Expansion and Functional Diversification of Long-Wavelength-Sensitive Opsin in Anabantoid Fishes

Abstract

Gene duplication is one of the most important sources of novel genotypic diversity and the subsequent evolution of phenotypic diversity. Determining the evolutionary history and functional changes of duplicated genes is crucial for a comprehensive understanding of adaptive evolution. The evolutionary history of visual opsin genes is very dynamic, with repeated duplication events followed by sub- or neofunctionalization. While duplication of the green-sensitive opsins rh2 is common in teleost fish, fewer cases of multiple duplication events of the red-sensitive opsin lws are known. In this study, we investigate the visual opsin gene repertoire of the anabantoid fishes, focusing on the five lws opsin genes found in the genus Betta. We determine the evolutionary history of the lws opsin gene by taking advantage of whole-genome sequences of nine anabantoid species, including the newly assembled genome of Betta imbellis. Our results show that at least two independent duplications of lws occurred in the Betta lineage. The analysis of amino acid sequences of the lws paralogs of Betta revealed high levels of diversification in four of the seven transmembrane regions of the lws protein. Amino acid substitutions at two key-tuning sites are predicted to lead to differentiation of absorption maxima (λ_max) between the paralogs within Betta. Finally, eye transcriptomics of B. splendens at different developmental stages revealed expression shifts between paralogs for all cone opsin classes. The lws genes are expressed according to their relative position in the lws opsin cluster throughout ontogeny. We conclude that temporal collinearity of lws expression might have facilitated subfunctionalization of lws in Betta and teleost opsins in general.

Keywords: Gene duplication; Gene expression; Opsin gene; Siamese fighting fish.

Fig. 1

Evolutionary history of opsin genes in the Anabantiformes. A A cladogram depicting the relationship of the anabantiform species and outgroups used in this study with illustration of opsin genes found in the genome of each taxon. Species of the genus Betta are depicted as a single taxon as there is no evidence of opsin gene number variation within the genus. Physical linkage among opsin genes is depicted using connecting lines. The orientation of the opsin genes to neighboring genes is shown by the shape of the illustrated genes in a 5’—3’ direction. B A phylogenetic reconstruction of the coding sequences of all lws paralogs suggests that lws gene expansion occurred in a common ancestor of the genus Betta

Fig. 2

Gene conversion masks the evolutionary relationship of lws paralogs in Abantiformes. A and B Two trees constructed from two fragments bound by the breakpoints identified by GARD show a topology consistent with a common duplication event in the ancestor of all anabantoid fishes (dark green arrow), followed by gene conversion and at least two, possibly three duplication events in the genus Betta (light green arrows). C Pairwise synonymous substitution rates (Ks) between lws1a and lws1b in four genera of anabantoid fishes. Decreases in Ks to zero in different regions of the coding sequence of lws1 paralogs suggest gene conversion. D A phylogenetic reconstruction based only on fragments of the coding region where Ks is relatively high in all species (concatenation of 0–150 and 600–700 bp) is also consistent with a single duplication event in the ancestor of Anabantiformes (dark green arrow) followed by three duplications in the genus Betta (light green arrows) (Color figure online)

Fig. 3

Synteny dot plots of genomic region harboring lws1a and lws1b. Synteny between A C. argus and H. temminckii, B H. temminckii and A. testudineus, C C. argus and A. testudineus, D B. splendens and A. testudineus, E B. splendens and C. argus, and F B. splendens and H. temminckii. Each black dot indicates a 25 bp window having a sequence identity of 65% or higher between the two compared genomes. Gray boxes indicate the positions of exons of the two lws paralogs

Fig. 4

Sliding window analysis of LWS paralogs amino acid sequence identity. Each red line represents A the mean divergence between one LWS paralog of all Betta species and LWS sequences of a set of representative teleost species (see table S2) or B the mean divergence between one LWS paralog of all Betta species and the remaining four paralogs in Betta. Gray areas indicate the position of the seven transmembrane domains of LWS (labeled on top from TM1–TM7). Dark vertical bars indicate the positions of LWS key-tuning sites (based on Yokoyama et al. 2008b) (Color figure online)

Fig. 5

Ontogenetic change of relative expression of different opsin paralogs in B. splendens. A Stacked bars plot showing the sum of relative expression of all paralogs of each opsin class at each of the eight studied ontogenetic stages. Bars in panels B–D represent the relative expression of each paralog of lws (B), rh2a (C), and sws2 (D) at each studied ontogenetic stage. Error bars at stages 120 and 270 indicate standard deviation. dpf = days post-fertilization