Introduction
How organisms are able to adapt to changing environmental circumstances is still a central problem in biology (Chevin et al. 2010; Delgado and Ruzzante 2020), as many fine-tuned mechanisms in one environment can be selected against in slightly different ecological settings. If possible, many organisms are able to use behavioural movements to avoid local stressful conditions (Sih et al. 2011; Wong and Candolin 2015); fish can for instance avoid warm surface water by moving to deeper and colder areas, or evade fresher top layers in a fjord during snow melting in the spring. If movement is restricted, adaptive physiological responses may revert the organisms back to steady states within the new environment, in response to the new environmental cues, a phenomenon referred to as phenotypic plasticity (Pigliucci 2001; Sangiao-Alvarellos et al. 2005). Although sounding like an optimal evolutionary force on the short scale, there is an ongoing debate wether phenotypic plasticity is hindering or facilitating genetic adaptation in the long term (Ghalambor et al. 2015); because plasticity may change the phenotypes available for selection after exposure to a new environment, and consequetly influence further genetic adaptations. Further, plastic responses can be either adaptive or non-adaptive (Svensson et al. 2020) with respect to the local phenotypic optimum, and is generally assumed to have an influence on evolutionary trajectories through the altered distributions of phenotypes (e.g expressional profiles) upon which selection can act. Having a wide possibility for plastic responses is often viewed as beneficial for adaption to new environments, as this gives a higher chance of expressing a “new phenotypic optima” directly, which then can be genetically assimilated in the new environment (Levis and Pfennig 2016; Waddington 1961), i.e., the induced expression pattern in the novel environment will become “fixated”. Fixating environmentally-induced plasticity through genetic assimilation should hence reduce genetic- and plastic diversity in the derived population, where the rate of stabilizing selection depend on the number of loci that contribute to the additive genetic variance of the character(s) (Lande 1976). Expression patterns of transcriptomes have been recognized as being plastic (Evans 2015), implying that genetically similar individuals can have different transcriptome profiles (phenotypes) as a response to environmental cues (Mäkinen et al. 2016; Papakostas et al. 2014). Indeed, changes in gene expression can evolve very rapidly in many species, including fish (Roberge et al. 2008) and could therefore play an important role in the early steps of population divergence (Wolf et al. 2010).
For aquatic organisms such as fish, the difference between saltwater and freshwater represents considerably different selective forces. Most fish cells maintain a constant ion concentration, and few species are able to cross the salinity gradient between fresh and salt water (Delgado and Ruzzante 2020). Most fish are therefore stenohaline, where the osmoregulatory machinery only operates within relatively narrow salinity boundaries (Hoar and Randall 1984). Only about 3-5% of all fish species are euryhaline, meaning that they posess physiological mechanisms that allow them to adjust to a wide range of salinities (McCormick et al. 2012). Shortly, in saltwater, a fish will have a lower concentration of inorganic ions and hence a lower osmotic pressure compared to the environment, and the fish will passively gain ions and loose water (Evans et al. 2005; Rankin and Jensen 1993). The situation for a freshwater fish is reversed, as the fish now has a higher concentration of ions when compared to the surroundings, and the fish passively gain water and loose inorganic ions. Consequently, to maintain homeostasis, saltwater fish drink saltwater, where excessive salts are actively secreted at the gills and water is absorbed in the intestine (Evans et al. 2005; Hoar and Randall 1984), and freshwater fish actively absorb ions at their gills, minimizes ion loss at their body surfaces, and actively reabsorb ions in their kidney to minimize urinary ion loss (Evans et al. 2005). Altogether, the cost of osmoregulation is highly variable, depending on salinity, oxygen and temperature (Ern et al. 2014), where the total cost range between a few percent up to 30-50% of the total energy budget (Boeuf and Payan 2001; Ern et al. 2014). In total, about 7% of the total energy budget can be spent in the gill tissue alone (Mommsen 1984).
The threespine stickleback Gasterosteus aculeatus (hereafter stickleback) is a small fish known to have a wide salinity tolerance (Bell and Foster 1994) at a seemingly low osmoregulatory cost (Grøtan et al. 2012). Originally of marine origin (Bell and Richkind 1981), the stickleback has invaded and established populations in numerous freshwater habitats since the last glaciation in the northern hemisphere (Bell and Foster 1994). Thus, stickleback populations are found at a wide osmotic range, spanning marine oceanic ecosystems, costal brackish water systems, freshwater rivers and lakes. Many populations have become landlocked after freshwater colonization, typically due to isostatic uplifting of the land following deglaciation, often restricting the gene flow between the founders and the derived populations. With reduced gene flow, and with freshwater habitats having a stable salinity compared to costal waters, one would expect strong directional selection on traits that facilitate local adaptation to low salinity. Furthemore, one would also expect traits promoting a broad salinity tolerance (being euryhaline) to be selected against, due to the cost of sustaining characters that has not been required for many generations, and the low genetic variation typically found in derived populations (Schultz and McCormick 2012). Genetic comparisons of marine and freshwater stickleback populations show signs of strong selection, and several outlier loci are identified by comparing whole genome sequences (Jones et al. 2012b), SNPs (Guo et al. 2015; Hohenlohe et al. 2010; Jones et al. 2012a) and microsatellite genotypes (DeFaveri et al. 2011; Taugbøl et al. 2014b). Genetic studies further indicate that the frequency of freshwater-linked alleles can increase rapidly in newly colonized freshwater habitats (Lescak et al. 2015). However, with respect to gene expression and potential gene-regulatory adaptations in response to salinity, previous experiments have either tested candidate genes (McCairns and Bernatchez 2010; Taugbøl et al. 2014a); compared populations directly without exposure to non-native environments (Jones et al. 2012b; Rastorguev et al. 2018); or tested for transcriptomic expression differences after a longer period of acclimatization in the non-native environments (30 days to 3 months; Gibbons et al. 2017; Wang et al. 2014). However, less is known of the immediate expressional patterns following acute exposure to contrasting salinities. The objective of this study was to assess transciptomic expression and compare regulatory changes in genes between marine and freshwater sticklebacks subjected to abrupt salinity transfers.