Environ. Sci. Technol. 2006, 40, 1076-1083
Gene Expression Profiling in Common Cormorant Liver with an Oligo Array: Assessing the Potential Toxic Effects of Environmental Contaminants K E I N A K A Y A M A , † H I S A T O I W A T A , * ,† EUN-YOUNG KIM,† KOSUKE TASHIRO,‡ AND SHINSUKE TANABE† Center for Marine Environmental Studies (CMES), Ehime University, Bunkyo-cho 2-5, Matsuyama 790-8577, Japan, and Faculty of Agriculture, Graduate School, Kyushu University, Hakozaki 6-10-1, Higashi-ku, Fukuoka 812-8581, Japan
To establish a monitoring system for gene expression profiles related to chemical contamination in wild common cormorants (Phalacrocorax carbo), the present study constructed an oligo array designed from expressed sequence tag (EST) sequences of the cormorant liver, where 1061 unique oligonucleotides were spotted. Common cormorants were collected from Lake Biwa, Japan in May 2001 and 2002. With the use of this oligo array, gene expression profiles in the liver of individual specimens were evaluated. To determine the expression patterns of genes altered by environmental contaminants, relationships between concentrations of persistent organochlorines including polychlorinated dibenzo-p-dioxins, furans, polychlorinated biphenyls, 1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane and its metabolites (DDTs), hexachlorocyclohexane isomers (HCHs), chlordane compounds (CHLs), butyltins, and bisphenol A (BPA) and expression levels of each gene in the cormorant liver were examined using stepwise multiple regression analysis. The reliability of data obtained by the oligo array was further confirmed by quantifying the expression levels of certain genes using real-time RTPCR. The 2,3,7,8-tetrachlorodibenzo-p-dioxin toxic equivalent (TEQ) level was positively correlated with both cytochrome P4501A4 and 1A5 gene expression. In addition, the mRNA level of an antioxidant enzyme, Cu/Zn superoxide dismutase, was negatively correlated with hepatic total TEQ. Other antioxidant enzymes, glutathione peroxidase 3 and glutathione S-transferase class µ, were negatively correlated with HCHs and BPA levels, respectively. The mRNA expression level of a nonenzymatic antioxidant, haptoglobin, was negatively but not significantly correlated with CHLs. These results led to a hypothesis that wild cormorant population may suffer from oxidative stress due to chemically induced formation of reactive oxygen species and subsequent reduction of antioxidant resistance. Thus, the cormorant oligo array may be a useful monitoring tool to identify specific gene expression profiles altered * Corresponding author phone/fax: +81-89-927-8172; e-mail:
[email protected]. † Ehime University. ‡ Kyushu University. 1076
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 3, 2006
by various environmental contaminants. Although further research is required to clarify a definitive cause-and-effect relationship, the current study provides valuable information on contaminant-responsive genes to predict potential effects on wildlife in a real environment.
Introduction Persistent organic pollutants (POPs) are biomagnified at higher trophic levels through food webs. The common cormorant (Phalacrocorax carbo), a fish-eating bird, is a top predator in the ecosystem of Lake Biwa, which is the largest freshwater reservoir in Japan. For these reasons, the Japanese Ministry of the Environment has been using the cormorant as a model organism to monitor environmental contaminants, especially dioxin-like compounds (“Survey on the State of Dioxin Accumulation in Wildlife”). One of our recent studies reported that cormorants collected from Lake Biwa were highly contaminated with dioxin-like compounds, such as polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), and coplanar polychlorinated biphenyls (co-PCBs) (1). Moreover, other organochlorines, butyltin compounds, and estrogenic chemicals were also detected in the liver of the same specimens (2, 3). Thus, the cormorant population inhabiting Lake Biwa has been exposed to a mixture of environmental contaminants, and this species is therefore a good model in which to observe the contamination of chemicals and their toxic effects. However, effects exhibited by these toxic chemicals in this species are not fully understood. It is well-known that dioxin-like compounds induce cytochrome P450 (CYP) 1A via the aryl hydrocarbon receptor, and CYP1A induction has been used as a biomarker to detect the contamination of dioxin-like compounds. Our previous research demonstrated that CYP1A protein and its catalytic activity were positively correlated with 2,3,7,8-tetrachlorodibenzo-p-dioxin toxic equivalent (TEQ) levels in the livers of cormorants from Lake Biwa, and the study also suggested a CYP1A-dependent metabolism or hepatic sequestration of dioxins (2). However, information on toxic effects of environmental contaminants on wildlife species is quite limited, and few studies addressing mixture toxicity on wildlife have been reported. Because organisms generally react to chemical exposure by altering the expression levels of mutliple genes, a wide variety of molecular changes should be monitored in order to predict potential toxic effects and their mechanisms. Global gene expression analysis can be used to evaluate chemical exposure and further toxic effects associated with the alteration of gene expression. However, the DNA sequences of only a few contaminant-responsive genes are currently available in wild species. Therefore, it is important to collect the information of these genes in wildlife. DNA array technology has become a standard tool in both molecular biology and clinical diagnostics. Because of its potential utility in global gene expression analysis, growing interest in this technology has led to its application to toxicological research, in the form of toxico- and ecotoxicogenomics (4). Commercially available and originally designed microarrays have been used in experimental animals such as rats and mice. Previous studies reported expression of dioxin-responsive genes in multiple mouse organs (5), in rat livers (6), and in human HepG2 cells (7). In addition, effects on gene expression profiles in the liver or other organs also have been investigated using rats exposed 10.1021/es051386m CCC: $33.50
2006 American Chemical Society Published on Web 01/06/2006
to other organochlorines, such as PCBs (6) or hexachlorobenzene (HCB; 8). Furthermore, Adeeko et al. (9) reported that a mixture of POPs altered the expression levels of a wide variety of genes in rat dams and fetuses. However, experimental designs of these exposure tests are a single or repeated dose(s) of chemical(s) via various administration routes over a short period of time, indicating that these designs do not completely reflect a real environment. In addition, chemical effects on gene expression may vary among species or even strains of the same species. Therefore, it is quite important to investigate the effects on gene expression profiles in animals inhabiting a real environment. From the aspect described above, microarray technology has been also applied to field-collected organisms. Williams et al. (10) detected up- or down-regulated genes in the liver of European flounder (Platichthys flesus) collected from a contaminated site, compared to those from a cleaner reference site. Thus, microarray technology is a powerful tool to detect chemically altered gene expression in wild animals as well as experimental animals. However, only a few microarray studies addressing alteration of gene expression profiles related to environmental contaminants in higher trophic wild species have been reported to date. To screen contaminant-responsive genes, to predict potential toxic effects, and to understand their mechanisms at molecular level in wild common cormorants, we constructed an oligo array targeting genes expressed in the liver and analyzed gene expression profiles in the liver of birds collected from Lake Biwa. Relationships between concentrations of environmental contaminants and gene expression patterns were also examined to screen for contaminantresponsive genes. To confirm the expression levels of genes screened by oligo array analysis, mRNA levels were further quantified by real-time RT-PCR. We also compared the microarray data to the real-time RT-PCR data to assess the consistency of target gene levels quantified by two different methods.
Materials and Methods Samples. Common cormorants were randomly collected from Lake Biwa in May 2001 and 2002. Liver samples were immediately removed after the measurement of biometry and stored at -20 °C for chemical analysis. Subsamples of livers were frozen in liquid nitrogen and stored at -80 °C until RNA preparation. The concentrations of environmental contaminants in cormorant liver were described in previous reports (1-3). Polychlorinated dibenzo-p-dioxins (PCDDs), polychlorinated dibenzofurans (PCDFs), polychlorinated biphenyls (PCBs), 1,1,1-trichloro-2,2-bis(p-chlorophenyl)ethane and its metabolites (DDTs; p,p′-DDT, -DDE, -DDD, o,p′-DDT, -DDE, and -DDD), hexachlorocyclohexane isomers (HCHs; R-, β-, and γ-HCH isomers), chlordane compounds (CHLs; cis-chlordane, trans-chlordane, oxychlordane, and trans-nonachlor), butyltin compounds (BTs; tri-, di-, and monobutyltins), and bisphenol A (BPA) were measured in liver tissue. PCDDs/DFs and coplanar PCBs were measured in samples collected during both 2001 and 2002 (n ) 20, 10 males and 10 females). Other contaminants were measured only in 2001 samples (n ) 16, 7 males and 9 females). The TEQ of dioxin-like compounds in the liver of cormorants was estimated using the toxic equivalency factor (TEF) for birds (11). Construction of an Oligo Array. A cormorant liver cDNA library was prepared for the construction of an oligo array (TOYOBO, Tsuruga, Japan). To characterize the cDNA library, 6930 randomly selected clones were sequenced. Following a BLAST homology search of cDNA sequences, approximately 2500 cDNA clones whose sequences revealed high identities with genes in the GenBank database were obtained. We designed 1061 unique 70-mer oligonucleotides. These oli-
TABLE 1. List of Genes Spotted on an Oligo Array gene species
n
oncogene transcription factor translation initiation factor elongation factor receptor transporter immune system synthetase kinase phosphatase xenobiotic-metabolizing enzyme antioxidant enzyme other enzymes notch homeobox growth factor heat shock protein ribosomal protein glycoprotein KIAA RIKEN MGC ATP-related protein haptoglobin selenoprotein cathepsin metallothionein others
27 49 7 5 81 16 37 38 53 11 40 7 205 3 5 5 8 35 16 61 49 69 18 3 6 9 1 197
total
1061
gonucleotides were spotted in duplicate onto TaKaRa-Hubble slide glass (TAKARA BIO INC., Shiga, Japan). Spotted genes on the oligo array were functionally categorized as listed in Table 1. Preparation of Antisense RNA Probe. Total RNA was extracted from the liver tissue (e400 mg) with TRIzol reagent (Invitrogen, Carlsbad, CA), and the quality of each RNA sample was checked by denaturing agarose gel electrophoresis. RNA samples from three specimens that contained relatively low TEQ levels of dioxin-like compounds were pooled to use as a common reference. We used 5 µg of total RNA to amplify amino allyl antisense RNA (aRNA) using an Amino Allyl MessageAmp aRNA kit (Ambion Inc., Austin, TX). To prepare RNA probes, 5 µg of aRNA was coupled with Cy3 (data; n ) 15) or Cy5 (reference) (Amersham Biosciences, Piscataway, NJ). Excess Cy dye was removed using a QIAquick PCR purification kit (QIAGEN K. K., Tokyo, Japan). Cy3- and Cy5-labeled aRNA was then mixed and fragmented with fragmentation buffer (5× fragmentation buffer: 0.5 M potassium acetate, 0.15 M magnesium acetate in 0.2 M Trisacetate [pH 8.1]). Labeled aRNA was purified and concentrated with Microcon YM-10 filters (Millipore, Billerica, MA) for use in hybridization. Hybridization, Wash, and Scan. For hybridization, the following target solution was used: Cy dye-labeled aRNA solution in 6× SSC, 0.2% SDS, 5× Denhardt’s solution (Eppendorf, Hamburg, Germany), and 0.1 mg mL-1 human cot-I (Invitrogen). The target solution was heated at 95 °C for 2 min to denature aRNA, cooled on ice, and then incubated at 65 °C for 5 min. After incubation, the target solution was placed on a glass slide, and the aRNA probes were hybridized with oligonucleotides at 65 °C for 17 h. Following the hybridization, the slides were washed twice with 0.2% SDS in 2× SSC at 55 °C and once with 0.2% SDS in 2× SSC at 65 °C, rinsed twice with 0.2× SSC at room temperature, and then dried by centrifugation at 800 rpm for 3 min. The washed slides were then scanned using a fluor-image analyzer (FLA8000, Fuji Photofilm Co. Ltd, Tokyo, Japan) at 532 nm (Cy3) and at 635 nm (Cy5). VOL. 40, NO. 3, 2006 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1077
TABLE 2. Primer Pairs and Probes Designed for Quantification of mRNA Expression Levels by Real-Time RT-PCR gene name
primer (upper, forward; lower, reverse)
Cu/Zn superoxide dismutase
5′-GGAGTGGCAGAAGTGGAAATAGA-3′ 5′-CAGGTCATCACGTTTTTCATGG-3′ glutathione peroxidase 3 5′-GTGCAAGGCACCATTTATGACTAC-3′ 5′-CACATTGACGAAGAGCACCATC-3′ hemopexin 5′-CACAAAAAACCGGAGGCA-3′ 5′-AGACCTCCTCACCTCTGAAGAAG-3′ haptoglobin 5′-CGGAGATTGAACATGGCTACAT-3′ 5′-TCACCCACTCATGATCCTTGTC-3′ ovotransferrin 5′-AACATGCAAGCAGAGGCTGTT-3′ 5′-CAAGGTAATGGCATCTGCTTCA-3′ glutathione S-transferase µ 1 5′-TTGAGGCGCTGGAGAAGATCTC-3′ 5′-CTTGGTGTTGCACCACTTTGCT-3′ glutathione S-transferase µ 3 5′-TCACCTTCGTGGACTTCCTGAT-3′ 5′-AAAGCGGTCCATGAAGTCCTT-3′ actin polymerase inhibitor 5′-CACCTTGGATGTCAACCACTTC-3′ 5′-AGCCATGCTCATCCTGCTTCT-3′ cysteine dioxygenase 5′-ATTTGTACAGCCCACCCTTCG-3′ 5′-TCCTTTCTCCAAACTGGCTGTAG-3′ estrogen receptor-related protein 5′-TGACCCTACTGTTCCAGACAGTGA-3′ 5′-GGAATATGCTTAGCCCATCCAA-3′
Microarray Data Analyses. Fluorescent intensities were quantified by Array Vision (Imaging Research Inc., St. Catharines, Ontario, Canada). The intensities on four corners of each spot were used as background. The fluorescence units of each spot were corrected using the mean value of the intensities on negative control spots. Expression levels of each gene were represented as data (Cy3)/reference (Cy5) ratios. The ratios were normalized using the Locfit (LOWESS) function in TIGR MIDAS (version 2.19; 12). Genes with an intensity less than 5000 were not included in the subsequent analyses. A preliminary test in which three samples were analyzed in triplicate on separate arrays indicated that the interslide variation of the gene expression levels was within 20%, if the spots with weak intensity were removed. Real-Time RT-PCR. Total RNA extracted from each liver sample (year 2001; n ) 16) was treated with DNase prior to quantification. Quantitative PCR was performed using the TaqMan one-step RT-PCR method (Applied Biosystems, Foster City, CA) and an ABI PRISM 7700 Sequence Detector System (Applied Biosystems). Primers and probes used in target gene quantification are listed in Table 2. Quantitative values were obtained from the threshold PCR cycle number (Ct) at which the increase in signal associated with an exponential growth of PCR products was detected. The relative mRNA levels in each sample were normalized to ribosomal RNA content. Statistical Analyses. To analyze relationships between gene expression and contaminant levels, a Spearman ranksum test was preliminarily performed. Correlation among contaminants was analyzed by the Pearson correlation test. Prior to analysis, contaminant concentrations and gene expression levels were logarithmically transformed. The association of contaminants with gene expression levels was examined by stepwise multiple linear regression analysis. Sex differences in contaminant concentrations and mRNA expression levels of each gene were analyzed by the Student’s t-test. Correlations of the microarray data to the real-time RT-PCR data in each gene quantified were also analyzed by the Spearman rank-sum test. All statistical analyses were performed using StatView 5 (SAS Institute Inc., Cary, NC) and SPSS 12.0J (SPSS Japan, Tokyo, Japan).
Results and Discussion Chemical Analysis. Concentration ranges and medians of each contaminant measured in the liver of common cormorants from Lake Biwa, Japan are shown in Table 3. All the measured contaminants except BPA were detectable in all 1078
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 3, 2006
amplicon (bp)
probe 5′-CTCTTTGTGGACCACACAACATCATTGGAC-3′
102
5′-CGAGTACATCCCCTTTAGGAATTACGCAGG-3′
102
5′-AGCTTTGATGCCGTCACACTCAGCGAGA-3′
160
5′-CACCTGATCAAGTACCACTGCAACCCGTAC-3′
117
5′-TTGACTTGCCTGCAGAAAGCAATGTACCTC-3′
102
5′-TCATGAAGACCCCCATTTTCTGGCGCA-3′
98
5′-CCAAGTGCCTGGAGCCCTTCAAGAA-3′
103
5′-ATAACATCGTGGAGATAACCGGCAAACACG-3′
107
5′-ACACCTTTGATCAGAGGACTGGACACAAGC-3′
102
5′-AAGCGCTTACCACTCTTTGTGACCTGGC-3′
102
TABLE 3. Concentrations of Environmental Contaminants Detected in the Liver of Common Cormorants from Lake Biwa, Japan (n ) 16) median (range) concentration contaminanta
male (n ) 7)
ΣTEQb
130
PCBs DDTs HCHs CHLs BTs BPA
400 140 14 14 140 0.53
female (n ) 9) (pg/g Wet Weight) (12-960) 62 (ng/g Wet Weight) (23-4000) 160 (13-1200) 150 (6.5-130) 22 (2.1-62) 12 (87-670) 150 (0.11-0.89) 0.36
(20-160) (36-740) (26-910) (6.3-52) (1.9-61) (43-680) (
