Epigenetics underpins phenotypic plasticity of protandrous sex change in fish

Abstract

Phenotypic plasticity is an important driver of species resilience. Often mediated by epigenetic changes, phenotypic plasticity enables individual genotypes to express variable phenotypes in response to environmental change. Barramundi (Lates calcarifer) are a protandrous (male-first) sequential hermaphrodite that exhibits plasticity in length-at-sex change between geographic regions. This plasticity is likely to be mediated by changes in DNA methylation (DNAm), a well-studied epigenetic modification. To investigate the relationships between length, sex, and DNAm in a sequential hermaphrodite, here, we compare DNAm in four conserved vertebrate sex-determining genes in male and female barramundi of differing lengths from three geographic regions of northern Australia. Barramundi first mature as male and later sex change to female upon the attainment of a larger body size; however, a general pattern of increasing female-specific DNAm markers with increasing length was not observed. Significant differences in DNAm between males and females of similar lengths suggest that female-specific DNAm arises rapidly during sex change, rather than gradually with fish growth. The findings also reveal that region-specific differences in length-at-sex change are accompanied by differences in DNAm and are consistent with variability in remotely sensed sea temperature and salinity. Together, these findings provide the first in situ evidence for epigenetically and environmentally mediated sex change in a protandrous hermaphrodite and offer significant insight into the molecular and ecological processes governing the marked and unique plasticity of sex in fish.

Keywords: DNA methylation; ecological adaptation; phenotypic plasticity; sex change; teleost; temperature.

FIGURE 1

(a) Map of Queensland, Australia, showing sampling locations (colored markers) as well as distribution of the genetically distinguishable subpopulations of barramundi (Lates calcarifer) identified by Jerry et al. (; dashed circles). (b) Principal component analysis summarizing genetic diversity among the barramundi sampled and analyzed in the present study, using the same genetic markers as in the aforementioned study. (c) Average austral summer yearly (October through September) salinity and temperature estimated by the Hybrid Coordinate Ocean Model (HYCOM) for positions 8.88 km offshore of the north Queensland east coast (green), mid‐northern GoC (orange) and southern GoC (blue) barramundi catch locations. Data are represented as mean point with standard error bars

FIGURE 2

Comparison of DNA methylation levels between male and female barramundi (Lates calcarifer) from Queensland, Australia (all regions pooled), by amplicon. (a) Box plots demonstrating low methylation of male‐associated amplicons dmrt1 and nr5a2 and high methylation of female‐associated amplicons cyp19a1a and esr1 and in males (blue), and the reciprocal relationship in females (orange). Letters denote significant differences between males and females resulting in the Mann–Whitney tests (p < .001, Table S8). (b) Hypothesized relationship between DNA methylation, length, and sex based on the results in a. (c) Preliminary generalized linear model of the relationship between DNA methylation and length (as total length, anterior to posterior), showing the overall relationship (black line), as well as the relationship for males (blue dashed line) and females (orange dashed line) separately

FIGURE 3

Proportion of DNA methylation in male (left column) and female (right column) barramundi (Lates calcarifer) explained by length (as total length in centimeters) and CpG site (as base pair position) for amplicons of male‐associated amplicons nr5a2 and dmrt1 and female‐associated amplicons cyp19a1a and esr1. Fitted curves correspond to beta regression with logit link for three regions in Queensland, Australia (indicated by color). Curves were evaluated at varying lengths, and the CpG with intercept value closest to the average value is shown. Model: Proportion methylated ~CpG site + region + total length + region: total length with logit link

FIGURE 4

Analysis of length (as total length, anterior to posterior), sex, and age in barramundi (Lates calcarifer) from the north Qld east coast (n = 1247; green), mid‐northern Gulf of Carpentaria (GoC; n = 2319; orange), and southern GoC (n = 3169; blue) >50 cm, Australia, collected between 2000 and 2017. (a) Fitted binomial GLM (solid lines) with standard error (gray shading) showing predicted proportion of females at a given length for each region. Gray dashed and dotted vertical lines indicate previously reported size ranges for sex change in the northern GoC (68–90 cm) and north Qld east coast (85–100 cm), respectively. (b) Length frequency distribution showing the total number of male (light color shades) and female (dark color shades) barramundi of each 10‐cm size class sampled for length, age, and sex in each region. (c) Fitted binomial GLM (solid lines) with standard error (gray shading) showing predicted proportion of females at a given age for each region