Gene expression in male and female stickleback from populations with convergent and divergent throat coloration

Abstract

Understanding of genetic mechanisms underlying variation in sexual dichromatism remains limited, especially for carotenoid-based colors. We addressed this knowledge gap in a gene expression study with threespine stickleback. We compared male and female throat tissues across five populations, including two in which female red coloration has evolved convergently. We found that the expression of individual genes, gene ontologies, and coexpression networks associated with red female color within a population differed between California and British Columbia populations, suggesting differences in underlying mechanisms. Comparing females from each of these populations to females from populations dominated by dull females, we again found extensive expression differences. For each population, genes and networks associated with female red color showed the same patterns for males only inconsistently. The functional roles of genes showing correlated expression with female color are unclear within populations, whereas genes highlighted through inter-population comparisons include some previously suggested to function in carotenoid pathways. Among these, the most consistent patterns involved TTC39B (Tetratricopeptide Repeat Domain 39B), which is within a known red coloration QTL in stickleback and implicated in red coloration in other taxa.

Keywords: carotenoid; gene expression; sexual dimorphism; stickleback.

FIGURE 1

Map showing study populations and images of males and females of each population. Sites, images from left to right [sample size]: (a) Anadromous male [4], female [4]; (b) Salmon River male [5], female [4]; (c) Bonsall Creek male [4], female [4]; (d) Little Campbell Stream male [4, but no reflectance data for one], red female [5], dull female [3]; (e) Matadero Creek male [6], red female [4], dull female [4]

FIGURE 2

Red chroma for populations, sexes, and color groupings in this study. Fish classified a priori as red in red, dull in gray. Grouped by population (LC Stream, Little Campbell Stream‐resident; Matadero from California, others British Columbia), sex (M, male; F, female), and color pattern. Sample size = 50, with 3–6 fish in each sub‐category, shown as points in boxplots; boxplots show median, quartiles, ranges

FIGURE 3

DESeq2 PC1 vs. PC2 for all populations. Males: square symbols; females: round symbols; red: Matadero stream (California; all other populations from British Columbia), includes red‐throated females; blue: Anadromous (only males red‐throated); black: Bonsall stream (all lack red throats); orange: Little Campbell Stream‐resident (includes red‐throated females), purple: Salmon River (only males red‐throated)

FIGURE 4

Relationship between red chroma and module eigengene value (first principal component) for “orangered3” in WGCNA analysis of Matadero population, n=14. Females in red, males in dark grey

FIGURE 5

Expression of two genes strongly associated with chroma in females from one or both populations possessing red females, relative to females from populations in which females are mainly dull. Fish classified a priori as red in red, dull in gray. Grouped by population, sex (F = female, M = male), and color pattern (red fill indicates colored red in throat, gray indicates dull). Sample size = 51, with 3–6 fish in each sub‐category; (a) si:ch211‐114n24.6; (b) bco2b

FIGURE 6

Expression of genes strongly correlated with chroma in inter‐population analyses of males and females. Fish classified a priori as red in red, dull in gray. Grouped by population, sex (F = female, M = male). Sample size = 51, with 3–6 fish in each sub‐category. (a) TTC39B; (b) ENSGACG00000010436