Article pubs.acs.org/est
Analysis of the Elodea nuttallii Transcriptome in Response to Mercury and Cadmium Pollution: Development of Sensitive Tools for Rapid Ecotoxicological Testing Nicole Regier,† Loïc Baerlocher,‡ Martin Münsterkötter,§ Laurent Farinelli,‡ and Claudia Cosio†,* †
Institut F.-A. Forel, University of Geneva, 10 route de Suisse, CP416, 1290 Versoix, Switzerland Fasteris SA, 109 chemin du Pont-du-Centenaire, CP28, 1228 Plan-les-Ouates, Switzerland § Institute of Bioinformatics and Systems Biology, Helmholtz Centre Munich, Ingolstädter Landstr. 1, 85764 Neuherberg, Germany ‡
S Supporting Information *
ABSTRACT: Toxic metals polluting aquatic ecosystems are taken up by inhabitants and accumulate in the food web, affecting species at all trophic levels. It is therefore important to have good tools to assess the level of risk represented by toxic metals in the environment. Macrophytes are potential organisms for the identification of metal-responsive biomarkers but are still underrepresented in ecotoxicology. In the present study, we used next-generation sequencing to investigate the transcriptomic response of Elodea nuttallii exposed to enhanced concentrations of Hg and Cd. We de novo assembled more than 60 000 contigs, of which we found 170 to be regulated dose-dependently by Hg and 212 by Cd. Functional analysis showed that these genes were notably related to energy and metal homeostasis. Expression analysis using nCounter of a subset of genes showed that the gene expression pattern was able to assess toxic metal exposure in complex environmental samples and was more sensitive than other end points (e.g., bioaccumulation, photosynthesis, etc.). In conclusion, we demonstrate the feasibility of using gene expression signatures for the assessment of environmental contamination, using an organism without previous genetic information. This is of interest to ecotoxicology in a wider sense given the possibility to develop specific and sensitive bioassays.
■
INTRODUCTION Human activity has led to a strongly increased input of toxic metals into the environment. In aquatic ecosystems, toxic metals have received special attention because of their toxicity to biota and bioaccumulation in aquatic food webs.1,2 Mercury (Hg) and cadmium (Cd) are two of the most toxic metals found on earth and belong to the most important pollutants worldwide. Furthermore, they are considered to be among the most hazardous elements accumulating in food webs, implicating risks also for human health.3 Rooted freshwater macrophytes are a key player of lake ecosystems.4 Elodea nuttallii is a rooted, submerged macrophyte native to Nothern America, which is constantly spreading worldwide.5 Elodea sp. are known to be able to adapt to a broad range of environmental conditions. They are exposed to both water and sediments and can accumulate high amounts of toxic metals. At the same time, they are a potential food source for herbivores, such as macroinvertebrates and fish, and may pass toxic metals into food webs.6−8 Due to its occurrence in a wide variety of aquatic environments, E. nuttalli is a good candidate test species for ecotoxicology. In the present work, we tested the possibility of using altered gene expression as fast and sensitive response to exposure to © 2013 American Chemical Society
metals. Due to the rapidity of transcriptomic responses in general, they allow obtaining results already after exposure to a contaminated environment for several hours up to one day and are promising to be used as specific biomarker of stress for environmental testing. Therefore, transcriptomic analyses could complement current ecotoxicological tests that use responses such as effects on growth and survival, which comprise biomarkers for general stress and usually require several days to weeks (e.g., refs 9−11). Several recent studies have used transcriptomics to classify trace contaminants and predict their mode of toxicity by measuring clusters of signature genes.12−14 For example, by means of expression patterns in Chlamydomonas reinhardtii, Chironomus tentans, and Daphnia magna, it was possible to discriminate between different toxic polyaromatics, organic pollutants, and metals.14−17 However, few studies have confirmed ecotoxicogenomic data in the field. In the current study, we aimed to develop an ecotoxicogenomic tool for the rooted macrophyte E. nuttallii and to apply it in the field. Received: Revised: Accepted: Published: 8825
March 11, 2013 June 4, 2013 June 26, 2013 June 26, 2013 dx.doi.org/10.1021/es401082h | Environ. Sci. Technol. 2013, 47, 8825−8834
Environmental Science & Technology
Article
However, as E. nuttallii is not yet established as a model species, there was practically no genomic sequence information available. Recent advances in sequencing technology such as RNA-Seq allow generating efficiently large amounts of genomic information and provide ecotoxicology research with promising new tools to study the effects of toxicants on the transcriptomic level without previously available sequence information, which is especially important when studying nonmodel organisms.18−21 Despite the importance of, and the problems caused by, toxic metal pollution in freshwater environments and the crucial role of macrophytes within those ecosystems,22−24 nothing is known about transcriptomic responses of aquatic macrophytes to toxic metals. Most studies that have assessed the effect of toxic metals on gene expression in plants on a transcriptomewide basis have used model species exposed under laboratory conditions such as poplar,25 tobacco,26 Arabidopsis thaliana,27 or Thlaspi caerulescens, which is known as metal hyperaccumulator.28 Moreover, no whole transcriptome analysis in response to Hg has been conducted in macrophytes. In the present study, we used next generation sequencing on the Illumina platform to characterize the transcriptome of E. nuttallii in response to increasing concentrations of Hg and Cd, allowing us to provide valuable sequence information for future research on this ecologically important macrophyte. Subsequently, we used NanoString nCounter technology29 to further analyze a differentially expressed subset of genes, which allowed us to detect metal-responsive expression patterns and gene expression signatures and therefore enabled us to develop a sensitive tool for future ecotoxicological testing.
of the discharge channel of a local chlor-alkali plant (GPS coordinates 44°55′13″N, 24°14′40″E).30 Plants were suspended in the water column for 24 h at 1, 4, and 6 m depth, corresponding to Hg concentrations of 3.63 ± 0.12, 2.9 ± 0.96, and 3.09 ± 0.61 ng L−1. pH, conductivity (μS cm−1), and temperature (°C) varied from 8.07 ± 0.34, 976 ± 17, 26.06 ± 1.01 at 1 m to 7.01 ± 0.28, 1321 ± 84, 24.17 ± 0.41 at 6 m, respectively. Plants in unmodified bottles, filled with water collected upstream of the chlor-alkali plant and exposed at the same depths, served as controls for field-exposed plants (0.3 ± 0.05 ng L−1 Hg, pH 8.49 ± 0.10, conductivity 332 ± 3.79 μS cm−1). Contaminated sediments from the Babeni reservoir (Hg concentration 1.8 ± 0.03 mg kg−1) were collected and brought to the laboratory. A layer of sediment was put into 1 L plastic beakers, which were then filled with Lake Geneva water and 10 cm long shoots were planted in the sediment for 24 h. Finally, a second set of nCounter analyses was conducted on plants exposed to stepwise increasing concentrations of Hg for 24 h to study dose−response effects on gene expression, where effective Hg concentrations were as follows: 0.056 ± 0.02, 0.18 ± 0.08, 0.21 ± 0.05, 0.25 ± 0.09, 1.21 ± 0.08, 4.18 ± 1.34, 35.6 ± 2.7, 522.6 ± 15.4 ng L−1 and 6.28 ± 0.28, 54.25 ± 8.98 μg L−1. RNA Isolation. Total RNA was extracted from shoots using the TRI-reagent and further purified using the RNeasy Plant Mini kit (Qiagen) according to the protocols supplied by the manufacturers. RNA concentration was determined spectrophotometrically at 260 nm and quality was assessed by agarose gel electrophoresis. Illumina mRNA Sequencing. Samples for Illumina mRNA sequencing (RNA-Seq) were prepared using the mRNA-Seq sample prep kit according to the manufacturer’s instructions (Illumina Inc.). Briefly, during this procedure the mRNA was purified from total RNA, double-stranded cDNA was synthesized and bar-coded adapters were ligated to both ends of the cDNA fragments. After electrophoresis on a 2% agarose gel, the cDNA fraction with a size of 200 ± 25 bp was excised from the gel and enriched by PCR using adapter-specific primers. The cDNA libraries were attached onto the surface of a Solexa flow cell, multiplexed in a single channel, and sequenced on an Illumina Genome Analyzer GAIIx for 2 × 54 cycles using the Chrysalis 36 cycles sequencing kit (paired-end runs, PE) or for 75 cycles (single-end runs, SE; Table 1). De Novo Transcriptome Assembly. De novo assembly of the transcriptome was achieved with the publicly available program VELVET (version 1.0.12; http://www.ebi.ac.uk/
■
EXPERIMENTAL SECTION Plant Material and Exposure. Cultures of Elodea nuttallii were established and maintained as described in Regier et al. (2013).8 Exposure to metals was done on 10 cm long shoots in 1 L plastic beakers containing filtered water from Lake Geneva. Plants were exposed under standardized conditions (20 °C, 16/ 8 h light/dark cycle) for 24 h. All experiments were started at the same time of the day in order to avoid detection of genes regulated diurnally by the circadian clock. Plants were exposed to the following metal concentrations: 500 and 5000 μg L−1 CdCl2 (effective concentrations 576 and 5752 μg L−1 Cd), 200 ng L−1, 200 μg L−1, and 2 mg L−1 HgCl2 (effective concentrations 70 ng L−1, 80 μg L−1, and 800 μg L−1 Hg). Shoots exposed to Lake Geneva water without metal spike were used as a control (0.05 ± 0.01 ng L−1 Hg and 4.67 ± 0.71 ng L−1 Cd). A first exposure to all metal concentrations mentioned was performed for RNA-Seq, and a second exposure to 500 and 5000 μg L−1 CdCl2 and to 200 ng L−1 and 200 μg L−1 HgCl2 as well as control was conducted for RT-qPCR. A third exposure was conducted for expression analysis in response to more different conditions using the nCounter system (NanoString Inc., Seattle WA, U.S.A.). The following metal concentrations were used for nCounter: control (no spike), 10, 50, and 200 ng L−1 HgCl2, 200 μg L−1 HgCl2, 30 ng L−1 CH3HgCl, 65 ng L−1 and 325 μg L−1 CuCl2, 200 ng L−1 HgCl2 + 30 ng L−1 CH3HgCl, 200 ng L−1 HgCl2 + 65 ng L−1 CuCl2, 500 μg L−1 CdCl2 and 200 ng L−1 HgCl2 + 500 μg L−1 CdCl2. To evaluate the metal-specificity of the response, we tested three distinct parameters: salinity (4 g L−1 NaCl), temperature (10 °C), and darkness. Additionally, plants in bottles with holes of 2 mm diameter, were exposed to the Hg contaminated Babeni reservoir in Romania, located on the Olt River downstream
Table 1. Number of Total Reads Obtained Per Sample As Well As Number and Percentage of Reads Per Sample Which Could Be Mapped on the Assembled Transcriptomea
a
8826
sample
read type
no. reads
mapped reads
% mapped
control control 70 ng L−1 Hg 80 μg L−1 Hg 800 μg L−1 Hg 576 μg L−1 Cd 5762 μg L−1 Cd total total
PE SE PE SE SE PE PE SE PE
11 897 050 9 796 939 15 995 086 2 943 621 1 677 801 10 764 094 11 531 568 14 418 361 50 187 798
5 068 227 4 827 611 4 651 925 1 536 494 909 949 4 886 273 4 294 478 7 274 054 18 900 903
42.6 49.3 29.1 52.2 54.2 45.4 37.2 50.4 37.7
PE: paired-end reads. SE: single-end reads. dx.doi.org/10.1021/es401082h | Environ. Sci. Technol. 2013, 47, 8825−8834
Environmental Science & Technology
Article
∼zerbino/velvet/),31 and the VELVET extension OASES (version 0.1.15; http://www.ebi.ac.uk/∼zerbino/oases/). The average insert size was set to 200 bp, and the lengths of sequence overlaps (referred to as k-mers) were varied in order to obtain the best possible results. Mapping of Sequence Reads onto Transcripts and Quantification of Gene Expression. The MAQ program (Mapping and Assembly with Qualities; version 0.7.1) 32 was used to map the reads of each sample on the assembled contigs in order to quantify gene expression. To be able to compare between the samples, mapping results were normalized to the sample with the lowest number of total reads. Transcripts with an arbitrary minimal expression difference of 2-fold were identified in single comparisons between two samples at a time. Those subsets were further used to identify transcripts that were regulated in response to Hg and Cd in a dose-dependent manner (e.g., control
