Ecological proteomics: Finding molecular markers that matter



Anne Dalziel, Fax: 604 822 2416; E-mail:


It is becoming increasingly clear that local adaptation can occur even in the face of high gene flow and limited overall genomic differentiation among populations (reviewed by Nosil et al. 2009). Thus, one important task for molecular ecologists is to sift through genomic data to identify the genes that matter for local adaptation (Hoffmann & Willi 2008; Stapley et al. 2010). Recent advances in high-throughput molecular technologies have facilitated this search, and a variety of approaches can be applied, including those grounded in population genetics [e.g. outlier analysis (Pavlidis et al. 2008)], classical and quantitative genetics [e.g. quantitative trait locus analysis (MacKay et al. 2009)], and cellular and molecular biology [e.g. transcriptomics (Larsen et al. 2011)]. However, applying these approaches in nonmodel organisms that lack extensive genetic and genomic resources has been a formidable challenge. In this issue, Papakostas et al. (2012). demonstrate how one such approach – high-throughput label-free proteomics (reviewed by Gstaiger & Aebersold 2009; Domon & Aebersold 2010) – can be applied to detect genes that may be involved in local adaptation in a species with limited genomic resources. Using this approach, they identified genes that may be implicated in local adaptation to salinity in European whitefish (Coregonus lavaretus L.) and provide insight into the mechanisms by which fish cope with changes in this critically important environmental parameter.

The European whitefish (Coregonus lavaretus L.) is a widely distributed salmonid found primarily in freshwater habitats in Northern and Central Europe. The species exhibits substantial morphological variation among populations and is characterized by multiple parallel radiations in postglacial lakes (Ostbye et al. 2006; Vonlanthen et al. 2012). In addition to these freshwater ecotypes, there are some populations that spawn in brackish and saltwater habitats, particularly in and around the Baltic Sea (Fig. 1). The evolutionary history of European whitefish is complex, but is thought to involve postglacial colonization of freshwater habitats by a marine or anadromous ancestor, followed by adaptive radiations in freshwater and subsequent re-colonization of brackishwater or saltwater habitats (Etheridge et al. 2012). However, the loci involved in allowing these freshwater fish to colonize (or recolonize) saltwater habitats remain largely unknown.

Figure 1.

 The brackishwater ecotype of the European whitefish (Coregonus lavaretus L.). Unlike the many freshwater populations of this species, this ecotype breeds in the saline waters of the Baltic Sea (Photograph credit: Mikael Himberg).

Adaptation to a novel salinity environment poses a substantial challenge for teleost fishes, which must maintain the osmolarity and ion composition of their internal fluids (including blood, extracellular fluids and cytoplasm) within a very narrow range to survive. When a fish is in saltwater, its body has a much lower osmolarity than does the environment, so the fish tends to passively gain ions from (and lose water to) the environment. Alternatively, when a fish is in freshwater, its body has a much higher osmolarity than does the environment, so the fish tends to passively lose ions to (and gain water from) the environment. These ionic and osmotic imbalances must be compensated for by actively pumping out ions (when in saltwater) or actively taking up ions (when in freshwater). Maintaining osmotic homeostasis is energetically costly and requires substantially different intracellular machinery in each of these environments (reviewed by Evans 2008; Hwang et al. 2011). As a result, most species of fishes are specialized to live within a narrow salinity range in either fresh water or salt water, and a complete remodelling of osmoregulatory tissues such as the gill is required to make a transition from one habitat to the other.

The first step in identifying genes involved in local adaptation to an environmental parameter such as salinity is to demonstrate that phenotypic differentiation has in fact occurred in the direction predicted (reviewed by Kawecki & Ebert 2004). To do this in European whitefish, Papakostas et al. (2012) conducted a ‘reciprocal-transplant’ style experiment and bred and reared fish from fresh (∼ 0 ppt) and brackishwater (∼ 4–8 ppt) whitefish populations in the laboratory at a range of ecologically relevant salinities. As predicted for local adaptation, increased salinity (>2 ppt) had a highly significant negative effect on both fertilization success and survival during early development in the freshwater whitefish, whereas brackishwater whitefish had nearly equal success at all salinities. Together, these data argue for local adaptation to higher salinity in the brackishwater population. However, Papakostas et al. (2012) also detected a significant influence of maternal effects on these traits within populations, suggesting that further studies must be undertaken to determine the importance of parental effects on the observed phenotypic differences.

Papakostas et al. (2012) then used a high-throughput label-free proteomic approach to measure protein levels in newly hatched fry from each salinity treatment and test for differences in the proteomic response to salinity among populations. Such differentially regulated proteins represent candidate genes that may have been subject to selection in these populations. The advantage of using a proteomic screen, compared with a transcriptomic screen, is that protein levels may be more reflective of whole-animal phenotype and performance than are mRNA levels (Diz et al. 2012). Unfortunately, it has been difficult to use proteomic technologies on nongenomic model organisms until relatively recently (e.g. Looso et al. 2010). The ‘label-free’ proteomics approach used by Papakostas et al. (2012) has the advantage of providing greater coverage of the proteome and a higher dynamic range in quantification compared with traditional proteomics approaches (Gstaiger & Aebersold 2009). Papakostas et al. (2012) found that many proteins vary with changes in salinity, but that the identity of the salinity-responsive proteins differed between populations such that only 6 of 115 proteins with a significant change in expression level were similar between the two ecotypes. In particular, the brackishwater population displayed changes in proteins that were previously known to be critical for acclimation to salinity change in salmonids (e.g. the Na+, K+ ATPase alpha subunit; Bystriansky et al. 2006), for stress tolerance in animals (e.g. Hsc70), for osmoregulation in elasmobranchs (tumour necrosis factor) and for salinity sensing in fishes (polyvalent cation receptors). In addition, Papakostas et al. (2012) were able to use protein–protein interaction models and network analysis (e.g. Malik et al. 2010) to identify novel candidate genes, biochemical pathways and signalling cascades involved in the differential responses to salinity between the ecotypes.

The observation that this ‘shotgun’ or unbiased proteomics screen identified many genes that would have been a priori candidates for local adaptation supports the idea that physiologically informed studies of the molecular basis for local adaptation can be useful (Dalziel et al. 2009). However, not all of the divergent proteins identified by Papakostas et al. (2012) would have been predicted to be candidates for salinity adaptation from prior physiological knowledge (e.g. heterogenous nuclear ribonucleoprotein M and stress-induced phosphoprotein 1). These results clearly demonstrate that ‘unbiased’ proteomic methods have an important place in the quest to find the loci that matter for local adaptation.

A.C.D. has just completed her PhD thesis in P.M.S.'s laboratory and studies the physiological and molecular mechanisms by which animals adapt to their environment. P.M.S. takes an integrative approach to understand the evolution and mechanistic basis of inter- and intra-specific variation in physiological traits.