Biochemical Genetics

, Volume 50, Issue 5, pp 454–466

Differences in Salinity Tolerance and Gene Expression Between Two Populations of Atlantic Cod (Gadus morhua) in Response to Salinity Stress


    • Department of Biological SciencesAarhus University
    • National Institute of Aquatic Resources, Technical University of Denmark
  • E. E. Nielsen
    • National Institute of Aquatic Resources, Technical University of Denmark
  • K. Meier
    • National Institute of Aquatic Resources, Technical University of Denmark
  • P. A. Olsvik
    • National Institute of Nutrition and Seafood Research
  • M. M. Hansen
    • Department of Biological SciencesAarhus University
  • V. Loeschcke
    • Department of Biological SciencesAarhus University

DOI: 10.1007/s10528-011-9490-0

Cite this article as:
Larsen, P.F., Nielsen, E.E., Meier, K. et al. Biochem Genet (2012) 50: 454. doi:10.1007/s10528-011-9490-0


Populations of marine fish, even from contrasting habitats, generally show low genetic differentiation at neutral genetic markers. Nevertheless, there is increasing evidence for differences in gene expression among populations that may be ascribed to adaptive divergence. Studying variation in salinity tolerance and gene expression among Atlantic cod (Gadus morhua) from two populations distributed across a steep salinity gradient, we observed high mortality (45% North Sea cod and 80% Baltic Sea cod) in a reciprocal common garden setup. Quantitative RT-PCR assays for expression of hsp70 and Na/K-ATPase α genes demonstrated significant differences in gene regulation within and between populations and treatment groups despite low sample sizes. Most interesting are the significant differences observed in expression of the Na/K-ATPase α gene in gill tissue between North Sea and Baltic cod. The findings strongly suggest that Atlantic cod are adapted to local saline conditions, despite relatively low levels of neutral genetic divergence between populations.


Gene expression variationPopulation structureSalinity gradientSalinity tolerance


Many marine organisms are widely distributed in the oceans, inhabiting an array of different physical and biological environmental conditions (Waples 1998). Temperature and salinity are commonly known to vary among geographic locations, and these factors are therefore expected to have a significant impact on distribution of species and populations (McKay et al.2001; Doebeli and Dieckmann 2003). Marine organisms experiencing different levels of temperature and/or osmotic stress within their distributional range could therefore be expected to show local evolutionary responses in terms of adaptive differences among populations (Garcia de Leaniz et al. 2007). The underlying genetic background of trait differences could originate from variation within the coding sequence of genes important for the adaptive trait (Sick 1965; Vasemägi and Primmer 2005) or alternatively from genetic variation in gene expression (Schulte 2001; Oleksiak et al. 2002; Babu and Aravind 2006). In general, short-term fluctuations in the physical environment may result in plastic acclimation responses that include increased activity of cellular defense mechanisms such as up-regulation of heat shock proteins (Deane et al. 2002; Fangue et al. 2006; Larsen et al. 2008a, b). On the other hand, long-term responses to potential sublethal conditions may include heritable adaptations leading to increased population fitness evolving in stressful environments (Lee and Petersen 2002; DeLong and Karl 2005) and ultimately speciation (Doebeli and Dieckmann 2003; West-Eberhard 2005).

In fish, one of the best-described examples of adaptation to environmental conditions is the killifish (Fundulus heteroclitus) distributed along a steep thermal gradient on the east coast of North America (Schulte 2007; Whitehead 2009, 2010). Killifish populations inhabiting areas with differences in mean temperature of more than 15°C showed 2-fold variation in gene expression of the Ldh-B gene, but intraspecific differences have also been observed in hsp70 and Na/K-ATPase expression when fish were subjected to the same environmental conditions in a common garden setup (Schulte 2001, 2007; Scott et al.2004; Scott and Schulte 2005; Fangue et al.2006). Furthermore, microarray experiments conducted on the same populations demonstrated that the expression of numerous genes was affected by natural selection (Oleksiak et al.2002; Whitehead and Crawford 2006). Larsen et al. (2007) showed that the extreme euryhalinity observed in European flounder (Platichthys flesus) is accompanied by salinity-dependent expression of Na/K-ATPase genes and up-regulation of heat shock protein genes following reciprocal transplantation into extreme saline conditions (Larsen et al. 2007, 2008b). Similarly, Fangue et al. (2006) demonstrated a strong correlation between temperature tolerance and hsp70 expression in killifish from extreme habitats. Thus, studies of variation in gene expression among natural populations subjected to environmental challenges under controlled conditions may reveal important functional physiological differences involved in adaptation to different local environmental conditions (Cossins and Crawford 2005; Gilad et al.2006; Whitehead and Crawford 2006; Larsen et al.2011).

Steep natural salinity gradients are ideally suited natural laboratories for investigating adaptive differences related to osmotic stress. A well-known gradient, going from near oceanic salinities to brackish and almost freshwater conditions, is found in the transition zone between the North Sea and the Baltic Sea (reviewed by Johannesson and Andre 2006). Atlantic cod (Gadus morhua) are found in relatively large numbers throughout this region. Besides highly variable environmental conditions within the species range, tagging as well as molecular studies conducted within the North Sea and Baltic Sea region have shown that cod tend to have low migration rates (Svedang and Svenson 2006), resulting in significant genetic differences between North Sea and Baltic Sea populations (Nielsen et al.2001a, 2003, 2005; Poulsen et al.2006). Hence, the occurrence of local populations with semi-independent evolutionary trajectories provides ample opportunities for local adaptations.

Some of the most promising candidate genes for studying physiological stress are heat shock protein (hsp) genes since they are known to be induced by osmotic stress (Deane et al.2002; Deane and Woo 2004). Particularly hsp70 is commonly used as an indicator of physiological stress in fish (Smith et al.1999; Basu et al.2001; Lejeusne et al.2006; Larsen et al.2008b). The Na/K-ATPase genes constitute another group of genes previously identified as important for ion balance maintenance in fish (McCormick 2001; Evans et al.2005; Larsen et al.2008a).

In this study, mortality rates of cod sampled in the North Sea and the Baltic Sea were quantified following long-term acclimation in a reciprocal transplantation experiment mimicking natural salinities in the North Sea and the Baltic Sea. Intra- and inter-population variation in expression of hsp70 and Na/K-ATPase α were quantified following short-term and long-term acclimation, testing the hypothesis that intraspecific variation in gene expression and salinity tolerance is related to the local environmental conditions experienced by cod. Results are discussed in relation to population structure, local adaptation, management, and potential re-establishment of cod populations in the future.

Materials and Methods

Tank Experiment

Live cod were sampled using long lines and fish traps for two consecutive years (2006 and 2007) in the North Sea at Skagen, Denmark, and in the Baltic Sea around the island of Bornholm, Denmark (Fig. 1). At the termination of the experiment, cod varied in length from 34.5 to 42.7 cm, and no significant difference in length was observed between locations/acclimation groups (one-way ANOVA, P > 0.05). Both North Sea and Baltic Sea cod were transported to aquarium facilities at Aarhus University, Denmark, and held at 10 ± 0.5°C in 1,000-liter tanks containing recirculated water mimicking natural salinities of the North Sea (33 ± 0.5 ppt) and Baltic Sea (9 ± 0.5 ppt). All cod were maintained on a cycle of 16 h of light and 8 h of darkness and allowed to acclimate to laboratory conditions for at least 1 month under conditions corresponding to their original environment. Following acclimation, 7–10 cod from each population were randomly assigned to either a control group or a reciprocal transplantation group. In the transplantation experiments, tank water salinity was gradually reduced from 33 to 9 ppt, or elevated from 9 to 33 ppt, over 2 h, in order to minimize acute stress responses. Cod were fed daily with rainbow trout (Oncorhynchus mykiss) and mussels (Mytilus edulis) cut into pieces, up to 5 days before they were killed (in the short-term experiment no food was supplied). Cod were sampled following 1 and 50 days at reciprocal salinities to study both short- and long-term gene regulation related to environmental salinity. Only cod collected in 2006 were used for gene expression quantification, since only one Baltic Sea cod survived the reciprocal transplantation experiment in 2007. All experiments were carried out according to Danish legislation concerning the use of experimental animals.
Fig. 1

Sampling localities in Denmark of North Sea cod (Skagen) and Baltic Sea cod (Bornholm)

Tissue Collection and RNA Extraction

Gill and kidney are known as the main organs for osmoregulation in fish (Sangiao-Alvarellos et al.2003), and accordingly these tissues were targeted. Fish were killed with a hard blow to the head, and tissue samples (gill filaments and kidney) were dissected and immediately placed in RNAlater, following the manufacturer’s instructions (Qiagen, Hilden, Germany). RNAlater preserved samples were stored at −20°C until RNA extraction. Total RNA was extracted using the RNeasy minikit (Qiagen, Hilden, Germany). Extractions were performed as recommended by the manufacturer, except for an additional DNase treatment step for 15 min at 25°C to remove any remaining genomic DNA in the RNeasy extracted samples. Total RNA was stored at −20°C. Concentration of extracted RNA was determined at 260 nm in a standard Hellma cuvette (path length 10 mm) using the GeneQuant II (RNA/DNA Calculator, Pharmacia Biotech). Total RNA quality was routinely analyzed using 2% agarose gel electrophoresis.

Reverse Transcription of RNA and qRT-PCR

Reverse transcription of total RNA into cDNA was performed using the SuperScriptII RNase H-Reverse Transcriptase kit (Invitrogen, Carlsbad, USA) in a reaction volume of 20 μl containing 1× reaction buffer, 5 mM MgCl2, 1 mM dNTP mixture, 0.3 μl anchored oligo-(dT)20 primer (2.5 μg/μl), 0.9 μl SuperScriptII reverse transcriptase, and 3 μg total RNA. Samples were then diluted 10-fold and stored at −20°C until qRT-PCR analysis.

Primers were designed using the Primer-3 software (Rozen and Skaletsky 2000) based on sequences (Table 1) from EST libraries produced by the CodGen project, Bergen, Norway ( PCR products from all primer pairs were checked on a 2% agarose gel to verify primer specificity and PCR product length and to insure that they produced only a single amplicon when cod cDNA served as template. Furthermore, dilution series were performed to verify amplification efficiency of primer pairs (Stahlberg et al.2004).
Table 1

Primer sequences used for qRT-PCR in this study


Nucleotide sequence (5′–3′)

Accession no.

Na/K-ATPase α












A Lightcycler 1.2 was used with SYBR Green chemistry (Lightcycler FastStart DNA Master SYBR Green I kit) and the Lightcycler relative quantification software 3.5 (Roche Diagnostics, Mannheim, Germany) to perform qRT-PCR. This software includes the second derivative maximum calculation, a fully automated method for CT determination. The CT value is defined as the cycle in which the second derivative is at its maximum; ideally, this should always be in the heart of the log-linear portion of the reaction. Moreover, this method is more objective and reproducible than human-mediated crossing point settings (Lightcycler relative quantification software manual). All qRT-PCR reactions were performed as follows: 10 min at 95°C, followed by 40 cycles of 95°C for 15 s, 60°C for 10 s, and 72°C for 10 s. Melting curve analysis was performed following each reaction to confirm the presence of only a single product in the reaction. Negative control reactions were performed for all samples using RNA that had not been reverse-transcribed to control for the possible presence of genomic DNA contamination. Genomic DNA contamination was present in all samples but never constituted more than 1:4,500 of starting cDNA copy numbers (data not shown).

Statistical Analysis

Relative gene expression values were calculated using the comparative method according to Livak and Schmittgen (2001). Data are presented as mean expression levels ± SE. Gene expression levels were compared using one-way ANOVA, and multiple pairwise comparisons were conducted using the Tukey test between salinities (within populations) or between populations (maintained at the same salinity). A comparison of mortality between groups was conducted using a Kruskal–Wallis test based on pairwise Mann–Whitney comparisons, applying the standard Bonferroni correction method. Analyses were conducted using the Past statistics software (Hammer et al. 2001), and throughout the significance level was set at α = 0.05.


Low mortality (0–15%) was observed when cod were maintained at their natural salinities. In contrast, high mortality was observed in the transplantation groups; 45% for North Sea cod and 80% for Baltic Sea cod (Fig. 2). A Kruskal–Wallis test based on pairwise Mann–Whitney comparisons demonstrated a significantly higher mortality in transplanted Baltic Sea cod (P < 0.01) but not in North Sea cod (P = 0.85) compared with the control group maintained at natural salinities.
Fig. 2

Survival of cod from the North Sea (NS) and the Baltic Sea (BS) following long-term acclimation to brackish seawater (9 ppt) and full-strength seawater (33 ppt) at 10°C (n = 2–10). * Significant difference between salinity treatments within populations (P < 0.05)

Long-Term Acclimation

In the long-term acclimation experiment, significant differences were observed between North Sea and Baltic Sea cod in expression of the Na/K-ATPaseα gene in gill tissue at 33 ppt (P = 0.014). Moreover, significant changes were observed in the expression of the Na/K-ATPase α gene in North Sea cod following acclimation to the different saline conditions in both gill (P = 0.042) and kidney tissues (P < 0.01) (Fig. 3). For Baltic Sea cod, there were no significant differences in expression of the Na/K-ATPaseα gene in any tissue (kidney, P = 0.11; gill, P = 0.68). In the long-term acclimation experiment, no significant differences in expression of hsp70 were found in any of the cod populations for any tissue following acclimation to reciprocal salinities compared to control cod kept at their native salinities (Fig. 4).
Fig. 3

Expression of Na/K-ATPase α in kidney (left) and gill (right) tissue of cod from the North Sea (NS) and the Baltic Sea (BS) following long-term acclimation to brackish seawater (9 ppt) and full-strength seawater (33 ppt) at 10°C (n = 2–8). Amounts of Na/K-ATPase α mRNA are normalized to the corresponding EF1α abundance from the same sample, and mean values are expressed in arbitrary units ± SE. * Indicates significant difference between salinity treatments within populations (P < 0.05) and # indicates significant difference between populations at the same salinity (P < 0.05)
Fig. 4

Expression of hsp70 in kidney (left) and gill (right) tissue of cod from the North Sea (NS) and the Baltic Sea (BS) following long-term acclimation to brackish seawater (9 ppt) and full-strength seawater (33 ppt) at 10°C (n = 2–8). Amounts of hsp70 mRNA are normalized to the corresponding EF1α abundance from the same sample, and mean values are expressed in arbitrary units ± SE. * Significant difference between salinity treatments within populations (P < 0.05). # Significant difference between populations at the same salinity (P < 0.05)

Short-Term Acclimation

In North Sea cod, significant up-regulation of hsp70 was observed in both gill (P < 0.01) and kidney (P = 0.031) tissue following 1 day at reciprocal salinities (Fig. 5). In Baltic Sea cod, expression of hsp70 was significantly up-regulated only in gill tissue (P < 0.01) and not in kidney tissue, although a strong trend toward up-regulation was observed (P = 0.069). The Na/K-ATPase α gene expression showed no significant differences in either gill or kidney tissue following 1 day at reciprocal salinities in any population (data not shown).
Fig. 5

Expression of hsp70 in kidney (left) and gill (right) tissue of cod from the North Sea (NS) and the Baltic Sea (BS) following short-term acclimation to brackish seawater (9 ppt) and full-strength seawater (33 ppt) at 10°C (n = 6–8). Amounts of hsp70 mRNA are normalized to the corresponding EF1α abundance from the same sample, and mean values are expressed in arbitrary units ± SE. * Significant difference between salinity treatments within populations (P < 0.05)


In this experiment, mortality was high for transplanted cod in both populations (compared with cod maintained at their natural salinities) but statistically significant only for Baltic Sea cod. Furthermore, intraspecific variation in expression of the Na/K-ATPase α gene was found between populations (higher expression in North Sea cod than in Baltic Sea cod at 33 ppt), supporting the presumed presence of adaptations to local saline conditions as suggested on the basis of neutral markers. Finally, short-term induction of hsp70 was detected in both cod populations in response to transplantations (i.e., change in salinity); however, following long-term acclimation this apparent acute stress response (hsp70 up-regulation) disappeared.

Long-Term Variation in Expression of the Na/K-ATPase α Gene

Large and significant differences were observed between the two cod populations in expression of the Na/K-ATPase α gene in gill tissue following long-term acclimation to seawater salinities (33 ppt), demonstrating population-specific patterns of gene regulation depending on the origin of the population. Also, Na/K-ATPase α gene expression in gill and kidney tissue showed a clear plastic response to salinity acclimation, in concordance with previous studies (Deane and Woo 2004). Furthermore, in North Sea cod we detected significant differences in gene expression between salinity treatment groups in both kidney and gill tissues. The smaller difference between acclimation groups of Baltic Sea cod in expression of the Na/K-ATPase α gene is similar to that observed in a study of European flounder across the same salinity gradient (Larsen et al. 2008b). In that study, Baltic Sea flounder showed much smaller variation in expression of the Na/K-ATPase α gene between salinity treatments compared with North Sea flounders, potentially signaling adaptation to a more unstable environment with changing salinities. Thus, similar patterns in gene expression in response to salinity could indicate that Baltic Sea cod are better adapted to changing salinities, although their high mortality at high salinities indicates that they are not able to tolerate full-strength seawater to the same extent as North Sea cod. Thus, surviving Baltic Sea cod in this experiment are most likely the most tolerant to salinity stress, as they survived the transplantation treatment; therefore, we could potentially underestimate the actual response in Na/K-ATPase α gene expression in the “real” population. Direction of Na/K-ATPase gene expression in both gill and kidney tissue is in line with similar studies on salinity acclimation (Deane and Woo 2004) and other population-specific studies from the same region in European flounder (Larsen et al.2008b). Furthermore, Na/K-ATPase gene expression has been shown to be a good proxy for Na/K-ATPase protein levels and Na/K-ATPase activity in both gill and kidney tissues (Deane and Woo 2004). Therefore, the results of this study most likely reflect functional differences between the two cod populations.

Expression of hsp70 and Intraspecific Variation in Salinity Tolerance

The observed initial stress response (hsp70 mRNA induction) in both North Sea and Baltic Sea cod following short-term exposure to reciprocal salinities shows that the cod are highly cellular stressed following transfer. This initial stress response disappears, however, under long-term acclimation when osmoregulative processes presumably regain homeostasis (Sangiao-Alvarellos et al.2005). No stress responses (in terms of differences in expression of hsp70 in transplanted cod compared with control cod maintained at natural salinities) were observed in the long-term acclimation experiment. Thus, it is important to remember that in the transplantation groups, particularly in the Baltic Sea group, mortality was very high. Therefore, gene expression results are only from individuals that were able to acclimate and survive; since only two individuals in the Baltic Sea transplantation group survived, this estimate is likely underestimated. The acclimation experiment was repeated in two consecutive years, however, and so we are confident with the results of variation in salinity tolerance (mortality in reciprocal habitats). A common garden project based on a family breeding design needs to be implemented to estimate the intrapopulation heritability of salinity adaptation.

The population-specific differences in gene regulation observed in this study could possibly be expected, due to the relatively high levels of genetic differentiation measured between cod populations in the North Sea and the Baltic Sea using neutral microsatellite markers (FST = 0.045; Nielsen et al.2001b, 2003; Poulsen et al.2006). Furthermore, the extremely large differences in salinity tolerance and subsequent low fitness of transplanted cod from both populations may be the result of genetically based adaptive differences between populations. Previously, evidence for local adaptation has been presented in studies on Atlantic cod, suggesting adaptation to different salinities in a number of important reproductive parameters between western Baltic and Baltic Sea cod. These studies showed limited plasticity in several reproductive parameters, such as salinity required for activation of spermatozoa, fertilization rates, and salinity at which eggs became neutrally buoyant, for local cod populations acclimated to different naturally occurring salinities (Nissling and Westin 1997). Moreover, other studies have demonstrated local adaptation in cod using common garden conditions, thereby presenting evidence of divergence among populations in important life history traits such as countergradient variation in body shape between populations (Marcil et al.2006) and differences in spawning time from four regions of Norway (Ottera et al.2006). The observed differences in salinity tolerance and gene expression therefore support the results of previous studies and highlight the importance of recognizing adaptive differences among marine fish populations on several levels of organization, instead of focusing on the generally low levels of genetic structuring observed from neutral genetic markers in marine fish.

Sampling in Nature

Whether experimental acclimation has influenced the survival of transplanted cod is impossible to rule out. Sampled cod, however, were of the same length, and they were acclimated at their natural salinities in the laboratory for a minimum of 1 month to insure that they had started feeding and that background mortality from injured or stressed fish had passed. It is therefore likely that the salinity challenge in the reciprocal transplantation groups was simply too large for the cod to acclimate, and this may explain the large and drastic mortalities in both populations. Particularly in the long-term transplanted Baltic Sea cod, we observed high mortality (one survivor in 2007 and two in 2006). As a result, it is important to consider that we quantified gene expression only in surviving cod from this group (2006) and that our results account only for this group of cod. Naturally, we are unable to predict whether the gene expression regulation in dead cod was different from surviving individuals. Similar studies on European flounder (Larsen et al.2008b) sampled at almost the same localities and maintained at the same salinities as the cod in this experiment showed no mortalities, demonstrating better osmoregulative capabilities for flounder compared with cod.

A number of factors in addition to genetics may also affect the results, such as early developmental acclimation, maternal factors, and epigenetic effects. Other authors have argued, however, that such nongenetic factors should not contribute significantly to gene regulation (Whitehead and Crawford 2006), in particular following proper acclimation (Deane and Woo 2004). Thus, still more research is warranted to elucidate the effect of epigenetics and developmental acclimation on gene expression (Goetz and MacKenzie 2008). Finally, low sample size of transplanted Baltic Sea cod due to high mortality imposes a problem regarding the statistical analysis (low power) and may result in uncertainty of estimates.

Conservation and Fisheries Management Perspectives

In a conservation perspective, our results support the idea that adaptive differences exist between the two sampled cod populations, despite relatively low levels of genetic differentiation using neutral markers. Although the sample sizes were limited in this study (but conducted for two consecutive years), the results point to the large adaptive importance of environmental gradients for creating the population structure in an environment with no obvious barriers to migration. Based on this and previous studies (e.g., Nissling and Westin 1997; Larsen et al.2007, 2008b), it is clear that there are several distinct populations of marine fish in the transition zone from the highly saline North Sea into the brackish Baltic Sea that are adapted to different salinities. Finally, this study illustrates that the Baltic Sea cod is evolutionarily and adaptively distinct, a distinction that is likely to be pronounced for many other Baltic populations of marine fish. This highlights the importance of the North Sea–Baltic Sea transition zone as a hot spot for intraspecific marine biodiversity and underlines the importance of conserving this marine region and the populations that inhabit it. Genetically effective immigration of cod and other species from neighboring populations (e.g., the North Sea) is unlikely due to the lack of adaptation to nonnative salinities, both at the gene expression level and in physiological and several reproductive characters.


This study was supported by grants to Peter Foged Larsen from the SLIP Research School under the Danish Network for Fisheries and Aquaculture Research financed by the Danish Ministry for Food, Agriculture and Fisheries and the Danish Agricultural and Veterinary Research Council, the Danish Institute for Fisheries Research, the Elisabeth and Knud Petersens Foundation, and the Idella Foundation. The authors thank Palle Holm Hansen for guidance on fish maintenance, Karen-Lise Dons Mensberg and Dorte Meldrup for assistance in the laboratory, and Henrik Baktoft for graphical assistance.

Copyright information

© Springer Science+Business Media, LLC 2011