations of immunoreagents are required in competitiveELISAs to obtain good sensitivities. Criteria used to eval-uate the optimization were maximum optical density (B0),dynamic range, concentration at midpoint (parameterCof the 4PL curve-fitting) and detection limit (LOD). Thebest combination of immunoreagents used was found tobe a combination of diluting the antiserum 1:50 000 andemploy a concentration of the synthesized Bio-LC-EE2 of0.025 pmol ml−1. Further dilution of the Bio-LC-EE2 didnot yield an improved sensitivity of the assay. This indicatesthat the affinity limit of the antibodies in the antiserumwas reached. Dilution of streptavidin-peroxidase conjugatewas chosen according to Bodmer et al.[26], TweenTM

250 C. Schneider et al. / Steroids 69 (2004) 245–253

Fig. 2. Master-calibration curve for ethynylestradiol obtained with spiked ultrapure water. Results of the four-parametric logistic curve fitting: (A) 100%;(B) 0.6�g l−1; (C) 0.136�g l−1; (D) 23%. Values averaged from 13 individual calibration curves; error bars calculated according to Gaussian errorpropagation based on standard deviation at each concentration level in the 13 individual calibration curves.

20 had to be added to reduce nonspecific binding of theconjugate.

Based on the optimized conditions, the EE2 calibra-tion curve was constructed in the concentration range of0.001–1000�g l−1. A mastercurve calculated from 13 cal-ibration curves consecutively performed during 3 weeksis shown inFig. 2. The midpoints varied in the range of111–177 ng l−1 (interassay CV 25%), the relative standarddeviation of the measured optical density for three repli-cates at each standard concentration was lower than 5%.The LOD at a signal-to-noise ratio (S/N) of 3 is found to be14 ng l−1. A suitable analytical working range of the assayis 0.022–1.2�g l−1 (part of the curve between 40 and 80%inhibition).

3.2.2. Cross reactivitiesEstradiol, several metabolites of estradiol and ethynylestra-

diol as well as the immunogen and a synthetic precursorof the immunogen and the biotinylated ethynylestradiol(EE2-6-carboxymethyloxime) were tested for cross re-activity (Table 1). The immunoassay proved to be veryspecific for ethynylestradiol as only negligible reactivitieswere observed for most of the endogenous compounds(CR < 0.5%). Substantial cross reactivities existed onlyfor EE2 metabolites conjugated at ring position 3; EE2-3-sulphate sodium salt and EE2-3-glucuronide sodium saltexhibiting cross reactivities of 37 and 17%, respectively.Binding of the EE2-6-carboxymethyloxime equals bindingof the ethynylestradiol, this indicates only the steroid corebeing recognized by the antibodies and the oxime-linkerbeing discriminated. As expected the immunogen itself ex-hibits a very high affinity, at least 4000 times higher than

ethynylestradiol. This can be explained by partial recogni-tion of the protein moiety in the immunogen.

These findings clearly demonstrate that the problem ofhomologous bridge recognition is not observed in this assay.Pronounced binding of the immunogen and lower selectiv-ities of the antibodies towards ethynylestradiol metabolitesconjugated at position 3 indicate a deficient recognitionwhich can be explained by the steric requirements of theethynylestradiol–BSA conjugate employed during the im-munization. Proximity of the carrier-protein, bound viathe oxime-linker at position 6, to the hydroxy group atposition 3 impedes a good discrimination of sulphate andglucuronide conjugated to position 3 of ethynylestradiol.In contrast, excellent selectivities are obtained for the moredistant hydroxy group at position 17. These findings aresupported by results of Agasan et al.[16], who reportedsimilar cross reactivities towards the 3- and 17-sulphate ofEE2 using the same type of conjugate for immunization.

3.2.3. Stability testsProteins, such as antibodies, are relatively susceptible

to matrix effects and harsh conditions[27]. Variation ofthe matrix in environmental analysis may be considerableand investigations of the matrix interferences should be anintegral part of assay development. Compatibility of theimmunoassay with organic solvents is important as envi-ronmental samples might need to be extracted and enrichedprior to quantification.

A common effect was observed both for humic acidand organic solvents, i.e. with increasing concentration inthe sample the standard curve shifted gradually to higherEE2 concentrations and, therefore, a loss in sensitivity was

C. Schneider et al. / Steroids 69 (2004) 245–253 251

Fig. 3. Effect of a model humic acid (mHA) on the calibration curve for ethynylestradiol obtained with spiked water samples: (�) ultrapure water, (�)0.2 mg l−1, (�) 5 mg l−1, (�) 10 mg l−1. Error bars: standard deviation (except ultrapure water: Gaussian error propagation based on master-calibrationcurve).

Fig. 4. Effect of organic solvents on the calibration curve for ethynylestradiol obtained with spiked water samples: (�) methanol, (�) acetone, (�)acetonitrile. Error bars: standard deviation.

252 C. Schneider et al. / Steroids 69 (2004) 245–253

observed. Interference with humic acid lead to flattenedcalibration curves of the immunoassay causing an overes-timation of EE2 in dependance of the HA concentration(Fig. 3). With increasing amounts of HA in the samplethe sigmoidal shape of the calibration curve was steadilylost. The intrinsic mechanism of this interference is un-known, we assume that it may be due to the unspecificbinding of EE2 to HA. Increasing amounts of organic sol-vents in the sample lead to a clear decrease in sensitivity(Fig. 4). Furthermore absorbance values fitted poorly to thefour-parametric logistic equation when increasing amountsof solvent were added to the sample. The extent of de-cline in sensitivity depends on the solvent used; similarto humic acids, organic solvents lead to an overestima-tion of EE2. These results are in good agreement withdata published by Knopp and co-workers[28] on matrixinterference.

Our goal in this study was to develop a straightforwardsynthesis for Bio-LC-EE2 and set up an ELISA based onpolyclonal antibodies. We did not immerse ourselves indiscriminating isomeric E- and Z-forms of the biotinylatedethynylestradiol derivative as a distinction of the isomersplayed only a role in immunoassays using monoclonal an-tibodies directed against only one of the two isomers. Inthe present work we have presented the facile synthesisof a biotinylated ethynynlestradiol and developed an op-timized ELISA for ethynynlestradiol. A further increasein assay sensitivity would be expected if solid-phase ex-traction for enrichment of samples prior to measurementis employed, which is currently under investigation. Thisimmunoassay could provide an alternative approach forquantifying ethynynlestradiol in environmental samples,such as surface waters and effluents from sewage treatmentplants.

Acknowledgements

C. Schneider thanks Mr. Thorsten Christian for help insetting up the preparative HPLC, and the German FederalEnvironmental Foundation (DBU) for financial support.

References

[1] Stone R. Environmental estrogens stir debate. Science 1994;265:308–10.

[2] Gutendorf B, Westendorf J. Comparison of an array of in vitroassays for the assessment of the estrogenic potential of natural andsynthetic estrogens. Toxicology 2001;166:79–89.

[3] Belfroid AC, Van der Horst A, Vethaak AD, Schafer AJ, Rijs GB,Wegener J, et al. Analysis and occurrence of estrogenic hormonesand their glucuronides in surface water and waste water in TheNetherlands. Sci Total Environ 1999;225:101–8.

[4] Desbrow C, Routledge EJ, Brighty GC, Sumpter JP, Waldock M.Identification of estrogenic chemicals in STW effluent. 1. Chemicalfractionation and in vitro biological screening. Environ Sci Technol1998;32:1549–58.

[5] Solé M, Lopez de Alda MJ, Castillo M, Porte C, Ladegaard-PedersenK, Barceló D. Estrogenicity determination in sewage treatment plantsand surface waters from the Catalonian area (NE Spain). EnvironSci Technol 2000;34:5076–83.

[6] Ternes TA, Stumpf M, Mueller J, Haberer K, Wilken RD, Servos M.Behavior and occurrence of estrogens in municipal sewage treatmentplants. Investigations in Germany, Canada and Brazil. Sci TotalEnviron 1999;225:81–90.

[7] Oh SM, Choung SY, Sheen YY, Chung KH. Quantitative assessmentof estrogenic activity in the water environment of Korea by the E-screen assay. Sci Total Environ 2000;263:161–9.

[8] Huang CH, Sedlak DL. Analysis of estrogenic hormones inmunicipal wastewater effluent and surface water using enzyme-linked immunosorbent assay and gas chromatography/tandem massspectrometry. Environ Toxicol Chem 2001;20:133–9.

[9] Kuch HM, Ballschmiter K. Determination of endogenous andexogenous estrogens in effluents from sewage treatment plants at theng/l-level. Fresenius J Anal Chem 2000;366:392–5.

[10] Routledge EJ, Sheahan D, Desbrow C, Brighty GC, Waldock M,Sumpter JP. Identification of estrogenic chemicals in STW effluent.2. In vivo responses in trout and roach. Environ Sci Technol1998;32:1559–65.

[11] Johnson AC, Belfroid A, DiCorcia AD. Estimating steroid oestrogeninputs into activated sludge treatment works and observations ontheir removal from the effluent. Sci Total Environ 2000;256:163–73.

[12] Xiao XY, McCalley DV, McEvoy J. Analysis of estrogens inriver water and effluents using solid-phase extraction and gaschromatography-negative chemical ionisation mass spectrometry ofthe pentafluorobenzoyl derivatives. J Chromatogr A 2001;923:195–204.

[13] Baronti C, Curini R, d’Ascenzo G, DiCorcia A, Gentili A, SamperiR. Monitoring natural and synthetic estrogens at activated sludgesewage treatment plants and in receiving river water. Environ SciTechnol 2000;34:5059–66.

[14] Sumpter JP, Purdom CE. Estrogenic effects of effluents from sewagetreatment works. Chem Ecol 1994;8:275–85.

[15] Agasan AL, Stewart BJ, Watson TG. Development of aradioimmunoassay method for ethynylestradiol in plasma using amonoclonal antibody. J Immunol Methods 1994;177:251–60.

[16] Agasan AL, Briggs MH, Hewish DR, Stewart B. Developmentof a monoclonal antibody for use in radioimmunoassay ofethynylestradiol. Steroids 1986;47:295–306.

[17] Fotherby K, Akpoviroro JO, Siekmann L, Breuer H. Measurementof ethynyloestradiol by radioimmunoassay and by isotope dilution-mass spectrometry. J Steroid Biochem 1981;14:499–500.

[18] de la Pena A, Chenault CB, Goldzieher JW. Radioimmunoassay ofunconjugated plasma ethynylestradiol in women given a single oraldose of ethynylestradiol or mestranol. Steroids 1975;25:773–80.

[19] Tacey RL, Harman WJ, Kelly LL. Development of a highly sensitiveand specific assay for plasma ethinylestradiol using combinedextraction, liquid chromatography and radioimmunoassay. J PharmBiomed Anal 1994;12:1303–10.

[20] Costello SM, Felix RT, Giese RW. Enhancement of immunecellular agglutination by use of an avidin–biotin system. Clin Chem1979;25:1572–80.

[21] Leary JJ, Brigati DJ, Ward DC. Rapid and sensitive colorimetricmethod for visualizing biotin-labeled DNA probes hybridized toDNA or RNA immobilized on nitrocellulose: Bio-blots. Proc NatlAcad Sci USA 1983;80:4045–9.

[22] Frey A, Meckelein B, Externest D, Schmidt MA. A stableand highly sensitive 3,3′,5,5′-tetramethylbenzidine-based substratereagent for enzyme-linked immunosorbent assays. J ImmunolMethods 2000;233:47–56.

[23] Dudley RA, Edwards P, Ekins RP, Finney DJ, McKenzie IGM, RaabGM, et al. Guidelines for immunoassay data processing. Clin Chem1985;31:1264–71.

C. Schneider et al. / Steroids 69 (2004) 245–253 253

[24] Munro C, Stabenfeldt G. Development of a microtitre plate enzymeimmunoassay for the determination of progesterone. J Endocrinol1984;101:41–9.

[25] Tiefenauer LX, Andres RY. Biotinyl-estradiol derivatives in enzymeimmunoassays: structural requirements for optimal antibody binding.J Steroid Biochem 1990;35:633–9.

[26] Bodmer DM, Tiefenauer LX, Andres RY. Antigen- versus antibody-immobilized ELISA procedures based on a biotinyl-estradiolconjugate. J Steroid Biochem 1998;33:1161–6.

[27] Pfortner P, Weller MG, Niessner R. Immunological method for thedetection of nitroaromatic residues covalently bound to humic acids.Fresenius J Anal Chem 1998;360:192–8.

[28] Deng A, Himmelsbach M, Zhu Q-Z, Frey S, Sengl M, BuchbergerW, et al. Residue analysis of the pharmaceutical diclofenac indifferent water types using ELISA and GC-MS. Environ Sci Technol2003;37:3422–9.