ARTICLE pubs.acs.org/est
Metabolomics of Microliter Hemolymph Samples Enables an Improved Understanding of the Combined Metabolic and Transcriptional Responses of Daphnia magna to Cadmium Helen C. Poynton,^,†,‡ Nadine S. Taylor,^,§ Joshua Hicks,|| Kimberly Colson,|| Sarah Chan,‡ Candace Clark,‡ Leona Scanlan,‡ Alexandre V. Loguinov,‡ Chris Vulpe,#,‡ and Mark R. Viant*,#,§ †
National Exposure Research Laboratory, U.S. Environmental Protection Agency, Cincinnati, Ohio, United States Department of Nutritional Sciences and Toxicology, UC Berkeley, Berkeley, California, United States § School of Biosciences, University of Birmingham, Edgbaston, Birmingham B15 2TT, U.K. Bruker-Biospin, Billerica, Massachusetts, United States
)
‡
bS Supporting Information ABSTRACT: Omic technologies offer unprecedented opportunities to better understand mode(s)-oftoxicity and downstream secondary effects by providing a holistic view of the molecular changes underlying physiological disruption. Crustacean hemolymph represents a largely untapped biochemical resource for such toxicity studies. We sought to characterize changes in the hemolymph metabolome and whole-body transcriptome to reveal early processes leading to chronic toxicity in the indicator species, Daphnia magna, after 24-h sublethal cadmium exposure (18 μg/L, corresponding to 1/10 LC50). We first confirmed that metabolites can be detected and identified in small volumes (∼36 μL) of D. magna hemolymph using Fourier transform ion cyclotron resonance mass spectrometry and NMR spectroscopy. Subsequently, mass spectrometry based metabolomics of hemolymph identified disruption to two major classes of metabolites: amino acids and fatty acids. These findings were compared to differentially expressed genes identified by a D. magna 44k oligonucleotide microarray, which included decreased levels of digestive enzymes and increased expression of cuticle proteins and oxidative stress response genes. The combination of metabolic and transcriptional changes revealed through KEGG pathway analysis and gene ontology, respectively, enabled a more complete understanding of how cadmium disrupts nutrient uptake and metabolism, ultimately resulting in decreased energy reserves and chronic toxicity.
’ INTRODUCTION Chemicals causing chronic sublethal toxicity at low levels pose a significant challenge for ecological risk assessment, requiring innovative approaches that consider mode(s)-of-action (MOA).1 Recently, the incorporation of omics technologies into toxicity testing has been proposed to increase our understanding of MOAs of environmental pollutants.2 Approaches such as transcriptomics, proteomics, and metabolomics provide a holistic overview of the molecular changes underlying physiological processes affected by a toxicant without biases to anticipated pathways.3 Discovery of perturbed pathways, whether these be directly related to the MOA or downstream secondary stress responses, help to tailor toxicity testing strategies toward relevant end points. Additionally, omic tools can potentially help screen for biomarkers of sublethal toxicity in environmental monitoring. However, conclusive linkages must be made between molecular responses, such as gene expression and changing metabolite levels, and adverse outcomes on individual or even population levels.2 Daphnia sp. have become important model organisms in both ecology and toxicology due to their wide geographic distribution, central role in freshwater food webs, ability to adapt to a range of habitats,4 and sensitivity to anthropogenic chemicals.5 They have r 2011 American Chemical Society
been widely incorporated into toxicity testing by the U.S. Environmental Protection Agency (U.S. EPA) and the international Organisation for Economic Cooperation and Development (OECD).4 Furthermore the Daphnia Genome Consortium (http://daphnia.cgb.indiana.edu/) was established to coordinate the development of genomic tools for these species. Several studies have demonstrated the value of transcriptomics to improve our understanding of chemical toxicity in Daphnia sp., particularly in response to cadmium.610 While these studies reveal which genes are affected, a more complete understanding of the molecular changes underlying the physiological responses requires the combination of several molecular hierarchies. The combination of transcriptional and metabolic measurements has the potential to reveal both regulatory processes as well as those more closely linked to organism fitness.11 Metabolomics is increasingly used in (eco)toxicology12,13 to measure the small molecule composition of an organism, or Received: February 17, 2010 Accepted: March 2, 2011 Revised: February 28, 2011 Published: March 18, 2011 3710
dx.doi.org/10.1021/es1037222 | Environ. Sci. Technol. 2011, 45, 3710–3717
Environmental Science & Technology compartment within, enabling a nonbiased investigation of the metabolic response to chemical stress. Given the importance of Daphnia sp., the limited number of published daphnid metabolomics studies is surprising. Metabolomic responses in whole daphnids have been investigated after copper exposure using Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS)14 and recently, following exposure to PAHs using nuclear magnetic resonance (NMR) spectroscopy and GCMS.15 Metabolomic studies in mammalian toxicology typically analyze biofluids such as plasma or urine.16 Biofluids benefit from being intimately linked to the cellular function of the organs which are perfused but lack the molecular complexity of wholebody homogenates. Despite these advantages, in environmental metabolomics invertebrate biofluids have rarely been studied by NMR spectroscopy and never by MS. The few studies include identification of potential biomarkers of toxicity in earthworm (Eisenia veneta) coelomic fluid,17 investigation of metabolic changes in hemolymph during tobacco hornworm (Manduca sexta) development,18 and the study of withering syndrome in abalone hemolymph (Haliotis rufescens).19 These studies demonstrate the potential of invertebrate hemolymph to provide novel biochemical information, warranting a metabolomics investigation of D. magna hemolymph as a potentially powerful approach for toxicity testing. Cadmium (Cd) is a nonessential metal used as a model toxicant in several ecotoxicogenomic studies. Transcriptomic studies in Daphnia sp. have identified classes of genes that correlate with Cd-induced sublethal effects to growth and reproduction,7,20 energy budgets,8 and population growth,9 all at concentrations 10100 times lower than those causing acute toxicity. This chemical was selected here based upon the wealth of gene expression and toxicological data for Cd-exposed Daphnia sp. Our study therefore comprised of three principal objectives. First, to assess the feasibility of sampling small hemolymph volumes from individual adult daphnids and to measure their metabolic profiles by FT-ICR MS and NMR spectroscopy. Second, to determine which metabolic pathways are significantly perturbed in hemolymph from Cd-exposed daphnids, using FTICR MS. And finally, to measure the whole-body transcriptomic responses of daphnids to Cd and, using both our own and published gene expression data, to create a model of the early molecular changes associated with cadmium toxicity in D. magna.
’ MATERIALS AND METHODS Maintenance and Chemical Exposures of Cultures. Daphnia magna were maintained using standard protocols described in the U.S. EPA Acute and Chronic Testing Methods.21 For metabolomics, the exposure was performed with adult (following release of second brood, 1014 days old) D. magna placed in COMBO media21 for 24 h. Animals were exposed to 0 or 18 μg/ L cadmium sulfate (n = 18 per group), the latter corresponding to 1/10 LC50, a sublethal concentration resulting in impaired reproduction during chronic exposures.20 For transcriptomics, the exposure was conducted using the same experimental setup, strain of D. magna and experimentalist, except (i) six replicate exposures were performed with twelve individuals, and (ii) moderately hard reconstituted water (MHRW) replaced the COMBO media (we confirmed this did not alter the acute toxicity of Cd: 24-h LC50 in COMBO media was 180 μg/L (95% CI of 160200 μg/L) and 24-h LC50 in MHRW was 143.6 μg/L (95% CI of 122169 μg/L)). Additional culturing details and
ARTICLE
water quality parameters are available in the Supporting Information. Following 24-h exposure, D. magna were collected and hemolymph or RNA was extracted as described below. Hemolymph Extraction. Hemolymph was extracted as described by Mucklow et al.22 Adult daphnids were removed from the exposure vessels individually and blotted dry. The heart of each D. magna was pricked with a needle, and 1 μL of hemolymph was collected from the body cavity. Hemolymph was flash-frozen in liquid nitrogen and stored at 80 C until analysis. Mass Spectrometry-Based Metabolomics. To achieve sufficient material for analysis, hemolymph from three animals was pooled for each biological replicate (i.e., 18 control and 18 Cdexposed daphnids resulted in n = 6 biological replicates per group). Samples were prepared by methanol precipitation. Specifically, 30 μL of 80:20 methanol:water (HPLC grade) containing 20 mM ammonium acetate was added to each 3-μL hemolymph sample, vortexed, and stored at 20 C overnight. Samples were centrifuged (5000-g, 10 min, 4 C) prior to analysis. An ‘extract blank’ was similarly prepared but without addition of hemolymph. Metabolite profiles were measured in negative ion mode using an FT-ICR mass spectrometer (LTQ FT, Thermo Scientific, Bremen, Germany) with a Triversa direct infusion nanoelectrospray ion source (Advion Biosciences, Ithaca, NY, USA). Nanoelectrospray conditions and data acquisition methods were as described previously14 with each sample analyzed in triplicate utilizing the SIM-stitching approach from m/z 70500.23 Mass spectra were processed using customwritten software, as previously described,14 including SIMstitching,23 addition of missing values, probabilistic quotient normalized,24 and, specifically for the multivariate analysis, generalized log transformed.25 Peaks were identified with an empirical formula(e) based upon their accurate mass, and, where available, natural abundance 12C/13C intensity ratios were used to calculate the number of carbon atoms. In many cases each empirical formula was putatively assigned26 a metabolite name from selected pathways in the KEGG database corresponding to central metabolism (see Table S3), using “single-peak search” in the MI-Pack software.27 Principal components analysis (PCA) was used to assess the similarities and differences between metabolic profiles of the treatment groups using PLS_Toolbox (Eigenvector Research, Wenatchee, WA, USA) in MatLab (The MathWorks, Cambridge, UK); the PC scores were tested for significance using Student’s t test (in MatLab, using the Statistics Toolbox). Additionally, using the non-generalized log transformed data, the significance of the intensity changes in each individual peak was examined using Student’s t test (p-values adjusted for False Discovery Rate (FDR)
