Rapid Neural DNA Methylation Responses to Predation Stress in Trinidadian Guppies

Abstract

DNA methylation (DNAm) is a well-studied epigenetic mechanism implicated in environmentally induced phenotypes and phenotypic plasticity. However, few studies investigate the timescale of DNAm shifts. Thus, it is uncertain whether DNAm can change on timescales relevant for rapid phenotypic shifts, such as during the expression of short-term behavioural plasticity. DNAm could be especially reactive in the brain, potentially increasing its relevance for behavioural plasticity. Most research investigating neural changes in methylation has been conducted in mammalian systems, on isolated individuals, and using stressors that are less ecologically relevant, reducing their generalisability to other natural systems. We exposed pairs of male and female Trinidadian guppies (Poecilia reticulata) to alarm cue, conspecific skin extract that reliably induces anti-predator behaviour, or a control cue. Whole-genome bisulphite sequencing on whole brains at various time points following cue exposure (0.5, 1, 4, 24, and 72 h) allowed us to uncover the timescale of neural DNAm responses. Males and females both showed rapid shifts in DNAm in as little as 0.5 h. However, males and females differed in the time course of their responses: both sexes showed a peak in the number of loci showing significant responses at 4 h, but males showed an additional peak at 72 h. We suggest that this finding could be due to the differing longer-term plastic responses between the sexes. This study shows that DNAm can be rapidly induced by an ecologically relevant stressor in fish and suggests that DNAm could be involved in short-term behavioural plasticity.

Keywords: behaviour; epigenetics; fish; plasticity.

FIGURE 1

Change in proportion of (A) substrate use and (B) mating behaviour after cue exposure. Substrate use was measured as the amount of time fish spent in the lower third of the tank minus the time spent foraging. Mating behaviour was measured as the amount of time males spent pursuing females and performing sigmoidal displays. Change in proportion of substrate use and mating effort were calculated by subtracting the proportion before cue exposure from the proportion after cue exposure such that a positive number indicates an increase after cue exposure and a negative number indicates a decrease. Boxplots show the interquartile range with the median indicated, and lines show the maximum and minimum values.

FIGURE 2

Number of identified differentially methylated sites (DMSs) and regions (DMRs) at each time point comparison.

FIGURE 3

Upset plots showing overlap between time points in (A and B) differentially methylated sites (DMSs) and (C and D) regions (DMRs) for females (dark purple) and males (light blue). (A) DMSs in females, (B) DMSs in males, (C) DMRs in females, and (D) DMRs in males. The bottom section of each plot indicates the intersection being shown for each bar with points indicating the time points involved in the overlap. Bars in the top portion of the plots show the size of overlap for each overlap. Bars are ordered from left to right by degree of overlap, thus note the order of time overlaps differs between figures A through D.

FIGURE 4

Heatmaps with cluster results for differentially methylated regions (DMRs) identified at each time point for females. Each row shows the relative methylation of a DMR identified at (A) 0.5 h, (B) 1 h, (C) 4 h, (D) 24 h, and (E) 72 h. Thus, each row represents a different DMR in figures A throughE. Each column is an individual fish. Colour scale indicates the scaled percent methylation levels from high (blue) to low (yellow). Hierarchical clustering with Euclidean distance and Ward's linkage was run on samples, as shown above each heatmap.

FIGURE 5

Distribution of differentially methylated sites (DMSs) and regions (DMRs) identified compared to a null distribution of all CpGs at each time point for females (A) and males (B). Asterisks denote significant differences from the null distribution as tested using G tests.