Multiple Stressors in the Anthropocene: Urban Evolutionary History Modifies Sensitivity to the Toxic Effects of Crude Oil Exposure in Killifish

Abstract

Persistence of wild species in human-altered environments is difficult, in part because challenges to fitness are complex when multiple environmental changes occur simultaneously, which is common in the Anthropocene. This complexity is difficult to conceptualize because the nature of environmental change is often highly context specific. A mechanism-guided approach may help to shape intuition and predictions about complexity; fitness challenges posed by co-occurring stressors with similar mechanisms of action may be less severe than for those with different mechanisms of action. We approach these considerations within the context of ecotoxicology because this field is built upon a rich mechanistic foundation. We hypothesized that evolved resistance to one class of common toxicants would afford resilience to the fitness impacts of another class of common toxicants that shares mechanisms of toxicity. Fundulus killifish populations in urban estuaries have repeatedly evolved resistance to persistent organic pollutants including PCBs. Since PCBs and some of the toxicants that constitute crude oil (e.g., high molecular weight PAHs) exert toxicity through perturbation of AHR signaling, we predicted that PCB-resistant populations would also be resilient to crude oil toxicity. Common garden comparative oil exposure experiments, including killifish populations with different exposure histories, showed that most killifish populations were sensitive to fitness impacts (reproduction and development) caused by oil exposure, but that fish from the PCB-resistant population were insensitive. Population differences in toxic outcomes were not compatible with random-neutral expectations. Transcriptomics revealed that the molecular mechanisms that contributed to population variation in PAH resilience were shared with those that contribute to evolved variation in PCB resilience. We conclude that the fitness challenge posed by environmental pollutants is effectively reduced when those chemicals share mechanisms that affect fitness. Mechanistic considerations may help to scale predictions regarding the fitness challenges posed by stressors that may co-occur in human-altered environments.

Keywords: Fundulus grandis; RNA‐seq; cross‐resistance; crude oil; deepwater horizon oil spill; ecological genomics; evolutionary genomics; toxicogenomics.

FIGURE 1

Adult Fundulus grandis fish were collected from four populations comprising two geographic pairs, from Louisiana (LA) and Texas (TX). Each geographic pair consists of a population from a clean reference site (‐Reference) and a population from a polluted site (‐Polluted). The TX‐Polluted population resides in an urban estuary in the Houston Ship Channel that has been polluted with persistent organic pollutants since at least the 1970's, where resident fish have evolved resistance to the toxic effects of PCBs. The LA‐Polluted site was contaminated with crude oil during the Deepwater Horizon oil spill in 2010. To our knowledge, there is no evidence that these fish have evolved resistance to oil toxicity.

FIGURE 2

Percent of eggs with successful fertilization after adult exposure to clean water and water spiked with crude oil. White and dark filled bars indicate no‐oil control or oil exposure conditions, respectively, grouped by source population. Oil exposure impaired fertilization success by nearly 70% in most populations, except for the PCB‐resistant TX‐Polluted population for which oil exposure did not affect fertilization success.

FIGURE 3

Population variation in toxicity outcomes following oil exposure during development. (A) Exposure to increasing oil concentrations (x‐axis, including no‐oil control “con” and three increasing oil concentrations [low], [medium], and [high]) impaired heart rate (beats per minute; BPM) in a dose‐responsive manner in developing embryos (7 dpf) in all populations except for the PCB‐resistant TX‐Polluted population for which oil exposure did not impair embryonic heart rate. (B–E) Exposure to increasing oil concentrations (x‐axis) induced cardiovascular system deformities in a dose‐responsive manner in developing embryos (10 dpf) in all populations except for the PCB‐resistant TX‐Polluted population (C) for which oil exposure did not induce deformities. Stacked bars indicate proportion of embryos within the treatment that had no observable abnormalities (PA score = 0; light gray), moderate abnormalities (PA score = 1–2; medium gray), or severe abnormalities (PA score = 3–4; dark gray).

FIGURE 4

Principal components analysis (PCA) for genes that varied in expression between populations. We included 5167 genes that varied in expression between populations, which excluded genes that also varied with oil exposure or had significant exposure‐by‐population interactions (p < 0.1). PCA was performed using mean expression values among replicate individuals (n = 5) within each treatment (oil concentration by population).

FIGURE 5

Population‐dependent transcriptional responses to embryonic oil exposure. (A) Heatmap includes genes that had oil exposure effects on transcription, but where those effects varied between populations (512 genes with significant oil exposure‐by‐population interaction, p < 0.05). The four panels, left to right, show oil concentration‐responsive genes for the LA‐Reference, LA‐Polluted, TX‐Reference, and TX‐Polluted populations, respectively. Within each population panel, increasing oil concentrations are organized starting from no‐oil controls (Con) on the left to the highest concentration of oil on the right. Individual genes are the rows. Genes (rows) were hierarchically clustered (Pearson correlation). For each gene within each population, expression was normalized to the control condition (black), where up‐regulation and downregulation relative to control conditions is indicated in yellow and blue, respectively. Color intensity relates to fold‐increase or decrease of log2 expression (see color scale bar). (B) The first two principal components for the 512 genes with significant oil exposure‐by‐population interaction (p < 0.05). Arrows indicate the trajectory of gene expression change with increasing concentration for each population. The base of each arrow represents the no‐oil control condition for that population, where the arrow trajectory tracks gene expression change with increasing oil concentrations, such that the tip of arrow represents the highest oil concentration. Populations share similar gene expression under control conditions, but expression diverges between populations as the concentration of oil exposure increases. The PCB‐resistant TX‐Polluted population shows the most divergent response to oil. The DHOS‐exposed LA‐Polluted population shows a blunted response to oil at the highest oil concentrations compared with the two regional reference populations that showed nearly identical transcriptional responses to oil exposure.

FIGURE 6

Population variation in the transcriptional responses to oil exposure for genes involved in the AHR signaling pathway, which is a key pathway that mediates toxicity. (A) Heatmap includes genes that are components of the AHR signaling pathway. The four panels, left to right, show oil concentration‐responsive genes for the LA‐Reference, LA‐Polluted, TX‐Reference, and TX‐Polluted populations, respectively. Within each population panel, increasing oil concentrations are organized starting from no‐oil controls (Con) on the left to the highest concentration of oil on the right. Individual genes are the rows. Genes (rows) were hierarchically clustered (Pearson correlation). Higher and lower transcript abundance is indicated in yellow and blue, respectively. Color intensity relates to fold‐increase or decrease of log2 expression (see color scale bar). (B) The first two principal components for the AHR‐regulated genes. Arrows indicate the trajectory of gene expression change with increasing concentration for each population. The base of each arrow represents the no‐oil control condition for that population, where the arrow trajectory tracks gene expression change with increasing oil concentrations, such that the tip of arrow represents the highest oil concentration. Populations tend to share similar AHR‐mediated gene expression responses to oil exposure, except for the PCB‐resistant TX‐Polluted population which shows a divergent response to oil. The LA‐Polluted population also shows a blunted response especially at higher oil concentrations.