Functional mutations in the thyroid-stimulating hormone receptor in natural stickleback populations at sites identical to human disease-causing mutations

Background

Thyroid hormones regulate multiple physiological functions, including metabolism, reproduction, and metamorphosis. Although there are variations in thyroid hormone signaling between populations and species, the causative mutations underlying these variations have rarely been identified. Here, we investigated whether information regarding the causative genes and mutations responsible for human thyroid diseases could assist with the identification of functional mutations in natural stickleback populations, which vary in thyroid hormone signaling between marine and stream-resident ecotypes. We first determined whether Japanese stickleback populations carry mutations at orthologous sites to those carrying non-synonymous mutations causing thyroid diseases in humans and then evaluated their effects using a heterologous mammalian cell line.

Results

We found that several stickleback populations carry non-synonymous mutations in the thyroid-stimulating hormone receptor 2 (Tshr2) gene. Using a heterologous cell culture system and recombinant stickleback thyroid-stimulating hormone (TSH) 1 and TSH2, we first showed that TSHR2 responds to TSH2, but not TSH1. We also found that amino acid changes in TSHR2 at orthologous sites to those at which loss-of-function mutations have been reported in humans similarly reduce TSHR2 function in the stickleback. In contrast, an amino acid change at the site of a gain-of-function mutation in humans increased receptor function. Furthermore, we also found that TSHR1 and TSHR2 are expressed in the throat area and the brain, respectively, suggesting subfunctionalization.

Conclusion

Natural stickleback populations carry functional mutations in a gene involved in thyroid hormone signaling at orthologous sites to those that are responsible for disease in humans. These results suggest that human disease-causing mutations can be informative in the search for functional mutations in natural animal populations.

Hormones regulate diverse aspects of physiology. Therefore, evolutionary changes in hormonal signaling may underlie the adaptive divergence of multiple phenotypic traits [1,2,3]. The identification of the causative mutations that are responsible for naturally occurring variation in hormone signaling pathways will lead to a better understanding of the genetic mechanisms underlying the coordinated evolutionary changes in multiple traits.

Thyroid hormones have multiple physiological functions in animals [4]. First, they regulate metabolism and thermal acclimation. Although less is known about the roles of thyroid hormones in metabolic and thermal regulation in ectothermic vertebrates than in mammals [5], several studies have shown that thyroid hormones alter the metabolic rates of fish [6,7,8]. Second, thyroid hormone signaling pathways regulate seasonal reproduction [9,10,11,12]. Changes in photoperiod alter the expression of thyroid-stimulating hormone (TSH) and the metabolism of thyroid hormones in the pituitary gland and/or hypothalamus, which regulates reproduction. Third, thyroid hormones can induce metamorphosis—the transformation of larvae into juveniles—in several vertebrate species [4, 13,14,15]. Taken together, these findings imply that thyroid hormones mediate ontogenetic transitions to enable animals to cope with changes in environmental conditions and the availability of ecologic resources [4]. Because the environmental conditions and ecological resources of habitats vary, the optimal levels of thyroid hormone signaling may differ between populations in different habitats [4, 6, 16].

Several candidate mutations have been identified that may underlie the variations in thyroid hormone signaling among populations and closely related species. First, sequence variations in the cis-regulatory region of the Tshb gene may underlie interspecies variability in the responsiveness of reproduction to photoperiod in mammals [17]. Second, the Thyroid-stimulating hormone receptor (Tshr) gene shows a signature of selective sweep in the chicken [18]. An amino acid mutation in the Tshr gene may reduce the response of reproduction to photoperiod, and thereby have released their breeding from seasonal dependency during domestication [18, 19]. Third, spring- and autumn-spawning Atlantic herring differ in several amino acids in TSHR, one of which alters the efficiency of cAMP production in a mammalian cell line [20]. Thus, there are documented examples of the TSH–TSHR pathway being a target of selection. However, more examples of mutations that underlie variations in thyroid hormone signaling need to be identified, to better understand how frequently this pathway has changed during adaptive evolution.

In the present study, we searched for functional mutations that alter thyroid hormone signaling in natural stickleback populations. Previous studies have shown that stream-resident threespine sticklebacks (Gasterosteus aculeatus) have lower blood thyroid hormone concentrations than the ancestral marine sticklebacks [6, 21]. A previous common garden experiment revealed that this difference is genetically determined [6]. However, the mutations that are responsible for the heritable reduction in thyroid hormone concentration in stream-resident sticklebacks have yet to be identified. In temperate regions, freshwaters have lower productivity than oceans [22, 23]. Furthermore, some freshwater habitats of sticklebacks are deficient in oxygen [6]. Therefore, it has been hypothesized that hypothyroidism, which reduces the metabolic rate, may be adaptive for survival in freshwater habitats that sometimes become deficient in energy sources and oxygen [6]. Although the roles of thyroid hormone signaling in reproduction have not been well characterized in the stickleback, variation exists in the timing and duration of reproduction among stickleback populations [24], which may be associated with variation in the thyroid hormone signaling pathway. For example, several stream-resident populations inhabiting stable environments have longer breeding seasons than marine populations [24]. The year-round reproduction is similar to that of domesticated animals.

To screen for mutations that might affect thyroid hormone signaling in natural stickleback populations, we took advantage of the accumulating information regarding the causative mutations for human thyroid diseases, such as congenital hypothyroidism [25,26,27,28]. Although such mutations are generally deleterious in humans, alteration of thyroid hormone signaling may be adaptive under certain environmental conditions, as illustrated by the example of stream-resident sticklebacks described above [6]. Here, we first bioinformatically searched for amino acid-changing mutations at the orthologous sites where thyroid disease-causing mutations occur in humans. Next, we experimentally tested the effects of these mutations using a human heterologous cell line, as no stickleback cell lines are yet available.

Screening for candidate functional mutations in thyroid-related genes

We first searched for mutations that cause congenital hypothyroidism in humans using the Human Gene Mutation Database (HGMD) [29]. There were 1125 mutations in 23 genes responsible for congenital hypothyroidism (Table S1) (the free version of the HGMD accessed in April 2021). To identify mutations at the orthologous sites in stickleback orthologs, we downloaded stickleback orthologs using Ensembl (https://ensemblgenomes.org/), except for SLC26A4 (NP_000432), DUOX2 (NP_054799), GLIS3 (NP_689842), and TTF2 (NM_003594.4), which do not have stickleback orthologs. The downloaded amino acid sequences of the human and stickleback orthologs were aligned using MAFFT v7.222 [30].

Next, we checked whether the orthologous sites of human disease-causing mutation sites have any SNPs in the Japanese stickleback populations, using previously identified non-synonymous SNP information [31] available from Dryad (https://doi.org/10.5061/dryad.v8rq0). In Yoshida et al. (2020) [31], non-synonymous mutations were analyzed in a Japanese Pacific Ocean anadromous threespine stickleback (N = 1), five freshwater threespine stickleback populations (N = 3 for the Gifu population and N = 1 for each of the Ono, Nasu, Aizu, and Shiga populations), and a closely related anadromous species, Japan Sea stickleback (Gasterosteus nipponicus) (N = 2). To identify the orthologous amino acid sites between human and stickleback orthologs, we confirmed that the surrounding amino acids of a focal site are conserved using the following three criteria, but the use of three different criteria detected the same sites as candidate functional mutations: i) 70% or more of the 2 upstream and 2 downstream amino acids are conserved, ii) 30% or more of the 1 upstream and 1 downstream amino acids are conserved, and iii) 20% or more of the 5 upstream and 5 downstream amino acids are conserved.

We additionally used the Thyroid Stimulating Hormone Receptor (TSHR) Mutation Database (https://www.tsh-receptor-mutation-database.org/), which includes all published TSHR mutations identified in humans up to 2018 and is more comprehensive regarding the TSHR mutations [32]. We investigated whether any non-synonymous mutations at evolutionarily conserved sites with |PROVEAN scores|> 2.5 [33] or SIFT scores < 0.05 [34] exist at orthologous sites in the Japanese stickleback populations. Amino acid-changing mutations are described following de Dunnen and Antonarakis’s nomenclature [35] with a reference amino acid, position in the protein, and an alternate amino acid: for example, C455W indicates that Cysteine at the 455 position is replaced by Tryptophan.

Since these screening steps were based on one or a few individuals per population, we next investigated whether the identified mutations were fixed or polymorphic within each population. We analyzed the previously determined whole-genome sequences of multiple individuals (N = 8–10 in total; Table 1) except the Shiga population [36,37,38]. Because the WGS data of only one individual was available for the Shiga population, we performed Sanger sequencing of PCR products spanning the target mutation: six individuals collected from Shiori River, Shiga, Japan [39], were used. After genomic DNA isolation with a DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany), we amplified the PCR fragments spanning the substitution site with the following primers using KAPA2G Fast Multiplex (KAPA Biosystems, Woburn, MA, USA) following the manufacturer’s instruction: forward primer, CACATTAACTTTAATCAGTCTTCAGC; reverse primer, ACACCAACCAATGGAAGCAT. The PCR products were purified using the MinElute PCR Purification Kit (Qiagen, Hilden, Germany), followed by Sanger sequencing using the forward primer at Eurofins Genomics (Tokyo, Japan).

Phylogenetic analysis of TSHR and TSHB

The stickleback has two Tshr genes, which we call Tshr1 (ENSGACG00000005233) on Chromosome 18 and Tshr2 (ENSGACG00000018635) on Chromosome 14. To investigate their phylogenetic relationships in ray-finned fishes, we conducted a phylogenetic analysis. The protein sequences of orthologs in ray-finned fishes were obtained using Ensembl (accessed on March 22, 2025). Only protein sequences longer than 500 amino acids were used for subsequent analysis. Duplicated orthologs were removed using seqkit [40]. Human TSHR (ENSP00000298171) was used as an outgroup. We used online MAFFT version 7 (https://mafft.cbrc.jp/alignment/server/index.html) with default settings for multiple sequence alignments and constructing a neighbor-joining tree with 100 bootstraps [41]. After exporting the newick tree file, a phylogenetic tree was drawn using ggtree [42]. To confirm the results of the phylogenetic analysis using the neighbor-joining method [43], we also created a phylogenetic tree of a subset of species using the maximum-likelihood method with IQ-TREE with 100 bootstraps [44]. We conducted the same analysis for two stickleback genes encoding the beta subunit of TSH [6, 45]: Tshb1 (ENSGACG00000005276) on Chromosome 17 and Tshb2 (ENSGACG00000009897) on Chromosome 12. All the obtained Tshb orthologs were longer than 100 amino acids and were used for the analysis. The human TSHB protein (ENSP00000256592) was used as an outgroup.

Cloning and in vitro mutagenesis of TSHR

All the recombinant DNA experiments were approved by the National Institute of Genetics (R6-22, R3-5, 28–16). We cloned the cDNAs encoding TSHR1 and TSHR2 from G. aculeatus and G. nipponicus. Using SuperScript IV Reverse Transcriptase (ThermoFisher, Waltham, MA, USA), oligo dT-primed cDNA was synthesized from total RNA of the pituitary gland of a Pacific Ocean individual collected from Bekanbeushi River, Akkeshi, Hokkaido [46] and a G. nipponicus individual collected from a tidepool in Hamanaka, Hokkaido, Japan [47]. To amplify the Tshr1 and Tshr2 cDNAs, high-fidelity PCR was performed using PrimeSTAR GXL DNA Polymerase (Takara, Shiga, Japan) with the following primers: tshr1_F, GGGACATTTGGGGACATTTA; tshr1_R, ACTTGGCCTGCAAACTGAAG; tshr2_F, CTTGGGAGTCTCCACTTCTGA; tshr2_R: TTACTGTACTTTAACATTTTGGGAT. After A-tailing with ExTaq (Takara, Shiga, Japan), the PCR products were ligated into the pGEM-T easy vector (Promega, Madison, WI, USA). The cloned cDNAs were sequenced with the Sanger method at Eurofins Genomics (Tokyo, Japan), and the obtained cDNA sequences are available at DDBJ (accession numbers LC856560-LC856563).

For in vitro mutagenesis introducing non-synonymous mutations identified in G. nipponicus, the Gifu stream-resident population, and the Ono stream-resident population (Fig. 1A; see the Results), we introduced point mutations into the cloned G. aculeatus Tshr2 gene. In vitro mutagenesis was conducted using Takara PrimeSTAR GXL DNA Polymerase as described previously [48] with the following primers: JS-mut-F, GACTTTTGGATGGGAATATACTTACTG; JS-mut-R, TCCCATCCAAAAGTCTGCAAATGCAAG; Gifu-mut-F, GCAGACCTGAAAATTAGGTTACGACA; Gifu-mut-R, AATTTTCAGGTCTGCCCTCATTGCGT; Ono-mut-F, CAGTTCCAATACCAGCATGGCCAAAC; Ono-mut-R, CTGGTATTGGAACTGCTGGACTGGTG. After mutagenesis, the mutagenized cDNAs were sequenced by the Sanger method at Eurofins Genomics (Tokyo, Japan) to confirm that the expected mutation was introduced without any PCR errors. All cDNAs were inserted into the mammalian expression vector pCAGGS [49].

Multiple protein sequence alignment of the transmembrane (TM) and intracellular (IC) domains of thyroid-stimulating hormone receptor (TSHR). Human TSHR (NM_000369.5), G. aculeatus TSHR1 (LC856560), G. aculeatus TSHR2 (LC856561), G. nipponicus TSHR1 (LC856562), G. nipponicus TSHR2 (LC856563), and Atlantic herring TSHR (ENSCHAP00020046165) were aligned with MAFTT version 7 (https://mafft.cbrc.jp/alignment/server/index.html). The positions of candidate functional mutations identified in the stickleback (this study) and the functional mutation identified in the Atlantic herring [20] are shown by red boxes. Amino acid-altering mutations are described following de Dunnen and Antonarakis [35] with an amino acid of the reference sequence, position in the protein, and the alternate amino acid

Full size image

Purification of recombinant stickleback TSH

TSH is a heterodimer composed of alpha and beta subunits with glycosylation [50]. Previous studies have shown that a recombinant single-chain protein of alpha and beta subunits with a C-terminal glycosylation signal derived from the human chorionic gonadotropin (hCTP) produced in mammalian cells can act on G protein-coupled receptors of fishes [51, 52]. Following these previous studies, we designed fusion proteins: between the stickleback beta-subunit and alpha-subunit, the hCTP, spacer sequences, and 6xHis tag were inserted (Fig. S1). Because we did not know whether either TSH1, TSH2, or both, react on TSHR1 and TSHR2, we designed both TSH1 and TSH2 recombinant proteins (Fig. S1). DNAs encoding the fusion proteins were synthesized by Eurofins Genomics (Tokyo, Japan) after codon optimization for Homo sapiens to increase the translation efficiency of the recombinant proteins in human cell lines. The synthesized DNAs were inserted into the pCAGGS mammalian expression plasmid [49].

For recombinant protein expression, the FreeStyle 293 Expression System (ThermoFisher, Waltham, MA, USA) was used to express proteins in suspension cultures of the human embryonic kidney (HEK) 293 cell line. The 293fectin (ThermoFisher) was used for transfection of the expression plasmids. After 4 days, we collected the cell cultures in 50 ml conical tubes, and the supernatants were collected after centrifugation at 100 × g for 5 min, followed by two additional centrifugation steps at 17,000 × g for 15 min.

Substances smaller than 10 kDa were first removed from the supernatant using an Amicon Ultra-15 10 K ultrafiltration system (Merk Millipore, Burlington, MA., USA). Subsequently, the 6xHis-tagged proteins were purified using the Ni–NTA Agarose (Qiagen, Hilden, Germany), as described previously [53]. The purified proteins were dialyzed against phosphate-buffered saline pH 7.5 without calcium or magnesium using Slide-A-Lyzer Dialysis Cassette, 2 K MWCP (ThermoFisher). Protein quantification was conducted using the Micro BCA Protein Assay Kit (ThermoFisher). We isolated 162 µg of TSH1 and 324 µg of TSH2 from 30 ml of cell culture. The integrity of the purified proteins was checked by Coomassie brilliant blue staining with Takara CBB Protein Safe Stain and western blotting with an HRP-conjugated anti-6xHis monoclonal antibody and Western BLoT Hyper HRP Substrate (Takara, Shiga, Japan) (Fig. S2).

Receptor functional assay

The receptor functional assay was conducted using an adherent HEK293 cell line. HEK293 cells were seeded on a poly-D-lysine-coated 96-well plate at a density of 10,000 cells/well. One day later, the TSHR-expression plasmids were transfected into the cells using the FuGENE HD Transfection Reagent (Promega). One day later, serially diluted TSH ligands were applied to each well with serum-free HE100 medium (Gmep, Kyushu, Japan) and incubated for 3 h. cAMP concentrations in each well were measured using a cAMP Glo Max Assay (Promega) and a plate reader EnSpire (PerkinElmer, Waltham, MA, USA). The experiments were performed in triplicate by TechnoPro (Tokyo, Japan). The effects of the ligands were tested using heteroscedastic two-way ANOVA [54] of the R package twowaytests [55] with the cAMP concentration as a response variable and the plasmid construct and the ligand concentration as independent variables.

Analysis of the expression patterns of TSHR1 and TSHR2 in multiple tissues

To investigate the tissue distribution patterns of TSHR1 and TSHR2, we conducted RNA-seq of multiple tissues from four individuals (2 males and 2 females) of an F₁ hybrid family made by crossing a Japanese Pacific Ocean G. aculeatus female derived from Bekanbeushi River, Hokkaido, Japan [46] and a G. nipponicus male derived from a tidepool in Hamanaka, Hokkaido, Japan [47]. The aim was just to know which tissues express the genes, but not to compare expression levels between tissues or between the sexes. These fish were one and a half year-old adults maintained under a short photoperiod of L:D = 8:16. After euthanasia with 0.5 g/L of ethyl 3-aminobenzoate methanesulfonate (MS-222), we collected the brain without the pituitary gland, the gill, the gut, the kidney, the liver, the trunk muscle, and the throat region containing the thyroid gland and stored them in RNAlater Stabilization Solution (Thermo Fisher Scientific, Waltham, MA, USA) until use. All animal experiments were approved by the Institutional Animal Care and Use Committee of the National Institute of Genetics (R6-18).

RNA was isolated using the RNeasy Mini Kit (Qiagen, Hilden, Germany). RNA-seq libraries were constructed using the NEBNext Ultra II RNA Prep Kit (NEB, Ipswich, MA), as described previously [36, 56]. The libraries were run on a NovaSeq X Plus in 150 bp paired-end mode at Macrogen Japan (Tokyo, Japan). The obtained read counts were 43.7–60.0 M for the brain, 37.8–77.5 M for the gill, 7.3–8.6 M for the gut, 28.1–54.7 M for the kidney, 36.6–56.0 M for the liver, 10.6–14.9 M for the muscle, and 11.3–14.6 M for the throat. These data are available from DDBJ (accession numbers DRR629039-DRR629066).

To calculate the transcripts per million (TPM) of each transcript, we mapped fastq files to the stickleback cds (Gasterosteus_aculeatus.GAculeatus_UGA_version5.cds.all.fa) downloaded from Ensembl [57] using the salmon (version 1.10.3) with the default settings [58]. All five transcripts annotated with the Tshr1 gene (ENSGACG00000005233) were summed: ENSGACT00000006966.2, ENSGACT00000030244.1, ENSGACT00000080222.1, ENSGACT00000030370, and ENSGACT00000006955.2. Only one transcript (ENSGACT00000024683.2) was annotated with Tshr2 (ENSGACG00000018635), whose reads were counted.

Allele-specific expression analysis of Tshr1 and Tshr2

To investigate whether G. aculeatus-derived and G. nipponicus-derived alleles are differentially expressed in the F₁ hybrid, we conducted allele-specific expression analysis of Tshr1 and Tshr2 using the RNA-seq data obtained above. We first made a reference genome with species-specific SNPs being masked to remove a mapping bias, because the stickleback reference sequence is based on G. aculeatus. Comparing the cloned Tshr1 and Tshr2 cDNAs, 15 and 19 nucleotides differed between G. aculeatus and G. nipponicus, respectively. To identify SNPs that are divergent between species but fixed within species, previously published whole-genome sequencing data of 10 G. aculeatus (DRX026609-DRX026618) fish and 10 G. nipponicus fish (DRX0266607, DRX0266608, DRX026619-DRX026626) [37] were first mapped to the stickleback reference genome (Gasterosteus_aculeatus.GAculeatus_UGA_version5.dna.toplevel.fa) downloaded from Ensembl 114 using BWA-mem2 version 2.2.1 [59], followed by visual examination of SNPs with the IGV [60]. We identified 5 and 16 species-diagnostic SNPs for Tshr1 and Tshr2, respectively. These conserved SNP sites were masked as N using maskfasta option of bedtools [61]. RNA-seq fastq files were then mapped to the masked reference genome using STAR (version 2.7.11b) [62] with default parameters. The resulting BAM files were split into two BAM files, each corresponding to one parental species, by SNPsplit (version 0.3.2) [63] using a VCF file containing the information of species-diagnostic SNPs for Tshr1 and Tshr2 described above. We counted the mapped reads on Tshr1 (ENSGACG00000005233) and Tshr2 (ENSGACG00000018635) in each BAM file, using featureCounts (version 2.0.6) [64].

To confirm the results, we also conducted reverse transcription PCR (RT-PCR), followed by cloning and Sanger sequencing. cDNAs were made from 100 ng of total RNA isolated above using random hexamers with SuperScript IV First-Strand cDNA Synthesis Kit (Life Technologies, Carlsbad, USA). PCR was conducted using 1 μl of throat cDNA for Tshr1and brain cDNA for Tshr2 with KAPA2G Fast Multiplex (KAPA Biosystems, Woburn, MA, USA). The primers used were as follows: TSHR1-ASE-F, CTTCGGCGAGACTCTCAAGA; TSHR1-ASE-R, CTCTTCTGCCTCCTCACACA; TSHR2-ASE-F, TTGTTCATCCTCCTGACCAGC; TSHR2-ASE-R, CAGCCACCTTCTGTTTGCC. The PCR products were cloned using the TArget Clone (TOYOBO, Osaka, Japan) and subjected to Sanger sequencing at Eurofins Genomics (Tokyo, Japan). For each fish, 12 clones were sequenced. For each fish, 12 clones were sequenced. In Tshr1, G. aculeatus and G. nipponicus have G and A, respectively, at 2,150,999 bp of Chromosome 18. In Tshr2, G. aculeatus and G. nipponicus have T and G, respectively, at 13,794,026 bp of Chromosome 14.

To compare mRNA expression levels of Tshr2 between pure species, we downloaded previously conducted brain microarray data from Dryad (https://doi.org/10.5061/dryad.40nk2). This microarray expression analysis was conducted on the brains of eight adult G. aculeatus and eight adult G. nipponicus individuals [37]. Normalized expression values of Tshr2 transcript (ENSGACT00000024683) in the brain were available and obtained. Normalization was conducted as follows: different fluorescence signals among arrays were first normalized by the 75th percentile-shift normalization, followed by dividing each value by the median of each probe and log₂-transformation [65]. The thyroid glands are diffusely distributed in the throat of the stickleback [66], so it is difficult to isolate an equal amount of thyroid tissues from different individuals, so we did not compare the gene expression levels of tshr1 in the throat between species.

Candidate functional amino-acid substitutions in the stickleback Tshr2 gene

Our screening using the HGMD database revealed that two stickleback genes, Tshr2 (ENSGACG00000018635) and NK-2 transcription factor related, locus 5 (Nkx2-5) (ENSGACG00000018433), carry non-synonymous mutations at the sites where mutations cause congenital hypothyroidism in humans (Fig. 1 and Fig. S3; Table 1). Taking advantage of an extensive TSHR mutation database, we further expanded our screening for TSHR mutations by conducting an additional search using this database. We identified two additional mutations in Tshr2 in G. nipponicus and the Ono population (Fig. 1; Table 1).

The Nkx2-5 gene is a homeobox-containing gene that plays an important role in the regulation of heart and thyroid development [67,68,69]. In humans, an amino acid-changing mutation at the orthologous site leads to developmental defects in the thyroid gland [70]. All the individuals from all the populations examined were heterozygous at this site (Supplementary Fig. S3B), indicating that gene duplication likely occurred in the stickleback or any common ancestors and that one of the duplicated copies acquired the mutation. Since the focus of this study was on interpopulation variation, we did not pursue this gene further in the present study.

The stickleback has two Tshr genes. A phylogenetic tree of Tshr1 and Tshr2 orthologs revealed that multiple ray-finned fishes have orthologs of both Tshr1 and Tshr2 (Fig. 2A and Fig. S4), suggesting that these two genes are likely products of teleost-specific genome duplication [71, 72]. The Tshr gene carrying a functional mutation in the Atlantic herring [20] belongs to the Tshr1 clade rather than the Tshr2 clade (Fig. 2A and Fig. S4).

Phylogenetic analyses of TSHR and TSHB. A Maximum likelihood phylogenetic tree of TSHR proteins of several ray-finned fish species. Human TSHR (ENSP00000298171) was used as an outgroup. TSHR1 and TSHR2 of the threespine stickleback and TSHR of the Atlantic herring are indicated by arrows. The scale bar indicates 0.05 amino acid substitutions per site. Bootstrap values larger than 50 are shown at each node. Branch tip labels indicate the Ensembl protein IDs, species names, and common names. The clades containing the stickleback TSHR1 and TSHR2 are shown in blue and red, respectively. Residues at the two amino acids in the second transmembrane domain (TM2) highlighted in Fig. 1 are shown on the right side. Two other amino acids highlighted in Fig. 1 are conserved. A neighbor-joining tree using all orthologs of ray-finned fishes downloaded from Ensembl is available in Fig. S4. B Maximum likelihood phylogenetic tree of TSHB proteins of several ray-finned fish species. Human TSHB (ENSP00000256592) was used as an outgroup. TSHB1 and TSHB2 of the threespine stickleback are indicated by arrows. TSHB of the Atlantic herring used for a previous functional assay [20] is also indicated by the arrow. The scale bar indicates 0.1 amino acid substitutions per site. Bootstrap values larger than 50 are shown at each node. Branch tip labels indicate the Ensembl protein IDs, species names, and common names. The clades containing the stickleback TSHB1 and TSHB2 are shown in blue and red, respectively. A neighbor-joining tree using all orthologs of ray-finned fishes downloaded from Ensembl is available in Fig. S5

Full size image

In the Tshr2 gene, amino acid-altering mutations at the sites where G. nipponicus (C455W), the Gifu population (R524L), and the Shiga population (R524W) carry non-synonymous mutations have been reported to be loss-of-function mutations responsible for congenital hypothyroidism in humans [73]. The C455W mutation is located in the second transmembrane domain (TM2 in Fig. 1). This position is just one amino acid after the site where a functional mutation was identified in the Atlantic herring (Fig. 1). The R524L and R524W mutations are located in the second intracellular loop (IC2 in Fig. 1). In contrast, a gain-of-function mutation has been identified in hyperfunctioning thyroid adenomas in humans at the site where the D612N mutation occurs in the Ono population [74], which is located in the third intracellular loop (IC3). The potentially functional mutation identified in the chicken was located in the fourth transmembrane domain [18], where no mutations were found in the sticklebacks examined (Fig. 1). Mutations identified in G. nipponicus, the Gifu, and the Ono populations were fixed within their populations, whereas the mutation in the Shiga population was polymorphic, with four homozygotes and two heterozygotes (Table 1).

Functional alterations in the Tshr2 gene in several stickleback populations

To investigate the functional effects of the mutations in Tshr2, we next conducted a TSHR functional assay. The stickleback has two Tshb genes (Fig. 2B and Fig. S5). Our phylogenetic analysis showed that multiple ray-finned fishes have orthologs of both Tshb1 and Tshb2 (Fig. 2B), suggesting that these two genes are also likely products of teleost-specific genome duplication [71, 72]. Since we did not know how TSH1 and TSH2 differ in their functions or which one acts on TSHR1 and TSHR2, we tested the functions of both TSHR1 and TSHR2 in response to TSH1 and TSH2.

We found that the wild-type TSHR1 of both G. aculeatus and G. nipponicus produced cAMP even in the absence of TSH ligands, but TSH1 further increased cAMP production in the cells expressing TSHR1 (Fig. 3A; heteroscedastic two-way ANOVA, P = 0.001 for the effect of the ligand), with no significant difference between G. aculeatus and G. nipponicus (P = 0.0984). In contrast, TSH1 did not stimulate TSHR2 (Fig. 3A).

Functional assays of TSHR candidate mutations. A TSHR1 responded to TSH1. B Amino acid changes altered TSHR2 functions. TSHR2 carrying the G. nipponicus-specific mutation (C455W) and the Gifu-specific mutation (R524L) did not respond to TSH2, whereas the wild-type G. aculeatus TSHR2 responded. TSHR2 carrying the Ono-specific mutation (D612N) increased cAMP more than the wild-type G. aculeatus TSHR2. Cells transfected with empty vectors responded to neither TSH1 nor TSH2

Full size image

TSH2 did not stimulate TSHR1 of either species (Fig. 3B) (heteroscedastic two-way ANOVA, P = 0.2466 for the effect of the ligand). In contrast, TSH2 stimulated the wild-type TSHR2 of G. aculeatus (Fig. 3B). Interestingly, TSHR2 of G. nipponicus did not respond to TSH2 (Fig. 3B). TSHR2 with point mutations identified in G. nipponicus and the Gifu population also did not respond to TSH2 (Fig. 3B). These results are consistent with the finding that mutations at these two sites cause a loss of function in humans [73]. In contrast, TSHR2 with the Ono-type mutation increased cAMP more than the wild-type TSHR2 (Fig. 3B), which is also consistent with the finding that a mutation at this site is a gain of function in humans [74].

Different tissue distributions of TSHR1 and TSHR2

Since TSHR1 and TSHR2 differed in ligand specificity, we hypothesized that TSHR1 and TSHR2 serve different physiological functions in the stickleback. Although both ligands, TSH1 and TSH2, are expressed in the pituitary [6, 45, 75], we did not know which tissues express their receptors. By RNA-seq, we have found that TSHR1 is expressed mainly in the throat region containing the thyroid gland, whereas TSHR2 is expressed only in the brain (Fig. 4A). These expression patterns are consistent with the idea that these two receptors exert different physiological functions.

Different tissue distributions of the Tshr1 and Tshr2 genes. A Boxplots of transcripts per million (TPM) in F₁ hybrids between a Japanese Pacific Ocean G. aculeatus female and a G. nipponicus male

Full size image

Using the interspecies hybrid, we investigated which species’ alleles are expressed at higher levels. Using the RNA-seq data, we found a trend of G. nipponicus Tshr2 allele expressed at higher levels than G. aculeatus allele in the brain, while no apparent difference was found between alleles for Tshr1 in the throat (Table S2). Because read counts of these receptors were low in RNA-seq, we confirmed this result by sequencing the RT-PCR products (Fig. S6A and Table S2). Brain microarray analysis of pure species showed that Tshr2 was expressed at higher levels in G. nipponicus than in G. aculeatus (Fig. S6B).

Functional mutations in the stickleback TSHR2

Using information on human genetic diseases, we could identify naturally occurring functional mutations in the stickleback Tshr2 gene. Because Tshr2 is expressed in the brain but not in the throat region containing the thyroid gland (Fig. 4A), its ligand, TSH2, appears to act on the brain rather than the thyroid gland, possibly regulating seasonal reproduction. The role of TSH2 in seasonal reproduction is consistent with previous findings that Tshb2, which encodes the beta subunit of TSH2, shows photoperiod-dependent changes in gene expression in the stickleback pituitary [6, 45].

Previous studies have shown that the Gifu population exhibits nearly year-round breeding [76], suggesting that a loss-of-function mutation in TSHR2 may be involved in photoperiod-independent reproduction. However, the Ono population also shows nearly year-round breeding [77], despite the presence of a gain-of-function mutation in TSHR2. Therefore, other mutations may be responsible for its year-round breeding. Alternatively, both constitutively active and inactive TSHR2 may lead to year-round breeding. Additionally, G. nipponicus exhibits seasonal migration and has a limited breeding season [46, 47, 78], even though it carries a loss-of-function mutation in TSHR2. There may be compensatory mutations in non-coding regions of the Tshr2 gene or other genes that restore TSHR2 function. For example, we found that gene expression levels increased in G. nipponicus (Fig. 4B and C). Alternatively, other environmental cues, such as temperature [24], may play a more significant role in inducing reproduction in G. nipponicus than photoperiodic changes. Thus, we do not yet know how these identified mutations are associated with the phenotype. Furthermore, because we do not know the blood circulating levels of the ligand, we are not sure that the concentrations tested in this study are within the physiological range. Genome editing to alter a single nucleotide is now possible in fish [79], which will help answer this question.

Functional divergence between TSHR1 and TSHR2

In contrast to TSHR2, TSHR1 responds to TSH1 and is expressed mainly in the throat region containing the thyroid gland. Therefore, the TSH1-TSHR1 pathway may be involved in regulating blood-circulating thyroid hormone levels. Since we did not find any functional amino acid-changing mutations in TSHR1, hypothyroidism in the stream sticklebacks is caused by other mutations likely in other genes. Linkage mapping will help identify the genetic basis of variations in thyroid hormone levels.

In the present study, TSHR1 produced cAMP without TSH1 to some extent. Since several wild-type G protein-coupled receptors are reported to be constitutively active [80], this is not surprising. However, we cannot exclude the possibility that the constitutive activity of stickleback TSHR1 is due to the heterologous expression system. For example, certain substances released from human cells may act on the stickleback TSHR1, or some inhibitory molecules present in the stickleback may be absent in human cells. Several other caveats also exist in the use of heterologous expression systems, such as overexpression and/or misfolding of the gene product [81].

In the Atlantic herring, alleles associated with the timing of reproduction are found in the TSHR belonging to the TSHR1 clade (Fig. 2), and no other paralogs belonging to the TSHR2 clade are present in the Atlantic herring genome (Fig. 2). The Atlantic herring has two TSHB paralogs, and a previous functional assay [20] used a paralog from the TSHB1 clade (Fig. 2). In the stickleback, TSHB2 and TSHR2, rather than TSH1 and TSHR1, appear to be involved in seasonal reproduction. Thus, the two paralogs of Tshb may have diverged in function in different ways between the Atlantic herring and the stickleback. In mice, the TSH molecules that act on the hypothalamus and the thyroid are differentially glycosylated, exerting different functions [82]. Further studies are needed to understand how the multiple functions exerted by thyroid hormones are coupled or decoupled at the molecular and physiological levels in different species.

Information on human genetic diseases helps identify functional mutations in natural animal populations

Some natural animal populations exhibit phenotypes similar to the symptoms observed in human diseases, and these phenotypes may be adaptive under certain ecological and environmental conditions [83,84,85,86]. Even within humans, some phenotypes that are deleterious in modern society may have been adaptive in the past [87,88,89,90,91]. Our study demonstrates that the causative genes and mutations responsible for human genetic diseases may be good candidates for the genes and mutations underlying similar phenotypic variation observed in natural animal populations.

With the advent of genome sequencing technologies, it has become increasingly easy to conduct whole-genome sequencing of natural populations and identify nucleotide sequence variants among populations. One of the important and challenging next steps is to identify functionally important mutations among these numerous variants and test their functional effects [92]. To this end, we can take advantage of the large amount of information on causative mutations of human genetic diseases. Future studies linking human genetic diseases and naturally occurring phenotypic variations in non-human animals will be useful not only for identifying causative mutations of phenotypic variations in non-human animals but also for better understanding the environmental conditions that may mitigate human disease symptoms.

The TSHR cDNA sequences are available at DDBJ (accession numbers LC856560-LC856563). Raw short read sequences are also available from DDBJ (accession numbers DRR629039-DRR629066). Non-synonymous SNP information is available from Dryad (https://doi.org/10.5061/dryad.v8rq0).

TSH:

Thyroid-stimulating hormone

TSHR:

Thyroid-stimulating hormone receptor

HGMD:

Human Gene Mutation Database

Nkx2-5:

NK-2 transcription factor related, locus 5

We thank Shinji Kanda and all members of the Kitano Lab for discussion and Haruka Yamazaki for help with allele-specific expression analysis.

This research was supported by JSPS Kakenhi (22H04983) and JST CREST (JPMJCR20S2) (J.K.).

Authors and Affiliations

1. Ecological Genetics Laboratory, National Institute of Genetics, Yata 1111, Mishima, Shizuoka, 411-8540, Japan

   Jun Kitano, Mana Sato, Hiyu Kanbe, Genta Okude & Asano Ishikawa

2. The Graduate University for Advanced Studies, SOKENDAI, Mishima, Shizuoka, Japan

   Jun Kitano, Hiyu Kanbe & Asano Ishikawa

3. Present address: Division of Applied Biosciences, Graduate School of Agriculture, Kyoto University, Kyoto, Japan

   Genta Okude

4. Present address: Graduate School of Frontier Sciences, The University of Tokyo, Kashiwanoha, Chiba, Japan

   Asano Ishikawa

5. Minamiizu Field Station, Fisheries Technology Institute, Japan Fisheries Research and Education Agency, Kamo, Shizuoka, Japan

   Yukinori Kazeto

6. Department of Molecular and Chemical Life Sciences, Graduate School of Life Sciences, Tohoku University, Sendai, Japan

   Takashi Makino

Contributions

Conception and Design of the Study (J.K., A.I., Y.K., T.M.); Data Collection and Analysis (J.K., M.S., H.K., G.O., T.M.), Interpretation of Results (J.K., T.M.), Drafting the Manuscript (J.K.), Critical Revision and Editing (J.K., M.S., H.K., G.O., Y.K., T.M.); Final Approval of the Version to be Published (J.K., M.S., H.K., G.O., A.I., Y.K., T.M.); Funding Acquisition (J.K.).

Corresponding author

Correspondence to Jun Kitano.

Ethics approval and consent to participate

All animal experiments were approved by the Institutional Animal Care and Use Committee of the National Institute of Genetics (R6-18). All the recombinant DNA experiments were approved by the National Institute of Genetics (R6-22, R3-5, 28–16).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/.

Reprints and permissions