Proteomic Profiles of White Sucker (Catostomus commersonii

Jan 17, 2012 - White sucker (Catostomus commersonii) sampled from the Thunder Bay Area of Concern were assessed for health using a shotgun approach to...
0 downloads 6 Views 814KB Size
Download PDF
Article pubs.acs.org/est

Proteomic Profiles of White Sucker (Catostomus commersonii) Sampled from within the Thunder Bay Area of Concern Reveal UpRegulation of Proteins Associated with Tumor Formation and Exposure to Environmental Estrogens Denina B. D. Simmons,*,† Niels C. Bols,† Bernard P. Duncker,† Mark McMaster,‡ Jason Miller,‡ and James P. Sherry‡ †

Department of Biology, University of Waterloo, 200 University Avenue West, Waterloo, ON, Canada Environment Canada, WSTD, 867 Lakeshore Road, Burlington, ON, Canada

‡

S Supporting Information *

ABSTRACT: White sucker (Catostomus commersonii) sampled from the Thunder Bay Area of Concern were assessed for health using a shotgun approach to compile proteomic profiles. Plasma proteins were sampled from male and female fish from a reference location, an area in recovery within Thunder Bay Harbour, and a site at the mouth of the Kaministiquia River where water and sediment quality has been degraded by industrial activities. The proteins were characterized using reverse-phase liquid chromatography tandem to a quadrupole-time-of-flight (LC-Q-TOF) mass spectrometer and were identified by searching in peptide databases. In total, 1086 unique proteins were identified. The identified proteins were then examined by means of a bioinformatics pathway analysis to gain insight into the biological functions and disease pathways that were represented and to assess whether there were any significant changes in protein expression due to sampling location. Female white sucker exhibited significant (p = 0.00183) site-specific changes in the number of plasma proteins that were related to tumor formation, reproductive system disease, and neurological disease. Male fish plasma had a significantly different (p < 0.0001) number of proteins related to neurological disease and tumor formation. Plasma concentrations of vitellogenin were significantly elevated in females from the Kaministiquia River compared to the Thunder Bay Harbour and reference sites. The protein expression profiles indicate that white sucker health has benefited from the remediation of the Thunder Bay Harbour site, whereas white sucker from the Kaministiquia River site are impacted by ongoing contaminant discharges.

■

INTRODUCTION Thunder Bay, the largest city in northwestern Ontario, is located on the northern shore of Lake Superior.1 The Great Lakes Water Quality Agreement has defined 39 regions as Areas of Concern (AOC) to be monitored by the International Joint Commission (IJC) (Annex 2 of the 1987 Protocol). Specifically, the Thunder Bay Area of Concern is listed as one of these locations due in part to declining fish populations, restrictions on fish consumption, loss of species health, reduced abundance and diversity, reduced recreational opportunities, and degradation of aesthetics. The area covers 28 km along the shoreline of Lake Superior, up to 9 km offshore from the City of Thunder Bay, the Kaministiquia River system, and a number of smaller rivers and creeks which drain the Thunder Bay watershed. Contaminated sediments containing polycyclic aromatic hydrocarbons, chlorophenols, dioxins, and furans were identified by the IJC as a major contributor to impaired © 2012 American Chemical Society

water quality in the Thunder Bay Harbour. Remediation efforts to clean the sediments and restore fish habitat in the Thunder Bay Harbour, known as Northern Wood Preservers Alternative Remediation Concept (NOWPARC) have been completed and are now being monitored. The incorporation of secondary treatment and 100% chlorine dioxide substitution at the Abiti Bowater pulp and paper mill, which discharges into the Kaministiquia River, have improved effluent quality and meet Lake Superior loading requirements.2 Municipal wastewater infrastructure is being improved and an additional remediation action is planned for the sediment in the Thunder Bay Harbour surrounding the site of the now closed Cascades Inc. mill Received: Revised: Accepted: Published: 1886

November 21, 2011 January 16, 2012 January 17, 2012 January 17, 2012 dx.doi.org/10.1021/es204131r | Environ. Sci. Technol. 2012, 46, 1886−1894

Environmental Science & Technology

Article

mill (48.368948, −89.249843, Kaministiquia River, Thunder Bay, ON), and a remediated location near a previously active paper mill (48.450881, −89.17622, Thunder Bay Harbour, Thunder Bay, ON) (Figure S1, Supporting Information) The Kaministiquia River and Thunder Bay Harbour sites lie within the Thunder Bay AOC. All fish were anaesthetized in a bath of clove oil and ethanol (10−15 drops of clove oil and 6 drops of ethanol in 10−15 L of water taken from the sampling location). Mass and fork length were measured and recorded. Blood was drawn from each fish by caudal puncture using a sterile, chilled, and heparinized 21-gauge needle and syringe, and then ejected after removal of the needle into a cold microcentrifuge tube which was then stored on ice for no longer than 1 h. Plasma was separated from the blood samples by centrifugation (9300g, 4 min) at 4 °C, transferred to cryogenic vials, flash frozen in liquid nitrogen, and subsequently stored at −80 °C prior to analysis. Gonads and liver were removed by dissection, and fork length (±0.1 cm), body weight (±1.0 g), liver and gonad weights (±0.01 g) were recorded. A 1-g ovarian sample was collected to calculate fecundity (number of eggs per fish). Sample Preparation. Plasma samples were thawed and transferred into a low-retention microcentrifuge tube so that the total volume of plasma was 15 μL. The plasma samples were then evaporated to near dryness using a centrifugal evaporator (Savant Instruments, Inc., model AES1010-120) for 10 min. The dry pellets were then resuspended in 50 μL of 50 mM tris-HCl (pH 7.8), followed by the addition of 2.65 μL of 100 mM TCEP in 50 mM tris-HCl (pH 7.8) to each tube. The tubes were then vortex mixed and left to incubate for 1 h at room temperature to reduce disulfide bonds. To acetylate cysteine residues, 2.8 μL of 200 mM IAM in 50 mM tris-HCl (pH 7.8) was added to each tube; the tube contents were vortex mixed and then left to incubate for another hour at room temperature. Fifty μL of 20% v/v formic acid was added to each tube, mixed thoroughly by vortex, and then left to digest on a heating block for 30 min at 115 °C. The resulting peptide mixtures in each tube were then evaporated to near dryness using centrifugal evaporation for approximately 45 min and then resuspended by vortex in 20 μL of 2% acetonitrile, 98% water, and 0.1% formic acid. Finally, the tubes were centrifuged for 10 min at 15 000g to remove debris, and the supernatant was transferred from each tube into glass chromatography vials containing 200-μL polypropylene conical vial inserts for subsequent analysis by liquid chromatography tandem mass spectrometry (LC-MS/MS). Sample Analysis. Peptides were separated by Agilent 1260 Infinity Binary LC on a Zorbax, 300SB-C18, 1.0 × 50 mm 3.5 μm column, with a thermostat controlled column temperature of 40 °C. The Agilent 6520 Accurate-Mass Quadrupole Timeof-Flight (Q-TOF) was used as the detector in tandem to the Agilent 1260 system. The pump timetable was 0−2 min 2% Solvent B, 2−22 min 2−40% solvent B, 22−27 min 40−60% solvent B, 27−32 min 90% solvent B, 32−50 min 2% solvent B (Solvent A; 5% acetonitrile 0.1% formic acid, and Solvent B; 95% acetonitrile 0.1% formic acid). For VTG quantification, an AQUA workflow was employed using the peptide K[Pro(13C5; 15N)]VYRRWLLD as an internal standard (97.21% purity, synthesized by Thermo Fisher Scientific, Ottawa, ON) to monitor the concentration of the endogenous VTG peptide KPVYRRWLLD. The internal VTG standard was spiked into all plasma digest samples. Endogenous VTG peptide molar concentrations were then calculated as a ratio to the internal standard’s molar concentration. White sucker VTG has yet to

(formerly Provincial Papers).2 The white sucker (Catostomus commersonii), a benthic fish indigenous to North America and found commonly throughout the Great Lakes,3 has been used for environmental effects monitoring (EEM) on Lake Superior for decades.4 The incidence of liver tumors and deformities in white sucker from Thunder Bay was a primary consideration when the AOC’s remedial action plan was formed. After the completion of NOWPARC it appeared as though the incidence of liver tumors in the white sucker population was reduced below 5%, but still required further assessment.2 Shotgun proteomics (also known as bottom-up proteomics) is a method of identifying proteins in complex biological samples.5 In shotgun proteomics, the proteins in a biological sample are digested and the resulting peptides are separated by liquid chromatography and characterized using tandem mass spectrometry. By comparing the masses of the peptides and their tandem mass spectra with those predicted from a sequence database or found annotated within a peptide spectral library, peptides can be identified and matched to a known protein. Absolute quantification (AQUA) of proteins can also be performed simultaneously by using synthesized peptides with incorporated stable isotopes as ideal internal standards to mimic endogenous peptides formed by protein digestion.6,7 The AQUA internal standard peptides are then used to accurately and precisely measure the absolute levels of the targeted endogenous protein through a selected reaction monitoring analysis. To our knowledge, the present study is the first to employ shotgun proteomic profiling of white sucker. To illustrate the power of shotgun proteomic profiling in environmental toxicology, an analysis was performed on the plasma proteins of adult white sucker, which had been sampled from the Thunder Bay AOC for the purpose of assessing the health of the parent population. An AQUA assay was developed for vitellogenin (VTG), which is an established biomarker that is a measure of the reproductive health of the female fish and of the potential exposure of males to environmental estrogens. Therefore, the overall objective of this study was to assess the utility of proteomic profiling for biomonitoring wild fish and to use this approach to develop methods for the quantitative assay of important biomarker proteins, such as VTG.

■

MATERIALS AND METHODS Tris(2-carboxyethyl)phosphine hydrochloride solution (TCEP) (0.5 M, pH 7.0), iodoacetamide (IAM) (≥99%), formic acid (puriss. p.a. ≥98%), Heparin (Ammonium salt), and HPLC peptide standard mixture (H2016-1VL) were purchased from Sigma-Aldrich Canada Ltd. (Oakville, ON). Tris(hydroxymethyl)aminomethane (tris) (electrophoresis grade) was purchased from Bio-Rad Laboratories (Canada) Ltd. (Mississauga, ON). Bovine serum albumin tryptic digest standard, chromatography column, column filter, column fittings, and instrumental calibration and reference solutions were purchased from Agilent Technologies Canada Inc. (Mississauga, ON). Aura Cocia Brand Clove Bud (100% pure essential oil) was purchased from a local grocery. 85% Ethanol was purchased from Caledon Laboratories Ltd. (Georgetown, ON) and HPLC-grade water and optima-grade acetonitrile (ACN) were purchased from Thermo Fisher Scientific Company (Ottawa, ON). Field Sampling and Plasma Collection. Sexually mature white sucker were sampled by gill nets from a reference location (48.918479, −87.767976, Little Gravel Creek, Terrace Bay, ON), a location downstream of an active pulp and paper 1887

dx.doi.org/10.1021/es204131r | Environ. Sci. Technol. 2012, 46, 1886−1894

Environmental Science & Technology

Article

Figure 1. Distribution of proteins found in white sucker from the Thunder Bay Harbour (TBH), the Kaministiquia River (Kam), and Little Gravel Creek (Gravel). (a) Validated protein distribution among site locations, (b) functional distribution of proteins in male white sucker, (c) functional distribution of proteins in female white sucker, (d) significant functional overlap, (e) number of fish expressing overlapping proteins from (d) by site location and sex.

be fully sequenced so the exact molar mass is unknown; molar concentrations of VTG were converted to mass concentrations based on the molecular mass of vtg1 from Danio rerio (accession Q8JH37). Each analytical run included a blank, peptide standard, and BSA digest standard injection every 8 samples in order to monitor baseline, carry-over, drift, and sensitivity during the runtime. Reference mass correction was enabled using the dual-ESI reference nebulizer for m/z 121.050873 and m/z 922.009798 (method variables for the separation and instrumental detection are summarized in Table S1, Supporting Information). Sodium-dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) with silver staining was conducted according to the guidelines in Current Protocols in Cell Biology8 for each white sucker with a dilution of 1:10 for male and 1:100 for female plasma. A protein standard ladder (Novex Sharp Unstained Protein Standard, Invitrogen Canada Inc., Burlington, ON) was separated in a separate lane on each gel for calibration purposes. Data Analysis. Both centroid and profile mass spectral data files were collected using Mass Hunter Data Acquisition Software (Version B.02.00). Data files for each fish were pooled

into groups by sex and location, and each group was analyzed separately. Spectrum Mill Software (Version A.03.03 SR4) was used to extract good quality spectra, filter noisy and poor quality spectra from raw data files, sequence peptides, and search protein databases. All Spectrum Mill MS/MS search settings were left to the manufacturer’s default except that spectra were searched using reversed database scores and dynamic peak thresholding, precursor mass tolerance was ±20 ppm, and product mass tolerance was ±50 ppm. Mixed acetamide (C), oxidized methionine, pyroglutamic acid (Nterm Q), deamidated (N), phosphorylated (S, T, and Y), and ubiquitination-GG (K) were specified as variable modifications. Proteins were searched using a species subset database (Ameirus nebulosus, Caenorhabditis elegans, Carassius auratus, Catostomus commersonii, Cyprinus carpio, Danio rerio, Gadus morhua, Gambusia af f inis, Morone americana, Mya arenaria, Mytilus edulis, Notprois atherinoides, Onchorhynchus mykiss, Oreochromis niloticus, Oryzias latipes, Perca f lavescens, Pimephales promelas, Salmo salar, and Xenopus laevis) within the NCBInr and Swissprot databases. Proteins were validated manually and accepted when they matched multiple members of a protein 1888

dx.doi.org/10.1021/es204131r | Environ. Sci. Technol. 2012, 46, 1886−1894

Environmental Science & Technology

Article

Figure 2. Differential expression profiles of plasma proteins from white sucker at the Thunder Bay Harbour (TBH), the Kaministiquia River (Kam), and Little Gravel Creek (Gravel). (a) Mean spectral intensities of proteins with significant functional overlap related to tumorigenesis, neurological disease, and reproductive system disease; (b) number of fish with differentially expressed proteins; (c) mean peptide spectral intensities of proteins with significant differential expression. Bars represent standard error, * indicates significantly increased expression (p < 0.05), and ** indicates significantly decreased expression (p < 0.05).

calculate a p-value determining the probability that each biological function and/or disease assigned to that data set is due to chance alone. Additional tests of significance were performed using GraphPad Prism 4, Version 4.03 (GraphPad Software, Inc.)

family or if they had a score greater than 5 and a minimum of one peptide with a %SPI of greater than 60% (as recommended for Agilent Q-TOF data). Mean spectral intensity values were calculated by the Spectrum Mill Software and were defined as the mean intensity of all peptides assigned to that protein. Peptide intensities were calculated from extracted ion chromatograms (EIC), or the sum of the precursor m/z abundance from the MS scans (∼ chromatographic peak area). Agilent Mass Hunter Quantitative Analysis Software (Version B.01.04, Build 1.4.126.2) was used to quantify VTG peptides using batch analysis. Product ion transitions used for quantification were 449.254 → 609.342 and 451.260 → 612.350 for the endogenous and labeled AQUA peptides, respectively. Densitometry scans of SDS-PAGE gels were analyzed by means of Fluorochem version 2.01 (Alpha Innotech Corporation). Valid protein hits were analyzed through the use of Ingenuity Pathways Analysis (Ingenuity Systems, www.ingenuity.com). Pooled data for each treatment with corresponding gene identifiers and expression values (mean spectral intensity values) were uploaded into the application. Each identifier was mapped to its corresponding object in Ingenuity’s Knowledge Base. These molecules, called network eligible molecules, were overlaid onto a global molecular network developed from information contained in Ingenuity’s Knowledge Base and then algorithmically generated based on their connectivity. Functional analysis identified the biological functions and/or diseases that were most significant to the data set. The right-tailed Fisher’s exact test was used to

■

RESULTS The percent yield of validated spectra for each group of replicates ranged from 4.1 to 6.1% of total filtered spectra (Table S2, Supporting Information). The mean number of validated proteins for each replicate within the six pooled searches ranged from 45 to 67 whereas the number of unique validated proteins for each pooled search ranged from 132 to 219 (Table S2, Supporting Information). Thus, the number of validated proteins in each replicate was about 30% of the protein diversity of its parent group. Of the 1086 unique proteins that were validated for all fish from the pooled sites, a total of 799 were assignable to a gene identifier that was eligible for mapping within Ingenuity’s Knowledge Base. At the time of this study, there were only 12 reviewed protein entries in the UniProt database and 119 protein entries in the NCBI protein database for C. commersonii, therefore the majority of the identified proteins were matched to fish species other than white sucker. It is likely that identification yields will improve in the future as the annotation of protein sequences and gene identifiers in available databases progresses for teleost species.9 The total number of proteins identified in females increased significantly from Little Gravel Creek, to Thunder Bay Harbour and then to the Kaministiquia River site (Chi-square test for 1889

dx.doi.org/10.1021/es204131r | Environ. Sci. Technol. 2012, 46, 1886−1894

Environmental Science & Technology

Article

Figure 3. (a) Mean plasma vitellogenin concentrations in female white sucker as determined by AQUA. (b) Mean plasma vitellogenin as percent of total protein in female white sucker as determined by SDS-PAGE. (c) Mean fecundity in female white sucker. (d) Mean gonadal somatic indices of white sucker in males and females. Thunder Bay Harbour (TBH, Stripe), the Kaministiquia River (Kam, Black), and Little Gravel Creek (Gravel, White). Bars represent standard error and the asterisks denote significant differences (p < 0.0001).

trend, p = 0.0113). This trend was even more pronounced for “unique to site” proteins for females, (Chi-square test for trend, p = 0.0018) but not for male fish; suggesting that location had a greater affect on protein expression in female fish (Figure 1a). Functional analysis by Ingenuity’s Knowledge Base identified 11 categories of biological function and disease among the data set that were considered significant (Fisher’s exact test, α ≤ 0.01, see Table S3, Supporting Information, for p-values), many of which also demonstrated apparent differential expression patterns between sex and among sites (Figure 1b and c). The highest numbers of mapped proteins for the entire data set were related to the following diseases: genetic disorder > gastrointestinal disease > cancer (specifically tumorigenesis) > immunological disease> neurological disease in rank order without implied significance. Notably, the number of detected proteins associated with neurological disease was significantly greater in males from the Thunder Bay Harbour and the Kaministiquia River sites than at Little Gravel Creek (Chi square test for trend, p < 0.0001). Additionally, there was significant functional overlap among the proteins associated with reproductive system disease, neurological disease, and cancer (Fisher’s exact test, p < 0.0001). Of the 75 proteins present in the data set that were associated with reproductive disease, 52% were also associated with tumorigenesis, and 45% were associated with both tumorigenesis and neurological disease (Figure 1d). Among the proteins with functional overlap among reproductive system disease, neurological disease, and tumorigenesis, there was significantly more expression in females from the Kaministiquia River and the Thunder Bay Harbour than from Little Gravel Creek (Chisquare test for trend, p = 0.0423) (Figure 1e). The mean peptide spectral intensities of differentially expressed proteins which were relevant to reproductive system disease, tumorigenesis, and/or neurological disease are

presented in Figure 2a. The chi-square test and two-way ANOVA (α = 0.05) were performed on the proteins Reticulon1 (RTN-1), Mast/stem cell growth factor receptor (KIT), Aromatase (CYP19A1), Tumor suppressor p53-binding protein 1 (TP53BP1), and E3 ubiquitin-protein ligase (HUWE1) to test for significant differences in the number of individual fish expressing the proteins and the mean relative plasma protein expression (using mean and standard error of peptide spectral intensity values from the individuals within a group). There was a significant increase in the number of individuals that expressed RTN-1 and TP53BP1 and a significant decrease in the number of individuals that expressed HUWE1 in females sampled from the Thunder Bay Harbour compared to Little Gravel Creek (Chi-square test, prtn‑1 = 0.0010, ptp53bp1 = 0.0391, phuwe1 = 0.0118) (Figure 2b). There was also a significant increase in the number of individuals that expressed TP53BP1 and a significant decrease in the number of individuals that expressed HUWE1 in males sampled from the Kaministiquia River and the Thunder Bay Harbour compared to Little Gravel Creek (Chi-square test, ptp53bp1 = 0.0019, phuwe1 = 0.0055) (Figure 2b). Figure 2c displays the mean relative plasma protein expression in male and female fish for RTN-1, KIT, CYP19A1, TP53BP1, and HUWE1 at all locations. Comparing male plasma from the Thunder Bay Harbour with the Kaministiquia River, RTN-1 expression was significantly increased and KIT expression was significantly depressed, CYP19A1 was undetectable in males from the Thunder Bay Harbour, HUWE1 was undetectable in males from the Kaministiquia River, and there were no significant differences in the expression of TP53BP1 in males between locations. HUWE1 expression was increased significantly in the plasma of females from Little Gravel Creek compared to the Thunder Bay Harbour and was undetectable in females from the Kaministiquia River. TP53BP1 plasma expression was signifi1890

dx.doi.org/10.1021/es204131r | Environ. Sci. Technol. 2012, 46, 1886−1894

Environmental Science & Technology

Article

leukemia cell differentiation protein (Mcl-1) are inactive.13 Thus, the differential expression pattern of CDK2, KIT, and HUWE1 suggest that tumorigenesis is occurring in fish from the Kaministiquia River. In contrast, the up-regulation of LYN (resulting in cell cycle arrest), TOP2B (involved in doublestranded DNA break/repair), and TP53BP1 (enhances TP53mediated transcriptional activation) in fish sampled from the Thunder Bay Harbour suggests cell-cycle repair mechanisms are being regulated by TP53 activation,9,10,14,15 supporting the observations of a 2006 study that fish tumors are declining at that site.2 Neurological Function of Fish. Recent studies have observed neuroendocrine effects in fish that have been exposed to pulp and paper mill effluents,16−18 but a clear mechanistic link was not found using this initial application of shotgun proteomics. Evidence from gene expression arrays and receptor/enzyme binding assays suggest that the polar components of pulp and paper mill effluents act upon the dopamine system in a manner that could interfere with luteinizing hormone (LH) release.16−18 In teleosts, LH is produced by gonadotrophs in the anterior pituitary regulating gonadal growth cycles, sex steroid and hormone synthesis, and triggers ovulation in females and sperm production in males.19 In the present study a significant increase in proteins associated with neurological disease was observed for male fish in the Thunder Bay Area of Concern. However, no proteins were identified in plasmas of white sucker from Thunder Bay that have direct interaction with monoamine oxidase (MAO), glutamic acid decarboxylase (GAD), or gamma-aminobutyric acid (GABA). Nor were other proteins that are involved in the release of LH identified. Conspicuously, the expression of neuroendocrine-specific protein (RTN-1) was significantly higher in male fish from the Thunder Bay Harbour compared to both Little Gravel Creek and Kaministiquia River. RTN-1 is correlated with neuronal differentiation and is thought to be involved in neuroendocrine secretion and membrane trafficking in neuroendocrine cells.20 Other differentially expressed proteins related to neurological function were Dystrobrevin alpha (DTNA) in female fish from Thunder Bay Harbour and Rho-associated protein kinase 2 (ROCK2) in female fish from Kaministiquia River. DTNA is thought to be involved in synapse formation and stability, and in the clustering of nicotinic acetylcholine receptors,21 while ROCK2 is a protein kinase that plays a critical role in the regulation of spine and synaptic properties in the hippocampus.22 Although a clear mechanism of neuroendocrine disruption was not apparent from the examination of networked pathways generated from functional analyses by the Ingenuity software, a link to the observed changes in expression of RTN-1, DTNA, and ROCK2 to a specific function may become possible in the future as knowledge of the fish neuroendocrine system is increased. Reproductive Health of Fish. Vitellogenin (VTG) protein is produced in oviparous organisms and is typically found in high concentrations in the plasma of ovulating female fish. Increased plasma VTG concentrations in fish, particularly males, is considered an indicator that fish have been exposed to endocrine disrupting chemicals. Plasma VTG concentrations were significantly elevated in the female fish from the Kaministiquia River compared to Little Gravel Creek and TBH, but there were no significant differences between the plasma VTG concentrations of females from the Thunder Bay Harbour and Little Gravel Creek. Previous studies assessing the reproductive health of pulp and paper effluent exposed fish in

cantly greater in females at the Thunder Bay Harbour and was undetectable in females from Little Gravel Creek. KIT and RTN-1 were undetectable in plasma of females from Little Gravel Creek and there were no significant differences due to location between the expression of CYP19A1 in females. The calculated method detection limit for the absolute quantification (AQUA) of VTG in white sucker plasma was 8.8 fmol/μL (1.1 μg/mL). Measured concentrations of plasma VTG in females from all sampling locations ranged from 2.488 to 21.430 μg/mL, whereas plasma vitellogenin was not detected in male fish from any of the sampling locations. Mean plasma vitellogenin concentration in female white sucker were significantly greater from the Kaministiquia River than from the Thunder Bay Harbour and Little Gravel Creek (one-way ANOVA, Tukey’s multiple comparison test, p < 0.05) (Figure 3a). The endogenous and internal standard peptides fragmented at two peptide bonds, consistently resulting in four product ions: b1 (m/z = 129.10), y92+ (m/z = 609.35; 612.35), y93+ (m/z = 407.25; 409.25), and y82+ (m/z = 561.15) (see Figure S2, Supporting Information for example chromatograms, spectra, and linear range). The average plasma VTG in female white sucker from each site as determined by AQUA followed a trend similar to the average percent plasma vitellogenin values obtained by SDS-PAGE (ranging from 11.76 to 12.91%, Figure 3b), although the intersite differences were not significant for the SDS-PAGE data (ANOVA, p > 0.05). Fecundity followed a significant linear trend by location and was greatest in females from the Kaministiquia River, then Thunder Bay Harbour, and then Little Gravel Creek (ANOVA, post-test for linear trend, p < 0.0001) (Figure 3c). The gonadosomatic index (GSI = gonad mass ÷ body mass ×100) was significantly greater in females than males (two-way ANOVA, Bonferroni post-test, p < 0.0001), however there were no significant differences between sites (Figure 3d).

■

DISCUSSION Tumorigenesis in Fish. The observed increase of cancerrelated proteins that were specifically attributed to tumorigenesis in fish sampled from the Thunder Bay Area of Concern relative to Little Gravel Creek suggests that tumors are still occurring in wild fish. However, in the fish from the Thunder Bay Harbour (NOWPARC area in recovery) there were less proteins related to tumorigenesis. Namely, the differentially expressed proteins DNA topoisomerase 2-beta (TOP2B), tumor suppressor p53-binding protein 1 (TP53BP1), mast/ stem cell growth factor receptor (KIT), cyclin-dependent kinase 2 (CDK2), E3 ubiquitin-protein ligase (HUWE1), and tyrosine-protein kinase (LYN) point to tumor formation and suppression mechanisms that involve the cellular tumor antigen p53 (TP53) (Figure S3, Supporting Information, describes the interprotein relationships). LYN controls the cell cycle by phosphorylating CDK2 at tyr-15, thus inactivating the enzyme and causing cell cycle arrest.9,10 LYN also phosphorylates KIT.11 CDK2 is a protein kinase that is essential for cell cycle G1/S phase transition. TP53 reduces CDK2 activity through p21, also resulting in cell-cycle arrest. However, CDK2 protein expression was up-regulated in fish from the Kaministiquia river site, implying that cell proliferation was uncontrolled by LYN or TP53. 12 Furthermore, increased expression of the oncoprotein KIT in fish from the Kaministiquia River also suggests increased cell proliferation.11 The simultaneous downregulation of HUWE1 in fish from the Kaministiquia River indicates that apoptosis mechanisms involving the myeloid 1891

dx.doi.org/10.1021/es204131r | Environ. Sci. Technol. 2012, 46, 1886−1894

Environmental Science & Technology

Article

of detection limit is not considered as relevant to modern mass spectrometer instruments as the method detection limit (MDL, defined as the minimum concentration of a substance that can be measured and reported with 99% confidence to be greater than zero). MDL is a more accurate reflection of sensitivity and precision because it accounts for variation in sample matrix, preparation method, and instrumental noise. At this time ELISAs for VTG have lower MDLs than AQUA methods, however improvements in sample preparation and instrumentation will likely lead to a narrower performance gap between the methods in coming years. The use of AQUA for the determination of white sucker VTG was highly effective, comparable in sensitivity to alternative SDS-PAGE methods available, and offered specificity similar to ELISA methods, but with the advantage of not requiring a priori knowledge of biological species. Utility of Shotgun Proteomics for the Biomonitoring of Wild Fish Populations. Although the model organism, zebrafish, has been the subject of proteomic studies,27,28 ours is the first application of shotgun proteomic profiling in the environmental toxicology of a wild fish species and is also the first to employ shotgun proteomic profiling on the plasma of the white sucker. The laboratory sample preparation and instrumental analysis of all plasma samples was completed within 72 h, costing at least 5-fold less per sample than comparable gel-based methods (not including hardware investment), thus demonstrating the competitive time and cost efficiency of the shotgun approach. Regardless of the apparently low % yield of validated spectra for white sucker plasma samples, the identification of 1088 proteins was favorably comparable with other studies that used similar detection methods and involved the fully sequenced zebrafish: the number of identifications ranged from 293 to 5716.27 The shotgun approach in whole proteome analysis is considered more likely to produce a greater number of protein identifications and can more easily detect low abundance proteins than gel-based and top-down methods.27,29 Thus, shotgun methods could be a better approach to use for organisms that are not yet fully sequenced and annotated. A disadvantage of the shotgun approach, however, is that for every protein present in a sample to be accurately identified and quantified, the assumption is that each protein is fully digested and that the ratio of a peptide to its parent protein is 1:1, which means that information is invariably lost in a bottom-up approach.7 Additionally, when working with a species that is underrepresented in sequence databases (e.g., white sucker), search and validation strategies are adjusted to account for less sequence coverage and have a greater reliance on homology to other organisms, thus increasing the possibility of a greater number of false positives and of failed matches. Consequently it is important to prefilter MS/MS data to remove nonpeptide fragmentation spectra and to apply postsearch criteria that ensure correct protein validation and high stringency matches.27 To increase sequence coverage we searched more than one protein database for matches to a subset of teleost fish species. To gain greater confidence in protein identifications we pooled samples from the same groups during spectral filtering and database searching and used several criteria filters. The use of shotgun proteomic profiling with the simultaneous monitoring of key biomarker proteins (such as the accurate quantitative analysis of VTG) appears be well-suited for assessing the health of wild fish populations and could be applied to future monitoring studies within the Great Lakes

Canada have reported increased VTG concentrations and egg production, however these effects appear to be specific to the type of mill effluent.23 Our results suggest that effluents being released into the Kaministiquia River may contain endocrine disrupting chemicals that are affecting the production of VTG and egg production, but not significantly impacting GSI.24 The production of VTG by white sucker has been considered a less sensitive indicator of endocrine disrupting substances than for other wild fish species, and in general white sucker populations have been reported to be insensitive to the impacts of environmental estrogens,25 which may explain why male plasma did not contain detectable amounts of VTG. Because both plasma VTG and egg production were elevated at the Kaministiquia River in white sucker females, it is plausible that egg production in other fish populations might also be affected. Therefore further investigation into the reproductive success of other wild fish species from the Kaministiquia River is warranted. Enzyme-linked immune-assay (ELISA) methods are currently the mainstay for the quantification of VTG in fish. However at the start of our assessment of fish health at AOCs on the Canadian side of the Great Lakes we knew that the wild fish species at the various AOCs would vary among locations. The development of an accurate ELISA for the quantification of VTG requires the generation of species-specific antibodies and a considerable time investment which would be justified if the species occurred at multiple AOCs or at AOCs that were impacted by environmental estrogens. For those reasons, we did not develop an ELISA for white sucker VTG. Thus, the use of mass spectrometry for the quantitative analysis of white sucker VTG was an attractive alternative method because it measures the protein directly, overcoming the problems associated with antibody binding and specificity, and the method could be developed without prior knowledge of biological species with a relatively small time investment. While we were not able to directly compare our quantification for white sucker VTG by LC-MS/MS to an established ELISA, we have compared our results to the relative VTG abundance values determined by scanning densiometric analysis of the protein bands within a SDS-PAGE lane8 (Figure 3a and b). SDS-PAGE with silver staining achieved a detection limit of 0.7 ng of purified rainbow trout VTG per lane (≡ 0.7 μg/mL). The SDS-PAGE provided a slightly better estimated sensitivity (MDL ≅ 3.82 fmol/μL based on the MW of Rainbow Trout VTG) for white sucker plasma VTG than the AQUA method (MDL = 8.8 fmol/μL). However, the SDS-PAGE method could be less accurate compared to AQUA because it is a nonspecific protein assay and may include other proteins that are not VTG, which could explain why there were greater amounts of VTG when quantified by SDS-PAGE in females at all sampling locations. An AQUA method nearly identical to the one used in the present study was developed for the measurement of fathead minnow VTG by Zhang et al.26 which was highly specific, linear, and was reported to have performance criteria comparable to the detection of fathead minnow VTG by ELISA. However, Zhang et al.26 based their estimated detection limit on the mass of the target peptide and not on the mass of its parent VTG molecule, which is the moiety measured by ELISA. Thus, their comparison was biased in favor of the AQUA method. Our method’s performance exceeded that of Zhang et al.26 by 4-fold, in that we achieved a peptide detection limit (defined as signal-to-noise ratio of 3 to 1) of 0.5 μg/mL versus 2 μg/mL. However, that determination 1892

dx.doi.org/10.1021/es204131r | Environ. Sci. Technol. 2012, 46, 1886−1894

Environmental Science & Technology

Article

(10) Grimmler, M.; Wang, Y.; Mund, T.; Cilensek, Z.; Keidel, E. M.; Waddell, M. B.; Jäkel, H.; Kullmann, M.; Kriwacki, R. W.; Hengst, L. Cdk-inhibitory activity and stability of p27Kip1 are directly regulated by oncogenic tyrosine kinases. Cell 2007, 128 (2), 269−80. (11) Pisters, P. W. T.; O’Sullivan, B. Soft-Tissue Sarcomas. In HollandFrei Cancer Medicine; Kufe, D. W., Pollock, R. E., Weichselbaum, R. R., Bast, R. C. Jr., Gansler, T. S., Holland, J. F., Frei, E., Eds; BC Decker: Hamilton, 2003. (12) National Center for Biotechnology Information. The p53 tumor suppressor protein; Genes and Disease; http://www.ncbi.nlm.nih.gov/ books/NBK22268/. (13) UniProtKB/Swiss-Prot. Q7Z6Z7 (HUWE1_HUMAN); http:// www.uniprot.org/uniprot/Q7Z6Z7. (14) UniProtKB/Swiss-Prot. Q02880 (TOP2B_HUMAN); http:// www.uniprot.org/uniprot/Q02880. (15) UniProtKB/Swiss-Prot. Q12888 (TP53B_HUMAN); http:// www.uniprot.org/uniprot/Q12888. (16) Basu, N.; Ta, C. A.; Waye, A.; Mao, J. Q.; Hewitt, M.; Arnason, J. T.; Trudeau, V. L. Pulp and Paper Mill Effluents Contain Neuroactive Substances That Potentially Disrupt Neuroendocrine Control of Fish Reproduction. Environ. Sci. Technol. 2009, 43 (5), 1635−1641. (17) Popesku, J. T.; Tan, E. Y. Z.; Martel, P. H.; Kovacs, T. G.; Rowan-Carroll, A.; Williams, A.; Yauk, C.; Trudeau, V. L. Gene expression profiling of the fathead minnow (Pimephales promelas) neuroendocrine brain in response to pulp and paper mill effluents. Aquat. Toxicol. 2010, 99 (3), 379−388. (18) Milestone, C. B.; Orrego, R.; Scott, P. D.; Waye, A.; Kohli, J.; O’Connor, B. I.; Smith, B.; Engelhardt, H. E.; Smith, D. S.; Servos, M. R.; MacLatchy, D. L.; Trudeau, V. L.; Arnason, J. T.; Kovacs, T.; Furley, T. H.; Slade, A. H.; Holdway, D. A.; Hewitt, L. M. A comparison of effluents and wood feedstocks from pulp and paper mills in Brazil, Canada, and New Zealand to affect fish reproduction: Chemical profiling and in vitro assessments. Environ. Sci. Technol. 2011, DOI: 10.1021/es203382c. (19) Popesku, J. T.; Martyniuk, C. J.; Mennigen, J.; Xiong, H.; Zhang, D.; Xia, X.; Cossins, A. R.; Trudeau, V. L. The goldfish (Carassius auratus) as a model for neuroendocrine signaling. Mol. Cell. Endocrinol. 2008, 293, 43−56. (20) UniProtKB/Swiss-Prot. Q16799 (RTN1_HUMAN); http:// www.uniprot.org/uniprot/Q16799. (21) UniProtKB/Swiss-Prot. Q9Y4J8 (DTNA_HUMAN); http:// www.uniprot.org/uniprot/Q9Y4J8. (22) UniProtKB/Swiss-Prot. O75116 (ROCK2_HUMAN); http:// www.uniprot.org/uniprot/O75116. (23) Parrott, J. L.; McMaster, M. E.; Hewitt, L. M. A decade of research on the environmental impacts of pulp and paper mill effluents in Canada: Development and application of fish bioassays. J. Toxicol. Environ. Health, Part B 2006, 9 (4), 297−317. (24) Bosker, T.; Munkittrick, K. R.; MacLatchy, D. L. Challenges and opportunities with the use of biomarkers to predict reproductive impairment in fishes exposed to endocrine disrupting substances. Aquat. Toxicol 2010, 100 (1), 9−16. (25) Palace, V. P.; Evans, R. E.; Wautier, K. G.; Mills, K. H.; Blanchfield, P. J.; Park, B. J.; Baron, C. L.; Kidd, K. A. Interspecies differences in biochemical, histopathological, and population responses in four wild fish species exposed to ethynylestradiol added to a whole lake. Can. J. Fish. Aquat. Sci. 2009, 66 (11), 1920−1935. (26) Zhang, F.; Bartels, M. J.; Brodeur, J. C.; Woodburn, K. B. Quantitative measurement of fathead minnow vitellogenin by liquid chromatography combined with tandem mass spectrometry using a signature peptide of vitellogenin. Environ. Toxicol. Chem. 2004, 23 (6), 1408−1415. (27) Forne, I.; Abian, J.; Cerda, J. Fish proteome analysis: Model organisms and non-sequenced species. Proteomics 2010, 10 (4), 858− 872. (28) Martyniuk, C. J.; Denslow, N. D. Towards functional genomics in fish using quantitative proteomics. Gen. Comp. Endrocrinol. 2009, 164 (2−3), 135−141.

Areas of Concern to determine the response of fish to remediation efforts in areas that are in recovery. The approach provided insights into tumorigenesis, neurological function, and reproductive health by demonstrating significant relationships between location and the functional analysis of plasma proteins from white sucker sampled from the Thunder Bay Area of Concern. Those data and relationships support decades of previous research which established that fish health has been negatively impacted by environmental degradation of the Thunder Bay Region. By investigating the differential expression of proteins, we suggest that the health of wild fish populations in the Thunder Bay Harbour have improved due to the NOWPARC remedial action plan, but that the health of fish in the Kaministquia River remains stressed.

■

ASSOCIATED CONTENT

S Supporting Information *

Tables S1−S3 and Figures S1−S3. This information is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Maria Villela, Gerald Treteault, and Lisa Heikilla for their field sampling and laboratory assistance. We thank Great Lakes Action Plan and Great Lakes Ecosystem Initiative for funding this research and the National Engineering and Sciences Research Council for their Post-Doctoral Fellowship to D.S.

■

REFERENCES

(1) City of Thunder Bay Official Website; http://www.thunderbay.ca/ Home.htm. (2) Thunder Bay Area of Concern - Status of Beneficial Use Impairments; En164-22/15-2011E-PDF; Environment Canada and the Ontario Ministry of the Environment, 2011. (3) Scott, W. B. Crossman, E.J., Freshwater Fishes of Canada; Bulletin 184; Fisheries Research Board of Canada: Ottawa, ON, 1973. (4) Bowron, L. K.; Munkittrick, K. R.; McMaster, M. E.; Tetreault, G.; Hewitt, L. M. Responses of white sucker (Catostomus commersoni) to 20 years of process and waste treatment changes at a bleached kraft pulp mill, and to mill shutdown. Aquat. Toxicol. 2009, 95 (2), 117− 132. (5) Wu, C. C.; MacCoss, M. J. Shotgun proteomics: Tools for the analysis of complex biological systems. Curr. Opin. Mol. Ther. 2002, 4 (3), 242−250. (6) Pan, S.; Aebersold, R.; Chen, R.; Rush, J.; Goodlett, D. R.; McIntosh, M. W.; Zhang, J.; Brentnall, T. A. Mass Spectrometry Based Targeted Protein Quantification: Methods and Applications. J. Proteome Res. 2009, 8 (2), 787−797. (7) Elliott, M. H.; Smith, D. S.; Parker, C. E.; Borchers, C. Current trends in quantitative proteomics. J. Mass Spectrom. 2009, 44 (12), 1637−1660. (8) Dell’Angelica, E. C. J. S. Bonifacino. Staining Proteins in Gels. In Current Protocols in Cell Biology; Bonifacino, J. S., Dasso, M., Harford, J. B., Lippincott-Schwartz, J., Yamada, K. M., Eds; John Wiley and Sons, Inc: New York, 2001; Unit 6.6. (9) Yuan, Z.; Huang, Y.; Kraeft, S.; Chen, L.; Kharbanda, S.; Kufe, D. Interaction of cyclin-dependent kinase 2 and the Lyn tyrosine kinase in cells treated with 1-beta-D-arabinofuranosylcytosine. Oncogene 1996, 13 (5), 939−46. 1893

dx.doi.org/10.1021/es204131r | Environ. Sci. Technol. 2012, 46, 1886−1894

Environmental Science & Technology

Article

(29) Khan, M. F.; Bennett, M. J.; Jumper, C. C.; Percy, A. J.; Silva, L. P.; Schriemer, D. C. Proteomics by mass spectrometry - Go big or go home? J. Pharm. Biomed. Anal. 2011, 55 (4), 832−841.

1894

dx.doi.org/10.1021/es204131r | Environ. Sci. Technol. 2012, 46, 1886−1894