ABSTRACT: Global climate change increasingly polarizes environments, presenting unprecedented challenges to many organisms (Smol, 2012). Polarization occurs not only in the spatial dimension, producing greater desert drought and tropical rainfall, for example, but also in the temporal dimension by making a local environment more variable over time. Many organisms survive these fluctuating environmental conditions by manifesting multiple distinct phenotypes through developmental processes that enable phenotypic plasticity (Pigliucci et al., 2006; Parsons et al., 2011). As with early development, these processes are expected to strictly regulate gene expression to canalize phenotype, despite the genetic diversity within populations (Alberch, 1982; Riska, 1986, Pigliucci et al., 1996). For plasticity to evolve, natural selection must act on genes that regulate trait variation, e.g, those conferring norms of reaction to a specific set of conditions. Despite the importance of these reaction norms for coping with environmental challenges, the genetic framework underlying phenotypic plasticity remains poorly defined, making it impossible to study how they function, differ among natural populations, and evolve. Here we used arsenic, a chemical inhibitor of salinity acclimation, to identify genes involved in transforming the gill from its freshwater to its seawater architecture in the euryhaline teleost Fundulus heteroclitus. Linear model interaction terms associated with the combined effect of arsenic and salinity challenge revealed an antagonistic relationship between arsenic exposure and salinity acclimation Exposure to arsenic during salinity acclimation yielded gene expression values similar to those observed in unexposed fish that remained in a stable environment, demonstrating that arsenic prevents changes in gene expression that normally enable osmotic plasticity. The gene sets defined by the interaction terms showed reduced inter-individual variation, suggesting unusually tight control, consistent with the hypothesis that they participate in a canalized developmental response. Evidence that natural selection acts to preserve their canalized gene expression was obtained by referencing three populations that differ in their adaptive tolerance to salinity changes (Whitehead et al., 2011). Specifically, populations adapted to withstand the widest salinity range showed both reduced transcriptional variation in genes enabling gill plasticity and an increased osmoregulatory capacity, highlighted by more stable plasma chloride concentrations in response to an osmotic challenge. Finally, we observed significantly fewer associations between genes underlying trait variation and their transcriptional regulators compared to genes that responded to only arsenic or salinity. Collectively, our results demonstrate that phenotypic plasticity converges on a molecular solution that parallels early development, in which the expression of phenotypic plasticity genes and phenotypes are canalized in part by reducing trans-regulatory complexity. 36 Sample comparisons with fish gills exposed to freshwater, freshwater to seawater for 1 hour, freshwater to seawater for 1 hour with arsenic, freshwater to seawater for 24 hours, freshwater to seawater for 24 hours with arsenic, and freshwater with arsenic for 48 hours
