Evolution of Transient Receptor Potential (TRP) Ion Channels in Antarctic Fishes (Cryonotothenioidea) and Identification of Putative Thermosensors

Abstract

Animals rely on their sensory systems to inform them of ecologically relevant environmental variation. In the Southern Ocean, the thermal environment has remained between -1.9 and 5 °C for 15 Myr, yet we have no knowledge of how an Antarctic marine organism might sense their thermal habitat as we have yet to discover a thermosensitive ion channel that gates (opens/closes) below 10 °C. Here, we investigate the evolutionary dynamics of transient receptor potential (TRP) channels, which are the primary thermosensors in animals, within cryonotothenioid fishes-the dominant fish fauna of the Southern Ocean. We found cryonotothenioids have a similar complement of TRP channels as other teleosts (∼28 genes). Previous work has shown that thermosensitive gating in a given channel is species specific, and multiple channels act together to sense the thermal environment. Therefore, we combined evidence of changes in selective pressure, gene gain/loss dynamics, and the first sensory ganglion transcriptome in this clade to identify the best candidate TRP channels that might have a functional dynamic range relevant for frigid Antarctic temperatures. We concluded that TRPV1a, TRPA1b, and TRPM4 are the likeliest putative thermosensors, and found evidence of diversifying selection at sites across these proteins. We also put forward hypotheses for molecular mechanisms of other cryonotothenioid adaptations, such as reduced skeletal calcium deposition, sensing oxidative stress, and unusual magnesium homeostasis. By completing a comprehensive and unbiased survey of these genes, we lay the groundwork for functional characterization and answering long-standing thermodynamic questions of thermosensitive gating and protein adaptation to low temperatures.

Keywords: Antarctica; TRP channels; cold evolution; notothenioids.

Fig. 1.

Estimated divergence times and phylogenetic relationships between the families of the notothenioidei suborder (notothenioids), with oxygen isotope composition (δ¹⁸O) of benthic foraminifera, which is a proxy for temperature and also ice volume (Shevenell et al. 2004). Gray-shaded area indicates the range in δ¹⁸O values measured. Branch where antifreeze glycoproteins (AFGPs) arose is indicated. Note that Nototheniidae is paraphyletic. Colored boxes indicate epochs. Phylogeny and isotope data adapted from Near et al. (2012) and Near et al. (2015); species numbers from Eastman and Eakin (2021).

Fig. 2.

Maximum likelihood phylogenetic relationships of all predicted protein sequences from existing transcriptomic notothenioid data and reference TRP sequences. Colors indicate TRP channel subfamilies. Notothenioids clusters indicated by asterisks (see Supplementary Material online for tip labels, species, data sources, and sequence accessions). No notothenioid proteins clustered with TRPP or TRPN subfamilies, but all other subfamilies were represented in notothenioid transcriptomic data.

Fig. 3.

Omega (ω) shifts for TRPV1a, TRPM4, TRPM7, and TRPA1b channels comparing cryonotothenioid branches (colors) with all other reference branches (black) generated with RELAX (see Materials and Methods). RELAX classifies codons into three classes (ω₁, ω₂, and ω₃) based on nonsynonymous/synonymous substitution rates (dN/dS or ω) and then compares the shifts in distribution of these classes (arrows) in test branches compared with reference branches. All four of these genes showed evidence of significantly intensified selection (k values > 1; see table 1 for more details). Colors indicate gene subfamilies, as in figure 2.

Fig. 4.

TRPA1 gene tree with synteny. Gene duplications or losses are indicated by asterisks. Four potential independent duplications of TRPA1b (pink asterisks) were found within the cryonotothenioid clade (blue). Some genes were pieced together from several scaffolds (“scattered on various scaffolds”). Node support values are bootstraps. Gray dotted lines indicate connected chromosome or scaffold, arrows indicate gene directionality. TRPA1 orthologs or paralogs are indicated by green arrows.

Fig. 5.

Gene expression data comparing (a, and inset expansion b) TRP channel expression in the Harpagifer antarcticus trigeminal ganglion and whole brain and (c) ranked expression levels in Astatotilapia burtoni and H. antarcticus. Expression levels are in transcripts per million (TPM). Diagonal lines indicate equal expression levels in both comparisons, genes above the line in (a) and (b) have higher expression in the trigeminal ganglion relative to the whole brain or, as in C, higher ranked expression in H. antarcticus compared with A. burtoni. (d) The five TRP channels with significantly higher expression in the H. antarcticus trigeminal ganglion (TG) compared with the whole brain based on counts normalized in DESeq2, adjusted P values shown. Colors indicate gene subfamilies, as in figure 2.

Fig. 6.

Schematics of selected TRP channels with sites under significant positive selection. Black stars indicate location of sites as determined by the location of the homologous site on channel from a species for which a structure has been determined. TRPA1b site location determined from human TRPA1 (UniProtKB-O75762; schematic adapted from Paulsen et al. 2015). TRPV1a site location determined from human TRPV1 (UniProtKB-Q8NER1; schematic adapted from Liao et al. [2013]). TRPM4 site location determined from human TRPM4 (UniProtKB-Q8TD43; schematic is adapted from Autzen et al. [2018]). TRPM7 site location determined from human TRPM7 (UniProtKB-Q96QT4; schematic is adapted from the mouse channel in Duan et al. [2018]). See supplementary figure S7, Supplementary Material online for the alignments of these sites.