KY EPSCoR: Transforming Kentucky's New Economy with EPSCoR: Microevolutionary Response of Two Salamander Species to Climate Change in an Isolated Seasonal Wetland

Microevolutionary response of two salamander species to climate change in an isolated seasonal wetland PROJECT RATIONALE: Climate change can strongly affect natural populations. Through its influence, many species will experience changes in their population demographics, shifts in the seasonal timing of life history characteristics (phenology), and possibly even extinction (Franks et al. 2007; Todd et al. 2010; Lane et al. 2011; Bálint et al. 2011). Global temperatures are projected to increase by several degrees (+1.8 to +4.0C) over the next century and models are needed to predict how species will respond (Meehl et al. 2007); however, little empirical information exists concerning the phenology, demographics, and genetics of species to inform the development of these models. Current changes in species' phenology could serve as an indication of what species are responding to climate change, and what species might persist or decline in response (Cleland et al. 2012). While changes in phenology of individuals allow short- term population persistence, it is unclear what impact this will have on genetic variation, which is the basis for adaptation and long-term survival (Lande & Shannon 1996). The study of responses to environmental change is especially important in species with limited dispersal abilities, such as amphibians, as range shifts may be limited. Findings from a 30-year study monitoring the amphibian community at Rainbow Bay (RB), an ephemeral wetland in eastern South Carolina, found increasing temperature to be positively correlated with shifts in the timing of breeding migrations in some species, negatively correlated in other species, and uncorrelated with phenology in other species (Todd et al. 2010). As a potential compounding effect, increased droughts at RB from 1978-2004 have resulted in decreased hydroperiods and have been shown to be significantly correlated with population declines in amphibian species adapted to long hydroperiods (Daszak et al. 2005). The resolution of phenological responses and population declines related to climate-associated factors is important, but the equally important question of how these changes have impacted demography and the loss or maintenance of genetic variation remains unclear (Franks et al. 2007; Lane et al. 2011). Collectively, these climate-associated factors have impacted the population dynamics of the amphibian community at RB by causing a community dominance shift among species. For example, the salamander Ambystoma opacum has advanced its breeding time by 15.3 days, and the population has steadily expanded in numbers since colonizing RB in 1980. In contrast, A. talpoideum has not shifted its breeding time and its population has steadily declined since 1980. The response to climate change of these two species over time can address many questions related to the ecological genomics of climate change adaptation. Here, we propose research objectives aimed at understanding how phenological and de- mographic responses to climate change impacts the population genetics and evolution of amphib- ian species. Rainbow Bay provides a unique opportunity for this work because amphibian spe- cies have been continuously sampled over the last 30 years, providing a valuable time series across a period of years in which climatological and ecological conditions have changed signifi- cantly. Using restriction-site-associated DNA (RAD) sequencing, we will (1) generate genomic surveys of single nucleotide polymorphisms (SNPs) from a salamander species (A. opacum) that has demonstrated a population expansion at RB and a salamander species (A. talpoideum) that has declined in numbers. For each species, genomic data will be collected from three different time points, representing early, intermediate, and late stages across the last 30 years of climato- logical and ecological change at RB. We will use these data to then (2) assess intraspecific trends in population genetic parameters over time and compare interspecific trends. Finally, we will (3) use the time series of SNP variation within each species to identify loci that may be potentially influenced by selection as a result of climate change. These results will not only be informative at a local scale, but will be useful in beginning to model the genetic responses of amphibian pop- ulations at a regional-scale. MATERIALS AND METHODS Study Site and Species - This study will take place at a 1-ha seasonal wetland, Rainbow Bay, on the US Department of Energy's (DOE) Savannah River Site in eastern South Carolina. Rainbow Bay was completely encircled by a drift-fence in September 1978 and amphibians and reptiles entering and leaving the wetland have since been censused daily (for a full description see Pechman 1991). All captured animals are cohort marked, sexed, and released on the opposite side of the drift fence. Individuals found dead in traps have been preserved at -70C, providing a wealth of genetic samples from numerous amphibian species over 30 years of study For this study, we will focus on the salamanders A. opacum and A. talpoideum, both common amphibians in the southeast U.S. (Petranka 1998) that have been well sampled at RB over time. These species migrate to seasonal wetlands in their respective breeding seasons, deposit eggs either in water (A. talpoideum) or on land (A. opacum), and then migrate back to the forested floodplain. Genetic Sampling - For each species, genomic data will be collected from three different time points, representing early, intermediate, and late stages across the last 30 years of climatological and ecological change at RB. We will sample a minimum of 10 individuals per species for each of the three time points. To assure a large enough sample size at a particular time point, if needed, we will bin individuals from two-consecutive years. DNA Extraction and Genomic Data Generation - Genomic DNA will be extracted using either a Qiagen® DNEasy Blood and Tissue Kit or a standard and phenol-chloroform method. We will use a double-digest RAD sequencing (ddRADSeq) approach, as described in Hohenlohe et al. (2010) and Peterson et al. (2012) to generate genome-wide panels of single nucleotide polymorphism (SNP) data for each individual. RAD-based SNP generation is expected to yield thousands of orthologous SNPs across individuals and will provide genomic estimates of population genetic parameters. Genetic Analyses - Observed (HO) and expected (HE) heterozygosity, and standardized allelic richness with rarefaction will be estimated for each time point in FSTAT v 2.9.3 (Goudet, 1995). Similarly, deviations from Hardy-Weinberg equilibrium and deviations from linkage disequilibrium between all pairs of loci with Bonferroni corrections will be tested using GenePop v 4.0.1 (Raymond and Rousset, 1995). Differences in genetic variation between time points will be evaluated using Independent-Samples Kruskal-Wallis Tests in SPSS v 18.0 (IBM Corporation). To test for genetic differentiation across years, FST (Weir and Cockerham, 1984) will be calculated in Arlequin v 3.5.1.2 (Excoffier et al., 2010). In addition, we will use an approximate Bayesian computation (ABC) approach to test among alternate demographic models for expanding, stable, and declining population sizes using my genomic data (Sousa et al. 2012). Finally, the genome-wide SNP will also provide an opportunity to test for signatures of selection that change across sampled time points. To identify candidate loci under selection, we will use an FST outlier approach, treating the three separate time points as separate populations, in order to compare allele frequency change over time (Beaumont and Nichols 1996). This project relates to the EPSCoR ecological genomics research initiative by utilizing next-generation sequencing and innovative analyses of genomic data to investigate species response to climate change.