Recent advances in genomics and bioinformatics are revolutionizing the process of discovering genes whose regulation has important consequences to the fitness of individuals 12. These resources for genetic investigations include functional genomic tools that compare the expression profiles of thousands of genes under different conditions. Microarrays are proving to be a particularly important modern resource for identifying genes in context of their complex regulation 3. Until recently, the application of microarrays has largely focused on studies from a select set of organisms (Saccharomyces cerevisiae, Caenorhabditis elegans, Drosophila melanogaster, Danio rerio, Mus musculus) that constitute some of the most well characterized laboratory models in the life sciences. While these model organisms are well suited for developmental, cellular and molecular studies – contributing a staggering amount of biological knowledge – it is difficult to relate environmental controls of gene regulation of these organisms (phenotype) to higher-level (population) responses because so little is known about their ecology. This chasm presents a challenge for toxicological genomic applications, especially those related to environmental toxicology where the goal is often to identify population and ecosystem-level responses in the context of environment change. Thus, the aim of discovering genes that are expressed as a function of ambient conditions (therefore, anchoring potential genetic biomarkers to biological/ecological functions) requires applying functional genomic approaches to keystone species with accessible natural populations and tractable ecologies, such as the ubiquitous aquatic micro-crustacean Daphnia pulex.

Commonly known as water fleas, Daphnia are familiar and ubiquitous inhabitants of ponds and lakes throughout the globe and have been the focus of study by limnologists for well over a century 456. As a dominant member of the planktonic community, Daphnia play a central role in aquatic food webs, serving as the primary grazers of algae, bacteria and protozoans, and as the primary forage for fish. As a result, Daphnia are long recognized as a sentinel/indicator species in freshwater ecosystems and they are routinely used to determine the toxicity of aqueous solutions and to gauge the quality of inland waters 7. Therefore, the gross-level responses of Daphnia to a number of environmental pollutants are well characterized 8, and these responses are referenced by environmental protection agencies (e.g., United States Environmental Protection Agency, Environment Canada, Organization for Economic Co-operation and Development) to define regulatory limits 9, monitor the quality of industrial and municipal effluents 1011 and estimate the risk of environmental toxins on natural environments. For these reasons, the present study extends the "road-map" for applications of DNA microarrays to studies of non-model organisms described by Gracey et al. 12 and later by others (reviewed in 1314), by characterizing the response of this critical aquatic indicator species to environmental stress. These investigations utilize D. pulex; a freshwater crustacean species that is ubiquitous throughout North America 15, sensitive to metals and metal mixtures 16, and supported by a major genome sequencing project 17. They also complement recent efforts by others to develop a microarray platform for D. magna 18; 19; 20.

We focus our present investigations on cadmium as a model environmental stressor, but the technologies described are pertinent to a much wider range of ecological and toxicological applications. Cadmium is a ubiquitous environmental contaminant 21. It ranks eighth on the Agency for Toxic Substances and Disease Registry 22 list of the top 50 priority pollutants, is one of the most common contaminants found in the U.S. EPA Superfund sites 23, and is highly toxic to aquatic life. Cadmium is extremely persistent in the environment and, as a result, bioaccumulates within food webs 24. In aquatic animals, cadmium is a potent calcium antagonist that disrupts calcium uptake and homeostasis 252627. Cadmium also induces oxidative stress, resulting in lipid peroxidation, damage to membranes, impaired cellular functions, and tissue damage 28. It is a Class B metal that has ten outer shell electrons in the d orbital and is highly reactive, tending to preferentially form covalent bonds with S>N>P>O 2930. Thus, complexation with anions controls exposure in the water column 31 and ultimately toxicity in the animal 32. Within many organisms, the major ligands for cadmium are small metal binding proteins known as metallothionein (MT) 3233.

The metallothioneins are a family of unusual and highly conserved small cytosolic proteins, characterized by their low molecular weight (i.e., 6000–7000 Daltons), lack of secondary structure in the absence of metal ligands, absence of histidine or aromatic residues, and high cysteine content (typically 30–33% of the protein). The MTs principally bind metals from groups 1B and 2B of the Periodic Table of Elements (e.g., cadmium, copper, mercury, silver, zinc; 3435), and in so doing form two metal-binding domains. The metal binding properties of MT are conferred by the large number, spacing and metal coordination of the cysteine residues within the protein. It is generally recognized that the physiological functions of MT are to regulate the intra-cellular concentrations of essential metals such as copper and zinc; to activate and deactivate zinc-regulated proteins; and to scavenge free radicals 3536. However, MT can also confer protection from metal poisoning by binding certain free metal ions or undergoing exchange reactions with metals bound to other ligands 3537. The synthesis of MT is also regulated by metals with MT mRNA expression increasing in response to elevated metals in tissues 3839. Because of its high specificity and sensitivity for metal induction, MT levels have been successfully used to diagnose copper, zinc, and cadmium exposures in numerous studies ranging from temperate freshwater environments to tropical marine systems 40. While MTs have been characterized in several Malacostraca crustacean species 41, with the exception of partial cDNA sequences of two MTs from D. magna 19 there is little sequence information available for these genes and their regulatory regions in the Branchiopoda, such as Daphnia.

We report a series of studies designed to test whether the exposure to sub-lethal chemical stressors results in identifiable changes in gene expression, exposing genes that respond to these conditions and providing a means of identifying potential biomarkers of response to specific exposures. Here, we developed a cDNA microarray platform for D. pulex to investigate differences in gene-response profiles for this aquatic indicator species following environmental perturbation by cadmium. Expression profiling successfully aided in the discovery of genes regulated in response to cadmium exposure, including the important metal biomarker, MT. Gene-response profiling provided mechanistic information that related to the observed cadmium-induced phenotypes and population-level responses. Finally, we identified and characterized the primary structure of the D. pulex MT sequences in context of their genomic architecture, translated sequence and phylogenetic relationship to other crustacean MTs.

In this paper we describe studies with newly developed D. pulex cDNA microarrays investigating the expressed mRNA responses of daphniids to the environmental stressor cadmium. Although microarrays have been successfully used to identify patterns of genes responding to metal exposure in organisms for which there is an abundant amount of sequence information 5253, it is often difficult to relate the identified genes to expressed phenotypes. These challenges exist in part because the environmental contributions to phenotype in most common laboratory species are poorly understood. However, as noted by Gracey et al., 12, abundant sequence data are not a pre-requisite for gene-expression profiling. Thus, one aim of the present study was to develop and apply microarray technologies to the ecologically tractable aquatic micro-crustacean, D. pulex. These studies revealed patterns of gene response that provide insight into the biological and potentially toxicological responses to this important environmental contaminant. The data emphasize the potential of this approach as a tool for discovering genes regulated by an environmental stressor, but in a species for which gene expression profiles can be interpreted in context of individual and population-level effects. Associations with cadmium induced phenotypes, identification of genes reported in the literature to be regulated following cadmium exposure, and independent validations via Q-RT-PCR provided support for their utility in this role. In addition, these studies led to the discovery D. pulex MT mRNAs and gene sequences, providing an important new biomarker for Daphnia studies, while also extending the phylogeny of this class of genes. The successful identification of the metal binding protein, MT, is highlighted, not only due to its critical biological functions, but also because the blind microarray approach provided advantages over cloning methods that require significant sequence similarity to isolate genes.

We identified the acute (48-h) and chronic (21-d) responses of our D. pulex isolate to cadmium to better define sub-lethal exposure concentrations and cadmium-induced phenotypes. The present study is in agreement with those of others 1621 that indicate D. pulex is one of the most acutely sensitive aquatic species to cadmium 54. Reported LC50 values are comparable to data published by others (i.e., range from 46 to 90 μg Cd/L; 16). As others have noted, chronic exposure of D. pulex to cadmium reduced individual fitness parameters, such as length and lipid-ovary indices, and inhibited population endpoints, such as number of clutches, cumulative reproduction, and per capita birth rate (Table 1; 5556). These endpoints have been quantitatively linked to population success 57.

For example, storage of lipids, vitellogenisis, and ecdysis (i.e., molt) are physiologically and chronologically associated with reproductive success and survivorship 57. Adipocytes are incorporated into the ovary during egg development providing energy for reproduction. Parthenogenetic embryos are then transferred into the brood chamber that is contained within the carapace for further development and the process ends with newborn Daphnia released into the water column during the molt. Also, molt directly relates to growth, as Daphnia size is constrained by the carapace. Therefore, decreased length, lowered lipid-ovary indices, and fewer clutches comprised of fewer offspring can be explained by impaired vitellogenisis and/or ecdysis. Cadmium has been reported by others to inhibit ecdysis 58 and vitellogenisis 5960 in arthropods.

Cadmium may also decrease daphniid size by limiting calcium intake. Calcium content is directly related to the mass of the carapace and, subsequently size 61 and cadmium is known to impair calcium uptake and metabolism via substitution 62. However, several studies have shown that while the period of increased calcium flux during molt can result in increased cadmium uptake 6364, cadmium does not affect calcium accumulation or content 64. Yet interestingly, our studies revealed two cDNAs that responded to cadmium that represent putative calcium binding proteins (singlet 97 and contig 220), suggesting that cadmium may affect calcium regulatory pathways in more subtle ways that contribute to its overall effects.

The present study is one of the first to apply microarray technologies to D. pulex. These "blind arrays" identified 99 elements for which gene-expression was positively regulated and 30 elements that were negatively regulated by cadmium. However, sequence analysis revealed that these probes included several redundant features. When they were assembled into EST contigs to isolate unique features, 30 up-regulated and 12 down-regulated genes were identified (Table 2). Given the redundancy that existed on the array and the 'blind' approach-employed, the observation of redundant, cadmium-regulated elements (e.g., cuticle protein) was reassuring. In addition, the number of unique elements regulated following exposure to cadmium only represents a very small percentage (~2%) of the estimated 1,550 unique elements on the array, which is similar to other microarray studies involving sub-toxic cadmium exposures (2%, 52; 1%, 65; 1–3%, 66 and as suggested by Andrew et al.52 is consistent with expectations for low, non-toxic exposures that tend to induce specific pathways.

The D. pulex microarrays and their genomics database have proven to be insightful experimental tools for identifying genes regulated by cadmium. Yet, relating the gene sequence to putative gene function is made difficult by the excessively large phylogenetic distances between Daphnia and its closest relatives among the classical genomic model systems. For example, the best model system to Daphnia with extensive functional genomic information is Drosophila, which last shared a common ancestor some 600 M years ago. There are also few ESTs that share sequence conservation with proteins in public databases; crustacean proteins represent only 0.1% of 6.9 million records in the NCBI taxonomic database. Of the 42 unique cadmium-regulated genes identified, only three were homologous to known Daphnia genes. Likewise, almost 36% of the elements identified were ESTs showing no sequence similarity with known proteins. Gracey et al.12 reported similar successes with 'blind' microarrays, as approximately 40% of identified elements were unidentifiable ESTs. However, as these authors noted, novel ESTs may represent untranslated regions of previously identified genes or multiple distinct regions of unknown genes. These possibilities cannot be dismissed; we estimate that 1/4 of the orphan genes are unknown because of sequences not extending into recognizable functional domains of the gene 67. The remaining 27 elements regulated in response to cadmium exposure on the microarray were putatively identified by similarity with known proteins.

The genes regulated in response to cadmium exposure identified in these experiments provided several measures of the utility of this approach for gene discovery and its reliability in identifying biologically sensible patterns of gene regulation. For example, many of the genes responding to cadmium exposure were part of common physiological pathways (i.e., ecdysis, metal detoxication) and few were indicative of general stress response (e.g., heat shock proteins, heme-oxidase) or overt cellular toxicity (e.g., housekeeping genes).

Ecdysteroid-responsive genes and other molt related genes that included chitinases 1 and 2, chitotriosidase, BcNCP 14.9, cuticle proteins 1 through 5, and singlet 251, comprised the majority of genes regulated in response to cadmium exposure (Figure 8, Table 2). This finding is critical given that altered expression of these genes, which are associated with the exoskeleton, provides a direct mechanistic link with demographic endpoints and individual fitness parameters that indicated ecdysis (i.e., molt) is impaired in Daphnia following cadmium exposure (Figure 1, Table 1). Deep sequencing of cDNA libraries constructed from RNA isolated from cadmium-exposed D. pulex provided support of these findings (Colbourne JK, personal communication). Cadmium derived libraries were enriched with genes that were identified as structural constituents of the cuticle. This observation suggests that ecdysis and molt related regulatory pathways are generally influenced by cadmium, which has been observed by others 6364. It is interesting that short-term exposures to cadmium of the Daphnia used in the array experiment provided patterns of response that were meaningful in interpreting demographic experiments that involved longer-term exposures to lower cadmium concentrations. This observation raises the possibility that an approach such as this could provide biomarkers that serve as early indicators of chronic exposures, which is an area we are currently exploring, but beyond the scope of the current manuscript.

These experiments also identified altered expression-levels of genes known to be regulated by cadmium (Figure 8). These included the metal binding protein, MT, which plays a central role in cadmium detoxication and is known to be induced by cadmium 194168. In addition, the microarrays identified a gene responsible for oxygen transport and iron metabolism in Daphnia, hemoglobin 69. Expression of hemoglobin is known to be induced by anoxia, limiting iron supplies, and cadmium 197071; unpublished results). As discussed above, the arrays also identified structural components of the cuticle and associated regulatory components (i.e., chitinase) and chitinase activity has been shown to positively correlate with cadmium concentrations in previous studies 1972. Likewise, opsin was found to be regulated in response to cadmium on the microarrays, and cadmium has been shown to interfere with photo-behavior in daphniids, which is a complex genetically linked trait involving kairomone signalling, photoreceptor detection (e.g., opsin, 73), and negative phototaxis 7475. At the cellular-level cadmium is known to induce oxidative stress 28. Glutathione S-transferase (GST), a phase II enzyme that catalyses the conjugation of reduced glutathione with electrophilic compounds, is an important antioxidant defence enzyme that works to detoxify the products of oxidative stress 76and was up-regulated on the array. GST gene transcription 19 and enzyme activity 77 have been shown by others to increase following exposure to cadmium.

Metallothionein induction is the principal adaptive response associated with organism survival during exposure to elevated cadmium concentrations 4178. The discovery of D. pulex MTs highlight the advantages of this approach in identifying genes regulated by cadmium. While the conservation of cysteine residues ensures thiolate cluster formation, metal binding and provides both the structural and functional basis to identify these genes as MTs, cloning methods such as hybridization or amplification using degenerate oligonucleotides were ineffective because of its lack of similarity with other known crustacean MT sequences (Figure 6). The present study is the first to provide genomic DNA sequence information for a non-malacostracan crustacean, the first to report the promoter region of any crustacean MT, and one of the few to identify the metal-responsive element (MRE) consensus core sequence in any invertebrate. MREs were identified in the 5'flanking region of D. pulex MTs (Fig 6), Both Dpu Mtn1 and Dpu Mtn3 contained MREs within 200 bp of the transcriptional start site. The promoter region of the sea urchin MT gene was found to contain two MREs residing within 300 bp of the start of transcription 7980. There was only genomic (primary) sequence provided for one crustacean MT (i.e., green shore crab, Carcinus maenas) in the entire NCBI repository, but it lacked any flanking sequences (i.e., upstream and downstream regulatory regions). Green shore crab MT has a structure common to Dpu Mtn1 and Dpu Mtn2 with three exons separated by two introns 81. In fact, similar architectures have been reported for the urchin, trout and several mammals. Also, D. pulex have much smaller intronic regions (Figure 6), likely a consequence of its compact genome size (~200 Mb). Although the architectures of these genes were similar to others, little sequence similarity was observed within their coding sequence for any known MT.

The translated daphniid sequences (i.e. D. pulex; D. magna, Poynton et al.19) also showed little similarity to other MT genes (Table 4, Fig 7), including other crustaceans. This includes the N-terminal crustacean motif (P- [GD]-P-C-C-x(3 or 4)-C-X-C;4148. However, since daphniid MTs represent the first MTs isolated from a non-malacostracan crustacean, the lack of phylogenetic representation in the NCBI repository could account for some these differences. There was considerable similarity in terms of number and distribution of cysteine residues. All crustacean MTs sequenced to date are comprised of 18 cysteine residues with the exceptions of the copper MT isolated from the green crab 41, D. magna MT 19 and Dpu Mtn3 identified in this investigation that each contain 19 cysteines. The conservation of cysteine residues is expected, given that they are responsible for a great deal of the tertiary structure and metal-binding functions of the protein 78. In fact, Valls et al.41 has suggested using the binding properties (e.g., stoichiometry of metal- MT species) to strengthen current classifications that are based on phylogeny.

The current classification procedure for MTs 82 is based on phylogenetic relationships rather than amino acid composition and cysteine content. This nomenclature protocol was established to better identify related functional properties in which MTs are divided into families based on evolutionary conservation. In this classification scheme, crustaceans constitute a single family, given the name number three, which is comprised of three sub-families (crustacean one, c1; crustacean two, c2; crustacean, c), where c1 and c2 genes each constitute monophyletic clades. The third family termed, c, is reserved for crustacean MTs that are different from these others. Daphnia pulex MTs do not align well with malacostracan type 1 or type 2 MTs and thus, according to this nomenclature are designated Dpu Mtn1, 3, c; Dpu Mtn2, 3, c; and Dpu Mtn3, 3, c.

In summary, the development of D. pulex cDNA microarrays and associated sequence information has provided a useful first generation functional genomic tool for examining biological responses of this key sentinel species to environmental agents and other stressors. Treatment of D. pulex with sub-toxic levels of cadmium revealed a specific pattern of gene expression changes that provide new insights into their biological and toxicological responses to this environmental contaminant. Moreover, microarray responses to cadmium led to the discovery of D. pulex MTs, whose gene structure and cysteine content clearly place it in this gene family, but whose sequence divergence reveals that classical cloning and sequencing techniques based on similarity were likely to fail. Further identification of D. pulex genes that are responsive to various experimental treatments through use of these genomics tools will provide new and important insights into their biology. Advances in Daphnia genomics will enable the further development of this species as a model organism for a wide variety of biological investigations.

JRS, JKC, JCD, CYC, CLF and JWH conceived the study and designed the experiments and subsequent analyses. JRS and SPG performed cadmium exposures and toxicity assays, JRS and JCD conducted the microarray experiments and THH performed the statistical analyses. JRS drafted the manuscript and all authors contributed to, improved upon, and read and approved the final version.