Gene expression analysis in Mytilus chilensis populations reveals local patterns associated with ocean environmental conditions

Abstract

Marine ecosystems involve relationships between genomic interactions of marine populations with shared biogeographic ranges and the environmental conditions. These relationships, studied mainly through neutral DNA markers, are not always consistent with actual biogeographic patterns or the evolutionary history of marine species. In addition, increased information at functional genomic level from non-model species allows the study of adaptive responses in marine populations. This work reports local transcriptomic patterns in populations of the mussel Mytilus chilensis and their correspondence with oceanographic variability in southern Chile. Analysis of gene expression patterns was conducted through qPCR of seven candidate genes involved in the response to environmental stress (HSP70, HSP90), iron metabolism (Ferritin), pathogens (Mytilin B, Defensin) and oxidative stress (SOD-CuZn, Catalase) in at five study sites located in southern Chile, from Valdivia (39°56′S–73°36′W) to Melinka (43°52′S–73°44′W). Multivariate and correlation analyses were used to assess the relationship between levels of individual gene expression and site features characterized using satellite data on surface temperature, chlorophyll concentration and total suspended sediments. Two main groups of sites with differential patterns of gene expression were identified. Individuals exposed to higher temperatures showed an overexpression of HSP70, HSP90 and Ferritin genes. The expression of SOD-CuZn and Catalase was correlated with local chlorophyll-a (i.e. food availability for mussels), although with opposite correlations. In addition, Mytilin B showed higher levels of expression in areas with higher freshwater influence. Patterns of gene expression across the region of interest suggest that spatial variability in environmental conditions induce phenotypic changes in different populations of the same mussel species. In addition, the analysis of expression patterns in candidate genes can reveal local patterns in populations where other molecular markers show no genetic structure.

Introduction

Understanding the effect of environmental variability on the biogeographic patterns of marine species has been one of the main purposes of marine physiology and molecular ecology (Osovitz and Hofmann, 2007). Within this framework, topics such as climate change in oceanic and coastal areas have become very important in studies of marine ecosystem dynamics (Harley et al., 2006, Parmesan, 2006), latitudinal gradients affecting speciation and biodiversity rates (Allen and Gillooly, 2006), larval development under different conditions of temperature and food availability (Hoeghguldberg and Pearse, 1995, O'Connor et al., 2007), and relationships between the dispersal ability of species and their biogeographic range (Lester et al., 2007).

Undoubtedly, genetic studies based on molecular DNA markers have made it possible to identify the relationship between biogeographic variables and population genetic structures in marine environments (Ayers and Waters, 2005, Kelly and Eernisse, 2007, Osovitz and Hofmann, 2007). However, few studies have associated phenotypic differences with genotypic changes determined by latitudinal patterns in environmental conditions (e.g. Lee and Petersen, 2002). Therefore, there is currently no certainty about the correlations between phenotypic variations and genomic changes inferred from neutral genetic markers (or markers assumed to be unaffected by natural selection). In some cases, these correlations may occur as a result of genetic processes such as heritability, genetic flow, hybridization among others; however, they are also affected by non-genetic processes, such as disruption of inter-specific interactions and habitat degradation (Ouborg et al., 2010). Furthermore, different types of molecular DNA markers may show contrasting results, or may not detect the population polymorphisms due to low genome coverage. In this context, microsatellite variability is widely used to infer levels of genetic diversity in natural populations (Lopes-Cunha et al., 2012, Loukovitis et al., 2012, Ni et al., 2011, Varela et al., 2009). However, the ascertainment bias caused by typically selecting only the most polymorphic markers in the genome may lead to reduce the sensitivity to judge genome-wide levels of genetic diversity (Väli et al., 2008). Thus, conservation genomics can incorporate these approaches to study the genetic basis of local adaptation or inbreeding depression. By contrast, predicting a population's variability or capacity to adapt to climate change based on genomic information will require not only the identification of relevant loci, but also a quantitative estimate of their connection to fitness and demographic vital rates. One of the most promising sources of genomic tools can be found at transcriptome level by characterization of Expressed Sequence Tags (EST) (see details Allendorf et al., 2010, Ouborg et al., 2010).

ESTs are cDNA sequences from regions of the transcriptome which are mainly obtained through cDNA libraries generated by cloning and sequencing (Bouck and Vision, 2007). Additionally at present, large amounts of ESTs can be obtained using next-generation sequencing technologies (Harismendy et al., 2009). These sequences are available in several public databases, such as GenBank (http://www.ncbi.nlm.nih.gov/genbank/), EMBL-EBI (http://www.ebi.ac.uk/embl/) and DDBJ (http://www.ddbj.nig.ac.jp/), where to date there are more than 67 million sequences available. EST sequences can provide a fast method to identify candidate genes in different species, allowing the development of studies for local adaptation to loci with certain phenotypic differentiation (Bouck and Vision, 2007). Candidate gene expression studies have a great potential to understand complex ecological phenomena, such as phenotypic plasticity, changes in habitat and local adaptations (Gibson, 2002). In this context, species with a wide biogeographical range have been studied to understand phenotypic plasticity through gene expression analysis (Dutton and Hofmann, 2009). Some widely studied candidate genes within this context are HSP70 and HSP90 genes, which encode for molecular chaperones involved in thermal shock response among other stressors (Lindquist, 1986). Additionally, other candidate genes are involved in the innate immune response, such as Mytilin B and Defensin, which are evolutionarily conserved genes that encode for antimicrobial peptides (AMP) (Mitta et al., 2000). Another candidate gene could be Ferritin, which is the main gene involved in iron storage at cellular level, storing up to 4500 iron atoms (Orino and Watanabe, 2008). However, the implication of this gene in different functions at cellular and molecular level is still under discussion. Finally, SOD-CuZn and Catalase are genes involved in the response to the oxidative stress generated by an excess of reactive oxygen species (ROS) in the cell (Mruk et al., 2002), and therefore could be considered as candidate genes in this context.

Intertidal benthic invertebrates from the southern Chile shoreline are an attractive system for biogeographical studies of stress and abundance. Intertidal invertebrates typically have broad geographical ranges that may span thousands of kilometers, whereas their longitudinal and tidal height ranges may be restricted to one degree and a few meters, respectively. Such patterns of geographic distribution can be approximately described as one-dimensional domains with two endpoints (rather than a continuous boundary), and provide simplified systems to draw up hypotheses about the responses of organisms to environmental gradients (Sagarin and Somero, 2006). Moreover, intertidal benthic invertebrates have little ability to find refuge during episodes of extreme environmental conditions, and must cope with sharp variations in temperature and salinity due to alternating periods of emersion and immersion associated to tides (Jones et al., 2010).

The mussel Mytilus chilensis is distributed along the Chilean coast and is most abundant in southern shores (Brattström and Johanssen, 1983). Over its wide geographic range of distribution, some individuals inhabit areas with substantial differences in local environmental conditions, such as temperature and freshwater inputs that affect surface salinity, phytoplankton productivity, and marine bacterial loads. Furthermore, anthropogenic factors, such as aquaculture, cause variations in the natural distribution of species due to the displacement of seeds and breeders from one location to another with different environmental conditions. This scenario increases the probability of losing local adaptation due to the reduction of genetic variation at inter-population level (Allendorf et al., 2008). Hence, the aim of this study was to assess the presence of local patterns of gene expression in 5 populations of M. chilensis that span a gradient in oceanographic variability and local environmental conditions. Our genetic assessment was based on the expression analysis of the above-described 7 candidate genes involved in environmental stress (HSP70 and HSP90), antimicrobial response (Mytilin B and Defensin), oxidative stress (Catalase and SOD-CuZn) and iron metabolism response (Ferritin).