One challenge facing ecotoxicology is to understand the mechanism of action of toxicants on living organisms (Snape et al. 2004
). Such knowledge would help to develop predictive simulation models of toxic effects, to link molecular biomarkers with population-level effects, and then to anticipate ecologic risk assessment issues for new chemicals. Gene expression profiles represent the primary level of integration between environmental factors and the genome, providing the basis for protein synthesis, which ultimately guides the response of organisms to external changes. Thus, the analysis of gene expression changes is a powerful tool both to diagnose major stressors in natural populations and to analyze the mechanisms of such stress responses.
Using gene expression profiles in ecotoxicology requires careful planning of experimental protocols that should take into proper account possible sources of variations in gene expression in natural populations. In fact, differences in gene expression due to sex, genotype, age, and intrinsic genetic variability can be quite substantial (Jin et al. 2001
; Oleksiak et al. 2002
; Ranz et al. 2003
; Townsend et al. 2003
DNA microarray technology has been applied extensively to the analyses of natural and anthropogenic factors in yeast for which whole-genome chips have been available for a few years (Causton et al. 2001
; Gasch et al. 2000
; Momose and Iwahashi 2001
). Causton et al. (2001)
analyzed how the whole genome of yeast is remodeled in response to environmental stressors such as temperature, pH, oxidation, and nutrients. The stress response was dependent on the level of the stress and showed an additive effect for multiple stressors. Similar results were found when using different stressors such as temperature shock, amino acid starvation, nitrogen source depletion (Gasch et al. 2000
), and cadmium (Momose and Iwahashi 2001
). The same approach has been used to characterize the alteration of gene expression in yeast induced by the pesticide thiuram (Kitagawa et al. 2002
). The results obtained for stress response in yeast likely will provide a reference frame for similar experiments with more complex organisms.
Custom-made microarrays have been used to understand responses to endocrine modulators in zebrafish (Hoyt et al. 2003
). Zebrafish embryos were exposed in vitro
to the environmental contaminant 4-nonylphenol (4NP). The gene expression profiling (using a custom-made microarray with 230 genes) identified a set of 9 genes associated with the function of estrogen response that is indicative of embryo exposure to 4NP even at low concentrations. A similar approach has been used to study the gene expression profiling in response to environmental stressors in the typical plant model organism Arabidopsis thaliana
. Using a cDNA microarray containing about 7,000 genes, Seki et al. (2002)
determined the expression profiles under drought, cold, and high-salinity conditions. Their analysis revealed a subset of 53, 277, and 194 genes that were differentially expressed > 5-fold after cold, drought, and high-salinity treatments, respectively. A set of 22 stress-inducible genes was found to respond to all three stress conditions. In a similar study the oxidative stress caused by high ozone concentrations has also been analyzed in A. thaliana
(Ludwikow et al. 2004
) with DNA microarray. A review of the applications of DNA microarrays for expression profiling under stress conditions in A. thaliana
has been recently published by Seki et al. (2004)
of the Riken Genomic Sciences Center in Kanagawa, Japan.
The application of gene expression profiles is not limited to model organisms for which the complete (or almost complete genome) is available. Several strategies are available to apply a genomic approach to species for which only a limited amount of genomic information is available. One possibility is heterologous hybridization. In fact, because of the length of the probes, cDNA microarrays can be used in heterologous hybridizations across strains and closely related species as long as sequence divergence is limited for a given gene (Rise et al. 2004b
). For example, Hittel and Storey (2001)
have used this approach to study the molecular basis of traits, such as hibernation, not present in model species. More recently, Renn et al. (2004)
have used heterologous hybridization to study gene expression profiling across a wide range of different species of African cichlid fish.
Another possible approach consists of identifying stress-induced genes using special techniques based on PCR, such as differential display PCR, suppressive subtractive hybridization PCR, and representational difference analyses. The application of these techniques to ecotoxicology has been reviewed recently by Snell et al. (2003)
Gracey et al. (2001)
used cDNA microarrays to identify hypoxia-induced genes in a nonmodel fish for which sequence data were unavailable. Their analysis revealed that although some changes in gene expression mirror the changes that occur in mammals, novel genes are differentially expressed in fish and tissue-specific patterns of gene expression occur during hypoxia.
Larkin et al. (2003)
described an expression profiling model system for endocrine-disrupting compounds that mimic estrogens. The research group created a gene array by cloning 30 genes from sheepshead minnows. The genes had been previously identified by differential display reverse transcriptase PCR, a method that screens thousands of RNA messages to identify genes that are turned on or off by specific treatments. They treated the fish with a constant concentration of weak and strong environmental estrogens and then determined which genes were differentially expressed in the livers of treated and control fish. They found a group of genes that were up-regulated by all the tested compounds, while other genes showed differential expression only in response to a specific compound. Exposure to different concentrations of the strong estrogen 17α-ethynyl estradiol revealed that changes in gene expression levels are dose sensitive and that exposure thresholds vary for different genes. A similar approach has been used to identify alterations in gene expression due to exposure to androgen hormones in largemouth bass fishes (Blum et al. 2004
Williams et al. (2003)
used a cDNA microarray-based approach to analyze the expression level changes of recognized biomarkers in a relevant fish species, European flounder (Platichthys flesus
). They arrayed 160 genes, of which 110 were already documented in the literature as biomarkers of toxic response in fishes and mammals. Five adult males and five adult feral females P. flesus
were caught from the Tyne (polluted) and the Alde (unpolluted) estuaries in the United Kingdom. Gene expression analysis revealed that 11genes were expressed differently in males between the Tyne and Alde. Such differences were not statistically significant in females, probably because of interindividual variations. Vitellogenin levels differed radically among the female fish, suggesting that their reproductive cycles were at different stages.
Despite the lack of extensive genomic data, invertebrates are the subject of increased interest. Because of their characteristics, estuarine amphipods typically are used to assess the ecologic risk associated with contaminated sediments. Perkins and Lotufo (2003)
isolated several genes from Leptocheirus plumulosus
and developed a quantitative assay to measure the effects of water exposure to 2,4,6-trinitrotoluene and phenanthrene on gene expression. They found that expression of the genes for actin and a retrotransposone element, hopper, were dependent on the exposure and tissue concentrations of those chemicals. Diener et al. (2004)
have optimized a protocol for differential display PCR to investigate gene expression in Daphnia magna
. Their protocol requires submicrogram amounts of total RNA and fewer than 10 animals and thus could provide a significant technical improvement for gene expression analyses of toxicant exposure.
Several efforts are focusing on the detection of pathogen infection in different animal species. Panicker et al. (2004)
developed a gene array for detection of pathogenic Vibrio
species, which are natural inhabitants of warm coastal waters and shellfish. Recently, microarray analysis has also been applied successfully to identify molecular markers of pathogen infection in salmon (Rise et al. 2004a
). This analysis identified transcripts induced and repressed by the pathogen, thus providing insights into the host response to the infection and into the mechanisms used by the pathogen to evade such response.