-.- Page 1 of 1

From: OAO <elucas@goliat.espol.edu.ec> To: <robbyviIlanueva@hotmail.com> Sent: Tuesday, October 17,2000 4:lO PM Subject: WWW ILL Request

elucas@goliat.espol.edu.ec (OAO) sent the following ILL request:

Author: J. P. Nash, B. D. Cuisset, S. Bhattacharyya, H. C. Suter, F. LeMenn and D. E. Kime, Book Title: Fish Physiology and Biochemistry Journal Article: An enzyme linked immunosorbant assay (ELISA) for testosterone, estradiol, and 1 7,20 beta-dihydroxy-4-pregenen-3-one using acetylcholinesterase as tracer: application to measurement of die1 patterns in rainbow trout (Oncorhynchus mykiss). Year: 2000 Volume: 22 Pages: 355-363 Location: Rush: on Needed By: User Name: OAO E-Mail address: elucas@oliat.eqol.edu.ec Phone #: 17-1 0-00

Server Protocol: HTTP/1 .l Remote Host: (200.10.149.1 1)

1 1/5/00

Fish Phyiology and Biochemisrn 22: 355-363. 2000. 0 2OOO Kluu.er .4cademir Publishers. Prinred in rhe Nerherlands. 35s

An enzyme linked immunosorbant assay (ELISA) for testosterone, estradiol, and 17,20/l-dihydroxy-4-pregenen-3-one using acetylcholinesterase as tracer: application to measurement of die1 patterns in rainbow trout (Oncorhynchus mykiss)

J. P. Nash’. B. Davail-Cuisset’, S. Bhattacharyya]. H. C. Suter’, F. Le Menn’ and D. E. Kirne’.* ’Department of Animal and Plant Sciences. The Unirversih of Shefield, Shefield S I O 2TN: UK: ‘INRA, Biologie de la Reproduction des Poissons, UniversitP de Bordeaux I , Av. dG’ Facultes, 33405 Talence, France; *Author for correspondence (Phone: f 4 4 114 222 4316; Fax: +44 114 222 0002; E-mail: d.kime@shefJield.ac.uk)

Accepted: March 27. 2000

Key Hiords: ELISA. fish, immunoassay. steroids

Abstract

A simple and rapid Enzyme Linked ImrnunoSorbant Assay (ELISA) is described and validated for testosterone. estradiol. and I 7.20p-dihydroxy-4-pregnen-3-one (1 7,20#?P). A general procedure for preparation of the acetyl- cholinesterase labeled steroid is described which is applicable to any steroid. Use of acetylcholinesterase tracer increased the sensitivity of assay so that reliable measurements of each steroid could be achieved with only 10 pl of plasma. The ELISA was applied to measurement of all three steroids every hour for over 24 hours in a female trout using cannulation of the dorsal aorta. This h g h sampling frequency revealed several short-term ( t 2 h) episodic pulses of testosterone and estradiol.

Introduction

Although radioimmunoassay has been the standard method for measurement of steroids and other hor- mones in plasma for over 20 years. it has a number of serious disadvantages that restrict its applicabil- ity: ( I ) Use and disposal of the radiolabeled tracer is an increasing problem in many countries where it is restricted to sites and personnel approved by na- tional regulator) authorities. ( 3 ) it requires expensive and sophisticated equipment for measurement. ( 3 ) ra- dioactive counting of large numbers of samples is time-consuming and requires large amounts of scin- tillation fluid. and ( 3 ) the availability of radiolabeled steroids from commercial sources is very limited and severely restricts the range of steroids that can be measured. especially in non-mammalian vertebrates in which the reproductive hormones differ from those in mammals.

During the last I5 years there have been a number of publications in which Enzyme Linked ImmunoSor-

bent Assay (ELISA) using microtiter plates have been described as an alternative to radioimmunoassay for the mammalian steroids testosterone and estradiol (El- der and Lewis 1985: Marcus and Durnford 1986. 1988: El Jabri 1991: Silvan et al. 1993a. b). Although the ovaries of female fish produce testosterone and estradiol. 1 1 -ketotestosterone is the major androgen in males, while the progestogens. 17.20B-dihydrox> -3- pregnen-3-one ( 17.30pP) and 17.20/3.2 1 -trihydroxj - 3-pregnen-3-one ( 2 O B S ) have been implicated in the induction of final gamete maturation in both sexes (Kime. 1993). Neither antisera. nor the tracers for these steroids are commercially available.

In addition to interest in measurement of steroids in fish plasma as an aid to aquaculture. there is in- creasing interest in such measurements to monitor the effects of endocrine disrupting chemicals on their re- productive function. A recent report (Tattersfield et al. 1997) has highlighted the need for a simple. rapid and reliable method for the measurement of hormones spe- cific to fish and other non-mammalian vertebrates as

356

the basis for regulatory testing of the effects of poten- tial endocrine disrupters on wildlife. Many of the fish commonly used for such testing (zebrafish. Danio re- rio: fathead minnows. Pimephalespromelas) are small and assay of steroids in plasma, which is restricted to a few microliters. requires very high sensitivity. Cuisset et al. ( 1994) have described a simple and very sensitive ELISA method for the assay of I 1 -ketotestosterone us- ing acetylcholinesterase as tracer. but the applicability was limited by use of electric eel as a source of acetyl- cholinesterase. In this communication we extend the applicability of the methodology to two common ver- tebrate steroids. estradiol and testosterone. and to the specific tish progestogen. 17.2O#lP and describe a gen- eral preparation procedure for the enzyme label which is applicable to any steroid. As an illustration of an ap- plication of the method. the serial sampling of steroid hormones by cannulation of a single fish is described since short-term changes in hormone levels could be missed in studies using the lower sampling frequen- cies necessitated by a less sensitive assay. The method was applied to the determination of die1 patterns of testosterone, estradiol and 17,20#lP in an individual rainbow trout using cannulation of the dorsal aorta. The highly sensitive assay enabled measurement of all three steroids from samples taken every hour for over 24 hours.

Materials and methods

Chemicals and equipmenr

Acetylthiochoiine. 5,5'-dithio-bis-2-nitrobenzoic acid (DTNB). acetylcholinesterase (Sigma C2888. 1OOO- 20oO units mg-I), trypsin and all steroids were ob- tained from Sigma (Poole, UK). The antiserum to testosterone was raised in rabbits, using testosterone- 3-CMO-BSA conjugate (Sigma), for which we grate- fully acknowledge the assistance of Dr P.M. In- gleton. (Institute of Cancer Studies, Hdianphire Hospital. Sheffield). Antisera to estradiol (raised again5t estradiol-6-CMO-BSA in rabbit ). and I7.20bP (against 3-CMO-BSA in sheep) were kindly donated by Dr Z. Yaron (Tel Aviv) and Dr A. P. Scott (Lowest- oft). Cross reactivities of these antisera are described in Bogomolnaya and Yaron (1984). and Scott et al. ( 1982).

Ninety-six well polystyrene high-binding mi- crotiter plates (Costar Catalogue No. 3590) were used for the ELISA. Plates were sealed for storage

with Anachem Sealplate sealing film. Plates were processed using a Chelsea Instruments shaker, an Anthos Model AW-1 plate washer and read with an Anthos Model HTII Plate Reader. 96-well low binding microtiter plates (Costar Cat. No. 2501 were used for sample dilution and preparation of standard curves.

Estradiol-6-CMO and testosterone-3-CMO were pur- chased from Sigma. The preparation of the 3-CMO of 17.1OBP was based on the method of Simp- son and Wright (1977). 17.2O#lP ( 1 . 1 mg) and carboxymethoxylamine hemihydrochloride ( 1 mg: Aldrich) were incubated with sodium acetate ( 3 m p ) in I 0 0 pl methanol at 4' overnight. Two hundred and fifty microliters of deionised water. acidified to pH 2 with glacial acetic acid. was added. mixed well and left for 10 min to allow the steroid-CMO to precipitate out. The oxime was extracted with 3 ml ethyl acetate. and the organic layer evaporated to give 0.5-1 rnp of the steroid-CMO. Completion of reaction was con- tirmed by the presence of a single uv absorbing area at the origin of thin-layer chromatography plates (Merck keselgel GF254) developed in chloroform-methanol ( 9 5 5 ) .

Generation of G4-acetylcholinesterase

Acetylcholinesterase (AChE. 1 mg) in 500 p1 0.1 M phosphate buffer pH 7.0 was treated overnight with 25 p1 trypsin solution (25 pg ml-' in 0.1 M phosphate buffer, pH 7.0) at room temperature. The solution was dialysed against 100 mlO.1 M borate buffer pH 8.5 us- ing an 80 mm length of dialysis tubing (Sigma D2272) for 24 h using several changes of borate buffer. The contents of the dialysis tubing were removed and the tubing washed with clean borate buffer to give a final volume of G4-AChE solution of I ml which could be stored at 4" for at least two weeks.

Preparation of steroid-CMO-AChE conjicgute

Four hundred nmol (174 p g ) of the steroid-CMO in 38 pl of freshly made N-hydroxysuccinimide solu- tion ( 1 mg ml-' in anhydrous dimethylformamide) was treated overnight with 32 pl freshly made N.N'- dicyclohexyIcarbodiimide solution ( 2 mg ml- ' in an- hydrous dimethylformamide) in the dark (Cuisset et al. 1994). It is essential that all traces of moisture are excluded from this step. Thirty microliters of this so- lution was reacted for 2 h in the dark with 200 pl of

357

the 1 mg ml-' G4-AChE stock and then mixed with 1 ml of steroid assaq buffer (0.1 M KzHPO? '3H:O. 0.1 M KHzPO4. 0.15 M NaCl. 1 mM EDTA, 0.1% bovine serum albumin (BSA). 0.15 mM sodium azide. pH 7.1 1. The product was chromatographed on a 900 x IS mm column of BioGel A 15-m (BioRad) with elu- tion buffer (0.01 M Tris. 0.01 M MgC12. 1 M NaCI. 0.15 mM NaN1 pH 7.4) at a flow rate of 300 p1 min-' . The eluate was discarded for 5 h. then eighty 1.5 mi fractions collected into tubes containing 1 ml steroid assay buffer. Five microliters of every 5th fraction was treated with 200 1.11 Ellmans reagent (215 mg 5.5'- dithiobis(2-nitrobenzoic acid) (DTNB) and 200 mg acetylthiocholine in 20 ml 1 M potassium phosphate buffered saline (PPBS) kept as 400 pl frozen stock aliquots for up to six weeks and diluted 150 for use) and absorption read at 405 nm after 2 h. In the region of peak activity. 5 p1 from all fractions was tested with Ellmans reagent and an activity curve plotted. Frac- tions showing > 30% of the maximum Ellmar? activity were combined and stored in stock aliquots at -20". The column was flushed overnight with elution buffer, stored at 4" and flushed again overnight before re-use.

More recently we have also successfully used dis- posable Pharmacia PDlO columns, using the same elution buffer with which acetylcholinesterase and the labelled steroid are eluted in the first twelve 0.2 ml fractions after the bed volume.

Steroid e.xtraction

Steroids were extracted from SO yl rainbow trout plasma with 5 ml dichloromethane. the solvent evap- orated and the residue redissolved in 500 p1 ethanol. The alcohol solution was transferred to an uncoated lo^ -binding microtiter plate. the solvent allowed to evaporate and 250 yl of steroid assay buffer added. Twent\,-five microliters of this ( 5 p1 plasma equiva- lent) gave a satisfactory point on the standard curve for samples with steroid levels in the range of 0.5- 10 ng ml-' plasma. Smaller volumes of buffer were added or serial dilutions made on the plate if lower or higher steroid levels were expected. If neces sq the plasma volume was increased to 250 pl. If sam- ples were not to be assayed immediately. the plate was sealed and refrigerated in a humid container ( 1-2 days) or frozen ( > 2 days) until required.

Plate coating

ELISA plates were coated with either anti-rabbit or anti-sheep according to the species in which the anti-

bodies to the steroid had been raised. Eight microliters of polyclonal goat anti-rabbit IgG (affinity purified. Sigma R3004: 1 mg ml-I) or 16 p1 polyclonal don- key anti-sheep IgG (affinity purified. Sigma S2763: 1 mg m1-l) was dissolved in 15 ml 0.05 M potasium phosphate buffer pH 7.4. 150 pl added to each well of the plate and incubated overnight at room temper- ature in a humid container. The plate was blocked by addition of a further 100 pl of steroid blocking buffer (0.1 M PPBS pH 7.4 containing 3% bovine serum al- bumin (BSA). l mM EDTA and 25 mM sodium azide) and incubation overnight at room temperature. Plates were sealed and stored at 4" in a humid container. Plates were washed 3 times with wash buffer ( 100 ml PPBS and 2.5 ml Tween-20 in 5 1 water) immediately before use.

A.

Checkerboard titration

A checkerboard titration was carried out for each anti- serum and after each preparation of new AChE label. Serial dilutions of the steroid antiserum (from 1 : 1000 dilution; plate columns 1 - 10) were made against serial dilutions of AChE label (from 1:4; plate rows A-H) in a microtiter plate. Plates were processed as in the Standard Assay Procedure and plots made of AChE dilution against absorption for each antiserum dilution to determine optimum dilutions of the two components (see Diamandis and Christopoulos 1996; Price and Newman 1997 ). For the tracers and antisera described here. these were: 1 : 1 SO. 1 :20. 120 for AChE tracer and 1:3,00000. 1:4OOO. and 1:30000 for antisera to estradiol. testosterone. and I 7.2OBP. respectively.

Standard assa!~ procedure

Serial dilutions of standard steroid (400 to 0.78 pg) in 100 1-11 assay buffer were made in the first 10 wells of rows A and B of the coated plate. The remaining 2 wells in these rows were reserved for BO t O pg) and non-specific bound (NSB ). Twenty-five micro- liters of the extracted steroids from the plasma samples ( 5 pl plasma equivalent) was pipetted into the remain- ing 6 rows in duplicate. Twenty-five microliters of the diluted steroid label was added to all wells. and 25 pl antiserum to all wells except for the NSB at the dilutions determined by the checkerboard titration. All wells were made up to IS0 pl in steroid assay buffer, the plates sealed and incubated 2 h at room temperature in a humid chamber. Plates were then washed three times with wash buffer. 200 1.11 per well of Ellmans reagent was added. the plates incubated

358

overnight in the dark at room temperature and read at 405 nm. Picograms per well were calculated for the samples from the standard curve using Stingray software (Dazdaq. Ringmer, UK).

Assay IU 1 idu tion

A pool of female rainbow trout ( 0ncorhynchii.s mykiss) plasma (500 p1) was treated with activated charcoal as described by Cuisset et al. ( 1994) and extracted twice with 5 ml dichloromethane. The com- bined extracts were evaporated and reconstituted in 3 ml assay buffer. One hundred microliter aliquots were pipetted in duplicate into wells on two rows a microtiter plate. Steroid was added to one pair of wells to give a concentration of 300 pg/well. serial dilutions ( x 2 ) performed and the samples assayed according to standard procedure.

Meusiirement of diei hormone prc$lesfrorn (in indii.idiiul femule ririnbovc. trout using ciinnulation of the dorsal uortu

The three year old female rainbow trout (2.61 kg) used for cannulation was held at the University of Gottingen's fish farm at Rellihausen, Germany. Mean water temperature at the time of cannulation was 4.5". The tank was exposed to natural daylight and thus ambient photoperiod for latitude 5 1'59' N and longi- tude 09'91' E. This particular genetic strain of trout normally spawns midwinter (January) and this indi- vidual had ovulated but not spawned at the time of cannulation (i.e., eggs were free running).

Blood samples were taken using cannulation of the dorsal aorta. based on the procedures of Axels- son and Fritsche (1994), Houston ( 1990) and Zohar (1980). The cannula was inserted on the 2 February 1994 at 1 1 : 15 and after an initial sample to check cannula function at 1225 the fish was left to recover for a 3.5 h without disturbance or blood removal. At 16:OO 300 pl of blood was removed via the cjmnula and then every hour for the following 24 hours. Night- time sampling was done using a head torch with 5 W tungsten bulb covered with a red acetate filter. At no time whilst removing blood did the fish show signs of stress or behavioural changes that might indicate it was aware of any procedure being performed. The sam- ples were centrifuged at 10 OOO x g for 6 minutes and plasma immediately frozen at -20 "C . For analysis of plasma steroid levels. 50 p l of plasma was extracted as described above.

Results

Cross- reactitin/ of anti-testosterone

The antiserum to testosterone cross-reacted with 5a- dihydrotestosterone (46%). 5p-dihvdrotestosterone ( 19% 1, 5a-androstane-3a. 17p-diol (3.7% 1. I 1 p- hydroxytestosterone (3.3% ). 5a-androstane-i,3,17p- diol (2.7%). 5p-androstane-3a. 17p-diol (2.5% ). 1 1 - ketotestosterone (0.859 ). estradiol (0.54% ). 4- androstenedione (0.379 ). 4-androstenetrione (0.3 1% and 17.20pP (0.18% ) at the 50% displacement level.

Stundurd curves and assay sensiti1,ih

Typical standard curves for all three steroids showed a workable range from 0.8 to 300 pg/well (Figure I ) . The sensitivities of the assays. taken as the concen- tration of steroid that induced 90% of BBo. were: testosterone. 0.6 pg: estradiol. 1.6 pg: I7.20PP. I .5 pg.

inter-assay variations measured on female trout plasma samples that gave approximately 50"c dis- placement were: testosterone, 7.2%: estradiol. 8.99: 17,20PP, 9.2% for 8 replicates.

Intra-assay variations were determined from stan- dard curves in which 8 replicates of each concentration were used (Figure 2). The co-efficient of variation was less than 10% over the central part of the standard curve between 3 and 200 pg/well.

Assay accuracy

There was a strong positive correlation ( r > 0.999) between the amount of steroid added to plasma and the amount found (Figure 3).

Application to measuring diel projles of three concurrent steroids in an individual fish

Estradiol, testosterone and 17.20,9P were measured in duplicate by ELISA using 10 pl of plasma equiva- lent for each steroid. '411 three steroids showed a large mount of variation over the 29 hours of sampling (Figure 3 ) . Apart from two peaks during early sam- pling. 17.20PP showed a fairly constant basal level. Testosterone and estradiol. however. showed several short-term ( < 2 h ) episodic pulses at various times throughout the sampling period. During these pulses the steroid concentration increased from basal levels of approximately 3 ng ml-' to peak levels of up to 18 ng ml-'. The peaks showed no synchrony either

359

100 1

I h

100

80

60

40

20

o i . 0

\

100

80

60

40

__I 0

4 '?\

'0 .. 0 1 1 10 100 1000 01 1 10 100' 1000 0 1 1 10 100 1000

Testosterone (pglwell) Estradiol (pglwell) 1720pP (pghvell) Figure 1 Typical standard curie\ for ELISA of testosterone. estradiol. and 17.2OBP Values are means of duplicate assays

with scotoperiod or between steroids. A similar pat- tern of pulsatile hormone levels was shown in several other individuals and there was no synchrony between profiles of different fish.

Discussion

Assay validation

We have followed closely the procedure described by Cuisset et al. (1994) for validation of the 11- ketotestosterone ELISA using the same enzyme label.

Sensitiiin

Sensitivity of the assays. which has been taken as 90% of the binding with 0 pg. is less than 1 pg for testos- terone and just over 1 pg for the other two steroids. Using identical conditions for label preparation and assay we also obtained a sensitivity of 0.39 pg for I 1 -ketotestosterone comparable to that obtained by Cuisset et al. 1994). As with radioimmunoassay. the sensitivity is a function of both the specific activity of the label and the quality of the antiserum. and will therefore vary with the material available to the user. We have. in this study. used the antisera that were available to us but have little reason to expect that any antisera which gives good results with RIA will not give similar results to those that we describe for ELISA.

Previous reports of ELISA for steroid hormones have generally used horse-radish peroxidase as the enzyme label (Elder and Lewis 1985: Marcus and Durnford 1986. 1988; El Jabri 1991; Silvan et al. 1993a, b). This has a major disadvantage in that the

enzymatic reaction must be irreversibly stopped be- fore photometric measurement can be made. A major advantage of the use of acetylcholinesterase is that such an arrest is unnecessary and the sensitivity can be greatly increased by using even smaller concentrations of enzyme label and allowing the reaction to continue for several days until sufficient colour has developed. Plates can, in fact, be read daily until it is judged that the absorption is sufficient. The higher turnover rate of acetylcholinesterase compared to other enzyme labels also gives a potentially more sensitive assay (Pradelles et al. 1985). In this paper we describe a method that is sufficiently sensitive for the rapid measurement of steroid hormones in mammals or in common fish species such as salmonids and cyprinids. and present an example of such an application. Adaptation of the methodology for assay of hormones in plasma and tis- sue of small species such as zebrafish and minnows of interest in studies on endocrine disruption will be described elsewhere.

A major advantage of ELISA is the abilit! to min- imise errors resulting from multiple pipetting by use of multichannel pipettes. Inter- and intra-assay coeffi- cient of variations (CV) reported in RIA are frequently 15-20% even when measured at the steepest part of the curve. For the three assays reported here we ob- tain intra-assay CVs around 5% in the steep pan of the curve, rising to 15-2052 at the flatter end re- gions (Figure 2). while interassay variation (at 50% displacement) was around 852, comparable to that re- ported by Cuisset et al. (1994) for 1 1 -ketotestosterone. There was a highly significant correlation between the

360

201 16 1 y\

0 8 ; \ '* /J - 1

> I

44

12 4

'\J 1

0 ' I

1 10 100 1000

Testosterone (pghell)

m i 16 , *\ '\

i

i 0. '

1 10 100 1000

Estradiol (pghell)

1 10 100 lo00

17,2OpP(pg/well)

Figure 2. Intra-assay coefficient of variance (CV) for ELISA of testosterone. estradiol. and 17.2OPP determined from a standard curve with 8 replicates at each concentration.

steroid added to and that measured in stripped trout (Figure 3) and carp (data not shown) plasma. A further advantage of ELISA is that sample handling can also be made more rapid and reproducible by the use of low binding microtiter plates for sample storage and dispensing with multichannel pipettes.

Tracer preparation

A major problem with the assay of steroids in plasma of non-mammalian vertebrates is the availability of tracer. While testosterone and estradiol are readily available in tritiated form. 1 I-ketotestostero&: and 17.20pP must be prepared chemically from other materials and must then be purified chromatograph- ically. Such conversion of radioactive materials is beyond the expertise of many laboratories and has severely restricted measurement of these hormones in fish plasma. Availability of commercial enzymati- cally labelled steroids. including acetylcholinesterase labels (Cayman Chemical Company. Ann Arbor). is similarly restricted to those of mammalian steroids which are routinely assayed in clinical labs. Cuis-

set et al. (1994) prepared the acetylcholinesterase from electric eel. but in this communication we de- scribe a simple procedure for making enzyme la- bel from commercially available acetylcholinesterase which requires no facility more complicated than a simple chromatographic column. Using such a pro- cedure we have made acetylcholinesterase labelled 1 1 -ketotestosterone which has similar sensitivity to that described by Cuisset et al. (1994).

A major advantage of ELISA is that. since the steroid is coupled to the enzyme by exactly the same reaction as that used to link it to the bovine serum albumin to stimulate the immunological reaction in the host animal. it is possible to make an enzyme label for any steroid for which it is possible to raise an antiserum. In addition to the steroids described in this paper we have also prepared acetylcholinesterase label and obtained standard curves of similar sensitiv- ity for 17.2Oa-dihydroxy-4-pregnen-3-one ( 17.20aP). 1 1 -deoxycortisol. 17.20fi.1 l-trihydroxy-4-pregnen-3- one ( 20pS) . cortisol and 17-hydroxyprogesterone which demonstrates the widespread applicability of

361

lo00 U

3 In

e!

100 i? 2 10

P) t

3 In 0 cn Q) C l

.l-

/ 1 y=l .oox4.054; FO.999

1 10 100 1wo Testosterone added (pg)

lo00 h

0

U

ZI cn rn

n

g loo v

10 n

-- r l

c1 0

b

1 10 100 lo00

Estradiol added (pg) .*_

1 10 100 lo00

17,20pP added (pg)

Figure 3. Assay accuracy for testosterone. estradiol. and 17,20BP in trout plasma. A known amount of the steroid was added to an aliquot of a steroid free extract of pooled female trout plasma and 2 x serial dilutions made with the extract. Assays were performed in duplicate.

the methodology. This greatly expands the numbers of steroids that it is possible measure by ELISA. since only the labels for testosterone and estradiol are available commercially. Such flexibility is particularly important in fish in which a large number of steroids are produced by gonadal tissue (Kime 1993) which cannot be measured by radioimmunoassay due to lack of availability of radiolabel except by a prohibitively expensive custom preparation involving very high le\.- els of radioactivity. ELISA ma! therefore play a major part in clarifying the role. for example. of progesto- gens and Sa- and SB-reduced steroids in fish. As with radioimmunoassay it is essential that ELISA is ful ly validated for each new species used. This is especially true for a new method since i t is not known whether other chemicals extracted from plasma may interfere with binding or enzymatic processes. The present val- idation suggests that there are no such problems in salmonid and cyprinid fish.

Application to the measurement qf die1 hormone patterns

To demonstrate an application of the ELISA method we have measured the steroids testosterone. estradiol and 17.20BP concurrently in a series of hourly sam- ples removed from a single adult female trout via a cannula over a 24-h period. Testosterone and estra- diol showed a clear pattern of pulsatile secretion. even though these samples were taken at a point in the re- productive cycle when these steroids are considered to be low or basal (Scott and Sumpter 1983). While levels were generally low there were times during the diel cycle when several sharp peaks or pulses occurred which would have been missed with a longer sampling frequency or in seasonal studies where measurement was taken at a single time in the day.

Remarkably. few studies on fish have examined in- dividual diel profiles of reproductive steroids. There ha5 been only one concerted attempt at measuring the short-term changes of a reproductive steroid us- ing cannulation (Zohar et al. 1986). These workers

362

i a

16

n

14 1-

E p 12 W

0

-1

+ Estradiol .- .v. . Testoste m e + 17,ZOPP n

10 12 14 16 18 20 22 00 02 04 06 08 10 12 14 16

Time of Day (h) Figure 4. Die1 profiles of plasma estradiol. testosterone and 17.20bP in an individual female rainbow trout. The samples were taken on 2 February when the fish had ovulated but not yet spawned. Bar indicates hours of darkness.

measured esuadiol and estrone in combination with gonadotropin (GtH) over various short-term periods using cannulation, however, estradiol was measured at either 4 hour intervals or for a maximum period of only 12 hours (Zohar et al. 1986). While they did reveal very similar episodic pulses of GtH at a slightly later stage in the reproductive cycle they did not show any pulses of esuadiol or estrone. The re- sults we have obtained therefore present new data on short-term fluctuations in steroid levels and may help in further understanding the temporal endocrine mech- anisms involved in chronobiological processes. The data indicate that the exact time of sampling'and the numbers of specimens sampled may have a large effect on the measured steroid level, and that single point measurement should not be over-interpreted.

In this communication we present the data from a single fish as an illustration of the additional informa- tion that can be obtained using the higher sensitivity of ELISA using acetylcholinesterase as enzyme label compared with radioimmunoassay. These data form part of a much larger study in which around 40-fish were cannulated at two different periods of the repro-

ductive cycle. These results will be described in full in a separate publication.

Without the increased sensitivity and ability to use a low sample volume the concurrent measurement of all three steroids would have been impossible (es- pecially at this time of the year when steroid levels are naturally low). The rapid determination of several steroids from these very small volumes of blood al- lows the examination of profiles of steroids at even higher sampling frequencies or over longer periods than those described (the cannulation procedure used produced over 150 pI of plasma. leaving 100 p1 for measurement of other hormones ). This methodology will therefore be of great interest to workers examin- ing the temporal endocrine mechanisms involved in photoperiodic. circadian and other chronobiological processes.

Cortisol levels remained stable and low (5.9 x 3.2 ng ml-' ) throughout the sampling period. indica- tive of an unstressed tish. and showed no correlation with the pattern of reproductive steroids. This is in agreement with the finding of Zohar ( 1980) that the physiological levels of GtH were unaffected by a simi-

363

lar cannulation procedure and were not correlated with cortisol.

In addition to interest in seasonal and Die1 cycles in fish in aquaculture. there is now increasing concern about the effects of endocrine disrupting chemicals on both human and wildlife populations (Kime 1998). Fish provide an excellent biomonitor of aquatic pol- lution. but study of many of the effects of potential endocrine disrupters has been restricted by lack of availability of simple. rapid and sensitive methods co measure the steroids of major importance. Recent rec- ommendations for statutory testing of the effects of potential endocrine disrupters on wildlife noted the lack of availability of such methods to monitor both in \Gvo and in v i m steroid production (Tatters field et al. 1997). The assays described in this communication provide a firm basis for such regulatory tests.

Acknowledgements

We gratefully acknowledge grants from the Com- mission of the European Community, ERBCI- PACT930067 (to DEK, F. Ollevier (Leuven) and P. Epler (Krakow)) and AQ.3.735 (to DEK and W. Holtz. Gottingen). We also thank Melanie Zerulla (Schering, Berlin) for helpful suggestions on the use of PDlO columns for preparation of label, and Prof. W. Holtz (Gottingen) for providing facilities for fish sampling.

References

Axelsson. M . and Fntsche. R. 1994. Cannulation techniques. In: biochemist^ and Molecular Biology of Fi\hes: Vol 2. Analyt- ical Techniques. pp. 7-36. Edited b! P Hochachka W. and T.P. Mommsen. Elsevier. Amsterdam.

Bogomolnaya. A. and Yaron. Z. 1984. Stimulation in iYrro of estra- diol secretion b! the m a n of a cichlid fish. Sarothcrodon u m u . Gen. Comp. Endocrinol. 53: 187-196.

Cuisset. B.. Pradelles. P.. Kime. D.E.. Kuhn. E.R.. Babin. P.. Davail. S. and Le Menn. F. 1991. Enzyme immunoassa! for I I - ketotestosterone using acetylcholinesterase as lahel: application to the measurement of 1 I-ketotestosterone in plasma of Sibenan sturgeon. Comp. Biochem. Physiol. 108C: 229-341.

Diamandis. E.P. and Chnstopoulo\. T.K. I996 Immunoassa!. Aca- demic Press. San Diego.

El Jahri. J . 1991. Enzyme immunoassay for plasma estradiol using a monoclonal antibody. J . Steroid Biochem. Mol. Biol. 38: 339- 343.

Elder. P.A. and Lewis. J.G. 1985. An enzyme-linked immunosor- bent assa! (ELIS.41 for plasma testosterone. J . Steroid Biochem. 12: 635638.

Houston. .4.H. 1990. Blood and circulation. f r 7 : Methods in Fish Biology. pp. 273-334. Edited b) C.B. Schrecl, and P.B. Moyle. American Fisheries Societ!. Bethesda. Maryland.

Kime. D.E. 1993. 'Classical' and 'non-classical' steroids in teleost reproduction. Re!. Fish BioI. Fish. 3: 160-180.

Kime. D.E. 1998. Endocrine Disruption in Fi\h. Kluwer .Academic Publishers. Dordrecht.

Marc&.G-l. and Durnford. R. 1986. A simple enzyme-linked immunosorbent assa! for testosterone. Steroids 46: 975-986.

Marcus, G.J. and Durnford. R. 1988. Estradiol assa! by microtitre plate enzyme immunoassa). J . Steroid Biochem. 29: 207-212.

Pradelles. P.. Grassi. J. and Maclouf. J. 1985. Enzyme immunoas- says of eicosanoids using acetylcholinesteme as label: an alternative to radioimmunoassay. Anal. Chem. 57: 1170-1 173.

Price. C.P. and Kewman. D.I. 1997. Principles and Practice of Immunoassa!. Groves-Dictionaries. New York.

Scott A.P. and Sumpter 1.P. 1983. A comparison of the female re- productive cycle of the .4utumn-spawning and Winter-spawning strains of rainbow trout (SaImo gazrdneri Richardson). Gen. Comp. Endocrinol. 52: 79-85.

Scott. A.P.. Sheldnck. E.L. and Flint. A.P.F. 1981. Measurement of I7a.20B-dihydroxy-4-pregnen-3-one in plasma of trout (Salrno gairdneri Richardson): Seasonal changes and response to salmon pituitary extract. Gen. Comp. Endocrinol. 46: 444-45 1.

Silvan. G., Illera, I.C. and Illera, M. 1993a. Application of enzyme- linked immunosorhant assay ELISA) for determining testos- terone levels in the follicular fluid of prepubertal and pubertal heifers. Anim. Reprod. Sci. 32: 375-28 1.

Silvrin. G.. Illera. I.C. and Illera. M. 1993b. Determination of fol- licular fluid estradiol levels by enzyme-linked immunosorbent assa!. Steroids 58: 324-329.

Simpson. T.H. and Wright. R.S. 1977. A radioimmunoassa! for 1 1 - ketotestosterone: its application in the measurement of levels in hlood serum of rainbou trout (Salmrj ,gairdnerii. Steroids 29: 383-398.

Tattersfield. L.. Matthiessen. P.. Campbell. P.. Grandy. N. and Lange. R. (Editors ). 1997. SETAC-Europe/OECD/EC Expert Workshop on Endocrine Modulators and Wildlife: Assessment and Testing (EMWAT). Veldhoven. The Netherlands. l & I 3 .April 1997. SET,4C-Europe. Amsterdam.

Zohar. I'. 1980. Dorsal aorta cathetenzation in rainhou trout tSulmo ,guirdneri). 1. Its validit! in the htud! of h l d gonadotropin patterns. Reprod. Nutr. Develop. 10: I8 1 1 -1 823

Zohar. Y.. Breton. B. and Fostier. A. 1986. Short-term profiles of plasma gonadotropin and estradiol-17p le\elz in the female rainbow trout (Salrno guirdtieri). from earl) ovarian recrudes- cence and throughout vitellogenesi\. Gen. Comp. Endocrinol. 64: 171-188.