Environmental Toxicology and Chemistry, Vol. 30, No. 2, pp.

477486, 2011 # 2010 SETAC Printed in the USA DOI: 10.1002/etc.390

DEVELOPMENT OF AN ENZYME-LINKED IMMUNOSORBENT ASSAY FOR QUANTIFYING VITELLOGENIN IN PACIFIC SALMON AND ASSESSMENT OF FIELD EXPOSURE TO ENVIRONMENTAL ESTROGENS
KAREN A. PECK,* DANIEL P. LOMAX, O. PAUL OLSON, SEAN Y. SOL, PENNY SWANSON, and LYNDAL L. JOHNSON
Northwest Fisheries Science Center, National Oceanic and Atmospheric Administration, Seattle, Washington (Submitted 5 May 2010; Returned for Revision 30 June 2010; Accepted 7 September 2010) Abstract A competitive enzyme-linked immunosorbent assay was developed to quantitate vitellogenin (VTG) in plasma and serum of coho (Oncorhynchus kisutch) and chinook (O. tshawytscha) salmon. The working range of the assay was 9 to 313 ng/ml (8020% binding), with 50% binding at 54 ng/ml. The intra-assay and interassay variations at approximately 50% binding were 8.1% (n 9) and 9.0% (n 9), respectively. Dilution curves of plasma or serum from coho and chinook females and estrogen-treated males were parallel to the puried coho VTG standard curve. Male plasma samples could be assayed at a minimum dilution of 1:40 (chinook) or 1:75 (coho) without assay interference because of high sample concentration, whereas minimum acceptable dilutions of male serum samples were 1:200 (chinook) or 1:600 (coho). Identication of proper techniques for preserving VTG integrity in plasma and serum samples showed that VTG from both species was robust; both sample types required no protease inhibitor despite subjection to two freezethaw cycles. To test its applicability, this assay was used to measure VTG in out-migrating juvenile chinook that were collected from urban and nonurban areas in Puget Sound, Washington, USA. Results showed a small but signicant plasma VTG elevation at two urban sites, suggesting that these juveniles may be exposed to environmental estrogens at an early life stage. Also, wild sh tended to have higher plasma VTG levels than hatchery sh collected in the eld. Elevation of mean VTG levels was similar to that previously reported in male English sole from the same area, where both males and females exhibited alterations in timing of spawning. Environ. Toxicol. Chem. 2011;30:477486. # 2010 SETAC KeywordsVitellogenin Enzyme-linked immunosorbent assay Salmon Endocrine disruptors Puget Sound

The environmental occurrence of endocrine-disrupting contaminants and their adverse effects on wildlife and humans is an important concern. Within this contaminant class, the estrogenic compounds have received the greatest attention, partly because of early observations of their dramatic feminizing effects on male sh in rivers in the United Kingdom [1]. Estrogenic compounds mimic the natural hormone estrogen, thus binding the cellular estrogen receptor and initiating the regulation of various genes that can ultimately impact natural hormonal functions. The effects of exposure of aquatic life to estrogenic compounds can manifest as reproductive, developmental, behavioral, or smoltication abnormalities. Numerous anthropogenic contaminants have been shown to possess estrogenic activity, including natural and synthetic hormones, alkyl phenolic chemicals (surfactants), phthalates (plasticizers), and certain polychlorinated biphenyls and organochlorine pesticides [1]. Laboratory studies in rainbow trout and other teleosts have demonstrated that environmentally relevant concentrations of estrogenic compounds can have deleterious effects on sh reproductive development, function, and behavior, and these effects co-occur with plasma vitellogenin induction in male sh [24]. Vitellogenin (VTG) is an egg yolk precursor protein, normally expressed in oviparous female sh in response to endogenous estrogen. However, because VTG induction in males or juveniles can occur because of exposure to exogenous
* To whom correspondence may be addressed (karen.a.peck@noaa.gov). Published online 29 October 2010 in Wiley Online Library (wileyonlinelibrary.com). 477

estrogen, it has been used extensively as a biomarker of estrogenic exposure in eld and laboratory studies. Elevated VTG levels have been observed in males of several sh species collected from various marine and freshwater sites in the United States [57]. Specically, Johnson et al. [6] reported an association between VTG induction in male English sole and alterations in timing of spawning in both males and females at some sites in Puget Sound, Washington, USA. Pacic salmon, including chinook, which is listed as threatened under the Endangered Species Act, and coho, which is a species of concern [8], outmigrate through these same areas. In addition, salmonid upstream freshwater habitat receives agricultural runoff, which may be a source of estrogenic compounds. The presence of these compounds in both marine and fresh waters in Puget Sound has been conrmed ([9]; http://your.kingcounty.gov/ dnrp/library/2007/kcr1976.pdf), with occurrence at concentrations that have been shown to elicit adverse reproductive effects in laboratory sh [24]. Because estrogenic exposure during early life stages has been shown to lead to adverse reproductive effects [3,4,10,11], these outmigrating juvenile salmon may be exposed at critically sensitive stages during early development. To assess whether Pacic salmon are being exposed to environmental estrogens, a competitive enzyme-linked immunosorbent assay (ELISA) was developed to measure plasma or serum VTG in these sh, with emphasis on assay sensitivity and reproducibility, evaluation of assay inhibition by high concentrations of plasma or serum, and identication of proper preservation of sample integrity. This ELISA was then validated by measuring plasma VTG levels in juvenile chinook outmigrating through contaminant-impacted areas in Puget Sound.

478

Environ. Toxicol. Chem. 30, 2011 MATERIALS AND METHODS

K.A. Peck et al.

Fish

For ELISA development and validation, coho salmon (Oncorhynchus kisutch) eggs were obtained from the University of Washington Hatchery and were hatched and reared to 2 or 3 adults at our facility. Chinook salmon (O. tshawytscha) were obtained from the Stillaguamish Hatchery (Arlington, Washington, USA) as returning adults in late summer of 2006 and 2007. For assessment of VTG levels in eld-caught sh, outmigrating juvenile chinook salmon were captured by beach seining or tow netting in various locations in Puget Sound (Fig. 1). Fish that had a hatchery mark (n clip) were recorded as

hatchery sh, and those without a hatchery mark were presumed wild. Laboratory control chinook were hatched and reared at the National Marine Fisheries Service facility in Mukilteo, Washington, from eggs obtained from the University of Washington Hatchery. These juveniles went through smoltication and were transferred to seawater approximately ve to six weeks before plasma collection (James Meador, National Marine Fisheries Service, Seattle, Washington, USA, personal communication).
Source of puried VTG

All chemicals were from Sigma-Aldrich unless otherwise specied. Coho VTG for use in ELISAs was puried from

Fig. 1. Map of Puget Sound (WA, USA) indicating sampling sites for juvenile chinook. Latitudes and longitudes are: Snohomish River Mainstem 478 56.0000 N 1228 10.1430 W; Skagit BayHope Island 488 24.1760 N 1228 34.1580 W; Elliott BayMyrtle Edwards Park 478 37.0200 N 1228 21.5050 W; Duwamish River Kellog Island 478 33.4210 N 1228 20.7380 W; Snohomish RiverLangus Pier 488 0.1080 N 1228 10.6790 W.

Quantification of vitellogenin in pacific salmonids

Environ. Toxicol. Chem. 30, 2011

479

serum according to Hara et al. [12]. Briey, an adult 2 female coho was anesthetized with buffered tricaine methanesulfonate (0.05% MS-222; Argent Chemical) and injected intraperitoneally on day 1 and day 7 with 10 mg 17b-estradiol/kg body weight. The 17b-estradiol was rst dissolved in coconut oil as described [13]. On day 20, blood was collected from the anesthetized sh by caudal puncture, and serum was prepared and stored at 808C. Vitellogenin was puried from serum via gel ltration chromatography [12] and quantitated using the Bradford Reagent (Sigma-Aldrich) according to the manufacturers directions. Puried VTG was aliquoted for single use and stored at 808C with the addition of 0.5% w/v bovine serum albumin (Invitrogen) for use with ELISA standards, or without bovine serum albumin when used for coating ELISA microplates. Puried rainbow trout VTG (Biosense) was reconstituted in phosphate-buffered saline (0.02 M sodium phosphate, 0.15 M sodium chloride, pH 7.2), followed by addition of 0.5% w/v bovine serum albumin. It was then aliquoted for single use and stored at 808C.
Blood sampling and storage

General VTG ELISA procedure

The method presented here is a competition ELISA for measuring salmonid VTG in plasma or serum samples and is a modication of the procedure of Lomax et al. [13] and Rodriguez et al. [16]. Puried coho VTG coated on wells of a 96-well microplate competes with free salmonid VTG in solution (the standard or sample) for binding to an anti-salmon VTG polyclonal antibody. The amount of this primary antibody bound to the puried coho VTG coated on the plate (herein referred to as bound primary antibody) was measured via the addition of an enzyme-labeled secondary antibody that was specic for the primary antibody. The enzyme activity was measured colorimetrically and is inversely proportional to the amount of free VTG in the standard or sample. The sample VTG concentration was determined by extrapolation from a standard curve composed of a dilution series of quantitated, puried free coho VTG.
Stepwise VTG ELISA procedure

For routine analysis, blood was collected from adult coho and chinook via caudal puncture, and serum and plasma were prepared as described [12,13]. For eld-caught juvenile chinook, the tail was sectioned at the caudal peduncle, followed by blood collection in a heparinized capillary tube, centrifugation at room temperature for 10 min at 12,700 g, and aspiration of plasma. All serum and plasma samples were stored at 808C until use. Calf serum was from Sigma-Aldrich. For experiments involving comparisons between VTG measurement in serum versus plasma, and the addition of protease inhibitors or additives to protect VTG integrity, both serum and plasma samples were prepared from the same adult female (either coho or chinook) and were additionally treated as follows: no protease inhibitor/additive was added; the protease inhibitor aprotinin was added to a nal concentration of 5 IU/ml [14]; the protease inhibitor phenylmethylsulfonyl uoride (PMSF) was added to a nal concentration of 1 mM [13]; or samples were diluted 1:100 in the additive 10 mM citrate, pH 6.5, plus 16 mg/ml polyethylene glycol 8000 (PEG) [15]. These same samples were also used to assess the effects of various storage temperatures on VTG integrity, and were further handled as follows before use in a VTG ELISA. One set of samples was prepared and stored on ice for several hours before being assayed (fresh). These samples were subjected to the least possible handling and therefore would be expected to have the best VTG integrity. A second set of samples was stored at 808C followed by thawing at 48C immediately before assaying (one freezethaw cycle: frz/thaw 1)the typical method used for eld-collected samples. To identify optimum storage temperature conditions for samples that might need to be re-assayed (and thus subjected to additional handling after the rst freezethaw cycle), an additional set of samples was subjected to two freezethaw cycles (frz/thaw 2) before being assayed. For comparison, a nal set of samples was subjected to intentional VTG degradation via one freezethaw cycle followed by 2 d at room temperature (frz/thaw 2d@RT). In addition, plasma samples from individual adult male chinook were subjected to two freezethaw cycles before use in a VTG ELISA.

VTG coating and blocking of microplates. For assay optimization steps, frozen puried coho VTG was thawed at 48C and serially diluted two fold in coating buffer (sodium carbonate 0.05 M, pH 9.6) from 600 ng/ml to 9.4 ng/ml. For routine assays, a VTG coating concentration of 25 ng/ml was eventually found to be optimal. Diluted VTG was added to 96-well Corning enzyme immunoassay/radioimmunoassay microplates (No. 3590) at 100 ml per well. In addition, the nonspecic binding (NSB) of the VTG-specic primary antibody was determined by coating three wells per plate with control male coho plasma or serum that was diluted in coating buffer to achieve the same amount of total protein per well as the puried VTG. Total protein in plasma or serum was quantitated using the Bradford Reagent, following manufacturers directions. Three blank wells containing only coating buffer were also included as a plate blank. Plates were sealed and incubated at 48C overnight, followed by washing four times (by manual inversion and relling) with 200 ml per well of phosphatebuffered saline plus 0.05% v/v Tween-20 (PBST). Wells were then blocked by adding 200 ml per well of blocking solution (PBST plus 2% w/v nonfat dry milk powder), and plates were sealed and incubated overnight at 48C. Finally, wells were again washed four times with PBST, and sealed plates containing 200 ml per well PBST were stored at 808C until use. This method allowed for large multiples of plates to be prepared simultaneously. Preparation of standards, samples, and primary antibody. All frozen standards and samples were thawed at 48C before use. The standards, samples, and primary antibody were diluted in assay buffer (PBST plus 1% w/v nonfat dry milk powder) to a 2 nal assay concentration. For the standard curve, the puried coho VTG (stored with 0.5% w/v bovine serum albumin added) was serially diluted twofold, with a nal assay concentration range of 625 to 4.9 ng/ml VTG. A quality control samplevitellogenic coho female serum (aliquoted for single use and stored at 808C)was diluted to achieve a nal assay dilution of 1:1,000,000, which yielded approximately 50% binding of the primary antibody. The primary antibody was a rabbit anti-Atlantic salmon-VTG polyclonal antibody (Biosense). A commercial antibody was chosen for this step because of the poor performance of a rabbit anti-coho salmon-VTG polyclonal antibody that we had produced and tested. For assay optimization steps, this antibody was serially diluted to achieve a nal assay concentration of 1:1,000 to 1:7,000. For routine

480

Environ. Toxicol. Chem. 30, 2011

K.A. Peck et al.

assays, a nal assay dilution of 1:1,000 for the primary antibody was eventually found to be optimal. Per well, 50 ml of the diluted primary antibody (2 nal concentration) was mixed with 50 ml of each diluted standard or sample (2 nal concentration) to achieve a 1 nal assay concentration of all components, and then preincubated in microfuge tubes at 48C overnight. A zero-vitellogenin standard (i.e., assay blank) was also included, which contained primary antibody plus PBST only. VTG-specic antibody incubation. Frozen microplates that had been coated and blocked were thawed overnight at 48C and then emptied. One hundred microliters of each standard or sample preincubation mixture was added per VTG-coated well. In addition, the plate blank and NSB wells received 100 ml each of the zero-vitellogenin standard preincubation mixture. All blank, standard, and sample analyses were performed in triplicate. Plates were then sealed and incubated for 2 h at 378C and 250 rpm on a Jitterbug-4 microplate incubator shaker (Boekel Scientic). Secondary antibody incubation. Wells were washed four times with PBST before addition of 100 ml/well secondary antibody, an anti-rabbit immunoglobulin G (H L) horseradish peroxidase conjugate (Promega) diluted 1:2,500 in assay buffer. Plates were then sealed and incubated for 45 min at 378C and 250 rpm. Peroxidase activity determination. Wells were washed four times with PBST before addition of 100 ml/well tetramethylbenzidine substrate solution (Pierce), prepared just before use. Plates were incubated for 30 min at room temperature, followed by termination of the reaction by the addition of 100 ml/well 2 M sulfuric acid. Absorbance values were measured at 450 nm with a VERSAmax Tunable microplate reader (Molecular Devices).
Expression of results and statistical analysis

iance [18] was used to examine the effect of sh origin (hatchery vs wild) on VTG concentrations while adjusting for the effect of site of collection. For all analyses, VTG concentrations were log-transformed to produce a normal distribution. All statistical analyses were performed with the JMPTM statistical package (SAS Institute).
RESULTS

Assay optimization, sensitivity, and reproducibility

A checkerboard titration approach was applied to determine the optimum reagent concentrations of puried coho VTG used to coat microplate wells and of primary antibody (Fig. 2). A target OD of 1.7 was chosen to obtain consistent results within the microplate reader linear range. The VTG concentrations that produced an OD of 1.7 at various primary antibody dilutions were identied. Next, a standard curve was included in assays containing these optimum combinations of primary antibody and VTG coating concentrations, and the working ranges of the assays were determined. The combination that elicited the highest assay sensitivity was chosen for routine assays: 1:1,000 primary antibody dilution plus 25 ng/ml VTG coating. These conditions produced a working range of 9 to 313 ng/ml VTG (8020% binding), with 50% binding at 54 ng/ml VTG. At approximately 20% binding, the intra-assay and interassay coefcients of variation were 18.0% and 11.6%, respectively; at approximately 50% binding they were 8.1% and 9.0%; and at approximately 80% binding they were 5.9% and 13.1% (n 9 for all values). Good parallelism was shown when the slopes from nine standard curves, performed over several months, were compared using F test of mean squares ( Fobs 0.21 < F0.05 2.18 with degrees of freedom 8, 39). Nonspecic binding wells elicited approximately 0 to 2% binding.
Plasma and serum dilution curves

All blank, standard, and sample assays were performed in triplicate. The mean absorbance value of the three plate-blank wells, expressed as optical density (OD), was subtracted from the OD of all other wells. The mean OD of each triplicate set of standard or sample wells (ODs) was used to calculate the proportion of bound primary antibody (B), expressed as a percentage of the bound primary antibody in the zero-vitellogenin standard (OD0, or maximum binding of primary antibody): B (%) [(ODs NSB)/(OD0 NSB)] 100, where NSB is the mean OD of the nonspecic binding wells. To achieve a linear transformation of standard and plasma or serum dilution curves, B values were logit transformed [17]: logit B loge X [B/(100 B)], and plotted against log VTG concentration or log sample dilution. Analysis of covariance was used to assess parallelism between the standard curve and various plasma or serum sample dilution curves [18]. Parallelism indicates that the anti-VTG antibody binds in a similar manner to the puried coho VTG and to native VTG in samples. Sample VTG concentration was interpolated from the standard curve. The effects of sample collection procedures and storage conditions on VTG measurement were evaluated by analysis of variance and Dunnetts comparison to control method for testing for signicant differences among means [18]. For these comparisons, fresh plasma samples were designated as the control. The same statistical tests were used to evaluate the effect of collection site on VTG concentration, using the VTG concentration in laboratory control sh as the control. Additionally, for the eld-collected sh, two-way analysis of var-

Dilution curves were compared against the puried coho VTG standard curve to assess the primary antibodys ability to recognize native VTG in plasma versus serum from coho males and females (Fig. 3A). Data points in the approximately 80 to 20% binding range were linearized using logit/log transforma-

4 3.5

Absorbance (450 nm)

3 2.5 2 1.5 1 0.5 0 0 100 200 300 400 500 600

Primary Antibody Dilution

1:1,000 1:2,000 1:3,000 1:4,000 1:7,000

Coho VTG coating (ng/ml)

Fig. 2. Checkerboard titration of primary antibody versus coho vitellogenin (VTG) coating for optimization of coho enzyme-linked immunosorbent assay.

Quantification of vitellogenin in pacific salmonids

Environ. Toxicol. Chem. 30, 2011

481

tion (Fig. 3B). For coho females and estrogen-treated males, good parallelism was demonstrated between the standard curve and plasma or serum at all dilutions tested. Dilution curves of normal coho male plasma and serum indicated a lack of measurable VTG at dilutions at or above 1:75 and 1:600, respectively. However, at lower dilutions, high sample concentration interfered with the assay to give false-positive results. Therefore, routine assays of coho male plasma and serum were performed at minimum dilutions of 1:75 and 1:600, respectively, which corresponds to a detection limit of approximately 675 ng/ml VTG for plasma and approximately 5,400 ng/ml VTG for serum. Calf serum, which also lacks VTG, behaved similarly to coho male plasma and serum by giving falsepositive results for VTG measurement as sample concentration increased. To assess whether the primary antibody would cross-react with VTG from other salmonids in a parallel fashion, dilution curves spanning the 80 to 20% binding range of puried rain-

bow trout VTG and of male and female chinook plasma versus serum were compared against the coho VTG standard curve (Fig. 4A). Good parallelism was shown with puried rainbow trout VTG, and for female chinook plasma or serum at all dilutions tested (Fig. 4B). Dilution curves of normal chinook male plasma and serum indicated a lack of measurable VTG at dilutions at or above 1:40 and 1:200, respectively. Accordingly, these two dilutions were chosen as minimum dilutions for routine assays, with corresponding detection limits for male chinook plasma at approximately 360 ng/ml VTG, and approximately 1,800 ng/ml VTG for serum.
Plasma versus serum, and sample storage requirements

To assess whether VTG measurements were similar in plasma versus serum samples, and to identify optimum storage requirements for samples, plasma and serum were collected from a single female (coho or chinook) and subjected to various conditions before VTG concentrations were measured. Samples were stored with nothing added to them (alone), or with aprotinin, PMSF, or citrate/PEG buffer added directly after

Fig. 3. (A) Binding displacement curves of puried coho vitellogenin (VTG) standard and serial dilutions of coho female or male plasma versus serum, as well as serum from a 17b-estradioltreated coho male, and calf serum. Standard curve data are representative of a typical curve. Serum or plasma was from a single sh, except for coho male plasma and serum, which was combined from three sh. (B) Linearization of binding curves by logit transformation. Parallelism was demonstrated between the coho VTG standard curve (s 0.78, r2 0.97) and coho female plasma (s 0.78, r2 0.98), coho female serum (s 0.78, r2 0.96), and 17b-estradiol treated coho male serum (s 0.75, r2 1.00).

Fig. 4. (A) Binding displacement curves of puried coho vitellogenin (VTG) standard, puried rainbow trout VTG, and serial dilutions of chinook female or male plasma versus serum. Standard curve data are representative of a typical curve. Serum or plasma was combined from three to six sh. (B) Linearization of binding curves by logit transformation. Parallelism was demonstrated between the coho VTG standard curve (s 0.78, r2 0.97) and puried rainbow trout VTG (s 0.79, r2 0.93), chinook female plasma (s 0.79, r2 0.98), and chinook female serum (s 0.77, r2 0.98).

482

Environ. Toxicol. Chem. 30, 2011

K.A. Peck et al.

sample preparation to potentially improve VTG integrity during storage. See the Materials and Methods section for details. In addition, optimum primary storage temperatures of these samples were assessed as described in the Materials and Methods section. Vitellogenin degradation would be expected to manifest as an increase in measured VTG concentration, because VTG degradation products are immunoreactive [13,14,19]. Results in Figure 5A indicate that coho VTG stability is robust. Regardless of whether plasma or serum was the VTG source, VTG concentrations were equivalent (no statistically signicant difference), and no protease inhibitors were required to protect VTG integrity. This was demonstrated by the fact that

VTG concentrations were not signicantly different between fresh samples that had nothing added, versus those with aprotinin or PMSF additions. However, the addition of citrate/PEG buffer gave erratic results, especially with serum samples, and was actually deleterious to VTG, causing an increase in VTG concentration (i.e., degradation) under some storage temperature conditions. Also, in the absence of citrate/PEG buffer, storage temperature did not signicantly alter VTG integrity, even when storage was at room temperature for 2 d. Chinook VTG was also found to be quite robust (Fig. 5B). Although serum VTG concentrations were somewhat lower than plasma VTG concentrations under some conditions, these differences were not statistically signicant. Also, the VTG concentration in plasma and serum was not signicantly different regardless of the addition of protease inhibitors or additives. This was true of all storage temperature conditions except room temperature, in which a signicant increase in VTG concentration/degradation was indeed detected in all plasma samples, as well as serum samples with either no additions or with aprotinin added. However, both PMSF and citrate/PEG buffer were effective in preserving serum VTG integrity even at room temperature. Plasma samples from adult male chinook also were subjected to two freezethaw cycles before use in a VTG ELISA. Vitellogenin measurements were not signicantly different after one versus two freezethaw cycles (data not shown), indicating that male samples with low VTG concentrations compared with female samples also resisted degradation under these storage conditions. Also, in juvenile and adult male chinook, VTG integrity did not differ when PMSF was added to whole blood versus plasma (data not shown).
VTG measurement in juvenile chinook from Puget Sound

Fig. 5. Comparison of coho (A) or chinook (B) mean vitellogenin (VTG) concentration in plasma versus serum samples stored under different conditions. Samples were stored with nothing added to them (alone), or with the addition of aprotinin, phenylmethylsulfonyl uoride (PMSF), or citrate/ polyethylene glycol 8000 (PEG) buffer ( buffer). Also, before assaying, samples were either stored on ice for several hours (fresh), subjected to one freezethaw cycle involving storage at 808C followed by thawing for approximately 1 h at 48C (frz/thaw 1), subjected to two such freezethaw cycles (frz/thaw 2), or subjected to one freezethaw cycle plus 2 d at room temperature (frz/thaw 2d@RT). See Materials and Methods section for details. Error bars indicate standard deviation from the mean. Fresh plasma was the control mean for all other plasma or serum samples. Assessment of signicant difference from the control mean was performed by analysis of variance and Dunnetts comparison with control method for testing for signicant differences among means. 1 p < 0.05, 2 p < 0.005, 3 p < 0.001.

To assess VTG levels in eld-caught sh from sites with potential estrogenic contamination, available plasma samples from juvenile chinook collected from various urban sites in Puget Sound in the summer of 2006 were analyzed. These sites included Elliott BayMyrtle Edwards, Duwamish River Kellog Island, and Snohomish RiverLangus Pier (Fig. 1). Samples from unexposed juvenile chinook that were hatched and raised under laboratory conditions were collected around the same time and used for comparison. Plasma samples from sh from less contaminated nonurban sites were available the following year (2007). These sites included Snohomish RiverMainstem and Skagit BayHope Island. Some sh from the nonurban sites were found to have plasma VTG levels that were higher than the highest levels found in any of the laboratory control sh (Table 1). Therefore, mean plasma VTG levels of sh from the Puget Sound sites were statistically compared with those of the laboratory control sh rather than with those of nonurban sites. Plasma VTG levels in samples collected at Elliott Bay Myrtle Edwards on June 14, 2006 were not signicantly different from laboratory controls (mean VTG concentration of 941 vs 694 ng/ml, respectively) (Table 1). However, sh collected two weeks later (June 28, 2006) from the same site, and sh from the Duwamish RiverKellog Island (June 14, 2006), had a statistically signicant elevation of mean VTG that was twofold to threefold higher than that of the laboratory controls (mean VTG concentration of 1,967 and 1,509 ng/ml, respectively). Juvenile chinook that were collected in 2007 from the nonurban sites Skagit BayHope Island and Snohomish RiverMainstem had mean VTG concentrations of 1,054

Table 1. Comparison of mean plasma vitellogenin (VTG) levels from juvenile chinook collected from various sites in Puget Sound (WA, USA) No. wild sh / No. hatchery sh: Mean St plasma Dev VTGa (ng/ml) Wild shb (N) 0-999 1,000-1,999 2,000-2,999 0/3 1/0 0/4 0/1 1/1 15/17 58% 56% 29% 2/9 8% 15% 1/0 Hatchery sh (N) ng/ml plasma VTG 3,000-3,999 4,000-4,999 5,000-5,999

Quantification of vitellogenin in pacific salmonids

Site

Collection date

Total sh (N)

Mean sh weight (g)

Laboratory control (Hatchery sh)c Snohomish R. Mainstem Skagit Bay Hope Is. Ellliott Bay Myrtle Edwards Ellliott Bay Myrtle Edwards Duwamish R. Kellog Is. Snohomish R. Langus Pier 10 25 20 8 10 8 14 3 9 1 1 3 9 26 7/33 27% 19.2 5.5 13.5 7.7 8.9 6.0 11.2 694 997 1,054 941 1,967d 1,509e 1,320 423 710 1,054 631 914 409 525 Total: 9 0/16 3/8 0/5 0/1 0/1 4/2 1 3/3 4/3 1/2 0/4 3/3 4/2 10 22 11 7 9 5 5 59 Total: Percentage of total no. of wild sh: Percentage of total no. of hatchery sh:

27 June 21 June 25 July 14 June 28 June 14 June 24 Aug

2006 2007 2007 2006 2006 2006 2006

1/0

1/0 4%

1/0 4%

Assessment of signicant difference from laboratory control mean was performed by analysis of variance and Dunnetts comparison with testing signicant differences among means. Fish that did not have a hatchery mark (n clip) were presumed to be wild. c Indicates plasma samples from unexposed juvenile chinook that were obtained as eggs from a hatchery, then hatched and raised under laboratory conditions. d p < 0.005. e p < 0.05. Environ. Toxicol. Chem. 30, 2011 483

484

Environ. Toxicol. Chem. 30, 2011

K.A. Peck et al.

and 997 ng/ml VTG, respectively, which were not signicantly different from the laboratory control levels. More specically, plasma VTG levels in laboratory control sh ranged from 360 to 1,831 ng/ml. All of the Puget Sound sites had sh with VTG levels above that range (2,000 ng/ml or higher) (Table 1). In particular, the June 28, 2006 collection at Elliott BayMyrtle Edwards was notable because 50% of the sh collected from that site had plasma VTG levels at 2,000 ng/ ml or higher. Interestingly, within the entire current study, the two sh with the highest plasma VTG levels were both wild sh, one from a nonurban site (Skagit BayHope Island) and the other from the June 28, 2006, collection at Elliott BayMyrtle Edwards. When comparing all sh collected from Puget Sound, regardless of date or site, 74% of the wild sh had plasma VTG levels of 1,000 ng/ml or higher, compared with only 44% of the hatchery sh. A two-way analysis of variance testing the effects of site of capture and sh origin (i.e., wild vs hatchery) on VTG levels of all sh collected in Puget Sound showed that wild sh had signicantly higher levels of plasma VTG than hatchery sh, even when the effect of site of capture was taken into account.
DISCUSSION

A competitive ELISA method has been developed and validated to measure VTG in plasma or serum from coho or chinook salmon. This assay has a sensitivity range of 9 to 313 ng/ml VTG (8020% binding), which is similar to that of VTG ELISAs for other salmonids such as rainbow trout, brown trout, and brook trout [14,2022] and other teleosts, such as common sole, zebrash, and English sole [13,16,23]. The demonstration of parallelism between the puried coho VTG standard curve and dilution curves of coho and chinook female plasma or serum, estrogen-treated coho male serum, and puried rainbow trout VTG indicates that the anti-VTG antibody binds in a similar manner to native VTG in these plasma or serum samples and to puried coho and rainbow trout VTG. Also, no statistically signicant difference was seen between the measurable VTG level in plasma versus serum from a single female coho or chinook. Therefore, VTG measurement can be conducted regardless of whether a study design can accommodate centrifugation of heparinized blood in the eld (for immediate plasma preparation and freezing) or if nonheparinized blood must rst be transported on ice to the laboratory before centrifugation (for delayed serum preparation and freezing). However, the collection of plasma would be preferred over serum for ELISA analysis of males or juveniles. This is attributable to the fact that assay interference (and thus false positives) caused by high sample concentration was much less pronounced with plasma versus serum, and the ability to dilute samples as little as possible maximizes detection sensitivity, especially for samples with lower VTG levels. Why plasma versus serum would behave differently in this respect is not clear. However, other investigators have also reported assay interference or inhibition in various VTG ELISAs [13,14,20,21,23], usually necessitating plasma dilutions of 1:10 to 1:100 for accurate VTG measurement. Various researchers have noted the instability of VTG in samples from different sh species, such as rainbow trout, cod, turbot, and wolfsh [19], fathead minnow [15], and tilapia [24]. Therefore, protease inhibitors are often added during plasma and serum preparation to prevent VTG degradation. For example, aprotinin has been used with several salmonids, such as rainbow trout, brook trout, masu salmon [14,22,25], and other

teleosts [15,23], PMSF has been used with a variety of teleosts [13,16,24], and citrate/PEG buffer has been used with fathead minnows [15]. Because native VTG and its degradation products are immunoreactive in ELISAs [13,14,19], the possibility for articially elevated VTG measurements, especially in eld studies of male or juvenile sh potentially exposed to estrogenic compounds, is of concern. Therefore, the effects of various protease inhibitors and storage temperatures on the integrity of coho and chinook VTG in plasma and serum samples were carefully evaluated. We found that VTG in samples from both species was in fact quite robust and could be subjected to multiple freezethaw cycles, thereby allowing easy re-assaying of samples, without the need for the addition of expensive or hazardous protease inhibitors such as aprotinin or PMSF. Even though Brodeur et al. [15] have reported the protective nature of citrate/PEG buffer with fathead minnow VTG, this buffer actually compromised the integrity of coho VTG under some conditions. This nding supports the notion that a stabilization approach for individual proteins is not easily generalized, and usually benets from a trial-and-error approach ([15] and references therein). Alternatively, aprotinin and PMSF did not compromise coho or chinook VTG integrity. Also, the fact that VTG measurements in plasma and serum were not statistically different obviates the need for a centrifuge in the eld for preparing plasma. Rather, blood can easily be stored on ice and prepared as serum on subsequent return to a laboratory. Validation of this ELISA involved plasma VTG measurement of juvenile chinook from urban sites in Puget Sound that may contain estrogenic compounds, as well as sh from less contaminated nonurban sites and laboratory control sh. The mean VTG concentration in laboratory control sh was similar to that found in studies of juvenile rainbow trout [26,27] and was not signicantly different from that of the nonurban sites. Fish collected from two Puget Sound sites in the current study, Elliott BayMyrtle Edwards and the Duwamish RiverKellog Island, had a small but statistically signicant increase (two to three fold) in mean plasma VTG compared with laboratory control sh, suggesting exposure to estrogenic compounds. A similar level of mean VTG elevation was observed in adult male English sole collected by Johnson et al. [6] from the same areas. Juvenile chinook salmon may spend several weeks in Puget Sound estuaries before migrating to the ocean [28], and it is reasonable to expect that estrogenic exposure and VTG elevation could occur during this time frame. Interestingly, elevated VTG levels in adult male English sole were associated with alterations in timing of spawning in both males and females. Both chinook and English sole from Myrtle Edwards, a site that is near several combined sewer and storm drain overows (CSOs), had the highest mean plasma VTG levels compared with other sites. One possibility for the increase in mean plasma VTG levels in juvenile chinook from Myrtle Edwards between June 14, 2006 and June 28, 2006 is that the sh may have lingered at this site, thereby prolonging their exposure to possible contaminants there. Also, in June 2006, two treated CSO events occurred at the nearby Elliott West CSO Treatment Facility, with a combined volume of 35.3 million gallons, and one untreated CSO event occurred at the nearby King Street CSO, with a volume of one million gallons [29]. These events may have been a source of estrogenic contaminants. None of the CSOs that are in the vicinity of any of the other Puget Sound sites in the current study were reported to have overows around the times of sampling.

Quantification of vitellogenin in pacific salmonids

Environ. Toxicol. Chem. 30, 2011

485

In rainbow trout, as well as Atlantic and masu salmon, plasma VTG increases within hours or days of estrogenic exposure, and remains elevated for weeks after exposure cessation [25,30,31]. Therefore, whether juvenile chinook in the current study had elevated VTG because of exposure to estrogenic contaminants at the collection sites or at upstream or nearby areas is unknown. Prior exposure could explain why the sh with the highest plasma VTG level in the current study was collected from a nonurban site (Skagit BayHope Island). Also, differences may exist in contaminant exposure between wild and hatchery sh, because wild sh in Puget Sound were found to have statistically higher levels of plasma VTG than hatchery sh. Because wild sh tend to have longer residence times in estuary environments than hatchery sh [32], they are more likely to be affected by environmental contaminants. The fact that juvenile chinook from the Duwamish River had elevated mean VTG levels, whereas English sole from the Duwamish did not [6], could be attributed to juvenile chinook being more sensitive to VTG induction, having greater exposure to estrogenic compounds at this site, or experiencing estrogenic exposure before reaching this area. Differences in exposure are certainly possible, because the two species occupy different habitats; salmon would mainly be in the water column in shallow near-shore areas, whereas English sole would mainly be in deeper water in benthic areas consisting of ne-grained sediment. However, benthic organisms are typically more exposed to contaminants than are those in the water column. Also, sole spend limited time in some parts of the Duwamish River, where water salinity is low, and would tend to move up and down the river with the salt wedge [33]. The range of VTG levels and the magnitude of VTG increase in the current study were similar to those reported in a marine study on wild ounder in Japan [34] and in a freshwater study on brown trout in Denmark [35]. The range and magnitude were also similar to the lower end of those reported for wild ounder in the United Kingdom or the Netherlands [36,37] and in freshwater studies involving either wild roach [38] or caged rainbow trout [39] held for three to four weeks downstream of sewage efuent outfalls in the United Kingdom. Also, although some studies on rainbow trout have reported adverse reproductive effects associated with 10- to 30-fold VTG elevation [2,11], adverse reproductive effects also have been reported with only two to three fold VTG elevation, which are levels similar to those observed in the current study. For example, Hashimoto et al. [34] observed an increase in intersex and decrease in testicular growth in wild male ounder, and Bennetau-Pelissero et al. [10] observed a decrease in spermatocrit, sperm motility, spawning females, and viable eggs in immature rainbow trout exposed to the phytoestrogen genistein. Therefore, juvenile chinook with elevated VTG during a sensitive early life stage could experience delayed reproductive effects such as those observed in ounder or rainbow trout. All of the Puget Sound sites had some sh with plasma VTG levels that were above the highest values observed in laboratory control sh, suggesting potential xenoestrogen exposure. Future work will involve an assessment of plasma VTG levels in juvenile coho and chinook from additional urban, upstream, and nonurban sites, and more work is needed to actually identify and measure the estrogenic contaminants in water, sediment, and sh in Puget Sound. Vitellogenin measurement in these salmonids also will be useful in laboratory exposure studies designed to evaluate the estrogenic potential of these contaminants and their possible adverse effects on other reproductive parameters. In addition to studying juvenile sal-

monids, assessment of estrogenic exposure via plasma VTG measurement of sexually maturing salmonids also may be informative, because rainbow trout males at this stage have been shown to have increased levels of aneuploid sperm, and a corresponding decrease in embryonic survival, when exposed to an environmentally relevant concentration of the synthetic estrogen 17a-ethynylestradiol [40]. We also intend to assess the usefulness of this ELISA for measuring VTG in other Pacic salmonids, such as chum, pink, and sockeye.
AcknowledgementThe authors thank our colleagues at the National Oceanic and Atmospheric Administrations Northwest Fisheries Science Center, including Jon Dickey, Ron Johnson, and James Meador for providing us with coho and chinook plasma and serum samples, David Baldwin for help with graphs, and James Meador and Bich-Thuy Eberhart for reviewing this manuscript. REFERENCES 1. Tyler CR, Jobling S, Sumpter JP. 1998. Endocrine disruption in wildlife: A critical review of the evidence. Crit Rev Toxicol 28:319361. 2. Schwaiger J, Mallow U, Ferling H, Knoerr S, Braunbeck T, Kalbfus W, Negele RD. 2002. How estrogenic is nonylphenol? A transgenerational study using rainbow trout (Oncorhynchus mykiss) as a test organism. Aquat Toxicol 59:177189. 3. Orn S, Holbech H, Madsen TH, Norrgren L, Petersen GI. 2003. Gonad development and vitellogenin production in zebrash (Danio rerio) exposed to ethinylestradiol and methyltestosterone. Aquat Toxicol 65:397411. 4. Brion F, Tyler CR, Palazzi X, Laillet B, Porcher JM, Garric J, Flammarion P. 2004. Impacts of 17b-estradiol, including environmentally relevant concentrations, on reproduction after exposure during embryo-, larval-, juvenile- and adult-life stages in zebrash (Danio rerio). Aquat Toxicol 68:193217. 5. Deng X, Rempel MA, Armstrong J, Schlenk D. 2007. Seasonal evaluation of reproductive status and exposure to environmental estrogens in hornyhead turbot at the municipal wastewater outfall of Orange County, CA. Environ Toxicol 22:464471. 6. Johnson LL, Lomax DP, Myers MS, Olson OP, ONeill SM, West J, Collier TK. 2008. Xenoestrogen exposure and effects in English sole (Parophrys vetulus) from Puget Sound WA. Aquat Toxicol 88:2938. 7. Hinck JE, Schmitt CJ, Blazer VS, Denslow ND, Bartish TM, Anderson PJ, Coyle JJ, Dethloff FM, Tillit DE. 2006. Environmental contaminants and biomarker responses in sh from the Columbia River and its tributaries: Spatial and temporal trends. Sci Total Environ 366:549 578. 8. Good TP, Waples RS, Adams P, eds. 2005. Updated status of federally listed ESUs of West Coast salmon and steelhead. NMFS-NWFSC-66. NOAA Technical Memo. U.S. Department of Commerce, Seattle, WA. 9. King County. 2007. Survey of endocrine disruptors in King County surface waters. Water and Land Resources Division, Seattle, WA, USA. 10. Benetau-Pelissero C, Breton B, Bennetau B, Corraze G, Le Menn F, Davail-Cuisset B, Helou C, Kaushik SJ. 2001. Effect of genisteinenriched diets on the endocrine process of gametogenesis and on reproduction efciency of the rainbow trout Oncorhynchus mykiss. Gen Comp Endocrinol 121:173187. 11. Le Gac F, Thomas JL, Mourot B, Loir M. 2001. In vivo and in vitro effects of prochloraz and nonylphenol ethoxylates on trout spermatogenesis. Aquat Toxicol 53:187200. 12. Hara A, Sullivan CV, Dickhoff WW. 1993. Isolation and some characterization of vitellogenin and its related egg yolk proteins from coho salmon (Oncorhynchus kisutch). Zool Sci 10:245256. 13. Lomax DP, Roubal WT, Moore JD, Johnson LL. 1998. An enzymelinked immunosorbent assay (ELISA) for measuring vitellogenin in English sole (Pleuronectes vetulus): Development, validation and crossreactivity with other pleuronectids. Comp Biochem Physiol B 121:425 436. 14. Tyler CR, van Aerle R, Nilsen MV, Blackwell R, Maddix S, Nilsen BM, Berg K, Hutchinson TH, Goksyr A. 2002. Monoclonal antibody enzyme-linked immunosorbent assay to quantify vitellogenin for studies on environmental estrogens in the rainbow trout (Oncorhynchus mykiss). Environ Toxicol Chem 21:4754. 15. Brodeur JC, Woodburn KB, Zhang F, Bartels MJ, Klecka GM. 2006. Plasma sampling and freezing procedures inuence vitellogenin measurements by enzyme-linked immunoassay in the fathead minnow (Pimephales promelas). Environ Toxicol Chem 25:337348.

486

Environ. Toxicol. Chem. 30, 2011

K.A. Peck et al. 29. King County. 2007. Combined Sewer Overow Control Program. 2006 2007 Annual Report. Department of Natural Resources and Parks, Wastewater Treatment Division, Seattle, WA, USA. 30. Schultz IR, Orner G, Merdink JL, Skillman A. 2001. Dose-response relationships and pharmacokinetics of vitellogenin in rainbow trout after intravascular administration of 17a-ethynylestradiol. Aquat Toxicol 51:305318. 31. Arukwe A, Re K. 2008. Molecular and cellular detection of expression of vitellogenin and zona radiata protein in liver and skin of juvenile salmon (Salmo salar) exposed to nonylphenol. Cell Tissue Res 331:701 712. 32. Fresh KL, Casillas E, Johnson LL, Bottom DL. 2005. Role of the estuary in the recovery of Columbia River Basin salmon and steelhead: An evaluation of the effects of selected factors on salmonid population viability. NMFS-NWFSC-69. NOAA Technical Memo. U.S. Department of Commerce, Seattle, WA. 33. Lassuy DR. 1989. Species proles: Life histories and environmental requirements of coastal shes and invertebrates (Pacic Northwest) English sole. U.S. Fish Wildl Serv Biol Rep 82(11.101); TR EL- 824. U.S. Army Corps of Engineers, Corvallis, OR. 34. Hashimoto S, Bessho H, Nakamura M, Iguchi T, Fujita K. 2000. Elevated serum vitellogenin levels and gonadal abnormalities in wild male ounder (Pleuronectes yokohamae) from Tokyo Bay, Japan. Mar Environ Res 49:3753. 35. Bjerregaard P, Hansen PR, Larsen KJ, Erratico C, Korsgaard B, Holbech H. 2008. Vitellogenin as a biomarker for estrogenic effects in brown trout, Salmo trutta: Laboratory and eld investigations. Environ Toxicol Chem 27:23872396. 36. Kirby MF, Allen YT, Dyer RA, Feist SW, Katsiadaki I, Matthiessen P, Scott AP, Smith A, Stentiford GD, Thain JE, Thomas KV, Tolhurst L, Waldock MJ. 2004. Surveys of plasma vitellogenin and intersex in male ounder (Platichthys esus) as measures of endocrine disruption by estrogenic contamination in United Kingdom estuaries: Temporal trends, 1996 to 2001. Environ Toxicol Chem 23:748758. 37. Vethaak AD, Lahr J, Schrap SM, Belfroid AC, Rijs GBJ, Gerritsen A, de Boer J, Bulder AS, Grinwis GCM, Kuiper RV, Legler J, Murk TAJ, Peijnenburg W, Verhaar HJM, de Voogt P. 2005. An integrated assessment of estrogenic contamination and biological effects in the aquatic environment of the Netherlands. Chemosphere 59:511524. 38. Jobling S, Nolan M, Tyler CR, Brighty G, Sumpter JP. 1998. Widespread sexual disruption in wild sh. Environ Sci Technol 32:24982506. 39. Harries JE, Sheahan DA, Jobling S, Matthiessen P, Neall P, Routledge EJ, Rycroft R, Sumpter JP, Tylor T. 1996. A survey of estrogenic activity in United Kingdom inland waters. Environ Toxicol Chem 15:19932002. 40. Brown KH, Schultz IR, Cloud JG, Nagler JJ. 2008. Aneuploid sperm formation in rainbow trout exposed to the environmental estrogen 17aethynylestradiol. Proc Natl Acad Sci 105:1978619791.

16. Rodriguez JN, Kah O, Geffard M, Le Menn F. 1989. Enzyme-linked immunosorbent assay (ELISA) for sole (Solea vulgaris) vitellogenin. Comp Biochem Physiol B 92:741746. 17. Rodbard D, Lewald JE. 1970. Computer analysis of radioligand assay and radioimmunoassay data. Acta Endocrinol Suppl 147:79103. 18. Zar JH. 1984. Biostatistical Analysis. Prentice-Hall, Englewood Cliffs, NJ, USA. 19. Silversand C, Hyllner SJ, Haux C. 1993. Isolation, immunochemical detection, and observations of the instability of vitellogenin from four teleosts. J Exp Zool 267:587597. 20. Bon E, Barbe U, Rodriguez JN, Cuisset B, Pelissero C, Sumpter JP, Le Menn F. 1997. Plasma vitellogenin levels during the annual reproductive cycle of the female rainbow trout (Oncorhynchus mykiss): Establishment and validation of an ELISA. Comp Biochem Physiol B 117:7584. 21. Sherry J, Gamble A, Fielden M, Hodson P, Burnison B, Solomon K. 1999. An ELISA for brown trout (Salmo trutta) vitellogenin and its use in bioassays for environmental estrogens. Sci Total Environ 225: 1331. 22. Schafhauser-Smith D, Benfey TJ. 2002. The purication and development of a quantitative enzyme linked immunosorbent assay (ELISA) for the measurement of vitellogenin in diploid and triploid brook trout (Salvelinus fontinalis). Fish Physiol Biochem 24:287298. 23. Fenske M, van Aerle R, Brack S, Tyler CR, Segner H. 2001. Development and validation of a homologous zebrash (Danio rerio Hamilton-Buchanan) vitellogenin enzyme-linked immunosorbent assay (ELISA) and its application for studies on estrogenic chemicals. Comp Biochem Physiol C 129:217232. 24. Inaba K, Buerano CC, Natividad FF, Morisawa M. 1997. Degradation of vitellogenins by 170kDa trypsin-like protease in the plasma of the tilapia, Oreochromis niloticus. Comp Biochem Physiol 118:8590. 25. Fujita T, Fukada H, Shimizu M, Hiramatsu N, Hara A. 2004. Quantication of serum levels of precursors to vitelline envelope proteins (choriogenins) and vitellogenin in estrogen treated masu salmon, Oncorhynchus masou. Gen Comp Endocrin 136:4957. 26. Pedersen SN, Christiansen LB, Pedersen KL, Korsgaard B, Bjerregaard P. 1999. In vivo estrogenic activity of branched and linear alkylphenols in rainbow trout (Oncorhynchus mykiss). Sci Total Environ 233: 8996. 27. Lindholst C, Pedersen KL, Pedersen SN. 2000. Estrogenic response of bisphenol A in rainbow trout (Oncorhynchus mykiss). Aquat Toxicol 48:8794. 28. Ruggerone GT, Nelson TS, Hall J, Jeanes E. 2006. 2005 Juvenile Chinook Duwamish River Studies. Study 1: Habitat utilization, migration timing, growth, and diet of juvenile Chinook salmon in the Duwamish River and estuary. King Conservation District and Salmon Recovery Funding Board, Seattle, WA, USA.