M. T. Thébault, Laboratoire des Sciences de l'Environnement Marin, UMR-CNRS 6539, Institut Universitaire Européen de la Mer, Université de Bretagne Occidentale, Place Nicolas Copernic, 29280 Plouzané, France Fax: +33 2 98 49 86 45 Tel: +33 2 98 49 86 12 E-mail: firstname.lastname@example.org
Groups of oysters (Crassostrea gigas) were exposed to 25 °C for 24 days (controls to 13 °C) to explore the biochemical and molecular pathways affected by prolonged thermal stress. This temperature is 4 °C above the summer seawater temperature encountered in western Brittany, France where the animals were collected. Suppression subtractive hybridization was used to identify specific up- and downregulated genes in gill and mantle tissues after 7–10 and 24 days of exposure. The resulting libraries contain 858 different sequences that potentially represent highly expressed genes in thermally stressed oysters. Expression of 17 genes identified in these libraries was studied using real-time PCR in gills and mantle at different time points over the course of the thermal stress. Differential gene expression levels were much higher in gills than in the mantle, showing that gills are more sensitive to thermal stress. Expression of most transcripts (mainly heat shock proteins and genes involved in cellular homeostasis) showed a high and rapid increase at 3–7 days of exposure, followed by a decrease at 14 days, and a second, less-pronounced increase at 17–24 days. A slow-down in protein synthesis occurred after 24 days of thermal stress.
The fluctuating thermal nature of the marine environment induces physiological changes in ectotherms that require molecular and gene expression adjustments . Comparative gene expression studies can be used to characterize these adjustments and lead to a better understanding of organismal responses to environmental change. Gene expression datasets can be clustered into groups of genes that represent different compartments of cellular function, and changes in the expression of genes from these clusters can be used to formulate hypotheses as to how different tissues and whole organisms respond to particular biotic or abiotic stresses. Few studies have addressed changes in gene expression in response to temperature variation on marine organisms. Alterations in gene expression have been observed in fish acclimated to constant temperatures and then exposed to daily temperature fluctuations  or to a strong heat stress . However, few molecular investigations have focused on the thermal stress response in marine invertebrates [4,5], particularly in the context of global changes and the potential effects on marine invertebrates [6,7].
The Pacific oyster Crassostrea gigas is a eurythermic bivalve mollusc that colonizes most of the western coast of Europe. This species prefers sheltered estuarine waters, where it is found in intertidal and shallow subtidal zones. Within their geographic range, oysters typically experience and respond to seasonal temperatures ranging from 4 to 24 °C . In the coldest regions inhabited by C. gigas, such as Brittany, France, summer water temperatures only occasionally reach 21 °C for short periods of up to a few days. Recent studies on C. gigas suggest that complex interactions between temperature and food quality and quantity affect gametogenesis [9,10]. Reproductive development is temperature dependent in C. gigas and typically occurs between February and September along the western Atlantic coast of France . Spawning occurs at a minimum seawater temperature of ∼ 19 °C [12,13]. One study focused on the expression of heat shock proteins (HSPs) in C. gigas, at the molecular and physiological level, in response to heat stress throughout the year .
The aim of this study was to improve our knowledge of differentially expressed genes in C. gigas exposed to a temperature slightly above the upper natural water temperature in the area. Our study is the first to apply an overall genomic approach to the study of the response of C. gigas to prolonged heat stress. We used animals outside the season of reproductive development and spawning in order to measure temperature stress without the onset of reproduction. Using suppression substractive hybridization (SSH), we identified genes that were up- and downregulated 7–10 and 24 days after transfer from 13 to 25 °C. Subsequently, genes likely to be associated with thermal stress were quantified using quantitative real-time PCR.
Suppression subtractive hybridization
SSH libraries were constructed from pooled gills and mantle of C. gigas after 7–10 and 24 days of exposure to different temperature treatments. The search for homology using the blastx program revealed 858 different sequences, of which 536 (∼ 62%) remain unidentified. Expressed sequence tags (ESTs) similar to genes potentially involved in a thermal response were subsequently clustered into 15 distinct functional categories: cell differentiation (including cell migration, adhesion, proliferation and apoptosis), cellular communication (including signal transduction), cellular stress (including inflammation and immune response), cytoskeleton and cell structure (including cellular matrix and cellular trafficking), detoxification, energetic metabolism, lipid metabolism, receptors and channels, regulation of nucleosides, nucleotides and acid nucleic metabolism, reproduction, respiratory chain, transcriptional processing, translational and post-translational processing, general metabolism and other functions and ribosomal proteins (Fig. 1 and supplementary Tables S1–S4). Among the 322 recognized protein-coding genes, 191 new sequences were obtained in C. gigas and 131 had been identified previously, of which 88 genes encode ribosomal proteins in both forward and reverse libraries. Among the newly known sequences, only one corresponded to a gene specific for mantle (mantle gene 4). No cellular signalling genes were identified in samples taken after 24 days. Cytoskeletal genes, translation and ribosomal genes were less abundant on warming. Only respiratory genes were more abundant on warming.
Gene expression patterns from different functional categories during temperature acclimation
Using real-time PCR we conducted a time-course study to compare transcript expression in oysters exposed to thermal stress at 25 °C relative to control animals maintained at 13 °C. Seventeen transcripts analysed after 0, 3, 7, 14, 17 and 24 days of exposure (Fig. 2) belonged to categories previously implicated in the stress response: (a) cell proliferation and differentiation: metastasis-associated protein 1 (MTA-1), Huntingtin-interacting protein K (HYPK), cystatin B, cathepsin L (CTSL, EC 184.108.40.206) (these proteins are known tumor markers) [14,15], QM protein (transcriptional control of cell differentiation and proliferation) [16,17], Ras family GTP-binding protein Rho1p (differentiation); (b) cellular stress: HSP70, HSP70 kDa protein 12A, HSP23, chaperonin-containing TCP1 (alternative name CCT) subunit 7, isoform b (chaperones), inhibitor of kappa light polypeptide (inflammation); (c) antioxidant defense: non-selenium glutathione peroxidase (EC 220.127.116.11); (d) metabolism of nitrogen and ammonia detoxification: glutamine synthetase (EC 18.104.22.168); (e) membrane fluidity: Δ9 desaturase (EC 22.214.171.124); (f) energetic metabolism: d-lactate dehydrogenase (d-LDH, EC 126.96.36.199, anaerobic metabolism), citrate synthase (EC 188.8.131.52, aerobic metabolism); and (g) translational processing (translation initiation factor eIF-2B delta subunit). Normalized expression data are summarized in Table 1.
Table 1. Expression patterns in gills (G) and mantle (M) throughout the experiment at 25 versus 13 °C. For each gene, + (or –) represents significant relative upregulation (or downregulation): +/− from 1.2- to 2-fold; ++/−− from 2- to 5-fold; +++/−− from 5- to 10-fold; ++++ > 10-fold. NS, not significant. *P < 0.05, **P < 0.01, ***P < 0.001.
In the gills, all transcripts selected in the forward SSH library at 25 °C, except citrate synthase, showed an initial expression peak at days 3–7, followed by a decrease at day 14, and then a smaller increase at days 17–24 at 25 °C compared with controls (13 °C) (Fig. 2A and Table 1). The most differentially expressed transcripts at 25 °C were HSPs, MTA-1 protein, chaperonin-containing TCP1 subunit 7, isoform b and d-LDH. Gene expression was less pronounced in mantle relative to the gills (Fig. 2). In the mantle, some transcripts (HSP70, HSP23, MTA-1 protein, Rho1p, Δ9-desaturase and glutamine synthetase) showed a peak of overexpression at days 14 and/or 17. In contrast to observations on gill tissue, mantle levels of HSP12A, MTA-1 protein, Rho1p, Δ9-desaturase and citrate synthase transcripts were significantly lower at 25 °C on day 3. Variation in the expression of d-LDH and chaperonin-containing TCP1 subunit 7, isoform b genes was not significant.
Different profiles were observed for transcripts selected from the reverse SSH library at 13 °C (Fig. 2B and Table 1). In gill tissue, inhibitor of kappa light polypeptide, non-selenium glutathione peroxidase, CTSL and translation initiation factor eIF-2B were highly expressed at days 3 and/or 7 at 25 °C relative to control (13 °C). CTSL, non-selenium glutathione peroxidase and translation initiation factor eIF-2B were downregulated at day 24 in gills at 25 °C. Expression of the HYPK gene did not show much variation, although it increased significantly from the beginning to the end of the experiment. Cystatin B, inhibitor of kappa light polypeptide and QM protein (60S ribosomal protein) showed a biphasic expression pattern. In mantle cells, variation in expression levels of CTSL and inhibitor of kappa light polypeptide genes was not significant. Overexpression of cystatin B, non-selenium glutathione peroxidase and translation initiation factor eIF-2B transcripts was limited to a relatively short time-window at days 14 and/or 17. HYPK and QM protein were underexpressed at days 3, 7 and 24 and day 7, respectively.
Of 322 genes identified in this study, 191 partial sequences had not been identified previously in C. gigas. Of the 131 known genes, 88 encode ribosomal proteins and had been identified previously in C. gigas responding to environmental stresses such as hydrocarbons, pesticides and hypoxia [18–20]. Thus, their gene products appear to be important for metabolic adjustments during stress in general. However, the expression patterns observed were tissue specific with gills being more responsive than mantle. We hypothesize that the observed patterns reflect functional differences between these two tissues. A number of genes that were highly expressed in gills showed a biphasic expression pattern, consisting of a strong short- and a moderate long-term response. Moreover, after 7–10 days of exposure, we detected differential expression of a number of genes that encode elements of the transcription and translation machinery, including transcription factors, ribosomal proteins and elongation factors. After 24 days of exposure to elevated temperature, the differential expression profile was dominated by strong downregulation of genes involved in protein synthesis, such as the translation initiation factor eIF-2B delta subunit, suggesting a slowing of protein synthesis. These findings may suggest that transcription factors are regulated though a feedback mechanism, inducing their own inactivation . Changes in gene expression of organisms subjected to thermal stress are known to involve major adjustments in the expression of ribosomal genes and genes coding for proteins involved in RNA metabolism and protein synthesis. In fact, protein synthesis in marine snails was inactivated at temperatures approaching lethal values .
Under mild thermal stress at 25 °C, genes coding for antistress proteins were differentially expressed. Protection against cellular stress, inflammation and stimulation of immune function appear to be important components of responses to thermal stress. HSPs play an essential role in maintaining protein homeostasis during exposure to proteotoxic stressors . They function by interacting with stress-denatured proteins and preventing their aggregation and/or degradation . HSP induction may therefore have an adaptative value for organisms facing thermal stress and significant ecological and biogeographical implications for species distribution and their thermotolerance limits [22,25]. Tissue-specific de novo HSPs synthesis was induced in C. gigas following exposure to 25 °C, which is 4 °C higher than the highest sea surface temperatures recorded in its distribution range. Two HSPs (HSP70 and HSP23) were greatly and rapidly upregulated in gills but slightly less and later in mantle. One inducible and two constitutive isoforms of HSP70 are synthesized in the gills and mantle of C. gigas. The expression level of the constitutive forms increases after thermal stress, whereas the inducible one is expressed only after exposure to 32 °C . These results suggest that the overexpression of HSP70 we observed might correspond to the constitutive form. HSP23 is a small heat shock protein highly induced following stress. Small HSPs are differentially expressed between tissues and through the different stages of development . A third HSP, HSP12A (alternative name 150 kDa oxygen-regulated protein; ORP150), was induced in gill tissue only. This chaperone, located in the endoplasmic reticulum, plays an important role in maintaining cell viability in response to stress . Among other chaperones, the chaperonin-containing TCP1 (subunit 7, isoform b) presented the same expression pattern in response to heat stress in our study. This complex, involved in folding actin, tubulin and cyclin E, among other proteins , is also upregulated in response to chemical stress . Hence TCP1 may play an important role in the recovery of cells after protein damage, by assisting the folding of cytoskeletal proteins that are actively synthesized and/or renatured under these conditions. The upregulation of all of these chaperones confirms the severity of the thermal stress under our experimental conditions.
A number of genes encoding structural components of the cytoskeleton and proteins involved in contractile functions (including actin, tubulin myosin and profilin) were differentially expressed in C. gigas in response to prolonged heat stress, some were induced and some repressed. Rho1p, for example, encodes for a protein involved in numerous processes including actin filament organization and is expressed in response to environmental changes . In this study, Rholp was rapidly upregulated in gills and later in mantle during warming. These results suggest that extensive cytoskeletal reorganization occurs in response to heat stress, as reported for fish gills .
Furthermore, several genes associated with the regulation of cell homeostasis were differentially expressed in our study. Some genes were differentially expressed in both tissues, showing increased apoptotic/autophagic activity. Among these genes, MTA-1 was strongly expressed in gills during warming, whereas it was initially downregulated in mantle. CTSL, a highly potent endoprotease involved in lysosomal bulk proteolysis, was strongly expressed, but only in gills. The upregulation of CTSL, combined with the downregulation of its reversible binding inhibitor cystatin B, implies that active protein degradation was taking place in the gills upon warming to 25 °C. Two less well-known genes, putatively involved in proliferation and apoptosis, QM protein and HYPK were both differentially expressed between the tissues. QM protein, also known as ribosomal protein L10, is a transcription cofactor that inhibits activation of AP-1 transcription factors. QM protein is implicated in the conversion of a broad variety of extracellular signals generated by growth factors, tumour promoters or genotoxic drugs [16,32]. In the sponge, Suberites domuncula, QM protein expression was significantly higher in tissues undergoing induced apopotosis . On warming to 25 °C, upregulation of QM protein occurred rapidly in gills and later in mantle. HYPK, identified as a antiapoptotic protein , displayed the same expression pattern as QM protein in both tissues. These results suggest that, in C. gigas, these proteins are induced to prevent pathologies such as inflammation and tumorigenesis during prolonged thermal stress.
The cellular stress response has an energetic cost and control of the balance between ATP supply and demand in the ciliated gill may become altered during thermal stress. In C. gigas, stressors such as hydrocarbons, herbicides, parasite infection or hypoxia [18–20,35], affect the expression of genes involved in energetic metabolism and our results show that changes in transcript levels of a number of genes involved in metabolic regulation also occur in response to temperature. In gill tissue, prolonged heat stress resulted in the rapid induction of several ATP-generating enzymes including the tricarboxylic acid cycle citrate synthase, suggesting that there was a need for rapid aerobic ATP production. In the early phase of warming, the rapid induction of LDH in gills may indicate that anaerobic metabolism is required. The LDH that we identified was d-specific. Many systematic studies have shown that d- or l-specific LDHs are present in all invertebrate groups . We also observed that the glutamine synthetase gene was upregulated in both tissues, as previously observed in response to hydrocarbons, herbicides or hypoxia [18–20]. In vertebrates, glutamine synthetase occupies a central position in nitrogen metabolism and is linked to amino acid turnover, nitrogen detoxification, nucleotide biosynthesis and more generally to growth . Although the capacity for glutamine biosynthesis is generally weak or absent in molluscs , a recent study reported the accumulation of glutamine associated with an upregulation of glutamine synthetase in a bivalve species in response to aerial exposure . Glutamine synthesis may also be an ammonia detoxification mechanism in invertebrates.
Genes involved in fatty acid metabolism are expected to be affected by temperature. Among these, Δ9 desaturase has been extensively studied in numerous animal groups including mammals, chicken, fish and insects [40,41]. High temperatures typically increase membrane fluidity in temperate eurytherms . The upregulation of Δ9 desaturase that we observed in gills, and later in mantle, agrees with this pattern. A similar pattern was previously observed in C. gigas in response to experimental hypoxia, suggesting that the regulation of this enzyme may be affected primarily by oxidative stress . Recent studies on intertidal bivalves show that critical warming may exacerbate cellular oxidative stress . In many species, the increase in lipid peroxidation and reactive oxygen species concentration in cells following heat stress has already been shown to modify the activity of antioxidant enzymes such as GPx . Non-selenium glutathione peroxidase is involved in detoxification by reducing fatty acid hydroxyperoxides and H2O2. In our study, the level of the non-selenium glutathione peroxidase transcript appears strongly upregulated in gills during the first week of thermal exposure. A similar thermal stress response, associated with oxidative stress, has also been observed in other marine poikilothermic species including molluscs [43,46–48].
Our results represent the first stages of investigation into the molecular response of oysters to high temperatures, focusing on early winter, outside the gametogenesis period. Future efforts will focus on the search for functional polymorphism in some of the genes potentially regulated by temperature in oyster populations located at the limits of the species distribution area.
Thermal acclimation and experimental design
Adult oysters (length 85 ± 5 mm) were collected from La Pointe du Château (Brittany, France) in November 2004 (seawater temperature 13 °C) and kept constantly immersed at ambient temperature (∼ 13 °C) in aerated 0.22-µm filtered seawater tanks for 21 days. Groups of oysters were then exposed to two laboratory-controlled temperature regimes in 40 L tanks: 60 oysters were acclimated for 4 weeks to 25 °C (4 °C above the temperature of seawater encountered in summer in southern Brittany), and a control group of 70 oysters was maintained in seawater at 13 °C. Oysters were fed three times a week with a microalgal suspension (containing Isochrysis galbana and Pavlova lutheri). No oysters died during the experiment.
For each of the experimental conditions, oysters were sampled at 0, 3, 7, 10, 14, 17 and 24 days following the start of the treatments. Gill and mantle tissues were dissected, rapidly frozen in liquid nitrogen and stored at −80 °C until analysis. Pools of gill and mantle were prepared on these sampling dates by taking 50 mg of each tissue from each of 10 individuals.
Total RNA was extracted using TRIzol® Reagent (Invitrogen, Carlsbad, CA) with 1 mL/50 mg of tissue. For SSH experiments, polyadenylated RNA was isolated using the PolyATtract®mRNA Isolation System (Promega, Madison, WI) according to the manufacturer's instructions. RNAs were resuspended in RNase-free water and their quantity was assessed by spectrophotometry.
Suppression subtractive hybridization
Messenger RNA was extracted from mantle and gills of oysters exposed to 13 or 25 °C after 7–10 and 24 days. Two micrograms of mRNA (1 µg from the gills and 1 µg from the mantle) were used as the template for SSH following the PCR-select cDNA subtraction kit procedure (Clontech, Palo Alto, CA). Hybridization and subtraction steps were carried out in both directions. For forward subtraction, the 25 °C sample (tester) was subtracted with 13 °C sample (driver) and the opposite was done for reverse subtraction. Four libraries (two forward and two reverse) were thus constructed. PCR products were then purified and cloned into pGEM-T vector (Promega). Five hundred white colonies per library were grown on Luria–Bertani medium (with 100 mg·L−1 ampicillin). A total of 2000 randomly selected clones were single-pass sequenced using an ABI 3730 sequencer with the sequencing kit ABI Big dye terminator version 3.1 at the Genoscope Sequencing Center (Evry, France). Sequences were then analyzed using BlastX algorithm available from the National Center for Biotechnology Information (NCBI) and the EST sequences were then submitted to its dbEST and GenBank databases (see supplementary Tables S1–S4).
Real-time PCR analyses
Real-time PCR was used to analyse the expression profiles of some selected genes involved in cell proliferation and differentiation, cellular stress, antioxidant defence, metabolism of nitrogen and ammonia detoxification, membrane fluidity, energetic metabolism and translational processing. Total RNA was extracted from gills and mantle of 10 oysters exposed to 13 and 25 °C for 0, 3, 7, 10, 14, 17 and 24 days. A pool of the 10 RNA samples was made for each tissue at each sampling point in a proportional manner according to the amount of total RNA collected from each animal. Reverse transcription was performed on 20 µg RNA from each pool using the oligo(dT) anchor primer (5′-GACCACGCGTATCGATGTCGACT(16)V-3′) and Moloney murine leukaemia virus (M-MLV) reverse transcriptase (Promega). Real-time PCR was performed in triplicate with 5 µL cDNA (1/20 dilution) in a total volume of 20 µL, using a 7300 Real-Time PCR System (Applied Biosystems, Foster City, CA). The concentrations of the reaction components were as follows: 1× Absolute QPCR SYBR Green ROX Mix (ABgene, Epsom, UK) and 70 nm of each primer. Oligonucleotide primer sequences used to amplify specific gene products are shown in Table 2. Reactions were realized with activation of Thermo-Start® DNA polymerase at 95 °C for 15 min followed by amplification of the target cDNA (45 cycles of denaturation at 95 °C for 30 s, annealing and extension at 60 °C for 1 min) and a melting curve programme from 95 to 70 °C decreasing by 0.5 °C every 10 s. Readings were taken at 60 °C. Each run included a positive cDNA control (one 13 °C sample from the present experiment analyzed in each amplification plate), a negative control (nonreverse-transcribed total RNA) and blank controls (water) for each primer pair. PCR products were then purified, cloned and sequenced for confirmation.
Table 2. Combinations of primers used in real-time PCR expression analysis.
For gene expression calculation, the threshold value (Ct) was determined for each target as the number of cycles at which the fluorescence rose appreciably above the background fluorescence. PCR efficiency (E) was calculated for each primer pair by determining the slope of standard curves obtained from serial dilution analysis of cDNA from different experimental samples (treatment and control), using the method described by Yuan et al.. Individual real-time PCR efficiencies (E) for target or reference genes were calculated according to: E = 10(−1/slope). Results are presented here as changes in relative expression normalized to the reference gene (ribosomal 18S), using the method described by Pfaffl  and determined using the equation:
where Etarget is the amplification efficiency of the target or gene of interest, Eref is the amplification efficiency of the reference (ribosomal 18S) and Ct is the crossing threshold.
The variations in gene expression were analyzed with Student's t-tests between oysters exposed to 13 and 25 °C, using statistica software (Statsoft, Maison-Alfort, France). These statistical analyses were performed using the triplicate real-time PCR assay values obtained for each sample; the graphs (Fig. 2) present the mean values with standard deviations.
This research program was financially supported by the national program PROGIG (Prolifération de Crassostrea gigas, LITEAU II) and the PolyGIGAS program of the Bureau des Ressources Génétiques (n°05/5210460/YF). The authors are grateful to Helen McCombie and Carolyn Friedman for English corrections.