NF stage 54 tadpoles from US Environmental Protection Agency, Mid-Continent Ecology Division in house culture
Treatment protocol
Experiment 1 (T4 and T3) Twenty-eight early prometamorphic (NF stage 54;(Nieuwkoop and Faber, 1956; Shi, 2000)) tadpoles were continuously exposed to three separate T4 concentrations (10, 20.1, and 40.3 nM), and a LSW control in one exposure set, or five different T3 concentrations (0.48, 0.97, 1.92, 3.84, and 7.68 nM), and a LSW control in the second exposure set as described in detail in Zhang et al (2006). Each chemical exposure concentration was replicated twice along with the associated LSW control. At 24 h, 48 h and 96 h two animals per exposure replicate (four animals total per each individual treatment) were randomly selected, euthanized in MS-222, and preserved in RNAlater (Ambion Inc., Austin, Texas, USA) for analysis of gene expression. On exposure day 14, all remaining organisms were euthanized in MS-222, weighed, and developmentally staged in a blind evaluation. Animals exposed to either chemical showed an acceleration of metamorphosis which was published previously (Zhang et al., 2006).
Experiment 2 (MMI, PTU and PER) NF stage 54 tadpoles were continuously exposed to a single concentration of PTU (20 mg/L), MMI (100 mg/L) or PER (4 mg/L). We have previously shown that exposure to these concentrations resulted in an increase in thyroid gland size at day 8 and significantly delayed metamorphosis at 14 days (Degitz et al., 2005; Tietge et al., 2005). The exposure regimen details are recorded elsewhere (Zhang et al., 2006). Briefly, tadpoles were randomly placed into 24 tanks (20 tadpoles/tank) and exposed (six tanks/chemical) to a single concentration of each chemical. At 24 h, 48 h and 96 h 5 tadpoles from two of the 6 tanks (10 tadpoles per each individual treatment) were randomly selected, euthanized in MS-222, and preserved in RNAlater (Ambion Inc., Austin, Texas, USA) for analysis of gene expression.
Extracted molecule
total RNA
Extraction protocol
Brain tissue was collected from each individual tadpole and total RNA was isolated using TRIzol reagent as described by the manufacturer (Invitrogen Canada Inc., Burlington, Ontario, Canada).
Label
P32
Label protocol
Amplified RNA (aRNA) was produced using the MessageAmp aRNA Kit as per the manufacturer’s protocol (Ambion) from one microgram brain total RNA isolated from an individual (for the T3 and T4 experimental sets) or from three separate pools of total RNA from individuals (for the inhibitor experimental set). This aRNA (500 ng) was annealed with 500 ng random hexamer oligonucleotide (Amersham Biosciences Inc, Baie d’Urfé, QC, Canada), and cDNA was prepared using MMLV RNase H- Superscript II reverse transcriptase as described by the manufacturer (Invitrogen) with the following modifications: the dNTP mix consisted of 500 µM each of dGTP, dTTP, and dCTP, 4 µM dATP, and 50 µCi [alpha-32P] dATP (PerkinElmer Life Sciences, Inc, Boston, MA, USA). RNA was removed from the radiolabeled cDNA by addition of 10 ul 1M NaOH and incubation at 70 °C for 10 min. Samples were cooled to room temperature and 10 ul of 1M HCl were added to neutralize the reactions. The radiolabeled cDNA was purified using a QIAquick PCR purification kit (Qiagen) and, immediately prior to hybridization, was heat denatured for 5 min at 95°C and then quickly cooled on ice for 5 min.
Hybridization protocol
Prehybridization, hybridization, and posthybridization washes were performed at 65°C. The radiolabeled cDNA targets from MMI-, PTU-, and PER-exposed animals and their time-matched experimental controls were each divided into 3 pools of (n=3, 3, and 4) for each treatment time point such that three arrays were probed per treatment and time point. The radiolabeled cDNA targets from T3 and T4-exposed animals with their respective time-matched control animals, were derived from 3 individual animals such that one array was used for each individual. Hybridizations were carried out in 20 ml of hybridization solution containing 4x SSC, 10% (w/v) dextran sulfate, 1.0% (w/v) SDS, and 0.5% (w/v) Blotto. Prewarmed hybridization solution was added to each hybridization tube (35 mm internal diameter x 150 mm length; Amersham) containing the array membrane and prehybridization was allowed to continue for 2 h. Radiolabeled cDNA samples were then added to a final concentration of 5 x 105 cpm/ml and allowed to hybridize overnight. After hybridization, the membranes were rinsed briefly with 50 ml 2x SSC, then washed twice with 50 ml 2x SSC/0.1% SDS for 15 min, once with 50 ml 0.1x SSC/1.0% SDS for 25 min, and rinsed with 50 ml 0.1x SSC. The arrays were placed on 3MM filter paper (Rose Scientific Ltd, Edmonton, AB, Canada) soaked with ddH20 and wrapped with plastic wrap. Each processed membrane was exposed to a phosphorimager screen (Molecular Dynamics Inc., Sunnyvale, CA, USA) for five days.
Scan protocol
Hybridization signals were collected using a Storm 820 Gel and Blot Imaging System (Amersham) at 50 µm resolution.
Description
Gene expression data sets obtained from three biological replicate arrays for each treatment and time point were analyzed for consistency and absolute agreement using intra-class correlation with a 2-way random effects model (SPSS Version 12.0, Chicago, IL, USA). The averages of the intra-class correlation coefficients over the replicate sets was not less than 0.82 with most values greater than 0.9 (Supplementary Table 3). Fold change response to each treatment was determined for each gene relative to the control and genes exhibiting fold change values greater than or equal to two were identified. As an additional quality measure and to reduce the chance of making a type I error, we omitted genes which displayed large variability over the replicate sets. For each gene/replicate set, variability was measured using the equation: (maximum value-minimum value)/2 which is appropriate for samples up to 5 per group (Montgomery, 1991). If the variation estimate for the gene/replicate set was greater than 1.1-1.7 (depending on the data set), the gene was removed from the final list of genes affected following chemical treatment for that time point. If less than 4 observations were available for a given time point and treatment, that gene was also removed from the final list of genes.
Data processing
The resulting image data were converted to a standard 8-bit TIFF file using Photoshop V5.0 (Adobe Systems Inc, San Jose, CA, USA). Both non-auto- and auto-level images were prepared for analysis in order to account for signal saturation. Non-auto-leveled images provide a linear range of strong signal intensities (such as alpha-actin) while auto-leveled images allow for analysis of the remaining signals. Relative expression for each gene target was collected from the image data using ImaGene Version 5.6.1 (BioDiscovery Inc, El Segundo, CA, USA). Signal intensities for each gene and blank position were determined from the median spot pixel intensities and corrected by subtracting the local median background pixel intensities. Signal intensities that were derived from areas of non-specific hybridization on the arrays were not included in the final analysis. A non-signal background was determined from the median intensity value plus one standard deviation of blank positions across the auto-leveled data set, and signal intensities for gene positions exhibiting values below the no signal were adjusted to this value. Saturated gene positions identified in auto-level data were replaced across all data sets by the corresponding values obtained in the non-auto-level analysis. Data for the three sets were analyzed by time point. Both non-auto- and auto-leveled data for each array were normalized using a geometric mean derived from the median signal intensities from the following genes: ribosomal proteins L8 and S10, GAPDH, ferritin, ubiquitin, NM23/dinucleotide phosphate kinase, cytoplasmic alpha-actin, and elongation factor-1 ? chain. The choice of which of these transcripts were used for normalization within a time point was dictated by spot quality and consistency (see Statistics below). The gene transcripts used for normalization across a time point are indicated in Supplementary Table 1.
